Crystalline Molecular Complexes and Compounds: Structure and Principles 2 Volume Set (International Union of Crystallography Monographs on Crystallography)

INTERNATIONAL UNION OF CRYSTALLOGRAPHY BOOK SERIES IUCr BOOK SERIES COMMITTEE E. N. Baker, New Zealand J. Bernstein, I...

Author: Frank H. Herbstein

37 downloads 1204 Views 17MB 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

INTERNATIONAL UNION OF CRYSTALLOGRAPHY BOOK SERIES

IUCr BOOK SERIES COMMITTEE E. N. Baker, New Zealand J. Bernstein, Israel P. Coppens, USA G. R. Desiraju, India E. Dodson, UK A. M. Glazer, UK J. R. Helliwell, UK P. Paufler, Germany H. Schenk (Chairman), The Netherlands IUCr Monographs on Crystallography 1 Accurate molecular structures A. Domenicano, I. Hargittai, editors 2 P.P. Ewald and his dynamical theory of X-ray diffraction D.W.J. Cruickshank, H.J. Juretschke, N. Kato, editors 3 Electron diffraction techniques, Vol. 1 J.M. Cowley, editor 4 Electron diffraction techniques, Vol. 2 J.M. Cowley, editor 5 The Rietveld method R.A. Young, editor 6 Introduction to crystallographic statistics U. Shmueli, G.H. Weiss 7 Crystallographic instrumentation L.A. Aslanov, G.V. Fetisov, J.A.K. Howard 8 Direct phasing in crystallography C. Giacovazzo 9 The weak hydrogen bond G.R. Desiraju, T. Steiner 10 Defect and microstructure analysis by diffraction R.L. Snyder, J. Fiala and H.J. Bunge 11 Dynamical theory of X-ray diffraction A. Authier 12 The chemical bond in inorganic chemistry I.D. Brown 13 Structure determination from powder diffraction data W.I.F. David, K. Shankland, L.B. McCusker, Ch. Baerlocher, editors 14 Polymorphism in molecular crystals J. Bernstein

15 16 17 18

Crystallography of modular materials G. Ferraris, E. Makovicky, S. Merlino Diffuse x-ray scattering and models of disorder T.R. Welberry Crystallography of the polymethylene chain: an inquiry into the structure of waxes D.L. Dorset Crystalline molecular complexes and compounds: structures and principles F.H. Herbstein

IUCr Texts on Crystallography 1 The solid state A. Guinier, R. Julien 4 X-ray charge densities and chemical bonding P. Coppens 5 The basics of crystallography and diffraction, second edition C. Hammond 6 Crystal structure analysis: principles and practice W. Clegg, editor 7 Fundamentals of crystallography, second edition C. Giacovazzo, editor

Crystalline Molecular Complexes and Compounds Structures and Principles Volume 1

F RA N K H . H E R B S T E I N Emeritus Professor of Chemistry, Technion-Israel Institute of Technology, Israel

1

1

Great Clarendon Street, Oxford OX2 6DP Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York # Oxford University Press 2005 The moral rights of the author have been asserted Database right Oxford University Press (maker)

First published 2005 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose the same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Data available Typeset by Newgen Imaging Systems (P) Ltd., Chennai, India Printed in Great Britain on acid-free paper by Biddles Ltd., King’s Lynn ISBN 0–19–856893–2 (Vol 1) 978–0–19–856893–3 ISBN 0–19–856894–0 (Vol 2) 978–0–19–856894–0 ISBN 0–19–852660–1 (Set) 978–0–19–852660–5 10 9 8 7 6 5 4 3 2 1

TWENTY: Give up learning and put an end to your troubles. From TAO TE CHING by Lao Tsu (A new translation by Gia-fu Feng and Jane English Wildwood House, London 1973)

OR How charming is divine philosophy, Not harsh and crabbed as dull fools suppose, But musical as Apollo’s lute, And a perpetual feast of nectared sweets, Where no crude surfeit reigns, John Milton (1608–74): Comus, 476.

This book has been written in many places and over too many years. It would never have been completed without the help and support of my wife Any. It is dedicated to her and to four teachers and friends: R. W. James, FRS, formerly Professor of Physics in the University of Cape Town G. M. J. Schmidt, formerly Professor of Chemistry in the Weizmann Institute of Science J. D. Dunitz FRS, emeritus Professor of Chemical Crystallography in the Eidgeno¨ssische Technische Hochschule, Zu¨rich Sir Aaron Klug P-PRS, OM, NL. MRC Laboratory of Molecular Biology, Cambridge, UK

This page intentionally left blank

Preface My intention is to give an account of the structure and properties of crystalline binary adducts, perhaps better known as molecular compounds and complexes, which are a broad group of materials whose several members are of interest to chemists (e.g. separations via crown ethers and identifications via charge transfer compounds), physicists (e.g. high-conductivity organics), biologists (is not DNA an excellent example of a hydrogenbonded molecular compound?) and technologists (zeolites for separations and as catalysts). I have tried to cater to them all; extensive inclusion of chemical formulae (for the nonchemists) and stereodiagrams (for the noncrystallographers) will hopefully make it easier to assimilate some of the unavoidable complexities. Most emphasis will be given to geometrical structures derived from crystal structure analyses for here lies the bulk of the available information. I refer to interactions between the components wherever this is possible, including both thermodynamic and electronic aspects. Consideration of the relation between structure and properties will be principally confined to the solid state and the implications of solid state results for understanding chemical reactivity in other phases will not be pursued. I restrict myself to crystalline materials because the results, and their meanings, are least unequivocal for this state of matter. Interactions between components in the fluid states are undoubtedly important but I leave these aspects to others. The word ‘‘molecular’’ appears in the title because most of the relevant materials are indeed molecular, but many contain charged entities and I have licensed myself to include what seems relevant, regardless of the formal restrictions of the title. Most of the substances considered are organic, some are inorganic and many have both organic and inorganic parts. I discuss representatives of many of the various types of molecular compound and complex, but I early realized that any attempt to cover all examples of all types would be self-defeating. The problems of choice beset us all, at all levels of our lives, and an author, struggling to compress the vast expanses of knowledge into a practical physical confine, is no exception. I have given preference to mature areas, where what is known presents a model for the treatment of those regions as yet unexplored, and I have emphasized the fundamentals – structure and thermodynamics. I guess (no other word seems realistic) that I have managed to include about 20% of what is available in the literature. The series Inclusion Complexes, dealing with less than half the topics covered here runs to five volumes and over two thousand pages. The series Comprehensive Supramolecular Chemistry, again with half the present coverage, stretches to eleven volumes and five thousand pages. Perhaps the most serious of my many acts of omission is exclusion of material on ‘‘Zeolites’’. This is not because of any lack of importance of the subject but because it is adequately covered in a number of books and an on-going journal. There is also little said about complexes between large biomolecules – this would have required a separate book.

viii

PREFACE

An overall theory is hardly possible but some areas have had sophisticated theories applied to them, e.g. the quantum-mechanical treatments of charge transfer interactions and the statistical mechanical treatments of some phase diagrams. In order to provide some unifying factors, I have given special emphasis to structural relationships and the classification scheme used is structural rather than chemical in nature. The classification scheme proposed here should not be regarded as more than a convenient framework – Nature is too complex and subtle for the imposition of straitjackets. Titles of books and review articles have been included in the (close to 4000) references, which are attached to each chapter for the convenience of the reader. Much of the information is conveyed through tables, some 200 in number. The tabulated material shows that considerable systematization is possible, but also that considerable variety remains as exceptions to those rules that I (and many others) have succeeded in developing. Crystal packing and other diagrams (some 600) are a challenge to author and reader alike. Colour would have helped but was ruled out as impractical. I am most grateful to authors, editors and publishers for permission to use published material. Some more detailed acknowledgments are made in the text. My thanks go to many friends and to Caltech, Northwestern, Cambridge (U.K.), The Royal Institution, the Universities of Cape Town and Witwatersrand and, last but not least, Technion for help and facilities. Needless to say, the responsibility for the contents is entirely mine. Haifa, November 2004

Frank H. Herbstein

Contents Volume 1 PART I

SOME PRELIMINARIES

1 Structural principles in the classification of binary adducts 1.1 Introduction 1.2 Structural classification of binary adducts 1.2.1 General considerations 1.2.2 Molecular complexes 1.2.2.1 Inclusion complexes 1.2.2.2 Moieties within molecules 1.2.2.3 Frameworks with guest participation and/or linkage 1.2.2.4 Segregated stack charge transfer complexes 1.2.2.5 Packing complexes 1.2.3 Molecular compounds 1.3 Other classifications 1.4 How many binary adducts are there? 1.5 Organic and inorganic supramolecular chemistry References

3 4 5 5 6 6 9 9 9 10 10 10 11 12 12

2 Historical outline References

15 19

PART II

MOIETIES WITHIN MOLECULES

Introduction to Part II 3 The enclosure species – crown ethers, cryptands and related molecules – as hosts 3.1 Introduction 3.2 Doubly bridged cyclophanes and analogous molecules as hosts for intramolecular guests 3.3 Cleft molecules as hosts 3.3.1 Single-cleft hosts 3.3.2 Double-cleft hosts 3.4 Container molecules as hosts 3.4.1 Introduction

27 28 30 44 44 47 48 48

x

CONT ENTS

3.4.2 Cavitands and caviplexes 3.4.3 Hemispherands and hemispheraplexes 3.4.4 Triply bridged cyclophanes and analogous molecules as three-dimensional hosts for intramolecular guests 3.4.5 Spherands and spheraplexes 3.4.6 Carcerands and carceplexes 3.5 Hemicarcerands and hemicarceplexes 3.5.1 Overview 3.5.2 The taming of cyclobutadiene, and of o-benzyne 3.5.3 Molecular mechanics and dynamics studies on the complexation and decomplexation processes 3.6 Comparisons of concepts References

48 50

4 Cyclodextrins, and some analogs, as hosts 4.1 Introduction 4.2 -Cyclodextrins as host 4.2.1 -Cyclodextrin as host in clathrate inclusion complexes 4.2.2 -Cyclodextrin as host in tunnel inclusion complexes 4.2.3 Chemically modified -cyclodextrins as hosts in inclusion complexes 4.3 -Cyclodextrins as host 4.3.1 -Cyclodextrin as host in clathrate inclusion complexes 4.3.2 -Cyclodextrin as host in tunnel inclusion complexes 4.3.3 Exceptional -cyclodextrin structures 4.3.4 Chemically modified -cyclodextrins as hosts in inclusion complexes 4.4 Rotaxanes and catenanes of cyclodextrins 4.5 -Cyclodextrins as host 4.5.1 -Cyclodextrin as host in clathrate inclusion complexes 4.5.2 -Cyclodextrin as host in tunnel inclusion complexes 4.5.3 Chemically modified -cyclodextrins as hosts in inclusion complexes 4.6 Larger cyclodextrins 4.7 Cyclic oligosaccharides as cyclodextrin analogs References

73 74 79 80 84

5 Crystal chemistry of some DNA oligonucleotides and their complexes 5.1 Introduction 5.2 Fundamentals of oligonucleotide structure 5.2.1 General aspects 5.2.2 Single crystal x-ray diffraction studies of oligonucleotides 5.3 Crystal chemistry of oligonucleotides and oligonucleotide-guest structures 5.3.1 Polymorphism, isomorphism, and heteromorphism

51 59 59 61 61 64 66 67 68

90 95 97 100 114 114 117 118 118 119 122 123 123 124

133 134 136 136 140 142 142

CONTENTS

5.4

5.5 5.6

5.7

5.3.2 Phase rule relationships 5.3.3 Applications of these concepts Intercalated hexanucleotide-drug complexes with B-DNA structures 5.4.1 The anthracycline drugs 5.4.2 Nogalamycin and derivatives 5.4.3 The 9-aminoacridine drugs 5.4.4 Native hexanucleotides and comparison of crystal structures Isomorphism and polymorphism of A-DNA octanucleotides and the binding of spermine 5.5.1 Octameric oligonucleotides Minor groove binders 5.6.1 Drug molecules that enter the minor groove 5.6.2 Decameric oligonucleotides 5.6.3 Polymorphs or intermediate phases? An example from the decanucleotides 5.6.4 Dodecameric oligonucleotides General survey of the crystal chemistry of oligonucleotide and oligonucleotide-drug complexes References

PART III

xi

143 144 145 145 151 154 156 158 158 167 167 171 178 183 187 189

HOST–GUEST INCLUSION COMPLEXES

Introduction to Part III 6 Tunnel inclusion complexes formed by hosts of lesser versatility 6.1 Introduction 6.2 Tunnel inclusion complexes with directionally bonded hosts 6.2.1 Urea, thiourea and selenourea as hosts 6.2.1.1 Introduction 6.2.1.2 Types of guest in hexagonal urea inclusion complexes 6.2.1.3 Guests which give rhombohedral urea inclusion complexes 6.2.1.4 Guests which give rhombohedral thiourea inclusion complexes 6.2.1.5 Hermann’s comprehensive structural model 6.2.1.6 Diffraction patterns from tunnel inclusion complexes 6.2.1.7 Hexagonal urea tunnel inclusion complexes 6.2.1.8 Determination of guest molecule conformation from diffuse x-ray scattering

203 204 206 206 206 207 208 209 210 212 215 218

CONT ENTS

xii

6.2.1.9

Variation of structure with temperature, with particular reference to {3(urea)[1/4(n-hexadecane)]} 6.2.1.10 Interruption of urea framework by host–guest hydrogen bonding 6.2.1.11 Rhombohedral urea, thiourea and selenourea tunnel inclusion complexes 6.2.1.12 Monoclinic complexes derived from the rhombohedral complexes 6.2.1.13 Behavior of some rhombohedral inclusion complexes on cooling 6.2.1.14 The orthorhombic Type 4 urea tunnel inclusion complexes 6.2.1.15 The hypothetical Type 5 orthorhombic tunnel inclusion complexes 6.2.1.16 The crystal structure of selenourea and its relation to the structures of its tunnel inclusion complexes 6.2.1.17 Thermodynamics of the formation of the tunnel inclusion complexes 6.2.2 The Bishop–Dance hosts – exo-2,exo-6-dihydroxy-2,6dimethylbicyclo[3.3.1]nonane and analogs 6.2.2.1 Introduction 6.2.2.2 The helical tubuland structures 6.2.2.3 The ellipsoidal tetragonal clathrate complexes of some Bishop–Dance hosts 6.2.2.4 Derived structures 6.2.3 Ta4P4S29 – an inorganic framework containing sulphur chains 6.2.4 The tunnel hydrates 6.2.4.1 Tunnel hydrates with several water molecules per tunnel cross-section 6.2.4.2 Tunnel hydrates with one water molecule per tunnel cross-section 6.3 Tunnel inclusion complexes with van der Waals bonded hosts 6.3.1 Tunnel inclusion and other complexes of deoxycholic acid and related compounds 6.3.2 Substituted spirocyclophosphazenes as hosts 6.3.3 Tritriptycene – a C62H38 hydrocarbon of D3h symmetry with three U-shaped bays 6.3.4 trans-anti-trans-anti-trans-Perhydrotriphenylene as host 6.3.5 N-(p-tolyl)tetrachloro-phthalimide as host 6.4 Comparison of the various types of tunnel inclusion complexes References 7 Clathrate inclusion complexes formed by hosts of lesser versatility 7.1 Introduction 7.2 Directionally bonded hosts

219 227 231 235 236 245 245 247 247 251 251 251 264 267 268 269 269 271 272 272 291 297 298 307 310 311 321 323 323

CONTENTS

7.2.1 Quinol (hydroquinone, 1,4-dihydroxybenzene) as host 7.2.1.1 Crystal structures of quinol polymorphs and -quinol clathrates 7.2.1.2 Low temperature phase transitions in -quinol clathrates 7.2.1.3 Introduction to statistical thermodynamics of clathrate structures and application to the quinol clathrates 7.2.2 Crystal structure of {(6H2O)[hexamethylene tetramine]} 7.2.3 Clathrates derived from existing structures 7.2.3.1 Helium hexahydrate 7.2.3.2 Cadmium cyanide clathrates 7.2.4 Overview of the polyhedral clathrates (including metalloid structures, clathrasils, gas hydrates, clathrate and semiclathrate hydrates) 7.2.4.1 Historical and general introduction 7.2.4.2 Restrictions on the shapes of the polyhedra 7.2.4.3 Packing of pentagonal dodecahedra 7.2.5 Metalloid structures 7.2.6 Clathrasils 7.2.7 Gas hydrates (structures with pentagonal dodecahedra) 7.2.7.1 Relation between guest type and structure type in the gas hydrates 7.2.7.2 Stoichiometry and thermodynamics of the gas hydrates 7.2.7.3 Prototype CS-I and CS-II crystal structures at low temperatures 7.2.7.4 Br286H2O is the only bromine hydrate, and the implications of this result 7.2.7.5 Gas hydrates with charged frameworks (ionic clathrate hydrates) 7.2.8 Peralkylonium hydrates and related structures 7.2.8.1 Introduction 7.2.8.2 Structures based on the CS-I structure 7.2.8.3 Structures based on the CS-II structure 7.2.8.4 Structures based on the HS-II structure 7.2.8.5 Structures based on the HS-I structure and its superstructure SHS-I 7.2.8.6 Structures based on the OS-I structure 7.2.8.7 Structures based on the TS-I structure 7.2.8.8 The effectiveness of the alkyl substituents in forming hydrates 7.2.9 Varieties of structures formed by a particular guest 7.2.10 The alkylamine hydrates 7.2.11 Structures without pentagonal dodecahedra (some with charged frameworks)

xiii

323 323 331

333 345 346 346 347

348 348 353 355 360 363 370 370 372 379 381 383 383 383 384 385 385 385 387 387 389 389 389 392

xiv

CONT ENTS

7.3 Hosts with a combination of directional bonds and van der Waals interactions 7.3.1 Phenol (and related compounds) as hosts 7.3.1.1 Phenol 7.3.1.2 Guayacanin as host 7.3.2 Dianin’s compound (4-p-hydroxyphenyl-2,2,4trimethylchroman) and related compounds as hosts 7.4 Van der Waals linked hosts 7.4.1 Tetraphenylene as host 7.5 Hexahosts and related compounds 7.6 Conclusions and a perspective view References 8 Inclusion complexes formed by versatile hosts 8.1 Introduction 8.2 Tri-o-thymotide and analogs as hosts 8.2.1 Crystallography of tri-o-thymotide and its complexes 8.2.1.1 The trigonal clathrate inclusion complexes 8.2.1.2 The hexagonal tunnel inclusion complexes 8.2.1.3 Tunnel inclusion complexes with organometallic guests 8.2.1.4 Crossed tunnel triclinic inclusion complexes 8.2.1.5 Miscellaneous inclusion complexes 8.2.2 Analogs of tri-o-thymotide 8.3 Trimesic acid and analogs as hosts 8.3.1 Introduction 8.3.2 Host–guest tunnel inclusion complexes based on noncatenated unary hexagonal networks 8.3.2.1 TMA as host 8.3.2.2 Two coordination complexes as potential hosts 8.3.3 Host–guest tunnel inclusion complexes based on catenated hexagonal unary networks 8.3.4 Host–guest clathrate interstitial inclusion complexes based on catenated hexagonal unary networks 8.3.5 Generalization of the concept of ‘‘interruption’’ to give binary networks 8.3.5.1 TMAH2O networks 8.3.5.2 Catenated neutral binary networks 8.3.5.3 Ionic binary networks 8.3.6 Hydrogen-bonded TMA binary complexes 8.4 The Heilbron complexes 8.5 Gossypol and its inclusion complexes 8.6 Tris(5-acetyl-3-thienyl)methane (TATM) as host 8.6.1 Introduction 8.6.2 Chemistry of TATM and its inclusion complexes 8.6.3 Conformations taken up by the TATM molecule in the various crystallographic structure types 8.6.4 Crystallography of the inclusion complexes of TATM

396 396 396 398 399 406 406 408 410 411 421 423 423 423 425 429 431 433 435 436 437 437 437 437 443 446 448 448 449 450 450 452 456 459 469 469 469 470 474

CONTENTS

8.6.5 Formation of the inclusion complexes 8.6.6 Dynamics of guest molecules in the complexes 8.6.7 Other examples 8.6.8 Summary 8.7 (5,10,15,20)-Tetraphenylmetalloporphyrins and complexes 8.7.1 Introduction 8.7.2 Crystallography of (5,10,15,20)tetraphenylmetalloporphyrin coordination complexes 8.7.2.1 Introduction 8.7.2.2 The four-coordinate coordination complexes 8.7.2.3 The five-coordinate coordination complexes 8.7.2.4 The six-coordinate coordination complexes 8.7.3 Crystallography of (5,10,15,20)tetraphenylmetalloporphyrin inclusion complexes 8.7.3.1 Crystallography of four-coordinate (5,10,15,20)tetraphenylmetalloporphyrin inclusion complexes 8.7.3.2 Crystallography of five-coordinate (5,10,15,20)tetraphenylmetalloporphyrin inclusion complexes 8.7.3.3 Crystallography of six-coordinate (5,10,15,20)tetraphenylmetalloporphyrin inclusion complexes 8.7.4 Comparative crystallography of the (5,10,15,20)tetraphenylmetalloporphyrin coordination and inclusion complexes 8.7.5 Questions of nomenclature and description 8.7.6 Can ‘‘sponge’’ structures be inferred from the chemical nature of the second component? References 9 Intercalation complexes 9.1 Introduction 9.2 Intercalation complexes of graphite (GICs) 9.2.1 Introduction 9.2.2 Alkali metals as guests (preparation at atmospheric pressure) 9.2.3 Alkali metals as guests (preparation at high pressures) 9.2.4 Alkaline earth and lanthanide metals as guests 9.2.5 Halogens as guests 9.2.6 Uses, actual and potential, of GICs 9.2.7 Summary for GICs 9.3 Intercalation complexes of inorganic hosts 9.3.1 Introduction 9.3.2 2H--TaS2 as host 9.3.3 Zirconium phosphates and phosphonates as hosts

xv

476 477 477 477 478 478 480 480 481 483 486 493

493

500

502

505 506 507 508 515 516 517 517 520 532 533 533 539 539 540 540 540 546

CONT ENTS

xvi

9.4 Concluding summary References

PART IV

552 552

PACKING COMPLEXES

Introduction to Part IV 10 Packing complexes 10.1 Introduction 10.2 Components are isomers of various types 10.3 The (stereoisomeric) components are enantiomers 10.3.1 Background 10.3.2 Types of binary phase diagram found with enantiomers as components 10.3.3 Formation of conglomerates 10.3.4 Comparing the stability of enantiomorphs and racemic compounds 10.3.5 The structural chemistry of systems with isolatable enantiomers 10.3.5.1 Components are rigid molecules 10.3.5.2 Complications due to conformational isomerism of component molecules 10.3.6 Enantiomorphs and racemic compounds of nonisolatable conformers 10.3.7 Racemic enantiomorphs 10.3.8 Solid solutions of enantiomers 10.4 The (stereoisomeric) components are diastereoisomers 10.4.1 Background 10.4.2 Diastereoisomers as components 10.4.3 Conformers as components 10.4.4 Cis-trans isomers as components 10.4.5 Cocrystallization of tautomers 10.4.6 Interallogon complexes 10.5 Components are positional isomers 10.6 Components have different chemical compositions 10.6.1 Substitutional solid solutions 10.6.1.1 Conditions for formation 10.6.1.2 The information desired 10.6.2 Systems with complete (or a wide range of ) mutual miscibility 10.6.3 Some binary phase diagrams involving phases (primary and intermediate) with extensive composition ranges 10.6.3.1 Dibenzyl – trans-stilbene 10.6.3.2 Diphenyl sulphoxide–diphenyl sulfone 10.6.3.3 p-Dibromobenzene–p-diiodobenzene 10.6.3.4 p-Dibromobenzene–p-chloronitrobenzene

563 564 565 566 566 568 569 571 574 574 577 582 584 585 590 590 590 592 597 600 603 605 605 605 605 607 608 615 615 615 617 617

CONTENTS

10.6.3.5 Benzoic acid–p-fluorobenzoic acid 10.6.3.6 1,2-4,5-Tetrachlorobenzene–1,2-4,5tetrabromobenzene 10.6.4 Evidence for nonrandom substitution in primary solid solutions 10.6.4.1 (trans-Stilbene)–diphenylmercury and tolane–diphenylmercury 10.6.4.2 2,3-Dimethylnaphthalene–anthracene 10.6.5 Inhomogeneity of some solid solution crystals 10.7 Interblock solid solubility 10.8 Primary interstitial solid solutions 10.9 Ordered packing complexes 10.9.1 Metal coordination complexes 10.9.1.1 Components of different composition and configuration but in the same oxidation state 10.9.1.2 Components with different compositions, configurations and oxidation states 10.9.1.3 The components are oligomers 10.9.2 Complexes in which a moiety plays more than one structural role 10.9.3 Miscellaneous packing complexes 10.9.3.1 Packing complexes without specific interactions 10.9.3.2 Packing complexes with incipient specific interactions 10.9.3.3 Packing complexes of the fullerenes C60-Ih, 6 C70-d5h and C76 References Book index

xvii

617 617 618 618 620 620 622 622 623 623 624 624 624 625 628 628 634 634 667 [1]

Volume 2 PART V

MOLECULAR COMPOUNDS WITH LOCALIZED INTERACTIONS

Introduction to Part V 11 Donor–acceptor molecular compounds (essentially localized interactions) 11.1 Introduction and classification

683 684

Part 1: Pure acceptors 11.2 n-Donors and s-acceptors 11.2.1 N, O, S containing ligands as donors and AgI salts as acceptors

687 687

CONT ENTS

xviii

11.3

11.4 11.5 11.6

11.7

11.8

n-Donors and *-acceptors 11.3.1 N, O, S or Se containing donors and dihalogens or halogenated molecules as acceptors 11.3.2 S containing molecules as donors and iodine molecules as acceptors (the polyiodines) 11.3.3. Physical measurements on molecular compounds of the type discussed above 11.3.4 Halogenated molecules as donors and dihalogens as acceptors 11.3.5 Self-complexes – N, O, S, Se to halogen interactions in one-component systems n-Donors and p-acceptors 11.4.1 N, O or S containing ligands as donors and Group VA metal halides as acceptors n-Donors and *-acceptors -Donors and *-acceptors 11.6.1 Aromatic molecules as donors and dihalogens as acceptors 11.6.2 Aromatic molecules as donors and polyhalogenated methanes as acceptors -Donors and p-acceptors 11.7.1 Aluminum tribromide as an acceptor 11.7.2 Miscellany-mainly MX3 (M ¼ As, Sb; X ¼ Cl, Br) as acceptors and aromatic molecules as donors -Donors and (localized) *-acceptors

688 688 706 708 712 713 717 717 723 727 727 731 733 733 734 736

Part 2: Self-interacting acceptors 11.9

n-Donors and s-acceptors 11.9.1 N, O, S containing ligands as donors and AgI salts as acceptors 11.9.2 N, O, S containing ligands as donors and HgX2 (X ¼ Cl, Br, I) as acceptors 11.10 n-Donors and p-acceptors 11.10.1 N, O, S containing ligands as donors and MX3 (M ¼ As, Sb; X ¼ Cl, I) as acceptors 11.11 -Donors and s-acceptors 11.11.1 Aromatics as donors and Ag(I) salts as acceptors; also {benzene CuAlCl4} 11.11.2 Olefins as donors and Ag(I) salts as acceptors 11.11.3 Some general structural principles emerging from Sections 11.11.1 and 11.11.2 11.11.4 Acetylides as donors and Ag(I) salts as acceptors 11.11.5 Acetylides as donors and Cu(I) salts as acceptors 11.11.6 Aromatics as donors and Hg(II) salts as acceptors 11.12 -Donors and p-acceptors n n n

737 737 739 761 761 764 765 777 783 784 784 786 788

CONTENTS

11.12.1 Aromatics as donors and MX3 (M ¼ Sb, Bi; X ¼ Cl, Br) as acceptors 11.12.2 Aromatics as donors and np3 metal ions (GaI, InI, TlI, SnII, PbII) as acceptors 11.13 Summary References 12 Hydrogen bonded molecular complexes and compounds 12.1 Introductory survey 12.1.1 Introduction 12.1.2 The characteristic features of hydrogen bonds 12.2 Application of graph theory to the description of hydrogen bond patterns 12.3 Statistics of hydrogen bond patterns 12.3.1 Methodology 12.3.2 Statistics of ring formation 12.4 Appendage structures (one component forms a hydrogen bonded framework, to which the second component is appended by hydrogen bonding) 12.5 Alternating framework structures (the components, in hydrogen-bonded alternating array, form a mixed framework) 12.5.1 Zero-dimensional frameworks 12.5.1.1 Structures with discrete pairs (A–B) of components 12.5.1.2 Structures with discrete triples (B–A–B) of components 12.5.1.3 Larger discrete groupings of components 12.5.2 One-dimensional frameworks (linear chains of alternating components) 12.5.2.1 Component A has two donor groups and the single acceptor of component B can accept two hydrogen bonds 12.5.2.2 Component A has two hydrogen bond donor groups and component B two acceptor groups 12.5.2.3 Both components have both hydrogen bond donor and acceptor functions 12.5.3 Two-dimensional frameworks (layer arrangements of alternating components) 12.5.4 Three-dimensional frameworks (arrangements of alternating components in space) 12.5.5 Accounting for formation of a molecular compound 12.6 Crystal engineering with hydrogen bonds 12.7 Charged or neutral moieties – when is there hydrogen transfer between the components? References

xix

788 822 836 836 851 852 852 853 861 862 863 864

864

867 867 867 877 881 884

885 889 893 896 901 904 905 908 911

CONT ENTS

xx

PART VI

MOLECULAR COMPOUNDS WITH DELOCALIZED INTERACTIONS

Introduction to Part VI 13 Charge transfer molecular compounds with delocalized –* interactions – introduction and general survey 13.1 Introduction and historical development 13.2 Classification 13.2.1 General considerations 13.2.2 Intramolecular -compounds and self-complexes 13.3 Chemical nature of donors and acceptors 13.3.1 Introduction 13.3.2 Donors 13.3.3 Acceptors 13.3.4 Quasi-acceptors 13.3.5 Ionization potentials of donors and electron affinities of acceptors 13.3.6 Determination of degree of charge transfer 13.4 Binary and quasi-binary donor–acceptor systems 13.4.1 Phase diagrams 13.4.2 Component ratios in binary donor–acceptor systems 13.5 Ternary -molecular compounds References 14 Layered molecules with intra-molecular donor–acceptor interactions 14.1 Introduction 14.2 Molecules of the paracyclophane type 14.2.1 Molecules derived from [n.n]paracyclophanes 14.2.2 Systems related to [n.n]paracyclophanes 14.2.3 Multi-layered systems 14.3 Molecules of the metaparacyclophane type 14.4 Molecules of the metacyclophane type 14.5 Some other systems 14.6 Concluding summary References 15 Crystal chemistry of mixed-stack –* molecular compounds 15.1 Introduction 15.2 Nonstacked structures containing structural groups of limited size 15.3 The crystallochemical families found for 1:1 –* molecular compounds

925 926 927 927 930 932 932 933 935 939 940 944 948 948 952 953 954

959 959 961 961 972 974 976 980 984 986 986 989 990 993 994

CONTENTS

15.4 Packing arrangements in n : m –* molecular compounds 15.5 Some special features of packing arrangements in –* molecular compounds 15.5.1 Crystals where one of the components is also found in interstitial positions 15.5.2 Noncentrosymmetric crystals of -molecular compounds 15.5.3 Acceptors based on polynitrofluorene 15.5.4 Resolution of helicenes by formation of diastereoisomeric charge transfer molecular compounds with enantiomeric acceptors 15.6 Structurally important interactions between polarizable and polar groups 15.7 Mixed-stack crystals with both charge transfer and hydrogen bonding interactions 15.7.1 The quinhydrones as a crystallochemical family 15.7.2 Molecular compounds of the flavins 15.7.3 Other crystals with both charge transfer and hydrogen bonding interactions 15.8 Mixed-stack crystals with both delocalized and localized charge transfer interactions 15.9 Donors and acceptors with special chemical features 15.9.1 Fluorinated aromatics as quasi-acceptors 15.9.2 1,3,5,7-tetramethyluric acid (TMU) as quasi-acceptor 15.9.3 Acceptor is a metal coordination complex 15.9.4 Donor is a metal coordination complex 15.9.5 Donors based on phenazine 15.10 Mixed-stack donor–acceptor molecular compounds with ionized ground states 15.10.1 Mixed-stack closed-shell charge transfer salts 15.10.2 Ion-radical salts 15.11 Isomeric (polymorphic) molecular compounds 15.11.1 Type 1 – isomerism due to different types of interaction without change of moiety structure 15.11.2 Type 2 – isomerism due to electron transfer 15.11.3 Type 3 – isomerism due to proton transfer or to –* electron transfer 15.11.4 Isomerism stabilized by both charge (–*) and proton transfer (CPT compounds) 15.12 Self-complexes 15.13 Conclusions 15.13.1 Structural variety in –* molecular compounds 15.13.2 How should the packing arrangements in –* molecular compounds be described?

xxi

1001 1005 1005 1007 1009

1010 1011 1013 1013 1022 1026 1030 1032 1032 1040 1040 1042 1044 1047 1047 1048 1052 1052 1054 1055 1058 1059 1064 1064 1065

CONT ENTS

xxii

15.13.3 Structural consequences of –* interactions References (Note. The components in the ground states of these molecular compounds are taken to be neutral unless explicitly stated otherwise). 16 Crystal (structural) physics of mixed stack –* molecular compounds 16.1 Introduction 16.2 Thermodynamic parameters 16.3 Spectroscopic measurements on the excited state 16.4 Crystals with disorder ) order transformations on cooling – modern treatments of second order phase transitions 16.4.1 General introduction 16.4.2 The Ehrenfest order of a phase transition 16.4.3 Landau theory of phase transitions 16.4.4 The critical exponents 16.4.5 The permitted symmetries of a low symmetry phase derived from a particular high symmetry phase 16.4.6 Temperature dependence of the order parameter 16.4.7 Pressure dependence of the critical temperature for ordering 16.5 Thermodynamic, structural and kinetic investigations of various systems showing second order transitions on cooling 16.5.1 The crystal structure of {Pyrene PMDA}(PYRPMA) and evidence for an order , disorder phase transition at 160K 16.5.2 The crystal structure of {Naphthalene TCNB} and evidence for an order , disorder phase transition at 72K 16.5.3 The crystal structure of {Anthracene TCNB} and evidence for an order , disorder phase transition at 213K 16.5.4 Other examples of second order transitions 16.6 Crystals with first order transformations on cooling 16.6.1 {Cycl[3.2.2]azine TNB} 16.6.2 Other examples 16.7 Physical nature of the disordered phase 16.8 Transformation to quasi-plastic phase(s) on heating 16.9 Transformation of the ground state from neutral ) ionic on cooling and/or application of pressure (NI transitions) 16.9.1 Introduction 16.9.2 {TTF chloranil} 16.9.3 {DMTTF chloranil} 16.9.4 Other examples 16.9.5 Concluding summary References

1066 1068

1081 1082 1083 1086 1090 1090 1091 1092 1093 1094 1096 1097 1097

n n n

1097

n n n

1105

n n n

n n n

n n n

n n n

1115 1119 1120 1120 1122 1122 1126 1128 1128 1129 1137 1139 1142 1142

CONTENTS

17 Segregated stack -molecular complexes 17.1 Introduction 17.2 Chemistry of donors and acceptors that participate in segregated stacks 17.2.1 Introduction 17.2.2 Donors 17.2.3 Acceptors 17.2.4 Preparation of crystals 17.3 Structures of cation-radical salts 17.3.1 Introduction 17.3.2 Cations are polycyclic aromatic hydrocarbons 17.3.3 TTF and related compounds as cations 17.3.4 TMPD salts containing -dimerized cation radicals 17.4 Structures of TCNQ anion-radical salts 17.4.1 Mutual arrangements of approximately plane-parallel TCNQ moieties 17.4.2 Structures with stacks of limited length 17.4.3 TCNQ anion radical salts in which the cations are metals 17.4.4 Stacked structures with –e average charge on the TCNQ moieties 17.4.5 Stacked structures with –0.8e average charge on the TCNQ moieties 17.4.6 Stacked structures with –2/3e average charge on the TCNQ moieties 17.4.7 Stacked structures with –0.5e average charge on the TCNQ moieties 17.4.8 Stacked structures with –0.4e average charge on the TCNQ moieties 17.4.9 Systems studied over a wide range of temperatures 17.4.10 Conclusions drawn from a survey of structural results for TCNQ anion radical salts 17.5 Other anion-radical salts 17.5.1 Alkali-metal chloranil salts 17.5.2 M(dmit)2 and M(mnt)2 as anion radicals in various guises 17.6 Structures of cation-radical anion-radical salts 17.6.1 General survey 17.6.2 Cation : anion ratio 1 : 1; monad stacks 17.6.3 Cation : anion ratio 1 : 1; diad stacks 17.6.4 Cation : anion ratio 2 : 1 or 1 : 2; monad stacks 17.6.5 Cation : anion ratio 2 : 1 or 1 : 2; diad stacks 17.7 Electron density studies of segregated stack complexes

xxiii

1147 1148 1151 1151 1152 1157 1161 1162 1162 1163 1167 1175 1177 1177 1180 1187 1189 1192 1193 1196 1202 1205 1211 1214 1214 1215 1220 1220 1220 1227 1229 1230 1232

xxiv

CONT ENTS

17.8

Theoretical studies of some segregated stack complexes 17.9 Studies of {[TTF][TCNQ]} and some related materials 17.10 Concluding summary References Appendix Book index

1234 1235 1252 1253 1267 [1]

Acknowledgements The author wishes to thank the following for permission to reproduce published material. Academic Press: Inclusion Compounds. Figs. 6.39, 6.65, 7.1, 7.24(a), 7.24(b), 7.38, 7.39, 7.40, 7.42, 7.43, 9.19, 9.22. Non-Stoichiometric Compounds: Fig. 8.4. American Association for the Advancement of Science Science. Fig. 10.24. American Chemical Society Accts. Chem. Res. Figs. 7.36, 10.15. Biochemistry. Figs. 5.7, 5.8, 5.9, 5.12. Chemistry of Materials: Figs. 6.3(b), 6.14, 6.20, 6.30, 8.6. Cryst. Growth & Design: Figs. 12.35, 12.36. Inorg. Chem. Figs. 9.18, 9.21, 10.19, 11.41, 11.58, 11.59, 11.71, 11.100, 11.119, 15.27, 17.38. J. Am. Chem. Soc. Figs. 3.4(a), (b), 3.8, 3.9, 3.17, 3.19, 3.25, 3.26, 3.28, 4.4, 4.15, 6.55, 6.60, 6.61, 6.62, 6.69, 7.5, 7.6, 7.32, 8.8, 8.53, 10.13, 11.16, 11.31, 11.43, 11.60, 11.63, 11.116, 12.6, 12.12, 12.14, 12.16, 12.18, 12.21, 12.27, 15.13, 15.32, 15.33, 16.43, 16.44, 17.9, 17.15, 17.19, 17.23, 17.42, 17.53. J. Chem. Educ. Figs. 4.1, 7.29. J. Med. Chem. Fig. 5.21. J. Org. Chem. Figs. 11.3, 11.72. J. Phys. Chem. Figs. 6.19, 6.26, 6.27, 6.28, 7.26, 17.22(b), 17.37. Organometallics: Figs.11.113, 11.121. American Crystallographic Association Fig. 3.13. American Institute of Physics J. Chem. Phys. Figs. 6.17, 6.25, 7.7, 7.15, 7.34, 7.35, 7.37, 15.29, 16.5, 16.20, 16.23. Sov. Phys. Crystallogr. Figs. 10.20, 10.22 American Physical Society Phys. Rev. Letts.: Figs. 13.4, 16.36, 17.46. Phys. Rev.: Fig. 16.38. Elsevier Adv. Organometall. Chem.: Fig. 9.6. CALPHAD: Fig. 10.21. Carbohydrate Research: Figs. 4.5(a), 4.11(a).

xxvi

ACKNOWLEDGEMENTS

Carbon: Figs. 9.10, 9.11. Chem. Phys. Letts: Fig. 6.11. Comp. Rend. Acad. Sci. (Paris), Ser. C,: Fig. 17.18. Coord. Chem. Revs. Fig. 10.40. FEBS Letters: Fig. 5.23. Inorg. Chim. Acta: Figs. 11.46, 11.53. J. Chromatography: Fig. 15.9. J. Mol. Biol. Figs. 5.6, 5.11, 5.17, 5.18, 5.20. J. Organometall. Chem.: Figs. 10.34, 11.120. J. Phys. Chem. Solids: Figs. 6.13, 17.52. Mater. Sci.: Fig. 16.14. Sol. State Comm. Fig. 16.39. Synth. Mets. Fig. 16.42. International Union of Crystallography Acta Crystallographica, Figs. 6.31, 6.51, 10.14, 10.16, 11.1, 11.15. Acta Crystallographica, B. Figs. 4.3, 4.16, 4.17, 4.19, 6.1, 6.3(a), 6.8, 6.50, 6.53, 6.58, 7.2, 7.25, 7.31, 7.41,10.7,10.8, 10.26, 11.40, 12.2, 12.5, 12.9, 12.10, 12.30, 12.31, 15.2, 15.3, 15.5, 15.7, 15.15, 15.18, 15.21, 16.4, 16.9, 16.10, 16.11, 16.12, 16.13, 16.22, 16.32, 17.17, 17.20, 17.22(a), 17.24, 17.27, 17.29, 17.32, 17.33, 17.43(a), 17.48, 17.49. Acta Crystallographica, C. Figs. 6.18, 8.7, 8.31, 8.37, 8.38, 10.4, 10.17, 10.17A, 10.32(a), 10.32(b), 12.24, 12.28, 15.6, 17.6, 17.7. Acta Crystallographica, D. Fig. 5.16. IUCr Monographs on Crystallography: Figs. 15.34, 15.35. J. Appl. Cryst. Fig. 7.27. Kluwer Academic Publishers J. Incl. Phenom. Figs. 3.11, 6.43, 7.13, 7.14, 7.24, 9.13. Macmillan Publishers Nature. Figs. 5.15, 7.16, 10.33. National Academy of Sciences U. S. A.: Proceedings. Figs. 5.13, 5.14, 5.22. NRC Research Press (Canada) Can. J. Chem.: Figs. 11.45, 11.52. Oldenbourg Verlag Z. Kristallogr. Fig. 6.21. Pergamon Comprehensive Supramolecular Chemistry. Figs. 4.2, 6.59 (Vol. 6). Tetrahedron Letters: Figs. 3.12, 3.22. Tetrahedron: Fig. 12.13. Plenum Water–a comprehensive treatise. Fig. 7.30. J. Cryst. Spectroscop. Res.: Fig. 10.31.

A C KN O W L E D G E M E NT S

xxvii

Professor M. Le Cointe. Ph.D. thesis, University of Rennes I: Figs. 16.35, 16.37, 16.40. RIA-Novosti, Paris La Recherche: Fig. 7.18 (I am grateful to Professor Rose Marx, Saclay, for her help in obtaining this figure). Springer: Monatshefte Chem.: Fig. 10.10. Springer Series in Materials Science No 18: Figs. 9.7, 9.8, 9.9. Topics in Current Chemistry. Figs. 6.41, 12.26. Taylor and Francis: Adv. Phys.: Fig. 9.2. Contemp. Phys.: Fig. 17.57. Mol. Cryst. Liq. Cryst. Figs. 6.12, 6.22, 6.23, 6.33, 10.2, 10.23, 11.25, 15.30, 16.29, 17.3. The Chemical Society of Japan Bull. Chem. Soc. Jpn. Figs. 4.5(b), 4.6(a), (b), 4.7, 4.9, 4.10, 4.13, 4.18, 11.42, 13.9, 13.10, 13.11, 16.33, 16.34, 17.10, 17.16. Chem. Letts. Figs. 7.17, 10.46, 14.6, 17.44. The Physical Society of Japan J. Phys. Soc. Jpn.: Figs. 13.7, 17.54, 17.55. The Royal Society of Chemistry Chemical Communications: Figs. 3.3, 3.4(c), 3.5, 3.16, 3.23, 6.47, 6.48, 8.21, 8.22, 10.30, 10.35, 10.41, 12.19. Chem. Soc. Revs.: Fig. 9.17. Chemistry in Britain, Fig. 1.2. J. Chem. Soc. A: Fig. 11.49. J. Chem. Soc. B: Figs. 6.67, 6.68. JCS Dalton. Figs. 8.15, 8.16, 8.17, 8.18, 9.23, 9.24, 11.12, 11.33, 11.39, 11.101, 17.11, 17.36. JCS Perkin II: Figs. 3.15, 6.40, 6.42, 7.3, 12.15, 15.14. JCS Trans. Farad. Soc. Figs. 6.4, 6.5, 6.6. J. Mater. Chem.: Figs. 17.39, 17.40. New J. Chem.: Fig. 12.3. The Royal Society of London: Proceedings, Ser. A: Figs. 8.19, 9.12, 15.4, 16.8, 16.30, 16.31. Verlag Chemie-Wiley Angew. Chem. Intl. Ed.: Figs. 3.27, 5.3, 10.36, 11.115, 14.2. Chem. Ber. Figs. 10.28, 11.35, 14.8, 14.12, 14.13, 14,14, 16.45. Chemistry Eur. J. Fig. 6.56. Prog. Inorg. Chem. Figs. 7.21, 7.23.

xxviii

ACKNOWLEDGEMENTS

Verlag Helvetica Chimica Acta: Helvetica Chimica Acta. Figs. 5.19, 7.4, 10.11. Verlag der Zeitschrift fu¨r Naturforschung Z. Naturforsch. (b): Figs. 11.110, 11.114, 11.118. Worth Publishers New York Lehninger Biochemistry, 2nd edition. Figs. 5.1, 5.2. Various Acta Chem. Scand.: Figs. 11.7, 11.8, 11.9, 11.30. Acta Chem. Scand A: Figs. 11.10. 11.11. J. Phys. D: Fig. 9.1. J. Struct. Chem. USSR: Figs. 6.2, 7.33. Liebigs Annalen: Fig. 15.20. Molecular Complexes: Fig. 13.8. Phys. Chem. Low-dimens. Materials: Fig. 9.15. I am grateful to Dr Moshe Kapon and Dr Mark Botoshansky for help of many kinds and to the staff of the Chemistry-Biology Library at Technion for their assistance in tracking down material.

Part I Some preliminaries


Chapter 1 Structural principles in the classification of binary adducts

The author . . . thereby extended an old chemical tradition, of calling complexes all those compounds which are in some way odd or unusual, and which do not readily fit the line-for-every-electron-pair representations that are the hallmark of our language. Very often the term complex persists even after the nature of the species has been elucidated – thus we have the complex inorganic ions, the -complexes, s-complexes, charge-transfer complexes, organo-metallic complexes, clathrate complexes, hydrogen-bonded complexes, and so on. W. J. le Noble (1974). Highlights of Organic Chemistry, An Advanced Textbook, Dekker New York Chapter 23, ‘‘Complexes’’, pp. 841–842

Summary: When the properties of the individual components are largely conserved in the primary or intermediate crystalline phases of a two-component (A and B) system, then these phases are called ‘‘binary adducts’’, which is a more formal name for what are usually called ‘‘molecular compounds and complexes.’’ The various types of binary adduct are classified in terms of those interactions between the components that determine the component arrangement in the crystal. Thus A . . . A interactions dominate in inclusion complexes, A . . . A and B . . . B interactions are equally important in segregated-stack charge transfer complexes, all interactions are of roughly equal importance in packing complexes, and A . . . B interactions dominate in molecular compounds. This classification is compared to other complementary schemes.

1.1 Introduction 1.2 Structural classification of binary adducts 1.2.1 General considerations 1.2.2 Molecular complexes 1.2.2.1 Inclusion complexes 1.2.2.2 Moieties within molecules 1.2.2.3 Frameworks with guest participation and/or linkage 1.2.2.4 Segregated stack charge transfer complexes 1.2.2.5 Packing complexes 1.2.3 Molecular compounds 1.3 Other classifications 1.4 How many binary adducts are there? 1.5 Organic and inorganic supramolecular chemistry References

4 5 5 6 6 9 9 9 10 10 10 11 12 12

4

1.1

PRINCIPLES IN THE CLASSIFICATION OF BINARY ADDUCTS

Introduction

The section at 1 bar through the pressure–temperature–composition (P–T–x) phase diagram of a binary (two component) system shows the conditions of stability of the various crystalline phases of composition AxBy. These phases may be primary solid solutions of B in A (or conversely), or AxBy compounds, with congruent or incongruent melting points, with crystal structures different from those of A and/or B. We define a ‘‘binary adduct’’ as a crystalline two-component phase, relatively easily separated into its components, in which the properties of the individual components are very largely conserved. In this definition, which is based on some comments of Ketelaar (1958), the words ‘‘binary,’’ ‘‘component’’ and ‘‘phase’’ have the meanings of the Phase Rule (Ricci, 1966). The phrase concerning the separation of the components refers to the status of intertwined species (see Chapter 3). The collective term ‘‘binary adduct’’ refers to all those chemical species which are the subject matter of this book. We shall be faithful to this definition after our own fashion for there are occasions when the advantages of a broader treatment justify some straying from the strait and narrow paths of excessive conformity. An immediate consequence of the conservation of the properties of the individual components is that there cannot be covalent bonding between the two components, although ionic and ion–dipole interactions are allowed. The binary adduct may have the same crystal structure as one of the components, when there will be solid solution of the minor (guest) component B in the major (host) component A, or it may have a different crystal structure and thus appear as a compound in the A–B phase diagram. This is an important distinction because the solid solution crystal maintains its structure on decomposition of the adduct into its components, while the phase diagram compound decomposes into separate crystals of one (when one component is a gas or liquid under ambient conditions) or both components. Harris (1997) has suggested that the two types of host should be called ‘‘hard’’ (solid solution) and ‘‘soft’’ (phase diagram compound) respectively. We prefer to keep the connection to well-established phase-diagram principles rather than adapt terms already used in other areas of chemistry. Not all phase diagram compounds are binary adducts in our present sense; for example, the phase KF2Al(C2H5)3 is composed of Kþ cations and [(C2H5)3Al-F-Al(C2H5)3] anions (Allegro and Perego, 1963), the properties of the individual components clearly not being ‘‘very largely conserved.’’ In contrast, the neutral molecule–salt complex [(C2H5)4NþBr]2(succinimide)] (Powell and Wait, 1958) could be included because the moieties1 of the individual components appear also in the complex, although with altered mutual arrangement. The principles governing binary adducts can be carried over without much change to ternary and higher adducts and some of these will also be discussed. There is a considerable resemblance between the phase diagrams and thermodynamics of molecular systems on the one hand and those of metal alloy systems on the other; in both instances the properties of the individual components are largely conserved. This is generally not so in purely inorganic systems, where rearrangement of the ions can occur. But the resemblance between molecular and metallic systems hardly extends to structural features and modes of interaction between the components. Metal atoms are 1 One dictionary defines moiety as a portion of indefinite size; we use it as a convenient term for molecule and/or ion.

STRUCTURAL CLASSIFICATION OF BINARY ADDUCTS

5

approximately spherical while molecules usually have complicated shapes; indeed formation of some binary adducts can be ascribed to particular features of these complicated shapes. Interactions between metal atoms in alloys and intermetallic compounds are approximately isotropic and, at the simplest level, are ascribed to delocalisation of one (or a few) electrons per atom over the volume of the substance; interactions between the entities in binary molecular adducts are usually highly anisotropic and directional and are ascribed, to different extents in different adducts, to hydrogen bonding, to localized or delocalized charge transfer, to ionic and ion–dipole forces and to the ubiquitous dispersion forces, possibly all acting in combination. It will be immediately apparent that we have here formalized the definition of ‘‘molecular complexes and compounds,’’ restricting ourselves to the crystalline state. The interactions between the components of binary adducts that occur in solution (or in the vapor phase) are traditionally considered to be of a transient, contact nature which can affect physical properties but are difficult to define in structural terms. However, over the past thirty years, there have been two important developments, one of which concerns us directly and immediately while the second is likely to be of great importance in the future. The first of these developments is the synthesis of an important new group of adducts in which the propinquity of the components persists in solution; the crown ethers represent the first examples of this type of adduct. We use the overall term ‘‘moieties within molecules’’ to define this group. The second concerns the explosive improvement in techniques of studying the structures of adducts formed in the gas phase. Earlier work had employed spectroscopic methods to infer structures of gas-phase adducts (Tamres and Strong, 1979) but the great advance has come from the use of supersonic beams to cool the adducts formed to very low effective temperatures. Most of this work has so far concerned adducts which exist only in the gas phase (Hutson, 1990) but this will surely be extended to adducts which also exist in the crystalline state, thus permitting inference of the importance of aggregation on structure and properties.

1.2 Structural classification of binary adducts 1.2.1 General considerations The primary level of our classification (Herbstein, 1993) is based on the relative structural importance of the interactions between the components A and B in determining the component arrangement in crystalline AxBy. The various hierarchies possible for A . . . A, B . . . B and A . . . B interactions can be completely listed as shown in Table 1.1. We distinguish between molecular compounds and complexes and define ‘‘molecular compounds’’ as those adducts in which A . . . B interactions determine the structure. The rest are ‘‘molecular complexes,’’ which are further subdivided into a number of groups. We distinguish three different types of molecular complex, in all of which the structural pattern in the crystal is not determined primarily by interactions between the two different components, although these may play some role. Le Noble (see above) has already drawn attention to the ambiguous way in which the term ‘‘complex’’ (from the Latin complectere, to entwine, braid, embrace) is used in chemical nomenclature. A material is first called a complex when its structure is not known, but the term generally persists even

6


Table 1.1. The classification of binary adducts into molecular complexes and molecular compounds Molecular complexes A . . . A dominant A . . . A dominant and A . . . B important A . . . A and B . . . B dominant A . . . A, A . . . B and B . . . B about equally important Molecular compounds A . . . B dominant

Inclusion complexes Inclusion complexes Segregated stack complexes Packing complexes Localized interactions Delocalized interactions

after the structure has been clarified. Thus we have such disparate usages as complex inorganic ions, s-complexes, -complexes, organometallic complexes and so on. The word ‘‘complex’’ is so entrenched in the chemical literature that it seems impossible to banish it, but we shall try to use it only in a clearly defined sense. The second level of our classification refers to the type of interaction denoted by the three dots linking the two components in the A . . . A, etc. notation. This interaction can be due to hydrogen bonding, van der Waals forces, ion and ion–dipole interactions, localized or delocalized charge transfer interactions and perhaps other interactions yet to be identified; we shall use the symbol in later chapters to indicate localized or delocalized charge transfer interactions. Such a classification scheme is summarized in Fig. 1.1. Our classification is based on structural features rather than on the chemical nature of the components, because a particular chemical entity can form adducts of different kinds, dependent on the nature of the second component. Urea provides a simple example; hydrogen bonded urea molecules form spirals enclosing paraffin hydrocarbons (and other types of guest) in typical channel2 inclusion complexes, and can also hydrogen bond to other molecules to form hydrogen bonded molecular compounds, such as hyperol (urea . . . H2O2; structure by X-ray diffraction at 295K (Lu, Hughes and Giguere, 1941) and by neutron diffraction at 85K (Fritchie and McMullan, 1981)). The metal coordination complexes of urea are not binary adducts in our present sense. We shall now discuss the various categories in Fig. 1.1, proceeding from top to bottom. n n n

1.2.2 1.2.2.1

Molecular complexes Inclusion complexes

We use the term ‘‘inclusion complex’’ in a similar but somewhat broader sense than has been customary in the past; parenthetically we remark that Powell (1984) used both ‘‘inclusion’’ and ‘‘enclosure’’ in his comprehensive introduction to the multivolume series Inclusion Compounds, but we have resisted the temptation to introduce yet another term into this already confused field. Thus ‘‘inclusion complexes’’ are all those crystalline twocomponent systems in which A . . . A interactions dominate and host (A) and guest (B) 2 Although the word ‘‘channel’’ has wide usage in the literature, ‘‘tunnel’’ is pictorially a far more appropriate term. The latter is beginning to replace the former in the recent literature, and we dare to join this trend.


7

Binary adducts Moieties within molecules Zeolites

A…A dominant

Inclusion complexes

Clathrates Channel inclusion complexes Lamellar complexes

A … A dominant, A … B important A … A and B … B equally important

Frameworks with guest participation and/or linkage Segregated stack complexes Packing complexes

A…B dominant

Localized interactions

A … A, B … B and A … B all about equally important

H-bonded interactions

Molecular compounds

Charge transfer interactions

Delocalized interactions

Charge transfer interactions

Fig. 1.1. Classification scheme for binary adducts, based on the hierarchy of A . . . A, etc. interactions in determining the arrangement of the A and B components in the crystalline adduct. Some second level aspects are also shown.

components can be distinguished without ambiguity; A . . . B and B . . . B interactions are of lesser importance. We denote the inclusion complexes by the symbolism {host[guest]} or {A[B]}. The terms ‘‘host’’ and ‘‘guest’’ have been used in a wider sense by Cram (see below) but we prefer to apply them only to inclusion complexes. These complexes are further subdivided on the basis of the dimensionality of the inclusion into clathrate (cage), tunnel and intercalation (layer) inclusion complexes. The dimensionality is neatly encapsulated in a diagram first used in an article in Scientific American (Brown, 1962) and now used as the logo of the Journal of Inclusion Phenomena and Molecular Recognition (Davies, Kemula, Powell and Smith, 1983); we use the diagram as presented by Harris (1993). Most inclusion complexes have frameworks in which the host molecules are linked to one another by hydrogen bonds or van der Waals forces. More recent work has shown that it is convenient to add here hosts where primary chemical bonds play an important role.

8


This occurs in one of two ways; either the host is an individual molecule that encloses a guest, (host plus guest) then crystallizing as an entity to form the crystal, or the host framework constitutes the whole crystal, the guests being enclosed in interstices. We call the first of these types ‘‘moieties within molecules’’ and discuss them below. Zeolites (Barrer, 1983), in which specific host–guest interactions can be important, are an example of the second type. Usually, however, host–guest and guest–guest interactions are of the van der Waals type; guest–guest interactions are often neglected. The term ‘‘clathrate’’ was introduced by Powell (1948), who defined such complexes as a ‘‘structural combination of two substances which remain associated not through strong attraction between them but because strong mutual binding of the molecules of one sort only makes possible the firm enclosure of the other (‘clathratus’ – enclosed or protected by cross bars of a grating).’’ Zeolites also fit this definition. The guest molecules in clathrate complexes are enclosed by the framework of host molecules and are localized at or about points within the framework – thus these complexes can be assigned a dimensionality of ‘‘zero’’. In the tunnel inclusion complexes the guests are enclosed in (essentially) one-dimensional tunnels in the matrix of host molecules. There are inclusion complexes that cannot be assigned unambiguously between the tunnel and clathrate types because of the occurrence of constrictions in the tunnels. In the intercalation complexes the guest molecules are located between layers of host molecules, and hence are considered ‘‘two-dimensional’’; the complexes of graphite and alkali metals (Dresselhaus and Dresselhaus, 1981) are typical examples. The discerning reader will already have noticed that the graphite intercalation complexes, with covalent bonding within the quasi-infinite graphite sheets, constitute two-dimensional analogs to the zeolites with their quasi-infinite three-dimensional frameworks.

Isolated cages

Linear nonintersecting tunnels (channels)

Interconnected cages

Two-dimensional interlamellar regions

Intersecting tunnels

Fig. 1.2. Some typical topologies of inclusion cavities in crystalline host solids. The smallest dimension of each of these cavities is comparable with molecular dimensions. (Reproduced with permission from Harris (1993)).


9

1.2.2.2 Moieties within molecules A very important development during the past twenty years, recognised by the award of the 1987 Nobel Prize in Chemistry to Cram (1988), Lehn (1988) and Pedersen (1988), is the purposeful synthesis of molecules designed to enclose (or include) other molecules or ions in which the propinquity of the components persists in solution; the crown ethers (Pedersen, 1988) represent the first examples of this type of adduct and the fullerenes with appropriate guests perhaps the most recent (Shinohara, Sato, Ohkoohchi, Ando, Kodama, Shida, Kato and Saito, 1992). We use the overall term ‘‘moieties within molecules’’ to define this group; in the formal terms of our classification these are molecular complexes because the A . . . A (host–host) interactions predominate, even though here the interactions are by covalent bonding. An additional emphasis stresses the relationship between ‘‘moieties within molecules’’ and ‘‘host–guest inclusion complexes’’ in the spirit of the development of inclusion chemistry, which has been broadly defined as incorporating ‘‘all those chemical species, whether they be continuous solids or discrete molecules, having voids or cavities of molecular dimension, and all of the related chemical association and other reaction chemistry. Implicit in the name is the capability of including some other molecular entities within the cavities’’ (Ramprasad, Lin, Goldsby and Busch, 1988). This point of view is recognized by inserting ‘‘Moieties within molecules’’ and ‘‘Zeolites’’ in Fig. 1.1 together with the now-classical clathrate and tunnel inclusion complexes as parts of the broader category of ‘‘Inclusion Complexes.’’ As the guests are not entirely contained within the hosts in the crystal structures of some members of the ‘‘moieties within molecules’’ group (e.g. in some cyclodextrin inclusion complexes), classification can sometimes be ambiguous. 1.2.2.3 Frameworks with guest participation and/or linkage Atoms or ions of guest moieties forming parts of frameworks, or linked to the framework by hydrogen bonding, were first discovered by Jeffrey (1984) and coworkers when analyzing the crystal structures of the peralkylonium salt hydrates and alkylamine hydrates; the latter were called semiclathrate hydrates. A simpler example is provided by trimesic acid.dimethyl sulphoxide (Herbstein, Kapon and Wasserman 1978), where the oxygen atoms of the dimethyl sulphoxide ‘‘guests’’ are parts of the walls of tunnels formed by the hydrogen bonded trimesic acid molecules, while the methyl groups lie within the tunnels. In these examples A . . . A interactions dominate but A . . . B interactions are important. 1.2.2.4 Segregated stack charge transfer complexes The members of the next group have A . . . A and B . . . B interactions of approximately equal importance, with smaller A . . . B interactions. The only complexes known which fit into this category have the components segregated into separate stacks. Although small in number, the group is of great interest and importance because most of its (organic) members have high electrical conductivities, with tetrathiafulvalene: tetracyanoquinodimethane (TTF–TCNQ) as the most famous example (crystal structure by Kistenmacher, Phillips and Cowan (1974)). There are many chemical and physical resemblances to the (delocalized) –* mixed-stack donor–acceptor molecular

10


compounds (see below), and the two groups will be considered together in Part VI (as is usually done), but with retention of the designation ‘‘complexes.’’ 1.2.2.5 Packing complexes Proceeding among the hierarchical possibilities, the next group has A . . . A, A . . . B and B . . . B interactions all of approximately equal importance; in other words, no one of these interactions is dominant in determining the structure. These are known as ‘packing complexes’. A typical example is sym-tetrabromobenzene:hexabromobenzene (Gafner and Herbstein, 1964). 1.2.3

Molecular compounds

Finally we arrive at the ‘‘molecular compounds.’’ Here the dominant A . . . B interactions can be localized in nature (Hassel, 1972), as in the acetone – bromine charge transfer compound where there is an n–s* interaction (Hassel and Strømme, 1972), or as in the hydrogen-bonded purine and pyrimidine bases in the DNA double helix (Watson and Crick, 1953). Alternatively the A . . . B interaction can be delocalized, as in the –* interactions found in anthracene picric acid (Fritzsche, 1858; Herbstein and Kaftory, 1976). n n n

1.3

Other classifications

Many proposals have been made for the classification and nomenclature of molecular compounds and complexes, and there is much confusion, overlapping and redundancy. We note here other approaches that complement the scheme set out above. It is perhaps too much to hope that any classification and nomenclature scheme will be able to match the complexity of reality and undoubtedly many binary adducts will be found to straddle classificatory boundaries. The terms host, guest, complex and their binding forces were defined in 1977 by Cram and coworkers in the following way: ‘‘Complexes are composed of two or more molecules or ions held together in unique structural relationships by electrostatic forces other than those of full covalent bonds . . . molecular complexes are usually held together by hydrogen bonding, by ion pairing, by -acid to -base interactions, by metal to ligand binding, by van der Waals attractive forces, by solvent reorganizing, and by partially made and broken covalent bonds (transition states) . . . high structural organization is usually produced only through multiple binding sites . . . a highly structured molecular complex is composed of at least one host and one guest component . . . a host–guest relationship involves a complementary stereoelectronic arrangement of binding sites in host and guest . . . the host component is defined as an organic molecule or ion whose bonding sites converge in the complex . . . the guest component is defined as any molecule or ion whose binding sites diverge in the complex (Kyba, Helgeson, Madan, Gokel, Tarnowski, Moore and Cram, 1977).’’ To which was added later: ‘‘In these definitions, hosts are synthetic counterparts of the receptor sites of biological chemistry, and guests the counterparts of substrates, inhibitors or co-factors.’’ The principle of complementarity

HOW MANY BINARY ADDUCTS ARE THERE?

11

was also emphasised: [in order] ‘‘to complex, hosts must have binding sites which cooperatively contact and attract binding sites of guests without generating strong nonbonded repulsions.’’ This very broad definition was directed towards complexes in solution but can well be extended towards crystalline complexes. Lehn (1988) has introduced the term ‘‘supramolecular chemistry . . . defined as ‘chemistry beyond the molecule’ bearing on the organized entities of higher complexity that result from the association of two or more chemical species held together by intermolecular forces.’’ The partners of a supramolecular species are called ‘‘molecular receptor’’ and ‘‘substrate,’’ in analogy to Cram’s use of ‘‘host’’ and ‘‘guest.’’ The complementarity of a receptor for a given substrate, leading to molecular recognition, depends on energetic (electronic) as well as geometrical features, and extends the celebrated ‘‘lock and key’’ steric fit concept enunciated by Emil Fischer (1894).3 A rather elaborate proposal for the classification and nomenclature of ‘‘host–guest-type compounds’’ has been made by Weber and Josel (1983), based on criteria of host–guest type and interaction, the topology of the host–guest aggregate and the number of components in the aggregate. The nomenclature used by Weber and Josel is rather different from our present proposal; for example the urea-n-paraffin tunnel inclusion complex is a tubulato-clathrate and the graphite-potassium lamellar intercalate is an intercalatoclathrate. Although we have preferred not to adopt the complete Weber–Josel proposal, we have found it convenient to use some of its parts.

1.4 How many binary adducts are there? What proportion of reported crystal structures can be classified as ‘binary (or higher) adducts’? The available statistics are limited to organic (ORG) and metalloorganic (MORG) structures, and refer to the inclusion of water and solvent molecules in such structures (Go¨rbitz and Hersleth, 2000). Using the October, 1998 release of the Cambridge Structural Database, they found that 8% of the 77 000 (nonduplicated) entries for organic compounds could be classed as ‘hydrates’ and 7% as ‘solvates’; the corresponding figures for the 91 000 metalloorganic structures were 10.5 and 17%. Clathrates and other molecular compounds and complexes were excluded from the survey, which thus somewhat underestimates the proportion of structures relevant in the context of this book. The proportion of published structures containing co-crystallized organic molecules has risen from very low values before 1950 to 11% (ORG) and 23% (MORG) for the three year period 1995–1997. Which molecule is the most prolific former of adducts? The principal contender appears to be sulfathiazole, which forms more than 100 solvates, plus many related two-component systems (Bingham et al., 2001). Two broad classes have been identified – inclusion phases, in which the main function of the guest is cavity filling, and co-crystals in which the partner molecule forms an essential part of the hydrogen-bonded framework. Some 60 crystals structures have been reported. Cyclotetramethylene tetranitramine (HMX) (George et al., 1965; Selig, 1982) has also been reported to form more than 100 solvates. 3 ‘‘The restricted action of the enzymes on glucosides may therefore be explained by the assumption that only in the case of similar geometrical structure can the molecules so closely approach each other as to initiate a chemical action. To use a picture I would like to say that enzyme and glucoside have to fit together like lock and key in order to exert a chemical effect on each other.’’ (quoted from Lichtenthaler, 1994).

12

1.5


Organic and inorganic supramolecular chemistry

In this book we have placed more emphasis on organic than on inorganic aspects of our subject, although the latter have not been ignored. Some redress of this imbalance can be obtained from the review article of Mu¨ller, Reuter and Dillinger (1995) that, presumably deliberately, has been placed adjacent to Desiraju’s (1995) classic article on Crystal Engineering. As in most of modern chemistry, the distinction between ‘‘organic’’ and ‘‘inorganic’’ will undoubtedly become more and more blurred in the future. References Allegra, G. and Perego, G. (1963). Acta Cryst., 16, 185–190. Barrer, R. M. (1983). J. Incl. Phenom., 1, 105–123. Bingham, A. L., Hughes, D. S., Hursthouse, M. B., Lancaster, R. W., Taverner, S. and Threfall, T. L. (2001). Chem. Commun., pp. 603–604. Brown, J. F., Jr. (1962). Scientific American, 207, pp. 82–92 (July, 1962). Cram, D. J. (1988). J. Incl. Phenom., 6, 397–413 (Nobel Lecture). Davies, J. E. D., Kemula, W., Powell, H. M. and Smith, N. O. (1983). J. Incl. Phenom., 1, 3–44. Desiraju, G. R. (1995). Angew. Chem. Int. Ed. Engl., 34, 2311–2327. Dresselhaus, M. S. and Dresselhaus, G. (1981). Adv. Phys., 30, 139–326. Fischer, E. (1894). Ber. Deutsch. Chem. Gesell., 27, 2985–2993. Fritchie, C. J., Jr. and McMullan, R. K. (1981). Acta Cryst., B37, 1086–1091. Fritzsche, J. v. (1858). J. prakt. Chem., 73, 282–292. Gafner, G. and Herbstein, F. H. (1964). J. Chem. Soc., pp. 5290–5302. George, R. S., Cady, H. H., Rogers, R. N. and Rohwer, R. K. (1965). Ind. Eng. Chem. Prod. Res. Dev. 4, 209–214. Go¨rbitz, C. H. and Hersleth, H.-P. (2000). Acta Cryst., B56, 526–534. Harris, K. D. M. (1993). Chem. Brit., 29, 132–136. Harris, K. D. M. (1997). Chem. Soc. Rev., 26, 279–290. Hassel, O. (1972). ‘‘Structural aspects of interatomic charge-transfer bonding.’’ Nobel Lectures in Chemistry 1963–1970, Elsevier, Amsterdam (1969 Nobel Lecture published in 1972). Hassel, O. and Strømme, K. O. (1959). Acta Chem. Scand., 13, 275–280. Herbstein, F. H. (1993). Acta Chim. Hung. Models in Chemistry, 130, 377–387. Herbstein, F. H. and Kaftory, M. (1976). Acta Cryst., B32, 387–396. Herbstein, F. H., Kapon, M. and Wasserman, S. (1978). Acta Cryst., B34, 1613–1617. Hutson, J. M. (1990). Ann. Rev. Phys. Chem., 41, 123–154. Jeffrey, G. A. (1984). ‘‘Hydrate inclusion compounds,’’ in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Academic Press, London, Vol. 1, pp. 135–190. Ketelaar, J. A. A. (1958). Chemical Constitution. Elsevier, Amsterdam, p. 363. Kistenmacher, T. A., Phillips, T. E. and Cowan, D. O. (1974). Acta Cryst., B30, 763–768. Kyba, E. P., Helgeson, H. C., Madan, K., Gokel, G. W., Tarnowski, T. L., Moore, S. S. and Cram, D. J. (1977). J. Am. Chem. Soc., 99, 2564–2571. Lehn, J. M. (1988). J. Incl. Phenom., 6, 351–396 (Nobel Lecture). Lichtenthaler, F. W. (1994). Angew. Chem. Int. Ed. Engl., 33, 2364–2374. Lu, C.-S., Hughes, E. W. and Giguere, P. A. (1941). J. Am. Chem. Soc., 63, 1507–1513. Mu¨ller, A., Reuter, H. and Dillinger, S. (1995). Angew. Chem. Int. Ed. Engl., 34, 2328–2361. Pedersen, C. J. (1988). J. Incl. Phenom., 6, 337–350 (Nobel Lecture). Powell, H. M. (1948). J. Chem. Soc., pp. 61–73.

REFERENCES

13

Powell, H. M. (1984), ‘‘Introduction,’’ in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies, and D. D. MacNicol, Academic Press, London, Vol. 1, pp. 1–28. Powell, H. M. and Wait, E. (1958). J. Chem. Soc., pp. 1866–1872. Ramprasad, D., Lin, W.-K., Goldsby, K. A. and Busch, D. H. (1988). J. Am. Chem. Soc., 110, 1480–1487. Ricci, J. E. (1966). The Phase Rule and Heterogenous Equilibrium. Dover, New York (1951, reprinted 1966), Chapter 1. Selig, W. (1982). Propell. Explos., 7, 70–77. Shinohara, H., Sato, H., Ohkoohchi, M., Ando, Y., Kodama, T., Shida, T., Kato, T. and Saito, Y. (1992). Nature, 357, 52–54. Tamres, M. L. and Strong, R. L. (1979). Mol. Assoc., 2, 331–456. Watson, J. D. and Crick, F. H. C. (1953). Nature, 171, 737–738. Weber, E. and Josel, H.-P. (1983). J. Incl. Phenom., 1, 79–85.


Chapter 2 Historical outline

Fir’d at first sight with what the Muse imparts, In fearless youth we tempt the heights of Arts, While from the bounded level of our mind, Short views we take, nor see the lengths behind; But more advanc’d, behold with strange surprize New distant scenes of endless science rise! So pleas’d at first the tow’ring Alps we try, Mount o’er the vales, and seem to tread the sky, Th’ eternal snows appear already past, And the first clouds and mountains seem the last: But, those attain’d, we tremble to survey The growing labours of the lengthened way, Th’ increasing prospect tries our wand’ring eyes, Hills peep o’er hills, and Alps on Alps arise. Alexander Pope: An Essay on Criticism

Summary: The history of our subject goes back to the beginning of the nineteenth century, although the compounds concerned remained curiosities, outside the main stream of the development of chemistry, until the first structures were determined by X-ray diffraction in the 1940s. We are now immersed in an explosion of interest and application, ranging from condensed-matter physics to structural biology.

Having defined our terms in Chapter 1, we can now sketch out the historical background. The first molecular complexes to be reported had illustrious parentages. The hydrate of sulphur dioxide appears to have been prepared in 1777–1778 by Joseph Priestley, while chlorine hydrate was reported in the following terms by Sir Humphry Davy in 1811 ‘it is generally stated in chemical books, that oxymuriatic gas [the original name of chlorine – F.H.H.] is capable of being condensed and crystallized at low temperature; I have found by several experiments that this is not the case. The solution of oxymuriatic gas in water freezes more readily than pure water, but the pure gas dried by muriate of lime undergoes no change whatever, at a temperature of 40 below 0 of FAHRENHEIT. The mistake seems to have arisen from exposure of the gas to cold in bottles containing moisture’ (Davy, 1811).

The complex was assigned the composition 10H2OCl2 by Michael Faraday (1823); the currently preferred formulation is about 66H2OCl2, the exact composition depending on conditions of preparation (see Chapter 7).

16

HISTORICAL OUTLINE

The first molecular compound to be reported was quinhydrone (hydroquinone: p-benzoquinone) by Wo¨hler (1844), who noted its unexpected colour. Continuation of these studies led Wo¨hler (1848) to the preparation of the first clathrates, with quinol (hydroquinone) as host and H2S as guest; the reported compositions of 4(quinol)H2S and 3(quinol)H2S are close to modern values. The SO2 clathrate was reported 10 years later (Clemm, 1859); Mylius (1886), who prepared the CO, HCN and formic acid clathrates somewhat later, suggested possible enclosure of guests by quinol without chemical combination. The picrates of benzene, naphthalene and anthracene were also prepared in the middle of the nineteenth century (Fritzsche, 1858), at a time when the atomic weight of carbon was still taken as 6. The first tunnel (channel) inclusion complex was perhaps the 2(thiourea)diethyl oxalate complex reported by Nencki (1874), although this would appear to require checking. There followed many disparate observations, unrelated to main currents of contemporary organic chemistry and perplexing from a structural point of view. The classical period of development culminated with the publication of the second edition of Paul Pfeiffer’s Organische Moleku¨lverbindungen (1927, First Edition 1921), which remains a mine of useful factual information although some of the structural ideas are fanciful in terms of current knowledge (Fig. 2.1). The modern period was inaugurated by the application of X-ray diffraction methods to the determination of key crystal structures. The first of these appears to have been Phenochinon OH

O

H

H

H

H5C6O H

H H

H H5C6O

H H OC6H6

OC6H5

H O

OH

Enolformen

H

O

Ketoformen

H

O H

H

H

H

OH H

H

O H

H

H

HO

H

H

H

H

O H

O

H O Chinhydron

H

Bei dem heutigen Stand der Chromophortheorie bleibt aber – entgegen der Ansicht Posners – die tiefe Farbe der Chinhydrone bei dieser Formulierung ganz ra¨tselhaft; auch die leichte Spaltbarkeit der Chinhydrone in ihre Komponenten, schon durch sog. indifferente Lo¨sungsmittel, la¨ßt sich schwer mit der Posnerschen Theorie in Einklang bringen. Fig. 2.1. Pfeiffer’s representation of Posner’s (1904) suggestion for the structures of phenoquinone and quinhydrone. From the short quotation attached, it is clear that Pfeiffer was not an enthusiastic supporter of the proposal. The diagram has been copied from p. 276 of Pfeiffer’s book (1927).

HIS T OR I C AL OUT L IN E

17

{{[(CH3)3AsPdBr2]2}-[dioxane]} (Wells, 1938), which is a tunnel inclusion complex. However, the real breakthrough came during and soon after the end of the Second World War when H. M. Powell and coworkers at Oxford reported the structures of p-iodoaniline 1,3,5-trinitrobenzene (Powell, Huse and Cooke, 1943), and {{3(quinol)}[CH3OH]} (Palin and Powell, 1945; cf. Davies, 1998), representative of charge transfer molecular compounds and clathrate complexes respectively. Other important structural elucidations of about the same time were those of gas hydrates (Stackelberg, 1949a,b; Pauling and Marsh, 1952), urea and thiourea tunnel inclusion complexes (Smith, 1950; Hermann and Lenne, 1952), and n–s* (localized) charge transfer compounds of halogens with donors containing oxygen or nitrogen (Hassel and Rømming, 1962). The important theoretical studies of charge transfer compounds by Mulliken (1952a,b) were complemented by thermodynamic and statistical-mechanical studies of clathrates (van der Waals and Plateeuw, 1959). The current period is distinguished by a number of themes. Firstly, there is extensive activity in the area of crystal structure analysis, which is the major experimental tool. The adducts range in size and complexity from the combination of small organic molecules found in hyperol (urea:hydrogen peroxide) (Lu, Hughes and Giguere, 1941) to complexes of large biomolecules, with the complex of lysozyme and tri(N-acetylglucosamine) (Phillips, 1966) as an early example now far surpassed in complexity. Secondly, there is considerable study of interactions between components using a wide range of spectroscopic techniques. Thirdly, there has been renewed interest in physical properties, especially electrical conductivity, and in theoretical explanations for the various types of physical behaviour that have been found. A major gap in current knowledge, in regard to both theory and experimental data, relates to the energetics and thermodynamics of the interactions between the components. One must also note that an important new direction of investigation has developed with the explosive growth of the branch of host-guest chemistry which we have called ‘‘moieties within molecules,’’ the moieties being ions as well as molecules. As we noted in Chapter 1, the importance of these substances rests on their occurrence in solution as well as in the crystalline state. Perhaps the first examples to be studied structurally as well as chemically were the cyclodextrin inclusion complexes (Cramer, 1954), which can enclose many different types of moiety within the hydrophobic interior of the doughnut-shaped molecule. The antibiotics such as enniatin constitute a somewhat similar group. These two sorts of host compound are natural in origin; the first purely synthetic examples, the macrocyclic ‘‘crown’’ ethers, were reported by Pedersen (1967) and since then this has become one of the most rapidly growing areas in the general field of inclusion complexes. The crown ethers were followed by the cryptands and many other variations on this theme and there seem to be few limits to the ingenuity of the organic chemist in the tailoring of particular hosts for the inclusion of specific guests. There have been many applications in analytical chemistry, synthetic organic chemistry and in biochemistry. The analogies to the behaviour of many biomolecules is striking and the phrases ‘‘molecular recognition’’ and ‘‘supramolecular chemistry’’ have become established in the literature (Lehn, 1995; Desiraju, 1995; Nangia and Desiraju, 1998). An important potential contribution of a study of crystalline supramolecular systems (‘‘binary adducts,’’ in a more old-fashioned language) is to provide detailed structural information, both static and dynamic, leading to an understanding of the interactions which are fundamental to molecular recognition, and n n n

18

HISTORICAL OUTLINE

thus hopefully, to the enhancement of our capabilities as molecular engineers, designing desired structures from first principles. The ability to form binary adducts is not limited to small molecules and considerable progress has been made in preparing adducts of large biomolecules, an early example of which has already been mentioned (Phillips, 1966). An important application is in the area of design of drugs that have a capability of recognising receptors in proteins and DNA. Both small-molecule model compounds and biomolecule complexes are subjects of active study. Biomolecule complexes of various kinds are hardly mentioned in this book – their variety and importance demand a book in its own right. However, there is no reason to believe that the interactions involved are fundamentally different from those described here. Highlights in the historical development of the scientific study of binary adducts are summarized in Table 2.1 (references are given in the body of the text). The task of this book is to weave these varied themes into whole cloth in as coherent and cohesive a manner as possible. Table 2.1. Some highlights in the study of binary adducts. The dates are only approximate and have generally been chosen to indicate publication of a particularly significant paper or book, or to mark some special event. Usually the contributions of the authors cited (and their coworkers) extend over many years Approximate date

Author(s)

Achievement

1777–8 1811 1823 1841 1849 1858

Joseph Priestley Humphry Davy Michael Faraday C. Schafthaült F. Wo¨hler J. von Fritzsche

1891

A. Villiers

1893 1897

H. W. Pickering A. W. Hofmann

1916

H. Wieland and H. Sorge

1926 1927

J. Martinet and L. Bornand P. Pfeiffer

1930

E. Hertel

1938

A. F. Wells

1940

M. F. Bengen

First observation of a gas hydrate (of SO2). Observation of the gas hydrate of Cl2. Analysis of the gas hydrate of Cl2. Preparation of graphite intercalates. Preparation of quinol clathrate of H2S. Preparation of first mixed-stack donor–acceptor compounds (benzene, naphthalene and anthracene with picric acid). Preparation of cyclodextrin inclusion complexes. Preparation of alkylamine hydrates. Preparation of nickel ammonium cyanide inclusion complex of benzene. Preparation of choleic acid inclusion complexes. Qualitative donor–acceptor theory of –* molecular compounds. Second edition of Organische Moleku¨lverbindungen. Early crystallographic studies of molecular compounds. Crystal structure of {{[(CH3)3AsPdBr2]2} [dioxane]} tunnel inclusion complex. Preparation of urea-hydrocarbon tunnel inclusion complexes.

19

REFERENCES

Table 2.1. (Continued) Approximate date

Author(s)

Achievement

1943

H. M. Powell

1945

H. M. Powell

1949 1950–2

1964

G. Briegleb A. E. Smith; C. Herrman and H.-U. Lenne M. von Stackelberg; L. Pauling and R. E. Marsh; W. F. Claussen. G. A. Jeffery, Yu. A. Dyadin, and their schools J. D. Watson and F. H. C. Crick F. Cramer W. Saenger; K. Harata J. H. van der Waals and J. C. Plateeuw L. Mandelcorn (editor)

Crystal structure of p-iodoaniline 1,3,5trinitrobenzene. Crystal structure of quinol clathrate of CH3OH. Spectroscopic studies of binary adducts. Crystal structures of urea-hydrocarbon tunnel inclusion complexes. Crystal structures of gas hydrates.

1966

R. S. Mulliken

1966 1969

D. C. Phillips O. Hassel

1983

J. L. Atwood and J. E. D. Davies (editors) D. J. Cram, J.-M. Lehn, C. J. Pedersen. J.-M. Lehn (chair, editorial board)

1951

1953 1954

1959

1987 1996

n n n

Further development of crystal chemistry of gas hydrates and related complexes. Structure of DNA (purine/pyrimidine hydrogen-bonded molecular compound). Publication of Einschlussverbindungen. Crystallography of cyclodextrin complexes. Statistical mechanics of clathrates. Publication of Non-Stoichiometric Compounds. Nobel Prize in Chemistry (inter alia theory of charge transfer interactions). Crystallography of lysozyme complexes. Nobel Prize in Chemistry (crystal structures of localized donor–acceptor molecular compounds). First issue of J. Inclus. Phenom. Nobel Prize in Chemistry (development of supramolecular chemistry). Publication of Comprehensive Supramolecular Chemistry in 11 volumes.

References Clemm, A. (1859). Ann. Chem., 110, 345–349. Cramer, F. (1954). Einschlussverbindungen. Springer, Heidelberg. Davies, J. E. D. (1998). J. Incl. Phenom. and Mol. Recogn. Chem., 32, 499–504. Davy, H. (1811). Phil. Trans. Roy. Soc., 101, 1–35, (see p. 30). Desiraju. G. R. (1995). Angew. Chem. Int. Ed. Engl., 34, 2311–2327. Faraday, M. (1823). Quart. J. Sci. Lit. and Arts, 15, 71–74. Fritzsche, J. von, (1858). J. prakt. Chem., 73, 282–292.

20

HISTORICAL OUTLINE

Hassel, O. and Rømming, C. (1962). Quart. Rev., 16, 1–18. Hermann, C. and Lenne, H.-U. (1952). Naturwiss., 39, 234–235. Lehn, J.-M. (1995). Supramolecular Chemistry – Concepts and Perspectives, VCH, Weinheim. Lu, C.-S., Hughes, E. W. and Giguere, P. A. (1941). J. Am. Chem. Soc., 63, 1507–1513. Mulliken, R. S. (1952a). J. Am. Chem. Soc., 72, 600–608. Mulliken, R. S. (1952b). J. Phys. Chem., 56, 801–822. Mylius, F. (1886). Chem. Ber., 19, 999–1009. Nangia, A. and Desiraju, G. R. (1998). Acta Cryst., A54, 934–944. Nencki, M. (1874). Ber. Deut. Chem. Gesell., 7, 779–780. Palin, D. E. and Powell, H. M. (1945). Nature, 156, 334–335. Pauling, L. and Marsh, R. E. (1952). Proc. Nat. Acad. Sci., 38, 112–118. Pedersen, C. J. (1967). J. Am. Chem. Soc., 89, 7017–7036. Pfeiffer, P. (1927). Organische Moleku¨lverbindungen. Enke, Stuttgart, 2nd Edition. Phillips, D. C. (1966). Scientific American, pp. 78–90 (November, 1966). Powell, H. M., Huse, G. and Cooke, P. W. (1943). J. Chem. Soc., pp. 153–157. Smith, A. E. (1950). J. Chem. Phys., 18, 150–151. Stackelberg, M. von, (1949a). Naturwiss., 36, 327–333. Stackelberg, M. von, (1949b). Naturwiss., 36, 359–362. Waals, J. H. van der, and Platteeuw, J. C. (1959). Adv. Chem. Phys., 51, 1–59. Wells, A. F. (1938). Proc. Roy. Soc. Lond. A, 167, 169–189. Wo¨hler, F. (1844). Ann. Chem., 51, 145–163. Wo¨hler, F. (1849). Ann. Chem., 69, 294–300.

Part II Moieties within molecules


Introduction to Part II Moieties within molecules

Chemistry is going to move into bigger molecules and new types of reactions. And the chemist is going to have to pay far more attention than he or she has done in the past to noncovalent bonding. Hitherto the chemist has grown accustomed to hydrogen bonding. But other forces will have to be considered, forces that make these big molecules adopt certain shapes . . . I think there is great hope for the future in work going on in the inclusion of small molecules in big ones. Lord Alexander Todd Chemical and Engineering News October, 1980

The unifying theme of Part II is that the binary adducts considered exist in solution as well as in the crystalline state – the guest moieties are enclosed within the confines of the host molecule and held in tight enough embrace to maintain the integrity of the combination despite the potentially disintegrative buffeting forces of the solvent molecules. Therefore our choice of title – moieties within molecules. Alternative terms, meeting with growing acceptance, are ‘‘supermolecule’’ (noun) and ‘‘supramolecular’’ (adjective). The difference between ‘‘supermolecules’’ and the ‘‘molecular compounds and complexes’’ considered in the later sections of this book is that the latter have at most a transient, if any, existence in solution. The supermolecules can have a variety of chemical forms, with the most useful distinction being made between two major geometrical groups – the rings and the threedimensional cages. However, even rings tend to wrap around guests as much as possible. Guests are chosen to match the geometrical shapes of the cavities of the hosts thus imparting selectivity to the systems. This is one aspect of another theme of currently growing importance – that of molecular recognition, which also finds its expression in the matched hydrogen-bonded systems discussed especially in Chapter 12. Perhaps 90% of the binary adducts described in this book have been obtained by chance rather than deliberate design, and this includes those dealt with in Chapters 4 and 5. Not so for the subjects of Chapter 3 – crown ethers, cryptands and other types of enclosure hosts – for these are adducts where the host molecule has been designed to fulfill the function of enclosing within its boundaries a guest of a particular kind. However, the first steps in this direction taken by C. J. Pedersen in the 1960s were not designed but serendipitous, with chance favoring his prepared mind. Pedersen, investigating the effects of bi- and multidentate phenolic ligands on the catalytic properties of the vanadyl group VO, obtained a small quantity (0.4%) of an unknown material which trapped sodium ions. His conclusion – ‘‘thus did I discover dibenzo-18-crown-6, the first crown ether and the first neutral synthetic compound capable of complexing the alkali metal cations’’ – has an almost Biblical ring of triumph (Pedersen, 1988). This discovery sparked off many

24

MOIETIES WIT HIN MOLECULES

investigations, some deepening our knowledge and understanding of the crown ethers, and others, especially those of Lehn (1988) and Cram (1988), moving off in new directions. Formation of complexes is most efficient if the host has a built-in conformation adapted to enclosure of the guest, but this is not a sine qua non for there are many examples showing considerable differences between the conformation of the neat host and that taken up in the complex. Thermodynamic measurements (generally carried out by NMR techniques) on the host–guest combination in solution provide fundamental information about the stability of the complexes formed, and the energetics of the formation and break-up of the host–guest combinations. Both rings and three-dimensional cages have been synthesized, and the guest species include metallic cations, a few anions, organic cations and anions and neutral molecules. We devote most attention to the latter. An informative survey of synthetic supramolecular chemistry has been given by Fyfe and Stoddart (1997). In Chapter 4 (Cyclodextrins and their complexes) we consider a particular group of host molecules – the , and g cyclodextrins – which have toroidal forms of three different sizes; these are natural products obtained from the degradation of starch. Because the host does not entirely enclose the guest, these complexes are less stable in solution than the enclosure species considered in Chapter 3 and we pay more attention to the varieties of arrangement found in the crystals. Cyclodextrins have found widespread use in the chemical and pharmaceutical industries. In Chapter 5 (DNA and its complexes) we consider a limited group of the inclusion complexes formed by another type of host molecule obtained from a natural product – DNA. The complexes of DNA with other biological molecules such as proteins constitute an area of research which seems at present to have limits neither to its size nor its importance but is too large for discussion in a single chapter of this book. For this reason we have limited ourselves to considering complexes of ‘‘small’’ molecules with DNA oligomers of limited length and various compositions. These complexes are essentially of two different kinds – those formed by intercalation of guests between the base pairs of the double helix, and those formed by inclusion of the guests in the small or large grooves on the periphery of the helical DNA fragment. Maverick and Cram (1996) have made some interesting remarks on the role of crystal structure analysis in the prosecution of the type of research described in these chapters (and elsewhere). We quote (with permission) an abbreviated version: ‘‘The . . . analyses presented many problems . . . the solvents used for isolation and purification . . . tend to disorder in the crystals . . . Almost without exception, crystals of carcerands, empty or complexed, were weak diffracters, resulting in very poor data-to-parameter ratios. Often the symmetry of the host cavity and . . . of the guest were incompatible, resulting in inaccuracies just where the greatest accuracy is required. For these reasons, the structures presented . . . are crude by some crystallographic standards.’’ The reader will find a comparison with the even more severe situation described in Chapter 9 (Intercalation Complexes) instructive. Nevertheless, ‘‘Crystal structure determinations . . . were of great importance [because] (i) . . . they were the . . . final criteria for our success in synthesizing these new types of complexes. (ii) [they] provided a wealth of information about preferred conformations . . . [suggesting] explanations for some of the observed binding phenomena . . . (vi) Crystal structures provide a simple, direct and independent means of convincing yourself and skeptics that ‘you know what you are doing.’’’ Few changes are required to make these remarks applicable to many analogous situations described in this book.

REFERENCES

25

Most of the molecules considered in this section have been described in a review volume edited by Semlyen (1997).

References Cram, D. J. (1988). J. Incl. Phenom., 6, 397–413. Fyfe, M. C. T. and Stoddart, J. F. (1997). Accts. Chem. Res., 30, 393–401. Lehn, J.-M. (1988). J. Incl. Phenom., 6, 351–396. Maverick, E. Cram, D. J. (1996). ‘‘Carcerands and Hemicarcerands: Hosts that imprison molecular guests,’’ in Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, Vol. 2, pp. 367–418. Pedersen, C. J. (1988). J. Incl. Phenom., 6, 337–350. Semlyen, J. A. (Editor) (1997). Large Ring Molecules, Wiley, Bognor Regis.


Chapter 3 The enclosure species – crown ethers, cryptands and related molecules – as hosts

‘‘A gentle guest, a willing host, Affection deeply planted – It’s strange how those we miss the most Are those we take for granted.’’ Sir John Betjeman ‘‘The Hon. Sec.’’ (1986)

Summary: Crown ethers are prototype examples of host molecules of the ring type where small ring sizes interact with metallic cations to form widely studied ion-molecule complexes, dealt with here only in passing, while the larger ring sizes interact with neutral molecules and organic cations of various types to form intramolecular inclusion complexes, rotaxanes and even catenanes. Although formally two-dimensional rings, the actual conformations taken up by the larger crown ethers are more complicated and include the formation of molecular clefts. The overall shapes of threedimensional cage molecules are approximately ellipsoidal; the cages are closed to lesser or greater extents, thus permitting ingress and egress of guest molecules tailored in size and shape to match the available portals and cavities. Thermodynamic measurements in solution, principally using NMR methods, provide insight into the energetics of these processes. Although considerable ingenuity has been expended in the synthesis of the hosts, these are more easily obtainable than might be imagined at first thought, and this holds great promise for their widespread use in the future. A crowning achievement of this type of host–guest chemistry is the synthesis of stable but reactive cyclobutadiene incarcerated within the cavity of a hemicarcerand, followed by the analogous preparation of o-benzyne.

3.1 Introduction 3.2 Doubly bridged cyclophanes and analogous molecules as hosts for intramolecular guests 3.3 Cleft molecules as hosts 3.3.1 Single-cleft hosts 3.3.2 Double-cleft hosts 3.4 Container molecules as hosts 3.4.1 Introduction 3.4.2 Cavitands and caviplexes 3.4.3 Hemispherands and hemispheraplexes 3.4.4 Triply bridged cyclophanes and analogous molecules as three-dimensional hosts for intramolecular guests 3.4.5 Spherands and spheraplexes 3.4.6 Carcerands and carceplexes

28 30 44 44 47 48 48 48 50 51 59 59

28

CROWN ETHERS, CRYPTANDS AND RE LATED MOLECUL ES AS HOST S

3.5

Hemicarcerands and hemicarceplexes 3.5.1 Overview 3.5.2 The taming of cyclobutadiene, and of o-benzyne 3.5.3 Molecular mechanics and dynamics studies on the complexation and decomplexation processes 3.6 Comparisons of concepts References

3.1

61 61 64 66 67 68

Introduction

Intramolecular enclosure of guests by hosts, often specifically and purposefully designed, has progressed remarkably rapidly over the last thirty years, spearheaded by the efforts of Pedersen, Lehn, Cram and their coworkers, and formally recognized by the joint award of the Nobel Prize for Chemistry to the trio in 1987. Many different host systems have been synthesized and the strengths of host–guest interactions in solution have been measured, and an understanding developed of the factors contributing to these interactions. In parallel, the determination of a number of crystal structures has led to establishment of a sounder geometrical foundation for the assessment of the various contributions to host– guest interaction. All this has been accompanied by growth of a nomenclature and jargon, especially as the systematic names of most of the host compounds are very complicated. We first introduce representatives of the main types of host (Fig. 3.1) and then discuss the structures of the complexes formed, emphasizing crystallographic results and noting synthetic methods and studies of solution thermodynamics only peripherally. The hosts are given the suffix ‘‘-and’’ (corand – often still called ‘‘crown ethers’’ – cryptand, . . . etc.) and the corresponding complexes are sometimes given the suffix ‘‘-ate’’ and are coronates, cryptates, . . . etc., or else the suffix ‘‘-plex’’ and become caviplexes, hemispheraplexes, . . . etc.). A distinction has been made between cavitands, which are hosts containing enforced cavities, and speleands, where the host combines the elements of a rigid, lipophilic cavity with polar binding sites. The classical origins of these names is obvious: Latin corona wreath, crown; Greek krupto hide; Greek speliaon cave; Latin cavus hollow; Latin carcer prison. When considering the formation of complexes in solution, two important factors are generally taken into account. These are (i) the principle of stereo-electronic complementarity between host and guest, and (ii) the principle of preorganization of a binding site before complexation. The first is often stated to be a modern formulation of Emil Fischer’s ‘‘lock and key’’ principle (Behr, 1994). It implies that there is some special interaction, perhaps charge transfer or hydrogen bonding, between host and guest, as well as geometrical complementarity. The second has been defined by Cram and coworkers in a long series of elegant investigations which we describe in more detail below. Essentially this principle states that binding sites are best organized to be complementary in a stereoelectronic sense prior to complexation; if not, then reorganization is necessary on complexation and the price paid in free energy may outweigh the gain achieved by complexation. If so, then complexation will not take place. In general, complexation requires replacement of solvent molecules by guest molecules. Preorganization also enhances selectivity by introduction of particular chemical features into an host molecule enhancing

29

I NT RO D UC T I O N

O O

O

O

O

N O

H3C

O

O

H

H

N

O O

O Corand (crown ether; 18-crown-6)

n

O

CH3 Podand

Cryptand

O O O

N

O X

X

O O

X

X

H3C

CH3

X

N X

H3C

CH3

CH3

CH3

Hemispherand (X = OCH3)

Cryptaspherand (X = OCH3) X

(H2C)n

X

(CH2)n

X

Doubly bridged cyclophanes CH3 CH3

O S

CH3

Si O

O Si

O CH3

CH3

CH3

CH3

O

O

O

O Si

S O

O

O

O O

O

S

O

O

CH3

CH3

Cavitand

O

O

O

CH3

O

CH3

O

O CH3 Si

Triply bridged cyclophanes

O

S O

Carcerand

Fig. 3.1. Representatives of the main types of host classified in chemical terms. It is important to note that the types of complex formed depend on the nature of the guest and, most importantly, on the geometry as well as the chemistry of the host. The filled circles in the carcerand represent methyl groups.

its ability to form complexes with special guests. However, as we shall see below, there can often be a considerable change of shape in an host molecule as it changes from uncomplexed to complexed conformation. Such comparisons require crystal structures of the neat host as well as those of the complex, and some are indeed available. In crystalline complexes it is also necessary to consider the possibility of intermolecular complexation of guest and/or solvent as an additional stabilizing factor.

30


Analysis of a crystal structure shows in detail the host–guest relationship in the solid state but it is often not clear to what extent this geometrical situation persists in long-lived fashion in solution. Thermodynamic and spectroscopic (especially NMR methods) studies of the solutions provide evidence to probe these questions. Here we draw special attention to five massive compilations of solution thermodynamic data, which also provide an exhaustive listing of the structural formulae of macrocyclic hosts. These deal with intramolecular complexation of metallic cations and anions (Christensen, Eatough and Izatt, 1974; 217 references; Izatt, Bradshaw, Nielsen, Lamb, Christensen and Sen, 1985; 340 references), organic cations and anions (Izatt, Pawlak, Bradshaw and Bruening, 1991; 1173 references), and neutral molecules (Izatt, Bradshaw, Pawlak, Bruening and Tarbet, 1992; 307 references); this information has been revised and extended (Izatt, Pawlak, Bradshaw and Bruening, 1995; 478 references). There does not seem to be any comparable material dealing with the thermodynamics of the solid complexes and their components, which would be even more relevant here in the context of our emphasis on the solid state. It will be observed that we pay much more attention in this chapter to thermodynamic parameters derived from measurements in solution than we do in most other chapters, where the emphasis is placed on the thermodynamics of the crystalline complexes. The reason is that the thermodynamic measurements refer to the process of intramolecular complexation and this will, we presume, be little altered if it occurs in solution or in the solid state. We start this chapter by considering intramolecular enclosure systems where the host molecules are perhaps more nearly two- rather than three-dimensional; enclosure is often less complete with two-dimensional hosts. We then proceed to the complete enclosure found with three-dimensional hosts. Our classification is largely based on the shapes taken up by the host molecules and thus crosses boundaries of chemical type. We use a mnemonic nomenclature – the supramolecular macrocyclic complex is denoted by parentheses{M[X]}, with the included guest within square brackets [X]. This hopefully lessens the confusion, especially when counterions and solvent molecules are present in the crystal but not included within the confines of the host. There are a number of excellent reviews (Cram and Cram, 1994; Jasat and Sherman, 1999; Hof, Craig, Nuckolls and Rebek, 2002; Rudkevich, 2002) which describe the historical development of the field, the synthetic aspects and the solution chemistry; these are complementary to the material given in this chapter, where the crystal-chemical aspects are emphasized.

3.2

Doubly bridged cyclophanes and analogous molecules as hosts for intramolecular guests

Background to the use of cyclophanes as hosts for the enclosure of neutral molecules has been given (Diederich, 1988; Odashima and Koga, 1983). The necessary, but not always sufficient, condition for intramolecular inclusion of a guest species within the internal cavity of a suitable host molecule is that the cavity should be large enough to accommodate the guest. The classical studies of crown ether complexes deal mostly with those having inorganic cations as guests (Dobler, 1981; Hilgenfeld and Saenger, 1982; Bradshaw, Izatt, Bordunov, Zhu and Hathaway, 1996); we shall not make more than passing reference to these for

DOUBLY BRIDGED CYCLOPHANES AND ANALOGOUS MOLECULES

31

reasons of space. Many complexes of the smaller crown ethers, especially 18-crown-6, with neutral molecules are known, but crystal structure analysis shows that these are intermolecular complexes with hydrogen bonding between the components and not intramolecular inclusion complexes; many are discussed in Chapter 12. From a consideration of the available crystallographic results, Uiterwijk et al. (1986) concluded that ‘‘urea and urea analogs like uronium and guanidinium cations can only form encapsulated complexes with complementary H-bonding schemes if the crown ether has at least 27 ring atoms. With smaller rings, perching complexes are formed.’’ It may well be posssible to generalize this conclusion to other potential guests. An example of complexation by a derivative of 18-crown-6 (3.1) presents a cautionary tale. This forms a highly stable 1 : 1 complex with nitromethane in which the methyl hydrogens are directed towards the emphasised N and O atoms of 3.1 (Weber, Franken, Puff and Ahrendt, 1986; DIZTIP). The difficulties of definition and interpretation appear when the stereoview of the nitromethane environment (Fig. 3.2) is examined. Firstly, the host molecule is not planar and indeed approximates to the cleft shape discussed in Section 3.3 below; secondly, the guest is not entirely enclosed within the host but rather between a group of four host molecules. One should beware of drawing conclusions from projections onto the mean plane of possibly nonplanar molecules. An interesting feature of {3.1[nitromethane]} is that spontaneous resolution has taken place on crystallization, the space group of the complex being P212121; spontaneous resolution does not occur in the two polymorphs of the neat host (TAFYOO, CARXOI), nor in the complexes with acetonitrile (FUCFOY), phenyl cyanide (YURROS), or in the bis(methanol) monohydrate complex (CARXOU). Chloroacetonitrile forms a 1 : 1 complex while the acetonitrile complex has a 1 : 2 host : guest ratio, both complexes being less stable than the nitromethane complex. Dimethylformamide, dimethyl sulphoxide, acetone, benzene and toluene do not form complexes.

N O

O

O

O O

O

3.1

Fig. 3.2. The formula of 3.1 (the N and two O atoms involved in bonding to the methyl hydrogens are emphasised) is shown, and a stereoview of the surroundings of the nitromethane molecule in the crystal of {3.1[nitromethane]}. The host molecules are shown as line drawings and the guest in space-filling representation. (Adapted from Weber, Franken, Puff and Ahrendt, 1986.)

32


The larger crown ethers1 can and do act as hosts in intramolecular inclusion complexes; a wide range of host shapes and degrees of flexibility can be attained by choice of suitable compositions. Most of the published results refer to (PF6) salts in which the guests are the dications bipyridinium (3.2), diquat (3.3) and paraquat (3.4). One series of publications carries the intriguing title ‘‘Molecular Meccano’’ (Anelli et al., 1992; Part 2, Amabilino, Ashton et al., 1995) and another invokes the name of ‘‘Molecular Lego’’ (Kohnke, Mathias and Stoddart, 1989), both trade names of children’s toys in which complicated structures are constructed by combination of a limited number of simpler units. We shall summarize some of the structural results, starting with smaller host molecules such as bis-m-phenylene-32-crown-10 (C28H40O10; BMP32C10; 3.5; Allwood, Shahriari-Zavareh, Stoddart and Williams, 1987). Note that we shall often use the acronyms of the original papers instead of systematic names. The neat compound (FIKWAX) crystallizes in space group P21/a (Z ¼ 2) and the molecules are centrosymmetric, with ˚ 2. Minimal an open conformation having a free central passage of dimensions 7.8 4.9 A rotations about four single bonds are required to extend the central passage so that a diquat cation can be complexed intramolecularly; the crystals have composition {C28H40O10[C12H12N22þ]}(PF6)2(CH3)2CO (space group P1, Z ¼ 2; FIKWEB). The conformations of uncomplexed and complexed hosts are compared in Fig. 3.3; the two pyridinium rings of the diquat cation have a twist angle of 22 .

NH+

+HN

Bipyridinium (3.2)

N+

N+

Diquat (3.3)

N+–CH3

H3CN+

Paraquat (3.4)

Scheme 3.1

Bis-p-phenylene-34-crown-10 (BPP34C10; 3.6; the formula is shown in Fig. 3.4(a) below2) also forms intramolecular inclusion complexes (Allwood, Spencer, ShahriariZavareh, Stoddart and Williams, 1987a). The neat compound (FIKVEA) crystallizes in space group P21/c (Z ¼ 4), with two molecules of somewhat different conformation at independent centres of symmetry; both conformations are open rather than self-filling. This is in contrast to the self-filling conformation found in tetramethoxy-BPP34C10 (Owen, 1984; CIDLOQ, C2/c, Z ¼ 4). The second conformation (designated II) of the 1 The crown ethers are named as n-crown-m where ‘n’ is the number of atoms in the macrocycle and ‘m’ is the number of (ether) oxygen atoms. Thus 3.5 is 32-crown-10 (or 32C10 for short), the macrocycle having 22 carbons and 10 oxygens. 3.5 contains two m-phenylenes and thus the overall number of carbons is 22 þ (2 3) giving the composition as C28H40O10. The acronym is BMP32C10 where BMP represents bis(m-phenylene) and 32 is the number of atoms (C þ O) in the macrocycle. 2 Although BMP32C10 and BPP34C10 have the same composition (C28H40O10), their macrocycles have different sizes and connectivities.


61 173 173 –174

74 –67 –154

68

N

16 73 167

–63

N

22 –66

77

33

94

–62 –171 164 –168 –67

9

7

175

–154 175 18 –62 64 173

Fig. 3.3. Comparison of the centrosymmetric shape of BMP32C10 (3.5) in its neat crystals (on the right) and in its complex with diquat (on the left); the oxygens are darkened; diquat is twisted by 22 about the central bond. The considerable difference in the two conformations is shown by the torsion angles about the various bonds (only torsion angles differing by >5 from 0, 180 are shown). (Reproduced from Allwood, Shahriari-Zavareh, Stoddart and Williams, 1987.)

neat crystals is unusually open and remains so when it encloses a bipyridinium cation, the terminal N–H groups being hydrogen bonded to opposing oxygens of the crown ether (Ashton, Philp et al., 1991; KOLMAZ). When the para hydrogens of the bipyridinium cation are successively replaced by methyls, n-propyls and n-butyls, then a series of rotaxanes3 is obtained, exemplified by the complex of composition {C28H40O10[C12H14N22þ]}(PF6)22[(CH3)2CO] (Fig. 3.4) in which the BPP34C10 macrocycle is threaded by a paraquat (PQT2þ) cation (P1, Z ¼ 2; Allwood, Spencer et al., 1987a; FIRXOT); the 4,4 0 -bi(2-hydroxyethylpyridinium) complex is KOLMED and the 4,4 0 -bi(2-(2-hydroxyethoxy)ethylpyridinium) complex is KOLMIH. As the conformation of the macrocycle hardly changes from one complex to the next, these macrocycles are examples of almost perfect preorganization. A rather similar arrangement and illustration of preorganization is found in the intramolecular paraquat inclusion complex of 1,5-dinaphtho-38-crown-10, which is stabilized by –* interactions between the electron-rich naphthalene and the electron-poor bipyridinium rings, possibly supplemented by weak electrostatic interactions (Ashton, Chrystal et al., 1987; FUVBAZ; 3 Rotaxanes (rota (L) ¼ wheel, axis (L) ¼ axle) have been defined (Schill, 1971) as molecules in which a cyclic structure is threaded by a chain or other linear subunit having bulky ends that prevent the dissociation (unthreading) of the cyclic and linear components. Pseudorotaxanes (Stoddart, 1991) are supramolecules in which the wheel is free to dissociate from the axle – thus the components are held together by nonbonded interactions rather than the mechanical interference of true rotaxanes. Catenanes (L. catena, chain) are molecules which contain two or more interlocked rings which are inseparable without the breaking of a covalent bond. The number of catenated rings is designated by [n] where n 2. These are all examples of interlocked or intertwined species (Amabalino and Stoddart, 1995). Here we return to a point of nomenclature. In Chapter 1 we required ‘binary adducts’ to be relatively easily separable into their components. In terms of this requirement, pseudorotaxanes are binary adducts while rotaxanes and catenanes are not. However, it is convenient to include here some examples of the latter in order to illustrate particular points.

34


(a)

O O

O

(b)

O

N

O O

O

N O

N

O

O N

O

O

O

O

O

O

O O

O O (c) N

N

Fig. 3.4. Ball and stick (a) and space-filling representations (b) of the centrosymmetric rotaxane complex {BPP34C10[PQT2þ]}(PF6)22(CH3)2CO; the diagrams show two different orientations. The host crown ether 3.6 has an S-shape when viewed from a direction normal to that of (b); (c) the best least squares fit is shown between conformation II of neat BPP34C10 and that found in the paraquat complex. (Parts (a) and (b) reproduced from Anelli et al. (1992) and part (c) from Allwood, Spencer et al., 1987b.)

{C36H44O10[C12H14N2]2þ2(PF6)-2((CH3)2CO)}). Other rotaxanes and pseudorotaxanes are discussed below. The diquat monohydrate complex of BPP34C10 introduces a new feature, with both cation and water molecule located within the torus of the host molecule (Allwood, Spencer, Shahriari-Zavareh, Stoddart and Williams, 1987b; FIKVIE); thus the composition is best represented as {BPP34C10[diquatH2O]}(PF6)2 and the conformation of the host crown ether is again similar to that of conformation II of neat BPP34C10. The crystal structure of {BPP31C9[diquat]}(PF6)2 also shows intramolecular inclusion of diquat by the crown ether host (Ashton, Slawin, Spencer, Stoddart and Williams, 1987; FIRXUZ). The diquat cations are both twisted, as noted above. The crystal structures of neat crown ethers of general formula BPP(3n þ 4)Cn have been determined (Slawin, Spencer, Stoddart and Williams, 1987) for n ¼ 7, 8, 10 (BPP34C10), 11 and 12, and some edge-on views are shown in Fig. 3.5 (analogous diagrams for BPP34C10 have been shown above). It is clear that considerable reorganization of the ring conformations is needed before complexes can be formed, except for the two conformations of BPP34C10 (especially conformation II). An important interaction in these complexes is the edge-to-face stacking between aromatic rings shown for BPP25C7 (Fig. 3.5(a)) and discussed theoretically by Burley and Pesko (1986). It is inferred that the edge-to-face disposition is a result of an electrostatic attraction between a partial positive charge on the intruding hydrogen and the negatively charged


(a)

(b)

(c)

(d)

35

Fig. 3.5. Edge-on views of the molecules BPP25C7 (FIKTIC), BPP28C8 (FIKTOI), BPP37C11 (FIKTUO) and BPP40C12 (FIKVAW) in their neat crystals. (Reproduced from Slawin, Spencer, Stoddart and Williams, 1987.)

-cloud of the aromatic ring; energies of 4–8 kJ/mol are involved. The molecule of bis(1,5-dihydroxynaphtho)-35-crown-9 in its neat crystals provides another example (Ashton, Chrystal et al., 1987; FUTZUP); the central cavity is filled by a T-disposition of the two naphtho moieties, with one in an edge-to-face relation to the other. Similar dispositions have also been observed in intermolecular arrangements. Another new feature is introduced when the host molecule is 1,5-dinaphtho-44-crown-12; the neat molecule crystallizes in space group P21/a, Z ¼ 2 (KEBLEI). A complex of 1 : 2 composition is formed with paraquat which can be represented as {1,5-DN44C12 [PQT2þ]}(PQT2þ)(PF6) 4 (CH3)2CO (KEBLAE; P1, Z ¼ 1). One of the paraquat cations is included within the macrocycle and the second is stacked between the macrocycles, the whole assembly forming an alternating arrangement of donor and acceptor groups (Fig. 3.6; Ortholand, Slawin, Spencer, Stoddart and Williams, 1989). This is very similar in overall form to the well-known . . . DADADA . . . stacking of alternating -donors and *-acceptors found in molecular compounds of polycyclic aromatic hydrocarbons and a variety of electron acceptor molecules (see Chapter 15). In the double or [3]catenane (Ashton, Brown, Chrystal et al., 1991) described below the donor and acceptor units have a . . . DADDAD . . . sequence. Many related examples are given by Amabilino and Stoddart, (1995). It was noted (Ashton, Reddington et al., 1988; Ashton, Goodnow et al., 1989; Bu¨hner, Geuder, Gries, Hu¨nig, Kock and Poll, 1988) that formation of intramolecular inclusion complexes stabilized by charge transfer interactions from a donor host to an acceptor guest, as in most of the examples described above, implied that the converse situation – acceptor host and donor guest – should also lead to formation of complexes. This is shown schematically in Fig. 3.7 and was realized in practice by using the bis(paraquat) derivative (formula 3.7, also shown on the right hand side of Fig. 3.7), abbreviated as

36


O B Paraquat within crown ether

N

Paraquat between crown ethers

Naphtho donor

Polyoxomethylene chain

O y z

A

C

x

y

z

Intramolecular paraquat acceptor

Naphtho donor

x

Fig. 3.6. The stacking arrangement of the paraquat moieties within and between the naphtho-crown ether rings in {DN44C12[PQT2þ]}(PQT2þ) (PF6)4(CH3)2CO is shown on the left of the diagram; an edge-on view of the arrangement within the {DN44C12[PQT2þ]} portion is shown on the right. (Data from Ortholand, Slawin, Spencer, Stoddart and Williams, 1989.)

[BBIPYBIXYCY][PF6]4. The systematic name of the cyclobis(paraquat-p-phenylene) tetracation is 5,12,19,26-tetraazoniaheptacyclo-[24.2.2.22,527,10212,15216,19221,24] tetraconta-1(28),2,4,7,9,12,14,16,18,21,23,26,29,31,33,35,-37,39-octadecaene (Odell, Reddington, Slawin, Spencer, Stoddart and Williams, 1988), which illustrates why abbreviations are preferred. 4+ N

N

N

N

[BBIPYBIXYCY] 3.7

Scheme 3.2

The host system is quite rigid and the conformation taken up in the solvated salt of composition (BBIPYBIXYCY)4þ(PF 6 )43CH3CN (VAFRID10) is rather strictly preserved in the 1,2- (Odell, Reddington et al., 1988) and 1,4-dimethoxybenzene inclusion complexes, the compositions of which were given as {BBIPYBIXYCY [C8H10O2]}4þ(PF 6 )43CH3CN (the solvent content can also be 2CH3NO2.H2O) (Ashton, Reddington et al., 1988); the 1,4-dimethoxybenzene inclusion complex is shown schematically on the right hand side of Fig. 3.7 and in space-filling mode on the left side of Fig. 3.8. The 1,2-dimethoxybenzene guest is disordered over two orientations and approximates in appearance in the crystal to 1,5-dimethoxynaphthalene, which also can be


O O

O

O

O CH3 N+

N+ CH3 O

37

O N+

O

OMe

O

N+

MeO N+

N+

O

Fig. 3.7. Schematic representation of the donor host–acceptor guest intramolecular inclusion complex dication {BPP34C10[PQT]2þ} on the left and the converse situation on the right with donor and acceptor roles interchanged; here the tetracation is {(BBIPYBIXYCY)4þ [C8H10O2]}, with the host designated 3.7. The donor benzene rings are hatched. Space-filling views of the tetracation are shown in Fig. 3.8. (Adapted from Fig. 1 of Ashton, Goodnow et al., 1989.) ˚ , deg, A ˚ 3) for some intramolecular inclusion complexes of formula Table 3.1. Cell dimensions (A 4þ {BBIPYBIXYCY* [guest]} (PF6 )43(solvent). The su’sy of the cell edges are 1 in 5000, and of 0.03 . The crystals all have space group P21/n, with Z ¼ 2; the host:guest ratios are all 1:1 except for the 1,3-bis(5-hydroxy-1-naphthyloxy)propane guest, where it is 1 : 0.5 Guest and refcode

a

b

c

Cell volume

No guest; MeCN solvate VAFRID10 No guest; monohydrate VAFROJ 1,2-dimethoxybenzene VAFSAW 1,4-dimethoxybenzene VAFRUP 1,5-dimethoxynaphthalene KIRTEK 1,3-bis(5-hydroxy-1-naphthyloxy)propane (1:0.5) KIRTIO

10.805 10.478 11.076 10.948 11.218 10.881

19.819 20.152 19.805 19.869 19.756 20.043

14.027 13.757 13.962 13.886 13.980 14.066

109.36 106.45 111.57 110.55 111.23 110.09

2834 2786 2848 2828 2888 2881

* CSD name cyclobis(paraquat-p-phenylene) y

su is standard uncertainty

included. Furthermore, the longer but analogous guest molecule 1,3-bis(5-hydroxy-1naphthyloxy)propane can also be included, with a host : guest ratio of 1 : 0.5 (Reddington, Slawin, Spencer, Stoddart, Vicent and Williams, 1991). All these crystals are essentially isomorphous, as is shown by their similar cell dimensions and identical space groups (Table 3.1). The conformation of [BBIPYBIXYCY]4þ is also maintained in its other complexes. An inclusion complex is also produced with tetrathiafulvalene, an easily oxidized electron donor (see Chapter 13; oxidation potential 0.4 V); the crystals have composition {BBIPYBIXYCY[C6H4S2]}4þ(PF 6 )44CH3CN but are triclinic (P1, Z ¼ 1; VOLMEO) and hence different from the group listed in Table 3.1. The guest is indeed inserted into the cavity of the host but anions and solvent molecules are differently disposed (Philp, Slawin, Spencer, Stoddart and Williams, 1991). All these crystals have an arrangement of rectangular doughnut (or ‘‘bagel’’) shaped host cations alternating with

38


OH O O

N O

O

N

N

O

O

O

O

O

N

N

N

OH

O

O

Fig. 3.8. Plan views of space-filling representations of the host–guest arrangement in the 1,4dimethoxybenzene intramolecular inclusion complex {BBIPYBIXYCY[C8H10O2]}(PF6)43CH3CN (on the left), and in the rotaxane {BBIPYBIXYCY[HOC6H4(CH2CH2O)4C6H4(OCH2CH2)4 C6H4OH]}(PF6)42CH3CN (on the right). (Reproduced from Anelli et al., 1992.)

layers of anions plus solvent molecules. The host cavity is empty in the parent structure and filled by the included guests in the supermolecules, with ‘‘threaded’’ as perhaps a more appropriate description for the last, and longest, of the guests. Hu¨bner et al. (1989) have shown that analogous hosts can include a variety of guests, among which are anthracene, phenanthrene, pyrene and a number of substituted naphthalenes; no crystal structures have been reported. The ‘‘threading’’ motif is also maintained in a variety of self-assembling rotaxane and pseudorotaxane complexes of [BBIPYBIXYCY]4þ, such as that illustrated on the right hand side of Fig. 3.8, and found as solvated salts. We give a number of guests with refcodes of the crystal structures (Anelli, Ashton, Spencer, Slawin, Stoddart and Williams, 1991; Anelli, Ashton, Ballardini et al., 1992): R

O

O

O

O

O

O

O

O

R

VOTPAV, as shown, with R ¼ Si(i-Pr)3 SOVLEU, as shown with R ¼ p-phenol SOVKOD, as shown, with R ¼ 2,5-dimethoxyphenyl VOTNUN, with three oxygens per chain, and R ¼ Si(i-Pr)3 SOVKOJ, with five oxygens per chain and R ¼ p-phenol VOTNEX, with three oxygens per chain and R ¼ H VOTNIB, as shown, and R ¼ H VOTNOH with five oxygens per chain and R ¼ H Scheme 3.3

We reproduce here two other structures, which lie on the borders of the subject matter of this book, but perhaps could be described as the ultimate in ‘‘molecule-within-molecule’’ inclusion complexes – the [2] catenane composed of the two host molecules BPP34C10

39


N

N

O

O

O

O O

O

O

N

N

O

O

O

O

O

O O

O

N N

N O

N

O

O

O

Fig. 3.9. The intertwined macrocyclic rings found in the catenane {[2]-[BPP34C10][BBIPYBIXYCY]}[PF6]45CH3CN (KEBKUX10). The ball-and-stick and space-filling views are from somewhat different orientations. The C atoms and rings in the cation have been shaded for clarity. (Reproduced from Anelli et al., 1992.)

Macrocyclic tetracation

Crown ether

Crown ether

y z x

Macrocyclic tetracation

Fig. 3.10. Double interlacing of two crown ether rings with the macrocyclic tetracation as found in the [3]catenane SOVLIY C28H40O10C48H40N44þC28H40O104(PF6). (Data from Ashton, Brown, Chrystal et al., 1991.)

and [BBIPYBIXYCY]4þ (Fig. 3.9) and the double or [3]catenane composed of BPP34C10 and the cation analogous to [BBIPYBIXYCY]4þ but made up of four bipyridyl units (Fig. 3.10). As noted above, in the double or [3]catenane (Ashton, Brown, Chrystal et al., 1991) the donor and acceptor units have a . . . DADDAD . . . sequence. One important crystallographic feature of the above group of complexes, perhaps considered trivial by chemists, is that they are all ternary complexes. In addition to the intramolecular inclusion of guests within the cavities of the host molecules, there is also intermolecular inclusion of solvent molecules in the interstices between the host molecules. These solvent molecules make an essential contribution to the cohesion of the crystals, as do the interactions between the ions. If there is disorder of the

40


anions, then this will have a stabilizing effect due to the additional configurational entropy. Crystal structures of pseudorotaxanes, rotaxanes and catenanes illustrating and extending the principles described above continue to be reported (e.g. Amabilino and Stoddart 1995; Ashton, Ballardini, Balzani, Be´loradsk’y, Gandolfi, Philp, Prodi, Raymo, Reddington, Spencer, Stoddart, Venturi and Williams, 1996; Asakawa, Brown, Menzer, Raymo, Stoddart and Williams, 1997 (RISSER, RISSIV, RISSOB, RISSUH, RISTAO, RISTES, RISTIW, TEYMAL10); Asakawa, Ashton, Hayes, Janssen, Meijer, Menzer, Pasini, Stoddart, White and Williams, 1998 (PUKNOY, PUKNUE, PUKPAM)). The crystals described above are all of salts but [2]catenanes with both partners neutral are also known (Hamilton, Feeder, Prodi, Teat, Clegg and Sanders, 1998; NIFLAP as the perdeuterodimethylsulfoxide solvate C34H32O7C32H12N4O6C2D6OS). N,N 0 ,N00 -tritosyl-5,8,14,17,23,26-hexamethyl-2,11,20-triaza[3.3.3]paracyclophane (formula not shown) forms a 1 : 1 intermolecular inclusion complex with dichloromethane (Bottino, Finocchiaro, Lipkowski, Mamo and Pappalardo, 1991; SOVKAP), whereas a number of different guests form a series of intramolecular inclusion complexes with 2,11,20,29–tetramethyl–2,11,20,29–tetraaza-[3333]-para-cyclophane (3.8) (Tabushi, Yamamura, Nonoguchi, Horotsu and Higuchi, 1984a, b). 3.8 is enantiomeric with exact or approximate C2-2 symmetry. The enantiomers are rapidly interconverted in solution but appear separately in the crystals, among which both racemates and enantiomorphs are found (Table 3.2). H3C

CH3

N

N

N

N

H3C

CH3 3.8

H2N+

(CH2)4

NH2+

H2N+

(CH2)4

NH2+

3.9

Scheme 3.4

Crystal structures have been reported for the complexes with dioxane (Abbott, Barrett, Godfrey, Kalindjian, Simpson and Williams, 1982), CHCl3, CO2, CH3CN and CH2Cl2, (Tabushi, Yamamura, Nonoguchi, Hirotsu and Higuchi, 1984a, b; Hirotsu, Kamitori, Higuchi, Tabushi, Yamamura and Nonoguchi, 1984) and CH2BrCl (Nonoguchi, Yamamura, Tabushi, Higuchi and Hirotsu, 1992). This information, together with that summarised in Table 3.2, shows some interesting features. The host molecule is found to have C2-2 symmetry in all the complexes (crystallographically imposed for the dioxane, CHCl3, CH3CN and CO2 complexes), and is therefore chiral in the solid state (also, of course, in solution, where racemization will be very rapid). A feature common to the unit cells of the complexes is the similarity of the values of [010]. The structure analyses show


41

Table 3.2. Crystal data for 2,11,20,29–tetramethyl–2,11,20,29–tetraaza[3.3.3.3]paracyclophane (3.8) and some of its 1 : 1 intramolecular inclusion complexes with various guests (the compositions were determined directly by NMR and/or GLPC methods for the chloroform, methylene chloride, bromochloromethane and acetonitrile complexes). This table was adapted from a summary of results given by Nonoguchi et al. (1992), and in earlier papers Guest

˚ a/A

˚ b/A

˚ c/A

None DERFIP

21.491

15.123

10.020

12.694 89.86 5.772

13.496 80.30 13.514

5.668

Vol./ asymmetric unit (A˚3)

Vol./ guest (A˚3)*

Space group

Z

807.6

–

P21/a

4

–

893.2

85.6

P1

1

112.41

898.6

91.0

P21

4

13.438

111.31

892.9

85.3

C2

2

5.691

13.440

112.06

889.3

81.7

C2

2

25.544

5.406

26.948

109.90

874.8

67.2

C2/c

4

25.590

5.420

26.900

109.86

877.3

69.7

C2/c

4

25.314

5.536

53.508

111.07

874.6

67.0

C2/c

8

25.373

5.486

53.759

111.00

873.3

65.7

C2/c

8

Enantiomorphs pyridinex 5.712 JOPVOZ 68.16 C6H6 49.846 JOPVIT 25.166 CHCl3 CEYHIX Dioxane 25.091 BIJJOT Racemates CH3CN DERFUB CO2 DERFOV CH2BrCl JOPVEP CH2Cl2 CEYHOD

ß/

97.26

x Reduced cell. ˚ 3 from the volume of the asymmetric unit; these volumes will * Nominal values obtained by subtracting 807.6 A be overestimated because of the difference in the shape of the host in its neat crystals and in those of the ˚ 3 in its crystals so the composition is probably nearer complexes. Note that benzene has a molar volume of 122 A 1:0.5, and this is also likely for the pyridine and dioxane complexes.

that all have stacks of homochiral host molecules along [010]. The structures in Table 3.2 can be divided into two groups (a) those with chiral space groups where spontaneous resolution has occurred on crystallization (the pyridine, CHCl3 and dioxane complexes, and, possibly the benzene complex), and (b) those with racemic space groups. The principal difference between the two groups lies in the mutual arrangements of the stacks, and how adjacent homochiral stacks interact, or stacks of different senses of chirality interact. The interactions within the stacks have been calculated by molecular mechanics (MM2) for the CH2Cl2 complex (Hirotsu et al., 1984); the total interaction energy of a column is 138 kJ/mol, with host–guest interaction amounting to 56 kJ/mol and guest–guest interaction to 4 kJ/mol. A further subdivision can be made in terms of unit cells and space groups: thus the CHCl3 and dioxane complexes belong together, as do the CH3CN and CO2 complexes, and the CH2Cl2 and CH2BrCl complexes. Comparison of the crystal structure of the neat host with those of the complexes shows that the host takes up a

42


Fig. 3.11. Comparison of the conformations of the macrocycle 3.8 in its complexes (on the left) and in its neat crystals (on the right). (Reproduced from Hirotsu, Kamitori, Higuchi, Tabushi, Yamamura and Nonoguchi, 1984.)

more compact form in its neat crystals than in the complexes, where its shape has to allow for inclusion of guests (Fig. 3.11). Guest exchange experiments show that about 40% of the guest in the CHCl3 complex can be replaced by CH2BrCl before there is a change in crystal structure. Weakness of host-guest binding is shown by the fact that guest exchange occurs. The 1,6,20,25-tetraammonium[6.1.6.1]paracyclophane cation 3.9 forms intra-molecular complexes with 1,3-dihydroxynaphthalene, 2,7-dihydroxynaphthalene, p-xylene and durene when crystallized from aqueous solutions at pH < 2 (Odashima, Itai, Iitaka and Koga, 1980). The crystal structure of the salt {3.9[durene]}4Cl4H2O has been reported ˚ , ¼ 97.23(4) , V ¼ 2360 A ˚ 3, space group (a ¼ 14.552(7), b ¼ 22.58(1), c ¼ 7.238(3) A P21/n, Z ¼ 2; ACPHDR); the durene is included within the torus of the cation while the anions and water molecules are situated between the cations. The cation is centrosymmetric, in contrast to 3.8. An analog to 3.9, 1 0 ,100 -dimethyldispiro-[1,6,20,25tetraoxa[6161]paracyclophane-13,4 0 : 32,400 -bispiperidine] (3.10, formula not shown), forms a complex of composition {3.10[C6H6]}C6H6H2O (DENFOR; P1, Z ¼ 1) in which one of the benzenes is enclosed within the host cavity while the other is located in tunnels between the host molecules. Other complexes of 3.10, with toluene and p-xylene (DENFUR), are intermolecular (Krieger and Diederich, 1985). The macrocyclic tetraimide shown in Fig. 3.12, which is composed of two 1,4,5,8naphthalenetetracarboxylic acid diimide subunits joined by two (CH2)8 chains, forms an ‘‘intercalative molecular cryptate’’ with nitrobenzene as guest (Jaswinski, Blacker, Lehn, Cesario, Guilhem and Pascard, 1987); here the polyether chains of crown ethers are replaced by aliphatic chains of methylene groups. The nitrobenzene takes up two orientations in the cavity 165 apart and the two aliphatic (CH2)8 chains have somewhat different conformations and types of disorder; hence the non-centrosymmetric space group P1 was assigned to the triclinic crystals. An example of the effects of the flexibility of an host molecule is illustrated by a tantalizingly brief report (Itai, Tanaka and Iitaka, 1979) about [26]metacyclophane (3.11) and its intramolecular complexes; a full report appears not to have been published. The diagrams reproduced in Fig. 3.13 show that the host molecule is very flexible; in the neat P 1 polymorph (there is a second polymorph, space group Pbca, about which no


O N O A O N O

43

O N O + O N – O O N O

A

Fig. 3.12. Three diagrams showing the host–guest relationship for the macrocyclic tetraimide composed of two 1,4,5,8-naphthalenetetracarboxylic acid diimide subunits joined by two (CH2)8 chains; the chemical formula (with A ¼ (CH2)8) is shown in the lower part of the diagram, a line outline at top left and a space-filling model at top right. Only one of the two orientations found for nitrobenzene is shown. (Reproduced from Jaswinski, Blacker, Lehn, Cesario, Guilhem and Pascard, 1987.)

information is available) the conformer has a squashed shape without an interior cavity (c); the p- and o-xylene inclusion complexes (space group Pnmn) have the guests parallel to the mean plane of the host (a), whereas the guests penetrate the host in the benzene, m-xylene, n-heptane, cyclohexane and geraniol complexes (b; space group I2/a). It is not known whether the crystals of the complexes in these groups are isomorphous.

(CH2)2

(CH2)2

(CH2)2

(CH2)2

(CH2)2

(CH2)2

[26] metacyclophane (3.11)

Scheme 3.5

44


(a)

(b)

(c)

Fig. 3.13. Mode of inclusion of the guests in (a) the p- and o-xylene inclusion complexes of [26]metacyclophane, and (b) of the benzene, m-xylene, n-heptane, cyclohexane and geraniol complexes; (c) the conformation of the neat host in its triclinic polymorph. (Reproduced from Itai, Tanaka and Iitaka, 1979.)

3.3

Cleft molecules as hosts

3.3.1

Single-cleft hosts

Most of the structures described above have doughnut (or ‘‘bagel’’) molecules as hosts, with the guest filling the central hole. The hosts in the structures to be described in this section have a tweezer (or U, cake-server, hairpin or horseshoe) shape with the guests enclosed between the two arms of the tweezer. Dibenzocrown ethers of the general type DB3nCn (n ¼ 6–12, with the phenyl groups linked 1,2 (ortho) to O) form stable adducts with [Pt(bpy)(NH3)2]2þ (bpy ¼ bipyridyl), [Rh(cod)(NH3)2]þ (cod ¼ 1,5-cyclooctadiene) and [Rh(nbd)(NH3)2]þ (nbd ¼ norbornadiene), the counterion being (PF6). One of the first hosts studied was dibenzo-24-crown-8, which forms adducts with these three metal complexes. The unsymmetrical location of guest in the Pt adduct (Colquhoun, Doughty, Maud, Stoddart, Williams and Wolstenholme, 1985; DATNOB, P1, Z ¼ 24) compared to its symmetrical location in the Rh adducts (Colquhoun, Doughty, Stoddart and Williams, 1984; cod complex COCPEP, P21/c, Z ¼ 4; nbd complex, COCPIT, Pbca, Z ¼ 4) is evidence for different types of host–guest interaction – charge transfer interactions between the benzene ring of the crown ether and the bipyridyl ligand stabilize the adducts with the Pt complex while the Rh complexes interact with the crown ether through a substantial number of weak hydrogen bonds and electrostatic interactions and their additive effect leads to stabilization (cf. caption to Fig. 3.14). The 1 : 1 complexes of the smaller host dibenzo-18-crown (DB18C6) with pyridinium NEXLOR) and 1-aminopyridinium BF4 (NEXLUX) are isostructural (both Cc, Z ¼ 4, with similar cell dimemsions), and have the cations located within the cleft (La¨msa¨, Huuskonen, Rissanen and Pursiainen, 1998). Crystal structures have been reported (Bush and Truter, 1972) for DB30C10 itself (P21/c, Z ¼ 2, hence centrosymmetric; DBTCAD) and DB30C10KI (Pnna, Z ¼ 4, the [host. Kþ] moiety having a twofold axis; KIBDOT10) and for DB30C10 intramolecular inclusion complexes. DB30C10 has an elongated shape in its neat crystals but the host molecules have U-shapes in all the complexes, including that with KI (Fig. 3.15). Thus there is a remarkable conformational change on complexation, indicating considerable 4

There is also (loc. cit.) a [Pt(bpy)(NH3)2]2þ monohydrate complex of DB30C10 (P21/n, Z ¼ 4; BEFHUP10).

45

CLEFT MOLECUL ES AS HOST S

norbornadiene

2,2'bipypridyl

Pt

Rh O

NH3 x

z

y

N

z x

y dibenzo-crown ether

dibenzo-crown ether

Fig. 3.14. The intramolecular complexes of DB24C8 with two metal coordination-complex cations. The unsymmetrical disposition on the left (DATNOB) is indicative of charge transfer interactions between the components while the symmetrical disposition on the right (COCPIT) suggests the combined effect of a large number of weak hydrogen-bonding and electrostatic interactions. (Data from Colquhoun, Doughty, Maud, Stoddart, Williams and Wolstenholme, (1985) and Colquhoun, Doughty, Stoddart and Williams (1984).)

K

Fig. 3.15. Comparison of the centrosymmetric shape of DB30C10 in its neat crystals (on the left) with the shape found in its complex with potassium iodide (on the right); the molecule has a twofold axis, shown by the vertical line. Oxygens are darkened. The U-shaped conformation of DB30C10 in the KI complex is rather similar to those conformations found in the intramolecular inclusion complexes shown in Figs. 3.14 and 3.16. (Reproduced from Bush and Truter, 1972.)

flexibility of DB30C10. The analogous di-(1,2)-benzo-30-crown-10 (Colquhoun, Goodings, Maud, Stoddart, Wolstenholme and Williams, 1985; CAKMEG10) and di-(2,3)-naphtho-30-crown-10 (3.13) hosts both take up a U-shape in their crystals. For example, in {dinaphtho-30-crown-10[diquat]2þ}(PF6)21/2H2O (Fig. 3.16) (Allwood, Colquhoun, Doughty, Kohnke, Slawin, Stoddart, Williams and Zarzycki, 1987; FIKVUG), both host and guest are located on a crystallographic twofold axis. The host-guest complex has much the same shape when diquat2þ is replaced by (Pt(bipyridyl)(NH3)2)2þ (FIKVOK; here the crystallographic results were of poorer quality because of disorder). The differences between the mutual arrangements of host and guest in the two complexes

46


Fig. 3.16. Edge-on view of space filling diagram of {DN30C10[Diquat]2þ} cation in its hexafluorophosphate hemihydrate salt. The diquat moiety is enclosed by the jaws of the tweezer host 3.12. The arrangement in {dibenzo-30-crown-10[diquat]2þ}[PF6] 21/2H2O is very similar. (Reproduced from Allwood, Colquhoun et al., 1987.)

were ascribed, as above, to maximization of Coulomb interactions in the diquat complex and of overlap of -arene systems in the Pt complex. Structures have also been reported (Kohnke, Stoddart, Allwood and Williams, 1985) for {2,20-bisformyldibenzo-30-crown-10[diquat]2þ}(PF6)2 0.75CH3CN (CULCAN) and {2,20-bismethyldibenzo-30-crown-10[diquat.]2þ}(PF6) 2H2O (CULCIV), with results rather similar to those already described. {2,20-Bishydroxymethyldibenzo-30-crown-10 [diquat]2þ}(PF6) 2 (CULCER) was also prepared but the structure was not determined because of its complexity; the crystals are of interest because it is possible that spontaneous resolution of the crown ether had occurred on crystallization (the space group is P2221, Z ¼ 8). O O

O

O 3.13

Scheme 3.6

Another type of molecular tweezer 3.13 has been synthesized by Harmata and Barnes (1990); this molecule has a U-shape despite its planar representation. Although crystals of the neat compound were not obtained, it could be crystallized together with a molecule of 1,3,5-trinitrobenzene (JESCAL; P21/c, Z ¼ 4) to give a cleft structure analogous to that shown in Fig. 3.16, the TNB molecule being held within the jaws of the host. It was suggested that the face-to-face donor–acceptor interactions (see Chapter 15) were supplemented by a face-to-edge interaction between a positively charged hydrogen of TNB pointing towards the centre of the -system of the central benzene ring; such interactions have been proposed for analogous systems (Burley and Pesko, 1986). 3.13a has a similar behaviour; it could only be crystallized as a 1 : 1 complex with nitrobenzene, ‘‘probably because different conformations precipitated together so that a single crystalline product could not form.’’ The guest molecule was located within the cleft of

47

CLEFT MOLECUL ES AS HOST S

the ‘‘naphthalene-walled clip’’ (Reek, Engelkamp, Rowan, Elemans and Nolte, 1998; BIZRAD, P1, Z ¼ 2; no 3D coordinates), as shown on the right. nitrobenzene guest OMe O MeO N Ph N OMe

N Ph N

O

O

O MeO 3.13a

N1

C1

C21

C1⬘ N1⬘ C21⬘

Scheme 3.7

Related molecules have been used for the synthesis of belt and collar shaped potential hosts for intermolecular complex formation. Crystal structures have been reported for two cylindrical hosts but not, as yet, for any complexes (Ashton, Brown, Isaacs et al., 1992; FUJGOG10; GIJWIF10). 3.3.2 Double-cleft hosts Saddle-shaped hosts, with two approximately mutually perpendicular clefts, have been reported (Schwartz, Knobler and Cram, 1992). As an example we reproduce a stereodiagram (Fig. 3.17) of one of these hosts with two enclosed molecules of benzene (JUSXIE). The

Fig. 3.17. Stereodiagram of a double-cleft host with a benzene guest in each cleft. (Reproduced from Schwartz, Knobler and Cram, 1992.)

48


crystals are tetragonal, space group I41/a and each molecule has S4 symmetry, the two clefts thus being equivalent. 3.4 3.4.1

Container molecules as hosts Introduction

An appreciable portion of the rest of this chapter is concerned with the work of D. J. Cram and colleagues at the University of California at Los Angeles. This has been described recently (Cram and Cram, 1994) in more detail than we can accommodate here; the series ‘‘Host-Guest Complexation’’ had reached No. 68 by 1997. Cram and coworkers use the following abbreviation (X)G where (X) is the container molecule and G the encapsulated guest, and (X)[G where the guest is perching or accommodated between (rather than within) the host molecules; we shall use our own notation described earlier. We (again) briefly summarise the nomenclature. The container molecule is given the suffix ‘‘-and’’ and the complex the suffix ‘‘-ate’’ or, more usually, ‘‘-plex.’’ Thus, at one extreme of the classification, cavitands are bowl-shaped molecular vessels, essentially closed at one end but open at the other; they contain enforced concave interior surfaces of molecular dimensions (Cram, Karbach et al., 1988). They form caviplexes without steric inhibition to either the complexation or decomplexation processes. At the other extreme one has the carcerands, which entrap guest molecules permanently within their shells by constrictive binding; this implies that the guests are incorporated during synthesis and presumably also have some template role. The hemicarcerands are intermediate, and have portals through which the guest molecules can enter or escape the host cavity, with energy barriers large enough to allow isolation and characterization of the complexes at normal temperatures. Cram has described these quasi-spherical molecules as having polar caps, generally rigid, joined by chains of atoms which can be rigid or flexible, and cover the temperate and equatorial zones. A generalized diagram is given in Fig. 3.20 below, where the polar caps are called ‘‘spacers’’ and the linking chains of atoms ‘‘connectors.’’ 3.4.2

Cavitands and caviplexes

A general formula for cavitands is shown in the insert (3.14, n ¼ 1, 2, 3; R ¼ CH3, Br, I); the overall dimensions can be changed by introducing methylene bridges of varying length while the solubility can be manipulated by replacement of the methyls by suitable groups. The methylene bridges have been replaced by quinoxaline (or analogous) flaps; these can be all axial with respect to the rest of the molecule, giving a vase conformation, or all equatorial, giving a kite conformation (Moran, Ericson, Dalcanale, Bryant, Knobler and Cram, 1991). A remarkable result was obtained for a derivative of host 3.14 in which the methylene bridges were replaced by quinoxalines and the pendant methyl groups by menthoxy (OOC(–)menthyl). Complexes of this host with aromatic guests such as benzene and toluene were shown, by desorption chemical ionization (DCI) mass spectrometry, to exist in the gas phase provided the temperature of the ion source was below 100 C. It was suggested that these complexes were formed by complexation of neutral benzene and neutral cavitand molecules in the gas phase, followed by ionization (Vincenti, Dalcanale, Soncini and Guglielmetti, 1990).

49

C ONT AINE R M OL E C UL E S AS HOSTS

Crystal structures were determined for nine caviplexes with methylene bridges, and seven of these contain uncomplexed, intermolecular, guests in addition to complexed, intramolecular, guests (Cram, Karbach et al., 1988). The space groups of two of the caviplexes show that spontaneous resolution of the enantiomeric host has occurred on crystallization. The conical cavitands are supported on the four methyl groups. The nearly closed bases of the cavities are defined by a 16-membered [1.1.1.1] metacyclophane ˚ , too small to allow passage of guest macroring, with an internal diameter of about 3 A molecules. The open tops are defined by [mmmm] meta-cyclophane macrorings con˚. taining eight oxygens and from 24 to 35 ring atoms, with internal diameters of 9–10 A We note some special features: R ¼ H, n ¼ 1, guest CH2Cl2, C2/c, Z ¼ 8. The guest perches above the open end of the bowl. When the guest is CH3CN, the space group is the noncentrosymmetric P4. An additional CH3CN molecule is located interstitially between the host molecules. R ¼ CH3, n ¼ 1, intramolecular guest cyclohexane, intermolecular guest benzene, P212121, Z ¼ 4; spontaneous resolution (of the enantiomeric host) has occurred on crystallization. The cyclohexane, perched above the open end of the bowl, is said to be in the boat conformation but this requires confirmation, especially as the R factor was 15%. R ¼ Br, n ¼ 1, inner methyls replaced by CH2CH2C6H5 (for solubilization), H2O present both as intramolecular and intermolecular guest, space group Pnma, Z ¼ 4 (the host molecule has a mirror plane) (Sherman, Knobler and Cram, 1991; JILZEJ). This caviplex is important in the synthesis of carceplexes (see Section 3.4.6).

R (CH2)n O

O (CH2)n

O

O CH3

H

H

R O

CH3

CH3 H

H

R

CH3

O

O

N

R'

O

N

R'

O (CH2)n

(CH2)n O R

Substituted quinoxaline

3.14

Scheme 3.8

A particularly rigid host is obtained by replacing the CH2 groups bridging the oxygens by Si(CH3)2 to give C40H48Si4O8 (5,10;12,17;19,24;26,3-tetrakis-(dimethylsiladioxa)1,8,15,22-tetramethyl[14]metacyclophane, the cavitand shown in Fig. 3.1). In the crystal structure of {C40H48Si4O8[CS2]}CS2 (Goldberg, 1986; CUYXEZ10) one guest molecule lies along the central, (approximately) fourfold axis of the host, while the second lies between adjacent host molecules in the [100] direction (Fig. 3.18), illustrating the delicacy of the balance between intra- and intermolecular enclosure. The enclosed guest is in a

50


x z y

x z

y

Fig. 3.18. Stereoviews, parallel (a) and perpendicular (b) to the cavity surface, of the 1 : 2 molecular inclusion complex of the cavitand C40H48Si4O8 with CS2. There is another CS2 molecule between adjacent hosts. (Data from Goldberg, 1986.)

lipophilic environment composed of phenyl groups within the cavity and methyl groups on and above its surface. The host has a dish-like shape, and is thus representative also of the cavitands described above. 3.4.3

Hemispherands and hemispheraplexes

Hemispherands are receptor molecules in which at least half of the binding sites for intramolecular complexation are pre-organized, and complexes are known with alkali metal and alkylammonium cations and neutral organic molecules. The factors governing the binding of malononitrile (CH2(CN)2) to four hemispherands have been studied by combining the results of crystal structure analyses, molecular mechanics calculations and solution thermodynamic measurements (Grootenhuis, van Eerden, Dijkstra, Harkema and Reinhoudt, 1987). We take one example (3.15, formula with n ¼ 1; C30H36O6). The crystal structures of 3.15 (at 113 K, P21/c, Z ¼ 4; MOMODT, Goldberg, 1980) and {3.15[CH2(CN)2]}0.60 (C2H5)2O (C2/m, Z ¼ 4; Grootenhuis, van Eerden, Dijkstra, Harkema and Reinhoudt, 1987; DUTDOL10) have been reported; the conformations of the host are compared in Fig. 3.19. Molecular mechanics calculations using these conformations give strain energies of 57 and 64 kJ/mol, indicating that the host is more strained when it has the conformation taken up in the complex. NMR measurements (240–320 K; in CDCl3 solution) give the association constant as 28 M1

51


CH3 CH3

H 3C X

X

O X = OCH3

012

08

X O

01

04

O n

Fig. 3.19. Comparison of the conformations taken up by 3.15 (n ¼ 1) in its neat crystals (centre) and in its complex {3.15[CH2(CN)2]}0.60 (C2H5)2O (right). (Reproduced from Grootenhuis, van Eerden, Dijkstra, Harkema and Reinhoudt, 1987.)

at 298 K, H ¼ 35.1, TS ¼ 26.8 and G ¼ 8.4 kJ/mol. The complex is enthalpy-stabilized in solution and, presumably, also in the solid state. The interaction energy between host and guest must be about 42 kJ/mol (we have ignored any influence of solvent molecules in the crystals of the complex). This will be due to four bifurcated C–H . . . O hydrogen bonds from guest to host (Fig. 3.19), and van der Waals interactions. t-Butylammonium perchlorate ((CH3)3CNH3þClO4) forms a complex with 3.15 (n ¼ 1) in which there are three Nþ–H . . . O hydrogen bonds to the oxygens of the three methoxy groups; there is also a conformational change on complexation but the strain energy of the macrocycle does not seem to have been calculated. The free energy of formation of the complex is 32.1 kJ/mol (in CDCl3 solution) (Koenig, Lein, Stuckler, Kaneda and Cram, 1979), four times greater than that of the malononitrile complex; values of H and TS have not yet been reported.

3.4.4 Triply bridged cyclophanes and analogous molecules as three-dimensional hosts for intramolecular guests The first enclosure (or container) molecule to be purposefully synthesized was that of the [2.2.2] cryptand (Dietrich, Lehn and Sauvage, 1968) shown in Fig. 3.20. For reasons of space, we shall discuss only some aspects of cryptand structural chemistry and refer the reader to the broader account given by Dietrich (1996). A general formulation of an enclosure molecule is shown in Fig. 3.20; the synthetic methods used for translating the schematic into real molecules have been summarised by Seel and Vo¨gtle (1992). A formally simple realization of the schematic of Fig. 3.20 (left) is shown in Fig. 3.20 (right) and has been achieved by O’Krongly, Denmeade, Chiang and Breslow (1985; R ¼ CH3) and Vo¨gtle, Berscheid and Schnick (1991; R ¼ H). We now consider crystallographic data on the modes of enclosure. The crystal structures of {C56H38O6.[CH3CN]} (P21/c, Z ¼ 4; JIHJEP, Vo¨gtle et al.) and {C57H40O6[C6H6]}2C6H6 (triclinic, P1, Z ¼ 2); DECWOX, O’Krongly et al.) have been reported in these two papers; the spacer is RC(p-C6H4O–)3 and the connector –C C–C C–. The shapes taken up by the host molecules are appreciably different in the two crystals; O’Krongly et al. (1985) noted that

52


O

O O N

O N O

O

[2.2.2]cryptand Spacer R

o

o o Connectors o

o

o

R

Spacer

Fig. 3.20. (above) Line diagram of [2.2.2]cryptand. (Below left) General formulation of an enclosure molecule. The spacers illustrated have three links, as in sym-C6H3(CH2NH2)3, but four and six link spacers have also been used successfully. The chemical nature of the connectors is matched to that of the spacers. The spacers can be different, as can the connectors. (Below right) Example of the translations of the general formulation into practice, as achieved by Breslow (R ¼ CH3) and Vo¨gtle (R ¼ H) and their coworkers. (Adapted from Seel and Vo¨gtle, 1992.)

their compound was ‘‘self-adjusting’’. One of the phenyl groups is twisted so as to interact with the included benzene molecule in a face-to-edge fashion, whereas the -acceptor acetonitrile molecule, disordered over two positions, has its nitrile -orbitals interacting with phenyl groups in both face-to-edge and face-to-face fashion. The spacer of an analogous host is 1,3,5-C6H3(O–)3; in this supramolecule, with acetonitrile as guest, it is the methyl group of the guest which interacts with a phenyl of the host (Berscheid, Nieger and Vo¨gtle, 1991; JODMOE, P21/c, Z ¼ 4). Two triply substituted aromatic rings, linked by poly(methylene ether) chains, form enclosures which can take up guests in intramolecular fashion. One example is 3.16 (shown schematically in the formula of Scheme 3.9 and in its correct shape in Fig. 3.21), which encloses diquat2þ in an hexafluorophosphate salt (Allwood, Kohnke, Stoddart and Williams, 1985); here the two spacers are benzenes linked between 1- and 2- positions by two -O-[(CH2)2-O]- chains and between 5-positions by a -(CH2)-O-(C6H4)-O-(CH2)chain. The crystal structures of the neat neutral host (orthorhombic, Pcab, Z ¼ 8; DAXMOE) and of the salt {3.16[Diquat2þ]}(PF6)2 (triclinic P1, Z ¼ 2; DAXMUK) have been reported. Comparison of the host conformation in the two crystals shows that some reorganization occurs on complexation (Fig. 3.21).


53

[Diquat]2+

y

x

z

y

z

x

Fig. 3.21. On the right is shown the host-guest arrangement in the salt {3.16[diquat]}(PF6)2 (DAXMUK) and on the left the somewhat different conformations of the host in the neat material. (Data from Allwood, Kohnke, Stoddart and Williams, 1985.)

R2 R1

R1 3.16 R1 = –O–(CH2)2–O-; R2 = –CH2–O–(C6H4)–O–CH2–

Scheme 3.9

The triply linked macrocyclic polyammonium receptor 3.17 binds dicarboxylate substrates (Lehn, Me´ric, Vigneron, Bkouche-Waksman and Pascard, 1991); linear substrates of formula O2C(CH2)nCO 2 (n ¼ 2–8) and fumarate and maleate have stability constants Ks ranging from 1400–4100 M1, while terephthalate has a Ks value more than an order of magnitude greater. Determination of the crystal structure of the terephthalate complex, which has space group Pnna, Z ¼ 4 (SIRZAU) shows that it indeed contains a ‘‘supramolecular species of cryptate nature in which two binding subunits of the ditopic coreceptor molecule cooperate in substrate binding,’’ but is in fact considerably more complicated, as is made clear by Lehn et al. (1991). The asymmetric unit consists of the intramolecular complex shown in Fig. 3.22, made up of the macrocycle and an enclosed terephthalate dianion (T(1)2), the complex having a two fold axis of symmetry normal to the N . . . N axis, a terephthalate dianion on a center of symmetry (T(2)2), another terephthalate dianion disordered over two orientations (T(3)2), five ordered water

54


Fig. 3.22. The intramolecular complex made up of the macrocycle 3.17 and enclosed terephthalate dianion, viewed from the side in outline (left) and space-filling (right) modes and, in the centre, along the N . . . N axis, which is normal to the complex’s crystallographic twofold axis of symmetry. O and N atoms are shaded. (Reproduced from Lehn, Me´ric, Vigneron, Bkouche-Waksman and Pascard, 1991.)

molecules and two more waters disordered over four sites. The asymmetric unit should thus be formulated as 1=2f3:176þ ½Tð1Þ2 g Tð2Þ2 Tð3Þ2 ð5 þ ð4X1=2ÞÞ H2 O, or 1=2fC57 H78 N8 6þ ½ O2 CðC6 H4 ÞCO2 g ð O2 CðC6 H4 ÞCO2 Þ ð O2 CðC6 H4 Þ CO2 Þ ð5 þ ð4X1=2ÞÞH2 O:

An intricate arrangement of hydrogen bonds links all these units. The four oxygens of the included terephthalate dianion are hydrogen bonded to the four ammonium groups, and also to neighboring water molecules; thus the intramolecular interaction is largely, but not entirely, one-dimensional. The view down the N . . . N axis reinforces the point made previously that it can be misleading to draw inferences about the shape of a complex from projections onto mean planes, as in the left- and right-hand parts of Fig. 3.22. Cavitand hosts capable of complexing small organic molecules have been synthesized using cyclotriveratrylene as spacer and –OCH2CH2CH2O– as connectors; these have been called cryptophanes (Collet, 1987). Cryptophane C (3.18; C54H51O9) has a stability constant Ks5 of 300 M1 for complexation with CH2Cl2 (Canceill, Cesario, Collet, Guilhem and Pascard, 1985) in CDCl3 solution, while that for Cryptophane D (3.19) is about twelve times smaller (both measured by 200 MHz NMR in solution) (Canceill, Cesario, Collet, Guilhem, Riche and Pascard, 1986); the values given in the earlier reports have been corrected by Collet (1987) following Canceill, Lacombe and Collet (1987). The energy barriers for inclusion and extrusion of the CH2Cl2 guest in both hosts were determined from the maximum broadening of the NMR signal (Table 3.3). CH2Cl2 seems 5 * The true stability constant is Ks ¼ Ks0(1 þ KsCDCl3[CDCl3]) where Ks0 is the apparent stability constant and KsCDCl3 is the stability constant of the CDCl3 cavitate ( 0.1 M1 for cryptophane C) and [CDCl3] ¼ 12.4 M. At high temperatures (310–330 K), Ks 2.1 Ks0 .

55


Table 3.3. Energy barriers (kJ/mol) for inclusion and extrusion of CH2Cl2 guest in Cryptophane C and D hosts in CDCl3 solution Host molecule

G# (inclusion)

G# (extrusion)

Cryptophane C Cryptophane D

45.2 49.5

47.0 46.1

Table 3.4. Thermodynamic parameters for intramolecular complexes of Cryptophane E with various guests. G refers to 300K; units for G and H are kJ/mol and for S J/mol K. The volumes were described as ‘van der Waals’ volumes without further definition Guest

˚ 3) V (A

G

H

CH3I CH2Cl2 CH2Br2 CHCl3 CHCl2Br CH(CH3)3 CHClBr2 CHBr3 CCl4 C(CH3)Cl3 C(CH3)2Cl2 C(CH3)3Cl CH3COCH3

54.5 57.6 65.5 72.2 76.1 79.4 80.1 84.0 86.8 89.2 91.6 93.9 70.0

10.0 11.7 12.6 15.5 14.2 11.7 12.1 9.6 5.0 0.8 0.4 3.3 5.4

S

4.2

25

25.1 21.8 15.9 6.3 5.9

29 25 13 17 17

Type of stabilization entropy enthalpy enthalpy enthalpy enthalpy and entropy enthalpy and entropy

c⬘ b⬘ O O Me

O

O Me

O Me

O

O

O

O

O Me

O O Me

O

O

O

O Me

b

Cl 13 c Cl 23

O

Cryptophane C (3.18) ((–) enantiomer)

O

a⬘ a

Cryptophane D (3.19) ((+) enantiomer)

Fig. 3.23. Left – schematic formulae of 3.18 and 3.19. Right – {Cryptophane D[CH2Cl2]} cavitate, showing the view down the pseudo-C3 axis, with 50% probability anisotropic thermal ellipsoids for the guest molecule (radii of other atoms are fixed arbitrarily). (Reproduced from Canceill, Cesario, Collet, Guilhem and Pascard, 1985.)

56


to enter the cavity of Cryptophane C more easily than that of Cryptophane D and is retained there more strongly. There is not much difference in the rates of leaving (extrusion). The crystal structures of these two isomeric intramolecular complexes have been reported. The crystals of {Cryptophane C[CH2Cl2]}2CH2Cl2 have two additional CH2Cl2 molecules located between the molecules of the complex and are monoclinic (P21/n, Z ¼ 4; CUSCEY) and racemic. The packing consists of alternate homochiral layers of ( þ ) and ( ) enantiomers of Cryptophane C. The cycloveratrylene spacers are mutually rotated by about 60 so the host has a pseudo-C3 axis (Fig. 3.23). The intermolecular dichloromethanes are accommodated in pairs in relatively large voids between groups of host molecules. The orthorhombic crystals of {( þ )-Cryptophane D[CH2Cl2]}1.25CH2Cl2 (space group P212121, Z ¼ 4; DIJJUB) have two additional dichloromethane sites located between the host molecules with occupation factors of 0.5 and 0.75; the space group shows that the macrocycle has resolved spontaneously on crystallization. The dichloromethane of the ‘‘supermolecule’’ shows no signs of disorder, but the interstitial dichloromethanes are loosely held in the crystal, which is efflorescent. Cryptophane E (three additional methoxy groups compared to C and D) forms a chloroform inclusion complex of composition {C57H60O12[CHCl3]}0.2CHCl30.2H2O (space group P21/n, Z ¼ 4; Canceill, Cesario, Collet, Guilhem, Lacombe, Lozach and Pascard, 1989; SEDPOG); this cryptophane also has a pseudo-C3 axis. Thermodynamic parameters have been measured for complexation of Cryptophane E and various guests (1H NMR measurements at 300 and 330 K, 1,1,2,2-tetrachloro-1,2-deuterioethane solvent); these parameters are listed against molecular volume in Table 3.4. A plot of G against volume for the various substituted methanes has a roughly parabolic shape, with a ˚ 3; the most stable complex is formed by chloroform, presumably minimum at 73 A because it fits most snugly into the cavity. The type of stabilization is also listed for those molecular complexes where separate H and S values are available. The methylene chloride complex is exceptional in that it is entropy stabilized. The acetone complex does not fit into the substituted methane series. Remarkably, Cryptophane E is easily oxidized þ to give a crystalline radical cation salt of composition {C57H60O12 [CHCl3]}PF6, the chloroform remaining included in the host cation (Renault, Talham, Canceill, Batail, Collet and Lajzerowicz, 1989; cf. Chapter 13). An isomeric pair of molecules analogous to the cryptophanes, with similar but not identical spacers and –OCH2–C C–C C–CH2O– connectors, has been synthesized, one of which is racemic (3.20; C66H54O62(CH2Cl2)) and the other (3.21; C66H54O6(CHCl3)) meso (Cram, Tanner, Keipert and Knobler, 1991; JOHGES, P1, Z ¼ 2). The crystal structure of 3.20 (space group Pbca, Z ¼ 8; JOHGAO) shows that the molecule has a compact form and near-spherical cavity capable of enclosing small molecules such as CHCl3, (CH3)3COH, CH2Cl2, cubane, propylene oxide and benzene. The value of Ks for benzene at 20 is 103 M1 and the activation free energy for decomplexation of benzene at 20 is 50 kJ/mol. Meso-3.21, which has an ellipsoidal cavity, does not complex the above guests in (CCl3)2CO solution, possibly because of competition from the solvent. There is also a more compact pair (3.22, 3.23) with shorter connectors (Tanner, Knobler and Cram, 1990; Fig. 3.24). Crystal structures have been determined (Tanner, Knobler and Cram, 1992). for {3.22[CH3CN]}CH3CN (triclinic, P1, Z ¼ 2; JORVER) and the isomorphous crystals of {3.23[CH3OH]}CH3OH (JORVIV) and {3.23[empty]}CH2Cl2 (JORVOB) (both rhombohedral, space group R 3, hexagonal cell dimensions for the

tosylate Ts

Ts

Ts

N

N

N

O

tosylate

S N

• CH3CN CH3CN CH3O

OCH3

methoxy

OCH3

tosylate

enclosed acetonitrile methoxy

acetonitrile of crystallization

methoxy

y

3.22. [CH3CN]·CH3CN (JORVER) z

x N

H

H

H

N

N

N

• CH3Cl2

N

methoxy methoxy

CH3O

OCH3

3.23. [empty]·CH2Cl2 (JORVIV)

methoxy

OCH3

disordered methylene chloride

x

y z

Fig. 3.24. Line formulae and molecular diagrams for the two hemicarceplexes 3.22 and 3.23. The CSD has erroneously interchanged the sets of coordinates given for isomorphous JORVIV and JORVOB. The former, labeled as containing both included methanol and methanol of crystallization, actually lists only methylene chloride of crystallization: the latter does not include coordinates for either type of methanol. (Data from Tanner, Knobler and Cram 1992.)

58


CH2Cl2 complex and (bracketed) the CH3OH complex: a ¼ 14.301(3) (14.262(2)), ˚ , V ¼ 7950 (7912) A ˚ 3, Z ¼ 6 (1/3 of a molecule in the asymc ¼ 44.65(1) (44.914(6)) A metric unit)). Stereodiagrams of the host – guest complexes are shown in Fig. 3.24. 3.23 in CDCl3 binds O2, N2 and H2O weakly and CH3OH with Ka 47 M1 (G295 ¼ 10 kJ/ mol, H 36 kJ/mol and S 88 J/mol K). Comparison of the two isomorphous rhombohedral structures shows that 3.23 is rigid enough for the presence or absence of a guest not to affect its shape. Very strong binding of p-nitrophenol in CD2Cl2 is shown by the triply linked hosts 3.24 and 3.25; the value of Kassoc is about 3.5 104 mol1 for 3.24 and about 10 104 mol1 for 3.25 (Cochran, Parrott, Whitlock and Whitlock, 1992). The crystal structures of the 1 : 1 intramolecular complexes of 3.24 and 3.25 with p-nitrophenol have been determined (as the bis(1,2-dichloroethane) (SUBZOE; C45H34N2O8C6H5NO32(C2H4Cl2); P21/a, Z ¼ 4) and 1,2-dichloroethane (SUBZIY; C45H34N2O8C6H5NO3(C2H4Cl2); P1, Z ¼ 2) solvates respectively), and show similar dispositions of guest to host (Fig. 3.25). NMR measurements confirm that these dispositions are maintained in solution. Hydrogen bonding ˚ between the hydroxyl of the guest and the pyridinium N of the host (d(O–H . . . N) ¼ 2.69 A ˚ (3.24) and 2.77 A (3.25)), noted in earlier work on analogous compounds (Sheridan and Whitlock, 1986, 1988; Whitlock and Whitlock, 1990), was inexplicably not mentioned in a later claim (Cochran, Parrott, Whitlock and Whitlock, 1992) that the detailed

O

O

O

O

O

O

O N O

O

N O

O

O

O

O

O 3.24

∗

3.25

N N

∗

N

N O

CH3 ∗ N CH3

∗

N N

CH3 N CH3

{3.25·[p-nitrophenol]}

Fig. 3.25. Upper – line formulae of 3.24 and 3.25. Lower – stereodiagram showing the disposition of guest and host in the supramolecular complex of 3.25 with p-nitrophenol; the triple bonds of the host molecule are marked with asterisks and oxygens are shown by hatched circles. In terms of the nomenclature used in Fig. 3.20(a), the substituted naphthalenes are the spacers of the host molecules, which have three different connectors. (Reproduced from Cochran, Parrott, Whitlock and Whitlock, 1992.)


59

geometry of these complexes is dominated by formation of a hydrogen bond involving the electron-rich p-xylene connector and the acidic o-nitro proton of the guest. As the OH . . . N hydrogen bonding has an energy of 30 kJ/mol, while the hydrogen bond contributes only about one-quarter of this to the intramolecular cohesion, it is clear that the former dominates.

3.4.5 Spherands and spheraplexes Spherands are macrocyclic or macropolycyclic systems where the ligands are organised, during synthesis and prior to complexation, in such a way that the unshared electron pairs of the binding sites line a roughly spherical cavity maintained by auxiliary covalent bonding (Cram, Kaneda et al., 1985; Cram and Trueblood, 1981; Cram, 1986). A comprehensive account, with some forty stereodiagrams of spherands, spheraplexes, hemispherands and hemispheraplexes, has been given by Maverick and Cram (1996a); we are perforce more modest. An essential requirement is that the spherand molecule should be rigid and not turned back on itself, with the interior cavity relatively unsolvated. A test of this concept is provided by comparing the geometrical structures of a neat spherand and those of the corresponding spheraplexes. This has been done for spherand 3.26 (C48H48O6; formula not shown), {3.26[Liþ]}Cl and {3.26[Naþ]}CH3SO4C6H5CH3 (Trueblood, Maverick and Knobler, 1991). The spherand moiety maintains nearly the same shape in all three crystals; the cavity is large enough to enclose only lithium or sodium ions. In one type of chemical modification of the prototype spherand the six methoxy groups of the interior were replaced by fluorines; the structure of an intermolecular complex of composition C42H30F62CH2Cl2, with interstitial disordered methylene chlorides, has been reported (Trueblood, Maverick and Knobler, 1991). In another type of modification, augmented spherands were prepared by replacing two pairs of ‘‘meta’’ methoxy groups by –OXO– bridges, where X ¼ –CH2CH2OCH2CH2–, –CH2CH2CH2– or –CH2CH2CH2CH2–. Crystal structures were reported for C52H52O8 (VOWVOS), C52H52O8LiCl3H2O (CAWRIB20). C50H48O6LiFeCl40.5CH2Cl2 (CAWREX20), C52H52O6LiCl3(C6H6) (VOWWEJ) and C50H48LiOþ 6 Cl 3(toluene) (VOWWIN). (VOWWEJ (reported as C2/c) and VOWWIN (reported as I2/a) are isomorphous (similar reduced triclinic cells). Lithium ions are enclosed in the interiors of all these hosts. These structures are extremely rigid, and the aromatic rings are very deformed; ˚ , 10% less than the sum of some intramolecular oxygen . . . oxygen contacts are about 2.5 A the van der Waals radii (Knobler, Maverick and Trueblood, 1992).

3.4.6 Carcerands and carceplexes Carcerands are noncollapsible molecular cells whose interiors are large enough to contain molecules or ions as guests and whose closed surfaces contain pores that are too small to allow guest molecules to enter or depart from their interiors without making or breaking covalent bonds; the concept was first proposed in 1983 (Cram, 1983) and a general account of the reduction of the idea to practice has been given by Maverick and Cram (1996b). A carceplex is composed of a carcerand containing at least one guest (prisoner)

60


molecule in its interior. The carceplexes were made by shell-closing two identical bowl-shaped cavitands, each containing on their rims four phenolic hydroxyls (Scheme 3.10; Sherman, Knobler and Cram, 1991). The cavitands also contain pendant functional groups for enhancement of their solubilities. In each synthesis, one molecule of solvent was incarcerated, the shell closure being templated by the guest ultimately enclosed. C4

H

HO O

2X

O

O

OH

O

O

OH

O

O

OH O

H

C CH2 CH2

C

H

CH2

H

C CH2

CH2

CH2

C

CH2 CH2 CH2 CH2 H C C

CH2 H

CH2

CH2

CH2

C

C

H

O O

CH2BrCl

H CH2

O

O C2 O

O

O

O

O

O G

CH2

O

CH2

O

O

CH2

O

O CH2

O O

O

O

CH2

O O

O

σh

O O

H

C CH2 CH2

3.27

C CH2 CH2

H

H

C

C

CH2

CH2

CH2

H

CH2

3.28

Scheme 3.10

Carceplexes containing (CH3)2NCOCH3, (CH3)2NCHO and dimethyl sulphoxide were prepared in solution; in earlier work with a different carcerand, insoluble complexes with (CH3)2NCHO, (CH2)4O, Csþ, argon and ClCF2CF2Cl had been prepared (Cram, Karbach, Kim, Baczynskyj and Kalleymeyn, 1985; Cram, Karbach, Kim, Baczynskyj, Marti, Sampson and Kalleymeyn, 1988). The crystal structure of {3.28.[(CH3)2NCOCH3]}0.5 CHCl3 (additional, but not fully identified, solvent molecules ˚, were present) has been determined (triclinic, a ¼ 16.302(3), b ¼ 18.940(3), c ¼ 23.182(4) A 1; R ¼ 0.184 at time of publication; ¼ 90.22(1), ¼ 93.97(1), g ¼ 102.38(1) , Z ¼ 2, P JILZIN). The {carcerand.[guest]} complex is shown in the stereodiagram of Fig. 3.26. ˚ and The cavity has the shape of a prolate ellipsoid of revolution with long axis 10.9 A ˚ . The upper half of the host is rotated by 15 with respect to the lower short axis 6.2 A half, thus making the molecule chiral. That the guest is well and truly incarcerated was demonstrated by heating the complex in C6D5NO2 solution to 160 , subsequent cooling producing no change in the NMR spectrum. In Cram’s view the interior of the carceplex is a definable mixture of free space and space-filling guests – ‘‘guest plus vacuum in varying proportions’’ – and thus constitutes a new phase of matter (see Cram and Cram (1994), p. 148). The two reacting tetrols (3.27) in Scheme 3.10 can be joined by hydrogen bonds rather than by covalent linkages. This has been done by carrying out the reaction in the presence of 1,8-diazabicyclo[5.4.0]undec-7-ene, which acts as a proton acceptor and forms cations (C9H17N2þ) while the two halves of the anionic carcerand are joined by charged hydrogen bonds (Chapman, Olovsson, Trotter and Sherman, 1998). Pyrazine was found to be the most strongly bound guest of those tried, and its complex (HIMJES; P4cc, Z ¼ 4; {C36H32O12(C36H28O12)4[C4H4N2]4(C9H17Nþ 2 )2(C6H5NO2)4H2O}) was about

HEMICARCERANDS AND HEMICARCEPLEXES

61

{3.28·[(CH3)2NCOCH3]}

Fig. 3.26. Stereoscopic side view of the {carcerand[guest]} complex {3.28[(CH3)2NCOCH3]} (JILZIN); the guest (dark outline) was not well enough defined to allow distinction among C, N and O atoms. Hydrogens have been omitted in this diagram; if included, then it is clear that there is no chance of the guest escaping without breaking bonds. (Reproduced from Sherman, Knobler and Cram, 1991.)

2000 times more stable than that of benzene. Crystal structure analysis showed that the pyrazines (disordered over two orientations) had their nitrogens in the equatorial plane of the host; they were involved in weak hydrogen bonding to half of the hydrogens of the intrabowl methylenes.

3.5 Hemicarcerands and hemicarceplexes 3.5.1 Overview A hemicarcerand is a host molecule that combines an enforced inner cavity large enough to accommodate solvent molecules with a portal which allows guest entry and exit at temperatures above 100 . This definition has a functional rather than a chemical structural basis and two types have so far been synthesized. In the first type, the reacting tetrols of Scheme 3.10 (above) are replaced by triols, with the consequence that the dimer product of the shell-closing reaction will have an opening at one corner of its waistband, the inner space of the cavity no longer being entirely cut off from its surroundings as happens in the carcerands. 3.29 is an example of a Type I hemicarcerand host molecule

62


(Cram, Tanner and Knobler, 1991). It was suggested that the term constrictive binding be used for the steric repulsions (decomplexation activation energy) that must be overcome in the decomplexation process.

H2C H

O

H2C

H2C CH2 CH2 H C

H

C

O O

O

C

O

CH2 Hb O

O

C H CH2 CH2 H2C H2C H

O O

G

Ha O

O

O

O

O O CH2

H2C CH2 CH2 H C C

O

O

O

O O CH2 O O

C C H CH2 CH2 H2C H2C H

3.29

Scheme 3.11

Type I hemicarceplexes with 3.29 as host and containing (CH3)2NCOCH3, (CH3)2NCHO, dimethyl sulphoxide, (CH3)CN, CH2Cl2, CH2Br2, CS2, Xe (Ks 200 M1) and Ar were prepared in solution, isolated and characterized. The crystal structure of {3.29[(CH3)2NCOCH3]}2CH3CN2CHCl3 was determined (orthorhombic, ˚ , Z ¼ 8, Pbna; R ¼ 0.168 at time of a ¼ 20.455(5), b ¼ 20.773(5), c ¼ 30.307(8) A publication; KEHDAC10; Cram, Tanner and Knobler, 1991). The {hemicarcerand[guest]} complex has an overall shape similar to that shown in the stereodiagram of Fig. 3.26. NMR studies showed the formation in solution of hemicarceplexes with nitrogen (Ks 180 M1), oxygen (Ks 44 M1) and H2O. Type II hemicarcerands have been prepared in which the relatively short equatorial linking –OCH2O– groups were replaced by the longer –CH2SCH2– groups (Bryant, Blanda, Vincenti and Cram, 1991). Enantiomerically pure analogs were made by using (R)- or (S)-2,2’-bis(bromomethyl)-1,10 -binaphthyl as linking groups. In somewhat later work analogs were produced in which the four linking groups were (1,3-CH¼N–C6H4– N¼CH–), which is rigid, and where portals and inner-space volume were large enough to permit entry and occupation by hosts such as [2.2]paracyclophane, ferrocene, adamantane and camphor (Quan and Cram, 1991). Other analogs have the semi-mobile o-xylyl {1,2–OCH2C6H4CH2O–)} as linking group (Cram, Blanda, Paek and Knobler, 1991; VURBUF; Pcan, Z ¼ 4; C160H136O2418(C4H9NO)) (hemicarcerand abbreviated as 3.30, formula not given; C4H9NO is dimethylacetamide); a more recent example encapsulates hydrocarbons with molecular weights greater than 200 (Cram, Jaeger and Deshayes, 1993). The hosts are all tetra-linked and hence are analogous, from the standpoint of chemical structure, to the carcerand 3.28. Nevertheless, they are named as hemicarcerands on the functional ground that entry and exit of the guest occurs without breaking covalent bonds. Qualitative information about structural recognition of guest by host 3.29 is provided by the facile incarceration of p-xylene compared to the nonincarceration of the other two


63

Table 3.5. Thermodynamic parameters (solution in 1,2(CD3)2C6D4 at 100 ) of the 3.30 hemicarceplexes referred to in the text. The units of G and H are kJ/ mol and of S J/mol K. Data from Cram, Blanda, Paek and Knobler, 1991 Guest

G

H

S

(CH3)2NCOCH3 CH3CH2O2CCH3 CH3COCH2CH3 C6H5CH3

15.5 15.9 22.2 14.2

6.3 13.0 10.5 þ9.2

þ25 þ8 þ31 þ63

xylene isomers. Some interesting thermodynamic parameters have been measured for hemicarceplexes of the 3.30 host (Cram, Blanda, Paek and Knobler, 1991), using 500 Mhz 1 H NMR to measure equilibrium constants for complexation and their temperature dependences, giving free energies, enthalpies and entropies of complexation (Table 3.5). The first three of these hemicarceplexes are both enthalpy and entropy stabilized; the toluene guest is entropy stabilized. The explanation advanced is that the positive carbons of the carbonyl groups of the first three guest molecules are electronically complementary to the 16 inward-turned unshared electron pairs of the oxygens of the eight ArOCH2 groups of the host. The unfavourable enthalpy of the toluene complex was attributed to the lack of complementary binding between the flat surface of the guest and the concave inner surfaces of the host. The positive entropies were attributed to two sources: firstly, solvated guest releases solvent molecules which become dispersed in the solvent, increasing the entropy, and, secondly, there is a contribution from the entropy of dilution of the empty space within the cavity – the large empty space of the cavity is broken up into many smaller empty spaces scattered among the solvent molecules. The kinetics of decomplexation, and their dependence on temperature were also measured by NMR and the free energies, enthalpies and entropies of activation for # ¼ disassociation determined. The constrictive binding free energy was defined as Gassoc # Gdisassoc (G ) and is the free energy of the transition state for association (relative to the uncomplexed state). It appears to depend little on the nature of the guest and derives mostly from the change from the unwrapped state of the empty host to the wrapped state # taken up by the hemicarceplex with included guest. The values found for Gassoc were about 100 kJ/mol, about 3/4 of which could be ascribed to an enthalpic contribution and 1/4 to an entropic contribution. Perhaps the most versatile of the currently available hemicarcerands is 3.31 (formula not shown) which differs from 3.29 in that the equatorial belt has four (instead of one) CH2 groups in each of the four (instead of three) vertical linkages (Robbins, Knobler, Bellew and Cram, 1994); again the term ‘‘hemicarcerand’’ is used here for functional rather than structural reasons. 3.31 is the host used in the work described below on o-benzyne (Warmuth, 1997). Thirty complexes of 3.31 were isolated and characterised, and crystal structures reported for six of these. Five were isomorphous; the general formula was {3.31[guest]}2C6H5NO2, space group P21/c, cell dimensions (for p-I2C6H4 guest at ˚ , ¼ 98.199(4) , Z ¼ 2 (PIHYEK); 156 K) a ¼ 16.777(2), b ¼ 19.795(2), c ¼ 20.327(2) A the other guests in this group were C6H5NO2 (PIHYIO), o-BrC6H4OH (PIHYOU),

64


p-(CH3)2C6H4(PIHZAH) and (CH3)2NCOCH3(PIHZEL). The 3.31 host is here centrosymmetric, and guests lacking a centre must be disordered. The sixth crystal structure was of {3.31[6H2O]}4(o-xylene); these crystals were triclinic, with two formula units in space group P1 (is this a misprint for P 1?). There are no symmetry requirements on the 3.31 molecule and it was twisted by 15 about its polar axis; the water molecules apparently take up an octahedral arrangement within the cavity. Crystal structures of cavitands, caviplexes and hemicarcerands and hemicarceplexes with a variety of guests continue to be reported in illustration of the principles described above (e.g. Helgeson, Paek, Knobler, Maverick and Cram, 1996 (TENLED, TENLIH, TENLON, TENLUT, TENMAA). For example, TENLED is {C128H168O24[C2H6OS] 2(C6H5NO2)} and is described as a ‘‘dimethyl sulfoxide clathrate nitrobenzene solvate.’’ ˚ , ¼ 69.36, ¼ 88.81, The reduced triclinic cell is a ¼ 12.984, b ¼ 14.664, c ¼ 20.336 A g ¼ 65.18 , P1, Z ¼ 1. The host is centrosymmmetric with partial disorder of some n-pentyl groups; the DMSO is disordered over two sites related by a center. Other examples, given by Helgeson, Knobler, and Cram (1997), are RAGXAY, RAGXEC, RACXIG, and RAGXOM. A general survey of Carceplexes and Hemicarceplexes has been given by Jasat and Sherman (1999). 3.5.2

The taming of cyclobutadiene, and of o-benzyne

Sherman, Knobler and Cram (1991) closed their report on carcerands with the prescient remark that ‘‘the inner phases of carcerands are unusual and interesting places where chemical reactions might be carried out.’’ A sequel was not long delayed – cyclobutadiene (C4H4) was reported as tamed but reactive within the ‘‘inner space’’ of 3.29 (Cram, Tanner and Thomas, 1991). The cyclobutadiene was synthesized in situ (within the inner cavity of the hemicarcerand) by irradiation (75 W xenon lamp, 25 ) of the 1 : 1 hemicarcerplex of 3.29 with -pyrone. O O

O hν

hν O

–CO2

Scheme 3.12

The -pyrone guest decomposes to C4H4 and CO2, the latter escaping from the cavity. The (Z)-OHCCH¼CHCHO was produced by reaction of C4H4 with O2 gas. The reactions were followed by NMR (Fig. 3.27). The sharp singlet in the NMR spectrum at ¼ 2.3 is assigned to singlet state cyclobutadiene on the basis of the following evidence. The host’s inward-pointing Ha and Hb protons give sharp doublets in the NMR spectrum of {3.29[C4H4]} at ¼ 4.27, 4.36, whereas triplet C4H4 would broaden and shift these signals, as does triplet oxygen in the spectrum of {3.29[O2]}. The sharpness of the signals at ¼ 4.27, 4.36 shows that the cyclobutadiene is rotating rapidly on the 1H NMR time scale about all its axes. {3.29[C4H4]} heated in (D8)THF at 220 gave free cyclooctatetraene (identified, inter alia, by its ‘‘pungent and characteristic odor’’), through intermediate formation of the cyclobutadiene dimer.

65


3.29

C2H2

O

O

H

H

3.29 H

7.0

6.5

6.0

5.5

5.0

H

4.5 d

4.0

3.5

3.0

2.5

2.0

Fig. 3.27. 1H NMR spectra (500 MHz, CDCl3, 60 ) of the hemicarcerplexes of 3.29 with cyclobutadiene (CH)4 and (Z)-OHCCH¼CHCHO. (Reproduced from Cram, Tanner and Thomas, 1991.)

This remarkable achievement has been followed by the synthesis of o-benzyne (C6H4; IUPAC name 1,2-didehydrobenzene) within the inner cavity of a hemicarcerand 3.31. The precursor was benzocyclobutendione, which was complexed with 3.31 through a molten phase and then photolysed ( > 400 nm) at 77K to give {3.31[benzocyclopropenone]} as the only product, the structure being confirmed by (an as-yet unpublished) crystal structure analysis. This intermediate complex was then further photolysed (77K, < 297 nm, 88 hours, 300 W xenon arc lamp) to give incarcerated benzyne, identified by its 1H and 13C NMR spectra. If the {3.31[o-benzyne]} product complex is allowed to warm up above 77K, then a host–guest reaction occurs in which the nominal triple bond of the o-benzyne is added across one of the aryl ether bars of the host cage (Warmuth, 1998). O O O

Scheme 3.13

˚ Singlet cyclobutadiene has a rectangular shape, with bond lengths of 1.34 and 1.60 A (Dunitz et al., 1988). Current very high level calculations of the structure of o-benzyne suggest that there is no bond length alternation, a conclusion in conflict with deductions from the NMR spectra which suggest a cumulene-type structure. These geometrical

66


details could possibly be identified from diffraction measurements on the hemicarcerplexes at low temperature. Studies of ‘‘chemistry in the inner space’’ continue, e.g. Warmuth, 2001.

3.5.3

Molecular mechanics and dynamics studies on the complexation and decomplexation processes

There is a wealth of information available about which hosts will complex (or not) with which guests, including studies of the chemistry, thermodynamics and kinetics. One would wish to translate these qualitative and quantitative results into molecular R

R

R

R

rigid ‘polar’ region OO

O

O O

O O

O

OO

O

O

O

O

O

O

OO

O

O

R

O

O O O

O

R

O

R

flexible ‘temperate’ and ‘equatorial’ regions

OO

R

1. R = PHCH2CH2 2. R = H

(a)

(b)

(c)

Fig. 3.28. (Above) Schematic diagram of the hemicarcerand; the pendant groups R ¼ PhCH2CH2 act to increase the solubility of the host 3.32 #1 in the experimental studies, while R is set to H in the stripped-down version of the molecule 3.32 #2 used for the calculations. (Below) Side views of three low-energy conformations of a space-filling model of 3.32 #2. (Reproduced from Sheu and Houk, 1996.)

COMPARISONS OF CONCEPTS

67

terms – how do the guests enter the hosts, how are they enabled to remain in the cavities and how can they leave? The hemicarcerands are very suitable substrates for this sort of study – the molecules have both rigid and flexible portions, and the portals have reasonably well defined shapes and sizes. The methods of molecular mechanics and molecular dynamics have been applied by Sheu and Houk (1996) to answer these questions, using the ‘‘stripped down’’ hemicarcerand (3.32 #2) as host (Fig. 3.28). The global minimum conformation of the host (2a in Fig. 3.28) has D2h symmetry; 2b has C2h symmetry and is 31.8 kJ/mol higher in energy than 2a, while 2c has D2h symmetry and is 38.5 kJ/mol higher in energy than 2a. Structure 2b is very similar to the analogous hemicarcerand structures found by X-ray diffraction. Sheu and Houk (1996) quote twelve examples of mono- and bicyclic guests which form isolable complexes with 3.32 #1, and twenty-eight examples of similar molecules which do not; 1,2-dimethoxy-4-bromobenzene is an example from the first group and 1,2-dimethoxybenzene from the second. It is the orientation of the intrahemispheric bridges (–OCH2CH2OCH2CH2O–) which is important in determining how easy it is for guest molecules to pass in and out through the portals of the host. Cram has pointed out that preorganization of the shape of the host and stereo-electronic complementarity between host and guest are the important factors determining the binding power of a particular host-guest combination. 3.32 is highly preorganized with two rigid polar caps, while the flexibility of the four intrahemispheric bridges allows the shape of the portals and cavity to range from square to rhomboid to rectangular (conformers 2a, 2b, 2c). Three processes must be considered. Firstly, can potential guests bind within the cavities? Calculation showed that all 12 guests and 28 nonguests mentioned above could form complexes with binding energies of 50–100 kJ/mol. Thus there must be reasons other than thermodynamics that allow some of these potential guests to form isolable complexes, while others do not. The second factor is whether the potential guest can enter the cavity through the available portals, taking into account the flexibility of the gates formed by the intrahemispheric bridges. For many guests, the barrier to entry is high even with the gates open. The third factor is possible escape of complexed guest during the purification process. This is considered to be the reason why many guests with otherwise favourable properties do not form isolable complexes.

3.6 Comparisons of concepts The different types of enclosure molecules introduced in this chapter, their unfamiliar nomenclature and their resemblances to and differences from established concepts in host–guest solid state chemistry can be somewhat bewildering. We have used the succint comparison of concepts made by Sherman, Knobler and Cram (1991) (see also Cram and Cram (1994; pp. 147–148)) as the basis for a more extended comparison organised in parallel with the treatment of this and subsequent chapters. The crown ethers considered here are single-ring hosts which wrap around cations and form intermolecular hydrogen-bonded complexes with suitably substituted organic guests. True intramolecular enclosure occurs only when the crown ether is large enough to bend back on itself to form a cleft into which the guest enters; a number of hosts (not necessarily crown ethers) of this ‘‘molecular tweezer’’ type have been prepared. Larger crown ethers, particularly benzocrown ethers, can, in addition to the act of enclosure, interact

68


with suitable guests by charge–transfer forces familiar from the classical chemistry (solution and solid state) of polycyclic aromatic hydrocarbons. Stoddart’s inception and extension of these ideas may well be leading to a new era of supramolecular chemistry based on manipulation of smaller units. Cryptands have three rings (connectors) joined at two apexes (spacers) and form triangular baskets with open sides; four connectors would give baskets of rectangular cross section. Here entry and exit of guests will be relatively easy, and conformational reorganization can be expected during the complexation process. The especial contribution of the Cram group lies in synthesizing molecular baskets with much more varied shapes and, usually, with restricted possibilities of entry and exit. Spherands are hollow and rigid baskets with pole–dipole interactions with guests; so far only cations or relatively small organic molecules have been enclosed in spherands. Carcerplexes contain guests which cannot escape without breaking covalent bonds of the carcerand host molecule; these baskets have a very tight weave. Hemicarcerands are analogous baskets but contain portals which are flexible enough to allow entry and exit of guests at higher temperatures and thus exchange between interior and exterior occurs without bond cleavage. To quote from Sherman et al. (1991) – ‘‘[The] existence and stability [of carciplexes] do not depend on host–guest attractions . . . other than gross size complementarity, but on physical envelopment of guests during [the] shell closures leading to [the formation] of carciplexes.’’ Cram contends that the interiors of carciplexes constitute a new form of matter. The comparison can be completed by considering the place of other types of enclosure. Clathrates are crystalline compounds where the guests are included in cavities left between host molecules linked by hydrogen bonds or van der Waals forces. Generally the sublattice of host molecules has a structure different from that of the neat host, and hence clathrates are secondary solid solutions of guest in host (if of variable composition), or ‘‘phase rule compounds’’ (if of fixed composition). There are some rare examples of primary solid solution. There is no special interaction between host and guest in solution. Zeolites constitute a special case of clathrates where the host structural units are linked by rigid covalent (or ionic) bonds, leading to the crystal becoming a ‘‘giant molecule’’ with many interstices; the complex is then a primary solid solution of guest in host.

References Abbott, S. J., Barrett, A. G. M., Godfrey, C. R. A., Kalindjian, S. B., Simpson, G. W. and Williams, D. J. (1982). J. Chem. Soc., Chem. Commun., pp. 796–797. Allwood, B. L., Colquhoun, H. M., Doughty, S. M., Kohnke, F. H., Slawin, A. M. Z., Stoddart, J. F., Williams, D. J. and Zarzycki, R. (1987). J. Chem. Soc., Chem. Commun., pp. 1054 –1058. Allwood, B. L., Kohnke, F. H., Stoddart, J. F. and Williams, D. J. (1985). Angew. Chem. Int. Ed. Engl., 24, 581–583. Allwood, B. L., Shahriari-Zavareh, H., Stoddart, J. F. and Williams, D. J. (1987). J. Chem. Soc., Chem. Commun., pp. 1058–1061. Allwood, B. L., Spencer, N., Shahriari-Zavareh, H., Stoddart, J. F. and Williams, D. J. (1987a). J. Chem. Soc., Chem. Commun., pp. 1064–1067.

REFERENCES

69

Allwood, B. L., Spencer, N., Shahriari-Zavareh, H., Stoddart, J. F. and Williams, D. J. (1987b). J. Chem. Soc., Chem. Commun., pp. 1061–1064. Amabilino, D. B., Ashton, P. R., Brown, C. L., Co´rdova, E., Godıńez, L. A., Goodnow, T. T., Kaifer, A. E., Newton, S. P., Pietraszkiewicz, M., Philp, D., Raymo, F. M., Reder, A. S., Rutland, M. T., Slawin, A. M. Z., Spencer, N., Stoddart, J. F. and Williams, D. J. (1995). J. Am. Chem. Soc., 117, 1271–1293. Amabilino, D. B. and Stoddart, J. F. (1995). Chem. Rev., 95, 2725–2828. Anelli, P. L., Ashton, P. R., Ballardini, R., Balzani, V., Delgado, M., Gandolfi, M. T., Goodnow, T. T., Kaifer, A. E., Philp, D., Pietraszkiewicz, M., Prodi, L., Reddington, M. V., Slawin, A. M. Z., Spencer, N., Stoddart, J. F., Vicent, C. and Williams, D. J. (1992). J. Am. Chem. Soc., 114, 193–218. Anelli, P. L., Ashton, P. R., Spencer, N., Slawin, A. M. Z., Stoddart, J. F. and Williams, D. J. (1991). Angew. Chem. Int. Ed. Engl., 30, 1036–1039. Asakawa, M., Ashton, P. R., Hayes, W., Janssen, H. M., Meijer, E. W., Menzer, S., Pasini, D., Stoddart, J. F., White, A. J. P. and Williams, D. J. (1998). J. Am. Chem. Soc., 120, 920–931. Asakawa, M., Brown, C. L., Menzer, S., Raymo, F. M., Stoddart, J. F. and Williams, D. J. (1997). J. Am. Chem. Soc., 119, 2614–2627. Ashton, P. R., Ballardini, R., Balzani, V., Be´loradský, M., Gandolfi, M. T., Philp, D., Prodi, L., Raymo, F. M., Reddington, M. V., Spencer, N., Stoddart, J. F., Venturi, M. and Williams, D. J. (1996). J. Am. Chem. Soc., 118, 4931–4951. Ashton,P.R.,Brown,C.L.,Chrystal,E.J.T.,Goodnow,T.T.,Kaifer,A.E.,Parry,K.P.,Slawin,A.M.Z., Spencer, N., Stoddart, J. F. and Williams, D. J. (1991). Angew. Chem. Int. Ed. Engl., 30, 1039–1042. Ashton, P. R., Brown, G. R., Isaacs, N,. S., Guiffrida, D., Kohnke, F. H., Matthias, J. P., Slawin, A. M. Z., Smith, D. R., Stoddart, J. F. and Williams, D. J. (1992). J. Am. Chem. Soc., 114, 6330–6353. Ashton, P. R., Chrystal, E. J. T., Mathias, J. T., Parry, K. P., Slawin, A. M. Z., Spencer, N., Stoddart, J. F. and Williams, D. J. (1987). Tetr. Letts., 28, 6367–6370. Ashton, P. R., Goodnow, T. T., Kaifer, A. E., Reddington, M. V., Slawin, A. M. Z., Spencer, N., Stoddart, J. F., Vicent, C. and Williams, D. J. (1989). Angew. Chem. Int. Ed. Engl., 28, 1396–1399. Ashton, P. R., Philp, D., Reddington, M. V., Slawin, A. M. Z., Spencer, N., Stoddart, J. F. and Williams, D. J. (1991). J. Chem. Soc., Chem. Commun., pp. 1680–1683. Ashton, P. R., Reddington, M. V., Slawin, A. M. Z., Stoddart, J. F. and Williams, D. J. (1988). Angew. Chem. Int. Ed. Engl., 27, 1550–1553. Ashton, P. R., Slawin, A. M. Z., Spencer, N., Stoddart, J. F. and Williams, D. J. (1987). J. Chem. Soc., Chem. Commun., pp. 1066–1069. Behr, J.-P. (1994). Editor, The Lock and Key Principle. The State of the Art – 100 Years On, Wiley, Chichester. Berscheid, B., Nieger, M. and Vo¨gtle, F. (1991). J. Chem. Soc., Chem. Commun., pp. 1364–1366. Bottino, F., Finocchiaro, P., Lipkowski, J., Mamo, A. and Pappalardo, S. (1991). J. Incl. Phenom. and Recogn. in Chem., 11, 41–48. Bradshaw, J. S., Izatt, R. M., Bordunov, A. V., Zhu, C. Y. and Hathaway, J. K. (1996). ‘‘Crown ethers,’’ in Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, Vol. 1, pp. 35–95. Bryant, J. A., Blanda, M. T., Vincenti, M. and Cram, D. J. (1991). J. Am. Chem. Soc., 113, 2167–2172. Bu¨hner, M., Geuder, W., Gries, W.-K., Hu¨nig, S., Kock, M. and Poll, T. (1988). Angew. Chem. Int. Ed. Engl., 27, 1553–1556. Burley, S. K. and Pesko, G. A. (1986). J. Am. Chem. Soc., 108, 7995–8001. Bush, M. A. and Truter, M. R. (1972). J. Chem. Soc., Perkin II, pp. 345–350.

70


Canceill, J., Cesario, M., Collet, A., Guilhem, J. and Pascard, C. (1985). J. Chem. Soc., Chem. Commun., pp. 361–363. Canceill, J., Cesario, M., Collet, A., Guilhem, J. Riche, C. and Pascard, C. (1986). J. Chem. Soc., Chem. Commun., pp. 339–341. Canceill, J., Cesario, M., Collet, A., Guilhem, J., Lacombe, L., Lozach, B. and Pascard, C. (1989). Angew. Chem. Int. Ed. Engl., 28, 1246–1248. Canceill, J., Lacombe L. and Collet, A. (1987). J. Am. Chem. Soc., 107, 6993–6996. Chapman, R. G., Olovsson, G., Trotter, J. and Sherman, J. C. (1998). J. Am. Chem. Soc., 120, 6252–6260. Christensen, J. J., Eatough, D. J. and Izatt, R. M. (1974). Chem. Rev., 74, 351–384. Cochran, J. E., Parrott, T. J., Whitlock, B. J. and Whitlock, H. W. (1992). J. Am. Chem. Soc., 114, 2269–2270. Collet, A. (1987). Tetrahedron (Report No. 226), 43, 5725–5759. Colquhoun, H. M., Doughty, S. M., Maud, J. M., Stoddart, J. F., Williams, D. J. and Wolstenholme, J. B. (1985). Isr. J. Chem., 25, 15–26. Colquhoun, H. M., Doughty, S. M., Stoddart, J. F. and Williams, D. J. (1984). Angew. Chem. Int. Ed. Engl., 23, 235–236. Colquhoun, H. M., Goodings, E. P., Maud, J. M., Stoddart, J. F., Wolstenholme, J. B. and Williams, D. J. (1985). J. Chem. Soc., Perkin II, pp. 607–624. Cram, D. J. (1983). Science, 219, 1177–1183. Cram, D. J. (1986). Angew. Chem. Int. Ed. Engl., 25, 1039–1057. Cram, D. J. and Cram, J. M. (1994). Container Molecules and their Guests, No. 4 of Monographs in Supramolecular Chemistry, Series Editor, J. Fraser Stoddart, London: The Royal Society of Chemistry. Cram, D. J., Blanda, M. T., Paek, K. and Knobler, C. B. (1992). J. Am. Chem. Soc., 114, 7765–7773. Cram, D. J., Jaeger, R. and Deshayes, K. (1993). J. Am. Chem. Soc., 115, 10111–10116. Cram, D. J., Kaneda, T., Helgeson, R. C., Brown, S. B., Knobler, C. B., Maverick, E. F. and Trueblood, K. N. (1985). J. Am. Chem. Soc., 107, 3645–3657. Cram, D. J., Karbach, S., Kim, Y. H., Baczynskyj, L. and Kalleymeyn, G. W. (1985). J. Am. Chem. Soc., 107, 2575–2576. Cram, D. J., Karbach, S., Kim, Y. H., Baczynskyj, L., Marti, K., Sampson, R. M. and Kalleymeyn, G. W. (1988). J. Am. Chem. Soc., 110, 2554–2560. Cram, D. J., Karbach, S., Kim. H.-E., Knobler, C. B., Maverick, E. M., Ericson, J. L. and Helgeson, R. E. (1988). J. Am. Chem. Soc., 110, 2229–2237. Cram, D. J., Tanner, M. A., Keipert, S. J. and Knobler, C. B. (1991). J. Am. Chem. Soc., 113, 8909–8916. Cram, D. J., Tanner, M. A. and Knobler, C. B. (1991). J. Am. Chem. Soc., 113, 7717–7727. Cram, D. J., Tanner, M. A. and Thomas, R. (1991). Angew. Chem. Int. Ed. Engl., 30, 1024–1027. Cram, D. J. and Trueblood, K. N. (1981). Top. Curr. Chem., 98, 43–106. Diederich, F. (1988). Angew. Chem. Int. Ed. Engl., 27, 362–386. Dietrich, B. (1996). ‘‘Cryptands,’’ in Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, Vol. 1, pp. 153–211. Dietrich, B., Lehn, J. -M. and Sauvage, J. -P. (1968). Tetrahedron Letts, pp. 2885–2888. Dobler, M. (1981). Ionophores and their Structures, Wiley, New York. Dunitz, J. D., Kru¨ger, C., Irngartinger, H., Maverick, E. F., Wang, Y. and Nixdorf, M. (1988). Angew. Chem. Int. Ed. Engl., 27, 387–389. Goldberg, I. (1980). Cryst. Struct. Commun., 9, 1201–1207. Goldberg, I. (1986). J. Incl. Phenom., 4, 191–198.

REFERENCES

71

Grootenhuis, P. D. J., Eerden, van, J., Dijkstra, P. J., Harkema, S. and Reinhoudt, D. N. (1987). J. Am. Chem. Soc., 109, 8044–8051. Hamilton, D. G., Feeder, N., Prodi, L., Teat, S. J., Clegg, W. and Sanders, J. K. M. (1998). J. Am. Chem. Soc., 120, 1096–1097. Harmata, M. and Barnes, C. L. (1990). J. Am. Chem. Soc., 112, 5655–5657. Helgeson, R., Knobler, C. B. and Cram, D. J. (1997). J. Am. Chem. Soc., 119, 3229–3244. Helgeson, R., Paek, K., Knobler, C. B., Maverick, E. F. and Cram, D. J. (1996). J. Am. Chem. Soc., 118, 5590–5604. Hilgenfeld, R. and Saenger, W. (1982). Top. Curr. Chem., 101, 1–82. Hirotsu, K., Kamitori, T., Higuchi, T., Tabushi, I., Yamamura, K. and Nonoguchi, H. (1984). J. Incl. Phenom., 2, 215–222. Hof, F., Craig, S. L., Nuckolls, C. and Rebek, J., Jr. (2002). Angew. Chem., Int. Ed. Eng., 41, 1488–1508. Hu¨bner, M., Geuder, W., Gries, W.-K., Hu¨nig,S., Kock, M. and Pohl, T. (1989). Angew. Chem. Int. Ed. Engl., 27, 1553–1556. Itai, A., Tanaka, Y. and Iitaka, Y. (1979), American Crystallographic Association, Winter Meeting. Abstract PA32. Izatt, R. M., Bradshaw, R. D., Nielsen, S. A., Lamb, J. D., Christensen, J. J. and Sen, D. (1985). Chem. Revs., 85, 271–339. Izatt, R. M., Bradshaw, R. D., Pawlak, K., Bruening, R. L. and Tarbet, B. J. (1992). Chem. Revs., 92, 1261–1354. Izatt, R. M., Pawlak, K., Bradshaw, R. D. and Bruening, R. L. (1991). Chem. Revs., 91, 1721–2085. Izatt, R. M., Pawlak, K., Bradshaw, R. D. and Bruening, R. L. (1995). Chem. Revs., 95, 2529–2586. Jasat, A. and Sherman, J. C. (1999). Chem. Revs., 99, 931–967. Jaswinski, J., Blacker, A. J., Lehn, J.-M., Cesario, M., Guilhem, J. and Pascard, C. (1987). Tetrahedr. Letts., 28, 6057–6060. Knobler, C. B., Maverick, E. F. and Trueblood, K. N. (1992). J. Incl. Phenom., 12, 341–360. Koenig, K. E., Lein, G. M., Stuckler, P., Kaneda, T. and Cram, D. J. (1979). J. Am. Chem. Soc., 101, 3553–3566. Kohnke, F. H., Mathias, J. P. and Stoddart, J. F. (1989). Angew. Chem. Int. Ed. Engl., Adv. Mat., 28, 1103–1110. Kohnke, F. H., Stoddart, J. F., Allwood, B. L. and Williams, D. J. (1985). Tetr. Letts., 26, 1681–1684. Krieger, C. and Diederich, F. (1985). Chem. Ber., 118, 3620–3631. La¨msa¨, M., Huuskonen, J., Rissanen, K. and Pursiainen, J. (1998). Chem. Eur. J., 4, 84–92. Lehn, J.-M., Me´ric, R., Vigneron, J.-P., Bkouche-Waksman, I. and Pascard, C. (1991). J. Chem. Soc., Chem. Commun., pp. 62–64. Maverick, E. and Cram, D. J. (1996a). ‘‘Spherands: Hosts preorganized for binding cations,’’ in Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, Vol. 1, pp. 213–244. Maverick, E. and Cram, D. J. (1996b). ‘‘Carcerands and Hemicarcerands: Hosts that imprison molecular guests,’’ in Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, Vol. 2, pp. 367–418. Moran, J. R., Ericson, J. L., Dalcanale, E., Bryant, J. A., Knobler, C. B. and Cram, D. J. (1991). J. Am. Chem. Soc., 113, 5707–5714. Nonoguchi, H., Yamamura, K., Tabushi, I., Higuchi, T. and Hirotsu, K. (1992). Bull. Chem. Soc. Jpn., 65, 805–811. Odashima, K. and Koga, K. (1983). ‘‘Cyclophanes in host-guest chemistry’’ in Cyclophanes, Vol. 2, edited by P. M. Keehn and S. M. Rosenfeld, Academic Press, New York etc., (Vol. 45 of Organic Chemistry, edited by H. H. Wasserman), Chapter 11, pp. 629–678.

72


Odashima, K., Itai, A., Iitaka, Y. and Koga, K. (1980). J. Am. Chem. Soc., 102, 2504–2505. Odell, B., Reddington, M. V., Slawin, A. M. Z., Spencer, N., Stoddart, J. F. and Williams, D. J. (1988). Angew. Chem. Int. Ed. Engl., 27, 1547–1550. O’Krongly, D., Denmeade, S. R., Chiang, M. Y. and Breslow, R. (1985). J. Am. Chem. Soc., 107, 5544–5545. Ortholand, J.-Y., Slawin, A. M. Z., Spencer, N., Stoddart, J. F. and Williams, D. J. (1989). Angew. Chem. Int. Ed. Engl., 28, 1394–1395. Owen, J. D. (1984). Acta Cryst., C40, 951–955. Philp, D., Slawin, A. M. Z., Spencer, N., Stoddart, J. F. and Williams, D. J. (1991). J. Chem. Soc., Chem. Commun., pp. 1584–1586. Quan, M. L. C. and Cram, D. J. (1991). J. Am. Chem. Soc., 113, 2754. Reddington, M. V., Slawin, A. M. Z., Spencer, N., Stoddart, J. F., Vicent, C. and Williams, D. J. (1991). J. Chem. Soc., Chem. Commun., pp. 630–634. Reek, J. N. H., Engelkamp, H., Rowan, A. E., Elemans, J. A. A. W. and Nolte, R. J. M. (1998). Chem. Eur. J., 4, 716–722. Renault, A., Talham, D., Conceill, J., Batail, P., Collet, A. and Lajzerowicz, Janine, (1989). Angew. Chem. Int. Ed. Engl., 101, 1251–1252. Robbins, T. A., Knobler, C. B., Bellew, D. R. and Cram, D. J. (1994). J. Am. Chem. Soc., 116, 111–122. Rudkevich, D. M. (2002). Bull. Chem. Soc. Jpn., 75, 393–414. Schill, G. (1971). Catenanes, Rotaxanes and Knots, Academic Press, New York. Schwartz, E. B., Knobler, C. B. and Cram. D. J. (1992). J. Am. Chem. Soc., 114, 10775–10784. Seel, C. and Vo¨gtle, F. (1992). Angew. Chem. Int. Ed. Engl., 31, 528–549. Sheridan, R. E. and Whitlock, H. W. (1986). J. Am. Chem. Soc., 108, 7210–7211. Sheridan, R. E. and Whitlock, H. W. (1988). J. Am. Chem. Soc., 110, 4071–4073. Sherman, J. C., Knobler, C. B. and Cram. D. J. (1991). J. Am. Chem. Soc., 113, 2194–2204. Sheu, C. and Houk, K. N. (1996). J. Am. Chem. Soc., 118, 8056–8070. Slawin, A. M. Z., Spencer, N., Stoddart, J. F. and Williams, D. J. (1987). J. Chem. Soc., Chem. Commun., pp. 1070–1072. Stoddart, J. F. (1991) in Host–Guest Interactions: from Chemistry to Biology, Ciba Foundation Symposium 158, Wiley, Chichester, p. 5. Tabushi, I., Yamamura, K., Nonoguchi, H., Hirotsu, K. and Higuchi, T. (1984a). J. Am. Chem. Soc., 106, 2621–2625. Tabushi, I., Yamamura, K., Nonoguchi, H., Hirotsu, K. and Higuchi, T. (1984b). J. Incl. Phenom., 2, 85–89. Tanner, M. E., Knobler, C. B. and Cram, D. L. (1990). J. Am. Chem. Soc., 112, 1659–1600. Tanner, M. E., Knobler, C. B. and Cram, D. L. (1992). J. Org. Chem., 57, 40–46. Trueblood, K. N., Maverick, E. M. and Knobler, C. B. (1991). Acta Cryst., B47, 389–398. Uiterwijk, J. W. H. M., Staveren, C. J. van, Reinhoudt, D. N., Hertog, H. J. den, Jr., Kruise, L. and Harkema, S. (1986). J. Org. Chem., 51, 1575–1587. Vincenti, M., Dalcanale, E., Soncini, P. and Guigliemetti, G. (1990). J. Am. Chem. Soc., 112, 445–447. Vo¨gtle, F., Berscheid, R. and Schnick, W. (1991). J. Chem. Soc., Chem. Commun., pp. 414–416. Warmuth, R. (1997). Angew. Chem. Int. Ed. Engl., 36, 1347–1350. Warmuth, R. (1998). J. Chem. Soc., Chem. Commun., pp. 59–60. Warmuth, R. (2001). Chem. Eur. J., 7, 1209–1220. Weber, E., Franken, S., Puff, H. and Ahrendt, J. (1986). J. Chem. Soc., Chem. Commun., pp. 467–469. Whitlock, B. J. and Whitlock, H. W. (1990). J. Am. Chem. Soc., 112, 3910–3915.

Chapter 4 Cyclodextrins, and some analogs, as hosts

. . . The Gem, The Diadem, The Ring Enclosing All T.Traherne ((?)1636–1674) Summary: The -, - and -cyclodextrins are oligosaccharides containing respectively six, seven and eight glucose units linked between 1 and 4 positions to form macrocyclic rings of overall toroidal or truncated-cone shape; the secondary hydroxyl side is conventionally defined as the ‘‘head’’ of the cone. The inner and outer surfaces of the toroids are hydrophobic while the upper and lower faces are hydrophilic. When crystallized from aqueous solutions the cyclodextrins form inclusion complexes with a large variety of guest molecules, ranging in type from inert gases through polyiodide salts to aromatics. The crystals contain an appreciable number of water molecules which are hydrogen bonded together and also to the host molecules; the amount of water varies from complex to complex. Two types of general structure can be distinguished – the clathrates and the tunnel inclusion complexes – and each of these types contains a number of isomorphous or isostructural classes; about two-thirds of the complexes of known structure are of the tunnel type. The guest molecules are generally enclosed within the cavities of the toroids in a variety of packings; there are sometimes water molecules within the cavities in addition to those located between the cyclodextrins, and guests are sometimes found between the host molecules. If the guest is a salt, linear anions such as polyiodides are found within the cavities while the counter cations are included in the water network. Formation of hydrogen-bonded cyclodextrin dimers is a feature of the tunnel inclusion complexes and these dimers are mainly head-to-tail in the -cyclodextrin complexes, predominantly head-to-head in the -cyclodextrin complexes and with all three possibilities in the (mostly isomorphous) -cyclodextrin complexes so far studied. The catalytic properties of cyclodextrins in organic reactions and their mimicry of enzyme behavior is ascribed to the manner in which the guest is held in a fixed position and orientation by the host even in solution. Cyclodextrins have found wide use in the pharmaceutical industry because of their water solubility, the innocuous nature of their degradation products and, most importantly, their ability to encapsulate drug molecules and release these slowly. Analogs to the cyclodextrins are currently being developed.

4.1 Introduction 4.2 -Cyclodextrins as host 4.2.1 -Cyclodextrin as host in clathrate inclusion complexes 4.2.2 -Cyclodextrin as host in tunnel inclusion complexes 4.2.3 Chemically modified -cyclodextrins as hosts in inclusion complexes 4.3 -Cyclodextrins as host 4.3.1 -Cyclodextrin as host in clathrate inclusion complexes 4.3.2 -Cyclodextrin as host in tunnel inclusion complexes

74 79 80 84 90 95 97 100

74

C Y C L O D E X T R I NS , A ND S OM E AN A L O G S , A S H O S T S

4.3.3 Exceptional -cyclodextrin structures 4.3.4 Chemically modified -cyclodextrins as hosts in inclusion complexes 4.4 Rotaxanes and catenanes of cyclodextrins 4.5 -Cyclodextrins as host 4.5.1 -Cyclodextrin as host in clathrate inclusion complexes 4.5.2 -Cyclodextrin as host in tunnel inclusion complexes 4.5.3 Chemically modified -cyclodextrins as hosts in inclusion complexes 4.6 Larger cyclodextrins 4.7 Cyclic oligosaccharides as cyclodextrin analogs References

4.1

114 114 117 118 118 119 122 123 123 124

Introduction

Towards the end of the nineteenth century Villiers (1891) reported the isolation (in 3% yield) of a group of unusual nonreducing oligosaccharides from cultures containing Bacillus macerans grown on a medium rich in amylose; the enzymes involved are cyclodextrin glycosyltransferases (CGTases EC 2.4.1.19). The product compounds (now obtained in 25% yield) were shown by Schardinger (1904) to be cyclic oligosaccharides, of toroidal or doughnut shape, containing from six to twelve glucose units. They have been called Schardinger dextrins, cyclo[n]amyloses (CAn), cyclomalto-oligosaccharides and cyclodextrins ( for n ¼ 6, for n ¼ 7 and for n ¼ 8, often abbreviated as -, - and -CD). The enzymatic breakdown of starch to give cyclodextrins is illustrated in Fig. 4.1. We follow Chemical Abstracts in using the -, - and -cyclodextrin (CD) nomenclature, corresponding to cyclomaltohexaose, cyclomaltoheptaose and cyclomalto-octaose respectively. There are also some larger cyclodextrins, whose properties are now being explored (see Section 4.6 below); these are referred to by Greek letters or as CAn, with the latter system to be preferred. However, we have generally retained the CD system because of its familiarity. There is also currently an active development of synthetic cyclic oligosaccharides (Gattuso, Nepogodiev and Stoddart, 1998), which show similarities and differences to the cyclodextrins. The water solubility of the cyclodextrins is an important feature of their properties – 14.5 g/100 ml. water at 25 C, 1.85 and 23.2. All three show increasing solubility with temperature (0–90 C), with solubility at 85 C in the order > > (Diaz, Vargas-Baca and Gracia-Mora, 1994). The solubility is ascribed to dipole interactions between the hydrophilic upper and lower surfaces of the tori and bulk water molecules. Complexes (inclusion complexes of one kind or another, in which molecules of many kinds as well as salts can be found as guests) have been reported for -, - and -cyclodextrins and a considerable amount of structural work has been reported, in addition to much chemical and biochemical study, all of which has been accompanied by important industrial developments. Formation of the complexes is ascribed to replacement of water molecules within the hydrophobic inner surface of the torus by hydrophobic guest molecules. The guests in the cyclodextrin complexes are retained within the rings in the solid state and even in solution (at least on a dynamic basis) and this has made the complexes useful as catalysts in organic chemistry and as model systems for many enzyme reactions. The


75

14.6 Å 5.3 Å

OH O

O

HO

O HO OH

OH O HO

OH

O OH O

O

HO

OH

7.8 Å

HO O

HO

OH O HO

O

HO OH O

OH

O

O HO

O O O

O

O

O

O O

O

O O O

OH O OH HO

OH O HO

O O

CGTase (Enzyme)

O

O

O HO

OH

O HO

HO O

O

O

OH

O HO

O

O

15.4 Å 6.5 Å

O

O

HO

O

HO

O

OH O

OH O

OH HO O

O

O

7.8 Å

HO OH

O

O

OH

OH

O O O

O

O

O O

O

O

O

OH O

O

O OH HO

O

OOH O

Starch

OH O

O

OH

HO

HO

HO

O OH

O OH

OH

HO

17.5 Å 8.3 Å

OH

HO

HO OH O OH

7.8 Å

O

OH O OH HO O

O

O

OH

HO

α– cyclodextrin (top) Molecular Weight 973 Glucose Units 6 Specific Rotation [a]D25 150.5 Cavity Volume mL/g 0.10 6 H2O molecules included (in solution)

β–cyclodextrin (center) Molecular Weight 1135 Glucose Units 7 Specific Rotation [a]D25 162.5 Cavity Volume mL/g 0.14 11 H2O molecules included (in solution)

γ–cyclodextrin (center) Molecular Weight 1297 Glucose Units 8 Specific Rotation [a]D25 177.8 Cavity Volume mL/g 0.20 17 H2O molecules included (in solution)

Fig. 4.1. Enzymatic breakdown of starch to give -, - and -cyclodextrins, together with a summary of their properties. (Reproduced from Diaz,Vargas-Baca and Gracia-Mora, 1994.)

cyclodextrins have been used as encapsulating agents for slow-release drugs and for other pharmaceutical purposes. Cramer (1954, 1987) played a key role in initiating many of these developments. We shall emphasise the structural chemistry of the crystalline complexes and not attempt to duplicate the coverage of other aspects in the extensive contemporary literature (Senti and Erlander, 1964; Bender and Komiyama, 1978; Saenger, 1980, 1984; Saenger, Jacob, Gessler, Steiner, Hoffmann, Sanbe, Koizumi,

76


Smith and Takaha, 1998; Saenger and Steiner, 1998; Szejtli, 1982, 1988, 1998; Ducheńe, 1987; Stoddart and Zarycki, 1989; Harata, 1991, 1998). A whole volume (No. 3) of Comprehensive Supramolecular Chemistry, edited by Szejtli and Osa (1996), deals with ‘cyclodextrins’ in 22 chapters extending over 693 pages; there are 3 700 references. Szejtli (1998) mentions that over 15 000 publications had appeared by the end of 1997 (duplication has not been taken into account). An accompanying development has been the preparation of chemically modified cyclodextrins, generally by methylating or acetylating some or all of the hydroxyl groups (Bender and Komiyama, 1978; Tabushi, 1982); these derivatives also form inclusion complexes, which are discussed together with those of the parent compounds. One fundamental crystallographic aspect must be emphasized at the outset. Apart from a few exceptions among chemically modified cyclodextrins, all the crystals discussed below contain varying amounts of water molecules essential to their existence, and crystals of anhydrous cyclodextrins have not been found. In this sense they resemble most crystalline biomolecules such as proteins. This situation differs from that found in virtually all the other molecular complexes and compounds dealt with here, where anhydrous hosts (for complexes) or parent components (for compounds) are the species of reference. -, - and -cyclodextrins all form both clathrate and tunnel inclusion complexes. Saenger, Jacob et al. (1998) suggest that, for -CD complexes, small guest molecules form clathrates and larger guests tunnel complexes, i.e. there is a size selectivity. There are not yet analogous generalizations for the - and -CD complexes. Cell dimensions have been reported for more than 300 inclusion complexes of the cyclodextrins and their methylated derivatives and the crystal structures of most of these compounds have been determined. These structures are complicated because of the possibilities of conformational differences in the macrocycles due to interactions with the included guests and, especially because of the variety of hydrogen bonding arrangements possible, particularly at room temperature, and the accompanying disorder. Thus nonstoichiometric amounts of water are often found in the structures, with the water molecules distributed over a number of sites with partial occupancies. In structural terms this means that the arrangements in different unit cells are different, the crystal structure analysis giving only a picture averaged over the whole crystal. It would be desirable to determine crystal structures at the lowest possible temperatures (currently 10K) by neutron diffraction, preferably also on deuterated crystals and then study the changes that ensue on heating. Not much attention has yet been paid to possible phase changes. Such a programme has been started but results are available for only a few complexes; some disorder often remains even at 15K. We shall generally describe the room-temperature structures in broad terms and not attempt to enter into all the details of conformational differences and hydrogen bonding; Harata (1996; 1998) covers much the same material from a somewhat different point of view. Thus, in essence, we concentrate on showing how the complexes can be classified into various structural (crystallographic) groups, pointing out both the great similarities within each group and the subtle differences among its members. The variety, in a chemical sense, of guests in particular structural families is quite remarkable. The next stage in an overall structural analysis should be to compare in detail the host–guest–water arrangements and interactions. This vast task is on the verge of practicality (Le Bas, Rysanek et al. (1988)) but we leave its realization to future explorers.


