Modern Coordination Chemistry 2002 Leigh Winterton

Modern Coordination Chemistry
The Legacy of Joseph Chatt

Modern Coordination Chemistry
The Legacy of Joseph Chatt
Edited by
G. J. Leigh
University of Sussex, Brighton, U K
N. Winterton
University of Liverpool, Liverpool, U K
ROYAL SOCIETY OF CHEMISTRY

ISBN 0-85404-469-8
A catalogue record for this book is available from the British Library
0The Royal Society of Chemistry 2002
All Rights Reserved.
Apart from any fair dealing for the purposes of research or private study, or criticism or
review as permitted under the terms of the UK Copyright, Designs and Patents Act, 1988, this
publication may not be reproduced, stored or transmitted, in any form or by any means,
without the prior permission in writing of The Royal Society of Chemistry, or in the case of
reprographic reproduction only in accordance with the terms of the licences issued by the
Copyright Licensing Agency in the UK, or in accordance with the terms of the licences issued
by the apropriate Reproduction Rights Organization outside the UK. Enquiries concerning
reproduction outside the terms stated here should be sent to The Royal Society of Chemistry at
the address printed on this page.
Published by The Royal Society of Chemistry,

Thomas Graham House, Science Park, Milton Road, Cambridge CB4 OWF, UK
Registered Charity Number 207890
For further information see our web site at www.rsc.org
Typeset in Great Britain by Vision Typesetting, Manchester, UK

Printed and bound by MPG Books Ltd, Bodmin, Cornwall, UK
Foreword
‘The Child is Father of the Man’

William Wordsworth (1770-1850)
This book is dedicated to the work of Joseph Chatt. In this Foreword I should
like to place Chatt’s contribution in the perspective of chemistry when he entered
the field, recognising that it was a very different field from that of the present day.
Using the quotation above of another who was brought up in Cumberland,
Wordsworth, I believe that his initial training in chemistry was an influence,
perhaps the most important, on his attitude and approach to the subject. Chatt
always talked with affection and gratitude of his mentor F. G. Mann. Mann was
shy, but a remarkable chemist who viewed chemistry as a single subject and not
as three branches: organic, inorganic and physical. He was a very careful and
precise chemist who enjoyed producing crystals of high purity. It is reputed that
the micro-analytical group in Cambridge often used his compounds to test their
apparatus. Another characteristic that I found with F. G. was that he was never
prepared to accept the immediate answer to any problem in his chemistry and
viewed all reasonable alternatives. These are some of the characteristics that
Joseph Chatt also exhibited.
The School of Chemistry in Cambridge was led by Sir William Pope, an
organic chemist who was interested in the optical activity of elements other than
carbon. F. G. Mann was his research assistant, and although Chatt was officially
registered as a student with Pope his training was with Mann. It is of interest that
by this time much of the work being carried out in the Cambridge laboratories
was inorganic in nature but it was still classified as organic.
Joe Chatt was one of the leading inorganic chemists of his day. However, as
was common at that time, his initial training was primarily in organic chemistry.
His initial research project involved the preparation and bridge-splitting reac-
tions of some dipalladium halogen-bridged phosphine complexes. The prepara-
tion of metal complexes of this type arose because Mann was interested in the use
of phosphines in the preparation of optically-active phosphorus compounds and
the platinum metals provided useful means of purifying and crystallising the
phosphine compounds. After his initial work on these compounds, Chatt
changed his research direction to a study of the stereochemistry of related arsenic
compounds. This work led to the preparation of phenylene- 1,2-bis(dimethyl-
arsine), a compound utilised in coordination chemistry to great effect by Ron
Nyholm. It was also during his period in Cambridge that Chatt became
vi Foreword
acquainted with the olefin-platinum compounds that were to be of major
interest when he started his independent career in inorganic chemistry.
On graduating, his interest in pursuing academic work was interrupted by the
war and he was recruited into the scientific Civil Service. He was given initially a
problem involving the preparation of organic nitro-compounds similar to TNT.
This was an attempt to produce better explosives, an idea that originated with
Robert Robinson, the leading British organic chemist of the day. However, the
compounds proved to be of no major improvement on the explosives then
available. Chatt’s career then followed a chequered path within the Government
service, from Swansea to Woolwich Arsenal, a major scientific laboratory in
London, to a position as Deputy Chief Chemist at Peter Spence & Sons Ltd at
Widnes in the North of England. This company was concerned with the produc-
tion of aluminium chloride and oxide. However, he was not happy in these
appointments and sought to return to an academic post in inorganic chemistry.
To appreciate Chatt’s position at this time it is perhaps important to place the
study of inorganic chemistry in perspective. Before the Second World War, it was
very much the Cinderella of chemistry, often of minor concern in University
chemistry courses. The subject was either omitted from the courses or taught by
staff who had little interest in the subject, particularly from the point of view of
research. Thus M. G. Evans, a leading physical chemist of his day and holder of
the Chair of Physical Chemistry in Manchester, excluded any major teaching of
the subject from the chemistry degree course.
When it came to the research in inorganic chemistry, which was not very
extensive in the UK, the primary interest was in the study of the non-transition
elements, with a particular emphasis on the similarity of the chemistry to that of
related organic compounds. In many instances comparison with the carbon
compounds was the prime aim. Thus there was a considerable interest in the
chemistry of boron and silicon compounds and the relationship to their carbon
analogues. As the compounds of these elements are particularly sensitive to
water, and often to dioxygen, the main work involved the use of vacuum line
techniques. In addition, the preparation of volatile compounds allowed the
application of a range of physical methods that at that time were developed
mainly to deal with volatile compounds. Methods for the determination of the
structures of compounds in the solid and liquid states were not well developed.
The X-ray structure of even relatively simple compounds could take years to
complete. Paradoxically, Chatt thus obtained a very good training in an area of
transition-metal chemistry, under the cloak of research in organic chemistry,
when such chemistry was not being studied in any inorganic department. As his
research work showed, this proved to be an extremely good training for the study
of what was to become the coordination chemistry of the later transition ele-
ments.
The activities in the war brought about significant changes of emphasis in
the study and role of chemistry, particularly of inorganic chemistry. The
atomic-bomb project led to the preparation, isolation and study of the chemis-
try of the transuranic elements. This focused on a completely new area of
inorganic chemistry. New techniques for the isolation and handling of chemical
Foreword vii
compounds were developed and, in collaboration with physicists, new ap-
proaches to many problems were discovered. In particular, techniques such as
ESR spectroscopy were developed which opened up the whole area of reson-
ance methods.
This work led to a complete reassessment of inorganic chemistry, and, in
particular, of the structural aspects of the subject, as chemistry, both inorganic
and organic, was in those days dominated by the isolation and determination of
the structures of compounds. It clearly took a while for these changes to be
appreciated in the university system but it was with this changing approach to
inorganic chemistry that Chatt took up his first university appointment as an ICI
Fellow at Imperial College in 1946. This Fellowship was part of a scheme
designed by ICI to make appointments in University departments for the best
researchers until permanent staff appointments were available. However, Chatt
found the environment at Imperial unsatisfactory and transferred to a new
enterprise for fundamental research that was also being set up by TCI in a
country house, The Frythe, just north of London. This laboratory, the Butter-
wick Laboratory, was subsequently called the Akers Laboratory. This was a
great opportunity that was fully appreciated by Chatt. Initially there was an
attempt to include him in the organic section, reflecting the view of the establish-
ment as to the status of inorganic chemistry but Chatt put up a strong fight to
have a separate inorganic section. This gave him the opportunity to research, for
the first time, his own ideas and he chose to look at the chemistry of metal olefin
and acetylene compounds. Chatt’s work in these laboratories was at a time when
there was the beginning of a surge of interest in inorganic chemistry in the UK
and Europe in general. At this time Nyholm was also starting his work in
coordination chemistry at University College in London. These two workers
initiated what has been referred to as the Renaissance of Inorganic Chemistry in
the UK. There was an air of expansion within the universities, and in chemistry
this particularly applied to inorganic chemistry. It is perhaps important to
recognise that in 1946 for the UK only 4% of the potential student cohort went
to university and there were less than 30 universities or their equivalents in the
country. If we compare this with the present situation of nearly 40% of the
cohort and more than 100 universities, the changes have been enormous, and
within the field of chemistry this is particularly true for inorganic chemistry.
The work described in this volume illustrates the wide range of contributions
that Chatt made to inorganic chemistry, with over-spill into organic chemistry
and biochemistry. His work attracted world attention and a large number of the
most eminent inorganic chemists visited and worked in his laboratories, both at
The Frythe and in Sussex. He attracted large numbers of postdoctoral workers,
many of whom went on to make their own mark within chemistry and have
contributed to this volume. On a purely personal level, in 1952 I debated long
between an offer to work at The Frythe and a university appointment, finally
deciding to take the university post.
The work carried out at The Frythe pioneered the basic chemistry that
pervades much of organometallic chemistry today, with its overtures into cataly-
sis. The work carried out at the University of Sussex was a major study in the
...
Vlll Foreword
nitrogen fixation problem and led to fascinating developments in the chemistry
of dinitrogen transition metal compounds.
Joseph Chatt was at the forefront in the development of inorganic chemistry,
from what may be considered ‘the dark ages’ to the present day. His influence,
both through his studies and the researchers with whom he worked, will be with
us for decades. In addition to his many contributions to chemistry, Joseph will be
remembered by many as a nice man and a good friend.
Lord Lewis of Newnham

Introduction
It is very appropriate that this volume, which deals with the scientific legacy of
Joseph Chatt, should appear at this time. A decision was taken very early in the
planning of the 34th International Coordination Chemistry Conference
(ICCC34) to have a theme dealing with ‘Joe Chatt Chemistry’. ICCC34 was held
in the United Kingdom to celebrate the Golden Jubilee of the first meeting
organised by Joseph Chatt at The Frythe, Welwyn, near London, which was
then a corporate research laboratory of Imperial Chemical Industries. Since this
first meeting ICCCs have grown steadily in size and importance with recent
meetings attracting delegates from more than 50 countries.
As ICCC34 coincided with the New Millennium, it was appropriate to use the
Joseph Chatt theme to provide a historical perspective of the development of
coordination chemistry because, as this volume clearly demonstrates, Chatt was
involved with an extraordinarily broad range of research activities. This his-
torical perspective was backed by an exhibition organised by Paul OBrien
giving details of earlier meetings and many of the personalities who have been
involved over the years.* More importantly, ‘Joe Chatt Chemistry’ was also used
to provide a prediction of the likely future developments in coordination chemis-
try. Here the organisers faced a difficulty. Joseph Chatt’s contributions to the
subject have impacted on so many areas that it became difficult to draw a line
defining the boundaries of topics that were to be considered under the theme of
‘Joe Chatt Chemisty’. In many cases papers which could have been presented
under this heading were included in the other sessions: Structure and Dynamics,
2 1st Century Materials, Biotechnology and Medicine, Technological Advances,
and Chemistry of Life. The splendid range of papers included in this book
illustrates how broadly Joseph Chatt contributed to the development of the
subject. The editors should be congratulated on bringing together such an
interesting and representative collection of papers defining his legacy.
Peter Tasker
* For more information, contact Professor P. OBrien, the Manchester Materials Science Centre and
the Chemistry Department, University of Manchester, Oxford Road, Manchester M 13 9PL, UK.
Preface
The historic threads of scientificenquiry that are woven into the fabric that is our
understanding of the natural world (and the benefits that stem from its applica-
tion in technology) should not simply be the province of scientific historians (nor
even, heaven forbid, sociologists of science or cultural theorists); they should
inform practitioners and the wider public alike of the debt today’s scientists owe
to their scientific forbears as well as demonstrating that the body of scientific
knowledge is continually growing and some portion of it changing. Indeed, the
processes, requirements and timescales of science, and the distinction to be made
between them and those of technological development, need to be much better
understood, particularly by those who formulate and execute public policy and
those who direct and manage industrial, commercial and financial enterprises
and not least by scientists themselves.
At the beginning of the 20th century, inorganic chemistry was overshadowed
by developments in organic and physical chemistry, developments in both of
which were to lay the foundations for the re-invigoration of inorganic chemistry
and the sub-disciplines of coordination and organometallic chemistry that char-
acterised the latter half of the century.
It is our purpose to provide a perspective of this formative period (and its
manifestation in certain areas of contemporary inorganic chemistry) through the
contributions of one of the foremost practitioners, Joseph Chatt.
Isaac Newton suggested that his own scientific vision was so great because he
had ‘stood on the shoulders of giants’. In a humbler sense, all scientists see further
and deeper than they otherwise might, because of the work done by those who
went before. However, with an ever-increasing volume of scientific publication, it
is difficult for today’s scientist to keep up with new material in his or her own
field, let alone to explore and appreciate the wider significance of the earlier
literature. While this is understandable, not only might they fail properly to
acknowledge work with a bearing on their own, but they also lose sight of the
methods (often very limited) available in the past that provide a proper testament
to the magnitude of earlier contributions. We should know what it was that
ensured that this earlier work has stood the test of time (particularly the test of
modern scientific methodology). In addition, the over-dependence on com-
puterised literature searching methods tends to reduce awareness of material
published before the mid-1960s. This is significant, as the post-war period up to
that point was the time when Joseph Chatt and his collaborators (as well as other
major contemporary figures) were especially productive.
xii Preface
Today’s science is much more of a team (one might even say a corporate)
activity, with the idea of scientific leadership of the sort provided by Chatt (and
amply exemplified by the various contributions in this volume from his former
students and co-workers) being seen increasingly as old-fashioned. However, for
today’s research students and their supervisors at least, it should be informative if
not instructive to be offered an appreciation of what it was like to work for such
an individual, particularly the testing and challenging but essentially supportive
environment in which Chatt’s students researched.
There are those who look at the close involvement that exists today between
science and business and are concerned that science is damaged by such interac-
tions. It is significant then to draw attention to Chatt’s employment and support
by Imperial Chemical Industries Ltd during the period 1949 to 1962. (Chatt’s
internal ICI reports for the period have been archived at the John Innes Centre,
Norwich, UK.) It says much about Chatt that his independence and strength of
character saw to it that he produced fundamental work of the highest standard.
It says much about ICI’s very senior staff that they believed in the importance of
participation in the scientific enterprise at its most fundamental. On the other
hand one is also forced to conclude that there were others in ICI who tended to
dismiss the workers at The Frythe as pampered academics and who did not
regard Chatt’s work as being of any value to them. Whether this was so, or
simply appeared so to those unable to conceive of or appreciate its significance
or potential, cannot now be gauged. One may simply speculate whether a greater
understanding on both sides of the gulf that inevitably separates the output of
purely scientific work from that needed to secure a commercial opportunity from
a technological development based upon it (and the presence of some mechanism
within the company purposefully designed to bridge this gap) might have led ICI
to gain an advantage from the developments in coordination and organometallic
chemistry potentially available to them. Suffice it to say that ICI was not the only
large chemical corporation that found difficulty in reconciling the shorter-term
demands of commercial operation and financial performance with the longer-
term needs of research and development. Indeed, the problem remains a general
and contemporary one.
Today’s generation of academic scientists, particularly in the UK, have oper-
ated in an environment in which they have felt obliged to secure funding for their
work primarily by stressing its ‘relevance’ or its potential for application, even
when this has been at the expense of purely curiosity-driven research. Some
argue that academics themselves are partly to blame because of their unwilling-
ness to recognise and support creativity and adventurousness through the peer
review process. Others argue that there is no incompatability between funda-
mental work done for its own sake and that done with some end in view. Indeed,
this principle was well exemplified by the overall purpose of the Nitrogen
Fixation Unit set up by the Agricultural Research Council at the University of
Sussex under Chatt’s direction. That it was so successful from a scientific point of
view arose from a further principle that characterised Chatt’s stewardship of the
Nitrogen Fixation Unit, namely the recognition of the contributions to be made
by different disciplines and by the purposeful interactions between them.
Preface xiii
The gap between scientists and those seeking to develop the results of their
science that was characterised in Chatt’s time by ignorance and indifference may
have been narrowed, though it may well now be more characterised by suspicion
and antagonism arising from a failure of scientists and their sponsors to recog-
nise the motivations, methods, language and timescales of each other. The wiser
manager and investor do indeed recognise the long-term importance of ‘pure’
research (as well as understanding that the timescale is often measured in decades
and not reflected by the demands of quarterly reporting of business perform-
ance!): there are just not enough of them to provide, business-wide, scientifically
literate and technologically creative management. Today’s successful academic
should be at ease with a portfolio of fundamental research driven both by
curiosity (particularly at the interfaces with other disciplines) and by technologi-
cal or societal need and must be willing to appreciate the intellectual challenges
associated with technological development (as well as its inevitable shorter-term
focus). There are not enough of such academics either.
It is important to look back at the contributions of major figures of the past to
assess, with a longer historical perspective, whether their reputation stands the
test of time. We know that such assessments have been made about major figures
such as Newton (and Hooke’s contributions) and Darwin (and Wallace’s contri-
butions). In one particular sense, this volume provides a starting point for that
assessment as far as Chatt is concerned.
As Peter Tasker’s comments in his Introduction indicate, this volume grew out
of the 34th International Coordination Chemistry Conference held at the Uni-
versity of Edinburgh, in July 2000. While most of the contributions presented to
the ‘Joe Chatt Chemistry’ sessions are reproduced in expanded form here, we
have also sought contributions from Chatt’s contemporaries and students,
whose reminiscences give a true measure of the man. There are, in addition,
contributions which concentrate more on the chemistry with which Chatt was
associated and, in various degrees, link Chatt’s work, particularly on phosphine-,
hydride-, olefin- and dinitrogen-metal complex chemistry and chemistry, bio-
chemistry and biology of nitrogen fixation, with the very latest developments in
these topics. These include chapters from recipients of the Chatt Lectureship,
conferred by the Royal Society of Chemistry.
While this book is not solely biography, history, scientific text or conference
proceedings, we hope that the mix of each will be of interest to many. We thank
all the authors who have provided contributions and believe that they truly
reflect the legacy of a great chemist to modern coordination chemistry.
G. J. Leigh and N. Winterton

Contents
Abbreviations xxvi
Section A: Reminiscences of Joseph Chatt Drawn from Conversations
and from the Recollections of Co-workers 1
A Memoir of Joseph Chatt 3
G. J. Leigh
A Memorable Start to a Career 8
R. G. Wilkins
Joseph Chatt and The Frythe: A Memoir of the Early 1950s 11
G. A. Gamlen
Recollections of Life with Joseph Chatt at The Frythe,
1958-62 18
D. M . Adams
Joseph Chatt F R S Some Memories of his Work at The Akers
Research Laboratories of Imperial Chemical Industries Ltd,
The Frythe, during the 1950s 25
L. A. Duncanson
Section B Recent Developments in the Synthesis, Bonding Modes and

Reactivity of Hydrido and Dihydrogen Complexes 29
Hydrides, Hydrogen Bonding and Dihydrogen Activation 31

R. H. Crabtree
1 Introduction 31
2 Intramolecular Interact ions 33
3 The Nature of A-H-H-E Hydrogen Bonding 35
4 Intermolecular Interactions 37
5 Reactivity 38
6 Intramolecular Effects of Pendant Groups 39
7 Heterolytic H-H Activation 41
8 Consequences of He-H Bond Formation 41
9 Conclusion 42
10 Acknowledgements 42
xv1 Contents
11 References and Notes 42
Hydrides and Dihydrogen Ruthenium Complexes:

a Continuation of Joe Chatt’s Chemistry 45
S. Sabo-Etienne and B. Chaudret
1 Introduction 45
2 Historical Aspects 45
3 Ruthenium Hydride and Dihydrogen Complexes 47
3.1 [RuH,(dppm),] Dihydrogen Bonds and Proton
Transfer 48
3.2 [RuH,(H,),(PCY3),1 48
3.3 Hydrogen Transfer and Hydrocarbon Activation 49
3.4 16-Electron Hydrido(dihydr0gen) Complexes:
Reactions of [RuH2(H2)(PCy ,),I with Halocarbons
and Related Reactions 49
3.5 18-Electron Hydrido(dihydr0gen) Complexes,
Proton Transfer and C-H Activation 50
3.6 Silane Complexes and Substitution Reactions of
CRuH,(H,)(PCy 3)21 53
4 Conclusion 56
6 References 57
Hydrido Complexes of Group 6 Transition Metals -

Formation of the Pentadentate Ligand with a P-P-Si-P-P
Framework 59
T. I t o
1 Introduction 59
2 Reactions of [MH,(dppe),] with Primary and Secondary
Silanes 60
3 Some Reactions Involving the Trihydro Complex
[MoH,(Ph,PCH,CH2P(Ph)C6H4-o],(Ar)Si-P,P7P,P,Si)] 63
4 References 65
Section C: The Chemistry of Phosphines 67
Some New Insights into the Steric Effects of Tertiary

Phosphine Ligands via Data Mining 69
D. M . P. Mingos
1 Introduction 69
2 Historical Background 70
3 Discussion 72
5 References 78
Contents xvii
Synthesisof New and Unusual Metal Complexes from the
Reaction of Dihalogen Adducts of Tertiary Phosphines with
Unactivated Metal Powders 79
S. M . Godfrey and C. A. McAulifle
1 Introduction 79
2 Background 80
3 Work at The University of Manchester Institute of
Science and Technology (UMIST) 80
3.1 Nature of PR,X, (X = C1, Br or I) 81
3.2 The Activation of Crude Metal Powders by
Reagents PR,X, 82
4 References 87
NMR Studies of Metal Complexes and Clusters with

Carbonyls and Phosphines 89
B. T. Heaton
1 Introduction 89
2 Metal Carbonyls 90
3 Mechanistic in situ Homogeneous Catalytic Studies on
Metal-Phosphine Complexes 95
5 References 99
Section D: Transition Metal Complexes of Olefins, Acetylenes, Arenes

and Related Isolobal Ligands 101
Some Notes on the Early Development of Models of Bonding

in Olefin-Metal Complexes 103
N. Winterton
1 Models of the Olefin-Metal Bond 103
2 Addendum 108
3 References 108
The Dewar-Chatt-Duncanson Bonding Model of Transition

Metal-Olefin Complexes Examined by Modern Quantum
Chemical Methods 111
G. Frenking
1 Introduction 111
2 Platinum Complexes of Strained Olefins 113
3 Olefin Complexes versus Metallacyclopropane 115
4 Electrostatically-bound Olefin Complexes 119
5 Summary 120
7 References and Notes 121
xviii Contents
CycloadditionReactions with Metalla-l,3-Dipoles 123

H.- W. Friihauf
1 Introduction 123
2 1,3-Dipolar Cycloaddition of Activated Alkyne to the
Fe-N=C Fragment 124
2.1 The Reaction of [Fe(CO),(dab)] with Dimethyl
Acetylenedicarboxylate 124
2.2 The Isolobal Relation of the Fe--N=C Fragment
with an Azomethine Ylide 125
3 Adding Variety to the 1,3-Dipolar System 126
3.1 Changing the Metal 126
3.1.1 Reactions of [Ru(CO),(dab)] 126
3.1.2 [Mn(CO),(dab)].: A test case for the isolobal
relation 127
3.2 Changing the Additional Ligands: Isocyanides for
Carbon Monoxide 129
3.2.1 Reactivity of [Fe(CO),(RNC)(dab)] 129
3.2.2 Reactions of [Fe(RNC),(dab)] with
dmad, olefins and heteroallenes 129
3.3 Changing the Heteroatoms 134
3.3.1 Oxygen instead of nitrogen: a-imino ketones
and a-imino esters 134
3.3.2 Sulfur instead of nitrogen: dithiooxamide 136
5 References 138
A Journey in Metal-Ligand Multiple Bond Chemistry 140

V. C. Gibson
1 Introduction 140
2 The Elusive [M(O)Cl,Cp] Complexes of Niobium and
Tantalum 141
3 Half-sandwich Imido Compounds of Niobium and
Tantalum 142
4 Exploiting the Isolobal Relationship between
Cyclopentadienyl and Imido Ligands 144
5 Alkyl and Alkylidine Complexes of Molybdenum:
Routes to Olefin Metathesis Catalysts 147
6 Olefin Polymerisation Catalysts 148
7 Distortional Isomerism 149
9 References 152
Contents xix
Synthesis, Characterisationand Catalytic Activity of
Heterobimetal Complexes 154
M . Abou Rida and A . K . Smith
1 Introduction 154
2 Synthesis of Ruthenium-Rhodium Heterobimetal
Complexes 154
3 Hydroformylation of Oct-1-ene 157
4 Conclusion 161
5 References 161
Tethered Arene Complexes of Ruthenium 163

M . A . Bennett and J . R. Harper
1 Introduction 163
2 Results and Discussion 164
4 References 168
Section E: Chemistry Related to Dinitrogen Complexes 169
Chemistry at the Unit of Nitrogen Fixation 171

R. L. Richards
1 Introduction: The Early Years and the Search for
Dinitrogen Complexes 171
2 The Search for Reactions of Coordinated Dinitrogen 173
3 Production of Ammonia and Search for a Catalytic
System 175
4 The Formation of Nitrogen-Carbon Bonds 177
5 Formation of Amines 180
6 Reduction of Alternative Substrates 180
7 Further Developments 182
8 Conclusions 183
9 References 184
Dinitrogen Activation by Early Transition Metal-

Amidophosphine Complexes 187
M . D. Fryzuk
1 Preamble 187
2 Background 187
3 Our First Dinitrogen Complexes 189
4 The Macrocycle 191
5 Trying to Cleave the N-N Bond 194
6 Conclusions 195
8 References 196
xx Contents
Metal-D initrogen Chemistry after Chatt 198

C. Floriani
1 Introduction 198
2 Coordinative Dinitrogen Binding 198
3 Reductive Dinitrogen Binding 199
4 N-N Triple Bond Cleavage to Nitride 200
6 References 206
Novel Chemical Transformationsat Diruthenium Centres

Bridged by Thiolate Ligands 208
M . Hidai, Y. Ishii and S. Kuwata
1 Introduction 208
2 Reactions of Hydrazines at Thiolate-bridged
Dirut henium Centres 208
3 Transformations of Small Organic Molecules at
Thiolate-bridged Diruthenium Centres 209
3,l Stoichiometric Transformations of Alkynes 210
3.2 Catalytic Head-to-head 2 Dimerisation of Terminal
Alk ynes 212
3.3 Transformation of Propargyl Alcohols 213
3.4 Other Catalytic Reactions 214
4 Concluding Remarks 215
5 References 216
The Chemistry and Applicationsof Complexes with Sulfur

Ligands 217
J . R. Dilworth, P. Arnold, D. Morales, Y.-L. Wong
and Y. Zheng
1 Introduction 217
2 Sterically-hindered Thiolates 218
3 Phosphinothiolate and Related Ligands 219
4 Polydentate N,S Donor Ligands 223
5 Thiosemicarbazone Ligands 226
5.1 Complexes of Ru, Rh and Ir 226
5.2Copper Bisthiosemicarbazone Complexes as
Hypoxic Selective Agents 228
6 Conclusions 229
7 References 229
Contents xxi
Section F: The Biological Work of the ARC Unit of Nitrogen Fixation
at the University of Sussex, and Later Developments 231
Biological Nitrogen Fixation 233

J . Postgate
1 The Unit of Nitrogen Fixation 233
1.1 Nitrogen Fixation 233
1.2 The Unit’s Origins 235
1.3 Earliest days of the U N F 236
2 Biochemical Research 238
2.1 Structural Analogues of Dinitrogen 239
2.2 Enzyme Studies 239
2.3 Studies of Mechanism 240
3 Physiological Research 242
3.1 Dioxygen Exclusion 243
3.2 The Biological Reductant(s) 245
3.3 The Demand for ATP 245
3.4 Dihydrogen Recycling 245
4 Genetical Research 246
4.1 The nifGenes 246
4.2 Regulation of nif 249
5 Interdisciplinary Interactions 249
6 Envoi 250
7 References 250
Vanadium, Molybdenum and Iron Complexes Based on a

Trithiolate Ligand 252
J . R. Sanders
1 Introduction - Phosphine and Thiolate Ligands 252
2 Recent Studies on Nitrogenase 253
3 Bulky Ligands and Tripodal Ligands 254
3.1 Ligands with Four-sulfur Donor Sets 254
3.2 Ligands with Three-nitrogen Donor Sets 254
3.3 Ligands with Three-sulfur Donor Sets 255
4 A Simple Tripodal Ligand 255
4.1 Introduction to the NS, Site 255
4.2 Vanadium and Molybdenum Complexes of NS, 255
4.3 Iron Chemistry of NS,-Carbonyl, Nitrosyl and
Isocyanide Complexes 257
4.4 Models for Hydrogenases 260
4.5 Dinuclear and Tetranuclear Iron-dinitrosyl
Compounds 260
5 Acknowledgements 26 1
6 References 26 1
xxii Contents
The Nature of Molybdenum and Tungsten Centres in
0x0-transfer Enzymes 263
C. D. Garner
1 Introduction 263
2 Nature of the Catalytic Centres of Mo and W 0x0-
transfer Enzymes 265
2.1 Coordination of the Metal Atom 265
2.2 Function of the Metal Centre 266
3 W-substituted DMSO Reductase 267
3.1 Production and Characterisation of W-DMSOR 267
3.2 Assays of Mo- and W-DMSO Reductase Activity 268
4 Chemistry Related to that of the Catalytic Centres of
Mo and W 0x0-transfer Enzymes 269
4.1 Towards the Synthesis of Molybdopterin 269
4.2 Dithiolene Complexes 27 1
5 Conclusions 272
7 References 275
Iron-Imide Clusters and Nitrogenase: Abiological Chemistry

of Biological Relevance? 278
S. C. Lee
1 Nitrogenase: Problem and Approach 278
2 Iron@)Reduction of the N-N Bond 280
3 Iron-Imide Cluster Chemistry 281
4 Analysis and Speculation 284
6 References 286
Determinants of the Reduction Potential in Rubredoxins,

the Simplest Iron-Sulfur Electron-transferProteins 288
Z. X i a o and A . G. Wedd
1 Introduction 288
2 Determinants of Reduction Potential 290
3 The Ionisation Energy Term, I E 29 1
4 The Solvation Energy Term, U 294
4.1 Proximity of Charged Residues 295
4.2 Proximity and Orientation of Dipoles (Including
NH S Hydrogen Bonds) 296
4.3 Solvent Access 297
5 Acknowledgement 300
6 References 300
Contents xxiii
Section G: Patterns and Generalisations in Stability and Reactivity 303
Hard and Soft Acids and Bases and Joe Chatt 305
R. G. Pearson
1 Introduction 305
2 Hard and Soft Acids and Bases (HSAB) 307
3 Chemical Hardness 309
4 References 3 12
Mechanisms of Platinum Reactions 313

F . Basolo
1 Chemistry of Platinum@) Complexes 3 14
2 Kinetics and Mechanisms of Ligand Substitution
Reactions of Platinum(x1) Complexes 3 14
3 Collaborative Research with Chatt on the Kinetic
Trans Efect of the Nickel Triad Metal Complexes 321
4 The Kinetics and Mechanisms of Ligand Displacement
Reactions of Platinum(1v) Complexes 322
5 The Application of Reaction Mechanisms for the
Synthesis of Metal Complexes 324
6 Joe Chatt the Person 326
7 References 326
Tuning Rhodium(1) Metal Centre Accessibility in Iodomethane

Oxidative Addition to Vaska-type Complexes by Interchanging
Tertiary Phosphine for Arsine and Stibine 328
S. Otto, S. N . Mzamane and A . Roodt
1 Introduction 328
2 Experimental 329
3 Reaction Scheme 329
4 Structural Aspects 330
4.1 Rhodium(1) Complexes 330
4.2 Rhodium(I1x) Alkyl Complexes 333
5 Rate Laws for Reaction 334
6 Reactivity and Reaction Parameters for Iodomethane
Oxidative Addition 337
7 Solvent Effects 338
8 Concluding Remarks 338
10 References 339
xxiv Contents
Section H: Other Papers Presented at the 34th International Conference
on Coordination Chemistry, Edinburgh, Scotland, July 2000 341
Formaldehyde Elimination from Methoxylated Transition

Metal Carbonyl Clusters 343
P. J . Dyson, B. F. G. Johnson, J . S. Mclndoe, D. Sambrook
and P. R. R. Langridge-Smith
1 Alkoxylation of Carbonyl Ligands 343
2 Chemical Derivatisation 344
3 Energy-dependent Electrospray Mass Spectrometry 344
4 Energy-dependent Electrospray Tandem Mass
Spectrometry 347
5 Formaldehyde Elimination 349
7 References 353
Exploring New Structures Based on Chatt’s {Pt,S,) Core for

Nucleation of Intermetallic Growth 355
2. Li, S.-W. Audi Fong, J . S. L. Yeo, W. Henderson, K . F. Mok
and T. S. A. Hor
1 Historical Perspective 355
2 Current Perspectives 357
2.1 Stereochemical Changes 357
2.2 Different Degrees of Site-anchoring 358
2.3 Bifacial Addition and Dissociation 358
2.4 Formation of Bimetal Complexes from Early and
Late Transition Metals 359
2.5 Introduction of an Unsaturated Functionality 359
2.6 Enhancing Electrochemical Activity 360
2.7 Electrospray Mass Spectrometry (ESMS) Analysis 360
4 References 364
A Rational Design of HeteropolynuclearSquarate

Complexes 366
F. Dumitru, D. Berger, N . Stanica, I . Ciocoiu and C. Guran
1 Introduction 366
2 Experimental 369
2.1 Materials 369
2.2 Compound Preparations 369
2.3 Physical Measurements 370
2.4 Results and Discussion 370
2.5 IR Spectra 370
2.6 Electronic Spectra and Magnetic Measurements 371
Contents xxv
3 Conclusions 3 72
4 References 373
Index of People and Places 375

Subject Index 379
A bbveviations
Hacac acet y lacet one; pent an-2,4-dione

Ala alanine or alanate
ArF perfluoroaryl
ASP aspartic acid or aspartate
ATP adenosine triphosphate
H,atsm H,NC( S)NHN=CMeCMe=NNHC( S)NH2
2,2’-bipyridyl
2,2’-bipyrimidine
7,8-benzoquinoline
CDA charge decomposition analysis

cod cycloocta- 1,Sdiene
cot cycloocta-l,3,5-triene
CP cy clopen t adien yl
CP* pent amet hylcyclopen t adien yl
Hcupf cupferron; N-nitrosophenylhydroxylamine
CY cyclohexyl
CYS cysteine or cysteinate
dab N,N’-disubstituted 1,4-diazabuta- 1,3-diene

DCPIP dichlorophenolindophenol
Hdct 2,6-dichlorothiophenol
depe 1,2-bis(diethyIphosphino)ethane
diars 1,2-bis(dimethy1arsino)benzene
digly diethyleneglycol dimethyl ether
dmad dimethyl acetylenedicarboxylate
Hdmavk dimethyl-P-aminovinylket one
dme 1,2-dimethoxyethane
dmpe 1,2-bk(dimethylphosphino)ethane
dmso or DMSO dimethyl sulfoxide
DMSOR dimethyl sulfoxide reductase
DMS dimethyl sulfide
Hdmt 2,6-dimethylthiophenol
dPPe 1,2-bis(dipheny1phosphino)et hane
dPPf CFe(C,H,PPh2)21
Abbreviations xxvii
dPPm bis(dipheny1phosphino)methane
dPPP 1,3-bis(diphenylphosphino)propane
dta dithiooxamide
en ethane- 1,2-diamine
EDESI-MS energy-dependent electrospray ionisation mass
spectrometry
ESI-MS electrospray ionisation mass spectrometry
HFc ferrocene; [Fe(q5-C H J2]
FeMoco iron-molybdenum cofactor of nitrogenase
Glu glutamic acid

H2gts H,NC( S)NHN=CHCH=NNHC(S)NH,
HiPIP high potential iron protein

His histidine
hmpa hexamet hylphosphoramide
HOMO highest occupied molecular orbital
INS inelastic neutron scattering

IR infrared
Leu leucine
LUMO lowest unoccupied molecular orbital
LYS lysine
Me,[16]aneS4 2,4,6,8,10,12,14,16-octamethyl-1,5,9,13-
tetrathiahexadecane
MeP methyl propanoate
Mes mesityl
mP methyl propynoate
MPT mol ybdopterin
nuclear magnetic resonance

,
N(C H CH ,NH SiMe,) ,
1,2-HSC6H4SCH,CH,NHCH2CH,SC,H,SH-1,2
N(CH,CH,SH),
[PhP(CH, SiMe,NH Ph),]
Hoxin 8-hydroxyquinoline
H2(2-pedt) (2-pyridy1)ethylenedi-1,2-thiol
o-phen 1,lO-phenanthroline
H(PN P) NH(SiMe,CH,PR,),
XXVlll Abbreviations
H,(P,N,) PhP(CH,SiMe,NHSiMe,CH,),PPh
HPS) 2-Ph,PC,H,SH
H(PSP) 2,6-(dipheny1phosphino)thiophenol
H,(PS,) PhP(C,H,SH-2),
H3(PS3) P(C ,H4S H -2)3
H,(NH,- pt edt) 2-amino-3-methyl-4-oxopteridinyl)ethylenedi- 1,2-thi01
H (NC(H)N M e ,- (2-N,N-dimethylaminomethyleneamino)-3-methyl-4-
ptedt) oxopteridiny1)ethylenedi-1,2-thiol
H,ptsm H MeNC(S)NHN=CMeCH=NNHC( S)NHMe
PY pyridine
H,qedt 1,2-
(2-(N,N-dimethylimino)-4-(qinoxalin-2-yl)ethylenedi-
thiol
Hquin 2-quinaldic acid; 2-quinolinecarboxylic acid
RCM ring-closing metathesis

Rd rubredoxin
ROMP ring-opening metathesis polymerisation
R-pyca 2-pyridine N-aryl carbaldimine
H2(S4) 1,2-HSC,H,SCH2CH,SC,H,SH- 1,2

H,salen N,N’-bis(salicy1idene)ethane1,2-diamine
H2sdt (2-C6 H JC( SH)=C H(SH )
Ser serine or serinate
SHE standard hydrogen electrode
terPY 2,2’:6’,2’’-t erpyridine

TfOH trifluoromethanesulfonic acid or triflic acid
t hf tet rahydrofuran
thPP 1,3,3a76a-tetrahydropyrrolo[ 3,2-b]pyrrole
tht tetrah ydrot hiophene
Htipt 2,4,6-tri-iso-propylthiophenol
TMAO trimethylamine N-oxide
TMA t rimet h ylamine
tmen N,N,N’,N’-tetramethylethane- 1,2-diamine
To1 tolyl; 4-methylphenyl
tren tris(2-aminoet hy1)amine
tripod tris(dipheny1phosphino)methane
TYr tyrosine
uv ultraviolet
Val valine
SECTION A:
Reminiscences of Joseph Chatt

Drawn from Conversations and
from the Recollections of
Co-workers
Joseph Chatt worked in several diflerent establishments during his career, and he
has also given an account of his early life in a recorded conversation. Clearly he was
very fondly regarded by those with whom he worked, and they continue to regard
their time with him as a high point in their careers. To provide a background against
which those who did not know him might wish to assess his life and work, we decided
to ask some of those involved to write an account of their experiences working in the
Chatt group. None of those approached needed any encouragement, and their
contributions are presented in this section with only very light editing to avoid
excessive overlap.
A Memoir of Joseph Chatt
G. J. LEIGH
School of Chemistry, Physics and Environmental Science, University of Sussex,

Brighton BN1 9QJ, UK
Joseph Chatt was born in the North East of England at Hordern, County
Durham, on November 6,1914. However, most of his early life was spent in the
North West, in Cumberland. He once mentioned that, together with his father,
he had observed German warships bombarding Hartlepool, though he must
have been very young when that occurred, which was during the First World
War. Nevertheless, he did have a fine memory, and his knowledge of the
fundamental facts of inorganic chemistry was one of his great strengths. He had
an almost personal understanding of the chemistry of the metallic elements, and
his comments and advice on chemistry were generally sound, even when he could
not explain why he had made them. His understanding was intuitive and
inspirational, and though he was not afraid to admit he was wrong or even that
he had made mistakes, everything he published was prepared with the utmost
care and circumspection, often to the annoyance of his less patient and meticu-
lous collaborators.
He inherited many of those qualities that are often regarded as characteristic
of the northern Englishman. He was hard working, meticulous, blunt and honest.
He was also very determined, and shrewd in his evaluation of people, even if he
also had some related prejudices. He only slowly came to accept that women
might also become good researchers, and he had a strong suspicion of men who
sported beards. However, he could be persuaded by a demonstration of scientific
ability, even if his initial reaction to some people was less than favourable. What
sometimes inhibited his relationship with younger researchers, especially grad-
uate students, was his unquestioning assumption that everyone in the laboratory
had his interest in chemistry and also his enormous accumulation of knowledge.
We, his colleagues, always called him Joseph, even after he retired. It was
generally only colleagues from the United States who presumed to call him Joe.
Whatever he undertook to do he did with great thoroughness, whether it was
chemistry or not. At one time he was interested in antique furniture, and I can
recall being lectured on how to determine the age of a chair by examining the
4 A Memoir of Joseph Chatt
construction of the seat. This was intended to help me in my purchase of
furniture for a newly acquired house, though there was little enough money to
spare on any furniture, let alone on antiques. He was actually very self-centred
and rather insensitive, though certainly not mean. It was always necessary to ask
him for help if you required it, because he was seldom aware that it might be
necessary. However, once he was asked for help he always did his utmost to
assist.
He was an avid collector of coins from his youth, and specialised in English
and British Empire and Commonwealth coins. He had sets of Maundy money
from every English (British) sovereign who had issued such coins. At his death he
was working on a catalogue of Peace Medals, medals that were struck in towns
all over Britain to celebrate the end of the First World War, but which had never
been listed.
He was also very lucky. This showed not only in his career and perhaps in his
chemical intuition, but also, sometimes, in his hobbies. He once bought a
jardiniere at a sale in aid of Sussex churches. It was said to be Sevres, and it was
not expensive. Subsequently he decided to take it to the Victoria and Albert
Museum in London, to confirm that it was an imitation and not really a Sevres
product. The curator initially agreed that it could not be genuine, so Joseph
asked to be shown the Museum collection, which, of course, consists only of
genuine items. As they proceeded to inspect the display, starting at the later and
moving to the earlier, Joseph’s jardiniere resembled more and more the Museum
items, and finally the curator had to admit that the jardiniere was a genuine and
very rare early Sevres piece, and certainly worth much more than he had paid for
it.
When he was ten years old his family moved to a farm in Cumberland, at
Welton, just south of Carlisle, and there began a formative time in his develop-
ment. The farm was eventually inherited by his brother, and Joseph never
became a farmer, even though the agriculturally pertinent topic of nitrogen
fixation ultimately became his major professional interest. He enjoyed fell-walk-
ing and cycling, and was clearly very active. During this time he sustained
damage to his leg that left him semi-crippled, though it was only towards the end
of his life that he was forced to use a stick as well as special shoes and a leg-iron.
Cumbria and the Lake District are heavily mineralised, and the abundance of
minerals in the rocks all around his home stimulated his interest in chemistry and
also in exploring the hills. In this he was also aided by his uncle who was Chief
Chemist in a steel works in Newcastle-upon-Tyne. The young Joseph visited him
often and was given the run of the analytical laboratory. It was during this time
that he developed a very refined experimental technique and also an interest in
experimentation generally. Joseph learned how to make working models in glass
of Hero’s engine, starting with glass tubing. This required considerable skill, even
more so when one realises that he must have been obliged to use soda glass. Until
quite late in his career he required a bench to be kept in the lab for his particular
use, though he seldom had the time to indulge himself in laboratory work.
Encouraged by learning that the Romans had once found gold in the Lake
District, Joseph hunted for it in the fells around the family farm, though for once
G.J . Leigh 5
he appears to have been unsuccessful. He set up a lighting system based on a
dichromate cell in his bedroom, the control being the raising of an electrode from
the acid bath using a piece of string. At that time one could buy chemicals from
the local chemist (pharmacist) and the young Chatt was able to purchase things
such as metallic sodium and aqua regia. He once claimed that he was one of the
very few people to have observed the reaction of metallic sodium with aqua regia,
and one would not wish to query this claim! It seems to have been rather
spectacular. Somewhat later the destruction of some more of his sodium metal by
throwing it into the water at Silloth Dock on the Cumbrian coast led to reports
that the port had suffered an IRA attack, and apparently these have never been
corrected. This story is told in detail elsewhere (G. J. Leigh, Coord. Chem. Rev.,
1991, 108, 4), but Joseph was hesitant to publish the true details even in 1991,
some fifty years after the event.
Joseph’s family was not academic, and his father did not understand how the
education system functioned at that time. Consequently, Joseph remained at the
village school until he was fourteen. It is apparent that the family was never very
flush with money, and spending some on educating a child beyond the minimum
legal age in a private establishment required a considerable sacrifice. Neverthe-
less, at the age of fourteen he was admitted to the Nelson School in nearby
Wigton, for which the normal age of entry was eleven. His promise was very
quickly recognised, and he was given every encouragement by several of the
teachers, some of whom are named in the above reference. As well as receiving
local scholarships for financial support, he was given specific help with his
mineralogy and chemistry, and he matriculated within two years instead of the
usual four. Matriculation is a term no longer used, but it was then a qualification
obtained at the age of fifteen or so, and was a necessary preliminary requirement
for further study at a university.
Chatt’s arrival at Cambridge owed everything to the mathematics master at
the Nelson School. He thought that Joseph would benefit even more from
Cambridge than he would from studying at Durham, which had been the
original plan, and so in the late summer of 1935 Mr Burns went to Cambridge to
try to encourage his own old college, St John’s, to admit this promising young
man who had already achieved so much and had already won at least two
scholarships. The college was full, so he tried others, finally asking at Emmanuel
College. They were also full, but the admissions tutor said that they generally lost
a potential student ‘falling down an Alp or something, in the long vacation’. So
Joseph was given a place, on the understanding that he would obtain the
required qualification in Latin, which was at that time regarded as an indispens-
able part of every person’s University education. This he did in very short time,
and he graduated in the summer of 1937.
He carried on with graduate work with F. G. Mann, for whom he always had
the highest personal and scientific regard. Mann was one of those few pioneers
who persisted with inorganic chemistry throughout the 1920s and 1930s, at
which time inorganic chemistry was generally believed to have been finished and
exhausted, and of no further interest. Mann wished to prepare organophosphine
complexes of transition elements, because phosphines were known to produce
6 A Memoir of Joseph Chatt
low-melting adducts with transition-metal salts and these adducts could be used
to measure parachors. The parachor was expected to give information about the
nature of the coordinate bond, and these two directions, phosphines and the
coordinate bond, were to inform much of Chatt’s earlier work at The Frythe (see
below, pp. 8, 11, 18 and 25). Joseph’s first six papers, published jointly with
Mann, concerned phosphines and arsines and their complexes with elements
such as palladium. The platinum metals also became an area of early interest.
Chatt graduated when the Second World War was beginning, and he was
drafted immediately into work for the war effort. Because 1,3-dinitrobenzene
proved to be a better explosive than had been expected, it was felt (apparently
originally by Sir Robert Robinson) that 1,3,5,7-tetranitronaphthalenemight be
exceptionally good. There are 24 tetranitronaphthalenes, and Joseph was asked
to make 200 g of this particular isomer in the Cambridge University laboratory.
Unfortunately, the eight months’ effort he needed to produce the compound
were wasted as the compound proved to be a very disappointing explosive. In
truth, the Ministry of Works, his then employer, had little idea of what a man of
Chatt’s calibre might achieve and a little later they were reasonably happy to let
him resign. He took up employment with Peter Spence and Sons Ltd at Widnes,
where he worked on various war-related projects such as the reduction of
titanium tetrachloride to give aqueous titanium trichloride and the properties of
activated alumina. He ultimately became Chief Chemist. Even then, he pursued
his own private researches on olefin complexes in his spare time. It was at Widnes
that he met his future wife, Ethel Williams.
Joseph had originally planned to pursue an academic career, and had even
arranged to teach heterocyclic chemistry at St Andrews, but the war had put a
stop to that. Once the war had ended he decided to try academe once more, and
finally took up an appointment in inorganic chemistry at Imperial College. He
found the life and the facilities at Imperial College completely unsatisfying. He
was not able to do any research work, and to the end of his life he never really
ever believed that a university chemistry department was the best place to do
chemistry research. Afterwards, at the ICI laboratory at The Frythe and later, in
the Nitrogen Fixation Unit at Sussex, he made great efforts to ensure that
everyone realised that research was the prime and only significant activity of the
laboratory, and that every person and thing in the organisation should operate
in order to make the research as fruitful as possible. He was extremely successful,
and his students and collaborators have carried the message across the world.
As a result of the frustrations he experienced at Imperial College, Chatt
decided to go back to industry, and approached Dr R. M. Winter, who was then
Controller of Research at ICI, for a job. As a result, he was appointed to a
position at the ICI Butterwick Research laboratories at The Frythe, a large
country house, near Welwyn Garden City. These laboratories were intended by
ICI to be for fundamental research. After some local in-fighting, Chatt obtained
the separate inorganic department he had been promised. With himself and just a
single assistant he started what was probably the most productive period of his
research career. Personal accounts of that period and the subsequent work at the
G. J . Leigh 7
Unit of Nitrogen Fixation and the University of Sussex are to be found elsewhere
in this book.
After his formal retirement from the Unit of Nitrogen Fixation in 1980, he
moved his office into the then School of Chemistry and Molecular Science at the
University of Sussex. He continued to pursue research, though he found it
difficult to maintain his degree of productivity without the permanent profes-
sional support provided in the Unit. He was still a regular attender at chemistry
seminars, and his understanding and insights were as keen as ever, even if his
knowledge was becoming a little dated. He still enjoyed travelling, especially sea
cruises, and had taken up painting. He was particularly pleased when, unex-
pectedly, someone bought one of his paintings at an exhibition. He was in
reasonable health for his age, and supported his wife in her activities in local art
societies.
Joseph died suddenly on May 19, 1994, whilst preparing himself for a joint
photograph with the six Fellows of the Royal Society who were then working in
chemistry at the University of Sussex. The sadness at his passing was tempered
by the realisation that he had led a long, full and rewarding life, and had enriched
the world of chemistry by his discoveries and example, perhaps by more than any
of his contemporaries.
A Memorable Start to a Career
RALPH G. WILKINS
Department of Chemistry and Biochemistry, New Mexico State University,
Las Cruces, NM 88003-8001, USA
In September 1949, I was just completing a PhD supervised by Richard Burkin,

at University College, Southampton, examining the coordination chemistry of
copper(1) complexes of long-chain aliphatic amines. As luck would have it, a
position tailor-made for me had become available at the ICI Butterwick Re-
search Laboratories in Welwyn. The work would involve fundamental research
in a small group headed by Dr J. Chatt. I knew of his work with F. G. Mann at
Cambridge before the war, and after an interview, I was delighted and not a little
surprised, to be offered the position, which I accepted with alacrity! Years later,
Joseph told me that my enthusiasm had won the day over some stiff opposition,
which says more about him than about me.
The inorganic chemistry group headed by Dr Chatt consisted of Alan Will-
iams and Alan Hart who were working towards degrees, and I was therefore the
first of many graduates to join him. We were housed in a small building (hut
might be a better description) of which there were many in the grounds of a
Victorian mansion, The Frythe. It is appropriate here to say something about the
establishment. There were a number of research groups. As well as inorganic,
some others that I recall were organic, physical, microbiology and toxicology.
My memory may be a little hazy about these and other matters! Most of us lived
on-site in small huts, each housing a few people, and we were fed quite well in the
main house. Some of the (mainly senior) scientists travelled from London daily
and were picked up at Welwyn Garden City station. Joseph and Ethel Chatt had
a house in St Albans.
It was a wonderful environment in which to get to know other scientists in
different disciplines. I remember particularly the organic chemists who were
working on the antibiotic griseofulvin. Some of these were John Grove, Jake
Macmillan and Dunc (L. A. Duncanson, who was to enjoy a very fruitful
collaboration with the inorganic group). There were ‘relevant’ overtones to our
researches, but a fundamental approach was encouraged.
I can only endorse the views expressed by others’ that Dr Chatt was friendly,
Ralph G. Wilkins 9
helpful, kind - and formal. He was Dr Chatt to me until some years after I left
The Frythe, then Joseph, and occasionally Joe. It was a very friendly group that I
joined and Joseph gave me an interesting problem that he had himself gone some
way to resolving. I think that in this way he gave me some confidence and an
early publication., Beautiful canary yellow crystals are deposited when C,H, is
passed into a concentrated solution of [{PtCl,(C,H,)),] in acetone at - 70°C.
The product was shown to be dichlorobis(q2-ethene)platinum(Ir), [PtCl,-
(C2H4)J. The complex is thermally unstable and reverts to the orange dimer
with loss of C2H4on warming above about - 6°C. In a cursory examination of
the literature I have found little further reference to the compound and this is
surprising since a structural determination, although tricky, would be wor-
thwhile. Joseph thought the compound would have a trans conformation.
In September 1950, Joseph decided to invite about 30 coordination chemists
to a meeting at The Frythe. This was to become the first International Coordina-
tion Chemistry Conference (international because both K. A. Jensen (Danish)
and Gerold Schwarzenbach (Swiss) attended!). Little need be said about this
meeting since it was featured in the exhibition of historical material from early
ICCCs at the 34th ICCC (Golden Jubilee) held in 2000 at the University of
Edinburgh. In 1950, all eleven talks took place over two days in the lounge of the
main house at The Frythe. We also had a nice dinner and social evening at the
Clock Restaurant in Welwyn. N. V. Sidgwick, who had just completed his
mammoth two-volume work on the chemical elements,, gave a short talk and I
recall that he reprimanded one speaker for using the phrase ‘data is’! It was a
splendid idea of Joseph’s to bring together active workers in the area and I
remember very well to this day the meeting and the impression it made on me.
Now Leslie Orgel and I are the only speakers still around. I gave a talk about the
preparation and equilibria between cis- and trans-complexes [PtCl,(ER,),]
(E = P, As or Sb), the study of which occupied most of my attention during my
stay at The Frythe. It formed the basis of four papers4 and Joseph presented the
work at the ICCC in Copenhagen in August, 1953. We followed the equilibria
using the vastly different dipoles of the two isomers.
We had a very sensitive meter, built in the excellent workshops on the site, to
determine the dielectric constants of benzene solutions of the complexes from
which dipole moments were calculated. Important findings were (1)that the total
bond energy of the cis isomer is about 10 kcal greater than that of the corre-
sponding trans isomer; (2) that the entropy differences between isomers, arising
from the high dipole moment of the cis isomer, can control the positions of
the equilibria; and (3) that traces of proligand (R,E), either added (E = P) or
autoproduced (E = As or Sb), are needed to establish the equilibria. These
studies were the first of their kind and evoked immediate interest.
Since a good deal of the pioneering work on equilibria of complex ions was
carried out in Sweden and Denmark, Joseph thought it would be a good idea to
visit and talk with some of the people there who were known to us only as names.
I was pleased when Joseph asked me to accompany him. We had an enjoyable
and fruitful trip in June of 1952. We were fortunate to meet both Bjerrums (Nils
and Jannik), Kai Jensen (who had done early studies of tertiary phosphine, arsine
10 A Memorable Start to a Career
and stibine complexes) and Sture Fronaeus, amongst others. Sten Ahrland has
written about our meeting at Lund, which resulted in his joining the group in
1953 and starting a ‘happy collaboration.. . with a great scientist and also a true
gentleman’.
After an enjoyable and rewarding three years at The Frythe, I decided that I
needed a change of direction. I went to the University of Southern California to
learn about kinetics from Arthur Adamson. I believe, perhaps wrongly, that
Joseph did not much like kinetics, indeed that he was uneasy with conclusions
based on rate data. At the end of the year at USC, and mainly as a result of a chat
between Joseph and Professor R. D. Howarth, chemistry department head at the
University of Sheffield, I was encouraged to apply for a vacant lectureship there
and was appointed, sight unseen (imagine that nowadays)!
I learnt a great deal in my three years with Joseph. Probably most important
was scientific integrity. Facts published should be correct and as accurate as
possible. Speculation should reside in ideas, which might very well turn out to be
inaccurate.
References
1 C. Eaborn and G. J. Leigh, Joseph Chatt, CBE, The Royal Society, 1996.
2 J. Chatt and R. G. Wilkins, Nature, 1950,165,859.
3 N. V. Sidgwick, Chemical Elements and Their Compounds, Oxford University Press,
Oxford, 1950.
4 J. Chatt and R. G. Wilkins, J . Chem. Soc. 1951,2532; 1952,273; 1952,4300; 1953,70.
5 S. Ahrland, Coord. Chem. Rev., 1996,154, 13.
Joseph Chatt and The Frythe: A
Memoir of the Early 1950s
GEORGE A. GAMLEN
One-time ICI employee at The Frythe
Joseph Chatt was successful and became famous because he was the right man in
the right place at the right time. He deserved his glory because he knew what he
wanted and he worked hard to get it. It goes without saying that he was a
brilliant chemist, but in addition he had a phenomenal intuition rooted in his
encyclopaedic knowledge of the subject.
In 1946, The Frythe was set-up by Sir Wallace Akers, ICI’s Research Director,
as a central research laboratory. It was hoped that chemists in ICT’s manufactur-
ing Divisions would follow the progress made at The Frythe, and pick up any
results which might be of commercial value. Chatt arrived there in 1947, bub-
bling with enthusiasm and with a grandiose plan; he wanted to establish a
world-class centre for inorganic chemistry that would rescue the subject from the
doldrums. His main theme would be the nature of the coordinate link using
complexes of the platinum-group metals as models. The first thing he needed was
a very good laboratory. Drawing on his experience in Cambridge with organo-
derivatives of the Group 15 and Group 16 elements, he installed powerful fume-
cupboards including one floor-to-ceiling walk-in cupboard. Large quantities of
compounds such as phosphines, arsines, selenides and tellurides were made in
this and sealed in glass phials of convenient size. It was a test for newcomers that
their first experimental work was usually to prepare one of these on the five-litre
scale, using a Grignard reagent. Success, or the lack of it, was used as a measure
of competence and potential usefulness to the department.
Chatt was very adept at experimental work and, for example, could usually
induce crystallisation when others had failed. He kept a magnifying glass in his
pocket to examine solutions for signs of crystals and if he spotted anyone
scratching with a glass rod, he would take over. Holding the glass and the tube
close to his eye, he would inspect the contents before beginning his own particu-
lar magic. Meanwhile, it was a matter of waiting with bated breath for the
outcome - which was usually greeted with a sigh of relief. It used to be said that
12 Joseph Chatt and The Frythe: A Memoir of the Early 1950s
nineteenth-century chemistry professors had large beards because they were a
reservoir for seed crystals and I used to wonder whether the old-fashioned 1930s
moustache (somewhere between Adolf and Groucho) that adorned Joseph’s
upper lip served the same purpose.
As the experimental work was so demanding, he carefully selected the school-
leavers who were to be his lab. assistants and personally trained them in
experimental methods. Some of these entrants later rose to senior academic or
industrial positions as a result of their ability and good fortune in having such a
fine teacher.
The utmost care was taken during preparations, but despite all the washing
and scrubbing, the repulsive smell of the products seemed to stick to the skin and
clothing. In this respect, the tellurides were particularly offensive. Lab. assistants
going home used to note how seats around them on buses would quickly empty.
There was bench space in the laboratory for about eight workers who set up
their apparatus in large enamel photographic trays. Any spillage could thus be
caught and the precious metal recovered. Even so, there was the occasional loss
which was signaled by a doleful rendering of a song beginning ‘My plat’num has
gone down the plug-’ole’ (tune: ‘My bonny lies over the ocean’) but not if Chatt
was in earshot. Precious metal recovery was a requirement of the experimental
work and all the used filter papers and melting point tubes, ground up, were
added to the residues.
The solvents commonly used were diethyl ether, benzene (not then recognised
as a carcinogen), ethyl alcohol and acetone. Once the staff numbered four or five,
a set-up to distil-off solvent from the collected reaction solutions ran every day so
that the solid residues could be returned to Johnson Matthey for recovery.
Chatt’s first secretary still has a vivid recollection of the numerous fires that
occurred (about once a month) but fortunately these were always put out before
serious damage was done. On one occasion, when Chatt was bringing a distin-
guished visitor from the house to the lab., he noticed from a distance that a thick
cloud of black smoke was pouring out of the fume-cupboard chimney and took
the visitor to the arboretum to see an outstanding example of a maidenhair fern
instead.
Chatt already had it in mind that physical measurements would be an import-
ant part of his work so there was a constant-temperature room which contained
the dielectric constant machine and later a UV-visible spectrophotometer and a
home-made calorimeter for measuring heats of cisltrans-isomerisation of com-
plexes. L. A. Duncanson and his infrared spectrophotometer, with which the
crucial spectra of the olefin-platinum complexes were measured, formed part of
the Organic Chemistry Department in the next hut. As soon as he could, Chatt
added an X-ray crystallography unit, run by his deputy, P. G. Owston, which
was located in the old stables of the main house.
The site had its own microanalytical lab. with two first-class analysts. Some of
the metal complexes were difficult to analyse accurately by standard methods
because of the presence of the metal. The analysts’ skill in solving these problems
was a great help when new compounds were discovered.
Chatt’s office was of modest size and he sat with his back to a wall that was
pierced by a hatch. His secretary sat on the other side of the wall and as Joseph
George A . Gamlen 13
finished each page of a new manuscript he would hand it through the hatch to the
secretary for typing. He always wrote in pencil, selecting one from a jarful on the
desk, and was careful to re-use the back of any paper that had been used for a
previous draft. A small and grubby rubber was used (rather imperfectly) to make
corrections and new material was inserted by writing above existing text. Some-
times he wrote a further insertion over the top of the first insertion and, as even
his best writing was quite difficult to read, his long-suffering secretary struggled
to cope. One day, when he had produced a particularly intricate and messy draft,
he put it through the hatch, closed it and said, smiling, ‘You know, I think that if I
just scribbled a wavy line on the page, Celeste would be able to write the paper
for me’.
Having established his base and begun experimental work at the bench, Chatt
began to recruit staff. He was helped in this by knowing many of the Heads of
university departments who recommended some of their best students to him.
He loved scientific meetings and discussing chemistry with people and was
delighted when the laboratory manager, M. T. Sampson, affectionately known as
Sammy, encouraged him to hold a conference on coordination chemistry at The
Frythe in 1950. This was later recognised as the First International Conference
on Co-ordination Chemistry. With Ralph Wilkins, the first PhD member of his
staff, he invited to attend everyone in Great Britain who had published on the
subject, and some key figures from overseas as well. There were 23 from British
institutions including the Australians, Craig, Maccoll and Nyholm; and three
from abroad, Albert (Australia, temporarily based in London), Jensen (Den-
mark) and Schwarzenbach (Switzerland). There were also six from ICI Divi-
sions in addition to the eight who attended from all the departments at The
Frythe.
Two of the lectures were seminal for Chatt; Irving spoke on the stability
constants of coordination compounds in aqueous solution (the Irving-Williams
series) and Orgel on the significance of d-orbital hybridisation in coordination
compounds. Chatt could see the possibility of another series of compounds akin
to the Irving-Williams series based on donors other than oxygen and nitrogen,
specifically the donors of Groups 15 and 16 and he knew how to get round the
the problem of the insolubility in water of the platinum and palladium complexes
by sulfonating the ligands. What he needed now was a good physical chemist,
preferably with experience of determining stability constants. The first candidate
he tried was British but proved unequal to the task and he had to look further
afield. He found the ideal man, Sten Ahrland, at Lund University.
At that time, one of the main geographical areas where complex chemistry was
a major topic of research was Scandinavia. In Denmark there was a long
tradition of chemical excellence both on the preparative and physical side. Zeise,
who in 1827 was the first to prepare an ethylene-platinum complex, was the first
professor of chemistry at the University of Copenhagen. By a great coincidence,
Jarrgensen, who later was Werner’s principal adversary, was not only born in the
same village, Slagelse,as Zeise, but in a house on the same street. Jannik Bjerrum,
son of Nils Bjerrum, signalled a major advance when he published his PhD thesis
in 1941. In it, he described the stepwise formation of complex ions and the
determination of stability constants. This, and other important papers which
were published in Denmark during the war, were well known in the other
Scandinavian countries. Further groundbreaking work was done, for example,
by Leden in Sweden and Flood in Norway.
Complex ions, a term which my mother always willfully misread, (‘Oh, yes, my
son is studying complexions’) became a highly popular Nordic research field
with Lars Gunnar Sillen in Sweden as the dominant figure until his sudden and
premature death. Schwarzenbach in Switzerland was an early pioneer too, but,
because of the war, the importance of complex ions in solution remained largely
unknown in the UK for six or seven years. It was Irving and Williams who drew
attention to Bjerrum’s work late in the 1940s.
Sten Ahrland joined the team in 1953. It was a cold winter that year and I
thought that, coming from Sweden, he would be used to it. In that I was wrong.
Ahrland never did get used to how cold England was, and he explained to me
how flats in Sweden were triple-glazed, and required by law to maintain a decent
temperature whether they were occupied or not. It was particularly bad on
Monday mornings at The Frythe as the heating was turned off over the weekend
and the huts were poorly insulated. Ahrland would sit in his office wearing his
overcoat, hat, scarf and gloves and place a two-bar electric fire on the desk in
front of him, trying to keep warm. He was a very gentle giant with a highly
developed sense of humour and a ready laugh. We did not meet again after he left
The Frythe but corresponded until his death.
Luigi Venanzi arrived at The Frythe before Sten Ahrland. His antecedents
seemed shrouded in some mystery, as he did not talk about them. He said he
came from Trieste and had worked in North Africa and Germany. He arrived in
England unable to speak a word of English but he had a gift for languages as we1.l
as for chemistry and within six months was speaking it fluently. It was only when
he said ‘abroad’ which he pronounced ‘abrode’, that you knew he was not a
native speaker. Venanzi adapted quickly to life in England. He was very elegant,
carried a rolled umbrella when he went out and smoked a fine pipe. He became
the epitome of an English gentleman with beautiful manners. In the laboratory
he was immaculate in everything he did. His lab. coat was always clean, with no
creases and completely stain free. His bench was always tidy and his notebooks
written in a neat hand. He got through his work at a tremendous rate and made
two or three new compounds every week. Venanzi occupied the bench adjacent
to the front door so he was often the first person to greet the many visitors who
now came to the lab.
Ralph Wilkins decided to join the brain drain and, encouraged by Fred
Basolo, moved to the University of New York at Buffalo, so in 1952 I became the
second Technical Officer that Chatt recruited onto his staff. When I went to The
Frythe for interview, I was first taken to meet Mr. Sampson, who was in charge of
The Frythe. He made two points to me: (a) that if I were appointed, he simply
wanted me to do the best academic research that I could; and (b) that I should
regard him as though he were my tutor and come to see him whenever I had a
problem. I found it difficult to believe him on either score. I could not see why a
large commercial organisation should pay people just to do academic research
and Mr Sampson had a Roman nose in the middle of a face which might have
belonged to an emperor. There was also a touch of hauteur in his manner that
seemed to belie his friendly words. Then I went down to the lab. and met Chatt.
In those days he was shy and diffident when meeting people but he greeted me
with ‘Well, you’re a pretty rare species’ and was avuncular and helpful. I was thus
emboldened to ask him if it was really true that we were only expected to do
academic research. He confirmed this and showed me the papers that they had
already published - a substantial collection for such a short time.
Later, I found that I had misjudged Sammy on the second count, too. At the
end of the day, Chatt told Sammy that he wanted to offer the post to me and they
immediately arranged a final interview at the ICI Headquarters in London, in
two days time. It only took ten days from the time that I saw the job advertised in
the paper to the visit to London when I was formally offered the post. My only
worry was that I had already been accepted for the Scientific Civil Service at a
salary of E830 a year and ICI only offered me E700. I explained this to Chatt in a
letter and already felt on such good terms with him that I asked him what he
thought I should do. He wrote a long reply filled with wise words, pointing out,
for example, that ICI was such a big slice of the chemical industry that whatever
they paid was going to be the average. He mentioned that he had also had
experience of working for the Scientific Civil Service and ICI offered very much
better working conditions. Finally he said that he wanted me to join him and if
he could get me another E50 a year would I give a firm undertaking to accept the
offer? I agreed, and never regretted the decision. In 1956, I transferred to one of
ICI’s main Divisional Research laboratories to study metal complexes as cata-
lysts. Three weeks later there was a re-organisation of the management because
the Division had decided to exploit the new fibre-reactive dyes that the Research
Department had found. My new boss sent for me and said that the Division was
not interested in my work. He gave me the option of starting dyestuffs research or
leaving the Company, and he didn’t mind which I chose. As there was really no
choice, I started dyestuff research. Fortunately, I found that there were some
good dyes which were cobalt and chromium complexes, so I worked on them. It
turned out to be rather easy to make mixed-ligand metal complex dyestuffs.
These gave tertiary shades at your whim, e.g. in an octahedral complex of
chromium or cobalt, if three apices are occupied by a tridentate yellow chromo-
phore and the other three by a blue the result is a green dyestuff. (Similarly, red
plus yellow gives orange, etc.) Addition of a colourless reactive system to one of
the chromophores then gives a fibre-reactive dye with excellent fastness proper-
ties. Useful, but trivial compared to what might have been. Seven years later,
when Chatt left the Company, I was asked to take over the team that he would
have led at the newly founded Corporate Research Labs., but by then the
substantial time advantage we could have had was lost.
Joe Chatt and The Frythe were as great a success academically as Sir Wallace
Akers, the ICI Director of Research responsible for setting up the laboratory,
had hoped for, but people always ask, ‘Did the Company get anything out of it?’
They certainly did not get a mega-invention such as polythene (which had been
developed in one of their Divisional labs.). There were certainly near-misses; at
the Corporate Labs., Steve O’Brien found that zirconium tetra(n-allyl) was a
high-mileage catalyst for polythene. Somehow it didn’t fit into the Company’s
operations, and the years dragged by. In the end, polythene was sold off, so the
catalyst was never used.
The most disappointing miss for me, however, concerned Pharmaceuticals
Division. While I was at The Frythe, I was responsible for sending samples of all
the compounds we made to them for biological testing. Six months later and
after well over a hundred compounds had been dispatched, we still had no results
from them. I spoke to the man in charge of testing and he said that there were no
results to send. None of the compounds was active so he hadn’t thought it
necessary to reply. In view of some of the elements that were contained in these
compounds, it seemed surprising to me (as a non-expert) that none showed any
activity. My opposite number replied that he was not a bit surprised as they
normally tested thousands of compounds before they found one that was active.
Keen to be sure, I asked whether not a single one had been of interest. It seems
that one of them, a platinum complex, caused the test bacteria to swell up to
many times their original volume, but that when the complex was washed out,
the bacteria went back to their normal size. This compound was actually one I
had made, cis-diammineplatinum dichloride! It seems that they already had so
many other compounds with apparent potential that one exhibiting only ‘tem-
porary’ activity was not of any great interest.
Squashed, I let it go at that. Years later, and especially when I found out the
extraordinarily circuitous path by which cisplatin, the anti-tumour drug, was
discovered, I realised with regret what an opportunity had been lost.
During the last years of The Frythe, some attempts were made to relate the
chemistry to industrial problems and to surmount communications problems
between The Frythe and the Divisions; for example, some staff from Divisional
research labs. were seconded to The Frythe. They found that it was easier to
make academic advances than to find useful new catalysts or processes. It was
not realised that the evaluation and development work needed a much bigger
specialist team than the pure chemistry research.
It was also decided to move responsibility for Chatt’s group into the Heavy
Organic Chemicals (HOC) Division, who were using the OX0 process to make
aldehydes with a cobalt catalyst and developing a process for vinyl acetate based
on Wacker-type chemistry. At one of the liaison meetings, someone from Chatt’s
group mentioned an elderly Russian paper that described the catalytic activity of
the platinum metals group and pointed out that rhodium was unusually active.
The HOC chemists tried it and found that not only was rhodium more active
than cobalt, so that milder conditions could be used, but that the product had a
much more useful normal-to-is0 ratio. There was only one snag: the existence of
the Russian paper meant that the process could not be patented. However, the
best efforts of the ICI physical and analytical chemists were unable to find any
trace of rhodium in the polymer product so it was decided to keep the use of
rhodium as a closely guarded secret. When Wilkinson published his famous
paper on the rhodium catalyst that bears his name, there was an inscrutable
smile on the face of the HOC chemists.
It may be said that the best reward for the Company as a result of setting up
The Frythe was the highly talented people who were attracted to join the
company as a result of Chatt’s work. Many of them later moved into academic
life and a good number were appointed to Chairs. People from The Frythe and
The Corporate Labs. who moved onto the industrial side also did very well. Two
reached Directorial level inside the Company and several did equally well
out side,
Was Sir Wallace Akers wrong to set up The Frythe as he did? In the sense that
it proved unrealistic to expect the Divisions to see the possibilities for research
that they themselves had not done, it may have been. On the other hand, it may
take decades before an academic advance fructifies into a major invention. It is
only now that we are beginning to see the emergence of new materials based on
nanotechnology that are of great importance to the electronics and computing
industries. Some of these are metal complexes and descendants of the chemistry
that Chatt did so much to forward.
Sir Wallace’s vision may thus still be realised, but it may not be ICI that
benefits financially. This would not bother him unduly: his ultimate goal was to
benefit mankind and the work at The Frythe most certainly did that.
Acknowledgements
My thanks are due to colleagues from The Frythe and the Corporate Research
Labs. for their help with personal memories of that time, especially Dr P. G.
Owston, Dr D. T. Thompson, Dr K. A. Taylor and Dr G. Booth. Any errors or
infelicities are mine.
Recollections of Lve with Joseph
Chatt at The Frythe, 1958-62
DAVID M. ADAMS
Joseph Chatt and The Frythe will always be associated in my mind, for it was
there that I spent four and a half years in his Inorganic Research Group, a time of
learning, fruitful collaboration, and real enjoyment for which I remain extremely
grateful. Indeed, when ICI, our mutual employers, saw fit to close the Inorganic
Group at the end of 1962, Joe and I were the last two members of the research
group actually present on the final day, which I believe was Christmas Eve. The
laboratories had been closed, and Joe had been banished for his few remaining
days to a remote office upstairs in the old house that stood at the centre of the
site. There it was that the two of us put the finishing touches to a paper for
publication in J . Chem. SOC.It was the end of a remarkable era during which Joe
had made outstanding contributions to the new field of transition-metal or-
ganometallic chemistry that had earned him election to the Fellowship of the
Royal Society, as he became one of the UK’s leading scientists, embellished the
reputation of a none-too-grateful ICI, and launched a large group of co-workers
into distinguished careers in industry and the academic world.
I had joined ICI at The Frythe in 1958 as a physical chemist, following a PhD
plus post-doctoral year under the direction of one of the great scientists of the
previous generation, Professor Sir Eric IS.Rideal FRS. During my post-doctoral
year I had applied infrared spectroscopy to the study of heterogeneous catalysts,
which was then a technique in its infancy insofar as the field of catalysis was
concerned. At that time we were able to study only the high frequency internal
vibrations of the adsorbed species. A simple calculation based upon the entropy
of adsorption showed that if we were to detect the stretching modes of the
adsorbed species as it vibrated against its metal support, it would be necessary to
work in what was termed the far-infrared, the region that lies between the
infrared and microwave parts of the electromagnetic spectrum. No commercial
instruments then existed for that region, so I began to design one.
It so happened that my daily travelling companion during that year was a
David M . Adams 19
certain assistant lecturer at Imperial College, London, Dr Jack Lewis (now
Professor Lord Lewis). One day he arrived on the railway platform with the news
that a group in ICI was looking for someone to build an instrument to do
far-infrared spectroscopy, and to apply it in transition-metal organometallic
chemistry. Was I interested? About a week later I found myself in the delightful
surroundings of The Frythe being interviewed by Joe Chatt and various of his
co-workers. I was suspicious of entering industry, being bent on an academic
research career, but the opportunity seemed too good to miss, and the money
was good for those days - which was a consideration in only the second year of
my marriage.
Joseph recognised early in his work on organometallic chemistry that struc-
tural and spectroscopic methods would be of particular importance to the field.
It is greatly to his credit that he built in ICI a preparative organometallic
research group with considerable strength in physical techniques of characterisa-
tion. Moreover, those of us on the physical methods side of the group were
encouraged to work at the cutting edge of our respective fields and to pursue our
own lines of research, whilst never forgetting that our primary reason for
existence was support of the preparative studies.
During our time at The Frythe, X-ray methods were primitive by today’s
standards. Raw data were collected on film and then laboriously turned, spot by
spot, into intensities. Rows of young ladies sat at mechanical calculating machin-
es, processing the resultant numbers, writing each new result on a slip of card
before moving to the next one. A single cycle of refinement could take three
months. Not surprisingly, the same young ladies were notable for their liveliness
at staff parties!
At such events Chatt was an avuncular presence. Unable to take an active part
because of damage to his foot, he would hold court in a corner, telling anecdotes
whilst sipping judiciously at a drink. Now and again, he and Ethel Chatt would
invite the whole group to their home and these were always delightful times.
They usually ended with extended viewings of Joe’s slides from his latest Ameri-
can trip. On one such occasion, I recall him running back and forth through a
series of slides on Niagara Falls and the related river system, until he was sure
that all of us, from secretaries to senior section leaders, had grasped the details of
the geology and hydrography of the region, meanwhile fielding hints from Ethel
to the effect that, quite probably, these nice people had already grasped the
essentials!
Life at The Frythe was good. The Frythe had a walled garden and many acres
of surrounding land. During the Second World War it was used as a hush-hush
research and development site by the Admiralty who, in particular, installed
excellent machining facilities, and turned the walled garden into a car park. After
the war ICI developed the site and built new state-of-the-art laboratories that
were eventually occupied mainly by Joe Chatt’s group. Complete with a reason-
ably well-stocked library, and a librarian, plus a good range of support staff, not
to mention the absence of a teaching load, conditions for academic research were
almost ideal. True, we lacked the daily contact with the wider community that a
university affords, but conference attendance was encouraged and paid for, and
20 Recollections of Life with Joseph Chatt at The Frythe, 1958-62
our daily tea and meal breaks were a fair approximation to the real thing. Our
publication rate was the envy of many a university.
The Inorganic Group was not alone at The Frythe. There was a distinguished
organic research group, also led by a Fellow of the Royal Society (Dr Bryant),
which specialised in plant growth hormones, such as giberellins, and other
natural products. These had obvious commercial prospects and therefore made
the group popular with senior people in the Company. The site also housed the
Industrial Hygiene Unit, a Company-wide resource entrusted with keeping an
archive of data on chemical toxicities, especially with respect to ICI’s manufac-
turing interests.
This mix of disciplines made for an interesting social life. Christmas celebra-
tions at The Frythe always included devoting the final afternoon before closing
for the festival to a concert party, more in the style of a pier show than a concert
hall. The site harboured some remarkably talented individuals. In Joe’s group
alone we discovered one year a brilliant jazz pianist who could also act, and a
gifted left-handed guitarist, recently joined from Oxford University, who an-
nounced in a bored voice that he was about to sing some madrigals that he had
found in a dusty manuscript in the Bodleian Library. The music it is true, could
have come from madrigals, but the words would have probably have been
considered a shade too ripe even for Carmina Burana, although their inventive-
ness could not be faulted. One chronicled the exploits of a mediaeval knight,
another the proclivities, ultimately tragic, of a warm-hearted Russian lady
named Olga. The number of encores was a record.
Catering at The Frythe was of a high standard, being under the direction of a
personable and somewhat formidable lady whose authority in her own field was
fully the equal of Joe’s in his sphere. My memory is of a constant stream of
visitors of many nationalities to the Inorganic Group. Most of us would be called
upon at some time to help out with this none-too-onerous task of entertaining, in
the process becoming almost addicted to the standard Company tipple of gin
and bitter lemon.
As a result of the supposed scientific standing and expertise of people in Joe’s
group, now and again one or more of us would be required to visit one of ICI’s
manufacturing Divisions for a spot of internal consulting. The conditions under
which we travelled were in strict contrast to those most of us had left not far
behind as research students. One set off from The Frythe, which was on the edge
of Welwyn village, to the railway station in one of the Company’s chauffeur-
driven Rovers, all walnut fascia and leather upholstery. First-class tickets were
permitted to those with five or more years service, or when travelling with a more
senior employee, so one generally tried to engineer that. Met on arrival by
another chauffeur, one would be whisked effortlessly to some dark satanic mill,
there to do one’s best, before the inevitable trek to the hospitality suite. It was a
hard life.
Being experimental scientists, we also felt obliged to probe the limits of our
expense accounts. This led on several occasions to strained interviews with
senior staff,during which our scientific training in arguing a case and leading the
reader to agree with our conclusions proved invaluable. It was never clear
David M . Adams 21
whether or not the annual salary review was influenced by these discussions. Be
that as it may, Joseph held the ultimate sanction within his Group of recommen-
ding whether, and, if so, how much, each of us was awarded annually, and he
used it.
From time to time Joe would feel the urge to mount a small conference at The
Frythe, generally aimed at widening our understanding of the field in which we
worked. These were small, select affairs,with a distinguished visitor playing a key
role. A. F. Wells, the noted crystallographer and a long-time employee of the
Dyestuffs Division of ICI, was one such. He visited us not long before he was
seduced to retire early and emigrate to the USA. Jack Lewis, then a lecturer at
University College, delivered a series of lectures on physical-inorganic chemistry,
with some emphasis on magneto-chemistry, all of which was very helpful as there
was still not a lot in print at undergraduate level on that sort of thing. Leslie
Orgel, then at the height of his fame at Oxford, was another welcome visitor, who
also acted as a consultant before vanishing to California to study ageing before it
was too late.
In retrospect, it was a remarkable time, a golden era, although probably few of
us realised it then. Joe’s group at The Frythe was just part of the tremendous
world-wide renaissance of inorganic and organometallic chemistry, and a high
proportion of those then involved, both in our group and elsewhere, later made it
to positions of prominence and even glory. Being so close to London, it was easy
to visit and participate in the Chemical Society meetings and conferences at
Burlington House, Piccadilly, and therefore to meet most of the great and good
as they passed through, or to drop into the Chemistry Department at University
College, London for the latest news. There, Ron Nyholm, not yet an Fellow of
the Royal Society, reigned genially over the Inorganic Department, and had
gathered about himself an extraordinarily talented bunch of ex-patriot Aussies,
every one of whom would later become distinguished in his own field. And
probably as many research ideas were conceived at Schmidt’s, a much-favoured
German restaurant famous in season for its Strasbourg game pie, in nearby
Goodge Street, as in the Department.
Now and again we would be caught up in some Company-wide scientific
event. One of these jamborees was held at The Frythe about 1961. It included a
banquet at a local restaurant that considered itself to be at the upper end of the
market, a relative term even in those days. It so happened that one of the major
scientific participants also doubled as the wine buyer for all the Company’s
canteens and hospitality suites. Service was good. We had all been served our
main course and a glass of red wine when there was a commotion at the top table.
Our wine buff had declared the vintage ‘corked’ and unsuitable for ICI em-
ployees. All of it was withdrawn and replaced. The manager was not a happy
man, and there were more than a hundred of us in the party.
Quite early in the life of his group at The Frythe, Joe achieved a ‘first’ of
considerable note in the field, and one that gave him much satisfaction. Specifi-
cally, he and Bernard Shaw synthesised the first complex metal hydride without
supporting carbon ligands. This molecule contains a strong, stable rnetal-to-
hydrogen bond, in this case platinum-to-hydrogen. A considerable series of
related platinum and palladium complex hydrides was made. The metal-hydro-
gen bond was shown to absorb infrared radiation near 2200 cm-l, and to vary in
frequency with the nature of the groups around the metal atom. Whilst this,
together with other physical data, was convincing evidence of their claim, the
final proof came in the form of an X-ray crystal structure analysis, also done
in-house, and which itself was proof of Joe’s foresight in establishing under one
roof almost all the then-known physical methods needed for characterisation of
organometallics.
This is not quite the end of the story. Joe’s group was not alone in trying to
make complex metal hydrides. Geoff Wilkinson, then recently arrived at Im-
perial College, London, following a distinguished period at Harvard and in the
Atomic Energy programme, was also working on the problem. Joe achieved a
double first in that he got into print ahead of the competition and, unusually for
those days, provided the clinching evidence of the X-ray structure. Wilkinson
was not happy. Having just come from Imperial College myself, I was able to give
Joe the possibly apocryphal news that the Imperial College copy of the journal in
which Joe’s announcement appeared bore across his article the imprint of a
muddy boot. True or not, it gave Joe a moment of the most exquisite pleasure,
not to be repeated, to my knowledge, until the announcement of his Fellowship
of the Royal Society.
Support services at The Frythe were excellent, and contributed greatly to the
Group’s research productivity, so much so that those of us who eventually
returned to academia sometimes suffered re-entry problems. It was never necess-
ary to chase an outside order: a man in the Purchasing Section did that auto-
matically. Library services, glassblowing, machining and carpentry were all on
tap. And there was that invaluable lower level of research help, the Assistant
Experimental Officer (AEO) who might or might not be a graduate, available to
do technically-demanding donkey work in support of the Group.
One such AEO underpinned the entire research effort of Joe’s Group and was
really the unsung hero of it. The principal stabilising ligands used in Joe’s
research were tertiary-, and sometimes di-tertiary- organophosphines. Synthesis-
ing these was generally far from trivial, particularly in the case of the di-tertiary
members, all of which are highly air-sensitive and can be prepared only under an
inert atmosphere. Due to their toxicity and penetrating smell, these materials
were made in a downstairs laboratory lined with fume cupboards, dedicated to
that purpose. There Mike Searle worked in glorious isolation. He devoted
several years of his life to this task, becoming very expert in their production, and
maintaining a steady supply of old and new types for the preparative team. As is
inevitable in such work, Mike absorbed a proportion of these chemicals, prob-
ably though the skin. A tall, cheerful and convivial man with a strong West
Country burr, he nevertheless found that this placed serious limitations on close
social contacts of almost every kind.
Joe’s election to the Fellowship of the Royal Society (FRS) gave much pleasure
to the whole group, past and present. Many of us were headed for academic
careers and it did us no harm to be seen to be part of such a distinguished team.
Moreover, the superb conditions under which we worked had allowed us to
David M . Adams 23
generate substantial numbers of publications that filled out our curricula and
gave us a head start amongst our contemporaries.
Quite rightly, Chatt kept the pressure on in respect of getting our work into
print. In part that was probably his way of keeping ICI’s senior management
quiet by pointing to the undeniable flow of publications in quality journals. In
practice, none of us needed much urging as it was in our own interests as much as
his and the Company’s, to get the stuff out. The Company required us to issue
each new paper first as an internal report (in green covers) that would be
circulated within the Divisions for their perusal. Presumably this was to allow
anyone within ICI to shout ‘Hold it!’ if they saw something of commercial value.
Permission given, the paper would then follow the usual route into print. The
number of publications from Joe’s group was remarkable by any standards. I
cannot quote a figure for the final tally, but I do recall a large set of filing shelves
in his secretary’s office, full from top to bottom with rows of reprints.
I remember the day the postman delivered the package of material that comes
to every FRS upon his election. There were items bound in the classic maroon
colour used by the Royal Society, there were things on parchment, a document to
be filled in with his biographical details to form a factual basis for his eventual
obituary, and goodness knows what else. He sat at his desk, dressed as usual in a
smart suit of American origin, with these things spread out around him, doing
very little but just quietly glowing with satisfaction, as well he might. And as we
congratulated him again, his response was: ‘There’s no reason why any of you
shouldn’t do the same.’ At least one did.
The powers that be at ICI acknowledged Joe’s election to the Royal Society by
throwing a banquet at their Millbank headquarters on the Thames, near the
Houses of Parliament, and we and our wives all trooped down to celebrate. No
expense was spared, and even the beer drinkers amongst us remarked on the
quality of the wines served. Joe was not ICI’s favourite son, but they were in a
cleft stick and they did what was expected of them. True, the Main Board
Research Director of the day, a lugubrious and inscrutable Scotsman, in a speech
notable for the delicacy with which he placed his feet, did describe Joe as ‘a
complex character’. One had the suspicion that no double-entendre was meant.
Writing a paper for publication jointly with Joe was a process. The individ-
ual(s) concerned would write the first draft and have it typed, double-spaced. A
copy went to Joe. In due time, one was called to his office and there would work
through it, sentence by sentence, paragraph by paragraph, considering in turn
the science, the accuracy of presentation, and the style. The majority of our work
went to one or other of the journals of the Chemical Society, which at the time
enjoyed the services of a senior editor of the old school whose mission in life was
to train a generation of chemists in the correct, precise, and especially, concise
use of English. I, personally, remain indebted to him for the training, although
that was not the immediate response at the time.
Joe was a natural ally of the editor in this process, and he insisted that every
detail be checked. I well remember one occasion on which two of us sat with him
and were arrested at a sentence that contained the word ‘co-ordination’. Joe
thought that it should be written as ‘coordination’. We instantly agreed with him,
having no axe to grind, and a pressing desire to go to lunch. We were rumbled.

There followed an excruciating half-hour in which the three of us researched the
use of the word: to this day I forget the outcome, but the trauma remains as fresh
as ever!
A manuscript would commonly go through several revisions, being re-typed at
each stage, complete with references. This, remember, was before the days of
word processors, and copying machines were in short supply, so mostly we
worked from carbon copies. Basically, secretaries were on the consumables
budget! The final version of a manuscript would be checked for spelling and
other infelicities by two or more of its authors, who were required by Joe to read
the thing to each other in reverse word order, the theory being that since it made
no sense that way, we would more readily spot the errors. Mostly it worked,
although I do recall one instance in which a colleague and I had done just that.
We took the faultless manuscript proudly to Joe, only to find that in our relief at
completing the miserable task, we had failed to correct a typing error in the very
first word. Joe was pleased!
Why did it end? Basically because the bean-counters in ICI couldn’t see a
return on their investment. Joe was never really at home in an industrial
environment, not that The Frythe could even remotely be termed industrial in
ambience. He was an academic to his fingertips and flourished in that atmos-
phere, as indeed he had done at Cambridge, and did later at Sussex University
when he headed the Nitrogen Fixation Unit of the Agricultural Research Coun-
cil. He never cultivated relationships within the company, finding the political
side of life distasteful. Consequently, he was unprotected when a chill wind began
to blow.
In retrospect, what was lacking in the Inorganic Research Group was any
sustained attempt to develop applications of interest to the Company. Most of us
in the group were new to organometallic chemistry when we joined Joe. It was an
exciting time, being in at an early stage in a field that was taking off. There were
so many new things to try, the Group was growing rapidly in international
standing, and Joe was always flying offsomewhere to spread the message. There
were things that ICI could and should have profited from, but the Group was
seen within the Company as too ivory tower, too far removed from the profit
motive, and finding lucrative industrial applications was not Joe’s forte or
motivation. The Company should have added to the group personnel experi-
enced in commercial development. In the end they just ran out of patience, and
we ran out of time. But it was good while it lasted!
Joseph Chatt FRS: Some
Memories of his Work at The
Akers Research Laboratories of
Imperial Chemical Industries
Ltd, The Frythe, during the
1950s
L. A. DUNCANSON
When given the opportunity to contribute to this publication commemorating

the life and work of Joseph Chatt, three of his many attributes came immediately
to my mind. First I remember his intense love of chemistry. Socially, Joe (as we
referred to him informally) was hardly one to display his heart on his sleeve but in
the research environment his passion for chemistry was an inspiration to all
those fortunate enough to have the opportunity to collaborate with him. Second-
ly, I recall his outstanding empathy for people coupled with his profound
understanding of human nature. Although his style of management was not
flamboyant, he had a strong commitment to the personal well-being of his staff
and to the development of their full potential as individuals. This, coupled with
the creative working environment he provided, fostered a very strong sense of
loyalty amongst the members of his research team whether they were prima
donna scientists, laboratory technicians, or that unsung heroine, his secretary of
many years, Inga Schmidt. Thirdly, I remember vividly Joe’s sheer common
sense and pragmatic approach to problem solving. It was always productive and
rewarding to discuss, and argue, about scientific or other issues with Joe, and his
counsel was always wise. In the 1950s it was not the fashion in industry to employ
management consultants, but with people like Joe around they were not needed.
26 Joseph Chatt FRS: Some Memories ofhis Work at The Akers Research Luboratories
In short, Joe was not only a nice person but, in his quiet way, an excellent
manager and a powerful scientific leader.
I was first made aware of Joe Chatt’s work by the late Professor Ron Nyholm,
one of his then sparring partners in the field of organometallic chemistry.
Knowing that I wanted to marry the daughter of one of his Australian cricketing
colleagues, but couldn’t afford to on a part-time demonstrator’s salary whilst
wanting to continue with research, he suggested that I applied for a job in
Imperial Chemical Industries at The Frythe. On arriving there I was asked to
become an infrared spectroscopist, in support of the organic and inorganic
chemistry being carried out in the laboratory. This turned out to be a most
rewarding opportunity, not so much financially but in terms of exciting chemis-
try.
At that time one of Chatt’s interests was the nature of bonding in olefin
coordination to platinum. He was sure that the high mobility of ligands in the
position trans to a coordinated olefin in square-planar platinum(I1) complexes
(the trans-effect) indicated a double bond between the olefin and the metal,
involving a a-bond between an electron pair of the olefin and a n-bond formed
by back donation of electrons from a filled d-orbital of the metal. The infrared
spectra of ethylene complexes were inconclusive regarding his original sugges-
tion that the olefin had rearranged to form a methylidene radical in order to
accept d-electrons into a vacant carbon orbital, producing a structure
(CH,-CH-Pt). However, propyleneplatinum(I1) chloride was synthesised
with some difficulty and found to have a strong absorption band in its infrared
spectrum at 1504 cm-’. This was easily assignable to a carbon-carbon double-
bond stretching vibration. Re-examination of spectra of the ethylene complexes
revealed only a very weak absorption band at this wavelength, indicating that
the olefin was symmetrically bonded to the metal. This and other evidence led to
the conclusion that the olefins were double-bonded to the platinum atom
through sharing of the olefin’s n-electrons to form a o-bond and back donation
of d-electrons into the vacant anti-bonding n-orbital of the olefin, similar to the
structure proposed by Dewar for olefin complexes of silver ions. In the plati-
num(I1) complexes the n-bond would be strengthened by hybridisation of a 5d
orbital with a vacant 6p orbital of the metal.
Some time later, Rob Guy, whose chemical skills were much respected by the
research team and who was also envied for his ability to charm ladies, managed
to synthesise a range of acetylene complexes of platinum. Their infrared spectra
indicated that bonding of the hydrocarbon to the metal was very similar to that
in olefin complexes. Interestingly, it was observed that a-hydroxyacetylenes are
chelated to platinum through interaction of oxygen lone-pair electrons with a
vacant 6p orbital of the metal.
After this we embarked upon a comprehensive study of the infrared spectra of
amine complexes of platinum and palladium. By we, I don’t just mean Joe, the
leader, and myself, but Alan Williams, George Gamlen, Bernard Shaw and last,
but far from least, Luigi Venanzi who had the gift of being able to synthesise any
compound you could think of which might have an interesting infrared spec-
trum. Incidentally I remember him not just for his chemical skills but as the most
L. A . Duncanson 27
outstandingly English Italian I have ever met. After he left The Frythe to become
a fellow of Magdalene College Oxford, I vividly remember meeting him walking
along The High with my very pretty baby daughter cradled in one arm and a
furled umbrella on the other, not just looking the part, but actually being the
perfect English gentleman.
The objective of the work we set out upon was to examine how the N H
stretching frequencies of coordinated amines could provide information about
the reactivity of transition metal complexes, with particular reference to the
trans-effect. To be brief, let me just say that we measured the frequencies and
absorption intensities of NH stretching modes in primary and secondary amine
complexes of platinum and palladium containing a wide range of ligands in the
position trans to the amines. These measurements led to the conclusion that the
trans-effect operates by two mechanisms. First, an inductive effect upon the
o-bonding of the amine lone pair to the metal, but secondly a conjugative
(mesomeric) effect involving the filled d-orbitals of the metal itself. These two
factors influenced in different ways the mechanisms of substitution reactions of
these coordination complexes.
We discovered also that there is inter-molecular hydrogen-bonding between
the coordinated nitrogen atom of the amine and the Pt-Cl bonds of other
molecules. We also observed a rather strange intermolecular hydrogen bond
between the amine hydrogen atoms and the d-electrons of the coordinated
platinum atom, whereby orbital following of the platinum d-electrons reduced
the transition moment and hence the absorption intensity of N H stretching
modes. In addition it was observed that conformational effects associated with
restricted rotation about the Pt-N bond strongly influenced both the inter- and
intra-molecular hydrogen bonding equilibria.
Although tempted, I will not bore the reader with more details of what still
excites me personally as a rewarding field of research. Nor will I burden the
typesetter with a lengthy list of references. The easy way to enrich and check the
veracity of the above brief summary is to search under the key-words Chatt,
Venanzi and infrared in the chemical literature of the 1950s and 1960s.
There followed many other exciting adventures involving studies of the in-
frared spectra of transition-metal complexes, but I will mention just one which I
find particularly memorable. This followed the discovery by Bernard Shaw of a
remarkably stable volatile platinum complex produced by reduction of
[PtCl,(PEt,),], the spectrum of which had a very strong and sharp absorption
band near 2200 cm -'.
This seemed only assignable to a Pt-H stretching vibration. Joe, being as
rigorous as ever, was not completely convinced until a similar strong, sharp,
absorption was found at 1600 cm-' in the spectrum of the corresponding
deuterium compound, precisely where expected from the mass difference be-
tween hydrogen and deuterium. The complex [PtClH(PEt,),] was the first in
which the stretching frequency of a metal-ligand bond was observable in the
spectral range available to us at the time. We used this to make direct measure-
ments of the bond strength of a coordinate link. It was found that the Pt-H
stretching frequencies of a range of hydride complexes with different anionic
28 Joseph Chatt FRS: Some Memories of his Work at The Akers Research Laboratories
ligands showed a strong correlation with the magnitudes of the anionic trans-
effects as established by Chernyaev in 1927.
Incidentally, on this question of why are we what we are, I remember being
told by a previous mentor, C. K. Ingold, that he took up chemistry because he
had been told at Imperial College in the early 1900s that Maxwell had already
finished physics! I would not want to enter into discussion about the relative
impacts of Chatt and Ingold on chemistry, but in my personal opinion based on
experience of both of them, they were level pegging in their research leadership
qualities. Be that as it may, I welcome this opportunity to remember and record
the strong bonds of affection and respect which I am certain all of my co-workers
at The Frythe still have for Joseph, as I know his wife Ethel always preferred us to
address him. I sincerely hope that, after all this time, she will forgive me for using
the affectionate diminutive that seemed so natural to us all.
SECTION B:
Recent Developments in the

