Advances in Inorganic Chemistry Volume46

Advances in
INORGANIC CHEM1STRY
Volume 46
ADVISORY BOARD
1. Bertini
D. M. P. Mingos, FRS
Universita degli Stud; di Firenze

Florence, ltaly
lmperial College of Science,

Technology, and Medicine
London, United Kingdom
A. H. Cowley, FRS
J. Reedijk
University of Texas
Austin, Texas, USA
H. B. Gray
California Institute of Technology
Pasadena, California, USA
M. L. H. Green, FRS
University of Oxford
Oxford, United Kingdom
0. Kahn
lnsfitut de Chimie de la Matiere
Condensee de Bordeaux
Pessac, France
Andre
E. Merbach
lnsfitut de Chimie
Minerale et Analyfique
Universite de Lausanne
Lausanne, Switzerland
Leiden University
Leiden, The Netherlands
A. M. Sargeson, FRS
The Australian National University
Canberra, Australia
Y. Sasaki
Hokkaido University
Sapporo, Japan
D.
F. Shriver
Northwestern University
Evanston, Illinois, USA
R. van Eldik
Universitat Erlangen-Numberg
Erlangen, Germany
K. Wieghardt
Max-flanck lnstitut
Mulheim, Germany
Advances in
INORGANIC CHEMISTRY
Including Bioinorganic Studies
EDITED BY
A. G. Sykes
Department of Chemistry
The University of Newcastle
Newcastle upon Tyne
United Kingdom
VOLUME 46
ACADEMIC PRESS
San Diego London Boston

New York Sydney Tokyo Toronto
This book is printed on acid-free paper.
Copyright 0 1999 by ACADEMIC PRESS

All Rights Reserved.
No part of this publication may be reproduced or transmitted in any form or by any
means, electronic or mechanical, including photocopy, recording, or any information
storage and retrieval system, without permission in writing from the Publisher.
The appearance of the code at the bottom of the first page of a chapter in this book
indicates the Publishers consent that copies of the chapter may be made for
personal or internal use of specific clients. This consent is given on the condition,
however, that the copier pay the stated per copy fee through the Copyright Clearance
Center, Inc. (222 Rosewood Drive, Danvers, Massachusetts 01923), for copying
beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent
does not extend to other kinds of copying, such as copying for general distribution, for
advertising or promotional purposes, for creating new collective works, or for resale.
Copy fees for pre-1999 chapters are as shown on the title pages. If no fee code
appears on the title page, the copy fee is the same as for current chapters.
0898-8838/99 $25.00
Academic Press
u division of Hurcouri Bruce & Company
525 B Street, Suite 1900, San Diego, California 92101-4495, USA

http://www.apnet.com
Academic Press
24-28 Oval Road, London NW I 7DX, UK
http://www.hbuk.co.uk/ap/
International Standard Book Number: 0-12-023646-X
PRINTED IN THE UNITED STATES OF AMEIUCA
98 99 0 0 0 1 02 0 3 B B 9 8 7 6
3 2
CONTENTS
The Octahedral MsY8and M6Y,* Clusters of Group 4 and
5 Transition Metals
NICHOLAS
PKOKOPLIK
AND D. F. SHRIVER
1. Introduction
.
.
11. Group 6 .
111. Group 5 Metal Halide Clusters
.
IV. Materials Chemistry Derived from Soluble Metal Halide Clusters
References
.
.
.
.
.
1
3
24
35
44
Recent Advances in Noble-Gas Chemistry

JOHN H. HOLLOWAYA N D ERICG. H o w
I. Introduction
Recent Review Literature

.
The Possibility of Argon Chemistry .
Krypton Chemistry
.
Xenon Chemistry
.
VI. Radon Chemistry
.
References
.
11.
111.
TV.
V.
.
.
.
.
.
.
.
51
53
54
55
61
91
93
Coming to Grips with Reactive Intermediates

ANTHONY
J. DOWNSA N D TIMOTHY
M. GREENE
Introduction
.
Reaction Intermediates: Nerve Centers of Chemical Reactions .
Experimental Characterization of Reaction Intermediates: Retardation
Experimental Characterization of Reaction Intermediates:
Time-Resolved Methods
V. Experimental Characterization of Reaction Intermediates: Flow and
Other Methods .
VI. Conclusions
.
.
References
I.
11.
111.
IV.
.
.
101
106
107
136
155
163
166
vi
CONTENTS
Toward the Construction of Functional Solid-State Supramolecular

Metal Complexes Containing Copper(1) and Silver(1)
MEGUMU
MUNAKATA,
LIANGPINGWU, AND TAKAYOSHI
KURODA-SOWA
I.
11.
111.
IV.
V.
VI.
VII.
VIII.
IX.
.
Introduction
Helical Frameworks .
S----SContact-Assembled Frameworks
.
Hexagonal Frameworks and Graphite-like Structures
Hydrogen-Bond-Assembled Frameworks
.
a-a-Interaction-Assembled Frameworks
.
Diamondoid Frameworks
.
Other Frameworks Based on Covalent Bonds
,
Concluding Remarks .
.
References
.
.
.
.
.
.
.
.
.
.
.
174
176
192
204
219
228
240
251
292
293
Manganese Redox Enzymes and Model Systems: Properties,

Structures, and Reactivity
NEILA. LAW,M. TYLERCAUDLE,
AND VINCENT
L. PECORARO
I. Introduction
.
11. The Enzymes .
111. Structural Models
IV. Physical Properties
V. Reactivity
.
VI. Conclusion
.
References
.
.
.
.
.
.
.
.
305
310
343
379
393
424
425
441
443
456
470
485
485
Calcium-Binding Proteins
BRYANE. FINNAND TORBJORN
DRAKENBERG
I.
11.
111.
IV.
V.
Introduction
.
Intracellular EF-Hand Calcium-Binding Proteins .
Calcium-Mediated Membrane-Binding Proteins
.
Extracellular Calcium-Binding Proteins
,
Summary
.
References
.
.
.
.
.
.
vii
CONTENTS
Leghemoglobin: Properties and Reactions

MICHAEL
J.
D A V I E S , CHRISTEL M A T H I E U , A N D
.
I. Introduction
11. Structure
.
111. Biological Localization .
IV. Reactions with Different Molecules .
V. Oxidation of Fe(I1) Leghemoglobin .
VI. Oxidation of Fe(II1) Leghemoglobin .
VII. Reduction of Fe(IV)=O Leghemoglobin
VIII. Reduction of Fe(II1) Leghemoglobin .
Reactions of Globin-Derived Radicals
References
.
ALAINPUPPO
495
497
499
501
507
511
519
524
527
538
.
.
.
Ix.
INDEX .
CONTENTS
OF PREVIOUS
VOLUMES
.
.
543
555
This Page Intention ally Left Blank
ADVANCES IN INORGANIC CHEMISTRY, VO1.
46
THE OCTAHEDRAL M6Y8 AND M6Y12 CLUSTERS OF

GROUP 4 AND 5 TRANSITION METALS
NICHOLAS PROKOPUK and D. F. SHRIVER
Department of Chemistry, Northwestern University, Evanston, Illinois 60208
I. Introduction
11. Group 6
A. Synthesis of Group 6 Clusters
B. Axial Ligand Chemistry
C. Inner Ligand Chemistry
D. Redox Chemistry and Photophysics of the Group 6 Metal Halide Clusters
E. Molecular and Electronic Structure of the Group 6 Metal Halid Clusters
A. Synthesis of Cluster Core
B. Redox Chemistry of the Group 5 Clusters
C. Ligand Substitution
D. Electronic and Molecular Structure
E. Niobium Iodide Clusters {Nb,I,}"+
A. Higher Nuclearity Clusters
B. Supported Cluster Materials
C. Charge-Transfer Salt Complexes
D. Extended Solids
E. Chemically Modified Surfaces
References
I. Introduction
The early transition metal halide clusters (MfiY8}"+(M = Mo, W;

Y = C1, Br, I; M = Nb; Y = I) and (MfiYl2}"+(M = Nb, Ta; Y = C1,
Br) are the basis for a diverse solid-state chemistry in which cluster
fragments are interconnected by halide bridges. Discrete clusters containing the {MsYs}"+and {MfiYln}"+
units are obtained from these materials by addition of ligands that disrupt the halide bridges. Control of
the coordination environment about the metal octahedron in the soluble clusters provides control of the chemical and physical properties
1
Copyright C 1999 bv Academic Press
All lights 01 reproduction in any farm reserved
os9s-smwY $25 on
PROKOPUK AND SHRIVER
of the cluster. In addition, the discrete clusters provide building

blocks for the construction of new types of extended cluster arrays
and cluster based materials.
Octahedral clusters of the electropositive metals, groups 3 to 7, are
stabilized by a-donor ligands such as halides, chalcogenides, and alkoxides, but the majority accessible to solution chemistry are the halide complexes. These highly symmetric and aesthetically pleasing
{M6Y8}"' and {M6Y12}n+clusters contain a robust core of six metal
atoms and either 8 face-capping {M,Y8}"+or 12 edge-bridging {M6Y12}"+
halide ligands Y. An additional 6 ligands X are terminally bound, one
to each metal center, completing the structures [M6Y8Y61n-6z
and
[M6Y12&61n-6Z
(Fig. 1).In the present discussion it is convenient to refer t o the charge of only the {M6Y8}"+
or {M6Y,2}"+
unit instead of referring to the entire cluster and its full complement of axial ligands.
This notation is especially useful when there is ambiguity as to the
type and charge of the ligands occupying the axial sites or when the
properties of the cluster core are independent of the axial ligands.
Nomenclature developed by Schafer and von Schnering (1) denotes
the bridging ligands as inner (superscript i) and the terminal ligands
outer (from ausser) or axial (superscript a); thus, [Mo&~,,]~is noted
[Mo6Cl!!Ckjl". Soluble octahedral clusters {M6Y8}ntand {M6Y12}n+
with
inner halide ligands have been reported for Nb, Ta, Mo, and W. Molybdenum and tungsten are observed predominately in the {M,Y,}"+
geometry, and the group 5 metals are found with the {M6Y12}nt
arrangement. One notable exception is the niobium iodide clusters such
as "b618(NH2CH3)6]in which the iodide ligands occupy face-capping
sites on the Nb, octahedron (2).
Solid compounds with {M6Ys}"+or (M6Yl2)"' units and halide ligands
have been observed for other metals, including those of the rare earth
OM
QY
OX
FIG. 1. Structures of [MQY&l' A and [M6Y1&10B.
OCTAHEDRAL CLUSTERS OF TRANSITION METALS
Am Cm Rk Cf Es Fm Md No Lr
FIG.2. Periodic table of metals (shaded) found in {MM,Y,P'or {M,Y,J"+ geometry.
and late d-block elements (Fig. 2). Extensive intercluster bridging has
confined the chemistry of the majority of these compounds to the condensed phases. By contrast, the weak intercluster bridging in the
group 5 and 6 derivatives provides solid-state materials that are easily dissolved into molecular cluster species. Subsequent solution
chemistry has revealed rich photophysical and redox properties for
these compounds. The relationship between the ligation of the group
5 and 6 metal halide clusters and their chemical and physical properties, as well as the subsequent materials chemistry that has evolved
from this association, is the subject of this review. The solid-state and
solution chemistry of the related zirconium halide clusters {Zr,Y,,}"+,
which occur with interstitial atoms occupying the center of the metal
octahedron or hydrides ligands bridging the octahedron faces, is relatively new and has been the subject of a number of reviews ( 3 , 4 ) .
11. Group 6
Recent interest in halide clusters containing the {M,Y8}"+ unit

stems in part from its structural relationship to the superconducting
Chevral phases PbMosQs (Q = S, Se, Te) in which eight chalcogenide
ligands Q occupy the face-capping positions ( 5 , 6). Early development
of the ligand substitution chemistry of the group 6 clusters had led
to recent applications of these compounds to catalysis and materials
chemistry. Molybdenum(I1) chloride was discovered over 100 years

ago (71,and the related cluster M O ~ C ~ ~ ( O H ) ~was
( H one
~ O )of~ the first
complexes characterized by X-ray diffraction (8).Renewed interest in
the substitution chemistry of the inner ligands arises from the promise of generating {MO&g}2- (& = s, Se, Te) units in solution, which
are potential precursors to Chevral phases. Also of current interest is
the discovery of the long-lived excited state exhibited by the molybdenum and tungsten clusters (9, 10).
A. SYNTHESIS
OF GROUP6 CLUSTERS
Early synthetic procedures for the group 6 metal halide clusters
relied on disproportionation reactions of the intermediate halides,
MoC13,MoBr3, WCl,, and WBr4, at elevated temperatures to produce
the clusters MsXlz(11,12). These reactions suffered from inherently
poor yields and the need to synthesize and isolate the reactive intermediate metal halide. McCarley and co-workers ( 1 3 ) introduced the
use of sodium tetrahaloaluminate melts to reduce the halides MoX5
and WX, (X = C1, Br) with aluminum at lower temperatures:
6MoC15+ 6Al- Mo6Cllz+ 6AlC1,
6WCls + 8Al+ W&!l,
+ 8AlC1,.
(1)
(2)
Near quantitative yields are achieved for the molybdenum halides,

but substantially lower yields (50%) are obtained in the synthesis of
the tungsten clusters, which require higher temperatures and the use
of Vycor reaction vessels. A higher yield low-temperature synthesis
was devised by McCarley and co-workers for W6C1,2using iron as the
reducing agent (14).
The chemical reduction of the higher Mo and W halides provides
good yields of the octahedral clusters, but the mechanism is obscure.
By contrast, chemical oxidation of zero-valent Mo and W leads t o the
bromo and iodo cluster species in poor yields but provides considerable insight into the formation of these cluster compounds. The reaction of the hexacarbonyls Mo(CO)$or W(C0l6 with Iz at moderately
low temperatures produces a mixture of metal halide phases (15, 16).
In the reaction W(CO), with Iz, lower nuclearity clusters have been
isolated as reaction intermediates that lead to w6 species. The tri-,
tetra-, and pentanuclear tungsten iodide species are obtained from
W(CO)s and I2 by varying reaction times and temperatures, and the
OCTAHEDRAL. CLUSTERS OF TRANSITION METALS
following reaction sequence was proposed by Franolic, Long, and

Holm:
The lower nuclearity clusters [ M O ~ C I ~were

~ I ~isolated
from MoClSin
AlCl,/KCl/BiCl,/Bi melts that led to the hexanuclear species ( 17).
This is the only intermediate isolated from reaction mixture that pro~ } ~ Solution
+
chemistry has yielded Moq, Mag,
duces the { M O ~ C ~core.
and M o ~halide species from mononuclear molybdenum complexes,
suggesting that a nucleation process similar to that proposed for the
tungsten systems may take place during the formation of hexanuclear
molybdenum clusters ( 18-22).
Extraction of the cluster from the products of the solid-state reactions with HX yields the hydronium salts of the molecular species
(H30),(M&&).6H,0, which are converted into the polymeric material M 6 m & under vacuum at 200C (12, 23). The structure of
M&X$X&, also designated MGX12, consists of a {M6&}4+
core with two
terminal axial ligands X2 and four axial ligands bridged to neighboring clusters Xjp2(11)(Fig. 3). Both M a l zand (H3O),(M&&).6H20
may serve as precursors to new molecular species containing the
metal halide cluster core (MO&}~.
B. AXIALLIGAND
CHEMISTRY
The ligand substitution chemistry of the axial ligands of the group
6 clusters is considerably more developed than the substitution chem-
FIG.3. Structure of Mo,C1,C12Cl,, . Reprinted with permission from Ref. (130).
istry of the inner ligands. This is due largely to the robust nature of
the {M6Y8}*+core, which requires substantially harsher conditions to
displace the inner ligands. The increased lability of the axial ligands
enables the coordination environment about the {M6Y8}4t framework
to be altered without disrupting the metal octahedron on the Y8 cube.
1. Mechanism of Ligand Exchange

Kinetic studies of the ligand exchange reaction
were performed by Sheldon with 36C1-labeled[Mo6C18C1612in aqueous

solutions. The rate-determining step was found to be first order with
respect to the molybdenum cluster and independent of halide concentration, and hence the aquation step
was proposed as rate limiting (24). Schafer found similar results with
the exchange reactions of [Mo6Cl8Cl6I2and [w6c18c1612-with Br- and
I- (25).Two significant findings from Schafers work are the equivalency of the six axial positions and the rate dependence on the metal
itself, which show that reactions of molybdenum clusters proceed
faster than those of the analogous tungsten complexes.
Preetz substantiated the equivalency of the axial positions for the
ligands F-, C1-, Br-, I-, and NCS- by monitoring the 19Fand 15Nand
95MoNMR spectra of the ligand exchange reactions
and
Reaction rates were found to increase with the series X = C1 < Br <
I < SCN < F, equilibration of the thiocyanate cluster was achieved
in 10 h, and 2-3 days were required for the formation of the fluoride
cluster at room temperature (26,271. All the clusters [ M O ~ C ~ ~ F ~ - ~ X I ~ and [Mo&~,(NCS)~-,X,]~including structural isomers Ke., mer or fac
of [Mo&18F3X3I2>
were identified by NMR. In all cases a statistical
distribution of all possible compounds, inclusive of the ratios for the

structural isomers, was obtained. Importantly, no evidence of isomersteering electronic effects was observed.
2. Synthesis of Molecular {M6Y8}Clusters
Various types of neutral ligands, including nitrogen donors (12,23,

28, 29), oxygen donors (12, 28, 30, 311, phosphines (28, 32, 331, and
solvent molecules (28, 34, 351, disrupt the extended MosCllz and
WsCllzsolid structure. Under harsher conditions terminal chloride ligands of MosCllzare displaced by neutral donor ligand, L:
For instance, addition of PPh3 to MosCllz yields M o , C ~ , C ~ , ( P Pat