77

Secondary hydroxy O(2)-H, O(3)-H rim Head

Internal tunnel

Tail Primary hydroxy O(6)-H rim (a)

(c)

o

a

o

c

b b (b)

(d) o

a

c o

b c (e)

c

o

b

a

Fig. 4.2. Schematic diagram showing (above): form of the cyclodextrin truncated cone molecule, with conventional definitions; there are twice as many secondary hydroxyls as primary hydroxyls (below) : the broad ways of describing cyclodextrin inclusion complexes: (a) head-to-head tunnel type, showing dimers found particularly in -CD complexes; (b) head-to-tail tunnel type; (c) cage herringbone type; (d) cage brickwork-like or slipped tunnel type; (e) slipped tunnel type with headto-head dimers. (Adapted from Harata, 1996.)

78


Table 4.1. Statistics of reported cyclodextrin inclusion complex structures (as of end 2002). The herringbone type of clathrate is denoted by A1 and the slipped tunnel or brickwork type by A2; the head-to-tail arrangement in the tunnel complexes is denoted by B1 and the head-to-head arrangement by B2 Type of cyclodextrin

Polymethyl- Polymethyl-

Cage structures

Tunnel structures

A1

A2

B1

B2

16 – 18 4 1

10 – 2 –

18 4 1 2

27 – 71 11 12(mixed)

Following Saenger (1984), the crystal structures can be classified, into two broad groups (Fig. 4.2), each of which can be further subdivided: Group A: Clathrate or cage structures, where the arrangement of cyclodextrin molecules is such that the cavities of the macrocyclic hosts are blocked off by neighboring host molecules and the guests are therefore enclosed in these cavities. The two subdivisions are: Group A1: herringbone arrangement of host molecules, Group A2: slipped-tunnel or brickwork-like arrangement of host molecules. This group can formally be considered to derive from the tunnel structures (see below) by mutual displacement of adjacent layers of host molecules. Group B: Tunnel structures, where the arrangement of cyclodextrin molecules is such that the toroidal cavities are lined up through the crystals to form approximately linear tunnels in which the guests are accommodated. The mutual arrangement of cyclodextrin molecules has been found to be either head-to-tail or head-to-head in the - and -cyclodextrin complexes studied, and a combination of both in the (mostly isomorphous) -cyclodextrin complexes studied. These two groups, in turn, can be further subdivided into smaller groups of isomorphous or isostructural complexes as has been illustrated by Caira (2002). Our emphasis in this chapter is on structures of CD complexes determined by single crystal methods but one should remember that many such complexes are only obtainable as microcrystalline powders. Here powder diffraction has an important role to play, especially for identification, and Caira (2002) has collected together typical powder patterns for the various isostructural groups. One should also note progress in the capability to solve crystal structures from powder data alone (Pop, Goubitz et al., 2002), and this will undoubtedly also be a feature of future developments. The November 2002 version of the CSD gives 358 hits for ‘cyclodextrin’, of which 116 refer to ‘alpha-cyclodextrin’, 214 to ‘beta-cyclodextrin’ and 17 to ‘gamma-cyclodextrin’. A statistical breakdown of the isomorphous and isostructural structures (i.e. those in Tables 4.2–4.19) among the various categories is shown in Table 4.1; about two-thirds of these structures are of the tunnel type. A measure of the rate of development of the field can be gleaned from the fact that Saenger listed 37 structures in his 1984 and 1985

-CYCLODEXTRIN AS HOST

79

reviews, while Dodds (1999) found 74 -CD structures in Version 5.12, (October, 1998) of the Cambridge Structural Database and 11 -CD structures. Our list of structures is representative but certainly not complete, nor can it be, with the continuing interest in cyclodextrin structures. 4.2 a-Cyclodextrin as host In -cyclodextrin (C36H60O30) the -D-glucose moieties are all in the pyranose staggered chair form with C1 conformation (1a 2e 3e 4e 5e); the glucose moieties are linked betwen 1- and 4-positions. The important structural features are the toroidal, truncatedcone shape of the macrocyclic molecule, its hydrophobic cavity and curved outer surfaces, and its hydrophilic upper and lower faces. The interior diameter of the cavity ˚ . The (primary) 6-hydroxyl face (defined conventionally as the ‘‘tail’’ of is about 5.2 A the molecule) is somewhat narrower than the 2,3-hydroxyl face (the ‘‘head’’ of the molecule); this is illustrated schematically in Fig. 4.2.1 There are intramolecular hydrogen bonds O(3)H . . . O(2) and O(3) . . . HO(2) between the secondary hydroxyl groups around the macrocyclic ring and these play an important role in its stabilization. X-ray data show that the mean O(2) . . . O(3) distances in -, - and -cyclodextrins are ˚ respectively; the interactions in -CD are weaker and its macrocycle 3.00, 2.86 and 2.81 A is more flexible than those of - and -CD. The C(6)-O(6) bonds are preferentially directed away from the centre of the ring (torsion angle O(5)-C(5)-C(6)-O(6) is ( )gauche; some of these bonds can turn inwards, the torsion angle becoming ( þ )gauche, with formation of hydrogen bonds between the O(6)H group and the guest molecule. These features are illustrated in the diagrams of the various structures given below. The solution thermodynamics of the formation of 1 : 1 inclusion complexes (predominantly in water at 298K) has been extensively studied. Rekharsky and Inoue (1998) have summarised and discussed values for the stability constants, standard free energies, enthalpy and entropy changes, and, for some examples, heat capacities. Some 600 values are given for -CD, 400 for -CD and 50 for -CD complexes. Many multiple values are given (for various temperatures, pHs and so on) so the number of guests involved for each cyclodextrin is about two-thirds of the number of separate values. There are also many values for chemically modified cyclodextrins. Corresponding measurements on crystalline complexes do not appear to have been reported, apart from an adiabatic calorimeter study of {-CD11H2O} over the range 13–300K (Hanabata, Matsuo and Suga, 1987), where a first-order phase change occurs at 226K. The solid-state heat capacities at 298.15K of the three common cyclodextrins and some defined hydrates have also been measured (Briggner and Wadso¨, 1990). 1 The secondary hydroxyls are at the broad end of the truncated cone, and the primary hydroxyls at the narrower end. However, the authorities do not agree on which is ‘‘head’’ and which ‘‘tail.’’ We follow Hamilton (Hamilton and Chen, 1988a), Tsoucaris (Mentzafos et al., 1991) and Harata (1996; see p. 290) in designating the broad end as ‘‘head’’ and the narrower end as ‘‘tail.’’ For example, the first of these references has ‘‘ . . . a head-tohead dimer is formed by means of extensive hydrogen bonding across the secondary hydroxyl ends of two adjacent -CD monomers . . . ’’. We use the same definition for - and -CD. Saenger (e.g. Steiner and Saenger, 1998a; see p. 454) and Kamitori et al (1998) interchange ‘‘head’’ and ‘‘tail.’’

80

4.2.1


a-Cyclodextrin as host in clathrate inclusion complexes

The crystalline inclusion complexes of -cyclodextrin (C36H60O30) fall into a number of isomorphous or isostructural groups and it is thus quite a versatile host. The chemical compositions within a group are reasonably similar, but there are differences among groups, especially in regard to degree of hydration, and this may indicate one of the sources of the structural differences. In our classification we largely follow Saenger (1985), discussing first the clathrates (Tables 4.2 and 4.3) and then the tunnel inclusion complexes (Tables 4.4–4.9). References and crystal data for earlier work given in extenso by Saenger are summarised in Tables 4.2 to 4.8, and in the text. We find that about two thirds of the structures reported in the CSD fall into the various isomorphous or isostructural groups detailed in the Tables. -Cyclodextrin forms clathrate inclusion complexes with smaller guests which are grouped together in Table 4.2; the gas molecules Cl2, O2, CO2, C2H4, CH4, propane and butane can be enclosed (Cramer and Henglein, 1957) in addition to the guests listed in Table 4.2. Herringbone arrangement of -CD moieties in clathrate inclusion complexes of -cyclodextrin;. Unit cell dimensions: a 9.5, b 14.3, c 37.5; space group P212121; Z ¼ 4; ˚ 3 ; compositions are given in the form {volume per asymmetric unit 1275 A CD[m(guest)]n(H2O)}, where ‘-CD’ is generally omitted for brevity Guest composition

Refcode; reference

a

b

c

V/FU

[7.6H2O] (form III) 0.48Kr58H2O 0.78Kr53H2O I24H2O n-propanol 48H2O CH3OH57H2O 0.8(CH3CN)57H2O CH3NO25H2O [butyric acid]43H2O; [1-butanol] 55H2O [3-iodopropionic acid]5H2O trans-2-butenoic acid5H2O pyrrole5H2O Acetic acid Propionic acid [60H2O] (form I) DMSO2CH3OH2H2O

BANXUJ; CS81 CYDXKR10; SN76 CDEXKR10; SN76 CDEXTI10; MSFM73b CDXPRO; SMFM74 CDEXME10; HS76 GEVTOQ; AJWH98 GULTUC; NILL00 MSFM73a MSFM73a BUPDEV; HUOH83a JECPEM; TSN90 QOHMEF; SRPG00 S85 S85 CHXAMH; MS74, KHS80* ACDMSM; H78 (P21; Z ¼ 2)*

9.400 9.470 9.446 9.558 9.393 9.465 9.479 9.452 9.45 9.44 9.685 9.43 9.404 9.427 9.46 9.529 9.505

14.356 14.299 14.377 14.240 14.292 14.339 14.323 14.299 14.29 14.382 13.508 14.406 14.293 14.34 14.29 14.858 14.150 102.88

37.536 37.489 37.402 36.014 37.515 37.365 37.397 37.380 38.11 37.99 39.581 38.174 37.265 37.62 38.11 34.038 19.738

1266 1261 1266 1225 1259 1268 1269 1263 1287 1289 1323 1296 1252 1272 1287 1205 1294

* These complexes are structurally related but not isomorphous. References: AJWH98 – Aree, Jacob, Saenger and Hoier, 1998; CS81 – Chacko and Saenger, 1981; H78 – Harata, 1978; HS76 – Hingerty and Saenger, 1976; HUOH83a – Harata, Uekama, Otagiri and Hirayama, 1983a; KSH80 – Klar, Hingerty and Saenger, 1980 (XRD and ND; CHXAMH02); MS74 – Manor and Saenger, 1974; MSFM73b – McMullan, Saenger, Fayos and Mootz, 1973b; NILL00 – Nakagawa et al., 2000; S85 – Saenger, 1985; SMFM74 – Saenger, McMullan, Fayos and Mootz, 1974; SN76 – Saenger and Noltemeyer, 1976; SPRG00 – Storsberg et al., 2000.


81

Table 4.2 but detailed crystallographic information is lacking. The crystal data show that these crystals are all isomorphous, despite small differences in water content. The small differences in cell dimensions depend on the nature of the guest and thus there is a fair degree of adaptability among the complexes of Table 4.2. The guest molecules are contained in the internal tunnels (tori) of the -CD molecules and the packing of the hosts is such that free exit of the guests from the tori is blocked by contiguous host molecules held together by an intricate network of hydrogen bonds between hydroxyl groups; the arrangement is shown rather well in the schematic diagram of Fig. 4.2(c). The water molecules lie between the host molecules and participate in the hydrogen bonding, small differences in water content leading to differences of detail in the hydrogen bonding schemes. In addition to the {-CD[I2]4H2O} clathrate complex, the detailed structure of which has been determined, there is a tunnel complex of composition {-CD0.5[I2]4H2O} (Table 4.6). The ternary complex {-CD[2(CH3)2SO.-(CH3OH)]}(CH3OH)2H2O (ACDMSM) also belongs to this group of slipped-tunnel structures; its monoclinic cell can be transformed to give a pseudo-orthorhombic unit cell with the monoclinic b axis coincident with the b axis of the orthorhombic cell of the other examples in Table 4.2. One methanol lies within the torus and the other outside, together with the water molecules. Perhaps surprisingly, the undecahydrate also belongs to this group of complexes. The complexes in Table 4.3 are clathrates, but derived from tunnels by mutual offset of adjacent layers, giving in Saenger’s phrase a ‘‘brickwork-like’’ layer arrangement (Fig. 4.3) or, in an alternative nomenclature, a ‘‘slipped tunnel’’ structure. In these complexes the rings (or methyl groups, for the dimethylformamide guest) are enclosed within the torus with the amino or hydroxyl group of the guest protruding from the 2,3-dihydroxyl face. The space between the -CD molecules is occupied by water molecules, some of which are disordered. There are four distinct -CD hydrates, two being hexahydrates (Form I (BANXUJ) is related to the group of isomorphous crystals of Table 4.2 while Form II (a ¼ 13.70, ˚ , space group P212121; Z ¼ 4; Lindner and Saenger, 1982a) is on its b ¼ 29.35, c ¼ 11.92 A Table 4.3. Clathrate structures, brickwork-like layers in the (001) plane; space group P212121; ˚ ; volume per formula unit 1280 A ˚3 Z ¼ 4. Unit cell dimensions: a 13.6 b 15.3, c 24.5 A Guest composition

Refcode; reference

a

b

c

V/FU

4-chlorophenol5H2O 4-bromophenol5H2O 4-iodophenol3H2O 4-nitrophenol3H2O 2-fluoro-4-nitrophenol3H2O p-iodoaniline3H2O p-hydroxybenzoic acid3.0H2O N,N-dimethylformamide5H2O 2-pyrrolidone 5H2O 11H2O

WEXKOZ; MMKO99 MESYEO; KTM01 CHAIPL; H76 ACDPNP; H77b ZEJDEX; ShSe94 CDEXIA; SBM76, H75 ACDHBA; H77b ACDMFM; H79 ACDPRO; H79 GOQZUH; PMP98

13.447 13.456 13.477 13.431 13.431 13.681 13.356 13.750 13.852 13.839

15.299 15.317 15.373 15.299 15.299 15.475 15.342 15.318 15.373 15.398

24.795 24.733 24.573 24.788 24.780 24.569 24.896 24.544 24.353 24.209

1275 1274 1273 1273 1273 1300 1275 1293 1297 1290

References: H75 – Harata, 1975; H76a – Harata, 1976a; H77b – Harata, 1977b; H79 – Harata, 1979; H82a – Harata, 1982a; HNI96 – Harata, Nagano, Ikeda et al., 1996; KTM01 – Kamitori, Toyama and Matsuzaka, 2001; MMKO99 – Muraoka et al., 1999; PMP98 – Puliti et al., 1998; SBM76 – Saenger, Beyer and Manor, 1976; ShSe94 – Shibakami and Sekiya, 1994.

82


(a)

a

1/4 (b)

1/4

1/4

N

b

N 1/4 N

N

N

1/4 c

N a N

Fig. 4.3. A schematic drawing showing the packing in {-CD[p-iodoaniline]}3H2O, a representative of the slipped tunnel structures of Table 4.3. Two cages, one intra-torus and occupied by guest molecules, and the other intermolecular and occupied by waters of hydration, are indicated by light and dense hatching respectively. In (a) the outer contour circles represent the O(2), O(3) rim (‘‘head’’) of the -CD molecules while the middle circles represent the O(6) rim (‘‘tail’’). Circles and molecules drawn with heavier lines are closer to the observer than those with lighter lines. This is also the structure of {-CD11H2O}. (Reproduced from Saenger, Beyer and Manor, 1976.)

own); Form III, approximately an octahydrate, is a ‘‘full’’ member of the group of isomorphous crystals of Table 4.2, and Form IV (undecahydrate) is classified in Table 4.3. The -CD moieties in both hexahydrates have strained, high-energy conformations, different from those in the other two complexes; two of the primary O(6) hydroxyl groups are in the gauche, trans conformation directed towards the center of the doughnut shaped molecule (Fig. 4.4) and two of the O(2) . . . O(3) hydrogen bonds are broken, leading to a ‘‘dented’’ shape for the macrocycle. In contrast, in both Forms III and IV (and also in the isomorphous I2 complex) all the O(6) hydroxyl groups are in the gauche, gauche conformation directed away from the centre of the macrocyclic ring which has a ‘‘round,’’ unstrained shape. In the octahydrate there are 2.6 water molecules within the cavity disordered over 4 sites while in the undecahydrate five water molecules are disordered. In Form I two ordered water molecules are included within the torus, and one water molecule in Form II. This comparison can be extended beyond the hydrates to a number of other -CD complexes (Saenger, Noltemeyer, Manor, Hingerty and Klar, 1976). Thus Saenger


83

Form III

Form I (b)

(a)

0(4)4

0(4)4

0(4)5

0(4)5

0(4)6

0(4)6

0(4)3

)3

0(4

0(4)2

0(4)1 0(4)2

(d)

(c)

(f)

(e)

0(4)1

Fig. 4.4. PLUTO (Motherwell, 1978) space-filling computer-drawn diagrams showing the conformation of the -CD ring in the Form II hexahydrate on the left and the (so-called ‘‘Form III’’) octahydrate on the right. In (a) and (b) the molecules are viewed, from the head side (O(2) . . . O(3)), normal to the hexagon of O(4) atoms of the various glucose residues; in (c) and (d) from the side and in (e) and (f) from the tail of the molecule. In the octahydrate the molecule has a relaxed, round shape but there are distortions in Forms I and II of the hexahydrate especially in the vicinity of glucoses 1 and 6, where O(2) . . . O(3) hydrogen bonds are broken in Form I. In the hexahydrate Form I two O(6) hydroxyls point into the cavity, but not in the octahydrate. (Reproduced from Chacko and Saenger, 1981.)

(1984) contends that -CD also has the strained conformation in aqueous solution and that adduct formation takes place by displacement of the two intra-torus waters by the guest molecule, accompanied by change of the -CD molecule from a ‘‘tense’’ to a more ‘‘relaxed’’ conformation. This mechanism is similar to the ‘‘induced-fit’’ mechanism

84


proposed for the interaction of enzymic proteins with their cofactors and substrates (Koshland, 1970), thus indicating that cyclodextrins and their complexes should provide appropriate models for study of enzyme behaviour. The structures of four 1 : 1 crystalline complexes of -CD with metal coordination complexes (and approximately six associated molecules of water) have been reported. Classification presents problems as only one packing arrangement (as opposed to the host–guest relationships) has been described, and the cell dimensions do not show definite similarities to any of the other groups listed in Table 4.2. There is an isomorphous pair of structures [{-CD[Rh(cod)(NH3)2]-PF66H2O}, a ¼ 14.033(4), b ¼ 19.517(6), ˚ (Alston et al., 1985a; DAXTEB) and {-CD [Co(Hdmg)2 c ¼ 23.803(6) A (n-C4H9)H2O)]7H2O}, where H2dmg ¼ dimethylglyoxime, a ¼ 14.083(3), b ¼ 19.237(4), ˚ (Chen et al., 2000; LOVVAT)] and two individual structures {-CD. c ¼ 24.517(5) A ˚ (Luo [Co(Hdmg)2 (n-C3H7)H2O)]7H2O}, a ¼ 13.440(3), b ¼ 17.593(4), c ¼ 29.009(6) A et al., 1996; RAXPOV) and {-CD[C6H12N2O4Pt)]5.5H2O}, a ¼ 10.102(4), b ¼ 13.526(4), ˚ (Alston et al., 1985b; DEGTEO). All these structures crystallize in space c ¼ 41.971(9) A group P212121 with Z ¼ 4, and this suggests that they may all be clathrates. This has been confirmed for the complex with [Co(Hdmg)2(n-C3H7)H2O)] as guest which is the only complex for which the crystal packing has been described. 4.2.2

a-Cyclodextrin as host in tunnel inclusion complexes

The tunnel inclusion complexes of -CD fall into a number of isomorphous or isostructural classes. An immediate distinction can be made between head-to-tail Table 4.4. Tunnel inclusion complexes of -CD with head-to-tail dimers. The space groups are all P21212; Z ¼ 2, and unit cells have been reoriented so as to make [001] the stack axis. Compositions as in Table 4.2, i.e. in the form {-CD[m(guest)]n(H2O)}, where ‘-CD’ is generally omitted for brevity. The square brackets indicate that the guest is included in the torus Guest composition

Refcode; reference

a

b

c

V/FU

[m-Nitrophenol]m-Nitrophenol6H2O* [Benzyl alcohol]Benzyl alcohol6H2O* Na-1-propanesulfonate97H2O [Methyl orange]Naþ9.8H2O [Methyl orange]Kþ9.8H2O [-aminobutyrate]Kþ10.0H2O] [1.5(acetate)]15Kþ9.8H2O [Benzenesulphonate] Naþ10.0H2O 2,5-dihydroxybenzoic acid3.5H2O [Hexanoate] Naþ11.0H2O (Z ¼ 4; see text)

ACDMNP; HUT78 WILJAC; SG94 ACDPRS; H77a CDXSOM; H76a CDXKOM; H76a CDKABA; TAM81 HRW65 CDXBZS; H76c WIZQEB; MM00 MSFM73a

22.231 22.189 21.608 22.099 22.120 21.861 21.89 21.832 21.939 21.94

16.865 16.602 16.700 16.359 16.419 16.624 16.54 16.529 16.786 16.53

8.152 8.265 8.302 8.296 8.292 8.279 8.30 8.356 8.273 16.56

1528 1527 1498 1500 1506 1504 1506 1508 1523 1503

* One guest within each -CD cavity, and one between the stacks. References: H76a – Harata, 1976a; H76c – Harata, 1976c; H77a – Harata, 1977a; HRW65 – Hybl, Rundle and Williams, 1965; HUT78 – Harata, Uedaira and Tanaka, 1978; MM00 – Mele and Malpezzi, 2000; MSFM73a – McMullan, Saenger, Fayos and Mootz, 1973a; SG94 – Steiner and Gessler, 1994; TAM81 – Tokuoka, Abe, Matsumoto, Shirakawa, Fujiwara and Tomita, 1981.


85

and head-to-head arrangements of the -CD molecules in a stack. The crystals of Table 4.4 are built up of -CD molecules stacked head-to-tail along [001], with the guest anions within the tori and the water molecules and cations outside (Fig. 4.5 (left)). The heads and tails of adjacent -CD molecules are hydrogen bonded together in these tunnel complexes, water molecules being located on the peripheries but not between the -CD’s (Fig. 4.6). The Class IIB structures (Fig. 4.5(right)) are lower symmetry variants of the Class IIA type, with differences of detail rather than of principle. The -CD molecules have exact or approximate C2-2 symmetry and closely similar conformations in all the complexes of Table 4.4, although there is some elliptical distortion in the benzenesulphonate and Methyl Orange complexes because of the flattened cross-sections of these guest molecules. The hydrophobic anions are located within the hydrophobic portions of the cavities but are hydrogen bonded to the host molecule at its 6-hydroxyl end; there is disorder about the two-fold axes. There are some differences among the individual complexes; for example the long Methyl Orange anions extend through two -CD molecules (in a partially disordered arrangement, as can be inferred from diffuse scattering on appropriate diffraction photographs), while the acetate groups of the potassium acetate complex are so small that water molecules are accommodated within as well as outside the cavities (Fig. 4.5). The hexanoate anion also extends through two -CD molecules but here the arrangement must be ordered as ˚ . The m-nitrophenol complex shows one the periodicity along [001] is doubled to 16.5 A striking difference of arrangement from the rest of this group, already hinted at by the compositional difference; one guest molecule is found within the torus while the other is located between the host molecules. There are a few analogous examples among other complexes; one possible consequence is different physico-chemical behavior for the two guest molecules. (a)

(b)

(c)

(d)

O6 H C6 H C5 O4 H C3 O3 H

Fig. 4.5. A comparison of the arrangements of guest molecules in the tunnels of some complexes of Table 4.4. The van der Waals envelopes of the guest molecules are shown; broken lines denote hydrogen bonds. The guests are (a) Methyl Orange (H3C)2NC6H4-N¼N–C6H4SO3; (b) Sodium benzenesulphonate NaC6H5SO3; (c) Sodium propanesulphonate NaC3H7SO3; (d) Potassium acetate KCH3COO. Compare with the schematic arrangement shown on the left side of Fig. 4.6. (Reproduced from Harata, 1976a,b.)

86


Table 4.5. Tunnel inclusion complexes of -CD with head-to-tail dimers. Space groups are all P21; Z ¼ 2, with stacks along [100] Guest composition

Refcode/reference

a

[m-Nitroaniline]6H2O

CDNOAN; H80

8.054

[Benzaldehyde]6H2O

BAJJAX; HUO81

7.932

[4-fluorophenol]6H2O

JUMYOF; ShSe92

7.845

[2-fluorophenol]5H2O

JUMYUL; ShSe92

7.842

Hydroquinone. 6H2O

PUPTEZ; StSa94

7.909

[p-cresol]6H2O

WEXLEQ; MMKO99

7.927

3-methyl-1-butanol7.1 H2O

S85

(R,S)-1-phenylethanol4H2O

BIJHOR; H82a

8.176

Mono(3-amino-3-deoxy-) -CD 5.5H2O

TOVLOF; HNI96

7.955

16.64

b/

c

V/FU

13.508 94.58 13.500 90.85 13.587 91.75 13.615 92.27 13.505 90.76 13.568 90.41 13.95 95.0 23.930 106.69 24.989 98.15

24.668

1338

24.704

1323

24.557

1308

24.550

1310

24.706

1319

24.54

1320

23.64

1368

13.853

1298

13.106

1290

Notes: BIJHOR has a head-to-tail arrangement in which -CD moieties are inclined at 17 to stack axis; guests sandwiched between -CD molecules. CSD gives p-benzoquinone diagram and formula for PUPTEZ; this is wrong. TOVLOF is a chemically-modified -CD. References: H80 – Harata, 1980; H82a – 1982a; HNI96 – Harata, Nagano, Ikela et al., 1996; HUO81 – Harata, Uekama, Otagiri, Hirayama and Ogino, 1981; MMKO99 – Muraoka et al., 1999; S85 – Saenger 1985; ShSe92 – Shibakami and Sekiya, 1992; StSa94 – Steiner and Saenger, 1994.

For most tables in this chapter, we have listed the variety of guests found in a particular host framework; a converse approach in used in Table 4.6, where we list the variety of frameworks found for hosts of the same chemical type, here ionic polyiodides of various compositions. The variety is quite surprising. Detailed crystal structures have been reported for ZZZANG10 (Cd pentaiodide) and for CYDXLI (Li pentaiodide), which has been inserted in Table 4.7 rather than in Table 4.6 in order to emphasize structural (unit cell) resemblances. A square arrangement of head-to-head dimers stacked in linear fashion along [001] is found in the tetragonal crystals of composition {2(-CD) 0.5(Cdþþ) I3 I2 27 H2O} ˚ 3; ˚ ; space group P41212; Z ¼ 8; volume per formula unit ¼ 1533 A (a ¼ 19.93, c ¼ 30.88 A Noltemeyer and Saenger, 1980; ZZZANG10). The guest anion is [I5 ] and other examples are found with the cations Kþ, NH4þ, Hþ and some divalent species; this is the stable form for Kþ and NH4þ cations. Isostructural crystals are formed with the neutral guest 4,4 0 -biphenyldicarboxylic acid [{2(-CD) C14H10O4 14H2O} (a ¼ 19.609, ˚ ; space group P41212; Z ¼ 8; volume per formula unit ¼ 1550 A ˚ 3; Kamitori c ¼ 32.257 A 0 et al., 1998; CAQPAL). The 4,4 -biphenyldicarboxylic acid guest is entirely enclosed


b

87

0

b a

c

1/4

1/4

1/4

1/4

1/4

1/4

a

c a

b

0

Fig. 4.6. (Left) General arrangement of -CD molecules in the tunnel inclusion complexes of Table 4.4, illustrated for the [1-propane sulphonate]. Naþ9.7H2O complex. Stack axis [001]. (Right) General arrangement of -CD molecules in the tunnel inclusion complexes of Table 4.5. illustrated for the [benzaldehyde] 6H2O complex. Stack axis [100]. (Adapted from Harata (1977a) and Harata, Uekama, Otagiri, Hirayama and Ogino, 1981.)

Fig. 4.7. A stereoview showing the head-to-tail stacking of –CD molecules in {-CD [benzaldehyde]6.0H2O}, illustrating the arrangement in the tunnel inclusion complexes of Table 4.5. Here water molecules also participate in the intracolumn stacking but in other examples the ˚ interactions are only between -CD molecules. Intermolecular O . . . O contacts of less than 3.1 A are shown by thin lines. The benzaldehyde guests can be seen inside the columns. (Reproduced from Harata, Uekama, Otagiri, Hirayama and Ogino, 1981.)

88


Table 4.6. Various crystals with polyiodide anions showing the variety of crystal types obtained. The approximately linear anions are held within the -CD torus, with counterions and water molecules between the stacks of -CD head-to-head dimers. The formulae are based on {2(-CD) – note the change from the standard formulation used elsewhere Overall composition

Refcode/ reference

2(-CD NaI3.I2.)8H2O

CYDXTF; NS80 CYDXTE

2(-CD)[I2]8H2O 2(-CD)CsI3I212H2O Alternatively, hexagonal

CYDXTC; NS80

Other cations

a/

b/

c/

Space group

19.590

NH4þ, Kþ, Rbþ

24.440 15.750 space group 109.30 not given 12.970 12.970 7.910 120 15.8 27.37 40.2 C222 15.8

2(-CD)Ba2(I)I212H2O ZZZAND10; 13.71 NS80 2(-CD)Cd0.5 I3I227H2O ZZZANG10; H3Oþ, NH4þ, 19.90 NS80; Kþ, Rbþ, Mg2 þ , Ca2þ, Sr2 þ , Ba2þ, Zn2 þ , Cd2 þ 2(-CD) CAQPAL; 19.609 KMKO98 [4,4 0 -biphenyldicarboxylic acid]14H2O

15.8 13.71 19.90

40.2 120 17.04 120 30.80 90

P62, P6222, P31 P622 P42212

19.609 32.257 P42212

References: KMKO98 – Kamitori et al., 1998; MSFM73a – McMullan, Saenger, Fayos and Mootz, 1973a; NS80 – Noltemeyer and Saenger, 1980.

within the head-to-head dimer, with carboxyls of succeeding guest molecules linked through a pair of water molecules. When -CD was crystallized from a solution containing 4,4 0 -biphenyldicarboxylic acid and CaCl2, then orthorhombic crystals [{2(-CD) ˚ ; space group P21212; Z ¼ 8; C14H10O414H2O} (a ¼ 35.397, b ¼ 24.577, c ¼ 27.969 A ˚ 3; Kamitori et al., 1998; CAQPEP) were obtained. volume per formula unit ¼ 1550 A Structure analysis showed that there were short columns composed of two head-to-head dimers instead of the continuous one-dimensional stacks; however, the guest molecules were enclosed in much the same pattern in both forms. Are these polymorphs? The word was not mentioned by Kamitori et al. The isomorphous structures of the first group in Table 4.7 (the organometallic guests) have been succinctly described by Klingert and Rihs (1991a) as follows: ‘‘ . . . the -CD molecules are arranged head-to-head to form a dimer, with all secondary hydroxy groups of one macrocycle linked to the adjacent one by direct hydrogen bonds. The dimers are stacked along the crystallographic c axis forming parallel channels in which the guest cations and anions are lined up alternately. The cations are encapsulated within the cavity of the dimers while the PF6 anions . . . are centred between the primary hydroxyl faces of adjacent dimers.’’ This description applies well to the complex with neutral Fe(cp)2 as guest (where the PF6 anion is absent), and also to the crystals with neutral molecular guests. The lithium


89

Table 4.7. Isomorphous triclinic crystals (the unit cells are reduced) with guests of various kinds. The space group is P1, Z ¼ 2 unless stated otherwise. In the first group of complexes there are organometallic cations held within the torus of the -CD head-to-head dimers, the PF6 counterions between the tail ends of adjacent dimers and water molecules between the stacks. The general formula is {2(-CD) [Mþ][PF 6 ]8H2O} (note the change from the standard formulation used elsewhere); (-CD) and PF6 are omitted for brevity, while Mþ represents the organometallic cation. The next group has two molecules and one salt as guest. The 1-octanol and valeric acid complexes have head-to-head dimers and hexagonal packing of stacks. The last two examples are presumably isostructural to the first group Organometallic guest cation

Refcode; reference

a/

b/

c/

Unit cell volume

Rh(cp)2þ8H2O

KIWZEV; KR91a,b KIWZAR; KR91a,b KIWYUK; KR91a,b JEMGUD; OH90 PEPBUH; MS-ES-F93 JEHYIE; KR90 KOGKEW; KR91b XIGBOE; S-RPLB01 ZASYOH; NVCdR94 CYDXLI10; NS80 S85

13.756 91.33 13.810 91.06 13.806 91.43 13.836 92.22 13.768 90.98 13.815 91.41 13.845 91.94 13.83 93.01 13.852 93.01 13.830 91.91 13.86

13.833 93.08 13.833 92.94 13.839 92.77 13.864 92.13 13.911 93.46 13.891 92.85 13.861 91.98 13.88 91.72 13.878 91.98 13.855 92.20 13.86

2565

S85

13.85

13.85

MSFM73a; S85 KR91a

11.911 91.3 13.892 89.63

13.870 93.75 13.926 89.20

15.561 119.67 15.560 119.76 15.520 119.80 15.694 119.76 15.601 119.47 15.639 119.64 15.641 119.85 15.72 119.58 15.719 119.32 15.690 119.82 15.63 120 15.62 120 15.669 115.01 16.585 60.26

Co(cp)2þ8H2O Fe(cp)2þ8H2O Fe(cp)29H2O CpRu(C6H6)þ8H2O Fe(cp)(C6H6)þ8H2O Fe(cp)(C7H8)þ8H2O {2(-CD) [n-butylisothiocyanate]9H2O} {2(-CD) [acetone]9H2O} 2(-CD)LiI3 I28H2O 1-octanol (Laue symmetry 6/mm) Valeric acid (Laue symmetry 6/mm) [Diethyl ether]4.1H2O Fe(cp)(C8H8)þnH2O

2574 2566 2606 2593 2601 2597 2615 2625 2602 1300 1298 2606 2786

Notes: ˚ , 90.02 1. KOGKEW gives a metrically-C-centered monoclinic cell with dimensions 13.884 23.977 15.641 A 93.91 90.08 (Z ¼ 4). Coordinates are available for checking. ˚ , 89.96 2. Fe(cp)(C8H8) gives a metrically-C-centered monoclinic cell with dimensions 13.892 24.183 16.585 A 90.80 90.18 (Z ¼ 4). Coordinates are not available for checking. ˚ , 89.95 94.30 3. JEMGUD gives a metrically C-centered monoclinic cell with dimensions 13.900 23.960 15.694 A 90.13 (Z ¼ 4). Coordinates are available for checking (Odakagi et al., 1990). 4. Na and Tl give crystals isomorphous with 2(-CD)LiI3I28H2O. References: KR90 – Klingert and Rihs, 1990; KR91a – Klingert and Rihs, 1991a; KR91b – Klingert and Rihs, 1991b; MSFM73a – McMullan, Saenger, Fayos and Mootz, 1973a; MS-FS-E93 – Meister, Stoeckli-Evans and Su¨ssFink, 1993; NS80 – Noltemeyer and Saenger, 1980; NVCdR94 – Nicolis, Villain et al., 1995; OH90 – Odagaki et al., 1990; S85 – Saenger 1985; S-RPLB01 – Sicard-Roselli et al., 2001.

90


Table 4.8. Trigonal crystals with organometallic cations held within the -CD head-to-head dimers and PF6.counterions held between the dimers; the water molecules lie between the stacks of dimers. The general formula is {2(-CD) [Mþ] PF6nH2O, with n 13} (note the change from the standard formulation used elsewhere). In this table only the organometallic guest cation is listed. The original data come from Klingert and Rihs (1991b; T ¼ 180K) except where noted otherwise Organometallic guest cation

Refcode

a

c

Unit cell volume

Fe(cp)(indan)þ Fe(cp)(cumene)þ Fe-bis(cp)propaneþ Fe(cp)(o-xylene)þ Fe(cp)(anisole)þ Ru(cp)(acetophenone)þ (water content not given)

KOGKIA KOGKOG KOGKUM KOGLAT KOGLEX PEPCAO; MS-ES-F93

13.836 13.846 13.812 13.867 13.907 13.918

50.464 50.272 50.356 50.468 49.700 49.77

8366 8347 8319 8404 8324 8349

pentaiodide complex has the anti-isomorphous structure,2 with the charged species interchanged. The complexes of Table 4.7 provide an excellent illustration of how a particular framework can accommodate a variety of chemically different guests, and this generalization applies not only to cyclodextrin complexes. We have some understanding of the principles governing the construction of the framework but little knowledge of what determines which guests will be accepted. The examples in Table 4.8 (another group of isomorphous crystals) show that small changes in the nature of the organometallic cations can lead to a different type of crystal structure, albeit still based on dimers enclosing cations and anions between stacks. About two-thirds of the -CD structures listed in the CSD can be accommodated in isomorphous or isostructural groups (Tables 4.2 to 4.8, excluding Table 4.7). The others are found in space groups P212121,3 P214 and P21212,5 with a few exceptions. Judging from the cell dimensions, these crystals are all different but we have not made a detailed structural comparison. A ‘‘one-of-a-kind’’ structure that warrants special mention is {-CD(cyclopentanone)4.6 H2O}. This was first studied by XRD at room temperature by Le Bas (see Tsoucaris et al., 1987; FERCIO) and then by neutron dif˚ , space group P6, Z ¼ 6; Le Bas and Mason, 1994; fraction at 20K (a ¼ 23.725, c ¼ 7.935 A KIRJOK10). Diffuse scattering at 300K has also been studied (Le Bas and Doucet, 1997). Although there is static disorder even at 20K, the high symmetry and low temperature reveal features not yet found in other -CD structures. 4.2.3

Chemically modified -cyclodextrins as hosts in inclusion complexes

It has been found that chemical modification of cyclodextrins, primarily by methylation or acetylation of hydroxyl groups, leads to hosts with properties different from those of the 2 ‘‘Two substances are said to be anti-isomorphous when their crystal structure[s] are geometrically identical but with the positions of corresponding atoms or ions interchanged.’’ (Evans, 1966, p. 194). 3 HOGCIP, KOBLOC, MEWFUP, QOYLEV, RAXPOV, ROQVUO, TEVCEC, VEHQAA, ZIBWIQ, ZIBWOW. 4 BOLVUT, CDXNEH, OBUMED, RIQDIE, RIQDOK. 5 HEHQAM, MESYAK.


91

native cyclodextrins. The compounds are shown in the schematic formula, with their trivial names (R00 ¼ CH3 for all formulae). R⬙O6C6H3C O3R⬘ 5 O4

4

1

O5 3

2 O2R

n 6 6 7 7

R R0 CH3 H CH3 CH3 CH3 H CH3 CH3 Scheme 4.1

n

name per-dimethyl--CD per-trimethyl--CD per-dimethyl--CD per-trimethyl--CD

The crystal structure of 6A,6C,6E-tri-O-methyl--CD6.7H2O (i.e. three alternate CH2OH groups methylated) has been reported (Durier, Buisson, Dueé, Driguez and Taravel, 1992; PEZKAG), but not (as yet) of any of its inclusion complexes. Crystal data ˚ , V(asymmetric unit) ¼ 1321.0(2) A ˚ 3, are a ¼ 13.975(1), b ¼ 29.162(2), c ¼ 12.965(1) A space group P212121, Z ¼ 4. The molecules are packed in herringbone fashion and there is some resemblance to -CD7.6H2O (form III). Two water molecules are ordered inside the torus, and one ordered, and 3.7 disordered, outside it. The chemically modified -cyclodextrins whose inclusion complexes have been studied crystallographically are hexakis(2,6-di-O-methyl)--CD (C48H84O30, called per-dimethyl-CD) and hexakis(2,3,6-tri-O-methyl)--CD (C54H96O30, per-trimethyl--CD). In per-dimethyl--CD the O(3) hydroxyl groups form O(3)-H . . . O(2) intramolecular ˚ . The crystal structures of the hydrogen bonds which vary in length from 2.87 to 3.16 A per-dimethyl--CD 1 : 1 complexes with I2, acetone (ROQVOI) and 1–propanol (KAF˚, GAZ10) are isomorphous (for the I2 complex: a ¼ 14.124, b ¼ 10.667, c ¼ 21.443 A ˚ 3; KAFFUS10) ¼ 106.3 , space group P21, Z ¼ 2, volume of the asymmetric unit 1552 A and have been shown to be clathrates, with the guests enclosed within the tori of the host molecules (Harata, 1990b). The 1 : 1 complex of per-dimethyl--CD with 3-iodopropionic ˚ , volume of acid is orthorhombic (VERVET; a ¼ 10,707, b ¼ 14.504, c ¼ 41.448 A ˚ 3, space group P212121) and also has a clathrate structure (Harata, asymmetric unit 1616 A 1989). It is striking that these three complexes are anhydrous although crystallized from water, in contrast to the hydrated complexes generally formed. Per-dimethyl-CDacetonitrile dihydrate (Aree, Hoier, Schulz, Reck and Saenger, 2000a; WEXKIT) is also orthorhombic P212121 but not isomorphous with VERVET. Among the interesting properties of the permethylated cyclodextrins is their solubility behavior, summarized as follows by Aree, Uson et al. (1999): ‘‘ . . . the solubility coefficients in water become negative, i.e. they are better soluble in cold than in hot water, where they precipitate or crystallize (Uekama and Irie, 1987) . . . methylated CDs have been crystallized from hot and cold water at 60–89 C and at 4–18 C . . . The crystals that could be obtained from hot water are anhydrous hexakis(2,6-di-O-methyl)-–CD


92

(Harata, 1995; Steiner, Hirayama and Saenger, 1996), anhydrous hexakis(2, 3, 6-tri-Omethyl)-–CD and anhydrous heptakis(2, 6-di-O-methyl)-–CD (Steiner and Saenger, 1995a), heptakis(2, 3, 6-tri-O-methyl)-–CDH2O (Caira, Griffith et al., 1994b) and octakis(2, 3, 6-tri-O-methyl)-–CD 2 H2O (Steiner and Saenger, 1998b),’’ Many of these structures (and others) are noted in the following pages. In per-trimethyl--CD all hydroxyl groups have been methylated and the methoxy oxygens can act only as hydrogen bond acceptors. The methoxy groups extend above and below the mean plane of the macrocyclic ring, thus increasing the height of the ˚ (instead of 8 A ˚ for -CD) and decreasing the cross-section truncated cone to 11 A at the 6-methoxy end (Fig. 4.8); this leads to a vase shape (cf. Section 3.4 on cavitands and caviplexes). There is no intramolecular hydrogen bonding and thus the macrocyclic rings are more flexible than those of -CD, with consequent enhancement of the possibilities for chiral discrimination, as has been demonstrated for the diastereoisomeric complexes of R- and S-mandelic acid [PhCH(OH)COOH] (Harata, Uekama, Otagiri and Hirayama, 1987). The four complexes whose structures have been determined fall into three groups (Table 4.9). The limited number of water molecules present is a striking feature of the compositions compared to those of the -CD complexes. The packing of pertrimethyl--CD molecules in the tunnel complexes is compared in Fig. 4.9. The principal

z y x

Fig. 4.8. Stereoview of hexakis(2,3,6-tri-O-methyl)--CD as found in its per-trimethyl--CD. [ p-iodoaniline]H2O complex (Table 4.9). The bracketed moieties are shown included in the torus. The 2,3-methoxy groups are at the bottom of the diagram (‘head’ of the molecule) and the 6-methoxy groups at the top (‘tail’). (Data from Harata, Uekama, Otagiri and Hirayama, 1984.)

sinb

I

II

III

Fig. 4.9. Schematic diagrams of the stacking arrangements in the tunnel inclusion complexes of pertrimethyl--CD: (I) p-iodoaniline; (II) benzaldehyde; (III) p-nitrophenol. (Reproduced from Harata, Uekama, Otagiri, Hirayama and Sugiyama, 1982.)


93

Table 4.9. Structural classification of the inclusion complexes of hexakis(2,3,6-tri-O-methyl)-˚ , deg. V/FU ¼ volume CD (per-trimethyl--CD); compositions as in Table 4.2; cell dimensions in A ˚ 3) per formula unit. These are all tunnel structures, with head-to-tail packing along [001] and (A reported in space group P21; Z ¼ 2. However, note that CECMAY has been reinterpreted as C2221 (Marsh et al., 2002), and that the space group of the last complex is P212121; Z ¼ 4 Guest and water content

Refcode/reference

a

b

c

V/FU

Class IA:[C6H6NI] H2O [p-iodoaniline] [R-mandelic acid]2H2O

BEYLOG; HUO82a CECMEC10; HUO87 BOHWUQ; HUOHS82 BUPDIZ, JEJWOK; H90a JEJXAX; H90a BUPDIZ; HUOH83a BUDKEQ; HUOH82b CECMAY10; HUO82a CECMAY11; MHKH02 MYM01

11.440

23.674

13.531

91.90

1831

11.624

23.739

13.786

106.56

1823

11.604

23.832

13.593

106.11

1799

11.604

23.669

13.824

106.72

1818

11.586

23.641

13.762

106.45

1808

11.59

23.285

13.901

106.98

1794

11.307

14.578

22.118

96.36

1812

13.123

23.187

13.113

107.19

1906

15.571

21.116

23.187

14.636

21.637

23.45

[benzaldehyde]

[(R)-phenylethanol] [(S)-phenylethanol]H2O Iodoacetic acidH2O Class IB: [p-nitrophenol]H2O Class II: [S-mandelic acid]3H2O Reinterpretation with space group C2221 (R)-1,7-dioxaspiro [5.5]-undecane5.1 H2O

1857

References: H90a – Harata 1990a; HUO82a – Harata, Uekama, Otagiri and Hirayama, 1982a; HUOH82b – Harata, Uekama, Otagiri and Hirayama, 1982b, 1984; HUOH83a – Harata, Uekama, Otagiri and Hirayama, 1983a; HUO87 – Harata, Uekama, Otagiri and Hirayama, 1987; HUOHS82 – Harata, Uekama, Otagiri, Hirayama and Sugiyama, 1982; MHKH02 – Marsh et al., 2002; MYM01 – Makedonopoulou, Yannakopoulou et al., 2001.

interactions between host molecules are of the van der Waals type, in contrast to the other cyclodextrin complexes where the hosts are blanketed by large numbers of water molecules. Here, if present, the few water molecules are hydrogen bonded to guest and host within the host tunnel and do not link host molecules. Despite the overall similarities in molecular arrangement, there are subtle differences of detail which are first revealed by small differences in cell dimensions. There is a considerable measure of adaptability in the per-trimethyl--CD tunnel inclusion complexes. A remarkable enantiospecific separation of the enantiomers of the racemic olive fly pheromone 1,7-dioxaspiro[5.5]undecane [] by complexation of the (R)-enantiomer with hexakis(pertrimethyl)--CD and of the [S]-enantiomer with heptakis(pertrimethyl)-CD has been reported by Mentzafos, Mavridis and Yannakopoulou (1999) and Makedonopoulou, Yannakopoulou et al. (2001). The binding constants of the two complexes determined by NMR in aqueous solution are respectively 6600 M1 for the


94

[R]-enantiomer – hexakis(per-trimethyl)--CD complex and 935 M1 for the [S]enantiomer – heptakis(per-trimethyl)--CD complex. These differences can be explained by the details of the interactions between enantiomer and cyclodextrin in the respective pairs. In the complex of the (R)-enantiomer with hexakis(pertrimethyl)--CD the guest interacts not only with the enclosing host but also with other hosts in the crystal lattice; C–H . . . O interactions appear to be particularly important. On the other hand, the diastereomers of hexakis(pertrimethyl)--CD with R-and S-phenylethanol are isomorphous (JEJWOK and JEJXAX in Table 4.9) and enantiomeric separation does not appear feasible. O

O O

O

[S] enantiomer

[R]-enantiomer

Scheme 4.2

The mode of insertion of the same guest molecules in the tunnel inclusion complexes of -CD and per-trimethyl--CD is compared in Fig. 4.10. There are similarities for p-iodoaniline but not for the other two guests; however, it would be premature to generalize on the basis of the meagre available evidence. The conformations of per-trimethyl--CD and per-trimethyl--CD have been discussed in detail together with the host–guest inclusion geometry (Harata, Uekama, Otagiri and Hirayama, 1982b, 1984). W I H H H

C

C

C C

C C O

O

H H

H

C H

H

Benzaldehyde

C

C C

C C

C

H C C C H

H

C

H

H

C H

H

C

C N

H

H

H

p-Iodoaniline

H

I H

C

C

C C

C C H

N

H H

H

W O H H

C

N C

O C C

C C

H H

H C

H

O C C

H C C

O H

H

C H

O N O

A p-Nitrophenol

B

Fig. 4.10. Schematic diagrams of the modes of inclusion of the same guest molecules in -CD tunnel inclusion complexes (on the left) and per-trimethyl--CD tunnel inclusion complexes (on the right). The O(2,3) hydroxyls are at the lower ends of the truncated cones. Note the water molecules included in two of the right hand diagrams. (Reproduced from Harata, Uekama, Otagiri and Hirayama, 1982b.)


95

4.3 b-Cyclodextrin as host The -cyclodextrin molecule (C42H70O35) consists of seven 1,4--linked D-glucose residues in the 4C1 chair conformation (Fig. 4.11). The secondary hydroxyls are hydrogen bonded around the macrocyclic ring, increasing its stability and reducing its flexibility. The primary hydroxyls have conformations which depend on the host–guest relationship; in general these hydroxyls have a gauche–gauche conformation and point away from the centre of the macrocycle. However, in some instances the C(6) hydroxyls of glucose residues G3 and G4 have the gauche–trans conformation and point towards the centre of the macrocycle; {-CD[1,4-diazabicyclo[2.2.2]octane]13H2O} (Table 4.10) is an example. This conformation is maintained by hydrogen bonding linking the hydroxyls through a water molecule.

Fig. 4.11.(a) Stereoview of -CD looking downwards from the tail (O(6) end) of the -CD molecule in its complex with [2,5-diiodobenzoic acid]7H2O. All the primary hydroxyls are shown with gauche–gauche conformations pointing away from the centre of the macrocycle; the intramolecular hydrogen bonding between adjacent secondary hydroxyls is clearly seen at the head of the molecule at the bottom of the diagram. Complex is listed in Table 4.14. (Reproduced from Hamilton, Sabesan and Steinrauf, 1981.)

Fig. 4.11.(b) Side-view of the -CD conformation in -CD[1,4-diazabicyclo[2.2.2]octane]13H2O complex, showing the C(6) hydroxyls of G3 and G4 with gauche–trans conformation, on the right hand side of the molecule, pointing towards the centre of the macrocycle. The 6-hydroxyls (the narrower tail of the molecule) are at the top of the diagram. Complex is listed in Table 4.10; oxygens cross-hatched. (Data from Harata, 1982b.)

96


˚ , deg.) Table 4.10. Crystal data for isomorphous -CD clathrate inclusion complexes. Unit cells (A have been reoriented where required to maintain a common pattern. All crystals have space group P21 with Z ¼ 2, except where stated otherwise. Compositions are expressed as {-CD[m(guest)] nH2O} with -CD being omitted for brevity. Measurements by XRD at room temperature (nominal 298K) except where stated otherwise; ND ¼ neutron diffraction. V/FU ¼ volume per formula unit ˚3 in A Guest Group I (a) (clathrates) 11H2O (ND at 298K)* 11H2O (ND at 120K) 1.8 HCl8H2O 2HI8H2O CH3OH6.5H2O C2H5OH8H2O (ND – 298K; deuterated crystal) C2H5OH8H2O (ND – 15K; deuterated crystal) KOH9H2O 0.5(DMSO)7.35 H2O Benzyl alcohol [C6H6N2O]6H2O (nicotinamide) [diethanolamine]6.4H2O [but-2-yne-1,4-diol]6.2H2O [propane-1,3-diol]7H2O [C4H10O2]6H2O (1,4-butanediol) (1,5-pentanediol)6.2H2O [diethylene glycol]6H2O [glycerol]7.2H2O [ethylene glycol]8 H2O (squaric acid)6.65 H2O

Refcode; reference

a

b/

c

V/FU

BUVSEQ02; BSHB84 BUVSEQ03; ZSM86 ZZZBVD; SB77 BOBPEN; LS82b BOBPIR; LS82b SIGHOF01; SMS91 SIGHOF02; SMS91 KOBRIC; CNV91 VACZIJ; AC02 DEBGOG; HUOHO85 CACCOY; HKFO83 YIYSII; SKGS95 ZIGZIY; SKGS95 YIYTAB; SKGS95 KUTKOZ; SKS92 YIYSOO; SKGS95 YIYSUU; SKGS95 PIJGOE; GSKS93 PIJGIY; GSK93 MIFHAK; CFM01

21.26

10.31 112.3 10.03 112.5 10.27 109.0 10.28 113.3 10.11 111.0 10.21 111.472 10.00 109.0 10.58 108.4 10.285 109.86 10.101 112.81 10.37 110.5 9.987 111.85 10.092 111.30 9.976 110.88 9.973 110.87 10.014 111.25 9.969 111.62 9.954 111.20 10.021 111.47 10.068 110.16

15.30

1533

14.89

1491

15.04

1515

15.30

1538

15.33

1521

15.215

1527

15.23

1473

15.22

1544

15.155

1532

15.356

1522

15.37

1525

15.247

1506

15.223

1544

15.274

1503

15.271

1508

15.240

1526

15.276

1507

15.251

1509

15.208

1504

15.231

1520

21.62 20.75 21.25 21.03 21.125 20.46 20.20 20.906 21.287 20.43 21.310 20.988 21.116 21.199 21.451 21.288 21.322 21.212 21.117


97

Table 4.10. (Continued ) 0.3(formic acid)7.7 H2O

ASR03

20.986

0.4(acetic acid)7.7 H2O

ASR03

21.044

DIRVOP; H84 POVSIC; StSa98c

20.12

Group I (b) (clathrates) [C6H12N4]6H2O (hexamethylene-tetramine) Trans-cyclohexane-1,4-diol5.4 H2O Group II (brickwork or slipped-tunnel) [C6H12N2]13H2O (diazabicyclo[2.2.2]octane) [C11H11N3SO2]8.3H2O (sulfathiazole) Miscellaneous: space group P212121, Z ¼ 4 2{(Mg(H2O)6 Cl2}3.5H2O (CaCl2)211H2O

20.042

10.169 110.92 10.157 110.67

15.171

1512

15.263

1526

15.29

1557

15.123

1572

16.60 117.4 16.50 117.3

15.44

1752

15.56

1741

10.35 102.1 10.378 102.30

BISTAY; H82b LILLUN; CGN94c

15.40

ZEZTED; NCCR95 HIDZAV; NCCR96

15.95

18.61

23.36

1734

15.88

17.58

24.27

1694

15.26

* A 12-hydrate has been reported (Lindner and Saenger, 1982b; Hamilton, Steinrauf and Van Etten, 1968 (BCDEXD, 01–05, 10; QQQAEV)) with approximately the same cell dimensions as the 11-hydratePerhaps this is an example of concomitant small compositional and structural variationsThe reversible dehydration has been studied (Steiner and Koellner, 1994)The XRD structure at 300K has been reported (Steiner and Koellner, 1994; BUVSEQ01). References: AC02 – Aree and Chaichit, 2002; ASR03 – Aree, Schulz and Reck, 2003; BSHB84 – Betzel, Saenger, Hingerty and Brown, 1984; CFM01 – Crisma et al., 2001; CGN94c – Caira, Griffith, Nassimbeni and van Oudtshoorn, 1994c; CNV91 – Charpin, Nicolis, Villian, Rango and Coleman, 1991; GSKS93 – Gessler, Steiner, Koellner and Saenger, 1993; H82b – Harata, 1982b; H84 – Harata, 1984; HKFO83 – Harata, Kawano, Fukunaga and Ohtani, 1983; HUOHO85 – Harata, Uekama, Otagiri, Hirayama and Ohtani, 1985; LS82b – Lindner and Saenger, 1982b; NCCR95 – Nicolis, Coleman, Charpin and de Rango, 1995; NCCR96 – Nicolis, Coleman, Charpin and de Rango, 1996; SB77 – Szjetli and Budai, 1977; SKGS95 – Steiner, Koellner, Gessler and Saenger, 1995; SKS92 – Steiner, Koellner and Saenger, 1992; SMS91 – Steiner, Mason and Saenger, 1991; StSa95 – Steiner and Saenger,1995; StSa98c – Steiner and Saenger,1998c; ZSM86 – Zabel, Saenger and Mason, 1986;

4.3.1 -Cyclodextrin as host in clathrate inclusion complexes The crystal data are given in Table 4.10; there are some twenty isomorphous (or nearly so) cage complexes (Class I) and one slipped-tunnel (brickwork) structure (Class II), which are illustrated in Figs. 4.12 and 4.13. The -CD clathrates resemble those of -CD in that both can be divided into similar structural types. The arrangement in Class II is head-totail but the slippage between adjacent planes is such that there is very little overlap between successive -CD molecules and hence only minimal hydrogen bonding between them. The two C(6) hydroxyls, which have gauche–trans conformations, point in towards the centre of the macrocycle in the undecahydrate and in the diazabicyclo[2.2.2]octane complex but without disrupting any of the intramolecular hydrogen bonds made by the secondary hydroxyls and without disturbing the round shape of the

98


macrocycle. Gessler, Steiner, Koellner and Saenger (1993) have concluded that ‘‘the general behaviour of the guest molecules is similar for all inclusion complexes of -CD with small hydrophilic molecules studied so far. If the complexed molecule is too small to fill the cavity, several water molecules are also included, thereby forming a cluster of hydrogen-bonded guest molecules. This cluster is dynamically disordered: the molecules are mobile, and perform jumps between different alternative (but discrete) sites. In solution, such rearrangements certainly occur in an even more pronounced way than in the crystalline solid state.’’ One structure that does not fit into any of the tabulated groups is {-CD2,7˚ , ¼ 109.28 , dihydroxynaphthalene4.6 H2O} (a ¼ 14.082. b ¼ 19.079, c ¼ 12.417 A space group P21, Z ¼ 2; Anibarro et al., 2001; CACQED); nevertheless, the packing principles are similar. The 2,7-dihydroxynaphthalene molecule is completely included in the -CD cavity with its long axis along the -CD molecular axis, leading to marked elliptical distortion. These moieties are arranged in herring-bone fashion with the 4.6 water molecules distributed at the narrow (tail) end of the -CD cavity. Another exceptional structure is that of {-CDmefenamic acidxH2O} (a ¼ 15.480. b ¼ 25.589, ˚ , ¼ 98.976 , space group P21, Z ¼ 2; Pop et al., 2001; MUPNEQ) where c ¼ 9.297 A mefenamic acid is 2-[2,3-dimethylphenyl-amino]benzoic acid; the moieties are arranged in herring-bone fashion. This structure is noteworthy because it was solved using highresolution room-temperature synchrotron data from polycrystalline samples; this is the first such structure determination and carries important implications for study of CD complexes unobtainable as single crystals. There is a glass transition at about 150 K and a first-order transition in {-CD11H2O} at 226K, with H ¼ 10.2 kJ/mol and S ¼ 451 J K1 mol1 (Hanabata, Matsuo and Suga, 1987); the first-order nature of the transition is hinted at by the appreciably different cell dimensions at 298 and 120K (Table 4.10). Neutron diffraction (Betzel, Saenger, Hingerty and Brown, 1984; Zabel, Saenger and Mason, 1986) shows that the phase change is due to a cooperative ordering of hydrogen bonds, which are dynamically disordered above 226K; the glass transition is ascribed to freezing of the configuration of protons participating in a four-membered ring of water molecules found only in the low temperature phase. Partially deuterated {-CD[C2H5OH3H2O]5H2O} has also been studied at 15K by neutron diffraction (Steiner, Mason and Saenger, 1989); at room temperature most hydroxyl groups and water molecules, especially those in the cavity region containing [C2H5OH3H2O], are extensively disordered. However, a well ordered network of hydrogen bonds is found at 15K. Later work (Steiner, Mason and Saenger, 1991) used fully deuterated crystals. There are two structures with divalent metal cations which have -CD packings related to the herringbone structure type. These are {-CD2{(Mg(H2O)6Cl2}3.5H2O} (Nicolis, Coleman, Charpin and Rango, 1995; ZEZTED) and {-CD(CaCl2)211H2O} (Nicolis, Coleman, Charpin and Rango, 1996; HIDZAV) (Table 4.10, Miscellaneous). The presence of metal cations (also in the KOH complex) involves new structural considerations. Kþ can be included without much disturbance of the overall structure, as is shown by the resemblance of the cell dimensions and symmetry of the KOH complex to those parameters of the large group of -CD complexes with organic guests. However, divalent cations require more drastic changes for their accommodation. Indeed, Nicolis, Coleman et al. (1995) comment: ‘‘The structure [of the Mg complex] may be considered


99

z y

Fig. 4.12. Diagram of the crystal structure of hexamethylenetetramine (HMT) complex of -CD (DIRVOP) viewed down the [100] axis. This is a representative of the clathrate structures (Table 4.10, Group Ib). The HMT guests (emphasized) are enclosed within the -CD tori, which have a herringbone arrangement. The water molecules have been omitted for clarity and the intermolecular hydrogen bonding is not shown. (Data from Harata, 1984.)

b

b

a

c a

c sin b

Fig. 4.13. The slipped-tunnel or brickwork structure of {-CD[diazabicyclo[2.2.2]-octane] 13H2O}, illustrating structures listed in Table 4.10 (Group II) (cf. Fig. 4.2(c)). (Reproduced from Harata, 1982b.)

in two ways: first, a novel arrangement of –CD monomers into which an inorganic motif is inserted or second, a space-filling arrangement of a –CD assembly within an inorganic matrix . . . The nature of the structure is such that it is difficult to say which substructure, organic or inorganic, determines the overall packing.’’ We add a second quotation about HIDVAZ (Nicolis, Coleman et al. (1996): ‘‘Suitable crystals . . . were obtained after 1 year, by slow evaporation of a highly concentrated CaCl2/–CD aqueous solution.’’ A novel complex of composition {-CD8(pyridine)3H2O}, crystallized from a CD–pyridine–water gel system is noted here because the -CD molecules are present as monomers (Rango, Charpin, Navaza, Keller, Nicolis, Villain and Coleman, 1992;

100


˚ , ¼ 101.87 , P21, KUFHOI). The cell dimensions are a ¼ 14.70, b ¼ 14.74, c ¼ 21.78 A Z ¼ 2. One pyridine is enclosed within the -CD cavity, one is at the primary hydroxyl level and six, together with the water molecules, are in tunnels between the host molecules; thus the composition is better expressed as {-CD[2(C5H5N)]6(C5H5N)3H2O}. The crystals contain 40% pyridine by weight in a low density packing of -CD molecules. This has been described as a ‘‘brickwork’’ structure by Dodds (1999). 4.3.2 4.3.2.1

-Cyclodextrin as host in tunnel inclusion complexes Structural data

The tunnel inclusion complexes of -CD are isostructural and can be divided (with perhaps one known exception) into the five isomorphous groups given in Tables 4.11 to 4.15. This classification is due to Mentzafos, Mavridis, Le Bas and Tsoucaris (1991), who have made other important contributions, discussed below, to the systematization of the crystal chemistry of this family of complexes. Perhaps the most striking feature of the tunnel inclusion complexes is the appearance of head-to-head -CD dimers, with direct hydrogen bonding between the O(2,3) ends of the truncated cones (Fig. 4.14). These dimers are the structural units in the -CD tunnel inclusion complexes and are linked to one another in stacks by hydrogen bonds via water molecules. The driving force for dimer formation appears to be the creation of a large apolar environment for enclosing hydrophobic guests. The various complexes are then distinguished by different modes of stacking of the dimers. We shall first give the facts – crystal data and packing arrangements for the five isomorphous groups, and then show how Mentzafos et al. (1991) have accounted for these arrangements in terms of different packings of essentially similar layers. In the first group of tunnel (CH) structures (Table 4.11) adjacent dimers in a stack are slightly laterally displaced but their axes remain parallell. The reduced triclinic unit cells in Table 4.12 fall into two groups when the standard ˚ , 113, 99, 103 ordering a < b < c is used. The first group has 15.2, 15.5, 18 A ˚ , 99, 113, 103 . This presumably and the second group has 15.2, 15.5, 18 A implies that the structural roles of a and b are interchanged between the two groups. The ternary complexes of pyrene with octanol and cyclohexanol (Table 4.14) have particularly interesting structures, illustrated for the octanol complex in Fig. 4.15 (p. 110). The pyrene molecules are enclosed not within the internal tunnels of the -CD molecules but within the hydrogen bonded portion of the dimer, and without appreciably perturbing the structure of the dimer. A tentative generalization was made that, in analogous complexes, other aromatic molecules could occupy the same site as pyrene and that the third component need not be an alcohol. The second group (IM; Table 4.12) has been listed separately because these structures are of the slipped-tunnel (brickwork) type), the two barbital complexes differing in having one-dimer and two-dimer lengths of chain respectively. In the screw tunnel (SC) structures (Table 4.13; Fig. 4.16) the integrity of the dimers is maintained but these are now both mutually displaced and tilted. Further evidence for the importance of the dimers is given by the ternary {-CD[acetylsalicylic acid0.5(salicylic acid)]11.7H2O} complex where the three guest molecules are enclosed within a -CD dimer (DIFHOP, Table 4.12, Group II, Nishioka, Nakanishi, Fujiwara and Tomita, 1984).


101

Fig. 4.14.(a) Stereoview of the stacking of -CD head-to-head dimers in the tunnel inclusion structure of -CD[2,5-diiodobenzoic acid]7H2O; the disordered guest molecule and the water molecules are not shown. This is an example of the second group of isomorphous -CD tunnel inclusion complexes with a tunnel (CH) structure, listed in Table 4.14. The space group for this group is C2, with four formula units in the unit cell. (Reproduced from Hamilton, Sabesan and Steinrauf, 1981.)

Fig. 4.14.(b) Two stereoviews of a space-filling model of the -CD dimer with enclosed guest molecules in the {-CD[1-adamantanecarboxylic acid]16H2O} structure. The upper view is looking down from the tail of the molecule (i.e. the primary hydroxyl end) and the lower view is normal to the sevenfold axis of -CD; water molecules are not shown. The two -CD moieties and the two 1-adamantanecarboxylic acid moieties (almost entirely enclosed in the -CD cavity, with one carboxyl group protruding more than the other) are both crystallographically independent. The interlocking of secondary hydroxyl oxygen atoms which link the -CDs to form hydrogen bonded head-to-head dimers is clearly shown in the lower diagram. This example comes from the group of isomorphous -CD tunnel inclusion complexes with an intermediate (IM) structure listed in Table 4.12. The space group is P1, Z ¼ 2. These are also called slipped tunnel or brickwork structures. (Reproduced from Hamilton and Sabesan, 1982b.)

102


Table 4.11. Crystal data for some -CD tunnel inclusion complexes with a tunnel (CH) structure; compositions as in Table 4.5. The space group is P1, Z ¼ 2. The reduced cells are given, with standard orientation (a < b < c); these show that the Group I and Group II structures are isostructural Guest

Refcode/ reference

Group I. [C8H8N2O3]7H2O; ( p-nitro-acetanilide)* [C8H8NOBr]13.5H2O; p-bromoacetanilide (benzoic acid)0.35(ethanol)10.3 H2O Group II. [(n-C3H7OH)]9H2O; n-propanol (metastable phase, at 120K; polymorph II). 1,7-dioxaspiro-5,5-undecane# 0.7(Z-tetradecane-7-en-1-al)11.6 H2Ox

a/

b/

c/

CHANAO; HMP78; XADHIT; CD99 AC03a

15.13 88.65 15.20 87.16 15.210 89.13

15.54 81.84 15.61 81.71 15.678 74.64

15.69 76.86 15.74 76.61 15.687 76.40

BCYDPR; JS79 FERCOU; TLRV87 YRM02 XUBXUN

15.316 104.23 15.60 101.4 15.475 101.856

15.461 103.98 15.72 101.7 15.466 101.909

15.575 100.92 15.93 103.2 15.720 103.769

V/FU

1778 1798 1751

1673 1800 1722

Notes: * Pseudo-monoclinic C2 symmetry; # 1,7-dioxaspiro-5, 5-undecane is the pheromone of the olive fly. x (Z-tetradecane-7-en-1-al) is the sex pheromone of the olive pest Prays olea; one molecule is within the dimer and the other between the dimers. References: AC03a – Aree and Chaichit, 2003; CD99 – Caira and Dodds, 1999; HMP78 – Harding, Mclennan and Paton, 1978; JS79 – Jogun and Stezowski, 1979; TLRV87 – Tsoucaris, Le Bas, Rysanek and Villain, 1987; YRM02 – Yannakopoulou et al., 2002.

Tunnels are no longer evident in the chessboard group (CB; Table 4.14). In flurbiprofen A is:

CH3 A

C

In fenoprofen A is :

H

COOH S-(+) enantiomer

F O

Scheme 4.3

Some indication of the influence of guest on overall structure can be obtained from the -CD complexes with flurbiprofen (2-(2-fluoro-4-biphenylyl)propionic acid) and fenoprofen (2-(3-phenoxyphenyl)propionic acid), where structures have been reported with most of the possible combinations of racemic and enantiomeric guests.

Table 4.12. Crystal data for the second isomorphous group of -CD tunnel inclusion complexes with an intermediate (IM) structure (also called slipped tunnel or brickwork structures); compositions as in Table 4.10. The space group is P1, Z ¼ 2The reduced cells are given, with standard orientation (a < b < c); these show that the Group I structures are isomorphous, as are the Group II structures, but Groups I and II are not isomorphous Guest; water content Group I [C11H16O2]15H2O; 1-adamantanecarboxylic acid at 108K. (Note (a)) [C7H8O].xH2O m-cresol (120K); [()-[C15H13O2F]10H2O; ()-(2-(2-fluoro-4-biphenylyl)-propionic acid; flurbiprofen) [(S)-( þ )-[C15H13O2F]10.5H2O; (S)-( þ )-(2-(2-fluoro-4-biphenylyl)propionic acid) [1.5(p-HOC6H4I)]12H2O; p-iodophenol; 153K

Refcode; reference

a/

b/

c/

V/FU

BOGCAB; HS82

15.255 113.54 15.327 113.07 15.420 113.63

15.491 98.87 15.366 99.68 15.490 99.36

17.747 102.54 17.887 102.64 18.633 103.05

1807

15.446 113.52 15.352 113.10 15.440 113.02 15.389 113.43 15.31 112.66 15.40 112.86 15.446 112.99 15.45 113.1

15.513 99.32 15.363 99.40 15.530 98.99 15.519 99.07 15.50 99.47 15.519 99.07 15.452 99.35 15.47 99.4

18.107 102.89 17.985 102.80 18.060 103.44 17.896 103.01 18.30 103.32 17.896 103.01 18.056 103.16 18.08 103.1

1861

15.45 113.6

15.55 99.4

18.09 103.5

1854

BCDMPH; JMS79 CEDMUT; UHI83

CIGXOF; UIH84 BCDPIH10; SJEB78

N-acetyl-L-phenylalanine]12H2O

AGAZIR; ACBS02

[C12H15NO3]13H2O (110K); N-acetylphenylalanine methyl ester N-acety-p-methoxy-L-phenylalanine11.5 H2O

DOCVUM03; S85 CBS01

N-acetyl-L-phenylalanine amide 13H2O

CS01

0.8(nonanoic acid)12 H2O

TEJHAR; RM96b

[0.5(C20H32O4)]10.3H2O; Iloprost is a 1:1 mixture of 16R and S diastereoisomers 4-biphenylacetic acid.xH2O

SHSH89

HHA92

1814 1843

1827 1863 1836 1879 1836 1854 1860

Table 4.12. (Continued ) Guest; water content

Refcode; reference

Group II 0.5[C11H10O2]9H2O; (ethyl cinnamate

BIDMOQ; HST-PU85

[1.5(n-C3H7OH]12H2O; n-propanol at 153K (note (b)) [(C9H8O4)0.5(C7H6O3)]11.7H2O; (acetylsalicylic acid0.5 (salicylic acid)) (3,4-xylidine complex stated to be isomorphous); note (c). 0.5{[C13H24O2]2(C2H6O]24.5 [H2O]}; 1,13-tridecanedioic acid; 173K 0.5{[C14H26O4]0.2(C2H6O]10.93 [H2O]}; 1,14-tetradecanedioic acid 0.5{[C12H22O4]1.16(C2H6O]13.7 [H2O]}; 1,12-dodecanoic acid; 100K. 0.5{[C12H22O4]0.5 (C2H6O]11.1[H2O]}; 1,12-dodecanedioic acid [1.5(R,S)()-(C8H10OS)]17H2O; (R, S)()-methyl-p-tolylsulphoxide [(C8H12N2O3)15.5H2O; (5,5-diethylbarbituric acid (barbital)), Form I N-acetyl-L-phenylalanine]12H2O

CDEXPR, BCDNPR10; JS79, SJEB78 DIFHOP; NNFT84

LONGIE; MTM99; CACPOM; MM01 WISREV; MM00; WISRIZ; MM00 GESVUV; FTMV88; DEVVAB; NAFT84; AGAZOX; ACBS02

N-acetyl-R-phenylalanine]12H2O

AGAZUD; ACBS02

Tert-butyl benzoic acid

HEGXUM; RMHD94

a/

b/

c/

V/FU

15.392 99.74 15.299 99.40 15.247 99.99

15.486 113.61 15.424 113.50 15.475 112.71

18.186 102.78 17.980 103.00 18.310 102.63

1854

15.280 99.79 15.436 99.78 15.251 99.90 15.409 99.71 15.432 99.25 15.497 99.70 15.390 99.27 15.410 98.74 15.417 99.69

15.510 113.13 15.492 113.01 15.456 113.13 15.488 113.12 15.476 113.29 15.549 112.30 15.390 113.39 15.470 113.72 15.476 113.08

18.207 103.02 18.242 102.75 18.153 102.81 18.220 102.90 17.984 102.96 18.123 103.63 17.890 102.75 18.150 103.53 18.244 102.94

1818 1865

1851 1877 1837 1868 1846 1881 1823 1851 1870

7-hydroxy-4-methylcoumarinxH2O

BAS00

7-hydroxyoumarinxH2O

BAS00

2( p-hydroxybenzaldehyde)4.72 H2O; note (d)

BAI02

[(C8H12N2O3)12.5H2O; (5,5-diethylbarbituric acid (barbital)), Form II (4 formula units per cell); note (d)

DEVVEF; NAFT84

15.37 99.47 15.46 99.22 15.262 92.67 15.529 99.89

18.01 113.21 15.55 113.21 15.728 96.97 15.568 93.47

18.01 103.35 18.16 103.23 16.350 103.31 32.327 103.82

1858

Notes: (a) Two other cells have been given for this complex but their relation to the structure solved is not clear (see QQQAEY). (b) 15.229 is also given. (c) In the (acetylsalicylic acid)2(salicylic acid)) complex, the salicylic acid molecule is in the centre of the -CD dimer. (d) The last two complexes have structure combining features of the tunnel and slipped-tunnel arrangements, and are placed here for convenience. They are not isomorphous with the other entries. References: ACBS02 – Alexander et al., 2002; BAI02 – Braga, Aree, Imamura et al., 2002; BAS00 – Brett, Alexander and Stezowski, 2000; CBS01 – Clark, Booth and Stezowski, 2001; CS01 – Clark and Stezowski, 2001; FTMV88 – Fujiwara, Tomita, Marseigne and Vicens, 1988; Vicens, Fujiwara and Tomita, 1988; HHA92 – Harata, unpublished, noted in Harata, Hirayama, Arima, Uekama and Miyayi, 1992; HS82 – Hamilton and Sabesan, 1982; HST-PU82 – Hursthouse, Smith, Thornton-Petit and Utley, 1982; JMS79 – Jogun, McLennan and Stezowski, 1979; MM00 – Makedonopoulou and Mavridis, 2000; MM01 – Makedonopoulou and Mavridis, 2001; MTM99 – Makedonopoulou, Tulinsky and Mavridis, 1999; NAFT84 – Nakanishi, Arai, Fujiwara and Tomita, 1984; NNFT84 – Nishioka, Nakanishi, Fujiwara and Tomita, 1984; RMHD94 – Rontoyianni, Mavridis, Hadjoudis and Duisenberg, 1994; RM96b – Rontoyianni and Mavridis, 1996; S79 – Stezowski, 1985; SHSH89 – Steiner, Hingrichs, Saenger and Hoyer, 1989; SJEB78 – Stezowski, Jogun, Eckles and Bartels, 1978; UHI83 – Uekama, Hirayama, Imai, Otogiri and Harata, 1983; UIH84 – Uekama, Imai, Hirayama, Otogiri and Harata, 1984.

106


Table 4.13. Crystal data for the isomorphous group of -CD tunnel inclusion complexes with a screw tunnel (SC) structure; compositions as in Table 4.5. The space group for this group is P21, with four formula units in the unit cell Guest; water content;

Refcode; reference

a

b/

c

V/FU

[C8H11N]16H2O (p -ethyl-aniline);

CDETAN; TFT81 TFT81

15.30

15.58

1881

15.53

1883

NIZGUY; CGNO96 DUTLIN10; HC88b GETPEA; HC88a GETPAW; HC88a CIVBUE; NFT84 KIFPAQ; S-R99 SAJPIC; HHUT88

15.342

32.31 102.4 32.49 102.6 32.54 102.44 32.23 101.2 32.12 100.8 32.76 101.5 33.189 104.85 32.545 103.56 35.31 102.7

15.324

1868

15.32

1850

15.28

1845

15.35

1880

15.562

1900

15.437

1884

15.50

2069

15.255

1821

15.621

1868

15.47

1902

15.609

1890

[C6H6NI]16H2O (p-iodo-aniline); (L-menthol)14.5H2O [R,S-()-C15H14O3]15H2O; (R,S-()-fenoprofen); at 123K [S-( þ )-C15H14O3]12H2O; (S-( þ )-fenoprofen); at 138K [R-( )-C15H14O3]13H2O; (R-( )-fenoprofen) [phenobarbital]13.5 H2O Adamantone [C11H15O3N3F]9.7H2O; (1-hexylcarbamoyl-5-fluorouracil) (carmofur) C12H10N2O212.7H2O; p-amino-p 0 -nitrophenyl 4,4 0 -diaminobiphenyl9.7H2O [S-( )-(C8H10OS)]17H2O; (S-( )-methyl-p-tolyl-sulphoxide) (Z ¼ 8) [1.5(cyclizine)]12.5H2O; (Z ¼ 8).

QACXEX; BLCS99; GYM03 GESWAC10; VFT88 CDN01

15.29

15.28 15.31 15.26 15.229 15.428 15.51

15.454 15.394 15.495 15.246

31.693 102.92 31.995 103.74 65.04 102.6 65.075 102.6

Note: fenoprofen is (3-phenoxyphenyl)propionic acid. References: BLCS99 – Brett, Liu, Coppens and Stezowski, 1999; CDN01 – Caira, Dodds and Nassimbeni, 2001; CGNO96 – Caira, Griffith, Nassimbeni and Oudtshoorn, 1996; GYM03 – Giastas et al., 2003; HC88a – Hamilton and Chen, 1988a; HC88b – Hamilton and Chen, 1988b; HHUT88 – Harata, Hirayama, Uekama and Tsoucaris, 1988; NFT84 – Nakanishi, Fujiwara and Tomita, 1984; S-R99 – Sanchez-Ruiz et al., 1999; the chemical composition was not given; TFT81 – Tokuoka, Fujiwara and Tomita, 1981; VFT88 – Vicens, Fujiwara and Tomita, 1988.

{-CD[()-flurbiprofen]10H2O} (Uekama, Hirayama, Imai, Otogiri and Harata, 1983) and {-CD[(S)-( þ )-flurbiprofen]10.5H2O} (Uekama, Imai, Hirayama, Otogiri and Harata, 1984) have very similar structures (Table 4.12); in the first of these the R-( ) enantiomer is found in one half of the -CD dimer and the S-( þ ) enantiomer in the other half, with a head-to-head arrangement of the guests in which the carboxyls are hydrogen bonded to a primary hydroxyl of the -CD and to a water molecule. A very similar

Table 4.14. Crystal data for the isomorphous group of -CD tunnel inclusion complexes with a tunnel (CH) structure; compositions as in Table 4.10. The space group for this group is C2, with four formula units in the unit cell Guest and water content

Refcode; reference

a

b/

c

V/FU

0.5(KI79H2O) (actually P21 but pseudo C2)

COCMIQ; BHN83

19.61

15.80

1790

NaI38H2O (details not given)

CYDXTF; NS80

19.58

15.75

1778

0.5{[Fe(cp)(mesitylene)]þPF66H2O}

KOGLIB; KR91

19.241

15.768

1745

0.5{[Fe(cp)(biphenyl)]þP F68H2O}; (actually P21 but pseudo C2) [ethyl p-aminobenzoate]7.5H2O

KOGLOH; KR91

19.206

15.674

1733

BIHJEH; HS82

18.75

15.66

1690

[C8H8O2]11H2O (m-toluic acid)

QQQAFA; HSV68

18.86

15.76

1728

[C7H5O2I]11H2O (m-iodobenzoic acid)

QQQAFG; HSV68

18.88

15.69

1724

[(C7H3I2O2]9H2O (2,5-diiodobenzoic acid)

QQQAFJ; HSS81, HSV68

19.19

15.74

1761

[C8H9NO2]13H2O (acetaminophen)¤

CD00

19.207

15.700

1740

[0.5(cimetidine)]11H2O

D99

19.22

15.75

1757

[C10H13BrO]9H2O (2-bromo-5-t-butylphenol)

QQQAFS; HSS81, HSV68

19.24

16.02

1798

[C10H13BrO]9H2O (2-bromo-4-t-butylphenol);

HSV68, HSS81

19.19

15.97

1779

[C7H5O2Br]13H2O (m-bromobenzoic acid)

QQQAFP; HSV68

19.23

15.80

1760

[3,3-dimethylbutylamine]11H2O

VIJXAN; MHT91

19.19

15.89

1773

[0.94(3,5-dimethylbenzoic acid)]9.1H2O

YOVVIO; RM94

19.37

16.00

1811

[4-t-butyltoluene]8.5H2O

KUTJUE; MH92

19.24

24.51 109.5 24.44 109.3 24.415 109.55 24.334 108.88 24.53 110.2 24.67 109.5 24.77 110.0 24.76 109.6 24.48 109.52 24.57 109.1 24.66 108.9 24.58 109.17 24.58 109.5 24.56 108.8 24.71 108.9 24.47 109.9

15.84

1753

Table 4.14. (Continued ) Guest and water content

Refcode; reference

a

b/

c

V/FU

[0.5(cyclopentadienyl mesitylene iron PF6)] H2O

KOGLIB; KR91

19.24

15.77

1745

[0.5(cyclopentadienyl biphenyl iron PF6)] H2O

KOGLOH; KR91

19.21

15.67

1732

[benzophenone] (water content not given)

DEVTED; LBdE84

19.24

15.94

1775

[biphenyl] (water content not given)

DEVTIH; LBdE84

19.34

15.80

1760

spiroacetal

TEMCIX; RLVT96

19.368

15.940

1787

[0.75(octanol)0.5(pyrene)].7.25H2O;

PUKPIU; UR98

19.326

15.922

1778

[1.5(cyclohexanol)0.5(pyrene)].5.25H2O;

PUKPOA; UR98

19.254

15.914

1767

[0.5 ((Z)-9-dodecen-1-ol)]9.6H2O0.5C2H5OH

ZUZXOH; MMH96

19.238

15.790

1752

0.5[C13H26O2].9.8H2O Tridecanoic acid

SOBHUM, 01, 02; MPAM00

19.363

15.937

1799

0.5[C14H26O2].97.5H2O (Z)-Tetradec-7-enoic acid

SOBJEY, 01, 02; MPAM00

19.316

15.936

1788

0.5(-naphthyloxyacetic acid)

ODEJOW; KYMM01

19.341

15.975

1801

Poly(tris(ethylene glycol))8H2O

BEZLAT; REFER; 173K

18.726

15.398

1653

C42H70O351.625(C3H6O)13H2O Poly(trimethylene oxide)13H2O C42H70O351.625(C3H6O)13H2O Poly(propylene glycol)13H2O 0.3(ethanol)12 H2O

KMK00 110K

19.369

15.983

1796

KMK00 110K

19.332

15.961

1880

AC03b

19.292

15.884

1785

1.5(C9H6O2)12 H2O; coumarin

GOSQOU; BAC99

19.322

24.42 109.6 24.33 108.9 24.56 109.5 24.49 109.8 24.450 108.72 24.441 109.00 24.467 109.47 24.477 109.52 24.597 108.55 24.564 108.98 24.632 108.77 24.475 110.48 24.540 108.92 24.572 109.00 24.691 109.35 24.641 106.76

16.050

1829

6-methylcoumarin

BAS00

19.210

7-methylcoumarin

BAS00

19.348

7-methoxycoumarin

BAS00

20.058

Butyrophenone11.5H2O

DGPOO; BS00

19.352

Valerophenone11H2O

DOGPUU; BS00

19.339

Clofibric acid Trans-cinnamic acid

CBM01 (not seen) XERTET; KMM00

19.422

4,7-dimethylcoumarin. 17H2O#

MASBAJ; BAS00

19.513

1,2-bis(4-aminophenyl)ethane12.35H2O

GYM03

19.319

[C10H11NO3]14.2H2O (diacetamate) (Z ¼ 8)

D99

19.275

[C8H15N7O2S3]10.5H2O (famotidine) (Z ¼ 8)

D99

37.72

¤ #

24.600 109.46 24.582 109.84 24.441 109.84 24.599 109.38 25.581 109.08 24.461 108.65 24.024 104.49 24.19 103.92 24.187 109.12 15.50 102.2

15.726

1752

15.784

1765

15.662

1806

15.916

1787

16.010

1871

15.941

1794

18.414

1862

33.315

1889

34.289

1888

26.91

1922*

The ibuprofen complex is isomorphous (Brown, 1997). Structure not yet determined. cf. MASBIR, MASBOX for other substituted coumarins.

References: BAC99 – Brett, Alexander, Ckark, Ross, Harbison and Stezowski, 1999; BAS00 – Brett, Alexander and Stezowski, 2000; BHN83 – Betzel, Hingerty, Noltemeyer, Weber, Saenger and Hamilton, 1983; BS00 – Brett and Stezowski, 2000; CBM01 – Caira, Bourne and Mvula, 2001 (not seen); CD00 – Caira and Dodds, 2000; D99 – Dodds, 1999; GYM03 – Giastas, Yannakopoulou and Mavridis, 2003; HS82 – Hamilton and Sabesan, 1982; HSS81 – Hamilton, Sabesan and Steinrauf, 1981; HSV68 – Hamilton, Steinrauf and Van Etten, 1968; KR91 – Klingert and Rihs, 1991b; KMM00 – Kokkinou et al., 2000; KYMM01 – Kokkinou, Yannakopoulou, Mavridis and Mentzafo, 2001 LdR84 – Le Bas, de Rango, Ryanek and Tsoucaris, 1984; MPAM00 – Reported as triclinic by Makedenopoulou et al (2000) and corrected to monoclinic by Marsh et al. (2002); MHT91 – Mavridis, Hadjoudis and Tsoucaris, 1991; MMH96 – Mentzafos, Mavridis and Hursthouse, 1996; NS80 – Noltemeyer and Saenger, 1980; RLVT96 – Rysanek et al., 1996; RM94 – originally given as triclinic by Rontoyianni and Mavridis (1994) and corrected to monoclinic by Herbstein and Marsh (1998); UR98 – Udachin and Ripmeester, 1998; UWR00 – Udachin, Wilson and Ripmeester, 2000.

110


Octanol

Pyrene

Octanol

Fig. 4.15. The tunnel (CH) structure of {-CD[0.75octanol0.5pyrene]7.25H2O} viewed approximately along [010]. The hydrogen-bonded -CD head-to-head dimers are clearly visible, the pyrene is in a hydrophobic site between the molecules of the dimer, and the octanol molecule, whose OH and methyl ends could not be distinguished, extends between dimers. Note that here, and in Fig. 4.16, the tunnel axis runs horizontally and not, as is more usual in such diagrams, vertically. (Reproduced from Udachin and Ripmeester, 1998.)

w

c sin b

A– b

B A⬘

Fig. 4.16. A schematic representation of the molecular packing of the head-to-head dimers in {-CD[ p-ethylaniline]16H2O} (CDETAN) seen along [100]. A and B represent the crystallographically independent -CD molecules and the solid circles the water molecules. The guest molecules, enclosed within the macrocyclic rings, are not shown explicitly. This is an example of the screw tunnel (SC) structures listed in Table 4.13. (Reproduced from Tokuoka, Fujiwara and Tomita, 1981.)

arrangement is found when the guest is the S-( þ ) enantiomer. The -CD complexes with (R,S)- (Hamilton and Chen, 1988b), R-( )-(Hamilton and Chen, 1988a) or S-( þ )fenoprofen (Hamilton and Chen, 1988a) present a somewhat different picture; although isomorphous (Table 4.13) there are differences in cell dimensions and water content. This is in accord with the fact that these complexes are diastereoisomers. In the R-( ) complex the guests are in a head-to-head arrangement while in the S-( þ ) complex there is a head-to-tail arrangement. The complex with (R,S) guests actually has a S : R ratio of 3 : 1, showing that there has been a pair of S guests in head-to-tail arrangement, or a pair of


111

Table 4.15. Crystal data for the isomorphous group of -CD tunnel inclusion complexes with a chessboard (CB) structure; compositions as in Table 4.10. The space group for this group is C2221, with eight formula units in the unit cell Guest and water content

Refcode/reference

a

b

c

V/FU

[benzil] (water content not specified [phenylethylmalonic acid] (water content not specified) [1-hydroxymethyl-adamantane] 11H2O trans-{Pt(PMe3)Cl2(NH3)] 5.5H2O [C11H16O]10H2O (4-tert-butylbenzyl alcohol) 2-aza-5-hydroxy-adamantane 10.5 H2O [2-methyl-2,4-pentanediol] 11.5H2O [C10H14O]13H2O (m-t-butylphenol), Z ¼ 16; probably a superstructural version. [C8H8O2Hg]12H2O phenylmercuric acetate, space group P22121, Z ¼ 4

DEVTON; MM91

19.58

24.00

32.84

1929

DEVTUT; LBdR84

19.09

24.27

32.58

1887

FASXUS; H85

19.16

24.27

32.58

1894

GIPFEQ; ASS88

19.43

24.08

32.50

1901

KOFJEU; MM91

19.196 24.393 32.808 1920

MECQUK; BGKB00 19.144 23.950 32.670 1872 TECYIJ; ZTK96

19.69

24.13

36.61

2174

QQQAFM; HSV68

19.15

24.33

62.78

1828

QQQAFD; HSV68

17.66

11.45

32.74

1655

References: ASS88 – Alston, Slawin, Stoddart, Williams and Zarycki, 1988; BGKB00 – Bobek et al., 2000; H85 – Hamilton, 1985; HSV68 – Hamilton, Steinrauf and Van Etten, 1968; placed here because of possible relationships with cell dimensions of other complexes in this table. LBdR84 – Le Bas, de Rango, Ryanek and Tsoucaris, 1984; MM91 – Mentzafos, Mavridis, Le Bas and Tsoucaris, 1991; ZTK96 – Zhukhlistova, Tischenko, Kuranova, Vainshtein, Mattson and Korpella, 1996;

R guests in head-to-head arrangement. Thus structural features of the diastereoisomers are copied to the pseudo-racemate. 4.3.2.2 An overall structural description Mentzafos et al. (1991) have shown that it is possible to transform the unit cells of the various groups of isomorphous complexes to a partially common measure. Thus the unit cells of the intermediate group (IM, Table 4.12) transform to a (non-standard) cell ˚ , ¼ 98.7, ¼ 116.3, ¼ 89.8 . The screw tunnel (SC; with a ¼ 19.2, b ¼ 24.2, c ¼ 18.1 A Table 4.13) structures can also be transformed to cells with dimensions approximately ˚ , ¼ 90.6 , space group P21 (here the unique monoclinic a ¼ 19.3, b ¼ 23.9, c ¼ 32.5 A axis is [001]). Similar results can be obtained for the first (triclinic) group of tunnel structures (Table 4.11), and it will immediately be noticed that the orthogonal a and b axes in the C2 (CH; Table 4.14) and C2221 (CB; Table 4.15) groups have similar values. Thus ˚ 2; the invariant packing unit in all these structures is a C-centred layer with an area of 460 A there is a pseudo-close-packed (or hexagonal) arrangement of -CD dimers, with each dimer ˚ and four others at 15.4 A ˚ ( ¼ 0.5 [(192 þ 242)1/2]). surrounded by two others at 19.3 A

112


CH mode

SC mode

CB mode IM mode

Fig. 4.17. Schematic view of the packing of -CD dimers. In the upper part of each of the various modes each -CD monomer is represented by a heptagon of O4 atoms. In the lower part the lateral displacement between two consecutive -CD dimers is shown with each heptagon representing a -CD dimer. (Reproduced from Giastas et al. (2003)).

The four packing modes shown in Fig. 4.17 have been clearly described by Giastas et al., (2003; p. 296): ‘‘ . . . All have the common feature that they consist of close-packed layers of -CD dimers and are generated by the two-dimensional invariable layers’ different relative positioning. In the CH mode the dimer’s sevenfold axis forms an angle of approximately 10 with the stacking axis. The two-dimensional layers stack in parallel so that the dimers align almost on top of each other to form channels, slightly deformed in the interface between dimers (interdimer interface). The channels are hydrophobic and the guests inside them are shielded from the water environment. The lateral displacement ˚ . On the contrary, in CB of two consecutive -CD dimers along the channel is 2.7 A ˚ , every dimer mode the lateral distance between two dimers of successive layers is 8.9 A being surrounded by solvent molecules. The primary faces and therefore the guest are exposed to the polar environment of water molecules and neighbouring hydroxy groups. The dimer’s sevenfold axis forms an angle of approximately 10 with the stacking axis, but the two-dimensional layers are related by a twofold screw axis and


113

they are not parallel (dihedral angle of 20 ). In the IM mode, a case between the CH and CB modes, adjacent layers are parallel but the dimer’s sevenfold axis forms an angle of approximately 20 with the stacking axis. Consequently, dimers are far from ˚ (the inner diameter of the primary exactly aligned, their lateral displacement being 6 A face of -CD). Thus, a breaking of the channel is observed that leaves parts of the guests free to interact with hydroxy groups of adjacent hosts, as well as with water molecules. Finally, in the SC mode, although the lateral displacement between two ˚ as in the CH mode, the dimer’s consecutive -CD dimers along a channel is only 2.7 A sevenfold axis forms an approximate angle of 10 with the stacking axis but the two-dimensional layers are related by a twofold screw axis and they are not parallel, therefore the guests interact with water molecules and hydroxy groups of adjacent host channels. The guest(s) emerging from the two primary faces of the dimer, is (are) situated between the previously mentioned practically invariant layers and therefore plays a crucial role in the packing. . . . ’’ The important role of the guests in determining the arrangement of the layers was illustrated (Makedonopoulou and Mavridis, 2000) by comparing the structures of -CD dimer complexes containing long aliphatic monocarboxylic and ,!-dicarboxylic acids with host : guest ratio 2 : 1; one dimer is threaded by one molecule of a long guest. Aliphatic monoacids with 12–16 C atoms induce tunnel packing (CH; Table 4.14) while aliphatic diacids with 10–16 C atoms have an intermediate packing mode (IM; Table 4.12). Their suggestion is that ‘‘ in the case of the aliphatic monoacids the hydrophobic end of the amphiphilic guests influences the packing towards the channel [CH] mode in order to protect that part of the guest from the polar aqueous environment that surrounds the dimers. The carboxyl groups, found entrapped in the hydrophobic channel, self-associate forming carboxylic dimers and thus stabilizing the whole system. In contrast, in the case of the diacids the two polar ends of the guest are free to interact with the solvent environment.’’ The number of water molecules per -CD ranges from 7.5 (biphenyl) to 13.4 (1-adamantanecarboxylic acid), spread over 12–16 sites. Although the water molecules have high displacement factors, indicating disorder and appreciable thermal motion, Tsoucaris and coworkers (Mentzafos et al., 1991) deduced that there does exist a quasiinvariant water network organized in layers. There are two separate subnetworks, in one there is hydrogen bonding of water to primary hydroxyls and in the other to secondary hydroxyls. These networks make an important contribution to the overall cohesion of the crystal because there are few direct hydrogen bonds between dimers. Between one and three waters are linked only to other waters (the so-called secondary hydration sphere). There is evidence from the displacement factors that the waters are more tightly linked than the guest molecules. The disposition of the guest molecules depends on their nature – on their size and where the hydrophobic and hydrophilic groups are located and how these interact with the -CD hosts and water molecules. The overall resemblances between the different structures in this family have been demonstrated; the differences, small and not-so-small, derive from the detailed nature of the host–guest interactions, and those involving water molecules. In contrast to the -CD complexes, where only about two-thirds of the structures fall into isostructural or isomophous groups, most of the -CD complexes fit into one of the categories given in the above Tables.

114

4.3.3


Exceptional -cyclodextrin structures

There are, at the time of writing, three other -CD structures that do not fit into the classification scheme described above. The first is {-CD[C14H10Cl2NO2]Naþ 11H2O}, where [C14H10Cl2NO2]Na þ is the sodium salt of the anti-inflammatory agent diclofenac, 2-[(2,6-dichlorophenyl)amino]benzeneacetic acid (Caira, Griffith, Nassimbeni and van Oudtshoorn, 1994a (HEHJEJ); Caira, Griffith and Nassimbeni, 1998; ˚ (this is MANNOE). The space group is P61, with Z ¼ 6, a ¼ 15.956(8) and c ¼ 50.95(1) A the first -CD complex found with a hexagonal space group although there are a number of -CD complexes with hexagonal space groups (Table 4.3)). The units of the complex are arranged in a regular head-to-tail fashion about the 61 axis, which is approximately normal to the mean plane of the -CD ring. The guest molecules are inserted between the hosts; Caira et al. describe this as the formation of ‘‘an endless helical host channel with a ˚ .’’ Formation of -CD dimers is presumably inhibited because the 2,6pitch of 51 A dichlorophenyl groups are too bulky to enter the tori of the -CD molecules. The second is {-CD[C14H10Cl2NO2]Naþ16H2O}, where [C14H10Cl2NO2]Naþ .is the sodium salt of the anti-inflammatory agent meclofenamate, 2-[(3-methyl-2,6dichlorophenyl)amino]benzoic acid; the crystals are orthorhombic, space group P212121, ˚ (Caira, Griffith and Nassimbeni, with Z ¼ 4, a ¼ 15.087(2), b ¼ 17.967(2), c ¼ 20.634(4) A 1998; MANNOE). The crystal structure has some resemblances to those of the complexes of per-trimethyl--CD (next paragraph). The third is {-CD[0.5(C13H22N4O3S)] 12.5H2O}, where the guest is ranitidine; the crystals are orthorhombic, space group P212121, with Z ¼ 4, ˚ (Dodds, 1999). The structure is not known. a ¼ 15.10, b ¼ 15.37, c ¼ 37.03 A

4.3.4


-Cyclodextrins chemically modified by methylation have advantages as pharmaceutical carrier molecules over the parent compound because of higher aqueous solubility and greater protection against hydrolysis both in solution and the solid state. The chemically modified -cyclodextrins whose inclusion complexes have been studied crystallographically are heptakis(2,6-di-O-methyl)--CD (C56H98O35, abbreviated as per-dimethyl--CD) and heptakis(2,3,6-tri-O-methyl)--CD (C63H112O35, abbreviated as per-trimethyl--CD). In per-dimethyl--CD the O(3) hydroxyl groups form O(3)˚ and the H . . . O(2) intramolecular hydrogen bonds with an average length of 2.85(3) A macrocycle has a round and nearly symmetrical structure rather like that of the parent -CD. Five crystal structures have been reported for complexes with per-dimethyl-CD as host; the cell dimensions (Table 4.16) show that there are three different structural arrangements. The conformation of the host molecule in the adamantol complex is reported to be similar to that of -CD in its complexes but the nature of the packing was not described. The isomorphous pair of p-iodophenol and p-nitrophenol complexes have clathrate structures in which, remarkably, the organic guests are located between the host molecules and only the water molecules are included within the tori (Fig. 4.18). The carmofur (1-hexylcarbamoyl-5-fluorouracil) complex has the guest disordered between two sites, with the hexyl group of one guest inserted into the torus from the secondary hydroxyl side, while the other guest is located between host molecules. The only other example of intercalation of a guest between CD hosts is


115

Table 4.16. Crystal data for clathrate inclusion complexes with per-dimethyl--CD as host Guest and water content

Refcode/reference

a

b/

c

V/FU

13.821 13.976 14.163 14.797 14.779 15.463 14.163 24.21 11.080

17.424 20.763 20.828 18.853 18.965 18.922 23.096 19.33 14.932

29.610 28.807 29.261 28.989 28.741 27.852 27.641 18.27 44.906

1783 2090 2158 2022 2014 2037 2260 2138

14.278

15.731

31.149

1749

Group II. Space group P21; Z ¼ 2 [C11H15O3N3F]3H2O; carmofur SAJPOI; HHUET88

15.70

15.95

2014

2H2O

CEQCUW; ASLH99

15.241

23.324

1851

0.5 (acetic acid)1.5 H2O

NITSIS; SN97

15.165

18.53 106.6 10.639 101.80 10.613 102.02

23.188

1825

Group I. Space group P212121 (Z ¼ 4) Anhydride ZULQAY; StSa95a 14.7H2O (at 100K). BOYFOK03; SPHG01. 14.7H2O (at 300K). BOYFOK04; AHSRS00 DEZMOK10; H88 6H5O[IC]2H2O; p-iodophenol DEZMIE10; H88 [ p-nitrophenol]2H2O [C11H8O2]3H2O; 2-naphthoic acid WAGHAN; H93, 99 Prostaglandin PGF2 BOYCAX; SCE81 BEFJOL; CES81 [adamantol]12H2O (at 120K) m-cresol acetate COFLOY; PG84 Clofibric acid CBM01 (not seen) CH3CN2H2O AHSRS00a

References: AHSRS00a – Aree, Hoier, Schultz, Reck and Saenger, 2000a; ASLH99 – Aree, Saenger, Leibnitz and Hoier, 1999; CBM01 – Caira, Bourne and Mvula, 2001 (not seen); CES81 – Czugler, Eckle and Stezowski, 1981; H88 – Harata, 1988; H93 – Harata, 1993; thermal motion of the guest has been analyzed (H99 Harata, 1999); HHUET88 – Harata, Hirayama, Uekama and Tsoucaris, 1988; PG84 – Pohlmann, Gdaniec, Eckle, Geiger and Stezowski, 1984; SCE81 – Stezowski, Czugler and Eckle, 1981; SN77 – Selkti et al., 1997; SPHG01 – Stezowski, Parker, Hilgenkamp and Gdenic, 2001; StSa95a – Steiner and Saenger, 1995a.

o

b

a

Fig. 4.18. A schematic diagram of the crystal structure of the isomorphous complexes of the perdimethylated--CD and p-iodophenol or p-nitrophenol, showing the water molecules included in the tori and the guest molecules located between the hosts. (Reproduced from Harata, 1988.)

{-CD[m-nitrophenol]m-nitrophenol6H2O} (see Table 4.4). No explanation has yet been advanced for these exceptional arrangements. An interesting serendipitous result was obtained when postulated per-dimethyl -CD was complexed with synthetic (hence racemic) 1,7-dioxaspiro-[5.5]undecane, which is the

116


Table 4.17. Crystal data for the tunnel inclusion complexes with per-trimethyl--CD as host; space groups P212121 with Z ¼ 4 Guest and water content

Refcode; reference

a

b

c

V/FU

Uncomplexed monohydrate

HEZWAK; CGN94b, SS98 GELKEN10; HHA92 RONWOG; BCNO96 CAMPIP; HUOH83 PAFSOE; HHA92 ZIFQOU; CGN94/5 COYXET20; HUI88 COYXAP10; HUI88 PINMAA; MMS94 QOYLIZ; MYM01 XAQJII; RMIB98, CP01 NIZHAF; CGNO96 MODHUI; CP01 CBM01 (not seen) BEYLOG; HUOH82a

14.823

19.382

26.534

1902

15.67

20.80

25.49

2077

15.232

21.327

27.597

2241

15.00

21.37

28.21

2261

14.890

21.407

28.540

2275

15.18

21.41

27.67

2248

15.271

21.451

27.895

2284

15.092

21.714

28.269

2316

14.796

22.444

27.720

2301

10.936

25.530

29.640

2069

11.149

25.664

29.427

2105

11.060

26.138

29.669

2144

11.190

26.080

29.185

2129

11.440

23.674

13.531

1831

[C6H5IO] m-iodophenol [C13H18O2] (S)-ibuprofen [C6H5OI]4H2O p-iodophenol [C14H12O2]H2O 4-biphenylacetic acid [C14H14O3](S)-naproxen* [C15H13O2F] (S)-( þ )-Flurbiprofen [C15H13O2F]H2O (R)-( )-Flurbiprofen [C14H28O2]2H2O ethyl laurate [S]-1,7-dioxaspiro-[5.5] undecane0.57 H2O methylcyclohexane (L-menthol]2H2O (248K) (R)-5-ethyl-1,3,5-trimethylhydrantoin Clofibric acid p-iodoanilineH2O

* (S)-6-methoxy--methyl-2-naphthaleneacetic acid. References: BCNO96 – Brown, Caira, Nassimbeni and Oudtshoorn, 1996 CBM01 – Caira, Bourne and Mvula, 2001 (not seen); CGN94/5 – Caira, Griffith, Nassimbeni and van Oudtshoorn, 1994/5; CGN94b – Caira, Griffith, Nassimbeni and van Oudtshoorn, 1994b; CP01 – Cardinael et al., 2001; HHA92 – Harata, Hirayama, Arima, Uekama and Miyaji, 1992; HUI88 – Harata, Uekama, Imai, Hirayama and Otagiri, 1988; HUOH82a Harata, Uekama, Otagiri and Hirayama, 1982a; HUOH83 Harata, Uekama, Otagiri and Hirayama, 198a, b; MMS94 – Mentzafos, Mavridis and Schenk, 1994; MYM01 – Makedonopoulou, Yannakopoulou et al., 2001; RMIB98 – Rontoyianni, Mavridis, Israel and Beurskens, 1998; SS98 – Steiner and Saenger, 1998.

major component of the pheromone of the olive fruit fly; the guest is a volatile liquid at room temperature (cf. Table 4.11). Structure analysis of the triclinic crystals (P1, Z ¼ 1; Rysanek, Le Bas, Villain and Tsoucaris, 1992; JOSWOD), which contained also a methanol molecule, showed that the host was in fact 2a,2b,2c,2d,2e,2f,3a,3g,6a,6b,6c,6d,6e,6f,6g-pentadeca-O-methyl--cyclodextrin. Thus

ROTAXANES AND CATENANES OF CYCLODEXTRINS

117

five rings had been 2,6-dimethylated, ring A trimethylated and ring G 3,6-dimethylated; the overall formula of the complex was {C57H100O35[C9H16O2]CH3OH}. The guest was totally enclosed within the macrocycle and resolution had occurred on crystallization, as only the S-enantiomer was found. The uncomplexed per-trimethyl--CD monohydrate and the anhydrous m-iodophenol complex (first two in Table 4.17) have cell dimensions which are related to those of the following six, but nevertheless are significantly different. In the monohydrate, the pertrimethyl--CD macrocycle has six rings in the 4C1 conformation, while G2 is in the unusual 1C4 conformation; in the m-iodophenol complex one trimethylglucose unit adopts the uncommon 0S2 high-energy skew-boat conformation, intermediate between 4C1 and 1 C4. These conformational differences must be taken into account in any assessment of the energetics of complexation. The other fully methylated complexes shown in Table 4.17 are essentially isostructural despite differences in water content, and have a tunnel structure with stacking of head-to-tail host molecules along [010]; however, adjacent molecules in the stack are shifted laterally so that the stack axis is zigzag rather than straight. In all these complexes the host molecule is considerably distorted from the regular, round shape of -CD itself towards an elliptical cross section because intramolecular hydrogen bonds cannot be formed and because of the steric hindrance of the methyl groups; the phenyl ring of the guest is within the cavity, with the phenolic OH protruding and hydrogen bonded to two water molecules. The fully methylated host does not form a complex with racemic flurbiprofen but instead separate complexes are formed with R- and S-flurbiprofen. The flurbiprofen guests in the fully methylated complexes have a head-to-tail arrangement in contrast to the head-to-head arrangement found in their -CD complexes (Section 4.3.2). The enantiospecific separation of the enantiomers of the racemic olive fly pheromone 1,7-dioxaspiro[5.5]undecane [] by complexation of the (R)-enantiomer with hexakis(pertrimethyl)--CD and of the [S]-enantiomer with heptakis(pertrimethyl)--CD has already been noted. The binding constants of the two complexes determined by NMR in aqueous solution are respectively 6600 M-1 for the (R)-enantiomer – hexakis (pertrimethyl)--CD complex and 935 M-1 for the [S]-enantiomer – heptakis (pertrimethyl)--CD complex. These differences can be explained by the details of the interactions between enantiomer and cyclodextrin in the respective pairs.

4.4 Rotaxanes and catenanes of cyclodextrins As noted earlier, these compounds, reviewed by Nepogodiev and Stoddart (1998), straddle the borders of present relevance; we consider only those compounds where crystal structures have been reported. Pseudorotaxanes of composition {2(-CD)[1,12diaminododecane]14 H2O)} (Rontoyianni and Mavridis, 1999; BOLVUT) and {2(CD)[12-diaminododecanoic acid]0.6(ethanol)14.4H2O} (Eliadou et al., 1999; VEHQAA) have the long-chain molecules threaded through the torus of head-to-head -CD dimers; the complexes are not isomorphous. These are true complexes in our present sense. Some other examples are probably out-of-bounds. It has been contended that cyclodextrins require chemical modification by replacement of hydroxyl by bulky substituents

118


in order to behave as useful models for enzymes. This has been done for -CD by monosubstitution of an hydroxyl by a bulky hydrophobic group; specifically, among the substituents used have been t-butyl-thio (Hirotsu, Higuchi, Fujita, Ueda, Shinoda, Imoto and Tabushi, 1982), phenylthio (Kamitori, Hirotsu, Higuchi, Fujita, Yamamura, Imoto and Tabushi, 1987) and –CH2NH(CH2)6NH2 (Dimitrius, Terzis, Coleman and de Rango, 1996). Unsymmetrical disubstitution has also been used to give 6A,6D-deoxy-6A-(tbutylthio)--CD (Fujita, Matsunaga, Imoto, Hirotsu, Kamitori and Higuchi, 1987). Crystal structure analysis shows that these are all self complexes, in which the -CD portion of the molecule acts as host and the substituent as guest, the substituent on one host being enclosed within the macrocycle of an adjacent molecule. O

O

O

O

NH O

per-dimethyl-b-CD O O

O

O

O

NH

bislactam

Scheme 4.4

The 2-catenane of heptakis(2,6-di-O-methyl)--CD and a macrocyclic bislactam (Armspach, Ashton, Moore, Spencer, Stoddart, Wear and Williams, 1993; YAPSEN) resembles more the 2-catenane shown in Fig. 3.9 than it does the -CD clathrate and tunnel inclusion complexes described above; thus Fig. 3.9 gives a better impression of the molecular shape than the schematic diagram shown above. Again we have crossed the border.

4.5 4.5.1

-Cyclodextrin as host

-Cyclodextrin as host in clathrate inclusion complexes

The structure of {-cyclodextrin[nH2O]} (-CD is C48H80O40) has been determined in three separate studies and a comparison is illuminating. The reports agree in regard to the overall structure, which is of the cage type and resembles the -CD cage structures; the -CD molecules are stacked along [010]. However, small differences in water content and cell dimensions were reported (Table 4.18), as well as differences in the distribution of water molecules in the unit cell. Another difference is that CYOCAM has one of the glucose residues disordered (even though the measurements were made at 120 K) while the macrocycle is completely ordered in CIWMIE10. Similar minor differences in crystals from different batches have been encountered with -CD complexes (Fujiwara, Yamazaki, Tomizu, Tokuoka, Tomita, Matsuo, Suga and Saenger, 1983; Steiner, 1990) and in


119

Table 4.18. Comparison of crystal data reported in determinations of the crystal structure of hydrated -CD. The complexes crystallize in space group P21 with Z ¼ 2 Water content, and temperature

Refcode; reference

a

b/

c

V/FU

[17 H2O] (at 120K)

CYOCAM; MS80

20.253(8)

16.892(6)

1731

[14 H2O] (at 300K)

CIWMIE10; H87

20.271(2)

16.847(2)

1831

[11 H2O] (at 300K) ND

CIMSAS; HBS84

20.287(10)

10.494(5) 105.32(1) 11.098(2) 104.97(1) 22.079(7) 105.07(4)

16.858(12)

1823

References: HBS84 – Hingerty, Betzel and Saenger, 1984; MS80 – Mclennan and Stezowski, 1980; H87 – Harata, 1987.

{-CD[n-propanol]17H2O} (see below). These differences presumably reflect real differences between the samples used (often in the detailed nature of low-occupancy water sites) and point up the need for caution in discussing the fine details of complicated chemical and crystal structures. 4.5.2 -Cyclodextrin as host in tunnel inclusion complexes In addition to the cage structures (only one example to date) there is a group of almost isomorphous tunnel structures (Table 4.19). The crystal structure of these tetragonal crystals deserves special comment, as has been emphasized by Steiner and Saenger (1998a) who discuss the overall packing and the rarity of space group P4212 especially among organic structures (0.006% of those listed in the Spring 1997 update of the Cambridge Structural Database). The crystal structure of {-CDn-propanol17H2O} (this is the ‘‘formula unit’’) has been described in detail by Ding, Steiner and Saenger (1991), who also compare results obtained for two different crystals of the n-propanol complex (SIBJAO, SIBJES). We describe the family of structures in overall terms, starting with the first four entries in Table 4.19. There are 6 formula units in the unit cell, with three crystallographically-independent -CD molecules stacked one above the other along [001] in Wyckoff positions (c) at 0, 1/2, z etc. (three different values of z). The asymmetric unit consists of three groups of two adjacent (linked) glucose rings, with each -CD molecule being obtained by the operation of the fourfold axis on the appropriate pair of glucose rings. Successive -CD molecules in a stack are hydrogen bonded in a sequence of head-to-head (A–B), tail-to-tail (B–C) and head-to-tail (C–A) interactions (A, B, C refer to succeeding molecules up the [001] axis in Fig. 4.19; this ABC notation is not standardized and we follow Steiner and Saenger (1998) in their Fig. 3, but not in their usage of ‘‘head’’ and ‘‘tail;’’ for our usage see Fig. 4.2 and footnote on p. 79)). The headto-head pair corresponds to the dimer found in some -CD and, particularly, in -CD tunnel inclusion complexes (see, for example, Fig. 4.16). The water and alcohol molecules are contained, in a disordered fashion which varies from crystal to crystal, both within the -CD tori, and in the columnar interstitial spaces between the -CD molecules (Fig. 4.19). Now consider what happens when -CD is crystallized from a solution of 12-crown-4, as

120


was first done by Vo¨gtle and Mu¨ller (1979). Crown ether complexes are obtained which are essentially isomorphous with the first group of four (Kamitori, Hirotsu and Higuchi, 1986). The 12-crown-4 (CREt) molecules are enclosed within the heads of the -CD molecules; thus the sequence along [001], shown in the left hand column of Fig. 4.20, has no enclosed CREt when the -CDs are juxtaposed tail-to-tail, single CREt molecules when the contact is tail-to-head, and a pair of CREts when the contact is head-to-head. The crown ether molecules are not disordered because they have intrinsic fourfold symmetry. Successive crown ether molecules in a stack are mutually rotated, with the angles between the pairs A/B, B/C and C/A 9.8, 11.2 and 21.0 (the zero sum is required to maintain translational symmetry). When crystallization is carried out in the presence of the salts LiSCN or KCl (Kamitori, Hirotsu and Higuchi, 1987), the pair of CREts contains the cation as {(CREt)Mþ(CREt)]} and this arrangement is also found with NaCl (Kamitori, Hirotsu and Higuchi, 1988); thus the crown ethers are present as neutral molecules and in (12-crown-4)2Mþ complexes. Rb does not form an analogous complex; NH4þ does not appear to have been tried. The anions are located, together with the waters, in the interstitial tunnels between the columns of -CD molecules. All these complexes, with their wide tunnels, lose solvent very easily.

Table 4.19. Crystal data (at room temperature) for some nearly isomorphous tetragonal tunnel inclusion complexes of -CD; space group P4212; Z ¼ 6. The compositions are given as {-CD[x(guest)nH2O} Guest and water content

Refcode; reference

a

c

V/FU

CH3OH C2H5OH n-C3H7OH17H2O

NUNRIX; SS98

23.808(4) 23.823(9) 23.840(5)

23.140(3) 23.227(5) 23.227(6)

2170(1) 2193(1) 2200(1)

23.805(2) 23.808(2) 23.75(2) 23.842(2) 23.816(4) 23.824(1) 23.77 23.80 23.82 23.37

23.196(7) 23.175(2) 22.92(2) 23.132(2) 23.072(3) 23.083(1) 23.15 23.22 23.21 23.91

2182(1) 2189 2155 2192 2181 2184 2181 2192 2195 2176

C6H5CH2OH C8H16O4*9H2O C8H16O4*1/3(LiSCN)7.7H2O C8H16O4*1/3(KCl)9H2O C8H16O4*1/3(NaCl)7.7H2O [1/3(C18H22N2)16.4H2O (cyclizine)# [2/3(C10H16N6S)14.7H2O (cimetidine) [2/3(C8H15N7O2S3)18.7H2O (famotidine) [2/3(C13H22N4O2S)16.3H2O (ranitidine) Polyethylene glycol Clofibric acid

CYDXPL, LS80; SIBJAO, SIBJES, DSS91 DOCYID; KHH86 FEJFIJ; KHH87 FEJFOP; KHH87 SAJNAS; KHH88 D99 D99 D99 D99 UWR00 CBM01 (not seen)

* 12-crown-4 (CREt) References: CBM01 – Caira, Bourne and Mvula, 2001 (not seen); D99 – Dodds, 1999DSS91 – Ding, Steiner and Saenger, 1991; KHH86 – Kamitori, Hirotsu and Higuchi, 1986; KHH87 – Kamitori, Hirotsu and Higuchi, 1987; KHH88 – Kamitori, Hirotsu and Higuchi, 1988; LS80 – Lindner and Saenger, 1980; SS98 – Steiner and Saenger, 1998; water content not given specifically, but it must be close to 17 H2O; UWR00 – Udachin, Wilson and Ripmeester, 2000.


N

N

121

CH3

Cyclizine

Scheme 4.5

Now consider the tetragonal 3 : 1 -CD cyclizine structure. The guest is incorporated into the space between the molecules of the head-to-head dimer, with one guest molecule for each repeat of three -CD molecules up the [001 axis. The cyclizine is disordered in a number of ways – firstly, the piperazine ring can point either up or down, secondly, the cyclizine molecule as a whole can take up one of the four rotationally equivalent positions. There is not enough room within a stack for incorporation of a second cyclizine. However, the tetragonal complexes of -CD with the smaller cimetidine, famotidine and ranitidine guests have 3 : 2 compositions (Table 4.19). Full structure analyses were not carried out for these complexes but we may surmise that two cimetidines, etc. replace one cyclizine within the head-to-head dimer, with appropriate disorder. The crystal structures of the cimetidine, famotidine, ranitidine and cyclizine complexes are of particular interest – their isomorphism with the other -CD complexes strongly

a b

Fig. 4.19. View of part of the crystal structure of the propanol tunnel inclusion complex of -CD, seen down [001]. The symmetry elements of the space group are shown. For simplicity only the A -CD molecules (defined in Fig. 4.20) are shown but this is adequate to show the arrangement of the stacks. The small circles show the water molecules within the internal tunnels of the -CD molecules, while the larger circles show the water molecules in the interstitial tunnels between the -CD molecules. (Reproduced from Ding, Steiner and Saenger, 1991.)

122


HEAD

CREt C

TAIL TAIL HEAD HEAD

B

B CREt C

CREt

C

A

TAIL HEAD

CREt C

TAIL TAIL

A

B

B CREt

HEAD HEAD TAIL

A

CREt 0

A

C

ab

Fig. 4.20. Schematic diagram of the arrangement of the -CD molecules in the stacks in the tunnel inclusion complexes; a slice parallel to the ð110Þ plane is shown. The heads and tails of the -CD molecules in the left-hand column are indicated. The designation ‘CREt’ shows that the 12-crown-4 molecules are included in the heads of the -CD molecule; an isolated ‘CREt’ indicates a neutral crown ether molecule, while a juxtaposed pair shows either a pair of neutral molecules or, when crystallization is carried out in the presence of LiSCN, NaCl or KCl, the cationic species {(CREt)2Mþ}. (Modified from Steiner and Saenger, 1998a.)

suggests that in fact there is only one packing mode for all -CD complexes. This could be tested by checking cell dimensions and symmetry for other -CD complexes already reported. Among these are the 15 different complexes, reported by Vo¨gtle and Mu¨ller (1979); a variety of crown ethers, cryptands, cryptates and coronates were guests, and compositions included 1 : 1, 2 : 1, 1 : 1 : 1, 1 : 1 : 2, 2 : 1 : 1 ratios. Also, crystalline 1 : 1 complexes of -CD with ferrocene, acetylferrocene and 1,1 0 -diacetylferrocene have been reported by Harada, Hu, Yamamoto and Takahashi (1988). Crystal data or structures, apart from those noted above, do not appear to have been reported. Steiner and Saenger (1998) have pointed out that the disorder in the tetragonal -CD complexes makes it difficult to determine the molecular dimensions of -CD with the same precision that has been achieved for - and -CD, nor could the interactions among the moieties be defined in detail. Thus caution is needed before drawing far-reaching conclusions from the limited information currently available. 4.5.3


Following the trend set earlier with chemically-modified - and -cyclodextrins, a number of permethylated -cyclodextrins have been prepared, and three different hydrates studied crystallographically (Table 4.20). Octakis(2,3,6-tri-O-methyl) -cyclodextrin is C72H128O40.

CYCLIC OLIGOSACCHARIDES AS CYCLODEXTRIN ANALOGS

123

Table 4.20. Crystal data for three permethylated -cyclodextrin hydrates. The compositions are given as {(C72H128O40).nH2O}. The last host is (C48H64O32) Composition

Refcode; reference

a

2.5H2O

GIWMAA; StSa98b XERSIW; AHSRS00 BEBJAT; AU99 ZUSYOB; YM96

17.730

4.5H2O 4.8H2O Octakis(3,6-anhydro)-CD14H2O

c

V/FU

Space group; Z

16.875

32.172

2271

P212121; 4

10.788

29.058

32.291

2531

P212121; 4

28.872

18.018 98.15 10.305 123.08

33.170

2209

P21; 8

15.406

1477

C2; 2

22.200

b/

References: AHSRS00b – Aree et al., 2000b AU99 – Aree, Uson et al., 1999; StSa98b – Steiner and Saenger, 1998b; YM96 – Yamamura et al., 1996.

4.6 Larger cyclodextrins Larger cyclodextrins ( for n ¼ 9 glucose units, " for n ¼ 10 and for n ¼ 11), were first isolated by French in 1957 (French, Pulley, Effendberger, Rougvie and Abdullah, 1965) and crystal structures have been reported for the (Fujiwara, Tanaka and Kobayashi, 1990; SIYKOA, 13.75 hydrate), " (Jacob, Gessler, Hoffmann, Sanbe, Koizumi, Smith, Takaha and Saenger, 1998; NOBBOV, 20.7 hydrate), (14; Jacob et al., 1998; NOBBUB, 27.3 hydrate) oligomers and for that with 26 glucose units (Gessler, Uson, Takaha, Krauss, Smith, Okada, Sheldrick and Saenger, 1998). Even larger cyclodextrins with 100 or more glucose units in the ring have been prepared (Takaha, Yanase, Takata, Okada and Smith, 1996). Whether inclusion complexes are formed by the larger cyclodextrins remains an open question at the time of writing (2003). The larger cyclodextrins do not have regular truncated-cone structures but are folded back upon themselves; thus guest inclusion seems unlikely. However, measurement of inclusion complexing capacity by capillary electrophoresis (Larsen, Ueda and Zimmerman, 1997) suggests that complexes may be formed by some guests. For example, the formation constant for the -CD–ibuprofen complex is 2600 M1, while corresponding values for the - and -CD ibuprofen complexes are 1013 and 225 M1 respectively.

4.7 Cyclic oligosaccharides as cyclodextrin analogs A series of novel CD analogs composed of alternating D- and L- rhamnopyranose (R) and mannopyranose (M) residues has been synthesized (Ashton, Cantrill, Gattuso, Menzer, Nepogodiev, Shipway, Stoddart and Williams, 1997), and we give some information about crystal structures that have been published (Table 4.21). 1-MM has a clathrate type structure while the other three molecules have tunnel arrangements. Although it has been mentioned that inclusion complexes are formed by


124

Table 4.21. Crystallographic information about cyclic oligosaccharides that may be cyclodextrin analogs n; CD analog

1R

2R

Crystal data

1-MM

3; -CD

CH2OH

CH2OH

2-RM

4; -CD

CH3

CH2OH

2-RR

4; -CD

CH3

CH3

3-RR

5; "-CD

CH3

CH3

C36H42 D18O309D2O; C2/c, Z ¼ 4 (molecules are centrosymmetric); ˚, a ¼ 28.005(2), b ¼ 9.807(2), c ¼ 20.853(2) A ˚ 3. NOHKEA ¼ 117.09(1) , V ¼ 5098.8(9) A (Ashton, Cantrill et al., 1997). 2{(C12H20O9)4}67H2O; P4, Z ¼ 1 (two independent molecules with four fold symmetry in the asymmetric unit); ˚, a ¼ b ¼ 24.200(5), c ¼ 7.918(3) A ˚ 3. TAHREZ V ¼ 4637(2) A (Ashton, Brown et al., 1996). C48H80O326Me2CO13H2O; C2/c, Z ¼ 4 (molecules have two fold axes); a ¼ 34.432(5), ˚ , ¼ 116.981) , b ¼ 7.986(2), c ¼ 31.910(3 A 3 ˚ V ¼ 7820(2) A . NOHKAW (Ashton, Cantrillet al., 1997). C60H100O406Me2CO10H2O; P1, Z ¼ 2 (two independent centrosymmetric molecules); ˚, a ¼ 15.130(6), b ¼ 19.077(5), c ¼ 20.017(5) A ¼ 72.74(2), ¼ 87.60(3), ¼ 88.60(3) , ˚ 3 NIHDIR V ¼ 5513(3) A (Gattuso, Menzer et al., 1997),

these hosts6, details have not yet been published. It does not seem too hazardous to predict that the crystallography of these inclusion complexes will prove as interesting as that of the cyclodextrins. References Alexander, J. M., Clark, J. L., Brett, T. J. and Stegowski, J. J. (2001). Proc. Nat. Acad, Sci., 99, 5115–5120. Alston, D. R., Slawin, A. M. Z., Stoddart, J. F. and Williams, D. J. (1985a). Angew. Chem. Int. Ed. Engl., 24, 786–788. Alston, D. R., Slawin, A. M. Z., Stoddart, J. F. and Williams, D. J. (1985b). J. Chem. Soc., Chem. Commun., pp. 1602–1604. Alston, D. R., Slawin, A. M. Z., Stoddart, J. F., Williams, D. J. and Zarycki, R. (1988). Angew. Chem. Int. Ed. Engl., 27, 1184–1185. Anibarro, M., Gessler, K., Uson, I., Sheldrick, G. M. and Saenger, W. (2001). Carbohydr. Res., 333, 251–256. 6 ‘‘Given the fact that these compounds share with the naturally-occurring CDs an ability to form inclusion complexes with other molecules, they could find themselves enjoying many biomedically-directed and materialsoriented applications in the future’’ (Gattuso, Menzer, Nepogodiev, Stoddart and Williams, 1997).

REFERENCES

125

Aree, T. and Chaichit, N. (2002). Carbohydr. Res., 337, 2487–2494. Aree, T. and Chaichit, N. (2003a). Carbohydr. Res., 338, 439–446. Aree, T. and Chaichit, N. (2003b). Carbohydr. Res., 338, 1581–1589 Aree, T., Hoier, H., Schulz, B., Reck, G. and Saenger, W. (2000a). Carbohydr. Res., 323, 245–253. Aree, T., Hoier, H., Schulz, B., Reck, G. and Saenger, W. (2000b). Carbohydr. Res., 328, 399–407. Aree, T., Hoier, H., Schulz, B., Reck, G. and Saenger, W. (2000c). Angew. Chem. Int. Ed. Engl., 39, 897–899. Aree, T., Jacob, J., Saenger, W. and Hoier, H. (1998). Carbohydr. Res., 307, 191–197. Aree, T., Saenger, W., Liebnitz, P. and Hoier, H. (1999). Carbohydr. Res., 315, 199–205. Aree, T., Schulz, B. and Reck, G. (2003). J. Incl. Phenom. Macrocyclic Chem., 47, 39–45. Aree, T., Uson, I., Schulz, B., Reck, G, Hoier, H., Sheldrick, G. M. and Saenger, W. (1999). J. Am. Chem. Soc., 121, 3321–3327. Armspach, D., Ashton, P. R., Moore, C. P., Spencer, N., Stoddart, J. F., Wear, T. J. and Williams, D. J. (1993). Angew. Chem. Int. Ed. Engl., 32, 854–858. Ashton, P. R., Brown, C. L., Menzer, S., Nepogodiev, S. A., Shipway, A. N., Stoddart, J. F. and Williams, D. J. (1996). Chem. Eur. J., 2, 580–591. Ashton, P. R., Cantrill, S. J., Gattuso, G., Menzer, S., Nepogodiev, S. A., Stoddart, J. F. and Williams, D. J. (1997). Chem. Eur. J., 3, 1299–1314. Bender, M. L. and Komiyama, M. (1978). Cyclodextrin Chemistry. Springer Verlag, New York. Betzel, C., Hingerty, B., Noltemeyer, M., Weber, G., Saenger, W. and Hamilton, J. A. (1983). J. Incl. Phenom., 1, 181–191. Betzel, C., Saenger, W., Hingerty, B. E. and Brown, G. M. (1984). J. Am. Chem. Soc., 106, 7545–7557. Bobek, M. M., Giester,G., Ka¨hlig, H. and Brinker, U. H. (2000). Tetr. Lett., 41, 5663–5667. Braga, S. S., Aree, T., Imamura, K., Vertut, P., Boal-Palheiros, I., Saenger, W. and Teixeira-Diaz, J. J. C. (2002). J. Incl. Phenom. Macrocyclic Chem., 43, 115–125. Brett, T. J. and Stezowski, J. J. (2000). Chem. Commun., pp. 551–552. Brett, T. J., Alexander, J. M. and Stezowski, J. J. (2000). J. Chem. Soc., Perkin 2, pp. 1095–1103. Brett, T. J., Alexander, J. M., Clark, J. L., Ross, C. R. II, Harbison, G. S. and Stezowski, J. J. (1999). Chem. Commun., pp. 1275–1276. Brett, T. J., Liu, S., Coppens, P. and Stezowski, J. J. (1999). Chem. Commun., pp. 551–552. Briggner, L.-E. and Wadso¨, I. (1990). J. Chem. Thermodyn., 22, 1067–1074. Brown, G. R. (1997). M.Sc. Thesis, Department of Chemistry, University of Cape Town. Brown, G. R., Caira, M. R., Nassimbeni, L. R. and Oudtshoorn, B. van. (1996). J. Incl. Phenom. Molec. Recogn. Chem., 26, 281–294. Caira, M. R. (2002). Rev. Roum. Chim., 46, 3761–386. Caira, M. R. and Dodds, D. R. (1999). J. Incl. Phenom. Molec. Recogn. Chem., 34, 19–29. Caira, M. R. and Dodds, D. R. (2000). J. Incl. Phenom. Molec. Recogn. Chem., 38, 75–84. Caira, M. R., Dodds, D. R. and Nassimbeni, L. R. (2001). Supramol. Chem., 13, 61–70. Caira, M. R., Bourne, S. A. and Mvula, E. (1999). J. Therm. Anal. and Calorimetry, 56, 1329–1334. Caira, M. R., Bourne, S. A. and Mvula, E. (2001). Biol. Zh. Armen., 53, 148–158. Caira, M. R., Griffith, V. J. and Nassimbeni, L. R. (1998). J. Incl. Phenom. Molec. Recogn. Chem., 32, 461–476. Caira, M. R., Griffith, V. J., Nassimbeni, L. R. and Oudtshoorn, B. van (1994a). J. Chem. Soc., Chem. Comm., pp. 1061–1062. Caira, M. R., Griffith, V. J., Nassimbeni, L. R. and Oudtshoorn, B. van (1994b). J. Chem. Soc., Perkin 2, pp. 2071–2072. Caira, M. R., Griffith, V. J., Nassimbeni, L. R. and Oudtshoorn, B. van (1994c). J. Incl. Phenom. Molec. Recogn. Chem., 17, 187–201. Caira, M. R., Griffith, V. J., Nassimbeni, L. R. and Oudtshoorn, B. van (1994/1995). J. Incl. Phenom. Molec. Recogn. Chem., 20, 277–290.

126


Caira, M. R., Griffith, V. J., Nassimbeni, L. R. and Oudtshoorn, B. van (1996). Supramol. Chem., 7, 119–124. Cardinael, P., Poulon, V., Perez, G., Coquard, G. and Toupet, L. (2001). J. Incl. Phenom. Molec. Recogn. Chem., 39, 159–167. Chacko, K. K. and Saenger, W. (1981). J. Am. Chem. Soc., 103, 1708–1715. Charpin, P., Nicolis, I., Villian, P., Rango, C. de and Coleman, A. W. (1991). Acta Cryst., C47, 1829–1833. Chen, Y., Luo, L., Chen, H., Hu, C., Chen, J. and Zheng, P. (2000). Bull. Chem. Soc. Jpn., 73, 1375–1378. Clark, J. L. and Stezowski, J. J. (2001). J. Am. Chem. Soc., 123, 9880–9888. Clark, J. L., Booth, B. R. and Stezowski, J. J. (2001). J. Am. Chem. Soc., 123, 9889–9895. Cramer, F. (1954). Einschlussverbindungen. Springer Verlag, Berlin. Cramer, F. (1987). In Cyclodextrins and Their Industrial Uses, D. Ducheˆne (editor), Editions de Sante´, Paris, pp. 11–20. Cramer, F. and Henglein, F. M. (1957). Chem. Ber., 90, 2572–2575. Crisma, M., Fomasier, R. and Marcuzzi, F. (2001). Carbohydr. Res., 333, 145–151. Czugler, M., Eckle, E. and Stezowski, J. J. (1981). J. Chem. Soc. Chem. Comm., pp. 1291–1292. Diaz, D., Vargas-Baca, I. and Gracia-Mora, J. (1994). J. Chem. Educ., 71, 708–714. Ding, J., Steiner, T. and Saenger, W. (1991). Acta Cryst., B47, 731–738. Dimitrius, M., Terzis, A., Coleman, A. W. and de Rango, C. (1996). Carbohydr. Res., 282, 125–135. Dodds, D. R. (1999). Ph.D. Thesis, Department of Chemistry, University of Cape Town. Ducheˆne, D. (1987). (Ed.). Cyclodextrins and Their Industrial Uses, Editions de Sante´, Paris. Durier, V., Buisson, G., Dueé, E., Driguez, H. and Taravel, F.-R. (1993). J. Chem. Soc., Perkin 2, pp. 1931–1935. Eliadou, K., Yannakopoulou, K., Rontoyianni, A., Mavridis, I. M. (1999). J. Org. Chem., 64, 6217–6226. Evans, R. C. (1966). An introduction to crystal chemistry, 2nd edition. Cambridge University Press. French, D., Pulley, A. O., Effendberger, J., Rougvie, M. and Abdullah, M. (1965). Arch. Biochem. Biophys., 111, 153–160. Fujita, K., Matsunaga, A., Imoto, T., Hirotsu, K., Kamitori, S. and Higuchi, T. (1987). J. Am. Chem Soc., 107, 1790–1791. Fujiwara, T., Tanaka, N. and Kobayashi, S. (1990). Chem. Letts, pp. 739–742. Fujiwara, T., Tomita, K., Marseigne, I. and Vicens, J. (1988). Mol. Cryst. Liq. Cryst. Inc. Nonlin. Opt., 156, 393–404. Fujiwara, T., Yamazaki, M., Tomizu, Y., Tokuoka, R., Tomita, K., Matsuo, T., Suga, H. and Saenger, W. (1983). Bull. Chem. Soc. Jpn., 56, 181–187. Gattuso, G., Nepogodiev, S. A. and Stoddart, J. F. (1998). Chem. Revs., 98, 1919–1958. Gattuso, G., Menzer, S., Nepogodiev, S. A., Stoddart, J. F. and Williams, D. J. (1997). Angew. Chem. Int. Ed. Engl., 36, 1451–1454. Gessler, K., Steiner, T., Koellner, G. and Saenger, W. (1993). Carbohydr. Res., 249, 327–344. Gessler, K., Uson, I., Takaha, T., Krauss, N., Smith, S. M., Okada, S., Sheldrick, G. M. and Saenger, W. (1998). Proc. Nat. Acad. Sci., 37, 605–609. Giastas, P., Yannakopoulou, K. and Mavridis, I. M. (2003). Acta Cryst., B59, 287–299. Hamilton, J. A. (1985). Carbohydr. Res., 142, 21–37. Hamilton, J. A. and Chen, L. (1988a). J. Am. Chem. Soc., 110, 4379–4391. Hamilton, J. A. and Chen, L. (1988b). J. Am. Chem. Soc., 110, 5833–5841. Hamilton, J. A. and Sabesan, M. N. (1982a). Carbohydr. Res., 102, 31–46. Hamilton, J. A. and Sabesan, M. N. (1982b). Acta Cryst., B38, 3063–3069. Hamilton, J. A., Sabesan, M. N. and Steinrauf, L. K. (1981). Carbohydr. Res., 89, 33–53. Hamilton, J. A., Steinrauf, L. K. and Van Etten, R. L. (1968). Acta Cryst., B24, 1560–1562.

REFERENCES

127

Hanabata, H., Matsuo, T. and Suga, H. (1987). J. Incl. Phenom., 5, 325–333. Harada, A., Hu, Y., Yamamoto, S. and Takahishi, S. (1988). J. Chem. Soc., Dalton, pp. 729–732. Harata, K. (1975). Bull. Chem. Soc. Jpn., 48, 2409–2413. Harata, K. (1976a). Carbohydr. Res., 48, 265–270. Harata, K. (1976b). Bull. Chem. Soc. Jpn., 49, 1493–1501. Harata, K. (1976c). Bull. Chem. Soc. Jpn., 49, 2066–2072. Harata, K. (1977a). Bull. Chem. Soc. Jpn., 50, 1259–1266. Harata, K. (1977b). Bull. Chem. Soc. Jpn., 50, 1416–1423. Harata, K. (1978). Bull. Chem. Soc. Jpn., 51, 1644–1648. Harata, K. (1979). Bull. Chem. Soc. Jpn., 52, 2451–2459. Harata, K. (1980). Bull. Chem. Soc. Jpn., 53, 2782–2786. Harata, K. (1982a). Bull. Chem. Soc. Jpn., 55, 1367–1371. Harata, K. (1982b). Bull. Chem. Soc. Jpn., 55, 2315–2321. Harata, K. (1984). Bull. Chem. Soc. Jpn., 57, 2596–2599. Harata, K. (1987). Bull. Chem. Soc. Jpn., 60, 2763–2767. Harata, K. (1988). Bull. Chem. Soc. Jpn., 61, 1939–1944. Harata, K. (1989). Carbohydr. Res., 192, 33–42. Harata, K. (1990a). J. Chem. Soc., Perkin 2, pp. 799–804. Harata, K. (1990b). Bull. Chem. Soc. Jpn., 63, 2481–2486. Harata, K. (1991). ‘‘Recent advances in the x-ray analysis of cyclodextrin complexes,’’ in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Vol. 5 (Inorganic and Physical Aspects of Inclusion), Oxford University Press, Oxford etc., pp. 311–344. Harata, K, (1993). Chem. Commun., pp. 546–547. Harata, K. (1995). Supramol. Chem., 5, 231–236. Harata, K. (1996). ‘‘Crystallographic studies,’’ in Cyclodextrins, Volume 3 of Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, pp. 279–304. Harata, K. (1998). Chem. Revs., 98, 1803–1827. Harata, K. (1999). Chem. Commun., pp. 191–192. Harata, K., Hirayama, F., Uekama, K. and Tsoucaris, G. (1988). Chem. Lett., pp. 1585–1588. Harata, K., Hirayama, F., Arima, H., Uekama, K. and Miyaji, T. (1992). J. Chem. Soc., Perkin 2, pp. 1159–1166. Harata, K., Kawano, K., Fukunaga, K. and Ohtani, Y. (1983). Chem. Pharm. Bull., 31, 1428–1430. Harata, K., Nagano, Y., Ikeda, H., Ikeda, T., Ueno, A. and Toda, F. (1996). Chem. Commun., pp. 2347–2348. Harata, K., Uedaira, H. and Tanaka, J. (1978). Bull. Chem. Soc. Jpn., 51, 1627–1634. Harata, K., Uekama, K., Imai, T., Hirayama, F. and Otagiri, M. (1988). J. Incl. Phenom., 6, 443–460. Harata, K., Uekama, K., Otagiri, M. and Hirayama, F. (1982a). Bull. Chem. Soc. Jpn., 55, 407–410. Harata, K., Uekama, K. Otagiri, M. and Hirayama, F. (1982b). Bull. Chem. Soc. Jpn., 55, 3904– 3910. Harata, K., Uekama, K., Otagiri, M. and Hirayama, F. (1983a). Nippon Kagaku Kaishi, pp. 173–180. Harata, K., Uekama, K., Otagiri, M. and Hirayama, F. (1983b). Bull. Chem. Soc. Jpn., 56, 1732–1736 Harata, K., Uekama, K. Otagiri, M. and Hirayama, F. (1984). J. Incl. Phenom., 1, 279–293. Harata, K., Uekama, K., Otagiri, M. and Hirayama, F. (1987). Bull. Chem. Soc. Jpn., 60, 497–502. Harata, K., Uekama, K., Otagiri, M., Hirayama, F. and Ogino, H. (1981). Bull. Chem. Soc. Jpn., 54, 1954–1959. Harata, K., Uekama, K., Otagiri, M., Hirayama, F. and Ohtani, Y. (1985). Bull. Chem. Soc. Jpn., 58, 1234–1238.

128


Harata, K., Uekama, K. Otagiri, M., Hirayama, F. and Sugiyama, Y. (1982). Bull. Chem. Soc. Jpn., 55, 3386–3389. Harding, M. M., Mclennan, J. M. and Paton, R. M. (1978). Nature, Lond., 274, 621–623. Herbstein, F. H. and Marsh, R. E. (1998). Acta Cryst., B54, 677–686. Hingerty, B. and Saenger, W. (1976). J. Am. Chem. Soc., 98, 3357–3365. Hingerty, B., Betzel, C. and Saenger, W. (1984). American Crystallographic Association Abstracts (Winter), p. 37. Hirotsu, K., Higuchi, T., Fujita, K., Ueda, T., Shinoda, A., Imoto, T. and Tabushi, I. (1982). J. Org. Chem., 47, 1143–1144. Hursthouse, M. B., Smith, C. Z., Thornton-Pett, M. and Utley, J. H. P. (1982). J. Chem. Soc. Chem. Comm., pp. 881–882. Hybl, A., Rundle, R. E. and Williams, D. E. (1965). J. Am. Chem. Soc., 87, 2779–2788. Jacob, J., Gessler, K., Hoffmann, D., Sanbe, H., Koizumi, K., Smith, S. M., Takaha, T. and Saenger, W. (1998). Angew. Chem., Int. Ed. Engl., 37, 606–609. Jogun, K. H. and Stezowski, J. J. (1979). Nature, Lond., 278, 667–668. Jogun, K. H., Mclennan, J. M. and Stezowski, J. J. (1979). Abstracts of 5th European Crystallographic Meeting, Copenhagen, p. 34. Kamitori, S., Hirotsu, K. and Higuchi, T. (1986). J. Chem. Soc. Chem. Comm., pp. 690–691. Kamitori, S., Hirotsu, K. and Higuchi, T. (1987). J. Am. Chem. Soc., 109, 2409–2414. Kamitori, S., Hirotsu, K. and Higuchi, T. (1988). Bull. Chem. Soc. Jpn., 61, 3825–3830. Kamitori, S., Hirotsu, K., Higuchi, T., Fujita, K., Yamamura, H., Imoto, T. and Tabushi, I. (1987). J. Chem. Soc. Perkin Trans. II, pp. 7–14. Kamitori, S., Maruoka, S., Kondo, S. and Okuyama, K. (1998). Carbohydr. Res., 312, 177–181. Kamitori, S., Matsuzaka, O., Kondo, S., Maruoka, S., Okuyama, K., Noguchi, K., Okada, M. and Harada, A. (2000). Macromolecules, 33, 1500–1502. Kamitori, S., Toyama, T. and Matsuzaka, O. (2001). Carbohydr. Res., 332, 235–240. Klar, B., Hingerty, B. and Saenger, W. (1980). Acta Cryst., B36, 1154–1165. Klingert, B. and Rihs, G. (1990). Organometallics, 9, 1135–1141. Klingert, B. and Rihs, G. (1991a). J. Incl. Phenom., 10, 255–265. Klingert, B. and Rihs, G. (1991b). J. Chem. Soc., Dalton, pp. 2749–2760. Kokkinou, A., Makedonopoulou, S. and Mavridis, I. M. (2000). Carbohydr. Res., 328, 135–140. Kokkinou, A., Yannakopoulou, K., Mavridis, I. M. and Mentzafos, D. (2001). Carbohydr. Res., 332, 85–94. Koshland, D. E. (1970). ‘‘The molecular basis for enzyme regulation.’’ In The Enzymes, Third Edition, edited by P. D. Boyer, Vol. 1, pp. 342–396. Academic Press, New York and London . Larsen, K. L., Ueda, H. and Zimmerman, W. (1997). 8th European Congress on Biotechnology, Budapest, 17–21 August. Lautsch, W., Weichert, R. and Lehmann, H. (1954). Kolloid-Z., 135, 134–136. Le Bas, G. and Doucel, J. (1997). Acta Cryst., B53, 673–679. Le Bas, G. and Mason, S. A. (1994). Acta Cryst., B50, 717–724. Le Bas, G., Rango, C. de, Ryanek, N. and Tsoucaris, G. (1984). J. Incl. Phenom., 2, 861–867. Le Bas, G., Rysanek, N., Royer, F. and Tsoucaris, G. (1988). Mol. Cryst. Liq. Cryst., 161, 363–372. Lindner, K. and Saenger, W. (1980). Biochem. Biophys. Res. Comm., 92, 933–938. Lindner, K. and Saenger, W. (1982a). Carbohydr. Res., 99, 103–115. Lindner, K. and Saenger, W. (1982b). Carbohydr. Res., 107, 7–16. Lindner, K. and Saenger, W. (1982c). Acta Cryst., B38, 203–210. Luo, L. B., Chen, H. L., Tang, W. X., Zhang, Z. Y. and Mak, T. C. W. (1996). J. Chem. Soc., Dalton, pp. 4425–4430. Makedonopoulou, S. and Mavridis, I. M. (2000). Acta Cryst., B56, 322–331. Makedonopoulou, S. and Mavridis, I. M. (2001). Carbohydr. Res., 335, 213–220.

REFERENCES

129

Makedonopoulou, S., Tulinsky, A. and Mavridis, I. M. (1999). Supramol. Chem., 11, 73–80. Makedonopoulou, S., Papaioannou, J., Argyroglou, I. and Mavridis, I. M. (2000). J. Incl. Phenom., 36, 191–215. Makedonopoulou, S., Yannakopoulou, K., Mentzafos, D., Lamzin, V., Popov, A. and Mavridis, I. M. (2001). Acta Cryst., B57, 399–409. Manor, P. C. and Saenger, W. (1974). J. Am. Chem. Soc., 96, 3630–3639. Mclennan, J. M. and Stezowski, J. J. (1980). Biochem. Biophys. Res. Comm., 92, 926–932. McMullan, R. K., Saenger, W., Fayos, J. and Mootz, D. (1973a). Carbohydr. Res., 31, 37–46. McMullan, R. K., Saenger, W., Fayos, J. and Mootz, D. (1973b). Carbohydr. Res., 31, 211–227. Marsh, R. E., Kapon, M. Hu, S. and Herbstein, F. H. (2002). Acta Cryst., B58, 62–77. Mavridis, I. M. and Hadjoudis, E. (1992). Carbohydr. Res., 229, 1–15. Mavridis, I. M., Hadjoudis, E. and Tsoucaris, G. (1991). Carbohydr. Res., 220, 11–21. Meister, G., Stoeckli-Evans, H. and Su¨ss-Fink, G. (1993). J. Organometall. Chem., 453, 249–253. Mele, A. and Malpezzi, L. (2000). J. Amer. Soc. Mass Spectrom., 11, 228–238. Mentzafos, D., Mavridis, I. M. and Hursthouse, M. B. (1996). Acta Cryst., C52, 1220–1223. Mentzafos, D., Mavridis, I. M. and Schenk, H. (1994). Carbohydr. Res., 253, 39–50. Mentzafos, D., Mavridis, I. M. and Yannakopoulou, K. (1999). J. Incl. Phenom., 33, 321–330. Mentzafos, D., Mavridis, I. M., Le Bas, G. and Tsoucaris, G. (1991). Acta Cryst., B47, 746–757. Motherwell, S. (1978). ‘‘PLUTO, Program for Plotting Molecular and Crystal Structures’’, University Chemical Laboratories, Cambridge, England. Muraoka, S., Matsuzaka, O., Kamitori, S. and Okuyama, K. (1999). Carbohydr. Res., 320, 261–266. Nakagawa, T., Immel, S., Lichtenthaler, F. W. and Lindner, H. J. (2000). Carbohydr. Res., 324, 141–146. Nakanishi, I., Arai, M., Fujiwara, T. and Tomita, K. (1984). J. Incl. Phenom., 2, 689–699. Nakanishi, I., Fujiwara, T. and Tomita, K. (1984). Acta Cryst., A40, C78. Nepogodiev, S. A. and Stoddart, J. F. (1998). Chem. Revs., 98, 1959–1976. Nicolis, I., Coleman, A. W., Charpin, P. and Rango, C. de, (1995). Angew. Chem., Int. Ed. Engl., 34, 2381–2383. Nicolis, I., Coleman, A. W., Charpin, P. and Rango, C. de, (1996). Acta Cryst., B52, 122–130. Nicolis, I., Villain, F., Coleman, A. W., and Rango, C. de, (1995). Supramol. Chem., 3, 251–259. Nishioka, F., Nakanishi, I., Fujiwara, T. and Tomita, K.-I. (1984). J. Incl. Phenom., 2, 701–714. Noltemeyer, M. and Saenger, W. (1980). J. Am. Chem. Soc., 102, 2710–2722. Odagaki, Y., Hirotsu, K., Harada, A. and Takahashi, S. (1990). J. Chem. Soc., Perkin I, pp. 1230–1231. Pohlmann, H., Gdaniec, M., Eckle, E., Geiger, G. and Stezowski, J. J. (1984). Acta Cryst., A40, C276. Pop, M.M., Goubitz, K., Borodi, G., Bogdan, M., De Ridder, D. J. A., Peschar, R. and Schenk, H. (2002). Acta Cryst., B58, 1036–1043. Puliti, R., Mattia, C. A. and Paduano, L. (1998). Carbohydr. Res., 310, 1–8. Rango, C. de, Charpin, C., Navaza, J., Keller, N., Nicolis, I., Villain, F. and Coleman, A. W. (1992). J. Am. Chem. Soc., 114, 5475–5476. Rekharsky, M. V. and Inoue, Y. (1998). Chem. Rev., 98, 1875–1917. Rontoyianni, A. and Mavridis, I. M. (1994). J. Incl. Phenom. Mol. Recogn. Chem., 18, 211–227. Rontoyianni, A. and Mavridis, I. M. (1996b). Acta Cryst., C52, 2277–2281. Rontoyianni, A. and Mavridis, I. M. (1999). Supramol. Chem., 10, 213–218. Rontoyianni, A., Mavridis, I. M., Hadjoudis, E. and Duisenberg, A. J. M. (1994). Carbohydr. Res., 252, 19–32. Rontoyianni, A., Mavridis, I. M., Israel, R. and Beurskens, G. (1998). J. Incl. Phenom. Mol. Recogn. Chem., 32, 415–428. Rysanek, N., Le Bas, G., Villain, F. and Tsoucaris, G. (1992). Acta Cryst., C48, 1466–1471. Rysanek, N., Le Bas, G., Villain, F. and Tsoucaris, G. (1996). Acta Cryst., C52, 2932–2936.

130


Saenger, W. (1980). Angew. Chem. Int. Ed. Engl., 19, 344–362. Saenger, W. (1984). ‘‘Structural aspects of cyclodextrins and their inclusion complexes’’ in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Vol. 2, pp. 231–259, Academic Press, London etc. Saenger, W. (1985). Isr. J. Chem., 25, 43–50. Saenger, W., Beyer, K. and Manor, P. C. (1976). Acta Cryst., B32, 120–128. Saenger, W., Jacob, J., Gessler, K., Steiner, T., Hoffmann, D., Sanbe, H., Koizumi, K., Smith, S. M. and Takaha, T. (1998). Chem. Rev., 98, 1787–1802. Saenger, W., McMullan, R. K., Fayos, J. and Mootz, D. (1974). Acta Cryst., B30, 2019–2028. Saenger, W. and Noltemeyer, M. (1976). Chem. Ber., 109, 503–517. Saenger, W., Noltemeyer, M., Manor, P. C., Hingerty, B. and Klar, B. (1976). Bioorg. Chem., 5, 187–195. Saenger, W. and Steiner, T. (1998). Acta Cryst., A54, 798–805. Sanchez-Ruiz, X., Alvarez-Larena, A., Jaime, C., Piniella, J. F. Redondo, J., Virgili, A., SanchezFerrando, F., Germain, G. and Baert, F. (1999).Supramol. Chem., 10, 219–223. Schardinger, F. (1904). Wien Klin. Wochenschi, 17, 207. Selkti, M., Navazza, A., Villain, F., Charpin, P. and De Rango, C. (1997). J. Incl. Phenom. Molec. Recogn. Chem., 27, 1–12. Senti, F. R. and Erlander, S. R. (1964). ‘‘Carbohydrates’’ (Chapter 9) in Non-Stoichiometric Compounds, edited by L. Mandelcorn, Academic Press, London and New York. Shibakami, M. and Sekiya, A. (1992). J. Chem. Soc., Chem. Comm., pp. 1742–1743. Shibakami, M. and Sekiya, A. (1994). Carbohydr. Res., 260, 169–179. Sicard-Roselli, C., Perly, B. and Le Bas, G. (2001). J. Incl. Phenom. Molec. Recogn. Chem., 39, 333–337. Steiner, T. (1990). Ph. D. Thesis, Freie Univ. Berlin, Germany. Steiner, T., Hinrichs, W., Saenger, W. and Hoyer, G-A. (1989). Carbohydr. Res., 192, 43–49. Steiner, T., Hirayama, K. and Saenger, W. (1996). Carbohydr. Res., 296, 69–82. Steiner, T. and Gessler, K. (1994). Carbohydr. Res., 260, 27–38. Steiner, T. and Koellner, G. (1994). J. Am. Chem. Soc., 116, 5122–5128. Steiner, T., Koellner, G. and Saenger, W. (1992). Carbohydr. Res., 228, 321–332. Steiner, T., Koellner, G., Ali, S., Zakim, D. and Saenger, W. (1992). Biochem. Biophys. Res. Comm., 188, 1060–1066. Steiner, T., Koellner, G., Gessler, K. and Saenger, W. (1995). J. Chem. Soc., Chem. Comm., pp. 511–512. Steiner, T., Mason, S. A. and Saenger, W. (1989). Z. Krist., 186, 281–282. Steiner, T., Mason, S. A. and Saenger, W. (1991). J. Am. Chem. Soc., 113, 5676–5687. Steiner, T. and Saenger, W. (1994b). Carbohydr. Lett., 1, 143–150. Steiner, T. and Saenger, W. (1995a). Carbohydr. Res., 275, 73–82. Steiner, T. and Saenger, W. (1995b). Chem. Commun., pp. 2087–2088. Steiner, T. and Saenger, W. (1998a). Acta Cryst., B54, 450–455. Steiner, T. and Saenger, W. (1998b). Angew. Chem. Int. Ed. Engl., 37, 3404–3407. Steiner, T. and Saenger, W. (1998c). J. Chem. Soc., Perkin 2, pp. 371–377. Stezowski, J. J. (1985). Trans. Am. Cryst. AsSoc., 20, 73–82. Stezowski, J. J., Czugler, M. and Eckle, E. (1981). Proc. I. S. Cyclodextrins, p. 151. Stezowski, J. J., Jogun, K. H., Eckle, E. and Bartels, K. (1978). Nature, Lond., 274, 617–619. Stezowski, J. J., Parker, W., Hilgenkamp, S. and Gdenic, M. (2001). J. Am. Chem. Soc., 123, 3919–3926. Stoddart, J. F. and Zarycki, R. (1989). Cyclodextrins – Monographs in Supramolecular Chemistry. Royal Society of Chemistry, London. Storsberg, J., Ritter, H., Pielartzik, H. and Groenendaal, L. (2000). Adv. Mater, 12, 567–569.

REFERENCES

131

Szejtli, J. 1982). Cyclodextrins and Their Inclusion Compounds. Akade´miai Kiado´, Budapest. Szejtli, J. (1988). Cyclodextrin Technology, pp. 450. Kluwer Academic Publishers, Dordrecht. Szetjli, J. (1998). Chem. Rev., 98, 1743–1753. Szetjli, J. and Budai, Zs. (1977). Acta Chim. Acad. Sci. Hung., 94, 383–390. Szetjli, J. and Osa, T. (1996). Editors, Cyclodextrins, Volume 3 of Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford. Tabushi, I. (1982). Accts. Chem. Res., 15, 66–72. Takaha, T., Yanase, M., Takata, S., Okada, S. and Smith, S. M. (1996). J. Biol. Chem., 271, 2902–2908. Tanaka, Y., Sakuraba, H. and Nakanishi, H. (1990). J. Org. Chem., 55, 564–567. Tokuoka, R., Fujiwara, T. and Tomita, K. (1981). Acta Cryst., B37, 1158–1160. Tokuoka, R., Abe, M. Fujiwara, T., Tomita, K. and Saenger, W. (1980). Chem. Letts., pp. 491–494. Tsoucaris, G., Le Bas, G., Rysanek, N. and Villain, F. (1987). J. Incl. Phenom., 5, 77–84. Udachin, K. A. and Ripmeester, J. A. (1998). J. Am. Chem. Soc., 120, 1080–1081. Udachin, K. A., Wilson, L. D. and Ripmeester, J. A. (2000). J. Am. Chem. Soc., 122, 12375–12376. Uekama, K. and Irie, T. (1987). In Cyclodextrins and their Industrial Uses, Duchene, D., Ed.; Editions de Sante, Paris, pp. 395–439. Uekama, K., Hirayama, F., Imai, T., Otogiri, M. and Harata, K. (1983). Chem. Pharm. Bull., 31, 3363–3365. Uekama, K., Imai, T., Hirayama, F., Otogiri, M. and Harata, K. (1984). Chem. Pharm. Bull., 32, 1662–1664. Vicens, J., Fujiwara, T. and Tomita, K. (1988). J. Incl. Phenom., 6, 577–581. Villiers, A. (1891). Compt. Rend. Acad. Sci. Paris, 112, 536–538. Vo¨gtle, F. and Mu¨ller, W. M. (1979). Angew. Chem. Int. Ed. Engl., 18, 623–624. Yamamura, H., Masuda, H., Kawase, Y., Kawai, m., Butsugan, Y. and Einaga, H. (1996). Chem. Commun., pp. 1069–1070. Yannakopoulou, K., Ripmeester, J. A. and Mavridis, I. M. (2002). J. Chem. Soc., Perkin 2, pp. 1639–1644. Zabel, V., Saenger, W. and Mason, S. A. (1986). J. Am. Chem. Soc., 108, 3664–3673. Zhukhlistova, N. E., Tischenko, G. N., Kuranova, I. P., Vainshtein, B. K., Mattson, P. and Korpella, T. (1996). Kristallografiya, 41, 97–108.


Chapter 5 Crystal chemistry of some DNA oligonucleotides and their complexes

The interaction between DNA and drugs is of great importance in molecular biology and medicinal chemistry. Drugs that target nucleic acids have wide application in nucleic acid recognition, regulation of biological processes and the development of therapeutic agents against cancers and virus-related diseases. Deng, Pan and Sundaralingam, 2003 . . . and, to crown all, Kinnaird and I had to conduct Sheridan down a damned corkscrew staircase, which had certainly been constructed before the discovery of fermented liquors, and to which no legs, however crooked, could possibly accommodate themselves. Lord Byron 31 October, 1815

Summary: We start with brief introductions to the fundamentals of oligonucleotide structure and to some current aspects of crystal chemistry common to cyclodextrin complexes, tetraphenylporphyrin-metal complexes and oligonucleotides and their complexes. This is followed by an outline of the special features of X-ray crystal structure analysis as applied to oligonucleotides, including the limitations on the resolution attainable. The implicit assumption usually made that native oligonucleotides are unary (single component) phases while their complexes (intercalation and minor groove binders are the two principal types) are binary can be a useful working hypothesis but neglects the important effects of solvent and ion content. In contrast to most of the published work, where priority has been given to features of biological importance, here we stress the crystal chemical aspects of hexamer, octamer, decamer and dodecamer oligonucleotide crystals, covering both native and complexed modifications. Many of the crystal structures are, group wise, isomorphous (same crystal structures despite differences in chemical nature). Differences in base-pair sequence sometimes break the isomorphism but usually not. Other different crystal types are often found in addition to the major isomorphous families. An important problem is that the chemical compositions (solvent content, presence and number of ions) are often not known so it can be difficult to decide whether particular clusters of related crystals are polymorphs (different crystal structures although the chemical compositions are the same) or have different compositions. Comparison of unit cell volumes is a useful but limited tool in making such decisions. Attempts are made to classify particular crystalline complexes in phase rule terms as primary solid solutions or phase rule compounds. There are resemblances in behavior and structural features to the cyclodextrins.

5.1 Introduction 5.2 Fundamentals of oligonucleotide structure 5.2.1 General aspects 5.2.2 Single crystal x-ray diffraction studies of oligonucleotides

134 136 136 140

134

CRYSTAL CHE MISTRY OF SOME DNA OLIGONUCLEOTIDES

5.3

Crystal chemistry of oligonucleotides and oligonucleotide-guest structures 5.3.1 Polymorphism, isomorphism, and heteromorphism 5.3.2 Phase rule relationships 5.3.3 Applications of these concepts 5.4 Intercalated hexanucleotide-drug complexes with B-DNA structures 5.4.1 The anthracycline drugs 5.4.2 Nogalamycin and derivatives 5.4.3 The 9-aminoacridine drugs 5.4.4 Native hexanucleotides and comparison of crystal structures 5.5 Isomorphism and polymorphism of A-DNA octanucleotides and the binding of spermine 5.6 Minor groove binders 5.6.1 Drug molecules that enter the minor groove 5.6.2 Decameric oligonucleotides 5.6.3 Polymorphs or intermediate phases? An example from the decanucleotides 5.6.4 Dodecameric oligonucleotides 5.7 General survey of the crystal chemistry of oligonucleotide and oligonucleotide-guest structures References

5.1

142 142 143 144 145 145 151 154 156 158 167 167 171 178 183 187 189

Introduction

The double-helix structure of deoxyribose nucleic acid (DNA), first reported by Watson and Crick in 1952, has attracted more attention and had more influence on the development of modern molecular biology than any other structure determination. The genetic material is composed of DNA, and the complementary nature of the two sets of base pairs in the two strands of the double helix provide a natural way of transmitting genetic information. DNA itself has only been obtained in the form of well oriented fibres up to some microns in length and so the amount of detail that can be extracted from structure analyses is severely limited. Thus a great advance was made when short sequences of DNA bases (synthetic oligonucleotides, also referred to as ‘‘DNA oligomers’’) were first synthesized in the 1970s and crystallized, thus allowing the execution of crystal structure analyses on fragments of relatively small size. This has shown that the original model of DNA derived from fiber diffraction is, not unexpectedly, somewhat oversimplified and has allowed the acquisition of much more structural detail (Neidle, Schneider and Berman, 2003). Our interest here is in the structures of the molecular complexes formed between drug molecules of various kinds and various DNA oligomers (hexamers, octamers, decamers and dodecamers) and specifically in their crystal chemistry. Thus we first review apposite fundamentals of DNA structure and then introduce those aspects of phase diagrams, isomorphism and polymorphism relevant here. The complexes can be classified in two main groups – the


135

first has the drug molecules intercalated between two successive base pairs and the second has the drug molecule nestling in the minor groove of the DNA oligomer; we treat these separately, In these complexes the DNA fragment is the ‘‘host’’ and the drug molecule the ‘‘guest;’’ some attention is given to the structures of the native hosts (i.e. without guests) because the variety that is found impinges on the structures of the complexes. There are a number of excellent reviews that give more detail than is possible here. The Oxford Handbook of Nucleic Acid Structures (Neidle, 1999) is one recent compendium; another, edited by Demeunynck, Bailly and Wilson (2003), has 25 chapters by 27 authors; besides its broad coverage of topics relevant to this chapter, and far beyond, it includes an illuminating introductory chapter by Waring and Wakelin (2003) and a very useful collection of some thousands of references with titles. The principal data resource is the Nucleic Acid Database (Berman, Westbrook et al., (2002); there were some 2500 entries by late 2004, not all relevant here. Bloomfield et al. (2000) give an overall survey in their Nucleic Acids. Structures, Properties and Functions. Yang and Wang (1999) give an overview of DNA-drug interactions. A Google search for ‘‘nucleic acid databases’’ gave 240 000 hits. Our coverage is selective rather than comprehensive – we omit many important topics outside the mainstream of past and current activity. Most authors emphasize the effects of base sequence on the detailed structure of the oligonucleotides and how the details of the binding of drug molecules differ from one group of oligonucleotides to the next; thus most structural papers include exhaustive analyses of oligonucleotide structure. Their goal has been to account for the large differences found in drug behavior despite small differences in chemical structure. Packing effects on oligonucleotide structure were once considered to be minimal with only passing attention paid to crystal structures as such. Accumulation of information has led to a change in attitude and most structures show appreciable environmental influence. Readers are directed to the original studies for detailed analyses of the geometries and binding interactions in the native oligonucleotides and their drug complexes, including discussions of water structure and the role of metal ions. This material must eventually find application also in discussions of crystal structure, but that stage of development has not yet been reached. Here we draw attention to an approximation that pervades this whole chapter. It is convenient to refer to the complexes as oligonucleotide-drug complexes, with the subliminal implication that the systems are binary. However, about half (by weight) of the crystals consists of solvent (mainly water, more or less organized) and there are also metal cations and spermine. It is surely correct to assign the major structural significance to the oligonucleotide-drug portion but the residue may not be neglected, as has been pointed out by many authors (e.g. Guerri, Simpson and Neidle, 1998). Also, one should not forget that an extrapolation is involved when making inferences about DNA structure and behavior based on results obtained for DNA oligomers. However, there seem to be good reasons to believe that such extrapolations are generally permissible. There is an important area that impinges on the subject matter of this chapter – the formation of complexes between proteins and DNA. However, this would require a book on its own and we give only a leading reference (Goldman, 1996).


136

5.2

Fundamentals of oligonucleotide structure

5.2.1

General aspects

The gross structure of DNA is well known, and the DNA oligomers have structures based on the same general principles. The description is conveniently given in terms of a spiral staircase, where the helical outer frame is composed of deoxyribose phosphate groups, and the treads (steps) of hydrogen-bonded pyrimidine–purine pairs. The sugar phosphate groups are negatively charged and so counterions, generally sodium and/or magnesium (calcium and barium are also possibilities, as is spermine (Scheme 5.4)), must accompany the double helix, and are usually embedded in an (often ill-defined) envelope of water molecules; DNA is usually about 50% water by mass. Thus, in formal terms, the oligonucleotides are salts or salt-molecule complexes, depending on the ionization state of the drug molecules. DNA has been found to occur in three stereochemical forms – in A and B the helices are right-handed while left-handed helices are found in the Z form. In earlier work counterions were only positively identified in the Z-DNA hexamers, the B-DNA decamers and some of the DNA-drug complexes,1 but this situation is changing with the increased resolution being attained in crystal structure analyses. We shall first discuss the individual components of the whole association, and then consider how they are combined into an oligomer of particular conformation. The base pairing has received most attention in the past. Watson and Crick showed that treads of the required dimensions to bridge between the two sugar phosphate spirals were obtained by hydrogen bonding guanine–cytosine (GC) and thymine–adenine (TA) pyrimidine–purine base pairs (Fig. 5.1). These are the complementary Watson–Crick base pairs and the implication is that each C in one strand can only be hydrogen bonded to G in the other strand, and similarly for the TA pair. Thus the two strands of the double helix are antiparallel (Fig. 5.2). Although Watson–Crick base pairing is by far the most common, work during the last forty years has shown that other base pairings do occur and are relevant to the structures of the DNA oligomers. The first of these alternative pairings was encountered by Hoogsteen (1959) in the crystal structure of the 1:1 complex of 1-methylthymine and 9-methyladenine. Among other examples of mismatched pairs (‘‘wobble’’ base pairs) are the pyrimidine–purine pairs CHþG (CHþ symbolizes protonated cytosine), TG (with three bridging water molecules included in the hydrogen bonding scheme), CA (one bridging water) and the purine–purine pair GA, which is found in three variations G(anti)A(syn), G(syn)A(anti) and G(anti)A(anti). The oligonucleotides discussed below are mostly self-complementary and based on Watson–Crick base pairing. However, the strands are not required to be identical – a complementary example with nonidentical strands is (5 0 -CGCAAAAAAGCG-3 0 ) (5 0 -CGCTTTTTTGCG-3 0 ) (Nelson, Finch et al., 1987; Table 5.13) while a noncomplementary example (also with non-identical strands) is {(5 0 -CG[5BrC]ATATTTCGC-3 0 ) þ (5 0 -CGCAAATATGCG-3 0 )}, which has GT and CA non-Watson–Crick base pairs (Aymami, Nunn and Neidle, 1999; Table 5.13).

1

The conformations denoted by Z, A and B will be defined below.

FUNDAMENTALS OF OLIGONUCLEOTIDE STRUCTURE H

H

H

H C C

O

C N 50°

To chain

0.2

8n

C

m

H

T

Thymine

137

N

0n

H

C

H

N

0.3

Adenine

m

C

N

N

C

A

O C

1.1

C C

1 nm H

N

H

N 51° To chain

52 ° H

H

H

C

N

C

C

Cytosine To chain

N 52°

0.2

H

C 0n

N

H 0.2

m

m

N 8n

m

Guanine C

N

9n

H 1.0

m

O

0.3

C O

9n

G C

C C

N

N

C

H

N 54° To chain

H

Fig. 5.1. The complementary T-A and CG hydrogen bonded Watson–Crick pyrimidine-purine base pairs. Line drawings with dimensions are on the left and space-filling models on the right. (Adapted from Fig. 31.3 of Lehninger, 1977.)

3⬘

5⬘

Strand I

Strand II

C

G

G

C

A

T

T

A

C

G

G

C

5⬘

3⬘

Fig. 5.2. A model of the base pairing in a hexameric oligonucleotide; base sequence is listed from 5 0 to 3 0 . The dashed lines joining the base pairs represent the hydrogen bonding schemes shown in more detail in Fig. 5.1. In this example, and in most of the other oligonucleotides discussed here, the two strands are antiparallel and identical, but identity is not required. The hexameric oligomer is designated d(5 0 -GCTAGC-3 0 )2. (Adapted from Fig. 31.4 of Lehninger, 1977.)


138

The sugar–phosphate arrangement in a single strand is shown schematically in Fig. 5.3. The bases are linked by 3 0 ,5 0 -phosphodiester bonds; the particular fragment illustrated can be written as d(5 0 -ApGpCpT-3 0 ), where d represents deoxy (RNA fragments, with an hydroxyl on the 2 0 ring carbon, are shown as ‘r’ or by lower case letters (‘c’ instead of ‘C’)), ‘p’ represents phosphate and A etc are abbreviations for the bases; a more usual notation is d(AGCT), with the duplex nature of the oligonucleotides discussed here emphasized by using the notation d(AGCT)2. The chain has a certain amount of conformational flexibility. We quote the concise descriptions of the conformations of A, B and Z forms given by Kennard and Hunter (1991; see also Kennard and Salisbury, 1993): ‘‘The A form has 11

H N

O O

P

n–1

O

X

C H C H H C

O n

C

A

N

H C

H O

C

H

N

C

N

C

N

H

O H C C H

O

H

P

O

O6 N 8 5⬘

H

6

9

4 3

N

N

4⬘

1⬘

H

N

2⬘

3⬘

G

1N 2

O 4⬘

O

5

7

N2 H H

H

C

N

O 1

O

P

n+1

O

N

O

O

O O4 CH3 O

5 4 6 1

O

P

n+ 2

O

N

O

N

H

T

3 2

O2

O

Y

O O

P

O

O

Fig. 5.3. Schematic arrangement of the four Watson–Crick bases and sugar-phosphate groups in a single strand of a fragment of DNA. Y shows the direction of the chain that runs from the 5 0 carbon to the 3 0 carbon. X indicates a nucleotide residue. A full notation for the duplex would be d(5 0 -ApGpCpT-3 0 )2. (Reproduced from Fig. 26 of Kennard and Hunter (1991)).

FUNDAMENTALS OF OLIGONUCLEOTIDE STRUCTURE

139

base pairs per turn of helix with 32.7 twist between adjacent base pairs. The rise per ˚ . The bases are oriented anti about the glycoside bond and the furanose base pair is 2.7 A ring adopts the C3 0 endo conformation. The global helix axis lies in the major groove with the bases displaced some distance away . . . . The B form has 10 base pairs per turn, ˚ . The glycoside orientation is anti a helix twist of 36.0 and a rise per base pair of 3.4 A 0 and the sugar conformation tends to C2 endo. The global helix axis is through the base pairs . . . . The Z form is a left-handed helix with a zigzag sugar–phosphate backbone, hence the name. The Z form is mostly adopted by alternating cytosine/guanine sequences and can be considered as a repeat of d(CpG) steps. There are 12 base pairs per ˚ . The helix twist, glycoside orientation and sugar conformation turn and a rise of 3.7 A depend on whether the pyrimidine or purine is being considered or whether it is the first or second base pair of the dinucleotide step.’’ In fibers, whether the A or B form is obtained depends on environmental conditions and this led to the notion that these conformations represent discontinuous states, only stable in very different environments. This point of view has been revised and it is now accepted that A and B forms have similar energies, and stabilization of one or other depends on sequence as well as on oligomer arrangements in the crystals; thus a continuum of right-handed DNA conformations, spanning the range from A to B, could be expected. An example is given in Fig. 5.9. A rough first approximation to the shape of oligonucleotides describes them as right cylinders; this is shown in some packing diagrams such as Fig. 5.7. A better approximation has the surface of the molecules molded into helical minor and major grooves (Fig. 5.4). The larger oligonucleotides (e.g the dodecanucleotides) are often kinked in the sense that

Minor groove

Major Groove

Fig. 5.4. Schematic representation of the major and minor grooves of a B-oligonucleotide. The helical laths represent the sugar-phosphate periphery and the cross-bars the base pairs. (Diagram kindly supplied by Prof. Noam Adir (Technion, 2004)).

140


they can be divided into two quasi-cylindrical units with axes inclined to one another; the d(CGCCCGCGGGCG) dodecamer (Table 5.5; Malinina, Fernandez, Hyunh-Dinh and Subirana, 1999) is an example. We quote from Tidwell and Boykin (2003): ‘‘In principle, molecules can bind to either the major or minor grooves of DNA. Due to the great difference . . . in the dimensions of the two grooves, targeting them requires vastly different shaped molecules. The major groove . . . is much wider than the minor groove; the groove width values for ˚ , respectively . . . the major groove is averaged-sequence B-form DNA are 11.6 and 6.0 A the site of binding of many DNA interacting proteins . . . [with only] limited reports of non-protein molecules that bind to the DNA major groove . . . Minor groove binding usually involves greater binding affinity and higher sequence specificity than that of intercalator binding.’’ 5.2.2

Single crystal x-ray diffraction studies of oligonucleotides

The detail obtainable from the results of crystal structure analyses depends ultimately on the nature of the samples, assuming that instrumentation and computing facilities are ‘‘state of the art.’’ The crystals considered in this book cover a wide spectrum of ‘‘diffraction quality’’ ranging from ‘‘exceptional’’ ({benzeneAgClO4} Section 11.11.1) through ‘‘standard’’ (most of the structures) to ‘‘poor’’ (some of the intercalation complexes of Chapter 9). The crystalline oligonucleotides are usually somewhere between ‘‘standard’’ and ‘‘poor.’’ The unit cells are relatively large (volumes of ˚ 3), the number of reflections some thousands and the R factors around 50 000 A obtained around 15–20%. Some improvement can be obtained by structure determination at low temperatures (typically 100K) but this is still the exception rather than the rule. We also note that the small (or minute) quantities of material available do not facilitate crystallochemical studies. There are indications from cell dimension measurements that not all crystals of some complexes are identical. For example, independently prepared crystals of {d(CGCGAATTCGCG)2-Hoechst 33258} measured ˚ 3 whereas two other at 173, 248 and 273K all have cell volumes close to 61 000 A ˚ 3 (see Table 5.13). A crystals (measured at 288 and 298K) had volumes of 67 000 A difference of 10% is surely to be ascribed to compositional or structural differences rather than experimental error. Similar examples are to be found below in the tables of crystallographic data (e.g. d(CCCGCGGG) in Table 5.7 (Fernandez, Subirana et al., 1997)). In small-molecule crystal structure analysis the usual procedure is to define the unit cell contents by a chemical formula (obtained by chemical analysis) or, at least, a mass content obtained through the measured density. Neither chemical analysis nor measured density is generally reported for biomolecular crystals and we use the cell volume as a designator of content. This is a measurable and useful entity despite its incomplete information tally. The assumption is usually made that equality of cell volumes (or their relationship by a small integer) implies similarity of chemical composition. A landmark review by Dickerson (1992), written when about 100 crystal structures of DNA oligonucleotides and complexes had been reported, contains a treatment of

FUNDAMENTALS OF OLIGONUCLEOTIDE STRUCTURE

141

the structure analysis process and much structural information that is still relevant. Some we shall extract here; that part omitted will repay study. Dickerson’s main classification was in terms of helical configuration (A, B, Z2 and irregular) with a secondary division in terms of space group. As our main interest is in the complexes our first division is between the two main groups of complexes – intercalation complexes and minor groove complexes;3 fortunately there is little overlap. Our next criterion is oligonucleotide length (hexamers, octamers, decamers and dodecamers), with a further subdivision in terms of sequence. We have grouped together isomorphous crystals and attempted to identify polymorphism. Native crystals and complexes are treated together because there is considerable overlap between some of these groups. Other problems of definition stem from the limited resolution of many data sets. This may be a result of inadequate crystal (diffraction) quality or of real disorder effects. One criterion of ‘‘resolution’’ is how far out (in reciprocal space) the measured reflections ˚; 1A ˚ atomic resolution extend. This is usually expressed as the direct space value X A corresponds to a Bragg angle of 60 for Cu K radiation. Other ‘‘diffraction quality’’ criteria are discussed by Dickerson (1992). One consequence of disorder is found among minor groove complexes (see below) that have been divided into two types of model – in Class I models the drug molecule takes up a single position in the minor groove while in Class II models there is apparent end-to-end disorder of the (almost symmetrical) drug molecule that may result from incorrect analysis because of the limited resolution of the data set (Goodsell, Kopka and Dickerson, 1995). Thus reports of Class II structures should be treated with some reserve. An authenticated example is the monoimidazole lexitropsin complex of the dodecamer d(CGCGAATTCGCG) (Goodsell, Ng et al., 1995). The next landmark is surely the establishment of the Nucleic Acid Database (NDB; Berman, Olson et al., 1992) and its subsequent development. The NDB contains detailed experimental and structural information (from crystal growth through diffraction techniques through geometrical results) for more than 2 000 structures (NMR studies are included although not considered here), as well as tools for analyzing this treasure trove. In our Table 5.1 we reproduce the description given in Table 1 of Berman, Westbrook et al. (2002). We have used the NDB to provide data for the limited number of topics covered in this chapter. The NDB does not give information about solvent content of reported structures (perhaps because this may change as analyses improve). The limitations of earlier and current results have been stressed in the paragraphs ˚ are above but the face of the future can also be seen. Resolutions of 0.7 A now being achieved both in protein and nucleic acid crystallography. For example, the B-DNA double helix structure of 5 0 -CCAGTACTGG-3 0 has been ˚ using a combination of liquid nitrogen temdetermined to a resolution of 0.74 A ˚ ) (Kielkopf, Ding, Kuhn and perature and synchrotron radiation (wavelength 0.78 A Rees, 2000). 2 3

We do not discuss Z-DNA because of limitations of space. Sometimes called ‘‘parallel’’ (to the planes of the base pairs) and ‘‘perpendicular’’.

142


Table 5.1. The information content of the NDB (reproduced from Berman, Westbrook et al. 2002) with some additions Primary experimental information stored in the NDB Structure summary – descriptor NDB, PDB and CSD names; coordinate availability; modifications, mismatch and drug binding. Stuctural description – sequence; structure type; descriptions about modifications, mismatches and drugs; description of asymmetric and biological units. Citations – authors, title, journal, volume, pages, year. Crystal data – cell dimensions, space group. Data collection description – radiation source and wavelength; data-collection device; temperature; resolution range; total and unique number of reflections. Crystallization description – method; temperature; pH value; solution composition. Refinement information – method; program; number of reflections used for refinement; data cutoff; resolution range; R factor; refinement of temperature factors and occupancies. Coordinate information – atomic coordinates, occupancies and temperature factors for asymmetric unit; coordinates for symmetry-related strands; coordinates for unit cell; symmetry-related coordinates; orthogonal or fractional coordinates. Derivative information stored in the NDB Distances – chemical bond lengths; virtual bonds (involving P atoms) Torsions – backbone and side-chain torion angles; pseudo-rotational parameters. Angles – valence bond angles, virtual angles (involving P atoms) Base morphology – parameters calculated by different algorithms Non-bonded contacts Hydrogen bonding classification Valence geometry RMS deviations from small molecule standards Sequence pattern statistics CIF information – coordinates, structure factors.

5.3 5.3.1

Crystal chemistry of oligonucleotide and oligonucleotide-guest structures Polymorphism, isomorphism, and heteromorphism

There are many striking resemblances, and also some differences, in the crystal chemistries of cyclodextrin inclusion complexes (Chapter 4), tetraphenylporphyrin-MII inclusion complexes (Chapter 8) and oligonucleotide native crystals and complexes. The established concepts of polymorphism and isomorphism are basic to the discussion and we find it convenient to introduce a new term ‘‘heteromorphism.’’ We define these terms immediately below, and then illustrate and compare them for particular examples from the three groups of cyclodextrins, tetraphenylporphyrin-MII moieties and oligonucleotides. Our approach is that terms and definitions currently accepted in small-moiety crystallography should be applied to inclusion complexes and biomolecules with minimal change. 1. Polymorphism. Experience and calculation show that many different packing arrangements have similar lattice energies4 (Lommerse et al., 2000) and this gives rise 4 Free energies rather than enthalpies should be considered but entropy is more difficult to calculate and is often explicitly or implicitly ignored.

OLIGONUCLEOTIDE AND OLIGONUCLEOTIDE-GUEST STRUCTURES

143

to the phenomenon of polymorphism. A formal definition of polymorphism is: when the same chemical compound occurs in two different crystalline forms these are called polymorphs; we use ‘two’ for simplicity but polymorph clusters can be larger than pairs. When a first order phase transformation occurs from one phase to the other between low temperature and the melting point, then the two phases are said to be enantiotropically related. If such a phase transformation does not occur then the two phases are monotropically related. A particular chemical substance can show both enantiotropic and monotropic behavior; quartz is one of many examples. The implication of the word ‘same’ is that there is no difference in chemical formula between the two phases. Should a difference in chemical nature (e.g. diamond Sn and metallic Sn) be allowed? These definitions were developed for small-moiety crystallography (including here inorganic crystals and metals) and not much attention has been given to assessing their applicability to biomolecular crystallography. The matter becomes much more complicated when dealing with biological molecules because of the presence of such large quantities of solvent (generally water) in the crystals, as well as ions. Analogous problems can occur with minerals. Polymorphism seen in the context of classical physical chemistry has been reviewed recently (Herbstein, 2004). 2. Isomorphism. Two crystals are defined as isomorphous5 when they crystallize in the same space group, in very similar unit cells and with very similar atomic coordinates. The term ‘‘very similar’’ has not been defined in quantitative terms, although progress, summarized by Dziubek and Katrusiak (2003), has been made in this direction. When the two structures are similar (without defining ‘‘how similar’’) then the two materials are said to be isostructural. Are Cu, Ni and Au isomorphous or isostructural? This question is discussed in Section 10.6.1. Perhaps the best answer is that latitude and flexibility in applying the definitions are preferable to rigidity. 3. Heteromorphism. This term is introduced to cover the situation where different crystals have formula units of the same volume and but different chemical compositions. These are not polymorphs because the different crystals have different compositions. This concept is found to be useful in the three categories of substance discussed here. ‘‘Hetero’’ refers to the different crystal structures and, for oligonucleotides, to different base pair sequences. In a sentence, polymorphism occurs when different crystal structures are found for the same composition, isomorphism when the same crystal structure is found despite differences in composition, heteromorphism when analogous formula units (of the same volume but different chemical compositions) occur in different crystal structures. The keywords ‘‘same’’ and ‘‘different’’ are open to interpretation. 5.3.2 Phase rule relationships In host–guest inclusion complexes, if the host crystal structure (no guest, native) is isomorphous with that of the inclusion complex then the complex is a primary solid solution phase. It is generally not known whether the host:guest ratio is fixed or variable. One example where the host–guest ratio has been shown to cover a wide range is bromine 5

Our distinction between ‘‘isomorphous’’ and ‘‘isostructural’’ is not used by all authors.

144


hydrate (Section 7.2.7.4); this is an intermediate phase rather than a primary solid solution. An example of a true primary solid solution is trimesic acid/bromine (Section 8.3.4), but here the composition range has not been established. If the complex and native host are not isomorphous then the host–guest complex is an intermediate phase in the ‘‘binary’’ phase diagram. For understandable reasons, no one seems as yet to have tried to determine a binary oligonucleotide-second component phase diagram, where ‘‘second component’’ is, for example, a drug molecule. 5.3.3

Applications of these concepts

1. Cyclodextrin complexes. A large amount of information is embedded in the tables of Chapter 4. We take the clathrate inclusion complexes of -cyclodextrin as our illustrative example. A herringbone arrangement is found in 15 complexes of Table 4.2. ˚ , space group These complexes are isomorphous with a 9.5, b 14.3 and c 37.5 A P212121, Z ¼ 4; the guests cover a range of chemical types. There is a second group of ˚ , space 10 isomorphous complexes in Table 4.3, with a 13.6, b 15.3 and c 24.5 A group P212121, Z ¼ 4; the guests cover a range of chemical types. That the two groups have the same space group is coincidental. The two groups are heteromorphic, both unit ˚ 3. There are no polymorphs. Both groups are cell volumes being around 51 000 A different intermediate phases in the hypothetical -cyclodextrin/guest/water phase diagram. In the cyclodextrins the host molecule is invariant while the nature (perhaps number) of guest molecules differs from complex to complex, as does the water content, which is sometimes not well established. 2. {Tetraphenylporphyrin-MII(guest)2} complexes. A large amount of information is embedded in the tables of Chapter 8. In the {tetraphenylporphyrin-MII(guest)2} inclusion complexes, the host molecules differ to a small extent because of differing metal ions while the different guest molecules appear to have a greater influence on crystal structure. The chemical compositions of the tetraphenylporphyrin-MII complexes are generally well established. We take the four-coordinate inclusion complexes listed in Table 8.17 as specific examples. The second to the tenth entry in ˚ , 106, 112, Table 8.17 are all isomorphous (a 10.5, b 11.2, c 12.0 A 103 ; Type II triclinic cell (all angles 90 , space group P1, Z ¼ 1). This group is followed by a group of 20, isostructural rather than isomorphous with the first group. These two groups are followed by a third, smaller, group with a Type I triclinic cell (all angles O > N > C. (Reproduced from Moore, Hunter et al., 1989.)

3/8

3/8 1/4

1/4

1/4

1/4 1/8

1/8

1/4

1/4 1/8

1/8 1/4

1/4

1/4 3/8

1/4 3/8

Fig. 5.7. Schematic view down [001] shows the packing arrangement of the DNA-drug complexes in the unit cell with space group P41212. The molecular twofold axis coincides with the crystallographic twofold axis at z ¼ 0, which runs diagonally across the square. The two complexes, stacked end-toend along the two fold screw axis in the [001] direction, are related by the twofold axis at z ¼ 1/4. The solvent channels are in the areas near the 41 axes. The dashed circles represent end views of the elongated hexamer-drug complexes. (Reproduced from Wang, Ughetto et al., 1987.)

I N T E RC AL AT E D HE XANUC L E OT IDE - DR UG COMPLEXES

151

are about 10% and may suggest that the various crystals have different (solvent) compositions rather than being heteromorphs. The {(hexamer duplex)2(daunomycin)} complex is shown as a ball-and-stick stereopair in Fig. 5.6 and as a schematic packing diagram in Fig. 5.7. The molecules ˚ and height 26.5 A ˚. can be roughly represented as cylinders of diameter 20 A These cylinders, whose axes lie along [001], are approximately close packed in two dimensions. 5.4.2 Nogalamycin and derivatives Nogalamycin and derivatives (Scheme 5.2) also belong to the anthracycline family of antibiotics and are active against a number of tumor lines. However, difficulties in administering the drugs have discouraged clinical trials. Nevertheless, interaction of nogalamycins with DNA has remained of interest. Crystal structures of a number of hexanucleotide-nogalamycin complexes have been described (Table 5.3).

R1

R2

Nogalamycin

COOCH3

CH3

Disnogalamycin

H

CH3

U-58872

COOCH3

CHO

H

R2

OH

N+ H3C aminoglucose (positively HO charged) H3C

O

O

O

R1 OH

D

C

R2

O

B

A

O

O

CH3

H H3C

O

H3CO OCH3

Scheme 5.2

nogalose (neutral) OCH3

152


Table 5.3. Intercalation of nogalamycin or derivatives in hexameric oligonucleotides (some modified); duplex:nogalamycin ratio is 1 : 2. All crystals have distorted right-handed B-DNA-like ˚, A ˚ 3. Chemical formulae in duplex structures with Watson-Crick base pairs. Dimensions in A Scheme 5.3. (pS) indicates that the phosphate group at the TpA step. has been replaced by phosphorothioate. [m5C] is 5-methylcytosine (5-MeC is also used). Structures at RT unless stated otherwise Hexameric oligonucleotide

Intercalated drug

a

b

c

V

NDB ID; Reference

Tetragonal P41212, Z ¼ 8 d(TGATCA)2 at 150K

nogalamycin 37.29 37.29 71.12

98 895 DDF049, 052, 063; SST96

Orthorhombic P212121; Z ¼ 4 d(TGTACA)2 at 150K

nogalamycin 26.30 51.98 67.08

91 703 DDF064; SBM96

Orthorhombic C2221; Z ¼ 8 d(CGT(pS)ACG)2

nogalamycin 22.98 47.27 64.44

69 999 LGR89

Hexagonal; P61 Z ¼ 6 d([m5C]GT(pS)A[m5C]G)2 d([m5C]GT(pS)A[m5C]G)2

nogalamycin 26.31 26.31 100.26 60 104 GLRW90 U-58872 26.27 26.27 100.23 59 903 GLRW90

Hexagonal; P6122; Z ¼ 6 d([m5C]GT(pS)A[m5C]G)239W nogalamycin 26.30 26.30 100.01 59 908 WEG90 References: GLRW90 – Gao, Liaw, Robinson and Wang, 1990; LGR89 – Liaw, Gao, Robinson, van der Marel, van Boom and Wang, 1989; SBM96 – Smith, Brannigan and Moore, 1996; SST96 – Schuerman, Smith, Turkenburg, Dettmar, Van Meervelt and Moore, 1996; WEG90 – Williams, Egli, Gao, Bash, van der Marel, van Boom, Rich and Frederick, 1990.

The hexamer-duplex/drug ratio is 1 : 2 in all the crystals. The asymmetric units of the orthorhombic and P61 hexagonal modifications contain two hexanucleotides, two nogalamycins, two hydrated magnesium ions and 113 water molecules. The P61 hexagonal modification approximates to P6122 symmetry and, under different crystallization conditions, a higher symmetry modification with this space group is obtained, where one strand of the DNA duplex and one drug molecule form the asymmetric unit. 39 water molecules were located. The DNA-drug interaction in both hexagonal and orthorhombic modifications has been described as follows by Liaw, Gao et al. (1989): ‘‘two nogalamycins bind to the DNA double helix in a 2 : 1 ratio with aglycon chromophore intercalated between the CpG steps at both ends of the helix. The nogalose and aminoglucose sugars lie in the minor and major grooves, respectively, of the distorted B-DNA double helix.’’ The hexagonal structures suggest that the volume per formula unit is about ˚ 3, which does not fit the unit cell volumes obtained for the tetragonal and 10 000 A orthorhombic structures. The packing in the tetragonal [d(TGATCA)–nogalamycin} complex has been shown by Smith, Davies, Dodson and Moore (1995) (Fig. 5.8; a somewhat different view is given by Schuerman et al. (1996); their Fig. 4). In overall terms Fig. 5.8 is remarkably similar to the analogous figure for the d([m5C]GT(pS)A[m5C]G)2 hexagonal P61 modification (Fig. 5.9).


153

a

Fig. 5.8. Partial view of the packing diagram of the P41212 crystal form of the nogalamycin/ ˚ ), d(TGATCA) complex, seen approximately along [101]. There are four layers along [001] (¼ 71 A which is horizontal in the diagram. (Reproduced from Smith, Davies, Dodson and Moore (1995).)

A

B

c/6

Fig. 5.9. Packing diagram of the P61 crystal form of the nogalamycin/d([m5C]GT(pS)A[m5C]CG) complex. The complexes are stacked end-over-end along a (and b) axes to form a sheet. One of these ˚ ). The packing of the sheets along the c direction sheets is shown in A, viewed down the c axis (100 A is shown in B; the thiophospho group abuts adjacent complexes. There are six sheets along the c axis. (Reproduced from Gao, Liaw, Robinson and Wang, 1990.)

154


In both structures there are layers of hexamer-drug complex moieties in (001) planes, with four layers along [001] in the tetragonal complex and six in the hexagonal complex. However, the hexagonal and orthorhombic phases are not accepted as heteromorphs because they do not have the same chemical composition as inferred from the unit cell volumes. The relation between the P61 and P6122 phases is not entirely clear; perhaps a small composition difference accompanies the small symmetry difference. The distortion of the hexamer duplex consequent on the intercalation is greater than that found in the anthracycline drug complexes – hence, a fortiori, the hexamer-nogalamycin phases are separate phases in the DNA-drug system.

5.4.3

The 9-aminoacridine drugs

The 9-aminoacridine-4-carboxamide class of compounds were originally explored as part of a broad attempt to develop acridune derivatives as anticancer drugs. Because these compounds exist as dications under physiological conditions, they are tight binding DNA intercalating agents. They are also potent cytoxins. This general background comes from Denny (2003). NH2

5

N H2

H

R O

N H

R=

H CH3 N+ CH3

R = HN +

H2

9-amino-[N-(2-methylamino)ethyl]acridine4-carboxamide

Morpholino-9-amino-[N-(2-methylamino)ethyl] acridine-4-carboxamide

Scheme 5.3

The structures listed in Table 5.4 are isomorphous. The asymmetric unit (using the 9-amino-DACA complex as example) consists of a single strand of the hexanucleotide, one intercalated 9-amino-DACA molecule, 27 ordered water molecules and, on a crystallographic two fold axis, an additional ‘end-stacked’ 9-amino-DACA molecule (necessarily disordered) and two water molecules. No metal atoms or spermine appeared. Thus the composition is {(hexamer duplex)3(9-amino-DACA)}. The two strands of DNA


155

Table 5.4. Intercalation of 9-aminoacridine derivatives in hexameric oligonucleotides. All crystals have distorted right-handed B-DNA-like duplex structures with Watson-Crick base pairs; the ˚, A ˚3 common space group is P64, Z ¼ 6. Dimensions in A Hexanucleotide

Intercalated drug

a

c

Cell Volume NDB ID; Reference

d(CGTACG)2 at 110K d(CGTACG)2 at 110K d(CGTACG)221W at 110K d(CGTACG)221W at 110K d(CG5-BrUACG)2 at 110K d(CCTAGG)2 at RT

9-amino-DACA

30.16

39.69

34 336

morpholinyl

30.239 39.340 34 214

5-fluoro-9-amino-DACA* 30.14

39.40

34 040

5-bromo-9-amino-DACA* 30.19

39.44

34 188

DD0015; AGCDW99 DD0048; AGDW02 DD0023; AGCDPW00 DD0051; TTT

6-bromo-9-amino-DACA

30.087 39.316 33 848

TATDWC99

cryptolepine

29.96

DD0047; LCP02

39.65

33 848

Notes: 9-amino-DACA is 9-amino[N-(2-dimethylamino)ethyl]acridine-4-carboxamide morpholinyl is 9-amino-N-(2-(4-morpholinyl)ethyl)acridine-4-carboxamide cryptolepine is 5-methyl indolo[2.3b]-quinoline * occupancies of drug molecules appear to be appreciably less than 1. References: AGCDW99 – Adams, Guss, Collyer, Denny and Wakelin, 1999; AGCDPW00 – Adams, Guss, Collyer, Denny, Prakash and Wakelin, 2000; AGDW02 – Adams, Guss, Denny and Wakelin, 2002; LCP02 – Lisgarten, Coll, Portugal et al., 2002; TATDWC99 – Todd, Adams, Thorpe, Denny, Wakelin and Cardin, 1999; TTT – Teixera, Thorpe, Todd et al., to be published (DD0051).

are related by a dyad axis and form a right-handed DNA duplex with Watson–Crick base pairing. The intercalated 9-amino-DACA molecule is located between each of the CpG base pair steps with its side chain in the major groove. The additional 9-aminoDACA molecules on the two fold axis stack at the ends of each DNA helix and link one duplex to the next by hydrogen bonds to form a continuous column of duplexes in the ab plane (Fig. 5.10). Because the 9-aminoacridine complexes have both drug molecules intercalated in the hexameric duplexes and additional drug-molecules between the duplexes, these complexes are separate phases in the hexamer-drug system. ˚ 3, somewhat The volume of the asymmetric unit in these complexes is about 5 700 A larger than the value obtained for the P6122 nogalamycin complex of Table 5.2; in both examples the asymmetric unit contains a single strand of the hexamer – whether these are heteromorphs remains moot. A partial view of the packing arrangement is shown in Fig. 5.10. Layers of molecules are formed, one layer being shown in Fig. 5.10 together with the space group symmetry elements which enable one to complete the packing diagram. Successive layers along [001] are obtained by operation of the 31 screw axis.

156


B A

A B

Fig. 5.10. One layer of {d(CG5-BrUACG)2(6-bromo-9-amino-DACA)} projected down the [001] axis, looking directly into the major groove. The space group is P64 and the symmetry elements are shown. The helix axis of the duplex is along [100] and is intersected by a two fold axis along [001]. A and B are the intercalated and ‘‘end stacked’’ DACA molecules respectively. (Adapted from Todd, Adams, Thorpe, Denny, Wakelin and Cardin, 1999.)

5.4.4

Native hexanucleotides and comparison of crystal structures

The B-DNA structure of the native hexanucleotide d(CTCGAC)2 has been reported ˚, (Wahl, Rao and Sundaralingam, 1996; BDF068). The hexagonal crystals (40.14, 44.47 A ˚ 3, P6222, Z ¼ 6) show an A-DNA like conformation at the termini of V ¼ 62052 A the duplex. The volume per asymmetric unit is much the same as for the crystals of the various sequences complexed with a variety of drug molecules (Table 5.2) but the two types of hexanucleotide are not isostructural. A native hexanucleotide derivative Rp-d(Gp(S)CpGp(S)CpGp(S)C), where ‘‘p’’ represents a phosphate group and ‘‘p(S)’’ a phosphorothioate group, crystallizes in a modification not yet encountered ˚ , V ¼ 23201 A ˚ 3, Z ¼ 4, P212121) elsewhere (orthorhombic, 34.90, 39.15, 20.64 A (Cruse, Salisbury et al., 1988). This is more densely packed than d(CTCGAC)2, as is shown in Fig. 5.11. The helical duplexes, considered as quasi-cylinders, are approximately close packed. The principal interactions are between cations and phosphate and phosphorothioate groups. Strand 1 of the duplex interacts mostly with symmetry-related strand 1* and similarly for strand 2. A number of possible hydrogen bonds were located.


157

Fig. 5.11. Stereodiagram of the packing arrangement in Rp-d(Gp(S)CpGp(S)CpGp(S)C), viewed down the [001] axis, which coincides with the helix axis. The solvent molecules are shown as open circles. 72 solvent molecules, of the approximately 143 in the asymmetric unit, have been located; some of these may be sodium or magnesium ions. (Reproduced from Cruse, Salisbury et al., 1988.)

Another native hexanucleotide has a Z-configuration – indeed it was the prototype of the Z configuration (Wang, Quigley, et al., 1979); this is not relevant for comparison with the B-DNA hexamers considered here. Would a hypothetical native B-DNA hexamer be expected to be isomorphous with the complexes listed in Table 5.2? Probably not, in view of the large perturbation of the hexameric duplex by the two intercalated drug molecules. If so, then the complexes are not to be classified as solid solutions but rather as separate phases in the binary hexamer– drug phase diagram. The packing arrangements in the anthracycline drug complexes differ from those in the nogalamycin and 9-aminoacridine complexes. In the anthracycline drug complexes the quasi-cylindrical molecules have their long axes along the [001] axis of the space group, leading to approximate close packing of parallel cylinders. In the hexagonal nogalamycin and acridine complexes the quasi-cylindrical molecules are arranged in the basal plane, and the layers of cylinders are arranged along the [001] axis by the operation of 31 and 64 screw axes respectively. The packing in the tetragonal and orthorhombic crystals does not appear to have been described in detail. The versatility of the d(CGATCG)2 base pair sequence should be noted. This forms some 15 isomorphous tetragonal complexes with different daunomycin-type guests (Table 5.2), monoclinic and orthorhombic crystals with dd-MOX as guest (Notes to Table 5.2), a triclinic 1 : 2 complex with 3 0 -desamino-2 0 -(2-methoxy-4-morpholinyl)daunomycin, (NDB-ID DDF041; Cirilli et al., 1993) and isomorphous hexagonal complexes with various 9-amino-DACA derivatives (Table 5.4). It is not known whether other base pair sequences behave similarly.

158

5.5


Isomorphism and polymorphism of A-DNA octanucleotides and the binding of spermine

The self-complementary octanucleotide d(ACGTACCG)2, where the asymmetric unit is a single strand of octamer, together with about 20 water molecules, crystallizes in space group P43212 (Table 5.5). The two octamer strands are related by a crystallographic twofold axis to form a right handed duplex. Among other differently sequenced octanucleotide duplexes crystallizing in the same space group are d(GCCCGGGC)2 (Heinemann et al., 1987), d(ATGCGCAT)2 (Clark et al., 1990), d(GTCTAGAC)2 (Cervi et al., 1992) and d(GTGCGCAT)2 (Bingham et al., 1992). The brominated analog d(ACGTACC[5 BrU])2 is isomorphous, as is the spermine complex of the differentlysequenced d(GTGTACAC)2. spermine di[!-aminopropyl]-tetramethylene diamine C10H26N4 H2N–(CH2)3–NH–(CH2)4–NH–(CH2)3NH2 spermidine

!-aminobutyl-!-aminopropylamine C7H19N3 H2N–(CH2)3–NH–(CH2)4–NH2

Scheme 5.4 Formulae of spermine and spermidine

Table 5.5. The group (space group P43212, Z ¼ 8) of isomorphous tetragonal octanucleotides, mostly native but with two examples of spermine complexation. The crystals have distorted right˚, A ˚ 3. Struchanded A-DNA-like duplex structures with Watson-Crick base pairs. Dimensions in A tures at RT unless stated otherwise Octanucleotide

Intercalated molecule

a

c

Cell Volume

NDB ID; Reference

d(CCCCGGGG)2

native

43.36

24.83

46683

d(CCCGCGGG)2 d(GCCCGGGC)2 d(GGCCGGCC)2 d(GGCCGGCC)2 d(GGCCGGCC)2 d(GGGCGCCC)2

265K 288K 255K 293K

native native native native native native

41.77 43.25 42.06 42.04 40.51 43.28

25.15 24.61 25.17 25.09 24.67 24.66

43880 46035 44527 44343 40485 46192

d(GGGCGCCC)2 at 115K

native

42.74

24.57

44882

d(GTGCGTAC)2 d(GTCTAGAC)2 d(GTGTACAC)243W d(CCCTAGGG)2 d(CTCTAGAG)2 d(GTACGTAC)2

native native spermine native native native

42.22 42.56 42.43 42.22 42.52 42.50

25.07 24.41 24.75 24.90 24.33 24.79

44688 44215 44558 44382 43892 44777

ADH012; HSWR87 ADH056; ES90 ADH0106; FSV97 ADH008; HLFB87 ADH013; WFBR82 ADH058; ES95 ADH098; WFBR82 ADH026; RHES88; SGGEFR89 ADH027; ADH057; EHHFSR88; SGGEFR89 ADH047; BLZS92 ADH041; CLH92 ADH014; JZS89 ADH078; TS96 ADH020; HLK89 ADH023; CDH90

at at at at

ISOMORPHISM AND POLYMORPHISM OF A-DNA OCTANUCLEOTIDES

159

Table 5.5. (Continued ) Octanucleotide


a

c

Cell Volume

NDB ID; Reference

d(GTACGTAC)2 at RT d(ACGTACC[5-BrU])283W at 100K; C72H81BrN30O32P7 d(ACGTACGT)2 d(ATGCGCAT)243W d(ATGCGCAT)243W

native native

42.32 43.597

25.04 26.268

44 846 49 927.5

ADH024; T90 TAP99

native native spermine

42.84 42.41 42.53

24.80 24.90 24.92

45 515 44 785 45 075

ADH070; WACW96 ADH033; CBS90 ADH032; CBS90

References: BLZS92 – Bingham, Li, Zon and Sundaralingam, 1992; CBS90 – Clark, Brown, Sanderson, Chwalinski, Neidle, Veal, Jones, Wilson, Garman and Stuart, 1990; CDH90 – Courseille, Dautant, Hospital, Langlois d’Estaintot, Precigoux, Molko and Teoule, 1990; CLH92 – Cervi, Langlois d’Estaintot and Hunter, 1992; EHHFSR88 – Eisenstein, Hope, Haran, Frolow, Shakked and Rabinovich, 1988; ES95 – Eisenstein and Shakked, 1995; FSV97 – Fernandez, Subirana, Verdaguer et al. 1997; HLFB87 – Heinemann, Lauble, Frank and Blo¨cker, 1987; HLK89 – Hunter, Langlois d’Estaintot and Kennard, 1989; HSWR87 – Haran, Shakked, Wang and Rich, 1987; JZS89 – Jain, Zon and Sundarlingam, 1989; RHES88 – Rabinovich, Haran, Eisenstein and Shakked, 1988; SGGEFR89 – Shakked, Guerstein-Guzikevich, Eisenstein, Frolow and Rabinovich,1989; T90 – Tagusagawa, 1990; TAP99 – Todd, Adams, Powell, Wilcock, Thorpe, Lausi, Zanini, Wakelin and Cardin, 1999; WACW96 – Wilcocks, Adams, Cardin and Wakelin, 1996; WFBR82 – Wang, Fujii, van Boom and Rich, 1982.

Both phases of true polymorphs of oligonucleotides must have the same chemical composition, including base pair sequence (known from the method of preparation). True polymorphism of oligonucleotides appears to be quite rare; native d(CCCGCGGG)2 is one of the few examples encountered among the oligonucleotides considered here. The two polymorphs crystallize in space group P43212 (Z ¼ 8, cell volume ˚ 3 (Table 5.7; ˚ 3 (Table 5.5; ADH0106)) and P21212 (Z ¼ 4, cell volume 42 479 A 43 880 A ADH103–5). The similar cell volumes suggest that the chemical compositions may well be the same. The octamer d(CGCTAGCG) crystallizes in two forms – orthorhombic (P212121, 24.77 ˚ , cell volume 119 187 A ˚ 3, 3 duplexes and 34 waters in the asymmetric unit) 41.52 115.89 A ˚ ˚ 3, four duplexes in the asymand hexagonal (P61, 48.7 115.9 A, cell volume 238 052 A metric unit) (Tereshko, Urpi et al., 1996). The structure of the first of these has been reported but not (yet) that of the second, although the comment was made that the packing in the two forms was very similar. As the ratio of the cell volumes is 1 : 2, the two forms could well be polymorphs. We first consider the behaviour of the octanucleotides as a group and then discuss the sub-groups in more detail. Six different groups of isomorphous crystals have been reported: ˚ 3; tetragonal, P43212, 18 examples; unit cell volume 44 000 A 3 ˚ ; hexagonal A, P61, 7 examples, unit cell volume 80 000 A ˚ 3; hexagonal B, P6122, 3 examples, unit cell volume 72 000 A ˚ 3; trigonal, R3, 1 example, unit cell volume (hexagonal cell) 230 000 A orthorhombic A, 4 examples (all with the same base pair sequence) P212121, unit cell ˚3 volume 41 000 A ˚ 3). 6. orthorhombic B, P21212, 1 example, unit cell volume 42 000 A 1. 2. 3. 4. 5.

160


Table 5.6. Two different groups of isomorphous hexagonal octanucleotides (space groups P61 and P6122, both Z ¼ 6) and one trigonal group (space group R3), all native. The crystals have distorted right-handed A-DNA-like duplex structures with Watson-Crick base pairs. All of the known P61 structures are of sequence d(GGX4CC) with the four central bases comprising C, G, A, T, U and I. ˚, A ˚ 3. Structures at RT unless stated otherwise. Note that, in contrast to Table 5.3, Dimensions in A the P61 and P6122 structures here are quite different Octanucleotide Hexagonal A P61, Z ¼ 6 d(GGGGCCCC)2 d(GGGATCCC)2 d(GGGTACCC)2 d(GGGTACCC)2 at 100K d(GGGTGCCC)2 d(GGTATACC)2 d(GGCATGCC)2 d(GG[BrU]A[BrU]ACC)2


a

c

Cell Volume

NDB ID; Reference

Native (1) native native native

45.32 46.83 46.80 46.06

42.25 44.49 44.52 44.09

82 530 84 494 84 446 81 006

ADH006; ADH007; ADH030; ADH031;

native native native native

45.62 44.97 46.29 45.05

40.99 41.76 42.97 41.72

81 133 80 056 79 739 73 425

RHES88 ADH010; SRK85 ADH076; NN97 SRC81; SRK85; DB89

32.40 32.18 32.34

79.25 78.51 78.49

72 048 70 409 71 093

ADH078; JS89 ADH038; TLBS93 ADH039; TLBS93

Hexagonal B P6122 Z ¼ 6 d(GTGTACAC)2 Native (2) d(GTGTACAC)2 From spermine (2) d(GTGTACAC)2 From spermidine (2)

MBK85 LFBH88 EFSR90 EFSR90

Notes: 1. 106 ordered solvent molecules were found around each double helix; if a density of 1.5 g cm3 is assumed, then there should be another 260 disordered solvent molecules per asymmetric unit. See Fig. 5.14 for the d(GTGCGCAC)2 packing diagram. 2. These three crystals are essentially identical; neither spermine nor spermidine was detected in the structure analysis. The best results were for the crystal grown in the presence of spermine. References: CLH92 – Cervi, Langlois d’Estantoit and Hunter, 1992; DB89 – Doucet, Benoit, Cruse, Prange and Kennard, 1989; EFSR90 – Eisenstein, Frolow, Shakked and Rabinovich, 1990; JS89 – Jain and Sundaralingam, 1989; LFBH88 – Lauble, Frank, Blo¨cker and Heinemann, 1988; MBK85 – McCall, Brown and Kennard, 1985; NN97 – Nunn and Neidle, 1997; RHES88 – Rabinovich, Haran, Eisenstein and Shakked, 1988; SRC81 – Shakked, Rabinovich, Cruse, Egert, Kennard, Sala, Salisbury and Viswamitra, 1981; SRK85 – Shakked, Rabinovich, Kennard, Cruse, Salisbury and Viswamitra, 1985; TLBS93 – Thota, Li, Bingham and Sundaralingam, 1993; TS96 – Tippin and Sundarlingam, 1996.

The tetragonal, hexagonal A, orthorhombic A and orthorhombic B unit cell volumes ˚ 3; the hexagonal B and trigonal unit cell volumes are are all around 43 000 or 80 000 A related by a factor of 3, but do not fit with the first group. It seems reasonable to infer that there are two separate groups of nominally heteromorphic structures – one comprising tetragonal, hexagonal A, orthorhombic A and orthorhombic B, and the second hexagonal B and trigonal. The orthorhombic A group has been noted by its investigators (Fernandez et al., 1997) as showing an unusual spread in cell dimensions. Also, the same sequence (d(CCCGCGGG)2) crystallizes in tetragonal (ADH0106;


161

Table 5.7. Two different groups of isomorphous orthorhombic octanucleotides; the space groups are P212121 and P21212; both Z ¼ 4. All crystals have distorted right-handed A-DNA-like duplex ˚, A ˚ 3. Structures at RT unless stated structures with Watson-Crick base pairs. Dimensions in A otherwise Octanucleotide

Intercalated drug

a

b

c

V

NDB code and Reference

Orthorhombic A P212121; Z ¼ 4 d(CCCGCGGG)2 native d(CCCGCGGG)2 native d(CCCGCGGG)2 native d(CCCGCGGG)2 native

21.84 23.13 23.14 24.63

35.40 40.82 40.12 40.59

41.17 42.52 41.90 42.49

31 830 40 146 38 899 42 479

ADH0102; ADH0103; ADH0104; ADH0105;

Orthorhombic B P21212; Z ¼ 4 native d(GTACGTAC)2

38.60

50.82

21.74

42 227

ADH059; LDCP93

FSV97 FSV97 FSV97 FSV97

References: FSV97 – Fernandez, Subirana, Verdaguer, Pyshnyi, Campos and Malinina, 1997; LDCP93 – Langlois d’Estantoit, Dautant, Courseille and Precigoux, 1993.

A13

A13

T12

T12 T4 A5

T4 A5

Fig. 5.12. Stereopair of the d(GTGTACAC)2spermine complex, viewed into the major groove down the dyad axis. The bonds of the spermine are emphasized. This is hexagonal B structure ADH039 (Table 5.6). (Reproduced from Jain, Zon and Sundaralingam, 1989.)

Table 5.5) and orthorhombic A (ADH0105; Table 5.7) structures with cell volumes of ˚ 3 and 42 479 A ˚ 3 respectively. Here the difference fits the usual behavior of 43 800 A polymorphic phases. Another same-sequence pair (GTACGTAC) crystallizes as tetragonal and orthorhombic B forms; these could also be a pair of polymorphs. The two structures have been compared in detail by Langlois d’Estantoit, Dautant, Courseille and Precigoux (1993). One marked difference is in the bend angle of the duplexes in

162


the different environments – this is 69 in the tetragonal form and 39 in the orthorhombic B form. The sequence d(ATGCGCAT)2 forms isomorphous tetragonal native (ADH033) and spermine-complex (ADH032) crystals; tetragonal d(GTGTACAC)2 forms only a spermine complex (ADH014; Table 5.5). The hexagonal B d(GTGTACAC)2 sequence forms isomorphous native, spermine and spermidine complexes (Table 5.6; Fig. 5.12). In all these examples, there is no obvious effect of complexation on cell dimensions; the guests were found by structure analysis. The spermine (spermidine) complexes are isomorphous with the native duplexes and thus should be considered as primary solid solution phases. It is not known whether the (relative) spermine content can range from zero (native duplex) to one (maximum spermine content of isomorphous complex). Some interesting features appeared in the first structure determination in this area – that of tetragonal d(GGCCGGCC)2 by Wang, Fujii et al., 1982; Table 5.5. The detailed conformation of the duplex was temperature dependent, approximating more closely to B-DNA at 265K and to A-DNA at 255K (Fig. 5.13). This is a theme that has resonated since the early papers, and composition (base pairs and formation of complexes), sequence, hydration state and environment are now recognized as important factors in determining the details of oligonucleotide structures (Shakked, 1991). Our discussion uses primarily information from cell dimensions and space groups but much more detailed studies based on information gleaned from crystal structures

Minor groove

Minor groove A–DNA

a

–8°C

–18°C

b

c

8–DNA

d

Fig. 5.13. (upper row) Space filling diagrams of four DNA models, all with the sequence GGCCGGCC; the P and O atoms are emphasized. The helix axis is vertical, and a horizontal twofold axis is located in the plane of the paper; (lower row) skeletal views down the helix axis. The complexes are tetragonal, space group P43212, details in Table 5.5. (Reproduced from Wang, Fujii et al., 1982.)


163

is becoming available. An example is the study of hydration patterns and intermolecular interactions in the isomorphous tetragonal octameric A-DNA complexes with sequences GGGCGCCC (at 293 and 115K), GGCCGGCC, CCCCGGGG and GCCCGGGC (Eisenstein and Shakked, 1995). We quote from the Abstract: ‘‘The A-DNA major groove is extensively hydrated and together with the hydration shells of the sugar–phosphate backbone can form an ordered network of fused polygons. The water structure of the phosphate backbone is less conserved than that of the grooves. Characteristic hydration patterns are associated with specific base sequences. The A-DNA minor groove provides sites for intermolecular contacts through hydrophobic and polar interactions. Well-ordered water molecules mediate interduplex interactions that involve either the grooves or the backbone, or both. The direct and water-mediated intermolecular interactions observed in the A-DNA crystal structures are relevant to various recognition motifs between DNA and other molecules.’’

Fig. 5.14. The arrangement of d(GTGCGCAC) duplexes in the tetragonal crystal (space group no. 96, P43212) projected down the [001] axis. The elliptical solvent channels run along the vertical ˚ in cross-section; these are filled with solvent molecules 21 axes and are approximately 20 10 A that are not shown. The only direct interaction between neighboring duplexes involves the abutment of the terminal base pairs of one molecule against the sugar phosphate backbone of symmetry related molecules. Note that ‘‘fourfold axis’’ in the original caption should be replaced by ‘‘twofold axis’’. (Reproduced from Wang, Fujii et al., (1982)).

164


c

b

a

Fig. 5.15. Projection of the structure of d(GG(BrU)A(BrU)CC)2 down [001] axis (hexagonal A, Table 5.6). The octamer crystallizes in space group P61 with one double helix in the asymmetric unit. The octamers have the A-DNA conformation with 10.9 base pairs per turn, a mean tilt ˚ . Consideration of the angle of 18 and an average separation between base pairs of 3.4 A diffuse scattering shows that the channels contain a me´lange of water molecules and B-DNA. A companion diagram of the isomorphous d(GGGGCCCC)2 given by McCall, Brown and Kennard (1985) (not reproduced here) shows six octameric molecules in an infinite spiral viewed in projection down the 61 screw axis. (Reproduced from Doucet, Benoit, Cvruse, Prange and Kennard (1989).)