Synthesis, Bonding Modes and
Reactivity of Hydrido and
Dihydrogen Complexes
Hydrido complexes of transition metals have been known since the early 193Os,
largely from the pioneering work of Hieber on the preparation of the hydridometal
carbonyls [FeH,(CO),], [MnH(CO),] and [CoH(CO),]. Subsequent work was
driven in part by the recognition that transition-metal hydride complexes are
important intermediates in a range of homogeneously catalysed reactions, such as
synthetically and industrially important hydrogenations, olejin isomerisations and
polymerisations, hydroformylations and hydrosilations.
Chatt’s discovery and characterisation of [PtCEH(PEt,),] and related com-
plexes, reportedjirst in 1957, and Wilkinson’s report of [ReHCp,] in 1955 repre-
sent further major advances in transition metal hydride chemistry. Such work
opened up t h e j e l d which has led to a greater understanding of the structures,
bonding, stereochemistry and reactivity of transition metal hydrides and of the
nature of the metal-hydrogen bond. The discovery of [PtClH(PEt,),] arose seren-
dipitously from attempts to obtain cyclobutadiene complexes of platinum(0). The
product, an air stable, monomeric and sublimable solid which analysed as
[PtCl(PEt,),], could also be isolated from aqueous solution from the reaction of
[PtCl,(PEt,),] with hydrazine. At the time, the oxidative and thermal stabilities of
[PtClH(PEt,),] (it can be distilled unchanged at 130°C/0.01 mmHg) were unique.
Typical thoroughness was shown in the structural and spectroscopic characterisa-
tion of these materials (as described above by Adams), including an early applica-
tion of nuclear magnetic resonance spectroscopy (carried out in C.E.H. Bawn’s
laboratory at the University of Liverpool), a technique which has played a critical
role in more recent developments. The availability of a large series of tractable
complexes enabled the role of the hydride ligand to befully characterised, including
the demonstration of its high trans-efect and ligandjeld strength. The demonstra-
tion of the reaction of [PtClH(PEt,),] with ethylene to give [Pt(C,H,)Cl(PEt,),],
30 Recent Developments
another milestone in organometallic chemistry,followed shortly. It was knownfrom
the early work of Hieber and others that the hydrogen in metal hydride complexes
can display both hydridic and protonic character. Crabtree summarises recent
developments on metal hydride complexes in which interactions between hydridic
and protonic hydrogens (‘dihydrogen bonding’ or ‘proton hydride interactions’) are
evident. Chaudret and Ito describe further developments in the chemistry of the
interaction of hydrogen and transition metals, including the synthesis, characterisa-
tion and reactivity of polyhydride, dihydrogen and silane complexes, particularly of
ruthenium and molybdenum.
Hydrides, Hydrogen Bonding and
Dihydrogen A ctivation
ROBERT H. CRABTREE
Yale Chemistry Laboratory, 225 Prospect St, New Haven, CT
06520-8107, USA
1 Introduction
Chattl" was one of the principal pioneers of hydride chemistry in the third
quarter of the 20th century, when the chemistry of terminal and bridging
hydrides was mainly developed. The area has expanded very greatly since the
mid-1980s. The most important advance was the recognition that molecular
hydrogen complexes (M-H,) are not only common but can also have an import-
ant role in the reactivity of molecular hydrogen in metal complexes. Chattlbwas
also a pioneer in bioinorganic chemistry, so it is of particular interest that M-H
and M-H, intermediates have been proposed2 as the key species in Ni-Fe
hydrogenases, the enzymes that convert H, to protons and electrons.
As a result of my being a doctoral student in his group during his Sussex
University period, our own subsequent work has been heavily influenced by the
Chatt legacy. Of our first two problems at Yale, our work on hydride complexes
started as a direct development of Chatt chemistry. It was the exceptionally high
formal oxidation states and coordination numbers in metal polyhydrides that
persuaded me that they must have some interesting physical or reactivity proper-
ties. As a testament to their hold on my attention, of the many fields we have
tackled, this is the only area that has been a focus of continuous work from my
very first paper,3 resulting from undergraduate work in Malcolm Green's lab.,
right up to the present. The other area, C-H activation, was something Chatt
held as an important goal for the future even if he had himself only carried out a
few early but influential experiments in the field.4
After completing our work on C-H activation in 1985,5 we participated in the
development of the chemistry of dihydrogen complexes6 by showing their gener-
ality and that they can be synthesised by protonation of compounds with a
terminal M-H bond (Equation l).7a
32 Hydrides, Hydrogen Bonding and Dihydrogen Activation
(Cy = cyclohexyl)
Such compounds can be hard to characterise by classical methods, so we
suggested7buse of the excess T , relaxation of the hydride signal in the 'H NMR
spectrum as a criterion for M-H, binding. This work7' has long been published
so in the present review we concentrate on developments in the hydride area
since 1994, with the rise of dihydrogen bonding. This term refers to the attractive
interaction between a protonic and a hydridic hydrogen in an M-H.-H-N or
M-H-H-0 group. These interactions are distinct from classical hydrogen
bonds,' A-H-B (AH = acid; B = base) in that the proton acceptor is an M-H
hydride instead of a lone pair of the base, B.
The first indications in this area date from 1990, when attractive A-H-H-M
interactions were first proposed to explain the close contact (H-eH, 2.4 A)
between the IrOH proton and the Ir-H hydrogen in a neutron diffraction study of
ci~-[1rH(OH)(PMe,),l.~ This H-H distance is rather long and about equal to the
sum of the van der Waals' radii for two H atoms and so the interaction may be
relatively weak. A truly short d(H-H) of 1.86 A was found by neutron diffrac-
tion'Oa in mer-[Fe(H),(H,)(PEt,Ph),], a study originally carried out to test our
earlier spectroscopic assignment" of this species as a dihydrogen complex. The
H, ligand was found in a canted orientation that Eisenstein recognised from an
extended Huckel analysis'" and from an ab initio c a l ~ u l a t i o nas~ a~ compromise
(2)between 1 (H-H Ito cis-Fe-H) that leads to the strongest back donation and
3 (H-H 11 to cis-Fe-H) which leads to the maximal attraction between the
protonic Fe-H, group and the hydridic Fe-H. She used the term 'cis-effect' for
this phenomenon, which can now alternatively be considered as dihydrogen
bonding because of the very short H-H distance (1.86 A).
P
I
P-tI4i-P
I
P
I
PA.-P
IH
p-l-p
P
H H H
1 2 3
From 1995, we and Morris' group found a long series of metal hydrides in
which a transition metal M-H bond acts as the weak base (proton acceptor) in a
hydrogen bond with O H or N H protons ( = A-H) as the weak acid partners.
This was shown by the short H-H distance (ca. 1.8 A) and by studies that
identified the A-H.-H-M interaction strength as ca. 4-8 kcal mol-'. Classical
hydrogen bonds,8aA-He-B, require a lone pair on the base, B, and both A and B
are electronegative. In A-H-H-M, the lone pair of the weak base, B, is apparent-
ly replaced by the M-H a-bond, and both M and H are much more electroposi-
tive than the N, 0 or F atoms common in the classical type, so we have a new
class of hydrogen bond. Morrissb has used an alternative descriptive term,
proton-hydride interaction.
The conformational preference usually seen in compounds having a dihydro-
Robert H . Crabtree 33
gen bonded He-H interaction is such that the N H or OH bond approaches the
M-H bond in a side-on direction (as in 4), although linear examples were also
found in cases where conformational or steric effects hinder a bent geometry.
Indeed, Eisenstein found in DFT calculations that the potential energy surface is
rather flat, so distortions from the ideal geometry are not very costly.I2
A A
/
4 4a
2 Intramolecular Interactions
The effect was easiest to study first in intramolecular cases because this allows
NMR to be used to best effect. Complex 5 proved to be a useful test bed so as to
get some idea of the energetics of the interaction.
Q N,H
H-p-H
Q,I H
Q HI
5 (Q = PPh,)
Our NMR method for estimating the H-H bond energy involves looking at
the C-NH, rotation barrier by variable temperature NMR in species such as 5.
In the transition state for 5 5’, shown in Equation 2, the Ha-H bond is broken
+
and the energy for the delocalisation of the N lone pair into the pyridine ring is
lost. We estimated the intrinsic C-N rotation barrier in the absence of dihydro-
gen bonding using a combination of experimental data and Hartree-Fock calcu-
lations. By measuring the barrier and subtracting our estimate of the delocalisa-
tion energy, we arrive at a reasonable estimate of the H.-H bond energy: 5.0 kcal
mol-’ for 5.13
5 rotation 5’
transition state
One of the trans hydrides in 5 can readily be replaced by any of a variety of