~~)~
room temperature, but it is necessary to reflux MosCllzwith an excess
of PPh3 in tetrahydrofuran to produce the triphosphine cluster
[ M O ~ C ~ ~ C ~ ~ ((28).
P P ~Similar
~ ) ~ I Cexperiments
~
by Walton and coworkers with PEt3 and Mo6ClIzin refluxing ethanol resulted in the
reduced cluster M o ~ C ~ , C ~ , ( P Ewhich
~ , ) ~ , contains a {Mo,C~,}~+
core.
Magnetic susceptibility measurements conducted on M o , C ~ , C ~ , ( P E ~ ~ ) ~
indicate that the cluster is diamagnetic, a result inconsistent with the
expected odd number of electrons in the {Mo,C~~}~+
core. The compound was subsequently reformulated as the reduced species
[ M O & ~ ~ ( P E ~ ~which
) ~ ] ~ + ,crystallizes with the counteranion
[Mo6Cl8Cl6l2-(321. The introduction of two triphenyl phosphine ligands increases the lability of the remaining chloride ligand in the
biphosphine cluster Mo6C18C1,(PPh3),. Both ethanol and pyridine dissolve Mo6Cl12,forming the solvent complexes Mo6C18Cl,(solvent)z,but
dissolution of the biphosphine cluster M O ~ C ~ , C ~ , ( Pin
P ~these
, ) ~ solvents to leads to the species [Mo,C1,C1z(PPh3)2(HOEt)212+
and
[MosCl&12(PPhs)A py)z12+ (281.
Neutral clusters with the formula Mo6ClsCl4L2can exist as cis or
trans isomers. Differences in polarity of the cis and trans structures
were employed by Saito and co-workers to separate the isomers of
[Mo6C1&1,(P(n-C3H7)3]2by chromatography to produce 3 and 15%
yields of the cis and trans respectively ( 3 3 ) .The less polar trans isomers of MO&!~&~~(PR~)~
(R = n-C,H9, n-C,HI1) have also been isolated
in greater than 40% yields (Fig. 4). Shifts in the 31PNMR spectra
reveal a preference for the formation of the all-trans isomer. DiSalvo
and co-workers effected the isomerization of trans-MosC18C14[P(nC5Hl113]2into ~ ~ ~ - M O ~ C ~ ~ C ~
by, immersing
[ P ( C ~ Hthe
~ )trans-P(n~~~
C,H,1)3 cluster in a THF solution of P(C,H,), (Fig. 5). A 95% conversion
ii
FIG.4. Structure of tralzs-MofiClsC1,tP(n-C4Hs)r12.
to the cis isomer was obtained (36).Alkylation of the trans phosphine

clusters
with trialkylaluminum reagents AIR; (R =
CH, , CzHs, n-C3H7,n-C4H9,and n-C6HI3)leads to the replacement of
two axial chloride ligands with R groups, M O & ~ & ~ ~ ( P (RIz
R , ) ~(Fig.
6). Additional AIR; reacts with M O ~ C ~ ~ C ~ ~ ( P R ,producing
) , ( R ) ~ the trialkyl M O ~ C ~ ~ C ~ ( P R ~and
) ~ (both
R ) mer
~ , and fac isomers were shown
to be present by 31PNMR. These organometallic derivatives are the
only examples in which carbon atoms are directly bound to the metal
atoms of the { M o ~ C ~framework.
~}~+
Addition of two equivalents of sodium methoxide to Mo6ClI2in
methanol leads to [ M O ~ C ~ ~ C ~ ~ ( O(37).
C H Single-crystal
~)~]~X-ray diffraction reveals that the methoxide ligands occupy axial positions in
a trans geometry (Fig. 7). However, it is not clear whether the trans
FIG.5. Structure of cis-MofiC1,C1,[P(C,H,),l,.
FIG.6. Structure of all-trans-Mo~Cl,CI,(C2H,),[P(C,H,),I,.
isomer is formed exclusively or only crystallizes preferentially because neither solution NMR nor X-ray powder pattern data on the
bulk material was reported. In contrast to the bisphosphine clusters
MO,$~~C~,(PR,),
, in which both electronic and steric factors may affect
the cisltrans ratio of the products, only electronic effects are likely to
influence the formation of the trans isomer of [Mo,Cl,Cl,(OCH,),].
The trans isomer of [ M O ~ C ~ ~ C was
~ ~ Bprepared
~ ~ ] from crystalline
MoGClHC14(H20)2 in a heterogeneous reaction with Br-. Homogeneous mixtures of Mo6C1RC1,(H20)2and LiBrl[(CfiH5)4AslBrlead
to bromide-rich clusters [(CfiH5)4As12[Mo6C18Cll,yBr4.81
after only 30
min (35).
FIG.7. Structure of trans-[M0,CI,C1,(0CH,,),]~
10
PROKOPUK AND S H R I W R
Mixed halo clusters (H30)2M6C&Y6

(M = Mo, W; Y = F, Br, I) have
been prepared from the aqueous acids HY and Mo6Cllz(12, 27, 32,
35). It is likely the substitution results from an excess of the acid,
HX, rather than the relative affinities of the { M O ~ Ccore
~ ~ for
} ~ the
~
halide, X-, ions. Dehydration of (H30)z[M6C18Y6]
leads to the mixed
ligand bridged systems Mo6C18YzY41a,
which are analogous in structure and reactivity to MoGCllZin that addition of 2L produces
Mo6Cl8Y4L2.
One of the first comparative studies of the mixed halide
clusters MosC18YzY41z
(Y = C1, Br, and I) revealed that the axial iodide
ligands of M O ~ C ~ , are
I ~ Imore
~ , ~ susceptible to displacement by nitrobenzene than the terminal bromide or chloride lands of Mo6C18YzY41z
(Y = C1, Br) ( 2 3 ) .Dissolution of Mo6C181z141z
in nitrobenzene produces
and the bromo and chloro derivatives yield bisnitrobenzene complexes M O ~ C ~ , Y ~ ( N ~ ~ C ~ H ~ ) ~ .
Strongly basic anions such as hydroxide and alkoxide easily displace the axial chloride ligands of [Mo6C18C1612-;thus, titration of
[MO~C~,C~,]~with OH- initially precipitates the tetrahydroxo cluster
M O ~ C ~ ~ ( O H ) ~14H20,
( H ~ O )which
redissolves as the hexahydroxo cluster [MO&l&OH)6]2-with increasing pH (23). Similarly, when Mo6Cllz
is refluxed with sodium alkoxides in alcohol solution, the alkoxide
substitutes for all of the axial chloride ligands (35, 3 9 ) . Methoxide,
ethoxide, and pentafluorophenoxide clusters [MO~C~,(OR)~]~have also
been generated by this method (35, 38).
Increased lability of the axial positions has been achieved by introducing weakly coordinating ligands such as triflate (trifluoromethanesulfonate) (40, 41 1, tetrafluoroborate (26, 42-45), nitrate (43, 46,
4 7 ) tosylate (401, trifluoroacetate (40,481,and perchlorate ( 3 4 ) .(Caution: perchlorate complexes are hazardously explosive.) Reactions of the
chloro cluster [Mo6C18C1612with the silver salt or the protonated form
of the weakly coordinating base B- generate the reactive species
[MO&18B6]2-:
Only the axial chloride ligands were abstracted by the silver salts or
acids, leaving the inner ligands and the (MO,C~,}~~
core intact. The
weakly coordinating ligands are easily displaced by anionic, X-, or
neutral ligands, L, generating the hexasubstituted clusters
Cotton and Curtis first employed this
[Mo6C1&l2- or [Mo~C~,L,]~+.
strategy with AgC104 to prepare the first tetracationic clusters with
11
the ( M O ~ C ~ ~unit,
} ~ ' [ M ~ ~ C l ~ ( s o l v e n t ) ~ l ( C(solvent
l O ~ ) ~ = DMF or
DMSO) (34). These compounds demonstrate the utility of using
weakly coordinating ions to alter the coordination environments
about the { M O ~ C ~core,
~ } ~but
+ they are potentially explosive.
The triflate derivative [ M O ~ C ~ ~ ( O S O ~(Fig.
C F 81,
~ ~ has
~ I ~been
used
extensively in our laboratory to prepare new cluster derivatives with
neutral ligands not previously observed to bind to the ( M O ~ C ~core
~}~'
in all six axial positions. Limits of the substitutional lability of the
triflate ligand of [Mo~C~,(OSO,CF~),~~have been explored with 19F
NMR (49). Diagnostic shifts in the 19FNMR signal of coordinated and
uncoordinated (free) OSOzCF3- may be employed to monitor the displacement reactions. As expected, only one peak is observed in the 19F
NMR spectrum of [Mo6C18(
OSOzCF3)6]"in noncoordinating methylene
chloride. Two peaks are observed with THF, acetonitrile, and acetone,
suggesting incomplete substitution of the triflate ligand by solvent
molecules. A single peak is observed for [MO&18(OSO$F3)6]2- in both
DMSO and methanol, indicating that the triflate ligands are completely displaced in these solvents (49). Titration of [Mo6C18
(OS02CF3)6]2in methylene chloride with phosphine ligands PR3 (R =
Et, t-butyl, Ph) generates all possible species [MosCls
(OS0zCF3)6-x(PR3),1'2-"'( 0 5 x 5 6) as determined by 31PNMR. Unlike
the ligand exchange reactions between [MO&18F612- and [Mo6C18Y612(Y = C1, Br, I, and NCS), in which a statistical distribution of products was formed (26,271, the x = 2 species is disfavored in the phosphine titration reactions. This may be due in part to the polar CH,Cl,
solvent, which may disfavor the neutral species. Interestingly species
13
FIG.8. Structure of ~Mo6Cl8(OSO,CF3),]'
12
PROKOPUK AND SHRrVER
with x < 6 persist in the presence of excess of the phosphines

P[C(CH3)3]S
and P(C2H5)3( 4 9 ) .It is unclear whether steric or electronic factors prevent full substitution of the triflate ligands by the
phosphines.
The acetonitrile ligated cluster [ M O & ~ ~ ( N C C His
~ )an
~ ]even
~ + more
labile cluster than
DiSalvo and co-workers prepared the hexaacetonitrile complex as the [SbCl,]- salt by treating
[ B U , N ] ~ [ M O , C ~ , ( ~ S O ~ C Fwith
~ ) ~ ] KSH to generate [Bu4N12
[MO~C~~(SH
which
) ~ I , reacts with SbCl, in CH3CN to produce
[ M O ~ C ~ ~ ( N C C H ~ ) in
~ ] less
[ S ~than
C ~ ~5%
] ~ yield (50).Some decomposition of the { M O ~ C ~core
~ } ~ 'is
. required to provide the sixth chlorine
atom for the formation of the [SbCl,]- counteranion. A similar abstraction of inner chloride ligands may have occurred in the alkylation reactions of the phosphine clusters
with trialkyl
aluminum reagents studied by Saito ( 3 3 ) . The [Mo&~&NCCH&~~+
cation also can be prepared in near quantitative yield by reaction of
the azide ligands of [ M O ~ C ~ & Nwith
~ ) ~NO'
] ~ - in acetonitrile; however,
the CH3CN ligands are removed under vacuum and the BF4- ions
then bridge molybdenum centers to fill the coordination sphere of the
{Mo~C~,}~+
units ( 4 9 ) .
The methoxide derivatives [MO&~&OCHS),#- and [ M o , C ~ & ~ ~
(OCHS)212provide an alternative strategy for generating new clusters
with the { M o & ~ ~ core
} ~ + (51-55). The strongly basic OCH,- ions are
easily abstracted by Bronsted acids, liberating methanol. Nannelli et
al. (54)first demonstrated the lability of the methoxide ligands with
metathesis reactions between [Mo~C~,(OCH,)~]~and phenol:
The trans isomer of [Mo,C~,(OCH,A~)~]~(Ar = g-anthracenemethano11 was prepared from trUnS-[MOsC18C14(0CH3)212- (Fig. 91, indicating
that little or no isomerization or ligand exchange between clusters
occurs (37).This strategy is convenient in that readily available proton sources including alcohols, phenols, thiols, and carboxylic acids
can be used, and the only by-products are solvent molecules. Other
organic ligands with pK, values comparable to phenol, such as sulfanamides, benzamide, pentafluoroanaline, acetylacetonate, and nitromethane, do not undergo the expected metathesis reactions with
[MO&l~(OCH&12-(56). Interestingly, the hexaphenoxide derivative
[ M O ~ C ~ ~ ( O C reacts
~ H ~ ) with
~ ] ~ -benzoic acid to produce the benzoate
cluster:
13
e
FIG.9. Structure of trans-lMo,Cl,CI,(OCH,C, ,HqI2lL
However, no reaction is observed between [Mo&!~~(OC~H,),]~and sodium benzoate, suggesting that formation of the alcohol or phenol is
the driving force for these reactions (56).
C. INNERLIGANDCHEMISTRY
1. Mechanism of Ligand Exchange
Substitution of the inner ligands of {MO&!~~}~+

and {WtiC18}4+
clusters
requires more forcing conditions than those used t o exchange axial
ligands. Neither boiling aqua regia nor fuming sulfuric acid affect the
{Mo6ClXJ4+
core (57). Basic solutions, however, disrupt the cubic array
of inner ligands. The inner halide ligands of [ M O ~ C ~ ~ ( O Hand
)~I
[Mo,Br,(OH),I2- are susceptible to partial displacement by hydroxide
ions, [MO&-,(OH),(OH)~I~.
Kinetic studies of the substitution of the
inner chloride and bromide ligands by OH- reveal that the process
occurs with a second-order rate law, which is first order with respect
to both cluster and the hydroxide ion. Interestingly, substitution of
by OH- was found to be autothe bromide ligands of [Mo6BrH(OH),I2
catalytic, suggesting that steric factors play a role in displacement of
the bromides (58,59).
2. Inner Ligand Exchange
As implied in the previous discussion, the inner ligands have lower

lability than their outer counterparts, but full substitution of the in-
14
PROKOPUK AND SHRTVER
ner chloride ligands by methoxide ions can be effected by boiling solutions of sodium methoxide and MosCllzto dryness. The resulting cluster, Naz[Mos(OCH3),(0CH3)6],
which has face-capping methoxide ions,
is pyrophoric (39, 60).
Briickner et al. prepared mixed inner ligand species by heating the
mixed halide clusters MO&l&&/z (X = Br, I) to 400C (44, 45). The
resulting mixtures of clusters were converted to the fluoride derivatives [Bu4NIz[Mos(Y,C1,-,)F6]
and analyzed by NMR spectroscopy. Deconvolution of the 1D and 2D 19FNMR spectra revealed a mixture of
products and structural isomers, all of which were identified in the
NMR spectra. A statistical distribution of products is obtained from
A contrasting result was observed when
Mo6Cl~BrzBr41z.
was heated to produce the z = 3 and 4 species as the most abundant
species. Complexes containing iodide ligands in close proximity are
favored. Complete exchange of both inner and outer chloride ligands
of M&1&1zC14/~
(M = Mo, W)occurs in fused salt mixtures of LWKX
(X = Br, I) (12, 61).
The similarity between the structure of {MosC18}4+
and the MosQs
(Q = chalcogenide) unit of the superconducting Chevral phases has
stimulated interest in the substitution chemistry of the inner ligands
and, in particular, substitution of the inner chloride ligands by a chalcogenide. A number of mixed halide-chalcogenide clusters
{Mo6(Y8-,&)}(X = C1, Br, I; Q = S, Se, Te; z = 1, 3, 4, 6, 8) have
been prepared from Mo6C18C1zC141z
and the elemental chalcogenide at
elevated temperatures (9OOOC) (62-64). The clusters are linked into
extended arrays through both inner and outer bridging ligands
{Mo6(Y,-,Qz)}(Fig. 10). Diamagnetic and dielectric properties of these
solids reveal insulator to metal to superconductor transitions upon
increasing chalcogenide content, halide size, and intercluster
bridging.
84
FIG.10. Intercluster bridging between {MQYJ clusters. The vertices of the cube
represent the inner ligands Y.
15
Lower temperature methods for introducing chalcogenides into the

face-capping positions of MosCllzwere developed in McCarley's laboratory, where MosClle in pyridine was refluxed with NaSH. The resulting soluble mixed-capped cluster (pyH),[Mo,(C1,S)Clsl was obtained in 41% yield (65). Saito used similar conditions followed by
extensive chromatographic separation to prepare the mixed halidechalcogenide clusters [Mo6(Y7Q)&I3-(X = Y = C1, Br; Q = S, Se) and
the disubstituted species [ M O , ( C ~ ~ S ~ ~ ) C(66,
~ ~ I67).
" ' ~ -The chemistry
of these interesting chalcogenide clusters has not been pursued. The
observation that chromatography of the clusters with a chloride-deficient eluent produces the aqua complex [Mo,(C~,Q)(H,O),~~+
indicates
that the axial chloro ligands of the chalcogenide-doped clusters are
more labile than those of the all-chloro complex [Mo6C18C1612-.
Tungsten clusters with inner halide-chalcogenide ligands have
been prepared by the reaction of W6Cllzand AzSe (A = Na, K) (68).
When W6Cl12and a n alkali metal chalcogenide are refluxed in toluene
with triethyl phosphine, clusters with mixed ligands are produced,
[Ws(Se,C18-,)(PEt,),]1t ( z = 7.0, A = Na; z = 6.4, A = K). Slight
changes in reaction conditions or chalcogenide source significantly affect the halide/chalcogenide ratio of the substituted product.
Full substitution of the inner chloride ligands of both MosCllz and
W,Clle by chalcogenides produce molecular clusters with the {M6Q8}n+
cores. When W,Cl,, and elemental sulfur are heated to 300"C, a n uncharacterized cluster results, which reacts with magnesium in the
presence of triethylphosphine to generate W6S8(PEt3),in 10% yield
(69). Better yields for {MsS8}"+(M = Mo, W) are obtained when
M,Cl12,NaSH, and NaOBu are refluxed in pyridine. In these reactions
the NaOBu serves as a proton abstractor and the chalcogenide clusters are isolated as the pyridine adducts ( 1 4 , 70, 71). The chemistry
and structural variations of the group 6 metal chalcogenide clusters
were recently reviewed in this series, and the reader is referred to
this publication for further information on these compounds (72).
D. Redox Chemistry and Photophysics of the Group 6 Metal

Halide Clusters
The robust nature of the (MO,C~~}~+
core in combination with its
photophysical and redox properties is ideal for solar energy storage
(73). Voltammetry on [Mo6Cl8Cl6I2in acetonitrile reveals two l-electron steps at - 1.53 and 1.56 V (vs SCE), corresponding to generation
of the oxidized and reduced species [Mo6C1&!ls]'- and [Mo6Cl8Cl,l3-,
respectively ( 74). Electrolysis of [Mo6ClsCl6l2-a t potentials more posi-
16
tive than that of the 2-/1- couple generates the oxidized cluster
[Mo6C18C16]'-,with no evidence for decomposition or any new electroactive species (75). Coulometric measurements are consistent with
quantitative one-electron oxidation of the cluster; however, the oxidized species [Mo6C18C16]'with a {MO,C~,)~'
core was not isolated (10).
Voltammetry experiments on [Mo6Cl8Clsl2in basic chloroaluminate
melts reveal a reversible oxidation couple corresponding to generation
of [Mo6C1,C161'-. In contrast t o the voltammetry experiments in acetonitrile, the [ M O ~ C ~ ~ C ~ ~ ] ~ ~couple
/ [ MinO the
~ Cionic
~ , Cmelt
~ ~is]chem~~
ically irreversible, presumably due to dissociation of axial chloride ligands (76). The oxidized clusters [MsY&]l- (M = Mo, W; Y = C1, Br,
core have been generated electroI; X = C1, Br, I) with the
chemically in organic electrolytes (Table I) (10,77, 78).
The tungsten clusters [wsY&]'- (X, Y = C1, Br) have been produced by chemical oxidation of [W6Y&l2- with NO+ (79). Oxidation
of WsBr12 with liquid bromine in a net two-electron process, and the
extended structure, WSBr8Br4(Br4)2,2,
results consisting of {W6Br8}6+
units bridged in a linear array by (Br4Y anions (80, 81).The rearrangement of inner halide ligands to edge-bridging positions takes
place with w&l&12c14/2 at elevated temperatures in the presence of
chlorine (821. The resulting core {w&l12}6+contains 12 edge-bridging
chloride ligands, and 6 axial chlorides complete the structure of
W&ll&l6. A n alternative route to w6c1&16 employs octachlorocyclopentene, C5C18,as both the oxidant and the chloride source (83). Oxidation of M o ~ C ~ ~ C ~by, C
C,C&
~ ~ ,leads
,
to M0,$1&13, which can be converted to the tetraethylammonium salt of [Mo6Cl12C1613(83). The
clusters w6c1&16 and [Mo6Cl1,Clsl3-are the only known examples of
TABLE I
REDOX
POTENTIALSO
FOR
1.60
1.38
1.14
0.99
0.93
0.80
0.57
0.56
SOME{MBY,P+CLUSTERS
- 1.56
I0
10
10
78
77
78
78
78
All values were obtained from voltammetry experiments in CH,CN and are referenced t o SCE.
17
group 6 metal halide clusters adopting the {MHY1l}lt

geometry. Conversion of the {M6YR}4+
core to {M,Y,,}+has only been observed in
these reactions, and no examples are known in which the reverse
transformation from {M6Y12}T
to {M6Y8}4+
occurs.
Clusters containing chalcogenide ligands {Mo6(ClR-2Q2)}f1+
are somewhat more substitutionally labile than the all-halo complexes and undergo oxidation at potentials about 0.87 V less positive than that
needed to oxidize the parent perhalo cluster (Table 1). The oxidation
potentials of the chalcogenide-substituted clusters change very little
with the identity of the chalcogenide, but significant solvent dependence is observed for the chalcogen-containing clusters. The cyclic
voltammograms of [ M O , ( C ~ ~ S ) Cin~ ~the
~ noncoordinating solvent
methylene chloride reveal a quasi-reversible wave corresponding to
reduction of the cluster a t 0.14 V vs. the ferrocenelferrocenium couple. In acetonitrile this wave is shifted to 0.28 V, indicating that partial substitution of the axial ligands by CH3CN has occurred. New
waves grow in upon changing the potential on acetonitrile solutions of
[Mo6(C17S)C161,suggesting that further changes in the coordination
environment of the {Mo,(C17S)}+core take place upon switching the
oxidation state of the cluster (66, 84).
In the early 1980s Gray and co-workers discovered an unusually
long-lived excited-state lifetime, 180 p s , for the cluster anion
[Mo&~,C~,]~-.
The emission spectrum is substantially red shifted from
the absorption bands (9) (Fig. 11).This is one of the longest observed
lifetimes of any purely inorganic compound. Shifts in the near-UV
absorption band upon variation of the Y ligands in { M o ~ Y ~and
}~+
{W6Y8}4were initially taken as an indication that the transition originates from an (inner) ligand-to-metal charge transfer; however, the
,/-
FIG. 11. Absorption (full) and emission (dashed) spectra of IBu,NIL[Mo,C1&l,,I in