Table 5.8. Two different crystal types of 1:2 octamer–distamycin complexes. All crystals have distorted right-handed B-DNA-like duplex structures with modified Watson-Crick base pairs. Structures at RT unless stated otherwise Octameric oligonucleotide

Intercalated drug

a

c

V

Tetragonal P4122, Z ¼ 8; asymmetric unit contains one DNA strand, one distamycin molecule and one Mg cation d(ICICICIC)2 distamycin 27.93 58.62 45 729 d(ICATATIC)2 distamycin 27.86 58.62 45 500 d(ICITACIC)2 distamycin 28.03 58.04 45 601 d(IcICICIC)2 distamycin 27.92 57.70 44 979 d(IcIcICIC)2 distamycin 27.93 57.47 44 799

NDB code / Reference

GDHB25; CRRS94 GDLB50; CRS97 GDLB51; CRS97 CRS95 CRS95


165

Table 5.8. (Continued ) Octameric oligonucleotide

Intercalated drug

a

b/

c

V

Monoclinic I2; Z ¼ 2; asymmetric unit contains one DNA strand and one distamycin molecule d(ITITACAC)236W distamycin 28.30 25.05 29.14 19 509 109.2 distamycin 28.11 25.33 30.88 20 491 d(ICATATIC)2 111.26

NDB code / Reference

DD0043; DPS03 GDLB49; CRS97

Notes: Lower case letters represent RNA residues. The unit cells are given for the two I2 structures; these are isomorphous. The space group used in the original publication was C2.

OH N

N

N

N

CH2OH

O H

H HO

OH

I is inosine (hypoxanthine riboside) References: CRRS94 – Chen, Ramakrishnan, Rao and Sundaralingam, 1997; CRS95 – Chen, Ramakrishnan and Sundaralingam, 1995; CRS97 – Chen, Ramakrishnan and Sundaralingam, 1997; DPS03 – Deng, Pan and Sundaralingam, 2003.

Exceptionally, the packing arrangement has been discussed in some detail for this group of isomorphous tetragonal complexes (Table 5.5). Analogous packing diagrams have been given by Bingham, Li, Zon and Sundaralingam. (1992) and Wang, Fujii et al. (1982; their Fig. 3 reproduced in Fig. 5.14). The packing in the hexagonal A family (Table 5.6) is shown in Fig. 5.15; the arrangement of spirals of octameric duplexes should be compared with the left hand diagram in Fig. 5.20 showing a spiral arrangement of decanucleotides. The synchrotron diffraction patterns from d(GG[BrU]A[BrU]ACC)2 have both a (sharp) Bragg and a diffuse component (Doucet, Benoit, Cruse, Prange and Kennard, 1989); the Bragg pattern gives an arrangement of A-DNA helices isomorphous with the other structures in Table 5.6 while the diffuse pattern was only compatible with a disordered array of B-DNA helices contained in the A-DNA framework These replace about 10%

166


of the water molecules found in the other structures. These unusual results were interpreted to show that ‘‘under certain conditions of crystallization the A and B conformations of the same DNA fragments can co-exist in the highly hydrated environment of the crystal lattice.’’ Indeed the arrangement described could be classified as a tunnel inclusion complex (Chapter 6), with some resemblance to neat selenourea. Some octanucleotides form minor-groove binding complexes with the natural antibiotic distamycin (Scheme 5.5). Two different crystal types (tetragonal, and monoclinic) have been found (Table 5.8). The molecular structure of the complex is shown in Fig. 5.16; both phases have essentially the same host-guest structure. The asymmetric unit in the monoclinic crystals consists of one octamer strand and one distamycin molecule; in the tetragonal crystals one double helix and two distamycin molecules. The distamycin complex of d(ICATATIC)2 crystallizes in the two space ˚3 groups P4122 and C2 (Fig. 5.17 a and b), with cell volumes of 45 500 and 20 747 A respectively; the ratio is 2.2 and it is possible (but not certain) that the two phases are heteromorphs.

Fig. 5.16. Stereodiagram of the 1 : 2 d(ITITACAC)2 complex with distamycin, showing the antiparallel side-by-side arrangement of the two distamycin molecules in the minor groove (monoclinic, Table 5.8). Complexed hairpin dimers have been shown by Kielkopf et al. (1998) (see Fig. 5.19). An example of a single propamidine molecule in the minor groove of the dodecameric d(CGCGAATTCGCG)2 duplex is shown in Fig. 5.21. (Reproduced from Deng, Pan and Sundaralingam, 2003.)

MI NOR GR OOVE BI NDE RS

167

(a)

(b)

Fig. 5.17. Stereo diagrams for (a) the tetragonal d(ICITACIC)2 structure (GDLB51) and (b) the monoclinic B d(ICATATIC)2 structure (GDLB49). The tetragonal crystals, viewed down [010] contain Mgþþ(H2O)6 cations, appearing in the form of asterisks. These ions are absent from the monoclinic crystals. This supports the conclusion from comparison of cell volumes that these two modifications are not heteromorphs (see text). The tetragonal crystals have sheets of quasi-cylindrical complex molecules in (004) planes with residual channels containing disordered solvent molecules. In the monoclinic B crystals the molecules are arranged in sheets on (011) planes. (Reproduced from Chen, Ramakrishnan and Sundaralingam, 1997.)

5.6 Minor groove binders 5.6.1 Drug molecules that enter the minor groove The general background has been given by Tidwell and Boykin (2003) and Moravek, Neidle and Schneider (2002); as noted earlier, in principle molecules can bind to either the major or minor grooves of DNA, the major groove being the binding site for many DNA interacting proteins while the minor groove is suited to small molecule binding. The oligonucleotides that complex with minor groove binders are decamers and dodecamers. These are considered separately – the decamers present a rather complex and diffuse

168


pattern of structures, while the pattern for dodecamers – native species and complexes alike – is much more compact. We first group together in Scheme 5.5 the minor groove binders that have been reported. The curved profile of these molecules, adapted to the general shape of the Scheme 5.5. Drug molecules that enter the minor groove N N R

N

2O

N

H

OR1

H H3C

N+

R2 =

R1 = H Hoechst 33 258

H

R1 = ethyl

Hoechst 33 342

N R1 = H Compound 16 N H NH2

H

+ H2N

NH2

N 6

4⬘

+ NH2

DAPI 4⬘,6-diamidino-2-phenylindole

CH3

O

N Guanidinium O

amide 2 N

N

CH3

H N H amide 1

N

H

Netropsin

O H

N

Amidinium NH2

H2N

+

+

amide 3 NH2

NH2


CH3

169

O

N N N

O

CH3

N

H

N O

H N

N NH2 +

+

H2N

H

H Imidazole-Pyrrole Lexitropsin

NH2

NH2 CH3

O

N

amide 3 N

O

CH3

N

H N

amide 2

H H

N

Distamycin O

amide 4 +

amide 1 N

O

NH2

NH2

H H

Naturally occurring netropsin has two pyrrole rings; the guanidinium head at the left and the amidinium tail at the right are both positively charged. Imidazole-pyrrole lexitropsin is identical to netropsin except that the left pyrrole ring has been replaced by imidazole. Distamycin has an uncharged –CHO head on the left and charged amidinium tail on the right. (after Goodsell, Ng et al., 1995).

O

O R

NH2

H2N +

+ NH2

R = trimethylene; propamidine R = pentamethylene; pentamidine H2 C

R = gamma-oxa-pentamidine C H2

H2 C O

C H2

NH2

170


H N

N N H2N

NH2

berenil +

+

NH2

NH2

O NH2

H2N +

+ NH R

R = H; furamidine R = ethyl; BEF R = isopropyl; BIPF R = cyclopropyl; BCPF

HN R

N N N H3N+

N H

N H N

TRIBIZ N H

O H3C

minor groove, is quite striking. Aromatic diamidines have broad spectrum effectiveness against a range of microbial species. However, so far only pentamidine, despite some negative features, has found clinical use. For reasons of space we hardly consider situations where the drug is covalently linked to the oligonucleotide (e.g. Kopka, Goodsell, Baikalov et al., 1994).


imidazole Im1 H N

hydroxy-pyrrole Hp2 pyrrole Py3

O H N

N

171

N H

O

H N

N H

amide Am1

OH

amide Am2

pyrrole Py4 O

H N R

N H amide Am3

O R= N H

O

HN + N H

CH3 CH3

ImHpPyPy Dp The polyamide sequence shown is ImHpPyPy Dp where is -alanine and Dp is dimethylaminopropylamide. The other guest is ImPyPy Dp (i.e. hydroxyl of Hp2 replaced by H).

5.6.2 Decameric oligonucleotides The decameric oligonucleotides present a rather complicated structural picture, summarized some years ago by Dickerson, Goodsell and Kopka (1996) for fourteen examples. At the time of writing (October, 2004) some twenty-odd decameric structures have been reported (including one complex) and they are found in six different groups (Table 5.9). ˚, 1. Nine are in a monoclinic cell, space group C2, Z ¼ 2, a 32.3, b 25.5, c 34.4 A 3 ˚ ¼ 113 , V ¼ 26 000 A , with one strand per asymmetric unit. This has been termed the CG family by Quintana, Grzeskowiak et al. (1992); decamers with the sequence type CCARxxYTGG tend to belong to this group. ˚ , V 54 000 A ˚ 3, 2. Two are in space group P212121, Z ¼ 4, a 36.6, b 42.5, c 34.7 A with one helix per asymmetric unit. The native and drug complex pair are isomorphous so the drug complex is a primary solid solution phase. There are two di-imidazole lexitropsin drug molecules side-by-side in the minor groove. 3. Four are in another orthorhombic cell, space group P212121, Z ¼ 4, a 38.9, b 39.4, ˚ , V ¼ 51 000 A ˚ 3, with one helix per asymmetric unit. This has been termed c 33.3 A the KK family by Quintana, Grzeskowiak et al. (1992); decamers with the sequence type CGAxxxxTCG tend to belong to this group.

˚,A ˚3 Table 5.9. Decameric oligonucleotides; all crystals are distorted right-handed B-DNA-like duplexes with Watson-Crick base pairs. Dimensions in A Decameric Oligonucleotide

Drug in the minor groove

Group I: monoclinic, C2, Z ¼ 2; CQ family Native d(CCAGGCCTGG)2, Native, from d(CCAACITTGG)2, 248K Ca2þ solution native d(CCAAGATTGG)2, d(CCAACGTTGG)2, native native d(CCAACGTTGG)2, native d(CCAGTACTGG)2, d(CCAGCGCTGG)2, native native d(CCAAIATTGG)2, d(CCAGGCCTGG)2, native Group II: orthorhombic, P212121, Z ¼ 4 d(CATGGCCATG)2, native d(CATGGCCATG)2, di-imidazole lexitropsin Group III: orthorhombic, P212121, Z ¼ 4; KK family native d(CGATCGATCG)2, d(CGATTAATCG)2, native

a

b/

c

Cell Volume

NDB ID/Reference

31.25 31.87

25.49 116.7 25.69 114.1

34.82 34.21

24 779 25 833

HA89 LKK93

32.52 32.25 32.00 31.73 32.24 32.21 32.15

26.17 25.53 25.37 25.79 25.35 25.14 25.49

34.30 34.38 33.63 34.22 34.19 34.14 34.82

25 556 25 978 26 737 27 022 26 933 25 116 25 490

PHC87; PYD91 PYD91; YPD91 BD0033; CD00 BD0023; KDKR00 BD0035; CD00 BD0055; LLKKD BDJ017; HA89

36.60 36.65

42.49 42.64

34.69 34.68

53 948 54 196

BDJ051; CKCD93 GDJ054; KGH97

38.93 38.60

39.36 39.10

33.30 33.07

51 025 49 911

BDJ025; GYPD91 BDJ031; QGYD92

118.9 113.4 112.98 116.9 117.17 114.70 116.71

d(CGATATATCG)2, d(CGATATATCG)2,

Native (Ca) Native (Mg)

38.76 38.69

40.06 39.56

33.73 33.64

52 373 51 489

YQD92; BDJ036 BDJ037; YQD92

Group IV: hexagonal, P6, Z ¼ 6 d(CCAGGC[5-MeC]TGG)2 d(CCAAGCTTGG)2

native Native (Ca)

53.77 53.08

53.77 53.08

39.35 34.32

98 527 83 741

HA91, HH92 BDJ052; GGK93

Group V: hexagonal, P61, Z ¼ 6 d(CCAAGCTTGG)2

native

45.32

45.32

42.25

75 151

MBK85

Group VI: trigonal, P3221, Z ¼ 6 d(CGATCG(5-MeA)TCG)2 d(CCATTAATGG)2 d(CCACTAGTGG)2 d(CCAACITTGG)2

native native native Native (Mg)

33.38 33.20 32.90 33.23

33.38 33.20 32.90 33.23

98.3 96.04 95.10 94.77

94 001 91 677 89 146 90 628

BGY93 BDJ055; GKD94 BDJ061; SG-G94 LKK93

References: BGY93 – Baikalov, Grzeskowiak, Yanagi, Quintana and Dickerson, 1993; CD00 – Chiu and Dickerson, 2000; CKCD93 – Goodsell, Kopka, Cascio and Dickerson, 1993; GKD94 – Goodsell, Kaczor-Grzeskowiak and Dickerson, 1994; GYPD91 – Grzeskowiak, Yanagi, Prive´ and Dickerson, 1991; GGK93 – Grzeskowiak, Goodsell, KaczorGrzeskowiak, Cascio and Dickerson, 1993; HA89 – Heinemann and Alings, 1989; HA91 – Heinemann and Alings, 1991; HH92 – Heinemann and Hahn, 1992; KDKR00 – Kielkopf, Ding et al., 2000; KGH97 – Kopka, Goodsell, Han, Chiu, Lown and Dickerson, 1997; LKK93 – Lipanov, Kopka, Kaczor-Grzeskowiak, Quintana and Dickerson, 1993; LLKDD – Lisgarten, Lipanov, Kopka et al., to be published; NDB BD0055; MBK85 – McCall, Brown and Kennard, 1985; PHC87 – Prive´, Heinemann, Chandrasegaran, Kan, Kopka and Dickerson, 1987; PYD91 – Prive´, Yanagi and Dickerson, 1991; QGYD92 – Quintana, Grzeskowiak, Yanagi and Dickerson, 1992; SG-G94 – Shakked, Guzikevich-Guerstein et al., 1994; YPD91 – Yanagi, Prive´ and Dickerson, 1991; YQD92 – Yuan, Quintana and Dickerson, 1992.

174


˚ , V 84 000 A ˚ 3, 4. Two are hexagonal, space group P6, Z ¼ 6, a 53.1, c 34.3 A with one duplex per asymmetric unit. 5. One (d(CCAAGCTTGG) 2) is in a P61 hexagonal cell, Z ¼ 6, a 45.32, c ˚ , V 75 000 A ˚ 3, This same sequence is found in P6 but the difference in cell 42.25 A volumes suggests that these are not polymorphs. ˚ , V 95 000 A ˚ 3, 6. Four are trigonal, space group P3221, Z ¼ 6, a 33.4, c 98.3 A with one helix per asymmetric unit. None of these cells match those given in Tables 5.10 and 5.11 for different groups of decamer-drug complexes. Using volume per asymmetric unit as a criterion, one may surmise that Groups I, II, III and V are heteromorphic; all have volumes per asymmetric ˚ 3. There is no evidence allowing a distinction between enantiotropes unit around 13000 A and monotropes. When molecules are flexible, different conformers are often found in different polymorphs. For DNA oligomers, base pair sequence can affect conformation; additionally, solvent and ion arrangement play an important role in determining crystal structures, which consequently are not easily accounted for. One path towards unscrambling the various factors is to examine ‘‘different sequences in the same environment and the same sequence in different environments’’ (Lipanov, Kopka et al., 1993). In terms of our definitions, ‘‘the same sequence in different environments’’ implies the possibility of true polymorphism, while ‘‘different sequences in the same environment’’ implies the possibility of heteromorphism. In both situations, the two phases should have the same overall composition. The data in Table 5.9 provide some leads. Baikalov et al. (1993) have noted that the two orthorhombic phases and the monoclinic phase have similar packing densities while the ˚ 3 in the trigonal phase is more loosely packed; the ‘‘molecular’’ volumes are about 6250 A first two groups and about 10% larger in the third, presumably due to differences in composition. In orthorhombic group III of Table 5.9, the d(CGATATATCG)2 phase has only slightly different cell dimensions when grown from solutions containing Mg2þ or Ca2þ. However, d(CCAACITTGG)2 gives a monoclinic phase from a Ca2þ solution and a trigonal phase from a Mg2þ solution. Because of the composition differences, these are not polymorphs, nor even heteromorphs – the monoclinic phase has five base pairs, one cacodylate ion7, 72 waters and one hepta-coordinated Ca2þ in the asymmetric unit, while the trigonal phase has ten base pairs, 36 water molecuiles and one octahedral Mg2þ in its asymmetric unit. Perhaps surprisingly, the trigonal phase has the simpler chemical composition. Clearly care is needed when making generalizations based on comparison of unit cell volumes and inferring structural arrangements about metal ions. The packing in the KK (d(CGAxxxxTCG)2) family is shown in Fig. 5.18. The decamers stack one on top of the other along the c axis to form pseudo-continuous helices arranged in square array, with extensive lateral contacts along the a and b directions. Besides extensive van der Waals interactions and hydrogen bonding, the octahedral Mg2þ(H2O)6 cations play a role in the cohesion; details are given by Quintana, Grzeskowiak, Yanagi and Dickerson (1992). Polyamide hairpin dimers containing the aromatic groups imidazole, hydroxypyrrole and pyrrole provide a means, through analysis of hydrogen bonding, for discriminating 7

Cacodylic acid is tetrahedral (CH3)2As¼O(OH).


175

Fig. 5.18. Stereodiagram of four columns of stacked KK-type decamer helices viewed along b, with a from left to right and c from bottom to top. A skeletal representation is given above and a spacefilling representation below. This diagram illustrates how the minor grooves join to build continuous diagonal channels through the crystal. (Reproduced from Quintana, Grzeskowiak, Yanagi and Dickerson (1992).)

among the four Watson–Crick base pairs. An imidazole/pyrrole (Im/Py) pair (the abbreviations are shown above in Scheme 5.5) distinguishes G-C from C-G and both of these from A-T and T-A, as has been demonstrated through the structure of decameric {d(5 0 -CCAGGCCTGG-3 0 ) (ImImPyPy--Dp)} (Kielkopf, Baird, Dervan and Rees, 1998). The distinction between A-T and T-A has been made through a detailed comparison of hydrogen bonding and other interactions in the ImHpPyPy--Dp Hp/Py and ImPyPyPy--Dp decameric oligonucleotides (i.e Hp/Py compared to Py/Py) (Kielkopf, White, Szewczyk et al., 1998) (crystal data in Table 5.10). The particular relevance here is that a second type of side-by-side packing of the polyamide in the minor groove has been

176


Table 5.10. Binding in the minor groove – the decameric oligonucleotides are all distorted right˚ and A ˚ 3. The cells handed B-DNA-like duplexes with Watson-Crick base pairs. Dimensions are in A are all reduced and the crystallography is discussed in the text Decameric Oligonucleotide

Drug in the minor groove/spce group

a

b/

c

Cell Vol.

Reference

25.7 116.9 26.90 98.91 30.51 103.09 30.20 101.09

34.2

24848

KDKR00

53.62

61987

KBDR98

53.48

67961

51.49

66418

BDD002; KW98 BDD003; KW98

d(5 0 -CCAGGCCTGG-3 0 )2

Native; C2

31.7

d(5 0 -CCAGGCCTGG-3 0 )2

ImImPyPy--Dp;I2

43.50

d(5 0 -CCAGTACTGG-3 0 )2

ImHpPyPy--Dp;I2

42.76

d(5 0 -CCAGTACTGG-3 0 )2

ImPyPyPy--Dp; I2

43.46

Notes: The 5-bromouridine substituted crystal was isomorphous and crystallized in space group C2 (explicit cell dimensions were not given). 5-Bromocytosine substitution leads to crystallization in space group P212121 ˚ , cell volume ¼ 59 567 A ˚ 3 (KBDR98); this could be isostructural with with a ¼ 34.4, b ¼ 39.0 and c ¼ 44.4 A orthorhombic Group II of Table 5.9. The first entry is a member of the CQ family (Table 5.9). References: KBDR98 – Kielkopf, Baird, Dervan and Rees, 1998; KDKR00 – Kielkopf, Ding, Kuhn and Rees, 2000; KW98 – Kielkopf, White et al., 1998.

demonstrated (Fig. 5.19). The structure of the d(5 0 -CCAGGCCTGG-3 0 ) (native) oligo˚ nucleotide was determined, at liquid nitrogen temperature, to the high resolution of 0.74 A ˚ . This made it possible to identify alternative using synchrotron radiation with ¼ 0.78 A conformations for phosphates, calcium ions and networks of water molecules; also some hydrogen atoms of base pairs could be identified. The oligonucleotide sequence d(5 0 -CCAGGCCTGG-3 0 )2 has already been met in ˚ 3), Table 5.9, the native form crystallizing in space group C2 (unit cell volume 24 779 A and with dimensions only slightly different from those of the first entry in Table 5.10; presumably these are essentially the same material, The following three entries can be classified as isostructural. Reduction of the cells gives changes in cell dimensions, bringing the values closer to 90 , and a change of space group from C2 to I2. The complexes are separate phases in the phase diagram. End-to-end binding of netropsin is found (Chen, Mitra et al., 1998; GDJ059) in the minor groove of decameric d(CCCCCIIIII)2 which crystallizes in a triclinic cell (P1, Z ¼ 1, ˚ 3,). The structural unit is ˚ , 86.30, 84.50, 68.58 , V ¼ 36 992 A 32.56, 32.59, 37.64 A one decameric duplex plus two netropsin molecules; There are two such duplexes in the asymmetric unit. We have limited our coverage to duplexes with complete Watson–Crick base pairing but some exceptions are briefly mentioned. There is a group of decanucleotides (Table 5.11) that actually have an octameric duplex arrangement with the first and last bases of the sequence interacting with similar moieties in other octanucleotides. For example the decamer d(5 0 -GGCCAATTGG-3 0 )2 has a Watson–Crick base-paired B-DNA octamer duplex with the two terminal 5 0 -G and G-3 0 bases of each single strand lying within the minor groove of a symmetry-related duplex. This structure has similarities to


177

Fig. 5.19. Structure of the decanucleotide d(5 0 -CCAGGCCTGG-3 0 )2 complexed with polyamide, showing the antiparallel, side-by-side arrangement of the guests in the minor groove. This diagram appeared on the cover of Helvetica Chimica Acta in illustration of a paper by Marques et al. (2002); more details are given by Kielkopf, White et al. (1998).

that of the native sequence d(5 0 -CGCAATTGCG-3 0 ) (Spink, Nunn et al., 1995) that has a two fold symmetry axis through the centre of the duplex and one DNA strand in the asymmetric unit. Despite similarities in the unit cell dimensions of the native and complexed CGCCAATTCG species (entries 2 and 3 of Table 5.11), the crystals are not isomorphous because of the difference in symmetry and the complex is a separate phase in the phase diagram. There is primary solid solubility of DAPI in the decameric oligonucleotide GGCCAATTGG (entries 4 and 5 of Table 5.11); here the asymmetric unit consists of one duplex decameric oligonucleotide and two crystallographically independent (but geometrically similar) polyamide molecules. Another example just over our borderline is d(CGACGATCGT)2; the crystals are ˚ , ¼ 113.45 (Qiu, monoclinic, space group P21, a ¼ 26.45, b ¼ 34.66, c ¼ 32.17 A Dewan and Seeman, 1997), The crystallographic asymmetric unit contains one B-DNA

178


Table 5.11. Unit cells of a group of isomorphous decameric oligonucleotides with octameric Watson-Crick duplex structure. The drug molecules are located in the minor groove. The first pair ˚ and A ˚ 3. has space group I212121 and the others have space group P212121 Dimensions are in A Structures at room temperature unless stated otherwise Decanucleotide Space group I212121 d(5 0 -CG(5-IC)AATTGCG-3 0 )2 d(5 0 -CGCAATTGCG-3 0 )2 Space group P212121 d(5 0 -CGCAATTGCG-3 0 )2 d(5 0 -GGCCAATTGG-3 0 )2 116W2Mg2þ, at 120K d(5 0 -GGCCAATTGG-3 0 )2


a

b

c

Cell Vol.

NDB ID; Reference

Native Native (1)

27.52 27.01

38.64 39.26

53.49 54.01

57 273 57 273

SNVBN95 SNVBN95

Netropsin (2) Native

24.84 26.11

39.80 36.46

54.08 52.56

53 465 49 606

DAPI (2)

25.616

36.565

52.961

49 606

NGN97 BD0006; VTM99 VSM99

Notes: 1. Berenil was present in the crystallizing medium but was not found in the electron density maps whereas netropsin did appear. 2. For formulae of berenil, DAPI and netropsin see Scheme 5.5. References: NGN97 – Nunn, Garman and Neidle, 1997; SNVBN95 – Spink, Nunn, Vojtechovsky, Berman and Neidle, 1995; VSM99 – Vlieghe, Sponer and Van Meervelt, 1999; VTM99 – Vlieghe, Turkenburg and Van Meervelt, 1999.

double helix. The nucleotide has eight Watson–Crick base pairs and a two nucleotide 5 0 -sticky end at each end of the duplex. The sequence GGCCAATTGG crystallizes as a fully base-paired Watson–Crick duplex ˚ , 105.2 ; Wood, in a monoclinic cell (space group C2, reduced cell 38.29 24.70 61.84 A Nunn, Trent and Neidle, 1997) that is not isomorphous with the monoclinic cell of ˚ , 93.91 ; Goodsell, CTCTCGAGAG (space group C2, reduced cell 40.83 24.32 49.83 A Grzeskowiak and Dickerson, 1995), These, and other examples, have crossed helix structures that we shall not explore further. 5.6.3

Polymorphs or intermediate phases? An example from the decanucleotides

We have noted previously that it is often difficult to define the relationships among different crystal modifications because their compositions are not known. This has been illustrated above (Fig. 5.17) for the tetragonal d(ICITACIC)2 and monoclinic B d(ICATATIC)2 octanucleotide structures. A rather complete example of what is required is provided by Tippin and Sundaralingam (1997) for the crystals of the A-DNA decamers d(CCGGGCCCGG), d(CCGGCC[m5C]GG), d(C[m5C]GGGCC[m5C]GG) and d(CCGGGCC[m(Br)5C]GG). The crystal data are given in Table 5.12; there are three isomorphous groups–hexagonal, orthorhombic A and orthorhombic B. Packing diagrams are given in Fig. 5.20 from which it is clear that the packing density increases from hexagonal through orthorhombic A to orthorhombic B. The spermine content, determined from the structure analyses, also increases in the same order from zero molecules per double helix through 1 to 2. Although Tippin and Sundaralingam


179

Table 5.12. Crystal data for various modifications of decameric oligonucleotides, some complexed with spermine. All crystals have distorted right-handed A-DNA-like duplex structures with modified ˚, A ˚ 3. All data from Tippin and Sundaralingam (1997) Watson-Crick base pairs. Dimensions in A Decameric oligonucleotide

a

b

Hexagonal P61, Z ¼ 6; double helix to spermine ratio d(CCGGGCC[m5C]GG)2 55.03 55.03 54.71 54.71 d(C[m5C]GGGCC[m5C]GG)2 55.68 55.68 d(CCGGGCC[(Br)5C]GG)2 Orthorhombic A P212121; Z ¼ 4; d(CCGGGCCCGC)2 d(CCGGGCC[m5C]GG)2 d(CCGGGCC[(Br)5C]GG)2 d(C[m5C]GGGCC[m5C]GG)2

c

V

NDB code

1:0; solvent content 67% 45.88 120 317 45.82 118 770 46.24 124 147

double helix to spermine ratio 24.91 44.87 48.14 24.78 44.57 48.00 24.36 44.29 48.05 24.83 44.72 47.91

1:1; solvent content 40% 53 816 53 015 51 849 53 199

Orthorhombic B P212121; Z ¼ 4; double helix to spermine ratio 1:2; solvent content 24% 23.74 40.89 43.60 42 316 d(CCGGGCCCGC)2 d(CCGGGCC[m5C]GG)2 23.64 40.82 43.44 41 929

P61

Ortho 1

Ortho 2

Fig. 5.20. Unit cell views of the hexagonal and orthorhombic A and B modifications of d(CCGGGCC[m5C]GG)2. The hexagonal cell is seen down the c axis, with the 61 screw axis at the origin; both orthorhombic cells are seen down their a axes. The view of the hexagonal cell down the c axis should be compared with the arrangement of the d(GGGGCCCC)2 octamers given by McCall, Brown and Kennard (1985) (see caption to Fig. 5.15). The space group is P61 in both examples. (Reproduced from Tippin and Sundaralingam, 1997.)

(1997) use the term ‘‘polymorphic’’ in their title and list of keywords, it is clear that the three types are ‘‘native’’, ‘‘monospermine’’ and ‘‘dispermine’’ phase rule compounds, no different in principle from ‘‘anhydrate’’, ‘‘monohydrate’’ and ‘‘dihydrate’’. A more exact formulation would require knowledge of the degree of hydration. Note that the base pair sequence d(C[m5C]GGGCC[m5C]GG)2 occurs in hexagonal and orthorhombic A phases, as does the sequence d(CCGGGCC[(Br)5C]GG)2, The base pair sequence d(CCGGGCCCGC)2 occurs in the two orthorhombic phases while the sequence d(CCGGGCC[m5C]GG)2 occurs in all three phases.

Table 5.13. Native dodecameric oligonucleotides and the isomorphous complexes binding drugs in the minor groove. All crystals have space group ˚,A ˚ 3. We use ‘native’ while NDB uses P212121, Z ¼ 4 and are distorted right-handed B-DNA-like duplexes with Watson-Crick base pairs. Dimensions in A ‘plain’ Dodecameric Oligonucleotide


a

b

c

Cell Vol.

NDB ID; Reference

d(5 0 -CGCGAATTCGCG-3 0 )2 d(5 0 -CGCGAATTCGCG-3 0 )2 at 16K (see text) d(5 0 -CGCGAATTCGCG-3 0 )2 d(5 0 -CGCGAATTCGCG-3 0 )2 d(5 0 -CGCGAATTCGCG-3 0 )2 d(5 0 -CGCGAATTCGCG-3 0 )2 160W at 137K d(5 0 -CGCGAATTCGCG-3 0 )2 at 248K d(CGCGAATTCGCG)2220 W, at 100K d(CGCGAATTCGCG)2 54W d(CGCGAATTCGCG)2 73W d(CGCGAATTCGCG)279W d(CGCGAATTCGCG)2 d(CGCGAATTCGCG)2 d(CGCGAATTCGCG)272W d(CGCGAATTCGCG)2 d(CGCGAATTCGCG)2 at 300K; also at 273, 248 and 173K* d(CGCGAATTCGCG)2 d(CGCGAATTCGCG)2 d(CGCGAATTCGCG)236W

native native

24.87 23.44

40.39 39.31

66.20 65.26

66 498 60 132

BDL001; DrDi81 BDL002; DSD82; HDK85

native native native Spermine, Mg, Na

25.30 25.59 25.94 25.186

40.24 40.82 40.74 40.208

65.94 66.67 66.20 65.656

67 132 69 642 69 960 66 488

BD0005; BD0029; DrDi81 BD0041; SHMVW00 BD0054; HSVW01 BDL084; SMHW98

Monoimidiazole lexitropsin furamidine

24.03

39.26

66.30

62 549

GNKLD95

24.24

39.94

65.88

63 781

GSN98

furamidine propamidine pentamidine BIPF BCPF -oxapentamidine# 16 Hoechst-33258

25.28 25.00 24.37 24.60 25.43 24.69 24.59 25.04

40.69 40.88 40.00 40.07 40.66 40.33 40.44 40.33

66.73 67.28 66.07 65.45 66.13 66.20 65.76 65.85

68 641 68 760 64 405 64 516 68 377 65 918 65 393 66 500.

LTNB96 NJN934b EJN92 TCKW96 TCKW96 GDL027; NJN94a CB95 GDL006; QLD91

cisplatin cisplatin netropsin

24.36 24.33 24.27

40.05 40.08 39.62

66.13 66.26 63.57

64 518 64 613 61 127

d(CGCGAATTCGCG)2 d(CGCGAATTCGCG)2

berenil DAPI

24.51 25.25

39.98 40.71

66.23 66.53

64 899 68 659

#WPDD84 #WPDD84 KYG85a; SMR92b, GKD95 BSSJ90 TGC89

d(CGCGAATTCGCG)2

Cells with slightly different dimensions are given in DD0005 (SBC00), 6, 8, 9, 13 for ‘benzimidazole derivative’ see also WNTN95 d(CGCGAATTCGCG)2 Bis-(piperidino-ethyl)furamidine 25.70 40.73 66.38 69484 DD0025; NS d(CGCGAATTCGCG)2 Cells with slightly different dimensions are given in DD0034, 35 for ‘Bis-phenylfuran derivative’ d(CGCGAATT[5BrC]GCG)2 Native 24.71 40.56 65.62 65767 BDLB03; FKKDD82 d(CGCGAATT[5BrC]GCG)2 netropsin 24.27 39.62 63.57 61127 KYG85a; GKD95 d(5 0 -CGCATATAGCG-3 0 )2 native 23.54 38.86 66.57 60896 BDL007; YPGD88 d(5 0 -CGCAAAAATGCG-3 0 )2 native 24.54 40.32 65.86 65165 BDL015; DSS89 d(CGCAAATTTGCG)2 native 24.87 40.90 65.64 66768 BDL038; EBSSN92 native 25.20 41.65 65.81 69073 CFWR87 d(CGCAAATTTGCG)2, propamidine 24.78 41.16 65.51 66817 GDL032; NN95; d(CGCAAATTTGCG)2, d(CGCAAATTTGCG)2, 65W berenil 24.64 40.61 65.07 65111 BDL016; BSGN92 distamycin 25.20 41.07 64.65 66910 CFWR87 d(CGCAAATTTGCG2) d(CGCAAATTTGCG)2 II netropsin 26.48 41.26 66.88 73071 CAM89 d(CGCAAATTTGCG)2 netropsin 25.65 42.03 65.33 70430 QLD91 TRIBIZ 24.70 40.82 64.90 65436 GDL039; CGNLL96 d(CGCAAATTTGCG)2 25.27 41.32 65.11 67985 SBSN94; VGA94 d(5 0 -CGCAAATTTGCG-3 0 )2x61W I Hoechst-33258 native 24.28 39.35 66.37 63411 BDL078; S-SA97 d(CGCGATATCGCG)2 netropsin 25.48 41.26 66.88 70311 GDL001; CAM89 d(CGCGATATCGCG)2 d(CGCGATATCGCG)2 Hoechst-33258 25.59 40.56 67.10 69645 CCAW89 d(5 0 -CGC[iG]AATTTGCG-3 0 )2 Hoechst-33342 25.77 41.10 64.30 68103 RGB98 d(CGC[et6G]AATTCGCG)289W Hoechst-33342 25.686 41.065 66.416 70055 GDL021; SMR92a d(CGC[et6G]AATTCGCG)289W Hoechst-33258 25.32 40.58 66.08 67896 GDL022; SMR92a d(CGCAAGCTGGCG)2 native 25.29 41.78 64.76 68426 BDL022; WSS90 d(CGCGTTAAGCGC)2 native 25.7 40.5 67.0 69737 BDL059; BRZS95 d(CGCGTTAAGCGC)2 netropsin 26.49 40.87 67.02 69820 BRZS95 d(CGTGAATTCACG)2 native 24.78 40.85 65.67 66475 BDL028; LKD90; NGRB91 native 25.08 39.91 66.47 66533 BD0057; TN (5 0 -CGCTTATATGCG-3 0 ) þ (5 0 -CGCATATAAGCG-3 0 ) native 25.40 40.70 65.80 68023 BDL006; NFLK87 (5 0 -CGCAAAAAAGCG-3 0 ) þ (5 0 -CGCTTTTTTGCG-3 0 )

Table 5.13. (Continued ) Dodecameric Oligonucleotide


a

b

c

Cell Vol.

NDB ID; Reference

Non-self-complementary {(5 0 -CG[5BrC]ATATTTCGC-3 0 ) þ (5 0 -CGCAAATATGCG-3 0 )} at 110 K Non-self-complementary {(5 0 -CG[5BrC]ATATTTCGC-3 0 ) þ (5 0 -CGCAAATATGCG-3 0 )} at 110 K d(5 0 -CGCGAAUUCGCG-3 0 )2

Native þ 78 W

25.19

40.58

66.01

67476

ANN99

TRIBIZ þ 90 W

25.50

40.42

65.55

67563

DD0014; ANN99

native

25.43

39.74

65.25

65941

BDL075; PS96

Notes: * see x5.2.2 # 5-bis(4-aminophenoxy)pentane (NSC 620107) [iG] is isoguanine (2-hydroxyadenine) furamidine is 2,5-bis(4-guanylphenyl)furan TRIBIZ see Scheme 5.5 References: ANN99 – Aymami, Nunn and Neidle, 1999; BRSZS95 – Balendiran, Rao et al., 1995; BSGN92 – Brown, Sanderson, Garman and Neidle, 1992; BSSJ90 – Brown, Sanderson, Skelly et al., 1990; CCAW89 – Carrondo et al., 1989; CAM89 – Coll, Aymami, van der Marel, van Boom, Rich and Wang, 1989; CB95 – Czarny, Boykin et al., 1995; CFWR87 – Coll, Frederick, Wang and Rich, 1987; CGNLL96 – Clark, Gray et al., 1996; DSD82 – Drew, Samson and Dickerson, 1982; DSS89 – DiGabriele, Sanderson and Steitz, 1989; DrDi81 – Drew and Dickerson, 1981; EBSSN92 – Edwards, Brown, Spink, Skelly and Neidle, 1992; EJN92 – Edwards, Jenkins and Neidle, 1992; FKKDD82 – Fratini, Kopka, Drew et al., 1982; GKD95 – Goodsell, Kopka and Dickerson, 1995; GNKLD95 – Goodsell, Ng, Kopka et al., 1995; GSN98 – Guerri, Simpson and Neidle, 1998; HDK85 – Holbrook, Dickerson and Kim, 1985; HSVW01 – Howerton, Sines, VanDerveer and Williams, 2001; KYG85a – Kopka, Yoon, Goodsell, Pjura and Dickerson, 1985a; KYG85b – Kopka, Yoon, Goodsell, Pjura and Dickerson, 1985b; LGC89 – Larsen, Goodsell, Cascio et al., 1989; LKD90 – Larsen, Kopka and Dickerson, 1990; LTNB96 – Laughton, Tanious, Nunn, Boykin, Wilson and Neidle, 1996; NFLK87 – Nelson, Finch, Luisi and Klug, 1987; NGRB91 – Narayana, Ginnell, Russu and Berman, 1991; NJN94a – Nunn, Jenkins and Neidle, 1994a; NJN94b – Nunn, Jenkins and Neidle, 1994b; NN95 – Nunn and Neidle, 1995; NS – Neidle and Simpsom, to be published; QLD91 – Quintana, Lipanov and Dickerson, 1991; PS96 – Partridge and Salisbury, 1996 (to be published); RGB98 – Robinson, Gao et al., 1998; SBSN94 – Spink, Brown, Skelly and Neidle, 1994; SBC00 – Squire, Baker, Clark et al., 2000; SMHVW00 – Sines et al., 2000; SMHW98 – Shui, McFail-Imon, Hu and Williams, 1998; SMR92a, b – Sriram, van der Marel, Roelen, van Boom and Wang, 1992a, b; S-SA97 Shatzky-Schwartz, Arbuckle et al., 1997; TCKW96 – Trent, Clark, Kumar, Wilson, Boykin, Hall, Tidwell, Blagburn and Neidle, 1996; TN – Todd and Neidle, to be published; TVC93 – Tabernero, Verdaguer, Coll, Fita, van der Marel, van Boom, Rich and Wang, 1993; VGA94 – Vega, Garcia Saez, Aymami et al., 1994; WPDD84 – Wing, Pijura, Drew and Dickerson, 1984; WNTN95 – Wood, Nunn, Czarny, Boykin and Neidle, 1995; WSS90 – Webster, Sanderson, Skelly, Neidle, Swann, Li and Tickle, 1990; YPGD88 – Yoon, Prive, Goodsell and Dickerson, 1988.


183

5.6.4 Dodecameric oligonucleotides Most dodecameric oligonucleotides are found in the B-DNA form and these are listed in the first part of this section (Table 5.13) and also in Table 5.14. The relatively few examples crystallizing in the A-form are listed towards the end of the section (Table 5.15). Especially interesting are those few examples where both A and B forms are found in the same oligonucleotide, reinforcing earlier examples of an intermediate range of structures between classic A and B forms. The formulae of most of the dodecamer duplexes can be written as {d(CGCXGCG)2} where X is severally A3T3 Tabernero et al., 1993; GT2A2C Balendiran, Rao et al., 1995; GATATC Coll, Aymami et al., 1989; GA2T2C Sriram et al., 1992b; [e6G]A2T2C Sriram et al., 1992a; GA2T2 [BrC] Kopka et al., 1985b. In all of these host structures the accessible minor groove is six base pairs in length, and in all the examples the four base pairs at the centre of the duplex are AT base pairs. The unit ˚ 3. cell volumes span the range from about 64 000 to 73 000 A

Fig. 5.21. Propamidine (with its bonds emphasized) in the minor groove of the dodecameric duplex d(CGCGAATTCGCG)2. The mode of complexation should be contrasted with the side-by-side arrangement noted earlier (Figs. 5.16 and 5.18). (Reproduced from Nunn and Neidle, 1995.)

184


The structure of the native dodecamer (d(CGCGAATTCGCG)2) was first determined by Drew and Dickerson (1981). The asymmetric unit consists of one double stranded helix; in addition, one spermine and 72 water molecules were identified. The density of the crystals could not be measured because ‘‘the crystals fell apart;’’ however, that the crystals were 50% water was used as a working hypothesis. Redetermination of this structure at atomic resolution (Tereshko, Minasov and Egli, 1999) and 120K has shown the important structural role of ordered Mg2þions (two hexahydrates and one pentahydrate per asymmetric unit). Dickerson, Goodsell and Kopka (1996) have summarized the crystal chemistry of fifteen differently sequenced native B-DNA dodecamers; all crystallize (from a variety of conditions) in space group P212121, Z ¼ 4, a 24.9, b 40.4, c ˚ , This is the cell given in Table 5.13 for the dodecamer-drug complexes. Thus the 66.2 A dodecamer-drug complexes should be considered to be primary solid solutions of drug molecules in the native B-DNA dodecamer. The structure determination of d(5 0 -GGCCAATTCGCG-3 0 )2 at 16K is, so far, the only example of a structure determination at such a low temperature, and the only one that has been followed at intervals of a few degrees both in the cooling and heating regimens (Fig. 5.22). The curve of cell volume against temperature shown in Fig. 5.22 is quite different from analogous measurements such as those given in Chapter 16 (Figs. 16.9, 16.15 and 16.24). Here the crystal appears to have been frozen into a static condition at about 200K, below which the volume does not change despite cooling to 16K; this is remarkable indeed. Drew et al. comment ‘‘This break [in the V–T ] curve at 200K marks what we shall term the solvent solidification point, leaving open the question of whether this represents a true phase change.’’ Another surprise occurred on heating – at about ˚ 3, accompanied by a change 210K the volume decreased sharply from 61 000 to 55 000 A to monoclinic symmetry, which appears to have persisted to room temperature. These curious phenomena do not appear to have been studied in detail, nor do other workers carrying out structure determinations at, say, 100K appear to have searched for similar

Cell volume, Å3

75 000 a = 25.3 b =40.8 c = 66.5

70 000 a=23.4 b=39.3 c=65.3

65 000 60 000

a = 22.7 b = 38.5 c = 62.6 β = 86.4°

55 000 50 000

0

100

200

300

Temperature, K

Fig. 5.22. Cell volume versus temperature for the dodecamer d(CGCGAATTCGCG). Lattice ˚ ) for the P212121 cell are shown at both ends of the cooling curve. The crystal adopts parameters (A monoclinic symmetry on warming to 220K. Filled circles for cooling, open circles for heating. (Reproduced from Drew, Samson and Dickerson, 1982.)

Table 5.14. Crystal data for various dodecameric oligonucleotides. All crystals have right-handed B-DNA-like duplexes with Watson-Crick base pairs. ˚ , degrees, A ˚ 3. We use ‘native’ while NDB uses ‘plain’ Triclinic cells have been reduced. Dimensions in A Dodecanucleotide

Space group

a/

b/

c/

Cell Volume

NDB ID; Reference

d(ACCGAATTCGGT)2

P1; 3 independent duplexes in unit cell P1; 3 independent duplexes in unit cell C2, Z ¼ 8

39.46 119.54 40.11 116.20 64.83

39.93 103.56 40.54 97.38 25.35

52 197

BD0052; HR01

56 761

BD0002; HR01

58 068

LH93

P212121, Z ¼ 8 R3 R3

44.8 64.067 65.89

39.82 92.08 40.47 99.32 35.36 92.24 66.1 64.067 65.89

42.9 44.679 47.09

127 039 158 819 177 051

DGS93 BD0001; HR01 BDL035; TVM91

R3 P43, Z ¼ 8 P41212, Z ¼ 8

41.96 40.207 40.197

41.96 40.207 40.197

101.4 57.575 77.336

154 611 93 076 124 959

BD0004; LM98 BD0003; RR98 BD0026; NKD00

d(ACCGACGTCGGT)2 d(CGTAGATCTACG)2 d(CGCGAAAAAA/CGa)2, d(ACCGACGTCGGT)2 (ACCGCCGGCGCT)/ (GGCGCCGGCGGT) d(CGCGAATTCGCG)2 d(ACCGGTACCGGT)2 d(CATGGGCCCATG)2

References: DGS93 – DiGabriele and Steitz, 1993; HR01 – Hizver, Rozenberg, Frolow, Rabinovich and Shakked, 2001; LH93 – Leonard and Hunter, 1993; LM98 – Liu, Malinina, Hyunh-Dinh and Subirana, 1998; NKD00 – Ng, Kopka and Dickerson, 2000; RR98 – Rozenberg, Rabinovich, Frolow, Hegde and Shakked, 2001; TVM91 – Timsit, Vilbois and Moras, 1991.

186


effects, perhaps because flash cooling has become the method of choice. The 16K, 300K and brominated cytosine structures have been compared by Kopka, Fratini, Drew and Dickerson (1983) particularly with respect to water structure. Crystal packing of the native oligonucleotide dodecamer has been described by Dickerson and Drew (1981) – the two terminal CG base pairs at each end of the duplex interact with those of a neigboring duplex by interlocking of their minor grooves through interduplex N–H . . . N hydrogen bonds between the guanine bases across the minor groove. This packing arrangement serves to stabilize the DNA and provides an A þ T rich region in the centres of the oligonucleotides that is free of packing effects and to which ligands can bind. In overall terms this description holds for all the isomorphous structures listed in Table 5.14, with details differing from example to example. The native dodecamers crystallize in a number of other structures, mostly one-of-a-kind and these are summarized in Table 5.14, with crystal symmetries running from triclinic through to rhombohedral. The Drew–Dickerson d(CGCGAATTCGCG)2 duplex, the prototype for so many orthorhombic crystals of different sequence but similar crystal structure, has been crys˚ , cell volume ¼ 154 611 A ˚ 3; tallized in the rhombohedral system (R3, 41.96, 101.40 A Liu, Malinina et al., 1998). The provenance of the new modification is ascribed to the presence of Ca2þ in the crystallizing solution instead of the more usual Mg2þ. The rhombohedral crystal structure has a quasi-hexagonal arrangement of the duplexes

Fig. 5.23. The pseudo-hexagonal arrangement of molecules at the same layer of columns shown in projection down the c axis. There are three different sets of molecules about the threefold axes. ˚ between phosphate groups. The calcium ion is on the Dotted lines show contacts of less than 7 A threefold axis. (Reproduced from Liu et al. 1998.)

GENERAL SURVEY OF THE CRYSTAL CHEMISTRY OF OLIGONUCLEOTIDE 187

Table 5.15. Crystal data for various dodecameric oligonucleotides. The crystals have right-handed ˚, A-DNA-like duplexes with Watson-Crick base pairs or mixed A and B forms. Dimensions in A ˚3 degrees, A Dodecanucleotide

Space group

a

b

c

Cell Volume

NDB ID; Reference

d(CCCCCGCGGGGG)2 d(CCGTACGTACGG)2 d(GCGTACGTACGC)2, d(CGCCCGCGGGCG)2

P3221 P6122 P6122 P212121, Z¼4 P41212, Z¼8

45.2 46.2 46.2 31.29

45.2 46.2 46.2 42.60

65.0 71.5 71.5 46.00

115 006 132 166 132 166 61 316

ADL025; VAF91 ADL045; BZS92 ADL046; BJZS92 MFHS99

40.197

40.197

77.336

124 959

BD0026; NKD00

d(CATGGGCCCATG)2

References: BJZS92 – Bingham, Jain, Zon and Sundaralingam, 1992; BZS92 – Bingham, Zon and Sundaralingam, 1992; MFHS99 – Malinina, Fernandez et al., 1999. VAF91 – Verdaguer, Aymami, Fernandez-Forner et al., 1991.

(Fig. 5.23). In the dodecamer molecule only the central decamer is in the B-form, the terminal cytosines being highly disordered. This reminds one of similar behaviour described earlier for some decamers (see Table 5.11). Dodecamers in the A-form are much less frequent than those in the B-form; some examples are given in Table 5.15. The first three entries have more or less the standard A conformation. The d(CGCCCGCGGGCG)2 sequence ‘‘shows a unique conformation, quite different from all previously studied oligonucleotide duplexes; the central octamer has an A conformation but with a sharp 65 kink in the centre; the terminal base steps have a B-like conformation; the major groove is completely closed in the centre, a hollow molecule is thus found. The results obtained confirm the high degree of variability of DNA structure.’’ (Malinina et al., 1999). The tetragonal d(CATGGGCCCATG)2 duplex (Ng, Kopka and Dickerson, 2000) has interesting crystallochemical features. The crystals have an unusually high solvent content – some 60% by weight. The structure shows archetypical features of both A- and B-DNA. The authors comment ‘‘ . . . crystals of G3C3 are not isomorphous with any previous oligonucleotide structure, whether A, B or Z . . . trapping of a stable intermediate [structure] suggests that the A- and B-DNA are not discrete, as previously believed’’ (cf. Section 5.2.1).

5.7 General survey of the crystal chemistry of oligonucleotide and oligonucleotide-guest structures It is simplest to start with the dodecameric oligonucleotides and work backwards. The entries in Table 5.13 show that a large group of native dodecamers are all isomorphous ˚ , space group P212121, Z ¼ 4) despite a wide variability in (a 26, b 41, c 66 A composition and sequence. This applies even to the noncomplementary example {(CG[5BrC]GAATTCGCG) þ (CGCAAATTTGCG)}. This isomorphism extends to all the minor groove drug complexes, again with a wide range of variation in composition and sequence of the base pairs and the chemical nature of the drugs. A conclusion is that the

188


overall crystal structure is determined by the general shape and interactions of the duplexes and that differences of composition and sequence, and presence or absence of a second component, provide only minor perturbations not sufficient to disturb the overall structure. These complexes are primary solid solution phases. However, one caveat is needed. There are a few examples of other phases (of the native duplexes) that do not fit with the thirty-odd entries of Table 5.13; they may be heteromorphs. These dodecamers all have the B-DNA conformation. There are a few dodecamers that have the A-DNA conformation, and two that have mixed conformations. The decamers present a more complicated picture. The twenty-odd differently sequenced decamers (14 were discussed by Dickerson, Goodsell and Kopka, 1996) crystallize in six different space groups (Table 5.9). In terms of cell volume the two orthorhombic and one monoclinic modification may well be heteromorphs (54 000, ˚ 3 (in round figures)) but the hexagonal and trigonal modifications 25 500 and 26 000 A ˚ 3). A decision must be based on the detailed crystal differ somewhat (84 000 and 95 000 A structures. Interestingly, there are two groups of natı¨ve structures and one native plus two complexes (DAPI and netropsin) that crystallize in space group P212121, (Z ¼ 4, one helix per asymmetric unit), but these groups all have different detailed structures. The three polyamide complexes of Table 5.10 are isostructural and crystallize in space group I2, but their structure differs in detail from the C2 structures (Group III) of Table 5.9. The polyamide minor groove binder is apparently large enough to make a considerable contribution to the cell volume. Both B-DNA (Tables 5.9 to 5.11) and A-DNA (Table 5.12) conformations are found. The octanucleotides resemble the dodecanucleotides in that there are many native structures, almost all isomorphous despite differences of composition and sequence (Table 5.5); the one spermine complex in Table 5.5 is a primary solid solution phase. However, there is also a resemblance to the decamers, with their non-isomorphous hexagonal and trigonal phases (Groups V and VI of Table 5.9). A difference between dodecamers and octamers is that there do not seem to be any octamer complexes (either intercalation or minor groove), the spermine and distamycin complexes excepted. The octanucleotides have A-DNA conformations (Tables 5.5, 5.6 and 5.7) but the octamerdistamycin complexes of Table 5.8 have the B-DNA conformation. The hexamer intercalation complexes fall into three groups according to the nature of the intercalated molecules. The native structures that have been reported are not directly comparable with the complexes but it seems unlikely that appreciably distorted intercalates will be isomorphous with (as yet hypothetical) undistorted native structures. Thus these are three different types of intermediate phase. The hexanucleotide – intercalation complexes have B-DNA conformations (Tables 5.2, 5.3 and 5.4). Some wider comparisons can be made. There is a resemblance to the ‘moiety within molecule’ structures described in Chapter 3 in that host and guest in the oligonucleotide complexes probably remain associated in solution, although NMR results suggest that there may be compositional and structural differences. There is a resemblance to the cyclodextrin complexes of Chapter 4 for similar reasons, and also in the occurrence of large groups of isomorphous structures within each of the separate -, - and cyclodextrin families. It is perhaps not too far-fetched to see an analogy between the structural roles of the host cyclodextrins and DNA duplexes, on the one hand, and the guests (drug molecules) on the other. However, cyclodextrins are invariably found in

REFERENCES

189

hydrated forms so it is more difficult to apply the phase rule. One final comparison between the cyclodextrin and oligonucleotide structures – it seems unlikely that either will be properly understood without low-temperature measurements on crystals of diffraction quality high enough to resolve solvent structure and locate ions. Although some hundreds of oligonucleotides and their complexes have been crystallized and their crystal structures reported, one does not know the extent to which this sample is representative of the global population. Thus generalizations should be presented with caution.

References Adams, A., Guss, J. M., Denny, W. A. and Wakelin, L. P. G. (2002). Nucl. Acids Res., 30, 719–725. Adams, A., Guss, J. M., Collyer, C. A., Denny, W. A. and Wakelin, L. P. G. (1999). Biochemistry, 38, 9221–9233. Adams, A., Guss, J. M., Collyer, C. A., Denny, W. A. and Wakelin, L. P. G. (2000). Nucl. Acids Res., 28, 4244–4253. Adams, A., Guss, J. M., Collyer, C. A., Denny, W. A., Prakash, A. S. and Wakelin, L. P. G. (2000). Molec. Pharmacol., 58, 649–658. Aymami, J., Nunn, C. M. and Neidle, S. (1999). Nucl. Acids Res., 27, 2691–2698. Baikalov, I., Grzeskowiak, K., Yanagi, K., Quintana, J. and Dickerson, R. E. (1993). J. Mol. Biol., 231, 768–784. Balendiran, K., Rao, S. T., Sekharudu, C. Y., Zon, G. and Sundaralingam, M. (1995). Acta Cryst., D51, 190–198. Berger, I., Su, L., Spitzner, J. R., Kang, C., Burke, T. G. and Rich, A. (1995). Nucl. Acids Res., 23, 4488–4494. Berman, H. M., Olson, W. K., Beveridge, D. L., Westbrook, J., Gelbin, A., Demeny, T., Hsieh, S-H., Srinivasan, A. R. and Schneoder, B. (1992). Biophys. J., 63, 751–759. Berman, H. M.,Westbrook, J., Feng, Z., Iype, L., Schneider, B. and Zardecki, C. (2002). Acta Cryst., D58, 889–898. Bingman, C., Zon, G.and Sundaralingam, M. (1992). J. Mol. Biol., 227, 738–756. Bingman, C., Jain, S., Zon, G. and Sundaralingam, M. (1992). Nucleic Acids Res., 20, 6637–6647. Bingham, C., Li, X., Zon, G. and Sundaralingam, M. (1992). Biochemistry, 31, 12803–12812. Bloomfield, V. A., Crothers, D. M. and Tinoco. I. Jr. (2000) ‘‘Nucleic Acids. Structures, Properties and Functions’’ University Science Books, Sausalito, CA 794 pp. Brown, D. G., Sanderson, M., Garman, E. and Neidle, S. (1992). J Mol. Biol., 226, 481–490. Brown, D. G., Sanderson, M., Skelly, J. V., Jenkins, T. C., Brown, T., Garman, E., Stuart, D. I. and Neidle, S. (1990). EMBO J., 9, 1329–1334. Carrondo, M. A. A. F. de C. T., Coll, M., Aymami, J., Wang, A. H-J., van der Marel, G. A., van Boom, J. H. and Rich, A. (1989). Biochemistry, 28, 7849–7859. Cervi, A., Langlois d’Estantoit, B. and Hunter, W. N. (1992). Acta Cryst., D48, 714–718. Chen, X., Ramakrishnan, B. and Sundaralingam, M. (1995). Nature Struct. Biol., 2, 733–735. Chen, X., Ramakrishnan, B. and Sundaralingam, M. (1997). J. Mol. Biol., 267, 1157–1170. Chen, X., Ramakrishnan, B., Rao, S. T. and Sundaralingam, M. (1994). Nature Struct, Biol., 1, 169–175. Chen, X., Mitra, S. N., Rao, S. T., Sekar, K. and Sundaralingam, M. (1998). Nucl. Acids Res., 26, 5464–5471. Chiu, T. K. and Dickerson, R. E. (2000). J. Mol. Biol., 301, 915–945

190


Cirilli, M., Bachechi, F., Ughetto, G., Colonna, F. P. and Capobianco, M. L. (1993). J. Mol. Biol., 230, 878–889. Clark, G. R., Brown, D. G., Sanderson, M. R., Chwalinski, T., Neidle, S., Veal, J. M., Jones, R. L., Wilson, W. D., Garman, E. and Stuart, D. I. (1990). Nucl. Acids Res., 18, 5521–5528. Clark, G. R., Gray. J., Neidle, S., Li, Y-H. and Leupin, W. (1996). Biochemistry, 35, 13745–13752. Coll, M., Frederick, C. A., Wang, A. H-J. and Rich, A. (1987). Proc. Natl. Acad. Sci., USA 84, 8385–8389. Coll, M., Aymami, J., van der Marel, G. A., van Boom, J. H., Rich, A. and Wang, A. H-J. (1989). Biochemistry, 28, 310–320. Courseille, C., Dautant, A., Hospital, M., Langlois d’Estaintot, B., Precigoux, G., Molko, D. and Teoule, R. (1990). Acta Cryst., A46, FC9–FC12. Cruse, W. B. T., Salisbury, S. A., Brown, T., Cosstick, R., Eckstein, F. and Kennard, O. (1988). J. Mol. Biol., 192, 891–905. Czarny, A., Boykin, D. W., Wood, A. A., Nunn, C. M., Neidle, S., Zhao, M. and Wilson, W. D. (1995). J. Am. Chem. Soc., 117, 4716–4717. Demeunynck, M., Bailly, C. and Wilson, W. D. (2003). Editors: DNA and RNA Binders—from small molecules to drugs. Wiley–VCH, Weinheim, 2 vols. Deng, J., Pan, B. and Sundaralingam, M. (2003). Acta Cryst., D59, 2342–2344. Denny, W. A. (2003). ‘‘Acridine-4-carboxcamides and the concept of minimal DNA intercalators’’ in DNA and RNA Binders – from small molecules to drugs. Edited by Demeunynck, M., Bailly, C. and Wilson, W. D. (2003). Weinheim, Vol. 2, pp. 482–502. Dickerson, R. E. (1992). Meth. Enzym., 211, 67–111. Dickerson, R. E. and Drew, H. R. (1981). J. Mol. Biol., 149, 761–786. Dickerson, R. E. Goodsell, D. and Kopka, M. L. (1996). J. Mol. Biol., 256, 108–125. DiGabriele, A. D. and Steitz, T. A. (1992). J. Mol. Biol., 231, 1024–1039. DiGabriele, A. D., Sanderson, M. R. and Steitz, T. A. (1989). Proc. Nat. Acad. Sci., USA, 86, 1816–1820 Doucet, J., Benoit, J.-P., Cruse, W. B. T., Prange, T. and Kennard, O. (1989). Nature, 337, 190–192. Drew, H. R. and Dickerson, R. E, (1981). J. Mol. Biol., 151, 535–556. Drew, H. R., Samson, S. and Dickerson, R. E. (1982). Proc. Natl. Acad. Sci., USA 79, 4040–4044. Dziubek, K. F. and Katrusiak, A. (2004). Z. Kristallogr., 219, 1–11. Edwards, K. J., Jenkins, T. C. and Neidle, S. (1992). Biochemistry, 31, 7104–7109. Edwards, K. J., Brown, D. G., Spink, N., Skelly, J. V. and Neidle, S. (1992). J. Mol. Biol., 226, 1161–1173. Eisenstein, M. and Shakked, Z. (1995). J. Mol. Biol., 242, 662–678. Eisenstein, M., Hope, H., Haran, T. E., Frolow, F., Shakked, Z. and Rabinovich, D. (1988). Acta Cryst., B44, 625–628. Eisenstein, M., Frolow, F., Shakked, Z. and Rabinovich, D. (1990). Nucl. Acids Res., 18, 3185–3194. Ettorre, A., Cirilli, M., and Ughetto, G. (1998). Eur. J. Biochem., 258, 350–354. Fernandez, L.G., Subirana, J.A., Verdaguer, N., Pyshnyi, D., Campos, L. and Malinina, L. (1997). J. Biomol. Struct Dyn., 15, 151–163. Fratini, A. V., Kopka, M. L., Drew, H. R., Dickerson, R. E. (1982). J. Biol. Chem., 257, 14686–14707, Frederick, C. A., Williams, L. D., Ughetto, G., van der Marel, G. A., van Boom, J. H., Rich, A. and Wang, A. H-J. (1990). Biochemistry, 29, 2538–2549. Gallois, B., Langlois d’Estantoit, B., Brown, T. and Hunter, W. N. (1993). Acta Cryst., D49, 311–317. Gao, Y-G. and Wang, A. H-J. (1991). Anti-Cancer Drug Des., 6, 137–149. Gao, Y-G. and Wang, A. H-J. (1995). J. Biomol. Struct. Dynamics, 13, 103–117.

REFERENCES

191

Gao, Y.-G., Liaw, Y.-C., Li, Y.-K., van der Marel, G. A., Van Boom, J. H., Wang, A. H.-J. (1991). Proc. Nat. Acad. Sci., USA, 88, 4845–4849. Gao, Y-G., Liaw, Y-C., Robinson, H. R. and Wang, A. H-J. (1990). Biochemistry, 29, 10307–10316. Gao, Y.-G., Priebe, W. and Wang, A. H.-J. (1996). Eur. J. Biochem., 240, 331–335. Goldman, A. ‘‘Interaction of proteins with nucleic acids.’’ (1996). Nucleic acids in chemistry and biology, edited by M. Blackburn and M. J. Gait, 2nd edition, Oxford University Press, pp. 376–441. Goodsell, D. S., Kopka, M. L., Cascio, D. and Dickerson, R. E. (1993). Proc. Natl. Acad. Sci., USA 90, 2930–2934. Goodsell, D. S., Grzeskowiak, K. and Dickerson, R. E. (1994). J. Mol. Biol., 239, 79–97. Goodsell, D. S., Kopka, M. L. and Dickerson, R. E. (1995). Biochemistry, 34, 4983–4993. Goodsell, D. S., Kaczor-Grzeskowiak, M. and Dickerson, R. E. (1995). Biochemistry, 34, 1022–1029. Goodsell, D. S., Ng. H. L., Kopka, M. L., Lown, J. W. and Dickerson, R. E. (1995). Biochemistry, 34, 16654–16661. Grzeskowiak, K,, Goodsell D. S., Kaczor-Grzeskowiak, M,, Cascio, D. and Dickerson, R. E. (1993). Biochemistry, 32, 8923–8931. Grzeskowiak, K., Yanagi, K., Prive´, G. G. and Dickerson, R. E. (1991). J. Biol. Chem., 266, 8861–8883. Guerri, A., Simpson, I. J. and Neidle, S. (1998). Nucl. Acids Res., 26, 2873–2878. Haran, T., Shakked, Z., Wang, A. H-J. and Rich, A. (1987). J. Biomol. Struct. Dynamics, 5, 199–217. Heinemann, U., Lauble, H., Frank, R. and Blo¨cker, H. (1987). Nucl. Acids Res., 15, 9531–9550. Heinemann, U. and Alings, C. (1989). J. Mol. Biol., 210, 369–381. Heinemann, U. and Alings, C. (1991). EMBO J., 10, 35–43. Heinemann, U. and Hahn, M. (1992). J. Biol. Chem., 267, 7332–7341. Herbstein, F. H. (2004). Cryst., Growth Des., 4, 1419–1429. Hizver, J., Rozenberg, H., Frolow, F., Rabinovich, D. and Shakked, Z. (2001). Proc. Nat. Acad. Sci., USA, 98, 8490–8495. Holbrook, S. R., Dickerson, R. E. and Kim, S. H. (1985). Acta Cryst., B41, 255–262. Hoogsteen, K. (1959). Acta Cryst. 12, 822–823. Howerton, S. B., Sines, C. C., VanDerveer, D. and Williams, L. D. (2001). Biochemistry, 40, 10023–10031. Hunter, W. N., Langlois d’Estaintot, B. and Kennard, O. (1989). Biochemistry, 28, 2444–2451. Jain, S. and Sundaralingam, M. (1989). J. Biol. Chem., 264, 12780–12784. Jain, S., Zon, G. and Sundaralingam, M. (1989). Biochemistry, 28, 2360–2364. Kennard, O. and Hunter, W. N. (1991) Angew. Chem. Int. Ed. Engl., 30, 1254–1277. Kennard, O. and Salisbury, S. A. (1993) J. Biol. Chem., 268, 10701–10704. Kielkopf, C. L., Baird, E. E., Dervan, P. B. and Rees, D. C. (1998). Nature Struct. Biol., 5, 104–109. Kielkopf, C. L., Ding, S., Kuhn, P. and Rees, D. C. (2000). J. Mol. Biol., 296, 787–801. Kielkopf, C. L., White, S., Szewczyk, J. W., Turner, J. M., Baird, E. E., Dervan, P. B. and Rees, D. C. (1998). Science, 282, 111–115. Kopka, M. L., Fratini, A. V., Drew, H. R. and Dickerson, R. E. (1983). J. Mol. Biol., 163, 129–146. Kopka, M. L., Goodsell, D. S., Baikalov, I, Grzeskowiak, K., Cascio, D., and Dickerson, R. E. (1994). Biochemistry, 28, 13593–13610. Kopka, M. L., Goodsell, D. S., Han, G. W., Chiu, T. K., Lown, J. W. and Dickerson, R. E. (1997). Structure (London), 5, 1033–1046. Kopka, M. L., Yoon, C., Goodsell, D. S., Pjura, P. and Dickerson, R. E. (1985a). Proc. Natl. Acad. Sci., USA 82, 1376–1380. Kopka, M. L., Yoon, C., Goodsell, D. S., Pjura, P. and Dickerson, R. E. (1985b). J. Mol. Biol., 183, 553–563.

192


Langlois d’Estantoit, B., Gallois, B., Brown, T. and Hunter, W. N. (1992). Nucleic Acids Res., 20, 3561–3566. Langlois d’Estantoit, B., Dautant, A., Courseille, C. and Precigoux, G. (1993). Eur. J. Biochem., 213, 673–682. Larsen, T. A., Goodsell, D. S., Cascio, D., Greskowiak, K. and Dickerson, R. E. (1989). J. Biomol. Struct. Dynamics, 7, 477–491. Larsen, T. A., Kopka, M. L. and Dickerson, R. E. (1996). Biochemistry, 30, 4443–4449. Lauble, H., Frank, R., Blo¨cker, H. and Heinemann, U. (1988). Nucleic Acids Res., 16, 7799–7816. Laughton, C. A., Tanious, F., Nunn, C. M., Boykin, D. W., Wilson, W. D. and Neidle, S. (1996). Biochemistry, 35, 5655–5661. Lehninger, A. L. Biochemistry, 2nd edition, 1977, Worth, New York. Leonard, G. A. and Hunter, W. N. (1993). J. Mol. Biol., 234, 198–208. Leonard, G. A., Brown, T. and Hunter, W. N. (1992). Eur. J. Biochem., 204, 69–74. Leonard, G. A., Hambley, T. W., McAuley-Hecht, K., Brown, T. and Hunter, W. N. (1993). Acta Cryst., D49, 458–467. Lerman, L. S. (1961). J. Mol. Biol., 3, 18–30. Liaw, Y-C., Gao, Y.-G., Robinson, H. R., van der Marel, G. A., van Boom, J. H. and Wang, A. H.-J. (1989). Biochemistry, 28, 9913–9918. Lipanov, A. A., Kopka, M. L., Kaczor-Grzeskowiak, M., Quintana, M. and Dickerson, R. E. (1993). Biochemistry, 32, 1373–1389. Lipscomb, L. A., Peek, M. E., Zhou, F. X., Bertrand, J. A., VanDerveer, D. and Williams, L. D. (1994). Biochemistry, 33, 3649–3659. Lisgarten, J. N., Coll, M., Portugal, J., Wright, C. W. and Aymami, J. (2002). Nature Struct. Biol., 9, 57–60. Lisgarten, J. N., Lipanov, A., Kopka, M. L., Kaczor-Grzeskowiak, M., and Dickerson, R. E. ‘‘The structure of a B-DNA decamer with an I/A mismatch and comparison with a G/A mismatch.’’ To be published (NDB ID BD0055). Liu, J., Malinina, L., Huynh-Dinh, T., and Subirana, J. A. (1998). FEBS Lett., 438, 211–214. Lommerse. J. P. M., Motherwell, W. D. S., Ammon, H. I., Dunitz, J. D., Gavezzotti, A., Hofmann, D. W. M., Leusen, F. J. J., Mooij, W. T. M., Price, S. L., Schweizer, B., Schmidt, M. U., Eijck, B. P. van, Verwer, P. and Williams, D. I. (2000). Acta Cryst., B56, 697–714. Malinina, L., Fernandez, L. G., Huynh-Dinh, T. and Subirana, J. A. (1999). J. Mol. Biol., 285, 1679–1690. Marques, M. A., Doss, R. M., Urbach, A. A. and Dervan, P. B. (2002). Helv. Chim. Acta, 85, 4485–4517. McCall, M., Brown, T. and Kennard, O. (1985). J. Mol. Biol., 183, 385–396. Moore, M. H., Hunter, W. N., Langlois d’Estantoit, B. and Kennard, O. (1989). J. Mol. Biol., 206, 693–705. Moravek, Z., Neidle, S. and Schneider, B. (2002). Nucleic Acids Res., 30, 1182–1191. Narayana, N., Ginell, S. L., Russu, I. M. and Berman, H. M. (1991). Biochemistry, 30, 4449–4455. Neidle, S. (1997). Biopolymers, 44, 105–121. Neidle, Stephen. (1999). Editor The Oxford Handbook of Nucleic Acid Structures, Oxford University Press, 662 pp. Neidle, S., and Simpson, I. J. ‘‘Structure of dCGCGAATTCGCG bis[piperidino-ethyl]-furamidine complex.’’ To be Published. Neidle, S., Schneider, B. and Berman, H. M. (2003). Methods of Biochemical Analysis (structural bioinformatics), 44, 41–73. Nelson, H. C. M., Finch, J. T., Luisi, B. F. and Klug, A. (1987). Nature, 330, 221–226,. Ng, H. L., Kopka, M. L. and Dickerson, R.E. (2000). Proc. Nat. Acad. Sci., USA, 97, 2035–2039.

REFERENCES

193

Nunn, C. M. and Neidle, S. (1995). J. Med. Chem., 38, 2317–2325. Nunn, C. M. and Neidle, S. (1997). Acta Cryst., D53, 269–274. Nunn, C. M. and Neidle, S. (1997). Biochemistry, 36, 4792–4799. Nunn, C. M., Jenkins, T. C. and Neidle, S. (1994a). Eur. J. Biochem., 226, 953–961. Nunn, C. M., Jenkins, T. C. and Neidle, S. (1994b). Biochemistry, 32, 13838–13843. Nunn, C. M., van Meervelt, L., Zhang, S., Moore, M. H. and Kennard, O. (1991). J. Mol. Biol., 222, 167–177. Partridge, B. L. and Salisbury, S. A. (1996). To be published. Prive´, G. G., Heinemann, U., Chandrasegaran, S-S., Kan, L., Kopka, M. I. and Dickerson, R. E. (1987). Science, 238, 498–504. Prive´, G. G., Yanagi, K. and Dickerson, R. E. (1991). J. Mol. Biol., 217, 177–199. Qiu, H., Dewan, J. C. and Seeman, N. C. (1997). J. Mol. Biol., 267, 881–898. Quigley, G. J., Wang, A. H-J., Ughetto, G., van der Marel, G. A., van Boom, J. H. and Rich, A. (1980). Proc. Natl. Acad. Sci., USA 77, 7204–7208. Quintana, J. R., Lipanov, A. A. and Dickerson, R. E. (1991). Biochemistry, 30, 10294–10306. Quintana, J. R., Grzeskowiak, Yanagi, K. and Dickerson, R. E. (1992). J. Mol. Biol., 225, 379–395. Rabinovich, D., Haran, T., Eisenstein, M. and Shakked, Z. (1988). J. Mol. Biol., 200, 151–161. Robinson, H., Gao, Y-G., Bauer, C., Roberts, C., Switzer, C. and Wang, A. H-J. (1998). Biochemistry, 37, 10897–10905. Rozenberg, H., Rabinovich, D., Frolow, F., Hegde, R. S. and Shakked, Z. (1998). Proc. Nat. Acad. Sci., USA. 95, 15194–15199. Saminadin, P., Dautant, A., Mondon, M., Langlois d’Estantoit, B., Courseille, C. and Precigous, G. (2000). Eur. J. Biochem., 267, 457–464. Schuerman, G. S., Smith, C. K., Turkenburg, J. P., Dettmar, A. N., van Meerveldt, L. and Moore, M. H. (1996). Acta Cryst., D52, 299–314. Shakked, Z. (1991). Curr. Opin. Struct. Biol., 1, 446–453. Shakked, Z., Rabinovich, D., Cruse, W. B. T., Egert, E., Kennard, O., Sala, G., Salisbury, S. A. and Viswamitra, M. A. (1981). Proc. Roy. Soc., B213, 479–487. Shakked, Z., Rabinovich, D., Kennard, O., Cruse, W. B. T., Salisbury, S. A. and Viswamitra, M. A. (1985). J. Mol. Biol., 166, 183–201. Shakked, Z., Guerstein-Guzikevich, G., Eisenstein, M. Frolow, F. and Rabinovich, D. (1989). Nature, 342, 456–460. Shakked, Z., Guerstein-Guzikevich, G., Frolow, F., Rabinovich, D., Joachimiak, R. and Sigler, P. B. (1994). Nature, 368, 469–473. Shatzky-Schwartz, M., Arbuckle, N. D., Eisenstein, M., Rabinovich, D., Bareket-Samish, A., Haran, T. E., Luisi, B. F. and Shakked, Z. (1997). J. Mol. Biol. 267, 595–623. Shi, K., Pan, B. and Sundaralingam, M. (2003). Acta Cryst., D59, 458–467. Shui, X., McFail-Isom, L., Hu, G. G. and Williams, L. D. (1998). Biochemistry, 37, 8341–8355. Sines, C. C., McFail-Isom, L., Howerton, S. B., VanDerveer, D. and Williams, L. D. (2000). J. Am. Chem. Soc., 122, 11048–11056. Smith, C. K., Davies, G. J., Dodson, E. J., and Moore, M. H. (1995). Biochemistry, 34, 415–425. Smith, C. K., Brannigan, J. A. and Moore, M. H. (1996). J. Mol. Biol., 263, 237–258. Spink, N., Brown, D. G., Skelly, J. V. and Neidle, S. (1994). Nucl. Acids Res., 22, 1607–1612. Spink, N., Nunn, C. M., Vojtechovsky, J. Berman, H. M. and Neidle, S. (1995). Proc. Natl. Acad. Sci., USA, 92, 10767–10771. Squire, C. J., Baker, L. J., Clark, G. R., Martin, R. F. and White, J. (2000). Nucl. Acids Res., 28, 1252–1258. Sriram, M., Van Der Marel, G. A., Roelen, H. L. P. F., Van Boom, J. H. and Wang, A. H.-J. (1992a). Embo J., 11, 225–232.

194


Sriram, M., van der Marel, G. A., Roelen, H. L. P. F., van Boom, J. H. and Wang, A. H-J. (1992b). Biochemistry, 31, 11823–11834. Tabernero, I., Verdaguer, N., Coll, M., Fita, I., van der Marel, G. A., van Boom, J. H., Rich, A. and Wang, A. H-J. (1993). Biochemistry, 32, 8403–8410. Takusagawa, F. (1990). J. Biomol. Struct. Dyn., 7, 795–809. Tereshko, V., Urpi, L., Malinina, L., Hyunh-Dinh, T. and Subirana, J. A. (1996). Biochemistry, 35, 11589–11595. Teixeira, S. C. M., Thorpe, J. H., Todd, A. K. and Cardin, C. J. ‘‘Heavy Going: the Atomic Resolution Structure of a Topoisomerase II Poison Intercalating into DNA in Multiple Orientations also Reveals a Multiple-Adenine-Thymine Hydrogen Bonding Pattern.’’ To be Published. Thota, N., Li, X. H., Bingman, C. and Sundaralingam, M. (1993). Acta Cryst., D49, 282–291. Tidwell, R. R. and Boykin, D. W. (2003). ‘‘Dicationic DNA minor groove binders as antimicrobial agents’’ in DNA and RNA Binders – from small molecules to drugs. Edited by Demeunynck, M., Bailly, C. and Wilson, W. D. (2003). Wiley VCH, Weinheim, Vol. 2, pp. 414–460. Timsit, Y., Vilbois, E. and Moras, D. (1991). Nature, 354, 167–170. Tippin, D. B. and Sundarlingam, M. (1996). Acta Cryst., D52, 997–1003. Tippin, D. B. and Sundaralingam, M. (1997). J. Mol. Biol, 267, 1171–1185. Todd, A. K. and Neidle, S. ‘‘Crystal Structure of a Non-Self Complementary DNA Dodecamer Containing an A/T Tract: Analysis of the Effect of Crystal Environment on Local Helical Parameters.’’ To be Published. Todd, A. K., Adams, A., Thorpe, J. H., Denny, W. A., Wakelin, L. P. G. and Cardin, C. J. (1999). J. Med. Chem., 42, 536–540. Todd, A. K., Adams, A., Powell, H. R., Wilcock, D. J., Thorpe, J. H., Lausi, A., Zanini, F., L. P. G. and Cardin, C. J. (1999). Acta Cryst, D55, 729–735. Trent, J. O., Clark, G. R., Kumar. A., Wilson, W. D., Boykin, D. W., Hall, J. E., Tidwell, R. R., Blagburn, B. L. and Neidle, S. (1996). J. Med. Chem., 39, 4554–4562. Vega, M. C., Garcia Saez, I., Aymami, J., Eritia, R., van der Marel, G. A., van Boom, J. H., Rich, A. and Coll, M. (1994). Eur. J. Biochem., 222, 721–726. Verdaguer, N., Aymamı´, J., Fernańdez-Forner, D., Fita, I., Coll, M., Huynh-Dinh, T., Igolen, J. and Subirana, J. A. (1991). J. Mol. Biol., 221, 623–635. Vlieghe, D., Sponer, S. and Van Meervelt, L. V. (1999). Biochemistry, 38, 16443–16451. Vlieghe, D., Turkenburg, J. P. and Van Meervelt, L. V. (1999). Acta Cryst., D55, 1495–1502. Wahl, M. C., Rao, S. T. and Sundaralingam, M. (1996). Biophys. J., 70, 2857–2866. Wang, A. H-J., Ughetto, G., van der Marel, G. A., van Boom, J. H. and Rich, A. (1987). Biochemistry, 26, 1152–1163. Wang, A. H-J., Fujii. S, van Boom, J. H. and Rich, A. (1982). Proc. Natl. Acad. Sci., USA 79, 3968–3972. Wang, A.H.-J., Gao, Y.-G., Liaw, Y.-C. and Li, Y.-K. (1991). Biochemistry, 30, 3812–3815, Waring, M. J. and Wakelin, L. P. G. ‘‘Forty years on.’’(2003). in DNA and RNA Binders – from small molecules to drugs. Edited by Demeunynck, M., Bailly, C. and Wilson, W. D. (2003). VCH, Weinheim, Vol. 1, pp. 1–17. Watson, J. D. and Crick, F. H. C. (1953). Nature, (Lond.), 171, 737–738. Webster, G. D., Sanderson, M. R., Skelly, J. V., Neidle, S., Swann, P. F., Li, B. F. and Tickle, I. J. (1990). Proc. Natl. Acad. Sci., USA, 87, 6693–6697. Wilcocks, D. J., Adams, A., Cardin, C. J. and Wakelin, L. P. G. (1996). Acta Cryst., D52, 481–485. Williams, L. D., Egli, M., Gao, Q., Bash, P., van der Marel, G. A., van Boom, J. H., Rich, A. and Frederick, C. A. (1990). Proc. Natl. Acad. Sci., USA, 87, 2225–2229. Williams, L. D., Frederick, C. A., Ughetto, G. and Rich, A. (1990). Nucl. Acids Res., 18, 5533–5541. Wing, R. M., Pjura, P., Drew, H. R. and Dickerson, R. E. (1984). EMBO J., 3, 1201–1206.

REFERENCES

195

Wood, A. A., Nunn, C. M., Czarny, A., Boykin, D. W. and Neidle, S. (1995). Nucl. Acids Res., 23, 3678–3684. Wood, A. A., Nunn, C. M., Trent, J. D. and Neidle, S. (1997). J. Mol. Biol, 269, 827–841. Yanagi, K., Prive´, G. G. and Dickerson, R. E. (1991). J. Mol. Biol., 217, 201–214. Yang, X.-L. and Wang, A. H.-J. (1999). Pharmacology and Therapeutics, 83, 181–215. Yoon, C., Prive, G. G., Goodsell, D. S., and Dickerson, R. E. 1988. Proc. Nat. Acad. Sci., USA, 85, 6332–6336. Yuan, H., Quintana, J. R. and Dickerson, R. E. (1991). Biochemistry, 31, 8009–8021.


Part III Host–guest inclusion complexes


Introduction to Part III Host–guest inclusion complexes

I came home today – and this guy was sitting there and I said ‘‘Hey, pal! What’s going on here?’’ – and when he smiled he had these big white teeth like luxury hotels on the Florida coastline. And when he closed his mouth it looked like a big scar. And I said to myself ‘‘Holy smokes – looks like some kind of guest-host relationship to me!’’ Laurie Anderson quoted in TIME Magazine (Music) February 21, 1983

The molecular complexes to be discussed in Part III are crystalline and their inclusion properties derive from the arrangement of host and guest moieties in the solid state. In this sense they differ from the complexes considered in Part II, which are considered to exist (to greater or lesser extent) in solution as well as in the solid state. A necessary requirement for classification as a host–guest inclusion complex is that host and guest can be distinguished; generally the much larger host component forms an array that includes the guest. As we have noted earlier (our definition of ‘complex’ in Chapter 1), the overall crystal structure is determined by host–host interactions, although host–guest interactions may make an essential contribution to the stability of the complex. Thus our emphasis is placed on the nature of the host and its arrangement, with the guests filling a secondary role. However, a more general view requires that this approach be modified when there is strong host–guest interaction, and we return to this point below. Two principal types of host arrangement can be discerned – the tunnel inclusion complexes (Fetterly, 1964) and the clathrates (Powell, 1964). In the first type the guests are included in tunnels between the host molecules and can be in mutual contact (head-tohead or head-to-tail) while in the second the guests are in cages separated one from the other by intervening host molecules. Guest–guest interactions can generally be ignored in the clathrates but may make a small contribution to overall enthalpy in tunnel complexes. The distinction between the groups is often not clearcut – for example tunnels tending towards hourglass shape may behave effectively as cages towards larger guest molecules but not towards smaller ones. Two principal types of host interaction can be discerned, and these distinctions cut across the tunnel/clathrate boundary. The host can interact through directional bonding of various kinds, of which hydrogen bonding is undoubtedly the most important in terms of current knowledge, or through nondirectional forces, such as the ubiquitous dispersion forces (often referred to, also here, as van der Waals’ forces). Host–guest interactions can be similarly divided, with van der Waals interactions being encountered more often than host–guest hydrogen bonding. Much recent work has shown that particular hosts can form complexes of different types, not necessarily all inclusion complexes, with different sorts of guests, and we shall use the term polyfunctionality to encompass this ability; usually only one of these

200

H O S T – G U E S T I N C L US I ON C OM P L E XE S

functionalities is relevant to the formation of inclusion complexes. For example, thiourea forms S-linked coordination complexes with main group and transition metals, hydrogenbonded complexes with small molecules such as parabanic acid (also other sulphurcontaining molecules can form donor-acceptor compounds with acceptors such as iodine) and NH . . . S hydrogen bonds to itself to give inclusion complexes. Only the mutual hydrogen-bonding functionality is important in the context of inclusion complexes. Two questions arise at this point. Firstly, what are the parameters inducing a particular component (here called a host) to form an inclusion complex with a second component of one type (here called a guest) but, say, a hydrogen-bonded pair with another type of component, this being a molecular compound with the host–guest nomenclature no longer appropriate? The short answer is that we really do not have an adequate explanation, and this holds also for the presumably simpler one-component systems (Gavezzotti, 1994, 1998). Secondly, how important are the host–guest interactions in inclusion complexes? Remember that initially, in the days of Powell’s great contributions, the emphasis was on inclusion and enclosure, and host–guest interactions, although never forgotten, were recognized only sotto voce. This imbalance was redressed by the schools of Cram and Lehn, with their stress on the supramolecular nature of molecular complexes and compounds. We shall see that in inclusion complexes the host–guest interactions can range from minimal to sufficiently important to distort the host framework. We describe such frameworks as being ‘‘interrupted’’ by the intervention of the guest. We use the more restricted term versatility to refer to the ability of a particular host molecule to form inclusion complexes of various structural types. Thus thiourea is a nonversatile molecule because it forms only rhombohedral tunnel inclusion complexes while urea is more versatile because it forms hexagonal, rhombohedral and orthorhombic tunnel inclusion complexes. Urea shows even more versatility with the homologous ,!dinitriles (NC(CH2)nNC) as guests; at least eight different types of structure were found for n ¼ 1–8, 10, 12 and there is a clear change between n ¼ 5 and 6 from layered hydrogenbonded 1 : 1 molecular compounds (where the host/guest distinction does not apply) to tunnel inclusion complexes (Hollingsworth, Santarsiero and Harris, 1994). Trimesic acid, whose functionality as an acid is not important in the present context even though it does form salt-molecule complexes with amino and other acids, is even more versatile because, as neat host, it forms interstitial clathrates and also tunnel complexes of two different kinds; the monohydrate forms another kind of tunnel complex. Versatility stems from the mutual adaptation of host to guest, leading to the possibility of forming inclusion complexes of different structural types; it is intended for use in a broad rather than narrow sense. Within the framework of a group of isomorphous or isostructural inclusion complexes we use the term adaptability to refer to the adjustment of a particular type of host structure to various guests; this word has been borrowed from Powell (1964, see p. 469) who writes ‘‘A greater adaptability [of the dimensions of the unit cell] to the shape of the enclosed molecule is to be expected.’’ Thus we shall apply this term to the variation of cell dimensions with guest type for a group of isomorphous structures, well illustrated for the inclusion complexes of tri-o-thymotide (Lawton and Powell, 1958; Powell, 1964; Chapter 8.2) and extend its use to isostructural situations where there is similarity of structural arrangement without the identity of space group and near identity of cell dimensions demanded of isomorphous structures; many illustrations can be found among the cyclodextrin inclusion complexes (Chapter 4) and elsewhere.

HOST–GUEST IN CLUSION COMPLEXES

201

It is useful to consider what we would ideally wish to know about a particular host that has been shown to form inclusion complexes; these considerations have also guided our choice of subjects for inclusion in the chapters of Part III. One of the first questions to be answered is the capability of the host to form inclusion complexes – what guests are compatible with the host and are there any chemical limitations on the types of guests. It would be best if this information, usually obtained from preparative experiments, was reinforced by determination of phase diagrams, both for the binary host–guest systems and for ternary host–guest–solvent systems (for some systems it would desirable to have pressure included as a variable). Crystal structures of appropriate complexes should be determined at room temperature and at low temperatures chosen after measurement of specific heat–temperature curves (or of other physical properties dependent on temperature). Low temperature structures would give information about guests that are often disordered at room temperature. The Cp–T curves would also show the occurrence of phase transformations below room temperature. Suitable neutron diffraction measurements would give information about the mechanisms of phase transformations. Solid state NMR measurements (1H, 2H, 13C and other nuclei) give information about dynamics of host and guest not generally obtainable by XRD. Thermodynamic measurements of free energies, enthalpies and entropies of formation (in the solid state) are required for an understanding of the interactions between the components governing formation and stability of the complexes. Again it would desirable to have these parameters measured over a range of temperatures. This is quite a tall order and the only systems that come close to meeting all these requirements are the tunnel inclusion complexes of urea and the polyhedral clathrate hydrates (more specifically the gas hydrates). A thermodynamic distinction between two types of inclusion complex is noted here. An inclusion complex may be either a primary solid solution of guest in host, or it may be a separate phase in the host–guest phase diagram. The crystal structure of the solid solution will be that of the pristine host but the presence of the guest introduces distortions and irregularities that may well require application of sophisticated nonstandard diffraction techniques for their elucidation. If the inclusion complex is a separate phase then the thermodynamics of the transformation from the (empty) pristine host arrangement to the (empty) host arrangement in the new phase is an important factor to be taken into account together with contributions of the host–guest interactions to the overall stability. Primary solid solution is relatively rare among inclusion complexes; examples are the inclusion of hydrogen and helium in some high pressure ice phases (Section 7.2.3), and complexes of Dianin’s compound (4-p-hydroxyphenyl-2,2,4-trimethylchroman; Section 7.4.2). Most host–guest inclusion complexes have crystal structures different from those of the pristine hosts; thermodynamic and structural studies have been coordinated particularly for urea and thiourea tunnel inclusion complexes (Chapter 6), the complexes of quinol (Section 7.2.1) and the gas hydrates and related clathrate hydrates (Section 7.3). The preparation of a large number of individual inclusion complexes, often referred to as solvates, has been reported and the crystal structures of some of these have been determined (Davies, Finochiarro and Herbstein, 1984). However, most of these occurrences have not yet been the subjects of systematic chemical or crystallographic study and thus have not been included. We have preferred to restrict our choice of examples to those families of inclusion complexes where the chemical and structural foundations have been reasonably well established, or where interesting new chemical or structural principles are

202

H O S T – G U E S T I N C L US I ON C OM P L E XE S

being revealed. For each group – the tunnel inclusion complexes and the clathrates – we have distinguished between less versatile and more versatile hosts. These are essentially qualitative and time dependent concepts, particularly as a less versatile host may well become more versatile as investigation is pursued. Although the distinction has proved useful in organising the large amount of available material, its limitations should not be forgotten. The first chapters in Part III deal with tunnel inclusion (Chapter 6) and clathrate complexes (Chapter 7) of less versatile hosts; it follows from our definition of ‘versatility’ that the distinction between these two types of complex can be made. This is followed by Chapter 8 dealing with inclusion complexes formed by more versatile hosts; here the distinctions made above – between tunnel and clathrate complexes, and between directionally bonded and nondirectionally bonded hosts – largely fall away. The concept of topology of inclusion complexes has already been introduced in Chapter 1 (particularly see Fig. 1.2). Thus Chapter 6 deals with zero-dimensional and Chapter 7 with onedimensional guest sites. The trio of possibilities is completed in Chapter 9, where twodimensional arrays of guests are sandwiched between lamellae of host molecules.

References Davies, J. E. D., Finochiarro, P. and Herbstein, F. H. ‘‘Inclusion compounds formed by other host lattices’’, in Inclusion Compounds, Vol. 2, pp. 407–453, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Academic Press, London. (1984). Fetterly, L. C. (1964). ‘‘Organic adducts’’, in Non-Stoichiometric Compounds, edited by L. Mandelcorn, pp. 491–567, Academic Press, New York etc. Gavezzotti, A. (1994). Accts. Chem. Res., 27, 309–314. Gavezzotti, A. (1998). Cryst. Revs., 7, 5–121. Hollingsworth, M. D., Santarsiero, B. D. and Harris, K. D. M. (1994). Angew. Chem. Int. Ed. Engl., 33, 649–652. Lawton, D. and Powell, H. M. (1958). J. Chem. Soc., pp. 2339–2357. Powell, H. M. (1964). ‘‘Clathrates’’, in Non-Stoichiometric Compounds, edited by L. Mandelcorn, pp. 438–490, Academic Press, New York. Powell, H. M. (1984). ‘‘Introduction’’, in Inclusion Compounds, Vol. 1, pp. 1–28, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Academic Press, London.

Chapter 6 Tunnel inclusion complexes formed by hosts of lesser versatility

and shades Of trellis-work in long arcades, and cirque and crescent framed by wall William Wordsworth

Summary: Crystalline tunnel inclusion complexes have one-dimensional tunnels in their structures; the host molecules, which constitute the matrix, may be bonded together by directional bonds (generally hydrogen bonds) or van der Waals forces while the guest molecules in the tunnels generally interact with the host molecules by van der Waals forces, although some examples are known of host–guest hydrogen bonding. The nature of the guest appears to be limited only by the size and shape of the tunnel available to it, although there are some examples of specific host–guest interactions. The arrangements of the host molecules in the matrix are generally different from those in the neat hosts and in this situation the tunnel inclusion complexes are separate phases. The Bishop–Dance complexes – which are primary solid solution of guest in host – are exceptions to this rule. Most of the examples quoted show effects of interaction between host and guest which manifest themselves as minor changes in cell dimensions of isomorphous crystals or as distortions of a basic crystal structure type. Guest molecules are generally disordered in the tunnels at room temperature but order on cooling, often in a number of stages, accompanied by interaction with and distortion of the framework, usually manifested as phase transformations. Diffraction patterns from many complexes are composed of separate contributions from the host framework and the guest arrangement, both being modified by mutual interaction of host and guest. The complicated effects that can ensue are illustrated by studies of {3urea[1/4(n-hexadecane)]} over the temperature range 30–400K.

6.1 Introduction 6.2 Tunnel inclusion complexes with directionally bonded hosts 6.2.1 Urea, thiourea and selenourea as hosts 6.2.1.1 Introduction 6.2.1.2 Types of guest in hexagonal urea inclusion complexes 6.2.1.3 Guests which give rhombohedral urea inclusion complexes 6.2.1.4 Guests which give rhombohedral thiourea inclusion complexes 6.2.1.5 Hermann’s comprehensive structural model 6.2.1.6 Diffraction patterns from tunnel inclusion complexes 6.2.1.7 Hexagonal urea tunnel inclusion complexes 6.2.1.8 Determination of guest molecule conformation from diffuse x-ray scattering

204 206 206 206 207 208 209 210 212 215 218

T UN N E L I N C L US I O N C O M P L E XE S

204

6.2.1.9 Variation of structure with temperature, with particular reference to {3(urea)[1/4(n-hexadecane)]} 6.2.1.10 Interruption of urea framework by host–guest hydrogen bonding 6.2.1.11 Rhombohedral urea, thiourea and selenourea tunnel inclusion complexes 6.2.1.12 Monoclinic complexes derived from the rhombohedral complexes 6.2.1.13 Behavior of some rhombohedral inclusion complexes on cooling 6.2.1.14 The orthorhombic Type 4 urea tunnel inclusion complexes 6.2.1.15 The hypothetical Type 5 orthorhombic tunnel inclusion complexes 6.2.1.16 The crystal structure of selenourea and its relation to the structures of the tunnel inclusion complexes 6.2.1.17 Thermodynamics of the formation of the tunnel inclusion complexes 6.2.2 The Bishop–Dance hosts – exo-2,exo-6-dihydroxy2,6-dimethylbicyclo[3.3.1]nonane and analogs 6.2.2.1 Introduction 6.2.2.2 The helical tubuland structures 6.2.2.3 The ellipsoidal tetragonal clathrate complexes of some Bishop–Dance hosts 6.2.2.4 Derived structures 6.2.3 Ta4P4S29 – an inorganic framework containing sulphur chains 6.2.4 The tunnel hydrates 6.2.4.1 Tunnel hydrates with several water molecules per tunnel cross-section 6.2.4.2 Tunnel hydrates with one water molecule per tunnel cross-section 6.3 Tunnel inclusion complexes with van der Waals bonded hosts 6.3.1 Tunnel inclusion and other complexes of deoxycholic acid and related compounds 6.3.1.1 The complexes of deoxycholic acid 6.3.1.2 The complexes of cholic acid 6.3.2 Substituted spirocyclophosphazenes as hosts 6.3.3 Tritriptycene – a C62H38 hydrocarbon of D3h symmetry with three U-shaped bays 6.3.4 Trans-anti-trans-anti-trans-Perhydrotriphenylene as host 6.3.5 N-(p-tolyl)tetrachloro-phthalimide as host 6.4 Comparison of the various tunnel inclusion complexes References

6.1

219 227 231 235 236 245 245 247 247 251 251 251 264 267 268 269 269 271 272 272 273 281 291 297 298 307 310 311

Introduction

Perhaps the first crystalline tunnel inclusion complex to be studied by modern diffraction methods was the dioxane complex of {(CH3)3AsPdBr2}2 (Wells, 1938); however, this field did not attract widespread attention until after the Second World War, when Bengen’s (1951) work became known. We quote from Fetterly’s (1964) description: ‘‘The discovery that urea forms crystalline adducts with long, straight-chain organic compounds was made accidentally by Bengen in 1940 (German Patent Application OZ123438,


205

18 March, 1940) while conducting tests with urea in a study of its action on proteins in pasteurized milk. He noted under certain conditions the fat separates out in such a form as to justify the use of urea in a method for determining the fat content in milk. When he was bothered by what appeared to be frothing and an emulsion, he added a small amount of n-octyl alcohol and set aside his samples. Later he observed long crystals at the interface of the liquid layers. On attempting to reproduce these unexpected crystals, he made the classic discovery that they also form when saturated aqueous urea solution is mixed with n-octanol. From this point, investigation soon extended the adduct formation to include higher alcohols, acids and finally n-paraffins and other straight-chain compounds.’’ These observations led to the opening of a whole new area of investigation and applications (Farina, 1984), including the discovery of analogous complexes formed by thiourea and selenourea. We are not able to resist one other quotation: in these complexes ‘‘The urea and thiourea molecules form [a] honeycomb arrangement . . . and the hydrocarbon molecules form the honey.’’(Rutherford and Calvo, 1969). The tunnel inclusion complexes of urea, thiourea (first reported by Angla, 1947a) and selenourea (first reported by Bekkum, Remijnse and Wepster, 1967) form the first group where the hosts are directionally bonded and are discussed together because of their strong chemical and structural resemblances. However, there is a basic structural distinction between the hexagonal tunnel inclusion complexes of urea on the one hand and the rhombohedral tunnel inclusion complexes of urea, thiourea and selenourea on the other. The tunnel dimensions in these two types of crystal are similar but there are important differences in the arrangements of the host molecules in the tunnel walls, the hexagonal complexes having a helical, chiral arrangement of host molecules while there is a layered, non-chiral arrangement of host molecules in the rhombohedral complexes. The vital difference is that in the first group the force field is rather uniform both along the direction of the tunnel axis and normal to it. Thus guest molecules with quasi-cylindrical symmetry do not have preferred locations along the tunnel axis or azimuthal orientations about it when kT is greater than the host–guest interaction energy. In contrast, the force field along ˚ in the rhombohedral complexes and guests the tunnel axis has a periodicity of 5.5 A with similar periodicities will lock in to the host structure and favour formation of rhombohedral rather than hexagonal complexes (see Table 6.1 below; Lenne´, Mez and Schlenk, 1968). We consider the hexagonal inclusion complexes to be helical tubulates, other examples of which are exo-2,exo-6-dihydroxy-2,6-dimethylbicyclo-[3.3.1]nonane and its analogs (see Section 6.2.2) and Ta4P4S29. The term ‘tubulate’ was first introduced by Weber and Jossel (1983) to describe an one-dimensional open-tunnel structure. We define a helical tubulate as a structure in which the host moieties are arranged in a helical framework leaving tunnels in which guest molecules are accommodated. Although the three examples discussed here are all chiral, racemic arrangements are also possible. In similar vein one can describe the rhombohedral inclusion complexes as having cylindrical tubulate structures. The urea and thiourea tunnel inclusion complexes have been quite thoroughly investigated from both chemical and structural points of view. A rational overall structural picture has been proposed and the available chemical information fits into this framework. However, there is a vast amount of detail to be added, particularly in regard to crystallographic changes that occur on cooling, when most complexes seem to behave in ways that are specific to the guest involved. The other systems have not been studied to the same


206

extent, either chemically or structurally. One can rationalize the directional interactions of the hosts of Section 6.2 but it seems difficult to predict the behavior of the rather disparate compounds of Section 6.3; for example, it remains a mystery (to me) why only one of the ten stereoisomers of perhydrotriphenylene should form inclusion complexes. The following general considerations (adapted from Yeo, Harris and Guillaume, 1997) are important for determining the structures of tunnel inclusion complexes – (1) host–host interactions, (2) host–guest interactions, (3) intratunnel guest–guest interactions, (4) intertunnel guest–guest interactions, (5) host conformations, (6) guest conformations. Their relative importance depends on the nature of host and guest, and the temperature (and pressure). The urea inclusion complexes probably constitute the best studied family. At one extreme, when host–guest interactions are weak (as for paraffin guests) and the temperature high (say 300K), then host–host interactions dominate. Cooling leads to phase changes due to perturbation of host structure by increasingly important (compared to thermal fluctuations) host–guest interactions. At the other extreme, host–guest interactions can be strong enough to produce appreciable modification of the basic host framework. And there is a vast, and little explored, area between these two extremes.

6.2

Tunnel inclusion complexes with directionally bonded hosts

6.2.1 6.2.1.1

Urea, thiourea and selenourea as hosts Introduction

The historical background to the formation of related tunnel inclusion complexes by the homologous molecules urea, thiourea and selenourea has been sketched above; the guests cover a wide range of different chemical types. The crystal structures of the inclusion complexes differ from those of the pure hosts (apart from selenourea, where this statement needs some qualification) and thus the complexes form separate phases in the phase diagrams. Urea forms hexagonal complexes with n-paraffins (Smith, 1950, 1952) and rhombohedral complexes with bulkier guests (Lenne´, Mez and Schlenk, 1968). For some guests both hexagonal and rhombohedral complexes can be obtained, depending on the details of the crystallization process. Thiourea (Lenne´, 1954) and selenourea (Bekkum, Remijnse and Wepster, 1967) form rhombohedral complexes isostructural with the rhombohedral urea complexes. Some guests can form rhombohedral complexes with both urea and thiourea and some others show similar behaviour vis-a-vis thiourea and selenourea. The differences in the shapes of the tunnels in the urea and thiourea complexes are somewhat larger than those between thiourea and selenourea complexes. Preparation of the complexes is generally by simple mixing of components, with or without a solvent, and crystallization by slow cooling. Much of our knowledge of the preparative and structural chemistry of the urea and thiourea complexes, and their thermodynamics, comes from the work, similar in scope and carried out more or less simultaneously, of groups at two industrial research laboratories, Shell Development Co. in the United States and BASF in Germany. The contributions of W. Schlenk, Jr. and coworkers are especially noteworthy. There have been a number of reviews (Fetterly, 1964; Schlenk, 1965; Bhatnagar, 1967; Takemoto and Sonoda, 1984; Hollingsworth and Harris, 1996).

DIRECTIONALLY BONDED HOSTS

207

Fig. 6.1. Stereodiagram of the crystal structure of urea viewed approximately along the c axis. (Reproduced from Swaminathan, Craven and McMullan, 1984.)

˚ at 12K; space The crystal structure of tetragonal urea (a ¼ 5.565(1), c ¼ 4.684(1) A group P4 21 m, Z ¼ 2) is shown in Fig. 6.1 (Swaminathan, Craven and McMullan, 1984; UREAXX). The molecules and hydrogen bonds are confined to the (110) and (110) planes of the unit cell which intersect and leave tunnels having a square cross-section of overall ˚ ; urea, in its neat crystals and in its tunnel inclusion complexes, is the size 3.94 3.94 A only molecule known where a carbonyl O atom accepts four hydrogen bonds. The two ˚ at 12K. The independent N . . . O hydrogen bond distances are 2.985(1) and 2.955(1) A diameter of the tunnel available to a potential guest molecule in neat urea would be, at ˚ after van der Waals radii have been taken into account. Thus tetragonal urea most, 1.5 A does not form tunnel inclusion complexes – the topology is right but not the dimensions. Thiourea is a ferroelectric which has many higher-order incommensurate phases between the room-temperature paraelectric and the low-temperature (Tc ¼ 202K) ferroelectric phase. At 295K it crystallizes in an orthorhombic cell (a ¼ 5.488(3), b ¼ 7.663(3), ˚ , Z ¼ 4, space group Pbnm, molecular symmetry Cs-m) in a closely packed c ¼ 8.564 A hydrogen-bonded structure without any tunnel-like features (Takahashi, Onodera and Shiozaki, 1990; THIOUR). The structure of selenourea is discussed briefly below. 6.2.1.2 Types of guest in hexagonal urea inclusion complexes The formation of hexagonal urea tunnel inclusion complexes with over three hundred different guests has been described (Lenne´, Mez and Schlenk, 1970). The potential guests studied included homologous series of the following structural types: n-paraffins, cyclohexylalkenes, secondary alkyl alcohols and chlorides, ketones, ethers, thioethers, carboxylic and dicarboxylic acids and their esters, 1,3-diglycerides. Unbranched paraffins from n-hexane onwards form crystalline complexes with urea. If the n-alkyl chain is sufficiently long, the presence of terminal ring groups does not prevent formation of complexes (e.g. 1-cyclohexyl- and 1-phenyleicosane, H11C6– (CH2)18–CH3 and H5C6-(CH2)18-CH3, form complexes) whereas a methyl group centered on a C13 chain inhibits complex formation (Schiessler and Flitter, 1952).


208

1-Iodopentane and 1-bromoheptane are the lowest n-alkyl halides to form complexes (Radell, Bradman and Bergmann, 1964), n-butyric acid is the smallest carboxylic acid and acetone the smallest ketone (Radell and Hunt, 1958). Diacyl peroxides (R–C(¼O)–O–O–C(¼O)–R; e.g. R ¼ C7H15, C10H21, C11H23) form complexes (Harris and Hollingsworth, 1990). The alkyl silanes RsiH3 (R¼n–C6H11 and above) and the di-n-alkyl silanes R2SiH2 (R ¼ C2H5 and above) form tunnel inclusion complexes (Muller and Meier, 1964a), as do linear di-n-alkoxy silanes (RO)2SiH2 (R ¼ C2H5 and above) (Muller and Meier, 1964b), the latter being generally more stable than the former. Crystalline urea inclusion complexes are formed from alcoholic solutions of urea and HgR2 (where R ¼ C2H5 and above) but not from RHgBr or compounds such as C4H9(CH2)5HgC4H9 (Bahr and Meier, 1958) n-Paraffin and di-n-alkyl silane guests are accommodated in the tunnels in head-to-tail fashion while alkyl carboxylic acids and n-alkyl silanes pack in head-to-head, tail-totail fashion.

6.2.1.3 Guests which give rhombohedral urea inclusion complexes These are much less common than the hexagonal variety, with perhaps 20–30 examples having been recorded, compared to the hundreds of hexagonal complexes known. There are two types of guest which appear to favour the rhombohedral structure. The first type comprises small molecules such as trioxane, which does give a rhombohedral complex (Cle´ment, Mazieres and Guibe´, 1972; QQQEVP) and acetone and dioxane, where the situation has to be clarified (‘‘Auswertung noch nicht abgeschlossen’’ (see p. 2439 in Lenne´, Mez and Schlenk, 1968). The second type comprises molecules whose length is ˚ , the half-periodicity of the tunnel in the rhomclose to some integral multiple of 5.5 A bohedral complexes; sometimes both hexagonal and rhombohedral complexes are formed. Some illustrations are given in Table 6.1; about a dozen other examples are known where both crystal types have been obtained (Lenne´, Mez and Schlenk, 1968; Hadicke and Schlenk, 1972). ˚ ) of guest molecule and the type of urea inclusion complex Table 6.1. Relation between length (A formed Guest

Nominal length of guest molecule

Nearest integral ˚ multiple of 5.5 A needed to favor rhombohedral structure

Structure type (H ¼ hexagonal, R ¼ rhombohedral)

2,13-dimethyltetradecane 2,14-dimethylpentadecane 2,15-dimethylhexadecane 2,16-dimethylheptadecane 2,17-dimethyloctadecane 2,18-dimethylnonadecane 2,19-dimethyleicosane

20.0 21.3 22.6 23.9 25.2 26.5 27.8

– 4 5.5 ¼ 22.0 4 5.5 ¼ 22.0 – – 5 5.5 ¼ 27.5 5 5.5 ¼ 27.5

H H and R R H H R R


209

6.2.1.4 Guests which give rhombohedral thiourea inclusion complexes Thiourea forms tunnel inclusion complexes with a much wider variety of guests than does urea, whose guests are essentially n-paraffins and their derivatives. The most stable thiourea complexes are formed with guests which have a branched chain or otherwise non linear structure, or are globular in shape or are substituted with fairly bulky groups. Schlenk (1951) has listed 108 guests divided into a number of categories, to which we add some further examples (with separate references) without pretending to present a complete list: Paraffins and olefins: e.g. 2,3-dimethylbutane through 2,2,4-trimethylbutane to 2,6,9,12,15-pentamethylheptadecane Alkyl halides, alcohols, ketones, acids, esters: e.g. isopropyl iodide, isobutyl chloride, 2-bromooctane Cycloparaffins and related ring systems: e.g. cyclopentane, cyclohexane, cyclooctane, cyclo-hexene, cyclohexa-1,4-diene, cyclooctatetraene Aromatic derivatives: e.g. isobutylbenzene, 2,3-dimethylnaphthalene, 1,6-dimethylnaphthalene, perhydrofluorene, perhydroanthracene. n-alkycyclohexanes (with the number of carbon atoms in the alkyl chain varying from 0 to 14) (Schlenk, 1951) dicyclohexyl-!,! 0 -polymethylenes (with the number of carbon atoms in the polymethylene chain varying from 0 to 9) (Lenne´, Mez and Schlenk, 1968) chlorocyclohexane (Harris and Thomas, 1990a); various trans-1,4-disubstituted cyclohexanes; cis-and trans–decalin; adamantane (van Bekkum, Palm, Verkade and Wepster, 1970); durene (Teter and Hettinger, 1955); ferrocene (Cle´ment, Claude and Mazieres, 1974; Hough and Nicolson, 1978); CHCl3 (Angla, 1947a); CCl4 (Angla, 1947a); alicyclic alcohols and ketones (Angla, 1947b). Schiessler and Flitter (1952) tested some 50 hydrocarbons of varying types for formation of inclusion complexes with thiourea: some 25 such complexes were obtained, leading to the conclusion that the cross-section of the guest should be within the limits ˚ , as measured on Fisher–Hirschfelder models. Complexes were (6.8 0.3) (5.8 0.5) A not formed readily if the dimensions were near the tolerance limits and, if formed, tended to be unstable. Phenyl groups tended to interfere with the formation of thiourea (and urea) complexes; n-paraffin chains attached to otherwise suitable molecules had similar deleterious effects on the formation of thiourea complexes. Although it has often been stated that n-paraffins do not form inclusion complexes with thiourea, it has been reported (McLaughlin and McClenahan, 1952) that such complexes can be obtained when n 12; a possible explanation is coiling of the paraffin chain within the tunnel. Tunnel inclusion complexes of selenourea have been prepared (Bekkum, Remijnse and Wepster, 1967) with 11 different guests (e.g. 4-t-butyl-1-neopentylbenzene, trans-1,4isopropylcyclohexane, adamantane, camphor), all of which also form complexes with thiourea. The cell dimensions of the selenourea complexes vary more with different guests than do those of the urea and thiourea complexes, perhaps because the weaker Se . . . H-N hydrogen bonds allow easier adaptation of the host matrix to the steric requirements of the guests. However, selenourea appears to be more selective in its choice of guest compounds than thiourea; for example, thiourea forms complexes with both cis and trans

210


isomers of 1-t -butyl-4-neopentylcyclohexane, whereas selenourea forms complexes only with the latter, thus enabling separation of isomers. 6.2.1.5 Hermann’s comprehensive structural model The structural chemistry of the urea and thiourea tunnel inclusion complexes can be summarised in terms of a structural model put forward by C. Hermann and described by Otto (1972). All these complexes are based on tunnels of hexagonal cross-section with the guest molecules disposed about the axes of the tunnels, and the walls of the tunnels containing the host molecules. The host molecules are arranged in spirals whose axes lie along the line of intersection of each group of three adjacent tunnels; these spirals are the basic structural motifs. A typical spiral, that of urea, is shown in Fig. 6.2, where one notes that there is hydrogen bonding of adjacent molecules within a spiral and also between different spirals; all possible hydrogen bonds are formed, oxygen acting as acceptor to four N–H bonds. The host molecules in the spirals are arranged about threefold screw axes; these are chiral, with a right-handed screw (31 axis) designated as R and a lefthanded screw (32 axis) as L. Hermann pointed out that there were only five ways (Fig. 6.3) of arranging such spirals to give the required structures, assuming left and right handed spirals to be equally probable. Types 1 and 2 are the two enantiomorphic forms of hexagonal urea complexes; Type 3 occurs in the rhombohedral urea, thiourea and selenourea tunnel inclusion complexes; Type 4 occurs in the orthorhombic crystals of urea with 1,4-dichlorobutane URDCLB), 1,5-dichloropentane or 1,6-dibromohexane as guest (Otto, 1972);Type 5 has not yet been reported. Trigonal selenourea has the Type 1 structure but the position of the

N C O L

L

N

Fig. 6.2. The L (32) screw axis illustrated for the hexagonal urea tunnel inclusion complexes, the dashed lines showing hydrogen bonds between O (large circles) and NH2 groups. The heavy vertical lines show the direction of the c axis. This diagram corresponds to Diagram 1 of Fig. 6.3. (Reproduced from Bhatnagar, 1967.)


211

R

L L

L

R

R

1

2 R

L

L

R

L

R

L R

R 3 L

L R

L

R R

R 4

5 L

L L

R

R

R

L L

Fig. 6.3. The five ways of arranging spirals of 31(R) and/or 32(L) symmetry to give the structures found for urea, thiourea and selenourea tunnel inclusion complexes. One is looking down the spiral shown in Fig. 6.2; hydrogen atoms of NH2 groups have been omitted. (Diagram adapted from Otto (1972) and Harris and Thomas (1990a).) Projection diagrams such as those shown above can be misleading as they do not show the three-dimensional arrangement of the hydrogen bonds, which is different in each structure type. This is illustrated by the stereodiagram of the left-handed helical arrangement (coordinates from Harris and Thomas, 1990). The chiral ribbons, which are defined by the anti N–H . . . O hydrogen bonds, run antiparallel to each other along the vertical c axis. The hexagonal channels are linked by their edges to form the solid-state honeycomb structure. At each z coordinate, the two urea molecules from separate helices are related by a two fold axis perpendicular to the channel axis. (Reproduced from Brown, Chaney, Santasiero and Hollingsworth, 1996.)

212


Fig 6.3 (Continued )

guest is occupied by a selenourea spiral similar to that at the corners of the hexagon (Rutherford and Calvo, 1969) so that no distinction can be made between host and guest. The five structural types described above are not interconvertible without breaking of hydrogen bonds and interchange of 31 and 32 spirals, the occurrence of a particular structure type depending on the nature of the guest. However, both hexagonal and rhombohedral urea complexes of a few guests have been found. The crystallographic and thermodynamic relations between such pairs do not appear to have been investigated; presumably they are polymorphs (if the compositions are identical) but relative stabilities have not been determined nor whether they transform enantiotropically or monotropically. Each structure type can itself undergo deformation to a derived structure of different symmetry, as a consequence of ordering of the guest arrangement on cooling, and concomitant changes in the host matrix. However, drastic changes in the positions of the spirals do not occur in such transformations, which are generally single crystal to single crystal, possibly accompanied by twinning. 6.2.1.6 Diffraction patterns from tunnel inclusion complexes We consider the diffraction patterns to be expected, in general terms, from structures of the tunnel inclusion complexes before describing, in more detail, each of the structure types, and some of the phase transformations that have been encountered. The general principles of such diffraction patterns have been understood implicitly since the first


213

structure determinations were carried out in the early 1950s; in particular, Laves and coworkers extracted much information from the diffuse scattering, as described below. A general discussion has been given by Frey (1997). We follow here the explicit and detailed description given by Harris and Thomas (1990b) but with some emendations. In one limit, the guest molecules are ordered with respect to those of the host and the X-ray diffraction patterns are those of a single crystal containing the two components in ordered array; an approximation to this situation is provided by {3(thiourea)[adamantane]} and {3(selenourea)[adamantane]} described below. In the other limit, the guest arrangement is disordered while that of the host is ordered, under the constraints imposed by the overall structure of the complexes. Thus X-ray diffraction photographs from crystals of tunnel inclusion complexes usually show two different types of scattering in reciprocal space, one being a regular lattice consisting of sharp (Bragg) reflections while the second component, incommensurate with the first, is much more diffuse. The Bragg diffraction pattern (the H diffraction pattern) derives, in first approximation, from the host matrix while the diffuse pattern (the G diffraction pattern) comes from the guest arrangement. A large variety of beautiful photographs has been presented by Nicolaides and Laves (1963), and by Harris and Thomas (1990b); we show one example (Fig. 6.4). Measurement of geometry and intensities from the H pattern will give the structure of the host matrix by application of the standard methods of crystal structure analysis, while analysis of the G pattern will give information, of a more or less limited kind, about the guest arrangement. A linear periodic array of scatterers in a tunnel will give a diffraction pattern consisting of an array of sheets in reciprocal space (Fig. 6.5). The average molecular length Lc can be calculated from the (reciprocal space) periodicity m of the sheets as Lc ¼ m/ m, where is the wavelength used. In fact, the scatterers are not points (as in Fig. 6.5) but have structure, which can be derived from the intensities of the sheets in reciprocal space through calculation of one-dimensional structure factors as F(00m) ¼ fn[cos 2mzn þ i sin 2mzn], where zn are the atomic coordinates projected onto the direction of the tunnel axis, and fn are the atomic scattering factors; the summation runs from n ¼ 1 to n ¼ N where N is the number of atoms in the repeat unit along the tunnel axis. Applications are discussed below. Up to this point we have assumed that host and guest arrangements do not interact and, hence, that H and G diffraction patterns will each give information only about their source

(hk 3)h (hk 2)h (hk 1)h (hk 0) – (hk 1)h – (hk 2)h – (hk 3)h

(hk 5)g (hk 3)g (hk 1)g – (hk 1)g – (hk 3)g – (hk 5)g

Fig. 6.4. X-ray diffraction photograph (Ni-filtered Cu K radiation) of {urea[x(dioctanoyl peroxide)]} oscillated about the c axis. Indexing of layer lines of the H pattern is shown on the left and of the G pattern on the right; the resolution of the latter into fairly sharp reflections is more pronounced than is usual. (Reproduced from Harris and Thomas 1990b.)


214

C

C*

Fig. 6.5. A one-dimensional array of point scatterers in direct space (representing the guest) is shown on the left, and the corresponding diffraction pattern of sheets in reciprocal space, on the right. Any perturbation in the periodicity of the one-dimensional array will lead to broadening of the sheets in the c* direction. The relationship between host and guest periodicities along c can be expressed as njcgj ¼ mjchj; when m and n are small integers then the two periodicities are commensurate; if n/m is irrational then the periodicities are incommensurate. (Reproduced from Harris and Thomas (1990b).)

C*

C

l=3 l=2 l=1 (hk 0) – l=1 – l=2 – l=3

these two-dimensional lattices are reciprocal to each other

b a

Fig. 6.6. The left hand portion of the diagram shows an arrangement of one-dimensional scatterers in real space without longitudinal correlation between adjacent tunnels (the host framework is omitted for clarity). The right hand portion of the diagram shows that, for m 6¼ 0, there are sheets of scattering in reciprocal space, as before. In projection (i.e. when m ¼ 0), the guest sublattice has the same periodicity as the host sublattice and we are in the single-crystal limit where both host and guest sublattices contribute to the hk0 reflections of the diffraction pattern. When there is correlation between the arrangement of the guests in the tunnels, then the sheets of scattering will show structure; this is illustrated in Fig. 6.4. (Reproduced from Harris and Thomas 1990b.)

structures (apart from the hk0 reflections, as noted in the caption to Fig. 6.6). Experience shows that this approximation holds reasonably well at room temperature, but it begins to fail when the crystals are cooled. In principle, the host arrangement will be modulated by the guests, and conversely. An excellent, but complicated, example is provided by the behaviour of {3(urea)[1/4(hexadecane)]}, which is discussed below.


6.2.1.7

215

Hexagonal urea tunnel inclusion complexes

(i) Chemical compositions: The chemical formula is generally expressed as {3(urea)[x(guest)]}, where x is not necessarily integral and the square brackets denote here, as elsewhere, inclusion of the guest in the host matrix. The non-stoichiometric compositions of the tunnel inclusion complexes of urea and a variety of guests have been summarised for different guests by Fetterly (1964). The molar ratios are given by the following equations: n-paraffins m ¼ 0.65n þ 1.51; n-acids m ¼ 0.71n þ 1.08; n-alcohols m ¼ 0.66n þ 1.55. where m ¼ urea/guest and n is number of carbon atoms in the guest molecule. We consider the n-paraffins as an example; m will have nearly integral values only for n ¼ 10 (8.01), 13 (9.96) and 16 (11.91) (larger values of n not considered here). Thus the structures will in general be ‘incommensurate,’ a term defined in the caption to Fig. 6.5 and discussed below. This will be expressed experimentally by the presence of diffuse scattering in the diffraction patterns. We have not found information in the literature to test this somewhat simple-minded prediction except for hexadecane, where the description below shows that it is not entirely confirmed. A more sophisticated approach due to Rennie and Harris (1990) applies to an idealized one-dimensional structural model with infinite, rigid and periodic host framework (repeat distance ch) and equally spaced rigid, identical guest molecules (repeat distance cg) lined up along the tunnel axis (Fig. 6.7). A practical definition of ‘commensurate’ is that sufficiently small integers p and q can be found such pch qcg, all othersituations leading to ‘incommensurability.’ Broadly speaking, the inclusion complex is incommensurate if the energy of interaction between host and guest substructures does not depend upon the position of the guest along the tunnel while commensurability will be found if there are preferred positions for the guests (i.e. the guest can ‘lock in’ to the host). Another important parameter is the guest–guest offset, Dg. This will be Dg in 2/3 of all pairs of adjacent tunnels, and 2Dg for the remaining pairs. In principle, the offset concept can be extended to a third dimension and this would provide a complete description of the host–guest structure; in practice, there is invariably too little experimental information for reliable conclusions to be drawn. If the guest host periodicity ch

guest–guest offset ∆g

host guest host guest host cg guest periodicity

Fig. 6.7. Two dimensional representation of a tunnel inclusion complex showing host and guest repeat distances, and the guest–guest offset. (Adapted from Rennie and Harris, 1990.)

216


symmetry is lower than that of the host framework, then, for the urea inclusion complexes, there are six different arrangements of these two kinds of offset, and this leads to formation of domains in the crystal. Rennie and Harris (1990, 1992) have applied their mathematical analysis to a calculation of the guest periodicity in the urea–n-hexadecane complex, using Lennard-Jones equations for the host–guest and guest-guest interactions. ˚ obtained is in excellent agreement with recent experimental The value of 22.6 0.1 A ˚ less than that calculated taking only end-to-end guest results. This periodicity is 0.5 A interactions into account: in other words, the guest–guest interaction is repulsive. The overall structure is incommensurate and there is no lock-in. As we shall see below, this model breaks down when the temperature is reduced and the interaction energies become larger than kT. (ii) Cell dimensions and crystal structure at room temperature. The cell dimensions of ˚ , space group the hexagonal urea tunnel inclusion complexes are a 8.2, c 11.0 A P6122 (Type 2 of Fig. 6.3) or P6522 (Type 1); Z ¼ 2. The six urea molecules in the unit cell are located on twofold axes; thus the two nitrogens are crystallographically equivalent, a point important in interpreting some of the spectroscopic results, especially the 14 N pure NQR spectra of the complexes (see below). p The hexagonal tunnel has a height of ˚ (¼ c) and an edge of 4.8 A ˚ (¼ a / 3), leaving an internal cross-section 11.0 A ˚ (i.e. after taking the van der Waals radii of the urea available to the guest of 5.3 6.0 A molecules into account (Schlenk, 1955)). Full room-temperature structure determinations, based on the H (Bragg) diffraction patterns, have been reported for the urea inclusion complexes of n-hexadecane (Smith, 1952; Harris and Thomas, 1990b; Forst, Jagodzinski, Boysen and Frey, 1990; ZZZAKG 03 (XRD), 04 (ND)), 1,10-dichloro-n-decane (and the 1,10-dibromo analog (Harris, Smart and Hollingsworth, 1991)), dioctanoyl peroxide, diundecanoyl peroxide, lauroyl peroxide and bis-(6-bromohexanoyl) peroxide (Harris and Thomas, 1990b). The same host matrix structure appears in all these crystals at room temperature but there are differences at low temperatures, which are discussed below. The crystal structure of {3(urea)[1/4(n-hexadecane)]} has been studied more extensively than that of any other tunnel inclusion complex. The first determination of the room temperature structure of the urea framework, by Smith (1952), using photographically measured intensities, has been followed by an X-ray diffractometric study (Harris and Thomas, 1990b), and, most sophisticated of all, a combined X-ray and neutron diffractometric study (Forst, Jagodzinski, Boysen and Frey, 1990). There is no doubt about the overall arrangement of the urea molecules in the framework but various points of detail remain to be settled, the problems arising from the fact that H and G diffraction patterns are not completely independent, as was assumed in the earlier studies. We shall first be satisfied with an overall description, essentially following Smith (Fig. 6.8); the changes that occur on cooling are described below. Two-thirds of the urea molecules lie in the walls of a particular tunnel and one-third point away from it (but, of course, lie in the walls of other tunnels). As noted earlier, an important consequence of the spiral arrangement of the urea molecules in the tunnel walls is that the force field in the tunnel is rather uniform both along the direction of the tunnel axis and, to a lesser extent, normal to it; n-paraffin guest molecules with extended planar conformation do not have preferred locations along the tunnel axis at temperatures where kT is greater than the host-guest interaction energy but there are six lower-energy azimuthal orientations with molecular planes at multiples of 60 to the a axis; these energies depend on the position of guest


c

217

(b)

b

c

a

(a)

(c)

Fig. 6.8. Crystal structure of {3(urea)[1/4(n-hexadecane)]}: (a) View down tunnel axis [0001]. Hydrogen bonds are indicated by dashed lines. The hexagonal and orthohexagonal cells are shown by broken and dotted lines respectively. (b) Projection along the orthohexagonal a axis, showing hydrogen bonding. (c) View along the orthohexagonal b axis. The hydrogen bonds are shown by heavy broken lines, while the light double lines show hydrogen bonds connecting atoms directly below those shown in the figure; the disordered positions of the n-hexadecane chains are suggested by the zigzag lines. Although the hexagonal structure is drawn in space group P6122 (61 and 31 axes; Type 2 of Fig. 6.3), the absolute configuration of the crystal used was not determined. The spirals along c (cf. Fig. 6.2) run horizontally in (c). The oxygens are shown as the large open or stippled circles and nitrogens as circles of intermediate size. The atomic numbering, given in the original figure, has been removed to enhance the clarity of the diagrams. (Reproduced from Smith, 1952.)

along the tunnel axis (Parsonage and Pemberton, 1967). Each oxygen atom is hydrogen ˚ and the longer bonds bonded to four nitrogens, the shorter bonds each being about 2.93 A ˚ ; for comparison, we note that in (neat) urea each oxygen is hydrogen each about 3.04 A ˚ and the longer bonds bonded to four nitrogens, the shorter bonds being each 2.955(1) A ˚ (at 12K). Thus the hydrogen bonding of the urea molecules is only each 2.985(1) A slightly different in the two different crystalline environments. The diffraction patterns show sharp Bragg reflections with half-l values, indicating that ˚ , the host and guest arrangements originally being the true repeat along c is 22.02 A considered as commensurate. Later work (see Table 6.2 below) suggests that some modification may be necessary. The n-hexadecane molecule is slightly shorter in the ˚ ); this was accounted for by complex than in its fully extended conformation (22.84 A rotation of the end methyl groups or small deviations from full extension of the hydrocarbon chain. The volume occupied by an urea molecule in tetragonal urea at 298K is ˚ 3, and that of n-hexadecane in triclinic n-hexadecane (Norman and Mathisen, 1972; 75.6 A ˚ 3; thus the volume occupied by six urea molecules and 1/2 of an QQQFBP) is 406 A

218


˚ 3, which should be compared n-hexadecane molecule in their respective crystals is 656.6 A 3 ˚ with 645.5 A , the unit cell volume of the inclusion complex. There is a small contraction ( 1.7%) on formation of the complex at room temperature. (iii) Absolute configuration: The absolute configuration of the urea sublattice of the hexagonal urea inclusion complexes (which is another way of expressing the choice between the two enantiomorphic space groups, or of defining the sense of rotation of the spirals of urea molecules in the intersections of the tunnel walls) has been determined by the following, rather intricate, method (Schlenk, 1968, 1973a, b, c). Inclusion complexes with various long chain ketones (e.g. undecan-1-one) as guests were found to have the plate-like habit needed for measurement of the specific rotation of the crystals about the c axis (for theory see Nye (1957), Chapter 14). The specific rotation measured over the wavelength range 230–578 nm was found to be independent of the nature of the guest (chosen to be achiral). About half of the crystals measured were found to be laevorotatory; these all had Type A structures, which had earlier been defined (Schlenk, 1968) as those which, when seeding a racemic solution of urea and -methylbutyric acid decyl ester, gave a preponderant amount of inclusion complex containing the dextrorotatory (þ) ester. Such definition is essential as not all crystals of urea inclusion complexes have habits that permit measurement of the specific rotation of the crystal. The next step was to decide whether the Type A crystals had left or right handed urea spirals. This was done by studying, with the aid of models and diagrams, whether 1,2- and 1,3-dimethylalkanes of known optical configuration (e.g.(þ)(9R,10R)-9,10-dimethyloctadecane) which formed Type A crystals, fitted best into left or right handed urea spirals. The study was carried out for six 1,2- and three 1,3-dimethylalkanes and for various groups of methylalkanones, and it was shown that Type A crystals contained right handed spirals, i.e. they crystallized in space group P6122 (Type 2 of Fig. 6.3). These crystals, although containing right handed spirals, are laevorotatory i.e. they rotate the plane of polarized light counterclockwise. It would be useful (perhaps essential) to confirm these conclusions by carrying out a full crystal structure analysis of an ordered hexagonal urea inclusion complex containing a guest of known optical configuration. Alternatively the Bijvoet method could be applied, using the anomalous scattering of oxygen (Engel, 1972); this could be a rather taxing experiment (Rabinovich and Hope, 1980). 6.2.1.8 Determination of guest molecule conformation from diffuse x-ray scattering The theoretical background has been outlined above. Most studies have been made with oscillation or stationary crystal photographic methods, using filtered or (preferably) monochromatic radiation. The length of a guest molecule in a tunnel depends (Laves, Nicolaides and Peng, 1965) on (a) the number of atoms in the chain, (b) the amount of overlapping of the ends of the molecules in the tunnel, (c) whether the guest is locked into structural features of the host framework, (d) details of the guest conformation such as the degree of coiling, and (e) details of the guest configuration such as presence of cis (Z) or trans (E) isomers (Nicolaides and Laves, 1958). The reference state is usually taken as the extended form of the corresponding n-alkane. Molecular lengths are obtained from measurements of spacings while structures, projected one-dimensionally, are derived from the intensities of the reflections; the model is varied to obtain best agreement between observed and calculated structure factors. The method was first applied, for urea and


219

thiourea tunnel inclusion complexes, to the spacings of diffuse lines (Nicolaides and Laves, 1954) and then extended to comparisons of intensities (Nicolaides and Laves, 1956); it can of course be used for any system where host and guest sublattices are incommensurable and a one-dimensional approximation is applicable (e.g. the pentaiodide ion in trimesic acid pentaiodide (see Section 10.3)). Applications include demonstration of the all trans configuration of squalene in its thiourea complex (Nicolaides and Laves, 1965), location of substituents in long-chain compounds using intensity calculations (Nicolaides and Laves, 1954), determinations of molecular lengths in a wide variety of complexes (Nicolaides and Laves, 1963) and comparison of the lengths of particular guests in their urea and thiourea complexes (Laves, Nicolaides and Peng, 1965). The lengths of about 400 guest molecules in 35 different categories (e.g. n-alkanes, symmetrical monomethylalkanes, secondary alcohols, dialkylketones, chloroalkanes) were measured in one application (Lenne´, Mez and Schlenk, 1970) and later extended to other types of guest (Hadicke and Schlenk, 1972). Detailed investigations of the G (diffuse) diffraction patterns from urea inclusion complexes of diacyl peroxides (Harris and Hollingsworth, 1990) and n–C24H50 (Fukao, 1994a,b) have been reported. 6.2.1.9

Variation of structure with temperature, with particular reference to {3(urea)[1/4(n-hexadecane)]}

Interrelated static and dynamic changes take place in urea tunnel inclusion complexes (and, of course, in many other types of complex) on cooling, and these have been studied by a variety of techniques. In general terms there are clear behavioral resemblances among the various hexagonal urea inclusion complexes but the details depend on the nature of the guest. For example, many investigations have shown that urea-n-paraffin systems all appear to show the common feature of a first order transition on cooling. The temperatures of the major transition as measured by DTA (Chatani, Anraku and Taki, 1978; ZZZAKG02) are shown in Fig. 6.9, and this provides additional evidence for influence of the guest. Only two values are available for odd-numbered paraffins (n ¼ 11, 15), and this is not sufficient to decide whether odd and even paraffins behave differently, especially in view of the spread of values for the even paraffins. We briefly recall some of the salient features of the physical methods that have been used; these remarks, adapted from some of the papers to be quoted later, will also apply to analogous studies of many other types of molecular complex and compound. Diffraction measurements will be most successful in providing information about the ordered parts of the structure; here, specifically, about the urea or thiourea framework over the whole temperature range and about the guest molecules at temperatures low enough for the guests to be ordered. Analyses based on measurements of (X-ray or neutron) Bragg reflections alone give information about the average structure, while studies of diffuse scattering give information about the disorder in the crystal. It is not possible to distinguish between static and dynamic disorder using X-ray diffraction but this can be done by neutron diffraction, measurements of elastic scattering (i.e. without change of wavelength (energy)) giving information about static disorder while measurements of inelasticscattering give information about dynamic disorder. Calorimetric measurements will provide information about thermal processes occurring in both host and guest sublattices; inferences about changes in molecular arrangement occurring in a transition are often

220


220 200 180 Ttr(K) 160 140 120 100 9

14

24 29 19 ‘n’ in Cn H2n + 2

34

Fig. 6.9. Temperatures of the (apparently) first order phase transition (hexagonal to orthorhombic) in {urea[n-paraffin]} tunnel inclusion complexes plotted against number of carbons in the guest. Additional values (up to n ¼ 45), that do not change the overall picture, are given by Fukao, Horiuchi, Taki and Matsushige, 1990. (Adapted from Chatani, Anraku and Taki, 1978.)

checked by comparing measured and calculated entropy changes. Calorimetric and diffraction studies refer to the system as a whole. NMR measurements provide a very powerful technique for studying molecular dynamics in the solid state. Proton NMR spectroscopy has been extensively used in the past but has the disadvantage that line width, second moment and relaxation data reflect average properties over the whole spin system. The 2H NMR spectrum reflects, because of the quadrupole interaction involved, the behaviour of isolated nuclei and is essentially not affected by magnetic interactions with other nuclei. For systems like the urea and thiourea tunnel inclusion complexes there is a particular advantage in being able to deuterate either host or guest and hence study one or other by 2H NMR spectroscopy. Motion in the range 103 to 108 Hz may be described as of ‘‘intermediate’’ rate on the 2H NMR quadrupolar time scale and such motion affects the 2H NMR line shape in a manner dependent upon its 3 angular extent and also its rate. Motion of rate < 10 Hz8 has no effect on the line shape and is said to be in the ‘‘slow’’ regime. Motion of rate > 10 Hz is said to be in the ‘‘fast’’ limit, and the line shape becomes insensitive to further increase in the rate of motion, though still characteristic of the angular extent of the motion. In most studies a particular technique is applied to a number of complexes but there is more to be learned when the results obtained by a variety of techniques are compared for a particular complex; we shall do this for {3(urea)[1/4(n-hexadecane)]} (referred to as {urea[n-C16H34]}), which has been studied more comprehensively than other complexes. The molar ratio of urea to n-C16H34 has been given as 11.91, and the complex could therefore be expected to be dimensionally commensurate, but the mismatch, although small, is important, as is shown by the diffraction studies discussed below. We start with the results of physical measurements, which have been gathered together in Fig. 6.10, using the same temperature scale to facilitate comparison among the results. The Cp–T values (Pemberton and Parsonage, 1965) show deviations from a smooth curve only in the 130–160K region, with an apparently first-order transition appearing at 152K, and some


221

T(K) 100

200

300 (a)

a (Å)

8.24 8.19 8.14

14.0

XRD

13.8

1295

(c)

1255

(e)

203 J/mol K

16 130

8

140

80 75 70

Cp

12

85

150

Cp(J/mol deg)

1275

160 T(K)

(d) 1H

4 0

12

NMR

(d)

8

6 4

4 2 0 100

200 T(K)

LINE WIDTH (gauss)

SECOND MOMENT (gauss2)

Vol (Å3)

14.2 b(Å)

(b)

0 300

Fig. 6.10. Comparison of the results of various types of physical measurements on {urea [n-C16H34]}. (a, b, c) The XRD measurements of cell dimensions are from Chatani, Anraku and Taki (1978). The high temperature form is p given as an orthohexagonal cell with twice the volume of the ˚ conventional hexagonal cell (bO ¼ 3(aH). The c-axis varies smoothly with temperature (11.015 A ˚ at 153K and 10.988 A ˚ at 98K) and does not show a discontinuity at 150K. (d) The at 298K, 11.007 A 1 H wide line NMR measurements (left: Second moment; right: line width) come from Umemoto and Danyluk (1967) and refer only to the behavior of the guest as d4-urea was used. (e) The specific heat measurements, from Pemberton and Parsonage (1965), refer to one mol of urea. Comparison with the Cp–T and 2H NMR (Fig. 6.11) measurements shows that finer temperature intervals are needed for the XRD and 1H NMR measurements. The curves are guides to the eye.

anomalous behaviour around 135K; these results are shown on an expanded temperature scale. NMR 1H line widths and second moments for {urea-d4[n-C16H34]} increase rapidly below 140K (Gilson and McDowell, 1961; Umemoto and Danyluk, 1967). The rigid lattice approximation appears to apply (by extrapolation) below about 100K but observed values are appreciably smaller than calculated values above 140K; this will be

222


considered below. The temperature dependence of the bands in the C–H stretching region of {urea[n-C16H34]} has been studied by Raman spectroscopy (Snyder, Scherer and Gaber, 1980). All bands become significantly narrower on cooling but the intensity of the intense methylene antisymmetric stretching fundamental at 2885 cm1 shows, anomalously, a sharp change of slope at 148K when plotted against 1/T; presumably this corresponds to the major peak in the Cp–T curve. 2 H NMR spectra of polycrystalline {urea[n-C16D34]} have been measured over the temperature range 295–115K (Harris and Jonsen, 1989) (Fig. 6.11).

ω-reorientations 60° jumps

D3C

At 115K, the CD3 group is rotating rapidly about its symmetry axis, but there is also evidence for a combination of 60 jumps and !-reorientations (not simultaneous) below 146K, where there is a striking change in the appearance of the spectra, associated with the major peak in the Cp–T curve. The motions remain essentially unchanged up to 295K, where they include 60 jumps about the long axis of the molecule, torsional libration (of approximately 25 ) about the penultimate C–C bond and rotation of the CD3 group. The most comprehensive diffraction studies yet made of any urea inclusion complex by a particular group are of {urea[n-C16H34]}, using Bragg and diffuse scattering of both X-rays and neutrons (Forst, Jagodzinski, Boysen and Frey, 1990; Forst, Boysen, Frey, Jagodzinski and Zeyen, 1986; Forst, Jagodzinski, Boysen and Frey, 1987). Even so, as a b, b⬘

b, b⬘ 295 K

b a

b

a

b 147 K

b

146 K

b

115 K

b a

100

0 –100 kHz

Fig. 6.11. 2H NMR patterns of polycrystalline {urea[n-C16D34]}. Three superimposed powder patterns can be identified, labeled 0 a 0 (CD3 group), 0 b 0 (CD2 adjacent to CD3) and 0 b 0 (all other CD2); 0 b 0 and 0 b 0 become resolved only at higher temperatures. (Reproduced from Harris and Jonsen, 1989.)


223

Table 6.2. Summary of structural data for {urea[n-C16H34]} (reproduced from Forst, Boysen, Frey, Jagodzinski and Zeyen, 1986) T (K)

Phase

>380 380

I

365

II

148

III

120

IV

Average structure Decomposition to tetragonal urea þ n-C16H34 Hexagonal P6122, a0 8.2, ˚ c0 11 A Trigonal P312, ˚ ¼ 2c0 c 22 A Orthorhombic P 212121; a 8.2, ˚ b 14.2, c 11 A Triclinic (two lattices?)

Order/disorder

Longitudinal disorder in adjacent tunnels; orientational disorder in each tunnel. Longitudinal disorder in adjacent tunnels; orientational disorder in each tunnel. Mutual longitudinal deformation of end groups; (microdomains in ab plane). Lateral orientational order of guests in adjacent tunnels. ˚ k c. Domain structure (host) 200 A

these authors note, complete clarification was not obtained and there is scope for further work. There are also studies by other groups, using both single crystal (Harris, Smart and Hollingsworth, 1991; Chatani, Taki and Todokoro, 1977; ZZZAKG) and polycrystalline samples (Harris, Gameson and Thomas, 1990). We summarize the major conclusions here and refer the reader to the original papers for additional detail. The phase relations found in the system are given in Table 6.2. The hexagonal structure stable from 380 to 365K is that characteristic of most urea tunnel inclusion complexes at room temperature, and is the structure analyzed for {urea[n-C16H34]} at room temperature (Section 6.2.1.7(ii)). The I , II transition is found only for the n-hexadecane guest and is perhaps due to similarity between the length ˚ , including allowance for Van der Waals radii of of a hexadecane molecule ( 22.6 A ˚ ); it is outside the temperature terminal methyl groups) and twice the c-axis repeat (22.0 A range covered in Fig. 6.10. As the terminal methyl groups cannot exactly match the positions of the potential wells in the urea matrix there is a mutual interaction between framework and guest molecules leading to a doubling of the c-axis repeat and more flexibility for lock-in. One should also note that there is evidence from 2H NMR spectroscopy for 14% gauche content in the C2–C3 bond and 9% in the C3–C4 bond of n-hexadecane (in {urea[n-C16H34]}) at room temperature (Cannarozzi, Meresi, Vold and Vold, 1991). The additional degrees of motional freedom of the guest may account for the lower values of the second moment of the hexadecane 1H NMR lines (1.5 G2 at 240K), compared both to calculation (6.0 G2, with free rotation of the hydrocarbon chains) and to the experimental values obtained for the dodecane (2.2 G2), tetradecane (2.5 G2) and octadecane (2.9 G2) complexes. Domain boundaries may occur when the mismatch is such that methyl groups are close to potential maxima. The II , III transition at 148K, analogs to which occur in the other urea[n-paraffin] complexes, corresponds to the peak in the Cp–T curve at 152K, to the break in the plot of I(2885 cm1) against 1/T, to striking changes in the 2H NMR spectra, and in the Debye–Scherrer X-ray powder patterns obtained by Harris, Gameson and Thomas (1990). Major changes in the 1H NMR spectra also occur in this temperature region. This major

224


transition occurs when the thermal energy (RT 1.25 kJ/mol) becomes comparable with the potential barrier to the free rotation of the guest molecules; the values for n-CnH2nþ2 guests (n ¼ 10 to 36) vary from 111 to 218K (Fig. 6.9). According to Chatani et al. (1977), the orthorhombic crystals obtained after transformation from the hexagonal (or trigonal) phase have space group P212121 (Z ¼ 12), and are triply twinned; however, Forst et al. (1990) show that there are six orientations, characterized by a small angular deviation ( 1 at 78K) of the orthorhombic a, b axes from the corresponding orthohexagonal axes. The twinning implies that a full structure determination of the orthorhombic form would be difficult, as both Harris et al. (1991) and Forst et al. (1990) have noted. The structure of the isomorphous orthorhombic {urea[1,3-butadiene]} has been reported (Chatani and Kuwata, 1975). The temperature dependence of the cell dimensions is shown in Fig. 6.10. There are no signs of discontinuities at 135K, where there is a minor peak in the Cp–T curve or at 120K, where Forst et al. (1990) have reported a first order transformation (see below). However, there are abrupt changes (in opposite directions) in a and b at 145K, which correlate well with the change in the 2H NMR spectrum at 146K (Fig. 6.11). The c axis and the volume appear to be continuous (although more detailed measurements would be desirable) and this accords with the remark (Chatani et al., 1977) that the transformation does not show any hysteresis. The II , III transformation at 150K is sometimes referred to as an order-disorder transformation, presumably because ordering of the guests occurs. The entropy of transition at 152K is 3 J/mol K for the formula unit {3(urea)[1/4 (n-hexadecane)]}; testing this value against R ln m gives 1.43 for m. The transformation appears, from a thermodynamic point of view, to have both first and second order features. The II , III transition is associated with lateral ordering of the guests in neighbouring tunnels. Below the transition temperature only fluxional oscillations are possible about the mean guest molecule orientation in the tunnels, and the urea framework distorts to conform to this orientation (Fig. 6.12). At 120K there is a further ordering process III , IV, shown by a sharpening of the diffuse scattering and splitting of Bragg reflections. Neutron diffraction shows that this is a first order transition involving mainly static processes (i.e. there is essentially no inelastic neutron scattering). There is additional longitudinal ordering of the guest molecules in adjacent tunnels which induces higher-order modulations into the urea framework leading to formation of a domain structure. The domains ˚ in length along c, thus containing chains of about nine n-hexadecane are about 200 A molecules. Curiously, there is no evidence for this transition in the specific heat or spectroscopic measurements, although it does occur in a region of rapid change in the NMR spectra; conversely, the activity in the Cp–T curve around 135K does not appear to be reflected in structural changes. Clearly there is room for further study of the Cp–T behaviour. The complexity of the structure of phase IV at 32K is illustrated by the remarkable oscillation photograph shown in Fig. 6.13, and this also shows the wealth of information potentially available from appropriate diffraction photographs. There are four layer lines (l ¼ 1 4, marked by L) due to the periodicity in the c direction of the host urea ˚ c-axis repeat). There matrix (note that the various l indices are defined in terms of the 11 A are nine layer lines (l 0 ¼ 1 9, marked by L 0 ) due to the periodicity in the c direction of the hexadecane guest; at lower diffraction angles the L and L 0 patterns cannot be resolved. The L 0 pattern corresponds to the diffuse s layer pattern appearing at higher temperatures


225

bh

b0 a0

6

9.7

9.1

0

4.7

7

9.26

4.71

Fig. 6.12. Schematic structures of {urea[n-C16H34]}. Upper diagram: The hexagonal structure viewed down the tunnel axis; the paraffin chain is shown in various orientations in order to emphasise its disorder. The orthohexagonal cell is also outlined. Lower diagram: The orthorhombic structure viewed down the tunnel axis, showing the distortion introduced by ordering of the orientations of the paraffin chains (schematic representation). No attempt has been made to represent the multiple orientations of the orthorhombic cell with respect to the orthohexagonal cell. (Reproduced from Chatani, Anraku and Taki, 1978.)

and identified as due to the guest (Forst, Jagodzinski, Boysen and Frey, 1990). There is one layer line (l00 ¼ 1, marked by L00 ) which corresponds to the d diffuse layer system perpendicular to c* and considerably broadened in the c* direction, which is found at ˚ C–C–C repeat period of the higher temperatures. This has been ascribed to the 2.56 A almost-periodic alkane molecule. The fine splitting of some of the L reflections is ascribed ˚ domain structure) induced by to higher-order modulations of the host lattice (the 200 A longitudinal ordering of C–C–C portions of the guest in neighbouring tunnels. The various ordering processes that occur as the temperature is reduced can be understood in overall terms as a progressive reduction in the motional freedom of the guest molecules leading towards their crystallization as a separate sublattice. However, the residual diffuse scattering on Fig. 6.13 shows that a completely ordered superstructure of host and guest is not achieved even at the lowest temperatures. Although the guest

226


sublattice remains incommensurate with the host framework, there is a mutual interaction which gives rise to the complicated domain structure. Forst et al. (1987) conclude that it would ‘‘be more adequate to term this class of substance as paraffin enclosure compounds instead of urea inclusion compounds.’’ Our emphasis has been, of course, in quite the opposite direction. The details of the diffraction patterns from urea inclusion compounds containing n-alkane guests depends on the nature of the guest, and a wealth of information is emerging from such studies. An excellent summary comes from Weber, Boysen and Frey (2000), which we quote in abbreviated form: ‘‘Urea inclusion compounds (UICs) . . . consist of a honeycomb-like urea host structure with a quasi-hexagonal lattice, forming open tunnels parallel to c in which various [guests, here n-alkanes] are embedded. UICs belong to the class of composite crystals, where host and guest substructures have different translational and/or point symmetry. In particular, they show complex disorder phenomena for various reasons: (i) both substructures have a different ‘‘dimensionality,’’ viz. in the ‘‘tubes’’ of the three-dimensional ordered urea–host framework, a predominantly one-dimensional guest structure, the alkane chains, are embedded; (ii) the orthorhombic or monoclinic eigen symmetry of the alkane molecules is in competition with the hexagonal symmetry of the urea host structure; (iii) both substructures have – along the unique axis – non-matching, in general incommensurate, translational periods . . . [The] interactions [between both substructures] are responsible for frustrations which become evident from mutual (incommensurate) modulations and disordering, [which is] highly temperature dependent, including possible structural phase transformations. The basic structural features of the host and guest structures are reflected by typical diffraction patterns: sharp Bragg reflections of the three dimensional urea host and narrow diffuse layers (‘s-layers’) perpendicular to the c axis corresponding to the onedimensional guest structure. As the s-layers show some intensity modulations and are superimposed by weak Bragg-like reflections, however, this rough division is not fully adequate, and lateral correlations between the guest molecules cannot be neglected. In addition, mutual modulations of host and guest lattices give rise to three-dimensional and one-dimensional satellite scattering accompanying the Bragg reflections and the s-layers respectively . . . .Another characteristic diffuse diffraction feature of UICs is the set of so-called [diffuse] ‘d-bands’ . . . explained by a longitudinal and lateral disorder of the guest molecules.’’ The structures of the low-temperature orthorhombic phases of the urea inclusion complexes of 1,10-dibromo-n-decane (ROPQOC) and 1,12-dibromo-n-dodecane (ROPQUI) have been determined (Yeo and Harris, 1997)), and of that with 1,10decanedicarboxylic acid (Yeo, Harris and Guillaume, 1997; NATNIF). The same host matrix structure appears in all these crystals at room temperature but there are differences at low temperatures due to different manners of ordering of the guests. The first two ˚ ) discomplexes are isomorphous and have the P212121 structure (cell 8 14 11 A ˚ ); the cussed above, but the third transforms to space group C2221 (cell 16 28 11 A ˚ for all three crystals. These more complete structure analyses were tunnel axis is 11 A made possible because the phase transitions were ‘‘single crystal’’ to ‘‘single crystal,’’ complications due to the multiple twinning found in other systems being absent. In all three low-temperature structures the lateral ordering is fairly complete but longitudinal ordering (the location of the guest molecules along the tunnel axes) is still lacking, i.e. the


227

Table 6.3. Crystal data for urea inclusion complexes showing deformation of the room-temperature hexagonal urea framework to orthorhombic or monoclinic at lower temperatures. The tunnel axes are shown in bold. N ¼ number of formula units in the unit cell Formula {(urea)3[0.21(1,10-dibromodecane)]} at 108K {(urea)30.27[(1,12-dibromododecane)]} at 108K {(urea)6[(1,10-decanedicarboxylic acid)]} (at 173K) {(urea)3x[1,6-dibromohexane]}

a/

b/

c/

N Space group

Reference

8.080 13.945* 11.007

4

P212121 (1)

8.186 14.133* 10.930

4

P212121 (1)

16.305 28.321* 11.000

8

C2221

(2)

13.381* 4

P21/n

(3)

8.560 10.889 92.82

p ˚ 14 A ˚ , or multiple thereof, thus pointing up the relationship between orthorhombic and ortho* 38A hexagonal cells. Notes: (1) Yeo and Harris, 1997; the axes have been reoriented from those in the original paper in order to emphasize the relationship between RT and LT structures. (2) Yeo, Harris and Guillaume, 1997; this entry should be compared with that for 1,10-undecanedione guest in Table 6.4, showing that the ordering in the a, b, c directions differs in the two crystals although both have the same C2221 space group. (3) see Hollingsworth and Harris, 1996, Table 4.

structure remains incommensurate. The coordinates given for the atoms of the guests in these structure determinations are unlikely to be reliable, and coordinates are not even given for the bromines of the first two structures! Many lessons can be learned from these results. One is not surprising – the experiments of twenty and thirty years ago bear repeating in order to take advantage of the much greater power now available from new technologies. For example, the cell dimensions determined laboriously from diffraction photographs of single crystals at rather wide temperature intervals could be measured at much closer intervals and with greater accuracy by using neutron diffraction or synchrotron radiation on polycrystalline samples (Finney, 1995). A second is the wealth of detail obtainable from single crystal XRD photographs using monochromatic radiation and stationary and/or oscillating samples; surely area detector techniques have much to offer here. A third is the synergistic effect of investigating the same set of phenomena by as many different methods as possible, and comparing the results in detail, as we have tried to illustrate in Figs. 6.10 and 6.24. A fourth is that study of commensurate structures yields much more, and more firmly based, information than does that of incommensurate structures, but for most guests of a homologous series only the latter are available. 6.2.1.10 Interruption of urea framework by host–guest hydrogen bonding In the previous section we have reviewed deformation of the hexagonal urea framework normally found at room temperature to lower-symmetry arrangements. However, we emphasize that the urea framework remains intact despite the deformation; urea–urea

228


L

4.5

L⬘

9

4.25 4

L⬙

1

8 7

3

6 5

2

4 3

1

2

0.5 0.25 0

1

Fig. 6.13. Section of a normal beam oscillation photograph (Cu K1 radiation) of {urea[n-C16H34]} at 32K taken with c vertical showing the complicated structure of phase IV. Indices L mark the Bragg layers of the host, L 0 those of the guest (corresponding to the diffuse s sheets) and L00 the position of the d–band; s and d are defined by Forst, Boysen, Frey, Jagodzinski and Zeyen (1986) and Forst, Jagodzinski, Boysen and Frey (1987) and in the text. (Reproduced from Forst, Boysen, Frey, Jagodzinski and Zeyen (1986).) (I am greatful to Dr Hans Boysen for an original print.)

hydrogen bonds are not broken. The deformation results from the increased host–guest interaction that occurs on cooling. More drastic effects can also occur. These are of two kinds. In the first, inclusion complexes are not formed but instead mixtures of urea and a second component crystallize to give binary arrangements of many different sorts. This often, but not only, occurs with lower members of a homologous series (see Table 7 of Hollingsworth and Harris, 1996); for example, 1,3-dicyanopropane, 1,4-dicyanobutane and 1,5-dicyanopentane form 1 : 1 hydrogen bonded compounds with urea (not of the tunnel variety), while tunnel inclusion complexes are formed by some of the higher ,!-dicyanoalkanes (that with 1,6dicyanohexane is monoclinically distorted). We do not discuss here the structures of binary adducts that are not tunnel inclusion complexes. We now consider the second kind of effect – formation of interrupted structures in which some framework urea–urea hydrogen bonds are replaced by urea–guest hydrogen bonding. These are examples of the general phenomenon found in tunnel inclusion complexes where the host-to-guest interactions are sufficiently important to interrupt, but


229

Table 6.4. Crystal data for urea inclusion complexes showing interruption of the urea framework. The tunnel axes are shown in bold. N ¼ number of formula units in the unit cell Formula unit

a/

b/

c/

N

Space group

Reference

{(urea)7[2,7-octanedione]} {(urea)8[2,9-decanedione]} {(urea)9[2,10-undecanedione]} {(urea)6[1,8-dicyanoo¨ctane#]}

8.211 8.229 8.345 15.125

8.211* 8.229* 13.939 7.487/ 104.10

76.91 44.16 32.982 25.815

6 3 4 4

P6122 P3112 C2221 C2/c

(1) (2) (3) (4)

p ˚ (or some * To obtain the orthohexagonal cell these values must be multiplied by 3, and approximate to 14 A multiple thereof ). # sebaconitrile; structure determined at 98K; there is no tunnel axis in simple terms (see below). Notes: (1) Brown, Chaney, Santarsiero and Hollingsworth, 1996; structure determined at 291K. (2) Hollingsworth, Brown, Hillier, Santasiero and Chaney, 1996. (3) Brown and Hollingsworth, 1995. (4) Hollingsworth, Santarsiero and Harris, 1994.

not destroy, the host framework by formation of hydrogen bonds between host and guest (cf. some of the tunnel inclusion complexes of trimesic acid). Hollingsworth and Harris (1996) have discussed the urea inclusion complexes of alkanediones, ,!-dicyanoalkanes and ,!-dihaloalkanes by comparing the structures of the inclusion complexes found within each homologous series (for a particular family of guests) and showing how the very interesting physical properties of the crystals can be explained in terms of their structures. We shall not repeat this material but rather emphasize the host–guest interactions in particular commensurate crystals of these groups as illustrations of the wider topic of inclusion complexes where the usual structures are distorted or deformed because of appreciable host–guest interactions. We first discuss the complexes with host-guest hydrogen bonding. Detailed results have been given for {(urea)7[2,7-octanedione], for which atomic coordinates are available (TOZHOF; Brown, Chaney, Santarsiero and Hollingsworth, 1996). There are 3 1/2 urea molecules and 1/2 2,7-octanedione molecules in the asymmetric unit, and thus 42 ureas and six guests in the unit cell. The 1/2 urea molecule is located at Wyckoff positions (b) (2x, x, 1/12, etc), with two fold symmetry, while the other ureas are in general positions. The guest molecules are located about Wyckoff positions (b) (x, 0, 0 etc.) and also ˚ have two fold symmetry. The hydrogen bonds between ureas range from 2.94 to 3.08 A ˚ in length. A limited view down while that between carbonyl oxygen and urea N is 3.08 A the tunnel axis is shown in Fig. 6.14. The usual structure is modified, but not drastically so. Packing diagrams have been given for the {(urea)8[2,9-decanedione]} and {(urea)9[2,10undecanedione]} structures and these show host–guest hydrogen bonding similar to that described for the octanedione complex. However, atomic coordinates have not been deposited and a detailed description is not possible. The departures from the usual structural arrangement are greater in {(urea)6 [1,8-dicyanoo¨ctane]} (LEMHIU, coordinates available; Hollingsworth, Santarsiero and Harris, 1994), and this makes the structure rather difficult to depict. There are three ureas in the asymmetric unit, and 1/2 of a guest molecule, which has two fold symmetry and is


230

Fig. 6.14. View down the tunnel axis of {(urea)7[2,7-octanedione]}, showing a single guest molecule hydrogen bonded to two urea molecules of different helices (in the sense of the lower part of Fig. 6.3). These two urea molecules are turned away from the tunnel axis by 38.5 . The O¼C . . . C¼O torsion angle of the guest molecule is 160 . (Reproduced from Brown, Chaney, Santarsiero and Hollingsworth, 1996.) C

C

z y x

A

A

Fig. 6.15. Stereoview down the tunnel axis of two unit cells of {(urea)6 [1,8-dicyanoo¨ctane]} (LEMHIU). Many atoms have been removed for clarity. The urea atoms are shown by smaller circles and the atoms of the guest by larger circles. N . . . O hydrogen bonds between molecules of the urea framework and N . . . N hydrogen bonds linking urea to guest are shown. Mutual displacement of parts of the urea framework are shown. (Data from Hollingsworth, Santarsiero and Harris, 1994.)


231

urea-urea O...N H-bonds C

G-urea N...N H-bonds

B

1,8-dicyanooctane guest (G)

z A y

x

Fig. 6.16. Part of the {(urea)6 [1,8-dicyanoo¨ctane]} structure viewed down [010]. Many atoms have been removed for clarity. Hydrogen bonding of the guest molecules (larger circles) to the urea framework (smaller circles) is shown. (Data from Hollingsworth, Santarsiero and Harris, 1994.).

located about Wyckoff position (e). The stereoview of Fig. 6.15 shows that the hexagonal framework is not complete, and the projection of Fig. 6.16 shows that the guest molecules are offset from one another, i.e. there are zigzag tunnels instead of the linear tunnels of the usual structural arrangement. The three independent oxygens are hydrogen bonded to ˚ ; the cyano nitrogen is linked to nitrogens of ureas with lengths ranging from 2.88 to 3.01 A ˚ ; as before, the host . . . two different urea nitrogens with N . . . H–N distances of 3.19 A guest interaction does not seem to be particularly strong. We shall not describe this rather complicated pattern of hydrogen bonds in more detail. 6.2.1.11 Rhombohedral urea, thiourea and selenourea tunnel inclusion complexes (i) Cell dimensions and structure: These complexes all have the structure shown as Type 3 in Fig. 6.3. The chemical formula is expressed as before as {3(host)[x(guest)]}; ˚ , 104.5 ; thiourea the rhombohedral cell dimensions are: urea complexes, a 9.0 A ˚ , 104.3 ; space ˚ , 104.0 ; selenourea complexes, a 10.4 A complexes, a 10.0 A group R 3 c, Z ¼ 2. The host molecules are located on two fold axes. The structures of these complexes are usually described in terms of the triply primitive hexagonal unit cell containing 18 host molecules (i.e. Z ¼ 6), with dimensions (in the same order as above) of ˚ . The hexagonal tunnel has height 11.0 a 14.2 (15.8; 16.5); c 11.0 (12.5; 12.9) A ˚ and edge 4.8 (5.37; 5.5) A ˚ . The relationship between the rhombohedral (12.5; 12.9) A unit cell and the hexagonal tunnel is shown in Fig. 6.17. Despite the resemblance to the (primitive) hexagonal urea structure, the rhombohedral structure differs in that the arrangement in the tunnel walls is not spiral but layered; the rhombohedral crystals are centrosymmetric, not chiral. There is a layer of sulphur atoms (using a thiourea complex as an example; see Fig. 6.19) pointing into the tunnel at z ¼ 0, followed by two

232


C

O

b

a Z

Y X

a

Fig. 6.17. Relationship between the rhombohedral and (low-temperature) monoclinic unit cells and the outline of the hexagonal tunnel, illustrated for thiourea complexes. The axes of the rhombohedral cell are shown by the full lines x, y and z while c ¼ x þ y þ z is the threefold axis of the corresponding triply-primitive hexagonal cell. Only a few of the thiourea molecules are shown. The axes of the monoclinic cell are a ¼ x, b ¼ y þ z, c. (Reproduced from Cle´ment, Mazieres, Gourdji and Guibe´, 1977.)

layers of thiourea molecules in the walls of this tunnel at z ¼ 1/6 and 2/6. Thus the potential field experienced by a guest molecule within the tunnel varies with z and this gives rise to a tendency to localize guest molecules in particular regions along the tunnel axis, and to favor inclusion of molecules whose dimensions along the tunnel axis are multiples of c/2; for example in the {3(thiourea)[cyclohexane]} complex (Lenne´, 1954) the cyclohexane molecules are localised in the vicinity of the sulphur atoms at z ¼ 0 and 1/2. We mention here that there is more variability in the cell dimensions of the isostructural rhombohedral tunnel inclusion complexes than in the analogous hexagonal complexes; also there do not appear to be any examples of host-guest interactions which interrupt the host framework. Full X-ray diffraction structure analyses have been made of the isomorphous {3(thiourea)[adamantane]} (VADWUS) and {3(selenourea)[adamantane]} (VADXAZ) complexes (Gopal, Robertson and Rutherford, 1989) and a stereodiagram of the first is shown in Fig. 6.18. The adamantane guest is two fold disordered in the tunnels. Each S (Se) atom forms four (two pairs of ) hydrogen bonds, with lengths 3.462(7) and ˚ (3.51(2) and 3.65(2) A ˚ ). The NH . . . Se hydrogen bonds are weaker than their 3.496(5) A oxygen and sulphur counterparts, and this leads to greater variability in cell dimensions of selenourea inclusion complexes. 2H NMR studies show that there are no phase transitions between 333 and 119K (MacIntosh, Frazer, Gruwel, Wasylishen and Cameron, 1992).


233

Fig. 6.18. Stereoview of a tunnel in {3(thiourea)[adamantane]}. Only one orientation of the twofold disordered adamantane is shown in each site. (Reproduced from Gopal, Robertson and Rutherford, 1989.)

The structure of {3(tu)[CCl4]} at 170K has been determined (a ¼ 15.539, ˚ ; Fait, Fitzgerald, Caughlan and McCandless, 1991; FABTAD10); this is c ¼ 12.529 A phase III (although the space group was reported as R 3 rather than the usual R3c, this has been reinterpreted by Marsh et al., 2002 (FABTAD11)). The guest molecule is disordered. Specific heat measurements (Sekii, Matsuo and Suga, 1990) show a first order transition (from phase I to II) at 41.3K (DHtrans ¼ 149 J/mol; DStrans ¼ 3.7 J/mol K) and a second order transition (from phase II to III) at 67.2K (DHtrans ¼ 241 J/mol; DStrans ¼ 3.9 J/mol K). The 35Cl NQR spectrum shows two lines at 7.4K, i.e in the phase stable in the lowesttemperature region (Adolphi, Conradi and Matsuo, 1994). This has been interpreted in terms of the ordered structure shown in Fig. 6.19, the CCl4 molecule being located at a site of two fold symmetry. At 170K, the CCl4 molecule is disordered over three sites, and this is compatible with the total entropy change between phases I and III (R ln 2.6). There is as yet no information about the structures of phases I and II from diffraction. A14N NQR study of the II to III transition does not seem to have given easily-interpretible results (El Ghallali, Gourdji, Guibe´ and Peńeau, 1994). A new structural feature appears in the isomorphous 3 : 1 complexes of thiourea with (6-benzene)CrCO3 (GARTIC10), (4-trimethylenemethane)-FeCO3 (SESLUX) and (5-cyclohexadienyl)MnCO3 (GARTUO10); these have essentially the same structure as, say, {3(thiourea)[adamantane]} but the guest molecules are ordered in a polar array and the space group is R3c (Fig. 6.20). These complexes were designed to be noncentrosymmetric in order to be capable of second harmonic generation (Tam, Eaton, Calabrese, Williams, Wang and Anderson, 1989). The {3(thiourea)[(6-benzene) CrCO3]} complex has about twice the SHG efficiency of urea (which provides a widely used reference standard).

234


Layer 1

Layer 4

S C In-plane CI Out-of-plane CI

Layer 1 Tunnel axis

Fig. 6.19. Postulated structure of {3(tu)[CCl4]} in its lowest-temperaturephase I, based on 35Cl NQR measurements. The twofold axis of the guest molecule is normal to the tunnel axis. (Reproduced from Adolphi, Conradi and Matsuo, 1994.)

b

a

b c

Fig. 6.20. Crystal structure of {3(thiourea)[(6-benzene)Cr(CO)3]} at 203K; a ¼ 16.130, ˚ , space group R3c, Z ¼ 6. The upper diagram shows the view down the tunnel c ¼ 12.569 A axis, with two guest molecules in each tunnel; the lower diagram is the view normal to the tunnel axis, showing only the polar arrangement of guest molecules, thioureas being omitted for clarity. (Reproduced from Tam, Eaton, Calabrese, Williams, Wang and Anderson, 1989.)


235

6.2.1.12 Monoclinic complexes derived from the rhombohedral complexes Either cooling or the particular shape of the guest can lead to distortion of the rhombohedral lattice to monoclinic; a number of examples are known and it is convenient to summarize the structural results before discussing the behavior of the complexes on cooling. We give five examples in Table 6.5; the tunnel axis is [001] in all the examples and the disordered guests are all located about crystallographic centers. The first four structures are isomorphous, and the fifth isostructural. The structure of {3(thiourea)[H2C¼C(CH3)–C(CH3)¼CH2]} has been determined at 143K (the complex is unstable at room temperature, while the guest polymerizes rapidly under X-radiation at 193K). The relationship between triply primitive hexagonal axes and the monoclinic axes are: am ¼ 2=3ah þ 1=3bh 1=3ch ;

bm bh ;

cm ch :

The appreciable deformation (Fig. 6.21) of the previously hexagonal tunnels was ascribed to the ordered longitudinal packing of the guest molecules in the tunnels. The 2,3-dichloro-1,3-butadiene complex is isomorphous with that of dimethyl-butadiene, as are the complexes with polymerized guests (Chatani and Nakatani, 1972). Although the cell dimensions of the 2,3-dimethyl-1,3-butadiene and 1,5-cyclooctadiene complexes are similar, the symmetry elements of the space group are differently oriented and the tunnel axis is along [001] for the first of these complexes and along [101] for the second (structure at 173K).

˚ , deg, A ˚ 3) for monoclinic {3(thiourea)[guest]} complexes; the space Table 6.5. Crystal data (A group is P21/a for the five examples, and all have Z ¼ 4 Guest/T

Refcode; reference

a

b/

c

Cell volume

2,3-dimethyl-1,3-butadiene; 143K Chlorocyclohexane; 85K

CN76

9.52(3)

12.55(4)

1683

JSH96

9.651

12.478

1753

1,5-cyclooctadiene; 173K

9.630(4)

12.615(5)

1770(1)

x(Squalene); 298K

WINZUO; GRB95 NL65

12.72

1837

0.5[2,6-diethyl-naphthalene)]; 298K

PIDRID; SSOK93

14.643(3)

15.42(3) 114.0(4) 15.964 114.22 16.05(5) 114.83(3) 15.72 115 9.282(2) 92.14(2)

12.571(3)

1707

10.14

References: GRB95 – Garneau, Raymond and Brisse, 1995; a and c have been interchanged to increase the similarity to the other structures; JSH96 – Jones, Shannon and Harris, 1996; Rietveld refinement at 85K; rhombohedral to monoclinic transition at 190–194K; NL65 – Nicolaides and Laves, 1965; composition must take into account the extension of the long squalene molecule through several thiourea unit cells; NL76 – Chatani and Nakatani, 1976; SSOK93 – Shindo, Shindo, Ohnuma and Kabuto, 1993 (reoriented).


236

b N

sin

(a)

N C

5.24

S

5.15

5.77 5.28

q 4.30 p r

(b)

Fig. 6.21. (a) The structure of {3(thiourea)[H2C ¼ C(CH3)––C(CH3) ¼ CH2]} shown in projection down [001]. The distortion of the original hexagonal structure is shown by the values of the angles p, q, r which are respectively 141, 115.5 and 103.5 , and by the dimensions inserted. (b) Relationship of the unit cell in projection down [001] (outlined) to the overall arrangement of the distorted tunnels. (Reproduced from Chatani and Nakatani, 1976.)

6.2.1.13

Behavior of some rhombohedral inclusion complexes on cooling

As examples we shall use the isostructural {3(thiourea)[ferrocene]} (abbreviated {3tu[(C5H5)2Fe]} (FERTUR01), {3(thiourea)[cyclohexane]} ({3tu[C6H12]}) and {3(urea)[trioxane]} ({3u[C3H6O3]}) complexes. The information is comparable with that given for (hexagonal){3(urea)[1/4 n-hexadecane]} although gaps remain in some of the crystallographic details. The crystal structure of {3tu[(C5H5)2Fe]} at room temperature is isomorphous with ˚ instead of the more those of other rhombohedral thiourea complexes, but with a ¼ 16.36 A ˚ . The Fe atom is at 0, 0, 1/4, a site with 32 symmetry; the five-membered rings usual 15.8 A are three-dimensionally disordered (Hough and Nicolson, 1978). The crystals are orange at room temperature, and become yellow on cooling, as does ferrocene itself. The diffraction pattern at 100K showed broadened peaks, suggesting the occurrence of phase transformations to lower symmetry phases. Nothing substantial is yet known about the crystallography of any processes occurring on cooling, but there are calorimetric


237

(Sorai, Ogasahara and Suga, 1981; Sorai and Shiomi, 1986), Mo¨ssbauer (Gibb, 1976; Lowery, Wittebort, Sorai and Hendrickson, 1990) and NMR (1H (Cle´ment, Gourdji and Guibe´, 1980), 2H (Lowery, Wittebort, Sorai and Hendrickson, 1990; Heyes, Clayden and Dobson, 1991) and 13C (Nakai, Terao, Imashiro and Saika, 1986)) studies, which we shall now describe; the calorimetric, Mo¨ssbauer and 2H NMR studies are remarkably detailed. The excess heat capacities (i.e. with respect to an interpolated smooth curve) arising from the phase transitions are shown in Fig. 6.22; the enthalpies and entropies of the five transitions were measured and are included in the diagram (Sorai, Ogasahara and Suga, 1981). The corresponding measurement for the (C5D5)2Fe complex has also been reported (Sorai and Shiomi, 1986), the two diagrams being very similar. The phase transition V , IV showed hysteresis and hence was classified as first order, while the order of the VI , V transition, despite indications of hysteresis, remained ambiguous; the other three transitions were second order. The transitions are associated with motions of the ferrocene molecules in the tunnels among a set of four orientations, which are shown schematically in Fig. 6.23. It should be noted that there is no indication, from any of the available physical measurements, which extend down to 87K, of freezing-in of the intramolecular rotations of the cyclopentadienyl rings. We shall now attempt to integrate the results of all these physical measurements, working upwards from the lowest temperatures. The information is about the reorientational motion of the ferrocenes, as very little is known about changes in the thiourea matrix. Comparison with other thiourea and urea systems suggests that there will be concomitant (and coupled) changes in the details of host and guest arrangements. A contrary view is that of Lowery et al. (1990) who,

60

1474 9.23

∆CP /JK–1 mol–1

50

263 1.79 14 0.08

40

30

Phase VI

V

IV

35 0.19 III

77 0.36 II

I

20 TC3 10

TC4

TC5

TC1 TC2

0 120

140

160

180 T/K

200

220

240

Fig. 6.22. The excess heat capacities arising from the phase transitions of {3tu[(C5H5)2Fe]}. The critical temperatures are (from TC1 to TC5) 147.2, 159.8, 171.4, 185.5 and 220K. The enthalpies (J/mol) and entropies (J/mol K) of the transitions are shown next to the peaks in the DCP/T curve. (Reproduced from Sorai, Ogasahara and Suga, 1981.)


238

basing themselves on the 2H NMR spectra of {3tu[(C5D5)2Fe]}, have concluded that ‘‘aside from rapid small amplitude reorientations at the higher temperatures, the thiourea molecules forming the hexagonal tunnels are stationary from 140 to 298K.’’ It is generally assumed that below 120K the parallel and perpendicular orientations (of the ferrocenes) are approximately equally populated, without there being any reorientation among them (Fig. 6.24); diffraction evidence is needed for more detailed definition of the low temperature crystal structure(s). The major reorientation takes place in the two associated phase changes in the region 120–163K. Sorai, Ogasahara and Suga (1981) consider in detail how the measured entropy changes shown on Fig. 6.22 should be allocated among the various reorientational processes; for example, 1.79 J/mol K was assigned to the VI)V transition.

(A) View down channel axis Perpendicular, or equatorial orientations

Parallel, or axial orientation

z

View normal to channel axis

(B)

z

(C)

Fig. 6.23. Schematic diagram showing the inclusion of the ferrocene molecules in the tunnels of the thiourea matrix. There is evidence from a single crystal 2H NMR study (Lowery, Wittebort, Sorai and Hendrickson, 1990) that the parallel orientation shown in the diagram is an approximation and that there are actually three orientations (related by a threefold axis) inclined at 17 to the tunnel axis. At high temperatures the ferrocenes are randomly oriented among the , , and z orientations. (Reproduced from Sorai, Ogasahara and Suga, 1981.)


239

This entropy change can be accounted for by a model (Fig. 6.23) in which reorientation between the three perpendicular sites (, , ) takes place within a group of three ferrocene molecules, which are perhaps located one above the other within a tunnel, giving a calculated entropy change of 1/3 R ln 3 (¼ 3.04 J/mol K); this is the ‘‘fast jump’’ process identified by Gibb (1976) from Mo¨ssbauer measurements. The first order change V , IV involves the disordering of such groups as well as interchange between perpendicular and parallel orientations (this is Gibb’s ‘‘slow jump’’ process, the rate being 108 Hz), with 65% of the molecules in the three perpendicular orientations and 35% in the parallel orientation. By 340K, the ferrocenes are equally distributed among the four orientations and tumbling rapidly. Descriptions at a molecular level have not yet been proposed for the other three phase transitions, which occur in a region of rapid change in the NMR and Mo¨ssbauer spectra and probably reflect the highly correlated nature of the interchanges between parallel and perpendicular orientations (Lowery, Wittebort, Sorai and Hendrickson, 1990). The results available about the thermal behavior of {3tu[C6H12]} are even more complete than those for {3tu[(C5H5)2Fe]}. When comparing these results (Fig. 6.24), one should note that cell dimensions and 14N pure quadrupole resonance frequencies will be most sensitive to changes in the thiourea framework, 1H NMR and Cp–T curves to both host and guest, and 2H NMR to the motions of the deuterated compound(s), which can be host and/or guest; of course, the host–guest interaction somewhat blurs these distinctions. The specific heat measurements (Cope, Gannon and Parsonage, 1972a) show one major thermal event at 130K and two much smaller events at 148K and 170K, and otherwise a smooth variation of Cp with T; only the anomalous region is shown in Fig. 6.24. The peak at 130K is surely to be identified with the first order phase change shown in the graph of lattice parameters against T (Fig. 6.24; Clement, Mazieres, Gourdji and Guibe´, 1977). The values of DHtr (1577 J/mol) and DStr (12.1 J/mol K) obtained from a DSC measurement (Poupko, Furman, Mu¨ller and Luz, 1991) are not very different from those given above for the first order transition in {3tu[ferrocene]}(1474 J/mol and 9.23 J/mol K). The minor peak at 148K appears to correspond with the splitting of the NQR line at 150K and to the (not very well established) change in slope of the lattice parameter curves at 145K. The temperature dependence of the þ 14N pure quadrupole resonance frequencies1 (Cle´ment, Mazieres and Guibe´, 1971; Cle´ment, Gourdji and Guibe´, 1975) is in agreement with these assignments; from room temperature down to 150K there is only one line which then splits into six components and this number remains unchanged below 130K, where sharp changes in frequencies indicate a first order phase change; there are no changes at 1 14 N (I ¼ 1) gives three NQR lines at ¼ (3 )e2Qq/4 and 0 ¼ þ ; however, the latter line occurs at very low frequencies and intensities and is generally not observed.


9.20

(a)

9.10 III

b(Å)

9.00 15.8 15.4 15.0 14.6

II

(b)

I

(e) 2540

XRD 128 K

150 K

1770 V(Å3)

(c) 1730

2530

NQR

V+ (kHz)

d(100)(Å)

240

1690 120

2520

(d) CP(J/mol K)

110

800 J/mol K

100

110

140

170 T (K)

90 80 70 110

CP 130 150 T (K)

170

Fig. 6.24. (a, b, c) Cell dimensions of {3tu[C6H12]} as a function of temperature; the c dimension ˚ ) does not vary with temperature. (Adapted from Cle´ment, Mazieres, Gourdji and Guibe´, (12.5 A 1977.) (d) Calorimetric measurements for {3tu[C6H12]}. (Reproduced from Cope, Gannon and Parsonage, 1972a.) (e) The temperature dependence of the 14N þ spectrum. Only five lines of the spectrum of Phase III are shown, the absent line being at slightly higher frequencies – 2550 kHz at 128K to 2570 kHz at 77K. (Reproduced from Cle´ment, Gourdji and Guibe´, 1975.)

170K (Fig. 6.24). There are three thiourea molecules in the rhombohedral unit cell at room temperature with the six nitrogens related by crystallographic symmetry elements, thus leading to only one NQR line. Below 150K the unit cell doubles in volume giving six thioureas or twelve nitrogens per cell; as there are six independent lines the nitrogens must be equivalent in pairs. The determination of the space group is incomplete; it was stated (Cle´ment, Mazieres, Gourdji and Guibe´, 1977) that the space group corresponding to the thiourea sublattice was P2/c while that of the crystal as a whole was given as P1, P1 or P2. Phase III appears to have the same symmetry as Phase II which does not lead to a change in the number of NQR lines; the cell volume in Phase III is 2.5% less than in Phase II. As an overall explanation for all these results it has been suggested (Cle´ment, Mazieres, Gourdji and Guibe´, 1977) that there are just three phases in the temperature region between 130 and 300K, thus withdrawing an earlier proposal (Cle´ment, Gourdji and Guibe´, 1975) of an additional phase change at 241K.


(a)

(b)

168 K 100 T1 sec 10

123 K

S2 (gauss2)

241

Thiourea 130 K

3 149 K

1 2

134 K 0.1

Cyclohexane 127 K

1 150 K .01 73

93

113

133

153

173

4

193

T (K)

6 8 103/T

10

Fig. 6.25. (a) The temperature dependence of the second moment S2 of the proton NMR absorption line in {3tu[C6H12]} (dots). The squares are for a sample containing 50% C6H12 and 50% C6D12. (Reproduced from Cle´ment, Mazieres, Gourdji and Guibe´, (1977).) (b) The temperature dependence of the relaxation time T1 for protons of the thiourea and cyclohexane molecules in {3tu[C6H12]}. The temperatures of the abrupt changes in T1 are indicated. (Reproduced from Cle´ment, Gourdji and Guibe´, 1975.)

Phase III

133K

! Phase II

148K

! Phase I

NMR studies provide additional information about the behaviour of the cyclohexane guests. 1H NMR line widths show a change at 130K for the thiourea protons and at 230–300K for the cyclohexane protons; although line widths are good indicators of the onset of molecular motions, experience suggests that they are not always sensitive to the occurrence of phase changes. The experimental value of the second moment for {3tu[C6H12]} at 77K is 3.9 G2 (Fig. 6.25(a)) whereas that calculated for a rigid lattice is 18.4 1 G2. Reorientation of C6H12 about its triad axis would give S2intra ¼ 3.6 G2, even if the triad axis were tilted by up to 30 to the tunnel axis. Thus the cyclohexane molecules are reorienting even at the lowest temperature reached in the NMR measurements. The relaxation times T1 (Fig. 6.25(b)) for the thiourea sublattice show abrupt changes at 130 and 168K, corresponding to peaks in the specific heat curves; the T1 changes in the cyclohexane sublattice at 127 (sharp) and around 149K also match features of the Cp–T curves. Thus the different physical techniques give compatible and complementary indications of changes in structure, although descriptions at the molecular level are still lacking. The further reductions in S2 above 133K and again above 150K are due to additional motions, which have been elucidated in a detailed study by deuterium NMR spectroscopy covering the temperature range 116–333K (Poupko, Furman, Mu¨ller and Luz, 1991); 2 H NMR curves for the lower temperature range are shown in Fig. 6.26. The 2H NMR spectrum at 127K shows a superposition of two equally intense, axially symmetric powder patterns, which are interpreted in terms of cyclohexane molecules


242

Species B

Species C

141 K

213 K

135 K A+B

163 K C

133 K A+B

151 K B+C

131 K A+B

147 K B+C 145 K

127 K A 100

B+C 0

–100 50 Frequency (kHz)

0

–50

Fig. 6.26. 2H NMR spectra from 3tu[C6D12] samples at various temperatures, recorded by the quadrupole echo sequence method. The spectrum at 127K corresponds to Phase III, those between 131 and 151K are from Phase II, and those from 163 and 213K are from Phase I. The species A, B and C are identified in the text. (Reproduced from Poupko, Furman, Mu¨ller and Luz, 1991.)

rigidly fixed with their triad axes along the tunnel axes but rapidly reorienting about these axes i.e. there is axial rotation of the cyclohexanes below the first order phase transition. This motion leads to relaxation of the equatorial deuterons by a threefold jump process, while the axial deuterons do not change; this is referred to as Species A in Figs. 6.27 and 6.28. The parameters of the Arrhenius rate equation for a jump to one of the adjacent sites are A ¼ 1.83 1013 s1 and Ea ¼ 10.5 kJ/mol. Both the diffraction patterns and the 1H T1 relaxation time vs. T 1 curve show that there is a concomitant change in the thiourea matrix at the phase transition but the details have not been worked out. The motion of the cyclohexanes in the tunnels changes in the temperature region of stability of Phase II. From below the III ) II phase transition to immediately above it, only Species A is present, but its uniaxial reorientational motion is changed by gradual addition of a fast wobbling component in the biaxial potential of the channnel; this is Species B, which is the only species present at 143K. Above this temperature rapid chair–chair interconversion takes place (Species C) and this continues up to the highest temperature reached. The varying proportions of Species A, B and C are shown in Fig. 6.28. The order parameter can be calculated from the average quadrupolar splitting (Fig. 6.29). The NMR studies show that ordering of the cyclohexane molecules takes place on cooling over the range 300–220K, without appreciable crystallographic changes. The anomalous region in the specific heat starts at about 170K, where there is a change in T1 for the thiourea protons, but the nature of the corresponding physical change is not known. There is a crystallographic disorder ) order transition at 150K, due to changes in


243

1 0.9 Relative intensity

0.8 0.7

B A

C

0.6 0.5 0.4 0.3 0.2 0.1 0 125 130 135 140 145 150 155 160 T (K)

Fig. 6.27. Relative abundances of species A, B and C in phase II, as a function of temperature. The curves are guides to the eye. (Reproduced from Poupko, Furman, Mu¨ller and Luz, 1991.)

orientations or motions of the cyclohexane molecules. At about 130K there is a first order crystallographic transformation, with appreciable changes in the details of the arrangement and librations of the thiourea molecules. The cyclohexane molecules continue to reorient down to at least 77K. The curve of heat capacity vs. T has been calculated (Cope, Gannon and Parsonage, 1972b) on the basis of an order , disorder model in which the cyclohexane molecules can occupy any of six positions, interacting both with the host framework and neighbouring guest molecules in the same and adjacent tunnels. The predicted curve was much broader than the experimental curve and peaked at a lower temperature. The disagreement is not surprising as this transition is one in which the principal changes are due to the thiourea framework. The behaviour of {3(urea)[trioxane]} on cooling (Cle´ment, Mazieres and Guibe´, 1972; Claude, Cle´ment and Dworkin, 1977) is similar, but not identical, to that of the isostructural {3(thiourea)[C6H12]} (Fig. 6.28). The specific heat curves show three apparently first order peaks at 190, 203 and 244K, in good agreement with DTA and DSC studies. From the latter (Gelerinter, Luz, Poupko and Zimmerman, 1990) the following enthalpies and entropies of transition for {3(urea)[C3D6O3]} were obtained for the three transitions IV ! III (276 J/mol, 1.46 J/mol K), III ! II (276 J/mol, 1.39 J/mol K) and II ! I (1187 J/mol, 4.95 J/mol K). The 14N NQR spectrum has been measured over the temperature region 130 to 300K (Cle´ment, Mazieres and Guibe´, 1972) and shows only a smooth variation of frequency with T for two þ and one lines. It was suggested in explanation that the trioxane guest molecules are distributed over sites of 3 and 32 symmetry in different parts of the crystal and this subjects otherwise equivalent nitrogen atoms to different electric field gradients. This behaviour of the NQR spectrum is quite different from the results of the analogous measurements for the isostructural {3(thiourea)[C6H12]} complex (Fig. 6.23(a)) where only one þ line was found for the rhombohedral structure, which then split further on cooling through both second order and first order transitions. The NMR line width of the


244

140 120

A

(kHz)

100

Phase III

Phase II

Phase I

80 60 B

40

C

20 0 90

110

130 150 T (K)

170

190

Fig. 6.28. A plot of the average quadrupole splitting < Q> ¼ < Qax>, (left hand ordinate) of the axial deuterons of C6D12 in the various solid phases of 3tu[C6D12]. The value of < Q> at 333K is 11 kHz. The order parameter of the molecular C3 axis can be calculated from the relation S ¼ < Q>/(125 kHz). The asymmetry parameters for the spectra in the various phases are III ¼ 0; II ¼ 0.17–0.25; I ¼ 0. (Adapted from Poupko, Furman, Mu¨ller and Luz, 1991.)

800 2785

CP(J/mol K)

700 600 500

IV

III

II

I

400 300 200 160

180

200 220 T (K)

240

260

Fig. 6.29. Cp vs. T for {3(urea)[C3H6O3]}. The experimental points lie on smooth curves in the rest of the 0–300K range. The lines are guides to the eye. (Adapted from Claude, Cle´ment and Dworkin, 1977.)

trioxane protons shows changes at 120, 190 and 240K, the latter two changes corresponding to the first order transitions at 189 and 243K. However, the specific heat curve does not show anomalously high values in the region about 120K. The diffraction patterns were reported to show a rhombohedral ) monoclinic transformation at 244K but no changes at 201 and 189K. However, there is a further change at (unspecified) lower temperatures in which a contracts, b expands and c remains unchanged (cf. Fig. 6.23). There are thus four phases in {3(urea)[C3H6O3]}, and much is now known about the behaviour of the guest molecules as a result of 2H NMR studies on {3(urea)[C3D6O3]} and


245

{3(deutero-urea)[C3H6O3]} (Gelerinter, Luz, Poupko and Zimmerman, 1990). We start with phase IV, where the spectra show that, below 130K, the trioxane molecules are essentially static on the NMR time scale. As the temperature increases towards the IV ) III transition, the trioxane molecules start reorienting by a three-site jump mechanism for which the Arrhenius parameters for one-directional jumps are Ea ¼ 20.1 kJ/mol and A ¼ 1.1 1013 s 1; fast wobbling of the trioxanes about their C3 axes sets in above 170K Although there are distinct changes in NMR line shape going through the IV ) III transition, the line shape remains the same in phases III and II, indicating similar behaviour of the trioxane molecules. Actually two inequivalent species, A and B, of trioxane were found, with a relative abundance of about 2 : 1, A being ordered more or less as in phase IV while B is much less ordered (can these be the trioxane guest molecules distributed over sites of 3 and 32 symmetry?). In phase I the trioxane molecule is similar to species B referred to above; fast ring inversion sets in on further heating, with kinetic parameters similar to those found for trioxane in liquid-crystalline and solution environments, while in crystalline neat trioxane there is no ring inversion below the melting point (338K). 6.2.1.14 The orthorhombic Type 4 urea tunnel inclusion complexes The complexes with 1,4-dichlorobutane, 1,5-dichloropentane and 1,6-dichlorohexane are Type 4, with orthorhombic symmetry and space group Pbcn (Otto, 1972). The ideal unit cell dimensions for {3(urea)[1,4-dichlorobutane]} calculated from the dimensions of the hexagonal urea tunnel are as follows, with measured values in square brackets: p ˚ , b ¼ 3 4.75 ¼ 14.25 [14.15] A ˚ , c ¼ 11.00 [10.96] A ˚. a ¼ 3 4.75 ¼ 8.23 [8.34] A There are additional weak reflections in the 1,4-dichlorobutane diffraction patterns which indicate an ordering of the guest molecules but this was not explored in detail. The crystals with 1,6-dibromohexane as guest are monoclinic with a slightly distorted variant of the Type 4 structure. The complex {3(thiourea)[ 0.5(2,6-diethylnaphthalene]} crystallizes in a monoclinic ˚ , ¼ 92.14(2) , Z ¼ 4, space unit cell with a ¼ 14.643(3), b ¼ 9.282(2), c ¼ 12.571(3) A group P21/a (Shindo, Shindo, Ohnuma and Kabuto, 1993). Professor J. S. Rutherford (Bulawayo) has suggested that this is a distorted variant of the Type 4 structure. 2,6Diethylnaphthalene is important as a feedstock of speciality high-performance polymers such as poly(ethylene 2,6-naphthalenedicarboxylate). The isomers of diethylnaphthalene have similar physical properties and so are difficult to separate by distillation or crystallization but the 2,6-isomer selectively forms a thiourea complex – hence the interest in the structure. A noncentrosymmetric variant of this structure type is taken up by {3(thiourea) ˚ , Z ¼ 4, [(4-1,3-cyclohexadiene)Fe(CO)3]} (a ¼ 12.562(1), b ¼ 16.128(1), c ¼ 9.536(1) A space group Pna21) (Tam, Eaton, Calabrese, Williams, Wang and Anderson, 1989; GARTOI10) (Fig. 6.30). This complex has a second harmonic generation (SHG) efficiency of about 40% of that of urea. 6.2.1.15 The hypothetical Type 5 orthorhombic tunnel inclusion complexes No example of this type has yet been encountered; it would be expected to have space group Pbca (Fig. 6.31).


246

1/4

R

L a down b

R

L R

c

L

Fig. 6.30. Projection of {3(thiourea)[(4-1,3-cyclohexadiene)Fe(CO)3]} structure down [100]. The symmetry elements of space group Pna21 are shown. The handedness of the thiourea spirals is marked by L and R. (Adapted from Tam, Eaton, Calabrese, Williams, Wang and Anderson, 1989.)

1/4

R

L L

R

L

R

L

R

R

L R

1/4

L 1/4

1/4

Fig. 6.31. Hypothetical Type 5 tunnel inclusion complex, space group Pbca. The handedness of the spirals is marked by L and R. The cell dimensions predicted for a urea complex are: p ˚ , b ¼ 3 4.75 ¼ 14.25 A ˚ , c ¼ 11.00 A ˚ . (Reproduced from Otto, 1972.) a ¼ 2 3 4.75 ¼ 16.46 A


247

6.2.1.16

The crystal structure of selenourea and its relation to the structures of the tunnel inclusion complexes There is no relation between the crystal structures of tetragonal urea and orthorhombic thiourea, on the one hand, and those of their hexagonal or rhombohedral inclusion complexes on the other. Not so for selenourea which crystallizes in the trigonal system ˚ (both at 173K), space group P31 (or P32), Z ¼ 27). The (a ¼ 15.201(5), c ¼ 12.950(5) A crystal structure (Rutherford and Calvo, 1969; SEUREA) shows that there are nine selenourea spirals in the unit cell. If we consider, for the moment, only the spirals at the corners of the hexagons in Fig. 6.3, then the structure is of Type 1 (or 2, depending on absolute configuration of the crystal studied). However, there are also spirals (of the same handedness as those at the corners of the hexagons), which are located on crystallographic threefold screw axes along the central axes of the hexagons. The two sets of spirals are essentially structurally equivalent although this is not required by the symmetry of the space group. Thus selenourea could be described as a self-inclusion complex. The two interlocking, but nonbonded, sets of spirals constitute a one-dimensional analog of the nonbonded but interlocking three-dimensional networks found in the quinol clathrates. 6.2.1.17 Thermodynamics of the formation of the tunnel inclusion complexes This subject, much studied soon after the discovery of these complexes (Schlenk, 1949; Redlich, Gable, Dunlop and Millar, 1950; Redlich, Gable, Beason and Millar, 1950; Zimmerscheid, Dinerstein, Weitkamp and Marschner, 1950), has attracted only little recent attention (Parsonage and Staveley, 1984). As Schlenk (1949) noted, an inclusion complex is, from a thermodynamic point of view, a true chemical compound (‘‘so wie etwa die Substanz CuSO45H2O’’) and not an absorbate. Put differently, this means that these complexes form separate phases in the binary host–guest phase diagram with crystal structure different from those of host and guest, and are not solid solutions of the guest in the neat crystals of the host. We consider the thermodynamics of the following reaction, which has, for convenience, been written in terms of urea (c ¼ crystal) although it is easily generalized to other types of inclusion complex. UreaðcÞ þ x(Guest (liq)) ) Inclusion complex(c)

ð6:1Þ

Although most of the published results are expressed in terms of one mole of guest, it is often convenient to use one mole of host as reference, especially when it is necessary to emphasize the crystallographic similarity of the complexes. A number of methods have been used to measure the various thermodynamic quantities; vapor pressure (p) of guest in equilibrium with the complex has been measured by a dew point method and compared with that of the pure guest (p0), both over a range of temperatures. Then DHf, the enthalpy of formation of the inclusion complex, is obtained from the van’t Hoff equation, d(ln K)/ d(1/T) ¼ DHf/R, where K ¼ p/p0; we use the approximation that DHf is independent of temperature. Values of K can also be obtained from measurements on the solution equilibrium urea(solid or solution) þ guest(solution) , inclusion complex

ð6:2Þ


248

where an aqueous or non-aqueous solvent can be used. Values of log K for urea complexes of n-alkanes and for the thiourea complexes of cyclohexane and CCl4, are plotted against 1000/T in Fig. 6.32. We limit our discussion to these complexes, but measurements for many other types of guest have been reported (see Fetterly (1964) for summary). Values of DHf can also be obtained by direct calorimetry, a method used by Schlenk (1949) and Zimmerschied et al. (1950). Values of DSf can be obtained by combining DGf (¼RT ln K) and DHf values, and also (sometimes, when there are not too many intervening phase transitions) from measurements of the specific heats of complex and components. Using the values from Fig. 6.32 at 1000/T ¼ 3.36 (i.e. T ¼ 298K), we find that DGf (kJ per mol guest, 298K) ¼ 10.46 1.66n, where n is the number of carbons in the n-alkane; thus, in agreement with experiment, only complexes with n-heptane and above will be stable. We summarize DHf and DSf values from various sources in Table 6.6. The DTA measurements were carried out at temperatures of 50–120 C; nevertheless, from the semiquantitative (numerical) agreement of his DHdec values with DHf values from calorimetry and vapour pressures at lower temperatures, McAdie (1962) concluded that the complexes decomposed to tetragonal urea and liquid guest on heating. The formation of the inclusion complexes from ureaþliquid guest (equation 6.1 above) is an exothermic process and the complexes are enthalpy stabilized at 298K. If we consider the hexadecane complex, then we can estimate its enthalpy of formation from crystalline hexadecane by adding the enthalpy of fusion of hexadecane (DHfus ¼ 47.3 kJ/mol; m.pt. ¼ 291K); the enthalpy of formation of the complex is 44.4 kJ/mol guest, and the complex is still enthalpy stabilized. The entropies of formation of the complexes (from liquid guests) are all negative, and this implies, in a rough way, that the paraffins are more

3

log K (formation)

2.5

C(16)

2 C(10)

1.5

C(9)

C(12)

1 0.5

C(8)R

TU C(7)

0 –0.5 2.8

2.9

3

3.1

3.2 3.3 1000 /T

C(8)S

3.4

3.5

3.6

3.7

Fig. 6.32. Log K versus 103/T for formation of urea and thiourea complexes (equation 6.1); the equilibrium constants were obtained from vapor pressure measurements. The urea complexes are all of n-paraffins, denoted as Cn etc; C8(S) values are from Schlenk (1949) and all the other values from Redlich, Gable, Dunlop and Millar (1950) including C8(R). The thiourea complexes (TU) are of CCl4 (to the right) and cyclohexane (to the left) (Redlich, Gable, Beason and Millar, 1950), the experimental points lying along the same line. The temperature range is from 273 to 353K.


249

Table 6.6. Values of Hf and Sf from various sources for {urea[n-paraffin]} inclusion complexes, and two thiourea complexes. The preferred values for Hf are averages of ‘‘vapour pressure’’ and ‘‘calorimetry’’ values, and for Sf ‘‘specific heat’’ values, taken from Pemberton and Parsonage (1966), who estimate their error to be 4 J/mol K. The ‘‘vapour pressure’’ values, which seem to be systematically high, are bracketed in the Table Guest

Host/ Guest Ratio

C7H16 C8H18 C9H20 C10H22 C11H24 C12H26 C15H32 C16H34 tu–CCl4 tu-cyclohexane

6.1 6.73 7.4 8.1 10.0 9.3 11.75 12.0 3.0 3.0

Hf (kJ/mol)

Vapour pressure

Calorimetry

25.4 44.0 49.4 52.7

31.8

Sf (J/mol K)

DTA (Mc-Adie, 1962)

41.4 61.1

67.4 87.9 13.7 10.7

56.1 95.4

95.8

Hf/ mole guest

Hf/ mole host

Vapour pressure/ mole guest

28.2 44.0 49.4 52.7 61.1 67.4

4.55 6.53 6.67 6.51 6.11 7.25

(90.1) (138) (151) (158)

91.7 13.7 10.7

7.64 4.57 3.56

(200) (234*) (244) (40.6) (30.5)

Specific heats/ mole guest

148 151 185 207

Sf / mole host

(14.8) (20.5) (20.4) 18.3 16.2 15.7 17.3 (13.5) (10.2)

* interpolated value.

ordered in the tunnels than in the pure liquids. Referring again to the hexadecane complex, we obtain the entropy of formation of the complex from crystalline hexadecane (ignoring the effects of small differences in temperature) by adding the entropy of fusion of hexadecane, which is 159 J/mol K, giving a value of 48 J/mol (guest) K. This implies that hexadecane is more ordered in the tunnels of the complex than in its neat crystals, which seems surprising. The two thiourea complexes included here are both enthalpy stabilized, but to a lesser degree than the urea complexes; however, the differences are not that striking when considered on a ‘moles host’ rather than on a ‘moles guest’ basis. The formation of an inclusion complex, at a particular temperature and pressure, from its components can be envisaged to take place through the following steps (the calculation is formulated for a composition {Host[x(Guest)]}, and thus the thermodynamic parameters of the inclusion complex are given as J/mol host (or J/K mol host): (1)

The crystalline host is converted from its stable (neat compound) structure to the empty matrix (this implies that the complex is a different phase from the neat host and that formation of a primary solid solution is not considered here) Stable hostðcÞ ) Empty matrix(c) DXtrans ¼ Xmatrix Xstable , where X is free energy, enthalpy or entropy. This process is endothermic.


250

(2)

The guest is converted from its stable structure to the vapour phase DXvap

Guest (liq or cryst) ) Guest(vapor) or DXsubl , according as guest is liquid or crystalline:

This process is endothermic. (3)

The vapour of the guest enters the empty matrix to form the inclusion complex Empty matrix(c) þ x[Guest(vapor)] ) Inclusion complex(c) DXincl ðDXdec ¼ DXincl for the inverse process of decomposition): This process is exothermic, unless the complex is entropy-stabilized.

The overall process for formation of the crystalline inclusion complex is Stable host(c) þ x[Guest (liq or cryst)] ) Inclusion complex(c) DXform ¼ DXtrans þ x½DXvap ðor DXsubl Þ þ DX incl : This conceptualization is quite general and has been applied to a number of inclusion complexes, as will be noted at appropriate points. Of the quantities on the right hand side of the equation, DXincl and DXtrans are not known. Now, following Schlenk (1949), we show how DHincl and DHtrans can be estimated for urea inclusion complexes formed from tetragonal urea and liquid n-paraffin guests. This is ˚ ) be the length of urea tunnel most easily formulated in terms of moles of guest. Let lP(A ˚ between interacting with the included guest, and Qincl be the enthalpy of interaction per A ˚ ) be the corresponding overall length of urea tunnel tunnel and guest. Let (lPþ2.4) (A ˚ between paraffin chains, taken lengthwise) (i.e. there is a noninteracting distance of 2.4 A and Qtrans be the enthalpy of transformation from tetragonal to (hypothetical) hexagonal ˚ . Then urea per A DHf (per mole of guest) ¼ ðlP Qincl Þ þ ððlP þ 2:4Þ Qtrans Þ þ DHvap : ˚ , and DHvap can be obtained from standard tables; Qtrans and For n-octane, lP ¼ 10.4 A DHvap are endothermic, and DHf (values from Table 6.6) and Qincl exothermic. Similar equations were set up for n-heptane, n-decane and n-hexadecane; values of Qtrans ˚ ) and Qincl (¼14.4 kJ/A ˚ ) were obtained by linear regression. As one mole(¼5.45 kJ/A ˚ , DHtrans ¼ 10.2 kJ/mol urea; DHincl depends cule of urea occupies a tunnel length of 1.87 A on the nature of the guest but the interaction enthalpy per CH2 group of gaseous guest is 19 kJ/mol. These values are somewhat different from those given by Schlenk (DHtrans ¼ 3.7 kJ/mol urea, DHincl 15 kJ/mol per CH2 group) because of different DHf and DHvap values used in the two calculations. These DHtrans values are considerably larger than those for, say, - and -quinol (see Section 7.3.1). The p–T and T–composition projections of the phase diagram of urea–n-heptane calculated (Farina, Di Silvestro and Colombo, 1986) from Schlenk’s vapour pressure measurements are shown in Fig. 6.32. These diagrams will be discussed later together with those of perhydrotriphenylene inclusion complexes.

DIRECTIONALLY BONDED HOSTS 160

150

251

LH + LU

140

p (heptane) mm

120

t °C

Liquid

100

100

LH + U

Existence region of inclusion complex

80 60

50

40

Solid LH + C

20 0 0

10

20

30

40

t (°C)

50

0 H

0.5 mol fraction urea

C+U 1.0 U

Fig. 6.33. Urea–n-heptane phase diagrams; left, p–T projection and right, T–X projection. LH is liquid heptane (f.pt. –91 C; b.pt. 98 C). Note that n-heptane and urea are not miscible in the liquid phase. (Adapted from Farina, Di Silvestro and Colombo, 1986.)

6.2.2 The Bishop–Dance Hosts – exo-2,exo-6-dihydroxy2,6-dimethylbicyclo[3.3.1]nonane and analogs 6.2.2.1

Introduction

In 1979 Bishop and Dance reported that 2,6-dimethylbicyclo[3.3.1]nonane-exo-2,exo-6diol (6.1; we use the same numbering of hosts as used by Bishop et al. in their series of papers, but with the prefix 6 indicating our chapter number, details being given in the next section) formed a series of trigonal tunnel inclusion complexes with a chiral helical tubuland structure. Since then much effort has been invested in generalizing the chemical basis of this discovery and it has been found that ten related molecules form similar crystal structures, some of which are inclusion complexes, and some not. The general composition of the group of tunnel inclusion complexes is {3(host)[(guest)x]} (x 1) (Bishop, Dance, Hawkins and Lipari, 1984; Bishop, Dance, Hawkins and Scudder, 1987; Bishop and Dance, 1988, 1991). Two of the hosts (6.2 and 6.4) also form a different series of racemic tetragonal ellipsoidal clathrate complexes, the general composition being {4(host)[guest]}. Usually a particular guest forms only one of these structural types, but there are examples (1,2-dichlorobenzene, CHCl3, C6H5Br) where both types (trigonal and tetragonal) have been obtained, the type depending on crystallization conditions. For these examples, there are (at least) two complexes, of 4 : 1 and 3 : 1 compositions, in the 6.2 (or 6.4)–guest phase diagram. Thus two of the Bishop–Dance hosts, and in potential others, form two distinct structural families of inclusion complexes, and we shall treat each of these families separately. We shall refer to the overall group of hosts as Bishop–Dance hosts. A particularly comprehensive account has been given by Bishop (1996). 6.2.2.2

The helical tubuland structures

The molecules 6.1–6.6 (Fig. 6.34) form a group of related hosts, all of which crystallize in closely similar structures, and some of which form tunnel inclusion complexes. Two


252

C4 O

C3 C2

C5

C6 C1

C7 O

C8

O C3

C3 O O C3*

C3* O

Type A

Type B

Fig. 6.34. The Bishop–Dance hosts – a perspective view of 6.1 down its twofold axis is shown in the top part of the figure. In the lower part schematic views (also from above the central methylene group) are given of Type A (6.1, 6.3, 6.5, 6.6) and Type B (6.2, 6.4) molecules; some hydrogens have been removed for clarity. The Type A hosts have C3–C3* not bridged in 6.1, and bridged by –CH2–CH2– (ethano) in 6.3 and by –CH2–CH2–CH2– (propano) in 6.5. In 6.6 there is no bridging between C3–C3* but the central –CH2– (methano) is replaced by –CH2–CH2– (ethano). A diagram of 6.7 is given later and other representations of some of the molecules are shown in Fig. 6.40. The Type B hosts have C3–C3* bridged by –CH2–CH2– in 6.2 and by –CH2–CH2–CH2– in 6.4.

strategies were employed in the attempt to develop a family of chemically related hosts. The first was to replace the methyls in 6.1 by hydrogens or ethyls (Bishop, Choudhury and Dance, 1982); these compounds were found not to form inclusion complexes. The second strategy was to connect between C3 and C7 (C3*) by ethano and propano bridges, as shown in the formulae (Fig. 6.34); it was found that there was retention of stereochemistry at C2 and C6 in 6.3 and 6.5 and inversion in 6.2 and 6.4 (numbering as in 6.1). From a chemical point of view hosts 6.1, 6.3, 6.5 and 6.6 form one group (called Type A) and 6.2 and 6.4 another (called Type B). Remarkably, all six compounds gave isostructural chiral trigonal crystals (i.e. spontaneous resolution had taken place on crystallization) based on the formation of tight spirals of hydroxyl . . . hydroxyl hydrogen bonds. These were called helical tubulands and, as noted earlier (Section 6.1), we have appropriated this useful name for application to analogous structures with a wider variety of hosts. However, many chemically-related compounds crystallize in quite different arrangements (e.g. layer structures; Hawkins, Scudder, Craig, Rae, Raof, Bishop and Dance, 1990) and some progress has been made towards understanding the factors involved in the formation, or


253

not, of helical tubuland inclusion complexes (Dance, Bishop and Scudder, 1986; Bishop, Craig, Dance, Kim, Mallick, Pich and Scudder, 1993; Bishop, Craig, Dance, Scudder and Ung, 1993). We have alluded to the overall problem in the Introduction to Part III, and avoided discussing the specific problem concerning urea adducts (see Section 6.1.2.10). The essential features required for a molecule to form helical tubuland structures of the P3121 type have been summarized as follows (Bishop, Dance, Hawkins and Scudder, 1987; Bishop, Craig, Dance, Scudder, Marchand and Wang, 1993), to which we have added two extra requirements (#5 and #6): 1. The diol molecules must have two fold symmetry, as shown in Fig. 6.34; this symmetry can be exact or obtained by averaging of equivalent conformations. 2. Some variability is permitted in the alicyclic bridging skeletons, which are also required to have some flexibility. 3. Substituent groups around the periphery appear to be deleterious. 4. The tertiary alcohol groups must have a methyl substituent, which seems to have the correct properties to support the tunnel wall structure. ˚ and the O–C . . . 5. The intramolecular O . . . O distance must be approximately 5.7 A C–O torsion angle must be 75 for Type A molecules or 95 for Type B structures (see Table 6.7). 6.7, 6.8 and 6.9 do not meet these requirements and we make some comments about this later. 6. In order that Type A and B hosts can form the same structure types, they must have the same disposition of hydroxyls and methyls. This happens when a Type A molecule is viewed down the two fold axis through the central methylene group from above, or up from below (as shown in the equivalent views of ball and stick models in Fig. 6.35). TYPE A

TYPE B

6.1

6.2

6.3

6.4

Fig. 6.35. Ball-and stick models of some examples of Type A and B molecules viewed down their two fold axes. The molecules have been oriented so that the hydroxyl and methyl groups are similarly disposed in both types of molecule. This requires that the alicyclic skeletons be differently disposed in the two types, as explained in the text. Some hydrogens have been omitted for clarity. The oxygens are lightly shaded and the carbons dark shaded. The H–O–C–CH3 torsion angles are in the range 51–67 .


254

Table 6.7. Nomenclature of the Bishop-Dance hosts, and some geometrical features Number

Chemical composition

Type A host molecules 6.1 C11H20O2

6.3

C13H22O2

6.5

C14H24O2

6.6

C12H22O2

6.7

C13H18O2

6.8

C14H20O2

6.9

C14H20O2

6.10

C12H20O2S

Type B host molecules 6.2 C13H22O2

6.4

Neat 6.2 in SODVUC

C14H24O2

Systematic name

˚ d(O . . . .O 0 ) A

2,6-dimethylbicyclo [3.3.1]nonane-exo-2, exo-6-diol 2,7-dimethyltricyclo [4.3.1.13,8]-undecane-anti-, 2,anti-7-diol 2,8-dimethyltricyclo [5.3.1.13,9]-dodecane-anti-2, anti-8-diol 2,6-dimethylbicyclo [3.3.1]decane-exo-2, exo-6-diol 4,7-dimethylpentacyclo [6.3.0.02,6.03,10.05,9] undecane-anti-4,anti-7-diol 4,7-dimethylpentacyclo[6.4.0.02,6.03,10.05,9] dodecane-anti-4,anti-7-diol 4,7,11-trimethylpentacyclo [6.3.0.02,6.03,10.05,9] undecane-anti-4,anti-7-diol 2,7-dimethyl-9-thiatricyclo[4.3.1.13,8]-undecane-anti-, 2,anti-7-diol

5.59

73.4

5.53

79.2

5.51

71.3

5.98

163.6

6.01

174.1

5.98

163.6

5.43

88.3

5.55

94.3

5.68

97.4

5.38 5.29 5.36

81.7 84.5 90.1

2,7-dimethyltricyclo [4.3.1.13,8]-undec-ane-syn-2, syn-7-diol 2,8-dimethyltricyclo [5.3.1.13,9]-dodecane-syn-2, syn-8-diol

(O–C . . . .C 0 O 0 ) (deg.)

The descriptions syn and anti refer to the OH functions relative to the larger of the bridges across the molecular twofold axis.

Crystal data for the trigonal tunnel inclusion complexes of the Type A hosts 6.1 and 6.7, and of neat 6.3 and 6.5 are given in Table 6.8, and also of the trigonal tunnel inclusion complexes of the Type B hosts 6.2 and 6.4. The hosts 6.1 and 6.7 form isostructural crystalline tunnel inclusion complexes while the crystals of 6.3 and 6.5 (which belong to the same isostructural family) do not contain guest molecules; the hosts 6.2 and 6.4 form a second, closely related, isostructural group of complexes. The a axes of crystals of the


255

˚ and those of Type B hosts in the range Type A hosts lie in the range 11.9–12.5 A ˚ 13.2–13.8 A. The cell dimensions are unusual for the constancy of c (but with some exceptions) and the variation of a with the nature of both hosts and guests; similar behavior is found with some cyclophosphazene tunnel inclusion complexes (see Section 6.3.2). 6.1 has been shown to form isostructural inclusion complexes of formula {3(6.1)[(guest)x]} (0.75 x 1.2) with iodineþethanol* (x ¼ 0.5), acetonitrile*, 1,2dimethoxyethane*, 1,2-dichloroethane*, ethyl acetate*, chloroacetic acid*, propanoic acid*, trichloroethene, thiophene*, chlorobenzene*, toluene*, dioxane*, acetic acid, formamide, N-methylformamide, ethylbenzene, p-xylene, m-xylene, cyclohexene, diethylamine, mesityl oxide, acetoneþwater, dimethyl sulphoxide, bromobenzene, m-dichlorobenzene, 1,3-dibromopropane, chloroform, ethyl 2-bromopropanoate and 2,5dibromothiophene (Ung, Gizachew, Bishop, Scudder, Dance and Craig, 1995). Complexes of 6.1 with o-dichlorobenzene, o-xylene, cyclohexanone and trans-1,2-dibromocyclohexane were not obtained. Results of extended X-ray refinements which take the contributions of the (asterisked) guest molecules into account were given in Ung, Gizachew et al., 1995. 6.2 has been reported (Ung, Bishop, Craig, Dance and Scudder, 1992a) to form trigonal inclusion complexes with the following guests – ethyl acetate, CCl4, CBr4, 1,1,1trichloroethane, 1,2-dibromoethane, 1,3-dibromopropane, 1,4-dibromo-butane, o-xylene, m-xylene, p-xylene, o-dichlorobenzene, m-dichlorobenzene, ethylbenzene, n-butylbenzene, chlorocyclohexane, 1,2-dibromocyclohexane, diethyl ether, di-n-butyl ether, 1,2-dimethoxyethane, n-hexane. We have arranged the crystal data of Table 6.8 by host and then in order of ascending cell volume, which also provides a natural division between the Type A and B hosts; it is remarkable that the same overall structure type is retained over a (unit cell) volume range ˚ 3. How should these crystals be classified in terms of host-guest phase from 880 to 1151 A diagrams? In the Type A group, neat 6.1 (recrystallized from the non-complex forming solvent mesitylene) is reported (Ung, Gizachew et al., 1995) to have the same crystal structure as the trigonal inclusion complexes with various guests; thus the inclusion complexes of 6.1 are primary solid solutions of guest in host and so differ from most other tunnel inclusion complexes such as those of urea and thiourea (see Section 6.2.1.15). The question is not relevant for 6.3, 6.5, 6.8 and 6.9 (all Type A) as their tunnels are too small for them to form inclusion complexes (Fig. 6.39) (Hawkins, Bishop, Craig, Dance, Rae and Scudder, 1993). Crystal data do not appear to have been reported for neat 6.6 and 6.7. The behaviour of 6.2 as host is particularly interesting. Neat 6.2 has a hydrogen-bonded layer structure with three independent diol molecules in the asymmetric unit ˚ , ¼ 109.42(1) , Z ¼ 12, space group P21/c; (a ¼ 7.398(2), b ¼ 25.166(3), c ¼ 20.076(4) A Ung, Bishop, Craig, Dance and Scudder, 1991; SODVUC); some geometrical features (Table 6.7) differ for 6.2 molecules in the neat crystals and in the complexes. Although the guest molecules in the trigonal tunnel complexes are generally highly disordered at room temperature, those of 6.2 with Br(CH2)3Br, CCl4 and o-xylene are well ordered (Ung, Bishop, Craig, Dance and Scudder, 1992a). 6.2 shows a further complication – recrystallized from benzene it gives tetragonal crystals of composition {4(C13H22O2)[C6H6]}, which have a clathrate structure; o-dichlorobenzene gives trigonal (host:guest ratio 3 : x where x 1) or tetragonal (host : guest ratio 4 : x where x 1) crystals, depending on crystallization conditions. Thus the 6.2–o-dichlorobenzene phase diagram includes the neat 6.2 crystals and the trigonal and tetragonal inclusion complexes, all three as separate

256


Table 6.8. Crystal data (294K) for some isostructural neat crystals and/or inclusion complexes of 6.1–6.7. The space group is P3121 (no. 152, or enantiomorph P3221, no. 154) except for neat 6.3, for which it has not been finally established. There are three host molecules in the unit cell, which have symmetry C2–2 in the crystals, and are located at Wyckoff positions x, 0, 1/3 etc. Additional crystal data for 6.1 with acetonitrile, chloroacetic acid, propanoic acid, trichloroethene, chlorobenzene, toluene and dioxane are given by Ung, Gizachew et al., (1995) Formula Type A host molecule {3(6.1)[(C4H10O2)0.75]} 1,2-dimethoxyethane {3(6.1)[(I2.–C2H5OH)0.5]} {3(6.1)[(ClCH2.–CH2Cl)0.75]} {3(6.1)[C4H8O2]} ethyl acetate {3(6.1)[C4H4S]} thiophene Neat 6.3 Neat 6.5 (Note 1) {3(6.6)[(CHCl3)1.5]} chloroform {3(6.7)[(C3D6O)1.2]} deutero-acetone Neat 6.8 Neat 6.9 Type B host molecule {3(6.2)[C4H8O2]} ethyl acetate {3(6.2)[Br(CH2)3Br]} {3(6.2)[(CCl4)1.2]} {3(6.2)[C6H4Cl2]} o-dichlorobenzene {3(6.2)[(C8H10)1.2]} o-xylene {3(6.4)[(C30H50)0.23) squalene {3(6.4)[C6H6]} benzene {3(6.4)[(ferrocene)0.75]}

Refcode / reference

˚) a(A

˚) c(A

˚ 3) V(A

KUBYUB; UBC92b

12.0416(3)

7.0110(2)

880.39(4)

KUBZEM; UBC92b ZACSEY; UGB95 EXHNEA20; UGB95 KUBZAI10; UCB95 FALRIT; DBH86 PICHAK; HBC93 BCD93

12.068(2) 12.075(1) 12.165(1) 12.4083(5) 11.906(1) 12.3430(4) 13.383(1)

6.984(3) 6.987(1) 7.001(1) 6.9702(4) 6.990(1) 6.8288(3) 7.026(1)

880.8(4) 882 897.3(2) 929.39(4) 858.1(1) 900.99(4) 1089.8(1)

WALDOC; BCD93

12.4957(6)

7.3076(3)

988.1(1)

POHYEQ; ABC97 ZEHFIB; BCS95

12.329(1) 12.609(2)

7.508(1) 7.209(2)

988.4(1) 992.6(3)

BUXRER10; DBH86 PAPSOO; UBC92a PAPSII; UBC92a VUSYIR; UBC93

13.192 13.206(2) 13.2812(2) 13.3717(6)

6.914 6.915(2) 6.904(1) 6.9045(4)

1042.0 1044.4 1054.6 1069.14

PAPSUU; UBC92a PIKJEY; UBCDRS93 FALROZ; DBH86 PIKJAU; UBCDRS93

13.380(2) 13.677(1) 13.740 13.7480(6)

6.905(1) 7.0533(9) 7.030 7.0312(5)

1070.5 1142.6 1149.5 1150.9

Notes: ˚ , ¼ 92.13 , (1) 6.5 has a stereoisomer (syn-2,anti-8) SEWYEY with a ¼ 12.473, b ¼ 13.056, c ¼ 15.449 A P21/c, Z ¼ 8. References: ABC97 – Ahn, Bishop, Craig, Downing and Scudder, 1997; BCD93 – Bishop, Craig, Dance, Scudder, Marchand and Wang, 1993; BCS95 – Bishop, Craig, Scudder, Marchand and Liu, 1995; BDH86 – Bishop, Dance and Hawkins, 1983; DBC92b – Ung, Bishop, Craig, Dance and Scudder, 1992b; DBH86 – Dance, Bishop, Hawkins, Lipari, Scudder and Craig, 1986; HBC93 – Hawkins, Bishop, Craig, Dance, Rae and Scudder, 1993; UBC92a – Ung, Bishop, Craig, Dance and Scudder, 1992a; UBC93b – Ung, Bishop, Craig, Dance and Scudder, 1993b; UBCDRS93 – Ung, Bishop, Craig, Dance, Rae and Scudder, 1993.

phases (cf. Ung, Gizachew et al., 1995). This is discussed below (Section 6.2.2.3). The well-resolved powder patterns (Fig. 6.36) show that it will be possible to study changes occurring on heating/cooling of the separate phases, including phase changes and chemical transformation.