anions, Y, to give 6 (Equation 3). Unlike 5 and 5’, 6 and 6’ are not equivalent:
Table 1 Some H...H bond strengths for 6
Y He-H bond strength (kcal mol-l)
~~~
H- 5.0
co 3.7
CN- 3.4
I- 3.3
MeCN 3.1
Br- 3.0
c1- 2.9
F- < 2.9
data taken from ref. 13
while 6 has an H-H hydrogen bond, 6' has a classical Ha-Y hydrogen bond.
Apart from the bond strength method used for 5, estimates of the relative
H--H/H-Y hydrogen bond strengths were also possible in this case from the
ratio of 6 to 6 . These two species can be distinguished by NMR spectroscopy at
- 80°C in the equilibrium of Equation 3. The C-N rotation barriers and the
resulting Ha-H bond strengths in 6 were very strongly dependent on the nature of
the trans ligand, Y, indicating the presence of a substantial trans effect on the
H-.H interaction. Where Y is H-, the Ha-H bond energy was highest (5.0 kcal
mol-I). When Y becomes more electron withdrawing, the He-H interaction
energy falls until, for Y = F-, the energy is < 2.9 kcal mol-' (Table 1).Since H
ligands trans to high trans effect ligands tend to be particularly hydridic,l4 this
implies that a basic hydride is best for dihydrogen bonding, in accord with the
electrostatic bonding model mentioned above. Epstein, Berke and c o - w ~ r k e r s ' ~
have recently used the trans effect of a nitrosyl trans to H to encourage particu-
larly strong dihydrogen bonding in an intermolecular case.
The presence of the other isomer, 6', allowed the N-H-Y hydrogen bond
strengths to be determined and, even for Y = F-, this proved to be only just a
little more (5.2 kcal mol-') than for the N-H-H-Ir bond where Y = H- (5.0
kcal mol-I). The system was designed to have a conformation that is most
favourable for formation of an N-He-H-Ir dihydrogen bond, however, so there
is probably some size mismatch for the larger Y groups. This may be largely
compensated by the ability of the pyridine ring to rotate about the Ir-N bond
and so move out of coplanarity with the H-Ir-Y group and allow hydrogen
bonding even for large Y groups. Fluoride being very similar in size to hydride,
however, a valid comparison is probably possible in this case.
I,Q
Y-lr-H (3)
H
Q' I
H
6' 6
From the decrease in hydrogen-bonding energies' on moving from the classi-
cal lone pair H-bond, N-H-(lone pair) (4-8 kcal mol-I), to the N-H-T case in
which the proton acceptor is usually an arene x system, < 2 kcal mol-I, one
might expect that any N-H-o type, where the acceptor is a o bond, would have a
negligible bond energy < 1 kcal rno1-I. In contrast, we find N-H-H-E interac-
tion energies of 4-8 kcal mol-' which are almost as large as for the N-H-(lone
pair) case. This requires E to be an electropositive element such as B or a
transition metal, so that the hydrogen has significant hydridic character, and
even then we usually also need a high trans effect ligand trans to the H in
question. It is still not entirely clear why the dihydrogen bonding energies in
species such as 5 are quite as large as we find, however. The H-H distance of 1.8
A is essentially the same as the He-B distance in the classical hydrogen bond, so
an unusually close approach of donor and acceptor atoms is not the critical
factor.
We suggested fast T I relaxation as a criterion for the presence of close He-H
distances in dihydrogen complexes, and Morris et ~ 1 . ' and~ our own group"
have detected a substantial excess relaxation for He-H bonding. Making the
usual assumptions, the excess T , in cases such as 5 and 6 can be interpreted in
terms of an H.-H distance of about 1.8 A in all the cases studied, a value
consistent with the structural data in related systems.
3 The Nature of A-H-H-E Hydrogen Bonding

In the first A-He-H-M hydrogen bonds found by us" and by Morris,8dthe weak
acid AH was an acidic NH or OH group and M was a d6 transition metal (e.g. 5,
6). Since a d6 metal such as iridium(II1)has d, nonbonding electrons, these could
in principle interact with the A-H proton (Figure 1). The reason the AH proton is
always close to the M H hydride could be nothing to do with a proton-hydride
interaction. Instead, the AH might in fact interact with these d, nonbonding
electrons adjacent to the M-H bond simply because H is the sterically smallest
ligand present and allows the N H to approach the metal most closely. In that
case, the A-H-H-E bond is really no different from the classical A-He-B
Figure 1 Since the d6 metals involved in the M--H-.H-A interaction have d, electrons, the
hydrogen bond might have been of the classical A-H-flone pair) type. The work
discussed here suggests it is best described as an interaction between the A H
proton and the H-M bond
hydrogen bond and the true interaction is between AH and a nonbonded d,
electron pair on M (4a). On this idea, the H-H part of the interaction would be
repulsive.
A study of BH,-NH, and its derivatives was useful to resolve this question
because neither B nor N has nonbonding electrons. This approach suggested
itself because a sample of this compound happened to be located in a prominent
place in the laboratory. The striking feature of BH,.NH, is that it is a solid,
unlike its isoelectronic analogues, such as C,H,. Indeed, its melting point of
+ 104°C is almost 300 degrees higher than that of ethane (m.p., - 183°C).Such a
large m.p. elevation is similar to that found for H,O and CH,, a classic case in
which the difference is ascribed to hydrogen bonding, albeit of the classical
variety.
In this study, the Cambridge Crystallographic Database (CSD) provided data
on intermolecular N-H-mH-B hydrogen bonds15ain a series of organic amine-
boranes. The nature of the interaction proved essentially identical structurally in
the transition metal and main group examples and so the d, nonbonding
electrons play no more than a minor role. The H--H distances in both main
group and transition metal cases range from 1.7-2.2 A,which should be com-
pared with 2.4 the sum of the van der Waals' radii for two hydrogens.
At first we only found data for the organic derivatives. The unsubstituted
example, BH,.NH,, not being an organic compound, was in the inorganic
database. According to the reported coordinates, it had an entirely different
configuration from that shown in 4. This species had been examined by X-ray
crystallography on several prior occasions but the B-H-.aH-N configuration
found was the reverse of the one in 4: B-Ha-H appeared to be almost linear and
N-H-eH appeared strongly bent, In collaboration' 5 b with Klooster and Koetzle
of Brookhaven National Labs., we were able to look at BH,.NH, by neutron
diffraction (Figure 2). By comparison of the neutron and X-ray work it was clear
that the B and N had previously been misassigned as N and B, respectively; the
true assignments produced a normal BHHN configuration as expected. The B/N
assignment is definitive by neutron diffraction because the neutron scattering
diameters are so different for the two nuclei.
A DFT calculation on the BH,.NH, d i ~ e r shows ' ~ ~ a conformation for the
BH-HN group that is very similar to that of 4. The H-.H bond energy was
calculated to be 6.6 kcal rno1-I per bond, comparable to that seen for transition
metal dihydrogen bonding. The calculated charge distribution suggested that the
BH bonds are polarised on forming the H--H interaction, which may help
explain the relatively high interaction energy, and that the boron is more
negatively charged than the hydride. The latter is consistent with the presence of
a significant B--H+ component in the interaction, or alternatively, one could
view the entire B-H bond as being the true proton acceptor. The H-H bonding
is not entirely electrostatic, however, because 'J(H-.-H+) coupling in the range
2-7 Hz is seen in the 'H NMR spectrum of compounds such as 5 and 6. Similar
small couplings between the A-H proton and the base have recently been seen
for classical A-H-aB hydrogen bonds.'
Other theoretical work has proved valuable. Calhorda et al.I5" found a large
Figure 2 The neutron diflraction structure of BH,.NH, showing intermolecular

dihydrogen bonding
Reproduced from ref: 15b with the permission of the American Chemical Society
number of complexes in the CSD database with short MH-H(O,N) distances

and carried out a DFT study for [(IrH(OH)(PH,),]PFG that concluded that the
counter-ion must be included to obtain good agreement with experiment.
Lledos, Eisenstein et a1.17 have written an excellent review on theoretical
methods applied to hydride complexes.
4 Intermolecular Interactions
In 5 and 6, the NH bond is held in a rigid chelate conformation and this could
affect the conformation of the A-Ha-H-M substructure, so the metric par-
ameters for such systems might be artefacts. In addition, the measured interac-
tion energies for 5 and 6 might be affected by the rigid geometry. In an extreme
revisionist interpretation, the whole dihydrogen bonding phenomenon might be
considered an artefact.
To resolve this problem, we therefore studied intermolecular interactions such
as are found when an acid HA co-crystallises with a metal hydride (M). In this
38 H ydrides, Hydrogen Bonding and Dihydrogen Activation
case, each moiety can find its most appropriate orientation in the co-crystal.
However, in practice, HA and M most often crystallise separately, however. To
favour co-crystallisation, we chose an acid, indole (7), that is a liquid and
therefore cannot crystallise, and a base, [ReH,(PPh,),] (8) that forms poor
quality powders rather than good quality crystals on attempted crystallisation
on its own; this gave satisfactory results. An X-ray structure by Rheingold
confirmed the co-crystal formulation and suggested the H-H distance was short
(< 2 A). In one crystallisation attempt, large, very high quality crystals were
obtained for the adduct between 7 and 8. These allowed Koetzle and Albinati to
obtain a high quality neutron diffraction structure.', This showed essentially the
conformation 4 previously seen and confirmed that this was not a result of the
constraints of chelation. The He-H distance of 1.73 A in this structure remains
the smallest to have been reliably determined. The value is much smaller than the
sum of the van der Waals' radii for two hydrogens (2.4 A).''
Energetics were also estimated for the intermolecular case. Approximate
values were first found via IR spectroscopy with the modified Iogansen equa-
tion," relating the low energy shift of the v(NH) or v(0H) band in the IR
spectrum to the interaction energy. Applied', to [ReH,(PPh,), - indole] (3.6
kcal mol-I), and [ReH,(PPh,),.ArOH] (5.6-5.8 kcal mol-'), the results seem
reasonable (Table 2). They agree quite well with UV-vis data from full equilib-
rium studies,,' which give a AG of 5 kcal mol-' binding energy for
[ReH,(PPh,),(C,H,N)] and indole.
Epstein, Berke et aZ.14 have used the Iogansen method to obtain intermolecu-
lar association energies of ca. 5.5 kcal mol-' between acidic alcohols such as
(CF,),CHOH and the hydridic hydride, [WH(CO),(NO)(PMe,),]. Equilibrium
constants for the same systems gave an interaction energy of 4.9 kcal mol-'.
5 Reactivity
Using the information discussed above, protonation of metal hydrides can now
be considered as going via the following pathway (Equation 4).
A
H
I
In one case, Chaudret and Sabo-Etienne, l a have seen an equilibrium between

[RuH,(dppm),] [dppm = bis(dipheny1phosphino)methanel and a dihydrogen
complex formed as a result of proton transfer from an alcohol such as
(CF,),CHOH (Equation 5). Other related reactions have been reported., l b
Robert H . Crubtree 39
6 Intramolecular Effects of Pendant Groups
Enzymes often catalyse reactions because the hydrogen bonding and other
reactive groups that surround the active site stabilise the transition state for the
reaction. We felt that this biomimetic approach might usefully be extended to
organometallic chemistry and we have now added pendant reactive groups to a
cyclometallated benzoquinolate (bq) ligand in the hope of seeing binding and
reactivity effects including ones resulting from hydrogen bonding. A variety of
groups can easily be introduced at the 2-position of the benzoquinolinate.
Cyclometallation of the bq gives species of type 9, in which the amino group is
adjacent to the site trans to the bq carbon, where a variety of ligands L readily
bind. The rigid geometry prohibits the group from binding to the metal - it can
only interact with the ligand L bound to the metal at the cis binding site.
H
9
The labile aqua complex22a9 (L = H 2 0 )is a very useful precursor in this area,
but it was hard to characterise fully, even with an X-ray crystal structure, because
the hydrogen bonding pattern remained ambiguous: out of 9a-c, the experimen-
tal data led us to prefer 9a, but only marginally.
H
9a 9b 9c
The DFT (B3PW91) calculations of Clot and Eisenstein22b predict that 9a

should be preferred. Subsequent to the DFT work, cooling of 9 to - 80°C led to
Table 2 Some dihydrogen bond strengths (kcal mol-I), deduced from Av(NH) and
A v ( 0 H ) I R spectroscopic data,for intermolecular adducts of some d o and
d 2 complexes with typical proton donors, indicating that direct X-H-M
hydrogen bonding is not predominant; data takenfrom ref: 12b
indole 3.6 3.3

2,4,6-Me3C,H,0H 5.6 4.7
d" configuration d2 do
adppe = 1,2-bis(diphenylphosphino)ethane
the water peak in the proton NMR spectra being resolved into a 1: 1 pattern,
consistent with structure 9a.
The bqNH, system can also stabilise an H F complex - the first of its kind - by
hydrogen bonding (ll).22 Protonation of the neutral fluoride (10) at - 80°C in
CD,Cl, gave the new complex shown in Equation 6. The presence of a J(H,F) of
440 Hz in the NMR spectrum at - 80°C is only consistent with the presence of a
hydrogen-bonded H-F ligand. The J(H,F) in the parent 10 is a mere 52 Hz. The
J(F,P) and J(F,H) couplings seen in 11 but with reduced values relative to the
parent fluoride, 10, indicate that the H F is still bound to the metal in 11.
H H
10 11
The H F compound was too unstable to survive to room temperature or to be
crystallised, so we were not able to obtain an experimental structure. The DFT
(B3PW91) calculations22bpredict the structure (Figure 3) and that the H F binds
to the complex in a bidentate fashion involving a coordinate and a hydrogen
bond of about equal strengths and with a total binding energy of 28.2 kcal mol-',
a much larger value than when the NH, pendant group is replaced by H (18.0
kcal mol-') when no H F complex is detectable experimentally. The H F distance
of 1.042 A in 11 is only slightly elongated from free H F (0.922 A)but the IrF
distance elongates significantly from 2.123 A in the fluoride 10 to 2.262 A in the
H F complex, 11. The N-HF distance of 1.43 A in 11 indicates it contains a very
strong H-bond.
n 0
Figure 3 The calculated structure for the HF complex 11from DFT(B3P W91).
Adapted from ref 22b. P H , ligands not included
7 Heterolytic H-H Activation

In order to identify the pendant group effects more securely, we have compared
the behavior of the bqNH, complexes with analogues that lack a pendant group,
that is, with the bqH derivatives. For example, in Equation 7, H, displaces
water from the precursor to give a molecular hydrogen complex that is
deprotonated by external base. In Equation 8, the corresponding situation with
bqNH, and Q = PPh, yields the hydride 12b where the coordinated H, has
been deprotonated by the pendant amino group.,,
Clot and Eisenstein carried out DFT calculations on a model system with
Q = PH, which predicted that 12a should be stabler than 12b, in contrast with
experiment. To try to reconcile theory with experiment, we replaced the PPh,
with the more basic phosphine PBu",. The dihydrogen complex, 12a, now
proved to be the stable isomer. The same species 12a was also seen for
Q = PMePh, and PMe,Ph. When the theoretical model PH, was replaced by
weaker donors PFH,, PF,H and PF, in the calculations, 12b became the stabler
isomer, in line with experiment (Q = PPh,).
Since the Q = PCy, case shows splitting of the H,, factors other than basicity
seem to be at work here and we are continuing our studies on the problem to
resolve the situation. The advantage of the joint theory-experiment approach is
that in the absence of the theoretical result, we would not have gone beyond the
PPh, case and would not have seen the M-H, isomer.
8 Consequences of H-H Bond Formation

Many metal hydrides protonate to give H, c o r n p l e x e ~but
, ~ ~kinetic protonation
can take place on M-H to give an M-(H,) complex, even when protonation at
the metal is thermodynamically favoured. Protonation of [FeH(dppe)Cp*]
(Cp* = pentamethylcyclopentadienyl) gave the dihydrogen complex at - 80 "C,
followed by rearrangement to the dihydride at 25 0C.25Kinetic protonation by
A-H at M-H is consistent with the presence of a dihydrogen-bonded

A-H-H-M precursor adduct as intermediate. Proton transfer in the adduct
gives the H, complex and conversion to the trunsl6 dihydride is slower because
motion26 of the heavy atoms is needed.
Gatling and Jackson27have shown how dihydrogen bonding to an OH group
of a hydroxyketone can direct the attack of borohydride to one face of the
molecule; indeed, the product is 99.7% trans (Equation 9). The effect was
suppressed by addition of F-, a species that disrupts the hydrogen bonding.
Under the conditions used, formation of an intermediate borate ester can be
excluded.
99.7% trans
The role of dihydrogen bonding in stabilising the transition states for solid
state reactions involving loss of H,, such as in the thermal decomposition of
triethanolamine/LiBH, or of BH,.NH, has been emphasised very recently both
in experimental28 and t h e ~ r e t i c a work.
l~~
9 Conclusion
A combination of computational and experimental approaches gives an under-
standing of the proton-hydride interaction. This new type of hydrogen bond, the
dihydrogen bond, is shown to influence the physical properties and reactivities of
a number of main group and transition metal compounds. It seems to be
important in cases of protonation of hydrides by acids HA.
10 Acknowledgements
I thank the NSF for funding and Odile Eisenstein, Arnie Rheingold, Tom
Koetzle, Eduardo Peris, Jesse Lee, and Per Siegbahn, for their insights into the
problems described here.
11 References and Notes

a) J. Chatt, Science, 1968, 160, 723; b) J. Chatt, J. R. Dilworth and R. L. Richards,
Chem. Rev., 1978,78, 589.
M. Pavlov, P. E. M. Siegbahn, M. R. A. Blomberg and R. H. Crabtree, J . Am. Chem.
Soc., 1998,120, 548; R. H. Crabtree, Inorg. Chim. Acta, 1986,125, L7.
R. H. Crabtree, A. R. Dias, M. L. H. Green and P. J. Knowles, J . Chem. SOC.A , 1971,
1350.
J. Chatt and J. M. Davidson, J . Chem. Soc., 1965,843.
5 R. H. Crabtree, Chem. Rev., 1985,85,245.
6 G. J. Kubas, Acc. Chem. Res., 1988,21,120.
7 a) R. H. Crabtree, M. Lavin and L. Bonneviot, J . Am. Chem. SOC.,1986,108,4032; b) X.
L. Luo and R. H. Crabtree, Inorg. Chem., 1990, 29, 2788; c) R. H. Crabtree, Angew.
Chem. Int. Ed. Engl., 1993,32,789.
8 a) G. A. Jeffrey and W. Saenger, Hydrogen Bonding in Biological Structures, Springer,
Berlin, 1994; b) R. H. Morris, Can. J . Chem., 1996,74,1907; c) R. H. Crabtree, P. E. M.
Siegbahn, 0.Eisenstein, T. F. Koetzle and A. L. Rheingold, Ace. Chem. Res., 1996,29,
348; d) W. Xu, A.J. Lough and R. H. Morris, Inorg. Chem., 1996,35, 1549.
9 R. C. Stevens, R. Bau, D. Milstein, 0. Blum and T. F. Koetzle, J . Chem. SOC., Dalton
Trans., 1990,1429.
10 a) L. S. Van der Sluys, J. Eckert, 0.Eisenstein, J. H. Hall, J. C. Huffman, S. A. Jackson,
T. F. Koetzle, G. J. Kubas, P. J. Vergamini and K. G. Caulton, J . Am. Chem. SOC.,1990,
112,4831; b) J.-F. Riehl, M. Pelissier and 0. Eisenstein, Inorg. Chem., 1992,31, 3344.
11 R. H. Crabtree and D. G. Hamilton, J . Am. Chem. SOC.,1986,108,3124.
12 a) J. Wessel, J. C. Lee, Jr., E Peris, G. P. A. Yap, J. B. Fortin, J. S. Ricci, G. Sini,
A. Albinati, T. F. Koetzle, 0.Eisenstein, A. L. Rheingold and R. H. Crabtree, Angew.
Int. Ed. Engl., 1995,34,2507; b) E. Peris, J. Wessel, B. P. Patel and R. H. Crabtree, J .
Chem. SOC.,Chem. Commun., 1995,2175.
13 E. Peris, J. C. Lee, Jr., J. R. Rambo, 0. Eisenstein and R. H. Crabtree, J . Am. Chem.
SOC.,1995,117,3485; J. C. Lee, Jr., E. Peris, A. L. Rheingold and R. H. Crabtree, J . Am.
Chem. SOC., 1994,116,11014.
14 E. S. Shubina, N. V. Belkova, A. N. Krylov, E. V. Vorontsov, L. M. Epstein, D. G.
Gusev, M. Niedermann and H. Berke, J . Am. Chem. SOC.,1996,118,1105.
15 a) T. B. Richardson, S. deGala, R. H. Crabtree, and P. E. M. Siegbahn, J . Am. Chem.
SOC.,1995, 117, 12875; b) W. T. Klooster, T. F. Koetzle, P. E. M. Siegbahn, T. B.
Richardson and R. H. Crabtree, J . Am. Chem. SOC.,1999,121, 6337; c) D. Braga, F.
Grepioni, E. Tedesco, M. J. Calhorda and P. E. M. Lopes, New J . Chem., 1999,23,219.
16 H. Benedict, I. G. Shenderovich, 0. L. Malkina, V. G. Malkin, G. S. Denisov, N. S.
Golubev and H. H. Limbach, J . Am. Chem. SOC., 2000,122,1979.
17 F. Maseras, A. Lledos, E. Clot and 0.Eisenstein, Chem. Rev., 2000,100,601.
18 A. Bondi J . Phys. Chem., 1964,68,441.
19 A. V. Iogansen, G. A. Kurchi, V. M. Furman, V. P. Glazunov and S. E. Odinokov, Zh.
Prikl. Spectrosk., 1980,33,460; S. G. Kazarian, P. A. Hamley and M. Poliakoff, J . Am.
Chem. SOC.,1993,115,9069.
20 P. Desmurs, K. Kavallieratos, W. Yao and R. H. Crabtree, New J . Chem., 1999, 23,
1111.
21 a) J. A. Ayllon, C. Gervaux, S. Sabo-Etienne and B. Chaudret, Organometallics, 1997,
16, 2000; b) A. J. Toner, S. Grundemann, E. Clot, H. H. Limbach, B. Donnadieu, S.
Sabo-Etienne and B. Chaudret, J . Am. Chem. SOC.,2000,122,6777; J. A. Ayllon, S. F.
Sayers, S. Sabo-Etienne, B. Donnadieu, B. Chaudret and E. Clot, Organometallics,
1999,18,3981.
22 a) D.-H. Lee, H. J. Kwon, B. P. Patel, L. M. Liable-Sands, A. L. Rheingold and R. H.
Crabtree, Organometallics, 1999,18,1615;b) E. Clot, 0.Eisenstein and R. H. Crabtree,
New J . Chem., 2001,25,66.
23 a) D.-H. Lee, B. P. Patel, E. Clot, 0. Eisenstein and R. H. Crabtree, Chem. Comrnun.,
1999,297; b) A better comparison would involve the 2- and 4-substituted species but
we have not yet carried out these experiments.
24 P. G. Jessop and R. H. Morris, Coord. Chem. Rev., 1992,121,155.
25 P. Hamon, L. Toupet, J.-R. Hamon and C. Lapinte, Organometallics, 1992,11, 1429.
26 R. H. Crabtree, X.-L. Luo and D. Michos, Chemtracts Inorg. Chem., 1991,3,245.
27 S. C. Gatling and J. E. Jackson, J . Am. Chem. SOC.,
1999,121,8655.
28 R. Custelcean and J. E. Jackson, J . Am. Chem. SOC.,2000,122,5251.
29 S. A. Kulkarni, J . Phys. Chem. A, 1999,103,9330.
Hydrides and Dihydrogen
Ruthenium Complexes:
a Continuation of Joe Chatt's
Chemistry
SYLVIANE SABO-ETIENNE AND BRUNO CHAUDRET
Laboratoire de Chimie de Coordination CNRS, 205 Route de Narbonne,
F-3 1077 Toulouse-Cedex 4, France
1 Introduction
Joe Chatt has been a pioneer of modern organometallic chemistry involving
transition metal phosphine derivatives and in particular hydride complexes. In
this article, we will emphasise the link between the pioneering work of Joe Chatt
and the present interest in o-bond complexes and their reactivity. This article
concentrates on the chemistry of ruthenium derivatives and is divided into a brief
historical overview, giving a perspective of the evolution of the chemistry of
ruthenium hydride complexes and a part describing recent achievements in this
field, mainly from our research group.
2 Historical Aspects
The first transition metal hydride [FeH,(CO),] was discovered by Hieber in
1931,' but the organometallic chemistry of metal hydrides really started in the
mid-fifties with the synthesis of [ReHCp,] by Geoffrey Wilkinson et aL2 and of
[PtClH(PEt,),] by Joe Chatt et ~ 1 The . ruthenium
~ hydride phosphine chemis-
try, which is rich in a great number of catalysts for a variety of organic transform-
ations, started with a family of octahedral complexes of general formula
[RuClH(chel),] (chel = a chelating dihapto ligand such as dmpe [1,2-
bis(dimet hy1phosphino)et hane], depe [1,2-bis(diethy1phosphino)ethanel , diars
[1,2-bis(dimethylarsino)benzene], etc.) reported in 1961 by Chatt and H a ~ t e r . ~
These compounds have been used as starting materials for a variety of deriva-
46 Hydrides and Dihydrogen Ruthenium Complexes
tives among which are the hydrido alkyl and aryl complexes [RuRH(chel),] and
the dihydrides [R~H,(chel),].~The latter was the first polyhydride of ruthenium.
Among the compounds described in the early report of 1961 was the ruthe-
nium(0) complex, [Ru(dmpe),] able to undergo the tautomeric equilibrium
described in Scheme 1. This reaction is remarkable in several respects: (i) it is an
early example of a C-H activation reaction and led the DuPont group of Ittel,
Tolman and co-workers to study in detail C-H activation processes in the late
seventieq6 (ii) the reaction is far from being straightforward and implies an
approach of the two ruthenium centres and presumably the presence of an
intermediate in which C-H bonds of one ruthenium centre are weakly coor-
dinated to the other centre; (iii) it is also remarkable that the weak C-H
interactions present in the intermediate lead to a modification of the geometry at
ruthenium with the diphosphines now coordinated cis. This behaviour can also
be described as an early example of 'molecular recognition'.
Scheme 1
The development of homogeneous catalysis, following, for example, the dem-
onstration of the hydrogenation of olefins by [RhCl(PPh3)3],7 and the need for
available coordination sites has led to the synthesis of ruthenium complexes
accommodating different monophosphines, and in particular PPh,. The first
complex in the series was [RuC~,(PP~,),].~ This complex displays an interesting
early example of a weak C-H agostic interaction blocking a coordination site.
Such interactions are evidenced by X-ray crystallography but are usually diffi-
cult to characterise by other methods, such as NMR, since the coordination
energy of the C-H group is usually very low. However, recent NMR investiga-
tions have demonstrated the presence of such interactions in several cases,
including a ruthenium phenylpyridine complex described hereafter.
In order to investigate the mechanism of hydrogenation reactions, hydride
complexes were synthesised, including [RuCIH(PP~,),],~a fast and selective
catalyst for olefin hydrogenation. This was followed by the synthesis in 1970 by
the group of Yamamoto of the dihydride [RuH,(PPh,),] analogous to
[RuH,(chel),], but which, thanks to the easy dissociation of triphenylphosphine,
was found to be an extraordinary precursor for inorganic and organometallic
complexes as well as a catalyst for many organic transformations." However, the
presence of four triphenylphosphine ligands on [RuH2(PPh,),] was a problem in
some reactions. Other attempts to produce polyhydride derivatives led in the
early seventies to the synthesis of two other complexes, [RuH,(PPh,),] l1 and
[RuH,(N,)(PPh,),].12 The tetrahydride exhibited characteristic features of
polyhydrides such as [IrH,(PPh,),] or [ReH,(PPh,),], popular at that time.
However its high field 'H NMR spectrum showed only a broad signal and its
reactivity was similar to that of the dinitrogen compounds and resulting from an
Sylviane Sabo-Etienne and Bruno Chaudret 47
easy dissociation of one mole of dihydrogen from the complex. Similar observa-
tions on cationic ruthenium trihydrides led Singleton to propose that dihydrogen
could exist as such in the coordination sphere of ruthenium compound^.'^
The need for highly reactive complexes which could also be precursors of
transient species displaying a very low coordination number led us in 1982to the
synthesis of the 'hexahydride' [RuH,(PCy,),] (Cy = cy~lohexyl).'~ One of the
major goals was the generation of active species for C-H activation reactions.
The characterisation and reactivity of this compound and of related species
towards o-HX bonds (X = H, C, Si) will be briefly presented in the section below.
3 Ruthenium Hydride and Dihydrogen Complexes

During the early eighties, we developed a strategy for synthesising new hydride
derivatives, controlling precisely the quantity of phosphines added. This is based
on the hydrogenation of the zerovalent precursor [Ru(cod)(cot)] (cod = cyclo-
octa-1,5-diene; cot = cycloocta-1,3,5-triene) in the presence of the desired
amount of ligands. The complexes [RuH,(dppm),] (1) [dppm = bis(dipheny1-
phosphino)methane],' [RuH,(PCy ,),I (2),14a, '5,1 [RUHq(PR3)31(R = CY, 3%
Pri, 3b)15and [Ru,H,(PR,),] (R = Cy, 4a; Pr', 4b) were prepared in this way.I5
The demonstration by Kubas in 1984 of dihydrogen coordination without
dissociation in [M(CO),(H,)(PCy,),] (M = Mo or W)17 and the reinvestigation
of the structure of [RuH,(PPh,),] which led Crabtree to propose the formula-
tion [RuH2(H2)(PPh3),l1*prompted us to reconsider the structure of complexes
2-4 (see Scheme 2).
H
4
Scheme 2
3.1 [RuH,(dppm),], Dihydrogen Bonds and Proton Transfer
[RuH,(dppm),] is another example of bisfchelate) derivatives similar to those
originally reported by Chatt and co-workers. The compound exists as a mixture
of the cis- and trans-isomers, which are in equilibrium at room temperature in
s01ution.l~It was found in the eighties to be a good precursor for heterobimetal-
lic derivatives, the best examples being [R~RhH,(dppm),]'~ and [Ru-
MoH,(CO)4(dPPm),l.20
In contrast to other complexes of general formulation [RuH,(chel),] which
can be easily protonated to yield a series of trans hydrido dihydrogen derivatives,
[RuH(H,)(chel),] [RuH,(dppm),] does not produce a stable dihydrogen
complex upon protonation, but leads after dihydrogen evolution to the cationic
monohydride [RuH(S)(dppm),] (S = solvent). However, we have recently
+
studied the interaction of [RuH,(dppm),] with phenol in toluene solution and

demonstrated the presence of hydrogen bonding between one hydride and
phenol and, moreover, the presence of a dynamic proton transfer at low tempera-
ture ( < - 30°C):,,
trans-(dppm),HRuH * HOPh e [RuH(H,)(dppm),] '(0Ph)-
The reaction is reversible and upon allowing the solution to warm to room
temperature the dihydride is quantitatively recovered. This reaction is interest-
ing for several reasons: (i) it was the first NMR observation of such a dynamic
proton transfer; (ii) this method allows the observation of unstable species
otherwise not detectable; (iii) this reaction is specific since only the trans-isomer
reacts with phenol.
We have extended this method to another complex: [RuH,(PCy,)Cp*]
(Cp* = pentamethylcyclopentadienyl). This complex was the first derivative
reported to display 'quantum mechanical exchange couplings', a consequence of
the ortho-para transition of dihydrogen which is observable by NMR when the
barrier to exchange of two hydrogens is sufficiently low (ca. 10 kcal m ~ l - ' ) . , ~
Addition of an acidic alcohol to this complex induced hydrogen bonding be-
tween a hydride and the alcohol proton. This kind of interaction thus modifies
the exchange barriers of the hydrogens and therefore the magnitude of the
exchange couplings. We can in this way classify the strength of the hydrogen
bonding interaction of various alcohols, the strongest ones being observed with
the most acidic alcohols, namely hexafluoro-isopropanol and nonafluoro-tert-
butanol. In this case, proton transfer was observed at low temperature to give the
otherwise inaccessible complex [RuH,(H,)(PCy,)Cp*] .24 +
Complex (2) has long been the only thermally stable derivative containing two
dihydrogen ligands and one of the few complexes containing two coordinated
o-bonds. Its infrared spectrum displays two bands of strong intensity, presum-
ably because of the coupling between the Ru-H and Ru-H, stretches. The X-ray
crystal structure of the compound was solved recently and, although an ambi-
guity concerning the real space group of the molecule remains, the refinements in
both the symmetrical (Pl) and unsymmetrical (Pi) space groups are consistent
with the dihydride bis(dihydrogen) f~ rmu latio n.~ The presence of dihydrogen
ligands is confirmed by NMR (TI measurements) and by inelastic neutron
scattering (INS). The use of the latter technique for the characterisation of
dihydrogen complexes was developed by Eckert. It involves the measurement of
the ortho-para transition of dihydrogen and allows an estimation of the rotation
barrier of dihydrogen, which in the present case is close to 1 kcal mol-1.26
[RuH,(H,),(PCy,),] (2) has been found to be a very efficient starting material
for the preparation of a great variety of new ruthenium hydride derivatives, in
particular those accommodating o-bonds. l 6 We will not describe all the chemis-
try of this compound but concentrate on four examples: hydrogen transfer and
metallation reactions, reactions with halocarbons, reactions with proton donors
and reactions with silanes, and in each case related reactions.
3.3 Hydrogen Transfer and Hydrocarbon Activation

The coordinated dihydrogen molecules of 2 may be easily substituted sequen-
tially to give a variety of new hydride and dihydrogen derivatives. The dihyd-
rido(dihydrogen) complex [RuH,(H,)(PCy,),] can thus be obtained by substitu-
tion of H, by a phosphine or directly from [Ru(cod)(cot)]. This compound and
the analogous [RuH,(H,)(PPr',),] lose reversibly a phosphine in solution to
give a transient species (RuH,(H,)(PR,),} (R = Cy, Pri) able to catalyse rapid
H/D exchange between the deuterium of C,D, and the protons of the phosphine
ligands.' This behaviour suggests a high potential for C-H activation reactions,
confirmed by the reaction with ethylene. Thus, after substitution of H,, we can
observe in the presence of excess ethylene the sequential dehydrogenation
at room temperature of cyclohexyl groups present on 2 to give
[RuH(C,H,)(PCy,)(q3-c6H8Pcy2) (5).27In the presence of 3,3-dimethylbut- 1-
ene, it is possible to isolate successively [RuH,(PCy,)(q3-C6H,Pcy2)] which is
better described as a monohydride complex accommodating a stretched dihyd-
rogen ligand (see Figure 1)28aand [RUH(~~-C,H,PC~,)(~~-C,H~PC~~)] in
which both phosphine ligands have been dehydrogenated.28bInterestingly, 5 has
been found to be an excellent and selective catalyst for the dehydrogenative
silylation of ethylene.27
3.4 16-Electron Hydrido(dihydr0gen) Complexes: Reactions of

[RuH,(H,),(PCy,),] with Halocarbons and Related
Reactions
After substitution of H,, addition of CHJ to 2 produces the 16-electron dihydro-
gen complex [ R U H I ( H ~ ) ( P C ~ , ) ,Analogous
].~~ complexes of general formula-
tion [RuHX(H,)(PCy,),] (X = C1 or SR) have been obtained from the reaction
of 2 with dichloromethane and thiols. [RuHI(H,)(PCy,),] reacts with an excess
of dihydrogen to give the following equilibrium which can be detected by
NMR:,'
rRuHI(H,)(PCY,),l + H, CRuH~(H,),(PCY,),l
Further reaction with chloroform produces the ruthenium(1v) dihydride
C125
Figure 1 Molecular structure of [RuH(H,)(PCy3)(q3-C,HJCy,)]
[RuCl,H,(PCy,),] which exists as a mixture of isomers and in equilibrium with

the ruthenium(r1) dihydrogen derivative [RuCI,(H,)(PCy,),]. Interestingly,
Caulton has recently reported that the prolonged reaction of 2 with CH,C1,
produced the carbene complex [RuCI,(PC~,),(=CH,)],~ a member of Grubbs’
carbene family, and an excellent precursor for metathesis and ROMP reac-
tion~.~~
3.5 18-Electron Hydrido(dihydrogen) Complexes, Proton

Transfer and C-H Activation
Addition of proton donors to [RuH2(H2),(PCy3),] gives, after substitution of
the two dihydrogen ligands, new hydrido(dihydr0gen) derivatives, as is the case
with carboxylic acids:33
Similar reactions were carried out with various good o-donors which are
also proton donors such as hydroxo-, amino- or thio-pyridine or quinoline
(HE-L) thus producing a new series of complexes [RuH(H,)(E-L)(PCy,),] in
which, because of the high electron density induced by the presence of these
ligands, a significant vibrational mode of the H-H bond is observed in the IR
spectrum. In addition, an H-H distance of ca. 1.3 A was calculated for

[RuH(H2)(0C,H,N)(PCy,),l from the relaxation time TI of the hydride signal
and from the H-D coupling constant observed after partial deuteration of the
compound.34
Analogous reactions of 2 were carried out with phenyl derivatives, uiz.
acetophenone, benzophenone or phenylpyridine, to give, after activation
of a C-H bond, complexes containing a metal-carbon bond. A series
of compounds of general formula [RuH(H2)(Pcy,),(c6H,R)] (R = COMe, 6;
COPh, 7; C,H,N, 8) were prepared in this way (see Scheme 3). The chemistry
of these compounds is very rich but that of the complexes of nitrogen
donors (8, and also the analogous triisopropylphosphine complex
[RuH(H2)(PrPi,),(C6H4CsH4N)](9)and the corresponding benzoquinoline de-
rivative) is very different from that of the oxygen donors.35
.. PCY3
a 2 6 , R = Me; 7 , R = Ph
0
+ CHz=CH2 - 2 or 6
20 "C
Scheme 3
Dihydrogen can be eliminated from 9 to give the 16-electron species
[RuH(PPr',),(C6H4C,H4N)] or substituted by various small molecules to give
[RuH(L)(PPri,),(C6H,C,H,N) (L = N,, O,, C O or C2H4). Ethylene can easily
be removed from the last complex but no sign of insertion into the Ru-H or
Ru-C bonds is observed.
This is in marked contrast with the reactivity of 6 and 7 which catalyse the
insertion of ethylene into the ortho C-H bond of the phenyl ring of acetophenone
and benzophenone (see Scheme 3). This reaction has been previously described
by Murai using various ruthenium catalysts, the most efficient being
[RuH,(CO)(PP~,),.]~~ It is derived directly from the early metallation reactions
of Joe Chatt and displays a high potential in organic synthesis. Moreover, it
avoids the production of salts which are the by-products of similar alkylation or
arylation reactions such as Heck or Suzuki couplings. However, Murai's reac-
tion operates at 130°C probably because of the need for the complex to eliminate
one PPh, and one C O ligand for the reaction to proceed. Our system operates at
room temperature but the drawback is the formation of the inactive bis(meta1-
lated) complex [Ru(PC~,),(C,H,COR),].~ 5 b
[RuH(H2)(PPri,),(C6H4c5H4N) can be protonated to give a new complex,
[~uH(H,)(~-C,H,C,H,N)(PPr',),] -+ (lo), accommodating two o-bonds: a
dihydrogen molecule and an aromatic C-H bond (Figures 2 and 3).37 The
agostic phenyl ring exhibits a hindered rotation process on the NMR time-scale
Figure 2 Molecular structure of [RuH(H,)(PPri3),(C,H4C5H4N)]

(9)
with a barrier of activation, E, = 35.6 f 1.8 kJ mol-'; AGI = 42.0 & 5.6 kJ
mol-' at -30°C. This barrier represents the maximum value of the bonding
energy of the C-H bond to ruthenium. Interestingly, the corresponding carbonyl
complex [IhH(CO)(PPr'3),(fl-C,H,C,H,N)] shows a similar process with a
+
higher rotation barrier : E , = 44.5 1.8 kJ mol-l. Since all other electronic and
steric factors of the ligands are strictly the same, this result demonstrates the
better n-accepting properties of C O which allow a stronger coordination of the
agostic C-H bond.
10 undergoes a reversible proton transfer process in thf to give the metallated
dihydrogen derivative [Ru(H,)(thf)(PPri,),(C,H4C5H4N)] according to the
+
following equation:
t hf
[lh.d3(H2)(PPr13)2(f?
-C,H,C,H,N)] + [Ru(H,)(thf)(PPr',),(C,H,C,H,N)] '
H2
This process occurs through the transient formation of a cationic bis(dihydr0-
gen) complex which rapidly loses H,. This is a facile C-H activation reaction, the
mechanism of which may involve either a classical oxidative addition or a direct
proton transfer from carbon to hydride, a process analogous to a-bond meta-
thesis. However, DFT calculations by Eric Clot strongly suggest the occurrence
of a ruthenium@) derivative in the transition state and therefore of a classical
oxidative addition process.37
Sylviane Saho-Etienne and Bruno Chaudret 53
.21
Figure 3 Molecular structure of [kuH(H2)(H-C,H,C,H,N)(PPr',),] + (10)
3.6 Silane Complexes and Substitution Reactions of

IIRuH2(H2 ) 2(PCY3 )21
As observed earlier, the dihydrogen ligands of 2 are very labile and can be
reversibly substituted by weakly coordinating ligands, dihydrogen evolution to
the gas phase being the driving force of the reaction. For example, this is possible
with N,; the complex [RuH,(N,)~(PCY~)~] can be isolated and the mixed
[RuH2(H2)(N2)(PCy3),]may be observed.38 N o evidence for hydrocarbon ad-
ducts has been obtained but the reaction with bulky silanes leads cleanly to the
substitution of one dihydrogen molecule to give complexes accommodating both
a H-H and a Si-H a-bond, such as that with triphenylsilane [RuH2-
(H,)(HSiPh,)(PCy,),] (11) (see Scheme 4). This complex has been characterised
by different techniques including X-ray crystallography at low t e m p e r a t ~ r e . ~ ~
The most unexpected aspect of this structure is the bending of the P-Ru-P angle.
It is close to 180" in 2 and decreases to 109.71(5)"in 11. This is associated with
additional weak interactions between the terminal hydrides and silicon, one
'non-bonding' Si . . * H distance being very short: 1.83(3) A.The molecule was
modelled by Barthelat and co-workers3' who clearly evidenced the presence of
two additional weak Si-H interactions in the complex which stabilise the bent
form at the expense of the expected trans configuration. According to the level
of calculation, the bent form is more stable by ca. 8-17 kJ mol-'. A
54 H ydrides and Dihydrogen Ruthenium Complexes
13 11
Scheme 4
similar complex is obtained when a germanium hydride is used, viz.
[RuH,(H,)(HGePh,)(PCy,),] (l2).,* Whereas all the hydrides of 11 exchange
rapidly on the NMR time-scale, even at low temperature (lOO°C),a decoales-
cence is observed in the case of 12 between the germanium hydride, the two
hydrides and the dihydrogen molecule.
Using a disilane HSiR,XSiR,H, both dihydrogen molecules of 2 may be
substituted to yield the corresponding bis(si1ane) complexes accomodating two
a-Si-H bonds, [kuH,(fi-SiR,XSiR,-~)(PCy,),] (R = Me, X = 0, C2H4,
C,H,, C6H4 or OSiMe,O; R = Ph, X = 0)(13) (see Scheme 4 (X = OSiMe,O)
and Figure 4).40 Like 11, the new complexes 13 display a bent geometry with a
P-Ru-P angle of 104-108" due to the presence of four additional non-bonding
interactions between the terminal hydrides and the silicon atoms of the disilane
ligands. The chelating effect of the disilane ligand is only a minor factor in the
stabilisation of the corresponding bis(si1ane) complex. These complexes are
fluxional and rapid exchange between terminal and silicon hydrides may be
observed by NMR, the coalescence temperature depending upon the ligand.
These processes presumably involve a ruthenium-dihydrogen intermediate.
All these complexes are catalytically active for hydrosilylation and dehyd-
rogenative silylation of ethylene. High turnovers and high selectivity for vinyl
derivatives may be obtained using both mono- and bis-silanes as reagent^.,^.^'
However, perhaps the most interesting complexes have been prepared through
the reaction of 2 with silanes (see Scheme 4). Thus, a redistribution reaction
1, I
Figure 4 Molecular structure of [RuH,(H -SiMe,OSiMe,OSiMe,-H )(PCy,),] (13)
6 C31
Figure 5 Molecular structure of [(RuH,(PCy,),),(p-SiH,)I (14)

is observed at room temperature which leads to 11 and to the remarkable

complex [(RUH,(PC~,),}~(~-S~H~)] (14)(see Figure 5).,, The SiH, molecule lies
between two ruthenium dihydride units. The Ru-Si distances are the shortest
reported (2.1875(4)A)but no direct Ru-Si bond is present. The silane molecule is
attached to ruthenium by four a-bonds and, uniquely, the back-bonding occurs
into a a* orbital of Si-H bond linked to the other ruthenium. The ease of this
reaction and the rich reactivity of 14 will lead us to study in detail further
transformations of silanes in the coordination sphere of ruthenium.
Conclusion
Using ruthenium hydride chemistry to provide milestones, it is possible to
appreciate the innovative character of Joe Chatt’s research and to understand
the evolution of the field for the past 45 years. In the early days, it was important
to demonstrate the possibility for various ligands to coordinate to transition
metals. Some of these ligands introduced by Joe Chatt have become the proto-
types of ancillary ligands; this is the case for phosphines, for example. Other
ligands have proved to be extremely useful tools in organometallic chemistry.
This is particularly true for hydrides, the chemistry of which is also associated
with the name of Joe Chatt. The first purpose of the early days was the demon-
stration of the existence of various types of complex. This purpose is not
completely outdated since, for example, the quest for an isolable alkane complex
is still a target of several research groups. However, with the demonstration by
Kubas of the possibility for a-bonds to give stable complexes, we have probably
now reached the point of having defined and isolated all potential coordination
modes between a transition metal and various potential ligands. Hence, elec-
tronic doublets or electronic vacancies as well as bonds, whether o or n, and all
combinations of these elementary modes of bonding may be involved in ligand
coordination to transition metals. It is also evident that the famous Chatt, Dewar
and Duncanson model used for understanding ethylene coordination to a transi-
tion metal is also usable for understanding o-bond coordination. The SiH,
complex discussed above represents a new version of this model in a bimetal
system.42To these modes of binding we must add the closed sphere interactions,
famous for gold@)(aurophilic interactions) which have recently led to the isola-
tion of a gold-xenon complex.43
These results bring us to the main difference between the present research and
that of the late fifties. The main objective of present research is to find new species
displaying a selective reactivity, for example for the functionalisation of hydro-
carbons, whether unsaturated (hydroboration, hydrosilylation, selective and
enantioselective hydrogenation of various functions, polymerisation of various
substrates including stereospecific polymerisation of polypropylene, etc.) or
saturated. Weak interactions appear to play an increasingly important role in
this field, whether hydrogen bonding, interaction of a-bonds (‘agostic’ bonds),
hydride silicon interactions, etc. These weak interactions will eventually lead to
the tuning of the selectivity in catalytic reactions by modelling the coordination
sphere of the complexes. This is presently a very active area of research.
Syluiane Sabo-Etienne and Bruno Chaudret 57
5 Acknowledgements
The authors are grateful to CNRS for support of their research and warmly
thank all their co-workers and colleagues for their contribution to the work
presented in this chapter.
6 References
1 W. Hieber and F. Leutert, Naturwissenschuften, 1931,19,360.
2 G. Wilkinson and J. M. Birmingham, J . Am. Chem. Soc., 1955,77,3421.
3 J. Chatt, L. A. Duncanson and B. L. Shaw, Chem. Znd. (London),1958,859.
4 J. Chatt and R. G. Hayter, J . Chem. Soc., 1961,2605.
5 J. Chatt and J. M. Davidson, J . Chem. Soc., 1965,843.
6 C. A. Tolman, S. D. Ittel, A. D. English and J. P. Jesson, J . Am. Chem. Soc., 1979,101,
1742 and refs therein.
7 J. A. Osborn, F. H. Jardine, J. F. Young and G. Wilkinson, J . Chem. Soc. A , 1966,1711.
8 T. A. Stephenson and G. Wilkinson, J . Znorg. Nucl. Chem., 1966,28,945.
9 P. S. Hallman, B. R. McGarvey and G. Wilkinson, J . Chem. Sue. A, 1968,3143.
10 a) A. Yamamoto, S. Kitazume and S. Ikeda, J . Am. Chem. Soc., 1968,90,1089; b) T. Ito,
S. Kitazume, A. Yamamoto and S. Ikeda, J . Am. Chern. Soc., 1970,92,3011.
11 R. 0. Harris, N. K. Hota, L. Sadavoy and J. M. C. Yuen, J . Organomet. Chem., 1973,
54, 259.
12 W. H. Knoth, J . Am. Chem. Soc., 1968,90,7172.
13 T. V. Ashworth and E. Singleton, J . Chem. Soc. Chem. Commun., 1976,705.
14 a) B. Chaudret, G. Commenges and R. Poilblanc, J . Chern. Soc. Chem. Commun., 1982,
1388; b) B. Chaudret, G. Commenges and R. Poilblanc, J . Chem. Soc. Dalton Trans.,
1984,1635.
15 a) B. Chaudret and R. Poilblanc, Organometallics, 1985, 4, 1722; b) T. Burrow, S.
Sabo-Etienne and B. Chaudret, Znorg. Chem., 1995,34,2470.
16 S. Sabo-Etienne and B. Chaudret, Coord. Chem. Reu., 1998,178-180,381.
17 a) G. J. Kubas, R. R. Ryan, B. I. Swanson, P. J. Vergamini and H. J. Wasserman, J . Am.
Chem. Soc., 1984,106,451; b) G. J. Kubas, Acc. Chem. Res., 1988,21,120.
18 R. H. Crabtree, Angew. Chem. Znt. Ed. Engl., 1993,32,789.
19 B. Delavaux, B. Chaudret, N. J. Taylor, S. Arabi and R. Poilblanc, J . Chem. Soc. Chem.
Commun., 1985,805.
20 B. Chaudret, F. Dahan and S. Sabo, Organometallics, 1985,4, 1490.
21 M. T. Bautista, E. P. Cappellani, S. D. Drouin, R. H. Morris, C. T. Schweitzer, A. Sella
and J. Zubkowski, J . Am. Chem. Soc., 1991,113,4876.
22 J. A. Ayllon, C. Gervaux, S. Sabo-Etienne and B. Chaudret, Organometallics, 1997,16,
2000.
23 S. Sabo-Etienne and B. Chaudret, Chem. Rev., 1998,98,2077.
24 J. A. Ayllon, S. Sabo-Etienne, B. Chaudret, S. Ulrich and H.-H. Limbach, Znorg. Chim.
Acta, 1997,259, 1.
25 A. F. Borowski, B. Donnadieu, S. Sabo-Etienne and B. Chaudret, Chem. Commun.,
2000,543; ibid, 1697.
26 V. Rodriguez, S. Sabo-Etienne, B. Chaudret, J. Thoburn, S. Ulrich, H.-H. Limbach, J.
Eckert, J.-C. Barthelat, K. Hussein and C. J. Marsden, Znorg. Chem. 1998,37,3475.
27 M. L. Christ, S. Sabo-Etienne and B. Chaudret, Organometallics, 1995,14, 1082.
28 a) A. F. Borowski, S. Sabo-Etienne, B. Donnadieu and B. Chaudret, to be published;
b) A. F. Borowski, S. Sabo-Etienne, M. L. Christ, B. Donnadieu and B. Chaudret.
Organometallics 1996,15, 1427.
29 B. Chaudret, G. Chung, 0. Eisenstein, S. A. Jackson, F. J. Lahoz and J. A. Lopez, J .
Am. Chem. SOC.,1991,113,2314.
30 M. L. Christ, S. Sabo-Etienne and B. Chaudret, Organometallics 1994,13,3800.
31 M. Olivan and K. G. Caulton, Inorg. Chem. 1999,38,566.
32 P. Schwab, R. H. Grubbs and J. W. Ziller, J . Am. Chem. Soc., 1996,118,100.
33 a) M. L. Christ, S. Sabo-Etienne, G. Chung and B. Chaudret, Inorg. Chem. 1994,33,
5316; b) G. Chung, T. Arliguie and B. Chaudret, New J . Chem., 1992,16,369.
34 Y. Guari, S. Sabo-Etienne and B. Chaudret, Organometallics 1996,15,3471.
35 a) Y. Guari, S. Sabo-Etienne and B. Chaudret, J . Am. Chem. SOC.1998, 120, 4228;
b) Y. Guari, A. Castellanos, S. Sabo-Etienne and B. Chaudret, to be published.
36 a) S. Murai, F. Kakiuchi, S. Sekine, Y. Tanaka, A. Kamatani, M. Sonoda and N.
Chatani, Nature, 1993, 366, 529; b) N. Chatani, T. Fukuyama, F. Kakiuchi and S.
Murai, 1.Am. Chern. SOC.,1996,118,493.
37 A. J. Toner, S. Grundemann, E. Clot, H.-H. Limbach, B. Donnadieu, S. Sabo-Etienne
and B. Chaudret, J . Am. Chem. SOC., 2000,122,6777.
38 S. Sabo-Etienne, M. Hernandez, G. Chung, B. Chaudret and A. Castel, New J . Chem.,
1994,18, 175.
39 K. Hussein, C. J. Marsden, J.-C. Barthelat, V. Rodriguez, S. Conejero, S. Sabo-
Etienne, B. Donnadieu and B. Chaudret, Chem. Commun., 1999,1315.
40 a) F. Delpech, S. Sabo-Etienne, B. Chaudret and J.-C. Daran, J . Am. Chem. Soc. 1997,
119, 3167; b) F. Delpech, S. Sabo-Etienne, J.-C. Daran, B. Chaudret, K. Hussein, C. J.
Marsden and J.-C. Barthelat, J . Am. Chem. SOC.,1999,121,6668.
41 F. Delpech, J. Mansas, H. Leuser, S. Sabo-Etienne and B. Chaudret, Organometallics,
2000,19, 5750.
42 I. Atheaux, B. Donnadieu, V. Rodriguez, S. Sabo-Etienne, B. Chaudret, K. Hussein
and J.-C. Barthelat, J . Am. Chem. SOC.,2000,122,5664.
43 S. Seidel and K. Seppelt, Science, 2000,290, 117.
Hydrido Complexes of Group 6
Transition Metals - Formation of
the Pentadentate Ligand with a
P-P-Si-P-P Framework
TAKASHI I T 0
Department of Materials Chemistry, Graduate School of Engineering,
Yokohama National University, 79-5 Tokiwadai, Hodogaya-ku, Yokohama
240-8501, Japan
1 Introduction
One day at the end of June, 1970, I knocked at the door of Professor Joe Chatt’s
office at the Unit of Nitrogen Fixation in the University of Sussex and met him
for the first time. This was the very day I started research on the transition metal
hydride complexes. I cannot forget that Professor Chatt was so kind that he
drove me in his own car to find accommodation for my family, my wife and two
little daughters.
The research theme I undertook as a postdoctoral fellow during the two years
under the warm and severe supervision of Professor Chatt spread over a wide
range of transition metal complexes, studies of which were performed in collab-
oration mainly with Drs Jeff Leigh and Jon Dilworth.
On coming back to Professor Akio Yamamoto’s group in 1972, I started a
research programme on molybdenum chemistry, which has lasted until now.
My research programme on the Group 6 transition metal hydrides mainly
concerns two kinds of starting complexes, one being of the type [MH,Cp,]
originally prepared by Wilkinson, Green, et al.,’ and the other of the type
[MH,(dppe),] (1) [dppe = 172-bis(diphenylphosphino)ethane]which was pre-
pared for the first time by Pennella in 1971., In this article, our recent research
results relating to the latter hydrido complex will be described, focusing mainly
on its reactions with primary and secondary silanes.
The tetrahydrido complex 1 has shown versatile reactivities, some of which are
60 Hydrido Complexes of Group 6 Transition Metals
typical of polyhydrido complexes, especially due to a high reactivity of the
coordinatively unsaturated intermediate species, (MHJdppe),} or (Mo-
(dppe),}, generated on either light irradiation or heating at 110°C in solution.
Examples hitherto reported by us are shown in Scheme l.,
H /A2 H
I I
Scheme 1
2 Reactions of [MH,(dppe),] with Primary and

Secondary Silanes
A typical method of formation of a transition metal-silicon bond is the oxidative
addition of silane derivatives, involving either Si-H or Si-Si bond cleavage, to a
low valent, coordinatively unsaturated transition metal complex. Various kinds
of transition metal silyl or silylene complexes have so far been synthesised via this
method. However, in contrast to the rich chemistry of the late transition metal
complexes with an M-Si bond, a limited number of syntheses of Group 6 metal
complexes with an M-Si bond have been r e p ~ r t e d . ~
In expectation of obtaining oxidative addition products of the general formula
H-(Mo}-SiR,, which are readily envisaged from the series of reactions reported,
we examined the thermal reactions of tetrahydride 1 with silanes. On reaction of
1 with phenylsilane PhSiH, in refluxing toluene, an oxidative addition reaction
Takashi Ito 61
involving Si-H bond cleavage to give an Mo-Si bond did in fact occur, although
the final product isolated was a little more complicated than that expected from
simple oxidative addition.'
Thus, heating a toluene solution of 1 under reflux in the presence of more than
two equivalents of phenylsilane for 3 h afforded a yellow solid in 78% yield,
which was characterised as 2 (shown in Scheme 2) spectroscopically as well as by
an X-ray crystallographic study (Figure 1).The high-field region of the 'H-NMR
spectrum of 2 in C6D6 at room temperature showed two resonances with
integration 1 : 1; a broad multiplet at 6 - 4.5 and a broad, apparent triplet at 6
-5.4 ppm. The results indicate that there are two sets of magnetically in-
equivalent protons that may be assignable to two hydrido ligands. Complex 2
represents the first example of the simultaneous activation of an Si-H bond and
two aromatic C-H bonds giving rise to the formation of a pentadentate ligand
which contains a P-P-Si-P-P framework.
Later, the detailed study of this reaction revealed that reaction of 1 with only
one equivalent of PhSiH, under similar conditions afforded a greenish yellow
solid, which was assignable as the trihydride 3a in 87% yield. Complex 2 was also
derived from trihydride 3a by its reaction with an excess of PhSiH, in refluxing
toluene. When 0-or p-tolyl silanes were used in place of phenylsilane, trihydride
complexes 3b and 3c, respectively, were obtained, although the disilyl complexes
corresponding to 2 were found to be too unstable to be isolated (Scheme 2).
When secondary silanes such as Ph(Me)SiH, or Ph,SiH, were employed in a
similar reaction with 1, the trihydrido complex 4, instead of 3, with a tridentate
ligand P-P-Si was isolated, and was structurally characterised by X-ray analysis
(Scheme 3).6
A possible reaction path from 1 to 3 is shown in Scheme 4: PhSiH, may
Figure 1 Molecular structure of 2

@SiH3 toluene
toluene
reflux reflux (> 2 equiv.)
(1 equiv.)
‘ R ’
Ph H
=
S
J(-i3
(1 equiv.)
-
toluene
reflux
Q 2 yellow
Scheme 2 Thermal reactions of1 with primary arylsilanes
+
2 Q-Si(R)H2
R = Me (a), Ph (b) 4 yellow
Scheme 3 Thermal reactions of 1 with secondary arylsilanes
oxidatively add to the 16-electron reactive intermediate A, generated thermally

from 1 on release of one mole of H,, to give a phenylsilyl-molybdenum inter-
mediate B. Since the direct substitution of the ortho hydrogens of phenyl groups
in the dppe ligand with Si in the coordinated SiH,Ph ligand seems to be less
likely, intervention of a silylene intermediate such as C in the Scheme, formed via
an a-hydrogen elimination of the silyl group, seems to be more plausible.
Activation of ortho C-H bonds in the phenyl groups of dppe ligands with the
Mo=Si bond may result in the formation of the tridentate ligand complex
corresponding to 4 which is isolated when secondary silanes are employed. In the
case of a primary silane, further a-hydrogen elimination from 4 followed by the
insertion of an Si=Mo bond into the ortho C-H bond may lead to the formation
of the pentadentate ligand complex 3. The alternative path involving oxidative
Takashi Ito 63
Scheme 4 Possible reaction path from 1 to 3
addition of the ortho C-H bonds of dppe, reductive elimination of Mo-C and
Mo-Si bonds, and oxidative addition of an Si-H bond (the path involving E
through G in Scheme 4) without intervention of the silylene intermediate,
however, cannot be ruled out.'
3 Some Reactions Involving Trihydrido Complex

[MoH,{[Ph,PCH,CH,P(Ph)C,H,-o],(Ar)Si-
P,P,P,P,Si)l(3)
The trihydrido complex 3 was found to undergo versatile reactions, generally
maintaining the unique pentadentate ligand intact, as summarised in Scheme 5.
As is described above, 3a reacts with a further mole of phenylsilane to give 2.
Solutions of complexes 3 in thf were found to be very susceptible to air and
changed from yellow to greenish in the presence of even trace amounts of air.
When dioxygen was bubbled through a thf solution of 3a at room temperature, a
similar colour change was immediately observed. The work up of the solution
afforded green crystals which were analysed as a peroxo complex of the type
[MoH{[Ph,PCH,CH,P(Ph)C6H~-o]2(Ph)S~-P,P,~,~,~~}(~2-O2)] (5) on the
basis of spectral and X-ray structural analyses.6 The analogous dioxygen com-
plexes corresponding to 3a were obtained similarly for tolyl derivatives 3b and
3c. Evolution of a significant amount of H, was detected by GLC when the
reaction of 3 with 0, was conducted in a sealed system.6 The presence of the
P-P-Si-P-P type of pentadentate ligand in 5 seems to be crucial in stabilising
the peroxo-type dioxygen complex as suggested by the two observations: (1)
treatment of the parent complex [MoH,(dppe),] (1) with 0, in solution resulted
only in decomposition of the complex; (2) the trihydrido complexes with a
P-P-Si tridentate ligand (complexes 4 in Scheme 3) did not give any dioxygen
complex similar to 5 on their treatment with 0, but decomposed.
Carboxylic acids such as formic, acetic, and benzoic acids reacted with 3a to
give corresponding carboxylato complexes 6 in which the carboxylato ligand
RoQoR R = OEt, OPi, OPf, OCH2Ph

8
Scheme 5 Some reactions of trihydrido complex 3a
coordinates to the metal in a unidentate mode through one oxygen atom as

shown spectroscopically as well as by the single crystal X-ray a n a l y ~ i sIf. ~one
considers that most of the related carboxylatomolybdenum complexes so far
reported possess the bidentate-type carboxylato ligand, the rigid double chelate
framework consisting of Si, four phenyl carbons and two P atoms in 6 seems to
have hindered the coordination of the second oxygen atom of the carboxylato
ligand. The same unidentate formato complex 6 (R = H) was obtained by the
reaction of 3 with gaseous carbon dioxide in a toluene solution (Scheme 5).8
On heating a thf solution of 3a or 3b at 70°C in the presence of two moles of
CH,I, the reddish orange iodo complex 7 was obtained, accompanied by evol-
ution of methane (Scheme 5). The similarity between 6 and 7 in the pattern and
the chemical shift of the 'H NMR signal assignable to the hydrido ligand
strongly suggests that the latter has a similar molecular arrangement to the
former, the structure of which was confirmed by X-ray structure analysis.
The reaction of 3a with malonates in refluxing toluene, as shown in Scheme 5,
Takashi Ito 65
afforded the malonato complex 8 which has a unique unidentate malonato
ligand bonded via one enolic oxygen atom to the metal, with the pentadentate
P-P-Si-P-P framework being kept intact. The highly stable nature of the
framework seems to have hindered the chelation of the malonato ligand.
In contrast, when pentan-2,4-dione was allowed to react with 3a under similar
conditions, chelated pentan-2,4-dionato complex 9 was obtained in which one of
four phosphorus atoms of the framework is dissociated.’ This result of the
unique reaction was confirmed by the X-ray structure analysis of the complex 9.
Although the explanation of these results obtained for two types of P-dicarbonyl
compounds, malonate and pentan-2,4-dione7is not straightforward at present,
the difference in the steric congestion between two dicarbonyl compounds seems
to be responsible.
Since the trihydrido complex of the type 3 with a unique pentadentate ligand is
interesting in view not only of its formation itself but also of its reactivity, we are
now exploring the related reactions utilising the tungsten analogue of 1 as well as
the primary and secondary germanes.
4 References
1 M. L. H. Green, J. A. McCleverty, L. Pratt and G. Wilkinson, J . Chem. SOC., 1961,
4854.
2 F. Pennella, Chem. Commun., 1971, 158.
3 T. Ito, Bull, Chem. SOC.Jpn., 1999,72, 2365.
4 X.-L. Luo, G. J. Kubas, J. C. Bryan, C. J. Burns and C. J. Unkefer, J . Am. Chem. SOC.,
1994,116,10312; X.-L. Luo, G. J. Kubas, C. J. Burns, J. C. Bryan and C. J. Unkefer,
1995,117,1159.
5 D.-Y. Zhou, M. Minato, T. Ito, and M. Yamasaki, Chem. Lett., 1997,1017.
6 D.-Y. Zhou, L.-B. Zhang, M. Minato, T. Ito, and K. Osakada, Chem. Lett., 1998,187.
7 T. Ito, D.-Y. Zhou, M. Minato, Y. Yamaguchi, and K. Osakada, unpublished results.
8 T. Ito, D.-Y. Zhou, K. Sumiura, M. Minato, and Y. Yamaguchi, unpublished results.
9 T. Ito, K. Shimada, T. Ono, M. Minato, and Y. Yamaguchi, J . Organomet. Chem.,
2000,611,308.
SECTION C:
The Chemistry of Phosphines
When Chatt first became involved in phosphine chemistry it was a minority sport.
Apparently the interest in such compounds was then based upon the property of
solubility in organic solvents that they conferred upon their adducts with metal
halides. This opened up the possibility of determining parachors, a function that, it
was hoped, would shed some light on the nature of the co-ordinate bond.
That this never transpired is one of the ironies of chemical history. What
phosphine complexes have enabled us to do has been much more signijicant. Chatt’s
early work concentrated on platinum group metals, and this was consonant with the
interest in complexes containing hydrides, olefins, alkyl groups, or aryl groups.
When Chatt left The Frythe, the work had just about reached the iron group, with
minor excursions beyond.
I t was natural that the Unit of Nitrogen Fixation should continue to exploit
phosphine complexes, and this was richly rewarded. Apart from opening up osmium
and rhenium phosphine chemistry, work also advanced to Group 6. Progress further
to the left in the Periodic Table is much more dificult if one insists upon using
tertiary phosphines, but even so Group 6 gave series of dinitrogen complexes to add
to the series obtained with osmium and rhenium. In fact some chemists became so
blask that when a manuscript describing only the second extensive series ofdinitro-
gen-phosphine complexes, based upon rhenium, was submitted to Chemical Com-
munications, a referee recommended rejection, saying that this was just another
series of dinitrogen complexes. Chatt was no less angry than the rest of his
collaborators.
It is possible that our concentration upon the use of phosphines has unbalanced
current perceptions about the coordination chemistry of dinitrogen. However,
phosphines have opened so many new vistas in inorganic and organometallic
chemistry that, be it due to luck or to exceptional insight, the debt ofthe chemical
community to Chatt, and also to pioneers such as F . G. Mann, is difJlcult to
over-estimate.
This section contains three contributions: that of Mingos describes some of the
historical background as well as developing some recent ideas about cone angles;
that of McAuliffe describes some unexpected compounds stabilised by phosphines
68 The Chemistry of Phosphines
which do not, at first sight, obey the normal rules elucidated for them; and that of
Heaton deals with new materials based substantially upon phosphines. The story of
phosphine coordination chemistry still has a long way to run.
Some New Insights into the
Steric Eflects of Tertiary
Phosphine Ligands via Data
Mining
D. MICHAEL P. MINGOS
Principal’s Lodgings, St Edmund Hall, University of Oxford,
Oxford OX1 4AR, UK
1 Introduction
Joseph Chatt made many valuable contributions to coordination and or-
ganometallic chemistry, but perhaps his most significant and widely applicable
legacy resulted from his promotion of tertiary phosphines as flexible ligands in
transition metal chemistry.’ He demonstrated very clearly that their presence
enabled chemists to develop reactions at the metal centre which were unique and
resulted in the ligation at the metal of many unusual and novel organic ligands
and fragments. His PhD training with F. G. Mann at Cambridge’ had introduc-
ed him to the distinct advantages of using triethylphosphine as a ligand in the
preparation of air-stable and crystalline samples of platinum metal complexes.
He recognised that having complexes which were soluble in a range of organic
solvents opened up reactions which could only be undertaken in non-protic
solvents. Of course he also learned the skills associated with making and safely
handling such ligands. Indeed his postgraduate research had resulted in the first
synthesis of bis(dimethy1arsino)benzene (diars) - a ligand which was destined to
become the workhorse of the extensive research efforts of the Nyholm group
during the 1950s and 1 9 6 0 ~ ~ 3 ~
With the benefit of hindsight, it is instructive to look back at those features of
phosphine which excited Chatt’s enthusiasm and interest and also indicate how
he advanced the field so that their advantages became more widely appreciated
by other chemists all over the world. In the late 1930s the absence of spectro-
scopic techniques for characterising complexes and the extreme difficulty of
completing a single crystal X-ray crystallographic analysis5 because of the phase
70 Some New Insights into the Steric Eflects of Tertiary Phosphine Ligands
problem meant that it was essential to have a pure compound which could be
analysed by classical techniques. Indeed, the assignment of coordination geomet-
ries had been closely associated with the formation of distinct isomers which
could be separated by fractional crystallisation techniques. Therefore, Chatt and
Mann's research at that time represented an extension of the methodology which
had been so successfully developed by Werner at the turn of the century.
Werner's research had been based primarily on ammonia and amines coor-
dinated to substitutionally inert transition metal ions, and such complexes could
be made and recrystallised from aqueous solutions. Of course, Werner's classical
work on the separation of enantiomorphs of tris(ethy1enediamine)complexes of
cobalt(II1)required the introduction of organic substituents onto the nitrogen
ligands in such a way that the methodology of working in aqueous solutions
could be maintained. Chemists at this time had studied the complex formation of
tertiary amines, but discovered that they did not form complexes which were as
stable as the comparable complexes of ammonia. The first report of a tertiary
phosphine complex of a transition metal has been attributed to Hofmann in
1857.6 Mann came into the field in the 1930s and recognised that tertiary
phosphines and arsines formed particularly stable complexes and he extended
the range of known platinum and palladium complexe~.~
2 Historical Background
When he set up an independent research group at The Frythe, Chatt recognised
the following important qualities of tertiary phosphines as ligands:
1. They imparted organic solubility on the complexes which were formed from
them. The solubility in a particular solvent could also be modified by chang-
ing the length of the alkyl substituent and introducing phenyl substituents.
2. The resulting complexes would produce highly crystalline samples if the
appropriate phosphine were used. Of course, crystallising compounds was at
that time, and indeed still is, as much an art as a science and Chatt himself had
various intuitive feelings about the desirability of using specific phosphines.
When I worked with him he had a distinct preference for triethylphosphine
and diethylphenylphosphine as ligands which he thought formed highly
crystalline and therefore easily separable complexes. In contrast, he was not
at all in favour of using triphenylphosphine. The relative insolubility of
cis-[PtC12(PPh,),18 had led him to disfavour this ligand - an oversight
which perhaps allowed others to develop the important chemistries of the
following complexes: [Pt(PPh,),,,] (Malatesta, Allen);' [RhCl(PPh,),]
and [RuCl,(PPh,),] (Wilkinson);" [IrCl(CO)(PPh,),] (Vaska);' and
[CoH(PPh,),] (Yamamot 0, Sacco). 9
3. I think that the discovery that he could make complexes which were soluble in
non-polar and relatively unreactive solvents such as benzene (more widely
used in those days than now) and diethyl ether (tetrahydrofuran did not
become readily available until the late 1950s) proved to be a particularly
important observation, because it opened up the possibility of doing meta-
D.Michael P. Mingos 71
thetical reactions of chloro-complexes of the transition metals with reactive
organic reagents such as Grignards and lithium aluminium hydride. This was
a development of the types of reaction which had become familiar to him as a
PhD student when he was making tertiary phosphines and arsines.
4. The solubility in non-polar solvents was also important to Chatt because he
early recognised the value of making dipole moment measurements on com-
plexes in order to provide valuable stereochemical information and to gain
some insight into the polarities of the metal-ligand bonds. His subsequent
routine use of this technique built on the foundations which had been devel-
oped previously by Sutton and his group at O ~ f o r d . ' ~ ,Of ' ~ course, the
advent of first proton and then phosphorus NMR spectroscopy led coordina-
tion chemists, and particularly Bernard Shaw and his co-workers, to use
methyldiphenyl- and dimethylphenyl-phosphine more extensively. These
ligands have particularly simple and yet informative proton NMR spectra,
especially in the light of the virtual coupling phenomenon of trans phosphines
noted by Shaw and his co-workers.
5. Chatt appreciated early on that phosphine ligands are particularly effective in
forming stable complexes with the platinum metals. Indeed his determination
to put this characteristic on a more quantitative basis led to the important
Class (a) and Class (b) classification scheme of metal-ligand interactions
proposed jointly with Ahrland and Davies.' This work was subsequently
subsumed into Pearson's general 'Hard and Soft' acid and base classification
scheme.'
6. Chatt recognised that substituents on the phosphines could alter the steric
requirements of phosphine ligands, but it was not a central concern of his
research at that time. He was particularly interested in ligands which coor-
dinated strongly to transition metals and remained coordinated to the metal
when the metal was modified by introducing hydride, alkyl and aryl ligands
'
via metathetical reactions.' The ability of ligands such as triphenylphosphine
to generate empty and potentially reactive coordination sites was first in-
dicated by Malatesta's work on [Pt(PPh,),] which showed that it lost a
phosphine to give [Pt(PPh,),] on recrystallisation from benzene' and this
observation became a critical part of the mechanism proposed by Wilkinson
and Osborn for alkene hydrogenation catalysed by [RhCl(PPh,),].20 It was
also mechanistic studies by Tolman on the hydrocyanation catalyst
[Ni(PRJ4] (R = alkyl, aryl, alkoxy or aryloxy) which led him to quantify for
the first time the relative steric effects of phosphine and related ligands.2'
Therefore, it was Tolman rather than Chatt who was able to show how
steric effects associated with phosphine ligands could be put on a more
quantitative basis using the cone angle concept.22A mathematically defined
cone was projected from the metal atom to the surface of the ligand as defined
by the van der Waals' radii of the atoms on the surface of the ligand. I suspect
that Chatt, Wilkinson and Malatesta all probably had an appreciation of the
importance of steric effects, but in my opinion they lacked the mathematical
background to articulate it in a way which could be readily appreciated by
other coordination chemists.
7. In the early 1950s the relative stabilities of comparable tertiary amine and
phosphine adducts with Lewis acids were interpreted in terms of the availabi-
lity of 3d orbitals at phosphorus. This view no longer prevails,23 but it led
Chatt and Nyholm to propose that the retrodative bonding which had been
developed earlier for metal-alkene complexes could also be applied to phos-
phine ligands, with which the conventional Werner dative bond could be
augmented by back-donation from filled metal d orbitals into empty d or-
bitals on p h o s p h ~ r u sOf
.~~course, the relative strengths of the two compo-
nents could be modified by changing the substituents on the phosphorus.
Chatt recognised that on this basis PF, would be a superior n-acid ligand and
this led him to attempt to synthesise [Ni(PF3)4].2’ This work was taken up by
others who demonstrated that PF, was as good as C O at stabilising low
oxidation state transition metal compounds.26He also appreciated that PPh,
and P(OPh), may be able to stabilise low oxidation states of the platinum
metals. This led to the isolation of complexes such as [Pt(alk~ne)(PPh,),],~~
but Chatt did not explore this idea to any great extent until the mid-
1960s when it became apparent that the isolation of stable dinitrogen com-
plexes could only be achieved when low oxidation state complexes were
generated.
8. Chatt’s PhD work had led him to understand the importance of bidentate
phosphine and arsine ligands and he, together with Davidson and Wat-
s0n,28.29later used the ligand Me,PCH,CH,PMe, to develop some classical
organometallic chemistry which for example demonstrated how C-H bonds
may be activated by transition metals.
3 Discussion
The introduction and background presented above have underlined the import-
ant contributions which Chatt made to the development of phosphine-stabilised
transition metal chemistry and also suggest why it was Tolman rather than he
who quantified the relative steric contributions of phosphine ligands.
Tolman’s original proposal (see Figure 1) depended on the direct physical
measurement of cone angles from idealised space-filling CPK models.2 Subse-
quently other methodologies have been developed for calculating the steric
effects of ligands and the subject has been well documented in White and
Colville’s recent review.,’ These methods sought to rectify the basic disadvan-
tage of the Tolman method, which does not take into account the variation in
cone angle with ligand conformation. The cone angle concept has subsequently
been profitably extended to other important classes of ligands, notably amides
such as N(SiMe,), -,substituted cyclopentadienyl and arene ligands, substituted
polypyrazolylborates, substituted alkyls and aryls including mesityl and super-
A common theme of modern main group and transition metal chemis-
me~ i t yl . ~’
try has been the isolation of compounds with low coordination numbers and the
kinetic stabilisation of multiply-bonded compounds using sterically demanding
ligands.j2
D. Michael P . Mingos 73
radius
Figure 1 Illustration of the dejinition of the Tolman cone angle for a typical aryl
phosphine ligand
Since the Cambridge Crystallographic Database33has a wealth of structural

data on phosphine ligand complexes, Thomas Muller and 134 decided to deter-
mine whether it could provide a statistically based analysis of the variation of
cone angles from complex to complex. Just as Dunitz and Burgi used the results
of structural analyses to map out the reaction coordinates for many basic
reactions and polytopal rearrangement processes, so we felt that the calculation
of cone angles for all reliably determined structures of phosphine complexes in
the database could record how the cone angle changed in response to the
different environments in the whole range of complexes. O r ~ e had n ~previously
~
completed a Dunitz and Burgi-type analysis on phosphine ligands and showed
how the dihedral angles of the ligands varied in a concerted manner, but a similar
statistical analysis of cone angles had not previously been attempted.
The crystallographic data define the positions of the nuclei for heavy atoms
and the Tolman angle calculation requires a knowledge of the van der Waals’
surface of the ligand. The crystallographic coordinates and the Tolman cone
angle were inter-related by the geometric relationships illustrated in Figure 2.
To make the calculations consistent with those reported by Tolman, the
Figure 2 Calculation of the Tolman cone angle from X-ray crystallographic data
metal-phosphorus bond length was initially set constant at 2.28 A and a van der
Waals’ radius of 1.00 A was used for hydrogen. This leads to the calculation of
cone angles in individual structurally characterised molecules, rather than being
based on idealised molecular models or geometries derived from molecular
mechanics calculations. The calculations do introduce a small error since hydro-
gen atom positions are generally defined on the basis of calculated positions
which assume a C-H distance of 0.98 A.This distance gives a better description
of the electron density distribution in the C-H bond rather than the value of
1.08 A,which defines the internuclear distance. However, the small systematic
error introduced is more than compensated for by access to the thousands of
structures in the Cambridge Crystallographic Database. The following questions
become accessible by using such a statistical analysis.
1. Is there a direct correlation between these data and the original cone angles
proposed by Tolman?
2. Do the cone angles vary greatly from complex to complex?
3. Does this variation depend on the rotational freedom of the ligands?
4. When there is more than one phosphine ligand in a complex, do they share a
common cone angle? For a particular ligand, does a deviation from the
normal cone angle occur in those complexes where the phosphine ligands
intermesh and the phenyl ligands may occupy space which may also be
apportioned to the cone of an adjacent ligand?
5. Are there any systematic variations in cone angle either across or down the
periodic table?
Table 1 summarises the results of the cone angle calculations for more than
4000 phosphine-containing compounds based on the ligands PMe,Ph, -,, PEt,
and PCy, (Cy = cyclohexyl). The results of the statistical analyses may be
presented as bar charts, which may be idealised into normal distribution curves
(see Figure 3). Such an analysis provides a mean cone angle for each ligand and
also a standard deviation.
D . Michael P. Mingos 75
Table 1 Comparison of computed average cone angles for tertiary phosphine
ligands with their Tolman cone angles
PPh, PPh,Me PPhMe, P M e , PEt, PCy,
No. of structural 2388 185 547 1203 354 108
determinations of
ligands
Av. M-P-C (") 1 14.8(1.2) 1 1 5.2(1.3) 1 1 5.7(1.3) 1 1 6.8(1.5) 1 1 5.4( 1.5) 1 12.7(1.4)
Tolman cone angle 145 136 122 118 132 170
Crystallographic 148(5) 135(5) 120(5) 1 1 l(2) 137(5) 160(5)
cone angle (")
Range of cone 129-168 121-153 108-139 96-121 124-159 146-172
angles (")
PPh, PMePh2
Number of 500
ligand smctures
400
L
300
100
160 170
130 140 IS0
Calculated cone angle (")
PMe,Ph
ligand structures
40(
3oc
ZOC
IOC
Calculated cone angle
Figure 3 Histograms and idealised normal distribution functions for the cone angles of
some tertiary phosphine ligands
76 Some New Insights into the Steric EfSects of Tertiary Phosphine Ligands
It will be recalled that for an idealised Gaussian distribution three standard

deviations each side of the mean define a 99% probability of finding a compound
with a cone angle in that range. The statistically derived cone angles for these
phosphines generally correlate well with the Tolman cone angles. For example,
for PPh,, PMePh,, PMe,Ph and PEt, the differences are less than one standard
deviation. For PMe, and PCy, the difference is approximately two standard
deviations. Some have suggested previously that Tolman may have over-es-
timated the cone angle of PCy, and an alternative cone angle of 157" has been
proposed, which agrees better with our calculated value than the original Tol-
man cone angle.
The standard deviations of the cone angles provide an indication of the
variation of the cone angles from complex to complex and therefore reflect the
ability of a ligand to change its conformation in order to optimise the packing of
the ligands within the molecule and the crystal. In general, the phosphine ligands
have a standard deviation of 5", suggesting a spread of cone angles of 30". In
contrast the standard deviation for PMe, is only 2.4", reflecting the limited
degrees of rotational freedom available to this ligand. Consequently the spread
of cone angles for this ligand is only 15".
Even within one complex the cone angles can vary significantly. For example,
for complexes containing three triphenylphosphine ligands, the average cone
angle is 150", close to the global mean cone angle for this phosphine of 148".
However, in some of these complexes the cone angles of the PPh, ligands differ
by as much as 16", approximately three standard deviations. In such complexes,
two of the three triphenylphosphines have cone angles close to the global mean
and the third deviates greatly from it. In these sterically crowded molecules the
restricted degrees of freedom available to the ligands lead one of the phosphines
to adopt a conformation with a much larger cone angle. These differences
underline the limitations of using a cone to define the space occupied by ligands
which have a propeller-like structure.
The standard deviations associated with the cone angles therefore provide an
indication of the ability of the phosphines to adapt to different environments.
For example, trimethylphosphine, which has limited scope for intermeshing with
other ligands because of the absence of large degrees of freedom associated with
the methyl groups and the absence of the flat phenyl rings, has the smallest
standard deviation. Therefore this phosphine has well defined steric require-
ments. In contrast the other phosphines studied have a much wider spread of
cone angles, and this is reflected in the larger standard deviations.
There are groups of complexes for which the cone angle of the triphenylphos-
phine is much smaller than the global mean of 148" and cone angles as low as
120" have been noted. In many of these complexes a small cone angle is de-
manded by the presence of bulky and inflexible co-ligands. For example, in
carborane metal complexes containing triphenylphosphines, the steric pressure
of the demanding carborane ligand may result in a reduction of the cone angle of
the phosphine. Similarly, multidentate ligands which have limited degrees of
freedom when located trans to the triphenylphosphine ligands cause a reduction
D.Michael P . Mingos 77
of the cone angle. Examples of such ligands include thia-crown ethers and
porphyrins. The carborane and multidentate ligands share a disc-like structure
and a reluctance to distort in order to permit the phenyl rings of the triphenyl-
phosphine ligand to adopt a conformation with a cone angle close to the mean.
In contrast, those complexes for which the calculated cone angle of triphenyl-
phosphine is consistently larger than the mean cone angle also belong to an
identifiable group. Many of the examples are complexes of gold and mercury,
where the metal has a lowish coordination number and where the co-ligands are
sterically not very demanding. In these complexes the triphenylphosphine can
take advantage of the greater space around the metal atom to expand its cone
angle.
It has been noted previously that the metal-phosphorus-carbon (M-P-C)
bond angle varies significantly in triphenylphosphine complexes. Indeed there
are Periodic regularities. A plot of cone angle for PPh, against the Periodic
Group number shows a progressive change in the angle across the transition
series. The average cone angle increases from ca. 142" to 155", and this may be
related to the decrease in the average M-P-C bond angle. This trend has been
noted previously and has been interpreted in terms of the change in s character in
the P-C bond, which is larger for complexes of the later transition metals. This
change in s character results in an opening up of the M-P-C angle and a larger
cone angle. The high proportion of complexes of the later transition metals with
large cone angles may be related to this.
We have also investigated the effect on the cone angle calculations of using the
crystallographically determined M-P distance rather than the putative 2.28 A
distance proposed by Tolman. The 2.28 A distance is smaller than the mean M-P
distance of 2.32(9) A calculated for all triphenylphosphine complexes. The cal-
culated mean cone angle based on observed metal-phosphorus distances is
146.9(6)O which contrasts with an angle of 148(5)O based on the standard
metal-phosphorus bond length proposed by Tolman.
In summary, this statistical analysis has provided some interesting insights
into the Tolman cone angle concept. Specifically, it has demonstrated that the
cone angles in real complexes vary much more than previously believed and that
there are systematic periodic differences in the average cone angles. The cone
angles may also be affected by the steric requirements of the co-ligands and the
coordination number of the complex. More surprisingly, the analysis suggests
that even within one complex containing several phosphine ligands the cone
angle may vary considerably.
4 Acknowledgements
The Alexander von Humboldt Stiftung is thanked for the research prize which
enabled me to complete this work in Heidelberg. Professor Walter Siebert's
hospitality and enthusiastic discussions there were much appreciated. Thomas
Miiller is also thanked for initiating this work.
78 Some New Insights into the Steric Efects of Tertiary Phosphine Ligands
5 References
1 G. J. Leigh and J. Chatt, Coord. Chem. Rev., 1991,108, 1.
2 J. Chatt and F. G. Mann, J . Chem. SOC.,1938,1949; 1939,1622.
3 J. Chatt and F. G. Mann, J . Chem. SOC.,1939,610.
4 R. S. Nyholm, Proc. Chem. Soc., 1961,273.
5 J. Chatt, F. G. Mann and A. F. Wells, J . Chem. SOC.,1938, 2086, provides a rare
example of a structural determination at that time.
6 A. W. Hofmann, Ann. Chim. Liebigs, 1857,103,357.
7 G. Booth, Adv. Inorg. Radiochem., 1964, 6, 1, provides a detailed early review of
phosphine and arsine complexes.
8 K. A. Jensen, Z . Anorg. A&. Chem., 1936,229,225.
9 L. Malatesta and R. Ugo, J . Chem. SOC., 1963,2080; A. D. Allen and C. D. Cook, Proc.
Chem. Soc., 1962,218.
10 F. Jardine, Prog. Inorg. Chem., 1981,28,63;Prog. Inorg. Chem., 1984,31,265, provide
very detailed reviews of the chemistries of these compounds.
11 L. Vaska and J. W. Diluzio, J . Am. Chem. SOC.,1961,83,1262.
12 A. Yamomoto, J . Organomet. Chem., 1986,300,347.
13 A. Sacco and M. Rossi., J . Chem. SOC., Chem. Commun., 1967,316.
14 L. E. Sutton, Biog. Mem. Roy. SOC.,1994,40,369.
15 J. Chatt and F. A. Hart, J . Chem. Soc., 1958,1474.
16 B. L. Shaw and J. M. Jenkins, J . Chem. SOC.(A),1966,770.
17 S. Ahrland, J. Chatt and N. R. Davies, Quart. Rev. Chem. SOC.(London),1958,12,265.
18 R. G. Pearson, J . Am. Chem. SOC.,1963,85,3533.
19 J. Chatt, Proc. Chem. SOC.,1962, 318.
20 J. A. Osborn, F. H. Jardine, G. Wilkinson and J. F. Young, J . Chem. SOC. (A), 1966,
1711.
21 C. A. Tolman and W. C. Diedel, Ann. N . Y. Acad. Sci., 1983,415,201.
22 C. A. Tolman, Chem. Reu., 1977,77,31.
23 A. G. Orpen and N. G. Connelly, J . Chem. Soc., Chem. Commun., 1985,1310.
24 D. P. Craig, A. McColl, R. S. Nyholm, L. E. Orgel and L. E. Sutton, J . Chem. SOC.,
1954,332.
25 J. Chatt and A. A. Williams, J . Chem. SOC.,1950,3061.
26 F. Seal, K. Ballreich and R. Schmutzler, Chem. Ber., 1961,94, 1123.
27 J. Chatt, G. A. Rowe and A. A. Williams, Proc. Chem. SOC.,1957,208.
28 J. Chatt and J. M. Davidson, J . Chem. Soc., 1965,843.
29 J. Chatt and H. R. Watson, Proc. Chem. SOC., 1960,243.
30 D. White and N. J. Colville, Adv. Organomet. Chem., 1994,36,99.
31 J. P. Collman, L. S. Hegedus, J. R. Norton and R. G. Finke, Principles and Applications
of Organotransition Metal Chemistry, University Science Books, Mill Valley, Califor-
nia, USA, 1987, p. 58.
32 B. Twanley, S. T. Haubrich and P. P. Power, Adu. Organomet. Chem., 1999,44,1, and
references therein.
33 F. R. Allen, J. E. Davies, J. J. Galloy, 0.Johnson, 0. Kennard, 0. F. Macrae, E. M.
Mitchell, J. M. Smith and D. G. Watson, J . Chem. I n . Comp. Sci., 1991,31, 187.
34 T. E. Miiller and D. M. P. Mingos, Trans. Met. Chem., 1995,20,553.
35 J. D. Dunitz and H. Burgi, Ace. Chem. Res., 1983,16, 153.
36 S. E. Garner and A. G. Orpen, J . Chem. SOC.,Dalton Trans., 1993,533.
Synthesis of New and Unusual
Metal Complexes from the
Reaction of Dihalogen Adducts of
Tertiary Phosphines with
Unactivated Metal Po wders
STEPHEN M. GODFREY AND CHARLES A. McAULIFFE
Department of Chemistry, University of Manchester Institute of Science and
Technology, Manchester M60 lQD, UK
1 Introduction
In 1847 Paul Thenard synthesised trimethylphosphine from the reaction of
methyl chloride with impure calcium phosphide at 180-300°C.’ The subsequent
discovery of aliphatic amines established the link between amines and phos-
phines and stimulated Hofmann and Cahours2to develop this area of chemistry.
A major development was the synthesis of triphenylphosphine by Michaelis3 in
1885.
In 1973 one of us edited a volume on transition metal complexes of Group 15
l i g a n d ~When
.~ approached to get his permission to dedicate the compilation to
him, Joseph Chatt, with characteristic generosity, requested that the book be
dedicated to Frederick Mann. The latter, said Chatt, invented the area of metal
complexes of phosphines. A subsequent volume was dedicated to Chatt.’
Indeed, Mann’s studies initiated an explosion of interest in transition metal
phosphine chemistry.6 Chatt was Mann’s student, and it is interesting to reflect
that Chatt employed transition metal salts as a means of derivatising tertiary
phosphines. Little did they realise how this area would expand! Chatt’s influence
on metal-phosphine chemistry was profound, and out of it sprung the employ-
ment of phosphine complexes as tools for investigations into metal hydrides,
nitrogen fixation, the stabilisation of unusual oxidation states, synthesis of metal
alkyls and aryls, and the immense practical area of homogeneous and heterogen-
80 Reaction of Dihalogen Adducts of Tertiary Phosphines
eous catalysis. The Chatt ‘family’ is large: Shaw, Venanzi, Leigh, Richards, and
their students. If we may be forgiven a personal statement: we view our pedigree
as Mann -+ Chatt --+ Venanzi -+McAuliffe Godfrey and Levason. Reason-
-+
ably, this may be seen as a very parochial view, since this area has been globally
researched. However, it was Mann and Chatt’s British School with the Antipo-
deans Dwyer and Nyholm who spearheaded this now vast area.
2 Background
There are four comprehensive reference texts on phosphine coordination chem-
i ~ t r y . ~ *Two
~ , ~are
? ’ somewhat ~ u t - o f - d a t e ,but
~ > the
~ article by Levason is more
recent. However, all four give much useful information on synthetic procedures
for the preparation of primary, secondary and tertiary phosphines; diphosphines;
multidentate and macrocyclic phosphines; mixed Group 15 donor and mixed
Group 15/16 donor systems; and phosphorus-containing heterocycles. The coor-
dination chemistry of phosphorus has expanded to include P, fragments, P6
rings, and even naked phosphorus atoms, for example, nine-coordinate P and
V6-P6 systems.8
The cone-angle concept has been comprehensively reviewed (see Reference 8
and the preceding article by Mingos). It was Tolman’ who introduced the idea of
the cone angle to describe in quantifiable terms the bulk of (mainly) phosphine
ligands. This term is especially applicable to Group 15 proligands. On binding to
metals they increase the coordination number of the donor atom to four, with
tetrahedrally disposed bonds that, rotated around an apical axis, describe a cone.
It is this that makes the term particularly appropriate for pnictide ligands. It is
important to appreciate that cone angles are not necessarily the dominant
steric/structural effect on ligand behaviour, and that steric bulk and cone angle
are not necessarily synonymous.
There has been much discussion of the nature of the M-PR, bond in metal
complexes, and the relative importance of the 0 and 7c contributions. Useful
background to this can also be found in References 4,5, and 8.
3 Work at The University of Manchester Institute of

Science and Technology (UMIST)
One evening in 1989,one of us (C. A. M.) was enjoying an after-work drink in The
Lass O’Gowrie (a pub on Charles Street, near the UMIST campus) with a close
friend and colleague, Tony Mackie, and we began to discuss the synthesis of
metal-phosphine complexes. We recognised that the traditional and much em-
ployed synthesis was the reaction of a metal salt with a quantity of a tertiary
phosphine (Equation 1).
MX, + yPR, + MX,(PR,), (1)
We recognised that this was the reaction of an oxidised metal with a phos-
phorus(II1)centre, and wondered if it would be possible for an ‘unoxidised’metal
to react with an oxidised phosphine, i.e. a phosphorus(v) centre. We both
Stephen M . Godfrey and Charles A. McAulife 81
thought that this was unlikely, since most metals have a somewhat impenetrable
surface oxide coating, as those of us who have enjoyed the challenge of initiating
a Grignard reagent preparation will attest.
We therefore planned to investigate the reactions of X,PR, (X = Br or I) with
unactivated metals (Equation 2).
M + nX,PR, +?
( n = 1 or 2)
To our delight we found that the activation of crude metal powders by reagents
X,PR, in dry diethyl ether at about ambient temperatures is facile and led to the
isolation of known, and, more frequently, previously unknown and unexpected
products (see below).
3.1 Nature of PR,X, (X = C1, Br or I)

Because of the success we initially experienced in the synthesis of coordination
complexes, we decided to investigate the structural chemistry of the reagents
PR,X, isolated from the reaction of stoichiometric quantities of dihalogen with
tertiary phosphines in diethyl ether.
Prior to our investigations, there were two recognised structural groups for
PR,X, compounds: molecular trigonal-bipyramidal trans-[PR,X,] and the
ionic [PR,X]X. Somewhat to our surprise we found that a single crystal X-ray
analysis of PPh,I, revealed it to have a tetrahedral four-coordinate structure,
Ph,P-1-1, a charge-transfer complex of molecular iodine and triphenylphos-
phine. As expected, the 1-1 separation in molecular diiodine, 2.71 A, is
lengthened to 3.161 A in PPh3I,.l0 There had been much previous work on
dihalogen adducts of tertiary phosphines, such as the solution studies by Harris
and his co-workers’ and by Du Mont.’, A subsequent study of a large number
of PR,I, (R, = Ph,, substituted triaryl, mixed arylalkyl, or trialkyl) showed that
the four-coordinate ‘spoke’/charge transfer structure was common in the solid
state.13 Moreover, this ‘spoke’ structure in the solid state was also exhibited by
Ph3P-Br-Br.14 We studied 20 other R,PBr, compounds and found that in
CDCl, solution they exist in the ionic (R,PBr)Br form, except for (C,F,),PBr,
which has a molecular five-coordinate structure both in CDCl, solution and in
the solid state,’, presumably due to the very low basicity of the parent tertiary
phosphine.
These results posed the question: what are the solid-state structures of tertiary
phosphine adducts of mixed halogens, such as PR,IBr? For all compounds
(R, = (p-ClC,H,),, Ph,, Ph,Prn, Ph,Me, PhMe, or Bun,) the ‘spoke’ structure
predominates, but other structures are present in small proportions. In particu-
lar, PPh,IBr exists as Ph,P-I-Br, with a little Ph,P-Br-T.16
The conclusions from our investigations are that the solid-state structures of
PR,X, (X = Br or I) depend crucially on the nature of R and the solvent of
preparation. For example, when employing diethyl ether as the solvent the ionic
[(Me,N),PI]I is produced, but PPhMe,I, has the molecular ‘spoke’ structure.’
Despite the fact that compounds of stoichiometry R,PCl, have been known
for over 120 years,l their solid-state structural nature has remained largely
unexplored, even though PPh,C12 is a widely used chlorinating agent in
organic reactions and is commercially available. Our X-ray crystallographic
study" of PPh,C12 crystallised from CH,C12/Et20 revealed it to be
[Ph,PCl+ - - C1- * . - +ClPPh,]Cl, and not a molecular trigonal-bipyramidal
*
adduct, not a molecular charge-transfer complex, and not the simple ionic
species [Ph,PCl]Cl. This contrasted with conclusions from all spectroscopic
data recorded on compounds of stoichiometry PR,C12 by earlier workers2' and
represents the first compound of this formulation to be crystallographically
characterised. However, we have shown from ,'P-{H} NMR studies that these
compounds are ionic in CDCl, solution. For those parent tertiary phosphines
that are weakly basic (or more acidic), a trigonal-bipyramidal structure was
revealed in the solid state and this also persists in CDCl, solution. It is evident
that all reported PR,F, compounds contain five-coordinate phosphorus centres.
However, for PR,Cl, and PR,Br2 the nature of R crucially determines the
structure.21This phenomenon has also been observed for AsR,Br,: the parent
AsPh, is a relatively weak base and thus AsPh,Br2 is trigonal-bipyramidal,
whereas the stronger base, AsMe,, forms a molecular four-coordinate species
Me, As-Br-Br.
In conclusion, the structural chemistry of dihalogen adducts of tertiary phos-
phines is clearly diverse and is sensitive to the nature of R and X; moreover the
central atom E (E = P, As2, or Sb2,) can play a crucial role, as can the solvent of
preparation. There has been some discussion about the bonding of P to the X
moiety.2627 Finally, although we have discussed here 1: 1 adducts of R,E with
X,, different ratios can lead to more complex compounds. For example one
polymorph of PPh,1427 reveals a strong interaction between the cation [PPh,-
I]' and a triiodide anion, and the individual [PPh,I] units are further linked
into a polymer by weak interactions between the triiodide anions. Further, the
compound PPri,14 contains two [PPr',I] cations weakly linked independently
+
to one of the two terminal iodine atoms of the same triiodide anion.2s
3.2 The Activation of Crude Metal Powders by Reagents PR,X,

Our investigations have revealed several surprising results:
(i) crude, inactivated metal powders react readily with these reagents and form
metal complexes;
(ii) complexes of metals in high oxidation states result from a one-pot syn-
(iii) stable complexes which defy Chatt's class (a)/class (b) categorisation and the
HSAB principle, e.g. [CoI,(SbPh,),], can be synthesi~ed;,~
(iv) low-oxidation-state gallium(r1)and indium@)complexes are a c c e s ~ i b l e ; ~ ~ * ~ ~
(v) even a noble metal such as gold can be oxidised readily.36
We now exemplify briefly the above observations.
Iron. In contrast to the well-established coordination of nickel salts with
monotertiary phosphines, reports of corresponding iron complexes are relatively
Stephen M . Godfrey and Charles A . McAulifle 83
rare, and only [FeCl,(PMe,),] and [FeCl,(PPh,),] are known for the oxidation
state 1 1 1 . ~ ~
Iron powder reacts with two equivalents of PPhMe,Br, in Et,O to form a
white solid, which is acutely sensitive to di~xygen.~' At exposure to low 0,-
partial pressures ( < 100ppm) it is possible to isolate an intensely purple trigonal-
bipyramidal complex [FeBr,(PPhMe,),], with the phosphines trans to each
other and in the apical positions. It might have been expected that exposure of
the initial white material, presumably an iron@) species, would have yielded a
p-0x0 complex.
We were interested to know if other zerovalent iron species could be oxidised
by these reagents. Reaction with [Fe2(CO),] yielded a series of diverse products:
ionic [ER,X][FeX,(ER,)] (E = P, R, = Ph,, Me,Ph or (p-MeOC,H,),, X = I;
R3E = Me,As, X = Br or I), [(Ph,E),Br][FeBr,], and the surprising ionic
adduct, [SbPh,] [Fe14]-SbPh31,remarkable not only for phenyl migration to the
antimony atom, but also for the formation of the rare [Fe14]- anion from a
zerovalent iron carbonyl prec~rsor.~'
Manganese. The X-ray crystal structure of [MnI,(PPhMe,)] made by direct
reaction of MnI, with PPhMe, in E t 2 0 has been reported by King and co-
w o r k e r ~ , ~and
' shows an infinite chain of MnI, units with two PPhMe, mol-
ecules additionally co-ordinated to alternate manganese atoms, producing a
sequence of manganese coordination numbers 4,6,4,6.. . However, when
PPhMe,IBr reacts with crude manganese powder the reaction of Equation 3
occurs.
2PPhMe,IBr + 2Mn -+ 2[MnBr,(PPhMe,)] + [MnI,(PPhMe,)] (3)
An X-ray crystal structure analysis,' of the iodo-product revealed that it, too,
was polymeric, but with a 5,5,5,5 structure, indicating that our new synthetic
method can produce isomeric forms of the complexes made by more conven-
tional means. Both isomers, however, react with 0, to form a 1: 1 adduct.
That subtle changes can significantly influence4, the product is illustrated in
Equation 4.
2PPh,I, + 2Mn + [MnI,(PPh,),] + MnI, (4)
Further structural variety was found in the product from the reaction in Equa-
tion 5.
P(NMe,),I, + Mn -+ [Mn(P(NMe,),}I,] (5)
The adduct 1 has the previously recognised 1 : 1 Mn1,:phosphine stoichiornetry,
but it is a dimer rather than polymeric (Scheme 1). Even more surprising were
the results of the 1 : 1 reaction of manganese powder with PMe,I,.,, This
produced a polymeric complex of 1: 1 stoichiometry, MnI,(PMe,), 2, but
this had the 4,6,4,6 structure. Complex 2 showed intriguing reactions with
molecular oxygen. [Mn,I,(PMe,),], 3, is unique in two respects: although
mixed-valence manganese complexes are well established with ligands involv-
ing 0- and N-donor systems, no such system was known for P-donors.
84 Reaction of DihaEogen Adducts of Tertiary Phosphines
(Me2N)3p\ /I\ /'

I /""\
/""\
I P(NMe2)3
1
2
\
/ \
Atmospheric O2 Trace 02,<100 ppm
+ MnIz
PMe3 PMe3
3
Scheme 1
Additionally, there is a PMe, molecule in the lattice, but it is surprising that the
material is not pyrophoric in air.
Cobalt. The reaction31 of cobalt metal with two mole equivalents of PMe,I,
yields [CoI,(PMe,),] (Equation 6).
Co + 2PMe,I, -+ [CoI,(PMe,),] + +I2 (6)
This complex is trigonal-bipyramidal. In contrast, when the reagent PBu",I,
was employed an intermediate species with the unusual formula [Co18(PBu",),]
was isolated. This proved to be [(Bu",PI),(p-I)] [(BU"~PI>(~-I)COI,], 4, and
provides a 'snapshot' of the reaction of these reagents with bare metals. Note that
cobalt is 'capturing' iodines from the reagent, and that at this point no Co-P
bonds exist.
The reagent AsPh,I, oxidises [co,(c0)8] (6: 1 molar ratio) to produce the
novel [(Ph,AsI),(I)][CoI,(AsPh,)] in quantitative yield. As well as being the first
crystallographically characterised complex containing a cobalt(r1)tertiary arsine
Stephen M . Godfrey and Charles A. McAulifle 85
/I\
P "Bu3
" B U ~P- I -I-co-I

4 -I'
\I
bond in the tetrahedral [CoI,(AsPh,)] - anion, it also contains the unusual

linear cation [Ph,As-I-I-I-AsPh,] +.43 When two mole equivalents of PMe,I,
react with cobalt (or nickel) powder, Equation 7, the MIr1complexes are pro-
duced quantitatively in a simple one-step r e a ~ t i o n . ~ ~ . ~ ~
2PMe,I, + M + [MI,(PMe,),] +*I2 (7)
In contrast, AsMe,I, and Co produce a mixed product, Equation 8.
4AsMe,I, + 2Co + (AsMe,I)[CoI,(AsMe,)] + [CoI,(AsMe,),] + +I2 (8)
Textbooks refer to Chatt's class (a) and class (b) acceptors, a concept subse-
quently developed in Pearson's HSAB principle. In contrast to this perceived
wisdom, the complex [CoI,(SbPh,),], which contains a hard metal centre and
extremely soft iodide and SbPh, ligands, is quite stable.34 Its formation is
thought-provoking.
2I,SbPh, + CO -+ [Ph,SbI][CoI,(SbPh,)] -+ [CoI,(SbPh,),] + $1, (9)
Nickel. The synthesis of unusual metal complexes by our new synthetic method is
shown by the formation of the square-planar cation44 in [Ph,PI] [Ni(PPh,)I,],
whereas, according to conventional wisdom, 'large' ligands such as PPh, and
iodide should force tetrahedral geometry around nickel@).
Gold. Gold has always been considered as the most noble of all metals. Only one
oxide of gold is known, Au,O,, and even this is unstable, decomposing to
produce gold metal at the moderate temperature of 160°C. Consequently, to
produce gold complexes it is necessary to solubilise the metal, usually by using
rather severe conditions, such as treatment with aqua regia and/or cyanide.
Clearly, the development of a new reaction route starting with elemental gold to
produce gold complexes under ambient conditions in a chemically inert solvent
is highly desirable.
We found36the following reaction, Equation 10.
3I,AsMe, + 2Au -+ 2[AuI,(AsMe,)] + Me,As (10)
The product is square-planar. Even more interesting is the reaction shown in
Equation 11.
2I,PMe, + Au --+ [Au13(PMe3),] + + I 2 (11)
This complex has a trigonal-bipyramidal structure with the phosphines in the
trans-apical positions, and is apparently the first structure where this geometry is
not forced upon the gold(rI1) by the steric requirements of the ligands.
These reactions indicate not only that even gold can be activated and oxidised
by EMe,I, (E = P or As), but they show that distinctly different products form
as a result of quite subtle changes in the reagents. AsMe,I, yields square-planar
[AuI,(AsMe,)] and I,PMe, yields trigonal-bipyramidal [Au(PMe,),I,], com-
plexes that differ both in stoichiometry and geometry.
Tin. The r e a c t i ~ n ~of~a yseries
~ ~ of PR,I, reagents with tin powder has yielded
complexes of empirical formula SnI,( PR,),, indicating that these reagents can
oxidise tin(0) to tin(1v) in one step. Tin(1v) complexes have hitherto proven
difficult to synthesise by conventional methods. The X-ray crystal structure of
octahedral trans-[SnI,(PPr",),] shows that all the iodine atoms are in the same
plane.
Gallium and Indium. Virtually nothing was known about the coordination chem-
istry of indium with tertiary phosphines until the pioneering work of Carty and
Tuck.,, We allowed I,PR, (R = Ph, Pr" or Pr') to react with indium metal
powder, and obtained a series of interesting and novel complexes, Equations
12-14.
3PPh,I, + 21n -+ [InI,(PPh,),~InI,(PPh3)] (12)

5
2PPr",I, + 21n --+ [{1nI2(PPrn3)},] (13)
Complex 5 was shown by X-ray crystallography to contain both tetrahedral, 6,

and trigonal-bipyramidal, 7, indium(II1) in the unit cell.39 No complex of in-
dium(@ of this stoichiometry had previously been identified. Finally, a reaction
PPh3 PPh3
I
I-&,
I ,,I
I I'
PPh3
6 7
with gallium metal powder38was performed with the diiodine adducts of tertiary
arsines, As(p-MeOC,H,),I, and AsEt,I,, to produce [GaI,((p-
MeOC,H,),As} ,] and [Ga,I,(AsEt,),], respectively. The last is a unique
example of a gallium-tertiary arsine complex containing a Ga-Ga bond, and
these complexes serve to illustrate the subtle effect of the organic substituent on
the chemistry of the arsenic atom.
Stephen M . Godfrey and Charles A. McAulifle 87
4 References
1 P. Thenard, C. R. Hebd. Seances Acad. Sci., Ser. C , 1847,25,892.
2 A. W. Hoffman and A. Cahours, Q. J . Chem. SOC.,1859,11,56.
3 A. Michealis and H. V. Soden, Ann., 1885,229,295.
4 C. A. McAuliffe (ed.), Transition Metal Complexes of Phosphine, Arsine and Stibine
Ligands, MacMillan, London, 1973.
5 C. A. McAuliffe and W. Levason, Phosphine, Arsine and Stibine Complexes of the
Transition Elements, Elsevier, Amsterdam, 1979.
7 W. Levason in The Chemistry of Organophosphorus Compounds, Vol. 1, F. R. Hartley
(ed.),John Wiley and Sons, New York, 1990, p. 567.
8 C. A. McAuliffe, ‘Phosphorus, Arsenic, Antimony and Bismuth Ligands’, in Compre-
hensive Coordination Chemistry, Vol. 2, G. Wilkinson, R. D. Gillard and J. A.
McCleverty (ed.),Pergamon Press, Oxford, 1987, p. 989.
9 C. A. Tolman, Chem. Rev., 1977,77,313.
10 S. M. Godfrey, D. G. Kelly, C. A. McAuliffe, A. G. Mackie, R. G. Pritchard and S. M.
Watson, J . Chem. SOC.,Chern. Commun., 1991,1447.
11 See References 10 and 11 in Reference 10 quoted above.
12 See, for example: W. W. Du Mont, M. Batcher, S. Pohl and W. Saak, Angew. Chem,
Znt. Ed. Engl., 1987,26, 912.
13 N. Bricklebank, S. M. Godfrey, A. G. Mackie, C. A. McAuliffe, R. G. Pritchard and P.
J. Kobryn, J . Chem. SOC.,Chern. Commun., 1993,101.
14 N. Bricklebank, S. M. Godfrey, A. G. Mackie, C. A. McAuliffe and R. G. Pritchard, J .
Chem. SOC.,Chem. Comm., 1992,355.
15 S. M. Godfrey, C. A. McAuliffe, I. Mushtaq, R. G. Pritchard and J. M. Sheffield, J .
Chem. Soc., Dalton Trans., 1998,3815.
16 N. Bricklebank, S. M. Godfrey, C. A. McAuliffe and R. G. Pritchard, J . Chem. SOC.,
Dalton Trans., 1993,2261.
17 N. Bricklebank, S. M. Godfrey, H. P. Lane, C. A. McAuliffe, R. G. Pritchard and J.-M.
Moreno, J . Chem. SOC.,Dalton Trans., 1995,2421.
18 A. Michaelis, Ann., 1876,181,256.
19 S. M. Godfrey, C. A. McAuliffe, R. G. Pritchard and J. M. Sheffield, J . Chem. Soc.,
Chem. Commun., 1996,2521.
20 K. B. Dillon and T. C. Waddington, Spectrochim. Acta, 1971,27A, 2381; A. Finch, P.
N. Gates and A. S. Muir, J . Raman Spec., 1988, 19, 91; J. Goubeau and R. Baum-
gartner, 2. Electrochem., 1960,64, 598.
21 S. M. Godfrey, C. A. McAuliffe, R. G. Pritchard, J. M. Sheffield and G. M. Thompson,
J . Chem. SOC.,Dalton Trans., 1997,4823.
24 P. Deplano, S. M. Godfrey, F. Isaia, C. A. McAuliffe, M. L. Mercuri and E. F. Trogu,
Chem. Z3er.l Recueil, 1997,130, 299.
25 R. Deeth, J . Chem. SOC.,Dalton Trans., 1997, 1995.
26 F. A. Demartin, A. Devillanova, F. Isaia and V. Lippolis, J . Chem. Soc., Dalton Trans.,
2000,1959.
27 F. A. Cotton and P. A. Kibala, J . Am. Chem. SOC., 1987,109,3308.
28 W. I. Cross, S. M. Godfrey, C. A. McAuliffe, R. G. Pritchard, J. M. Sheffield and G. M.
Thompson, J . Chem. Soc., Dalton Trans., 1999,2795.
29 S. M. Godfrey, D. G. Kelly, A. G. Mackie, P. P. MacRory, C. A. McAuliffe, R. G.
Pritchard and S. M. Watson, J . Chem. SOC., Chem. Commun., 1991,1447.
30 H. P. Lane, S. M. Godfrey, C. A. McAuliffe and R. G. Pritchard, J . Chem. SOC., Dalton
Trans., 1994,3249.
31 C. A. McAuliffe, S. M. Godfrey, A. G. Mackie and R. G. Pritchard, Angew. Chem. Int.
Ed. Engl., 1992,31,919.
32 C. A. McAuliffe, S. M. Godfrey, A. G. Mackie and R. G. Pritchard, J . Chem. SOC.,
Dalton Trans., 1996, 157.
34 S. M. Godfrey, C. A. McAuliffe and R. G. Pritchard, J . Chem. SOC.,Chem. Commun.,
1994,45.
35 S. M. Godfrey, C. A. McAuliffe and R. G. Pritchard, J . Chem. SOC.,Dalton Trans., 1993,
2875.
36 S. M. Godfrey, N. Ho, C. A. McAuliffe and R. G. Pritchard, Angew. Chem., Int. Ed.
Engl., 1996,35,2343.
38 B. Beagley, S. M. Godfrey, K. J. Kelly, S. Kungwankunakorn, C. A. McAuliffe and R.
G. Pritchard, J . Chem. SOC.,Chem. Commun., 1996,2179.
39 S. M. Godfrey, K. J. Kelly, P. Kramkowski, C. A. McAuliffe and R. G. Pritchard,
40 J. D. Walker and R. Poli, Inorg. Chem., 1989,28, 1793.
41 B. Beagley, J. C. Briggs, A. Hosseiny, W. E. Hill, T. J. King, C. A. McAuliffe and K.
Minten, J . Chem. SOC., Chem. Commun., 1991,1449.
42 S. M. Godfrey, C. A. McAuliffeand R. G. Pritchard, J . Chem. Soc., Dalton Trans., 1993,
371.
43 S. M. Godfrey, H. P. Lane, A. G. Mackie and R. G. Pritchard, J . Chem. SOC.,Chem.
Commun., 1993,1190.
45 N. Bricklebank, S. M. Godfrey, C. A. McAuliffe and K. C. Molloy, J . Chem. SOC.,
46 A. J. Carty and D. G. Tuck, Prog. Inorg. Chem., 1975,19,243, and references therein.
NMR Studies of Metal
Complexes and Clusters with
Carbonyls and Phosphines
B. T. HEATON
Department of Chemistry, University of Liverpool, Liverpool L60 7ZD, UK
1 Introduction
It was with great pleasure that I accepted the invitation to participate in the
ICCC at Edinburgh in the session devoted to Joe Chatt and to write this article
about the enduring influence he had on the chemistry I have been involved with
since I started my D. Phil. with him at Queen Mary College (QMC), London in
July 1964. At that time, Joe Chatt had just started the Unit of Nitrogen Fixation.
At The Frythe, he had built up a large internationally-recognised team, including
experts such as Bernard Shaw and Luigi Venanzi, and it must have been very
difficult for him at QMC to have only an inexperienced group of three research
students. This was cushioned to some extent by being able to retain his former
secretary from ICI, Inga Wass (nke Schmidt), who helped him enormously.
On my first day I was asked if I would like to move to the University of Sussex
since there was not enough space to develop the new Unit of Nitrogen Fixation
at QMC. As a result, we moved to Sussex in October, 1964 and our group was
among the first to occupy the purpose-built chemical laboratories at Sussex.
These were opened officially later in the year by the Queen. Designed by Sir Basil
Spence, this new University was a very exciting and stimulating place to be.
Nevertheless, although the architecture on the University campus was very
pleasing, there were initial problems in the laboratory design, This resulted in
Chatt and his collaborators spending much time with the architects in designing
the new Unit of Nitrogen Fixation, which was occupied subsequently after 1967.
At the end of my first post-graduate year, I did not have too many results!
Most of my time had been devoted to establishing the laboratory and to
developing the necessary preparative expertise. During this period, Chatt’s
group started to expand rapidly and my development was greatly assisted by the
arrival of permanent nitrogen fixation chemists, Ray Richards and Jeff Leigh and
90 NMR Studies of Metal Complexes and Clusters with Carbonyls and Phosphines
overseas visitors such as Paolo Chini who spent six months in 1965 with Chatt as
a NATO fellow. He and I worked on benches opposite each other, and I learned
much of my preparative technique from him. It was natural some years later to
start our long and productive collaboration on NMR studies of clusters, dis-
cussed below.
It is interesting to reflect on the state-of-the-art equipment being used during
my D. Phil. work, which involved the preparation of water soluble phosphorus-
ligand complexes via the hydrolysis of ligands containing P-Cl bonds. Far IR
spectrometers had just become available. This enabled the assignment of v(Pt-X)
in, for example, [PtX,(PR,),] (X = C1 or Br) so that cis and trans isomers could
be distinguished. This allowed the very time-consuming dipole moment
measurement, which had previously been routinely used at The Frythe, to be
dispensed with. The modern NMR methods were still a long way ORWe were
using continuous wave NMR spectrometers with either permanent ('H, 60
MHz) or electromagnets ('H, 100 MHz) and a Computer of Average Transients
(CAT) for ,'P NMR measurements! Using a CAT was a pretty hit-and-miss
exercise because we did not know which region to scan and because of the very
small spectral sweep width. Nevertheless, I was taught the fundamentals of NMR
spectroscopy by Alan Pidcock who had just taken up a Lectureship at Sussex
having obtained his doctorate with Luigi Venanzi, then at the University of
Oxford.
These early measurements stimulated my interest in NMR spectroscopy, and,
on moving to the University of Kent at Canterbury (1972), we were lucky to be
able to buy the first Fourier Transform NMR spectrometer in the UK. This
instrument was still based on an electromagnet ('H, 100 MHz) but allowed faster
acquisition of NMR spectra and enabled the development of multinuclear NMR
spectroscopy. This permitted me to start collaborating with Paolo Chini who
had taken up an appointment at the University of Milan where he was develop-
ing metal carbonyl cluster chemistry. In Milan, Chini had access only to an IR
spectrometer that aided the clean preparation and subsequent crystallisation of
clusters, and, importantly, an X-ray diffractometer for their structural character-
isation.
In this article are reviewed some of my more important multinuclear NMR
contributions made in the areas of metal carbonyl and metal phosphine com-
plexes involved in homogeneous catalysis.
2 Metal Carbonyls
Chatt's group attracted Chini because Booth and Chatt had noticed that addi-
tion of base to [PtCl,(CO),] produced various intense colours.' Work in 1965 at
Sussex allowed the isolation of bright red [Pt3(p-C0),Lx] (x = 3, L = PPh,Bz;
x = 4, L = PPh,) and black [Pt3(p-C0)4L3] (L = PPh, or PMe,Ph).2 These
were all isolated by careful fractional crystallisation using a variety of solvents.
Subsequently, the initial observation of Booth and Chatt was shown to be due to
the formation of Pt,-triangular stacks, [Pt,,(p-CO),,(CO),,] ( n = 1-6).334 This
remarkable series of clusters showed that the Pt,-triangles were almost eclipsed
B. T. Heaton 91
Figure 1 Schematic representation of the X-ray structure of [ P t 9 ( ~ - C O ) g ( C O ) g ]and

2-
the rotation of the [ P t 3 ( ~ - C 0 ) 3 ( C 0 )groups
3] about the pseudo-three-fold axis
in solution
in the solid state but 13C and 195PtNMR studies in solution showed that, for
n = 3, there was ready rotation of the intact [Pt,(p-CO),(CO),]-groups about
the pseudo-3-fold axis (Figure l), together with inter-exchange of [Pt3(p-
CO),(CO),] groups between the Pt, and Pt,, clusters (Equation l).5
This was the first example of a facile, intra-molecular, metal polyhedral rearran-
gement in a transition metal-carbonyl cluster. Other work showed that related
metal polyhedral rearrangements occur when the metal polyhedra are not well
close-packed, as when they contain a hetero-interstitial atom, e.g.
[Rh9E(C0)2,]2- (E = P or As),6 [Rh,,E(CO),,]"- (n = 3, E = P or As; n = 2,
92 N M R Studies of Metal Complexes and Clusters with Carbonyls and Phosphines
E= S)677 and [Rh12Sb(C0)2,].3-6

The initial work on Rh-containing carbonyl clusters relied heavily upon the
coupling patterns and relative intensities obtained from one-dimensional NMR
spectra, together with appropriate isotopic-labelling experiments. Thus, it was
possible to determine unambiguously the source of the interstitial carbide in the
trigonal-prismatic cluster, [Rh,C(CO), J 2 - (Equation 2).8
6[Rh(CO),]- + l3cCl, +. [Rh613C(C0),,]2- + 4C1- + 9CO (2)
Related work proved the presence and determined the source of the interstitial
nitride in the isoelectronic cluster [Rh,N(CO), ,]- , 9 and the facile interconver-
sion of the trigonal prismatic and octahedral clusters (Equation 3)."
-2co
[Rh6C(C0)l .51* [Rhf.jC(CO)l3 1 2 -
+ 2co
trigonal prism octahedron (3)
Interestingly, the observed migration in solution of seven COs around one of
the three Rh,-square planes of the Rh,-octahedron in [Rh,C(C0),3]2- is exact-
ly in accord with the librations of the same COs around the same Rh,-square
plane in the solid state." This provided the first example of a correlation
between C O fluxionality in solution and thermal motion in the solid state.
However, there is still much more to be done to be able to predict both the
migrational pathways and to clarify the exact mechanism - even in small
clusters.
Acquisition of lo3Rh NMR data using direct methods originally required the
use of 15 mm tubes containing ca. 1 g of cluster and the use of high field
spectrometers because of the insensitivity of this nucleus, despite being 100%
naturally abundant. In the 1970s and 1980s measurements were done on the then
highest field spectrometer (360 MHz) available, which was provided by the
EPSRC service at Edinburgh. Clearly, this was neither very convenient nor easy
to arrange from Canterbury and indirect methods were developed; a
13C-{103Rh} probe was developed at the University of and a
'H-{lo3Rh} INDOR probe at the University of Bristol.I6 The 13C-('03Rh}
NMR measurements allowed unambiguous assignment of resonances and clear
elucidation of C O migrational pathways.17 'H-( Io3Rh}INDOR measurements
also provided important information about both H-site occupancies and H
migration in clusters.16*' A particularly interesting series of clusters is
[H,Rh, 3(C0)12(p-C0)1 2](5 -,)- (x = 1-4) which, in solution, all exhibit inter-
stitial H-migration and migration of all but the three bridging COs in the
Rh,-hexagonal plane. At low temperature, it was possible to stop both H and
C O migration for x = 3 and show that the hydrogens occupy square-faces of the
hexagonal close-packed Rh, skeleton. It was reasoned that the preferential site
occupancy of these different square-faces decreases with increasing y-CO occu-
pancy of the square-face and this hypothesis has recently been substantiated by
the neutron diffraction determination of the structure of [H,Rh, 3-
(CO)12(p-CO)12]3- (Figure 2).18 For x = 2 in solution it was impossible, even at
B. T. Heaton 93
Figure 2 Schematic representation of the neutron difraction structure of

[H2Rh13(p-C0),2(C0)12]3-; each H is co-planar with a Rh,-square-face
very low temperature, to stop H-migration. However, H-migration is much less

facile in the solid state and, as expected from the static neutron diffraction
structure (Figure 2), two equally intense high-field 'H resonances are observed in
the solid-state 'H NMR spectrum. Comparative variable temperature solution
and solid-state NMR studies on other clusters substantiate this general view but
probably the most spectacular result is observed for [HCo,(CO), 5] -. In this
case, neutron diffraction studies show the presence of an interstitial hydride
which gives rise to 6(H)at + 1 ppm whereas in solution the resonance appears at
+ 23.2 ppm, due to H-migration from inside to outside the cluster.20
Recently, indirect 2D NMR methods have been developed for obtaining
Io3RhNMR data (HMQC measurements) for rhodium clusters containing edge-
'
and face-bridging ~ a r b o n y l s . ~$ 2.2~For such clusters, multiple quantum effects
become important because of the presence of 100% lo3Rhand non-conventional
delays are required. These methods are now used routinely to collect data on
substituted clusters, which, since there is a reference point (substituted Rh), allow
unambiguous assignments. ' '
7 3 2
The predominant patterns of monodentate phosphine substitution in

tetrahedral and octahedral rhodium clusters have been e ~ t a b l i s h e d ' ~ and '~
recent work on trigonal prismatic clusters suggests the following pattern (see
Figure 3).22324Thus, only one of the three possible isomers of [Rh,C(p-
CO),(CO),(PPh,),] - is detected in solution; further substitution occurs, but
the structures of these derivatives have not yet been established. There are few
reported ligand-substituted clusters containing seven or more metal atoms and
Figure 3 Schematic representation of the solution structures of

[Rh6C(p-C0)9(C0)6-,(PPh,)x]2 - ( x = 1 or 2)
there is clearly much more work to be done on both hetero- and homo-metallic
lower and higher nuclearity clusters to understand better non-carbonyl mono-
and multi-dentate ligand site occupancies.
For N-donor ligands (e.g. pyridine and dipyridyl, py and bipy), reaction with
[Rh,(CO), ,] results surprisingly in disproportionation (Equations 4 and 5), and
the solution structures of the products (established from VT HMQC measure-
ments) are subtly different, depending upon whether the ligand is mono- or
bi-dentate (see
Figure 4).24
All of the above work involved neutral or anionic clusters with v(C0) < 2145
cm-l. Recently, there has been a significant number of new metal carbonyls with
v(C0) > 2145 cm-'.25326Most of these complexes are mononuclear and contain
metals in relatively high oxidation states. We recently reported only the second
example of a dinuclear carbonyl in this class. Surprisingly, [Pt2(CO),] 2+ 1,
which is colourless and very moisture-sensitive, is formed by the slow dissolution
of black PtO, in concentrated H,SO, on stirring under an atmosphere of C O for
a few days at room temperature (Equation 6).