CH,Cl,.
18
independence of the emission spectra of the {Mo&}~+unit on changing the identity of inner ligands suggests that the transition responsible for radiative decay is between purely metal-based orbitals (10,77,
85, 86). The {W6Y8}4+clusters display emission spectra that vary only
slightly with changes in the inner and outer ligands. A summary of
the photophysical data for various molybdenum and tungsten clusters
is given in Table 11. The low rate constants for the red-shifted radiative decay are attributed to a spin- and symmetry-forbidden character
of the transition. Also important to the low rate constants is the absence of high-energy vibrational modes capable of deactivating the
excited state. Additionally, the halide ligands appear to shield the
Mo6 and w6 core and thereby inhibit energy transfer t o solvent molecules.
From the photophysical and electrochemical data on [ M O ~ C ~ & ~ ~ ] ~ - ,
a modified Latimer diagram can be constructed relating the excited
cluster [ M o ~ C ~ , C ~to
~ ~the
" * oxidized and reduced ground states ( 7 4 )
(Fig. 12). When [Mo6C1&l6I2-*is quenched with an electron donor
such as benzoquinone, [ M O ~ C ~ , C ~is~ ] generated.
~Similarly,
[Mo6C18C1612-*
is quenched by the electron acceptor phenothiazine to
produce [M0,C1&1~1'-.The long excited-state lifetime of [Mo6C1&1612-*
and the strong reducing and oxidizing nature of the electronically
TABLE I1
PHOTOPHYSICAL
DATA^ OF {M6Yap+CLUSTERS
Cluster
A,,~,,,
805
825
880
814
802
766
758
752
701
698
698
7(ps)
180
140
86
190
110
71
84
2.2
4.4
5.6
15
15
19
16
27
30
k , (103 s-1)
k , (103 s-1)
1.1
4.5
2.1
7.0
1.9
13
10
23
10
8
17
11
11
13
Values are reported with acetonitrile solutions of cluster.
10
650
490
310
93
59
50
89
34
20
Refs.
10, 85
89
89
89
10, 85
89
85
10, 77, 89
85, 89
77, 89
77, 89
77, 84
89
77, 89
77, 89
77, 84
pIO,CI,CI,]'~-
- I .J3
t1.30
19
vlo,CI,CI,]'
FIG.12. Latimer diagram of [Mo6C1,C1~I". Excited state energy in eV; electrode potentials estimated vs. SCE in CH,CN.
quenched clusters should be useful for converting solar energy to

chemical energy ( 73).
The potentials in the excited-state Latimer diagram demonstrate
that the annihilation reactions of [Mo6C18C1613and [Mo6C18C161'-liberate enough energy to generate [Mo6Cl8Cl6l"* and electrogenerated
chemiluminescence (ECL). This phenomenon has been used extensively to study electron transfer reactions between the excited states
of a broad range of the halide clusters
(M = Mo, W; X =
C1, Br, I; Y = C1, Br, I) in combination with both electron donors
and acceptors (78, 87, 88). Chemical oxidation of benzyl alcohol and
2-propanol by electrochemically generated [Mo6ClsC1611-has been
demonstrated; however, photochemically generated [MosC18C161'~
is
less effective for the oxidation of alcohols (75). The excited cluster
[Mo6C18C1612-*
is quenched by molecular oxygen through an energy
transfer mechanism to generate singlet oxygen, which was identified
by reaction with l-methylcyclohexene (891.
E. MOLECULAR
AND ELECTRONIC
STRUCTURE
OF THE GROUP
6 METAL
HALIDE
CLUSTERS
1. Electronic Structure
Electronic structure calculations carried out at a variety of levels
on these octahedral metal-halide clusters vary in detail (90-104). Our
view of the metal-based orbitals responsible for metal-metal bonding
in the {M6Y8}4+
unit has evolved very little since Cotton and Haas'
semiempirical treatment in 1964 (90). Undoubtedly the importance of
symmetry in the calculations has led to the consistent description of
the bonding scheme. Only a brief summary emphasizing the relation
between the proposed models to the physical and chemical properties
of the clusters is presented here. For a more detailed discussion of the
bonding numerous reviews are available ( 105-1091.
20
PROKOPUK AND SHRIWR
Orbitals responsible for the metal-metal bonding in {M6Y8}4+

have
t2", and eg symmetry and are directed along the edges of
alg,tlu,tZg,
the M6 framework, generating 12 two-electron bonds. A total of 24
cluster bonding electrons, CBE (CBE number is equal to the number
of valence electrons of the metal atoms minus the number of anionic
ligands and the charge of the cluster) fill the bonding orbitals, creating "closed shell" diamagnetic compounds (Fig. 13). ESR measurements are consistent with the proposed diamagnetism of [M6Y8&I2(M = Mo, W) clusters and paramagnetism of the oxidized species
CM~Y&l*-(10, 79).
Some dispute persists as to the size of the highest occupied molecular orbital (HOMO)-lowest unoccupied molecular orbital (LUMO) gap
in the closed-shell 24-electron species. This aspect of the molecular
orbital diagram is of particular interest for its implications on the
photophysical properties of the clusters. Electrochemical measurements are consistent with a large, 2- to 3-eV, separation; however,
the exact nature of the excited state remains uncertain.
2. Molecular Structure
These highly symmetric compounds provide aesthetically pleasing
structures with near perfect Mo6and
octahedrons that expand and
contract to accommodate different coordination environments. Comprehensive studies of the structures of the halide clusters [Mo6Y&I2-
FIG.13. A molecular orbital diagram for {M,Y,}4+ based on a n Extended Hiickel calculation, adapted from reference (108).
21
(Y = C1, Br, I; X = F, C1, Br, I) by Preetz reveals that both electronic

and steric effects of the inner and outer ligands influence the Mo-Mo,
Mo-Y, and Mo-X bond lengths (42, 43, 110). For a given {Mo,Y,}~+
(Y = C1, Br, I) core, the Mo-Mo bond length increases with the axial
ligands series X = F < C1 < Br < I (Table 111). Expansion of the Mo6
octahedron is attributed to changes in the electronegativity of the
axial ligands. The 95Mo NMR spectra of the various halide clusters
reveal that resonances occur at the lowest field for complexes having
six axial fluoride ligands. A high field shift occurs with increasing
polarizability of the axial ligands, indicating that the electron density
on the metal centers increases with decreasing electronegativity of
the outer ligands. Because this increased electron density coincides
with an expansion of the Mo6 octahedron, it is reasonable to assume
that the added electron density increases occupation of the antibonding orbitals of the Mo6 framework. In addition to increasing the electron density on the metal centers, changes in the identity of the axial
ligands X from F to I also result in a reduced HOMO-LUMO gap as
the ligand field strength of X decreases. The 95MoNMR shift of the
halide clusters varies linearly with the electronegativity of the axial
ligands (110), and a similar relationship exists between the 95Mo
NMR shift and the Mo-Mo bond length (Fig. 14).
TABLE I11
SELECTED
BONDLENGTHS"
OF SOME
LMo6Y8&]*-CLUSTERS
Cluster
d(Mo-Mo)
2.593
[Mo,CI,F,I"
[MofiC1$ls12
2.602
[ M O ~ C I ~ B ~ ~ ~2.604
~ ~
[Mo~CI,IJ2.615
[MotiBr,F61*2.618
[MosBrxClsl"
2.636
[MotiBr8Brtil"
2.640
[MotiBr,I,l"
2.644
[Mo&F,I"
2.650
[Mo,I,CI,]'~
2.655
[MotiIRBrhlZ.
2.670
[Mo&I,I'2.675
IMos(CI$)CIJ
2.609
[MoB(C17Se)ClJ3
2.616
[Mo6(Cl6Se,)C1,1"
2.610
~
" All distances are in A.
d(Mo-Y)
d(Mo-X)
d(Y-Y)
d(Y-X)
Ref.
2.488
2.469
2.465
2.466
2.622
2.604
2.600
2.598
2.798
2.775
2.775
2.767
2.479"
2.501b
2.513"
1.993
2.420
2.565
2.788
2.012
2.451
2.491
2.815
2.008
2.460
2.616
2.846
2.468
2.471
2.496
3.517
3.489
3.483
3.484
3.710
3.680
3.674
3.673
4.006
3.923
3.924
3.911
3.234
3.523
3.629
3.801
3.305
3.591
3.675
3.854
3.394
3.656
3.764
3.923
42
42
42
42
43
43
43
43
3.559
3.591
'' Chalcogens are disordered over the capping positions.
110
43
43
43
66
66
67
22
X.
r
.
Average Mo-Mo bond length I Angstroms
FIG.14. Plot of 95Mo NMR shift vs. average Mo-Mo bond length. Data are taken
from references (43, 66).
The expansion of the Mo6 octahedron correlates with contraction of

the cubic array of the eight inner ligands and shortening of the Mo-Y
bonds. The net result of these distortions is a deviation in the planarity of the Mo-Y4 units on the face of the Y8 cube (Fig. 15). The Mo
atoms of the clusters [MO,C~~&]~(X = F, C1, Br, I) protrude outside
the halide cube with progressively increasing distances as the axial
ligand is changed from F through I. At the other extreme, the metal
atoms lie within the Is cube and approach the Y4 face as X increases
in size for [Mo61&12-. The [Mo6Brs&12-clusters exhibit intermediate
structures, with the metal atoms of [Mo6BrsF612inside the halide
cube and those of [Mo6BrsBr6I2and [ M O ~ B ~ &outside.
]~With the ex~ ] ~ [- M O ~ B ~ ~ overlap
F ~ ] ~ - of
, the inner and
ception of [ M O & ~ ~ Fand
outer ligands induces a strain within the cluster framework, and this
strain becomes more pronounced as the size of either X or Y increases.
The strain may account for the increased reactivity in clusters with
larger ligands (26, 27).
tI
FIG. 15. Distortions of Y8 cube and X, octahedron with changes in X or Y.
23
Changes in the inner ligands Y of the clusters [M6Y8X,12from C1 to

I results in an expanded Ms octahedron and a Y8 cube. Substitution
of chalcogenide atoms for the inner halide ligands has only a modest
effect on the average Mo-Y and Mo-Mo bond length (Table III), with
the chalcogenide ligands distributed among two or more face-capping
sites. However, the terminal chloride-molybdenum bond lengths are
elongated compared t o the all-halide clusters. These bond length
changes have been attributed to decreased electrostatic attraction between the axial chloride ligands and the less positive halidechalcogenide cores { M o ~ Y ~ - , Q ~ ) ~ ~ ~ ~ + .
The pseudo-halide derivatives [Mo&YJ- (X = C1, Br; Y = N3 ( 4 6 ) ,
NCO (471, NCS (X = C1 only) (40, 111, 1121, and NCSe ( 4 7 ) ) have
been reported and crystallographically characterized. Some discrepancies exist in the assignment of the disposition of the ambidentate
ligands. For instance, both N- and O-bound cyanate are indicated by
14NNMR in tetrachloroethane; however, crystallographic studies are
consistent with N-bound cyanate ligands. Selenocyanate is found to
be exclusively N-bound by I4N NMR, IR, and X-ray crystallography.
Thiocyanate ligands of most [ M O ~ C ~ ~ ( S C N
salts
)~I~
are
- assigned as
N-bound to the cluster, with the exception of K2[Mo6C18(SCN)61,
which
has been identified as S-bound in both the solid state and solution.
Interestingly, structures of the cluster anions [Mo6Br8(N3),1vary
with the type of counterion, including relatively noninteracting
[Bu4N]+and [Ph3P=N=PPh3]+.The cluster anions of these salts have
regular octahedral geometries; however, the Mo-Mo and Mo-Br
bonds vary by as much as 8% between the two structures.
Several vibrational studies of the [MsY,&12- clusters have been reported (12, 53, 113-117). The structures of these compounds led
Hogue and McCarley to predict that extensive mixing of the symmetry coordinates in the normal modes would prevent the application of
group-frequency concepts ( 1 2 ) ;however, diagnostic M-M, M-X, M-Y,
and X-M-Y stretching frequencies have been identified, and they
find use in the characterization of the clusters. Selection rules for the
high-symmetry (0,)halide clusters [M6YM&I2-(Y = C1, Br, I; X = F,
C1, Br, I) predict a total of ten Raman active (3A,, + 3E, + 4Tz,) and
five IR (5T1,) active modes. The three totally symmetric (Alg)bands,
which consist of the breathing mode of the M6octahedron, the YMcube,
and the terminal X , ligands, are virtually unmixed and thus provide
diagnostic features that are easily identified by their totally polarized
Raman bands. Decreased symmetry in the environment about the MI;
core, owing to site symmetry in the solid or the presence of axial ligands having lower symmetry, does not prevent application of groupfrequency concepts.
24
Pyridine N-oxide derivatives, M O ~ C ~ , C ~ ~ ( O N C ,have

H ~ Rbeen
) ~ , prepared, and the v(Mo-0) band was assigned around 1200-1270 cm-'.
The v(M-0) frequencies correlated with the Hammett constant (T of
the R group. As the R group becomes more electron withdrawing (or
increases in u),the v(Mo-0) stretching frequency decreases. This observation contrasts with trends in the v(M-0) frequency with latetransition metal complexes such as Ni, Co, and Cu, in which the M-0
vibrational energy increases with the electron-withdrawing nature of
the R group and indicates that n-bonding is unimportant in the axial
ligand-metal bonds (31).
Oxidation of the {M6Y8}4tcore removes electrons from the bonding
orbitals of the M6framework, causing an expansion of the metal octahedron. Few of these oxidized clusters have been isolated and structurally characterized. The EPR spectrum of [Mo6C1&1,]'~indicates
that the compound is distorted along one axis; however, single-crystal
X-ray data is not available to confirm this distortion. The EPR spectrum of the oxidized tungsten cluster [W6Br8Br6l1- contrasts with the
spectrum of [Mo6C1&1,]'- in that the values are consistent with an
undistorted octahedron, which has been verified by the crystal structure of [(Ph3P)2Nl[W6Br8Br61
(10, 79).
Clusters containing the {Ta&112}2t and {Nb6C112}2t

cores have been
known for nearly as long as those of the group 6 metals, and their
structures were first reported by Linus Pauling (118).Despite the
long history of the group 5 metal halide clusters {MsY12}"+(M = Nb,
Ta) the ligand substitution chemistry is not as well developed as that
of the analogous group 6 metal compounds {M6Y8}4+
(M = Mo, W),
because of the extensive redox chemistry associated with the {M6Y1,}"+
core. Five different oxidation states have been observed for the Nb
and Ta cores, with compounds in the n = 2-4 states being most common. The n = 1 and n = 5 states are electrochemically accessible;
however, clusters in these oxidation states have not been isolated. As
the average oxidation state changes for the metals in the cluster core
(M6Y,,}"', the affinity for the axial ligands is altered. In the more electron-rich (reduced) state, the clusters have higher affinity for IT-acid
ligands such as phosphines, and in the more electron-deficient state,
the core has a greater affinity for n-donor ligands such as chloride.
Thus, the stability of a particular oxidation state is altered by the
coordination environment about the {M6Y,,}"+core.
25
A. SYNTHESIS
OF CLUSTER
CORE
clusters of niobium and tanThe most convenient route to {M6Y12}2+
talum is the conproportionation of the pentahalides NbY5 and TaY5
(Y = C1, Br) with excess of the metal in molten alkali halide:
20NaY + 14MY5+ 16M + 5Na4M,Y18.
(13)
Formation of lower nuclearity compounds can be avoided by using a

10 to 7 ratio of NaY to MY5 (119).The conproportionation reaction is
convenient for the synthesis of the group 5 metal clusters because it
allows the use of glass reaction vessels as opposed to electron-beamsealed metal tubes, it provides almost quantitative yield of cluster (in
excess of go%), and it requires only 24 h for the reaction to reach
completion, as opposed to 4-6 days. Work-up of the product in aqueous acid H Y solutions with small amounts of stannous chloride to
prevent oxidation produces M6Y14.8H20.
Unlike the group 6 polymeric
materials MsYlz, M,Y14*8H,O consists of discrete molecules with a
{M6Y12}+core coordinated by two halide ligands and four water molecules, and is accurately described as MsH12Y2(Hz0)4.4H20.
The soluble tertiary salts &Nb6Y12Yshave been used as precursors to other
cluster compounds with the {Nb6YI2)a+
core, with limited success
(120, 121).
B. REDOX
CHEMISTRY
OF THE GROUP
5 CLUSTERS
Initial studies on the redox properties of the {M6Y12}n+
clusters were
reported on poorly defined complexes simply noted as [TasC1,,]2 or
[NbsCIJt (1221. Aqueous cluster solutions were prepared from the
sulfate salts Ta6C1,,S04 and Nb6C1,,SO, with various electrolytes.
Some combination of water molecules and anions fills the coordination sphere of the {M6Y12}+cluster. Despite these ambiguities, the
oxidized compounds {M6Y12}4.C
were obtained with various chemical oxidants by a two-electron transfer process (123):
{M6Y12}2+
+ 20x -+ {MsY12}4+
+ 2Red.
(14)
Subsequent investigations ( 124-126) revealed an equilibrium between the intermediate {MfiY12}3+

state and the {M6Y,2}2+and
{M6Y12}4 clusters,
+
26
and a slow disproportionation of the oxidized { T E ~ ~ Y