257

tetragonal clathrate inclusion complex

trigonal tunnel inclusion complex

0

10

20 degrees 2

30

40

Fig. 6.36. Powder diffraction patterns (Cu K radiation) of the trigonal and tetragonal complexes of 6.2 with o-dichlorobenzene. (Adapted from Ung, Bishop, Craig, Dance and Scudder, (1993b).)

Single crystals of neat 6.4 have not yet been obtained but powder X-ray diffraction and IR spectroscopy indicate that its structure differs from those of the trigonal inclusion complexes; trigonal and neat crystals are thus separate phases. No information is available about the reported tetragonal clathrates. The relationships between the structures with 6.1–6.4 as hosts have been discussed in considerable depth (Bishop and Dance, 1988; Dance, Bishop and Scudder, 1986). 6.1 is ‘‘a potent host molecule. Helical tubulates are formed with a wide variety of small guests including alkenes, aromatic hydrocarbons, haloaromatics, ketones, ethers, esters, sulfides, amines and nitriles’’ (Ung, Bishop, Craig, Dance and Scudder, 1992b); the other diols show similar behaviour and it is clear that spatial rather than chemical properties of the guests are dominant. We shall describe the structure of 6.2 as representative and only briefly consider the other structures. The projection down [001] is shown in Fig. 6.37. The wall of a particular tunnel is made up of three host molecules in one orientation (>CH2 pointing inwards) and three in the opposite orientation (–CH2–CH2– pointing inwards); adjacent tunnels have converse arrangements. Parenthetically, it is this ordered combination of opposite orientations, together with the small range of intramolecular d(O . . . O) distances, that allows hosts of both types to form such similar arrangements in the solid state. The diol molecules are hydrogen bonded together along spirals in which each diol functions, within the spiral, as a double hydrogen bond donor or as a double hydrogen bond acceptor in alternating sequence. Part of a spiral is shown schematically as a linear diagram (the ring system of 6.2 is shown as a heavy line):

donor ....HOC(2)

acceptor donor acceptor C(6)OH....OC CO....HOC COH....OC CO.... H

H

H

H

258


Fig. 6.37. Projection of the trigonal complex of 6.2 with o-dichlorobenzene viewed down a threefold screw axis axis. For clarity, only hydroxyl hydrogens have been inserted (the smallest circles) and the guest molecules have been omitted. The limits of the diagram are 0 a 2, 0 b 2, 0.25 c 1.25; a single unit cell is emphasized.

˚ and that in 6.2 is 5.43 A ˚ ; thus inversion of The O . . . O distance2 within 6.1 is 5.60 A configuration at the substituted carbons (C(2) and C(6)) between 6.1 (representative of Type A host) and 6.2 (Type B host) does not change a fundamental geometrical parameter of these molecules. It is important to note that the triangle shown in projection in Fig. 6.37 is not planar, and is actually the projection of the spirals shown in Fig. 6.38. Now let us return to Fig. 6.37; if one starts with the molecule at the top left corner of the unit cell, the apex CH2 group is located at z ¼ 3/6. Moving clockwise around the tunnel wall the next molecule, which is hydrogen bonded to the first, has its apex CH2 group at z ¼ 5/6, i.e. a shift along z by 2/6. This is the requirement that two adjacent molecules in the tunnel wall should be hydrogen bonded. If one moves around the six molecules of the tunnel wall, and upwards along z, one returns to the top left hand corner (equivalent) molecule at z ¼ 15/6, i.e. a shift of 2c. This implies that the tunnel wall is made up of a spiral of hydrogen bonded molecules, with a pitch of 2c. However, the crystallographic periodicity is c, implying that there is a second spiral displaced along c by one period. The two coaxial spirals are internally but not mutually hydrogen bonded. The double spiral is shown in Fig. 6.39. One spiral is shown in space-filling form in Fig. 6.40, where the 2

These distances vary somewhat from complex to complex because of packing effects.


259

3c

2c

c

Representation of host molecule

b

Origin

Spiral about a three-fold screw axis

Fig. 6.38. Schematic diagram of the hydrogen-bonded spirals in the trigonal 6.2 tunnel inclusion complexes. For clarity, the molecule is represented as a V, with a CH2 group at the apex and the hydroxyls at the ends of the two arms. Complete representations are in the center of the diagram and ˚, half representations on the sides. The hydrogen bonds are shown as broken lines; d(O . . . O) ¼ 2.99 A < O . . . O . . . O ¼ 112.8 , (O . . . O . . . O . . . O) ¼ 106.2 .

Fig. 6.39. Stereodiagram of the double spiral of hydrogen bonded diol molecules forming the walls of one helical tunnel in the trigonal crystal structure of 6.1. [x(guest)]. The view is along the c axis and the spirals are left handed; one spiral starts at the top of the diagram and the other at the bottom. The spirals ˚ (¼ c) and are not hydrogen bonded to one another. The are separated in the z direction by 7.00 A generality of this arrangement among the trigonal Bishop-Dance complexes is demonstrated by using 6.1 as host in this diagram, and P3221 space group. (Reproduced from Andreetti, 1984.)

central tubular tunnel which contains the guests is also apparent. The traveled reader, en route to the Cistine Chapel, will no doubt have noticed that the double staircase leading to the Vatican Museum has just this structure; another architectural example is ‘Le Grand Escalier’ in the sixteenth century Chateau Chambord in the Loire Valley. Each of the donor hydroxyls in a spiral can act as an acceptor and conversely, and this available hydrogen bonding capability is used to link the spirals together laterally to form

260


the three dimensional crystal. This is shown in Fig. 6.41 where the ‘spines’ of the structure are enclosed in circles, with 32 axes (not shown) running normal to the page through their centres; it is important to remember that the three hydrogen bonds shown within any circle are at different heights along z and that diol molecules related by the 32 axes

Fig. 6.40. Stereodiagram of a space filling representation of one of the spirals shown in Fig. 6.35, from the same viewpoint. (Reproduced from Dance, Bishop and Scudder, (1986).)

a

b

Fig. 6.41. Projection view, parallel to the threefold screw axes, of the diol network in the crystals of 6.1 (and its inclusion complexes); the filled circles and dotted lines represent OH hydrogen atoms and hydrogen bonds respectively; other hydrogen atoms were omitted for clarity. The hydrogen bonded spines are circled and the tunnels are outlined as triangles. This figure should be compared with Fig. 6.37; despite the difference in the hosts (6.1 and 6.2) the overall structures are the same. (Reproduced from Bishop and Dance, 1988.)


261

running along the c edges of the unit cell are in different spirals of the double spiral (Figs. 6.39 and 6.40). Six structures of Table 6.8 are compared in Fig. 6.42; the outlined inner regions show the projected tunnel boundaries giving the ‘‘unobstructed cross-sectional areas’’ (UCA, previously designated Aun) that are available to guest species of any length for movement along the tunnel without steric impediment. Recent values (Bishop, 1996, Table 1) for the ˚2 nine helical structures 6.1–6.9 are 19.8, 29.2, 2.8#, 32.3, 1.2#, 34.0, 22.7, 9.9# and 8.9# A # (the values marked by are for the guest-free material). The common feature of the structures is the hydrogen bonding in the ‘spine’ regions; it is perhaps surprising that these hydrogen bonds are of appreciably different strengths, with d(O . . . O) ¼ 2.81, 2.98, 2.81 ˚ respectively. It is clear from the values of UCA that 6.3, 6.5, 6.8 and 6.9, and 3.05 A although isostructural with the other complexes, have constrictions in their tunnels which prevent the formation of tunnel inclusion complexes. 6.7 has a suitable UCA and does form complexes despite some shape differences (Table 6.7). The structures of complexes of ferrocene ((C5H5)2Fe) and squalene (2,6,10,15,19, 23-hexamethyl-2,6,10,14,18,22-tetracosahexaene, C30H50) with 6.4 have also been reported (Ung, Bishop, Craig, Dance, Rae and Scudder, 1993). These complexes, which have compositions (by NMR) of {3(6.4)(ferrocene)0.75} and {3(6.4)(squalene)0.21} respectively, are isomorphous with {3(6.4)(benzene)}; the anomalous scattering from the

OH

HO H3C

CH3

CH3

H3C

OH

HO

6.1 6.2

HO

OH

H 3C

CH3

OH

HO

6.3

6.4

OH

HO

CH3

H3C

HO

OH

H3C

CH3

CH3

H3C

6.5

6.6

Fig. 6.42. Comparative projections down [001] of one tunnel only for the inclusion complexes of 6.1, 6.2, 6.4 and 6.6, and of the neat crystals of 6.3 and 6.5. All six diagrams are on the same scale. Key hydrogen atoms defining the van der Waals surface of the host tunnels are shown as black dots. For other details see the caption to Fig. 6.41. (Reproduced from Bishop, Craig, Dance, Scudder, Marchand and Wang, 1993.)


262

Fig. 6.43. Side view of one tunnel in the {(6.4)3[(squalene)0.21]} complex showing the included squalene molecule. The oxygens of the diol molecules are emphasized. One column of diol molecules has been removed for clarity. The double spiral of diol molecules is seen edge-on. (Reproduced from Ung, Bishop, Craig, Dance, Rae and Scudder, 1993.)

Fe of ferrocene allowed the space group of the crystal used to be established as P3121 (such a determination of absolute configuration has limited value unless related to other chiral physical properties of the crystal, for example the face development). The guest molecules are disordered in both these complexes and special techniques were used for their refinement. It was found that three ferrocene molecules were disordered over four unit cell repeats in the [001] direction, that its five fold axis was inclined at 66 to [001], and that it was also disordered orientationally. The arrangement found for the squalene complex is shown in Fig. 6.43. That the disposition of the hydroxyls is vital to the ability to form trigonal helical tubuland complexes while the nature of the atomic arrangement bridging them is secondary is neatly shown by the fact that 6.7 (Fig. 6.44) forms such complexes with ethyl acetate and deutero-acetone (Table 6.8). The distance between oxygen atoms in 6.7 ˚ (compared to 5.6 A ˚ in 6.1). 6.8 and 6.9 form isostructural helical complexes is 5.98 A (Table 6.8) but the values of the unobstructed cross-sectional areas (UCA) are so small (9.9 ˚ 2 respectively) that only very small guests will perhaps form complexes. and 8.9 A H

OH

HO CH3 6.7

OH HO

CH3

CH3

CH3 6.8

OH

HO

CH3 6.9

CH3


263

Fig. 6.44. Formulae 6.7–6.9 shown above as conventional line diagrams (some hydrogens omitted for clarity), with 6.7 below viewed down the molecular twofold axis (all hydrogens are shown). This twofold axis is not immediately obvious in the line diagram.

In the composition formulae {3(host)[guest]x}) of the tunnel inclusion complexes, ‘x’ takes on only a limited set of values – these are 0.75, 0.86, 1, 1.2 and 1.5 (squalene is an understandable exception at 0.21). The structural explanation is as follows. A single unit cell, with three host molecules, has a single tunnel (see Figs. 6.36 and 6.40); thus ‘x’ gives the ‘number’ of guest molecules in a single period of the tunnel along z. When x ¼ 0.75 (6.1 with monoglyme, 1,2-dichloroethane; 6.2 with ferrocene), there are three guest molecules in four unit cells; for 0.86 (6.1 with trichloroethylene, toluene), there are six guests in seven unit cells; for 1 (6.1 with acetonitrile, ethyl acetate, thiophene, chlorobenzene, dioxane; 6.2 with ethyl acetate, 1,3-dibromopropane, CBr2F2, o-dichlorobenzene; 6.4 with benzene), there is one guest per unit cell; for 1.2 (6.1 with 1,2-dichloroethane, propanoic acid; 6.2 with CCl4, o-xylene; 6.7 with acetone-d6), six guests in five unit cells; for 1.5 (6.6 with CHCl3), three guests in two unit cells. The guest molecules are necessarily orientationally disordered within the tunnels and also longitudinally, except for the x ¼ 1 situation. However, the compositions are commensurate in the sense used for urea inclusion complexes; thus one wonders whether analogous diffuse scattering occurs at room temperature, and phase transitions on cooling. We have not encountered such reports. A comparison of the hexagonal (not rhombohedral) urea and the trigonal Bishop–Dance tunnel inclusion complexes (respectively, UTIC and BDTIC) is rewarding: 1. Both are chiral, UTIC because of the chiral arrangement of the achiral urea host molecules, and BDTIC because of spontaneous resolution of chiral host molecules and their arrangement in a chiral space group.


264

2.

3.

4.

5.

Some BDTIC are primary solid solutions of guest in host, others are separate phases in the host-guest phase diagram. All UTIC are separate phases in the host-guest phase diagram. The UTIC often show spectacular diffuse (X-ray) scattering, the details of which vary with the nature of the guest. There are also phase changes at low temperatures. Neither of these features has been reported for the BDTIC, perhaps because they have not been sought. The UTIC can be commensurate or incommensurate, depending on the nature of the guest. The BDTIC appear to be commensurate and only a limited number of specific host:guest ratios have been reported, as described above. Examples are 3 : 0.75, 3 : 0.86, 3 : 1, 3 : 1.2 and 3 : 1.5. Strong hydrogen bonding between host and guest can lead to partial deformation of the host framework in both types of complex (for BDTIC see 6.2.2.4 ‘‘Derived structures’’ below).

6.2.2.3 The ellipsoidal tetragonal clathrate complexes of some Bishop–Dance hosts These structures should logically be discussed in Chapter 7 but it is convenient to place them here because the hosts belong to the same chemical family. We have noted above that 6.2 recrystallized from benzene gives the prototype ellipsoidal clathrate {4(6.2)[C6H6]}, which is tetragonal, space group I41/acd, a ¼ 23.021, ˚ , V ¼ 10011 A ˚ 3, Z ¼ 8(BUXRIV10; Hawkins, Bishop, Dance, Lipari, Craig c ¼ 18.889 A and Scudder, 1993). Other guests giving the ellipsoidal clathrate structure with 6.2 include acetone, acetonitrile, dichloromethane (LORQOY; Bishop, Craig, Dance, Scudder and Ung, 1999), toluene, nitrobenzene, benzylcyanide, chlorobenzene, bromobenzene, o-dichlorobenzene, p-dichloro- and p-dibromobenzene, p-bromo- and p-nitrotoluene (Ung, Bishop, Craig, Dance and Scudder, 1992a). Chloroform and bromobenzene form both types (trigonal and tetragonal) of inclusion complex with 6.2 ((Ung, 1993). 6.10 is another host which comes into this group, forming an isomorphous clathrate with CHCl3 as guest. Note that 6.10 is a Type A host and (according to available reports) forms only a tetragonal ellipsoidal clathrate (with chloroform) but not trigonal helical tunnel inclusion complexes, in contrast to the other Type A hosts. Crystal data and references are summarized in Table 6.9.

HO

OH

CH3

CH3

6.10

S

In contrast to the trigonal complexes of ethyl acetate and other guests with 6.1, 6.2 and 6.4, spontaneous resolution does not occur on crystallization of 6.2 with benzene but a


265

Table 6.9. Crystal data for the tetragonal ellipsoidal clathrates of two Bishop–Dance hosts (cf. Table 4 of Bishop, 1996) Formula

Refcode / reference

˚) a(A

˚) c(A

˚ 3) V(A

{4(6.2)[CS2]} {4(6.2)[CH3CN]} {4(6.2)[CH2Cl2]} {4(6.2)[C6H6]} {4(6.2)[o-C6H4Cl2]} {4(6.10)[CHCl3]}

POLFIF; BCMS94 LORQUE; BCDSU99 LORQOY; BCDSU99 BUXRIV10; HBDLCS93 VUSYEN; UBCDS93b QULLAK; BCDKMPS93

23.031(2) 22.979(1) 23.007(1) 23.021(1) 23.442(4) 23.042(3)

18.773(1) 18.864(1) 18.869(2) 18.889(2) 18.928(4) 19.022(3)

9958(2) 9961(1) 9988(1) 10010(1) 10401(3) 10099(2)

References: BCDKMPS93 – Bishop, Craig, Dance, Kim, Mallick, Pich and Scudder, 1993; BCDSU99 – Bishop, Craig, Danc, Scudder and Ung, 1999; BCMS94 – Bishop, Craig, Marougkas and Scudder, 1994; HBDLCS93 – Hawkins, Bishop, Dance, Lipari, Craig and Scudder, 1993; UBCDS93b – Ung, Bishop, Craig, Dance and Scudder, 1993b.

racemic structure is obtained. It would be interesting to see whether a complex is obtained by crystallization of resolved 6.2 from benzene. The resolved 6.2 could, at least in principle, be obtained by hand separation of enantiomorphic crystals of {6.2[1/3(ethyl acetate)]}; presumably neat 6.2 would appear instead of a complex. It is interesting to note that 2,6-dimethylideneadamantane-1,3,5,7-tetracarboxylic acid forms 1 : 2 inclusion compounds with mesitylene, with space group I41/acd and structural similarity to the ellipsoidal clathrates (Ermer and Lindenberg, 1991; VOBFUN; cf. VOBFOH, VOBGAU). Using 1,2-dichlorobenzene as guest allows preparation of either the helical tubuland type or the ellipsoidal clathrate type of crystal, depending on crystallization conditions (Ung, Bishop, Craig, Dance and Scudder, 1993b). It has been reported that 6.2 can also form both types of inclusion complex with CHCl3 and C6H5Br as guests (Ung, Bishop, Craig, Dance and Scudder, 1992b). These authors studied the relative stability of the 6.2– 1,2-dichlorobenzene polymorphs (sic) by heating samples in sealed tubes at 60–65 for 20 hours, and found that the tetragonal form was stable under these conditions but that the trigonal form had almost completely transformed to the tetragonal form. They inferred that the ‘‘ellipsoidal clathrate structure is of lower energy and the one preferred when allowed by constraints of guest size and shape.’’ We note that the two forms are not polymorphs3 but phases of different composition (4 : 1 and 3 : 1 respectively) in the binary 6.2–1,2-dichlorobenzene phase diagram. Furthermore the ellipsoidal form is racemic while the helical tubuland form is enantiomorphic; racemization of 6.2 during transformation seems extremely unlikely as this would require breaking covalent bonds. A possible explanation is that the sample of 6.2 used was a conglomerate (mixture of both enantiomorphs). The matter is both complicated and interesting, and requires further investigation. Although these phenomena remind one of the urea channel inclusion complexes which, for some guests, can be obtained as either hexagonal (enantiomorphic) or rhombohedral (racemic) (Table 6.1), the resemblance is superficial. The behavior of the urea complexes appears to be true polymorphism, and the change from enantiomorphic to 3

Polymorphism is defined as ‘‘the appearance of different crystal structures for the same chemical entity.’’


266

c

origin

a

Fig. 6.45. Schematic representation of the tetragonal ellipsoidal clathrate structure viewed down [100]. The projected material lies within the limits 0 x 0.50; 0.25 y 0.75; 0.25 z 1.00. The guests are shown for convenience as spheres, but it should be noted that this hides their mutual orientation as shown for the benzene complex below (Fig. 6.47). The hosts are represented as (H)O–C—C–O(H); in the double spiral on the left (O–C–C–O) is 88 , and –88 in the double spiral in the center. Hydrogen bonds are shown by dashed lines. Left and centre spirals are linked by slightly non-planar quadrilaterals (torsion angles about the bonds 9 ).

c

a

b

Fig. 6.46. A ball and stick representation of the tetragonal ellipsoidal clathrate structure viewed down [100]; only hydroxyl hydrogens have been included. The projected material lies within the limits 0 x 0.50; 0.25 y 0.75; 0.25 z 1.00. The guests are shown for convenience as spheres, but it should be noted that this hides their mutual orientation as shown for the benzene complex below (Fig. 6.47). This diagram is to be compared with Fig. 6.45.

racemic involves a change of arrangement, only hydrogen bonds being broken and reformed; however, here too further investigation seems desirable. The crystal structure of the tetragonal clathrates is complicated; detailed descriptions have been given by Bishop (1996) and in earlier papers by this group. Host diol molecules occupy the 32 general positions of the centrosymmetric space group I41/acd, with the


267

Fig. 6.47. Cross-sectional representation of the cavities in {4(6.2)[C6H6]}, linked along the twofold axes parallel to [001], showing the van der Waals surface due to the hydrogen atoms of the host molecules, and the major (80%) orientation of the benzene guest molecules (the minor orientation is rotated 30 about the benzene sixfold axis). The benzenes are at 222 (D2) sites and contiguous sites along [001] are related by the 4 (S4) operation. (Reproduced from Bishop, Dance and Hawkins, 1983.)

guest molecules (possibly disordered) at the eight positions with D2-222 symmetry. The host molecules are segregated by enantiomer around 41 and 43 axes where they form double spirals (Ung, Bishop, Craig, Dance and Scudder, 1993b). This is illustrated schematically in Figs. 6.45 and 6.46 for the isomorphous complexes of 6.2 listed in Table 6.9. The structure of {4(6.10)[CHCl3]} is similar. The tunnels in these complexes are so constricted that they are better described as having clathrate rather than tunnel structures; this is shown in Fig. 6.47 for the benzene complex of 6.2, where the benzene guests are 80% ordered in one orientation. 6.2.2.4 Derived structures When 6.1 is crystallized together with p-chlorophenol or hydroquinone, or 6.2 with p-methoxyphenol, structural arrangements related to those of the helical tubuland family are obtained. Because of this relationship it is more convenient to discuss them here rather than in Chapter 12 (Hydrogen Bonded Molecular Complexes and Compounds), where they strictly belong. {6.1p-chlorophenol} (SUFFOO10) and {6.10.5(hydroquinone)} (SUFFUU10) are ˚, isostructural (a ¼ 6.927(1) [6.864(1)], b ¼ 12.696(1) [12.829(1)], c ¼ 19.286(4) [15.974(1)] A ˚ 3, both P21/c, Z ¼ 4) (Ung, ¼ 94.83(1) [103.099(3)] , U ¼ 1690.1(4) [1370.0(2)] A Bishop, Craig, Dance and Scudder, 1994); {6.2p-methoxyphenol} is also isostructural (Bishop, Craig, Dance, Scudder and Ung, 1994). In {6.1p-chlorophenol} the hydroxyl of the p-chlorophenol molecule occupies one of the positions of the hydrogen bonded spine, forming two dimensional sheets instead of tunnels (Fig. 6.48). This is a mixed framework structure in the nomenclature of Chapter 12. {6.2p-methoxyphenol} has a very similar structure, while in {6.10.5(hydroquinone)} the bifunctional hydroquinone bridges

268

(a)


(b)

p-chlorophenol

Fig. 6.48. Derived structures: (a) the trigonal hydrogen bonded spine . . . OH . . . OH . . . OH.. constitutes the structural core of the helical tubuland lattice of 6.1 and its tunnel inclusion complexes; (b) the similar spine found in {6.1p-chlorophenol} and {6.2p-methoxyphenol} where one of the sites is taken up by the X–C6H4–OH molecule. (Reproduced from Ung, Bishop, Craig, Dance and Scudder, 1993a.)

between two hydrogen bonded spines. These derivative structures are racemic, in contrast to the usual run of trigonal helical tubuland structures. 6.2.3

Ta4P4S29 – an inorganic framework containing sulphur chains

Metal-grey single crystals of tantalum disulphide thiophosphate were synthesized from the elements and found (Fiechter, Kuhs and Nitsche, 1980) to crystallize in the tetragonal ˚. system in the centrosymmetric space group I41/acd with a ¼ 15.849(3), c ¼ 13.143(4) A The TaPS6 compound was formulated as Ta[PS4jS2], the notation being intended to emphasize the presence of two kinds of sulphur atoms. The structure, shown in Fig. 6.49, led the authors to comment ‘‘Another interesting feature of the Ta[PS4jS2] structure is the existence of tunnels extending along the fourfold screw axes. It appears probable that foreign atoms or small molecules can be inserted into and move in these tunnels, the free ˚ .’’ diameter of which is about 4.65 A This prophecy was realized by preparation of a compound of composition Ta4P4S29 by heating stoichiometric quantities of the elements in an evacuated tube for 10 days at 500 C (Evain, Queignec, Brec and Rouxel, 1985). The black crystals were tetragonal with ˚ , V ¼ 3309.9 A ˚ 3, Z ¼ 4. The space group of the crystal used for a ¼ 15.571, c ¼ 13.652 A the analysis was shown to be P43212, which is chiral. The basic framework of the structure has composition TaPS6 and is made up of bicapped biprismatic Ta2S12 units, including sulphur pairs, bonded to each other through tetrahedral PS4 groups which share sulphurs. Thus the two structures are made up of the same [Ta2S12] and [PS4} units linked in the same way, but the [Ta2S12] biprisms are differently tilted. The large tunnels running through the framework in the [001] direction contained S10 chains (average ˚ , <S–S–S ¼ 105.8 ) in the form of right-handed helices; there were only d(S–S) ¼ 2.052 A


269

x y

Fig. 6.49. View of the contents of the unit cell of Ta[PS4jS2], looking down [001]. Ta filled, S hatched, P quartered. (Data from Fiechter, Kuhs and Nitsche, 1980.)

van der Waals interactions between framework and chains. Fibrous sulphur contains similar ˚ , <S–S–S ¼ 106 ) but of both senses. The semiconducting and helices (d(S–S) ¼ 2.07 A diamagnetic inclusion complex was formulated as {TaV4PV4(SII)16-(S2II)4[1/2S010]}. There are other analogous structures. Cycloo¨ctasulphur has been shown to be encapsulated in an open metal-sulphide framework grown from molten cesium polysulphide fluxes (Marking and Kanatzidis, 1995). The complex has the formula {Cs2Sn3S7[1/2S8]} and is monoclinic (space group C2/c, Z ¼ 8); neat Cs2Sn3S7 has not yet been prepared. The structure of the complex is based on Sn3S4 defect cubane units connected by (-S)2 bridges forming large open rings in Sn3S72 anionic layers. These layers are stacked along [001] and mutually aligned so that tunnels are formed. The Csþ cations are located between the layers, while the S8 molecules are in the tunnels in a disordered arrangement of two conformations. A Te8 ring has been reported in Cs3Te22 (Sheldrick and Wachtold, 1995), and Se and Te chains have been injected into the zeolite mordenite (Bogomolov, Poborchoy, Romanov and Shagin, 1985; Terasaki, Shiokawa, Ito, Watanabe and Thomas, 1989). 6.2.4 The tunnel hydrates Most families of tunnel inclusion complexes are characterized by a particular host which forms a structure in which many different kinds of guest can be accommodated. The tunnel hydrates are characterized by a variety of hosts which all crystallize in sheets pierced by tunnels in which water molecules, the common guest, are accommodated. The group of tunnel hydrates to be discussed here constitute a different phase from the anhydrous crystals of the host, with further subdivisions depending on the degree of host– water hydrogen bonding and on the number of water molecules per tunnel. There are also crystals in which the water is absorbed zeolitically but these will not be considered here. 6.2.4.1

Tunnel hydrates with several water molecules per tunnel cross-section

In this group of complexes the tunnels have lateral dimensions large enough to accommodate several water molecules which are bonded both to one another (across and along

270


the tunnel) and also to the host framework. One example is N,N 0 -ethylenediaminesuccinic acid pentahydrate (EDDS; 6.11). EDDS is a hexadentate chelating agent which is isomeric with ethylenediamine tetraacetic acid (EDTA) and chemically quite similar to it. However, a major difference between EDDS and EDTA is that the former has two asymmetric carbon atoms; if EDDS is synthesized from, say, l-aspartic acid then both these carbons (asterisked in the formula, where all except the hydroxyl hydrogens have been omitted) will have the same chirality. The pentahydrate of EDDS synthesised in this way crystallizes in space group P21212 with Z ¼ 2 (Scarbrough and Voet, 1976; ENSUCP); thus the EDDS molecule has two fold symmetry in the crystal. The EDDS molecules are hydrogen bonded together to form networks with tunnels (along [001]) large enough to contain five water molecules (Fig. 6.50). Although the water molecules are not part of the network, they are hydrogen bonded both to EDDS molecules and to one another. The pentahydrate loses water to the atmosphere rather easily. O –O O

+ * N

+ N *

OH O

HO O – O

Rather similar networks of host molecules which leave central tunnels containing water molecules are found in {(þ)-isoo¨livil[acetoneH2O]}, where the acetone is also located in the tunnels (Wong, Manners and Palmer, 1977; ISOVIL10), and in tetracycline hexahydrate (Caira, Nassimbeni and Russell, 1977; TETCYH01). The water molecules are

Fig. 6.50. Stereoview of the unit cell of EDDS pentahydrate, looking down [001], with [100] horizontal and [010] vertical. Water oxygen atoms are represented as 50% probability ellipsoids. Intermolecular hydrogen bonds between EDDS molecules are shown as broken lines; hydrogen atoms have been omitted for clarity. (Reproduced from Scarbrough and Voet, 1976.)


271

bonded both to one another and to the surrounding network, of which they do not, however, form a part. 6.2.4.2 Tunnel hydrates with one water molecule per tunnel cross-section ˚ ), {caffeine[0.8 H2O]} Theophylline monohydrate (Sutor, 1958b; THEOPH4) (c ¼ 4.50 A 5 ˚ (Sutor, 1958a (CAFINE ); Gerdil and Marsh, 1960) (c ¼ 3.97 A), {thymine[0.8 H2O]} ˚ )), and {biuret[0.8 H2O]} (Hughes, Yakel and (Gerdil, 1961; THYMMH) (c ¼ 3.65 A ˚ ) all have rather Freeman, 1961 (BIUHYD); Craven, 1973 (BIUHYD10)) (c ¼ 3.82 A similar structures in which the host molecules lie in pleated sheets pierced by tunnels whose cross-sections are such as to permit accommodation of zigzag chains of water molecules. O H3C

O

N

N N

O

R

N

CH3 R=H thophyline = CH3 caffeine

H

O

O CH3

N N

H

O

N

NH2 NH2

H thymine

biuret

The common feature in their cell dimensions is the shortness of the c axes, which are about equal to the thickness of the pleated sheets. The water molecules are hydrogen bonded both to the host molecules and to one another, as illustrated in Fig. 6.51 for {thymine[0.8H2O]}. In theophylline monohydrate successive water molecules in a chain hydrogen bond to successive theophylline molecules and are thus ordered, but in {caffeine[0.8H2O]} one out of every five water molecules in a chain, on the average, appears to be missing. {Deutero-biuret[0.8H2O]} was also studied by neutron diffraction (Craven, 1973) but the nature of the chains could not be defined more clearly. It would seem that these structures are incommensurate at room temperature, presumably because of differences between the thickness of the pleated sheets of host molecules and the periodicity of the chains of water molecules; however, the water molecules are essential for maintaining the pleated sheets. Other possible examples are trimesic acid trihydrate ˚ ) (Herbstein and Marsh, 1977), the isomorphous pair 2,4,6which has a short c axis (3.68 A trinitro-1,3-benzenediol2/3H2O (BOCNEM) and the corresponding triol (BOCNIQ) (Pierce-Butler, 1982), and salazopyrine1/2(N,N-dimethylformamide)2.25H2O (van der Sluis and Spek, 1990; SENCIX), which differ in detail from the previous examples. The basic chromium acetate compound {(OCr3(CH3COO)63H2O)þ[Cl .6H2O]} has its anions and (crystal) water molecules in tunnels (Chang and Jeffrey, 1970; CRACOP11). The related tris(5-acetyl-3thienyl)methane dihydrate {(TATM)2H2O} is noted in Section 8.6.

4 THEOPH10 is an independent re-determination of the structure by Sun et al. (2002), with change of the space group from P21 to P21/n. 5 Update: CAFINE01 (Edwards et al., 1997).


272

b z

x

y

M4 M3

H2O M5

H2O

M2 1 2

1 4

N2 O M1

O2 N1

–1 4

–1 2

O1

Fig. 6.51. A clinographic projection of the crystal structure of thymine[0.8H2O], with the chains of water molecules emphasized. Molecules M1, M2, M3 and M4 are successively related by centers of symmetry at 0,0,0; 1/2,0,1/2 and 101 respectively. M5 is related to M2 by the glide plane (dotted parallelogram) at y ¼ 1/4, and to M3 by a twofold screw axis at 1/2, y, 3/4. (Reproduced from Gerdil, 1961.)

6.3 6.3.1

Tunnel inclusion complexes with van der Waals bonded hosts Tunnel inclusion and other complexes of deoxycholic acid and related compounds

The term ‘‘choleic acid,’’ coined by Demarçay (1838), was used by Lachinov (1885) (Latschinoff ) for the crystalline compound isolated from ox bile, which was later shown to be an inclusion complex of deoxycholic acid containing a mixture of palmitic and stearic acids. Thereafter this term was applied in a more restricted sense to describe the molecular complexes of deoxycholic acid and apocholic acid; however, sufficient information is now available to make the use of such overall terms unsatisfactory and we shall avoid ‘‘choleic acid’’ where possible. The formulae of the steroids DCA (deoxycholic acid), CA (cholic acid), ACA (apocholic acid) and CDCA (chenodeoxycholic acid) are shown below, as well as a perspective, side-on view of the DCA molecule which emphasizes that it is far from planar (Scheme I). All these hosts form tunnel inclusion complexes of related structural types, and also other complexes which are not inclusion complexes. We discuss them together here because of the chemical relationships among the hosts, in a hopefully pardonable retreat from our strict constructionist structural principles! There are a number of general reviews with a stress on the historical background (Sobotka, 1934; Fieser and Fieser, 1959; Herndon, 1967) while the structural chemistry has also been discussed (Giglio, 1984); there is a comprehensive chemical and structural review by Miyata and Sada, 1996). The preparation of inclusion complexes of DCA was first reported in 1916 (Wieland and Sorge, 1916). The crystal structures of many DCA and cholic acid inclusion complexes are now known (84 hits for DCA and 100 for cholic acid in November, 2002 version of CSD), and structures have also been reported for some examples of cholanamide, ACA and CDCA complexes.

VAN DER WAALS BONDED HOSTS

H3C 21

R1

20

CH3

22 23

17 1

C

CH3

273

D HO

O

HO

O

2

HO

A

B

4

6

R2

6.12 OH

H3C CH3

C

CH3

A

D

B

HO 6.13

There are a number of variations on the formula 6.12: 1. R1 ¼ OH, R2 ¼ H deoxycholic acid (DCA); 3,12-dihydroxy-5ß-cholan-24-oic acid; C24H40O4. 2. R1 ¼ H, R2 ¼ OH chenodeoxycholic acid (CDCA ); 3,7-dihydroxy-5ß-cholan-24-oic acid; C24H40O4. 3. R1 ¼ OH, R2 ¼ OH cholic acid (CA); 3,7,12-trihydroxy-5ß-cholan-24-oic acid; C24H40O5. 4. In cholanamide (3,7,12-trihydroxy-5ß-cholan-24-amide; C24H41NO4) the hydroxyl of the carboxyl group of cholic acid is replaced by amide NH2. 6.13 is apocholic acid (ACA), C24H38O4.

6.3.1.1 The complexes of deoxycholic acid There is some confusion about the crystal structure of neat DCA, suitable single crystals being difficult to prepare. Go and Kratky (1934; ZZZRES) reported an orthorhombic cell ˚ , without space group or structure. However, Ferro, with a ¼ 31.37, b ¼ 49.90, c ¼ 14.12 A Quagliata, Giglio and Piacente (1981) reported that microcrystals give powder patterns closely similar to those of orthorhombic {(DCA)2[phenanthrene]}. This possibly implies that the orthorhombic DCA complexes are primary solid solutions of guest in host, contrary to evidence from the phase diagrams noted below. A 13C cross polarization/ magic angle spinning NMR spectrum has been reported for a polycrystalline sample of neat DCA (Heyes and Dobson, 1990), from which it is possible to infer that there is one molecule in the asymmetric unit. This crystallographic information is compatible with the known structures of the orthorhombic Group 1A and 1B complexes (Table 6.10), but does not exclude other possibilities.


274

Hydrophobic side C18 C19 Head

Ring D

Ring B

C7

Tail Ring C

C12

C21 =O

Ring A

OH C24 Hydrophilic side

OH

C3

OH

The DCA molecule is viewed approximately in the mean plane of Rings B, C and D. The DCA molecule has an hydrophobic side with three protruding methyl groups and an hydrophilic (polar) side with two hydroxyls and one carboxyl. The standard numbering of the atoms is shown in part, as well as the distinction between hydrophobic and hydrophilic sides of the molecule. The torsion angle ¼ (C17–C20–C22–C23) as shown in the uppermost diagram of the Scheme; this is chemical numbering for DCA and may differ for other situations.

DCA forms orthorhombic, tetragonal and hexagonal inclusion complexes (and sodium deoxycholate forms helical macromolecular associations in solution (Blow and Rich, 1960)); there are also a few monoclinic DCA complexes. The inclusion nature of the tetragonal and hexagonal complexes is not as marked as that of the orthorhombic complexes. The guests in these molecular complexes include such varied substances as organic acids (mono and dibasic), aliphatic and aromatic hydrocarbons, alkaloids, alcohols, azo dyes, esters, ethers, phenols, -carotene, methyl orange. In the group of orthorhombic complexes the hydrophobic sides of the DCA molecule make up the tunnel walls (and also in the tetragonal crystals, but in a somewhat different sense as described below), while the converse occurs in the hexagonal crystals. The crystal structure of CDCA (Lindley, Mahmoud, Watson and Jones, 1980; CHNOCH) has been reported; all the hydroxyl and carboxyl groups are involved in the network of hydrogen bonds between the molecules. The crystal structures of a number of inclusion complexes of cholic acid (Section 6.3.1.3 below) and of some non-inclusion type complexes have been reported; the structure of cholic acid itself is based on an extensive network of hydrogen bonds (Miki, Kasai, Shibakami, Chirachanchai, Takemoto and Miyata, 1990; JEWDEY). Phase diagrams (melting and thaw points) have been determined for the following systems: ACA with montanic acid (CH3(CH2)25CH2COOH) and stearic acid (CH3(CH2)15CH2COOH) (Rheinboldt, Pieper and Zervas, 1927), ACA with palmitic acid (CH3(CH2)13CH2-COOH) and cetyl alcohol (CH3(CH2)14CH2OH) (Rheinboldt, Flume and Ko¨nig, 1929), ACA with camphor (Rheinboldt, Ko¨nig and Flume, 1929) and DCA with stearic acid, palmitic acid and cetyl alcohol (Rheinboldt, Flume and Ko¨nig, 1929), DCA with camphor (Rheinboldt, Ko¨nig and Flume, 1929). All the diagrams have very


275

200 180 Melting points

t (°C)

160

Eutectic

140

Thaw points

120

Inclusion complex

100 80 Eutectic

60 0

0.2

0.4 0.6 0.8 Mol. Fraction ACA

1

Fig. 6.52. Phase diagram of ACA (C24H40O4) with montanic acid, redrawn from Rheinboldt, Pieper and Zervas (1927). There are eutectic points at 1 and 99 mol.% ACA. The 8 : 1 compound of ACA and montanic acid is at 88.9 mol.% ACA, in the region of the maximum melting point. The lines are guides to the eye.

similar appearances, which rather resemble the phase diagram of perhydrotriphenylene with n-heptane (Fig. 6.65; Farina and Di Silvestro, 1980, 1982). We have recalculated the phase diagram of ACA with montanic acid, plotting temperature against mole fraction of ACA (instead of weight fraction) (Fig. 6.54); the complex is a phase different from neat ACA. The guests in the orthorhombic group of DCA complexes include the following types: 1. alkanes (branched and unbranched pentanes to decanes) (Huntress and Philips, 1949); hexadecane and dodecane; 2. fatty acids (acetic through stearic, including branched chains) (Herndon, 1967); 3. aromatic hydrocarbons (toluene, the xylenes, naphthalene, acenaphthene, phenanthrene, 1,2-benzanthracene, 1,2,5,6-dibenzanthracene, methylcholanthrene, hexahydromethylcholanthrene (Fieser and Newman, 1935); complexes were not formed with chrysene, pyrene, triphenylene, perylene and 1,2-benzpyrene. Although any explanation of negative results is necessarily inconclusive, it seems probable that only aromatic molecules with relatively small cross-sections can form complexes while larger aromatics do not – the criterion is geometric rather than chemical, the shapes of the tunnels in the structures described below being such that only smaller molecules can enter. A similar situation is encountered with the tunnel inclusion complexes of N-(p-tolyl)tetrachlorophthalimide (Section 6.3.5). The orthorhombic inclusion complexes all contain the same structural motif, which is a pleated bilayer sheet composed of hydrogen-bonded DCA molecules lying about the (200) planes (Fig. 6.53). Hydrogen bonds between hydroxyls O(3), O(12) and the carboxyl oxygens link DCA molecules both in the [010] and [001] directions; if one uses the O(3)– H . . . O(24) ¼ C hydrogen bond along [010] as designator, then this can be called a ‘headto-tail’ complex, but the real situation is more complicated. The hydrophilic portions of

276


O

b

S D G

H V

V

C

B

T

a

Fig. 6.53. Schematic structure of orthorhombic DCA molecular complexes, using the following symbols: B bilayer of DCA molecules (shown stippled); G guest molecule; H helical hydrogen bonding scheme holding hydrophilic sides of DCA in bilayer; V van der Waals contacts between hydrophobic sides of DCA bilayer; C tunnel along [001] with hydrophobic walls, containing guest molecules; O origin of unit cell; S 21 axis at z ¼ 0 in P21212 or at z ¼ 1/4 in P212121; T twofold axis in P21212 or 21 axis in P212121, including both exact 21 and approximate 2 axes in those examples with true space group P212121 but which approximate to P21212. (Reproduced from Jones, Schwarzbaum, Lessinger and Low, 1982.)

the molecules face into the interiors of the bilayers and thus hydrogen bonding of guest to host is not possible. Adjacent bilayer sheets are juxtaposed so as to leave tunnels in the [001] direction, with only van der Waals interactions between adjacent bilayers. To quote Jones, Schwarzbaum, Lessinger and Low (1982): ‘‘The remarkable capacity of DCA for accommodating such a wide variety of guest molecules in the channels of these orthorhombic crystals does not result from any conformational changes in the individual DCA molecules, but is possible because the same basic DCA bilayers can be shifted in major and subtle ways relative to each other along a and/or b and/or c such that the combined host-guest crystal complex achieves a stable configuration. Since the guest molecules are packed into hydrophobic channels, the accommodations seem to be those which maximise van der Waals attraction, that is, those which result in the closest packing in the crystal as a whole.’’ The torsion angle shown in Scheme I is (þ)-gauche, i.e. around 60 . The only exception encountered is for COXDEY (Table 6.10), which is ()-gauche. To a first approximation all the orthorhombic DCA tunnel inclusion complexes form ˚ , with one isostructural group with a 25.5–27.35, b 13.35–13.81 and c 7.1–7.2 A ˚ 3. Further 4 {DCA-x[guest]} units in a unit cell whose volume ranges from 1250 to 1350 A subdivisions can be made in terms of space group (Craven and de Titta, 1972). Complexes ˚ are Group IA, those with space group P212121 and with space group P21212 and c 7.2 A ˚ are Group IB (both Z ¼ 4, one DCA molecule in the asymmetric unit) and those c 7.2 A ˚ are Group IC (Z ¼ 8, two DCA molecules in the with space group P212121 and c 14.4 A asymmetric unit). Some of the orthorhombic complexes whose structures have been reported are classified in Table 6.10. Although this sample may well not be representative of the total population, we note that the numbers of reported structures are 6, 18 and 7 in Groups IA, B and C respectively, suggesting that Group IB structures are most widespread; however, this does not take into account the many DCA-fatty acid complexes, most of which belong, at least to a first approximation, in Group IA.


277

In the Group IA complexes, exemplified by 2 : 1 DCA-[(þ)camphor], the adjacent bilayer sheets are at the same height along z because they are related by the twofold axis parallel to [001]. The guest (þ)camphor molecule is twofold disordered and fits into one translation along [001]. The 2 : 1 DCA-[()camphor] complex has a very similar structure, with three to six sites for the disordered camphor molecule; it seems unlikely that these enantiomers could be separated by formation of DCA complexes. Go and Kratky (1934) have reported that many fatty acid complexes crystallize in Group IA (palmitic (ZZZQCS), stearic (ZZZQHO), lauric (ZZZQIC), caprylic (ZZZQIM), heptanoic (ZZZQIO) and propionic (ZZZQIU) acids; the DCA:guest ratios are 4 : x with x depending on the nature of the guest. ˚ (c/2) in the In the Group IB complexes adjacent bilayers are mutually shifted by 3.6 A [001] direction. An example is the 1 : 1 {DCA-[acetic acid]} complex where the acetic acid molecules form hydrogen bonded chains along the tunnels. The p-diiodobenzene, phenanthrene and palmitic acid-ethanol complexes have closely similar structures; the trebling of the c axis in the phenanthrene complex is due to the disposition of the guest, not of the host, molecules. In Group IC (exemplified by {2DCA[cyclohexanone]}), opposing walls of the DCA tunnel are related by a crystallographic twofold screw axis and also by an approximate two fold axis (because the two crystallographically independent DCA ˚ (c/2) translation along c, and because c is molecules are related approximately by a 7.07 A now double the length found in Groups IA and IB). Thus the space group is strictly ˚ , and approximately P21212 with c ¼ 7.07 A ˚ . A similar P212121 with c ¼ 14.14 A arrangement is found in {2DCA-[norbornadiene]}. Continuity of the tunnels along [001] was demonstrated (Suheiro, 1988) by preparation of DCA complexes of polyethers (of composition [–(CH2)m–1–CHR–O–]n, where R ¼ H, m ¼ 2, 3, 4, 6 and R ¼ CH3, m ¼ 2); the polyethers had molecular weights of 3000. X-ray powder diffraction patterns indicated that the complexes had the usual type of orthorhombic structure, although finer distinctions could not, of course, be made. One striking feature of the orthorhombic deoxycholic acid inclusion complexes, especially those with long-chain fatty acids as guests, is that the DCA:guest ratio is always one of small integers; for example 4 : 1 for C3 to C7 fatty acids, 6 : 1 for C8 to C15 and 8 : 1 for C16 and higher. Thus the situation is quite different from that in the urea tunnel inclusion complexes where the ratio is, in general, incommensurate. Earlier workers subjected this point to much experimental testing by analysis of samples prepared from solution and the consistency of the result was puzzling, especially at a time when the structures were not known. Even today the implication is that there must be empty spaces in the tunnels whenever the ratio of the [001] translation period to the length of the guest molecule cannot be expressed as a ratio of small integers; for example, palmitic acid would be expected to occupy three unit cells and a small part of a fourth along [001]. On the basis of their determination of the structure of the palmitic acid complex, Coiro, Giglio, Mazza, Pavel and Pochetti (1982; CHOPAL) suggested that an ethanol molecule is present between the palmitic acid molecules and thus the complex is ternary with composition {8DCA[palmitic acidethanol]}. Although the two types of guest were not found directly by structure analysis because of disorder, some generalisation of the proposal appears permissible because of the reported presence of ethanol in many such complexes (Giacomello and Bianchi, 1943). However, the phase diagrams noted above also indicate (but perhaps not with the necessary precision) compositions such as {8DCA[palmitic


278

Table 6.10. Classification of orthorhombic DCA inclusion complexes in terms of periodicity along c and space group. The composition is given as {xDCA[y(guest)]} Guest ˚) Group IA (P21212, c 7.2 A DCA/guest (x : y) 2:1 (þ)-camphor ()-camphor ferrocene di-t-butylthioketone thiocamphenilone Dibromoethane ˚) Group IB (P212121, c 7.2 A DCA/guest (x : y) 1:1 Acetic acid 2:1 ethylmethylketone diethylketone m-chloroacetophenone (103K) p-diiodobenzene Ethyl acetate(163K) pinacolone phenylacetylene styrene naphthalene 5:3 Acetone (103K) 5:2 acetophenone 8:3 3:1

4:1 8:1:1

p-fluoroacetophenone methylpentylketone phenanthrene 1,2-benzanthracene Propiophenone (103K) p-fluoropropiophenone (103K) p-dimethylamino-azobenezene Palmitic acidethanol

˚) Group IC (P21212, c 14.4 A DCA/guest (x : y) 2:1 cyclohexanone (R)-3-methylcyclohexanone (S)-3-methylcyclohexanone norbornadiene

Torsion Refcode angle

Reference

63.8 63.4 63.4

JSLL82 CCMP95 MEPSH00

CHOLCM HICHIK FEHYAS01 (360K) COXDEY

65.1 (Note 2) 62.0 FIYGAV

62.2 58.5 62.3

DECHAC# DANSOA DANSIU DAJLEF

61.8 61.0 61.7

DCPBID# JIFSOG COFNEQ DOSKEB

PVR84 PVR87 BG62

CdT72 CCMP95 CCMP95 TCP-B85

62.6 62.2

CGPQ72 NNSZ86 CMPGP85 GMS85 FQGP81 FQGP81 DXCHAC P-BCTSLL80 DAJLAB, TCP-B85; BARJOT10 CP-BLL87 BEGHOK10 W-LV87; CP-BLL87 DANSUG P-BTCLL85 DCPHEN CGPQ72 FIQ83 FEHMEK W-LVP-BCMFLL87 FEHMIO W-LVP-BCMFLL87

63.2 62.8

BIVKAS CHOPAL

CGM82 CDG80

65.0 58.3

DANSEQ

P-BTCLL85

DCANBD

T79 T79 DFGMP81

61.6 62.8 63.0 62.8 61.0 62.1

66.2, 66.7


279

Table 6.10. (Continued ) Guest di-t-butyldiperoxycarbonate quadricyclane ferrocene Cis-N-nitroso-2, 6-dimethylpiperidine (130K) N-nitrosopiperidine (130K)

Torsion angle 63.4, 66.1 62.7, 67.2 66.4, 66.8 66.9 61.5

Refcode

Reference

DXCHBC# DANFUT

FLLP-BTZ75 CGMP84

RAZGII (100–294K) JOQJAA

MEPSH00 GMP99

JOQHUS

GMP99

Notes: (1) Limited crystal data have been reported for a number of complexes without full structure determinations; see, for example, Giacomello and Bianchi (1943) and Go and Kratky (1934). (2) This is the only example encountered of a ( )-I conformation for DCA. # no coordinates References: BG62 – Bonamico and Giacomello, 1962; CCMP95 – Candeloro de Sanctis, Coiro, Mazza and Pochetti, 1995; CDG80 – Coiro, D’Andrea and Giglio, 1980; CdT72 – Craven and de Titta, 1972; CGMP84 – Coiro, Giglio, Mazza and Pavel, 1984; CGMPP82 – Coiro, Giglio, Mazza, Pavel and Pochetti, 1982; CGPQ72 – Candeloro de Sanctis, Giglio, Pavel and Quagliata, 1972; CMPGP85 – Coiro, Mazza, Pochetti, Giglio and Pavel, 1985; CPBLL87 – Chang, Popobvitz-Biro, Lahav and Leiserowitz, 1987; DFGMP81 – D’Andrea, Fedeli, Giglio, Mazza and Pavel, 1981; FIQ83 – Ferro, Imperatori and Quagliata, 1983; FLLP-BTZ75 – Friedman, Lahav, Leiserowitz, Popovitz-Biro, Tang and Zaretzkii, 1975. FQGP81 – Ferro, Quagliata, Giglio and Piacente, 1981; GMS85 – Giglio, Mazza and Scaramuzza, 1985; GMP99 – Gdaniec, Milewska and Polonski, 1999; JSLL82 – Jones, Schwarzbaum, Lessinger and Low, 1982; MEPSH00 – Mu¨ller, Edwards, Prout, Simpson and Heyes, 2000; NNSZ86 – Nassimbeni, Niven, Stuart and Zemke, 1986; P-BCTSLL80 – Popovitz-Biro, Chang, Tang, Shochet, Lahav and Leiserowitz, 1980; PRV87 – Padmanabhan, Ramamurthy and Venkatesan, 1987; PVR84 – Padmanabhan, Venkatesan and Ramamurthy, 1984; T79 – Tang, 1979; TCP-B85 – Tang, Chang et al., 1985; W-LVP-BCMFLL87 – Weissinger-Levin, Vaida et al., 1987.

acid]} and here the presence of ethanol cannot be supposed because of the high temperatures involved. It would be worthwhile to carefully compare (preferably at low temperature) the structures of solution-grown and melt-grown single crystals of one of these inclusion complexes. An exception to the orthorhombic structures described above is the monoclinic struc˚, ture found for {2DCA[o-xylene]} (a ¼ 7.238(7), b ¼ 26.171(12), c ¼ 13.510(9) A ˚ 3, Z ¼ 2, P21; WALHUM; ¼ 60.1, 60.3 ). Inspection of ¼ 90.91(9) , V ¼ 2559(3) A the cell dimensions shows at once the resemblance to the orthorhombic structures and this has been confirmed by determination of the crystal structure (Candeloro de Sanctis and Giglio, 1979; Cerrini, Pochetti, Gallese and Possagno, 1993). The 13C CP/MAS NMR spectroscopy of {2DCA[ferrocene]} has been studied over the temperature range 160–350K (Heyes and Dobson, 1990), and extended by additional XRD and NMR studies (Mu¨ller et al., 2000), the results of which are now briefly described.


19.4

2680

19.2

2660

19

Cell volume (cub. A)

Chemical shift (ppm) for C21

280

18.8 18.6 18.4 18.2

2640 2620 2600 2580 2560

18

2540

17.8 90

140

190

240 T (K)

290

340

90

140

190

240 T (K)

290

340

Fig. 6.54. {2DCAferrocene): (a) Plot of chemical shift for the two peaks of C21 versus T. It was not possible to distinguish between one and two peak models in the temperature range from 330 to 360K and so this region has been left blank. (b) Plot of V vs. T; V has been normalized to the hightemperature P22121 cell but volume should be doubled below 305K. There is a change of slope around 305K but no clear volume discontinuity, hence Mu¨ller et al. categorize the phase transformation as second order. (Data from Mu¨ller et al. (2000); Dr C. K. Prout (Oxford) is thanked for helpful correspondence.)

The 13C NMR resonance from C21 (of a methyl group, see Scheme 1) is split into two lines which coalesce at 360K (Fig. 6.54(a)); this is in keeping with there being two ˚ ) and one crystallographically independent DCA molecules at low temperature (a 14 A ˚ at high temperature (a 7 A). Cell dimension–temperature plots show essentially identical behaviour for all three axes (perhaps surprisingly for such an anisotropic structure) and so can be summarized by a volume–T plot (Fig. 6.54(b)). The onset of the change is at ˚ structure) 305K, above which temperature the intensities of h odd reflections (in the 14 A gradually decrease, and then disappear at 360K. Differential scanning calorimetry shows no distinct endo- or exothermic event in the range 293–360K. Thus both these techniques suggest that the phase change is second order in the Ehrenfest (1933) sense, with a gradual progression from an ordered low-temperature phase to a disordered hightemperature phase, with Tc 305K. Much detail has been omitted but two points can be made. Firstly, the order–disorder transition occurs when a degree of motion of one of the components increases significantly; this also occurs in the order–disorder transitions discussed in Chapter 16. Secondly, there are many indications that behavior similar to that of {2DCAferrocene} occurs in many other DCA–guest complexes, the temperaturedependent behaviour of which surely warrant study at a similar level of detail. The results of van der Waals energy calculations on the orthorhombic DCA complexes have been summarised by Giglio (1984); many subtleties of the bilayer shifts can be reproduced despite neglect of host-guest interactions, while disordered guest molecules can be located in the tunnels when these are taken into account. Similar calculations have been made for the hexagonal complexes (see below). The DCA complexes are rather stable against loss of guest – for example, the xylene complex loses xylene in air only at the melting point of the complex (423K).


281

Fig. 6.55. Stereoscopic views of the 2 : 1 DCA-[acetone] (left) and 1 : 1 ACA-[acetone] (right) complexes, looking down [001]. The cell parameters are, respectively, a ¼ 25.809(5), b ¼ ˚ , space group P212121, four molecules of DCA and two molecules of 13.610(2), c ¼ 7.233(1) A ˚ , space group P212121, acetone (disordered) per cell, and a ¼ 24.47, b ¼ 14.26, c ¼ 7.50 A 4 molecules of ACA and of acetone (disordered) per cell. Note that this figure is rotated by 90 compared to Fig. 6.53, and only half of the bilayers are shown above. (Reproduced from Lahav, Leiserowitz, Popovitz-Biro and Tang, 1979.)

Dissolution of a carboxylic acid complex in ethanol and recrystallization will lead to replacement of the acid by ethanol, although a number of recrystallizations may be needed for complete replacement. In similar fashion xylene will replace ethanol and acetic acid xylene; however, these processes do not appear to have been studied quantitatively. Vapor pressure measurements have been made for the complexes of DCA with styrene and naphthalene (Ferro, Quagliata, Giglio and Piacente, 1981) and also with phenanthrene, 1,2-benzanthracene (Ferro, Imperatori and Quagliata, 1983) and 11,12benzofluoranthene (Ferro, Quagliata and Conte, 1983); only the guest vaporizes in the temperature ranges investigated. The enthalpies of formation of the crystalline complexes are given as 15 5, 24 5, 49 8, 22 10 and 5 5 kJ/mol; thus these complexes are enthalpy-stabilized. The enhanced stability of the phenanthrene complex is in accord with the ordering of the guest molecules in its crystal structure. The structure of only one complex of ACA appears to have been reported, that with acetone (composition nominally 1 : 1) (Popovitz-Biro, Tang et al., 1985; APCHAC10; ¼ 62.2 ). The similarity to the Group IB structures is shown in Fig. 6.55.

6.3.1.2

The complexes of cholic acid

Many guests form complexes with cholic acid (but not fatty acids, according to Rheinboldt and Lauber (1929)). A short list has been given by Nakano, Sada and Miyata (1994) and a more extensive list by Nakano, Sada, Kurozumi and Miyata, (2001). Among the guests are: 1:1. benzene, toluene, ethylbenzene, 1,5-hexadiene, 2,5-norbornadiene, chlorobenzene;

282


2 : 1. 1-methylnaphthalene, myrcene, ethylcyclohexane, dibutyl ether, -ionone, methyl decanoate. 3 : 2. tetralin, cyclohexane. The crystal structures of many of these complexes have been reported (Table 6.11). Neat cholic acid has a structure different from those of the complexes so that these represent separate phases in the cholic acid–guest phase diagrams (which do not appear to have been reported). The overall situation, based on 23 new crystal structures and 28 from the literature, has been neatly summarized by Nakano et al. (2001): ‘‘The facially amphiphilic molecular structure of cholic acid gives rise to the bilayer structure by means of van der Waals association of lipophilic faces and hydrogen bonding between hydrophilic faces.’’ Although there are considerable resemblance to the DCA complexes, there are also marked differences. Nakano et al. (2001) distinguish four groups that result from different conformations of the cholic acid side chain (values for ranging around ( )180 (trans), and (þ)60 ((þ)-gauche)6 and different interdigitations of the methyl groups ( and ) in ˚ gives the the lipophilic faces; sliding of the upper layer of the -type stacking by 4.5 A -type stacking). Four types, (-trans, -(þ)-gauche, -trans and -(þ)-gauche) result from the combination of the two factors, and these are shown by subtle differences among the cell dimensions of the types of complex despite their overall resemblances (Table 6.11). All these crystals have the same basic arrangement of cholic acid bilayers but the stacking modes, and the details of the inclusion of the guests, are somewhat different; examples of each type are shown in Fig. 6.56 and more detail is given for 1 : 1 {cholic acid-[acetophenone]} (Miki, Masui et al., 1988; VABSOG) in Fig. 6.57. The cholic acid molecules are hydrogen bonded to give hydrophilic bilayers (interacting by dispersion forces) in the (100) planes; these layers are mutually arranged so as to provide tunnels along [010] in which the acetophenone molecules are accommodated. The headto-tail bilayers have a parallel arrangement in the DCA complexes, but are antiparallel in the cholic acid complexes. Some 120 inclusion complexes of cholanamide with various guests have been reported, while some 50 compounds did not form inclusion complexes (Sada, Kondo, Uchioda et al., 1998). For reasons of space we do not discuss these in detail but some useful comparisons can be made with those of cholic acid. Notice, first, that the three cholanamide complexes in Table 6.11 have cell dimensions very close to those of the monoclinic -trans group, particularly for {cholic acid[-(S)-valerolactone]}. Thus the ‘‘assembly mode of the hosts’’ (a phrase of Sada et al. (1993)) is the same in these two groups of complexes. But there is an important difference – the additional hydrogen of cholanamide (–NH2 compared to –OH) allows for hydrogen bonding of guests to the host framework. We note here a number of other complexes of DCA and analogs, which can hardly be called inclusion complexes; however, it is more convenient to place them here than ˚ , c 49 A ˚, elsewhere. The structures of two isomorphous tetragonal complexes (a 14 A space group P41212, Z ¼ 8) have been reported; the compositions are 2 : 3 DCA:H2O (DOCHHY; Coiro, D’Andrea and Giglio, 1979; Tang et al., 1979) and 2 : 1 : 1 DCA : ethanol : H2O (DCAETO; Candeloro de Sanctis, Coiro, Giglio, Pagliucca, Pavel 6

The ( )-gauche conformation is found but is rare.


283

(a)

(b)

(c)

(d)

Fig. 6.56. The four structural patterns found for the inclusion complexes of cholic acid: (a) -(þ)gauche (the diagram is specifically for benzene as the guest); (b) -trans (phenetol); (c) -gauche (ethynylbenzene); (d) -trans(benzyl alcohol). Compare Table 6.11. (Adapted from Nakano et al., 2001.)


284

CA hydroxyl head

CA carboxyl tail

hydrophobic layer guest tunnel hydrophilic layer

guest tunnel

z

x

Fig. 6.57. Crystal structure of 1 : 1 {cholic acid-[acetophenone]} in projection down [010]. Unit cell parameters are in Table 6.11. The hydrogen bonding network between the oxygen atoms is shown by broken lines. The acetophenone oxygens are not hydrogen bonded. The head-to-tail directions of cholic acid molecules run in opposite directions in the two partners of the bilayer; this is the converse of the situation in the DCA inclusion complexes (cf. Fig. 6.55). The torsion angle about the bond marked as is 58.4 (full definition in Scheme 6.1). (Adapted from Miki, Masui, et al., 1988.)

and Quagliata, 1978) and the DCA arrangements are essentially identical, with only small differences in the dispositions of the guest molecules. As Giglio (1984) points out, there are relations between unit cell dimensions of orthorhombic and tetragonal crystals: atetr borth 2 corth ;

ctetr 2 aorth :

These stem mainly from the size of the DCA molecules and do not reflect resemblances in ˚ ) lies parallel to borth, atetr (and crystal structure. Thus the DCA molecule (length 14 A btetr) and determines the lengths of these axes; the separation between two rows of the ˚ , accounting for the lengths of corth and of 1/2(atetr); ctetr is same monolayer is about 7 A about four times the approximate distance between two bilayers in both the tetragonal and orthorhombic crystals. The conformation of the DCA molecules is similar to that in the orthorhombic crystals. However, in the tetragonal crystals hydrophilic sides of the DCA molecules face one another and are hydrogen bonded to form tunnel walls enclosing water and/or ethanol molecules, which also participate in the hydrogen bonding.


285

Table 6.11. Classification of CA inclusion complexes (composition {CA. [guest]}) on the basis of ˚ , deg.) and unit cell volume (A ˚ 3). The CA : guest ratio is 1 : 1 except where unit cell dimensions (A noted otherwise. The 1 : 1 complexes are generally monoclinic, space group P21, Z ¼ 2, except for a few triclinic complexes, triclinic and monoclinic structures being related. The 2 : 3 complex has space group P21, Z ¼ 4. There are also a few related orthorhombic complexes. Three cholanamide complexes have been included in the Table for comparison Composition -trans; 180 Orthorhombic No guest Acrylonitrile Polymorph I Triclinic Methyl acetate

Refcode / reference

a/

b/

c/

Space group

V

Torsion angle

JEYDEW; MKSCTM90 PEMZAI01; NSM96

16.477(4)

8.394(3)

16.993(3)

2350

167.3

16.882

8.497

17.019

P212121, Z¼4 P212121, Z¼4

2528

171.3

12.223(2) 90.18(1) 12.279 90.41 12.53 90.69 12.655 91.93 12.289 90.39 12.474 91.09

8.189(1) 105.72(2) 8.245 105.69 8.28 107.2 8.354 106.02 8.238 105.83 8.252 106.31

14.204(2) 94.03(1) 14.157 94.23 14.16 94.9 14.125 94.68 14.246 94.97 14.214 94.43

P1, Z¼2 P1, Z¼2 P1, Z¼2 P1, Z¼2 P1, Z¼2 P1, Z¼2

1364

159.3 166.6 167.9 158.4 157.9 165.8 174.3 155.8 157.9 165.8 155.6 171.8

12.183

7.878 104.16 8.003(2) 104.76(2) 8.117 103.81 7.909 106.03 8.609 105.18 8.610(5) 105.21(2)

14.300

P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2

1331

162.3

1414

169.5

1407

166.2

1394

175.0

1459

159.7

1449

154.6

ethyl acetate (polymorph I) methylacrylonitrile

PIWKIP; CNS94a PIWKOV02; NKSM01 NSKM01

Acetone (CA : acetone 1 : 1.5) 0.5(Methyl acetate) 0.5(isopropyl acetate) 0.5(acetone) þ 0.5 (1,2-dichloro-benzene)

ZEJDUN; CNS94b NOJPIL; S96 NOJPOR; S96

Monoclinic 105 Acrylonitrile Polymorph II -(S)-valerolactone; diethyl ketone N-nitroso-piperidine (100K) benzyl alcohol 4-fluorobenzyl alcohol

PEMZAI02; NSM96 JOLFIZ; MKSTM91 ZEJFEY; CNS94b JOQHEC; GMP99 GUNPIO; NSKM01 YUNZAI; STS95

-(þ)-gauche; 60 ; monoclinic, Ethyl acetate PIWKOV; Polymorph II CNS94a PIWKOV01; NKSM01 Benzene WEYNUJ; NSM94 Fluorobenzene GUNLOQ; NSKM01

13.010(3) 12.787 13.258 12.636 12.632(5) 115 13.668(3)

13.627(4) 13.57

14.049(4) 13.960 13.818 13.900 13.806(3)

1375 1396 1428 1382 1399

7.824(4) 113.53(1)

14.095(2)

P21, Z¼2

1385

65.2

8.038(9) 114.25(2) 8.06 114.4

14.076(4)

P21, Z¼2 P21, Z¼2

1406

62.7

1405

62

14.10


286

Table 6.11. (Continued ) Composition

Refcode / reference

a/

b/

c/

Space group

V

Torsion angle

Chlorobenzene

GUNLUW; NSKM01 GUNKOP; NSKM01 GUNMAD; NSKM01 GUNMEH; NSKM01 GUNLAC; NSKM01 LAFCAW; CNS93 LAFCEA; CNS93; SS94b VABSOG; MMK88 GUNMUX; KSKM01 BIFQAI; S97; NSKMOI BIFQEM; S97

13.66

14.01

60

1421

61

1413

65

1423

61

1431

60

1411

60.3

1418

60.0

1447

58.4

1416

59.0

1423

58.2

1460

62.6

13.709

P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2

1411

ECARAB; BBFFP01 MIWTIV; PSGNH01 RABKEG; STS95c TEMWOX; S96b ZUZDON; STS95c ZUZDUI; STS95a ZUZFAB; STS95c YUNYOV; STS95b YUNYOV01; STS95 PIWKUB; CNS94a GUNNIM; NSKM01

8.10 114.6 8.083 114.05 8.10 114.6 8.10 114.2 8.15 114.6 8.049(1) 115.20(2) 8.106(3) 113.52(1) 8.093(1) 113.69(1) 8.111(1) 113.65(1) 8.133 114.12 8.266 114.61 8.150 114.10 8.019 112.81 8.154 113.27 8.078 114.42 8.147 113.71 8.109 115.54 8.135 115.94 8.549 113.02 8.510(4) 113.15(2) 7.969(1) 113.53(1) 8.085 114.75

1430

61.8

1386

62.2

1456

59.9

1416

61.2

1425

57.8

1416

61.8

1429

62.5

1466

62.6

1467

58.3

1411

65.7

1417

59.2

P21, Z¼2 P21, Z¼2

1455

164

1441

167

toluene bromobenzene iodobenzene styrene Aniline Nitrobenzene Acetophenone Benzaldehyde benzonitrile p-nitrotoluene Bicyclo(3.2.0)hept2-en-6-one 0.5(benzil) 4-fluoro-acetophenone p-toluidine 2-fluoroaniline 4-fluoroaniline 3,4-difluoroaniline 2-fluorobenzyl alcohol (note 1) 2-fluorobenzyl alcohol (note 1) Ethyl propionate anisole

13.740 13.69 13.68 13.57 13.742(2) 13.579(2) 13.719(2) 13.565(2) 13.642 13.495

13.392 13.540 13.577 13.613 13.834 13.928 13.351 13.421(5) 13.571(1) 13.566

˚ -trans; 180 ; monoclinic, c 16 A phenol GUNPEP; 12.07 NSKM01 N-methylaniline GUNPOU; 12.00 NSKM01

7.92 111.9 7.96 111.6

14.01 14.01 14.09 14.24 14.095(2) 14.048(1) 14.229(2) 14.055(6) 14.055 14.398 14.024 14.002 14.353 14.182 14.034 13.993 14.029 13.966 13.978(3) 14.237(1) 14.225

16.39 16.22


287


Refcode / reference

a/

b/

c/

Space group

V

Torsion angle

Benzyl chloride

GUNMIL; NSKM01 GUNMOR; NSKM01 GUNKUV; NSKM01 GUNNUY; NSKM01 GUNLIK; NSKM01 POZPUP; GP98 POZPUP01; S97a POZPUP02; NSKM01 GUNNEI; NSKM01 GUNNAE; NSKM01 GUNNOS; NSKM01 GUNPUA; NSKM01 RABJUV; STS95 RABKAC; STS95 JOQHIG; GMP99

12.34

7.82 111.7 7.83 111.5 7.83 111.8 7.94 109.9 7.881 112.06 7.940 109.20

16.24

P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2

1456

174

1459

173.0

1469

174.8

1468

170

1483

173.4

1498

170

P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2

1498

170.0

1478

179.5

1475

171

1481

171

1492

167.6

1487

175.6

1447

167.8

Benzyl bromide ethylbenzene Benzylmethyl ether Allyl benzene n-propiophenone

Benzyl formate Phenyl acetate phenetol N-ethylaniline 3-chloroaceto-phenone 3-fluoroaceto-phenone N-nitroso-4methylpiperidine (100K) cis-N-nitroso-2, 6-dimethylpiperidine (300K)

JOQHOM; GMP99

12.29 12.41 12.11 12.434 12.300

12.085 12.256 12.11 12.14 12.580 12.794 12.353

12.754

16.30 16.28 16.24 16.328 16.240

7.961 109.70 7.897 109.67 7.97 108.5 7.95 108.2 7.984 112.23 7.813 113.13 7.675 111.09

16.208

7.881 111.97

11.355

P21, Z¼2

1525

166.8

16.798(4)

P21, Z¼2 P21, Z¼2

1414

65.2

1423

62.8

15.46

P21, Z¼2

1417

69.8

8.056

P212121 Z¼4 P21, Z¼4 P21, Z¼2

2934

59.7

2661

52.5 60.3 177.3

˚ ; 118 . -gauche-A; 60 ; monoclinic, c 16 A n-Propyl acetate PIWLAI; 12.101(2) 7.884(3) CNS94a 118.04(2) i-Propyl acetate PIWLEM; 12.141(2) 7.979(3) CNS94a 117.76(2) ˚ ; 106 . -gauche-B; 60 ; monoclinic, c 15.5 A ethynylbenzene GUNLEG; 12.264 7.803 NSKM01 106.80 Miscellaneous, some with related unit cells m-chloroaniline QOQFAD; 14.624 24.903 YC01 0.5(o-xylene) YOYFIB; 13.827 25.612 NSM95 90.99 Acetone3H2O LEVTAH; 13.111 7.759 CNS94c 105.70

16.212 16.12 16.14 16.045 16.172 16.359

16.604(3)

7.515 14.893

1459


288


Refcode / reference

a/

b/

c/

Space group

V

Torsion angle

Bis(acetic acid)

ZUPKIE; NSM96 ZZZKJE; NH65 ZUPKEA; NSM96 YUNYUB; STS95 ZEJFAV; CNS94b LISYUH; D00 YAZTIC; SS94b

13.353

8.189 112.17 7.849 113.53 15.363

13.993

1417

176.9 166.1 No coordinates 176.9

8.485(4) 101.30(2) 8.171 105.40 7.620 97.53 7.847 106.58

28.512(5)

P21, Z¼2 P21, Z¼2 P212121 Z¼4 P21, Z¼4 P21, Z¼4 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2 P21, Z¼2

4H2O Acetic acid 1 : 1.5 3-fluorobenzyl alcohol (note 2) methyl ethyl ketone 4-aminopyridine 3.18H2O 2(3-fluoroaniline)

1 : 1 {Cholanamideguest} complexes; Propan-2-ol LILGES01; SKU98 Rac-butan-2-ol TEHPUR; BKSMM96 (R)-butan-2-ol TEHPOL; BKSMM96 (S)-butan-2-ol TEHPIF; BKSMM96 1,4-dioxane LAHGAC01; SKU98

13.696 13.831 13.459(5) 12.597 10.975 14.326

monoclinic; trans 13.103 7.799 104.66 13.251(2) 7.869(1) 104.82(1) 13.286(1) 7.853(1) 105.12(1) 13.226(1) 7.871(1) 104.7(1) 13.170 7.868 104.97

14.043 12.400

27.954 19.749 16.125

14.092 14.045(1) 14.075(1) 14.028(1) 14.098

1384 2634 3193

1637

64.0 167.3 156.3 165.3 156.9

1737

176.3

1393

156.6

1416

159.1

1418

162.4

1412

159.8

1405

162.9

2780

Notes: (1) Crystals of YUNYOV01 were grown by standard crystallization from solvent whereas those of YUNYOV were prepared by an absorption method. The cell dimensions are very similar but the CA conformations differ. (2) YUNYUB differs from all the other CA inclusion complexes in having different conformations for the two crystallographically independent CA molecules. References: BBFFP01 – Bertolasi et al., 2001; BKSMM96 – Briozzo, Kondo, Sada, Miyata and Miki, 1996; CNS93 – Caira, Nassimbeni and Scott, 1993; CNS94b – Caira, Nassimbeni and Scott, 1994b; CNS94b – Caira, Nassimbeni and Scott, 1994b; CNS94b – Caira, Nassimbeni and Scott, 1994c; D00 – Dastidar, 2000; GMP99 – Gdaniec, Milewska and Polonski, 1999; GP98 – Gdaniec and Polonski, 1998; MKSCTM90 – Miki, Kasai, Shibakami, Chirachanchai, Takemoto and Miyata, 1990; MKSTM91 – Miki, Kasai, Shibakami, Takemoto and Miyata, 1991; this is a rare example of enantiomer discrimination; MMK88 – Miki, Matsui et al., 1988; NKSM01 – Nakano, Katsuta, Sada and Miyata, 2001; NSKM01 – Nakano, Sada, Kurozumi and Miyata, 2001; NSM94 – Nakano, Sada and Miyata (1994); NSM96 – Nakano, Sada and Miyata (1996); PSGNH01 – Polonski et al., 2001; S95 – Scott, 1995; S96a, b – Scott, 1996a, b; S97a, b – Scott, 1997a, b; SKMTM93 – Sada, Kondo, Miyata, Tamada and Miki, 1993 SKU98 – Sada, Kondo, Miyata and Miki, 1994; SS94b – Shibakami and Sekiya, 1994b; STS95c – Shibakami, Tamura and Sekiya, 1995; TP-BLL79 – Tang, Popovitz-Biro, Lahav and Leiserowitz, 1979; YC01 – Yoswathananont et al., 2001.


289

˚ , space The structures of two isomorphous hexagonal complexes (a 15.1, c 18.7 A group P65, Z ¼ 6) have been reported; the compositions are DCA: [ethanol:H2O] ¼ 3 : 2 : 1 (H2O (Candeloro de Sanctis, Giglio, Petri and Quagliata, 1979) and DCA:[dimethyl sulphoxide:H2O] ¼ 2 : 1 : 1 (Giglio and Quagliata, 1975). In these two complexes the DCA ˚ around the 65 axis of molecules are hydrogen bonded to form a spiral of diameter 4 A the crystal; the hydrogen bonding is quite different from that in the orthorhombic crystals and O(12) is not involved in the bonding between host molecules. The guest ethanol and water molecules are contained within the interior tunnels of the spirals and are hydrogen bonded to the host molecules. The spirals themselves are arranged in hexagonally closepacked fashion in the crystal, with the hydrophobic regions of the molecules in contact; there are only van der Waals forces between the spirals. One consequence of the differences between the hydrogen bonding schemes in the orthorhombic and hexagonal crystals is that the DCA molecules take up somewhat different conformations in these two types of complex. Calculations (Giglio and Quagliata, 1975) suggest that it is the energy gains from the various possible hydrogen bonding schemes which provide the driving force for the various arrangements and that these offset the energy losses expended in altering the DCA conformations. The interior of the tunnels in the hexagonal crystals is hydrophilic; in the {2DCA[ethanolH2O]} complex there appear to be short ethanol–water–ethanol chains which give additional stabilization to the complex by forming hydrogen bonds to O(12). The dimensions of the tunnels are more strictly defined in the hexagonal than in the orthorhombic complexes. The range of guests accommodated in the hexagonal is much more limited than that found in the orthorhombic complexes because hydrogen bonding is both stronger than van der Waals bonding and has more stringent directional requirements. Helical arrangements of hydrogen bonds rather similar to those in the hexagonal DCA complexes are found in the so-called ‘‘low temperature polymorph’’ of CDCA (see

Fig. 6.58. CDCA-solvent complex, viewed down the c axis; the isolated circles indicate the guest ˚ , Z ¼ 6, space group P65. There are region. The crystals are hexagonal, a ¼ 22.25, c ¼ 10.26 A twofold screw axes at the centers of the cell edges and of the long diagonal. (Reproduced from Rizkallah, Harding, Lindley, Aigner and Bauer, 1991.)

290


formula 6.12, note 2), which is actually a tunnel inclusion complex with a variable guest content (Fig. 6.56; Rizkallah, Harding, Lindley, Aigner and Bauer, 1990; the bromobenzene ‘clathrate’ is listed as VEFYAG). The structure determination is noteworthy because it was carried out on a crystal of dimensions 500 40 60 , using synchrotron ˚ ). The tunnels, which have hydrophobic internal surfaces, are wide radiation ( ¼ 0.895 A ˚. enough to accommodate molecules with van der Waals diameters up to 8 A The physiological role of DCA-fatty acid complexes is not certain but they may have a part in excretory processes (Smith, 1973). Certain human intestinal stones have been found to consist chiefly of DCA-fatty acid complexes (8 : 1 ratio) (Fowweather, 1949); DCA also plays an important role in the formation and solubilization of gallstones (Dirscherl and Gerhards, 1961). There is spectroscopic evidence for interaction between DCA and organic molecules also in solution; solubilization of hydrocarbons in alkaline aqueous solutions by addition of DCA may be due to micelle formation in which the hydrogen bonded bilayers found in the inclusion complexes could play some role. A detailed picture has not yet been given. The 1 : 1 complexes of cholic acid with methanol, ethanol, 1-propanol and three fluoroethanols are isostructural (Table 6.12). There are no bilayers and the guest molecules are contained within cavities interconnected by narrow tunnels; there is an intricate arrangement of five hydrogen bonds between three cholic acid molecules and the guest molecules. These complexes are not inclusion complexes in the present sense. Jones and Nassimbeni mention that 1-butanol and 2-propanol form complexes with cholic acid (space group P6522) but details were not given; Nakano, Sada and Miyata (1994) consider 1-butanol and 2-propanol to be non-complexing solvents, and the contradiction remains to be resolved. There are hydrates of cholic acid, a monohydrate in space group P21 (Z ¼ 2; Lessinger, 1982; Shibakami, Tamura and Sekiya, 1995; BUGJES), a hydrate in P6522 (Z ¼ 12; Lessinger and Low, 1993; Shibakami, Tamura and Sekiya, 1995b; DIPFAJ10; Lessinger gives this as a hemihydrate and Shibakami as a monohydrate) and a tetrahydrate (ZZZKIE) included in Table 6.12. There is also a methanol monohydrate (NUMVEW) and an acetic acid complex (ZUPKEA). Compositions similar to those found among the orthorhombic DCA inclusion complexes of long chain guests are indicated by the melting point diagrams reported for some Table 6.12. Crystal data for 1 : 1 Cholic acid complexes with some alcohols and fluoroethanols. The crystals are isomorphous with spce group P212121 and Z ¼ 4. The CA molecules all have the trans conformation Composition

Refcode/Reference

a

b

c

V

Torsion angle

Methanol Ethanol 1-propanol 2-fluoroethanol 2,2-difluoro-ethanol 2,2,2-trifluoro-ethanol

JEGMIR; JS72, JN90 CHOLET; JN90, JS72 JEGMUD; JN90 ZIPYUS; SS94a ZIPZED; SS94a ZIPZIH; SS94a

14.560 14.653 14.951 14.585 14.461 14.562

15.198 15.045 15.026 15.319 15.314 15.41

11.625 11.739 11.864 11.792 11.848 11.72

2572 2588 1665 2603 2623 2645

166.2 169.9 167.8 170.5 170.2 171.5

References: JN90 – Jones and Nassimbeni, 1990; JS72 – Johnson and Schaefer, 1972; SS94a – Shibakami and Sekiya, 1994a.


291

Table 6.13. Compositions (from phase diagrams) of some steroid complexes of unknown structure Host dehydro-epi-androsterone dehydro-epi-androsterone dehydro-epi-androsterone dehydro-epi-androsterone dehydro-epi-androsterone androsterone cholesterin

acetate acetate acetate acetate

Guest

Host : guest molecular ratio

stearic acid stearic acid lauric acid benzoic acid phenylurethane stearic acid stearic acid

8:1 16 : 1 and 8 : 1 6:1 4:1 6:1 4:1 4:1

steroid systems (Dirscherl and Gerhards, 1961). Some typical compositions are given in Table 6.13. Structural evidence is lacking and analogies should be drawn with caution because the forms of the phase diagrams are quite different from those found for DCAand ACA-fatty acid systems. There are tunnels of rectangular cross-section in the 5 : 1 tunnel inclusion complex of zeyloxanthonone (C28H34O5) with palmitic acid (Dean, Herbstein, Kapon and Reisner, 1991; KIFWUR). The crystals are orthorhombic with a ¼ 29.095(4), b ¼ 9.508(2), ˚ , space group Pbca, Z ¼ 8. The molecules are linked by intermolecular c ¼ 20.718(2) A ˚ ) in one dircarbonyl . . . hydroxyl hydrogen bonds (d(O(5)–H . . . O(3) ¼ C) ¼ 2.70(2) A ection, and dispersion forces in the other two. Thus these complexes are classified as intermediate between the urea and thiourea complexes (all hydrogen bonded) and those of DCA and perhydrotriphenylene.

OH

O O

A HO

B

C

O

Zeyloxanthonone: hydroxyl O(5) is at the left of the diagram and carbonyl O(3) at the right

6.3.2 Substituted spirocyclophosphazenes as hosts The substituted spirocyclophosphazenes 6.14–6.16 (but not 6.17) form tunnel inclusion complexes with a variety of guests (Allcock, Allen, Bissell, Smeltz and Teeter, 1976), as does 6.18 (Alcock et al., 2000). The subject has been reviewed (Allcock, 1978, 1984; Kobayashi, Isoda and Kubono, 1996). The crystal structures of neat 6.14–6.16 have been reported (Table 6.14); because of the spiro nature of the arrangements about the P atoms of the central ring, the substituent rings are approximately perpendicular to the planes of the phosphazene rings and there is considerable flexibility about the linkages


292

120°

140° 120°

60°

110°

Benzene tunnel inclusion complex

Guest free

p-Xylene clathrate

Fig. 6.59. The host molecule 6.15 in different environments viewed normal to the plane of the phosphazene ring. The large differences in molecular shape are quite remarkable. (Reproduced from Kobayashi, Isoda and Kubono, 1996.)

through oxygen. An example of this structural feature is shown in Fig. 6.59, where the host 6.15 in its neat crystals and in two complexes is viewed normal to the plane of the phosphazene ring. Host 6.14 forms hexagonal inclusion complexes with tetralin, cumene, norbornadiene, iso-octene, trans-decalin, cyclohexane, n-heptane, CS2, styrene, ethyl acetate, CCl4, ethanol and various substituted benzenes (Allcock and Siegel, 1964). The neat crystals of 6.14 are monoclinic (Allcock, Levin and Whittle, 1986); thus the inclusion complexes are separate phases in the host-guest phase diagrams. An early report of polymorphism does not appear to have been substantiated. The amount of guest retained at room temperature depends on its nature, e.g. the ideal molar ratio 6.14/guest is 0.5 (based on the crystal structure noted below) while values of 0.59 and 0.11 were obtained for o-xylene and ethanol respectively. The guests are disordered in the tunnels. More recent results show that 6.14 forms the same sort of hexagonal inclusion complexes with substituted ethylenes (eight examples ranging from trans-stilbene to trans-trans-trans-1,6-diphenyl-1,3,5-hexatriene) and two quasi-linear tetrathiophenes (Sozzani et al., 2004a) and also with polyethylene, polyethylene oxide and polytetrahydrofuran (Sozzani et al., 2004b). The molecular complexes have melting points some 150–200K higher than those of the guests themselves. Advanced NMR methods were used to show that the enhanced stability of the complexes was due to the cooperative effect of multiple (although weak) CH . . . interactions.

O

O P

N O

P O

N N

P

O

O

6.14: tris(o-phenylenedioxy)cyclotriphosphazene


293

Host 6.15 forms 1 : 3 inclusion complexes with benzene and 1 : 1 complexes with p-xylene and p-chlorotoluene, while 6.16 gives inclusion complexes when recrystallized from toluene, p-xylene, cycloalkenes, alkenes or alcohols (Allcock and Walsh, 1971; Allcock, Stein and Bissell, 1974); only the structure of the p-xylene complex appears to have been reported. 6.18 forms inclusion complexes with o-xylene, 1,2-dichlorobenzene, p-xylene, benzene, tetrahydrofuran and cyclohexane; the first two are isomorphous but the other structures all differ (Alcock et al., 2000). The inclusion complexes can be prepared by recrystallization from appropriate organic solvents or by a spontaneous process which occurs when the neat solid is brought into contact with the liquid or vapour of many organic compounds (Allcock and Siegel, 1964). More recently, amorphous thin films of host exposed to vapours of guest have been employed, followed by examination by electron microscopy (Kobayashi, Isoda and Kubono, 1996). A somewhat similar process of complex formation appears to occur with potassium benzene sulphonate and organic guests (Barrer, Drake and Whittam, 1953).

O O

O P

N O

P

O

N O

N

N

P

P

O

N

P

O P O

O

N

O

O

6.15: tris(2,3-naphthalenedioxy)cyclotriphosphazene

6.16: tris(1,8-naphthalenedioxy)cyclotriphosphazene

Crystal data are summarized in Table 6.14; one notes a wide range of structural possibilities. The known tunnel inclusion complexes of 6.14 are all isomorphous, with the same space group (P63/m) and similar cell dimensions. A portion of a representative crystal structure, that of {6.14[0.5C 6H6]}, is shown in Fig. 6.60 (a). Although water has very different chemical properties from those of the organic guests, it nevertheless fits (perhaps surprisingly) into the same structural pattern. 6.15 forms a hexagonal complex with benzene, with a different composition (1 : 3) from the hexagonal 6.14 complexes, a similar value of [001] but a different value of [100]. These are all tunnel inclusion complexes, as is triclinic {6.16[p-xylene]}. The general disposition of the three host molecules along the tunnel axes is rather similar in the three types of complex but the host molecules are differently interleaved and the tunnel cross-sections are appreciably different. In particular, the tunnels in the complexes of 6.14 and 6.15 have similar shapes but ˚ respectively) while those in the complexes different sizes (diameters of 5 and 9 to 10 A ˚ ). of 6.16 have both different sizes and shapes (the cross-section varies from 5.2 to 7.0 A Many of the physical properties can be understood in terms of these structural features.


294

˚ , deg. A ˚ 3) for the tunnel inclusion complexes of some Table 6.14. Room temperature crystal data (A spirocyclophosphazenes. Ideal compositions observed so far are {6.14[0.5(guest)]}, {6.15[3(guest)]} and {6.15[guest]}, and {6.16[0.5(guest)]} Host

Guest

Refcode; reference

6.14

Monoclinic, guest-free Water

p-chlorotoluene C6F6

DOFSUM; ALW86 DOFTAT; ALW86 PXZBEN; AABST76 PHOPNB; AABST76 PXZOXY; AABST76 ZZZADM; AABST76 ZZZADJ; AABST76 AFL82 AFL82 HAVHAN; KAIK94 NDXPZB10; AS74 JULJAB; KAIKT93 LEPSAA; KATIK94 KKIK93

6.16

p-xylene

ASB74

1 : 0.5

6.18

1,2-dichlorobenzene p-xylene

AP00

1:2

AP00

1:1

Benzene Bromobenzene o-xylenex p-xylene Mesitylene

6.15

Styrene 4-bromostyrene Guest-free Benzene# p-xylene

Host/ guest ratio

T(K)*

1:2

a/

b/

c/

Unit cell volume

Space group; Z

25.086

25.913

3821

P21/n; 8

11.606

5.911 95.97 –

10.087

1177

P63/m; 2

1:x

6.16 but it is thought that guest-guest interactions probably lower the mobility in the 6.15 complex. The high stability of {6.16[0.5 (p-xylene)]} was attributed to the confining effects of constrictions in the channels, which convert the tunnels almost into cavities. For a particular host the motion of the guest becomes more restricted as the guest molecule increases in size. It seems probable that there will be phase changes on cooling, but this issue does not seem to have been explored. 6.15 also forms clathrate (cage) complexes with p-xylene and p-chlorotoluene; these are noted here for convenience. Despite the geometrical similarity of the guests (and the well-known structural interchangeability of chlorine and methyl), the complexes are not isomorphous, although there are obvious crystallographic resemblances. In the crystal structure of 6.17 the biphenylenedioxy ring systems are inclined at angles of 50 to the mean plane of the cyclophosphazene ring system (Allcock, Stein and Stanko, 1971), while in the other hosts the substituent ring systems are perpendicular, or nearly so, to the mean cyclophosphazene ring plane. It has been suggested that 6.17 cannot

I

II

0

5

III

Fig. 6.60. (a) Stereodiagram of the tunnel formed by the host molecules in the benzene tunnel inclusion complex of 6.14, viewed down [0001]. (b) Comparison of the tunnel cross-sections in the inclusion complexes of 6.14 (I), 6.15 (II) and 6.16 (III). (Both diagrams reproduced from Allcock, Allen, Bissell, Smeltz and Teeter, 1976.)

296


form tunnel inclusion complexes because it lacks the ‘‘paddle wheel’’ structure required for the formation of tunnel walls.

O

O P

N O

N

P O

P N

O

O

6.17: tris(2,2'-biphenylenedioxy)cyclotriphosphazene

Thermodynamic parameters for formation of inclusion complexes of a number of cyclophosphazenes have been determined (Table 6.15) from measurements of vapor pressures according to the following decomposition reaction: fH ½nG gs , Hs þ nGg where H is the host, G is guest and the s, g subscripts refer to solid and vapor states. If one takes into account the enthalpy of sublimation of benzene (44.0 kJ/mol), then the enthalpy of formation of {6.15[2(benzene)]} according to the hypothetical reaction Hs þ nGs , fH ½nG gs is 30.8 (4.2) kJ/mol (at an unspecified temperature). However, there is a problem here because the 6.15 to benzene ratio is given as 1 : 3 by Allcock and Stein (1974) and 1 : 2 by Whitaker, Simon and Victor (1971), and the measurements cannot be properly interpreted until the composition is established. The structure of the 1 : 3 complex of hexaphenylcyclotriphosphazene (Ph2PN)3 with 1,1,2,2-tetrachloroethane is not known. Table 6.15. Values of thermodynamic parameters (kJ/mol; J/mol K) at 298K for the decomposition reaction given above Complex

G

H

S

{6.15[2(benzene)]} (WSV71) {(Ph2PN)3[3(1,1,2,2-tetrachloroethane)]} (WBFGS68) {2,2 0 -dichloro-4,4,-6,6-tetraphenylcyclo-triphosphazene CH3CN} (WG69)

36(4) 46.9(0.8) 7.1(0.2)

119(4) 109(8) 49(2)

279(11) 209(21) 139(6)

References: WBFGS68 – Whitaker, Barreiro, Furman, Guida and Stallings, 1968; WG69 – Whitaker and Guida, 1969; WSV7 – Whitaker, Simon and Victor, 1971.


O

297

O P N

N O

P

N

O

P

O

O

6.18

6.3.3 Tritriptycene – a C62H38 hydrocarbon of D3h symmetry with three U-shaped bays Tritriptycene (6.19) has been synthesized and the crystal structure of the 1 : 1 acetone tunnel inclusion complex determined (Bashir-Hashemi, Hart and Ward, 1986; FATSOI). The molecule has a paddle-wheel shape (Fig. 6.61) not too different from those of the cyclophosphazene derivatives discussed in the preceding section, but is much more rigid. The net distance (i.e. allowing for the van der Waals diameters) across the U-shaped bays ˚ . The tunnel structure is shown clearly in Fig. 6.62, the acetone guests are is 5.5 A orientationally disordered and each guest site is only 1/3 occupied. It was suggested that tritriptycene forms a 1 : 3 complex with water, but details were not given. The occurrence of a family of isostructural complexes seems not unlikely.

Fig. 6.61. Stereoview of tritriptycene, with 30% probability ellipsoids. (Reproduced from BashirHashemi, Hart and Ward, 1986.)


298

A

O

B

A O C

B

B

Fig. 6.62. Stereoview of four unit cells of {tritriptycene[acetone]}, with 20% probability ellipsoids, ˚ , Z ¼ 2, P63/m. viewed down [0001]. The crystal data are a ¼ 16.646(3), c ¼ 11.553(4) A (Reproduced from Bashir-Hashemi, Hart and Ward, 1986.)

6.3.4

Trans-anti-trans-anti-trans-perhydrotriphenylene as host

Perhydrotriphenylene (C18H30) has ten stereoisomers, of which six are from three enantiomeric pairs and four are meso forms (Farina and Audisio, 1970a). The most stable of these is the enantiomeric trans-anti-trans-anti-trans stereoisomer 6.20, which we designate as PHTP, following general practice; only one of the two enantiomers is shown. PHTP constitutes about 60% of the product yield in the synthesis and is also the most easily isolated as it is the least soluble of the isomers. The enantiomers have been separated (Farina and Audisio, 1970b). Both racemic and enantiomeric PHTP form tunnel inclusion complexes with a large variety of guest molecules (Farina and Audisio, 1970b; Farina, Allegra and Natta, 1964); most of the published work (Farina,1984) concerns the complexes of racemic PHTP and it is to these that we shall refer in general, as it appears, from many hints in the literature, that diffraction quality crystals of complexes of enantiomeric PHTP have not yet been obtained. Aspects of the physical and structural chemistry of the complexes, including the polymerization of monomers in the tunnels, have been reviewed (Farina, Di Silvestro and Sozzani, 1996). Some 400 potential guests have been checked for formation of tunnel inclusion complexes with PHTP and some 80 were found to form such complexes (Fro¨mming and Oppermann, 1974). The following list, which is certainly not complete, is taken from a number of sources: 1. 2. 3. 4. 5. 6. 7.

n-alkanes, from C5 through C36; aliphatic dienes, including butadiene, cis- and trans-pentadiene, trans,trans -hexa2,4-diene, 2,5-dimethylhexa-2,4-diene; saturated monocarboxylic acids, from hexanoic (C6) to hexacosanoic (C26); saturated dicarboxylic acids from hexanedicarboxylic (C6) through to hexadecanedicarboxylic (C16); unsaturated carboxylic acids, including sorbic (C6) and four C18 acids; alcohols and diols, from heptanol (C7), octanol, dl-octanol-2, and others up to octadecan-9-diol-8,11 (C18); aliphatic and aromatic aldehydes, including heptanal, octanal, 2-methylundecanal, 4-methoxybenzaldehyde;


H H

H

299

H

H H

Fig. 6.63. Different views of PHTP (6.20). Most hydrogens of methylene groups have been omitted for clarity. The central ring, with its hydrogens, is emphasized.

8. 9. 10. 11.

12. 13. 14.

halogenated compounds, including CHCl3, CCl4, heptyl and octyl bromide, 11-bromo- undecanoic acid; aliphatic ketones, including 2-, 4-, 5-, 6- and 7-tridecanone (C13), 1-octan-5-one (C8), 1-decan-3-one(C10); aromatic carboxylic acids, including benzoic and 4-methylbenzoic acid, cinnamic acid and its hydroxy and 4-methoxy derivatives; aromatics, including benzene, toluene, 1-phenyl-2-methylpropane, 1-methyl7 and 1-ethylnaphthalene, 1,6-diphenyl-1,3,5-hexatriene, trans-stilbene, 2,2,11,11tetramethyl-3,5,7,9-dodecatetraene; cyclic compounds, including dioxan, cyclohexane, decalin and tetralin; linear macromolecules such as polyethylene, cis- and trans-1,4-polybutadiene, polyoxyethyleneglycol; twenty molecules of varying kinds, all with non-linear optical (NLO) properties, form tunnel inclusion complexes with PHTP (Hoss, Ko¨nig, Kramer-Hoss, Berger, Rogin and Hulliger, 1996);

The number and variety of the guests remind one of the tunnel inclusion complexes of urea and thiourea; both urea and PHTP show a certain selectivity towards linear longchain guests, but urea does not form complexes with CHCl3 or CCl4, although {3(thiourea)[CCl4]} is well known. PHTP has two polymorphic phases (Table 6.16); the P21/n phase (crystal structure at 173K by Harlow and Desiraju, 1990) is stable up to the melting point (398K) while the C2/c phase (m.pt. 390K) is monotropic with respect to it (Farina, Di Silvestro, Grassi and Sozzani, quoted in Farina (1984)). Neither of these phases has a structure similar to those of the inclusion complexes, so the latter constitute separate phases in the host–guest phase diagrams. This is illustrated by the melting point diagram for the PHTP-transstilbene system given by Fro¨mming and Oppermann (1974) and that for the PHTP-nheptane system given by Farina (1984) (Figs. 6.64 and 6.65)). As Farina (1984) points out, the inclusion complexes of PHTP are low-stability binary adducts and will be expected to behave, from the standpoint of the phase rule, similarly to hydrates and other binary systems where the stability of the phase rule compound formed is limited to the solid state and melting or dissolution leads to decomposition of the compound (Farina and 7 Note that Harlow and Desiraju (1990) crystallized neat PHTP from this solvent, and an inclusion complex was not formed.


300

130 125

temperature (°C)

120

liquid (L)

115 stilbene + L

C+ PHTP

L+C

110 105 100

stilbene + C

95 0 stilbene

20

40

60

80

100 PHTP

mol % PHTP

Fig. 6.64. T–x phase diagrams for the binary PHTP system with trans-stilbene as guest. (Adapted from Fro¨mming and Oppermann (1974).)

L

C

100

L+C

C+ PHTP

50

0

0 n-heptane

50 mole% PHTP

100 PHTP

Fig. 6.65. T–x phase diagrams for the binary PHTP system with n-heptane as guest. (Reproduced from Farina, 1984).

Di Silvestro, 1980, 1982). The phase diagrams each show the formation of only one inclusion complex; similar results were obtained for systems with n-alkanes and polyethylene as guests, and there is some resemblance to the phase diagram of apocholic acid and montanic acid (Fig. 6.52). The form of the phase diagrams for the PHTP inclusion


301

165

M.Pt. (°C)

155 Complex with polyethylene as guest melts at 178 °C

145

135

125

115 5

10

15

20 25 30 n of CnH2n+2

35

40

Fig. 6.66. Melting points of the n-alkane tunnel inclusion complexes of PHTP plotted against the number of carbon atoms in the guest molecule; complexes with n even are shown by filled squares and with n odd by open circles. As the melting points appear to lie on two separate curves, there may be a structural change between n ¼ 20 and 30. Data taken from Table 5 of Farina, 1984.

complexes is quite different from those of the urea inclusion complexes (Fig. 6.29), where incongruent decomposition is found below the melting point of urea (132 C). The melting points of PHTP–n-alkane inclusion complexes have been determined (Farina, Di Silvestro, Grassi and Sozzani, 1984) and are shown in Fig. 6.66; there is no distinction between the melting points of odd and even alkanes as is found, for example, for the melting points of the analogous urea inclusion complexes. The tunnel inclusion complexes have structures based on a common pattern of hexagonally arranged stacks of superimposed PHTP molecules, with tunnels containing the guest molecules running between the stacks. Details of unit cell size and space group depend on the nature of the guest molecule. It is clear from the cell dimensions given in Table 6.16 that the fundamental structure is based on the stacking of PHTP molecules one ˚ . This is the structure implied for above the other to give a stack axis periodicity of 4.8 A ˚ in space {PHTP-n-heptane]} by the hexagonal cell dimensions of a ¼ 14.40, c ¼ 4.78 A group P63/m. Furthermore Zhost ¼ 2 requires that the PHTP molecules have symmetry Cs-m incompatible with their structure. Thus the PHTP molecules are probably stacked in short sequences of (þ) and () enantiomers, with an average site occupancy of 50% for each. When the separated (þ) or () enantiomer is used as host the space group is P63 and such averaging is no longer necessary (Farina and Audisio, 1970b). The n-heptane ˚ channel-axis periodicity and must therefore be disordered guest is longer than the 4.78 A both azimuthally and longitudinally. Photographs of crystals rotated or oscillated about the tunnel-axis direction show continuous diffuse layer lines corresponding to a period˚ , about 0.5 A ˚ less than that expected for the planar extended conformation of icity of 10.7 A n-heptane. Possible conformational models for the included molecules have been investigated from the diffuse scattering and by testing various ways of fitting the guests into the channels (Allegra, Farina, Immirzi, Colombo, Rossi, Broggi and Natta, 1967). The resemblances to the diffraction patterns obtained from the {urea–[n-alkane]} tunnel


302

˚ , deg.) for the two polymorphs and some tunnel inclusion complexes of Table 6.16. Crystal data (A racemic PHTP. In this table the cells of monoclinic crystals have c unique to facilitate comparison with the hexagonal and rhombohedral cells. Z gives the number of molecules of each component in the unit cell b

c

Guest

a

Polymorph I PHTRPL03 HD90 Polymorph II PHTRPL01 AFI67 n-C7H16 PHTPHP AFI67 CHCl3 PHTPCF AFI67 dioxane QQQCQG AFI67 cyclohexane TPTRPC AFI67 benzene AFI67 butadiene CA71 1-(4-nitrophenyl)-piperazine (NPP)NOVWOK; KBA96

18.315 15.319 5.298 95.53 P21/n

4

–

16.94

10.41

9.73

113.5 C2/c

4

–

14.40 25.08 25.11 25.55

14.40 25.08 25.11 25.55

4.78 4.78 28.68 43.02

120 120 120 120

2 6 36 54

0.45 3 15 21

2. 3.

0.76 1 0.8

4.

15.70 13.9 4.78 121 13.35 14.72 4.78 115 15.023 23.198 4.730 90

Space group

P63/m P63/m R 3 R 3

Zhost Zguest Notes

2 P21/m 2 Cmc21 4

1.

5; at 100K, see text.

Notes: 1. At 173K. The polymorph was crystallized from 1-methylnaphthalene, which is noted above as forming an inclusion complex with PHTP. The contradiction needs resolution. PHTRPL02 (Luca et al., 1983) appears identical to PHTRPL03 but a space group was not given. 2. This structure type is also found for complexes with other n-hydrocarbons, n-ethers, n-carboxylic acids, n-esters, iso-octane and CCl4 as guests. There is no coherence between host and guest sub-lattices. 3. Room temperature form. 4. Cell dimensions are average values for the three guests, benzene, toluene, bromoform. The crystals were described as monoclinic without details of space group. 5. The twenty guests with NLO properties referred to above give analogous crystals with monoclinic or orthorhombic symmetry (Hoss, Ko¨nig, Kramer-Hoss, Berger, Rogin and Hulliger, 1996). References: AFI67 – Allegra, Farina, Immirzi, Colombo, Rossi, Broggi and Natta, 1967; CA71 – Colombo and Allegra, 1971; HD90 – Harlow and Desiraju, 1990; KBA97 – Ko¨nig, Bu¨rgi, Armbruster, Hulliger and Weber, 1997.

inclusion complexes are obvious, and one would expect analogous behaviour, involving ordering of the guests and consequent phase changes, on cooling PHTP complexes. Such studies have not yet been reported8 nor have specific heats (or other physical properties) been measured at low temperatures. The chloroform complex is an ordered version of the fundamental structure (see Fig. 6.67); the two channels have different symmetries with the guest molecules ordered in one set and disordered about centres of symmetry in the other. There is a first order transition at 58 C in which this distinction may be removed and both channels include ˚ axis periodicity means that the pseudo-mirror symmetry of disordered guests. The 4.78 A the PHTP molecules remains. This restriction falls away in the rhombohedral dioxane and cyclohexane complexes where the c-axis has to be multiplied by factors of 6 and 9 8

The structure of PHTP-NPP, described below, did not show changes between 55–300K.


303

(a)

b⬘

a⬘

=a

⬘

b= b⬘/ √3

y x

y x a = a⬘/ √3

(b)

b=

0

0

2 1

Y

x

4Å

Fig. 6.67. Schematic comparison of the molecular packing in the tunnel inclusion complexes of PHTP with (top) n-heptane and (below) CHCl3. n-Heptane is represented by circles and chloroform by three short segments at 120 . The geometrical relationship between the two unit cells in projection is shown in the upper diagram. The small arrows show the shifts of the PHTP molecules of the CHCl3 complex from their positions in the heptane complex; this leads to an increase in cell volume by a factor of 3.035 (25.08/14.4 ¼ 1.742). Reproduced from Allegra, Farina, Immirzi, Colombo, Rossi, Broggi and Natta, 1967.

304


3.7 3.7

4.02

4.1

3.92

3.7

1

0 3.9 4.28

3.5

3.8 3.4 3.3 3.3 Z Y 2Å

Fig. 6.68. The modulated structure of PHTP-cyclohexane viewed in projection down the [010] axis. Only part of the unit cell is shown. (Reproduced from Allegra, Farina, Immirzi, Colombo, Rossi, Broggi and Natta, 1967.)

respectively. The PHTP positions along the tunnel axis in the cyclohexane complex have ˚ ; the cyclohexane molecules been shown to be helically modulated with a radius of 0.4 A are orientationally disordered but interlocked with the displaced PHTP molecules (Fig. 6.68). The disorder in projection has been modeled by a very computer-intensive method developed by Welberry and coworkers (Mayo, Proffen, Bown and Welberry, 1999), which has also been applied to modeling of the diffuse scattering from urea-nalkane channel inclusion complexes. The model takes into account both disorder of the cyclohexane guest molecules and distortion of the host matrix. The latter is an important feature because the rhombohedral crystals transform reversibly into a micro-twinned monoclinic structure at 273K. The benzene and butadiene complexes appear to be isostructural; crystal quality was adequate for structure determination only for the butadiene complex (Colombo and Allegra, 1971). The principal difference from the structures described earlier is that the butadiene molecules are here ordered in the channels; their propinquity allows for facile polymerization under the influence of X- or -rays. The inclusion complex of 1,4-transpolybutadiene with PHTP has two polymorphic forms, with a transition at 70 C (Iwayanagi, Sakurai, Sakurai and Seto, 1968). Among other macromolecules that form inclusion complexes with PHTP are polyethylene and polyoxyethylene. All these crystals are isomorphous with {PHTP[n-heptane]}, but those with isotactic 1,4-trans-polypentadiene and 1,4-trans-poly-2,3-dimethylbutadiene as guests appear to have different structures not yet reported. PHTP is an especially interesting host for the formation of crystals with nonlinear optical properties as co-crystallization experiments show that it includes many of the most efficient NLO molecules currently available (Hoss, Ko¨nig, Kramer-Hoss, Berger, Rogin and Hulliger, 1997; Ko¨nig and Hulliger, 1996). The complex {PHTP[NPP]0.2} (NPP ¼ 1-(4-nitrophenyl)-piperazine (6.21)) has been studied in some detail (Ko¨nig,


305

Bu¨rgi, Armbruster, Hulliger and Weber, 1997), and some of their principal results are given here as a paradigm for the study of polar tunnel inclusion complexes.

NO2

N N NPP; 6.21

The crystals form as yellowish needles, up to several millimeters in length; they melt at 429K whereas ( )PHTP melts at 403K and NPP at 410K (cf. Fig. 6.64). They show second harmonic generation (SHG), pyroelectric and electro-optic (EO) properties. Oscillation and precession X-ray photographs (essentially unchanged over a temperature range from 55 to 300K) show a repeating pattern of equally spaced planes in reciprocal space (along c*) with diffuse and sharp scattering features. Every fifth plane (cor˚ ) shows only Bragg-like, sharp diffraction spots, while the responding to d(001) ¼ 4.73 A scattering associated with the four planes in between is essentially continuous within the planes and sharp normal to them. A set of relatively distinct but weak reflections is superimposed on the diffuse planes. The structure was determined at 100K. There are stacks of disordered (þ)- and ( )PHTP molecules along [001], analogous to those shown in Fig. 6.69; the PHTP molecules conform to space group Cmcm. However, homochiral PHTP molecules fit together better in a stack than heterochiral molecules, and so each particular stack is homochiral and the averaging is done over stacks as a whole. Further analysis, based principally on the weak Bragg reflections, showed that the polar NPP molecules were lined up in a head-to-tail arrangement, and that the chains in adjacent tunnels were parallel (Fig. 6.69); the NPP molecules in a tunnel are located so that the ˚ with the mean plane of a neighcentres of the phenyl rings coincide to within 0.05 A bouring PHTP molecule. Additional experiments showed that crystals containing 1.4 mol% of the tailor-made growth additive 1-(p-tolyl)piperazine (TP) entirely lost polar physical properties. The conclusions from the experimental study can be summarized as follows: 1. Occurrence of SHG, EO and pyroelectricity confirm polarity along [001], and show that the space group of the crystal as a whole is Cmc21. 2. Reversal of the pyroelectric effect along [001] and the lack of polar growth properties indicate twinning across (001). ˚ (i.e. about 10 NPP 3. The polar domains along [001] have a length of at least 100 A molecules). 4. Consistency of the physical properties of crystals grown under different conditions showed that these properties are not due to chance but must be attributed to special features of the nucleation and growth process. Consider the growth process in a particular tunnel. If a nitro group of NPP is exposed, then the (increasing) relative strengths of nitro . . . nitro, NH . . . NH and NH . . . nitro interactions indicate that the next NPP molecule will be linked, tail-to-head, by nitro . . . H–N hydrogen bonding, and so on, giving a polar chain. If a N–H group is exposed then the chain will be of opposite polarity, but the feasibility of N–H . . . N–H


306

c9/5

P H T P

Fig. 6.69. Schematic diagram of {PHTP[(NPP) 0.2]} showing NPP molecules in the tunnels. The ˚ and the repeat distance along [001] (vertical) is 5 4.73 ¼ arrow at upper left represents 4.73 A ˚ , the length of two NPP molecules. The NPP molecules are hydrogen bonded between nitro 23.65 A groups (at the top of each molecule) and H–N groups (at lower end). The chains of NPP molecules in the different tunnels are parallel. (Reproduced from Ko¨nig, Bu¨rgi, Armbruster, Hulliger and Weber, 1997.)

hydrogen bonding makes it possible to reverse the chain sequence. Thus nitro groups will be exposed at both ends of a growing tunnel, with the macroscopic crystal being a twin of two antiparallel polar domains. As growth in all the tunnels will be governed by the same mechanism, the polar chains in adjacent tunnels will be parallel (and not randomly ˚ . Application of Markov statistics has allowed oriented) despite their separation by 15 A these qualitative considerations to be given a quantitative hue. A rather similar treatment has been developed independently for polar urea inclusion complexes by Harris and Jupp (1997a, b). Vapour pressure measurements as a function of temperature have given values (Table 6.17) for the enthalpies of the following reactions: ðInclusion complexÞcryst: ) ðPHTPÞc þ ðguestÞvapor

DHv


307

Table 6.17. Enthalpies of reaction (kJ/mol) for some PHTP inclusion complexes. Guest

Hv

Hdec

(Hv – Hdec)

Hvap of guest

n-heptane cyclohexane dioxan chloroforma chloroformb

45.2 31.8 34.7 38.9 28.9

11.7 0.8 0.8 8.4 0.4

33.5 31.0 35.5 30.5 29.3

31.9 32.8

a b

31.4

polymorph stable at room temperature; polymorph stable above 58 C.

ðInclusion complexÞcryst: ) ðPHTPÞc þ ðguestÞliquid

DHdec

The measured values of (DHv–DHdec) should be equal to the enthalpy of vaporization of the various guests, and this is tested with satisfactory results in the final column of Table 6.17. The tunnel inclusion complexes of urea and thiourea follow a single structural pattern and individual structures are marked by great crystallographic similarity. Thus urea and thiourea show little versatility or adaptability in the structures of their complexes. In contrast, the tunnel inclusion complexes of perhydrotriphenylene show both versatility and considerable adaptability. 6.3.5 N-(p-tolyl)tetrachloro-phthalimide as host Pratt and Perkins (1918) reported that N-(p-tolyl)tetrachlorophthalimide (TTP; 6.22) crystallized from ethanol and glacial acetic acid as white plates of the neat compound but from aromatic solvents, or from solutions containing aromatic molecules of various kinds, as yellow needles of composition {4(TTP)[guest]}; they recognized that these were molecular complexes. Cl

O

Cl CH3

6.22

Cl Cl

O

The findings of Pratt and Perkins about the chemical nature of the TTP complexes have been confirmed, a number of structures have been determined (Herbstein and Kaftory, 1981) and the results have been put into a broader context (Herbstein, 1987). The guests must have some aromatic character (e.g. tetralin forms a complex but decalin and cyclohexane do not) but the size of permitted aromatics is limited (pyrene is the aromatic molecule with the largest cross section to form a complex). Heteroaromatics (e.g. pyridine and -picoline are permissible guests but not aromatic quinones (e.g. p-benzoquinone and 2-methyl-p-benzoquinone do not form complexes). TTP has a possible functionality as an acceptor in –* charge transfer compounds (by analogy to the well-known acceptor tetrachlorophthalic anhydride) and the complexes with perylene and fluoranthene

308


appeared to be of the charge transfer type (structures were not determined); tetrabenzonaphthalene did not form an adduct with TTP while phenothiazine was the only compound tested (out of 28) which formed adducts of both types. These results are explicable in terms of the crystal structure of the complexes. All the tunnel inclusion complexes were found to be isomorphous with cell dimensions which ˚ from one guest to another – there is minimal varied only by a few hundredths of an A ˚ , ¼ 106.4 , with adaptability. The cell dimensions are a ¼ 22.27, b ¼ 3.91, c ¼ 19.91 A 4 TTP molecules in the unit cell; the space group of the overall structure (i.e. without taking the effects of the guest into account, see below) is C2, which is a Sohnke group (o-dichlorobenzene guest BAVVAV; o-xylene guest BAVVEZ). The TTP molecules are arranged so as to leave a tunnel of elliptical cross section with dimensions close to those of a benzene ring viewed edge-on. The walls of the tunnel are lined with Cl and O atoms of the host molecules. The reason for the limitation in size of the guest molecules is immediately apparent – they must be able to fit into the tunnels. However, it is not at all clear why the guests must be aromatic; unfortunately the guests could not be located along the tunnel axis in the room-temperature structure determinations and thus there are no clues about their mode of interaction with the atoms of the tunnel walls. The chirality appears to be a solid state conformational effect – the p-tolyl groups are rotated by about 55 about the C–N bonds and the TTP molecules are thus chiral; spontaneous resolution of the TTP conformational enantiomers has taken place on crystallization. The yellow color of the complexes was ascribed to intramolecular charge transfer, there being limited resonance interaction between the donor and acceptor portions of the nonplanar TTP molecule. There is no intermolecular charge transfer interaction between host and guest. The crystallography is in fact more complicated than as described above. This is ˚ , most complexes because the true tunnel-axis periodicity of the unit cell is 7.8 A showing sets of weak layer lines corresponding to the doubled periodicity. In some complexes the reflections on these weak layer lines conform to the monoclinic subcell of the TTP molecules but in others their symmetry is lowered to triclinic, and an example of such a complex is shown in Fig. 6.70. The weak reflections due to the guest molecules provide, in potential, information about their disposition but it has not yet been found possible to exploit this to determine the full crystal structures. As in so many other examples of molecular compounds and complexes, low temperature studies would also be of value here. TTP has an interesting polymorphism with some relation to the structures of the inclusion complexes. The -polymorph (stable up to 173 C; space group Cmca, Z ¼ 8) has a structure (Kaftory, 1978; TOCPIM) very different from that of the complexes; there are intermolecular localized C ¼ O . . . Cl charge transfer interactions while the two ring systems of which the molecules are composed are mutually perpendicular, the lack of colour being ascribed to the consequent lack of intramolecular charge transfer interaction between donor and acceptor portions. However, the -phase (stable from 173 C to the melting point at 214 C, quenchable to room temperature; space group P21/c, Z ¼ 8; ˚ , very TOCPIM11) has a structure composed of stacks of TTP molecules, with b ¼ 3.97 A similar to those in the tunnel inclusion complexes; however, there are two kinds of stacks, composed of two differently linked sets of molecules (Fig. 6.71). The structure consists of sheets of molecular stacks in the (100) planes; in one set of sheets there are intermolecular C ¼ O . . . Cl charge transfer interactions as in the -phase, while in the interleaved set of sheets there is van der Waals bonding between molecules disposed approximately as in


309

(a) b† = 2bm

Q2

Q⬘2 P2

P⬘2

Q⬘1

Q1

P1

P⬘1

1 a sin b m 2 m 1 a sin b † 2 † 1 1 C = c sin 2 m 2 †

(b) P(xyz)

a†

CL1 C1 CL2

C6 C7

C2 C3 1 a = 2 m 1 a sin 2 †

CL3

origin

g†

bm = 12

Q( 12 + x, 12 + y, z)

C13

C15 C12 C9 C11 C10

C14

O1

N

Q( – 12 + x, 12 + y, z)

C5 O2 C4 CL4

b†up

P⬘(x,y,z)

0

5Å

Fig. 6.70. Projections of the structure of the {4(TTP)[o-xylene]} tunnel inclusion complex: Above View down the monoclinic c axis (which is the same as the triclinic c axis), showing two monoclinic cells (lightly outlined) and one triclinic cell (heavily outlined). The eight crystallographically independent TTP molecules in the triclinic cell are shown schematically, labelled so as to show their relationship to the four TTP molecules of the monoclinic subcell shown in the lower part of the diagram. Below Projection down the [010] axis of the overall unit cell of monoclinic symmetry; the monoclinic am and cm axes are in the plane of the page, as is ct (because t ¼ 90 ), but at runs down to the left below the plane of the page. The coordinates shown with the labels of the molecules refer to the monoclinic subcell. The van der Waals envelopes of the atoms lining the tunnels of elliptical cross-section are shown schematically; the actual cross-section of the tunnel depends on the y coordinate. The molecules at y are emphasised. (Reproduced from Herbstein and Kaftory, 1981.)


310

A molecules

B molecules

4

O(2) 4

A molecules

3.47 3 2.91

3 4 3.60

3.74 3.73

c

b down

a

Fig. 6.71. -TTP, projection of the unit cell contents onto (010). The structure consists of interleaved sheets of stacks of A and B molecules, which are conformationally identical but interact differently, as noted in the text. The short O . . . Cl contacts in the A strips are indicated by the ˚ . (Reproduced from Herbstein and Kaftory, 1981.) distance 2.91 A

the molecular complex, adjacent molecules being related by centres of symmetry rather than two fold axes. The two sets of sheets are linked by van der Waals interactions. The conformation of the TTP molecules in the -phase is essentially identical to that in the complexes, and the yellow colour of both is attributed to the same intramolecular charge transfer interactions. A difference is that the complexes are chiral while the -phase is achiral. Thus one can see that structural elements are carried over, in part, from one polymorph to the next as well as to the molecular complexes.

6.4

Comparison of the various tunnel inclusion complexes

Tunnel inclusion complexes consist of a host framework, with guests included in essentially linear tunnels. In all our examples the empty host framework is unstable with respect to the pristine host; thus the tunnel inclusion complexes are separate phases in the

REFERENCES

311

host–guest phase diagram. No example is known where the guest is in solid solution in a pristine host structure; this contrasts with the situation in the clathrates, where a few examples of primary solid solutions are known. Other behavior depends on the strength of the host–guest interaction, and on the possibility that host–guest interaction is sometimes stronger than host–host interaction. This latter effect shows itself most vividly in the urea complexes, where, for certain guests, the host framework is interrupted by formation of host–guest hydrogen bonds. Such interruption has not yet been encountered in the thiourea inclusion complexes. Host–guest interaction also manifests itself in the form of phase changes that occur on cooling. These have been widely studied in the urea and thiourea complexes, with their hydrogen bonded frameworks. {2(DCA)ferrocene} is the only example so far encountered in complexes with van der Waals bonded frameworks. However, there seems no reason to doubt wide occurrence also in complexes with van der Waals bonded frameworks, and variable-temperature studies of these complexes should provide interesting results.

References Adolphi, N. L., Conradi, M. S. and Matsuo, T. (1994). J. Phys. Chem., 98, 1968–1973. Ahn, P. D., Bishop, R., Craig, D. C., Downing, G. A. and Scudder, M. L. (1997). Aust. J. Chem., 50, 1053–1059. Allcock, H. R. (1978). Accts. Chem. Res., 11, 81–87. Allcock, H. R. (1984). ‘‘Cyclophosphazene inclusion adducts.’’ In Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Academic Press, London. Vol. 1, pp. 351–374. Allcock, H. R., Allen, R. W., Bissell, E. C., Smeltz, L. A. and Teeter, M. (1976). J. Am. Chem. Soc., 98, 5120–5125. Allcock, H. R., Ferrar, W. T. and Levin, M. L. (1982). Macromolecules, 15, 697–703. Allcock, H. R., Levin, M. L. and Whittle, R. R. (1986). Inorg. Chem., 25, 41–47. Allcock, H. R., Primrose, A. P., Silverberg, E. N., Visscher, K. B., Rheingold, A. L., Guzei, I. A. and Parvez, M. (2000). Chem. Mater., 12, 2530–2536. Allcock, H. R. and Siegel, L. A. (1964). J. Am. Chem. Soc., 86, 5140–5144. Allcock, H. R. and Stein, M. T. (1974). J. Am. Chem. Soc., 96, 49–52. Allcock, H. R., Stein, M. T. and Bissell, E. C. (1974). J. Am. Chem. Soc., 96, 4795–4800. Allcock, H. R., Stein, M. T. and Stanko, J. A. (1971). J. Am. Chem. Soc., 93, 3173–3178. Allcock, H. R. and Walsh, E. J. (1971). Inorg. Chem., 10, 1643–1647. Allegra, G., Farina, M., Immirzi, A., Colombo, A., Rossi, U., Broggi, R. and Natta, G. (1967). J. Chem. Soc. (B), pp. 1020–1028. Andreetti, G. D. (1984). In Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Academic Press, London, Vol. 3, pp. 129–146. Angla, B. (1947a). Compt. Rend. Acad. Sci. Paris, 224, 402–404. Angla, B. (1947b). Compt. Rend. Acad. Sci. Paris, 224, 1166–1167. Bahr, G. and Meier, G. (1958). Z. anorg. allgem. Chem., 294, 22–49. Barrer, R. M., Drake, J. and Whittam, T. V. (1953). Proc. R. Soc. Lond., A219, 32–53. Bashir-Hashemi, A., Hart, H. and Ward, D. L. (1986). J. Am. Chem. Soc., 108, 6675–6679. Bekkum, H. van, Palm, J. H., Verkade, P. E. and Wepster, B. M. (1970). Rec. Trav. Chim. Pays-Bas, 89, 913–919. Bekkum, H. van, Remijnse, J. D. and Wepster, B. M. (1967). J. Chem. Soc. Chem. Comm., pp. 67–68.

312


Bengen, M. F. (1951). Angew. Chem., 63, 207–208. Bertolasi, V., Bortolini, O., Fogagnolo, M., Fantin, G. and Pedrini, P. (2001). Tetrahedron Asymmetry, 12, 1479–1483. Bhatnagar, V. M. (1967). J. Struct. Chem. U.S.S.R., 8, 568–586. Bishop, R. (1996). ‘‘Helical host lattices formed by alicyclic diols,’’ in Solid State Supramolecular Engineering; Crystal Engineering, Volume 6 of Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, pp. 85–115. Bishop, R., Choudhury, S. and Dance, I. G. (1982). J. Chem. Soc. Perkin Trans. II, pp. 1159–1168. Bishop, R., Craig, D. C., Dance, I. G., Kim, S., Mallick, M. A. I., Pich, K. C. and Scudder, M. L. (1993). Supramol. Chem., 1, 171–178. Bishop, R., Craig, D. C., Dance, I. G., Scudder, M. L., Marchand, A. P. and Wang, Y. (1993). J. Chem. Soc., Perkin 2, pp. 937–943. Bishop, R., Craig, D. C., Dance, I. G., Scudder, M. L. and Ung, A. T. (1993). Supramol. Chem., 2, 123–131. Bishop, R., Craig, D. C., Dance, I. G., Scudder, M. L. and Ung, A. T. (1994). Mol. Cryst. Liq. Cryst., 240, 113–119. Bishop, R., Craig, D. C., Dance, I. G., Scudder, M. L. and Ung, A. T. (1999). J. Struct. Chem., 40, 633–671. Bishop, R., Craig, D. C., Marougkas, A. and Scudder, M. L. (1994). Tetrahedron, 50, 8749–8756. Bishop, R., Craig, D. C., Scudder, M. L., Marchand, A. P. and Liu, Z. (1995). J. Chem. Soc., Perkin 2, pp. 1295–1300. Bishop, R. and Dance, I. G. (1979). J. Chem. Soc. Chem. Comm., pp. 992–993. Bishop, R. and Dance, I. G. (1988). Top. Curr. Chem., 149, 137–188. Bishop, R. and Dance, I. G. (1991). ‘‘Design, synthesis, and structure of alicyclic diol hosts,’’ In Inclusion Compounds edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Oxford University Press, Vol. 4, pp. 1–26. Bishop, R., Dance, I. G. and Hawkins, S. C. (1983). J. Chem. Soc. Chem. Comm., pp. 889–891. Bishop, R., Dance, I. G., Hawkins, S. C. and Lipari, T. (1984). J. Incl. Phenom., 2, 75–83. Bishop, R., Dance, I. G., Hawkins, S. C. and Scudder, M. L. (1987) J. Incl. Phenom., 5, 229–232. Blow, D. M. and Rich, A. (1960). J. Am. Chem. Soc., 82, 3566–3571. Bogomolov, V. N., Poborchoy, V. V., Romanov, S. G. and Shagin, S. I. (1985). J. Phys. C, 18, L313–L317. Bonamico, M. and Giacomello, G. (1963). Gazz. Chim. Ital., 92, 647–650 (1962); Chem. Abstr. 58, 2919. Briozzo, P., Kondo, T., Sada, K., Miyata, M. and Miki, K. (1996). Acta Cryst., B52, 728–733. Brown, M. E., Chaney, J. D., Santarsiero, B. D. and Hollingsworth, M. D. (1996). Chem. Mater., 8, 1588–1591. Brown, M. E. and Hollingsworth, M. D. (1995). Nature, 376, 323–327. Caira, M. R., Nassimbeni, L. R. and Russell, J. C. (1977). Acta Cryst., B33, 1171–1176. Caira, M. R., Nassimbeni, L. R. and Scott, J. L. (1993), J. Chem. Soc., Chem. Commun., pp. 612–614. Caira, M. R., Nassimbeni, L. R. and Scott, J. L. (1994a). J. Chem. Soc., Perkin 2, pp. 623–628. Caira, M. R., Nassimbeni, L. R. and Scott, J. L. (1994b). J. Chem. Crystallogr., 24, 783–791. Caira, M. R., Nassimbeni, L. R. and Scott, J. L. (1994c). J. Chem. Soc., Perkin 2, pp. 1403–1405. Candeloro de Sanctis, S., Coiro, V. M., Giglio, E., Pagliuca, S., Pavel, N. V. and Quagliata, C. (1978). Acta Cryst., B34, 1928–1933. Candeloro de Sanctis, S., Coiro, V. M., Mazza, F. and Pochetti, G. (1995). Acta Cryst., B51, 81–89. Candeloro de Sanctis, S. and Giglio, E. (1979). Acta Cryst., B35, 2650–2655. Candeloro de Sanctis, S., Giglio, E., Pavel, N. V. and Quagliata, C. (1972). Acta Cryst., B28, 3656–3661.

REFERENCES

313

Candeloro de Sanctis, S., Giglio, E., Petri, F. and Quagliata, C. (1979). Acta Cryst., B35, 226–228. Cannarozzi, G. M., Meresi, G. H., Vold, R. L. and Vold, R. R. (1991). J. Phys. Chem., 95, 1525–1527. Cerrini, S., Pochetti, G., Gallese, F. and Possagno, E. (1993). Acta Cryst., C49, 1087–1092. Chang, H. C., Popovitz-Biro, R. Lahan, M. and Leiserowitz, L. (1987). J. Am. Chem. Soc., 109, 3883–3893. Chang, S. C. and Jeffrey, G. A. (1970). Acta Cryst., B26, 673–683. Chatani, Y., Anraku, H. and Taki, Y. (1978). Mol. Cryst. Liq. Cryst., 48, 219–231. Chatani, Y. and Kuwata, S. (1975). Macromolecules, 8, 12–18. Chatani, Y. and Nakatani, S. (1972). Macromolecules, 5, 597–604. Chatani, Y. and Nakatani, S. (1976). Z. Kristallogr., 144, 175–188. Chatani, Y., Taki, Y. and Todokoro, H. (1977). Acta Cryst., B33, 309–311. Claude, R., Cle´ment, R. and Dworkin, A. (1977). J. Chem. Thermodynamics, 9, 1199–1204. Cle´ment, R., Claude, R. and Mazieres, C. (1974). J. Chem. Soc. Chem. Comm., pp. 654–655. Cle´ment, R., Gourdji, M. and Guibe´, L. (1975). J. Magn. Reson., 20, 345–350. Cle´ment, R., Gourdji, M. and Guibe´, L. (1980). Chem. Phys. Lett., 72, 466–468. Cle´ment, R., Mazieres, C., Gourdji, M. and Guibe´, L. (1977). J. Chem. Phys., 67, 5381–5385. Cle´ment, R., Mazieres, C. and Guibe´, L. (1971). Mol. Phys., 21, 247–255. Cle´ment, R., Mazieres, C. and Guibe´, L. (1972). J. Solid State Chem., 5, 436–440. Coiro, V. M., D’Andrea, A. and Giglio, E. (1979). Acta Cryst., B35, 2941–2944. Coiro, V. M., D’Andrea, A. and Giglio, E. (1980). Acta Cryst., B36, 848–852. Coiro, V. M., Giglio, E., Mazza, F. and Pavel, N. V. (1984). J. Incl. Phenom., 1, 329–337. Coiro, V. M., Giglio, E., Mazza, F., Pavel, N. V. and Pochetti, G. (1982). Acta Cryst., B38, 2615–2620. Coiro, V. M., Mazza, F., Pochetti, G., Giglio, E. and Pavel, N. V. (1985). Acta Cryst., C41, 229–232. Colombo, A. and Allegra, G. (1971). Macromolecules, 4, 579–584. Cope, A. F. G., Gannon, D. J. and Parsonage, N. G. (1972a). J. Chem. Thermodynamics, 4, 829–842. Cope, A. F. G., Gannon, D. J. and Parsonage, N. G. (1972b). J. Chem. Thermodynamics, 4, 843–848. Craven, B. M. (1973). Acta Cryst., B29, 1525–1528. Craven, B. M. and Titta, G. de, (1972). J. Chem. Soc. Chem. Comm., pp. 530–531. D’Andrea, A., Fedeli, W., Giglio, E., Mazza, F. and Pavel, N. V. (1981). Acta Cryst., B37, 368–372. Dance, I. G., Bishop, R., Hawkins, S. C. Lipari, T., Scudder, M. L. and Craig, D. C. (1986). J. Chem. Soc. Perkin Trans. II, pp. 1299–1307. Dance, I. G., Bishop, R. and Scudder, M. L. (1986). J. Chem. Soc. Perkin Trans. II, pp. 1309–1318. Dastidar, P. (2000). CrystEngComm., No. 8. Dean, F. M., Herbstein, F. H., Kapon, M. and Reisner, G. M. (1991). Acta Cryst., C47, 137–141. Demarçay, H. (1838). Ann., 27, 270–291. Dirscherl, W. and Gerhards, E. (1961). Liebigs Ann. Chem., 639, 181–194. Edwards, H. G. M., Lawson, E., Matas, M. de, Shields, L. and York, P. (1997). J. Chem. Soc. Perkin Trans. II, pp. 1985–1990. Ehrenfest, P. (1933). Proc. Amsterdam Acad., 36, 153–157. El Ghallali, J., Gourdji, M., Guibe´, L. and Peńeau, A. (1994). Z. Naturforsch., A49, 433–438. Engel, D. W. (1972). Acta Cryst., B28, 1496–1509. Ermer, O. and Lindenberg, L. (1991). Helv. Chim. Acta, 74, 825–877. Evain, M., Queignec, M., Brec, R. and Rouxel, J. (1985). J. Sol. State Chem., 56, 148–157. Fait, J. F., Fitzgerald, A., Caughlan, C. N. and McCandless, F. P. (1991). Acta Cryst., C47, 332–337. Farina, M. (1984). ‘‘Inclusion polymerization.’’ in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Academic Press, London., Vol. 3, pp. 297–329. Farina, M., Allegra, G. and Natta, G. (1964). J. Am. Chem. Soc., 86, 516–518. Farina, M. and Audisio, G. (1970a). Tetrahedron, 26, 1827–1837.

314


Farina, M. and Audisio, G. (1970b). Tetrahedron, 26, 1839–1844. Farina, M. and Di Silvestro, G. (1980). J. Chem. Soc. Perkin Trans. II, pp. 1406–1410. Farina, M. and Di Silvestro, G. (1982). Gazz. Chim. Ital., 112, 91–97. Farina, M., Di Silvestro, G. and Colombo, A. (1986). Mol. Cryst. Liq. Cryst., 137, 265–277. Farina, M., Di Silvestro, G. and Sozzani, P. (1996). ‘‘Perhydrotriphenylene: A D3 symmetric host,’’ in Solid State Supramolecular Engineering; Crystal Engineering, Volume 6 of Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, Chapter 12, pp. 371–398. Ferro, D., Imperatori, P. and Quagliata, C. (1983). J. Chem. Eng., Data, 28, 242–244. Ferro, D., Quagliata, C. and Conte, M. R. (1983). Thermochim. Acta, 60, 211–217. Ferro, D., Quagliata, C., Giglio, E. and Piacente, V. (1981). J. Chem. Eng. Data, 26, 192–195. Fetterly, L. C. (1964). ‘‘Organic adducts’’, in Non-Stoichiometric Compounds, edited by L. Mandelcorn, pp. 491–567, Academic Press, New York and London. Fiechter, S., Kuhs, W. F. and Nitsche, R. (1980). Acta Cryst., B36, 2217–2220. Fieser, L. F. and Fieser, M. (1959). Steroids., pp. 55–61, Reinhold, New York. Fieser, L. F. and Newman, M. S. (1935). J. Am. Chem. Soc., 57, 1602–1604. Finney, J. L. (1995). Acta Cryst., B51, 447–467. Forst, R., Boysen, H., Frey, F., Jagodzinski, H. and Zeyen, C. (1986). J. Phys. Chem. Solids, 47, 1089–1097. Forst, R., Jagodzinski, H., Boysen, H. and Frey, F. (1987). Acta Cryst., B43, 187–197. Forst, R., Jagodzinski, H., Boysen, H. and Frey, F. (1990). Acta Cryst., B46, 70–78. Fowweather, F. S. (1949). Biochem. J., 44, 607–610. Friedman, N., Lahav, M., Leiserowitz, L., Popovitz-Biro, R., Tang, C. P. and Zaretzkii, Z. (1975). J. Chem. Soc. Chem. Comm., pp. 864–865. Frey, F. (1997). Z. Kristallogr., 212, 257–282. Fro¨mming, K.-H. and Opperman, J. (1974). Arch. Pharmaz., 307, 18–23. Fukao, K. (1994a). J. Chem. Phys., 101, 7882–7892. Fukao, K. (1994b). J. Chem. Phys., 101, 7893–7903. Fukao, K., Horiuchi, T., Taki, S. and Matsushige, K. (1990). Mol. Cryst. Liq. Cryst., 180B, 405–416. Garneau, I., Raymond, S. and Brisse, F. (1995). Acta Cryst., C51, 538–541. Gdaniec, M., Milewska, M. J. and Polonski, T. (1999). Angew. Chem. Int. Ed. Engl., 38, 392–XXX. Gdaniec, M. and Polonski, T. (1998). J. Am. Chem. Soc., 120, 7353–7354. Gelerinter, E., Luz, Z., Poupko, R. and Zimmerman, H. (1990). J. Phys. Chem., 94, 5391–5398. Gerdil, R. (1961). Acta Cryst., 14, 333–344. Gerdil, R. and Marsh, R. E. (1960). Acta Cryst., 13, 166–167. Giacomello, G. and Bianchi, E. (1943). Gazz. Chim. Ital., 73, 3–13. Gibb, T. C. (1976). J. Phys. C: Solid State Phys., 9, 2627–2642. Giglio, E. (1984). ‘‘Inclusion compounds of deoxycholic acid.’’ in Inclusion Compounds, Vol. 2, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Academic Press, Orlando., pp. 207–229. Giglio, E. and Quagliata, C. (1975). Acta Cryst., B31, 743–746. Giglio, E., Mazza, F. and Scaramuzza, L. (1985). J. Incl. Phenom., 3, 437–445. Gilson, D. F. R. and McDowell, C. A. (1961). Mol. Phys., 4, 125–134. Go and Kratky (1934). Z. phys. Chem. (Leipzig), B26, 439–452. Gopal, R., Robertson, B. E. and Rutherford, J. S. (1989). Acta Cryst., C45, 257–259. Hadicke, E. and Schlenk, W. Jr. (1972). Ann., 764, 103–111. Harlow, R. L. and Desiraju, G. R. (1990). Acta Cryst., C46, 1054–1055. Harris, K. D. M., (1998). J. Sol. State Chem., 84, 280–284.

REFERENCES

315

Harris, K. D. M., Gameson, I. and Thomas, J. M. (1990). J. Chem. Soc., Faraday Trans., 86, 3135–3143. Harris, K. D. M. and Hollingsworth, M. D. (1990). Proc. R. Soc. Lond., Ser. A, 431, 245–269. Harris, K. D. M. and Jonsen, P. (1989). Chem. Phys. Lett., 154, 593–598. Harris, K. D. M. and Jupp, P. E. (1997a). Proc. R. Soc. Lond., Ser. A, 453, 333–352. Harris, K. D. M. and Jupp, P. E. (1997b). Chem. Phys. Lett., 274, 525–534. Harris, K. D. M., Smart, S. P. and Hollingsworth, M. D. (1991). J. Chem. Soc. Farad. Trans., 87, 3423–3429. Harris, K. D. M. and Thomas, J. M. (1990a). J. Chem. Soc. Faraday Trans., 86, 1095–1101. Harris, K. D. M. and Thomas, J. M. (1990b). J. Chem. Soc. Faraday Trans., 86, 2985–2996. Hawkins, S. C., Bishop, R., Craig, D. C., Dance, I. G., Rae, A. D. and Scudder, M. L. (1993). J. Chem. Soc., Perkin 2, pp. 1737–1745. Hawkins, S. C., Bishop, R., Dance, I. G., Lipari, T., Craig, D. C., and Scudder, M. L. (1993). J. Chem. Soc., Perkin Trans. II, pp. 1729–1735. Hawkins, S. C., Scudder, M. L., Craig, D. C., Rae, A. D., Raof, R. B. A., Bishop, R. and Dance, I. G. (1990). J. Chem. Soc., Perkin 2, pp. 855–870. Herbstein, F. H. (1987). Top. Curr. Chem., 140, 107–139. Herbstein, F. H. and Kaftory, M. (1981). Z. Kristallogr, 157, 1–25. Herbstein, F. H. and Marsh, R. E. (1977). Acta Cryst., B33, 2358–2367. Herndon, W. C. (1967). J. Chem. Educ., 44, 724–728. Heyes, S. J. and Dobson, C. M. (1990). Magn. Reson. in Chem., 28, 37–46. Heyes, S. J., Clayden, N. J. and Dobson, C. M. (1991). J. Phys. Chem., 95, 1547–1554. Hollingsworth, M. D., Brown, M. E., Hillier, A. C., Santarsiero, B. D. and Chaney, J. D. (1996). Science, 273, 1355–1359. Hollingsworth, M. D. and Harris, K. D. M. (1996). ‘‘Urea, thiourea, and selenourea,’’ in Solid State Supramolecular Engineering; Crystal Engineering, Volume 6 of Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, pp. 177–237. Hollingsworth, M. D., Santarsiero, B. D. and Harris, K. D. M. (1994). Angew. Chem. Int. Ed. Engl., 33, 649–652. Hoss, R., Ko¨nig, O., Kramer-Hoss, V., Berger, U., Rogin, P. and Hulliger, R. (1996). Angew. Chem. Int. Ed. Engl., 35, 1664–1666. Hough, E. and Nicholson, D. G. (1978). J. Chem. Soc. Dalton Trans., pp. 15–18. Hughes, E. W., Yakel, H. L. and Freeman, H. C. (1961). Acta Cryst., 14, 345–352. Huntress, E. H. and Philips, R. F. (1949). J. Am. Chem. Soc., 71, 458–460. Iwayanagi, S., Sakurai, I., Sakurai, T., and Seto, T. (1968). J. Macromolec. Sci. Phys., 2, 163–177. Johnson, P. L. and Schaefer, J. P. (1972). Acta Cryst., B28, 3083–3088. Jones, E. L. and Nassimbeni, L. R. (1990). Acta Cryst., B46, 399–405. Jones, J. G., Schwarzbaum, S., Lessinger, L. and Low, B. W. (1982). Acta Cryst., B38, 1207–1215. Jones, M. J., Shannon, I. J. and Harris, K. D. M. (1996). J. Chem. Soc., Faraday Trans., 92, 273–279. Kaftory, M. (1978). Acta Cryst., B34, 471–475. Kobayashi, T., Isoda, S. and Kubono, K. (1996). ‘‘Cyclophosphazenes: Structures and molecular selectivities,’’ in Solid State Supramolecular Engineering; Crystal Engineering, Chapter 13 in Volume 6 of Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, pp. 399–419. Ko¨nig, O., Bu¨rgi, H.-B., Armbruster, T., Hulliger, R. and Weber, T. (1997). J. Am. Chem. Soc., 119, 10632–10640.

316


Ko¨nig, O. and Hulliger, R. (1996). Nonlinear Optics (Mol. Cryst. Liq. Cryst., Sci. Technol. Sect. B), 17, 217–139. Kubono, K., Asaka, N., Isoda, S. and Kobayashi, T. (1994). Acta Cryst., C50, 324–326. Kubono, K., Asaka, N., Isoda, S., Kobayashi, T. and Taga, T. (1993). Acta Cryst., C49, 404–406. Kubono, K., Asaka, N., Taga, T., Isoda, S. and Kobayashi, T. (1994). J. Mater. Chem., 4, 291–297. Kubono, K., Karata, H., Isoda, S. and Kobayashi, T. (1993). J. Mat. Chem., 3, 615–621. Lachinov, P. A. (1885). Chem. Ber., 18, 3039–3047. Lahav, M., Leiserowitz, L., Popovitz-Biro, R. and Tang, C. P. (1978). J. Am. Chem. Soc., 100, 2542–2544. Laves, F., Nicolaides, N. and Peng, K. C. (1965). Z. Kristallogr., 121, 258–282. Lenne´, H.-U. (1954). Acta Cryst., 7, 1–15. Lenne´, H.-U., Mez, H. C. and Schlenk, W., Jr. (1968). Chem. Ber., 101, 2435–2444. Lenne´, H.-U., Mez, H. C. and Schlenk, W., Jr. (1970). Ann., 732, 70–96. Lessinger, L. (1982). Cryst. Struct. Commun., 11, 1787–1792. Lessinger, L. and Low, B. W. (1993). J. Cryst. Spectr. Res., 23, 85–92. Lindley, P. F., Mahmoud, M. M., Watson, F. E. and Jones, W. A. (1980). Acta Cryst., B36, 1893–1897. Lowery, M. D., Wittebort, R. J., Sorai, M. and Hendrickson, D. N. (1990). J. Am. Chem. Soc., 112, 4214–4225. Luca, C., Popa, A., Bilba, N. and Mikhaila, G. (1983). Rev. Roum. Chem., 28, 211–216. MacIntosh, M. R., Frazer, B., Gruwel, M. L. H., Wasylishen, R. E. and Cameron, T. S. (1992). J. Phys. Chem., 96, 8572–8577. McAdie, H. G. (1962). Can. J. Chem., 40, 2195–2203. McLaughlin, R. L. and McClenahan, W. S. (1952). J. Am. Chem. Soc., 74, 5804–5805. Marking, G. A. and Kanatzidis, M. G. (1995). Chem. Mater., 7, 1915–1921. Marsh, R. E., Kapon, M., Hu, S. and Herbstein, F. H. (2002). Acta Cryst., B58, 62–77. Mayo, S. C., Proffen, Th., Bown, M. and Welberry, T. R. (1999). J. Appl. Cryst., 32, 464–471. Miki, K., Kasai, N., Shibakami, M., Chirachanchai, S., Takemoto, K. and Miyata, M. (1990). Acta Cryst., C46, 2442–2445. Miki, K., Kasai, N., Shibakami, M., Takemoto, K. and Miyata, M. (1991). J. Chem. Soc. Chem. Comm., pp. 1757–1759. Miki, K., Masui, A., Kasai, N., Miyata, M., Shibakami, M. and Takemoto, K. (1988). J. Am. Chem. Soc., 110, 6594–6596. Miyata, M. and Sada, K. (1996). ‘‘Deoxycholic acid and related hosts,’’ in Solid State Supramolecular Engineering; Crystal Engineering, Volume 6 of Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, pp. 147–176. Mu¨ller, R., Edwards, A. J., Prout, K., Simpson, W. M. and Heyes, S. J. (2000). Chem. Mater., 12, 1314–1322 Muller, R. and Meier, G. (1964a). Z. anorg. allgem. Chem., 332, 81–94. Muller, R. and Meier, G. (1964b). Z. anorg. allgem. Chem., 332, 281–286. Nakai, T., Terao, T., Imashiro, F. and Saika, A. (1986). Chem. Phys. Lett., 132, 554–557. Nakano, K., Katsuta, M., Sada, K., and Miyata, M. (2001). Cryst. Eng. Comm., No. 11. Nakano, K., Sada, K. and Miyata, M. (1994). Chem. Letts., pp. 137–140. Nakano, K., Sada, K. and Miyata, M. (1995). Chem. Commun., pp. 953–954. Nakano, K., Sada, K. and Miyata, M. (1996). Chem. Commun., pp. 989–990. Nakano, K., Sada, K., Kurozumi, Y. and Miyata, M. (2001). Chem. Eur. J., 7, 209–220. Nassimbeni, L. R., Niven, M. L., Stuart, D. A. and Zemke, K. J. (1986). J. Cryst. Spectr. Res., 16, 557–567. Nicolaides, N. and Laves, F. (1954). J. Am. Chem. Soc., 76, 2596–2597.

REFERENCES

317

Nicolaides, N. and Laves, F. (1956). J. Am. Chem. Soc., 78, 6416–6417. Nicolaides, N. and Laves, F. (1958). J. Am. Chem. Soc., 80, 5752–5756. Nicolaides, N. and Laves, F. (1963). J. Am. Oil Chem. Soc., 40, 400–413. Nicolaides, N. and Laves, F. (1965). Z. Kristallogr., 121, 283–296. Norman, N. and Mathisen, H. (1972). Acta Chem. Scand., 26, 3913–3916. Nye, J. F. (1957). Physical Properties of Crystals. Oxford, Clarendon Press. Otto, J. (1972). Acta Cryst., B28, 543–551. Padmanabhan, K., Venkatesan, K. and Ramamurthy, V. (1984). Can. J. Chem., 62, 2025–2028. Padmanabhan, K., Venkatesan, K. and Ramamurthy, V. (1987). J. Incl. Phenom., 5, 315–323. Parsonage, N. G. and Pemberton, R. C. (1967). Trans. Farad. Soc., 63, 311–328. Parsonage, N. G. and Staveley, L. A. K. (1984). ‘‘Thermodynamic studies of inclusion compounds’’. in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Academic Press, London etc., Vol. 3, pp. 1–36. Pemberton, R. C. and Parsonage, N. G. (1965). Trans. Farad. Soc., 61, 2112–2121. Pemberton, R. C. and Parsonage, N. G. (1966). Trans. Farad. Soc., 62, 553–557. Pierce-Butler, M. A. (1984). Acta Cryst., B38, 3097–3100. Polonski, T., Szyrszyng, M., Gdeniec, M., Nowak, E. and Herman, A. (2001). Tetrahedron Asymmetry, 12, 797–800. Popovitz-Biro, R., Chang, H. C. Tang, C. P., Shochet, N. R., Lahav, M. and Leiserowitz, L. (1980). Pure Appl. Chem., 52, 2693–2704. Popovitz-Biro, R., Tang, C. P., Chang, H. C., Lahav, M. and Leiserowitz, L. (1985). J. Am. Chem. Soc., 107, 4043–4058. Poupko, R., Furman, E., Mu¨ller, K. and Luz, Z. (1991). J. Phys. Chem., 95, 407–413. Pratt, D. S. and Perkins, G. A. (1918). J. Am. Chem. Soc., 40, 198–214. Rabinovich, D. and Hope, H. (1980). Acta Cryst., A36, 670–678. Radell, J. and Hunt, P. D. (1958). J. Am. Chem. Soc., 80, 2683–2685. Radell, J., Bradman, B. W. and Bergmann, E. D. (1964). Can. J. Chem., 42, 1069–1072. Redlich, O., Gable, C. M., Beason, L. R. and Millar, R. W. (1950). J. Am. Chem. Soc., 72, 4161–4162. Redlich, O., Gable, C. M., Dunlop, A. K. and Millar, R. W. (1950). J. Am. Chem. Soc., 72, 4153–4160. Rennie, A. J. O. and Harris, K. D. M. (1990). Proc. Roy. Soc. Lond., A430, 615–640. Rennie, A. J. O. and Harris, K. D. M. (1992). J. Chem. Phys., 96, 7117–7124. Rheinboldt, H., Flume, E. and Ko¨nig, O. (1929). Hoppe-Seyler’s Z. physiol. Chem., 180, 180–186. Rheinboldt, H., Ko¨nig, O. and Flume, E. (1929). Hoppe-Seyler’s Z. physiol. Chem., 184, 219–224. Rheinboldt, H. and Lauber, A. (1929). Hoppe-Seyler’s Z. physiol. Chem., 182, 255–258. Rheinboldt, H., Pieper, H. and Zervas, P. (1927). Liebigs Ann. Chem., 451, 256–273. Rizkallah, P. J., Harding, M. M., Lindley, P. F., Aigner, A. and Bauer, A. (1990). Acta Cryst., B46, 262–266. Rutherford, J. S. and Calvo, C. (1969). Z. Kristallogr., 128, 229–258. Sada, K., Kondo, T., Miyata, M. and Miki, K. (1994). Chem. Mater., 6, 1103–1105. Sada, K., Kondo, T., Miyata, M., Tamada, T. and Miki, K. (1993). J. Chem. Soc., Chem. Commun., pp. 753–755. Sada, K., Kondo, T., Ushioda, M., Matsuura, Y., Nakano, K.Miyata, M. and Miki, K. (1998). Bull. Chem. Soc. Jpn., 71, 1931–1937. Scarbrough, F. E. and Voet, D. (1976). Acta Cryst., B32, 2715–2717. Schiessler, R. W. and Flitter, D. (1952). J. Am. Chem. Soc., 74, 1720–1723. Schlenk, W., Jr. (1949). Ann. Chem., 565, 204–240. Schlenk, W., Jr. (1951). Ann., 573, 142–162. Schlenk, W., Jr. (1955). Svensk. Kem. Tidsskr., 67, 435–462.

318


Schlenk, W. Jr. (1965). Angew. Chem. Int. Ed. Engl., 4, 139–145. Schlenk, W., Jr. (1968). Chem. Ber., 101, 2445–2449. Schlenk, W., Jr. (1973a). Liebig’s Ann. Chem., pp. 1145–1155. Schlenk, W., Jr. (1973b). Liebig’s Ann. Chem., pp. 1156–1178. Schlenk, W., Jr. (1973c). Liebig’s Ann. Chem., pp. 1195–1209. Scott, J. L. (1995). J. Chem. Soc., Perkin 2, pp. 495–502. Scott, J. L. (1996a). Supramol. Chem., 7, 201–207. Scott, J. L. (1996b). J. Chem. Cryst., 26, 185–189. Scott, J. L. (1997a). Supramol. Chem., 8, 231–239. Scott, J. L. (1997b). Supramol. Chem., 8, 241–251. Sekii, M., Matsuo, T. and Suga, H. (1990). J. Incl. Phenom. and Mol. Recogn. in Chem., 9, 243–251. Sheldrick, W. S. and Wachtold, M. (1995). Angew. Chem. Int. Ed. Engl., 34, 450–452. Shibakami, M. and Sekiya, A. (1994a). J. Incl. Phenom. and Mol. Recogn. in Chem., 16, 397–412. Shibakami, M. and Sekiya, A. (1994b). Chem. Commun., pp. 429–430. Shibakami, M., Tamura, M. and Sekiya, A. (1995a). J. Incl. Phenom. and Mol. Recogn. in Chem., 22, 155–168. Shibakami, M., Tamura, M. and Sekiya, A. (1995b). J. Am. Chem. Soc., 117, 4499–4505. Shibakami, M., Tamura, M. and Sekiya, A. (1995c). J. Incl. Phenom. and Mol. Recogn. in Chem., 22, 299–311. Shindo, T., Shindo, M., Ohnuma, H. and Kabuto, K. (1993). Bull. Chem. Soc. Jpn., 66, 1914–1918. Sluis, P. L. van der, and Spek, A. L. (1990). Acta Cryst., C46, 883–886. Smith, A. E. (1950). J. Chem. Phys., 18, 150–151. Smith, A. E. (1952). Acta Cryst., 5, 224–235. Smith, R. L. (1973). The Excretory Function of Bile, Chapman and Hall, London. Snyder, R. G., Scherer, J. R. and Gaber, B. P. (1980). Biochim. Biophys. Acta, 601, 47–53. Sobotka, H. (1934). Chem. Revs., 15, 311–375. Sorai, M. and Shiomi, Y. (1986). Thermochim. Acta, 109, 29–44. Sorai, M., Ogasahara, K. and Suga, H. (1981). Mol. Cryst. Liq. Cryst., 73, 231–254. Sozzani, P., Comotti, A., Bracco, S. and Simonutti, R. (2004). Chem. Commun., 768–769. Sozzani, P., Comotti, A., Bracco, S. and Simonutti, R. (2004). Angew. Chem. Int. Ed. Engl., 43, 2792–2797. Suehiro, K. (1988). J. Incl. Phenom., 6, 9–13. Sun, C., Zhou, D., Grant, D. J. W. and Young, V. G. Jr. (2002). Acta Cryst., E58, o368–o370. Sutor, D. J. (1958a). Acta Cryst., 11, 83–87. Sutor, D. J. (1958b). Acta Cryst., 11, 453–458. Swaminathan, S., Craven, B. M. and McMullan, R. K. (1984). Acta Cryst., B40, 300–306. Takahashi, I., Onodera, A. and Shiozaki, Y. (1990). Acta Cryst., B46, 661–664. Takemoto, K. and Sonoda, N. (1984). ‘‘Inclusion compounds of urea, thiourea and selenourea.’’ in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Academic Press, London, Vol. 2, pp. 47–67. Tam, W., Eaton, D. F., Calabrese, J. C., Williams, I. D., Wang, Y. and Anderson, A. G. (1989). Chem. Mat., 1, 128–140. Tang, C. P. (1979). Ph. D. Thesis, Feinberg Graduate School, The Weizmann Institute of Science, Rehovoth, Israel. Tang, C. P., Popovitz-Biro, R., Lahav, M. and Leiserowitz, L. (1979). Isr. J. Chem., 18, 385–389. Tang, C. P., Chang, H. C., Popovitz-Biro, R., Frolow, F., Lahav, M. Leiserowitz, L. and McMullan, R. K. (1985). J. Am. Chem. Soc., 107, 4058–4070. Terasaki, O., Shiokawa, K., Ito, M., Watanabe, D. and Thomas, J. M. (1989). In ‘‘Zeolites: Facts, figures, future,’’ edited by P. A. Jacobs and R. A. van Santen, Elsevier, Amsterdam, pp. 431–438.

REFERENCES

319

Teter, J. W. and Hettinger, W. P., Jr. (1955). J. Am. Chem. Soc., 77, 6695. Umemoto, K. and Danyluk, S. S. (1967) J. Phys. Chem., 71, 3757–3764. Ung, A. T. (1993). Ph. D. Thesis, University of New South Wales, Sydney, Australia. Ung, A. T., Bishop, R., Craig, D. C., Dance, I. G., Rae, A. D. and Scudder, M. L. (1993). J. Incl. Phenom. and Mol. Recognit. Chem., 15, 385–398. Ung, A. T., Bishop, R., Craig, D. C., Dance, I. G. and Scudder, M. L. (1991). J. Chem. Soc. Chem. Comm., pp. 1012–1014. Ung, A. T., Bishop, R., Craig, D. C., Dance, I. G. and Scudder, M. L. (1992a). Struct. Chem., 3, 59–61. Ung, A. T., Bishop, R., Craig, D. C., Dance, I. G. and Scudder, M. L. (1992b). J. Chem. Soc., Perkin 2, pp. 861–862. Ung, A. T., Bishop, R., Craig, D. C., Dance, I. G. and Scudder, M. L. (1993a). J. Chem. Soc. Chem. Comm., pp. 322–323. Ung, A. T., Bishop, R., Craig, D. C., Dance, I. G. and Scudder, M. L. (1993b). Tetrahedron, 49, 639–648. Ung, A. T., Bishop, R., Craig, D. C., Dance, I. G. and Scudder, M. L. (1994). Chem. Mater., 6, 1269–1281. Ung, A. T., Gizachew, D., Bishop, R., Scudder, M. L., Dance, I. G. and Craig, D. C., (1995). J. Am. Chem. Soc., 117, 8745–8756. Volkov, A., Huizhe, Z., Shuncheng, L., White, M. A. and Coppens, P. (1998). Acta Cryst., C54, IUC980026 (CIF access paper). Weber, E. and Jossel, H.-P. (1983). J. Incl. Phenom., 1, 79–85. Weber, T., Boysen, H. and Frey, F. (2000). Acta Cryst., B56, 132–141. Weisinger-Levin, Y., Vaida, M., Popovitz-Biro, R., Chang, H. C., Mannig, F., Fro;ow, F., Lahav, M. and Leiserowitz, L. (x985). Tetrahedron, 43, 1449–1457. Wells, A. F. (1938). Proc. R. Soc. Lond., Ser. A, 167, 169–189. Whitaker, R. D., Barreiro, A. J., Furman, P. A., Guida, W. C. and Stallings, E. S. (1968). J. Inorg. Nucl. Chem., 30, 2921–2925. Whitaker, R. D. and Guida, W. C. (1969). J. Inorg. Nucl. Chem., 31, 875–877. Whitaker, R. D., Simon, J. and Victor, D. (1971). J. Inorg. Nucl. Chem., 33, 2677–2678. Wieland, H. and Sorge, H. (1916). Hoppe-Seyler’s Z. fu¨r Physiolog. Chem., 97, 1–27. Wong, R. Y., Manners, G. D. and Palmer, K. J. (1977). Acta Cryst., B33, 970–974. Yeo, L. and Harris, K. D. M. (1997). Acta Cryst., B53, 822–830. Yeo, L., Harris, K. D. M. and Guillaume, F. (1997). J. Sol. State Chem., 128, 273–281. Yoswathananont, N., Chirachanchai, S., Tashiro, K., Nakano, K., Sada, K. and Miyata, M. (2001). CrystEngComm., No. 19, 1–4. Zimmerschied, W. J., Dinerstein, R. A., Weitkamp, A. W. and Marschner, R. F. (1950). Ind. Eng. Chem., 42, 1300–1306.


Chapter 7 Clathrate inclusion complexes formed by hosts of lesser versatility

So from the beginning man was conscious of shell, peel and skin. Also he had a highly developed hand directly usable as an enclosing device and basic to much achievement since then. Successor man has used this hand to make enclosures for himself and his family . . . some of these things are today repeated for the individual who is so familiar with the idea that he uses it in myth and metaphor, claiming, sometimes, to be cribbed, cabined and confined, bound in (but by his own fears) or, maybe, seeing himself or another as a prisoner (but of system, doctrine, conscience and the like). In one way or another everyone understands what it is like to be hemmed in. H. M. Powell, 1984

Summary : Cavities in frameworks constructed from linked host molecules can be filled (all or in part) by guest molecules of appropriate size and shape, thus forming clathrate inclusion complexes. In general (but there are some exceptions), there are no covalent or hydrogen bonds between host and guest and the host-guest interactions are due to dispersion and polar forces. Guest-guest interactions (often, but not always justifiably, neglected) are through dipolar and dispersion forces. In one group of clathrates the interaction among the host molecules is strongly directional, aiding in the formation of periodic arrays of cavities. One prevalent motif is the four-connected tetrahedral node found in the polyhedral clathrates, a group which includes the metalloid clathrates (with covalent bonding between pairs of framework atoms (Si, Ge, . . . )), the clathrasils (with ionic-polar bonding between SiO2 groups) and the gas hydrates (of different kinds), where there is hydrogen bonding between adjacent framework water molecules and, sometimes, additional linkaging. A remarkable concatenation of geometrical and chemical factors governs the formation of the frameworks of these clathrates. A second prevalent motif is the ring of six hydrogen-bonded hydroxyl groups found in the quinol and phenol clathrates, and those of Dianin’s compound. The statistical mechanics of the formation of solid solutions of guests in the host frameworks has been worked out and applied particularly to the -quinol clathrates and, in lesser detail, to the gas hydrates; a substantial number of phase diagrams have been determined for these systems. Most of these clathrates have crystal structures different from those of the pristine hosts and hence are intermediate phases in the binary phase diagrams; Dianin’s compound provides a fairly rare example of formation of a primary solid solution. Clathrates can also be formed when there are only nondirectional dispersion forces between host molecules, and the inclusion complexes of tetraphenylene are used as an example. Finally, the hydrogen-bonded motifs found in the quinol etc. clathrates have inspired the synthesis of covalently bonded analogs, and these hexahosts also form clathrates.

7.1 Introduction 7.2 Directionally bonded hosts

323 323

CL AT HRATE INCLUS ION COMP LEXES

322

7.2.1

Quinol 7.2.1.1 7.2.1.2 7.2.1.3

(hydroquinone; 1,4-dihydroxybenzene) as host Crystal structures of quinol polymorphs and -quinol clathrates Low temperature phase transitions in -quinol clathrates Introduction to statistical thermodynamics of clathrate structures and application to the quinol clathrates 7.2.2 Crystal structure of {(6H2O) [hexamethylene tetramine]} 7.2.3 Clathrates derived from existing structures 7.2.3.1 Helium hexahydrate 7.2.3.2 Cadmium cyanide clathrates 7.2.4 Overview of the polyhedral clathrates (including metalloid structures, clathrasils, gas hydrates, clathrate and semiclathrate hydrates) 7.2.4.1 Historical and general introduction 7.2.4.2 Restrictions on the shapes of the polyhedra 7.2.4.3 Packing of pentagonal dodecahedra 7.2.5 Metalloid structures 7.2.6 Clathrasils 7.2.7 Gas hydrates (structures with pentagonal dodecahedra) 7.2.7.1 Relation between guest type and structure type in the gas hydrates 7.2.7.2 Stoichiometry and thermodynamics of the gas hydrates 7.2.7.3 Prototype CS-I and CS-II crystal structures at low temperatures 7.2.7.4 Br2 86H2O is the only bromine hydrate, and the implications of this result 7.2.7.5 Gas hydrates with charged frameworks (ionic clathrate hydrates) 7.2.8 Peralkylonium hydrates and related structures 7.2.8.1 Introduction 7.2.8.2 Structures based on the CS-I structure 7.2.8.3 Structures based on the CS-II structure 7.2.8.4 Structures based on the HS-II structure 7.2.8.5 Structures based on the HS-I structure and its superstructure SHS-I 7.2.8.6 Structures based on the OS-I structure 7.2.8.7 Structures based on the TS-I structure 7.2.8.8 The effectiveness of the alkyl substituents in forming hydrates 7.2.9 Varieties of structures formed by a particular guest 7.2.10 The alkylamine hydrates 7.2.11 Structures without pentagonal dodecahedra (some with charged frameworks) 7.3 Hosts with a combination of directional bonds and van der Waals interactions 7.3.1 Phenol (and related compounds) as hosts 7.3.1.1 Phenol 7.3.1.2 Guayacanin as host 7.3.2 Dianin’s compound (4-p-hydroxyphenyl-2,2,4-trimethylchroman) and related compounds as hosts 7.4 Van der Waals linked hosts 7.4.1 Tetraphenylene as host 7.5 Hexahosts and related compounds 7.6 Conclusions and a perspective view References

323 323 331 333 345 346 346 347 348 348 353 355 360 363 370 370 372 379 381 383 383 383 384 385 385 385 387 387 389 389 389 392 396 396 396 398 399 406 406 408 410 411


323

7.1 Introduction Molecules of complex shape often pack together in crystals such that cavities remain between them; similar situations occur for molecules of simpler shapes joined by directed (often hydrogen) bonding. If the cavities remain empty then these arrangements are usually thermodynamically unstable with respect to a denser, cavity-free packing, but filling the cavities, even partially, leads to stabilization. In such clathrate structures the guest component is entirely enclosed in a framework formed by the host. It is geometrical rather than chemical factors which determine whether a given molecular species can be enclathrated in a particular host framework. The ‘‘clathrate’’ nomenclature was introduced by H. M. Powell in 1948 (see Chapter 2 and Davies (1999)). One type of clathrate complex – the gas hydrates – was first encountered almost two hundred years ago and a second – the quinol clathrates – almost a century and a half ago, although they were only characterized as such some fifty-odd years ago. The intermolecular bonding in both the water and quinol frameworks is via hydrogen bonds, but van der Waals interactions are often sufficient for the formation of frameworks stabilized by the presence of guests, e.g. in the tri-o-thymotide (Section 8.2) and tetraphenylene clathrates. Both types of interaction occur in the complexes of phenol while there are directed covalent interactions in the metalloid structures and clathrasils. The literature is very extensive; we note here only one of the earlier books (Hagan, 1962) and some general reviews (Parsonage and Staveley, 1984; Davidson and Ripmeester, 1984; Sixou and Dansas, 1976; Ripmeester and Ratcliffe, 1991) while more specific references are given in the body of the text. This chapter has been arranged so that a substantial part of the experimental information is summarized before the statistical thermodynamical treatment of the stability of the clathrates is discussed in detail. Relatively little attention is paid here to the motions of guests (and hosts) and the reader is referred to the more extensive coverage by Sixou and Dansas (1976) and Ripmeester and Ratcliffe (1991). As noted earlier, the integration of structural results (from diffraction) with thermodynamical and statistical thermodynamical principles has been developed more substantially in the field of the clathrates than for any other group of inclusion structures. The pioneering work of van der Waals and Platteeuw has been very substantially advanced by the Canadian (Davidson, Ripmeester) and Siberian (Dyadin) schools. After some trepidation, we have decided that the most transparent means of introducing this material is to provide a general introduction immediately after discussing the structural results for the quinol clathrates, and then to use the theory and associated experimental measurements to discuss the stability of these materials. Later applications to the clathrate hydrates will be made after having described the structural background for these materials. 7.2 Directionally bonded hosts 7.2.1 Quinol (hydroquinone; 1,4-dihydroxybenzene) as host 7.2.1.1 Crystal structures of quinol polymorphs and -quinol clathrates Although the first quinol clathrate was prepared by Wo¨hler (1849) some 150 years ago, the principles underlying the structures in this family of complexes were clarified only after

324


the end of the Second World War by crystallographic (Palin and Powell, 1947, 1948) and thermodynamic studies (van der Waals and Platteeuw, 1959); there are excellent introductions to the subject (Powell, 1964, 1984; MacNicol, 1984a) and a more recent comprehensive account (Mak and Bracke, 1996). The quinol system appears to be monotropic with -quinol stable up to the melting point, but there are two additional polymorphs, both metastable under ambient conditions. The first of these is -quinol, which can be crystallized from n-octane as an empty clathrate structure (Lindemann, Shklover and Struchkov, 1981); another method is by recrystallization from air-free propanol in the presence of a seed of the argon clathrate, giving almost pure -quinol (99.8%) (Evans and Richards, 1952). In general the -quinol structure requires stabilization by addition of suitable guest molecules. The second is -quinol, which does not have any clathrating properties. We first consider the structures of the polymorphs and their inter-relations and then pass on to the structures and thermodynamics of the clathrates. The crystal data for the three polymorphs are given in Table 7.1; the - and -polymorphs have much the same density of packing but this is considerably less in (empty) -quinol. The relationships among the structures of the three polymorphs of quinol are quite remarkable. The -polymorph (Wallwork and Powell, 1980) has a structure consisting of two parts, linked together (Fig. 7.1). One part is a double helix of chains of hydrogen bonded molecules around threefold screw (31) axes, and the other is an interpenetrating cage structure of hydrogen bonded molecules around centres of symmetry on threefold inversion ( 3) axes. The interpenetrating cage structure is very similar to that of the -quinol clathrates described below and provides an explanation for the fact that -quinol has been shown to take up small amounts of gaseous guests; the structure is such that the maximum stoichiometry is one guest per 18 quinol molecules. Such ‘‘structural parsimony’’ – the use of structural elements in more than one context – is found in other systems, for example trimesic acid and its polymorphs and complexes (Section 8.4).

Fig. 7.1. Stereodiagram of part of the crystal structure of -quinol – the interpenetrating cage structure so similar to that in -quinol is shown on the right hand side of the diagram and the double helix of quinol molecules on the left, the line indicating the 31 axis. (Reproduced from MacNicol (1984a).)


325

Fig. 7.2. ORTEP sterodiagram of the -quinol clathrate of HCl. Hydrogen bonds are stippled. The whole unit cell is shown (but not the H atom of HCl), the direction of view being approximately perpendicular to that in Fig. 7.3. (Reproduced from Boeyens and Pretorius (1977).)

Fig. 7.3. Schematic diagram of parts of the two interpenetrating frameworks of the -quinol structure. The larger circles represent OH groups, linked together in hexagons by hydrogen bonds shown as horizontal lines. The benzene rings bridging the hydrogen bonded hexagons are shown. An acetonitrile molecule is shown enclosed in the cage formed by the interpenetration of the two frameworks. (Reproduced from Mak (1982).)

The structure of the empty -polymorph is analogous to that of the HCl clathrate shown in Fig. 7.2; the -quinol framework is centrosymmetric but the crystal as a whole will be polar if the guest is in a polar arrangement, as for HCl and CH3OH, or centrosymmetric if the guest arrangement is also centrosymmetric (either because of inherent symmetry or disorder). The first structure of this series to be determined was that of the methanol clathrate. The structure is based on the formation of two interpenetrating frameworks (Fig. 7.3) which are not bonded to each other – in other words they are catenated

326


(cf. Section 8.3). The hydroxyl groups of six quinol molecules form a planar hexagon with alternate quinol molecules pointing diagonally above and below the hexagon. Interpenetration of two such frameworks gives an overall structure with cavities occupied by guest molecules. Details of the cage structure are shown by the relationship of the two partial networks (Fig. 7.3). We use our usual nomenclature for inclusion complexes – {n(host) [m(guest)]}. The third quinol polymorph, the -form, has a rather different structure with spirals of hydrogen bonded quinol molecules formed about two fold screw axes, the resulting sheets being weakly linked by van der Waals forces to give an easily cleaved layer structure. It is not included as it is not relevant in the present context. The crystallization diagram (not a phase diagram as equilibrium was not attained) of the system quinol–SO2–H2O at about room temperature has been reported (Chekhova, Polyanskaya, Dyadin and Alekseev, 1975) and shows a primary solid solution of -quinol containing up to 2 mol% SO2 and a range of -quinol solid solutions containing from 13.5–31.5 mol% SO2. The -phase can contain He, Ne, Ar, Xe, CH4, N2, O2 or SO2 in solid solution (for references see Chekhova et al. (1975)), although in much more restricted amount than in -quinol; a crystal of composition {0.021SO2 p-C6H4(OH)2} has about the limiting amount and its cell dimensions are very close to ˚ less in a and 0.01 A ˚ larger in c) although the volume per those of -quinol itself (0.02 A formula unit is, surprisingly, a little less; however, these differences may well not be significant. Quinol–guest appears to be the only system encountered so far that shows both primary solid solution of guest in pristine host and formation of a separate clathrate phase (cf. introduction to Part III). The -quinol clathrates have been found to crystallize in three groups of isomorphous structures, which form an isostructural family (Table 7.1). Among the guests are Ar, Kr, Xe, HCOOH, CO2, HBr, C2H2, CH4, S2O (Kutty, Sharma and Murthy, 1971), N2O, C3H8 (Peyronel and Barbieri, 1958), as well as mixtures such as CH3OH þ SO2 and SO2 þ HCl. The composition given above for the -phase clathrate of SO2 does not take into account the possible presence of water – for example, a crystal of the -phase with cell parameters ˚ , dm ¼ 1.458 g cm3, Z ¼ 3 can be calculated to have the a ¼ 16.308(5), c ¼ 5.810(10) A composition {3(quinol) [0.2H2O 0.9SO2]}; analysis gave 0.6H2O, the discrepancy presumably being due to occluded water (Lindemann, Shklover and Struchkov, 1981). When the guest molecules are themselves either quasi-spherical or disordered then the cavities remain quasi-spherical, giving Type I complexes (space group R3). Distortion occurs when the guest interacts with the framework, as do HCl and methanol, giving Type II complexes (space group R3). The cavities are all of the same kind and size in complexes of Types I and II but this no longer holds for Type III complexes (space group P3). In the CH3CN clathrate the ordered anisotropic guest leads to significant changes in cell dimensions and reduction in lattice symmetry giving rise to three crystallographically independent cavities, with one guest molecule pointing in the opposite direction to the other two. -{3(quinol) [CH3NC]} presents a problem in classification – in terms of space group it belongs in Type II but unit cell shape would place it in Type III. Palin and Powell (1948) have commented ‘‘The slight variation in the unit cell dimensions for the different compounds are related to the dimensions of the enclosed molecule . . . a ¼ 16.58, c ¼ 5.47 . . . represent the natural dimensions determined by the equilibrium of the parts of the -quinol framework alone . . . Simultaneously with increase in c there is in nearly

Table 7.1. Crystal data for the polymorphs of quinol and the - and -quinol clathrates. The classification of the -quinol clathrates as Types I, II, III follows MacNicol (1984a) and is based on space group Compound/Reference/ REFCODE Quinol Polymorphs -quinol WP80; HYQUIN02 -quinol (Type I) LSS81; HYQUIN05 -quinol M-M66

Crystal class

˚) a (A

˚) c (A

Vol./formula ˚ 3)# unit (A

Z

dm g cm3

Space group

R*

38.46

5.65

134.0

54

1.35

R 3

R

16.613

5.475

145.4

9

1.26

R 3

4

1.33

P21/c

1.366

R 3

˚ 3; ˚ , ¼ 107 , V ¼ 132.5 A monoclinic, a ¼ 8.07, b ¼ 5.20, c ¼ 13.20 A pairs of molecules at independent centers of symmetry.

- and -Quinol Clathrate Inclusion Complexes -{quinol [0.016SO2]} R 38.26 Type I -{3(quinol) [0.62C2H2]} PP48; ZZZVKY -{3(quinol) [0.74CO2]} PP47; ZZZVLW -{3(quinol) [0.9SO2]} PP48 -{3(quinol) [HCOOH]} PP47; ZZZVLO -{3(quinol) [0.87Xe]} BFSV89; JAMKEN -{3(quinol) [0.87H2S]} HM83; ZZZVLG11 Type II -{3(quinol) [0.89 CH3OH]} M82; ZZZVLI01

5.605

132.3

54

R

16.63

5.46

145.3

3

R 3

R

16.17

5.82

146.4

3

R 3

R

16.29

5.81

148.4

3

R 3

R

16.42

5.65

146.6

3

R 3

R

16.610

5.524

146.6

3

R 3

R

16.67

5.518

147.6

3

R 3

R

16.62

5.562

147.9

3

R3

Table 7.1. (Continued ) dm g cm3

Compound/Reference/ REFCODE

Crystal class

˚) a (A

˚) c (A

Vol./formula ˚ 3)# unit (A

Z

-{3(quinol) [0.87 HCl]} BP77; HYQHCL (XRD); HYQHCL01(ND) -{3(quinol) [0.884 SO2]} GPDA75, PA82; QUOLSO01 x -{3(quinol) [0.9 SO2 0.6 H2O]}; PA82; ZZZAPG x -{3(quinol) [CH3NC]} CM83; BUSPAG

R

16.65

5.453

148.3

3

R

16.311

5.821

149.0

3

1.46

R3

R

16.31

5.810

155.3

3

1.46

R3

R

15.95

6.348

155.4

3

trigonal

16.003

6.245

153.9

3

1.34

P3

R

16.215

13.846

350.3

3

1.65

R 3m

trigonal

17.102 (5)

23.904 (9)

336.4

4

1.54

P 3m1

Type III -{3(quinol) [CH3CN]} CM83; HQUACN01 Miscellaneous 3{(quinol) [C60]} E91,; SOMGIK {4.5 (quinol) [C70] [C6H6]} ER93; HASWIH

Space group R3

R3

Notes: x The SO2 clathrate analyzed by Palin and Powell (1948) was reported to be Type I. The original 300K study of SOMGIK had R ¼ 0.196; this has been superceded by SOMGIK01 to 04 (BRS00) at 100, 200, 300 and 373K, with R 0.04. * R ¼ rhombohedral. References: BFSV89 – Birchall, Frampton, Schrobilgen and Valsdo´ttir (1989); BP77 – Boeyens and Pretorius (1977); BRS00 – Blanc, Restori, Schwartzenbach et al., 2000; CM83 – Chan and Mak (1983); E91 – Ermer (1991); ER93 – Ermer and Ro¨bke (1993); GPDA75 – Chekhova et al., 1975; HM83 – Ho and Mak (1983) (7); LSS81 – Lindemann, Shklover and Struchkov (1981); M82 – Mak (1982); M-M66 – Maartman-Moe (1966); PA82 – Polyanskaya, Alekseev, Bakakin and Chekhova (1982); PA82 – Polyanskaya, Andrianov, Alekseev, Bakakin, Dyadin, and Chekhova (1982); PP47 – Palin and Powell (1947); PP48 – Palin and Powell (1948); WP80 – Wallwork and Powell (1980).


329

every case a decrease in a . . . ’’ Inspection of comparable unit cell dimensions reported for various -quinol clathrates suggests that considerable caution must be exercised before drawing conclusions from their variations; the measurements are of various degrees of precision and the compositions are not always unequivocal. Calculation of unit cell dimensions is discussed below. The single framework shown in Fig. 7.3 has been termed a super-cube or a superpolonium network by Ermer (1991); we remind the reader that Po is the only element to crystallize in a simple cubic structure. In this nomenclature the centres of the (OH)6 rings are the corners of the cubes and the p-phenylene groups are the cube edges; there is an appreciable trigonal distortion in {3(HQ) [C60]}. The cavities in such a single network ˚ , which is just the diameter of the C60 football. Simple mixing of have a diameter of 10 A quinol and C60 in 3 : 1 ratio in benzene and slow evaporation gave large (up to 7 mm in length), shiny black (dark brownish red in transmission) crystals of the 3 : 1 complex. Their structure (Fig. 7.4) is just as anticipated, but with remarkable additional detail. The C60 molecules are contained in a single network of the same type as found in -quinol but considerably distorted – for example, the hydrogen-bonded hexagons that are planar in -quinol are here almost as puckered as in cyclohexane (both have D3d symmetry, and bond and torsion angles are respectively 112.9 , 50.5 ({3(HQ) [C60]}) and 111.4 , 54.9 (C6H12)) and the cages are considerably elongated (compared to those in -quinol) in order to accommodate the fullerene guest. The stability of the complex, and its colour, are ascribed to charge transfer interactions between quinol as -donor and C60 as -acceptor. One cannot do better than quote Ermer : ‘‘The ‘giant’ C60 molecules enforcing the single host network . . . have a molecular mass 2.2 times larger than that of the three HQ molecules, and occupy 50% of the crystal volume. (HQ)3C60 represents one of the rare inclusion compounds with more guest than host . . . One is irresistibly reminded to a tropical snake devouring a prey of disproportionately larger cross-section than that of her (sic!) normally slim body . . . the question arises whether the labels ‘host’ and ‘guest’ should be reversed . . . ’’

5Å

5Å

Fig. 7.4. Ball-and-stick stereoview of the super-polonium architecture of {3(HQ) [C60]}. Note particularly the puckered rings of oxygen atoms and contrast these with the analogous planar rings in Figs. 7.2 and 7.3. (Reproduced from Ermer (1991).)


330

Ermer (1991) found that the [C60] molecule was disordered in its enclosure, the molecular site symmetry being 2/m, and the details have now been studied further by determining the crystal structure at 100, 200, 293 and 373K (Blanc et al., 2000). No phase transition was detected over this range of temperatures. The most favourable orientation (a)

(b)

(c)

Fig. 7.5. Stereoviews of the three H-bonded quinol cages found in {4.5(HQ) [C70] [C6H6]}. The diagrams are all to the same scale. The H-bonds are shown as open sticks. (a) Twin peanut enclosure of two A type C70 molecules; (b) Single enclosure of B type C70 molecule; (c) Tetrahedral enclosure containing dimeric benzene sandwich. (Reproduced from Ermer and Ro¨bke (1993).)


c

o a

331

c

o

b

b

a

Fig. 7.6. A stereoscopic schematic representation of the H-bonded superoctahedral quinol host network of {4.5(HQ) [C70] [C6H6]}. The unit cell edges (O, a, b, c) are shown by thinner lines. The small spheres symbolize the (OH)6 rings (at their midpoints) acting as octahedral centres and the thick rods represent the HQ p-phenylene bridges. A twin cage (‘‘peanut’’) is shown on the left with its two A-type C70 guests and a single cage at the top with a single C70 guest. The tetrahedral cage containing the twin benzene sandwich is shown at center front. Single orientations of the actually disordered guest molecules are shown. (Reproduced from Ermer and Ro¨bke (1993).)

of [C60] with respect to the phenyl rings of HQ has [C60] C–C bonds pointing to the centres of the phenyl rings. However, this does not match the 3m symmetry of the cage, hence the orientational disorder. The remarkable {3(HQ).[C60]} structure is outshone by the even more remarkable {4.5(HQ) [C70] [C6H6]} structure. Mixing of quinol and C70 in 6 : 1 ratio in benzene and slow evaporation gave shiny black hexagonal plates, with a clear brownish red colour in transmission; crystal data are given in Table 7.1. Two C70 molecules are enclosed in the peanut shaped cavity shown in Fig. 7.5(a). One C70 molecule is in a medium-sized single cage (Fig. 7.5(b)) and two benzene molecules are in a tetrahedral cage (Fig. 7.5(c)), although this requires confirmation. The combinatory arrangement of these three enclosures to give the crystal is shown in Fig. 7.6. The enclosing frameworks are made up of quinol molecules with their hydroxyl groups linked to form six-rings which are appreciably distorted from planarity. The basic structural principles employed in the simple -quinol clathrates are carried over to the two quinol-fullerene complexes, but with dramatic extensions. 7.2.1.2 Low temperature phase transitions in -quinol clathrates There are phase transitions on cooling in some -quinol clathrates (for summary see Parsonage and Staveley (1984, pp. 33–34) and also Matsuo and Suga (1984)). Thus {3(quinol) [HCN]} has a sharp peak in the Cp–T curve at 177.8 0.1K, with S 0.69 R corresponding to a disordering of two orientations (ln 2 ¼ 0.693) (Matsuo, Suga and Seki, 1970), and the {3(quinol) [0.95(H2S)]} complex a peak at 7.6K (the only one in the range 1–300K) (Ukegawa, Matsuo and Suga, 1985). The (calorimetrically determined) transitions in {3(quinol) [x(CH3OH)]} depend on composition and are absent for x5 *0.55; the temperatures of the transition are inversely dependent on composition, Tc ¼ 65.7, 61.0,


332

54.4 and 44.4K for x ¼ 0.974, 0.897, 0.837 and 0.728 (Matsuo, 1970). Framework and guest behaviour in {3(quinol) [0.98(CH3OH)]} have been studied in considerable detail (Matsui, Terao and Saika, 1982) over the range 4.2–363K by a combination of 13C high resolution solid state NMR spectroscopy and measurements of T1 proton relaxation times (the sample composition was not given explicitly but has been inferred here from the Tc value of 67K); single crystals of size 3 5 10 mm were used. Above 67K a single and axially symmetric 13C chemical shift tensor was observed for the trapped methanol, showing that the molecules reorient themselves among three potential minima in the cavity. Analysis of the T1 results gave the motional parameters for the C3 reorientation as Ea ¼ 2.9 0.8 kJ/mol, t0 ¼ 3.7 1012 s, and for inversion about the centre of the cavity as Ea ¼ 10.0 0.4 kJ/mol, t0 ¼ 2.5 1014 s. Each 13C absorption line for both quinol and methanol splits into three below 67K, indicating that the methanol molecules ordered into six orientations while the symmetry of the quinol sublattice was reduced below R3. The directions and librational frequencies of the C–O bond axes were determined from the

TP

θ (degrees)

40 C-axis θ

C–O axis

θ

30

0

100

(a)

C-axis

1

2

200 T (K)

(b) 3

θ

300

C-axis

O θ

H

δ(CO)

C–O bond

P δ(Z)

6

5

4

H

C HH

Fig. 7.7. -{3(quinol).[CH3OH]} clathrate : (a) the temperature dependence of the angle between the c axis and the C–O bond direction of the guest methanol molecule; (b) details of the positioning of the methanol guest. (Reproduced from Matsui, Terao and Saika (1982).)


333

13

C methanol chemical shift tensor; ordering of the methanol dipoles, methyl tunneling and residual C3 reorientational motion were also investigated. The temperature dependence of the angle between the C–O bond of the methanol and the c-axis of the crystal is shown in Fig. 7.7. There is good agreement between these NMR results and room temperature X-ray diffraction results (Mak, 1982) which show that the C–O bond of the methanol is tilted by 35 from the c axis giving close contact with three of the oxygen atoms of the sixmembered OH ring; angles of 40 at 300K and 32 below 100K were deduced from dielectric susceptibility measurements (Ripmeester, Hawkins and Davidson, 1979). The structure below the phase transition is ferroelectric with the spontaneous polarization along the c axis (Murakami, Komukae, Osaka and Makita, 1990). Similar structural results were obtained for the HCl clathrate where the Cl atoms interact with the hydroxyl oxygens and the HCl molecule lies on a cone with its generator inclined at 33 to the c axis (Boeyens and Pretorius, 1977). However, spectroscopic, calorimetric and dielectric susceptibility measurements show a large degree of angular freedom of the guest down to the lowest temperatures (Matsui, Terao and Saika, 1982), so there appears to be a contradiction between the diffraction and other measurements. No phase transitions have been found in the HCl quinol clathrate in the range 1.5–300K. In the isomorphous SO2 clathrate there is also threefold disorder of the guest but only weak interaction with the framework. Here one set of dielectric constant measurements indicates a phase change at 40K (on cooling, there being some hysteresis) (Schneider, Tornau, Vlasova and Gurskas, 1985), but not another (Davidson, Davies, Gough, Leaist and Ripmeester, 1990). Having studied a variety of samples, the latter authors ‘‘emphasize the need for well-characterized, uniform samples if meaningful results are to be obtained and interpreted.’’ 7.2.1.3

Introduction to statistical thermodynamics of clathrate structures and application to the quinol clathrates We begin with a very simple physical model. Consider a crystalline molecular host framework containing empty cavities of one kind, and allow this to absorb a monatomic gaseous guest, with one molecule (atom) in each occupied cavity but not all cavities necessarily occupied; the guest molecules do not interact with one another. A solid solution is formed, the stability of which will be determined by the interplay of enthalpy and entropy, G ¼ H – TS having to be negative for the solid solution to be stable. A necessary requirement for formation of the solid solution is that the interaction between the molecules of the guest and the cavity walls should be exothermic (H negative). The entropy change is determined by a number of factors; there will be a large and negative contribution from the entropy difference between freely translating gas molecules and those confined within the cavities. If the cavities are not completely filled then the negative change in translational entropy will be partially offset by a positive configurational entropy resulting from the distribution of gas molecules over the cavities. The equilibrium will be temperature and pressure dependent. This model is analogous to Langmuir’s model of monolayer gas adsorption on a solid surface (Berry, Rice and Ross, 1980), as was pointed out by van der Waals and Platteeuw (1959).

334


This simplified model needs to be extended in the following ways: 1. 2. 3. 4.

5. 6.

The empty framework may not be stable with respect to the normal (pristine) crystalline form of the host material. The free energy of transformation must be included. The guests may distort the host framework. The host framework may have cavities of more than one kind, with different capacities to accommodate guest molecules. The guest molecules will not in general be monatomic and their orientations and librations within the cavities must be taken into account; in some clathrates more than one molecule may enter a particular cavity (but so far this has not been incorporated into the theory). The guest molecules in different cavities may interact, possibly in anisotropic fashion. Classical statistics is used in all treatments; this means that the results do not apply at low temperatures.

Earlier studies on the statistical thermodynamics of clathrate solid solutions (one type of cavity and one guest (van der Waals, 1956; van der Waals and Platteeuw, 1956); two types of cavity and one guest (Platteeuw and van der Waals, 1959); two types of cavity and different types of guest (Barrer and Stuart, 1957)) were consolidated by van der Waals and Platteeuw (1959) into a paper which remains the fundamental reference despite the elapse of more than forty years, and is distinguished by its clear exposition of the theory and its bringing together of theory and experiment. Other useful accounts are by Powell (1964) and Davidson (1973). Many of the earlier simplifying assumptions have been eliminated in a second wave of treatments from Dyadin and colleagues in Novosibirsk; we include only those available in English. Phase diagrams of clathrate solid solutions, taking guest-guest interactions into account, were calculated by Belosludov, Dyadin and Chekhova (1984) and Belosludov, Dyadin, Chekhova, Kolesov and Fadeev (1985) for quinol-noble gas systems. A related approach was taken by Schneider, Tornau, Vlasova and Gurskas (1985). A recent review is by Belosludov, Lavrentiev and Dyadin (1991). Application of the theory to clathrate hydrates has been made by Dyadin and Belosludov (1996). We first consider experimental results for the -quinol clathrates, and then relate these to theory, followed by discussion of how the theories lead to calculation of the phase diagrams of these clathrate systems. Application to clathrate hydrates is given later after discussion of those structures. Formation of -quinol clathrate solid solution from the empty -quinol structure by addition of guest molecules is symbolized by the overall reaction: c þ (guest)g , c (solid solution)

ð7:1Þ

This can be divided into two stages: c (empty) ) c (empty)

ð7:2Þ

c (empty) þ (guest)g ) c (solid solution):

ð7:3Þ

Reaction (7.1) is an univariant equilibrium, the vapor pressure having a definite value at any particular temperature. The equilibrium composition of the -solid solution depends on temperature and pressure of guest and on the interaction between guest and quinol. Experimental investigation of these solid solutions is greatly complicated by kinetic


335

factors; for example, the quinol-argon clathrate has a dissociation pressure of 3.4 atm. at 298K but can be kept indefinitely in a stoppered vial. Clearly argon atoms have difficulty in escaping from their cages. This leads to very slow attainment of equilibrium, a problem overcome by van der Waals and Platteeuw (1956) by addition to the system of a third component (a liquid whose vapour pressure was too low to affect the equilibria in the -quinol–(-quinol clathrate)–gas system and whose molecules were too large to be enclathrated), and by Barrer and Ruzicka (1962) by shaking the (quinol þ guest) mixture with small glass or steel balls, giving rapid equilibration even at 77K. It is convenient to express the reactions involved in a uniform and consistent manner so that numerical values of the thermodynamic quantities are directly comparable; each phase is enclosed in parentheses and followed by a descriptor. The reactions given above can be written more explicitly as 3fðHQÞgðc; formÞ þ qðMðgÞÞ ) f3ðHQÞ ½qðMÞ gðc; formÞ

ð7:1aÞ

fXððqÞð0ÞÞg: 3fðHQÞgðc; formÞ ) f3ðHQÞ ½0 gðc, empty formÞ fXðð0Þð0ÞÞg:

ð7:2aÞ

f3ðHQÞ ½0 gðc; formÞ þ qðMðgÞÞ ) f3ðHQÞ ½qðMÞ gðc; formÞ fXððqÞð0ÞÞg:

ð7:3aÞ

Here HQ ¼ quinol, M ¼ guest, q is the fraction of cavities occupied in general, 3 is the ratio of HQ molecules to cavities in the -quinol clathrate; q ¼ qc (the critical fraction of cavities occupied) at the three-phase invariant point (-HQ, -clathrate, gas M(g)); subscripts ‘‘c’’, ‘‘g’’, . . . represent crystal, gas, . . . The thermodynamic quantities pertaining to reaction (7.1a) are written as X((q)(0)), where X ¼ G, H, S, V and the first-mentioned phase ((q)) is product and the second () reactant, i.e. X ¼ X(product) – X(reactant). When q ¼ 0 then we have X((0)(0)) for the transformation (HQ)c ) (HQ, empty)c (equation 7.2a). To avoid misunderstandings, we shall often explicitly state the number of moles of quinol or gas involved in particular reactions. It is often convenient to express the standard reaction as having all the cavities filled (i.e. q ¼ 1); thus one mole of guest would be the gaseous product of the decomposition reaction. We first consider the measurement of the thermodynamic quantities pertaining to the empty - and -quinol frameworks, and then the interaction of guest with -quinol. The enthalpy of the transformation c ) c has been derived from the difference in the enthalpies of solution of - and (empty) -quinol measured directly in a twin calorimeter (Evans and Richards, 1952), and also by using -quinol clathrates with a range of guest concentrations, prepared by using different pressures of the gaseous guest. H((q)(0)) was then obtained by extrapolating to zero guest concentration (Fig. 7.8). H((q)(0)) is a linear function of q when the guests are small molecules such as Ar and O2 (Evans and Richards, 1954). The direct and extrapolation methods gave H((0)(0)) ¼ 0.54(13) and 0.88 kJ/mol quinol respectively. Parsonage and Staveley (1960) give 0.80 kJ/mol quinol from extrapolation of enthalpy of solution measurements of methane clathrates of different compositions.


336

30.0 25.0

–∆H (kJ)

20.0 15.0 10.0 5.0 qc = 0.34 0.0 –5.0 0

0.2

0.4 0.6 q (mol guest)

0.8

1

Fig. 7.8. Values of enthalpy of formation H((q)(0)) (equation (7.1a); i.e for 3 mols quinol) measured by Evans and Richards (1954) at 298K for -quinol clathrates of Ar (open and filled circles) and O2 (crosses). The slope gives H((1)(0)) and the intercept H((0)(0)). The equation of the best straight line through all the points is H (kJ/mol guest) ¼ 24.79q–2.65 (R2 ¼ 0.9953). We have here ignored the possibility, considered by Evans and Richards, that slope and intercept depend on the nature of the Ar or O2 guest. However, the values for the methane clathrate (small dots enclosed in ellipses) (Parsonage and Staveley, 1960) show that H((1)(0)) does vary for different guests. The value for the krypton clathrate (q ¼ 1) is 24(3) kJ/mol (differential calorimeter; Grey, Parsonage and Staveley, 1962)) and provides supporting evidence in the same direction.

The free energy G((0)(0)) (¼G – G) of the c ) c transformation can be derived (van der Waals and Platteeuw, 1959; Deming, Carlisle, Lauerman, Muckerman, Muirhead and Child, 1969) from measurements at various temperatures of the dissociation pressure and values of qc at the three-phase invariant point 3 3 fHQ ½M gðs; formÞ ¼ HQðs; formÞ þ MðgÞ; qc qc the symbols being defined above. The theory of ideal solid solutions (van der Waals and Platteeuw, 1959) gives the following relationships: 1 Gðð0Þð0ÞÞ ¼ ½RT lnð1 qc Þ

3 Hðð0Þð0ÞÞ þC Gðð0Þð0ÞÞ ¼ RT

at a particular temperature; for measurements made at a number of temperatures.

Van der Waals and Platteeuw (1959) measured qc (¼0.34, irrespective of the nature of the guest) for clathrates with the small molecule guests Ar and Kr and found


337

G((0)(0)) ¼ 343 J/mol quinol. Combining this with the mean of the earlier values of H((0)(0)) (710 J/mol quinol) gives S((0)(0)) ¼ 1.23 J/K mol quinol. Neither enthalpy nor entropy values are expected to be markedly temperature dependent. -Quinol is enthalpy stabilized with respect to (empty) -quinol at 298K; all the evidence suggests that these two polymorphs are monotropically related. Deming et al. (1969) have made measurements of qc and dissociation pressures over a range of temperatures for methane (10–26 C) and methyl fluoride (25–50 C) clathrates (Table 7.3). It has been suggested that the values of H((0)(0)) increase with guest size because of distortion of the -framework but cell dimensions do not support this suggestion; for example, the cell dimensions of the empty -quinol clathrate and that containing 0.87Xe ˚ , V(per quinol) ¼ 145.3 (146.6) A ˚ 3. as guest are a ¼ 16.613 (16.610), c ¼ 5.474 (5.524) A Cell dimensions (Table 7.1) hardly differ for the H2S (R3) and CH3OH and HCl (both R3) clathrates but there are more substantial changes for the SO2 (R3) and, especially, CH3CN (P3) clathrates. A measurement of H((0)(0)) for the latter by the extrapolation method would be very interesting. Comparison of specific heats for (empty) - and -quinol would give direct values for H((0)(0)) and S((0)(0)). S(298) for 1 mol of (empty) -quinol is 140 J/K (from Cp measurements on argon (Parsonage and Staveley, 1959) and methane (Parsonage and Staveley, 1960) clathrates of different concentrations and extrapolation to q ¼ 0); the two derived curves agree within 5% and the entropies at 298K agree well (139.2 and 140.0 J/K per mole of -quinol respectively; Parsonage and Staveley, 1960). Measurements for -quinol do not appear to have been made; it seems unlikely that the precision would be greater than that of the values given in Table 7.2. We have paid some attention to the thermodynamics of the - and the -quinol polymorphs because this seems to be the only example where the empty clathrate framework has been isolated and subjected to measurement. The parameters defining the relative stability of the two polymorphs are fundamental to understanding the stability of the overall clathrate system. The available results for quinol do not appear to be very accurate, possibly due to difficulties in producing suitable samples. Evans and Richards (1954) considered the free energy of formation (G ¼ H – TS) of the -clathrates in terms of the enthalpy and entropy of the reaction {-3(quinol)} þ Mgas ¼ {-(3(quinol)) [M]}. The H values in Table 7.3 range from 20 to 55 kJ/mol. If the -quinol framework is undisturbed on formation of the

Table 7.2. Summary of reported values (for 1 mol of quinol) of thermodynamic parameters for X((0)(0)). The units of G and H are kJ/mol and for S J/K mol Clathrate Empty polymorphs (Type I) Ar and oxygen (Type I) CH4 (Type I) CH3F (Type II)

H((0)(0))

G((0)(0))

S((0)(0))

1

0.54(12) 0.71 (Ar)2; 0.77(O2)2 0.213(109)3; 0.504 0.586(180)3

0.3475 0.410(4)3 0.444(12)3

1.761 0.67(38)3 0.46(67)3

References: (1) Evans and Richards, 1952; (2) Evans and Richards, 1954; (3) Deming et al., 1969; the large errors are ascribed to the narrow temperature ranges used; (4) Parsonage and Staveley, 1960; (5) Child, 1964b.

338


Table 7.3. Values of thermodynamic parameters X((1)(0)) for the formation of -quinol clathrates. The units of G and H are kJ/mol and for S J/K mol. The values for X((1)(0)) are for 3 mols quinol and 1 mol guest; this is equation 7.3a, with q ¼ 1. Only clathrates that have been classified crystallographically have been included in this table Guest Type I Ar Kr Xe CH4 oxygen nitrogen HCOOH CO2 H2S C2H2 Type II CH3F CH3OH SO2 HCl

qc (298K)

H((1)(0))

S((1)(0))

0.345 0.345

25.11 23.3(33)3 40.66 29.2(3)2 23.11 24.31 51.01 55.64 43.54 39.74

84

38.52 461 (approximate) 43.54 38.51

1012

0.392

0.422 0.475

Type III CH3CN

31.44

Others HBr

42.71

862

(1) Evans and Richards, 1954; (2) Deming et al., 1969; (3) Grey, Parsonage and Staveley, 1961; (4) McAdie, 1963, 1966; van der Waals and Platteeuw, 1959; (6) Allison and Barrer, 1968.

clathrate, then S will be equal to the difference in the entropy of the guest molecules M in the gas phase and in the clathrate. In general this will be large and negative, e.g. the translational entropy of HCl gas is 153.6 J/mol K at 298K, corresponding to TS of 45.8 kJ/mol; the actual value will be somewhat larger (less negative) because of residual entropy due to libration and disorder of HCl within the cavity and, if the cavities are not all filled, the contribution of the configurational entropy. Thus it is clear that the enthalpy and entropy contributions to G are opposed and G will be expected to be small. It has been reported that reversible vapor pressures are obtained on heating SO2 and CH3OH clathrates; the derived free energies were small (Wynne-Jones and Anderson, 1952). Values of H((1)(0)) of 59 and 38 kJ/mol were obtained from plots of ln p against 1/T; the latter agrees reasonably well with the corresponding value in Table 7.3. Measurements have been made of specific heats of the -quinol clathrates of Ar (Parsonage and Staveley, 1959), CH4 (Parsonage and Staveley, 1960), Kr (Grey, Parsonage and Staveley, 1961), and CO, N2 and O2 (Grey and Staveley, 1963–4). Subtraction of the contribution of the empty -quinol framework (Parsonage and Staveley, 1960) gave the enthalpy contributions of the guests as a function of temperature (Fig. 7.9). Cp of empty -quinol is about 420 J/3 mol quinol deg., so the contributions of one mole of


339

guest is only a few percent of the measured Cp and cannot be expected to be highly precise. The resemblance in the curves for nitrogen and CO should also be noted; this is not the first example of similarities in behavior of these two gases. The differences among the partial specific heat curves reflect differences in the modes of guest motion in the cavities over the temperature range. The curves are usually interpreted in terms of the cell model, where it is assumed that the motion of the guest is entirely decoupled from that of the host framework. While this may be a reasonable assumption for the lighter guests, it does not apply to heavier guests such as Kr. Here 83Kr Mo¨ssbauer measurements by Hazony and Ruby (1968) have been interpreted in terms of rattling of the guest atoms within the cages, and modification of the dynamics of the framework because of the coupling to the Kr guests. Measurement of the properties of the empty frameworks of clathrates and of their interaction with guests is generally not feasible as crystals cannot be obtained; as noted above, the quinol-guest system is the exception rather than the rule. Thus it is important to develop methods of calculating the required properties. As a first step in this direction, the AMBER force field (developed by Weiner, Kollman, Nguyen and Case, 1986) was used by Zubkus, Shamovsky and Tornau (1992) to perform empirical force field calculations of the interaction of the -quinol clathrate framework with 27 different gases; use of the CVFF force field was also studied but the results were less satisfactory and will not be considered here. The -quinol clathrate structure was assumed throughout to be Type I; as the results cited in Table 7.1 show, this is an oversimplification, at least for CH3OH, HCl and CH3CN of the guests considered. The total energy of the crystal was equal to the potential energy which includes allowance for bond stretching, angle bending, torsional distortion, out of plane bending, van der 100

200

300

40 CO

CP J/mole gas K

40 N2

30 30

O2 20 Kr 0

100

200 T (K)

300

Fig. 7.9. Cp values for one mol of guest in -quinol clathrates; data from references in the text. Lines are guides to the eye; note the shifts along the ordinate. The values (not shown) for Ar are similar to those for Kr, and those for methane to those for oxygen.


340

Calculated ∆H( (0)(1)) kJ/mol guest

80 SO2

70 CH3OH

60 50 N2

40 O2

30

CH4

20 10 0 0

10 20 30 40 50 60 Measured ∆H((0)(1)) kJ/mol guest

70

Fig. 7.10. Comparison of measured and calculated (Zubkus, Shamovsky and Tornau, 1992) values of H((0)(1)) for some quinol clathrates. Note that these values are endothermic and refer to the decomposition reaction.

Waals interactions (Lennard-Jones 6-12 potential), electrostatic interactions and hydrogen bonding. Comparisons of observed and calculated interaction energies (Fig. 7.10) and unit cell dimensions were made. The interaction energies agree reasonably well, although calculated values are systematically some 8 kJ/mol higher than observed values. The measured and calculated values for the cell dimensions of the empty -quinol ˚ ), structure do not agree well (a ¼ 16.61 (meas), 16.44 (calc); c ¼ 5.47 (meas), 5.73 (calc) A and there is no correlation (not shown) of the measured and calculated deviations from these separate baseline values. The value of the inclination angle between c axis and O . . . O vector of the -quinol molecule was also calculated; Palin and Powellp(1948) have shown, assuming standard molecular dimensions, that 5.5(1 þ cos ) ¼ a/ 3 and 5.5 sin ¼ (2/3)c. Thus is not independent of the cell dimensions. Orientations of the guest molecules within the cavities were also calculated but here experimental values are lacking for comparison. Child (1964a) has discussed the values of the enthalpies and entropies of dissociation of the quinol clathrates compared to those of vaporization of the liquid guests at their boiling points. We first consider enthalpies, shown in Fig. 7.11. The relationship is H((1)(0)) ¼ 1.85H(vap) þ 11.91 (R2 ¼ 0.9591). Methanol is clearly anomalous, perhaps due to strong hydrogen bonding in the liquid phase. Child inferred that H((1)(0)) 2H(vap) and suggested that vaporization of a liquid could be envisaged as occurring in two hypothetical steps in which removal of 1 mole of liquid and simultaneous creation of 1 mole of holes is accompanied by an enthalpy increase of 2H(vap) (offset by an enthalpy decrease of –H(vap) when the holes collapse to give a normal liquid). In the clathrate the holes remain. Details of the interaction between guest and framework molecules would have to be taken into account in a quantitative explanation and this has been done in the results summarized in Fig. 7.11.


341

70

∆H( (0)(1)) kJ/mol

60 50 40 CH3OH 30 20 10 0 0

10

20 ∆H(Vap) kJ/mol

30

40

Fig. 7.11. H((0)(1)) versus H(vap), units of kJ/mol. H(vap) refers to liquid guest at boiling point. From left to right along the abscissa, the filled circles are for Ar, Kr, N2, O2, SO2, CH4, HCl, HBr, HCOOH; CH3OH was not included in determining the straight line.

The entropies of vaporization of nonassociated liquids (molecular weights 100) are 100 J/mol K; this is Trouton’s rule, first proposed in 1884. Child (1964a) has pointed out that the values of S((1)(0)) are not very different for the quinol clathrates with small molecule guests noted in Fig. 7.11 and Table 7.3 (it does not matter much whether equation (7.1a) or (7.3a) is used as S((0)(0)) 5 J/3 mol quinol K). This leads to the important conclusion that the clathrates are not entropy-stabilized. The entropy change corresponding to the inverse of equation (7.1a) (i.e. dissociation), normalized to 1 mole of gaseous guest, has been shown (Child, 1964b) to be Sðð0ÞðqÞÞ ¼ R lnðV=Vf Þ RTð@ ln Vf =@TÞv þ R þ R ln q þ Rðð1 qÞ=qÞ lnð1 qÞ

ð7:4Þ

under the conditions that there are only dispersion and repulsion terms in the host-guest interaction; V is the molar volume of gaseous guest in the standard state, Vf is the ‘‘free’’ volume per mole in the clathrate and is a measure of the empty space in the cavities, and q is the fraction of cavities occupied. The first term on the right hand side of equation (7.4) is the increase in entropy on expanding a gas from Vf to V; the second term is related to the departure of the host-guest potential from a square well; R derives from the communal entropy of the gas molecules which the enclathrated guest molecules do not have; the last two terms give the negative of the configurational entropy, Sc, of the clathrate. For the Ar clathrate, qc ¼ 0.34 and Sc ¼ 15.6 J/mol K. Thus we can estimate S((0)(q)) 80.8 þ 0 þ 8.3 – 15.6 73.5 J/mol K (we have used V ¼ 22.4 l and Vf ¼ 1.29 cm3). This is in reasonable agreement with the measured value (84 J/mol K (Table 7.3)) and the entropy of vaporization of Ar (S(vap) ¼ H(vap)/Tb ¼ 6530/87.3 ¼ 74.8 J/mol K). The major problem with this calculation comes from the uncertainty in estimating the appropriate value of Vf.

342


We now move on to the problem of calculating the phase diagram of the quinol–guest systems. The theory is first set out in a rather general form and will then be particularized for the quinol clathrate system; this is essentially an abbreviated version of the presentation of Belosludov, Dyadin, Chekhova and Fadeev (1984). In the i phase (i ¼ , , ) the chemical potentials of host component (Q) and guest (G) are given by 2 iQ ¼ 0i Q þ i ½kT lnð1 yi Þ 1=2yi Ui

ð7:5Þ

i þ kT ln½ yi ð1 þ Si Þ=2hi ð1 yi Þ

i ¼ yi U

ð7:6Þ

where 0i Q is the chemical potential of the (empty) host framework in the i phase; i is the number of cavities of type i per host molecule Q ( ¼ 1/18 for -quinol, 1/3 for -quinol, 1/23 and 3/23 for the two types of cavity in the Type I gas hydrate, 2/17 and 1/17 for the two type of cavity in the Type II gas hydrate); yi is the fractional degree of filling of the cavities by the guest molecules (0 yi 1); 1/2Ui is the dispersion interaction energy between guest molecules in the cavities; 1/2UiD is the dipolar interaction energy between guest molecules in the cavities, with the guest molecules all oriented (in one of two opposite directions) along some common axis; i ¼ Ui þ Si UiD ; Ui ¼ Ui þ S2i UiD ; U hi ¼ 2a3i giØi (T) exp[–Wi(0)/kT], where ai is the radius of the cavity, Wi(0) is the potential energy of the guest molecule in the cavity of the host framework (the host–guest interaction); 2a3i gi/V is the ratio of the free volume of the guest molecule in the i cavity to its molecular volume in the gas phase; Øi ðTÞ ¼ ð2m kT=h2 Þ3=2 j ðTÞ where m is the mass of the guest molecule and j(T) is its internal partition function (Øi (T) is the molecular partition function of the gaseous guest, with the volume factor removed); D S1 i ln½ð1 þ Si Þ ¼ 2yi Ui =kT½ð1 þ Si Þ=ð1 Si Þ ; Si being the order parameter in the i phase. In order to simplify the representation of the three variables (P, T, concentration (xi ¼ iyi/ (1 þ iyi)) governing the behaviour of a two component system (host plus guest1), the vapor pressure is plotted against temperature for all the possible states with three phases in equilibrium. Application of the phase rule (F ¼ C – P þ 2 ¼ 2 – 3 þ 2 ¼ 1) shows that the system will have a definite vapor pressure at every temperature and hence the data can be plotted on a two-dimensional diagram. A clear discussion is given by Glasstone (1947) for the Na2SO4–H2O system. The three-phase curves meet at a point, the quadruple point (P0, T0), where all four phases are in equilibrium, i.e. the chemical potentials of the host component in the four phases are all equal, and those of the guest component are all equal, i.e. Q ¼ Q ¼ Q ¼ G Q;

¼ ¼ ¼ G :

ð7:7Þ

The chemical potentials of the components in the gas phase (total pressure P) and their partial pressures are governed by the ideal gas laws. The equations governing these parameters together with equations (7.5) and (7.6) are inserted in (7.7) and give a group of 1

We consider only one kind of guest but generalization to many kinds has been included in some treatments.


343

six transcendental equations in y, y, y , xG, P, T, which have to be solved in order to calculate the phase diagram of the system. This is discussed in some detail by Belosludov et al. (1985) but will not be considered here. The present problem is first to calculate P0 and T0, and then the three-phase curves for the quinol-noble gas system, where certain approximations are permissible (e.g. the guests are not dipoles so UiD ¼ 0; the phase is liquid quinol, in which the guest has negligible solubility; in the vapor phase the guest behaves as an ideal gas, while the host has zero partial pressure; the guest–host interactions in the - and -phases are the same; the guest– guest interaction is neglected here but its influence was considered later). Thus g g ¼ g; W W ¼ W and hence h ¼ h; y y ¼ y; Ø (T) ¼ Ø (T) ¼ ØG (T). Following Belosludov et al. (1984) the parameters of the quadruple point are (to a first approximation): y0 ¼ 1 exp½0 ðÞðÞ=ð ÞkT

T0 ¼ Tm f1 þ ½ 0 ðÞðÞ=ð Þ =HmL g (subscript ‘‘m’’ refers to melting point) P0 ¼

kT0 y0 expðW=kTÞ: 3 2a g ð1 y0 Þ

The approximate expressions for y0 and T0 do not contain parameters dependent on the nature of the guest, while the changes caused by using the full equations are very small. Belosludov et al. (1984) give values for y0 varying between 0.398 and 0.401 for the guests Ar, Kr, Xe, N2, O2, CH4 (all nonpolar) and HCl (dipolar). Similarly T0 lies between 446.7 and 448.9K for the same guests (Tm ¼ 445.46K). This constancy does not hold for P0, which is dependent on the nature of the guest through a, g and W. A convenient parameter for assessing the guest dependence of P0 is the Lennard-Jones–Devonshire ", where the interaction between two atoms separated by R is given by u(R) ¼ 4"{(/R)12 – (/R)6}; the parameter is the distance for which u(R) ¼ 0 and " is the value of u(R) at its minimum (R ¼ 21/6). The P0–" relationship is shown in Fig. 7.12. There is reasonably good agreement between calculated and measured values of P0 (Kr 15.2 (calc), 13.8 atm (meas); Xe 5.57, 5.8); obviously experimental tests of theory will proceed more sensitively through P0 than through y0 and T0. Phase diagrams for quinol-xenon are shown in Figs. 7.13 and 7.14. The phase diagrams for quinol-krypton (Dyadin, Chekhova and Sokolova, 1987) and quinol-argon (van der Waals and Platteeuw, 1959) are qualitatively similar, but quantitatively different. The equilibrium concentration of methanol in -quinol/methanol at 298K is 0.49 instead of 0.34, as calculated for the noble gas clathrates, and this has been ascribed to a larger value of ((0)(0)) consequent on distortion of the cavities by the larger methanol molecules. However, this has been disputed by Belosludov, Dyadin, Chekhova and Fadeev (1984), who point out that the measured distortions of the cavities are often small (see Table 7.1). Instead they suggest that it is guest-guest interaction which causes the increase in the value of ((0)(0)); a model calculation using the Lennard-Jones–Devonshire formula for Xe completely filling the cavities gives 1/2U ¼ 1200 J mol1. The matter is somewhat controversial as this approach requires that the absolute value of H((0)(0)) is not more than 40–60 J mol1, whereas direct measurement gives 883 J (mol quinol)1.


344

90 N2

80 70

P0(atm)

60 Ar

50

O2

40 30 CH4 20 Kr Xe

10

HCl

0 0

100

200 ε (K)

300

400

Fig. 7.12. The variation of the calculated quadruple point pressure P0 with the Lennard-Jones parameter ". (data from Table 1 of Belosludov et al. (1984).)

450 b – L– G

a–b–L

T (K)

a–L–G quadruple point

445

a–b–G

metastable 440 0

5

10

15

P, atm.

Fig. 7.13. P–T phase diagram for the quinol–xenon system. The continuous lines show results of calculations for the stable three-phase equilibria, and the dashed lines the metastable equilibria. The dots show the experimental results of Kazankin, Palladiev and Trofimov, 1972. (Reproduced from Belosludov et al. (1991).)


T = 298 K

345

P = 5 atm

P atm

TK 446.6 440.8

4 420

a

2

b

380

a

b

0.075 0

x nα 0.1

0.2 x nβ

0 x nα 0.1

0.2 x nβ

Fig. 7.14. An isobaric and an isothermal section of the P–T–x phase diagram for the quinol–xenon system. The broken vertical lines show the concentrations of xenon for the hypothetical situations where the cavities in the - and -quinol phases are completely filled. (Adapted from Belosludov, Dyadin, Chekhova and Fadeev (1984).)