Conc H,SO,
-k co RT/1 atm ' CPt2(CO)61+,
The structure of complex 1, which has not yet been isolated in the solid state,
was established by 13C and 195PtNMR measurements in solution using both
B. T. Heaton 95
I
2+
I
COB-Pt -
.CoA
,.;
I /Pt-coB
I co*
COA
-
1
natural-abundance 3 C 0 and 99%-enrichment. This formulation was subse-

quently substantiated by Raman, EXAFS and other measurements.
Even for dinuclear platinum complexes, both 13C and 195PtNMR spectra
quickly become very complicated with increasing 3C-enrichment due to the
increasing abundance of the different isotopomers, which results in long-range
coupling and the appearance of second-order spectra (see Figure 5). This prob-
ably also accounts for our lack of good NMR results on higher nuclearity,
crystallographically characterised cluster^^'.^^ such as [Pt 19(CO)22]4- and
[HNi,,Pt,(CO),,] 5- although the influence of metal character with increasing
cluster size could also be important. In this connection, it is perhaps significant
that, even for clusters containing metals without a spin, no NMR data have
been reported on any high nuclearity ( > ca. 30 metal atoms) structurally-
characterised cluster, including the recently reported three-shell cluster,
[Pdi 4 5 ( ~ ~ ) ~ ( ~ ~ ~ 3 ) 3 0 1 - ~ ~
3 Mechanistic in situ Homogeneous Catalytic Studies

on Metal-Phosphine Complexes
Chatt was one of the early pioneers in furthering our understanding of the nature
of the coordinate link in metal complexes containing phosphine l i g a n d ~ . ~ '
I " ' I " " 1 " " I " " I " " I " ' 7 I " . ' I . I 1 I
' ' , ' I ' , ' . I ' 3 "
180.0 175.0 170.0 165.0 160.0 155.0 150.0 145.0 140.0 135.0
Figure 5 N M R spectra o f [ P t 2 ( C O ) J 2 +(a)

: Observed 13C at 1.1% 13CO;
(b) Observed/simulated I3C at 99% 13CO;(c)Observed 195Ptat 1.1% 13CO;
(d)Observed/simulated 195Ptat 99% I3CO 6 (CO,) 158.7,6 (COB)166.3,6 ( P t )
-21 1.2, 'J(Pt-CO,) 1595.7, 'J(Pt-COB) 1281.5, 'J(Pt-Pt') 550.9, 2J(Pt-C0,)
-26.2, 2J(Pt-COB)199.6, 2J(Co,-COB) = 3J(co,-cOB) =0,
3J(COB-COB)= 19.8
Nowadays, metal-phosphine complexes are widely used as catalysts for a variety

of industrially important homogeneous catalytic reactions, which often require
an elevated pressure of reactant gas.
In order to improve the selectivity and efficiency of the catalytic reaction, it is
important to have a detailed knowledge both of the nature of the intermediates
and of the rates of the individual steps involved in the catalytic cycle in situ.
Possession of this knowledge also avoids patent litigation.
B. T. Heaton 97
The initial NMR measurement under a high pressure of gas, aimed at spectro-
scopic characterisation of species present under these conditions, was carried out
in the early 1 9 8 0 ~ this
; ~ ~was followed by Roe's introduction of the now
commonly used sapphire tube.33 Although the sapphire tube is easy to use, it
suffers from the following problems.
1. Poor gas/solution mixing, which often results in rate-limiting step gas dissol-
ution. Furthermore, without agitation dissolution of H, (30 atm) in CHCl, at
room temperature can take up to ca. 10 h before equilibrium is reached.
2. Inability to measure/maintain constant pressure. This presents serious prob-
lems for rapid reactions because of gas depletion in both 5 and 10 mm NMR
tubes.
All these problems have been overcome by Jon Iggo at Liverpool through the
design of a flow cell34 which allows multinuclear NMR measurements to be
made at constant pressure ( < 200 atm) from - 40 to + 190°C. This cell can be
used not only to characterise structurally catalytic intermediates but also to
carry out meaningful kinetic measurements, which compare favourably with
kinetic data from reactions carried out in a stirred autoclave. For rhodium-
catalysed hydroformylation reactions, the rates determined in the flow cell are an
order of magnitude higher than those obtained from a sapphire tube.
Recently, this cell, together with complementary high-pressure IR measure-
ments, has been used to investigate the mechanism of the rhodium-catalysed
hydroformylation using a monodentate biuret-based phosphorus diamide
ligand, 2.35 The major rhodium acyl species observed in situ is
[Rh(C(O)R}(CO),L,] which has a trigonal-bipyramidal structure with two
equatorial ligands, L. The equatorial C O is found to exchange with dissolved C O
at a much faster rate than the apical CO. Although the generally accepted
mechanism proposed by W i l k i n ~ o n is
, ~confirmed, the overall kinetics are rather
complicated; the rate-determining stage of the hydroformylation reaction for this
system is not due to a single step but is strongly dependent on the conditions
used.
There is much current industrial interest in the palladium-catalysed

methoxycarbonylation of ethene to give methyl propanoate (MeP), the precur-
sor to methyl methacrylate. The Pdo catalyst precursor contains a novel
bis(phosphine), for example, 1,2-(CH2PBut2),CsH4,and, under reaction condi-
tions, gives MeP with high turnover and selectivity (99.98%).37We have inves-
tigated this reaction and have spectoscopically characterised all the Pd" inter-
mediates involved in the catalytic cycle (see Scheme l).38For none of the
intermediates in Scheme 1 is there evidence for dissociation of the bidentate
phosphine and it is surprising that the Pd-Et intermediate exhibits a P-C-H
interaction, even in polar solvents such as RCN (R = Me or Et); in both the other
intermediates shown in Scheme 1, the fourth site is occupied by solvent. There is
no evidence for the involvement of Pd-OMe species which, together with Pd-Hs,
have been proposed to be involved in the related copolymerisation of alkenes
with C O to give the commercially available copolymer, K a r i 1 0 n . ~ ~
x = solv do
PBut2
L-L =
PBu'~
Scheme 1 Spectroscopically characterised intermediates involved in the Pd-catalysed

methoxycarbonylation of ethene
The characterisation of all the species in Scheme 1 provides one of the few
homogeneous transition metal-catalysed cycles which has been completely eluci-
dated.
Earlier work on the in situ characterisation of rhodium-phosphine/hydrazine
complexes was instrumental in the successful outcome of patent litigation be-
tween Hovione (Portugal) and Pfizer (USA). Here, we were able to prove
conclusively by 31P,"N, and lo5RhNMR spectroscogies that in the hydrogena-
tion of doxycycline to methacycline, rhodium complexes with N-donor ligands
such as 3 and 4 were involved, rather than [RhCl(PPh3)3].40*41
[ 1' [
ph3p\R;NH2NH2
Ph3P' \NH2NH2
Ph P
Ph3P'
\Rh/
NH~YNH,
C
, H2
'NH-NH2
4
,Rh\
PPh3
PPh3
12+
B. T. Heaton 99
Chatt showed that hydrazine is one of the intermediates involved in some
nitrogen fixation pathways and it was a pleasant coincidence for me to carry out
this work on hydrazine complexes and to be able to prove their involvement in
catalytic hydrogenation reactions.
4 Acknowledgements
For the cluster work, it is a pleasure to thank the groups in Milan, Bologna,
Osaka, Canterbury, Liverpool, and more recently St Petersburg, who have
contributed to this work over the years; I am most grateful to all the people
whose names appear in the references. For the catalytic work, my thanks go to
Taro Eguchi for his help with building/designing/tuning probes in the early
high-pressure NMR experiments and for passing on his expertise, to Jon Iggo
who has developed the unique high-pressure NMR flow cell, to Chacco Jacob for
involvement in the catalytic hydrogenation work, to Stefan0 Zacchini for eluci-
dating the mechanism of the methoxycarbonylation of ethene and to Robin
Whyman for his general contributions to our catalytic studies and for his
expertise on high-pressure IR spectroscopy. I am also grateful to EPSRC,
Hovione, Enios Acrylics, and BP Chemicals for providing the necessary financial
support.
5 References
1 J. Chatt and G. Booth, unpublished results quoted in G. Booth, Adv. Inorg. Chem.
Radiochem., 1964,6,47.
2 G. Booth, J. Chatt and P. Chini, J . Chem. Soc., Chem. Commun., 1965,639.
3 J. C. Calabrese, L. F. Dahl, P. Chini, G. Longoni and S. Martinengo, 1. Am. Chem.
SOC., 1974, 96, 2614.
4 G. Longoni and P. Chini, J . Am. Chem. Soc., 1976,78,7225.
5 C. Brown, B. T. Heaton, P. Chini, A. Fumagalli and G. Longoni, J . Chem. Soc., Chem.
Commun., 1977, 309; C. Brown, B. T. Heaton, A. D. C. Towl, P. Chini, A. Fumagalli
and G. Longoni, J . Organomet. Chem., 1979,181,233.
6 B. T. Heaton, L. Strona, R. Della Pergola, J. L. Vidal and R. Schoening, J . Chem. Soc.,
7 L. Garlaschelli, A. Fumagalli, S. Martinengo, B. T. Heaton, D. 0.Smith and L. Strona,
J . Chem. SOC.,Dalton Trans., 1982,2265.
8 V. G. Albano, P. Chini, S. Martinengo, D. J. A. McCaffrey, D. Strumolo and B. T.
Heaton, J . Am. Chem. Soc., 1974,96,8106.
9 S. Martinengo, G. Ciani, A. Sironi, B. T. Heaton and J. Mason, J . Am. Chem. Soc.,
1979,101,7095.
10 B. T. Heaton, L. Strona and S. Martinengo, J . Organomet. Chem., 1981,215,415.
11 D. Braga and B. T. Heaton, J . Chem. Soc., Chem. Commun., 1987,608.
12 B. E. Mann, J . Chem. Soc., Dalton Trans., 1997,1457.
13 B. F. G. Johnson, J . Chem. Soc., Dalton Trans., 1997, 1473.
14 C. Brown, B. T. Heaton, L. Longhetti, D. 0. Smith, P. Chini and S. Martinengo, J .
Organomet. Chem., 1979,169,309.
15 C. Brown, B. T. Heaton, L. Longhetti, W. T. Povey and D. 0. Smith, J . Organomet.
Chem., 1980,192,93.
16 S. Martinengo, B. T. Heaton, R. J. Goodfellow and P. Chini, J . Chem. SOC. Chem.
Commun., 1977,39; C. Allevi, B. T. Heaton, C. Seregni, L. Strona, R. J. Goodfellow, P.
Chini and S. Martinengo, J . Chem. SOC.,Dalton Trans., 1986,1375.
17 B. T. Heaton, J. A. Iggo, I. S. Podkorytov, D. J. Smawfield and S. P. Tunik, in Metal
Clusters in Chemistry, ed. P. Braunstein, L. Oro and P. Raithby, Wiley-VCH, 1999,
p. 960.
18 R. Bau, M. H. Drabnis, L. Garlaschelli, W. T. Klooster, Z. Zie, T. F. Koetzle and S.
Martinengo, Science, 1997,275, 1099.
19 T. Eguchi, R. A. Harding, B. T. Heaton, G. Longoni, K. Myagi, J. Nahring, N.
Nakamura, H. Nakayama, T. Pakkannen, J. Pursianinen and A. K. Smith, J . Chem.
SOC.,Dalton Trans., 1996 625.
20 B. T. Heaton and T. Eguchi, J . Chem. SOC.,Dalton Trans., 1999,3523.
21 B. T. Heaton, J. A. Iggo, I. S. Podkorytov, D. J. Smawfield, S. P. Tunik and R.
Whyman, J . Chem. SOC.,Dalton Trans., 1999, 1917.
22 J. S. Z. Sabounchei, B. T. Heaton, J. A. Iggo, C. Jacob and I. S. Podkorytov, J . Cluster
Chem., 2001,12,347.
23 B. T. Heaton, L. Strona, R. Della Pergola, L. Garlaschelli, U. Sartorelli and I. H.
Sadler, J . Chem. SOC.,Dalton Trans., 1983,173; M. Bojczuck, B. T. Heaton, S. Johnson,
C. A. Ghilardi and A. Orlandini, J . Organomet. Chem., 1988,341,473.
24 J. S. Z. Sabounchei, B. T. Heaton, J. A. Iggo and C. Jacob, unpublished data.
25 H. Willner and F. Aubke, Angew. Chem., Int. Ed. Engl., 1997,36,2402.
26 S. H. Srauss, J . Chem. SOC.,Dalton Trans., 2000, 1.
27 Q. Xu, Y. Souma, B. T. Heaton, C. Jacob and K. Kanamori, Angew. Chem. Int. Ed.
Engl., 2000,39,208; Q. Xu, B.T. Heaton, C. Jacob, K. Mori, Y. Ichihashi,Y. Souma, K.
Kanamori and T. Eguchi, J . Am. Chem. SOC.,2000,122,6862.
28 D. M. Washecheck, E. J. Wucherer, L. F. Dahl, A. Cemotti, G. Longoni, M. Manas-
sero, M. Sansoni and P. Chini, J . Am. Chem. Soc., 1979,101,6110.
29 A. Cerriotti, F. Demartin, G. Longoni, M. Manassero, M. Marchionna, G. Piva and
M. Sansoni, Angew. Chem. Int. Ed. EngE., 1985,24,697.
30 N. T. Trau, D. R. Powell and L. F. Dahl, Angew. Chem. Int. Ed. Engl., 2000,39,4121.
31 J. Chatt, Nature, 1950,637; J . Chem. Soc., 1951,652.
32 B. T. Heaton, J. Jonas, T. Eguchi and G. A. Hoffmann, J . Chem. SOC.,Chem. Cornmun.,
1981,331; B. T. Heaton, L. Strona, J. Jonas, T. Eguchi and G. A. Hoffman, J . Chern.
Soc., Dalton Trans., 1982, 1159.
33 D. C. Roe, J . Mag. Res., 1985,63,388.
34 J. A. Iggo, D. Shirley and N. C. Tong, New J . Chem., 1998,1043.
35 S. C. van der Slot, P. C. Kamer, P. W. N. M. van Leeuwen, J. A. Iggo and B. T. Heaton,
Organometallics, 2001, 20,430.
36 D. Evans, J. A. Osborn and G. Wilkinson, J . Chem. SOC.(A),1968,3133.
37 W. Clegg, G. R. Eastham, M. R. J. Elsegood, R. P. Tooze, X. L. Wang and K. W.
Whiston, Chem. Commun., 1999, 1877.
38 G. R. Eastham, B. T. Heaton, J. A. Iggo, R. P. Tooze, R. Whyman and S . Zacchini,
39 E. Drent, J. A. M. Broekhaven and P. H. M. Budzelaar, Applied Homogeneous
Catalysis with Organometallic Compounds, ed. B. Cornils and W. A. Herrman, VCH,
Weinheim, 1996, p. 333.
40 B. T. Heaton, C. Jacob, G. L. Monks, M. B. Hursthouse, I. Ghatak, R. G. Sommerville,
W. Heggie, P. R. Page and I. Villax, J . Chem. SOC.,Dalton Trans., 1996,61.
41 B. T. Heaton, C. Jacob and P. R. Page, Coord. Chern. Reu., 1996,154,193.
SECTION D:
Transition Metal Complexes of

Olefins, Acetylenes, Arenes and
Related Isolobal Ligands
Modern organometallic chemistry of the transition metals has its origins in ma-
terialsjrst described in the 19th and the early 20th centuries. These include Zeise’s
salt, early examples of metal carbonyl compounds such as Mond’s nickel tetracar-
bony1 discovered in 1888, and the so-called polyphenylchromium compounds in
1919. Further milestones of the 20th century include the discovery in 1930 of
butadienetricarbonyliron and of ferrocene described in 1951.
It was Zeise’s salt, a compound believed to contain C,H4 bound to platinum,
which hadjrst caught Chatt’s young eye after reading Keller’s Chemical Review on
metal-olefin complexes published in 1941. He had been intrigued by Gel’man’s
suggestion that four electrons were involved in bonding between Pt and C2H4,
including two platinum d electrons. Chatt was also struck by the divergence of view
between that he recollected from F. G. Mann’s undergraduate lectures on olefin
complexes in which Mann had suggested that the C=C n-electrons acted as a lone
pair in a donor bond to the metal and Keller’s scepticism that such bonding or
Gel‘man’s model were possible.
Elucidating the constitution and structure of K[PtCl,(C,H,)]-H20 had repre-
sented a challenge to chemists ever since it had been first isolated from the reaction
of platinum(II) chloride and ethanol by William Christopher Zeise in 1825. From
early arguments about its constitution to later eflorts to characterise its structure
and describe the nature of the metal olefin bond some 125 years after its discovery,
Zeise’s salt continued to fascinate chemists. As is o f e n the case, such fascination
brought forth new and important ideas on chemical bonding as well as deeper
understanding that grew to have technological and industrial significance. Some
of this early work, particularly relating to the contributions of M . J . S. Dewar
and Chatt, is reviewed by N . Winterton. G. Frenking describes more recent
developments arising from the application of modern quantum chemical methods to
the so-called Dewar-Chatt-Duncanson model of bonding of olefins to transition
metals.
102 Transition Metal Complexes
The scope of discovery that arose following the later work of Chatt and others
and the patterns of bonding and reactivity among isoelectronic ligands and metals is
rejected in the other contributions brought together in this section. These include
chapters by H.-W. Fruhauf (on 1,3-dipolar additions to a series of 1,4-dia-
zabutadienemetal carbonyl and isonitrile complexes), by V. C. Gibson, the Royal
Society of Chemistry’s Chatt Lecturer for 2001-2002 (on studies of metal imido
complexes which reveal and explore their isolobal relationship with metal cyclopen-
tadienyl complexes), by M . Abou Rida and A. K . Smith (on the preparation of
ruthenium-rhodium bimetal complexes and their use in olejin hydroformylation)
and by M . A . Bennett and J . R. Harper (on tethered arene-metal complexes).
Gibson’s chapter also includes a brief historical commentary on the ‘distortional’
or ‘bond-stretch’ isomerism debate concerning green and blue forms of
[Mo(O)CE,(PMe,Ph),] initiated by Chatt, Manojlovic-Muir and Muir in 1971.
Some Notes on the Early
Development of Models of
Bonding in Olepn-Metal
Complexes
N. WINTERTON
Leverhulme Centre for Innovative Catalysis, Department of Chemistry,
University of Liverpool, Liverpool L69 7ZD, UK
1 Models of the Olefin-Metal Bond

In 1951, in a discussion' following his paper2 reviewing n-complex theory,
M. J. S. Dewar proposed a model (Figure 1)to describe the bonding of an olefin
to silver@)or copper@).The model suggested that, in addition to o-donation of
olefin n-bonding electrons to the metal, d, electrons on the metal would also
interact with antibonding orbitals of n-symmetry on the olefin. No experimental
evidence was provided to support this proposal nor, indeed, was explicit mention
made of Zeise's salt or other platinum-olefin complexes. From a study of
Chemical Abstracts 41 (1947)-75 (1971), it would appear that Dewar did not
follow up his proposal with more detailed studies, possibly indicating that this
was not a prime focus of his own interests in M O theory and its application to
Figure 1 Model of bonding of olefins with silver(1) proposed by M . J . S. Dewar (takenfrom

ref: 1 )
104 Early Development of Models of Bonding in Olejin-Metal Complexes
Figure 2 Valence bond description of bonding of olejins with silver(r) (takenfrom Winstein
and Lucas, reJ6)
organic ~hemistry.~" Indeed, in the brief discussion3bof bonding in metal-olefin

complexes in his classic text3"(published in 1969),he cites only ref. 1 and no later
theoretical or experimental work (either his own or that of other workers) on this
topic.
The interactions of unsaturated hydrocarbons and metal ions, particularly
silver(I),was an active area of research and had been reviewed by Keller4 in 1941.
A highly relevant paper5 appeared in the same year (though not in time to be
included in Keller's review) in which Taufen, Murray and Cleveland provided
convincing spectroscopic evidence of olefin coordination to silver (albeit in
solution) in a mode that was consistent with the valence-bond formulation
(Figure 2) proposed by Winstein and Lucas.6 Such evidence would also have
supported Dewar's own formulation, though the paper by Taufen et al. was cited
in neither Dewar's Bull. SOC.Chim. Fr. paper2 nor in the subsequent discussion.'
However, it should be noted that the discussion section is not supported by
references. If Dewar was aware of Taufen et al.'s work and its significance it is
odd that he made no mention of it.
Chatt and Duncanson, who developed Dewar's proposal in their seminal
paper7**of 1953 on metal-olefin complexes, do cite the paper of Taufen et aL5 (as
well as of Dewar), though Chatt had not done so in an earlier relevant paper of
1951.9'10Precisely when Chatt became aware of Taufen's work in not known.
While ref. 7 has been cited more than 500 times since 1981, Dewar's paper2
about 250 times and the associated discussion' more than 200 times during the
same period, the Taufen et al. paper, ref. 5, has been cited only seven times, an
astonishing result bearing in mind the paper's significance. This work indeed
merits wider recognition.
While it is surprising that Dewar and Chatt may not immediately have
become aware of a paper from such a well known journal on its publication in
1941 (and, apparently, were unaware of it ten years later), it must remembered
that sea-borne communication at the critical time (1941 onwards) was subject to
wartime hazards, and the material may only have become more generally known
after the war.
According to an interview,'' Chatt had become interested in olefin complexes
long before joining ICI in 1947 and had been fascinated by the suggestion by
Anna Gel'manI2 that the metal-olefin bond involved four electrons, including
two d electrons on platinum as well as the n-electrons of the olefin. Gel'man
considered 'unsaturated molecules . . . as acceptors and donors at the same time',
though in the model the metal was formulated as platinum(1v).
Early on, Chatt had not been able to find evidence of interaction between
trimethylborane and ethylene.'*1° He concluded from this indirect experimental
N . Winterton 105
evidence that simple donation of electrons from the olefin to a transition metal
was insufficient to be responsible for bonding of an olefin to a transition-metal
ion and that bonds involving metal d orbitals were required, similar to those
invoked in models developed for transition metal-CO and -NO bonding.
Chatt had first suggested in a note13,14 (though without new supporting
experimental detail) that olefin coordination in complexes of the sort
K[PtCl,( RCH=CH,)] involved coordination of an ethylidene moiety (Figure 3),
drawing analogies with the role of d orbitals on platinum in 7c-bonding to the
carbon of coordinated CO. After discussions with C. K. Ingold," Chatt was
persuaded to have another look at the structure and this proposal was subse-
quently abandoned.
Concurrently, Chatt pursued other highly productive studies on the constitu-
tion, structure and bonding of related complexes, developing ideas (such as the
origin of the trans effect) concerning the role of n-bonding of organophosphine,
-arsine and -stibine and carbon monoxide ligands proposed earlier. During this
period, with Williams, Chatt' '316 made an important contribution by demon-
strating that the role of d orbitals in bonding might be enhanced by the presence
of strongly electronegative groups attached to the ligating atom and this was
demonstrated by the synthesis and isolation of platinum complexes of phos-
phorus trifluoride, a development also reported at about the same time by
Wilkinson.17 The suggestion had originally been made by Chatt in 1949 in a
notel8.l9in Nature.
As Chatt acknowledges in the abstract of his paper with Duncanson7(and had
also noted in a paper presented to a conference in March 195220)only the Dewar
model of olefin coordination was supported by his new infra-red evidence
adduced for platinum complexes. As Chatt's interview" with Leigh reveals, this
change came about following conversations with Dewar, who pointed out to
Chatt the discussion section' following the Bull. SOC. Chim. Fr. paper,2 which
Chatt had not seen. Chatt's particular contribution was to develop the model,
providing both the basis for predictions regarding its structural consequences
(orientation of the C-C group with respect to the plane defined by the metal and
the other donor atoms, changes in C-C bond length change, state of hybridisa-
tion at C and C-C-H bond angle changes) as well as experimental and spectro-
scopic evidence in its support. The Figure used in ref. 7 (Figure 4) has been widely
reproduced. A crystal structure study of Zeise's salt, published in 195421with a
Figure 3 Early proposal by 1.Chatt of bonding between ethylene and platinum(r1) in

Zeise's salt (takenfrom reJ: 13)
106 Early Development of Models of Bonding in Olejn-Metal Complexes
Orbitals used in the combi.utation of ethylew

with filafinzmt.
8d - Type bond 7r-Type bond
Figure 4 M O description of bonding between ethylene and platinum(r1) in Zeise’s salt by

Chatt and Duncanson (takenfrom ref: 7 )
correction in 1955,22first2’ corrects Chatt’s structural inferences in ref. 13 and

then offers support22to the Chatt-Duncanson proposal in ref. 7 . The key point
for Chatt, as he says, was that, ‘this was t h e j r s t time that any structure for the
olejn compounds had been suggested and reasonably proven. Of course, there had
been many others, all suggestions, as listed in R. N . Keller’s review4 and in a review
by A . D. German, none proven any more than had been Dewar’s proposed silver
ion-olejin structure’. This comment may be considered questionable in the light
of the Taufen paper, though, at that particular time, compared with today, the
weight given to evidence for proposed structures of species in solution may have
been much less than that given to those of stable isolated solids. There had also
been a brief report in 1950 of the crystal structure of [Ag(C6H6)]C104,23though,
at the time, Chatt was not convinced of the relevance of such structures to
bonding in transition metal-olefin complexes.
However, there is no doubt that Chatt always acknowledged Dewar’s vital
24,25
role, as he did in the key paper with Duncanson7 and
So how, since this time, have chemists viewed the relative contributions of
Dewar and Chatt to what has become known26 as the Dewar-Chatt-Duncan-
son model (a term to which Dewar subsequently took exception27and one which
Chatt himself apparently never used)?
The crystal structures of Zeise’s salt21,22and of two analogous pallad-
ium-olefin complexes were p ~ b l i s h e d , shortly
~ ~ , ~ ~after the publication of the
Chatt-Duncanson paper. They confirmed the structural proposals made by
Chatt, but none of these papers cites Dewar,1’2 though metal-olefin bonding
models were not discussed. Two short reviews on the history of Zeise’s ~ a l t , ~ ” ~ ’
(one31part of a more general discussion of the history of organometallic chemis-
try) also only refer to Chatt’s contribution, though neither specifically address
questions of bonding.
Emeleus and Anderson32 discuss both the structure and bonding in plati-
num-olefin complexes citing explicitly Chatt and Duncanson’s 1953 paper and
N . Winterton 107
their infra-red studies (as well as the crystallographic studies21,22,28,29 ). They
reproduce a version of the Chatt-Duncanson (and not Dewar’s original) MO
description. Dewar’s paper is cited at the end of this section, curiously prefaced
with ‘cf’.
On the other hand, in an organic text,,, in which the role of transition
metal-olefin complex intermediates is reviewed,34 the bonding scheme is de-
scribed (in a section heading) as ‘Dewar’s M O picture’, which, in later text, was
‘soon extended by Chatt and Duncanson to platinum complexes.. .’.
in 1960 writes ‘The geometry ofthe olefin-metal system.. . is in accord
with the type of bonding discussed by Chatt and Duncanson following the proposal
by Dewar’. In Cotton’s review the work of Taufen et aL5 is given due prominence. ~
Guy and S h a ~in, ~ 1962 also indicate the prior contribution of Dewar, though
forbear from characterising the scheme as the Dewar-Chatt-Duncanson model.
Similarly, Green37 in 1968 says beneath an MO diagram, ‘. . .description
originally proposed by Dewar and in a modijied form by Chatt’.
The various editions of Cotton and W i l k i n ~ o n ,deal~ with the matter in
different ways: in the first and second editions the model is attributed to Dewar
alone, though Chatt and Duncanson’s spectroscopic work is cited; the third
edition refers to the Dewar-Chatt description; in the fourth edition, Dewar only
is mentioned.
Later r e v i e ~ s ~ ~published
g~’ in the 1970s and a key paper4’ dealing with a
neutron diffraction study of Zeise’s salt continue to refer to the Dewar-
Chatt-Duncanson model. Even very recent such as the 13C and 2H
NMR study of Zeise’s salt in the solid state, cites both Dewarl and Chatt and
Dun~anson.~
An interesting final perspective on the foregoing is provided by the theoretical
work of Bohme, Wagener and Frenking4, (elaborated further by Frenking in the
next chapter44) who conclude from their calculations that bonding in
{Cu(C,H,)+) is better described by an electrostatic model rather than one
involving back bonding. Some authors have reached similar conclusions for
{Ag(C2H4)’). However, there continues to be much discussion on this point (not
reviewed here but see refs. 45 and 46). As in the case of Chatt’s ideas concerning
the n-acidity of empty 3d orbitals on the phosphorus of PY, bound to a suitable
transition metal, which more recent ~ o r k has ~ ~shown
y ~ possibly
~ to be less
important than interactions involving empty o* orbitals of PY,, it now seems
possible that the systems about which Dewar speculated’ are not the best
exemplars of back-bonding involving alkenes. However, in both cases the earlier
thinking created the opportunity for improved models to be developed and
refined.
In my view, Greenwood and E a r n ~ h a wsummarise~~ matters most appro-
priately: ‘The key to our present understanding of the bonding in Zeise’s salt and all
other alkene complexes stems from the perceptive suggestion by M . J . S . Dewar in
1951 that the bonding involves electron donation from the n-bond of the akene into
a vacant metal orbital of o symmetry; this idea was modijied and elaborated by
J . Chatt and L. A . Duncanson in a seminal paper in 1953 and the Dewar-
Chatt-Duncanson theoryforms the basis of most subsequent discussion’.
Perhaps, this is how history, finally, should judge this important joint contri-
108 Early Development of Models of Bonding in OleJin-Metal Complexes
bution of theory and experiment from Dewar and Chatt and those that followed.
2 Addendum
After this brief survey was written, Seyferth” published a further perspective on
Zeise’s salt, reviewing again much of the early discussions of its synthesis and
disputes about its composition and formulation. In a section devoted to the
development of bonding models and more recent structural studies, Seyferth
cites an earlier review by 0 1 s s 0 n ~that
~ covers much of the discussion of the
present paper, and reaches a broadly similar conclusion concerning Chatt’s
contribution in developing Dewar’s earlier suggestion. I thank L. I. Elding for
providing a copy of Olsson’s publication. I also gratefully acknowledge receipt
from Professor D. M. P. Mingos of a copy of a manuscript prior to p ~ b l i c a t i o n . ~ ~
3 References
1 Discussion, p.C79, following M. J. S. Dewar, Bull. SOC.Chim. Fr., 1951, C 71.
2 ‘A Review of n-Complex Theory’, M. J. S. Dewar, Bull. SOC.Chim. Fr., 1951, C 71.
3 (a) M. J. S. Dewar, The Molecular Orbital Theory of Organic Chemistry, McGraw-Hill,
1969; (b) pp. 349-353; 358-359.
4 ‘Coordination Compounds of Olefins with Metals’, R. N. Keller, Chem. Reu., 1941,28,
229.
5 ‘Effect of Silver Ion Coordination upon the Raman Spectra of Some Unsaturated
Compounds’, H. J. Taufen, M. J. Murray and F. F. Cleveland, J . Am. Chem. SOC., 1941,
63, 3500.
6 ‘The Coordination of Silver Ion with Unsaturated Compounds’, S. Winstein and H. J.
Lucas, J . Am. Chem. Soc., 1938,60,836.
7 ‘Olefin Coordination Compounds. Part 111. Infra-red Spectra and Structure: Attem-
pted Preparation of Acetylene Complexes’, J. Chatt and L. A. Duncanson, J . Chem.
SOC., 1953,2939.
8 The original manuscript for reference 7 was published internally by ICI as Butterwick
Research Laboratories Report No. BRL/230, March 1953.
9 ‘Olefin Coordination Compounds. Part 1. Discussion of Proposed Structures. The
System Ethylene-Trimethylborine’, J. Chatt, J . Chem. SOC.,1949,3340.
10 The original manuscript for reference 9 was published internally by ICI as Butterwick
Research Laboratories Report No. BRLj81, June 1949.
11 ‘An Interview with Professor Joseph Chatt’, G. J. Leigh, first published in Coord.
Chem. Rev. 1991,108, 1, taken from A Celebration oflnorganic Lives, Elsevier, 2000.
12 ‘Contribution to the Theory of Complex Compounds’, A. Gelman, Compt. Rend.
Acad. Sci. URSS, 1939,24, 549.
13 ‘A New Structure for Olefin Coordination Compounds’, J. Chatt, Research, 1951,4,
180.
14 The original manuscript for reference 13 was published internally by ICI as Butter-
wick Research Laboratories Report No. BRL/l29,24 November 1950.
15 ‘The Nature of the Coordinate Link. Part IV. Complex Formation by Phosphorus
Trifluoride’, J. Chatt and A. A. Williams, J . Chem. Soc., 1951, 3061.
wick Research Laboratories Report No. BRL/150,19 June, 1951.
N . Winterton 109
17 ‘The Preparation and Properties of TetrakistribromophosphineNickel and Tetrakis-
trifluorophosphine Nickel’, G. Wilkinson, J . Am. Chem. Soc., 1951,73,5501.
18 ‘The Coordinate Link in Chemistry’, J. Chatt, Nature, 1950,165,637.
wick Research Laboratories Report No. BRL/87, 12 October 1949.
20 ‘The General Chemistry of Olefin Complexes with Metallic Salts’, J. Chatt, Butterwick
Research Laboratories Report BRL/190, April 1952.
21 ‘A Note on the Crystal Structure of Zeise’s Salt’, J. A. Wunderlich and D. P. Mellor,
Acta Cryst., 1954,7, 130.
22 ‘A Correction and a Supplement to a Note on the Crystal Structure of Zeise’s salt’, J.
A. Wunderlich and D. P. Mellor, Acta Cryst., 1955,8, 57.
23 ‘Structure of the Silver Perchlorate-Benzene Complex’, R. E. Rundle and J. H. Goring,
J . Am. Chem. Soc., 1950,72,5337.
24 ‘A Visit to Copenhagen for a Symposium on Coordination Chemistry’, G. A. Gamlen,
L. M. Venanzi and J. Chatt, Butterwick Research Laboratories Report No. BRL/254,
December 1953.
25 ‘The Work of the Inorganic Chemistry Department at The Frythe’, J. Chatt, Butter-
wick Research Laboratories Report No. BRL/293, August 1954.
26 ‘Organo-Transition Metal Compounds and Related Aspects of Homogeneous Cata-
lysis’, B. L. Shaw and N. 1. Tucker, Chapter 53, Comprehensive Inorganic Chemistry,
Pergamon Press, 1973.
27 ‘Relationship between Olefinic n Complexes and Three-Membered Rings’, M. J. S.
Dewar and G. P. Ford, J . Am. Chem. Soc., 1979,101,783.
28 ‘The Crystal Structure of an Ethylene-Palladium Chloride Complex’, J. N. Dempsey
and N. C. Baenziger, J . Am. Chem. Soc., 1955,77,4984.
29 ‘The Crystal Structure of Styrene-Palladium Chloride’, J. R. Holden and N. C.
Baenziger, J . Am. Chem. Soc., 1955,77,4987.
30 ‘Historical Origins of Organometallic Chemistry. Part 1. Zeise’s Salt’, J. S. Thayer, 1.
Chem. Educ., 1969,46,442.
3 1 ‘Organometallic Chemistry: A Historical Perspective’, J. S. Thayer, Adv. Organomet.
Chem., 1975,13, 1.
32 Modern Aspects oflnorganic Chemistry, H. J. Emelkus and J. S. Anderson, Routledge
& Kegan Paul, London, Third Edition, 1960, pp. 307-310.
33 The Chemistry ofAlkenes,Ed. J. Zabicky, part of the series The Chemistry of Functional
Groups, Ed. S. Patai, Interscience, 1970.
34 ‘Alkene Complexes of Transition Metals as Reactive Intermediates’, J.-F. Biellman, H.
Hemmer and J. Levisalles, in The Chemistry ofAlkenes, Ed. J . Zabicky, Vol 2, Ch. 5,
pp. 215 et seq., Interscience, 1970.
35 ‘The Infra-red Spectra of Transition Metal Complexes’, F. A, Cotton, in Modern
Coordination Chemistry, Eds. J. Lewis and R. G. Wilkins, Interscience, New York,
1960, pp. 376 et seq.
36 ‘Olefin, Acetylene and p-Allylic Complexes of Transition Metals’, R. G. Guy and B. L.
Shaw, Adv. Inorg. Chem., 1962,4,77.
37 Organometallic Compounds, 3rd Edn., Vol. 2, The Transition Elements, M. L. H. Green,
Methuen & Co., London, 1968, p. 17.
38 Advanced Inorganic Chemistry, F. A. Cotton and G. Wilkinson, Interscience, New
York, 1st edn. 1962,2nd edn. 1966,3rd edn. 1972,4th edn. 1980.
39 ‘Coordination of Unsaturated Molecules to Transition Metals’, S. D. Ittel and J. A.
Ibers, Adv. Organometal. Chem., 1976,14, 33.
40 ‘Recent Developments in Theoretical Organometallic Chemistry’, D. M. P. Mingos,
110 Early Development of Models of Bonding in Olefin-Metal Complexes
Adv. Organometal. Chem., 1977,15, 1.
41 ‘Neutron Diffraction Study of the Structure of Zeise’s Salt, KPtCl,(C,H,).H,O’, R. A.
Love, T. F. Koetzle, G. J. B. Williams, L. C. Andrews and R. Bau, Inorg. Chem., 1975,
14, 11.
42 ‘A Carbon-13 and Deuterium NMR Investigation of Solid Platinum-Ethylene Com-
plexes: Zeise’s Salt and Pt(q2-C2H4)(PPh,),)’,G. M. Bernard, R. E. Wasylishen and A.
D. Phillips, J . Phys. Chem., A , 2000,104, 8131.
43 ‘Copper-substituted ethanes as a model for copper-acetylene interactions on the metal
surface. Quantum mechanical study of the structure and bonding of copper-acetylene
and copper-ethylene compounds Cu,(C,H2) (n = 1, 2, 4), Cu(C2H2)+,Cu,(C,H,)
(n = 1, 2) and Cu(C2H4)+,M. Bohme, T. Wagener and G. Frenking, J . Organomet.
Chem., 1996,520, 31.
44 G. Frenking, this volume pp. 1 1 1-122.
45 ‘The bonding strengths of Ag+(C2H4)and Ag+(C,H,), complexes’, B. C. Guo and A.
W. Castleman, Chem. Phys. Lett., 1991,181, 16.
46 ‘A Comparative Computational Study of Cationic Coinage Metal-Ethylene Com-
plexes (C2H4)Mf (M = Cu, Ag, Au)’, R. H. Hartwig, W. Koch, D. Schroder, H.
Schwarz, J. Hrusak and P. Schwerdtfeger, J . Phys. Chem., 1996,100,12253.
47 ‘n-Accepting Abilities of Phosphines in Transition Metal Complexes’, D. S. Marynick,
J . Am. Chem. SOC.,1984,106,4064.
48 ‘Structural Evidence for the Participation of P-X r ~ * Orbitals in Metal-PX, Bonding’,
A. G. Orpen and N. G. Connelly, J . Chem. SOC.,Chem. Commun., 1985,1310.
49 Chemistry o f t h e Elements, N. N. Greenwood and A. Earnshaw, Pergamon Press,
Oxford, 1984, p. 360.
50 ‘[(C2H4)PtC13]-, The Anion of Zeise’s Salt, K[(C,H4)PtC1,].H20’, D. Seyferth,
Organometallics, 2001,20,2.
51 ‘Zeise’s Salt - from Curiosity to Model Substance’, L. F. Olsson, in William Chris-
topher Zeise - en dansk kemiker, ed. T. Morsing, Dansk Selskab for Historisk Kemi,
Historisk-kemiske skrijter nr. 2, 1990, 57-72.
52 ‘A Historic Perspective of Dewar’s Landmark Contribution to Organometallic Chem-
istry’, D. M. P. Mingos, J . Organornet. Chem., submitted for publication.
The De war- Chatt-Duncanson
Bonding Model of Transition
Metal-Olepn Complexes
Examined by Modem Quantum
Chemical Methods
GERNOT FRENKING
Fachbereich Chemie, Philipps-Universitat Marburg, Hans-Meerwein-Strasse,
D-35032 Marburg, Germany
1 Introduction
There is probably no bonding model that has proven to be more useful for
explaining metal-ligand interactions in transition metal (TM) complexes than
the suggestion which was originally made by Dewar to describe the bonding of
'
an olefin coordinated to silver@ and copper($ Chatt and Duncanson2 realised
that this model and their spectroscopic data of Zeise's salt gave a coherent picture
and, for the first time, an understanding of the bonding of a compound which had
been a puzzle for chemists ever since it was accidentally synthesised in 1825.3
Figure l a shows a schematic representation of the Dewar-Chatt-Duncanson
(DCD) model. The pivotal idea is that the olefin serves as a donor and an
acceptor at the same time. There is ligand + metal donation and metal + ligand
back-donation. The former interaction involves a donor orbital of the ligand
which has n symmetry in the free ligand but c symmetry in the complex. The
metal acceptor orbital is mainly the d,, orbital of the metal. Quantum chemical
calculations have shown that the valence s orbital of the metal is less important
as an acceptor orbital than the d,, 0rbita1.~The metal -+ ligand back-donation
takes place via a d(n) orbital of the metal and the n* orbital of the olefin.
The picture shown in Figure l a has become a standard model in textbooks of
inorganic and organometallic chemistry to explain the bonding in TM com-
pounds. However, it was soon realised that there are TM compounds with olefin
112 The Dewav-Chatt-Duncanson Bonding Model of Transition Metal-OEeJin Complexes
H H
o TM
g
c-
G
H H
TM I
Figure 1 Schematic representation of (a)the DCD model for TM-olejn bonding and (b)a
metallacyclopropane
ligands which would better be described as metallacyclopropanes as shown in

Figure lb. The latter compounds have two electron-sharing covalent5
metal-carbon bonds instead of a donor-acceptor bond. The dichotomy of two
bonding modes, i.e. donor-acceptor bond and electron-sharing covalent bond,
has also become a very helpful device for the understanding of T M complexes
with other ligands such as alkynes, carbenes and carbynes6
The DCD model has been used for decades as a heuristic device to explain the
structures and properties of T M complexes without the nature of the
metal-ligand interactions having been examined by theoretical methods. The
progress in quantum chemical methods in the last decade has made it possible for
calculations on T M compounds to be carried out with high a c c ~ r a c y The .~
electronic structure can then be analysed with modern charge and energy par-
titioning methods.498A charge-partitioning method has been developed with
respect to the DCD model which makes it possible to quantify the amount of
ligand -+ metal donation and metal -+ ligand b a c k - d ~ n a t i o n Application
.~ of
charge decomposition analysis (CDA) to a wide range of T M compounds shows
that it can be used as a quantitative expression of the DCD Details of
CDA can be found in the l i t e r a t ~ r e . ~ ~ ~ ~ ~ ' *
In this paper we present a summary of three theoretical investigations of T M
olefin complexes, which show that the original bonding model of Dewar' and its
ingenious application to metal-olefin interactions by Chatt and Duncanson2
Gernot Frenking 113
stand the test of examination by quantum chemical methods. The strength of the
DCD model comes to the fore as a result of calculations which show that changes
in the donor or acceptor moiety of the complex induce concomitant changes in
the metal-ligand interactions and in the properties of the complexes, which can
be understood by the model. Analysis of the calculated data shows the scope but
also the limitation of the DCD bonding model. The calculations reveal examples
of olefin complexes which are not properly described by the DCD model. We will
present only results which are important in the context of this paper. For details
we refer the reader to the original publications.
2 Platinum Complexes of Strained Olefins

Olefins become more reactive when steric constraints enforce a nonplanar
geometry on the R,C=CR, moiety. A twisted or pyramidal geometry of the
olefin weakens the double bond. The energy level of the occupied n orbital
increases while the unoccupied n* orbital becomes lower in energy. The change
in the energy levels of the 7c and n* orbitals should enhance the metal-ligand
donation and back-donation and thus should strengthen the metal-olefin bond.
We studied this effect by calculating the structures and bond energies of the
platinum complexes [PtL(PH,),] where L is ethylene or a strained tricyclic
olefin as shown in Figure 2.l'
The strain in the olefin moiety of the tricyclic compounds becomes larger when
the chain length rz of the bridging (CH,), group becomes smaller. This becomes
evident from the calculated pyramidalisation angle 8 and the C=C bond lengths
of the free olefins which are shown in Table 1. The calculations also show that the
C=C distances of the olefinic moiety become longer in the complexes and that
the pyramidalisation angle 8 becomes bigger than in the free olefin. The increase
H3P PH3 H,P PH3
1 2 ( n = 3)
3 ( n = 2)
4 ( n = 1)
5 ( n = 0)
Figure 2 Calculated platinum olefin complexes 1-5 and dejinition of the pyramidalisation
angle 0
114 The Dewar-Chatt--Duncanson Bonding Model of Transition Metal-OleJn Complexes
Table S Theoretically predicted pyramidalisation angles 6 (degrees)and C=C
bond lengths (A)of the platinum complexes 1-5 and the free olefins.
Calculated metal-olefin binding energies D, (kcal mol-I). Results of the
CDA calculations of ligand + metal donation d , metal + ligand
back-donation b and ratio d/b"
Molecule Sym. 8(A8)b C=C(AC=C)b D, d b d/b
1.427 (0.096) 99.9 0.511 0.383 1.33

1.446 (0.104) 147.1 0.477 0.396 1.20
1.460 (0.111) 200.3 0.498 0.429 1.16
1.480(0.118) 244.7 0.504 0.460 1.10
1.513(0.133) 293.5 0.517 0.500 1.03
1.331
1.342
1.349
1.362
1.380
aTaken from reference 11
bThe values in parentheses give the differences between the values in the complex and
the free olefin
of the C=C distance upon complexation has the same trend as the bond dissocia-
tion energies of the Pt-L bonds S < 2 < 3 < 4 < 5 (Table 1).This means that the
change in the carbon-carbon bond lengths of the coordinated olefin reflects
nicely the bond strength. We want to point out that this holds only within a series
of compounds which has the same kind of bonding interactions. It will be shown
below that the change of geometry does not always correlate with the bond
strength because different types of metal-olefin bonds may have a different
influence on the geometry. Note that the further increase of the pyramidalisation
angle 8 of the olefin ligands in the complexes becomes less for the more strongly
strained ligands (Table 1).
What about the correlation of the DCD model of metal-olefin bonding with
the calculated changes of the geometries and the bond dissociation energies?
Table 1 also shows the CDA results of the olefin + Pt(PH,), donation d and the
(PH,),Pt + olefin back-donation b. We wish to point out that the absolute
values of the donation and back-donation have little meaning. It is the ratio d/b
of the two terms which should be used for a comparison of different complexes or
different ligands. Table 1 shows that the ethylene complex 1 has d/b = 1.33, i.e.
the calculated charge donation is larger than the back-donation. The d/b values
of the other olefin complexes decrease in a regular fashion with the trend
1 > 2 > 3 > 4 > 5. It means that the (PH,),Pt + olefin back-donation becomes
more important when the coordinated olefin is more strained. It follows that the
energy lowering of the LUMO of the olefin is more important for increasing the
bond energy than the energy increase of the HOMO. Figure 3 shows a diagram
Gernot Frenking 115
1.35 7 1
-
1.3
1.25 -
1.2 -
e
1.15 -
-0
1.1 -
1.05 -
I . I ~~ -1
45 50 60 65 70
55 0
Figure 3 Plot of the d/b ratio of olejin -+Pt donation and Pt -+ olejin back-donation
against pyramidalisation angle 0 (degree)of compounds 1-5
Reproduced with permission from reference 11
1.35 - 1
1.3 -
1.25 -
1.2 -
% 1.15 -
1.1 -
1.05 -
I ,
30 40 50 60
D, , kcal mol-'
Figure 4 Plot of the d/b ratio of olejin -+ Pt donation and Pt + olejin back-donation
against the bond dissociation energies D , (kcal mol-l) of compounds 1-5
Reproduced with permission from reference 1 1
where the d/b ratio of 1-5 given by the CDA is correlated with the pyramidalisa-
tion angles 6. Figure 4 shows a plot of the calculated bond energies and the d/b
ratio. The diagrams nicely demonstrate that partitioning of the electronic struc-
ture of molecules in terms of the DCD model yields a quantitative correlation
between the d/b ratio of the metal-ligand interactions and the geometries and
bond energies of the compounds.
3 Olefin Complex versus Metallacyclopropane

The dichotomy of the two bonding models which are sketched in Figure 1 has
been investigated by us with the help of CDA calculations using the model
compounds [W(CO),(C,H,)] (6)and [WCl,(C,H,)] (7) as examples.8a~'2Figure
5 shows the calculated geometries and the theoretically predicted bond dissocia-
116 The Dewar-Chatt-Duncanson Bonding Model of Transition Metal-Olejin Complexes
D, = 12.1 kcal mol-'
D, = 41.4 kcal mol-*
8
1.166
Figure 5 Calculated geometries and W-C,H, bond dissociation energies (kcal mol-') of
[ WCl,(C,H,)] 7 and [( W(CO),(C,S,)] 6. Bond lengths are given in A
Reproduced with permission from reference 12b
tion energies of the molecules. It becomes obvious that the W-C distance of 7
(2.103 A)is substantially shorter than in 6 (2.372 A).The shorter W-C bond
length of the former compound suggests stronger metal-ethylene interactions
than in the latter molecule. Stronger tungsten-ethylene bonding is also indicated
by the C-C distances in the two compounds. The C-C bond length in 7 (1.459 A)
is clearly longer than in 6 (1.402 A). A direct proof for stronger metal-ligand
interactions in the former complex comes from the topological analysis of the
electron density distribution.13 Figure 6 shows the contour line diagrams of the
Laplacian V2p(r) in the plane which contains the WC, moiety. Areas of charge
concentration (V2p(r) < 0) are depicted with solid lines while areas of charge
depletion (V2p(r) > 0) are drawn with dashed lines. Visual inspection of the
shape of the Laplacian shows that the electronic charge of the ethylene ligand in
7 is much more distorted than in 6 . There is a much larger area of charge
Gernot Frenking 117
I \ J
Figure 6 Contour line diagrams of the Laplacian distribution V2p(r) of (a) [ WCl,(C,H,)]
7 and (b) [ W(CO)5(C2H4)] 6 in the plane containing the W C , moiety. Dashed
lines indicate charge depletion (V2p(r) > 0 )and solid lines indicate charge
concentration (V’p(r) < 0). The solid lines connecting the atomic nuclei are the
bond paths, and the solid lines separating the atomic nuclei indicate the zero-flux
surfaces in the plane. The crossing points of the bond paths and the zero-Jux
surfaces are the bond critical points
Reproduced with permission from reference 8a
concentration at the ligand carbon atoms in the former compound pointing

toward the tungsten atom than in the latter species.
The calculated W-C,H, bond dissociation energies (BDEs) do not correlate
with the geometrical parameters. The compound 7 has a surprisingly low BDE
D, = 12.1 kcal mo1-l while 6 has a much stronger bonded ethylene ligand with
D, = 41.4 kcal mol-’. What is the reason for the much stronger metal-ethylene
interactions in the former compound leading to such a low BDE? It is important
to recognise that strong electronic interactions do not necessarily mean a strong
118 The Dewar-Chatt-Duncanson Bonding Model of Transition Metal-OleJin Complexes
bond in a thermodynamic sense. The BDE is the energy difference between the
molecule and the fragments in the electronic ground states, which may not be the
electronic reference state in the molecule. The analysis of the electronic structure
in terms of the two bonding models shown in Figure 1 did not give only an
understanding of the bonding situation in the two ethylene complexes. It also
gave a plausible explanation for the low BDE of 7. However, we must give a short
outline of the CDA in order that the results of the calculations can be under-
stood.
The CDA method considers the bonding in a complex in terms of fragment
molecular orbital interactions between two closed-shell fragments. In the present
case, the fragments are {W(CO),} and C2H4 for 6 and {WCl,} and C2H4 for 7 .
The mixing of the occupied orbitals of C2H4 and the unoccupied orbitals of
{W(CO),} or {WCl,} gives the electron donation d. The mixing of the unoc-
cupied orbitals of C2H4and the occupied orbitals of {W(CO),} or {WCl,} gives
the back donation b. The mixing of the occupied orbitals of the two fragments
gives the repulsive polarisation r. The mixing of the vacant orbitals of the
fragments gives the residual term A. It should be zero because mixing of unoc-
cupied orbitals cannot physically contribute to the electron density. It was found
that the residual term A is a sensitive probe which shows if the electronic
structure of a compound can reasonably be described by donor-acceptor inter-
actions of the chosen fragments. For example, the electron density of ethylene
cannot be described by interaction of two methylene fragments in the ‘A, state
but only in the 3B, state. A CDA of ethylene using (‘Al) CH2 as building blocks
gives values for A which are very large. Thus, a significant deviation of A = 0
beyond numerical noise indicates that the compounds cannot be described by
the interactions of the chosen fragments.
Table 2 shows the CDA results of 6 and 7. The data suggest for 6 that ethylene
is a stronger donor than acceptor. The total amounts of donation and back-
donation are larger than in [Pt(C2H,)(PH3),] (Table 1) but the d/b ratio of the
two complexes is similar. The calculated value for A is close to zero which
indicates that the bonding situation in 6 can be discussed in terms of donor-
acceptor interactions between {W(CO),) and C2H4. The results for 7 are very
different. The calculated amounts of donation and back-donation are negative,
which is a physically unreasonable result. The same conclusion comes from the
calculated residual term A = 0.351. The CDA results clearly show that the
bonding situation in 7 should not be discussed in terms of donor-acceptor
Table 2 Charge decomposition analysis of [ WCl,(C,H,)] (7)

and [ W(CO),C2H,I (6)”
d -0.263 0.225
b -0.194 0.148
r -0.318 - 0.422
A 0.351 - 0.025
aTaken from reference 8a
Gernot Frenkiny 119
interactions between the closed-shell fragments (WCl,) and C2H4. The latter
compound should rather be considered as a metallacyclopropane where the
W-C bonds arise from the electron-sharing interactions between two open-shell
fragments. This explains why the BDE of the W-C,H, bonds are so low.
Ethylene has a closed-shell electronic ground state. In order to promote C,H,
from the ground state to the triplet excited state which is the reference state for
the binding interactions with (WCl,} (which has a triplet ground state) a large
amount of excitation energy is necessary. The singlet + triplet excitation of
ethylene is -100 kcal mol-l. Thus, the interaction energy between (WCl,} and
C2H4 is much higher (D,= 12 kcal mol-' plus the excitation energy) than the
-
BDE. The value of 112 kcal rno1-l correlates nicely with the calculated bond
lengths of 7.
4 Electrostatically-bound Olefin Complexes

The olefin complexes discussed above could either be classified as belonging to
the DCD bonding scheme or metallacyclopropanes. Now we want to introduce
yet another type of strongly bonded TM-olefin complex which is held together
mainly by electrostatic interactions.
In the course of a theoretical investigation of model compounds for ethylene
and acetylene bonded to copper atoms on a metal surface we calculated the
structures and bond energies of (Cu(C,H,)} and {CU+(C,H,)}.'~Figure 7 shows
the optimised geometries of the molecules. The C-C distance is slightly longer
than the calculated value of free ethylene (1.336A). We want to point out that the
carbon-carbon bond length and the pyramidalisation of the ethylene ligand in
(Cu(C,H,)} and {Cuf(C2H,)>are nearly the same and that the Cu-C distance of
the latter compound (2.095 A) is only slightly shorter than in the former molecule
(2.122 A). However, the positively charged species has a much higher BDE
D, = 43.9 kcal mol-' than the neutral compound which has only BDE = 4.2
kcal mol-l. The weak copper-ethylene bond of {Cu(C,H,)) suggests that the
compound should be considered as van der Waals' complex. But what about the
bonding in {Cu+(C,H,)}?
Figure 7 Calculated interatomic distances ( A )and bond dissociation energies D,(kcal

mol-') of {Cu(C,H,)} and {Cu+(C2H4)}
Reproduced with permission from reference 11
120 The Dewar-Chatt-Duncanson Bonding Model of Transition Metal-OleJn Complexes
Figure 8 Contour line diagrams of the Laplacian distribution V2p(r)of (a){ Cu(C2H4))and
(b){ C u + ( C 2 H 4 ) )
Reproduced with permissionfrom reference 1 1
Figure 8 shows the Laplacian distribution of {Cu(C,H,)} and (Cu+(C,H,)}.

Visual inspection of the shapes of the areas of charge concentration (solid lines)
and charge depletion (dashed lines) shows hardly any difference between the two
compounds. The most prominent difference is that in neutral {Cu(C,H,)} there
are bond paths between each C atom and Cu, besides the bond path between the
carbon atoms. There is also a ring critical point in the CuC, moiety of
{Cu(C,H,)}. Thus, the topological analysis defines the latter compound as a
cyclic molecule. The positively charged (Cu+(C,H,)}, however, does not have
bond paths which connect C u + with the carbon atoms. There is only a bond path
from Cu+ to the midpoint of the C-C bond. Thus, the topological analysis
suggests that (Cu+(C,H,)} has a T-shaped structure and that it should not be
considered as a cyclic compound. The bonding interactions arise from the charge
attraction between positively charged C u + and the negative charge accumula-
tion in the C-C bond region. The Natural Bond Orbital analysis15 of
{Cu+(C,H,)} shows that the copper atom carries a positive charge of + 0.91.
There is very little ethylene-+Cu+ charge donation in the compound.
{Cu+(C,H,)} is neither a donor-acceptor complex which can be described by
the DCD model nor a metallacyclopropane. Rather, it is a electrostatically
bonded species which is held together mainly by Coulomb attraction between
C u + and the n charge of ethylene.
5 Summary
The DCD bonding model of T M olefin complexes is supported by accurate
quantum chemical calculations which quantify the amount of ligand -+ metal
donation and metal -+ ligand back-donation. The geometries and bond energies
Gernot Frenking 121
of olefin complexes which can be described by the DCD model show a nice
correlation with the strengths of the donation and back-donation. Modern
charge partitioning schemes show that there are other types of olefin complex for
which bonding interactions are not well described by the DCD model. The
bonding in metallacyclopropanes comes from electron-sharing covalent bonding
between a metal fragment and the olefin which have electronic triplet states.
There are also compounds such as (Cu+(C,H,)) where the bonding between the
metal and the olefin is mainly due to electrostatic attraction. The different types
of TM-olefin complexes exhibit a different correlation between bond strength
and bond length.
6 Acknowledgements
The author thanks his co-workers who contributed to the publications which are
described in this work. This research has been supported by the Fonds der
Chemischen Industrie and the Deutsche Forschungsgemeinschaft.
7 References and Notes

1 M. J. S. Dewar, Bull. Soc. Chim. Fr., 1951, C79.
2 J. Chatt and L. A. Duncanson, J . Chem. Soc., 1953,2929.
3 a) W. C. Zeise, Overs. K . Dan. Vidensk. Selsk. Forh. 1825-26, 13; b) W. C. Zeise,
Poggendorfs Ann. Phys. Chem., 1831,21,497.
4 A. Diefenbach, F. M. Bickelhaupt and G. Frenking. J . Am. Chem. Soc., 2000,122,6449.
5 Donor-acceptor bonds may also have large covalent contributions. The difference
between the electron-sharing bond and the donor-acceptor bond, which is also
sometimes called a dative bond, is that the former bond dissociates homolytically and
the latter heterolytically. The bonding fragments of a donor-acceptor bond provide a
donor electron pair and an empty acceptor orbital while the bonding fragments of an
electron-sharing bond provide one electron each.
6 G. Frenking and N. Frohlich, Chem. Rev., 2000,100,717.
7 a) G. Frenking, I. Antes, M. Bohme, S. Dapprich, A. W. Ehlers, V. Jonas, A. Neuhaus,
M. Otto, R. Stegmann, A. Veldkamp and S. F. Vyboishchikov, in Reviews in Computa-
tional Chemistry, Vol. 8, K. B. Lipkowitz and D. B. Boyd (Eds), VCH, New York, 1996,
p. 63-144; b) T. R. Cundari, M. T. Benson, M. L. Lutz and S. 0.Sommerer, Reviews in
Computational Chemistry, Vol. 8, K. B. Lipkowitz and D.B. Boyd (Eds), VCH, New
York, p. 145-202,1996; c) M. Diedenhofen, T. Wagener and G. Frenking, in Computa-
tionul Organometallic Chemistry, T. Cundari (Ed), Marcel Dekker, New York, 2001,
p. 69-121.
8 a) G. Frenking and U. Pidun, J. Chem. Soc., Dalton Trans., 1997, 1653; b) N. Frohlich
and G. Frenking, in Solid State Organometallic Chemistry: Methods und Application,
M. Gielen and B. Wrackmeyer (Eds), Wiley-VCH, New York, 1999, p. 173-226; c) C.
Boehme, J. Uddin and G. Frenking, Coord. Chem. Rev. 2000,197,249; d) J. Uddin and
G. Frenking, J . Am. Chem. Soc., 2001,123,1683.
9 S. Dapprich and G. Frenking, J . Phys. Chem., 1995,99,9352.
10 a) S. Dapprich and G. Frenking, Organometallics 1996, 15, 4547; b) G. Frenking, S.
Dapprich, K. F. Kohler, W. Koch and J. R. Collins, Mol. Phys., 1996, 89, 1245; c) S.
Dapprich and Frenking, Angew. Chem., 1995, 107,383; Angew. Chem., Int. Ed. Engl.,
122 The Dewar-Chatt-Duncanson Bonding Model of Transition Metal-OleJin Complexes
1995,34,354; d) S. F. Vyboishchikov and G. Frenking, Chem. Eur. J., 1998,4,1439; e)
S. F. Vyboishchikov and Frenking, Chem. Eur. J., 1998,4,1428.
11 The calculations were carried out at B3LYP using a quasi-relativistic small-core ECP
with a DZP-quality basis set for Pt and 6-31G(d) for C and H. The bond energies were
approximated CCSD(T) values. For details see the original paper: J. Uddin, S.
Dapprich, G. Frenking and B. F. Yates, Organometallics, 1999,18,457.
12 The calculations were carried out the MP2 level using a quasi-relativistic small-core
ECP with a DZP-quality basis set for W and 6-31G(d)for C, 0 and H. An ECP with a
(31/31/1) valence basis set was employed for C1. The bond energies were calculated at
CCSD(T). For details see the original papers: a) U. Pidun and G. Frenking, Or-
ganometallics, 1995, 14, 5325; b) U. Pidun and G. Frenking, J . Organomet. Chem.,
1996,525,269.
13 R. F. W. Bader, Atoms in Molecules: A Quantum Theory, Oxford University Press,
Oxford, 1990.
14 The calculations were carried out at the MP2 level using a quasi-relativistic small-core
ECP with a TZP-quality basis set for Cu and 6-31G(d)for C and H. The bond energies
were calculated at CCSD(T)using larger valence basis sets. For details see the original
paper: M. Bohme, T. Wagener and G. Frenking, J . Organomet. Chem., 1996,520,31.
15 A. E. Reed, L. A. Curtiss, and F. Weinhold, Chem. Reu., 1988,88,899.
Cycloaddition Reactions with
Me talla-l,3-Dipoles
HANS-WERNER FRUHAUF
Institute of Molecular Chemistry, University of Amsterdam, Nieuwe
Achtergracht 166,1018 WV Amsterdam, The Netherlands
1 Introduction
Much of the work described in this article has previously been presented at
International Coordination Chemistry Conferences, including the latest piece
(Section 3.3.1), in the session on Joe Chatt Chemistry at the 34th ICCC in
Edinburgh, which resulted in the invitation to write this paper.
When the reader scans through this introduction and the rest of the paper, he
will find ligands such as olefins, acetylenes, or phosphines/phosphites. They
could be taken from a key-word list of the published work of Joseph Chatt and
make clear the connection to his seminal work.
The chemistry that is described in this account arose from an attempt to
prepare the first iron(0) y2-alkyne complex. In an investigation on [Fe(C0),(y2-
olefin)(1,4-diaza-1,3-diene)] complexes' it was found that the stability of the
iron-olefin bond increased strongly with increasing n-acidity of the olefin. So the
idea was, in analogy with the preparation of the olefin complexes, to substitute
photochemically a CO-ligand in [Fe(CO),(diazadiene)] for the supposedly very
good n-accepting dimethyl acetylenedicarboxylate (dmad). As it turned out, it
was not possible to irradiate the tricarbonyl complex in the presence of dmad,
because a very fast thermal reaction immediately consumed the reactants.
This was readily recognised from the immediate disappearance of the intense
colour of the [Fe(CO),(diazadiene)] complex on the addition of dmad.
All [Fe(CO),(diazadiene)] complexes have a very intense MLCT transition in
the visible region which, depending on the diazadiene, makes them intensely red,
violet (N,N'-disubstituted 1,4-diazabuta-1,3-dienes, dab), blue (2-pyridine-N-
aryl carbaldimines, R-Pyca), or green (bipy, o-phen). The fascinating chemistry
that arose from this initial observation will be summarised in this account.
124 Cycloadd it ion Reactions with M eta1la-1,3-Dipol es
2 1,3=DipolarCycloaddition of Activated Alkyne to the

Fe-N=C Fragment
2.1 The Reaction of [Fe(CO),(dab)] with Dimethyl
Acetylenedicarboxylate
The original reaction mentioned in the introduction is shown in Scheme 1. In the
'H NMR of the reaction mixture, two apparently isomeric products could be
identified in a ratio of ca. 9: 1. In an attempt to separate them by column
chromatography, the isomer ratio was inverted to 1: 9. From this it was obvious
that 4 was an intermediate and 5 was the final product. Compound 5 could be
isolated and fully characterised, including by X-ray structural analysis.2 So it was
concluded that the isomerisation consisted of a reductive elimination to form the
1,5-dihydropyrrol-2-0ne and recoordination of its olefinic double bond.
2 3
1
1
+L'
Me09C.
5 4
Scheme 1
The thermally labile intermediate 4 could also be isolated (and was character-
ised by X-ray analysis3) when the reaction was performed in the presence of
trimethyl phosphite L' in place of carbon monoxide. As a consequence, the metal
becomes more electron-rich and reductive elimination does not occur. With the
structure of 4 being firmly established, the pathway shown in Scheme 1 was
proposed and it was realised that the Fe-N=C fragment resembles a 1,3-dip0le.~
By 1 3 C 0labelling studies5 and from the X-ray structure of 4 with L' = P(OMe),
it was shown that the additional ligand L' is always stereospecifically incorpor-
ated trans to the inserted carbonyl. The initial bicyclo[2.2.1] cycloadduct 2 could
not be observed even at very low temperatures - evidently because C O insertion
is faster than its formation. Later, when the 1,3-dipole was modified (cJ: Section
3), stable analogues of 2 could be obtained. The crystal and molecular structures
of homologues of 2 could be determined in the iron ~ y s t e m ,for ~ ' ~ruthenium7
and manganese.8
Huns- Werner Fruhauf 125
With the formation of 2, two interdependent stereocentres are formed, the
bridgehead iron and carbon atoms. The imine carbon atoms in 1 are prochiral,
and the alkyne can approach either of them from either the re- or the si-face.
With chiral substituents at the nitrogen atoms, the approach of one face may be
favoured over the other. This would lead to diastereoselectivity which has been
investigated with different types of diazadiene ligands and a series of chiral
N-substituents.' Depending on both the type of diazadiene (C, and non-C,
symmetric) and the chiral N-substituents, diastereoselectivities from 0 to > 99%
have been observed.
2.2 The Isolobal Relation of the Fe-N=C Fragment with an

Azomethine Ylide
Huisgen defined a 1,3-dip0le'~~"as a species that may be described (as in
Equation 1) by zwitterionic octet structures 6a 6b,* and which may undergo
t--)
cycloaddition reactions of the type 3 + 2 + 5 with suitable multiple bonds

(dipolarophile d=e) to give a neutral five-membered ring 7.
6a 6b 7
This general description of a 1,3-dipole can be easily recognised in the

iron-imine system with the aid of the isolobal relation 8 and 9 (Equation 2) and,
derived from it, 10 and 11 (Equation 3).
8 9
10 a 11
In the well-known isolobal a n a l ~ g y ' ~between

.'~ a d8 ML, fragment 8 and a
carbene 9, the four ligands L at the iron correspond to, e.g., the three C O ligands
and one of the two imine units in 1. If, by way of N-donation, an imine is added to
8 and 9, respectively, structures 10 and 11 are obtained, of which 10 is equivalent
to 1, while 11 resembles one resonance structure of an azomethine ylide, a
classical 1,3-dipole. When the relevant atom arrangement in 10 is substituted
into the general formula 6a 6b of a 1,3-dipole, the resulting organometal
t)
* The formal 173-dipoleresults from localising two of the four electrons at the onium centre b',
which gives rise to the sextet structures +a-b--c- H -a--b-c+. They demonstrate the ambivalence
of 1,3-dipoles7which may have both nucleophilic and electrophilic properties at each end.
126 Cycloaddition Reactions with Metalla-l,3-Dipoles
1,3-dipole 12a t) 12b in fact represents a valence-bond description of

the dab. Earlier M O calculations on the model compound
[Fe(CO),(HN=CH-CH=NH)] l4 had shown that the frontier orbitals of 1 are
C=N-Fe based, i.e. both the HOMO and LUMO have strongly mixed metal d
and ligand rc* character.
12a 12b
From this it seemed reasonable to formulate the initial reaction step (Scheme
1) in terms of a 1,3-dipolar cycloaddition reaction of the alkyne across the
Fe-N=C fragment, resulting in the (unobserved) ferrabicyclo[2.2.1] intermediate
2 from which the isolated products 4 and 5 are readily derived.
3 Adding Variety to the 1,3-Dipolar System

The concept of 1,3-dipolar cycloadditions was supported by CAS-SCF calcula-
tions by Dedieu and Liddel16 on both the organic and inorganic species. Both
are nucleophilic dipoles with relatively high HOMO and LUMO energies, the
metalla- 1,3-dipoles at slightly higher energy. Cycloaddition reactions of nuc-
leophilic 1,3-dipoles are HOMO-controlled, i.e., the interaction of the dipole
HO M O with the dipolarophile LUMO is predominant, and the reactivity of the
dipole can be increased either by decreasing the HOMO-LUMO gap, and/or
increasing the H O MO energy.
Variations to the system, to explore their influence on reactivity, can be
brought about by changing the metal atom, the hetero atom, or the electronic
properties of the additional ligands on the metal.
3.1 Changing the Metal

3.1.1 Reactions of [Ru(CO),(dab)]
With ruthenium instead of the 1,3-dipolar reactivity markedly in-
creased. The same reaction pathway is followed, but the activation barriers
further along the reaction coordinate are different. The reaction of 13 with one
equivalent of dmad (Scheme 2)7proceeds instantly at - 78 "Cand, in contrast to
the iron system, the bicycloC2.2.11 adduct 14 is stable at that temperature. After
protonation of the amido nitrogen with HBF,, which inhibits CO-insertion, 16 is
stable at room temperature, and its X-ray structure could be determined. On
warming the solution of 14 to room temperature in the presence of CO or PPh,,
CO-insertion occurs, as in the iron system, to give 15.
3.1.1.1 Double cycloaddition. In the presence of an excess of dmad, the

increased 1,3-dipolar reactivity of 13, or rather 14, leads to a second cycloaddi-
Huns- Werner Frzihuuf 127
R R
16 17
Scheme 2
tion to give the 1,4,3a,6a-tetrahydropyrrolo[3,2-b]pyrrole(thpp) complex 17. In
17, the coordination of thpp to the metal is highly d y n a m i ~ ~ * and
' ~ ? at
'~ T
> 30 "C it readily dissociates from the metal. The { Ru(CO),} fragments immedi-
ately form [Ru,(CO),,] and, unfortunately, cannot be intercepted by an excess
of diazadiene to regenerate 13 in order to make the formation of thpp catalytic.
In the iron system, the double cycloaddition to thpp could only be observed
when the dipole was further activated by exchanging the CO ligands for isocyan-
ides (see Section 3.2).
3.2.1.2 Olefins as dipolarophiles. The reactivity of 13 is also sufficient to

cycloadd olefins such as dimethyl fumarate and maleate.' The conservation of
the cis- and trans-configuration of the olefins in products formed on N-protona-
tion and insertion of C O into the Ru-N bond of the cycloaddition product
indicates that the reaction is stereospecific and most likely to be a concerted
process. The C O ligands in 13 are thermally labile, for which reason 13 is only
stable in solution under an atmosphere of CO.,' It is therefore not surprising
that the substitution products in which a CO is replaced by dimethyl fumarate or
maleate are formed as side-products.
3.1.2 [Mn(CO),(dab)]-:a test-case for the isolobal relation

Stufkens et al.21-22had shown by means of ESR spectroscopy that in
[Mn(CO),(dab)]' radicals, obtained photochemically by homolytic cleavage of
the metal-metal bond in dinuclear complexes [Mn,(CO),(dab)], the unpaired
electron does not reside at the metal atom, but is mainly localised in the
n*-orbital of the dab. The radical [Mn(CO),(dab)]' (18) is therefore best de-
scribed as a 16-electron d6 manganese@ species, i.e. [Mn '(dab'-)(CO),]. Com-
paring this with the homologous [Fe(CO),(dab)] (1) shows that now we have a
d6 ML, fragment instead of the d8 ML, fragment in 1 which was an isolobal
analogue of carbene. If the isolobal concept holds, the C=N-MnlL, present
in 18 should not behave as a 1,3-dipole. Indeed, when [Mn,(CO),(dab)]
was irradiated in hexane (to generate 18) in the presence of dmad or
methyl propynoate (mp) as dipolarophiles, no cycloaddition took place but
extensive decomposition.8 The only product that could be identified was
[Mn,(CO),(dab)], which is also formed in the absence of any other reactants or
ligands.,' When the same experiments were performed instead in thf, a co-
ordinating solvent, the bicyclo[2.2.1] cycloaddition products 20, and in case of
R = Pri, a mixture of 19/20a (Scheme 3) was obtained in 60-70% yields. A donor
molecule (thf)can coordinate weakly to the 16-electron species 18, and hence act
temporarily as an additional ligand in 180thf.In that case, a d6 ML, fragment is
obtained, which is isolobally related to a d8 ML, fragment, and consequently the
corresponding L,Mn-N=C fragment is isolobal with the azomethine ylide and
hence a 1,3-dipole. The bicyclo[2.2.1] cycloadducts 19/20 cannot undergo mi-
gratory CO-insertion because the nitrogen has no lone pair with which to attack
nucleophilically a carbonyl carbon. Unfortunately, the only crystalline material
suitable for X-ray analysis turned out to be a mixed crystal with 19 and 20a both
present in the unit cell. The two could not be distinguished individually. The
geometry of the nitrogen bridge was therefore disordered and averaged between
sp2 and sp3.
R = Pi, pTol
18
1/2 Mn(, CO),
Complexes [M(CO),(dab)] (M = Cr, Mo or W), which like 18*thfcontain a d6

ML, fragment, do not react with dipolarophiles. This can be rationalised in the
following way: (i) the extra electron in the LUMO, or rather SOMO, of the dab
ligand in 18othf increases the reactivity of the 1,3-dipole sufficiently, and (ii) the
carbonyl ligands in [M(CO),(dab)] bind too strongly to the metal and cannot be
lost as easily as thf from manganese in order to avoid an increase of the
coordination number to seven during the cycloaddition.
Hans- Werner Fruhauf 129
3.2 Changing the Additional Ligands: Isocyanides for Carbon
Monoxide
The decision to replace carbonyls in 1 for isonitriles is based on two reasons: (i)
while C O is a pronounced n-acceptor, isocyanides R-NC are better a-dona-
and
ting/less n - a ~ c e p t i n g ,2~8 ~ - their electronic and steric properties can be varied
by the choice of R. Increased donation to the metal should raise the HOMO-
energy of the 1,3-dipole and increase its reactivity; (ii) like CO, isocyanides are
known to undergo insertion reaction^,^^-^^*^^ and their introduction opened the
possibility of obtaining new, isocyanide inserted compounds, and further, in
mixed complexes, of studying a possible intramolecular competition between
CO and isocyanide insertion in the initial cycloadducts.
3.2.1 Reactivity of [Fe(CO),(RNC)(dab)](21)

With Pd/C catalysis, one C O in complex 1 can be selectively substituted for an
isonitrile R-NC to give 21.30When 21 is reacted with dmad at - 60°C, three
types of product, 22-24, are obtained in 60-95% total yields after column
separation.'
22 23 24
The product distribution depends on the isocyanide used. Only with aromatic
(R = o-tolyl or 2,6-xylyl), and not with aliphatic isocyanides (R = But, Bus or
benzyl) are isonitrile insertion products 24 formed. Aryl isocyanide insertion is
obviously very fast since the competing formation of 23 from CO insertion is not
observed; complexes 22 are only minor products. With aliphatic isocyanides, the
thpp complexes 22, from double cycloaddition of dmad (cf: Section 3.1.1), are the
major products (70 to >95% of the product mixture) and indicate a strongly
increased 1,3-dipolar reactivity, i.e. the intermolecular second cycloaddition is
preferred to the intramolecular CO insertion. Compared with the ruthenium
compound 17, the thpp in 22 is strongly bound to the metal and can only be
decomplexed oxidatively with cerium@), or under 80 bar of CO.
3.2.2 Reactions of [Fe(RNC),(dab)](25) with drnad,,' 0 1 e f i n s ~ ~ ' ~ ~

and h e t e r ~ a l l e n e s ~ ~ - ~ ~
Complexes 25 had been described in the literature3 with spectroscopic evidence
only, and were supposedly very labile. However, it was possible to prepare 25a-c
(a: R = 2,6-xylyl, b: R = But, c: R = Cy) in synthetically useful amount^.^' In
particular 25a proved to be the optimally activated 1,3-dipole, which not only
reacts with the greatest variety of dipolarophiles, but also exhibits some very
surprising and interesting consecutive reactions.
The reaction of 25a-c with dmad3' proceeds in a manner completely analog-
ous to the reaction of 21, i.e. giving thpp complexes by double cycloaddition in
case of the aliphatic isonitriles (b, c), and isonitrile insertion with the aromatic
2,6-xylyl isonitrile a. Just as complexes 4 with L' = trimethyl phosphite (Scheme
1) and 23, the corresponding trisisonitrile bicyclic complex does not reductively
eliminate a pyrrolinone imine, probably again due to the o-donor capacity of the
ligands that increase the electron density at the iron@) centre. However in the
presence of water, this trisisonitrile bicyclic analogue of 23 finds an alternative
pathway for which a plausible sequence of steps involves hydrolysis of one ester
function to the acid. The hydroxyl group then oxidatively adds to the electron
rich iron(@ centre forming an intermediate with a formal iron@) centre. Simul-
taneously, the imine moiety is displaced from the metal. From the high oxidation
state, a reductive elimination regenerates the iron(I1)oxidation state by forming a
1,5-dihydro-2-iminopyrrole followed by recoordination of the double bond. The
stable tricyclic end product 26, the structure of which has been established by
X-ray crystallography, is formed by formal insertion of the ring double bond into
the metal-hydrogen bond and recoordination of the pendant imine group. That
the new proton originates from water has been confirmed by performing the
reaction in D 2 0 instead of H 2 0 . 2H NMR unequivocally proved the incorpor-
ation of deuterium at the former alkyne carbon atom.
26 R = 2,6-~ylyl
E = CQMe
The reaction of 25a, containing the aromatic 2,6-xylyl isocyanide ligands, with
dimethyl maleate in pentane (Scheme 4)32*33 proceeds with stereospecific cyc-
loaddition of the cis-olefin and subsequent isocyanide insertion to form the
bicylo[2.2.2] complex 27, which precipitates as a yellow powder in 95% isolated
yield. The structure of 27 has been confirmed by X-ray crystallography. Very
surprisingly, at slightly elevated temperature the whole reaction sequence is
cleanly reversed, a process that can be monitored by temperature-dependent
NMR spectroscopy. When a solution of pure 27 in C,D, is kept at 60°C, the
signals of 27 disappear and the signals of 25a, dimethyl maleate, and 2,6-xylyl
isocyanide grow in and finally replace them until the solution is cooled down
again. It is an indication of the microscopic reversibility of the reaction that the
reverse reaction is also stereospecific, with only dimethyl maleate and no fumar-
ate being formed. This temperature-controlled molecular self-assembly and dis-
assembly is unique, considering all the steps involved, and the ease with which
these steps occur. During the process, C-C, C-N, Fe-C o-bonds, and an
Fe-CNR donor bond are formed and broken. Another intriguing aspect of the
Huns- Werner Frzihauf 131
reverse disassembly reaction is that the reaction represents the first unequivo-
cally established case of an isocyanide deinsertion reaction.
m . E E
Pi-N, ,N-Pt' +
Fe; -
RNC' I CNR E E
CNR -u
25 R = 2,5-~ylyl
I
:[ ]
de-insertion insertion
Me02C
Pt'\N*
RN A n - p +-
Me02C Fe,
I\
CNR
+CNR
-CNR - $ Me0,C
Me02C Fey Pt'

RNC CNR
27
Scheme 4
The influence of the type of isonitrile on the reactivity of 25 towards olefins is
dramatic. With the aliphatic isonitriles b and c, a totally different reaction with
dimethyl maleate is observed. Scheme 5 shows the isolated products 28 and 29b,c
and the proposed reaction pathway.33
The observation that the reaction is strongly slowed down by an excess of
isocyanide suggests that the reaction starts with the substitution of an isocyanide
by an alkene. The rest of the reaction sequence is more or less speculative, but in
any case there is a high degree of stereoselectivity. The final tricyclic complex 29
contains five chiral centres, yet only one diastereomer is observed, the structure
of which has been confirmed by X-ray crystallography. The relative amount of
the organic olefin dimer 28 depends on the reaction temperature. At - 30°C, the
ratio 28/29 is 1/4, and at 0°C it is 6/4, i.e. at higher temperature, reductive
elimination of 28 becomes faster than olefin insertion into the Fe-H bond. The
fact that complexes 25b,c, with three strongly electron-donating aliphatic
isocyanides, give only cycloaddition with the very reactive dipolarophile dmad,
and not with dimethyl maleate, must be a consequence of the high electron
density at the metal and the weaker metal-isocyanide coordinative bond in 25b,c
(weaker n back-bonding as compared to the aromatic isocyanides). The substitu-
tion of one of the terminal aliphatic isocyanides for a better n-accepting olefin via
a dissociative pathway apparently becomes much faster than a cycloaddition
reaction.
Complexes 25 also cycloadd the C=S double bonds in heteroallenes CS,,
COS,34and aryl i s o t h i o c y a n a t e ~ .The
~ ~ *reaction
~~ of 25a with CS, (Scheme 6)
gave an 80% yield of the expected bicyclo[2.2.2] complex 30, the structure of
which has been established by X-ray ~ r y s t a l l o g r a p h yWhen
. ~ ~ the reaction was
performed in the presence of water or HBF,, the amido nitrogen bridge in the
initial bicycloC2.2.13 adduct was protonated, which inhibited the insertion of
isocyanide, and 31 was isolated in 80% yield.
E
C: R = BU'
E-CH2
I
H2C-E
28
\
E
+cNRl
E
-
p-elimn.
29b,c
Scheme 5
+ HBF4or H20 30
31
Scheme 6
With aryl isothiocyanates, 4-XC6H,NCS, (X = H, Me, OMe or NO,), com-

plex 25a reacted in the familiar way by cycloaddition of the C-S bond and
isocyanide insertion to form the bicyclo[2.2.2] adducts 32.35The structure of 32,
X = H, has been determined by X-ray crystallography. In the presence of a
second equivalent of isocyanide, the bicyclo[3.2.2] products 33, resulting from a
second isocyanide insertion, could be isolated. This reaction is cleanly thermally
reversible, and for the first time, the thermodynamic parameters of an isocyanide
insertion/deinsertion reaction could be determined by temperature-dependent
NMR spectroscopy. At high temperatures the reaction sequence is reversed all
the way to 25. However, at these temperatures the thermal stability of 25 is not
sufficient to establish a stable equilibrium. The crystal structure of 33 could not
be determined, because the equilibrium between 32 and 33 was obviously not
frozen even at very low temperatures. When a saturated solution of pure 33,
X = OMe, in ether/dichloromethane (5/1) was stored at - 80°C, after several
months crystals of 32, X = OMe, were obtained. So, the dynamic equilibrium
had completely shifted towards the side of 32, X = OMe, as a result of its lower
solubility.
The labilisation of the coordinative bonds due to the strong a-donation of the
aliphatic isocyanide ligands in 25a,b, which has already been mentioned above,
obviously also plays a role again in the reaction of 25a,b with aryl isothiocyan-
ates (Scheme 7).36The X-ray structure of 34c, L = Bu‘NC indicates a C3.2.01
bicyclic complex with a coordinated amido nitrogen and an uncoordinated
imine nitrogen (A). However, temperature-dependent NMR spectroscopy indi-
cates a dynamic competition of these two nitrogen atoms for the coordination
site. As in all previous cases, the terminally coordinated aliphatic isocyanides do
not insert in the initial bicycloC2.2.1) cycloadduct, but rather an external iso-
thiocyanate is inserted, most likely after precoordination by displacement of the
imino group.
1
R‘
.Pi
Pi,
B 34a-d A
Scheme 7
3.3 Changing the Heteroatoms
Like the variations of the metal (Section 3.1) and of the additional ligands
(Section 3.2), the variation of the heteroatom in the 1,3-dipole cannot be done
arbitrarily. A suitable synthetic route to and the stability of the resulting starting
compound are the limiting factors. So, like the complexes [Ru(CO),(dab)] (13)
and [Fe(RNC),(dab)] (25), several of the 1,3-dipoles in this section had to be
prepared and reacted in situ. Their identity, however, was beyond doubt - either
through spectroscopic characterisation and comparison with stable representa-
tives of the same type, or from a complete characterisation of their reaction
products.
3.3.1 Oxygen instead of nitrogen: a-imino ketones and a-imino

esters
The previously mentioned CAS-SCF calculations16 had already shown that the
isolobal analogy between an azomethine ylide and the L,Fe-N=C fragment
(Equation 3) may be extended to an L,Fe-O=C fragment 35 and a carbonyl ylide
36 (Equation 4), and that the oxygen homologues would have slightly higher
HOMO and LUMO levels, and a smaller HOMO-LUMO gap. From these
properties, an increased 1,3-dipolar reactivity could be expected.
35 36
In order to compare the 1,3-dipolar reactivity of the L,Fe-N=C and the
L,Fe-0-L fragments, i.e. to investigate the chemoselectivity of a dipolarophile
towards them, a series of suitable complexes with an Fe-N==C and an Fe-0%
fragment within the same molecule, namely [Fe(CO),(a-imino ketone)] (37)38939
and [Fe(CO),(a-imino ester)] (41),40p42 were prepared and allowed to react with
activated alkynes (Schemes 8-10).
When complexes 37a,b reacted in pentane solution at -78°C under an
atmosphere of C O with dmad (Scheme 8), the butenolide complexes 38a,b were
formed in clean reactions and could be isolated in 85 and 65% yields,,' i.e. the
reaction proceeded with complete chemoselectivity for the Fe-O=C fragment.
When the same reaction was performed at - 50°C and - 30"C, respectively,
, ~ ~butenolide complexes 39a,b
with methyl propynoate (mp) as d i p ~ l a r o p h i l ethe
with the hydrogen next to the inserted carbonyl group were formed with com-
plete chemo- and regio-selectivity. Complexes that had incorporated two moles
of the alkyne (40a,b)were observed as minor side-products. In the absence of CO,
and with two equivalents of mp, the tricyclic complexes 40a,b were formed
exclusively in moderate yields. Complexes 40a,b can also be prepared in almost
quantitative yields by irradiation of the corresponding complexes 39a,b in the
presence of an excess of mp at room temperature, an observation that strongly
supports the pathway indicated in Scheme 8. X-Ray structural analysis of 40b
has confirmed that, of the possible regioisomers, only the one shown in Scheme 8
is formed.
Hans- Werner Frti'hauf 135
37 R' = a: Ph
R2 = C02Me
b: Me
R3 = C02Me, H
The effect of reaction temperature on the cycloadditions of 37a,b with dmad

and mp (vide supra) already indicated that not only has the dipolarophile a
marked influence on the reactivity, but so has the substituent R' (phenyl us.
methyl) in the 1,3-dipole. However, in both cases the cycloaddition was com-
pletely selective for the Fe-0% moiety. In order to probe how far the reactivity
of the Fe-0% dipole could be further attenuated by the choice of R', an
extended series of imino ester complexes 414' was reacted with the two
dipolarophiles dmad and Instead of the aryl or alkyl groups R' in 37,
complexes 41 bear an electronegative oxygen atom. The less reactive
dipolarophile mp again reacted exclusively with the more reactive Fe-O=C
dipole. However, with the more reactive dmad, the Fe-N=C dipole becomes
competitive, and both the butenolide complexes 45 (Scheme 9) and pyrrolinone
complexes 49 (Scheme 10) are formed. The weakly coordinating ester carbonyl
group in 47 is displaced by an extra CO in 49. The initial bicyclo[2.2.1] cycload-
ducts 42 and 46 also undergo a side reaction with an {Fe(CO),) species present
from the in situ preparation of 41, to give the binuclear products 44 and 48. All
product structures have been confirmed by X-ray crystallography.
42
R'
a: OMe
b: OMe
c: OMe
R2
Pr"
neepentyl
Prl
[Fe(CO),]
i -2CO
R'
d: OEt. Pr'
e: Opt' Pt'
f: OMe BU'
g: OEt, But
h: Opt' But
i: OMe amd
j: OEt amyl' 44a-j
Scheme 9
R' CO insertn.,
R3 +co,
h'N-R'
0'
cycloaddn. red. elimn.
-+-
' /
oc'Fe\
kOC0
R3 = CQMe
41a-j 46 47
48a,c 49%
Scheme 10
The steric bulk of both substituents R1and R2 clearly influences the product
distribution. Increasing the bulk of R1,e.g. in the series c -+e, or f -+ h, results in
an increasing preference for the Fe-N=C fragment, while increasing the bulk of
R2 (a, c, f) favours the Fe-O=C fragment.
In the stable and isolable imino ketone complexes 37a and 37c (Scheme 1 l), at
room temperature, one or two carbonyls could be exchanged for phosphorus
ligands k-n [k = P(OMe),, 1 = PPh,, m = PEt,, n = PPr",] or dppe
(Ph2PCH2CH2PPh2).6 The resulting complexes 50/51 were prepared in the
expectation that the increased donor capacity of the phosphorus ligands, as
opposed to CO, would increase x-back-donation into the imino ketone LUMO
and thus increase the reactivity of the 1,3-dipole. This was indeed the case, and
complexes 50/51 were found reactive towards acetylenes (dmad, mp, phenyl
acetylene), olefins (dimethyl maleate) and aryl isothiocyanate. Not all cycloaddi-
tion products were isolable though, and could only be characterised spectro-
scopically. The most remarkable result, however, was that for the first time in the
iron system the initial bicyclo[2.2.1] adduct was directly observable. In the
reaction of the dppe complex 51a with dmad (Equation 5), complex 52 was the
stable end product of which single crystals could be grown to determine its X-ray
structure. CO insertion does not occur in this case, which is obviously due to the
trans disposition of the oxygen bridge and the single carbonyl ligand, as
evidenced by the crystal structure of 52.
3.3.2 Sulfur instead of nitrogen: dithiooxamide

The coordination chemistry of dta towards carbonyl iron has been studied to
find out if it was possible to prepare at least in situ mononuclear chelate
[Fe(CO),(a-S,o-S'-dta)] (dta = dithiooxamide) complexes (53) in order to inves-
tigate their 1,3-dipolar b e h a ~ i o u r . ~
This
, proved successfulwith dta a 4 (Scheme
12). However, only the reaction of 53a with dmad gave an isolable [3 + 21
R3 R3
HR2
0,,N-R' PR3or dppe
* HR2
0,,N-R'
Fe rt
Fe
oc/(i;co OC/(i!.PR3
37 50a,c,k-n
a: R'= But, F?= H, F& Ph k: R=OMe
c: R'= Me, F?= Ph, @= Ph I: R = P h
m:R = Et
n: R = P f
Scheme 11
Ph Ph
L
cycloadduct 54a. In 54a, an (Fe(CO),) fragment, present in solution from the in

situ generation of 53a, has coordinated to the sulfide bridge. This inhibits
carbonyl insertion and stabilises the cycloadduct. Extensive decomposition is
observed when the reaction with dmad is done with pure 53a, i.e. in the absence
of {Fe(CO),), and the only isolable product is a small amount of 54a. When 53d
reacts with an excess of dmad, no [3 + 21 cycloaddition takes place, and only
small amounts of 55d formed by two [2 + 21 cyloaddition reactions can be
isolated.
I
E-E Me"
Fe2(co)9
____Lc
s' X ' sN
-N -Fe(CO)5
54
40
a: R = Me, R'-R' = CH2-CH2 excess
b: R = H, R' = PI'
E-E
c: R = H, R' = Benzyl
d: R = R' = Et E = C02Me
55
Scheme 12
4 Acknowledgements
I wish to thank my very creative and prolific former graduate students Frank
Seils, Maarten van Wijnkoop, Wouter (P. P. M.) de Lange, Nantko Feiken and
Ron Siebenlist, and the accompanying undergraduates named in the references.
It is their work which is collected here, and they have made it a very pleasant and
fruitful period. However, the work would not have been possible without the
countless structure determinations by the crystallographers Carl Kriiger (Miil-
heim), Ton Spek (Utrecht), and Kees Goubitz (Amsterdam), and their co-
workers. Finally, the successive financial support by the German (DFG) and
Dutch (SON-NOW) Science Foundations is gratefully acknowledged.
5 References
1 H.-W. Fruhauf, I. Pein and F. Seils, Organometallics, 1987,6, 1613.
2 H.-W. Friihauf, F. Seils, M. J. Romio and R. J. Goddard, Angew. Chem. Int. Ed. Engl.,
1983,22,992.
3 H.-W. Fruhauf, F. Seils, R. J. Goddard and M. J. Romiio, Organometallics, 1985, 4,
948.
4 H.-W. Friihauf, F. Seils and C. H. Stam, Organometallics, 1989,8,2338.
5 H.-W. Fruhauf and F. Seils, J . Organomet. Chem., 1986,302,59.
6 R. Siebenlist, PhD Thesis, Chapter 4, pp. 125-157, Universiteit van Amsterdam,
Amsterdam, The Netherlands, 1996.
7 M. van Wijnkoop, P. P. M. de Lange, H.-W. Fruhauf, K. Vrieze, Y. Wang, K. Goubitz
and C. H. Stam, Organometallics, 1992,11, 3607.
8 M. van Wijnkoop, PhD Thesis, Chapter 3, pp. 71-88, Universiteit van Amsterdam,
Amsterdam, The Netherlands, 1992.
9 N. Feiken, P. Schreuder, R. Siebenlist, H.-W. Fruhauf, K. Vrieze, H. Kooijman, N.
Veldman, A. L. Spek, J. Fraanje and K. Goubitz, Organometallics, 1996, 15,2148.
10 R. Huisgen, 1,3-Dipolar Cycloaddition Chemistry, ed. A. Padwa, Wiley & Sons Inter-
science, New York, NY, 1984.
11 R. Huisgen, Angew. Chem. Int. Ed., 1968,7,321.
12 R. Hoffmann, Angew. Chem. Int. Ed. Engl., 1982,21,711.
13 F. G. A. Stone, Angew. Chem. Int. Ed. Engl., 1984,23,89.
14 M. W. Kokkes, D. J. Stufkens and A. Oskam, J . Chem. Soc., Dalton Trans., 1983,439.
15 H. K. van Dijk, D. J. Stufkens and A. Oskam, J . Am. Chem. Soc., 1989,111,541.
16 A. Dedieu and M. J. Liddel, University Louis Pasteur, Strasbourg, France, 1992,
unpublished.
17 M. van Wijnkoop, P. P. M. de Lange, H.-W. Fruhauf, K. Vrieze, W. J. J. Smeets and A.
L. Spek, Organometallics, 1995, 14,4781.
18 P. P. M. de Lange, H.-W. Friihauf, M. J. A. Kraakman, M. van Wijnkoop, M.
Kranenburg, A. H. J. P. Groot, K. Vrieze, J. Fraanje, Y. Wang and M. Numan,
19 P. P. M. de Lange, H.-W. Friihauf, M. van Wijnkoop, K. Vrieze, Y. Wang, D.
Heijdenrijk and C. H. Stam, Organometallics, 1990,9, 1691.
20 W. P. Mul, C. J. Elsevier, H.-W. Friihauf, K. Vrieze, I. Pein, M. C. Zoutberg and C. H.
Stam, Inorg. Chem., 1990,29,2336.
21 T. van der Graaf, D. J. Stufkens, A. Oskam, and K. Goubitz, Inorg. Chem., 1991,30,
599.
22 R. R. Andrea, W. G. J. de Lange, T. van der Graaf, M. Rijkhoff, D. J. Stufkens and A.
Oskam, Organometallics, 1988,7, 1100.
23 E. Singleton and H. E. Oosthuizen, Adv. Organornet. Chem., 1983,22,209.
24 Y. Yamamoto, Coord. Chem. Rev., 1980,32,193.
25 L. Malatesta and F. Bonati, Isocyanide Complexes ofMetals, John Wiley & Sons, New
York, 1969.
26 P. Fantucci, L. Naldini, F. Cariati, V. Valenti and C. Bussetto, J . Organornet. Chern.,
1974,64, 109.
27 M. Minelli and W. Maley, Inorg. Chern., 1989,28,2954.
28 J. Y. Saillard, A. Le Beuze, G. Simmoneaux, P. Le Maux and G. Jaouen,
THEOCHEM - J . Mol. Strut., 1981,3, 149.
29 Y. Yamamoto and H. Yamazaki, Coord. Chern. Rev.,1972,8,225.
30 P. P. M. de Lange, M. J. A. Kraakman, M. van Wijnkoop, H.-W. Friihauf, K. Vrieze,
W. J. J. Smeets and A. L. Spek, Inorg. Chirn. Acta, 1992,196, 151.
31 P. P. M. de Lange, M. van Wijnkoop, H.-W. Fruhauf, K. Vrieze and K. Goubitz,
Organornetallics, 1993,12,428.
32 R. P. de Boer, P. P. M. de Lange, H.-W. Fruhauf and K. Vrieze, J . Chern. Soc., Chern.
Cornrnun., 1992,580.
33 P. P. M. de Lange, R. P. de Boer, M. van Wijnkoop, J. M. Ernsting, H.-W. Friihauf, K.
Vrieze, W. J. J. Smeets, A. L. Spek and K. Goubitz, Organornetallics, 1993,12,440.
34 P. P. M. de Lange, E. Alberts, M. van Wijnkoop, H.-W. Friihauf, K. Vrieze, H.
Kooijman and A. L. Spek, J . Organornet. Chern., 1994,465,241.
35 N. Feiken, H.-W. Friihauf, K. Vrieze, J. Fraanje and K. Goubitz, Organornetallics,
1994,13,2825.
36 N. Feiken, H.-W. Friihauf, K. Vrieze, N. Veldman and A. L. Spek, J . Organornet.
Chern., 1996,511,281.
37 H. tom Dieck, R. Diercks, L. Stamp, H. Bruder and T. Schuld, Chern. Ber., 1987,120,
1943.
38 M. van Wijnkoop, R. Siebenlist, J. M. Ernsting, P. P. M. de Lange, H.-W. Fruhauf, E.
Horn and A. L. Spek, J . Organornet. Chern., 1994,482,99.
39 M. van Wijnkoop, R. Siebenlist, P. P. M. de Lange, H.-W. Fruhauf, K. Vrieze, W. J. J.
Smeets and A. L. Spek, Organornetallics, 1993,12,4172.
40 R. Siebenlist, H.-W. Friihauf, K. Vrieze, W. J. J. Smeets and A. L. Spek, Eur. J . Inorg.
Chern., 2000,907.
41 R. Siebenlist, H.-W. Fruhauf, K. Vrieze, H. Kooijman, W. J. J. Smeets and A. L. Spek,
Organornetallics, 2000,19,3016.
42 R. Siebenlist, M. de Beurs, N. Feiken, H.-W. Friihauf, K. Vrieze, H. Kooijman, N.
Veldman, M. T. Lakin and A. L. Spek, Organornetallics, 2000, 19,3032.
43 R. Siebenlist, H.-W. Fruhauf, H. Kooijman, N. Veldman, A. L. Spek, K. Goubitz and J.
Fraanje, Inorg. Chern. Acta, in press.
A Journey in Metal-Ligand
Multiple Bond Chemistry
VERNON C . GIBSON
Department of Chemistry, Imperial College of Science, Technology and
Medicine, Exhibition Road, South Kensington, London SW7 2QA, UK
1 Introduction
In the late 1970s, whilst an undergraduate at the University of Sheffield, and
under the tutelage of Maitlis, McCleverty and Fenton, I was first introduced to
the many seminal contributions to inorganic and organometallic chemistry of
Joseph Chatt: the simple but ‘universal’ bonding model for alkenes binding to a
transition metal centre, the chemistry of metal-ligand multiple bonds, car-
bon-hydrogen bond activation, and not least of all his work on dinitrogen
complexes at the Nitrogen Fixation Unit, to highlight but a few. At that time, as
an undergraduate, I could only dream that I would myself one day be research-
ing some of the very same areas into which Joe Chatt had provided such great
insight and inspiration. The chemistry outlined in this chapter is a personal
journey which started in the mid-1980s with my appointment to a lectureship in
inorganic chemistry at the University of Durham and which led me along a path
that would enter some of the subject areas which had been so dramatically
influenced by Chatt.
Starting with a desire to answer a simple question in metal 0x0 chemistry, the
journey quickly broadened to take in metal imido systems and to explore
applications of these complexes in olefin metathesis, especially ring-opening
metathesis polymerisation (ROMP), and later in a-olefin polymerisation. The
structure of this chapter broadly reflects the original itinerary, with a few
diversions along the way. It is inevitable that an account of this type glosses over
the many important contributions by other researchers in the field. I apologise
unreservedly to them and hope readers will refer to the citations of their work
contained within the accompanying references.
Vernon C. Gibson 141
2 The Elusive [M(O)Cl,Cp] Complexes of Niobium

and Tantalum
Embarking on an academic career in the mid-l980s, one of the first questions I
wanted answered was why terminal 0x0-ligands are so abundant for certain
transition metals but not for others? The explanation for late transition metals
was relatively straightforward since the hard n-basic 0x0-ligand is less compat-
ible with d electron-rich metals, leading to the filling of anti-bonding M-L
molecular orbitals. However, the answer was less clear for early transition metal
systems and especially so it seemed for the Group 5 metals, vanadium, niobium
and tantalum. [V(O)Cl,Cp] (l),a volatile and thermally robust three-legged
piano stool molecule (Figure l), had been described by E. 0.Fischer in 1958,' yet
30 years later the niobium and tantalum analogues (2 and 3) had still not been
described. These, therefore, presented tantalising targets.
My first PhD student, Terry Kee (now on the faculty at the University of
Leeds) set about trying to make the niobium and tantalum compounds by
reaction of M(0)C13 with a variety of Cp sources. We tried them all! All failed,
and all we were able to salvage from this initial endeavour were improved
syntheses of the M(0)C13 starting materials2 Other obvious routes, e.g. via
reaction of [MCl,Cp] with (Me,Si),O also proved unsuccessful, in the case of
[TaCl,Cp*] (Cp* = pentamethylcyclopentadienyl) affording [(TaC13Cp*},(p-
O)] and [TaC13(OSiMe3)Cp*].3Terry, however, would not be thwarted, and he
switched from looking at high-valent routes to exploring the potential for
low-valent precursors. He was subsequently able to show to our surprise that the
half-sandwich tertiary phosphine tantalum(II1) complex, [TaC1,(PMe3),Cp*]
(4), reacted with carbon dioxide to yield [(Ta(O)Cl,Cp*},] (5), along with
[TaCl,(CO),(PMe,)Cp*] (6)as the main by-product3 (Figure 2). The character-
ising data for 5 revealed a dinuclear structure with bridging rather than terminal
0x0 ligands. It was about this time that Alan Shaw joined the group and he and
Terry were able to provide the answer as to the elusivity of these corn pound^^^^
by showing that 5 is in fact unstable converting, at room temperature in chloro-
form solution and at elevated temperature in toluene, to a mixture of the
trinuclear oxide cluster, [Ta304C14Cp*3] (7),5 and the dinuclear species,
[(TaCl,Cp*),(p-O)] (8). It was also found to react with (Me,Si),O to give a
mixture of the same dinuclear compound and [TaCl,(OSiMe,)Cp*) (9),which
explained the failure of our earlier attempts to prepare [Ta(O)Cl,Cp*] via the
o+v; - CI
CI
1
E.0.Fischer
1958 unknown
Figure 1 Half-sandwich oxychlorides of the Group 5 metals

142 A Journey in Metal-Ligand Multiple Bond Chemistry
4
5
cHYI Me3SiCl
I
+
0+~%-0Ar
OAr
[Cp*TaC13(OSiMe3)1
10
9
Ar = 2,6-Me2C~H3
Figure 2 The low valent route to [Ta(O)Cl,Cp*] and derivative chemistry
reaction of [TaCl,Cp*] with (Me,Si),O. Ultimately, Alan was able to show that
it is possible to stabilise terminal 0x0 half-sandwich tantalum species by ex-
changing the chlorides for phenoxide ligands (compound lo)., However, the
question was essentially answered: for tantalum, at least, the elusive nature of
terminal 0x0 complexes was attributable to the greater stability afforded by
bridging 0x0 ligand environments. Unfortunately, this low-valent pathway pro-
ved to be unsuited to the niobium analogue.
3 Half-sandwich Imido Compounds of Niobium and

Tantalum
The lack of stability of terminal 0x0 compounds of the heavier Group 5 metals
led us naturally to the door of the isoelectronic imido (NR) ligand, where the
availability of a substituent attached to the multiply-bonded group would allow
both steric and electronic modulation of the products’ stability and reactivity.
There had been a handful of half-sandwich imido complexes of the Group 5
metals synthesised by other workers, especially for vanadium6 and t a n t a l ~ m , ~
but at that time none were known for niobium. A half-sandwich imido com-
pound of niobium we considered, therefore, a prime target.
David Williams was able to show that heptamethyldisilazane, (Me,%),-
NMe, reacts with [NbCl,Cp] to give moisture-sensitive, yellow-orange
[Nb(NMe)Cl,Cp] (11),* ‘the crystal structure of which revealed a monomeric
three-legged piano stool structure (Figure 3). By subsequent judicious choice of
silylated amine reagents, he was able to generalise this synthetic entry to access a
wide family of half-sandwich niobium and tantalum imido compounds (12,
Figure 3) and set about developing their derivative chemistry.’
Vernon C . Gibson 143
Figure 3 Half-sandwich imido complexes of the heavier Group 5 metals, and the molecular
structure of [Nb(NMe)CI,Cp]
Figure 4 The isolobal relationship between imido and cyclopentadienyl species
At about this time, we were enjoying a collaboration with Dick Schrock at

MIT, applying his well-defined molybdenum metathesis catalysts to new ma-
terials synthesis (in conjunction with Jim Feast at Durham), and we had often
found ourselves, over various liquid refreshments, discussing the relationship
between the simple imido moiety and cyclopentadienyl ligands. At first sight they
might appear as different as chalk and cheese, but on closer examination some
striking similarities become apparent. Although possessing different formal
charges, these six-electron ligand fragments should present the same symmetry
combinations of frontier orbitals to a metal centre (Figure 4), thereby allowing an
intriguing isolobal parallel to be drawn between the Group 4 metallocenes,
half-sandwich imido compounds of the Group 5 metals and bis(imido) com-
pounds of the Group 6 metals (Figure 4). The relationship appeared to work in
bis(imido)tungsten systems under investigation in Dick’s labs at MITI’ and we
directed our attention to testing this hypothesis on the half-sandwich Group 5
metal imido system.
’
Elegant theoretical studies by Lauher and Hoffmann’ had established the
molecular orbital description of the Group 4 metallocenes and so David Will-
iams turned his hand to M O calculations on the [Nb(NMe)Cp] system. The
outcome was a frontier orbital picture showing only minor differences to the
(MCp2) fragment.’ Moreover, the interaction of trimethylphosphine with
[Nb(NMe)Cl,Cp] provided experimental evidence for the similar orientations of
the complex LUMOs in [MX,Cp,] and [M(NR)X,CP].~
4 Exploiting the Isolobal Relationship Between

Cyclopentadienyl and Imido Ligands
In order to place the isolobal analogy between Group 4 metallocenes, half-
sandwich imido compounds of the Group 5 metals and bis(imido)metal com-
plexes of the Group 6 metals on a firm experimental footing, we set about
probing similarities and differences in their structures and reactivity. Andrew
Poole was able to show that the bis(pheny1) complex, [Nb(N-2,6-
Pr’,C,H,)(Ph),Cp*] reacts in much the same way as [ZrPh,Cp,] upon treat-
ment with trimethylphosphine.’ ,*’
In both cases, elimination of benzene occurs
resulting in formation of a stable benzyne complex (see 13, Figure 5 for the
niobium derivative). The close similarity of the bond lengths within the benzyne
ring and the orientation of the benzyne ligand provided further support for the
metallocene-like nature of the frontier orbitals of the (Nb(NR)Cp*) fragment.
In an analogous reaction, the bis(benzy1) species [Nb(N-2,6-
Pr’,C,H,)(CH,Ph),Cp*] was found to eliminate toluene upon treatment with
trimethylphosphine to give the benzylidene complex, [Nb(N-2,6-
Pr’,C,H,)(=CHPh)(PMe,)Cp*] (14).12,’ Here, the substituents of the al-
kylidene ligand are orientated towards the Cp* and imido ligands allowing
overlap of the carbene carbon p orbital with the d, symmetry ‘metallocene-like’
frontier orbital of the (Nb(NR)Cp*) fragment. The olefin complexes, [Nb(N-2,6-
Pr’,C,H,)(CH, =CHR)(PMe,)Cp’] (Cp’ = Cp or Cp*; R = H or Me)l47I5 e.g.
15, revealed similar orientational preferences consistent with ‘metallocene-like’
frontier orbitals for the half-sandwich imido fragment. Furthermore, their
method of synthesis, via treatment of the dihalide precursors with two equival-
ents of alkylmagnesium chloride in the presence of trimethylphosphine, is entire-
ly analogous to that employed in the preparation of their metallocene relatives. It
proved possible to extend the series to acetylene complexes (16) by two routes:
either by direct displacement of the olefin14 or via exchange of an acetylene for
one of the PMe, ligands in [Nb(NR)(PMe3),Cp*].15
Having synthesised a family of ‘metallocene look-alikes’, we were interested to
RN
13 R = 2,6-Pt‘,CeH3 14 16
e Nb- - PMe3
BU”
RN//
15
> 17 18
Figure 5 Metal-imido relatives of the zirconocene family

explore their derivative chemistry, especially potentially useful C-C bond-form-
ing reactions. However, we had not reckoned with the over-zealous binding of
the trimethylphosphine ligand which we unsuccessfully attempted to displace
from the niobium centre with a variety of unsaturated organic substrates. This is
most graphically illustrated by the reaction of [Nb(NAr)(y2-C3H6)(PMe3)Cp]
(19) with butadiene (Figure 6). Although the propylene can be displaced to give
an y2-butadiene complex (20, as four isomers),16 prolonged warming did not
result in displacement of the trimethylphosphine to give the q4-butadiene prod-
uct. This is in contrast to the zirconocene system which not only readily forms
the y4-butadiene species, but also has been shown to furnish synthetically useful
carbon-carbon bond forming reactions.17
Two ways forward appeared open to us: either we could find an alternative less
strongly binding ligand which could be displaced from niobium, or we could
look to its neighbour tantalum, which, according to the diagonal relationship,
should bear a closer electronic resemblance to zirconium. Andrew Poole decided
to set about developing the chemistry of the tantalum system. Disappoint-
ments were in store, however. For example, the tantalum benzyli-
dene species was not formed cleanly on thermolysis of [Ta(N-2,6-
Pri2C6H3)(CH2Ph)2Cp*] in the presence of trimethylphosphine, nor could evi-
dence be obtained for the formation of a tantalum-benzyne species on thermoly-
sis of [Ta(N-2,6-P1-'~c,H,)(Ph)~cp*]. Only paramagnetic decomposition prod-
ucts were produced.I6 These disappointments turned out to be a good omen,
reflecting as they did the greater lability of the tertiary phosphine ligand.
The tantalum-olefin derivatives provided the advance we were seeking. The
ethylene complex, [Ta(N-2,6-Pri2C6H3)( C, H4)(PMe ,)Cp *] (21) was found to
react with an excess of ethylene to give the tantalacyclopentane (22) complex
shown in Figure 7 . I 6 9 I 8 It also proved possible to synthesise the same complex in
better yield by reaction of [Ta(N-2,6-Pri2C6H3)C12cp*] with two equivalents of
ethylmagnesium chloride in the presence of an excess of ethylene. In behaviour
reminiscent of zirconacyclopentanes, the half-sandwich tantalacyclopentane
species does not undergo facile p-elimination to afford but-1-ene due to the
conformational constraints within the MC, metallacycle that prevent the
P(C-H) bonds from accessing the metal-centred LUMO. It does, however,
undergo slow exchange with C2D4 to give the perdeuteriated metallacycle
@a8), indicating that the reverse p-C-C bond cleavage occurs to generate the
bis(a1kene) species. Two of the C-C bond coupling processes, with acetonitrile
Figure 6 The displacement of propylene to give an y2-butadiene species

r 1
L 22 R = 2,6-Pt2CeH3 22de
1 MeCN
23 24 21
Figure 7 Some reactions of the tantalacyclopentane complex 22
Ar
a)
11
AtN=Mo'
N
I 'Ph
PMe,
.Ph
- A
AtN
ArN
25 26
ArN ArN
27 R = CH$'h,
CH&Me3
CH2CMe2Ph
Figure 8 Attempts to synthesise a) benzyne and b) alkylidene complexes of molybdenum
(23)and carbon monoxide (24),are shown in Figure 7, revealing the potential for
new C-C coupling reactions.
As part of his PhD studies, Phil Dyer (now on the faculty at the University of
Leicester) was given the task of extending the relationship to the bis-
(imido)molybdenum system. He was able to synthesise the four-coordinate
olefin and acetylene complexes, [Mo(NBu'),(C,H,)(PMe3)] (17) and
[Mo(NBut),(PhC=CPh)(PMe3)], (18)20but benzyne and alkylidene derivatives
proved elusive. Although potential precursors such as the bis(pheny1) complex
[Mo(NAr),(Ph),(PMe,)] (25) could be made quite readily, the resultant ther-
molysis product was not the anticipated benzyne species. Rather, in this case
biphenyl is generated, along with the MoIVcomplex, [IMO(NA~),(PM~,),]~~ (26,
Figure S), which Phil had isolated earlier from the magnesium reduction of
[Mo(NAr),Cl,(dme)] in the presence of PMe,.,,
A range of dialkyls (27), including the dibenzyl, were also investigated as
potential precursors to alkylidene products, but these proved to be too stable.
28 29 30
31
Figure 9 Chromium(v1)alkylidene species
Heating the samples above 100°C (up to 150°C) resulted in decomposition

rather than in the formation of an isolable alkylidene species2, (Figure 8). In fact,
it was quite some time before a bis(imido) alkylidene complex of the Group 6
metals proved accessible and, to our surprise, chromium proved to be the metal
that provided the solution (Figure 9). Martyn Coles (now on the faculty at the
University of Sussex) synthesised the neopentyl complex, [Cr(NAr),-
(CH2CMeJ2] (28),and found that, in thf solvent at room temperature, a smooth
elimination of neopentane occurred to give the Crvl alkylidene species,
[Cr(=CHCMe,)(NAr),(thf)] (29).24 The PMe, derivative, [Cr(=€HCMe,)-
(NAr),(PMe,)] (30),can be prepared straightforwardly by subsequent treatment
of the thf adduct with one equivalent of the phosphine. These represented
the first Crvlalkylidene species to prove isolable. To our surprise, in the absence
of a donor solvent, the transiently formed alkylidene species is capable
of activating one of the C-H bonds of benzene to give a neopentyl/phenyl
product (31, Figure 9), which was first identified as a result of activation of the
deuterio-benzene NMR solvent.24 A bigger surprise was the fact that this same
C-H bond activation occurs in the presence of acyclic olefins and norbornenes.
This was unexpected since high-valent alkylidene complexes of the other Group
6 metals, molybdenum and tungsten, are invariably efficient olefin metathesis
catalysts.
5 Alkyl and Alkylidene Complexes of Molybdenum:

Routes to Olefin Metathesis Catalysts
The work of O ~ b o r n ,and
~ Schrock26in the late 1980s had established synthetic
entries into low-coordinate molybdenum and tungsten alkylidene species which
were to prove of tremendous significance in the field of olefin metathesis, afford-
ing ‘living’ polymerisation systems for the ROMP (ring-opening metathesis
polymerisation) of cyclic olefin monomers, as well as the more recently develop-
ed application of ring-closing metathesis (RCM). The route they developed to the
catalysts was based on protonation of an imido ligand followed by an a-H
abstraction with elimination of alkane, to give the stable alkylidene complex
(Figure 10a,b). Osborn found that, for tert-butylimido derivatives, the products
A Journey in Metal-Liyand Multiple Bond Chemistry
NBU‘
Buhl
\ \ - Y
NAr
Figure 10 Osborn’s (a), Schrock’s (b) and the selective protonation route (c)to
molybdenum(v1) alkylidene complexes
are usually oils which can lead to handling difficulties, and the methodology was
restricted to alkoxides that are relatively electron-withdrawing. Schrock’s ap-
proach employs precursors containing 2,6-diisopropylphenylimido ligands
whose alkylidene products are, in general, more amenable to manipulation.
However, a drawback in the preparation of arylimido catalysts is the necessity
for triflic acid (CF,SO,H, HOTf), an expensive and potentially hazardous
reagent, to ‘protonate off’ the much less basic arylimido ligand.
During our investigations on the half-sandwich Group 5 metal imido and
Group 6 bis(imido) metal systems, we found that a convenient way of changing
the imido ligand was to treat the relatively electron rich tert-butylimido metal
species with a variety of anilines. This, for example, allowed us to access conveni-
ently a range of new imido complexes of vanadium,27chromium2* and molyb-
d e n ~ m systems.
~ ~ Ed Marshall adapted this methodology to the synthesis of
molybdenum metathesis catalysts (Figure 1Oc). He synthesised bis(imido)molyb-
denum precursors containing a combination of tert-butylimido and 2,6-diisop-
ropylphenylimido ligands and was able to show that a less strong acid such as
pentafluorophenol can be used to ‘protonate off’ selectively the more electron-
rich tert-butylimido group thereby generating the target arylimidomolybdenum
alkylidene species.
6 Olefin Polymerisation Catalysts

The isolobal analogy between cyclopentadienyl and imido ligands held promise
for the development of new a-olefin polymerisation catalysts based on imido
ligation. However, our early endeavours were to prove disappointing. Cationic
Vernon C . Gibson 149
Figure 11 Well-defined cationic chromium alkyl catalystsfor ethylene polymerisation
alkyl complexes based on niobium and molybdenum, of the type

[Nb(NAr)RCp] and [Mo(NAr),R] respectively, gave disappointingly low
-+ +-
activities in ethylene p~lymerisation.~'We tried a tantalum analogue of the

niobium system and this too proved to be of very low activity.,' There seemed
only one way to go for imido systems from there and that was to chromium, a
metal with a track record of producing highly active ethylene polymerisation
catalysts when supported on silica. Bis(imido)chromium chemistry was less well
developed than for its heavier Group 6 congeners molybdenum and tungsten.
Nevertheless, Martin Coles and Chris Dalby, in a project sponsored by BP
Chemicals, soon showed that it was possible to prepare the complexes
[Cr(NBu'),(CH,Ph),] (32) and [Cr(NAr),(CH,Ph),] (33) and they showed that
the cationic monobenzyl species [Cr(NBu'),(CH,Ph)](B(C,F,),) (34) could be
generated using (Ph,C)(B(C,F,),) or (PhNMe,H)(B(C,F,),).32 In the case of the
anilinium reagent, mono- and bis-amine adducts are formed with the liberated
N,N-dimethylaniline. The compounds were tested for ethylene polymerisation
activity in conjunction with BP Chemicals and were found to give reasonable
activities and with negligible loss of performance over three hour^.^' 9 3 2
The relatively short journey from the Group 4 metallocenes on the left hand
side of the transition series to Group 6 bis(imido) metal systems in the middle of
the series has led to a continuing journey towards the late transition metals in a
quest for new olefin polymerisation catalysts. This journey is still in progress and
so an account of the outcome of these studies will have to wait until another
occasion.
7 Distortional Isomerism
This account would not be complete without mention of distortional isomerism, a
term Joe Chatt was to coin in describing two isomeric forms of the six-coordinate
compounds, [Mo(O)Cl,(PMe,Ph),] (35, Figure 12) in a short communication
published in Chem. Cornmun. in 1971, but which later was to spark an intense
debate into the reality or otherwise of the phenomenon of bond stretch isomerism.
We followed the debate with some interest, since we had inadvertently stumbled
across seven-coordinate 0x0 and sulfido compounds of niobium of the type
[Nb(E)Cl,(PMe,),] (E = 0 or S; 36) which seemed to show a closely related
effect.
In their 1971 paper Chatt, Manojlovic-Muir and Muir3, proposed the term
distortional isomerism to describe two forms of [Mo(O)Cl,(PMe,Ph),], one blue
0
PhMe2P- hocCI
~
PhMe2P‘ I PMe2Ph
CI
35 36 E = O , S
Figure 12 Six-coordinate molybdenum and seven-coordinate niobium complexes which

appeared to display the phenomenon of bond-stretch isomerism
with v(Mo=O) at 954 cm-l and the other green with v(Mo=O) at 943 cm-’.
These were members of a range of analogous blue and green oxomolybdenum
compounds with high or low (Mo=O) vibrational frequencies, respectively,
reported earlier by Butcher and Chatt and thought at first to be cis and trans
isomers., X-ray structure determinations showed that blue [Mo(O)Cl,-
(PMe,Ph),] and the green diethylphosphine derivative [v(Mo=O) = 940 cm-’]
in fact possessed similar cis-mer configurations, but with strikingly different
organophosphine orientations and markedly different Mo=O bond lengths.
Based on these observations Chatt and co-workers proposed a new form of
isomerism involving ‘two equilibrium arrangements of ligands which differ in the
distortions of the highly strained coordination polyhedron of the metal’. They
suggested that the blue and green forms of cis-mer [Mo(O)Cl,(PMe,Ph),],
which they concluded had different organophosphine orientations (resulting in
C, symmetry in the blue complex and C, in the green), exhibited this type of
distortional isomerism.
In his study of the system, Parkin was able to show that small M=O bond
length differences can be caused by an isomorphous contaminant,
[MoCl,(PMe,Ph),] in this case.36 Also, since [MoCl,(PMe,Ph),] is yellow, a
mixture of blue [Mo(O)Cl,(PMe,Ph),] and yellow [MoCl,(PMe,Ph),] would
readily account for the green ‘isomer’. This explanation, however, did not pro-
vide an answer as to why two stretches were seen in the IR spectra recorded by
Chatt (since only one would be expected), and led Enemark37to conclude that
Chatt’s assignment of two Mo=O stretches had been erroneous. It was on this
basis that the simple and persuasive contaminant theory gained widespread
acceptance as providing an explanation for all the spectroscopic and crystallo-
graphic observations surrounding Chatt’s distortional isomer system.
We, however, were sceptical that Chatt would have made an error of this kind,
especially given the importance attached to IR spectroscopy as a characterisa-
tion technique at the time the original study was carried out. We resolved to
investigate further. Mary McPartlin (who had determined the structures of our
niobium 0x0 compounds) and I discussed the issue with Joe, and learnt that
Tony Butcher, who had carried out the original work in Chatt’s Sussex labora-
tory, had made a successful career in business, and was living in Cambridgeshire.
After making contact, Tony was adamant that there had been two different
stretches in the IR spectra of the blue and green ‘isomers’. Indeed, his keenness to
resolve the issue led him to make available various spectroscopic facilities at his
Company, and he even volunteered to do some of the work himself! Oliver
Robinson, a PhD student in my group set about repeating the synthesis of the
Chatt-Butcher compounds following the original procedure and was able to
make and grow crystals of the blue isomer (Figure 13). A structure determination
carried out by Mary McPartlin revealed different phosphine orientations from
the blue isomer studied by Parkin and Enemark, and moreover the new blue
isomer had a Mo=O stretch at 954 cm-’, the same value observed by Chatt. We
then determined the structure of the blue isomer prepared under the conditions
described by Enemark and Parkin and also the structures of the two bromo
isomer analogues (Figure 14). This confirmed beyond all doubt the existence of
two isomeric forms of pure [Mo(O)X,(PMe,Ph),] (X = C1 or Br).38Thus, two
stretches in the IR could be explained by the presence of two different blue
isomers, and that the only difference between these two isomers is the orientation
of the phosphine ligands. Whatever one’s view of the phenomenon of bond
stretch isomerism the description distortional isomers would still seem to be an
appropriate description of the relationship between these compounds.
~ 0 - rl
EtOH
~ 1 ~
green .-!?!?!-+
PMe2Ph blue isomer
(excess)
j H*0
u(Mo=O) = 943 crn-’
(excess) ~(MO=O) = 954 cm-1
Figure 13 Routes to the two blue isomers of [Mo(O)C1,(PMe2Ph,),]
Figure 14 The structures of the two forms of [Mo(O)X,(PMe,Ph),] ( X = Cl or Br),

illustrated for X = Br: (a) the structure of C, symmetry for the high-frequency
form and (b) the structure of C, symmetry for the low-frequency form
8 Acknowledgements
This author is indebted to the many students and postdocs who, through their
dedication and hard work, have enlightened and enlivened the journey described
here. The largest debt of gratitude, however, is owed to Joe Chatt for the many
fundamental insights he provided into the stucture, bonding and reactivity of
transition metal coordination and organometallic compounds, and for his
boundless enthusiasm which has inspired so many others to enter the field.
9 References
1 E. 0.Fischer and S. Vigoureux, Chem. Ber., 1958,91,1342.
2 V. C. Gibson, T. P. Kee and A. Shaw, Polyhedron, 1988,7,2217.
3 V. C. Gibson and T. P. Kee, J . Chem. SOC., Chem. Commun., 1989,656.
4 A. Shaw, PhD Thesis, University of Durham, 1989.
5 P. Jernakoff, C. de M. de Bellefon, G. L. Geoffroy, A. L. Rheingold and S. J. Geib,
6 D. D. Devore, J. D. Lichtenhan, F. Takusagawa and E. A. Maatta, J . Am. Chem. SOC.,
1987,109,7408.
7 J. M. Mayer, C. J. Curtis and J. E. Bercaw, J . Am. Chem. SOC., 1983,105,2651.
8 V. C. Gibson, D. N. Williams, W. Clegg and D. C. R. Hockless, Polyhedron, 1989,8,
1819.
9 D. N. Williams, J. P. Mitchell, A. D. Poole, U. Siemeling, W. Clegg, D. C. R. Hockless,
P. A. O’Neil and V. C. Gibson, J . Chem. SOC.,Dalton Trans., 1992,739.
10 D. S. Williams, M. H. Schofield, J. T. Anhaus and R. R. Schrock, J . Am. Chem. SOC.,
1990,112,6728.
11 J. W. Lauher and R. Hoffmann, J . Am. Chem. SOC., 1976,98,1729.
12 J. K. Cockcroft, V. C. Gibson, J. A. K. Howard, A. D. Poole, U. Siemeling and C.
Wilson, J . Chem. SOC. Chem. Commun., 1992,1668.
13 M. C. W. Chan, J. M. Cole, V. C. Gibson, J. A. K. Howard, C. Lehmann, A. D. Poole
and U. Siemeling, J. Chem. SOC.Dalton Trans., 1998, 103.
14 A. D. Poole, V. C. Gibson and W. Clegg, J . Chem. SOC.Chem. Commun., 1992,237.
15 U. Siemeling and V. C. Gibson, J . Organomet. Chem., 1992,426, C25.
16 A. D. Poole, PhD Thesis, University of Durham, 1992.
17 For an early review see, G. Erker, C. Kriiger and G. Miiller, Adu. Organomet. Chem.,
1985,24, 1.
18 V. C. Gibson and A. D. Poole, J . Chem. SOC.Chem. Commun., 1995,2261.
19 P. W. Dyer, V. C. Gibson, J. A. K. Howard, B. Whittle and C. Wilson, J . Chem. SOC.
20 P. W. Dyer, V. C. Gibson, J. A. K. Howard, B. Whittle and C. Wilson, Polyhedron,
1995,14, 103.
21 P. W. Dyer, PhD Thesis, University of Durham, 1993.
22 P. W. Dyer, V. C. Gibson and W. Clegg. J . Chem. SOC.Dalton Trans., 1995,3313.
23 G. L. P. Walker, PhD Thesis, Imperial College, London, 1997.
24 M. P. Coles, V. C. Gibson, W. Clegg, M. R. J. Elsegood and P. A. Porrelli, Chem.
Commun., 1996,1963.
25 G. Schoettel, J. Kress and J. A. Osborn, J . Chem. SOC.Chem. Commun., 1989,1062.
26 R. R. Schrock, J. S. Murdzek, G. C. Bazan, J. Robbins, M. DiMare and M. O’Regan, J.
Am. Chem. SOC.,1990,112,3875.
27 M. C. W. Chan, J. M. Cole, V. C. Gibson and J. A. K. Howard, 1. Chem. Soc., Chem.
Commun., 1997,2345.
28 M. P. Coles, C. I. Dalby, V. C. Gibson, W. Clegg and M. R. J. Elsegood, Polyhedron,
1995,14,2455.
29 A. Bell, W. Clegg, P. W. Dyer, M. R. J. Elsegood, V. C. Gibson and E. L. Marshall, J .
Chem. Soc. Chem. Commun., 1994,2247.
30 A. Bell, W. Clegg, P. W. Dyer, M. R. J. Elsegood, V. C. Gibson and E. L. Marshall, J .
Chem. SOC.Chem. Commun., 1994,2547.
31 M. P. Coles, C. I. Dalby, V. C. Gibson. I. R. Little, E. L. Marshall, M. H. R. da Costa
and S. Mastroianni, J . Organomet. Chem., 1999,591,78.
32 M. P. Coles, C. I. Dalby, V. C. Gibson, W. Clegg and M. R. J. Elsegood, J . Chem. SOC.
33 A. Bashall, V. C. Gibson, T. P. Kee, M. McPartlin, 0. B. Robinson and A. Shaw,
Angew. Chem., 1991,103,1021; Angew. Chem., Int. Ed., Engl., 1991,30,980.
34 J. Chatt, Lj. Manojlovic-Muir and K. W. Muir, J . Chem. Soc. Chem. Commun., 1971,
655.
35 A. V. Butcher and J. Chatt, J . Chem. Soc. A , 1970,2652.
36 K. Yoon, G. Parkin and A. L. Rheingold, J . Am. Chem. Soc., 1991,113,1437.
37 P. J. Desrochers, K. W. Nebesny, M. J. LaBarre, S. E. Lincoln, T. M. Loehr and J. H.
Enemark, J . Am. Chem. Soc., 1991,113,9193.
38 A. P. Bashall, S. W. A. Bligh, A. J. Edwards, V. C. Gibson, M. McPartlin and 0. B.
Robinson, Angew. Chem., Int. Ed., Engl., 1992,31, 1607.
Synthesis, Characterisation and
Catalytic Activity of
Heterobimetal Complexes
MAURICE ABOU RIDA AND ANTHONY K. SMITH
Laboratoire de Chimie Organometallique de Surface, CNRS-UMR 9966, Ecole
Superieure de Chimie Physique Electronique de Lyon, 43 Boulevard du 11