(Y~ =z }C1,
~ ~Br)
cluster to mononuclear Ta(V) complexes and the reduced clusters
{Ta6Y12}2tand {Ta6Y1~}~'.
Cyclic voltammetry studies (127) on {TasBr12}2treveal two reversible one-electron oxidations corresponding to generation of the n = 3
and n = 4 states,
*
-e
-e
{Ta6Br1z}2t
te
{Ta6Br12}3+ {Ta6Br12}4t,
+e
(16)
suggesting only a minimal structural rearrangement upon changing

oxidation state. Contrasting results were obtained by Hussey and coworkers (128, 129) who investigated the cyclic voltammetry response
of [TasCllzC16]'6-n)in acetonitrile. Three 1-electron transfer processes
were observed, corresponding to generation of the n = 2-5 species.
The reduced cluster [Ta6C11zC16]4undergoes facile substitution by an
acetonitrile molecule, generating [Ta6CllZClS(CH3CN)l3(Fig. 16),
which is detected in the voltammogram along with the liberated chloride ion. Noncoordinating electrolytes in methylene chloride prevent
the loss of the chloride ligand, and the redox couples are chemically
reversible (Fig. 16). Interestingly, [TasC11S(OSOzCF3)6]2undergoes
two reversible one-electron reductions in acetonitrile to generate
[TasC11z(0S02CF3)s13-and [TasCl,~(oso,CF,),]~-,
with no evidence for
-1.5
0.5 1.0 1.5

Potential in V vs. Ag/AgCI
-1.0 -0.5 0.0
FIG. 16. Cyclic voltammogram of [Bu,Nl,[TasC1,,Clsl in (A) CHZC1, and

0.1 M [Bu,NlBF,. Sweep rate is 0.10V/s.
(B)CH,CN
27
ligand displacement on the cyclic voltammetry time scale. Presumably the strong (T- and n-donor properties of the chloride ions weaken
the Ta-C1 bonds of [Ta6Cl12C1614-,
making the chloride ligands susceptible to displacement. In contrast, the reduced {TasC1,2}2+core of
[TasC112(OS02CF3)6]4is stabilized by the weak donor properties of the
triflate ligands.
In acidic chloroaluminate melts, the redox potentials of the
{Ta&112}2+3+and {Ta&!112}3+14+couples are anodically shifted by one
volt compared to values obtained in acetonitrile electrolytes, suggesting that the majority of the axial chloride ligands have been replaced by [AlClJ ions (129).A similar shift in the redox potentials of
couples is observed when changing
the {Nb6C112}2fi3t
and {Nb6C112}3+14
from a basic chloroaluminate melt to an acidic one. In C1--rich molten
salts, the highly reduced cluster [Nb6c112c1615with a {Nb6clI2}+core
is accessible (128),but this species is not accessible in acidic melts or
acetonitrile (130).At the other extreme, the highly oxidized cluster
[Nb6C112Cls]1with a {Nb6C112}5+
core is generated at 1.76 V vs. SCE in
acetonitrile electrolytes.
+
C. LIGANDSUBSTITUTION
The ligand substitution chemistry of the b6Y12&I6-n- and
[Ta6Y12&](6-1)clusters is complicated by disproportionation reactions
and dependence of the ligand affinity on the oxidation state of the
cluster. Unless measures are taken to control chemical potentials, unwanted redox reactions can produce mixtures of partially substituted
clusters in a mixture of oxidation states. Consequently, no strategies
for exchanging all six axial ligands with a wide range of donor groups
exist. Instead, a number of less versatile methods have been developed that introduce a limited range of ligands a t the axial positions
and provide moderate control of the redox chemistry.
The axial ligands of &Nb6YrLY6(A = alkali metal; Y = C1, Br) undergo simple metathesis with azide and thiocyanate. Dissolution of
&NbsYI2Y6in alcohol solutions of N,: or NCS- produces the crystalline
compounds &NbsYl2& (X = NB, NCS) (131-133). Long reaction
times, on the order of weeks, are required when the A4Nb6Y12Y6
clusters are used as precursors to new cluster compounds. Also, the relative affinity of the {M6Y12}+ core for different anionic and neutral ligands is unknown and may restrict the types of clusters accessible
from &Nb6Y12Y6.
The hydrated complexes M6Yl2Y2(H2O),-4H20
(M6Y14) provide more
versatile precursors t o new {M6Y12}+ (M = Nb, Ta) derivatives. The
28
aqua ligands are slowly displaced by neutral L or anionic X- donor

ligands, generating M6YlzYzL4or [M6Y12Y2&l4-(31, 134-1361. The
range of ligands amenable to this strategy has not been delineated,
but if care is taken to prevent oxidation of the cluster, a variety of
8-acids are likely to bind to the reduced {M6Ylz}2+
core. Saito and coworkers used chromatography to isolate the cis and trans isomers of
the phosphine complexes [M6C112C1z(PR3)4]
in moderate yields (1337%) (135) (Fig. 17). Also isolated in the separation procedure was
the trisubstituted cluster [Ta6C112C13(PEt3)31,
indicating that substitution of the water molecules of Ta6Cll,C1,(H20),.4H20is more complex
than first proposed.
Repeated addition of aliphatic alcohols ROH to M6Y12Yz(H20)4.4Hz0
(M = Nb, Ta; Y = C1, Br) under an oxygen-free atmosphere eventually
yields [M6Ylz(ROH)61X2
(137).Interestingly, 'H NMR and conductivity
measurements indicate that the axial halide ligands (especially bromide) are displaced by methanol before the axial water ligands (138).
Under anaerobic conditions [M6Ylz~OCH3)z(HOCHz)41
can be obtained
from methanolic solutions of M6YlzY2(Hz0)4.4Hz0
and two equivalents
of sodium methoxide (139). Care must be taken to exclude air from
these reaction mixtures to avoid unwanted oxidation of the cluster.
In basic solutions the {M6Y12}'+ core is rapidly oxidized upon exposure to air. Presumably the strong u- and 8-donor properties of hydroxide and alkoxide ligands increase electron density on the
FIG.17. Structure of t r a n ~ - T ~ C l ~ , C l ~ ~ P ( C ~ H ~ ) ~ 1 ~ .
29
{M6Yl,}2rcore, making the clusters [M6Y,2(OR)s14(R = H, CHd susceptible to oxidation. Titration of M6YIzYa(Hz0)4
with OH- precipitates
the neutral complex M,Y,,( OH),(H,O), . 6Hz0under Na (140).Isolation
of M6Y,,(OH),(H20)4.6H,0indicates that the axial chloride ligands are
displaced in preference to the axial H,O ligands. Subsequent addition
of OH- yields &M,Y,,(OH), . In oxygen-containing atmospheres, MfiY12
(OH),(H,O), precipitates and then redissolves as A2MsY,2(OH)fi
with
increasing pH ( 141 1. The methoxide clusters M,Y,2(OCH3),(HOCH3)4
and A,M,Y,,(OCH,), have been prepared in a similar fashion (139)
(Fig. 18).
In acid environments the {M6Y12}2+
core of MsY12Y2(H,0)4.4Hz0
is
depending on the metal, inner
oxidized to either {M6Y,2}3'or {M6Y12}4t
to
ligand, and acid source. Hydrochloric acid oxidizes TafiC112Cl,(H20)4
Ta6C1,,C1,(H,O), with a {Ta6C11,}4+core (142); however, NbsCllzCl~
core (143, 144).
(HzO), is oxidized to [Nb6C1&1,1S- with a {Nb6C112}3+
Synthesis of the two-electron oxidized cluster [Nb6CllzCls12~
requires
(143, 145).
chlorine gas to oxidize Nb,ClIaCla(H~O),
Introduction of the weakly coordinating triflate ion at the axial positions of the {M6y12}"+cluster is accomplished by the reactions
[ B U ~ N I J ~ " ~ C+~xsHOSO2CFS
~~C~~]
+ [Bu4N1~[Ta,C1,,(0S0~CF:~~~1
+ 6HCl(g). (17)
The resulting cluster is an electron deficient species (146) (Fig. 19).
The niobium analog [Bu4Nls[Nb,Cl,a(OS0,CF3)61has also been prepared. Cyclic voltammetry on solutions of [Bu4Nla[Ta6Cllz(OSOzCFB)61
reveals two 1-electron transfers a t 0.89 V and 0.29 V vs. AgIAgC1,
corresponding to generation of [Ta,Cl,,(OSOaCF,),]3- and [Ta6Clln
(OS02CFS)fi]4-,
respectively. These reduction potentials are among the
FIG.18. Structure of [Ta,C1,,(OCH,),JL .
30
PROKOPUK AND SHRIWR
FIG. 19. Structure of [Ta,Cl,,(OSO,CF,),IZ-
most positive observed for the {TasC11z}4+

and {TagC11~)~+
clusters and
attest to the poor donor strength of the triflate ion. The strong oxidizO~CF,)~
the
] simple meing character of [ B U ~ N ] ~ [ T ~ , C ~ ~ ~ ( O Sprevents
tathesis reactions with anionic ligands X- such as C1-, Br-, I-, NCS-,
and CN- that are observed with the analogous [Bu4Nlz[MosCls
(oso2cF,),]
(40). Instead, a complex redox process accompanies the
substitution reactions and reduced clusters [Bu4N13[Ta6Cl12&I(X =
C1, Br, I, NCS) are obtained. A slight excess of X- is required for
complete substitution of the triflate ligands, and the reaction rate increases with the ligand series C1- < Br- < NCS- < I- < CN-. This
trend does not follow the basicity or polarizability of the ligand, so
other factors, possibly oxidation potentials of X-,influence the redox/
substitution process. In fact the reduced cluster [Ta&Il2(OSO2CF3),I3is observed spectroscopically as an intermediate. Reduction of
[Ta6C11z(OSOzCF3)s]2by C1- (1.29 V vs. Ag/AgCl) would not be predicted from redox potentials; however, refluxing methylene chloride
solutions of [ B u , ~ ] z [ ~ a 6 ~ 1 1 z ( ~and
~ ~ z[Bu,NICl
~ F 3 ) 6 yields
]
the reduced product [ B U ~ N ] ~ [ T ~ , C ~ ,It
~ Cis~likely
,].
that oxidation of C1- is
promoted by loss of Clz due to a slow stream of Nz passed over the
reaction mixture. In mixtures of [BU~N],[T~,C~~~(~S~~CF,),I
and the
more reducing ligands I-, Br-, and NCS-, the reduced cluster core
{TasC112}2+
is detected spectroscopically and the proposed reaction/substitution scheme in Fig. 20 is proposed. Surprisingly, the simple metathesis reaction
31
FIG.20. Reaction scheme of [TasCl~~(OSOzCF3),12with X
is not observed for any anionic, X-, or neutral, L, ligands including

phosphine oxides.
D. ELECTRONIC
AND MOLECULAR
STRUCTURE
Numerous electronic structure calculations on the {MfiY12}n+
cluster
have been performed at various levels (90, 91, 93, 94, 96, 98, 104).
The molecular orbital structure obtained from extended Huckel calculations indicates that the eight metal-based orbitals responsible for
M-M bonding in the {MfiY12}n+
unit have alg,tZg,
tlu,and a2,, symmetry
(147)(Fig. 21). In contrast to the directionality of the bonding orbitals
in {M6Ys}4+
(see Section II.E.l), the bonding orbitals of {M6Ylz}"+are
directed along the faces of the metal octahedron, creating 8 threecenter, two-electron bonds. The number of cluster-bonding electrons
available to fill the eight bonding orbitals is equal to the number of
valence electrons of the metals minus the number of anionic ligands
and the charge of the cluster. Thus, [Nb6Cl12C16]4is electron precise,
with 16 CBE [30 - 18 - (-4)1, and [Nb6Cll2Cl6l4is electron deficient
with 14 CBE.
The exact bonding or antibonding character of the a2,,level differs
among the various calculations depending on how the contribution of
the axial ligands to the molecular orbitals is considered. Oxidation of
the {Nb6cll2}" core to {NbfiC112}4+
corresponds to depopulation of the
32
FIG.21. A molecular orbital diagram for {M,Y,,j"+ based on an extended Hiickel calculation, adapted from reference (108).
a2ulevel; thus, the aZulevel is the HOMO or LUMO of the n = 2 or

n = 4 clusters respectively. Magnetic susceptibility measurements
and ESR spectra of the clusters with n = 2-4 confirm the assigned
aZusymmetry of the HOMO/LUMO (148).Hyperfine splitting of the
unpaired electron of [Nb6Y&I3- delocalized over the six niobium centers (spin 9/2) leads to impressively split ESR spectra (Fig. 22). X-ray
2.8
3.0
3.2
3.4
3.6
3.8
4.0
kG
FIG.22. EPR spectrum of [Bu4NI~,[NbsCll2BrGI
in CHaClaa t room temperature.
OCTAHEDRAL, CLUSTERS OF TRANSITION METALS
33
crystal structure determinations for the cluster series [NbsCl12Clsl(n-6)reveal that the average Nb-Nb bond length decreases from 3.029 to
2.899 A as the oxidation state on the {NbsC112}'L+
core is lowered from
n = 4 to n = 2 (149). This trend is consistent with the assignment of
the aZulevel as predominately bonding in character for the all-chloro
clusters. The high energy of the a2"orbital explains the relative stability of the oxidized clusters {M6Y12}3+
and {M6Y12}4+.
The average oxidation state of the metal atoms ranges from 2.33 ( n = 2) to 2.66 ( n = 4).
Numerous UV-vis absorption studies on the Nb and Ta clusters
have been reported; however, despite the existence of one-to-one correspondence in the spectra of the two metals, there is little agreement
on the assignments for the metal-based transitions (135, 150-152).
Solutions of both {TasY12}4+ and {Ta6Y,2}3+
are orange-brown in color,
and the niobium analogs {Nb6Y,2}4t
and {NbsY12}3t
are yellow-brown in
solution. The reduced states {Nb6Y12}2+
and {TasY~z}~~
are olive green
and green, respectively. Attempts to assign the electronic transitions
of the clusters have been complicated by the similarity in the spectra
of {M6Y12}3+
and {M6Y12}4+
clusters, which makes identifying mixtures
of clusters in these oxidation states difficult. Additionally the
{MsY12}2-i
state is intensely colored and therefore easily obscures the
spectra of the oxidized clusters.
E. NIOBIUMIODIDE
CLUSTERS
{NbsIs}"+
Hexanuclear niobium iodide clusters {Nb61B}f1+
are the lone exception to the usual edge-bridging halide geometry adopted by the group
5 metal halide clusters. With eight face-capping iodide ligands, the
{NbsIs}"+core is isostructural with the group 6 metal halide clusters
{MO~C~
and
~ }{W6Clg}4+
~+
(153).The similarity in the metal-ligand configurations permits the use of the molecular orbital structure derived
for the group 6 metal halides {M6Y8}"+
to be applied to the {Nb61R}n+
unit. The Nb61B13
is extremely electron deficient, with 19 CBE available for 12 metal-metal bonding orbitals. This unique electronic
structure gives rise to the distinctive chemical and physical properties
of { N ~ ~ I J J I + .
The binary phase Nbs1813is prepared (154) from the decomposition
of Nb318at 950C:
34
PROKOPUK AND SHRIWR
Increased yields are obtained when Niobium metal is added to the

reaction mixture to reduce the NbI, by-product:
11NbI4 + 13Nb+ 4Nb&I3.
(20)
The addition of CsI to the reaction mixture produces the reduced cluster CsNb61813,
with 20 CBE (155).
The electron deficiency of Nb61813and csNb61813 allows these compounds to absorb hydrogen in the solid state at 300C and atmospheric pressure, generating HNb61813and CsHNb61813( 155, 156).
Low-temperature neutron diffraction studies on Nb61813, HNbs1813,
and DNb61813
reveal that the hydrogen atom is slightly displaced from
the center of gravity of the surrounding Nb6 octahedron (157). The
size of the displacement is enough to accommodate multiple hydrogen
atoms, suggesting that more than one interstitial hydrogen may reside in the cage in high-temperature phases of HNb61813.Adsorption
of hydrogen into the interstitial position is unusual, and this is the
only case in which an interstitial atom is implanted in a preexisting
M6 octahedron of a metal halide cluster. The uptake of hydrogen by
Nb61813increases the number of CBE from 19 to 20, and thus a decrease in the average Nb-Nb bond length is expected. However, single-crystal X-ray diffraction studies of Nbs1813and HNb61813reveal a
slight expansion of the Nb6 cage of HNb61813compared t o the intersticontial free cluster (158) (Table Iv). By contrast, Zr6HCl12(EtNH2)6
tains interstitial hydrogen atoms but an average Zr-Zr bond length
is -0.3 A shorter than the interstitial free [zr&1&C16I2-, which has
a comparable CBE count (both at 13) (159, 160).
Low-temperature magnetic susceptibility measurements on Nb61813
and HNb61813
reveal ground spin states of S = 1/2 and S = 0, respecTABLE IV
NIOBIUM-NIOBIUM
BONDDISTANCES
IN SOME
{Nb6I8P+CLUSTERS
Cluster
CBE
d(Nb-Nb)(AP
Ref.
Average Nb-Nb bond lengths are reported with

crystals at room temperature except HNb61813(211 K).
35
tively. At higher temperatures the former has a ground state S = 3/2

and the latter S = 1 (161).The change in spin states is not due to
simple transitions from the ground to excited state but a coupled electronic and structural phase change (158,161, 162). Heat capacity,
X-ray diffraction, and magnetic susceptibility measurements indicate
that a second-order or higher phase change occurs at 274 K for Nb61813
and 324 K for HNb61813.The structural deformation, a rotation of one
Nb, face of the octahedron by 7", is accompanied by the crossing of
electronic levels, which leads to a reduction of the spin degeneracy.
This spin-crossover transition is made possible by competition between the effects of a spin flip and changes in the molecular orbital
levels that result from the cluster distortion (163).
Two discrete molecular compounds containing {Nb618}"+
units were
obtained from reactions of Nb61813and NH,R (R = CH, , C,H,) (2).The
amines displace all the bridging axial iodide ligands, generating the
neutral complexes [Nb618(NH,R)6].These clusters are electron-rich
with 22 CBE compared to the 19- and 20-CBE {Nb618}n+
cores obtained
from solid-state reactions. Consequently, the increased number of
electrons available for cluster bonding orbitals results in shorter
Nb-Nb distances (Table IV). Magnetic susceptibility measurements
indicate that [Nb618(NH2CH3)6]
is diamagnetic. This observation is
contrary to the prediction of a half-filled e, HOMO for a 22-CBE species with a {M6Y8}"+
geometry.
A. HIGHER
NUCLEARITY
CLUSTERS
Mixed metal clusters have been generated by reaction of organometallic species with ambidentate cyanide ligands and [Mo6C18
(OSO&F,)6]2- (411. Both [CpMn(C0)2CNl- and CpRu(PPh,),CN displace the triflate ligands of [ M o ~ C ~ ~ ( O ~ O ~ forming
C F , ) ~ ]the
~ ~ ,12
metal clusters { M o ~ C ~ ~ [ ( ~ - C N ) M ~ ( C O )and
~ C ~{Mo6C18[(p-CN)
]~}~R U ( P P ~ ~ ) ~ Crespectively.
~ ] ~ } ~ + , Single-crystal X-ray structure determinations on (PPN)2{Mo6C18[(p-CN)Mn(C0)2Cp]6}
reveals that the
[CpMn(C0)2CNl-fragment is coordinated to the {MO,C~,)~+
core via
bridging cyanide ligands (Fig. 23). Interestingly, the electronic spectra of these mixed metal clusters show an extremely intense absorbance that most likely corresponds to a charge transfer band [possibly intervalence charge transfer from the CpMn(C0)2CN- ligandl.
The intensity of this band cannot be accounted for from a simple com-
36
FIG.23. Structure of { M O ~ C ~ ~ [ ( ~ - C N ) C ~ M ~ ( C O ) ~ I ~ } .
bination of the absorbencies of the manganese centers and the