7.2.2 Crystal structure of {(6H2O [hexamethylene tetramine]} The hydrogen bonded hexagon with alternate molecules pointing up and down is an important motif in many clathrate structures to be described later; it is found in the clathrates of phenol and of Dianin’s compound and has been used in the design of the hexa-host clathrate forming molecules (Section 7.5) (MacNicol, 1984b); it is also found in the ice-I (disordered protons) and ice-II (ordered protons) structures. We note here one specific example where the host molecules are water and form a framework analogous to one of the frameworks in the quinol clathrates. The guest is hexamethylene tetramine [HMT; (CH2)6N4] which is unusual in that its solubility in water increases with decreasing temperature, a property shared with some other tertiary amines and explained in terms of an increasing association with water at lower temperatures. The existence of a crystalline hexahydrate decomposing at about 13.5 C has been known for many years (Cambier and Brochet, 1895; Delepine, 1895, 1897). The crystals are ˚ , c ¼ 8.670 A ˚ R3m, Z ¼ 3; Mak, 1965; HXMTHT) and the rhombohedral (a ¼ 11.620 A structural framework is based on a slightly-puckered ring of hydrogen-bonded water molecules with the mean ring plane parallel to (0001). These rings then form hexagonal cages by being stacked one above the other, separated by the hexagonal c spacing; neighbouring columns are staggered in the [0001] direction, in accordance with the rhombohedral symmetry. The host framework is then completed by cross-linkage of neighboring columns by hydrogen bonds (Fig. 7.15), giving rise to to an arrangement equivalent to that of a single framework in the -quinol structure, each benzene ring there (Fig. 7.2) being replaced here by an hydrogen bond; one proton of each water molecule is disordered over two sites. The HMT molecules occupy cavities in the water framework, each guest molecule being suspended ‘‘bat like’’ (in Mak’s vivid phrase) from the upper walls of the cavity. This material is classified as a semi-clathrate hydrate (see Section 7.2.4 below) because the


346

C

2

2

1 2

1 2

2

1

2

1 2

1 1

2

1 1

1

1 H(4)

2 2

C C N C N N

1

1

C

C 1 1

2

1 1

1

C

N

2

2

1

2 2

1

2 2

1 b

2 2

1

2

1

O(1) 1 2

2 2

O(2) 2

1

a

Fig. 7.15. Perspective drawing of the crystal structure of HMT hexahydrate. The eight (H2O)6 rings that form the immediate surroundings of a guest molecule are shown. Hydrogen atoms, except those in the H-bonds to HMT, are omitted for clarity. The axes are of the triply primitive hexagonal cell; c is along [111] of the rhombohedral cell. Note that only three of the four N atoms of HMT act as H-bond acceptors. (Reproduced from Mak (1965).)

guest molecule is hydrogen bonded to the water framework. Isostructural complexes are unlikely to be formed because of the specific geometry required in this structure. 7.2.3

Clathrates derived from existing structures

In general, existing structures are too closely packed to be able to accommodate guest molecules. However, one may envisage exceptions to this rule if the guest molecules are very small or if the host frameworks are very extended. The clathrates would then be primary solid solutions. Helium hexahydrate constitutes a true example of the first type and Cd(CN)2 clathrates quasi-examples of the second type. 7.2.3.1

Helium hexahydrate

Helium, which is too small to form a stable Type I or II gas hydrate (see below), forms an hydrate of ideal composition {6H2O [He]} under pressure (62% occupancy at 0.29 GPa, 79% at 0.47 GPa) (Londono, Kuhs and Finney, 1988; Londono, Finney and Kuhs, 1992). The crystal structure was determined by powder neutron diffraction at 195K (and the ˚ , space pressures noted) and shown to be rhombohedral, a ¼ 12.934(1), c ¼ 6.216(1) A group R 3. This is the structure of ice II, with ordered hydrogens in the H-bonds; the He atoms are enclosed between the six-membered rings of water molecules (Fig. 7.16). It is


347

Fig. 7.16. Structure of {6H2O [xHe]} showing the He atoms occupying sites approximately halfway between the two types of six-membered ring in the ice-II structure. The two types of crystallographically independent water molecule are shown in black and white respectively. The rhombohedral [111] axis is horizontal (diagram by SCHAKAL, written by E. Keller, University of Freiburg, FRG). (Reproduced from Londono, Kuhs and Finney (1988).)

considered a true example because it is derived from an existing structure. In terms of the phase rule, this is a primary solid solution of helium in ice II. The dependence of occupancy ( He) on pressure follows ideal solution behaviour

He ¼

CHe PHe 1 þ CHe PHe

where CHe is the Langmuir constant (¼ 0.07(1) GPa1) and PHe is the helium fugacity. The Ih (ordinary ice) structure can accommodate H2, He or Ne in solid solution under pressure, and phase diagrams have been calculated (Dyadin and Belosludov, 1996). No structural work appears to have been carried out. 7.2.3.2 Cadmium cyanide clathrates Cd(CN)2 is a prototype for the formation of tetrahedral arrangements analogous to those formed by other ABn (n ¼ 0, 2) moieties in which there is a tetrahedral disposition of bonds about A and the B groups can form linear (or quasi-linear) links in both directions; examples of ABn to be encountered later are Si, Ge, OH2 and SiO2 Cd(CN)2 itself has an anticuprite structure in which there are two identical interpenetrating but not-linked diamond-type networks (more accurately -cristobalite-type networks). If one of these networks is removed then a single network with adamantane-like cavities is formed (Fig. 7.17). These cavities are occupied, in a series of isomorphous cubic crystals, by a variety of hydrocarbon and halocarbon molecules such as CHCl3, CH3CHCl2, ˚ ; SAJREA), C2H5CHClCH3, CHCl2CHCl2, CH3CCl3, CHCl2CHCl2 (a ¼ 12.691(2) A ˚ ; SAJRIE) and methylcyclohexane (Kitazawa, CClF2CCl2F, cyclohexane (a ¼ 12.685(2) A Nishikiori, Kuroda and Iwamoto, 1988). Although the crystals lose guest fairly easily at room temperature, the structure of {Cd(CN)2 [CCl4]} could be determined at ˚ , Z ¼ 8, space group Fd 300K (a ¼ 12.668(2) A 3m; SAJRAW); the CCl4 molecules are

348


Fig. 7.17. Structure of {Cd(CN)2 [CCl4]}, showing only the Cd(CN)2 framework; the origin has been shifted by 1/4 along each of three cube-axis directions. The adamantane-like cage is shown by the solid lines; the large circles are Cd and the small circles C or N. (Reproduced from Kitazawa, Nishikiori, Kuroda and Iwamoto, (1988).)

disordered in the cavities. This is considered to be a quasi-example as the single-network Cd(CN)2 structure does not exist. ˚ ; the isostructural Zn(CN)2, with a Zn–CN–Zn The Cd–CN–Cd span length is 5.485 A ˚ span length of 5.12 A does not give analogous clathrates. More complicated structural frameworks based on the Cd(CN)2 network have been synthesized (Abrahams, Hoskins, Liu and Robson, 1991) but these are beyond the scope of this book. 7.2.4

Overview of the polyhedral clathrates (including metalloid structures, clathrasils, gas hydrates, clathrate and semiclathrate hydrates)

7.2.4.1 Historical and general introduction The first gas hydrate encountered seems to have been that of SO2 – Joseph Priestley noticed in 1777–8 that the ‘‘ice’’ formed by cooling aqueous solutions of SO2 sank; this anomalous ice was analyzed as SO2 10H2O some fifty years later (La Rive, 1829). The compound H2SO3 6H2O is now described as the gas hydrate SO2 7H2O in a well known textbook of inorganic chemistry (Cotton and Wilkinson, 1980). Early characterizations of chlorine hydrate as solid chlorine by Pelletier in 1785 and Carsten in 1786 were shown to be wrong by Davy (1811), who obtained the solid hydrate by bubbling chlorine gas into cold water. The hydrate was analysed as Cl2 10H2O by Faraday (1823) and more accurately later, as discussed below (historical references in Chapter 2). Bromine hydrate was reported in 1829 by Lo¨wig but without an analysis; in fact, four hydrates had


349

been thought to have been identified over the years, with 12, 10, 8.6 and 7 molecules of H2O for each Br2 (Dyadin and Belosludov, 1996), but it has now been shown that there is only one bromine hydrate (Section 7.2.7.4). Many other gas hydrates were prepared in the period between 1880 and 1925, principally by de Forcrand and Villard; for example, argon hydrate was reported within one year of the discovery of argon (Villard, 1896, 1897). The historical background up to about 1925 has been summarised by Schroeder (1926). However, the nature of the gas hydrates was clarified only in the early 1950s by model building studies (Claussen, 1951) and crystal structure analyses of key compounds by von Stackelberg and coworkers, and Pauling and Marsh (1952). Many of the earlier crystallographic studies were carried out by von Stackelberg and his school (1949, 1954), following their still earlier preparative and thermodynamic studies; we do not discuss these results in detail but note that the structure of the SO2 gas hydrate was determined in 1942 by H. Fru¨hbuss (Dissertation, Bonn) and that of the double hydrate of CHCl3 and H2S by J. Pieutuchowsky in 1941 (Diplomarbeit, Bonn). Hydrates of methane and other constituents of natural gas (Makogon, 1974, 1987; Cox, 1983; Sloan, 1998) have caused occasional problems in pipelines by forming solid deposits at unexpectedly high temperatures (Fig. 7.18). More importantly, such deposits have been discovered in nature (Chersky, Makogon and Medovski, 1970), in permafrost regions of the Arctic on land and on continental slopes and rises offshore; for example, the Deep Sea Drilling Project off the Pacific Coast of Guatemala found the zone from 1926 to 1955 metres to consist almost entirely of solid gas hydrate. Some twenty five sites had been identified by 1980 (Kvenvolden and McMenamin, 1980). Some estimates of the worldwide reservoir are as high as 1.5 1018m3, large enough to suggest that natural gas hydrates form a potential energy source (Chersky, Makogon and Medovski, 1970; Pearson, Halleck, McGuire, Hermes and Wright, 1983; Katz, 1971a,b; Byk, Makogon and Fomina, 1980; Kvenvolden, 1994). Interest continues, as news items entitled Fire and Ice Under the Deep Sea Floor (Appenzeller, 1991) and Fire from Ice (Adam, 2002) attest. Pressure is an important variable in gas hydrate chemistry but not considered here because of space limitations; a leading reference is Kuhs (2004). Metalloid analogs to the gas hydrates, exemplified by compositions such as K8Ge46, were first reported in 1965 (Kasper, Hagenmuller, Pouchard and Gros, 1965), when the resemblance between the structure of the rare mineral melanophlogite and those of the gas hydrates was also recognized (Kamb, 1965). This led somewhat later to the development of a structurally related group of silica-based compounds, the clathrasils (Gies, Liebau and Gerke, 1982; Gies, 1991; Gies and Marler, 1996). The analogy between H2O and SiO2 based structures is, of course, well known; for example, hexagonal ice Ih is isostructural with -tridymite and cubic ice Ic with -cristobalite. There had been parallel advances in apparently unrelated fields. Alkylamine hydrates, some of which contained surprisingly large numbers of water molecules, had been studied by Pickering (1893) by phase diagrams and reinvestigated (in part) later (Somerville, 1931), while polyhydrates of many peralkylammonium salts had been prepared (Fowler, Loebenstein, Pall and Kraus, 1940). The structural inter-relations between these two groups of compounds and the gas hydrates were only revealed by the extensive crystallographic studies of Jeffrey2 and coworkers. Much work, especially on the hydrates of 2

George A. Jeffrey 1915–2000.

350


Fig. 7.18. A huge pillar of solidified (natural) gas hydrate produced in a methane pipeline rupture under below-zero (63 C) conditions; the height of the pillar can be estimated, by comparison with the hut, as about 60 m. The photograph was taken in Siberia by Novosti Press Agency and is reproduced from Makogon (1987).

peralkylammonium salts and analogs (referred to together as peralkylonium salt hydrates) and the trialkylamine oxides and analogs (referred to together as the trialkyline oxide hydrates), has been carried out by Dyadin3 and coworkers in the Soviet Union and also more recently by Nakayama and coworkers in Japan. There have been a number of excellent reviews (van der Waals and Plateeuw, 1959; Powell, 1964; Jeffrey and McMullan, 1967; Jeffrey, 1969; Davidson, 1973; Jeffrey, 1984a; Jeffrey, 1984b; Dyadin and Udachin, 1984, 1987; Dyadin and Belosludov, 1996; Jeffrey, 1996) and books (Bercez and Bala-Achs, 1983; Sloan, 1998). The polyhedral clathrates,4 a convenient phrase covering the gas hydrates and analogs (including the 3 4

Yuri A. Dyadin 1935–2002. The series of crystal structure analyses by Jeffrey and coworkers is entitled Polyhedral Clathrate Hydrates.


351

metalloids and clathrasils), the alkylamine hydrates, the peralkylonium salt hydrates and the trialkyline oxide hydrates, are one of the most widely and deeply studied families of molecular compounds and complexes, with roughly equal amounts of attention having been given to the structural and the physico-chemical aspects. Jeffrey (1996) has used the following classification of the polyhedral clathrates: (i) The ‘‘true clathrates’’ have uninterrupted frameworks and the guest molecules or salts do not participate in the framework structure; there are van der Waals interactions between neutral framework and neutral guest. The gas hydrates and analoges are the prime examples. (ii) In the ‘‘ionic clathrates’’ the cation (or anion) participates in the framework which may undergo considerable modification in order to accommodate the bulky anion (or cation). The framework–guest interaction is ionic. (iii) In the ‘‘semiclathrates’’ functional groups of the guest molecules enter into a specific hydrogen-bonding relationship with the water molecules of the framework, while the hydrophobic groups occupy voids in the structure as in the ‘‘true clathrates’’. The hydrates of the trialkylamine (phosphine, stibine) oxides belong formally to this category.

Polyhedral clathrates

Semiclathrates

True clathrates

Interrupted frameworks

Uninterrupted frameworks

Metalloid structures

Clathrasils

(guests not linked to framework)

Gas hydrates

Peralkylonium salt hydrates (anions part of framework; cations interrupt framework)

Alkylamine polyhydrates

Trialkyline oxide hydrates

(guests H-bonded to framework)

(guests bonded to framework through participating oxygen atoms)

Fig. 7.19. Diagram showing relationships among the various types of polyhedral clathrate structures. The classification is based on a combination of chemical and structural considerations.


352

The structures are marvellous examples of the adaptation of pure geometrical principles to the demands of real chemical molecules. Perhaps Coleridge in Kubla Khan told us more than he could have suspected: It was a miracle of rare device, A sunny pleasure-dome with caves of ice.

More prosaically, Jeffery (1996) has commented that ‘‘ . . . present knowledge is likely to be only the tip of the iceberg in hydrate structural chemistry . . . ’’ Our description of the structures of the polyhedral clathrates will be based to a large extent on the accounts of Jeffrey, Davidson, Dyadin and their coworkers. We shall first describe the overall geometrical features of these structures, leading to definition of some of the polyhedra encountered and introduction of the distinction between those structures that contain pentagonal dodecahedra and those that do not. We then consider the structural group containing pentagonal dodecahedra and uninterrupted frameworks, starting with the metalloid and clathrasil analogs to the gas hydrates and proceeding to the cubic gas hydrates themselves. The thermodynamics and stoichiometry of the gas hydrates is considered in Section 7.2.7.2. We then broaden the treatment by including structures with interrupted frameworks, starting with those of the peralkylonium salt hydrates and trialkyline oxide hydrates and then consider the alkylamine hydrates. Finally, hydrates not based on pentagonal dodecahedra are discussed. Thus we weave together structural and chemical considerations but make no claim to complete consistency – the discerning Table 7.4. Various nomenclatures used to describe the polyhedral clathrates Symbol

Name

Space group Previous name/ example

Structures containing dodecahedral cavities CS-I (SCS-I) Cubic Structure I Pm3n (Superstructure of CS-I) CS-II Cubic Structure II Fd3m HS-I HS-II HS-III

Hexagonal Structure I Hexagonal Structure II Hexagonal Structure III

TS-I RS-I OS-I

Tetragonal Structure I P42/mnm Rhombohedral Structure I R3m Orthorhombic Structure I Pbam

Structures not containing dodecahedral OS-II Orthorhombic Structure II Tet-II Tetragonal II Cub-I Cubic I

P6/mmm P63/mmc P6/mmm

cavities Fmmm I4/mcm I43d

Cub-II

Cubic II

Im3m

Tet-III

Tetragonal III

I41/amd

Structure I or Type I gas hydrate Structure II or Type II gas hydrate No example known No example known Structure H hydrate; dodecasil 1H Bromine hydrate dodecasil 3R RS-I (Dyadin and Udachin, 1987)

Jeffrey (1984a) (Table 2) I II IV V Not considered by Jeffrey III

nonasils 16X 156H2O VI (X¼ (CH3)3CHNH2) VII HEF6 5H2O HF (E¼P, As, Sb) Sigma-2


353

reader will note, for example, that the clathrasils are all treated together ignoring distinctions due to presence or absence of pentagonal dodecahedra. The relationships among the various structural types and groups of compounds which comprise the polyhedral clathrates are set out in Fig. 7.19; there is some overlap between the structural and chemical bases of this classification in the sense that there are some (but very few) alkylamine polyhydrates with true clathrate structures. We shall first consider the arrangements found in the gas hydrates and their analogs, including those few structures not based on pentagonal dodecahedra, and later proceed to discussion of the polyhedra found in the peralkylonium salt hydrates and semiclathrate hydrates. We shall use a nomenclature for the polyhedral clathrates based on that used by Dyadin and Udachin (1987) to describe in general terms their crystallography, but extended here to cover the whole field as it is currently known. The relationship of this nomenclature to those used by other authors is set out in Table 7.4. 7.2.4.2

Restrictions on the shapes of the polyhedra

The structures are all based on polyhedral frameworks of four-connected tetrahedral moieties, which we shall refer to generally as T 4; specifically T 4 can be Si or Ge, SiO2or H2O. These will give neutral frameworks but participation of some other atoms can give charged frameworks. The tetrahedral disposition of the bonds about T4 implies that each vertex of the polyhedra comprising the overall structure must have an order of exactly three (the order is the number of bonds directed (on the same side of a supporting plane) so they can be edges of the same polyhedron) (King, 1972). Furthermore, the angles between the edges of the polyhedra must be close to tetrahedral in order to minimise angle strain, leading to the following order of preference for the shapes of regular plane faces: pentagonal (most favourable) > hexagonal > quadrilateral triangular. Angle strain will not be appreciably altered by moderate deviations from regularity because decrease of some angles (further from tetrahedral) must be accompanied by increase of others (closer to tetrahedral) as the sum of the internal angles of a planar n-gon is constant at (n – 2). Appreciable bending of the faces of the polyhedra is also not to be expected. Although these considerations show that the pentagonal face will tend to be preferred in all polyhedra and that the preferred polyhedral shape will be the slightly distorted pentagonal dodecahedron, nonetheless experience shows that other types of face and polyhedron are also found. Two groups of structures can be distinguished – those based on various packings of pentagonal dodecahedra and those which do not contain any pentagonal dodecahedra (Fig. 7.20). These two groups cut across the boundaries of guest type and uninterrupted or interrrupted frameworks and will be discussed separately, starting with a consideration of the packing of pentagonal dodecahedra. The pentagonal dodecahedron is one of the platonic solids, with 12 regular pentagonal faces, 20 vertices (T 4 moieties) and 30 edges. The faces must be planar in the ideal structure and the bond angles, at 108 , are nearly tetrahedral with only small deviations from tetrahedrality encountered in real structures (for example the H–O–H angle in water is ideally 104 ). The pentagonal dodecahedron cannot by itself fill space (its fivefold axes are not compatible with the requirements for long range order in a crystal, that can only be


354

Polyhedral clathrates

Structures based on packing of Pentagonal dodecahedra

Structures without pentagonal dodecahedra

CS-I, CS-II TS-I HS-I, HS-II, HS-III OS-I

Cub-I, Cub-II Tet-II OS-II (nonasil)

Fig. 7.20. Classification of the polyhedral clathrates into two structural groups in terms of presence or absence of pentagonal dodecahedra. Some examples of structural types (defined and described in the following sections) are given in each category.

fulfilled by bodies with 2, 3, 4 or 6 fold symmetry)5 and therefore must be associated, in an ordered fashion, with other polyhedra (that can and generally do have nonplanar faces and unequal edges and/or angles) in order to form an arrangement which satisfactorily fills space in the crystal. The following relationships hold for the parameters describing closed convex threedimensional polyhedra with all vertices of order 3. King (1972), who listed all convex polyhedra of order 3 with up to 18 faces, and Wells (1975) give good summaries of the relevant solid geometry. FþV¼Eþ2 Euler’s rule 3V ¼ 2E Relationship between edges and vertices i (i Fi) ¼ 2E Relationship between edges and faces Totality of faces i Fi ¼ F (Here Fi refers to the number of faces with i sides or edges; for example F4 is the number of quadrilateral faces.)

1. 2. 3. 4.

For a given set of V, E and F values there are many polyhedra corresponding to different solutions of equations (3) and (4). We shall list the polyhedra found experimentally to date in structures with uninterrupted frameworks; these are described in terms of the number of faces of a given kind, together with the numbers of vertices, edges and faces which, of course, satisfy Euler’s rule. Thus the pentagonal dodecahedron is described as 512 (i.e. it has twelve pentagonal faces), with 12F þ 20V ¼ 30E þ 2. 1. 2. 3.

8-hedra 9-hedra 12-hedra (D) 5

(4454), (5464) (4158) (512), (435663)

8F þ 12V ¼ 18E þ 2 9F þ 14V ¼ 21E þ 2 12F þ 20V ¼ 30E þ 2

This sentence was phrased in the pre-icosahedral era; time will tell how it needs to be modified.


(a)

(b)

(c)

(d)

355

Fig. 7.21. (a) pentagonal dodecahedron (D), (b) 14-hedron (T), (c) 15-hedron (P) and (d) 16-hedron (H). These are the polyhedra found, either in association with the pentagonal dodecahedron, or separately, in the gas hydrate and analogous structures. Reproduced from Jeffrey and McMullan (1967).

4. 5. 6. 7. 8. 9.

14-hedra (T) 15-hedra (P) 16-hedra (H) 17-hedra 18-hedra 20-hedra

(51262), (4668) (51263) (51264) (43596273) (51266) (58612), (51268)

14F þ 24V ¼ 36E þ 2 15F þ 26V ¼ 39E þ 2 16F þ 28V ¼ 42E þ 2 17F þ 30V ¼ 45E þ 2 18F þ 32V ¼ 48E þ 2 20F þ 36V ¼ 54E þ 2

The polyhedra 3–6 are shown in Fig. 7.21. The next problem is to consider how the dodecahedra can pack together to form closely packed ordered structures. 7.2.4.3

Packing of pentagonal dodecahedra

Jeffrey and McMullan (1967) point out that there are only two ways of joining pentagonal dodecahedra which preserve fourfold first-neighbor coordination about T 4 moieties at the vertices. These are: (a) formation of hydrogen bonds between vertices of different dodecahedra, either directly or through additional T 4 moieties (sharing edges or corners would give five or sixfold coordination), (b) sharing common faces of adjacent dodecahedra. On this basis the possible structures can be divided into four groups (Fig. 7.22) of which the first and third are immediately relevant to the discussion of the cubic gas hydrates and their analogs.

356


Polyhedral clathrates based on pentagonal dodecahedra

Vertex linked in three dimensions

Face sharing in two dimensions with vertex linking in the third

Face sharing in three dimensions

Face sharing within limited groups of dodedahedra, with vertex linking between groups

Fig. 7.22. Possible ways of packing pentagonal dodecahedra so as to fill space.

(i) Vertex linking in three dimensions. Consider eight pentagonal dodecahedra at the corners of a cube, surrounding a dodecahedron at the center. When the central dodecahedron is rotated by 90 (about a cube axis) with respect to those at the corners, then eight vertex to vertex links can be formed from central to surrounding dodecahedra. The structure (space group Pm3n – Oh3)6 is analogous to that of CsCl, but with the different ions replaced by differently-oriented dodecahedra. This is the structure of cubic 2[(nC4H9)3SF] 40H2O, to be discussed later. The Type I gas hydrate structure (CS-I; see Table 7.4 for nomenclature) is obtained by addition of six T 4 moieties at the Wyckoff (d) positions (1/4,1/2,0, etc) of the space group; these then form bonds to the remaining twelve vertices of the central dodecahedron. The unit cell contains 46 T 4 moieties, which are distributed as follows: 6 at positions (c) with symmetry 4 2m; 16 at positions (i), with symmetry 3; 24 at positions (k), with symmetry m. The T 4 moieties can be considered to be arranged in six pentagonal dodecahedra (D or M ), centred at the twofold positions (a) of symmetry mmm, and six 14-hedra (T or M14), centred at the sixfold positions (d) of symmetry 42m (Fig. 7.23). The hydrogen bond O . . . O distance in the water framework of the Type I gas hydrate ˚ , and this gives a cell edge of 12 A ˚ , with the pentagonal dodecahedron (M12) is 2.8 A ˚ ˚ 3 while the 14-hedron has an having a net diameter of 5.2 A and a net volume of 74 A ˚ ˚ 3 (van der ellipsoidal cavity of dimensions 6.2 3.2 A with a net volume of 266 A Waals radii of the framework atoms are taken into account in calculating these volumes). These values, which depend on the nature of the T 4 moiety, set limits to the sizes of guest molecules which can be accommodated in the cavities, which are not necessarily all occupied. The traditional way of expressing composition is in terms of ‘hydration 12

6 The space group number assigned by the International Union of Crystallography is No. 223; the symbol Pm3n used in International Tables for X-Ray Crystallography (First edition, 1952; Second edition 1965) has been replaced by Pm 3n in International Tables for Crystallography (First edition, 1983; Fifth revised edition 1998). Both symbols are used here interchangeably.


357

Fig. 7.23. The polyhedral T 4 moiety framework of the Type I gas hydrate structure; only a limited number of the polyhedra have been included in the diagram in order to keep matters simple. The pentagonal dodecahedra are lightly outlined and the 14-hedra emphasized. The center of the cell at 1=2,1=2,1=2 is marked by the cross. (Reproduced from Jeffrey and McMullan (1967).)

number’ Gð¼ 46=ðm þ nÞ ¼ 5:75 for a Type I gas hydrate with all cavities occupied). Alternatively the content of the unit cell is given as {46T 4 [n(M14)] [m(M12)]}, i.e. n guests in the M14 cages and m in the M12cages, where n ¼ 6 and m ¼ 2 in the ideal, fully-occupied structure; the fractional occupancies of the large (M14) and small (M12) cages are often denoted as qL and qS (sometimes L and S ). As we have done elsewhere, we place the whole formula within parentheses, with framework former first and then the enclathrated species within square brackets. It seems more logical to express compositions in terms of unit cell contents than in other ways, but hydration numbers are so well entrenched in the gas hydrate literature that it is often convenient to use them as well. (ii) Single layer packing. Slightly distorted pentagonal dodecahedra (angle distortions 3 ) can be packed together to give a two-dimensional arrangement of hexagonal symmetry, each dodecahdron sharing faces with four other dodecahedra. The layers are stacked one on the other with sequence . . . AA . . . ; in addition to the [512] (D) dodecahedra within the layers there are two other types of cage – two [435663] (D00 ) cages and one eicosahedral [51268] (E) cage per unit cell – between the layers (Fig. 7.24). The E cage is large and D and D00 are small. This is the dodecasil 1H structure, the simplest of these polytypes. The space group is P6/mmm and the ideal composition is expressed as 0 {34 T 4 [1(M20)] [2(M12 )] [3(M12)]}.


358

A third major type of gas hydrate (see Section 7.2.7 below; HS-III in addition to CS-I and CS-II), called structure H hydrate, was discovered in 1987 using NMR methods (Ripmeester, Tse, Ratcliffe and Powell, 1987). The guests include methylcyclohexane, methylcyclopentane, 2-methylbutane, 2,3-dimethylbutane, 2,2-dimethylbutane, 2,2,3-trimethylbutane, hexamethylethane, hexachloroethane, 2,2-dimethylpentane, 3,3dimethylpentane, cycloheptene, cyclooctane, cis-cyclooctene, adamantane, bicyclo[2.2.2]-oct-2-ene, 2,3-dimethyl-2-butene, 2,3-dimethyl-1-butene, 3,3-dimethyl-1-butene, 3,3-dimethyl-1-butyne, cis-1,2-dimethylcyclohexane, t-butylmethylether, 2-adamantone, tetra-methylsilane and isoamyl alcohol. Indexed X-ray and neutron powder diffraction

a c⬘ c

a⬘

S16L8·136H2O

Structure II Fd 3m 512

51264

S3S2L·34H2O

Structure H p6/mmm 512

435663

51268

Fig. 7.24. (upper) The hexagonal layer obtained by packing slightly distorted pentagonal dodecahedra as described in the text; (lower) the cages between . . . ABC . . . stacked cubic layers [(111) planes] and . . . AA . . . stacked hexagonal layers. (Reproduced from Jeffrey (1984b) and Ripmeester and Ratcliffe, 1991.)


359

patterns indicate that this class of gas hydrate is isomorphous with dodecasil 1H. Confirmation has come from determination of the structure of the hydrate {34H2O [2,2-dimethylpentane] [5Xe, 5H2S]} at 173K (a ¼ 12.212(2), c ¼ 10.143(2) ˚ , P6/mmm, Z ¼ 1; Udachin, Ratcliffe, Enright and Ripmeester, 1997b; PEXQIS). Why A did it take so long? ‘‘Crystals were grown in a sealed tube from the four phase mixture of 2,2-dimethylpentane, Xe, H2S and ice at approximately –20 C for 6 years. On the walls . . . there formed transparent well-edged hexagonal prisms . . . ’’ (my italics). An example of a composite structure, incorporating elements of both the CS-II and HS-III structures, has been found in {30.33 H2O [0.86 choline hydroxide]. ˚ , trigonal tetra-n-propylammonium fluoride} at 243K (a ¼ 12.5335(1), c ¼ 90.525(1) A R3; Z ¼ 12; Udachin and Ripmeester, 1999a; XAWPAM). (Choline is (CH3)3N þ CH2CH2OH.) We describe the structure in somewhat oversimplified terms. The cholines are located in two types of large cages, and the cations in large supercages formed by coalescence of dodecahedral and other types of cage; choline is hydrogen bonded to a displaced water molecule in one of the supercages but not in the other. Along the (unprecedently long) c axis, there is a sequence of CS-II and HS-III domains arranged in layers with a complicated sequence. It is suggested that variants of this hydrate structure may play an important role in natural settings. (iii) Triple layer packing (face sharing in three dimensions). If additional T4 moieties are placed between the layers described in the previous section so as to give additional dodecahedra and 16-hedra, then a cubic structure is formed with an . . . ABCABC . . . sequence of layers in which there is three-dimensional face sharing between dodecahedra leaving 16-hedral voids. When T 4¼H2O then the 16-hedron, which is almost spherical in ˚ and a net volume of 151 A ˚ 3. The normal to the layer shape, has a net diameter of 6.6 A is the [111] axis of the cube; the space group is Fd3m. There are [51264] cages between the layers, centred at the eightfold positions (b) of symmetry 4 3m, while the dodecahedra are centred at the 16-fold positions (c) of symmetry 4 3m. The 136 T 4 moieties are

Fig. 7.25. Stereodiagram illustrating the characteristic packing of oxygen polyhedra in the CS-II type structure. Two 16-hedra are centred at (3/8, 3/8, 3/8) and (5/8, 5/8, 5/8) and two clusters of four 12-hedra centred at (1/8, 1/8, 1/8) and (7/8, 7/8, 7/8). The view is down the a axis. The solid circles show the centres of 12- and 16-hedra. (Reproduced from McMullan and Kvick (1990).)

360


distributed with 8 in (a), with symmetry 4 3m; 32 in (e) with symmetry 3m and 96 in (g) with symmetry m. The composition is expressed as {136T 4 [8(M16) 16(M12)]}; i.e. {136D2O [8CCl4] [8Xe]}; KELKUH at 13K and KELKUH01 at 100K. This is the structure of the Type II gas hydrate (Fig. 7.25) and of the dodecasil 3C clathrasil. In analogy with cubic and hexagonal close packing, an . . . ABABAB . . . hexagonal arrangement of the layers is also possible. Such a structure would have space group P63/mmc and the same stoichiometry as the cubic modification. The semiclathrate iso-propylamine hydrate (CH3)2CHNH2 8H2O has a distorted version of this structure. We shall now discuss the metalloid structures, the clathrasils and the gas hydrates as realizations of these geometrical principles. 7.2.5

Metalloid structures

The first hint of these structures came from studies of the thermal decomposition of compounds formed between alkali metals and Si (or Ge), suggesting the formation of compounds reported tentatively (Hohmann, 1948) as KSi8, RbSi8 and CsSi8, and later as KSi6, RbSi6 and CsSi8 (Scha¨fer and Klemm, 1961). Definitive results were obtained by thermal degradation of alkali silicides and germanides of formula ME (M¼Na, K, Rb, Cs; E¼Si, Ge) (Kasper, Hagenmuller, Pouchard and Cros, 1965; Cros, Pouchard and Hagenmuller, 1970), or exposure of powdered Group IV metal to alkali metal vapors under argon at 600–700 (Gallmeier, Scha¨fer and Weiss, 1969) to give intermediate phases (as single crystals) of composition {Si46 [Na8]} and {Si136 [Nax]} (x < 11) in the Si–Na system, and analogous E46M8 phases in other systems. {Si46 [Na8]} is cubic ˚ , space group Pm (a ¼ 10.19(2) A 3n) and has the CS-I (Type I) gas hydrate structure with ˚ (cf. all the cavities filled by Na atoms. The Si–Si distances in the framework are 2.37 A ˚ 2.35 A in diamond-structure Si), while the Na to Si distances are too large for any bonding. The electrical and magnetic properties suggest that there is a certain localization of the 3s electron of Na. The less well defined {Si136 [Nax]} (x < 11) has the CS-II gas ˚ , space group Fd3m), and the formula proposed hydrate structure (a ¼ 14.62(2) A was {136Si [6 3Na] [3 2Na]}. Both types of compound have chemical properties similar to those of silicon, the alkali metals being chemically inert; neither material reacts with strong acids, apart from HF which also attacks silicon. Phases of the type {(XY)46 [M8]} (M¼Na, K; X¼Al, Ga, In; Y¼Si, Ge) have been prepared (Westerhaus and Schuster, 1977) and are similarly chemically inert. When the framework atoms are ordered in the XY (X 6¼ Y) compounds, the space group is P43n; if X and Y cannot be distinguished then the space group is Pm 3n. In agreement with Pauling and Marsh’s (1952) results for CS-I chlorine hydrate, the polyhedra are not regular; for example, in ˚ and one of 2.43 A ˚ , with two angles {Si46 [K8]} the dodecahedron has four sides of 2.38 A at 107 and three at 110 , while the hexagonal face of the 14-hedron has four angles of 124 and two of 111 ; similar, but more accurate, results have been reported ˚ ; < Si–Si– for {Si46 [Na8]} (d(Si–Si) ¼ 2.306(2), 2.371(2), 2.373(2). 2.393(3) A Si ¼ 105.2(1), 105.8(1), 105.9(1), 124.8(1) ) (Reny, Gravereau, Cros and Pouchard, 1998). Defect structures are also known; in {Sn44 [K8]} one-third of the Sn atoms which lie across the shared hexagons of the CS-I structure are left empty (at random) and the compound should be formulated as {(3b-Sn1)8(4b-Sn0)36 [Kþ]8} (Zhao and Corbett, 1994); Corbett notes analogous but unpublished examples from von Schnering’s


361

laboratory. Reports of superconductivity (Tc 6K) in CS-I type Si-based structures containing Ba and transition metals have attracted attention; there are also possible thermoelectric applications of Ge clathrates. The cell dimensions for many of this group of compounds, and for the halogen analogs (see below) are collected in Table 7.5. The data given in this table enable one to ascertain the effects on cell size of changes in both framework and enclathrated atoms. Thus ˚ and by Sn to an expansion of 1.32 A ˚; replacing Ge by Si leads to a contraction of 0.51 A ˚ replacing Na by K leads to an expansion of 0.11 A. Halogen analogs of the alkali metal–Group IV compounds have been prepared (Menke and von Schnering, 1973) by reaction of the elements in sealed tubes at 700 C, giving shiny single crystals of composition {Ge38A8 [X8]}, where X¼Cl, Br or I; A¼P, As or Sb. Single crystal X-ray diffraction studies of {Ge38P8 [Br8], {Ge38P8 [I8]} and {Ge38As8 [I8]} showed that these had CS-I gas hydrate structures, with ordered arrangements of the framework atoms in space group P43n, there being 8Ge in 8(e) positions with x 0.18 etc. and 8A atoms in 8(e) with x 0.82 etc. (in space group Pm3n these would be the 16-fold (i) positions). Presumably the ordering was demonstrated for the phosphorus compounds and inferred for the others. This ordering was not confirmed in a briefly described neutron diffraction study of {(Ge38As8)8þ [8I1]}, which was assigned space group Pm3n (Chu, Chu and Ray, 1982); neutron diffraction showed that eight Ge and eight As atoms were disordered over the 16-fold (i) positions, with the remaining Ge atoms in the six fold (c) and 24-fold (k) positions (Chu, Chu, Rosenstein and McMullan, 1982). This material is an n-type semiconductor, with a resistivity of 1 ohm cm at 300K. Analogous results were obtained for {Ge14(GaSb)12Sb8 [I8]} (Menke and von Schnering, 1976); the bracketed Ga and Sb atoms were disordered over the 6(c) positions at 0, 1/4,1/2 etc. In {Ge43.33I2.67 [I8]} the framework iodines are disordered over the 6(c) sites and are presumed to be present as tetrahedral I3þ (Nesper, Curda and von Schnering, 1986). This compound was formed during an unsuccessful attempt to synthesize ˚ when Br is replaced by I, by 0.11 A ˚ when As {Ge46 [Xe8]}! The cell expands by 0.09 A ˚ when Sb replaces As. replaces P and by 0.26 A The compounds K8Sn25, K6Sn23Bi2 and Ba8Ga16Sn30 are not clathrate metalloids but there are certain structural resemblances (references in Nolas et al., 1999). After a period of some dormancy, interest in these materials has revived, particularly in the CS-II type silicides and germanides. Single crystals (polyhedral in shape, bluish, with metallic lustre) have been prepared of {Si136 [Na16][Cs8]} and {Ge136 [Na16][Cs8]} (Bobev and Sevov, 1999; Ramachandran, Diefenbacher, Sankey, Sharma, Marzke, O’Keefe, Gryko and McMillan, 1999). The materials are remarkably air- and moisturestable. The smaller Na completely occupies the smaller Si(Ge)20 cage, and the larger Cs the larger Si(Ge)28 cage. These are the 16-fold dodecahedral (512) and the eightfold 16-hedral (51264) cages in the nomenclature used in the previous section (Section 7.2.4.3) and later in discussing the gas hydrates. Rb is too small for the large cages and K too large for the small cavities. The Na–Cs pair is so far the only successful combination of guests. Powder studies (Rietveld analyses) on {Si46 [Na8]} and {Si136 [Nax]} (Reny, Gravereau, Cros and Pouchard, 1998) showed that both dodecahedral and 14-hedral cages were fully occupied in the CS-I compound, but that, in the CS-II compound, the ‘‘sodium atoms are exclusively, not only preferentially, located in the eight large Si28 sites for x 8, and that, for 8 x 24, the smaller Si20 sites are progressively occupied with

Table 7.5. Cell dimensions of E46M8 (E¼Si, Ge and other combinations) (CS-I structure) and E136XxYy (CS-II structure) compounds. Standard ˚ and for the halogen analogues 0.001 A ˚ ; more precise values are indicated uncertainties of measurements for alkali metal compounds are 0.01–0.02 A specifically. Space group P4 3n is denoted by #, while the space group for the other CS-I compounds is Pm3n. The space group for the CS-II compounds is Fd3m Compound

˚) a (A

Compound

˚) a (A

Compound

˚) a (A

CS-I structure type {Si46 [Na8]} (GSW69) {Si46 [Na8]} (RGCP98)

10.19 10.1983 (2)

{Si46 [K8]} (GSW69)

10.30

{Si38Ga8 [K8]}

10.427

{Si38Ga8 [Rb8]

10.469

{Al18Ge28 [Na8]} (G70) {Ge46 [K8] (GSW69) {Al23Ge23 [K8]} (WS77) {Al23Ge23 [K8]} (WS77) {In18Ge28 [K8]} (WS77) {Ge38Al8 [Rb8]} (vS98) {Ge38Ga8 [Rb8]} (vS98) {Ge38Ga8 [Cs8]} (vS98)

10.70# 10.71# 10.80# 10.76 – 10.822 10.783 10.835

{Sn46 [Na8]} (GSW69) {Sn46 [K8]} (N99) {Sn44.6 [Rb8]} (ZC94) {Sn44 [K1.6 Cs6.4]} (ZC94) {Sn38Ga8 [K8]} (vS98) {Sn38Ga8 [Rb8]} (vS98) {Sn38Ga8 [Cs8]} (vS98)

12.03 12.030 12.054(1) 12.084(1) 11.935 11.964 12.006; 12.079 at 300K (N00)

{Ge38In8 [K8]} (vS98) {Ge38In8 [Rb8]} (vS98) {Ge38P8 [Cl8]} (MvS73) {Ge38P8 [Br8]} (MvS73) {Ge38P8 [I8]} (MvS73) {Ge38As8 [Br8]} (MvS73) {Ge38As8 [I8]} (CCR82; CCRM82) {Ge38Sb8 [Br8]} (MvS73) {Ge38Sb8 [I8]} (MvS73) {Ge14(GaSb)12Sb8 [I8]} (MvS76) {Ge43.33I2.67 [I8]} (NCvS86)

10.997 11.033 10.351# 10.407# 10.507# 10.516# 10.616 10.789 10.870 11.273# 10.814

{Sn38Al8 [Rb8]} (vS98)

12.036

{Sn42Zn4 [Cs8]} (N00)

12.093 at 11K

Ge136[Na16][Cs8] (BS99)

15.4805 (6)

Sn104Ga32[Ba16] (vS98)

17.054

CS-II structure type Si136[Na16][Cs8] (BS99) Si136[Na16][Na4] (RGCP98) (see text)

14.7560 (4) 14.7030 (5)

References: BS99 – Bobev and Sevov, 1999; single crystal measurements; CCR82 – Chu, Chu and Ray, 1982; CCRM82 – Chu, Chu, Rosenstein and McMullan, 1982; GSW69 – Gallmeier, Scha¨fer and Weiss, 1969; MvS73 – Menke and von Schnering, 1973; MvS76 – Menke and von Schnering, 1976; N99 – Nolas et al., 1999; this contains earlier references; N00 – Nolas et al., 2000; NCvS86 – Nesper, Curda and von Schnering, 1986; RGCP98 – Reny, Gravereau, Cros and Pouchard, 1998; vS98 – von Schnering and coworkers (1998); WS77 – Westerhaus and Schuster, 1977; ZC94 – Zhao and Corbett, 1994.


363

increasing x.’’ These results have been confirmed by 23Na NMR spectroscopy. The ˚ for x 11, and then increases to 14.72 A ˚ cell dimension is constant at a ¼ 14.642 A for x ¼ 24. Thermodynamic studies do not appear to have been reported but the same theory as used for the gas hydrates should apply here essentially unchanged. There are differences in cage occupation between the gas hydrate and metalloid structures. Using the Xe CS-I gas hydrate as example, its composition has been measured as {46H2O [5.89Xe] [1.43 Xe]} (Section 7.2.7.2), i.e. partial occupancy of both types of cage. Only full occupancy has been found in all the {E46 [M8]} structures noted above. In the CS-II gas hydrate, with ideal formula {136H2O [8M16] [16M12]}, complete occupancy of the 16-hedral cages was found for eighteen CS-II hydrates studied (Section 7.2.7.2). In {Si136 [Nax]} Reny et al. (1998) found that the Na atoms first filled the eightfold 16-hedral (51264) cages, without changing the cell dimension, and then entered the smaller 16-fold dodecahedral (512) cages, with appreciable effect on a. The remarkable difference is that a specimen with composition {Si136 [Na1]} could be prepared, i.e that the empty {Si136 []} framework required very little stabilization from host-guest interaction, or, in other words, that the free energy difference between the ‘‘polymorphs’’, ‘‘diamond Si’’ and ‘‘empty {Si136 []} framework,’’ although positive, is rather small. A caveat is necessary; these samples have been prepared at high temperatures and examined after quenching to room temperature – but what are the relevant equilibrium conditions? 7.2.6 Clathrasils We give a brief introduction to the classification of clathrasils for the reader unfamiliar with the overall mineralogical background (Liebau, 1985). One starts with the tectosilicates (Liebau, Gies, Gunawardane and Marler, 1986), which are phases containing threedimensional frameworks of corner linked [TO4] tetrahedra. A distinction is then made between dense and porous tectosilicates, the latter having less than 21 [TO4] tetrahedra per ˚ 3. Porosils are a subclass of the porous tectosilicates and have the general formula 1000 A SiO2 nM, indicating three-dimensional four-connected SiO2 frameworks with guest molecules M occupying voids in the framework. The clathrasils have closed pores, i.e. the windows between polyhedra are too small for passage of guest molecules without decomposition of the whole structure, while zeosils have open pores, the windows between the polyhedra being large enough for the passage of guest molecules. The very important zeolites, which will not be discussed here for reasons noted earlier (see Preface), have partial replacement of Si by Al and have the general formula Axþ y/x[Aly Si1yO2] mH2O nM. The smallest polyhedral building units from which the clathrasil framework can be generated are called fundamental polyhedra. The names are assigned as follows: Nonasils are frameworks generated from 9-hedra. Dodecasils are frameworks generated from 12-hedra. Deca-dodecasils have 10-hedra and 12-hedra as fundamental polyhedra. A polytypic symbol is then added : thus dodecasil 1H has an hexagonal one-layer structure, while dodecasil 3C has a cubic three-layer structure based on the same hexagonal layers as in 1H. The name can be followed by a list of the polyhedra in the structure (with


364

fundamental polyhedra enclosed in double brackets) and by a designation of the crystallographic symmetry, as space group (e.g. Fmmm), point group (mmm) or crystal class (orthorhombic). Thus the general formula for the nonasil family is 88SiO2 8M8 8[[M9]] 4M20 (Fmmm) where Mmi is the polyhedron with mi faces and n is the number of such polyhedra in the structure. The (chemical) contents of the polyhedral cages can also be designated if required. The nomenclature, due to Liebau and coworkers, is used in full or abbreviated form as required. The framework densities of the various silica modifications and the clathrasils form a continuous series (Table 7.6) but with a ratio of almost 2 : 1 between that of coesite, the highest density silica, and silica sodalite, the lowest density porosil. The general method of synthesis of clathrasils is from aqueous solutions of silica under hydrothermal conditions using appropriate template molecules and temperature and pressure conditions in order to obtain the clathrasil desired (Table 7.7); Gies (1991) notes that ‘‘the guest species truly act as templates giving rise to pores in the host framework which reflect the geometry of the templates.’’ Durations as long as three months were necessary for some samples. Synthesis in the absence of air showed that ‘‘hilfsgassen’’ (small molecules such as nitrogen and oxygen which play an important role in stabilizing, especially, CS-II gas hydrates (see below)) do not play an important role in stabilising clathrasils (Gunawardane, Gies and Liebau, 1987). Dodecasil 1H has been studied by CP MAS NMR and found to give narrow 29Si lines indicating a well crystallized sample, while the 13C spectrum showed that the template molecule 1-amino-adamantane was retained intact in the structure (Groenen, Alma, Bastein, Hays, Huis and Kortbeek, 1983). The crystals are well formed morphologically and have the hardness of quartz. Structural information for the different types of clathrasils is summarized in Table 7.8. Investigation of the clathrasils has its origins in the study of the rare mineral melanophlogite, found in Sicily (first description by Lasaulx, 1876), Italy, Bohemia, California and FSU and always associated with an appreciable amount of hydrocarbon impurity e.g. the mineral from Bohemia analyzes as C2.2H17.3O5.4S0.1 46SiO2 (Zak, 1972). Kamb (1965) suggested that the material had the CS-I gas hydrate structure and this has been confirmed by later work (Gies, Gerke and Liebau, 1982; Gies, 1983b). At room temperature natural melanophlogite is tetragonal and microtwinned and there is a displacive phase transformation to the cubic phase at 65 C (exact temperature depends on ˚ , space group Pm3n; the sample used composition of sample). At 200 C a ¼ 13.436(3) A ˚ 3) of Table 7.6. Framework densities (Dfr ¼ number of tetrahedra per 1000 A different silica modifications and some porosils. (Reproduced from Gies, Liebau and Gerke, 1982) Phase

Dfr

Phase

Dfr

Coesite Quartz Keatite Cristabolite Tridymite Silica glass Silica ZSM-48

29.2 26.5 25.0 23.3 22.6 22.0 19.8

Nonasil Melanophlogite Dodecasil 3C Dodecasil 1H Decadodecasil 3R Silica sodalite

19.4 18.9 18.6 18.4 17.6 17.4


365

Table 7.7. Experimental conditions for synthesis of some clathrasils (medium is H2O in absence of air; for effects of ethylenediamine and air see Groenen, Alma, Bastein, Hays, Huis and Kortbeek, 1983) t C

Name of clathrasil

Template molecules

Melanophlogite

Kr, Xe 160 60 CH3NH2 Am > Pr > iso-Pr; iso-Bu > Et; Me > Hex, Hep. It was concluded, on the basis of the geometry of the cavities, that the neo-hexyl radical should have the greatest stabilizing ability but no data have yet been presented in support of this contention. 7.2.9 Varieties of structures formed by a particular guest We have already noted that particular guests generally form only one type of gas hydrate because of limitations imposed by the size of the guest, but that the size boundary is diffuse rather than sharp and thus there are some guests which can form both CS-I and CS-II structures (Table 7.10). Such versatility becomes enhanced in the more complicated structure types and different framework arrangements appear at different compositions. The phase diagram between (n-C4H9)4PBr and water provides an example – compounds with hydration numbers of 37.5, 32, 26 and 24 appear (Dyadin, Zelenina, Zelenin and Yakovlev, 1973). 7.2.10 The alkylamine hydrates Preparative work and phase diagrams have been reported by Pickering (1893), Somerville (1931) and Favier, Rosso and Carbonnel (1981). Crystal structures have been determined by Jeffrey and coworkers, and Jeffrey (1969, 1984a) has emphasized the relationships between these structures and those of the gas hydrates and the peralkylonium hydrates (Fig. 7.36). Compositions and some other information about the alkylamine hydrates are summarised in Table 7.15. The gas hydrate structures are based on variations of two (or three) principal arrangements, and the peralkylonium salt hydrates are mainly derived from these, with the

390


large peralkylonium cations acting as structure determining elements. The alkylamine hydrates come from a much greater variety of structural types, each of which represents a different compromise between the demands of clathration and hydrogen bonding. These complexities make it very difficult to predict structure, and even precise stoichiometry, in the alkylamine hydrates. The various modes of interaction of alkylamine molecules with the surrounding water framework (Fig. 7.36) have been summarized as follows by Jeffrey (1969, 1984a) (see also Table 7 of Jeffrey, 1984a): 1.

2.

3.

True clathrate, no hydrogen bonds to the water framework, as in {156H2O [16{(CH3)3CNH2}]} (McMullan, Jeffrey and Jordan, 1967). This is discussed below, with other structures that do not contain dodecahedral cavities. Formation of one donor and one acceptor hydrogen bond which bridge the water oxygen vertices at the opposite ends of a void, as in {104H2O [12{(CH3CH2)2NH}]} (Jordan and Mak, 1967; DETHHC20). Formation of two donor hydrogen bonds across two adjacent oxygens that could form an edge of a polyhedron of the regular gas hydrate type, as in {80H2O

Table 7.15. The alkylamine hydrates and their compositions (see Notes for References S, P, etc.). Nominal compositions are taken from determinations of phase diagrams and older preparative work, and more exact values from crystal structure analyses; some of the compositions refer to idealized structures. Complexes where both nominal compositions and results from crystal structure analyses are available are bracketed together. The lower hydrates are not polyhedral clathrate hydrates but are included for completeness (a) Primary amines as guests MeNH2 EtNH2 CH3(CH2)2NH2 (CH3)2CHNH2 CH3(CH2)3NH2 (CH3)2CHCH2NH2 CH3CH2CH(CH3)NH2 (CH3)3CNH2 CH3(CH2)4NH2 CH3(CH2)5NH2

3(S); 10(P) 0.5(P); f5:5ðP; SÞ; ½5:45 *g; 0.5 (P, C); 3.5 (P, C); f6:5ðCÞ; ½6:5 *g; f8ðP; CÞ; ½7:96 *g 0.5 (C); 1.5 ( C); 3.5 (P, C); 6.5 ( C); f8ðCÞ; ½7:53 *g; 0.5 (C); 1 (C); 1.5 (C); 38(C) 1(C); 1.5(C); 6.5(C); 12(C); 36(P) 0.5 (C); 1 (C); 4 (C); 6.5(C); 21(C) 0.33(C); ).5(C); 1(C); 2.25(C); 6.5(C); f9:75ðCÞ; ½9:75 *g 0.5(C); 2(C); 5.5(P); 37(P) 0.5(C); 1(C);

(b) Secondary amines as guests Me2NH Et2NH (CH3(CH2)2)2NH

f7ðP; SÞ; ½6:9 *g f6:66ðGÞ; ½6:8 *g; f8ðP; S;Þ; 8:10ðGÞ; ½8:66 *g 0.5(P,C); 5.5(P)

(c) Tertiary amines as guests Me3N Et3N

2(P); 7(P); 10(S) 2(S?,C); 3(C); 8(P); f10ðGÞ; ½10:22 *g; 57(C)

Notes: 1 References are given in brackets and are as follows : P Pickering (1893); S Somerville (1931); C Favier, Rosso and Carbonnel (1981), G Glew (1965). 2 The complexes analyzed crystallographically have compositions inserted in square brackets and are marked with asterisks. Literature references are given in the discussion below.


391

[10{(CH3)2CHNH2}]} (McMullan, Jeffrey and Panke, 1970); the structure is related to the CS-II type. 4. Formation of two acceptor hydrogen bonds from a bridging water oxygen, as in {41H2O.[4{(CH3)3N}]} (Panke, 1968) (see Section 7.2.8.5). 5. Replacement of a water oxygen vertex and bridging across a void by a hydrogen bonded dimer of two amine molecules, as in {104H2O [16{CH3CH2CH2NH2}]} (Brickenkamp and Panke, 1973; at 173K; PROAMH) (see Fig. 7.36).

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 7.36. Environments of alkylamine molecules in water clathrate cages : (a) t-Butylamine guest clathrated within a 17-hedron (43596273), in {156H2O.[16{(CH3)3CNH2}]}. This complex is a true clathrate. (b) Diethylamine hydrogen bonded within 18-hedron (51266), in {104H2O [12{(CH3CH2)2NH}]}; there are also irregular cavities (435866) but no dodecahedral cavities. The guests are hydrogen bonded within both types of cavity. (c) and (d) In {41H2O.[4{(CH3)3N}]} there are dodecahedral (512), 15-hedral (51263) and 26-hedral (52462) cavities, the latter being formed from the fusion of two 14-hedra. The hydrogen bonding of trimethylamine within a distorted 15-hedron is shown in (c); in (d) the hydrogen bonding of two trimethylamine molecules and additional water molecules within the 26-hedron is shown. Compare Fig. 7.33. (e) and (f) In {80H2O [10{(CH3)2CHNH2}]} there are 8-hedra (4662), 12-hedra (512), 14-hedra (425864) and 16-hedra (51264). An isopropylamine molecule hydrogen bonded within a 16-hedron is shown in (e) and an isopropylamine molecule hydrogen bonded within a 14-hedron in (f). This structure is related to the HS-II structure type which contains only 12-hedra and 16-hedra. The C and N atoms of the amine molecules are shaded; the water oxygens are the solid vertices and the solid edges are O . . . O hydrogen bonds. The N–H . . . O hydrogen bonds are open. Hydrogen positions are not known. (Reproduced from Jeffrey, 1969.)

392


Fig. 7.37. {104H2O [16(n-propylamine)]} – ORTEP stereoview of water framework. The structure contains 11-hedra (425861) and large irregular cavities formed by fusion of 14-hedra (51262) and 16-hedra (51264). The guest molecule (not shown) is hydrogen bonded to the framework from within the large cavities. (Reproduced from Brickenkamp and Panke, 1973.)

A number of tri-n-alkylamine (phosphine, arsine, stibine) oxides form clathrate hydrates (references in Dyadin and Udachin (1987)), and the crystal structure of tri-nbutylphosphine oxide 34.5-hydrate has been reported (see Section 7.2.8.6). 7.2.11

Structures without pentagonal dodecahedra (some with charged frameworks)

The division of polyhedral clathrate structures into those derived from packing of pentagonal dodecahedra and those without these energetically-favored polyhedra has already been noted (Fig. 7.20). Having considered the first group in some detail, we now pass on to the second, where rather less information is available. These structures are based on the packing of truncated octahedra (this term is used for the (4668) polyhedron, while the cubo-octahedron is (3846) and the truncated cube is (3866); McMullan and Jeffrey, 1965). In contrast to the many possible space-filling combinations of pentagonal dodecahedra with other polyhedra, only one space-filling arrangement of truncated octahedra is possible and thus only one structure type is to be expected. The silica sodalite containing ethylene glycol is one example of such a (neutral framework) structure which has already been mentioned (Fig. 7.26; Richardson et al., 1988). Before embarking on descriptions of the structures, it is worthwhile summarizing new features that become important when the frameworks contain other groups that replace some of the water molecules. The four-connected water molecules in ice and the true clathrates are linked by hydrogen bonds containing disordered half-hydrogens. Replacement of H2O by OH or F results in proton deficiency, while introduction of H3Oþ gives an extra hydrogen; thus Jeffrey (1996) entitles a section dealing with this area ‘‘Proton Disorder, Deficiency, or Excess in Host Lattices.’’ There are many experimental difficulties in determining details of these crystal structures. Hydrogen positions (especially if disordered) can seldom be identified from ambient temperature x-ray diffraction studies, nor can F and O be differentiated, while low-temperature neutron diffraction studies are entirely lacking. The distances between oxygens in weak hydrogen bonds overlap those in O . . . O van der Waals contacts; sometimes polyhedra are clearly broken (a particular hydrogen bond is replaced by a van der Waals contact), but there are


393

Fig. 7.38. Diagram of the packing of truncated octahedra in the body-centered cubic crystals of silica sodalite (where the vertices are the Si atoms of SiO2 groups, cf. Fig. 7.26) and HPF6 5H2O HF (where the vertices are the oxygen atoms of the water molecules). The twofold disordered PF6 anions are enclosed within the truncated octahedra. The significance of the arrows is explained in the caption to Fig. 7.39. (Reproduced from Jeffrey, 1984b.)

many examples of disorder. NMR (H, F) can provide information not obtainable from diffraction. Two systems with charged frameworks – the clathrate hydrates of the strong acids HEF6 (E¼P, As, Sb), HBF4 and HClO4 and the tetramethylammonium hydrates – have been studied in detail; although we discuss them separately, we shall emphasize their interconnections. After a somewhat tortured history, the compound with the nominal composition HPF6 6H2O (structure first determined from powder data by Bode and Teufer, 1955) has been reformulated as 2{4H2O HF H3Oþ [PF6]} per unit cell. The crystals are cubic ˚ at 107K, Z ¼ 2, space group Im (a ¼ 7.544(2) A 3m; Wiebcke and Mootz, 1986); water and HF molecules are disordered over the vertices of a truncated octahedron (4668), which contains the fourfold disordered PF6 anion (Fig. 7.38). The water framework is cationic, in contrast to gas hydrate and alkylamine hydrate frameworks that are neutral, or those of the alkylonium salt hydrates which are anionic. The isotypes, with As, Sb replacing P, have the same structure. The (nominal) HPF6 7.67H2O hydrate has the CS-I gas hydrate structure, with the anions in the 14-hedra (see also Section 7.2.7.5). There are 92 hydrogen bonds with 98 protons. It was suggested that the unit cell composition was {34H2O 6HF 6H3Oþ 6[PF6]}, giving the required 92 hydrogen-bonding protons. Somewhat similar, but less well-established, proposals have been made for (nominal) HBF4 5.75H2O and HClO4 5.75H2O. The PF6 anion is too large to fit into the dodecahedral cavities, which can contain the smaller BF4 and ClO4 anions; hence the compositional differences. Finally in this section, we note the isotypic hexagonal structures of (nominal) HEF6 6H2O (E¼P, As, Sb), which are derived from the HS-I (hypothetical Type IV clathrate hydrate) structure type (see Table 7.4). The HEF6 6H2O structures were

394


determined without taking the superstructure into account (Wiebcke and Mootz, 1988). ˚, The hexagonal unit cell (for HAsF6 6H2O at 148K) has a ¼ 23.428, c ¼ 13.841 A space group P6/mmm. The composition is {46H2O [7X]}, and the polyhedron formula is {46H2O [4662] 3[4258(6)4] 2[51263]}; anions occupy all except the smallest ([4662]) cages. The (6) notation indicates that pairs of vertices are not connected, i.e the 14-hedron is broken. Detailed crystal structure analyses show that there are differences in the stoichiometries of the As and Sb compounds ({33H2O 7H3Oþ 7[AsF6]} and {35H2O 7H3Oþ 7[SbF6]}); perhaps addition of HF will provide the answer but this issue is still open. The existence of pentagonal dodecahedra in the CS-I and HS-I structures points up the need for a flexible attitude to classification schemes. The Me4NOH–H2O phase diagram (Mootz and Sta¨ben, 1992) has, inter alia, clathrate hydrate phases with hydration numbers 4, 4.6 (dimorphic), 5 (dimorphic), 6.67, 7.5 (dimorphic), 8.75 and 10. The 4.6, 5, 6.67 and 7.5 phases have incomplete polyhedra, while the others (apart from 4 (structure not known) and 5, stable below 207K, triclinic, Hesse and Jansen, 1991; TMAMOH03) have complete polyhedra but are proton deficient. The first example of an incomplete polyhedron was found by McMullan, Mak and Jeffrey (1966) in the orthorhombic crystals of the form of (CH3)4NOH.5H2O (space group Cmcm, Z ¼ 4). This is a distorted version of the body-centered cubic structure found in {4H2O HF H3Oþ [PF6]} (see above); the analogous formulation is {5H2O OH [(CH3)4Nþ]}. The water framework (in which the hydroxyl is incorporated, in some as yet undefined way, making it anionic) has 24 hydrogen bonds but there are only 22 hydrogens. The space-filling 14-hedra (4668) of the basic structure are changed into broken polyhedra ˚ ) in order to allow for (44(4)2 66(6)2) with certain edges are expanded (from 2.78 to 4.36 A the proton deficiency and to contain the twofold disordered (CH3)4Nþ cations (Fig. 7.39). The orthorhombic structure has been further refined at 300K by Hesse and Jansen, 1991; TMAMOH04). The monoclinic 6.67 structure (P21/m, Z ¼ 6; SOXLEW (Mootz and Sta¨ben, 1992)) has two broken polyhedra, occupied by cations, and a third type which is vacant. The framework of the stable form of {7.5H2O.[{(CH3)4NOH}]} (m. pt. 277K; tetragonal, space group I4/mcm, Z ¼ 8; FOXLOT01) consists of eight 15-hedra, containing the cations, and four vacant decahedra, the unit cell formula being {60H2O 8OH 8[51263] 4[4258]}; the framework is proton-deficient with 128 protons for 136 hydrogen bonds. The host structure is isostructural with the hypothetical clathrate hydrate called ‘‘tetragonal II’’ (Table 7.4), derived from the crystal structure of {32H2O [i-C5H11PBr]} (Solodovnikov, Polyanskaya, Alekseev, Aladko, Dyadin and Bakakin, 1982). The isomorphous ternary hydroxide, {14H2O 3(OH) [Csþ 2{(CH3)4Nþ}]} (m. pt. 330K; ˚ , Z ¼ 4, intensities measured at 223K) provides the first a ¼ 15.242(8), c ¼ 11.819(6) A example of a metal cation in a polyhedral clathrate hydrate cage (Mootz and Sta¨ben, 1994; WEZXOO). The framework is proton deficient, with 124 protons for 136 hydrogen bonds; the four Cs ions occupy the decahedra. The metastable form (triclinic; FOXLOT10) is appreciably modified from the form and not all the framework atoms are fourconnected. The decahydrate contains a 17-hedron accommodating the twofold disordered cation, and a vacant nonahedron; the unit cell formula is {40H2O 4OH 4[4151066] 4[4356)}(Pnma, Z ¼ 4; Mootz and Seidel, 1990; FOXLUZ10). Mootz and Seidel (1990) have pointed out that the anionic host framework of the ˚ at 323K, Z ¼ 2, high-temperature form of (CH3)4NOH 46H2O (cubic, a ¼ 8.146(3) A


395

b c a

Fig. 7.39. The mode of incorporation of the tetramethylammoniun cation in the water framework of {5H2O OH [{(CH3)4N}þ]}. Later work (Mootz and Seidel, 1990) suggests that the cation is twofold rather than rotationally disordered. In the previous diagram (Fig. 7.38) the arrows show the hydrogen bonds of the analogous {4H2O HF H3Oþ [PF6]} structure which are broken in order to accommodate the tetramethylammoniun cation. (Reproduced from Jeffrey, 1984b.)

m; Mootz and Seidel, 1988; TMAMOH02) is isostructural with the space group Im3 cationic host framework of HPF6 5H2O HF, and can be derived from the hypothetical gas hydrate structure Cub-II (Jeffrey’s VII) (Table 7.4) by statistical substitution of one-sixth of the H2O molecules by hydroxyl ions. The unit cell formula for the host structure is {10H2O 2OH [4668]}; refinement showed that the oxygen positions were 93% occupied. This cubic structure is proton deficient rather than having the incomplete polyhedra of the analogous 5 hydrate. The low-temperature 4.6 ˚ , Pa3, Z ¼ 40; SOXLAS) and has incomplete hydrate is (primitive) cubic (a ¼ 21.493 A polyhedra. ˚ (space group I43d; SOXLIA) and is The 8.75 hydrate is cubic with a ¼ 18.38(2) A isostructural with the alkylamine hydrate {156H2O.[16{(CH3)3CHNH2}]} (McMullan, Jeffrey and Jordan, 1967), which is a true clathrate without hydrogen bonding of guest to the atoms of the framework. The analogous formulation of the 8.75 hydrate is {140H2O 16OH [16{(CH3)4Nþ}]}; the cations are contained within 17-hedral (43596273) cavities, while smaller octahedral (4454) cavities are vacant. The framework is proton deficient with 296 protons for 312 hydrogen-bonded edges. We close this section by noting an uncharged structure not containing pentagonal dodecahedra – the trigonal (T) hydrate structure of Udachin, Ratcliffe and Ripmeester (2001a). This is the dimethyl ether (DME) clathrate hydrate of composition DME.7H2O

396


(there is also a CS-II DME hydrate of composition DME 17H2O) DME 7H2O is tri˚ , space group P321. The structure can be described as gonal with a ¼ 34.995, c ¼ 12.368 A 0 {12P 12T 24T 12U 348 H2O}, where P is the 51263 cage known from bromine hydrate, T is the 51263 cage from the CS-I structure, T 0 is a previously unobserved cage 4351063 and U is a small cage designated 425861 encountered previously in a propylamine hydrate and devoid of guest.

7.3

7.3.1

Hosts with a combination of directional bonds and van der Waals interactions Phenol (and related compounds) as hosts

7.3.1.1 Phenol ˚ , ¼ 90.36 , P21, Phenol itself is monoclinic (a ¼ 6.050(1), b ¼ 8.925(2), c ¼ 14.594(3) A Z ¼ 6, at 123K; [001] taken as unique axis) (Zavodnik, Bel’skii and Zorkii, 1988; PHENOL03). There are pseudo-31 axes along [100] and the molecules are hydrogen bonded ˚. into an helical arrangement around these axes; d(O . . . O) 2.66 A Isomorphous clathrates are formed with a range of small molecules which include Xe, HCl, HBr, HI, H2S, H2Se, SO2, CO2, COS, CS2, CH3Br, CH2Cl2, CH3CHF2 and ˚, H2C¼CHF (Nikitin, 1939). The crystals are rhombohedral, space group R3, a ¼ 12.05 A ¼ 85–86 (depending on guest) (von Stackelberg, Hoverath and Scheringer, 1958). The unit cell contains 12 phenol molecules and, normally, four guest molecules; however, more smaller (e.g. HCl, HBr) or fewer larger molecules (e.g. CS2) may be enclathrated. The molecular arrangement resembles that of the quinol clathrates in that the hydroxyls of

Fig. 7.40. Stereodiagram of the rhombohedral unit cell of the phenol clathrates, where pairs of hexameric units around the rhombohedral lattice points nearest to, and furthest from the observer, have been omitted in order to show the ellipsoidal cage more clearly. Hexagons of hydrogen-bonded hydroxyl groups are shown around the corners of the unit cell; the phenyl rings have been omitted for clarity and are indicated only by lines extending from the rings of oxygens. The cavities containing only one molecule are located about the unit cell corners while the ellipsoidal cavity containing three guest molecules is centred at the unit cell centre. The crystal data come from von Stackelberg, Hoverath and Scheringer, 1958. (Reproduced from MacNicol, 1984a.)

HOSTS WITH DIRECTIONAL INTERACTIONS AND VAN DER WAALS BONDS 397

six phenols are hydrogen bonded together to form a planar hexagon, with the phenyl rings protruding alternately above and below the plane of this hexagon. Two such hexagons are aligned parallel, with enough space between them to enclose a smaller guest molecule (Fig. 7.40); this accounts for the twelve phenols of the unit cell. The structure as a whole is formed by repetition of such groupings about the eight corners of the rhombohedral unit cell, enclosing an ellipsoidal cavity between them. Thus there are two types of cavity in the cell – a larger one, which normally contains three guest molecules, and a smaller one, which normally encloses one guest molecule sandwiched between hexagons of oxygen atoms. With larger guests only the larger cavities are occupied, perhaps with less than the full complement of three guests per cavity, while with smaller guests both cavities are occupied and the larger cavity can contain up to four molecules. The phenol clathrates are appreciably less stable than those of quinol; dissociation temperatures (i.e. the temperature at which the partial pressure of guest reaches 1 atm.) and enthalpies of formation and enclathration have been measured (Table 7.16). Clathration in -phenol is a non-stoichiometric, zeolitic process, the absorption isotherms not following Langmuir’s equation, possibly because the guests accommodated in the elongated cavity interact mutually. The values of H((0)(0)) are not constant from one clathrate to the next and this has been ascribed (Allison and Barrer, 1968) to Table 7.16. Enthalpies of reaction (kJ/mol) for phenol clathrates with different guests Guest

HD AB68

Kr Xe CH4 C2H6 C2H4 CO2 SO2 HCl H2S HBr

22.4 34.3 25.9 35.3 32.0 37.9 44.8

TD(K)

HI

C

Temp. range (K)

H((0)(0))

24.7 44.4 27.2 47.7 60.7 64.9

0.7 0.83 0.65 0.88 0.89 0.86

195–228 238–252 195–223 228–244 227–241 241–249

0.59 2.7 0.25 3.7 13 7.5

NK52

30.7

277

30.6 37.7 30.9 35.4 35.6

263 312 279 298 302

Notes: 1. HD is the enthalpy of the reaction : 1 mole gas(M) þ n/ moles -phenol ) 1 mole fðn= Þ phenol [M]g; HD is the enthalpy of dissociation of the clathrate containing one mole of gas (M) as guest. 2. HI is the enthalpy of the enclathration reaction in -phenol : 1 mole gas(M) þ n/ moles -phenol ) 1 mole fðn= Þ phenol [M]g. 3. H((0)(0)) is the enthalpy of the reaction : -phenol ) -phenol and is equal to ( /n)[HDHI]. 4. TD is the dissociation temperatures (i.e. the temperature at which the partial pressure of guest reaches 1 atm.). 5. C is the critical guest composition required for the clathrate to be stable; these values are assumed to be constant over the temperature ranges shown. References: AB68 – Allison and Barrer, 1968; NK52 – Nikitin and Kovalskaya, 1952.

398


perturbation of the host structure in different ways by different guests. The larger the value of c required before the clathrate phase will form, the more endothermic is H((0)(0)). The statistical thermodynamic theory of clathration due to van der Waals and Plateeuw is only approximately applicable to the phenol clathrates because among its basic assumptions are that the host framework should not be affected by the clathration and that the cavities be occupied by single guest molecules; neither applies to the phenol clathrates. Comparison of pairs of HD values in Table 7.16 suggests that there is a systematic difference between the values from the two sources. 7.3.1.2 Guayacanin as host This compound, isolated from the heartwood of Tabebuia guayacan Hemsl., has an essentially planar molecule which crystallizes as the acetone clathrate {C30H24O4 [1/ 3 with hexagonal cell dimensions 3(C3H6O)]} in the rhombohedral space group R ˚ , Z ¼ 18 (Wong, Palmer, Manners and Jurd, 1976; GUAYAC). a ¼ 24.68, c ¼ 20.59 A ˚ ) in a chair-shaped sixThe molecules are hydrogen bonded (d(O . . . O) ¼ 2.82(1) A membered ring via their hydroxyl groups, alternate molecules pointing up and down (Fig. 7.41). The cage formed by two such groupings contains two acetone molecules in an ordered arrangement; presumably clathrates would also be formed with other suitably sized guests. The major difference from the analogous clathrates of phenol

O

HO

CH3 CH3 H3C

CH3

Fig. 7.41. Stereoview (looking down c, with a horizontal) of the molecular packing in the guayacanin-acetone clathrate. The thermal ellipsoids are drawn at the 50% probability level except ˚ 2 was for atoms of the (disordered) acetone molecules where an arbitary temperature factor of 1.0 A used. (Reproduced from Wong, Palmer, Manners and Jurd, 1976.)


and Dianin’s compound is the nonplanarity of the ring of oxygens, these being displaced ˚. from the mean ring plane by 0.92 A 7.3.2 Dianin’s compound (4-p-hydroxyphenyl-2,2,4-trimethylchroman) and related compounds as hosts This compound (7.1 below) was first prepared by Dianin (1914), who noted its remarkable ability to crystallize together with many organic solvents which were tightly retained in the crystals of the molecular complexes so formed. Complexes with some 50 different guests were reported when the subject was revived after the Second World War (Baker, Floyd, McOmie, Pope, Weaving and Wild, 1956). Among the guests were Ar, SO2, I2, NH3, ethanol, CHCl3, CCl4, n-heptanol, di-t-butylacetylene, 2,2,5-trimethylhex-3-yn-2-ol, cyclopentane, cyclooctane, decalin, glycerol, SF6 and di-t-butylnitroxide. Parallel surveytype crystallographic studies (Powell and Wetters, 1956; Powell, 1964) suggested that these complexes were true clathrates and this has been confirmed by later concerted synthetic studies and crystal structure analyses, principally by MacNicol and coworkers in Glasgow (MacNicol, McKendrick and Wilson, 1978; MacNicol, 1984a). These studies Table 7.17. Definition of Dianin’s compound and its analogues. For Dianin’s compound (7.1) the configuration shown is ( )-4(S) R8 R2

X

R7

R2' R6 CH3

HY

X

Y

R2

R2 0

R6

R7

R8

Serial number

Are clathrates formed?

O S O O O S O S S O O O S S Se

O O S O O O O O O NH O O S NH O

CH3 CH3 CH3 CH3 H CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3

CH3 CH3 CH3 H CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3

H H H H H CH3 H H H H CH3 H H H H

H H H H H H H H CH3 H H CH3 H H H

H H H H H H CH3 CH3 H H H H H H H

7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 7.10 7.11 7.12 7.13 7.14 7.15

Yes Yes Yes Yes (quasi-racemate) Yes (quasi-racemate) Yes Yes Yes No No No No No No No


400

have proved important for a number of reasons. Firstly, directed syntheses have shown that a number of derivatives of Dianin’s compound form clathrates isostructural with the original group, the differences being understandable in terms of the differences in the molecular structures. This is an interesting example of molecular engineering. Secondly, these complexes are structurally similar to the quinol and phenol clathrates, being based on the formation of hexagonal rings of hydrogen-bonded hydroxyl groups, and thus appreciably broaden this structural family. Thirdly, the importance of this structural type has led to suggestions for the design of complex-forming hosts in which the hydrogen bonds are replaced by covalent bonds; these are the ‘‘hexa-hosts’’ discussed below (Section 7.5). Although most of the results have been obtained for Dianin’s compound itself, we shall treat the family as a whole in order to emphasize the relationships among the various members. The molecules to be considered can all be represented by a single structural formula and the individual compounds are defined as in Table 7.17. The chemical numbering starts from 1 at X¼O (S) in the pyrone (thiapyrone) ring. All the molecules have a chiral centre at C(4) (see for example Dianin’s compound; 7.1); 7.4 and 7.5 have in addition chiral centers at C(2). We first consider the behavior of the pure compounds on crystallization from a racemic solution (both enantiomers present in 1 : 1 proportion) because this is relevant to the definition of the clathrates in terms of the phase rule. Dianin’s compound crystallizes as the racemate as do 7.11 and 7.12; all the other compounds are spontaneously resolved on crystallization. Crystal data are summarized in Table 7.18, including results for Dianin’s compound which has been resolved chemically (Brienne and Jacques, 1975; Collet and Jacques, 1976–7). Table 7.18 thus contains three compounds which are essentially isomorphous and one that is isostructural. Dianin’s compound 7.1 is not spontaneously resolved on crystallization but crystallizes in a racemic rhombohedral space group (Table 7.19). Compounds 7.11 and 7.12 are not spontaneously resolved on crystallization but crystallize in achiral monoclinic space groups, which are not isostructural. A sine qua non for complex formation is that both enantiomers are present in equal amounts. This is immediately understandable in terms of the crystal structure of the

˚ ) for Dianin’s compound and analogues. The space Table 7.18. Crystal data (dimensions in A groups are all P212121 and Z ¼ 4. These data are for the neat compounds and not for the clathrates Compound

a

b

c

Resolved?

Are clathrates formed by racemic host?

(S)-7.1 7.3

10.60 10.66

13.30 13.55

10.08 10.50

Yes Yes

7.10 7.9

10.42 11.78

13.69 16.50

10.37 8.48

Chemically Spontaneously (from C6H12) Spontaneously Spontaneously

No No

Reference / REFCODE HMM79a; MPTMCH HMM79c; HPMTCM

References: HMM79a – Hardy, McKendrick and MacNicol, 1979a; HMM79c – Hardy, McKendrick and MacNicol, 1979c.

Table 7.19. Crystal data (300K unless noted otherwise) for the isostructural clathrates of Dianin’s compound and its analogues. The space groups are all ˚ ) are given for the hexagonal cell containing 18 host molecules R3 except where noted otherwise and dimensions (A Host

Reference/REFCODE

Guest

Host/guest ratio

( )7.1

M84a FKK70; DIANET FK71; DIANHP10 FKK70; DIANCH FK71, AD90, PL90; SIHJEY; SIHJEY10 LG88; GIRBOY

None Ethanol n-Heptanol CHCl3 CCl4 (at 140K)

6 6 6 6

( )7.2

( )7.15

MMW69; HPTHCR MW71; TCHHXO HMMK79 MMKW87; FIYFAU MMKW87; FIYFEY

( )7.3 ( )(S) 7.1þ(þ)(2R,4R)7.4 (þ)(2R,4S) 7.5þ(þ)(2R,4R)7.4 ( )7.5 ( )7.6

HMMW79; MPMCHR10 M84 M84 GHMM79; PMHBZP10 HMM79; METCCP

( )7.7 ( )7.8

M84a HMM79; MSOCYO10 (cyclooctane)

Xe x Ten carboxylic acids (see below) Ethanol 2,2,5-Trimethylhex-3-yn-2-ol Di-t -butylacetylene n-Hexane Ethanol (72 host molecules in unit cell) CCl4 CCl4 CCl4 CCl4 Cyclopentane (also with guests n-pentane and cyclo-hexane, both 6 : 1) Presumably none Cyclooctane (also with guests cyclopropene, cyclopentane, cyclohexane, CCl4 (all 6 : 1); C6H6 (6 : 1.2); toluene, cycloheptane (6 : 1.33)

a

c

6 : 1.4

26.94 26.969 27.12 27.116 27.147 26.912 27.023

10.94 10.990 11.02 11.023 10.939 10.901 10.922

6 6 6 6 6

: : : : :

2 1 1 1 2

27.81 27.91 28.00 28.225 57.42

10.90 10.99 11.08 10.859 10.817

6 6 6 6 6

: : : : :

2 1 1 1 1

27.063 26.94 26.64 26.936 29.22

12.074 11.19* 11.24* 10.796 10.82

32.392 33.629

8.423 8.239

: : : :

2 1 1 1

Notes and References: * Quasiracemates; space group R3. x Structures of ten carboxylic acid complexes (formic, acetic, trifluoroacetic, propanoic, pentanoic, isobutanoic, hexanoic, heptanoic and octanoic) have been reported by Small (2003). The ˚ . The dispositions of the compositions (( ) 7.1) : (acid) are 6 : 2 for formic and acetic acid guests and 6 : 1 for the others, and the cell dimensions lie in the range 27.13–27.30, 10.93–11.20 A disordered guest in the cavities were investigated by difference electron density maps. AD90 – Abriel, DuBois, Zakrzewski and White, 1990; FK71 – Flippen and Karle, 1971; FKK71 – Flippen, Karle and Karle, 1971; GHMM79 – Gall, Hardy, McKendrick and MacNicol, 1979; HMM79 – Hardy, McKendrick and MacNicol, 1979; HMMW79 – Hardy, McKendrick, MacNicol and Wilson, 1979; LG88 – Lee, Gabe, Tse and Ripmeester, 1988; M84 – Hardy and MacNicol, unpublished – see MacNicol, 1984; M84a – Mills, MacNicol and Wilson, unpublished – see MacNicol, 1984a; M84a – MacNicol, 1984a; MMKW87 – MacNicol, Mallinson, Keates, and Wilson, 1987; MMW69 – MacNicol, Mills and Wilson, 1969; MW71 – MacNicol and Wilson, 1971; PL90 – Pang, Lucken and Bernardinelli, 1990.

402


complexes of Dianin’s compound, and by extension, those of its analogs. These fall into a number of isomorphous groups, all of which taken together are isostructural (Table 7.19). Racemic Dianin’s compound (without guests) and its (many) clathrates are all isomorphous; thus the clathrates are primary interstitial solid solutions of guests in the host framework. This is an unusual but not unprecedented situation in the crystal chemistry of organic inclusion complexes – other examples are the clathrates of -quinol (Section 7.2.1.1) and some clathrates of -trimesic acid (Section 10.3). For all these examples addition of the guest is not necessary in order to stabilize the empty host framework, in contrast to the situation in most other inclusion complexes. The only other compound, of those listed in Table 7.19, which has been shown directly to form primary interstitial solid solutions is 7.7. The other clathrates of Table 7.19 are possibly also primary interstitial solid solutions but crystal data for the neat hosts appears to be lacking. There is spontaneous resolution on crystallization for compound 7.3 (Table 7.18) so the (empty) racemate is less stable than the conglomerate of enantiomers; addition of guest leads to stabilization of the racemate. The clathrates all have essentially the same structure, based on the linking of six molecules by hydrogen bonds to form centrosymmetric hexagonal rings of hydroxyl (or thiol) groups, with alternate molecules (of alternating chirality) pointing up and down (this description must be slightly modified for the two pseudo-racemates). The groups of six molecules are stacked one above the other, with the mean planes of the hexagonal rings normal to the c -axis direction of the hexagonal unit cell. Thus a cage is formed with one six-ring of oxygens (or sulphurs) forming the floor of the cage and the next six-ring, one unit cell away along c , forming the ceiling of this cage (and the floor of the next). The walls of the cage are formed by the remaining portions of the host molecules, three pointing up from the floor of the cage and three pointing down from the ceiling (Fig. 7.42). The cage can contain three molecules of a smaller guest (e.g. methanol), two of larger guests (e.g. Xe, ethanol or CHCl3) or one of a still larger guest, such as n-heptanol. The smaller guests are often disordered while n-heptanol was found in a gauche conformation, different from its usual extended conformation.12 The linear acetylene 2,2,5-trimethylhex3-yn-2-ol lies along the 3 axis of the unit cell, with statistical disorder of OH and CH3 to conform to this symmetry. In the Xe complex, the two Xe atoms in a cage were found at the (symmetry-related) positions z 0.3 and 0.7. As the average occupancy was 0.71 for the sample used, some cages must be singly occupied with a single location for Xe. Cell dimensions of 7.1 and {7.1.[1.8(ethanol)]}have been measured (using polycrystalline samples) over the range 10–300K (Zakrzewski, White and Abriel, 1990); extrapolation to 0K (Table 7.20) allows examination of the result of incorporation of guest without the complications of anharmonic effects present at room temperature. The effects are rather small at 0K, a small contraction in a and a small expansion in c, giving a contraction of 0.2% in volume. It seems that the pre-prepared cavity accepts the guest molecules essentially without change in its shape. The major change on heating to 300K is the increase in the size of a, and the thermal vibrations of the guest lead to an appreciable increase in overall volume (2.5% for 7.1 and 3.4% for the clathrate). The effects of 12 n-Heptane is not completely extended in its channel inclusion complexes with 1-phenylalkyl-9-anthroates (Lahav, Leiserowitz, Roitman and Tang, 1977; PPANTR, ZZZBKM).


temperature are appreciably larger than those of clathration. Both these measurements, and those of specific heat (also for the CCl4 clathrate) (White and Zakrzewski, 1990) show that there are no phase changes in the 0–300K range. The molar heat capacity of the enclathrated guests can be calculated from the measurements and compared with values for the neat guests (cf. Fig. 7.9). CCl4 behaves in the clathrate much as in the bulk, but the pair of hydrogen bonded ethanol molecules in the cage shows differences from bulk (solid and liquid) ethanol. The arrangement described above resembles those found in the clathrates of phenol and guayacanin. Comparison of the dimensions of the six-membered rings (Table 7.21) shows that there are appreciable differences among them, both in regard to O . . . O distances (a measure of hydrogen bond strength) and the degree of puckering. Clearly there is more variability in the strengths of the hydrogen bonds in the clathrates of Dianin’s compound and its analogs than in the quinol clathrates. Values for the phenol complex are not given because of the (inevitable) lack of precision of an analysis carried out in 1958. In the clathrates of Dianin’s compound (7.1) a waist is formed at z c/2 by the 2-methyl group syn to the p-hydroxyphenyl moiety, and the cavity has the form of an ˚ long, with maximum and minimum net diameters (i.e. after allowing hourglass about 11 A ˚ and 4.3 A ˚ for the van der Waals envelopes of the atoms of the host molecules) of 6.1 A respectively (Figs. 7.42(a) and 7.43(a)). Very similar cavity shapes are found in the closely isomorphous clathrates of 7.2 and 7.15, with ether O replaced respectively by S and Se. When the 2-methyl group is removed to give 7.5, the new compound still acts as an host but the cavity is broadened at the waist (Figs. 7.42(b) and 7.43(b)). However, the greatest flattening of the cavity is found in the clathrates of 7.7 (addition of methyl in ˚ and the cavity has position 8) (Figs. 7.42(c) and 43(c)), where the c-axis shortens to 8.2 A the shape of a ‘‘Chinese lantern.’’ MacNicol notes that ‘‘this change in cavity geometry is reflected in selective clathration properties.’’ Cyclopentane is selectively enclathrated in 7.7 when recrystallized from an equimolar mixture of cyclopentane, cyclohexane and cycloheptane, while 7.8, with a rounder cavity, favors cyclohexane. Although the clathrates of 7.1 and 7.2 are isomorphous and structurally very similar, their 6-methyl analogs 7.6 and 7.11 pack in quite different ways. Now, referring back to Table 7.17 and bearing in mind the descriptions given above of the various clathrates, it is of interest to see why six potential host compounds do not ˚ ) and cell volumes (A ˚ 3) of 7.1 and {7.1.[1.8(ethanol)]} extrapolated Table 7.20. Cell dimensions (A to 0K, and 300K values (from Table 7.17) Compound

a(0)

( )7.1 26.69 {( )7.1 [1.8(ethanol)]} 26.60 Effects of clathration* a(0) 0.09 Effects of temperature a(300) a(0) ( )7.1 0.25 {( )7.1 [1.8(ethanol)]} 0.37 * X(T)¼X(clathrate, T) X(neat host, T).

a(300)

c(0)

c(300)

V(0)

V(300)

26.94 26.969 a(300) 0.029 c(300) c(0) 0.07 0.07

10.87 10.92 c(0) 0.05 V(300) V(0) 170 231

10.94 10.990 c(300) 0.05

6706 6691 V(0) 15

6876 6922 V(300) 46


404

˚ ) of the six-membered rings of hydroxyl (thiol) Table 7.21. Comparison of the dimensions (A ˚ ) is groups in clathrates in which these rings are important structural elements. ‘‘Puckering’’ (A defined as the mean (unsigned) displacement of oxygen (sulphur) atoms from their mean plane; representative values are given. For host/guest ratios see Table 7.19; references to quinol complexes are in Table 7.1 and to Dianin-type complexes in Table 7.19 Host

Guest

d(O . . . O)

Puckering

-quinol -quinol

none none H2S SO2 CH3OH HCl CH3NC CH3CN Xe acetone ethanol CHCl3 ethanol 2,2,5-trimethylhex-3-yn-2-ol di-t–butyl-acetylene n-hexane CCl4 CCl4 CHCl3 cyclooctane

2.677(3) 2.678(3) 2.696(1) 2.727(6), 2.733(6) 2.653(5), 2.779(5) 2.61(1), 2.77(1) 2.779(6), 2.800(6) 2.792, 2.788; 2.785, 2.782; 2.745, 2.773 2.705(2) 2.81(2) 2.85 2.877(3) 2.96(1) 3.03(1) 3.07(1) 2.975(3) 3.76(1) (d(S . . . S)) 2.767(3)

0.059 0.036

guayacanin 7.1 7.2

7.15 7.3 7.5 7.7 7.8

2.78(1)

0.049

0.035 0.92 0.21 0.22 0.24 0.01 0.26 0.38 0.35

form clathrates. The two amino compounds can be eliminated at once as this group cannot form a six-membered hydrogen bonded ring of the type required. It would appear that methyl groups in position 7 interfere with the lateral packing of the columns of cages so this may account for the failure of 7.9 and 7.12 to form clathrates. However, 7.2 and 7.13 differ only in having OH and SH groups respectively, which difference is no hindrance to formation of clathrates by both 7.2 and 7.3; similarly 7.6 and 7.11 have thioether and ether rings respectively, which difference is no hindrance to formation of clathrates by both 7.1 and 7.3. It seems that the tuning is fine indeed and that quantitative calculations rather than qualitative considerations will be needed to account for these delicate distinctions. We have emphasized above that the formation of six-membered rings of hydroxyl groups (also one thiol group) is a structural feature common to the hosts discussed in this section and to the quinol clathrates. How uniform are these rings? We have listed in Table 7.21 the values of d(O–H . . . O) and the degree of puckering (the average (unsigned) deviation of the oxygens (S) from their best plane). The strength of the ˚ ) to weak (d > 3.0 A ˚ ); the rings range hydrogen bonding ranges from strong (d < 2.7 A from planar (e.g. -quinol) to appreciably puckered ({3(guayacanin) [acetone]}. Any


(a)

(b)

(c)

Fig. 7.42. Comparative stereoviews of the host packing in (a) the CCl4 clathrate of 7.1, (b) the CCl4 clathrate of 7.5, and (c) the neat framework of 7.7. The guest molecules are not shown. (Reproduced from MacNicol, 1984a.)

complete discussion of the thermodynamics of these clathrates would have to take into account variations in the framework, a feature usually neglected. Molecular mechanics and thermodynamic calculations have been made for n-alkane complexes (C5H12 to C9H20) of Dianin’s compound (Iwashiro, 1993). These will be stable if Gincl þ Gvap < 0, and this was found to hold for appropriate mixtures of conformations of the C5 to C8 guests. The all-anti conformation of C5H12 and C6H14


406

(a)

(b)

+C

(c) +C

1.0

1.0

7.1 Å 4.2 Å (7.1 Å) 6.1 Å

0.5

7.7 Å

0.3

6.3 Å

0.5 0.3

5.2 Å 2.6 Å

2.8 Å

2.6 Å

0.0

0.0

0 1 2 3Å

Fig. 7.43. Vertical sections through the cavities in the clathrates of (a) 7.1, (b) 7.5 and (c) 7.7. (Reproduced from MacNicol, 1984a.)

comprises 90% of all the probable conformations in the cavity but mixtures of other conformations are found for C7H16 and C8H18. The calculations indicate that the inclusion complexes are enthalpy stabilized at 298K.

7.4 7.4.1

Van der Waals linked hosts Tetraphenylene as host

The formation of complexes of tetraphenylene (tetrabenzo[a,c,e,g] cyclooctatetraene) was reported at the time of its first synthesis (Rapson, Shuttleworth and van Niekerk, 1943); these were all of composition {2C24H16 [G]}, where G ¼ benzene, CCl4, dioxan, pyridine, CHCl3 or acetone. Crystal structures have been determined at 300K for a number of complexes (Herbstein, Mak, Reisner and Wong, 1984; Wong, Luh and Mak, 1984), and the subject has been reviewed (Mak and Wong, 1987, 1996). The crystals are isomorphous ˚ , c 18.8 A ˚ , space group P42/n, Z ¼ 2). The structure of {2C24H16.[C6H6]} is (a 9.8 A shown in Fig. 7.44, where it will be seen that the tetragonal crystals contain stacks of tetraphenylene molecules separated by channels parallel to [001] in which the disordered guest molecules (located on sites of 4 symmetry) are included. Even those guest molecules (e.g. CCl4) which could conform to this point symmetry are found to be disordered. ˚ , ¼ 100.56 , As neat tetraphenylene is monoclinic (a ¼ 15.628, b ¼ 13.126, c ¼ 16.369 A 3 ˚ , space group C2/c, Z ¼ 8; Irngartinger and Reibel, 1981; BASCIH), the V ¼ 3301 A complexes constitute separate phases in the binary tetraphenylene–guest phase diagrams. The unit cell dimensions (Table 7.22) show an interesting adaptation of host location to the spatial requirements of the guests; the isomorphous structures can be divided into three groups according to the nature of the guest. The quasi-planar molecules of Group A form a fair progression, as do the quasitetrahedral molecules of Group B; however, the unit cell of tetrahedral CCl4 is more

VAN DER WAALS LINKED HOSTS

407

Fig. 7.44. Stereodiagram of the crystal structure of {2(tetraphenylene).[C6H6]}, viewed approximately down the [001] axis of the tetragonal unit cell. Both orientations of the guest molecule are shown. (Reproduced from Herbstein, Mak, Reisner and Wong, 1984.)

˚ ) of tetraphenylene channel inclusion complexes arranged according Table 7.22. Cell dimensions (A ˚3 to nature of guest. The units of volume are A Group Guest

a

c

A

tetrahydrofuran dioxane benzene

9.906(1) 9.968(1) 10.069(1)

18.503(5) 1815.7 18.553(5) 1843.5 18.431(5) 1868.6

77 97 109

119

cyclohexane

10.073(1)

18.712(3) 1898.6

124

140

18.46(1) 18.491(6) 18.546(6) 18.593(3) 18.633(5) 18.647(4) 18.932(6)

78 81 90 91 101 108 108

82

B

C

CH2Cl2 (CH3)2CO CH2Br2 CHCl3 CH3CHBrCH3 CH2BrCH2CH3 CCl4

9.892(5) 9.902(2) 9.935(2) 9.952(2) 9.973(1) 10.004(1) 9.930(2)

Unit cell Cavity Guest volume volume volume

1806 1813.0 1830.6 1831.5 1853.3 1866.2 1866.8

95 104

117

Reference/ REFCODE

HM82; BESXEC HM82; BESXOM (138K) HM82, HMRW84; BESYIH10 (115K) HM82, HMRW84; BESYON10 (153K) HM82; BESWUR HM82, BESXAY (183K) HM82; BESXIG (185K) HM82; BESWOL HM82; BESXUS HM82; BESYUZ (10 kBar) WLM84; BESYED10

References: HM82 – Huang and Mak, 1982; HMRW84 – Herbstein, Mak, Reisner and Wong, 1984; WLM84 – Wong, Luh and Mak, 1984.

elongated along [001] than one might have expected. Low temperature structures in which the guests are (hopefully) ordered would be needed to account for these anomalies. The approximate cavity volumes (¼1/2(Vcomplex – Vtetra), where Vtetra is the molecular volume of tetraphenylene as obtained from the cell dimensions of neat


408

tetraphenylene) match those of the pure guests apart from a systematic discrepancy of 8%, which could indicate that the value used for Vtetra is not quite appropriate for the complexes. Two complexes of substituted tetraphenylenes are noted here as possible springboards for further study. Tetranitrotetraphenylene (positions of nitro groups unknown) gives a complex C24H12N4O8.CCl4 of unknown structure (Rapson, Shuttleworth and van Niekerk, 1943). Perfluorotetraphenylene forms a 1 : 1 complex with ferrocene, the structure of which has been reported (OCUJIF) but not discussed.

7.5

Hexahosts and related compounds

On the basis of the widespread occurrence of the six-membered hydrogen ring of hydroxyl groups in clathrates (Section 7.2), it has been argued that hexa-substituted benzenes should be able to function as hosts in clathrate inclusion complexes. The analogy is illustrated in Fig. 7.45 and has been extensively tested by synthetic and structural studies. About half of some 70 potential hosts formed inclusion complexes, and about a dozen crystal structures had already been reported many years ago (MacNicol, 1984b). The hexa-host analogy and related topics have been reviewed by MacNicol and Downing (1996). We give only two illustrative examples of clathrate formation. Hexakis(phenylthio)benzene (–ZR¼–S–C6H5) forms a series of isomorphous rhombohedral complexes with host/guest ratio 1 : 2 with the similar guest molecules CCl4, CCl3Me, CCl3Br, CCl3NO2 C H

O

S

S S CH2

S

S S

Fig. 7.45. The hexahost analogy illustrated – on the left the hydrogen-bonded hexamer typical of quinol and phenol clathrates is shown, with the hexa-substituted benzene analogue on the right. The six-membered rings are planar or approximately so, and the substituents point alternately above and below these mean planes. The diagrams are approximately to scale. The CH2S groups may be represented more generally as ZR and some examples are: XAr; where X ¼ O, S, Se and Ar represents a variety of substituted phenyl rings CH2 XAr CH2 X½CH2 n Ar CH2 SO2 C*HðCH3 ÞPh, giving a chiral host.

HEXAHOSTS AND RELATED COMPOUNDS

409

and CCl3CN (Table 7.24). The crystals of the neat host are triclinic so, again, the complexes are separate phases in the host–guest phase diagram (Pang, Brisse and Lucken, 1995); two independent determinations at different temperatures are in good agreement (Table 7.23). There are clear resemblances to the Dianin family. Our second example of a hexa-host is hexakis( p-t-butylphenylthiomethyl)benzene, where –ZR ¼ –CH2S–p-C6H4–C(CH3)3. This forms clathrates with host/guest ratio of 1 : 2 with cyclohexane, cycloheptane, cyclooctane, toluene, iodobenzene, phenylacetylene, 1-methylnaphthalene, 2-methylnaphthalene and bromoform, and with host/guest ratio of 2 : 1 with hexamethyldisilane and squalene (MacNicol, 1984b). When this host was recrystallized from an equimolar mixture of o- and p-xylene, the clathrate was found to contain 95% o-xylene. The 2 : 1 squalene clathrate (Freer, Gilmore, MacNicol and Wilson, 1980; SQUBPT) is triclinic with reduced cell a ¼ 14.710, b ¼ 15.773, c ¼ 19.745 ˚ , ¼ 101.81, ¼ 109.03, ¼ 98.07 , Z ¼ 1 (2 molecules of host and 1 of guest in the A unit cell), space group P 1. The squalene molecules are accommodated in continuous channels running through the crystal; the disorder of the squalene molecules was resolved and it was found that the squalene conformation in the clathrate was different from that in neat squalene at 163K. An interesting, but perverse, example is provided by hexakis( p-hydroxyphenyloxy)benzene, which forms a rhombohedral adduct with six pyridine molecules ˚ , deg, A ˚ 3) for neat crystals of hexakis(phenylthio)benzene (space group P Table 7.23. Data (A 1, ˚, Z¼1. For comparison, the rhombohedral cell of the complexes (Table 7.24) has a 10.75 A 83.2 (Z ¼ 1) Temperature

a/

b/

c/

Molecular volume

Reference/ REFCODE

300K

9.589 68.45 9.561 68.45

10.256 76.92 10.209 76.98

10.645 65.52 10.619 65.52

883

MWB95; ZERJEL

875

PBL95; ZERJEL01

220K

˚ , deg, A ˚ 3) for some inclusion complexes of hexakis(phenylthio)benzene Table 7.24. Crystal data (A (all measurements at 220K except for the CCl4 and CBr4 complexes (300K)). The compositions are C42H30S6 2(guest). The space group is R3, and the hexagonal cell contains 3 formula units Guest

a

c

Molecular volume

Reference/ REFCODE

CCl4 CBr4 Cl3C–CH3 Cl3C–Br Cl3C–NO2 Cl3C–CN

14.263 14.327 14.203 14.184 14.207 14.321

20.715 20.666 20.571 20.623 20.606 20.474

1216 1225 1198 1198 1201 1212

HMW79; HPTBZC MWB95; ZERJIP PBL95; ZAPCUQ PBL95; ZARDAX PBL95; ZARDEB PBL95; ZARDIF

References: HMW79 – Hardy, MacNicol and Wilson, 1979; MWB95 – Michalski, White et al., 1995; PBL95 – Pang, Brisse and Lucken, 1995.

410


per host, and also contains one molecule of water (MacNicol, Mallinson, Murphy and Robertson, 1987; FOPHEX). The perversity arises from the fact that the pyridines are hydrogen bonded to the hydroxy groups and thus this molecular compound really belongs in Chapter 12; however, the water molecules are enclathrated. The analogy has been pushed further by synthesis of octa-substituted naphthalenes; for example, it has been found that octakis(m-tolylthio)naphthalene forms a 1 : 1 clathrate with dioxane and octakis( p-tolylthio)naphthalene a 1 : 2 clathrate (DEFCIA). Octakis(m-tolylthio)naphthalene (DEFCAS) and its dioxane clathrate (DEFCEN) are isomorphous, both crystallizing in space group P4/ncc with Z ¼ 4 (MacNicol, Mallinson and Robertson, 1985). Thus this clathrate is another example of a primary interstitial solid solution. MacNicol (1984b, pp.125–6) has commented that ‘‘an analogy may have its interest but the central questions always are, of course, whether the analogy has any validity and particularly crucial, whether it leads anywhere useful.’’ There is no doubt about the usefulness of the analogy, as even the above very small sample of interesting new hosts and clathrates demonstrates. However, there is not much resemblance between the detailed structures of the clathrates of the hexahosts and those of, say, phenol or Dianin’s compound and its analogs. The hexahosts are bulky molecules with complicated shapes, and often these cannot be packed into a crystal without leaving interstices that can be filled by a variety of guests.

7.6

Conclusions and a perspective view

Having marshalled the facts currently known to us (and bearing in mind Jeffrey’s remark about these being only the tip of an iceberg), we summarize what is known about two fundamental, and interconnected, issues raised by Dyadin, Bondaryuk and Aladko (1996) – the stoichiometry of the clathrates and their nature as phases. These authors have suggested that three types of solid solution can be distinguished : these are (a) iskhoric of Types I and II, (b) alloxenic and (c) allokiric. In Type I iskhoric (GK. " !! penetrate) solid solutions the host framework is stable even without the presence of guests. These are usually called primary (or ) solid solutions and are well known in metallurgy as substitutional solid solutions (for example, CCP Cu can take up to 33 atomic % Zn in solid solution without change of crystal structure, forming the -brasses). In the present context these would be interstitial rather than substitutional. An example is racemic Dianin’s compound 7.1 (see Table 7.19). Although no definitive studies appear to have been made, it is quite possible that there is a range of compositions i.e there is no fixed host : guest ratio and the complexes are nonstoichiometric. In Type II iskhoric solid solutions, the clathrate phase has a different crystal structure from that of the pure host, and is only stable when the interstitial cavities have been filled to a certain degree. Examples of hosts are -hydroquinone (not forgetting that -hydroquinone forms Type I iskhoric solid solutions), and H2O (at atmospheric pressure – to avoid complications caused by very small guests such as hydrogen and helium). For both these hosts, the equilibrium composition depends on the ambient conditions and the nature of the guest. The complexes are nonstoichiometric. Tetragonal bromine hydrate is an excellent example (Section 7.2.7.4).

REFERENCES

411

In ternary (or higher) systems, additional possibilities arise. Alloxenic (Greek o another and " o guest) solid solutions have substitution of one kind of guest by another. A simple example is the mixed argon–krypton hydrate (CS-II) where a complete range of solid solutions is formed. This is analogous to a system like Cu–Ni, where a complete range of solid solutions is formed. Other combinations of metals can give phase diagrams with limited solid solubility and formation of intermediate phases; analogous situations occur in the polyhedral clathrates. Dyadin et al. (1995) consider ‘‘binary hydrates,’’ where there is segregation of the two kinds of guest into cavities of different kinds (sizes), to differ from alloxenic solid solutions. Presumably the matter should be argued on the basis of the particular system being considered. In allokiric (Greek o host) solid solutions units of the host framework are substituted by units of another kind. Metalloid CS-I structures like {Si38Ga8 [K8]} (Section 7.2.5) are examples; it has not yet been established whether a range of compositions is possible. A more complicated situation occurs in (nominal) HPF6 6H2O, where HF and H3O þ replace framework water molecules (Section 7.2.11). Here the requirements of charge balance limit compositional variability. In order to place our present treatment of Polyhedral Clathrates in a wider perspective, it is useful to compare it with the recent comprehensive accounts of Jeffrey (1996; hydrate inclusion compounds) and Dyadin and Belosludov (1996; stoichiometry and thermodynamics of clathrate hydrates). Our treament is at a more introductory level than either of these, but it is broader in the sense that we have included metalloid structures and the clathrasils in addition to inclusion hydrates of various kinds; we have also attempted to weave together structural chemistry and thermodynamics. Jeffrey has written a very comprehensive account of the structural chemistry of his title theme, and who could hope to match his experience and understanding in this area? Dyadin and Belosludov have included a vast amount of material on the physical chemistry of systems ranging across the whole field, and connected this thermodynamic approach with the structural chemistry, based on many years of theoretical and experimental studies in both areas. After obtaining an overall picture from the present chapter, a reader could not do better than progress to the deeper and more extensive accounts given by these authors.

References Abrahams, B. F., Hoskins, B. F., Liu, J. and Robson, X. X. (1991). J. Am. Chem. Soc., 113, 3045–3051. Abriel, W., DuBois, A., Zakrzewski, M. and White, A. (1990). Can. J. Chem., 68, 1352–1356. Adam, D. (2002). Nature, 415, 913–914. Alekseev, V. I., Gatilov, Yu. V., Polyanskaya, T. M., Bakakin, V. V., Dyadin, Yu. A. and Gaponenko, L. A. (1982). J. Struct. Chem., USSR, 23, 395–399. Allen, K. W. and Jeffrey, G. A. (1963). J. Chem. Phys., 38, 2304–2305. Allison, S. A. and Barrer, R. M. (1968). Trans. Farad. Soc., 64, 549–556. Appenzeller, T. (1991). Science, 252, 1790–1796. Baker, W., Floyd, A. J., McOmie, J. F. W., Pope, G., Weaving, A. S. and Wild, J. H. (1956). J. Chem. Soc., pp. 2010–2017. Barrer, R. M. and Edge, A. V. J. (1967). Proc. Roy. Soc. Lond., A300, 1–24. Barrer, R. M. and Ruzicka, D. J. (1962). Trans. Farad. Soc., 58, 2262–2271.

412


Barrer, R. M. and Stuart, W. I. (1957). Proc. Roy. Soc. Lond., A243, 172–189. Belosludov, V. R., Dyadin, Yu. A., Chekhova, G. N. and Fadeev, S. I. (1984). J. Incl. Phenom., 1, 251–262. Belosludov, V. R., Dyadin, Yu. A., Chekhova, G. N., Kolesov, B. A. and Fadeev, S. I. (1985). J. Incl. Phenom., 3, 243–260. Belosludov, V. R., Lavrentiev, M. Yu. and Dyadin, Yu. A. (1991). J. Incl. Phenom. Molec. Recognit., 10, 399–422. Berecz, E. and Balla-Achs, M. (1983). Gas Hydrates. Elsevier, Amsterdam and New York, pp. 343. Berry, R. S., Rice, S. A. and Ross, J. (1980). Physical Chemistry, Wiley, New York etc., see pp. 1245–1249. Bertie, J. E. and Jacobs, S. M. (1982). J. Chem. Phys., 77, 3230–3232. Beurskens, P. T. and Jeffrey, G. A. (1964). J. Chem. Phys., 40, 2800–2810. Bibby, D. M. and Dale, M. P. (1985). Nature (Lond.), 317, 157–158. Birchall, T., Frampton, C. S., Schrobilgen, G. J. and Valsdo´ttir, J. (1989). Acta Cryst., C45, 944–946. Blanc, E., Restori, R., Schwarzenbach, D., Bu¨rgi, H.-B., Fo¨rtsch, M., Venugopalan, P. and Ermer, O. (2000). Acta Cryst., B56, 1003–1010. Bobev, S. and Sevov, S. C. (1999). J. Am. Chem. Soc., 121, 3795–3796. Bode, H. and Teufer, G. (1955). Acta Cryst., 8, 611–614. Boeyens, J. C. A. and Pretorius, J. A. (1977). Acta Cryst., B33, 2120–2124. Bonamico, M., Jeffrey, G. A. and McMullan, R. K. (1962). J. Chem. Phys., 37, 2219–2231. Brammer, L. and McMullan, R. K. (1993). ACA Annual Meeting, Albuquerque, New Mexico, Abstract II01. Brickenkamp, C. S. and Panke, D. (1973). J. Chem. Phys., 58, 5284–5295. Brienne, M. J. and Jacques, J. (1975). Tetr. Letts., pp. 2349–2352. Byk, S. Sh., Makogon, Yu. F. and Fomina, V. I. (1980). Gas Hydrates (in Russian), Khimiya, Moscow. Cady, G. H. (1981). J. Phys. Chem., 85, 3225–3230. Cady, G. H. (1983a). J. Chem. Educ., 60, 915–918. Cady, G. H. (1983b). J. Phys. Chem., 87, 4437–4441. Cady, G. H. (1985). J. Phys. Chem., 89, 3302–3304. Calvert, L. D. and Srivastava, P. (1969). Acta Cryst., A25, S131. Cambier, R. and Brochet, A. (1895). Bull. Soc. Chim. France [3], 13, 392–418. Carbonnel, L. and Rosso, J. C. (1973). J. Sol. State Chem., 8, 304–311. Chan, T. L. and Mak, T. C. W. (1983). J. Chem. Soc., Perkin Trans. II, pp. 777–781. Chekhova, G. N., Polyanskaya, T. M., Dyadin, Yu. A. and Alekseev, V. I. (1975). J. Struct. Chem. USSR, 16, 966–971. Chersky, N. N., Makogon, Yu. F. and Medovskii, D. I. (1970). Chem. Abstr., 73, 79123f. Child, W. C., jun. (1964a). Quart. Revs., 18, 321–346. Child, W. C., Jr., (1964b). J. Phys. Chem., 68, 1834–1838. Chu, T. L., Chu, S. S. C. and Ray, R. L. (1982). J. Appl. Phys., 53, 7102–7103. Chu, T. L., Chu, S. S. C., Rosenstein, R. D. and McMullan, R. K. (1982). ACA Summer Meeting, Abstract B7. Claussen, W. F. (1951). J. Chem. Phys., 19, 259–260, (erratum p. 662); 1425–1426. Collet, A. and Jacques, J. (1976–7). Isr. J. Chem., 15, 82–83. Collins, M. J., Ratcliffe, C. I. and Ripmeester, J. A. (1990). J. Phys. Chem., 94, 157–162. Cotton, F. A. and Wilkinson, G. (1980). Advanced Inorganic Chemistry – A Comprehensive Text, 4th Edition, Wiley, New York etc., p. 532. Cox, J. L. ( (1983). Editor), Natural Gas Hydrates : Properties, Occurrence and Recovery, Butterworths, London. Cros, C., Pouchard, M. and Hagenmuller, P. (1970). J. Sol. State Chem., 2, 570–581.

REFERENCES

413

Davidson, D. W. (1973). ‘‘Clathrate hydrates’’ in Water – A Comprehensive Treatise, edited by F. Franks, Plenum Press, London and New York, Vol. 2, pp. 115–234. Davidson, D. W., Calvert, L. D., Lee, F. and Ripmeester, J. A. (1981). Inorg. Chem., 20, 2013–2016. Davidson, D. W., Davies, J. E. D., Gough, S. R., Leaist, D and Ripmeester, J. A. (1990). J. Incl. Phenom., 8, 187–197. Davidson, D. W., Davies, M. M. and Williams, K. (1964). J. Chem. Phys., 40, 3449–3450. Davidson, D. W., Desando, M. A., Gough, S. R., Handa, Y. P., Ratcliffe, C. I., Ripmeester, J. A. and Tse, J. S. (1987a). J. Incl. Phenom., 5, 219–223. Davidson, D. W., Desando, M. A., Gough, S. R., Handa, Y. P., Ratcliffe, C. I., Ripmeester, J. A. and Tse, J. S. (1987b). Nature (Lond.), 328, 418–419. Davidson, D. W., Gough, S. R., Lee, F. and Ripmeester, J. A. (1976). Rev. Chim. Mine´rale, 14, 447–463. Davidson, D. W., Handa, Y. P., Ratcliffe, C. I., Tse, J. S. and Powell, B. M. (1984). Nature (Lond.), 311, 142–143. Davidson, D. W., Handa, Y. P. and Ripmeester, J. A. (1986). J. Phys. Chem., 90, 6549–6552. Davidson, D. W., Handa, Y. P., Ripmeester, J. A., Tse, J. S., Dahn, J. R., Lee, F. and Calvert, L. D. (1986). Mol. Cryst. Liq. Cryst., 141, 141–149. Davidson, D. W. and Ripmeester, J. A. (1984). ‘‘NMR, NQR and dielectric properties of clathrates,’’ in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Vol. 3, pp. 69–128, Academic Press, London etc. Davies, J. E. D. (1999). J. Incl. Phenom. Mol. Recognit. Chem., 32, 499–504. Delepine, M. (1895). Bull. Soc. Chim. France [3], 13, 352–365. Delepine, M. (1897). Bull. Soc. Chim. France [3], 17, 110–113. Deming, R. L., Carlisle, T. L., Lauerman, B. J., Muckerman, J. T., Muirhead, A. R. and Child, W. C., Jr. (1969). J. Phys. Chem., 73, 1762–1766. Desando, M. A., Handa, Y. P., Hawkins, R. E., Ratcliffe, C. I. and Ripmeester, J. A. (1990). J. Incl. Phenom. Mol. Recognit. Chem., 8, 3–16. Dharmawardhana, P. B., Parrish, W. R. and Sloan, E. D. (1980). Ind. Eng. Chem. Fundam., 19, 410–414. Dianin, A. P. (1914). J. Russ. Phys. Chem. Soc., 46, 1310–1319; Chem. Abstr., 9, 1903. (1914). Dyadin, Yu. A. and Aladko, L. S. (1977). J. Struct. Chem. USSR, 18, 41–47. Dyadin, Yu. A. and Belosludov, V. R. (1996). ‘‘Stoichiometry and thermodynamics of clathrate hydrates,’’ in Solid State Supramolecular Engineering; Crystal Engineering, Volume 6 of Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, pp. 789–824. Dyadin, Yu. A., Bondaryuk, I. V. and Aladko, L. S. (1995). J. Struct. Chem. USSR, 36, 995–1045. Dyadin, Yu. A., Chekhova, G. N. and Sokolova, N. P. (1987). J. Incl. Phenom., 5, 187–194. Dyadin, Yu. A., Gaponenko, L. A., Aladko, L. S. and Bogatyrova, S. V. (1984). J. Incl. Phenom., 2, 259–266. Dyadin, Yu. A. and Udachin, K. A. (1984). J. Incl. Phenom., 2, 61–72. Dyadin, Yu. A. and Udachin, K. A. (1987). J. Struct. Chem. USSR, 28, 394–432. Dyadin, Yu. A., Zelenina, L. S., Zelenin, Yu. M. and Yakovlev, I. I. (1973). Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. Nauk, pp. 30–33; Chem. Abstr., 79, 115682x (1973). Ermer, O. (1991). Helv. Chim. Acta, 74, 1339–1351. Ermer, O. and Ro¨bke, C. (1993). J. Am. Chem. Soc., 115, 10077–10082. Evans, D. F. and Richards, R. E. (1952). J. Chem. Soc., pp. 3932–3936. Evans, D. F. and Richards, R. E. (1954). Proc. Roy. Soc. Lond., A223, 238–250. Favier, R., Rosso, J.-C. and Carbonnel, L. (1981). Bull. Soc. Chim. France, pp. 225–236.

414


Feil, D. and Jeffrey, G. A. (1961). J. Chem. Phys., 35, 1863–1873. Flippen, J. L. and Karle, J. (1971). J. Phys. Chem., 75, 3566–3567. Flippen, J. L., Karle, J. and Karle, I. L. (1970). J. Am. Chem. Soc., 92, 3749–3755. Fowler, D. L., Loebenstein, W. V., Pall, D. B. and Kraus, C. A. (1940). J. Am. Chem. Soc., 62, 1140–1142. Freer, A., Gilmore, C. J., MacNicol, D. D. and Wilson, D. R. (1980). Tetr. Lett., 21, 1159–1162. Fyfe, C. A. and Gies, H. (1990). J. Incl. Phenom., 8, 235–239. Gall, J. H., Hardy, A. D. U., McKendrick, J. J. and MacNicol, D. D. (1979). J. Chem. Soc., Perkin Trans. II, 376–380. Gallmeier, J., Scha¨fer, H. and Weiss, A. (1969). Z. Naturforsch., 24B, 665–667. Gaponenko, L. A., Polyanskaya, T. M., Alekseev, V. I. and Dyadin, Yu. A. (1983). Third All-Union Conference on the Crystal Chemistry of Inorganic and Coordination Compounds (in Russian), Novosibirsk, p. 57. Gerke, H. and Gies, H. (1984). Z. Kristallogr., 166, 11–22. Gerke, H., Gies, H. and Liebau, F. (1982). In ‘‘Soluble Silicates’’ edited by J. S. Falcone, Jr., ACS Symposium Series 194, pp. 305–318. Gies, H. (1983a). Z. Kristallogr., 162, 82–84. Gies, H. (1983b). Z. Kristallogr., 164, 247–257. Gies, H. (1984). Z. Kristallogr., 167, 73–82. Gies, H. (1986a). J. Incl. Phenom., 4, 85–91. Gies, H. (1986b). Z. Kristallogr., 175, 93–104. Gies, H. (1991). ‘‘Clathrasils and zeosils : inclusion compounds with silica host frameworks,’’ in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Vol. 5, pp. 1–36, Oxford etc., Oxford University Press. Gies, H., Gerke, H. and Liebau, F. (1982). Neues Jahr. Mineral. Monatsh., 3, 119–124. Gies, H., Liebau, F. and Gerke, H. (1982). Angew. Chem. Int. Ed. Engl., 21, 206–207. Gies, H. and Marler, B. (1996). ‘‘Crystalline microporous silicas as host–guest systems,’’ in Solid State Supramolecular Engineering; Crystal Engineering, Volume 6 of Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, pp. 851–883. Glasstone, S. (1947). Textbook of Physical Chemistry, Van Nostrand, New York, see pp. 786–790. Glew, D. N. (1965). Trans. Farad. Soc., 61, 30–36. Glew, D. N. and Rath, N. S. (1966). J. Chem. Phys., 44, 1710–1711. Gough, S. R. and Davidson, D. W. (1971). Can. J. Chem., 49, 2691–2699. Gough, S. R., Handa, Y. P., Ratcliffe, C. I., Ripmeester, J. A. and Tse, J. S. (1987). J. Phys. Colloq., C1, 537–542. Grey, N. R., Parsonage, N. G. and Staveley, L. A. K. (1962). Mol. Phys., 5, 153–159. Grey, N. R. and Staveley, L. A. K. (1963–4). Mol. Phys., 7, 83–95. Groenen, E. J. J., Alma, N. C. M., Bastein, A. G. T. M., Hays, G. R., Huis, R. and Kortbeek, A. G. T. G. (1983). J. Chem. Soc., Chem. Comm., pp. 1360–1362. Gunawardane, R. P., Gies, H. and Liebau, F. (1987). Z. anorg. allgem. Chem., 546, 189–198. Hafemann, D. R. and Miller, S. L. (1969a). J. Phys. Chem., 73, 1392–1397. Hafemann, D. R. and Miller, S. L. (1969b). J. Phys. Chem., 73, 1398–1401. Hagan, M. (1962). Clathrate Inclusion Compounds, pp. 189, Reinhold, New York. Handa, Y. P. (1985). J. Chem. Thermodynamics, 17, 201–208. Handa, Y. P. (1986). J. Chem. Thermodynamics, 18, 891–902. Handa, Y. P. and Tse, J. S. (1986). J. Phys. Chem., 90, 5917–5921. Hardy, A. D. U. and MacNicol, D. D. Unpublished–see MacNicol (1984). Hardy, A. D. U., MacNicol, D. D. and Wilson, D. R. (1979). J. Chem. Soc., Perkin Trans. II, pp. 1011–1019.

REFERENCES

415

Hardy, A. D. U., McKendrick, J. J. and MacNicol, D. D. (1979). J. Chem. Soc., Perkin Trans. II, 1072–1077. Hardy, A. D. U., McKendrick, J. J., MacNicol, D. D. and Wilson, D. R. (1979). J. Chem. Soc., Perkin Trans. II, 729–734. Hazony, Y. and Ruby, S. L. (1968). J. Chem. Phys., 40, 1478–1481. Henkel, G. unpublished. Herbstein, F. H., Mak, T. C. W., Reisner, G. M. and Wong, H. N. C. (1984). J. Incl. Phenom., 1, 301–308. Hesse, W. and Jansen, M. (1991). Z. anorg. allgem. Chem., 595, 115–130. Ho, W. C. and Mak, T. C. W. (1983). Z. Kristallogr., 161, 87–90. Hohmann, H. (1948). Z. anorg. allgem. Chem., 257, 113–130. Hollander, F. and Jeffrey, G. A. (1977). J. Chem. Phys., 66, 4699–4705. Hondoh, T., Anzai, H., Goto, A., Mae, S., Higashi, A. and Langway, C. C. (1990). Jr. J. Incl. Phenom. Mol. Recognit. Chem., 8, 17–24. Huang, N. Z and Mak, T. C. W. (1982). Chem. Commun., pp. 543–544. Irngartinger, H. and Reibel, W. R. K. (1981). Acta Cryst., B37, 1724–1728. Iwashiro, F. (1993). J. Am. Chem. Soc., 115, 2231–2235. Jeffrey, G. A. (1969). Accts. Chem. Res., 2, 344–352. Jeffrey, G. A. (1984a). J. Incl. Phenom., 1, 211–222. Jeffrey, G. A. (1984b). ‘‘Hydrate inclusion compounds,’’ in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Vol. 1, pp. 135–190, Academic Press, London etc. Jeffrey, G. A. (1996). ‘‘Hydrate inclusion compounds,’’ in Solid State Supramolecular Engineering; Crystal Engineering, Volume 6 of Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, pp. 757–788. Jeffrey, G. A. and McMullan, R. K. (1962). J. Chem. Phys., 37, 2231–2239. Jeffrey, G. A. and McMullan, R. K. (1967). Prog. Inorg. Chem., 8, 43–108. Jordan, T. H. and Mak, T. C. W. (1967). J. Chem. Phys., 47, 1222–1228. Kamb, B. (1965). Science, 148, 232–234. Kasper, J., Hagenmuller, P., Pouchard, M. and Cros, C. (1965). Science, 150, 1713–1714. Katz, D. L. (1971a). J. Pet. Technol., 23, 419–423. Katz, D. L. (1971b). J. Pet. Technol., 24, 557–558. Kazankin, Yu. N., Palladiev, A. A. and Trofimov, A. M. Radiokhim.iya, (1972). 14, 847–854; Chem. Abstr., 78, 76317g (1973). King, R. B. (1972). Theor. Chim. Acta (Berlin), 25, 309–318. Kirchner, M. T., Boese, R., Billups, W. E. and Norman, L. R. (2004). J. Am. Chem. Soc., 126, 9407–9412. Kitazawa, T., Nishikiori, S., Kuroda, R. and Iwamoto, T. (1988). Chem. Lett., pp. 1729–1732. Kuhs, W. F. (2004). ‘‘The high pressure crystallography of gas hydrates.’’ in High Pressure Crystallography edited by A. Katrusiak and P. McMillan, NATO Science Series II. Vol. 140, pp. 475–494. Kuhs, W. F. and Lehman, M. S. (1987). J. Phys. (Paris) Colloq. C1, 48, 3–8. Kutty, K. R. N., Sharma, D. K. and Murthy, A. R. V. (1971). J. Ind. Inst. Sci., 53, 154–159. Kvenvolden, K. A. (1994). Ann. N. Y. Acad. Sci., 715, 232–247. Kvenvolden, K. A. and McMenamin, M. A. (1980). U. S. Geol. Surv. Circ., 825, 14 pp. La Rive, A. de, (1829). Ann. Phys. Chim., 40, 401–412. Lahav, M., Leiserowitz, L., Roitman, L. and Tang, C.-P. (1977). J. Chem. Soc., Chem. Comm., pp. 928–929. Lasaulx, A. L. (1876). Neues Jahr. Mineral., p. 250. Lee, F., Gabe, E., Tse, J. S. and Ripmeester, J. A. (1988). J. Am. Chem. Soc., 110, 6014–6019.

416


Liebau, F. (1985). The Structural Chemistry of the Silicates (Structure, Bonding and Classification), Springer-Verlag, Berlin, Heidelberg etc. Liebau, F., Gies, H., Gunawardane, R. P. and Marler, B. (1986). Zeolites, 6, 373–377. Lindeman, S. V., Shklover, V. E. and Struchkov, Yu. T. (1981). Cryst. Struct. Comm., 10, 1173–1179. Londono, D., Finney, J. I. and Kuhs, W. F. (1992). J. Chem. Phys. 97, 547–552. Londono, D., Kuhs, W. F. and Finney, J. L. (1988). Nature (Lond.), 332, 141–142. Long, Y., Zheng, P., Wu, G. and Wang, B. (1987). J. Incl. Phenom. Macrocyclic Chem., 5, 355–362. Lo¨wig, C. (1829). Ann. Phys. Chem., 42, 113–119; Ann. Phys. Chem. (Poggendorf), 16, 376–380. Maartman-Moe, K. (1966). Acta Cryst., 21, 979–982. MacNicol, D. D. (1984a). ‘‘Structure and design of inclusion compounds : the clathrates of hydroquinone, phenol, Dianin’s compound and related systems’’, in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Vol. 2, pp. 1–45, Academic Press, London etc. MacNicol, D. D. (1984b). ‘‘Structure and design of inclusion compounds : the hexa-hosts and symmetry considerations’’ in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Vol. 2, pp. 123–168, Academic Press, London etc. MacNicol, D. D. and Downing, G. A. (1996). ‘‘Symmetry in the evolution of host design,’’ in Solid State Supramolecular Engineering; Crystal Engineering, Volume 6 of Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, pp. 421–464. MacNicol, D. D., Mallinson, P. R., Keates, R. A. B. and Wilson, F. B. (1987). J. Incl. Phenom., 5, 373–377. MacNicol, D. D., Mallinson, P. R., Murphy, A. and Robertson, C. D. (1987). J. Incl. Phenom., 5, 233–239. MacNicol, D. D., Mallinson, P. R. and Robertson, R. B. (1985). J. Chem. Soc., Chem. Comm., pp. 1649–1651. MacNicol, D. D., McKendrick, J. J. and Wilson, D. R. (1978). Chem. Soc. Rev., 7, 65–87. MacNicol, D. D., Mills, H. H. and Wilson, F. B. (1969). J. Chem. Soc. (D), pp. 1332–1333. MacNicol, D. D. and Wilson, F. B. (1971). J. Chem. Soc. (D), pp. 786–787. Mak, T. C. W. (1965). J. Chem. Phys., 43, 2799–2805. Mak, T. C. W. (1982). J. Chem. Soc., Perkin Trans. II, pp. 1435–1437. Mak, T. C. W. and Bracke, B. F. R. (1996). ‘‘Hydroquinone clathrates and diamondoid host lattices,’’ in Solid State Supramolecular Engineering; Crystal Engineering, Volume 6 of Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, pp. 23–60. Mak, T. C. W. and McMullan, R. K. (1965). J. Chem. Phys., 42, 2732–2737. Mak, T. C. W. and Wong, H. N. C. (1987). Top. Curr. Chem., 140, 141–164. Mak, T. C. W. and Wong, H. N. C. (1996). ‘‘Tetraphenylene and related hosts,’’ in Solid State Supramolecular Engineering; Crystal Engineering, Volume 6 of Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol and F. Vo¨gtle, Pergamon, Oxford, pp. 351–369. Makogon, Yu. F. (1974). Hydrates of Natural Gas (in Russian), Nedra, Moscow. Makogon, Yu. F. (1987). La Recherche, 18, 1192–1200. Marler, B., Dehnbostel, N., Eulert, H.-H., Gies, H. and Liebau, F. (1986). J. Incl. Phenom., 4, 339–349. Matsui, S., Terao, T. and Saika, A. (1982). J. Chem. Phys., 77, 1788–1799. Matsuo, T. (1970). J. Phys. Soc. Jpn., 30, 785–793. Matsuo, T. and Suga, H. (1984). J. Incl. Phenom., 2, 49–59. Matsuo, T., Suga, H. and Seki, S. (1970). J. Phys. Soc. Jpn., 30, 785–793.

REFERENCES

417

McAdie, H. G. (1963). Canad. J. Chem., 41, 2137–2143. McAdie, H. G. (1966). Canad. J. Chem., 44, 1373–1385. McCusker, L. (1988). J. Appl. Cryst., 21, 305–310. McMullan, R. K., Bonamico, M. and Jeffrey, G. A. (1963). J. Chem. Phys., 39, 3295–3310. McMullan, R. K. and Jeffrey, G. A. (1965). J. Chem. Phys., 42, 2725–2732. McMullan, R. K., Jeffrey, G. A. and Jordan, T. H. (1967). J. Chem. Phys., 47, 1229–1234. McMullan, R. K., Jeffrey, G. A. and Panke, D. (1970). J. Chem. Phys., 53, 3568–3577. ˚ . (1990). Acta Cryst., B46, 390–399. McMullan, R. K. and Kvick, A McMullan, R. K., Mak, T. C. W. and Jeffrey, G. A. (1966). J. Chem. Phys., 44, 2338–2345. Menke, H. and Schnering, H. G. von, (1973). Z. anorg. allgem. Chem., 395, 223–238. Menke, H. and Schnering, H. G. von, (1976). Z. anorg. allgem. Chem., 424, 108–114. Michalski, D., White, M. A., Bakshi, P. K., Cameron, T. S. and Swainson, I. (1995). Can. J. Chem., 73, 513–521. Mills, H. H., MacNicol, D. D. and Wilson, F. B. Unpublished–see MacNicol (1984). Mootz, D., Oellers, E. J. and Wiebicke, M. (1987). J. Am. Chem. Soc., 109, 1200–1202. Mootz, D. and Seidel, R. (1990). J. Incl. Phenom. Molec. Recognit. Chem., 8, 139–157. Mootz, D. and Sta¨ben, D. (1992). Z. Naturforsch. Teil B, 47, 263–274. Mootz, D. and Sta¨ben, D. (1994). J. Am. Chem. Soc., 116, 4141–4142. Murakami, E., Komukae, M., Osaka, T. and Makita, Y. (1990). J. Phys. Soc. Jpn., 59, 1147–1149. Nakayama, H. (1979). Bull. Chem. Soc. Jpn., 52, 52–56. Nakayama, H. (1987). Bull. Chem. Soc. Jpn., 60, 2319–2326. Nakayama, H. and Usui, H. (1986). Bull. Chem. Soc. Jpn., 59, 833–837. Nakayama, H. and Watanabe, K. (1978). Bull. Chem. Soc. Jpn., 51, 2518–2522. Nesper, R., Curda, J., and Schnering, H. von, (1986). Angew. Chem. Int. Ed. Engl., 25, 350–352. Nikitin, B. A. (1939). Dokl. Akad. Nauk SSSR, 24, 565–567. Nikitin, B. A. and Kovalskaya, M. P. (1952). Izv. Akad. Nauk SSSR, Otdel. Khim. Nauk, pp. 24–30. Nolas, G. S., Weakley, T. J. R. and Cohn, J. L. (1999). Chem. Mater., 11, 2470–2473. Nolas, G. S., Chakoumakos, B. C., Mahieu, B., Long, G. J. and Weakley, T. J. R. (2000). Chem. Mater., 12, 1947–1953. Palin, D. E. and Powell, H. M. (1947). J. Chem. Soc., pp. 208–221. Palin, D. E. and Powell, H. M. (1948). J. Chem. Soc., pp. 815–821. Pang, L., Brisse, F. and Lucken, E. A. C. (1995). Can. J. Chem., 73, 351–361. Pang, L., Lucken, E. A. C. and Bernardinelli, G. (1990). J. Am. Chem. Soc., 112, 8754–8764. Panke, D. (1968). J. Chem. Phys., 48, 2990–2996. Parrish, W. R. and Prausnitz, J. M. (1972). Ind. Eng. Chem., Process Des. Develop., 11, 26–35. Parsonage, N. G. and Staveley, L. A. K. (1959). Mol. Phys., 2, 212–222. Parsonage, N. G. and Staveley, L. A. K. (1960). Mol. Phys., 3, 59–66. Parsonage, N. G. and Staveley, L. A. K., ‘‘Thermodynamic Studies of Inclusion Compounds’’ in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Vol. 3, pp. 1–36, Academic Press, London etc. (1984). Pauling, L. and Marsh, R. E. (1952). Proc. Nat. Acad. Sci. USA, 38, 112–120. Pearson, C. F., Halleck, P. M., McGuire, P. L., Hermes, R. and Wright, A. (1983). J. Phys. Chem., 87, 4180–4185. Peyronel, G. and Barbieri, G. (1958). J. Inorg. Nucl. Chem., 8, 582–585. Pickering, S. U. (1893). J. Chem. Soc. London, 63, 141–195. Platteeuw, J. C. and Waals, J. H. van der, (1959). Rec. trav. Chim., 78, 126–133. Polyanskaya, T. M., Alekseev, V. I., Bakakin, V. V. and Chekhova, G. N. (1982). J. Struct. Chem. USSR, 23, 101–104. Polyanskaya, T. M., Andrianov, V. I., Alekseev, V. I., Bakakin, V. V., Dyadin, Yu. A. and Chekhova, G. N. (1982). Dokl. Akad. Nauk SSSR, 266, 349–353.

418


Powell, H. M. (1964). ‘‘Clathrates’’, in Non-Stoichiometric Compounds, edited by L. Mandelcorn, Academic Press, New York and London, pp. 438–490. Powell, H. M. (1984). ‘‘Introduction’’, in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Vol. 1, pp. 1–28, Academic Press, London etc. Powell, H. M. and Wetters, B. D. P. (1956). Chem. Ind., pp. 256–257. Ramachandran, G. K., Diefenbacher, J., Sankey, O. F., Sharma, R., Marzke, R. F., O’Keefe, M., Gryko, J. and McMillan, P. F. (1999). Mater. Res. Sym. Proc., 507, 483–486. Rapson, W. S., Shuttleworth, R. G. and van Niekerk, J. N. (1943). J. Chem. Soc., pp. 326–327. Reny, E., Gravereau, P., Cros, C. and Pouchard, M. (1998). J. Mater. Chem., 8, 2839–2844. Richardson, J. W., Jr., Pluth, J. J., Smith, J. V., Dytrych, W. J. and Bibby, D. M. (1988). J. Phys. Chem., 92, 243–247. Ripmeester, J. A., Hawkins, R. E. and Davidson, D. W. (1979). J. Chem. Phys., 71, 1889–1898. Ripmeester, J. A. and Ratcliffe, C. I. (1991). ‘‘Solid state NMR studies of inclusion compounds,’’ in Inclusion Compounds, edited by J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Vol. 5, pp. 37–89, Academic Press, London etc. Ripmeester, J. A., Tse, J. S., Ratcliffe, C. I. and Powell, B. M. (1987). Nature, 325, 135–136. Rodger, P. Mark. (1990). J. Phys. Chem., 94, 6080–6089. Roozeboom, H. W. B. (1884). Rec. Trav. C him. Pays-Bas, 3, 26. Roozeboom, H. W. B. (1885). Rec. Trav. C him. Pays-Bas, 4, 65. Scha¨fer, R. and Klemm, W. (1961). Z. anorg. allgem. Chem., 312, 214–220. Schlenker, J. L., Dwyer, F. G., Jenkins, E. E., Rohrbaugh, W. J., Kokotailo, G. T. and Meier, W. M. (1981). Nature (Lond.), 294, 340–342. Schneider, V. E., Tornau, E. E., Vlasova, A. A. and Gurskas, A. A. (1985). J. Incl. Phenom., 3, 235–242. Schnering, H. G. von, and coworkers. (1998). Z. Kristallogr. – New Crystal Structures, 213, 664–678. Schroeder, W. (1926). Die Geschichte des Gashydrate, 98 pp. F. Enke, Stuttgart. Sixou, P. and Dansas, P. (1976). Ber. Bunsen-Gesell. Physik. Chem., 80, 364–389. Sloan, E. D. (1998). Clathrate Hydrates of the Natural Gases, 2nd Ed., Revised and Expanded, Marcel Dekker, New York. Small, R. W. H. (2003). Acta Cryst., B59, 141–148. Solodovnikov, S. F., Polyanskaya, T. M., Alekseev, V. I., Aladko, L. S., Dyadin, Y. A. and Bakakin, V. V. (1982). Sov. Phys. Crystallogr., 27, 151–155. Somerville, W. C. (1931). J. Phys. Chem., 35, 2412–2433. Stackelberg, M. von, and Meinhold, W. (1954). Z. Elektrochem., 58, 40–45. Stackelberg, M. von, and Mu¨ller, H. R. (1954). Z. Elektrochem., 58, 25–39. Stackelberg, M. von, Hoverath, A. and Scheringer, C. (1958). Z. Elektrochem., 62, 123–130. Stackelberg, M. von. (1949). Naturwiss., 36, 327–333, 359–362. Tammann, G. and Krige, G. J. R. (1925). Z. anorg. allgem. Chem., 146, 179–195. Tse, J. S., McKinnon, W. R. and Marchi, M. (1987). J. Phys. Chem., 91, 4188–4193. Tse, J. S., Ratcliffe, C. I., Handa, Y. P. and Powell, B. M. (1986). J. Incl. Phenom., 4, 235–240. Udachin, K. A., Ratcliffe, C. I., Enright, G. D. and Ripmeester, J. A. (1997). Supramol. Chem., 8, 173–178. Udachin, K. A., Enright, G. D., Ratcliffe, C. I. and Ripmeester, J. A. (1997). J. Am. Chem. Soc., 119, 11481–11486. Udachin, K. A., Ratcliffe, C. I. and Ripmeester, J. A. (2001a). Angew. Chem. Int. Ed., 40, 1303–1305. Udachin, K. A., Ratcliffe, C. I. and Ripmeester, J. A. (2001b). J. Phys. Chem., B105, 4200–4204. Udachin, K. A., Ratcliffe, C. I. and Ripmeester, J. A. (2002). J. Supramol. Chem., 2, 400–508. Udachin, K. A. and Ripmeester, J. A. (1999a). Nature (Lond.), 397, 420–423.

REFERENCES

419

Udachin, K. A. and Ripmeester, J. A. (1999b). Angew. Chem. Int. Ed., 38, 1983–1984. Ukegawa, H., Matsuo, T. and Suga, H. (1985). J. Incl. Phenom., 3, 261–267. Villard, P. (1896). Compt. Rend., 123, 377–379. Villard, P. (1897). Ann. Phys. Chem. [7], 11, 289–394. Waals, J. H. van der, and Platteeuw, J. C. (1956). Rec. Trav. Chim., 75, 912–918. Waals, J. H. van der, and Platteeuw, J. C. (1959). Adv. Chem. Phys., 2, 1–57. Wallwork, S. C. and Powell, H. M. (1980). J. Chem. Soc., Perkin Trans. II, pp. 641–646. Weiner, S. J., Kollman, P. A., Nguyen, D. T. and Case, D. A. (1986). J. Comput. Chem., 7, 230–252. Wells, A. F. (1975). Structural Inorganic Chemistry, 4th Edition, see Chapter 3 (Polyhedra and Nets), Clarendon Press, Oxford. Westerhaus, W. and Schuster, H. U. (1977). Z. Naturforsch., 32B, 1365–1367. White, M. A. and Zakrzewski, M., (1990). J. Incl. Phenom. Mol. Recognit. Chem., 8, 215–225. Wiebcke, M. and Mootz, D. (1986). Z. Kristallogr., 177, 291–299. Wiebcke, M. and Mootz, D. (1988). Z. Kristallogr., 183, 1–13. Wo¨hler, F. (1849). Ann. Chem., 69, 294–300. Wong, H. N. C., Luh, T.-Y. and Mak, T. C. W. (1984). Acta Cryst., C40, 1721–1723. Wong, R. Y., Palmer, K. J., Manners, G. D. and Jurd, L. (1976). Acta Cryst., B32, 2396–2400. Wynne-Jones, W. F. K. and Anderson, A. R. (1952). Chem. Abstr., 48, 34c (1954). Za´k, L. (1972). Amer. Mineral., 57, 779–796. Zakrzewski, M., White, M. A. and Abriel, W. (1990). J. Phys. Chem., 94, 2203–2206. Zavodnik, V. E., Bel’skii, V. K. and Zorkii, P. M. (1988). J. Struct. Chem. USSR, 28, 793–795. Zhao, J.-T. and Corbett, J. D. (1994). Inorg. Chem., 33, 5721–5726. Zubkus, V. E., Shamovsky, I. L. and Tornau, E. E. (1992). J. Chem. Phys., 97, 8617–8627.


Chapter 8 Inclusion complexes formed by versatile hosts

. . . As clay in the hands of the potter, Who contracts and expands it at will, So are we in Thy hands, O gracious Lord . . . From the Evening Prayer on the Day of Atonement.

Summary: Versatility in a host can be achieved in a number of different ways, all leading to the capability of a single host species to form a variety of inclusion complexes. A small number of examples has been chosen for illustration. Tri-o-thymotide (TOT) interacts with its neighbors (hosts and guests) only via van der Waals forces, forming clathrates with smaller guests and tunnel inclusion complexes of a number of types with larger or elongated guests; some 70 complexes of various types have been recorded. Hydrogen-bonded trimesic acid (TMA) gives tunnel inclusion complexes, with guests ranging from long-chain hydrocarbons to polyiodides, interstitial complexes with halogens and various small molecules as guests; it also forms hydrogen-bonded compounds with suitable acceptors and salts with components having some basic functionality. The Heilbron host E,E-1-[p-dimethylaminophenyl]-5-[o-hydroxyphenyl]-penta-1,4-dien-3-one has at its disposal two possible conformations and can act both as hydrogen-bond donor and acceptor; these potentialities are all exploited in the different complexes studied, which have CHCl3, m-dinitrobenzene and p-dimethylaminobenzaldehyde as second component. Over 100 complexes of racemic gossypol have been prepared and most of these can be grouped into eight different types of hydrogen-bonded arrangement, with more types undoubtedly awaiting discovery. The tripod molecule tris(5-acetyl-3-thienyl)methane (TATM) is a so-far rare example of a flexile molecule (i.e. one that can occur in many conformationally isomeric states (conformers)) which forms host–guest inclusion complexes with a large variety of guests (solvents). Some forty odd different types of guest have been reported to form inclusion complexes. Five different types of crystal structure, with nine different guests, have been reported. Analysis of this substantial but nevertheless incomplete data base shows that each group of crystallographically isomorphous structures contains a particular TATM conformer with characteristic torsion angles. Finally over 200 complexes of a wide variety of guests have been prepared with unsubstituted tetraphenylmetalloporphyrins (containing metals such as Zn, Mn, Fe, 2H) as hosts; many structural resemblances suggest that there is a common interaction in all these complexes, possibly based on charge transfer guest–core and guest–metal interactions.

8.1 Introduction 8.2 Tri-o-thymotide and analogs as hosts 8.2.1 Crystallography of tri-o-thymotide and its complexes 8.2.1.1 The trigonal clathrate inclusion complexes 8.2.1.2 The hexagonal tunnel inclusion complexes

423 423 423 425 429

422

8.3

8.4 8.5 8.6

8.7

INCLUSION COMPLEXES FORMED BY VERSATILE HOSTS

8.2.1.3 Tunnel inclusion complexes with organometallic guests 8.2.1.4 Crossed tunnel triclinic inclusion complexes 8.2.1.5 Miscellaneous inclusion complexes 8.2.2 Analogs of tri-o-thymotide Trimesic acid and analogs as hosts 8.3.1 Introduction 8.3.2 Host–guest tunnel inclusion complexes based on noncatenated unary hexagonal networks 8.3.2.1 TMA as host 8.3.2.2 Two coordination complexes as potential hosts. 8.3.3 Host–guest tunnel inclusion complexes based on catenated hexagonal unary networks 8.3.4 Host–guest clathrate interstitial inclusion complexes based on catenated hexagonal unary networks 8.3.5 Generalization of the concept of ‘‘interruption’’ to give binary networks 8.3.5.1 TMA.H2O networks 8.3.5.2 Catenated neutral binary networks 8.3.5.3 Ionic binary networks 8.3.6 Hydrogen-bonded TMA binary complexes The Heilbron complexes Gossypol and its inclusion complexes Tris(5-acetyl-3-thienyl)methane (TATM) as host 8.6.1 Introduction 8.6.2 Chemistry of TATM and its inclusion complexes 8.6.3 Conformations taken up by the TATM molecule in the various crystallographic structure types 8.6.4 Crystallography of the inclusion complexes of TATM 8.6.5 Formation of the inclusion complexes 8.6.6 Dynamics of guest molecules in the complexes 8.6.7 Other examples (5,10,15,20)-Tetraphenylmetalloporphyrins and complexes 8.7.1 Introduction 8.7.2 Crystallography of (5,10,15,20)-tetraphenylmetalloporphyrin coordination complexes 8.7.2.1 Introduction 8.7.2.2 The four-coordinate coordination complexes 8.7.2.3 The five-coordinate coordination complexes 8.7.2.4 The six-coordinate coordination complexes 8.7.3 Crystallography of (5,10,15,20)-tetraphenylmetalloporphyrin inclusion complexes 8.7.3.1 Crystallography of four-coordinate (5,10,15,20)tetraphenylmetalloporphyrin inclusion complexes 8.7.3.2 Crystallography of five-coordinate (5,10,15,20)tetraphenylmetalloporphyrin inclusion complexes 8.7.3.3 Crystallography of six-coordinate (5,10,15,20)tetraphenylmetalloporphyrin inclusion complexes 8.7.4 Comparative crystallography of the (5,10,15,20)tetraphenylmetalloporphyrin coordination and inclusion complexes 8.7.5 Questions of nomenclature and description

431 433 435 436 437 437 437 437 443 446 448 448 449 450 450 452 456 459 469 469 469 470 474 476 477 477 478 478 480 480 481 483 486 493 493 500 502 505 506

T RI - O-T HY MOTIDE AND ANALOGS AS HOSTS

423

8.7.6 Can ‘‘sponge’’ structures be inferred from the chemical nature of the second component? References

507 508

8.1 Introduction In this chapter we shall discuss hosts characterized by their ability to form inclusion complexes of a number of different crystallographic types, including the possibility that the host will form both tunnel and clathrate complexes. Usually this only occurs with different guests but there are some examples where a particular guest forms both types, as in some tri-o-thymotide complexes. Another possibility is that host–host interactions for a particular host will be based on directional forces (generally hydrogen bonding) in some complexes, on nondirectional forces (generally van der Waals bonding) in others and on a combination of the two in yet others; this occurs in the Heilbron complexes. Such versatility is in contrast to the greater degree of similarity found in the structural behavior of the hosts discussed in the two preceding chapters. Again we note that this distinction may well disappear, or at least be reduced, with the passage of time. In this chapter we treat all the complexes of a particular host together, in contrast to our practice elsewhere of giving priority to structural rather than chemical resemblances. Clearly many structures discussed in this chapter could have been placed under other headings. 8.2 Tri-o-thymotide and analogs as hosts 8.2.1 Crystallography of tri-o-thymotide and its complexes Tri-o-thymotide (TOT; C33H36O3, 8.1, see also Table 8.4; Chemical Abstracts name 6H,12H,18H-tribenzo(b,f,j))[1,5,9]trioxacyclododecin-6,12,18-trione,1,7,13-tri-methyl-4,10, 16-tris(1-methylethyl)-; CSD name ‘‘tri-o-thymotide’’), first synthesized by Naquet in 1865, forms inclusion complexes with a large variety of guests (Baker, Gilbert and Ollis, 1952), the first of these (with benzene) having been noted by Spallino and Provenzal in 1909. The most recent, and most comprehensive, review is by Gerdil (1996). The November 2002 version of the CSD gives 32 hits for ‘‘tri-o-thymotide.’’ CH3

O C

(H3C)2HC O

O

CH(CH3)2

O

C H 3C

O C O CH(CH3)2

CH3

424


Fig. 8.1. Stereodiagram of the TOT molecule in the (M)-(–) propeller conformation with all three carbonyl groups pointing out above the mean molecular plane. (This diagram was made available by Professor D. J. Williams (Imperial College, London); see Williams and Lawton (1975) for acknowledgement.)

The TOT molecule has been shown by NMR to exist in solution as a mixture of a major, propeller, conformation (approximate molecular symmetry C3-3) and a minor, helical conformation (C1-1). However, in its neat crystals and in its complexes it always has the shape of a somewhat flattened three-bladed propeller, with all three carbonyl oxygens on the same side of the mean ring plane (Fig. 8.1). There is some flexibility in the details of its conformation (for example, the angles between the normals to the three phenyl rings are 27, 43 and 43 in the pyridine complex and 30, 37 and 48 in the neat compound) and it has been suggested that such flexibility plays a role in the ability of TOT to form a range of inclusion complexes. The crystals of neat TOT are racemic, having the achiral but non-centrosymmetric space group Pna21 (Brunie and Tsoucaris, 1974; Williams and Lawton, 1975). The structures of all the complexes noted below are different from those of neat TOT, and hence constitute different phases in the binary phase diagrams. The relation between the sense of optical rotation in solution and the absolute optical configuration of TOT was first determined from the chiro-optical properties of the (þ)-isomer, which was assigned the M-configuration (left handed propeller form) by Downing et al., (1968). Two independent crystallographic studies have given the opposite assignment. Thus the structure of ()TOT[0.5((R)-2-butanol)] at 123K gave the M-configuration for TOT on the basis of the known absolute configuration of (R)-2-butanol (Gerdil and Allemand, 1979; Allemand and Gerdil, 1981; Gerdil, 1987). The experiment was not without its complications as the guest molecule took up two orientations in the cavity, related by a crystallographic two fold axis, and it was also possible to grow single crystals of (P)(þ)-{TOT.[0.5((R)-2butanol)]}. Confirmatory results were obtained for TOT clathrates grown from optically enriched (S)-(þ)-2-bromobutane and optically pure (R,R)-(þ)-trans-2,3-dimethylthiirane (Arad-Yellin, Green, Knossow and Tsoucaris, 1980). Crystallographic studies of more than 70 TOT inclusion complexes have shown that these can be divided into at least four groups on the basis of the ways in which the guests are included (Tables 8.1 to 8.5). A distinction was first made between clathrate (Table 8.1) and tunnel (Table 8.2) complexes (Lawton and Powell, 1958; Powell, 1964; Gerdil, 1987); since then a number of other crystallographic types have been found but not enough structures have yet been analyzed to permit finality in their classification. There is a


425

group of linear tunnel complexes with metallorganic guests (Table 8.3) and two group of crossed tunnel complexes where benzene, cis- and trans-stilbene and methyl-cis- and methyl-trans-cinnamate have been found as guests (Table 8.4). In addition there are isostructural groups of miscellaneous complexes of as yet unknown structure (examples in Table 8.5). It seems that any alterations of arrangement needed to obtain maximum packing efficiency for different included guests are obtained by a combination of minor conformational changes and adjustments of packing, which are slight within each group but major between groups. The packing efficiency appears to be greater in the trigonal clathrates than in the hexagonal tunnel complexes and the cell dimensions of the former vary more with guest size than do those of the latter group. Powell (1964, 1984; see below) has correlated the variations in the cell dimensions with the sizes of the guests. The volume ˚ 3 in neat, racemic TOT and about 7% less than this in the per TOT molecule is 849 A enantiomeric pyridine (trigonal) clathrate. Is this an example of closer packing of enantiomers than of pairs of racemic molecules, which would be contrary to Wallach’s ‘‘Rule’’ discussed in Chapter 10? 8.2.1.1 The trigonal clathrate inclusion complexes The first group to be discussed is that of the trigonal clathrate complexes (Table 8.1), where a necessary, but not sufficient, condition for formation is that the largest molecular ˚ . The ideal composition of the clathrates is dimension should not exceed 9.5 A {2(TOT)[guest]}, and this is usually found, although C2H4I2 as guest is an exception. The guests are often disordered at room temperature. The following are among the guests forming clathrates: 1. Primary alcohols – ethanol to pentanol; methanol forms a clathrate only in the ˚ and it also presence of acetone while the largest dimension of pentanol is 9.5 A forms tunnel complexes. 2. Alkyl halides – CH3Br to n-C4H9Br (n-C4H9I forms a tunnel complex). 3. Dihalogenoalkanes – X(CH2)nX for X ¼ Br, n ¼ 1 to 3; for X ¼ I, n ¼ 1. 4. Miscellaneous – pyridine, diethyl ether, acetone, C2H4I2 (90% of the cages are empty), I2.

˚, A ˚ 3) for neat TOT (a) and its trigonal clathrate complexes (b) (at room Table 8.1. Crystal data (A temperature, unless stated otherwise). The space groups (S. G.) of the trigonal clathrates are all P3121 (absolute configuration not implied), with Z ¼ 6. Asterisks in the table denote crystals for which full structure analyses have been reported. Angles (90 , 120 ) determined by symmetry are not included Table 8.1(a) Substance (formula unit)

Reference/Refcode

a

b

c

V/Z

TOT* Z ¼ 4 Pna21

BT74 TOTHYM WL75 TOTHYM01

16.05 16.049

13.39 13.424

13.94 13.909

749 752

References: BT74 – Brunie and Tsoucaris, 1974; WL75 – Williams and Lawton, 1975.

426


Table 8.1(b) Substance (formula unit) Trigonal clathrate complexes TOT[0.5(ethanol)]* [0.5(acetone)]* [0.5(pyridine)]* [0.5((R)(–)-2-butanol at 123K* [0.5(C5H11OH)] [0.5(bromocyclohexane)) [0.5(chlorocyclohexane)) at 158K* [0.5(thiophene))* [0.5(2-bromobutane)] at 125K [0.5(ethylmethylsulfoxide)] [0.5(benzene)] [0.5(chloroform)] ( þ )-TOT[0.5(( þ )-2-bromobutane)] at 223K [0.5(3,4-epoxycyclopentanone)] [0.5((R)-4-hydroxycyclopent-2-enone)] [0.5( þ )-trans-2, 3-dimethylthiirane)] (racemate) [0.5(S,S)-( )-2,3-dimethylthiirane [0.5(trans-(R,R)-( þ )-2, 3-dimethyloxirane [0.5(R,R)-( þ )-2,3-dimethylthiirane TSBS [0.5((S)( þ )-butanol]* at 90K

Reference; Refcode

a

c

V/ Z

LP58; WL75; TOTETL FRHP92; PENPIH BNTDG77; TOTPYD10 AG83; OTHYMD WL75 GF85; DITRON GF85; DITRIH PB96; TOYBIS AG82; BERWIE; THYBBU is stereoisomer AG82; BIGDAW NP52; ZZZWAK NP52; ZZZWAG A-Y,GKT80; THYBBU

13.443 13.46 13.67 13.642 13.70 13.794 13.604 13.585 13.620

30.143 30.30 29.90 30.180 30.74 30.876 30.605 29.914 30.075

786 792 806 785 832 848 817 797 805

13.538 13.70 13.55 13.72

30.598 29.90 30.30 30.24

809 810 803 822

GLB99; CIDDOI GLB99; CIDFEA CABRAY; A-YGKT80

13.660 13.765 13.611

30.304 30.057 30.340

816 822 811

CABREC; A-YGKT80 CABPUG; A-YGKT80

13.603 13.484

30.440 30.440

813 799

THYMTI; A-YGKT80 GG92; PABJAD

13.600 13.556

30.280 31.705

808 841

Notes: (1) – The compositions given are nominal compositions not always realized in practice. In PB96 it is stated correctly that the unit cell contains six TOT molecules, and that the host : guest ratio is 2 : 1, but the number of thiophenes in the cell should be 3 and not 1, as stated in the Abstract. It is the cavity which contains a single thiophene. (2) – Cell dimensions have been given (Arad-Yellin, Green, Knossow and Tsoucaris, 1983) for trigonal clathrate complexes (space group P3121) with the following guests: 2-chlorobutane, 2-bromobutane* (at 225K; there is an independent structure analysis at 125K (Allemand and Gerdil, 1983)), 2-iodobutane, trans-(R,R)-2,3-dimethyloxirane* (CABPUG, structure at 223K), trans-2,3-dimethylthiirane*, trans-2,3-dimethyloxetane, trans-2,3dimethylthietane, propylene oxide, 2-methyltetrahydrofuran, methyl methanesulfinate, 2,3,3-trimethyloxaziridine. References: AG81 – Allemand and Gerdil, 1982; AG81 – Allemand and Gerdil, 1983; A-Y, GKT80 – Arad-Yellin, Green, Knossow and Tsoucaris, 1980; BNTDG77 – Brunie, Navaza, Tsoucaris, Declercq and Germain, 1977; FRHP92 – Facey, Ratcliffe, Hynes and Ripmeester, 1992; GF85 – Gerdil and Frew, 1985; GG92 – Gnaim, Green, AradYellin, Vyas, Frolow and Keehn, 1992; TSBS is tri-3-(2-butyl)-6-methylsalicylide; GLB99 – Gerdil, Liu and Bernardinelli, 1999; LP58 – Lawton and Powell, 1958; NP52 – Newman and Powell, 1952; PB96 – Pang and Brisse, 1996; WL75 – Williams and Lawton, 1975.

In the trigonal clathrate complexes, the cell dimensions show a marked dependence on the nature of the guest; this excellent illustration of ‘‘adaptability’’ is shown in Fig. 8.2 where the a and c dimensions are plotted for an eclectic group of guests. There is a only rough linear relation between a and c, with many exceptions. However, the linear


427

relationship is obeyed well for the homologous series of {TOT[0.5(n-alkanol)]} complexes (data from Lawton and Powell, 1958). Thus, (to quote Powell (1964, p. 475)), ‘‘the steady increase in the clathrate [unit cell size] is [due to] . . . an adaptation of the cage size to the increasing size of its imprisoned molecule.’’ The roughness of the overall linear relationship shows that shape of guest, in addition to size, must be taken into account. The careful measurements of cell dimensions and crystal densities by Lawton and Powell (1958) (some of the densities were reported with standard uncertainties as small as 0.0003 g cm3!) demonstrated that the unit cells of these complexes always contain 6 TOT molecules and 3 guest molecules, thus implying some empty space in the cavities when smaller molecules are guests. Other guests show similar behaviour but the data are not as complete. The apparent appearance of empty space in the tunnels reminds one of the behavior of the deoxycholic acid tunnel inclusion complexes of fatty acids (Section 6.3.1). ˚ and the guest molecule is The cage has approximate dimensions 9.0 5.8 7.9 A surrounded by eight TOT molecules (Fig. 8.3). The TOT clathrates with C3H7Br (one of two polymorphs) and C4H9Br show doubled values of the a dimension, implying a lateral ordering of the guests. A structure analysis of the 2 : 1 clathrate of TOT with 2-chlorocyclohexane showed that the guest was disordered over the axial 2-chloro chair and axial 2-chloro boat conformations in 2 : 1 ratio; both the axial and the boat 31.2

31

30.8

c (Å)

30.6

30.4

30.2

30

29.8 13.4

13.5

13.6 a (Å)

13.7

13.8

Fig. 8.2. 54 pairs of a and c cell dimensions are plotted for trigonal TOT clathrates, the values being taken from Table 8.1, literature sources and Table 1 of Gerdil (1996). The filled circles are for the homologous series of TOT[0.5(n-alkanol)] complexes, with the points in order from left to right for the guests C2H5OH, CH3OH, n-C3H7OH, n-C4H9OH, n-C5H11OH. The methanol values are anomalous because this complex also contains acetone. The open circles are for a catholic variety of guests.

428


Fig. 8.3. Stereodiagram of the eight TOT molecules surrounding the CH2I2 guest (only iodines shown) in TOT[0.5(CH2I2)]. (For acknowledgements see Williams and Lawton, 1975.)

conformations are thermodynamically disfavoured (Allemand and Gerdil, 1982). These conclusions about the guest conformations were confirmed by IR spectroscopy. The crystal structure (at 298K) of the {TOT[0.5(acetone)]} clathrate and the guest dynamics (over the temperature range 77–298K) have been studied by a combination of XRD, solid state NMR and molecular modelling (Facey, Ratcliffe, Hynes and Ripmeester, 1992). The crystals are isomorphous with the other examples of the trigonal clathrate type, the acetone carbonyl bond lying along the two fold axis of the space group. Although the displacement factors of the atoms of the acetone and of the carbonyl groups of the host are larger than those of the other atoms, there was no large scale disorder. The solid state 2 H NMR study was carried out using acetone-d6, and the 13C study using acetone with the carbonyl C enriched to 12 mol%. Their main conclusions, with particular relation to the dynamics of the acetone molecule, were summarized by Facey et al., as follows. A number of dynamic processes occurred. The methyl groups of the enclathrated acetone undergo rotation at rates 107 Hz at 77K. Above this temperature a new slow motion sets in, which is primarily a two fold flip of the acetone molecule about the carbonyl bond with an activation energy of 13.6 0.8 kJ/mol. There is a secondary site, energetically less favourable by 4.4 0.5 kJ/mol, at an angle of 63 10 from the favored site. The general shape of the potential was confirmed by molecular mechanics calculations on a group of 410 atoms comprising the eight TOT molecules surrounding an enclathrated acetone, but quantitative agreement was difficult to obtain because of the need to take into account the dynamics of the guest molecule and of the flexible host molecules. The success of the study was ascribed to the use of a combination of complementary techniques, whereas each on its own was not able to provide more than a part of the overall picture. Because of the spontaneous resolution of TOT on crystallization with guests, considerable attention has been given to the possibilities of using TOT as a resolving agent for enantiomeric mixtures of suitable guests (Arad-Yellin, Green, Knossow, Ryanek and Tsoucaris, 1985). A measure of the enantioselectivity is given by the enantiomeric excess (e.e.) of the guest in a single TOT crystal of given handedness. The clathrates give e.e.


429

values ranging from 2 to 83% while the tunnel inclusion complexes give small but still significant values around 5%. Amplification of the optical purity of the guest can be obtained by successive recrystallizations. Chiral recognition using TOT complexation has been discussed in detail by Gerdil (1987, 1996). 8.2.1.2

The hexagonal tunnel inclusion complexes

Hexagonal linear tunnel complexes (Table 8.2) are formed when the longest molecular ˚ but with the other dimensions such that the guest can fit into dimension exceeds 9.5 A ˚ . The ratio TOT : guest is generally a tunnel with cross-sectional diameter of about 4.3 A not that of two small integers because the guest molecules are longer than the periodicity of the host matrix in the tunnel axis direction. The following are among the guests forming tunnel inclusion complexes: 1. 2. 3. 4. 5. 6.

RX with R ¼ C5H11 and above, X ¼ OH, Br. RI with R ¼ C4H9 and above. I(CH2)6I and higher ,!-dihalogenoalkanes. ROR’, with R ¼ CH3 or C2H5, R’ ¼ C4H9. Hg(C2H5)2. The tetraketone CH3(CH2)7COCH2CO(CH2)7COCH2CO(CH2)7CH3.

˚ , deg., A ˚ 3) for the hexagonal linear tunnel complexes of TOT (at Table 8.2. Selected crystal data (A room temperature). Asterisks in the table denote crystals for which full structure analyses have been reported. The absolute structure was determined only for the last entry Hexagonal linear tunnel inclusion complexes: (long-chain guests) Composition/Reference/Refcode a c TOT[0.5((R(–))-2-butanol)] LP58 0.2(TOT)[(n-C16H33OH)] (cetyl alcohol)* WL75; TOTCET [0.5 C4H9I] LP58 [0.5Br(CH2)10Br]* S-GH 99 [0.5I(CH2)8I]* S-GH99 [0.5I(CH2)10I]* S-GH99 [0.5(n-hexane)] NP52; ZZZWAG [0.5Br(CH2)8Br]* S-GH00 (123K) (see Note (3))

V/Z

Z

Space group

14.31 14.31

28.99 29.02

858 858

6 6

P62 P61

14.25 14.294 14.319 14.285 14.20 37.162

29.03 29.039 29.111 29.123 28.90 29.207

851 856 862 858 841 832

6 6 6 6 6 42

P31 P61 P61 P62 P62 P65

Notes: (1) The compositions given are nominal compositions not always realized in practice, e.g. the cetyl alcohol complex was 6 : 1.3. (2) Cell dimensions have been given (Arad-Yellin, Green, Knossow and Tsoucaris, 1983) for tunnel complexes (space group P61) with the following guests: 2-chloro-octane, 2-bromo-octane, 3-bromo-octane, 2-bromononane, 2-bromodecane. (3) a ¼ 3as þ bs; b ¼ as þ 2bs, c ¼ cs, where the ‘‘s’’ subscript refers to the cell used for the other complexes. p a ¼ 7as. The space group is for the crystal used in the structure analysis. No Refcode. References: (1) LP58 – Lawton and Powell, 1958; NP52 – Newman and Powell, 1952; S-GH99 – Serrano-Gonzalez and Harris, 1999; S-GH00 – Serrano-Gonzalez and Harris, 2000; WL75 – Williams and Lawton, 1975.


430

7. 8.

n-C12H25COOCH3. CH3C(H)C6H13X, with X¼OH, Br.

In the hexagonal linear tunnel inclusion complexes the dimensions of the unit cells are ˚ . Three space groups have been encountered – all similar, with a 14.2 and c 29.2 A P31, P61 and P62 (and enantiomorphs); among a group of 28 crystals of the tunnel type with unbranched guests, 19 had P61, 5 had P31 and 4 had P62 (Lawton and Powell, 1958). The changes in arrangement due to adoption of one of the other space groups are expected to be minor. The changes in cell dimensions with the nature of the guest are appreciably smaller than those found for the trigonal TOT clathrates but periodic variations do occur which depend on the nature of the guest. These are plotted against the molecular length in Fig. 8.4 and have been discussed in detail by Powell (1964); briefly stated, the periodic variations are ascribed to the better fit that occurs between host matrix and guest when the guest approaches certain critical lengths. The structures consist of a spiral arrangement of the roughly disc-shaped TOT molecules around the [001] axis (Fig. 8.5), leaving a tunnel ˚ ; the unhindered nature of the tunnel (Fig. 8.6) is draof approximate diameter 4.3 A matically shown by the formation of tunnel inclusion complexes of the tetraketone CH3(CH2)7COCH2CO(CH2)7COCH2CO(CH2)7CH3. Some crystals give one dimensional diffuse X-ray scattering from the guests when the periodicities of host molecule matrix and included guest molecules along the axis of the 14.35 C7Br C3OH

C7OH

C6Br

14.30

C12OH C16OH

C5I C4I

14.25

C18I C18OH

n-c12H25Co2Me

C5Br

2C/6

14.20

C18Br

C10OH

C8I

C6OH

C16I

C8Br

C8OH

a(Å)

C16Br

C7I

4C/6

3C/6

5C/6

5150 Vol.(Å)3 5100

Molecular length(Å)

29.15

C5OH

10 C5Br

15

C5I

20 C10OH

C6Br

c(Å)

25

n-c12H25Co2Me

29.10

C18Br C16I

29.05

C7I

C7OH C4I

C7Br C8OH

29.00

C18OH

C16OH C8Br

C18I

C16Br

C8I C12OH

C5OH

Fig. 8.4. Variation of cell dimensions with the (nominal maximum) length of the included molecule in the tunnel inclusion complexes of TOT. Alcohols are shown by circles, bromides as squares and iodides as triangles. (Reproduced from Powell, 1964.)


431

Z X Y

Fig. 8.5. Stereodiagram of the packing of the TOT host molecules in the {TOT[0.5(cetyl alcohol)]} tunnel inclusion complex. The axis of the tunnel is along c. The disordered guest molecule is not shown. (For acknowledgements see Williams and Lawton, 1975.)

tunnels are incommensurable (Lawton and Powell, 1958). Detailed studies of these diffraction effects (at 293K) have been reported by Serrano-Gonzalez and Harris (1999), who applied methods similar to those used in the study of {urea[n-alkane]} complexes and discussed earlier (Section 6.2.1). There is evidence for ordering at lower temperatures, analogous to the phenomena found in the urea complexes. The four structures determined (Table 8.2) include that of {TOT[0.5(cetyl alcohol)]} (Fig. 8.5), where the guest molecule is disordered in the tunnel at room temperature; the more precise structures are those with ,!-dihalogenoalkanes as guests. {TOT[1,8-dibromo-octane]} differs from the previous examples in that there is a commensurate relationship between host and guest sub-systems; the 293K structure does not change essentially on cooling to 123K, apart from an increase in the degree of order of the guest. There is a superstructure involving a seven-fold increase in unit cell volume (Serrano-Gonzalez and Harris, 2000). The hexagonal linear TOT tunnel inclusion complexes can be classified as helical tubulands (Chapter 6). The methyl and isopropyl groups dominate the interior of the tunnel, providing a hydrophobic environment for the guest molecules. Less attention has been paid to the structures of the tunnel inclusion complexes than to those of the trigonal clathrates, perhaps because of the interest in the potentialities of the latter as resolving agents. 8.2.1.3

Tunnel inclusion complexes with organometallic guests

TOT tunnel inclusion complexes with organometallic compounds as guests were prepared in the course of a search for complexes showing second-harmonic generation (see Chapter 6 for analogous work on thiourea inclusion complexes); desired materials must crystallize in a noncentrosymmetric space group. Of the five materials studied crystallographically

432


(Table 8.3), only one fills this criterion. There is a zigzag arrangement of host molecules in the {TOT[0.5(W(CO)5-pyridine)(MeOH)]} complex which leaves tunnels along [100] containing the organometallic guests (Fig. 8.6) aligned as centrosymmetric head-to-tail pairs. ˚ , deg., A ˚ 3) for the linear tunnel inclusion complexes with Table 8.3. Selected crystal data (A organometallic guests (at room temperature). Only angles not determined by symmetry have been inserted in the Table. Asterisks in the Table denote crystals for which full structure analyses have been reported. Data are taken from Tam, Eaton, Calabrese, Williams, Wang and Anderson, 1989. Compound/Refcode

a

b/

c

V/ Z

I

Space group

TOT{0.5[W(CO)5pyridine]}{0.5MeOH}* (see Note 1); SESMIN TOT [W(CO)5-4-picoline]* (see Note 2); SESMUY TOT[6-tetralin-Cr(CO)3]; SESMEI

22.91

13.71/ 95.22 23.91/ 102.62 20.36

23.36

913.8

8

I2/a

14.26

1069

4

P21/c

13.76

1018

4

Pca21

12.85 14.53

Notes: (1) The 2 : 1 TOT W(CO)5-pyridine and W(CO)5-4-aminopyridine (SESMOS) complexes are isomorphous. (2) The 1 : 1 TOT W(CO)5-4-picoline and W(CO)5-5-ethylpyridine (SESNAF) complexes are isomorphous.

b c

Fig. 8.6. The crystal structure of {TOT[0.5(W(CO)5-pyridine)(MeOH)]} viewed down [100]. The tunnels containing the organometallic guests run normal to the plane of the page. The disordered MeOH has been omitted for clarity. (Reproduced from Tam, Eaton, Calabrese, Williams, Wang and Anderson, 1989.)


433

8.2.1.4 Crossed tunnel triclinic inclusion complexes The crossed tunnel inclusion complexes crystallize in two groups of isostructural structures (use of reduced cells is needed to demonstrate this unequivocally) and contain a somewhat surprising variety of guests – cis-stilbene, benzene, methyl cis- and transcinnamate in Group A, and trans-stilbene and trans-p-(dimethylamino)cinnamaldehyde) in Group B (Table 8.4). The crystal structures of the benzene and trans-stilbene tunnel inclusion complexes, which belong in Groups A and B respectively, are shown in Figs. 8.7 and 8.8. The cell edges of these two complexes are very similar and they have the same space group and

˚ , deg., A ˚ 3) for two different groups (A and B) Table 8.4. Crystal data (reduced cells) are given; A of isomorphous triclinic crossed tunnel inclusion complexes (at room temperature, unless stated otherwise). Asterisks in the table and notes denote crystals for which full structure analyses have been reported

Group A TOT[0.5(cis-stilbene)]* BNTDG77; A-YBGKT79; ZZZAVX10 TOT [1.25(C6H6)]* AG83; BOLGOY TOT [0.5(methyl-cis-cinnamate)] A-YBGKT79; TOTCCI TOT [0.5(methyl-trans-cinnamate)] A-YBGKT79; TOTTCI TOT[0.75(1,3-disilylbenzene)]; BGMPR00; OBICAD Group B TOT [0.5(trans-stilbene)] BNTDG77; A-YBGKT79; ZZZAXS10* TOT[0.5(trans-p-(dimethylamino)cinnamaldehyde) Tam et al., 1989; SESNEY TOT[0.75(phenylsilane)]; BGMPR00; OBICEH Group C TOT[0.5(isopropyl-Ndintrophenylvalinate)] BNTDG77; ZZZAXP

a/

b/

c/

V/Z

Z

Space group

11.32 93.89 11.312 94.45 11.3 94.0 11.51 96.6 11.290 94.04

13.17 102.68 13.147 102.91 13.0 102.0 13.04 101.8 13.055 102.94

24.76 93.61 24.983 93.72 25.0 92.0 24.19 91.3 24.707 93.24

895

4

P 1

900

4

P 1

895

4

P 1

882

4

P 1

885

4

P 1

11.639 83.95 11.54 81.63

13.027 76.79 13.20 79.09

24.409 84.81 23.47 85.34

893

4

P 1

867

4

P 1

11.082 84.01

13.154 77.66

24.616 87.09

871

4

P 1

11.360 89.76

13.790 89.98

23.360 89.18

915

4

P 1

Notes: The compositions given are nominal compositions not always realized in practice, e.g. the triclinic benzene complex actually had the composition {TOT[0.8(benzene)]} and the cis-stilbene complex was 1 : 0.4. The results for ZZZAXP would appear to require checking. References: AG83 – Allemand and Gerdil, 1983; A-YBGKT79 – Arad-Yellin, Brunie, Green, Knossow and Tsoucaris, 1979; BGMPR00 – Borisenko et al., 2000; BNTDG77 – Brunie, Navaza, Tsoucaris, Declercq and Germain, 1977.

434


Fig. 8.7. Stereoview of the molecular arrangement in the TOT[1.25(benzene)] tunnel inclusion complex. The origin is at the rear upper left hand corner of the cell, the [100] axis comes out towards the viewer, [010] is down and [001] is towards the right. One set of benzenes is centered at the origin, a pair of benzenes is located about the centre of the cell at 1/2, 1/2, 1/2 and another pair about 1/2, 1/2, 0 (all with equivalent positions). (Reproduced from Allemand and Gerdil, 1983.)

Fig. 8.8. Stereoview of the molecular arrangement in the TOT[0.5trans-stilbene] tunnel inclusion complex. The origin is at the rear lower left hand corner of the cell, the [100] axis comes out towards the viewer, [010] is towards the right and [001] is upward. One stilbene is centred about 1/2, 1/2, 1/2 and another about 0, 1/2, 0 (all with equivalent positions). The first of these is equivalent to the second pair of benzenes as given in the caption to Fig. 8.7; the second does not have a benzene counterpart. (Reproduced from Arad-Yellin, Brunie, Green, Knossow and Tsoucaris, 1979.)

similar unit cell volumes, but the cell angles are different – hence they are not isomorphous. The structures are not identical because of different distributions of the guest molecules, as noted in the captions to these two figures. In both there are two sets of tunnels that are approximately mutually orthogonal but there is a third set of benzene molecules located in


435

a third tunnel. The hexagonal linear tunnel complexes and the linear tunnel complexes containing organometallic guests, on the one hand, and the triclinic tunnel inclusion complexes, on the other, are structurally quite distinct, the first two types having one set of parallel tunnels and the third two (or three) sets of mutually perpendicular tunnels. 8.2.1.5 Miscellaneous inclusion complexes There are a number of other groups of isomorphous or isostructural inclusion complexes whose structures have not yet been reported (Table 8.5). Thus TOT exhibits considerable versatility as a host. We have noted above that neat TOT forms racemic crystals but their space groups show that spontaneous resolution takes place on formation of both the trigonal clathrate and hexagonal linear tunnel inclusion complexes; thus, in the absence of complicating factors such as twinning, domain formation or disorder, a particular crystal of one of these types of complex will contain only one of the enantiomers of TOT (Powell, 1952). This was strikingly demonstrated when Newman and Powell (1952) grew large single crystals of the clathrates of TOT with ˚ , deg., A ˚ 3) for three different groups of TOT Table 8.5. Crystal data (reduced cells are given; A inclusion complexes (at room temperature). These results have been taken from Gerdil (1996, Table 3); a few miscellaneous complexes given by Gerdil have been omitted. No structure analyses have been reported. Isomorphous trigonal P31 complexes (host : guest 1 : 1) a/ b/

c/

V

Z

TOT[norbornene] TOT[norbornadiene] TOT[2,3-dihydrofuran] TOT[3,4-epoxyoxolane] TOT[2,3-epoxycyclopentan-1-one] TOT[2,3-epoxycyclohexan-1-one] (see Note 1)

10.22 10.25 10.199 10.12 10.150 10.263/ 119.9

864 853 824 822 832 865

3 3 3 3 3 3

Pbca complexes (host : guest 1 : 1) 13.184 23.086 23.977 13.19 23.59 23.83 13.319 23.427 23.760 13.332 23.751 24.250

912 927 927 960

8 8 8 8

B group of isostructural orthorhombic Pbca complexes (host : guest 1 : 1) TOT[2-cyclohepten-1-one] 13.827 37.191 14.281 TOT[2,3-epoxycycloheptan-1-one] 13.857 36.334 14.668

918 923

8 8

A group of isostructural orthorhombic TOT.[4-chloro-2-cyclopenten-1-one] TOT[3-chloro-1-cyclohexene] TOT[3-chloro-2-methylbut-1-ene] TOT[trans-2-chlorocyclohexan-1-ol]

17.11 16.98 16.730 16.88 16.747 17.071/ 90.30

17.099/ 89.87

Notes: (1) This compound was described as pseudo-trigonal but actually triclinic. Further study of the crystal symmetry seems necessary. (2) Space groups (but not other data) have been given for TOT complexes with the following guests: dl-2,3dibromobutane (Pbcn); meso -2,3-dibromobutane (P21/c); 3-bromooctane (C2/c), fluothane (Pbca), meso2,3-butanediol carbonate (P21) (Arad-Yellin, Green, Knossow and Tsoucaris, 1983). (3) The 2-chlorotetrahydropyran complex crystallizes in a pseudo-hexagonal R system with a ¼ b ¼ 17.142(3), ˚ (Gerdil and Frew, 1985; DITRUT) the Br(CH2)3Br complex crystallizes in an unspecified c ¼ 10.270(2) A orthorhombic unit cell (Lawton and Powell, 1958). The structures have not been reported.

436


benzene,1 CHCl3 and n-hexane and showed by polarimetry that solutions made from individual crystals were optically active, the specific rotation diminishing with time because of racemization. The optical activity results from hindered rotation about the single bonds in the twelve membered ring of the TOT molecule; the activation enthalpy for racemization is about 88 kJ/mol. The trigonal 1 : 0.5 (Table 8.1) and 1 : 1 (Table 8.5) complexes of TOT and the hexagonal complexes of Table 8.2 are chiral but the others are all racemates. The reason for this difference in behavior is not understood but the distinction resides in the nature of the guest and thus it must stem from subtle details of the host–guest recognition process (cf. Section 8.6 below). 8.2.2

Analogs of tri-o-thymotide

Many potential chemical modifications of the TOT molecule are possible and their syntheses and structural properties have been reviewed in detail (Ollis and Stoddard, 1984). Some of these are sketched in Table 8.6. No trisalicylide, apart from TOT, has yet been found to crystallize as an inclusion complex, not even tri-o-carvacrotide where only methyl and i-propyl groups are interchanged with respect to TOT. Tri-o-thiothymotide, the direct sulphur analog to TOT, could not be synthesized. Among this whole series (some sixty compounds were discussed by Ollis and Stoddard, 1984) the only complexes so far reported are the 1 : 1 inclusion complexes of N,N 0 -dimethyl-N00 -benzyltri-3methyltrianthranalide and N-methyl-N 0 -benzyltri-3-methyl-trianthranalide with toluene; crystal structure analysis (Edge et al., 1981; MANTRN) of the first of these showed it to be an enantiomorphic tunnel inclusion complex (space group P212121; Z ¼ 4) with the host in a helical conformation. However, more recently it has been reported that tri-3(2-butyl)-6-methylsalicylide (TSBS) has complex-forming properties similar to those of TOT and that analogous types of trigonal and hexagonal complexes are formed (Gnaim, Green, Arad-Yellin, Vyas, Frolow and Keehn, 1992). Trigonal clathrates are formed with nitromethane, ethyl acetate, 2-butanol (in Table 8.1), trifluoroacetic acid, 2-chlorobutane and probably with CHCl3, CCl4, 1-bromobutane and CH2I2. A tunnel inclusion complex is formed with 2-octanone. 3

R X

6 R

6

Y

Y X

3

Y X R

3

6 8.2 1 This must have been the enantiomorphic trigonal clathrate reported by Gerdil (1996, Table 3) and not the racemic triclinic crossed tunnel complex noted here in Table 8.4.

TRIMESIC ACID AND ANALOGS AS HOSTS

437

Table 8.6. Variations on the substituted trisalicylide structure, some of which have been tested for formation of inclusion complexes. The asterisked compound was not synthesized X

R

Y

3

6

Name of compound

Type name

C C C

¼O ¼O ¼O

O O O

Me CHMe2 CH3

CHMe2 Me CH(CH3)CH2CH3

Trisalicylide Trisalicylide Trisalicylide

C N

¼O H

S O

Me H

CHMe2 H

Tri-o-thymotide (TOT) Tri-o-carvacrotide Tri-3-(2-butyl)-6methylsalicylide (TSBS) Tri-o-thiothymotide* Trianthranilide

Trithiosalicylide Trianthranilide

8.3 Trimesic acid and analogs as hosts 8.3.1 Introduction Neutral trimesic acid (benzene-1,3,5-tricarboxylic acid; TMA) is a rather versatile host for formation of inclusion complexes because of the ability of the carboxylic acid groups to form hydrogen bonds both with one another and also to other groups such as water. Thus unary networks, containing only TMA can be formed, and also binary networks containing TMA linked in ordered fashion to other molecules. There are some examples where unary and binary networks are combined in a single compound. Networks analogous to those of TMA but based on linked coordination complexes are also possible hosts for formation of inclusion complexes. The possibilities are further extended when anionic TMA species are incorporated together with inorganic cations and, especially, organic cations. We shall describe these complexes in a logical order rather than in the chronological sequence of their discovery; the structural chemistry has been reviewed (Davies, Finochiarro and Herbstein, 1984; Herbstein, 1987, 1996), and since then considerably expanded. The November 2002 version of the CSD gives 37 hits for ‘‘trimesic acid’’ (including anions). 8.3.2 Host–guest tunnel inclusion complexes based on noncatenated hexagonal networks 8.3.2.1

TMA as host

We start with a simple analogy – isophthalic acid (1,3-benzenedicarboxylic acid) could possibly form hexagonal rings of hydrogen bonded molecules but, in fact, forms a ribbon motif of hydrogen bonded dimers (Alcala and Martinez-Carrera, 1972; BENZCD). However, when isophthalic acid is substituted in the 5-position by a bulky group such as decanol to give 5-decyloxyisophthalic acid, then hexagonal tunnels are formed with ˚ (Yang, Marendez, Gelb and Hamilton, 1994; Fig. 8.9; PIWHAE). diameters of 14 A The isophthalic acid core is planar while the decyloxy chains take up alternating up and down positions. When the bulky substituent in position 5 is replaced by a third carboxylic acid group to give trimesic acid, and this is crystallized from acetone or other suitable solvent (not water) in the presence of appropriate guests, isostructural crystalline complexes are obtained in which the TMA molecules form hexagonal sheets (‘‘chicken wire’’)


438

R

O O

H

R O

H

O

R

R = OC10H21

14 Å

R

R

R

Fig. 8.9. Schematic view of a ring of six hydrogen bonded molecules in the crystal structure of 5-decyloxyisophthalic acid; only one of the dimeric carboxyl groups is shown in detail.

Fig. 8.10. The basic ‘‘chicken wire’’ motif in the uninterrupted TMA network is a two-dimensional arrangement of six-molecule rings, the hydrogen bonds between carboxyl groups being represented by dashed lines. (Reproduced from Herbstein, Kapon and Reisner, 1987.)


439

˚ (Fig. 8.10). These are the simplest inclusion which contain tunnels of net diameter 14 A complexes of trimesic acid from a structural point of view. They are called ‘‘hexagonal TMA tunnel inclusion complexes’’ (Herbstein, Kapon and Reisner, 1987) and can be classified in terms of the number of sheets in a crystallographic repeat in the direction of the tunnel axis, and in terms of the crystallography of the unit cell; ‘‘hexagonal’’ is used as a descriptor and exact symmetry is not implied. So far twolayer, three-layer and five-layer repeats have been found. The guest molecules are contained in the tunnels and are usually disordered at room temperature. Representative crystal data are summarized in Table 8.7. As noted in Table 8.7, there are actually two groups of complexes, Group (a) with uninterrupted TMA networks (as shown in Figs. 8.10 and 8.11) and Group (b) in which

˚ , deg., A ˚ 3) for the ‘‘hexagonal’’ TMA tunnel inclusion Table 8.7. Representative crystal data (A complexes; only complexes whose structures have been determined are included except for the fivelayer octanol complex. Values of angles fixed by cell symmetry are not given Composition

a/

b/

(a) Complexes with uninterrupted TMA TMA[isooctane]a 28.60 16.60 FOPGIA 102.6 16.50 16.50 2TMA [tetradecane]b FOPGEW 2TMA[1.33(octanol)] 18.00 18.01 106.1 100.2

c/

Volume of formula unit

networks 6.93 802

Z

Space group

Unit-cell contents

Number of layers

8

C2/c

8TMA þ 8(C8H18) 6TMA þ 3(C14H30) 12TMA þ 8(C8H17OH)

2 (X3.47)

6TMA þ 2(C18H35OH)] 8TMA 8(C10H16O) þ 16(H2O) 16TMA þ 8(pyrene) þ 16 EtOH

3 (X3.49)

10.07

791.5

3

P31

17.02 116.4

741

6

P1

789

3

P1

976

4

P21212

(b) Complexes with interrupted 16.44 2TMA 79.1 [0.67(oleylalchol)c 2TMA[2(camphor)] 32.44 4(H2O)

TMA networks 15.97 10.46 83.9 61.8 16.90 7.12

2TMA[pyrene] 2(C2H5OH) at 198K; SURYUZ; KT99 2TMA[1.5(pyrene)] 2(CH3OH); UCUKUY; HKS01

28.13

16.55 95.13

14.73

854

8

C2/c

9.533 93.88

13.540 95.02

14.644 90.14

940

2

P1

2 (X3.36) 5 (X3.40)

2 (X3.56)

4TMA þ 3(pyrene) þ 4(methanol)

Notes: a There is an isomorphous complex of composition 2TMA[0.5(squalene)]. b There are isomorphous complexes of composition 2TMA[2(isooctane)]; 2TMA[2(1-octene)]; 2TMA [epichlorohydrin]. c Cell is reduced apart from choice of origin. d The absolute configurations of the crystals with chiral space groups were not determined. References: HKS01 – Herbstein, Kapon and Shteiman, 2001; KT99 – Kolotuchin et al., 1999.

440


guest or water molecules are included in the TMA network, which we call ‘interrupted.’ We first discuss Group (a). The four isomorphous three-layer complexes show nearly identical cell dimensions, indicating minimal influence of the guests on the shape of the TMA network. These crystals are chiral but the absolute configuration of the {2TMA[n-tetradecane]} complex (Fig. 8.11) was not determined. The TMA network appears to be centrosymmetric but there are indications (but no more than this from a room temperature determination in which the conformation of the guest molecule could not be ascertained with certainty) that the tetradecane has a coiled conformation and is thus not centrosymmetric; if so, then this may be an example of spontaneous resolution of conformational enantiomers. In the {2TMA[0.5(squalene)]} complex the periodicity along the tunnel axis is much less than the length of the squalene molecule, which is possibly not linearly extended in the cavity, as it is in its inclusion complex with urea, but folded into some sort of curved conformation, similar to the oleyl alcohol complex discussed below. A detailed structure determination was not carried out. The {TMA[isooctane]} complex, the structure of which was determined, is isomorphous but the guest molecules could not be found. The examples with interrupted TMA networks (Herbstein, Kapon and Reisner, 1988) are particularly interesting, and more widespread than expected. In the camphor complex, which is a two layer structure, water molecules intervene in two of the six carboxyl bridges of each macrocyclic ring giving a tunnel with an elliptical rather than cylindrical cross section (Fig. 8.13); unfortunately the camphor molecules could not be located in this room temperature structure analysis so that the role of the guest in engendering this distortion is not known. The space group is chiral and the ring of TMA and water molecules is not flat; the absolute configuration of the crystal used was not determined. Double interruption of a carboxylic acid dimer by water molecules was encountered in 1934 in oxalic acid dihydrate (Zachariasen, 1934), much studied since then; however,

b c a

Fig. 8.11. Stereoview of the {2TMA[n-tetradecane]} structure down [001], along which there are three layers. The TMA layers are appreciably nonplanar and this eliminates the congeneric centrosymmetric space group P31/m. The n-tetradecane molecules are appreciably disordered in the tunnels and their representation is only schematic. (Reproduced from Herbstein, Kapon and Reisner, 1987.)


441

(COOH)22H2O forms hydrogen-bonded sheets rather than chains. In the three-layer oleyl alcohol complex the hydroxyl of the guest molecule, which has a U-shaped conformation, intervenes, from both sides, in two of the six carboxyl bridges of the central of the three layers in the [001] direction. Thus only one-third of the layers are interrupted in this structure. Despite non-planarity of the layers, this complex crystallizes in a centrosymmetric space group. Double interruption of a carboxylic acid dimer by hydroxyls was encountered in the 1 : 2 complex of 1,1 0 -binaphthyl-2,2 0 -dicarboxylic acid with methanol

A

C

A

C 0

B

0

B

Fig. 8.12. Stereoview of the {2TMA[camphor]4H2O} structure down [001], along which there are two layers. The TMA layers are appreciably nonplanar and the space group is chiral, in accord with the chirality of the camphor molecules. The disordered camphor guests are not shown. (Reproduced from Herbstein, Kapon and Reisner, 1988.)

B

B

0 C

A

0 C

A

Fig. 8.13. Stereoview of the {2TMA[0.67(oleyl alcohol)]} structure down [001], along which there are three layers. Although the TMA layers are appreciably nonplanar, the space group is centrosymmetric. The oleyl alcohol molecules are fairly well ordered in the tunnels. (Reproduced from Herbstein, Kapon and Reisner, 1988.)


442

interrupted dimer

carboxylic acid dimer

2C

0

A

B

A B

B

2B B

A A

B

A 2A

A B

z y x

Fig. 8.14. TMPYME. Diagram of a single layer showing the hydrogen bonding scheme. Pyrene A is at general positions and is (essentially) coplanar with the TMA and the hydroxyls of the methanol molecules; pyrene B, which extends out of the plane of the page between the layers, is located about centers of symmetry. The crystallographically independent TMA molecules are designated A and B. Adjacent layers are mutually offset, the degree depending on whether they are separated by pyrene B or methanol methyls. The methyl group of only one of the methanols in the interrupted dimer is shown. (Reproduced from Herbstein, Kapon and Shteiman, 2001.)

(Weber, Cso¨regh, Stensland and Czugler, 1984; CILLUE) (see Chapter 12, Fig. 12.15), the arrangement being very similar to that in 2TMA[0.67(oleyl alcohol)], and in sulfazecin : methanol (Kamiya, Takamoto, Wada and Asai, 1981; FULFZC). The guests forming tunnel inclusion complexes with TMA are a rather eclectic group; there are unbranched (n-tetradecane) and branched (isooctane) paraffins, long chain alcohols of various kinds (heptanol, octanol, decanol and oleyl alcohol), alkenes (1-octene, squalene), alicyclics (camphor) and miscellaneous (methoxy ethyl ether, propellane, epichlorhydrin). However, a comprehensive survey establishing various possible guest types has not yet been made. In this (extended) group, the most detailed structure reports are for the well ordered crystals of {2TMApyrene2(ethanol)} at 198K (SURYUZ) and {2TMA1.5(pyrene) 2(methanol)} at 300K (abbreviated as TMPYME; refcode UCUKUY) (crystal data in Table 8.7). Both complexes have layer structures, the (essentially planar) layers being


443

constructed from rings of six TMA molecules, hydrogen bonded through four ‘‘carboxyl dimers’’ and two ‘‘interrupted dimers’’ where ethanol (methanol) is included in the R44 (12) (graph set) ring. The packing of the layers differs in the two complexes, leading to different three-dimensional structures. In the methanol complex, one pyrene molecule is located within the layer and the other, at a center of symmetry, between the layers in one type of interlayer space, while the methyls of the methanol protrude into the other type of interlayer space (Fig. 8.14). In the ethanol complex, the superpositioning of the layers is such that two types of stack are formed; one of these is mixed, containing pyrene and one of the independent TMA molecules in alternating sequence, while the other stack contains only the second type of TMA. Spectroscopic study is needed to establish whether the partial mixed stack arrangement in the crystalline ethanol complex implies donor–acceptor interaction. 8.3.2.2

Two coordination complexes as potential hosts

This is an opportune point to draw attention to two structures, both based on copper coordination complexes, in which there are analogies to the quasi-hexagonal TMA channels described above. In both examples the channels contain water or other solvent molecules and the intriguing question arises whether these could not perhaps be replaced by guest molecules of other types. There is no doubt that the geometry is right but it is not clear whether the chemical problems can be overcome; no one seems yet to have tried to find out. Interest in the first of these examples was originally stimulated when it was found that a series of copper complexes of composition Cu3L3(OH)X2xH2O (where HL is pyridine-2-carbaldehyde oxime (8.3) and X ¼ 1/2SO42, NO3, ClO4, or OH) had anomalously low magnetic N

N

OH

H 8.3 Pyridine-2-carbaldehyde oxime

moments (Beckett and Hoskins, 1972; CUPRAL10). The problem of the magnetic moment was solved by determining the crystal structure of {Cu3L3(OH)SO4163H2O} ˚ , space group P (trigonal, a ¼ 18.05, c ¼ 7.25 A 3, Z ¼ 2), which was found to have trinuclear (Cu3L3(OH)X2xH2O) groups in which strong metal–metal interaction had caused pairing of two of the three formally unpaired electrons; these units are analogous to TMA. {Cu3L3(OH)SO4163H2O} has trimeric units hydrogen bonded to one another in the ˚ , where the shortening is probably an [001] direction (d(O . . . O) is as short as 2.36(2) A artifact due to systematic errors) while the lateral contacts between the trimers are due to van der Waals interactions. In contrast to the TMA networks, here the interactions normal to the network plane seem to be appreciably stronger than the in-plane interactions. The channel cross section is only approximately circular, with an internal diameter of roughly ˚ . Beckett and Hoskins concluded ‘‘ . . . that most of the water of crystallization (con9A tained in the channels) is liquid in nature, moving randomly throughout the containing volume. Even though the water molecules are not rigidly held in the lattice, they seem essential to the structure of the crystal . . . ’’

444


S

y z

O(H)

CuI

CuI Cu

Cu

z

O(H)

S

C N O Cu

Fig. 8.15. (left) A projection of the trimer unit of Cu3L3(OH)SO42xH2O (where HL is pyridine-2carbaldehyde oxime (8.3)) viewed down its three fold axis; the sulphato group has been omitted for clarity; (right) a schematic view perpendicular to the threefold axis showing the relationship of the Cu3 core, the hydroxy group and the sulphato group. (Reproduced from Beckett and Hoskins, 1972.)

b

y 0

0

x

a

Fig. 8.16. Simplified diagram of the crystal structure of Cu3L3(OH)SO42xH2O (where HL is pyridine-2-carbaldehyde oxime (8.3)) viewed down [001]. The sulphato groups have been omitted for clarity; the copper atoms are shown by the filled circles. The channels are occupied by disordered solvent molecules. (Reproduced from Beckett and Hoskins, 1972.)


445

The second example was prepared by allowing ascorbic acid to oxidize in acidic aqueous solution in the presence of Cu(II) (Norman, Rose and Stenkamp, 1987; JEDKIM). Blue trapezoidal prisms of composition [(Cu2þ)9(HCl)2(cpa3)6(H2O)3]xH2O were obtained by allowing the reaction mixture to stand for three days; the composition was inferred from the results of the crystal structure analysis described below (cpa is an acronym for the branched chain dicarboxylic acid 2-carboxypentonic acid (1,2,3,4tetrahydroxybutane-1,1-dicarboxylic acid (8.4)). In the coordination complex the asterisked hydroxyls of the two carboxylic groups and the 1-hydroxyl group lose protons and are linked to Cu atoms. The compound is insoluble in water and organic solvents, and decomposes in concentrated acids or bases. O OH OH∗

∗HO C

∗HO

CH CH

O

CH2OH

OH

8.4 cpa or 1,2,3,4-tetrahydorxy-butane-1, 1-dicarboxylic acid

O C Cu CL

Fig. 8.17. A view of the structural unit in [(Cu2þ)9(HCl)2(cpa3)6(H2O)3]xH2O, (cpa is an acronym for the branched chain dicarboxylic acid 2-carboxypentonic acid (1,2,3,4tetrahydroxybutane-1,1-dicarboxylic acid (8.4)), emphasizing the two different copper sites. The Cu3Cl(cpa)3 unit with the copper ligating atoms is shown, but the two terminal alcohol units of each cpa have been omitted. (Reproduced from Norman, Rose and Stenkamp, 1987.)

446


Fig. 8.18. A view of one polymeric layer in the crystal structure of [(Cu2þ)9(HCl)2 (cpa3)6(H2O)3]xH2O, the channels being occupied by disordered solvent molecules. The axes ˚ ) and b axes of the trigonal cell (c ¼ 7.98 A ˚ , space group P321). shown are the a ( ¼ 21.27 A (Reproduced from Norman, Rose and Stenkamp, 1987.)

The blue trapezoidal prisms were found to be trigonal, with a ¼ 21.274(5), ˚ , space group P321. The space group is one of the Sohncke group,2 but the c ¼ 7.9766(7) A absolute configuration of the crystal used was not determined. Direct methods showed that one Cu atom was in a general position and one on a twofold axis, while Cl was located on a threefold axis. The Y-shaped Cu6Cl(cpa)3 unit (Fig. 8.17) is essentially equivalent to the trimesic acid molecule from a structural point of view, and the Y-units are linked similarly to those in the quasi-hexagonal TMA network of Fig. 8.11 to form an hexagonal ˚ filled with arrangement (Fig. 8.18), with cylindrical channels of net diameter about 17 A disordered solvent molecules. The diameter of the cylindrical channels is not very dif˚ ) for the TMA analog. ferent from that found ( 14 A 8.3.3

Host–guest tunnel inclusion complexes based on catenated hexagonal unary networks

The hexagonal networks of Fig. 8.11 appear in two other types of inclusion complex in a rather special arrangement. Consider for the moment six TMA molecules hydrogen bonded together in a ring to form the hexagonal motif of the TMA network shown in Fig. 8.11. The cylindrical hole in this ring is large enough to accommodate three similar 2 A Sohncke space group has symmetry elements only of the first kind. We here follow Flack (2003), see especially pp. 914–915.


447

rings threaded through roughly perpendicular to the original ring; and the hole formed by the new rings can in its turn accommodate two additional rings making three in all. This gives an interlinking arrangement of triply triply catenated rings, a catenane or linkedchain structure (latin: catena ¼ chain) being one in which macrocyclic molecules are linked together mechanically without the aid of a chemical bond (Schill, 1971). Now extend this arrangement in space using intersecting networks rather than single rings. The arrangement shown schematically in Fig. 8.19 results, where the networks are seen edgeon. The TMA matrix, if composed of planar networks, cannot fill space and tunnels are left which are parallel to the axes of intersection of the networks. Furthermore the overall arrangement is chiral if the network triplets are not mutually perpendicular – a simple analogy is provided by the biphenyl molecule which is only achiral if the two rings are coplanar or mutually perpendicular. Two examples of this structure type are known. The first is ‘‘trimesic acid pentaiodide’’ (BZHTIB) and its analog in which the I 5 ion is replaced by Br5 or IBr2 . The water molecules in this compound (see caption to Fig. 8.19 for composition) are included between the networks and are not hydrogen bonded to them; it seems that their structural role is to provide a location for the proton counterions. The anions in the tunnels are

y 1/4

1/4

x

1/4

1/4

1/4

1/4

1/4

1/4

0 2Å

Fig. 8.19. Schematic diagram of the triply triply catenated hexagonal TMA networks as described in the text. The direction of view is along the plane of the networks. This is the structure of the ‘‘trimesic acid pentaiodide’’ tunnel inclusion complex, whose composition is TMA0.7H2O [0.09HI5], and its HBr5 and HIBr2 analog (Herbstein, Kapon and Reisner, 1981). It is also the structure of the interstitial inclusion complex ‘‘-TMA’’, whose composition is TMA [0.04C6H4(COOH)20.04TMA] (Herbstein, Kapon and Reisner, 1985). These complexes all crystallize in space group I222, the symmetry elements of which are shown in projection onto (001), and are isostructural. The tunnels at the corners and centre of the diagram, represented by hatched circles, are empty in -TMA but are occupied by polyhalide chains viewed end-on in the TMA polyhalides. (Reproduced from Herbstein, Kapon and Reisner, 1981.)

448


incommensurable with the TMA matrix and give rise to striking diffuse scattering patterns from which it is possible to infer their structures. This is entirely analogous to the situations found in the incommensurable urea, thiourea (Section 6.2.1) and TOT tunnel inclusion complexes (Section 8.2). The second example, so-called -TMA (DAZBOV), was produced by heating TMA to about 300 C, where the compound decomposes slightly as it flash-sublimes and forms an interstitial clathrate complex in which stabilization comes from inclusion of benzene1,3-dicarboxylic acid decomposition products and residual TMA molecules between the networks, while the channels occupied by the anions of the TMA pentahalides remain empty. This is one of the few examples known of a binary adduct not made by crystallization from solution. 8.3.4

Host–guest clathrate interstitial inclusion complexes based on catenated hexagonal unary networks

There is a group of interstitial clathrate complexes based on anhydrous TMA, the structure of which was determined (Duchamp and Marsh, 1969; BTCOAC) many years before the work described above. Anhydrous TMA (called -TMA), crystallized from water, has a complicated structure based on nonplanar triply–triply catenated hexagonal TMA networks (space group C2/c, 48 molecules in the unit cell); the nonplanarity of the networks allows them to be essentially space filling, as can be seen from the relatively high density of 1.46 g cm3. Nevertheless, TMA, when crystallized from water containing Br2, acetone (DAZBUB), I2 (DAZCAI), resorcinol (DAZCEM) or hydroquinone (DAZCIQ), forms interstitial complexes of composition {TMA[pX]} where p ¼ 1/6 for the first two guests and 1/12 for the latter three; the interstitial nature is shown by their densities being greater than 1.46 g cm3. These interstitial complexes are isomorphous with -TMA, the dif˚ ; thus the degree of ferences in cell dimensions being only a few hundredths of an A adaptability is small. The structure of {-TMA[1/6Br2]} has been determined (Herbstein, Kapon and Reisner, 1985; DAZBIP), the disordered bromine molecules being located between the intersecting networks (see Section 10.8). It is not clear whether p can take up any value between 0 and 1/6 (or 1/12), giving true interstitial solid solutions, or whether it is restricted to 1/6 (or 1/12), giving a compound in the phase-rule sense. There is possibly an analogy here with the bromine hydrate structure, where it has been shown that the variable bromine content depends on the conditions under which the crystals are grown (Section 7.2.7.4). 8.3.5

Generalization of the concept of ‘‘interruption’’ to give binary networks

An ‘‘interrupted’’ TMA network has a ring size larger than that achievable with the standard carboxyl dimer linkage. Thus ‘‘interruption’’ allows greater flexibility in designing networks. The simple examples given above can be extended and generalized by following a proposal made by Meleńdez and Zaworotko (1997) that a binary network should be considered as having two components, a ‘‘director’’ and a ‘‘propagator’’, which must be bifunctional. Using {2TMA[2(camphor)]4H2O} as example, TMA is the director and water the propagator. The network need not be planar, and tetrahedral directors can be envisaged (and have been used). The concept has previously been illustrated for a


449

single container molecule in Fig. 3.20, where the director was called a ‘‘spacer’’ and the propagator a ‘‘connector’’. Another possible way of achieving larger ring size in TMA complexes would be to replace –COOH by –CH2COOH, but benzene-1,3,5-triacetic acid is likely to be much less rigid than TMA; perhaps –C C–COOH could be used. 8.3.5.1 TMAH2O networks The complexes described above are all based on hexagonal ‘‘chicken wire’’ unary networks of TMA molecules. However, the networks can contain a second component in addition to TMA, giving a binary network. One example is the group of tunnel inclusion complexes based on planar networks of composition TMAH2O, where the water molecules are an integral part of the network. The simplest of these complexes is TMA3H2O, which has already been noted as a possible tunnel hydrate (Section 6.2.4).

L y

A

B

a x U2

C

D

U1

K E F G

c

e

d

b

H J

I

L

A

B

K

Fig. 8.20. (a) Features of a difference density projection for {TMAH2O[2/9PA]}, with TMA molecules outlined and water molecules shown as circles. The TMAH2O layer has one carboxyl group per molecule hydrogen bonded across centres of symmetry to form the usual carboxylic acid dimer arrangement but the other two carboxyl groups are hydrogen bonded via water molecules (in the centre of the cell). The difference density which represents the included PA molecules is ˚ 3. In {TMAH2O[2/9 PA]} the molecules lie in the ( contoured at levels of 1, 2, 3 e A 1 1 1) planes and the hydrogen bonds do not lie in the plane of projection but are directed out of this plane. For TMA5/6H2O this diagram serves as a slightly distorted representation of the arrangement in the TMAH2O layers. (b) The structure of TMA5/6H2O (space group P1, Z ¼ 12), showing the zigzag chains of included molecules K and L and the stacking of the framework TMAH2O networks A–J. All twelve TMA molecules are crystallographically independent. The view is along c*, with a vertical. (Reproduced from Herbstein and Marsh, 1977.)

450


Although this structure has not been studied in detail, it is clear from what follows that it should be written as {TMAH2O[2H2O]}, where one water molecule is an integral part of the network and the other two are contained in tunnels of rectangular cross section. The complex of composition {TMAH2O.[2/9 picric acid]} has been studied in more detail (Herbstein and Marsh, 1977) and it was found that the picric acid (PA) molecules (originally chosen in the (vain) hope that they would form a substitutional solid solution by replacing TMA molecules in the TMAH2O network) were contained in the rectangular tunnels piercing the TMAH2O network (Fig. 8.20(a)). Striking diffuse scattering on oscillation photographs attested to a partially disordered arrangement of two PA guests ˚ . The unusual composition follows from the along a tunnel, with a periodicity of 16.5 A ratio of four PA molecules to nine TMAH2O layers (each layer contains two TMAH2O ˚ , 2 16.5 ¼ 33 A ˚ . Durene and mesitylene units); 9 3.64 ([001] periodicity) ¼ 32.8 A (1,3,5-trimethylbenzene) are guests in isomorphous complexes, which have not been studied in detail. An ordered chain of hydrogen-bonded methanol molecules, one ˚ within the [CH3OH] replacing [2H2O]} with a rather short O . . . O distance of 2.43 A ˚, chain, is found in an isomorphous complex (reduced cell: 3.679(1), 8.971(1), 18.038(3) A 77.76(1), 86.86(1), 88.04(1) , P1, Z ¼ 2) (Chatterjee et al., 2000; XAVQEQ); a similar complex contains chains of acetone molecules, which cannot be hydrogen bonded. The hydrate with the unusual composition TMA5/6H2O is isostructural with {TMAH2O[2/9PA]}. Its structure is revealed by rewriting the composition as {5(TMAH2O)[TMA]}, indicating that the additional TMA molecule is present as a guest in the rectangular tunnels piercing the TMAH2O layers. However, adjacent TMA guests are hydrogen bonded through carboxyl groups in the meta positions and the chain so formed is zigzag rather than linear. This is accommodated in the crystals by offsetting the TMAH2O layers in sequence (Fig. 8.20(b)). Although both the TMAH2O layers and the zigzag TMA chains are separately centrosymmetric, their arrangement lacks a centre and the space group is the P1. Comparison of reduced cells shows that these crystals are isostructural rather than isomorphous. 8.3.5.2 Catenated neutral binary networks 4,4’-bipyridine was used as a propagator molecule by Sharma and Zaworotko (1996; RAPHAR), with possibly unexpected results. The {(C9H6O6)1.5(C10H8N2)} crystals are ˚ , ¼ 99.95(6) , space monoclinic, with a ¼ 11.105(4), b ¼ 10.132(4), c ¼ 18.889(7) A group P21/c, Z ¼ 4; one bipyridine is at a centre of symmetry and one in a general position. The carboxyl dimers of neat TMA are replaced by strong N . . . H–O and weak C–H . . . O¼C links. The non-planar macrocyclic rings have chair conformations and are doubly doubly catenated (Figs. 8.21 and 8.22) to give the overall crystal structure. The calculated density (it was apparently not measured) is relatively high at 1.41 Mg m3; the presence of void space after catenation of the networks is specifically mentioned but there is no mention of possible additional enclathration of small solvent molecules. 8.3.5.3 Ionic binary networks A number of attempts have been made to engineer porous organic solids by combining cationic propagators with anionic TMA directors. Water molecules can also participate


451

4,4′-bipyridine TMA

Fig. 8.21. The nonplanar binary hexagonal ring motif formed from TMA and 4,4 0 -bipyridine (BP) moieties. The N . . . H–O and C–H . . . O hydrogen bonds are shown by dashed lines. (Adapted from Sharma and Zaworotko, 1996.)

BP

BP A

B

C

TMA

Fig. 8.22. The nonplanar binary hexagonal ring motif formed from TMA and 4.4 0 -bipyridine (BP) moieties. The three independent doubly doubly catenated binary networks are denoted by A, B, C. (Adapted from Sharma and Zaworotko, 1996.)

in the networks formed. Binary honeycomb networks involving various cation–anion combinations were found in the salts tris(dicyclohexylammonium) trimesate methanol solvate {3[(H2N(cyclo-C6H11)2]þ[(C6H3(CO (TOZZUD; at 173K, 2 )3]}xCH3OH ˚ , space group P63, Z ¼ 6) and dimethylammonium trimesate a ¼ 17.609, c ¼ 17.677 A

452


{12[(H2N(CH3)2]þ}{[(C6H3(CO2H)3]3[(C6H3(CO2H)2(CO2)]3[(C6H3(CO2H)(CO 2 )] ˚ , space group R3, Z ¼ 6, (Meleń[(C6H3(CO2)3]} (TUBBAT; a ¼ 33.412, c ¼ 17.465 A dez, Sharma, Zaworotko, Bauer and Rogers, 1996). The honeycomb network in the first of these salts is neutral and is based on interactions of the type shown below; there are no carboxylate dimers. The cyclohexyl groups of the ammonium ions project above and below the plane of the honeycomb sheets and almost close the cavities, which contain disordered solvent molecules. Methanols are also hydrogen bonded to oxygens of carboxylate groups. The (remarkable) second salt contains together neutral TMA and its singly, doubly and triply deprotonated anions; there are two independent honeycomb grids, one of which has [(C6H3(CO2H)3] molecules and [(C6H3(CO2H)(CO 2 )2] in 1 : 3 ratio, while the second has [(C6H3(CO 2 )3] and [(C6H3(CO2H)2(CO2 )] anions in 1 : 3 ratio. The cations cross link adjacent sheets by N–H . . . O hydrogen bonds to the anions rather than to the neutral molecules. c

c N

C R

+

H

O

H

R O

O H

H

+

C O

N c

c

c = cyclohexyl

N,N,N 0 ,N 0 -tetraalkylethylenediammonium cations (alkyl ¼ methyl, ethyl) have also been used as propagators with TMA anions as directors, and these give binary chains which are hydrogen bonded in a series of different arrangements (Meleńdez and Zaworotko, 1997); the salts are N,N,N 0 ,N 0 -tetramethylethylenediammonium bis(benzene3,5-dicarboxylic acid-1-carboxylate) dihydrate (PIDQOI), N,N,N 0 ,N 0 -tetraethylethylenediammonium bis(benzene-3,5-dicarboxylic acid-1-carboxylate) (PIDQUO) and N,N,N 0 ,N 0 -tetramethylethylenediammonium benzene-5-dicarboxylic acid-1,3 – dicarboxylate (PIDSUQ). M(II)acetate hydrate (M ¼ Co, Ni, Zn) also gives binary chains with TMA anions (Yaghi, Li and Groy, 1996), and analogous arrangements are found in [Co(5-C5H5)2]þ (TMA)(TMA)2H2O (PUNSEW)) and [Co(5-C5H5)2]þ [Co(H2O)6]2þ (TMA3) (PUNSIA) (Braga, Angeloni, Tagliavini and Grepioni, 1998). These structures, which can be considered as steps towards the ultimate goal, will not be described in detail here. The binary networks illustrated here could alternatively have been included under the category of ‘‘Mixed Framework Structures’’ as outlined in Chapter 12.

8.3.6

Hydrogen-bonded TMA binary complexes

The crystal structure of TMAdimethyl sulphoxide (DMSO) (Herbstein, Kapon and Wasserman, 1978; TMADMS) is unusual in that the TMA molecules are mutually linked


453

B 3.689 3.807

3.630 3.629

A

O(4)

O(5) C(5)

2.656

2.559

C(8) O(4)

2.635

3.739

117.9

C(3)

O(3)

C(9) O(6)

C(6)

C(2) C(1) C(7)

a cos 10°

O(2) O(1) c cos 10°

b up

Fig. 8.23. One layer of the crystal structure of {TMADMSO} projected approximately onto (010). (Reproduced from Herbstein, Kapon and Wasserman, 1977.)

through single >C–OH . . . O¼C < hydrogen bonds between TMA molecules along [100] (and thus carboxylic acid dimers are not formed), and through the oxygen of DMSO along [001] (Fig. 8.23), the remaining atoms of the DMSO molecule filling the tunnels between the TMA molecules. Thus this is a layer structure with the so-called guest DMSO playing partly the role of host and partly that of guest. Ambiguity remains about the space group which could be P21 or P21/m. {TMADMSO} is on the borderline of ‘‘inclusion complexes’’, but, for completeness, we note here some other TMA complexes with neutral components and organic salts of TMA (metal trimesate salts are excluded). In the first two of these examples there is hydrogen bonding between TMA and the second component, and these could have been included in Chapter 12. Examples of hydrogen bonded TMA complexes with dimethylformamide (XAVPOZ), dimethylamine (XAVPUF), and N,N,N 0 ,N 0 -tetramethylethylenediamine (PIDQOI01) have been given by Chatterjee et al. (2000), where both strong and weak hydrogen bonds appear to play important roles in determining the packing arrangements, which are all different. {TMA(Ph3PO)2} (prepared by refluxing TMA and triphenylphosphine in 1 : 2 ratio in toluene) is triclinic (Lynch, Smith, Byriel and Kennard, 1992; P1, Z ¼ 2; KUCCUG). The crystal structure can be described as made up of units of {TMA(Ph3PO)2}, linked in chains. The two Ph3PO molecules of the asymmetric unit are linked, as acceptors, to two different carboxyls of a particular TMA molecule by fairly strong hydrogen bonds ˚ ). The OH of the third carboxyl of this TMA molecule acts (d(O . . . O) ¼ 2.50(1), 2.54(1) A

454


OPPh3

O

H O

OPPh3 O

Ph3PO

H

O O

H O

O

H O

O

O O

H

O

H

Ph3PO

Fig. 8.24. Schematic view of part of the TMA(Ph3PO)2 structure showing linkage of TMA and Ph3PO molecules in chains along [100] with Ph3PO molecules appended (cf. Chapter 12). The hydrogen bonds linking the TMA molecules are emphasized while the Ph3PO molecules are boxed to show up the appendage nature of the arrangement of the two moieties in the crystal.

as donor to a carbonyl oxygen of one of the first mentioned carboxyls ˚ ). The unionized state of all three carboxyls is confirmed by the (d(O . . . O) ¼ 2.61(1) A ˚ between C–OH and C¼O bond lengths. Most hydrogen distinct differences of 0.1 A bonded binary adducts with a chain structure have the components arranged in alternating fashion . . . A . . . B . . . A . . . B . . . A . . . In TMA(Ph3PO)2 there are chains of hydrogen bonded TMA molecules to which (Ph3PO)2 units are appended; the triphenylphospines do not participate in the chains. This is shown schematically in Fig. 8.24. The chains interact mutually through van der Waals linkages and are not hydrogen bonded to one another. Appendage structures are also discussed in Chapter 12. TMA and urea form three crystalline binary adducts, with 1 : 1, 1 : 2 and 1 : 3 compositions. Structural information is available only for the 1 : 3 composition (VidenovaAdrabrinska, 1994, 1996). The crystals are monoclinic (a ¼ 6.718(5), b ¼ 20.412(8), ˚ , ¼ 92.70(6) , Z ¼ 4, space group P21/n; CEKSIU); the somewhat high c ¼ 12.730(4) A RF factor of 10.4% was ascribed to polysynthetic twinning. There are seventeen independent hydrogen bonds in the structure and we shall limit ourselves here to a brief description. The formula unit is also the structural unit, with two carboxyl groups bonded to ureas, each through a pair of C–O–H . . . O¼C hydrogen bonds (d ¼ 2.479(5), 2.493(6) ˚ ; these are shown as 1 and 2 in Fig. 8.25) and C¼O . . . (syn)H–N hydrogen bonds A ˚ ) (TMA atoms on the left in all formulae where TMA . . . urea (d ¼ 2.991(6), 2.971(6) A hydrogen bonds are concerned). The TMA and these two ureas are essentially coplanar and form a set of sinusoidal ribbons which run through the crystal, with urea . . . urea ˚ ) linking the C¼O . . . H–N hydrogen bonds (d ¼ 3.035(6), 2.837(6), 2.929(6), 3.299(6) A planar portions of the structural units; it is the anti N–H groups of the ureas which participate in the formation of these chains. One could describe the chains as consisting of hydrogen-bonded urea dimers (emphasized for clarity in Fig. 8.25) linking meta


455

B

2 3 1

0A

C

Fig. 8.25. The slice of the TMA(urea)3 crystal structure lying in (or about) the plane of the TMA molecules. The hydrogen bonded pair of urea molecules within this plane is emphasized, while the third urea molecule (out of the plane) is seen approximately edge-on. The hydrogen bonds referred to in the text are denoted by 1, 2, 3. One sinusoidal chain of molecules is enclosed within the hatched area, but it must be remembered that hydrogen bonds between chains cross the chain boundaries. (Adapted from Videnova-Adrabrinska, 1994.)

carboxyl groups of TMA. The third urea is nearly perpendicular to this layer and is ˚ ) to the third hydroxyl of TMA. All these C–O–H . . . O¼C hydrogen bonded (d ¼ 2.545(5) A hydrogen bonds are ordered, as is shown by the occurrence of distinct single and double C–O bonds in the three carboxyl groups. Although description in terms of sinusoidal ribbons is convenient, this is not the whole story because the ribbons are not isolated but are also mutually linked by hydrogen bonding. Furthermore, the third urea, perpendicular to the ribbon plane, links adjacent layers, above and below, by hydrogen bonds; we do not give details. The crystalline complex is thus a three-dimensional hydrogen-bonded structure. TMA forms a triclinic 1 : 1 : 1 ternary complex with dioxane and water, which has pleated sheets of (neutral) TMA and water molecules, bridged by dioxanes (Herbstein and Kapon, 1978; TMESAD). There is also a ternary complex of composition glycinetrimesic acid monohydrate (space group Pna21, Z ¼ 4; Herbstein, Kapon, Maor and Reisner, 1981; GLYTMS), where glycine is present as a zwitterion, with the cationic charge localized on the NHþ 3 group. The balancing negative charge is shared between one of the carboxyls ˚ ), almost of TMA and that of glycine, which are linked by a very short (2.464(3) A symmetrical hydrogen bond. This is also reflected in the C–O bond lengths of these ˚ for TMA and 1.232(2) and carboxyl groups, which are 1.223(4) and 1.284(4)* A ˚ for glycine (hydrogen between asterisked oxygens). The other two carboxyls 1.275(4)* A of TMA have regular C–O bond lengths (1.215(4) and 1.318(4), and 1.212(3) and 1.303(4)

456


˚ ). There are no carboxylic acid dimers but one pair of carboxyl carbonyl acceptors of A TMA is linked through the donor hydroxyls of the water molecule. TMA and glycine are acids of similar strengths (pKa ¼ 2.12 and 2.34 respectively) and it was suggested that this was the reason for the formation of an almost symmetrical hydrogen bond. Histidine forms a salt with TMA and both {L-histidinium trimesate.1/3acetone} (LHISTM) and {DL-histidinium trimesate1/3acetone} (DLHTMS) have essentially– identical cell dimensions and related space groups (P212121 and Pna21 respectively) (Herbstein and Kapon, 1979); the acetone could be replaced by water but the crystals were of poorer diffraction quality. The structures contain ribbons of cations and anions extending along [010]. Each ribbon has histidinium ions of one sense of chirality only; in the L-salt all ribbons are congruent while in the DL-salt the sense of chirality alternates in the [100] direction, normal to the ribbon axis. 8.4

The Heilbron complexes

Soon after the end of the First World War it was reported by Heilbron and Buck (1921) that the compound now named as E,E-1-[ p-dimethylaminophenyl]-5-[o-hydroxyphenyl]penta-1,4-dien-3-one (for structure diagram see Fig. 8.26) formed molecular complexes with a wide variety of guests, including ethanol, chloroform, acetic acid, m-dinitrobenzene, p-dimethylaminobenzaldehyde (PDMB), benzene, 4-methoxybenzaldehyde (anisaldehyde), and 2,4,6-trinitrotoluene; Heilbron and Buck used the name 4 0 -dimethylamino-2-hydroxydistyryl ketone, which we abbreviate for our use here as DHDK. Indeed Heilbron and Buck noted that ‘‘the ketone is very difficult to obtain in the free state as it tenaciously retains traces of solvent;’’ diffraction-quality single crystals of neat DHDK have not yet been obtained (cf. deoxycholic acid, Section 6.3.1; TATM Section 8.6). Thus the crystal structure of neat DHDK is not known but those of its complexes with ethanol (DADYUC) and chloroform (DADYOW) (isomorphous, 2 : 1), m-dinitrobenzene (1 : 1) (DADZIR) and PDMB (1 : 1) (DADZOX) have been reported (Herbstein, Kapon, Reisner and Rubin, 1984); also crystal data for the acetic acid (DADZAJ) and methanol (DADZEN) complexes. The DHDK molecule can take up a number of conformations, although these are not all equally probable; the two so far reported are shown in Fig. 8.26. The DHDK molecule has three functionalities – the hydroxyl and carbonyl groups can act as hydrogen bond donors HO

HO

O N CH3 CH3

O

CH3 N CH3

Fig. 8.26. The s-trans,trans (on left) and s-cis,trans (on right) conformations of DHDK. The first of these conformations is found in the {DHDK[0.5X]} clathrates and in {DHDK p-dimethylaminobenzaldehyde}, and the second in {DHDKm-dinitrobenzene}. The single bond about which there is a difference of conformation is marked. (Reproduced from Herbstein, Kapon, Reisner and Rubin, 1984.)