Novembre 1918,69619 Villeurbanne Cedex, France
1 Introduction
Heterobimetal complexes have been the subject of much interest in view of the
catalytic potential associated with the proximity of the metal atoms even in the
absence of any metal-metal bond.’ Ring-opening (or metal-insertion) reactions
of chelated bidentate phosphine ligands provide a useful route to ligand-bridged
heterobimetal c ~ m p l e x e s . ~In- ~these reactions the reactant is usually a mono-
nuclear complex containing a chelating phosphine as part of a four-membered
ring. On reaction with an appropriate metal complex, this four-membered ring
opens to produce a ligand-bridged bimetal complex in which the bidentate
ligand becomes part of a less-strained five-membered ring. We have exploited
this synthetic route to produce a range of heterobimetal complexes, using both
bidentate and tridentate phosphine ligands. In addition, a major goal of this
research was to demonstrate that two or more metal centres can cooperate in the
homogeneous catalysis of the hydroformylation of alkenes.
2 Synthesis of Ruthenium-Rhodium Heterobimetal

Complexes
The synthesis of a number of heterobimetal complexes of the type
[(RhRuC1,Cp)(p-CO),(p-q2-L-L)](Rh-Ru) (L-L = dppm, [dppm = bis-
(diphenylpho~phino)methane]~ Ph,PC(=CH,)-PPh,,’ Ph,PNHPPh, or dppp6
[dppp = bis(dipheny1phosphino)propanel) has been carried out. These com-
plexes are readily prepared by treatment of the chelating phosphine complexes
Maurice Abou Rida and Anthony K . Smith 155
Y = CH2. NH, CMe2, CHPPh2
Figure 1 Synthesis of ruthenium-rhodium heterobimetal complexes
Me2SAuCI
-CI
Figure 2 Synthesis route to the ruthenium-rhodium-gold trimetal complex 3
,
[RuCl(L-L)Cp] with [RhCl(CO),] (Figure 1). The reaction proceeds rapidly
and quantitatively at room temperature. Many diphosphines that give four-
membered rings upon chelation can be used in this reaction. Ligands that chelate
to a metal atom forming a five-membered ring do not generally undergo this
ring-opening reaction, except for dppe, which undergoes a ring-opening reaction
to produce a rather unstable heterobimetal complex which was characterised
spectroscopically, but which could not be isolated in a pure state. These hetero-
bimetal complexes, which are not very soluble in organic solvents, may be
regarded as complexes of a rhodium(1) anion and a ruthenium@) cation with a
donor-acceptor ruthenium-rhodium bond.
We have also reported some of the reactions of these complexes with CO
under a range of conditions of temperature and pressure. A notable reaction is
the completely reversible reaction of [CpRu(p-CO),(p-q' :q'-L-L)RhCl,] with
CO to give [RhCl,(CO),] - and [Ru(CO),(ql-L-L)Cp] + . 7
156 Synthesis, Characterisation and Catalytic Activity of Heterobimetal Complexes
tripod 1 ,/
'Ph2 1OTf-
Figure 3 Synthesis of the q3-tripod complex 4 and the reaction of complex 4 with
[RhCl,(CO),]-. HOTf = HOSO,CF,
The tridentate ligand l,l,l-tris(diphenylphosphino)methane,CH(PPh,),,

known as tripod, offers attractive potential and a variety of modes for coordinat-
ing to transition metals. Many complexes in which CH(PPh,), can act as an
y l-monodentate ligand', an y2-chelating ligand*>', an q3-p2-bridging-chelating
Iigand8," or an y3-p,-bridging ligand1lP1' have been synthesised. In addition,
the dangling arm or arms provide potential coordination sites for other metals.
As an example of the utility of tripod in constructing heterometal complexes,
we have recently investigated the reaction of the monometal complex [RuCl(y2-
HC(PPh,),)Cp] 1 with [RhCl(CO),],. This reaction gives rise to the hetero-
binuclear complex [CpRu(p-CO), ( p q l :y '-HC(PPh,),}RhCl,] 2. Subsequent
reaction of complex 2 with [AuCl(SMe,)] gives the heterotrinuclear complex
[CpRu(p-CO),(p-yl :y l : y1-HC(PPh,),)RhAuCI3] 3 (Figure 2).
The y2-tripod complex 1 is converted into the y3-tripod complex [RuCp(y3-
HC(PPh,),)](OTf) 4 by treatment with Ag(0Tf). When complex 4 is treated with
the anionic rhodium complex [RhCl,(CO),]-, a ring-opening and a P-C bond
cleavage reaction occurs to generate the known p-dppm heterobimetal complex
[CpRu(p-CO),(p-yl :yl-dppm)RhC1,] 5 (Figure 3).
3 Hydroformylationof Oct-1-ene
The hydroformylation of oct-1-ene was studied at 115°C using complexes 1 and
5 as catalysts. The reaction was carried out with an oct-1-ene concentration of
9.55 x mol ~ m - catalyst
~ , concentration of 1.9 x lop5 mol cm-3 and a
+
total pressure of (Pco PH2)60 atm (at ambient temperature; CO: H,, 1: 1).
With complex 5 as the catalyst, straight chain and branched aldehydes are
formed. These are the only reaction products under these reaction conditions
(Figure 4). About 50% of the oct-1-ene is converted after 40 hours. Nonanal was
the major product with a relatively constant +so ratio of 2.8. At higher tempera-
tures, isomerisation and hydrogenation reactions take place in addition to
hydroformylation.
The catalytic activity and selectivity of 2 was also studied towards the hydro-
formylation of oct-1-ene. The mole number-time profile of oct-1-ene and reac-
tion products is shown in Figure 5. With this bridging tripod complex the
activity is about the same as that of dppm complex 5 (50% conversion after 40
hours); however, the n/iso ratio is much higher, between 8 and 9. The presence of
the extra phosphine on the central carbon atom clearly has a marked effect on
the selectivity of the reaction.
Since the regioselectivity (expressed as +so ratio) to linear aldehyde using the
tripod bridging complex 1 is higher than that of the analogous bridging dppm
complex 5, we suggest that the presence of the free diphenylphosphine group on
the central carbon atom in 1 has a strong effect in controlling the selectivity of the
reaction. This may be explained essentially by steric effects.
This result is in accordance with virtually all other phosphine- or phosphite-
coordinated rhodium hydroformylation catalysts, where an excess of phosphine
0.01
0.009
0.008
u)
u
$ 0.007
c
g 0.006
5
w-
0.005
0 0.004
za,
0.0°3
0.002
0.001
0
0 20 40 60 80 100
Time in hours
+ oct-1-ene A nonanal .A 2-methyloctanal
Figure 4 Hydroformylation of oct-1 -ene with complex 5 as catalyst

0.01
0.009
0.008
v)
+
5K 0.007
8 0.006
5 0.005
.
I
::
-
a)
0.004
- 0.003
r" 0.002
0.001
0
0 20 40 60 80 100
Time in hours
+ oct-1-ene A nonanal 5. 2-methyloctanal
Figure 5 Hydroformylation of oct-1 -ene with complex 2 as catalyst
(PPh,) is needed to maintain good selectivity or stability. The need for excess
PPh, in mononuclear rhodium catalysts arises from the relatively weak
Rh-PPh, (or phosphite) bonding. In order to maintain the coordination of two
PPh, ligands, which are required for good regioselectivity, a large excess of PPh,
is required to force the dissociation equilibrium to favour [RhH(CO)(PPh3)2].20
In our case the free dangling phosphino group of the tripod seems to play an
analogous role.
It is important to note that complexes 1 and 5 are recovered unchanged from
the reaction mixture following catalysis, and 31P NMR spectroscopy of the
product solution showed no other phosphorus-containing species present.
We explored the hydroformylation reaction for monometal model complexes
that represent one 'half' of the bimetal catalysts 1 or 5. These tests give us an idea
about whether each metal centre is functioning as a conventional mononuclear
catalyst or whether there is some cooperativity. Thus the catalytic activity and
selectivity of the complexes [RuCl(q2-Ph,PCH,PPh,)Cp], [RuCl{q2-
HC(PPh,),}Cp] and [(RhCl(CO),},] were studied under the same conditions
as described above.
The complexes [RuCl(q ,-Ph, PCH,PPh,)Cp] and [RuCl { q ,-HC( PPh,),} -
Cp] are extremely poor hydroformylation catalysts. They showed less than 0.5%
conversion of alkene to aldehyde after 30 hours, linear to branched ratios of one
or less, and undesirable amounts of alkene isomerisation and hydrogenation
products.
The complex [{RhCl(CO),),] is a much more active catalyst with a relatively
high turnover frequency of 156 (at 50% conversion), and 97% conversion of
oct-1-ene after three hours. The products are exclusively the straight chain and
the branched aldehydes with an nliso ratio of approximately one.
The reactivity of [{RhCl(CO),),] in the presence of various amounts of added
tripod was also investigated. It was found that the presence of an excess of tripod
had very little effect on the n/iso ratio of the aldehydes, with the maximum not
exceeding 1.5 even with a Rh/tripod ratio of 1: 4.
The activity and regioselectivity of 1 and 5 therefore contrast to those of
monometal rhodium or ruthenium complexes, and the regioselectivity of com-
plex 1 is particularly noteworthy. These results indicate that the active species
uses some sort of bimetal cooperativity to effect high regioselectivities.
Any discussion of mechanisms should take into account the early work on
cobalt-catalysed hydroformylation.21 A monometal mechanism was proposed
that has become the generally accepted pathway 2 2 for both cobalt and rhodium
catalysts. A more speculative mechanism was also suggested involving an inter-
molecular hydride transfer from [CoH(CO),] to [Co(acyl)(CO),]. Elimination
CI
-Alkene
1
+ Alkene
Rh
Figure 6 Proposed mechanism for alkene hydroformylation with complex 2 as catalyst

160 Synthesis, Characterisat ion and Catalytic Activity of Heterobimetal Complexes
of the aldehyde product then produces [Co,(CO),], which reacts with H, to

break the Co-Co bond to reform two [CoH(CO),] molecules. This suggests an
interesting mechanistic possibility for the heterobimetal catalysts 1 and 5.
The constrained proximity of the two metal centres, held together by the
tripod ligand, should increase the probability of an intramolecular hydride
transfer. Thus a mechanism can be proposed (Figure 6) in which bimetal
cooperativity, via an intramolecular hydride transfer, facilitates the elimination
of aldehyde from the acyl intermediate.
The first steps in the proposed mechanism are essentially the same as those
established for monometal Rh/PPh, catalysts, except that the proposed addition
of H, oxidises two metal centres. In monometal systems, the final steps are the
addition of H, to a rhodium(1) to produce a rhodium(II1) dihydride species that
can then eliminate aldehyde product. The ruthenium-rhodium intermediate
avoids this problem by having a proximate Ru-H moiety, which can intra-
molecularly transfer a hydride to facilitate the aldehyde elimination. Thus the
final steps of the mechanism are H and CO bridge formation between the
1
.Alkene + Alkene
Figure 7 Alternative mechanismfor aEkene hydroformylation using complex 2 as catalyst

ruthenium and the rhodium atoms and then elimination of the aldehyde and
reformation of 5. A similar mechanism can be proposed for complex 1.
The bimetal cooperativity in the proposed mechanism represents a very
effective way of performing hydroformylation. The fundamental concept of a
hydride transfer between two metal centres has been studied and shown to occur
in stoichiometric model reactions by numerous groups.23 Hidai and Mat-
~ u z a k a ,attributed
~ the synergistic effect observed in the hydroformylation of
olefins by the [Co,(CO),]-[Ru3(CO), ,] bimetal system to a 'dinuclear reductive
elimination of aldehydes from cobalt acyls and ruthenium hydride(s)'. [Rh,(p-
SBu'),(CO),(PPh,),] is another hydroformylation catalyst for which bimetal
cooperativity has been p r ~ p o s e d .However,
~ the reaction rates and regioselec-
tivities of [ R ~ , ( ~ - S B U ' ) ~ ( C O > , ( Pvery
P ~ ~closely
) ~ ] resemble those of Rh/PR,
monometal catalysts, indicating that the active catalyst may be monometal in
nature,26 quite unlike 1 and 5.
A mechanism in which the catalytic reaction takes place entirely at the
rhodium atom, as shown in Figure 7, cannot however be discounted. In this case
the cooperativity between the metal atoms arises from the ruthenium moiety
acting as a labile ligand.
4 Conclusion
Bidentate or tridentate phosphine ligands that chelate to a metal atom to form a
four-membered ring can undergo facile ring-opening reactions to produce
heterobimetal complexes. Two of these complexes have been shown to be active
and selective hydroformylation catalysts. The heterodinuclear complexes are
recovered unchanged at the end of the reaction. Two possible mechanisms for the
catalytic reaction have been proposed, one in which the initial addition of H,
involves the formation of both Rh-H and Ru-H bonds, while the other mechan-
ism involves oxidative addition of H, to the rhodium atom, with the ruthenium
moiety playing the role of a labile ligand in the catalytic cycle. The bridging
diphosphine or triphosphine ligands increase the selectivity of the reaction
towards the linear aldehyde. Catalytic studies of possible monometal fragments
from the bimetal complex indicate that these fragments are not the active species
in the reaction. Further, in situ, studies will need to be carried out to determine
the mechanistic pathway.
5 References
1 P. Braunstein, L. A. Oro and P. R. Raithby, eds., Metal Clusters in Chemistry, Wiley,
New York, 1999; C. P. Casey, R. M. Bullock, W. C. Fultz and A. L. Rheingold,
Organometallics, 1982,1,1591; R. G. Finke, G. Gaughan, C. Pierpont and M. E. Cass,
J . Am. Chem. Soc., 1981,103,1394; B. Chaudret, B. Delavaux and R. Poilblanc, Coord.
Chem. Rev., 1988,86, 191 and refs. therein; A. Maisonnet, J. P. Farr, M. M. Olmstead,
C. T. Hunt and A. L. Balch, Inorg. Chem., 1982,21,3961.
2 B. L. Shaw, M. J. Smith, G. N. Stretton and M. Thornton-Pett, J . Chern. Soc., Dalton
Trans., 1988,2099.
3 B. Delavaux, B. Chaudret, F. Dahan and R. Poilblanc, Organometallics, 1985,4,935.
4 R. H. Dawson and A. K. Smith, J . Organomet. Chem., 1986,309, C56.
5 M. P. Brown, D. Burns, R. Das, P. A. Dolby, M. M. Harding, R. W. Jones, E. J.
Robinson and A. K. Smith, J . Chem. SOC.,Dalton Trans., 1991,351.
6 J. V. Barkley, J. C. Grimshaw, S. J. Higgins, P. B. Hoare, M. K. McCart and A. K.
Smith, J . Chem. SOC.,Dalton Trans., 1995,2901 and references therein.
7 P. S. Bearman, A. K. Smith, N. C. Tong and R. Whyman, Chem. Commun., 1996,2061.
8 J. T. Mague and S. E. Dessens, J . Organomet. Chem., 1984,262,347.
9 K. J. Beckett and S. J. Loeb, Can. J . Chem., 1988,66 1073.
10 A. A. Bahsoun, J. A. Osborn, P. H. Bird, D. Nucciarone and A. V. Peters, J . Chem. SOC.,
11 D. J. Darensbourg, D. J. Zalewski, A. L. Rheingold and R. L. Durney, Inorg. Chem.,
1986,25,3281.
12 D. J. Darensbourg and D. J. Zalewski, Organometallics, 1985,4,92.
13 D. J. Darensbourg, D. J. Zalewski and T. Delord, Organometallics, 1984,3, 1210.
14 J. A. Clucas, M. M. Harding, B. S. Nicholls and A. K. Smith, J . Chem. Soc., Chem.
Commun., 1984,319.
15 A. A. Bahsoun, J. A. Osborn, J. P. Kintzinger, P. H. Bird and U. Siriwardane, Nouv. J .
Chem., 1984,8, 125.
16 J. A. Osborn and G. G. Stanley, Angew. Chem., Int. Ed. Engl., 1980,19,1024.
17 A. A. Bahsoun, J. A. Osborn, C. Voelker, J. J. Bonnet and G. Lavigne, Organometallics,
1982,1, 1114.
18 M. M. Harding, B. S. Nicholls and A. K. Smith, J . Chem. Soc., Dalton. Trans, 1983,
1479.
19 M. M. Harding, B. S. Nicholls and A. K. Smith, J . Organomet. Chem., 1982,226, C17.
20 R. L. Pruett and J. A. Smith, J . Org. Chem., 1969,34,327; J. M. Brown and A. G. Kent,
J . Chem. SOC.,Perkin Trans. I I , 1987, 1597.
21 D. S. Breslow and R. F. Heck, Chem. Ind. (London), 1960,467; R. F. Heck and D. S.
Breslow, J . Am. Chem. SOC., 1961,83,4023.
22 M. F. Mirbach, J . Organomet. Chem., 1984,265, 205; W. R. Moser, A h . Chem. Ser.,
1992,230, 3.
23 W. D. Jones, J. M. Huggins and R. G. Bergman, J . Am. Chem. Soc.,1981,103,4415; M.
J . Nappa, R. Santi and J. Halpern, Organometallics, 1985, 4, 34; B. D. Martin, K. E.
Warner and J. R. Norton, J . Am. Chem. Soc., 1986, 108, 33; K. E. Warner and J. R.
Norton, Organometallics, 1985,4,2150.
24 M. Hidai and H. Matsuzaka, Polyhedron, 1988,7,2369.
25 P. Kalck, Polyhedron, 1988,7,2441.
26 R. Davis, J. W. Epton and T. G. Southern, J . Mol. Catal., 1992,77, 159.
Tethered Arene Complexes of
Ruthenium
MARTIN A. BENNETT" AND JOANNE R. HARPERa9b
aResearch School of Chemistry, Australian National University, Canberra,
ACT 0200, Australia
Department of Chemistry, University of Cambridge, Lensfield Road,
Cambridge, CB2 lEW, UK
1 Introduction
In this memorial volume devoted to one of the pioneers of modern coordination
and organometallic chemistry, it is worth recalling that Joseph Chatt's classic
papers in the 1950s dealing with the complexes of platinum(I1) with alkenes and
with tertiary phosphines appeared at about the same time as those of another
great contemporary, Ronald (later Sir Ronald) Nyholm at University College
London, on the closely related complexes of tertiary arsines with platinum@)
and the later transition elements.' Although Chatt and Nyholm worked inde-
pendently during their careers, they were firm friends and exchanged information
and ideas, especially during the early years of the development of the subject.
There are other interesting historical similarities and connections between them.
Both spent their formative childhood years in mineral-rich areas, Chatt in
Cumberland, in the north west of England, Nyholm in Broken Hill, New South
Wales, and their interest in minerals undoubtedly contributed to their interest
in science.2 The bidentate ligand o-phenylenebis(dimethylarsine), o-
C,H,(AsMe,),, which Nyholm used to such good effect to stabilise unusual
oxidation states and stereochemistries for the d-block elements, had been made
first by Chatt in 1939 when he was a PhD student with F. G. Mann in Cam-
bridge.3 Nyholm inherited his interest in the coordination chemistry of tertiary
arsines from the supervisor of his BSc. Honours project at the University of
Sydney, G. J. Burrows. The latter had worked earlier with E. E. Turner who, like
F. G. Mann, had been a Research Assistant at the University of Cambridge,
working on organoarsenic chemistry with Sir William Jackson Pope.4
Potentially chelating ligands containing both alkene and tertiary phosphine
or arsine donor centres were first synthesised in Nyholm's group in 1961 with the
164 Tethered Arene Complexes of Ruthenium
ideas of stabilising alkene coordination and studying the reactivity of the alkene
when it was either coordinated to, or held in the vicinity of, a transition metal
atom.,., It was Chatt who first suggested that the anomalous platinum(1v)
compounds that Nyholm's group had isolated from the addition of bromine
to the platinum(I1) complexes of (2-~inylphenyl)dimethylarsine, [PtBr,(o-
Me,AsC,H,CH=cH,),], and (2-allylphenyl)dimethylarsine, [PtBr,(o-
Me2AsC6H4CH,CH=CH,),], contained a metal-carbon o-bond in a chelate
ring; this led to a rare joint publication by these two eminent c h e m i ~ t s . ~
Ligands containing both C- and classical N-, 0-,P- or As-donors are now
commonplace and have found extensive application in the study of 'strapped' or
'tethered' cyclopentadienyl complexes, for Metal-arene complexes
are less numerous and, -in general, less stable than their metal-cyclopentadienyl
counterparts, in part because the neutral arene is lost more easily from the
coordination sphere. Thus, tethered arene complexes offer the possibility of
stabilising arene coordination for a range of metals and oxidation states. Mirkin
~ ' ~shown that Ph,P(CH,),XPh (X = CH, or 0)form
and c o - ~ o r k e r s ' ~have
tethered arenerhodium(1) cations such as [Rh(ql : q6-Ph2P(CH2),XC6H,}( q l -
Ph,P(CH,),XC,H,}], which undergo reversible electrochemical one-electron
oxidation, presumably generating the corresponding arenerhodium(I1) cations.
These appear to be kinetically more stable than their unstrapped counterparts,
but they have not so far been isolated.
Although complexes of the type [RuX,(PR,)(q'-arene)] (X = C1 or Me) are
'
known to undergo electrochemical one-electron ~ x id a tio n ,'~ ' the arene in the
oxidised species is likely to be labile, rendering isolation difficult. One way to
circumvent this problem is to use tethered arene-phosphine ligands, the idea
being that dissociation of the arene from the higher-oxidation-state complex
may be slow enough to allow the complex to be isolated. Similarly, chelating
unsaturated tertiary amines and tertiary phosphines such as o-
CH,=CHC,H,NMe,, CH,=CH(CH,),PMe, and o-PhC=cC,H,NMe, have
been used to generate chelation-stabilised complexes of bis(acety1acetonato)-
ruthenium@) and -rutheniurn(m).' 6 ,
2 Results and Discussion
The ligands employed so far in our work are C,H,(CH,),PR, (R = Ph or Me),
C,Me,(CH,),PPh,, 2,4,6-Me,C6H,(CH,),PPh, and C,H,SiMe,CH,PPh,.
They have been made in yields of 40-80% by the standard methods outlined in
Schemes 1 and 2. The reaction of Ph,PCl with C,H,SiMe,CH,MgC1 gave some
diphenylmethylphosphine as a result of Si-CH, bond cleavage; this was separ-
ated from C,H,SiMe,CH,PPh, by vacuum sublimation.
X = CH2, n - 2 , Y = Br X = CH2, n= 2, R = Me or Ph
X = SiMe2, n= 1, Y = CI X = SiMe,, n= 1, R = Ph
Scheme 1
Martin A . Bennett and Joanne R. Harper 165
R R=H,Me
k R
Reagents: (i) Br(CI+),Br, CuBr, hmpdthf; (ii) Mg, thf; (iii) PbPCI
Scheme 2
The methyl o-toluate complex [RuCl,(~6-1,2-MeC,H4~~2Me)]2 (1) em-

ployed in earlier studies on planar chiral areneruthenium complexes'8 proved to
be a suitably labile precursor to the tethered arene complexes. It reacts with the
donors mentioned above in a 1:2 mol ratio in dichloromethane at room tem-
perature to give quantitatively the corresponding P-bonded adducts [RuCl,(y6-
1,2-MeC6H,CO,Me)(L)], which lose the methyl o-toluate on heating in
dichloromethane or dichloromethane-thf at 120"C for 24-72 h. The resulting
tethered complexes 2-6 can be isolated in the form of air-stable, orange, crystal-
line solids in yields ranging from 6O-8OYO for C,H,(CH,),PR, [R = Me (2) or
Ph (3)] and C,H,SiMe,CH,PPh, (4), through ca. 18% for 2,4,6-
Me,C,H,(CH,),PPh, (5), to ca. 7% for c,Me,(CH,),PPh, (6) (Schemes 3 and
4). In the case of 6, the yield can be increased to 35% by use of di-n-butyl ether in
X = CH2, n= 2, R = Me (2)
X = CH2, n= 2, R = Ph (3)
X = SiMe2, n = 1, R = Ph (4)
Scheme 3
' 'P- (CH&

Ph2 vMe
Me' 'R' R' = H (5),Me (6)
Scheme 4
place of dichloromethane. The beneficial effect of ether solvents has also been
observed in the synthesis of (@-arene)chromium tricarbonyls.' 9,20 Attempts to
form the tethered complexes 2 and 3 by UV-irradiation of solutions of the methyl
o-toluate complexes at room temperature led only to decomposition. In the
thermal reaction, the methyl o-toluate complex precursors of products 2-4 need
not be isolated but can be generated in situ from solutions of complex 1 and the
ligand.
Smith and Wright2' have reported that complex 3 is formed in 50%
yield by heating the p-cymene complex [RuC12(r6-1,4-MeC,H4CHMe,)-
(Ph,P(CH,),C,H,)] in chlorobenzene at 130°C and, in higher yield, by exhaus-

tive anodic oxidation and subsequent spontaneous in situ reduction of the
presumed labile ruthenium(II1) species. However, we have consistently been
unable to reproduce the thermal reaction, either under the stated conditions or in
CH,Cl,/thf at 120°C. In refluxing chlorobenzene the p-cymene is displaced and
we have obtained NMR spectroscopic evidence for the formation of a y6-
chlorobenzene complex, but attempts to isolate it failed.
Tethered areneruthenium(I1) complexes similar to those described here have
recently been investigated by several groups as catalyst precursors for ring-
opening metathesis polymerisation and for cyclopropanation, so it is important
to develop reliable syntheses. During the course of our work, two groups have
reported that tethered complexes such as [RuCl,(yl :y6-
Cy,P(CH,),C,H,)] (Cy = cyclohexyl) can be obtained in 80-90% yield by
heating the p-cymene P-donor adduct in chlorobenzene at 120-140°C. Possibly
the formation of the tethered complex in this case is favoured by the bulky
cyclohexyl groups on phosphorus, cf: the promotion of the formation of four-
membered ring cyclometallated complexes by bulky substituents such as t-butyl
on p h o s p h ~ r u sOur
. ~ ~results are, however, in agreement with those reported by
Ward et al.25-27and Rieger et a1.28,who have used the more labile ruthenium@)
complex of an aromatic ester, ethyl benzoate, as a precursor to tethered
areneruthenium(I1) complexes containing the CH,CH,PPh, strap.
An alternative method of generating complexes containing a (CH,),PPh,
strap has recently been reported in which the diphenylvinylphosphine adduct of
a (methylarene)ruthenium(II) complex undergoes an intramolecular base-promo-
ted Michael addition reaction (Scheme 5).29
R R
KOBU'
MeCN
R = H, pMe, pCHMf+, 3,5-Me2 or Me,
Scheme 5
The shifts to low frequency of the 'H and 13C NMR resonances associated
with the arene ring in complexes 2-6 provide clear evidence for arene coordina-
tion and this has been confirmed by single crystal X-ray studies. The complexes
show a typical half-sandwich structure with Ru-Cl distances of 2.40-2.42 A and
Ru-P distances of ca. 2.30 A. The Ru-C(arene) distances trans to chloride
(2.16-2.21 A) are significantly less than those trans to the phosphorus donor
(2.24-2.28 A).These features are also evident in the structure of the non-tethered
complex [RUC1,(PMePh2)(y"-C6H6)]30 and reflect the relative trans-influences
of C1 and PR,. In all the complexes containing Ar(CH,),PR,, the trimethylene
strap allows simultaneously without strain an almost trigonal geometry for the
RuC1,P fragment and coplanarity of the benzylic carbon atom with the arene
Martin A. Bennett and Joanne R. Harper 167
Table 1 Electrochemical data
Compound E1/2 v AEp (mV s-')
+ 1.26" 70
+ 1.32" 60
+ 1.34" 60
+ 1.2Ob 80
+ l.lob 80
a Measured in 0.5 M (Bu,N)PF,/CH,Cl, at 293 K with a scan rate of 100 mV sC1, referenced to
Ag/AgCI.
* Measured in 0.2 M (Bu,N)PF,/CH,Cl, at 253 K with a scan rate of 100 mV sK1, referenced to
Ag/AgCl.
carbon atoms. In contrast, the two atom strap in complex 4 causes a bending of
the Si-C(C,H,) bond out of the aromatic plane by ca. 14".
As expected, the presence of the tether does hinder the release of the arene from
the coordination sphere at the ruthenium@) level. Thus, whereas benzene is
displaced completely from the unstrapped complex [RuCl,(q6-C,H6)(q1-
Ph,P(CH,),Ph)] in refluxing acetonitrile over 48 hours, displacement of the
tethered arene in complex 3 does not proceed to completion, even after 11 days;
in both cases, the product is a mixture of cis- and trans-isomers of [RuCl(ql-
Ph,P(CH,),Ph)(NCMe),]Cl.
All the tethered complexes show fully reversible, one-electron Ru"/"' redox
couples in CH,Cl, (see Table 1).The El,, values are in the range 1.34-1.10 V vs.
Ag/AgCl, being reduced by increasing methyl substitution in the ring, as ob-
served also for tethered arene-rhodium and non-tethered arene-ruthenium sys-
t e m ~ . ~They
~ . ' are
~ also reduced by replacement of PPh, by the more electron-
donating PMe,. The potentials appear to be consistently somewhat greater (ca.
200 mV) than the quasi-reversible potentials of 0.89-1.09 V (us. SCE) reported for
[RuCl,(q6-arene)(L)] complexes in a~etonitrile,'~ indicating that the presence of
the strap may slightly stabilise ruthenium(@ relative to ruthenium(II1). Since the
potentials of the tethered complexes are clearly too high to allow isolation of the
presumed ruthenium(II1)species, we are currently attempting to prepare strapped
dimethylruthenium(I1) derivatives, which are expected to oxidise at a lower
potential. For example, treatment of complex 4 with dimethylzinc gives
[RuMe,(rjl :q6-Ph,PCH,SiMe,C6H,)], identified tentatively on the basis of
NMR data [G(Ru-CH,) - 0.32; G(Ru-CH,) - 4.75 (,JPc = 17 Hz)]. We also
plan to generate strapped areneruthenium(0) complexes in order to compare
their C-H activation properties with those of their unstrapped counterparts.
3 Acknowledgements
This work has been supported by an Australian Postgraduate Award to Joanne
Harper. We also thank Professor Brian F. G. Johnson, FRS, University of
Cambridge for the provision of laboratory facilities and for helpful discussion.
4 References
1 R. S. Nyholm, J . Chem. Soc., 1950,843,848,851,857, and subsequent papers.
2 G. J. Leigh, Interview with Joseph Chatt in Coord. Chem. Rev., 1991,108,4.
4 A. F. Masters, Appl. Organomet. Chem., 1990,4,389.
5 H. W. Kouwenhoven, J. Lewis and R. S. Nyholm, Proc. Chem. SOC.,1961,220.
6 M. A. Bennett, H. W. Kouwenhoven, J. Lewis and R. S. Nyholm, J . Chem. Soc., 1964,
4570.
7 M. A. Bennett, J. Chatt, G. J. Erskine, J. Lewis, R. F. Long and R. S. Nyholm, J . Chem.
Soc. A , 1967,501.
8 J. Okuda, Comments Inorg. Chem., 1994,16,185.
9 U. Siemeling, Chem. Rev., 2000,100, 1495.
10 H. Butenschon, Chem. Rev., 2000,100,1527.
11 E. T. Singewald, C. A. Mirkin, A. D. Levy, and C. L. Stern, Angew. Chem., Int. Ed.
Engl., 1994,33,2473.
12 E. T. Singewald, X. Shi, C. A. Mirkin, S. J. Schofer and C. L. Stern, Organometallics,
1996,15,3062.
13 E. T. Singewald, C. S. Slone, C. L. Stern, C. A. Mirkin, G. P. A. Yap, L. M.
Liable-Sands and A. L. Rheingold, J . Am. Chem. Soc., 1997,119,3048.
14 D. Devanne and P. H. Dixneuf, J . Organomet. Chem., 1990,390,371.
15 A. Ceccanti, P. Diversi, G. Ingrosso, F. Laschi, A. Lucherini, S. Magagna and P.
Zanello, J . Organomet. Chem., 1996,526,251.
16 M. A. Bennett, G. A. Heath, D. C. R. Hockless, I. Kovacik and A. C. Willis, J . Am.
Chem. SOC.,1998,120,932; Organometallics, 1998,17, 5867.
17 I. Kovacik, PhD Thesis, Australian National University, 1995.
18 P. Pertici, P. Salvadori, A. Biasci, G. Vitulli, M. A. Bennett and L. A. P. Kane-Maguire,
J . Chem. Soc., Dalton Trans., 1988,315.
19 C. A. L. Mahaffy and P. L. Pauson, Inorg. Synth., 1979,19,154.
20 R. Davis and L. A. P. Kane-Maguire. Comprehensive Organometallic Chemistry, ed. G.
Wilkinson, F. G. A. Stone and E. W. Abel, Pergamon, Oxford, 1982, Vol. 3, p. 1001.
21 P. D. Smith and A. H. Wright, J . Organomet. Chem., 1998,559,141.
22 A. Fiirstner, M. Liebl, C. W. Lehmann, M. Picquet, R. Kunz, C. Bruneau, D. Touchard
and P. H. Dixneuf, Chem. Eur. J., 2000,6,1847.
23 D. Jan, L. Delaude, F. Simal, A. Demonceau and A. F. Noels, J . Organomet. Chem.,
2000,606, 55.
24 B. L. Shaw, J . Am. Chem. SOC.,1975,97,3856; J . Organomet. Chem., 1980,200,307.
25 B. Therrien, T. R. Ward, M. Pilkington, C. Hoffmann, F. Gilardoni and J. Weber,
26 B. Therrien, A. Konig and T. R. Ward, Organometallics, 1999,18, 1565.
27 B. Therrien and T. R. Ward, Angew. Chem., Int. Ed. Engl., 1999,38,405.
28 A. Abele, R. Wursche, M. Klinga and B. Rieger, J . Mol. Catal. A , 2000,160,23.
29 K. Y. Ghebreyessus and J. H. Nelson, Organometallics, 2000,19,3387.
30 M. A. Bennett, G. B. Robertson and A. K. Smith, J . Organomet. Chem., 1972,43, C41.
SECTION E:
Chemistry Related to Dinitrogen

Complexes
Thefirst complex of dinitrogen was announced in 1965, and despite the speculation
that dinitrogen complexes analogous to the well known carbonyl complexes should
exist, none had been hitherto forthcoming. Bert Allen announced the$rst, and that
took some courage, since his evidence was primarily spectroscopic. Like
many seminal discoveries, this was by chance (he was attempting to prepare
[ R u ( N H 3 ) J 2 + )but it is only the prepared and enquiring mind that is open to
appreciate the signijicance of unexpected observations.
After that dinitrogen complexes came thick and fast, and the Chatt group, with
their sure foundation in phosphine chemistry, were ideally placed to exploit them.
This is explained in the initial contribution from Richards. Inevitably the impetus of
the Chatt groupfaded with the years and the group itselfsplit up. Others took up the
torch, and the beautiful work of Fryzuk, and the magnijicent achievements of
Floriani are reviewed here. Their work should be regarded as building on the
foundations laid at the Unit of Nitrogen Fixation in Sussex.
Further developmentsfrom the same basis are detailed by Hidai and by Dilworth.
Hidai was involved in the synthesis of one of the first dinitrogen complexes, and he
spent some time working in Sussex. His account shows how many discoveries,
initially quite unforeseeable, can arise from the same set of observations when they
are analysed by the minds of diflerent workers. The same is true of Dilworth, who
obtained his D. Phil. in the Unit and who discovered there one of theJirst large
groups of homologous dinitrogen complexes. Once he left the Unit, he exploited the
ideas and principles he had absorbed there in his own individual way. His contribu-
tion is a summary of work that is yet another fruitful tree springing from the same
seeds.
The reviews in this section provide a succinct summary of the Unit’s dinitrogen
chemistry, of some of the dinitrogen chemistry that sprang from it, and of the
developments it helped to set in train. It represents a commendable legacy of the
Unit of Nitrogen Fixation. The biological work of the Unit of Nitrogen Fixation
and some of its influences are described in Section F.
Chemistry at the Unit of
Nitrogen Fixation
RAYMOND L. RICHARDS
School of Chemistry, Physics and Environmental Science, University of Sussex,
Brighton, Sussex BN1 9QJ, UK
1 Introduction; The Early Years and the Search for

Dinitrogen Complexes
I joined the Unit of Nitrogen Fixation (UNF) in 1964,actually taking up my post
at the University of Sussex before the embryo UNF, then based in London at
Queen Mary College (Chemistry) and the Royal Veterinary College (Biology),
moved to Sussex. I joined directly from Manchester as a somewhat naive holder
of a fresh PhD. As I left Manchester the then Head of Department, Jack Lewis,
told me that Joseph Chatt had very high standards and joining him was a bit like
jumping in at the deep end, but that I would benefit enormously from working
with him. This proved to be entirely accurate. Although my original plan was to
spend a short time at the UNF before moving on, I so much enjoyed the
atmosphere and working ethos in the U N F that I remained there. I was to spend
the next 16 years working with Joseph until his retirement in 1980 and this article
is an account of the chemistry that developed during that period, with some
indication of directions of progress after that date.
The U N F chemistry was set under way at Sussex by a small initial group
consisting of myself, Brian Heaton, Rosemary Paske and David Newman,
strengthened soon after by the arrival of Jeff Leigh and Mike Mingos. Joseph
took a very close interest in all our projects. He had an impressive grasp of what
we all were doing and a disconcertingly extensive memory for detail. He expected
rigour in our work and kept us on our toes by recalling exactly what we told him
in the previous discussion. We were immediately called to account if our dis-
cussions of progress lacked consistency! Nevertheless, Joseph’s enthusiasm for
chemistry was infectious and we all beavered away at our pet projects with
enjoyment as well as care.
172 Chemistry at the Unit of Nitrogen Fixation
A major cloud was on our horizon, however. Although we knew from the
emerging biochemistry that the action of nitrogenase must involve interaction of
N, with iron or molybdenum or both, we had no routes to complex compounds
that would establish this. In 1965, however, Allen in Toronto produced the first
example of such an interaction' between Ru and N, in the complex ion
[Ru(NH,),(N,)]~+.This was followed by the preparation^^.^ of [CoH(N,)-
(PPhJ3] and [IrCl(N,)(PPh,),] and although these discoveries were an encour-
agement in the sense that binding of dinitrogen in transition metal complexes
appeared to be a general phenomenon, this was cold comfort since we had not
been involved in making any of them!
The Ru compound was also reported (erroneously as it turned out) to give
ammonia rather easily on treatment with sodium borohydride and at this stage it
appeared that much of the chemistry required to understand the action of
nitrogenase had been done. Indeed, I recall that at that time Joseph was con-
sidering what alternative project we might undertake; one that he was very
interested in was C-H bond activation. He had in fact published a seminal,
pioneering paper on this topic, also involving ruthenium ~ h e m i s t r yI. ~
had read
and re-read this paper and had been greatly impressed by the meticulous work
and the techniques used in it, such as detailed use of isotope labelling, so that it
proved inspirational for me, as it did for others who subsequently worked in
C-H activation ~ h e m i s t r y . ~
However, what we decided to do was to broaden our strategy. As well as direct
attempts to prepare dinitrogen complexes, which was proving difficult, (the
above compounds had in any case been obtained by accident), we set out to
examine the reactivity of the known dinitrogen complexes, develop the chemistry
of metal complexes which might prove to be precursors of new dinitrogen
compounds and also to look at the chemistry of dinitrogen analogues such as
isocyanides, which might also be substrates of nitrogenase. This approach was
typical of Joseph Chatt's management style, which led to so much success, in the
following ways. First, the projects that were undertaken developed from group
and interdisciplinary discussions in which we all took part, second we were
encouraged to work around a topic following pathways that might have only an
intuitive connection to our major interest, the only restriction being that the
science should be of a good standard. This approach allowed us to follow
hunches, kept us from becoming stale and in the end led to new chemistry of
direct relevance to our remit of elucidating the function of nitrogenase.
For me particularly, the study of isocyanide complexes was a case in point. As
well as competition in the chemical aspects of work at the UNF, there was strong
competition in biochemical work. In particular, examination of alternative
substrates for dinitrogen was in vogue and this had led to the seminal discovery
by Mike Dilworth in Australia that nitrogenase would reduce acetylene to
ethylene.6 In an interdisciplnary effort which was to prove the hallmark of the
U N F approach, the UNF biologists, John Postgate and Michael Kelly, and I
demonstrated that MeNC was also a substrate and that the enzymatic reduction,
to give methane and methylamine, could be mimicked if the MeNC was first
coordinated to a metal.7 More of this type of study later.
Raymond L. Richards 173
The dinitrogen complexes that had been prepared at this point carried phos-
phine co-ligands and it seemed reasonable for us to develop phosphine chemistry
of the early transition metals, since this should lead to suitable sites to bind
dinitrogen at biologically relevant metals such as iron and molybdenum. This
approach also had an advantage for us in that we were building on the extensive
background in such chemistry that had been built up by Joseph and his col-
leagues in his earlier career, particularly at ICI in the The Frythe laboratory.
A large range of complexes was prepared; in particular, halide-phosphine
complexes in relatively high oxidation states provided the starting materials that
eventually yielded dinitrogen-binding sites upon reductive elimination of hali-
d e . * Two
~ ~ examples are shown in Equations (1) and (2), each system giving an
extended series of dinitrogen complexes.'O~''
(X = Cl or Br; R = alkyl or aryl)
(M = Mo or W; R = alkyl or aryl)
2 The Search for Reactions of Coordinated Dinitrogen

Other groups around the world were also producing dinitrogen complexes at a
great rate and one of my interests was to try to produce ammonia from them by
various means. Up to this point, as we reached the 1970s, it had turned out that
none were capable of doing this by the methods we and others had used.',
Nevertheless, we were convinced that bound dinitrogen should be reactive
towards attack by protons. Amongst other considerations, two pieces of chemis-
try stood out in supporting this view by demonstating that bound dinitrogen was
weakly basic. The first was the observation that Allen's original dinitrogen
complex, although not in fact able to produce ammonia, was able to bind a
second R U ( N H ~ ) ~ , +
group to give the binuclear complex [(NH3)5-
RuN,Ru(NH,)~]~+ [Equation (3)].13
[(NH,),RuN,Ru(NH,),]~" + H,O (3)

The second was that the complex truns-[ReCl(N,)(PMe,Ph),l, prepared by Jon
Dilworth during his D. Phil. work in the UNF,14 was able to produce a whole
range of adducts, such as shown in Equations (4) and (5).l59I6
trans-[ReCl(N,)(PMe,Ph),l + [CrCl,(thf),] -+
[( PMe, Ph),ClRe(N,)CrCl,(t hf),] + t hf (4)

2 trans-[ReCl(N,)(PMePh),]+ [MoCl,(PPh,),] -+
trans-[MoCl,{(N,)ReCl(PMePh),},] + 2PPh, (5)

Disappointingly, the rhenium complex also was not capable of producing
ammonia. Most generally, protonation of dinitrogen complexes led to evolution
of the dinitrogen as the gas, often with the production of metal hydrides, a not
unexpected result in view of the electron-richness of the metal sites to which
dinitrogen was bound. That we were eventually successful was due to a combina-
tion of factors, all important in research, as seasoned practitioners will know.
These were persistence, preparedness and serendipity.
Because the metals important in nitrogenase were known to be molybdenum
and iron, we had attempted to protonate some of the dinitrogen complexes
known at that time. For example, on treatment with acids, the ~ o m p l e x ' ~
trans- [FeH(N,)( PEt ,CH ,CH2PEt2),]( BPh,) evolved N, and the molybdenum
compound trans-[Mo(N,),(dppe),] (dppe = PPh,CH,CH,PPh,) first prepared
by Hidai,'* gave a hydride, trans-[MoH,Cl,(dppe),], on treatment with HCl,I9
and no appreciable amount of ammonia or hydrazine was observed. The last
result was particularly perplexing because we had shown that the dinitrogen in
the molybdenum complex is weakly basic, acting as a donor to AlEt,.,'
We appeared to have reached stalemate when Jeff Leigh and Graham Heath
decided to look at the reactions of organic chlorides with trans-[W(N,),(dppe),].
Instead of loss of N, with formation of W-C bonded species, they observed the
formation of nitrogen-carbon bonds, as shown in Equation (6). The resulting
acetyldiazenido-complex could be protonated reversibly to give a hydra-
zido(2 - ) species.21
trans-[W(N,),(dppe),] + CH,COCl -+
trans-[WCl(N,COCH,)(dppe),] + N, (6)
This led us immediately to re-examine our discouraging earlier work on the
lack of protonation of dinitrogen in trans-[Mo(N,),(dppe),]. We found that
serendipity had worked in reverse, in that we had been correct in our approach,
but had chosen the wrong metal to use with HCl as reagent and the wrong anion
to use with tran~-[Mo(N,)~(dppe)~] as reagent! Under the correct conditions,
HCl gave trans-[WCl(NNH,)(dppe),]Cl from trans-[W(N,),(dppe),] and
HBr gave trans-[MoBr(NNH,)(dppe),]Br from trans-[Mo(N,),(dppe),]
[Equation (7)]. 9
+ 2HX -+
trans-[M(N,),(dppe),]
trans-[MX(NNH,)(dppe),]X + N, (7)
(M = W, X = C1; M = Mo, X = Br)
Thus we had reactivity of coordinated dinitrogen to make both N-H and N-C
bonds, opening up the prospect of routes to ammonia, hydrazine and organo-
nitrogen componds using dinitrogen complexes as catalysts or catalyst precur-
sors. For convenience, I will outline the advances made in the two areas separate-
ly, although they moved forward together of course, progress in one aiding that
in the other.
3 Production of Ammonia and Search for a Catalytic

System
Despite our demonstration of the partial reduction of coordinated N,, outlined
above, achieving the goal of ammonia production proved frustratingly slow. A
lot of effort was expended in trying various ways of converting the NNH, ligand
in the above complexes into ammonia, but without success. In these experiments
we restricted ourselves to relatively mild conditions, bearing in mind that we
wished to have some semblance of biological conditions.22
At that time Alan Pearman was working with me as a D. Phil. student and I set
him the task of looking at the chemistry of the NNH, compounds. He examined
their deprotonation to give NNH complexes and the displacement of the three-
electron donor NNH ligand by the isoelectronic NO ligand to give nitrosyl
complexes, as illustrated in Equations (8) and (9).,,
trans-[WF(NNH,)(dppe),](BF,) + NEt, -+
trans-[WF(NNH)(dppe),] + (NEt,H)(BF,) (8)
trans-[WF(NNH)(dppe),] + NO 4 trans-[WF(NO)(dppe),] + ‘HNO’ (9)

He also rang the changes on the counter ions that could be used in the system,
the aim being no more elevated than producing a good quantity of data for his
thesis. This entailed both metathetical change of counter ion [Equation (lo)] and
use of various acids, including H,SO,.,,
trans-[MX(NNH,)(dppe),]X + NaY +
trans-[MX(NNH,)(dppe),]Y + NaX (10)
(M = Mo or W; X = C1 or Br; Y = BPh,, ClO, or PF,)
Having exhausted our enthusiasm for the dppe system, we decided to move on
to monophosphine complexes, expecting to observe similar behaviour, perhaps
with more lability in the products. It was our habit to run these reactions in
sealed flasks and measure the gas evolved to give a mass balance.
Memorably, on a Friday afternoon in October 1974, we added H,SO, to
cis-[W(N,),(PMe,Ph),] in methanol and observed a rapid reaction to evolve a
gas and deposit a blue solid. We were immediately very excited by this because it
signified a different reaction pathway from those that we had seen before, and we
set about measuring the amount of gas produced. With typical intuition Joseph
came into the lab., to see how we were getting on, just as we found that only one
mole of gas had been produced (we assumed correctly that it was N,). We were
all convinced that the remaining N, had been lost from the metal as ammonia or
hydrazine, but we had to bite our nails until the following week, when the
completed analysis of the system showed that we had indeed produced two moles
of ammonia per tungsten atom [Equation (1l)]. We quickly showed that the Mo
analogue also produced ammonia under the same conditions but with only
about 0.7 moles of ammonia per Mo. Thus this success was a combination of
serendipity in the matter of choosing the correct solvent and acid and sheer
-
dogged persistence!
MeOH
cis-[W(N,),(PR,),] N, + 2NH3 + Wvlproducts
H P 4
+ 4(PHR3)(HS0,)
We confirmed this conversion by preparing C~~-[M(~~N,),(PR,),] and show-
ing that it gave "NH,; then we published our preliminary results.24 At around
the same time other workers showed that ammonia could be obtained from the
dppe system if rather more drastic conditions were used than the mild ones that
we had employed.2sIn addition, it was also demonstrated around this time also
that hydrazine and some ammonia could be obtained from complexes having
dinitrogen bridging two metals (e.g. N2H4 from [{ZrCp*,(N,)) 2(N2)])26
(Cp*=C,Me,) so the way seemed clear to using dinitrogen complexes in a
catalytic, low-temperature system to produce ammonia and hydrazine on a
commercial scale.
We now set out with high optimism first to characterise our system as far as
possible, and then to use this knowledge to design a commercial catalyst, not
realising at the time that attainment of a catalytic system that would work even
under laboratory conditions was to prove elusive for some ten years.
Characterisation of the ammonia-producing reactions proved to be complex,
and Richard Henderson in particular spent many hours in detailed kinetic
studies of the various systems that we had studied. He demonstrated the different
pathways that could be adopted, to give either hydrides, with or without dinitro-
gen loss, or to give hydrazides or ammonia, depending on the type of complex,
the protic acid and the solvent used.27Nevertheless, the overall picture was for
Joseph gratifyingly close to that which he had proposed for dinitrogen reduction
at a single metal some years before.', We were able, by judicious choice of
reagent and solvent, to isolate and fully characterise examples of the intermedi-
ate species MNNH, MNNH,, MNNH,, MN, MNH and MNH, from the
ammonia-producing reactions, and also to observe their interconversion and
degradation in reaction solutions by ',N NMR spectroscopy and other tech-
n i q u e ~ . , ~ -This
~ ' led to the proposal of the cycle of reduction of N, to ammonia
at a single metal site, shown in Scheme 1, which has been dubbed the 'Chatt
cycle',,, and the suggestion that it could be operative in nitrogenase if a single
metal site is involved, which is still unclear.33
Using this chemistry as the basis, Chris Pickett and Jean Talarmin were later
able to develop later an electrochemically-driven cycle for the reduction of N, to
NH3,34thus establishing in principle the vision that Joseph had held from early
days in the UNF, that a low-technology electrochemical process for generating
ammonia for agriculture could be achieved.
Raymond L. Richards
H+
MoN, y MoNNH y MoNNH,
e
H+
e
-
H+
MoNNH,'?
e
MoN + NH3
177
H+ H+ H+ N
MoN MoNH MoNH, Mo + NH, --% MoN,
e e e
Scheme 1
4 The Formation of Nitrogen-Carbon Bonds

As pointed out above, the first formation of a nitrogen-carbon bond from a
dinitrogen complex was achieved in 1972, [Equation (6)] and provided the
stimulus for the protonation studies which lead to NH, formation. The work on
N-C bond formation was not strictly in the remit of the UNF, whose members
were supposed to devote their working hours to the study of nitrogenase, so this
work was carried out by colleagues, directed by Joseph and Jeff Leigh, in an
associated laboratory in the University of Sussex, the funding coming from
sources outside the primary funding body of the UNF. Nevertheless, both
groups kept a close interest in each other's work, scientific discussion was free
and uninhibited and mutual benefit was enjoyed by all.
Initially it was thought that the mechanism of both protonation and alkyla-
tion would be the same, involving electrophilic attack by a proton or a carbo-
cation such as RCO'. It soon became clear, however, that the mechanism of
N-C forming reactions was much more complicated. Moreover, several types of
reaction were discovered.
Alkyl bromides and iodides were found to react with the trans-
[M(N,),(dppe),] complexes to give diazenido-compounds; these could proto-
nate at the terminal nitrogen to give hydrazido(2 -)-complexes which reverted to
the diazenido parents on treatment with base [Equation (12)].35
HX
trans- [MX(N,R)(dppe),] s trans- [MX(NNHR)(dppe),]X (12)
Base
(M = M o o r W,X = Br or])
The range of metals was also extended to include rhenium in the preparation
of [ReC1,(N2COR)(PMe,Ph),1 from [ReC1(N,)(PMe,Ph),],36 but other dinit-
rogen complexes proved resistant.
In the above reactions and related ones, such as the formation of
[MoCl(N,COPh)(dppe),] from [Mo(PhCN)(N,)(dppe),], the halide of the
attacking reagent was included in the complex product. However, the reaction
of [Mo(SCN)(N,)(dppe),] - with Bu"1 produced [Mo(SCN)(N,Bu")(dppe),],
which suggested that the mechanism of the reaction must differ in detail from
those of the previously described The studies which led to a detailed
understanding of the mechanism of N-C bond forming reactions are described
below, but first the discovery of further types of complex will be described.
One type was essentially found by accident. When the complexes trans-
[M(N,),(dppe),] were treated with MeBr in thf rather than benzene (the previ-
ous solvent of choice) instead of methyldiazenido-compounds, another type of
product was isolated. After some puzzlement, X-ray analysis proved the reaction
products to be diazobutanol complexes, [MBr(N,CH(CH,),OH}(dppe),l+- and
other solvents such as tetrahydropyran and tetrahydrothiophene gave analog-
ous materials.38
This reaction therefore gave a route to diazoalkane complexes by reaction of
ligating dinitrogen and its discovery was particularly gratifying at the time,
because attempts to produce such complexes directly by interaction of a diazo-
alkane with a metal complex generally caused decomposition of the diazo-
alkane.39
Once the diazoalkane products had been recognised, new routes were found to
them. One involved a gem-dibromide as the reagent as shown in Equation ( 13).,’
trans-[M(N,),(dppe),] + RR’CBr, -+[MBr(N,CRR’)(dppe),]Br + N, (13)
(M = Mo or W)
These complexes did not react with acids, but rather with nucleophilic re-
agents such as LiMe, to give diazenido-complexes [Equation (14)].
[MBr(N,CRR’)(dppe),]Br + LiMe -+ [MBr(N,CRR’Me)(dppe),] + LiBr (14)
A more general reaction gave examples of diazoalkane complexes which could
not be obtained from gem-dibromides and this involved the acid-catalysed
condensation of hydrazido(2 -) complexes with aldehydes or ketones.41 The
hydrazido(2 -) complexes were obtained from protonation of dinitrogen com-
plexes as described above. The condensation reaction was originally discovered
by Masanobu Hidai, who had been involved in the discovery of the first molyb-
denum dinitrogen complex trans-[Mo(N,),(dppe),], and in 1977 he was able to
spend a year working at Sussex, where he extended the range of diazoalkane
complexes by making use of the array of hydrazide complexes that we had
In particular, the use of the monophosphine precursors
[MX,(NNH,)(PMe,Ph),] (X = halide, M = Mo or W) allowed access to the
complexes [MX,(NNCMe,)(PMe,Ph)3],43 a useful way of accessing N-C bonds
for the monophosphine series, since reaction of the complexes cis-
[M(N,),(PMe,Ph),] with alkyl halides generally leads to loss of all dinitrogen as
the gas. The diazoalkane complexes also react with protic acids to give organonit-
rogen compounds, together with hydrazine, as shown for example in (15).37
[WBr,(N,CMe2)(PMe,Ph),1 + HBr -+
[WBr,(PMe,Ph),] + [PMe,PhH]Br + N,H, + Me,CNNCMe, (15)
We conclude this survey of the types of reaction that were discovered by
mentioning the variety of products obtained from reaction of a,w-dibromides,
Br(CH,),Br, with trans-[M(N,),(dppe),], which depended on the value of n. As
we have already noted, for n = 1, a diazoalkane complex was formed for
M = Mo. However, for M = W the complex [{WBr(dppe),},(p-N2CH2N2)]
was also isolated. In general, more than one type of complex was always formed,
the long-chain diazenido-complexes evidently acting as alkyl bromides and
attacking further dinitrogen c ~ m p l e x e s . ~ ~ * ~ ’
This wide variety of reactions naturally raised questions as to the mechanism
of these reactions and detailed studies established that a wide range of substitu-
tion, exchange and alkylation reactions all proceeded at the same rate, and that,
under pseudo first-order conditions, the reactions are first-order in the molyb-
denum complex. It also turned out that tungsten compounds showed similar
behaviour, although irradiation is also necessary. Thus a common rate-controll-
ing step is implied for these reactions, which can only be loss of dinitrogen as in
(16).44
(M = Mo or W)
In substitution and exchange reactions there is competition between the new
ligands and N, for the vacant site. Thus in all the alkylation reactions, the alkyl
halide also competes with N,, forming an unstable intermediate [M(N,)-
(XR)(dppe),], which homolyses to generate an Mi species, [MX(N,)(dppe),], and
a radical R', which for a dibromide can be a bromoalkyl radical
Br(CH,),-,CH,'. If the radical is reasonably stable and the solvent is inert, then
the radical appears to stay within the solvent cage containing both it and the MI
species, until it reacts with the remaining N, ligand as in (17).
If the radical is too unstable (e.g. BrCH,CH,') it will decompose before N-C
bond formation. Alternatively, a radical which is too stable will not attack N,. It
appears that alkylation and acylation of coordinated dinitrogen have similar
mechanism^.^^
This mechanism is satisfying but not universal, as has been hinted at above.
For example an alternative initiation step was discovered for the reaction of
[M(SCN)(N,)(dppe),] - with Bu"1. The reaction was first-order in each reactant
and the product, [M(SCN)(N,Bu")(dppe),]- did not include the halogen from
the alkyl halide. The mechanism shown in (18) was favoured on the grounds that
the redox potential of [M(SCN)(N,)(dppe),] - is about a volt more negative than
that of [M(N2),(dppe),], so the anion is more likely to transfer an electron to the
halide. Moreover, the electron transfer step would generate an MI species plus a
radical, which is exactly what results from metal-assisted homolysis of the alkyl
halide.
CM(SCN)(N,)(dPPe),l- + Bu"I + CM(SCN)(N,)(dPPe),l

+ C,H,' + I - --+ products (18)
It is clear from the products isolated from reactions in thf and related solvents
discussed earlier, that as well as the alkylation reactions being a function of metal
complex and halide, there is also a dependence upon solvent. If the solvent
contains hydrogen atoms which are easily removed by radicals then attack on
solvent might occur and the radical thereby generated can attack the MI species
thus yielding a diazenido-complex. This is the case, for example with thf as
shown in (19), followed by a reversible reaction with protic acids to give a
diazoalkane complex [Equation (20)].
5 Formation of Amines
As might be expected from the parallel work on N, protonation, a major target
of the N-C formation work was to develop a system to produce amines by a
catalytic process involving an organic feedstock and N,, via a dinitrogen com-
plex intermediate. This concept was certainly established in principle by the
above work, in that ligating dinitrogen had been transformed into an organic
fragment at the metal. There still remained much to do if a catalytic system was
to be viable, not least of which was the necessity to show that the metal-bound
species could be removed from the metal, preferably by a mild hydrogenation
step, so that the metal remained bound in a stable compound.
Although the removal of organic fragments was achieved, this was only at the
expense of the metal complex, which was generally completely degraded because
of the rather vigorous conditions that were necessary. Thus [MoBr(N,Bu")-
(dppe),] gave NH, and Bu"NH, on treatment with Na[BH,] in a sealed tube at
100°C but no metal complex was isolated.45 Hydrazido(2 -) complexes were
converted into anilines by various destructive methods, including treatment with
sulfuric acid in propylene carbonate, reaction with Li[AlH,] and distillation
from strong base;46diazoalkane complexes also gave amines on treatment with
Li[AlH4].43
The potential for a cyclic system using electroreductive conditions was demon-
strated by controlled potential electrolysis of [ M o B r { N N m H,)(dppe),]
Br, in thf under N,, which gave N-aminopyrrolidine and the parent dinitrogen
complex trans-[ Mo(N,),(dppe),] .47
Work continued albeit at a lower pace after the above discoveries. For
example the formation of N-Si bonds was established in the laboratory of Hidai
and in this case a catalytic system was developed for formation of trimethylsilyl
species from Me,SiCl, using trans-[Mo(N,),(dppe),] and other dinitrogen com-
plexes as catalysts
[M(N2)2(dppe)21
N2+ Na + SiMeCI * N(SiM%)3+ HN(SiM%), + Me3SiSiMe, (21)
thf
6 Reaction of Alternative Substrates

As I mentioned at the beginning of this article, from the early days at the UNF,
the chemists were aware of the versatility of nitrogenase, in that a number of
substrates other than dinitrogen, such as CH,NC (to CH, and NH,),7 cyanide
ion (to CH, and NH,),' and C2H2(to C2H4), could be reduced by this enzyme.6
These reactions and the observation that carbon monoxide inhibits the reduc-
tion of all substrates except the proton, established the organometallic nature of
the enzyme. In the absence of any other substrate, protons are reduced to H,;
there is a corresponding reduction in the molecular yield of H, in the presence of
substrate. The detailed biochemistry of these processes and their inter-relation-
ships have been discussed in terms of an elegant kinetic study by the U N F
biochemists Lowe and Thorneley and it is not my purpose to give an account of
that work since it is well presented el~ewhere,~' but to emphasise the stimulus
that the interaction with biochemists gave us to study the metal chemistry of
dinitrogen and its alternative substrates. I will concentrate on only two aspects,
since they illustrate the general approach, these being the reactivity of coordinat-
ing isocyanides and the behaviour of these ligands and of alkynes at dinitrogen-
binding site^.^'^^^
Our first entry into the study of alternative substrates was the observation that
MeNC was reduced by nitr~genase.~ At that time there was still debate over
whether or not a metal was involved in the enzyme action and it was a small
triumph to be able to demonstrate that MeNC would only undergo reductive
cleavage reactions to CH, and MeNH, after first binding a metal, otherwise the
product of reduction was Me,NH.7 The work also lead us into some interesting
organometallic chemistry. One of the metals we had used in our efforts to mimic
nitrogenase action was platinum, and although its use was more inspired by
Joseph's background in organometallic platinum chemistry than any biological
consideration, we found that reductive cleavage of MeNC with reducing agents
such as Na[BH,] did occur in platinum complexes. We never attempted to work
out the detail of these reactions, but in some preliminary experiments our student
Lisa Badley found that the ligating MeNC was succeptible to attack by reagents
such as alcohols and amines to give carbene (alkylidene) complexes as shown in
Equation (22), and this reaction proved to be general for a wide range of
metals.' 5 4
9
+ HR3 --+ c ~ s - [ P ~ C ~ ~ { C ( N H R ' ) R ~ }(22)

c~s-[P~CI~(R'NC)(PR~J] (PR~~)]
(R', R2, and R5 = alkyl or aryl; R3 = OR4 or NHR', R4 = alkyl)
However, this work pre-dated the discovery of protonation of dinitrogen in

the molybdenum and tungsten complexes and, once this had come to light, we
turned our attention to the behaviour of isocyanides and alkynes at those sites
that activated dinitrogen to reduction. For isocyanides, the first step was to bind
these ligands in place of dinitrogen in the complexes trans-[M(N,),(dppe),],
which was done by a displacement reaction [Equation (23)l."
trans-[IM(N,),(dppe)~] + 2RNC + + 2N, (23)

tr~ns-[M(RNC),(dppe)~]
(R = alkyl or aryl)
Immediately evident, from the pronounced lowering of the N-C stretching

frequency of the isocyanides on binding, was the extent to which these sites are
electron-releasing. This is of course the basis of the activation of N, at these sites
towards protic attack at the position p to the metal, and together with Armando
Pombeiro, then a graduate student, we were able to demonstrate a very similar
pattern of protic attack on RNC at these sites, to give carbyne (alkylidyne)
c o m p l e x e ~ ,as
~ ~shown
. ~ ~ in Scheme 2; a similar reaction also occurs in trans-
[ReC1(RNC)(PMe,Ph)4].52357 Under appropriate conditions coupling of car-
byne fragments can also occur to give alkyne complexes (Scheme 2).58Under
similar conditions to those used to produce NH, from N, in cis-
[M(N,),(PMe,Ph),], RNC gives RNH,, NH, and low yields of hydrocarbons
from complexes of the type [M(RNC),(PMe,Ph)4].51
HX HX
trans-[M(RNC),(dppe),] --+ [M(dNHR)(RNC)(dppe),]X --+
[M(=CNHR),(dppe),]X, + [MX(q2-MeHNC=CNHMe)(dppe),]X
Scheme 2
The reactions of alkynes at these nitrogen-binding sites are complicated since
a variety of pathways can be followed, including polymerisation, and although
the work I shall now briefly describe was initiated while the UNF chemistry that
concerns us here was being developed, the range of chemistry which flowed from
it took many years to mature. Suffice to say that Scheme 3 shows one aspect of
the chemistry of alkynes which relates closely to that shown by N, and RNC at
these sites; the field has been reviewed.59
Scheme 3
The outcome of these studies was a reasonably coherent picture of reduction of
dinitrogen and alternative substrates at a metal site, which appeared to relate at
least in principle to nitrogenase action.
7 Further Developments
Here it is my purpose simply to point out the directions that ‘UNF chemistry’
took following the period covered by this review. Two major discoveries had
great influence. The first was the demonstration, by genetic techniques, that
alternative nitrogenases existed in which molybdenum is replaced by vanadium
or iron.60,61This led to renewed interest in producing ammonia from iron
dinitrogen complexes62 and the development of dinitrogen complexes of
Much of the work involved phosphine co-ligands and could be
seen as a further extension of the work begun at the UNF. However, because the
metals in nitrogenase have a sulfur-ligand environment, efforts have been made
to prepare complexes of dinitrogen at metal centres where the co-ligands have
sulfur-donor atoms. This has proved to be a difficult task, although the number
of sulfur-ligated dinitrogen complexes is increasing, notable examples being
[Re(SC,H,Pri,-2,4,6),(N,)(PPh,)l”s and the most successful system in terms
of binding N,, the macrocyclic thioether complex truns-[Mo(N,),(Me,[ 161
aneS,)] Its dinitrogen ligands can be reduced, but the yields of ammonia are
low and the susceptibility of the thioether ligand to degradation make an
electrochemically-driven cycle for ammonia based on this system unlikely at
present. Further developments in this area have involved the use of chelating
thiolate ligands and are discussed in the article by Roger Sanders elsewhere in
this book (p. 252).
The second major advance was the determination of the definitive X-ray
crystal structure of nitrogenase, which showed the active centre of molybdenum
nitrogenase to consist of an Mo-Fe-S cluster (FeMoco), which has a unique
extended FeS-bridged structure and appears to be the centre at which the
reduction of N, and related molecules occurs. The central Fe atoms in FeMoco
are coordinated by three sulfides in an approximate planar environment and the
molybdenum atom is also coordinated to three sulfurs, as well as two oxygens
from homocitrate and one histidine n i t r ~ g e n . “It~is thought that the vanadium
atom in vanadium nitrogenase and one iron atom in the ‘iron-only’ nitrogenase
are in similar environments. This knowledge increased further the interest of
chemists in iron-sulfur clusters, whose chemistry has been outstandingly devel-
oped by Holm, Coucouvanis and other^.^^>^^ The U N F and its later manifesta-
tions, by virtue of their multidisciplinary environment, were particularly well
equipped to investigate the chemistry of FeMoco itself and the efforts of Barry
Smith and his colleagues have led to detailed study of isolated FeMoco and
demonstration of its ability to function outside the enzyme, to interact with CO,
and to reduce protons.70971
8 Conclusions
Joseph Chatt founded a laboratory with an environment that was multidisci-
plinary and flexible. It was an ideal place for a chemist such as myself to learn his
trade and contribute towards the solution of a major problem, the function of
nitrogenase. Certainly the work of the U N F chemists during the period that I
have covered played a leading role in developing the area of dinitrogen chemis-
try. Whether that chemistry is indeed that chosen by nature remains to be
established, as do catalytic processes based on it. I thoroughly enjoyed my time
as a U N F chemist attempting to find such processes and I wish all success to
current and future workers who attempt to complete the search. I also wish to
thank all my colleagues during my time at the U N F for their help and friendship.
9 References
1 A. D. Allen and G. V. Senoff, Chem. Commun., 1965,621; A. D. Allen, F. Bottomley, R.
0.Harris, V. P. Reinsalu and C. V. Senoff, J . Am. Chem. Soc., 1967,89,555.
2 A. Yamamoto, S. Kitazume, L. S. Pu and S. Ikeda, J . Chem. Soc., Chem. Commun.,
1967,79; A. Sacco and M. Rossi, Chem. Commun., 1967,316.
3 J. P. Collman, M. Kubota, F. D. Vastine, J. Y. Sun and J. W. Kang, J . Am. Chem. Soc.,
1968,90,5430.
4 J. Chatt and J. M. Davidson, J. Chem. Soc., 1965,843.
5 R. H. Crabtree, Chem. Rev., 1985,85,245.
6 M. J. Dilworth, Biochim. Biophys. Acta, 1966,127, 285.
7 M. Kelly, J. R. Postgate and R. L. Richards, Biochem. J., 1967,102, lc.
8 J. Chatt, G. J. Leigh, D. M. P. Mingos and R. J. Paske, J . Chem. Soc. (A),1968,2636.
9 A. V. Butcher, J. Chatt, G. J. Leigh and P. L. Richards, J . Chem. Soc. Dalton Trans.,
1972,1064.
10 J. Chatt, G. J. Leigh and R. L. Richards, Chem. Commun., 1969,515.
11 B. Bell, J. Chatt and G. J. Leigh, Chem. Commun., 1970, 842.
12 J. Chatt and R. L. Richards, in The Chemistry and Biochemistry of Nitrogen Fixation,
ed. J. R. Postgate, Plenum Press, London, 1971, p. 19.
13 D. F. Harrison, E. Weissberger and H. Taube, Science, 1968,159,320.
14 J. Chatt, J. R. Dilworth and G. J. Leigh, J . Chem. Soc., Chem. Commun., 1969,687.
15 J. Chatt, R. C. Fay and R. L. Richards, J. Chem. Soc. ( A ) ,1971,702.
16 P. D. Cradwick, J. Chatt, R. H. Crabtree and R. L. Richards, J . Chem. Soc., Chem.
Commun., 1975,351.
17 G. M. Bancroft, M. J. Mays and B. E. Prater, J . Chem. Soc., Chem. Commun., 1969,589.
18 M. Hidai, K. Tominari, Y. Ushida and A. Misono, J . Chem. Soc., Chem. Commun.,
1969,1392.
19 J. Chatt, G. A. Heath and R. L. Richards, J . Chem. Soc., Chem. Commun., 1972,1010.
20 J. Chatt, R. H. Crabtree and R. L. Richards, J . Chem. Soc., Chem. Commun., 1972,
534.
21 J. Chatt, G. A. Heath and G. J. Leigh, J . Chem. Soc., Chem. Commun., 1972,444.
22 J. Chatt, G. A. Heath and R. L. Richards, J . Chem. Soc., Dalton Trans., 1974,2079.
23 J. Chatt, A. J. Pearman and R. L. Richards, J . Chem. Soc., Dalton Trans., 1976,1520.
24 J. Chatt, A. J. Pearman and R. L. Richards, Nature, 1975,39.
25 C. R. BrQlet and E. E. Van Tamelen, J. Am. Chem. Soc., 1975,97,911.
26 J. M. Manriquez, R. D. Sanner, R. E. Marsh and J. E. Bercaw, J . Am. Chem. Soc., 1976,
98, 8351.
27 R. A. Henderson, G. J. Leigh and C. J. Pickett, Adv. Inorg. Chem. Radiochem., 1984,27,
197.
28 J. Chatt, J. R. Dilworth and R. L. Richards, Chem. Rev., 1978,78,29.
29 R. L. Richards, in New Trends in the Chemistry ofNitrogen Fixation, eds. J. Chatt, L.
M. da Camara Pina and R. L .Richards, Academic Press, London, 1980, p. 199.
30 A. Galindo, A. Hills, D. L. Hughes, R. L. Richards, M. Hughes and J. Mason, J . Chem.,
Soc. Dalton Trans., 1990,283.
31 S. N. Anderson, M. E. Fakely, R. L. Richards and J. Chatt, J . Chem. Soc., Dalton
Trans., 1981,1973.
32 C. J. Pickett, JBIC, 1996,1, 601.
33 R. L. Richards in Biological Nitrogen Fixation for the 21st Century, eds. C. Elmerich,
A. Kondorosi and W. E. Newton, Kluwer, 1998, p. 17.
34 C. J. Pickett, K. S. Ryder and J. Talarmin, J . Chem. Soc., Dalton Trans., 1986, 1453.
35 J. Chatt, A. A. Diamantis, G. A. Heath, N. E. Hooper and G. J. Leigh, J . Chem. Soc.,
36 J. Chatt, G. A. Heath, N. E. Hooper and G. J. Leigh, J . Organomet Chem., 1973,57,
C67.
37 G. J. Leigh, in New Trends in the Chemistry of Nitrogen Fixation, eds. J. Chatt, L. M.
da Camara Pina and R. L. Richards, Academic Press, London, 1980, p. 215.
38 P. C. Bevan, J. Chatt, R. A. Head, P. B. Hitchcock and G. J. Leigh, J . Chem. Soc., Chem.
Commun., 1976,509.
39 M. L. Ziegler, K. Wiedenhammer, H. Zeuner, P. S. Skell and W. A. Herrmann, Angew.
Chem. Int. Ed. Engl., 1976,15,695 and references therein.
40 R. Ben-Shoshan, J. Chatt, W. Hussain and G. J. Leigh, J . Organomet. Chem., 1976,112,
c9.
41 M. Hidai, Y. Mizobe and Y. Ushida, J . Am. Chem. Soc., 1976,98,7824; M. Hidai, Y.
Mizobe, M. Sato, T. Kodama and Y. Ushida, J . Am. Chem. SOC.,1978,100,5740.
42 J. Chatt, A. J. Pearman and R. L. Richards, J . Chem. Soc., Dalton Trans., 1978,4720.
43 P. C. Bevan, J. Chatt, M. Hidai and G. J. Leigh, J . Organomet. Chem., 1978, 160, 165.
44 J. Chatt, R. A. Head, G. J. Leigh and C. J. Pickett, J . Chem. Soc., Chem. Commun., 1977,
299; J . Chem. Soc., Dalton Trans., 1979, 1638.
45 D. C. Busby and T. A. George, J . Organomet. Chem., 1978,149, L 273.
46 P. C . Bevan, J. Chatt, G. J. Leigh and E. G. Leelamani, J . Organomet. Chem., 1977,139,
c59.
47 C. J. Pickett and G. J. Leigh, J . Chem. Soc., Chem. Commun., 1981,1033.
48 M. Hidai and Y. Mizobe, Chem. Rev., 1995,95,1115.
49 W. R. Hardy and E. Knight, Biochim. Biophys. Acta., 1967,139,69.
50 R. N. F. Thorneley and D. J. Lowe, in Molybdenum Enzymes, ed. T. G. Spiro, Wiley,
New York, p. 221.
51 A. J. L. Pombeiro, in New Trends in the Chemistry of Nitrogen Fixation, eds. J. Chatt,
L. M. da Camara Pina and R. L. Richards, Academic Press, London, 1980, p. 249.
52 A. J. L. Pombeiro and R. L. Richards, Coord. Chem. Rev., 1990,104,13.
53 E. M. Badley, J. Chatt and R. L. Richards, J . Chem. Soc ( A ) ,1971,21.
54 J. Chatt, R. L. Richards and G. H. D. Royston, J . Chem. Soc., Dalton Trans., 1973,
1433.
55 J. Chatt, C. E. Elson, A. J. L. Pombeiro, R. L. Richards and G. H. D. Royston, J . Chem.
56 J. Chatt, A. J. L. Pombeiro and R. L .Richards, J . Chem. Soc., Dalton Trans., 1980,492.
57 A. J. L. Pombeiro, C. J. Pickett and R. L. Richards, J . Organomet. Chem., 1982,224,
285.
58 Y. Wang, J. J. R. Frausto da Silva, A. J. L. Pombeiro, M. A. Pellinghelli, A. Tiripicchio,
R. A. Henderson and R. L. Richards, J . Chem. Soc., Dalton Trans., 1995, 1183.
59 R. L. Richards, in Biology and Biochemistry ofNitrogen Fixation, eds. M. J. Dilworth
and A. R. Glenn, Elsevier, Amsterdam, 1991, p. 58.
60 R. N. Pau in Biology and Biochemistry ofNitrogen Fixation, eds. M. J. Dilworth and A.
R. Glenn, Elsevier, Amsterdam, 1991, p. 37.
61 R. R. Eady, Adv. Inorg. Chem., 1991,36,77.
62 G. J. Leigh, Acc. Chem. Res., 1992,25, 177.
63 J. J. H. Edema, A. Meetsma and S. Gambarrota, J . Am. Chem. Soc., 1989,111,6878.
64 D. Rehder, C. Woitha, W. Preibsch and H. Gailus, J . Chem. SOC., Chem. Commun.,
1992,364.
65 J. R. Dilworth, J. Hu, R. Thompson and D. L. Hughes, J . Chem. Soc., Chem. Commun.,
1992,551.
66 T. Yoshida, T. Adachi, T. Ueda, M. Kaminaka, N. Sasaki, T. Higuchi, T. Aoshima, I.
Mega, Y. Mizobe and M. Hidai, Angew. Chem., int. Ed. Engl., 1989,28,1040.
67 D. C. Rees, M. K. Chan and J. Kim, Adu. Znorg. Chem., 1993,40,89.
68 R. E. Palermo, R. Singh, J. K. Basin and R. H. Holm, J. Am. Chem. SOC., 1994, 106,
2600.
69 K. D. Demadis, S. M. Malinek and D. Coucouvanis, Znorg. Chem., 1996,35,414038.
70 I. Harvey, R. W. Strange, R. Schneider, C. A. Gormal, C. D. Garner, S. S. Hasnain, R.
L. Richards and B. E. Smith, Znorg. Chim. Acta, 1998,275-276, 150.
71 T. Le Gall, S. K. Ibrahim, C. A. Gormal, B. E. Smith and C. J. Pickett, J . Chem. SOC.,
Dinitrogen Activation by Early
Transition Metal-Amido
Phosphine Complexes
MICHAEL D. FRYZUK
Department of Chemistry, University of British Columbia, 2036 Main Mall,
Vancouver, B. C. V6T 1Z1, Canada
1 Preamble
The following describes work on dinitrogen activation by Group 4 and 5
complexes carried out at the University of British Columbia over the last decade
or so in the Fryzuk group. While we are relative newcomers to this area, it is clear
that our work builds on a strong foundation masterfully constructed by the
Nitrogen Fixation Unit led by Professor Joseph Chatt while at Sussex Univer-
sity. There are also many other important contributors to the area of dinitrogen
activation during the last 35 years, for example, the Hidai group from Japan,
Schrock's group from MIT, and many others.' Before describing our contribu-
tions, I have included a short background section on how some of our work
started.
2 Background
As a graduate student with Professor B. Bosnich at the University of Toronto
during the period 1974-78, I had the pleasure of interacting with Professor A. D.
Allen. During a graduate course, he recounted the story of the discovery2 of the
first ever dinitrogen complex [Ru(NH3)JN2)I2+1 by his graduate student C. V.
Senoff. It was fascinating, mainly because it went to the heart of scientific
discovery through serendipity, curiosity, and i n ~ i g h t A
. ~ little later during my
graduate career, in 1977, I attended a joint Chemical Institute of Canada-
American Chemical Society conference in Montreal and there heard a fasci-
nating lecture on zirconium hydrides by a young assistant professor from Cal-
tech by the name of John B e r ~ a wAt
. ~ the heart of this work was the dinuclear
zirconium dinitrogen complex [( ( Z ~ ( N , ) C P * ~ ) ~ ( ~ 2- N(Cp*
~ ) ] = C,Me,) that
188 Dinitrogen Activation by Early Transition Metal-Amido Phosphine Complexes
served as both a versatile starting material for Zr" via displacement of coor-
dinated N, but also for Z P complexes by oxidative addition, for example, by
dihydr~gen.~?~
2+
N
Cp*2Zr-N=N- ifjrCp*,
N
H3
Ill
N
1 2 (Cp* = q5-CSMe,)
I pretty much decided after that lecture that I wanted to learn more about early
transition metal chemistry to augment my late transition metal background and
applied to Bercaw for a postdoctoral position. I should point out that my
graduate work was on the preparation of chiral diphosphine ligands (chiraphos
and prophos) and their use in the asymmetric hydrogenation of amino acid
precursors using rhodium(1) complexes. Bosnich was a fantastic supervisor who
taught me the basics of ligand design, but more importantly showed me that
chemistry was a craft.
Once I took up my independent career in 1979 at the University of British
Columbia, dinitrogen chemistry was not one of my initial goals despite those
early influences by pioneers in this area. I did have a fascination for early
transition metal derivatives that I had picked up from the Bercaw experience but
I also wanted to keep my fingers in the late transition metal area. To satisfy both
cravings, I designed and synthesised a mixed-donor multidentate ligand that we
subsequently called PNP.7 However, it wasn't until 1988 that dinitrogen com-
plexes were discovered in my laboratory. A general formula for a PNP complex
is shown below.
The design of the PNP amidodiphosphine ligand was based on the idea that
the combination of both hard and soft donors should allow access to a variety of
oxidation states for different transition metal complexes. For the late metals, the
soft phosphine donors would serve to stabilise the hard amido-ligand bond while
for the early transition elements, the amido donor would anchor the soft phos-
Michael D.Fryzuk 189
phine donors to these metals. In both cases, it was expected that the combination
of mixed donors would facilitate changes in oxidation states and even allow for
the preparation of uncommon oxidation levels for certain metals.
3 Our First Dinitrogen Complexes

When Tim Haddad joined my group in 1985 his task was to examine the
organometallic chemistry of Group 3 and Group 4 complexes stabilised by
the P N P ligand. On one project, he examined the reaction of [MCl,-
{N(SiMe,CH,PR,),}] (M = Zr or Hf; R = Me, Pr', or But) with 'magnesium
butadiene' (MgC,H,*thf) to generate the Group 4 butadiene complexes
[M(C,H,)Cl{N(SiMe,CH,PR,),)I (M = Zr or Hf). While these could be con-
sidered as ZrlI and Hf" species, clearly this was a formalism. Tim also showed
that these same complexes could be prepared by reductive techniques using
Na/Hg and 173-butadiene. Unbeknownst to me, Tim also investigated the
Na/Hg reduction of the zirconium(1v) precursors under dinitrogen. His first
positive result was from the reduction of [ZrCl, { N(SiMe,CH,PBu',),}] under
N, from which he generated a deep green complex which he declared to be
[[{(Bu',PCH,SiMe,),N}ZrCl]2(~-N2)]. Unfortunately, he was never able to
grow X-ray quality crystals of this material so its structure remains uncertain.
Attempts to reduce the related complex with the smaller methyl substituents at
phosphorus, [ZrCl,{ N(SiMe,CH,PMe,),)], failed completely; no colour
changes were observed and recovery of starting material was generally achieved.
He was pretty sure that he had a dinitrogen complex with the tertiary butyl
derivative, but I remained sceptical. My input was to suggest changing the
substituents at phosphorus to isopropyl, which I hoped would make things more
crystalline. Fortunately, that lucky guess paid off and we did obtain deep blue
single crystals of [[{(Pri,PCH,SiMe,),N)ZrCl],(p- q 2 : q2-N2)]3, Reaction (1).
Me,
2 Na/Hg, N,
D
toluene
(-2NaCI)
Me,
3 (dark blue)
Complex 3 was a real surprise. It contained a side-on bound dinitrogen

fragment bridging the two zirconium centres with a N-N bond distance of
1.548(7) A,the longest N-N bond distance in a dinitrogen complex. When we
communicated this work in 1990,8 it was only the second example of a side-on
bound dinitrogen complex structurally characterised, the other being the dinuc-
lear samarium derivative [{ SmCp*,},(p-y2: y2-N2)]4.9
What was intriguing about these two complexes was the very different N-N
bond lengths for the coordinated dinitrogen unit; in contrast to the very long
bond distance in our complex, the disamarium complex showed a rather short
N-N bond length of 1.088(12)A,virtually unchanged from that of free dinitrogen
at 1.0947(5)A.Another more obvious difference between our zirconium dinitro-
gen complex and the samarium derivative was that dinitrogen was irreversibly
bound to the two ZrIVcentres whereas N, was easily lost from [(SmCp*,),(p-
y2: y2-N2)].Given the long N-N bond length in 3, it seemed reasonable to assign
formal oxidation states such that the complex consisted of two ZrIVcentres and
the bridging dinitrogen unit was N,4-. Addition of an excess of acid released
exactly one equivalent of hydrazine, which supported this notion.* Also support-
ive of the presence of a highly reduced form of N, was the lack of reactivity with
H,; unlike the Bercaw dinitrogen complex 2 that reacts with H, to form
[ZrH,Cp*,] with release of N,, our side-on N, derivative 3 was impervious to
H2.
An obvious question that arose from this work was 'Why side-on bound
dinitrogen rather than end-on?' We were able to show that the ancillary ligands
play a major role in determining which orbitals on the metal centre are available
to bind with the dinitrogen n* orbitals; since the amide donor of the P N P ligand
could overlap with one of the nonbonding d orbitals, this effectively removed one
of the possible n-symmetry combinations found in the end-on mode of bonding.
As a result the side-on mode became more favourable." We were able to test this
simple model by the preparation of related complexes, exemplified by the aryl-
oxide complex [[Zr((Pri,PCH2SiMe,),N}(O-2,6-Me2C6H3)],(p-y2: y2-N,)] 5,
which was synthesised by M. Mylvaganam. This complex also showed a side-on
bound N, unit with a very long bond length of 1.528(7)A." In collaboration
with Thomas Loehr demonstrated that resonance Raman spectroscopy was
extremely effective in distinguishing between the end-on and side-on modes of
bonding for coordinated dinitrogen.'
Despite the fact that we had in our hands a strongly activated dinitrogen
fragment and a relatively unexploited bonding mode for N,, reactivity studies
were hampered by the fact that the P N P ligand system was prone to phosphorus
decoordination. Attempts to add HCl in a controlled fashion generally resulted
in complex mixtures from which no identifiable products could be obtained. In
an effort to circumvent this phosphine dissociation problem, we redesigned our
ancillary ligand.
4 The Macrocycle
If phosphorus decoordination were responsible for preventing clean, unique
reactions of our dinitrogen complexes, then a strategy to prevent this process was
necessary. One approach was to change from an acyclic tridentate ligand system
to a tetradentate macrocyclic ligand that would mimic the PNPX ligand frame-
work found for 3 (X = C1) and 5 (X = OAr). The rationale was straightforward:
incorporation of the phosphine donors into such an array would prevent disso-
ciation since the macrocycle is rigid and not able to decoordinate any donor
without the whole ligand being released. In 1995, a new postdoctoral fellow in
my group, Jason Love, decided that he wanted to try and prepare the P,N,
macrocycle. At this point we had never successfully prepared a macrocycle
containing phosphine donors and I reasoned that this might end up being a fairly
arduous task. I cautioned Jason that this might be difficult and suggested that he
use ‘high dilution’ conditions (what little I knew about macrocycles suggested
that this might be prudent). He ignored me.
Jason’s strategy was to prepare the P2N2 macrocycle by isolating the P N P
intermediate and then to close the ring using the starting material for PNP,
1,3-bis(chloromethyl)tetramethyldisilazane,HN(SiMe,CH,Cl),. His first at-
tempt was partially successful; the reaction of the intermediate
HN(SiMe,CH,PHPh), with one equivalent of HN(SiMe,CH,Cl), and two
equivalents of LiBu produced a small amount of crystalline material with a very
unusual 31P{1H) NMR spectrum. Instead of the expected singlet(s) Jason ob-
served 1: 1: 1: 1 quartet patterns. When he showed this spectrum to me, I immedi-
ately realised what had happened; instead of isolating the expected neutral
diamine macrocycle PhP(CH,SiMe,NHSiMe,CH,),PPh, he had clearly pre-
pared the dilithio derivatives which were showing 7Li coupling to the phos-
phorus nuclei. My suggestion to him was to repeat the reaction but add a total of
four equivalents of LiBu; this worked amazingly well (yields were SO-90%).
What was remarkable to me was that high dilution techniques were not necess-
ary for the success of this reaction (Scheme 1). Our hypothesis still is that this
reaction proceeds via lithium templating. There are some details that need to be
included here; the reaction is both temperature- and solvent-dependent. Two
forms of the macrocycle are generated because of the presence of the trigonal
pyramidal phosphine donors. In thf, lower temperatures favoured the formation
of the anti derivative whereas in Et,O only the syn derivative was observed. As it
turned out, the syn derivative was the most useful form for the preparation of
early transition metal derivatives and so the use of diethyl ether is the method of
choice in my group for the preparation of the P,N, macrocyclic system.
Jason then turned his attention to the coordination chemistry of syn-
Me2 Me2 Me2 Me2
2 LiPHPh, thf
0 "C
CI (>go%) Ph-p P-Ph
I
syn-Li2(thf)(P2N2) 6 anti-Li2(thf)2(P2N2)
Scheme 1
Li,(thf)(P,N,) 6 (P,N, = PhP(CH,SiMe,NSiMe,CH,),PPh). One of his first

reactions was with [ZrCl,(tht),] (tht = tetrahydrothiophene) to generate the
ZrIv-containingprecursor [ZrCl,(P,N,)] 7.The chemistry of this ZrIVderivative
is very extensive,14and it can also be reduced to generate the dinitrogen complex
[{(P2N2)Zr},(p-q2:y2-N,)] 8. We were able to confirm that the change in ligand
design had not dramatically affected the coordination mode of the N, unit as it
was bound in a side-on manner.15 However, the N-N bond length was only
1.43(1) A, considerably shorter than the other side-on bound dinitrogen com-
plexes that we had isolated thus far using the P N P ancillary ligand.
One of the first reactions we investigated was the reaction of 8 with dihydro-
gen. Now one might ask why this reaction should be pursued since our original
dinitrogen complex 3 was unreactive to H,. I guess the answer is that we thought
that perhaps the shorter N-N bond length found in 8 might imply a less
activated N, unit that might be displaced by H, like the Bercaw dinitrogen
complex 2. What in fact occurred was unexpected and certainly more remarkable
than displacement of coordinated N,. The reaction of H, with 8 resulted in the
cleavage of H, and the formation of a N-H bond and a bridging hydride
[Reaction (2)].
The evidence for this transformation was initially based on 'H NMR spectros-
copy. The formation of [{Zr(P2N2)},(p-q2:q2-N2H)(p-H)] 9 was clearly
evidenced by a loss of the deep blue-green colour of 8 to generate yellow-orange
9. The solution 'H NMR spectrum of 9 showed two new resonances at 5.53 ppm
and at 2.07 ppm due to the N-H and the Zr,(p-H) units, respectively. These
assignments were confirmed by 15N and ,H labelling. We also extended this
reaction to silanes; the addition of phenyl- or butyl-silane (RSiH,; R = Ph or Bu)
to 8 resulted in the formation of the silylated analogues. One of these (R = Bu)
was crystalline enough to confirm that the solid state structure was identical to
that found in solution [Reaction (3)].
RSiH .?
(3)
toluene
(Methyls on Si omitted)
8 10 R = P h o r B u
We were able to grow crystals of 9 and these were submitted to Victor G.

Young, Jr., at the University of Minnesota for X-ray analysis using CCD
detection and low-temperature data collection. The result was startling. Victor
informed us that the structure had refined to indicate a coordinated dihydrogen
unit bridging the two Zr centres and parallel to the side-on bound N, moiety.
What this suggested was that in the solid state a different structure was present
from that found in solution. In retrospect, we should have been much more
circumspect, but the simplicity of this new bonding mode for dihydrogen found
in the solid state structure dared us to over interpret this result. And unfortunate-
ly we did. Our report in Science in 1997 sparked a lot of interest and it didn’t take
too long before we had Thomas Koetzle and Albert0 Albinati interested in
collaborating to obtain a neutron structure of this crystalline material. Before
Jason Love left my group to return to Sussex University, he grew some crystals of
9 but these weren’t quite big enough for the neutron study. A new postdoctoral
fellow in my group, Wolfram Seidel, was able to grow a variety of larger crystals
of 9 suitable for transport to Grenoble. The result was pretty embarrassing. No
bridging dihydrogen unit, just the same structure that we observed in solution.
About the same time that we got this result, I attended an ACS meeting in Dallas,
Texas where Djamaladdin Musaev, a research associate with Keiji Morokuma
at Emory University, came up to me and told me how hard they had tried to
model the bridging dihydrogen moiety from the low temperature X-ray structure
using DFT calculations but they found that it was unstable. I then recounted
how we had just confirmed that the solution and solid state structures were in
fact identical and did not involve a coordinated dihydrogen unit. As a result of
this chance meeting, we published a joint paper with the results of the DFT
calculations and the neutron structural analysis.16 While the whole story did
eventually get told, it was clear that my over-interpretation of the X-ray data was
unfortunate, despite the happy ending.
5 Trying to Cleave the N-N Bond

A drawback of the above dinuclear zirconium dinitrogen complexes is that one is
limited to a total of four electrons that can be delivered to the dinitrogen
fragment. In other words, even with the strongly activated dinitrogen fragments
that we have been able to isolate, they still contain a N-N bond, albeit a single
bond assuming the N24- formalism. We have tried adding reducing agent (KC,)
to 3 in an effort to cleave the N-N bond, but we were unable to crystallise the
product. One way to circumvent having to add the extra electrons at the end was
to move to Group 5 and generate systems that might intrinsically be able to
supply more electrons to the dinitrogen fragment. Our target molecule was the
tantalum nitride complex, [TaN(P,N,)], which might be accessible by formation
of the tantalum(x1) fragment (Ta(P2N2))and subsequent reaction with N,. We
hoped to mimic the dinitrogen cleavage discovered by the Cummins group at
MIT using [MO(NRA~),].'~,'~ Despite many attempts, our efforts have not met
with success.
However, we have been able to prepare the dinuclear niobium complex
[(Nb(P,N,)),(p-N,)] 11 by reduction of the niobium(II1) precursor
[NbCl(P,N,)] with KC, under N2;l complex 11 contains an essentially linear
end-on bridging dinitrogen fragment. In this case each niobium centre is formal-
ly NbIVand contains one unpaired electron. Attempts to prepare the tantalum
analogue of 11 have been completely unsuccessful, largely because of our inabil-
ity to prepare a TaIII precursor. While much new organometallic chemistry has
been discovered using [TaMe,(P2N2)] as a starting we could not
find ways to access dinitrogen complexes of tantalum using P2N2as an ancillary
ligand. What we did find was that the P2N2macrocycle tended to generate very
unreactive complexes of tantalum, so we redesigned our ligand system.
Sam Johnson's PhD project was the ill-fated tantalum nitride synthesis de-
scribed above. He struggled to generate reactive Ta(P,N,) precursors and in
every case was thwarted by the apparent coordinative saturation that the macro-
cyclic ligand engendered. He reasoned that a new ancillary ligand with fewer
donors was necessary to generate more reactive tantalum precursors. He decided
to prepare the N P N donor shown below. As it turns out, Jason Love and I were
also talking about this same ligand system in the context of Group 4 chemistry
and polymerisation catalysis. While the latter project never panned out (we got
scooped by Schrock on this aspect), Sam's idea worked incredibly well, better
than even I could have expected.
The tantalum(v) precursor [TaMe,(NPN)], prepared from TaMe,Cl, and

(NPN)Li,*thf, reacts smoothly with H, to produce the dinuclear Tar” tetra-
hydride complex [((NPN)Ta),(p-H),], completely analogous to that found for
the related [TaMe,(P,N,)] derivative. However, whereas the tetrahydride pro-
duced from the macrocyclic system is unreactive to ligand addition, exposure of
[((NPN)Ta),(p-H),] to a dinitrogen atmosphere results in a colour change and
the formation of the dinitrogen complex [((NPN)Ta}2(p-y2-y1-N,)(p-H),] 12
that displays a new bonding mode, 2 2 that of side-on and end-on as shown in
Scheme 2.
While we still have some work to do to cleave the N-N bond, the fact that this
change in ligand design allowed us to generate a dinitrogen complex of Group 5
is very encouraging. Work continues in our laboratory to probe the reactivity
patterns of this side-on end-on bound N, unit, as well as examine the prepara-
tion of other dinitrogen complexes using these kinds of ligand designs.
EtO-thf
TaMe,CI,
Ph
(-2LiCI)
Ph
Ph 12
Scheme 2
6 Conclusions
I remember reading in the Huheey texbook, Inorganic about the
headlines that travelled around the world after Joe Chatt showed that protona-
196 Dinitrogen Activation b y Early Transition Metal-Amido Phosphine Complexes
tion of Group 6 dinitrogen complexes generated ammonia stoichiometrically.
The final headline of ‘Basic life processes created in UK lab’ that appeared here
in Vancouver in the daily newspaper The Province in January of 1975 had
certainly promised a lot. The whole area of nitrogen fixation, while vibrant in the
70s and 80s’ seemed to lose some momentum in the late 80s. However, this topic
would appear to be undergoing a renaissance in many different quarters. The
report of the X-ray structure of the active site of the iron-molybdenum nitro-
genase protein from Doug Rees’ g r o ~ p the , ~cleavage
~ ~ ~of~the dinitrogen by
transition metal complexes by a number of different groups, and the discovery of
new reactivity patterns for coordinated dinitrogen, are all helping to fuel re-
newed interest in this important area of research. But it all comes back to a
strong foundation. For that we owe a great debt to Joseph Chatt and the
Nitrogen Fixation Unit. Without their commitment to curiosity-driven research,
my own contributions would lack contextual substance.
7 Acknowledgements
I am grateful to the continued funding of the Natural Sciences and Engineering
Research Council of Canada and to the Petroleum Research Fund, administered
by the American Chemical Society. I especially want to thank all of my talented
co-workers over the past years who have helped make this area of research fun.
8 References
1 M. D. Fryzuk and S. A. Johnson, Coord. Chem. Rev., 2000,200-202,379.
2 A. D. Allen and C. V. Senoff,Chem. Commun., 1965,621.
3 C. V. Senoff, J . Chem. Ed., 1990,67,368.
4 J. E. Bercaw, in Transition Metal Hydrides, ed. R. Bau, Washington, D.C., 1978,
pp. 136.
5 J. M. Manriquez and J. E. Bercaw, J . Am. Chem. Soc., 1974,96,6229.
6 J. M. Manriquez, R. D. Sanner, R. E. Marsh and J. E. Bercaw, J . Am. Chem. Soc., 1976,
98,3042.
7 M. D. Fryzuk, Can. J . Chem., 1992,70,2839.
8 M. D. Fryzuk, T. S. Haddad and S. J. Rettig, J . Am. Chem. Soc., 1990,112,8185.
9 W. J. Evans, T. A. Ulibarri and J. W. Ziller, J . Am. Chem. Soc., 1988,110,6877.
10 M. D. Fryzuk, T. S. Haddad, M. Mylvaganam, D. H. McConville and S. J. Rettig, J .
Am. Chem. SOC.,1993,115,2782.
11 M. D. Fryzuk, M. Mylvaganam, S. J. Rettig, T. M. Loehr and J. D. Cohen, Inorg.
Chem., 1998,37,112.
12 J. D. Cohen, M. D. Fryzuk, M. Mylvaganam and T. M. Loehr, J . Am. Chem. Soc.,
1994,116,9529.
13 M. D. Fryzuk, J. B. Love and S. J. Rettig, J . Chem. Soc., Chem. Cornmun., 1996,2783.
14 M. D. Fryzuk, J. B. Love and S. J. Rettig, Organometallics, 1998, 17, 846.
15 M. D. Fryzuk, J. B. Love, S. J. Rettig and V. G. Young, Jr., Science, 1997,275, 1445.
16 H. Basch, D. G. Musaev, K. Morokuma, M. D. Fryzuk, J. B. Love, W. W. Seidel, A.
Albinati, T. F. Koetzle, W. T. Klooster, S. A. Mason and J. Eckert, J . Am. Chern. Soc.,
1999,121,523.
Michael D. Fryzuk 197
17 C. E. LaPlaza and C. C. Cummins, Science, 1995,268,861.
18 C. E. LaPlaza, M. J. A. Johnson, J. C. Peters, A. L. Odom, E. Kim, C. C. Cummins, G.
N. George and I. J. Pickering, J . Am. Chem. Soc., 1996,118,8623.
19 M. D. Fryzuk, M. A. Bowdridge, and C. M. Kozak, unpublished results.
20 M. D. Fryzuk, S. A. Johnson and S. J. Rettig, Organometallics, 1999,18,4059.
21 M. D. Fryzuk, S. A. Johnson and S. J. Rettig, Organometallics, 2000,19, 3931.
22 M. D. Fryzuk, S. A. Johnson and S. J. Rettig, J . Am. Chem. Soc., 1998,120,11024.
23 J. E. Huheey, E. A. Keiter and R. L. Keiter, Inorganic Chemistry, HarperCollins, New
York, 1993.
24 M. M. Georgiadis, H. Komiya, P. Chakrabarti, D. Woo, J. J. Komuc and D. C. Rees,
Science, 1992,257, 1653.
25 J. Kim and D. C. Rees, Nature, 1992,360,563.
Metal-Dinitrogen Chemistry
After Chatt
CARL0 FLORIANI
Institut de Chimie Minerale et Analytique, BCH 3307, Universite de Lausanne,
CH- 1015 Lausanne, Switzerland
1 Introduction
Reviewing the status of the art of dinitrogen activation is rather difficult, owing
to the variety of widely spread results in the literature, and, generally speaking, to
the prior lack of a methodological approach to the problem.' Even so, it is
possible today to single out some major directions in the historical development
of the field. Bearing in mind that we are still at the stage of acquiring fundamental
knowledge before the dinitrogen molecule will become a useful chemical reagent,
some of the objectives in the use of dinitrogen are: (i) to reduce and hydrogenate
it under mild conditions; (ii) to functionalise organic substrates; (iii) to form
metal-nitrido compounds analogous to metal-carbides in material science; (iv)
to mimic the polymeric forms and the properties of metalla-acetylides. Then, the
quite difficult move will be from stoichiometric to catalytic chemistry. We have
to confess that the bio-oriented approach in the field, in terms of practical results,
has so far been quite disappointing.2 None of the claimed bio-inspired models2
has been shown to be capable even to fix the dinitrogen molecule. Therefore, we
will follow the historical development of biologically unrelated transition metal
compounds devoted to dinitrogen activation. Three major advances can be
singled out in the area, that have arisen since the discovery that a metal complex
was able to bind d i n i t r ~ g e n They
.~ are: (i) coordinative dinitrogen binding;
(ii) reductive dinitrogen binding; (iii) N-N triple bond cleavage to nitride.
2 Coordinative Dinitrogen Binding

The coordination of dinitrogen was initially pursued just for the pleasure of
having such an inert molecule interact with a metal centre.' Then the perception
Carlo Floriani 199
dawned that this aesthetic operation was chemically very important. Although
isoelectronic with CO, the high-energy n* antibonding orbitals of N, prevent
good binding to an electron-rich metal, because of the very limited back dona-
tion. Although electron transfer to N, is rather weak because of the limited back
donation, the binding to the metal has the important consequence of introducing
a significant polarisation into the N, molecule, thus making accessible kinetic
pathways for its reactivity. We should remember that in many cases thermo-
dynamics are not the major problem for the chemical activation of dinitrogen, as
is the case for CO,, for example. The dinitrogen polarisation, which depends on
the extent of the n back donation from the metal, has been proved by the
quantitative study of the formation of acid-base adducts between metal-bonded
dinitrogen and Lewis acids, in pioneering work by Chatt’s scho01.~The elegant
work by Sellmann’ completed our understanding in this regard by transforming
N, into Me-N=N-Me by a sequential reaction of an Mn-N, functionality with
LiMe followed by MeX. These investigations revealed an electron-richness of the
terminal nitrogen, which can be quantitatively related to the polarisation of the
N, moiety in diazoalkanes,6 opening perspectives in two major directions,
namely the protonation7 and the alkylation’d,s of N,.
In the former case, the protonation appeared to be a quite complex reaction,
with a major message: even in the absence of a significant electron-transfer from
metal, the protonation can draw electrons from the metal, and the overall result
is the formation of some NH, and N,H4.7 Chatt’s group provided a major
contribution to this research domain. The protonation of the metal-bonded N,
opened the door to its reaction with organic electrophiles,1d.8the beginning of
the organic functionalisation using the metal-activated forms of N,. The pioneer-
ing work by Chatt’s group was actively pursued mainly by Hidai’s group in
Japan.1d98c Using molybdenum and tungsten complexes having phosphines as
ancillary ligands, this group demonstrated a variety of organic functionalisations
employing both the protonation and the alkylation of metal-bonded dinitro-
gen.
3 Reductive Dinitrogen Binding

The historical move to early transition metals in rather high oxidation states and
the use of 0- and n-donor ligands, contrary to what was observed in the late and
low-oxidation-state transition metals, has been the preferred approach to
achieve a real electron-transfer from the metal to the N, ligand. In this context,
two major strategies have been particularly successful, namely the use of the
ancillary ligands mentioned above and the macrocyclic effect.
Such an approach allows not only the pursuit of the two- but also the more
important four-electron reduction of dinitrogen to a m e t a l l a - h y d r a ~ i d e In
.~~~~
the case of early transition metals, the major strategy has been to take advantage
of both the high energy of the filled d orbitals and the absence of the n-acceptor
co-ligands around the metal. In this way, a much better matching has been
achieved between the high energy of the n* dinitrogen orbitals and the appropri-
ate d orbitals of the metal. This can be exemplified by the end-on bridging
200 Metal-Dinitrogen Chemistry After Chatt
bonding modes of dinitrogen, moving from the form a, seen in the coordinative
bonding mode, to b and c (Chart l).9d31
Chart 1
Such a move is associated with the electron-richness of the metal, the nature of
the donor atoms and the set of the metal frontier orbitals, which should be
preorganised to form one o and one or two n bonds with a nitrogen atom of N,.
This is greatly enhanced by geometrical constraints, which have been provided
mainly by a polydentate or macrocyclic ligand with nitrogen or oxygen donor
atoms. A very significant example in the literature has been reported by Schrock
with his class of trisamidoamine ligands.9a*9by9e*12 In the case of electron-rich
early transition metals supported by conformationally very flexible monoden-
tate or polydentate ligands, the electron transfer to dinitrogen takes place with
side-on bonding modes to the metals."*13 Such a bonding mode sometimes
makes the metal --+ N, electron-transfer process more efficient, as revealed by the
longest N-N bonds (-1.50 A) found in the literature, even longer than in
hydrazine derivatives and hydrazine itself.' The bonding modes shown in Chart
1 are full of implications for new perspectives in dinitrogen chemistry, particular-
ly in the use of N, as the electronic equivalent of [C=C12 - . I 4 The limiting forms
reported in Chart 1 have also been reported for the dimetalla-acetylide deriva-
tives. Monometalla-acetylides and dimetalla-acetylides are considered as the
building blocks of metalla-cumulenes; their potential in material science is just
now dawning.' Analogously, the metal-dinitrogen moiety can be considered as
a potential building block for a variety of metal-dinitrogen po1yme1-s.'~
4 N-N Triple Bond Cleavage to Nitride

N-N triple bond cleavage has always been assumed as a sort of a dream in
dinitrogen activation, due to the major practical use of N, in producing ammo-
nia, and to the fact that in the actual use of heterogeneous catalysts the formation
of nitrides is unanimously considered as the intermediate step in the formation of
NH, by dinitrogen hydrogenation. '
Nevertheless, N-N triple bond cleavage in the homogeneous phase, employ-
ing well-defined transition metal complexes, is an avenue in dinitrogen chemistry
which is quite recent and there are as yet neither many nor diverse examples.
Roughly speaking, the a priori requirements in order to achieve this goal are an
early transition metal having a d2 or d3 configuration, ligands acting with
o-donor atoms without n-accepting properties, and the preorganised arrange-
ment of the donor atoms expressed by macrocyclic or sterically hindered mono-
dentate ligands. Under these conditions the form c (see Chart 1) of the
metal-dinitrogen unit seems suitable to undergo a further two-electron reduc-
tion with the complete cleavage of the residual N-N single bond. Three examples
Carlo Floriani 20 1
have appeared so far in the literature, with some of the common features
mentioned above. Chronologically, the first one is the report by C. C. Cum-
mins' on the spontaneous cleavage of the N-N triple bond occurring according
to Scheme 1. Such a result has been achieved using sterically hindered amido
groups assuming a trigonal planar geometry in binding to a MolI1d3 ion. The
pathway leading to the formation of molybdenum-nitride from N, has also been
accurately investigated. The latest report in the field is by Cloke,', who used a
tridentate silylated diamidoamine ligand for supporting a vanadium(m)-d2,
chloride-bridged dimer. Its reduction with KC, under N, led to the formation of
a dinuclear nitrido species.
\,\")Ar
Ar(R)N-Mo,'
N(R)Ar
Scheme 1
A detailed account of extensive research by the author's group will be given

here.lg This includes numerous aspects of the problems related to metal-assisted
N-N triple bond cleavage. The a priori choice has been a metal with high-energy
d orbitals having either a d2 or d3 electronic configuration and capable of
establishing a multiple bond in a well-defined direction. This has been achieved
using a calixr4larene moiety as an ancillary ligand. Such a choice introduces a
number of unprecedented novelties. They are: (i) a set of oxygen donor atoms
playing the role of 0- and n-donors: (ii) their arrangement in a quasi-planar
geometry; (iii)a single reactive site accessible axially. In addition, by using a set of
oxygen donor atoms which can exercise additional binding towards other metal
ions, we took advantage of secondary thermodynamic driving forces derived
from the solvation of alkali cations associated with the main structure in the
ion-pair or ion-separated forms.
This rather complex strategic approach allowed us in this domain of chemical
synthesis for the first time (i) to achieve the stepwise supply of electrons to
dinitrogen up to the complete cleavage of the N-N triple bond; (ii) to understand
how dinitrogen binding two metal ions rearranges its bonding mode, thus being
ready to cleave to nitride; (iii) to mimic a metal-oxo surface; and (iv) highlight the
relevance of the bifunctionality of the systems. A complete and very detailed
report on stepwise reduction of dinitrogen by the use of [Nb111-calix[4]arene]

dimer 1 has already been p~b1ished.l~ The [Nb111)-calix[4]arene] (see complex 1
in Scheme 2) is a powerful reducing agent for N-N multiple bonds. The reaction
of 1 with dinitrogen led to the formation of the dinuclear metalla-hydrazide, 2,
with four-electron reduction of N, (Scheme 2). This reaction is, however, strongly
dependent on the nature of the solvent used, which should be either thf or dme.
The complete cleavage of the residual N-N single bond in 2 is achieved when two
electrons are provided to the system from the reaction with sodium metal
(Scheme 2). It has been shown that this first step in the cleavage of N-N bonds
can follow two different pathways, depending on the reaction solvents, which are
summarised in Chart 2. The reduction in hydrocarbon solution (see below)
follows a different pathway.
-0,
n,o-
I
1.75 A -111
N
2.66 A
Na Nb-N, 1.91 A
\o//Nt\o/ a) N2, thf, oNb,/\o 2Na, thf
Nb- - -Nb, 2.80 A
N- - -N, 2.61 A
b) digly dme Na
/ \
-0 0-
U
1 2 3
A =dme
-0 no'-=7
digly 0--0 0-
0
I 0 o ?
K a = CpBut-calix[4]-04}tetraanion
Scheme 2
Regardless of whether the solvent is thf or dme (dme = dimethoxyethane), the

two electrons reduce the two NbVto NbIV,with the reduction of the Nb-N triple
bond. At this stage, the Nb=N-N=Nb skeleton undergoes a transoid rearrange-
ment (thf) followed by cleavage to the monomeric nitrido species, in equilibrium
in solution with the corresponding dimer.lg In the case of dme, the
NkN-N=Nb skeleton undergoes a cisoid rearrangement followed by the for-
mation of an Nb-Nb bond (Chart 2). Thus, the two electrons introduced in the
system are temporarily stored in the metal-metal bond. At the same time, the N,
molecule moves from an end-on to a side-on bonding mode, thus being preor-
ganised for the cleavage to the dimeric nitrido species 3, the latter occurring
Carlo Floriani 203
Nb
Nb'
Calix[4Jarenetetraanion around Nb omitted

Chart 2
upon heating. The experimental sequence related to the pathway b in Chart 2

is shown in Scheme 3. The bonding rearrangement of N, does not affect its
extent of reduction, the N-N bond remaining very long (1.43 A).The isolation of
4 mimics not only the rearrangment of N, over a metal-oxo surface, but also
raises the possibility of using the two electrons stored in the Nb-Nb bond for
introducing a further functionality at metals close to the reduced N, moiety.
The isolation of 5 with 0, or pyridine-N-oxide (py0) exemplifies this (Scheme 3).
Scheme 3
The active species (Nb=Nb), 1, can perform the six-electron reduction of N,

without the addition of any further reducing agent when the reaction is carried
out in toluene. The reaction leads to the trinuclear 1.1,-bisnitrido species, which is
rather labile in the presence of solvents binding alkali cations, leading to the
compounds 7 and 8 (Scheme 4). A detailed report has been published on why the
solvent can drastically affect the reduction pathway of N, mediated by metalla-
cali~[4]arenes.’~The bifunctionality of the complexes used is such that the
solvation of the alkali cation can be an important driving force and at the same
time the presence of tight-ion pair or separated-ion forms can affect the kinetic
pathways (see Chart 3).
In conclusion, this section reports a number of novelties in the metal-dinitro-
gen chemistry scenario.’ For the first time, it has been shown that the active
species performing the reduction of N, is a very reactive M=M functionality, and
that an ancillary ligand containing exclusively oxygen donor atoms can be
successfully employed. The presence of alkali metal ions in the active bifunc-
tional species reinforces the role of the solvent in dinitrogen reduction assisted by
transition metal complexes. In particular, the solvent allows one to select either
the four- (thf or dme) or the six-electron (toluene) reduction of dinitrogen to
hydrazine or to ammonia, respectively. In the former case, a fine tuning of the
toluene
tmen
tmen = M@NCH2CH2NMe2 \
Scheme 4
Carlo Floriani 205
2 I- k2
Nb
k2 (thf, dme)
Nb 1 4 - Calix(4larene tetraanion
around Nb omitted
N4 N
‘
“i?
Chart 3
+2e-
___z
thf
’”’ A
/ 1.43
I 8, ‘I
{Nb),
{N’bf
Y‘NI +2e-
dme
2.87 A 2.05A
{Nb} stands for the Nb-calix[4]arene moiety
Chart 4
solvent (dme or thf) drives the ultimate two-electron reduction of N, to nitride

via different intermediates. Particularly relevant in this context is the rearrange-
ment of the bonding mode of dinitrogen from end-on to p-q2-q2over a metal-oxo
surface modelled by the (Nb-calixC41arene) fragment. A comprehensive sum-
mary of the different stepwise pathways leading to the reduction of dinitrogen to
nitrides is given in Chart 4. The results presented here for a challenging area of
~ ~* ~such
r e ~ e a r c h ,7,1 ~ ' as dinitrogen activation, show that, using appropriate
model compounds, we were able to detect unprecedented pathways in dinitrogen
reduction and to open up novel perspectives in the field.
5 Acknowledgements
We thank the Fonds National Suisse de la Recherche ScientiJique (Bern, Switzer-
land, Grant No. 20-61'246.00) for financial support.
6 References
1 For a general review of transition metal dinitrogen complexes see: (a) A. D. Allen, R. 0.
Harris, B. R. Loescher, J. R. Stevens and R. N. Whiteley, Chem. Rev., 1973,73,11; (b)J .
Chatt, J. R. Dilworth and R. L. Richards, Chem. Rev., 1978, 78, 589; (c) R. A.
Henderson, G. J. Leigh and C. J. Pickett, Adv. Inorg. Chem. Radiochem., 1983,27, 197;
(d) M. Hidai and Y. Mizobe, Chem. Rev., 1995, 95, 1115; (e) A. E. Shilov, Metal
Complexes in Biomimetic Chemical Reactions, CRC, Boca Raton, FL, 1997; (f) F.
Tuczek, Angew. Chem., Znt. Ed. Engl., 1998, 37, 2636; (8) G. J. Leigh, New J . Chem.
1994,18, 157.
2 (a) M. K. Chan, J. Kim and D. C. Rees, Science, 1993,260,792; (b) J. B. Howard and D.
C. Rees, Chem. Rev., 1996,96,2965; (c) B. K. Burgess and D. J. Lowe, Chem. Rev., 1996,
96,2983; (d) R. R. Eady, Chem. Rev., 1996,96,3013; (e) D. Sellmann and J. Sutter, Acc.
Chem. Res., 1997,30,460; (f) Transition Metal Sulfur Chemistry, ed. E. 1. Stiefel and K.
Matsumoto, ACS Symposium Series, American Chemical Society, Washington, DC,
1996, p. 653: (8)J. A. Ibers and R. A. C. Holm, Science, 1980,209,223; (h) K. K. Stavrev
and M. C. Zerner, Chem. Eur. J., 1996, 2, 83; (i) H. Deng and R. Hoffmann, Angew.
Chem., Int. Ed. Engl., 1993, 32, 1062; ('j)D. Sellmann, Angew. Chem., Int. Ed. Engl.,
1993,32, 64.
3 A. D. Allen and C. V. Senoff, Chem. Comm., 1965,621.
4 J. Chatt, R. H. Crabtree, E. A. Jeffery and R. L. Richards, J . Chem. Soc., Dalton Trans.,
1973,1167.
5 D. Sellmann and W. Weiss, Angew. Chem., Znt. Ed. Engl., 1977,16, 880; 1978,17,269.
6 D. Sutton, Chem. Rev., 1993,93,995.
7 (a) G. J. Leigh, Acc. Chem. Res., 1992,25,177;(b) D. A. Hall and G. J. Leigh, J . Chem.
Soc., Dalton Trans., 1996, 3539; (c) B. E. Smith, M. C. Durrant, S. A. Fairhurst, C. A.
Gormal, K. L. C. Grondberg, R. A. Henderson, S. K. Ibrahim, T. Le Gall and C. J.
Pickett, Coord. Chem. Rev., 1999,185186,669; (d) M. A. Tyson and D. Coucouvanis,
Znorg. Chem., 1997,37,3808; (e) R. L. Richards, Pure A p p l . Chem., 1996,68,1521; (f) Y.
Nishibayashi, I. Shotaro and M. Hidai, Science, 1998,279,540.
8 (a) J. Chatt, G. A. Heath and G. J. Leigh, J . Chem. Soc., Chem. Commun., 1972,444; (b)
J. Chatt, A. A. Diamantis, G. A. Heath, N. E. Hooper and G. J. Leigh, J . Chem. Soc.,
Dalton Trans., 1977,688; (c) M. Hidai, Coord. Chem. Rev. 1999,185-186,99.
9 (a) M. B. O'Donoghue, N. C. Zanetti, W. M. Davis, and R. R. Schrock, J . Am. Chem.
Soc., 1997,119,2753; (b) K-Y. Shih, R. R. Schrock and R. Kempe, J . Am. Chem. SOC.,
1994,116,8804; (c) A. L. Odom, P. L. Arnold, and C . C. Cummins, J . Am. Chem. Soc.,
1998,120,5836; (d) R. Ferguson, E. Solari, C. Floriani, D. Osella, M. Ravera, N. Re, A.
Carlo Floriani 207
Chiesi-Villa and C. Rizzoli, J . Am. Chem. Soc., 1997, 119, 10104; (e) M. B. O’Do-
noghue, W. M. Davis and R. R. Schrock, Znorg. Chem., 1998,37,5149.
10 (a) M. B. O’Regan, A. H. Liu, W. C. Finch, R. R. Schrock and W. M. Davis, J . Am.
Chem. Soc., 1990,112,4331; (b) R. R. Schrock, M. Wesolek, A. H. Liu, K. C. Wallace
and J. C. Dewan, Znorg. Chem., 1988,27,2050; (c) R. R. Schrock, T. E. Glassman, M. G.
Vale and M. Kol, J . Am. Chem. Soc., 1993, 115, 1760; (d) S. M. Rocklage and R. R.
Schrock, J . Am. Chem. SOC., 1982,104,3077; (e) S. M. Rocklage, H. W. Turner, J. D.
Fellmann, and R. R. Schrock, Organometallics, 1982, 1, 703; (f) H. W. Turner, J. D.
Fellmann, S. M. Rocklage, R. R. Schrock, M. R. Churchill and H. J. Wasserman, J .
Am. Chem. Soc., 1980,102,7809; (g) M. R. Churchill, Y.-J. Li, K. H. Theopold and R. R.
Schrock, Znorg. Chem., 1984,23,4472;(h) R. R. Schrock, R. M. Kolodziej, A. H. Liu,
W. M. Davis and M. G. Vale, J . Am. Chem. Soc., 1990,112,4338.
11 M. D. Fryzuk, T. S. Haddad, M. Mylvaganam, D. H. McConville and S. J. Rettig, J .
Am. Chem. Soc., 1993,115,2782.
12 (a) R. R. Schrock, Acc. Chem. Res. 1997,30,9, and references therein; (b) M. Kol, R. R.
Schrock, R. Kempe and W. M. Davis, J . Am. Chem. Soc., 1994, 116, 4382; (c) M. B.
O’Donoghue, W. M. Davis, R. R. Schrock and W. M. Reiff, Znorg. Chem., 1999,38,
243.
13 (a) J. D. Cohen, M. D. Fryzuk, T. M. Loehr, M. Mylvaganam, and S. J. Rettig, Znorg.
Chem., 1998,37, 112; (b) M. D. Fryzuk, T. S. Haddad and S. J. Rettig, J . Am. Chem.
Soc., 1990, 112, 8185; (c) M. D. Fryzuk, J. B. Love, S. J. Rettig and V. G. Young,
Science, 1997, 275, 1445; (d) R. Duchateau, S. Gambarotta, N. Beydoun and C.
Bensimon, J . Am. Chem. Soc., 1991,113,8986;(e) M. D. Fryzuk, S. A. Johnson and S. J.
Rettig, J . Am. Chem. Soc., 1998,120,11024; (f) E. Campazzi, E. Solari, C. Floriani and
R. Scopelliti, Chem. Commun., 1998,2603.
14 E. Solari, J. Hesschenbrouck, R. Scopelliti, C. Floriani and N. Re, Angew. Chem., Znt.
Ed. Engl., 2001,40, 932.
15 Modern Acetylene Chemistry, eds. J. P. Stang and F. Diederich, VCH, Weinheim, 1995.
16 (a) Catalytic Ammonia Synthesis: Fundamentals and Practice, ed. J. R. Jenning, Plenum,
New York, 1991, Vol. 1; (b) Ammonia: Catalysis and Manufacture, ed. A. Nielsen,
Springer, New York, 1995; (c) T. Trevis, Chim. Znd., 1993, 581.
17 (a) C. E. Laplaza, M. J. A. Johnson, J. C. Peters, A. L. Odom, E. Kim, C. C. Cummins,
G. N. George and I. J. Pickering, J . Am. Chem. Soc., 1996,118,8623,and references
therein; (b) J. C. Peters, J.-P. F. Cherry, J. C. Thomas, L. Baraldo, D. J. Mindiola, W.
M. Davis and C. C. Cummins, J . Am. Chem. Soc., 1999,121,10053.
18 G. K. B. Clentsmith, V. M. E. Bates, P. B. Hitchcock and F. G. N. Cloke, J . Am. Chem.
Soc., 1999, 121, 10444.
19 (a) A. Zanotti-Gerosa, E. Solari, L. Giannini, C. Floriani, A. Chiesi-Villa and C.
Rizzoli, J . Am. Chem. Soc., 1998, 120, 437; (b) A. Caselli, E. Solari, R. Scopelliti, C.
Floriani, N. Re, C. Rizzoli and A. Chiesi-Villa, J . Am. Chem. Soc., 2000,122, 3652.
Novel Chemical Transformations
at Diruthenium Centres Bridged
by Thiolato Ligands
MASANOBU HIDAI," YOUICHI ISHIIbAND SHIGEKI
KUWATA~
aDepartment of Materials Science and Technology, Faculty of Industrial
Science and Technology, Science University of Tokyo, 2641 Yamazaki, Noda,
Chiba 278-8510, Japan
Department of Chemistry and Biotechnology, Graduate School of
Engineering, The University of Tokyo, Hongo, bunk yo-ku, Tokyo 113-8656,
Japan
1 Introduction
The transformation of small molecules on multimetal centres has received much
attention because it is expected to lead to novel reactions which are rarely
attainable at single metal centres. This might occur through the cooperative
interaction of two or more adjacent metal atoms with a substrate molecule, or
simultaneous activation of more than two substrates, or both. Indeed, unique
and efficient chemical transformations are realised on sulfur-bridged multimetal
active sites in metalloenzymes such as nitrogenase and hydrogenase. These
well-designed biological multimetal systems inspired us to investigate the prep-
aration of sulfur-bridged multimetal complexes and their reactivities towards
nitrogenase substrates and related small molecules, such as hydrazines, CO, and
alkynes. We have synthesised a series of thiolato-bridged diruthenium complexes
with structural diversity, as outlined in our earlier review articles.' We sum-
marise here our recent studies on the reactivities of these complexes.
2 Reactions of Hydrazines at Thiolato-bridged

Diruthenium Centres
Crystallographic studies on nitrogenase have revealed that its active site, the
Masanobu Hidai, Youichi lshii and Shigeki Kuwata 209
FeMo-cofactor, contains six coordinatively unsaturated, trigonal-planar Fe
centres bridged by sulfido ligands., This unique structural feature stimulated
us to examine the reactions of the doubly coordinatively unsaturated
diruthenium(I1)complex [Cp*Ru(p-SPr'),RuCp*] 1 (Cp* = y5-C,Me,) with ni-
trogenous substrates. Complex l is inert toward molecular dinitrogen itself.
However, treatment of 1 with phenylhydrazine affords the p-phenyldiazene
'
complex [Cp*Ru(p-y' :rj -PhN=NH)(p-SPr'),RuCp*] 2 with concurrent forma-
tion of aniline and ammonia (Equation l).3 The presence of the p-y':y'-
PhN=NH ligand in 2 was deduced by IR and 'H NMR spectroscopy as well as
by a preliminary X-ray diffraction study of the 2,4,6-triisopropylbenzenethiolato
analogue of 2, as depicted in Figure l(a).4
Furthermore, the disproportionation reaction of hydrazine into ammonia and
dinitrogen proceeds in toluene at 40°C in the presence of a catalytic amount of 1,
according to the stoichiometry shown in Equation 2.3The 'H NMR spectrum of
the reaction mixtures revealed the presence of the diazene complex [Cp*Ru(p-
y 1 :y '-HN=NH)(p-SPr'),RuCp*] 3, although isolation and full characterisation
of 3 were unsuccessful. A plausible catalytic cycle based on these observations is
depicted in Scheme 1. Considering that hydrazine and diazene species bound at
the multimetal centre of the FeMo-cofactor have been proposed as important
intermediate stages of the biological nitrogen f i ~ a t i o n the
, ~ dinuclear diazene
complex 3, which plays a key role in the N-N bond cleavage of hydrazine, is of
special interest from the mechanistic point of view of nitrogenase.
On the other hand, the reactions of the cationic RulI1complex [Cp*RuCl(p-
SPr'),Ru(OH,)Cp*](OTf) 4 (OTf = CF,SO,) with RNHNH, (R = Ph or H)
result in the formation of the terminal hydrazine complexes [Cp*RuCl(p-
SPr'),Ru(NH,NHR)Cp*][OTf] 5, which have been characterised spectroscopi-
cally as well as by a preliminary X-ray analysis for the phenylhydrazine complex
5a (R = Ph; Figure l(b)).4No disproportionation product has been detected in
these reaction mixtures, in contrast to the reactions shown in Equations 1 and 2.
Cp*Ru\\/,?cp* + 2PhNHNH2
-
PrlS' SPrl
pPN, =N, ,H
1
toluene
Cp*Ru, RuCp*+ PhNH2 + NH3 (1)
rt, 8 h \"
PrlS SP?
2
1 (2 mol%)
3N2H4 w N2 + 4NH3
toluene, 40 "C, 16 h
3 Transformations of Small Organic Molecules at

Thiolato-bridged Dirutheniurn Centres
Various substrates other than those containing nitrogen have also been demon-
210 Novel Chemical Transjormations at Diruthenium Centres
(a) (b)
Figure 1 Structures of (a) [Cp*Ru(p-q1:q'-PhN=NH)(p-SC,H,Pri,-2,4,6),RuCp*]
and (b) the cationic part of 5a
2N2H4 2 NH3
H. .H
N2 + 2NH3 N2H4
Scheme 1
strated to be activated and transformed on thiolato-bridged diruthenium com-

plex such as 1.' For example, we have isolated the bis(a1kene)diruthenium
complex [{C~*RU(H,C=CHCN)(~-SP~')),~~ as well as the butadiene complex
[(Cp*Ru(p-SFc)},(CH,=CHCH=CH,)](Fc = ferr~cenyl).~ In this section, we
describe the transformations of alkynes and other unsaturated organic molecules
on the thiolato-bridged diruthenium complexes.
3.1 Stoichiometric Transformation of Alkynes

We have already found that the reactions of the coordinatively unsaturated
diruthenium(I1) complex 1 with alkynes result in unique coupling of the alkynes
to afford a variety of ruthenacycles, depending significantly upon the substituent
on the alkynes.' On the other hand, the corresponding reactions of the
Masanobu Hidai, Youichi Ishii and Shigeki Kuwata 21 1
RulI/RulI1complex [Cp*Ru(p-SPr'),RuCp*] 6 give the bisalkynyl complexes
[Cp*Ru(C=CR)(p-SPr'),Ru(CmCR)Cp*] (R = aryl), which are further converted
into diruthenacyclopentadienoindan complexes upon protonation.' Here we
focus on the transformations of alkynes by the RulI1/RulI1complexes 4 and
[Cp*RuCl(p-SR),RuClCp*] 7 (R = Me or Et).
'
RT' -
1
Ru-RU
\ /
OTf
To&&
-
1
R"
12 (R"= OMe or Me)
OTf
-loTf
?u--Ru
o@OMe
Me0
R"
16
(R = COOMe)
-
1
Ru-Ru
OTf n lOTf
Ru-RU
n lOTf
Fiy-Ru
PoMe
R'
R f R
11 10 / 15
RCGH (R = H or COOMe)
-
1
yu-qu
OTf
RC=CH H O Ru-RU
L
-
1 OTf
____)
CI H20
~~
8a:R=H
8 b R = COOMe
8c: R = COMe
1- H+
-
1
qu--Fp
OTf
CI c (R = H)
14
8
+ \
HCI -TOTf
F?J--R/J
-TOTf -IoT
R' w R'
(R' = OMe or Me) l3
Me
Scheme 2
212 Novel Chemical Transformations at Diruthenium Centres
The reactions of the cationic complex 4 with acetylene or alkynes having

electron-withdrawing substituents afford the terminal vinylidene complexes
[C~*RUC~(,U-SP~~)~RU(=C=CHR)C~*](OT~) 8 (R = H, COOMe, or COMe) as
shown in Scheme 2.' When aromatic terminal alkynes are used, two molecules of
the alkyne are incorporated into the diruthenium centre of 4 to give the cationic
diruthenacycle complexes 9, which are probably formed via the alkynyl-vi-
nylidene and butenynyl intermediates 10 and 11 derived from 8.9 In fact, the
butenynyl complex 11 (R = R' = Fc) has been isolated from the reaction of 4 and