{ M O ~ C ~core;
~ } ~ however, intensity-stealing between a Mo6-based
d-d transition and a charge transfer band may lead to the increased
absorption (Fig. 24).
Cyclic voltammetry on [ B U ~ N ] ~ { M O ~ C ~ , [ ( ~ - C N ) M ~ (reveals
C~)~C~]G)
a broad redox wave assigned to the oxidation of the pendant manganese centers. This wave is anodically shifted 0.26 V from the free
ligand/metal complex and is consistent with reduced electron density
on the C ~ M n ( c 0fragment.
)~
The breadth of the redox wave is attributed to communication between the manganese and molybdenum
atoms through the bridging cyanide ligands. Chemical oxidation of
( M O ~ C ~ , [ ( ~ - C N ) M ~ ( C O ) ~with
C ~ ]NO+
, ) ~ - results in clean substitution
of one of the carbonyls of the manganese fragment by NO, generating (Mo6C1,[(~-CN)Mn(CO)(NO)Cp],)4+.
250
300
350
400
450
500
550
600
wavelength I nm
FIG. 24. Absorption spectra of (A) [ B U ~ N I [ C ~ M ~ ( C O ) ~ C(B)

N I , [Bu,NlZ[Mo6CI,
in CH2Clz.
(OSOZCFJ)61,
and (C) [Bu4NlZ{MosC181(~-CN)CpMn(CO)z16}
37
Reaction of NazMo6C18(
OCH3)s with the carboxylic acid H02CC5H4
FeCp produces the organometallic derivative Na2M06C18(02CC5H4
~}~'
FeCp), (55). The iron centers are attached to the { M O ~ C ~core
through the carboxylate moieties on the Cp ring (Fig. 25). In contrast
to voltammograms of {M0~Cl~[(~-CN)MnfC0)~Cpl,f2,
cyclic voltammograms of Na2MofiC18(OzCCsH4FeCp)6
in DMF display a single redox
wave shifted 0.10 V from that of the acid H02CC5H4FeCp,indicating
that the iron centers are independent and equivalent. Cyclic voltin DMSO reveals a chemically
ammetry on NazMosC18(02CC5H4FeCp)fi
irreversible oxidation of the iron centers. A reduction wave is observed substantially shifted to negative potentials. This chemical irreversibility is attributed to reduction of uncoordinated ferrocenium
carboxylate. A decrease in basicity of the carboxylate moiety upon
oxidation of the pendant ferrocenyl group should make this oxidized
ligand a weaker donor, which is susceptible to dissociation. Presumably, vacant coordination sites of {Mo,C~,}~+
are taken up by solvent
molecules.
B. SUPPORTED
CLUSTER
MATERIALS
A number of strategies have been developed to generate composite
materials by the incorporation of the { M O ~ C ~unit
~ } ~into
'
a host ma-
38
trix. Coordinating polymers such as poly(4-vinylpyridine) (PVP) provide sites to bind the { M o & ~ ~ units
} ~ + directly to the polymer backto
bone. Thus, the addition of Mo6Cl12or [Bu4NI2[Mo6C1,(0SO~CFg)61
PVP in ethanol secures the {MO~C~,}~+
unit to the pyridine moieties of
the polymer. Because the { M O ~ C ~unit
~ } ~acts
+ as a cross-linking agent,
the nature of the axial ligands on { M O , C ~influences
~}~~
the physical
properties of the cluster/polymer composite material ( 164 ). Dissolution of MosCllzin ethanol generates M O , C ~ ~ C ~ ~ ( Ewith
~OH
two
) ~ labile
,
axial sites capable of coordinating pyridine. In contrast [Bu4N12
[MO&1,(OSO&F&] has six sites available for pyridine to bind and
thus provides a more effective cross-linker. The number of accessible
coordination sites is reflected in the glass transition temperature, T,,
of the resulting material. Adsorption of Mo&l12 and [Bu4Nlz[Mo&18
(OS02CF3),]into PVP in a ratio that provides equal donor groups t o
cluster binding sites results in composite materials with a higher T,
than those derived from Mo6ClI2.Immobilization of the { M O ~ C ~core
~}~+
by the PVP does not drastically alter the emission spectrum of the
{ M O ~ C ~core;
~ } ~ however,
+
increasing the number of pyridine groups
coordinated to the cluster lowers the excited-state lifetime (164).The
photocatalytic properties of these materials have been demonstrated,
and singlet oxygen generated by quenching the excited cluster with
O2 has been used to oxidize alkenes (165).
DiSalvo and co-workers developed a monomer in solvent strategy
for preparing monodispersed { M O ~ C ~units
, } ~ ~in a polymer matrix
in a coordinating sol(166).By dissolving
vent with a polymerizable functional group such as N-vinylimidazole
(NVI), a monomer unit containing the {MO,C~~}~+
unit is generated,
[Mo,C~,(NVI)~](OSO,CF,),. The low solubility of these tetracationic
clusters is circumvented by using the monomer unit, in this case NVI,
as the solvent. Polymerization of the cluster/monomer solution produces a monodispersed cluster in an organic matrix. In addition t o
solvating the cluster cations, the uncoordinated monomer/solvent dilutes the highly cross-linking clusters. By adjusting the cluster concentration in the monomer/solvent, the degree of cross-linking can
be controlled.
Oxide supports have also been used to immobilize {Mo,Cl8}*+clusters with various degrees of success. Basic and acidic silica adsorb
[BU,N]~[MO,C~,(OSO~CF,),]
(167).Both electro[ B U ~ N ] ~ [ M O &and
~~C
~~]
static and covalent binding of the {M~,cl,}~+
core to the silica surface
can be obtained with proper choice of cluster, solvent, and silica treatment. The photophysical and photocatalytic properties of the cluster
are retained on the Si02support (168).
39
Intercalation of the aqua complex [Mo&~~(H@),]~+

into layered sodium montmorillonite creates a pillared structure (169). The cluster
cations increase the d-layer spacing from 12.5 A for sodium montmorillonite to 16.6 A for the intercalated composite material. The driving
force for the intercalation process is attributed to ion exchange with
the Na' of the montmorillonite; however, the amount of cluster adsorbed by the layered clay is larger than expected based on the sodium content. Similar observations have been made for the intercalation of {Nb6C112}2+'3+
and {Ta6C112}2+
into sodium montmorillonite,
which results in an increase in d-layer spacing and adsorption of excess cluster (170). The axial ligands attached to the intercalated
group 5 clusters were not unambiguously identified, thus the driving
force for intercalation is unclear.
C. CHARGE-TRANSFER
SALT
COMPLEXES
Organic-inorganic charge-transfer salt complexes containing radical organic cations such as tetrathiafulvalenium (TTF)' or tetramethyltetrathiafulvalenium (TMTTF)' and anionic clusters [M6C18C1612and [M6Cl12C16]'6~"'as counterions have been electrocrystallized.
Other organic-inorganic salt complexes have been prepared from
these organic radicals with smaller anions such as BF;, PF,, ReO;,
and SbF, ( 171). Interactions between the organic radical cations, especially overlap of the HOMO of neighboring cations, can produce
semiconducting, metallic, or superconducting properties in the molecular salt compounds. Using large cluster anions changes the anion/
cation volume ratio, thereby forcing the cation radicals into new packing arrangements. Also, the use of radical cluster anions creates the
possibility of anion-cation spin interactions.
Single crystals of the first charge-transfer salt complex containing
a hexanuclear metal halide cluster, (TMTTF)2[Mo6C18C16],
were prepared from TMTTF and [Et4N]2[Mo6C18C18]
(172). The structure of
(TMTTF)2[Mo6C18C16]
consists of cation dimers with the anionic metal
clusters in a distorted CsCl structure (Fig. 26). The CsCl arrangement
is likely a consequence of the comparable volumes of the cation $mer
and [MO~C~,C~,]~-.
Interdimer distances (S...S distance of 4.87 A) are
quite large, resulting in an insulating material. As expected no ESR
signal could be obtained because the cation radicals are paired.
The tetrakis(methy1thio)tetrathiafulvalenium (CH,S),TTF+ salt of
[Mo6C18(NCs)6]2-was prepared in the same manner as
(TMTTF)z[MosClsCls]and has the same CsC1-based structure with
isolated dimers of (CH,S),TTF'
(111). However, unlike
40
FIG.26. Unit cell of [TMTTF12[Mo6C18C161.
(TMTTF)2[Mo6C18C161,the cation dimers of [(CH3SI4TTFl2

[Mo6C1,(NCS),]interact with the axial ligands of the cluster. Sulfursulfur distances separating the (CH3S),TTFt dimer and
[Mo&~,(NCS)~]~are 3.793 and 3.755 These cluster-cation interactions stabilize a slipped planar configuration in the cation dimer,
which favors a mobile triplet exciton at room temperature. In effect,
the triplet state cation dimers are coupled through the
[Mo6C1,(NCS),l2-cluster .
Charge-transfer salt complexes containing the ESR-active
[NbfiCll2Cll2I3cluster with the organic cation radicals TTFt and
TMTTF' were prepared by electrocrystallization from CH3CN/CH2CI
solutions (130, 173). The TTF' salt, which was isolated as
(TTF)2(Bu4N)[Nb6C112Cllz].CH3CN,
contains isolated TTF' dimers. At
higher temperatures the triplet state of the cation dimer is populated
and interacts with the unpaired electron localized on the {NbfiCL12}3'
core. In contrast, the TMTTFt salt as well as the selenium analog
TMTSF were isolated as (TMTTF),[Nb6ClI2Cll2]
.0.5CH2C1, containing
a [(TMTTF),I3' cation and neutral TMTTF molecule. The TMTTF
units in the [(TMTTF)J3' cation are stacked into infinite layers surrounded by [Nb6C112C11213anions (Fig. 27). The fifth TMTTF molecule
is neutral and orthogonal to the cationic TMTTF species. Four partially oxidized TMTTF molecules are present. The HOMO of the organic stack is 5/8 filled (5 electrons in 4 orbitals), and metallic properties are expected. Conductivity measurements, however, indicate that
the hybrid material is semiconducting, with band gaps of 0.16 and
0.20 eV for the TMTTF and TMTSF, respectively. Inequivalency in
A.
41
FIG. 27. Unit cell of [TMTTFl,[Nb,Cl,,Cl,]; two clusters a t corners of the unit cell
are omitted for clarity. Reprinted with permission from Ref. (130).
the structures of the stacked cations is invoked to explain the electron

localization and the resulting semiconducting properties. A single
ESR signal is obtained for both (TMTTF),[Nb6Cl1,C1,,1 and
(TMTSF)S[NbsC112C112],
which indicates mixing of the localized spins
of the cluster and cation.
D. EXTENDED
SOLIDS
The most notable example of a covalently linked extended solid generated from a soluble precursor is Prussian blue (174). A number of
strategies for assembling discrete mononuclear complexes into ex-
42
tended solids have been adapted to the formation of an extended

metal halide cluster network. Interactions between counter cations
and the ligands of the metal halide clusters often result in solids with
bonding in one, two, or three dimensions. Hydrogen bonding between
ligands with donor or acceptor groups and water molecules in the
crystal lattice also provides intercluster linkage.
Alkali metal salts of [M6Y&12- and [MsY12&l'6-n'often contain cation-cluster interactions that fill the coordination sphere of the alkali
metal cation. When ligands from neighboring clusters coordinate to a
common cation, an extended network results. A recent example is
in
the ferrocenecarboxylate derivative, Na2[Mo6Cls(02CCSH4Fecp)6],
which the ferrocenyl moieties are connected to the { M O ~ C ~core
~}~+
through the carboxylate groups (55).The multidentate properties of
the carboxylate group enable the ligand to coordinate to the cluster
and the sodium counterions. Two carboxylate ligands from one cluster
are linked to a sodium counterion, and a third ligand from an adjacent cluster anion plus a methanol molecule fill the coordination
sphere of the cation. These interactions create a one-dimensional
array of [ M O ~ C ~ ~ ( O ~ C C S H ~anions
F ~ C ~(Fig.
) ~ ]28).
~Crystals of the pseudohalide complexes AM6Y12& (A = K, Rb, Cs;
Y = C1, Br; X = NB,NCS) produced from aqueous solutions incorporate water molecules in the unit cell (131-133). These water molecules as well as the inner halide and the pseudohalide ligands coordinate the alkali cations, resulting in an extended solid with a threedimensional array. Interestingly both the a- and P-nitrogens of the
azide ligand bind the cations, whereas only the sulfur atoms of the
thiocyanate groups coordinate to the alkali metals. In addition to cation bridges, water molecules connect the azide ligands of adjacent
clusters via hydrogen bonding.
FIG.28. Display of the linear chain of {Mo&&}~+

units connected through bridging
carboxylate ligands and sodium ions of Naz[Mo6Cls(OzCC6H4Fecp)61.
The ferrocenyl
moieties of the carboxylate groups are omitted for clarity.
43
Extensive hydrogen bonding has been observed in many structures

containing {M6Y12}n+units with OH-, HzO, or CHBOHligands in the
axial sites. The axial ligands act as both proton donors and acceptors.
Hydrogen bonding interactions between the axial ligands and solvent
molecules of the crystal lattice create three-dimensional cluster systems. A few structures containing extended networks of clusters
linked via hydrogen bonds are listed in Table V. It should be noted
that the hydrogens were not located in the structures but inferred
from short donor-acceptor distances. Two compounds, (Me,N)
(175) and [Nb6Brlz(HzO)61[HgBr41
12Hz0(1761, pos[Ta6Cllz(Hz0)61Br4
sess hydrogen bonds between water donors and bromide acceptors.
DC electrical measurements indicate that [Nb6Br12(Hz0)61[HgBr41.
12H20has semiconducting properties; however, the origin of this conductivity is not completely understood.
E. CHEMICALLY
MODIFIED
SURFACES
The affinity of the cluster anions [ B U , N ] ~ ~ ~ C ~ & ] ~
=-C1,
( X Br, I)
for gold and silver surfaces is determined by the identity of the axial
ligand X (177). Only [Nb6CllzBr613adsorbs in a monolayer fashion on
gold as determined by quartz crystal microgravimetry, X-ray photoelectron spectroscopy, and cyclic voltammetry. Redox potentials for
b6Cl12Br613-adsorbed to gold are anodically shifted 0.15 V relative
to the cluster in solution. This shift is consistent with the transfer of
electron density from the cluster to the metal surface. By contrast no
TABLE V
HYDROGEN-BONDED
CLUSTER
COMPLEXES
Cluster
H Bonding
Ref.
175
176
176
179
179
180
181
181
181
Bromide ligands from both the cluster cation and the anion partake in hydrogen bonding.
Cation and anion are involved in hydrogen bonding.
44
sign of niobium or chlorine is observed in the X P S spectrum of gold