T HE HE IL BR ON C OM PLEXES

457

and acceptors respectively, while the dimethylaminophenyl and hydroxyphenyl rings can act as donors and acceptors respectively in -* charge transfer compounds. Thus it is possible that the 1 : 1 TNT and TNB adducts are -* charge transfer compounds in which the dimethyl-aminophenyl ring acts as donor and TNT or TNB as acceptor; structures were not determined because of poor crystal quality. One can envisage hydrogen bonding of hydroxyl and carbonyl groups to form DHDK dimers or chains; in addition guests with suitable hydrogen bonding capabilities could link with the DHDK molecules. All these possibilities are realized in the structures determined. Centrosymmetric dimers of DHDK are formed by head-to-waist hydroxyl-carbonyl hydrogen bonding, the DHDK molecules being in the s-trans-trans conformation. These dimers form isomorphous 2 : 1 clathrates with ethanol and CHCl3 as guests (Fig. 8.27). It is possible that the 1 : 1 acetic acid complex is isostructural but the structure has not been determined. There are only van der Waals forces between DHDK dimers and between DHDK dimers and guests, so these complexes are analogous to the clathrates of tri-othymotide (Section 8.2). The {DHDK[m-dinitrobenzene]} complex, although a tunnel inclusion complex, is of a kind not yet encountered elsewhere. The molecules are arranged in sheets, with the DHDK molecules linked in chains by head to waist hydrogen bonds (Figs. 8.28 and 8.29). The remaining space in the sheets has a sinuous rather than linear shape and the m-dinitrobenzene molecules fit in rather neatly, with van der Waals interactions both to each other and to the DHDK hosts. These sheets are then superimposed in an offset manner which leads to the enclosure of the guest molecules in tunnels. The interactions between host molecules are a combination of hydrogen and van der Waals bonding.

C

A

C

A

Fig. 8.27. Stereodiagram of the {DHDK[0.5X]} crystal structure, where X ¼ C2H5OH or CHCl3. The guest molecules, which are disordered and not shown, are enclosed in the cavities centered at 0,1/2,0 and analogous positions. The hydrogen bonds are shown as dashed lines. (Reproduced from Herbstein, Kapon, Reisner and Rubin, 1984.)


458

C

B

A

z x

Fig. 8.28. DHDK-m-dinitrobenzene (DADZIR) crystal structure viewed down [010] showing the molecules (host and guest) located in the (201) planes. The DHDK molecules are linked by carbonyl . . . hydroxyl hydrogen bonds. The host . . . guest interaction is via van der Waals forces. The arrangement within the sheets is shown in Fig. 8.29. (Data from Herbstein, Kapon, Reisner and Rubin, 1984.)

B(666)

A(674) B(665)

B(764) b

A(665)

A(566) A(664) B(655)

[102]

0 B(656) A(556)

A(655)

Fig. 8.29. The {DHDK[m-dinitrobenzene]} structure showing sheets of molecules lying in the (201) planes. The two crystallographically independent molecules of each type are designated A and B. The reference molecules are denoted as 555, translations along the crystal axes being specified by adding or subtracting integers from the reference code, as in the ORTEP system. The rectangle shows the unit cell of the pg plane group. The hydrogen bonds are denoted by thin lines. Note the s-cis, trans conformation of DHDK. (Reproduced from Herbstein, Kapon, Reisner and Rubin, 1984.)

GOSSYPOL AND ITS INCLUSION COMPLEXES

459

DHDK

PDMB

B

A

0

y x carbonyl...hydroxyl H-bond (2.723Å)

Fig. 8.30. DHDK-p-dimethylaminobenzaldehyde (DADZOX) crystal structure viewed down [001]. The packing unit is the DHDK . . . PDMB pair linked by a carbonyl . . . hydroxyl hydrogen bond. The DHDK molecules are not hydrogen bonded to one another. The PDMB molecules have been emphasized. (Data from Herbstein, Kapon, Reisner and Rubin, 1984.)

Finally there is the {DHDKPDMB} structure (Fig. 8.30) (PDMB ¼ pdimethylaminobenzaldehyde), which is the most difficult of the DHDK complexes to categorize in terms of the classification used here. There is a hydrogen bond between DHDK hydroxyl and PDMB carbonyl; in addition the DHDK molecules are arranged so as to form the walls of tunnels which contain the guest molecules. However, it is hardly a classical tunnel inclusion complex; possibly dipole–dipole interactions, both host–host and guest–guest, play an important role in determining the overall arrangement. 8.5 Gossypol and its inclusion complexes Gossypol (8.5; 1,1 0 ,6,6 0 ,7,7 0 -hexahydroxy-5,5 0 -diisopropyl-3,3 0 -dimethyl[2,2 0 -binaphthalene]-8,8 0 -dicarboxaldehyde) is a yellow pigment isolated from cotton seed kernels as the racemate; formation of enantiomers is due to restricted rotation about the central bond. Gossypol was named by Marchlewski (1899), its formula determined by Adams, Morris, Geismann, Butterbaugh and Kirkpatrick (1936), and synthesized by Edwards (1958) (see also Adams, Geismann and Edwards, 1960). Gossypol has three possible tautomeric forms – aldehyde, lactol and quinoid – but has so far always appeared as the aldehyde in the crystalline state. It has antitumor, antiviral and antifertility properties. Enantiomeric gossypol ([]D19 ¼ þ445(10) ) has been isolated from Thespesia populnea (Bhakuni, Dhar and Sharma, 1968; King and de Silva, 1968) and is the first example of a natural - 0 -dinaphthyl derivative showing optical activity due to restricted rotation. The chiral material has also been obtained by chromatographic separation of the racemate (Clark, 1927; Matlin, Belengeur, Tyson and Brookes, 1987). The crystallography of Gossypol and its inclusion complexes has been comprehensively reviewed (Gdaniec, Ibragimov and Talipov, 1996). ‘‘Gossypol’’ gives 147 hits in the October 2002 version of the CSD.


460

O

OH

OH

HO

OH OH

HO

O

8.5

The ideal symmetry of the molecule would be C2–2, with the two fold axis normal to the central bond (Fig. 8.31). The torsion angle about this bond would be expected to be 90 ; in practice deviations of up to 15 from the perpendicular conformation are found. There are two strong intramolecular hydrogen bonds (O(3)–H . . . O(2) and O(7)–H . . . O(6)), forming six-membered rings, and two weaker ones (O(4)–H . . . O(3) and O(8)–H . . . O(7)), forming five-membered rings and participating weakly in threecenter hydrogen bonds. The polymorphism of gossypol has been discussed by Ibragimov and Talipov (1994) and we follow their notation; seven polymorphs in all have been identified. Two of these were prepared by recrystallization from suitable solvents. Monoclinic P1 (volume per molecule ˚ 3; Z ¼ 4; space group P21/c; Talipov, Ibragimov, Dadabaev, Aripov and V ¼ 627.9 A ˚ 3; C2/c, Sadykov, 1986; listed by CSD as BEMLOU 03 and 04). Monoclinic P2 (V ¼ 645.4 A Z ¼ 32) is not listed by the CSD. Five polymorphs (P3–P7), identified by their powder diffraction patterns (but without assignment of unit cells), were prepared by removal of guest molecules from various complexes. In addition, P3 was obtained (in rather intriguing fashion) as ‘‘monocrystals’’ from the decomposition of the CH2Cl2 complex of gossypol ˚ 3 (this volume includes empty tunnel space, see p. 467; C2/c, Z ¼ 8; Talipov, (V ¼ 710 A Ibragimov, Nazarov, Aripov and Sadykov, 1985; this is listed by the CSD as BEMLOU 02,

C2 C22

C26

C3

C29

C8

C30

C14

C28

01

C13

08

C21

C5

C4

C3 C11

C6 C10

C2

C20 C19

C16

C9

C1 C12

C15

C7

C3∗

C1∗ C2∗

C4

C24 C23 C25

05 C17

C18

C4∗ C6∗

C5∗ C27

07

C7∗ 06

Fig. 8.31. The spatial formula of gossypol. The hydrogen bonds are denoted by dashed lines. The crystallographic numbering is shown. The di-n-propyl ether guest shown has cross-hatched atoms (Reproduced from Gdaniec, 1991a.)


461

12 and 22). The CSD also lists monoclinic BEMLOU and triclinic BEMLOU 01. Although some interesting relationships appear to be implied by these results, we shall hardly consider them further because of lack of direct relevance to our concerns. About 100 molecular inclusion complexes have been prepared from racemic gossypol by recrystallization from various solvents (Ibragimov and Gdaniec, 1992). Single crystals have been grown for about 70 of these and crystal data determined; cell dimensions may vary somewhat from one report to another but we have not attempted to impose any uniformity. Enough crystal structures have now been reported to demonstrate the wealth of variety found in the arrangements of the components and to outline features of the main structural families. Ibragimov, Talipov and Aripov (1994) remark that ‘‘the total number of various gossypol clathrate isostructural groups discovered by us to date is 20.’’ There are indications that enantiomeric gossypol also forms molecular complexes (for example, with acetone but not with acetic acid) but these have not yet been explored in any depth (or, perhaps, reported in any detail). We shall classify the crystalline molecular complexes of racemic gossypol in terms of 1. gossypol : guest ratio ( 2 : 1, 1 : 1, 1 : 1.5, 1 : 2); 2. gossypol packing unit; 3. structural sub-family (tunnel, clathrate or layer type complexes); and give examples of nine different structural types. In most of the complexes studied until now, the ‘‘packing unit’’ is the centrosymmetric dimer made up of two gossypol enantiomers, with two fairly strong O(5)–H . . . O(3) (or O(1)–H . . . O(7)) hydrogen bonds ˚ ). (Parenthetically we note that such dimers cannot be ‘‘packing between them (d 2.8 A units’’ of enantiomeric gossypol). These dimers may then be hydrogen bonded to one another to form more extended arrays of various types, and may also be hydrogen bonded to the guests. There are often intermolecular hydrogen bonds of about the same strength as the intradimer bonds, so that some care must be taken with the descriptions. A few complexes have been found to contain individual gossypol molecules which then form more extended hydrogen-bonded arrays. The complexity of the arrangements can make clearcut assignment to a particular structural class difficult. Disorder of the guests can also complicate descriptions. 2 : 1 layer inclusion complexes of the gossypol : m-xylene type: Triclinic complexes of this isomorphous group are formed with o-, m- and p-xylene, ethylbenzene and p-chlorotoluene as guests; the crystal structure of the m-xylene and CCl4 complexes have been reported (Fig. 8.32; Ibragimov, Talipov, Aripov and Sadykov, 1990; Ibragimov, Talipov and Zorky, 1994). There are close similarities in cell edges and angles among all the crystals (Table 8.8). In general (see below) the cell angles provide a finer means of discrimination among the slightly differing structural types than cell edges; the gossypol complexes provide excellent examples of the value of cell dimensions in suggesting evidence of structural differences in advance of execution of full structure analyses. Of course, comparisons of this kind can only usefully be made among reduced cells. It is possible to identify the centrosymmetric dimers noted above as the ‘‘packing unit,’’ but there are other intermolecular hydrogen bonds of about the same strength as the intradimer bonds, leading to the formation of infinite columns of gossypol


462

molecules along [010]. Thus the term ‘‘packing unit’’ has formal rather than physical significance. The CCl4 guests are enclathrated between the host molecules and it is perhaps a matter of taste whether this structure type is to be described as a clathrate or a layer arrangement. ˚ , deg., A ˚ 3) for 2 : 1 gossypol : guest complexes of the Table 8.8. Crystal data (reduced cells, A ˚ , angles 0.01–0.03 , volumes 2(gossypol) : m-xylene type. S.u.’s of cell edges are 0.001–0.007 A 3 ˚ 0.4–1.0 A . The data have been ordered according to decreasing cell volumes within a group Guest

a

b

c

Cell volume

2(gossypol) with Ethylbenzene; VEVSUK10 m-xylene*; JIDSOE o-xylene; VEVSIY p-chlorotoluene; VEVTIZ p-xylene; VEVSOE

8.451 8.478 8.505 8.575 8.406

14.195 14.087 14.072 14.060 14.079

14.411 14.399 14.395 14.269 14.213

114.89 115.39 115.51 115.48 115.82

102.55 104.89 105.76 107.11 92.20

92.21 93.20 93.37 93.22 104.76

1513 1475 1466 1451 1443

gossypol with CCl4; CUVLAG Paraldehyde; VEVTEV

8.847 8.97

14.015 12.95

14.304 14.69

102.16 101.2

91.12 91.1

105.79 90.5

1663 1674

˚ 3 by Ibragimov, Talipov, Aripov and Note: the cell volume for the CCl4 complex is given erroneously as 1547 A Sadykov, 1990.

Intercalated carbon tetrachloride guests

C

Layers of H-bonded gossypol hosts

B

A

z y

Fig. 8.32. Triclinic 2(gossypol)[CCl4] projected down [100] (CUVLAG01). Hydrogen bonds are shown (as dashed lines). (Data from Ibragimov, Talipov and Zorky, 1994.)


463

2 : 1 layer inclusion complex gossypol: 0.5(benzene). The crystals (Gdaniec, Ibragimov and Talipov, 1990b) are triclinic (space group P1, Z ¼ 2) with a ¼ 11.241(3), ˚ , ¼ 98.89(2), ¼ 99.86(2), ¼ 98.91(2) , b ¼ 14.986(4), c ¼ 17.380(4) A 3 ˚ V ¼ 2800(2) A (reduced cell; CUVKIN; JIDTIZ). There are two crystallographically independent gossypol molecules in the asymmetric unit, A and B, with dihedral angles between naphthyl rings of 88.2(1) ) and 79.1(1) respectively. Centrosymmetric dimers of the familiar type are formed between pairs of A molecules and between pairs of B molecules. These dimers are linked into columns along [211] by ˚ ) hydrogen bonds (the O(1A)–H . . . O(8B) and O(8B)–H . . . O(4A) (d ¼ 2.77, 3.17 A corresponding bonds with A and B labels interchanged do not occur). 2 : 1 layer inclusion complexes of the gossypol: 0.5(amyl acrylate) type. Complexes of this isomorphous group are formed with amyl acrylate and amyl acetate crystals (Gdaniec, Ibragimov and Talipov, 1990a). The crystals are triclinic (space group P1, Z ¼ 2) with (for amyl acrylate, VEVMOY) a ¼ 14.425(2), b ¼ 15.519(1), ˚3 ˚ , ¼ 97.89(1), ¼ 117.80(1), ¼ 67.01(1) , V ¼ 2992.3(14) A c ¼ 16.409(2) A ˚ (reduced cell a ¼ 14.425(2), b ¼ 15.519(1), c ¼ 16.029(7) A, ¼ 77.78(2), ¼ 64.90(3), ¼ 67.01(1) ). There are two crystallographically independent gossypol molecules in the asymmetric unit, A and B, with dihedral angles between naphthyl rings of 97.6(1) ) and 105.5(1) respectively. Two A molecules related by a centre of symmetry form the dimer identified previously, but two additional gossypol molecules, of the B type, are hydrogen bonded to the gossypols of the dimer (d(O5A)– H . . . O(8B) ¼ 2.983(7), d(O8B)–H . . . O(8A) ¼ 2.813(5), d(O1B)–H . . . O(4A) ¼ ˚ ). Amyl acrylate guests are hydrogen bonded to A and B gossypols 3.125(5) A ˚ ). Presumably it is appropriate to describe this as a clathrate (d ¼ 2.75, 2.98, 3.01 A type structure (Fig. 8.33). 2 : 1 clathrate inclusion complexes of the gossypol:0.5(ethyl acetate) type. Crystallographic data have been reported for fourteen complexes of this group, the guests being ethyl acetate (VEVTUL*), n-butyl acetate (VEVVAT)*, acetylacetone, methyl propionate (VEVTOF*), ethyl acetoacetate (VEVWUO)*, acetyl acetone, methyl acrylate, ethyl bromoacetate (VEVVEX) (Ibragimov, Talipov and Gdaniec, 1990), ethyl acrylate (KIVCEX, given as 1 : 1), acetylacetone (VEVWIC, KIVCIB), propyl butyrate, di-n-propyl ether (KIVCAT*) and butyl ethyl ether (KIVCUN, given as 1 : 1) (Gdaniec, 1991a) (asterisks indicate that crystal structures have been reported). The isomorphous crystals are monoclinic, space group C2/c, Z ¼ 8, with the a parameter ˚ , from varying from 11.01 to 11.54, b from 30.54 to 30.77, c from 16.47 to 17.09 A ˚ 3. The complexes with isobutyl 90.1 to 92.4 , and cell volumes from 5601 to 5928 A acetate (VEVVUN) and methyl (S)-(–)-2-chloropropionate (KIVDAU*) as guests have very similar cells but the space groups are P21/n and the enantiomorphic C2;3 presumably the differences in arrangement are small. The packing unit is not the hydrogen-bonded dimer noted above but instead bimolecular layers are formed containing molecules of the same chirality sense, which are hydrogen bonded within the layers through the three hydroxyl groups of each molecule. These layers interact by van der Waals forces and there are no hydrogen bonds between them. Guest molecules having a carbonyl group in a chain of limited length (around seven 3

space group C2 in the original paper but C2/c in CSD.


464

atoms) can be accommodated in cavities formed between the layers, possibly in disordered fashion; the carbonyl group is hydrogen bonded to O(5) of the gossypol. The gossypol: 0.5(ethyl acetoacetate) structure is shown in Fig. 8.34. One wonders whether the gossypol molecules which complex with the chiral methyl-(S)-(–)-2-chloropropionate guest in the enantiomorphic C2 unit cell, have undergone spontaneous resolution on crystallization, giving all the layers the same sense of chirality. As there are two gossypol molecules in the asymmetric unit, presence of both enantiomers is allowed despite the enantiomorphism of the space group, nor is there any indication from the cell dimensions of such an effect. A full crystal structure analysis would be needed for clarification of this point. A

gossypol A dimer

gos sypol B

two guest molecules

x y C

B

Fig. 8.33. Gossypol : 0.5(amyl acrylate) structure viewed down [010] of the triclinic cell. Guest molecules emphasized; dashed lines are H bonds. (Data from Gdaniec, Ibragimov and Talipov, 1990a.) C

B

A

Fig. 8.34. Projection of the gossypol : 0.5(ethyl acetoacetate) structure down [100] of the monoclinic cell. The disordered guest molecules, which are hydrogen bonded to gossypol, are emphasized. Two hydrogen bonded bimolecular layers of gossypol molecules are shown; these interact by van der Waals forces. (Data from Ibragimov, Talipov and Gdaniec, 1990.)


465

1 : 1 tunnel inclusion complexes of the gossypol : acetone-type. Cell dimensions have been given for 16 complexes in this group (Ibragimov, Gdaniec and Dadabaev, 1990; Table 8.9). The crystals are all triclinic, space group P1, Z ¼ 2. The unit cells reported by the original authors are, except where noted otherwise below, reduced cells but with unconventional choices of origin; a conventional reduced cell has, inter alia, all interaxial angles obtuse or all acute and we have made the necessary changes in Table 8.9. The cell dimensions indicate that there are two groups of isomorphous complexes, and an additional miscellaneous group. The dihedral angles between the nearly planar naphthyl moieties are 82.9, 83.9, 83.0 and 86.0 in VEVRJX, VEVROD, VEVRUJ and JEGWAT respectively, suggesting that the latter differs somewhat from the other three. The packing unit is the centrosymmetric dimer, with the nearly planar naphthyl moieties approximately ˚ in mutually orthogonal, and intramolecular hydrogen bond lengths of 2.76 and 2.80 A ˚ , deg., A ˚ 3) for 1 : 1 gossypol : guest complexes. S.u.’s of Table 8.9. Crystal data (reduced cells, A ˚ ˚ 3. Within the groups the cell edges are 0.001–0.007 A, angles 0.01– 0.03 , volumes 0.51.0 A data have been ordered according to decreasing cell volume. Asterisks denote, as before, that full structure analyses have been reported Guest

a

Group I (gossypol-acetone type) 1-butanol; VEVNIT 11.090 2-methyl-1-propanol; 10.780 VEVPAN 2-butanone; VEVNAL 10.775 1-propanol; VEVNEP 10.841 prop-2-enol; VEVNOZ 10.861 acetone;* CUVKEJ 10.665 acetonitrile;* JEGWAT 10.938

b

c

Cell volume

11.090 11.204

14.510 14.399

102.60 103.19

110.90 107.53

101.10 101.61

1554 1545

11.114 11.073 11.035 11.135 10.982

14.421 14.205 14.142 14.379 14.162

102.95 102.32 101.60 103.53 102.01

108.74 109.87 110.49 108.67 112.24

101.23 101.13 100.45 102.28 102.75

1526 1500 1496 1494 1453

Group II (gossypol-butanal type). There seem to of structures. Pentanal; VEVPOB 10.343 11.643 3-methyl-1-butanol; VEVPER 10.048 11.677 1-butanal* VEVRUJ 10.190 11.335 3-buten-1-al; VEVPIV 10.258 11.271 Group III (Miscellaneous) Dioxane LOQSEP trichloroacetic acid; VEVRAP methacrylic acid; VEVPUH cyclohexanone* ; VEVROD tetrahydrofuran* ; VEVRJX 1-methylethanol; VEVNUF

10.905 11.178 10.996 10.803 10.788 10.585

11.055 11.425 11.065 11.157 10.979 11.152

be differences between the two pairs 14.928 15.426 14.665 14.455

108.83 110.08 106.96 106.47

106.47 107.64 103.74 102.41

95.40 94.69 98.93 98.97

1597 1585 1527 1523

13.772 13.139 13.452 14.692 13.880 14.017

107.07 107.39 98.24 104.61 99.89 101.53

96.72 97.07 107.09 104.73 103.87 108.80

98.68 95.06 94.30 103.34 102.04 97.64

1546 1575 1536 1573 1518 1500

References: Leading references are Talipov, Ibragimov, Tischenko and Aripov, 1989; Gdaniec, Ibragimov and Dadabaev, 1990; Ibragimov, Talipov and Zorky, 1994; Ibragimov, Gdaniec and Dadabaev, 1990; details can be obtained through the REFCODES.


466

the VEVRJX and VEVROD structures respectively. There is a guest molecule ˚ in the VEVRJX and VEVhydrogen bonded to O(1) of each host (d ¼ 2.64, 2.78 A ROD structures respectively). The arrangement is shown in the stereodiagram of the VEVROD structure (Fig. 8.35), which can be taken as a prototype of this group of structures. This packing unit, somewhat modified, persists in the other structures of this group; Ibragimov, Gdaniec and Dadabaev (1990) consider that there are threecentered hydrogen bonds between such packing units but the intermolecular distances indicate that these are surely very weak. Comparison of the detailed structures shows considerable variability in overall arrangement resulting from accommodation of rather different guests in a single basic structure type. 1 : 1 tunnel inclusion complexes of the gossypol : chloroform type. Crystallographic data have been reported for five complexes of this group, the guests being chloroform (Gdaniec, Ibragimov and Talipov, 1990b; CUVKUZ*), diiodomethane, 1,2-dichloroethane, 1,2-dibromoethane and isovaleric acid ((CH3)2CHCH2COOH)* (Gdaniec, Ibragimov and Dadabaev, 1990). The isomorphous crystals are monoclinic, space group C2/c, Z ¼ 8, with a varying from 28.22 to 28.84, b from 8.95 to 9.16, c ˚ , from 107.8 to 109.7 , and cell volumes from 6235 to 6615 A ˚ 3. from 25.80 to 26.88 A The packing unit is again the centrosymmetric gossypol dimer, and there is no host– guest hydrogen bonding. Isovaleric acid forms centrosymmetric dimers in its complex, thus explaining how it can be incorporated as a guest in the same structural group as the hydrophobic halomethanes and haloethanes. The overall arrangement is (roughly) one of gossypol dimers in ( 101) planes, interleaved by planes of guest molecules,

host guest H-bond

gossypol

cyclohexanone

x z y

Fig. 8.35. Stereodiagram of gossypol[cyclohexanone] projected down [110], as a representative ˚ ) is shown of the gossypol[acetone] group of structures. The host-guest hydrogen bond (2.775 A as a dashed line, and the guests are emphasized. (Data from Ibragimov, Gdaniec and Dadabaev, 1990.)


467

giving an intercalate type complex. The gossypol dimers form columns along [101] ˚ ); because the hydrophilic linked by O(1)–H . . . O(8) and O(8)–H . . . O(4) (d ¼ 2.94 A groups of the gossypol molecules are located within the columns, there is no hydrogen bonding between columns. 1 : 1 tunnel inclusion complexes of the gossypol : dichloromethane type. The guests CH2Cl2 and CH2Br2 form isomorphous 1 : 1 complexes with gossypol (for CH2Cl2 ˚ , ¼ 113.05(2) , Z ¼ 8, C2/c; a ¼ 21.320(4), b ¼ 19.129(6), c ¼ 15.765(2) A Ibragimov, Talipov and Aripov, 1994; JIDTOF). These are tunnel inclusion complexes which contain the centrosymmetric dimers so characteristic of the gossypol family of complexes. It is remarkable that these two complexes decompose (lose solvent at room temperature) by a single crystal to single crystal mode to give the P3 polymorph of gossypol (see p. 460). The volume decrease is 4% for the CH2Cl2 ˚ , a reduction of 3.2%) and 9.4% complex (due primarily to a change of c to 15.267 A for the CH2Br2 complex (reduction of c by 4.9%). A view down the tunnel axis is shown in Fig. 8.36. 1 : 1 inclusion complexes of gossypol with methyl acetate and acetic acid. These two examples are isomorphous (P 1, Z ¼ 2): ˚ , 92.34, 91.90, 98.71 , cell volume ¼ GOSPOL (acetic acid): 6.924, 14.276, 14.706 A 3 ˚ . 1434 A ˚ , 92.23, 91.70, 98.69 , cell VEVVIB (methyl acetate): 6.976, 14.313, 14.727 A 3 ˚ volume ¼ 1452 A . 1 : 1.5 clathrate inclusion complex of gossypol : 1.5(benzaldehyde). This has a triclinic ˚ , ¼ 73.62(1), ¼ 88.29(1), cell with a ¼ 10.959(2), b ¼ 11.418 (2), c ¼ 14.116(2) A 3 ˚ ¼ 87.73(1) , V ¼ 1693.0(5) A , space group P1, Z ¼ 2 (Gdaniec, Ibragimov and Talipov, 1991; cell is reduced). The dihedral angle between the naphthyl rings of the gossypol molecule is 86.1(3) . The two gossypol molecules in the unit cell are ˚ ) to form the familiar centrosymmetric hydrogen bonded (d(O(5) . . . O(3 0 ) ¼ 2.90 A dimers, and these are packed around two different kinds of centrosymmetric cage; there are no hydrogen bonds between dimers. The larger of the cages (T2) contains two benzaldehyde molecules related by a centre, and the second (T1) contains a single

B

C

0

B

A

C

0

A

Fig. 8.36. The diagram on the left shows the {gossypol : [CH2Cl2]} complex viewed down the axis of the tunnel which contains the halomethane guest. The diagram on the right shows the P3 gossypol polymorph obtained (as a single crystal) after loss of guest from the {gossypol[CH2Cl2]} complex. The tunnels shown as blank areas in fact contain atmospheric gases. (Adapted from Ibragimov, Talipov and Aripov, 1994.)

468


y x z

Fig. 8.37. Stereodiagram of gossypol1.5[benzaldehyde]. The hydrogen bonds are shown as dashed lines, oxygens are shown as circles, hydrogens have been omitted for clarity and the guests are emphasized. The cell in the figure is that given by Gdaniec et al. (1991), i.e. reduced but with nonstandard choice of axes. (Reproduced from Gdaniec, Ibragimov and Talipov, 1991.)

benzaldehyde in disordered array simulating a centre. The guests in the cages are hydrogen bonded to the hosts (Fig. 8.37). An alternative way of looking at this arrangement is to consider it as a cluster of two gossypols, forming a dimer. O(1) of gossypol is the donor in an hydrogen bond to the oxygen of one of the benzaldehydes in a T2 cage; the same arrangement is found on the other side of the dimer. This accounts for one benzaldehyde of the formula. The second benzaldehyde is hydrogen bonded to O(8) of gossypol, but only to one O(8) of a particular dimer, accounting for the remaining half benzaldehyde; this linkage takes up one of two alternative orientations. Overall, the second benzaldehyde is linked to one O(8) in half of the dimers and to its centrosymmetric congener in the other half, thus preserving composition and an average centre of symmetry. The clusters interact only by van der Waals forces. Analogous, but simpler, hydrogen-bonded clusters interacting by van der Waals forces are discussed in Chapter 12. 1 : 2 clathrate inclusion complex of gossypol : 2(salicylaldehyde). Triclinic crystals with a 1 : 1 ratio of gossypol to salicylaldehyde (2-hydroxybenzaldehyde) were obtained from neat salicylaldehyde while monoclinic crystals with a 1 : 2 ratio were obtained from salicylaldehyde/benzene mixtures. The structure of the triclinic 1 : 1 crystals (RIDNOH; Gdaniec, Talipov and Ibragimov, 1995) is not discussed here. The 1 : 2 monoclinic crystals (Gdaniec, 1991b; a ¼ 11.130(2), b ¼ 29.542(5), c ¼ 11.744(2) ˚ , ¼ 98.45(1), V ¼ 3820(1) A ˚ 3, space group P21/n, Z ¼ 4; JINFAN) have an interA esting structural relation to the gossypol : 1.5(benzaldehyde) hydrogen-bonded cluster. The gossypol molecules form the familiar centrosymmetric dimers; there is no hydrogen bonding between dimers. The hydroxyl oxygen of one of the salicylaldehyde guests (A in Fig. 8.38) is the acceptor in a bifurcated hydrogen bond from hydroxyl O(8) ˚ ), while the carbonyl oxygen of the second of one gossypol of a dimer (d ¼ 2.81 A salicylaldehyde guest (B in Fig. 8.38), is hydrogen bonded to O(1) of this gossypol, two orientations being found. This gives the 1 : 2 gossypol: salicylaldehyde ratio. The difference between gossypol : 1.5(benzaldehyde) and gossypol : 2(salicylaldehyde) is that

TRIS(5-ACETYL-3-THIENYL )M ETHANE (T AT M) AS HOST

469

B

A

Fig. 8.38. Diagram of one-half of the dimeric cluster of composition gossypol : 2[salicylaldehyde]. Hydrogen bonds are shown as dashed lines. The two crystallographicallyindependent guests are denoted as A and B respectively; B takes up two coplanar orientations in a 2 : 1 ratio. (Reproduced from Gdaniec, 1991b.)

both O(8)’s of the dimer in the latter are linked to salicylaldehydes, while only one of the two O(8)’s of the dimer is linked to benzaldehyde in the former. Disorder of orientation and arrangement complicates the situation.

8.6 Tris(5-acetyl-3-thienyl)methane (TATM) as host 8.6.1 Introduction When the host molecule is rigid, as has been the situation in most of the complexes considered previously, the possibilities of mutual host–guest adaptation are severely limited. However, the preparation of many inclusion complexes of the flexile (conformationally labile) host molecule tris(5-acetyl-3-thienyl)methane (TATM; 8.6), which has a tripod shape, and determination of a number of their crystal structures, now provides information (limited but indicative) about the mutual adaptation of a host and its guests (Herbstein, 1997a). Very few other examples of this type have yet been encountered.

8.6.2 Chemistry of TATM and its inclusion complexes TATM (C19H16O3S3; indexed in Chemical Abstracts under ‘‘Ethanone, 1,1 0 ,100 (methylidyntri-4,2-thiophendiyl) tris’’; CSD name ‘‘tris(5-acetyl-3-thienyl)methane’’) was first synthesized by Yakubov, Sudarushkin, Belenkii and Gold’farb (1973); these authors reported that the sublimed compound was an amorphous solid (m.pt. 50–58 C) which gave crystalline 2 : 1 adducts with benzene, ethanol and pyridine. An impressive list of inclusion complexes is due to Bin Din and Meth-Cohn (1977), who extended the earlier work and noted that ‘‘a systematic study has so far not revealed a solvent which is not

470


Scheme: formal representation of the TATM molecule, with those bonds emphasized about which torsion can occur. H1 attached to C1 is below the plane of the page and the apex C1 of the trigonal pyramid C1 C2 C8 C14 is pointing away from the observer. Positions of ring hydrogens have been distorted for convenience in drawing. The particular conformation illustrated here and in Fig. 8.39 is for Molecule A in the {TATM[1/2CCl4]} complex. The line formula of the Scheme is somewhat misleading as the molecule is not planar but has the tripod shape shown in Fig. 8.39. Torsion of five membered rings about the (H)CC bonds (t1, t2 and t3 are defined as H1–C1–C=C for each of the three rings) is one type of conformational variable and the second depends on whether the carbonyl oxygens are syn or anti to S in the rings; these torsion angles t4(S1–C4–C6–O1) and analogues (t5 and t6)) are all 0 or 180 . The conformations of the methyl groups, for which there is little experimental information, have not been included. C19H191,192,193

O3 C18

τ6

C16 S3

3

H17 C17

H3

C15 C14

H15

τ3 H11 C11 C8

H131,132,133C13 C12

H1 C3 τ1 1 C 1 C2 τ2

2 C 9 C10 S2 τ5

C5 H5 H9

S1 C4

τ4 C6

O1

C7H71,72,73

O2

8.6

incorporated’’ (my italics). This conclusion has been substantiated in later work (Sidhu and Ripmeester, 2001). Bin Din and Meth-Cohn (1977) made a variety of TATM analogs by replacement of the acetyl groups with functionalities such as OCOCH3; only the methyl ester showed any promise of clathration ability, while attempts at resolving a racemic mixture of 2-butanol with TATM were unsuccessful. A correlation of the thenavailable experimental data (crystal structures, TATM conformations) for the TATM inclusion complexes has been made (Herbstein 1997a) and this analysis is now brought up to date. The guests are collected together, classified in terms of host : guest ratio and chemical nature of the guest, in Table 8.10. Only binary TATM systems are considered here. It would be interesting to see if TATM could act as an acceptor for guests with hydrogen bonding capabilities; this has so far only been done for TATM2H2O.

8.6.3

Conformations taken up by the TATM molecule in the various crystallographic structure types

The conformation of the TATM molecule in various crystalline inclusion complexes provides one potential source for classification; another is the variety of crystal structures


471

Table 8.10. Crystalline {(TATM)x[Guest]y} inclusion complexes, classified according to host– guest ratio and chemical nature of the guest. TATM host–guest inclusion complexes have so far been reported in the literature with more than 60 different types of guest; nine different kinds of crystal structure have been reported, with 12 different guests. Bold type denotes complexes for which crystal structures have been reported. The dynamics of the (deuterated) guests has been studied by NMR for many of these complexes (Sidhu et al., 1996) Host–Guest ratio 1:2 water 1:1 cyclo-nonanone 2:1 C6H5R with R ¼ F, Cl, Br, I, CH3, C2H5, CH(CH3)2 (cumene); o-, m- and p-xylene, mesitylene benzene, naphthalene cyclohexane (2 polymorphs), cycloheptane, cycloo¨ctane, cycloo¨ctene methanol, ethanol, iso-propanol, sec-butanol, tert-butanol, n-decanol acetone, methyl ethyl ketone, methyl phenyl ketone methyl phenyl ether halocyclohexanes (X ¼ F, Cl, Br), CH2X2 (X ¼ Cl, Br, I), CHCl3, CCl4, 1,3-dichloropropane (5 polymorphs), also 1,3-dihalopropane (halo ¼ F, Br), 1,2-dichloroethane (2 polymorphs), 1,4-dichlorobutane pyridine, piperidine, triethylamine ethyl acetate, methyl cyanide, acetic acid, nitromethane, dimethyl sulphoxide, HCON(CH3)2, decalin, acetonitrile, dimethylformamide. 3:1 t-Butylbenzene, n-hexane, 1,5-dichloropentane, 1,6-dichlorohexane 4:1 1,8-dichlorooctane, 1,9-dichlorononane, 1,10-dichlorodecane Notes: Preparation of complexes has been reported by Yakubov, Sudarushkin, Belenkii and Gold’farb (1973), Bin Din and Meth-Cohn (1977). Roos and Dillen, (1992), Pang and Brisse (1994b), Pang and Brisse (1994a), Sidhu and Ripmeester (2001, 2003), Sidhu, Enright, Udachin and Ripmeester (2004).

(see next section) and whether these are isomorphous, isostructural or indifferent. Analysis should hopefully lead to uncovering the relation that must exist between these two structural aspects, but this largely remains to be achieved. Structural information is available for 30 crystallographically independent TATM molecules in ten different types of crystal structure (including the dihydrate). Specifically, the following questions can be asked, and answered, from an analysis of the published crystallographic results: 1. Is there a correlation between the TATM conformer and the crystal structure of the inclusion complexes, i.e. is the same TATM conformer found in all the members of a group of isomorphous crystals? The answer is ‘Yes’. 2. If so, then which TATM conformers occur in the various groups of isomorphous inclusion complexes, and what is the range of TATM molecular structural variation

472


Fig. 8.39. Ball and stick model of TATM molecule as found in its CCl4 complex (only one of the two independent molecules in the asymmetric unit is shown). The view is along the methine C–H bond, the hydrogen being behind the plane of the page. The hydrogens have been inserted in calculated positions. Notice that two oxygens (of C¼O groups) are syn to S and one anti.

3.

(specifically torsion angles) found within the several groups? This is answered below. What is the range of structural variation when there is more than one TATM molecule in the crystallographic asymmetric unit? This is answered below.

Potentially one can also hope to study how the nature of the guest influences the conformation taken up by the host, but this is not generally possible at present because the guests are disordered in most of the crystal structures that have been determined. Repetition and extension of the crystal structure determinations at low temperatures should be rewarding. The crystals included in Table 8.9 are all racemic and so absolute optical configurations are not required; nevertheless, meaningful comparison of the geometrical structures taken


473

up by TATM in its various inclusion complexes requires comparison of conformers of the same optical configuration. The enantiomer chosen (arbitrarily) for intercomparison has been drawn (see Scheme) so that H1 (attached to C1) is below the plane of the page and the apex C1 of the trigonal pyramid C1C2C8C14 points away from the observer. The torsion angles 1, 2 and 3 are defined (again arbitrarily but consistently) as 1 ¼ (H1C1C2¼C3) (C3 linked to S1), and correspondingly for 2 and 3. The TATM molecule will be achiral only for some special values of the torsion angles 1, 2 and 3; as these were not found, the TATM molecule is chiral. The torsion angle with the smallest absolute value has been taken as 1 (it is found that j 1j0 ) and rings 2 and 3 follow in clockwise sequence with the enantiomer oriented as described above. Further distinction among conformers depends on whether the carbonyl oxygens are syn (designated S) or anti (designated A) to sulfur in the rings; experimentally it is found that the torsion angles 4(S1–C4–C6 ¼ O1) and analogs, ( 5 and 6) are either 0 (syn) or 180 (anti). The results show that when there is more than one molecule in the asymmetric unit (i.e. in crystallographic groups III and IV), then analogous torsion angles in the crystallographically independent molecules do not differ by more than 10 (answer to x3 above). Also, analogous torsion angles within a group of isomorphous structures (i.e. in crystallographic groups III, IV, VII and VIII, and in TATM2H2O)) do not differ by more than 10 (answer to x2 above). Representative values for three specific inclusion complexes are given in Table 8.11. In fact, four conformations have so far been found for the TATM molecule considered as an entity: (a) Conformation 1, with 1 0 , 2 105 , 3 160 and SSS for the three acetyl groups (C1–SSS). The guests are ethyl acetate, ethanol, cyclohexane (polymorph A in Group II), 1,3-dichloropropane (polymorphs 1, 3, 4 and 5), 1,2-dichloroethane (polymorphs 1 and 2) and n-hexane. (b) Conformation 2, with 1 0 , 2 130 , 3 150 and SSA for the three acetyl groups (C2–SSA). The guests are benzene, CCl4, cyclohexane (polymorph B in Group III), cycloheptane, cycloo¨ctane and 1,3-dichloropropane (polymorph 2).

Table 8.11. Torsion angles (degrees) in the TATM host molecule as found in some of its inclusion complexes Inclusion complex

1

2

3

4

5

6

Conformation of carbonyl with respect to ring sulfur

Conformation 1 TATM0.5(ethyl acetate) 7.0 102.9 157.0 2.2 1.1 0.6 SSS Conformation 2 TATM0.5(benzene) 16.2 126.2 148.2 3.5 1.3 178.9 SSA Mol. A Conformation 3 TATMcyclononanone 7.3 101.8 15.2 0.8 10.3 4.3 SSS at 220K.

474


Conformation 3, with 1 0 , 2 100 , 3 15 and SSS for the three acetyl groups (C3–SSS). The guest is cyclononanone. (d) Conformation 4, with 1 0 , 2 24 , 3 100 and SSS for the three acetyl groups C4–SSS). This is the conformation found in TATM2H2O.

(c)

The overall conformation of the TATM molecule is determined by the values of the ring torsion angles 1, 2 and 3 (presumably the major factor), and by the values of the torsion angles 4, 5 and 6 of the acetyl group. We had earlier contended (Herbstein, 1997a) that C1 and C2 belong in the same well of the map of potential energy as a function of 2 and 3, and thus refer to one set of ring conformations. The additional examples that have since accrued suggest that C1 and C2 are indeed separate, and this applies with greater force to the overall conformations C1–SSS and C2–SSA. Syn and anti conformations of carbonyl O with respect to S appear, from the statistics of occurrence, to be of approximately equal energy but it is not known what factors govern the appearance of one or other conformation in a particular molecule, nor is it known why the other possible conformations of the acetyl group (SAS, ASS, AAA, AAS, ASA and SAA) do not appear. A molecular mechanics study is sorely lacking. 8.6.4

Crystallography of the inclusion complexes of TATM

Ten different kinds of crystal structures have been reported in the literature, with twelve different guests (Table 8.12). The shape of one example of the TATM molecule is shown in Fig. 8.39. The terms ‘‘clathrate,’’ ‘‘cage’’ and ‘‘tunnel’’ have been used to describe different varieties of these complexes. Structure analyses of the ethyl acetate complex (representative of our Group I) by van Rooyen and Roos (1991a), and of the polymorph A and polymorph B complexes of cyclohexane (representatives of our Groups II and III) by Pang and Brisse (1994b), show clearly that these are all tunnel complexes, with the linear tunnel axis of Groups I and II along the shortest cell dimension and the zigzag tunnel axis of Group III along the longest. The cyclohexane guest of the Group II (polymorph A) complex is located about a center of symmetry and is ordered; the cyclohexane of the Group III (polymorph B) complex is at a general position and takes up three orientations in 40 : 35 : 25 ratio. The cycloheptane and cycloo¨ctane molecules of the Group III complexes take up two orientations in approximate 2 : 1 ratio. Further distinctions can sometimes be made within a group – for example, the benzene and CCl4complexes are both in Group III, but the benzene molecules are ordered (only the second example of ordering of guests among the published crystal structures) while the CCl4 molecules have partially disordered arrangements in which one Cl atom is ordered and the three remaining Cl atoms are trigonally disordered in two orientations with 2 : 1 ratio. It is perhaps not surprising that the three alicyclic guests cyclohexane (Group III – polymorph B), cycloheptane and cycloo¨ctane form isomorphous crystals but the common features causing the benzene and CCl4 complexes to crystallize in this structure type are not clear. The structures of the {TATM0.5(1,3-trichloropropane)} polymorphs can be described in somewhat similar terms (Sidhu, Enright et al., 2005) but details of the resemblances are often obscure. The triclinic hexane complex (Group IV) and the monoclinic cyclononanone complex (Group V) are tunnel inclusion complexes. The first has the tunnel axis along [111]


475

˚ , degrees, A ˚ 3) for some inclusion complexes of TATM. Table 8.12 Crystallographic data (A Compositions are expressed as {TATM[n(guest)]}. Triclinic cells have been reduced. S.u.’s of cell ˚ , of angles 0.01–0.04 and of cell volumes 0.4–2 A ˚ 3. Analyses lengths are 0.002–0.006 A were at room temperature unless stated otherwise b

c

Cell volume

11.229 10.907

12.329 12.306

99.05 97.23

106.43 106.63

98.42 96.87

1057 1012

10.372 10.194

12.488 12.795

81.49 79.09

71.26 72.74

84.95 84.89

1010 1054

13.560 13.684 13.729

14.197 14.235 14.227

89.68 88.90 89.40

76.60 77.46 77.15

75.50 78.77 76.24

2089 2170 2157

13.734 14.013

14.177 13.986

89.22 89.46

76.84 77.15

76.01 75.22

2154 2198

12.647 12.684

12.694 12.704

20.604 85.89 20.572 102.71

74.31 103.95

86.36 91.35

3173 3128

10.994

19.464

13.417

90

109.4

90

2708

12.690

12.743

13.738

90

109.5

90

2100

12.715

12.723

13.893

90

109.97

90

2113

Group VII: P2/n, Z ¼ 8 0.5(1,3-dichloropropane), 173K, polymorph 2; SEUR05

14.319

13.499

22.328

90

104.08

90

4186

Group VIII: P2/n, Z ¼ 8 0.5(1,3-dichloropropane), 173K, polymorph 3; SEUR05

12.731

12.762

25.937

90

93.88

90

4204

Group IX: C2/c, Z ¼ 8 0.5(1,3-dichloropropane), 173K, polymorph 5; SEUR05

17.500

18.465

14.982

90

119.43

90

4217

n (guest), Refcode, reference

a

Group I: tunnel axis along [100], P1, Z ¼ 2 0.5(ethyl acetate) JIYVIW; RR91a 8.229 0.5 (1,2-dichloroethane) 173K, 8.033 polymorph 1; SERP02 Group II: tunnel axis along [100], P1, Z ¼ 2 0.5(ethanol) VUJZOP; DR92 0.5(cyclohexane), polymorph A at 220K; YIRVAW; PB94b

8.335 8.622

Group III: tunnel axis along [001], P1, Z ¼ 4 0.5(benzene) JIZFED; RR91b 11.538 11.638 0.5(CCl4) LEKREY; PHW94 0.5(cyclohexane), polymorph B 11.668 YIRVAW01; PB94b 0.5(cycloheptane) at 220K [c] YIRVIE 11.721 0.5(cyclooctane) at 220K YIRVOK 11.914 , Z ¼ 6 Group IV a and b: P1 a. 1/3(n-hexane) KUGZAN; RD92 b. 0.5(1,3-dichloropropane), 173K, polymorph 1; SEUR05 Group V: P21/c, Z ¼ 4 cyclononanone at 220K WIKCOI; PB94a Group VI: P21/c, Z ¼ 4 0.5(1,2-dichloroethane), 173K, polymorph 2; SERP02 0.5(1,3-dichloropropane), 173K, polymorph 4; SEUR05

Notes: (1) Space groups for the ethanol and ethyl acetate complexes were originally given as P1, but Marsh (1994) has corrected these to P 1 (see VUJZOP01 and JIYVIW01); (2) the complexes of Groups I and II are not isomorphous; however, there are resemblances between the two structure types, and both have tunnels along [100]. Similar remarks apply to Groups IVa and IVb. References: DR92 – Dillen and Roos (1992); PB94a – Pang and Brisse (1994a); PB94b – Pang and Brisse (1994b); PHW94 – Pang, Hynes and Whitehead (1994); RD92 Roos and Dillen (1992); RR91a – Rooyen and Roos (1991a); RR91b – Rooyen and Roos (1991b);SERP02 Sidhu, Enright et al., 2002; SEUR05 – Sidhu, Enright et al., 2005.


476

(possible disorder not mentioned) and the second along [100] (guest disordered over two orientations in 75 : 25 ratio). ˚ , 88.98, 89.57, 80.70 , P1, The TATM.2H2O complex (triclinic, 8.761, 10.793, 21.794 A Z ¼ 4) has two (ordered) waters as part of the framework and two (disordered) waters in tunnels.

8.6.5

Formation of the inclusion complexes

The different overall conformations of TATM inferred above will occur in TATM solutions in relative proportions given by the Boltzmann distribution; these proportions will depend on temperature but not on the nature of the solvent. The solution will become supersaturated on cooling and the appropriate inclusion complex will begin to crystallize. For example, if the solvent is benzene then {TATM[1/2(benzene)]} will crystallize and the solution will become depleted in molecules with Conformation C2– SSA; the Boltzmann distribution will be continuously re-established by conversion of molecules with the other conformations and this process will continue until the appropriate amount (determined by its solubility) of crystalline inclusion complex has been formed. What happens with cyclohexane as solvent? Here it is presumed that nuclei of cyclohexane with the differently shaped conformers C1–SSS and C2–SSA will be formed and these will give rise to the two polymorphic forms of {TATM[1/2C6H12]} designated above as Polymorph A of Group II and Polymorph B of Group III. Pang and Brisse (1994b) report that Polymorph B was obtained by relatively rapid cooling (1 C/h) and A by slow cooling (5 C/day); this suggests that B is metastable with respect to A, in ˚ 3). Presumably accordance with the respective volumes/formula unit (B ¼ 539 A Ostwald’s rule of successive reactions (e.g. Findlay, Campbell and Smith, 1951) applies to this system but this has not been explicitly stated. In terms of this description the overall process of formation of a crystalline inclusion complex can be divided into three stages: 1.

2. 3.

Selection of appropriate TATM conformer from the ensemble of conformers, requiring recognition between solvent molecule and the appropriate TATM conformer, i.e. the determining factor is host-guest interaction. Formation of nuclei from the TATM-solvent aggregates, requiring predominantly host conformer–host conformer recognition. Growth of nuclei to form crystals.

Thus the formation of crystalline inclusion complexes containing TATM molecules of different conformations (depending on the solvent partner) does not present any conceptual difficulties. This description, with the guest (solvent) molecule plucking out the appropriate TATM conformer from the Boltzmann distribution, is the converse of that given when a rigid host is used in selectivity experiments with a solution of, or containing, a mixture of guests. This description also provides a possible explanation for the occurrence of an amorphous sublimate; the presence of more than one conformer in the condensing sublimate prevents crystallization.


477

8.6.6 Dynamics of guest molecules in the complexes Deuterium NMR (labeling of guests) has been applied to these systems to study guest dynamics down to 112K (Sidhu et al., 1996). There is sixfold rotation of benzene in its TATM complex, with an activation energy of 4.1(4) kJ/mol, and no minimum in relaxation time down to 112K. These results are compatible with the crystal structure noted above. o-Xylene and p-xylene are rigidly held in their TATM complexes, with rapid rotation of the methyl groups. Dimethylsulfoxide is also held rigidly, but with rapid rotation of the methyl groups. Mesitylene occupies two sites in the TATM structure, but the motion of the methyl groups was not established. Acetonitrile and nitromethane both have a precession motion, with an activation energy of 11.5(5) kJ/mol. The behavior of the deuterated guest in triclinic {TATM0.5(1,2-dichloroethane)} was also studied (Sidhu, Enright et al., 2002). Correlation of NMR and XRD measurements led to a model in which the trans conformer of 1,2-dichloroethane performs 180 flips through the centre of symmetry of the guest and about an axis perpendicular to the Cl–C–C–Cl plane. 8.6.7 Other examples The host–guest complexes of E,E-1-[ p-dimethylamino-phenyl]-5-[o-hydroxy-phenyl]penta-1,4-dien-3-one (the Heilbron complexes (Herbstein, Kapon, Reisner and Rubin, 1984); see Section 8.4) are an example of selection by the guest of conformationallydistinct states of the host from solution, as shown by the fact that the host molecules have the s-trans, trans conformation in some complexes and the s-cis, trans in others; however, the number of such complexes of known crystal structure is limited, and generalization is not yet possible. 8.6.8 Summary Analysis of the published crystallographic data shows that the TATM molecules in a particular group of isomorphous host–guest inclusion complexes (with different guests) all have the same conformation, with numerical values of ring torsion angles not differing by more than 10 within the group of crystals. The same holds for comparisons between different host molecules in a particular inclusion complex when there is more than one molecule in the crystallographic asymmetric unit. This could be taken to imply that there is a 1 : 1 correlation between ring conformations and crystal structure but this is an oversimplification. The situation is complicated by the fact that, while the ring conformations are the primary factor in determining the energy of the TATM molecule, the conformations of the three acetyl groups (syn or anti relation of carbonyl oxygen to ring sulphur) are an essential component in determining the overall molecular shape, which is more important than ring conformation in determining the crystal structure of the inclusion complex. Thus the correlation is between overall molecular conformation (i.e. the combination of the conformations of rings and acetyl groups) and crystal structure. The syn and anti conformations of the acetyl groups appear to have similar energies from the statistics of their occurrence but only two of the possible combinations of ring conformation–acetyl group conformation have so far been encountered. It is not clear


478

whether this is because the available sample is unrepresentative of the population as a whole or because other factors contributing to overall molecular energy have not yet been recognized. 8.7

(5,10,15,20)-Tetraphenylmetalloporphyrins and complexes

8.7.1

Introduction

, , , (or 5,10,15,20)-Tetraphenylmetalloporphyrins4 form very many molecular complexes (the , , , notation is due to Fischer and the 5,10,15,20 notation is that recommended by IUPAC (Smith, 1984)); TPP is the abbreviation used for the dianion of 5,10,15,20-tetraphenylporphyrin itself and thus the metalloporphyrins are designated as (TPP–M2þ) (with obvious extensions for metal ions of different charge); the formula of (TPP–M2þ) is C44H28N4M. Many substances listed as ‘‘complexes,’’ are either coordination complexes or inclusion complexes comprised of coordination complexes with non coordinated guests; sometimes the same molecule behaves as a ligand and as a guest in an inclusion complex. (TPP–M2þ) molecules, often as ‘‘complexes,’’ have been extensively investigated from many points of view, including comprehensive crystallographic studies (Scheidt and Lee, 1987); the November 2002 Version 5.24 of the CSD (272066 entries) gave some 600 entries for TPP complexes and analogs. Many crystal structures have been reported using the conventional approach to their description, but the most substantial corpus of material is that accumulated by Strouse and his coworkers over almost a decade. The Strouse group has determined the structures of some 300 ‘‘complexes’’ (of all kinds), many at low temperatures (mostly around 100K but two at 15K). The extent of this contribution is shown by the following papers relevant to the content of this chapter: 1.

2.

3.

4.

Byrn, Curtis, Hsiou, Khan, Sawin, Tsurumi and Strouse, 1990 (B90); cell dimensions were given for 65 isostructural TPP complexes, and 52 packing diagrams were shown. Byrn, Curtis, Goldberg, Hsiou, Khan, Sawin, Tendik and Strouse, 1991 (B91): Crystal data were reported for 45 complexes, and ‘‘a detailed analysis [was made] of the molecular packing in over 100 TPP-based clathrates;’’ 56 packing diagrams were shown. Byrn, Curtis, Hsiou, Khan, Sawin, Tendik, Terzis and Strouse, 1993 (B93). Crystal data were provided for 75 new porphyrin-based clathrates, and the molecular packing was analyzed in over 200 tetraarylporphyrin-based lattice clathrates; some 150 packing diagrams were shown. Byrn, Curtis, Hsiou, Khan, Sawin, Terzis and Strouse (1996); this is an extensive summary review of earlier material.

By 1993, the structures of some 480 ‘‘porphyrin-based lattice clathrates’’ had been reported (Table 8.13, taken from Table I of B93). 4 In more general terms the host materials are tetraarylmetalloporphyrins (TAP), where the aryl groups may be substituted. We consider only the neutral (TPP–M2þ) species, but the reader should be aware of extensive additional material not covered here for reasons of space. There are related series where the host is TPP– Mnþ(n ¼ 3–6) and appropriate counterions. Examples of Mnþ are Mn3þ, Fe3þ, Co3þ, Rh3þ, Au3þ, Sn4þ, Ce4þ, U4þ, Nb5þ, Mo6þ, W6þ; there also some examples with Co, Fe, Feþ.

(5,10, 15,20)-TETRAPHENYLME TALLOPORPHYRINS AND COMPL EXES

479

Table 8.13. Crystallographic symmetry of porphyrin-based lattice clathrates for which X-ray structural data are available Crystal symmetry

Z

number

Crystal symmetry

Z

number

triclinic

1 2

139 81

orthorhombic

monoclinic

2 4 8

56 132 10

2 4 8

1 18 8

tetragonal

2 4 8

20 9 3

rhombohedral

6

3

B93 have commented that ‘‘Although the TPP molecule lacks any functionality that might be expected to dictate the relative disposition of neighboring molecules, one finds that the host structure is strongly conserved in porphyrin-based clathrates.’’ Perhaps the most remarkable features of the coordination complexes are the variety of metals that can be incorporated in the TPP macrocycle, and the catholic nature of the ligands/guests. The M2þ species include M ¼ 2H, Mg, Mn, Fe, Co, Ni, Cu, Zn, Mo, Ru, Pd, Ag, Cd, Sn. The ligands (actual or potential) include heterocyclics (pyridine, piperidine, 4-picoline, acridine, isoquinoline, 1,4-dioxane, trioxane), aromatic esters (methylbenzoate, tetramethylpyromellitate), quinones (anthrone, 9-xanthone, bianthrone), aldehydes (benzaldehyde, also with various substituents), aliphatic alcohols (2-propanol, hexanol, 6methyl-5-hepten-2-ol) and phenols. The guests include aromatic hydrocarbons (benzene, anthracene, phenanthrene, 1,2-benzanthracene, 2,3-benzofluorene, coronene) and substituted aromatics (halobenzenes, toluene, the xylenes, 1,2,4-trimethylbenzene, mesitylene, ethylbenzene, styrene, phenylacetylene). Although most of the guests have an aromatic functionality, this is not a requirement and the impression is that the hosts are very versatile indeed and able to adapt their packing arrangement in the crystal to the shape of the guests and, perhaps, to particular types of host–guest interaction. Some of the compounds noted above as ligands also appear as guests. Classification of the second component as ligand and/or guest is reconsidered at the end of this section after surveying the structural results. Discussion of the structural chemistry of the metalloporphyrin complexes (using ‘‘complex’’ in its most general sense) raises a number of issues, some formal, as in their classification, and some actual, as in their structure and behavior. The point of view taken in this book (see Chapter 2) is that binary adducts (alternatively, molecular complexes and compounds) are ‘‘crystalline two-component phases in which the properties of the components are very largely conserved.’’ This implies that there are no covalent bonds between the components and is shown in its purest form in Chapter 10 (‘‘Packing Complexes’’). However, this pure concept breaks down in Chapter 12 (‘‘Hydrogen Bonded Molecular Compounds and Complexes’’), and Chapter 11 (‘‘Donor–Acceptor Molecular Compounds (Essentially Localized Interactions)’’), where we admit the importance of partial bonds between the components. A similar process occurs in the metalloporphyrin complexes, where there is bonding (of varying strength) between metal

480


and ligand in the five- and six-coordinate species (the four-coordinate species do not have ligands); it is only when guests are present that one has inclusion complexes in the strict sense. However, the structural resemblances across the whole group are so strong that it is logical to treat all these species together, but only after they have been defined separately. The structural consequences have been neatly summarized by B90; p. 1871: . . . the materials . . . include 4-, 5- and 6-coordinate species, all of which exhibit the same porphyrin framework. In the structures of the 4-coordinate materials, solvate molecules occupy the open channels. With the 6-coordinate materials, the axial ligands occupy the channels. In both cases the metal atom resides on a crystallographic inversion centre. [This applies specifically to the triclinic (Z ¼ 1) and some monoclinic crystals (FHH).] In the 5-coordinate materials, the single axial ligand and one solvent molecule occupy alternate sites in the channel. . . . to a large extent the packing is controlled by the centrosymmetric TPP molecule . . .

Strouse and coworkers have found it convenient to describe the large number of structures that they have investigated in terms of non-standard body-centred triclinic cells containing 2 formula units per cell, and unusual choices of monoclinic angles for some monoclinic crystals. The host–guest relationship is then illustrated using a ‘‘lattice section’’ through these cells, generally in the plane of the porphyrin ring. Lattice sections for {(TPP–Mn2þ)(toluene)2} and {(TPP–Mg2þ)(4-picoline)2} are compared in Fig. 8.53, cell dimensions being given in Table 8.22. The problem with this approach is that, using lattice sections alone, it is often difficult to be certain whether a particular material is a coordination complex, or an inclusion complex based on a combination of coordination complex and included guest. A definitive decision can only be made on the basis of the complete structure analysis, usually determined by the Strouse group and available from the CSD, but seldom directly from the four papers noted above. We have preferred, in general, to follow most earlier authors and use a conventional crystallographic approach, emphasizing the advantages of using reduced cells for describing triclinic crystals, and have given standard descriptions for the other systems. We shall compare the two approaches for a few examples. We give many cross-references to the Strouse papers, noting, but not using, their classifications. Each group of related crystal structures deserves detailed analysis, and individual attention should also be given to exceptional examples. This has been done here but only partially; perhaps a full treatment will attract a dedicated author. Although these ‘complexes’ were described in the earlier studies as ‘‘extremely air sensitive,’’ it appears that decomposition was generally due to loss of guest (or solvent) molecules and later workers have grown crystals from solution and carried out diffraction studies without special precautions beyond enclosing crystals in capillaries or a plastic coating. 8.7.2

Crystallography of (5,10,15,20)-tetraphenylmetalloporphyrin coordination complexes

8.7.2.1 Introduction We start by considering the crystallography of the {host TPP-ligand} species as four-, five- and six-coordinate coordination complexes, organizing the crystal structures into isomorphous or isostructural groups on the basis of published crystallographic results. We then extend this treatment to the inclusion complexes, followed by some cross-correlation. Most of the available results are for triclinic crystals. The reader is reminded that,


481

although a triclinic unit cell can be chosen in many ways, it is only the ‘Niggli reduced cell’ that is unambiguously defined.5 Thus we have quoted results for reduced cells, making transformations where necessary. There are two types of triclinic cell, Type (I) with the origin chosen so that the three interaxial angles are acute, and Type (II), with the three interaxial angles obtuse; for more detail and references see Herbstein (1997b). The two types are not interconvertible. We have followed the convention that a < b < c where possible. However, many structure analyzes are reported in terms of nonconventional cell choices; in order to avoid confusion, we have followed the original authors in their choice of cell unless stated otherwise. Original estimates of precision are used when required (not often), even if these may sometimes appear to be over optimistic. The crystallographic classification is initially based on cell dimensions and space groups. These provide a necessary, but not sufficient, basis for classification because molecular arrangements are appreciably, but not entirely, determined by the packing of the (TPP–M2þ) moiety. Thus it is essential to examine the crystal structure to determine the mode of interaction between the (TPP–M2þ) moiety and the second component, i.e. is this a ligand linked to the metal, giving a coordination complex or a guest, without special linkage to the metal, giving an inclusion complex? Strouse et al. have called many of their structures ‘‘clathrates;’’ the CSD have distinguished between coordination metalloomplexes (explicit metal–ligand linkages) and ‘‘clathrates,’’ which encompasses both weaker metal–ligand linkages and inclusion complexes. We have treated the two latter groups separately. One must also note that the second components are disordered in some crystals and thus clearcut classification is not always possible. A number of variables have to be taken into account. One is the nature of the metal ion; we restrict ourselves to uncharged hosts so the metal ion is formally M2þ. Another variable is the nature of the second component, which can be a ligand, thus giving a coordination complex, or the guest, in an inclusion complex. Some of the coordination complexes can also contain guests, so structural classification can be quite complicated, especially when based on unit cell dimensions rather than full crystal structures. Much of our information comes from the Cambridge Structural Datafile (CSD; Version 5.24 of November, 2002; 272 066 entries), where reduced cells and packing diagrams are given. Atomic coordinates are available for most entries, thus enabling re-examination of brieflyreported structures. The metal–ligand distances (exemplified for Zn) fall (approximately) ˚ , and weaker with d(Zn–O/ into two groups – stronger binding with d(Zn–O/N) 2.2 A ˚ . We classify both as coordination complexes; longer distances imply inclusion N) 2.5 A rather than coordination. We first discuss the true coordination complexes. 8.7.2.2

The four-coordinate coordination complexes

The four-coordinate (TPP–M2þ) neat host molecules (i.e. without ligands or guest molecules) crystallize in three isomorphous groups of crystals: Tetragonal crystals for M ¼ Ni, Cu, Co (Madura and Scheidt, 1976), Pt, Pd, 2H (Hamor, Hamor and Hoard, 1964, TPHPOR10; Stone and Fleischer, 1968; 5 Reduced cells can be defined in a number of ways; we choose not to elaborate but refer the reader to authoritative sources such as the International Tables for Crystallography.

482


˚ , c: 13.89–14.04 A ˚ , space group I 42d (No. TPHPOR1X), Fe with a: 15.04–15.13 A 3 ˚ 122), Z ¼ 4, volume per formula unit : 788–798 A , molecular symmetry S4 4 (see Table XXVIa of Scheidt and Lee (1987) for details and references). Triclinic crystals for M ¼ 2H (Barker, Stanley and Fronczek, 2002, TPHPOR11), Zn, Ag, Cr, Mn; A full range of solid solutions is formed between TPP–Ag2þ and TPP–2Hþ (Donnay and Storm, 1967) with crystal data (Z ¼ 1, space group P1) given below. The three triclinic (TPP–M2þ) four-coordinate coordination complexes are ˚ , deg., A ˚ 3) are for reduced cells, but with nonisomorphous. The cell dimensions (A standard choice of origin.

Mnþ Reference; Refcode

a

b

c

Cell volume

2Hþ; SL67; TPHPOR01 Zn2þ; SME86; ZZZTAY02 Ag2þ; SME86; DOWRAI

10.420 10.382 10.503

12.410 12.421 12.485

6.440 6.443 6.351

99.14 98.30 97.73

101.12 101.15 100.68

96.06 96.47 97.15

799 798 801

References: SL67 – Silver and Tulinsky, 1967; SME86 – Scheidt, Mondal, Eigenbrot et al., 1986.

Triclinic crystals for M ¼ Cd. The reduced cell has a ¼ 10.096, b ¼ 12.446, ˚ 3, ˚ , ¼ 79.95, ¼ 75.81, ¼ 81.24 , volume per formula unit 800.6 A c ¼ 13.438 A space group P1, Z ¼ 2 (Hazell, 1986).

Only (TPP–2Hþ) is common to the first two groups. As the volumes per formula unit for ˚ 3 (triclinic)), this ˚ 3 (tetragonal) and 802 A the two polymorphs are not very different (792 A datum gives only a weak indication that the tetragonal crystals are the more stable. The volume occupied by the (TPP–M2þ) host molecule (in the unit cell) can be taken as ˚ 3to a first approximation, irrespective of the nature of the metal ion. An exception is 800 A ˚ 3 in its triclinic perhaps provided by neat (TPP–Zn2þ), which has a molar volume of 798 A 3 ˚ in a monoclinic form produced by high temperature crystallization crystals, but 841 A from 1,3,5-triisopropylbenzene (B93); a molar volume difference of 5% for polymorphs is unusual but not unprecedented. In the tetragonal crystals (which are not centrosymmetric but are racemic) the non4 symmetry with the pyrrole rings planar host porphyrin macrocycle molecule has S4 ˚ off tilted at 12 to (001). The carbons linked to the phenyls are displaced by 0.4 A (001). The planes of the phenyl rings, which librate about the C–C bonds with an amplitude of 9 , are tilted at 80 to (001). The structure consists of layers of molecules ˚ , mutually shifted in (or about) the (001) plane; succeeding layers are separated by 3.5 A by a/4 and are related by a two fold axis at z ¼ 1/8 along [100]. The triclinic (Z ¼ 1) crystals have a structure with crumpled layers about (010); all (TPP–M2þ) host molecules are translationally equivalent and the macrocycle is only slightly distorted from planarity. The ring dihedral angles have values of 90–60 in the triclinic (Z ¼ 1) group. Thus the tetragonal and triclinic (Z ¼ 1) groups can be described as having layer structures, although of rather different kinds. The metalloporphyrin macrocycles are appreciably distorted from planarity in the tetragonal polymorphs but hardly at all in the triclinic (Z ¼ 1) polymorphs. The roles of intramolecular (the nature of the metal atom) and intermolecular interactions (‘‘packing effects’’) in determining the relative stabilities of


483

the two structural groups do not yet appear to have been sorted out. This is important for understanding the thermodynamics of the formation of the molecular complexes and the nature of the products of their thermal decomposition. (TPP–Cd2þ) has a weak dimer structure and so is probably not directly comparable with the other structures. 8.7.2.3

The five-coordinate coordination complexes

The entries in Table 8.14 have been arranged in groups of isomorphous or isostructural crystals. The cell dimensions of the first pair of entries suggest that these two crystals are closely isomorphous. It is clear from Fig. 8.40 that {(TPP–Zn2þ) 9-xanthone} (HALWAS) is a five-coordinate coordination complex and not an inclusion complex. The five-coordinate coordination complexes have common geometrical features. The four links from the metal atom to the equatorial nitrogens are very similar; in HALWAS they ˚ . The metal atom is displaced from the plane through are 2.055 (twice), 2.058 and 2.063 A ˚ . The angle ¼ Neq–M–Xax the equatorial nitrogens towards the ligand by here 0.218 A is 98.28 ; there are four such angles but we shall only give one value. The Zn–O distance ˚ in the 9-xanthone complex is appreciably shorter than the values found in the of 2.205 A bianthrone and dibenzosuberone complexes (see below) and attests to stronger bonding ˚ ), in accord here. The anthrone complex has a very similar structure (d(Zn–O) ¼ 2.226 A with the appearance of Fig. 26 of B93, which has the caption ‘‘Stage 2 ‘Hybrid’ clathrates.’’ We have placed the entries for 3-nitroaniline and 5-octanoic lactone next because of similarities in axial lengths despite differences in interaxial angles. There is some similarity between the lattice sections for these two complexes in Fig. 21 of B93. The crystal structure of {(TPP–Zn2þ)(3-nitroaniline)} has been obtained from HAMLAI. It is a five-coordinate coordination complex with Zn linked to the amino group of ˚ , while the distances to the equatorial nitrogens are 3-nitroaniline (d(Zn–N) ¼ 2.314 A ˚ ˚ . Perhaps surprisingly, the bifunctional 1.967, 2.018, 2.046, 2.058 A); ¼ 0.244 A 2,5-hexanedione ligand has only one of its carbonyl groups linked to Zn (crystal structure ˚ ; d(Zn–N) ¼ 2.041, 2.056, 2.059, 2.088 A ˚; from HAMGUX), with d(Zn–O) ¼ 2.261 A ˚ ¼ 0.188 A and ¼ 99.37 . The reduced cells show that the next four entries are isostructural. The pair Grandlure III/IV and II6 are closely isomorphous, despite the differences in the chemical formulae of the guests; the lattice sections shown in Fig. 19 of B93 (captioned ‘‘ ‘normal’ stage 2 clathrates’’) are similar. The Grandlure I clathrate (lattice section in Fig. 17 of Byrn et al. (1996)) has a different arrangement. 4-Chlorophenol and 4-methylcyclohexanone have similar lattice sections (Fig. 19 (B93)) but the inter-axial angles differ appreciably. 6

Glossary of chemical names: 1. 2. 3. 4. 5. 6. 7.

eugenol is 1-allyl-4-hydroxy-3-methoxybenzene linalool is 3,7-dimethyl-1 : 6-octadien-3-ol trans-anethole is p-propenyl-phenyl methyl ether grandlure I is (1R,2S)-( þ )-cis-isopropenyl-1-methylcyclobutane-ethanol grandlure II is (Z)-3,3-dimethyl-1,–cyclohexane-ethanol grandlure III is (Z)-3,3-dimethyl-1,–cyclohexane-acetaldehyde grandlure IV is (E)-3,3-dimethyl-1,–cyclohexane-acetaldehyde

The last four entries are sex pheromones of the boll weevil; their Zn TPP complexes have been used as slow release formulations of these pheromones (Byrn et al., 1996).


484

Table 8.14. Crystal data for triclinic and monoclinic (TPP–Mnþ) five-coordinate coordination ˚ , deg., A ˚ 3) are for reduced cells. Italicized volumes indicate that complexes. The cell dimensions (A the cell dimensions were measured at low temperature (nominally 100K) Mnþ

Ligand

Triclinic, Z ¼ 2, space group P1 Zn2þ (anthrone) HAMDAA (9-xanthone) HALWAS Zn2þ Zn2þ 3-nitroaniline HAMLAI; ˚ d(Zn–N) ¼ 2.314 A Zn2þ 5-octanoic lactone HAMJAG; ˚ d(Zn–O) ¼ 2.234 A 2,5-hexanedione HAMGUX; Zn2þ ˚ d(Zn–O) ¼ 2.261 A Zn2þ 1-hexanol 4-chlorophenol; HAMFUW; Zn2þ ˚ d(Zn–O) ¼ 2.376 A Zn2þ 4-Cl-Ph-vinylidenecarbene GAMCAY Zn2þ phenethylpropionate Grandlure III/IV Zn2þ Zn2þ Grandlure II Mg2þ (2-propanol)2 HAMBEC 145K (see text) Mg2þ (H2O)(acetone)2 133K GEPBUY (see text) Mg2þ (H2O)(2-picoline)2 DUJKUO (see text) Zn2þ 6-methyl-5-hepten-2-ol 4-methylcyclohexanone Zn2þ HAMHIM; ˚ d(Zn–O) ¼ 2.306 A

a

b

c

Cell volume

11.138 11.100 12.358

12.798 12.850 12.781

15.307 15.500 14.943

81.21 80.05 103.54

85.56 86.14 111.03

80.11 78.89 106.92

2121 2136 1950

11.202

12.671

14.516

100.47

95.52

102.80

1956

12.677

11.300

14.204

94.45

96.13

100.54

1979

10.407 10.428

10.864 10.479

19.366 18.506

89.93 88.58

76.06 81.62

77.96 79.98

2076 1970

10.667

12.514

18.091

86.94

72.90

78.51

2262

11.187 10.818 10.882 11.154

11.310 11.097 10.906 12.519

18.731 19.375 19.924 16.564

77.27 94.73 94.99 104.87

87.79 100.29 100.86 108.35

73.37 105.59 105.00 101.84

2214 2184 2220 2016

10.694

12.925

15.631

105.18

90.27

102.27

2033

10.328

13.321

16.607

87.23

82.65

86.60

2260

10.377 10.791

11.026 10.804

19.627 18.035

91.40 93.18

102.29 94.27

102.94 102.35

2132 2043

10.931

90

102.57

90

4032

10.896

90

102.70

90

4033

15.478

90

109.80

90

4093

Monoclinic, Z ¼ 4, space group P21/a or orientational variant Zn2þ (3-penten-2-ol) (4) HALXOH; 10.553 35.812 ˚ d(Zn–O) ¼ 2.267 A Zn2þ 2-phenylethylamine (4) 10.688 35.499 HAMMIR 128K; ˚ d(Zn–N) ¼ 2.193 A 3-methylcyclohexanone (4) Zn2þ 15.405 18.245 P21/n Fe2þ NO Scheidt and Frisse, 1975; 13.48 13.48 ONFTPP

9.755

Tetragonal, Z ¼ 2, I4/m

References: All data (except where noted) were taken from B93: triclinic crystals from Tables III and IV; monoclinic crystals from Table V.

The (TPP–Mg2þ)(X)} complexes are particularly interesting, and illustrative of complications that may arise. The crystal structure determination for X ¼ [propan-2-ol]2 shows that this is a five-coordinate coordination complex, where the liganded species is a ˚ ); d(Mg–O) ¼ 2.076 A ˚ ). Cell hydrogen-bonded propan-2-ol dimer (d(O . . . O) ¼ 2.740 A


TPP

485

Zn

d(O–Zn) = 2.205 Å 9-xanthone TPP

1Zn–O=C = 164.37° 1N–O=C = 94.35°

y x

z

Fig. 8.40. Projection of crystal structure of (TPP–Zn2þ)(9-xanthone) down [001] (note that the crystal data were taken from HALWAS and have a non-standard choice of origin in the figure, but not in the table; indulgence is requested for the confused state of the literature).

dimensions (for a body-centred cell) were given in Table IV (B93) under the heading ‘‘triclinic Z ¼ 2 clathrates (stage 2), expanded a.’’ The structure was solved (HAMBEC) ˚ , 111.12 101.72 101.84 , Z ¼ 1), in a primitive triclinic cell (11.155 12.519 16.595 A where the cell edges are similar to those of the reduced cell (Table 8.14) but the interaxial angles differ. When X ¼ (H2O)(acetone)2, the water oxygen is linked to Mg (d(Mg– ˚ ) and to the two acetone molecules by hydrogen bonds (d(O . . . O) ¼ 2.71, O) ¼ 2.054 A ˚ ) (McKee and Rodley, 1988; GEPBUY). Thus the water oxygen is three-coord2.81 A inate. The structure for X ¼ (H2O)(2-picoline)3 is analogous (Ong, McKee and Rodley, ˚ , The values are 0.33 A ˚ (HAMBEC), 0.45 A ˚ (GEP1986); < d(Mg–N) > ¼ 2.088 A ˚ (DUJKUO). Although the compositions suggest that these complexes BUY) and 0.41 A could be six-coordinate, the ligands have composite structures resulting in five coordinatecomplexes, with nonplanar macrocycles. These complexes provide good examples of the dangers involved in the use of lattice sections. The structural diagrams (for example, those in Figs. 1–3 of McKee and Rodley, 1988) show the nature of the ligands very clearly, whereas this is not at all obvious from the corresponding lattice sections (Figs. 21 and 23 of B93, although Fig. 30 is somewhat more transparent). The (TPP–Zn2þ)(2-phenylethylamine)2 complex is a five-coordinate coordination ˚ to the four equatorial nitrogens, complex with distances of 2.086(twice), 2.079, 2.063 A ˚ ˚. and 2.193 A to the axial nitrogen of the ligand; ¼ 0.331 A 2þ The three complexes (TPP–Zn )(X), where X ¼ 4-nitro--picoline-N-oxide, 9-anthraldehyde and 5,12-naphthacenequinone are five-coordinate coordination complexes but, because of disorder, have cell dimensions isostructural with those of the sixcoordinate coordination complexes and are more conveniently discussed in the next section.

486


8.7.2.4 The six-coordinate coordination complexes For those complexes with triclinic unit cells, the (TPP–M2þ) macrocycles are translationally equivalent and located at crystallographic centres of symmetry when Z ¼ 1, and this also holds for a single (possibly disordered) ligand in the unit cell. When Z ¼ 2 there are no symmetry requirements for (TPP–M2þ) macrocycle or ligand. There are a number of monoclinic structures where the (TPP–M2þ) macrocycles are located at crystallographic centres of symmetry but are not all translationally equivalent, and a few of lower symmetry. There are a very few orthorhombic and tetragonal complexes. Our classification is mostly based on the many families of isomorphous crystals. The triclinic complexes with Z ¼ 1. Scheidt and Lee (1987) summarized all the information available to them in terms of the reduced cells (although the standard setting of axes with a b c was not used), making transformations where necessary. We have incorporated these, and other, cells into Table 8.15 after transformation to the standard settings. Table 8.15 has been split into three parts; Part A has monofunctional ligands and host : ligand ratio 1 : 2, and Part B bifunctional ligands with host : ligand ratio 1 : 1; some miscellaneous examples are placed in part C. Part A is further divided into two groups, the first with stronger metal–ligand interaction, and the second with weaker. The first six compounds in Table 8.15 (Part A, Group I) are isomorphous. Remembering that the TPP ˚3 ˚ 3, one sees that the ‘‘ligands’’ add 110 185 A molecule has a molar volume of 800 A (per ‘‘ligand’’) to the unit cell volumes (for triclinic, Z ¼ 1 structures). The available structure analyses7 show that the ‘‘ligands’’ occupy axial coordination positions and extend roughly normal to the plane of the core; the in-plane bonding is generally stronger than the out-of-plane bonding (for example, equatorial d(Mg–N) in {(TPP–Mg2þ)[piperidine]2} is ˚ whereas the axial value is 2.39 A ˚ ). These closely isomorphous structures have 2.07 A closely similar lattice sections, as given in the Byrn et al. papers (#1 Fig. 5 (B90); #2–7 Fig. 3 (B90); #8 Fig. 3 (B90)). The next three structures are isostructural rather than isomorphous. Examination of the crystal structure of {(TPP–Mg2þ)[4-picoline]2} (HAMFAC) shows that this is a six-coordinate coordination complexes. These triclinic unit cells are all Type II, i.e. in the standard setting all angles are either obtuse or 90 . After the ninth entry onwards, we encounter a group of four isomorphous crystals which are Type I, i.e. all the angles are acute in the standard setting. We have made some attempt to group together similar unit cells, but it is difficult to do this in an entirely consistent manner. These examples illustrate the delicacy of the balance among the various factors determining the crystal structure; similar molecules often, but not always, crystallize in isomorphous crystals. The effect on the crystal structure of changing the metal atom can be investigated in the pairs {(TPP–M2þ)[pyridine]2} (M ¼ Fe, Mg) and {(TPP–M2þ)[THF]2} (M ¼ Fe, Zn), and the triple {(TPP–M2þ)[piperidine]2} (M ¼ Fe, Co, Mg). The differences in cell edges do not exceed 1–2% and in angles 2 . However, note the striking differences between {(TPP– Mg2þ)[picoline]2} and {(TPP–Fe2þ)[picoline]2}). Similar behavior is found among the 7

Parenthetically we note that the crystal structure of {(TPP–Cd2þ)[dioxane]2} was reported in the non– centrosymmetric space group P1 with an unsymmetrical disposition of dioxane ligands and a large thermal (or disorder) displacement of Cd normal to the mean core plane (Rodesiler, Griffith, Ellis and Amma, 1980). Scheidt and Lee (1987, p. 58) have pointed out that the cell dimensions and overall structure fit so well with those of the other members of the isomorphous group that re-examination seems desirable.


487

inclusion complexes (next section), e.g. the pair {(TPP–M2þ)[toluene]2} (M ¼ Zn, Mn). The effects of changing the ligand while maintaining the identity of the metal can be appreciably larger (see the triple {(TPP–Fe2þ)(L)2 (L ¼ pyridine, piperidine, 4-picoline)) in Table 8.15. The {(TPP–M2þ)[THF]2} (M ¼ Zn, Fe) are six-coordinate coordination complexes, ˚ and d(Fe–O) ¼ 2.351 A ˚ . {(TPP–Mg2þ)[methylbenzoate]2} with d(Zn–O) ¼ 2.380 A ˚ . {(TPP– (JIVSAI) is a six-coordinate coordination complex, with d(Zn–O) ¼ 2.390 A 2þ Zn )[o-chloroaniline]2} (JIVNIL) is a six-coordinate coordination complex, with d(Zn– ˚ and f ¼ 91.27 . These are all on the lower border of Group II (weaker Nax) ¼ 2.460 A interaction) crystals. The crystal structure of {(TPP–Zn2þ)[bianthrone]} (HAMDUU) is shown in Fig. 8.41. Both moieties are translationally equivalent and located at independent centers of symmetry (i.e. Z ¼ 1, as described at the beginning of this section). The Zn–O distance of ˚ is similar to the values given for {(TPP–Zn2þ)[1,5-dihydroxyanthraquinone]} 2.572 A ˚ ). B91 (p. 6551) describe these ˚ (2.58 A) and {(TPP–Zn2þ)[ p-diacetylbenzene]} (2.50 A materials as ‘clathrates’ and ‘‘porphyrin sponges containing ‘cross-linking’ ligands.’’ A lattice section is shown in Fig. 7 of B93, with the following caption: ‘‘Double-row clathrates. The bianthrone clathrate has been included among these structures with the 2 : 1 [guest : host] stoichiometry because the single molecule occupies channel sites in two adjacent channels. Each anthrone moiety fills the role of a guest species.’’ The CSD uses the term ‘‘TPP–Zn(II)-bianthrone clathrate’’ (our emphasis). In our view these are incorrect descriptions of the structure. The bianthrone molecule is here a bifunctional ligand, and the material is a six-coordinate coordination complex, with the ligand crosslinking between (TPP–Zn2þ) moieties. The three (TPP–Zn2þ) complexes with substituted anthraquinones have very similar cell dimensions. Lattice sections are given for 1,5-diaminoanthraquinone ((Fig. 3 (B93)) and 1,5-dihydroxyanthraquinone (Fig. 10 (B91)), the ligand being described as ‘‘crosslinking,’’ and for 1,8-dihydroxyanthraquinone (Fig. 1 (B91), This suggests that the 1,8-dihydroxyanthraquinone moiety is disordered, but this is not shown in the lattice section, which presents a five-coordinate situation; coordinates for oxygens are not given in JIVMIK so this point cannot be pursued. Presumably, the anthraquinone moieties behave as cross-linking ligands in the same way as bianthrone. This has been shown directly for 1,5-dihydroxyanthraquinone, using the structural results given in JIVMAC. The link to Zn is via the hydroxyl oxygen and not the carbonyl oxygen; here equatorial ˚ and axial d(Zn–O) ¼ 2.584 A ˚ , with < N–Zn–O ¼ 90.7 . Careful d(Zn–N) ¼ 2.041 A inspection of Fig. 10 (B91) is in accord with this description. Thus links from Zn to carbonyl oxygen appear to be weaker than those to hydroxyl or amino groups. {(TPP–Zn2þ)[nitrobenzene]2} (JIVPIN) is a six-coordinate coordination complex, with ˚ , < N–Zn–O ¼ 90.1 . {(TPP–Zn2þ)[benzenethiol]2} (JIVNAD) is a d(Zn–O) ¼ 2.655 A ˚ , < N–Zn . . . S ¼ 93.18 ; Zn six-coordinate coordination complex, with d(Zn–S) ¼ 3.08 A ˚ ; < C–S . . . Zn ¼ 101.62 . to equatorial N distances are 2.041, 2.044, 2.050, 2.056 A The three complexes HALTUJ, JIVNEH and JIVPEJ are actually five-coordinate coordination complexes which crystallize in disordered fashion in unit cells isostructural with those of the six-coordinate coordination complexes. We illustrate for JIVPEJ (Fig. 8.42). As noted in the caption to that figure, problems remain in the refinement of the structure, and this also appears to be the situation with HALTUJ and JIVNEH. However, disorder is not shown in the JIVNEH and JIVPEJ lattice sections of Fig. 2 (B91).

˚ , deg., A ˚ 3) are for reduced cells, Table 8.15. Crystal data for triclinic (TPP–Mnþ) six-coordinate coordination complexes. The cell dimensions (A ˚ 3, space group P transformations having been made where necessary; these cells have Z ¼ 1, unit cell volumes (¼ formula unit volumes) 1100 A 1. Note that structure determinations were often carried out in nonreduced cells. Italicized volumes indicate that the cell dimensions were measured at low temperature (nominally 100K). References in brackets Mnþ

Ligand

a/

b/

c/

V

Part A: Triclinic, Z ¼ 1, space group P1; ratio of host to monofunctional ligand is 1 : 2. ˚) Group I: Coordination complexes with stronger metal-ligand (X) interactions (d(M–X) 2.2 A Mg2þ Fe2þ

(pyridine)2 (3) HAMFAC (pyridine)2 FUXTUN (2,4)

Zn2þ

(THF)2 DOBGOQ (5)

Mg2þ

(piperidine)2 CULXIQ (1,9)

Fe2þ

(4-picoline)2 JIVMEG; 128K (2)

Fe2þ

(piperidine)2 TPPFEP (1, 7)

Zn2þ

(3-methoxy-pyrazine)2; HAMJUA

Fe2þ

Mg2þ

(dimethylphenyl-phosphine)2 GIFJAG (2, 11); (tri-n-butyl-phosphine)2 KACGJE (2, 11); (4-picoline)2 CULXEM (10);

Fe2þ

(THF)2 PHTPFE (6)

Fe2þ

9.619 102.12 9.423 101.70 9.572 102.71 9.944 101.78 9.511 100.28

11.000 103.92 10.560 104.96 11.115 103.78 11.463 104.59 11.128 104.77

11.891 113.95 11.998 111.95 11.720 115.01 11.914 115.60 12.016 113.52

1046.8

9.797 101.02 10.737 93.37 10.305 105.83 12.039 98.13 10.146 65.63 9.688 64.95

11.113 105.67 10.962 108.85 11.149 111.61 12.499 116.39 11.210 76.32 11.354 76.08

12.071 113.70 11.041 111.52 12.341 100.75 12.528 109.79 11.643 67.42 11.804 65.16

1089.6

1008.4 1022.0 1106.4 1069.8

˚, d(Fe–N) ¼ 2.026 A but designated ‘‘clathrate’’ by CSD

1121 1201

˚ d(Fe–P) ¼ 2.284 A

1492

˚ d(Fe–P) ¼ 2.345 A

1109 1055.0

Cd2þ

(dioxane)2 PHPNCD (1,7)

Co2þ

(piperidine)2 PTPORC (9);

9.845 65.50 9.934 64.98

11.327 77.41 11.494 75.01

11.614 65.84 11.830 64.45

1073.5 1100

˚ ). Group II: Coordination complexes with weaker metal–ligand (X) interactions (d(M–X) 2.5 A These complexes have been called ‘clathrates’ by the CSD. Zn2þ Zn2þ Zn2þ Zn2þ Zn2þ Zn2þ Zn2þ Mg2þ Zn2þ Cd2þ Zn2þ Zn2þ

(2-acetyl-pyridine)2 HALXEX (3) (o-chloro-phenol)2 (2) (o-chloro-aniline)2 JIVNIL 123K (2) (o-methyl-aniline)2 HAMMEN (2); (aniline)2 (3) HAMCIH 123K (2,4-dichloro-phenol)2 JIVNOR 15K (3) (2,4,5-trichloro-phenol)2 HALWEW 15K (3) (methyl benzoate)2 JIVSAI (2) (methyl benzoate)2 SEMRIL (1) (methyl benzoate)2 JIVSOW (2) (acetophenone)2 JIVRIP (2) (o-hydroxy-acetophenone)2 JIVPOT (2)

10.545 110.85 10.716 113.40 10.781 113.74 10.731 113.94 10.822 109.44 10.753 106.83 8.883 92.74 11.052 106.37 10.994 106.93 10.998 107.72 10.601 111.19 10.228 111.54

10.670 107.69 10.826 106.39 10.980 107.31 11.017 106.53 11.125 112.46 10.912 110.25 11.874 104.53 11.236 109.44 11.125 109.25 11.025 109.56 10.643 107.75 10.595 106.71

12.238 101.51 11.480 102.21 11.560 102.47 11.783 101.94 11.360 103.37 11.302 103.48 12.040 111.12 11.410 105.57 11.605 104.89 11.874 104.27 12.325 100.93 12.333 99.59

1153 1090

˚ d(Zn–O) ¼ 2.493 A

1104

˚ d(Zn–O) ¼ 2.460 A

1135

˚ d(Zn–N) ¼ 2.522 A

1087

˚ d(Zn–N) ¼ 2.470 A

1106

˚ d(Zn–O) ¼ 2.479 A

1133

˚ d(Zn–O) ¼ 2.496 A

1175

˚ d(Mg–O) ¼ 2.390 A

1180

˚ d(Zn–O) ¼ 2.619 A

1191

˚ d(Cd–O) ¼ 2.795 A

1162

˚ d(Zn–O) ¼ 2.516 A CSD gives ‘‘clathrate’’ ˚ d(Zn–O) ¼ 2.512 A (to carbonyl). CSD gives ‘‘clathrate’’

1134

Table 8.15. (Continued) Mnþ

Ligand

a/

b/

c/

V

Part B: ratio of host to bifunctional ligand is 1:1 ˚ ), and have been called ‘‘clathrates’’ by the CSD. These coordination complexes have weaker metal–ligand (X) interactions (d(M–X) 2.5 A Zn2þ Zn2þ Zn2þ Zn2þ Zn2þ

Zn2þ

Zn2þ Zn2þ

1,5-diamino-anthraquinone HAMKUB 123K (3) 4-diacetyl-benzene (2) bianthrone 128K HAMDUU x(3) # (tetramethyl pyromellitate) HALVAR (3) (4-nitro--picoline-N-oxide) at 123K HALTUJ (1)

1,8-dihydroxyanthraquinone JIVMIK (2) 9-anthraldehyde JIVNEH (2) 5,12-Naphthacenequinone JIVPEJ (2)

10.760 105.78 8.745 105.27 12.738 98.89 10.164 62.12 8.136 105.60

10.853 107.49 10.085 102.12 12.965 103.31 11.319 73.66 9.819 104.00

11.364 111.78 12.196 105.99 8.159 98.76 11.876 79.37 12.804 94.60

1060

10.578 106.72

10.794 105.61

11.559 109.49

1088

10.924 102.97 10.668 105.08

11.591 103.72 11.691 106.21

9.696 105.06 10.285 103.46

1097

˚. d(Zn–N(H)) ¼ 2.614 A CSD gives ‘‘clathrate’’

1044 1270 1157 944.4

1123

d(Zn–O) ¼ 2.572 ˚ ; see text. A d(Zn–O) ¼ 2.540 ˚ . CSD gives ‘‘clathrate’’ A Disordered but CSD gives d(Zn–O) ¼ 2.527 ˚ , and classifies as A ‘‘5-coordinate coordination complex.’’ Guest disordered, no conclusion possible. Guest disordered, no conclusion possible. Guest disordered; CSD gives ‘‘clathrate’’

Zn2þ

1,5-dihydroxyanthraquinone* JIVMAC (2)

Part C.

Miscellaneous

Zn2þ

(methyl 4-nitrobenzoate) (3)

Fe2þ

(NO)(4-methyl piperidine) NIPORF (3)

10.590 107.37

10.857 104.79

11.294 111.29

1054

11.468 100.54 11.550 87.83

20.357 105.55 17.236 74.38

2460

T r i c l i n i c P 1 Z¼2 11.204 96.85 10.668 87.07

2042

˚, d(Fe–N(O)) ¼ 1.741 A ˚ d(Fe–N(H)) ¼ 2.463 A

For illustration we give both Niggli and Delaunay reduced cells for {(TPP–Co2þ)[piperidine]2} Co2þ

(piperidine)2 Niggli (9);

Co2þ

(piperidine)2 Delaunay (8);

9.934 11.494 9.934 11.503

11.830 67.73 11.830 101.49

78.01 64.46 101.99 115.64

1126 1126

* anthrarufin x HAMDUU is listed as a ‘clathrate’ by the CSD, but is clearly a six-coordinate coordination complex. No CSD entry found for {(TPP– Zn2þ) (3-methyl-2-cyclohexenone)2} (Fig. 6 of B93 suggests two orientations for 3-methyl-2-cyclohexenone, as noted in caption). References: (1) B90; (2) B91; (3) B93; (4) Li, Coppens and Landrum, 1988; (5) Schauer, Anderson, Eaton and Eaton, 1985; (6) Reed, Mashiko, Scheidt, Spartalian and Lang, 1980; (6) Rodesiler, Griffith, Ellis and Amma, 1980); (7) Radonovich, Bloom and Hoard, 1972; (8) Scheidt, 1974a; (9) McKee, Ong and Rodley, 1984; (10) Sodano, Simmoneaux and Toupet, 1988; (11) Belani, James, Dolphin and Rettig, 1988; (12) Scheidt, Brinegar et al., 1977.


492

3A (TPP-Zn2+)

bianthrone

d(Zn–O) = 2.572 Å x

B

z y

d(Zn–N) = 2.037, 2.051 Å

Fig. 8.41. The crystal structure of {(TPP–Zn2þ)(bianthrone)} at 128K, viewed down [001]. There are chains of (TPP–Zn2þ) and bianthrone moieties along [010] linked by bonds between Zn and carbonyl oxygen. Both moieties are at crystallographic centers of symmetry. There are resemblances to the arrangement in {(TPP–Zn2þ)(dibenzosuberone)3} (Fig. 8.43) but also appreciable differences. (Data from HAMDUU.) Table 8.16. Crystal data for monoclinic (TPP–Mnþ) six-coordinate coordination complexes. The ˚ , deg., A ˚ 3) are for reduced cells, transformations having been made where cell dimensions (A ˚ 3, space group P21/n unless stated necessary; these cells have Z ¼ 2, unit cell volumes 2000 A otherwise. Italicized volumes indicate that the cell dimensions were measured at low temperature (nominally 100K). The original cell dimensions come from Table V of B93, except for (1) McKee and Rodley, 1988; (2) McKee, Ong and Rodley, 1984 Mnþ

Ligand

a

b

c

Cell volume

Cu2þ Zn2þ

(picoline)2 HALWUM; 145K (2-methoxy-3-isobutyl-pyrazine)2 P21/c HAMBON; ˚. d(Zn–N) ¼ 2.925 A (acetophenone)2; HALZOJ (phenethylpropionate)2 (methanol)2 (1); GEBPIM (methanol)2(acetone (1) GEPBOS (1-methylimidazole)2 (2); ˚ CULXAI, d(Mg–N) ¼ 2.227 A

10.125 14.774

15.372 10.968

13.547 15.791

99.91 92.99

2077 2555

10.708 8.738 13.302 10.01 20.764

20.653 26.960 12.868 17.75 9.659

10.819 11.066 11.039 12.74 ˚3 4164 A

102.27 2338 99.60 2567 113.29 1736 110.7 2118 Tetragonal P42/n, Z ¼ 4

Mg2þ Zn2þ Mg2þ Mg2þ Mg2þ

The two complexes with phosphine ligands are pseudo-octahedral six-coordinate coordination complexes of the usual type. Only some of the complexes crystallizing in monoclinic space groups and listed in Table V of B93 are six-coordinate coordination complexes. These are given in Table 8.16, together with other relevant examples. Other entries in Table V, such as those for transanethole and 4-thiocyanatonitrobenzene are actually four-coordinate inclusion complexes (Table 8.18).


493

3C

1.343 Å

1.225 Å

2.656 Å

z x

A 0 y B

Fig. 8.42. Projection of JIVPEJ {(TPP–Zn2þ)[5,12-naphthacenequinone]} down [100]. There are chains of (TPP–Zn2þ) and [5,12-naphthacenequinone] moieties along [001] linked by Zn . . . O1 ˚ ). The [5,12-naphthacenequinone] moiety takes up two orientations as shown interactions (2.656 A by the ostensible presence of four carbonyl oxygens, instead of two. The oxygens of the second ˚ away from the Zn, too far for meaningful interaction. Furthermore, one C¼O orientation are 4.512 A ˚ , but the other is long at 1.343 A. Thus further study distance has the acceptable value of 1.225 A seems desirable.

The methanol and 1-methylimidazole complexes are regular six-coordinate coordination complexes; the ligand in the methanol-acetone complex is this hydrogen-bonded pair, with methanol linked to Mg. The Mg–O distances are 2.220 (methanol complex) and ˚ (methanol-acetone complex). 2.188 A 8.7.3 Crystallography of (5,10,15,20)-tetraphenylporphyrin inclusion complexes 8.7.3.1

Crystallography of four-coordinate (5,10,15,20)-tetraphenylporphyrin inclusion complexes There are three different groups of {(TPP-M2þ)[guest]2} structures which come into this overall category–(1) triclinic, Z ¼ 1 (Table 8.17); (2) triclinic, Z ¼ 2 (Table 8.18);


494

Mn tetraphenyl porphyrin

Mn hidden C

B

0 toluene

z

A

y x

Fig. 8.43. The packing in {TPP–Mn2þ}[toluene]2 viewed down [101]. The included toluene molecules have been increased in size and lightened in colour. (Data from Kirner, Reed and Scheidt, 1977.)

(3) monoclinic, Z ¼ 2 (Table 8.19). Strouse and coworkers use the term ‘clathrates’, while we prefer to restrict this word to its original usage. All the complexes in Table 8.17 have a 1 : 2 host guest ratio, apart from {(TPP–2Hþ)[ p-xylene]} (SEMNUT). The crystal structure of triclinic {(TPP–Mn2þ)[toluene]2} has been determined at 98K (Fig. 8.43) and has two toluene molecules, related by the center of symmetry at the centre of the ˚ from the mean plane of the macrocycle. The structural macrocycle, at distances of 3–3.4 A unit is the centrosymmetric combination (toluene)(TPP–Mn2þ)(toluene), where the toluene plane is nearly parallel to the plane of the TPP core. In Table 8.17 the entries #2–8 form an isomorphous group, which is in accord with the corresponding lattice sections given in Figs. 5, 6 and 7 (B90); a diagram has not been found for {(TPP–Zn2þ)[ p-xylene]2}. Perhaps surprisingly, the unit cells of {(TPP– M2þ)[toluene]2} (M ¼ Mn, Zn) and {(TPP–Zn2þ)[2-fluorotoluene]2} have remarkably similar dimensions. We note in the table four entries for different M, but all with m-xylene as guest; this follows Table 2 of B91, except that we compare reduced cells whereas they compared body-centred cells. We have chosen the anthracene complex, which is somewhat more closely packed than the others of this group, to illustrate the packing arrangement in these complexes (Fig. 8.44). The (TPP–Zn2þ) moiety is located at Wyckoff positions (a) (0,0,0) and the anthracene molecule at (d) 1/2,0,0. The plane of the anthracene molecule is not parallel to the plane of the porphyrin ring, and a charge transfer interaction would require substantiation by spectroscopic techniques. It seems more reasonable to describe this as an inclusion complex where anthracene is held by pairs of phenyls from translationallyrelated (TPP–Zn2þ) moieties.


495

˚ , deg., A ˚ 3) for triclinic (TPP–M2þ) four-coordinate inclusion complexes, Table 8.17. Crystal data (A ˚ 3, space group P with Z ¼ 1, unit cell volumes (¼ formula unit volumes) 1100 A 1. The cell dimensions are for reduced cells. References in brackets Mnþ

Guest

Type II 2Hþ Mn2þ Zn2þ Zn2þ Zn2þ Zn2þ Zn2 þ* Cd2þ Zn2þ Zn2 þ* Zn2þ* Zn2 þ* Cu2þ 2Hþ Zn2þ Zn2þ Zn2þ Zn2þ Zn2þ Zn2þ Zn2þ

triclinic unit cells [p-xylene] SEMNUT (1) (toluene)2 TPPMNT10 (1) (toluene)2 ZNPORT (1) (2-fluorotoluene)2 HAMLOW (3) (indene)2 HAMJEK(3) (indole)2 HAMJIO (3) (o-xylene)2 SEMPEF (1) (m-xylene)2 JIVRUB (2) (m-xylene)2 SEMNON (1) (p-xylene)2128K JIVPUZ; (2) (ethylbenzene)2 SEMMAY (1) (styrene)2 HAMKAH (1) (m-xylene)2 JIVPAF 128K (2) (m-xylene)2 SEMNIH (1) phenanthrene SEMMUS (1) anthracene HALYAU (3) coronene HAMGAD (3) benzacephenanthrylene HALYUO (3) 1,2-benzanthracene HALYIC (3) 2,3-benzfluorene HALZEZ (3) (bromobenzene)2 JIVNUX 193K (2) (phenyl-acetylene)2 SEMNAZ (1) Zn2þ Zn2þ (1,2,4-trimethyl-benzene)2 SEMMEC (1) Zn2þ (anisole)2 HALTIX (3) Zn2 þ* (3-methyl-anisole)2 SEMROR (1) Zn2 þ* (m-ethylvinyl-benzene)2 (1) Zn2þ 2,5-dimethyl-2,4-hexadiene (1) Zn2þ (3-hydroxy-acetophenone)2 JIVREL (2) Zn2þ (phenylisocyanate)2 HAMFIK 145K (3) (nitrobenzene)2 JIVPIN Zn2þ

Type I triclinic unit cells (4-vinyl-anisole)2(3) Zn2þ Zn2þ (phenazine)2 HAMMOX (3) Zn2þ (acridine)2 JIVVAL 128K (2) (1,5-cyclo-octadiene)2 SEMPIJ (1) Zn2þ Zn2þ 9,10-bis(phenethynylanthracene) x HAMDEE 128K(3) Zn2þ ((bicycloheptene)-COOH)4 JIVVEP (2)

a

b

c

V

7.879 10.487 10.502 10.480 10.541 10.766 10.550 10.926 10.992 10.076 10.932 10.408 9.924 10.172 9.529 9.458 10.491 10.372 10.593 10.590 10.758

10.418 11.320 11.349 11.390 11.162 11.040 10.980 10.995 11.174 11.297 11.487 11.613 10.570 10.749 10.715 11.990 10.570 10.901 11.146 11.182 11.047

12.956 11.465 11.404 11.462 11.912 11.902 12.057 12.067 11.796 11.885 11.488 11.648 11.984 12.020 11.790 10.655 12.497 11.971 11.651 11.530 11.501

103.73 107.80 107.65 107.85 106.81 106.71 105.83 105.52 105.65 108.30 106.91 106.69 104.46 104.28 106.13 103.52 110.77 108.21 107.37 106.54 106.90

102.19 110.63 110.48 109.95 112.62 114.38 114.34 114.55 113.42 109.92 105.88 110.18 109.54 110.21 103.49 105.19 104.59 106.60 105.51 104.60 106.75

97.66 103.34 103.87 103.50 102.23 102.70 101.12 103.56 105.19 101.43 110.58 103.45 97.45 98.73 104.38 106.30 101.61 107.25 110.88 111.85 110.20

991 1122 1119 1134 1153 1138 1148 1168 1164 1134 1173 1174 1115 1154 1060 1056 1187 1110 1114 1111 1107

10.359 10.837

10.768 11.162

11.611 12.359

106.61 99.73

104.29 108.95

105.40 116.51

1121 1175

10.726 11.18 11.340 10.851 11.416

11.350 11.48 11.506 11.370 11.644

11.360 11.68 11.590 11.995 10.465

107.85 107.29 106.33 108.28 109.62

105.31 111.75 107.24 99.10 104.37

109.63 107.38 111.19 114.19 107.49

1131 1180 1212 1210 1152

10.439

10.686

11.584

107.79

103.52

107.39

1097

10.792

11.110

11.329

107.72

104.44

109.12

1127

10.630 11.145 11.106 11.147 9.855

18.941 11.355 11.284 11.518 10.291

11.542 11.616 11.664 11.632 13.022

75.39 66.19 67.41 60.44 79.98

74.30 76.63 71.74 82.45 77.00

71.11 72.11 70.80 62.75 81.33

1203 1244 1245 1146 1259

11.674

11.676

12.751

86.00

63.69

77.14

1518

Notes: x Table II of B93; ‘‘layered’’ clathrate, see Fig. 27. References: DR92 – Dillen and Roos (1992); PB94a – Pang and Brisse (1994a); PB94b – Pang and Brisse (1994b); PHW94 – Pang, Hynes and Whitehead (1994); RD92 Roos and Dillen (1992);RR91a – Rooyen and Roos (1991a); RR91b – Rooyen and Roos (1991b); SERP02 Sidhu, Enright et al., 2002; SEUR05 – Sidhu, Enright et al., 2005.

496


C Zn 0B Zn anthracene

Zn A x

Zn y z

Fig. 8.44. The packing arrangement in the triclinic (Z ¼ 1) {(TPP–Zn2þ) [anthracene] fourcoordinate inclusion complex (HALYAU). A ‘‘lattice section’’ version of this diagram is given in Fig. 3 of B93. (Data from HALYAU.)

When a centrosymmetric guest such as anthracene is replaced as guest by a noncentrosymmetric molecule such as phenanthrene, and the latter occupies the same site, then it is necessarily disordered, as B93 pointed out in the caption to their Fig. 4. A related example is provided by (TPP–Zn2þ)(9,10-bis(phenylethynyl)-anthracene), where the reduced cell has appreciably different dimensions from the other entries in Table 8.17; the lattice section shown in Fig. 27 (B93) has been described as that of a ‘‘layered’’ clathrate. Comparison of Figs. 8.44 and 8.45 shows that these two inclusion complexes are isostructural, the stacked arrangement of (TPP–Zn2þ) and anthracene moieties (vertical direction in the two diagrams) being the same. The arrangement in a horizontal direction is different because of the difference in size between the two guest molecules. Triclinic HALTIX (guest anisole) has an analogous structure; here the two anisole molecules are related by the center of symmetry; this also holds for HAMMOX (guest: phenazine; Fig. 7 of B93), and JIVVAL (guest: acridine, Fig. 7 of B91). Both are described as. ‘‘double-row clathrates’’ When the guest is 3-methylanisole, the methyl and methoxy groups are disordered (B90, p. 1866; Fig. 6) but the composition remains 1 : 2. The last entry in Table 8.17 has an unusual (but not unprecedented) composition. B91 (p. 6550) have already noted that pairs of hydrogen-bonded dimers of bicycloheptenecarboxylic acid are formed. These dimers constitute the guest species in a four-coordinate inclusion complex (Fig. 8.46). However, there is a complication, because there is also a ˚ ). Thus weak interaction between Zn and one of the carboxyl oxygens (d(Zn–O) ¼ 2.89 A


497

C Zn 9,10(bis(phenylethenyl)anthracene 1.198 Å 0

A

z x

y

3B

Fig. 8.45. The packing arrangement in the triclinic (Z ¼ 1) {(TPP–Zn2þ) [9,10-bis (phenylethynyl)anthracene] four-coordinate inclusion complex (HAMDEE). The location of the triple bond is indicated by its bond length. A ‘‘lattice section’’ version of this diagram is given in Fig. 27 of B93. (Data from HAMDEE.)

{(TPP–Zn2þ)[bicycloheptenecarboxylic acid]4} comes somewhere between a fourcoordinate inclusion complex and a six-coordinate coordination complex. Carboxylic acid dimers are also found when the guests are o- and m-toluic acid, but without the additional interaction. Another example is {(TPP–Zn2þ)[3-hydroxyacetophenone]2}, where the hydrogen-bonded 3-hydroxyacetophenone dimer is shown in Fig. 8.47. Similarly in {(TPP–Zn2þ)[o-isopropylphenol]2}, where the guest molecules are hydrogen bonded as a linear pair; these situations were shown clearly in Figs. 5 and 4 of B91, but not mentioned in the text. We also note that the cell dimensions for JIVREL and the three isomorphous {(TPP–M2þ)[methyl benzoate]2} complexes (M ¼ Mg, Zn Cd) (Table 8.15) are very similar. Nevertheless, the first of these is an inclusion complex and the other three are six-coordinate coordination comples. Structure cannot necessarily be inferred from composition or similarity of cell dimensions. The second to the seventh entries of Table 8.18 are isomorphous to a fair approximation; this is confirmed by comparison of crystal structures, and is in accord with the similar lattice sections shown for these complexes in Fig. 29 of B93. The last four entries also form an isostructural group. These two groups (and some other entries) are treated ˚ together by B93, with the title ‘‘ ‘Herring-bone’ monoclinic Z ¼ 2 clathrate with a 19 A chain’’), without taking into account that the b axes differ for these two groups. This is also not immediately obvious from the lattice sections. (TPP–Zn2þ)(diphenylacetylene) (HAMHUY; Fig. 8.48) is a monoclinic analog of the (TPP–Zn2þ) (anthracene) structure, with the (TPP–Zn2þ) centred at 0,0,0 and the guest at 1/2,0,0. The two guest molecules in the 1 : 2 structures are related by centers of symmetry, as is shown in the layered structure of HAMCAZ; the oxygens of the 4-thiocyanato˚ from the Zn, and the sulfurs are even further. In the nitrobenzene guests are more than 4 A


498

bicycloheptene carboxylic acid dimer

N bicycloheptene Zn

O d (Zn–O) 2.89 Å

COOC6 H5 > NO2 : Substitution in the 1-position tended to favor molecular compound formation while substitution in the 2-position sometimes hindered. Somewhat similar comparisons have been made for the picric acid compounds of polymethylbenzenes and of polyhydroxybenzenes and naphthalenes, and of some aromatic hydrocarbons (Baril and Hauber, 1951). For benzene derivatives the tendency to compound formation is greater the larger the number of methyl substituents and the more symmetrical their disposition in the ring. Methyl groups were found to be even more effective in side chains than in the ring, a perhaps surprising result. Unsaturation in the side chains produced very explosive picrates! However, it should be appreciated that these considerations are intrinsically qualitative as they rest on the formation of crystalline molecular compounds, without taking into account their structures, which may be very varied. Polycyclic hydrocarbons forming molecular compounds can be divided into the following groups: (i) Planar aromatic hydrocarbons, e.g. benzene, naphthalene, anthracene, phenanthrene, pyrene, perylene. (ii) Nonplanar aromatic hydrocarbons, e.g. the helicenes, heterohelicenes, 9,10-dihydroanthracene. (iii) Substituted aromatic hydrocarbons, e.g.

CHEMICAL NATURE OF DONORS AND ACCEPTORS

935

NH2

(H3C)2N

N(CH3)2

2,7-diaminopyrene

TMPD NH2

(iv) Heteroaromatics and related compounds, e.g. N

S N

acridine

S

S

S

S

tetrathiafulvalene (TTF)

H phenothiazine

X

X

X X tetra-X-tetracene: X = S thia Se selena Te tellura

N

O

H 1,2,3,4-tetrahydro1-oxocarbazole

N OH carbostyril (lactam form)

Heteroaromatic donors, mainly with nitrogen and oxygen substitution, have been known for many years while sulphur-containing molecules have become very important recently. Tetrathiafulavalene (TTF) and derivatives act as donors in many radical-cation salts and, of course, in {[TTF][TCNQ]} (see Chapter 17 for formulation), arguably the most-widely studied of the ‘organic metals’. The effects of chemical modification of TTF will be illustrated later (Chapter 17). Another important donor containing sulphur is tetrathiatetracene; Se and Te analogs have also been prepared (Sandman et al., 1982). Among other donors are various metallocenes, including ferrocene, dibenzene chromium and tricarbonylchromium anisole, porphyrins and coordination complexes (e.g. metal oxinates with Cu(II), Pd(II) and Ni(II) as the metal atoms). 13.3.3 Acceptors Acceptors are conveniently grouped on the basis of fundamental structures to which substituents have been appended in various ways. The fundamental structures are ethylene, p- and o-benzoquinone, cyclopentadienone, benzene, naphthalene and fluorene. The most important electron-withdrawing substituents are nitro, cyano, halo, anhydride, furoxan and furazan groups. We give below a selection of proven and potential acceptors. Synergistic interaction of various substituents seems to be an important factor in the formation of electron acceptors. For example, maleic anhydride, phthalic anhydride and

CHARGE TRANSFER MOLECULAR COMPOUNDS

936

hexachlorobenzene do not form stable -compounds with aromatic hydrocarbons while tetrachlorophthalic anhydride (TCPA) is a powerful electron acceptor. Unsymmetrical substitution, as in DDQ, enhances the effectiveness of electron acceptors, presumably because dipole and polarization interactions also contribute to the stability of the molecular compounds, in addition to the charge transfer interactions. (a) Substituted olefins Examples are TCNE, TCNQ and hexacyanobutadiene (HCBD). Of these TCNQ is undoubtedly the most prominent and occurs not only as the radical anion in many charge transfer salts and organic metals but also as a neutral acceptor in many mixed stack -molecular compounds (for some other formulae see Section 17.2.3). O O

O

O R'

NC

CN

NC

CN

CN

NC

O

O

O

CN

O

O

CN

NC

TCNE

HCBD

NC

R NC

R

CN CN

CN

R

NC

CN NC

a: R = R' = H; b: R = NO2, R' = H; c: R = R' = NO2.

CN

NC

TCNQ

CN TNAP

(b) Substituted aromatic hydrocarbons Substituted polynitrobenzenes (e.g. TNB, picric acid) are the best known members of this group but polynitro naphthalenes and fluorenes also form molecular compounds. Anhydrides of benzene and naphthalene polycarboxylic acids are important acceptors, as well as polycyanobenzenes. X O2N

NO2

O O2N

C O

X = H, OH, CONH2 NO2 X O2N

NO2 Y NO2

X = H, OH, CH3, Cl, Br, I, NH2, (N ≡ N); Y = H; X = Y = OH.

2


O2N

NO2

NO2

O2N

NO2 NO2 O2N

NO2

O O N

O

N

N

N

N

N

O O

O

N

N

N

N

O N

O

O

N

benzotrifuroxan ( BTF).

NC

CN

NC

CN

N N

CN

N CN

CN NC

CN

NC

CN

NC

CN

NC

CN TCNB

CN X

X

O

O

O O

O

O X

O

O

X

O

X = H, F, Cl, Br or I X = Cl: TCPA X

X

X

X O

O

various anhydrides

O

X

O O

O O

O

O

mellitic trianhydride

X

X

O

O

X X = H: PMDA

O

937


938

(c) Unsubstituted and substituted quinones Most attention has been paid to the tetra-substituted p-benzoquinones (see also Section 17.2.3 (v)), the less stable o-benzoquinones having been largely ignored; fluorenones, naphthoquinones, phenanthraquinones and anthraquinones have also been studied. Two vicinal triones are relatively new, powerful acceptors and have been shown to form 1 : 1 molecular compounds with pyrene (Gleiter and Schanz, 1980). We have already noted that the quinhydrones (Section 15.7.1), and {aromatic hydrocarbon quinone} molecular compounds (Section 15.6) are best treated as separate groups. n n n

O

O

O

X'

X

X'

NO2

O2N

O a: X = X' = F, Cl, Br, I, CN or CH3. b: X = Cl; X' = CN. O

O

O

X

O

O

Cl Cl

O O Cl

Cl

O Cl

O

O

O

O

Cl

Cl Cl

O

Cl

Cl O

O

O O

X

O

X

X

X X = Cl, CN

R

O O R

R' a: R = NO2, R' = H. b: R = H, R' = NO2.

R' Substituted o-benzoquinones and phenanthraquinones

O

O

O

O O

O

O

O

O

O O

O

1,2,3,5,6,7-s-hydrinacenehexone 1,2,3,6,7,8-pyrenehexone

(d) Cyclopentadienone systems The last of the molecules shown in this section is of particular interest because it contains an asymmetric carbon atom (asterisked); the presence of the carboxyl group allows resolution into enantiomers which can be used for the formation of diastereoisomeric -complexes with nonplanar overcrowded aromatic hydrocarbons. The resolved aromatics can then be recovered (Newman and Lutz, 1956).


R'

939

R

Cl

Cl

Cl

R

R

Cl Cl Cl

O a: R = R⬘= H; b: R = NO2, R⬘= H; c: R =R⬘= NO2.

O Cl

Cl

Cl R'

R

Cl Cl Cl O2N

O2N

O NO2

R

NO2

R O a: R = R⬘= H; b: R = NO2, R⬘= H; c: R = R⬘= NO2.

OC∗H(CH3)COOH

(e) Chelates Chelates act as acceptors in some -molecular compounds; examples are bis(difluoroborondimethylglyoximato)Ni(II) (on the left), metal bis(dithiolene) (on the right, where M ¼ Ni, Pd, Pt) and bis(cis-1,2-trifluoro- methylethylene-1,2-dithiolato)Ni(II). The chemistry of the dithiolenes and related species and their metal complexes has been comprehensively reviewed (Mueller-Westerhoff and Vance, 1987). F

F B O

O N

N

R

S

Ni N

R

M N

O

S

R

S

S

R

O B

F

F

13.3.4 Quasi-acceptors Three groups of molecular compounds are known which are structurally similar to the mixed stack -donor–acceptor molecular compounds but do not show charge transfer bands in their absorption spectra. The donors are regular donors of the -compound series and thus any difference in properties must derive from the second component, which we call a quasi-acceptor.


940

The first quasi-acceptor type is a halogenated aromatic hydrocarbon; the perfluorinated molecules hexafluorobenzene, octafluoronaphthalene and decafluoro-biphenyl are the best known examples, especially C6F6, and the structures of their molecular compounds will be described later (Section 15.9.1). Decachloropyrene perhaps behaves in a similar way but other examples are not known. O CH3

O

CH3 H

N

N

CH3

N O

O

N CH3

N

O

CH3

N

N

H

1,3,5,7-tetramethyluric acid (TMU)

caffeine

CH3 N

N

O

N N

CH3

O 1,3-dimethylalloxazine

The second type is based on the purines 1,3,7,9-tetramethyluric acid and caffeine; some pyrimidines may play a similar role. Mixed stack molecular compounds of aromatic hydrocarbons and heteroaromatics with TMU and caffeine have been prepared (WeilMalherbe, 1946; Booth and Boyland, 1953) and some crystal structures are known (Section 15.9.2). These have mixed stack structures but there is no charge transfer band in solution or solid state spectra. Pyrene, TCNE and DDQ form 1 : 2 molecular compounds with 1,3-dimethylalloxazine but the varying nature of the first component suggests that further study is needed before these can be reliably classified. The third group is based on the flavins as quasi-acceptors. Various iosalloxazine derivatives have been reported to form molecular compounds of different types with electron donors such as phenols. The chemistry and structures of these molecular compounds are discussed in Section 15.7.2. 13.3.5

Ionization potentials of donors and electron affinities of acceptors

The first ionization potential (or energy) of a molecule is defined as the energy which is required to remove an electron from the highest occupied molecular orbital (HOMO) of the neutral molecule in its ground state and is denoted I1. It is thus the enthalpy of the reaction D ! Dþ þ e (infinitely separated):


941

The electron affinity (EA) of a molecule is defined as the difference in energy between the ground state of the neutral molecule plus an electron at rest at infinity, and that of the negative ion. It is thus the enthalpy of the reaction A þ e ! A The electron affinity of a stable negative ion is defined as positive, although this contradicts the usual thermochemical sign convention. The difference in energy between the lowest vibrational level of the ground state of the neutral molecule and the corresponding level of the cation or anion is termed the ‘adiabatic ionization energy’, or ‘adiabatic electron affinity’, respectively. The energy difference between the energy level of the ground state and that part of the potential curve to which, on applying the Franck–Condon principle, transition is most likely to occur is denoted the ‘vertical ionization energy’ or the ‘vertical electron affinity’ as the case may be. Photoelectron spectroscopy, for example, gives the vertical ionization energy. Experimental and calculated values of ionization potentials and electron affinities have been tabulated (Blaunstein and Christopherou, 1971; Rosenstock et al., 1980). Reasonably accurate values of ionization potentials have been available for many materials for some time; in the last few years many new values have been obtained and there has been a marked increase in accuracy stemming from widespread application of photoelectron spectroscopy (Eilfeld and Schmidt, 1981). Selected values for donors of interest here are given in Table 13.3. There has been considerable confusion about the correct absolute values of electron affinities and two rather different sets of values were proposed at one point (Briegleb, 1964; Batley and Lyons, 1962). However, these can be correlated and put on an absolute scale (Chen and Wentworth, 1975) and absolute electron affinity values have been calculated for about 150 electron acceptors from measurements of charge transfer spectra and half-wave potentials (Table 13.4, which also includes some more recent values). It is illuminating to compare typical values of I1 and EA for organic donors and acceptors with those for alkali metals and halogens. For a typical organic donor I1 7 eV, whereas for Li I1 ¼ 5.4 eV and for Cs I1 ¼ 3.4 eV; for the best organic acceptors EA 2.8 eV, while EA for fluorine is 3.4 eV and for iodine 3.2 eV. Thus it is somewhat more difficult to form cations from organic donors than from alkali metal atoms, and also somewhat less advantageous to form anions from organic acceptors than from halogen atoms. The values of ionization potential and electron affinity given in Tables 13.3 and 13.4 refer to isolated gaseous molecules and we shall denote them as IG and EAG respectively. However, the appropriate ionization potentials for molecules in crystals (IC) will be less than those for isolated molecules because of polarization effects in the crystals. Similarly the electron affinity of a molecule in a crystal (EAC) will be greater than that of a molecule in the gas phase because of polarization. Thus we can write IC ¼ IG – P, where P is the polarization energy; similarly EAC ¼ EAG þ P*, where both P and P* are positive quantities. P and P* depend on the polarizabilities of the molecules involved as well as on the crystal structures. Gutmann and Lyons (1981) have given a detailed discussion; the following values (Table 13.5) for P (for a single charge) are quoted from them and other sources. A comprehensive set of values of P has been determined by ultraviolet photoelectron spectroscopy (examination of line widths of the spectra) for 44 organic


942

Table 13.3. First -ionization potentials (eV) for some aromatic hydrocarbons and other molecules. Values from Schmidt (1977) unless stated otherwise. Method is photoelectron spectroscopy (PES) unless stated otherwise Molecule

I1

Molecule

I1

Molecule

I1

benzene naphthalene

Perylene (CS77) Ovalene (BL62)

6.97 6.71

Coronene (BL62) Diphenyl (R81)

7.29 7.95

Biphenylene (BCS74)

7.61

hexamethylbenzene (R80)

7.8 (CTS)

tetracene

9.24 8.15 8.12 (adiabatic) (SSI81) 7.41 7.36 (adiabatic) (CBS72) 6.97

6.83

TDAE

6.5

pentacene

6.61

Tetrathiofulvalene (R80) Graphite (BL62)

4.9

6.28 (ETR)

phenanthrene

7.86

chrysene

7.59

triphenylene phenothiazine1

7.88 6.54 (ETR)

benzo[c]phen-anthrene N,N,N 0 ,N 0 -tetramethylbenzidine (CS77)

7.60 6.48 (ETR)

N,N-dimethyl-pphenylenediamine (BCS74) 1,2,7,8dibenzochrysene Pyrene (CS79)

anthracene

7.20 7.41

Notes: PES photoelectron spectroscopy (molecule in gas phase); accuracy 0.02 eV for I < 10 eV; vertical ionization potential is measured. CTS charge transfer spectrum (molecule in solution). ETR electron transfer reactions (molecule in solution); adiabatic ionization potential is measured. References: B74 – Boschi et al., 1974; BL62 – Batley and Lyons, 1962; C72 – Clark et al., 1972; CS77 – Clar and Schmidt, 1977; CS79 – Clar and Schmidt, 1979; R80 – Rosenstock et al., 1980; SSI81 – Sato et al., 1981.

solids (Sato, Seki et al., 1981); the values of P ranged from 0.9 to 3.0 eV. Planar condensed aromatic hydrocarbons appeared to have a common value of 1.7 eV. These values refer to the crystals of the pure components and somewhat different values would be expected to apply to these molecules in the crystals of the charge transfer molecular compounds. Measurements of photoelectric emission thresholds for neutral and ionic charge transfer compounds give values of P for anthracene and perylene in some of their molecular compounds as well as in their neat crystals (Batley and Lyons, 1968) and these values are also included in Table 13.5. Taken at face value these numbers would appear to mean that the donor qualities of anthracene are enhanced (by 0.2–0.4 eV) in its PMDA and TCNQ charge transfer molecular compounds compared to its behaviour in its neat crystals while those of perylene are diminished in its chloranil and PMDA -compounds. Analogous values for electron acceptors (P*) are hardly available as measurement of the solid state electron affinity is very difficult. Nevertheless the electron affinity of PMDA in PMDA molecular compounds has been estimated as 3.6(3) eV (Stezowski et al., 1986).


943

Table 13.4. Electron affinities (eV). Most of the values have been taken from the more extensive collection in Table III of Chen and Wentworth (1975); these are charge transfer spectrum (CTS) values except where noted otherwise (PDS photodetachment spectra, E1/2 half wave potentials) Molecule

EA

Molecule

EA

Molecule

EA

p-benzoquinone

1.83 2.62

1.09 (PDS) 2.57

p-bromanil

2.48

TCNB

2.00

DDQ (CS77) p-chloranil p-iodanil

hexacyanobenzene p-fluoranil hexacyanobutadiene

2.56 2.45 3.09

duroquinone

3.13 2.48 2.43 (E1/2) 1.67

2,4,7-trinitrofluoren-9-one 2,4,5,7-tetranitrofluoren-9-one 2,4,5,7-tetranitrofluorenemalononitrile TCNE TCNQ tetrafluoro-TCNQ

2.17

o-bromanil

Hexafluorobenzene (SH80) o-chloranil

2.82

2,5-difluoro-TCNQ (SF79)

tetranitromethane

1.63

2,3-dicyano-5,6dicyano-7-nitro-1, 4-naphthoquinone TNB

1.73

TNT 1,4,5,8naphthalene tetracarboxylic acid anhydride

1.67 2.28

TCPA mellitic trianhydride

1.72 2.38

tetrabromophthalic anhydride PMDA dibromo-PMDA

2.24 (E1/2) 2.56 2.77 2.84 3.22 (E1/2) 3.02

1.72 2.04 2.23

References: CS77 – Clar and Schmidt, 1977; SF79 – Saito and Ferraris, 1979; SH80 – Sowada and Holroyd, 1980.

Table 13.5. Values of polarization energy P (in eV) for aromatic hydrocarbon molecules in their neat crystals (values from Gutmann and Lyons (1981), except where noted otherwise) and in some -molecular compounds (values from Batley and Lyons (1962), except where noted otherwise) Molecule in its neat crystals

P

naphthalene phenanthrene tetracene chrysene pyrene coronene anthracene

1.3, 1.4 2.1, 2.1, 1.7, 2.4, 1.8,

Molecule in some -molecular compounds

P

anthracene PMDA anthracene TCNQ perylene PMDA perylene chloranil

2.16 2.39 1.81 1.57

1.7 (SSI81) 1.57 (S81) 1.7 (SSI81) 1.6 (SSI81) 1.7 (SSI81) 1.61 (KSS82)

n n n

n n n

perylene

1.9, 1.6 (SSI81)

n n n

n n n

References: KSS82 – Karl, Sato, Seki and Inokuchi, 1982; S81 – Silinski, 1981; SSI81 – Sato, Seki and Inokuchi, 1981.

944

13.3.6


Determination of degree of charge transfer

The degree of charge transfer (Z) from donor to acceptor, giving compositions DZþAZ (where 0 Z 1), is a quantity of prime importance in the discussion of delocalized charge transfer molecular compounds. The lattice energy, the conductivity, the nature of a possible Peierls transition, the type of ESR spectrum, are all heavily dependent on this parameter. A number of methods of general applicability have been used for determining Z. One method relies on diffraction studies and uses the differences in bond lengths between neutral and ionic forms of donor and acceptor moieties; two other methods rely on spectroscopic studies, one using the dependence of selected stretching frequencies on moiety charge, while the other uses the dependence of the oscillator strength of the charge transfer band on moiety charge. Photoelectron spectroscopy has been used to demonstrate the presence of two charge states but is not suitable for quantitative determination of Z. The most accurate method is based on analysis of diffuse x-ray (or neutron) scattering and will be discussed in Chapter 17. Dependence of moiety bond lengths on charge is shown in Table 13.6 for some donor and acceptor moieties. Clearly TMPD becomes more quinonoid and TCNQ more benzenoid on passing from neutral molecule to ion; this is in accordance with theoretical calculations for TMPD (Haddon, 1975) and TCNQ (Johanson, 1975). Four equations have been proposed for estimation of Z from changes in various bond lengths in TCNQ ((1) Flandrois and Chasseau, 1977; (2) Umland et al., 1988 extending an earlier proposal of Coppens and Guru Row, 1978; (3) Kistenmacher et al., 1982; (4) derived here from values in Table 13.6): ZTCNQ ¼ 7:25ðb cÞ 8:07ðc dÞ 1 ZTCNQ ¼ 26:24 29:92½ða þ cÞ=ðb þ dÞ

ð1Þ ð2Þ

ZTCNQ ¼ 19:83 41:67c=ðb þ dÞ ZTCNQ ¼ 1:374 þ 8:13fðb þ dÞ ða þ cÞg

ð3Þ ð4Þ

There seems little reason to prefer one or other of these equations at the available levels of precision of bond length measurements. One should note that all the equations assume a linear dependence of Z on dimensions, which is surely an over simplification. The values of the coefficients depend on the bond lengths used for neutral and ionic moieties, only one set being available for the neutral molecule but a number for the ionic moiety as it appears in different closed-shell salts; presumably it would be best to take a weighted average of the latter but such sophistication seems premature at the available levels of precision. It is possible to identify the most sensitive parameters (‘discriminators’) from the dimensions given in Table 13.6. We illustrate some of the problems by using {[TTF][TCNQ]} as an example, its crystal structure having been determined at a number of temperatures. We have used bonds (a) and (b) of TTF as standards although it has been suggested by Mayerle et al., (1979) that the ring double bond (d) is more sensitive, based ˚ in {[TTF][TCNQ]} and 1.336 A ˚ in TTFI0.71. Parenthetically, on values of 1.323(4) A we had considered TTFClO4 for use as our reference cation because four independent values are available for each bond length; however, using the ring double bonds as an ˚ , with a clearly example, we find values of 1.292(20), 1.305(15), 1.306(17) and 1.321(13) A unacceptable spread.


945

Table 13.6. Dimensions of neutral and ionic forms of representative donor and acceptor moieties. The ionic forms are for salts where the counterion is a closed shell ion with an assumed integral ˚ ) are not corrected for thermal motion and have been averaged single charge. The bond lengths (A assuming D2h symmetry. Values in brackets are experimental standard uncertainties divided by 1 ðn 1Þ2 where n independent values have been averaged. l ¼ lion–lneutral. Equations showing the dependence of moiety charge (Z) on bond lengths (a, b, c, d ) are given; the constants have dimensions such that Z is dimensionless. The values of a etc. in the equation for Z are those for the species under consideration Moiety

Bond

l (neutral molecule)

l (ion)

l

TMPD

a b c

A (IK79) 1.390(5) 1.397(5) 1.418(6)

B (dBV72) 1.363(2) 1.428(2) 1.345(2)

0.026(5) þ0.031(5) 0.073(6)

a b c d

C (CK71) 1.349(3) 1.757(2) 1.730(2) 1.314(3)

D (YN80) 1.404(10) 1.713(3) 1.725(4) 1.306(8)

þ0.055 0.042 0.005 0.008

a b c d C N

E (LST65) 1.346(3) 1.448(2) 1.374(3) 1.441(3) 1.140(2)

F (HSV72) 1.373(1) 1.423(3) 1.420(0) 1.416(8) 1.153(2)

0.027(3) þ 0.025(3) 0.046(2) þ 0.025(6) þ 0.013(3)

a b

Me2N

c

NMe2

ZTMPD ¼ 10.85 7.69(c þ a b) TTF S S

Sc b d a S

ZTTF ¼ 3.835 9.174(b a) TCNQ NC

c a

NC

b

CN d CN

See text for equations for ZTCNQ. Notes: A. TMPD0 at 298K; B. TMPDClO4 at 110K; C. TTF0; D. TTFClO4; E. TCNQ0 at 300K, libration corrected; F. Rb TCNQ at 113K. References: CK71 – Cooper, Kenny, Edmonds, Nagel, Wudl and Coppens, 1971; dBV72 – de Boer and Vos, 1972; HSV72 – Hoekstra, Spoelder and Vos, 1972; I79 – Ikemoto, Katagiri, Nishimura, Yakushi and Kuroda, 1979; LST65 – Long, Sparks and Trueblood, 1965; YN80 – Yakushi et al. 1980.

We illustrate possible use of discriminator bonds (a) and (b) of TTF and (c) and (d) of TCNQ by plotting these in Fig. 13.6, together with ‘standard’ values (300K) for the neutral and ionic moieties. The differences between paired bond lengths in the neutral and ionic moieties are only a few times the standard uncertainty of the measurements. If we make the approximation (permissible in terms of the available precision) that bond lengths are not dependent on temperature, then using mean values we calculate that jZTCNQj ¼ 0.41 (eq. 1), 0.37 (eq. 2), 0.44 (eq. 3) and 0.39 (eq. 4). Although the internal agreement is satisfactory, the accepted value of jZj (from X-ray diffuse scattering


946

TTF 1.42 1.40

Bond (a)

cation

TCNQ 1.46

1.38

Bond (d) 1.44

1.36

neutral neutral

1.42

anion

neutral

1.42

anion

1.34 1.75

1.40

1.74

1.38

1.73

1.36

1.72

Bond (b)

cation 1.71

Bond (c)

0

100

neutral 0

100 200 300 T (K)

200 300 T (K)

˚ ) of two discriminator bond lengths in each of Fig. 13.6. The ordinates show measured values (A the TTF and TCNQ moieties of {[TTF][TCNQ]} at 300K (Kistenmacher et al., 1974) and at 100, 60, 53 and 45K (Schulz et al., 1976). The standard values deduced for these bonds in the anionic and neutral species are shown at the left and right extremes of the temperature scale and the measured values at the appropriate temperatures. The horizontal broken lines show the bond lengths corresponding to Z ¼ 0.59 deduced from x-ray diffuse scattering measurements. See text for discussion.

(Section 17.7)) is 0.59. It is clear that much higher precision is required for all bond length measurements if meaningful conclusions are to be extracted. The spectroscopic methods have been applied mainly to TTF and TCNQ, and we summarize the conclusions in Table 13.7. Again a linear dependence of Z on frequency is assumed. For example, resonance Raman measurements of the frequency of the exocyclic C¼C bond stretch ( 4) show this to be 1454 cm1 for neutral TCNQ (Van Duyne et al., 1979) and 1379 cm1 for TCNQ (Matsuzaki, Kuwata and Toyoda, 1980). Linear interpolation gives ZTCNQ ¼ –0.01333 C¼C þ 19.39. As the resolution of the spectroscopic methods is 1–3 cm1, a precision of 0.01 in Z can be expected. However, solid-state (crystal field) effects can alter vibration frequencies by 10 cm1. We also note that (C N) is very sensitive to the degree of charge transfer, in contrast to d(C N). Photoelectron spectroscopy using deconvolution of N-1s and S-2p spectra demonstrates the presence of neutral and charged moieties in segregated stack molecular compounds. This is illustrated for {[TMTTF][TCNQ]} in Fig. 13.7 (Tokumoto, Koshizuka, Murata, Kinoshita, Anzai, Ishiguro and Mori, 1982; Tokumoto, Koshizuka, Anzai and Ishiguro, 1982). Unfortunately the method does not give reliable quantitative results (Ritsko et al., 1978).


947

Table 13.7. Determination of degree of charge transfer by spectroscopic techniques (neutral species) cm1

(ionic species) cm1

(Charge transfer)/frequency relationship

TTF Raman, ag 3, predominantly C¼C stretch 1515 (S80) 1413 (MMT80)

ZTTF ¼ –0.00980 C¼C þ 14.85.

TCNQ Resonance Raman, 4 exocyclic C¼C stretch 1454 (VD79) 1379 (MKT80)

ZTCNQ ¼ –0.01333 C¼C þ 19.39

TCNQ IR Absorption C N stretch 2227 (CB81) 2183 (CB81)

ZTCNQ ¼ –0.02273 C N þ 50.61

TMTTF-TCNQ

TMTTF+

S 2p

TMTTF0

170 165 BINDING ENERGY (eV)

EMISSION INTENSITY (ARB. UNITS)

EMISSION INTENSITY (ARB. UNITS)

References: CB81 – Chappell et al., 1981; MKT80 – Matsuzaki, Kuwata and Toyoda, 1980; MMT80 – Matsuzaki, Moriyama and Toyoda, 1980; S80 – Siedle et al., 1980; VD79 – Van Duyne et al., 1979.

TMTTF-TCNQ

N Is

TCNQ– TCNQ0

405

400 BINDING ENERGY (eV)

Fig. 13.7. Core level XPS spectra for {[TMTTF][TCNQ]} after data processing. Observed spectra are given by dotted lines, simulated spectra by full lines, and the individual moiety spectra by chain lines. The latter are obtained from TMTTF, {[TMTTF][(DDQ)2]}, TCNQ and K-TCNQ. (Reproduced from Tokumoto et al., 1982.)

It has also been suggested that the value of Z can be obtained from measurement of the oscillator strength ( f ) of the charge transfer band (Jacobsen and Torrance, 1983). The relation between them is f ¼ f82 mc CT j RDA j2 Zg=h ¼ 3:254 105 CT j RDA j2 Z ˚ ) between the centers of the donor and acceptor molecules where RDA is the distance (A and CT is the excitation energy of the charge transfer band in cm1. The oscillator strength is measured from Z 4mc 2 ð Þ cos2 d ; f ¼ Ne2 1

948


where N is the number of DA pairs per unit volume and is the angle between the direction of light polarization and the transition dipole moment (Wooten, 1972). Results obtained by this method are given in Section 17.4 to demonstrate that a neutral to ionic phase transition occurs in {TTF chloranil}. Some care is needed in using the values of Z derived by these methods as systematic errors can occur. Cross checking of one method against another is always desirable. n n n

13.4 13.4.1

Binary and quasi-binary donor–acceptor systems Phase diagrams

Most charge transfer molecular compounds have been prepared by what are essentially hit-or-miss techniques – mix the components in a particular ratio in a suitable solvent and hope that the binary compound will crystallize. However, careful studies, involving determination of the phase diagram of the system, are now becoming more common. Ideally both the binary phase diagram, which covers a wide range of temperatures, and the ternary (D-A-solvent) phase diagram, at a particular or limited number of temperatures, should be determined, but this has rarely been done. The ternary phase diagram will give more reliable results in the lower temperature regions where solid–solid transitions are likely to be slow. In older work melting point diagrams were determined, perhaps with the addition of thaw points; in modern work use of differential scanning calorimetry (DSC) has greatly improved both the precision and the reliability of the results. Quite complicated binary diagrams can sometimes be obtained, as exemplified by that of carbazole/ 2,4,7-trinitrofluoren-9-one (Fig. 13.8; Krajewska and Pigon, 1980); we maintain our D/A nomenclature despite the converse usage in Fig. 13.8. This shows such typical features as a 1 : 1 molecular compound with congruent melting point and a 1 : 2 molecular compound

500 T (K)

450

TNF

2:1

1:1

carbazole

mole per cent

Fig. 13.8. Phase diagram for the carbazole-TNF system. Liquidus and eutectic lines; peritectic lines; --ø--ø-- phase transition line. (Reproduced from Krajewska and Wasilewska, 1981.)

B I NA R Y A N D Q UA S I- B I NA R Y D O N O R – A C C E P T O R S YS T E M S

I

IX

140

L

II

Temperature, °C

A II

30

10

20

30

10

20

30

10

20

30

10

20

30

I

100 I

20

P P

P

P

10 P

120 I

949

II

I,I I

A

80 II II,II II

60 P

I

II

A

40 A A 0

20

40 60 Mol %

80

100

2

Fig. 13.9. (a) on left: Phase diagram for the mixed donor system (phenanthrene/anthracene)–picric acid: x single phase from Debye–Scherrer patterns; two phase from Debye–Scherrer patterns. Mol.% anthracene runs from zero on the left to 100% on the right. Greek letters (, , ) indicate phases stable at increasing temperatures; (b) on right: Debye–Scherrer patterns (Cu K) from the phases found in the phase diagram. In order from top to bottom these show (i) the three phases of phenanthrene-picric acid and solid solutions (ii) the three patterns from the phases of intermediate solid solution I (iii) the three patterns from the phases of intermediate solid solution II (iv) the two patterns from anthracene-picric acid and solid solutions. The very simple patterns of P and I have been interpreted as indicating that these are plastic phases (see Section 16.5). (Reproduced from Koizumi and Matsunaga, 1974.)

with incongruent melting point (in some systems ratios other than 1 : 1 or 1 : 2 are found but these are relatively rare). The 1 : 1 compound is more stable than the 1 : 2 compound, which is the usual situation. The composition ranges of both compounds appear negligibly small (although it should be remembered that the methods used are generally not sensitive to composition differences of less than 1%). The 1 : 1 compound shows a solid state transition at about 460K; transitions below room temperature are also often found (see Chapter 16). Despite the improvements in methodology, controversy has not been entirely eliminated; for example, a phase diagram has been reported for the pyrene/picryl chloride system which shows DA, D4A3, D2A and D4A compounds (only DA melts congruently) (Bando and Matsunaga, 1976) while other workers could find only DA and D3A2 compounds (Krajewska and Wasilewska, 1981). Sluggishness in the attainment of equilibrium and problems of identification of incongruently melting compounds from small breaks in the liquidus curves can give rise to many difficulties. The binary phase diagram summarizes an appreciable amount of information about a system, and is always a desirable preliminary to a detailed study of the structure and/or


950

L

L

S 200

t /°C

200 S

S

100

S

100 S

Py-TNB

Mol%

Py-TNP FI-TNB

FI-TNP

Mol%

Fig. 13.10. Phase diagrams for the mixed acceptor systems pyrene–(TNB, TNP) and fluoranthene– (TNB, TNP) (TNP ¼ 2,4,6-trinitrophenol or picric acid). The crystal structure of pyrene TNB is known. (Reproduced from Inabe et al., 1981.) n n n

properties of a particular charge transfer molecular compound. Many such diagrams are available for mixed stack systems (for examples, see Herbstein, 1971; D’Ans and Kaufmann, 1956; Kofler, 1956; Radomska and Radomski, 1980a, b). There do not appear to be any phase diagrams for segregated stack systems but there are some for ‘‘complex isomers’’ e.g. 1-bromo-2-naphthylamine/picric acid (Hertel, 1926); 4-bromo-1-naphthylamine/ 2,6-dinitrophenol (Hertel and Van Cleef, 1928); o-bromoaniline/picric acid (Komorowski et al., 1976). Although the donor : acceptor ratio appears to be fixed at a ratio of small integers in the charge transfer molecular compounds, it is possible to replace to a considerable extent one donor by another, and the same holds for suitable pairs of acceptors. Thus about 60% of the phenanthrene molecules can be replaced by anthracene in the {phenanthrene TNB} compound, and about 20% of the anthracene molecules in {anthracene TNB} can be replaced by phenanthrene (Lower, 1977). Similar extensive mutual solid solubility of anthracene and phenanthrene has been found in the (anthracene–phenanthrene)/TCNB (Wright et al., 1976) and (anthracene–phenanthrene)/picric acid (Koizumi and Matsunaga, 1974) systems. In the first of these there is complete mutual solid solubility of the two donors and the crystal structures of {anthracene TCNB} and {phenanthrene TCNB} are isostructural at room temperature (there is disorder of the phenanthrene molecule). In the second system (Fig. 13.9) the situation appears to be more complicated; in addition to limited solid solubility of anthracene in {phenanthrene picric acid} and of phenanthrene in {anthracene picric acid} two intermediate phases with fairly wide composition ranges appear. Only the crystal structure of {anthracene picric acid} is known (Herbstein and Kaftory, 1976). Among the mixed acceptor systems there are some (Fig. 13.10) with a full range of solid solubility at lower temperatures as well as at higher temperatures; in accordance with these phase diagrams the {pyrene picric n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

B I NA R Y A N D Q UA S I- B I NA R Y D O N O R – A C C E P T O R S YS T E M S

200

t /°c

L

L

L

S

100

951

S S

0 Py-NPA

Py-DNF

200

Mol%

L

L

S

S

Py-TNT

FI-TNT

L

t /°c

S

100

0 FI-TNT

FI-TNP

Ph-TNP

Ph-TNB

Mol%

Fig. 13.11. Phase diagrams for the systems Py-(NPA, DNF); Py-(DNF, TNT); (Py, Fl)-TNT; Fl-(TNT, TNP); (Fl, Ph)-TNP; and Ph-(TNP, TNB). Py ¼ pyrene; Fl ¼ fluoranthene; Ph ¼ phenanthrene; TNP ¼ 2,4,6-trinitrophenol (picric acid); NPA ¼ 2-nitrophthalic anhydride; DNF ¼ 2,4-dinitrofluorobenzene. (Reproduced from Inabe et al., 1981.)

acid} and {pyrene TNB} compounds were reported to be isomorphous at room temperature, and also the pair {fluoranthene picric acid} and {fluoranthene TNB} (Herbstein and Kaftory, 1975a). Other systems (Fig. 13.11) show complete miscibility only at high temperatures and complicated diagrams at lower temperatures. There are many indications of solid state transitions. Some information about relevant crystal structures is noted in the captions to these diagrams; references are given in Chapter 15. The phase diagrams for the mixed acceptor systems (Fig. 13.10 and 13.11) are compatible with the more limited information available from crystallographic studies despite the different temperature ranges of the two types of study. It is not known whether there is ordered or random substitution of one donor (or acceptor) for another in the mixed stacks of these molecular compounds. n n n

n n n

n n n


952

One ternary mixed-donor system has been studied – (anthracene, acridine, phenazine)/ PMDA (Karl et al., 1982). The three individual molecular compounds ({anthracene PMDA}, {acridine PMDA}, {phenazine PMDA}) are isomorphous (triclinic, P 1, Z ¼ 1; Table 15.2, Group 1a). Complete miscibility was found for all donor ratios over the temperature range 500–120K, the molecular compounds melting around 500K. n n n

n n n

13.4.2

n n n

Component ratios in binary donor–acceptor systems

Most crystalline -molecular compounds, be they mixed stack or segregated stack in structure, have a 1 : 1 donor : acceptor ratio and this has been the most extensively studied group, both in regard to crystal structures and physical properties. Many 1 : 2 and 2 : 1 compositions have been reported, and most of these occur as incongruently-melting compounds in systems where the 1 : 1 composition is the most stable. However, there are examples where the 1 : 2 (or 2 : 1) molecular compound appears to be the more stable e.g. (benzo[c]pyrene)2 TMU (Weil-Malherbe, 1946); relative stabilities of the different compositions in such systems do not appear to have been investigated. There are also reports of compositions other than 1 : 1, 1 : 2 or 2:1 (Table 13.8); these should be viewed with caution if based on chemical analyses alone. However, there are structural explanations for some unusual compositions. For example, in both n n n

Table 13.8. Some examples of compositions other than 1 : 1, 1 : 2 or 2 : 1 reported for -molecular compounds Donor (D)

Acceptor (A)

D:A

Reference

Bromodurene Tetralin Pyrene N,N-dibenzyl-m-toluidine Triphenylmethanol Phenanthrene Fluorene 1-Naphthylamine 2-Naphthylamine Phenanthrene Naphthalene 1-Naphthol p-Phenylenediamine Dibenzo[c,d]phenothiazine Benzene 1,4-Diphenylbutadiene Tetrathiotetracene(TTT) TTT TTT

BTF Nitrobenzodifuroxan PMDA TNB TNB p-Dinitrobenzene TNB 2,3-Dinitrophenol 2,3-Dinitrophenol 1,2,4,6-Tetranitrobenzene 1,2,4,6-Tetranitrobenzene Tetryl p-Benzoquinone DDQ o-Chloranil TNF o-Chloranil o-Bromanil TCNE

3:2 1:3 1:3 3:2 3:2 3:1 3:4 3:2 3:2 2:3 3:2 3:2 2:5 3:2 3:1 3:1 3:1 3:1 3:1

B60 BC58 HS69 KMC39 KHM12 K08 HKR76 S40b S40b S40a S40c S40c S09 M64 PJF17 OW46 M65 M64 M64

References: B60 – Bailey, 1960; BC58 – Bailey and Case, 1958; HKR76 – Herbstein, Kaftory and Regev, 1976; HS69 – Herbstein and Snyman, 1969; K08 – Kremann, 1908; KHM12 – Kremann, Hohl and Muller, 1912; KMC39 – Kent, McNeil and Cowper, 1939; M64 – Matsunaga, 1964; M65 – Matsunaga, 1965; OW46 – Orchin and Woolfolk, 1946; PFJ17 – Pfeiffer, Jowleff, Fischer, Monti and Muuly, 1917; S09 – Schlenk, 1909; S40a – Shinomiya, 1940a; S40b – Shinomiya, 1940b; S40c – Shinomiya, 1940c.

TERNARY -MOLECULAR COMPOUNDS

953

(pyrene)3 (picryl bromide)2 (Herbstein and Kaftory, 1975b) and (TMTTF)1.8 TCNQ (Kistenmacher et al., 1976) the molecules present over and above the 1 : 1 composition are included in the structure without participating in the charge transfer interaction, i.e. they are present as ‘‘molecules of crystallization’’. The remarkable structure of {(fluorene)3 (TNB)4} is discussed in Section 15.4. The segregated-stack radical-cation, radical-anion salts seem all to have equimolar compositions but the Mn(TCNQ)m salts, where M is a closed-shell cation, have a wide range of compositions depending on the nature of the cation. There are good structural explanations for compositions such as Cs2(TCNQ)3 (see Section 17.4.6) or N-(n-butylpyridinium)4(TCNQ)7 (see Section 17.4.2.3). n n n

n n n

n n n

13.5 Ternary -molecular compounds There are a number of molecular compounds which contain three components and we can distinguish three situations: (i) Noninteracting third component, where the third component appears to be present essentially as ‘‘molecules of crystallization’’ and does not participate in the charge transfer interaction. However, no relevant structures have been reported so some reserve must be maintained about examples such as azulene BTF(propionic acid)0.5 (Bailey, 1960); 2,2 0 ,4,4 0 hexamethylstilbene (picric acid)2benzene (Elbs, 1893); 1-naphthol hexachloro-1-indenone1/2X where X ¼ benzene or acetic acid (Pfeiffer et al., 1924); phenanthrene TNBx(benzene) and tetrabenznaphthalene picric acid. ethanol, where the third component may be present zeolitically in both examples (Herbstein, Kaftory and Regev, 1976). (ii) Interacting third component: one example where all three components interact is Kofler’s (1944) 1 : 1 : 1 ternary compound {1-naphthylamine pyridine picric acid}, the structure of which has been determined (Bernstein et al., 1980; see Section 15.11.4). A second example is the ordered ternary compound {3,3 0 -dimethylthioazolinocarbocyanine TCNQ 2,4,7-trinitrofluorenone} (Kaminskii et al., 1974), with mixed stacks of alternating TCNQ and TNF moieties. The pseudo-binary systems with mixed donors or acceptors presumably should be classified in this group. n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

Table 13.9. Examples of TNB molecular compounds formed by a particular donor and by its potassium salt. Melting points in C Donor

o-Aminobenzoic Acid m-Aminobenzoic Acid p-Aminobenzoic Acid 1-Anthrol

TNB molecular compound

TNB molecular compound of K salt

Crystals

M. Pt.

Crystals

M. Pt.

orange needles

192

red red-brown plates

151 161

deep-red needles red-brown needles red needles black needles

114 118 115(dec) 275

Notes: All the compounds are 1 : 1 (Sudborough and Beard, 1910) except for the K salt of anthrol. TNB which has composition C14H9OK {C6H3(NO2)3}2 (Cadre and Sudborough, 1916). n n n

954


(iii) There is another group of molecular compounds where the third component, although not participating in the charge transfer system, exerts a profound influence on the structure and its properties. These are the benzidine TCNQ.solvent compounds discussed in Section 15.7.3.3, which have both charge transfer and inclusion characteristics. (iv) Finally we draw attention to a group where both a particular donor, and its potassium salt, form molecular compounds with TNB (Table 13.9). As both neutral compounds and the salts are highly colored, it seems probable that charge transfer interactions occur in both. These pairs of compounds could provide an opportunity to compare the donor characteristics of molecules and their anions. n n n

References Andre`, J. J., Bieber, A. and Gautier, F. (1977). Ann. Phys. (Paris), 1, 145–256. Andrews, L. J. and Keefer, R. M. (1964). Molecular Complexes in Organic Chemistry, Holden-Day, San Francisco. Bailey, A. S. and Case, J. R. (1958). Tetrahedron, 3, 113–131. Bailey, A. S. (1960). J. Chem. Soc., pp. 4710–4712. Bando, M. and Matsunaga, Y. (1976). Bull. Chem. Soc. Jpn., 49, 3345–3346. Baril, O. L. and Hauber, E. S. (1931). J. Am. Chem. Soc., 53, 1087–1091. Batail, P., La Placa, S. J., Mayerle, J. J. and Torrance, J. B. (1981). J. Am. Chem. Soc., 103, 951–953. Batley, M. and Lyons, L. E. (1962). Nature (London), 196, 573–574. Batley, M. and Lyons, L. E. (1968). Mol. Cryst. Liq. Cryst., 3, 357–374. Bennett, G. M. and Willis, G. H. (1929). J. Chem. Soc., pp. 256–268. Bernstein, J., Herbstein, F. H. and Regev, H. (1980). Acta Cryst., B36, 1170–1175. Blaunstein, R. P. and Christopherou, L. G. (1971). Rad. Res. Rev., 3, 69–118. Bleidelis, J., Shvets, A. E. and Freimanis, J. (1976). Zh. Strukt. Khim., 17, 1096–1110. Boer, J. L. de and Vos, A. (1972). Acta Cryst., B28, 839–848. Booth, J. and Boyland, E. (1953). Biochim. Biophys. Acta, 12, 75–87. Boschi, R., Clar, E. and Schmidt, W. (1974). J. Chem. Phys., 60, 4406–4418. Brackman, W. (1949). Rec. Trav. Chim., 68, 147–159. Briegleb, G., Fo¨rster, Th., Friedrich-Freksa, H., Jordan, R., Kortu¨m, G., Mu¨nster, A., Scheibe, G. and Wirtz, K. (1948). Zwischenmolekulare Krafte, Karlsruhe, G. Braun, 142 pp. Briegleb, G. (1961). Elektronen-Donator-Acceptor-Komplexe. Springer, Berlin-Go¨ttingen-Heidelberg. Briegleb, G. (1964). Angew. Chem. Int. Ed. Engl., 3, 617–632. Cadre, S.T. and Sudborough, J. J. (1916). J. Chem. Soc., 109, 1349–1354. Carstensen-Oeser, E., Go¨ttlicher, S. and Habermehl, G. (1968). Chem. Ber., 101, 1648–1655. Chappell, J. S., Bloch, A. N., Bryden, W. A., Maxfield, M., Poehler, T. O. and Cowan, D. O. (1981). J. Am. Chem. Soc., 103, 2442–2443. Chen, E. C. M. and Wentworth, W. E. (1975). J. Chem. Phys., 63, 3183–3191. Clar, E. and Schmidt, W. (1977). Tetrahedron, 33, 2093–2097. Clar, E. and Schmidt, W. (1979). Tetrahedron, 35, 2673–2680. Clark, P. A., Brogli, F. and Heilbronner, E. (1972). Helv. Chim. Acta, 55, 1415–1428. Clayden, J., Greeves, N., Warren, S. and Wothers, P. (2001). Organic Chemistry. Oxford University Press, Oxford. 1512 pp. Coleman, L. B., Cohen, M. J., Sandman, D. J., Yamagishi, F. G., Garito, A. F. and Heeger, A. J. (1973). Sol. State Commun., 12, 1125–1132.

REFERENCES

955

Cooper, W. F., Kenny, N. C., Edmonds, J. W., Nagel, A., Wudl, F. and Coppens, P. (1971). J. Chem. Soc., Chem. Commun., pp. 889–890. Coppens, P. and Guru Row, T. N. (1978). Ann. N.Y. Acad. Sci., 313, 244–255. D’Ans, J. and Kaufmann, E. (1956). ‘‘Solubility equilibria of organic substances in organic substances’’ in Landolt-Bernstein Tables, Springer-Verlag, Berlin-Go¨ttingen-Heidelberg, Sixth Edition, Vol. II, Part 2, pp. 1–500. Dahm, D. J., Horn, P., Johnson, G. R., Miles, M. G. and Wilson, J. D. (1975). J. Cryst. Mol. Struct., 5, 27–34. Davis, K. M. C. and Symons, M. C. R. (1965). J. Chem. Soc., pp. 2079–2083. Eilfeld, P. and Schmidt, W. (1981). J. Electr. Spectroscop. Related Phenom., 24, 101–120. Elbs, K. (1893). J. prakt. Chem. (N.F.), 47, 44–79. Fieser, L. F and Fieser, Mary. (1961). Advanced Organic Chemistry. Reinhold, New York. 1155 pp. Flandrois, S. and Chasseau, D. (1977). Acta Cryst., B33, 2744–2750. Fritzsche, J. von (1858). J. prakt. Chem., 73, 282–292. Gleiter, R. and Schanz, P. (1980). Angew. Chem. Int. Ed. Engl., 19, 715–716. Gutmann, F. and Lyons, L. E. (1981). Organic Semiconductors, Part A, original edition published by Wiley, New York and reprinted, with corrections, by R. E. Kriejer, Malabar, Florida. Haddon, R. C. (1975). Austr. J. Chem., 28, 2333–2342. Hendrickson, J. B., Cram, D. J. and Hammond, G. S. (1970). Organic Chemistry (Third edition). McGraw-Hill Kogakusha. Tokyo. 1279 pp. Herbstein, F. H. (1971). ‘‘Crystalline -molecular compounds: chemistry, spectroscopy and crystallography’’ in Perspectives in Structural Chemistry, edited by J. D. Dunitz and J. A. Ibers, Wiley, New York etc., Vol. IV, pp. 166–395. Herbstein, F. H. and Kaftory, M. (1975a). Acta Cryst., B31, 60–67. Herbstein, F. H. and Kaftory, M. (1975b). Acta Cryst., B31, 68–75. Herbstein, F. H. and Kaftory, M. (1976). Acta Cryst., B32, 387–396. Herbstein, F. H. and Snyman, J. A. (1969). Phil. Trans. Roy. Soc. Lond., A264, 635–666. Herbstein, F. H., Kaftory, M. and Regev, H. (1976). J. Appl. Cryst., 9, 361–364. Hertel, E. and van Cleef, J. (1928). Ber., 61, 1545–1549. Hertel, E. (1926). Ann. Chem., 451, 179–208. Hoekstra, A., Spoelder, T. and Vos, A. (1972). Acta Cryst., B28, 14–25. Ikemoto, I., Katagiri, G., Nishimura, S., Yakushi, K. and Kuroda, H. (1979). Acta Cryst., B35, 2264–2265. Inabe, T., Matsunaga, Y. and Nanba, M. (1981). Bull. Chem. Soc. Jpn., 54, 2557–2564. Jacobsen, C. S. and Torrance, J. B. (1983). J. Chem. Phys., 78, 112–115. Johanson, H. (1975). Int. J. Quantum Chem., 9, 459–471. Kaminskii, V. F., Shibaeva, R. P. and Atovmyan, L. O. (1974). J. Struct. Chem., 15, 434–440. Karl, N., Ketterer, W. and Stezowski, J. J. (1982). Acta Cryst., B38, 2917–2919. Karl, N., Sato, N., Seki, K. and Inokuchi, H. (1982). J. Chem. Phys., 77, 4870–4878. Kent, A., McNeil, D. and Cowper, R. M. (1939). J. Chem. Soc., pp. 1858–1862. Kistenmacher, T. J., Emge, T. J., Bloch, A. N. and Cowan, D. O. (1982). Acta Cryst., B38, 1193–1199. Kistenmacher, T. J., Phillips, T. E. and Cowan, D. O. (1974). Acta Cryst., B30, 763–768. Kistenmacher, T. J., Phillips, T. E., Cowan, D. O., Ferraris, J. P., Bloch, A. N. and Poehler, T. O. (1976). Acta Cryst., B32, 539–547. Kofler, A. (1956). ‘‘Melting-point Equilibria in Organic Systems’’ in Landolt-Bernstein Tables, Springer-Verlag, Berlin-Go¨ttingen-Heidelberg, Sixth Edition, Vol. II, Part 3, pp. 350–403. Kofler, A. (1944). Z. Elektrochem., 50, 200–207. Koizumi, S. and Matsunaga, Y. (1974). Bull. Chem. Soc. Jpn., 47, 9–13. Komorowski, L., Krajewska, A. and Pigon, K. (1976). Mol. Cryst. Liq. Cryst., 36, 337–348.

956


Konno, M., Saito, Y., Yamada, K. and Kawazura, H. (1980). Acta Cryst., B36, 1680–1683. Krajewska, A. and Pigon, K. (1980). Thermochim. Acta, 41, 187–197. Krajewska, A. and Wasilewska, quoted by Pigon, K. and Chojnocki, H. (1981). ‘‘Electrical conductivity of solid molecular complexes’’ in Molecular Complexes, ed. by H. Ratajczak and W.J. Orville-Thomas, Vol. 2, pp. 451–492. Kremann, R. (1908). Monatsh., 29, 863–890. Kremann, R., Hohl, H. and Muller, R., II. (1912). Monatsh., 42, 199–220. Long, R. E., Sparks, R. A. and Trueblood, K. N. (1965). Acta Cryst., 18, 932–939. Lower, S. K. (1968). Mol. Cryst. Liq. Cryst., 5, 363–368. Lowry, T. M. (1924). Chem. and Ind., 48, 218–221. Mackay, I. R., Robertson, J. M. and Sime, J. G. (1969). J. Chem. Soc., Chem. Commun., pp. 1470–1471. Martinet, J. and Bornand, L. (1925). Rev. gen. sci., 36, 569–577; Chem. Abstr., 20, 861 (1926). Matsunaga, Y. (1964). J. Chem. Phys., 41, 1609–1613. Matsunaga, Y. (1965). J. Chem. Phys., 42, 2248–2249. Matsuo, T. and Higuchi, O. (1968). Bull. Chem. Soc. Jpn., 41, 518–519. Matsuzaki, S., Kuwata, R. and Toyoda, K. (1980). Sol. State Commun., 33, 403–405. Matsuzaki, S., Moriyama, T. and Toyoda, K. (1980). Sol. State Commun., 33, 857–859. Mayerle, J. J., Torrance, J. B. and Crowley, J. I. (1979). Acta Cryst., B35, 2988–2995. McConnell, H. M., Hoffman, B. M. and Metzger, R. M. (1965). Proc. Natl. Acad. Sci. U. S., 53, 46–50. Meyers, M. and Trueblood, K. N. (1969). Acta Cryst., B25, 2588–2599. Mulliken, R. S. (1952a). J. Am. Chem. Soc., 72, 600–608. Mulliken, R. S. (1952b). J. Phys. Chem., 56, 801–822. Mulliken, R. S. (1954). J. chim. Phys., 51, 341–344. Mulliken, R. S. and Person, W. B. (1969a). J. Am. Chem. Soc., 91, 3409–3413. Mulliken, R. S. and Person, W. B. (1969b). Molecular Complexes. Wiley-Interscience, New York, 498 pp. Mueller-Westerhoff, U. T. and Vance, B. (1987). Dithiolenes and Related Species, Chapter 16.5 in Comprehensive Coordination Chemistry, Vol. 2, 595–631, edited by G. Wilkinson, R. D. Gilard and J. A. McCleverty, Pergamon Press, Oxford etc. Newman, M. S. and Lutz, W. B. (1956). J. Am. Chem. Soc., 78, 2469–2473. Orchin, M. and Woolfolk, E. O. (1946). J. Am. Chem. Soc., 68, 1727–1729. Pfeiffer, P., Goebel, F. and Angern, O. (1924). Ann. Chem., 440, 241–264. Pfeiffer, P., Jowleff, W., Fischer, P., Monti, P. and Mully, H. (1917). Ann. Chem., 412, 253–335. Pfeiffer, P. (1927). OrganischeMolekulverbindungen. 2nd Edition. F. Enke, Stuttgart. Ponte Goncalez, A. M. (1980). Prog. Sol. State Chem., 13, 1–88. Powell, H. M., Huse, G. and Cooke, P. W. (1943). J. Chem. Soc., pp. 153–157. Prout, C. K. and Castellano, E. E. (1970). J. Chem. Soc. (A), pp. 2775–2778. Prout, C. K. and Wright, J. D. (1968). Angew. Chem., Int. Ed. Engl., 7, 659–667. Prout, C. K. and Kamenar, B. (1973). In Molecular Complexes 1, 151–207. Elek: London. Radomska, M. and Radomski, R. (1980a). Thermochim. Acta, 40, 405–414. Radomska, M. and Radomski, R. (1980b). Thermochim. Acta, 40, 415–425. Ritsko, J. J., Epstein, A. J., Salaneck, W. R. and Sandman, D. J. (1978). Phys. Rev., B17, 1506–1509. Roberts, J. D. and Caserio, M. C. (1964). Basic Principles of Organic Chemistry. Benjamin, New York. 1315 pp. Rose, J. (1967). Molecular Complexes. Pergamon Press, Oxford. Rosenstock, H. M., Sims, D., Shroyer, S. S. and Webb, W. (1980). J. National Standard Reference Data Series (U.S. Nat. Bur. Stands.) part I, no. 66. Saito, G. and Ferraris, J. P. (1979). J. Chem. Soc.Chem. Commun., pp. 1027–1029.

REFERENCES

957

Sandman, D. J., Stark, J. C., Hamill, G. P., Burke, W. A. and Foxman, B. M. (1982). Mol. Cryst. Liq. Cryst., 86, 1819–1825. Sato, N., Seki, K. and Inokuchi, H. (1981). J. Chem. Soc., Faraday Trans. 2, 77, 1621–1633. Schlenk, W. (1909). Ann. Chem., 368, 277–395. Schmidt, W. (1977).J. Chem. Phys., 66, 828–845. Schulz, A. J., Stucky, G. D., Blessing, R. H. and Coppens, P. (1976). J. Am. Chem. Soc., 98, 3194–3201. Shinomiya, C. (1940a). Bull. Chem. Soc. Jpn., 15, 92–103. Shinomiya, C. (1940b). Bull. Chem. Soc. Jpn., 15, 137–147. Shinomiya, C. (1940c). Bull. Chem. Soc. Jpn., 15, 259–270. Siedle, A. R., Kistenmacher, T. J., Metzger, R. M., Kuo, C.-S., Van Duyne, R. P. and Cope, T. (1980). Inorg. Chem., 19, 2048–2051. Silinski, E. A. (1981). Organic Molecular Crystals, Springer Verlag, Berlin. Solomons, T. W. G. and Fryhle, C. B. (2000). Organic Chemistry (Seventh edition). Wiley, New York. 1258 pp. Soos, Z. G. (1974). Ann. Rev. Phys. Chem., 25, 121–153. Soos, Z. G., Kuwajima, S. and Harding, R. H. (1986). J. Chem. Phys., 85, 601–610. Sowada, U. and Holroyd, R. A. (1980). J. Phys. Chem., 84, 1150–1154. Stezowski, J. J., Stigler, R.-D. and Karl, N. (1986). J. Chem. Phys., 84, 5162–5170. Strebel, P. J. and Soos, Z. G. (1970). J. Chem. Phys., 53, 4077–4090. Streitweiser, A., Heathcock, C. H. and Kosower, E. M. (1992). Introduction to Organic Chemistry (Fourth edition). Macmillan, New York. 1256 pp. Sudborough, J.J. and Beard, S.H. (1910). J. Chem. Soc., 97, 773–798. Tamres, M. L. and Strong, R. L. (1979). Mol. Assoc., 2, 331–456. Tokumoto, M., Koshizuka, N., Anzai, H. and Ishiguro, T. (1982). J. Phys. Soc. Jpn., 51, 332–338. Tokumoto, M., Koshizuka, N., Murata, K., Kinoshita, N., Anzai, H., Ishiguro, T. and Mori, N. (1982). Mol. Cryst. Liq. Cryst., 85, 195–202. Torrance, J. B., Vazquez, J. E., Mayerle, J. J. and Lee, V. Y. (1981). Phys. Rev. Lett., 46, 253–257. Umland, T. C., Allie, S., Kuhlmann, T. and Coppens, P. (1988). J. Phys. Chem., 92, 6456–6460. Van Duyne, R. P., Suchanski, M. R., Lakovits, J. M., Siedle, A. R., Parks, K. D. and Cotton, T. M. (1979). J. Am. Chem. Soc., 101, 2832–2837. Weil-Malherbe, H. (1946). Biochem. J., 40, 351–363. Weiss, J. (1942). J. Chem. Soc., pp. 245–252. Wiberg, N. (1968). Angew. Chem. Int. Ed. Engl., 7, 766–779. Wo¨hler, F. (1844). Ann. Chem., 51, 145–163. Wooten, F. (1972). Optical Properties of Solids, Academic Press, New York, p. 72. Wright, J. D., Ohta, T. and Kuroda, H. (1976). Bull. Chem. Soc. Jpn., 49, 2961–2966. Yakushi, K., Nishimura, S., Sugano, T., Kuroda, H. and Ikemoto, I. (1980). Acta Cryst., B36, 358–363.


Chapter 14 Layered molecules with intra-molecular donor–acceptor interactions

. . . lock’d in fierce embrace . . . Unknown.

Summary: Layered molecules (in the present context) are donor–acceptor cyclophanes in which the mutual orientation and/or location of the donor and acceptor ring systems can be altered in order to investigate the conditions for maximum overlap and hence charge transfer. The enhancement of charge transfer in pseudogeminal as opposed to pseudo-ortho cyclophanes has been demonstrated by comparison of UV-visible spectra for many diastereoisomeric pairs. Crystal structure determinations show that there is superposition of donor and acceptor moieties in the paracyclophane series, with appreciably more distortion in [2.2] than in [3.3] paracyclophanes. Charge transfer interactions persist in triply and quadruply layered donor–acceptor paracyclophanes but the orientation effects are lost. The charge transfer spectra of syn and anti diastereoisomers in the metacyclophane series are remarkably similar; crystal structure analyses show considerable intramolecular distortion and mutual displacement and nonparallelism of the donor and acceptor moieties. The ideas developed in the cyclophane studies are beginning to be applied to other areas of chemistry. 14.1 Introduction 14.2 Molecules of the paracyclophane type 14.2.1 Molecules derived from [n.n]paracyclophanes 14.2.2 Systems related to [n.n]paracyclophanes 14.2.3 Multi-layered systems 14.3 Molecules of the metaparacyclophane type 14.4 Molecules of the metacyclophane type 14.5 Some other systems 14.6 Concluding summary References

959 961 961 972 974 976 980 984 986 986

14.1 Introduction We have already emphasized that the more stable –* charge-transfer molecular compounds have equimolar ratios and that the most striking feature of their crystal structures is the alternating arrangement of donor and acceptor molecules, planes essentially parallel, in mixed stacks. However, the mutual donor–acceptor arrangements in the crystals, under which heading we include the aspects of D A interplanar distance, orientation and overlap, is not free of influence from neighbouring stacks and separation of the various n n n

INTRA-MOLECULAR DONOR–ACCEPTOR INTERACTIONS

960

contributory factors can be troublesome. Thus study of charge transfer in an intramolecular situation has many advantages because the donor–acceptor arrangement is well-defined. This approach appears to have been suggested first by Cram and Day (1966), who synthesized 14.1 (the [2.2] compound); the corresponding [3.3] (Shinmyozu, Inazu and Yoshino, 1977) and [4.4] (Cram and Day, 1966) compounds have also been reported. Cram and Day (1966) anticipated that 14.2 and 14.3 would differ because of the possibility of interannular hydrogen bonding in 14.3 but not in 14.2 (Fig. 14.1). Such hydrogen bonding has not yet been encountered. However, in the early 1970s Staab pointed out a more important feature – that 14.2 and 14.3 differ in the mutual orientation of the -systems of donor and acceptor moieties and thus permit a direct test of the validity of Mulliken’s ‘‘overlap and orientation’’ principle for a D A pair. Many cyclophanes suitable for this purpose have been synthesized in the last decade, principally by Staab and coworkers in Germany but also in Japan and elsewhere (Schroff, van der Weerdt et al., 1973; Schroff, Zsom et al., 1976) and we survey these results, which have been reviewed (Schwartz, 1990). The cyclophane systems needed were obtained by extensive synthetic programmes which will not be discussed here despite their many novel features. The UV-visible absorption spectra of many diasteroisomeric pairs were measured in dilute solution, with Beer’s law checked to ensure that the spectra were truly those of intramolecular species. Solutions in rigid glasses allow low-temperature spectroscopy while determination of the crystal structures gives details of molecular structure (the rings are generally deformed in these strongly interacting systems) and also shows if there are appreciable intermolecular interactions in the crystals. Approximate molecular orbital calculations aid in the interpretation of the results. The reader should be warned that what started out as a simple and clearcut means of testing Mulliken’s ‘‘overlap and orientation’’ principle has developed complications; Staab, Dohling and Krieger (1991) remark ‘‘that a general and strict correlation between CT absorption and ground-state stabilisation by CT interaction (which is still widely taken for granted in organic chemistry) cannot be expected.’’ This will be illustrated towards the end of this chapter. n n n

p(CH2) X O

O

p(CH2) O

X O q(CH2) 14.2

X

X q(CH2) 14.3

Fig. 14.1. Schematic diagram of the pseudo-ortho (14.2) and pseudo-geminal (14.3) diastereoisomers of [n,m]paracyclophane quinhydrones (X¼OH) and related molecules. The methylene linkages between the rings can be equal but are not required to be so; n ¼ p þ 2 and m ¼ q þ 2, where p and q have integer values 0. The para mode of bridging is shown for both rings but meta bridging, and mixed meta-para bridging, are also found, as are other types of ring. 14.1 has X ¼ H, p ¼ q ¼ 0; 14.2 and 14.3 have X ¼ OH and p ¼ q ¼ 0.

MOLECULES OF THE PARACYCLOPHANE T YP E

961

14.2 Molecules of the paracyclophane type 14.2.1 Molecules derived from [n.n]paracyclophanes The experimental results show that para- and metacyclophane systems differ in important respects and we consider them in separate sections, starting with the paracyclophanes. Nomenclature and schematic structures are summarised in Fig. 14.1. Initial emphasis was placed on the quinhydrone systems; analogous crystalline intermolecular systems have been extensively studied (Section 15.7.1). The pseudo-ortho (14.2) and pseudogeminal (14.3) diastereoisomers were synthesized (Rebafka and Staab, 1973, 1974), followed by other syntheses and studies of physical properties (Staab and Rebafka, 1977; Staab, Herz and Henke, 1977; Staab and Haffner, 1977; Staab and Taglieber, 1977). The absorption spectra of the diastereoisomeric pair 14.2 and 14.3 are shown in Fig. 14.2 and it is clear that the charge transfer band centred at ca. 500 nm is much more pronounced for the pseudogeminal than for the pseudo-ortho isomer, and thus the charge-transfer interaction is appreciably stronger when the donor and acceptor moieties are parallel rather than when they are 60 apart. This conclusion is reinforced by a number of other measurements. For example, NMR results (1H NMR, d6-DMSO solution, coalescence of singlets at ¼ 4.14 and 3.98 above 161 C) shows that there is exchange of oxidation states between the two rings with simultaneous proton exchange in 14.3 but not in 14.2 (Rebafka and Staab, 1974). As the crystal structure of 14.3 has not been reported, it is not known whether there is interannular hydrogen bonding in this molecule. The pseudo-ortho configuration of the compound obtained in a synthesis of the paracyclophane diene quinhydrone (rings linked by double rather than single bonds) was inferred from the essential identity of its charge

HO

Ig E

O OH

O

4 O

OH

HO

O

3

2 250

300

350

400

450

500

550

600

650

700

[nm]

Fig. 14.2. Absorption spectra (dotted 14.2, full 14.3) of the diastereoisomeric [2.2]paracyclo-phane quinhydrones (solvent CH3OH). (Reproduced from Rebafka and Staab, 1974.)

962


transfer absorption with that of 14.2 (Stobbe, Kirchmeyer, Adiwidjajaj and Meijere, 1986). The orientation dependence of the charge transfer excitations in the [2.2] and [3.3] quinhydrone paracyclophanes has been explained in terms of Hu¨ckel molecular orbital theory (Vogler, 1983a,b), and good agreement obtained with experiment. The enhancement of the charge-transfer band in the pseudogeminal [2.2]paracyclophane donor–acceptor diastereoisomer (14.3) over that in the corresponding pseudoortho diastereoisomer (14.2) is a quite general phenomenon (see spectral parameters summarized in Table 14.1). The [2.2]paracyclophane quinhydrones have the disadvantage, in the present context, of very strong interannular interactions, with marked deforma˚ between their mean planes, compared to the tions of the rings and distances of 2.9 A ˚ 3.2 A found in the crystalline quinhydrones. These complicating features do not appear in the [3.3]paracyclophane quinhydrones (Staab and Herz, 1977b; Staab, Herz, Krieger and Rentea, 1983); in these diastereoisomeric pairs the CT interaction is also strikingly greater in the pseudogeminal than in the pseudo-ortho diastereoisomer. One of the objectives of the synthetic programmes was the preparation of cyclophanes with very strong donor and acceptor moieties. One such molecule would contain TMPD as donor and TCNQ as acceptor. Paracyclophanes with TCNQ as acceptor and donors of various kinds have been reported (Staab and Knaus, 1979; Staab, Knaus, Henke and Krieger, 1983; Tatemitsu et al., 1978). They show the same orientation effects as described above for the quinhydrones. The importance of the methoxy substituents in enhancing the donor strength is shown by the much stronger CT absorption in the compound with a p-dimethoxybenzene donor group and TCNQ acceptor group than when the donor is a benzene ring (Staab, Knaus et al., 1983). The converse situation where TMPD is donor and there are various acceptors has been achieved in the [2.2] series (Staab, Reimann-Haus et al., 1983) and also in the [3.3]paracyclophane series (Staab, Gabel and Krieger, 1983), where specifically the pseudo-ortho and pseudogeminal diastereoisomers of N,N,N 0 ,N 0 -tetramethyl[3]-(2,5)p-benzoquinone[3]paracyclophane-5,8-diamine were prepared and the crystal structure of an analog to the latter compound (in which the p-benzoquinone group was replaced by p-dimethylcyanobenzene) has been determined. One diastereoisomer with the desired TMPD–TCNQ combination has been synthesized – the pseudogeminal N,N,N 0 ,N 0 tetramethyl[3]-(2,5)tetracycanoquinodimethane[3]paracyclophane-5,8-diamine (Staab, Hinz, Knaus and Krieger, 1983). Its absorption spectrum shows a particularly broad CT band stretching over the region from 600–1600 nm. The trans-annular coupling is so strong (because of low ID and high EA values) that there is no longer any localization of valence electrons in the two rings and the spectrum does not show features of the TMPD and TCNQ spectra at short wavelengths, such as are found in weaker donor-acceptor systems. There is little dependence of the absorption spectrum on solvent polarity, the C N stretching frequency in TCNQ is 2216 cm1 and there is no ESR spectrum; all these facts point to a neutral ground state for the molecule rather than the ionic ground state that might have been anticipated (cf. Section 15.10.2). The TCNQ moiety is also neutral in the corresponding pseudo-ortho diastereoisomer ( CN ¼ 2220 cm1). Another such strong donor-acceptor pair is TTF and TCNQ where considerable progress has been made in the synthesis of the separate systems (TCNQ as before, TTF as below) but the desired combination in the same cyclophane (14.4) has not yet been reported. However, [3]tetrathiafulvaleno-[3]paracyclophane (14.5.1) and an analogous


963

Table 14.1. Comparison of spectral parameters for charge-transfer absorption bands of various donor–acceptor cyclophanes in pseudogeminal and pseudo-ortho configurations (note that this distinction does not hold when the donor is a benzene ring). The max values are in nm Donor moiety

Acceptor moiety

Cyclophane Pseudogeminal max/"

Pseudoortho max/"

Solvent

Ref

p-Dimethoxybenzene (DMB) DMB Benzene DMB

TCNQ

[2.2]para

695/3225

730/258

CHCl3

SKHK83

[3.3]para [3.3]para [2.2]para

705/3450 530*/795 483/1329

670/117

CHCl3 CHCl3

475/3000 466/360 462/324

DMB

BQ

TMPD TMPD

BQ BQ

[3.3]para [4.4]para [2]para crown (3) [2]para crown (3) with Naþ [2.2]para [3.3]para

438/160 498/142 500*/100 445/97

SK79 SKHK83 SHH77

DMB DMB DMB

TCNQ TCNQ p-benzoquinone (BQ) BQ BQ BQ

TMPD p-Dihydroxybenzene (quinol) quinol p-trimethylsiloxybenzene DMB

TCNQ BQ

[3.3]para [2.2]para

1050/3160 495/1600 515/170

SH83 CH3OH SR77

BQ BQ

[3.3]para [2.2]para

500*/105 430/205

dioxane CHCI3

SHKR83 SHH77

pyrazine

[2.2]para

462/3210 444/1710 500*/116 440/405

460/103

CF3COOH

SA81

benzene benzene benzene DMB

BQ BQ BQ p-dinitrobenzene

[2.2]para [3.3]para [4.4]para [2.2]para

340/597 406/407 288/1290 468/414

* shoulder

#

CHCl3 CHCl3

478/874

577/1930 538/2455

SDK83 SDK83 SSK83 SSK83

595/160 CH2CI2 537/76 387/1480#

475/120

CHCl3

SR83 SGK83

CD66 SIY77 CD66 SH77

2nd CT band

References: CD66 – Cram and Day, 1966; SA81 – Staab and Appel, 1981; SDK83 – Staab, Do¨hling and Krieger, 1981; SGK83 – Staab, Gabel and Krieger, 1983; SH77 – Staab and Haffner, 1977; SH83 – Staab, Hinz, Knaus and Krieger, 1983; SHH77 – Staab, Herz and Henke, 1977; SHKR83 – Staab, Herz, Krieger and Rentea, 1983; SIY77 – Shinmyozu, Inazu and Yoshino, 1977; SK79 – Staab and Knaus, 1979; SKHK83 – Staab, Knaus, Henke and Krieger, 1983; SR77 – Staab and Rebafka, 1977; SR83 – Staab, Riemann-Haus, Ulrich and Krieger, 1983; SSK83 – Staab, Starker and Krieger, 1983.

[4.4] compound have been synthesized (Staab, Ippen et al., 1980) and crystal structure analyses (briefly) reported for both compounds; surprisingly the latter compound was found to be the isomer 14.6 rather than 14.5.2. The [2.2]tetrathiafulvalene isomers 14.7 and 14.8 have been synthesized but only the second was obtained in pure form (Ippen, Tao-pen, Starker, Schweitzer and Staab, 1980).


964

CN

NC (H2C)n

S

S

S

(H2C)n S

14.5.1: n = 3 14.5.2: n = 4

14.4 S

S

NC

S

(CH2)n

S

(CH2)n

CN

S

S

S

S

14.6

Its structure has been confirmed crystallographically; the molecule has a step-like anti conformation. The [3.3]tetra-thiafulvalenes (14.9 and 14.10) have been synthesized but not completely purified. Black needles of a complex of composition [3.3]tetrathiafulvalene:(TCNQ)4 were obtained and shown to be triclinic but the structure was not reported. Conductivity measurements gave values of about 10 2 S/cm along the needle axis at 300K the conductivity at 40K is ten orders of magnitude less than at 400K!

(H2C)n

S

S

S

S

S

S

(H2C)m

S

S

S S

S S

S

S

(CH2)m

(CH2)n S

S

14.7: n = 0; 14.9: n = 1.

14.8: m = 2; 14.10: m = 3.

Sometimes there are complications. For example, the pyrazine moiety acts as an acceptor in the diastereoisomeric 12,15-dimethoxy-4,7-diaza[2.2]paracyclophanes (pseudogeminal and pseudo-ortho) (Staab and Appel, 1981); however, there are only small differences, which are solvent dependent, between the absorption spectra of the two diastereoisomers in the CT region of the spectrum. Thus a straightforward explanation in terms of ring overlap is no longer possible, as already noted in the Introduction. The generalization that there is greater CT interaction in pseudogeminal than in pseudoortho diastereoisomers breaks down for [4.4]paracyclophanes and those with longer


965

methylene bridges (Staab, Do¨hling and Krieger, 1981; Staab, Starker and Krieger, 1983). Crystal structure analysis of pseudogeminal-6,9,16,19-tetramethoxy[4.4]paracyclophane (Staab, Do¨hling and Krieger, 1983) and pseudogeminal-7,10,18,21-tetramethoxy[5.5]paracyclophane (Staab, Starker and Krieger, 1983) shows that the interannular dis˚ respectively. There is no distortion of the rings and presumably tances are 4.01 and 5.11 A there is little CT interaction at such large distances. Further comparison of the relative strengths of the donor–acceptor interaction in the two diastereoisomers can be made through emission spectra and zero-field splitting (ZFS) parameters (Hausser and Wolf, 1976); in these experiments the molecules are held at 1.3K, either in rigid glasses or as their crystals. The diastereoisomers used were those of 4,7-dicyano-12,15-dimethoxy-[2.2]-paracyclophanes (Schweitzer, Hausser et al., 1976) and 12,15-dimethoxy-4,7-diaza[2.2]-paracyclophanes (Staab, Herz and Henke, 1977). Crystal structures have been determined for the first pair of compounds by Irngartinger and Merkert but details do not appear to have been published. Emission spectra were used as follows: the characteristic properties of the excited triplet state of CT compounds originate from the fact that the two triplet excitons have a high probability of being, at a given time, in two different orbitals separated in space, i.e. one in the HOMO of the donor and the other in the LUMO of the acceptor. Consequently the exchange integral is diminished, with the energy of the first excited triplet state being reduced less than that of the corresponding excited singlet state. Thus large spectral overlap occurs between phosphorescence (triplet to ground state transition) and fluorescence (first excited singlet to ground state transition) spectra. Furthermore, the absolute red shifts of the emission spectra in CT-cyclophanes are larger than those in CT molecular compounds because of the stronger interannular interactions in the cyclophanes. The ESR method is to be preferred to emission spectroscopy because the ZFS parameters (the D values) are more sensitive than emission spectra to differences in CT interaction. The larger the separation between the two triplet electrons, the smaller is jDj; small jDj values thus correspond to larger CT interactions. The trend in jDj values shown in Table 14.2 is entirely compatible with the absorption spectra and also with the results of Hu¨ckel MO calculations (Vogler, Ege and Staab, 1977). However, it is not clear why the jDj values from the crystals are lower than those from the rigid glasses. The spectroscopic and other results mentioned above are reinforced by the detailed molecular geometries obtained from crystal structure analyses (note that in some instances structures of analogs have been reported rather than those of the immediatelyrelevant molecules, presumably because of difficulties in obtaining suitable crystals of the molecules of primary interest). Only a limited number of representative structures are discussed here. The pseudogeminal and pseudo-ortho diastereoisomers of 4,7-dimethoxy[2](2,5)tetracyanoquinodimethane[2]paracyclophane (Staab, Knaus, Henke and Krieger, 1983) are shown in Figs. 14.3 and 14.4, the actual molecular structures showing a remarkable resemblance to schematic formulae such as 14.1. The geometries of the two diastereoisomers are very similar (apart from the mutual positioning of the substituents), even in respect to the distortions introduced by the interannular interactions. Thus it is probable that these geometries can be taken as representative of all diastereoisomeric pairs in the [2.2]paracyclophane series.

966


Table 14.2. Zero Field Splitting (ZFS) parameters measured by Optical Detection of Magnetic Resonance (ODMR); all measurements at 1.3K. n-Octane and PMMA are rigid glasses. The jDj values are given in units of cm1. Compound

a. 2,5-dimethylpyrazine b. 1,4-dimethoxy-2,5dimethylbenzene c. 4,7-diaza[2.2]paracyclophane (Combination a-b) d. 1,4-dicyano-2,5dimethylbenzene e. 4,7-dicyano-12,15dimethoxy[2.2]-paracyclophane (Combination b-d)

Component

Paracyclophane

Donor

Pseudogeminal

Acceptor

Remarks Pseudoortho

0.177

n-octane n-octane

0.116 0.0967

0.1022

0.1229

n-octane PMMA

0.0313 0.0213

0.0642 0.0259

PMMA single crystal

The results of crystal structure analyses of [3.3]paracyclophanes are similar to those in the [2.2] series, and we shall give only one example – pseudogeminal 14,17dimethoxy[3](2,5)-p-benzoquinone[3]paracyclophane (Staab and Knaus, 1979; Fig. 14.5). There is almost exact overlap of the two rings as in the [2.2] series but less ring distortion. This perhaps explains the somewhat higher extinction coefficients found for [3.3] than for [2.2]cyclophanes (Table 14.1). The methylene bridge is disordered in some of the crystal structures in this group (for example, Fig. 14.5) but not in all of them. When both rings of the [nn]paracyclophane are the same and both are paradisubstituted in the same way, then the pseudogeminal diastereoisomer will be centrosymmetric while the pseudo-ortho diastereoisomer will have a C2(2) axis normal to the mean ring planes (these symmetries can be exact or approximate). If the rings are different then both diastereoisomers will have twofold axes (or even lower symmetry). If the molecules are chiral, there is a possibility of spontaneous resolution if the compound crystallizes in a Sohnke space group (cf. Section 11.2.2.1). This has so far been reported, for paracyclophanes, only for pseudogeminal-N,N,N 0 ,N 0 -tetramethyl[3](2,5)-pbenzoquinone[3]paracyclophane-14,17-diamine (space group P21, Z ¼ 2, molecular symmetry C11; Staab, Gabel and Krieger, 1983; BUVRIT). The metacyclophane (see Section 14.4 below) syn-15,18-dimethoxy[3]-p-benzoquinone[3]metacyclophane also crystallizes in space group P21, Z ¼ 2; Staab, Herz, Do¨hling and Krieger, (1980). Correlation of absolute configuration and optical rotation has not been reported for these two compounds but has been effected for a metaparacyclophane (see Section 14.3 below). A rather detailed comparison of the effects of various factors on donor–acceptor interactions in paracyclophanes has been made possible by the synthesis of a series of [2.2], [3.3] and [4.4] paracyclophanes all containing 1,2,4,5-tetracyanobenzene as acceptor and with a variety of donor moieties (Staab, Wahl and Kay, 1987). As crystal structures have been determined for most of these molecules, the comparisons can be


967

1.572

TOP VIEW

N

methoxy dicyano

1.579

z y methoxy

1.579

SIDE VIEW 1.510

N dicyano z

y x

Fig. 14.3. Top and side views of the pseudogeminal diastereoisomer of 4,7-dimethoxy[2](2,5)tetracyanoquinodimethane[2]paracyclophane (BUZROD). Note the close superpositioning of ˚ . Full molecular donor and acceptor moieties, and also their appreciable deformation. Distances in A dimensions of all crystal structures discussed in this Chapter are given in the original papers. (Data from Staab, Knaus, Henke and Krieger, 1983.)

968


Fig. 14.4. Stereoview view of the pseudo-ortho diastereoisomer of 4,7-dimethoxy[2](2,5)-tetracyanoquinodimethane[2]paracyclophane (BICFAU01). The overall molecular structure is very similar to that of the pseudogeminal diastereoisomer in Fig. 14.3. (Diagram produced using data from Staab, Knaus, Henke and Krieger, 1983.)

made for systems of well-defined geometry. The spectroscopic results are summarized in Table 14.3. The crystallographic results (Staab, Krieger, Wahl and Kay, 1987) follow the pattern established previously – considerably more distortion in the [2.2] (FIPLOF, FIPLUL, FIPMAS) than in the [3.3]paracyclophanes (FIPMEW, FIPMIA, FIPMOG), while distortion is negligible in the [4.4]paracyclophanes; almost complete eclipse of the donor and acceptor ring systems in the [2.2]paracyclophanes but with some mutual lateral shift of rings in [3.3]paracyclophanes and rather more variability in [4.4]paracyclophanes; substituents are coplanar with associated rings except for the methoxy groups in the tetramethoxy-substituted [2.2]paracyclophanes where the methyl groups are roughly normal to the ring planes. As Staab, Krieger, Wahl and Kay, (1987) point out, these changes can be followed qualitatively by changes in the colour of the crystals and more quantitatively by changes in max and " of the charge transfer absorption bands (Table 14.3). For example, in the [2.2]paracyclophane series, the benzene donor is weakest while increasing the donor strength by substitution of four methyl groups or two methoxy groups leads to a deepening of the colour of the solutions; however, four methoxy groups cannot be coplanar, thus reducing the mesomeric effect and the donor strength. If the dimethoxy donor is kept constant while the methylene chain length is increased, then the colour weakens from deep violet to dark red to orange as one passes from the [2.2] to the [3.3] and then to the [4.4]paracyclophane. The paracyclophanes discussed above are intramolecular charge-transfer molecular compounds and, as such, have a donor face and an acceptor face (Janus-like molecules) Thus one can anticipate formation in the crystals of stacks of the following type: D

A

D

A

D

A

D

A

These occur in some crystals but not in others; certainly this type of stacking is not a feature of the crystal structures comparable in importance to the – D A D A D A – arrangement


969

p-dimethoxy portion 1.412

1.382

1.518

1.497

1.227 p-benzoquinone portion x z

y

0

A

dimethoxy above

benzoquinone below

B

x y

Fig. 14.5. (upper) Perspective view of the pseudogeminal diastereoisomer of 14,17-dimethoxy[3](2,5)p-benzoquinone[3]paracyclophane (CECTIN). The molecule has a twofold axis normal to the mean ˚ . There is some disorder, not shown, in the methylene bridges. Note the ring planes. Distances in A close superpositioning of donor and acceptor moieties, and also some deformation. ˚ , I41cd, Z ¼ 8. (Data (lower) Packing diagram viewed down [001]; tetragonal, 15.891(2), 13.223(2) A from Staab, Herz, Krieger and Rentzea, 1983.)

970


Table 14.3. Colours and parameters for charge transfer absorption bands in spectra of [2.2], [3.3] and [4.4]paracyclophanes, with 1,2,4,5-tetracycanobenzene as acceptor moiety and various donors. The data in this Table are taken from pp. 556–557 of Staab, Krieger, Wahl and Kay, (1987). Asterisks indicate shoulders Donor Moiety

1. 2. 3. 4. 5. 6. 7. 8.

[n.n]

Benzene p-dimethoxybenzene 1,2,4,5-tetramethylbenzene 1,2,4,5-tetramethoxybenzene benzene p-dimethoxybenzene 1,2,4,5-tetramethylbenzene p-dimethoxybenzene

Colour

[2.2]

Charge-transfer band

yellow deep violet red orange red

[3.3] dark red [4.4]

orange

max(nm)

"

395* 520 440 380* 416 508 434 495

437 240 537 1318 1288 347 575 89

found in the mixed stacks of most intermolecular donor–acceptor molecular compounds (see Chapter 15). However, we note that this sort of stacking is found in many of the donor– acceptor complexes discussed in Chapter 3. One may venture the suggestion that ternary molecular compounds could be formed with stacks having the following arrangement: D

A

D' A

D

A'

D

A

D' A

D

Synthesis of such molecular compounds does not appear to have been attempted – a neat balance of donor and acceptor strengths would appear to be needed for success; however, it should be noted that [3.3]paracyclophane forms a 1 : 1 : * molecular compound with TCNE (Bernstein and Trueblood, 1971; PACTCN10) (see also Section 15.4), so the suggestion may not be entirely fanciful. It has been argued that cyclophanes containing donor (or alternatively acceptor) moieties in both rings would not be likely to form mixed stacks with added acceptor (donor) molecules but that segregated stacks would be favored (Staab, Gabel and Krieger, 1987). This proposal has so far been tested only for pseudogeminal-5,8,14,17-tetrakis(dimethylamino)[3.3]paracyclophane, which is formulated as TMPD TMPD

and which we shall denote for convenience as [TMPD-(CH2)3]2. This material forms a black 1 : 2 molecular compound of metallic appearance with TCNQ, the roomtemperature conductivity along the ‘‘longer crystal axis’’ being 1.5 S/cm, about 106 times as large as that of TMPD TCNQ (in which there is a mixed stack arrangement of ionized moieties). The crystal structure has not yet been reported. n n n


971

The compound [TMPD-(CH2)3]2 would be expected to undergo oxidation in four stages, the first giving an analog of Wu¨rster’s blue cation. Cyclic voltametry shows that oxidation occurs at potentials of 0.242, 0.102 and 0.249 V (two unresolved stages); these values should be compared with potentials of 0.206 and 0.378 V for TMPD and 0.081 and 0.242 V for 2,5-dimethyl-TMPD. Thus oxidation to the radical cation is more facile in [TMPD-(CH2)3]2 than in the other two compounds. Fast electron exchange has been demonstrated in the radical cation according to the following scheme: TMPD +•

TMPD

TMPD

TMPD

+•

The very rigidly linked [2.2.2.2](1,2,4,5) cyclophane quinhydrone (14.11; Staab and Schwendeman, 1978) is strongly deformed, the interannular distance (as judged from the crystal structures of the corresponding cyclophane (Hanson, 1977; CYLOPH) and the ˚ , compared to about 3 A ˚ in 14.3; the rings are tetraquinone (Krieger, 1978) is about 2.69 A boat-shaped in both reference molecules, with the substituent groups pointing away from the transannular region rather than towards it as in the paracyclophanes. Nevertheless the absorption spectrum is very similar to that of 14.3 (for the CT band, m ¼ 491 nm and " ¼ 1280, compared to 492 nm and 1600 for 14.3). Another surprising feature is that there is no hydrogen exchange between donor and acceptor moieties, even on heating to 140 C; perhaps this difference from the behaviour of 14.3 is due to a different type of deformation in 14.11 which increases the distance between hydroxyl and quinone oxygens beyond the limit for which exchange is possible.

CN O

O

HO

OH

14.11

NC

C

NC

C

CN

14.12

Crystal structures have been reported (Mizuma, Miki, Kai, Yasuoka and Kasai, 1982) for 14.12 (as its benzene solvate, P 1, Z ¼ 2; BICDUM) and for the 14,17-dimethoxy derivative of 14.12 (Fdd2, Z ¼ 16; pseudo-ortho diastereoisomer; BICFAU). The molecular structures show the distortions familiar from earlier work on cyclophane systems; in particular the six-membered rings have boat forms and the C(CN)2 portions are slightly twisted away from coplanarity with the ring. The packing in both crystals is based on head-to-tail stacking of the molecules (cf. previous paragraph). The solvent molecules in the benzene solvate do not participate in the stacking but are located in sheets about the (100) planes, interleaving double sheets of stacks of cyclophane molecules.

972

14.2.2


Systems related to [nn]paracyclophanes

Enhancement of the CT absorption was obtained in intramolecular quinhydrones with oligo-oxaparacyclophane structures (14.13; Bauer, Briaire and Staab, 1983). The effect was most marked for n ¼ 3; the CT absorption in the region of 500 nm was rather weak for the neat molecules but was strongly enhanced when the Naþ crown ether complex was formed, the conclusion being that the complexation had forced the donor-acceptor quinhydrone pair into a more parallel alignment. The crown ether portion of the molecule can be adjusted to be selective for a particular cation, while clathration of the cation is shown by enhanced spectral absorption; thus the principle of a cation-selective ligand with a ‘‘built-in’’ charge-transfer indicator has been demonstrated. O O CH2

OCH3 O O

CH2

n

O CH3O 14.13 Note: the –CH2–O–CH2– group is repeated n times (for n ≥ 1).

Some work has also been done on cyclophanes containing naphthaleno systems, including synthesis of intramolecular quinhydrones based on [2](1,4)naphthaleno[2]paracyclophane (Herz and Staab, 1977) and [2.2](1,4)naphthalenophane. The syn and anti isomers 14.14 and 14.15 of the latter type have very similar absorption spectra, both showing broad CT bands between 500 and 700 nm (Staab and Herz, 1977a; not reproduced here). It was inferred that in these molecules direct donor–acceptor interaction through space was less important than the interaction through the strongly coupled [2.2]paracyclophane portions of the molecules. The spectra of 14.16 and 14.17 (Fig. 14.6) show even broader CT bands than 14.14 and 14.15. O

O

OH

OH O HO

O

HO 14.14 (syn isomer, dec 260 °C)

14.15 (anti isomer)

The crystal structure of 14.16 (space group Cc, Z ¼ 4) has been reported (Mizuma, Miki, Kai, Tanaka and Kasai, 1982; BIMNOA). One of the six-membered rings in the


973

5 Charge transfer bands 4 Log epsilon

14.20

14.16

3

14.17 2 2,5-dimethyl-TCNQ

14.12

1 300

700

500

900

Wavelength nm

Fig. 14.6. Electronic spectra (in CH2Cl2) of the two-layer molecules 14.12, 14.16 and 14.17, the three-layer molecule 14.20 and 2,5-dimethyl-TCNQ. (Adapted from Yoshida et al., 1978.)

naphthaleno portion is boat-shaped while the other, which protrudes, is planar. The crystals contain head-to-tail stacks of molecules. NC

NC

NC

NC CN

CN

CN 14.16

CN 14.17

(Note: some double bonds have been omitted in the TCNQ moieties).

Some work has been done on systems where the two rings have different sizes, one six-membered and the other seven-membered; para systems are discussed here and meta and mixed systems later. The [2]paracyclophane[2](3,7)p-tropoquinonophane (14.18) shows little evidence of CT absorption, and this is also true of the corresponding [3.3] compound (Kawamata, Fukazawa, Fujise and Ito, 1982a,b). However, there is strong charge transfer when the seven-membered ring is positively charged and so acts as a strong electron acceptor; for example [2.2](1,4)tropylioparacyclophane tetrafluoroborate (14.19) has a broad CT band centred at about 350 nm, with "m about 3100 (Horita, Otsubo, Sakata and Misumi, 1976; O’Connor and Keehn, 1976).

974


O

+ O

BF4–

O

14.18

14.2.3

14.19

Multi-layered systems

Considerable effort has been invested in the synthesis of multilayered systems and, so far, triply and quadruply layered molecules are known. In the triple-layer molecules 14.20–14.25 there are rather similar CT bands (14.21 and 14.23 have essentially identical spectra), which do not depend on the mutual orientation of donor and acceptor moieties (Staab, Zapf and Gurke, 1977; Machida, Tatemitsu, Sakat and Misumi, 1978; Staab and Zapf, 1978). 14.20 has the absorption maximum of its CT band in the 600 nm region (Fig. 14.6) but replacement of the TCNQ moiety in these molecules by p-benzoquinone leads to appreciable bathochromic shifts to the 450–500 nm region. On the basis of the spectroscopic results Machida, Tatemitsu, Sakat and Misumi (1978) inferred that ‘‘a sandwiched benzene ring functions as a sort of conductor for intramolecular donor–acceptor interaction and not as an insulator.’’ Presumably this is the reason why there is a distinct CT absorption for 14.20 but only a barely noticeable shoulder for 14.12 (Fig. 14.6). The generalisation holds also for the triple-layer charged tropylium system 14.24 (Horita, Otsubo, Sakata and Misumi, 1976). Thus it is intriguing to note that the CT interaction in 14.25 is weak (Tatemitsu, Otsubo, Sakata and Misumi, 1975). Unfortunately details of the absorption spectra have not been published for the p-dicyano triple-layer analogs (Yoshida, Tatemitsu, Sakata, Misumi, Masuhara and Mataga, 1976) of 14.21–14.23 (where the p-benzoquinone group has been replaced by p-dicyanobenzene), which appear to be the only other comparable group of molecules to have been synthesized. CN NC

CN CN

14.20


O

975

O

O

O

O

O MeO

MeO MeO

OMe OMe

OMe 14.21

14.22

pseudogeminal isomer 1

+

14.23 pseudo-ortho isomer 2

O –

BF4

O

14.24

14.25

There are four-layer systems (Staab and Zapf, 1978) with a p-benzoquinone moiety at one end as acceptor and a p-dimethoxybenzene moiety at the other end as donor, separated by two benzene rings. These compounds show intense CT absorption bands in the 350–550 nm range (a: m ¼ 447 nm, " ¼ 2500; b: m ¼ 450 nm, " ¼ 2490); the intramolecular nature of the absorption was checked from the concentration dependence of the spectra. Not only are the intensities of the CT bands higher than those of the comparable triple-layered [2.2]paracyclophane quinhydrones but they also show a distinct bathochromic shift, indicating overall enhancement of the donor strength of the -electron system interacting with the acceptor p-benzoquinone moiety. These molecules were synthesized as a diastereoisomeric pair but there was no evidence from their spectra of any dependence of the absorption on the mutual orientation of donor and acceptor. Thus the triple- and quadruple-layer systems behave differently in this regard from the simpler [2.2] and [3.3] paracyclophane quinhydrones. Simple Hu¨ckel molecular orbital theory breaks down when four-layer cyclophane double-quinhydrones are considered, the states being dominated by important (nearly first order) configuration interaction (Vogler, 1983b). The crystal structure of 14.25 has been reported (Toyoda, Tatemitsu, Sakata, Kasai and Misumi, 1986; FEFYIY) and also that of the compound in which the p-benzoquinone

976


moiety is replaced by bromobenzene (Koizumi, Toyoda, Miki, Kasai and Misumi, 1986; DOHXAZ); the crystals are isomorphous, Pbcn, Z ¼ 4, and the molecules have twofold axes along the O . . . O (Br . . . H) vectors. The outer benzene rings are deformed to boats and the inner ring has a twist shape in these triply layered molecules. A similar pattern of distortions for inner and outer rings is found in the centrosymmetric quadruply layered tetramethyl [2.2]cyclophane C40H14 (Mizuno et al., 1977; MPCPHT10).

14.3

Molecules of the metaparacyclophane type

The [2.2] metaparacyclophane quinhydrones and derivatives are conveniently treated here because they act as a bridge between the paracyclophanes, which they resemble in rigidity, and the metacyclophanes, to which they are perhaps closer in geometrical structure. 12,15-Dimethoxy[2](2,6)-p-benzoquinono[2]paracyclophane (14.26; DALTOZ) and 13,16-dimethoxy[2](2,5)-p-benzoquinono[2]metacyclophane (14.27) were first reported by Staab, Jo¨rns and Krieger in 1979 and, in more detail, some years later (Staab, Jo¨rns, Krieger and Rentzea, 1985). The quinhydrone analog of 14.27 has also been reported (Tashiro, Koya and Yamamoto, 1983). 14.26 has a broad CT band from 400–630 nm (m ¼ 490 nm, " ¼ 590, in CHCl3), whereas the CT band of 14.27 is blue-shifted by about 70 nm (m ¼ 420 nm, " ¼ 825, in CHCl3). O

O OMe OMe

MeO

O

14.26

MeO

O

14.27

14.26 crystallizes in three polymorphic forms, two racemic and one enantiomorphic (P212121, Z ¼ 4) and the stereochemical implications derived from the crystal structure of the latter have been investigated in particularly thorough fashion (Staab, Jo¨rns, Krieger and Rentzea, 1985). The meta-bridged quinone unit shows a much greater deformation from planarity than the para-bridged aromatic moiety (Fig. 14.7). The carbonyl group is located above the aromatic ring in a manner similar to that found in binary CT molecular compounds where the acceptor is a quinone, and also in self-complexes such as naphthoquinones (cf. Section 15.6). The spontaneous resolution of 14.26 into chiral crystals was exploited by handseparation of enantiomorphs under the microscope (the Pasteur method) and measurement of optical rotatory dispersion (ORD) and circular dichroism (CD) in CHCl3 solution (Fig. 14.8). In principle, at least, the absolute configurations of the crystals could

MOLECUL ES OF THE METAPARACYCLOPHANE TYPE

para bridge

1.377

977

p-dimethoxy portion

1.554

meta bridge

z x

p-benzoquinone portion

1.224 y

C

A

B

z y

Fig. 14.7. Molecular and crystal structure of 14.26 (DALTOZ). (upper): Perspective view of the ˚ ) in the two portions of the molecule are shown, as well as molecule. The different C–O distances (A a typical bridging C–C distance. Bond angle deformation is preferred to alteration of bond lengths. (lower): Packing arrangement shown in projection down [100]; hydrogens omitted for clarity. ˚ , P212121, Z ¼ 4. (Data from Staab, Jo¨rns, Krieger and Orthorhombic; 7.698(1) 8.205(1) 24.025(2) A Rentzea, 1985.)

be determined by the Bijvoet method using anomalous scattering from the oxygen atoms and then related to the signs of the optical rotation; however, this has not yet been done and is a challenging task, as has been demonstrated, for example, by Rabinovich and Hope (1980). Furthermore, ‘‘these systems offer, apparently for the first time, the

978


(+) O

40

OMe

30 MeO

O

20

[u]·10–3

10 A 0 300 10

400

500

600

l[nm]

B

20

30

40 (–)

Fig. 14.8. Circular dichroism of two enantiomorphic single crystals (of unknown optical purity) in CHCl3 solution (concentrations 2.8 106 (A) and 2.1 106 g/ml (B) respectively). The molar rotations (in deg. at 20 ) were []460 5215 (A) and þ 5811 (B); []334 35164 (A) and þ 41720 (B). (Reproduced from Staab, Jo¨rns, Krieger and Rentzea, 1985.)

opportunity of measuring ORD and CD related to a charge-transfer chromophore with well-defined and rigid donor–acceptor orientations’’ (Staab, Jo¨rns, Krieger and Rentzea, 1985). 14.26 and 14.27 are, of course, isomers which differ in that the aromatic ring is para bridged in the first of the pair and meta bridged in the second. The crystal structure of 14.27 has not been reported but those of 5,8,12,15-tetra-methoxy[2.2]metaparacyclophane (DALTUE) and the corresponding bis-quinone {[2.2](2,5)-(2,6)-p-benzoquinophane (DALVAN) have been determined (Staab, Jo¨rns, Krieger and Rentzea, 1985). All three molecules have similar overall shapes, and it is reasonable to assume that this holds for 14.27 as well. There have also been a number of investigations of compounds of the [3.3]metaparacyclophane series and both quinhydrone isomers have been synthesized (as the methoxy derivatives 14.28 and 14.29) (Staab, Jo¨rns, Krieger and Rentzea, 1985). In the corresponding benzene compounds it was inferred (Staab and Knaus, 1979) from NMR spectra that the rings were not parallel, thus accounting for the rather ill-defined CT bands in the UV-visible spectra. Definitive evidence about molecular structure comes from the crystal structure analysis of 14.28 (Fig. 14.9; DALVER).

MOLECUL ES OF THE METAPARACYCLOPHANE TYPE

979

p-dimethoxy portion

O4 para bridge

O3 1.536 O2

meta bridge p-benzoquinone portion

y x

O1 z

A

B

C

y z

Fig. 14.9. Molecular and crystal structure of 14.28 (DALVER). (upper) Perspective view; distances ˚ . The torsion angle O1–O2–O3–04 is 141.5 . in A (lower) Pseudo-hexagonal close packing of molecular stacks viewed along [100]; monoclinic, ˚ , ¼ 114.87(2) , P21/c, Z ¼ 4. (Data from Staab, Jo¨rns, Krieger and 8.555(1) 25.697(3) 8.370(1) A Rentzea, 1985.)

980


OMe

O

O OMe

MeO

O

MeO 14.28

O

14.29

The two rings are inclined at an angle of 12.6 , the individual rings being only mildly deformed; one bridge shows disorder of a methylene group over two sites (occupancy of major site 85%) similar to that encountered earlier (see Fig. 14.5). The spectrum of 14.28 has a CT band similar to that of 14.26 (m ¼ 485 nm, " ¼ 560, in CHCl3); however, there is also an absorption at shorter wavelength (m ¼ 375 nm, " ¼ 620, in CHCl3) which was ascribed to a second CT transition analogous to that found in the electronic spectra of pseudo-ortho [3.3]paracyclophane quinhydrones. The CT band of 14.29 is blue-shifted by about 120 nm (m ¼ 365 nm, " ¼ 800, in CHCl3), analogous to the behaviour of the 14.26 and 14.27 diastereoisomers. The rather complicated and distorted geometrical structures of the molecules in this series prevent explanation of the spectra by the simple HMO model that worked fairly well for the [2.2] and [3.3]paracyclophane quinhydrones. The analogous compounds containing tropoquinono rings have also been reported (Kawamata, Fukazawa, Fujise and Ito, 1982a,b); the spectra are very similar to those of the para compounds (e.g. 14.18), with ill-defined CT bands. 14.4

Molecules of the metacyclophane type

The anti- and syn-isomers of a number of [2.2]metacyclophanes, where both donor and acceptor groups are suitably-substituted aromatic moieties, have been investigated (Staab, Schanne, Krieger and Taglieber, 1985); the corresponding quinhydrones are also available (Staab, Reibel and Krieger, 1985). Crystal structures have been reported for representative molecules and there are appreciable geometrical differences between anti and syn isomers. The structure of anti-5,8-dimethoxy-13-nitro[2.2]metacyclophane (Staab, Schanne, Krieger and Taglieber, 1985; DAVHAJ) is shown in Fig. 14.10 and that of syn-13,16dimethoxy-[2](2,6)-p-benzoquinono[2]metacyclophane (Staab, Reibel and Krieger, 1985; DEBZEP) in Fig. 14.11. Meo x

OMe Meo x

OMe x

X = COOMe, CN, NO 2

O

O x

MeO

OMe

SYN

O

O Meo

OMe

ANTI

The anti-isomer shows little overlap of donor and acceptor portions and the tilt between them is relatively small at 15 . On the other hand, although the two rings of

MOLECULES OF THE METACYCLOPHANE TYPE

981

O1

1.563

p-dimethoxy portion O2

PERSPECTIVE VIEW

C16

2-meta bridges y N1 nitrobenzene portion

x

z

nitrobenzene portion

SIDE VIEW

meta bridges

z p-dimethoxy portion x y

Fig. 14.10. Molecular structure of anti-5,8-dimethoxy-13-nitro[2.2]metacyclophane (DAVHAJ). ˚. (upper) Perspective view of molecule; the torsion angle O1–O2–C16–N1 is 178.95 ; distances in A (lower) Side view. A stereoview of the molecule is shown in the original paper. (Data from Staab, Schanne, Krieger and Taglieber, 1985.)

the syn molecule appear to be appreciably overlapped in plan view, the side view shows that there is a tilt angle of 33 between them and that there are close interannular approaches only on one side of the molecule. Despite these geometrical differences, the spectra of the anti, syn pair of diastereoisomeric quinhydrones are remarkably similar (Fig. 14.12; Staab, Reibel and Krieger, 1985) and this holds also for other analogous pairs; investigation of the solvent dependence of the fluorescence from various molecules showed that the long wavelength absorption bands were indeed charge-transfer bands (Staab, Schanne, Krieger and Taglieber, 1985).


982

PERSPECTIVE VIEW p-dimethoxy portion

O4

O3

1.557

O2


x y

O1

z


O1

O2

SIDE VIEW 1.557 O4 O3 p-dimethoxy portion

x

y

z

Fig. 14.11. Molecular structure of syn-13,16-dimethoxy[2](2,6)-p-benzoquinone-[2]metacyclo-phane (DEBZEP). (upper) Perspective view of molecule; the torsion angle O1–O2–O4–O3 is 9.81 ; ˚ . (lower) Side view. A stereo view of the molecule is shown in the original paper. distances in A (Data from Staab, Reibel and Krieger, 1985.)

The similarity of the CT absorption bands from anti and syn donor–acceptor metacyclophanes despite the very different disposition of donor and acceptor moieties has been explained in terms of Hu¨ckel molecular orbital theory (Vogler, Schanne and Staab, 1985).

MOLECULES OF THE METACYCLOPHANE TYPE

983

lg e O

O

O

O OMe MeO

MeO

OMe

4.00

3.00

2.00

1.00 250

300

350

400

450

500

550

l (nm)

Fig. 14.12. The charge-transfer spectra (in CHCl3) of the anti -syn isomer pair of quinhydrones shown as inserts at the top of the diagram. (Reproduced from Staab, Schanne, Krieger and Taglieber, 1985.) lg ε

O MeO

O OMe

O

lg ε

O MeO

HO

OMe

4

4

3

3

2

2

300

400

500

λ (nm)

O

300

O OMe

400

O

O MeO

500

OH

λ (nm)

Fig. 14.13. Comparison of the spectra of the syn and anti diastereoisomers of the two [3.3]metacyclophanes shown as inserts at the top of the figure. Replacement of one methoxy group in each donor moiety by an hydroxyl has little effect on the spectra. (Reproduced from Staab, Herz, Do¨hling and Krieger, 1980.)