ferrocenylacetylene. Coupling of two different alkynes takes place in a similar
manner to give the diruthenacycle complexes 12 when the vinylidene complexes
8b and 8c are treated with p-tolylacetylene.'
The vinylidene complexes 8 also react with nucleophiles such as water and
methanol. Hydration of the unsubstituted vinylidene complex 8a leads to the
formation of the p-y' : yl-acetyl complex 13, whereas the corresponding reac-
tions of 8b and 8c result in the C-C bond cleavage, giving the cationic carbonyl
complex 14 and organocarbonyl products R'COMe ( R = OMe or Me).* On the
other hand, the reactions of 8 with methanol afford the methoxycarbene complex
15 and the vinyl complex 16, depending upon the substituent of the vinylidene
complexes 8. Intramolecular nucleophilic attack takes place in the reaction of 4
with 3-butyn- 1-01, giving the cyclic alkoxycarbene complex [Cp*RuCl(p
SPr'),Ru(=C(CH,),O } Cp*](OTf).'
3.2 Catalytic Head-to-head 2 Dimerisation of Terminal Alkynes

Interestingly, the butenynyl complex 11 (R = R' = Fc) has been found to be an
efficient catalyst for linear di- and trimerisation of ferrocenylacetylene (Equation
3).1° Our subsequent investigations have led to the finding that the
diruthenium(II1)complex [Cp*RuCl(p-SMe),RuClCp*] 7a with bridging MeS
ligands catalyses the head-to-head (2)-dimerisation of various terminal alkynes
stereoselectively (Equation 4).l
11 (R = R' = Fc)
5 mol%
Fc-H
CICH,CH,CI
Fc Fc
60 "C, 30 h
7a (6.5 mol%), NH4BF4

2 RH
- (4)
MeOH, 60 "C
R
Even aliphatic alkynes with substituents such as chloro, hydroxy, and ester
groups are effectively transformed into the corresponding head-to-head (2)-
dimers by 7a. Noteworthy is the strong dependence of the catalytic activity
Masanobu Hidai, Youichi Ishii and Shigeki Kuwata 213
HC$ ' ~
CI CI
2 RC&H
&HCl
f
7a
L
Scheme 3
upon the bridging thiolato ligands. Thus, the sterically hindered Pr'S-bridged
complex [Cp*RuCl(p-SPr'),RuClCp*] 7b exhibits only marginal activity in
contrast to the primary alkanethiolato complexes 7a and [Cp*RuCl(p-
SR),RuClCp*] (R = Et or Pr"). A plausible mechanism is shown in Scheme 3,
which shows a butenynyl intermediate.
These dimerisation reactions of terminal alkynes have been further extended
to the catalytic cyclisation of a,co-diynes. For example, treatment of 1,15-hexa-
decadiyne with 10 mol% of 7a affords the endo-macrocyclic product, (2)-1-
cyclohexadecen-3-yne with complete stereoselectivity (Equation 5).' This novel
cyclisation is of particular utility, because synthetic routes to endo-cyclic (2)-1-
en-3-ynes are extremely limited. A related palladium-catalysed cyclisation of
a,m-diynes to give the corresponding em-cyclic 1-en-3-ynes has been reported by
Trost and co-worker~.'~
7a (10 mol%), N b B F 4
MeOH, 60 "C
3.3 Transformation of Propargyl Alcohols

We have already revealed the unique coupling reactions of two moles of propar-
gyl alcohols on the RulI/RulI1complex 6, which give two types of diruthenacycle
complexes depending upon the substituents of the alcohols.15 Formation of the
terminal allen y lidene complexes [Cp*RuCl(p- SPr'), Ru(=C=C=CR,)Cp*] (OTf)
from the cationic diruthenium complex 4 and propargyl alcohols has also been
demonstrated.’ On the basis of these findings, we have recently developed novel

propargylic substitution reactions of propargyl alcohols catalysed by the
thiolato-bridged diruthenium complex 7a (Scheme 4).16 These reactions enjoy
several synthetic advantages. A variety of nucleophiles including alcohols,
amines, amides, and thiols can be employed to realise the substitution at the
propargylic position with complete regioselectivity. Allenic by-products, which
P(=O)Ph*
Scheme 4
are always produced by the classical propargylic substitutions, are not observed
at all. In addition, easily available propargyl alcohols are used as the substrates
without derivatisation to the corresponding halides or esters. The proposed
reaction mechanism involves the attack of nucleophiles at the y-carbon atom in
an allenylidene intermediate 17 (Scheme 5). These catalytic reactions are in sharp
contrast to the Nicholas reaction, in which a stoichiometric amount of
[co,(c0)8] is used to achieve the propargylic substitution. It is also to be
emphasised that these reactions provide some of the few examples of catalytic
reactions via allenylidene intermediates.
3.4 Other Catalytic Reactions

We have further examined the reactions of thiolato-bridged diruthenium com-
plexes with other unsaturated organic substrates. When the cationic Ru”’ com-
plex 4 is treated with cinnamyl alcohol in p-xylene, the allylated aromatic
compound 18 is obtained in good yield (Equation 6).17 We assume a n-ally1
intermediate because the reaction using 1-phenylprop-2-en-1-01gives, instead of
cinnamyl alcohol, the same product 18; however, the detailed reaction mechan-
ism is still obscure. This novel allylation reaction is halogen-free, and may
replace the conventional Friedel-Crafts alkylation.
The cationic RulI1complex 4 also promotes silylative dimerisation of aromatic
aldehydes with hydrosilanes. For example, the reaction of benzaldehyde and
triethylsilane in the presence of a catalytic amount of 4 affords the dimerisation
product 19 along with a small amount of the hydrosilylation product
PhCH,OSiEt, (Equation 7).18 This type of silylative dimerisation of aldehydes is
relatively scarce in the literature; common ruthenium complexes such as
[RuCl,(PPh,),] and [Ru,(CO), ,] give only the hydrosilylation products.
Masanobu Hidai, Youichi Ishii und Shigeki Kuwata 215
" 1
R qu-
A I+ f$ll
Nu
nl + K
L Nu
NU-ti
R
Scheme 5
PhvoH +- a 4 (5mot%)
140°C,2h
18
PhCHo + * HSiEt3
MeCN, mol%)
4 (4120 OC, 24 t-,*
ph)+ph
Et3Si0 OSiEt3
+ H2 (7)
19
4 Concluding Remarks
We have shown that the thiolato-bridged diruthenium centres promote a variety
of novel transformations of small molecules such as hydrazines and alkynes.
Apparently, some of these transformations involve the cooperation of the two
ruthenium atoms. Another important feature of these transformations is the
strong dependence of the reactivities upon the oxidation states of the metals and
the substituents on the bridging thiolato ligands. We believe that new types of
reactions on multimetal centres will be further developed by designing metal
frameworks and ancillary ligands.
5 References
1 M. Hidai, Y. Mizobe and H. Matsuzaka, J . Organomet. Chem., 1994,473, 1; M. Hidai
and Y. Mizobe. Transition Metal Sulfur Chemistry: Biological and Industrial Sign$-
cance, American Chemical Society, Washington DC, 1996, Chapter 19.
2 J. B. Howard and D. C. Rees, Chern. Reu., 1996,96,2965.
3 S. Kuwata, Y. Mizobe and M. Hidai, Inorg. Chem., 1994,33,3619.
4 S. Kuwata, Y. Mizobe and M. Hidai, J . Inorg. Biochem., 1997,67,281.
5 K. D. Demadis, S. M. Malinak and D. Coucouvanis, Inorg. Chem., 1996,35,4038.
6 J.-P. Qii, H. Matsuzaka, Y. Ishii and M. Hidai, Chem. Lett., 1996, 767.
7 H. Matsuzaka, J.-P. Qii, T. Ogino, M. Nishio, Y. Nishibayashi, Y. Ishii, S. Uemura
and M. Hidai, J . Chem. Soc., Dalton Trans., 1996,4307.
8 Y. Takagi, H. Matsuzaka, Y. Ishii and M. Hidai, Organornetallics, 1997,16,4445.
9 H. Matsuzaka, Y. Takagi and M. Hidai, Organometallics, 1994,13, 13.
10 H. Matsuzaka, Y. Takagi, Y. Ishii, M. Nishio and M. Hidai, Organometallics, 1995,14,
2153.
11 Y. Ishii, K. Ogio, M. Nishio, M. Retbarll, S. Kuwata, H. Matsuzaka and M. Hidai, J .
12 J.-P. Qii, D. Masui, Y. Ishii and M. Hidai, Chem. Lett., 1998,1003.
13 Y. Nishibayashi, M. Yamanashi, I. Wakiji and M. Hidai, Angew. Chem., Int. Ed. Engl.,
2000,39,2909.
14 B. M. Trost, S. Matsubara and J. J. Caringi, J . Am. Chem. Soc., 1989,111,8745.
15 H. Matsuzaka, H. Koizumi, Y. Takagi, M. Nishio and M. Hidai, J . Am. Chem. SOC.,
1993,115,10396.
16 Y. Nishibayashi, I. Wakiji and M. Hidai, J . Am. Chem. Soc., 2000,122, 11019.
17 Y. Nishibayashi, M. Yamanashi, Y. Takagi and M. Hidai, J . Chem. Soc., Chem.
Commun., 1997,859.
18 H. Shimada, J.-P. Qii, H. Matsuzaka, Y. Ishii and M. Hidai, Chem. Lett., 1995,671.
The Chemistry and Applications
of Complexes with Sulfur
Ligands
JONATHAN R. DILWORTH", PHILIP ARNOLD", DAVID
MORALE$ YEE-LOK WONG" AND YIFAN ZHENG"
aInorganic Chemistry Laboratory, South Parks Road, Oxford OX1 3QR, UK
Department of Chemistry, University of Hawaii, 2545 McCarthy Mall,
Honolulu, Hawaii 96822, USA
1 Introduction
One of us (JRD) had the great good fortune to work under Joseph Chatt's
supervision for some fifteen years at the then Unit of Nitrogen Fixation. It left an
indelible impression. Joseph had a truly comprehensive grasp of the chemistry
and periodicity of the elements and this was allied to an unerring instinct for
where interesting coordination chemistry was to be found. The remit of the
chemists in the Unit was of course to study nitrogen fixation, but the research
scientists were allowed a substantial measure of freedom in the individual areas
that were studied. Joseph believed firmly that an indirect approach often proves
more fruitful than a direct frontal assault. Indeed the first genuinely extensive
series of dinitrogen complexes (of rhenium) arose from a project directed to
understand the mechanism of formation of nitrides from hydrazine.' Many of
the early dinitrogen complexes contained tertiary phosphine co-ligands and the
rational design of such dinitrogen-binding sites became feasible. Although these
provided invaluable information on the mechanisms of dinitrogen reduction
they were structurally far removed from the sulfur-ligated metals in the
iron-molybdenum co-factor. This prompted a search that is still in progress to
establish the factors that determine how a sulfur-ligated metal ion will bind
dinitrogen. The initial work that we did in the Unit on metal-sulfur coordination
chemistry stimulated my long-term research interest in the binding and activa-
tion of small molecules at metal-sulfur sites and the potential applications of
such complexes in areas as diverse as catalysis and the treatment of cancer. This
218 The Chemistry and Applications of Complexes with Sulfur Ligands
review describes recent results on the coordination chemistry of a range of
anionic and neutral sulfur ligands with metals in Groups 6-1 1. Appropriately
perhaps this encompasses many of the elements, the chemistry of which Joseph
was instrumental in opening up in the 1960s.
2 Sterically-hindered Thiolates
Our initial studies of small molecule activation by sulfur-ligated metals involved
the use of sterically-hindered aromatic thiolates where the tendency for forma-
tion of inactive thiolate bridged species was minimised. The success of this
strategy was illustrated by the isolation of stable complexes such as
[Mo(tipt),(CO)] - (tipt = 2,4,6-triisopropylthiophenolate).This species has a
trigonal bipyramidal structure with apical C O ligands., The structure occurs
throughout the chemistry of such ligands and was repeated in the 14-electron
complex [Re(tipt)3(N2)(PPh3)],3still a very rare example of a metal thiolate
species which binds N,. We have recently been extending the reactions of bulky
thiolates to Ru, Os, Rh and Ir with the objective of investigating their catalytic
activity.
Square-planar mononuclear or binuclear complexes of the types [M( SAr)-
(CO)(PPh,),] and [(M(SAr)(CO)(PPh,)),] are obtained by the straightforward
reaction of [MX(CO)(PPh,),] (M = Rh, X = F; M = Ir, X = C1) with the free
thiol, base being required for Ir. The size of the thiol and the ease of loss of a PPh,
dictate whether a monomer of dimer is f ~ r m e dX-ray. ~ crystal structures of both
monomeric (M = Ir, SAr = 2,6-C1,C,H3S) and dimeric forms (M = Rh,
SAr = tipt) were reported. All the complexes were highly active for the hydrofor-
mylation of 1-heptene (comparable to rates for [RhCl(PPh,),] under the same
conditions) and showed good regioselectivity for linear as opposed to branched
aldehyde. More recently we have investigated the chemistry of ruthenium@)with
bulky aromatic thiolate ligands. There have been few systematic investigations of
the chemistry of Ru" with such ligands, and none of their catalytic activity.
Reaction of 2,6-dichlorothiophenol (Hdct) with [RuCl,(PPh,),] in methanol
gave the complex [Ru(dct),(PPh,),] in high yield. An X-ray crystal structure
(Figure 1) showed an octahedral structure with a chloro group on each thiolate
interacting with the metal by electron-pair donation, conferring an overall
18-electron configuration. This contrasts with the structure of the analogous
complex with pentafluorothiophenol where the electron deficiency of the metal
centre is relieved by agostic interactions with phosphine phenyl hydrogen^.^ The
difference reflects the greater electron-pair donating capability of chloride com-
pared to fluoride. The bulky triphenylphosphine ligands in the dct complex
are cis. The complex [Ru(dct),(PPh,),] is completely inactive for the hydro-
genation of hex-1-ene, but shows approximately three times the activity of
[RuCl,(PPh,),] for the hydrosilylation of benzophenone to phenylethanol
under comparable conditions. It appears that the Ru-C1 bonds are too strong
to permit the metal to activate dihydrogen and/or bind the substrate. Electro-
spray mass spectrometry of solutions of the complex in MeCN identified species
such as [Ru(dct)(MeCN),(PPh,),] +,suggesting that one thiolate ligand may be
Jonathan R. Dilworth et al. 219
Figure 1 The X-ray structure of [Ru(SC,H,C1,-2,3)(PPh3),]
lost during the catalytic cycle. We are currently investigating the chemistry of
other bulky thiols with a range of RuI1precursors with tertiary phosphine and
nitrogen donor ligands.
3 Phosphinothiolate and Related Ligands

In the course of our investigations of the chemistry of sterically encumbered
thiolate ligands we observed several examples where the substituents on the
aromatic group of the thiol were not mere spectator groups, but became coor-
dinated to the metal. These interactions ranged from agostic hydrogen interac-
tions in [Re(dmt),(PPh,),] (dmt = 2,6-dimethylthiophenolate) to y6-bonding of
an arene group in [MO((2-y6-C6H,)C6H,S-1-Ph-6}(SC6H,-2,6-Ph2)(CO)].6
The bonding interactions of the substituents with the metal range from weak to
very strong, raising the possibilities of both hemi-labile and stable chelate
behaviour for these classes of thiolate ligand. When starting our studies of
catalysis by metal thiolate complexes we were concerned at the possibility of
facile elimination of free thiol from possible thiolate hydride intermediates, and
the strong binding characteristics of tertiary phosphine groups prompted us to
initiate a study of phosphinothiolate proligands. We have recently published a
comprehensive review of this area7 which illustrates that such ligands chelate
strongly to an extremely wide range of metals and non-metals, and that stable
complexes can be formed even with elements that generally interact weakly with
thiolate sulfur and tertiary phosphines. The three examples from our recent work
below serve to illustrate the diversity of the coordination chemistry that is
accessible.
We recently reported4 the reactivity of the Rhl complex [Rh(PS)(CO)(PPh,)]
(PS = 2-Ph2PC6H4S) for oxidative addition reactions with a range of elec-
trophilic reagents. These reactions are summarised in Scheme 1. The reversible
protonation of the precursor complex with HBF, provides a unique example of

reversible formation of a dithiolate bridged system with an M-M bond (the
X-ray crystal structure appears in Figure 2, with phosphine phenyl groups
omitted for clarity).
Scheme 1
16
Figure 2 The structure of[(Rh(PS)(CO)(PPh,)2)2]2+,

anions and phosphine phenyls
omitted
Jonathan R. Dilworth et al. 22 1
Other electrophiles react at the sulfur of the PS ligand, showing that suscepti-
bility to attack at the metal or sulfur is a delicate balance of steric and electronic
effects. The quantitative formation of acetone by reaction with an excess of Me1
(together with a triiodide bridged dimer, Scheme 1) also illustrates how the
presence of anionic sulfur ligands can dramatically alter the course of a reaction
([RhCl(CO)(PPh,),] gives virtually no acetone under the same conditions) and
augurs well for the possibility of thiolate-coordinated metal sites generating
unique catalytic behaviour. We have also synthesised a binucleating phos-
phinothiolate proligand (HPSP) via a multistage synthesis starting from 2,6-
dimethylbromobenzene8 (Scheme 2). Such ligands support stable dinuclear
structures as illustrated by the X-ray crystal structure of the Ir complex shown in
Figure 3. Each Ir has a formal oxidation state of 11, and there is a bridging CO
group to complete a pseudo-octahedral geometry about each metal ion. We are
currently investigating the catalytic activity of such complexes.
Phosphinothiolate ligands display a particularly varied and novel chemistry
with gold.' Reaction of the potentially tridentate proligand PhP(C,H,SH-2),
(H,PS,) with [AuCl,] - gave the square planar AulI1complex [AuCl(PS,)]. The
strong chelation of the PS, ligand inhibits reduction to Aul with its preferred
linear geometry. A second product from the reaction is an unusual dimeric
species containing both Au' and AulI1 with an Au-Au bonding interaction
(see Figure 4). An isomeric form of this complex was also structurally character-
ised, and this differed only in the orientation of the linear S-AuI-S vector
with respect to the planar Aul" unit. The use of the polydentate PS, ligands
( n = 2 or 3) with the Au unlikely to provide coordination numbers
Br Br Br Br PPh2 Br PPh2 PPh2 SH PPh2 PPh2 SH PPh2

II II
S S
Scheme 2
greater than four permits a variety of bridging structures supporting Au-Au

bonding. The structure of one derived by using the potentially tetradentate PS,
system is shown in Figure 5 and comprises two square planar AulI1units held in
close proximity. There is some resemblance here to the well-known A-frame
dimers, but with an additional carbon in the bridge. Gold-gold interactions are
commonplace for Aul but the Aul-AulI1and AulI1-AulI1interactions seen here are
comparatively rare.
It seemed plausible that if the thiolate sulfur were replaced by neutral thioether
sulfur then the latter would be more weakly bound to the metal introducing the
possibility of hemi-labile coordination. We have therefore investigated the chem-
istry of some related phosphinothioether complexes of RulI and Pd". The
new ligand 2-MeSC,H4CH,PPh, (L) was prepared in reasonable yield in a
three stage synthesis starting from 2-bromotoluene by bromination to 2-
Figure 3 The structure of [(lrCl(PSP)),(p-CO)]
Ph
Figure 4 The structure of [Au,(PS,),]
Figure 5 The structure of [Au,(PS,),]

n 0
Figure 6 The X-ray crystal structure of [PdCl,(Ph,PCH,C,H,SMe)]
BrC,H,CH,Br and successive reactions with Na(Ph,P) and LiBu”/MeSSMe.

The proligand L reacted readily with RuCl,.xH,O in refluxing ethanol under N,
to give the yellow complex [RuCl,(L),] in good yield. The X-ray crystal struc-
ture showed the expected pseudo-octahedral geometry with cis chloride ligands.
The catalytic activity of this complex and that of the related 2-Ph2PC,H,SMe
ligand are being investigated.
The complex [PdCl,(L)] was also prepared in a straightforward manner in
good yield from [PdCl,(cod)] (cod = cycloocta-1,5-diene) in dry acetone. The
X-ray crystal structure of this complex is shown in Figure 6. The geometry about
the Pd is essentially square planar with the PdPCCCS six-membered chelate ring
having a boat-type configuration. This complex is one of the most active cata-
lysts yet reported for the Heck coupling reaction of aryl halides with olefins,
achieving turn-over numbers in excess of lo6. The air-stability of the precursor
means that the whole reaction can be carried out aerobically using reagent grade
solvents. The stereochemistry of addition is almost exclusively trans, and un-
usually the system is active for both aryl bromides and iodides, and does not
require the addition of co-catalysts such as silver salts to achieve very high
activities. The exact mechanism remains unclear, but 31PNMR spectroscopy
shows that the precursor complex is essentially intact at the end of the cycle apart
from exchange of the coordinated halide. This strongly suggests a cycle passing
through Pdl* and PdIV,but detailed mechanistic studies remain to be done.
4 Polydentate N,S Donor Ligands

Mention has already been made of the Re complex [Re(tipt),(N,)(PPh,)], and
we wished to see if this trigonal bipyramidal motif could be extended to the
activation of small molecules on other metals. We also wished to move away
from tertiary phosphine ligands and we have embarked on a systematic study of
asymmetric tripodal ligands with N,S,, N,OS and NOS, donor sets with
various backbone lengths. Symmetrical tetradentate N-capped ligands have
been investigated at great length and a recent series of elegant papers has
described the chemistry of vanadium,' molybdenum", and iron' with
(N[CH,CH,S],)3- and the binding of small molecules such as C O to the MNS,
core. The tripodal triamido ligand (N[CH,CH,NSiMe,] 3 ) 3 - has also been
shown to form both the mononuclear complex [Mo(NN,)(N,)]'~ and the in-
t riguing t et r anuclear species [Fe ((N,)M o(NN3)]3]. '
Asymmetric tripodal
ligands represent a far greater synthetic challenge, and have been comparatively
little explored,I5 despite the fact that certain of these can serve as models for the
asymmetric coordination found in metalloenzymes such as nitrile hydratase.
There have however been some reports of MoV1complexes of N,S, tripodal
ligands16 and MoVcomplexes of NOS, l i g a n d ~ . ' ~ ~
The methods used to synthesise the new compounds are summarised in
Schemes 3 and 4 and permit multigram quantities of the proligands to be
prepared with silica gel column purification of intermediates and the final
product. The pyridylmethylaminodithiol C,H,NCH,N(CH,CH,SH), (H,L')
was prepared by a slight modification of the literature method.17 These new
compounds were initially allowed to react with [MoO,(acac),] (acac = pentane-
2,4-dionate) or [WO,Cl,(dme)] (dme = 1,2-dimethoxyethane) in methanol in
the presence of triethylamine or KOH as base to give the yellow complexes
[MO,L] (M = Mo or W, L = L' - L6; see Schemes 3 and 4). The X-ray crystal
structures of two of the complexes (M = Mo, L = L' or L3) have been deter-
mined and a representation of the the first of these is shown in Figure 7. The
overall geometry about the Mo is psuedo-octahedral with the two 0x0-groups
in the usual cis configuration. Despite the asymmetry of the coordination
the bond distances are in the usual range found for MoV1dioxo-complexes.
The five-membered chelate ring systems cause a bending of the NSO
A
MeOH, reflux
R NH, 2.NaBH4
R
HzL2 R = H
no H,L3 R = M e
H,L4 R=BIJ'
Scheme 3
R' R
hydrolysis
R"
Scheme 4
Figure 7 The X-ray crystal structure of a M o v1 complex of an asymmetric N,SO ligand
donor set away from the 0x0-groups which have an 0-Mo-0 angle of close to
108".We are currently investigating the chemistry and detailed electrochemistry
of the dioxo-complexes and also reactions of the polydentate ligand systems with
MI1and MIv precursors. Mvldioxo-complexes of molybdenum and tungsten have
been investigated for their catalytic activity for reactions such as olefin oxida-
tion'* and epoxidation" and the results of our studies of catalysis with our
complexes will be reported elsewhere.
In parallel we have investigated the coordination chemistry of a series of new
tridentate HNNS proligands, the syntheses of which are summarised in Scheme
5. The use of the Me,Si protecting group enables the alkylation of the amine
nitrogen to be carried out in high yield. The di(amine)thiol proligands were
obtained as colourless liquids in yields of around 50% following column
chromatography. They deteriorate rapidly on standing, and are best freshly
prepared. The proligands react readily with [WO,Cl,(dme)] in methanol in the
A Me,SiCI
toluene Et,N, thf
HS
HL'
R-Br hydrolysis
Et,N, thf
Mop3 HS
R = CH2CeH5, CH2C6H4But or CH2C6H3(BUt)2
Scheme 5
presence of base to give the new complexes of the type [WO,Cl(NNS)], isolated
in high yield as yellow solids. These are reactive precursors for a wide range of
chemistry exemplified by the replacement of the chloride ligands by both alkyl
and aryl groups.
5 Thiosemicarbazone Ligands
Thiosemicarbazone ligands have been investigated extensively, the studies being
largely driven by the biological activity of many of their metal complexes.20They
are also of interest as they provide anionic sulfur donors with little tendency to
form thiolate bridged dimers. Despite this there have been surprisingly few
reports of their use in catalysis. A very rare example is the catalysis by a nickel
bisthiosemicarbazone complex of the oxidation of C O to CO, by water.,' We
have therefore recently been exploring the catalytic potential of thiosemicar-
bazone complexes of Ru, Os, Rh and Ir.
5.1 Complexes of Ru, Rh and Ir

We have prepared a range of bidentate thiosemicarbazone compounds of the
type R3,NC(S)NHN=CR'R2 (R1 = R2 = R3 = H, alkyl or aryl). These react
with [RuCl,(PPh,),] in acetonitrile at room temperature to give complexes of
the type [RuL,(PPh,),] (HL = thiosemicarbazone). The X-ray crystal structure
of the derivative for which R1 = H, R2 = Ph and R3 = H is shown in Figure 8. It
reveals an unusual bonding mode for the thiosemicarbazone ligands, each of
which is N,S-bound and monoanionic, but each of the chelate rings contains four
atoms rather than the five normally found. Since this work was done there have
been two other reported examples of this type of bonding, also for RulI.,, It has
Figure 8 The X-ray crystal structure of a Ru" bis(thiosemicarbazonate)complex

showing an unusual four-membered chelate ring
been suggested that the ligand adopts this particular mode of coordination
because of steric interactions of the metal with the thiosemicarbazone pheny1.22
At a stoichiometry of 1: 1 ligand-to-metal in acetonitrile the main products are
the new Ru" mono(thiosemicarbazonate) species [Ru(L)(MeCN),(PPh,),]Cl.
On the basis of 'H and 31PNMR spectra these are assigned a structure with the
MeCN and PPh, groups in a trans configuration. Reaction of the mono-
thiosemicarbazone cation (R3 = H, R2 = Ph, R' = Ph) with one equivalent of
triethylamine as base in MeCN results in deposition of a deep red solid of the
same stoichiometry, [Ru(L)(MeCN)(PPh,),] C1. This was recry st allised from
dichloromethane-methanol and a representation of X-ray crystal structure ap-
pears in Figure 9. This shows that a phenyl substituent of the thiosemicarbazone
has been ortho-metallated to give a planar tridentate ligand. The complex is
interestingly still cationic (the chloride counter-anion is not shown in Figure 9)
and the tridentate ligand still bears a charge of 1 -. It is therefore in the thione
form and, at least formally, a proton has been transferred from the phenyl group
to a nitrogen of the thiosemicarbazonate ligand, which must be strongly basic.
The thione form of bonding is reflected in the C-S bond lengths that are
significantly shorter than those of the complex shown in Figure 8.
Initial investigations of the catalytic activity of the monohydrazone complexes
for hydrosilylation have been encouraging with activities much higher than for
[RuC1,(PPh3),] under the same conditions. Reaction of the same thiosemicar-
bazones with [MCl(CO)(PPh,),] (M = Rh or Ir) gave complexes of the type
[M(L)(CO)(PPh,)] which, on the basis of spectroscopic data, have square-
planar structures with the thiosemicarbazone bonded as an anion. These are the
first reported thiosemicarbazones of Ir, and their chemistry and catalytic activity
will be reported elsewhere.
Figure 9 The X-ray crystal structure of a Ru" complex of a tridentate o-metallated

thiosemicarbazone
228 The Chemistry and Applications of Complexes with Su2fur Ligands
5.2 Copper Bisthiosemicarbazone Complexes as

Hypoxic Selective Agents
There is currently great interest in the development of complexes of radioactive
metals that can be specifically targeted to hypoxic tissue. Cancer cells, particular-
ly those located inside actively growing tumours are hypoxic, due to a lack of
blood supply. Such cells have significantly lower partial pressures of oxygen and
are therefore less oxidising than normal cells. Many types of cells have enzymatic
systems capable of reducing neutral metal complexes to anions, which are then
trapped within the cell by virtue of the acquired negative charge. The tendency
for the complex to be reoxidised to the neutral species, with consequent migra-
tion out of the cell by diffusion, will be significantly lower in hypoxic cells.
Therefore if a neutral complex can be engineered to have a reduction potential
within the range accessible within a cell, and a stable anion, it may well undergo
selective retention in hypoxic tissue.
The copper(1r)bisthiosemicarbazones defined in Figure 10 were initially pre-
pared in the 1960s,but it was some twenty years before their potential as hypoxic
selective agents was recognised through the work of Petering23 and
F ~ j i b a y a s h iThese
. ~ ~ complexes have antitumour activity even with 'cold' cop-
per, although the mechanism is largely unknown.20 Copper has a number of
medically useful radioactive isotopes such as 62Cu and 64Cu that are positron
emitters and there is much interest in the possibility of using these for PET
(positron emission tomography) imaging. Bisthiosemicarbazone complexes of
these isotopes have been used to provide images of blood flow in the brain25and
in tumours.26 Our interest as inorganic chemists has been focussed on the
remarkable dependence of the selectivity on the nature of the substituents on the
ligand backbone. The complex with hydrogens on the C, backbone [Cu(gts)]
shows no selective uptake in hypoxic cells whatsoever, and is irreversibly trapped
in both oxic and hypoxic cells. However, the complexes with one or two methyl
groups ([Cu(ptsm)] and [Cu(atsm)]) show high uptake selectivities for hypoxic
cells. A study of the cyclic voltammetry of a series of CulI bisthiosemicar-
bazones showed that the most hypoxic selective complexes showed the most
negative reduction potentials for formation of the Cu' anion.27However, it was
by no means clear if the redox potential was the principal factor in determining
the selectivity for hypoxic cells. We have undertaken a detailed study of the redox
chemistry of these systems with particular reference to the stability of the Cul
R' HN NHR'
R' = R2 = R3 = H, gts
R2 = R3 = Me, R' = H, atsm
R2 = Me, R3 = H, R' = Me, ptsm
Figure 10 CU" bisthiosemicarbazone complexes that show antitumour activity

anion, using density functional calculations, the pH dependence of cyclic volt-
ammetry and UV-visible spectroscopy.28
We found that for the hypoxic selective atsm and ptsm complexes the reduc-
tion process is reversible (Nernstian) and addition of small amounts of acid has
little effect. By contrast [Cu(gts)] shows a very much less reversible reduction
process which becomes completely irreversible on the addition of even traces of
acid. This suggests strongly that the stability of the Cu' state is crucial in
determining the selectivity and that irreversibility is caused by protic attack on
the anion. But the question of why backbone methylation should increase the
stability of the Cu' state has remained unanswered.
In an attempt to address this we have carried out calculations on the LUMO
and HOMO orbitals of 13 CulI bisthiosemicarbazones using DF methods.28The
complexes contain an unpaired electron and the calculations were done using the
spin-unrestricted formalism. The first striking result was that the HOMO-
LUMO gap was very small (around 0.05 eV) and was strongly substituent
dependent. The overall effect is that for the atsm (hypoxic selective) complex the
LUMO is metal-based whereas for the non-selective gts complex it is ligand-
based. Addition of the electron to the gts system may therefore occur to give a
formally triplet state. Since these ligand systems are planar and highly de-
localised, and addition of an electron to porphyrin complexes is well known to
occur often on the ligand to give radical anions, the formation of an analogous
species in the Cu bisthiosemicarbazones is not unreasonable. We intend to carry
out some low-temperature reduction/EPR experiments to try and obtain experi-
mental evidence for the formation of a radical species.
6 Conclusions
There is some way to go before the chemistry of sulfur ligands rivals in its extent
and applications that of the tertiary phosphines pioneered by Joseph Chatt.
Nevertheless, we hope that this brief survey of our work in the area will serve to
illustrate that sulfur-based ligands can generate unusual and novel chemistry in
terms of structures and reactivities. This project started in the Unit with the remit
of contributing to an understanding of nitrogenase with iron and molybdenum,
but has ultimately led to applications as diverse as highly active catalysts and the
development of metal complexes for imaging and therapy in medicine.
7 References
1 J. Chatt, J. R. Dilworth and G. J. Leigh, J . Chem. SOC.,Dalton Trans., 1973,612.
2 P. J. Blower, J. R. Dilworth, J. Hutchinson, T. Nicholson and J. Zubieta, J . Chem. SOC.,
Dalton Trans, 1985,2639.
3 J. R. Dilworth, J. Hu, J. R. Miller, D. L. Hughes, J. A. Zubieta and Q. Chen, J . Chem.
SOC.,Dalton Trans., 1995,3153.
4 J. R. Dilworth, D. Morales and Y. Zheng, J . Chem. SOC.,Dalton Trans, 2000,3007.
5 R. M. Catala, D. Cruz-Garritz, P. Sosa, P. Terreros, H. Torrens, A. Hills, D. L. Hughes
and R. L. Richards, J . Organomet. Chem., 1989,359,219.
6 P. T. Bishop, J. R. Dilworth, T. Nicholson and J. Zubieta, J . Chem. SOC.,Dalton Trans.,
1991,385.
7 J. R. Dilworth and N. Wheatley, Coord. Chem. Rev., 2000,199,89.
8 J. R. Dilworth, Y. Zheng and D. V. Griffiths, J . Chem. SOC.,Dalton Trans., 1999,1877.
9 K. Ortner, L. Hilditch, Y. Zheng, J. R. Dilworth and U. Abram, Inorg. Chem., 2000,39,
2801.
10 S. C. Davies, D. L. Hughes, Z. Janas, L. Jerzykiewicz, R. L. Richards, J. R. Saunders
and P. Sobota, Chem. Commun., 1997,1261.
11 S. C. Davies, D. L. Hughes, R. L. Richards and J. R. Sanders, J . Chem. Soc., Dalton
Trans., 2000,719.
12 S. C. Davies, D. L. Hughes, R. L. Richards and J. R. Sanders, Chem. Commun., 1998,
2699.
13 M. B. O’Donoghue, N. C. Zanetti, W. M. Davis and R. R. Schrock, J . Am. Chem. SOC.,
1997,119,2753.
14 M. B. O’Donoghue, W. M. Davis, R. R. Schrock and W. M. Reiff, Inorg. Chem., 1999,
38,243.
15 (a) R. L. Fanshawe and A. G. Blackman, Inorg. Chem., 1995, 34, 421; (b) A. M.
Dittler-Klingemann and F. E. Hahn, Inorg. Chem, 1996, 35, 1996; (c) C. Ochs, F. E.
Hahn and R. Frohlich, Eur. J . Chem., 2000,6,2193; (d) C. R. Cornman, T. C. Stauffer
and P. D. Boyle, J . Am. Chem. SOC.,1997,119,5986.
16 J. M. Berg, K. 0.Hodgson, A. E. Bruce, J. L. Corbin, N. Pariyadath and E. I. Stiefel,
Inorg. Chim. Acta, 1984,90,25.
17 M. Mikuriya and T. Kotera, Chem. Lett., 1998,971.
18 K. Sato, M. Aoki, J. Takagi and R. Noyori, Science, 1998, 281, 1646 and references
therein.
19 W. A. Herrmann, J. Fridgen, G. M. Lobmaier and M. Spiegler, Nouv. J . Chim., 1999,
23, 5.
20 D. X. West, A. E. Liberta, S. B. Padhye, R. C. Chikate, P. B. Sonawane, A. S. Kumbhar
and R. G. Yerande, Coord. Chem. Rev., 1993,123,49.
21 Z . Lu, A. White, A. L. Rheingold and R. H. Crabtree, Angew. Chem., Int. Ed. Engl.,
1993,32, 92.
22 (a) F. Basuli, S. M. Peng, and S. Battacharaya, Inorg. Chem., 1997, 36, 5645; (b) F.
Basuli, U. Ruf, C. G. Pierpont and S. Bhattacharya, Inorg. Chem., 1998,37,6113.
23 D. T. Mingel and D. H. Petering, Cancer Research, 1978,38,117.
24 Y. Fujibayashi, H. Taniuchi, K. Wada, Y. Yonekura, J. Konishi and A. Yokoyama,
Ann. Nucl. Med., 1995,9, 1, and references therein.
25 Y. Fujibayashi, K. Wada, H. Taniuchi, Y. Yonekura, J. Konishi and A. Yokoyama,
Biol. Pharm. Bull., 1993, 16, 146.
26 C. J. Mathias, M. J. Welch, D. J. Perry, A. H. McGuire, X. Zhu, J. M. Connett and M.
A. Green, Int. J . Rad. Appl. Instrum. B, 1991,18, 807.
27 J. L. J. Dearling, J. S. Lewis, D. W. McCarthy, M. J. Welch and P. J. Blower, Chem.
Commun., 1998,2531.
28 P. J. Blower, J. R. Dilworth, R. I. Maurer, G. D. Mullen, C. A. Reynolds and Y. Zheng,
J . Inorg. Biochem., 2001,85, 15.
SECTION F:
The Biological Work of the ARC

Unit of Nitrogen Fixation at the
University of Sussex, and Later
Developments
The Unit of Nitrogen Fixation at the University of Sussex became the foremost
laboratory in the world for the study of many aspects of the biological nitrogen
fixation problem. Perhaps the areas most successfully exploited were the chemical
(see Section E ) and the genetic. However, the biological work was considerably
broader than the genetical, and it says much for both Chatt and for his Deputy
Director (and later Director of the Unit) that such a diverse research programme
could be successfully carried out with such success.
I n fact, it was the policy of the ARC to establish a Unitfor the working lifetime of
a particularly distinguished individual, but to wind up the work once that person had
retired. When Chatt retired it was clear that it would have been an act intellectual
vandalism to close the Unit, and the reputation of the Unit and its principal
researchers continued to grow under Postgate’s enlightened direction. Perhaps this
owed much to the old-fashioned British attitude that informed the setting up of the
Unit:find a good set of chaps and let them get on with it. Sadly, this kind of
approach would not be acceptable today.
Included in this section is an account by Postgate of his time in the Unit and
another by Sanders to show how the biological work influenced, and still influences,
some related chemistry. There is also a contribution from Wedd, one of the many
distinguished scholars who passed through the Unit in itsformative days, and who is
happy to acknowledge the injuence it had upon him. The contribution by Lee shows
how the problems originally tackled in the Unit are still giving rise to original and
innovative research, and Garner, a Chatt Lecturer, describes some innovative work
in the area of bioinorganic chemistry, a subject that the combined biological and
chemical work in the Unit did much to stimulate.
Biological Nitrogen Fixation
JOHN POSTGATE
Houndean Lodge, 1 Houndean Rise, Lewes, East Sussex BN7 lEG, UK
Former Director, ARC Unit of Nitrogen Fixation
1 The Unit of Nitrogen Fixation

In 1963 Chatt became the first Director of the Agricultural Research Council’s
Unit of Nitrogen Fixation (UNF) and remained in the post until his retirement in
1980. The U N F comprised a multidisciplinary team of scientists whose research
gained it a formidable reputation. That research involved not only chemistry but
microbiology, biochemistry and molecular genetics, and the biological compo-
nent grew to be the major part of the Unit’s programme.
1.1 Nitrogen Fixation

The fixation of nitrogen is a process of transcendent importance in nature
because it compensates for a net loss of nitrogen from the biosphere to the
atmosphere which takes place during the normal breakdown and mineralisation
of organic matter in soil and water, mediated by certain types of bacteria: a
process called denitrification. Some natural non-biological processes, such as
combustion, lightning and irradiation, lead to modest fixation of nitrogen as
nitrogen oxides, which wash into soil and water as nitrates, but for many
millennia the major input of new nitrogen into the biosphere has taken place by
way of biological nitrogen fixation. During the twentieth century nitrogen
fertiliser manufactured industrially from atmospheric dinitrogen has made a
steadily increasing contribution. The biological process is a property of a hetero-
geneous group of bacteria called nitrogen-fixing bacteria, but of no other living
things. Even in the present era of chemically-produced nitrogenous fertiliser, it
provides over half of the global input of new nitrogen into soil and water.
Nitrogen fixation and denitrification both form part of the planetary cycling of
nitrogen, a complex of processes often formalised as the Nitrogen Cycle, of which
Figure 1 is a representative example.
Thus nitrogen fixation is fundamental to the persistence of the biosphere as we
know it today.lP3 It is also fundamentally important to world agriculture and
234 Biological Nitrogen Fixation
n
-
DlNlTROGEN
0 0
AMMONIF ICATION
Proteins efc.
AMMONIA of microbes,
(NH3) I $ plants and
animals
NITRITES
Figure 1 A version of the nitrogen cycle, a diagram that illustrates the chemical
transformations undergone by nitrogen atoms in the terrestrial biosphere.
(After Reference I)
forestry, and primarily for this reason it has been the subject of intensive research
in agriculturally-orientated Institutes and Universities throughout the world
since the existence of the biological process was finally confirmed in 1886.
In 1960 a group of microbial biochemists in the Central Research Labora-
tories of the Dupont de Nemours Chemical Corporation, USA, broke through a
barrier that had impeded researchers for at least two decades. They extracted
from a species of nitrogen-fixing bacteria a solution containing the enzyme which
is responsible for the activation of d i n i t r ~ g e nIt. ~was called nitrogenase, and, in
appropriate conditions, it bound N, from the atmosphere and reduced it to
ammonia.
Their extract was a highly impure proteinaceous solution obtained by disrupt-
ing the cells of a common soil bacterium, Clostridium pasteurianum. This organ-
ism belonged to the class of microbes called anaerobes: it would not grow in the
presence of oxygen. Two features proved to have delayed earlier attempts to
extract the enzyme: its nitrogen-fixing activity was irreversibly inactivated by
exposure to dioxygen, and it functioned only when provided with pyruvate as the
reducing agent. Within a couple of years the pyruvate was shown to have a dual
function: it acted both as a reductant and as a source of ATP, both of which were
essential for the biological conversion of N, into NH,. The experimental
achievement of the Dupont group lay largely in devising ways of extracting the
material and handling it anoxically, but it opened up to biochemists the possibil-
ity of purifying nitrogenase and discovering how it bound and reduced the
dinitrogen molecule.
Biological nitrogen fixation presented a chemical enigma. The two major
industrial procedures employed to cause atmospheric dinitrogen to enter chemi-
cal combination, like the few available laboratory methods, required drastic
conditions. The obsolete Birkeland-Eyde process required exposure of dinitro-
John Postgate 235
gen and dioxygen to an electric arc at some 2000 to 3000°C, followed by
exceptionally rapid cooling and scrubbing of the nitrogen oxides so formed; and
the widely-used Haber process required dinitrogen plus dihydrogen, catalysts,
high pressures and temperatures, and the total absence of dioxygen and water.
Yet nitrogen-fixing bacteria in soil brought dinitrogen into chemical combina-
tion at ordinary temperatures and pressures, in environments where both water
and air were abundant: indeed, the genus Azotobacter would only grow and fix
nitrogen in air.
One of the industrial processes used to fix nitrogen was oxidative, the other
reductive. Both had provided models for speculation on the mechanism of
biological nitrogen fixation, but Dupont’s advance settled the matter: because the
product was ammonia and because the least trace of dioxygen stopped the
biological reaction, oxidative pathways needed no longer to be considered
seriously as far as the biological process was concerned. Nevertheless, the details of
how the enzyme actually bound and reduced dinitrogen were still a closed book.
1.2 The Unit’s Origins

The historical circumstances surrounding the UNF’s formation have been de-
scribed in some detail elsewhere5 and will be repeated here only in outline.
One of Dupont’s commercial rivals, the multi-national petrochemical com-
pany Shell, was intrigued by the advance. Why was Dupont devoting a high
quality research effort to so fundamental a scientific topic? To discover whether
there was a possible industrial application that could be of interest, Shell com-
missioned the British microbiologist K. R. Butlin6 to look into and report upon
the state of the art in nitrogen fixation research, with special attention to
chemical and biochemical aspects.
Butlin visited the four major centres of research on nitrogen fixation then
operating in Britain, the USA’s major research centre at the University of
Wisconsin, Madison, and finally Dupont’s Central Research Laboratories in
Wilmington, Delaware, USA. There he talked freely with Dr Carnahan, the
senior member of Dupont’s team. Carnahan explained that it was good for the
Corporation’s public image to be seen to have advanced fundamental knowledge
of a process so crucially important in agriculture; they had also hoped to learn
enough about the enzyme to be able to imitate its action commercially, to patent
procedures for making agricultural fertiliser more cheaply, or perhaps to develop
new methods of making valuable nitrogen compounds directly from the atmos-
phere. However, they had already concluded that their discovery was unlikely to
affect the fertiliser or fine chemicals industry in the short or medium term, and
were planning to move closer to agriculture, into research on nitrogen-fixing
bacteria that associate with plants.
Butlin did not agree; in his report he considered that a serious possibility still
remained that understanding how nitrogenase worked would enable industrial
chemists to mimic the process, or develop new processes, and he recommended
Shell to make a substantial investment in biochemical research in this area.
However, the Company also decided that the economic prospects of such
research were too remote to be of commercial interest, but considered that
further research on the subject in a more open laboratory than Dupont’s would
be to their and everyone else’s advantage. So they passed Butlin’s report to the
then Secretary of the Agricultural Research Council, Professor E. G. (later Sir
Gordon) Cox FRS.
Cox, himself an inorganic chemist, was well aware of the perplexing chemical
questions presented by the biological fixation of nitrogen. He also recognised
that there was no centre in Britain with the combined biological and chemical
expertise needed to pick up and extend the Dupont group’s findings. So he urged
the Council to set up a dedicated research Unit to study the problem. He wanted
a research effort that would focus on the basic chemistry; he took the view that
research on plants and/or other steps of the nitrogen cycle would be distracting.
There had long been evidence that molybdenum was involved in some essential
way in nitrogen fixation, perhaps in the actual catalytic process conducted by
nitrogenase. It so happened that Joseph Chatt, already an FRS and distin-
guished for his research in metallo-organic complexes, had left Imperial Chemi-
cal Industries’ Frythe Laboratory in 1962 and was considering posts in the USA.
Chatt agreed to become Director of the Council’s new Unit.
The Unit’s remit would be to study the fundamental chemistry of biological
nitrogen fixation. It would be multidisciplinary, involving microbiologists and
biochemists as well as pure chemists, and would therefore be large. In the jargon
of the day, its research would be strategic: it would study the basic science
underlying a process of great practical importance. Its programme would defi-
nitely not be practically orientated: it would not involve itself in other steps of the
nitrogen cycle, nor with associations with plants, nor yet with immediate prob-
lems in agriculture. Any useful fall-out of the new Unit’s findings would be taken
care of by other stations, university groups or research agencies.
In retrospect it was remarkably enlightened of the Council to approve and
fund so basic a research project. In 1962-63 long-term strategic research was still
the primary remit of the Research Councils, but the political climate was chang-
ing and pressure was already building up to shift their research programmes
towards foreseeable practical objectives. Within a few years a fundamental
initiative on that scale by a British Government research agency would become
effectively impossible.
1.3 Earliest Days of the UNF

In 1963 the present writer was seeking a new research post. I had taken an
honours degree in chemistry at Oxford before turning to microbiology, and had
worked for nearly a decade in Butlin’s group at Teddington, which specialised in
Economic Microbiology, then briefly at the Microbiological Research Establish-
ment on Salisbury Plain. With Butlin I had become a leading authority on
sulfate-reducing bacteria, which play an important part in the planetary Sulfur
Cycle as well as being of substantial economic i m p ~ r t a n c eCox
. ~ duly arranged
for an interview with Chatt. I take up the story with an excerpt from a tribute I
spoke at Chatt’s funeral.
John Postgate 237
‘So Joseph and I were put in touch and, one chilly morning in the Spring of
that year, we met, in a distinctly austere room in the Abbey Hotel in New York
City. Why there? Because we were both travelling scientists at the time; I was
returning to the UK after a six-month visiting Professorship at the University of
Illinois, Joseph was en route to a short sabbatical at Pennsylvania State Univer-
sity.
In the space of some 40 minutes he explained to me his plans for the new Unit.
It would have ten research staff, half of them microbiologists - as he described
them - with appropriate support staff; graduate students and post-docs would
pass through. New laboratories would be built for it, but where was not yet clear;
finding temporary lab. accommodation would be a priority on his return to the
UK. I was impressed by his enthusiasm and shared his view of the need for basic
research in that area, but it all seemed a little tentative.
I do not recall much more about the interview. It was low key and we parted
thoughtfully.’
That last sentence glosses over the fact that we were each dubious about the
other. I was unimpressed by Chatt’s seemingly dismissive view of biological
science (a not uncommon attitude among chemists in those days), and I learned
much later that Chatt found me somewhat ‘bohemian’. However, the meeting
proved to be the beginning of nearly twenty years of very productive research
collaboration. For despite my lack of experience of nitrogen-fixing bacteria, Cox
and Chatt deemed my qualifications suitable - in the eyes of chemists someone
who could handle sulfur ought to be able to cope with nitrogen; they are not far
apart in the Periodic Table.’ In that year I was appointed Assistant Director of
the UNF-to-be, charged with planning and directing the biological side of the
research. That autumn Chatt and I moved into the Agricultural Research
Council’s crowded headquarters in London, with a desk in the waiting room for
me, and we set about planning a new laboratory, visiting possible sites, talking to
architects, picking up ideas and, above all, telephoning fellow scientists for
temporary laboratory space.
I discovered Chatt’s remarkable tenacity and firmness of purpose, once he had
decided what he wanted. I think the Council’s planning staff was disconcerted by
the juggernaut that had moved in among them, but they were admiring, too.
Within a couple of months he had found temporary laboratory space for
chemistry at Queen Mary College, and I had negotiated a conversion at the
Royal Veterinary College, where I could get the microbiology started. We moved
out of the Agricultural Research Council’s H Q building in 1964, to their polite
relief, and we worked separately for about a year, taking on a few staff. Finding a
permanent home for the Unit was something of a problem; for some months
Queen Mary College was the most probable host, but diverse practical consider-
ations led to the University of Sussex being preferred and there the biological
and chemical sides of the Unit came together, in temporary accommodation,
early in 1965. Its own purpose-built laboratories at Sussex became available in
1967.
The Unit moved gradually into the forefront of its research area and gained a
formidable reputation both nationally and internationally.’ It became a para-
digm of successful interdisciplinary research: chemists, biochemists, physiol-

ogists and geneticists homing in on a central problem. Many of its senior staff
became world leaders in their special research areas, while sustaining the prin-
ciple of collaboration, both internal and external. Visiting workers came from all
parts of the world and unanimously praised its collaborative atmosphere and
effectiveness.By the time Chatt retired in 1980 it had grown from its intended 24
to some 80, including visiting workers. It forged links with the University,
especially with its School of Molecular Sciences; its publications ranged from
abstruse chemical kinetics to advanced molecular genetics - several opened up
wholly new and exciting research areas.
As far as the biological side of the UNF’s research is concerned, I should put
on record a fact for which I am personally very grateful. Chatt planned and
directed the chemical side of the UNF’s research in considerable detail, but he
left the planning and the directing of the biological programme entirely to me. A
dichotomy between chemistry and biology was necessary at its outset, because
the Unit’s biological and chemical research thrusts were remote from each other
- though naturally they shared the ultimate objective of understanding the mode
of activation of the dinitrogen molecule by nitrogenase. Chatt and I were
confident that the two prongs of the Unit’s approach would in time converge
and, with this end in mind, the Unit held regular internal seminars at which
chemists and biologists would present, explain, and if appropriate justify, their
research to each other. These meetings brought out early problems in communi-
cation: just as biologists would be bewildered by chemical jargon, such as talk of
hard versus soft ligands, n-bonding, and charge states in metal complexes, so
chemists would be bothered by casual use of biological acronyms such as ATP,
ADP, NAD, not to mention that ghastly piece of biochemical jargon, ‘reverse
electron flow’. The experience was salutary for the linguistic discipline of both
sides, and we gradually learned each other’s languages and discarded imprecise
terminology.
The UNF’s biological research catalysed many important discoveries in other
laboratories, as well as prompting re-thinks in general enzymology, microbial
physiology, genetics and evolutionary theory. In the space available I can only
touch upon a few highlights without detailed citations (see Section 6).
2 Biochemical Research
By the time the U N F had settled into its laboratories at the University of Sussex
the Dupont group’s break-through had been followed up and extended in several
laboratories. In particular, the Charles Kettering Laboratory in the USA had
extracted a particulate nitrogenase (i.e. one which sedimented in a regular
high-speed centrifuge) from the aerobic nitrogen-fixing bacterium Azotobacter
vinelandii. This had three valuable properties for research. Firstly, the particulate
preparation was stable in air, provided it had no ATP and was therefore not
functioning; secondly, in vitro it would function (anoxically and given ATP, of
course) using sodium dithionite as an artificial reducing agent, to form ammonia
from dinitrogen, plus gaseous dihydrogen (its activity with pyruvate was only
John Postgate 239
marginal); thirdly, in the absence of dinitrogen ( i e . under argon or helium) it
generated exclusively dihydrogen. Manometric comparisons of the amounts of
dihydrogen evolved under dinitrogen compared with argon were widely used in
the mid 1950s as a quick and reliable measure of nitrogen fixation. Such findings
opened up a route to rigid enzymological studies on nitrogenase.
2.1 Structural Analogues of Dinitrogen

In the 1960s the use of structural analogues of substrates to probe the kinetics of
enzyme action had been commonplace for about two decades; sometimes they
would be inhibitors of enzyme activity, competitive or non-competitive as the
case might be, and sometimes they would be alternative substrates. Dinitrogen
monoxide, azides and cyanides had been tested with intact nitrogen-fixing
microbes and enzyme preparations and proved to be substrates alternative to
dinitrogen. Carbon monoxide was not: it inhibited nitrogen fixation but the
enzyme continued to form dihydrogen. Was a terminal nitrogen atom obligatory
for interaction with the dinitrogen-binding site? One of the earliest collabor-
ations between biologists and chemists within the U N F was the testing of methyl
isocyanide, prepared by the chemists for the biologists, with our own nitrogenase
preparation from Azotobacter chroococcum. It proved to be an alternative sub-
strate that competed with N,. It was reduced, not to dimethylamine but to
methylamine plus methane (plus small amounts of C, products whose origin was
the subject of now obsolete speculation); experiments in D,O showed that the
methane arose from the terminal C of the isocyanide, so the triple N-C bond had
been split.' Chatt was intrigued by these findings, since they supported specula-
tion that the substrate-binding site was a transition metal. He proposed a
long-shot experiment: would xenon interfere with the N,-binding site? A small
but expensive cylinder of xenon was procured and an economical experiment
designed, but no influence on N, reduction was detected.
Methyl isocyanide was reduced by living nitrogen-fixing bacteria, and this
finding suggested that the reaction could be used as a versatile and quick assay
for nitrogenase, since the methane produced could be quantitated readily by gas
chromatography.' Indeed, it worked well with live bacteria or with nodules from
nitrogen-fixing plant associations, as well as with enzyme preparations. How-
ever, fortunately for the progress of research in the world at large, because
isocyanides are exceptionally unpleasant to work with, a more agreeable and
equally versatile assay was developed almost simultaneously, jointly in Wiscon-
sin and Australia, using acetylene, which nitrogenase also reduces, to yield
ethylene."*" The 'acetylene test' for nitrogen fixation is still widely used in the
field as well as in laboratories.
2.2 Enzyme Studies

By 1965-66 research in the USA had shown that the nitrogenase in Dupont's
original extract from Clostridium pasteurianum, which was a true solution, was
composed of two distinct metalloproteins, both essential for activity, both irre-
versibly destroyed by exposure to dioxygen. The process needed magnesium ions
as well as ATP, and the ATP was converted to ADP, which inhibited the reaction
- in experimental work it had to be removed, or rather, recycled. The relatively
dioxygen-tolerant but particulate nitrogenase preparation from Azotobacter

vinelandii also consumed ATP and required Mg2+;when disrupted it yielded two
highly dioxygen-sensitive proteins very similar to those in the clostridial extract.
Procedures for purifying the two nitrogenase proteins by anoxic ion-exchange
chromatography were being developed. Facilities for the culture and harvesting
of bacteria on a large scale were essential for work of this kind, and the U N F
designed and developed its own all-glass culture facilities which served it satisfac-
torily for some twenty-five years.’
In the USA, the anaerobe C. pasteurianum and the aerobe A. vinelandii were
the favoured bacteria for biochemical research on nitrogen fixation. In order that
the UNF’s research might be complementary to the research of others rather
than competitive, it seemed logical to adopt different ‘work horses’. I chose
Azotobacter chroococcurn as our aerobe. Preliminary tests with an anaerobe with
which I was familiar from earlier years, Desulfovibrio desulfuricans, showed it to
be unsuitable because it was difficult to culture routinely in large amounts
(Desulfouibrio remained a secondary research subject for many years) and the
more amenable Klebsiella pneumoniae became our ‘pet’ anaerobe. K . pneumoniae
was not a true anaerobe; given fixed nitrogen as an ammonium salt it grew and
multiplied vigorously in air, but it would fix nitrogen only in the absence of
dioxygen. (The species name pneumoniae caused some consternation among our
non-microbiologist colleagues until they were assured that ours was a harmless
strain originally isolated from fermenting corn liquor.)
Nitrogenases were extracted from both microbes and proved also to be binary.
We purified the two nitrogenase proteins from each organism: they were very
One was a large (MW -

similar to each other and to analogous proteins being reported from the USA.
220000) a,P2 tetramer, and contained molybdenum
plus much iron, the latter as iron-sulfur clusters; the other was y2 dimeric and
contained only an iron-sulfur cluster. The former was termed the MoFe-protein
and the latter the Fe-protein. They were duly characterised and detailed data
published; comparable nitrogenase proteins from other genera and species of
bacteria were purified and characterised, some by visiting workers who had
come for that purpose, some by graduate students. They were very similar, and
the similarities of nitrogenases from diverse bacteria extended to function: one
could, for example, construct an active hybrid nitrogenase by mixing the larger
MoFe-protein from Klebsiella with the smaller Fe-protein from Azotobacter.
2.3 Studies of Mechanism

By the early 1970s it had become possible to ask mechanistic questions. Did the
two proteins act sequentially or as a complex? Where and in what order did N,
and ATP bind? How was Mg involved? Were identifiable intermediates formed
between dinitrogen and the end-product ammonia? Why was H, always evol-
ved?
John Postgate 24 1
The UV-visible spectra of the MoFe- and Fe-proteins lacked exploitable
features, but the behaviour of their electron paramagnetic resonance and Moss-
bauer spectra proved to be informative. For details of these studies, which
occupied several years of research in various laboratories, specialised texts
should be consulted. In outline, the Fe/S cluster in biologically active (i.e. not
dioxygen-damaged) Fe-protein could exist in a reduced or an oxidised state. In
the reduced protein it displayed a rhombic EPR signal which, in the presence of
ATP and Mg ions, changed to axial, a change accompanied by a lowering of the
protein’s standard redox potential and changes in other physico-chemical prop-
erties. Native MoFe-protein displayed three redox states, observable in its
Mossbauer spectra. It was normally isolated in the intermediate state, which also
displayed a unique EPR signal (the other two states were EPR-silent). The most
reduced, EPR-silent state appeared when the enzyme was actually functioning;
the most oxidised state was probably physiologically irrelevant. Tests with
substrate analogues gave hints that these influenced the spectra of the MoFe-
protein. The upshot of experiments on these lines was that the MoFe-protein
included the N,-binding site, and that the Fe-protein bound and hydrolysed
ATP, donating electrons to N, by way of FeS clusters in the M0Fe-pr0tein.l~
This view became accepted and a then contemporary suggestion from the USA
that the Fe-protein bound dinitrogen was abandoned.
Conventional quantitative enzymology confirmed that dihydrogen evolution
was an intrinsic part of biological dinitrogen reduction, and that the overall
stoichiometry should be written as shown.
N, + 16MgATP + 8H,O + 8e- +2NH, + 16MgADP + H, + 8 0 H -
This involves a net transfer of eight electrons and the consumption of an
unexpectedly large amount of ATP. This could hardly be a one-step reaction,
and the questions regarding pathway and the existence of a nitrogenase complex
remained.
High velocity sedimentation studies with mixtures of the separated proteins
from Klebsiella, albeit unsatisfactory because of their sensitivity to air, suggested
that the proteins associated readily, and steady-state kinetics suggested a
stoichiometry of two Fe-protein molecules to one of MoFe-protein. For several
years the UNF’s consensus favoured a real rather than transient nitrogenase
complex. However, proteins often associate in nitro and more convincing evi-
dence bearing upon complex formation emerged from detailed kinetic studies on
nitrogenase function.
Details of several years of U N F research exploiting rapid reaction, rapid
quench and stopped flow kinetics are in the scientific literature and would be
inappropriate here.I4 The upshot was a scheme which evolved over the period
and, with later refinements, is now considered to be correct in all its major
features. In brief, the obligatory evolution of dihydrogen is believed to arise from
the formation of a metal-dihydride complex within the reduced MoFe-protein, a
two-step process requiring the transfer of two electrons successively from the
MgATP-activated Fe-protein coupled with two protonation steps and the hy-
drolyses of four molecules of ATP to ADP; N, is bound by displacement of two
hydrido groups as H, (precedents for this reaction exist in coordination chemis-

try); subsequent reduction of bound N, takes place by further sequential electron
transfers from MgATP-activated Fe-protein molecules, with concomitant proto-
nations and hydrolyses of another 12 molecules of MgATP; short-lived bound
forms of N, in intermediate states of reduction precede the release of bound N as
NH,. Rate constants for many of the steps were estimated by computer
modelling of the experimental data; they compelled the conclusion that no
long-lived complex was formed; rather that a molecule of MoFe-protein charged
with dinitrogen reacted with successive molecules of MgATP-activated Fe-
protein. The rate of dissociation of the transient protein-protein associations
determined the turnover time of the complete reaction. Following suggestions
from the USA, the two nitrogenase proteins are now officially designated ‘dinit-
rogenase’ (the MoFe-protein), since it binds N,, and ‘dinitrogenase reductase’
(the Fe-protein), since it donates electrons to dinitrogenase.
Parallel with this research, Chatt and his team were deeply involved in their
exhaustive study of the chemistry of the dinitrogen complexes formed by various
transition metals. In particular, their successful search for complexes in which the
dinitrogen group could be protonated in mild conditions to yield ammonia (see
the contribution by R. L. Richards in Section E) indicated that hydrazine was
often formed instead of or alongside ammonia. It was logical for the UNF’s
biochemists to seek hydrazine as a biological intermediate, despite the fact that
earlier isotope dilution experiments in the USA using I5N-hydrazine had in-
dicated none. Rapid quenching of nitrogenase, with acid or alkali, in the first few
seconds of its functioning did indeed yield hydrazine, in amounts which peaked
in the pre-steady state and became virtually undetectable in the steady state.
Hence a bound species at the oxidation level of hydrazine was probably formed
during nitrogenase function, which, being bound, did not exchange with isotopi-
cally labelled hydrazine.15 By the later 1970s the Unit’s chemistry and biochem-
istry were converging.
Everything pointed to the N,-binding site being a metal atom in the MoFe-
protein, probably Mo itself, but it remained elusive. Evidence that the site had to
be sterically close to the Mo atom was in due course obtained using mutant
bacteria (see Section 5, below).
3 Physiological Research
As soon as the enzymology of nitrogenase began to be resolved, no less that four
obvious physiological questions arose. First, given their sensitivity to dioxygen,
how did nitrogen-fixing bacteria protect the two nitrogenase proteins from
denaturation in air? Secondly, how were electrons channelled from the cells’
metabolism into the reduction of dinitrogen? Thirdly, how did the bacteria cope
with the enzyme’s substantial demand for ATP? Fourthly, nitrogenase activity
was always accompanied by dihydrogen evolution; did this happen in nature? If
not, why not?16
John Postgate 243
3.1 Dioxygen Exclusion

The first question arose with particular force from studies on bacteria belonging
to the genus Azotobacter. These are very efficient and reliable nitrogen fixers yet
they require dioxygen for growth; they neither multiply nor fix dinitrogen
without it. The UNF approached the problem using continuous culture, which is
a powerful tool for studying physiological questions in microbiology. Imagine
a microbial culture - a vessel in which a population of bacteria is multiplying -
which is filled to the brim and fed continuously with fresh culture medium, the
excess of medium being allowed to overflow. Imagine now that the medium
contains an excess of everything the bacteria need to multiply but for one
component, glucose, for example. The bacterial population in the vessel multi-
plies at a rate determined by how fast fresh medium is supplied and the popula-
tion density is determined by the concentration of glucose in the influent me-
dium. The bacteria comprising the population are said to be ‘glucose-limited’.
Once a population is in a steady state its density, multiplication rate and
nutritional status remain constant for as long as the experimenter wishes and
these parameters, as well as aeration and temperature, can be altered indepen-
dently and the way the properties of the bacteria change can be examined.
(Parenthetically, much of the development work on continuous culture during
the 1950s and 1960s, in which I had some involvement, had been carried out at
the Microbiological Research Establishment, Porton Down, whose remit in-
cluded work on defence against biological warfare. This information reached the
University’s student body during its militant 1960s, and the UNF was briefly
picketed, suspected of secret bellicose research. Happily one of the editors of the
present volume (GJL), more patient than either Chatt or me, explained to the
demonstration’s leaders the importance of nitrogen fixation to world nutrition,
and it dispersed peacefully.)
Our continuous culture studies on Azotobacter chroococcum revealed that
nitrogen-fixing, glucose-limited populations were unexpectedly sensitive to oxy-
genation: if aerated vigorously they ceased multiplying and fixing nitrogen.
However, populations supplied with plenty of glucose were far from sensitive:
they flourished not only in air but even in hyperbaric dioxygen. It transpired that
they adjusted their respiration rate in proportion to the concentration of dissol-
ved dioxygen; under high oxygenation the bacteria evinced the fastest respiration
rates recorded for living things. But that adjustment took a little time; if they
were abruptly exposed to a dioxygen stress, they would ‘switch off’ their nitrogen
fixation and restore it rapidly when the stress was relieved (Figure 2).
Several years’ research along these lines led to the conclusion that respiration,
in addition to its normal function of generating energy, was exerting a protective
action on nitrogenase simply by scavenging dioxygen, and that, if the dioxygen
stress was greater than respiration could cope with immediately, the organism
was able to protect the enzyme from oxygen damage, though it became non-
functional in the protected state. We introduced the terms ‘respiratory protec-
tion’ and ‘conformational protection’ for these two processes. The latter term
became obsolete in the 1980s when workers in the Netherlands demonstrated its
10 20 30 40 50
Time (minutes)
Figure 2 ‘Switch on’ and ‘switchof of nitrogenase in response to aeration levels. Live
cells from a continuous culture of Azotobacter chroococcum were shaken
gently in air while their nitrogen-jixing activity was measured using the
acetylene test. Shaking, and therefore aeration, was intensijied at A and
returned to its original value at B. Notice the brief delay in reaching maximum
activity again. ( A f e r Reference I ) .
mechanism: the bacteria possess a non-haem iron protein which, under dioxygen
stress, forms a protected but enzymically inactive complex with the nitrogenase
proteins; the complex dissociates as the dissolved dioxygen tension approaches
zero.’ However, the essential incompatibility of nitrogen fixation and dioxygen,
which we had thus rationalised, influenced thinking in many laboratories and a
variety of stratagems for evading oxygen damage were discovered among nitro-
gen-fixing systems, including respiration, compartmentation, ‘switch on and off’
processes, cooperative screening from dioxygen, and simple evasion. In addition,
the number of recognised species of bacteria known to be able to fix nitrogen was
multiplied several-fold over the next decade with the realisation that many were
‘microaerobic’ fixers: they were not good at respiratory protection and only
functioned at low partial pressures of dioxygen. Dioxygen exclusion is now
accepted as being fundamental to the physiology of both symbiotic and free-
living nitrogen fixation.16
John Postgate 245
3.2 The Biological Reductant(s)

The standard redox potential of the Fe-protein is low, and it becomes even lower
when it interacts with MgATP. A specific low-potential electron donor channels
electrons from the cell’s general metabolism to nitrogenase. In the anaerobe
Clostridiumpasteurianum it was known to be a ferredoxin; the UNF’s particular
contribution in this area was to show, in parallel with workers in the Nether-
lands, that the equivalent electron donor in A . chroococcum was a flavodoxin,
shuttling between its reduced and semi-quinone forms.
3.3 The Demand for ATP

Nitrogenase consumes 16 molecules of ATP to fix one N, molecule, and more is
consumed in synthesising and maintaining the enzyme itself. This is a significant
drain on the cell’s metabolic resources, and most nitrogen-fixing bacteria do not
make nitrogenase if fixed nitrogen is available to them. In a continuous culture
limited by a source of both carbon and energy, such as glucose, the population
density is in effect being limited by the amounts of ATP the cells can divert into
biosynthesis. Any metabolic process with a high ATP demand will thus lower the
population density. By comparing steady-state glucose-limited populations of K .
pneumoniae with and without fixed nitrogen (as ammonium ion), research in the
U N F quantitated the energy cost of nitrogen fixation, quantitated the effect of
ammonium ion as a repressor of nitrogenase synthesis, established quantitatively
that dioxygen also repressed nitrogenase synthesis and, unexpectedly, demon-
strated that K . pneumoniae could be coaxed into fixing nitrogen in an oxic
atmosphere provided its respiratory activity kept the partial pressure of dissol-
ved dioxygen vanishingly low: a marginal example of respiratory protection.
Details of the mechanisms of repression of nitrogenase synthesis by dioxygen
or fixed nitrogen involved genetical research and are outined later in this article.
3.4 Dihydrogen Recycling

Bacteria such as K . pneumoniae and C. pasteurianum growing without air evolve
dihydrogen as a normal product of their anaerobic metabolism whether they are
fixing nitrogen or not, so the contribution made as a by-product of nitrogenase
function is not easily assessed. However, azotobacters such as A . chroococcum
evolve no dihydrogen, and continuous culture experiments analogous to those
outlined in Section 3.1 showed that they were remarkably efficient in terms of
substrate consumption, despite their need to sustain respiratory protection of
nitrogenase. We discovered that acetylene caused the organisms to evolve dihyd-
rogen; it acted by blocking an ‘uptake hydrogenase’, a type of enzyme which
oxidised dihydrogen to water as it was formed. This enzyme enhanced respir-
atory protection and also provided energy to regenerate ATP; it significantly
enhanced the efficiency of nitrogen fixation by these bacteria. These findings
were consistent with concurrent research in the USA on hydrogen recycling in
the symbiotic nitrogen-fixing bacteria Bradyhizobiumjaponicum.’
Genetical Research
The U N F was set up in order to study a problem in fundamental biological
chemistry. In 1968 nothing whatever was known about the genetics of nitro-
genase synthesis and it seemed to me that their study might open up a new
approach. I privily undertook a few experiments myself, and rapidly realised that
serious progress would only be made by someone with proper training. No staff
position was available but I had funds to recruit a graduate student willing to try
to open up the subject and, after an exacting period when we nearly gave up, he
had success. Coincidentally and unknown to me, similar thoughts had occurred
to Professor R. C. Valentine at the University of California, Davis, and he, too,
could finance only a graduate student - who also had success. Again by coinci-
dence, both of us were working with K . pneumoniae. Twenty years later the
genetics of nitrogen fixation had become a major topic in molecular biology,
demanding regular international conferences, and was occupying well over 300
research scientists in laboratories all over the world. It was all started by two
graduate students.
4.1 The nifGenes

Earlier work in the U N F had shown that laboratory cultures of many kinds of
bacteria could simulate nitrogen fixation by scavenging nitrogenous impurities
from the atmosphere, from dust, or from components of culture media. Many
scientists had been misled in earlier decades and as a result of our report (which
editors were at first reluctant to accept because of its negative character) numer-
ous putative nitrogen-fixing organisms were dismissed from the literature, in-
cluding all those that were not bacteria.20 But the examination of each isolate
was laborious, and the advance of genetics depended absolutely on devising a
quick, simple and reliable way of screening hundreds of bacterial colonies for
fixation. Both students, in California and in Sussex, solved the problem and both
prepared a few mutants of K . pneumoniae that were unable to fix nitrogen.
Standard genetical techniques (transformation in the USA, plasmid-mediated
conjugation in the UNF) were used to correct the mutations with DNA from the
parent organism; prompted by the US work, we confirmed that the mutations
lay close to genes coding for histidine biosynthesis in the bacterial chromo-
some.21 , 2 2 The genetical shorthand nif was adopted for nitrogen fixation genes
(and will be used henceforth here). The UNF's major break-through in this area,
made in 1971, resulted from the manipulation of the nifgenes of K . pneumoniae
from the bacterial chromosome to a genetical element called a plasmid: a circle of
DNA which carries genes which are not on the chromosome. The plasmid we
used belonged to a class capable of transferring itself from one cell to another
and, within limits, to other genera and species of bacteria. With it we were able to
transfer the ability to fix nitrogen to Escherischia coli, a species of bacteria whose
genetics was already substantially elucidated. In one of our hybrids, the nifgenes
had become incorporated into the E . coli chromosome: we had for the first time
created an entirely new species of nitrogen-fixing bacteria. The exciting possibil-
John Postgate 247
ity thus arose that nifgenes might be transferable to all sorts of creatures; the
economic ramifications of this possibility, including that of creating self-fertilis-
ing plants for agriculture, caused something of a sensation and brought the U N F
a brief spell of media attention.
The episode caught the Agricultural Research Council’s administration nap-
ping. Chatt had realised that biological nitrogen fixation was a much more
complex problem than we had suspected at the time of the UNF’s formation and
had accepted that genetics ought to form part of the biologists’ approach. He had
therefore sought an increase in staff, mentioning, among other evidence of good
progress, the then unpublished transfer of nifto E. coli. But the Administration
was unimpressed; his request was refused out of hand: no such increase was
financially possible. However, once the press furore got under way, Chatt moved
quickly; he arranged an interview in London with the Council itself, and they
were sufficiently impressed to increase the UNF’s senior staff by a third - from 10
to 15 - four of the new posts to reinforce the biological side of the programme,
one to expand the chemical side. Thus his U N F became the largest autonomous
research Unit in the Agricultural Research Council.
As far as research was concerned, the fundamental value of our gene transfer to
E. coli was that a huge repertoire of well-defined mutants throughout the E. coli
chromosome already existed and were available as backgrounds in which to
study nif expression, and procedures were well established for mapping and
characterising genes in E. coli. An additional bonus appeared within a year or so:
in 1973 two scientists in the USA developed techniques for cutting up and
manipulating DNA in vitro and inserting it into E. coli. The UNF’s geneticists
embarked upon a concerted programme of mapping the nifgenes of K . pneu-
moniae using both traditional methods and new-style genetic manipulation and,
where possible, establishing their functions. We were disconcerted when, in 1975,
US scientists sparked a widespread panic about the dangers of genetic manipula-
tion in vitro, with some bizarre political consequence^,^^ but the agricultural
potential of the UNF’s use of the technique happily saved us from censure in the
media.
We had originally expected klebsiella’s nifto comprise about five genes; by the
time Chatt retired we, along with a few other laboratories throughout the world,
had discovered some fifteen (Table l),all linked together as a chain of clusters of
genes. (The final number, established in the late 1980s, when the complete DNA
sequence of nifwas published, was twenty.) A much improved nifplasmid was
constructed, able to transfer itself to many genera of bacteria as well as being far
more stable than the earliest n i f p l a ~ m i d Named
.~~ pRD1 (earlier RP41), we
made it available to laboratories throughout the world and it became the basic
tool of most research on nifgenetics. With it, several new species and genera of
nitrogen-fixing bacteria were constructed in order to learn about the conditions
and genetic backgrounds in which nifcould be expressed. Many clones of DNA
were prepared from it, some carrying all the nifgenes at once, others carrying the
clusters within nif, or carrying each of the known genes separately, and yet others
with fragments of the genes, including their regulatory sites. Plasmids with
mutated nifgenes were also made.
Table 1 The nif gene cluster of Klebsiella pneumoniae as it was known about
1980
The individual genes clustered within nifwere assigned capital letters (agreed among the
research laboratories concerned); future research would reveal another five, and the
nature and functions of most nifgene products would be established. The vertical
arrows indicate subclusters of genes (called operons) which had been shown to be
transcribed in concert, and the directions of their transcription.
i hisD this gene, involved in histidine biosynthesis, is attached to one end of the
I
I nifcluster and plays no part in niffunction.
-----------
nifJ codes for pyruvate oxido-reductase, which generates electrons from
pyruvate.
nijH codes for the peptide subunit of the Fe-protein of nitrogenase.

niJa codes for the GC subunit of the MoFe-protein of nitrogenase.
nifl codes for the p subunit of that MoFe-protein.
nijE function uncertain, see nip.

nijN suspected of involvement, with ni@, in the synthesis or insertion of the
MoFe moiety in nitrogenase.
nifu function unknown.

nifs function unknown.
nip concerned with the specificity of the MoFe moiety.
nifM probably concerned with the FeS moiety of the Fe-protein.
niJF codes for the immediate electron donor to the Fe-protein.
nifA codes for a peptide which activates all the groups of nifgenes.
nifL codes for a repressor which over-rides activation by the nifA product.
niJB concerned with insertion of the MoFe moiety.

nifQ concerned with mobilising Mo for nitrogenase synthesis.
-----------
I shiA this gene, concerned with shikemic acid synthesis, lies at the other end of
nifand plays no part in its functioning.
The whole chain of operons was termed a regulon. It was activated by way of nifA and
switched off by way of nifL. The fact that all the nifgenes were linked into a single
regulon in Klebsiella was fortunate for research; future studies would show that, in most
nitrogen-fixing bacteria (including our A. chroococcurn),comparable nif operons exist
but they are dispersed about the genome.'
By the late 1970s, nifgenetics had become a popular subject for research and,
though we were leaders in this area, substantial advances were made in labora-
tories outside the UNF, as well as by visiting scientists working alongside the
UNF's geneticists. By then, with the genetics of K . pneumoniae nifas a guide, we
had started to investigate the genetics of nitrogen fixation in the aerobe A .
John Postgate 249
chroococcum, with special emphasis on its tolerance of oxygen and the fact that it
did not evolve dihydrogen. When Chatt retired in 1980 dioxygen-sensitive
mutants and mutants lacking hydrogenase had been obtained, despite oper-
ational difficulties in making mutants of this organism, but our major contribu-
tions to azotobacter genetics would be made later in the 1980s.
4.2 Regulation of nif

Gradually questions of how the expression of these genes was regulated arose:
how did external factors such as ammonium ions or oxygen supply switch nifon
or off! The emphasis of our research duly shifted from nifstructure towards nif
regulation. Plasmids were constructed bearing hybrid genes in which nif genes
with their regulatory sites were fused to an alien gene (lac) whose product
(P-galactosidase)gave a colour reaction with an appropriate substrate.26 These
were invaluable for studying the regulation of nifgenes, because they would give
a colour reaction when activated.
It was already clear that subclusters of genes within the nif cluster were
transcribed independently, processes regulated by the products of nifA and nifL
(Table 1). A complicated cascade of regulatory steps began to be revealed
whereby ammonium ions prevented activation of nifA and thus blocked tran-
scription of the whole of nif27 Studies with a construct in which the nifH gene
was fused with lac proved that dioxygen repressed nitrogenase synthesis in a
different way, seemingly by interacting directly with the nifL gene.28 In due
course the molecular details of these processes, and the nature of the stretches of
DNA which interacted with regulatory substances, would become the major
preoccupation of the UNF’s geneticists.
5 Interdisciplinary Interactions
The Unit’s research was essentially interactive and convergent. Just as conver-
gent chemical and biochemical thinking led to the discovery of a bound inter-
mediate at the oxidation level of hydrazine during N, reduction, so our bio-
chemical and physiological research benefited from each other’s progress, and
our genetics grew out of both - and benefited both. For example, I mentioned in
Section 2.3 that genetic mutants were recruited to prove that the FeMo-cluster of
the molybdoprotein carried the site which bound dintrogen. The n ip (Table 1)
mutant of K . pneumoniae made nitrogenase, but it was non-functional. If purified
MoFe-protein from normal K . pneumoniae was treated with acid, a fragment of
low molecular weight could be prepared which retained Mo, Fe and S as well as
exhibiting that protein’s characteristic EPR sprectrum. It was nicknamed
‘FeMoco’; it was a very unstable molecule and was rapidly destroyed by dioxy-
gen. When added to the defective n ip nitrogenase, FeMoco restored normal
activity. A second mutant, n i p , also made a defective nitrogenase but its defect
was different: it reduced acetylene but not N,. A FeMoco preparation could be
made from nzfVnitrogenase, and when this material was added to defective n ip
nitrogenase, it converted it to a nip-type enzyme: able to reduce acetylene but

not N,.
In later years studies on the chemical nature of FeMoco would involve
detailed collaboration between our biochemists and our chemists. Again, on the
physiological side, the nifH-lac fusion mentioned in Section 4.2 was exploited to
study the quantitative kinetics of oxygen repression? and another nif-lac fusion
was used to show that Mo had a regulatory effect on nitrogenase synthesis.
6 Envoi
The U N F was a multidisciplinary and cooperative research team; in addition, a
great deal of work involved collaboration with other laboratories throughout
the world. Moreover, the U N F was never without a few graduate students,
short-term students, visiting scientists and postdoctoral researchers?whose con-
tributions were often invaluable, sometimes ground-breaking. That is why I have
here attributed all our scientific advances to the UNF; it has been impracticable
to name names in this short chapter. The specific scientists involved were named
as authors of the UNF’s primary publications, most of which can be found in the
bibliography of a textbook written at about the time of Chatt’s retirement.,’
Chatt was at times perceptibly bemused by the directions the biological
research was taking and it says much for his understanding of the mechanics of
research that he accepted the dedication of his biological staff and gave them
their heads. The foregoing survey has given little more than a taste of the
systematic biological research during his fifteen-year Directorship of the UNF,
and further substantial advances were made in the years following his departure.
None the less, my account illustrates how effective a focussed multidisciplinary
project can be if the scientists participating are motivated to learn each other’s
languages and to interact cooperatively.
7 References
The UNF’s biological publications were so numerous during Joseph Chatt’s Directorship
that complete citations here would have been impracticable. Those below are signpost
references, principally to reviews, books or conference reports.
1 J. R. Postgate, Nitrogen Fixation, Cambridge University Press, London, New York

and Melbourne, 1998 (introductory).
2 G. Stacey, R. H. Burris and H. J. Evans (eds), Biological Nitrogen Fixation, Routledge,
New York and London, 1992 (advanced).
3 I. R. Kennedy and E. C. Cocking (eds),Global Nitrogen Fixation: The Global Challenge
and Future Needs, University of Sydney SUNfix Press, Sydney, 1997 (agronomic).
4 J. E. Carnahan, L. E. Mortenson, H. F. Mower and J. E. Castle, Biochim. Biophys.
Acta, 1960,44, 520.
5 J. R. Postgate, Notes Rec. Roy. SOC.Lond., 1998,52, 355.
6 J. R. Postgate, J . Gen. Microbiol., 1966’45, 1.
7 J. R. Postgate. The Sulphate-reducing Bacteria, 2nd edn., Cambridge Univrsity Press,
London, New York and Melbourne, 1984.
John Postgate 25 1
8 P. S. Nutman, Phil. Trans. Roy. Soc. Lond. B, 1987,317,69.
9 M. Kelly, J. R. Postgate and R. L. Richards, Biochem. J . , 1967,102, lc.
10 M. Dilworth, Biochim. Biophys. Acta, 1966,44,285.
11 R. Schollhorn and R. H. Burris, Proc. Nut. Acad. Sci. U S A , 1967,57,213.
12 K. Baker, Biotechno!. Bioeng., 1978,20, 1345.
13 R. R. Eady, B. E. Smith, R. N. F. Thorneley, D. A. Ware and J. R. Postgate, Biochem.
SOC.Trans., 1973,1, 37.
14 R. N. F. Thorneley, J. Chatt, R. R. Eady, D. J. Lowe, M. J. O’Donnell, J. R. Postgate, R.
L. Richards and B. E. Smith, in Nitrogen Fixation, Volume 1, W. E. Newton and W. H.
Orme-Johnson (eds.), University Park Press, Baltimore, 1980, p. 171.
15 R. N. F. Thorneley, R. R. Eady and D. J. Lowe, Nature, 1978,272,557.
16 M. G. Yates in Recent Developments in Nitrogen Fixation, W. Newton, J. R. Postgate
and C. Rodriguez-Barrueco (eds.), Academic Press, London, New York, San Fran-
cisco, 1977, p. 219.
17 J. R. Postgate, Laboratory Practice, 1965,14, 1140.
18 R. L. Robson and J. R. Postgate, Ann. Rev. Microbiol., 1980,34, 183.
19 H. J. Evans, T. Ruiz-Argueso, N. T. Jennings and J. Hanus in Genetic Engineeringfor
Nitrogen Fixation, A. Hollaender (ed.), Plenum Press, New York, 1977, p. 333.
20 S. Hill and J. R. Postgate, J . Gen. Microbiol., 1969,58,277.
21 S. Streicher, E. Gurney and R. C. Valentine, Proc. Nut. Acad. Sci. U S A , 1971,68,1174.
22 R. A. Dixon and J. R. Postgate, Nature, 1971,234,47.
23 R. A. Dixon and J. R. Postgate, Nature, 1971,237, 102.
24 J. D. Watson and J. Tooze, The DNA Story: A documentary History ofGene Cloning,
W. H. Freeman, San Francisco, 1981.
25 R. A. Dixon, F. C. Cannon and A. Kondorosi, Nature, 1976,260,268.
26 R. A. Dixon, R. R. Eady, G. Espin, S. Hill, M. Iaccarino, D. Kahn and M. Merrick,
Nature, 1980,286, 128.
27 G. Espin, A. Alvarez-Morales, F. Cannon, R. Dixon, and M. Merrick, Mol. Gen.
Genet., 1982,186, 518.
28 S. Hill, E. Kavanagh, R. B. Goldberg and R. Hanus, Nature, 1981,290,424.
29 J. R. Postgate. The Fundamentals of Nitrogen Fixation, Cambridge University Press,
London, New York and Melbourne, 1982.
Vanadium, Molybdenum and
Iron Complexes Based on a
Trithiolate Ligand
J. R. SANDERS
Department of Biological Chemistry, John Innes Centre, Colney Lane,
Norwich NR4 7UH, UK
1 Introduction- Phosphine and Thiolate Ligands

Joseph Chatt and his co-workers developed a functional model for nitrogenase
action at the ARC Unit of Nitrogen Fixation at the University of Sussex in the
1960s and 1970s. This involved coordination and reduction of dinitrogen at Mo
and W sites with monodentate and chelating phosphine co-ligands, and featured
the isolation and structural characterisation‘ of several intermediate complexes
in which the dinitrogen had been successively reduced to diazenide, hydrazide,
imide and nitride ligands, leading to the formulation of chemical2 and electro-
chemical3 cycles of reduction. In the course of this work a great deal of new
coordination chemistry of the metals Os, Re, Mo and W with ligands such as
hydride, carbon monoxide, nitric oxide, cyanide and isocyanides (which act as
inhibitors or alternative substrates for nitrogenase action) was developed. Subse-
quent research by Chatt’s group at Sussex and by his many ex-students and
collaborators contributed to the advance of coordination chemistry leftward
across the Periodic Table into the vanadium group. This advance, along with his
earlier discoveries in the field of platinum metals, has meant that one part of
Chatt’s legacy has been the establishment of the comparative coordination
chemistry of the transition metals; another part has been subsequent generations
of coordination chemists.
Concurrent studies on nitrogenases, including those carried out by biochem-
ists in the Unit at S U S S ~revealed
X,~ that they contain cofactors with significant
quantities of molybdenum, iron and sulfur (later it was found that in some
nitrogenases molybdenum is apparently replaced by vanadium or iron). Efforts
were therefore made to develop the nitrogen chemistry of these metals with
J . R. Sanders 253
co-ligands such as thiolates and thioethers rather than phosphines. Sulfur (unlike
phosphorus) tends to bridge between metals; while this results in the formation of
potentially very active cluster complexes, it means that the study of reactivity at
single-metal sites is difficult. The strategy used to counter this tendency of sulfur
to form bridges was to prepare compounds featuring sterically crowded ligands
such as triisopropylthiophenolate or macrocyclic compounds such as trithiacyc-
lononane. The chemistry of S-ligated metals such as iron, molybdenum and
vanadium with co-Iigands such as hydride, carbon monoxide, isocyanides and
hydrazines was developed in this way,' though a comprehensive scheme of
reduction of dinitrogen comparable to that on phosphine complexes was not
achieved.
2 Recent Studies on Nitrogenase

The structure of the nitrogenase cofactor FeMoco6 features an MoFe,S, core
(Figure 1). The Mo atom at one end of FeMoco is ligated by three core sulfurs,
two oxygens in a homocitrate ligand and one nitrogen (from histidine in the
protein). The iron at the other end is ligated by one cysteinyl sulfur and three core
sulfur atoms, while the six iron atoms in the middle are each ligated by three
sulfur atoms approximately in a plane, and apparently by no other ligands. On
extraction from the protein the histidine and cysteine ligands are replaced by
solvent, e.g. N-methylformamide.
There is still controversy about where on FeMoco dinitrogen, carbon monox-
ide and alternative substrates are bound and/or reduced. For example, it has
been proposed on the basis of molecular modelling studies and experiments on
structurally defined N, complexes7 that the homocitrate ligand on the molyb-
denum could become monodentate leaving a vacant site which would be suitable
for binding dinitrogen,8 and on the basis of recent density functional theory
calculations9 M. C. Durrant of this laboratory has suggested that the molyb-
denum is the preferred site for N, binding. The interaction of CO with FeMoco
has also been studied under turnover conditions by EPR'' and by stopped-flow
IR spectroscopy." Using the latter technique, bands at 1906, 1936, 1958 and
1880 cm-', appearing under different partial pressures of CO, have been assig-
ned to v(C0). Bands in this region are also observed when FeMoco is elec-
trochemically reduced in a thin layer cell under C 0 . l 2 It has been inferred from
these measurements that one CO molecule binds in a bridging mode between
Figure 1 The ( M o F e J , ) core of FeMoco

254 Vanadium,Molybdenum and Iron Complexes Bused on a Trithiolate Ligand
two Fe atoms at low partial CO pressure, but that at higher partial pressure two
CO molecules bind in a terminal mode on adjacent Fe atoms."
Semi-reduced, isolated FeMoco has not been observed to interact with N,, but
kinetic studies have shown that it binds thiols and selenols (probably at the
cysteinyl Fe), cyanide (at two sites, this Fe and the Mo) and protons (at the
bridging sulfurs). Reduction of acetylene to ethylene and of protons to H, has
also been demonstrated with various preparations of isolated FeMoco.'
3 Bulky Ligands and Tripodal Ligands

3.1 Ligands with Four-sulfur Donor Sets
The presence of sulfur groups in FeMoco has stimulated Sellman to prepare
ligands such as A and B (Figure 2) that are derived from benzenedithiol. B reacts
with iron(@ salts giving a sterically protected five-coordinate iron complex
[FeIIB] which binds CO and N2H4 (but not N,) to form six-coordinate com-
plexes. It also binds diazene N,H, which is produced by oxidation of bound
hydrazine in the dinuclear [(Fe11B),(p-N,H,)J.'4 This has led to a model for
binding of diazene between two iron atoms in a form of FeMoco in which one of
the central S bridges is lost and replaced by a bridging N,, diazene or hydrazine.
The two iron atoms in this opened-out FeMoco are assumed also to be ligated by
neighbouring glutamine and histidine groups in the protein, and the structure is
held together by N-H-S hydrogen bonds. A related bridging diazene complex is
[{ Ru'~A(PC~~)),(~-N,H,)] (Cy = cyclohexyl) which catalyses the N,-dependent
HD formation by D,/H+ exchange,15 a key feature of nitrogenase action.
Figure 2 The proligands A2- and B2 -
3.2 Ligands with Three-nitrogen Donor Sets

Two series of co-ligands used in syntheses of complexes that are able to react
with and reduce dinitrogen feature bulky substituted amido groups. Cummins
and co-workers have made [Mo(NRAr),] (R = t-butyl, Ar = 3,5-Me,C,H,), a
very crowded three-coordinate Mo"' compound that reacts with N, forming
[(Mo(NRA~)~),(~-N,)] and ultimately [MoN(NRAr),].'' Schrock and his col-
leagues have performed similar chemistry using trilithium salts of anions such as
([Me3SiNCH,CH,]3N)3- (NN3)3-.This reacts with simple salts of V, Mo and
Fe (among other metals) giving compounds in which the ligand binds in a
J . R . Sanders 255
tetradentate manner, with an almost planar MN, set. The metal has three
orbitals available to bind to one additional ligand; this facilitates the formation
of a triple bond to ligands such as N, (for Mo and Fe) and the bulky trimethyl-
silyl group provides steric protection for transformations of these ligands.
Thus [Mo(NN,)Cl] is reduced in thf with Mg to [{Mo(NN,)N=N),Mg(thf),]
and this reacts with FeCl, giving trigonal planar [{Mo(NN3)(N=N)),Fe].’*
3.3 Ligands with Three-sulfur Donor Sets

Seeking to combine the geometry of a ligand that provides a trigonal environ-
ment at a single metal site with the presence of sulfur donor atoms to mimic as far
as possible the coordination sphere of FeMoco, Power and co-workers syn-
thesised an iron(r1) complex anion [Fe(SC6H2-2,4,6-Bu‘),]- with very bulky
sulfur ligands,” but its interaction with small molecules such as N, and C O was
not reported. Other chemists have synthesised tripodal tetradentate ligands with
the three arms terminating in thiolate rather than amine groups. For example
Koch has used the ligands (P[C6H4-2-S]3)3-,20 (P[C6H3-2-S-3-Ph]3)3- and
(N[CH,C6H,-2-S],)3- 2 2 to stabilise a series of complexes of iron and nickel in
supporting ligands such as CO and CN, and George has made FeIVcomplexes
using the ligand (P[C6H,-2-S-3-Me3Si]3)3 -.23
4 A Simple Tripodal Ligand

4.1 Introduction to the NS, Site
We have explored the chemistry of the simple tripodal ligand (N[CH,CH,-
S],),- (NS,)3- with V, Mo and Fe, with respect to the binding of N,, nitrogen
species such as hydrazine, and other small molecules such as CO, CNMe, CN-
and NO. We have made this ligand on a large scale by adaptation of literature
~yntheses,,~ taking stringent safety precautions in handling the vesicant
tris(ch1oroethylamine) hydrochloride. Our results are presented in the remainder
of this review and illustrate the several different types of bonding exhibited in
complexes of NS,.
4.2 Vanadium and Molybdenum Complexes of NS,

We have studied the reactions of H,(NS,) with vanadium more thoroughly than
those with any other starting from the known [V(0)(NS3)].26Figure 3
shows that the V(NS,) site will support imide, hydrazido, hydrazine, ammonia
and nitride ligands, forming multiple bonds to nitrogen species that may be
involved in the later stages of reduction of N,. Several of these complexes are
256 Vanadium,Molybdenum and Iron Complexes Based on a Trithiolate Ligand
0
Y
p?
I (ii)
S,;-s
NSiMe3
Figure 3 Transformations of V(NS,) compounds. Reagents (i)N,H,; (ii)thf, 60 "C;

(iii)Me,SiNHNMe,; (iv)NUN,; (v) N,SiMe,; (vi) H,O; (vii)(NMe,)OH
interconvertible. The site does not bind N,, H, or CO in stable complexes,

though we have evidence for the transient formation of an unstable dinuclear
dinitrogen complex [{V(NS,)),(p-N,)], formed in the initial stages of the reac-
tion of [V(O)(NS,)] with hydrazine. It is noticeable that the V(NN,) site also
does not bind N,, despite the steric protection of the vanadium inside the pocket
formed by the bulky substituents on the NN, ligands.
The compounds shown in Figure 3 have all been characterised crystallog-
raphically. The bridging nitride anion [(V(NS,)) 2(p-N)]- features the first struc-
turally characterised linear symmetrical V-N-V bridge., There is no direct
bonding between the two vanadium(1v) atoms. There are no compounds with a
single-atom bridge between two metal atoms in NN, chemistry, where close
approach of the M(NN,) sites is prevented by the bulky substituents on the NN,
ligands.
We have made several other VII1complexes by displacing the hydrazine ligand
in [V(N,H,)(NS,)] with C1-, N3-, CN-, MeCN or CNBu', and have prepared a
series of imido-complexes [V(NAr)(NS,)] (Ar = various substituted phenyl
groups), and of hydrazido-complexes [V(NNR1R2)(NS,)] (R', R2 = methyl or
phenyl) from reactions of [V(O)(NS,)] with aryl isocyanates or 1,l-disubstituted
hydrazines, respectively. Structural studies on several of these compounds al-
ways reveal trigonal bipyramidal coordination about the V atom. The V-N
distance in the V(NS,) system, as a result of the shape of the NS,, is sensitive to
J . R. Sanders 257
the trans-influence of the other ligand. The range of compounds [VZ(NS,)]
allows comparisons of trans-influence within two series of complexes, paramag-
netic [V"IZ1(NS3)] and diamagnetic [VvZ2(NS,)]. For the first series, the V-N
distances in the M(NS,) system trans to Z' give an order of trans-influence of
C1- > NH, > N2H4 > MeCN, and for the second series the order is
0 > NSiMe, > NH > NC,H,Cl-4 > "Me, = NNMePh. The V-N distan-
ces in the latter are consistently higher than those in the former, reflecting the
higher trans influence of the multiply bonded Z 2 ligands (despite the smaller
ionic radius of Vv compared with that of VIII).
We have not studied the Mo(NS,) site as extensively as that of the vanadium
analogue,28but it is evident that the pattern of reactivity of molybdenum with
NS, is similar. Structurally characterised five-coordinate Mot" compounds,
prepared starting from [MoO,(acac),] (acac = pentane-2,4-dionate), include
[Mo(NO)(NS,)], [Mo(NNR)(NS,)], [Mo(NNR,)(NS,)](BF,) (R = Me or Ph)
and [{Mo(NS,)),(pS)] as a minor product. However, treatment of [(Mo(p-
Br)Br(CO),),] or of [WI,(CO),(MeCN),] with H,(NS,) gives [(MIV(NS,)},{p-
SCH,CH,N(CH,CH,SH),-S),] (M = Mo or W), which feature two bridging
diprotonated NS, ligands. Thus the NS, chemistry of both vanadium and
molybdenum is characterised by multiple bonding between the metal and the
ligand trans to the NS, ligand, and by the absence (so far) of any compounds in
which the metal atom has an oxidation state lower than 111.
4.3 Iron Chemistry of NS3-Carbonyl,Nitrosyl and Isocyanide

Complexes
Reaction of the very soluble [Fe(acac),] with (Et,N)Cl and H,(NS,) in MeCN
gives (Et,N)[Fe(NS,)Cl] (high-spin iron(@, S = 5/2). Unlike its vanadium
counterpart, this is reduced (Figure 4) by sodium amalgam in MeCN, giving a
yellow solution. This solution does not react with dinitrogen or with dihydrogen,
but addition of CO gives the green (Et,N)[Fe(NS,)(CO)], with v(C0) at 1910
cm-'. This is one of three five-coordinate iron@) carbonyl compounds dis-
covered recently; all have magnetic moments at 20°C characteristic of an S = 1
state.2 9,309 2 1
The Fe(NS,) site is the only one of the three M(NS,) sites that we have studied
which appears to be capable of binding carbon monoxide. In order to under-
stand better the characteristics of this binding, we have carried out quantum
calculations on [Fe(NS,)Cl] - and [Fe(NS,)(CO)] -, using density functional
theory. The calculations correctly predict the ground states of the anions, give
the v(C0) of the carbonyl anion as 1926 cm-' (close to the experimental value of
1910 cm-') and give good agreement between calculated and observed bond
distances and angles in the Fe(NS,) system. They also predict the C O binding
energy in the carbonyl anion to be - 102 kJ mol-'. We have carried out similar
calculations on the hypothetical [Fe(NS,)(N,)] -; these show that the ground
state has S = 2, that v(NN) is 2222 cm-' and that the N, binding energy is
- 29 kJ mol -'. The dinitrogen, unlike the carbonyl, is not a strong enough ligand
258 Vanadium, Molybdenum and Iron Complexes Based on a Trithiolate Ligand
[Fe(MeCOCHCOMe)3] + H3(NS3)
p] CNMe
\ (vii)
/(vii)
p] NO
Figure 4 Transformations ufFe(NS,) compounds. Reagents; (i)(Et,N)C1;

(ii)(Et,N)OAc + CO; (iii)(Et,N)OAc + C N M e ; (iu)Na/Hg + CO;
(v) N a / H g + C N M e ; (ui)N a / H g + N O ; (uii) N O
to enforce spin-pairing and the dinitrogen complex is too labile to be isolated.

The results show that if the trigonal iron sites in FeMoco are similar to the
Fe(NS,) site then C O binding at the former is entirely plausible, but that N,
binding is, at best, transient.
(Et,N)[Fe(NS,)(CO)] is also formed (Figure 4) if [Fe(acac),] is treated with
H,(NS,) in presence of tetraethylammonium acetate under an atmosphere of
carbon monoxide, when the excess of H,(NS,) acts as a reducing agent. If the
tetraethylammonium acetate is omitted, a complex of stoichiometry
Fe,(NS,),(CO), and structure [Fe{Fe(NS,)(CO)),-S,S’] (Figure 5 ) with a linear
trinuclear Fe,S, core is isolated. The value of v(C0) in the spectrum of this
trinuclear compound is 1938 cm-’. The following equilibrium holds in meth-
anol, and can be demonstrated by the reversible uptake of CO.
We found that the anion [Fe(NS,)(CO)]- can also be transformed into the
trinuclear compound [Fe{ Fe(NS,)(CO)-S,S’) ,] by condensation with FeCl,.
This prompted us to search for analogues with other metals and we have made
[Co{Fe(NS,)(CO)-S,S’),] and D. J. Evans in our laboratory has made
[Ni{Fe(NS,)(CO)-S,S} ,I, by starting from CoCl, and NiCl, respectively; these
have structures almost identical to that of the all-iron trimer. However, reactions
of the iron carbonyl anion with CuCl,, ZnC1, and VCl, produce only the all-iron
trimer.
J . R. Sanders 259
Figure 5 Crystal structure of [Fe(Fe(NS,)(CO)-S,S')2]
We have searched for interactions of the Fe(NS,) site with other small mole-
cules able to function as alternative substrates for dinitrogen. In reactions which
exactly parallelled those with carbon monoxide, we have obtained blue, mono-
nuclear, terminal RNC complexes (Et,N)[Fe(NS,)(CNR)] (R = Me, But or Cy)
either by treatment of [Fe(acac),] with RNC and H,(NS,) in the presence of
(Et,N)OAc, or by reduction of Et,N[Fe(NS,)Cl] in the presence of CNR.,l
They have NC stretching frequencies in the terminal range (2061-1940 cm-l),
values lower than those of the corresponding proligands (unlike in the case of the
PI1'compound [V(NS,)(CNBu')]). By treating these isocyanide-containing ani-
ons with metal chlorides in MeCN we have made homo- and hetero-metallic
cluster complexes [M{Fe(NS3)(CNR)-S,S),1 (M = Fe, Co, or Ni) with some-
what higher NC stretching frequencies than in the parent anions.
We have also prepared the nitrosyl compound (Et,N)[Fe(NS,)(NO)] from the
reaction of [Fe(acac),] with H,(NS,) in presence of Et,NOAc and either N O gas
or the N O source N-nitroso-N-methyl-p-toluenesulfonamide.32 The nitrosyl
anion has v(N0) at the relatively low frequency of 1621 cm-', a magnetic
moment corresponding to a state S = 3/2 and a Fe-N-0 angle of 154.4'. This
nitrosyl anion also reacts with metal halides to make clusters [M{Fe(NS,)(NO)-
S,S'),] (M = Fe, Co, or Ni), with rather higher values of v(N0) than in the parent
anion. The trinuclear carbonyl, isocyanide and nitrosyl complexes have subnor-
mal magnetic moments at 20 "C; variable-temperature magnetic measurements
are underway.
We have been unable to isolate a complex with a simple monodentate cyanide
ligand on the Fe(NS,) site but we have made an analogous complex of cobalt
(Et,N)[Co(NS,)(CN)] (v(CN) 2100 cm-l, spin S = 1) by the reaction of
[Co(acac),] with (Et,N)CN and H,(NS,). 29
The values of v(C0) in our anionic (1910 cm-l) and neutral (1938 cm-l)
carbonyls resemble those found (1906, 1936, 1958 and 1880 cm-') in the turn-
over reaction of FeMoco with CO," reinforcing the idea that the central
trigonally ligated iron atoms in FeMoco are those where C O binding occurs, and
suggesting that such binding (even at low pressures of CO) is terminal rather
than bridging. Comparable IR data for CN-, RNC and N O interactions with
the enzyme are not yet available.
4.4 Models for Hydrogenases

In the trinuclear compounds with Fe,S, clusters described above, two of the
three sulfur atoms in each NS, ligand are used to make auxiliary bonds to the
central metal atom. The individual ligation about each of the outer iron atoms is
however still trigonal bipyramidal, but this is not true of all NS, compounds.
Treatment (by D. J. Evans in our laboratory) of (Et,N)[Fe(NS,)(CO)] under C O
with [NiCl,(dppe),] (dppe = Ph,CH,CH,PPh,) gives the dinuclear complex
[{ Fe(NS,)(CO),-S,S’)NiCl(dppe)]in which the iron atom is octahedrally coor-
dinated.,, The core of this compound is dinuclear with nickel bound to iron by a
bis(thio1ate) bridge, and iron binding two C O molecules. This is as close a
structural model as exists of the active site of the NiFe-hydrogenase from
Desulfovibrio gigas which is a dinuclear t hiolate-bridged nickel-iron complex in
which the nickel atom is coordinated by four cysteinate-sulfur atoms, two of
which bridge to a six-coordinate iron atom.,, This hydrogenase model illus-
trates the use of metal-sulfur compounds in modelling cofactors of metal-sulfur
enzymes other than nitrogenase, of which hydrogenases are outstanding
examples. Related work by C. J. Pickett in our laboratory with the tridentate
tripodal proligands MeC(CH,SH), and its monomethylated derivative
MeC(CH,SH),(CH,SMe) has led to modelling of the all-iron hydrogenase
system.,’
4.5 Dinuclear and Tetranuclear Iron-dinitrosyl Compounds

The possibility of sulfur bridging in our Fe(NS,) systems, together with the lack
of steric protection of the metal atom afforded by bulky groups, makes isolation
and study of complexes containing simple ligands on iron more difficult than
when the Fe(NN,) site is used; on the other hand, these same properties give our
complexes extra synthetic potential not available to Fe(NN,) systems. We have
further exploited this synthetic potential using the iron dinitrosyl dimer
[(F~(NO),),(,U-I),].~~ If this is treated with the anions [Fe(NS,)(L)] - (L = CO,
CNR or NO) described above, then semi-Roussin salts [Fe(NO),( Fe(NS,)(L))-
S,S’], again with double-sulfur bridges, are formed. [{ Fe(NO),) ,(p-I)J has also
been treated with the yellow solution obtained from the reduction of the
[Fe(NS,)Cl] - anion (see above); this gives the tetranuclear [{Fe(NO),-
{Fe(NS,)}-S,S’},S,S’] (Figure 6) in which all the sulfur atoms of the NS, ligands
form auxiliary bonds to other iron atoms. While these dinuclear and tetranuclear
compounds are not obvious models for any metal-sulfur enzyme their further
elaboration and the development of the use of the iron dinitrosyl iodide reagent
present interesting synthetic challenges.
The NS, ligand thus displays different behaviour with iron from that with
vanadium and molybdenum; this is largely dictated by the stability of the iron(@
oxidation state. The reactions of H,(NS,) with iron compounds of very low
J . R. Sanders 26 1
Figure 6 Crystal structure of [Fe{Fe(NO),{ Fe(NS,)}-S,S’} ,S,s’]
oxidation state such as iron carbonyls have not yet been explored. Interesting
chemistry may also be expected if one of the (CH,CH,SH) arms of H,(NS,) is
derivatised to make a mixed thiolate-thioether proligand such as
N(CH,CH,SH),-CH,CH,SMe, similar to that in the all-iron hydrogenase
model ligand described above.
5 Acknowledgements
I should like to thank my colleagues in the Unit, later the Nitrogen Fixation
Laboratory, and now part of the Department of Biological Chemistry at the
John Innes Centre in Norwich, who have helped me in this research and some of
whose results I have presented here. Special thanks, in reverse order of age, go to
Marcus Durrant, Dave Evans, Chris Pickett, David Hughes and Ray Richards.
6 References
1 A. Galindo, A. Hills, D. L. Hughes, R. L. Richards, M. Hughes and J. Mason, J . Chem.
Soc., Dalton Trans., 1990,283.
2 R. A. Henderson, G. J. Leigh and C. J. Pickett, Adv. Inorg. Chem. Radiochem., 1983,27,
283.
3 C. J. Pickett and J. Talarmin, Nature, 1985,317,652.
4 See chapter by J. R. Postgate, this volume, pp. 233-251.
5 R. L. Richards, New J . Chem., 1997,21,727.
6 D. C. Rees, M. K. Chan and J. Kim, Adv. Inorg. Chem., 1993,40,89.
7 K. C. L. Gronberg, C . A. Gormal, M. C. Durrant, B. E. Smith and R. A. Henderson, J .
Am. Chem. Soc., 1998,120, 10613.
8 C. J. Pickett, J . Bioinorg. Chem., 1996,1,601.
9 M. C. Durrant, Inorg. Chem. Comm., 2001,4,60.
10 H.-I. Lee, L. M. Cameron, B. J. Hales and B. M. Hoffman, J . Am. Chem. Soc., 1997,119,
10121.
11 S. J. George, G. A. Ashby, C. W. Wharton and R. N. F. Thorneley, J . Am. Chem. Soc.,
1997,119,6450.
12 S. K. Ibrahim, K. Vincent, C. A. Gormal, B. E. Smith, S. P. Best and C. J. Pickett,
13 B. E. Smith, M. C. Durrant, S. A. Fairhurst, C. A. Gormal, K. C. L. Gronberg, R. A.
Henderson, S. K. Ibrahim, T. Le Gall and C. J. Pickett, Coord. Chem. Rev., 1999,
185186,669.
14 D. Sellman and J. Sutter, Acc. Chem. Res., 1997,30,460.
15 D. Sellman and A. Fiirsattel, Angew. Chem. Int. Ed. Engl., 1999,38,2023.
16 C. E. Laplaza, M. J. A. Johnson, J. C. Peters, A. L. Odom, E. Kim, C. C. Cummins, G.
N. George and I. J. Pickering, J . Am. Chem. SOC.,1996,118,8623.
17 R. R. Schrock, Acc. Chem. Res., 1997,30,9.
18 M. B. O’Donoghue, W. M. Davis, R. R. Schrock and W. M. Reiff, Inorg. Chem., 1999,
38,243.
19 F. M. MacDonnell, K. Ruhlandt-Senge, J. J. Ellison, R. H. Holm and P. P. Power,
Inorg. Chem., 1995,34,1815.
20 H.-F. Hsu, S. A. Koch, C. V. Popescu and E. Munck, J . Am. Chem. SOC.,1997, 119,
8371.
21 D. H. Nguyen, H.-F. Hsu, M. Millar, S. A. Koch, C. Achim, E. L. Bominaar and E.
Miinck, J . Am. Chem. SOC.,1996,118,8963.
22 N. Govindaswamy, D. A. Quarless and S. A. Koch, J . Am. Chem. SOC., 1995,117,8468.
23 J. D. Niemoth-Anderson, K. A. Clark, T. A. George and C. Ross, J . Am. Chem. SOC.,
2000,122,3977.
24 P. Barbaro, C. Bianchini, G. Scapacci, D. Masi and P. Zanello, J . Am. Chem. SOC.,
1994,33,3180.
25 S. C. Davies, D. L. Hughes, Z . Janas, L. Jerzykiewicz, R. L. Richards, J. R. Sanders, J.
E. Silverston and P. Sobota, Inorg. Chem., 2000,39,3485.
26 K. K. Nanda, E. Sinn and A. W. Addison, Inorg. Chem., 1996,35,1.
27 S. C. Davies, D. L. Hughes, Z. Janas, L. Jerzykiewicz, R. L. Richards, J. R. Sanders and
P. Sobota, J . Chem. SOC.,Chem. Commun., 1997,1261.
28 S. C. Davies, D. L. Hughes, R. L. Richards and J. R. Sanders, J . Chem. SOC., Dalton
Trans., 2000, 719.
29 S. C. Davies, D. L. Hughes, M. C. Durrant, R. L. Richards and J. R. Sanders, J . Chem.
SOC., Dalton Trans., 2000,4694.
30 M. Ray, A. P. Golombek, M. P. Hendrich, V. G. Young, Jr. and A. S. Borovik, J . Am.
Chem. SOC., 1996,118,6084.
31 L. Cooper, S. C. Davies, J. R. Dilworth, D. L. Hughes, M. Konkol, R. L. Richards, J. R.
Sanders and P. Sobota, Can. J . Chem., in press.
32 M. Konkol and J. R. Sanders, unpublished results.
33 S. C. Davies, D. J. Evans, D. L. Hughes, S. Longhurst and J. R. Sanders, Chem.
Commun., 1999,1935.
34 M. W. W. Adams and E. I. Stiefel, Curr. Opin. Chem. Biol., 2000,4,214.
35 C . J. Pickett and M. Razavet, Chem. Commun., 2001,847.
36 B. Haymore and R. D. Feltham, Inorg. Synth., 1973,14,81.
The Nature of Molybdenum and
Tungsten Centres in 0x0-transfer
Enzymes
C. DAVID GARNER
School of Chemistry, University of Nottingham, Nottingham NG7 2RD, UK
1 Introduction
In a perspective provided by structure of the Periodic Table, molybdenum and
tungsten are distinct in being the only 4d- and 5d-transition metals that are
required for the normal metabolism of biological systems. These metals play a
vital role as the catalytic centres of a wide variety of Mo was first
identified as an essential trace element in the 1930s,because of its role in nitrogen
; ~ metal is now known to be the catalytic centre of over 50 enzymes.
f i ~ a t i o nthis
Evidence for the involvement of W in biological systems has been obtained only
relatively recently, especially for enzymes of hyperthermophilic archea that
thrive near 100°C.5
There are (at least) two striking parallels between the nature and function of
Mo and W centres in enzymes. First, the net effect of the catalysis effected by
virtually all of these enzymes is the transfer of an oxygen atom to, or from, the
substrate - with the metal undergoing a concomitant redox change from Mvlto
MIv or vice versa (1).
X + H,O + Mvle MIv + 2H+ + XO (1)
Secondly, in each of the 0x0-transfer enzymes, a single Mo or W atom is bound
to one or two molecules of a special ligand, generally known as ‘molybdopterin’
(MPT).6This entity has been shown by a series of spectroscopic and degradative
studies to be present in all of the Mo 0x0-transfer enzymes. However, MPT was
first structurally characterised by protein crystallography in a W enzyme, the
aldehyde oxidoreductase from Pyrococcus f~riosus.~ Figure 1 shows the struc-
ture of MPT. It comprises a reduced pterin, fused to a pyran ring bearing a
dithiolene group that chelates Mo or W.
264 The Nature of Molybdenum and Tungsten Centres in 0x0-transfer Enzymes
O S-
Figure 1 Molybdopterin ( M P T ) ,one or two molecules of which ligate M o (or W) in the

M o (or W) enzymes; the phosphate group may be bound to a nucleotide ( R )
Despite the above important common denominators, and the similar size and
chemical properties of chemically equivalent W and Mo species,8 important
differences have evolved in the biological roles of these two metals. Two particu-
lar points highlight these differences: (i) the significance of these metals for
nitrogen fixation; (ii) the prevalence of Mo over W as the catalytic centre of
0x0-transfer enzymes.
Protein crystallographic studies of the FeMo-protein of nitrogenase have
identified Mo as an integral component of the Fe-Mo-S cluster that consti-
tutes the catalytic centre.' Although the role that Mo plays in the catalytic
process is not yet established, substitution of Mo by W leads to an inactive
enzyme." This difference is surprising because synthetic Fe-Mo-S and
Fe-W-S clusters manifest quite similar chemical and the
elegant studies of Chatt et clearly demonstrated the feasibility of reduc-
ing N, bound to a Wo centre under ambient conditions.
Mo is a trace element that is required by all forms of life, from bacteria,
through higher plants and animals to man,' whereas W has only been
identified in a few primitive organism^.^,^ Mo enzymes catalyse a wide
variety of conversions, as illustrated in Figure 2, several of which are crucial
to humans. Xanthine oxidase catalyses the oxidation of xanthine to uric acid
and also oxidises a range of aromatic heterocycles and simple a1deh~des.l~
Sulfite oxidases occur in animal and human livers and catalyse the oxidation
of SO,,- to SO4,-, as the final step in the oxidative degradation of the
sulfur-containing amino acids cysteine and methionine. Humans excrete ca.
1 g of SO4'- per day. Rare genetic deficiencies that lead to a lack of sulfite
oxidase produce severe neurological abnormalities, mental retardation, and
death in infancy.15Also, Mo is essential for both routes to fixed nitrogen in
the biosphere: the nitrogenases that convert N, to NH3 (see (i)) and the
nitrate reductases that catalyse the first step in nitrogen assimilation, conver-
0 0
Sulfite Oxidase SO
-: + H20+SO-,' + 2H+ + 2e-
Nitrate Reductase NQ2-+ 2H+ + 2 8 - NO2- + H20
DMSO Reductase (CH3)$30 + 2H+ + 2e-+ (CH3)2S+ Y O
Figure 2 Examples of reactions catalysed by M o enzymes

C. David Garner 265
sion of NO,- to NO2- (prior to the reduction of NO,- to NH, by some

nitrite reductases). l 6
In contrast to the Mo 0x0-transfer enzymes, the majority of the W enzymes
only catalyse aldehyde oxidation or its equivalent, i.e. oxygen atom transfer to a
~ a r b o n , with
~ . ~ acetylene hydratase being a notable exception.' However,
iso-enzymes, in which W is substituted for Mo -the metal normally incorporated
by biology in these systems - have been identified'8p20and these are extending
the range of W 0x0-transfer enzymes. Studies of the relative behaviours of these
two metals at the same catalytic site are now beginning to reveal interesting
differences in their kinetics and thermodynamics (see Section 3.2).
2 Nature of the Catalytic Centres of Mo and W

0x0-transfer Enzymes
2.1 Coordination of the Metal Atom
Several protein crystallographic investigations have been accomplished for Mo
and W enzymes. The results of these structural studies have provided important
details concerning the overall molecular architecture of the polypeptides in-
volved and the nature and relative disposition of the corresponding reaction
centres. Although proposed in advance of several of the protein crystallographic
studies, the classification of the nature of Mo centres in 0x0-transfer enzymes by
Hillel (Figure 3 ) serves as a very useful, if still a provisional, means of subdividing
this large group of enzymes.
Thus, in the oxidised forms of the enzymes:
members of the xanthine oxidase family have one MPT group bound to a
fac- M oOS(H, 0)cent re;
members of the sulfite oxidase family involve one MPT group bound to a
cis-MOO, unit;
members of the DMSO reductase family possess two MPT groups bound to a
cis-MoO(0Ser) centre - this group of enzymes may also involve a Mo=S or a
Mo=Se group in place of the Mo=O group and a Mo-SeCys group in place of
the Mo-OSer group.
Several crystallographic studies of the Mo DMSO reductase (Mo-DMSOR)
of Rhodobacter capsulatus and Rhodobacter sphaeroides have been accom-
Xanthine Oxidase Sulfite Oxidase DMSO Reductase

Family Family Family
Figure 3 Nature of the catalytic centres of some M o enzymes'

p l i ~ h e d . ~ An
' - ~ important
~ aspect of these structural studies has been a dis-
cussion concerning whether the oxidised enzyme has two or one '0x0'-groups.
The structure determination of R. sphaeroides DMSOR at 1.3 A resolution24 is
significant in this context since it indicated that the enzyme, as crystallised,
contains a disordered mixture of two types of Mo centre: one comprising a