surfaces that were exposed to [Bu~N]3[Nb6C112C~613~.
At the other extreme, solutions of [Bu,N]3[Nb6C11zC161
completely dissolve gold films.
The increased reactivity of the clusters with the heavier halides parallels the interaction of gold with the free halide ions., where C1- adsorbs the weakest and I- the strongest (178). Of the three clusters
only [Nb6C1,2C16]3-forms a monolayer on silver surfaces. Both
[Nb6CllzBr6]3and [Nb&llz16]3- dissolve silver. These findings demonstrate that halide ligands should be versatile anchors for immobilizing many types of metal complexes to metal surfaces.
Pyridine groups immobilized on gold surfaces through strong goldsulfur interactions displace the triflate ligands of [Ta6Cl12
(OS02CF3)fi14and result in a immobilized monolayer of cluster on the
gold surface. XPS reveals that an average of four triflate ions are
replaced by the surface-bound pyridines. Dilution of the pyridine surface with coadsorbed thiolphenol, a noncoordinating moiety, reduces
the amount of cluster adsorbed to the surface and the average number of triflate ligands displaced by surface pyridine. The number of
triflate ligands remaining on the surface-bound cluster may be adjusted with the thiolphenol/mercaptopyridine ratio. The different coordination environments about the {Ta&112)"t core in the resulting
material have been identified with differential pulse voltammetry. As
the number of coordinating pyridine ligands coordinated to the
{TafiC112}'1+
core decreases, the oxidation potentials of the clusters are
cathodically shifted. The shift reflects the n--acid character of the pyridine ligand.
ACKNOWLEDGEMENTS
We appreciate the many contributions of former graduate students and senior collaborators, who have contributed to our research in the area of inorganic metal cluster
chemistry: Laura Robinson, Dean Johnston, Joseph Hupp, and Donald Ellis. Our research in this area was supported by the National Science Foundation, inorganic chemistry program and the NSF supported Materials Research Center a t Northwestern University.
REFERENCES
1. Schafer, H.; Schnering, H.-G. v. Angew. Chem. Znt. Ed. Engl. 1964,76, 833.
2. Stollmaier, F.; Simon, A.Znorg. Chem. 1985,24, 168.
3. Ziebarth, R. P.; Corbett, J . D. Acc. Chem. Res. 1989,22, 256.
45
4. Rogel, F.; Zhang, J.; Payne, M. W.; Corbctt, J. D. Adu. Chem. Ser. 1990, 226, 369.
5. Chevrel, R.; Hirrien, M.; Sergent, M. Polyhedron 1986, 5, 87.
6. Hughbanks, T.; Hoffmann, R. J. J. A m . Chenz. Soc. 1983, 105, 1150.
7. Blomstrand, W. J . Prakt. Chem. 1859, 77, 88.
8. Brosset, C. Ark. Kemi. Mineral. Geol. 1943,20A, 1.
9. Maverick, A. W.; Gray, H. B. J. Am. Chem. Soc. 1981, 203, 1298.
10. Maverick, A. W.; Najdzionek, J . S.; MacKenzie, D.; Nocera, D. G.; Gray, H. 3. J .
Am. Chem. Soc. 1983, 205, 1878.
11. Schafer, H.; Schnering, H.-G. v.; Tillack, J.; Kuhnen. F.;Wohrle, H.; Baumann,
H. 2. Anorg. Allg. Chem. 1967,353, 281.
12. Hogue, R. D.; McCarley, R. E. Inoig. Chenr. 1970, 9, 1354.
23. Dorman, W. C.; McCarley, R. E. Inorg. Chenr. 1974, 13, 491.
14. Zhang, X.; McCarley, R. E. Inorg. Chem. 1995, 34, 2678.
15. Drobot, D. V.; Mikhailova, L. G.; Bol'shakiv, K. A.; Sbitnev, V. L. Russ. J . Inorg.
Chem. 1978,23, 643.
16. Franolic, J . D.; Long, J. R.; Holm, R. H. J . A m . Chem. Soc. 1995, 117, 8139.
17. Jiidden, K.; Schafer, H. 2. Anorg. Allg. Chem. 1977,430, 5.
28. Stensvad, S.; Helland, B. J.; Babich, M. W.; Jacobson, R. A,; McCarley, R. E.
J. Am. Chem. Soc. 1978, 100, 6257.
19. Aufdembrink, B. A,; McCarley, R. E. J . Am. Chem. Soc. 1986, 208, 2474.
20. Zietlow, T. C.; Gray, H. B. Iiiorg. Chein. 1986, 25, 631.
21. Cotton, F. A,; Poli, R. J . Am. Chem. Snc. 1988, 220, 830.
22. Burini, A.; Cotton, F. A,; Czuchajowska. J. Pdyhedron 1991, 10, 2145.
23. Sheldon, J. C. J. Ch,em. Soc. 1960, 1007.
24. Sheldon, J. C. J. Chem. Soc. 1960, 3106.
25. Lessmeister, P. v.; Schafer, H. 2. Anorg. Allg. C h e m 1975, 427, 171.
26. Harder, K.; Peters, G.; Preetz, W. 2. Anorg. Allg. Chem. 1991, 598/599, 139.
27. Preetz, W.; Braack, P.; Harder, K.; Peters, G. 2. Anorg. Allg. Chern. 1992, 612, 7.
28. Fergusson, J. E.; Robinson, B. H.; Wilkins. C. J. J. Chem. Soc. A. 1967, 486.
29. Carmichael, W. M.; Edwards, D. A. J. Inorg. Nucl. Chem. 1967,29, 1535.
30. Sheldon, J. C. J. Chem. Soc. 1961, 750.
32. Field, R. A,; Kepert, D. L.; Taylor, D. Inorg. Ch.rm. Acta 1970, 4, 113.
32. Hamer, A. D.; Smith, T. J.; Walton, R. A. Inorg. Ch,em. 1976, 15, 1014.
33. Saito, T.; Nishida, M.; Yamagnta, T.; Yamagata, Y.; Yamaguchi, Y. Inorg. Chem.
1986,25, 1111.
34. Cotton, F. A.; Curtis, N. F. Inorg. Ch,etn. 1965, 4, 241.
35. Schafer, H.; Plautz, H.; Abel, H.-J.; Lademann, D. 2. Anorg. Allg. Chem. 1985,
526, 168.
36. Ehrlich, G. M.; Deng, H.; Hill, L. I.; Steigerwald, M. L.; Squattrito, P. J.; DiSalvo,
F. J. Inorg. Chem. 1995, 34, 2480.
37. Perchenek, N.; Simon, A. 2. Anorg. Allg. Chem. 1993, 619, 98.
38. Schafer, H.; Brendel, C.; Henkcl, G.; Krebs, B. 2. Anorg. Allg. Chem. 1982, 491,
275.
39. Nannelli, P.; Block, B. P. Inorg. C h e m 1968, 7, 2423.
40. Johnston, D. H.; Gaswick, D. C.; Lonergan, M. C.; Stern, C. L.; Shriver, D. F.
Inorg. Chenr. 1992, 31, 1869.
41. Johnston, D. H.; Stern, C. L.; Shliver, D. F. Inorg. Chem. 1993,32, 5170.
42. Preetz, W.; Harder, K.; Schnering, H. G. v.; Niche, G.; Peters, K. J. Alloy. Comp.
1992, 283, 413.
46
PROKOPUK AND SHRJYER
43. Preetz, W.; Dublitz, D.; Schnering, H. G. v.; Safimannshausen, J. 2. Anorg. Allg.
Chem. 1994,620, 234.
44. Briickner, P.; Peters, G.; Preetz, W. 2. Anorg. Allg. Chem. 1993, 619, 551.
45. Briickner, P.; Peters, G.; Preetz, W. 2. Anorg. Allg. Chem. 1993,619, 1920.
46. Bublitz, D.; Preetz, W.; Simsek, M. K. 2. Anorg. Allg. Chem. 1997, 623, 1.
47. Simsek, M. K.; Preetz, W. 2. Anorg. Allg. Chem. 1997, 623, 515.
48. Harder, K.; Preetz, W. 2. Anorg. Allg. Chem. 1992,612, 97.
49. Johnston, D. H. Ph.D. Thesis, Northwestern University, Evanston, 1993.
50. Ehrlich, G. M.; Warren, C. J.; Haushalter, R. C.; DiSalvo, F. J. Znorg. Chem. 1996,
34, 4284.
51. Perchenek, N.; Simon, A. 2.Anorg. Allg. Chem. 1993, 619, 103.
52. Perchenek, N.; Simon, A. Acta Crystallogr. 1991, C47, 2354.
53. Schoonover, J. R.; Zietlow, T. C.; Clark, D. L.; Heppert, J . A,; Chisholm, M. H.;
Gray, H. B.; Sattelberger, A. P.; Woodruff, W. H. Znorg. Chem. 1996,35, 6606.
54. Nannelli, P.; Block, B. P. Znorg. Chem. 1969,8, 1767.
55. Prokopuk, N.; Shriver, D. F. Znorg. Chem. 1997,36, 5609.
56. Prokopuk, N.; Kennedy, V. 0.; Stern, C. L.; Shriver, D. V. Unpublished results.
57. Sheldon, J . C. Nature 1959, 184, 1210.
58. Sheldon, J. C. J. Chem. SOC.1964, 1287.
60. Chisholm, M. H.; Heppert, J . A.; Huffman, J. C. Polyhedron 1984,3, 475.
62. Perrin, C.; Sergent, M.; Traon, F. L.; Traon, A. L. J. Solid State Chem. 1978,
25, 197.
63. Pilet, J. C.; Traon, F. L.; Traon, A. L.; Perrin, C.; Perrin, A.; Leduc, L.; Sergent,
M. Surf Sci. 1985, 156, 359.
64. Perrin, A.; Perrin, C.; Sergent, M. J. Less-Common Met. 1988, 137, 241.
65. Michel, J . B.; McCarley, R. E. Znorg. Chem. 1982,21, 1864.
66. Ebihara, M.; Toriumi, K.; Saito, K. Inorg. Chem. 1988,27, 13.
67. Ebihara, M.; Toriumi, K.; Sasaki, Y.; Saito, K. Gazz. Chim. Ztal. 1995, 125, 87.
68. Xie, X.; McCarley, R. E. Znorg. Chem. 1997,36, 4011.
69. Saito, T.; Yoshikawa, A.; Yamagata, T.; Imoto, H.; Unoura, K. Znorg. Chem. 1989,
28, 3588.
70. Hilsenbeck, S. J.; Young, V. G.; McCarley, R. E. Znorg. Chem. 1994,33, 1822.
71. Ehrlich, G. M.; Warren, C. J.; Vennos, D. A.; Ho, D. M.; Haushalter, R. C.;
DiSalvo, F. J. Znorg. Chem. 1995,34, 4454.
72. Saito, T. Adv. Znorg. Chem. 1997, 44, 45.
73. Gray, H. B.; Maverick, A. W. Science 1981,214, 1201.
74. Nocera, D. G.; Gray, H. B. J. Am. Chem. SOC.1984, 106, 824.
75. Jackson, J . A.; Mussell, R. D.; Nocera, D. G. Znorg. Chem. 1993,32, 4643.
76. Barnard, P. A.; Sun, I.-W.; Hussey, C. L. Znorg. Chem. 1990,29, 3670.
77. Zietlow, T. C.; Nocera, D. G.; Gray, H. B. Inorg. Chem. 1986,25, 1351.
78. Mussell, R. D.; Nocera, D. G. Znorg. Chem. 1990,29, 3711.
79. Zietlow, T. C.; Schaefer, W. P.; Sadeghi, B.; Nocera, D. G.; Gray, H. B. Inorg.
Chem. 1986,25, 2198.
80. Schafer, H.; Siepmann, R. 2. Anorg. Allg. Chem. 1968,357, 273.
81. Siepmann, R.; Schnering, H. G. v. 2.Anorg. Allg. Chem. 1968,357, 289.
82. Siepmann, R.; Schnering, H. G. v.; Schafer, H. Angew. Chem. Int. Ed. Engl. 1967,
6, 637.
83. Kepert, D. L.; Marshall, R. E.; Taylor, D. J. C. S. Dalton Trans. 1974, 506.
47
84. Ebihara, M.; Isobe, K.; Sasaki, Y.; Saito, K. Znorg. Chem. 1992,31, 1644.
85. Zietlow, T. C.; Hopkins, M. D.; Gray, H. B. J. Solid State Chem. 1985,57, 112.
86. Zietlow, T. C.; Schaefer, W. P.; Sadeghi, B.; Hua, N.; Gray, H. B. Inorg. Chem.
1986,25, 2195.
87. Mussell, R. D.; Nocera, D. G. Polyhedron 1986, 5, 47.
88. Mussell, R. D.; Nocera, D. G. J . Am. Chem. SOC.1988, 110, 2764.
89. Jackson, J . A.; Turro, C.; Newsham, M. D.; Nocera, D. G. J . Phys. Chem. 1990,
94, 4500.
90. Cotton, F. A,; Haas, T. E. Inorg. Chem. 1964, 3, 10.
91. Kettle, S. F. A. Theor. Chim. Acta 1965,3, 211.
92. Guggenberger, J.; Sleight, A. W. Inorg. Chem. 1969,8, 2041.
93. Voronovich, N. S.; Korolkov, D. V. Zh. Strukt. Khim. 1971, 13, 458.
94. Voronovich, N. S.; Korolkov, D. V. Zh. Strukt. Khim. 1971,12, 613.
95. Miiller, H. 2. Phys. Chem. (Leipzig) 1972,249, 1.
96. Wirsich, J . Theor. Chim. Acta 1974, 34, 67.
97. Cotton, F. A,; Stanley, G. G. Chem. Phys. Lett. 1978,58, 450.
98. Bursten, B. E.; Cotton, F. A,; Stanley, G. G. Isr. J. Chem. 1980, 19, 132.
99. Seifert, G.; GroBmann, G.; Miiller, H. J. Mol. Struct. 1980, 64, 93.
100. Hughbanks, T.; Hoffmann, R. J. Am. Chem. SOC.1983,105, 1150.
101. Wooley, R. G. Inorg. Chem. 1985,24, 3519.
102. Hughbanks, T. Inorg. Chem. 1986,25, 1492.
103. Robinson, L. M.; Bain, R. L.; Shriver, D. F.; Ellis, D. E. Inorg. Chem. 1995, 34,
5588.
104. Lin, Z.; Williams, I. D. Polyhedron 1996, 15, 3277.
105. Cotton, F. A. Acc. Chem. Res. 1969,2, 240.
106. Corbett, J. D. Acc. Chem. Res. 1981, 14, 239.
107. Simon, A. Angew. Chem. Znt. Ed. Engl. 1988,27, 159.
108. Hughbanks, T. Prog. Solid State Chem. 1989, 19, 329.
109. Cotton, F. A,; Hughbanks, T.; Runyan Jr., C. E.; Wojtczak, W. A. In Early Transition Metal Clusters with pi-Donor Ligands; Chisholm, M. H., Ed.; VCH Publishers, Inc.: New York, 1995.
110. Briickner, P.; Preetz, W.; Piinjer, M. 2. Anorg. Allg. Chem. 1997, 623, 8.
111. Guirauden, A,; Johannsen, I.; Batail, P.; Coulon, C. Znorg. Chem. 1993,32, 2446.
112. Weissenhorn, R. G. Z. v. Anorg. Allg. Chem. 1976,426, 159.
113. Clark, R. J. H.; Kepert, D. L.; Nyholm, R. S.; Rodley, G. A. Spectrochim. Acta
1966,22, 1697.
114. Cotton, F. A.; Wing, R. M.; Zimmerman, R. A. Znorg. Chem. 1967,6, 11.
115. Zelverte, A,; Mancour, S.; Caillet, P. Spectrochim. Acta 1986,42A, 837.
116. Mancour, S.; Potel, M.; Caillet, P. J. Mol. Struct. 1987, 162, 1.
117. Mancour, S.; Caillet, P.; Jouan, M.; Dao, N. Q. J . Ramun Spect. 1987, 18.
118. Vaughan, P. A,; Sturdivant, J . H.; Pauling, L. J . Am. Chem. SOC.1950, 72, 5477.
119. Koknat, F. W.; Parsons, J . A.; Vongvusharintra, A. Znorg. Chem. 1974,13, 1699.
120. Recheweg, 0.;
Meyer, H.-J. 2. Krist. 1996, 211, 396.
121. Ueno, F.; Simon, A. Actu Crystallogr. 1985, C41, 308.
122. McCarley, R. E.; Hughes, B. G.; Cotton, F. A,; Zimmerman, R. Znorg. Chem. 1965,
4, 1491.
123. Espenson, J . H.; McCarley, R. E. J . Am. Chem. SOC.1966, 88, 1063.
124. Espenson, J. H. Znorg. Chem. 1968,4, 631.
125. Espenson, J. H.; Boone, D. J. Inorg. Chem. 1968,4, 636.
126. Eisenbraun, R.; Schafer, H. 2.Anorg. Allg. Chem. 1985,530, 222.
48
127.
128.
129.
130.
Cooke, N. E.; Kuwana, T.; Espenson, J. Inorg. Chem. 1971,10, 1081.

Quigley, R.; Barnard, P. A.; Hussey, C. L.; Seddon, R. Inorg. Chem. 1992,31, 1255.
Hussey, C. L.; Quigley, R.; Seddon, K. R. Inorg. Chem. 1995,34, 370.
Penicaud, A.; Batail, P.; Coulon, C.; Canadell, E.; Perrin, C. Chem. Muter. 1990,
2, 123.
131. Meyer, H.-J. 2. Anorg. Allg. Chem. 1995, 621, 921.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
Reckeweg, 0.;
Meyer, H.-J. Z. Naturforsch 1995, 50b, 1377.
Reckeweg, 0.;
Meyer, H.-J. Z. Anorg. Allg. Chem. 1996,622, 411.
Klendworth, D. D.; Walton, R. A. Inorg. Chem. 1981,20, 1151.
Imoto, H.; Hayakawa, S.; Morita, N.; Saito, T. Znorg. Chem. 1990,29, 2007.
Koknat, F. W.; McCarley, R. E. Inorg. Chem. 1972,11, 812.
Kashta, A,; BrniEeviC, N.; McCarley, R. E. Polyhedron 1991, 10, 2031.
BrniEeviC, N.; Planinic, P.; BaBic, I.; McCarley, R. E.; Rutar, V.; Xie, X. Inorg.
Chem. 1993,32, 3786.
BrniEeviC, N.; MuStoviC, F.; McCarley, R. E. Inorg. Chem. 1988,27, 4532.
BrniEeviC, N.; Schafer, H. Z. Anorg. Allg. Chem. 1978,441, 219.
BrniEeviC, N.; MesariC, S.; Schafer, H. Croat. Chem. Acta 1984, 57, 529.
Hughes, B. G.; Meyer, J . L.; Fleming, P. B.; McCarley, R. E. Inorg. Chem. 1970,
9, 1343.
143. Fleming, P. B.; Dougherty, T. A,; McCarley, R. E. J. Am. Chem. Soc. 1967.89, 159.
144. Koknat, F. W.; McCarley, R. E. Inorg. Chem. 1974, 13, 295.
145. Field, R. A.; Kepert, K. L.; Robinson, B. W.; White, A. H. J. C. S. Dalton Trans.
1973, 1858.
146. Kennedy, V. 0.;
Stern, C. L.; Shriver, D. F. Inorg. Chem. 1994,33, 5967.
147. Bond, M. R.; Hughbanks, T. Inorg. Chem. 1992,31, 5015.
148. Mackay, R. A,; Schneider, R. F. Inorg. Chem. 1967,6, 549.
149. Perrin, C.; Ihmai'ne, S.; Sergent, M. New J. Chem. 1988, 12, 321.
150. Schneider, R. F.; Mackay, R. A. J. Chem. Phys. 1968,48, 843.
151. Fleming, P. B.; McCarley, R. E. Inorg. Chem. 1970,9, 1347.
152. Robbins, D. J.; Thomson, A. J. J. C. S. Dalton Trans. 1972, 2350.
153. Bateman, L. R.; Blount, J. F.; Dahl, L. F. J. Am. Chem. Soc. 1966,88, 1082.
154. Simon, A.; Schnering, H.-G.; Schafer, H. Z. Anorg. Allg. Chem. 1967,355, 295.
155. Imoto, H.; Corbett, J. D. Inorg. Chem. 1980, 19, 1241.
156. Simon, A. 2. Anorg. Allg. Chem. 1967,355, 311.
157. Fitch, A. N.; Barrett, S. A,; Fender, B. E. F.; Simon, A. J. C. S. Dalton Trans.
1984, 501.
158. Imoto, H.; Simon, A. Inorg. Chem. 1982,21, 308.
159. Rogel, F.; Corbett, J . D. J . Am. Chem. Soc. 1990, 112, 8198.
160. Chen, L.; Cotton, F. A. Inorg. Chem. 1996,35, 7364.
161. Finley, J . J.; Camley, R. E.; Vogel, E. E.; Zevin, V.; Gmelin, E. Phys. Reu. B 1981,
24, 1323.
162. Finley, J . J.; Nohl, H.; Vogel, E. E.; Imoto, H.; Camley, R. E.; Zevin, V.; Andersen,
0. K.; Simon, A. Phys. Reu. Lett. 1981,46, 1472.
163. Nohl, H.; Anderson, 0. K. Znt. Phys. Conf Ser. 1980, No. 55, 61.
264. Robinson, L. M.; Shriver, D. F. Coord. Chem. Reu. 1996,37, 119.
165. Jackson, J. A.; Newsham, M. D.; Worsham, C.; Nocera, D. G. Chem. Muter. 1996,
8, 558.
166. Golden, J. H.; Deng, H.; DiSalvo, F. J.; Frechet, J. M.; Thompson, P. M. Science
1995,268, 1463.
167. Robinson, L. M.; Lu, H.; Hupp, J. T.; Shriver, D. F. Chem. Muter. 1995, 7, 43.
168.
169.
170.
171.
172.
173.
174.
175.
176.
177.
178.
179.
180.
181.
49
Robinson, L. M. Ph. D. Thesis, Northwestern Universit,y, Evanston, 1995.