Quinhydrones of the [3.3]metacyclophane series and related molecules have also been studied (Staab, Herz and Do¨hling, 1979; Staab and Do¨hling, 1979; Staab, Herz and Do¨hling, 1980; Staab, Herz, Do¨hling and Krieger, 1980). Spectra of anti and syn pairs of isomers show marked resemblances (Fig. 14.13). The spectrum of syn-15,18-dihydroxy[3](2,6)-p-benzoquinone[3]meta-cyclophane is very similar to that of pseudogeminal 14,17-dihydroxy[3](2,5)-p-benzoquinone-[3]

984


O

O

HO

lg ε

O

OH

O

HO

OH

4

3

2

300

400

500

600

λ (nm)

Fig. 14.14. Comparison of the UV-visible absorption spectra (in dioxane) of the pseudogeminal 14,17-dihydroxy[3](2,5)-p-benzoquinone[3]paracyclophane (on the left of the insert in the upper portion of the diagram) and of syn-15,18-dihydroxy[3](2,6)-p-benzoquinone[3]metacyclophane (on the right). (Reproduced from Staab, Herz, Do¨hling and Krieger, 1980.)

paracyclophane (Fig. 14.14). The overlaps of donor and acceptor portions of these two molecules are not very different, thus accounting for the similarity of their spectra; the structure of syn-15,18-dihydroxy[3](2,6)-p-benzoquinone[3]metacyclophane is shown in Fig. 3 of Staab, Herz, Do¨hling and Krieger (1980) while that of pseudogeminal 14,17-dihydroxy-[3](2,5)-p-benzoquinone[3]paracyclophane is in Fig. 14.5. The crystal structures of two syn isomers have been reported – of syn-15,18dimethoxy(2,6)-p-benzoquinono[3.3]metacyclophane (Staab, Herz, Do¨hling and Krieger, 1980; MXBQMP) and of syn-6,9-dimethoxy-15,18-dinitro-2,11-dithia[3.3]metacyclophane (Staab, Schanne, Krieger and Taglieber, 1985; DAVGUC). As would be expected, the molecules of the [3.3]metacyclophane series are less distorted (but some distortion remains) than those of the [2.2]metacyclophane series. 14.5

Some other systems

We give a few examples of other systems where the principles discussed above are being applied in order to illustrate how one may expect the particular features of cyclophane systems to be exploited in the future. The electron transfer properties of the verticallystacked porphyrin-quinone(1)-quinone(2) cyclophane, an analog to compounds involved in the primary process of biological photosynthesis, have been studied in order to determine which structural factors favour a consecutive, stepwise electron transfer and which an integrated process relying on electron coupling (Staab, Tercel, Fischer and Krieger, 1994). During the course of the synthesis of compound 14.30, the crystal structure of 14.31 (which is a substituted pseudo-ortho[3.3]-paracyclophane) was determined (WIHMIJ); the molecular structure resembles that shown in Fig. 14.5.

SOME OTHER SYSTEMS

985

Cl 0 pseudo-ortho [3.3]paracyclophane

0 Cl

Cl 0

0Me

Me0

Et

0Me

Cl

Et

Me

Me N NH

Me

HN

Me0

Me00C

N

C00Me

Me Et

Et

14.30

14.31

Another system (Cowan, Sanders, Beddard and Harrison, 1987), involving different dispositions of pyromellitimide (an electron acceptor) and porphyrin rings (the cofacial pair acting as an electron donor), demonstrates that ‘‘mere proximity between donor and acceptor is not a sufficient condition for electron transfer. There is also a strong geometrical requirement.’’ Picosecond fluorescence measurements show (by marked quenching of the fluorescence) that there is rapid electron transfer from the excited porphyrin pair to the pyromellitimide electron acceptor in 14.32 whereas there is relatively little fluorescence quenching in 14.33, and hence slow electron transfer. A rather similar rigid triple-ring molecule with a porphyrin sandwiched between two parallel p-benzoquinone units has also been reported (Weiser and Staab, 1984). There are clear analogies to triple-layer CT cyclophanes. Another triple layer system is the crystalline green cationic acceptor–donor–acceptor system prepared by Simonsen et al. (1998), with butane-1,4-diyl linkers; there are four PF6 counter-anions and three MeCN molecules of solvation. The acceptors are bipyridinium moieties and the donor is based on TTF. Crystal structure analysis shows a rather distorted molecule without particularly close interactions. Presumably shorter linker units are needed to enforce geometrical constraints. N

O

O

O

O

O

O

O O

N

O

N

O

O

N O O N NH O

O O

NH

N

N H N

HN N

N H N

HN O

N O

O

O

O

O

O

O (2) N NH

HN N

(1)

14.32

14.33


986

14.6

Concluding summary

Thus, in summary, the measurements on the cyclophane intramolecular donor–acceptor compounds have confirmed, for many two-layer molecules, a marked dependence of the degree of charge transfer on the mutual orientation of the donor and acceptor moieties. The charge transfer is greater when the long axes of the moieties are parallel than when they are inclined at an angle of 60 . For two-layer molecules the situation in the [2.2]paracyclophanes is complicated by the very strong interaction between the adjacent rings, but the orientation effect is just as marked in the [3.3]paracyclophanes where the distance between the rings is similar to non-bonded distances found in the corresponding crystalline compounds. Charge transfer interaction persists through the rings in three- and four-layer molecules but the orientation effect is lost. The orientation effects are not marked in systems such as the tropoquinophanes and the higher [n.n]cyclophanes (n > 3), where the CT interaction is rather weak. In the [n.n]metacyclophane series the geometrically-different anti and syn isomers give remarkably similar CT absorption spectra, a result which has been explained by HMO theory. One may anticipate that a wealth of information remains to be uncovered by spectroscopy at very low temperatures of molecules in rigid glasses, and that these results will eventually be interpreted by ab initio calculations based on the detailed geometrical structures obtained from diffraction studies. In addition one should note that cyclophanetype molecules lend themselves to study of a variety of secondary interactions between moieties in defined geometrical situations, and this is likely to be an important growth area in future research. References Bauer, H., Briaire, J. and Staab, H. A. (1983). Angew. Chem. Int. Ed. Engl., 22, 334–335. Bernstein, J. and Trueblood, K. N. (1971). Acta Cryst., B27, 2078–2089. Cowan, J. A., Sanders, J. K. M., Beddard, G. S. and Harrison, R. J. (1987). J. Chem. Soc., Chem. Comm., pp. 55–58. Cram, D. J. and Day, A. C. (1966). J. Org. Chem., 31, 1227–1232. Hanson, A. W. (1977). Acta Cryst., B33, 2003–2007. Hausser, K. H. and Wolf, H. C. (1976). Adv. Magn. Reson., 8, 85–121. Herz, C. P. and Staab, H. A. (1977). Angew.Chem. Int. Ed. Engl., 16, 394. Horita, H., Otsubo, T., Sakata, Y. and Misumi, S. (1976). Tetr. Letts., pp. 3899–3902. Ippen, J., Tao-pen, C., Starker, B., Schweitzer, D. and Staab, H. A. (1980). Angew. Chem. Int. Ed. Engl., 19, 67–69. Kawamata, A., Fukazawa, Y., Fujise, Y. and Ito, S. (1982a). Tetr. Letts., 23, 1083–1086. Kawamata, A., Fukazawa, Y., Fujise, Y. and Ito, S. (1982b). Tetr. Letts., 23, 4955–4958. Koizumi, Y., Toyoda, T., Miki, K., Kasai, N. and Misumi, S. (1986). Bull. Chem. Soc. Jpn., 59, 239–242. Krieger, C. (1978). Unpublished. Machida, H., Tatemitsu, H., Sakata, Y. and Misumi, S. (1978). Tetr. Letts., pp. 915–918. Mizuma, T., Miki, K., Kai, Y., Tanaka, N. and Kasai, N. (1982). Bull. Chem. Soc. Jpn., 55, 2026–2028.

REFERENCES

987

Mizuma, T., Miki, K., Kai, Y., Yasuoka, N. K. and Kasai, N. (1982). Bull. Chem. Soc. Jpn., 55, 979–984. Mizuno, H., Nishiguchi, K., Toyoda, T., Otsubo, T., Misumi, S. and Morimoto, M. (1977). Acta Cryst. B33, 329–334. O’Connor, J. G. and Keehn, P. M. (1976). J. Am. Chem. Soc., 98, 8446–8450. Rabinovich, D. and Hope, H. (1980). Acta Cryst., A36, 670–678. Rebafka, W. and Staab, H. A. (1973). Angew. Chem. Int. Ed. Engl., 12, 776–777. Rebafka, W. and Staab, H. A. (1974). Angew. Chem. Int. Ed. Engl., 13, 203–204. Schroff, L. G., Weerdt, A. J. A.van der, Staalman, D. J. H., Verhoeven, J. W. and de Boer, Th. J. (1973). Tetr. Letts., pp. 1649–1652. Schroff, L. G., Zsom, R. L. J., Weerdt, A. J. A. van der, Schrier, P. I., Geerts, J. P., Nibbering, N. M. M., Verhoeven, J. W. and de Boer, Th. J. (1976). Rec. Trav. Chim. Pays-Bas, 95, 89–93. Schwartz, M. H. (1990). J. Incl. Phenom., 9, 1–35. Schweitzer, D., Hausser, K. H., Taglieber, V. and Staab, H. A. (1976). Chem. Phys., 14, 183–187. Shinmyozu, T., Inazu, T. and Yoshino, T. (1977). Chem. Letts., pp. 1347–1350. Simonsen, K. B., Thorup, N., Cava, M. P. and Becher, J. (1998). Chem. Commun., pp. 901–902. Staab, H. A. and Appel, W. (1981). Liebig’s Ann. Chem., pp. 1065–1072. Staab, H. A. and Do¨hling, A. (1979). Tetr. Letts., pp. 2019–2022. Staab, H. A. and Haffner, H. (1977). Chem. Ber., 110, 3358–3365. Staab, H. A. and Herz, C. P. (1977a). Angew. Chem. Int. Ed. Engl., 16, 392–394. Staab, H. A. and Herz, C. P. (1977b). Angew. Chem. Int. Ed. Engl., 16, 799–801. Staab, H. A. and Knaus, G. H. (1979). Tetr. Letts., pp. 4261–4264. Staab, H. A. and Rebafka, W. (1977). Chem. Ber., 110, 3333–3350. Staab, H. A. and Schwendemann, V. (1978). Angew. Chem. Int. Ed. Engl., 17, 756–757. Staab, H. A. and Taglieber, V. (1977). Chem. Ber., 110, 3366–3376. Staab, H. A. and Zapf, U. (1978). Angew. Chem. Int. Ed. Engl., 17, 757–758. Staab, H. A., Do¨hling, A. and Krieger, C. (1981). Liebig’s Ann. Chem., pp. 1052–1064. Staab, H. A., Do¨hling, A. and Krieger, C. (1991). Tetr. Letts., 32, 2215–2218. Staab, H. A., Gabel, G. and Krieger, C. (1983). Chem. Ber., 116, 2827–2834. Staab, H. A., Gabel, G. and Krieger, C. (1987). Chem. Ber., 120, 269–273. Staab, H. A., Herz, C. P. and Do¨hling, A. (1979). Tetr. Letts., pp. 791–794. Staab, H. A., Herz, C. P. and Do¨hling, A. (1980). Chem. Ber., 113, 233–240. Staab, H. A., Herz, C. P. and Henke, H.-E. (1977). Chem. Ber., 110, 3351–3357. Staab, H. A., Herz, C. P., Do¨hling, A. and Krieger, C. (1980). Chem. Ber., 113, 241–254. Staab, H. A., Herz, C. P., Krieger, C. and Rentzea, M. (1983). Chem. Ber., 116, 3813–3830. Staab, H. A., Hinz, R., Knaus, G. H. and Krieger, C. (1983). Chem. Ber., 116, 2835–2847. Staab, H. A., Ippen, J., Tao-pen, C., Krieger, C. and Starker, B. (1980). Angew. Chem. Int. Ed. Engl., 19, 66–67. Staab, H. A., Jo¨rns, M. and Krieger, C. (1979). Tetr. Letts., pp. 2513–2516. Staab, H. A., Jo¨rns, M., Krieger, C. and Rentzea, M. (1985). Chem. Ber., 118, 796–813. Staab, H. A., Knaus, G. H., Henke, H. -E. and Krieger, C. (1983). Chem. Ber., 116, 2785–2807. Staab, H. A., Krieger, C., Wahl, P. and Kay, K-Y. (1987). Chem. Ber., 120, 551–558. Staab, H. A., Reibel, W. R. K. and Krieger, C. (1985). Chem. Ber., 118, 1230–1253. Staab, H. A., Reimann-Haus, R., Ulrich, P. and Krieger, C. (1983). Chem. Ber., 116, 2808–2826. Staab, H. A., Schanne, L., Krieger, C. and Taglieber, V. (1985). Chem. Ber., 118, 1204–1229. Staab, H. A., Starker, B. and Krieger, C. (1983). Chem. Ber., 116, 3831–3834. Staab, H. A., Tercel, M., Fischer, R. and Krieger, C. (1994). Angew. Chem. Int. Ed. Engl., 33, 1463–1466. Staab, H. A., Wahl, P. and Kay, K-Y. (1987). Chem. Ber., 120, 541–549. Staab, H. A., Zapf, U. and Gurke, A. (1977). Angew. Chem. Int. Ed. Engl., 16, 801–803.

988


Sto¨bbe, M., Kirchmeyer, S., Adiwidjaja, G. and Meijere, A. de, (1986). Angew. Chem. Int. Ed. Engl., 25, 171–173. Tashiro, M., Koya, K, and Yamato, T. (1983). J. Am. Chem. Soc., 105, 6650–6653. Tatemitsu, H., Natsume, B., Yoshida, M., Sakata, Y. and Misumi, S. (1978). Tetr. Letts., pp. 3459–3462. Tatemitsu, H., Otsubo, T., Sakata, Y. and Misumi, S. (1975). Tetr. Letts., pp. 3059–3062. Toyoda, T., Tatemitsu, H., Sakata, Y., Kasai, N. and Misumi, S. (1986). Bull. Chem. Soc. Jpn., 59, 3994–3996. Vogler, H. (1983a). Tetr. Letts., pp. 2159–2162. Vogler, H. (1983b). Z. Naturforschung, 38B, 1130–1135. Vogler, H., Ege, G. and Staab, H. A. (1977). Mol. Phys., 33, 923–932. Vogler, H., Schanne, L. and Staab, H. A. (1985). Chem. Ber., 118, 1254–1260. Weiser, J. and Staab, H. A. (1984). Angew. Chem. Int. Ed. Engl., 23, 623–625. Yoshida, M., Tatemitsu, H., Sakata, Y., Misumi, S., Masuhara, H. and Mataga, N. (1976). J. Chem. Soc. Chem. Comm., pp. 587–588. Yoshida, M., Tochiaki, H., Tatemitsu, H., Sakata, Y. and Misumi, S. (1978). Chem. Lett., pp. 829–832.

Chapter 15 Crystal chemistry of mixed-stack p–p* molecular compounds

Wie aus zahlreichen Versuchen hervorgeht, vereinigen sich die Nitroko¨rper der aliphatischen wie aromatischen Reihe mit den verschiedenartigsten organischen Verbindungen zu mehr oder weinigen tieffarbigen Additionsprodukten. Besonders gut charakterisiert sind vor allem die Verbindungen aromatischer Di-und Trinitroko¨rper mit aromatischen Kohlenwasserstoffen, Amine und Phenolen. Wie ein statistische Ueberschicht der bis heute dargestellten, etwa 700 Moleku¨l-verbindungen der Nitroko¨rper zeigt, haben diese in den allermeisten Fa¨llen, ganz unabha¨ngig davon, wieviele Nitrogruppen die nitroide Komponente entha¨lt, auch unabha¨ngig von der Zusammensetzung der benzoiden Komponente, die denkbar einfachste Zusammensetzung A1B1, indem auf 1 Moleku¨l des Nitroko¨rper 1 Moleku¨l des Kohlenwasserstoffs bzw. seiner Derivate kommt. Rund 85% der Verbindungen, von dene wenige aufgeza¨hlt seien, entsprechen diesen Typus. Paul Pfeiffer, 1927 (p. 336).

Summary: The room-temperature crystal structures of many mixed stack charge transfer molecular compounds can be grouped into a relatively small number of crystallochemical families. Although the mixed stack arrangement predominates, there are structures which deviate to a greater or lesser extent from such an arrangement, for reasons which are often not clear. We first review these maverick structures and then classify the more common mixed stack arrangements into a number of structural groups. Compounds with quinonoid acceptors sometimes have special structural features because of the mode of interaction of the carbonyl groups with aromatic rings. In the quinhydrone family the (aromatic ring) carbonyl interaction is supplemented by hydrogen bonding between carbonyl and hydroxyl oxygens and this leads to considerable structural homogeneity. Many CT compounds where the –* interaction is supplemented by hydrogen bonding show special structural features and physical properties different from those with only –* interaction. The mixed stack compounds with ionic ground states generally resemble those with neutral ground states in structural terms but have different physical properties. In the isomeric compounds the possibility of having both electron and proton transfer leads to variations on the mixed stack theme. n n n

15.1 15.2 15.3 15.4 15.5

Introduction Nonstacked structures containing structural groups of limited size The crystallochemical families found for 1 : 1 p–p* molecular compounds Packing arrangements in n : m p–p* molecular compounds Some special features of packing arrangements in p–p* molecular compounds 15.5.1 Crystals where one of the components is also found in interstitial positions 15.5.2 Noncentrosymmetric crystals of -molecular compounds

990 993 994 1001 1005 1005 1007

990

CRYSTAL CHEMISTRY OF MIXE D-STACK –* MOLECULAR COMPOUNDS

15.5.3 Acceptors based on polynitrofluorene 1009 15.5.4 Resolution of helicenes by formation of diastereoisomeric charge transfer molecular compounds with enantiomeric acceptors 1010 15.6 Structurally important interactions between polarizable and polar groups 1011 15.7 Mixed-stack crystals with both charge transfer and hydrogen bonding interactions 1013 15.7.1 The quinhydrones as a crystallochemical family 1013 15.7.2 Molecular compounds of the flavins 1022 15.7.3 Other crystals with both charge transfer and hydrogen bonding interactions 1026 15.8 Mixed-stack crystals with both delocalized and localized charge transfer interactions 1030 15.9 Donors and acceptors with special chemical features 1032 15.9.1 Fluorinated aromatics as quasi-acceptors 1032 15.9.2 1,3,5,7-Tetramethyluric acid (TMU) as quasi-acceptor 1040 15.9.3 Acceptor is a metal coordination complex 1040 15.9.4 Donor is a metal coordination complex 1042 15.9.5 Donors based on phenazine 1044 15.10 Mixed-stack donor–acceptor molecular compounds with ionized ground states 1047 15.10.1 Mixed-stack closed shell charge transfer salts 1047 15.10.2 Ion-radical salts 1048 15.11 Isomeric (polymorphic) molecular compounds 1052 15.11.1 Type 1 – isomerism due to different types of interaction without change of moiety structure 1052 15.11.2 Type 2 – isomerism due to electron transfer 1054 15.11.3 Type 3 – isomerism due to proton transfer or to –* electron transfer 1055 15.11.4 Isomerism stabilized by both charge (–*) and proton transfer (CPT compounds) 1058 15.12 Self-complexes 1059 15.13 Conclusions 1064 15.13.1 Structural variety in –* molecular compounds 1064 15.13.2 How should the packing arrangements in –* molecular compounds be described? 1065 15.13.3 Structural consequences of –* interactions 1066 (Note. The components in the ground states of these molecular compounds are taken to be neutral unless explicitly stated otherwise). References 1068

15.1

Introduction

Most –* molecular compounds have 1 : 1 donor acceptor (D A) ratios and the components crystallize in alternating array in mixed stacks, the stacks being close packed in a quasi-hexagonal arrangement of (approximate) cylinders (Fig. 15.1). We remind the reader of our conventions – donors always come first in the formulation and are linked to the acceptor by , indicating a –* charge transfer (CT) interaction; we have not been very strict in our usage of this indicator. Most donor and acceptor molecules are disk-like in shape, i.e. their cross-sectional area is appreciably greater than their thickness. ˚ and thus the The molecular thickness of aromatic donor and acceptors is about 3.5 A n n n

n n n

n n n


991

˚ for. . . . DADA. . . . stacks. The interplanar periodicity along the stack axis should be 7–8 A spacing, measured along the normals to the planes of the approximately parallel disks, ˚ less than the sum of the thickness of the components; this is is often found to be 0.1–0.2 A interpreted as evidence for a D A interaction along the normal to the component planes, or stack axis (see below), which is additional to the ubiquitous van der Waals interactions. There are a number of facts in support of this contention. Firstly, one can compare the transannular atom-to-atom distances, found from crystal structure analyses (Staab et al., 1983) in pseudo-geminal-5,8,14,17-tetramethoxy[3.3]paracyclophane and its quinhydrone analog pseudo-geminal-14,17-dimethoxy-[3](2,5)p-benzoquinone-[3]paracyclophane (see Chapter 14 for discussion of these compounds). The molecules have identical ˚ closer in the latter than in the former molecule, conformations but the two rings are 0.13 A in conformity with the existence of charge transfer interactions between the rings in the latter molecule but not in the former. A second line of evidence comes from comparison of Young’s modulus values in the direction of the stack axes, which are about ten times larger for crystals of -molecular compounds than for those of comparable aromatic hydrocarbons (Danno et al., 1967). Thus we choose to emphasize the mixed stacks as the essential and characteristic feature of the packing arrangements in -molecular compounds. The simple mixed stack description works remarkably well for most of the structures so far reported; however, it tends to break down when donor and acceptor molecules differ appreciably in disk (face) area or when there is additional bonding (e.g. hydrogen bonding or dipolar interactions) between the components. There are some crystals of this type where the structure can equally well be described in terms of layers of donor and acceptors. The extremes of packing type within the stacks can be described as ‘‘overlapped disks’’ (Fig.15.1(a)) and ‘‘slipped disks’’ (Fig.15.1(b)). However, intermediate situations are also found and there have been a number of essentially similar proposals (Fritchie and Arthur, 1966; Goldberg and Shmueli, 1973b; Visser et al., 1990) for a quantitative description of the arrangement within the stacks; we have adapted the proposal of Visser et al. (1990) to mixed stack compounds where the components lack symmetry. Right-handed orthonormal axial systems L, M, N are defined for the donor and acceptor molecules (or convenient portions, such as benzene rings); with L along the longest molecular axis, n n n

(a)

(b)

Stack Axis

(c)

Stack Axis

Fig. 15.1. Schematic representation of different types of stacking in crystalline mixed stack -compounds: (a) overlapped disk stacking, (b) slipped disk stacking and (c) quasi-hexagonal arrangement of stacks.

992


M perpendicular to L and in the best molecular plane, and N at right angles to the plane. The offset of the centre of gravity of the donor from the center of gravity of the acceptor is then given in terms of the components along the L, M and N axes of the acceptor (an arbitrary choice). A simplified situation for centrosymmetric molecules projected onto their mean planes is shown in Fig. 15.2. Results for a few molecular compounds are summarized in Table 15.1. Unfortunately most authors describe molecular overlap and calculate interplanar distances on a less well-defined basis. ˚ along the stack axis may require amendment in two The repetition period of 7 A ways : (a)

there are a number of crystals (Table 15.3) where the stack periodicity is doubled to ˚ and the arrangement along the stack is 14 A -------D1 A1 D2 A2 D1 A1 D2 A2------˚ | 14 A | The subscripts here refer to positional and/or orientational differences between donor (and acceptor) molecules, not to chemical differences. The causes of the ordering

Table 15.1. Examples of different stacking arrangements in mixed stack molecular compounds: I is {phenazine TCNQ} (Goldberg and Shmueli, 1973c; TCQPEN10); II is {dibenzop-dioxin TCNQ} (Goldberg and Shmueli, 1973b; TCQBDX); III is {anthracene TCNQ} ˚ and angles in deg (Williams and Wallwork, 1968; TCQANT). Distances in A n n n

n n n

n n n

Parameter

I

II

1. 2. 3. 4. 5.

8.57 38 1.85 1.89 3.38

7.04 11 0.51 0.48 3.46

Repeat distance along stack axis Angle between stack axis and plane normal Mutual offset along L Mutual offset along M Interplanar spacing

III 7.00 0 0.0 0.0 3.50

M N

N N

L

N

Fig. 15.2. {Phenazine TCNQ}, showing donor and acceptor molecules in a mixed stack projected onto their common mean plane. The long (L) and short (M) in-plane axes of TCNQ are shown. (Reproduced from Goldberg and Shmueli, 1973b.) n n n

NONSTACKED STRUCTURES

993

along the stack direction are not always clear; interactions between adjacent stacks are sometimes invoked. Higher degrees of ordering (i.e. periodicities greater than ˚ ) do not appear to have been encountered. 14 A (b) there are a number of crystals where the concept of infinite stacks no longer applies; instead the stack length is limited to one, two or three pairs. These finite stacks can be arranged in different ways. These ideas will be illustrated in more detail and also extended to -molecular compounds with D : A 6¼ 1.

15.2 Nonstacked structures containing structural groups of limited size The smallest structural group of limited size is the donor–acceptor pair, not incorporated in a stack. Such arrangements are found in {benz[a]-anthracene PMDA} (Foster et al., 1976; BZAPRM10) and {1,10-phenanthroline TCNQ} (Goldberg and Shmueli, 1977; TCQPAN10). The overall crystal structures can be described in terms of mutually shifted layers of donors and acceptors. Analogous donor–acceptor pairs are found in the crystal structures of tryptamine picrate (TRYPIC) and dl-tryptophan picratemethanol (TPTPCM) (Gartland et al., 1974). The red colour of these crystals attests to the occurrence of charge transfer interactions within the donor–acceptor pairs; neighbouring pairs are linked by hydrogen bonds and not by –* interactions.1 n n n

n n n

H tryptamine: R+ = –CH2CH2NH3+

N H

R+

tryptophan: R+ = –CH2CH(CO2H)NH3+

In {(9-ethylcarbazole)2 TCNE} (Lee and Wallwork, 1978; ETCZCE) the structure ˚ between donor and consists of centrosymmetric DAD units, with a distance of 3.24 A acceptor molecules. These sandwiches are arranged in face-centred pleated sheets, with mean plane (010); however, the pleats are alternately parallel to (021) and (021); there being an angle of 60 between these planes. Successive sheets are related by a c glide plane perpendicular to [010]. The analogous arrangement found in {(acridine)2 PMDA} (Karl, Binder et al., 1982; BIWVUY) is shown in Fig. 15.3. Centrosymmetric ADA sandwiches are found in the 1 : 2 compound of a Ni(II)etioporphyrin with 2,4,5,7-tetranitrofluorenone (P21/n, Z ¼ 2) (Grigg et al., 1978; ETPNFL); and in {2,7-bis(methylthio)1, Z ¼ 1) (Nakasuji, Sasaki et al., 1,6-dithiapyrene- (tetrahydrobarreleno-TCNQ)2} (P 1988; SANYOV). Effectively-isolated DAD triads are found in (toluene)2 tetraphenylporphyrinato-M(II) molecular compounds (M ¼ Cr, Mn, Zn; see Chapter 8 and Section 15.9.4). In the 2 : 1 compound of N,N,N 0 ,N 0 -tetramethylbenzidine and chloranil n n n

n n n

n n n

n n n

1 Mixed stacks are found in three tryptophan metabolite–picric acid molecular compounds; their red colour shows that there is charge-transfer interaction (Nagata et al., 1995).

994


Fig. 15.3. The crystal structure of {(acridine)2 PMDA} at 120K, showing a stereoscopic view down [100] of one layer of the structure ([001] is vertical and [010] runs from left to right). The tilt of the acridine molecules with respect to the PMDA molecules can be seen clearly. (Reproduced from Karl, Binder et al., 1982.) n n n

the two benzene rings of the benzidine moiety have an interplanar (twist) angle of 31 and the chloranil is sandwiched between two parallel rings of successive benzidines, the second ring of each benzidine molecule not participating in the charge transfer interaction (Yakushi, Ikemoto and Kuroda, 1971; TMBCAN). There are a number of analogous structures where the ground state is ionic and these are discussed in Section 15.10. No overall explanation has yet been put forward for these exceptional structures.

15.3

The crystallochemical families found for 1 : 1 p–p* molecular compounds

The crystal structures of most of the 1 : 1 –* molecular compounds can be grouped into a limited number of crystallochemical types; we use number of formula units in the unit cell (Z ); space group and stack axis direction as criteria for this classification, restricting ˚ (however, some exceptions are ourselves at this point to stack axis periodicities of 7 A ˚ ). The cell admitted when the donor or acceptor molecule is thicker than the usual 3.5 A dimensions in directions normal (or approximately so) to the stack axis are determined mainly by the cross-sectional dimensions of the stacks and their mode of arrangement. As noted above, there is generally quasi-hexagonal close packing of the stacks, all stack axes being parallel. A few exceptions have been found to this rule; for example, in {pyrene pbenzoquinone} (Bernstein et al., 1976; PYRBZQ) (tetragonal, space group P41, Z ¼ 4) ‘‘slipped disk’’ stacking of the usual kind is found but the stacks are arranged in layers one stack thick, successive layers being related by the 41 axes (Fig. 15.4). No explanation has n n n

C4

2.8

8

C11

C4 C1 C2

b

C9

995

O4

C11

2.4

5

3.4

9

O4 C2

O4

2.8 8

T HE CRYSTALLOCHE MICAL FAMILIES

2.4

C9

9

01

a 0

1

2Å

Fig. 15.4. Part of the {pyrene p-benzoquinone} structure, viewed down [001]. The stack axes are alternately along the [100] (open circles) and [010] (full circles) axes of the tetragonal unit cell, and the layers are related by the fourfold screw axis along [001]. Some short distances between ˚ would now be considered as stacks are shown. In particular, the (C)–H . . . O¼C distance of 2.49 A evidence for a weak hydrogen bond. (Reproduced from Bernstein et al., 1976.) n n n

been put forward for this unusual arrangement, which is chiral (but the absolute configuration of the crystal used in the analysis was not determined). A distorted version of this structure type is found in {9-methoxy-5,11-dimethyl-6Hpyrido[4,3-b]carbazole TCNQCH3CN} (P21/c, Z ¼ 4) (Viossat, Dung and Daran, 1988; GEZKIF), where there are mixed stacks with axes approximately along [110] and [110], arranged in successive layers about z 0 and 1/2. A similar arrangement is found in ˚ , Z ¼ 16, space group I41/a) (Bravic 1,3-indandione (tetragonal, a ¼ 14.361, c ¼ 13.631 A et al., 1976; INDDON); which can be considered to be a self-complex, with the two parts of the molecule having donor and acceptor properties respectively. It seems probable that there are both dipole–dipole and –* interactions. Most of the other structures can be classified as shown in Table 15.2; the scheme is based on that developed earlier (Herbstein, 1971; see Table 20) but the numbering of the groups has been changed to match the usual crystallographic hierarchy – triclinic, monoclinic, etc.; primitive, centred unit cells; molecules at special positions, molecules at general positions. Structures have been determined for most but not all of the compounds listed. The most prolific acceptor and donor components are TCNQ and TTF. There are 1161 hits for TCNQ in the October, 2002 issue of the CSD, and 253 for TTF, both numbers including derivatives and covering binary adducts of all kinds. Table 15.2, which includes only mixed-stack structures, has 48 examples with TCNQ as acceptor and 16 with TTF as donor. n n n

996


Table 15.2. Classification of 1 : 1 –* molecular compounds into structural groups. In each group the donors that form molecular compounds with particular acceptors are listed; in general only the most recent reference is given. Alternative orientations (e.g P21/c instead of P21/a with alternative stack axis) are not listed separately. Acronyms are listed in Table 14.2. Z is the number of formula units in the unit cell Acceptor

Donors

Group 1a: Triclinic, P1, Z ¼ 1, stack axis [001] TCNE C6(CH3)6, (Sahaki et al., 1976; MBZTCE; Maverick et al., 1978; MBZTCE10) [3,3]paracyclophane (Bernstein and Trueblood, 1971; PACTCN10); ferrocene (Adman et al., 1967; FERTCE); [2,2]metacyclophane (Cohen-Addad, Renault et al., 1988; GEBREK). 2,5-dimethyl-TCNQ Octamethylene-TTF (OMTTF) (Chasseau and Leroy, 1981; BESPEU). fluoranil TTF (Mayerle et al., 1979; TTFFAN); durene (Dahl and Sørensen, 1985); N,N-dimethylaniline (disordered over two orientations) (Dahl, 1981b; BAPLEJ). chloranil Pd(II)oxinate (Kamenar et al., 1965; CLAQPD); bis(8-hydroxyquinoline) (Prout and Wheeler, 1967; HQUCLA); perylene (Kozawa and Uchisa, 1983; CAFWAH); N,N,N 0 ,N 0 -tetramethylbenzidine (TMBD) (Yakushi et al., 1973; MBZDCN). Perylene (Kozawa and Uchida, 1979; PERPBQ); acceptor disordered 2,5-dibromo-3,6at inversion center. dichloro-pbenzoquinone TCNB Pd(II)oxinate (Kamenar, Prout and Wright, 1966; PDHQCB); TMPD (Ohashi et al., 1967; TPDTCB); biphenyl (Pasimeni et al., 1983; BUHSIG); acridine (disordered) (Marsh, 1990; KARKAP01). TCNQ Naphthalene (Shaanan et al., 1967; TCQNAP); chrysene (Munnoch and Wright, 1974; CHRTCQ); d14-p-terphenyl (Lisensky et al., 1976; TCQDTP10, ND); phenazine (Goldberg and Shmueli, 1973b; TCQPEN10); Cu(II)oxinate (Williams and Wallwork, 1967; TCQCUH); Pt(II)oxinate (Bergamini et al., 1987; FEFCAU); bis(1,2-benzoquinonedioximato)-Pd(II) (Keller et al., 1977; BCDPDQ); bis(1,2-benzoquinonedioximato)Ni(II) (Keller et al., 1977; ZZZATV); bis(propene-3-thione-1-thiolato)Pt(II) (Mayerle, 1977; PRTTCQ); dibenzofuran (may be ordered in P1) (Wright and Ahmed, 1981); TMTSF (Kistenmacher et al., 1982; SEOTCR); dibenzotetrathiafulvalene (DBTTF; ¼ 0.47e) (Kobayashi and Nakayama, 1981; Emge, Wijgul et al., 1982; BALNAD); 9,9-trans-bis-(telluraxanthenyl) (Lobovskaya et al., 1983); E-DMDBTTF (Shibaeva and Yarochkina, 1975); Octamethylene-TTF (OMTTF) (Chasseau et al., 1982; BESPEU); 2,2 0 ,5,5 0 -tetramethoxystilbene (Zobel and Ruban, 1983; TMXSTQ10). Tetrakis(methylthio)TTF (Mori, Wu et al., 1987; FIJYEC). Bis(ethylenedithio)TTF (Mori and Inkuchi, 1986; FAHLEF). There is also a monoclinic polymorph.


997

Table 15.2. (Continued ) Acceptor

Donors

2,5-difluoro-TCNQ PMDA

DibenzoTTF (Emge, Wijgul et al., 1982; BITROL) Anthracene (Robertson and Stezowski, 1978; ANTPML at 153 and 300K); phenazine (Bulgarovskaya et al., 1982; Karl, Ketterer and Stezowski, 1982; BECNUS02); acridine (disordered) (Binder et al., 1982; BIHBUP10) {previous three examples isomorphous}; tetracene (Bulgarovskaya et al., 1987a; FILHOK). p-xylene (Dahl, 1975a; PXYHFB); TMPD (Dahl, 1979; MPAHFB); C6(CH3)6 (at 233K) (Dahl, 1973; MBZFBZ01). Pyrene (Collings et al., 2001; ECUVIH). Perylene (Schmitt et al., 1969; PERNIT).

C6F6

octafluoronaphthalene bis(cis-1, 2-prefluoromethylethylene-1,2dithiolato)Ni(II) 2,3,5,6-tetracyanohydroquinone p-dinitrobenzene

Pyrene (Bock, Seitz et al., 1996; TEXPOB10) TTF (Bryce, Secco et al., 1982; BIRDIP).

Group 1b: Triclinic, P1, Z ¼ 2, stack axis [001] TCNQ Acenaphthene (Tickle and Prout, 1973c; ACNTCQ); 5-pheny1-,3-thiaselenole-2-thione (Kaminski et al., 1979; PTSTCQ); dibenzothiophene (Wright and Ahmed, 1981; BAHFEV). phenothiazine (Toupet and Karl, 1995; PTZTCQ01). PMDA Phenothiazine (Anthonj et al., 1980; PTZBMA). TCPA d8-naphthalene (Wilkerson et al., 1975; DNPCPH at 120K). TNB Pyrene (Prout and Tickle, 1973b; PYRTNB); acepleiadylene (Hanson, 1966; APANBZ); tetrabenznaphthalene (Herbstein et al., 1976); dibenzothiophene (Bechtel et al., 1977; DBTTNB); TTF (Bryce and Davies, 1987; GASGUC); phenanthrene-chromium tricarbonyl (De et al., 1979; CPCTNB); 12-imino-12H-benzimidazo [2,1-b]-[1,3]benzothiazine (Viossat et al., 1995; ZAYQEV). TCNB 2,3,6,7-tetramethoxythianthrene (Bock, Rauschenbach et al., 1996; RIKYUF) picric acid Tetrabenznapthalene (Herbstein et al., 1976; ZZZAHA). 3,5-dinitrobenzoic acid Phenothiazine (Fritchie and Trus, 1968; PHTNBA). BTF 13,14-dithiatricyclo[8,2,1,14,7]tetradeca-4,6,10,12-tetraene (Kamenar and Prout, 1965; BOXTET). 2,6-dichloro-N-tosyl-1,4Pyrene (Shvets et al., 1980; PYTQIM). benzoquinonemonoimine octafluoronaphthalene Triphenylene (Collings Roscoe et al., 2001; ECUVON). Group 2a: Monoclinic, P21/a, Z ¼ 2, stack axis [001] TCNE h10-and d10-pyrene (Larsen et al., 1975; PYRTCE); [2.2] (9,10)-anthracenophane (Masnovi et al., 1985; DIRKIY). Perylene (Ikemoto et al., 1970; PERTCE10)

998



Donors

TCNB

anthracene (below 200K) (Stezowski, 1980; ANTCYB11); pyrene (Prout, Morley et al., 1973; PYRCBZ at 178 and 300K); C6(CH3)6 (Niimura et al., 1968; CYBHMB); p-phenylenediamine (PD) (Tsuchiya et al., 1973; PDTCNB); biphenylene (Stezowski et al., 1986; Agostini et al., 1986; DURYUK);, durene (Lefebvre et al., 1989; KARHAM). Perylene (Yamachi et al., 1987). Perylene (Tickle and Prout, 1973a; PERTCQ); pyrene (Prout, Tickle and Wright, 1973; PYRTCQ); dibenzo-p-dioxin (Goldberg and Shmueli, 1973b; TCQBOX); 1,2-di(4-pyridyl)ethylene (Ashwell et al., 1983; BUKXOU) (red, non-conducting crystals; black crystals were also reported); 6,13-diacetyl-5,14-dimethyl-1,4,8,11-tetraazacyclotetradeca-4,6, 11,13-tetraenenickel(II) (Lopex-Morales et al., 1985); 4, 6, 8-trimethylazulene (Hansmann et al., 1997a; ROJYUK) (donor disordered at inversion center). Bis(ethylenedithio) TTF (Mori and Inokuchi, 1987; FAHLEF01); monoclinic polymorph. OMTTF (Chasseau and Hauw, 1980; OMTFNQ). 2,7-bis(methylthio)-1,6-dithiapyrene (Toyoda et al., 1993; PIGYUK).

HCBD TCNQ

Dimethoxy-TCNQ (tetracyano-2,6napthoquinodimethane (TNAP) PMDA

1,8-4,5-naphthalene tetracarboxylic dianhyride Pyromellitic dithioanhydride N,N 0 -dimethylpyromellitic di-imide) 1,4-dithiintetra carboxylic N,N 0 -dimethyldiimide p-benzoquinone fluoranil chloranil

BTF

pyrene (above 160K) (Herbstein and Snyman, 1969; Allen et al., 1989; PYRPMA); carbazole (disordered) (Stezowski, Binder and Karl, 1982; BIWVOS); biphenylene (red, stable above 400K. (Stezowski, Stiegler and Karl, 1986; DURZAR). Antharacene (Hoier et al., 1993; WABWEB); dibenz[a,h]anthracene (Zacharias, 1993; BZANTC10). Anthracene (Bulgarovskaya et al., 1974; TPYMAN); acridine (Bulgarovskaya et al., 1976; ACRTMA); dibenzothiophene (Bulgarovskaya et al., 1978; PMTABT); (all three isostructural). Anthracene (Bulgarovskaya et al., 1977; PMEANT). Acridine (Yamaguchi and Ueda, 1984; space group correction by Marsh, 1986; CEJTAM). TTF (Frankenbach et al., 1991; SIVBAA). Perylene (Hanson, 1963; PERFAN). Pyrene (Prout and Tickle, 1973c; PYRCLN); 9-methylanthracene (Prout and Tickle, 1973a, MANTCB) (disordered); TTF (Ohashi et al., 1967; Mayerle et al., 1979; TTFCAN). copper oxinate (BTF is disordered) (Prout and Powell, 1965; ZZZGDI)


999


Donors

benzo[1,2-c;4,5-c 0 ]bis[1,2,5]-thiadiazole4,8-dione 1,4,5,8-naphthalenetetrone 3,3 0 ,5,5 0 -tetrachlorodiphenoquinone octafluoronaphthalene

TTF (Gieren et al., 1984; CIYNUT)

C6F6 bis(difluoroborondi methylgloximato)Ni(II)

Pyrene (Herbstein and Reisner, 1984; CEKBUP) Anthracene (Starikova et al., 1980; ANPHXN). Naphthalene (Potenza and Mastropaolo, 1975; NPOFNP); tolan (Collings, Batsanov, et al., 2001; OCAYIA). anthracene (Collings, Roscoe, et al., 2001; ECUTUR). dimer of o-diethynylbenzene (Bunz and Enkelmann, 1999; JOCRIC); Perylene (Boeyens and Herbstein, 1965a; ZZZLJY) Anthracene (Stephens and Vagg, 1981; BADZOV).

Group 2b: Monoclinic, P21/a, Z ¼ 4, stack axis [001] TCNB N,N-dimethylphenylenediamine (Ohashi, 1973; DMPTCN); perylene (Bock, Seitz et al., 1996; REHMUM). PMDA Phenanthrene (Evans and Robinson, 1977; PENPYM). TNB s-triaminobenzene (Iwasaki and Saito, 1970; NIBZAM); p-iodoaniline (Powell, Huse and Cooke, 1943; IANNOB); tricarbonylchromium anisole (Carter et al., 1966; CCATNB); 3-formylbenzothiophene (Pascard and Pascard-Billy, 1972; TNBFTB); DBTTF (Lobovskaya et al., 1983). picric acid Anthracene (Herbstein and Kaftory, 1976; ANTPIIC). TCNQ dithieno (3,2-b 0 ;2 0 ,3-d)thiophene ((Zobel and Ruban, 1983; Bertinelli et al., 1984; CAPTOC); dithieno [3,4-b : 3 0 ,4 0 d]thiophene (Catellani and Porzio, 1991; VIGTAG); 2,4,7-trinitrofluorenone C6(CH3)6 (Brown, Cheung et al.,1974; TNFLMB). 2,4,6-trinitro-anisole pyrene (disordered over two orientations) (Barnes et al., 1984; CILRAQ). Group 3a: Monoclinic, P21/a, Z ¼ 2, stack axis [010] TCNE Perylene (Ikemoto et al., 1970; PERTCE10); DBTTF (Lobovskaya et al., 1983). TCNB Hydroquinone (Bock, Seitz et al., 1996; REHNAT) TCNQ bis (ethylenedithio)TTF (Mori and Inokuchi, 1987; FAHLEF01) (also triclinic isomer with segregated stack). TCNQF4 Trans-stilbene (Sato et al., 2001; QILZOA). PMDA Benzene (Boeyens and Herbstein, 1965a; ZZZKSM); naphthalene (orange polymorph) (Bar-Combe et al., 1979; NAPYMA); perylene (Boeyens and Herbstein, 1965b; PERPML); trans-stilbene (Kodama and Kumakura, 1974a; PYMAST); chrysene (Bulgarovskaya et al., 1987b; FILHIR).

1000 CRYSTAL CHEMISTRY OF MIXE D-STACK –* MOLECULAR COMPOUNDS


Donors

fluoranil

Pyrene (Bernstein and Regev, 1980; PYRFLR); chrysene (Munnoch and Wright, 1975; CHRFAN). Pyrene (Shchlegova et al., 1981; C1–BAZDAH, Br–BAZCUA).

3,3 0 ,5,5 0 -tetrahalodiphenoquinone (halo ¼ Cl,Br)

Group 3b: Monoclinic, P21/a, Z ¼ 4, stack axis [010] TCNQ (tetramethyl)porphyrinato)Ni(II) (Pace et al., 1982; BEGLUU); dibenzotellurophene (Singh et al., 1984). TNB Naphthalene (Herbstein and Kaftory, 1975a; PVVBKP); azulene (Hanson, 1965a; AZUNBZ; Brown and Wallwork, 1965); skatole (133K) (Hanson, 1964; SKINIB); indole (133K) ) (Hanson, 1964; INTNIB). picric acid Naphthalene (Banerjee and Brown, 1985; PVVBHJ01). picryl chloride 9-isopropylcarbazole (Cherin and Burack, 1966; ZZZMCI). picryl bromide fluoranthene (Herbstein and Kaftory, 1975a; FLABPC). BTF Perylene (Boeyens and Herbstein, 1965a; ZZZMHI); Benzene (Boeyens and Herbstein, 1965a; ZZZLDA). Group 4a: Monoclinic, C2/m, Z ¼ 2, stack axis [001] PMDA naphthalene (ordered yellow form) (Le Bars-Combe et al., 1979; NAPYMA01); 9,10-dibromoanthracene (Bulgarovskaya, Belsky et al., 1987; FILHEN). TCNE Naphthalene (Shmueli and Goldberg, 1974; CYENAP01). TCNB naphthalene (above 68K) (Kumakura et al., 1967); anthracene (above 200K) (Stezowski, 1980; ANTCYB12); phenanthrene (donor disordered) (Wright et al., 1978; PHTCBZ); -naphthol (NAPTCC10), -naphthol (BNATCB10) (both Couldwell and Prout, 1978). TCNQ Anthracene (Williams and Wallwork, 1968; TCQANT); benzidine (Yakushi et al., 1974a; BZTCNQ); N-methyl-phenothiazine (Kobayashi, 1973, 1987; TCQNMP10); dithienylethene (Zobel and Ruban, 1983; CAPTAO); trans-stilbene (Zobel and Ruban, 1983; STILTQ10); carbazole (donor disordered) (Mitkevich and Sukhodub, 1987). TMPD (Hanson, 1965b; QMEPHE) C6F6 Anthracene (ZZZGMW); pyrene (ZZZGKE) (both Boeyens and Herbstein, 1965a); durene (Dahl, 1975b; DURHFB). octafluoronaphthalene Phenanthrene (Collings, Batsanov et al., 2001; ECUVED) Note: {4,6,8-trimethylazulene TCNE} is a rare example crystallizing in space group P2/m, Z ¼ 4, stack axis [010] (Hansmann et al., 1997b). n; n n

PACKING ARRANGEMENTS IN n:m – * M OL ECULAR COMPOUNDS

1001

We cite three examples of structures of the 1 : 1 mixed stack type which crystallize in a space group other than those listed in Table 15.2. {5,8-Dimethoxy-2,11dithia[3,3]paracyclophane TCNE} crystallizes in space group C2/c, Z ¼ 4, with mixed stacks along [010] (Cohen-Addad, Consigny et al., 1988). Both components lie on two fold axes; the donor molecule is chiral and so are individual stacks as successive donors are related by translation; the space group is, of course, centrosymmetric. The interplanar spacings between TCNE and p-dimethoxyphenyl and phenyl rings are respectively 3.15(1) ˚ , indicating, as would be expected, that p-dimethoxyphenyl is a stronger and 3.33(1) A donor than phenyl. {TTF m-dinitrobenzene} also crystallizes in space group C2/c, Z ¼ 4, but the mixed stacks lie along [100], with TTF at centres of symmetry and mdinitrobenzene on two fold axes (Bryce et al., 1988). There is little overlap of components and their planes are mutually inclined at 8 ; nevertheless, the black color of the crystals suggests appreciable charge transfer (in the excited state). The low conductivity along the stack axis ( 109 S/cm) and other physical properties show that the ground state is neutral. {Chrysene TNB} crystallizes in space group Pna21, Z ¼ 4, with both components in general positions; there are mixed stacks along [001] (Zacharias et al., 1991; VIGKIF {3-Methylchrysene TNB} (VIGLAY) is isomorphous, with disorder of the donor. ˚ , except a few The structures in Table 15.2 all have stack axis periodicities of 7 A ˚ examples where donor and/or acceptor thickness is greater than 3.5 A. Although {trans˚ stack axis periodicity, stilbene PMDA} (Kodama and Kumakura, 1974a) has a 12.48 A the arrangement of the components justifies its inclusion in Table 15.2; both molecules lie about centres of symmetry, but there are stepped stacks, with each phenyl group of the donor being overlapped on one side only by an anhydride portion of the PMDA acceptor. However, in about 10% of 1 : 1 donor acceptor molecular compounds the stack axis ˚ , a possibility already noted in Section 15.1. A somewhat random periodicity is 14 A selection of examples is given in Table 15.3. Possibly some of these crystals have order-disorder transitions leading to doubling of the stack periodicity as occurs in {pyrene PMDA} (Herbstein and Samson, 1994); but this has hardly been investigated. n n n

n n n

n n n

n n n

n n n

n n n

n n n

15.4 Packing arrangements in n : m p–p* molecular compounds D : A ratios of 1 : 2 or 2 : 1 are found in a rather small fraction (roughly 5%) of the crystal structures that have been reported. Other ratios (e.g. 3 : 2, 3 : 4) are much less common. The pyrene-picryl chloride system is remarkable in that five molecular compounds with mole ratios of 4 : 1, 3 : 1, 2 : 1, 1 : 1 and 1 : 3 respectively have been reported on the basis of a DSC study (Bando and Matsunaga, 1976) (but see Section 13.4 for conflicting results). Crystal data have been reported only for the 1 : 1 compound (Herbstein and Kaftory, 1975a). The 1 : 2 (or 2 : 1) compounds are generally found in one or other of two structural groups. In the first group the donor (say) has a much larger cross-sectional area than the acceptor and can thus behave as a bifunctional donor, with two acceptor molecules sandwiched between each pair of donors. This arrangement is found in {copper oxinate2 (TCNB)2} and in {copper oxinate (picryl azide)2} (Bailey and Prout, 1965; n n n

2

n n n

Copper oxinate is bis (8-hydroxyquinolinato)Cu(II).


˚ periodicity along the stack Table 15.3. 1 : 1 donor-acceptor molecular compounds with 14 A axis; see also Section 15.5.1. Only the higher symmetry space group is given when the systematic absences do not unequivocally determine the space group Z

Crystal structure reported?

P1

2

No

13.94

P1

2

No

[101]

13.8

P1

2

Yes*

n n n

[001]

13.93

P1

4

No

n n n

[001]

14.5

P21/n

4

Yes*

[001]

13.7

P21/c

4

No

[001]

13.83

P21/c

4

Yes

[001]

14.57

P21/c

4

Yes

[001]

13.72

P21/c

4

Yes

[001]

14.61

P21/c

4

Yes

[001]

13.7

P21/c

8

No

[001]

13.02

C2/c

4

Yes

[001]

13.84

C2/c

4

Yes

n n n

[001]

14.4

A2/a

8

No

n n n

[001]

14.54

A2/a

8

No

[001]

14.13

Pcmm

8

No

[100]

14.0

Amam

4

No

[100]

14.63

Pnca

4

Yes

[001]

14.53

P21/c

4

Yes

[201]

14.73

P21/n

4

Yes

Molecular compound

Stack axis along

Stack axis periodicity ˚) (A

{naphthalene picryl chloride} (Herbstein and Kaftory, 1975a); {naphthalene picryl bromide} (Herbstein and Kaftory, 1975a); PVVBHG {pyrene dicyanomethylene-croconate} (Doherty et al., 1982); BEFGIC {pyrene picryl chloride} (polymorph II) (Herbstein and Kaftory, 1975a) {pyrene PMDA} (below 160K) (Herbstein et al., 1994) {phenanthrene TNB} (Herbstein and Kaftory, 1975a) {benzo[c]phenanthrene DDQ} (Bernstein et al., 1977); BZPCBQ {4-(2-hydroxyethyl)carbazole DDQ} (Qi et al., 1996); TEJGUK {2,3,7,8-tetramethoxythianthrene TCNQ} (D’yachenko et al., 1977); Hinrichs and Klar, 1982; MXTTCQ {(pyrene)3 (picryl bromide)2} (Herbstein and Kaftory, 1975b); PYRBPC {guiacol picric acid} (yellow polymorph) (Herbstein, Kaftory and Regev, 1976); ZZZAHG01 {anthracene TNB} (at 173K) (Brown et al., 1964); ANCTNB {trithia[5]heterohelicene TCNQ} (Konno et al., 1980); THLCTC {triphenylene picryl chloride} (polymorph II) (Herbstein and Kaftory, 1975a); PVVBEY01 {triphenylene picryl bromide} (polymorph II) (Herbstein and Kaftory, 1975a); PVVBEV {benzene picric acid} (Herbstein, Kaftory and Regev, 1975); ZZZAGV {C6(CH3)6 picryl chloride} (Powell and Huse, 1943)x {resorcinol p–benzoquinone} (Ito et al., 1970); BZQRES {5,10–dihydro-5,10-diethylphenazinium TCNQ} (Dietz et al., 1982); BEWBUA {dipyrido-1,3,4,6-tetraazapentalene TNF} (Groziak et al., 1986); FARMAM n n n

[001]

13.81

n n n

[001]

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

Space group

n n n

n n n

PACKING ARRANGEMENTS IN n:m – * M OL ECULAR COMPOUNDS

1003

Table 15.3. (Continued ) Molecular compound

Stack axis along

Stack axis periodicity ˚) (A

Space group

Z

Crystal structure reported?

{-2,7-bis(methylthio)-1,6dithiapyrene TCNQ} (Nakasuji et al., 1987); FUDTON01

[100]

15.08

P21/c

4

Yes

n n n

Notes: * two independent pyrenes at inversion centers; acceptor at general position. x diffuse scattering ignored. The C6(CH3)6 picryl bromide and picryl iodide molecular compounds are similar but the diffuse scattering is more complex. n n n

PAZQCU). In these two molecular compounds the stacks can be represented schematically as shown below and thus are essentially 1 : 1 in character; consequently the stack ˚ , as before. Analogous situations are found in axis periodicity will be about 7 A {bis(N-isopropyl-2-oxy-1-naphthylidene-aminato)Cu(II) (TCNQ)2} (Matsumoto et al., 1979; IPONTC) and in {9,10-dihydroanthracene (TNB)2} (Herbstein et al., 1986; ZZZAGS10). Relative donor and acceptor sizes cannot be the only factor in determining the stability of these 1 : 2 compounds as palladium oxinate forms a 1 : 1 compound with TCNB (Kamenar et al., 1966; PDHQCB). n n n

n n n

Donor A

A Donor

A

A

The second group of molecular compounds is characterized by stacks of -----DAD DAD DAD DAD----˚ (3 3.5 A ˚ ). type (for a 2 : 1 composition). The stack axis periodicity will be about 10.5 A A striking example is found in the {(perylene)3 TCNQ} compound (Fig. 15.5) (Hanson, 1978; TCQPER); the role of the perylene molecule outside the stacks is discussed below (Section 15.5.1). Rather similar stack arrangements (but without the additional interstitial molecules) are found in the isomorphous pair {benzo[c] pyrene (TMU)2} (Damiani, Giglio, Liquori and Ripamonti, 1967; TMUBZP10) and {coronene (TMU)2} (Damiani, Giglio, Liquori, Puliti and Ripamonti, 1967) and in {stilbene (TNB)2} (Bar and Bernstein, 1978; STINBZ), {DBA (TNB)2} (Zacharias, 1976), {(HMTSF)2 TCNQ} (Emge et al., 1982; BOWSUB), {(BTT)2 TCNQF4} (Sugano et al., 1988; SAJMEV) (BTT is the hexaradialene benzo[1,2-c : 3,4-c 0 : 5,6c 0 0 ]trithiophene), {(HMB)2 TCNE} (IR study, crystal structure has not been reported; Hall and Devlin, 1967), and {(TTM-TTF)2 TCNQ} (Mori, Wu et al., 1987; FIJYAY; there is also a 1 : 1 compound FIJYEC). n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n


x

y

Fig. 15.5. Projection down [100] of the {(perylene)3 TCNQ} crystal structure showing – DADDAD– stacks and interstitial perylene molecules. The crystals are triclinic and the stack axis ˚ . (Reproduced from Hanson, 1978.) periodicity is 10.422 A n n n

Charge-transfer mean axis

(111)

(010)

B iii C

iv A

TCNE(2)

C A i (i) x, y, z (ii) 1 + x, y, 1 + z (iii) 1 − x, 1 − y, 1 − z (iv) −x, 1 − y, −z

ii

(000) TCNE(1)

(101)

Fig. 15.6. Intermolecular arrangement in {[2.2.2]paracyclophane TCNE}; the molecules are projected onto the plane defined by the barycentres of the phenyl rings. (Reproduced from CohenAddad et al., 1984.) n n n

One method of forcing the formation of ---DADDAD--- type stacks is to use a cyclophane donor where the two potential donor portions have very different donor strengths. Thus in {(5,7-[12]-paracyclophanediyne)2 TCNE}, the TCNE acceptor is sandwiched between benzene rings of two different donor molecules, with an interplanar distance of n n n

SPECIAL FEATURE S O F PACKING ARRANGEM ENTS

1005

˚ , while there is no interaction between triple bonds and TCNE (Harata et al., 1972; 3.27 A PCDTCN). In contrast, mixed stacks of the usual type are found in {1,5naphthaleno(2)paracyclophane TCNE}, with the distances between TCNE and ben˚ (Irngartinger and zene and naphthalene planes essentially equal at 3.464 and 3.457 A Goldman, 1978); a similar situation is found in {[2.2.2.2]paracyclophane TCNE} (Cohen-Addad et al., 1984; COMHOB). An unusual variant on mixed stacking is found in {[2.2.2]paracyclophane TCNE} (Cohen-Addad et al., 1984; COMHIV), where only two of three benzene rings participate in charge transfer interactions with TCNE (Fig. 15.6). In the molecular compounds considered up to this point, the donor molecules have all been planar or approximately so. However, in the isomorphous {transazobenzene (TNB)2} (ABTNBA) and {N-benzylideneaniline (TNB)2} (ABTNBB) (Bar and Bernstein, 1981) there are angles of 42 and 48 between the planes of the phenyl rings in the donor molecules and thus there is some disruption of the stacking leading to a tendency to form DAD units. This tendency reaches an extreme in {(N,N,N 0 ,N 0 tetramethylbenzidine)2 chloranil} (Yakushi et al., 1971; TMBCAN) where isolated (nonstacked) DAD sandwiches are found (see Section 15.2). Finally we note that {trans-4-methylstilbene (PMDA)2} has a rather complicated disordered structure (Kodama and Kumakura, 1974b; PYMSTL). n n n

n n n

n n n

n n n

n n n

n n n

n n n

15.5 Some special features of packing arrangements in p–p* molecular compounds 15.5.1 Crystals where one of the components is also found in interstitial positions Perhaps some thirty -molecular compounds with compositions other than 1 : 1, 1 : 2 or 2 : 1 have been reported (for earlier work see Table 12 of Herbstein, 1971). Some of these compositions require authentication but crystal structure analysis does provide explanations for the unusual compositions of {(pyrene)3 (picryl bromide)2} (Herbstein and Kaftory, 1975b; PYRBPC) and {(perylene)3 TCNQ} (Hanson, 1978; TCQPER). In these two compounds there are respectively 1 : 1 and 2 : 1 donor acceptor stacks of the usual types, with the additional molecules in interstitial positions where they do not participate in the charge transfer interaction (see Fig. 15.5). There are a number of examples of interstitial donors among the molecular compounds of the flavins and these are noted in Section 15.7.2. The fact that a particular component can play two (or more) different roles in a crystal structure is not new; a classical example already noted is CuSO45H2O where four water oxygens and two sulphate oxygens are coordinated octahedrally about Cu and the fifth water molecule is present as solvent of crystallization (Beevers and Lipson, 1934). It is possible that {(coronene)3 TCNQ} (Truong and Bandrauk, 1977) has a structure similar to that of {(perylene)3 TCNQ}. {5,6-Dihydropyrimidino[5,4-c]carbazole}3 TCNQ2H2O also comes into this general group. There are mixed. . . . DADDAD . . . stacks (TCNQs at centers of symmetry) with additional donor molecules in interstitial positions. The two types of donor are mutually hydrogen bonded and also to the water molecules (Dung et al., 1986; DULFAR). {Phenothiazine (PMDA)2} (Brierley et al., 1982) is another example with 1 : 1 donor acceptor stacks, with the additional PMDA n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n


separating the stacks from one another. The stacking arrangements in the 1 : 1 stacks are appreciably different in {phenothiazine PMDA} and in {phenothiazine (PMDA)2}. {(TMTTF)1.3 (TCNQ)2}, which is a segregated stack molecular compound (Chapter 17), is mentioned here because the additional 0.3 molecule of TMTTF is inserted interstitially between the stacks of TMTTF and (TCNQ)2 with its molecular plane parallel to the stack axis (Kistenmacher et al., 1976; SEOTCR01). The isostructural -compounds {bis(3,6-dibromocarbazole) tris(PMDA)} ((Bulgarovskaya et al., 1989; VILFIF) and {bis(N-methyl-3,6-dibromocarbazole) tris(PMDA)} (Dzyabchenko et al., 1994; WEXKEP) and the (not isostructural) {3,6-dibromocarbazole bis(PMDA)} (Bulgarovskaya et al., 1989; VILFEB) show interesting resemblances and differences. The crystals are all triclinic and were reported in reduced cells, albeit with unconventional choices of origin. All three have 1 : 1 mixed donor acceptor stacks, with, however, different modes of overlap and differing dispositions of the additional PMDA molecules. In the isostructural pair there are sheets of donor acceptor stacks (stack axis [001]) arranged in the (002) planes, interleaved by sheets of stacks of PMDA molecules, located at crystallographic centres; these PMDA molecules are roughly coplanar with the components in the stacks. Thus VILFIF and WEXKEP could be said to have compositions {donor PMDA}0.5(PMDA). In VILFEB3 the additional PMDA molecules are located at two independent sets of symmetry centres with markedly different orientations with respect to the mixed stacks, although both have their molecular planes parallel to the stack axes. Thus VILFEB could be said to have composition {(3,6-dibromocarbazole) PMDA}[0.5(PMDA1) þ 0.5(PMDA2)], where ‘1’ and ‘2’ refer to crystallographically independent PMDAs. This arrangement has similarities to that found in (perylenium)2PF62/3(THF) shown in Fig. 17.6 and is also related to that of {(perylene)3 TCNQ}. One might guess that triclinic {(fluorene)3 (TNB)4} (Hertel and Bergk, 1936; Herbstein, Kaftory and Regev, 1976; ZZZAGP) has 1 : 1 stacks with the additional TNB molecules inserted interstitially; however, the crystal structure (Mariezcurrena et al., 1999; ZZZAGP02) shows a remarkable arrangement of mutually-perpendicular 1 : 1 and 1 : 2 stacks (Fig. 15.7). The 1 : 1 stacks, with orientationally-disordered fluorene about inversion centers, are similar to those found in {4-methylchrysene TNB} (Zacharias et al., 1991; VIGLEC) while the 1 : 2 stacks resemble those in {1-methylchrysene (TNB)2} (Zacharias et al., 1991; VIGKOL). Placing this structure description here is clearly somewhat arbitrary. The structure of the ternary compound {benzidine TNB}1/2(C6H6) (Yakushi, Tachikawa et al., 1975; BDTNNB) is somewhat different from those described above. Benzidine and TNB form mixed stacks rather similar to those found in {benzidine TNB} itself as crystallized from CHCl3 (Tachikawa et al., 1974; BNZTNB); however, there is some alteration in the mutual positioning of the stacks so as to leave channels in which the benzene molecules of solvation are accommodated at centres of symmetry with their molecular planes approximately perpendicular to those of the components in the stacks. The benzene molecules are quite strongly contained within the channels and require application of a vacuum for their removal. Other similar examples are discussed later. n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

3 VILFEB is incorrectly given a 2 : 3 composition by the CSD and incorrectly named as ‘bis(PMDA) 3,6dibromocarbazole PMDA solvate’.


1007

y O

z (a) O y

x

(b)

˚, Fig. 15.7. Triclinic {(fluorene)3 (TNB)4} (a ¼ 7.596(7), b ¼ 27.69(2), c ¼ 7.276(11) A ¼ 93.117(9), ¼ 91.114(11), ¼ 82.374(8) ; this is the Niggli reduced cell in a non-standard setting)) showing (above) the 1 : 1 stacking arranged along the [001] axis, and (below) the 1 : 2 stacking arranged along the [100] axis. Atomic displacement ellipsoids are at an arbitrary level. (Reproduced from Mariezcurrena et al., 1999.) n n n

15.5.2 Noncentrosymmetric crystals of -molecular compounds The vast majority of -molecular compounds crystallize in centrosymmetric space groups. Some of the exceptions are listed in Table 15.4, where a separation has been made between the noncentrosymmetrical crystals with enantiomorphic (Sohncke groups) and non-enantiomorphic space groups (see International Tables for X-ray Crystallography, 1965, Vol. I, pp. 41–43, for further discussion). The components in the listed crystals are achiral;4 thus the chirality for the first group, or lack of a centre for the second group, results from the details of the mixed-stack donor–acceptor arrangement. These crystals may well have interesting physical properties. The example of {1,5diaminonaphthalene chloranil} (Tamura and Ogawa, 1977; CANANP) is worthy of note because the component molecules are centrosymmetric; the noncentrosymmetric structure probably results from a compromise between the requirements of hydrogen bonding between different stacks and charge-transfer interactions within stacks. The 1,6-diaminopyrene – bromanil system (Fujinawa et al., 1999) has a number of points of interest. Firstly, the compound is polymorphic and the thermodynamically stable n n n

4 This holds, for some examples, only if the component molecule is planar (a situation not always realized in practice) or there is disorder.


Table 15.4. -Molecular compounds which crystallize in non-centrosymmetric space groups Molecular Compound/reference/refcode

Space Group

Z

Crystal Structure Reported ?

2

Yes

2

Yes

2

Yes

2

Yes

2

Yes

n n n

2

Yes

n n n

4

Yes

4 4 4

Yes Yes No

4

Yes

4

Yes

4

Yes

2

Yes

2

Yes

2

Yes

2

Yes

2

Yes

4

Yes

4

Yes

4 4

Yes Yes

I. Enantiomorphic Space Groups 1. benzo[c]pyrene (TMU)2 (Damiani, Giglio, P1 Liquori and Ripamonti, 1967); TMUBZP10. P1 2. coronene (TMU)2 (Damiani, Giglio, Liquori, Puliti and Ripamonti, 1967) 3. 7,8-benzoquinoline TCNQ P21 (Shaanan and Shmueli, 1980); BZQTCQ10. 4. 1-acetylskatole TNB (Surcouf and Delettre, P21 1978); ASKNBZ 5. TTF 2,7-dintro-9-fluorenone (Soriano-Garcia P21 et al., 1989); KARHOA. P21 6. 2-methylchrysene TNB (Zacharias et al., 1991); VIGKUR. P212121 7. 5-methylchrysene TNB (Zacharias et al., 1991); ZEGKIF10. P212121 8. carbazole TNB (Bechtel et al., 1976); CBZTNB P212121 9. perylene TNB (Hertel and Bergk, 1936); ZZZOZO. 10. anthracene BTF (Boeyens and Herbstein, 1965a); P212121 ZZZTOS. P212121 11. hydroquinone naphthaquinone (Thozet and Gaultier, 1977a); NPQHRQ. 12. acenaphthene 3,5-dimethylpicric acid P212121 (Chantooni and Britton, 1998); PUNYUS. 13. pyrene p–benzoquinone (Bernstein et al., 1976); P41 PYRBZQ. (or P43) n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

II. Non-enantiomorphic Space Groups Pc 1. pyrene TMU (Damiani et al., 1965); MURPYR. 2. N,N 0 -dimethyldihydrophenazine TCNQ Cm (Goldberg and Shmueli, 1973a); TCQMHP. Cc 3. trans-4-methylstilbene PMDA (disordered) (Kodama and Kumakura, 1974); PYMSTL. Pn 4. 1,5-diaminonaphthalene chloranil (Tamura and Ogawa, 1977); CANANP. Pn 5. 1,6-diaminopyrene bromanil (Fujinawa et al., 1999); QADGEH. Pca21 6. phenanthrene DDQ (Herbstein et al., 1978); PANCYQ. 7. benz[a]anthracene PMDA (Foster et al., 1976); Pna21 BZAPRM10. 8. chrysene TNB (Zacharias et al., 1991); VIGKIF. Pna21 9. 3-methylchrysene TNB (Zacharias et al., 1991); Pna21 VIGLAY; disordered. n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n


1009

delta-polymorph crystallizes in the noncentrosymmetric space group Pn (QADGEH). This polymorph resembles {1,5-diaminonaphthalene chloranil}. The other (metastable) polymorph (triclinic P 1, Z ¼ 4; QADGEH01) appears to have an intriguing structure but poor crystal quality prevented a definitive determination of its structure. Both polymorphs have interesting physical properties that we shall not discuss. We have deliberately excluded from consideration here molecular compounds where one or both of the components are themselves chiral; for example, crystals of some flavin compounds (Section 15.7.2) lack centres of symmetry but are not included because the flavin molecules are chiral and also because the charge-transfer interactions are probably weaker than the hydrogen bonding in these crystals. n n n

15.5.3 Acceptors based on polynitrofluorene Four acceptors of this type (2,7-dinitrofluoren-9-one, 2,4,7-trinitrofluoren-9-one, 2,4,5,7tetranitrofluoren-9-one and 2-(2,4,5,7-tetranitrofluoren-9-ylidenene)-propane-dinitrile) have been found to form donor-acceptor compounds with a variety of aromatic hydrocarbons (Table 15.5). An early use was the formation of molecular compounds with Table 15.5. Some molecular compounds of aromatic hydrocarbons with acceptors based on substituted polynitrofluorenes. These are all mixed stack structures Molecular compound

Space group

Reference/refcode

1. 1,12-dimethylbenzo[c]phenanthrene 4bromo-2,5,7-trinitrofluorenone 2. hexahelicene 4-bromo-2,5,7trinitrofluorenone 3. 2,6-dimethylnaphthalene 2,7dinitro-9-fluorenone 4. TTF 2,7-dinitro-9-fluorenone

Details not given Details not given P1

Ferguson et al., 1969

5. Hexamethylbenzene 2,4,7trinitrofluorene-9-one 6. 1-ethylnaphthalene 2,4,5,7tetranitrofluorene-9-one 7. 2-ethylnaphthalene 2,4,5,7tetranitrofluorene-9-one 8. 3,6-dimethylphenanthrene 2,4,5,7tetranitrofluorene-9-one 9. chlorobenzene 2-(2,4,5,7tetranitrofluorene-9-ylidene)propanedinitrile) 10. (chlorobenzene)2 2,4,5,7tetranitrofluorene-9-one 11. TTF 2-(2,4,5,7-tetranitrofluorene-9ylidene)-propanedinitrile) 12. benzonitrile 2,7-dicyano-9dicyanomethylene-4,5-dinitrofluorene

P21/a

n n n

n n n

n n n

n n n

n n n

n n n

P21/c

n n n

P21/c

n n n

n n n

n n n

n n n

n n n

P21

P21/n P21/c

P212121 Pna21 P21

Ferguson et al., 1969; HELFLU. Suzuki, Fuji et al., 1992; PARBIT. Soriano-Garcia et al., 1989; KARHOA. Brown et al., 1974 Baldwin and Baughman, 1993; LAVFOD. Shah and Baughman, 1994; LESZIS. Baldwin and Baughman, 1993; LAVFUJ. Batsanov, Perepichka et al., 2001; TIJTIP. Batsanov, Perepichka et al., 2001; TIJTOV. Perepichka, Kuz’mina et al., 1998 Batsanov and Perepichka, 2003.


overcrowded aromatic hydrocarbons such as hexahelicene (cf. Mackay, Robertson and Sime, 1969). More recent examples have some of the nitro groups replaced by cyano groups (Perepichka, Kuz’mina et al., 1998) or by butylsulfanyl, butylsulfinyl or butylsulfoxyl substituents (Perepichka, Popov et al., 2000). Presumably there is a synergistic effect between electron-withdrawing nitro and carbonyl (dicyanoethylene) groups, such as is also found between halogens and carbonyl groups (cf. 1,4-benzoqunone and chloranil). The two chlorobenzene molecular compounds are unusual in their mode of preparation (see original paper for details) and in the role of chlorobenzene as apparent donor, although it is usually considered a weak acceptor, All these crystal structures are of standard infinite mixed-stack type without aberrant features; overlap of the donor is over the polynitrofluorene ring without the carbonyl (or dicyanoethylene) substituent appearing to play an important role. 15.5.4

Resolution of helicenes by formation of diastereoisomeric charge transfer molecular compounds with enantiomeric acceptors

An acceptor of the type discussed in the previous section has an interesting application. Resolution of the helicenes (Martin, 1974) is an interesting example of the use of charge transfer interactions. Helicenes are nonplanar because of intramolecular overcrowding and hence exist in enantiomeric forms. The first steps towards their resolution were taken in 1956(!) when Newman and Lutz synthesized and resolved TAPA (2-[2,4,5,7-tetranitro9-fluorenylidene-aminoo¨xy]propionic acid; Fig. 15.8; see also Newman and Lednicer, 1956). R(–)TAPA forms a red complex in solution when mixed with racemic hexahelicene and M(–)hexahelicene5 can be recovered from the solution. These results imply that the P( þ )hexahelicene R(–)TAPA diastereoisomer is more stable in solution than the M(–)hexahelicene R(–)TAPA diastereoisomer. These principles have been adapted to n n n

n n n

CH2OH H O N

C

CO2H

H

OH OH

H NO2

O 2N

OH

H

X

O2N

H

H

H 3C

N

H3C

N

NO2 (a)

O

N N

H O

(b)

Fig. 15.8. Enantiomeric electron acceptors used in the resolution of helicenes by formation of diastereoisomeric charge transfer molecular compounds (a) TAPA and related compounds : X ¼ methyl, R(–)TAPA; X ¼ ethyl, R(–)TABA; X ¼ isopropyl, R(–)TAIVA; X ¼ butyl, R(–)TAHA. 5 The absolute configurations (e.g. that of TAPA (Kemmer et al., 1976)) were determined after Newman’s pioneering studies and have been added here for completeness.

INTERACTIONS BE TWEEN POLARIZABLE AND POLAR GROUPS

1011

[10] [8]

[12]

start

uv 280 nm

[6]

[14]

10

20

30

40

50

60

70

80

90 (min)

Fig. 15.9. Resolution of a mixture of the racemates of the [6]-, [8]-, [10]-, [12]-and [14]-helicenes. The microsilica column contained 25% R(–)TAPA covalently linked to the silica; the mobile phase was 25% dichloromethane–cyclohexane; U ¼ 0.26 cm/sec. In all instances the more strongly retained enantiomer was the P( þ )helicene. Similar results were obtained for the [7]-, [9]-, [11]-and [13]helicenes. [5]-Helicene was resolved only by a multipass technique. (Reproduced from Mikes et al., 1976.)

high performance liquid chromatography with great success (Mikes et al., 1976; Numan et al., 1976), (Fig. 15.9). The HPLC results confirm the deduction from the solution experiments that the P(þ)hexahelicene R(–)TAPA diastereoisomer is the more stable. Later work has shown that riboflavin, adenosine and adenylic acid coated on microsilica particles can also be used as the stationary phases. The M(–)helicenes are more strongly retained on riboflavin. Measurements have been made (in tetrachloroethane solution at 248K, using 1H NMR) of the stabilities of the (P)-[7]-thiaheterohelicene S(þ)-TAPA and (P)-[7]-thiaheterohelicene R(–)TAPA diastereoisomeric molecular compounds (Nakagawa et al., 1982); the P S diastereoisomer was found to be the more stable by DDH ¼ 1.03 kJ/mol and DDS ¼ 1.40 J/mol K. A1H NMR survey of the structures of the two diastereoisomers was said to indicate that the components packed better in the P S than in the P R diastereoisomer. Analogous crystal-structure comparisons do not appear to have been made. n n n

n n n

n n n

n n n

n n n

n n n

15.6 Structurally important interactions between polarizable and polar groups The juxtaposition of polar groups (such as >C¼O or R > C¼C < R 0 ) in acceptors and polarizable groups (such as benzene rings) in donors leads to dipole-induced dipole interaction between them. This interaction, first emphasized in the present context by Prout and Wallwork (1966), can be an important or even dominating influence in determining mutual donor-acceptor arrangement in unary and binary crystals (see also Gaultier et al., (1969)). {Perylene fluoranil} (Hanson, 1963; PERFAN) has a typical overlap n n n


diagram (Fig. 15.10(a)) which is also found in the molecular compounds of perylene with 2,5-dibromo-3,6-dichloro-p-benzoquinone (Kozawa and Uchida, 1979; PERPBQ) and TCNQ (Tickle and Prout, 1973a), and in {(perylene)3 TCNQ} (Hanson, 1978) (Fig. 15.10(c)); and also in the unary crystals of 1,4-naphthoquinone (Gaultier and Hauw, 1965; NAPHQU) and 1,4-anthraquinone (Dzyabchenko and Zavodnik, 1984; COBBIE04) (Fig. 15.10 (d) and (e)). Aromatic ring carbonyl interactions are also important in forming the DAD triples in {(TMBD)2 chloranil} (Yakushi, Ikemoto and Kuroda, 1971) (Fig. 15.10(f)) and in the mixed stacks of chrysene fluoranil (Munnoch and Wright, 1975; CHRFAN) and {1,5-diaminonaphthalene chloranil} (Tamura and Ogawa, 1977; CANANP). Direct overlap of donor and acceptor is also found in solvent-free and in solvated benzidine TCNQ (see Section 15.7.3.2). One should note the close dimensional correspondence between donor and acceptor molecules in all these molecular compounds; the distance between the centres of the benzene rings in a diphenyl-like portion of perylene or benzidine is closely equal to the O---O distance in the halo-anils or the distance between the centres of the extra-ring double bonds in TCNQ. n n n

n n n

n n n

n n n

n n n

n n n

(a)

(b)

(c) C1

F

(d)

(e)

O

O

O

O

O

O

O

O

(f )

C N O Cl

Fig. 15.10. Patterns of overlap which result from interactions between polarizable and polar groups (a) perylene fluoranil; (b) pyrene chloranil; (c) perylene TCNQ in (perylene)3 TCNQ; a somewhat similar overlap diagram is found in perylene TCNQ; (d) 1,4-naphthaquinone; (e) 1,4anthraquinone; (f) (TMBD)2 chloranil. n n n

n n n

n n n

n n n

n n n

n n n

CHARGE TRANSFER AND HYDROGEN BONDING INTERACTIONS

1013

Aromatic ring polar group overlap is less pronounced but still obvious in the molecular compounds of pyrene with chloranil (Prout and Tickle, 1973c; PYRCLN) (Fig. 15.10(b)), fluoranil (Bernstein and Regev, 1980; PYRFLR), p-benzoquinone (Bernstein et al., 1976) and 2,6-dichloro-N-tosyl-1,4-benzoquinonemonoimine (Shvets et al., 1980; PYTQIM), and in {acenaphthene chloranil} (Tickle and Prout, 1973a) and {acenaphthene TCNQ} (Tickle and Prout, 1973b; ACNTCQ). We end this section by stressing the obvious – there are many molecular compounds in which interaction between polarizable and polar groups could occur but does not. This is the situation for such molecular compounds as {HMB chloranil} (Jones and Marsh, 1962; CLAHMB), {TMBD chloranil} (Yakushi et al., 1973; MBZDCN), {anthracene 3,3 0 ,5,5 0 -tetrachlorodiphenoquinone} (Starikova et al., 1980; ANPHXN) and {pyrene 3,3 0 ,5,5 0 -tetrachlorodiphenoquinone} (Shchlegova et al., 1981; BAZDAH). Although there is no direct aromatic ring carbonyl overlap in the P21/n polymorph of 2,3-dichloronaphthazarin (Rubio et al. 1985; DCDHNQ01), nevertheless the molecules are arranged in stacks containing symmetry centres that are situated such that the quinoid part of one molecule partially overlaps a benzenoid part of a neighbouring symmetry-related molecule, the interplanar distance ˚ . Consequently this polymorph forms a self-complex in the solid state; being 3.40(3) A spectroscopic data indicate that such close associations also exist in solution. Finally, we note that in the few crystal structures that have been reported for molecular compounds with polynitrofluorenones (TNF and TENF) as acceptors, the acceptor behaves as a polynitroaromatic and aromatic ring carbonyl interactions do not seem to be important. n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

15.7 Mixed-stack crystals with both charge transfer and hydrogen bonding interactions 15.7.1 The quinhydrones as a crystallochemical family 15.7.1.1

Crystal structures

Quinhydrones (Patil et al., 1986) are molecular compounds formed by polyhydroxyaromatics (as donors) and aromatic quinones (as acceptors); quinhydrone itself is {hydroquinone p-benzoquinone} (hydroquinone is also called quinol or 1,4-dihydroxybenzene). We define members of the quinhydrone structural family as having crystal structures involving two specific interactions between the components n n n

(i) hydrogen bonding between hydroxyl and carbonyl groups (ii) charge transfer interaction between donor and acceptor ring systems. The crystal structures appear to be determined by an interplay between these factors, whereas the UV–visible spectroscopic properties are a consequence of the charge transfer interactions, which often show themselves structurally by a superposition of the carbonyl group of the acceptor over the (possibly substituted) aromatic ring of the donor. A consequence of this definition is that every molecular compound composed of a polyhydroxyaromatic and an aromatic quinone is not necessarily a quinhydrone. Donor : acceptor ratios of 2 : 1, 1 : 1 and 1 : 2 (and, exceptionally, 1 : 3) are known and can be seen to depend on the nature of the components. Much of the systematic preparative work dates back 60–80 years; we quote some of these results (Siegmund, 1908;


Table 15.6. Quinhydrones reported to be formed between polyhydroxy-aromatics and aromatic quinones Donor

D/A Ratio

(a) With p-benzoquinone as acceptor: phenol 2:1 p-chlorophenol 2 : 1, 1 : 1 p-bromophenol 2 : 1, 1 : 1 p-nitrophenol 1:1 1,2-dihydroxybenzene 2 : 1, 1 : 1 1,3-dihydroxybenzene 1:1

Donor

D/A Ratio

1,4-dihydroxybenzene(quinol) 1,2,3-trihydroxybenzene 1-naphthol 2-naphthol p-methylaminobenzene 2-naphthylamine

1:1 1:3 2 : 1, 1 : 1 2 : 1, 1 : 1 1:1 2 : 1, 1 : 1

(b) With other acceptors: 1-naphthol -naphthoquinone; 1-naphthol phenanthrenequinone n n n

n n n

Meyer, 1909; Kremann et al., 1922) in Table 15.6, which does not give a complete survey of the known compounds. This type of information, taken together with that about crystal structures in Table 15.7 (which is hopefully more comprehensive), shows both what has been learned and what gaps remain. The methods used include crystallization from solutions of the mixed components and determination of binary phase diagrams. Two aromatic amines are included among the donors because their molecular compounds may resemble the quinhydrones in structure. Similarity of intercomponent interactions leads to striking resemblances among many of the crystal structures and justifies treatment of the whole group as a single crystallochemical family. The known crystal structures (Bernstein, Cohen and Leiserowitz (1974) and later work) can be grouped as shown in Table 15.7. In this classification we have emphasized the importance of donor-acceptor stacking and distinguished between 1 : 1 and 1 : 2 (or 2 : 1) compositions of the stacks. This means that quinhydrones with overall compositions 1 : 2 (or 2 : 1) are classified according to the D : A ratio in the stack and not in terms of overall composition. The structural features of Group A are conveniently introduced by reference to the two polymorphs of quinhydrone ( triclinic, P 1, Z ¼ 1; , monoclinic, P21/c, Z ¼ 2). In these crystals a combination of charge-transfer interactions along [100] and hydrogen bonding along [120] gives molecular sheets, one molecule thick, parallel to (001). In the triclinic polymorph (Fig. 15.11(a)) all the molecular sheets are identical, whereas in the monoclinic polymorph (Fig. 15.11(b)) the direction of the hydrogen bonds in successive sheets changes from [120] to [1 20] in accord with the introduction of the c glide plane. The arrangements in the two polymorphs are formally similar to those found in the triclinic and monoclinic polymorphs of p-dichlorobenzene (Herbstein, 2001). The volume per formula unit in both quinhydrone polymorphs is less than the sum of the volumes (in their respective crystals) of the two components; the reduction is 9% for the triclinic polymorph and 7% for the monoclinic polymorph, suggesting that the former is the more stable at room temperature and pressure. Analogous reductions are found for almost all the quinhydrones but none is as large as that found for -quinhydrone; the only exception we have found is {(2,5-dimethylhydroquinone)2 (2,5-dimethyl-p-benzoquinone)} (CIKRAP), where there is a volume increase of 1.4%. The orientations of the stacks alternate in {resorcinol p-benzoquinone} (BZQRES) (space group Pnca, Z ¼ 4) (Fig. 15.11(c)) and the structure thus resembles that of n n n

n n n


1015

Table 15.7. Classification of the quinhydrones according to relationships among their crystal structures Type I: with DA stacks Group A: (donor : acceptor ratio 1 : 1). -quinhydrone (Sakurai, 1968; QUIDON01) and -quinhydrone (Sakurai, 1965; QUIDON); resorcinol p-benzoquinone (Ito et al., 1970; BZQRES); hydroquinone 1,4-naphthaquinone (Thozet and Gaultier, 1977a; NPQHRQ); 2-phenyl- (or 2-chlorophenyl)-benzhydroquinone 2phenyl- (or 2-chlorophenyl)-benzoquinone (Desiraju et al., 1979; PBZQHQ); durohydroquinone duroquinone [duroquinhydrone] (Patil et al., 1986; Pennington et al., 1986; FADCES), Group B: (donor : acceptor ratio 1 : 1, donor is a phenol). p-chloro- (or p-bromo-)phenol p-benzoquinone (isomorphous) (Shipley and Wallwork, 1967b; BNQCLP, BNQBRP). Group C: (donor : acceptor ratio 2 : 1; additional donor bridges between DA stacks) (hydroquinone)2 2,5-dimethylbenzoquinone) (Patil et al., 1984; CISCOW); (hydroquinone)2 duroquinone) (Pennington et al., 1986; FADCIW). n n n

n n n

n n n

n n n

n n n

n n n

n n n

Type II: (with --DAD-- or --ADA-- as repeat units in the stacks) Group D1: (donor : acceptor ratio 2 : 1). {(phenol)2 p-benzoquinone} [phenoquinone] (Sakurai, 1968; PHENQU); ({p-chloro(or p-bromo-) phenol}2 p-benzoquinone} (these two molecular compounds are isomorphous) (Shipley and Wallwork, 1967b; BNQDCP, BNQDPB). Group D2: (donor : acceptor ratio 1 : 2). {(1,3,5-trihydroxybenzene) (p-benzoquinone)2} (Sakurai and Tagawa, 1971; PHLBZQ); {1,4-dihydroxynaphthalene (1,4-naphthoquinone)2} (Artiga et al., 1978b; NPQNHQ); {-naphthol (2,3-dichloro-1,4-naphthoquinone)2} (Thozet and Gaultier, 1977b; CNQNPO); {(2,5-dimethyl-1,4-dihydroxybenzene) (2,5-dimethyl-1,4-benzenoquinone)2} (Patil et al., 1984). n n n

n n n

n n n

n n n

n n n

n n n

Notes: Cell dimensions have been reported for {(p-cresol)2 p-benzoquinone} and {-naphthoquinonequinhydrone} (ZZZOXY) (Anderson, 1937); the structure of {-naphthol 2-methyl-1,4-naphthoquinone} has been briefly described (Berthelon et al., 1979). n n n

n n n

monoclinic -quinhydrone; polymorphism has not been reported. There is also a remarkable resemblance (Herbstein, 1971, see p. 317) between the structures of {resorcinol p-benzoquinone} and {anthracene TNB} (Brown, Wallwork and Wilson, 1964; ANCTNB). There is superpositioning of carbonyl groups over aromatic rings in - and -quinhydrones (Fig. 15.11(d)) and this is also found in {p-ClC6H4OH C6H4O2}, phenoquinone, {resorcinol C6H4O2} and {(p-ClC6H4OH)2 C6H4O2} (C6H4O2 is p-benzoquinone). Two structures in this group have noteworthy features. {Hydroquinone 1,4naphthoquinone} has a structure (Thozet and Gaultier, 1977a; NPQHRQ) based on hydrogen bonding and charge-transfer interactions, with aromatic ring-carbonyl superpositioning. Individual sheets are analogous to those in -quinhydrone but four crisscrossed sheets are packed in each b-axis period. The crystals are chiral (space group P212121) although the components are not. The structures of 2-phenylquinhydrone and 2-(p-chloro)-phenylquinhydrone are based on the usual principles of hydrogen bonding and charge transfer interaction but are such that quinol and quinone components cannot n n n

n n n

n n n

n n n

n n n

n n n


(a)

(b)

a

b

o [101up]

a9 o

b 1Å (c)

o

O

b C up

2Å

4

2.7

OH

O

a/2

OH

Fig. 15.11. (a) Triclinic () quinhydrone in projection down [001]; (b) monoclinic () quinhydrone in projection down [101]; (c) {resorcinol p-benzoquinone} in projection down [001]; the resorcinol molecules are seen edge-on and the p-benzoquinone molecules are seen slightly tilted. The overlap diagram typical of many quinhydrones showing superposition of carbonyl group on aromatic ring is illustrated in the lower left portion of the diagram. Hydrogen bond lengths (d(O . . . O)) are ˚ in -quinhydrone and 2.71 A ˚ in the -polymorph. (Reproduced from Herbstein, 1971.) 2.739 A n n n

be distinguished in the crystal because of disorder (space group P21/c, Z ¼ 2); it was suggested that the crystals could be conglomerates of regions with true space group symmetry either P21 or Pc (Desiraju et al., 1979). The same structural principles apply in Group B as in Group A, except that the molecules are hydrogen bonded in pairs (Fig. 15.12) rather than ribbons because of the monofunctionality of the phenols. The structures of {(hydroquinone)2 (2,5-dimethyl-p-benzoquinone)} (CISCOW) and {(hydroquinone)2 (duroquinone)} (FADCIW) form a separate group in which the equimolar DA stacks are preserved with the second hydroquinone molecule bridging between the stacks; the details of the hydrogen bonding by which the bridging is effected differ in the two crystals. We first illustrate for {(hydroquinone)2 2,5-dimethyl-p-benzoquinone)} (Fig. 15.13); there is carbonyl-aromatic ring overlap within the stacks and the stacks are linked together by hydrogen bonding between hydroquinone and benzoquinone to give sheets analogous to those already noted in - and -quinhydrones. The interstitial hydroquinones hydrogen bond between hydroquinones of the stacks to give cohesion in the third dimension. On the other hand, the stacks in {(hydroquinone)2 (duroquinone)} are not hydrogen bonded into sheets but the cohesion is given by the bridging hydroquinones n n n

n n n

n n n

n n n


(a)

1017

c sin a

b up

C O Cl

a sin g c sin a (b) a down

b sin g

Fig. 15.12. Projections of the crystal structure of {p-chlorophenol p-benzoquinone} down (a) [010], (b) [100]. In (b) two hydrogen-bonded donor–acceptor pairs are shown, related by a centre of symmetry; one of these pairs has been omitted in (a) for clarity. (Reproduced from Herbstein, 1971.) n n n

which link four stacks by HQ--HQ and HQ--DQ hydrogen bonds (Fig. 15.14); there is no carbonyl-aromatic ring overlap within the stacks. Thus the first structure is essentially a true quinhydrone in terms of its resemblance to - and -quinhydrone while the second is noted here for reasons of chemical rather than structural resemblance. The 2 : 1 and 1 : 2 quinhydrones of Group D have somewhat similar structures and we describe that of phenoquinone ({(phenol)2 p-benzoquinone}; Fig. 15.15). The components are hydrogen bonded in groups of three, the minor component being flanked on each side by a molecule of the major component. The stacks have composition n n n

----DAD DAD DAD---˚ ; the D---A interplanar and thus the characteristic periodicity along the stack axis is 11 A distances are somewhat shorter than the D---D distances. Aromatic ring-carbonyl overlap occurs in phenoquinone, there is partial overlap in {(2,5-dimethylhydroquinone)2 (2,5dimethyl-p-benzoquinone)} (CIKRAP) but none in {1,4-dihydroxynaphthalene (1,4naphthoquinone)2} (NPQNHQ) nor in {-naphthol (2,3-dichloro-1,4-naphthoquinone)2} (CIQNPO). In the latter the -naphthol molecule is disordered across a centre of sym˚ ) are also important. metry and Cl---O interactions (d ¼ 2.95 A The {1,3,5-trihydroxybenzene (p-benzoquinone)2} compound (PHLBZQ) has an interesting structure in which there are independent branched zigzag chains of molecules n n n

n n n

n n n

n n n


Fig. 15.13. Stereoview of the {(hydroquinone)2 (2,5-dimethyl-p-benzoquinone)} structure, seen approximately down [001]; hydrogen bonds are shown as broken lines; all three component molecules are at independent centres of symmetry, with some symmetry-related molecules omitted to simplify the diagram. Ribbons of hydrogen-bonded hydroquinone 2,5-dimethyl-p-benzoquinone extend along the [101] direction as shown in the lower ac plane of the cell. The aromatic ringcarbonyl overlap occurs between superimposed ribbons of this type. The ribbons are linked by the ‘‘interstitial’’ hydroquinone molecules shown in the centre of the (100) faces. (Reproduced from Patil, Curtin and Paul, 1984.) n n n

n n n

Fig. 15.14. Stereodiagram of the crystal structure of {(hydroquinone)2 (duroquinone)} showing linking of four DA stacks by hydrogen bonding to the interstitial hydroquinone in the center of the cell. (Reproduced from Pennington et al., 1986.) n n n

in each molecular layer (Fig. 15.16). Two of the acceptor molecules bridge between different hydroxyl groups of the donor while the third p-benzoquinone is linked through one of its carbonyl groups to the third hydroxyl, while the second is unlinked; this arrangement accounts for the composition. The chains interact only by dispersion forces


(a)

1019

(b) O

3.16

b

O

3.32

3.28

3.42

3.25

c

3.32 3.42

a

a (c)

Fig. 15.15. Phenoquinone (a) projection down [001], (b) projection down [010], (c) overlap diagram showing superimposed DAD hydrogen-bonded triples; note the aromatic ring-carbonyl overlap. (Reproduced from Sakurai, 1968.)

0

c⬘

c

b⬘

Fig. 15.16. {1,3,5-trihydroxybenzene (p-benzoquinone)2} projected onto the mean molecular plane. The zigzag hydrogen-bonded ribbons, one of which has been shaded, extend along the axis labelled c. There is no hydrogen bonding between neighbouring ribbons in the b 0 direction. (Reproduced from Sakurai and Tagawa, 1971.) n n n

so that the crystal structure is a two-dimensional analog of the intersecting but non-bonded three-dimensional networks found in the quinol clathrates (Chapter 8) and some other crystals such as -trimesic acid (Section 9.3) and adamantane-1,3,5,7-tetracarboxylic acid (Ermer, 1988; GEJVEW).


The equimolar molecular compound of 6-hydroxydopamineHCl and its oxidized p-quinonoid form is a quinhydrone (Andersen et al., 1975; DOPAQC). The compound ˚, crystallizes in a monoclinic unit cell with a ¼ 6.815, b ¼ 13.108, c ¼ 20.365 A ¼ 95.43 , space group P21/n, Z ¼ 4. The component cations stack alternately along [100], with an angle of 3.8 between ring planes and interplanar spacings of 3.42 and ˚ , somewhat longer than in most quinhydrones. The two components are respect3.49 A ively colourless and yellow in their neat crystals and the red colour of the molecular compound is credibly ascribed to a charge transfer interaction. There is no carbonylaromatic ring overlap. OH

OH O

HO

O

OH CH2

CH2

CH2

CH2

NH3+

NH3+

The asymmetric unit consists of the two cations shown and two chlorides as counterions.

Crystal structures have been reported for a number of cyclophane quinhydrones (see Chapter 14 for details and references); for example, a stacked structure was found for pseudogeminal 14,17-dimethoxy[3](2,5)-p-benzoquinone, with an

AD AD AD AD ˚ , about arrangement along the stacks and an average intermolecular separation of 3.42 A ˚ larger than the intramolecular transannular separation (see Fig. 14.5). Hydrogen 0.23 A bonding between moieties in different stacks seems unlikely for steric reasons. 15.7.1.2

Thermodynamic studies of quinhydrones

A number of measurements have been made of thermodynamic parameters for various quinhydrones. The enthalpy of formation of a crystalline molecular compound of composition DmAn is defined as DH f ðDm An Þ ¼ HðDm An Þ fmHðDÞ þ nHðAÞg; where the enthalpies H(DmAn) etc. refer to the crystalline materials at the same (specified) temperature. Analogous equations and conditions apply to free energies DGf (DmAn) and entropies of formation DSf (DmAn). Available results for DHf (DmAn) are summarized in Table 15.8. We first consider the results for quinhydrone. Three polymorphs of hydroquinone and two of quinhydrone are now known and thus specification of polymorph, with regard to both quinhydrone and hydroquinone, is required; Suzuki and Seki (1953) checked that they used -hydroquinone but do not report which quinhydrone polymorph was used, while the converse situation applies to the results of Artiga et al. (1978a, b). The values of the crystal densities suggest that triclinic quinhydrone is stable with respect to monoclinic at room temperature but the quantitative situation is not known, nor is it known whether


1021

Table 15.8. Values of Hf(DmAn) (kJ/mol) for some quinhydrones as measured by a variety of methods Molecular Compound

–Hf(DmAn)

Method

Quinhydrone

20.2 (303K); 22.6 (297K) 19.7; 20.6(1) (295K) 33.9 23.2 (297K)

Dissolution (Artiga et al., 1978a; Suzuki and Seki, 1953) vapour pressure (Nitta et al., 1951; Kruif et al., 1981). Combustion (Schreiner, 1925) calculated from G, S (Schreiner, 1925) Estimated (Suzuki and Seki, 1953) Dissolution (Artiga et al., 1978a, b); vapour pressure (Kruif et al., 1981) Dissolution (Artiga et al., 1978b); vapour pressure (Kruif et al., 1981) vapour pressure (Kruif et al., 1981)

{hydroquinone 1,4naphthoquinone} {1,4-dihydroxynaphthalene (1,4-naphthoquinone)2} {1,4-dihydroxynaphthalene 1,4-naphthoquinone} n n n

n n n

n n n

22.0 (0K) 6.1 (303K) 8.0(2) (320K) 8.6 (303K) 9.6(4) (320K) 11.6(2) (330K)

the two polymorphs are enantiotropically or monotropically related. Artiga et al., report that triclinic quinhydrone is thermally stable up to its melting point at 443K, suggesting that the relationship for quinhydrone is monotropic. The precision reported for the various measurements is 1–3%, but intercomparison of the DHf values suggests that systematic errors can occur, apart from the uncertainty about which polymorph was used. The value of DHf for quinhydrone at 297K is about –21(2) kJ/mol, presumably applying to the more stable -polymorphs of hydroquinone and quinhydrone. The value of DHf at 0K was estimated Suzuki and Seki (1953) by correcting the 297K value via specific heat values measured by Lange (1924) and ‘‘put in order’’ by Schreiner (1925); as the specific heats were measured at rather large temperature intervals, these corrections are somewhat uncertain. The free energy and entropy of formation of quinhydrone at 297K were measured by an EMF method (Schreiner, 1925) and the following values obtained : DGf ¼ 15:3kJ=mol

DSf ¼ 26:3J=mol K:

The corresponding values from vapour pressures (Kruif et al., 1981) [–13.8(1) kJ/mol and –23(5) J/mol K] are in good agreement. One may conclude that the components are more strongly bound in crystalline quinhydrone than in the separate crystals of hydroquinone and p-benzoquinone, and that there is a concomitant reduction in the entropy, presumably because of a reduction in the amplitudes of thermal vibration. The negative value of DSf implies that quinhydrone becomes more stable with respect to its separated components as the temperature is lowered, it being assumed that DHf and DSf are only weakly temperature dependent; this is also the situation for anthracene picric acid (see Chapter 16). A value of DGf (330K) of –8.6(1) kJ/mol. has been calculated for {1,4-dihydroxynaphthalene 1,4-naphthoquinone} (crystal structure does not seem to have been reported) from vapour pressure measurements (Kruif et al., 1981) with DSf (330K) calculated n n n

n n n


as –9(5) J/mol K. The situation is similar to that found for quinhydrone but the molecular compound is less stable with respect to its components than is quinhydrone. Specific heats have been measured for {1,4-dihydroxynaphthalene (1,4-naphthoquinone)2} and its components from 4–300K and DSf (297K) calculated as 7.7 J/mol K, leading to a value of DGf (297K) of 10.9 kJ/mol (Artiga et al., 1978b). Vapor pressure measurements (Kruif et al., 1981) give DGf (320K) ¼ 12.9(10) kJ/mol., DHf (320K) ¼ 9.6(4) kJ/mol, and DSf(320K) ¼ 10(10) J/mol K. The positive value of DSf implies that {1,4-dihydroxynaphthalene (1,4-naphthoquinone)2} becomes less stable with respect to its separated components as the temperature is lowered, contrary to the situation in quinhydrone. The energies of complexation have been calculated for some of these molecular compounds using atom-atom potentials (Wit et al., 1980); the results were qualitatively correct but not very accurate. This is not surprising as atom-atom potentials hardly seem fitted to the quantitative representation of hydrogen bonding and charge transfer interactions. n n n

n n n

15.7.1.3 Spectroscopy Polarized absorption (polymorph not specified) (Nakamoto, 1952) and reflection spectra (monoclinic polymorph) (Anex and Parkhurst, 1963) have been measured from small single crystals of quinhydrone. Nakamoto (1952) interpreted his results in terms of enhanced absorption, due to charge transfer, when the radiation was polarized normal to the ring planes of the two component molecules but Anex and Parkhurst (1963) claimed that the appropriate polarization direction is along the stack axis and not along the ring normals, a result verified by both techniques used by them. They point out that Mulliken’s theory in fact suggests that the polarization vector for the charge transfer band should lie along the line joining the centres of the components. However, the occurrence of crisscrossed stacks in monoclinic quinhydrone (Fig. 15.10(b)) forces the polarization to lie along the average normal to the molecular planes, which is the stack axis. Measurements on the triclinic polymorph, where all plane normals are parallel, and inclined to the stack axis, would provide the decisive test. A measurement of this kind made on triclinic {anthracene PMDA} (ANTPML) shows unequivocally that the moment of the principal charge transfer transition lies along the anthracene to PMDA center-to-center vector or [001] axis and not along the plane normal (Merski and Eckhardt, 1981) (cf. Section 17.3). Polarized single-crystal spectra of {resorcinol p-benzoquinone} have been studied by Amano (quoted by Ito et al., 1970). Intense charge transfer bands were observed with light polarized along [100] while the crystal was almost transparent to light polarized along [010]. The orthorhombic crystal structure requires the charge-transfer moment to be along the stack axis as in monoclinic quinhydrone. n n n

n n n

15.7.2

Molecular compounds of the flavins

The flavo co-enzymes, which are the non-protein parts of flavoenzymes, are widely involved in oxidation-reduction processes in the cell. The two forms most commonly found in biological systems are flavin mononucleotide (FMN or riboflavin 5 0 -phosphate) and flavin adenine dinucleotide (FAD), whose molecular structures are given in Fig. 15.17.


1023

FAD FMN H

Riboflavin (Vitamin B2) H Lumiflavin H CH2 10

H3C H3C

6

N N 5

(CHOH)3 1 N

CH2O

O PO O–

O PO O–

CH2 N N

O O

NH O

HO

N

NH2 N

OH

Fig. 15.17. The molecular structures of several flavins. In iso-alloxazine the substituents at C(7), C(8) and N(10) are all hydrogens.

The iso-alloxine or flavin nucleus, which is common to all forms, is the site of electron exchange and is shown in the air-stable, fully oxidized or quinoid form. It is reduced enzymatically by one or two electrons to either the semiquinoid or the hydroquinoid state. Since many of the substrates, including reduced pyridine nucleotide, DPNH, are aromatic or quasi-aromatic, it is possible that the electron interchange reaction occurs by formation of an intermediate, transient, charge transfer complex (Szent-Gyorgi, 1960). In consequence considerable effort has been applied to the preparation of complexes between electron donors and flavins in various oxidation states. Cationic, neutral and anionic species can be obtained depending on the pH of the system (Kosower, 1960; Tollin, 1968). Neutral and charged flavoquinone complexes have been obtained, as well as charged flavosemiquinone complexes. A fairly complicated series of equilibria has been shown to exist between the fully oxidized and fully reduced forms of flavins (Kuhn and Stroeble, 1937), each intermediate being characterized by ‘‘excellent crystallizing power, vivid colours and sharply defined composition.’’ The best results were obtained with riboflavin where the equilibria can be represented as : flavin ! verdoflavin ! chloroflavin ! rhodoflavin ! leucoflavin. All the crystalline intermediates were paramagnetic (bulk susceptibility measurements), ionic (Naþ or Cl ions were necessary), and supposed to be composed of different mixtures of the various flavin oxidation states. Most studies to date of complexes of flavins are based on visible and ESR spectra of solutions but some crystalline compounds have been prepared. For convenience, neutral and charged species will be described together. The neutral crystalline compounds are usually orange to orange-red, while the protonated complexes range from green through


deep red to black. Charge transfer interaction between donor and flavin acceptor has been suggested for certain compounds, although this may not always have been appropriate. Neutral crystalline molecular compounds of 2,3- and 2,7-naphthalenediol with riboflavin have been prepared (Fleischmann and Tollin, 1965a) with phenol : flavin ratios of 2 : 1, although spectroscopic analysis indicated that the solution compositions were 1 : 1. There was an extra band in the spectra of mineral oil suspensions of powdered crystals, and this has been ascribed to charge transfer. Various indoles, such as 5-methylindole, carbazole and tryptophan, gave neutral 1 : 1 crystalline compounds with lumiflavin but not with riboflavin. Protonated complexes such as lumiflavin-tryptophan-HCl-H2O have also been obtained. Although the crystalline complexes are more highly colored than the parent flavins, the diffuseness of the additional bands in the spectra makes it difficult to decide whether there is indeed a charge transfer interaction (Pereira and Tollin, 1967). Crystalline complexes have also been prepared with compositions phenol : flavin : HCl, where the phenols include hydroquinone, 4-chloropyrocatechol, resorcinol and 1,2- and 1,4-dihydroxynaphthalene and the flavins riboflavin, lumiflavin and 9-methylisoalloxazine (Fleischmann and Tollin, 1965b). The spectra suggest charge-transfer interactions between the components. Mention must also be made of the paramagnetic semiquinone salts and complexes first prepared by Kuhn and Strobele (1937) and studied later by Fleischmann and Tollin (1965c), who showed that iso-alloxazine derivatives in concentrated hydroiodic acid were reduced to the semiquinone state and that some crystalline flavin hydroiodides could be obtained; ESR measurements showed that the latter had 100% unpaired spins. Riboflavin hydroiodide (no analysis given) gave pink crystalline platelets, while lumiflavin hydroiodide (composition FH22HI) was a dark-brown solid which could not be recrystallized. Black crystals (red in thin section) were obtained from saturated solutions of riboflavin in hydroiodic acid to which various phenols (e.g. hydroquinone, 2-naphthol, 2,3-, 1,7-, 2,7-, 1,5- and 1,4-dihydroxynaphthalene) had been added; a typical composition was FH3þI2,3-dihydroxynaphthalene. There were no indications of charge transfer interactions in the spectra (Tollin, 1968). A number of crystal structures have been determined (Table 15.9) and comparison shows many common features. 1. 2.

3.

There is always an extensive array of hydrogen bonds which is probably the primary factor in determining the overall arrangement. Most structures contain infinite stacks of donor and acceptor molecules but two have finite DAAD sequences (Fig. 15.18). The interplanar distances range from 3.3 to ˚ and there is some correlation between colour of the crystals (a rough measure of 3.5 A charge transfer, in the absence of spectroscopic results) and interplanar distance. However, the nature of the moieties is probably a more important factor here and it seems clear that the protonated flavins are appreciably stronger acceptors than the neutral flavins (in contrast one may note that picric acid molecule and picrate ion have similar acceptor strengths). In four of the compounds (out of eight) there are additional donor molecules in interstitial positions outside the mixed stacks, with the planes of the interstitial molecules approximately normal to those in the mixed stacks. The arrangement is similar to that encountered in a few mixed stack donor–acceptor molecular compounds

Table 15.9. Some structural details for crystalline flavin molecular compounds. Reports of crystal structure determinations after 1975 have not been found. Complex/reference/refcode

Colour of Crystals

Interplanar ˚) Spacings (A

Space Group

Z

Stacking Type

Interstitial Donors?

1. Adenine riboflavin. 3H2O (Fujii et al., 1977); ADRBFT10 2. (Naphthalene-2,3-diol) lumiflavin (Wells et al., 1974); LUMNPO20 3. 5 0 -Bromo-5 0 -deoxy-adenosine riboflavin3H2O (Voet and Rich, 1971); RIBBAD10 4. (Naphthalene-2,3-diol)2 (10-propyl-isoalloxazine) (Kuo et al., 1974); PALXND10 5. (Naphthalene-2,3-diol)2 lumiflavin3H2O (Fritchie and Johnson, 1975); LUNMPO10 6. (2,6-Diamino-9-ethylpurine)2 (lumiflavin)2 C2H5OHH2O (Scarbrough et al., 1976, 1977); LUFAEP 7. Quinol lumiflavin chloride (Karlsson, 1972); LUMIHQ 8. Quinol riboflavin(HBr)2(H2O)2 (Bear et al., 1973); BIBHQN10

dark yellow

3.4–3.5

P21

4

No

yellow

3.46; 3.48

P21/c

4

infinite mixed stacks ditto

Yes

orange-brown

3.45; 3.54

P212121

4

ditto

No

orange

3.38; 3.46

A2/a

8

ditto

Yes

reddish-orange

3.3

P1

2

ditto

Yes

deep red

3.4

P21/c

4

DAAD sequence (cf. Fig. 15.17)

No

n n n

green

P21/c

8

black

P21

4

blackish

P21/c

4

DAAD sequence (see Fig. 15.17) One stack of infinite mixed type; the other has pairs of donor and acceptor moieties. Infinite mixed stacks.

No

n n n

quinol/flavin 3.49; flavin/flavin 3.72 3.4 in infinite mixed stack; 3.26 and 3.52 in ‘‘paired’’ stack. 3.3

black

Not listed

P1

1

n n n

n n n

n n n

n n n

n n n

n n n

9. (Quinol)3 lumiflavinium bromide (Tillberg and Norrestam, 1972); LUMFHQ 10. (Naphthalene-2,7-diol)3 (10-methylisoalloxazinium bromide)22H2O (Langhoff and Fritchie, 1970); MAZNDO10 n n n

n n n

Charge transfer interaction suggested.

No

Yes Yes


x

y

Fig. 15.18. Schematic diagram showing one sheet of the molecular packing in quinol lumiflavin chloride extended over two unit cells. The long and short full lines represent the lumiflavin and hydroquinone molecules respectively; the dashed lines are hydrogen bonds and the circles Cl ions. The four-moiety groups, in the sequence hydroquinone lumiflavin lumiflavin hydroquinone, lie approximately in the (301) planes. Similar DAAD groupings are found in the structure of (2,6diamino-9-ethylpurine lumiflavin)2C2H5OH.H2O (Scarborough et al., 1977). (Reproduced from Karlsson, 1972.) n n n

n n n

n n n

n n n

n n n

(Section 15.5.1). A much larger sample of flavin complexes will have to be examined before drawing conclusions from the high frequency of interstitial donors among the complexes of Table 15.9. Judging from the colours of the complexes it appears that the flavins act as quasiacceptors in the first three complexes of Table 15.9 with essentially no charge transfer interaction with the donor; indeed adenine and 5 0 -bromo-5 0 -deoxyadenosine are both weak donors. The geometry of the -systems is rather variable and the flavins appear to belong to that group of acceptors in which the strength of the -interaction varies little with the mutual lateral position of donor and acceptor molecules. This implies that any part of the flavin nucleus can participate in charge transfer interaction and that the actual overlap that occurs in the crystals depends both on the nature of the partners and the requirements of the hydrogen bonding scheme. This may well be a requirement for a moiety so widely involved in biological and other oxidation-reduction reactions. The complexes containing riboflavin all crystallize in chiral space groups but this is just a consequence of the chirality of riboflavin itself. 15.7.3 15.7.3.1

Other crystals with both charge transfer and hydrogen bonding interactions Crystals without solvent of crystallization

In {1,5-diaminonaphthalene chloranil} (DAN CA) (Tamura and Ogawa, 1977; ˚ ), with CANANP) there are mixed stacks of DAN and CA molecules along [010] (7.95 A n n n

n n n


1027

mixed donor–acceptor stacks with hydrogen bonding between adjacent stacks chloranil

2.749 1,5-diaminonaphthalene

Cl

N1

O1

N2

2.969 O2

y x

Fig. 15.19. Crystal structure of (DAN CA) projected down [001]. The N atoms are shown cross˚ . The two kinds of hydrogen bonds between amino groups and chloranil hatched. Distances in A oxygens are shown by broken lines. (Data from Tamura and Ogawa, 1977.) n n n

hydrogen bonding between carbonyl and amino groups of molecules in different stacks separated by [001]. Thus sheets of molecules, linked by both charge transfer interactions (there is carbonyl group–aromatic ring overlap) and hydrogen bonds, are formed in the (100) planes. These sheets are then stacked along [100] (Fig. 15.19). The component molecules are, of course, centrosymmetric and so is their arrangement within the sheets, at least to a very good approximation. However, the space group is Pn, which is noncentrosymmetric (see Table 15.4). It seems that most efficient packing of the sheets is obtained when they are mutually offset along [010]. One may perhaps anticipate that both charge transfer interactions and hydrogen bonds will be found in {durendiamine chloranil} (Matsunaga, 1964) and in the two polymorphs of {1,6-diaminopyrene chloranil} (Matsunaga, 1966) but crystal structures have not been reported. Hydrogen bonding interactions have been used in an attempt to convert the herringbone arrangements of charge-transfer crystals into more planar arrangements, hopefully increasing the interaction between donor and acceptor moieties by reducing the interplanar distance and thus affording observable changes in moiety geometries (Bock, Seitz et al., 1996). Specifically the herring-bone structure of {perylene 1,2–4,5tetracyanobenzene} (P21/c, Z ¼ 4; REHMUM) was compared with that of {pyrene 2,3,5,6-tetracyanohydroquinone} (P 1, Z ¼ 1; TEXPOB10). Indeed, both planarization and (minor) shortening were achieved (Fig. 15.20), but there was no effect on moiety geometries. Perhaps more interesting is the way in which the formation of CN . . . HO hydrogen bonds gives a structure composed of pyrene and hydrogen-bonded tetracyanohydroqunone layers without removing the integrity of the donor–acceptor stacks. n n n

n n n

n n n

n n n


NC

CN

NC

CN

345 pm

OH NC

CN

NC

CN

340 pm `

OH

277 162 191

`

(A) (B) 290 164° 204

(C)

197 156 172 159 256 299

299 `

`

Fig. 15.20. Comparison of the crystal structures of {perylene 1,2–4,5-tetracyanobenzene} (P21/c, Z ¼ 4) and {pyrene 2,3,5,6-tetracyanohydroquinone} (P1, Z ¼ 1). Both are viewed normal to the stack axes. Note the central layer of hydrogen-bonded tetracyanohydroquinone moieties. In the lower part of the figure the hydrogen-bonding patterns found with the 2,3,5,6-tetracyanohydroquinone molecular compounds with pyrene (left) and perylene þ 2H2O are shown. (Reproduced from Bock, Seitz et al., 1996.) n n n

n n n

When perylene is co-crystallized with 2,3,5,6-tetracyanohydroquinone, two molecules of water are incorporated in the crystals (TEXPUH10) and the dimerized hydrogen-bonding pattern found in {pyrene 2,3,5,6-tetracyanohydroquinone} is replaced by an interrupted pattern (Fig. 15.20 above). The donor–acceptor stacks remain. n n n

15.7.3.2 Crystals with solvent of crystallization For some donor–acceptor combinations, the combined occurrence of charge transfer interactions and hydrogen bonding allows formation of a more open arrangement of molecules, with interstices or channels which can be occupied by solvent. This occurs in the TCNQ compounds of benzidine (BD), toluidine (TL) and diaminobenzidine, where solvent-free and solvated types are found (Ohmasa et al., 1971). The BD/TCNQ system is the most thoroughly studied and three types of crystal have been obtained.


1029

I – solvent-free BD TCNQ (crystal structure reported (Yakushi et al., 1974a; BZTCNQ); II – with aliphatic guests such as acetone, acetonitrile, CH2Cl2, CH2ClCH2Cl and CH2BrCH2Br. These crystals are all isostructural and the structure of BD TCNQ1.8CH2Cl2 has been determined in detail (Ikemoto, Chikaishi et al., 1972; BDTCQC10); III – with aromatic guests such as benzene and substituted benzenes (X ¼ CH3, Cl, Br, NO2, CN). The crystal structure of BD TCNQC6H6 has been determined (Yakushi, Ikemoto and Kuroda, 1974b; BDTCNB10). n n n

n n n

n n n

The crystal structures of {BD TCNQ1.8CH2Cl2} and {BD TCNQC6H6} are based on similar arrangements of hydrogen-bonded BD and TCNQ molecules in layers, as illustrated for {BD TCNQ1.8CH2Cl2} in Fig. 15.21. The stacking of these layers is rather different in the two types of inclusion complex. In the CH2Cl2 complex the layers are directly superimposed so that channel and stack axes coincide, while in the benzene complex a BD TCNQ pair forms the repeat unit in the stack and channel and stack axes are mutually inclined at an angle of 30 . The enthalpy of formation of crystalline {BD TCNQ1.8CH2Cl2} has been measured (Ohmasa, Kinoshita and Akamatsu, 1971), the reaction being: n n n

n n n

n n n

n n n

n n n

BD TCNQ(s) + 1.8CH2Cl2(g) ! BD n n n

n n n

TCNQ1.8CH2Cl2(s) + 83.6 kJ.

The reaction is thus exothermic and the inclusion complex is stable with respect to its components. The composition with 1.8 molecules of CH2Cl2 is the most stable, and the 10%

3.110 Å

C N Cl

Fig. 15.21. Projection of the {BD TCNQ1.8CH2Cl2} structure onto (001). The arrangement of BD and TCNQ molecules in a single layer of the benzene complex is very similar. (Reproduced from Ikemoto et al., 1972.) n n n


Table 15.10. Conductivities of benzidine/TCNQ inclusion complexes Type

Formula

Conductivity (ohm cm)1

Activation energy (eV)

I II III

{BD TCNQ} BD TCNQ1.8CH2Cl2 BD TCNQC6H6

109 104 106

0.54 0.11 0.28

n n n

n n n

n n n

of vacancies on solvent sites must contribute some entropy stabilization as well. All solvent sites are occupied in the benzene complex. It seems unlikely, however, that (in both types of complex) the solvent molecules are as ordered as depicted in the structure diagrams. Very different electrical conductivities and activation energies are found for the three types of crystal (Takahashi et al., 1976), the conductivities of the inclusion complexes being much larger than those of the solvent-free crystals (Table 15.10). The conductivity of {BD TCNQxCH2Cl2} is strongly dependent on the concentration of solvent vacancies; in the most stable crystals (x ¼ 1.8) there is extrinsic semiconduction due to the solvent vacancies. It is the presence of these vacancies that probably accounts for the larger conductivity and smaller activation energy of Type II as compared to Type III crystals. Polarized reflectance spectra of single crystals of {BD TCNQ} and {BD TCNQ1.8CH2Cl2} have been measured at 295 and 30K at ambient pressure and also at 295K, 56.7 kbar for {BD TCNQ}. The degree of charge transfer at 295K, 1 bar was estimated to be 0.3 and increased by 30% on cooling or application of pressure; however, the crystals retained their neutral ground states (Yakushi et al., 1985). n n n

n n n

n n n

n n n

N N

DPC N H

Monoclinic {[bis(dihydro-5,6-pyrimidino[5,4-c]carbazole)- TCNQ]-dihydro-5,6pyrimidino[5,4-c]carbazole dihydrate} (P21/c, Z ¼ 2; {[DPCTCNQDPC]DPC2H2O} for short; Dung et al., 1986; DULFAR) has mixed . . . DAD . . . stacks (shown within the ˚ ) linked by DPC and water molecules. Only the firstsquare brackets) along [100] (11.04 A mentioned DPC molecules are involved in donor–acceptor interactions. The mixed stacks retain their integrity despite the considerable perturbation of the structure. n n n

15.8

Mixed-stack crystals with both delocalized and localized charge transfer interactions

By analogy with the structures discussed in Section 15.7, one may anticipate mixed-stack structures in which the hydrogen bonding between stacks is replaced by localized charge transfer interactions, acting in addition to the delocalized charge transfer interactions

DEL OCALIZED AND LOCALIZED CHARGE TRANSFE R I NTERACTIONS

N(4)

a 3.15

N

N C

N(3) S(1)

3.37

C

1031

S(2) S(3)

3.54

N

0

N

S

c

S

N

N

N

C

C

N N(4)

bis(1,2,5-thiadiazolo) tetracyanoquinodimethane (BTDA-TCNQ)

3.15 S(2)

S(3) 3.58 S(1)

[b] – [101]

Fig. 15.22. Crystal structure of {TTF (BDTA-TCNQ)}, where the acceptor is bis(1,2,5-thiadiazolo)tetracyanoquinodimethane. (upper) ‘‘Sheet-like’’ network in the ac plane showing the secondary N S interactions between TTF and acceptor molecules and also between the acceptor molecules themselves; the TTF molecules are shaded. (lower) Arrangement of the corrugated sheets along the [010] (stack) axis, with -* interactions in the [010] direction. (Adapted from Susuki, Kabuto et al., 1987.) n n n

n n n

within the stacks. This is a striking feature of the structural arrangements in the Bechgaard salts, important as the first organic superconductors (Williams et al., 1985) but outside the scope of this book. So far only one example is known among the mixed stack structures – {TTF (BTDA-TCNQ)}, where preparation of the acceptor (and of a number of its molecular compounds (Yamashita et al., 1985)), determination of the crystal structure of the acceptor (Kabuto et al., 1986; FARSOG6 ) and that of the TTF molecular compound (Suzuki, Kabuto et al., 1987; FUVYEA), have been described. The neat acceptor (formula in Fig. 15.22) forms a two-dimensional sheet-like network with strong n n n

6

For reinterpretation in space group C 2/m see FARSOG01 (Suzuki, Fujii et al., 1992).


S . . . N C– interactions. The equimolar molecular compound is monoclinic (a ¼ ˚ , ¼ 91.312(1) , Z ¼ 2, space group P21/n) and 9.660(1), b ¼ 7.231(1), c ¼ 14.628(2) A has a neutral ground state. The interactions among the components are such that a threedimensional network is formed.

15.9

Donors and acceptors with special chemical features

We use the term ‘quasi-acceptor’ to describe the second component in binary systems where a donor and this component form mixed stack crystals but other evidence (generally spectroscopic) for a charge transfer interaction between the components is lacking. 15.9.1

Fluorinated aromatics as quasi-acceptors

Considerable effort has been devoted to the study of systems of aromatic hydrocarbons and polyfluorinated aromatics (Fenby, 1972; Swinton, 1974; Table 15.11); the spectroscopy and thermodynamics of solutions have been studied, phase diagrams have been reported and a number of crystal structures determined (Table 15.12; we include some other quasi-acceptors in this Table for convenience). Resurgence of interest in this area has been prompted by (expressed in currently fashionable terms) recognition of the areneperfluoroarene interaction as an important supramolecular synthon (Dai, Nguyan et al., 1999) with potential applications for solid-state chemistry, crystal engineering, molecular electronics, liquid crystals etc. Congruently melting 1 : 1 compounds have been found in the phase diagrams of C6F6 with benzene, toluene, p-xylene and mesitylene (Duncan and Swinton, 1966) and congruently melting 1 : 1 compounds in the phase diagrams of benzene and naphthalene with perfluoronaphthalene, and of benzene with perfluorobiphenyl, while the {biphenyl–C6F6} compound melts incongruently (McLaughlin and Messer, 1966). Equimolar molecular compounds have been found in most binary phase diagrams but not other compositions. No evidence of –* charge transfer interaction has been Table 15.11. Melting points of equimolar molecular compounds formed from various donors and hexafluorobenzene or octafluoronaphthalene Donor/hexafluorobenzene

M.pt. (K)

Donor/ octafluoronaphthalene

M.pt. (K)

Benzene (Patrick and Prosser, 1960) 1-methylnaphthalene (Griffith et al., 1983) 2-methylnaphthalene (Jackson and Morecombe, 1986) 2-ethylnaphthalene (Dahl, 1988)

297

Anthracene (Collings, Roscoe Phenanthrene (Collings, Roscoe Pyrene (Collings, Roscoe Triphenylene (Collings, Roscoe

447

Tetralin (Dahl, 1988) Quinoline (Dahl, 1988) iso-quinoline (Dahl, 1988)

373 335.7 306.7 266.7 318.1 285.4

et al., 2001) 445 et al., 2001) 522 et al., 2001) 478 et al., 2001)

DONORS AND ACCEPTORS WITH SPECIAL CHEMICAL FEATURES

1033

found in solutions of aromatic hydrocarbons and fluoroaromatics but there may be n–* charge transfer when nitrogen-containing heteroaromatics are used as donors (Beaumont and Davis, 1967); there is some crystallographic evidence for specific interactions between nitrogens of heteroaromatic donors and quasi-acceptors. Mixed stacks of donor and quasi-acceptor molecules have been found in almost all the crystal structures so far reported and thus these structures resemble those of -molecular compounds insofar as packing arrangements are concerned (Dahl, 1988). However, the absence of charge-transfer bands in the solution spectra, the colourless nature of the crystals, and the fact that lattice energy minimizations based on atom–atom potentials do not reproduce the observed crystal structures very well (Dahl, 1990) all indicate that directional forces of another kind must be invoked to account for the retained occurrence of mixed stacks despite replacement of true acceptors by quasi-acceptors. One possibility is localized dipole–dipole interaction between C-H and C-F bonds, expressed in terms of quadrupole interactions. The fact that both benzene and hexafluorobenzene have large molecular quadrupole moments (MQM), of similar magnitude but of opposite sign7, led to the suggestion (Williams, 1993) that the stacking motif is dictated by electrostatic quadrupole-quadrupole interactions (see also Clyburne et al., 2001). The negative values for aromatic hydrocarbons indicate that the negative charge is clustered in the center and above the molecular plane, while that for perfluoroaromatics is located at the molecular perimeter and therefore there is a depletion of electron density at the center of the ring (Fig. 15.23). Illustrations, apparently based on Williams (1993), have been given by Dunitz (1996) and Collings, Roscoe et al. (2001). Many of the crystals show low temperature phase transitions; for example, the structures of four 1 : 1 benzene–hexafluorobenzene phases have been determined (Table 15.13). The structural progression has been described as follows (Williams et al., 1992) ‘‘In phase I the molecules are rotating about the column axis, and the system behaves like parallel cylindrical rods. On lowering the temperature the molecules first tilt with respect to the column axis, leading to a monoclinic phase, followed by a distortion to a triclinic phase due to the freezing out of the rotations of the heavier C6F6 molecules. At this stage the unfavorable interaction between the columns is compensated by the rotational freedom of

F– F–

F– + + F–

H H

+

+

–

+ + +

F–

– F–

H

+

+

H

–

+

– – – H H

+

+

Fig. 15.23. Schematic charge distributions in hexafluorobenzene and benzene. 7

MQM of benzene 29.0, naphthalene 45, ferrocene 30 and hexafluorobenzene þ31.7, all 1040 C m2.

Table 15.12. Summary of information available about mixed-stack crystal structures of molecular compounds formed between aromatic hydrocarbons and quasi-acceptors Aromatic hydrocarbon/reference/refcode

Quasi-acceptor

Space group

Z

Remarks

1. Anthracene (Boeyens and Herbstein, 1965a); ZZZGMW

C6F6

C2/m

2

Only cell dimensions and space groups (at 300K) reported for these three compounds.

2. 3. 4. 5. 6. 7.

Pyrene (Boeyens and Herbstein, 1965a); ZZZGKE Perylene (Boeyens and Herbstein, 1965a); ZZZLJY p-Xylene (Dahl, 1975a); PXYHFB Mesitylene (Dahl, 1971a); MTYHFB Durene (Dahl, 1975a); DURHFB C6(CH3)6; MBZFBZ

C6F6 C6F6 C6F6 C6F6 C6F6 C6F6

C2/m P21/a P 1 Pnma C2/m R 3m; P 1

8. N,N-dimethylaniline (Dahl, 1981b); BAPLEJ 9. N,N-dimethyl-p-toluidine (Dahl, 1981a); METOFB 10. TMPD (Dahl, 1979); MPAHFB 11. Trans-stilbene (Batsanov, Howard et al., 2001); TIJTUB 12. Perdeuterobenzene (Overell and Pawley, 1982); BICVUE 13. Cr(6-C6H6)2 (Aspley et al., 1999); FIBGUS

C6F6 C6F6 C6F6 C6F6

P 1 I2/m P 1 P21/c

2 2 1 4 2 3 1 1 2 1 2

14. o-diethynylbenzene dimer (Bunz and Enkelmann, 1999); JOCRIC 15. hexamethylmelamine (Aroney et al., 1987) 16. Naphthalene (Potenza and Mastropaolo, 1975); NFOFNP 17. anthracene (Collings, Roscoe et al., 2001); ECUTUR

triclinic below 223K at 298K (Dahl, 1971b); at 233K (Dahl, 1973)

Colourless plates

C6F6

m; or R3 R3m P 1

C6F6

P21/a

4

C6F6 octafluoronaphthalene

C2/m P21/c

2 2

Poor, disordered crystals

octafluoronaphthalene

P21/a

2

At 120K

C6F6

1 1

at 299K; phase transitions at lower temperatures (see Table 15.13) Red-pink blocks; there are also yellow plates of rather similar structure not described in detail. At 165K.

18. phenanthrene (Collings, Roscoe et al., 2001); ECUVED 19. pyrene (Collings, Roscoe et al., 2001); ECUVIH 20. triphenylene (Collings, Roscoe et al., 2001); ECUVON 21. TTF (Batsanov, Collings et al., 2001); TIJVAJ 22. Diphenylacetylene (Collings, Batsanov et al., 2001; at 100K); OCAYIA 23. 2,6-dimethylnaphthalene (Birtle and Naae, 1980); BPHFPC 24. 1,8-diaminonaphthalene (Batsanov, Collings, Howard and Marder, 2001); EDAWAH 25. ferrocene (Clyburne et al., 2001); YEBQOL 26. ferrocene (Burdenuic et al., 1997) 27. bis(decamethylferrocene) (Beck et al., 1998); SIBQOJ 28. bis(N-(2,6-dimethylphenyl)N’-(1,3,4,5,6,7,8heptafluoro-naphth-2-yl)sulfurdiimide (Lork et al., 2001); NEMHOC 29. biphenyl (Lin and Naase, 1978); ZZZBRD 30. Biphenyl (Naae, 1979); BPPFBP 31. 4-bromobiphenyl (Birtle and Naase, 1980); BPHFBP 32. 4-methylbiphenyl (Birtle and Naase, 1980); BPHFPA 33. Naphthalene (Foss et al., 1984); CEKYUM 34. triphenylene (Weck et al., 1999); CUKXIP 35. Trans-stilbene (Bruce et al., 1987); SERQAH 36. Benzene (Hazell, 1978); DCLPYR 37. Diphenylbutadiyne (Coates et al., 1997). M.pt. 360K;


P21/a

2

Below 250K

octafluoronaphthalene octafluoronaphthalene

P1 P 1

1 2

At 120K At 120K

octafluoronaphthalene octafluoronaphthalene

P21/c P21/a

2 2

At 120K


P 1


P21

2

1.5(Octafluoronaphthalene) Perfluorophenanthrene Perfluorophenanthrene

P 1 P 1 P21/c

2 2 4

Pseudo-isostructural with naphthaleneoctafluoronaphthalene At 223K.

Octafluoronaphthalene

P 1

1

2,3,4,5,6-pentafluorobiphenyl decafluorobiphenyl decafluorobiphenyl

C*/c C2/c P21/c

4

decafluorobiphenyl

P21/c

decafluorobiphenyl Perfluorotriphenylene trans-decafluoro-azobenzene decachloropyrene Decafluorodiphenyl butadiyne. M.pt. 387K.

C2/c C2/c P 1 P21/c P 1

4 8 1 4 1

At 173K.

Orange crystals, m.pt. 428–430K. M.pt. 425K.

Table 15.12. (Continued ) Aromatic hydrocarbon/reference/refcode

Quasi-acceptor

38. Benzalazine (Vangala, Nangia and Lynch, 2002); EGAWEO 39. Pyrene (Damiani, De Santis et al., 1965) 40. Benzo[c]pyrene (Damiani, Giglio, Liquori and Ripamonti, 1967) 41. Coronene (Damiani, Giglio, Liquori, Puliti and Ripamonti, 1967) 42. Acridine (Yamaguchi and Ueda, 1984; Marsh, 1986); CEJTAM self complex

Bis(pentafluorophenylmethylidene) hydrazone TMU (TMU)2

P 1

2

Pc P1

2 1

(TMU)2

P1

1

1,4-dithiintetra-carboxylicN,N 0 -dimethyldiimide 2,3,4,5,6-pentafluorobiphenyl (Brock et al., 1978); PFBIPH

P21/n

2

C2221

4

Notes: (1) TMU is 1,3,5,7-tetramethyluric acid. (2) There is no evidence for charge transfer interaction in compound #43. (3) Beck et al. (1999) note the preparation of {(ferrocene)4octafluoronaphthalene} but a structure was not reported.

Space group

Z

Remarks


1037

˚ , deg.) for the four C6H6–C6F6 phases (angle values are given only if Table 15.13. Crystal data (A different from 90 ) Phase

T(K)

a/

b/

c/

V/formula ˚ 3) unit (A

Z

I (Overell and Pawley, 1982) II (Williams et al., 1992) III (Williams et al., 1992) IV (Williams et al., 1992)

279

11.952

11.952

7.238/120

299

3

R 3m

260

6.631

12.330/99.67

7.302

295

2

I2/m

215

6.380/93.99

12.338/96.37

7.395/91.85

284

2

P 1

30

9.516

7.429/95.60

7.537

265

2

P21/a

Space group

the C6H6 molecules. Reduction of their rotational energy leads to a realignment of the columns and the unusual transition from a triclinic to a monoclinic phase with decreasing temperature . . . The columnar nature of the structure[s] is a result of the strong attraction between the quadrupole moments of opposite phase of the benzene and hexafluorobenzene molecules.’’ Shorter range electrostatic interactions determine the details of the packing. The structure determinations are noteworthy because they were carried out by combined use of X-ray (synchrotron) and neutron diffraction on polycrystalline samples. The electron affinity of hexafluorobenzene has been measured as 0.86(3) eV (Wentworth et al., 1987). The enthalpy of formation of crystalline {C6H6 C6F6} has been measured by a differential scanning calorimetric method as þ1.0(3) kJ/mol (Brennan et al., 1974). This implies that the entropy of formation must be positive and greater than 3.3 J/mol K (at 300K). In contrast to this situation, most -molecular compounds are enthalpy-stabilized (see Section 16.5). The enthalpy of formation of {p-xylene C6F6} has been reported to be –0.08 0.20 kJ/mol (Ott et al., 1976). The {triphenylene–perfluorotriphenylene} molecular compound, first prepared by Smith and Massey (1969), has a melting point of 524K, compared with 472K for triphenylene and 382K for perfluorotriphenylene (structure by Hursthouse, Smith and Massey, 1977; Fdd2, Z ¼ 8; PPTRPH). The DSC trace of the molecular complex shows first order transitions at 294 and 377K. The crystal structure reported (but the phase was not specified) is particularly interesting (Fig. 15.24; Weck et al., 1999; 7.390 20.987 ˚ , ¼ 95.26 , C2/c, Z ¼ 8; CUKXIP)) and has many features typical of the mixed16.998 A stack structures. There are mixed stacks of the two components, seen edge-on in the upper part of Fig. 15.24. This also shows the appreciable distortions from planarity of both components; Weck et al. summarize these as 33 and 16 tilts in the skeletons of perfluorotriphenylene and triphenylene in the cocrystal, to be compared with 40 and 2 tilts in the neat crystals. In the lower part of the figure one notes the almost ideal superpositioning of the component molecules, an unusual situation in crystal packing ˚ , while arrangements. The centroid to centroid distances for the central rings is 3.698 A ˚. the shortest intermolecular C . . . C distance is 3.369 A n n n

n n n


Most of the structures listed in Table 15.12 are of the mixed stack type (as illustrated in Figs. 15.24 and 15.25), some having particular features needing comment. The decafluorobiphenyl (C12F10) molecule is not planar; the dihedral angles between the two rings are 59.5 in neat crystalline C12F10 (Gleason and Britton, 1976; DECFDP01) (and 64.4 in

B

mixed stacks (horizontal) triphenylene

perfluorotriphenylene

y

x

mixed stacks (vertical)

Fig. 15.24. The triphenylene . . . perfluorotriphenylene crystal structure viewed (above) edge-on to the stacks (fluorines are large cross-hatched circles and hydrogens are open circles) and (below) normal to the mean molecular planes, showing the close-packed arrangement of the mixed stacks; perfluorotriphenylene has been removed from the central mixed stack. (Data from Weck et al., 1999.)


1039

z

y

Fig. 15.24. (Continued )

the gas phase (Bastiansen et al., 1989)), 55.3 in the naphthalene compound, 50.8 in the biphenyl compound and 52.9 in the analogous molecule 2,3,4,5,6-pentafluorobiphenyl (C6H5–C6F5) (PFBIPH). In {C12H10 C12F10} (BPPFBP) the biphenyl molecule is also nonplanar (dihedral angle 36.6 ); however, mixed stacks are formed with an angle of 7.1 between juxtaposed halves of each molecule. A similar mixed stack arrangement is found in 2,3,4,5,6-pentafluorobiphenyl with superpositioning of C6H5 and C6F5 portions of the molecule; this compound must therefore be classed as a self-complex. Stacks are not formed in {naphthalene-decafluorobiphenyl}, where naphthalene is planar and decafluorobiphenyl nonplanar; instead the naphthalene molecule is sandwiched between C6F5 halves of neighboring C12F10 molecules, with an interplanar angle of 6 . The naphthalene in the complex is phosphorescent, in contrast to its lack of phosphorescence in its neat crystals. The mixed-stack {naphthalene octafluoronaphthalene} molecular compound has been studied by Raman spectroscopy at ambient pressure down to 10K, and at ambient temperature in the pressure range 1–80 kbar and found to be stable (i.e. no phase transitions or chemical reactions) under these conditions (Desgreniers et al., 1985). The unusual composition of {ferrocene1.5(octafluoronaphthalene)} (Clyburne et al., 2001) is n n n

n n n


ferrocence decafluorophenanthrene decamethylferrocence

Fig. 15.25. Comparison of the packing arrangements found in {ferrocene deca fluorophenanthrene} (on the left : a 2 þ 2 supersandwich) and {decamethylferrocene deca fluorophenanthrene} (SIBQOJ; Beck et al., 1998) (on the right : mixed stacks). (Adapted from Burdeniuc et al., 1997.) n n n

n n n

explained by noting that there are 1 : 1 mixed stacks of ferrocene and octafluoronaphthalene separated by ‘‘octafluoronaphthalene of crystallization’’, the molecules being located at centers of symmetry. The {ferrocene decafluorophenanthrene} structure (Burdeniuc et al., 1997; no refcode) is unusual in that there is a 2 þ 2 supersandwich with the ferrocenes located between the wingtip arene rings of the decafluorophenanthrene (Fig. 15.25). The {naphthalene (p-iodotetrafluorobenzene)2} (P1, Z ¼ 1) (Law and Prasad, 1982) molecular compound possibly belongs in this group. The crystal structure has not been reported but a Raman study shows that the crystals are ordered, with the components linked mainly by van der Waals interactions; no low-lying charge transfer band was found. Our inclusion of {benzene–decachloropyrene} in Table 15.12 is questionable; although mixed stacks are formed the material could be an inclusion complex of an unusual type. A spectroscopic study is needed to establish whether the orange color of {trans-stilbene trans-decafluoroazobenzene} is due to charge transfer interaction. n n n

n n n

n n n

15.9.2

1,3,5,7-Tetramethyluric acid (TMU) as quasi-acceptor

TMU forms crystalline molecular compounds with a number of aromatic hydrocarbons and some crystal structures have been reported (Table 15.10). These are all mixed stack structures based on ---DQDQ--- and ---QDQ QDQ--- sequences (where Q represents the quasi-acceptor TMU). There is considerable disorder and low-temperature studies would appear to be necessary for understanding the interaction between the components. No charge transfer bands are found in the solution or solid state spectra.

15.9.3

Acceptor is a metal coordination complex

Examples of metal chelate complexes which act as electron acceptors have been briefly noted in Chapter 13 and this material will now be expanded. The 1,2-ethylenebis (dithiolene)-M system (see Section 13.3.3, Table 13.4 for formula) forms many different


1041

salts and also molecular compounds with ionic and neutral ground states. For example, equimolar perylene and pyrene molecular compounds have been prepared with Ni dithiete (X ¼ CF3) as acceptor and found to have mixed stack structures with neutral ground states (Schmitt et al., 1969; PERNIT). The crystals are semiconductors with roomtemperature resistivities of 105 ohm cm; the activation energy for conduction in the perylene molecular compound is anisotropic with minimum activation energy along the stack axis. ,,,@-Tetraphenylporphyrinato-Zn(II) forms an equimolar neutral ground state molecular compound with Ni dithiete (from spectroscopy, crystal structure not known) but the analogous Co(II) porphyrin links to the Ni dithiete by a bridging Co–S bond to give a covalent molecule rather than a molecular compound (Shkolnik and Geiger, 1966). Ni dithiete is the anion in some ion-radical salts to be discussed later (Section 15.10.2). Bis(difluoroboronbenzimidazole)Ni(II) (Ni(dmgBF2)2; formula in Table 13.4) forms equimolar molecular compounds with the donors anthracene (Stephens and Vagg, 1981; BADZOV) and benzimidazole (Stephens and Vagg, 1980; FBGLNJ). The acceptor exists in the solid state and in solution as a weakly bonded dimer, with –* interaction between the two halves of the dimer. It also occurs as a dimer in the mixed stack molecular compound with benzimidazole. However, in the anthracene molecular compound it is the monomer which acts as acceptor. Analogous examples of the effect of molecular compound formation on donor or acceptor structure have been found in the -dimerization of TCNQ in some ion-radical salts (Table 15.12 below). In both anthracene and benzimidazole molecular compounds donor and acceptor are mutually located so as to avoid interference with the protruding fluorine atom of the BF2 groups. The metal atoms do not play a special role in any of these molecular compounds. In contrast to these nonconducting molecular compounds, {(perylene)2 MS4C4(CN)4} (M ¼ Ni, Cu, Pd) compounds have been found to be fairly good conductors (Alcacer and Maki, 1974); for example, the Pd molecular compound has a room temperature resisitivity of 50 ohm cm. Structures are not known but there are presumably stacks of (perylene)þ 2 moieties with the anions between the stacks (cf. Section 17.3). Molecular compounds are known with a 2 : 1 ratio of toluene to tetraphenylporphyrinato-M(II) [M ¼ Mn (Kirner et al., 1977); Cr (Scheidt and Reed, 1978); Zn (Scheidt et al., 1978)]. The Mn and Zn compounds are triclinic and isomorphous (P1, Z ¼ 1), while that with Cr is monoclinic (P21/c, Z ¼ 2). All three molecular compounds contain ˚ DAD sandwiches, with toluene and porphyrin planes approximately parallel and 3.5 A apart; the DAD triples are not stacked but are effectively isolated from one another (cf. Section 15.2). The bis(toluene)porphyrin arrangement is thought to be indicative of donor-acceptor interaction. We have discussed this group of molecular compounds in Chapter 8. The bis(arene)Fe(II) dication has been shown to act as an electron acceptor in the formation of mixed stack charge transfer crystals with the electron donor arenes ferrocene and durene, PF 6 acting as counterion. The structures of the carmine tetragonal crystals of [Cp2Fe, (durene)2Fe2þ, (PF 6 )2] (VIPJUZ) and of the red-orange triclinic crystals of [durene, (hexamethylbenzene)2Fe2þ, (PF 6 )2]acetone (VIPKAG) have been determined. There is spectroscopic evidence for the effectiveness of other arenes such as mesitylene, pentamethylbenzene, 1,4-dimethylnaphthalene and 9-methylanthracene (Lehmann and Kochi, 1991). n n n


15.9.4

Donor is a metal coordination complex

15.9.4.1 Metal chelates as donors The structures of some seven metal oxinate acceptor molecular compounds have been determined (for references see Table 15.2), as well as those of a number of other metal chelate acceptor molecular compounds. The formulae of the metal chelates are shown in Fig. 15.26. These are all equimolar (or effectively equimolar) mixed-stack, neutral ground state molecular compounds, diamagnetic (except where the metal atom is paramagnetic) and with high resistivities. The interaction between metal atom and acceptor is weak and plays little, if any, role in determining the crystal structures. This is shown dramatically by the resemblances between the component arrangements in {Pd(II) oxinate chloranil} (Kamenar, Prout and Wright, 1965; CLAQPD) and in the metal-free {(8-hydroxyquinolinol)2 chloranil}.(Prout and Wheeler, 1967; HQUCLA). However, there is at least one exception (Matsumoto et al., 1979) to this generalization. Three members of this group show special features. Using bis(N-alkyl-2oxynaphthylidene-aminato)Cu(II) and Ni(II) chelates as donors, a series of 1 : 2 molecular compounds was prepared (Matsumoto et al., 1979) with TCNQ and chloranil as acceptors, and crystal structures were determined for Cu(L-i-pr)2 itself and for {Cu(L-ipr)2 (TCNQ)2}, (IPONTC) where L-i pr is the isopropyl substituted ligand. In Cu(L-ipr)2 itself the coordination at Cu is distorted tetrahedral, with a dihedral angle of 39 between the two intersecting Cu(NO) planes; this distortion from the expected square planar arrangement was ascribed to steric hindrance between the two bulky iso-propyl groups. In {Cu(L-i-pr)2 (TCNQ)2} there are two antiparallel sets of mixed stacks, with ‘‘naphthalene’’ portions of the chelate and TCNQ molecules in alternating array. The geometry of the metal chelate has changed to a centrosymmetric stepped structure with square planar coordination about Cu. It was suggested (Matsumoto et al., 1979) that the overall arrangement represents a compromise in which the favoured coordination about Cu, as indicated by the arrangement taken up in the neat compound, is distorted minimally in order to achieve the most favourable stacking of ‘‘naphthalene’’ and TCNQ moieties. This is one of the most striking examples known of the effect of molecular compound formation on donor (or acceptor) geometry. In {[Cu(salphen)2]2 TCNQ} (Cassoux and Gleizes, 1980; PSALTQ) there appear, at first sight, to be segregated stacks of metal chelate and TCNQ molecules. However, the conductivity of powders is 108 S/cm and the moieties (especially TCNQ, as judged from bond lengths) appear to be neutral. The paradox is resolved by reference to the stereodiagram of the packing (Fig. 15.27). This shows that the two halves of the centrosymmetric TCNQ molecules interact with benzene rings of salicylaldiaminato groups of two different donors to form two separate ---DADA--- stacks, where A ¼ 1/2(TCNQ). The second salicylaldiaminato groups of each of these donors form a quasi-segregated stack, presumably without charge transfer interaction. In {Ni(gh)2 TCNQ} (Megnamisi-Belome and Endres, 1982; BEXZIN) there is an alternating arrangement of neutral moieties (as judged from the TCNQ bond lengths) but there is so little overlap between them that one can hardly refer to ‘‘mixed stacks.’’ Instead the black colour of the crystals is ascribed to Ni–NC interactions (d(Ni . . . N) ¼ 3.357(5) ˚ ) mediated through the conjugated system of the TCNQ molecules. A n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n


N O

O

R

N

bis(8-hydroxyquinolato) M(II) metal chelates (M = Cu, Ni, Pd)

O N

H

R

O

M(II)

1043

M(II) N O

N

bis(N-alkyl-2-oxynaphthylideneaminato M(II) with M = Cu, Ni and R = CH3, C2H5, i-C3H7, n-C3H7, t-C4H9.

O S

N

S Pt

M N N OH O

S

bis(1,2-benzoquinondioximato)M(II) with M(II) = Ni, Pd. Abbreviated as M(BCD)2

O

S

bis(propene-3-thione1-thiolate)Pt(II)

O Cu

HC N

N

CH

N,N'-(1,2-phenylene)bis(salicylaldiaminato) Cu(II) Abbreviated as Cu(salphen). O

H

N

O

CH3

N N

Ni N

H3C

N

N M

N

N

O

O H3C CH3 H bis(ethanediol dioximato)Ni(II) tetramethylporphyrinato Ni(II) Ni(gh)2

Fig. 15.26. Formulae of metal chelate compounds which act as donors in various molecular compounds.


Fig. 15.27. Stereodiagram of [Cu(salphen)2]2 TCNQ structure viewed down [010]. The [100] axis is horizontal. (Reproduced from Cassoux and Gleizes, 1980.) n n n

15.9.4.2 Metal porphyrins as donors Some one hundred molecular compounds of various porphyrins with a variety of electron acceptors have been reported, most by Treibs (1929) and some by Hill et al. (1967). The molecular compounds are in the main coloured powders but some single crystals were obtained. The crystal structure of {(tetramethylporphyrinato)Ni(II) TCNQ} has been determined (Pace et al., 1982; BEGLUU); it has a mixed stack structure with a neutral ground state, is diamagnetic and has a needle axis conductivity of . Soos et al. (1979) have shown that in a regular charge transfer solid DEm > 0 when Z < 0.68 0.01. For example, DEm is measured as 0.07 eV in {TMPD TCNQ}, which corresponds to Z0.60 – 0.65, in good agreement with values obtained by other methods. The larger values of DEm 0.13 eV in TMPD CA and PD CA imply somewhat smaller values of Z. When Z > 0.68 then DEm is predicted to be zero and the solid becomes paramagnetic. The {TMPD CA} structure is unusual in that there is complete eclipsing of donor and acceptor ions in the quasi-hexagonally close-packed stacks; such eclipsing is found elsewhere only in the (neutral ground state) molecular compound between 2,4,6tris(dimethylamino)-s-triazine and TNB (Williams and Wallwork, 1966). Both X-ray diffraction and ESR measurements show that {TMPD CA} undergoes a phase change on cooling below 250K but the low-temperature structure has not been reported. Another unusual feature is -dimerization9 of TCNQ in the formation of molecular compounds with N-ethylphenazinium and bis(dipyridyl)Pt(II) (nos. 15–17 in Table 15.14). ˚, The two parts of the (TCNQ–TCNQ)2 ion are joined by a covalent bond of length 1.63 A ˚ 0.09 A longer than the standard C–C single bond length; the arrangement at the linked carbons is approximately tetrahedral. Each half of the (TCNQ–TCNQ)2 ion, represented as vertical lines in the diagram, participates in charge transfer interactions in separate stacks while the central portion of the dianion only links the stacks. n n n

n n n

n n n

n n n

n n n

NC

CN

NC

CN

9 -Dimerization is used here to indicate covalent bond formation between two moieties; -dimerization is used to describe pairing of moieties by HOMO–LUMO interation without formation of a covalent bond.

MIXED-STACK DONOR–ACCEPTOR MOLECULAR COMPOUNDS

1049

Table 15.14. Some mixed-stack ion radical salts for which crystal structures have been reported. (tfd)2 is sometimes called dithiete. Phenazine moieties are discussed earlier and, apart from one example, have been omitted Ion radical salt/Reference/refcode

Z

Remarks

1. Tropylium Ni(tfd)2 (Wing and Schlupp, 1970); TRFSNI

1

disorder of cations; Curie–Weiss magnetic behavior, i.e. two independent spins per D þ A pair.

2. Phenoxazinium Ni(tfd)2 (Singhabandu et al., 1975); FMENPX 3. TTF Pt(tfd)2 (Kasper and Interrante, 1976); FMEPTF 4. TMPD chloranil (Boer and Vos, 1968); TMABCA 5. TTF fluoranil (Torrance, Vasquez et al., 1981); TTFFAN

1

n n n

n n n

n n n

n n n

1 0.60 Ionic form is stable at low temperatures or above 9 kBar at 300K

n n n

6. TTF chloranil (Torrance, Vasquez et al., 1981); TTFCAN 7. p-Phenylenediamine chloranil (Hughes and Soos, 1968) 8. Decamethylferroceniumþ DDQH (Gebert et al., 1982); MEFEQU10 9. TMPD TCNQ (Hanson, 1965b); QMEPHE 10. TMBTP TCNQ (Darocha et al., 1979); MBPTCR 11. Decamethylferrocenium TCNQ (Miller et al., 1987); MCFECT01 12. DBTSeF TCNQ (Emge, Bryden et al., 1982); BOWSUB 13. DBTTF 2,5-TCNQF2 (Emge, Wijgul et al., 1982); BITROL 14. OMTTF DBTCNQ (Akhtar et al., 1985) n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

-bonded TCNQ dimers 15. (N-ethylphenaziniumþ)2 (TCNQ–TCNQ)2 (Morosin et al., 1978) EPZTCD 16. [Bis(dipyridyl)Pt(II)]2þ (TCNQ–TCNQ)2 (Dong et al., 1977) TCQDPT; 17. [Bis(2,9-dimethyl-1,10-phenanthroline) Cu(I)] (TCNQ–TCNQ)2 (Hoffmann et al., 1983; CABKEV)

0.5 1

From Em; crystal structure known only in outline Both ions disordered

0.60 1 1 0.47 0.6 0.7 1

n n n

n n n

n n n

structure not stacked

structure not stacked


The ESR spectrum of (NEPþ)2 (TCNQ–TCNQ)2 has been studied (Harms, Keller et al., 1981); in addition to a number of features that were not clearly understood, there are thermally activated triplet spin exciton (TSE) lines (line width 0.5 mT) which show an orientation-dependent fine structure splitting that was identified with an S ¼ 1 excitation of the dianion, and was assigned to a transition to an excited state in which the long -bond was broken. The TSE is quasi-immobilized on the dianion, in contrast to TSEs in other TCNQ salts, which show extreme line narrowing because of the fast diffusional or hopping nature of the paramagnetic excitation along the TCNQ anion stack. In the {Pt(2,2 0 -dipy)2 (TCNQ-TCNQ)2} salt there is a phase transition at 87 C, accompanied by a color change and an enormous increase in paramagnetism (corresponding to two unpaired electrons per formula unit) which is ascribed to breaking of the long -bond. There are a number of molecular compounds where the components have ionic ground states and the stacks are mixed but where a simple alternation of moieties is not found. There are five examples of a –DþDþAADþDþAA– arrangement, which can be further subdivided in terms of the interactions between the moieties. The first example is {phenothiazine Ni(tfd)2}, which is a cation radical, anion radical salt (Geiger and Maki, 1971; Singhabhandhu et al., 1975; FMENPZ). The cation is nonplanar, with a dihedral angle of ˚ ; the 172 and there is strong cation–anion interaction with an interplanar distance of 3.36 A ˚ ˚ anion–anion distance is 3.83 A and the cation–cation distance varies from 3.4–3.9 A because of cation folding. Thus the structure is based on individual cation . . . anion pairs, largely isolated from other such pairs, and could be represented schematically as n n n

n n n

n n n

----Dþ Dþ A A Dþ Dþ A A ----: The magnetic susceptibility is consistent with a spin-paired singlet ground state and a thermally populated triplet excited state; a detailed ESR study has not yet been reported. {Phenoxazine Ni(tfd)2}, which could be expected to be isostructural, has in fact a different stacking (–DþA DþA DþA) and very different physical properties (Singhabhandhu et al., 1975; FMENPX). n n n

X

S

S

X

S M

N

X

S

S

X

H Phenothiazine

M = Ni, X = CF3 Bis(cis-1,2-perfluoromethylethylene-1,2-dithiolato)Ni(II), abbreviated as Ni(tfd)2. M = Pt, X = CN Bis(dicyanoethylene-1,2dithiolato)Pt(II).

The three compounds of the second subset {5-(1-butylphenazinium) tetrafluoroTCNQ} (NBP TCNQF4) (Metzger et al., 1982; BISWAB10); {NBP TCNQ} (Gundel et al., 1983; BISWUU10); {N,N 0 -dimethylbenzimidazolinium TCNQ} (Chasseau et al., 1972; MBZTCQ) are all characterized by strong anion . . . anion interactions. Although {NBP TCNQF4} and {NBP TCNQ} are not isomorphous (space groups P21/c and P1, respectively) the structures of their stacks are very similar. The anion radicals interact ˚ ) and both have ring-external bond (R-EB) overlap strongly (interplanar spacings 3.15 A n n n

n n n

n n n

n n n

n n n

n n n

MIXED-STACK DONOR–ACCEPTOR MOLECULAR COMPOUNDS

1051

while the other interactions are much weaker. ESR studies show that there are ‘‘quasiimmobile’’ Frenkel excitons localized on the pairs of adjacent anion radicals. A similar cation-anion radical arrangement is found in {N,N 0 -dimethylbenzimidazolinium TCNQ}, where the interplanar spacing between adjacent anion radicals is remarkably ˚ , implying strong coupling. Physical properties have not been measured short at 3.07 A apart from conductivity which is very low at 1010 S/cm. These compounds with strongly bound TCNQ -dimers are also noted in Section 17.4.2. The briefly-described {tetrakis(methylthio)TTF TCNQ}10 has a mixed stack ˚, –DDAADDAA– arrangement (Mori,Wu et al., 1987; FIJYEC) with d(D . . . D) ¼ 3.48 A ˚ ˚ d(A . . . A) ¼ 3.41 A, d(D . . . A) ¼ 3.58 A; thus -dimers do not appear to be formed. A similar stacking arrangement is found (Iwasaki, Hironaka, Yamazaki and Kobayashi, 1992) in {TTF 4,8-bis(dicyanomethylene)4,8-dihydrobenzo-[1,2-b : 4,5-b 0 ]dithiophene} ˚ , d(A . . . A) ¼ 3.49 A ˚ , d(D . . . A) ¼ 3.36 A ˚ , indicating that the with d(D . . . D) ¼ 3.62 A strongest interaction is between donor and acceptor units. The bond lengths in TTF suggest that the moieties are present as ions. There is a six-molecule periodicity along the stack axis in {4,4 0 ,5,5 0 -tetraethyltetrathiofulvalene (TCNQ)2} {(TETTF) (TCNQ)2} (Galigne´ et al., 1977; ETFTCQ). ˚ ). The arrangement The space group here is C2/c and the stack axis is [001] (¼22.61 A can be represented schematically as: n n n

n n n

n n n

n n n

1

TCNQ

TETTF C2

n n n

TCNQ

1

TCNQ

TETTF C2

TCNQ

1

TCNQ

TETTF C2

TCNQ

1

where 1 represents a crystallographic centre of symmetry and C2 a crystallographic two fold axis. The mixed stack arrangement explains the low conductivity. It was inferred, from the bond lengths in TCNQ, that 0.4 e had been transferred from donor to acceptor. ˚ and the angle between TETTF The interplanar distance between TCNQ moieties is 3.36 A and TCNQ planes is 9.3 . Thus a possible description is of weak TCNQ -dimers separated by TETTF moieties, with the -dimers having laterally displaced R/R overlap rather than the more usual R/EB overlap. Mixed stacks of neutral (TNF) and charged (TCVPDM) moieties are found in the crystal structure of {tetramethylammonium p-tricyanovinylphenyl-dicyanomethide 2,4,7trinitrofluorenone} {(CH3)4NþTCVPDMTNF} (Sandman et al., 1980; TCVPDA), with the tetramethylammonium acting as counterion. The charge transfer interaction is n n n

CN NO2

CN

NC

O2N

NC

CN

TCVPDM–

10

NO2 O TNF

{bis(tetrakis(methylthio)TTF) TCNQ}has also been studied(FIJYAY). n n n


somewhat limited as there is an angle of 16 between the planes of the two moieties in the stacks; it was considered that TCVPDM behaved as a closed-shell mono-anion electron donor and TNF as a neutral closed-shell electron acceptor (which is its usual role in -molecular compounds).

15.11

Isomeric (polymorphic) molecular compounds

There are a number of examples of binary molecular compounds of the same overall chemical formula occurring in different crystal structures. This is the usual definition of polymorphism, which does not take into account that the nature of the chemical entity in two (or more) polymorphs can vary, without change of chemical formula, from ‘‘hardly different’’ to ‘‘very different’’ (Herbstein, 2001). The charge transfer molecular compounds provide many interesting examples of this range of possibilities, leading to ‘isomerism’ of the molecular compounds, where the state of the components may differ appreciably in the two polymorphs (or isomers), usually as a result of electron or proton transfer. This is evidence that different types of interaction between the components predominate as the crystals are formed under different (but not always well defined) conditions. We call these ‘‘isomeric (polymorphic) molecular compounds’’, with ‘polymorphic’ usually being dropped for brevity. We restrict ourselves in this chapter to molecular compounds of the charge-transfer type, where three types of isomerism are known : Type 1 : If donor and acceptor can interact in different ways without appreciable change in their individual (chemical) structures, then different crystal structures can ensue, depending on which of the different types of intercomponent interaction (e.g. –*; n–*; hydrogen bonding) predominates in a particular isomeric molecular compound. Type 2 : Here the components differ chemically in the different isomers; among the examples are those where the components are neutral in the ground state of one isomeric molecular compound but ionic in the ground state of the other. Type 3 : Here the isomeric molecular compounds are distinguished by the occurrence of proton transfer from acceptor to donor in one isomer but not in the other, where electron transfer (usually virtual electron transfer, with the ground state neutral) takes place. A variety of isomeric molecular compounds is possible in principle in Type 1 but only pairs of isomers occur in Types 2 and 3. It is convenient to include here, as an extension of the discussion of Type 3 isomeric molecular compounds, a group of molecular compounds where both charge transfer and proton transfer occur; these CPT molecular compounds (Section 15.11.4) are not isomers. 15.11.1

Type 1 – isomerism due to different types of interaction without change of moiety structure

Two clearcut examples are currrently known : firstly, in {(9,10-diazaphenanthrene)2 TCNE} there is n–* interaction in the triclinic polymorph and –* in the monoclinic n n n

ISOMERIC (POLYMORPHIC) MOLECULAR COMPOUNDS

1053

polymorph (Shmueli and Mayorzik, 1980) (see also Section 12.4). Secondly, decamethylferrocene and TCNQ form two 1 : 1 salts (Miller, Zhang et al., 1987); the monoclinic salt (MCFETO01) is dark green, metamagnetic and has mixed . . . DADA . . . stacks, while the triclinic salt is purple, paramagnetic and has discrete stacks of DAAD dimeric units (MCFETO02; cf. Section 17.4.2). Formation of the mixed stack structure is kinetically controlled and that of the dimeric stacks thermodynamically.

Table 15.15. Cell dimensions at 120K for the high (red; HT) and low-temperature (black; LT) phases of biphenylene PMDA. The stack axis is [001] for both phases n n n

Phase Red (HT); DURZAR Black (LT); DURZAR01

˚) a(A

˚) b(A

(deg.)

˚) c(A

9.280(1)

11.869(2)

7.293(1)

98.68(2)

13.368(1)

5.809(1)

10.443(1)

102.30(1)

Volume/ formula unit

Calculated density g cm3

˚3 794.0 A

1.548

792.3

1.552

(a)

(b)

Fig. 15.29. ORTEP stereodiagrams of (a) donor–acceptor pair in red {biphenylenePMDA}; (b) donor–acceptor pair in black {biphenylene PMDA}. In both diagrams projection is onto the biphenylene plane. (Reproduced from Stezowski, Stigler and Karl, 1986.) n n n


A more subtle example, essentially polymorphic in nature, is provided by {biphenylene PMDA}, which has a first-order phase transition (with hysteresis) at 400K (Stezowski, Stigler and Karl, 1986). Cell dimensions (at 120K) for both phases are given in Table 15.15. The red phase has almost complete overlap of donor and acceptor molecules (Fig. 15.29(a)), with an interplanar angle of 4.1 and appears to be a –* donor–acceptor molecular compound of the standard type; there is typical overlapped disk packing. The black phase has donor and acceptor displaced along their long molecular axes, with an angle of 9.9 between molecular planes; this is typical slipped disk packing. Carbonyl groups of PMDA overlap benzene rings of biphenylene and it seems that both components behave as though composed of two virtually noninteracting halves. Thus one isomer has standard –* interaction whereas the second has more localized interaction between polar carbonyl and polarizable benzene ring. Typical slipped disk packing is also shown, for example, by {trans-stilbene PMDA} (Koduma and Kumakura, 1974a; PYMAST). n n n

n n n

15.11.2

Type 2 – isomerism due to electron transfer

There are two structural groups to be considered. In the first group mixed stack -compounds undergo neutral , ionic polymorphic transitions at temperatures and pressures depending on the components involved, the mixed stack structure remaining largely unchanged through the transitions (see Section 13.2.1). The best studied example is {TTF chloranil} which has a mixed stack structure at room temperature and pressure (P21/n, Z ¼ 2) (Mayerle et al., 1979; TTFCAN) and undergoes a neutral , ionic polymorphic transition at 84K and atmospheric pressure (Batail et al., 1981). The conductivity of {TTF chloranil} at 300K is 8 104 S/cm, in conformity with its mixed stack structure (Torrance, Mayerle et al., 1979). An extended discussion is given in Section 16.9. The second group has pairs of isomers, one of which has a mixed stack structure and the other a segregated stack structure. There is a growing number of examples. In the TMTSF/ TCNQ system the red, semiconducting, and the black, conducting phases (1 : 1 compositions) were crystallized under rather similar conditions. The black form (Bechgaard, Kistenmacher, Bloch and Cowan, 1977; SEOTCR) has a segregated stack structure with ˚ 3, while the red form has charge transfer of 0.6 e and volume/formula unit of 272.0 A a mixed stack structure with a charge transfer of 0.2 e and volume/formula unit of ˚ 3 (Kistenmacher, Emge et al., 1982; SEOTCR01). The red form is obtained by 282.5 A recrystallization of either black or red forms from hot CH3CN (Bechgaard et al., 1977). The {bis(ethylenedithio)tetrathiafulvalene/TCNQ} {BEDT-TTF/TCNQ} system provides another example. One isomer has a mixed stack structure (space group P21/n, Z ¼ 2; ˚ 3), a resistivity of 106 ohm-cm and a sharp IR volume/formula unit of 610.9(5) A spectrum, charge transfer 0.2–0.3 e from TCNQ bond lengths (Mori and Inokuchi, 1987; FAHLEF01); the other has segregated stacks (space group P1, Z ¼ 1; volume/ ˚ 3), high conductivity, a broadened IR spectrum and a metal , formula unit of 598.5(2) A insulator transition at around room temperature (Mori and Inokuchi, 1986; FAHLEF). Similar polymorphism has been reported in {2,7-bis(methylthio)-1,6-dithiapyrene/ TCNQ} {MTDTPY/TCNQ}; Nakasuji et al., 1987). The black mixed-stack crystals n n n

n n n


1055

˚ 3 while that for (monoclinic; FUDTON01) have a volume per formula unit of 608.0(2) A 3 ˚ the segregated stack structure (triclinic; FUDTON) is 596.4(5) A . In all these examples the segregated stack phase has a lower volume/formula unit than the mixed stack phase, suggesting that the former has the lower enthalpy at 300K; it remains to be seen whether this is a general feature confirmed for more examples and by direct measurement of the enthalpy differences. In {1,2-di(4-pyridyl)ethylene) TCNQ} there is a red, semiconducting form with a typical mixed stack, neutral ground state structure (BUKXOU) and a black, conducting form, the structure of which was not reported (Ashwell et al., 1983). Matsunaga (1978) has reported that {7-methylbenzo[a]phenazinium -TCNQ} is dark green when crystallized from ethanol and violet when crystallized from CH3CN. Recrystallization of the green form from CH3CN gave the violet form. Examination of the electronic, vibrational and ESR spectra suggested that the violet form is {NMBPþ TCNQ } and that the green form is non-ionic with a neutral ground state i.e. {NMBP TCNQ}. There was no N-H stretching vibration in the region of 3000– 3500 cm1 in the spectrum of either sample, thus eliminating the possibility that reduction of the diamagnetic NMBPþ cation to the HNMBPþ cation radical had occurred during preparation and that this was the cause of the apparent dimorphism (cf. Section 15.8.5). Thermodynamic results do not appear to be available for any of these systems. There are a number of actual or potential isomeric molecular compounds of the mixed stack/segregated stack type among the various TTF/haloanil binary systems. For example, {TTF fluoranil} (Mayerle et al., 1979; Torrance, Mayerle et al., 1979; TTFFAN) has a structure similar to that of {TTF chloranil} although triclinic (P1,Z ¼ 1). {TTF fluoranil} also appears as a micro-crystalline phase with RT ¼ 10 S/cm and, presumably, a segregated stack structure in analogy with that of {TMTTF bromanil} (RT ¼ 1 S/cm) (Mayerle and Torrance, 1981; TMFBRQ10). n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

15.11.3

Type 3 – isomerism due to proton transfer or to p–p* electron transfer

When the components of a binary system are acids and bases in both the Lowry-Brønsted (proton transfer from acid to base) and Lewis (electron transfer from base to acid) senses, then both types of interaction can occur. For the donor-acceptor combinations considered here the proton transfer will be actual for the ground state moieties but the electron transfer will be virtual, occurring only in the excited state. The comparative strengths of the Lowry–Brønsted and Lewis acid–base interactions will determine which structure occurs at a particular temperature, as the two interactions are temperature-dependent in the crystalline state, with the salt-like structure favoured at lower temperatures. Thus, 0 taking an amine–phenol pair as an example, the salt-like structure {RNHþ 3 – OR } will be favoured at lower temperatures but this may transform to a neutral -molecular compound form {[RNH2] [HOR 0 ]} on heating. The transition temperature can be anywhere between a very low temperature and the melting point, depending on the details of the system. In crystallographic terms the two isomers are polymorphs, the relationship being enantiotropic in most known examples but monotropic in some. Certain organic moieties function effectively as electron acceptors in both neutral and anionic states; for example, the picric acid–picrate anion pair (Matsunaga and Saito, 1972; Saito and Matsunaga, 1972; Saito and Matsunaga, 1973a, b). Thus both charge and n n n


proton transfer can take place in the same crystalline species in many binary systems: these are the CPT complexes discussed in the next section. Early suggestions that ‘isomeric complexes’ could occur were elaborated in a series of experimental studies by Hertel and coworkers in the 1930s. Later Briegleb and coworkers used UV–visible and IR spectroscopy to identify the salt-like and molecular-compound forms of various isomeric complexes. This work has been summarized by Herbstein (1971). Many of the more recent developments are due to Matsunaga and his colleagues. It does not seem possible to give a clearcut relation between the type of molecular compound formed and the pKa values of acid and base (for summaries see Kortum et al., 1961 and Perrin (1965)). This is because of the competition between proton transfer and charge transfer and the temperature-dependence of these processes; also many pKa values are not known. In illustration we note that anilinium picrate and pyridinium picrates are salts (pKa of aniline, pyridine and picric acid are 4.51, 5.23 and 0.42 respectively). There are cations and anions in the crystal structure of pyridinium picrate; the original structure determination (Talukdar and Chaudhuri, 1976; PYRPIC) is wrong and has been corrected (Botoshansky et al., 1994; PYRPIC02). The picrates of o-chloroaniline (pKa ¼ 1.96), m-nitroaniline (pKa ¼ 2.46), 3-nitro-4-methyl-aniline (pKa ¼ 2.96) and N,N-dimethylamino-p-benzaldehyde (pKa ¼ 1.62) are yellow crystalline salts which give red solutions in benzene and red melts. Thus proton transfer occurs in the solid state but charge transfer interactions predominate in (nonpolar) solutions and in the melt (Saito and Matsunaga, 1971). Charge transfer molecular compounds are formed at room temperature when the bases involved are very weak, as in the picric acid compounds of carbazole, skatole and indole (Briegleb and Delle, 1960) and these are also found with bases such as o-nitroaniline (pKa ¼ 0.62) and p-nitroaniline (pKa ¼ 0.99); similarly p-dimethylaminobenzaldehyde (pKa ¼ 1.62) and o-nitroaniline give molecular compounds with 2,4,6-trinitrobenzoic acid

Table 15.16. Polymorphic transitions in crystals of some isomeric compounds. All transitions are from a salt-like structure at lower temperatures to a charge-transfer molecular compound stable above the transition temperature. pKa values are given in brackets Components

ttr( C)

Htr (kJ/mol)

A. Picric acid (0.42) as acceptor: 1. 2,5-dichloroaniline (1.57) (Matsunaga et al., 1974) 2. o-bromoaniline (2.55) (Komorowski et al., 1976) 3. o-iodoaniline (2.60) (Matsunaga et al., 1975) 4. 1-chloro-2-naphthylamine (Matsunaga et al., 1974) 5. 1-bromo-2-naphthylamine (Matsunaga et al., 1975) 6. 1,6-dibromo-2-naphthylamine (Matsunaga et al., 1975)

74 102 100–107 128–138 118–125 99–108

17.6 24.7 28.0 33.5 34.3 14.2

B. 2,4,6-trinitrobenzoic acid (0.65) as acceptor: 1. o-chloroaniline (2.65) (Matsunaga et al., 1974)

133

9.2

C. 2,6-dinitrophenol(3.57) as acceptor: 1. 4-bromo-1-naphthylamine (3.21) (Hertel and Frank, 1934) 2. 4-chloro-1-naphthylamine (Matsunaga et al., 1974)

monotropic 76


T (K)

1057

400 P+L

?

L+C

E1

c C

C+L

P1

b ?

d

S+L

350 P+C

P+?

S

P2

? D+L

S+? a

E2

S+D

300

BrA·(PiOH)2? BrA·PiOH (BrA)2·PiOH PiOH

D+B

50 mpc

BrA

Fig. 15.30. Phase diagram of the o-bromoaniline(BrA)–picric acid (PiOH or P) system. L = liquid; C = solid solution of charge transfer molecular compound type; S = solid solution of salt type; D = (BrA)2 PiOH; insert – the region of the solid state transition in expanded scale. (Reproduced from Komorowski et al., 1976.) n n n

(pKa ¼ 0.65) (Matsunaga and Osawa, 1974)). Results of studies of polymorphic transitions in isomeric compounds are summarized in Table 15.16. A complete study of a particular binary system that includes isomeric complexes does not appear to have been made, but some partial studies, extending beyond identification by spectroscopic techniques, have been reported. For example, the phase diagram of the o-bromoaniline–picric acid system has been determined in some detail (Fig. 15.30) (Komorowski et al., 1976). Equimolar {o-bromoaniline picric acid} is an isomeric complex with a salt , molecular compound transition temperature at 102 C while the 2 : 1 compound is of the CPT type discussed in the next section. The appearance of solid solutions in both polymorphs of the equimolar material is quite unexpected and should be checked. Thermodynamic studies have been made of 4-bromo-1-naphthylamine 2,6-dinitrophenol (Hertel and Schneider, 1931), where the yellow (stable; m.pt. 91.5 ; measured density 1.654 g cm3 ) and red (metastable; m.pt. 84.5 ; measured density 1.56 g cm3) phases are monotropically related (Hertel and Frank, 1934). The enthalpies of formation of the yellow and red phases were determined calorimetrically (temperature not specified but presumably 300K) to be 13.4 and 0.4 kJ/mol respectively. The yellow phase has thus both lower enthalpy and higher density than the red phase, as would be expected from the stronger (Coulombic) binding between the components. The crystal structure of the (metastable) red phase has been determined at 298K (space group P21/a, Z ¼ 4, stack n n n

n n n


axis [010]); the familiar mixed stacks were found and there is no hydrogen bonding between the components (Carsten-Oeser et al., 1968; PABRAN). 15.11.4

Isomerism stabilized by both charge (p–p*) and proton transfer (CPT compounds)

In these molecular compounds the charge and proton transfer can be carried out in a number of different ways and currently three types can be identified : (a)

CPT-A compounds: these are 1 : 1 (electron) donor–acceptor molecular compounds, where a proton is transferred from electron acceptor to donor, thus giving rise to the possibility of cation–anion and charge transfer interaction occurring in the same binary adduct; examples are tryptophan picrate (Matsunaga, 1973) and seretonin picrate (Thewalt and Bugg, 1972); see also Section 15.2. The red colours of both molecular compounds are due to charge transfer interactions between the cation donors and the picrate anions, acting as acceptors. This has been demonstrated directly for seretonin picrate by determination of the crystal structure of the monohydrate (P21/c, Z ¼ 4) that contains mixed stacks along [010] (cf. Section 15.2). H

H N

N O OH H2C

H 2C NH3+ Tryptophan cation

(b)

NH3+ Serotonin cation

CPT-B compounds: these are essentially 2 : 1 donor–acceptor molecular compounds, where one of the donor molecules acts as proton acceptor while the second participates in the charge transfer interaction with the anion, which acts both as an electron acceptor and a proton donor; examples are the black {(benzidine)2 picrate} (Saito and Matsunaga, 1973a), {(aniline)2–(2,4,6-trinitrobenzoic acid} (TNBA)) (Matsunaga, Osawa and Osawa, 1975) and the red {(o-aminobenzoic acid)2 picrate} (Matsunaga and Usui, 1980). More complicated compositions have been encountered in other systems (e.g. {(2,4,5-trimethylaniline)3(3,3 0 ,5,5 0 tetranitrobiphenyl-4,4 0 -diol (TNBP))2} (Matsunaga and Osawa, 1974) and the dark orange {(N,N-diethylaniline)5 (TNBP)3} (Saito and Matsunaga, 1973b; Lloyd and Sudborough, 1899); spectroscopic studies show that these are CPT compounds but crystal structure analyses are needed to show how the charge and proton transfers are distributed among the moieties. One can distinguish a subgroup in which these functions are combined in a single molecule but localized in different parts thereof; here one half of a monoprotonated (electron) donor acts as proton acceptor and the other half participates in chargetransfer interaction with the anion acceptor; examples are the picrates of o-dianisidine and of tetramethylbenzidine (Saito and Matsunaga, 1973a) and the 3 : 2

S ELF -C O MPLE XE S

1059

Fig. 15.31. ORTEP stereoview of pyridinium-1-naphthylamine–picrate crystal structure. The asymmetric unit has been marked. Only one orientation of the disordered 1-naphthylamine molecule is shown. (Reproduced from Bernstein et al., 1980.)

compound of benzidine (pKa ¼ 4.95) with 2,6-dinitrophenol (pKa ¼ 4.09) (Saito and Matsunaga, 1974). (c) CPT-C compounds : here the functions are localized in different molecules and thus a ternary composition can be expected. Ko¨fler’s (1940) ternary complex {pyridine– 1-naphthylamine–picric acid} {PY-NA-PC} was shown, by spectroscopy, to be a CPT compound (Matsunaga and Saito, 1972). Crystal structure analysis (Bernstein et al., 1980; PYNPCR) showed that it was in fact {pyridinium–1-naphthylamine– picrate}, in which there was a herringbone arrangement of centrosymmetric planeto-plane stacks of composition PYþ NA PC ( 1) PC NA PYþ (Fig. 15.31). This arrangement has similarities to that found in {(acridine)2 . . . PMDA} (Fig. 15.3). Thus the proton transfer is from picric acid to pyridine while the principal charge transfer is from 1-naphthylamine to picrate anion. The 1-naphthylamine molecule was disordered over (at least) two orientations.

15.12 Self-complexes In a molecule that has regions with electron-donor properties and other regions with electron-acceptor properties, intramolecular and/or intermolecular charge transfer can occur. If nitrogen and/or oxygen is present, then there can be n–* interaction. Alternatively, with appropriate molecular conformation, –* interactions could occur. If these interactions are intramolecular, then charge-transfer bands would be expected in both solution and solid state spectra. If the charge transfer interactions are intermolecular then one would expect a charge-transfer band in the spectrum of the solid but not (or to a far lesser extent) in solution, and a pairwise or stacked arrangement in the solid, with propinquity of donor and acceptor regions of different molecules. Many examples have been studied and all the possibilities outlined above have been encountered (Table 15.17). Classification can be conveniently made in terms of acceptor type and it will be noted that virtually all the well-known acceptor types are represented. The molecular arrangement in the red polymorph of 2-(4 0 -methoxyphenyl)-1,4-benzoquinone is shown in Fig.15.32 (Desiraju et al., 1977; PANQUO). Early and more recent work has been reviewed (Bleidelis, Shvets and Freimanis, 1976; Chitkina and Belskii, 2002).


(a)

D

A

A

D

D

A

1– x, 1/2+y, 1/2–z

(b)

O1 C6

C1

C8

C2

C9 C10

C5

O3

C7 C4

C3

C12

C13

C11

1 – x, –1/2 + y, 1/2– z

O2 x, y, z

Fig. 15.32. –* Interactions in the red polymorph of 2-(4 0 -methoxyphenyl)-1,4-benzoquinone. (a) stereopair diagram of the alternating donor–acceptor interaction in a molecular stack with the schematic arrangement shown in the center; (b) Details of pairwise overlap. The molecules are projected onto the plane of the central benzoquinone ring (line shaded), with the molecule above darkened and that below shown with broken lines. (Reproduced from Desiraju et al., 1977.)

A few structures that do not fit easily into the framework of Table 15.17 will be discussed separately. The crystal structure of 1-(2-indol-3-ylethyl)-3-carbamidopyridinium chloride monohydrate has been determined (tetragonal, space group P41) (Herriott et al., 1974; INYECP); it serves as an intramolecular model of the nicotinamide adenine dinucleotide – tryptophan charge transfer compound. In the crystal the molecules are in the fully extended transoid conformation, with the two planar portions (the donor and acceptor portions) of a particular molecule essentially mutually perpendicular. An indole ring of one molecule is stacked between nicotinamide rings of two other molecules, and conversely (Fig. 15.33; p. 1061). Thus there are two sets of mutually perpendicular donor–acceptor stacks in the crystal. Within a stack the donor and ˚. acceptor molecules are approximately parallel, with an interplanar distance of 3.7 A N

O Cl–· H2O

+ N Indole ring is donor portion H2C

CH2

NH2 Nicotinamide ring is acceptor portion.


1061

Table 15.17. Self-complexes classified according to acceptor type (a) Polynitrobenzene acceptor portions NO2

NO2 H

O2N

N

CH2

O2N

N

NO2

CH3

N(Et)2

OCH3

1. Donor portion of one molecule overlaps with acceptor of adjacent molecule to form a stack (Shvets et al., 1974; NETPHN10).

R1

2. Donor and acceptor portions are approximately perpendicular, with hinge at tertiary nitrogen. ---DA DA DA--- stacks are formed (Shvets et al., 1975a). D

A

D

A

A

D

A

D

N R2 H

3. R1 = N(CH3)2; R2 = NO2. Dark brown crystals (Nakai et al., 1976; two polymorphs MABZNA (P1, Z = 4) and NBZMAA (P21/c, Z = 4)).

4. As for #3 but with R1 = NO2; R2 = N(CH3)2. Orange crystals. Stacking as for #3 (Shvets et al., 1974).

(b) p-Benzoquinone acceptor portions H3CO

O

O X

CH3 N

5.

O O

6.

R

5. Red crystals; stacks with overlap of donor and acceptor rings, with -* interaction between the rings. There is no overlap of carbonyl groups and aromatic rings (Desiraju, Paul and Curtin, 1977; PANQUO).

X = CH3, R = H. 6. Adjacent molecules are stacked antiparallel about two fold screw axes, thus allowing D---A interactions along the stack (Prout and Castellano, 1970).

7. As for #6 but with X = Cl, R = COOC2H5; orange crystals. Stacking as in #6 (Shvets et al., 1975b).

8. As for #6 but with X = Cl, R = CH3; orange crystals. Stacking as in #6, but with additional Cl...O interaction, ˚ (Shvets et al., 1979; d(Cl...O) = 3.29 A MAMNAQ).


Table 15.17. (Continued ) O

O Ph

C2H5

O O 9.

OCH3

N

N O H3C

N

O

10.

CH3

10. Molecule folded back on itself to allow intramolecular donor–acceptor interaction (Shvets et al., 1975c; AEPCNQ10).

9. Black crystals with metallic lustre. Molecules are grouped in isolated donor-acceptor pairs (Shvets et al., 1978).

11. 2,3-Dichloronaphthazarin (DCDHNQ01) – see Section 15.6. (c) TCNE type acceptor (for review see Chetkina and Bel’skii, 2002). CN NC CN

N(R1R2)

12. R1 = R2 = CH3. Violet crystals. Molecules are stacked antiparallel with donor–acceptor interactions between adjacent molecules (Chetkina et al., 1976; TCYVMA). 13. R1 = Ph, R2 = H. Dark-violet crystals. Partial overlap of donor and acceptor portions of the molecules (Popova et al., 1978). (d) TCNQ type acceptor NC

14, 15. NR2

NC CN

14. R = n-propyl; dark green crystals with metallic lustre. Pairwise interactions between TCNQ portions (Povot’eva et al., 1981; BAFGOE). 15. R = CH3; black crystals. Pairwise interactions between donor and acceptor portions of molecule (Povot’eva et al., 1981).

NC – NC

16. N+

CH3

H CN

16. Picolyl-tricyanoquinodimethan. Zwitterionic structure; diamagnetic. Angle of 30 between planes of benzene and pyridinium rings. The pyridinium nitrogen stacks above the dicyanomethide portion of the next molecule along the b axis (Popova, Chetkina and Bespalov, 1981).

(e) Hexafluorobenzene type acceptor 17. Pentafluorobiphenyl, see Section 15.9.1. 18. Two polymorphic forms, both of which have stacking arrangements similar to those shown for #3 and 4 earlier in this Table. (Lindeman et al., 1981; BANGOM).


1063

Table 15.17. (Continued ) F

F HO

F

N F

F

(f) Various 19. 1,3-Indanedione (Bravic et al., 1976; INDDON). There are stacks of antiparallel molecules, thus allowing maximum interaction between donor and acceptor portions of the molecules and also maximum dipolar interaction. The crystals are tetragonal and the arrangement of the stacks resembles that found in {pyrene p-benzoquinone} (see Fig. 15.4). n n n

4.156

3.342

3.535

3.881

3.515

3.875

3.489

4.012

3.936

2.904

3.1 92

3.572

Carbon Nitrogen Oxygen Chlorine Hydrogen bond Intermolecular contact

94

2. 9

07

3.1

3.169

Fig. 15.33. Projection of part of crystal structure of 1-(2-indol-3-ylethyl)-3-carbamido-pyridinium chloride monohydrate onto (100); two mutually-perpendicular donor–acceptor stacks are shown. (Reproduced from Herriott et al., 1974.)


The -orbital overlap between adjacent donor and acceptor portions appears nearly optimal, and the permanent dipoles are strongly coupled. It was suggested (Herriott et al., 1974) that unsymmetrical –* and dipole–dipole interactions could lead to a specific three-dimensional mutual orientation, provided that one of the faces of the molecule is blocked, and that such a mechanism could be important in enzyme-coenzyme interactions. The molecule of 2-(2-pyridylmethyldithio)benzoic acid has a hinged conformation in the solid, with pyridyl and phenyl rings approximately parallel and arranged in stacks ˚ ) (Karle et al., 1969; PYSBAC). Presumably the pyridyl ring acts as along [001] (¼7.37 A electron donor and the carboxyl-substituted phenyl ring as acceptor, and there are both intramolecular and intermolecular charge transfer interactions. The hinged conformation with parallel donor and acceptor portions is a feature that appears in many of the compounds discussed in this section. The compound 3,3 0 -diacetyl-5,5 0 -diethoxycarbonylglaucyrone gives black crystals from benzene, whose structure shows a stepwise arrangement of conjugated molecules connected by self charge transfer interactions (Baker et al., 1980; ETGLAU). The right hand side of the molecule in the diagram is considered to behave as the electron–donor portion and the left hand side as the electron–acceptor portion. These two portions are superposed in alternating fashion in the crystal. CO2Et

O

O

H3C

O

CH3

O H 3,3'-diacetyl-5,5'-diethoxycarbonylglaucyrone

15.13 15.13.1

O

CO2Et

O

Conclusions Structural variety in –* molecular compounds

One sees, from the survey above, that there is considerable structural variety in this family of molecular compounds. The simple picture of parallel mixed stacks of donors and acceptors in alternating array, although applicable to the majority of –* molecular compounds, and also to the quasi-acceptor molecular compounds, needs considerable emendation and extension. Stacks of limited size are found, stack axes are not necessarily parallel, and components with suitable functional groups can give rise to important lateral interactions. In particular, hydrogen bonding between like and/or unlike components, added to the primary stacking structural synthon, can provide possibilities for crystal engineering similar to those described in Chapter 14. Although the most usual donor : acceptor ratio is 1 : 1, other compositions are found. These generally maintain mixed stacks with the component in excess accommodated in various ways, often as ‘solvent of crystallization’ filling space but not playing any structural role.

CONCLUSIONS

15.13.2

1065

How should the packing arrangements in p–p* molecular compounds be described?

Despite the considerable structural variety in this family of molecular compounds, the essential feature appearing in virtually all structures is a plane-to-plane interaction between the two components. For most of the molecular compounds considered here, this is a donor–acceptor interaction leading to color changes on formation, and physical properties that stem from the anisotropic arrangement. However, the same feature appears also in the structures containing quasi-acceptors rather than true acceptors; as there is no sensible structural difference between these two groups, they can be treated together. A natural consequence is to describe the structures in terms of mixed stacks containing donors and acceptors (or quasi-acceptors) in alternating array. Some authors have, however, preferred a mixed-layer to a mixed-stack description. The matter could be settled if one could compare, for a particular structure, the one-dimensional interaction energy within mixed stacks to the two-dimensional interaction within layers but this information is not available. Description is a matter of choice, and we use three examples, from the extremes and the center, to illustrate the dilemmas. The interaction of a benzene ring (a polarizable donor group) with a polar acceptor group (e.g. a carbonyl group) often leads to a characteristic superposition (overlap diagram) within a mixed stack. An example is (monoclinic) {perylene fluoranil} (Hanson, 1963; PERFAN; Section 15.6 and Fig. 15.10), which has a typical herring-bone structure also shown by (monoclinic) {perylene 1,2–4,5-tetracyanobenzene} (Fig. 15.20; Bock, Seitz et al., 1996; REHMUM). How could one visualize the growth of such a crystal in formal (not necessarily physical) terms? Perhaps as preformed quasi-cylindrical stacks, with a given mutual donor-acceptor disposition, that then aggregate into a quasi close-packed arrangement of the quasi-cylindrical stacks. The mutual orientation of the stacks is determined by secondary cross-stack interactions; in the particular example of {perylene . . . fluoranil} these could be weak C–H . . . O¼C hydrogen bonds. At the other extreme one has a hydrogen-bonded layer such as that found in {hexamethylbenzene 1,3,5-tricyanobenzene} (LAGNAI) (Fig. 15.34; Reddy, Goud et al., 1993; Desiraju and Steiner, 1999; see p. 327 et seq.); each layer contains only a single component. Here the layers would be formed first and then stacked one over the other to give optimal plane-to-plane overlap of hexamethylbenzene donor and tricyanobenzene acceptor. {Pyrene 1,2–4,5-tetracyanohydroquinone} (Fig. 15.20; Bock, Seitz et al., 1996; TEXPOB10) provides a similar example, again with the layers each containing a single component and the –OH . . . NC–C hydrogen bonding in the tetracyanohydroquinone layer providing the dominant structural interaction. What happens in the middle? One example is {dibenz[a,c]anthracene -1,3,5trinitrobenzene} ({DBA TNB}; Carrell and Glusker, 1997; RULLUF). Here the layers, which are somewhat corrugated, contain both components, with molecules of the smaller TNB acceptors surrounded by molecules of the larger DBA donors (Fig. 15.35). Carrell and Glusker provide a well-balanced description of the structure as having ‘‘. . . layers containing both DBA and TNB molecules, interconnected within a layer by C–H . . . O interactions. Layers stack on one another so that DBA molecules are sandwiched between TNB molecules and vice versa. The average distance between molecules in these sand˚ ’’ (italics added). Are the (relatively) many weak hydrogen bonds more wiches is 3.23 A n n n

n n n

n n n

n n n

n n n

n n n


Fig. 15.34. The crystal structure of {hexamethylbenzene 1,3,5-tricyanobenzene} (15.207 8.839 ˚ , 110.48¯, Z = 4, C2/c; the HMB molecules are at inversion centers and the 1,3,5-tricyano14.460 A benzene molecules on two fold axes). The overlapped layers of hexamethylbenzene (darkened, below) and the hexagonal 1,3,5-tricyanobenzene network mediated by weak C–H . . . N hydrogen ˚) bonds (above) are shown. The weak C–H . . . NC–C hydrogen bonds (d(C . . . N) = 3.47, 3.52 A are indicated by dashed lines. The layers are parallel to (110). (Reproduced from Desiraju and Steiner, 1999.) n n n

important in determining the overall structure than the plane-to-plane interaction? At present there does not seem to be an answer to this question. However, stacking appears to be a feature occurring in all the structures of this family whereas layers can only be identified when there are appreciable lateral interactions. Thus, in our view, ‘stacking’ is the primary feature with ‘layering’ a possibly important secondary attribute. 15.13.3

Structural consequences of p–p* interactions

It is widely recognized that the donor–acceptor interactions in –* molecular compounds are too weak to cause changes in bond lengths that are measurable at current levels of precision. This applies even to measurements made at very low temperatures (e.g. {pyrene PMDA} at 19K (Herbstein, Marsh and Samson, 1994; PYRPMA04). n n n

CONCLUSIONS

1067

2.36 2.40

2.53 2.57

2.48

2.53 2.51 2.46

layer

TNB

DBA

Fig. 15.35. Structural features of the triclinic DBA TNB crystals (a ¼ 7.277(2), b ¼ 11.237(6), ˚ , = 104.13(4), = 96.04(3), = 95.15(2) , space group , Z = 2. (above) One layer c ¼ 13.902(5) A of the structure showing TNB surrounded by DBA molecules and linked by CH O interactions. (below) Stacking of layers showing that DBA and TNB are not entirely coplananr. (Reproduced from Desiraju and Steiner, 1999.) n n n

n n n

Charge density studies (Chap. 17.7), comparing neat components with such components in the molecular compounds, may provide evidence of interaction but this remains a task for the future. Deviations from planarity in flexible molecules are more easily accessible. A clearly discernible effect has been found in {9,10-dihydroanthracene TNB}, where the DHA molecule is folded (dihedral angle 146 ) in its neat crystals (DITBOX) but planar in the molecular compound (ZZZAGS10) (Herbstein, Kapon and Reisner, 1986). Appreciable differences in the shapes of components in their neat crystals and in their molecular compounds has been found in {triphenylene–perfluorotriphenylene} (Weck et al., 1999; CUKXIP; Fig. 15.24) and, to a lesser extent, for (nonplanar) benzo[c] phenanthrene in its neat crystals (Lakshman et al., 2000; BZPHAN01) and its DDQ molecular compound (Bernstein, Herbstein and Regev, 1977; BZPCPQ). n n n


References Adman, E., Rosenblum, M., Sullivan, S. and Margulis, T. N. (1967). J. Am. Chem. Soc., 89, 4540–4542. Agostini, G., Corvaja, C., Giacometti, G., Pasimeni, L., Clement, D. A. and Bandoli, G. (1986). Mol. Cryst. Liq. Cryst., 141, 165–178. Akhtar, S., Tanaka, J., Nakasuji, K. and Murata, I. (1985). Bull. Chem. Soc. Jpn., 58, 2279–2284. Alcacer, L. and Maki, A. H. (1974). J. Phys. Chem., 78, 215–217. Allen, C. C., Boeyens, J. C. A. and Levendis, D. C. (1989). S. Afr. J. Chem., 42, 38–42. Andersen, A. M., Mostad, A. and Rømming, C. (1975). Acta Chem. Scand., B29, 45–50. Anderson, J. S. (1937). Nature, Lond., 140, 583–584. Anex, B. G. and Parkhurst, L. J. (1963). J. Am. Chem. Soc., 85, 3301–3302. Anthonj, R., Karl, N., Robertson, B. E. and Stezowski, J. J. (1980). J. Chem. Phys., 72, 1244–1255. Aroney, M. J., Hambley, T. W., Patsilades, E., Piernes, R. K., Chan, M.-K. and Gonda, I. (1987). J. Chem. Soc., Perkin Trans. 2, pp. 1747–1752. Artiga, A., Gaultier, J., Haget, Y. and Chanh, N. B. (1978a). J. Chim. Phys., 75, 379–383. Artiga, A., Gaultier, J., Hauw, C. and Chanh, N. B. (1978b). Acta Cryst., B34, 1212–1218. Ashwell, G. J., Kennedy, D. A. and Nowell, I. W. (1983). Acta Cryst., C39, 733–734. Aspley, C. J., Boxwell, C., Buil, M. L., Higgitt, C. L., Long, C. and Perutz, R. N. (1999). Chem. Commun., 1027–1028. Bailey, A. S. and Prout, C. K. (1965). J. Chem. Soc., pp. 4867–4881. Baker, S. R., Begley, M. J. and Crombie, L. (1980). J. Chem. Soc. Chem. Commun., pp. 390–392. Baldwin, S. L. and Baughman, M. C. (1993). Acta Cryst., C49, 1840–1844. Bando, M. C. and Matsunaga, Y. (1976). Bull. Chem. Soc. Jpn., 49, 3345–3346. Banerjee, A. and Brown, C. J. (1985). Acta Cryst., C41, 82–84. Bar, I. and Bernstein, J. (1978). Acta Cryst., B34, 3438–3441. Bar, I. and Bernstein, J. (1981). Acta Cryst., B37, 569–575. Barnes, J. C., Chudek, J. A., Foster, R., Jarrett, F., Mackie, F., Paton, F. and Twiselton, D. R. (1984). Tetrahedron, 40, 1595–1601. Bars-Combe, M. le, Chion, B. and Lajzerowicz-Bonneteau, Janine. (1979). Acta Cryst., B35, 913–920. Bastiansen, O., Gunderson, S. and Samdal, S. (1989). Acta Chem. Scand., 43, 6–10. Batail, P., La Placa, S. J., Mayerle, J. J. and Torrance, J. B. (1981). J. Am. Chem. Soc., 103, 951–953. Batsanov, A. S. and Perepichka, I. F. (2003). Acta Cryst., E59, o1318–o1320. Batsanov, A. S., Collings, J. C., Howard, J. A. K. and Marder, T. B. (2001). Acta Cryst., C57, 1306–1307. Batsanov, A. S., Collings, J. C., Howard, J. A. K., Marder, T. B. and Perepichka, I. F. (2001). Acta Cryst., E57, o950–o952. Batsanov, A. S., Howard, J. A. K., Marder, T. B. and Robins, E. G. (2001). Acta Cryst., C57, 1303–1305. Batsanov, A. S., Perepichka, I. F., Bryce, M. R. and Howard, J. A. K., (2001). Acta Cryst., C57, 1299–1302. Bear, C. A., Waters, J. M. and Waters, T. N. (1973). J. Chem. Soc. Perkin Trans.2, pp. 1266–1271. Beaumont, T. G. and Davis, K. M. C. (1967). J. Chem. Soc. (B), pp. 1131–1134. Bechgaard, K., Kistenmacher, T. J., Bloch, A. N. and Cowan, D. O. (1977). Acta Cryst., B33, 417–422. Bechtel, F., Chasseau, D. and Gaultier, J. (1976). Cryst. Struct. Commun., 5, 297–300. Bechtel, F., Chasseau, D., Gaultier, J. and Hauw, C. (1977). Cryst. Struct. Commun., 6, 699–702.

REFERENCES

1069

Beck, C. M., Burdeniuc, J., Crabtree, R. H., Rheingold, A. L. and Yap, G. P. A. (1998). Inorg. Chim. Acta, 270, 559–562. Beevers, C. A. and Lipson, H. (1934). Proc. Roy. Soc. Lond., A146, 570–582. Bergamini, P., Bertolasi, V., Ferretti, V. and Sostero, S. (1987). Inorg. Chim. Acta, 126, 151–155. Bernstein, J. and Regev, H. (1980). Cryst. Struct. Commun., 9, 581–586. Bernstein, J. and Trueblood, K. N. (1971). Acta Cryst., B27, 2078–2089. Bernstein, J., Cohen, M. D. and Leiserowitz, L. (1974). ‘‘The Structural Chemistry of the Quinones’’, in The Chemistry of the Quinonoid Compounds, S. Patai, Editor, J. Wiley and Sons, New York. Bernstein, J., Herbstein, F. H. and Regev, H. (1977). Acta Cryst., B33, 1716–1724. Bernstein, J., Regev, H. and Herbstein, F. H. (1980). Acta Cryst., B36, 1170–1175. Bernstein, J., Regev, H., Herbstein, F. H., Main, P., Rizvi, S. H., Sasvari, K. and Turcsanyi, B. (1976). Proc. Roy. Soc. Lond., A347, 419–434. Bertholon, G., Perrin, R., Lamartine, R., Thozet, A., Perrin, M., Caillet, J. and Claverie, P. (1979). Mol. Cryst. Liq. Cryst., 52, 589–596. Bertinelli, F., Costa Bizzarri, F., Della Casa, C., Marchesini, A., Pelizzi, G., Zamboni, R. and Taliani, C. (1984). Mol. Cryst. Liq. Cryst., 109, 289–302. Binder, W., Karl, N. and Stezowski, J. J. (1982). Acta Cryst., B38, 2915–2916. Birtle, S. L. and Naae, D. G. (1980). ACA, Ser., 2, 7, 10. Bleidelis, J., Shvets, A. E. and Freimanis, J. (1976). Zh. Strukt. Khim., 17, 1096–1110. Bock, H., Rauschenbach, A., Nather, C., Kleine, M. and Bats, J. W. (1996). Phosphorus, sulfur, silicon, related elements, 115, 51–83. Bock. H., Seitz, W., Sievert, M., Kleine, M. and Bats, J. W. (1996). Liebigs Ann., 1929–1940. Boer, J. L. de, and Vos, A. (1968). Acta Cryst., 24, 720–725. Boeyens, J. C. A. and Herbstein, F. H. (1965a). J. Phys. Chem., 69, 2153–2159. Boeyens, J. C. A. and Herbstein, F. H. (1965b). J. Phys. Chem., 69, 2160–2176. Botoshansky, M., Herbstein, F. H. and Kapon, M. (1994). Acta Cryst., B50, 191–200. Bravic, G., Bechtel, F., Gaultier, J. and Hauw, C. (1976). Cryst. Struct. Commun., 5, 1–4. Brennan, J. S., Brown, N. M. D. and Swinton, F. L (1974). J. Chem. Soc. Farad. Trans. I, 70, 1965–1970. Briegleb, G. and Delle, H. (1960). Z. Elektrochem., 64, 347–355. Brierley, C., Barton, R., Robertson, B. E. and Karl, N. (1982). ACA Summer Meeting, La Jolla, Abstract PB–7. Brock, C. P., Naae, D. G., Goodland, N. and Hamor, T. A. (1978). Acta Cryst., B34, 3691–3696. Brown, D. S. and Wallwork, S. C. (1965). Acta Cryst., 19, 149. Brown, D., Wallwork, S. C. and Wilson, A. (1964). Acta Cryst., 17, 168–176. Brown, J. N., Cheung, L. D., Trefonas, L. M. and Majeste, R. J. (1974). J. Cryst. Mol. Struct., 4, 361–374. Bruce, M. I., Snow, M. R. and Tiekink, E. R. (1987) Acta Cryst. C43, 1640–1641. Bryce, M. R. and Davies, S. R. (1987). Synth. Metals, 20, 373–374. Bryce, M. R., Davies, S. R., Hursthouse, M. R. and Motevalli, M. (1988). J. Chem. Soc., Perkin Trans. II, pp. 1713–1716. Bryce, M. R., Secco, A. S., Trotter, J. and Weiler, L. (1982). Can. J. Chem., 60, 2057–2061. Bulgarovskaya, I. V. and Zvonkova, Z. V. (1976). Sov. Phys. Cryst., 21, 335–337; Kristallografiya, 21, 597–599. Bulgarovskaya, I. V., Smelyanskaya, E. M., Federov, Yu. G. and Zvonkova, Z. V. (1977). Sov. Phys. Cryst., 22, 104–106; Kristallografiya, 22, 184–187. Bulgarovskaya, I. V., Smelyanskoya, E. M., Federov, Yu. G. and Zvonkova, Z. V. (1974). Sov. Phys. Cryst., 19, 157–160; Kristallografiya, 19, 260–265. Bulgarovskaya, I. V., Belsky, V. K. and Vozzhennikov, V. M. (1987). Acta Cryst., C43, 768–770.


Bulgarovskaya, I. V., Vozzhennikov, V. M., Krasavin, V. P. and Kotov, B. V. (1982). Cryst. Struct. Commun., 11, 501–504. Bulgarovskaya, I. V., Zavodnik, V. E. and Vozzhennikhov, V. M. (1987a). Acta Cryst., C43, 764–766. Bulgarovskaya, I. V., Zavodnik, V. E. and Vozzhennikov, V. M. (1987b). Acta Cryst., C43, 766–768. Bulgarovskaya, I. V., Zavodnik, V. E., Bel’skii, V. K. and Vozzhennikov, V. M. (1989). Sov. Phys. Cryst., 34, 203–207; Kristallografiya, 34, 345–352. Bulgarovskaya, I. V., Zvonkova, Z. V. and Kolninov, O. V. (1978). Sov. Phys. Cryst., 23, 665–669; Kristallografiya, 23, 1175–1182. Bunz, U. H. F. and Enkelmann, V. (1999). Chem. Eur. J., 5, 263–266. Burdeniuc, J., Crabtree, R. H., Rheingold, A. L. and Yap, G. P. A. (1997). Bull. Soc. Chim. France, 134, 955–958. Butler, M. A., Wudl, F. and Soos, Z. G. (1975). Phys. Rev., B12, 4708–4719. Carrell, H. L. and Glusker, J. P. (1997). Struct. Chem., 8, 141–147. Carstensen-Oeser, E., Go¨ttlicher, S. and Habermehl, G. (1968). Chem. Ber., 101, 1648–1655. Carter, O. L., McPhail, A. T. and Sim, G. A. (1966). J. Chem. Soc. (A), pp. 822–838. Cassoux, P. and Gleizes, A. (1980). Inorg. Chem., 19, 665–672. Catellani, M. and Porzio, W. (1991). Acta Cryst., C47, 596–599. Chantooni, M. K., Jr. and Britton, D. (1998). J. Chem. Cryst., 28, 329–333. Chasseau, D. and Hauw, C. (1980). Acta Cryst., B36, 3131–3133. Chasseau, D. and Leroy, F. (1981). Acta Cryst., B37, 454–456. Chasseau, D., Gaultier, J. and Hauw, C. (1972). Compt. rend. Acad. Sci., Paris, Ser. C, 274, 1434–1437. Chasseau, D., Gaultier, J., Fabre, J. M. and Giral, L. (1982). Acta Cryst., B38, 1632–1635. Cherin, P. and Burack, M. (1966). J. Phys. Chem., 70, 1470–1472. Chetkina, L. A. and Bel’skii, V. K. (2002). Cryst. Reps., 47, 581–602. Chetkina, L. A., Popova, E. G., Kotov, B. V., Ginzburg, S. L. and Smelyanskaya, E. M. (1976). Zh. Strukt. Khim., 17, 1060–1066. Clyburne, J. A. C., Hamilton, T. and Jenkins, H. A. (2001). Cryst. Engin., 4, 1–9. Coates, G. W., Dunn, A. R., Henling, L. M., Dougherty, D. A. and Grubbs, R. H. (1997). Angew. Chem. Int. Ed. Engl., 36, 248–251. Cohen-Addad, C., Consigny, M., D’Assenza, G. and Baret, P. (1988). Acta Cryst., C44, 1924–1926. Cohen-Addad, C., Lebars, M., Renault, A. and Baret, P. (1984). Acta Cryst., C40, 1927–1931. Cohen-Addad, C., Renault, A., Communandeur, G. and Baret, P. (1988). Acta Cryst., C44, 914–916. Collings, J. C., Batsanov, A. S., Howard, J. A. K. and Marder, T. B. (2001). Acta Cryst., C57, 870–872. Collings, J. C., Roscoe, K. P., Thomas, R. Ll., Batsanaov, A. S., Stimson, L. M., Howard, J. A. K. and Marder, T. B. (2001). New J. Chem., 25, 1410–1417. Couldwell, M. C. and Prout, C. K. (1978). J. Chem. Soc. Perkin II, pp. 160–164. Dahl, T. (1971a). Acta Chem. Scand., 25, 1031–1039. Dahl, T. (1971b). Acta Chem. Scand., 26, 1569–1575. Dahl, T. (1973). Acta Chem. Scand., 27, 995–1003. Dahl, T. (1975a). Acta Chem. Scand., A29, 170–174. Dahl, T. (1975b). Acta Chem. Scand., A29, 699–705. Dahl, T. (1979). Acta Chem. Scand., A33, 665–669. Dahl, T. (1981a). Acta Cryst., B37, 98–101. Dahl, T. (1981b). Acta Chem. Scand., A35, 701–705. Dahl, T. (1988). Acta Chem. Scand., A42, 1–7. Dahl, T. (1990). Acta Cryst., B46, 283–288. Dahl, T. and Sørensen, B. (1985). Acta Chem. Scand., B39, 423–428.

REFERENCES

1071

Dai, C., Nguyen, P., Marder, T. B., Scott, A. I., Clegg, W. and Viney, C. (1999). Chem. Commun. 2493–2494. Damiani, A., De Santis, P., Giglio, E., Liquori, A. M. and Ripamonti, A. (1965). Acta Cryst., 19, 340–348. Damiani, A., Giglio, E., Liquori, A. M. and Ripamonti, A. (1967). Acta Cryst., 23, 675–681. Damiani, A., Giglio, E., Liquori, A. M., Puliti, R. and Ripamonti, A. (1967). J. Mol. Biol., 23, 113–115. Danno, T., Kajiwara, T. and Inokuchi, H. (1967). Bull. Chem. Soc. Jpn., 40, 2793–2795. Darocha, B. F., Titus, D. D. and Sandman, D. J. (1979). Acta Cryst., B35, 2445–2448. De, R. L., Seyerl, J. von, Zsolnai, L. and Huttner, G. (1979). J. Organometall. Chem., 175, 185–191. Desgreniers, S., Kourouklis, G. A., Jayaraman, A., Kaplan, M. L. and Schmitt, P. H. (1985). J. Chem. Phys., 83, 480–485. Desiraju, G. D., Paul, I. C. and Curtin, D.Y. (1977). J. Am. Chem. Soc., 99, 1594–1601. Desiraju, G. R., Curtin, D. Y. and Paul, I. C. (1979). Mol. Cryst. Liq. Cryst., 52, 259–266. Desiraju, G. R. and Steiner, T. (1999). The Weak Hydrogen Bond in Structural Chemistry and Biology, IUCr Monographs on Crystallography, Oxford University Press. Dietz, K., Endres, H., Keller, H. J., Moroni, W. and Wehe, D. (1982). Z. Naturforsch., 37b, 437–442. Doherty, R. M., Stewart, J. M., Mighell, A. D., Hubbard, C. R. and Patiadi, A. J. (1982). Acta Cryst., B38, 859–863. Dong, V., Endres, H., Keller, H. J., Moroni, W. and No¨the, D. (1977). Acta Cryst., B33, 2428–2431. Duncan, W. A. and Swinton, F. L. (1966). Trans. Farad. Soc., 62, 1082–1089. Dung, N.-H., Viossat, B., Lancelot, J.-C. and Robba, M. (1986). Acta Cryst., C42, 843–847. Dunitz, J. D. (1996). Persp. Supramol. Chem., 2, 1–30. D’yachenko, O. A., Atovmyan, L. O., Kovalev, A. A. and Soboleva, S. V. (1977). J. Struct. Chem., 18, 713–719; Zh. Strukt. Khim., 18, 898–907. Dzyabchenko, A. V., Bulgarovskaya, I. V., Zavodnik, V. E. and Stash, A. I. (1994). Kristallografiya, 39, 434–438. Dzyabchenko, A. V. and Zavodnik, V. E. (1984). Zh. Strukt. Khim., 25, 177–179. Emge, T. J., Bryden, W. A., Cowan, D. O. and Kistenmacher, T. J. (1982). Mol. Cryst. Liq. Cryst., 90, 173–184. Emge, T. J., Wijgul, F. M., Chappell, J. S., Bloch, A. N., Ferraris, J. P., Cowan, D. O. and Kistenmacher, T. J. (1982). Mol. Cryst. Liq. Cryst., 87, 137–161. Ermer, O. (1988). J. Am. Chem. Soc., 110, 3747–3754. Evans, D. L. and Robinson, W. T. (1977). Acta Cryst., B33, 2891–2893. Fenby, D. V. (1972). Rev. Pure Appl. Chem., 22, 55–65. Ferguson, G., Mackay, I. R., Pollard, D. R. and Robertson, J. M. (1969). Acta Cryst., A25, S132. Flandrois, S., Ludolf, K., Keller, H. J., No¨the, D., Bondeson, S. R., Soos, Z. G. and Wehe, D. (1983). Mol. Cryst. Liq. Cryst., 95, 149–164. Fleischmann, D. E. and Tollin, G. (1965a). Proc. Nat Acad. Sci., 53, 38–46. Fleischmann, D. E. and Tollin, G. (1965b). Proc. Nat Acad. Sci., 53, 237–242. Fleischmann, D. E. and Tollin, G. (1965c). Biochim. Biophys. Acta, 94, 248–257. Foss, L. I., Syed, A., Stevens, E. D. and Klein, C. L. (1984). Acta Cryst., C40, 272–274. Foster, R., Iball, J., Scrimgeour, S. N. and Williams, B. C. (1976). J. Chem. Soc. Perkin 2, pp. 682–685. Frankenbach, G. M., Beno, M. A. and Williams, J. M. (1991). Acta Cryst., C47, 762–764. Fritchie, C. J., Jr. (1966). Acta Cryst., 20, 892–898. Fritchie, C.J., Jr. and Arthur, P., Jr. (1966). Acta Cryst., 21, 139–145. Fritchie, C. J., Jr. and Johnston, R. M. (1975). Acta Cryst., B31, 454–461. Fritchie, C. J., Jr. and Trus, B. L. (1968). J. Chem. Soc. Chem. Commun., pp. 833–834.


Fujii, S., Kawasaki, K., Sato, A., Fujiwara, T. and Tomita, K.-I. (1977). Arch. Biochem. Biophys., 181, 363–370. Fujinawa, T., Goto, H., Naito, T., Inabe, T., Akutagawa, T. and Nakamura, T. (1999). Bull. Chem. Soc. Jpn., 72, 21–26. Galigne, T. L., Fabre, J. M. and Giral, L. (1977). Acta Cryst., B33, 3827–3831. Gartland, G. L., Freeman, G. R. and Bugg, C. E. (1974). Acta Cryst., B30, 1841–1849. Gaultier, J., Hauw, C. and Breton-Lacombe, M. (1969). Acta Cryst., B25, 231–237. Gaultier, J. and Hauw, C. (1965). Acta Cryst., 18, 179–183. Gebert, E., Reis, A. H., Jr., Miller, J. S., Rommelman, H. and Epstein, A. J. (1982). J. Am. Chem. Soc., 104, 4403–4410. Geiger, W. E., Jr. and Maki, A. H. (1971). J. Phys. Chem., 75, 2387–2394. Gieren, A., Lamm, V., Hu¨bner, T., Rabben, M., Neidlein, R. and Droste, D. (1984). Chem. Ber., 117, 1940–1953. Gleason, W. B. and Britton, D. (1976). Cryst. Struct. Commun., 5, 483–488. Goldberg, I. and Shmueli, U. (1973a). Acta Cryst., B29, 421–431. Goldberg, I. and Shmueli, U. (1973b). Acta Cryst., B29, 432–440. Goldberg, I. and Shmueli, U. (1973c). Acta Cryst., B29, 440–448. Goldberg, I. and Shmueli, U. (1977). Acta Cryst., B33, 2189–2197. Griffith, G., Jackson, P. R., Kenyon-Blair, E. and Morcom, K. W. (1983). J. Chem. Thermodynam., 15, 1001–1002. Grigg, R., Trocha-Grimshaw, J. and King, T. J. (1978). J. Chem. Soc. Chem. Commun., pp. 571–572. Groziak, M. P., Wilson, S. R., Clauson, G. R. and Leonard, N. J. (1986). J. Am. Chem. Soc., 108, 8002–8006. Gundel, D., Sixl, H., Metzger, R. M., Heimer, N. E., Harms, R. H., Keller, H. J., No¨the, D. and Wehe, D. (1983). J. Chem. Phys., 79, 3678–3688. Hall, B. and Devlin, J. P. (1967). J. Phys. Chem., 71, 465–466. Hansmann, C., Foro, S., Lindner, H. J. and Fuess, H. (1997a). Z. Kristallogr., 212, 79–80. Hansmann, C., Foro, S., Lindner, H. J. and Fuess, H. (1997b). Z. Kristallogr., 212, 81–82. Hanson, A. W. (1963). Acta Cryst., 16, 1147–1151. Hanson, A. W. (1964). Acta Cryst., 17, 559–568. Hanson, A. W. (1965a). Acta Cryst., 19, 19–26. Hanson, A. W. (1965b). Acta Cryst., 19, 610–613. Hanson, A. W. (1966). Acta Cryst., 20, 97–102. Hanson, A. W. (1978). Acta Cryst., B34, 2339–2341. Harata, K., Aono, T., Sakabe, N. and Tanaka, J. (1972). Acta Cryst., A28, S14. Harms, R. H., Keller, H. J., No¨the, D., Werner, M., Gundl, D., Sixl, H., Soos, Z. G. and Metzger, R. M. (1981). Mol. Cryst. Liq. Cryst., 65, 179–196. Hazell, A. C. (1978). Acta Cryst., B34, 466–471. Herbstein, F. H. (1971). ‘‘Crystalline -Molecular Compounds : Chemistry, Spectroscopy, and Crystallography’’, in Perspectives in Structural Chemistry, edited by J. D. Dunitz and J. A. Ibers, Wiley, New York, Vol. IV, pp. 166–395. Herbstein, F. H. (2001). ‘‘Varieties of polymorphism.’’ Advanced Methods in Strucutre Analysis, pp. 114–154. Eds. R. Kuzel and J. Hasak, CSCA. http : //www.xray/ecm/book/. Herbstein, F. H. and Kaftory, M. (1975a). Acta Cryst., B31, 60–67. Herbstein, F. H. and Kaftory, M. (1975b). Acta Cryst., B31, 68–75. Herbstein, F. H. and Kaftory, M. (1976). Acta Cryst., B32, 387–396. Herbstein, F. H. and Reisner, G. M. (1984). Acta Cryst., C40, 202–204. Herbstein, F. H. and Samson, S. (1994). Acta Cryst., B50, 182–191. Herbstein, F. H. and Snyman, J. A. (1969). Phil. Trans. Roy. Soc. Lond., A264, 635–666. Herbstein, F. H., Kaftory, M. and Regev, H. (1976). J. Appl. Cryst., 9, 361–364.