square pyramidal [MoO,(OSer)(MPT)] centre and the other a distorted trig-
onal prismatic [MoO(OSer)(MPT),] centre. The present consensus favours the
latter as the catalytic site.
The catalytic centres of W 0x0-transfer enzymes appear to belong to the
DMSOR family, as exemplified by the structure of the aldehyde oxidoreductase
of P. furiosus.
2.2 Function of the Metal Centre

The Mo centre of all of the Mo enzymes characterised to date is able to access the
MoV1,MoV,and MoIVoxidation states. Spectroscopic studies of these enzymes,
notably EPR investigations of the MoVstate, have clearly demonstrated that the
substrate interacts directly with the metal centre.25Bailey et al. have character-
ised the novel, pink form of DMSO reductase from R. capsulatus, produced by
the addition of a large excess of DMS to the oxidised enzyme. This study revealed
the presence of a complex with 'DMSO' bound to the Mo via its oxygen atom
(Figure 4).23The oxidation state of the metal and the electronic structure of this
centre have yet to be established. However, the clear implication of this struc-
tural investigation is that Mo-DMSOR catalyses the direct transfer of an oxygen
atom from an Mo=O group to DMS. Microscopic reversibility means that the
reduction of DMSO also involves direct transfer of an oxygen atom forming an
Mo=O group. These conclusions are compatible with the observation that
oxidation of the tertiary phosphine 1,3,5-triaza-7-phosphatricyclo[3.3.1. llde-
cane by DMS''0 mediated by R. sphaeroides DMSOR involves the transfer of
"0 from the sulfur to the phosphorus.26 The presumption is that the reaction
sequence is: DMS 8 ~ - ~ o DMSJ ~ ~ ; 'O=MOV[; R,P-' 'O=MOV';
R P=' 0-M01".
Figure 4 The complex with 'DMSO' bound to the M o via its oxygen atom at the heart of
the DMSO reductusefrom R. c a p s ~ l a t u s ~ ~
C. David Garner 267
The nature of the catalyses effected by Mo enzymes of the xanthine oxidase
and sulfite oxidase families is not yet fully established, but it is clear that they do
not function in the simple, direct, manner of the DMSO reductases.'
3 W-substituted DMSO Reductase

3.1 Production and Characterisation of W-DMSOR2'
The periplasmic dimeth ylsulfoxide reduct ases (DM SORs) of the photosynthetic
bacteria R. capsulatus and R. sphaeroides function in a respiratory chain with
DMSO as the terminal electron acceptor and catalyse the environmentally
important Reaction (2).2 7 3 2
DMSO + 2e- + 2 H ' e D M S + H 2 0 (2)
These enzymes have a high affinity for DMSO and will also catalyse the reduc-
tion of trimethylamine-N-oxide (TMAO) to trimethylamine (TMA).29 The
DMSORs are the simplest known Mo enzymes and are purified as monomers of
molecular weight (M,) ca. 85000 Da that comprise a single polypeptide chain
containing only one prosthetic group with an Mo atom bound to two pyranop-
terin guanine dinucleotides.
Stewart et al. have investigated whether the DMSOR of R. capsulatus is
capable of utilising W in place of M o . ~ ' Prior to this study, only limited
comparisons between the nature and properties of a W enzyme and those of the
corresponding Mo enzyme had been made.'8919Substitution of Mo by W was
accomplished readily via natural uptake under [W0412--rich, [Mo0412--
depleted, conditions and W:Mo ratios of > 99: 1 were produced. However, it is
important and interesting to note that the presence of a low concentration (6 nM)
of Na2[Mo0,] was essential for cell growth. In a separate experiment, involving
a growth medium containing equal quantities of the two metals (3 pM
Na2[Mo04] plus 3 pM Na2[W04]) the Mo: W ratio in the isolated DMSOR
was ca. 1.5: 1. Thus, the normal processes of metal uptake, delivery and/or
incorporation lead to only a slight preference for Mo over W in the DMSOR of
R. capsulatus.
In these investigations, the Mo- and W-grown cells were found to contain the
same amount of DMSOR. This result is in contrast to that of Santini et al., who
observed that the amount of W-substituted TMAO reductase, produced from E.
coli by genetic manipulation of the pathway for metal uptake, was ca. 15% of the
level found with M o . ~ '
R. capsulatus W-DMSOR was obtained with the metal in the Wvl oxidation
state.20Thus, the isolated enzyme exhibited no EPR signal and none could be
induced by the addition of K,[Fe(CN),]. However, an EPR signal, consistent
with the presence of Wv, was obtained by incubation of the enzyme with
dithionite for 6 10 minutes; further such incubation extinguished the EPR
signal, presumably because of the formation of WIV.The nature of the super-
hyperfine interaction observed in the EPR signal implies the presence of a
WV-OH group and this appears to be the first identification of this moiety in a
protein. The rhombicity, [(gl - g2)/(g1 - g3)], and the orientation of the Wv
g-values were both very similar to those of the MoV 'high-g split' signal of
Mo-DMSOR.~' Also, the two EPR signals display a similar magnitude and
orientation for metal-proton superhyperfine coupling tensors. Thus, the MoV
and W(v) centres of R. capsulatus DMSOR appear to experience essentially the
same ligand field. Furthermore, the UV-visible spectrum of oxidised W-
DMSOR has a profile similar to that of its Mo ~ o u n t e r p a r t , ~ with ~ ~ A,.,~ ~
~ ' the
values blue-shifted by ca. 150 nm (i.e. 3000-5000 cm-I). The absorptions are
considered to arise from ligand-to-metal charge-transfer transitions, from sulfur-
based orbitals to a do metal centre. Consistent with this view, the blue-shift
observed is similar to that (ca. 4350 cm-' from Mo to W) for the two lowest
energy transitions of the [MS,I2- (M = Mo or W) anions.33 These spec-
trochemical differences indicate that Wvl is more difficult to reduce than MoV1
(see Section 3.2), in this case by ligand-to-metal charge-transfer.
The structure of W-DMSOR from R. capsulatus2' corresponds directly to the
structure of the Mo-DMSOR isolated from this bacterium.23Thus, W-DMSOR
comprises a single polypeptide chain that envelops the prosthetic group that
comprises a WO(0Ser)centre bound to two MPT groups, each covalently linked
to a guanine dinucleotide (Figure 5). It is not clear whether there is a third oxygen
atom bound to the W because of the possible ripples in the electron density in the
vicinity of this heavy atom. However, the W LII,-edgeEXAFS was satisfactorily
interpreted by backscattering from two oxygen atoms, one at 1.76 A (W=O) and
another at 1.89 A (W-OSer), plus a shell of four sulfur atoms at 2.44 A from the
metaL2
3.2 Assays of Mo- and W-DMSO Reductase A ~ t i v i t y ~ ' , ~ ~

The activity of W-DMSOR has been measured, using procedures described
( f ? O ) ( H O ) ( O ) P O ~ ~ N yI "'y .NH2
Figure 5 Schematic representation of the nature of the metal centre of oxidised

W-substituted DMSO reductase of Rhodobacter capsulatus ( R =guanine
dinucleotide)20
C. David Garner 269
previously for M o - D M S O R . ~With
~ the dithionite-reduced dye, methyl violo-
gen, as the electron donor, the steady-state rate of oxidation of the dye was found
to be 52.8 & 1.6 s-' for Mo-DMSOR and 936 2 20 s-' for W-DMSOR. In the
reverse assay, measuring the rate of DMS oxidation using 2,6-dichlorophenol-
indophenol (DCPIP) as the oxidant, the activities of Mo-DMSOR and W-
DMSOR were found to be 8.5 & 0.1 s-l and < 0.05 s-', respectively. Thus, in
these assays, W-DMSOR reduces DMSO some 17 times faster than the normal
(Mo) enzyme but, in contrast to the behaviour of Mo-DMSOR, W-DMSOR
displays no discernible ability to catalyse the oxidation of DMS.
'H NMR spectroscopy provides a very convenient means of assaying the
activity of enzymes in intact ~ e l l s .This
~ ~ technique
,~ ~ has been used to monitor
the rate of turnover of DMSO and the alternative substrate TMAO by DMSO
reductase in R. capsulatus cells containing Mo-DMSOR or W-DMSOR.37In
each experiment, the initial 'H NMR spectrum was dominated by the DMSO (or
TMAO) singlet, which decreased steadily in amplitude over time with a con-
comitant growth of the DMS (or TMA) signal. Plots of the concentration of
substrate (or product) us. time were essentially linear, indicating a zero-order
process. The W-grown cells were found to reduce both DMSO and TMAO at ca.
9 and 22%, respectively, of the rate of Mo-grown cells. Nevertheless, these
experiments showed that the W-grown cells are clearly capable of turnover, with
either DMSO or TMAO acting as the terminal electron acceptor.
The lack of ability of the isolated W-DMSOR to catalyse the oxidation of
DMS with DCPIP as the oxidant suggested that this enzyme might not be
capable of physiological activity.20However, the 'H NMR study of the perform-
ance of this enzyme in R. capsulatus cells37 clearly demonstrated that W-
DMSOR is physiologically competent. The apparent difference can be explained
with reference to the relevant redox potential data. EPR potentiometric titra-
tions2' showed that the Wvl/Wv and Wv/WIv couples of W-DMSOR have
midpoint potentials of - 203 mV and - 105 mV (us. SHE) with each potential
being ca. 325 mV lower than that of the corresponding couple of Mo-DMSOR.~'
This difference defines the greater difficulty of reducing Wvlin the same enzyme
environment as MoV1and rationalises why oxidised (isolated) Mo-DMSOR is
reduced by DMS but oxidised (isolated)W-DMSOR is not. However, in uiuo, the
reduction of DMSOR by ubiquinol is mediated by the pentaheme c-type cyto-
chrome DorC, with midpoint potentials of - 34, - 128, - 184, - 185, and
-276 mV (us. SHE).38Therefore, DorC is capable of reducing oxidised W-
DMSOR, thereby allowing the protein to turnover inside a cell.
4 Chemistry Related to that of the Catalytic Centres of

Mo and W 0x0-transfer Enzymes
4.1 Towards the Synthesis of Molybdopterin
A general strategy for the synthesis of asymmetrically substituted dithiolene
ligands has been d e ~ e l o p e d ~ ' - (Figure
~' 6), since MPT (Figure 1) is such an
entity and virtually all dithiolene complexes previously investigated have in-
volved symmetrical ligands. A wide range of Ar groups have been incorporated
KSC(S)OPI‘ NaSC(S)NMe2
Ar Ar
S S
I H2S04 H2S04
I
Ar ps>=.
Ar
is- S- 4
OH-
Figure 6 General strategy for the synthesis of asymmetrically substituted dithiolenes
into this synthetic procedure, including quinoxaline derivatives, in which the

pyrazine ring has been reduced in a selective and controlled manner to produce a
stereochemistry that matches that of the ‘tetrahydro’-form present in the Mo and
W 0x0-transfer enzymes.42 An important aspect of this strategy has been the
formation of CpCo complexes to confirm the release of the dithiolene ligand
from the proligand, e.g. Figure These investigations have been ex-
7.39940,42
tended to substituted pteridine derivatives, involving substituents to ensure that

the compounds were reasonably soluble in the organic reaction media em-
p10ye d. ~~
In addition to addressing directly the significant challenge presented by the
synthesis of MPT using chemical procedures, these investigations are directed at
improving the understanding of the role of MPT in the catalyses accomplished
by the Mo and W 0x0-transfer enzymes. One intriguing aspect is to consider a
possible redox function for MPT. The protein crystallographic studies reported
for Mo and W enzyme^^*^'-^^ indicate that the pyrazine ring is at a ‘tetrahydro’-
level. However, the presence of the fused pyran ring means that the level of
oxidation of the pyrazine ring involved is actually equivalent to that of a
dihydro-pyrazine. This consideration leads to the suggestion that the potential
exists for a cooperation between the redox behaviour of Mo (or W) and that of
MPT. Specifically, this would involve changing the oxidation level of the pyra-
zine ring by opening and closing the pyran ring (Figure 8).44 This behaviour,
when transmitted to the metal via the dithiolene group, could complement the
redox changes of the metal that are required for the execution of catalysis.
C. David Garner 27 1
Me
. e
HO I H H
Bn02C
0
/I F P
-
CsOH
H20 CpCol,
*
I H H I H H
Bn02C Bn02C
Figure 7 Synthesis of a tricyclic compound, involving a CpCo centre bound to a

dithiolene group substituted onto a tetrahydropyran ring42
Figure 8 Conversion of a ‘tetrahydropterin’ to a dihydro-state by opening of the pyran

ring44
4.2 Dithiolene Complexes

The general synthetic strategy that produces asymmetrically substituted dithio-
lenes (Figure 6) has been utilised to prepare a range of [MO(dithiolene),12-
(M = Mo or W) complexes (Figure 9).45946The appropriate K,[MO,(CN),]
compound was employed as the starting material to prevent the synthesis of the
tris(dithio1ene) complex. Each of the [MO(dithiolene)J2 - complexes has a
square-pyramidal environment at the metal, as exemplified in Figure 10, and
involves essentially identical dimensions for corresponding Mo and W systems
(Table 1).
These complexes [MO(sdt),12- (M = Mo or W) represent minimal structural
analogues of the active sites of the Mo and W 0x0-transfer enzymes of the
DMSO reductase type (Figure 3). Furthermore, their redox properties show
similarities to those of the Mo and W centres of the enzymes (Figure 11). Thus,
each complex possesses a reversible [MO(dithiolene),] -/[MO(dithiolene),] -
couple and the potential of each WV/WIvcouple is ca. 225 mV lower than that of
the equivalent MoV/MoIVcouple; cf the potential difference of ca. 325 mV
observed for the MoV/MoIVand MoV1/MoVcouples, with Mo > W, for the
DMSO reductases of R. capsulatus (Section 3.2).20 Also, these systems MIv
complexes are converted to their Mvl counterparts by oxygen-atom transfer
from Me,NO and the [M0,(dithiolene),12- complexes are reduced to the
A r =
sdt
a 0
2-pedt 3-pedt
N9oly/
4-pedt qedt
M=MoorW
Me2N I'
NH,-ptedt NC(H)N(Meh-ptedt
Figure 9 Synthesis of [MO(dithiolene),12- ( M = M o or W) c ~ r n p l e x e s ~ ~ ~ ~ ~
Figure 10 Structure of the dianion in (PPh4)2[WO(sdt),].EtOH (the nature of sdt is

depicted in Figure 9); there is an H-bond between the W-4 and EtOH
groups46
corresponding [MO(dithiolene),] - by treatment with Ph,P. The redox and

spectroscopic properties of these systems are influenced by the nature of the
substituent on the dithiolene group, implying electronic communication via the
dithiolene groupand that M P T could play a role in modulating the properties of
the Mo or W centres of the 0x0-transfer enzymes.
5 Conclusions
Based on the knowledge presently available, it is interesting to speculate on the
different roles that have evolved for Mo and W in biological systems. A funda-
mental aspect is that, for an element to be involved in biology, it must be
C. David Garner 273
Table 1 A comparison of bond lengths (A) and angles (")forthe [M0(sdt),l2-

( M = M o or w)anions in their (PPh,)' salts.45746There are no
signijicant differences between the corresponding dimensions of the two
anions
[M oO(sd t),] - [ WO(sdt),] -
M=O 1.700(5) 1.724(7)

M-S 2.366(2)-2.385(2) 2.362(3)-2.383(3)
C=€ 1.325(9)-1.33(1) 1.32(2)
s-c 1.738(8)-1.774(7) 1.76(l t l.79( 1)
0-M-S 107.5(2)-110.7(2) 107.2(3)-110.5(3)
S-M-S 82.17(7)-86.92(7) 82.4(1k87.2(1)
Figure 11 Redox chemistry of complexes [MO(dithiolene),12- (M=Mo or W)4s346
available for incorporation. Starting from the premise that life evolved on this
planet, it is relevant to observe that both Mo and W are present in the Earth's
crust at a concentration of 1.5 ppm. Thus, both these elements are reasonably
abundant, but less so than Cu (55 ppm), Zn (70 ppm) and Fe (50000 ppm), metals
that are widely employed in biology. Despite the extensive similarities between
the chemistry of Mo and that of W, their geochemistry is quite different. Thus,
Mo occurs in the Earth's crust primarily as molybdenite, MoS,, whereas W
occurs as scheelite, CaWO,, and wolframite, (Fe, Mn)W0,.8
Mo and W are generally taken up by organisms as the [M0,l2- ion. Bacteria
have efficient systems for [Mo0,12- uptake and transport; the transport pro-
teins bind these anions by a series of H-bonds to the polypeptide chain and the
site is suitable for binding both [Mo0,12- and [W0,]2-.47 Therefore, the
relative availability of [Mo0,12- and [W0,l2- will be an important factor in
determining whether Mo or W is incorporated by an organism. Thus, R. cap-
sulatus (Section 3.1) produces a physiologically competent enzyme with W
replacing Mo under [WO,]'--rich, [Mo0,I2--depleted condition^.^^*^^ How-
ever, in a growth medium that contains an equal concentration of [Mo0,12-
and [W0,l2-, R. capsulatus shows only a modest preference for incorporating
Mo over W. One important caveat to emerge from this study was that a trace of
Mo is required for cell growth;20the essential role that Mo is required to fulfil is
unknown, but it would appear that W is not capable of performing this function.
The significantly greater presence of Mo over W at the catalytic centre of
enzymes can be understood by noting the relative concentrations of [Mo0412-
and [WO4I2- in seawater. Thus, [Mo0412- is present at the relatively high
concentration of 1 x lo-, mg dmP3,ca. 100 x the concentration of [W0412-.
Thus, when these metals are taken up from an aqueous medium where this
concentration difference applies, [Mo0,12 - would be expected to be incorpor-
ated by an organism in preference to [WO,]”-. However, the 100-fold excess of
[Mo0,12- over [WO4I2- in seawater has (probably) not always been the case.
MoS, is insoluble in water and, therefore, it would appear that Mo was relatively
unavailable for involvement in the early stages of evolution of life on Earth.
However, in an oxidising atmosphere (i.e. post photosynthesis) MoS, is con-
verted to [Mo0,12- (3).
2MoS, + 7 0 , + 2H,O --+ 2[Mo0,I2- + 4 s 0 2 + 4H’ (3)
[W0,l2-, arising from the dissolution of oxide sources, has probably always
been available in seawater.
These observations offer some rationalisation for the association of W with
primitive organisms and the significantly greater involvement of Mo in biologi-
cal systems currently observed. However, the differences in the chemical proper-
ties of Mo and W are likely to be important in determining the roles of these
metals in biological systems. Given the similarities in the size of corresponding
chemical systems of these two elements,8 discrimination between them is likely to
be achieved on the basis of redox potentials and/or bond strength consider-
ations. Some such differences have emerged from the assays of the activity of
Mo-DMSOR us. W-DMSOR (Section 3.2). W-substituted R. capsulatus cells
turn over at a slower rate than their Mo counterpart^,^^ even though reduction
of the substrate by the isolated enzymes proceeds at a significantly (ca. 17 times)
faster rate for W-DMSOR than Mo-DMSOR.~’Thus, for this system at least,
reduction of the oxidised state (MV1)to the reduced state (MIV)would appear to
contain the rate-determining step of the catalytic cycle employed by these
enzymes (Figure 12). This latter aspect is consistent with the preference W > Mo
for the VI oxidation state and the generalisation that W 0x0-transfer enzymes
catalyse reactions that have a low redox potential whereas Mo 0x0-transfer
~
enzymes operate at a higher redox
My involvement with this topic has spanned some thirty years. During this
time, considerable progress has been made in the development of our under-
standing of the nature and function of the Mo and W 0x0-transfer enzymes.
However, many significant challenges remain and, to address these successfully,
it will be necessary for the present pattern of interdisciplinary research to
continue. Thus, geneticists, biologists, biochemists, chemists, and physicists need
to maintain their synergic interactions that have long characterised this field. A
major goal will be to define the coordination chemistry of the catalytic centres of
these enzymes, to a standard that would be acceptable to Joseph Chatt - a
formidable task!
C. David Garner 275
Figure 12 Diagrammatic representation of the catalytic cycle of DMSO reductase in

Rhodobacter capsulatus, with either M o or W a t the active site and DMSO
or T M A O as the electron acceptor; reduction of the oxidized enzyme by
ubiquinol is mediated by the pentaheme c-type cytochrome DorC
6 Acknowledgements
I am very honoured by the award of the Joseph Chatt Medal and am pleased to
be associated with this great scientist, who made so many significant contribu-
tions to coordination chemistry. In respect of the research described above, I
wish to thank all of my colleagues who have made distinctive and important
contributions. My interest and involvement with Mo and W chemistry has been
sustained by stimulating collaborations of many friends, particularly David
Collison, John Joule, Frank Mabbs (at Manchester), John Enemark (at Tucson)
and Stephen Davies and Jon McMaster (at Nottingham). I gratefully acknowl-
edge the important contributions to the research described in this article made
by Professor Gareth Morris and Drs Elaine Armstrong, Michael Austerberry,
Jacqui Birks, Ben Bradshaw, Stephen Boyde, Sue Bailey, John Charnock, An-
drew Dinsrnore, Arefa Docrat, Richard Ellis, David Rowe, Lisa Stewart and
Clare Wilson.
7 References
1 R. Hille, Chem. Rev., 1996,96, 2757.
2 M. K. Johnson, D. C. Rees and M. W. W. Adams, Chem. Reu., 1996,96,2817.
3 W. R. Hagen and A. F. Arendson, Structure and Bonding, 1998,90,161.
4 H. Bortels, Arch. Mikrobiol., 1930,1, 333.
5 F. 0.Bryant and M. W. W. Adams, J . Biol. Chem., 1989,264,5070.
6 J. L. Johnson and K. V. Rajagopalan, Proc. Natl. Acad. Sci., USA, 1982,79,6856; J. C.
Hilton and K. V. Rajagopalan, Arch. Biochevn. Biophys., 1996,325,139.
7 M. K. Chan, S. Mukund, A. Kletzin, M. W. W. Adams and D. C. Rees, Science, 1995,
267,1463.
8 N. N. Greenwood and A. Earnshaw, The Chemistry ofthe Elements, Pergamon Press,
Oxford, 1984, p 1170.
9 R. N. F. Thorneley and D. J. Lowe, JBIC, 1996,1,576.
10 H. H. Nagatani and W. J. Brill, Biochim. Biophys. Acta, 1974,362,160.
11 R. H. Holm and E. D. Simhon Molybdenum Enzymes, ed. T. G. Spiro, John Wiley and
Sons, New York, 1985, pp 1-87.
12 C. D. Garner and R. M. Miller, J . Chem. Soc., Dalton Trans., 1981,1664.
13 J. Chatt, J. R. Dilworth and R. L. Richards, Chem. Rev., 1978,78,589.
14 K. V. Rajagopalan, Adv. Enzymol., 1991,64,215.
15 J. L. Johnson and S. K. Wadman, ‘Molybdenum Cofactor Deficiency,’in The Meta-
bolic Basis oflnherited Disease, 61h Edition, eds. C. R. Scriver, A. L. Beaudet, W. Sly
and D. L. Vallee, McGraw-Hill, New York, 1989,1463-1475.
16 D. J. Richardson and N. J. Watmough, Curr. Opin. Chem. Biol., 1999,3,207.
17 B. M. Rosner and B. Schink, J . Bacteriol., 1995,177,5767.
18 R. A. Schmitz, M. Richter, D. Linder and R. K. Thauer, Eur. J . Biochem., 1992, 207,
559; P. A. Bertram, R. A. Schmitz, D. Linder and R. K. Thauer, Arch. Microbiol., 1994,
161,220; P. A. Bertram, M. Karrasch, R. A. Schmitz, R. Bocher, S. P. J. Albracht and
R. K. Thauer, Eur. J . Biochem., 1994,220,477.
19 J. Buc, C.-L. Santini, R. Giordani, M. Czjek, L.-F. Wu and G. Giordani, Mol.
Microbiol., 1999,32, 159.
20 L. J. Stewart, S. Bailey, B. Bennett, J. M. Charnock, C. D. Garner and A. S. McAlpine,
J . Mol. Biol., 2000,299, 593.
21 H. Schindelin, C. Kisker, J. Hilton, K. V. Rajagopalan and D. C. Rees, Science, 1996,
272,1615; F. Schneider, J. Lowe, R. Huber, H. Schindelin, C. Kisker and J. Knablein,
J . Mol. Biol., 1996,263, 53.
22 A. S. McAlpine, A. G. McEwan, A. L. Shaw and S. Bailey, JBIC, 1997,2,690.
23 A. S. McAlpine, A. G. McEwan and S. Bailey, J . Mol. Biol.,1998,275,613.
24 H. K. Li, C. Temple, K. V. Rajagopalan and H. Schindelin, J . Am. Chem. SOC.,2000,
122,7673.
25 R. C. Bray, Quart. Rev. Biophys., 1988,21,299.
26 B. E. Schultz, R. Hille and R. H. Holm, J . Am. Chem. Soc., 1995,117,827.
27 N. R. Bastian, C. J. Kay, M. J. Barber and K. V. Rajagopalan, J . Bid. Chew., 1991,266,
45.
28 A. G. McEwan, S. J. Ferguson, J. B. Jackson, Biochem. J., 1991,274,305.
29 T. Sat0 and F. N. Kurihara, J . Biochem., (Tokyo), 1987,102,191.
30 C.-L. Santini, B. Ize, A. Chanal, M. Miiller, G. Giordano and L.-F. Wu, EMBO J.,
1998,17, 101.
31 B. Bennett, N. Benson, A. G. McEwan and R. C. Bray, Eur. J . Biochem., 1994, 225,
321.
32 G. N. George, J. Hilton, C. Temple, R. C. Prince and K. V. Rajagopalan, J . Am. Chem.
SOC., 1999, 121, 256.
33 E. Diemann and A. Muller, Coord. Chem. Rev., 1973,10,79.
34 B. Adams, A. T. Smith, S. Bailey, A. G. McEwan and R. C. Bray, Biochemistry, 1999,
38, 8501.
35 G. F. King, D. J. Richardson, J. B. Jackson and S. J. Ferguson, Arch. Microbiol., 1987,
149,47.
36 D. J. Richardson, G. F. King, D. J. Kelly, A. G. McEwan, S. J. Ferguson and J. B.
Jackson, Arch. Microbiol., 1988,150, 131.
C. David Garner 277
37 L. J. Stewart, S. Bailey, D. Collison, G. A. Morris, I. Preece and C. D. Garner, Chem.
Biochem., 2001,2, 703.
38 A. L. Shaw, A. Hochkoeppler, P. Bonora, D. Zannoni, G. R. Hanson and A. G.
McEwan, J . Biol. Chem., 1999,274,9911.
39 R. L. Beddoes, A. Dinsmore, M. Helliwell, C. D. Garner and J. A. Joule, Acta
Crystallogr., Sect. C, 1997,53, 213.
40 A. Dinsmore, J. H. Birks, C. D. Garner and J. A. Joule, J . Chem. Soc., Perkin Trans. I ,
1997,801.
41 A. Dinsmore, C. D. Garner and J. A. Joule, Tetrahedron, 1998,54,3291 and 9559.
42 B. Bradshaw, A. Dinsmore, C. D. Garner and J. A. Joule, Chem. Commun., 1998,417.
43 B. Bradshaw, D. Collison, C. D. Garner and J. A. Joule, Chem Commun., 2001,123.
44 D. Collison, C. D. Garner and J. A. Joule, Chem. Soc. Rev., 1996,25,25.
45 E. S. Davies, R. L. Beddoes, D. Collison, A. Dinsmore, A. Docrat, J. A. Joule, C. R.
Wilson and C. D. Garner, J . Chem. Soc., Dalton Trans., 1997,3985.
46 E. S. Davies, G. M. Aston, R. L. Beddoes, D. Collison, A. Dinsmore, A. Docrat, J. A.
Joule, C. R. Wilson and C. D. Garner, J . Chem. SOC.,Dalton Trans., 1998,3647.
47 D. M. Lawson, C. E. Williams, D. J. White, A. P. Choay, L. A. Mitchenhall and R. N.
Pau, 1. Chem. Soc., Dalton Trans., 1997,3981; D. M. Lawson, C. E. M. Williams, L. A.
Mitchenhall and R. N. Pau, Structure, 1998,8,1529.
Iron-hide Clusters and
Nitrogenase: A biological
Chemistry of Biological
Relevance?
SONNY C. LEE
Department of Chemistry, Princeton University, Princeton, New Jersey 08544,
USA
1 Nitrogenase: Problem and Approach

The biological reduction of dinitrogen to ammonia is accomplished by the
nitrogenase enzymes through chemistry that, despite decades of study, continues
to defy description at the atomic level.*p5The site of nitrogen reactivity within the
enzyme is almost certainly an unusual iron-sulfur (Fe-S) cluster that can also
contain a heterometal (either molybdenum or vanadium) depending on the
enzyme variant. The Mo-containing enzymes are the best characterised, with
macromolecular crystal structures of the Mo-nitrogenases providing the most
detailed molecular visualisation to date of the catalytic metallocluster (the
iron-molybdenum cofactor, FeMoco, Figure 1) in a resting ~ t a t e . ~ - ~
The nitrogenase enzymes present formidable challenges to molecular under-
standing.lp5 The enzyme cannot be studied in a poised ‘ready’ state; in the
absence of dinitrogen substrate, protons from water are reduced to dihydrogen
and the nature of the substrate-reducing form remains a mystery. Under turn-
over, the enzyme system undergoes a bewildering series of events: protein dock-
ing and release, electron transfer and storage, ATP hydrolysis, proton transfer
and multisite substrate binding and release with the various enzyme states
populated concurrently under experimental conditions; enzyme kinetics are
therefore complex, and active enzyme intermediates (and inhibited forms) are
generally ill-defined. FeMoco itself is structurally complex and unique in Fe-S
chemistry; moreover, the resting state structure may not reflect important as-
pects of the reactive species. The cofactor cluster can be extracted intact from the
Sonny C. Lee 279
H
/S \
FeMoco
Figure 1 Iron-molybdenum cofactor structures in resting (left)and hypothetical

dinitrogen-bound states (right)
native protein, but in this condition shows only limited ligand substitution
chemistry and no substrate reactivity." Because of these (and other) constraints,
oxidation state assignments for FeMoco remain contentious, the necessary
reaction stoichiometry for substrate reduction is uncertain and the exact loca-
tion of dinitrogen-binding is unknown.
The dearth of concrete molecular data has not discouraged mechanistic
debate. The protein-derived FeMoco structure in particular has inspired specu-
lation that dinitrogen is activated while bound as a bridging ligand to multiple
iron centers (Figure This conjecture is drawn primarily from two
l).7911912
observations: (1)the unusual structure of FeMoco, with six central iron sites that
appear three-coordinate and anomalously unsaturated; and (2) the probable
existence of an iron-only nitrogenase, indicating that the heterometal is unnec-
essary for dinitrogen activation. This hypothesis has been bolstered by computa-
tional models that also implicate dinitrogen interaction at the central iron
positions, although the exact binding mode differs from study to
Experimental support for this proposal, however, is unavailable from either the
enzyme system or from synthetic models.
We are investigating the fundamental cluster chemistry of the Fe-N bond to
define possible interactions of iron and nitrogen during biological nitrogen
fixation. Our strategy is loosely constrained; rather than seeking high-fidelity
models, we focus instead on minimal complexes with key features congruent with
those of the biological system. Thus, we have sought clusters of low-coordinate,
weak-field/high-spin FeIdlI1that incorporate nitrogen core ligands and/or me-
diate the multielectron reduction of N-N bonds. Through the study of these
congruent complexes, we hope to establish chemistry equivalent to that which
occurs, so far unobserved, in the natural system. Our initial progress using this
approach is highlighted here.
280 Iron-lmide Clusters and Nitrogenase: Abiological Chemistry of Biological Relevance?
2 Iron@) Reduction of the N-N Bond
Hydrazine N-N bond reduction is a transformation that can lend insight into
the final bond cleavage of nitrogen fixation. Under reducing conditions where
hydrazines are neutral or protonated, simple iron-sulfur systems, including
[Fe,S,] cubanes, are unable to reduce the N-N bond, whereas heterometallic
[MFe,S,] cubanes (M = Mo or V) can.17 This has led to the proposal that the
final N-N bond reduction occurs at the molybdenum site in FeMoco.17 How-
ever, the identities, and therefore protonation states, of various postulated
intermediates during biological nitrogen fixation are unknown. A plausible
alternative scenario could involve the multiple-bond reduction proceeding
through anionic-forms of the partially reduced dinitrogen, for example, hydra-
zide(2- ), with bond cleavage occurring prior to final protonation.
Fe-N'
1 I
r i i s N--Fe
Dh
I Ar
QS
/
Ph' I,N-l-Fe, / SAr
Fe-N
/ \
ArS Ph
3 4
(a) 1 ArSH; (b) thf (Ar = mesityl); (c) 1 ArSH; (d) 0.5 PhN(H)-N(H)Ph
Scheme 1
Our initial attempt to explore this possibility is outlined in Scheme l.'* Iron@)
bis(amide) precursor 1 serves as a synthetic starting point. The bis(trimethy1-
sily1)amide ligands perform two primary functions, providing steric control of
the high-spin metal environment and acting as a strong latent base for protolytic
metathesis chemistry; this last r6le allows incorporation of highly basic ligands
such as hydrazide without the use of aggressive 'free' anions. Introduction of one
equivalent of ortho-substituted aryl thiol allows the selective metathesis of amide
for thiolate without disproportionation to bis(thio1ate)2 and starting bis(amide)
1; the heteroleptic iron@)dimer 3 can be isolated when this reaction is conducted
in coordinating thf solvent. Addition of 0.5 equivalents of 1,2-diarylhydrazine to
a solution of 3 results in the formation of the tetrameric imidoiron cluster 4 in
good isolated yield. This cluster consists of four tetrahedral iron(rr1) centres, each
terminally ligated by one thiolate and bridged through three ,uu,-arylimidenitro-
gen atoms to form a heterocubane structure. Cluster 4 is paramagnetic with an
S = 2 ground state; a high-spin iron(r1r) formulation for the tetrahedral iron sites
was confirmed by Mossbauer measurement.
The formation of cubane 4 formally proceeds via the two-electron reduction of
the hydrazine N-N bond coupled to concomitant one-electron oxidation of Ferl
Sonny C. Lee 28 1
to FelI1and cluster assembly. To our knowledge, this is the first example of a
hydrazine cleavage mediated by Fell/Ferxloxidation. With respect to proposed
Fe-mediated nitrogenase mechanisms, this process offers a possible pathway for
the final reduction of core-bound dinitrogen substrate prior to ammonia release,
and therefore provides an opportunity to explore fundamental reaction chemis-
try of potential relevance to biological nitrogen fixation. The bridging core
ligands in cluster 4 are readily accessible, the addition of free p-toluidine to 4
resulting in the rapid (minutes) exchange of core ligands with retention of cluster
structure; in contrast, low-spin iron carbonyl imide clusters are essentially inert
to the exchange of their imide bridges.lg Detailed reactivity and mechanistic
studies are in progress.
3 Iron-Imide Cluster Chemistry

The synthesis of the imidoiron cubane 4 and the possibility of cluster-bound
imide ligands as intermediates during nitrogenase turnover has led us to a
general exploration of weak-field iron-imide (Fe-NR) cluster chemistry. Our
initial studies establish the existence of a rich reaction manifold that is readily
accessible from simple iron(Ir1)precursors.20-22A recent report from Fenske and
co-workers describes additional synthetic routes to Fe-NR clusters,23 further
demonstrating the generality of the chemistry.
cI\ ,But 1'-

+ +
5 7 6
thf, 80 "C
I
FeCI3 + 2 Li(NHBu')
1 toluene
7 6 8
Scheme 2
Fe-NR clusters can be synthesised directly from anion metathesis reactions of
iron(@ chloride and lithium t-butylamide, although reaction outcomes are
complex and sensitive to minor changes in reaction conditions (Scheme 2). Thus,
the reaction of FeCl, with two equivalents of Li(NHBu') in thf at 80°C results in
the formation of three characterised species,20 the major product being the
reduced 1Fe11/3Fe111 cubane 6. Two neutral products are also isolated in limited
yield, the all-iron(IIr)chloroimidocubane 5 and a C3", symmetric, site-differenti-
ated cubane 7. The latter cluster possesses three terminal chloride ligands and
one terminal t-butylimide. (Figure 2). The terminal Fe-NR moiety is nearly
linear (178.6(3)")with an Fe-N distance (1.635(4)A)characteristic of a multiple
bond. The neutral cluster possesses formally an oxidized 3Fe111/lFeIVcore; zero-
field Mossbauer analysis of polycrystalline material revealed a 3: 1 ratio of
valence-localised Fe1Ii/FelVsites at - 123"C, firmly establishing the oxidation
state assignment and therefore the presence of the terminal imide group in cluster
7. This is the only example of a characterised terminal imide ligand on iron. In
marked contrast to the well-studied, highly-reactive isoelectronic ferry1
([Fe=0l2') moiety, the terminal imide functionality in 7 is notable for surprising
stability that probably arises from the steric shielding afforded by the t-butyl
substituents; in addition, the potent nitrogen anion donors facilitate the
stabilisation of high-oxidation state FeIV.Note that multiply-bonded metal-
nitrogen groups (including the terminal metal-imide functionality) are essential
intermediates in the classic Chatt nitrogen fixation cycle;24 this pathway, for-
mulated from studies of low-oxidation state mononuclear phosphine complexes
of molybdenum and tungsten, represents yet another possible route for biologi-
cal dinitrogen reduction.
When the self-assembly reaction is conducted in toluene, a different product
distribution is obtained.21Neutral cluster products are still formed in low yields;
however, terminal imide cluster 7 is now the predominant neutral species, with
chlorocubane 5 virtually absent as judged by 'H NMR spectroscopy. The
Figure 2 The structure of [Pe,(p3-NBu')C1,], 7

Sonny C. Lee 283
majority of the iron-containing reaction mass is divided equally between two
anionic products, reduced cubane 6 and trinuclear cluster 8. The structure of this
new cluster is formally derived from cubane cluster 6 by removal of one iron(II1)
site (Figure 3); bond distances and nitrogen geometries reveal one p-imide ligand,
one p3-imide, and two p-NR bridges protonated as amides. Although we do not
yet understand the nature of the solvent effect on cluster formation, cluster 8 may
be an intermediate in the assembly pathway of the cubane core.
We have also explored alternative synthetic routes to Fe-NR clusters using
hindered iron(II1)amide precursors and protolytic metathesis chemistry (Scheme
3).22 Protolytic metathesis of the monomeric heteroleptic precursor 9 with one
equivalent of tert-butylamine results in the self-assembly of neutral cubane 5 in
modest (ca. 25%) yield. The equivalent reaction with aniline, however, produces
more complex chemistry, with formation of the double cluster salt 10 and
azobenzene. The cation structure in 10, comprised of two facially bridged octa-
hedral iron centres, has been reported previously as a diiron(I1) m o n ~ c a t i o n , ~
and we believe the same assignment is applicable here; the reduction to the
iron(@ state is presumably linked to the oxidative coupling that generates
azobenzene. This charge assignment dictates a 2 - charge for the accompanying
trinuclear cluster, which is formed from a linear array of edge-fused tetrahedral
iron centres with p-imide bridges. Formal oxidation states of 2Fe111/lFeIVare
required for the trinuclear dianion; we have isolated and structurally character-
ised the dilithium salt of this dianion (11, see below), providing support for the
high oxidation state formulation.
The presence of lithium chloride alters the chemistry, presumably by coor-
dination to intermediary species during cluster self-assembly. Thus, the addition
of one equivalent of LiCl to 9 yields the chloride adduct 12; further treatment
with one equivalent of t-butylamine leads to neutral cubane 5, but at a much
reduced rate compared to the reaction without LiCl. The reaction with aniline is
also retarded in the presence of LiC1, and the formation of azobenzene is
significantly diminished. We have identified three products from this reaction
Figure 3 The structure of [ F e , ( ~ - N H B u ' ) , ( ~ - N B u t ) ( ~ - ~ B u t-) C

8 1,l
FeCI3
(Me3Si),N
7- I
(Me3Si),N
;e.,...cI
/ 'CI
12
\d
\
(Me3Si)2N,
la
T
/ CI'
5
'B u
My 12-/ 9
PhN=NPh +
CI,
CI
F.e,
,N\
F
,e:,
CI
CI thf\ jC! ,thf PI Ph
1
Th
N ~
1
2-
I I0
Mes 15
mi
1-1 Ph plh
11 13 14
(a) 2 Na[N(SiMe,),]. thf; (b) 1 'BuNH,, thf; (c) 1 PhNH,, thf; (d) 1 LiCI, thf;
(e) 1 LICI, 1 PhNH,, thf; (f) 1 LiCI, 1 MesNH2,thf
Scheme 3
system, all of which are dianions: the linear trimeric cluster 11 noted previously;
the phenylimidocubane 13, with formal oxidation states of 2Fe"/2Fe1I1;and the
dimer 14, composed of edge-fused tetrahedral iron(m) centres bridged by imide
ligands. The principal product of this reaction is diiron(m) 14, which is consistent
with the suppression of redox chemistry evident in the decreased azobenzene
yield.
The nature of the imide substituent can also influence cluster assembly. For
example, treatment of 9 with LiCl and mesidine (2,4,6-trimethylaniline,
MesNH,) leads cleanly to the synthesis of iron(m) dimer 15 in high yield without
detectable formation of other clusters or of azomesitylene; the high selectivity in
this case probably stems from the steric bulk of the aryl substituent which
inhibits aggregation past the dimeric stage. In the absence of LiC1, the reaction
with mesidine gives a complex product mixture including azomesitylene.
4 Analysis and Speculation

Our preliminary survey of Fe-NR cluster chemistry reveals strong parallels
between imide and sulfide ligation on high-spin iron.22The dinuclear, trinuclear,
and tetranuclear geometries found in imidoiron clusters all have structural
counterparts in fundamental biologically-relevant Fe-S cluster motifs (Figure
Sonny C. Lee 285
Structural Parallels
2-
(CYS)SG, ,s\ S(Cys)
2-Fe Ferredoxin Fe
(Cys)S' yFe', ",S(Cys) 14,15
I
Ar
3-Fe Ferredoxin
(denatured)
3-Fe Ferredoxin a
4-Fe Ferredoxin
FeMocofactor [Co8(NPh)o(PPh&]'- 16
Chemical Parallels
Ar(H)N-N(H)Ar + 2 e - - 2ArN2- + 2 H+ (1)
PhN=NPh + 4 e -
RS-SR
SO
+ 2e-
+ 2 e-
- -
-
___L 2ArN2-
2RS-
[sCl2- s2-
(2)
(3)
(4)
Figure 4 Parallels between weak-jield Fe-S and Fe(Co)-NR chemistry
4).26The occurrence of protonated bridging amides in 8 may be analogous to the

behaviour of three-iron ferredoxins that undergo protonation upon reduction,
presumably at bridging sulfide^.^ The correspondence between Fe-NR and
Fe-S chemistry extends to the redox activity of the core ligands, where the two-
electron reduction of the N-N bond and the four-electron oxidative formation of
the N-N bond are fully equivalent to sulfur-based half-reactions prevalent in
286 Iron-Imide Clusters and Nitrogenase: Abiological Chemistry of Biological Relevance?
Fe-S chemistry (Figure 4, Equations 1-4).28 Some differences are also evident,
the most obvious being redox effects arising from the stronger donor ability of
nitrogen anions; this is manifested in the generally higher mean oxidation states
of Fe-NR clusters relative to their Fe-S analogues.
The relationship between imide and sulfide chemistry is reinforced by the
imidocobalt cluster 16, recently reported by Link and F e n ~ k ewhich
, ~ ~ is striking
in its resemblance to FeMoco (Figure 4). For weak-field environments,
metal-metal bonding and valence electron count are not thought to dictate
cluster geometry; instead, properties such as oxidation state, ligand environment,
and net charge are better indicators of structure. Thus, in weak-field metal-sulfur
chemistry, CoII is the prevalent oxidation state in tetrahedral geometries and
structural homologues of Co-S and Fe-S clusters occur when mean oxidation
states of I1 are approached.22Because of the parallel relationships between imide
and sulfide, and between tetrahedral Co" and Fe", we expect that an Fe-NR
analogue of the FeMoco should exist as a predominantly iron@) cluster; success-
ful synthesis of this analogue cluster will advance our understanding of nitro-
genase cofactor chemistry.
With respect to conjectured iron-mediated mechanisms of nitrogen fixation,
our studies reveal the intrinsic chemistry of nitrogen anions in weak-field iron
environments. High-spin, tetrahedral FeI1/II1centres can mediate nitrogen redox
chemistry at the N-N and N=N levels, and fully reduced, anionic nitrogen
fragments are readily incorporated as accessible core ligands of clusters
homologous to known Fe-S geometries. The possible convergence of this chem-
istry with equivalent, hypothetical events in nitrogenase offers an intriguing
starting point for further inquiry.
5 Acknowledgements
I am grateful for the contributions of my co-workers whose names appear in the
references. This work was made possible by the Beckman Foundation (Beckman
Young Investigator Award) and the National Science Foundation (CAREER
CHE-9984645).
6 References
1 D. C. Rees and J. B. Howard, Curr. Opin. Chem. Biol., 2000,4,559.
2 B. E. Smith, Adv. Inorg. Chem., 1999,47,159.
3 D. C. Rees and J. B. Howard, Chem. Rev., 1996,96,2965.
4 B. K. Burgess and D. J. Lowe, Chem. Rev., 1996,96,2983.
5 R. R. Eady, Chem. Rev., 1996,96,3013.
6 J. Kim and D. C. Rees, Nature, 1992,360,553.
7 M. K. Chan, J. Kim and D. C. Rees, Science, 1993,260,792.
8 T. Bolin, N. Campobasso, S. W. Muchmore, T. V. Morgan and L. E. Mortenson,
Molybdenum Enzymes, Cofactors, and Model Systems, American Chemical Society,
Washington, DC, 1993, p. 186.
9 S. M. Mayer, D. M. Lawson, C. A. Gormal, S. M. Roe and B. E. Smith, J . Mol. Biol.,
Sonny C. Lee 287
1999,292,871.
10 B. E. Smith, M. C. Durrant, S. A. Fairhurst, C. A. Gormal, K. L. C. Gronberg, R. A.
Henderson, S. K. lbrahim, R. L. Gall and C. J. Pickett, Coord. Chem. Rev., 1999, 186,
669.
11 I. Dance, J . Biol. Inorg. Chem., 1996,1, 581.
12 D. Sellmann and J. Sutter, J . Biol.Inorg. Chem., 1996,1, 587.
13 T. H. Rod and J. K. Narrskov, J . Am. Chem. Soc., 2000,122,12751.
14 P. E. M. Siegbahn, J. Westerberg, M. Svensson and R. H. Crabtree, J . Phys. Chem. B,
1998,102,1615.
15 K. K. Stavrev and M. C. Zerner, Int. J . Quant. Chem., 1998,70,1159.
16 I. Dance, Chern. Commun., 1997,165.
17 D. Coucouvanis, J . Biol.Inorg. Chem., 1996,1, 594, and references therein.
18 A. K. Verma and S. C. Lee, J . Am. Chem. Soc. 1999,121,10838.
19 For example: F. Ragaini, J.-S. Song, D. L. Ramage, G. L. Geoffroy, G. P. A. Yap and A.
L. Rheingold, Organometallics, 1995,14, 387.
20 A. K. Venna, T. N. Nazif, C. Achim and S. C. Lee, J . Am. Chem. Soc., 2000,122,11013.
21 M. C. Traub, A. K. Verma and S. C. Lee, unpublished work.
22 T. N. Nazif, J. S. Duncan, A. K. Verma and S. C. Lee, submitted for publication.
23 H. Link, A. Decker and D. Fenske, Z. Anorg. Allg. Chem., 2000,626,1567.
24 C. J. Pickett, J . B i d . Inorg. Chem., 1996,1,601.
25 Z. Janas, P. Sobota and T. Lis, J . Chem. Soc., Dalton Trans., 1991,2429.
26 H. Beinert, R. H. Holm and E. Munck, Science, 1997,277,653.
27. K. S. Chen, J. Hirst, R. Camba, C. A. Bonagura, C. D. Stout, B. K. Burgess and F. A.
Armstrong, Nature, 2000,405,814.
28 For example, K. S. Hagen, A. D. Watson and R. H. Holm, J . Am. Chem. Soc., 1983,105,
3905.
29 H. Link and D. Fenske, Z. Anorg. Allg. Chem., 1999,625,1878.
Determinants of the Reduction
Potential in Rubredoxins, the
Simplest Iron-SulJiur Electron-
transfer Proteins
ZHIGUANG XIAO AND ANTHONY G. WEDD
School of Chemistry, University of Melbourne, Parkville, Victoria 3010,
Australia
1 Introduction
Professor Joseph Chatt guided one of us (AGW) during a postdoctoral period in
1970-71. As Director of the ARC Unit of Nitrogen Fixation, he emphasised the
importance of connections across chemistry, biochemistry, microbiology and
genetics. As a chemist, he taught that synthesis of simple species could illuminate
the structure and reactivity of more complex systems and, in particular, the
nitrogenase enzyme. His influence, and that of his associates Jeff Leigh and Ray
Richards, upon the authors’ approach to the chemistry of 0x0-molybdenum
centres has been outlined briefly.’
Recent developments in molecular genetics and chemical synthesis have
brought a new sophistication to modelling of protein chemistry: the biomolecule
becomes its own model. For example, variation of amino acid residues via site-
directed mutagenesis allows variation of the ligands at a metal active site. Joe’s
eyes would have sparkled at the possibilities, especially that of rational design of
synthetic proteins containing sites of desirable catalytic He would
also have delighted in the continuing crucial role of experiment. For example, the
nature of the blue copper centre is understood in exquisite detail4 but attempts to
construct an artificial blue copper site encountered a number of unanticipated
experimental o b ~ t a c l e sThe
. ~ resultant exploration indicated that the tetrahedral
stereochemistry for CulI is achieved by destabilising the more favoured tetra-
Zhiguang Xiao and Anthony G. Wedd 289
(a 10 20 30 40 50
MKKMCTVCG YIYNPEDGDP DNGVNPGTDF KDIPDDWVCP LCGVGKDQFE EVEE
Figure 1 Structural representations of the Fe"'-rubredoxin from Clostridium

pasteurianum.
(a) Backbone (the Fe atom is represented by The pseudo-two-fold axis (see
0).
text) is in the plane of the page, passing though the Fe atom; (b) N H - . S
interactions (- - - -) around the Fe(S-Cys), centre in RdCp (generated from the
coordinates of pdb5rxn.ent in the Brookhaven Protein Databank). The
pseudo-two-fold axis is now perpendicular to the page, passing though the Fe
atom; (c)Amino acid sequence.
gonal arrangement rather than by imposing peptide-induced 'strain'. Exclusion

of solvent water is a key determinant.
A vigorous campaign is underway to understand the role of iron-sulfur
proteins in all their beautiful structural manifestations.6p8The present contribu-
tion reviews factors that define the reduction potential of the simplest of the
iron-sulfur electron transfer proteins, rubredoxin (Rd) which features an
Fe(S-Cys), site of tetrahedral geometry (Figure 1; residue numbering* is for Rd
from Clostridiurn pasteurianurn, RdCp, and the same numbering scheme for
-
residues is applied to the other Rds discussed in the text). It has the advantages of
being small ( 6000 Da) and robust enough to endure substitution of both metal
and ligands, as well as significant variation in the secondary coordination
This appears to be related to the presence of a hydrophobic core
adjacent to the metal site plus 17 charged surface residues that ensure high
solubility.
The structural parameters of the active site have been defined to good preci-
~ i 0 n . l The
~ ' ~active site forms a 'knuckle' sitting above the hydrophobic core of
the protein (Figure l(a)). The atoms of the p-loops of protein (residues 5-11,
38-44) which carry the cysteine ligands exhibit a pseudo two-fold symmetry
which includes six NH-S interactions (Figure 1, (a) and (b)). The symmetry-
related 'surface' ligands Cys9 and Cys42 are in a different environment from the
* For better readability, equivalent sequence positions for other Rds quoted in the text have been
changed to those of RdCp
290 Determinants of the Reduction Potential in Rubredoxins
'interior' ligands Cys6 and Cys39. In addition, mutation at one residue, at its
symmetry-related partner and, simultaneously, at both residues (the double
mutant) provides three closely related mutant proteins for comparison with the
native form.
2 Determinants of Reduction Potential

The reduction potential E"' of a protein active site P is defined by Equations
(1)-(3). It is determined by its 'ionisation energy', I E , governed by electronic
structure, and by its 'solvation energy', U , governed by reorganisation of the
active site environment upon reduction [Equation (2)].6
P" + ne- + p ( x - n ) (1)
AG"',, = - nFE"' = AH"',, - TAS"',, (2)
E"' =IE +U (3)
Note that Equations 2 and 3 each express E"' as a function of two variables. In
general, both I E and U will contain enthalpic and entropic contributions.
Major determinants of I E are the geometry of the ligand field and the donor
strength of the ligands that together determine the energy of the redox-active
orbitals and define the effective nuclear charge, Zeff.Low (negative) potentials
would be expected for the Fe1I1/FeI1couples of tetrahedral alkyl thiolate com-
plexes [Fe(SR),]' - 1 2 - as the strongly electron-donating ligands preferentially
stabilise the oxidised FelI1state. However, their potentials are very low indeed
(e.g. [Fe(SEt),l1 - I 2 - in MeCN: E"', - 1320 mV versus NHE)." This may be a
consequence of strong spin polarisation in the d5 configuration of the Felt'
form.6
The observed potentials of the native proteins of the family of simple ru-
bredoxins in aqueous solution cover a range of about 140 mV: + 60 to - 80 mV
versus NHE (Table 1).These are shifted positively by over one volt compared to
the model complexes in MeCN solution. Decreased covalency in the FelI1-S
bonds of three Rds relative to the complex [Fe(o-C,H,(CH,S),)),] - was detec-
ted from the intensities of sulfur K-edge X-ray absorption spectra." This is
consistent with the presence of the six NH-S hydrogen bonds in the proteins
(Figure l(b)), lowering electron density on the ligands and causing an increase in
Zeff,I E and E"'. However, differences in covalency among the three Rds exam-
ined did not follow the observed order of reduction potentials and it was
concluded that the solvation energy term U also makes important contributions
to the more positive potentials of the Rds.
A major contributor here is the low effective dielectric constant in protein sites
(E - 5-10 versus I N 38 for MeCN and 80 for water) which reduces the difference
in solvation energy between the oxidised and reduced states and increases E"'. In
addition, the geometry of the Fe(S-Cys),* site (including S-Fe-S bond angles)
and the NH-S hydrogen bonding pattern is retained in all Rds that have been
* Formulations such as Fe(S-Cys), are intended to imply direct bonds between iron and sulfur
atoms of the cysteinyl residues and not from iron to sulfur and thence to cysteinyls.
Zhiguang Xiao and Anthony G. Wedd 29 1
Table 1 Reduction potentials E"' for simple rubredoxins'
Class Source E"', mV Ref:
I (V44) Clostridium pasteurianum - 77, - 53 49,53
Chlorobium limicola' - 61 54
Butyribacterium methyltrophicum - 40 55
Helio bacillus mobilis - 46 56
Pyrococcus furiosus A44V - 58 48
Cp Pf chimeras3 - 46 to - 67 53
I1 (A44) Clostridium pasteurianum V44A - 24,+ 31 49,53
Pyrococcus furiosus Oto + 31 57, 58, 59
DesuEfovibrio vulgaris H4 0 50
Desulfovibrio vulgaris M' + 5 60
Desulfovibrio gigas + 6 61
Megasphaera elsdenii + 23 62
Cp Pf chimeras3 + 63 to 69+ 53
'versus NHE.
'f. sp. Thiosulfatophylum
3constructions of fused domains from Clostridium pasteurianum and Pyrococcusfuriosus
4strain Hilden borough
'strain Miyazaki
characterised structurally. Consequently, the effects of changes in ligand geo-

metry upon E"' would appear to be minimal, that is, the entatic state or rack
mechanisms suggested for blue copper and other centres4 do not appear to be
relevant here.
The temperature-dependence of the potentials allows estimation of the
contributions of the enthalpic ( -AH"',,/F = + 170 mV) and entropic
(T(AS"',,/F = - 250 mV) terms to E"' (Equation 2) for the RdCp couple.21.22
[FelI1(S-Cys),]- + e- + [Fe11(S-Cys),12- E"' = - 80(10)mV (4)
The enthalpy change AH"',, is negative ( - AH"',,/F is positive), favouring the Fell
oxidation state. The higher anionic charge of the [Fe(S-Cys),12 - centre seems to
be compatible with the rather hydrophilic environment (including expected
strengthening of the six NH-S hydrogen bonding interaction^)'^ and solvent
exposure of the two surface ligands Cys9 and Cys42 (Figure l(b)). In addition,
molecular dynamics calculations suggest that solvent access may be increased in
the reduced state.22The entropy change is negative, favouring the Fe"' redox
state. This will be determined by differences both in the flexibility of the protein
chain and in the solvation properties between the two redox states.
The influence of the terms I E and CI on the reduction potential for the 'simple'
case of Rd (Equation 4) will now be assessed.
3 The Ionisation Energy Term, IE

Mutant forms of RdCp have been generated with the aim of replacing each of the
four cysteine ligands, in turn, by ~ e r i n e . These ~ ~ proteins, C6S, C9S, C39S
~ ~ - four
and C42S, would constitute four geometric isomers in which FeOS, centres are
orientated differently within the same protein chain. The total data confirm the
presence of FeIII-O-Ser bonds in the surface ligand mutants C9S and C42S and
in the double mutant C9S/C42S.21,25The incorporation of the harder olate
ligand26g27leads to large negative shifts of ca. - 200 and - 450 mV in E"' in the
single and double mutants, respectively (Figure 2; Table 2). The protein has been
converted from a weak reducing agent to a strong one. In addition, the potentials
become dependent upon pH below characteristic pK, values, interpreted as a
consequence of protonation of the 0-Ser ligands of the FeI1centres. EXAFS data
reveal an increase of 0.1-0.2 A in the Fe-0 bond length upon reduction. The
data are consistent with the following couples (x = 1 or 2).
pH > pK,: [Fel"(S-C ys), -(, 0-Ser),] - +e - = [Fell(S-C ys), -(, 0-Ser),] -
(5)
pH < pK,: [Fell1(S-C ys), -(, 0-Ser),] - +H +e+ - = [Fe"(S-C ys), -
(0-Ser),- l{O(H)-Ser)] - (6)
For pH values above the pK, values, large increases in the reduction enthalpies
are responsible for the dramatic lowering of E"' (Table 2). The new olate ligand is
a stronger 0 donor and decreases Zeff, stabilising the oxidised form. In addition,
the structure of the Fe1I1-C42Sprotein indicates that the smaller size of oxygen
relative to sulfur leads to closer packing of atoms in the local surface region
-1 00
-200
5-
E -300
Y
W"
-400
-500
-600
5 6 7 8 9 10
PH
Figure 2 Vuriation of peak potential EP with pH for RdCp mutant proteins. Reproduced
with permission of the publishers and authors ofreJ: 21
Table 2 Changes in the reduction thermodynamics of RdCp on C y s to Ser

substitutiona'"
PH Mutant pK, A( - AH"',, /F)" AT(AS"',,/F)" AEc'b

(EO'rnut - Eo'wt)
( mV ) (mV ) ( mV )
Above p K ,
10 C6S 9 - 80 -5 - 85
c39s ( - 185) ( + 80) ( - 105)
8 c9s 7 - 220 + 10 - 210
C42S 7 - 225 + 35 - 190
10 C9S/C42S 2 9.3 - 540 + 85 - 455
Below pK,
8 C6S 9 + 130 - 170 - 40
c39s ( - 150) ( + 70) ( - 80)
6 c9s 7 - 230 + 60 - 170
C42S 7 - 225 + 105 - 120
8 C9SfC42S >, 9.3 - 640 + 250 - 390
a From ref. 21
The data are relative to those of the recombinant form, averaged in the range pH 6-10
" Enthalpic and entropic contributions to AE"'(see Equation 2). Differences of 3 50 mV
are considered to be significant
around position 42, a closer approach of the polar peptide chain to the iron
centre and loss of the NH-S-42 hydrogen bond.25This would modify interac-
tions at the metal-protein interface (see below). The magnitude of the enthalpic
contribution to AE"' for the double surface mutant C9S/C42S is more than twice
that for the single surface mutants, consistent with the presence of two 0-Ser
ligands. In contrast, the entropic contributions to AE"' are minor in each mutant
protein.
Relative changes at pH values below the p K i e dof the single mutant proteins
C9S and C42S show that the observed positive shifts in E"' from the pH
dependence (Figure 2) are entirely due to increases in the entropic term
T(AS"',,/ F). Structural changes induced by protonation of the Ser ligand appear
to drive this change.
The environments of the interior ligands Cys6 and Cys39 differ from those of
the surface ligands Cys9 and Cys42 (Figure 1).They are involved with two NH-S
hydrogen bonds and are adjacent to the hydrophobic core of the protein. The
C6S and C39S mutant proteins exhibit negative shifts in E"' which are only
about one half of those seen for the surface single mutants (Table 2). The other
properties of C39S are similar to those of the latter proteins and the primary
reason for the difference in E"' appears to be a longer Fe-(O-Ser) link (about 0.03
A)lessening the olate ligand donor strength in C39S.25A possible reason for the
lengthening is that the protein chain carrying residue 39 may have lower confor-
mational freedom as it is adjacent to the aromatic core of the p r ~ t e i n . ' " ~ ' ~
However, two NH-.S-Cys hydrogen bonds have been lost (as opposed to one for
the surface ligand mutants, Figure l(b)) and these are known to influence
metal-ligand bond lengths.28
The thermodynamic parameters of the C6S protein below its pK, 9 are N
unique for Rd systems: a large positive A (- AH"',,/F) value is coupled to a large

negative A( TAS"',,/ F ) value (Table 2). While the resonance Raman spectra of the
C6S and C39S proteins are very similar, indicating the presence of a similar
Fe1I1S3Ocentre in each protein, a v(Fe-OH) vibration of a Fe-OH, (x = 1 or 2)
fragment was detected at 617 cm-' for C6S but not for C39S.,' The weakness of
a longer Fe1I1-O(Ser)bond in C6S has led to its hydrolysis and the presence of a
[Fe1I1(S-Cys),(0H)]- centre and an unligated HO-Ser-6 residue.
The observed reversible electrochemistry and its pH dependence (Figure 2)
may then be interpreted by the processes shown in Equations 7 and 8.
pH > pK,: [FeII'(S-Cys),(OH)]- + e- = [F~II(S-C~S),(OH)]~- (7)
pH < pK,: [FerI1(S-Cys),(OH)]- + H f + e- = [Fe" (S-Cys),(OH),]-
The more positive reduction potential for C6S relative to those of C9S and C42S
is then attributable to the presence of a less-electron-donating HO - ligand
relative to Ser-0-.
A more subtle approach to assess the influence of electronic factors on E"'
addresses the conserved residue Tyrl 1 which is adjacent to the Fe(S-Cys), site
(Figure 3).29*30Its p-hydroxyl group is exposed to the solvent indicating that its
substitution will have a minimal impact on protein structure. Substitution of
-OH with -H, -F, -NO2 and -CN in this position in Rd from Pyrococcusfuriosus
induced a 30 mV (- 3 kJ) range in E"' and a correlation of 3 kJ per G~ unit with
the Hammett parameter for electron donation.29 More electron-withdrawing
groups on the ring induced more positive potentials (-NO,, -CN) while more
electron-donating groups induced more negative potentials (-H, -F, -OH). The
possible interactions responsible for the modulation are remote from the substi-
tution site. They are proposed to be either (a) an electrostatic interaction between
the positively charged perimeter of the aromatic ring and the negatively charged
ligand sulfur atom of Cys or (b) a perturbation of the ll-NH--S-Cys9 hydrogen
bond (Figures 1 and 3).
4 The Solvation Energy Term, U

The effects upon E"' of the mutations discussed in the previous section were
interpreted as consequences of changes primarily in the ionisation energy term
I E of Equation 3. Similar assumptions are made in many literature discussions.
The influence of the solvation energy term U is much more difficult to assess as it
is determined by the reorganisation of the active site environment upon reduction.
However, very few proteins have been characterised structurally in both oxida-
tion states. Consequently, discussions of U must remain speculative. In the
present case, we will proceed by assuming that, for each Fe(S-Cys), site in the
simple Rd proteins (Table l),the background dielectric constant is the same and
the intrinsic properties of the Fe-S bonds are the same. Then, differences in E"'
due to U (Equation 3) can be attributed to differences in specific charge interac-
tions in the vicinity of the active site as a result of varying amino acid sequence.
Figure 3 Environment of the T y r l l side-chain in rubredoxin
Under this imperfect assumption, these interactions will be mono- or di-polar in

nature and are listed below.
(1) Proximity of charged residues,
(2) proximity and orientation of dipoles (including NH-S hydrogen bonds),
(3) solvent access.
The challenge then is to devise experiments that will allow individual assessment
of these factors, especially as they are interdependent and the potentials are very
sensitive to small perturbations. The results of experiments designed to test
factors (1)<3) above will now be addressed in turn.
4.1 Proximity of Charged Residues

The high solubility of RdCp follows from its anionic charge of 9 units at pH 7.
The amino acid sequence contains 13 carboxylate (Asp, Glu) and four am-
monium (Lys) residues (Figure l(c))'' and their positioning relative to the active
site must be assessed. The localised charges of these residues are all confined to
the surface of the protein: none is buried in the interior. In addition, the surface
adjacent to the FelI1(S-Cys), unit is, in fact, hydrophobic and the closest charged
residues are Lys46 and Glu47 whose N Eand Oa atoms are, respectively, 8.0 and
7.7 from the S y ligand atom of Cys6. Such weak electrostatic effects could help
tune the site to the observed E"' (Equation 4).
Altered surface charges influence E"' in some proteins and not in others. This
appears to be related both to the distance of the charge from the active site and to
the influence of solvating water in masking the effects of the For
example, changes in surface negative charge by up to five units did not signifi-
cantly influence the reduction potential of an 8Fe-8S f e r r e d ~ x i n . ~ ~
Two alkyl side chains Val8 and Leu41 form part of the surface close to the
Fe(S-Cys), site of RdCp (Figure l(b)). The influence of surface charge has been
assessed by substituting these neutral residues with positively charged Arg and
negatively charged Asp.36The potentials of the mutant proteins V8R and L41R
showed increases of 40 and 58 mV, respectively, as expected if the positive charge
stabilises the reduced [Fe11(S-Cys),12- member of the couple at distances of
7.0-8.5 A.In addition, the effects were approximately additive as the shift for the
double mutant V8R/L41R was +85 mV. However, the mutant proteins with
negative charges in these positions, V8D and L41D also showed positive shifts
and the conclusion was that the proximity of these charges alone could not be
used to rationalise the observations. It was suggested that each mutation had
induced an increase in effective access by the polar solvent water. Increased
polarity would stabilise preferentially the reduced form with its higher anionic
charge (see Section 4.3 below), assuming that the different charges on the
side-chains were masked by solvation.
4.2 Proximity and Orientation of Dipoles (Including NH-S

Hydrogen Bonds)
The idea that these dipolar interactions were important in tuning E"' values in
proteins originated with Carter. He suggested that the differing numbers of
backbone peptide NH-S hydrogen bonds determined which Fe,S,(S-Cys),
couple was employed by reductant ferredoxins ([Fe,S,(S-Cys),] - / 3 -) or by
oxidant HiPIPs ([Fe,S,(S-Cys),] - I 2 -).3 Spectroscopic support was supplied
by Sanders-Loehr who also pointed out the hydrophobic nature of the HiPIP
sites (Section 4.3 below).38 Warshel and Stephens have laid emphasis on the
number and orientations of backbone peptide amides surrounding active
sites.39940
Low and Hill have synthesised two backbone engineered analogues of the
HiPIP from Rhodocyclus t e n ~ i sA. ~peptide ~ amide link is replaced by an ester,
thereby removing an NH-S-Cys hydrogen bond to the Fe,S,(S-Cys), cluster
with apparently minor structural implications. The [O]-Va142 and [0]-Ala57
proteins showed negative shifts of - 86 and - 126 mV, respectively, consistent
with the less polar environment favouring the oxidised state [Fe,S,(S-Cys),] -
of lower anionic charge. In related work, the S79P mutant of the HiPIP from
Chromatium vinosum was designed to reduce the partial positive charge on
79-NH, involved in one of the NH-S hydrogen bonds to the cluster.42 NMR
spectroscopy indicated that the overall protein folding and electron distribution
in the cluster is little affected in the S79P mutant. E"' was again shifted negatively
(- 104 mV). Similar effects have been seen in other FeS systems and in model
corn pound^.^^^^^
Careful structural monitoring is necessary when exploring the influence of
mutations aimed to vary backbone peptide orientation. The amide groups near
the Fe,S,(S-Cys), cluster were varied in ferredoxin I of Azotobacter v i n e l ~ n d i i . ~ ~
The P21G mutation alters the orientation of backbone amide relative to the
cluster and E"' increased by 42 mV. The 140Q mutation inserts a new amide
group near the cluster and E"' increased by 53 mV. Both results indicate a
sensitivity of E"' to the proximity of amide dipoles and are consistent with
theoretical predictions of contributions to enthalpy changes favouring the reduc-
ed form of the [Fe,S,(S-Cys),12 - I 3 - couple.
Two glycine residues adjacent to the cysteine ligands at positions 10 and 43 are
conserved in all rubredoxins, consistent with the proposal that a p carbon
substituent at these positions would eclipse adjacent peptide carbonyl groups
and prevent formation of the 11-NH--SY-9and 44-NH-.SY-42 hydrogen bonds
(Figure l(b)).46Incorporation of valine at GlO causes the 9-10 peptide link to
invert in the G10V/G43A double-mutant protein relieving steric interaction
between C9 0 and V10 Cp.47This drastic change in conformation is accom-
panied by other significant structural changes but the new conformation allows
the 11-NH-SY-9interaction to be maintained. The 9-10 peptide link now orients
its negative pole towards the iron atom and the combined structural changes
cause E"' to shift by at least - 40 mV. One may speculate that inverted polarity
of the 9-10 amide dipole dominates the shift: note that it is in the opposite sense
(favouring the oxidised form) to that observed in the ferredoxin I system dis-
cussed above. In that work, the targetted amide directed its positive pole towards
the active site.
For the V44A mutant of RdCp, E"' shifts by more than +50 mV while the
complementary mutation A44V in the Pyrococcus furiosus protein produces a
similar but negative shift.48949 The Cp atom of Ala 44 in V44A of RdCp lies very
close to the position of the Cy2atom of Val 44 in the native Rd (Figure 4). But the
lower overall steric demand of A (Me) relative to V (Pr') allows the protein chain
to move towards the iron site. The 44NH-342 distance contracts by 0.4(1) A,
consistent with significant strengthening of the hydrogen bond. Consequently, a
minor change in the steric requirements of a surface residue has been amplified to
a positive shift in E"' of at least 50 mV by the strengthening of the NH-S
hydrogen bond.
An Rd-like Fe(S-Cys), centre identified in the 'as isolated' rubrerythrin from
Desulfovibrio vulgaris exhibits an unusually high reduction potential (E"' = + 230
mV; cf. Table l).50 Recent structural data indicate that this centre features an
unusually strong seventh NH-S-Cys bond in addition to the six NH-S-Cys
bonds usually seen in Rd (Figure 1(b)).28The Fe-S distance involving this extra
seventh NH-S-Cys bond appears to be lengthened significantly, consistent with
relative stabilisation of the reduced form and the increase in E"'.
4.3 Solvent Access

Although the protein surface close to the active site in Rd is hydrophobic, the S y
ligand atoms of the two surface Cys ligands appear to have some contact with
solvent. The Pr' side chains of two adjacent valine residues, V8 and V44, define
Figure 4 Overlay of the structures of RdCp and the corresponding part of the V44A
protein. The iron atom and residues 6-8 and 42-44 are shown. The modeled
N H protons of residues 44 which are involved in the 44NH-aS42 hydrogen
bonds are labelled x in RdCp and y in V44A. The bond is 0.4(1)8,shorter in
the V44A mutant protein than in RdCp.
Reproduced with permission of the publishers and authors of re5 49.
the surface of the rubredoxin from Clostridium pasteurianum and control access
to its Fe(S-Cys), active site (Figure l(b)). To assess the effect of systematic change
of the steric bulk of the alkyl side chains, a number of single and double mutant
proteins were isolated which vary G (H), A (Me), V (Pri), L (Bu') and I (BuSec)at
those position^.^^*^^,^^
At a superficial level, the increased potentials correlate with decreased side-
chain volumes, consistent with increased solvent access (Table 3 and Figure 5).
The double mutant V8G/V44G exhibits the most positive shift in potential (116
mV). This correlates with the absence of side-chains at positions 8 and 44,
allowing maximum solvent penetration. On the other hand, the shift for the
double mutant with larger alkyl side chains V8I/V44I ( + 22 mV) is essentially the
sum of those of the single mutant constituents V8I ( - 4 mV) and V44I ( + 24 mV).
Structural data for Ferrlforms of the V44A and V44I proteins are available to
assist interpretation but conclusions must assume that no significant structural
variation occurs upon reduction.
Ichiye and Scott have drawn attention to two classes of Rd proteins from
bacterial sources.33348 The potentials of class I (Cl, Hm, Bm, Po) are negative and
similar to that of RdCp (- 77 mV) (Table 1).Those of class I1 (PA DOH,DuM, D g )
cluster around 0 mV. Ichiye observed that class I and I1 proteins have valine and
alanine, respectively, at position 44. The amide NH of position 44 acts as a
hydrogen bond donor to the Sy atom of ligand C42 (Figures l(b), 4). For RdCp,
calculations predicted that the V44A mutant protein would show a positive shift
in potential of about 40 mV, effectivelyconverting it from class I to class II.33The
lower steric demand of A (Me) compared to V (Pr') would drive an increase in
Table 3 Peak potentials for V8 and V44 mutants of RdCp
Protein E , mV A E , mVb
rRd - 77
V8G -7 + 70
V8A - 44 + 33
V8L - 82 -5
V81 - 81 -4
V44G 0 + 77
V44A - 24 + 53
V44L - 87 - 10
V44I - 53 + 24
V 8G/V44G + 39 + 116
V8I/V44G - 13 + 73
V8I/V44I - 55 + 22
a From reference 49.

Potential relative to rRd.
- 8
80
0 0 44
O L
I I I I
G A V I
L
aa
Figure 5 Dependence of peak potentials E , on the residue sidechain in V8 and V44

mutant Rd proteins.
Reproduced with permission of the publishers and authors of rej 49
polarity via a change in the conformation of the protein chain and its translation
towards the iron atom. Solvent access effects would be less important.
The discussion of the structure of the V44A protein under Section 4.2 above
provides some experimental support for this p r e d i ~ t i o n . ~ * , ~ ~
(i) The peak potential of V44A is 3 53 mV more positive than that of rRd
(Figure 5).
(ii) The backbone in the vicinity of residue 44 is shifted towards the iron site
(Figure 4) and the 44N-SY42 distance is decreased significantly by 0.4( 1) A.
In addition, the solvent-accessible surface areas of the side-chains of the C9
and C42 ligands are increased by about 10% in the V44A mutant relative to rRd.
Overall, the experimental results are consistent with a combination of backbone
structural change and solvent access contributing to a more polar environment
for the iron site in the V44A protein. Their relative contributions to the larger
shifts of + 70 to + 116 mV seen for the V (Pr’) to G (H)mutants (Figure 5; Table
2) must await further detailed structural information.
The difficulty in discriminating between these two influences is highlighted
when examining the structure of the Rd from Desulfovibrio des~lfuricuns.~’ As its
sequence is shorter than that of RdCp by seven residues (20-26), RdDd was not
included in the analysis of Ichiye and Scott. Its potential is about 0 mV (class 11)
but it features valine at the position equivalent to 44 (class I). A number of
separate structural features would appear to be at work here.
Estimation of the effects of solvent access has been the most difficult of the
factors to assess. The best understood systems from this point of view are the
HiPIP proteins where the aromatic core seems to stabilise the oxidised state
[Fe,S,(S-Cys),] - of the buried site by restricting solvent a c ~ e s s , ~but
, ’ ~this is,
of course, negative evidence.
5 Acknowledgement
AGW thanks the Australian Research Council for support.
6 References
1 A. G. Wedd, Coord. Chem. Rev., 1996,154,5.
2 B. I. Dayiyat, and S. L. Mayo, Science, 1997,278,82.
3 W. F. DeGrado, Science, 1997,278,80.
4 H. B. Gray, B. G. Malstrom and R. J. P. Williams, JBIC, 2000,5, 551.
5 H. W. Hellinga, J . Amer. Chem. Soc., 1998,120,10055.
6 R. H. Holm, P. Kennepohl and E. I. Solomon, Chem. Rev., 1996,96,2239.
7 F. Capozzi, S. Ciurli and C. Luchinat, Struct. Bonding, 1998,90, 127.
8 H. Beinert, JBIC, 2000,5,2.
9 W. Lovenberg and W. M. Williams, Biochemistry, 1969,8, 141.
10 I. Moura, M. Teixeira, J. LeGall and J. J. G. Moura, J . Inorg. Biochemistry, 1991, 44,
126.
11 I. Mathieu, J. Meyer and J.-M. Moulis, Biochemistry J., 1992,285,255.
12 M. K. Eidness, J. E. O’Dell, D. M. Kurtz Jr., R. L. Robson and R. A. Scott, Prot. Eng.,
1992,5, 367.
13 M. Ayhan, Z. Xiao, M. J. Lavery, A. M. Hamer, K. W. Nugent, S. D. B. Scrofani, M.
Guss and A. G. Wedd, Inorg. Chem., 1996,35,5902.
14 M. W. Day, B. T. Hsu, L. Joshua-Tor, J.-B. Park, Z. H. Zhou, M. W. W. Adams and D.
Zhiguang Xiao and Anthony G. Wedd 30 1
C. Rees, Protein Sci., 1992,1, 1494.
15 L. C. Sieker, R. E. Stenkamp and J. LeGall, Methods Enzymol., 1994,243,203.
16 D. E. Sayers, E. A. Stern and J. R. Herriott, 1. Chem. Phys., 1976,64,427.
17 G. N. George, I. J. Pickering, R. C. Prince, Z. H. Zhou and M. W. W. Adams, J . Biol.
Inorg. Chem., 1996,1,226.
18 L. E. Maelia, M. Millar and S. A. Koch, Inorg. Chem., 1992,31,4594.
19 K. Rose, S. E. Shadle, M. K. Eidness, D. M. Kurtz Jr., R. A. Scott, B. Hedman, K. 0.
Hodgson and E. I. Solomon, J . Am. Chem. Soc., 1998,120,10743.
20 G. Battistuzzi, M. Borsari, A. L. Macedo, J. J. G. Moura, P. Rodrigues and M. Sola,
JBIC, 2000,5,748.
21 Z. Xiao, A. R. Gardner, M. Cross, E. M. Maes, R. Czernuszewic, M. Sola and A. G.
Wedd, JBIC, 2001,6,638.
22 R. B. Yelle, N. Park and T. Ichiye, Proteins: Struct., Funct., Genet., 1995,22, 154.
23 J. Meyer, J. Gaillard and M. Lutz, Biochemistry Biophys. Res. Commun., 1995,212,827.
24 J. Meyer, J. Gagnon, J. Gaillard, M. Lutz, C. Achim, E. Munck, Y. Petillot, C. M.
Colangelo and R. A. Scott, Biochemistry, 1997,36, 13374.
25 Z. Xiao, M. J. Lavery, M. Ayhan, S. D. B. Scrofani, M. C. J. Wilce, J. M. Guss, P. A.
Tregloan, G. N. George and A. G. Wedd, J . Am. Chem. Soc., 1998,120,4135.
26 S. Ahrland, J. Chatt and N. R. Davies, Q. Rev. Chem. Soc., 1958,12,265.
27 R. G. Pearson, Hard and Soft Acids and Bases, Dowden, Hutchinson and Ross,
Stroudsburg, Pa, 1973.
28 L. Sieker, M. Holmes, 1. Le Trong, S. Turley, M.-Y. Liu, J. LeGall and R. E. Stenkamp,
JBIC, 2000,5,505.
29 D. W. Low and M. G. Hill, J . Am. Chem. Soc., 1998,120,11536.
30 R. Kummerle, H. Zhuang-Jackson, J. Gaillard and J.-M. Moulis, Biochemistry, 1997,
36,15983.
31 B. Shen, L. L. Martin, J. N. Butt, F. A. Armstrong, C. D. Stout, G. M. Jensen, P. J.
Stephens, G. N. La Mar, C . M. Gorst and B. K. Burgess, J . Biol. Chem., 1993, 268,
25928.
32 I. Quinkal, V. Davasse, J. Gaillard and J. M. Moulis, Protein. Eng., 1994,7,681.
33 P. D. Swartz, B. W. Beck and T. Ichye, Biophys. J., 1996,71,2959.
34 J. K. Hurley, A. M. Weber-Main, M. T. Stankovich, M. M. Benning, J. B. Thoden, J. L.
Vanhooke, H. M. Holden, Y. K. Chae, B. Xia, H. Cheng, J. L. Markley, M. M.
Martinez-Julvez, C. Gomez-Moreno,J. L. Schmeits and G. Tollin, Biochemistry, 1997,
36,11100.
35 P. S. Brereton, M. J. Maher, P. A. Tregloan and A. G. Wedd, Biochim. Biophys. Acta,
1999,1429,307.
36 Q. Zeng, E. T. Smith, D. M. Kurtz, Jr., and R. A. Scott, Inorg. Chim. Acta, 1996,242,
245.
37 C. W. J. Carter, J. Kraut, S. T. Freer and R. A. Alden, J . B i d . Chem., 1974,249,6339.
38 G. Backes, Y. Mino, T. M. Loehr, T. E. Meyer, M. A. Cusanovich, W. V. Sweeney, E. T.
Adman and J. Sanders-Loehr, J . Am. Chem. Soc., 1991,113,2055.
39 G. M. Jensen, A. Washel and P. J. Stephens, Biochemistry, 1994,33,10911.
40 P. J. Stephens, D. R. Jollie and A. Warshel, Chem. Rev., 1996,96,2491.
41 D. W. Low, and M. G. Hill, J . Am. Chem. Soc., 2000,122,11039.
42 E. Babini, M. Borsari, F. Capozzi, L. D. Eltis and C. Luchinat, JBIC, 1999,4,692.
43 E. Denke, T. Merbitz-Zahradnik, 0. M. Hatzfeld, C. H. Snyder, T. A. Link and B. L.
Trumpower, J . Biol. Chem., 1998,273,9085.
44 T.-A. Okamura, S. Takamizawa, N. Ueyama and A. Nakamura, Inorg. Chem., 1998,
37,2415.
45 K. Chen, G. J. Tilley, V. Sridhar, G. S. Prasad, C. D. Stout, F. A. Armstrong and B. K.
Burgess, J . Biol. Chem., 1999,274,36479.
46 E. Adman, K. Watenpaugh and L. H. Jensen, Proc. Natl. Acad. Sci. USA, 1975, 72,
4854.
47 M. J. Maher, Z. Xiao, M. C. J. Wilce, J. M. Guss and A. G. Wedd, Acta Cryst., 1999,
D55,962.
48 M. K. Eidness, A. E. Burden, K. A. Richie, D. M. Kurtz, Jr., R. A. Scott, E. T. Smith, T.
Ichiye, B. Beard, T. P. Min and C. H. Kang, Biochemistry, 1999,38,14803.
49 Z. Xiao, M. J. Maher, M. Cross, C. S. Bond, J. M. Guss and A. G. Wedd, JBIC, 2000,5,
75.
50 J. LeGall, B. C. Prickril, I. Moura, J. J. G. Moura, A. V. Xavier and B. H. Huynh,
Biochemistry, 1988,27, 1636.
51 W. K. Hormel, B. C. Pickril, K. Titani, J. LeGall and L. C. Sieker, FEBS Lett., 1986,
201, 147.
52 A. Soriano, D. Li, S. Bian, A. Argawal and J. A. Cowan, Biochemistry, 1996,35, 12479.
53 M. K. Eidness, K. A. Richie, A. E. Burden, D. M. Kurtz, Jr. and R. A. Scott,
Biochemistry, 1997,36, 10406.
54 K. J. Woolley and T. E. Meyer, Eur. J . Biochemistry, 1987,163, 161.
55 K. Saeki, Y. Yao, S. Wakabayashi, G. J. Shen, J. G. Zeikus and H. Matsubara,
J . Biochemistry, 1989,106,656.
56 W. Y. Lee, D. C. Brune, R. LoBrutto and R. E. Blankenship, Arch. Biochemistry
Biophys., 1995,318,80.
57 M. W. W. Adams, Adv. Inorg. Chem., 1992,38,341.
58 L. D. Gilles de Pelichy and E. T. Smith, Biochemistry, 1999,38,7874.
59 R. Christen, T. Jancic, Z. H. Zhou, M. W. W. Adams, J. M. Tomich and E. T. Smith,
J . Inorg. Biochem, 1997,65,53.
60 F. Shimizu, M. Ogata, T. Yagi, S. Wakabayashi and H. Matsubara, Biochimie, 1989,
71, 1171.
61 I. Moura, J. J. G. Moura, M. H. Santos, A. V. Xavier and J. LeGall, FEBS Lett., 1979,
107,419.
62 C. V. Dijk, J. W. V. Leeuwen, C. Veeger, J. P. G. M. Schreurs and E. Barendrecht,
Bioelectrochem. Bioenerg., 1982,99, 743.
SECTION G:
Patterns and Generalisations in

Stability and Reactivity
During the twenties and thirties, immediately prior to the beginning of Chatt’s
research at Cambridge under the direction of F. G. Mann, the study of inorganic
compounds was a relatively unfashionable area of chemical research. Even the early
contributions of Werner were being overshadowed by developments in organic
chemistry. This was, in part, associated with the largely descriptive nature of
inorganic chemistry. Apartfrom some notable exceptions, such as theformulation of
the periodic table of the elements and its rationalisation of chemical composition
and Lewis’ characterisation of the coordinate bond, there were few overarching
generalisations which might permit the organisation of the vast body of data on
inorganic compounds, the necessary precursor to a proper basis of fundamental
understanding. Limitations in the physical and spectroscopic techniques available
to investigate the synthesis, structure, bonding and reactivity of inorganic com-
pounds made their study slow and dificult. The growth in inorganic, particularly
coordination, chemistry which began after the Second World War, may be linked
with the developments in structural and spectroscopic techniques, such as X-ray
crystallography, I R and U V-visible spectroscopy and magnetic measurements,
coupled with developments in crystal and ligand field theories, and associated
models of metal-ligand bonding. I t was inevitable that the relatively small number
of those working in these areas should seek each other out to exchange ideas and to
develop research collaborations. More surprising, perhaps, was the support given
by I C I , Chatt’s employers, to the arranging and hosting of such a meeting in 1951,
later to be designated as the first International Conference on Coordination
Chemistry. The meeting at The Frythe brought together names that will always be
associated with the growth, contribution and impact of coordination chemistry in
the latter half of the twentieth century.
Chatt has recalled elsewhere his work on the synthesis of complexes of plati-
num(I1) and the implications for structure arising from their stereoisomerism and
the use of techniques such as the measurement of dipole moments to distinguish cis
and trans isomeric forms. The extensive synthetic studies from the Russian school
led by Chernyaev revealed the directing injuence of ligands bound to metals in
304 Patterns and Generalisations in Stability and Reactivity
which groups in the trans position were more readily substituted than those in the
cis positions. The availability of a range of tractable materials, their ready and
selective transformation and the availability of techniques able to monitor such
reactions provided the stimulus for the study of the kinetics of these processes and
proposals concerning their mechanisms. This permitted ideas to be formulated that
sought to rationalise and explain the origins of this labilising effect. This topic is
covered in the contribution from Professor Basolo, who has the distinction of being
the Royal Society of Chemistry’s j r s t Chatt Lecturer in 1995-1996. Professor
Basolo also reviews some of his work on platinum(1v) chemistry.
Considering Chatt’s contributions to synthetic and structural aspects of co-
ordination chemistry and his interest in seeking associated patterns and ration-
alisations, it was natural that he would have been stimulated by evidence for
stability trendsfor oxygen, nitrogen and halide donor atom ligands in coordination
complexes in aqueous solution described by Irving at the 1951 meeting at the
Frythe. This, and earlier summaries by Sidgwick in 1941, led him to study the
stability of coordination complexes involving a range of acceptors and ligands with
heavier donor atoms, such as sulfur, phosphorus, and arsenic. From an analysis of
the available data, Chatt, with Ahrland and Davies, recognised two important
generalisations:Jirst, the significant difference between the afinities of donor atoms
from thejrst and second periods of the Periodic Table, N and P, 0 and S, F and C1
and second, the existence of two classes of acceptor, those forming more stable
complexes with the lighter ligating atoms, N , 0 and I;, and those forming more
stable complexes with the heavier ligating elements, P, As, S and Cl. These
diferences were characterised as representing class (a) or class (b) behaviour and
were seen as a further manifestation of the role of metal-ligand n-bonding. The
classifications (and those of others, most particularly of J.O. Edwards and G.
Schwarzenbach) were to be later subsumed into Pearson’s hard and soft acid and
base ( H S A B )description, as described by Professor Pearson in the Jirst contribu-
tion in this section. In addition, Pearson also summarises later work, done in
collaboration with R. G. Purr, which has sought to give a theoretical underpinning
of the H S A B principle.
Finally, the paper by S. Otto, S. N . Mzamane and A. Roodt, explores the
consequences for reactivity of square planar rhodium@) complexes towards
oxidative addition and reductive elimination of changing a pair of P-donor ligands
with As-donor analogues. Medium effects are also investigated.
Hard and Soft Acids and Bases
and Joe Chatt
RALPH G. PEARSON
Chemistry Department, University of California, Santa Barbara CA 93 106,
USA
1 Introduction
I began my research career as a physical organic chemist. Very shortly, Fred
Basolo introduced me to the chemistry of coordination compounds. As a result I
knew a great deal of both organic and inorganic chemistry and could see the
fundamental similarities. This led to an appreciation of G. N. Lewis and his
concept of generalised acids and bases.
I realised that much of chemistry could be discussed in terms of the simple
equation
A + B = A:B (1)
Where A is an electron acceptor (acid), B is an electron donor (base) and A: B is
an acid-base complex. The latter could be almost any chemical species. For
example AgCl would be a combination of Ag' (acid), and C1- (base). CH,OH
would be C H plus OH - ,and so on.
+
Clearly the strength of the coordinate bond in A:B was of the greatest
importance in chemistry. Any empirical rules that could help to estimate AH of
Reaction 1, even qualitatively, would be very useful. My thoughts on this subject
were published in 1963 in a paper entitled 'Hard and Soft Acids and Bases'.' This
paper has been identified by Current Contents as one of the most widely quoted
in the scientific literature.
I would like to discuss next the people whose earlier work had a great influence
on me, in the above connection. One of these was John Edwards, who was
already expert in thinking of chemistry in terms of Reaction 1. He had published
a paper noteworthy for several features2 First, he combined data on both
organic reactions and metal ion complexation and secondly he combined both
equilibrium data and rate data to draw his conclusions. He concluded that two
properties of a base were important: one was the ordinary proton basicity and
306 Hard and Soft Acids and Bases and Joe Chatt
the other was ease of electron donation. Also, different Lewis acids (serving as
electrophilic centres) had different sensitivities to these two properties.
In 1961 Edwards spent some time at Northwestern University and we pub-
lished a paper on nucleophilic reactivity that identified the kinds of electrophiles
that would depend on proton basicity and those that would depend on electron
donation. We also changed the emphasis to polarisability of the base, rather than
ease of electron loss.,
We were very fortunate at this point in time because Harry Gray had collected
a great deal of rate data on substitution reactions of platinum(I1)complexes. This
was my first exposure to the more noble of the transition metals. I soon had a
great deal more because I was fortunate enough to be a collaborator with Joe
Chatt on the kinetics of substitution reactions of a series of organometallic
compounds of nickel@),palladium(I1) and platinum(~r).~
This was a good piece of work because we pointed out the great difference in
rates of reaction for the first, second and third transition series. The ratios were
5 x lo6, lo5, and 1, for Ni, Pd and Pt with compounds of the general formula
[MRCl(PR,),] prepared by Bernard Shaw. Of even greater importance to me
was the discovery of the stability of organometallic compounds of platinum(r1)
and palladium(II), compared to the great reactivity of compounds such as AIR,
and RMgX or even nickel@)analogues.
It will be recalled that the organometallic chemistry of the transition metals
was in its infancy even in 1960. It had essentially begun only in 1951, with the
discovery of ferrocene. By 1960 Chatt was arguably the foremost figure in this
important field. Besides leading the way by making dozens of new compounds,
he developed the guidelines for successful syntheses. He also made important
contributions to the theory of bonding in these novel compounds.
In terms of what I was doing, his most important paper was published in 1958,
with Sten Ahrland and Norman Davies.’ This paper was entitled ‘The Relative
Affinities of Ligand Atoms for Acceptor Molecules and Ions.’ The authors drew
on a large basis set of data available to them. This included equilibrium data in
aqueous solution and in the gas phase, but also data on the general stability and
ease of preparation of the transition metal organometallics. This enabled them to
add carbon donor ligands to the more common examples. Their acceptors
included the metal atoms in various oxidation states and Lewis acids of the type
GaMe, and BF,.
Their conclusions pointed out two regularities: (1) there is usually a great
*
difference in coordinating ability between the first and second element from each
of the groups of the Periodic Table; (2) there are two classes of acceptor: (a) those
that form their most stable complexes with the first ligand atom of each group,
and (b) those which form their most stable complexes with the second or a
subsequent ligand atom. These two classes were called (a) and (b), respectively,
and a large number of acceptors (or Lewis acids) were classified as (a) or (b) in
character.
In fact, something similar had been done even earlier by Gerold Schwarzen-
bath? His paper, published in German, apparently attracted little attention. He
was limited to metal ions in solution, in their normal oxidation states. Metal ions
Ralph G . Pearson 307
with a rare gas outer configuration were called class A, and the ones with d"
configuration were called class B. Writing the donor atom of the ligands in order
of decreasing electronegativity we find
F > 0 > N > C1> Br > I > P N C>S
The class A metal ions form their most stable complexes with the left side of the
order, and the class B with the right side.
The transition metal ions were classified as going from class A to class B, as the
number of d electrons increased, except for Zn. The explanation for all of this was
that ionic bonding was predominant in class A, and covalent bonding much
more evident in class B. Zinc failed because of its low ionisation potential, which
caused it to be more A in character.
2 Hard and Soft Acids and Bases (HSAB)

In the 1963 paper, Reaction 1 was actually discussed as the generalised nucleo-
philic displacement reaction:
A:B'+ B = A:B + B' (3)
The reason for this was partly because I was still using rates of reaction as input
data and partly because in the laboratory it was Reaction 3 which was usually
observed, and not Reaction 1. However B can be made constant and then (3)and
(1) become equivalent for our purposes.
Apart from the kinetic data, the criteria of Ahrland, Chatt and Davies were
used unchanged to categorise a large number of Lewis acids as either class (a) or
class (b). Many examples other than metal ions or atoms were included, such as
CH3+,RSO,+, RS' and I,. Looking at the two classes, it was clear that class (a)
acids were of low polarisability, and class (b) were of high polarisability. Since
polarisation means the distortion of the electron cloud of a chemical system by
an electric field, we can label class (a) systems as hard, and label class (b) systems
as soft. Hardness means resistance to change or distortion.
Bases were classified according to the order (2). On the left side the donor
atoms are of low polarisability, or hard, and on the right they are highly
polarisable, or soft. We then can summarise the data used by the HSAB Prin-
ciple, 'hard acids prefer to coordinate to hard bases and soft acids to soft bases.'
We could equally well have said that class (a) acids prefer class (a) bases and
class (b) acids prefer class (b) bases. I had several good reasons for the name
change, one being that the use of the comparative, or even the superlative, was
easier. One acid being harder that another was easier to understand than one
acid being more class (a) than another. Also hard and soft were simple, easily
visualised descriptions of the key properties.
The original choices of A, or (a), and B, or (b), were related to the subgroup
labels of the Periodic Table, IA us. IB, and so on. Unfortunately, in the United
States the labels A and B had become inverted from the convention in Europe,
except for IA and B and IIA and B. This could lead to some confusion. At any
rate, I decided to go with hard and soft, though I sometime regretted it.
It turned out that many scientists did not like the words hard and soft. They
would be acceptable in private conversations, but not in serious scientific papers.
Even today I find some people put quotes around ‘hard’ and ‘soft.’ When spoken
aloud, as in a lecture, it is conventional to smile. Nevertheless, the terms have
found their way into the literature, where they seem to fill a definite need.
Other work by Chatt was important to me in organising my thoughts on
generalised acids and bases. One useful idea was his n-bonding t h e ~ r y The .~
important feature, in his view, was the presence of loosely held outer d-electron in
class (b) metal ions. These could form n-bonds by donation to empty orbitals in
the ligands. These would include n-orbitals in C O or C,H,, and d orbitals in
ligands with P, A, S and Se as donors. Class (a) metal ions would have only
tightly held electrons in the valence shell and also empty orbitals, not too high in
energy, on the metal. Basic atoms such as 0 and F could form n-bonds by
electron donation from ligand to metal.
Of great importance was his realisation of the significance of oxidation state
on the hardness.’ Increased positive charge on the acceptor atom would lead to
increased hardness, as a rule, though there were a few exceptions. A key con-
clusion was that metal atoms in the zero oxidation state would always be class (b)
or soft. This would extend to bulk metals as well and led to an immediate
understanding of poisoning in such catalysts.
The effect of other attached ligands was also understood by Chatt. For
example BF, and BH, both had boron(II1) as the acceptor atom. But in BF, it
would be almost B3+ because of the high electronegativity of F. In BH, the
bonding between B and H would be very covalent, and we would have Bo. Hence
BF, is a hard acid and BH, is a soft acid. Klixbull Jarrgensen systematised this
effect and called it ‘Symbiosis’.’
Armed with so much data, and new insights, it was easy to predict that RSO,’
would be a hard acid, and RS’ would be a soft acid (the outer p electrons of
sulfur take the place of the d electrons of the transition metals). Finding that
RS0,F was stable, while RS0,I was almost impossible to make, and that RSSR
disulfides were stable, while sulfenic esters, RSOR, were very unstable, simply
confirmed the predictions.
The HSAB Principle is supposed to be a unifying concept which makes easier
the remembering of a vast body of chemical facts, and which allows predictions
of a limited nature. It is a summary of facts, and not a theory, though it is often
called that. The reasoning seems to be that its validity is dubious, like that of a
theory.
A common misconception is that HSAB says that only combination of hard
acids with hard bases, or soft acids with soft bases, can be stable. This is certainly
not the case in fact, nor is it implied in HSAB. Instead it only implies an extra
stabilisation in hard-hard or soft-soft combinations. The overall stability of an
A: B bond depends on the intrinsic strengths of A and B, determined by such
factors as size, charge and polarity, for the most part.
For example, in the gas phase F- always is a stronger base that I- and forms
stronger bonds to all metal ions. But in aqueous solution F- is more strongly
solvated than I- by 53 kcal mol-l. Both effects are due to size, and only the
Ralph G.Pearson 309
difference in the heats of hydration enables I - to compete with F- for various

metal ions in solution. For soft metal ions, like Hg2+,I- wins out. For hard ion
such as H', F- is the stronger base.
The really serious objection to the HSAB concept was that no exact definition
of hardness or softness was given. There was an operational definition as given
by Ahrland, Chatt and Davies, but it only put Lewis acids into one of two boxes.
There was no way to rank-order within the boxes, though there were borderline
cases. There was a rank-ordering of bases, given by (2), but it was only approxi-
mate and did not distinguish between all the bases with the same donor atom.
Finally all that was done was to put bases into one of the two boxes labelled hard
and soft.
An operational definition could be given using the equilibrium constant or
reaction heat, for a reaction such as (3). This would rank-order a series of acids or
bases with suitable choices for A: B. Many such definitions were suggested, but
they failed because somewhat different orders were found, depending on the
reference reactions chosen. Also, they lacked a theoretical basis linking them to
accepted theory. Any numbers obtained were valid only for the chosen reference,
and were not transferable.
One can certainly argue that ill-defined terms that cannot be measured and
quantified have no place in exact science. But it can also be argued that the
statement 'the substance is red' has some informational value, even though an
absorption spectrum would usually be better. Saying that an acid is soft also has
some value. The uses made of such statements are actually examples of 'fuzzy
logic.' In spite of its name, this is a respected branch of mathematics. It is a
method for making the best use of limited information." It is widely used in
technology, manufacturing and finance.
In spite of its limitations, the HSAB concept proved to be useful in almost all
areas of chemistry, and related fields. This was more true for chemists actually
working in the laboratory, trying to make definite compounds, or materials with
certain properties. A good example is given by the work of Chatt and his
co-workers. While probably not thinking in terms of hard and soft, they drew the
same conclusions. It hardly needs to be said how successful this kind of thinking
was.
Another area where the simple HSAB rules has been helpful is in the teaching
(or learning) of chemistry. Students today are overwhelmed with vast quantities
of material. Being introduced in high school and in the first year of general
chemistry to the HSAB Principle, pupils can systematise much of their informa-
tion. In my experience students always want a few simple rules that cover a lot of
territory.
3 Chemical Hardness
The problem of an exact definition of hardness, and a valid experimental
procedure to measure it, was solved for me in 1983. Bob Parr, the well-known
theoretician, spent a sabbatical quarter in Santa Barbara. He had already used
density functional theory to define the electronic chemical potential, p.'
In (4), E is the energy of any chemical system, N is the number of electrons, v is the
potential given by the nuclei, I and A are the ionisation potential and electron
affinity. These also define the Mulliken electronegativity, xw.
Density functional theory (DFT) is a branch of quantum mechanics that uses
the electron density function, p, to describe a chemical system, rather than the
more usual wave function. It had already been shown that the electron density
contained all the necessary information to calculate the ground state energy and
other properties. l 2 It is much easier to work with than wave functions.
Parr asked me if the quantity (ap/aN),was related to my ideas of hardness and
softness. This showed great chemical insight on his part since within a few hours I
was convinced that it was exactly what I meant by hardness. Within a few days
we wrote a definitive paper.13
I-A
(5)
Here y, the Greek letter h, is the hardness. The softness, B, is simply the reciprocal
of y.
The initial reason for my accepting (5) as the definition of hardness, was that
the few values of I and A that I could easily find were in agreement with what I
already knew. Class (a) metal ions did give large values of ( I - A), and class (b)
ions gave low values. Other virtues of (5) quickly became apparent. It gave a
procedure for calculating the hardness for any chemical system by simply
measuring I and A. Fortunately, just at this time there was a great deal of new
activity in measuring ionisation potentials and electron affinities. This could be
done for atoms, molecules and positive ions, and tables of hardness values, as
well as electronic chemical potentials, were made available.
An important factor for me was that ( 5 ) gave y the meaning of resistance to
change in the number of electrons, N . But this also implies that we can write
where the symbol 6 means the functional dependence of one variable upon
another. This follows because p is a function of N . Equation 6 has the meaning of
resistance to change, or deformation, of the electron cloud. But this was exactly
the meaning that I had for hardness!
There was another reason why ( 5 ) is a good definition of hardness, though I
was not aware of it in 1983. In molecular orbital theory, Koopman’s theorem
says that ( I - A) is the energy gap between the H O M O and the LUMO.” This
energy gap is the difference in energy between the ground state and the lowest
excited state of the same multiplicity. If this gap is small then it is easier to change
the electron density, according to quantum mechanical perturbation theory. For
example, a large gap means low polarisability and a small gap (y small) means
Ralph G.Pearson 31 1
high polarisability. But these were just the properties that I had associated with
hard and soft in the beginning.
The meaning of both ,u and y is also seen in describing the initial interaction of
two chemical systems, C and D. Since the electronic chemical potential of the
combined systems must be constant everywhere, electrons will flow from the
system of low electronegativity to that of higher electronegativity (see Equation
4). The number of electrons that are transferred is given by13
This is a chemical form of Ohm’s Law. The numerator gives the difference in
potential and the denominator gives the resistance to electron flow.
Parr and I had originally called y the ‘absolute’ hardness. The reason was that
it was a companion parameter to yM (or -p), called the ‘absolute’ electronegativ-
ity, because it had a sound basis in fundamental theory. While this made sense for
1.1, it seemed unnecessary for y. The other scientific use of the term hardness
would be for physical or mechanical hardness. Thus the name ‘chemical’ hard-
ness seems more appropriate for q.
The concepts of electronic chemical potential and chemical hardness have led
to many new ways of looking at chemical reactions and other phenomena of
interest to chemists. For example, the identification of hardness with the
HOMO-LUMO gap leads to the ‘Principle of Maximum Chemical Hardness.’
There seems to be a rule of nature that molecules arrange themselves to be as
hard as possible, that is, to have the largest gap between the occupied orbitals
and the empty ones.16
This is not the place for a discussion of the many new uses of chemical
hardness. A recent book gives a fairly detailed introduction and current sum-
mary. l 4 Applications are given ranging from atoms to solids. The case of solids is
particularly interesting because both ,u and y have long been a part of solid state
theory. Thus ,u is the Fermi energy and q is the band gap. Physical hardness is an
important property of solids and is related to chemical hardness.
Density functional theory is the basis for the definition and use of ,u and y.
DFT, of course, has revolutionised the whole theory of chemistry in the last
twenty years. The Nobel Prize in Chemistry was given to Walter Kohn in 1998
for his contributions. These include the original concept and an efficient method
for making energy calculations on molecules (and solids). But ,u and y are
concepts of DFT which are still new and where many more applications will
undoubtedly be found.
At this point we seem to have come a long way from the chemistry of Joe
Chatt. Is there really a connection between the stability of olefin complexes of
platinum(I1)and the semi-conducting properties of GaAs? In this short review I
have tried to show that there is a direct relationship between the two. The class
(a) and class (b) metal ions of Chatt and the two classes of electrophiles of
Edwards clearly are the parents of the Principle of Hard and Soft Acids and
Bases. The latter, in turn is one parent of Chemical Hardness, the other being
DFT. This puts hardness near the forefront of modern chemical theory. Hope-
fully, there will be many new uses for the concept. These, like myself, will owe a
debt to the scientists who went before.
4 References
1 R. G. Pearson, J . Am. Chem. SOC.,1963,85,3533.
2 J. 0.Edwards, J . Am. Chem. SOC.,1954,76,1540.
3 J. 0.Edwards and R. G. Pearson, J . Am. Chem. SOC.,1961,84,16.
4 F. Basolo, J. Chatt, H. B. Gray, R. G. Pearson and B. L. Shaw, J . Chem. SOC.,1961,
2207.
5 S. Ahrland, J. Chatt and N. R. Davies, Q . Rev. Chem. SOC.,1958,12,265.
6 G. Schwarzenbach, Experentia SuppE., 1956,5, 162.
7 J. Chatt, Nature, 1950, 165, 859; J. Chatt, Proc. Chem. SOC,1962,318.
8 C. K. Jarrgensen, Inorg. Chem., 1964,3,1201.
9 R. G. Pearson, Chem. Br., 1967,3,103.
10 For example, see D. H. Rouvray, Fuzzy Logic in Chemistry, Academic Press, New
York, 1997.
11 R. G. Parr, R. A. Donnelly, M. Levy and W. E. Palke, J . Chem. Phys., 1978,68,3801.
12 P. Hohenberg and W. Kohn, Phys. Rev., 1964,136, B864.
13 R. G. Parr and R. G. Pearson, J . Am. Chem. SOC.,1983,105,7512.
14 R. G. Pearson, Chemical Hardness, Wiley-VCH, Weinheim, 1997.
15 R. G. Pearson, Proc. Natl. Acad. Sci. USA, 1986,83,8440.
16 R. G. Pearson, J . Chem. Ed., 1987,64,561.
17 R. G. Parr and P. K. Chattaraj, J . Am. Chem. SOC.,1991,113,1854.
Mechanisms of Platinum
Reactions
F. BASOLO
Department of Chemistry, Northwestern University, 2145 Sheridan Road,
Evanston, Illinois 60208-3113, USA
Professor Joseph Chatt was one of the truly outstanding inorganic chemists of
the twentieth century.
I wish to begin by thanking the editors, G. J. Leigh and N. Winterton for
asking me to write a chapter in this book honouring my professional and
personal friend, Professor Joseph Chatt. He and I met for the first time in 1955
when my wife and I were on our initial tour of England. We drove making
unannounced stops at bed and breakfast places; starting from St Andrews where
I played a round of golf at the Royal and Ancient golf course, where the game was
first played. We continued to drive south, again making unannounced stops at
various universities where I wanted to meet some particular inorganic chemists,
whom I knew from their publications, and who knew me for the same reason.
One place I wanted finally to reach was The Frythe to meet Joseph, because I
had read most of his papers, and I felt his research problems were well designed
to get reliable information that answered important questions of fundamental
concern. Furthermore, it was clear from his publications that the experimental
work was done with great care, and that his interpretation of the experimental
results was almost always on target.
Not only did I get to meet Joseph, but my wife Mary met his wife Ethel. We all
had a friendly visit together, after which they took us to see some Roman
(archaeological) digs nearby. They asked us to be their house guests overnight,
which we gladly accepted. After a delightful dinner, we talked about personal
matters, but no chemistry. We learned they had two children, a girl and a boy; we
had the same but ours were a little younger. This way our families became
acquainted, and we developed a friendship that has lasted all these years - see the
end of this chapter for some of the friendly exchanges between Joseph and me.
3 14 Mechanisms of Platinum Reactions
1 Chemistry of Platinum@) Complexes
Immediately when chemists think of platinum(1r) complexes, they know it will
almost certainly involve Chatt chemistry. He could rightly be called the ‘Father
of Platinum(I1)Coordination Chemistry in the UK,’ because the scientific litera-
ture abounds with his publications on important aspects of platinum(@ chemis-
try. I do believe that his fundamental seminal research on these systems is second
to none.
In 1996, Coordination Chemistry Reviews (CCR) dedicated one of its issues to
the memory of Joseph Chatt. I published an article in that issue entitled,
‘Recollections of early studies on platinum(r1) complexes related to Chatt’s
contributions to coordination chemistry’. For this reason, and because some
authors will discuss platinum(r1)chemistry in other chapters, I will devote most
of this paper to our research on the kinetics and mechanisms of platinum@)
ligand displacement, and to the chemistry of platinum(1v) complexes. Since
Joseph did little research on platinum(1v) chemistry, it seems appropriate that I
review our work on it. Following this, I will briefly describe another area we
worked on that interested Chatt although he was not involved in such research.
This was that of making use of reaction mechanisms to help design the synthesis
of metal complexes.
2 Kinetics and Mechanisms of Ligand Substitution

Reactions of Platinum@) Complexes
We have published many papers on the kinetics and mechanisms of platinum@)
reactions, and the publications on this aspect of our research are reviewed in the
Basolo and Pearson book.2 It seems appropriate to summarise some of our work
on this topic, because it differs from the extrakinetic trans eflect properties of
platinum(I1) complexes studied extensively by Chatt. This means that he and his
research group investigated mostly the static properties of platinum(r1) com-
plexes. Instead our work is on the dynamics of the reactions of platinum(r1)
complexes, and we chose to call this the kinetic trans eflect. Thus, Chatt and his
research group considered measurements on the complexes’ stabilities, dipole
moments, infrared spectra, nuclear magnetic resonance (NMR) and platinum(I1)-
ligand bond distances measured by X-ray crystallography. They also noted the
qualitative rates of ligand substitution, and further observed that this reaction
occurs without stereochemical change. For example, these experimental obser-
vations resulted in data such as the bond distances (A) of cis- and trans-
[PtCl,(PR,),] (1 and 2).
PR3 PR3
CI-Pt
I -CI
2.29A
I I 2.30 8,
CI PR3
1 2
F. Basolo 315
Table 1 Dipole moments of trans-[PtLCl(PEt,),] complexes in benzene at
25 "C3
L Debye units
4.2
3.4
2.6
1.1
0.0
This shows that the Pt-Cl bond distance is greater if it is trans to PR, (1) than
if it is trans to Cl- (2).Such results are in accord with the fact that the greater rate
of C1- displacement is that of the C1- trans to PR,. Another approach used by
Chatt to obtain information on the nature of ligands trans to the ligand being
replaced was to measure the dipole moments of a group of analogous com-
pounds. The net dipole moment of a complex will depend on the polarity and the
geometry of the four groups surrounding the metal. Data of interest in this
discussion are shown in Table 1 for the compounds trans-[PtLCl(PEt,),].
For changes in L, the dipole moments decrease in the order
H > CH, > C,H, > p-ClC,H, > C1, which is the same as the order of decreas-
ing trans-effect for these ligands. The results suggest there is a larger transfer of
electrons from H-, CH, -, C,H, - and p-ClC,H,- toward platinum(1r). There-
fore, the Pt-C1 bond is more polar and presumably weaker in H-Pt-C1 than in
Cl-Pt-C1. This agrees with the bond distances derived from X-ray crystallogra-
phy and these results are consistent with the qualitative observations that the
reactivity of these complexes depends on the nature of the ligand in the trans
position.
The Russian School, particularly Chernyaev, did much of the early research on
platinum(I1) chemistry. This arose from the large platinum resources in some of
their natural minerals. So important was the availability of platinum to them
that they had an Institute devoted in large part to the chemistry of platinum.
Most of their research dealt with the syntheses and reactions of platinum
complexes. Their primary goal seemed to have been to enhance the extraction of
platinum from its mineral source. As early as 1926, Chernyaev, reported that
certain ligands in the position trans to the leaving group of square-planar
platinum(I1)complexes have a marked effect on its replacement substitution. He
used this with considerable success in the preparation of desired platinum(I1)
complexes. For example, he was able to prepare cis-[PtCl,(NO,)(NH,)] - by the
reactions shown in (1).
Cl Cl CI
Chernyaev designed this synthesis, knowing that C1- has a greater trans effect
316 Mechanisms of Platinum Reactions
than does NH,. Therefore, the NO2- will replace a Cl- opposite a C1- ligand,
whereas, if NH, had the greater trans eflect the NO2- group would replace the
C1- trans to the NH,. The Russian School attributed this to the activating group
being the more polarisable; thus, in the reactions above C1- is more polarisable
than NH,. During some 30 years that followed, ligands such as CO, N O and
C2H4 were found to have among the highest trans eflects. This could not be
explained by the Russian ideas concerning the importance of polarisability of the
trans ligands.
It is of interest to note that as early as 1828, a Danish pharmacist, Zeise,
prepared the stable compound K[PtCl,(C,H,)], commonly called Zeise’s salt.
Some 125 years later, the chemical bonding of ethylene (C2H,) to platinum(I1)
was finally understood. There is very little doubt that the unsaturated ligand
H2C=CH2 coordinated to platinum(I1) is held to the metal by a double bond,
where the n-bonding electrons from the organic molecule form a coordinate
bond. The metal in turn donates a pair of d electrons to the organic molecule
through a n bond. This description of the bonding was proposed by Chatt’” and
Dewar6 and is generally known as the Dewar-Chatt-Duncanson model. The
bonding type is very important in organometallic chemistry.
Again, I feel certain that other authors will give a more detailed account in this
book of the above n-bonding description. Therefore, returning to the very high
trans efSect of ligands such as C2H4, CO, and NO, Chatt proceeded to explain
how the n concept could be used to account for the bonding of C2H4to platinum
in Zeise’s salt. He wrote,5b ‘The operation of the rapid and readily reversible
elimination of groups trans to ligands A (3) of high trans-eflect (i.e. high double-
bonding capacity) is readily explained if we suppose that trans-substitution
occurs by an S,2 (bimolecular) mechanism. Increasing double bonding by A
increases the electron affinity of the metal atom and hence the ease of nuc-
leophilic attack. Also, because the electron withdrawal occasioned by A occurs
from the antinodes from A of the &-orbital, the attack takes place there,
preferentially displacing the ligand trans to A.’
A-
r
Pt -CI
I
CI
3
As stated by Chatt and co-workers, their early speculation on the role of
trans-n bonding groups in ligand substitution of platinum(I1) complexes was
based on the assumption that the reactions proceed by an S,2 mechanism.
However, at the time (1955) most of the observations reported on such reactions
were qualitative and little had been done to use detailed kinetic studies in
attempts to elucidate the mechanism of ligand ~ubstitution.~ Since the valence
bond theory in use then assigned dsp2 hybridisation to the square-planar plati-
num(I1) complexes, coordination chemists believed an entering nucleophile
would readily attack the low energy vacant p orbital on the metal and substitu-
tion would take place by an S,2 mechanism. Furthermore, a coordination
F . Basolo 317
number of four means little or no steric hindrance in going to coordination
number five. Many kinetic studies now support the sN2 mechanism, so much so
that when Romeo8aand co-workers8b*8c discovered s N 1 reactions of platinum(r1)
complexes, the referees delayed publication because of the strong belief that ‘all
platinum(I1)substitutions are sN2 as appears in Basolo’s and Pearson’s book’.
It was clear in the 1950s that there was a need for detailed kinetic studies of
ligand substitution reactions of platinum@) complexes, and our laboratory was
prepared to do this because it was engaged in such studies of octahedral substitu-
tion. However, only a brief account of our studies is given in this article. At about
this time Martin’ and his students initiated their investigations of aquation
reactions of chloroammineplatinum(I1) complexes.
There is now much kinetic data on substitution reactions of square-planar
complexes’ all of which are explained in terms of a bimolecular (sN2) displace-
ment mechanism. For reactions such as
H O
[MA3X]”+ +Y - a [MA3YIn++ X -
in water solution, a two-term rate law

Rate = kl[MA3Xnf] + k,[MA,X”+][Y] (3)
is generally followed, where k , and k, are first-order and second-order rate
constants, respectively. Under pseudo-first-order conditions with an excess of
Y - , the experimental first-order rate constant, kobs,is related to the individual
rate constants as shown by the equation
This requires that a plot of kobsversus [Y] be linear with an intercept of k, for
the reagent-independent path and a slope of k, for the reagent path. Plots of this
type are common for substitution reactions of square-planar complexes. Such a
plot is shown in Figure 2 for the reaction trans-[PtCl,(py),] (py = pyridine) with
a variety of different reagents.
The results shown in Figure 2 are consistent with Chatt’s classification that a
polarisable ligand atom such as S of SCN- is a much better nucleophile towards
a polarisable metal such as platinum(I1) forming Pt-SCN. l o His classification
refers to these metals as class (b) metals. Such metals do not interact strongly
with much less polarisable ligands. The other class of metals, (a), are much less
polarisable and interact more strongly with the much less polarisable ligand
atoms as would be the case for Al-F. My former colleague Pearson” has
proposed the nomenclature ‘soft’ for highly polarisable metals and ligand atoms,
and ‘hard’ for the less polarisable metals and ligands. Like Chatt, Pearson also
states that ‘soft/soft’ or ‘hard/hard’ interactions are more stable than are ‘soft/
hard’ systems. Both Chatt and Pearson would be the first to agree that this
classification is qualitative, albeit useful. I do recall that the first time Pearson
told me about his ‘hardlsoft’ classes, my response was ‘so what else is new?’ I was
aware of Chatt’s (a) and (b) classes, as well as the work on the stabilities of metal
A A
I ,.x -X
- I
A-M A-M-S
I 's fast I
A A
I
A-M-X 43
I +Y fast
A-M
A
I ,,x
I 'y
A
- -X
fast
A-M-Y
A
I
I
A
Figure 1 General S,2 mechanisms of ligand substitution reactions of square-planar metal

complexes, such as platinum(I1)compounds, where S is solvent and Y is entering
nucleophile
7 SCN-
350
300
250
h
7
I
W
0 200
z
v)
Z 150
a
A2
100
50
2 4 6 8 10
[Y], (M X IO-*)
Figure 2 Rates of reaction of trans-[PtCl,(py),], with diflerent nucleophiles in CH,OH

solvent at 30°C.
F . Basolo 319
complexes by Jannik Bjerrum and Gerold Schwarzenbach. I even knew about
the Swedish chemist Jons Jacob Berzelius (1779-1848) who had discovered the
elements selenium, silicon, thorium and zirconium. I am told he made the point
that certain metals are found on the Earth's crust as sulfides (HgS), others as
oxides (A1203). However, the names 'soft' and 'hard' have a real meaning, readily
understood, and this acid/base concept is even in elementary chemistry text
books, and in undergraduate courses in inorganic chemistry.
Physical organic chemists for years attempted to evaluate and quantify the
strengths of nucleophiles in S,2 reactions towards carbon. One approach that
met with some success was that reported by Swain and Scott.', They determined
a large number of nucleophilic reactivity constants n, using CH,Br as a stan-
dard. Although this proved reasonably satisfactory for nucleophilic displace-
ment reactions at carbon, it is well known that no one scale of nucleophilicity
exists and that the substrate is important in any determination of nucleophilic
strengths. Quantitative evaluation of the nucleophilic properties of various
reagents generally brings in their basicities toward the proton and a characteris-
tic which may be loosely defined as polarisability or electronegativity. The
nature of the electrophilic substrate determines which of the properties makes
the greatest contribution.
Although Edwards has had success with the use of electrode potentials to
e ~ t i m a t e 'nucleophilic
~ strengths, one disadvantage is that Eo values are not
known for many common reagents. Other attempts to quantify nucleophilicities
towards metals were less successful, so finally Belluco et al.I4 decided to use the
rate constants for C1- displacement from trans-[PtCl,(py),] by different nuc-
leophiles as standards. The procedure followed was essentially that used by
organic c h e m i s t ~ to
' ~ standardise reactivities towards CH,Br.
The nucleophilic reactivity constants nopt were defined by
where ky and k, refer to the constants for reactions of trans-[PtCl,(py),] in
-
C H 3 0 H at 30°C.
CH,OH
trans-[PtCl,(py),] +Y -/O
30 "C
trans-[PtCIY(py),lo/+ + C1- (6)
Table 2 gives values of nopt for several nucleophilic reagents. A plot of log k, for
other platinum complexes against nopt gives reasonably good straight lines
(Figure 3), which supports the linear free energy relationship
log k , = Dnopt + log k, (7)
The intercepts of plots such as Figure 3 are close to the values of k, for each
substrate, and the constant D is dependent on the nature of the substrate. It is a
nucleophile discrimination factor, and a large value of D means the complex is
very sensitive to changes in the nature of the nucleophilic reagent. Finally the
most significant point that can be made from values of nopt (Table 2) is that
platinum(I1) is a class (b) or soft metal. The strongest protonic base CH,O- in
-1
-2
Figure 3 Correlation of the rates of reaction of platinum(I1)complexes with noptfor various

nucleophiles: 0, trans-[PtCl,(py),] in methanol at 30°C, 0, [PtCl,(en)] in
water at 35 "C.
Table 2 Values of nopt and of p K , of some nucleophiles
Nucleophile n0pt PK,

CH,O- < 2.4 15.8
c1- 3.04 5.74
-
NH3 3.06 9.25
NO, - 3.22 3.33
1- 5.42 - 10.7
SCN- 6.65 - 1.8
S=C(NH, )z 7.17 - 0.96
s20,,- 7.34 1.9
PPh, 8.39 2.6 1
CH,OH solvent has the smallest nopt,whereas some of the weakest protonic
bases which are highly polarisable (I-, S=C(NH,),, S 2 0 , 2-) have the largest nopt
values. This came as no surprise, but it was important to quantify the fact and to
have a standard for values of nucleophilic strengths of reagents towards plati-
num@) complexes.
F . Basolo 321
3 Collaborative Research with Chatt on the Kinetic
Trans E’ect of the Nickel Triad Metal Complexes
By the late 1950s, Chatt and Bernard Shaw3 had succeeded in preparing several
alkyl-, aryl- and hydrido-metal complexes of the nickel triad. Because of our
mutual interest in the kinetic trans efect of platinum(I1) complexes, Chatt and I
decided to examine the rates of ligand substitution of these new organometallics.
Shaw prepared the compounds and Harry Gray did the kinetic studies. Years
later Chatt gave the following account of why the research was so rapidly
accomplished.
‘Shaw said, “that man Gray must be an absolute glutton for work. I did not
know it was possible to do kinetics so fast; as soon as I get a compound out of the
lab., the result is in and he is waiting for another. I cannot keep up with him”.
When I met Fred Basolo at our next conference, he told me Gray said “that guy
Shaw’s a worker. He nearly drives me mad, the compounds come so fast; as soon
as I have done one the next compound is waiting”.’ Joseph and I laughed, saying
if this competition continues, we will very soon have enough for a publication.
The results of this joint venture were reported for reaction (8) in ethanol