Christiano, S. P.; Pinnavaia, T. J. J . Solid State Chem. 1986, 64, 232.
Christiano, S. P.; Wang, J.; Pinnavaia, T. J. Inorg. Chem. 1985,24, 1222.
Ouahab, L. Chem. Muter. 1997,9, 1909.
Ouahab, L.; Batail, P.; Perrin, C.; Garrigou-Lagrange, C. Muter. Res. Bull. 1986,
21, 1223.
Penicaud, A,; Batail, P.; Davidson, P.; Levelut, A.-M.; Coulon, C.; Perrin, C. Chern.
Muter. 1990, 2, 117.
Keggin, J. F.; Miles, F. D. Nature 1936, 137, 577.
BrniEevic, N.; Ruzic-Toros, Z.; Kojic-Prodic, B. J . C. S . Dalton Trans. 1985, 455.
Vojnovic, M.; Antolic, S.; Kojic-Prodic, B.; Brnicevic, N.; Miljak, M.; Aviani, I.
2. Anorg. Allg. Chem. 1997, 623, 1247.
Prokopuk, N.; Shriver, D. F. Chem. Muter. 1998, 10, 10.
Deakin, M. R.; Li, T. T.; Melroy, 0. R. J. Electroanal. Chem. 1988,243, 343.
BrniEeviC, N.; Nothig-Hus, D.; Kojic-Prodic, B.; Ruzic-Toros, Z.; DaniloviC, Z.;
McCarley, R. E. Inorg. Chem. 1992,31, 3924.
BrniEevic, N.; McCarley, R. E.; Hilsenbeck, S.; Kojic-Prodic, B.; Acta Crystallogr.
1991, C47, 315.
Beck, U.; Simon, A,; Sirac, S.; BrniEeviC, N.; Z . Anorg. Allg. Chem. 1997, 623, 59.
ADVANCES IN INORGANIC CHEMISTRY, VOL.
46
RECENT ADVANCES IN NOBLE-GAS CHEMISTRY

JOHN H. HOLLOWAY and ERIC G. HOPE
Department of Chemistty, The University of Leicester, Leicester LEI 7RH, United Kingdom
I. Introduction
11. Recent Review Literature
111. The Possibility of Argon Chemistry
Krypton Chemistry
V. Xenon Chemistry
A. Core Chemistry and Recent Advances
B. Reactions of the Xenon Fluorides
VI . Radon Chemistry
References
rv.
I. Introduction
Until "XePtF," was reported in 1962 (I),the only compounds containing noble-gas elements in combination with other elements were
weakly bonded species of two main types: (i) gaseous cationic or excited-state species observed by spectroscopic means and (ii)species in
which the noble-gas atoms are adsorbed, enclathrated, or encapsulated on or within the lattices of other molecules. The former includes
mostly excited-state diatomic molecules containing like or unlike noble-gas atoms, or diatomics containing oxygen, nitrogen, or metal
atoms in combination with a noble-gas atom, a wealth of van der
Waals-bonded dimers and trimers, diatomic cationic species containing like or unlike noble gases, or noble-gas hydrides. These were
briefly reviewed in 1968 (2). The latter consists mainly of clathrates
in which the noble-gas atoms lie at the interstices of crystalline cages
formed by water, organic molecules, or double hydrates involving both
organic and water molecules [e.g., species such as R.2Ng.17H20 (R =
CH,COOH, CH2C12,CHC1, or CCll, and Ng = Ar, Kr, or Xe)l (2-41,
or where noble-gas encapsulation in zeolites occurs ( 4 ) .
Subsequent to the discovery of the xenon-platinum hexafluoride
51
Copyright c) 1999 by Academic P r a s
All rights of reproduction In a n y form reserved
0898-8838/99 525 00
52
HOLLOWAY AND HOPE
complex, simple binary fluorides of xenon (XeF,, XeF,, and XeF,),

krypton (KrF2),and radon (probably RnF2),have been prepared (2-41,
and the general chemistry of these species has been well established
and developed. Compounds isolable at ordinary temperatures have
been obtained only for the heavier noble gases, and the only stable
compounds that can be made by direct synthesis are the fluorides.
Most of the compounds that have been obtained consist of the noble
gas combined with the most electronegative elements, fluorine and
oxygen. However, compounds in which the noble gases are bonded to
nitrogen, carbon, or chlorine have been characterized, and there is a
growing and interesting chemistry evolving around some of these species. The chemistry of xenon is undoubtedly the most extensive, xenon
occurring in oxidation states +2 to +8, and stable species have been
isolated for each of these. For krypton and radon, the compounds
found so far are all in oxidation state +2 and, until quite recently, all
of the known species involved only krypton-to-fluorine or radon-tofluorine bonds. The most sensitive detection methods have provided
no definitive evidence for a stable argon compound. The inert status
for helium and neon remains unchallenged.
The chemistry of xenon stems from the reactions of the binary fluorides and falls into five main categories:
1. reactions in which the xenon fluoride acts as a fluorinating
agent,
2. hydrolysis reactions yielding oxide fluorides, oxides, and oxygencontaining salts,
3. fluoridelanion metathesis reactions between the fluoride and an
anhydrous acid,
4. reactions with fluoride ion acceptors, yielding fluorocations, and
5. reactions with fluoride ion donors, yielding fluoroanions.
Krypton chemistry parallels that of xenon but is, in general, more
limited. The bonds between krypton and other elements, even the
highly electronegative fluorine and oxygen atoms, are weaker than
those with xenon. The outcome is that although krypton-fluoride cationic species are known, Kr-0, Kr-N, and Kr-C bonds are difficult
to obtain and, where they exist, are unstable. On the other hand, the
weakness of the Kr-F bond, both in neutral KrF, and in the cations
[KrFI' and [Kr,F,I+, provides a means of carrying out useful low-temperature fluorinations, which have been exploited in producing novel
species that are often inaccessible using elemental fluorine.
Radon apparently forms a difluoride and some cationic complexes.
However, the evidence is based exclusively on radiochemical tracer
53
experiments because radon has no stable isotopes, and the work is

made even more difficult by radons intense radioactivity, which not
only makes it hazardous to manipulate but also leads to decomposition of the reagents involved.
A wealth of reviews covering various aspects of noble-gas chemistry
were written up to the end of 1979 and some of those written in English are listed in reference ( 5 ) .In addition to the books ( 2 ,3 ) and an
extensive and authoritative book chapter ( 4 )already mentioned, three
other books appeared in the early days of noble-gas chemistry,
and these are listed in reference ( 6 ) .There is also a valuable and authoritative bibliography covering the literature in the period 1962 to
1977 (7).
Because the main features of noble-gas chemistry are now well documented to the extent that they appear in all undergraduate inorganic textbooks, this review concerns itself only with the new chemistry that has evolved since 1979, and an effort has been made to pick
out the main features of this work and set it within the context of the
chemistry established up until that time.
II. Recent Review Literature
Since 1979, there have been a number of reviews of note. Among

them is an interesting summary of the history of the discovery of noble-gas chemistry in which the reasons for the failure of Yost and
Kaye t o obtain noble-gas compounds in the 1930s are examined. The
article also speculates whether there were others who envisioned noble-gas reactivity and discusses work that took place before Bartletts
key experiments (8).Other general summaries of noble-gas chemistry
were published in celebration of the 25th anniversary of the discovery
of chemical activity of xenon, a t the 100th anniversary of the discovery of fluorine (9) and very recently in a new Encyclopaedia of Inorganic Chemistry (10).A review covering aspects of the chemistry of
xenon, and especially its reactions with pentafluoro-oxoselenate and
pentafluoro-oxotellurate groups, was published a t the beginning of
the review period ( l l ) ,and a detailed and comprehensive academic
review of cationic and anionic complexes of the noble gases was published in 1984 (12).An important development in noble-gas chemistry
over the last 15 years has been the synthesis of compounds containing
stable bonds between noble gases and elements other than fluorine
and oxygen. This was foreseen in an article in 1982, which traced
54
HOLLOWAY AND HOPE
then-recent developments and discussed Xe-N bonding as well as the

fragmentary evidence for Xe-C and Xe-B bonds (13).
An authoritative review on the chemistry of radon, which includes
the formation of clathrate compounds, the simple and complex fluorides, ionic radon in solution, redox properties, and the various unsuccessful efforts to prepare an oxide and halides of radon other than the
fluoride, was published by the most significant contributor to this field
in 1983 (14).
A review article on van der Waals complexes of metal-atom noblegas complexes was published in 1995 (15).
111. The Possibility of Argon Chemistry
The existence of krypton and xenon compounds inevitably raises

the question as to whether noble-gas chemistry might be extended to
argon. The increasing range of krypton compounds that have been
synthesized over the last 10 years, and in particular the recent observation of Kr-0 and Kr-N species, will undoubtedly further stimulate
interest in this area.
Although some years ago ab initio calculations on [HeFI' and
[NeF]+ indicated that the ground states of these molecules are unstable, those for [ArF]' implied a stable ground state and suggested
that a likely complex to be synthesized would be [ArFI+[PtF,I- (16).
Evidence from photoelectron spectroscopy did indeed produce evidence of [ArFI+ and provided a measure of its dissociation energy
{D,([ArFI+)2 1.655 eV) but confirmed instability for the helium and
neon analogs (17). More recently new ab initio calculations at the
MP4C3DTQY6-311G(2df, 2pd)//MP2/6-31G(d,p) level predict that the
best counterions for the stabilization of [ArF]+appear to be [AuFJ or
[SbFJ (18). High-level calculations also predict that [HC=N-ArFl+
should be stable with a binding energy of 160 kJ.mol-' which is
closely similar to that of the recently observed krypton analog [HC=
N-KrFl+ (157 k J . mol-') (19).Interesting work, in which electronegativity has been redefined and adjustments to the well-established
electronegativity scales have been made, has demonstrated close
agreement between the four scales (Allred and Rochow, Sanderson,
Mulliken, and Pauling) and that the values for argon and krypton are
similar (20).All of this suggests that more effort to prepare an argon
fluoride complex might be worthwhile.
55
IV. Krypton Chemistry
The chemistry of krypton is less extensive than that of xenon, and

the well-established characteristics, summarized in detail in the book
chapters in 1973 by Bartlett and Sladky ( 4 ) and in 1984 by Selig and
Holloway (12) are dealt with in outline here, but the new developments since 1979 are referenced in detail.
The only stable binary fluoride of krypton is the difluoride, KrF2,
and all the known chemistry derives from this molecule. Early reports
of the preparation of KrF4have not been substantiated, and no simple
stable oxides or oxide fluorides have been isolated. However, small
amounts of the violet free radical, KrF, have been observed following
y-irradiation of KrFz (2-4). There has also now been direct observation of [KrO,]' ( n = 1,2) and [&OH]' by Fourier-transform ion cyclotron resonance (FT-ICR) mass spectrometry, for which ab initio theoretical calculations suggest that the two species are covalently bonded
with bond dissociation enthalpies of 237.7 and 169.5 k J . mol-' respectively (211.
The volatile, colorless crystalline difluoride was prepared initially
by subjecting mixtures of krypton and fluorine at low temperature
and pressure to an electric discharge or by irradiation of liquid
krypton/fluorine mixtures with UV light (22). The latter is a more
rapid and convenient way of preparing KrFz than the earlier discharge method, and it has since been found that laser as well as UV
photolysis is also effective (23). These photolytic methods are particularly useful for preparing small quantities (22,231. Because molecular
fluorine is not significantly dissociated until 500-600C, the preparation of thermally unstable compounds such as KrFz by a thermal
method did not seem to be an option. However, in 1976, Soviet workers reported the preparation of KrFz using catalytic-thermal generation of atomic fluorine from molecular fluorine (24). The nonequilibrium conditions necessary were created by generating atomic fluorine
on the surface of a nickel wire at 600-700C and allowing the atomic
fluorine to react with krypton in a low-temperature zone close to the
reactor wall a t -196C. The rate of production is high (-6 g/h) compared with other preparative methods. The reaction method and operating conditions of another high-yield, thermal catalytic method has
also recently been discussed (25).
Krypton difluoride has now been shown to exist in two crystallographic forms, a low-temperature ( - 196C) phase, referred to as
a-KrFz, and a high-temperature (-78C) phase, /3-KrF2(23). Study of
the vibrational spectra has shown that the X-ray structure reported
56
HOLLOWAY AND HOPE
in 1972 (261, which contains two linear KrFz molecules per tetragonal
unit cell (space group P4,lmnrn) aligned in planes perpendicular to
the tetrad axis, is that of P-KrF,. Raman data suggests that the
a-KrFpcrystallizes in the same space group as XeF, (271, but a definitive conclusion must await an X-ray structural investigation of the
krypton species at low temperature.
The adducts of KrF, are analogous to those of XeF, and until quite
recently were limited to cationic derivatives, [KrFl', [KrFl' .xKrF2,
or [Kr2F3]+.
KrFz, of Lewis acid pentafluorides (12,281 or the
[KrpF31+
related but somewhat less powerfully Lewis acidic metal oxide tetrafluorides (12,28,29). More recent reports have included the synthesis
of [KrF]'[MF,J
(M = As, Sb) in anhydrous HF (30), the preparation
of [K~F]'[AS~F~~]for the first time, and the observation of two solidstate forms of [KrFl+[SbpFJ (23), providing only the second example
of polymorphism in krypton complex chemistry {cf. [KrFl+[AsF61(31I}. The cationic complex of krypton difluoride, [Kr,F31+.xKrFz.
[BF41-,incorporating the [BF4]-anion, has been prepared for the first
time from the reaction of KrFz with BF3at -40C in HF. The presence
of HF to help bring about the ionization was shown to be essential in
this preparation because the complex is not produced a t temperatures
as low as -78C when KrFz and BF3 are brought directly together
(30). The use of anhydrous HF has also provided a means of synthesizing two new metal fluoride adducts, 2KrFzsMnF4 and KrFz.MnF4,
from the combination of KrFz and MnF, in HF solvent (32). The 2 : 1
adduct decomposes in dynamic vacuum at -45"C, yielding KrFz and
the 1: 1 adduct, which itself is stable up to -25C. Because neither
species provided a Raman spectrum, their exact nature is uncertain,
but it Seems likely that they are molecular adducts rather than complexes with cationidanionic character.
In general, the component molecules of all of the cationidanionic
complexes are linked together by fluorine bridges, which have a considerable degree of ionic character. However, Raman spectra of the
complexes derived from the metal oxide tetrafluorides indicate that
they are best formulated as essentially covalent structures (12). The
spectra of the [KrpF31'cation correlate well with those associated with
the analogous xenon species, but additional peaks, which have no
equivalents in the xenon spectra, indicate that, unlike [XezF3It,the
krypton cation is unsymmetrical (Fig. 1)with one short and strong
bond, %,-Fa, a weaker and longer bond, Krb-F,, and two bridging
bonds, Krh----Fhand Kr,----Fb,which are also of different lengths and
strengths (31). Alternatively, the cation may be regarded as a dis-
57
FIG.1. The structure of IKr2F1]'
torted KrFz molecule fluorine bridged to a [KrFI' cation,

[KrFz. KrF]' (33).
As in the case of xenon, one of the most interesting developments
in krypton chemistry has been the identification of bonds between
noble-gas atoms and elements other than fluorine. The first example
of such a species containing a Kr-0 bond, Kr(OTeF5In,has been prepared by the reaction of krypton difluoride with B(OTeF& a t -90 to
-112C in S02C1F solvent (341, the thermally unstable product decomposing to species analogous to those found in the decomposition
of Xe(OTeF&. The high solubility of B(OTeF& relative t o KrFz in
SOzCIF at the low temperatures required for the stabilization of
Kr(OTeF& is thought to serve to maintain B(OTeF& in excess and to
prevent the formation of FKr(OTeF5),which was not observed ( 3 4 ) .
Perhaps even more remarkable is the extension of krypton chemistry to include Kr-N-bonded species in the cations [HCrN-KrF]'
( 3 5 )and [R,C=N-KrFl+ (Rf = CF:<,C2F5,n-C,F,) (36).The first to be
prepared, the fluoro(hydrocyano)krypton(II) cation, [HC=N-KrFl',
was obtained from the low-temperature reaction of KrFy with
[HC~NH]+[AsFJin H F or BrF5 solvent rather than by direct combination of [KrFl+[AsF,]- with HC=N in H F or BrF5 (which was the
route used in the preparation of the xenon analog) because, not only
is [KrF]' strongly oxidizing toward HC=N and BrF5, it also tends to
undergo autoredox reactions in both solvents (35).Warming of [HC=
N-Kr-F]+[AsF,l~ in BrF5 solution to -58 to -55C for several hours
produced little evidence of decomposition, but brief warmings above
-50C led to rapid formation of [NF,]', CF4, and CF3H. However,
when the solid complex is warmed above -5O"C, krypton, NF3, and
CF, gases are rapidly evolved. Extension of the HCN work to include
other nitrogen bases that are oxidatively resistant and have first adiabatic ionization potentials close to or exceeding that of [KrFI' has
led to the preparation of three other examples of cationic Kr-Nbonded species. Derived from perfluoroalkyl nitriles, these species,
[R&=N-KrFI+[AsFJ
(Rf = CF3, C2Fs,n-C3F7),have been prepared
by reaction of R&=N/AsF5 mixtures with KrF2 in BrF5 a t tempera-
58
HOLLOWAY AND HOPE
tures between -57 and -61"C, and have been compared with their
xenon analogs, which have also been made (36).
Inevitably, the discovery of the krypton-nitrogen-bonded species
has led to the testing of a number of theoretical models. Nonrelativistic quantum chemical calculations, including electron correlation effects, have been done for the ground-state [HC=N--KrF]' cation, and
the computed geometrical structure, stability toward dissociation, and
harmonic vibrational spectrum agree closely with the experimental
data (37, 38). The calculations suggest a Kr-F bond stronger than
that in &Fa. The fact that the xenon analog was also made prompted
a comparison of the experimental properties of the bonds formed between nitrogen and fluorine to the noble gas atoms, which suggested
that the exceptional ability of the [NgFI+(Ng = Kr, Xe) ions to act as
Lewis acids is related to the presence of holes in the valence shell
charge concentrations of the krypton and xenon atoms, which expose
their cores. The study also provided reason to believe that the mechanism of formation of the Ng-N bonds in the adducts is similar to that
in the formation of hydrogen bonds (39).
The ability of noble-gas fluorides to bring about oxidation and/or
fluorination is in the order XeFz = covalent Xe(I1) derivatives <
XeF4 < XeFs < &Fa. Also, it is well known that the cationic derivatives, [XeFl.', [XezF31+,[XeF31t, [XeFJ+, [Xe2F1J+,and [KrFI' or
[Kr2F3]+have a higher fluorinating ability than their neutral parent
compounds (13).Clearly, from this comparison, it is evident that KrFz
and its cationic derivatives are the most powerful noble-gas fluorinating agents. Indeed, thermodynamic data indicate that reactions involving KrFz are about 50.2 kJ.mo1-I more exothermic than those
with elemental fluorine. This has two consequences: the first is that
such reactions have t o be undertaken with great care; the second is
that these reagents offer the potential to synthesize other novel highoxidation-state species (13). This important characteristic has been
exploited for some time. For example, KrFz is known to fluorinate
xenon to XeFs and iodine to IF7(331, whereas the [KrF]' cation reacts
spontaneously with oxygen to yield [O,]' and xenon to give [XeFl'
(31).The cations, [KrFl' and [Kr2F31+,
oxidize BrF, to IBrFJ' [40,411.
The fact that [KrFl' takes BrF5 to [BrF,]' but fails to take XeOF, to
[XeOFJ' (42) has been explained in terms of the [KrFI' attacking
BrF5 at the bromine because of the nucleophilic attraction of its nonbonding lone pair whereas the most nucleophilic part of the XeOF4
molecule is its oxygen atom. Thus, an intermediate XeF4-OF might
be expected (43). This is given further credence by the observation
of a yellow explosive intermediate in reactions of [KrF]' with
59
XeOF, (44). It is also noteworthy that XeOF, is not fluorinated by