REFERENCES

1073

Herbstein, F. H., Kapon, M. and Reisner, G. M. (1986). Acta Cryst., B42, 181–187. Herbstein, F. H., Kapon, M., Rzonzew, G. and Rabinovich, D. (1978). Acta Cryst., B34, 476–481. Herbstein, F. H., Marsh, R. E. and Samson, S. (1994). Acta Cryst., B50, 174–181. Herriott, J. R., Camerman, A. and Deranleau, D. A. (1974). J. Am. Chem. Soc., 96, 1585–1589. Hertel, E. and Bergk, H. W. (1936). Z. Phys. Chem., B33, 319–333. Hertel, E. and Frank, H. (1934). Z. Physik. Chem., B27, 460–466. Hertel, E. and Schneider, K. (1931). Z. Physik. Chem., B13, 387–399. Hill, H. A. O., MacFarlane, A. J. and Williams, R. J. P. (1967). J. Chem. Soc. Chem. Commun. p. 605. Hinrichs, W. and Klar, G. (1982). J. Chem. Res. (S), pp. 336–337. Hoffman, S. K., Corvan, P. J., Singh, P, Sethulekshmi, C. N., Metzger, R. M. and Hatfield, W. E. (1983). J. Am. Chem. Soc., 105, 4608–4617. Hoier, H., Zacharias, D. E., Carrell, H. and Glusker, J. P. (1993). Acta Cryst., C49, 523–526. Hughes, R. C. and Soos, Z. G. (1968). J. Chem. Phys., 48, 1066–1076. Hursthouse, M. B., Smith, V. B. and Massey, A. G. (1977). J. Fluor. Chem., 10, 145–156. Ikemoto, I., Chikaishi, K., Yakushi, K. and Kuroda, H. (1972). Acta Cryst., B28, 3502–3506. Ikemoto, I., Yakushi, K. and Kuroda, H. (1970). Acta Cryst., B26, 800–806. Irngartinger, H. and Goldmann, A. (1978). Z. Kristallogr., 149, 97. Ito, T., Minobe, M. and Sakurai, T. (1970). Acta Cryst., B26, 1145–1151. Iwasaki, F. and Saito, Y. (1970). Acta Cryst., B26, 251–260. Iwasaki, F., Hironaka, S., Yamazaki, N. and Kobayashi, K. (1992). Bull. Chem. Soc. Jpn., 65, 2180–2186. Jackson, P. R. and Morcom, K. W. (1986). J. Chem. Thermodynam., 18, 75–80. Jankowski, W. and Gdaniec, M. (2002). Acta Cryst., C58, o181–o182. Jones, N. B. and Marsh, R. E. (1962). Acta Cryst., 15, 809–810. Kabuto, C., Suzuki, T., Yamashita, Y. and Mukai, T. (1986). Chem. Letts., pp. 1433–1436. Kamenar, B. and Prout, C. K. (1965). J. Chem. Soc., pp. 4838–4851. Kamenar, B., Prout, C. K. and Wright, J. D. (1965). J. Chem. Soc., pp. 4851–4867. Kamenar, B., Prout, C. K. and Wright, J. D. (1966). J. Chem. Soc., (A) pp. 661–664. Kaminskii, V. F., Shibaeva, R. P., Aldoshina, M. Z., Lyubovskaya, R. N. and Khidekel, M. L. (1979). J. Struct. Chem., 20, 130–133; Zh. Strukt. Khim., 20, 157–160. Karl, N., Binder, W., Kollet, P. and Stezowski, J. J. (1982). Acta Cryst., B38, 2919–2921. Karl, N., Ketterer, W. and Stezowski, J. J. (1982). Acta Cryst., B38, 2917–2919. Karle, J., Karle, I. L. and Mitchell, D. (1969). Acta Cryst., B25, 866–871. Karlsson, R. (1972). Acta Cryst., B28, 2358–2364. Kasper, J. S. and Interrante, L. V. (1976). Acta Cryst., B32, 2914–2916. Keller, H. J. and Soos, Z. G. (1985). Top. Curr. Chem., 127, 169–216. Keller, H. J., Leichert, I., Megnamisi-Belombe, M., Nothe, D. and Weiss, J. (1977). Z. anorg. allgem. Chem., 429, 231–236. Keller, H. J., Moroni, W., No¨the, D., Scherz, M. and Weiss, J. (1978). Z. Naturforsch., 33b, 838–842. Kemmer, T., Sheldrick, W. S. and Brockmann, H. (1976). Angew. Chem. Int. Ed. Engl., 15, 115. Kirner, J. F., Reed, C. A. and Scheidt, W. R. (1977). J. Am. Chem. Soc., 99, 1093–1101. Kistenmacher, T. J., Emge, T. J., Bloch, A. N. and Cowan, D. O. (1982). Acta Cryst., B38, 1193–1199. Kistenmacher, T. J., Phillips, T. E., Cowan, D. O., Ferraris, J. P., Bloch, A. N. and Poehler, T. O. (1976). Acta Cryst., B32, 539–547. Kobayashi, H. and Nakayama, J. (1981). Bull. Chem. Soc. Jpn., 54, 2408–2411. Kobayashi, H. (1973). Bull. Chem. Soc. Jpn., 46, 2945–2949. Kobayashi, H. (1987). Synth. Metals, 19, 475–480. Kodama, T. and Kumakura, S. (1974a). Bull. Chem. Soc. Jpn., 47, 1081–1084.


Kodama, T. and Kumakura, S. (1974b). Bull. Chem. Soc. Jpn., 47, 2146–2151. Kofler, A. (1940). Z. Elektrochem., 50, 200–207. Komorowski, L., Krajewska, A. and Pignon, K. (1976). Mol. Cryst. Liq. Cryst., 36, 337–348. Konno, M., Saito, Y., Yamada, K. and Kawazura, H. (1980). Acta Cryst., B36, 1680–1683. Kortu¨m, G., Vogel, W. and Andrussow, K. (1961). ‘‘Dissociation Constants of Organic Acids in Aqueous Solutions.’’ Butterworths, London. Kosower, E.M. (1966). In ‘‘Flavins and Flavoproteins’’, E. C. Slater, editor, Elsevier, Amsterdam, pp. 1–14. Kozawa, K. and Uchida, T. (1979). Bull. Chem. Soc. Jpn., 52, 1555–1558. Kozawa, K. and Uchida, T. (1983). Acta Cryst., C39, 1233–1235. Kremann, R., Sutter, S., Sitte, F., Strzelba, H. and Dobotzky, A. (1922). Monatsh., 43, 269–313. Kruif, C. G. de, Smit, E. J. and Govers, H. A. J. (1981). J. Chem. Phys., 74, 5838–5841. Kuhn, R. and Strobele, R. (1937). Ber. Deut. Chem. Ges., 73, 753–760. Kumakura, S., Iwasaki, F. and Saito, Y. (1967). Bull. Chem. Soc. Jpn., 40, 1826–1833. Kuo, M. C., Dunn, J. B. R. and Fritchie, C. J., Jr. (1974). Acta Cryst., B30, 1766–1771. Lakshman, M. H., Kole, P. L., Chaturvedi, S., Saugier, J. H., Yeh, H. J. C., Glusker, J. P., Carrell, H. L., Katz, A. K., Afshar, C. E., Dashwood, W-M, Kenniston, G. and Baird, W. M. (2000). J. Am. Chem. Soc., 122, 12629–12636. Lange, F. (1924). Z. Physik. Chem., 110, 343–362. Langhoff, C. A. and Fritchie, C. J., Jr. (1970). J. Chem. Soc. Chem. Commun., pp. 20–21. Larsen, F. K., Little, R. G. and Coppens, P. (1975). Acta Cryst., B31, 430–440. Law, K. S. and Prasad, P. N. (1982). J. Chem. Phys., 77, 1107–1113. Lee, D. and Wallwork, S. C. (1978). Acta Cryst., B34, 3604–3608. Le Bars-Combe, M., Chion, B. and Lajzerowicz-Bonneteau, J. (1979). Acta Cryst., B35, 913–920. Lefebvre, J., Miniewicz, A. and Kowal, A. (1989). Acta Cryst., C45, 1372–1376. Lehmann, R. E. and Kochi, J. K. (1991). J. Am. Chem. Soc., 113, 501–512. Lin, T. and Naae, D. G. (1978). Tetr. Letts., pp. 1653–1656. Lindeman, S. V., Andrianov, V. G., Kravcheni, S. G., Potapov, V. M., Potekhin, K. A. and Struchkov, Yu. T. (1981). J. Struct. Chem. USSR, 22, 578–585; Zh. Strukt. Khim., 23, 123–131. Lisensky, G. C., Johnson, C. K. and Levy, H. A. (1976). Acta Cryst., B32, 2188–2197. Lloyd, L. L. and Sudborough, J. J. (1899). J. Chem. Soc., 75, 580. Lobovskaya, R. M., Shibaeva, R. P. and Ermenk, O. N. (1983). Sov. Phys. Cryst., 28, 161–163. Lopez-Morales, M. E., Soriano-Garcia, M., Gomez-Lara, J. and Toscano, R. A. (1985). Mol. Cryst. Liq. Cryst., 125, 421–427. Lork, E., Mews, R., Shakirov, M. M., Watson, P. G. and Zibarev, A. V. (2001). Eur. J. Inorg. Chem., pp. 2123–2134. Mackay, I. R., Robertson, J. M. and Sime, J. G. (1969). J. Chem. Soc., Chem. Commun., pp. 1470–1471. Mariezcurrena, R. A., Russi, S., Mombru´, A. W., Suescun, L., Pardo, H., Tombesi, O. L. and Frontera, M. A. (1999). Acta Cryst., C55, 1170–1173. Marsh, R. E. (1986). Acta Cryst., B42, 193–198. Marsh, R. E. (1990). Acta Cryst., C46, 1356–1357. Martin, R. H. (1974). Angew. Chem. Int. Ed. Engl., 13, 649–660. Masnovi, J. M., Kochi, J. K., Hilinski, E. F. and Rentzepis, P. M. (1985). J. Phys. Chem., 89, 5387–5395. Matsumoto, N., Nonaka, Y., Kida, S., Kawano and Ueda, I. (1979). Inorg. Chim. Acta, 37, 27–36. Matsunaga, Y. (1964). J. Chem. Phys., 41, 1609–1613.. Matsunaga, Y. (1966). Nature, Lond., 211, 183–184. Matsunaga, Y. (1973). Bull. Chem. Soc. Jpn., 46, 998–999. Matsunaga, Y. (1978). Bull. Chem. Soc. Jpn., 51, 3071–3072.

REFERENCES

1075

Matsunaga, Y. and Osawa, R. (1974). Bull. Chem. Soc. Jpn., 47, 1589–1592. Matsunaga, Y. and Saito, G. (1972). Bull. Chem. Soc. Jpn., 45, 963–964. Matsunaga, Y. and Usui, R. (1980). Bull. Chem. Soc. Jpn., 53, 3085–3088. Matsunaga, Y., Osawa, E. and Osawa, R. (1975). Bull. Chem. Soc. Jpn., 48, 37–40. Matsunaga, Y., Saito, G. and Sakai, N. (1974). Bull. Chem. Soc. Jpn., 47, 2873–2874. Maverick, E., Trueblood, K. N. and Bekoe, D. A. (1978). Acta Cryst., B34, 2777–2781. Mayerle, J. J. (1977). Inorg. Chem., 16, 916–919. Mayerle, J. J. and Torrance, J. B. (1981). Acta Cryst., B37, 2030–2034. Mayerle, J. J., Torrance, J. B. and Crowley, J. I. (1979). Acta Cryst., B35, 2988–2995. McLaughlin, E. and Messer, C. E. (1966). J. Chem. Soc. (A), pp. 1106–1110. Megnamisi–Belombe, M. and Endres, H. (1982). Acta Cryst., B38, 1826–1828. Merski, J. and Eckhardt, C. J. (1981). J. Chem. Phys., 75, 3705–3718. Metzger, R. M., Heimer, N. E., Gundel, D., Sixl, H., Harms, R. H., Keller, H. J., No¨the, D. and Wehe, D. (1982). J. Chem. Phys., 77, 6203–6214. Meyer, K. H. (1909). Chem. Ber., 42, 1149–1153. Mikes, F., Boshart, G. and Gil-Av, E. (1976). J. Chromatogr., 122, 205–221. Miller, J. S., Zhang, J. H., Reiff, W. M., Dixon, D. A., Preston, L. D., Reis, A. H., Jr., Gebert, E., Extine, M., Troup, J., Epstein, A. J. and Ward, M. D. (1987). J. Phys. Chem., 91, 4344–4360. Mitkevich, V. V. and Sukhodub, L. F. (1987). Sov. Phys. Cryst., 31, 483–484; Kristallografiya, 31, 815–817. Mori, T. and Inokuchi, H. (1986). Solid State Commun., 59, 355–359. Mori, T and Inokuchi, H. (1987). Bull. Chem. Soc. Jpn., 60, 402–404. Mori, T., Wu, P., Imaeda, K., Enoki, T., Inokuchi, H. and Saito, G. (1987). Synth. Metals, 19, 545–550. Morosin, B. (1976). Acta Cryst., B32, 1176–1179. Morosin, B. (1978). Acta Cryst., B34, 1905–1909. Morosin, B., Plastas, H. J., Coleman, L. B. and Stewart, J. M. (1978). Acta Cryst., B34, 540–543. Munnoch, P. J. and Wright, J. D. (1974). J. Chem. Soc. Perkin II, pp. 1397–1400. Munnoch, P. J. and Wright, J. D. (1975). J. Chem. Soc. Perkin II, pp. 1071–1074. Naae, D. G. (1979). Acta Cryst., B35, 2765–2768. Nagata, H., In, Y., Doi, M. and Ishida, T. (1995). Acta Cryst., B51, 1051–1068. Nakagawa, H., Tanaka, H., Yamada, K-I. and Kawazura, H. (1982). J. Phys. Chem., 86, 2311–2314. Nakai, H., Shiro, M., Ezumi, K., Sakata, S. and Kubota, T. (1976). Acta Cryst., B32, 1827–1833. Nakamoto, K. (1952). J. Am. Chem. Soc., 74, 1739–1742. Nakamura, K., Kai, Y., Yasuoka, N. and Kasai, N. (1981). Bull. Chem. Soc. Jpn., 54, 3300–3303. Nakasuji, K., Sasaki, M., Kotani, T., Murata, I., Enoki, T., Imaeda, K., Inokuchi, H., Kawamoto, A. and Tanaka, J. (1987). J. Am. Chem. Soc., 109, 6970–6975. Nakasuji, K., Sasaki, M., Murata, I., Kawamoto, A. and Tanaka, J. (1988). Bull. Chem. Soc. Jpn., 61, 4461–4463. Newman, M. S. and Lednicer, D. (1956). J. Am. Chem. Soc., 78, 4765–4770. Newman, M. S. and Lutz, W. B. (1956). J. Am. Chem. Soc., 78, 2469–2473. Niimura, N., Ohashi, Y. and Saito, Y. (1968). Bull. Chem. Soc. Jpn., 41, 1815–1820. Nitta, I., Seki, S., Chihara, H. and Suzuki, K. (1951 Sci. Pap. Osaka Univ. No. 29.). Numan, H., Helder, R. and Wynberg, H. (1976). Rec. Trav. Chim. Pays-bas, 95, 211–212. Ohashi, Y. (1973). Acta Cryst., B29, 2863–2871. Ohashi, Y., Iwasaki, H. and Saito, Y. (1967). Bull. Chem. Soc. Jpn., 40, 1789–1796. Ohmasa, M., Kinoshita, M. and Akamatu, H. (1971). Bull. Chem. Soc. Jpn., 44, 391–395. Ott, J. B., Goates, J. R and Cardon, D. L (1976). J. Chem. Thermodynam., 8, 505–512. Overell, J. S. W. and Pawley, G. S. (1982). Acta Cryst., B38, 1966–1972. Pace, L. J., Ulman, A. and Ibers, J. A. (1982). Inorg. Chem., 21, 199–207.


Pascard, R. and Pascard-Billy, C. (1972). Acta Cryst., B28, 1926–1935. Pasimeni, L., Guella, G., Corvaja, C., Clemente, D.A. and Vincentini, M. (1983). Mol. Cryst. Liq. Cryst., 91, 25–38. Patil, A. O., Curtin, D. Y. and Paul, I. C. (1984). J. Am. Chem. Soc., 106, 4010–4015. Patil, A. O., Curtin, D. Y. and Paul, I. C. (1986). J. Chem. Soc. Perkin Trans.2, pp. 1687–1692. Patil, A. O., Pennington, W. T., Desiraju, G. R., Curtin, D. Y. and Paul, I. C. (1986). Mol. Cryst. Liq. Cryst., 134, 279–304. Patil, A. O., Wilson, S. R., Curtin, D. Y. and Paul, I. C. (1984). J. Chem. Soc. Perkin Trans. 2, pp. 1107–1110. Patrick, C. R. and Prosser, G. S. (1960). Nature, Lond., 187, 1021. Pennington, W. T., Patil, A. O., Curtin, D. Y. and Paul, I. C. (1986), J. Chem. Soc. Perkin Trans.2, pp. 1693–1700. Pereira, J. F. and Tollin, G. (1967). Biochim. Biophys. Acta, 143, 79–87. Perepichka, I. F., Kuz’mina, L. G., Perepichka, D. F., Bryce, M. R., Goldenberg, L. M., Popov, A. F. and Howard, J. A. K. (1998). J. Org. Chem., 63, 6484–6493. Perepichka, I. F., Popov, A. F., Orekhova, T. V., Bryce, M. R., Andrievskii, A. M., Batsanov, A. S. and Howard, J. A. K. (2000). J. Org. Chem., 65, 3053–3063. Perrin, D. (1965). ‘‘Dissociation Constants of Organic Acids in Aqueous Solutions.’’ Butterworths, London. Popova, E. G., Chetkina, L. A. and Bespalov, B. P. (1981). J. Struct. Chem. USSR, 22, 586–589; Zh. Strukt. Khim., 22, 132–136. Popova, E. G., Chetkina, L. A. and Kotov, B. V. (1978). J. Struct. Chem. USSR, 19, 914–919; Zh. Strukt. Khim., 19, 1071–1079. Potenza, J. and Mastropaolo, D. (1975). Acta Cryst., B31, 2527–2529. Povot’eva, Z. P., Chetkina, L. A. and Bespalov, B. P. (1981). J. Struct. Chem. USSR, 22, 386–390; Zh. Strukt. Khim., 22, 94–99. Powell, H. M. and Huse, G. (1943). J. Chem. Soc., pp. 435–437. Powell, H. M., Huse, G. and Cooke, P. W. (1943). J. Chem. Soc., pp. 153–157. Prout, C. K. and Castellano, E. E. (1970). J. Chem. Soc. (A), pp. 2775–2778 Prout, C. K. and Powell, H. M. (1965). J. Chem. Soc., pp. 4882–4887. Prout, C. K. and Tickle, I. J. (1973a). J. Chem. Soc. Perkin II, pp. 731–734. Prout, C. K. and Tickle, I. J. (1973b). J. Chem. Soc. Perkin II, pp. 734–737. Prout, C. K. and Tickle, I. J. (1973c). J. Chem. Soc. Perkin II, pp. 1212–1215. Prout, C. K. and Wallwork, S. C. (1966). Acta Cryst., 21, 449–450. Prout, C. K. and Wheeler, A. G. (1967). J. Chem. Soc. (A), pp. 469–475. Prout, C. K., Morley, T., Tickle, I. J. and Wright, J. D. (1973). J. Chem. Soc. Perkin II, pp. 523–527. Prout, C. K., Tickle, I. J. and Wright, J. D. (1973). J. Chem. Soc. Perkin II, pp. 528–530. Qi, F., Yu, W.-T., Hong, L., Lin, X.-Y. and Yu, Z.-P. (1996). Acta Cryst., C52, 2274–2277. Reddy, D. S., Goud, B. S., Panneerselvam, K. and Desiraju, G. R. (1993). J. Chem. Soc., Chem. Commun., pp. 663–664. Robertson, B. E. and Stezowski, J. J. (1978). Acta Cryst., B34, 3005–3011. Rubio, P., Florencio, F., Garcia-Blanco, S. and Rodriguez, J. G. (1985). Acta Cryst., C41, 1797–1799. Sahaki, M., Yamada, H., Yoshioka, H. and Nakatsu, K. (1976). Acta Cryst., B32, 662–664. Saito, G. and Matsunaga, Y. (1971). Bull. Chem. Soc. Jpn., 44, 3328–3335. Saito, G. and Matsunaga, Y. (1972). Bull. Chem. Soc. Jpn., 45, 2214–2215. Saito, G. and Matsunaga, Y. (1973a). Bull. Chem. Soc. Jpn., 46, 714–718. Saito, G. and Matsunaga, Y. (1973b). Bull. Chem. Soc. Jpn., 46, 1609–1616. Saito, G. and Matsunaga, Y. (1974). Bull. Chem. Soc. Jpn., 47, 1020–1021.

REFERENCES

1077

Sakanoue, S., Yasuoka, N., Kasai, N. and Kakudo, M. (1971). Bull. Chem. Soc. Jpn., 44, 1–8. Sakurai, T. (1965). Acta Cryst., 19, 320–330. Sakurai, T. (1968). Acta Cryst., B24, 403–412. Sakurai, T. and Tagawa, H. (1971). Acta Cryst., B27, 1453–1459. Sandman, D. J., Grammatica, S. J., Holmes, T. J. and Richter, A. F. (1980). Mol. Cryst. Liq. Cryst., 59, 241–252. Sato, A., Okada, M., Saito, K. and Sorai, M. (2001). Acta Cryst., C57, 564–565. Scarbrough, F. E., Shieh, H.-S. and Voet, D. (1976). Proc. Nat. Acad. Sci., 73, 3807–3811. Scarbrough, F. E., Shieh, H.-S. and Voet, D. (1977). Acta Cryst., B33, 2512–2523. Scheidt, W. R. and Reed, C. A. (1978). Inorg. Chem., 17, 710–714. Scheidt, W. R., Kastner, M. E. and Hatano, K. (1978). Inorg. Chem., 17, 706–710. Schmitt, R. D., Wing, R. M. and Maki, A. H. (1969). J. Am. Chem. Soc., 91, 4394–4401. Schreiner, E. (1925). Z. physik. Chem., 117, 57–87. Shaanan, B. and Shmueli, U. (1980). Acta Cryst., B36, 2076–2082. Shaanan, B., Shmueli, U. and Rabinovich, D. (1976). Acta Cryst., B32, 2574–2580. Shah, M. C. and Baughman, M. C. (1994). Acta Cryst., C50, 1114–1117. Shchlegova, T. M., Starikova, Z. A., Trunov, V. K., Lantratova, O. B. and Pokrovskaya, I. E. (1981). J. Struct. Chem., 22, 553–557; Zh. Strukt. Khim., 22, 93–97. Shibaeva, R. P. and Yarochkina, O. Y. (1975). Sov. Phys. Dokl., 20, 304–305. Shipley, G. G. and Wallwork, S. C. (1967a). Acta Cryst., 22, 585–592. Shipley, G. G. and Wallwork, S. C. (1967b). Acta Cryst., 22, 593–601. Shkolnik, G. M. and Geiger, W. E., Jr. (1966). Inorg. Chem., 5, 1020–1025. Shmueli, U. and Goldberg, I. (1974). Acta Cryst., B30, 573–578. Shmueli, U. and Mayorzik, H. (1980). Abstract 1-A-37, ECM-6, Barcelona. Shvets, A. E., Bleidelis, Ya. Ya. and Freimanis, Ya. F. (1974). J. Struct. Chem. USSR, 15, 430–433; Zh. Strukt. Khim., 15, 504–508. Shvets, A. E., Bleidelis, Ya. Ya. and Freimanis, Ya. F. (1975a). Zh. Strukt. Khim., 16, 98–103. Shvets, A. E., Bleidelis, Ya. Ya. and Freimanis, Ya. F. (1975b). J. Struct. Chem. USSR, 16, 386–390; Zh. Strukt. Khim., 16, 604–606. Shvets, A. E., Bleidelis, Ya. Ya. and Freimanis, Ya. F. (1975c). J. Struct. Chem. USSR, 16, 592–596; Zh. Strukt. Khim., 16, 640–644. Shvets, A. E., Bleidelis, Ya. Ya. and Freimanis, Ya. F. (1975d). Zh. Strukt. Khim., 16, 415–419. Shvets, A. E., Bleidelis, Ya. Ya., Freimanis, Ya. F. and Bundule, M. F. (1978). J. Struct. Chem. USSR, 18, 84–87; Zh. Strukt. Khim., 18, 107–111. Shvets, A. E., Bleidelis, J. J., Markava, E. J., Freimanis, J. F. and Kanepae, D. V. (1980). J. Struct. Chem., 21, 559–563; Zh. Strukt. Khim., 21, 190–195. Shvets, A. E., Malmanis, A. J., Freimanis, Ya. F., Bleidelis, Ya. Ya. and Dregeris, J. F. (1979). J. Struct. Chem. USSR, 20, 414–419; Zh. Strukt. Khim., 20, 491–497. Siegmund, W. (1908). Monatsh., 29, 1089–1109. Singh, N. B., McWhinnie, W. R., Ziolo, R. F. and Jones, C. H. W. (1984). J. Chem. Soc., Dalton Trans., pp. 1267–1273. Singhabandhu, A., Robinson, P. D., Fang, J. H. and Geiger, W. E., Jr. (1975). Inorg. Chem., 14, 318–321. Smith, V. B. and Massey, A. G. (1969). Tetrahedron, 25, 5495–5501. Soos, Z. G., Keller, H. J., Ludolf, K., Queckbo¨rner, J., Wehe, D. and Flandrois, S. (1981). J. Chem. Phys., 74, 5287–5294. Soos, Z. G., Keller, H. J., Moroni, W. and Nothe, D. (1977). J. Am. Chem. Soc., 99, 5040–5044. Soos, Z. G., Keller, H. J., Moroni, W. and Nothe, D, (1978). Ann. N. Y. Acad. Sci., 313, 442–458. Soos, Z. G., Mazumdar, S. and Cheung, T. P. P. (1979). Mol. Cryst. Liq. Cryst., 52, 92–102.


Soriano-Garcia, M., Toscano, R. A., Robles Martinez, J. G., Salme´ron, U. A. and Lezama, R. R. (1989). Acta Cryst., C45, 2442–2444. Staab, H. A., Herz, C. P., Krieger, C. and Rentea, M. (1983). Chem. Ber., 116, 3813–3830. Starikova, Z. A., Shchegoleva, T. M., Trunov, V. K., Lantratova, O. B. and Pokrovskaya, J. E. (1980). J. Struct. Chem., 21, 181–184; Zh. Strukt. Khim., 21, 73–76. Stephens, F. S. and Vagg, R. S. (1980). Inorg. Chim. Acta, 43, 77–82. Stephens, F. S. and Vagg, R. S. (1981). Inorg. Chim. Acta, 51, 163–167. Stezowski, J. J. (1980). J. Chem. Phys., 73, 538–547. Stezowski, J. J., Binder, W. and Karl, N. (1982). Acta Cryst., B38, 2912–2914. Stezowski, J. J., Stigler, R.-D. and Karl, N. (1986). J. Chem. Phys., 84, 5162–5170. Sugano, T., Hashida, T., Kobayashi, A., Kobayashi, H. and Kinoshita, M. (1988). Bull. Chem. Soc. Jpn., 61, 2303–2308. Surcouf, E. and Delettre´, J. (1978). Acta Cryst., B34, 2173–2176. Suzuki, K. and Seki, S. (1953). Bull. Chem. Soc. Jpn., 26, 372–380. Suzuki, T., Fujii, H., Yamashita, Y., Kabuto, C., Tanaka, S., Harasawa, M., Mukai, T. and Miyashi, T. (1992). J. Am. Chem. Soc., 114, 3034–3043. Suzuki, T., Kabuto, C., Yamashita, Y., and Mukai, T. (1987). Bull. Chem. Soc. Jpn., 60, 2111–2115. Swinton, F. L. (1974). ‘‘Interactions in binary systems containing aromatic fluorocarbons’’, Molecular Complexes, edited by R. Foster, London, Elek, Vol. II, pp. 63–106. Szent-Gyorgi, A. (1960). ‘‘An Introduction to a Submolecular Biology’’, New York: Academic Press, p. 197. Tachikawa, N., Yakushi, K. and Kuroda, H. (1974). Acta Cryst., B30, 2770–2772. Takahashi, N., Yakushi, K., Ishii, K. and Kuroda, H. (1976). Bull. Chem. Soc. Jpn., 49, 182–187. Talukdar, A. N. and Chaudhuri, B. (1976). Acta Cryst., B32, 803–808. Tamamura, T., Yamane, T., Yasuoka, N. and Kasai, N. (1974). Bull. Chem. Soc. Jpn., 47, 832–837. Tamura, H. and Ogawa, K. (1977). Cryst. Struct. Commun., 6, 517–520. Thewalt, U. and Bugg, C. E. (1972). Acta Cryst., B28, 82–92. Thozet, A. and Gaultier, J. (1977a). Acta Cryst., B33, 1052–1057. Thozet, A. and Gaultier, J. (1977b). Acta Cryst., B33, 1058–1063. Tickle, I. J. and Prout, C. K. (1973a). J. Chem. Soc. Perkin II, pp. 720–723. Tickle, I. J. and Prout, C. K. (1973b). J. Chem. Soc., Perkin II, pp. 724–727. Tickle, I. J. and Prout, C. K. (1973c). J. Chem. Soc. Perkin II, pp. 727–731. Tillberg, O. and Norrestam, R. (1972). Acta Cryst., B28, 890–898. Tollin, G. (1968). In ‘‘Molecular Associations in Biology,’’ Proceedings of an International Symposium, B. Pullman, editor, Academic Press, New York, pp. 393–409. Torrance, J. B., Mayerle, J. J., Lee, V. J. and Bechgaard, K. (1979). J. Am. Chem. Soc., 100, 4747–4748. Torrance, J. B., Vazquez, J. E., Mayerle, J. J. and Lee, V. Y. (1981). Phys. Rev. Letts., 46, 253–257. Toupet, L. and Karl, N. (1995). Acta Cryst., C51, 249–251. Toyoda, J., Oda, A., Murata, I., Kawamoto, A., Tanaka, J. and Nakasuji, K. (1993). Bull. Chem. Soc. Jpn., 66, 2115–2117. Treibs, A. (1929). Ann., 476, 1–60. Truong, K. D. and Bandrauk, A. D. (1977). Can. J. Chem., 55, 3712–3716. Tsuchiya, H., Marumo, F. and Saito, Y. (1973). Acta Cryst., B29, 659–666. Vangala, V. R., Nangia, A. and Lynch, V. M. (2002). Chem. Commun., pp. 1304–1305. Viossat, B., Dung, N.-G. and Daran, J. C. (1988). Acta Cryst., C44, 1797–1800. Viossat, B., Tomas, A., Dung, N.-G., Mettey, Y. and Viervond, J. -M. (1995). Acta Cryst., C51, 1896–1898.

REFERENCES

1079

Visser, R. J. J., Bouwmeester, H. J. M., Boer, J. L. de. and Vos, A. (1990). Acta Cryst., C46, 852–856. Voet, D. and Rich, A. (1971). Proc. Nat. Acad. Sci., 68, 1151–1156. Weck, M., Dunn, A. R., Matsumoto, K., Coates, G. W., Lobkovsky, E. and Grubbs, R. H. (1999). Angew. Chem. Int. Ed. Engl., 38, 2741–2745. Wells, J. L., Trus, B. L., Johnston, R. M., Marsh, R. E. and Fritchie, C. J., Jr. (1974). Acta Cryst., B30, 1127–1134. Wentworth, W. E., Limero, T. and Chen, E. C. M. (1987). J. Phys. Chem., 91, 241–245. Wilkerson, A. K., Chodak, J. B. and Strouse, C. E. (1975). J. Am. Chem. Soc., 97, 3000–3004. Williams, J. H. (1993). Acc. Chem. Res., 26, 593–598. Williams, J. H., Cockcroft, J. K. and Finch, A. N. (1992). Angew. Chem. Int. Ed. Engl., 31, 1655–1657. Williams, J. M., Beno, M. A., Wang, H. H., Leung, P. C. W., Emge, T. J., Geiser, U. and Carlson, K. D. (1985). Acc. Chem. Res., 18, 261–267. Williams, R. M. and Wallwork, S. C. (1966). Acta Cryst., 21, 406–412. Williams, R. M. and Wallwork, S. C. (1967). Acta Cryst., 23, 448–455. Williams, R. M. and Wallwork, S. C. (1968). Acta Cryst., B24, 168–174. Wing, R. M. and Schlupp, R. L. (1970). Inorg. Chem., 9, 471–475. Wit, H. G. M. de, Klauw, C. H. M. van der, Derissen, J. L., Govers, H. A. J. and Chanh, N. B. (1980). Acta Cryst., A36, 490–492. Wozniak, K., Kariuki, B. and Jones, W. (1991). Acta Cryst., C47, 1113–1114. Wright, J. D. and Ahmed, Z. A. (1981). Acta Cryst., B37, 1848–1852. Wright, J. D., Yakushi, K. and Kuroda, H. (1978). Acta Cryst., B34, 1934–1938. Yakushi, K., Ikemoto, I. and Kuroda, H. (1971). Acta Cryst., B27, 1710–1718. Yakushi, K., Ikemoto, I. and Kuroda, H. (1973). Acta Cryst., B29, 2640–2641. Yakushi, K., Ikemoto, I. and Kuroda, H. (1974a). Acta Cryst., B30, 835–837. Yakushi, K., Ikemoto, I. and Kuroda, H. (1974b). Acta Cryst., B30, 1738–1742. Yakushi, K., Tachikawa, N., Ikemoto, I. and Kuroda, H. (1975). Acta Cryst., B31, 738–742. Yakushi, K., Uesaka, T. and Kuroda, H. (1985). Mol. Cryst. Liq. Cryst., 125, 355–363. Yamachi, H., Saito, G., Sugano, T., Katayama, C. and Tanaka, T. (1987). Synth. Metals, 19, 533–536. Yamaguchi, Y. and Ueda, I. (1984). Acta Cryst., C40, 113–115. Yamashita, Y., Suzuki, T., Mukai, T. and Saito, G. (1985). J. Chem. Soc. Chem. Commun., pp. 1044–1045. Zacharias, D. E. (1976). Abstracts of ACA Winter Meeting, Clemson, S. Carolina, p. 12. Zacharias, D. E. (1993). Acta Cryst., C49, 1082–1087. Zacharias, D. E., Prout, K., Myers, C. B. and Glusker, J. P. (1991). Acta Cryst., B47, 97–107. Zobel, D. and Ruban, G. (1983). Acta Cryst., B39, 638–645.


Chapter 16 Crystal (structural) physics of mixed stack –* molecular compounds

Be not curious in unnecessary matters for more things are shewed to you than men understand. Ecclesiaticus 3:23

Summary: Thermodynamic measurements for a limited sample of crystalline mixed-stack 1 : 1 -molecular compounds show that most are enthalpy-stabilized, some entropy-stabilized, and a few both enthalpy and entropy-stabilized. Correlation of these thermodynamic results with crystal structures remains a task for the future. Combination of optical spectroscopic methods at very low temperatures and electron spin resonance measurements have provided proof of Mulliken’s theory also for the solid state. The room temperature crystal structures of 1 : 1 -molecular compounds are not necessarily representative of the entire range of pressure–temperature behavior of these materials. There are often hints of disorder (usually of the donor) in the room temperature structures, and these have been correlated for a few systems with disorder-to-order transitions (thermodynamically second-order, following Ehrenfest) that occur on cooling; these have been studied by a combination of calorimetric, diffraction and resonance techniques. Despite overall similarities, each system surveyed has its own individual characteristics. A number of 1 : 1 -molecular compounds have been shown to transform to quasi-plastic phases on heating. A small number of mixed stack 1 : 1 -molecular compounds with neutral ground states have been shown to transform to ionic structures on cooling or application of pressure.

16.1 16.2 16.3 16.4

Introduction Thermodynamic parameters Spectroscopic measurements on the excited state Crystals with disorder ) order transformations on cooling – modern treatments of second order phase transitions 16.4.1 General introduction 16.4.2 The Ehrenfest order of a phase transition 16.4.3 Landau theory of phase transitions 16.4.4 The critical exponents 16.4.5 The permitted symmetries of a low symmetry phase derived from a particular high symmetry phase 16.4.6 Temperature dependence of the order parameter 16.4.7 Pressure dependence of the critical temperature for ordering 16.5 Thermodynamic, structural and kinetic investigations of various systems showing second order transitions on cooling

1082 1083 1086 1090 1090 1091 1092 1093 1094 1096 1097 1097

CRYSTAL (STRUCTURAL) PHYSICS

1082

16.5.1

The crystal structure of {Pyrene PMDA}(PYRPMA) and evidence for an order , disorder phase transition at 160K 16.5.2 The crystal structure of {Naphthalene TCNB} and evidence for an order , disorder phase transition around 72K 16.5.3 The crystal structure of {Anthracene TCNB} and evidence for an order , disorder phase transition at 213K 16.5.4 Other examples of second order transitions 16.6 Crystals with first order phase transformations on cooling 16.6.1 {Cycl[3.2.2]azine TNB} 16.6.2 Other examples 16.7 Physical nature of the disordered phase 16.8 Transformations to quasi-plastic phase(s) on heating 16.9 Transformation of the ground state from neutral ) ionic on cooling and/or application of pressure (NI transitions) 16.9.1 Introduction 16.9.2 {TTF chloranil} 16.9.3 {DMTTF chloranil} 16.9.4 Other examples 16.9.5 Concluding summary References n n n

1097

n n n

1105

n n n

n n n

n n n

n n n

16.1

1115 1119 1120 1120 1122 1122 1126 1128 1128 1129 1137 1139 1142 1142

Introduction

The mixed-stack model of crystalline -molecular compounds described in the preceding chapters requires amplification in many respects for it to provide a full picture of current knowledge. In this chapter we shall concentrate on some aspects of the structural physics of mixed-stack 1 : 1 -molecular crystals; treatment is limited to the 1 : 1 composition and infinite mixed stack arrangement because information is not available for other compositions and types of structure. We begin by considering the thermodynamic parameters reported for ambient temperature and pressure; this will give some overall feeling for the strengths of the interactions leading to the formation of the molecular compounds, and, in particular, whether these are enthalpy or entropy stabilized (or both) with respect to the individual components. In the next section we discuss low temperature spectroscopic studies that give information about the degree of charge transfer in the excited state and hence provide a test of the applicability of Mulliken’s theory to the solid state. The types of solid state transformation that occur on cooling are then discussed in some detail. Most of the transformations studied so far are of the second order, but there are also some first-order transformations (‘order’ defined below). This leads us to consideration of the nature of the structural disorder found in many crystalline -molecular compounds at room temperature. Then two types of solid state transformation pertinent to only a limited number of -molecular compounds are discussed – for some, to a quasi-plastic phase at high temperatures, and for some others to an ionic ground state on cooling or application of pressure. One aspect of the crystal physics of mixed-stack -molecular compounds has been entirely omitted – their electronic properties. This is not because of lack of importance but because it would move us too far from our essentially structural theme.

T HE R M O D Y N AM IC PA R A M E T E R S

1083

16.2 Thermodynamic parameters We note in Appendix 1 (section 1) that a crystalline binary adduct may be enthalpy and entropy stabilized with respect to its crystalline components (situation (i)), or entropy stabilized (situation (ii)), or enthalpy stabilized (situation (iii)). Values of Hc and Sc for three groups of -molecular compounds are plotted in Fig. 16.1. These values refer nominally to 298K and were determined by the electrochemical method (Appendix 1, Section 2.1.1), where Hc and Sc (assumed constant over the range 278–318K) are derived from the measured temperature dependence of Gc. The precision of the thermodynamic parameters is not high. Although the errors of the free energy values range from 0.01 to 0.18 kJ/mol, those of the derived Hc and Sc values range from 5% up to 40%. Fig. 16.1 shows that all three permitted quadrants of Hc – Sc space are populated, but not in equal measure. For the available sample, which is not necessarily randomly selected from the total population, most of the -molecular compounds are enthalpy stabilized, fewer are entropy stabilized and even fewer are both enthalpy and entropy stabilised. Abdel-Rehiem et al. (1975) pointed out that there is an approximately linear relationship between Hc and Sc for compound formation between a series of monosubstituted naphthalenes and picric acid; this is illustrated for a wider range 30 ✖ ✖

✖ ✖

20 Forbidden Region

✖

Entropy stabilized

✖✖✖✖ ✖

10 ∆Hc (kJ/mol)

✖

0

★ ● ✖ ● ★ ● ● ● ● ● ●★ ● ✖✖ ✖ ● ● ✖ ● ✖ ● ● ■■■ ● ● ■ ●■ ■■ ■■ ✖ ■ ■ ■ ■ ■■ ■■ ✖

–10 Enthalpy stabilized

–20 ■ ■

–30 –100

✖

Enthalpy and entropy stabilized

■ ■

■

–50

0

50

100

150

∆Sc (J/mol K)

Fig. 16.1. Plot of Hc against Sc for 60 -molecular compounds of the following groups: (a) crosses – Picric acid compounds of various aromatic hydrocarbons [19 values] (Shahidi and Farrell, 1980), denoted by diagonal crosses; (b) filled squares – Picric acid compounds of substituted naphthalenes [21 values] (Abdel-Reheim et al., 1975); (c) circles – Styphnic acid (1,3-dihydroxy2,4,6-trinitrobenzene) compounds of substituted naphthalenes [17 values] (Shahidi, Farrell and Westwood, 1980); (d) five-pointed stars – aromatic hydrocarbons with TNB (3 values). The detailed numerical values and identities of the compounds are given in the cited references.


1084

of compounds in Fig. 16.1. The values for the picric acid compounds of substituted naphthalenes and of aromatic hydrocarbons all fall on the same straight line; the first of these groups is enthalpy stabilized all the latter either enthalpy and entropy stabilized or entropy stabilized. The line for the styphnic acid compounds (of much the same group of substituted naphthalenes) is parallel but displaced to higher Hc values; these are all enthalpy stabilized. The linear relationship between Hc and Sc follows from the form of the free energy equation recast as Hc ¼ TSc þ Gc A linear plot of Hc against Sc requires that the slope should be T ( ¼ 298K for the present set of data) and that Gc should be constant; using all the points in Fig. 16.1, we obtain Hc (kJ/mol) ¼ 250 (K) Sc (J/mol K)6.13 (kJ/mol). The derived slope (R2 ¼ 0.917) is reasonably close to the required value. The distribution of Gc is shown in Fig. 16.2, where the different groups of molecular compounds have not been distinguished. If this is done then one finds that the mean for the picric acid compounds of substituted naphthalenes is 6.54(2.4) kJ/mol), 1.1(0.7) kJ/ mol for the styphnic acid compounds of substituted naphthalenes and 10.5(2.7) kJ/mol for the picric acid compounds of aromatic hydrocarbons. The picric acid compounds of substituted naphthalenes and of aromatic hydrocarbons do not differ significantly in terms of free energy values but are very different when enthalpies and entropies are compared; the least stable are the styphnic acid compounds of substituted naphthalenes. The deviations of individual points from the overall linear relationship must have a physical explanation, as must the location of particular types of compound in different quadrants of the diagram; for example (with the exceptions of indene and acenaphthylene) all the {aromatic hydrocarbon picric acid} compounds have large positive entropies of formation. Presumably the principal distinctions will have to be made among the compounds lying in the different quadrants, in terms of the different types of stabilization (enthalpy and/or entropy) that apply. Unfortunately there is not much overlap between the sample whose crystal structures have been determined and that for which thermodynamic n n n

20

Number of compounds

18 16 14 12 10 8 6 4 2 0 0) and discontinuous (B < 0) transitions (see Fig. 16.6). The condition for thermodynamic equilibrium is determined by the minima of the potentials, i.e. for @G/ @Q ¼ 0. The solutions for second order (2-4) and tricritical (2-6) potentials are 2-4: Q2 ¼ A/B (Tc.– T) and 2-4-6: Q4 ¼ A/C (Tc – T), T < Tc. In Landau’s original work the implicit assumption was made that the polynomial expansion of the potential was only valid in the vicinity of Tc, and hence for small values of Q, but Salje (1990) has proposed that ‘‘the polynomial expansion of G is a good approximation over an extended temperature interval and that the approximation also holds for larger values of the order parameter.’’ We consider in the next section how the results summarised here can be used to determine whether there is justification for applying Salje’s extension of Landau theory to binary molecular compounds. 16.4.4 The critical exponents In a more general approach, the behaviour of the system is described in terms of critical exponents , the values of which depend on the physical property under investigation and the nature of the transition (Berry, Rice and Ross, 1980, p. 863–864; Lifshitz and Pitaevskii, 1980; Stanley, 1971; Tole´dano and Tole´dano, 1987). Very general relationships among the various critical exponents can be derived that do not depend on the physical system. A physically measurable quantity such as spontaneous strain (defined below), intensity of a superlattice reflection (for these two quantities the symbol b is generally used for the exponent) or the excess specific heat Cp (here is generally used for the exponent) can be given as f(–") ¼ A(–") {1 þ B(– ")xþ }

(x > 0), where " ¼ (T – Tc)/Tc.

The correction terms drop out on taking the limit in order to obtain the critical exponent, which is defined as ¼ lim ½lnf ð"Þ=lnð"Þ : "!0


1094

is expressed as a function of temperature in the form f(1–T/Tc), which is given (to a first approximation) as f ð1 T=Tc Þ ¼ Að1 T=Tc Þ

ð16:2Þ

or, recast in log-log form, as log½ f ð1 T=Tc Þ ¼ log A þ log½ð1 T=Tc Þ :

ð16:3Þ

The critical exponents have generally been obtained from measurements made very close to Tc (perhaps within 0.5–1 ) because then the effect of fluctuations can be investigated. Ignoring the correction terms, one has f(– ") ¼ A(–"). We have used this power law form to fit various order parameters and obtain accompanying values for the exponents. The Salje extension implies that data over the whole available temperature range can be used; thus equation (16.3) should give a linear plot with slope . We shall find that this holds well for {pyrene PMDA} (PYRPMA), but that there are problems, discussed below, with {naphthalene TCNB} (NAPTCB) and {anthracene TCNB} (ANTCYB); however, these are hardly critical exponents in the sense of the theory, particularly because experimental values sufficiently close to Tc are usually lacking. We consider them convenient for description but their physical significance has still to be established. n n n

n n n

16.4.5

n n n

The permitted symmetries of a low symmetry phase derived from a particular high symmetry phase

The permissible space groups of the low temperature phase derived from a particular high temperature phase after a second order phase transition were first obtained by Landau and Lifshitz in 1937–1939 (see Lifshitz and Pitaevski, 1980, XIV x145) and, more extensively, by Lyubarski (1960) and Koci 0 nski (1983). The latter authors have treated some tetragonal and hexagonal space groups in detail. However, space groups of mixed stack -molecular compounds are usually monoclinic and, to the best of our knowledge, those relevant to second order phase transitions have been treated only by Bernstein (1967). We quote a limited set of Bernstein’s results, assuming that the high temperature phase has space group C2/m or P21/a, that the moieties are located at crystallographic centers, and that the transition occurs with preservation of axial directions. I. High temperature phase has space group C2/m. The conditions and possible space groups for the low temperature phase are: 1. Unit cell remains centered and cell volume remains unchanged (a) Cm, (b) C2, (c) C 1. 2. Unit cell becomes primitive but cell volume remains unchanged (a) P2/m, (b) P21/a, (c) P2/a, (d) P21/m. Examples of I.2(b) are {naphthalene TCNB}, {anthracene TCNB}, {pyrene C6F6}, and (perhaps) {anthracene TCNQ}. 3. Unit cell remains C-centered, cell volume is doubled and lattice becomes triclinic (P). The axial directions are not preserved in the reduced cell. 4. Unit cell becomes body centered monoclinic, c axis and cell volume are doubled. The axial directions are not preserved in the reduced cell. (a) I2/m, (b) I2/c. Example of I.4(b) is {anthracene C6F6}. 5. Unit cell remains C-centered monoclinic, c axis and cell volume are doubled. (a) C2/m, (b) C2/c. n n n

n n n

n n n

n n n

n n n

DISORDER ) ORDER TRANSFORMATIONS

1095

II. High temperature phase has space group P21/a. The conditions and possible space groups for the low temperature phase are: 1. Unit cell remains primitive monoclinic, but center of symmetry disappears (a) P21, (b) Pa. The C-polymorph of naphthazarin (of course, not a -molecular compound) belongs to II.1(b) (Herbstein et al., 1985), as does (TTF-CA} (x16.9.2). 2. Unit cell becomes primitive triclinic, without change of volume (a) P 1 A possible example of II.2(a) is {naphthalene TCNE}. 3. Unit cell remains primitive monoclinic, c axis and cell volume are doubled. (a) P21/a, (b) P21/n. Example of II.3(b) is {pyrene PMDA}. 4. Unit cell becomes pseudo-centered monoclinic, but the actual space group is P 1. The axial directions are not preserved in the reduced cell. (a) pseudo-C-centered (b) pseudo-I-centered. n n n

n n n

The piecemeal contributions described above, important as they were, have now been consolidated into the tables of Stokes and Hatch (1988). The parent (high symmetry) polymorphic phase with space group G0 transforms, with a physical generalized distortion (r), to a low symmetry phase with space group G, where G is a subgroup of G0. Stokes and Hatch (1988) have derived the G space groups for all the 230 G0 space groups. The G0 space groups of interest here are C2/m and P21/c and we excerpt in Table 16.3 the relevant information from the Stokes and Hatch tables; this is discussed together with phase transitions of the various compounds. The label ‘Spec(ies)’ in Table 16.3 refers to the ferroic species of the transition. For the three examples of interest here, this label is ‘nf’ nonferroic. A ‘nonferroic’ or ‘co-elastic’ crystal has ‘‘elastic and strain anomalies . . . correlated with the structural phase transition’’ (Salje, 1990). The Bernstein and Stokes and Hatch prescriptions can be reconciled when one remembers that in Bernstein the transition occurs with preservation of axial directions

Table 16.3. Information from the Stokes and Hatch (1988) tables for the G0 space groups C2/m and P21/c. The definitions of the various column headings are given by Stokes and Hatch. The columns relevant in the present context have been emphasized Space group G0

Parent Irrep.

Image

Lan

Lif

Subgroup G

Spec.

Dir

Size

Basis

Origin

C2/m

Yþ 2

A2a

0

0

14 P21/c

nf

P1**

2

(000)

Aþ 2

A2a

0

0

15 C2/c

nf

P1**

2

Yþ 2

A2a

0

0

14 P21/c

nf

P1**

2

(001); ð0 10Þ; (100) (100); (010); (002) (001); (010); ð 20 1Þ

P21/c

(000)

(000)

1096


while the space group can change (as we have done here for {pyrene PMDA}, whereas Stokes and Hatch maintain space group and allow axial directions to change. n n n

16.4.6

Temperature dependence of the order parameter

We now define ‘‘spontaneous strain’’ and show how it is measured, basing ourselves extensively on Salje (1990; see Chapter 4). In a ferroelastic (not relevant here) or co-elastic transition the shape of the crystal is changed and this creates a macroscopic spontaneous strain. Ignoring microstructures (justified by the absence of transitioninduced twinning in the crystals considered here), the macroscopic spontaneous strain can be replaced by the structural spontaneous strain, usually called the spontaneous strain, which is measured as the volume average of the deformation of the unit cell. Originally introduced for ferroelastic systems by Aizu (1970), this definition has been expanded to include all structural phase transitions that lead to variation of the shape of the crystallographic unit cell, especially for co-elastic systems. In order to measure the spontaneous strain the lattice parameters of the high-symmetry phase have to be extrapolated into the temperature regime of the low-symmetry phase, i.e. the high symmetry phase is the reference phase. This extrapolation represents that part of the thermal expansion that is not related to the structural phase transition and therefore does not contribute to the excess spontaneous strain. The numerical values of the spontaneous strain are now defined by the strain tensor that relates the low-symmetry unit cell to the high-symmetry unit cell when extrapolated to the same temperature. The spontaneous strain in the (low-symmetry) ordered phase is conveniently given in terms of the Vogt coefficients ej, calculable as a function of temperature from the measured cell dimensions of the ordered phase (not subscripted) and the extrapolated cell dimensions of the disordered phase (subscripted) using e1 ¼ a/a0 – 1, e2 ¼ b/b0 – 1, e3 ¼ (c sin b*/c0 sin b0*) – 1, e5 ¼ (a cos b*/a0 sin b0*) – (c cos b*/c0 sin b0*).

(16.4)

We give the equations only for monoclinic crystals as the three present examples are all monoclinic; Salje (1990) gives the equations for all the crystal families. The (scalar) spontaneous strain in the ordered phase is es ¼ (ej2)1/2. The long range order parameter Q refers to the crystal structure averaged over many unit cells; it can be determined from measurements on the intensities of superlattice reflections, from the spontaneous strain and from the ESR spectrum. Spontaneous strain in the ordered phase, and intensities of the superlattice reflections are both proportional to Q2. There is also a (structural) short range order parameter referring to the crystal structure averaged over a few unit cells, which will be only briefly considered in what follows; it can be determined from measurements on the intensities of the diffuse scattering of x-rays (or neutrons). The diffraction aspects of these parameters are clearly discussed by Warren (1969; see Chapter 12). The physical interpretation (i.e. at the molecular level) of the measured values is often controversial. The disordered state above Tc is usually described in terms of one or other of two extremes, statistical ‘static’ disorder where the potential energy diagram for a molecule in the field of its neighbors is a double-well potential with

C R Y S T AL S WI T H S E C OND OR DE R P HAS E T R ANSFOR MATI ONS I N C OOLING 1097

the barrier height appreciably higher than kT, or as ‘dynamic’ disorder where the potential energy diagram is a flat-bottomed well (definitions adapted from Lefebvre et al., 1989). One is able to build up a fairly satisfactory picture of the ideally ordered structures below the transition temperature but it is much more difficult to define the nature of the high temperature, disordered structure, and particularly to decide whether the disorder is dynamic or static. Some progress has been made by use of rigid-group refinement of diffraction data and solid state NMR methods, and by model calculations. These results are described in the final parts of this section. One important concept, mentioned here only in passing, is that of the ‘antiphase domain’ (Warren, 1990; pp. 216–227). These are well-ordered regions of the crystal separated by ‘change step’ boundaries. As the name suggests, neighbouring domains are mutually out-of-phase. The concept has been applied mostly to metallic alloys (e.g. Cu3Au) and hardly to organic systems. 16.4.7 Pressure dependence of critical temperature for ordering Applying Fig. 16.6 to PYRPMA, the point K has the coordinates P ¼ 1 atm. (1 bar), Tc 165K, the line separating ordered and disordered phases has a positive slope and, following Fig. 16.6, one would expect increase of pressure to favour ordering. The pressure dependence of Tc. can be calculated from Ehrenfest’s second order analog to the first order Clapeyron equation given by Pippard’s (1964), equations (8.19) and (9.1.) Ehrenfest’s equation can be written as dTc/dP ¼ Tc[(@V/@T)2 – (@V/@T)1]/[(CP)2 – (CP)1]

(16.5)

where the subscripts 1 and 2 refer to the state of the system just below and just above Tc. A value of [(@V/@T)2 – (@V/@T)1] is obtained from Fig. 16.10 and a value of [(CP)2 – (CP)1,] from Fig. 16.7; using these values we calculate that dTc/dP þ17 K/kbar, if the specific heat values given by Boerio-Goates and Westrum (1980) are used, and about half this amount using those of Dunn et al. (1978). Thus application of a pressure of a few kbar should produce the ordered phase of PYRPMA at room temperature. Anticipating later discussion we note that [(@V/@T)2 – (@V/@T)1] is negative for PYRPMA and ANTCYB and positive for NAPTCB. As [(CP)2 – (CP)1] is negative, application of pressure should increase Tc for PYRPMA and ANTCYB, and reduce it for NAPTCB. The only experimental test is for ANTCYB (Ecolivet et al., 1988; see x16.5.3(d)).

16.5 Thermodynamic, structural and kinetic investigations of various systems showing second order transitions on cooling3 16.5.1 The crystal structure of {Pyrene PMDA} (PYRPMA) and evidence for an order , disorder phase transition at 160K n n n

(a) Introduction. The melting point diagram (Herbstein and Snyman, 1969; HS69) shows three molecular compounds, the equimolar compound being by far the most stable of the three. The 160K phase transition in PYRPMA was first demonstrated by x-ray 3

This treatment is largely based on Herbstein (1996).


1098

35

∆Cp (J/mol K)

30 25 20 15 10 5 0 120

130

140 T (K)

150

120

140

160

∆Cp (J/mol K)

260

210

160 100

160

180

200

T (K)

Fig. 16.7. Part of the Cp – T curve for {pyrene PMDA}, showing the broad transition centred about 155K which corresponds to the second-order transition studied by x-ray diffraction. The filled circles show the results from Boerio-Goates and Westrum (1980) and the open circles those from Dunn et al., (1978). The specific heat is a smooth function of temperature outside the range shown. The enlarged inset shows CP after subtraction of background. The values of Cp1, and Cp2, the specific heats just below and above Tc, are 272 and 245 J/mol K. The curves are guides to the eye. (Adapted from HS93.) n n n

diffraction (HS69) but many of its parameters were established by two independent sets of calorimetric measurements (Dunn, Rahman and Staveley, 1978; Boerio-Goates and Westrum, 1980; Fig. 16.7), These showed a small -type anomaly at 155K, with the enthalpy and entropy of the transition calculated as 222 J/mol and 1.34 J/molK by BoerioGoates and Westrum. There is another transition at 353K about which very little is known. In a much more detailed study (Herbstein, Marsh and Samson, 1993; HMS93; Herbstein and Samson, 1993; HS93) the earlier structural results were confirmed, the structure of PYRPMA at 19K was determined and cell dimensions measured as a function of temperature, as well as intensities of some superlattice reflections. No hysteresis of intensities or cell dimensions was encountered. The experimental measurements were analysed using Ehrenfest’s criteria for determining the order of a transition and then, more quantitatively, in terms of Landau theory, using the temperature dependence of


spontaneous strain and superlattice intensities to determine the nature of the transition. The excellent agreement obtained for these two parameters between theory and experiment was then checked against the excess specific heat, but here agreement was incomplete. Detailed studies by resonance techniques and measurements of elastic constants as functions of temperature are lacking for PYRPMA. (b) Crystal structure of PYRPMA in ordered and disordered phases. The earlier x-ray studies on PYRPMA (HS69) showed that there were mixed stacks (stack axis [001]) of alternating pyrene and PMDA molecules; in the 295K structure, both moieties were located at crystallographic centres (space group P21/a, Z ¼ 2). The c axis was found to have doubled on cooling to 110K and the space group changed to P21/n. There were now two pairs of pyrenes at independent centres, with the four PMDA molecules in the cell at general positions, but only slightly displaced from their 295K positions. The transition was single crystal to single crystal, with conservation of axial directions. The crystal structure of PYRPMA at 295K has been redetermined by Allen, Boeyens and Levendis (1989) and described in terms of a model in which the pyrenes are statically disordered over three orientations. (c) Experimental study of the order–disorder transition. The most striking feature of the cell dimension–T curves (Fig. 16.9) is the expansion of b on cooling from 300 to 165K, followed by the more usual contraction on further cooling. b – T and b–T curve have a cusps at 165K; a–T and c–T curves show only changes of slope in this region. As V ¼ abc sin b, the V–T curve must also show a cusp at Tc. The cell dimensions given in Fig. 16.9 for the disordered phase were extrapolated (‘‘by eye’’) into the region of stability of the ordered phase for use in the calculation of the spontaneous strain. Superlattice reflections, corresponding to a doubling of the c axis (i.e. those with l odd in the ordered structure), appear below 165K (Fig. 16.11). A smooth increase of relative superlattice intensities (IT/I0)hkl from zero at Tc to unity at T ¼ 0K, all (IT/I0)hkl showing the same behavior, is characteristic of the occurrence of a second order phase transformation (Cowley, 1980). The behavior of the superlattice intensities on cooling is quite different from the usual temperature dependence of Bragg reflections; these may increase or decrease as the crystal is cooled, in accord with small changes in atomic positions, but uniform behavior is not found (the different behavior of fundamental and superlattice intensities is illustrated below for ANTCYB (Fig. 16.25)). (d) The Ehrenfest order of the transition. The following items of evidence taken together indicate that the transition is of the second order as defined by Ehrenfest. The V–T curve (Figs. 16.9 and 16.10) is continuous but has a change in slope at Tc ¼ 160K; all three thermal expansion coefficients (not shown) have discontinuities at 165K; there are no superheating or supercooling effects (i.e. no hysteresis); the specific heat CP has a peak (Fig. 16.7) but not a discontinuity. The detailed behaviour of the cell dimensions (Fig. 16.9) suggests that the real physical situation is more complicated than that implied by the formal definition. (e) Application of Landau theory. Analysis following Stokes and Hatch (1988; see p. 1–10) gives many possible subgroups G when the high-symmetry space group (G0)


1100

(i) sin

b

(ii)

sin

b

5Å carbon oxygen

Fig. 16.8. Crystal structure of {pyrene PMDA}, showing the projections down [001] in (i) the low-temperature ordered structure (ii) the room-temperature disordered structure. In (i) the atoms of the PMDA molecules near c/4 are represented by shaded circles and those near 3c/4 by open circles; ˚ (at 19K). The length of [001] is 7.3 A ˚ (at 300K). For clarity two the length of [001] is 14.4 A molecules at each corner of the projection have been omitted in (i) while in (ii) one pyrene and one PMDA molecule have been omitted; hydrogens omitted throughout for clarity. (Reproduced from Herbstein and Snyman, 1969.) n n n

is P21/c (Table 16.3). The description used here has maintained analogous cell axes for both phases with consequent differences in the space groups, whereas Stokes and Hatch follow the converse path. On our basis one of the possibilities for the space group of the low temperature ordered phase is P21/n with one axis doubled as reported by HS69 and HS93. The physically irreducible representations for this transition is Y2þ and


13.9

TC a (Å)

13.8 13.7

9.34 7.31

9.32

b (Å)

9.30

c (Å)

7.27

9.28 93.5 7.23 93.1 (deg)

950

92.7 92.3

940

V (Å3)

91.9

930 920

TC 0

300

100 200 T (K)

Fig. 16.9. a(T ) plotted against T(K); there is a change of slope in the region of Tc (165K) but no cusp. b(T ) plotted against T(K); there is a cusp at Tc. c(T) and c(T)/2 plotted against T(K). The ordinate values below Tc must be doubled. There are indications of a change of slope at Tc (167K). b(T) plotted against T(K). There is a cusp at Tc. V(T) (or V(T)/2) plotted against T(K). There is a cusp at Tc (160K). (Reproduced from HS93.)

Volume per formula unit (cubic Å)

466 Disordered phase (dV/dT )P = 0.065 Å3/K

465 Tc Ordered phase (dV/dT)P = 0.111 Å3/K 464 150

155

160 T (K)

165

170

Fig. 16.10. The variation in (volume/formula unit) of the ordered and disordered phases in the vicinity of Tc (¼160K), as calculated from quadratic expressions fitted to the cell dimension curves. (Adapted from HS93.)


1102

1.2 –441 –521 –131

1.0

IT I0

0.016 0.014

0.6 0.4

0.012 0.2

es

0.010 0.008

0.0

Normalized intensity

Spontaneous strain

0.8

0.006 0.004 0.002 0.000

Tc 0

50

100 T (K)

150

200

Fig. 16.11. Upper graph – variation of the normalized intensities (IT/I0) of superlattice reflections as a function of temperature (I0 was estimated for each reflection from a smooth extrapolation of the I–T curve to T ¼ 0K). Lower graph – the spontaneous strain in the ordered phase of PYRPMA plotted against temperature. The curves shown are guides to the eye. (Reproduced from HS93.)

a continuous change is permitted by both Landau and Lifshitz conditions and also under renormalization group analysis. The relationship between the space groups was discussed in HS69. The usual method of deriving the critical exponents depends on fitting a power law equation of the type given above with experimental values measured as close as possible to the critical temperature. As the measurements for PYRPMA covered a wide temperature range but were sparse close to Tc, the Salje extension of classical Landau theory to large values of the order parameter was used. The (scalar) spontaneous strain es ¼ (ej2)1/2 in the ordered phase calculated from the Vogt coefficients ej as a function of temperature is proportional to Q2, as is the superlattice intensity. Both curves have similar shapes (Fig. 16.11). If the phase transition is tricritical (the intermediate stage between continuous and discontinuous transitions), then it follows from Landau theory that the fourth power of the order parameter is linearly proportional to T; this is indeed found in the plot of (es)2 against T (Fig. 16.12). A further test of the tricritical nature of the transition can be made using the intensities of the superlattice reflections. The squares of the normalized intensities against T give a linear plot (Fig. 16.12). Finally dependence of the excess specific heat on temperature (Fig. 16.7 (insert)) was tested for compatibility with a tricritical transition by plotting log (CP) against log (Tc– T ), where Tc ¼ 155K. The log–log plot (Fig. 16.13) shows that CP as a function of (1 – T/ Tc) is not well represented by the power law equation CP ¼ A[1–(T/Tc)]. A forced linear fit gives ¼ 0.77, whereas ¼ 0.5 for a tricritical transition (Salje (1990, p. 120)).

1.2

– I 2N (44 1) – I 2N (521) –– I 2N (1 3 1)

2.0

1.0 0.8 0.6

1.5

0.4 0.2

1.0

0.0

–104e 2S

(Normalized intensity)2

(Spontaneous strain)2


0.5

0.0

0

50

100 T (K)

150

200

Fig. 16.12. Upper graph – the square of the normalised intensity (IT/I0)2 plotted against temperature for three superlattice reflections of the ordered phase of PYRPMA. The origin has been moved up for clarity. Lower graph – the square of the spontaneous strain (es)2 in the ordered phase of PYRPMA plotted against temperature. (Reproduced from HS93.)

1.6

log (∆Cp (J/mol K))

1.4 1.2 1 0.8 0.6 0.4 0.2 0 –0.2 –1 –0.5

0 0.5 1 log (Tc–T (K))

1.5

2

Fig. 16.13. Log–log plot of the excess heat capacity (Cp in Fig. 16.7) against deviation from critical temperature in the vicinity of the phase transition. (Reproduced from HS93.)

A problem with this calculation is that the two independent sets of CP measurements agree well over the range 0–300K except in the region of the peak. A standard calculation of the entropy of transition gives 2.87 J/molK (as there are two orientations, but for only one of the components, a specific heat anomaly of 1/2R ln 2 would be predicted); this is about twice the measured entropy of transition of 1.34 J/molK.

1104


The theory of an almost tricritical phase transition with a simple one-component order parameter indicates that the power-law exponents b, obtained from the temperature dependence of the spontaneous strain and the intensities of the superlattice reflections, should be 0.25 and , obtained from (CP), should be 0.5 (Salje, 1990). Good agreement with experiment was found for the first of these predictions but not for the second. There is a physical difference in that the diffraction measurements cover the whole temperature range below Tc while (Cp) is restricted to between Tc–35 and Tc. (f) Crystal structures of disordered and ordered phases. One way of describing the process in which the long range order of the ordered phase is decreased is by comparing ordered and disordered unit cells through their cell dimensions and molecular arrangements. Perhaps a natural approach is to note that the stack axis ([001]) doubles at Tc and to infer that one is dealing with a Peierls phenomenon,4 due to some change in the -electron HOMO–LUMO interaction between donor pyrene and acceptor PMDA molecules. However, this would appear to be ruled out by the fact that the [001] axial length changes smoothly with temperature. The most striking change in cell dimension behaviour at Tc is that in [010], where (abnormal) expansion on cooling (of the disordered phase) changes to (normal) contraction (of the ordered phase). This could imply that the interaction between adjacent stacks in the direction of [010] is normal from 0–165K, and then becomes (in some unspecified way) abnormal when the long range order along the stack axis disappears. Expansion on cooling is unusual but not unprecedented; however, neither NAPTCB nor ANTCYB behave in this way. More definite, but still tentative, indications were inferred from an analysis of the molecular arrangement at 19K (Herbstein, 1996, which should be consulted for details). There is as yet no consensus about the nature of the driving force for ordering. Wideline NMR studies have been made of {pyrene PMDA} (Fyfe, 1974a, b). Line width and (H)2 begin to fall at 195K, about 30 above Tc but there is no sign of any change in line width or (H)2 at Tc itself nor at 353K where there is a first-order transition to a phase of unknown structure. The measured value of (H)2 is 4.9 G2 at 77K, which is about 20% smaller than the calculated value for a rigid lattice based on the crystal structure, suggesting that there is still appreciable reorientational motion below Tc. A similar lack of correlation between wide-line NMR results and studies of solid-state transitions appears also for other systems discussed here. The explanation (Darmon and Brot, 1967) is that the proportion of molecules jumping from one orientation to another at any given instant is very small and the contribution of these molecules to the thermodynamic functions is essentially negligible. The values of (H)2 at high temperatures (2.0–2.4 G2 at 300–420K) are explicable (but still not entirely) only if reorientation of the pyrene molecules takes place by both small angle (20 ) and large angle (160 ) in-plane jumps. Measurements of spin-lattice relaxation times (Fyfe et al., 1976) give an activation energy of 57 kJ/mol for the large angle process; Ea could not be measured for the small angle jumps. n n n

4 The Peierls phenomenon (Peierls, 1954, see Chapter 5) refers to the doubling of the periodicity of an . . . ABABAB . . . stack at low temperatures because the ‘dimerized’ arrangement . . . AB AB AB . . . has lower energy.


16.5.2 The crystal structure of {Naphthalene TCNB} and evidence for an order , disorder phase transition around 72K n n n

(a) Introduction. The order-disorder transitions in the -molecular compounds {naphthalene TCNB} (NAPTCB) and {anthracene TCNB} (ANTCYB) are rather similar (Lefebvre, Odou, Muller, Mierzejewski and Luty, 1989 (LOMML89); Ripmeester, (1995)). We first discuss NAPTCB and then ANTCYB; comparisons are also made with the (somewhat different) behavior of PYRPMA. The order–disorder behavior in NAPTCB and ANTCYB is similar in the sense that the same space group changes occur, but Tc values are appreciably different at 76 and 212K (Fig. 16.14). n n n

n n n

CP (J/mol K)

(b) Crystallographic background and x-ray diffraction measurements of temperature dependence of degree of long range order. Cell dimensions have been measured by XRD at 294, 95 and 65K (LOMML89) and by neutron diffraction over the range 300–15K

125 75 K 100

NAPTCB

75 40

50

60 T (K)

70

80

CP (J/mol K)

350

300 ANTCYB 250 150

170

190 T (K)

210

230

Fig. 16.14. (Upper panel) the calorimetrically measured specific heat of NAPTCB in the region of the order–disorder transition. The CP–T curve is smooth over the rest of the temperature range 0–300K. (Lower panel) the specific heat of ANTCYB in the region of the order–disorder transition, measured by differential scanning calorimetry (DSC). Ecolivet et al. (1988) suggest that there may be a double peak (shown by arrows) in the DSC curve, corresponding to two events, but this remains to be proved. (Reproduced from Ecolivet, Lemeé, Delugeard, Girard, Bertault, Collet and Mierzejewski, 1988.)

1106


(Czarniecka et al., 1985) (Fig. 16.15). There are systematic differences between XRD and ND values and closer temperature intervals are required before definite inferences can be drawn about the order of the transition; the indications are that there is no discontinuity in cell volume at Tc, in accordance with a second order character for the transition. Calculation of the temperature dependence of the spontaneous strain requires a better set of cell dimensions. The NAPTCB and ANTCYB molecular compounds have closely related structures but are not isomorphous. An early room temperature structure analysis on a twinned crystal of NAPTCB (Kumakura, Iwasaki and Saito, 1967) showed that the components were arranged in the usual donor–acceptor mixed stacks, with stack axes along [001] (Fig. 16.16). These results were confirmed (and extended) in a later single crystal study of the structures at 300, 95 and 65K (LOMML89). The same formal description applies to NAPTCB and ANTCYB although the arrangement of the molecules, and their relationship to the cell axes, is different for the two compounds; one structural difference is that in NAPTCB the molecules lie very nearly in (102) planes, but in ANTCYB in ð102Þ planes (Fig. 16.17). NAPTCB TC

9.4 9.3 a (Å) 9.2

12.7 12.6 b (Å) 12.5 6.9 6.8 c (Å)

6.7

(deg)

108.0 107.5

780

V (Å3)

770 760 750 740 TC 0

100

200 T (K)

300

Fig. 16.15. NAPTCB – cell dimensions as a function of temperature, measured by XRD (centered circles, LOMML89) and ND (solid circles, Czarniecka et al., 1985).


At 300K the naphthalene molecules are disordered in equal measure over two orientations separated by 36 , while the TCNB molecules are completely ordered. On cooling, the space group changes from C2/m in the disordered phase to P21/a in the ordered phase, without appreciable change in unit cell dimensions. In the C2/m, Z ¼ 2 structure, the naphthalene centres are at Wyckoff positions (a) (000, 1/2 1/2 0) and TCNB centres at Wyckoff positions (c) (00 1/2, 1/2 1/2 1/2); the site symmetry of (a) and (c) positions is 2/m. In the P21/a, Z ¼ 2 structure, the naphthalene centres are at Wyckoff positions (a) (000, 1/2 1/2 0) and the TCNB centres at Wyckoff positions (b) (001/2, 1/2 1/2 1/2); the site symmetry of (a) and (b) positions is 1. The donor molecules take up one (averaged) orientation above Tc and two below; the TCNBs do not change. Above Tc the two stacks

N

N

N

N

a sin

O

N

N

N

N

b

N

N

N

N

N

N

N

N

N

N

N

N

Fig. 16.16. The ordered structure of NAPTCB, space group P21/a (hydrogens omitted for clarity). In the diagram the centers of symmetry at cell corners, centers of cell edges and at cell center have been omitted. The site symmetry is for both components. The naphthalenes are centred at Wyckoff positions (a) [000; 1/2,1/2,0] and the TCNB’s at Wyckoff positions (b) [00,1/2; 1/2,1/2,1/2]. The molecules lie nearly in ð102Þ planes. The disordered structure (space group C2/m) has essentially the same cell dimensions, but the units at cell corners and center are now identical (i.e. related by translations). The naphthalene molecule is disordered over the orientation shown and its mirror image, with equal occupancies. The site symmetry is 2/m for both components, the two fold axis running along [010] and the mirror plane lying in (010). (Adapted from Kumakura et al., 1967.)


1108 (a)

III

(b) II I

I II III D

c

A

c a

D

A b down a Stack axis Stack axis NAPTCB

ANTCYB

˚, Fig. 16.17. Comparison of the stack arrangements in NAPTCB (at 300K: 9.420 12.684 7.31A ˚ 107.4 , Z ¼ 2, C2/m) and ANTCYB (at 300K: 9.526 12.780 7.440 A, 92.36 , Z ¼ 2, C2/m), with the donor molecules (respectively naphthalene and anthracene) labelled as D and denoted by thick lines, while the acceptor molecules (TCNB) are labelled A and denoted by hatched rectangles. Stacks I and III are translationally equivalent, while stack II is shifted by b/2 behind the plane of the page and is related to I by the a-glide. (Adapted from LOMML89.)

(stack axis [001]) in the unit cell are related by the C-centering operation while below Tc the two stacks are related by the a-glide. The structure can be described as a close packed arrangement of stacks of elliptical cross section, with the packing of the stacks shown in Fig. 16.16. Within a particular stack, all donor molecules are translationally equivalent, as are all acceptor molecules; this applies to both disordered and ordered structures. In the disordered structure the arrangement of components along a stack is . . . ..hDi A hDi A hDi A hDi . . . where hDi represents the averaged orientation of the donor molecules in a stack. In the fully ordered structure the arrangement of components along stack I (Fig. 16.16) is . . . ..D1 A D1 A D1 A D1 . . . and that in stack II is . . . D2 A D2 A D2 A D2 . . . where D1 and D2 are the two orientations of the donor molecules, related by the a-glide; the TCNBs have the same orientations in both stacks but this is not required by the crystal symmetry. Thus the disordering process, viewed from ‘‘within stack I’’, is to produce sections of . . . D1 A D2 A D1 A D1 . . . and analogously for stack II. Analysis following Stokes and Hatch (1988; see p. 1.8–9) gives many possibilities for a contiuous transition from the high symmetry (G0) space group C2/m; one of these is to P21/a, without change of cell dimensions or shift of origin, the physically irreducible


I(323) (arbitary units)

10

1

0.1 0.0010

0.0100

0.1000

1.0000

(1–T/73.5)

Fig. 16.18. I(323) against (1 – T/73.5) for NAPTCB, both on a logarithmic scale; the intensity measurements (in arbitary units) extend over the range 60–75K. Apart from the two points closest to Tc, the measurements are well fitted by I(323) 1 (1 – T/73.5)0.334.

representation being Y2þ and a continuous change being permitted by both Landau and Lifshitz conditions and also under renormalization group analysis. The relationship between the space groups is discussed above. The critical exponent can be determined from the temperature dependence of the intensities of the superlattice reflections; the necessary measurements (by LOMML89) are available only for reflection (323). We have noted above (equation (16.2)) that Q2 Isuperlattice ¼ A(1–T/Tc)2b The log/log plot (Fig. 16.18) shows that the exponent 2b ¼ 0.334 except in the region closest to Tc, where a higher value could be appropriate. b ¼ 0.33 is generally considered to indicate an order–disorder transition so that there is a considerable discrepancy here, perhaps due to the fact that I(323) was measured over a range of only 13.5K (compared to ranges of 145K for the analogous measurements for PYRPMA and ANTCYB). The values of the order parameter Q at 65K from various sources can be compared as follows: an extrapolated value of I(323) ¼ 9 at 0K is obtained from Fig. 16.18, with I(323) ¼ 4.3 at 65K (from Fig. 6 of LOMML89). Thus Q(65K) ¼ (4.3/9)1/2 ¼ 0.7, which is not in good agreement with the NMR value (essentially unity) or that of LOMML89 from XRD (Q(65K) ¼ 0.90). The diffraction measurements of superlattice intensities at different temperatures for ANTCYB are more extensive than those for NAPTCB and we discuss extraction of order parameters and critical exponents below.

1110


(c) Other physical measurements. Specific heat measurements (Boerio-Goates, Westrum and Fyfe, 1978) (Fig. 16.14(a)) show that there is a gradual transition over the range 40–75K (Htrans ¼ 192 21 J/mol, Strans ¼ 3 0.3 J/mol K); Htrans is not very different from the value found for the transition in PYRPMA (222 J/mol) but the Strans values differ by a factor of 2 because of the difference in Tc values. On the basis of the structural results given above, we note that ordering of the molecules below Tc would give a specific heat anomaly of 1/2R ln 2, because there are two orientations, but only for one of the components. This prediction (2.87 J/molK) is in excellent agreement with the measured entropy of transition. Unfortunately the same argument applied to PYRPMA resulted in a discrepancy of 100%! The anomalous regions in the CP–T curves of NAPTCB and ANTCYB have strikingly similar shapes that differ from the classical shape found in PYRPMA. The significance of these differences is not clear. The occurrence of a phase transition was confirmed by Raman spectroscopy which placed Tc at 69K and 62K for the C10H8 and C10D8 molecular compounds respectively (Bernstein, Dalal, Murphy, Reddoch, Sunder and Williams, 1978). The donor molecule (labelled D in Fig. 16.17) at 000 (for NAPTCB) or 100 (for ANTCYB) impinges on the TCNB molecule at 1,0,1/2 (labelled A in Fig. 16.17); the H . . . N interactions between these two components are responsible for the ordering. This is illustrated for NAPTCB at 65K in Fig. 16.19. There are three N . . . H–(C) interactions at the ¼ 0 equilibrium position while there are only two such interactions if the naphthalene is rotated in its own plane to ¼ 36 . At both these positions the N . . . H nonbonded distances are close to the sum of the van der Waals radii. If the naphthalene molecule took up an intermediate orientation at ¼ 18 , ˚ less than the sum of the van the closest N . . . H nonbonded distances would be 0.2–0.3 A der Waals distances. This suggests that the transition is driven by intermolecular interactions (‘‘packing effects’’), in accordance with the lowering of Tc when hydrogen is replaced by deuterium (Bernstein, Dalal, Murphy, Reddoch, Sunder and Williams, 1978; Dalal, Ripmeester et al., 1978). These qualitative considerations can be rendered quantitative by theoretical calculations and the results of measurements by various resonance techniques. A double well potential energy curve for naphthalene in the field of the surrounding TCNB molecules was calculated by Shmueli and Goldberg (1973) with barrier height about 12kJ/mol and minima separated by 36 , in accordance with the 300K crystal structure; naphthalene to TCNB interactions were found to be the most important.

2.

68

A

2.

99

5

A

2

2.774 A

) plane of {naphthalene TCNB} (almost the plane Fig. 16.19. Two adjacent molecules in the (102 of the molecules), showing the closer N . . . H distances in the ordered crystal structure. (Reproduced from Lefebvre et al., 1989.) n n n


A converse, quasi-experimental, analysis was carried out by LOMML89. They calculated structure factors for the disordered structure at 294 and 95K in terms of an orientational probability function for the disordered molecule. For NAPTCB the best fit between observed and calculated structure factors was obtained when the orientational probability function had two maxima at 16 (at 294K) or 18.5 (at 95K). An angle of 36 between the energy minima was measured by 2H NMR spectroscopy (Ripmeester, 1982). Thus these three approaches agree in describing the disordered structure as statically disordered, with the two orientations separated by a barrier height of 12kJ/mol. Earlier wide line (Fyfe, 1974a, b) and pulsed NMR measurements (Fyfe, Harold-Smith and Ripmeester, 1976) have since been extended to lower temperatures, and supplemented by ESR measurements (Ripmeester, Reddoch and Dalal, 1981). We consider all these results together. Pulsed NMR measurements of relaxation times (Fig. 16.20; T1, the spin– lattice relaxation time, and T1 the spin–lattice relaxation time in the rotating frame) were used to obtain activation energies of 9.4 (Ea) and 42.6 kJ/mol (Eb) for small and large angle reorientations respectively, to be compared with calculated values of 8 and 46 kJ/mol (Fyfe, Harold-Smith and Ripmeester, 1976); only T1 changes noticeably at Tc. The spin–lattice relaxation times (Fig. 16.20) were also used to estimate E (the energy difference between sites 1 and 2) as a function of temperature (Ripmeester, Ratcliffe et al., 102 Naphthalene – TCNB Tc

101 TI 10 MHz TI 25.3 MHz TIr 25.3 MHz HI = 22 G

TI 25.3 MHz 100 TI/sec TIr /sec 10–1 TIr 25.3 MHz H1 = 10 G 10–2

1000K /T 10–3 2

4

6

8

10

12

14

16

18

Fig. 16.20. Static (T1) and rotating frame (T1) 1H spin-lattice relaxation times measured for NAPTCB. (Spin–lattice relaxation times vs. 1000/T.) Dashed lines represent measurements from Fyfe, Harold-Smith and Ripmeester, 1976. (Reproduced from Ripmeester, Dalal and Reddoch, 1981.)


1112

1995). These values increase from 1.5 kJ/mol to 7.15 kJ/mol over the range 71.5–63K, and can be fitted to E ¼ 6.72T þ 49.54 (R2 ¼ 0.99) (Fig. 16.20). Extrapolation gives Tc ¼ 73.7K (E ¼ 0), in good agreement with the value obtained from temperature of disappearance of the 323 superlattice reflection (Fig. 16.18; 73.5K; LOMML89). The anomalous increase of the specific heat occurs (on cooling) at 75K (Fig. 16.14(a)). Linear extrapolation, which seems inherently unlikely, would give E ¼ 49.5 kJ/mol at T ¼ 0K. The ratio of the occupancies of the two sites 1 and 2 is given by p2/p1 ¼ exp(E/RT) and can be calculated from the E–T dependence. At 71.4K, p1 ¼ 7.2%, falling to 0.2% at 68.6K and is essentially zero (i.e. complete ordering) at lower temperatures. LOMML89 measured p1 ¼ 4.8% from their 65K crystal structure analysis, using the heights of residual (difference) electron density peaks in the alternative orientation; possibly their temperature was slightly underestimated. A diagram of potential energy as a function of orientation of the naphthalene molecule in the unit cell is shown in Fig. 16.22 for the ordered and disordered phases. The correlation times (average time between instantaneous 36 jumps) have been measured by NMR (Ripmeester, Dalal and Reddoch, 1981), ESR (Grupp, Wolf and Schmid, 1982) and incoherent quasi-elastic neutron scattering (QNS) (Czarniecka et al., 1985). The values obtained by the different methods for the disordered structure all fall on the same straight line in a plot of log against 1000/T (Fig. 16.23). However, a distinct break occurs at Tc. The QNS measurements give an average residence time at room temperature of 8 ps. Density of states curves for the components and the molecular compound have been obtained by incoherent inelastic neutron scattering. The 15K ESR measurements of Erdle and Mo¨hwald (1979) showed that the triplet state was largely localized on naphthalene (Zt0.25 0.05) and that the normal to

8 ∆E

7

∆E (kJ/mol)

6 5 4 3 2 1 0 60

62

64

66

68

70

72

T (K)

Fig. 16.21. E (the energy difference between favorable and unfavorable orientations of naphthalene (sites 1 and 2) in the ordered phase) as a function of temperature for NAPTCB; these values were redrawn from those given by Ripmeester, Ratcliffe et al. (1995).


50 Energy (kJ/mol)

40

L

30

50

20

40

10 Ea

30 20

Disordered phase M

Eb

Ordered phase (~63 K) 1

2

10 –60

0

60 120 180 240 300 Clockwise angle of rotation of M axis from a sin direction (deg.)

360

Fig. 16.22. Potential for the reorientation of the naphthalene molecule in the disordered and ordered phases of NAPTCB; Ea, Eb and E are defined in the text. is 36 . In the disordered phase, a particular naphthalene molecule is located in the isoenergetic (and hence equally populated) minima about 0 (and 180 ) and separated by . In the ordered phase these two minima no longer have equal energies (energy difference as function of temperature shown in Fig. 16.21) and one is favored and the other disfavored. (Adapted from Ripmeester, Ratcliffe et al. (1995).)

1 × 10–6 QNS ESR NMR

1 × 10–7 1 × 10–8 tau (sec) 1 × 10–9 1 × 10–10 1 × 10–11 1 × 10–12

2

4

6

8

10 12

14

16

1000/T (K–1)

Fig. 16.23. Correlation time as a function of temperature for NAPTCB. NMR measurements from Ripmeester, Dalal and Reddoch, 1981; ESR from Grupp, Wolf and Schmid, 1982; QNS from Czarniecka et al., 1985. The latter authors note . . . ‘‘we conclude that the correlation times determined by NMR and QNS methods corroborate each other, and we are inclined not to consider the difference in slopes of the NMR and QNS results as real..’’ Arrow marks Tc. (Adapted from Czarniecka et al., 1985.)

1114


the plane of the triplet state naphthalene molecules was tilted by 10 with respect to [001]. However, this tilt angle was less than 2 in the 65K structure. This indicates that formation of the excited state leads to a geometrical change, a phenomenon for which there is considerable evidence from other sources (Ponte Goncalves, 1977). (d) The phase transition in summary. A rather complete picture emerges when the results from the various experimental methods are concatenated. The transition can be inferred to be second order (in agreement with LOMML89) from the form of the specific heat curve (Fig. 16.14(a)) and from the absence of a discontinuity in cell volume at Tc (¼73.5K) (Fig. 16.15). At very low temperatures the naphthalene and TCNB molecules are completely ordered in space group P21/a, each moiety being located at a crystallographic centre of symmetry. The long axes of the two naphthalenes are 36 apart and their mean molecular planes are mutually tilted by 2 . The TCNB molecules are so located that they are all parallel, although this is not required by the space group. The temperature dependence of the intensity of a superlattice reflection indicates that the long range order begins to fall below its 0K value of unity as the crystal is heated. There is some disagreement about the temperature at which this becomes appreciable, one interpretation of the XRD measurements suggesting fairly low temperatures, while another suggests 65K, while NMR relaxation times indicate that this process occurs even closer to the critical temperature of 73K. The average energy difference between a naphthalene molecule favorably and unfavorably oriented at an inversion centre is 7kJ/mol at 63K, falling to zero at Tc and rising to some tens of kJ/mol at 0K. There is a second-order ‘order to disorder’ transition at 73K, accompanied by an anomalous additional specific heat corresponding to a measured entropy of transition of 3 0.3 J/mol K, which is very close to the expected value of 1/2R ln 2. As the temperature approaches Tc from below, the probability of finding an unfavorably oriented naphthalene molecule increases. Unfavorably oriented molecules lead to a decrease in the energy difference between sites with correctly and unfavorably oriented molecules, and this cooperative effect leads to complete loss of long range order as Tc is reached. The driving force for disordering on heating is the small energy difference between correctly and unfavourably oriented molecules in a particular lattice site, which decreases as the temperature approaches Tc from below. Above Tc the space group changes to C2/m, without change in cell dimensions. The site symmetry of the sites occupied by naphthalene and TCNB molecules increases to 2/m. This is achieved for naphthalene by statistical occupation of a particular site by the molecule in two orientations separated by 36 , while the orientation of the TCNB molecule (which hardly changes from that in the ordered structure) is now fixed by the 2/m symmetry of its site. Above Tc the potential barrier between the two naphthalene orientations, which are of equal energy, is 8.0 kJ/mol. Nothing seems to be known about possible existence of short range order above Tc – in other words, the diffuse scattering in the diffraction pattern requires investigation. One anomaly awaits explanation – the pressure dependence of Tc. One can make an approximate calculation using the results given in Figs. 16.14(a) and 16.15 (‘approximate’ because of the nature of the experimental evidence) similar to that carried out for PYRPMA and this gives dTc/dP ¼ 31 K/kbar, contrary to the expectations from Fig. 16.6. No explanation has been offered.


16.5.3 The crystal structure of {Anthracene TCNB} and evidence for an order , disorder phase transition at 213K n n n

(a) Introduction. The molecular compound {anthracene TCNB} (ANTCYB) has been extensively studied. The specific heat has been measured by DSC (Fig. 16.14(b); Ecolivet, Bertault, Mierzejewski and Collet, 1987) and the anomalous region is remarkably similar in shape to that found for NAPTCB and shown in Fig. 16.14(a); Htrans ¼ 150 J/mol and Strans ¼ 0.7 J/mol K. The enthalpy of transition is similar to that found for NAPTCB (192 21 J/mol) but the entropy of transition is smaller by a factor of 4, again (as in PYRPMA) a reflection of the higher transition temperature. A Brillouin scattering study of the elastic anomalies at the phase transition confirms that Tc ¼ 212 0.5K. This study also allowed determination, in the ordered phase, of the order parameter relaxation time as [5 1011/(Tc T)] s K1 (Ecolivet and Mierzejewski, 1990). n n n

(b) Crystallographic background. The crystal structure at 300K was first determined by Tsuchiya, Marumo and Saito (1972) and structures above and below Tc first by Stezowski (1980; crystal structures at 297, 234, 226, 202, 170, 138 and 119K), and then by LOMML89 (crystal structures at 294, 225 and 65K). There is a comprehensive series of cell dimension – temperature measurements for ANTCYB that show very weak cusps at Tc but no discontinuities (Fig. 16.24). These results show that the spontaneous strain in the ordered phase is almost entirely in the [100] direction. The disordered structure is isostructural with that of NAPTCB but the two ordered structures differ in the gross sense (as noted above) that the naphthalene molecules are in (102) planes but the anthracene molecules in ( 102) planes, and in a subtle sense in that the TCNBs in ANTCYB change their orientations slightly (by 2 ) on ordering but not in NAPTCB. Stezowski used his intensity measurements to calculate molecular libration amplitudes using the rigid body model. Electron density plots at the different temperatures were also calculated, from which the angles between various axes of symmetry-related molecules were obtained; these angles are in good agreement with values deduced from ESR measurements by Mo¨hwald, Erdle and Thaer (1978). However, this approach seems to ignore the essential order-disorder features of the system, as pointed out by Park and Reddoch (1981); in particular, the electron density plots and the (inferred) libration amplitudes will contain contributions from both thermal vibrations and disorder. The LOMML89 and Stezowski measurements of the superlattice (h þ k odd) reflection (140) and the fundamental reflection (240) are in excellent agreement; the different modes of temperature dependence of these different reflection types are shown in Fig, 16.25. A check on whether the superlattice reflections all show the same dependence of Fobs on T, as required by theory, shows a considerable spread and more closely spaced measurements like those made for (140) appear to be required. (c) ESR measurements of the order parameter as a function of temperature. The temperature dependence of the order parameter has also been determined by ESR and can be compared with the diffraction measurements. Measurements on the ESR spectra of the triplet excitons produced by optical irradiation of ANTCYB in its charge transfer band give the temperature dependence of the long range order parameter. Below Tc, for a


1116

ANTCYB TC a (Å)

9.50

12.75 7.45

b (Å) 12.70

7.40 12.65

c (Å) 7.35 7.30

93.0 (deg)

92.5

900 V (Å)

890

TC

880 870 50

100 150 200 250 300 T (K)

Fig. 16.24. ANTCYB – cell dimensions as a function of temperature, including 65K values (centered circles), are from LOMML89. The cell dimensions given by Stezowski (1980) at 297, 234, ˚ , 0.2 ) with those of 226, 202, 170, 138 and 119K are in satisfactory agreement (to within 0.01 A LOMML89. The curves are guides to the eye. 350 300 (240) fundamental

10 F (obs)

250 200

(140) superlattice

150 100 50 0 0

50

100

150 200 T (K)

250

300

Fig. 16.25. ANTCYB: Dependence of Fobs of the 140 superlattice and 240 fundamental reflections on temperature. The measured values are from LOMML89 and Stezowski (1980). The curves are guides to the eye.


general orientation, the ESR spectrum for each of the ms ¼ 1 transitions consists of two lines, corresponding to the fine structure tensors of the two sublattices (which are related, as noted earlier, by the a glide plane). As the temperature is raised, these two lines approach and broaden, coalescing near Tc into a single line that becomes narrower at higher temperatures. The separation between the two lines gives the long range order parameter S ¼ S0(1 – 2p), where S0 would be the separation for a perfectly ordered crystal, and p is the fraction of sites occupied by unfavorably oriented molecules. This experiment was first done for ANTCYB by Mo¨hwald, Erdle and Thaer (1978), who identified their order parameter with the angle between the long axes of the anthracene molecules. Park and Reddoch (1981) pointed out that this approach suffers from the same deficiencies as that of Stezowski. We quote from Park and Reddoch ‘‘An exciton moving rapidly along a chain [stack in our nomenclature] will then sample this disorder [of the orientations of the anthracene molecules along the stack]. In the fast limit the position of its resonance line will be the average of the lines for the two orientations, weighted by the relative populations within a given chain. Such a fast-limit resonance can thus shift with temperature, but need not broaden. If this fast moving exciton now jumps at a slow rate to a neighboring chain which is predominantly of opposite orientation, line broadening may be expected . . . One way to treat this problem is to consider a four-site exchange model with two rate constants and four sites, consisting of the two orientations in each of the two sublattices . . . The separation [of the lines] S is . . . a statistical average weighted by the number of correctly [xA] and incorrectly [xB] oriented anthracenes in the chain [(xA þ xB) ¼ 1]. Thus S ¼ S0 (xA – xB) ¼ S0(1 – 2xB) where S0 would be the separation for a perfectly ordered crystal.’’ Park and Reddoch derived xB (our p) from their experimental measurements as a function of 1/T and their values have been replotted as Q (¼S/S0) against T, to which we have added the XRD values of Q derived from reflection 140 (see Fig. 16.26, which should be compared with Fig. 16.25). There is excellent agreement between Q(ESR) and Q(XRD) (from 140) showing that both have the same functional dependence on T/Tc. We next plot log(IT/I0) against log(1 – T/Tc) (a form that allows direct comparison with the NAPTCB analog in Fig. 16.18) to test whether the power law equation (16.2) is applicable; Fig. 16.27 shows that a linear relation is not obtained. Park and Reddoch reported b ¼ 0.34 for the range 198–208.3K and made a comprehensive comparison of their experimental results with various theoretical treatments and came to the conclusion that the renormalization group calculation gave best agreement with their value for b. This is not surprising as there is evidence for the occurrence of fluctuations in the temperature region for which their value was derived. Park and Reddoch determined the rate constant for interchain hopping as !12 ¼ [1.6 107 þ 1 1011 exp(Ea/kT)] s1, where Ea ¼ 1051 cm1 ( ¼ 12.6 kJ/mol). The second term was tentatively interpreted as due to scattering of triplet excitons by anthracene molecules undergoing large amplitude vibrations. Ea is the barrier between A and B orientations. The corresponding activation energy in NAPTCB was given as 9.4 kJ/mol. (d) The phase transition in summary. The weight of the evidence suggests that the driving force for the phase transition comes from intermolecular packing interactions, and


1118

1.00 Long range order parameter

0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 0

50

100

150

200

250

T(K)

Fig. 16.26. ANTCYB – long range order parameter Q as a function of temperature. The open circles show values derived from the splitting of ESR lines and are replotted from Fig. 1 of Park Reddoch (1981); the dots show XRD values for Q(140) and are transferred from Fig. 16.25.

Q2

1.000

0.100

0.010 0.001

0.010

0.100

1.000

(1–T/215)

Fig. 16.27. ANTCYB – ESR and XRD values of Q2 plotted against (1T/Tc) (Tc ¼ 215K); the scales are logarithmic. The data are replotted from Fig. 16.26.

not from orientation dependence of charge transfer interactions. One important indication comes from the lowering of Tc in ANTCYB with increasing deuteration of anthracene (h10 – 214.5 ( 0.5) K; b-d4 – 202.5; -d4 – 201.5; d10 – 198.5), that has been ascribed (Dalal, Haley, Northcott, Park, Reddoch, Ripmeester, Williams and Charlton, 1980) to reduction in intermolecular repulsions on deuteration (d(CD) < d(CH)). A similar effect found for NAPTCB has already been noted. There are resemblances between the NAPTCB and ANTCYB situations but details differ. At the two equilibrium positions separated by an in-plane rotation angle of 16 () the average N . . . H distance is close to the


sum of the van der Waals radii (in NAPTCB ¼ 36 , the average N . . . H distance again being close to the sum of the van der Waals radii). If the anthracene molecule took up an intermediate orientation at ¼ 8 , the closest N . . . .H non-bonded distances would be ˚ less than the sum of the van der Waals distances (0.28 A ˚ in NAPTCB). The 0.18 A ˚ difference of 0.1 A could lead one to expect dynamic disorder in ANTCYB instead of the static disorder found in NAPTCB. The potential energy curves calculated for an anthracene molecule librating in the field of its nearest neighbours have been found to show a broad single-well minimum rather than the double potential well noted above for NAPTCB. The orientational probability function determined by LOMML89 shows a single maximum at ¼ 0 . The differences between the two structures have led LOMML89 to suggest that the phase transition at 213K has elements of both a ‘displacive’ and a ‘disorder to order’ transition. In this sense ‘displacive’ means that the molecules (here only anthracenes) undergo a continuous change in orientation as the temperature is lowered below Tc (this is similar to the points of view expressed by Mo¨hwald and Stezowski) while ‘disorder to order’ means that the fraction of molecules in the correct orientation for a particular sub-lattice increases as T falls below Tc. The matter is not yet settled. A semiquantitative analysis of the pressure dependence of Tc can be carried out using equation (16.4). From Figs. 16.14(b) and 16.24, we calculate dTc/dP ¼ 80 K/kbar; a measured value of 32 K/kbar has been reported (Ecolivet, Lemeé, Delugeard, Girard, Bertault, Collet and Mierzejewski, 1988). The agreement seems reasonable considering the difficulties of interpreting the specific heat curve (Fig. 16.14(b)) and of inferring values of @V/@T from Fig. 16.24, where the slopes must be measured close to Tc and not averaged over the whole 0–300K temperature interval, which would give zero slope difference. We shall not discuss the many preliminary theoretical calculations made on ANTCYB but give two leading references (Brose, Luty and Eckhardt, 1990; Kuchta, Luty and Etters, 1990). 16.5.4 Other examples of second order transitions There is clear evidence for two orientations for anthracene in {anthracene TCNQ} at 300K (Williams and Wallwork, 1968). These crystals are isomorphous with those of the disordered structure of {naphthalene TCNB}, with [001] stack axis in both examples. It seems reasonable to infer that {anthracene TCNQ} will show a disorder-to-order transition on cooling, perhaps also to space group P21/a without appreciable change of cell dimensions. Heat capacity measurements on {pyrene TCNB} showed an approximately symmetrical anomaly in the range 220–250K; the details were difficult to reproduce even with samples of the same thermal history and there was facile supercooling of the 290K crystals down to at least 150K (Clayton et al., 1976). Thermal hysteresis of 40 has been reported and shattering of the crystals, on cooling through a transition at 183 3K. The crystal structures at 290 and 178K were found to be similar (Prout, Morley et al., 1973) and it is possible that supercooled crystals were used for the 178K diffraction measurements. The structure analysis showed the pyrene molecules to be azimuthally disordered and the TCNB molecules ordered; the available NMR measurements (Fyfe, 1974a, b) are not detailed enough for comparisons to be made n n n

n n n

n n n

n n n

1120


with the x-ray diffraction and Cp results. Further study is needed before conclusions can be drawn. There are also indications of pyrene disorder in {pyrene TCNQ} (Prout, Tickle and Wright, 1973). Low temperature phase transformations without change of crystal system have been demonstrated in {anthracene C6F6} and {pyrene C6F6} (see Bernstein (1967)) – these are probably disorder to order transitions. Measurements of reorientational motion of both components have been made in {pyrene C6F6}, using 19F in addition to proton resonance. The two molecules behave rather similarly. A decrease in the second moment of the 19 F line above 130K was attributed to onset of reorientation about the hexad axes of hexafluorobenzene, which would not be detectable by x-ray diffraction or Cp measurements; such reorientation has been detected by NMR in {benzene C6F6} (Gilson and McDowell, 1966). The crystal structures of two polymorphs (yellow and orange) of {naphthalene PMDA}, which are presumably monotropically related, have been reported (LeBarsCombe et al., 1979). The crystal structure shows that the yellow polymorph is ordered at 293K; specific heat and Raman scattering measurements (Macfarlane and Ushioda, 1977) show that there are no transitions in the range 2–300K. Donor and acceptor molecules have their long axes parallel over the whole temperature range; such an eclipsed mutual orientation is unusual. In the orange polymorph, for which only crystallographic results are available, two orientations (separated by 42 ) are found for the naphthalene molecules at a particular site, with unequal populations of 72 and 28%. Although one might expect a disorder to order transition on cooling (perhaps analogous to that in {pyrene PMDA), none has been found in the range 300–148K. n n n

n n n

n n n

n n n

n n n

n n n

n n n

16.6 16.6.1

Crystals with first order phase transformations on cooling {Cycl[3.2.2]azine TNB} n n n

The first order transformation in {cycl[3.2.2]azine TNB} (orange needles elongated along [001]) is monoclinic to monoclinic (Table 16.4), with clear discontinuities in the cell dimensions at T ¼ 143 3K (Fig. 16.28) (Hanson, 1978). The transition was n n n

˚, Table 16.4. {Cycl[3.3.2]azine TNB} – unit cell dimensions (A deg.) at 293 and 91K. Describing the 91K structure in terms of a non-standard B-centered cell shows that the a and c dimensions have approximately doubled as a result of the transformation n n n

a b c b ˚ 3) V(A Z Space group

B 293K

A 91K

14.37(1) 15.99(1) 6.682(4) 92.17(5) 1534.3 4 P21/n

centered 28.474(15) 15.636(8) 13.102(6) 91.39(4) 5831.5 16 B21/d

primitive 15.528(8) 15.636(8) 13.102(6) 113.56(5) 2915.9 8 P21/c

CRYSTALS WITH FIRST ORDER PHASE T RANSFORMATIONS ON COOL ING 1121

cycl[3.2.2]azine N

92.5 A 1.01

a sin beta

A

B

B

92 beta

2a sin beta

beta

1.00 91.5 1.02

1.01

1.05 b

b

4V V

1.00

1.00

1.02 2 1.01

– 1(202)

– 1(404)

c sin beta

2c sin beta

1 – 1(505)

1.00

– 1(2.5 0 2.5)

0 100

200 T(K)

300

100

200 T(K)

300

Fig. 16.28 {Cycl[3.2.2]azine TNB} – variation of some crystal properties with temperature. B refers to the P21/n disordered structure and A to the B21/d ordered structure given in Table 16.4. Cell lengths and volume are expressed as multiples of the values at 91K; the temperature dependence of one fundamental reflection and one ‘‘superlattice’’ reflection is illustrated. (Adapted from Hanson, 1978.) n n n

reversible but there was a small hysteresis between the transition temperature on cooling (140K) and heating (146K); the transition was single crystal to single crystal. The 293K structure has mixed stacks of alternating cycl[3.2.2]azine and TNB molecules along [001]; the former take up three (or more) orientations in disordered fashion, and calculations suggest that there is no steric reason why the cycl[3.2.2]azine molecule should not rotate (presumably in hindered fashion) in its own plane. In the ordered structure the stack axis periodicity has doubled and the two TNB molecules differ in orientation by 12 ; the two cycl[3.2.2]azine molecules also take up (to a good approximation) two orientations. The two TNB molecules change their orientations slightly as the ordered form is heated, and become indistinguishable above 143K; at the same time the constraints on the orientations of the cycl[3.2.2]azine molecules are relaxed and these molecules rotate, at random, into their high temperature orientations. Some of these rotations are through angles as large as 134 . There are resemblances to the {pyrene PMDA} system. n n n


1122

100

CP /calth K–1 mol–1

80

4

60

40 2 20 0 0

100

10 200

20

0 300

T/K

Fig. 16.29. Measured heat capacity for {naphthalene TCNE}; the dashed line shows the lattice heat capacity. The splitting of the peak at 155K is shown at top left, and the (very) lowtemperature specific heat at lower right. (Reproduced from Boerio-Goates and Westrum, 1980a) n n n

16.6.2

Other examples

There are transformations from monoclinic room temperature structures to triclinic low temperature structures in {TMPD chloranil} (Boer and Vos, 1968), {naphthalene TCNE} (Bernstein, 1967) and {TDT5 TNB} (Williams and Wallwork, 1966). These are first order transformations but may also have disorder–order features; low temperature crystal structures are not known. 14N NQR measurements show that there are two independent N atoms in {naphthalene TCNE} at 77K (Onda et al., 1973), which is compatible with the triclinic symmetry found by x-ray diffraction at a somewhat higher temperature. However, NMR measurements (Fyfe, 1974b) do not give any indication of the complicated phase behavior revealed by Cp measurements (Fig. 16.29; Boerio-Goates and Westrum, 1980a). n n n

n n n

n n n

n n n

16.7

Physical nature of the disordered phase

Disorder of one or other of the components is a not-unusual feature of mixed stack -molecular compounds. A key question is whether such disorder is static or dynamic. A condition of static disorder at low temperature may well change to one of dynamic disorder at higher temperature. It is perhaps worth emphasizing that it is much more difficult to describe a disordered than an ordered state. The experimental methods available can be divided into three groups: (a)

Those sensitive to the presence of disorder, but not to reorientational motions. The principal methods are x-ray and neutron diffraction (as employed for crystal structure analysis by use of Bragg reflections). (b) Those sensitive to reorientational motions but not to the presence of disorder. NMR line width and spin–lattice relaxation time measurements come into this category. 5

2,4,6-tris(dimethylamino)-1,3,5-triazine

PHYSICAL NATURE OF THE DISORDERED PHASE

1123

(c) Those sensitive both to disorder and to reorientational motion. Raman scattering and the behavior of triplet excitons are suitable methods. Elastic and inelastic scattering of neutrons are also powerful techniques in the study of phase transformations (Axe, 1971) but have hardly been applied as yet to the study of -molecular compounds. The most direct method of demonstrating the occurrence of disorder is via x-ray (or neutron) crystal structure analysis, particularly if the molecular compound has been studied at only one temperature, the most common current situation. Under these circumstances the disorder is revealed by the atoms of one or other of the components having abnormally large displacement factors. The room temperature structure of {naphthalene TCNE} (P21/a, Z ¼ 2, both components at centers of symmetry) provides a convenient example (Williams and Wallwork, 1966), which has been much discussed. The atomic displacement factors are much larger for naphthalene than for TCNE (Fig. 16.30) and the difference increases for atoms at the periphery of the molecule, suggesting large librational motions, or static disorder. These results can be shown in another way through electron density and difference electron density syntheses; the electron density synthesis is essentially independent of any postulated model. A comparison is made for the disordered naphthalene and ordered TCNE molecules in Fig. 16.31. Two models can be postulated at this stage: n n n

(a) large in-plane librations of naphthalene about the normal to the molecular plane, corresponding to the dynamic disorder defined above, (Fig. 16.32(a)); or (b) static disorder of naphthalene in two orientations (Fig. 16.32(b)).

Individual Debye–Woller factors (Å2)

16 C5

14 Naphthalene (300 K)

12 10 8

C4

6

C3

4 C1 2

C2

N TCNE (300 K)

0

1 2 3 Distance from mol. center (Å)

4

Fig. 16.30. The equivalent isotropic atomic displacement factors for the individual atoms in {naphthalene TCNE} at room temperature, plotted against the distance of the atom from the molecular centre. Separate curves are shown for the two molecules (naphthalene – C3 to C5; TCNE – C1 to N). (Reproduced from Herbstein and Snyman, 1969.) n n n

1124


0

0

1Å

1Å

Fig. 16.31. Electron density and difference syntheses in the planes of (a) naphthalene and (b) TCNE molecules in the crystals of {naphthalene TCNE} at room temperature. The contours of electron ˚ 3 and start at 0 eA ˚ 3, while the contours of difference density are at density are at intervals of 1 eA 3 ˚ intervals of 0.5 eA (thick lines zero and positive contours, thin lines negative contors). Two naphthalenes, mutually rotated by 12 , give a reasonable fit to the electron density and difference density contours. (Reproduced from Herbstein and Snyman, 1969.) n n n

A distinction can be made between the two models by calculating the potential energy (PE) curves for libration of the naphthalene molecule in the {naphthalene TCNE} structure (Shmueli and Goldberg, 1974). The complete PE calculations show minima at two orientations separated by 90 in-plane rotation (the N atoms of TCNE lie nearly at the n n n

PHYSICAL NATURE OF THE DISORDERED PHASE

(a)

1125

(b)

Fig. 16.32. {Naphthalene TCNE} – ORTEP diagrams illustrating (a) dynamic disorder (b) static disorder. (Reproduced from Shmueli and Goldberg, 1974.) n n n

corners of a square). The first orientation has a double well, indicating model (b). The second orientation is 8 kJ/mol higher in energy and is not populated in {naphthalene TCNE}. However, in {[3,3]paracyclophane TCNE} these two positions are occupied in 3 : 1 ratio (Bernstein and Trueblood, 1971) and in {(d8-pyrene) TCNE} at 105K in 93 : 7 ratio (Larsen et al., 1975), corresponding to at least 2.3 kJ/mol energy difference between the potential minima. Supporting evidence for static disorder in {naphthalene TCNE} came from the results of a constrained refinement of the x-ray diffraction data, in which the molecules were treated as rigid bodies of standard dimensions whose translational and librational motions were refined. The constrained refinement gave an R-factor of 9.0%, compared to 12.8% for a conventional refinement, in which positions and anisotropic displacement factors of the individual atoms were refined. The physical reason for favoring static disorder is that N. . . . H contacts are more repulsive for model (a) than (b), a situation similar to that illustrated above for {naphthalene TCNB} (Fig. 16.16). Two points should be noted about the calculations of librational potential energy. Firstly changes in charge transfer energy as a function of libration angle have been neglected; quantum mechanical calculations (Kuroda, Amano et al., 1967) indicate that the charge transfer interaction energy between naphthalene and TCNE increases by only 0.4 kJ/mol when naphthalene is rotated in-plane by 12 from the model (a) structure. Secondly, as Shmueli and Goldberg (1974) point out, the possible relaxation of the surroundings of a molecule during its libration was not taken into account in their PE calculations. Such correlation of molecular motions was included in later calculations by Allen, Boeyens and Levendis (1989). The disorder discussed above is a thermodynamic disorder characteristic of the equilibrium state of the crystal and thus temperature dependent. However, there are also some examples of nonequilibrium disorder, presumably frozen-in during growth of the crystals, and thus dependent on their history and not alterable by changes in temperature. One could expect two orientations, widely separated in azimuth, for one (or both) of the components, possibly unequally populated and perhaps not interconvertible by in-plane rotation. This type of disorder has been postulated for {azulene TNB} (structure determined at 300K (Brown and Wallwork, 1965) and 178K (Hanson, 1965)) and for {indole TNB} at 133K. However, the evidence does not seem to be conclusive – in {azulene TNB} there are indications of a phase change at lower temperatures, while in {indole TNB} the existence of the disorder depends on a distinction made between C and NH groups. A more established example is provide by depends on a distinction made between C and NH groups. A more established example is provide by {benzo[c] n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n

n n n


1126

phenanthrene DDQ} where there is disorder of the Cl and CN groups in the acceptor (Bernstein et al., 1977). Orientational disorder must certainly occur in the quasi-plastic phases of charge transfer molecular compounds (Section 16.8) and substitutional disorder in the solid solutions discussed earlier (Section 13.4) but detailed studies have not yet been reported in either of these areas. n n n

16.8

Transformations to quasi-plastic phase(s) on heating

The occurrence of disorder to order transformation on cooling in many charge transfer molecular compounds has been noted above. Conversely, it has been found that many charge transfer molecular compounds transform on heating to a quasi-plastic phase stable at temperatures close to their melting points (Inabe et al., 1981; see also Section 13.4). The systems investigated and the results obtained are summarized in Fig. 16.33(a). Three donors (pyrene, fluoranthene and phenanthrene) have been studied in combination with (a) A D Py

DNF

DNC

DNP

DNT

TNB

TNC

TNP

TNT

NPA

PMDA NBF

DNBF

BTF

st

st

st

st

st

st

st

st

st

st

st

st

st

st

st

st

st

st

st

mst

st

st

st

mst

mst

st

Fl

st

Ph F

Cl

OH

Py

NO2

DNF

DNC

DNP

TNB

O

TNC

CO O CO

OC OC

N

O2N

PMDA

NBF

O

O2N

N N

O DNBF

CO O CO

O2N

TNT NO2

N

CH3 NO2 NO2

TNP

O Ph

O2N

NO2

NO2 F1

NO2

O2N

NO2

O2N

DNT

OH NO2

O2N

NO2

NO2

Cl

NO2

CH3 NO2

NO2

NO2

NO2 NO2

st

NPA O NO N N

O

O

N N O NO

O

BTF

Fig. 16.33. (a) The -molecular compounds that give isomorphous quasi-plastic phases on heating, and the structural formulae of the compounds; ‘st’ refers to stable and ‘mst’ to metastable molecular compounds. (b) The structure proposed for the quasi-plastic phases. The molecules located at 000 and 2/3,1/3, 1/3 are different from those at 00,1/2; 2/3, 1/3, 5/6 and 1/3, 2/3, 1/6. (Reproduced from Inabe et al., 1981.)

TRANSFORMATIONS TO QUASI-PLASTIC PHASE(S) ON HE AT ING

1127

(b)

Fig. 16.33. (Continued )

6 4 2

a

b

c

d

e

f

g h

i

0 6 Second moment/G2

4 2 0 6 4 2 0 6 4 2 0 –50

50

150

–50

50

150

t/°C

Fig. 16.34. Second moments of the proton resonance for the fluoranthene compounds with (a) DNP; (b) TNB; (c) TNC; (d) TNP; (e) TNT; (f) NPA; (g) PMDA; (h) DNBF and (i) BTF. The vertical arrows indicate the transition temperatures to the quasi-plastic phases. (Reproduced from Inabe et al., 1981.)

13 different acceptors; of the 39 possible combinations, 28 transform to quasi-plastic phases characterized by very simple Debye–Scherrer x-ray diffraction patterns, of the type illustrated in Fig.13.9(b) (P and I@ ). The crystal structure proposed for the isomorphous quasi-plastic phases is shown in Fig. 16.33(b); the donor–acceptor stacks continue to exist


1128

but both donor and acceptor molecules are supposed to be azimuthally disordered so that the differences between the various donors, on the one hand, and the various acceptors, on the other, are eliminated by the rotational disorder. Such rotational disorder is feasible because of the disk-like shapes of the donor and acceptor molecules. The proton NMR spectra (Fig. 16.34) indicate that the disorder is dynamic rather than static; the second moments at lower temperatures (4 G2) are consistent with those crystal structures that are known, while the much reduced values of 1 G2 above the temperatures of transition to the quasi-plastic phases are consistent with rotation of the donor molecules in their own planes (there are too few protons in the acceptors for conclusions to be drawn about their state of motion). Another indication of the quasi-plastic nature of the high-temperature phases comes from the values measured for the entropies of transformation to these phases, which range up to 57 J/K mol and are comparable with the entropies of fusion of many of the compounds (46–67 J/K mol). Similar results have been obtained in a high-temperature infrared study of {pyrene 4nitrophthalic anhydride} (Swamy et al., 1983). n n n

16.9

16.9.1

Transformation of the ground state from neutral ) ionic on cooling and/or application of pressure (NI transitions) Introduction

We have already noted in Chapter 13, following McConnell, Soos and Torrance and their coworkers, that crystalline 1 : 1 mixed stack -molecular compounds are quite sharply divided into those with nominally neutral and those with nominally ionic ground states. However, some of the nominally neutral group are close to the neutral–ionic border (Fig. 13.2) and ten of the compounds in Table 13.1 (those with values of Pc appended) have been shown to transform reversibly to an ionic state on the application of pressure (pressure-induced NIT or PINIT; Torrance, Vasquez et al., 1981), as could well be expected on the basis of McConnell’s original ideas.6 The transformation pressure was determined as the onset of a distinct colour change from yellow or green (neutral) to red or brown (ionic). This phenomenon of a neutral , ionic transformation in the solid state has received considerable attention. Originally {TTF chloranil} was the only example of a transition induced on cooling (temperature-induced NIT or TINIT) but others have since been added, most having 1:1 ratios and being composed of TTF and chloranil derivatives or analogs. Transition temperatures are mostly below 100K, while transition pressures are a few kbar; the ionic states achieved by cooling or pressure may well differ in detail. Spectroscopic techniques are usually employed to show that a transition does occur and permit inferences about its nature. Diffraction studies appear mandatory for a proper understanding of the system but are still quite rare. The transition can also be photo-induced (Tanimura and Koshihara, 2001) but this will not be discussed here. Experiment shows that both first-order and second-order NITs are found. Most n n n

6 Girlando, Painelli et al. (1993) state that investigations carried out on members of the original series have not confirmed the occurrence of transitions. Discontinuous NI transitions have been found only for TTF-CA, TTF – fluoranil and tetramethylbenzidine – TCNQ.

NEUTRAL TO IONIC TRANSFORMATION OF THE GROUND STATE

1129

information is available for {TTF chloranil} (summarized by Le Cointe, LemeeĆailleau, Cailleau and Toudic, 1996; see also Bernstein (2002), pp. 195–197); this is one of the few organic crystals to have been studied over a range of temperatures and pressures and by a variety of experimental techniques. We shall find that there are two primary parameters describing (perhaps determining) the occurrence of a transition and its nature. The first is the charge transfer from neutral to ionic state, leading to a difference in ionicity between the two states. The N state has an ionicity of about 0.3 (quasi-neutral could be a preferable descriptor) and that in the I state is larger. The second parameter is the degree of dimerization – donor and acceptor molecules are equally spaced along the stack axis in the N state but unequally spaced in ˚ or less, dimerization is an the I state. As the difference between the two distances is 0.2 A unfortunate term but too well-established to be dislodged; some schematic diagrams should be viewed with caution. The ionicity is best determined spectroscopically and the degree of dimerization from comparison of N and I crystal structures. n n n

16.9.2 {TTF

n n n

chloranil}

The first study of the molecular compound (needles crystallized from acetonitrile) showed a mixed stack, neutral ground state crystal structure at 300K (a ¼ 7.411, b ¼ 7.621, ˚ , b ¼ 99.20 , P21/n, Z ¼ 2; stack axis [100]; degree of charge transfer c ¼ 14.571 A 9 20%; Mayerle et al., 1979; TTFCAN). In parallel studies, powder patterns (temperature range 10–300K at 1 bar) were indexed in terms of the same unit cell throughout; no additional reflections appeared on cooling to 4K nor was there evidence for a triclinic distortion. There were also cell dimension measurements, from single crystals at 300K, over the pressure range 0–20 kbar. Distinct changes of slope were seen at 84K and 1 bar, which was identified as the 1 bar neutral , ionic transformation temperature, and at 11 kbar and 300K, identified as the 300K neutral , ionic transformation pressure.7 This work was followed by a detailed neutron diffraction study using co-sublimed crystals (Le Cointe, 19948; Le Cointe, Lemee´-Cailleau, Cailleau, Toudic, Toupet et al., 1995; TTFCAN); other techniques (CP–T, 35Cl NQR, IR and Raman spectroscopy) were also used. The cell dimensions at various temperatures and pressures (up to 5 kbar) are shown in Fig. 16.35 (Le Cointe, 1994); at atmospheric pressure there are abrupt changes in [010] and [001] around 81K, and to a lesser extent in b and V, while [100] is continuous. The P–T phase diagram derived from these (and other) measurements (Lemeé-Cailleau, Le Cointe et al., 1997) is shown in Fig. 16.36. The space group below 84K is determined by the behaviour of the 030 reflection which is systematically absent above 84K.9 The intensity of this reflection increases 7 Some of the earlier results must be treated with caution. We give a few examples. Batail et al. (1981) give a curve of cell volume against T with a cusp at 84K, i.e. V ¼ 0 (their Fig. 1). This suggests that the NI transition is second order, contrary to all later views. Unfortunately the later neutron diffraction measurements are not complete enough to allow calculation of V–T curves. Kagoshima et al. (1985) give a (heating) curve (their Fig. 2) of I (31 1) against T with an abrupt fall to zero at 78.5 (5)K; it is not clear how this can be reconciled with the known structures. 8 I am grateful to Dr Marylise Buron-Le Cointe for a copy of her doctoral thesis (University of Rennes I). 9 Le Cointe, Lemee´-Cailleau, Cailleau, Toudic, Toupet et al., 1995; (see III B#2, Ionic phase; p. 3378) remark that ‘‘only the (070) superstructure reflection was clearly extracted from the background’’; however, it is the temperature dependence at 1 bar of I(030) that is shown in Fig. 2(a) of this reference. The temperature and pressure dependence of I(030) is shown in Fig. 3 of Le Cointe (1994), from which our Fig. 16.37 has been adapted.


1130

7.4 1 kbar

7.35 7.3

2.55 kbar 1 bar

a (Å)

7.25

5 kbar

7.2 7.15 7.1 7.05 – 7.62 7.6

1 kbar

7.58

2.55 kbar

b (Å)

7.56

5 kbar

1 bar

7.54 7.52 7.5 7.48 7.46

0

50

100

150 200 T(K)

250 300

14.51

c(Å)

14.5 14.49 14.48 14.47 79

80

81

82 83 T (K)

84

85

8

Fig. 16.35. The temperature and pressure dependence of the cell dimensions of TTF-CA (values for [001] were given only for the temperature range shown; b is 99.1 at 300K and 98.6 at 50K (both at 1 bar). From Le Cointe (1994).


1131

1500 TTF-CA 1000 P(MPa)

Ipara Iferro C

500

Npara

0 100

200

300

T (K)

Fig. 16.36. TTF-CA pressure-temperature phase diagram – circles ND, squares NQR, diamonds vibrational spectroscopy. The subscripts para and ferro are abbreviations for paraelectric and ferroelectric. C is the estimated critical point. Lines are guides to the eye. (Reproduced from LemeéCailleau, Le Cointe et al., 1997.)

Integrated intensity of (030) reflection (arbitrary units)

6 5 4 3 1 bar 2.55 1 kbar kbar

2

5 kbar

1 0 0

50

100

150

200

250

300

T(K)

Fig, 16.37. Neutron diffraction intensity of TTF-CA (030) reflection as a function of temperature and pressure. (Adapted from Le Cointe (1994).)

monotonically with falling temperature; similar behaviour is found over the range 0 to 5 kbar (Fig. 16.37). Thus the transition involves a change of space group from P21/n, Z ¼ 2 (N phase) to Pn, Z ¼ 2 (I phase). In the N phase the D and A molecules are located at centres of symmetry but this requirement is relaxed in the I phase. Confirmation of the loss of centrosymmetry comes from the 35Cl NQR spectra – two independent resonances in the N phase and four in the I phase (Gallier et al., 1993; Gourdji et al., 1991; Le Cointe, Gallier et al., 1995). The shape of the curves of I(030) against T (Fig. 16.37) requires some comment. These start out as though the transition was second order (cf. Fig. 16.25) and then fall abruptly to

1132


Table 16.5. Calorimetric measurements on TTF-CA Reference

H J/mol

S J/mol K

Kawamura et al. (1997) Lemeé-Cailleau et al. (1997)

504(5) 461(32)

6.12(6) 5.49(9)

Note: S measured on 0.2 mg samples was given as 4 J/mol K (Wolfe, 1982).

TTF-CA T≈70 K b I

N

a 100 µm

Fig. 16.38. TTF-CA – coexistence of the N and I phases at 70K observed by optical microscopy with unpolarized light. (Reproduced from Buron-Le Cointe et al. (2003; Fig. 6).)

zero, as one would expect for a first order transition. The overall shape, which appears to be qualitatively the same up to at least 5 kbar, is reminiscent of that found for long range order in Cu3Au (see Fig. 12.4 of Warren (1969)). There are a number of measurements of specific heat for TTF-CA; Kawamura et al. (1997) found a single peak while Lemeé-Cailleau, Le Cointe et al. (1997) found a split peak in the same temperature region; nevertheless, the transition enthalpy and entropy values are in reasonably good agreement (Table 16.5) and the thermodynamic parameters are in the expected range for a first order transition. The Clapeyron equation (dP/dT ¼ H/TcV, where dP/dT is the slope at Tc) provides a test of various interrelated quantities. A precise value of dP/dT ( ¼ 4.17 Mpa/K) is obtained from Fig. 16.36 and Tc is precise at 81K. However, neither H nor V is that precisely defined. We have estimated V from the cell dimensions above ˚ 3 (Lemeé-Cailleau, Le Cointe et al. and below Tc (Fig. 16.35) and obtain V ¼ 7.2 A 3 ˚ ). Using a mean value for H (Table 16.5) we obtain (dP/dT) calc ¼ (1997) give 4–5 A 1.37 Mpa/K, some 40% of the phase diagram value. Lemeé-Cailleau, Le Cointe et al. (1997) give a phase diagram value of 3.1 Mpa/K (3.6 Mpa/K from Fig. 8 of Le Cointe (1994)) and a derived value of 3.2 0.4 Mpa/K. The source of the discrepancies is not known. Buron-Le Cointe, Lemeé-Cailleau et al. (2003) show a micrograph of a TTF-CA needle crystal in which the interface ((010) plane) between the N and I phases can be clearly seen in Fig. 16.38. This is an example of Mnyukh’s (2001; see pp. 121–143 and Figs. 2.34 and 2.43) category of epitaxial growth when the two phases resemble one another. The hysteresis and related phenomena can all be described in terms of Mnyukh’s treatment of first-order enantiotropic phase transitions (see his Fig. 2.30) and have no direct connection with the neutral-to-ionic nature of the phase transformation.


1133

The details of the crystal structures of the N and I phases of TTF-CA (Le Cointe, Lemee´-Cailleau, Cailleau, Toudic, Toupet et al., 1995) require some comment. The N phase structure was determined by neutron diffraction at 300 and 90K; we consider only the 90K results (1223 independent reflections, R ¼ 3.4%, 118 parameters, goodness of fit 1.40, all nonhydrogen atoms refined anisotropically). Both component molecules have D2h–mmm symmetry within the precision of the measurements; bond lengths and angles have standard values. The I phase structure was determined by neutron diffraction at 40K (1636 independent reflections, R ¼ 5.0%, 105 parameters, goodness of fit 1.48, all atoms refined isotropically). There are large differences between chemically equivalent bonds; for example, the (intraring) C ¼ C bond lengths in ˚ , with / ¼ 29. Thus either there is a TTF are given as 1.323(1) and 1.364(1) A remarkable (and unprecedented) polarization effect in the I structure or the standard uncertainties have been underestimated by a factor of about 10;10 the second alternative seems more likely. Le Cointe, Lemee´-Cailleau, Cailleau, Toudic, Toupet et al. (1995) give considerable attention to differences in some intermolecular distances between N and I structures. They also assign NQR frequencies to specific Cl atoms in the I phase ˚ , all with s.u.s on the basis of C–Cl bond lengths (d(C–Cl) ¼ 1.724, 1.705,1.711,1.714 A of 0.001 A). These differences are less impressive if the true standard uncertainties are ˚. 0.01 A The main difference between N and I structures lies in the separation of the TTF and CA molecules along the [100] stack axis; these (center to center, not plane to plane) ˚ (at 90K, ¼ a/2) but unequal at 3.50 and distances are equal in the N structure at 3.61 A ˚ 3.69 A in the I structure at 40K. This alternation is called ‘dimerization’. Batail et al. (1981) pointed out that there was C–H . . . O hydrogen bonding along the [001] direction, at a time when such bonding was controversial. However, it seems disputable that the transformation is due to changes in C–H . . . O hydrogen bonding because hydrogenated and deuterated molecular compounds transform in essentially the same way (Ayache and Torrance, 1983), On the other hand, Oison et al. (2001) contend that charge transfer produces a strengthening of these hydrogen bonds. Lack of precision prevents the use of component dimensions to estimate the degree of charge transfer (ionicity) in the I phase, and this must be done by spectroscopic techniques (see below). The temperature (Girlando et al., 1983) and pressure (Tokura et al., 1986; Girlando et al., 1986) induced neutral to ionic phase transitions in {TTF chloranil} have been studied by IR spectroscopy (Fig. 16.39). The differences between the spectra of the neutral and ionic phases are shown by comparing the upper spectrum with those in the centre and lower positions. The resemblances between temperatureinduced and pressure-induced ionic phases is shown by a comparison of centre and lower parts – the low-temperature and high-pressure ionic phases are, at least, very similar in nature. The Le Cointe 300K ND measurements (Fig. 16.35, up to 5 kbar) indicate a first order transition but one that deviates somewhat from classic expectations. These measurements show a linear dependence of a and b cell dimensions on pressure in both phases; the values n n n

10 Le Cointe et al. (1995) remark ‘‘ . . . the standard deviations have particularly small values at 40K, which can be explained as an effect of the isotropic refinement.’’ It does not seem reasonable that a structure refined isotropically (even at lower temperature) will be more precise than one refined anisotropically, conditions of measurement being essentially the same.


1134

TTFCAN–neutral phase at 1 bar, 300K

Absorbance

PINI at 11 Kbar, 300 K

TINI 1 bar, 15 K

1800

1400

1000 – –1 v/cm

600

Fig. 16.39. IR spectra of {TTF chloranil}. (top), the neutral phase at room temperature and pressure; (center) the ionic phase at 300K and 11 kbar; (bottom) the ionic phase at 15K, 1 bar (Girlando et al., 1986). The spectra shown in the centre and lower panels were obtained with polarized IR – full line, electric vector parallel to stack axis; dashed line, perpendicular to it). The resemblance between the central and lower spectra implies that the same ionic phase is obtained by cooling to 15K at 1 bar and by compression to 11 kbar at 300K. (Reproduced from Girlando et al. (1986).) n n n

˚ /kbar in the N and I phases, and the corresponding of da/dP are 0.041 and0.025 A values for the b dimension are 0.013 and 0.012 respectively. At 1 bar the (020) reflection shows definite hysteresis between heating and cooling curves (Fig. 16.40; this is also found by other techniques such as 35Cl NQR and specific heat measurements), in


1135

7.58 7.57

b(Å)

7.56 7.55

1 bar: Hysteresis

TTF-CA (URennes) EFFECT OF PRESSURE ON HYSTERESIS

7.54 7.53 7.52 7.51 75

77

79

81

83

85

T(K) 7.6

1 kbar

7.58

2.55 kbar

b(Å)

7.56

5 kbar

1 bar

7.54 7.52 7.5 7.48 7.46 0

50 100 150 200 250 300 T(K) 7.54

b(Å)

7.53

5 Kbar: No hysteresis

7.52 7.51 7.5 195

200

205 T(K)

210

215

Fig. 16.40. The effect of pressure on the TINI in TTF-CA. The central panel is taken from Fig. 16.35, while the upper and lower panels are from the Le Cointe (1994) thesis (Figs. 7 and 16). The filled circles are for the heating regimen and the open circles for cooling.

contrast to what is found at 5 kbar (Fig. 16 from Le Cointe (1994)), where there is no hysteresis. This confirms the first order nature of the transition at 1 bar and suggests that the transition at 5 kbar is second order. If so, the tricritical point in the phase diagram (Fig. 16.36) must be shifted.


1136

7.5 7.4

Neutral

a (300K) (Å)

7.3 7.2 Intermediate region

7.1 7

Ionic

6.9

11 kbar 6.8 0

5

10

1700

15 P (kbar)

20

25

CA° 0.0 0.2

Wavenumber (cm–1)

N

0.4

I1

0.6 1600 TTF-CA

0.8

I2 1.0 K 1500

0

10

+

CA–

20 Pressure (kbar)

30

Fig, 16.41. Upper panel: Diffraction study of PINI transition in TTF-CA at 300K; the diagram has been redrawn from Metzger and Torrance (1985) who quote results (‘‘to be submitted’’) of King et al. The King results do not appear to have been submitted. Lower panel: frequencies of the C¼O stretch band at 290K as a function of hydrostatic pressure; the pressure dependences for neutral chloranil and for KþCA were used for calibration purposes. Approximate values deduced for the ionicity are shown by the right-hand ordinate (Tokura et al., 1986). (Adapted from Metzger and Torrance (1985) and Tokura et al., (1986).)

The PINI transition at 300K, 11 kbar is shown by diffraction (the variation of the [100] cell dimension with pressure; upper panel of Fig. 16.40) and by IR spectroscopy (lower panel of Fig. 16.41; Tokura et al., 1986); the transition is essentially first order but neither diffraction nor spectroscopy shows an ideally sharp transition. Some care is needed in the preparation of {TTF chloranil} as it crystallizes from solution as plates or needles, and, by co-sublimation, as parallelepipeds of up to 20 mm3 in n n n


1137

volume (Kawamura et al., 1997). It is the structure of the co-sublimed form (also obtainable from solution) that is discussed above. Two other phases of {TTF chloranil}, named as I and II, have been reported; I and II both appeared as black needles, with phase II shown to have a 1 : 1 composition, and this was presumed to hold also for phase I (Matsuzaki et al., 1983). The analogous black ‘‘snowflakes’’ (Girlando et al., 1983) appear, from comparison of 300K IR spectra, to be essentially II, perhaps with a small contamination of I. Kawamura et al. (1997) noted that their specific heat measurements required correction for the presence of more than one phase. It was inferred (Girlando et al., 1986) that phase I was a mixed valence ionic solid but detail is lacking. Phase II is a fully ionic solid containing (TTFþ)2 and (CA)2 self dimers, present either in distorted stacks or as discrete units (e.g. as in TTF.Br) (Girlando et al., 1986; Matsuzaki et al., 1983); IR and Raman spectra show that there is no phase transition between 300 and 15K (Matsuzaki et al., 1983). Electronic spectra for {TTF bromanil} suggest that phases I and II are also found in this system. No crystallographic information is available for these phases. n n n

n n n

16.9.3 {DMTTF chloranil} n n n

A TINI transition at 65 5K in {2,6-dimethyltetrathiafulvalene chloranil} was first proposed by Aoki et al. (1993) on the basis of studies of polarized visible reflection and IR spectra; a ‘‘most striking feature’’ was the appearance of a coexisting neutral-ionic phase at lower temperatures. Definitive structure determinations at 300, 75 and 45K by x-ray diffraction (Collet et al., 2001) clarified earlier confusions (Nogami et al., 1995) about the low temperature structure. Cell dimensions are given in Table 16.6; there is excellent agreement among the three independent 300K measurements and this applies also to the two crystal structure determinations. Nogami et al. have reported unusual behaviour of the cell dimensions at low temperatures (their Fig. 2, reproduced here as Fig. 16.42); the Collet values at 75 and 40K agree well with those Nogami, thus suggesting that the full set of Nogami cell dimensions can be relied upon to give valuable information about the nature of the transition. Information about the nature of the transition is also obtained from spectroscopy. The optical conductivity spectra for the C ¼ O stretch mode are shown in Fig. 16.43(left), while the wave numbers of the peaks of the spectra at different temperatures are shown in Fig. 16.43(right). The transition is not abrupt but is spread over some 50K, suggesting that it is second order. Similar conclusions can be drawn from the temperature dependence of the IR spectra of DMTTF-CA (Fig. 16.44) and from the cell dimension–T curves (Fig. 16.41), The low temperature phase has a unit cell doubled along [001] with noncentrosymmetric space group P1; thus there are two crystallographically independent molecules of each component in the unit cell. Detailed crystal structures have been reported by Collet et al. at 300, 75 and 40K. We compare the 75K (2395 independent reflections, R ¼ 2.82%, 125 parameters, goodness of fit 1.04, all nonhydrogen atoms refined anisotropically) and 40K (4693 independent reflections,11 R ¼ 3.44%, 485 parameters, n n n

11 2777 with F20 > 2(F20); 2362 of the 4693 were superstructure reflections, 765 with F20 > 2(F20). In the 40K determination the coordinates of one atom should have been fixed and the Friedel parameter should have been included in the refinement. The low internal R value for Friedel opposites (2.16%) suggests the possibility of polysynthetic twinning.


1138

˚ , , A ˚ 3) for DMTTF-CA from various sources and at Table 16.6. Cell dimensions (reduced cell; A different temperatures. The space group is P1, Z ¼ 1 at all temperatures except for 40K, where it is P1, Z ¼ 2 T(K)

a

b

c

V

Reference

300 300

7.272 7.285

7.666 7.678

8.512 8.521

95.91 95.90

103.89 103.89

91.89 91.91

457.4 459.4

300 75 75 40# 40

7.269 7.121 7.118 7.099 7.090

7.673 7.586 7.581 7.563 7.556

8.514 8.476 8.464 16.937 8.450 X2

95.87 95.87 96.00 95.77 95.83

103.91 104.07 104.02 104.21 104.17

91.87 90.92 90.83 91.02 90.98

457.2 441.4

Aoki et al. (1993) Horiuchi et al. (2001); UDEQUP01 Collet et al. (2001) Collet et al. (2001). Nogami et al. (1995)* Collet et al. (2001). Nogami et al. (1995)*

876.2

* Values read off Fig.2 of Nogami et al.; volumes deliberately omitted. ˚ , 83.78 79.84 88.99 . # Reduced cell 7.100 7.564 16.680 A

b

0 Lattice parameter increment [%]

c 0 a –1 b

d

–2 a

0

100

–1

200

Temperature [K]

300

0

100

200

300

Temperature [K]

Fig. 16.42. Cell dimension decrements for DMTTF–CA as a function of temperature. The space group is P1 above 65K and P1 below. (Reproduced from Nogami et al., 1995.)

goodness of fit 0.93, all nonhydrogen atoms refined anisotropically) structures. At 75K ˚ there is a . . . DMTTF..CA . . . stack along [100], the components being separated by 3.56 A ( ¼ a/2). At 40K there are two crystallographically independent . . . DMTTF–CA . . . ˚ in one stack, and 3.46 stacks, with the pairs of components separated by 3.49 and 3.61 A ˚ in the second. Collet et al. describe these as neutral and ionic stacks and 3.64 A respectively but critical scrutiny suggests that the component dimensions are not precise enough for such an identification. Although it was reported (Nogami et al., 1995) that structures had been determined at 293, 109, 48 and 29K, details do not appear to have been published.


(a)

1139

(b) DMTTFQCI4

Peak of optical conductivity spectra, C=O stretch mode, E perpendicular to DA stack (Horiuchi et al., 2001)

C=O stretch mode 10K

47 55 60 63.5 65 70 200 S cm–1 90 1600 1650 Wave number (reciprocal cm)

Wave number (reciprocal cm)

35

1635 1630 1625 1620 1615 1605 1605 1600 0 10 20 30 40 50 60 70 80 90 100 T (K)

Fig. 16.43. (left) The optical conductivity spectra of DMTTF-CA (E perpendicular to DA stack; C¼O stretch mode) over a range of temperatures. The dashed line is a guide to the eye. (right) Wave numbers of spectral peaks (from the dashed line in the left diagram) plotted against temperature. These values can be compared with the values given for CA0 (1685 cm1) and CA1 (in K CA) (1520 cm1) in Fig. 16.40 (lower panel). (Reproduced from Horiuchi et al., 2001.)

Aoki et al. (1995) report that the N phase transforms into a fully ionized I phase above 12 kbar at 300K. Thus there may be a difference in pressure and temperature induced transitions in this system. The NI transition in TTF-CA is essentially first order in nature while that in DMTTF– CA appears to be essentially second order. 16.9.4 Other examples (a) TTF – 2,5-dichloro-p-benzoquinone. This molecular compound (TTF–2,5-Cl2BQ) ˚ , 106.94 97.58 93.66 , crystallizes as dark green triclinic needles (7.935 7.216 6.844 A Z ¼ 1, space group P 1) (Girlando, Painelli et al., 1993). There are resemblances to the TTF-CA crystal structure. Raman and polarized IR spectra allowed the evaluation of many microscopic parameters involved in the contrasting behaviour of TTF-CA and TTF–2,5-Cl2BQ. However, we shall not dwell on these as study over a wide range of temperatures and pressures is lacking. A putative lower-symmetry structure was studied theoretically by Katan and Koenig (1999). (b) 2-chloro-5-methyl-p-phenylenediamine–2,5-dimethyldicyanoquinone-diimine. This molecular compound (abbreviated as ClMePD–DMeDCNQI) shows unique features not yet encountered in other molecular compounds that show neutral-to-ionic transitions. IR


1140

Ionicity

0.5 DMTTF-CA from C=0 stretch mode of CA

0.4

Mode Intensity L

67 K

1 DMTTF a(g) mode

0.5

0

50

150 100 Temperature (K)

200

Fig. 16.44. Upper curve – ionicity estimated from frequency of C¼O stretch mode versus temperature. Note that the ionicity extrapolates to 0.5 at 0K. Lower curve–mode intensity from DMTTF mode. (Adapted from Horiuchi et al. (2001).)

and visible spectra of powdered crystals over the range 50–350K led to the conclusion that a continuous change in molecular ionicity over a range of 200K is accompanied by dynamic distortions of the stacks at low temperatures (Aoki and Nakayama, 1997). The ˚ , 91.23 112.19 crystals (from dichloromethane) are triclinic needles, 7.463 7.504 7.191 A 96.91 , Z ¼ 1 (presumably at 300K). The space group was given as P1, although ‘‘to obtain suitable values of thermal factors in the analysis, 2- and 5-substitutional sites [in ClMePD] were assumed to be occupied by Cl atoms and methyl groups with equal probability.’’ The ClMePD and DMeDCNQI molecules formed mixed-stack columns in a direction parallel to the b axes. This description suggests that the space group could be P 1; however, no structural details have been published so this remains a speculation. These spectroscopic results were confirmed by a Raman and polarized IR study over the range 80–400K (Masino et al., 2001). Spectroscopic studies at various pressures have been carried out by Masino et al., (2003). A low temperature diffraction study seems essential. (c) 3,3’,5,5’-Tetramethylbenzidine–TCNQ. A TINI transition was found in this mixedstack molecular compound at around 205K. At 300K the golden-yellow cosublimed ˚ , 100.19 , Z ¼ 2, space group P21/n crystals are monoclinic, 6.722 21.873 8.108 A (Iwasa et al., 1990; details of the structure not reported). Polarized visible and IR spectra show a first-order transition on cooling at 200K and on heating at 228K; thus there is appreciable hysteresis. The ionicity is 0.59 in the high temperature N phase and


1141

1550

0.5

1500

Degree of Ionicity

Wav number/cm–1

0

1 1450

0

1 2 Pressure/GPa

3

Fig. 16.45. Pressure dependence of the C¼N stretching IR peak frequency at 300K (left hand scale) of JIXWES and ionicity (right hand scale). Note the hint of phase coexistence around 1 Gpa. (Reproduced from Aumu¨ller et al. (1991).)

0.69 in the low temperature I phase; we retain the N/I nomenclature despite its limitations here. Details of the spectra suggest that there are two types of TCNQ molecule in the I phase. A micrograph at 180K shows that the I phase has a striated appearance. ‘‘Narrow diagonal I domains appear in (a, c) planes whose number increases as the temperature decreases. Such coexistence is observed over several [tens of] degrees Kelvin. However, when large parts of the crystal are transformed,.. the transition is always accompanied by a sharp breaking of the crystal along the a and c axes.’’ (Buron-Le Cointe et al., 2003). It is difficult to obtain diffraction-quality crystals for further studies. There are two PINI transitions at 300K, one at 6 and the other at 20 kbar. (Iwasa et al.,1993). The 6 kbar transition is to an equal mixture of ‘‘ionic’’ and ‘‘neutral’’ molecules, while there are only ionic entities above 20 kbar. (d) TTF–N,N 0 -dicyano-2,5-dimethyl-1,4-benzoquinone-di-imine. NC N CH3 N,N⬘-dicyano-2,5-dimethyl-1,4-benzoquinone-di-imine H3 C N CN

The 300K crystal structure of this mixed-stack neutral molecular compound was reported ˚ , 101.41 , Z ¼ 2, P21/c; JIXWES). The by Aumu¨ller et al. (1991) ( 6.172 7.831 17.998 A high-pressure FT-IR and electronic spectra of powdered samples (300K, up to 3.5 GPa;


1142

Fig. 16.45) were measured, and also the electronic conductivity along the needle axis of a single crystal. There is a sharp PINI transition at 1.0 GPa from an essentially neutral state to an almost completely ionic state. The electrical conductivity increases by five orders of magnitude above about 1.5 GPa. One could anticipate a first order transition to an ionic state on cooling but this has not been investigated. JIXWES appears to resemble TTF-CA more closely than any of the other examples discussed above. 16.9.5

Concluding summary

What started off as a rather limited enterprise, exemplified only by TTF-CA, has blossomed into a multifacetted study with a growing number of subjects, each differing from TTF-CA in one or more aspects. The experimental work summarized above has been accompanied by many theoretical studies, mostly employing the twin concepts of ‘ionicity’ and ‘dimerization’ and neglected here for reasons of space.

References Abdel–Rehiem, A. G., Farrell, P. G. and Westwood, J. V. (1975). J. Chem. Soc. Faraday Trans., 1, 71, 1762–1771. Aizu, K. (1970). J. Phys. Soc. Jpn., 28, 706–716. Allen, C. C., Boeyens, J. C. A. and Levendis, D. C. (1989). S. Afr. J. Chem., 42, 38–42. Aoki, S. and Nakayama, T, (1997). Phys. Rev., 56, R2893–R2896. Aoki, S., Nakayama, T, and Miura, A. (1993). Phys. Rev., B48, 626–629. Aumu¨ller, A., Erk, P., Hu¨nig, S., Ha¨dicke, E., Peters, K. and Schnering, H. G. von (1991). Chem. Ber., 124, 2001–2004. Axe, J. D. (1971). Trans. Amer. Cryst. Assoc., 7, 89–106. Ayache, C. and Torrance, J. B. (1983). Solid State Commun., 47, 789–793. Basaki, S. and Matsuzaki, S. (1995). Synth. Met., 70, 1239–1240. Batail, P., LaPlaca, S. J., Mayerle, J. J. and Torrance, J. B. (1981). J. Am. Chem. Soc., 103, 951–953. Beasley, J. R., Doyle, D. F., Chen, L., Cohen, D. S., Fine, B. R. and Pielak, G. (2002). Proteins: Structure, Function and Genetics, 49, 398–402. Berry, R. S., Rice, S. A. and Ross, J. (1980). Physical Chemistry. John Wiley and Sons, New York etc. Bernstein, H. J., Dalal, N. S., Murphy, W. F., Reddoch, A. H., Sunder, S. and Williams, D. F. (1978). Chem. Phys. Lett., 57, 159–162. Bernstein, J. (2002). Polymorphism in molecular crystals. Clarendon Press, Oxford. pp. xiv þ410. Bernstein, J. and Trueblood, K. N. (1971). Acta Cryst., B27, 2078–2089. Bernstein, J., Regev, H. and Herbstein, F. H. (1977). Acta Cryst., B33, 1716–1724. Bernstein, T. (1967). M.Sc. Thesis, Department of Chemistry, Technion – Israel Institute of Technology, Haifa, Israel. Boer, J. L. de and Vos, A. (1968). Acta Cryst., B 24, 720–725. Boerio–Goates, J. and Westrum, E. F., Jr. (1979). Mol. Cryst. Liq. Cryst., 50, 259–268. Boerio–Goates, J. and Westrum, E. F., Jr. (1980a). Mol. Cryst. Liq. Cryst., 60, 237–248. Boerio–Goates, J. and Westrum, E. F., Jr. (1980b). Mol. Cryst. Liq. Cryst., 60, 249–266. Boerio–Goates, J., Westrum, E. F., Jr. and Fyfe, C.A. (1978). Mol. Cryst. Liq. Cryst., 48, 209–218.

REFERENCES

1143

Bragg, W. L. and Williams, E. J. (1935). Proc. Roy. Soc., London., A151, 540–546. Brose, K.–H., Luty, T. and Eckhardt, C. J. (1990). J. Chem. Phys., 93, 2016–2031. Brown, D. S. and Wallwork, S. C. (1965). Acta Cryst., 19, 149. Buron-Le Cointe, M., Lemeé-Cailleau, M. H., Cailleau, H., Toudic, B., Moreác, A., Moussa, F., Ayache, C and Karl, N, (2003). Phys. Rev., B68, 064103-1–064103-7. Clayton, P. R., Worswick, R. D. and Staveley, L. A. K. (1976). Mol. Cryst. Liq. Cryst., 36, 153–163. Collet, E., Buron-Le Cointe, M., Lemeé-Cailleau, M. H., Cailleau, H., Toupet, L., Meven, M., Mattauch, S., Heger, G, and Karl, N. (2001). Phys. Rev., B63, 054105-1–054105-12. Cowley, R. A. (1980). Adv. Phys., 29, 1–110. Czarniecka, K., Janik, J. M., Janik, J. A., Krawcyzk, J., Natkaniec, I., Wasicki, J., Kowal, R., Pigon, K. and Otnes, K. (1985). J. Chem. Phys., 85, 7289–7293. Czekalla, J. and Meyer, K. O. (1961). Z. Physik. Chem. (NF), 27, 185–198. Czekalla, J., Briegleb, G., Herre, W. and Glier, R. (1957). Z. Elektrochem., 61, 537–546. Dalal, N. S., Ripmeester, J. A., Reddoch, A. H. and Williams, D. F. (1978). Mol. Cryst. Liq. Cryst. Letts., 49, 55–59. Dalal, N., Haley, L. V., Northcott, D. J., Park, J. M., Reddoch, A. H., Ripmeester, J. A., Williams, D. F. and Charlton, J. L. (1980). J. Chem. Phys., 73, 2515–2517. Darmon, I. and Brot, C. (1967). Mol. Cryst., 2, 301–321. Dunitz, J. D. (1995). Chem. Biol., 2, 709–712. Dunn, A. G., Rahman, A. and Staveley, L. A. K. (1978). J. Chem. Thermodynam., 10, 787–796. Ecolivet, C. and Mierzejewski, A. (1990). Phys. Rev., B42, 8471–8481. Ecolivet, C., Bertault, M., Mierzejewski, A. and Collet, A. (1987). In Dynamics of Molecular Crystals, edited by J. Lacombe, pp. 187–192. Amsterdam: Elsevier. Ecolivet, C., Lemee´, M. H., Delugeard, Y., Girard, A., Bertault, M., Collet, A. and Mierzejewski, A. (1988). Mater. Sci., 14, 55–58. Ehrenfest, P. (1933). Proc. Acad. Sci. Amsterdam, 36, 153–157. Erdle, E. and Mo¨hwald, H. (1979). Chem. Phys., 36, 283–290. Fyfe, C. A. (1974a). J. Chem. Soc. Farad. Trans. 2, 70, 1633–1641. Fyfe, C. A. (1974b). J. Chem. Soc. Farad. Trans. 2, 70, 1642–1649. Fyfe, C. A., Harold–Smith, D. and Ripmeester, J. (1976). J. Chem. Soc. Farad. Trans. 2, 72, 2269–2282. Gallier, J., Toudic, B., De´lugeard, Y., Cailleau, H., Gourdji, M., Peńeau, A. and Guibe´, L. (1993). Phys. Rev., B47, 11688–11695. Gilson, D. F. and McDowell, C. A. (1966). Canad. J. Chem., 44, 945–952. Girlando, A., Marzola, F., Pecile, C. and Torrance, J. B. (1983). J. Chem. Phys., 79, 1075–1085. Girlando, A., Painelli, A., Pecile, C., Calestani, G., Rizzoli, C., and Metzger, R. M. (1993). J. Chem. Phys., 98, 7692–7698. Girlando, A., Pecile, C., Briliante, A. and Syassen, K. (1986). Solid State Commun., 57, 891–896. Gourdji, M., Guibe´, L. Peńeau, A., Gallier, J., Toudic, B. and Cailleau, H., (1991). Z. Naturforsch., 47a, 257–260. Grupp, A., Wolf, H. C. and Schmid, D. (1982). Chem. Phys. Lett., 85, 330–334. Haarer, D. and Karl, N. (1973). Chem. Phys. Lett., 21, 49–53. Haarer, D. (1974). Chem. Phys. Lett., 27, 91–95. Haarer, D. (1975). Chem. Phys. Lett., 31, 192–194. Haarer, D. (1977). J. Chem. Phys., 67, 4076–4085. Haarer, D., Philpott, M. R. and Morawitz, H. (1975). J. Chem. Phys., 63, 5238–5245. Hammick, D. Ll. and Hutchinson, H. P. (1955). J. Chem. Soc., pp. 89–91. Hanson, A. W. (1964). Acta Cryst., 17, 559–568. Hanson, A. W. (1965). Acta Cryst., 19, 19–26. Hanson, A. W. (1978). Acta Cryst., B34, 2195–2200.

1144


Herbstein, F. H. (1971). ‘‘Crystalline -molecular compounds’’ in Perspectives in Structural Chemistry, Wiley: London and New York, Vol. IV, pp. 166–395. Herbstein, F. H. (1996). Cryst. Revs., 5, 181–226. Herbstein, F. H. and Samson, S. (1994). Acta Cryst., B50, 182–191. Herbstein, F. H. and Snyman, J. A. (1969). Phil. Trans. Roy. Soc. Lond. A264, 635–666. Herbstein, F. H., Kapon, M., Reisner, G. M., Lehmann, M. S., Kress, R. B., Wilson, R. B., Shau, W.–I., Duesler, E. N., Paul, I. C. and Curtin, D. Y. (1985). Proc. Roy. Soc. Lond., A399, 295–319. Herbstein, F. H., Marsh, R. E. and Samson, S. (1994). Acta Cryst., B50, 174–181. Horiuchi, S., Okimoto, Y., Kumai, R. and Tokura, Y. (2001). J. Am. Chem. Soc., 123, 665–670. Inabe, T., Matsunaga, Y. and Nanba, M. (1981). Bull. Chem. Soc. Jpn., 54, 2557–2564. Iwasa, Y., Koda, T., Tokura, Y., Kobayashi, A., Iwasawa, N. and Saito, G. (1990). Phys. Rev., B42, 2374–2377. Iwasa, Y., Watasnabe, N., Koda, T. and Saito, G. (1993). Phys. Rev., B47, 2920–2377. Iwata, S., Tanaka, J. and Nagakura, S. (1967). J. Chem. Phys., 47, 2203–2923. Kagoshima, S., Kanai, Y., Tani, M., Tokura, Y. and Koda, T. (1985). Mol. Cryst. Liq. Cryst., 120, 9–15. Kanai, H., Tani, M., Kagoshima, S., Tokura, Y. and Koda, T. (1984). Synth. Metals, 10, 157–160. Katan, C. and Koenig, C. (1999). J. Phys. Cond. Matter, 11, 4163–4177. Kawamura, T., Miyazaki, Y. and Sorai, M. (1997). Chem. Phys. Lett., 273, 435–438. Koci 0 nski, J. (1983). Theory of Symmetry Changes at Continuous Phase Transitions, Elsevier, Amsterdam etc. and PWN–Polish Scientific Publications, Warsaw. Krzystek, J. and Schutz, J. von (1993). Adv. Chem. Phys., 86, 167–329. Kuchta, B., Luty, T. and Etters, R. D. (1990). J. Chem. Phys., 93, 5935–5939. Kumakura, S., Iwasaki, F. and Saito, Y. (1967). Bull. Chem. Soc. Jpn., 40, 1826–1833. Kuroda, H., Amano, T., Ikemoto, I. and Akamatu, H. (1967). J. Amer. Chem. Soc., 89, 6056–6063. Landau, L. D. and Lifshitz, E. M. (1980). Statistical Physics, being Volume 5 of Course of Theoretical Physics, Third Edition, revised and enlarged by E. M. Lifshitz and L. P. Pitaevskii, translated by J. B. Sykes and M. J. Kearsley, Pergamon Press, Oxford etc., Chapter XIV. Larsen, F. K., Little, R G. and Coppens, P. (1975). Acta Cryst., B31, 430–440. Le Bars–Combe, M., Chion, B. and Lajzerowicz–Bonneteau, J. (1979). Acta Cryst., B35, 913–920. Le Cointe, M. (1994). ‘‘La Transition neuter-ionique dans le TTF – p-chloranile: Aspects structuraux.’’ University of Rennes I, France. Le Cointe, M., Lemeé-Cailleau, M. H., Cailleau, H., and Toudic, B. (1996). J. Mol. Struct., 374, 147–153. Le Cointe, M., Gallier, J., Cailleau, H., Gourdji, M., Peńeau, A., and Guibe´, L. (1995). Sol. State Commun., 94, 455–459. Le Cointe, M., Lemeé-Cailleau, M. H., Cailleau, H., Toudic, B., Toupet, L., Heger, G., Moussa, F., Schweiss, P., Kraft, K. H. and Karl, N. (1995). Phys. Rev., B51, 3374–3386. Lefebvre, J., Odou, G., Muller, M., Mierzejewski, A. and Luty, T. (1989). Acta Cryst., B45, 323– 336. Lemeé-Cailleau, M. H., Le Cointe, M., Cailleau, H., Moussa, F., Roos, J., Brinkmann, D., Toudic, B., Ayache, C and Karl, N, (1997). Phys. Rev. Lett., 79, 1690–1693. Liu, L. and Guo, Q-X. (2001a). Chem. Rev., 101, 673–695. Liu, L. and Guo, Q-X. (2001b). Chin. J. Chem., 19, 670–674. Lyubarski, G. Ya. (1960). The Application of Group Theory in Physics, Pergamon Press, Oxford. MacFarlane, R. M. and Ushioda, S. (1977). J. Chem. Phys., 67, 3214–3220. Masino, M., Girlando, A., Farina, L. and Brillante, A. (2001). Phys. Chem. Chem. Phys., 3, 1904–1910. Masino, M., Farina, L., Brillante, A. and Girlando, A. (2003). Synth. Mets., 133–134, 629–631. Matsuzaki, S., Moriyama, T., Onomichi, M. and Toyoda, T. (1983). Bull. Chem. Soc. Jpn., 56, 369–374.

REFERENCES

1145

Mayerle, J. J., Torrance, J. B. and Crowley, J. I. (1979). Acta Cryst., B35, 2988–2995. Metzger, R. M. and Arafat, E. S. (1983). J. Chem. Phys., 78, 2696–2705. Metzger, R. M. and Torrance, J. B. (1985). J. Am. Chem. Soc., 107, 117–121. Mohwald, H., Erdle, E. and Thaer, A. (1978). Chem. Phys., 27, 79–87. Mnyukh, Y. (2001). Fundamentals of solid-state phase transitions, ferromagnetism and ferroelectricity. First Books. Nogami, Y., Taoda, M., Oshima, K., Aoki, S., Nakayama, T. and Miura, A. (1995). Synth. Mets., 70, 1219–1220. Oison, V., Katan, C. and Koenig, C. (2001). J. Phys. Chem., A105, 4300–4307. Onda, S., Ikeda, R. Nakemura, D. and Kubo, M. (1973). Bull. Chem. Soc. Jpn., 46, 2878–2879. Park, J. M. and Reddoch, A. H. (1981). J. Chem. Phys., 74, 1519–1525. Peierls, R. E. (1954). Quantum Theory of Solids. Oxford: Clarendon Press. Pippard, A. B. (1964). The Elements of Classical Thermodynamics. Cambridge University Press, London and New York. Ponte Goncalves, A. M. (1980). Prog. Solid State Chem., 13, 1–88. Ponte Goncalves, A. M. (1977). Chem. Phys., 19, 397–405. Prout, C. K., Morley, T., Tickle, I. J. and Wright, J. D. (1973). J. Chem. Soc. Perkin II, pp. 523–528. Prout, C. K., Tickle, I. J. and Wright, J. D. (1973). J. Chem. Soc. Perkin II, pp. 528–530. Ripmeester, J. A. (1982). J. Chem. Phys., 77, 1069–1070. Ripmeester, J. A., Ratcliffe, C. I., Enright, G. and Brouwer, E. (1995). Acta Cryst., B51, 513–522. Ripmeester, J. A., Reddoch, A. H. and Dalal, N. S. (1981). J. Chem. Phys., 74, 1526–1533. Robertson, B. E. and Stezowski, J. J. (1978). Acta Cryst., B34, 3005–3011. Salje, E. K. H. (1990). Phase transitions in ferroelastic and co-elastic crystals. Cambridge University Press: London, New York. Shahidi, F. and Farrell, P. G. (1978). J. Chem. Soc. Chem. Commun., pp. 455–45X. Shahidi, F. and Farrell, P. G. (1980). J. Chem. Research (S), pp. 214–215. Shahidi, F., Farrell, P. G. and Westwood, J. V. (1980). J. Chem. Research (S), p. 357. Shmueli, U. and Goldberg, I. (1973). Acta Cryst., B29, 2466–2471. Shmueli, U. and Goldberg, I. (1974). Acta Cryst., B30, 573–578. Stanley, H. E. (1971). Introduction to Phase Transitions and Critical Phenomena, Oxford University Press, New York, Oxford. Stezowski, J. J. (1980). J. Chem. Phys., 73, 538–547. Stokes, H. T. and Hatch, D. M. (1988). Isotropy subgroups of the 230 crystallographic space groups, World Scientific, Singapore, New Jersey, London, Hong Kong. Suzuki, K. and Seki, S. (1955). Bull. Chem. Soc. Jpn., 28, 417–421. Swamy, H. R., Ganguly, S. and Rao, C. N. R. (1963). Spectrochim. Acta, 39A, 23–28. Tanimura, K. and Koshihara, S. (2001). Phase Transitions, 74, 21–34. Tokura, Y., Okamoto, H., Koda, T., Mitani, T. and Saito, G. (1986). Solid State Commun., 57, 607–610. Toledano, J.-C. and Toledano, P. (1987). The Landau Theory of Phase Transitions. World Scientific, Singapore. Torrance, J. B., Vazquez, J. E., Mayerle, J. J. and Lee, V. Y. (1981). Phys. Rev. Lett., 46, 253–257. Tsuchiya, H., Marumo, F. and Saito, Y. (1972). Acta Cryst., B24, 1935–1941. Warren, B. E. (1990). X-ray Diffraction. Dover–Mineola, N. Y. Williams, R. M. and Wallwork, S. C. (1966). Acta Cryst., 21, 406– 412. Williams, R. M. and Wallwork, S. C. (1968). Acta Cryst., B24, 168–174. Wilson, K. G. (1971). Phys. Rev., B4, 3184–3205. Wolfe, C. R. (1982). Mol. Cryst. Liq. Cryst., 85, 337–343. Ziman, J. M. (1964). Principles of the Theory of Solids, Chapter 10, Magnetism, University Press, Cambridge.


Chapter 17 Segregated stack -molecular complexes

7,7,8,8-Tetracyanoquinodimethane (TCNQ) is a strong -acid which forms stable crystalline radical salts of type MþTCNQ and a new class of complex salts represented by MþTCNQ TCNQ which contain formally neutral TCNQ. The complex anion-radical salts have the highest electrical conductivities known for organic compounds, exhibiting volume electrical resistivities as low as 0.01 ohm cm at room temperature. Both the conductivity and electron paramagnetic absorption are anisotropic as determined by measurements along major crystal axes. Melby, Harder, Hertler, Mahler, Benson and Mochel, 1962 The term ‘organic metal’ is a misnomer because these solids are neither ductile nor malleable; they are frail organic crystals or relatively brittle polymers. The only properties reminiscent of metals are high reflectivity and relatively low room temperature resistivity, which decreases with decreasing temperature over a certain temperature range. F. Wudl, 1984

Summary: The unexpected phenomenon of high electrical conductivity in some organic crystals appears to be associated with stacked arrangements of radical ions of electron donors and/or acceptors. A variety of chemical types has been developed, principally based on the parent donor tetrathiafulvalene (TTF) and the parent acceptor tetracycanoquinodimethane (TCNQ). TTF cation radical salts and TCNQ anion radical salts have stacked structures with marked anisotropy of arrangement and physical properties and hence are termed ‘quasi-one dimensional’. However, lateral interactions between stacks lead to a measure of two-dimensional character in some of these salts, and this aspect is taking on increasing importance with the development of some of the newer types of donor and acceptor. The degree of stacking ranges from -dimeric pairs to stacks of infinite length; the longer stacks often contain dimeric or tetradic subgroupings, particularly in the TCNQ radical anion salts. In the cation radical-anion radical salts, of which the most famous example is {[TTF][TCNQ]}, there are segregated stacks of cation and anion radicals. {[TTF][TCNQ]} itself has monad stacks in its averaged structure, but there are also examples of diad stacks. The phase transitions in {[TTF][TCNQ]} below 54K are of an unusual type, the drastic drop in stack axis conductivity on cooling below 54K not being accompanied by appreciable changes in average moiety arrangement. Studies of the very weak diffuse scattering above 54K and the weak satellite reflections below 54K lead to a model in which the high resistivity below 54K is accounted for in terms of pinned charge density waves (CDW), which become mobile above the phase transitions. In the temperature region from 54K up to 300K the conductivity can be explained semiquantitatively by a combination of CDW and single phonon scattering.

1148

SEGREGAT ED STACK -MOLECULAR COMPLEXES

17.1 17.2

Introduction Chemistry of donors and acceptors that participate in segregated stacks 17.2.1 Introduction 17.2.2 Donors 17.2.3 Acceptors 17.2.4 Preparation of crystals 17.3 Structures of cation-radical salts 17.3.1 Introduction 17.3.2 Cations are polycyclic aromatic hydrocarbons 17.3.3 TTF and related compounds as cations 17.3.4 TMPD salts containing -dimerized cation radicals 17.4 Structure of TCNQ anion-radical salts 17.4.1 Mutual arrangements of approximately plane-parallel TCNQ moieties 17.4.2 Structures with stacks of limited length 17.4.3 TCNQ anion-radical salts in which the cations are metals 17.4.4 Stacked structures with e average charge on the TCNQ moieties 17.4.5 Stacked structures with 0.8e average charge on the TCNQ moieties 17.4.6 Stacked structures with 2/3e average charge on the TCNQ moieties 17.4.7 Stacked structures with 0.5e average charge on the TCNQ moieties 17.4.8 Stacked structures with 0.4e average charge on the TCNQ moieties 17.4.9 Systems studied over a wide range of temperatures 17.4.10 Conclusions drawn from a survey of the structural results for TCNQ anion-radical salts 17.5 Other anion-radical salts 17.5.1 Alkali-metal chloranil salts 17.5.2 M(dmit)2 and M(mnt)2 as anion radicals in various guises 17.6 Structures of cation-radical anion-radical salts 17.6.1 General survey 17.6.2 Cation : anion ratio 1 : 1; monad stacks 17.6.3 Cation : anion ratio 1 : 1; diad stacks 17.6.4 Cation : anion ratio 2 : 1 or 1 : 2; monad stacks 17.6.5 Cation : anion ratio 2 : 1 or 1 : 2; diad stacks 17.7 Electron density studies of some segregated stack complexes 17.8 Theoretical studies of some segregated stack complexes 17.9 Studies of {[TTF][TCNQ]} and some related materials 17.10 Concluding summary References

17.1

1148 1151 1151 1152 1157 1161 1162 1162 1163 1167 1175 1177 1177 1180 1187 1189 1192 1193 1196 1202 1205 1211 1214 1214 1215 1220 1220 1220 1227 1229 1230 1232 1234 1235 1252 1253

Introduction

Many, many studies have been stimulated by the discovery (Ferraris et al., 1973; Coleman et al., 1973) of a large, temperature-dependent electrical conductivity in {[TTF][TCNQ]} (nomenclature is discussed below) and the association of this striking physical property with the (then-novel) segregated (and not mixed-stack) arrangement of donor and acceptor moieties in the crystal. TTF (tetrathiafulvalene) was indexed in Chemical Abstracts under 0 ‘‘2,2 -Bi-1,3-dithiole’’ until 1972 and thereafter as ‘‘1,3-Dithiole, 2-(1,3-dithiol-2ylidene)’’; registry number 31366–25–3) and TCNQ (7,7,8,8-tetracyanoquinodimethane;


1149

0

first under ‘‘2,5-Cyclohexadiene-1,a:4,a -dimalonitrile’’ and then as ‘‘Propanedinitrile, 2,2 0 -(2,5-cyclohexadiene-1,4-diylidene)bis’’; registry number 1518-16-7); the twocomponent complex has registry number 51159-15-0. Considerable effort has been invested in chemical modification of TTF and TCNQ and in the study of analogous systems in the hope of finding materials with new or better physical properties. The many new materials synthesized can be classified as cation-radical salts (i.e. with closed shell anions), as anion-radical salts (i.e. with closed shell cations) and as cation-radical anionradical salts. Crystal structures have been determined for many of these new materials and their physical properties (especially electronic properties such as conductivity, thermoelectric power and magnetic susceptibility) have been measured over a wide range of temperatures and (sometimes) pressures. The donors and acceptors that form cation and anion radicals and give segregated stack crystal structures can also participate as neutral or ionic moieties in mixed-stack crystals, and also sometimes present crystal structures of new, albeit related, types. The behaviour of individual moieties in a binary system depends on the properties and interactions of both components. As noted in Chapter 1, the structure determining influences in segregated-stack binary molecular complexes are A . . . A and B . . . B and not A . . . B – hence ‘complexes’ and not ‘compounds’. A different symbolism is also needed. Thus we emphasize the segregation by replacing the inappropriate {TTF TCNQ} by {[TTF][TCNQ]}, in more general terms {[donor][acceptor]}; in some instances only one of the components is stacked and only it is placed in the square brackets. Problems arise with known structures that do not fit into the simple mixed stack/segregated stack framework and with unknown structures: then we replace ‘complexes’ and/or ‘compounds’ by the noncommittal ‘adduct’ and use a {donor–acceptor} symbolism. The sequence of events leading to the discovery of high conductivity in {[TTF][TCNQ]} seems to have been as follows. The synthesis of TCNQ and its ability to form highly-conducting organic salts (conductivities of 102 S/cm) had been reported in 1962 (Melby et al., 1962); rapid progress was made during the next decade in determining and relating crystal structures (Shibaeva and Atovmyan, 1972) and physical properties of relevant materials. Dibenzo-TTF (then called 2,2 0 -bis-1,3-benzdithiolene) had been prepared in 1926 (Hurtley and Smiles, 1926). TTF itself was reported by a number of groups towards the end of the 1960s (Prinzbach et al., 1965; Wudl et al., 1970). TTFþCl was prepared by Wudl and coworkers in 1970, its conductivity at 25 C being reported as 0.2 S/cm in 1972 (Wudl et al., 1972). The crucial step of combining TTF and TCNQ was taken very soon thereafter, essentially simultaneously, by research groups at Monsanto, Johns Hopkins and the University of Pennsylvania (who published their results (Miles et al., 1972; Ferraris et al., 1973; Coleman et al., 1973) and at SUNY-Buffalo (who did not). To quote from Wudl’s lively account: (Wudl, 1984) n n n

‘‘Having established that TTF salts are highly conducting, it did not take long before the marriage of the young donor with the old acceptor took place. The exact location of the ceremony is somewhat obscure. It is clear it was a fertile coupling as judged from the number of papers it generated between 1973 and the present. The sudden interest (1973–1975) in this solid was caused by two, almost simultaneous, events: the discovery of metal-like conductivity between room temperature and 56K and the report of ‘superconductivity’ just above 56K. Of these only the former survived scrutiny.’’

1150


8000

6 5

6000 4

{[TTF ] [ TCNQ]}

Conductivity {[TTF] [TCNQ]} 4000 [010] S/cm

3

Conductivity Lead MS/cm

2

2000 1 Pb 0

0 0

100

200 T(K)

300

400

Fig. 17.1. Schematic representations of the conductivities of Pb (Onnes and Tuyn, 1929) (righthand scale; the superconductivity of Pb below 7.2K has been omitted from the figure) and {[TTF][TCNQ]} (along the stack axis; left-hand scale) as functions of absolute temperature; note that Pb is about 1000 times more conductive than {[TTF][TCNQ]}Pb has been chosen for comparison because its Debye temperature is not very different from that of {[TTF][TCNQ]}, which is considered to have metallic behaviour down to 60K because its –T curve follows the same course as that of Pb and other metals.

While much interesting physics and chemistry has been developed in this area since 1973, it is perhaps fair to state that no system has yet been found with properties more striking than those of the classical {[TTF][TCNQ]} molecular complex, although the discovery of superconductivity in the Bechgaard salts (TMTSF)2X (Friedel and Jerome, 1982) may require some modification of this judgment. Our primary concern in this chapter is with the structures of segregated stack molecular complexes and related ion radical salts and the properties that derive from these structures rather than with the phenomenon of high conductivity as such, which has been the main interest of most investigators in this field (Howard, 1988). Thus we do not discuss many important high conductivity materials, such as the essentially inorganic materials reviewed elsewhere (Miller and Epstein, 1976). However, we have considered the TTF and TCNQ salts with closed shell counterions (and analogs) to come within our boundaries, both because of their intrinsic structural interest in the context of stacked arrangements and because of their relevance to {[TTF][TCNQ]} and related molecular complexes; similar latitude has been extended to some other ion radical salts. A rough picture of the changing interest in TCNQ, TTF and {[TTF][TCNQ]} can be obtained from Fig. 17.2, where the total number of entries under each of these three headings (in practice using the Registry Numbers) in the quinquennial indexes of Chemical Abstracts has been plotted against the end-year of the quinquennium. The interest in TCNQ and TTF appears to have peaked in the early 1990’s, as other donors (many based on TCNQ and TTF) and other types of ion radical salt have taken over part of the stage. {[TTF][TCNQ]} seems to have disappeared from the area of active research but surely it is not yet completely understood.

CHEMISTRY OF DONORS AND ACCEPTORS

1151

600 TCNQ

Number of publications

500

TTF 400 TTF–TCNQ 300 200 100 0 1977

1982

1987

1992

1997

1992

1997

2002

End of quinquennium

Fig. 17.2. The numbers of references to TCNQ, TTF and {[TTF][TCNQ]} in Chemical Abstracts for 5-year periods are plotted against the end-year of the quinquennium. The numbers of references have been obtained from the Registry Numbers for the three materials using Scifinder ScholarTM but no attempt has been made to check their relevance or eliminate duplication.

17.2 Chemistry of donors and acceptors that participate in segregated stacks 17.2.1 Introduction The donors and acceptors of interest here are oxidized or reduced by multistage electron transfers without the formation or rupture of electron pair single bonds. Such reactions have been summarized in terms of three equations by Deuchert and Hu¨nig (1978), whose treatment we follow closely. Their structural principle is stated as follows: ‘‘Reversible redox reactions with transfer of two electrons in two separate steps are to be expected with compounds in which the end groups X and Y of the reduced form 1. have free electron pairs or -systems available, and 2. are connected by vinylene groups (n ¼ 0, 1, 2 . . . ).’’ The various stages are represented by the reduced (RED), radical (‘semiquinone’ or SEM) and oxidized (OX) forms. The systems A and B differ only by two charges; otherwise they are iso--electronic. The SEM radical appears as a cation in A and an anion in B. System C, where the SEM stage is a neutral radical, is less relevant in the present context. All three redox systems are capable of many variations. Deuchert and Hu¨nig (1978) define two particular types of redox system which contain most of the examples of donors and acceptors to be considered below: Two-stage Wu¨rster-type redox systems: two-stage redox systems are said to be of the Wu¨rster type if their end groups are located outside a cyclic -system that has aromatic character in the reduced form.


1152

E1 –e $ þe

REDa

E2 –e $ þe

SEMaþ1

OXaþ2

Equation A X–(–CH¼CH–)n–X

–e $ þe

{X–(–CH¼CH–)n–X þ $ þX–(–CH¼CH–)n–X}

–e $

þ

X–(–CH¼CH–)n–Xþ

þe

Equation B –

Y–(–CH¼CH–)n–Y–

–e $ þe

{–Y–(–CH¼CH–)n–Y $ Y–(–CH¼CH–)n–Y–}

–e $

Y–(–CH¼CH–)n–Y

þe

Equation C X–(–CH¼CH–)n–Y

–e $ þe

{X–(–CH¼CH–)n–Y $ þX–(–CH¼CH–)n–Y–}

–e $

þ

X–(–CH¼CH–)n–Y

þe

Notes: 1. (electron loss (oxidation) occurs for the forward directions of the arrows, and conversely). 2. SEM is shown as a resonance hybrid.

Two-stage Weitz-type redox systems: two-stage redox systems are said to be of the Weitz type if their end groups form part of a cyclic -system that has aromatic character in the oxidized form. In the discussion that follows the chemical formulae of the donors and acceptors are given together with partial reference to the appropriate equations and values of n.

17.2.2

Donors

By the early 1970s it had become clear that the following donors had a high tendency to appear as cation radicals in appropriate systems (ionization potentials are given in Table 13.5): TMPD TTF TTT

N,N,N 0 ,N 0 -tetramethylphenylenediamine tetrathiafulvalene tetrathiotetracene.

Some effort has also been invested in the synthesis of related systems, especially those based on pyranylpyrans. An extensive review, especially of newer examples, is given by Yamashita and Tomura (1998). TMPD forms stacks in some of its cation-radical salts (see Section 17.3.4) but mixed rather than segregated stacks in most molecular compounds with acceptors. It has a Wu¨rster type redox system.


1153

Equation A: H

H N

H N

N

H

H reduced form

H +

N

H

+

H oxidised form

The major synthetic efforts applied to TTF, and to many other donors, have been in two directions: (i) replacement of S by its congeners Se and Te; for example, in the TTF and TTT systems the Se and Te congeners have been synthesized. Three important trends to be expected upon replacement of S or Se by Te have been noted (McCullough et al., 1987). We quote: The more diffuse p and d orbitals centred on tellurium should give larger conduction bandwidths due to increased interstack interactions and result in materials with reduced electron scattering and enhanced metallic electrical conductivity. In addition this increase in orbital spatial extension ought to increase the interchain interactions giving rise to more two- or three-dimensional character. This extended dimensionality should help suppress the various instabilities which often lead to insulating ground states in quasi-one-dimensional organic conductors. Finally, the greater polarizability of tellurium should reduce the on-site Coulombic repulsion and help support doubly-charged species. Unless the molecular component can support doubly-charged species, only a correlated type of conductivity is possible.

(ii) preparation of substituted TTFs and congeners; in particular, alkylthio substituents could alter the dimensionality of the crystals by increasing interstack interactions. The relevant complexes are shown below, together with notes and references; synthetic details are in the original papers. Most synthetic studies have been accompanied by the preparation of donor – TCNQ molecular complexes of various compositions and measurements of electrical conductivities on compacted powders or single crystals; a wide range of conductivities is often found (a list of over 400 crystalline conducting organic quasi-one-dimensional molecular complexes has been compiled by Howard (1988)). Other acceptors have sometimes also been used. Structural information for correlation with physical properties is slowly being accumulated. (i) 1,2,4,5-Tetrakis(dimethylamino)benzene is a powerful donor of the Wu¨rster type (Eqn. A, n ¼ 0) (Elbl et al., 1986). 1,6-Diaminopyrene – TCNQ has a conductivity of 2 S/cm (Scott et al., 1965); its structure is not known. 2,7-Bis(dimethylamino)tetrahydropyrene – TCNQ has a conductivity of 0.4 S/cm and has been estimated from the CN stretching frequency to have a charge transfer of 0.57; its structure is not known. A number of related molecular complexes of low conductivity have been prepared (Ueda et al., 1983). Fully aromatic 2,7-bis(dimethylamino)- and 1,3,6,8-tetrakis-(dimethylamino)-pyrene and partially-reduced 2,7-bis(dimethylamino)pyrene have been shown to be good electron donors (Sakata et al., 1984).


1154

(ii) Dithiapyrenes (Tilak, 1951) (Eqn. A, n ¼ 2) and dithiaperylenes (Eqn. A, n ¼ 3). These are Weitz type systems. S

R2 R1 S

S S R1 R2

1,6-dithiapyrene (DTPY; R1 ¼ R2 ¼ H). When R1 ¼ SCH3, R2 ¼ H the complex is 2,7-bis(methylthio)-1,6-dithiapyrene and when R1 ¼ H, R2 ¼ SCH3 it is 3,8-bis(methylthio)-1,6-dithiapyrene.

3,10-dithiaperylene (DTPR)

{[DTPY][TCNQ]} has a segregated stack crystal structure (see Table 17.10 ) (Thorup et al., 1985; DAKTIS); the structures of some salts of 2,11-diphenyl-DTPR have been reported (Nakasuji et al., 1986) and also of some molecular complexes of substituted DTPYs (see Table 17.10). 3,4:3 0 ,4 0 -Bibenzo[b]thiophene (BBT), the thiophene analog of DTPR (i.e. with the S atoms in five-membered rings) has been prepared (Wudl et al., 1978) and is isoelectronic in its neutral state with perylene. Structures of molecular complexes have not been reported. S S

S

S

S

S

S

S S

4,5:9,10-bis(ethanediylthio)1,6-dithiapyrene (Nakasuji et al., 1986)

S S S

2,3:7,8-bis(ethanediylthio)-1,6-dithiapyrene (ETDTPY) (Nakasuji et al., 1986)


1155

The crystal structures of the TCNQ (FUDTON) and chloranil (FUDTUT) adducts of 3,8-bis(methylthio)-1,6-dithiapyrene have been reported (Nakasuji et al., 1987). (iii) Benzotrichalcophenes (Cowan et al., 1982): X

X

X

X ¼ S (BTT) (Hart and Sasaoka, 1978), Se (BTS), Te (BTTe)

BTT forms intensely coloured, moisture-stable crystalline 1 : 1 charge transfer adducts with TCNE, DDQ, TCNQ and chloranil. Structures and physical properties do not appear to have been reported. (iv) Systems based on TTF and congeners (Eqn. A, n ¼ 1; these are Weitz type systems); syntheses have been extensively reviewed (Narita and Pitman, 1976; Krief, 1986; Schukat et al., 1987); Nielsen et al., 2000; Yamada and Sugimoto, 2004). The basic structure (below left) has been substituted in many ways, some of which are illustrated and others referenced (Fangha¨ngel et al., 1983). R

Ch

R

Ch

Ch

Ch

Ch

Ch

R

R R

Ch

Ch

R

TTF : Ch ¼ S, R ¼ H; TSF (or TSeF) : Ch ¼ Se, R ¼ H TTeF : Ch ¼ Te, R ¼ H (Narita and Pitman, 1976) Tetrakis(alkylthio)TTF (Mizuno et al., 1978): Ch ¼ S, R ¼ SCH3; TMTTF : Ch ¼ S, R ¼ CH3; TMTSF : Ch ¼ Se, Te, R ¼ CH3 (Iwasawa et al., 1987)

HMTSF (Berg et al., 1976) : Ch ¼ Se, R ¼ CH21; HMTTeF (Saito et al., 1983) : Ch ¼ Te, R ¼ CH2 DTTSF (Elbl et al., 1986): Ch ¼ Se, R ¼ S

H3C

CH3 S

S NR

RN S H3C

S CH3

Bis(2,5-dimethylpyrrolo[3,4-d]tetrathiafulvalene: R ¼ H (BP-TTF); Ph (BPP-TTF) (Chen etal., 1988) 1

2, 2 0 -bis(2,4-diselenabicyclo[3.3.0]octylidene).


1156

Ch

Ch

Ch

Ch

Ch

Ch

Ch

Ch

Ch

Ch

Ch

Ch

DBTTF : Ch ¼ S (Spencer et al., 1977) (Unsymmetrical versions with only one benzene ring have been prepared, and also symmetrically (E) and unsymmetrically (Z) substituted dimethyl complexes) (Shibaeva and Yarochkina, 1975; Tanaka et al., 1983; Nakano et al., 1989) DBTSF : Ch ¼ Se

Bis(ethylenedithiolo)TTF (BEDT-TTF) Ch ¼ S (Saito et al., 1982); Ch ¼ Se (Lee et al., 1983).

(iii) Systems based on pyran derivatives: R

R S X

X

S S

R

R

X ¼ O, S, Se, Te; (Alizon et al., 1976; Detty et al., 1983); X ¼ S, R ¼ H or CH3 (Sandman, Epstein et al., 1977). See also Sandman, Fisher et al., 1977. 4,4 0 -Bithiopyranylidene (BTP, X ¼ S, R ¼ H) is isoelectronic with TTF. Note: when X ¼ O and R ¼ C6H4C12H25, then charge transfer salts (counterions BF4, ClO4, TCNQ) with appreciable mesophase (liquid crystalline) temperature ranges are found (Saeva et al., 1982).

(iv) Systems based on p-quinobis(1,3-dithiole): R

S

S

R

R

S

S

R

R ¼ CH3, TMCHDT (Bis(4,5-dimethyl-2H-1,3-dithiolylidene-2)-1,4-cyclohexa-2,5-diene) (Fabre et al., 1978).

R

S

S

R

R

S

S

R

R ¼ H (Sato et al., 1978), H, CH3 (Ueno et al., 1978). Also analogues based on the anthracenediylidene system (Bryce et al., 1990). Dithiadiazafulvalenes have been shown to be strong electron acceptors (Tormos et al., 1995).


1157

(v) Systems based on tetrathiotetracene: X

X

X

X

X ¼ S TTN (dehydrotetrathianaphthazarin). Se TSN (naphtho[1,8-cd:4,5-c 0 d 0 ]bis(1,2-diselenole)

TTN was shown electrochemically to be a much poorer donor than TTF; however, {[TTN] [TCNQ]} crystallizes as black needles with a single crystal conductivity of 40 S/cm between 200–300K, which is higher than that of {[TTT][TCNQ]} (1 S/cm) and the same as the microwave conductivity of {[TTT][(TCNQ)2]} (Wudl et al., 1976). Crystal structures are discussed later. X

X

X

X

Equation A, n ¼ 2,3. Se TSA (tetraseleno-anthracene; anthra[9,1-cd:10,5-c 0 d 0 ]bis(1,2-diselenole)) (Endres et al., 1982)

X

X

X

X

X ¼ S TTT (naphthacene[5,6-cd: 11,12-c 0 d 0 ]bis(1,2-dithiole)) Se TSeT (Khidekel and Zhilyaeve, 1981) Te TTeT (Sandman et al., 1982)

(vi) Systems with both N and S atoms: S N R

.

S

R ¼ H, CH3 Benzo-1,3,2-dithiazol-2-yl and derivatives (Wo¨lmerhauser et al., 1984).

17.2.3 Acceptors The synthesis of TCNQ (Acker and Hertler, 1962) was accompanied by a thorough investigation of its ion radical salts (Melby et al., 1962) (the abstract of this paper is quoted at the head of this chapter); there are examples (Yamaguchi et al., 1989) of alternative syntheses. A number of substituted TCNQs have been synthesized as well as


1158

analogs of various kinds. A new family of segregated stack crystals is based on TTF with tetrahalo-p-benzoquinones as acceptors (Torrance et al., 1979). Bis(dithiolene)metal complexes behave as acceptors with many cations (Interrante et al., 1975); the crystal structures are rather different from those of TCNQ molecular complexes and the materials have interesting physical properties. The various acceptors are compiled below; structures and properties are discussed at appropriate places later in this chapter. (i) TCNQ and related complexes: Equation B: NC

NC

CN C–

C–

NC

CN C

NC

CN

CN oxidized form

reduced form

7,7,8,8-tetracyanoquinodimethane (TCNQ) (Equation B, n ¼ 3; these are Wu¨rster type systems). Among the derivatives synthesized are monofluoroTCNQ (Ferraris and Saito, 1979), 2,5difluoro-TCNQ (Saito and Ferraris, 1979), tetrafluoro-TCNQ (TCNQF4) (Wheland and Martin, 1975), and methyl-, ethyl- and 2,5dimethyl-TCNQ (Andersen and Jorgensen, 1979).

C

N N N N

11,11,12,12-tetracyanonaphthoquinodimethane (TNAP) (Diekman et al., 1963) (Equation B, n ¼ 4)

Some three-dimensionally modified TCNQ derivatives such as dihydro- and tetrahydrobarreleno-TCNQ, monobenzo- and dibenzobarrelleno-TCNQ have been NC

CN

NC

CN

Tetracyanodiphenoquinodimethane (Aharon-Shalom et al., 1979) (Equation B, n ¼ 7)

NC

CN

NC

CN

13,13,14,14-Tetracyano-4,5,9,10-tetrahydropyrenequinodimethane (TCNTP) (Aharon-Shalom et al., 1979; Maxfield et al., 1979) (Equation B, n ¼ 5); see also Suguira et al., 2000.


1159

prepared and shown to form molecular complexes of varying degrees of ionicity with TTF derivatives (Nakasuji et al., 1986). 11,11,12,12,13,13,14,14-octacyano-1,4:5,8-anthradiquinotetramethane{OCNAQ;1,4:5,8tetrakis-(dicyanomethylene)anthracene} has been prepared (Mitsuhasi et al., 1988) and shown to form molecular complexes with TTF, pyrene, phenothiazine, TMTTF and TTT. Structures (Inabe et al., 1988) are discussed later. The acceptor is somewhat nonplanar because of steric hindrance between adjacent C(CN)2 groups. (ii) Other polycyano complexes:

CN

NC

NC

N

N

N

N

CN

CN NC

NC NC

CN

H

Hexacyanobutadiene (HCBD), also called tetracyanomuconitrile (Webster, 1964).

NC

CN

NC

CN

CN

H

Tetracyanobi-imidazole (Rasmussen et al., 1982).

O2N

NO2

NC

CN

O2N

Tetracyanoethylene (TCNE) (Equation B, n ¼ 1)

2,4,5-Trinitro-9-(dicyanomethylene)fluorene (Dupuis and Neel, 1969)

(iii) S-heteroquinoid acceptors: The tetracyanothieno[3,2-b]thiophene system, where the tetracyano groups bracket a heteroquinoid nucleus, has been used as a basis for acceptors (Yui et al., 1989). The corresponding dicyano complexes (2,5-bis(cyanoimino)-2,5-dihydrothieno[3,2b]thiophenes (DCNTTs)) are also good acceptors in charge transfer complexes and, especially, in highly conductive salts (Aumu¨ller et al., 1988). R1 NC

R1 S

CN

NC

CN

S R2

CN

S N NC

N S DCNTT

R1 and R2 are various combinations of H, CH3, Cl, Br.

R2


1160

(iv) Other acceptors containing S and N atoms: N

N S

S N

N BDTA

Bis([1,2,5]thiadiazolo)[3,4-b;3 0 ,4 0 -e]pyrazine (BDTA; the formula is drawn to indicate that the molecule is best represented as a resonance hybrid) is a 14 electron heterocycle (Yamashita et al., 1988), with a higher electron affinity than that of BDTA-TCNQ below. Its crystal structure has been determined but formation of molecular complexes and radical ion salts has not been reported in detail. NC

CN

N

N

N

N

NC

CN

NC

CN

N S

S

NC

S N

CN

BDTA-TCNQ

TDA-TCNNQ

Two related acceptors are (bis[1,2,5]thiadiazolo)TCNQ (BDTA-TCNQ) (Se and S, Se analogs have also been reported (Suzuki et al., 1987)) and [1,2,5]thiadiazolo)tetracyanonaphtho-quinodimethane (TDA-TCNNQ). The electron donors TTF, TMTTF and TMTSF gave very low conductivity (109 S/cm) adducts with BDTA-TCNQ and these are probably mixed stack compounds; however, the TTN molecular compound had a conductivity of 1 S/cm and could have segregated stacks (Yamashita et al., 1985). Structures have not been reported. (v) Tetrahalo-p-benzoquinones and analogs: O X

Y

X

Y O

X ¼ Y F, Cl, Br, I X ¼ CN, Y ¼ Cl 2,3-Dichloro-5,6-dicyano-pbenzoquinone (DDQ)

NC

Various substituents CN

N,N 0 -dicyanoquinone diimines (Aumu¨ller and Hu¨nig, 1986a,b)


1161

(vi) Aromatic anhydrides: 1,4,5,8-naphthalenetetracarboxylic anhydride (NDTA) forms conducting salts with various cations. The salt (5,6-dihydro4a,6a-phenanthrolinium)2(NDTA)5 has a stacked structure (Heywang et al., 1989; VARKEE, P 1, Z ¼ 1; Born and Heywang, 1991; VARKEE10; see Section 17.4.5). (vii) 1,2-Ethylenebis(dithiolene)metal complexes (see Section 17.5.2): X

S

S

X

n– n– S

M

Rn+ X

S

S

X

n ¼ 0, 1, 2. X ¼ H, CH3, CN (abbreviated as mnt which is maleonitrile dithiolate or cis-2,3dimercapto-2-butenedinitrile), CF3 (abbreviated as tfd), C6H5 etc. M ¼ Ni, Pd, Pt, Fe, Co, Cu, Au etc.

S

Rn+ S

S

S

M S

S

S S

S

n ¼ 0.5, 1, 2. Abbreviated as dmit. H2(dmit) ¼ 4,5-dimercapto-1,3-dithiole2-thione.

17.2.4 Preparation of crystals A feature of most synthetic papers is that some charge transfer radical ion salts crystallize without difficulty, generally as needles of intense colour (deep green, blue or black) showing metallic reflectivity; on the other hand, some products are obtained only as microcrystalline powders. The reason for such behaviour is generally not understood. Standard methods are used for growth of crystals or recrystallization of powders – slow cooling of saturated solutions, slow evaporation, interdiffusion of components in U-tubes (Kaplan, 1976), simultaneous sublimation of components on to cold surfaces (Andersen, Engler and Bechgaard, 1978). However, many beautiful crystals show disappointing diffraction patterns, the apparent external perfection being accompanied by appreciable disorder at the unit cell level. Electrocrystallization (Chiang et al., 1971; Ristagro and Shine, 1971; Rosseinsky and Kathirgamanathan, 1982) is a technique of increasing popularity for growing crystals of highly conductive materials. It was first used for the perchlorates of pyrene, perylene and azulene and later extended to {[perylene][Ni dithiolate]} (Alcacer and Maki, 1974) and various salts of TCNQ and analogs (Kathirgamanathan et al., 1979, 1980, 1982). Both potentiostatic and galvanostatic arrangements have been used (Fig. 17.3), and it has been shown that the crystals growing on the anode function directly as electrodes (Enkelmann, Morra et al., 1982). The nature of the solvent can also be important (Anzai et al., 1982). The highly conductive materials considered here are usually very anisotropic, with the metallic conductivity generally restricted to the stack axis. Thus chemical impurities and lattice imperfections would be expected to strongly influence the one-dimensional electron transport, and more so than in three-dimensional metals. This anticipation is reinforced by the many reports that the conductivity of {[TTF][TCNQ]} is sample dependent (Fig. 17.46). The matter is somewhat controversial. One group (McGhie et al., 1974, 1978) claims that extreme precautions (gradient sublimation under argon, use of quartz


1162

– –

+

70~80

100~125

+

15

50 30

Fig. 17.3. Example of a crystal growth cell used for growing highly conductive crystals in a nitrogen atmosphere. Concentrations of 104 M were used, with voltages in the range 1–5 V and currents in the range 1–10 mA. Dimensions in mm. (Reproduced from Anzai et al., 1982.)

apparatus) are essential while another (Gemmer et al., 1975) suggests that the simple standard techniques of recrystallization and sublimation are adequate to yield TTF and TCNQ of excellent chemical purity (impurities at the ppm level as judged by the use of high performance liquid chromotography (HPLC)). Gemmer et al. (1975) conclude that the variations in the conductivity of different {[TTF][TCNQ]} samples are due to lattice defects. McGhie et al. (1974, 1978) emphasize that large well formed crystals of {[TTF][TCNQ]} are obtained only when high purity components are used and that the physical perfection of the crystals depends both on the chemical purity of the components and the growth conditions.

17.3 17.3.1

Structures of cation-radical salts Introduction

The structures of a fairly large number of cation-radical salts have been determined in the last few years and many of these are important in the context of high conductivity (Shchegolev and Yagubskii, 1982; Shibaeva, 1982)), or even superconducting (Williams, Beno et al., 1985) materials. Space limits us to considering here only the cation-radical salts derived from polycyclic aromatic hydrocarbons because of their fundamental position in the crystal chemistry of charge transfer molecular complexes, and the salts of TTF, which illustrate some of the principles found also in other cation radical salts, and where the cation participates in the much-studied {[TTF][TCNQ]}.

STRUCTURES OF CATION-RADICAL SALTS

1163

17.3.2 Cations are polycyclic aromatic hydrocarbons We shall treat these as one structural group with subdivisions; however, there are already enough structural types to suggest that the true picture may well be more complicated. Three structural subdivisions can be discerned – those in which all the radical cations are in stacks, those in which some are stacked and others not, and a non-stacked group.

17.3.2.1 Stacked radical cations {(Naphthalene)þ PF6} forms dark red-violet conducting ( ¼ 0.12 S/cm) crystals 2 ˚ , space group P42/n, (grown by electrocrystallization; tetragonal, a ¼ 11.56, c ¼ 6.40 A Z ¼ 2) that are stable at low temperature but decompose on warming to room temperature (Fritz et al., 1978; NAPHFP10); the AsF6 salt (ZZZBTG) is isomorphous. The structure consists of segregated stacks of (C10H8)2þ cations and PF6 anions along [001]. The

0

A hexafluorophosphate

B x y Fig. 17.4. Crystal structure of {(naphthalene)þ 2 PF6 } at 223K projected down [001]. The diagram shows the mutual arrangement of nearly coplanar cations and anions. The 1-9 and 2-3 bonds in the naphthalene cation are significantly shorter than those in neutral naphthalene at 90K (Brock and ˚ and 1.386(6) as against 1.415(2) A ˚ ). (Data from Fritz Dunitz, 1982) (1.408(6) as against 1.423(2) A et al., 1978.)

1164


cations lie in (002) planes and are separated by the remarkably small interplanar distance ˚ , successive cations along the 42 axis of a stack being mutually rotated by of 3.20 A 90 ; presumably there is a random arrangement of neutral molecules and cations in each stack and this accounts for the conductivity. The packing arrangement (Fig. 17.4) is similar to those of bis(diphenylglyoximato)Ni(II) iodide and Ni(phthalocyanine) iodide (Marks, 1978). The preparation of some of the above and related salts [e.g. (triphenylene)2X, X ¼ PF6 or AsF6; (perylene)2BF4] has been reported (Kro¨hnke et al., 1980). Conductivities ranged from 103 to 101 S/cm. The salts (fluoranthene) þ 2 X (X ¼ PF6 (FANTHP10), AsF6 (BOSJUO), SbF6 (BUNCAO)) are isomorphous (a ¼ 6.50, b ¼ 12.49, c ¼ ˚ , ¼ 104.0 , P21/c, Z ¼ 2 for hexafluoroarsenate at 120K) and, judging 14.75 A from the curves of cell dimensions versus T, all show second (or perhaps higher) order phase transitions at 200K (Enkelmann, Morra et al., 1982). The small differences between high and low temperature structures derive from small rotations of cations and anions; adjacent stacked fluoranthenium cation radicals are separated ˚ . The dimensions found for the fluoranthenium cation radical are by 3.22 and 3.28 A not precise enough for comparison with those of the neutral molecule. The salts are quasi-metallic above 200K and semiconducting below. In both (naphthalene) þ 2 X and (fluoranthene) þ X the cation planes are exactly or nearly normal to the 2 stack axes. Three isomorphous bis(perylenium) salts [(pe)2(X)x(Y)y, with X ¼ PF6 (CUWBAX01), AsF6 (PITDOL), x ¼ 0.8 1.5; Y ¼ CH2Cl2, C4H8O, y ¼ 0.5, 0.8] have been studied (Keller, No¨the et al., 1980); the black needles were grown by electrocrystallization. These have structures (orthorhombic, a ¼ 4.285, b ¼ 12.915, c ¼ ˚ , space group Pnmn, Z ¼ 1) different in many respects from that of 14.033 A (naphthalene) þ PF6. The translationally equivalent perylenium cation radicals are 2 ˚ and an inclination of 38 to stacked along [100] with an interplanar spacing of 3.40 A (100) and are mutually shifted along their long axes. The channels between the cation stacks are filled with an ill-defined melange of anions and solvent molecules. The stack axis DC conductivities are in the range 102–103 S/cm at room temperature, follow a metallic regime down to 200K and then diminish rapidly. Keller, No¨the et al. (1980) suggest that the rather high conductivities are due to enhancement of the mixed valence state of the cation radical stacks by non-stoichiometry arising from replacement of some anions by neutral solvent molecules. Other perylene charge transfer salts are discussed in Section 17.5.2. 17.3.2.2

Structures in which not all cations are stacked

There is a second conducting form of (pe)2(PF6)2/3(THF) ( 102 S/cm; CUWBAX) that has infinite stacks of perylene moieties, with a slight tendency to the formation of tetrads (Endres et al., 1985). However, only half the perylene moieties are in stacks, while the others flank the stacks on all four sides, the planes of the flanking perylenes being virtually perpendicular to the planes of those in the stacks (Fig. 17.5). Endres et al. (1985) note that ‘‘intuition would suggest that the charges are distributed in the stacks but there are some physical arguments against this idea.’’ Thus we shall not venture an explanation for the conductivity.


1165

Bis(perylenium) hexafluorophosphate. THF stack of perylenes

flanking perylenes with anions and solvent molecules

y x

Fig. 17.5. Crystal structure of (pe)2(PF6)2/3(THF) viewed down [001] (a ¼ 13.076, b ¼ 14.159, ˚ , ¼ 110.87 , space group P2/m, Z ¼ 3). (Data from Burggraf et al., 1995.) c ¼ 13.796 A

Fig. 17.6. Stereoview of the (pe)3ClO4 structure viewed along [001], showing the stacked but jogged tetrad and the flanking perylene moieties. The ClO4 anions are ordered. (Reproduced from Endres et al., 1985.)

1166


(Perylene)3ClO4 is a semiconductor in which there are two offset pairs (-dimers) of perylene moieties per triclinic unit cell, thus forming jogged stacks (Endres et al., 1985; CUWBEB); the interplanar distances within and between the -dimers are not sensibly ˚ . Adjacent stacks are separated by flanking perylenes (Fig. 17.6), different at 3.37(5) A very much as in (pe)2(PF6)2/3(THF). It was not possible to infer moiety charges from bond lengths because of disorder problems. In {(perylene)4[Co(mnt)2]3} there are three perylenes stacked as trimers, a trinuclear [Co(mnt)2]3 moiety and a flanking perylene approximately normal to the perylenes (Gama et al., 1993; see Section 17.5.2). Thus there are a number of examples of flanking perylenes and analogs have been encountered in other structures; (TTF)2-[Ni(S4C4H4)] (see Section 17.3.7) is a possibly relevant example for comparison. The overlap of the perylene pairs or triads, whether stacked or not, has been described as ‘graphite-like’ but ‘ring-over-bond’ is more appropriate. The structure of (quaterphenyl)12(quaterphenyl)4(SbF6)10 has been briefly reported (Enkelmann, Go¨ckelmann et al., 1985). The structure consists of stacks of quaterphenyl cations (the first group) separated by layers of anions, in which the second group of cations is incorporated. Thus there appear to be resemblances to (pe)2(PF6)2/3(THF) but a detailed comparison is not possible.

17.3.2.3 Structures in which the cations are not stacked (Pe)6ClO4 has three independent -dimers arranged about centers of symmetry of the triclinic unit cell; the overlap diagrams are very similar and the interplanar spacings ˚ (Fig. 17.7). This material is a semiconductor are 3.38(2), 3.44(2) and 3.46(3) A ( 102 S/cm (Endres et al., 1985; CUWBIF).

Fig. 17.7. (Pe)6ClO4, stereoview along [100], showing the herring-bone arrangement of -dimers. The disordered anions (at symmetry centres) have been omitted for clarity. (Reproduced from Endres et al., 1985.)


1167

17.3.2.4 Various The crystal structure of (pyrene)þ has not been reported. In (pheno2 (ClO4) þ thiazine)2 SbCl6 (BUFZEH) the cation radicals are isolated in the crystals and do not interact (Uchida et al., 1983). 17.3.2.5 Conclusions The structures of the aromatic hydrocarbon cation-radical salts show some interesting resemblances to the cation-radical salts of TTF discussed below. There are non-stacked and stacked -dimers and infinite monad and diad stacks. Formation of a cation-radical anion-radical salt analogous to {[TTF][TCNQ]}, in which the cation radical is an aromatic hydrocarbon and the anion radical is a suitable open shell acceptor and the stacks are segregated, has so far been achieved only for perylene–M(mnt)2 derivatives (see Section 17.5.2). 17.3.3 TTF and related compounds as cations TTF forms salts with many closed shell anions. Quite complex relations were found early on (Scott, La Placa et al., 1977) for the TTF halides (Fig. 17.8) and the situation becomes even more complicated when TTF salts of polyatomic anions are considered. Our classification is based on the overall structural arrangement, and we make a major distinction between nonstacked and stacked arrangements of the TTF moieties. Type A: Non-stacked arrangements of TTF moieties. Group 1: Salts of type (TTF2þ)(X)2 are isostructural for X ¼ Cl (TTFDCL), Br ˚ (for X ¼ Cl), space group (TTFDBR). These are tetragonal with a ¼ 13.56, c ¼ 10.10 A

4b

Chloride

Bromide 4b

Iodide

0.6

0.8 1.0 Effective Charge (y)

2.0

Fig. 17.8. Phases observed in the TTF-halide systems, indicated by vertical lines. The abscissa y is the halogen content defining the effective or average charge per cation site, thus the compositions are given as TTFXy. The rectangles show the composition ranges found for some salts. (Adapted from Scott et al., 1977.)


1168

I41/acd, Z ¼ 8 (Scott, La Placa et al., 1977). (TTF2þ)(ClO4)2 is monoclinic (Ashton et al., 1999, who give references to other (TTF2þ) structure determinations). In both instances, the cations are nonplanar with the two halves of the moiety mutually rotated by 60 about the central C–C bond. Ashton et al. comment ‘‘We believe, in all probability, that the conformation of (TTF2þ) is determined by the multiple interactions that determine its supramolecular order.’’ TTF(OCN)2 has been reported as a microcrystalline insulating powder (Kathirgamanathan and Rosseinsky, 1980). Group 2: Salts of type TTFX, where X ¼ Cl (TTFMCL), Br (TTFMBR). The iso˚ (for morphous halide salts are orthorhombic, with a ¼ 11.073, b ¼ 11.218, c ¼ 13.95 A X ¼ Cl), space group Pbca, Z ¼ 8 (Scott et al., 1977). There are centrosymmetric pairs ˚ (-dimers), and ordered anions of eclipsed cations with interplanar spacing 3.34 A (Fig. 17.9). Their importance in the present context is to suggest that adjacent neutral TTF molecules and TTFþ cation radicals can be present in a stack without seriously perturbing the stack. Other examples are TTFI3, (Teitelbaum et al., 1980; TTFIOD), TTF3Sn(CH3)2Cl3 (Matsubayashi et al., 1980) and TTFDDQ (Mayerle and Torrance, 1981b). The TTFI3 structure has pairs of cations (TTFþ)2 surrounded by noninteracting triiodide ions, while TTF3Sn(CH3)2Cl3 has similar (TTFþ)2 dimers in a matrix of centrosymmetric 2 chlorine-bridged dimeric [Cl2(CH3)2Sn(Cl)2Sn(CH3)2Cl2] moieties. TTFDDQ (dark red ˚ , ¼ 77.51(2), ¼ triclinic crystals, a ¼ 10.272(4), b ¼ 12.195(5), c ¼ 6.609(2) A 81.93(2), ¼ 87.30(4) , P1, Z ¼ 2; Fig. 17.10) has eclipsed TTF -dimers (interplanar ˚ ) and also stacks of DDQ -dimers with very short interplanar spacings of spacing 3.4 A ˚ ˚ between them. The axes of the two kinds of stacks 2.97 A within the -dimers and 3.56 A

c H

H C

C

S C

S

C

C

S

H

S C

H

Br

S

C H b

Fig. 17.9. Bounded a axis projection of the orthorhombic crystal structure of TTFBr; TTF moieties viewed edge-on. The chloride is isostructural. (Reproduced from Scott et al., 1977.)


b pair of ~ DDQ dimers

1169

TTF dimer

Å

4Å

2.97

3.

a ~

3.5

c ~ Cl

6Å

TTF dimer

C≡N DDQ

b ~

Fig. 17.10. Two projections of the crystal structure of TTFDDQ. The interplanar spacing between ˚ . (Reproduced from Mayerle and Torrance, 1981b.) the DDQ moieties is 2.97 A

are considerably inclined to one another. In these salts the moieties are fully ionized and the conductivities are correspondingly low ( 108 S/cm). TTF.DDQ is probably better described as a DDQ anion-radical salt, with the DDQ moieties arranged in jogged tetrad stacks and the (TTFþ)2 -dimers acting as (possibly spin paired) cations. TTFClO4 (ZZZBWA10) has a rather complicated structure (a ¼ 16.762(1), b ¼ ˚ , Z ¼ 16, space group Pbca) (Yakushi et al., 1980) in which 20.906(2), c ¼ 12.538(1) A ˚) there are pairs of essentially eclipsed cations (-dimers with interplanar spacing of 3.41 A arranged in tetrads. The structure is not stacked. TTF(OSO2CH3), TTF(SCN)1.4 and TTF(HSO4)1.2 have also been reported (Kathirgamanathan and Rosseinsky, 1980). but structures are not known. -Dimers are characteristic of many structures of this group but (TTF)32þ triads arranged 2 (Fig. 17.11) (Kondo et al., between (SnCl6)2 anions are found in (TTF)2þ 3 (SnCl6) 2þ 2 1984; CELREQ); (TTF)3 (PtCl6) (DERKAM) is isomorphous. Salts with [SnCl4R2]2 anions (R ¼ methyl, ethyl) are isostructural (Matsubayashi et al., 1985; ethyl is DERJUF). As the dimensions of the two independent TTF moieties of the triad do not differ significantly, delocalization of the charge across the triad was inferred. Powder conductivity of 2.4 103 S/cm was ascribed to in-layer transport through sulfur–sulfur contacts. Type B: Stacked arrangements. Group 3: (stacks with the long axes of the TTF moieties mutually perpendicular): This group comprises the salts TTFXy, where X ¼ Cl, NO3, SCN (disordered anions), Br, I (ordered anions). The values of y vary from 0.55 to 0.7 and depend on the nature and degree of ordering of the anion. The overall gross structural pattern common to all these salts is rather simple but there are many complications of detail. The basic structural pattern has stacks of eclipsed TTF cation moieties (with average charge Z < 1) arranged as shown in Fig. 17.12 (broken-line cell in the centre of the diagram). The alignment of TTF moieties in adjacent stacks is mutually perpendicular. The channels between the stacks of Fig. 17.12 contain the anions; as Johnson and

1170


S(12) TTF triad

2– (SnCl6)

S(4) S(2)

S(11)

S(3) S(1) Sn b a

Fig. 17.11. The non-stacked structure of (TTF)3SnCl6 showing the layers of (TTF)32þ triads ˚ , P4/mbm, Z ¼ 2. The interspersed among the anions. The crystals are tetragonal, 11.801, 11.861 A ˚ . (Reproduced from Kondo et al., 1984.) interplanar spacing in the triad is 3.49 A

Watson (1976) originally pointed out, anions in these channels with ionic radii greater ˚ will distort the arrangement of the TTF stacks. The prototype structure is than 1.8 A ˚, TTFCl0.7 (MTTFCL), which has a tetragonal unit cell with a ¼ 11.12, c ¼ 3.595 A Z ¼ 2, space group P42/mnm (we use the highest symmetry space group among the possibilities, unless there is good reason to the contrary). The salts TTFBr0.59 (ZZZBGJ), TTFI0.69 (Scott et al., 1977) and TTF(SCN)0.570.03 (Kobayashi and Kobayashi, 1977) have essentially the same structure, while there is a small monoclinic distortion in TTF(NO3)0.55. Two different modes of distortion come into play when the halide content increases slightly. In TTFCl0.7 the unit cell distorts to orthorhombic (see Fig. 17.12) while in TTFBr0.740.77 (LaPlaca et al., 1975) and TTFI0.700.72 (ZZZBGD) the TTF and halide subcells become incommensurable (Table 17.1). The discussion of the mutual interaction of TTF and I sublattices in TTFI0.70, as inferred by Johnson and Watson (1976) from the details of the very complicated x-ray diffraction patterns, is remarkable for its thoroughness but too complex for summary here. Measurements of the peak shapes of the S 2p peaks in the x-ray photoelectron spectra (Ikemoto, Yamada et al., 1980) indicate that the fraction of cations in TTFBr0.7 is 0.71(5) but 0.52(5) in TTFI0.7, suggesting that some larger polyiodide anions may be present in the latter; both salts contain TTF cations and neutral molecules but in different amounts. The system TTFCly (0.67 y 0.70) has been investigated in some detail over a range of temperature (Williams, Lowe Ma et al., 1980). All crystals in this composition range show the same temperature dependence of conductivity; however, x-ray diffraction (oscillation) photographs from crystals at the two extremes of composition show slightly different diffuse layer lines. These have their origins in the disordered Cl


1171

arrangements; the TTF and Cl sublattices are commensurable for y ¼ 0.67 (ZZZBGG) and incommensurable for y ¼ 0.70 (indeed the composition range was inferred from the detailed spacings of the diffuse layers). There is a phase change at 250K for TTFCl0.67 (cf. the conductivity behaviour in Fig. 17.13) and the diffuse lines sharpen and resolve into individual Bragg reflections indicating ordering of the Cl sublattice. The ordered monoclinic phase has composition (TTF)3(Cl)2 and commensurable TTF and Cl sublattices. Orientation relations between the monoclinic and tetragonal phases can be

Table 17.1. The TTF and halide sub-cells for TTFBr0.76 and TTFI0.72. These phases, shown as 4b in Fig. 17.9, have small ranges of composition and the values of c and vary with composition Subcell parameter

˚) a (A b c (deg)

TTFBr0.76 MTTFBR

TTFI0.72 MTTFID

TTF

Br

TTF

I

15.617 15.627 3.573 91.23

17.368 15.623 4.538 116.01

15.998 16.114 3.558 90.96

8.19 16.11 4.871 102.82

2b

2a sin b

Fig. 17.12. Projection down [001] of the structures of all the TTFXy (y < 1) phases, with the exception of TTFCl0.92. The broken-line cell in the center of the diagram is for the disordered ˚ ). The full broken-line rectangle shows tetragonal salts of Group 4a such as TTFCl0.68 (a ¼ 11.12 A ˚ ), while the TTF sublattices of the orthorhombic cell of TTFCl0.77 (a ¼ 10.77, b ¼ 22.10 A TTFBr0.76 and TTFI0.72 are shown in the top left quarter of the overall diagram. Minor distortions ˚ . (Adapted from Scott et al., 1977.) are not shown in the diagram. For all the cells c 3.6 A

1172


300 0

250

T(K)

200

0 –1 –1 260K 3.5

–2 260K

4.5 1000/T (K)–1

–3 167K –4 126K 2

4

6 8 1000/T (K)–1

10

12

Fig. 17.13. Temperature dependence of the conductivity of TTFCl0.67. The ordinate (/RT) is on a logarithmic scale, and the approximate temperatures of the breaks have been inserted into the diagram. (Adapted fromWilliams et al., 1980.)

summarized as follows: p amono ¼ bmono ¼ 2atetr ; cmono 3ctetr ; mono 92 (temperature dependent). Because the four h110i directions of the tetragonal cell are equivalent, four equally probable orientations will be obtained for the monoclinic cell. Thus the low temperature phase is a quintuple twin, consisting of the domains of the four monoclinic orientations and of unchanged tetragonal phase (twinning often accompanies first-order phase changes). Overlap of diffraction patterns prevented determination of the details of the low temperature structure. When TTFCl0.70 is cooled, it forms two types of low temperature phase with compositions (TTF)3(Cl)2 and (TTF)7(Cl)5 respectively; the TTF and Cl sublattices are commensurable in both. The ordered structure of (TSeF)3(ClO4)2 is a monoclinic distortion of the tetragonal subhalide structure, in which the (TSeF) moieties are arranged in triads with their long axes approximately mutually perpendicular (Shibaeva, 1983; CENBIG). Group 4: (stacks with the long axes of the TTF moieties mutually parallel): This group comprises the salts (TTF)3(BF4)2 (Legros et al., 1983; CELVIJ) and (TTF)3(I3)2 (Teitelbaum et al., 1980). The first of these (Fig. 17.14) has slightly jogged stacks of TTF moieties and orientationally disordered anions; bond lengths are said to differentiate between TTF cations and neutral molecules, and XPS measurements of the shape of sulphur 2p peak indicate that 2/3rds of the TTF moieties are cations (Ikemoto et al., 1980). The powder conductivity is 2 105 S/cm. Only a gross structure has been reported for (TTF)3(I3)2 (Johnson et al., 1975).


1173

TTFHgCl3 has a remarkable structure that combines in one crystal features encountered above in separate structures (Kistenmacher, Rossi et al., 1980; FTFHGC10). There are two types of layer (Fig. 17.15); in the layer about y ¼ 0 there is an inorganic polymer of composition HgCl3 and a stack of TTF cation radicals with an interplanar spacing ˚ , while in that about y ¼ 1/2 there are isolated (Hg2Cl6)2 ions and (TTFþ)2 of 3.6 A ˚ . The resonance Raman spectrum shows a -dimers, with an interplanar spacing of 3.43 A splitting of the TTF 3 mode, indicating the presence of two types of TTF cation; the difference between their charges was estimated to be

Crystalline Molecular Complexes and Compounds: Structure and Principles 2 Volume Set (International Union of Crystallography Monographs on Crystallography)

Crystal Structure Analysis: Principles and Practice (International Union of Crystallography Texts on Crystallography)

Molecular Aggregation: Structure Analysis and Molecular Simulation of Crystals and Liquids (International Union of Crystallography Monographs on Crystallography)

Crystallography of Modular Materials (International Union of Crystallography Monographs on Crystallography)

Crystallography and Surface Structure

Crystallography of the Polymethylene Chain: An Inquiry into the Structure of Waxes (International Union of Crystallography Monographs on Crystallography)

The Weak Hydrogen Bond: In Structural Chemistry and Biology (International Union of Crystallography Monographs on Crystallography)

Aperiodic Crystals: From Modulated Phases to Quasicrystals (International Union of Crystallography Monographs on Crystallography)

Crystal Structure Refinement: A Crystallographer's Guide to SHELXL (International Union of Crystallography Texts on Crystallography)

Crystallography

The Basics of Crystallography and Diffraction: Third Edition (International Lunion of Crystallography Texts on Crystallography)

Crystallography

Fundamentals of Crystallography (International Union of Crystallography Book Series, No. 2)

Fundamentals of Crystallography (International Union of Crystallography Book Series, No. 2)

Crystallography

Crystallography

Crystallography

International Tables for Crystallography, Vol.F: Crystallography of Biological Macromolecules

Advanced Structural Inorganic Chemistry (International Union of Crystallography Texts on Crystallography)

Data Mining in Crystallography (Structure and Bonding)

Macromolecular Crystallography Protocols: Structure Determination

Macromolecular Crystallography

Crystallography and Crystal Defects

Principles of Protein X-Ray Crystallography

Essentials of crystallography

Basic elements of crystallography

Macromolecular Crystallography

Structure of Materials: An Introduction to Crystallography, Diffraction and Symmetry

Electron Crystallography

Crystalline Molecular Complexes and Compounds: Structure and Principles 2 Volume Set (International Union of Crystallography Monographs on Crystallography)

Crystal Structure Analysis: Principles and Practice (International Union of Crystallography Texts on Crystallography)

Molecular Aggregation: Structure Analysis and Molecular Simulation of Crystals and Liquids (International Union of Crystallography Monographs on Crystallography)

Crystallography of Modular Materials (International Union of Crystallography Monographs on Crystallography)

Crystallography and Surface Structure

Crystallography of the Polymethylene Chain: An Inquiry into the Structure of Waxes (International Union of Crystallography Monographs on Crystallography)

The Weak Hydrogen Bond: In Structural Chemistry and Biology (International Union of Crystallography Monographs on Crystallography)

Aperiodic Crystals: From Modulated Phases to Quasicrystals (International Union of Crystallography Monographs on Crystallography)

Crystal Structure Refinement: A Crystallographer's Guide to SHELXL (International Union of Crystallography Texts on Crystallography)

Crystallography

The Basics of Crystallography and Diffraction: Third Edition (International Lunion of Crystallography Texts on Crystallography)

Crystallography

Fundamentals of Crystallography (International Union of Crystallography Book Series, No. 2)

Fundamentals of Crystallography (International Union of Crystallography Book Series, No. 2)

Crystallography

Crystallography

Crystallography

International Tables for Crystallography, Vol.F: Crystallography of Biological Macromolecules

Advanced Structural Inorganic Chemistry (International Union of Crystallography Texts on Crystallography)

Data Mining in Crystallography (Structure and Bonding)

Macromolecular Crystallography Protocols: Structure Determination

Macromolecular Crystallography

Crystallography and Crystal Defects

Principles of Protein X-Ray Crystallography

Essentials of crystallography

Basic elements of crystallography

Macromolecular Crystallography

Structure of Materials: An Introduction to Crystallography, Diffraction and Symmetry

Electron Crystallography

Recommend Documents