solution between pyridine and several planar compounds of the general formula
indicated.
trans-[MClL(PEt,),] + py + trans-[MLpy(PEt,),] i- + C1-
(M = Ni, Pd, or Pt; L = alkyl, aryl, or hydride)
The relative rates of reaction trans-[M(o-tolyl)Cl(PEt,),l are approximately

5 x lo6 for Ni, 1 x lo5 for Pd, and 1 for Pt. The large difference in lability
between platinum(I1) and nickel@) is in accord with a mechanism where ligands
above and below the plane move in to displace C1-, since nickel@) more readily
expands its coordination number than does platinum@). In further support is the
observation that trans-[Ni(mesityl)Cl(PEt,),l reacts only about 2 x lo4 times
faster than the corresponding platinum@) compound, because the mesityl ligand
blocks the coordination sites above and below the plane and retards the reaction
of nickel(@ more than of platinum(I1). In fact, the differences in rates for the
mesityl systems resemble those between octahedral cobalt(II1) and rhodium(II1)
complexes.
The results of this study further show that the trans-labilising ability of
the ligands L decrease in the order PMe, > H > Me > phenyl M p
methoxyphenyl M p-chlorophenyl > biphenyl > o-tolyl > mesityl M C1.
As mentioned above, there are two main hypotheses used to account for the
trans efect. One is that of the Russian S ~ h o o lwhich
~ , ~ considers it to be mostly
electrostatic in origin, depending largely on the polarisability of the ligand: the
more polarisable the ligand, the greater its trans efect. The second hypothesis is
that of the English School’ which suggests that large trans efects are produced
by ligands able to .n back-bond to the metal. This n back-bonding lowers the
electron density on the metal which enhances nucleophilic attack on the metal
and/or stabilises a five-coordinate transition state for reaction. The results of our
study with Chatt further support there being two types of ligand that have high
kinetic trans efects: those that seem to function via the polarisation theory and
those that involve rc bonding. For example, the relative rates of reaction of
compounds with the trans ligands H, Me, phenyl, and Cl are approximately
100000: 200: 30: 1. Since rc back-bonding is of no importance for these ligands, it
follows that this rapid rate decrease must be mainly electrostatic in origin. This is
supported by the large decrease in dipole moment for changes in the trans-ligand
L = H (4.2 D), Me (3.4 D), phenyl(2.6 D), and Cl(0 D). However, good rc bonding
ligands such as C,H, and CO do not have this polarisation effect, but they do
have a large kinetic trans efect best explained by the rc back-bonding concept.
This study was also the first to get rate data on the ligand substitution reactions
of homologous complexes of metals of the nickel triad, and to determine the
kinetic trans efects of many different ligands.
4 The Kinetics and Mechanisms of Ligand

Displacement Reactions of Platinum(1V) Complexes
All of the research we and others had done on ligand substitution reactions of
six-coordinate octahedral complexes showed them to react by a dissociative sN1
mechanism. Instead, four-coordinate square-planar metal complexes react by an
sN2 process, because of the ease with which their coordination numbers can
increase, and because of a low energy p-orbital for attack by the entering ligand.
Furthermore, all studies on the analogous d2sp3 six-coordinate cobalt(m) com-
plexes react by an S,1 mechanism. Why then would we want to investigate the
substitution reactions of tran~-[PtCl~(NH,),]~+? Our reasoning was that the
higher oxidation state of platinum(1v) might make it more electrophilic than
cobalt(m), and result in the platinum(1v)complexes being more susceptible to an
sN2 nucleophilic attack. We were not wrong about this but, as is now to be seen,
we also found some other more interesting chemistry of platinum(1v)complexes.
There is a considerable amount of information on the preparations and
reactions of platinum(1v)complexes, largely done by Russian chemists. In spite of
this, there seems to have been only one quantitative kinetic study reported on
substitution reactions on these systems. Unfortunately, as is pointed out later,
the results of this investigation are difficult to interpret because of the complica-
tions due to photosensitivity and platinum(I1) catalysis of these reactions. Thus
the correlations attempted by the Russian workers7 for the series of complexes
they investigated should be viewed with some caution.
We followed spectrophotometrically Reaction (9).
tr~ns-[PtBr,(en),]~++ C1- --+ [PtBrCl(en),12+ + Br- (9)
It was observed that this reaction is photocatalysed so that all studies were then
made on reaction mixtures which were kept in the dark. Using an excess of
hydrochloric acid, the data were found to give good pseudo-first-order plots and
the rates showed a first-order dependence on the concentration of chloride
ion. However, different preparations of trans-[PtBr,(en),]Br, (en = 1,2-
F . Basolo 323
diaminoethane), all identical in analysis, gave quite different rate constants.
Recrystallisation of a particular sample also gave fractions which differed in their
reactivity.
From the work of Taube and King,” it was suspected that the different
reactivities of these presumably identical compounds were caused by catalytic
amounts of [Pt(en),12’. This was confirmed16 both by the addition of the
catalyst and also by the addition of cerium@) to destroy the catalyst. Addition of
five mole per cent [Pt(en)2]2’ results in complete reaction within five minutes of
a complex which would otherwise react in the dark at 25°C with ti = 44 min.
This same complex in the presence of cerium(1v) showed no detectable reaction-
under these conditions over a period of 24 h.
The extremely slow exchange of chloride ion in trans-[PtC12(en),] 2+ is
catalysed by [Pt(en)2]2’ with the attendant rate law (10).
Rate =k [PtCl,(en)22’][Pt(en)22+] [Cl-] (10)
However, the results can be explained in terms of the inner-sphere mechanism
[Equations (11)and (12)].
There is now a good deal of evidence2 for the addition of other groups to
square-planar complexes such as 4. There is also X-ray evidence2 for the exist-
ence of bridged platinum(I1)-X-platinum(1v) halogen complexes of the type 5
postulated in Equation (12). Furthermore, there is ample evidence to support the
view that certain oxidation-reduction reactions proceed by an atom-transfer
mechanism through a bridged intermediate analogous to 5. Such a halogen
atom-transfer mechanism is applicable to a one-electron redox p r o c e ~ s , ~but
not to a two-electron process of the type described here for the plati-
num(rI)-platinum(Iv) system. In this case, where the bridged intermediate in-
volves only one halogen, the net effect requires a halogenium ion transfer.
It follows from the proposed mechanism of chloride ion exchange in this
system that platinum exchange between the platinum(II)-platinum(Iv) species
should proceed at the same rate as does the chloride ion exchange.” Another
way of observing the same thing would be to start with 195Ptin one of the
complexes and follow the rate at which it is distributed between the two different
oxidation states of platinum. These experiments” have been done and support
the proposed mechanism.
Finally, it should also be pointed out that the ligand which adds to platinum(I1)
as in Equation (11)need not be of the same kind as those attached to platinum(1v)
Therefore, it follows that there may be a very common mechanism for the
syntheses of platinum(1v) complexes involving platinum(r1) catalysis. For
example, the method of preparation of trans-[PtX,(NH,),]X, is the oxidation of
[Pt(NH3),I2+, as shown by Equation (13).
[Pt(NH3),I2+ + ',"' tr~ns-[PtX,(NH,),]~+

7
(13)
- x2
Having prepared tr~ns-[PtX,(NH~),]~+, it can be allowed to react with a

solution of [Pt(NH,),I2+ plus Y - to form the desired tr~ns-[PtY,(NH,),]~'.
This method was used successfully2' to prepare tr~ns-[Pt(sCN),(NH,),]~+.
Also, the Russians were able to allow solutions of [PtCl(NH,),]Cl, in aque-
ous HCl to react under mild conditions to give quantitative yields of trans-
[PtCl,(NH,),]Cl,. This puzzled them, because they knew that NH, should not
be displaced by Cl- under the experimental conditions used. Now we can be
certain that the reaction proceeded rapidly to give quantitative yields of the
desired product due to the presence of catalytic amounts of [Pt(NH3),I2'.
5 The Application of Reaction Mechanisms for the

Synthesis of Metal Complexes
Joseph was interested in our research on the application of reaction mechanisms
to design the syntheses of new metal complexes and of known compounds. He
was naturally more concerned with platinum compounds, but as our work was
mostly with cobalt complexes, he would also listen to me tell him about our
research on these compounds.
One example will suffice to show the value of making use of reaction mechan-
isms in syntheses. This approach allowed us to prepare the new nitrito
(M-ONO) complexes [M(ONO)(NH,),]2' of rhodium(m) and of iridium(m),
although their nitro (M-NO,) compounds had been known for about a century.
The success we had with this started with our detailed kinetic study21 of the
apparently simple Reaction (14).
+
[COCI(NH,)~]~' NO,- 5 [CO(NO,)(NH,),]~+ + C1- (14)
We found the reaction to be anything but simple. The following five steps
appear to be involved.
[CoC1(NH,)5]2++ H,O -+ [CO(NH,),(OH,)]~+ + C1- (15)

[c0(NH,)5(oH2)]3' + H,O -+ [Co(OH)(NH,),]'+ + H30t (16)
2HN0, -+ N 2 0 3 + H,O (17)
[(NH3)5Co-OH]2t -[
N o
(NH,),Co-O - - - H
0-N - - NO,
[(NH,),CO-ONO]~+ HNO, +
I 1-
2+
(18)
F . Basolo 325
[CO(ONO)NH,),]~+-+ [CO(NO,)(NH,),]~' (19)
Of particular significance in this reaction scheme is the O-nitrosation step (18),

which suggests that the Co-0 bond is not cleaved and that the kinetic product is
therefore the unstable nitrito (Co-ONO) isomer. This then rearranges to the
stable nitro (Co-NO,) linkage isomer. Oxygen- 18 labelling experiments of Mur-
mann and Taube22later confirmed that there is no rupture of the Co-0 bond
during this process.
This mechanistic concept for the formation of nitrito-complexes of cobalt(II1)
suggests that other analogous metal systems should yield similar materials.
However, the corresponding nitrito-complexes of rhodium(II1)and of iridium(m)
were not known. One reason that previous investigators had not been successful
in preparing these is that the platinum group metal complexes are usually very
slow to react and rather drastic reaction conditions had been used. As a result,
the stable nitro-product rather than the kinetic nitrito-product was isolated.
Since the formation of M-ON0 does not involve M-0 bond cleavage,
the reaction as shown in (18) is expected to occur even under rather mild
experimental conditions. This was found to be the case and salts of the new
complexes [M(ONO)(NH,),]"' where M = rhodiurn(m), iridium(II1) or plati-
num(rv) have been prepared.,,
Sometimes amusing things happen when dealing with graduate students doing
PhD research. That occurred in this case. After I talked about this problem to
Geneva Hammaker, one of our few female graduate students at the time, she
went to the library to get directions for the preparations of the two starting
materials. In a few days she returned asking for a new problem because she read
that an attempt had been made to prepare [Rh(ONO)(NH,)J2+, but that it had
failed. I asked about the conditions of the experiments, and was told the reaction
mixture was kept at 100°C for a few hours and the product isolated was the
nitro-complex [Rh(N02)(NH,),]2'. I explained to her that this was to be
expected, because under those conditions one gets the thermodynamically stable
nitro-compound. She was almost in tears. Rather than suggesting a different
problem, I described to her in detail why I felt this preparation should work,
because the metal-oxygen bond was not broken. This indicates that the metal
may only have a small effect so the reaction to give the desired nitrito kinetic
product should take place under mild conditions. She reluctantly agreed to try it.
A few days later she came to my office and was a different person - all smiles and
joyous with the satisfaction that she had been able to make a new compound,
[Rh(ONO)(NH,),]Cl,. She then proceeded to prepare the corresponding irid-
ium and platinum complexes. She followed this by investigating the rates of
rearrangement of each to the thermodynamic nitro complexes, [M(N02)-
(NH3) 512+ *
Her paper was submitted to the new journal. Inorganic Chemistry, and has the
honour of being its first paper with the reference that I can easily remember,
Inorg. Chem., 1962,1,1.
6 Joseph Chatt the Person
I have always considered Joseph a friend, one with whom I could always discuss
chemistry and family. I am so very pleased that the Royal Society of Chemistry
and the British inorganic chemists are paying him the honour he justly deserves.
It is my opinion that he should be considered, ‘the Father of platinum(I1)
chemistry.’
I think the first time Joseph came to the US was in the late 1950s when we
invited him to give a lecture at the Gordon Research Conference. He took that
occasion to visit me at our department and give a seminar. In England, he is
called Joseph - in the US, Joe. I know of no other chemist who conducts his
research with greater care. Research scientists say, ‘If a scientist does not want to
make a mistake, albeit minor, he should not do research.’ To my knowledge, Joe,
by his careful research, proved this statement to be false.
The seminar he gave was on some elegant pioneering work on the syntheses,
reactions, and properties of some new complexes. Joe was very likeable and
pleasant, but formal. Therefore, it surprised me when he started his lecture by
telling this story:
On a commuter train to London there were two passengers who always caught the train
at the same time, and who often sat across from one another. One of the passengers had a
pad of paper and kept throwing a sheet out the window during the ride. After some days of
seeing this, the other passenger asked, “Why are you doing this?” His answer was “To
keep the lions away.” The questioner said, “But there are no lions in England,” and the
response was, “Yes, so you see it works.”
Joe wanted to make the point that in basic science it is not enough to know it
works, but one also needs to know why it works. He then proceeded to illustrate
the point in the talk about his elegant research.
Joseph and I began to see less of one another after our retirements. However,
each Christmas we exchanged news letters about our doings during the year. His
letter was always much longer than mine, as he told me about his coin collection
and his and Ethel’s holiday cruises. I miss getting his letter each year, but I think
of him often during the season of Christmas and the New Year.
I am humble but very honoured to have received the first Joseph Chatt Medal
from the Royal Society of Chemistry.
7 References
1 F. Basolo, Coord. Chem. Rev., 1996,154, 151.
2 F. Basolo and R.G. Pearson, Mechanisms ofInorganic Reactions, John Wiley and
Sons, New York, 1’’ edn. 1958, 2”d.edn 1967, and references therein.
3 F. Basolo, J. Chatt, H. B. Gray, R. G. Pearson and B. L. Shaw, J . Chem. Soc., 1961,
2207.
4 I. I. Chernyaev, Ann. Inst. Platine (USSR),1926,4, 243,261.
5 (a) J. Chatt and L. A. Duncanson, J . Chem. SOC.,1953, 2939; (b) J. Chatt, L. A.
Duncanson and L. M. Venanzi, J . Chem. Soc., 1955,4456.
6 M. J. S. Dewar, Bull. SOC.Chim. Fr., 1951, C79.
F . Basolo 327
7 (a) A.A. Grinberg, Ann. Inst. Platine(USSR),1932,10, 58; (b) Y. K. Syrkin, Bull. Acad.
Sci. USSR Classe Sci. Chim., 1948,69.
8 (a) R. Romeo, personal comments; (b) S. Lanza, D. Minniti, R. Romeo, P. Moore, J.
Sachinidis and M. L. Tobe, J . Chem. Soc., Chem. Commun., 1984,542; (c) S. Lanza, D.
Minniti, P. Moore, J. Sachinidis, R. Romeo and M. L. Tobe, Znorg. Chem., 1984,23,
4428.
9 M. A. Tucker, C. B. Colvin and D. S. Martin Jr., Inorg. Chem., 1964,3, 1373.
10 S. Ahrland, J. Chatt and N. R. Davies, Q. Reu. Chem. Soc., 1958,12, 265.
11 R. G. Pearson, J . Am. Chem. Soc., 1963,85,3533; Science, 1966,151,172.
12 C. G. Swain and C. B. Scott, J . Am. Chem. Soc., 1953,75141.
13 J. 0.Edwards, J . Am. Chem. Soc., 1954,76,1540; J. 0. Edwards, 1956,78,1819.
14 U. Belluco, L. Cattalini, F. Basolo, R. G. Pearson and A. Turco, J Am. Chem. SOC.,
1965,87,241.
15 H. Taube and E. L. King, J . Am. Chem. Soc., 1954,76,4053.
16 F. Basolo, M. L. Morris and R. G. Pearson, Faraday Soc. Disc., 1960,29,80.
17 R. L. Rich and H.Taube, J . Am. Chem. Soc., 1954,76,2608.
18 F. Basolo, P. H. Wilks, R. G. Pearson and R. G. Wilkins, J . Znorg. Nucl. Chem., 1958,6,
161.
19 L. T. Cox, S. B. Collins and D. S. Martin, Jr., J . Znorg. Nucl. Chem., 1961,17, 383.
20 R. C. Johnson and F. Basolo, J . Inorg. Nucl. Chem., 1960,13,36.
21 R. G. Pearson, P. M. Henry, J. G. Bergmann and F. Basolo, J . Am. Chem. Sue., 1954,
76, 5920.
22 R. K. Murmann and H. Taube. J . Am. Chem. Soc., 1956,78,4886.
23 F. Basolo and G. S. Hammaker, Inorg. Chem., 1962,1, 1.
Tuning Rhodium(1) Metal
Centre Accessibility in
Iodomethane Oxidative Addition
to Vaska-type Complexes by
Inter changing Tertiary Phosphine
for Arsine and Stibine
S. OTTO, S. N. MZAMANE AND A. ROODT
University of the Free State, P.O. Box 339, Bloemfontein, 9300, South Africa
1 Introduction
The complex, truns-[IrCl(CO)(PPh),),1, which was first reported by Angoletta'
and later correctly formulated by Vaska,2 is one of the best known early
organometallic complexes exhibiting catalytic activity. Interestingly enough, at
that time the rhodium analogue was already known3 and had been investigated
to some e ~ t e n t . ~
These complexes are typical of those studied by Chatt, which enabled bonding
theories and their applications to be d e ~ e l o p e dand
, ~ were soon recognised as
important model systems for studies on homogeneous catalysis. These d8
square-planar systems thus undergo a range of reactions, such as oxidative
addition, with different substrates.6 Ironically, today complexes with the general
formula trans-[MCl(CO)(L),] (M = Rh or Ir; X = halide or pseudo halide;
L = neutral ligand) are still in many instances known as analogues of Vaska's
c ~ m p l e x ,in~ spite
* ~ of important earlier contributions by Chatt and co-workers.
The rhodium complexes are more resistant towards oxidative addition than
their iridium counterparts and this is believed to be linked to steric crowding,
especially when employing bulky tertiary phosphine ligands. Work by Wilkin-
son*p1oexplored several aspects of the steric and electronic properties of the
rhodium@)analogues, but definite crystal structural confirmation of the reaction
S. Otto, S. N . Mzamane and A . Roodt 329
products could not be achieved. In this paper we address several of these
'
aspects' and outline steric implications of analogous Group 15 donor ligands
(P,As and Sb) - including the incorporation of methyl substituents on the phenyl
rings of these ligands, which results in interesting isomorphic structures.
2 Experimental
For experimental procedures utilised in this work, i.e. synthetic,l29' UV-vis
spectr~photometry,'~ kinetic analyses,' IR/NMR spectroscopy' and X-Ray
the reader is referred to earlier work.14 The data for the
crystal structures reported herein were solved using standard techniques and
'
details have been reported. 8-21
3 Reaction Scheme
The general stoichiometry of the reaction is shown in Scheme 1. The equilibrium
reaction defining the first step (forward = oxidative addition; reverse = reduc-
tive elimination) was verified previously.' 0,14 Figure 1 illustrates the oxidative
addition of Me1 to [RhCl(C0){Y(p-T01),)~1.
Scheme 1 General reaction pathway for the iodomethane oxidative addition and successive
migratory CO insertion reactions in Vaska-type compounds (Y=P ( I ) , As (2))
The slow transformation of the tr~ns-[RhCl(CO){As(p-Tol)~}~] complex, 2,
(Figure 2) (v(C0) = 1973 cm-'), and the clean conversion to the intermediate
2100 1900 1700

Wavenumber / cm-'
Figure 1 Infrared spectra (25°C; 10 min intervals) of iodomethane oxidative addition to (a)
[RhCl(CO)(As(p-Tol),}J (2) and (b) [ R ~ C I ( C O ) ( P ( ~ - T O(l),
~ )illustrating
~)~]
the formation of the rhodium(llr)-alkyland rhodium(llI)-acylspecies; CHCI,,
[Rh] = 0.02 M ; [CH,I] = 3 M
330 Tuning Rhodium(1)
Figure 2 Ortep drawing of trans-[RhCl(CO)(As(p-Tol),} J (2) (50% probability ellipsoids;

hydrogen atoms omitted for clarity) showing the numbering scheme. Thefirst
digit of the carbon atom numbering corresponds to the phenyl ring number and
the second to the atom in the ring
alkyl species (v(C0) = 2050 cm-'), is apparent. However, in the case of the
trans-[RhCl(CO){ P(p-Tol),) 2 ] complex, 1, it is clear that under identical condi-
tions the reactant is converted to the intermediate alkyl species only to a small
extent, thus confirming its thermodynamic unfavourability. This probably also
explains why the intermediate P-alkyl complex could not be isolated, compared
to the corresponding As species, which has indeed been obtained in the solid
state and structurally characterised (see Figure 3).
4 Structural Aspects
The reactants as well as the intermediate alkyl species in this reaction have been
characterised. Thus, the difficulty associated with interpreting NMR spectra of
As and Sb complexes (arising from the unfavourable properties of the nuclei) has
been partially circumvented by X-ray structural analysis.
4.1 Rhodium(1) Complexes

The molecular structure of 2, shown in Figure 2, is isomorphous and thus
virtually identical to the P(p-Tol), analogue." 1 and 2 are compared in Table 1.
As shown in Table 1, the As-Rh-As moiety in 2 (4.824(1) A)is significantly
longer than that of P-Rh-P in 1 (4.665(1)A).This in turn means there is a larger
'cavity' for entering nucleophiles in the arsine Vaska-type complex compared to
the sterically more hindered phosphine analogue.
The increase in the length of Rh-Cl bonds from complex 1 and 2 suggests more
electron density on the Rh-centre in complex 2. The Rh-C(l) and CEO bond
distances in the isomorphous complexes 1 and 2 do not differ significantly.
S. Otto, S. N . Mzamane and A. Roodt 331
Table 1 Selected interatomic bond lengths (A) and angles (") in complexes 1 , 2
and 3
_ _ _ _ _ _ _ _ _ ~ -
Rh-C( 1) 1.798(5) 1.817(5) 1.841(13)

Rh-C1 2.358(2) 2.393(2) 2.403(3)
Rh-L( 1) 2.334(2) 2.4181(4) 2.4664(11)
Rh-L(2) 2.331(2) 2.4 169(4) 2.4727(11)
L(l)-C(n 1.821(5) 1.940(5) 1.944(9)
L(2)-C(n 1.823(5) 1.942(5) 1.942(9)
C(1)-0(1) 1.139(6) 1.082(5) 1.174(14)
Rh-C(2) -
2.280(8)
Rh-I __ -
2.7860(11)
L(l)-L(2) 4.665(5) 4.835(1) 4.939(2)
L(1)-Rh-L(2) 175.67(4) 175.9 1(2) 176.66(4)
C( l)-Rh-Cl 173.11(19) 174.2(2) 178.9(5)
C(1)-Rh-L(1) 90.64(15) 90.77(13) 9 1.0(4)
C( l)-Rh-L(2) 9 1.77(15) 92.11(13) 90.6(4)
C1-Rh-L(1) 89.84(4) 89.14(3) 89,59(6)
Cl-Rh-L(2) 88.19(4) 88.29(3) 88.76(6)
O(1)-C(1)-Rh 177.3(10) 174.4(5) 174.8(15)
C(1)-Rh-I - -
80.7(5)
AS(1)-Rh-I 93.53(4)
C1-Rh-I 100.32(7)
C(2)-Rh-I 165.6(2)
As(2)-Rh-I 89.62(4)
a) Ref 11
These compounds are some of the few structurally characterised Vaska-type

complexes not showing a disorder along the carbonyl/chloro axis. The
L(l)-Rh-L(2) bond angles for these are 175.91(2)"and 175.67(4)"for L = As@-
Tol), and P(p-Tol), respectively, which, being significantly smaller than 1SO",
enables them to crystallise in the non-centrosymmetric space group, Pna2,. The
Flack parameters are zero within experimental error, indicating that the correct
stereochemical isomer was refined.
As i n d i c a t e d " ~ ~and
~ compared in Table 2, not only are 1 and 2 isomor-
phous with one another, they are also isomorphous with trans-[IrCl(CO){ P(p-
Tol),) 2],1 trans-[Pt(CH,)Cl{As(p-Tol), ] 2 ] 2 2 and trans-[Ir(CH,)(CO){ P(p-
TO^),),].^^ This range of metal centres, oxidation states, donor atoms and
combinations of trans ligands is thus quite novel. It seems that incorporation
of the 4-methyl substituents on the aryl rings of the Group 15 L tertiary ligand
causes such space demands that, during crystallisation, the space requirements
of the halide, C O etc. moieties are overridden. Similar isomorphism is not
observed in the case of the corresponding [RhCl(CO)(YPh,),] complexes
(Y = P, As or Sb) .24
Of further interest is the fact that the Rh-P and Ir-P bond lengths in the
structures listed in Table 2 are virtually identical. However, changing the Group
15 donor atom results in significant increases in the Rh-L bond length {from 2.33
w
w
F3
Table 2 Comparison of bond data and spectroscopic properties in trans-[MX(CO)L,] (M = Rh or Ir;

X = Cl or C(O)CH,; n = 2 or 3)
M-X v ( C O ) ~ M-CO L-M-L Ref

Complex x" (A) (ern-') (A) (")
2.322( 1) 2.382(1) 1979 1.77(1) 1.140(2) 180.0(1) 25
2.349 l(7) 2.388(2) 1943 1.748(8) 1.163(7) 180 26
2.3344(14) 2.415(7) 1970 1.814(14) 1.056(14) 180 27
2.3426(7) 2.443(7) 1964 1.73l(9) 1.15(2) 180 27
2.333(2) 2.3581(12) 1976 1.798(5) 1.139(6) 175.67(4) 11
2.4175(4) 2.393(2) 1973 1.817(5 ) 1.082(5) 175.91(2) b
2.4226(4) 2.3538(14) 1975 2.017(7) 0.717(7) 175.97(6) 27
2.5655(2) 2.315(3) 1971 1.797(13) 1.175(13) 180 27
2.598l(5) 2.4094(18) 1971 1.87 5(7) 1.035(6) 119.97(2) 27
2.568(2) - 1710 1.91l(20) 1.12 l(25) 120.0(1) 28
2.330(1) 2.382(3) 1950 1.791(13) 29,30
2.33 l(2) 2.364(2) c 1.817(8) 1.134(10) 175.19(2) 31
2.345(2) 2.398(7) 1934 1.78(2) 1.10(2) 180 32
a Dichloromethane solution
This work
Not reported
(P) to 2.42 (As) to 2.58 A (Sb)). The relatively easy formation of five-coordinate
stibine-containing c ~ m p l e x e smay~ ~ be
* ~a ~consequence of these changes.
In Table 2, comparisons between the complexes investigated in this study and
a few other relevant iso-structural complexes from the literature suggest that the
introduction of ligands with heavier donor atoms, such as arsines and stibines,
leads to reduced steric crowding at the metal, creating more space for entering
moieties in reactions such as oxidative addition.
To further illustrate the steric effects induced by these tertiary Group 15
ligands, we have calculated the Tolman angles3, (&, using 2.28 A for standar-
dised Ni-P bonds) and the effective cone angles (OE, using true bond distance~).,~
8, = 140 and 133" and QE = 140 and 132" for AsPh, and SbPh,, respectively.
These values are more than 10"less than the corresponding values for PPh, (we
assume that para substitution has negligible effect on the cone angle). The
increase in P-C, As-C and S b C bond lengths (Table 1) from 1.82 to 1.94 to 2.10
A27and reduction in the C-L-C angles for the Group 15 elements, 116,109,102
and 97" for N to Sb, re~pectively,~' result in an increased ability of the ligand to
'fold back' from the metal.
Whether the decrease in the Rh-Cl bond length in [RhCl(CO)(YPh,),]
(Y = P, As, Sb; Table 2) is a result of the decrease in electron density at the metal
centre or arises from the decrease in steric crowding of the Group 15 donor
ligand as discussed above is, however, not clear at this point.
As the electron-donating ability of the Group 15 donor L ligand decreases, the
electron density on the metal decreases, resulting in a weaker M-CO bond, a
stronger CZEO bond and a small but significant increase in v(C0) (in di-
chloromethane). The net change in electron density is further illustrated by the
reactivity change of the complexes towards oxidative addition (Table 3).
4.2 Rhodium(II1) Alkyl Complexes

The molecular structure of the intermediate alkyl complex, trans-
[R~(CH,)C~(I)(CO)(AS(~-TO~)~)~]
3 (Figure 3), clearly shows that trans addition
of iodomethane has occurred, as has been found previously14 in systems with
sterically congested and electron-rich metal centres.
Complex 3 represents one of the very few characterised arsine complexes of
this type, and of interest is the fact that the Rh-As bonds were lengthened by
almost 0.05 A compared to its precursor, 2, increasing the As-Rh-As distance
from 4.835(1) to 4.959(2) A.The Rh-Cl bond stayed virtually unaffected. The
increase in the Rh-CO bond length, while of borderline statistical significance, is
consistent with the higher v(C0) value of 2050 cm-'.
The rhodium(II1)-I bond distance in 3 (trans to Me) is longer than those (trans
to PPh,) in [Rh(CH,)I(CO)(PPh,)(cupf)] (Hcupf = cupferron; N-nitrosophenyl-
) ~ ~ [Rh(CH,)I(CO)(PPh,)(quin)] (Hquin = 2-quinolinecar-
h y d r ~ x y l a m i n e and
boxylic acid)37(from the cis addition of iodomethane), though it is significantly
shorter than for other trans-CH,-Rh-I moieties, such as the average value of
2.803(1) A in [Rh(CH,)I(CO)(PPh,)(oxin)] (Hoxin = 8-hydroxyq~inoline)~~
and 2.849(1) A in [Rh(CH,)I(CO)(PPh,)(dmavk)] (Hdmavk = dimethyl p-
Figure 3 Ortep drawing of trans-[Rh(CH,)Cl(I)(CO){As(p-To/),}J (3) (50% probability

ellipsoids; hydrogen atoms omitted for clarity) showing the numbering scheme.
TheJirst digit of the carbon atom numbering corresponds to the As atom, the
second to the ring and the third to the atom in the ring
amin~vinylketone).~~ The metal centres in the latter examples display high

reactivity toward iodomethane, with half-lives (for [CH,I] = 1 M) in the second
range. The slower kinetics observed for iodomethane oxidative addition to 2 and
the shorter Rh-I bond length in 3 may be related.
5 Rate Laws for Reaction

Iodomethane oxidative addition to rhodium(1) complexes proceeds according to
the general m e c h a n i ~ m lgiven
~ . ~ ~in Scheme 1, followed by a migratory carbonyl
insertion. The reverse of the first step represents reductive elimination. The
observed pseudo-first-order rate constant ([CH,I] >> [Rh]) for the first step in
Scheme 1 is defined by Equation 1.
( k d =~~oA[CH,II
~ -k RE (1)
The kinetic constants k,, and kRE refer to the rates of the oxidative addition
and reductive elimination reactions, respectively. The equilibrium constant for
the first step is given by Equation 2.
KO, = koAlkRE (2)
Oxidative addition and reductive elimination were studied in a range of
solvents and different CH,I concentrations at three different temperatures using
different kinetic techniques (IR and UV-vis). The rate of rhodium(II1)-alkyl
formation shows a direct relationship of the pseudo-first-order rate constant on
[CH,I], as illustrated in Figure 4.
12
4 / 25 OC
7
Iv)
\
5
7n i o
Y
v
0
z 3
0 1 . 1 - 1 -
0.2 0.4 0.6 0.8

[CH31] I M
Figure 4 [ C H J ] and temperature dependence of the pseudo-Jirst-order rate constantfor

the iodomethane Oxidative addition to: (a)trans-[RhCl(CO){As(p-Tol),},] and
(b) trans-[RhCI(CO){P(p-Tol),},] in acetone
These results are also supported by time-resolved IR spectroscopy studied

under identical conditions (Figure 1).The conversion to the alkyl species from 1
is clearly less favoured than that from 2. In the latter case the reaction proceeds
quite efficiently to the alkyl species (v(C0) = 2050 cm-l), while the conversion of
[RhCl(CO){P(p-Tol),)J to the rhodium(II1)-alkyl complex is thermodynami-
cally less favoured.
An aim of this study was to investigate the effect of structural aspects on the
reactivity of the two Vaska-type complexes. It was shown crystallographically
that 2 is less sterically crowded than 1, since introduction of the larger arsenic
atom enlarges the ‘cavity’ in the complex to accommodate entering moieties,
such as in oxidative addition of iodomethane. This increase in ‘cavity’ size is
expected to shift the equilibrium more toward the intermediate, resulting in a
larger KO, for 2 than for 1 as shown in Table 3. Of interest is the fact that
oxidative addition is thermodynamically favoured for 2 compared to 1 in both
dichloromethane (KOA 9 us. 1.7 M-’) and acetone (15 us. 2.3 M-’). However, in
ethyl acetate and toluene values of KO, are, strangely enough, quite similar
though of lower precision.
The expression for the observed rate constant for the formation of the
rhodium(1rr)-acylspecies via migratory carbonyl insertion ([CH,I] >> [Rh]), as
w
Table 3 Temperature and solvent dependence of rate constants (Scheme 1)for the oxidative addition of CH,I to m
trans-[RhCl(CO){Y(p-Tol),},] ( Y = P , As and Sb"; AC = acetone, EA = ethyl acetate,
DCM = dichloromethane, TOL = toluene)
P(P-To03 A4P- To03

Temp
Constant ("C) AC EA DCM TOL AC EA DCM TOL
20.7 6.0 8.9 2.38 20.7 6.0 8.9 2.38
17.0 17.1 4d 0.1" 17.0 17.1 4d 0.1"
- - - - - - -
5 0.08(4)
15 OS(2) 0.09(3) 0.56(1) - 1.32(5) 0.1 l(4) 0.22(1) -
25 1.30(6) 0.24(2) 2.90(9) 0.20(1) 4.2(2) 0.84(4) 0.60(1) 0.05(2)

35 - - 13.3(4) 2.03(9) -
2.78(8) 6.86(6) 3.8(1)
- -
58(3) 72(12) 85U) 115(5) 87(6) -
- lOO(9) -48(30) - 86(15) .8(18) -
- 1(4)
5 - - - - - - -
0.20(1)
15 0.366(9) 0.362(4) 0.54(1) - 0.25( 2) 1.13(1) 0.029(4) -
25 0.75( 2) 0.880(3) 1.27(4) 1.63(1) 0.52(6) 5.1l(1) 0.041(4) 0.445(2)
35 1.22(3) - 1.61(4) - 142) 15.1(6) O.lO(4) -
- -
39(4) 30(12) 53(1) 83(6)
- 167(14) - 193(40) - - 123(5) 3(18)
25 1.7(1) 0.27(2) 2.3(1) 0.12(1) 9(1) 0.16(1) 15(1) 0.11(4)
25 1.8(3) 0.28(6) 2.5(5) - 0.17(3) 14(3) -
9(2)
a) Ref 27. In neat CH,I at 25"C, estimated values for SbPh, complex: k,, ca. 3 x M-'s-'; K O , > 100 M-'
') Ref 40
') Ref 41
d, Estimated from chloroform

e , Estimated from benzene
f, From Equation (2)
g, Average of three temperatures
illustrated in Scheme 1,is shown in Equation (3), where k,, and k-cl represent the
rate constants for the forward and reverse steps for the migratory carbonyl
insertion reaction, respectively.
The formation of the final rhodium(II1)-acyl species has been observed and
identified (Figure l), though was not studied in detail.
6 Reactivity and Activation Parameters for

Iodomethane Oxidative Addition
Depending on the solvent, the relative reactivities of 1 and 2 (as well as of
[RhCl(CO)(SbPh,),]) towards the oxidative addition and reductive elimination
span ranges of up to two orders-of-magnitude (Table 3). The half-lives for the
reaction at ambient temperature and [CH,I] = 1 M, range from a few minutes
to hours, and are comparable with those observed for [RhCl(CO)(YPh,),]
(Y = P or As)."
1 is more reactive in the less polar/coordinating solvents dichloromethane and
toluene, but the reaction is thermodynamically less favoured. In the more polar
and coordinating acetone and ethyl acetate the As complex is actually a factor of
4-5 more reactive than the P analogue, which suggests increased solvent contri-
bution in the As complex (see discussion below).
The fact that the [RhCl(CO)(SbPh,),] complex shows comparable reactivity
to 2 suggests similar electron density on the Rh-centre in both. However, the Sb
complex shows a significant increased reactivity toward solvents, which is at-
tributed to the smaller steric demand at the metal centre.27
The formation of the rhodium(rI1)-acylspecies (see Figure 1)was not studied in
any detail. However, at low [CH,I], it follows from Equation 3 that the rate of
formation of the acyl species is first-order in [CH,I], and directly proportional
to the equilibrium constant KoA. Thus, in the case of 1, the effective rate of
formation of the acyl species is decreased (by at least a factor of five), as the
conversion of 1 to the intermediate [R~(CH,)C~(I)(CO){P(~-TO~)~},] species is
thermodynamically less favoured.
The activation parameters (Table 3) are characterized by positive values of
AHfkoAand negative values of ASfk9, and indicate that bond-formation plays an
important role in forming the transition state. This is in agreement with previous
work which showed that oxidative addition proceeds via an associative mechan-
ism.42 For 2 in ethyl acetate, the oxidative addition rate constants could not be
determined accurately, but k,, could be used for the calculation of the activation
parameters. The values suggest less ordered transition states in which significant
solvent interaction may occur, but it is clear that additional research is still
required.
338 Tuning R hodium(1)
7 Solvent Effects
The arsine complex, 2, shows a two orders-of-magnitude variation in the koA, k,,
and K,, values in the four solvents studied. The P analogue, 1, however, shows
only ca. one order-of-magnitude variation in these constants, with a two-fold
variation in k,, in the four solvents.
An increase in solvent polarity (as manifested in the dielectric constant, E ) for
solvents with roughly similar solvent doriicity (from the donor number DN),that
is, comparing ethyl acetate to acetone and toluene to dichloromethane, results in
an increase in both k,, and K O , for both 1 and 2.However, kR, for 1 shows very
little solvent dependence (a decrease of only ca. 1.2-fold is observed from ethyl
acetate to acetone, and 1.5-fold from toluene to dichloromethane) whereas for
,k for 2 a one order-of-magnitude increase is observed. We conclude that the
reductive elimination is inhibited in solvents of increased polarity, whereas
oxidative addition is favoured. As a result K O , for 2 spans a larger range of
values compared to 1.
Increasing the coordinating ability of the solvent (i.e. dichloromethane to ethyl
acetate) with relative constant polarity increases k,, in the As complex by two
orders-of-magnitude, while leaving k,, virtually unaffected. In the P complex,
however, a change in the donicity of the solvent has a quite significant effect on
the values of k,, (about 15-fold increase), compared to that of kRE (only a ca.
two-fold increase). This might be indicative of competition between the CH,I
and the solvent in complex 2 where there is less steric congestion at the metal
centre. The different steps in the reaction sequence are thus apparently in-
fluenced differently in the two complexes, but the net effect on K O , is the same.
Moreover, it cannot be excluded that the intercept in the equivalent to Figure
4(b) for ethyl acetate (so large compared to the other rate constants in general)
actually incorporates contributions from both the reductive elimination as well
as a possible parallel solvent pathway.
Concluding Remarks
It was shown that these Vaska-type complexes of rhodium@)are excellent model
complexes to study basic effects, including the use of molecular structure infor-
mation to explain solution effects on reaction mechanisms. However, additional
research is still required to aid in understanding solution behaviour of these
systems.
9 Acknowledgements
Financial assistance from the South African N R F and the research fund of the
University of the Free State is gratefully acknowledged. Tania Hill is also
thanked for experimental assistance.
10 References
1 M. Angoletta, Gazz. Chim. Ital., 1959,89,2359.
2 L. Vaska and J. W. DiLuzio, J . Am. Chem. Soc., 1961,83,2784.
3 L. Vallarino, J . Chem. Soc., 1957,2287.
4 J. Chatt and B. L. Shaw, Chem. Ind., 1961,290.
5 J. Chatt and L. A. Duncanson, J . Chem. Soc., 1953,2939.
6 See for example: H. A. Zahalka and H. Alper, Organometallics, 1986,5,2497.
7 See for example: M. Selke, W. L. Karney, S. I. Kahn and C. S. Foote, Inorg. Chem.,
1995,34, 5715, and refs. within.
8 M. C. Baird and G. Wilkinson, J . Chem. SOC., Chem. Commun., 1966,267.
9 M. J. Mays and G. Wilkinson, J . Chem. SOC.,1965,6629.
10 I. C. Douek and G. Wilkinson, J . Chem. Soc. (A),1969,2604.
11 S. Otto, S. N. Mzamane and A. Roodt, Acta Cryst., 1999, C55,67.
12 Typical synthesis: trans-[RhCl(CO){ P(p-Tol),),] and trans-[RhCl(CO){As(p-
Tol),),]: [{Rh(p-Cl)(CO),},] (25 mg, 0.064 mmol) was dissolved in acetone (3 cm3)
and the appropriate ligand (0.28 mmol) also dissolved in acetone (4 cm3) was added.
Slight effervescence can be observed as C O is liberated from the starting compound.
Crystals suitable for X-ray analysis were obtained by slow evaporation of these
solutions. 1: v(CO)/CH,Cl, = 1976 cm-' v(CO)/KBr = 1968 cm-'; 'H NMR: 2.32
(singlet,para-CH,, 18 H); 7.14 (doublet, meta 12 H, 3JH, = 8 Hz); 7.57 (multiplet, ortho
12 H); 31P NMR: 54.36 (,J,,, = 118 Hz); 2 v(CO)/CH,Cl, = 1973 cm-';
v(CO)/KBr = 1966cm-'; 'H NMR: 2.37 (singlet, para-CH,, 18 H); 7.19 (doublet, meta
12 H, ,JHH = 8 Hz); 7.57 (doublet, ortho 12H, 3JHH= 8 Hz).
13 D. D. Perrin and W. L. F. Armarego, Purification ofLaboratory Chemicals, 1988,
Third Edition, Butterworth-Heineman Ltd, Oxford.
14 A. Roodt and G. J. J. Steyn. Res. Research Dev. Inorg. Chern., 2000, Vol. 2, Chapter 1,
S.G. Pandalai (Ed), Transworld Research Network, Trivandrum, p. 1.
15 SCIENTIST for Windows, 1990, program for least-squares parameter estimation,
version 4.00.950, Micromath, Utah, USA.
16 G. J. J. Steyn, A. Roodt, A. Poletaeva and Y. S. Varshavsky, J . Organomet. Chem.,
1997,536537,197.
17 General for 2 and 3; Three dimensional intensity data collected on a Bruker SMART
CCD diffractometer at 293(2)K; MoK, radiation (0.71073A);reflections corrected for
Lorentz and polarisation effects and absorption corrections applied using SADABS;
structures solved by direct method and successive Fourier synthesis (SHELXS-97 and
SHELXL-97); hydrogen atoms placed in calculated positions with fixed isotropic
thermal parameters; R = [(CAF)/(CF,)]; wR = C[W(F,~-F,~)~]/C[W(F~~)~]~;
2; Emp. formula C,,H,,As,ClORh; FW. 862.97; crystal system Orthorhombic; space
group Pna2,; a 21.8704(10); b 10.6662(5); c 16.8577(8) A; V/A3 3932.5(3); 2 4;
D,/g.cm-, 1.458; p/mm-l 2.203; TmaxlTmin 0.681/0.804; F(000) 1744; crystal
size/mm 0.68 x 0.54 x 0.34; 6 limit/" 1.86 to 28.27; index ranges - 12 2 h 5 28;
- 12 I k I 12; - 21 I 1 I 22; collected refl. 16086; independent refl. 7961; Rint
0.0217; Obs. refl.[I > 2a)I] 7161; data/restr./param. 7960/1/440; goodness of fit 1.088;
R/wR ( I > 4(0)1) 0.0289; 0.0590; R/wR (all data) 0.0361; 0.0629; Flack parameter
0.017(7);Ap,,,; Apmi,/ e k 3 0.466; - 0.403.
3; Emp. formula C,,H,,As,ClIORh; FW. 1004.90; crystal system Triclinic; space
group P I ; a 11.1449(6); b 12.5765(7); c 17.4452(10) A;o! 75.795(1); 87.296(1); y
75.042(1)"; V / A 3 2289.8(2); 2 2; D,/g cmP3 1.458; p/mm-' 2.568; Tmax/Tmin
0.458/0.861; F(000) 996; crystal size/mm 0.36 x 0.10 x 0.06; 6 limit/" 1.20 to 25.03;
index ranges - 13 I hI 12; - 14 I kI 14; - 20 I I < 14; collected refl. 12107;
independent refl. 7669; Rint 0.0241; obs. refl.[Z > 2411 6085; data/restr./param.
7669/0/459; goodness of fit 1.213; R/wR ( I > 4(0)I) 0.0712; 0.1347; R / w R (all data)
0.0885; 0.1431; Apmax;Apmin/e.k31.648; - 1.459.
18 G. M. Sheldrick, SADABS, 1999, program for absorption corrections to area detector
data, University of Gottingen, Germany.
19 G. M. Sheldrick, SHELXS-97,1997, program for solving crystal structures, University
of Gottingen, Germany.
20 G. M. Sheldrick, SHELXL-97, 1997, program for refining crystal structures, Univer-
sity of Gottingen, Germany.
21 K. Brandenburg, DIAMOND, 1998, computer graphics program, Germany.
22 S. Otto and A. Roodt, Acta Cryst. Sect. C, 1996,52,1636.
23 T. S. Janik, M. R. Churchill, R. F. See, S. L. Randall, J. M. McFarland and J. D.
Atwood, Acta Cryst. Sect. C , 1992,48, 1493.
24 S. Otto, PhD Thesis, Free State University, Bloemfontein, South Africa, 1999.
25 K. R. Dunbar and S. C. Haefner, Inorg. Chem., 1992,31,3676.
26 J. A. Venter, personal communication.
27 S. Otto and A. Roodt, Znorg. Chim. Acta, submitted for publication.
28 G. J. Lamprecht, C. P. Van Biljon and J. G. Leipoldt, Znorg. Chim. Acta, 1986,119, L1.
29 M. R. Churchill, J. C. Fettinger, L. A. Buttrey, M. D. Barkan and J. S. Thompson, J .
30 M. Selke, C. S. Foote and W. L. Karney, Inorg. Chem., 1993,32,5425.
31 M. R. Churchill, J. C. Fettinger, B. J. Rappoli and J. D. Atwood, Acta Cryst. Sect. C,
1987,43,1697.
32 E. Kuwabara and R. Bau, Acta Cryst. Sect. C, 1994,50,1409.
33 C. A. Tolman, Chem Rev. Sect. C , 1977,77,313.
34 S. Otto, A. Roodt and J. Smith, Znorg. Chim. Acta, 2000,303,295.
35 A. N. Sobolev, I. P. Romm, V. K. Belsky and E. N. Guryanova, J . Organomet. Chem.,
1979,179,153.
36 S. S. Basson, J. G. Leipoldt, A. Roodt and J. A. Venter, Znorg. Chirn. Acta, 1986, 118,
L45.
37 M. Cano, J. V. Heras, M. A. Lobo, E. Pinilla and M. A. Monge, Polyhedron, 1992,11,
2679.
38 K. van Aswegen, J. G. Leipoldt, 1. M. Potgieter, G. J. Lamprecht, A. Roodt and G. van
Zyl, Transition Met. Chem., 1991, 16, 369.
39 L. J. Damoense, PhD Thesis, Free State University, Bloemfontein, South Africa, 2000.
40 J. Rydberg, Principles and Practices of Solvent Extraction, 1992, J. Rydberg, C.
Musikas and G. Choppin (Eds.), Marcel Dekker Inc., New York, p. 23.
41 M. Sandstrom, I. Persson and P. Persson, Acta Chim. Scand., 1990,44,653.
42 See for example, J. A. Venter, J. G. Leipoldt and R. van Eldik, Znorg. Chem., 1991,30,
2207.
SECTION H:
Other Papers Presented at the

34th International Conference on
Coordination Chemistry,
Edinburgh, Scotland, July 2000
The health and dynamism of contemporary coordination chemistry is well exempli-

Jied in the many and varied lectures and posters presented at the 34th ICCC. This
would have pleased Joseph Chatt. He would also have been gratijied to see his own
contributions recognised in the session in his honour.
Papers presented in the ‘Joe Chatt Chemistry’sessionillustrate the breadth of his
contributions and their continuing legacy as well as the new directions, approaches
and techniques, unanticipated by Chatt, but which would have excited his interest.
Authors include many of Chatt ’s co-workers, collaborators and former students as
well as those acknowledging a scientijic debt to his pioneering work. It is sad to
recall that we had hoped that Luigi Venanzi would have been among them but his
illness and later passing prevented this.
Of the 22 plenary and invited lectures presented in the Joe Chatt Chemistry
session at Edinburgh, 18 are included in this volume, some in amended and expanded
form. Three of these papers we believe to be suficiently distinctive for it to be
inappropriate for them to be assigned to one of the earlier sections. These concern
the use of electrospray mass spectrometry to study the processes of eliminationfrom
carbonyl clusters of ruthenium and iridium described by J. S. McIndoe and co-
workers, the synthesis and characterisation of complexes containing the { Pt,S,}
core (also reporting the use of electrospray mass spectrometry) reported by T. S. A .
Hor and colleagues and of squarate coordination complexes, described by F.
Dumitru and colleagues.
Formaldehyde Elimination from
Methoxylated Transition Metal
Carbonyl Clusters
PAUL J. DYSON," BRIAN F. G. JOHNSON,bJ. SCOTT
McINDOE,' DUNCAN SAMBROOKbAND PATRICK R. R.
LANGRIDGE-SMITH'
aDepartment of Chemistry, The University of York, Heslington, York YO10
5DD, UK
Department of Chemistry, The University of Cambridge, Lensfield Road,
Cambridge CB2 lEW, UK
Department of Chemistry, The University of Edinburgh, West Mains Road,
Edinburgh EH9 355, UK
Joseph Chatt's interests in organometallic chemistry were wide-ranging, from

bonding theories to nitrogen fixation. While these areas may not seem of
immediate relevance to either carbonyl cluster chemistry or to electrospray mass
spectrometry (both of which play a major role in the work described herein), the
chemistry that is discussed broadly overlaps with Chatt's contributions in metal
hydride, metal phosphine and low-oxidation-state chemistry.
The following work describes our investigation into some unexpected chemis-
try resulting from our interest in unsaturated metal clusters.' We study such
systems primarily by mass spectrometry, and to assist these investigations we
developed the new data presentation technique of energy-dependent electros-
pray mass spectrometry (EDESI-MS). One of the first systems studied was that
of alkoxylated transition metal carbonyl clusters, which display loss of formalde-
hyde in their electrospray ionisation fragmentation process. We have been able
to correlate this behaviour with macroscopic chemical properties of these clus-
ters.
1 Alkoxylation of Carbonyl Ligands

A well-known reaction in transition metal carbonyl chemistry is nucleophilic
344 Formaldehyde Elimination from Methoxylated Transition Metal Carbonyl Clusters
attack of alkoxide ions, RO-, on the electropositive carbon atom of a carbonyl
ligand, affording an anionic alkoxycarbonyl species (Equation 1).
1-
M-CO + MeO- - M-C
/
OMe
\\o
(1 1
Alkoxylation occurs very rapidly and the product is thermodynamically fa-

The reaction between alkoxide ions and transition metal carbonyl
clusters, [M,(CO),], generates anionic species of general formula
[M,(COOR)(CO), - '1 -. A number have been isolated and crystallographically
characterised, including [Ir,(COOMe)(CO), s] - , 3 [Os,C(COOEt)H(CO),,] -
-:
and [OS,C(COOM~)I(CO),~] [Ir,(COOMe)(CO), '1 - and [Rh,(COOMe)-
(CO)' 51 - .,
2 Chemical Derivatisation
The alkoxylation reaction has been exploited in the in situ derivatisation of
neutral metal carbonyl complexes for analysis by electrospray ionisation mass
spectrometry (ESI-MS).7 ESI is a relatively new ionisation technique which
involves spraying a solution from a charged capillary into a strong electric field.
Tiny droplets are formed from which the solvent is evaporated by means of a
warm bath gas. Acquisition of charge by the target molecule usually takes place
by chemical ionisation, frequently addition of H from a protic solvent (typically
+
acetonitrile/water). However, neutral metal carbonyl compounds do not readily

undergo protonation as they are insufficiently basic. Derivatisation by alkoxide
ion was subsequently found to be a convenient method for chemically generating
[M + OR]- ions.' Charged organometallic species are readily analysed by
electrospray mass spectrometry,' and typically just a single envelope of peaks
corresponding to the parent is observed in the mass spectrum. In ESI-MS,
fragmentation is considerably reduced compared to more conventional ionisa-
tion techniques, such as electron impact. l o The alkoxide derivatisation method
works equally successfully for clusters, and, despite the presence of multiple
reaction sites, double alkoxylation giving [M + 2(0Me)12- ions has never been
observed by ESI-MS. Such a reaction is, however, not without precedent, as a
double alkoxylation product was recently isolated from the reaction between
NaOMe and [Ir6(CO)16] in methanol and the solid-state structure of the prod-
uct [Ir,(COOMe)2(C0),4]2 - was determined.' ' The two methoxycarbonyl
fragments are on adjacent metal vertices of the octahedral framework.
3 Energy-dependent Electrospray Mass Spectrometry

While fragmentation tends to be minimal for electrospray ionisation under
normal conditions, it can be increased very conveniently by changing the voltage
applied at the skimmer cones. Essentially, increasing the cone voltage causes
Paul J . Dyson et al. 345
collision-induced dissociation (CID) before the ions are directed into the mass
analyser. Analysis of the resulting fragmentation pattern can yield interesting
information on the compound in question, and we have been able to correlate
the information gained from the mass spectrometric studies with the compounds'
macroscopic chemical behaviour.
The conventional method of displaying fragmentation data from ESI-MS is to
stack a series of spectra gathered at different cone voltages.' Such an approach
is illustrated in Figure 1, which shows the negative-ion ESI mass spectra of
[Ir,(COOMe)(CO), ' 3 - 1 recorded at cone voltage settings of 25,75 and 150 V.
Each spectrum provides a snapshot of the ligand stripping process as a
function of increasing cone voltage, and presentation of all the possible data sets
in this fashion is clearly not practical. However, the entire fragmentation pattern
can be easily visualised using energy-dependent electrospray ionisation mass
spectrometry (EDESI-MS). This technique has recently been shown to be useful
for the analysis of fragmentation processes of cluster compounds, demonstrated
using [Rh,(COOMe)(CO), J - 2.l 3 A very large amount of data is generated in
such studies as a different spectrum is obtained at each increment. EDESI-MS
1135.0
I
1 1107.01
796.1
10h I
Figure 1 Negative-ion ESI mass spectra of [Ir,(COOMe)(CO), 1, showing the eflect

of the cone voltage setting on the fragmentation patterns; (a) 25 V,(b) 75 V;
(c) 150 V
100-
x-
200 =.
180-
i!u
160 -
11
Figure 2 The two-dimensional E D E S I - M S map generated from 201 negative-ion E S I - M S
spectra of [Ir,(COOMe)(CO), I] - 1 at cone voltage settings 0fO-200 V. The top
trace is a 1D spectrum generated by combining all 201 spectra together
involves plotting this huge amount of data (up to 201 spectra) in a two-dimen-
sional format, generating a map (with mass-to-charge ratio on the horizontal
axis and cone voltage on the vertical axis), the contours of which describe the
entire fragmentation pattern of the compound in question. An additional feature
is a spectrum generated by summing all the spectra used in the map; this
spectrum appears at the top of the map. Each cross peak in the EDESI map
represents a particular fragment ion, the most intense and/or long-lived of which
are generally regarded as having particular stability. For transition metal car-
bony1 clusters, the primary fragmentation route is via loss of the carbonyl ligands
as carbon monoxide.
Figure 2 shows the composite 1D/2D EDESI mass spectrum for 1. Due to the
timespan of the experiment, good signal-to-noise is obtained at the expense of
resolution.
The fragment peaks in the spectrum correspond to consecutive loss of CO
from the central Ir, core. A formaldehyde molecule, HCHO, is also lost. From
the EDESI-MS spectrum shown in Figure 2 it is not immediately clear where the
%-
or.. . . , . . . . , . . . . , . . .
910.1
I
1
Y
J
882.2 I 938.2
I
854.4
0 - W ' . .. +.
700 8h
Figure 3 Negative-ion ESI-MSIMS spectra of [Ir,(COOMe)(CO), ,]- 1, showing the

effect of the collision voltage setting on the fragmentation patterns; (a) 25 V,(b)
75 V;(c) 150 V
HCHO loss (m/z 30) channel occurs relative to the C O loss (m/z 28) channel, but
careful inspection reveals that the discontinuity probably occurs at m/z 969, i.e.
[Ir,(CO), + OMel- loses HCHO to form the [Ir,H(CO),]- at m/z 939. Un-
equivocal confirmation of formaldehyde loss is provided by acquiring a conven-
tional high-resolution mass spectrum at the appropriate cone voltage.
4 Energy-dependent Electrospray Tandem Mass

Spectrometry
Alternatively, the recent introduction of EDESI-MS/MS provides another useful
tool for analysing such systems.l4 Tandem mass spectrometry (MS/MS) allows
selection of a single ion using one mass analyser then introducing it to a collision
cell. Energetic collisions with an inert gas in this cell cause fragmentation of the
ion and a daughter ion spectrum is obtained. MS/MS techniques are especially
useful for the analysis of complex mixtures, but their application to molecules
"'1
0 1
750 8C0 850 900 950 1000 1050 1100 1150
m/z
Figure 4 The two-dimensional EDESI-MSIMS map generated from 201 negative-ion

daughter ion ESI-MSIMS spectra of [Ir,(COOMe)(CO), - 1 at collision
voltage settings of 0-200 V. The top trace is a 1 D spectrum generated by
combining all 201 spectra together
with complicated isotopomer envelopes is also useful, as instead of a broad,

near-Gaussian distribution of peaks for each ion, a single peak is produced
instead. This feature of MS/MS spectra is illustrated in Figure 3, which shows
negative-ion daughter ion ESI-MS/MS of [Ir,(COOMe)(CO), ,]-, recorded at
collision voltage settings of 25,75 and 150 V.
The principal difference between these spectra and those shown in Figure 1 is
the disappearance of the isotopomer envelopes; instead, a single peak is observed
for each ion. An EDESI-MS/MS map for [Ir,(COOMe)(CO), - can be gener-
ated by stacking all the spectra, collected at collision voltages of 0-200 V, in an
entirely analogous way to that in EDESI-MS (Figure 4).
Comparison between the two EDESI maps reveals the expected similarities,
but also some marked differences. In particular, the ability to fragment the
parent ion within the collision cell is markedly less than that achieved at the
skimmer cone. The EDESI map shows that the ion [Ir,H]- (in which all C O
ligands have been removed) makes its first appearance at a cone voltage of 132 V,
and by 175 V is the only ion present. In contrast, the most heavily fragmented ion
in the EDESI-MS/MS is [Ir,H(CO),]-, which only just appears at a high
collision voltage of 190 V. The same ion in the EDESI-MS map appears at 105 V
and disappears by 158 V. Despite this behaviour at high voltages, at low voltages
fragmentation is induced more readily in the collision cell, as a comparison of the
two maps at 20 V makes clear. In the EDESI-MS map, only the intact parent ion
[Ir,(COOMe)(CO), ,] - is present, whereas in the EDESI-MS/MS map, the
fragment ions [Ir,(COOMe)(CO),,] - ( n = 8-10) are already evident in signifi-
cant intensity. It should be noted that fragmentation in the collision cell can,
however, be increased by the simple expedient of using argon instead of dinitro-
gen as the collision gas.
Apart from the differences in fragmentation power, overall the EDESI-MS
and EDESI-MS/MS maps are qualitatively similar. Essentially the same pattern
of intensities for each daughter ion is observed, best represented by the summed
spectrum at the top of each map. This feature is not surprising given that the
mechanism for fragmentation is collision-induced dissociation by N, gas in both
cases. An advantage of the selection of a single ion is apparent in the EDESI-
MS/MS approach in that identification of the point at which HCHO loss versus
C O loss takes place is more straightforward.
5 Formaldehyde Elimination
We have studied a number of different anionic methoxycarbonyl clusters, and
found them all to undergo loss of HCHO at some point during their fragmenta-
tion processes. In some cases, we have correlated differences in fragmentation
patterns between the various clusters with their macroscopic chemical proper-
ties. Confirmation of peak assignments was carried out in some cases using
Fourier transform ion cyclotron resonance (FTICR).
The elimination of an aldehyde or ketone from a coordinated alkoxide is a well
known process in coordination chemistry. For example, treatment of metal
halide complexes with alcoholic base is a standard method for the preparation of
metal hydride complexes (Equation 2).' Labelling experiments have shown that
the a hydrogen is transformed into the hydride ligand.16
M-X + R,CHO- + M-H + R,CO + X - (2)
This process has also been observed in cluster chemistry. For example, the
cluster anion [Ru,IrH(OMe)(CO), ,]- eliminates HCHO under carbonyl
loss conditions (prolonged heating) to generate the cluster anion [Ru,IrH,-
(CO),,] - . I 7 Because fragmentation in the mass spectrometer also involves
carbonyl loss, it seems plausible that such a process might be simulated under
EDESI-MS conditions.
Methoxylation of the hexaruthenium carbide cluster [Ru,C(CO), 7] generates
the stable anionic cluster [Ru,C(COOMe)(CO), ,]- (3a), the negative-ion
EDESI mass spectrum of which is shown in Figure 5.18
As seen for 1 and 2, at the lowest cone voltages, the only peak observed is that
Figure 5 The negative-ion EDESI mass spectrum of [Ru,C(COOMe)(CO),,] - 3a
of the intact parent ion. Upon increasing the fragmentation energy, two C O
ligands are lost, and the ions [Ru6C(COOMe)(C0),]- (n = 14 or 15) appear in
the EDESI-MS map at very low intensity. The structure of these ions is not
obvious because the CO ligands may be lost from either the cluster shell or from
the methoxycarbonyl ligand. The third neutral molecule to be lost from the
cluster is HCHO rather than a C O ligand, to generate the hydride cluster
[RU,C(H)(CO),,] -. The remaining fifteen cross peaks correspond to the series
[Ru,C(H)(CO),] - (x = 0-14), and have roughly equal intensity, leading ulti-
mately to [Ru,C(H)] -. The closely related anion [Ru,C(COOEt)(C0),,] - (3b)
undergoes an analogous fragmentation sequence, except that CH,CHO is elim-
inated instead of HCHO. As expected, [Ru,C(COOPh)(CO),,] - (3c) does not
display similar behaviour, as the phenyl ring prevents formation of an exocyclic
C=O bond. The C O ligands are progressively stripped in the case of 342, with
complex fragmentation occurring at the highest cone voltages.
Compound 3a is quite stable and we have isolated (Ph,PNPPh,)[Ru,C-
(COOMe)(CO),,] and established its solid-state structure (see Figure 6 for the
structure of the anion);’* the structure of 2 is known., The phosphine-substituted
derivative of 3a, [Ru,C(COOMe)(CO), ,(PPh,)] - (4) was prepared by treatment
of [Ru,C(CO),,(PPh,)] with sodium methoxide. The EDESI mass spectra of 4 is
Figure 6 The molecular structure of the anion [Ru,C(COOMe)(CO),,] - 3a
shown in Figure 7.
The spectrum of 4 is very similar to that of 3a. The PPh, is clearly lost first, as
shown by the large space between cross peaks in the EDESI map, followed by a
single CO ligand and then rapid loss of HCHO. The remainder of the pattern
involves straightforward C O stripping, from [Ru,C(H)(CO), - down to the
[RU,C(H)] - core, as for 3a.
While caution must be applied to any direct comparisons between fragmenta-
tion patterns observed in the gas phase and chemical properties observed in
solution, in this case there is an obvious correlation. The early loss of formalde-
hyde from 3a (and 4) compared to 2 during the fragmentation process equates to
the differences in chemical reduction of the two clusters. Treatment of
[Ru,C(CO), 7] with methanolic KOH provides [RU,C(CO),,]~ - cleanly"
whereas reduction of [Rh,(CO), 6] requires stronger reducing agents such as
Na/Hg to produce the dianion. The hexaruthenium dianion [Ru,C(CO), ,I2-
is a widely-used precursor in cluster chemistry2' and the mechanism of its
formation presumably commences similarly to the reaction with NaOMe. Treat-
ment of [RU6C(COOMe)(C0),,] - with OH- quantitatively yields
[RU6C(C0)1,]2-.'2c Based on the EDESI data, we would also expect that
treating [RU,C(CO),,(PPh,)] with OH- should yield [Ru,C(CO),,(PP~J]~-,
and preliminary synthetic results show that this seems to be the case (though
some [Ru,C(C0)1,]2- is also formed).
It is reasonable to assume that the -COOMe group rearranges to form a
relatively strong multicentre bonding interaction, probably driven, in the first
instance, by the loss of a C O ligand (Scheme 1).Formation of an - 0 M e ligand is
likely to be the step prior to elimination of formaldehyde, and as already
mentioned, the cluster anion [Ru3Ir(H)(0Me)(CO),,] - is known to eliminate
352 Formaldehyde Elimination froin Methoxylated Transition Metal Carbonyl Clusters
1039 2
Figure 7 The negative-ion EDESI mass spectrum of[Ru,C(COOMe)(CO),5(PPh,)] - 4
HCHO under carbonyl loss conditions (prolonged heating) to generate the

cluster anion [Ru,IrH,(CO),,] -. The hydride ligand is likely to be abstracted
from the cluster by OH- (this step, of course, is not observed in the mass
spectrometer). The resulting cluster will be short of one CO ligand, but as two
have been lost, there will be plenty of C O present in solution for the unsaturated
cluster to pick up.
This HCHO elimination mechanism is different from one proposed earlier for
the reduction of [Ru,C(CO), 7], involving nucleophilic addition of OH- to a C O
ligand to form a T O O H intermediate, followed by expulsion of CO, and then
removal of H + by O H - to form [ R u , C ( C ~ ) , , ] ~ - . Further
~~ experiments are in
progress to confirm the mechanism.
6 Acknowledgements
We would like to thank the Royal Society for a University Research Fellowship
(P. J. D.). Thanks also to Paul Skelton for collecting the FTICR data.
H. 1-
1-
1-
1
@: -----;I+
+ HCHO
7 References
1 (a) G. Critchley, P. J. Dyson, B. F. G. Johnson, J. S . McIndoe, R. K. O’Reilly and P. R.
R. Langridge-Smith, Organometallics, 1999,18,4090;(b) P. J. Dyson, B. F. G. Johnson,
J. S . McIndoe and P. R. R. Langridge-Smith, Inorg. Chem., 2000,39,2430.
2 (a) D. C. Gross and P. C. Ford, J . Am. Chem. Soc., 1986,108,6100; (b) R. J. Trautman,
D. C. Gross and P. C. Ford, J . Am. Chem. Soc., 1985,107,2355; (c) D. C. Gross and P.
C. Ford, J . Am. Chem. Soc., 1985, 107, 585; (d) D. C. Gross and P. C. Ford, Inorg.
Chem., 1982,21,1702.
3 L. Garlaschelli, M. C. Malatesta, S. Martinengo, F. Demartin, M. Manassero and M.
Sansoni, J . Chem. Soc., Dalton Trans., 1986,777.
4 B. F. G. Johnson, J. Lewis, W. J. H. Nelson, J. N. Nicholls, M. D. Vargas, D. Braga, K.
Henrick and M. McPartlin, J . Chem. Soc., Dalton Trans., 1984, 1809.
5 L. Garlaschelli, S. Martinengo, P. Chini, F. Canziani and R. Bau, J . Organomet. Chem.,
1981,213,379.
6 G. Ciani, A. Sironi, P. Chini and S. Martinengo, J . Organomet. Chem., 1981,213, C37.
7 W. Henderson, J. S . McIndoe, B. K. Nicholson and P. J. Dyson, Chem. Commun., 1996,
1183.
8 W. Henderson, J. S. McIndoe, B. K. Nicholson and P. J. Dyson. J . Chem. Soc., Dalton
Trans., 1998,519.
9 W. Henderson, B. K. Nicholson and L. J. McCaffrey, Polyhedron, 1998,17,4291.
10 B. F. G. Johnson and J. S. McIndoe, Coord. Chem. Reu., 2000,200,901.
11 R. Della Pergola, L. Garlaschelli, S. Martinengo, M. Manassero and M. Sansoni, J .
Organomet. Chem., 2000,593-594,63.
12 (a) L. A. P. Kane-Maguire, R. Kanitz and M. M. Sheil, J . Organomet. Chem., 1995,486,
243;(b)A. van den Bergen, R. Colton, M. Percy and B. 0.West, Inorg. Chem. 1993,32,
3408.
13 P. J. Dyson, B. F. G. Johnson, J. S. McIndoe and P. R. R. Langridge-Smith, Rapid
Commun. Mass Spec., 2000,14,311.
14 P. J. Dyson, A. K. Hearley, B. F. G. Johnson, J. S. McIndoe, P. R. R. Langridge-Smith
and C. Whyte, Rapid Commun. Mass Spec., 2001,15, 895.
15 J. P. Collman, L. S. Hegedus, J. R. Norton and R. G. Finke, Principles and Applications
of Organotransition Metal Chemistry, University Science Books, Mill Valley CA, 1987.
16 L. Vaska and J. W. DiLuzio, J . Am. Chem. SOC.,1962,84,4989.
17 S. Haak, A. Neels, H. Stoeckli-Evans, G. Suss-Fink and C. M. Thomas, Chem.
Commun., 1999,1959.
18 P. J. Dyson, N. Feeder, B. F. G. Johnson, J. S. McTndoe and P. R. R. Langridge-Smith,
J . Chem. SOC., Dalton Trans., 2000,1813.
19 (a) D. Braga, F. Grepioni, P. J. Dyson, B. F. G. Johnson, P. Frediani, M. Bianchi and
F. Piacenti, J . Chem. SOC.,Dalton Trans., 1992,2565; (b) S. R. Drake, B. F. G. Johnson
and J. Lewis, J . Chem. SOC., Dalton Trans., 1989,243; (c) B. F. G. Johnson, J. Lewis, S.
W. Sankey, K. Wong, M. McPartlin and W. H. J. Nelson, J . Organornet. Chem., 1980,
191, C3.
20 P. J. Dyson, Adv. Organomet. Chem., 1998,43,43.
21 R. J. H. Clark, P. J. Dyson, D. G. Humphrey and B. F. G. Johnson, Polyhedron, 1998,
17,2985.
Exploring New Structures BasedA
on Chatt's (Pt2S2}Corefor I
Nucleation of Intermetallic
Growth
ZHAOHUI LI," S.-W. AUDI FONG," JEREMY S. L. YEO," W.
HENDERSON,bK. F. MOK" AND T. S. ANDY HOR"
aDepartment of Chemistry, National University of Singapore, 3 Science Drive
3, Singapore 117543
Department of Chemistry, University of Waikato, Private Bag 3 105,
Hamilton, New Zealand
1 Historical Perspective
The development of the chemistry of platinum sulfide complexes can be traced to
the early work of Chatt.1.2 Among his complexes in this field, those having a
significant impact on modern cluster chemistry are probably [{ Pt(PR,),),(p-
S),] (1). These compounds were synthesised independently by Chatt
(PR, = PPhMe,)2 and Ugo (PR, = PPh,), in the early 1970s. The triphenyl-
phosphine complex can be easily prepared in good yields but it is practically
insoluble in all common organic solvents, except CHC1, and CH2Cl, in which it
decomposes readily. ,2+6
The methylated derivative [Pt2(PR3)4(p-S)(p-SMe)]PF,3was among the ear-
liest crystallographically-characterised derivatives whereas the protonated ana-
logue [Pt2(PR3)4(p-s)(p-sH)]PF6 (Figure l) has been isolated and characterised
only very r e ~ e n t l y . ~
Although the related trinuclear derivative [Pt3(PR3)6(p3-s),]2+was syn-
thesised by Chatt and Mingos in their earlier work,2 its synthetic relationship
with 1 was not established until the work particularly of Mingos who explored a
series of metallation reactions using 1, R = Ph as a metalloligand.s~'oThis has
led to an exponential growth of medium- to high-nuclearity aggregates over the
past two To date, the metalloligand 1, R = Ph probably accounts
356 Exploring New Structures Based on Chatt's {Pt2S2}Core
lH+
SlA
PlA
Figure 1 Structure of [Pt,(PPh,),(~-S)(~-SH)]PF,

(For clarity, the P F , - counter-anion
is omitted)
not just for the largest number of heterometal aggregates, but also the largest
permutation of different heterometals.' Its ligating abilities towards p-, d-
andf-block, hard and soft metals, as well as late and early transition metals have
been demonstrated. Two significant predecessors are the pentanuclear [Pt,Pd-
(PPh,)8(p,-S),](BF4),9 and hexanuclear [Pt,Ag,(PPh3)8(p3-S)4](BF4)2.'o They
proved conclusively that two (Pt,S,) units can be brought together by 'naked'
metals, and that in doing so, they can support M-M bonding, as shown in the
latter. This finding prompted a vigorous exploration of other metal aggregates
with different nuclearities, with or without M-M bonds. This is summarised in
our recent review.16
Over the years, there have been three major milestones that mark the develop-
ment of this work (Figure 2):
(i) Capping of a single metal atom on the (Pt,S,} core - This is best exemplified
in the isolation of [Pt2T1(PPh,),(p,-S)2](PF6).'3It highlights the ability of this
metalloligand to support and stabilise coordinatively and electronically un-
saturated metal centres and, in some cases, to promote heterometal M-M
interactions.
(ii) Aggregate to cluster conversion - A range of aggregates has been made
using this methodology but the majority of them do not have M-M bonds.
However, a facile reduction process of [Pt,M(PPh,),(p,-S),]X (M = Cu or Ag)
to [Pt,MX(PPh,),(CO)(p,-S)] (X = halide) by C O raises the possibility of con-
verting these aggregates to a new class of cluster compound^.'^ It is remarkable
that these reduction and desulfurisation processes can be achieved in a single
step by such a mild reductant with no intractable by-products.
(iii) Tandem migration of a metal fragment across the S-S bridge - This is best
illustrated in [Pt2Au(PPh,)4(p3-S)2] which displays an interesting dynamic
+-
process with the metal moving across the sulfur sites with a low energy barrier
between an intermediate and two transition This metal migration
complements a ligand migration process that we reported earlier in a similar M,
core.'
Zhaohui Li et al. 357
Tllll
co
Figure 2 Major developments in the field of [Pt2(PPh3)4(pS)2]

chemistry
To date, we have extensively studied the metallation chemistry of the { Pt,S,)

core. Our immediate aim was to explore the materials chemistry of these aggre-
gates. The use of these aggregates as precursors to electronic and magnetic
materials is an interesting prospect. This system is especially suitable for such
development because of the almost infinite possibilities arising from combining
different metals and ligands. In doing so, we are able to tune the chemical and
physical properties to suit desired applications.
2 Current Perspectives
2.1 Stereochemical Changes
The (Pt2S2}plane in the parent complexes is usually flat whereas in the metal-
lated derivatives it is commonly hinged. There are however several notable
exceptions, one of which is [Pt2Hg2(N03)2(PPh3)4(p3-S)2]2+ which has a flat
{ Pt,S2) core (Figure 3). In this stereoconfiguration, metal addition takes place
both above and below the {Pt2S2}plane, thus resembling the growth in 'one-
dimensional' coordination polymers. In a folded configuration, metallation can
occur at the apical positions of both sulfur sites, with or without M-M bonding.
With a careful choice of metal and supporting ligands, one can envisage the
development of metallocycles with different stereochemistry and cavity sizes. An
example is found in [Pt2Hg,(PPh,),(p,-S)2(p-Cl)2]2+.
358 Exploring New Structures Based on Chatt’s (Pt2S2}Core
M6
0171
Figure 3 DifSerent stereochemistry ofthe ( P t 2 S 2 }core
2.2 Different Degrees of Site-anchoring

Although the { Pt,S,’, core undergoes simple single-site derivatisation easily, e.g.
to give [Pt2(PPh3),(p-S)(p-SMe)] and [Pt2Au(PPh3),(p3-S),1 +, there are
+
examples which show that, under certain conditions, both sulfurs can be metal-
lated. When this happens, they can support a single heterometal atom, e.g.
[Pt,Hg(C,H,)(PPh3),(p3-S),1 or two metals anchoring in close proximity e.g.
+
[Pt,Hg2(C2H,)2(PPh3)4(p3-S)2]2+ (Figure 4). This versatility gives an additional

degree of freedom to materials-growth based on the {Pt2S2)core.
2.3 Bifacial Addition and Dissociation

In a chelating mode, the metalloligand can exhibit different modes of adduct
formation. Addition of a binary compound gives rise to a neutral addition
compound, e.g. [Pt2HgC12(PPh3),(p3-S)2](Figure 5). Halide dissociation would
give rise to ionic adduct, e.g. [Pt2Hg(PPh3),(p3-S)2]2+.A 2: 1 addition would
lead to a bifacial attack and formation of a bis(che1ate) configuration, e.g.
[Pt,Hg(PPh,)s(p.,-s),l 2 + .
Double-site Anchor SingIe-site Anchor
Figure 4 Diferent degrees of site-anchoring on the ( P t 2 S 2 )core
2.4 Formation of Bimetal Complexes from Early and Late

Transition Metals
Most of the documented examples of aggregate formation with the {Pt2S2)core
are addition reactions involving metals from the middle and late transition
series. These are generally Lewis acidic systems that complement the basic
character of the metalloligand. In most cases, the soft ligating behaviour of the
sulfur donors also matches the softer character of the late metals. Our recent
efforts have been directed to the synthesis of late/early bimetal systems, a typical
example of which is the synthesis of [Pt2VO(OCH3)2(PPh3)4(p3-S)2]from +-
NH4V03. The lack of acidity of the vanadate substrate is remedied by a

methanolysis process (Figure 6).
2.5 Introduction of an Unsaturated Functionality

The { Pt2S2}core is formally coordinatively and electronically saturated since
the 16-electron d8 centre is effectively stable to addition reactions. However,
upon metallation, it is possible to introduce unsaturation to the system when the
metal carries along a labile ligand such as cycloocta-1,5-diene (cod). This is
exemplified in the synthesis of [Pt3(cod)(PPh3)4(p3-S)2]2+and [Pt,Pb-
(N03)2(PPh3)4(p3-S)2]. This transformation from a Lewis basic to a Lewis acidic
complex is accompanied by alteration of its chemical reactivity and coordination
360 Exploring New Structures Based on Chatt’s (Pt2S2)Core
T
Neutral Mono-Cationic
Direct Adduct Formation Addition with
Dissociation
Dicationic
Bifacial Attack
Figure 5 Diferent modes of adduct formation exhibited by [Pt2(PPh,),(p-S),]
behaviour. When a ‘naked’ metal such as T1’ is introduced, as in

[Pt2T1(PPh,)4(p3-S)2]+,the unsaturated behaviour is significantly enhanced.
2.6 Enhancing Electrochemical Activity

The electrochemical activity of some heterometallic aggregates has been re-
ported.” When different metals are in close proximity within a cluster core, and
especially when sulfide is a connecting ligand, electronic communication be-
comes a clear possibility. When each metal carries an electroactive ligand such as
the ferrocenyl group, in principle the complex becomes a multi-centred redox
system. The level of electrochemical activity, as well as M-M and M-L com-
munication, can be tuned easily by fitting the appropriate ligands and metals to
this core. This represents a relatively simple yet powerful approach to synthesise
electroactive aggregates, e.g. [Pt2Hg(Fc)(PPh3)4(p3-S)2] (Fc = (C,H,)-
+
(C5H4)Fe)and CPt2T1 (dpPf)&-%I+ [dppf = Fe(CsH4PPh2)2]-
2.7 Electrospray Mass Spectrometry (ESMS) Analysis

In view of the powerful nature of the {Pt,S2) core to couple to virtually all metals
in the Periodic Table, and because different metals can carry a range of function-
alities, there is an infinite number of possible permutations that one can attempt.
It is therefore important to develop a combinatorial-like tool which can be used
Zhaohui Li et al. 36 1
t-~N H ~ +
2MeOH H+
OMe
Lewis Basic - IH4OH Lewis Acidic
Figure 6 Formation of [Pt2VO(OCH,),(PPh3),(p3-S),1+,an example of the introduction

of an early transition metal to the heterometal system
to screen for positive reactions and to identify potentially stable and isolable
products. When such information is fed into the design and synthesis process,
one can maximise the productivity of synthetic experimentation. We have devel-
oped ESMS as such a combinatorial tool. It permits ‘in situ’ observation of
possible products when the substrates are mixed in solution. A good example is
illustrated in our study of the mercury(@ systems in which a number of Hg/Pt
aggregates were detected and subsequently synthesi~ed.~ (Figures 7, 8 and 9)
When this technique is applied to the V/Pt system, we observed metal-assisted
ligand transformation in these aggregates at different cone voltages. In addition,
the methanolysis process shown in Figure 6 for a reaction involving 1 and a
vanadium(v) system can be followed conveniently by ESMS. When this tech-
nique is be applied to a molybdenum(v1) system, using [MOO,]’- as a substrate
in a reaction with 1, R = Ph, one could also detect cyclometallation activity.
What started as a project involving simple Lewis acid/base addition reactions
has emerged into one that traverses materials synthesis to the development of a
technique that could change our approach to inorganic synthesis in general. The
contributions of Chatt in the early 1970s, Mingos in the 1980s and other
researchers in the past two decade^'-^'^-^^ made this evolution possible. Our
work in the past two decades proved that the (Pt,S,) system is probably the
most powerful, convenient and general synthetic precursor to heterometal aggre-
362 Exploring New Structures Based on Chatt's {Pt2S2}Core
p\ /p
/ Pt\
-12+ I
R-I
r r
rnlz 900 mlz 1739 rnlz 1603
Fluorescein
mercuric(11)
acetate
a b
rnlz 1117
a: mlz 1733 a: mlz 1781 (3a) a: mlz 1838 a rnlz 1868

b: mlz 981 b rnlz 1029 b' miz 1086 b: rnlz 1117
Figure 7 ESMS observed species in mixtures containing [Pt2(PPh3),(p-S),] and various

mercury(r1) species
gates and clusters. The materials, chemical and dynamic properties of a large
portion of these heterometal complexes remain largely unexplored. This is a
subject of our future investigations.
3 Acknowledgements
T. S. A. H. thanks D. M. P. Mingos (who started this research under the guidance
of J. Chatt) for introducing him to this fruitful area of research. We are grateful
for funding support from the National University of Singapore and the Univer-
sity of Waikato and to J. J. Vittal for X-ray structural characterisation of some
compounds. Technical assistance in the preparation of this manuscript from
Tengku Tjendekrawan is acknowledged.
m/z 1051 m/z 1603 m/z 1783 m/z 2182
m/z 983 m/z 1250 m/z 1603 m/z 1739
Figure 8 Major species observed in the mixtures of[Pt,(PPh3),(,u-S),] and

mercury-phosphine complexes under ESMS conditions
+ +
m/z 1045 m/z 1796 m/z 2124
&
HgCl
HgCl
@ -Fe- -$l+ Pt Fe
-Hg-s’ sH
‘-gJ&
+ Fe
\ /
P.7Pt $P dPt\,
P
m/z 1889 m/z 1137
Figure 9 Major species observed in the mixtures of [Pt2(PPhJ4(,u-S),]and

mercury-ferrocenyl complexes under ESMS conditions
364 Exploring New Structures Based on Chatt’s { Pt,S,} Core
4 References
1 J. Chatt and F. A. Hart, J . Chem. Soc. A , 1960,2807.
2 J. Chatt and D. M. P. Mingos, J . Chem. Soc. A , 1970,1243.
3 R. Ugo, G. La Monica, S. Cenini, A. Segre and F. Conti, J . Chem. Soc. A , 1971,522.
4 R. R. Gukathasan, R. H. Morris and A. Walker, Can. J . Chem., 1983,61,2490.
5 (a) C. E. Briant, C. J. Gardner, T. S. A. Hor, N. D. Howells and D. M. P. Mingos, J .
Chem. Soc., Dalton Trans., 1984,2645; (b) T. S. A. Hor and A. L. C. Tan, Inorg. Chim.
Acta, 1988,142, 173.
6 (a) A. Shaver, R. D. Lai, P. H. Bird and W. Wickramasinghe, Can. J . Chem., 1985,63,
2555; (b) V. W.-W. Yam, P. K.-Y. Yeung and K.-K. Cheung, J . Chem. Soc., Chem.
Commun., 1995,267.
7 (a) S.-W. A. Fong, J. J. Vittal, W. Henderson, T. S. A. Hor, A. G. Oliver and C. E. F.
Rickard, Chem. Comm., 2001,421; (b) S.-W. A. Fong, W. T. Yap, J. J. Vittal, T. S. A.
Hor, W. Henderson, A. G. Oliver and C. E. F. Rickard, J . Chem. Soc., Dalton Trans.,
2001,1986.
8 (a) C. E. Briant, M. J. Calhorda, T. S. A. Hor, N. D. Howells and D. M. P. Mingos, J .
Chem. Soc., Dalton Trans., 1983,1325; (b) C. E. Briant, D. I. Gilmour, M. A. Luke and
D. M. P. Mingos, J . Chem. Soc., Dalton Trans., 1985,851; (c) D. I. Gilmour, M. A. Luke
and D. M. P. Mingos, J . Chem. Soc., Dalton Trans., 1987, 335; (d) W. Bos, J. J. Bow,
P. P. J. Schlebos, P. Hageman, W. P. Bosman, J. M. M. Smits, J. A. C. van Wietmar-
schen and P. T. Beurskens, Inorg. Chim. Acta, 1986,119,141.
9 C. E. Briant, T. S. A. Hor, N. D. Howells and D. M. P. Mingos, J . Chem. Soc., Chem.
Commun., 1983,1118.
10 C. E. Briant, T. S. A. Hor, N. D. Howells and D. M. P. Mingos, J . Organomet. Chem.,
1983,256, C15.
11 (a) M. Capdevila, Y. Carrasco, W. Clegg, R. A. Coxall, P. Gonzalez-Duarte, A. Lledos,
J. Sola and G. Ujaque, Chem. Commun., 1998,597; (b) H. Liu, A. L. Tan, Y. Xu, K. F.
Mok and T. S. A. Hor, Polyhedron, 1997,16,377.
12 (a) H. Liu, A. L. Tan, C. R. Cheng, K. F. Mok and T. S. A. Hor, Inorg. Chem., 1997,36,
2916; (b) V. W.-W. Yam, P. K.-Y. Yeung and K.-K. Cheung, Angew. Chem. Int. Ed.
Engl., 1996,35, 739.
13 M. Zhou, Y. Xu, L.-L. Koh, K. F. Mok, P.-H. Leung and T. S. A. Hor, Inorg. Chem.,
1993,32,1875.
14 H. Liu, A. L. Tan, K. F. Mok, T. C. W. Mak, A. S. Bastanov, J. A. K. Howard and T. S.
A. Hor, J . Am. Chem. Soc., 1997,119,11006.
15 (a) B. H. Aw, K. K. Looh, H. S. 0. Chan, K. L. Tan and T. S. A. Hor, J . Chem. Soc.,
Dalton Trans., 1994, 3177; (b) M. S. Zhou, Y. Xu, C.-F. Lam, L.-L. Koh, K. F. Mok,
P.-H. Leung and T. S. A. Hor, Inorg. Chem., 1993,32,4660; (c) M. Zhou, Y. Xu, C.-F.
Lam, P.-H. Leung, L. -L. Koh, K. F. Mok and T. S. A. Hor, Inorg. Chem., 1994,33,
1572; (d) M. S. Zhou, A. L. Tan, Y. Xu, C.-F. Lam, P.-H. Leung, K. F. Mok, L.-L. Koh
and T. S. A. Hor, Polyhedron, 1997,16,2381; (e) M. Zhou, C.-F. Lam, K. F. Mok, P.-H.
Leung and T. S. A. Hor, J . Organomet. Chem., 1994,476, C32; (f) M. S. Zhou, Y. Xu,
A.-M. Tan, P.-H. Leung, K. F. Mok, L.-L. Koh and T. S. A. Hor, Inorg. Chem., 1995,
34,6425; ( g ) A. L. Tan, M. L. Chiew and T. S. A. Hor, J . Mol. Struct. (THEOCHEM),
1997,393,189; (h) H. Liu, A. L. Tan, K. F. Mok and T. S. A. Hor, J . Chem. Soc., Dalton
Trans., 1996,4023; (i) Z. Li, H. Liu, K. F. Mok, A. S. Batsanov, J. A. K. Howard and T.
S. A. Hor, J . Organomet. Chem., 1999,575,223.
16 S.-W. A. Fong and T. S. A. Hor, J . Chem. Soc., Dalton Trans., 1999,639.
17 Z . Li, Z.-H. Loh, K. F. Mok and T. S. A. Hor, Inorg. Chem., 2000,39,5299.
18 J. S. L. Yeo, G. Li, W.-H. Yip, W. Henderson, T. C. W. Mak and T. S. A. Hor, J . Chem.
Soc., Dalton Trans., 1999,435.
19 Z . Li, X. Xu, S. B. Khoo, K. F. Mok and T. S. A. Hor, J . Chem. Sue., Dalton Trans.,
2000,2901.
A Rational Design of
Heteropolynuclear Squarate
Complexes
F. DUMITRU," D. BERGER," N. STANICA,' I. CIOCOIU"
AND C. GURAN"
aDepartment of Inorganic Chemistry, 'Politehnica' University Bucharest,
Polizu 1,78126-Bucharest, Romania
Romanian Academy, Institute of Physical Chemistry, Bucharest, Romania
Faculty of Chemistry, A. I. Cuza University, Copou 11, Iasi, Romania
1 Introduction
The synthesis of the first divalent metal complexes of squarate ion (c404),- by
West and Niu' in 1963, has constituted the starting point for many researches of
the coordination chemistry of squarate dianion.
In their pioneering work, West and Niu' assumed that the squarate complexes
of divalent cations Mg2+,Ca2+,Mn2+,Fe2+,Co2+,Ni2+,Cu2+ and Zn2+ form
an isostructural series (Figure 1). They suggested a bis-chelating coordination
mode for the squarate dianion in all the squarate complexes, but their proposed
structure does not correspond to the actual coordination mode of squarate.
In the light of the structures of squarate-metal complexes reported so far it is
clear that chelation by this ligand is limited to some alkaline- and rare-earth
metal cation^.^-^ This behaviour is explained by the large bite parameter of
squarate dianion.
X. Solans and co-workers,' attempting to answer the question: 'does squarate
act in a bidentate manner?, calculated the value of the bite parameter
(b = d2/d, = d(0-O)/d(M-0), Figure 2) for the squarate ligand in a series of
copper complexes of formula [Cu(C40,)L], where L represents an N,N-biden-
tate ligand.
They estimated the bite parameter for a bidentate squarate ligand to be 1.70 A
and d, to be 2.47 A (using the value of the bite parameter of the oxalate and the
d, bite for squarate). From this they concluded that the squarate should act as a
F . Dumitru et al. 367
Figure 1 Structurefor divalent metal-squarate complexes ( M = M g , Ca, M n , Fe, Co, Ni,

Cu, Zn)
Figure 2 The bite parameter, d,/d,, of the squarate ion
bidentate ligand in either an asymmetric manner or in a symmetric one with an

average copper-oxygen distance of about 2.47 A.This value is outside the
normal range for copper-oxygen bond lengths.
Energy calculations have complemented these stereochemical considerations
and showed that in the case of bidentate squarate a minimum of energy is
obtained if the configuration around the metal ion is close to tetrahedral, a
situation that would be unusual for copper(I1).
These studies have helped to elucidate the coordination modes of squarate by
establishing the following facts:
(i) when the oxygen atom is replaced by a larger atom, such as sulfur, the
resulting ligands can be bidentate,g.'o
(ii) when metal ions with larger ionic radii (alkaline-earth cations5p7 and
cerium(rrr)2)are used, q4 chelation can be achieved. The average d, values in
cerium(m) complexes are 2.69 and 2.62 close to the predicted one (2.47 A)
for a chelating squarate ligand.
As far as squarate coordination chemistry with 3d ions is concerned, a
wide variety of modes has been found: monodentate,"" p-172-bis-
(monodentate),8.' '-'p-1,3-bi~(monodentate)~~'~~~ and tetrakis(monoden-
tate).2426
The structures of transition-metal squarate complexes consist of one-dimen-
sional metal squarate chains interlinked by hydrogen bonding. These complexes
could be possible precursors to low dimensional polymeric electrical conductors
368 A Rational Design of Heteropolynuclear Squarate Complexes
or molecular magnets, because the linear metal squarate chains would appear to
serve as a pathway for electron conduction or magnetic superexchange. This
ability arises from the structural features of squarate: a set of four oxygen donors,
planar stereochemistry, and n-electron delocalisation.
To control the polymerisation process and make possible the isolation of
species of desired nuclearity, polydentate ligands are used as blocking groups.
Most of them are polydentate ligands with delocalised .n systems (Table 1).These
Table 1 Some typical blocking ligands
Blocking ligand Complex compounds References
bipym
8
27
phen
/CHicHTNH2 30
N-CH$HrNH, 29
\
CH2-CH,NH,
tren
1
2-
19
LN 1
WN-
salen
F . Durnitru et al. 369
features allow the tailoring of the nuclearity of desired polynuclear compounds
and the tuning of the magnetic properties of such squarate-bridged complexes.
Starting from these considerations, we have chosen as blocking ligand [6-(4-
chlorophenyl)pyridazin-3-y1] hydrazine, L, and synthesised the building blocks
[Fe(C104)L(H20),](C104)21 and [Fe(C104)L2(H20)](C104)22. Then we as-
sembled the polynuclear systems by reaction of the building blocks 1 or 2 with
dipotassium squarate, resulting in the preparation of [Fe(C404)-
(c104)2 2 (H2 0 ) 4 l(c104)2 3, CFe 2 (c404(clo4)
) 2 L41(clo4)2 4, CFeCr(C404)-
(c104) 2L2-(H20)4](c104) 2 5 and CFeCr(C404)(C104) 2 L41 (C1O4)2 6.
2 Experimental
2.1 Materials
[6-(4-Chlorophenyl)pyridazin-3-y1]hydrazine was obtained following the litera-
ture p r o c e d ~ r el . Iron(m)
~ perchlorate hydrate, chromium(II1)perchlorate hexa-
hydrate and squaric acid were purchased from commercial sources and used
without any further purification. A dipotassium squarate solution was prepared
by adding the required quantity of solid potassium hydroxide to an aqueous
solution of squaric acid. The metal content was determined by atomic absorp-
tion spectrometry. The molar conductivities were measured in 10-3M CH,CN
solutions.
2.2 Compound Preparations

[Fe(ClO,)L( H 0)3] (CIO,), 1 was obtained in quantitative yield from refluxing
methanol solutions of [6-(4-chlorophenyl)pyridazin-3-yl]hydrazine,L, (1 mmol,
0.221 g) and iron perchlorate (1 mmol, 0.354 g). The resulting black powder was
collected by vacuum filtration, washed with water, methanol and diethyl ether
and stored over P,O,. (1, C,oH,,C14FeN40,5 Calcd. C 19.1, H 2.4, N 8.9, Fe 8.9;
found: C 19.0, H 2.45, N 9.0, Fe 8.75%)
[Fe(Cl0,)L2(H2O)](C1O4),2 was obtained in quantitative yield using a pro-
cedure identical to that used for 1 except that 2 mmol, 0.442 g L were used. (2,
C2,H2,C1,FeN80,3 Calcd. C 29.5, H 2.46, N 13.8, Fe 6.88; found: C 29.7, H 2.5,
N 13.4, Fe 6.65%)
[Fe2(C404) (C104)2L2(H20141(C104) 2 and [Fe2(C404)(C104)2L41(clo4) 2
were obtained in quantitative yield from refluxing an aqueous solution of
dipotassium squarate (1 mmol, 0.190 g) with methanolic solutions of 1 (2 mmol,
1.258 g) or 2 (2 mmol, 1.627 g). The solid was collected by vacuum filtration,
washed with water, methanol and diethyl ether and stored over P,O,. (3,
C24H26C16Fe2N8024 Calcd. C 25.4, H 2.29, N 9.87, Fe 9.87; found: C 25.7, H
2.35, N 10.1, Fe 9.76%; 4,C,,H3,Cl8Fe2N,,O2, Calcd.: C 35.1, H 2.39, N 14.89,
Fe 7.45; found: C 35.3, H 2.45, N 15.1, Fe 7.33%)
The heterodinuclear complexes [FeCr(C,04)(C104),L2(H20)4](C104)2 5 and
[FeCr(C104)2(C40,)L4](C104)26 were obtained in a similar manner, starting
from chromium analogues of 1 and 2.
370 A Rutional Design of Heteropolynucleur Squarate Complexes
2.3 Physical Measurements

IR spectra were recorded with a Nicolet 2DXFT-IR spectrophotometer as KBr
pellets in the 4000-500 cm-' region. The UV-visible and reflectance were run on
a VSU-2G spectrometer using MgO as the reference for the reflectance spectra.
Molar conductances were measured at room temperature on a Radelkis KFT
conductivity meter. Metal ions were determined on a Pye-Unicam atomic ab-
sorption spectrophotometer. Elemental analyses were done by combustion with
a Carlo Erba instrument CHNS Elemental Analyser Model 1106. Magnetic
measurements were carried out at room temperature with a Faraday-type mag-
netometer.
2.4 Results and Discussion

Elemental chemical analyses, molar conductivity data, IR and electronic spectra
and magnetic susceptibility values lead us to propose the structures in Figure 3.
2.5 IR Spectra
The most characteristic bands, useful in a diagnostic sense, are collected in Table
2.
The v ~ band
- ~ in 1 and 2 is shifted to higher values than in the free bases,
pointing to the involvement of the unsubstituted hydrazine N-atom in the
coordination process, as occurs when hydrazine acts as an unidentate ligand.32
In addition, calculation of charge density of L, using HyperChem 4.0, supports
the conclusions from the IR data, showing that for such unsymmetrically sub-
2+
1 2 + OCIO,
0
H,Oq
OH0
0 0
OCIO,
OCIO,
1:2 type electrolyte
Figure 3 Proposed structuresfor [FeCr(C404)(C104)2L2(H20)4](C104)2

5 and
CFeC~(C~O4)2(C,O4)~41(~~04)2 6
F . Durnitru et al. 371
Table 2 Characteristic bands in IR spectra of complexes 1 4 (v, ern-', KBr)
Compound 6NH 'CN 'OH VCIO V~~ vco

1610 1460 - - 827 __
1 1628 1416 3404 1083 833 -
625
2 1666 1452 3428 1093 843 -
625
3 1620 1430 3390 1090 855 1485
625 1725
4 1659 1450 3410 1170 850 1495
625 1710
5 1615 1455 3270 1135 850 1480
649 1690
6 1615 1430 3260 1130 852 1480
649 1690
stituted hydrazines the nitrogen atoms involved in coordination are the ones
located in the 1 and 3 positions (Figure 4).
We have also observed in the IR spectra of the squarate-containing complexes
3-6 the C-0 stretching bands specific to p-l,3-bis(monodentate) coordination of
squaric acid (ca. 1480 cm-' for C-0 and 1710 cm-l for localised double-bonded
C=O).
2.6 Electronic Spectra and Magnetic Measurements

For the mono- and di-nuclear iron(m) complexes 1-4 the transition bands
appear in the range of 373-503 nm, characteristic of Fe3+ in a distorted octahed-
ral geometry (Figure 5a). For heterodiriuclear complexes 5 and 6 the specific
Fe3+ transitions are partially superposed upon those of Cr3+.The position and
shape of these bands confirm octahedral coordination at both Fe3+ and Cr3+
metal ions (Figure 5b).
The observed transitions were assigned on the basis of the magnetic moments
CI I
+ -0.009
0.022
Figure 4 Optimised structure and charge density for [6-(4-chlorophenyl)-

pyridazin-3-yll hydrazine, L
372 A Rational Design of Heteropolynuclear Squarate Complexes
1 - I
25 -I - ____- I_ _ "
I
/
/
*
-
!
'
204 I - 9 " ' c
300 500 700 900
'TZg+ 4T,g(Fe3+) 4T2g(Fe3+) 4A2g+ 4T2g 'Azg-+ 4Tig 'TZg+ 4T2g
(Cr 3+) (Cr 3+) (Fe 3+)
503 nm 373 nm 585 nm 469 nm 368 nm
Figure 5 Electronic spectra of(a) [Fe2(C404)(C104)2L,(H20)4](CI0,),, 3;

(b) ~ F e C r ~ C ~ O ~ ~ 5 ~C~O~~~~~~H,O
Table 3 Magnetic moments f o r the complexes 1-6
Compound &ff' BM
1.42
2.78
4.65
4.73
5.33
4.06
(Table 3) which indicate that the iron compounds are low-spin complexes.
This supports again the proposed structures where iron is coordinated to
perchlorate, a strong ligand forcing spin pairing.
The magnetic moments of the compounds which contain both chromium(Ir1)
and iron(1rr) compounds are slightly lower than expected for low-spin and
high-spin complexes of these metals (Table 3) which suggests a weak interaction
between the two paramagnetic ions (Fe3+and Cr3+).
3 Conclusions
We have reported the synthesis and the characterisation of a new series of the
polynuclear complex compounds of iron(m) and chromium(rr1)where the bridg-
ing ligand is squarate dianion, coordinated in a p-1,3-bis(monodentate) manner.
Further studies (EPR and Mossbauer spectroscopies) will complete the struc-
tural characterisation of these complexes and perhaps confirm the ability of
squarate to serve as a pathway for magnetic superexchange interaction between
paramagnetic ions.
F. Dumitru et al. 373
4 References
1 R. West and H. Y. Niu, J . Am. Chem. Soc., 1963,85,2589.
2 J. C. Trombe, J. F. Petit and A. Gleizes, New J . Chem., 1988,12, 197.
3 J. C. Trombe, J. F. Petit and A. Gleizes, Inorg. Chim. Acta, 1990,167, 69.
4 J. C. Trombe, J. F. Petit and A. Gleizes, Eur. J . Solid State Inorg. Chem., 1991,28,669.
5 C. Robl and A. Weiss, 2. Naturforsch., Teil B, 1986,41, 1490.
6 C. Robl, V. Gnutzmann and A. Weiss, 2. Anorg. Allg. Chem., 1987,549, 187.
7 C. Robl and A. Weiss, Muter. Res. Bull., 1987,22, 373.
8 X. Solans, M. Aguilo, A. Gleizes, J. Faus, M. Julve and M. Verdaguer, Inorg. Chem.,
1990,29, 775.
9 (a) D. Coucouvanis, D. G. Holah and F. J. Hollander, Inorg. Chem., 1975,14,2657; (b)
D. Coucouvanis, D. Swenson, N. C. Baenziger, C. Murphy, D. G. Holah, N. Sfarnas,
A. Simopoulos and A. Kostikas, J . Am. Chem. Soc., 1981,103,3350.
10 R. Grenz, F. Gotzfried, U. Nagel and W. Beck, Chem. Ber., 1986,119,1217.
11 Q. Chen, S. Liu and J. Zubieta, Inorg. Chim. Acta, 1989, 164, 115.
12 Q. Chen, L. Ma, S. Liu and J. Zubieta, J . Am. Chem. Soc., 1989,111, 5944.
13 0.Simonsen and H. Toftlund, Inorg. Chem., 1981,20,4044.
14 G. Bernardinelli, P. Castan and R. Soules, Inorg. Chim. Acta, 1986,120,205.
15 R. Soules, A. Mosset, J. P. Laurent, P. Castan, G. Bernardinelli and M. Delamar,
Inorg. Chim. Acta, 1989,155, 105.
16 J. P. Chesick and F. Doany, Acta Crystallogr., Sect. B, 1981,37, 1076.
17 C. E. Xanthopoulos, M. P. Sigalas, G. A. Katsoulos, C. A. Tsipis, C. C. Hadjikostas, A.
Terzis and M. Mentzafos, Inorg. Chem., 1993,32,3743.
18 R. Soules, F. Dahan, J. P. Laurent and P. Castan, J . Chem. Soc., Dalton Trans., 1988,
587.
19 F. Lloret, M. Julve, J. Faus, X. Solans, Y. Journaux and I. Morgenstern-Badarau,
Inorg. Chem., 1990,29,2232.
20 G. M. Frankenbach, M. A. Beno, A. M. Kini, J. M. Williams, U. Welp, J. E. Thompson
and M. H. Whangbo, Inorg. Chim. Acta, 1992,192,195.
21 G. Bernardinelli, D. Deguenon, R. Soules and P. Castan, Can. J . Chem., 1989,67,1158.
22 I. Castro, J. Faus, M. Julve, Y. Journaux and J. Sletten, J . Chem. Soc., Dalton Trans.,
1991,2533.
23 J. A. C. van Ooijen, J. Reedijk and A. L. Spek, Inorg. Chem., 1979,18,1184.
24 Q. Chen, S. Liu and J. Zubieta, Angew. Chem., Int. Ed. Engl., 1990,29, 70.
25 M. Habenschuss and B. C. Gerstein, J . Chem. Phys., 1974,61,852.
26 A. Weiss, E. Riegler and C. Robl, 2. Naturforsch., Teil B, 1986,41, 1329.
27 J. T. Reinprecht, J. G. Miller, G. C. Vogel, M. S. Haddad and D. N. Hendrickson,
Inorg. Chem., 1980,19,927.
28 I. Castro, J. Sletten, L. K. Glzrum, J. Cano, F. Lloret, J. Faus and M. Julve, J . Chem.
29 I. Castro, J. Sletten, M. L. Calatayud, M. Julve, J. Cano, F. Lloret and A. Caneschi,
Inorg. Chem., 1995,34,4903.
30 I. Castro, M. L. Calatayud, J. Sletten, F. Lloret and M. Julve, J . Chem. Soc., Dalton
Trans., 1997, 811.
31 I. Zugravescu, M. Petrovanu, E. Rucinschi, Rev. Roum. Chirn., 1962,7, 1405.
32 B. T. Heaton, C. Jacob and P. Page, Coord. Chem. Rev., 1996,154,193.
Index of People and Places
People
Adamson, Arthur 10 Duncanson, L.A. 8,12,103 ff.
Albinati, Albert0 193 Durrant, M.C. 253
Akers, Sir Wallace 11, 15, 17 Dwyer, Frank 80
Albert,A. 13 Dyer, Phil 146
Allen, A.D. 71, 169, 18
Arhland, Sten 10, 13, 14,71,304,306 Edwards, J.O. 304,305
Evans, M.G. vi
Basolo, Fred 14,305 Evans, D.J. 258
Bawn, C.E.H. 29
Bercaw, John 187 Feast, J. 143
Berzelius, J.J. 319 Fenton, D.J. 140
Bjerrum, Jannik 9, 13,319 Flood,H. 14
Bjerrum, Nils 9, 13 Fronaeus, Sture 10
Bosnich, B. 187
Bradley, E.M. 181 Gamlen, George A. 26
Burkin, A. Richard 8 Gel’man, A. 101
Burrows, George J. 163 Gray, H.B. 306,321
Butcher, Tony 150 Green, M.L.H. 31
Butlin, K.R. 235 Grove,John 8
Burn,E.W. 5 Guy,R. 26
Carnahan, J.E. 235 Haddad, Tim 189

Chatt (nie Williams), Ethel 6,8, 19,28, Hammaker, Geneva 325
313 Hart,A. 8
Chatt, Joseph v ff., viii, xi, 8, 11, 18, 25, Heaton, B.T. 171
29, 31,45, 56, 59,67, 69, 79, 80,89, Heath, G. 174
101, 103 ff., 123, 140 ff. 163, 169, Henderson, R.A. 176
171 ff., 187, 199,217 ff., 233 ff., 252, Hidai, M. 178, 187, 199
288,303, 306 ff., 313 ff., 328,341, 343 Hieber, W. 29,30,45
Chernyaev, 1.1. 28, 303, 315 Hofmann, A.W. 70
Chini, P. 90 Hooke, Robert xiii
Coles, Martyn 147 Howarth, R.D. 10
Cox, G.E. (Sir Gordon Cox) 236
Craig, D.P. 13 Iggo,Jon 97
Ingold, C.K. 28,105
Dalby, Chris 149 Irving, H.M. 13,304
Darwin, Charles xiii
Davies, N.R. 71,304,306 Jensen, K.A. 9,13
Dewar, M.J.S. 26, 101, 103 ff. Johnson, Sam 194
Dilworth, J. 59, 173 Jargensen, C.K. 308
Dilworth, M. 172 Jarrgensen, S.M. 13
376 lndex of People and Places
Kee, Terry 141 Poole, Andrew 144
Kelly, M. 172 Pope, Sir William J. v, 163
Koetzle, T. 193 Postgate, J. 172
Kohn, W. 311
Kubas, G.J. 56 Richards, R.L. 80,89,288
Rideal, Sir Eric K. 18
Leden,Ido 14 Robinson, Oliver 150
Leigh, G.J. 59, 80, 89, 171 ff., 243,288 Robinson, Sir Robert vi, 6
Lewis, Jack 19,21, 171
Lewis, G.N. 303,304 Sacco,A. 71
Loehr,T. 190 Sampson, Michael T. 13,14
Love, J.B. 191 Schrock, R.R. 143,187
Lowe, D.J. 181 Schwarzenbach, G. 9,13,14,304,306,
319
Mackie,A. 80 Searle, Mike 22
Maccoll, A. 13 Seidel, Wolfram 193
McLeverty, J.A. 140 Senoff, C.V. 187
Macmillan, Jake 8 Shaw, Alan 141
McPartlin, Mary 150 Shaw, B.L. 21,26,27,71,80, 89,306,321
Maitlis, P.M. 140 Sidgwick, N.V. 9,304
Malatesta, L. 70, 71 Sillkn, L.G. 14
Mann, F.G. v, 5, 8,67,69,70, 80, 101, Spence, Sir Basil 89
163,303 Sutton, L.E. 71
Marshall, Ed. 148
Maxwell, J.C. 28 Talarmin, J. 176
Mingos, D.M.P. 80,171 Thorneley, R.N.F. 181
Mond,L. 101 Tolman, C.A. 71 ff., 80
Morokuma, K. 193 Turner, Eustace Ebenezer 163
Musaev, Djamaladdin 193
Mylvaganam, M. 190 Valentine, R.C. 246
Vaska, L. 70
Newman, D. 171 Venanzi, Luigi 14,26,80, 89,90, 341
Newton, Isaac xi, xiii
Nyholm, R.S. v, vii, 13,21,26,69,80, Wallace, Alfred Russel xiii
163 Wass (nee Schmidt), Inga 89
Wells, A.F. 21
O’Brien, Paul ix Werner, Alfred 13,70, 303
O’Brien, Steve 15 Wilkins, R.G. 13,14
Orgel, Leslie 9, 13, 21 Wilkinson, G. 16,22, 29,45, 70,71
Owston, P.G. 11 Williams, A. 8, 26
Williams, David 143
Parr, R. 304,309 Williams, R.J.P. 14
Paske, Rosemary 171 Winter, R.M. 6
Pearman, Alan 175
Pearson, R.G. 71 Yamomoto, Akio 59,71
Pickett, C.J. 176,260 Young, Jr., Victor G. 193
Pidcock, A. 90
Pombeiro, A. 182 Zeise, William Christopher 13, 101 ff.,
316
Index of People and Places 377
Agricultural Research Council xii, 24, Peter Spence & Sons Ltd vi, 6
233 ff., 288 PfizerInc. 98
American Chemical Society 187
Queen Mary College, London 89, 171,
BP Chemicals 149 237
California Institute of Technology 187 Royal Society 22,23

Charles Kettering Laboratory 238 Royal Society of Chemistry (see also The
The Chemical Society (see also Royal Chemical Society) xiii, 102,304,326
Society of Chemistry) 21,23 Royal Veterinary College 171,237
Chemistry Institute of Canada 187
Schmidt’s Restaurant 21
DuPont de Nemours Chemical Corp. 46, ShellLtd 235
234 ff. State University of New York, Buffalo
14
Emory University 193
University of California at Davis 246
Harvard University 22 University College, London vii, 21, 163
Horvione Sociedade Quimica SA 98 University College, Southampton 6
University of Bristol 92
Imperial College of Science and University of British Columbia 187
Technology, London vii, 6,22,28 University of Cambridge v, 6, 24,69,
Imperial Chemical Industries vii, ix, xii, 163,303
11, 15, 16, 18, 19,23,26, 173,236, Emmanuel College 5
303 St John’s College 5
ICI Frythe Laboratory vii, ix, xii, 6, 8, University of Copenhagen 13
11, 14, 16, 18, 19,26,67, 70,89, 173, University of Durham 140
236,303,313 University of Kent at Canterbury 90,92
ICI Akers Laboratory vii University of Leeds 141
ICI Butterwick Laboratory vii, 6 , 8 University of Leicester 146
Industrial Hygiene Unit 20 University of Liverpool 29
Dyestuffs Division 15,21 University of Lund 13
Corporate Laboratory 15 University of Manchester vi, 171
Pharmaceuticals Division 16 University of Milan 90
Heavy Organics Division 16 University of Minnesota 193
University of Oxford 71,90
John Tnnes Centre xii Magdalene College 27
Johnson, Matthey 12 University of Sheffield 10, 140
University of Southern California 10
Massachusetts Institute of Technology University of Sussex vii, xii, 6, 24, 31, 59,
(MIT) 143,187 89, 147, 171 ff., 187, 193, 231,237,252
Microbiological Research Establishment University of Sydney 163
236,243 University of Toronto 187
University of Wisconsin, Madison 235
Nelson School, Wigton 5 Victoria and Albert Museum 4
Nitrogen Fixation Unit xii, 6,24,59,67,
89, 140, 169, 171 ff., 187, 196,217, Woolwich Arsenal vi
231,233 ff., 252,288
Northwestern University 306 Yale University 3 1
Subject Index
Ab initio calculations 32 Bimetal complexes 154 ff.

Acetylene complexes 26 Bite parameters 368
Acetylene hydratase 265 Biuret-based phosphorus diamide 97
Acids and bases, general, hard and soft Blocking ligand 368,369
305 ff. Bond-stretch isomerism 149
Acids, class (a) and class (b) 305 ff. Bonding model 35, 110 ff.
A-frame dimers 221 Bond, coordinate 6
Agostic activation 46, 52 Bond dissociation energies 34, 115 ff.
Agostic interactions 218,219 Boranes
Aldehyde oxidoreductase 263,266 t rimet h y 1bo rane 104
Alkane coordination 164 Bradyhizobiumjaponicum 245
Alkoxylation 343,344 Butadienetricarbonyliron 102
alkoxycarbonyl complexes 344 ff.
Alkylidene/carbene 18 1 CalixC41arenes 201 ff.
Alkylidene (carbyne) complexes 182 Calorimetry 12
Amidophosphines 187 ff. Cambridge X-Ray Crystallographic
Amines Database 36,73,74
N-aminopyrrolidine 180 Carboxylic acids and esters
from dinitrogen complexes 180 acetic acid 64
2,4,6-trimethylaniline 254 benzoic acid 64
Arsines and arsine complexes iv, 5,6,9, diethyl malonate 45
10,69, 105, 329 ff. dimethyl acetylenedicarboxylate 123,
(2-allylpheny1)dimethylarsine 164 124,134 ff.
dimethyl(2-vinylpheny1)arsine 164 dimethyl fumarate 127
1,2-phenylenebisdimethylarsine vi, 46, dimethyl maleate 127, 130, 131, 136
69,163 formic acid 64
Aryl isothiocyanates 131-133, 136 methyl methacrylate 97
Atom transfer reactions 323 methyl propynoate 128,134-136
ATP demand in nitrogen fixation 241, methyl propanate 97
245 Carbon-carbon coupling reaction 145 ff.
Azotobacters 235,240,243 Carbon disulfide 129 ff.
Azotobacter chroococcum 239,240, Carbonyl clusters, nucleophilic attack on
243-245,248 344
Azotobacter vinelandii 238,240, 297 C-H activation 31,46,47,49, 62, 72, 147,
172
Bacterial culture 240 ff. C O exchange rates 97
continuous culture 243 ff. C O fluxionality 90 ff.
Benzoquinol 39 C O insertion reaction 135, 136
Beta-galactosidase 249 I3CO labelling 124
Bifacial addition 358 CAS-SCF Calculations 126, 134
Bifacial dissociation 358 Catalysis, homogeneous 95 ff., 164 ff.
Bipyridine 123,368 activity 154 ff.
Birkeland-Eyde process 234 selectivity 157 ff.
380 Subject Index
Centrifugation 238, 241 Cytochrome DorC 269,275
Charge decomposition analysis 114 ff.
Charge concentration 116-1 18 Denitrification 233
Chatt cycle 177 Density functional theory (DFT) 33 ff.,
Chatt Lectureship 263,304 52,194,229,257, 309 ff.
Chromatography FeMoco and 253
gas 239 Desulfovbrio desulfuricans 240,300
ion-exchange 240 Desulfovibrio yigas 260,29 1
Cinnamyl alcohol 214 Desulfovibrio vulgaris 291
Cis-effect 32 Desulfurization 356
Cisplatin 16 N,N’-dialkyl- 1,4-diazabuta-1,3-diene
Cis and trans isomerism 303 123 ff.
Cleavage, carbon-phosphorus bond 156 N,N’-diaryl-1,4-diazabuta-1,3-diene 123
Clostridium pasteurianum 234, 239, 240, Dewar model, metal-olefin bond 101,
245,288 ff. 103 ff., 106, 107, 110 ff., 316
Clusters Diastereoselectivity 125
carbonyl, in situ derivatisation of 344 Diazabutadiene, dimethyl 123ff.
carbonyl migration in 92 Diazenido-complexes 177 ff.
collision induced dissociation 345 Diazene, dimethyl 199
Fe,S, (see also ferredoxins) 280 Diazene, phenyl 209
heterometal aggregates 356 Diazoalkane complexes 178
iron-imide 280,281 ff. Dielectric constants 9, 12
metal carbonyls 90 ff., 343 ff. Dihydrogen activation, heterolytic 3 1 ff.,
metal clusters 356 41
metalloligands 355 Dihydrogen bonding 30,31 ff., 47 ff.
MoFe,S, 280 Dihydrogen complexes 29, 31ff., 45 ff.,
WFe,S, 280 47 ff.
Cobalt complexes 70, 84,271, 324 1,5-Dihydropyrro1-2-one 130
Cobalt metal, reactions of 84 Dimerisation
Cofactor alkyne 212
iron-molybdenum, of nitrogenase 209, a-o-diyne 2 13
217,253 silylative 214
ligation in 253 1,3-Dinitrobenzene 6
DFT applied to 253 Dinitrogen 67, 171 ff.
EPR spectroscopy 253 nitrogen fixation 171 ff., 278,280,281
stopped-flow IR spectroscopy 253 Chatt cycle 177
thiol binding 254 binding site in nitrogenases 242
selenol binding 254 Dinitrogen activation 200 ff.
cyanide binding 254 N=N cleavage 194 ff., 200,201,280 ff.
proton binding 254 catalytic 175
acetylene reduction 254 hydrogenation 193,200
model complexes, diazene binding Dinitrogen coordination 67, 171 ff.,
254 198 ff.
Co-crystallisation 37 Dinitrogen complexes 169, 173 ff., 187 ff.,
Collision induced dissocation (CID) 198 ff., 217
345 ff. alkylation of 177, 191, 199
Conductance 370 ancillary ligands, role in 190
Cone angle 71 ff. 80 electron transfer to 202 ff.
Copper complexes 228 end-on us. side-on bonding 189, 190,
Cyclometallation 39 204
Cycloaddition, end-on bridging mode 189,200 ff.
1,3 dipolar 123 ff. molybdenum 173 ff., 201
double 126 N-N bond length in 189
(Z)-l-Cyclohexadecen-3-yne 213 niobium 194 ff.
Cyclopropanation 166 osmium 173
Subject Index 38 1
protonation of 174,175, 177, 199, Gallium complexes 86
201 Gallium metal, reactions of 86
Raman resonance spectroscopy and Gene transformation 246
190 Genetics
rhenium 173 ff. Escherischia coli 246,247
ruthenium 173 ff., 187 gene transformation 246
side-on bridging mode 202 ff. nif, genetics of 246 ff.
silylation of 180 nifplasmids 247
tantalum 190 nif, regulation of 249
tungsten 173 ff. plasmid-mediated conjugation 246
vanadium 201 Giberellin 20
zirconium 187 ff. Gold complexes 75,85,221
Dinitrogen compounds viii Gold metal, reactions of 85
Dinitrogen reduction in nitrogenases, Griseofulvin 8
stoichiometry 241 Growth hormones 20
Dinitrogenase 242
Dinitrogenase reductase 242 Haber process 235
Disulfides, RSSR 308 Half-sandwich complexes 142 ff.
Dipole moments 9,71,314-316,322 Hard and Soft Acid and Bases 71,84,85,
1,3-Dipole 123 ff., 126, 134 188,304,305 ff.
Fe-O=C 134ff. class (a) and class (b) 71, 82, 85, 305 ff.,
Fe-N=C 124-127,129 317
Distortional isomerism 149 Hardness 305 ff.
Dithiolene complexes 271 ff. principle of chemical hardness 309
Dithiooxamide 136 Hartee-Fock calculations 33 ff.
a-a-Diynes 213 Heck coupling 51,223
DMSO reductase (DMSOR) 264 ff. Hemi-labile coordination 221
Donor-acceptor bonds 112 ff. Heteroallenes 129
Doxycycline 98 Heterobimetal complexes,
ruthenium-rhodium 154 ff.
Electroactive ligands 360 Heterotrimetal complexes,
Electrochemistry 164, 167,225, 228, 271, ruthenium-rhodium-gold 154 ff
290 ff. 1,lSHexadecadiyne 2 13
cyclic voltammetry 228 Hexafluoroisopropanol 48
titrations, potentiometric 269 HiPIP 296,300
Electrochemical oxidation 164, 167 from Chromatium uinosum, mutant of
Electronic chemical potential 309,310, 296
311 Homogeneous catalysis 95, 157
Epoxidation 225 Huckel calculations 32
EPR spectroscopy vi, 229,241,249, Hydrazine 280,281
266 ff. disproportionation of 209
Escherischia coli 246,247 production from N, 176
EXAFS 94,268,290,292 1,2-diphenyl 280
N-N cleavage in 280,281
Ferredoxin 185,245, 296,297 Fe"/Fe"' mediated cleavage of 280,
Azobacter vinelandii, from 297 28 1
Ferrocene 101 [6-(4-chlorophenyl)pyridazin-3-y1]
Flavodoxin 245 369
Formaldehyde, elimination of 349 phenyl 209
Fourier transform ion cyclotron Hydrazido(2-) complexes 174 ff.
resonance 349 Hydrido-complexes 45 ff.
Fragmentation, mass spectroscopic Hydrocarbon activation 49,50 ff.
344 ff. Hydrocyanation 71
Friedel-Crafts reaction 214 Hydroformylation 97, 154 ff., 218
Fuzzy logic 309 of hept-1-ene 218
382 Subject Index
Hydroformylation (cont.) Irowimide chemistry 279 ff.
of oct-1-ene 157 Irving-Williams series 13
catalysis of 159 ff., 218 Isocyanides 129 ff., 130, 131, 172, 181 ff.
mechanism of 159 Isolobal relationship 127, 128, 143, 144,
intramolecular hydride transfer and Isothiocyanates 131, 132, 133, 136
160 Isotope labelling 172, 176
regioselectivity 157, 159 2H labelling 239, 254
synergistic effects 160 lSNlabelling 176,242
Hydrogen exchange 54 l 8 0 labelling 266, 325
a-Hydrogen elimination 59 ff. 19sPtlabelling 323
Hydrogen fluoride complexes 40 Isotope labelling and NMR 92,94
Hydrogenase 260 Isotope exchange 254
Hydrogenation 46,99, 188,218
asymmetric 188 Karilon 98
Hydrogen bonding 27,31,33 ff. Kinetics, stopped flow 241
in rubredoxins 288 ff. Kinetics, rapid quench 241
Hydrogen bond energies 34 ff. Kinetics and mechanism 314ff., 321 ff.,
Ni-Fe hydrogenases 3 1 328 ff.
Hydrogen migration 92 Klebsiella pneumoniae 240,241,245-249
Hydrogen transfer 45 ff.
Hydrosilylation 218,227 Ligands, 3-nitrogen donors 254
Hydratases 265 3-sulfur donors 255
Hypothermophilic archea 263 4-sulfur donors 254
tungsten as essential element in 263, NS, donors 255 ff.
265 Linkage isomerism 325
Hypoxicity 228
Manganese complexes 82, 123,127 ff.
Imido half-sandwich compounds 142 Manganese metal, reactions of 82
Indium complexes 32,70, 86,226, 343 ff. Magnetic moments 257,372
Indium metal, reactions 86 variable temperature 259
Infrared spectroscopy 12, 18,22,26 ff., Manganese-diimine complexes 123 ff.
38, 105, 149 ff., 209,253,314, 329 ff., Manganese-imine complexes 123 ff.
370 Mass spectrometry, Electrospray
high pressure and 97 ionisation 218, 343 ff., 360 ff.
Iogansen equation and 38 Mass spectrometry, Energy-dependen t
time-resolved 329, 334 ff. electrospray ionisation 343 ff.
Insertion and deinsertion reactions tandem 347ff.
52 Mechanisms, dinitrogen alkylation 177,
CO insertion 129, 135, 136 179
competition of C O and RNC 129 oxidative addition 328 ff.
isonitrile insertion and deinsertion substitution reactions 314 ff.
129,131 Mercury complexes 75
olefin insertion 129, 130 Metal carbonyl clusters 45 ff., 89 ff.
Intermolecular interactions of pendant Metal hydrido complexes 29, 31 ff., 59ff.
groups 31 ff. reactivity 38
International Coordination Chemistry with silanes 59 ff.
Conference ix, xiii, 13, 303 Metal-olefin bonding 73, 101 ff.
Iogansen equation 38 103 ff., 106, 107, 110 ff., 316
Ion cyclotron resonance 349 Metal powders, reactions of 79
Iron complexes 41,45, 83, 123 ff., 129, Metalla-1,3-dipoles 123 ff.
262 ff. Metallahydrazides 200
nitrosyls 260 Metallaacetylides 200
with an NS, donor 257 Metallacumulenes 200
Iron-diimine complexes 123 ff. Metallacyclopropanes 123 ff.
Iron-imine complexes 123 ff. Metallocenes 149
Subject Index 383
Metalloligands 355 ATP demand in 241,245
Ortho-metallation 164,227 Azotobacter chroococcum 239,240,
Metathesis, olefin 50, 147 243-245,248
Methacycline 98 Azotobacter vinelandii 238,240
Methoxycarbonylation 97,98 Bradyhizobiumjaponicum 245
Methyl iodide 64,329 ff. Birkeland-Eyde process 234
Methyl viologen 269 Clostridium pasteurianum 234, 239,
Molecular dynamics calculations 29 1 240,245
Molecular hydrogen complexes 32 ff. conformation protection from 0, 243
M O calculations 103, 126, 143,229 continuous culture 243
bandgap 311 Desulfovbrio desulfuricans 240
CAS-SCF 126,134 enzyme inhibitors of 239
electron affinity 310 xenon 239
electron density function 310 ferredoxins in 245
electronegativity 31 1 flavodoxins in 245
electronic chemical potential 309, 310, Haber process 235
31 1 hydrogen evolution/recycling 245
Fermi energy 3 11 Klebsiella pneumoniae 240, 245-248
fuzzy logic 309 microaerobic 244
ionization potential 310 Nitrogen Cycle, and 233
Koopman’s theorem 3 10 Nitrogenase 264,279
Molybdenum complexes 59ff., 149, 150, cofactor, FeMoco 253
151,219, 253,ff., 263 ff. ligation in 253
with NS, donor 255 DFT calculations 253
carboxylato complexes 63 ff. EPR spectroscopy 253
Molybdenum-iron proteins, stopped-flow IR Spectroscopy 253
crytsallography 270 thiol binding to 254
Molybdopterin 263,264,270, 272 selenol binding to 254
syntheses leading to 269 ff. cyanide binding to 254
dithiolene ligands and 269 ff. proton binding to 254
Mossbauer spectroscopy 241,280,282 acetylene reduction by 254
Mutagenesis, site-directed 291 ff. models, diazene binding by 209,
serine for cysteine replacement by 293 254
EXAFS studies and 293 MoFe protein of, crystallography of
264
Natural Bond Order analysis 120 MoFe protein of, tungten substitution
Neutron diffraction 32,36,38,93,107, in 264
193 oxygen damage in 243
Neutron scattering, inelastic 49 pyruvate, in enzyme turn-over 234
Nicholas reaction 214 respiration protection 243
Nickel complexes 71, 72,85,95 Genetics
Nickel metal, reactions of 85 gene transformation 246
Nickel tetracarbonyl 101 nifgenetics 246 ff.
Nifgenetics 246 ff. nifplasmids 247
Nifregulation 249 nif, regulation of 249
Nifplasmids 246 plasmid-mediated conjugation 246
Escherischia coli 246 Escherischia coli 246
Niobium complexes 141, 194 ff. , Nitrosation 325
Nitrate reductase 224 Nitrite reductase 265
Nitrile hydratase 224 Nonafluoro-tert-butanol 48
Nitrogenase 172, 181, 183,238 ff., 279 ff. Nuclear magnetic resonance spectroscopy
alternative substrates for 239 (NMR) 29, 32 ff., 46,48, 52, 54, 71,
iron-molybdenum cofactor, FeMoco 90,116,296,314,329
279,280,286 13CNMR 91,94,166
Nitrogen fixation viii, 171 ff.,187 ff., 264 13C(lo3Rh)NMR 93
384 Subject Index
NMR spectroscopy (cont.) Olefin hydroformylation 157 ff.
Computer of Average Transients (CAT) Olefin metathesis 50, 141
90 Olefin polymerisation 141 ff.
continuous wave 90 Oxidases
1D 92 xanthine oxidase 264,265
2D 93 sulfite oxidase 264,265
FT 90 Oxidative addition 52, 60, 188,219,
'H 61 ff., 71, 124, 166, 192,209,227, 328 ff.
269,282 0 x 0 process 16
lH(lo3Rh)NMR 93 Rhodium us. cobalt catalysts 16
'H TI relaxation times 32 ff., 49, 51 Oxygen transfer enzymes 263 ff.
'H variable temperature 33,94, 133 molybdenum centres in 263
*H 130 tungsten centres in 263
H-D coupling constant 51 labelling studies of 266
high pressure 97
HMQC 93,94 Paladium catalysis 97
INDOR 93 Palladium complexes 95, 106
15N 176 Palladium and platinum hydrides 22,
31P 71,82, 158,223, 227 70 ff.
31P(1H}191 Parachor 6
195Pt 91, 94 Pentafluorothiophenol 218
lo3Rh 92, 93 o-Phenanthroline 123
solid state 93, 107 1-Phenylpropen-2-01 214
Nucleophilicity scales 307, 319 Phosphine(v) halides 79 ff.
Nucleophilic reactivity constants 319 Phosphines and phosphine complexes 5,
Nucleophilic discrimination factor 319 6,9, 10, 31 ff., 45ff., 67, 69 ff., 79 ff.,
173 ff., 306 ff.
Olefins as dipolarophiles 127 aminophosphines 154
Olefin complexes 9, 12,26, 111 ff., 140 Ph,PNHPPh, 154
bond dissociation energy 115-1 17, chiraphos 188
119 dimethylphenylphosphine 71,74-76
charge decomposition analysis 112, diphosphines 154
114,115,118 dppm 47,48,72,154
copper(1) 111, 119, 120 Ph,PC( = CH,)PPh, 154
Coulombic interactions 119 ff. dPPP 154
d(C=C) 113,114 dppe 59 ff., 137,176
Dewar-Chatt-Duncanson scheme methyldiphenylphosphine 7 1,74-76
112 ff. prophos 188
electron density distribution 117 tertiary phosphines, general 70 ff,, 136,
electrostatic attraction and 119, 120 163
ligand-to-metal donation, d 111, 112 tricyclohexylphosphine 47 ff., 74-76
metal to ligand back-donation, b 111, triethylphosphine 45, 75,76,95, 136,
112 137
metallacyclopropane model for 112, trimethylphosphine 74-76,79,82
115 triphenylphosphine 46, 51,70 ff., 75,
platinum(I1) 113, 114, 118 76,82, 84,90, 126, 136, 137
[Pt(PH,),(olefin)] 114, 118 triphosphines 156
pyramidalization angle, 8 113, 114 tripod 156
ratio d/b 114, 115 bonding modes of 156
silver(1) 111 P-C cleavage in 156
steric effects 113 tri-i-propylphosphine 49, 51, 52
strained olefins 113 ff. tri-n-propylphosphine 136, 137
symmetry considerations 112 Phosphite, trimethyl 124, 137
Zeise's salt 103,105, 106, 107, 111, 306 Phosphorus trifluoride 72,107
Olefin hydrogenation 46 Photochemistry 60, 165
Subject Index 385
Photocatalysis 332 ff. [Fe(S-Cys),] *-I2- 291
nbonding 103 ff., 308,316,322 [Fe(S-Cys),-,( 0-Ser),] 292
retrodative 72 [Fe(s-Cy~)~(oH)]~-/*- 292
Dewar-Chatt-Duncanson bonding natural rubredoxins 291
model 101,106,107,110 ff., 316 mutant rubredoxins 293
Plasmid-mediated conjugation 246 ff. Reductive elimination 124, 130,329
Platinum complexes vi, 9, 16,45 ff., 90 ff, Reflectance spectroscopy 370
113,114,118,164,313 ff. Regioselectivity 2 18
Platinum hydrides 21,22,28 Resonance Raman spectroscopy 190,
Platinum-sulfur clusters 358 ff. 294
Polarisability 310, 311,316,317 Rhodium complexes 70,90 ff., 164,218,
Polyhedral rearrangements 91 ff. 329,343
Polyhydrides 38,46, 59 ff., Rhodocyclus tenuis 296
Polymerisation 50,148 Rhodobacter capsulatus 265-269,
a-olefin, of 140, 148, 149 R hod0 bacter sphaeroides 26 5-267
ring opening metathesis (ROMP) 50, Ring-opening metathesis polymerization
140,147,166 (ROMP) 141
Polythene 15 Ring-closing metathesis (RCM) 147
Positron emission tomography 228 Ruberythrin 297
Propargyl alcohol 213 from Desulfouibrio vulgaris 297
Proton-hydride interaction 30,32 ff. Rubredoxin 288 ff.
Proton transfer 48,49 effective dielectric constant in 290,294
2-Pyridine-N-aryl carbaldimine 123 EXAFS 288ff.
Pyridine-N-oxide 203 [Fe(S-Cys),] site and 288 ff.
Pyrococcus furiosus 263,266,291,294 from Clostridium pasteurianum 291 ff.
active site of 289
Quantum chemical methods 111 ff. double mutants of 292,293
Quantum mechanical exchange coupling hydrophobic core of 289
48 single mutants of 291,293
surface charge and 289
Raman spectroscopy 95 from Desulfouibrio desulfuricans 300
Redox couples 167,261 from Pyrococcus furiosus 291,294,297
Reductases effect of OH substitution in T y r l l on
dmso reductase (DMSOR) 264 ff., 271, 294
274,275 reduction potential of 290,291
from Rhodobacter capsulatus, effect of cysteine replacement 292
[MoO,(Ser-O)(mpt)] centre in effect of pH 292-294
265,266 effect of OH substitution in Tyrll of
from Rhodobacter capsulatus, Rd(Pf) 294
[MOO(Ser-O)(mpt),] cent re in native proteins 290
265,266 solvation energy and 290
from Rhodobacter capsulatus, Mo temperature dependence of 291
replacement by W in 267-269, Ruthenium complexes 45 ff., 70, 126,
273 163 ff., 208 ff., 218, 319 ff
geochemistry of Mo and W in Ruthenium-silane adducts 53 ff.
relation to 273,274 Ruthenacycles 210
Mo form, activity of 266,267,274 Diruthenacycles 210
W form, activity of 267 alkynyl-vinylidene 2 10
molecular weight of 267 butenylyl 210
nitrate reductase 264
nitrite reductase 265 Sedimentation, high velocity 241
tmao reductase (TMAOR) 267 Sigma-bond complexes 38 ff., 47 ff.
Reduction potentials 290 ff. Si-H activation 60 ff.
[Fe(SEt)J-/*- 290 Si1anes
[F~{O-C,H,(CH,S),),]~-~~- 290 diphenvlsilane 6 1
386 Subject Index
Silanes (cont.) Tetrahydrothiophen 178,192
methylphenylsilane 61 1,3,5,7-Tetranitronaphthalene6
phenylsilane 60,61 Thiolato-complexes 208 ff.
2-tolylsilane 61 TMAO reductase 267,269
4-tolylsilane 61 Tin complexes 86
Silylation 53,59 ff., 214 Tin metal, reactions 86
Silylative dimerization 214 Topological analysis 120
Silver-olefin complexes 103 ff., 106 Trans-effect 34,257,305
Squaric acid 366 Trans-effectlextrakinetic 314-316,
Squarate complexes 376 ff. 321
bite angle 367 Trans-effectlkinetic 314,316,321,
bite parameter 367 322
bonding modes 367 Trans-influence 166
charge density calculations 370 1,3,5-Triaza-7-phosphatricyclo[3.3.1.1]
hydrogen bonding 367 decane 266
polymeric chains 367 Triflic acid 148,209
molecular magnets and 368 TNT vi
Stability constants 13 Tripodal complexes 254 ff.
Stibines and stibine complexes 9,105 Trisamidoamine ligands 200
Stopped-flow IR spectroscopy 253 Tungsten complexes 115-1 19,263ff.
Strapped complexes 163 ff.
Substitution mechanisms 27 Ubiquinol 269
Substitution, nucleophilic _ . 12,229,241,268,
UV/VIS spectroscopy
aquation reactions at Pt" 217 370
at PtI1 314 ff., 321 ff.
at PtIV 322 ff. Valence bond theory 316
PtT1catalysis of substitution at PtTV Vanadium complexes 253 ff.
323 with NS,donors 255
Sulfite oxidase 264,265 Vaska-type compounds 329 ff.
Sulfenic esters, RSOR 308
Sulfur cycle 236 Wacker process 16
Sulfur donors 217 ff.
bisthiosemicarbazones 228 Xanthine oxidase 264,265
phosphinothiolates 219 X-ray absorption spectra 290
thiolates 209 ff., 217 ff., 252 ff. NH-S hydrogen bonds in rubredoxins
thiooethers 253 and 292ff.
thiosemicarbazones 226,227 X-ray crystallography vi, 12,19,22,
triisopropylbenzenethiols and thiolates 36 ff., 46,48,53,61ff., 69,81 ff., 105,
209,218,253 106,124ff., 149 ff., 166,178,193,209,
tri t hiacyclononane 25 3 218 ff., 282,314,329ff., 350
Suzuki coupling 5 1 charge coupled detection and 193
Symbiosis of ligands 308 low temperature data collection and
193
Tantalum complexes 141 ff., 194 proteins and 264,266,270
Tethered complexes 164ff.
Tetrahydropyran 178 Zeise's salt 101,103,105-107,11 1, 316
1,3,3a,6a-Tetrahydropyrro10[3,2-b]pyrrol Zirconium complexes 144 ff., 187 ff.
127 Zirconium tetra-ally1 15

Modern Coordination Chemistry 2002 Leigh Winterton

Uploaded by

Copyright:

Available Formats

Modern Coordination Chemistry 2002 Leigh Winterton

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Modern Coordination Chemistry 2002 Leigh Winterton

Uploaded by

Copyright:

Available Formats

Modern Coordination Chemistry

The Legacy of Joseph Chatt

ROYAL SOCIETY OF CHEMISTRY

0The Royal Society of Chemistry 2002

All Rights Reserved.

Published by The Royal Society of Chemistry,

Registered Charity Number 207890

For further information see our web site at www.rsc.org

Typeset in Great Britain by Vision Typesetting, Manchester, UK

‘The Child is Father of the Man’

Lord Lewis of Newnham

G. J. Leigh and N. Winterton

Section B Recent Developments in the Synthesis, Bonding Modes and

Hydrides, Hydrogen Bonding and Dihydrogen Activation 31

Hydrides and Dihydrogen Ruthenium Complexes:

Hydrido Complexes of Group 6 Transition Metals -

Section C: The Chemistry of Phosphines 67

Some New Insights into the Steric Effects of Tertiary

NMR Studies of Metal Complexes and Clusters with

Section D: Transition Metal Complexes of Olefins, Acetylenes, Arenes

Some Notes on the Early Development of Models of Bonding

The Dewar-Chatt-Duncanson Bonding Model of Transition

CycloadditionReactions with Metalla-l,3-Dipoles 123

A Journey in Metal-Ligand Multiple Bond Chemistry 140

Tethered Arene Complexes of Ruthenium 163

Section E: Chemistry Related to Dinitrogen Complexes 169

Chemistry at the Unit of Nitrogen Fixation 171

Dinitrogen Activation by Early Transition Metal-

Metal-D initrogen Chemistry after Chatt 198

Novel Chemical Transformationsat Diruthenium Centres

The Chemistry and Applicationsof Complexes with Sulfur

Biological Nitrogen Fixation 233

Vanadium, Molybdenum and Iron Complexes Based on a

Iron-Imide Clusters and Nitrogenase: Abiological Chemistry

Determinants of the Reduction Potential in Rubredoxins,

Mechanisms of Platinum Reactions 313

Tuning Rhodium(1) Metal Centre Accessibility in Iodomethane

Formaldehyde Elimination from Methoxylated Transition

Exploring New Structures Based on Chatt’s {Pt,S,) Core for

A Rational Design of HeteropolynuclearSquarate

Index of People and Places 375

Hacac acet y lacet one; pent an-2,4-dione

CDA charge decomposition analysis

dab N,N’-disubstituted 1,4-diazabuta- 1,3-diene

Glu glutamic acid

HiPIP high potential iron protein

INS inelastic neutron scattering

nuclear magnetic resonance

RCM ring-closing metathesis

H2(S4) 1,2-HSC,H,SCH2CH,SC,H,SH- 1,2

terPY 2,2’:6’,2’’-t erpyridine

Reminiscences of Joseph Chatt

School of Chemistry, Physics and Environmental Science, University of Sussex,

In September 1949, I was just completing a PhD supervised by Richard Burkin,

having no axe to grind, and a pressing desire to go to lunch. We were rumbled.

When given the opportunity to contribute to this publication commemorating

Recent Developments in the

One of the trans hydrides in 5 can readily be replaced by any of a variety of

3 The Nature of A-H-H-E Hydrogen Bonding