PtFG (45).
The successful oxidation of BrF5 to [BrF,] led to investigations on
the oxidative fluorination of metals and metal fluorides. The first example was the successful oxidation of elemental gold to gold(V) (42),
the treatment of Au powder in anhydrous HF giving a pale yellow
solid, [KrFl+[AuFGI-,which is sparingly soluble in H F to give a pale
yellow solution:
7KrFz + 2Au
20C
aHF
2[KrFI+[AuFJ
+ 5Kr.
(1)
Pyrolysis of the krypton complex at between 60 and 65C gave gold

pentafluoride, AuF5:
[KrF]+[AuFJ
60-65C
8h
A u F ~+ Kr
+ F,.
(2)
Since then, the oxidative fluorinating abilities of KrF, have been

used to prepare CoF3 (46), MnF4 (321, and the lanthanum tetrafluorides, LnF4 (Ln = Ce, Pr, or Tb) (471, and their heptafluoro complexes,
Cs3LnF7(Ln = Ce, Pr, Nd, Tb, Dy, or Tm) (47, 48). Cobalt trifluoride
and the lanthanum compounds can also be prepared using xenon fluorides, except for the thermally unstable PrF4,which can only be obtained using KrF, (47). Further careful investigation of the reaction of KrFz with LnOz (Ln = Ce, Tb, or Pr) in 2 : 1 or 3 : 1 ratios has
shown that, here again, LnF, is produced, but reaction at a 1 : 1 ratio
gives a product with a composition close to that of LnOFz that
has been shown not t o be a mixture of Ln02 and LnF4 (49).Another
species in oxidation state four that has been prepared is the hexafluoronickelate(1V) anion, iF612-, as the [Xe2F,J$[NiF61z-salt, by
the reaction of NiFz, XeF6, and KrF, in anhydrous HF (50). Reaction of KrF, with [XeF51+[AuF,I-in anhydrous H F below 273 K gives
[XeF5]+[AuF6]-(51).
A number of high-oxidation-state fluorides and oxide fluorides have
also been prepared using KrFz. For example, Cr02F2can be converted
t o CrOF4 (29) [rather than CrOF3 which is the product of reaction
with XeF, (5211 and has been shown to form a stable 1 : 1 adduct with
KrF2 which has a n essentially covalent structure with a Kr----F----Cr
bridge (Fig. 2). A notable achievement has been the synthesis of AgF3
(53)by precipitation from anhydrous HF solutions of [AgF41- salts by
addition of fluoro acids such as BF,, PF,, or AsF5. It was also shown
60
HOLLOWAY AND HOPE
Kr
F'
FIG.2. The structure of CrOF4. KrFz.
that previous claims were in fact due to Ag(II)Ag(III)F, (53).There

has also been considerable interest in the derivation of high-oxidation-state fluorides from a number of the actinide oxides or fluorides.
Reactions of the oxides and tetrafluorides of uranium, neptunium,
and americium, and the oxfluorides of neptunium, with KrF, in BrF5
or HF were reported in the mid 1970s and early 1980s (53-55), and
brief mention has been made of the preparation of UF6 from UF4 (56),
the fluorination of NpOF4 to NpFs (571, and the oxidation of PuF3 to
PuF, (58) using KrF, at room temperature or lower. More recently
(591, it has been shown that not only uranium and neptunium oxides
and low-valent fluorides but also plutonium tetrafluoride can be converted to their hexafluorides at ambient temperature using KrF2vapor or KrF, in anhydrous HF. However, efforts to prepare AmFs from
AmO, in anhydrous HF failed, in contrast to the earlier report (55).
The highest oxidation-state species synthesized recently using KrF,
in anhydrous HF have been TcOF,, made by reaction with Tc02F3
(60), and the new osmium(VII1) oxide fluoride, cis-OsF402,prepared
by the fluorination of O s 0 4 (61).The latter was previously erroneously
characterized as OsF60 (62).
Another interesting aspect of the oxidative fluorinating ability of
krypton difluoride is its ability to produce coordinatively saturated
fluorocations. By 1974 three such species were known, namely,
[NF41+,[ClFsl', and [BrF,]', the last being produced using [KrF]' or
[Kr,F,I+ (41, 42). It was later shown that the [NF4]+salts could be
prepared using the [KrFl' cation, KrF, reacting with NF3 in the presence of the Lewis acids SbF,, NbF5, PtF5, TiF4, and BF, to yield
[NF41'[SbFsl-, [NF41'[NbF61-, [NF4]'[PtF6]-, [NF,I,"TiFsI2-, and
[NF41+[BF41respectively (63).In more recent studies, the syntheses
of [NF,I'[SbF,I-, [NF41'[AsF61- (301,and [NF41+[BFJ (25, 30) have
been studied under different conditions. The fluorination of NF, by
[KrF]'[SbFsl- in anhydrous HF has been shown to proceed quantitatively at temperatures as low as -31"C, indicating an ionic two-electron mechanism. It was shown that CIFBand BrF5 could also be con-
61
verted to their complex fluorocations by [KrFl'. In contrast, PtFs was

shown to be capable of oxidizing only NFB and ClF, but not BrF5.
Efforts to fluorinate OF2, CF3NF2, and [C1F401- with [KrFI' salts,
however, were not successful (30).
An unusual reaction of KrF, has been the observation that KrF,
decomposes on zeolites at 233 K, fluorinating the zeolite and encapsulating the krypton released. A proportion of the latter is retained even
upon heating to 773 K, demonstrating the potential for R5Krrecovery
using KrF,/zeolite systems (64).As might be expected, the fluorination of C,,-fullerene by KrF, is extreme, yielding species ranging from
C,,F,, to the hyperfluorinated species C6OF78, as determined by mass
spectrometry. An interesting observation in these experiments, however, is that in comparison with other fluorination methods, the concentration of oxygen-containing derivatives is low (65).
V. Xenon Chemistry
A. CORECHEMISTRY
AND RECENT
ADVANCES
With the exception of XeF, which is obtained as a n unstable free
radical (2-4), there is no evidence for the occurrence of xenon compounds in odd-numbered oxidation states. The only species that can
be synthesized directly from the elements are the three fluorides
XeF:,, XeF, , and XeFs . Early reports of the preparation of XeFHhave
not been substantiated.
There are a variety of methods for the preparation of XeF2 (2-4),
and recent additions include oxidation of xenon using the blue solutions of AgF, in H F (66)and a new thermal catalytic synthesis (25).
The latter is related to that recently devised for the preparation of
gram quantities of KrF2( 2 4 )and has the potential to become one of the
better methods for producing XeF,. However, this route in the hands
of other workers appears to yield rather pure XeFs (671, so the conditions need further study and appraisal. In the meantime, the best
methods to make the difluoride are by heating an excess of xenon
with fluorine (2-4) or by irradiating Xe/F2 mixtures with ultraviolet
light or sunlight (2-4, 68). Xenon tetrafluoride is the most difficult
xenon fluoride to prepare because, even under optimum conditions
(heating xenon and fluorine under pressure in a 1: 5 ratio), equilibrium concentrations of XeF2 and XeF, occur (2-4). The claim that
XeF,, free from contamination by XeFz or XeF6, can be prepared by
irradiation of a gaseous mixture of xenon and fluorine (molar ratio
62
HOLLOWAY AND HOPE
2 1:2) in a reactor with walls coated with NiF2 (69)is probably not
valid. However, the thermal combination of xenon and fluorine in a
1:20 ratio of 350C for two days followed by one day at 50C in the
presence of sodium fluoride yields mainly the involatile complex,
Naz[XeF8][Eq. (311 but with traces of XeFz and XeF, . After removal of
the volatile mixture of XeF2 and XeF4, the Na2[XeF81is decomposed
a t 350C to give XeF4 [Eq. (4)l (70):
Xe + 3F2+ 2NaF
350C (2 d)
50C (1 d)
Naz[XeFsl
350C
Naz[XeF81
XeF,
+ F2+ 2NaF.
(3)
(4)
Another route to high-purity XeF4that has been discovered recently

involves a modification of the thermal catalytic method (24)in which
the fluorinating agent is 02Fzrather than Fz (67).
The best preparative route to XeFs remains the thermal combination of xenon and fluorine in a 1:20 ratio at high pressure (2-4).
However, the thermal catalytic method using elemental fluorine also
seems to yield a pure product, albeit in a lower yield (67).
All of the xenon fluorides are colorless solids except the hexafluoride, which is yellow in the liquid and gaseous states. The di- and
tetrafluorides have low volatility and the expected linear (Dmh)and
square planar (D.,h) structures respectively in both the solid and vapor
states. The bonding in these structures has been accounted for using
a three-center, four-electron bond model similar to that employed to
explain the structure of C02.Although the structures and bonding for
XeFz and XeF4 have been widely accepted for many years, physicochemical studies continue t o be reported and refinements made. For
XeF2,new Raman spectra in the solid state and vapor phase at 325 K
have been reported, and force constants have been evaluated using a
SVFF approximation (71).Raman study of the molecular dynamics of
XeF2 in acetonitrile (721, hydrogen fluoride, and bromine pentafluoride (73) (which are the most widely used solvents in noble-gas fluoride
chemistry) at various concentrations has been carried out, and vibrational and rotational correlation functions, as well as the characteristic times, have been calculated.
The structure of XeFs in the solid, liquid and vapor states continues
to be discussed in detail. Out of all the known hexafluorides, it has
the highest boiling point, suggesting that, unlike other hexafluorides,
it might be polymeric in the liquid and solid states. Some of the uncertainty has arisen because of difficulties in manipulating the com-
63
pound. Its powerful fluorinating ability leads to it damaging container

and window materials and producing lower-xenon fluorides, which
contaminate the specimen. Another difficulty is its ready hydrolysis
to the dangerously explosive solid, Xe03. In the vapor state, it has
been clear for some time that an octahedral structure for this molecule could not have the lowest energy (74-76). The VSEPR model (74,
75) predicted that an extra lone pair should be stereochemically active
and distort the molecule to lower symmetry. Experimental and theoretical work has demonstrated that this is indeed the case, with the
molecule having local minima giving CBvand C2, geometries via a
pseudorotation mechanism in which the lone pair moves over the octahedral faces and edges (77, 78). Theoretical work in 1987 provided
additional information concerning Oh, C2", and Csv structures (791,
and more recent experimental work, using high-resolution photoelectron spectroscopy to resolve the ligand field splittings, yielded bond
angles for the CSvstructure that agreed with the theoretical work (80).
These studies also confirmed earlier work that indicated that the dipole moment is very small. However, both pieces of work referred to
the Cpvstructure as a local minimum, and no theoretical vibrational
frequency analysis was presented to support this. Now, the Oh, Czv,
and C3, structures have been studied with a significantly larger basis
set, and harmonic vibrational frequencies for each have been determined and analyzed (81). At the SCF level of theory, the Oh and Cpv
structures have been shown not to be minima but to be transition
states whose imaginary vibrational frequencies lead to the C13"
structure.
Nothing new has emerged recently about the structure of XeF, in
the liquid state and in solution, and the current state of knowledge
was well reviewed by Seppelt and Lentz in 1982 (13). In both the
molten state and in solution, monomeric XeFs is in equilibrium with
tetrameric ([XeF5]+F-)4
but at low temperatures the monomer disappears. The 25 fluorines of the tetramer are involved in a complex
scrambling mechanism, which is perhaps related to that in Rh,(CO),, ,
although it is clear that no xenon-xenon bonds equivalent to the
metal-metal bonds in the rhodium compound occur. In the solid
state, at least four phases exist, of which, because of lack of X-ray
intensities, absorption problems, and disorder, only the cubic structure, which contains 144 XeF6 molecules in the unit cell arranged
as 24 tetramers and 8 hexamers, has been determined. In each, the
fluorines attached to each xenon form square pyramidal configurations, with the remaining fluorine atoms contributing to a twofold
or threefold bridge linking adjacent xenons. Because the bridging
64
HOLLOWAY AND HOPE
F- ions are not close to the fourfold axis of the [XeF51+groups, it

has been suggested that lone pairs of electrons may occupy this position.
Thermochemical data on the xenon fluorides continues to accumulate. New melting and boiling points of XeF, at 136 2 5C and 338C
respectively, and an enthalpy of fusion of 14.04 2 2.09 kJ.mol-l, have
been obtained, and disproportion of XeFz to XeF, and Xe has been
observed above 350C (82). A new determination of the heat of formation of XeF, , obtained by the combustion of germanium in the difluoride, is in close agreement with an earlier value obtained by reaction
of XeFz with PF3(83). Existing thermodynamic data have been used
to calculate successive one-electron potentials for the reduction of
XeFa (84).
From the beginning of noble-gas chemistry there has been continued interest in theoretical and spectroscopic aspects. The more recent
include an ESCA study of XeF, to high statistics in which the satellites observed on the low-energy side of the xenon 3d and 4d primary
photolines have been attributed to monopole transitions from the occupied MOs of XeFz to the antibonding and/or Rydberg-state orbitals
(85). The nature of the satellite lines has been identified using theoretical SCF-Xa-SW calculations. This has been followed up by X-ray
photoelectron spectroscopy and theoretical studies on XeF, and XeFs
in which the Xe 4d and Xe 3d spectra have been reported. However, the main center of focus in this paper was on KrFz (86). Other
theoretical work that has been carried out includes SCF-MO studies
on XeFz and XeF, (87) and the application of the multiple-scattered-wave-SCF-X, method to estimate ionization potentials and
charge distribution in XeF,, XeF4, XeFs, and XeC1,. The role of 5d
orbitals, especially their importance in XeF, and XeFs binding, has
been discussed (88).
Many attempts have been made to understand theoretically the
electronic structures of XeF, ( n = 1, 2, 4, 6) species, the majority having been made since the mid-1970s (79, 81, 89-93). Generally, these
have utilized the conventional Hartree-Fock molecular-orbital approach and in some cases correlation effects have been included (81,
89). Relativistic effects have been ignored despite the fact that, due
to the high nuclear charge of xenon, they should be high. The spectra
have been studied by different techniques (e.g., 75, 80, 94). Now, allelectron Dirac-Fock and Dirac-Fock-Breit calculations for the series
have been caried out, and the role of relativistic effects on bond
lengths and dissociation energy has been reported (95).
65
B. REACTIONS
OF THE XENON
FLUORIDES
With the exception of the reactions of xenon with the powerful fluorinating agent PtF, and its relatives, all of the chemistry of xenon
has been derived from reactions of the binary fluorides.
1. Hydrolysis Reactions Yielding Oxide Fluorides, Oxides, and
Oxygen-Containing Salts
All of the known oxides, oxide fluorides and a number of oxygencontaining salts stem from hydrolysis reactions of XeF, or XeF,. It
has been well established for some time that no stable oxo-species can
be obtained from the difluoride, XeF2. Stable in acid or neutral solution, it decomposes instantaneously in base with liberation of xenon,
oxygen, and hydrogen fluoride. The tetrafluoride, on the other hand,
is instantly hydrolyzed by water and, depending on the conditions, up
to one third of the xenon may be retained in solution. There is also
evidence that some XeF2 may be liberated. There has been little recent work in this area, but the most recent (96)suggests that the
initial product may be XeOFz and that this decomposes to give XeFz
if the water supply is limited. In the presence of larger amounts of
water, XeOz and XeO may be produced and in turn can produce Xe03.
The reaction of XeF, with water is difficult to control and can result
in explosions. However, the reaction can be controlled by passing dry
nitrogen over crystalline XeF, a t room temperature to sweep the vapor into water, where the following reaction occurs:
XeFs + 3H20+ XeOB+ 6HF.
(5)
Hence, the aqueous hydrolysis of both XeF, and XeFs results in the
formation of the trigonal pyramidal trioxide, Xe09,in solution. Aqueous solutions of the trioxide, known as xenic acid, are fairly stable
and can be used as powerful but kinetically slow oxidizing agents.
These might have found wide application were it not for the fact that
solid XeOB,which is readily obtained from these solutions on evaporation, is a violent and sensitive explosive.
The addition of alkali to XeO, solutions yields xenate ions
[HXeOJ. Although salts of this species can be isolated, solutions
containing [HXe041- disproportionate to give the perxenate ion,
[XeO6I4-,and xenon:
2[HXe041- + 2[OH]- .--, LXe0614-+ Xe
+ O2 + H20.
(6)
66
HOLLOWAY AND HOPE
Within the time frame of this review, an improved synthesis of perxenate solutions has been devised in which XeFs is dissolved in excess of HOPOFz under carefully controlled conditions, the excess of
HOPOFz, HF, and OPF3 is removed, and the Xe03 remaining is dissolved slowly in 4 M NaOH at 0C (90). Both reactions have been
carried out on the 6- to 7-mmol scale without explosions occurring.
The addition of solid BazXe06 to cold concentrated sulfuric acid
yields a second xenon oxide, XeO,, as an unstable and explosive gas.
Not surprisingly, this has been little studied, but infrared spectroscopy and electron diffraction have shown that it has the expected tetrahedral geometry.
The controlled hydrolysis of XeFs with water in a 1:1 molar ratio
produces the colorless, stable, volatile liquid XeOF4, which has a
square pyramidal structure:
XeF,
+ HzO+ XeOF4 + 2HF.
(7)
Efforts to produce XeOzFzby the addition of a further stoichiometric

quantity of water have resulted only in the production of XeO,, and
no evidence for a dioxide difluoride intermediate has ever been obtained.
Of the known oxide fluorides, XeOF,, XeOF4, XeOzF2,and XeO3FZ
(2-4, 96), the most stable is XeOF4.The monoxide difluoride, XeOF,,
mentioned earlier, is unstable but can be obtained from the reaction
of XeF, with water at -80C. The dioxide difluoride is obtained by
mixing XeO, with XeOF4 at dry-ice temperature and fractionally distilled from the resulting mixture of XeOzF2,XeOF4, and XeFz. Although thermodynamically unstable with respect to XeF, and Oz, it is
sufficiently stable to be held in well-fluorinated Kel-F containers for
several days. The trioxide difluoride can be obtained by reaction of
XeFs with either Xe04 or Na4[Xe06]but is contaminated with other
xenon species, principally XeOF4 (3, 4, 96). The reactions of xenon
fluorides and oxide species have been systematized in order of descending acidity, XeFs > XeOZF4> XeO3FZ> XeO, > XeOF4 > XeF, >
XeOzFz> XeO, = XeFz (97).
The most recent advances in this area include new syntheses of
XeOF4 by the reaction of NaN03 (98) [Eq. (311 or POF3 (99) with a
very slight excess of XeFs. In the case of the former, an excess of
NaN03 has to be avoided to suppress the secondary reaction (98)[Eq.
(9)1, whereas in the latter, excess of POF, may lead to generation of
explosive XeO, (99):
+ FNO,
NaN03 + XeOF, + NaF + XeOzF2+ FNO, .
NaN03 + XeFG+ NaF + XeOF,
67
(8)
(9)
Reaction of CsN03 with a large excess of XeF6 at temperatures above

the melting point of XeF, proceeds quantitatively according to Eqs.
(10) and (11)to yield XeOF, (100). However, if in Eq. (10) the excess
of XeF, is reduced to less than loo%, some of the CsF can react with
XeOF, to yield Cs+[XeOF,I-. For this and other reasons, substitution
of CsN03 by NaN03 is advantageous (98,100). The reaction of CsN03
with a n excess of XeOF, yields FNOz and XeOzFzin quantitative yield
with CsF and CsXeOF6 as by-products. Provided a n excess of XeOF,
is used, a better new and safe synthesis for Xe02F2is the reaction of
XeOF, with N205(100):
CsN03 + XeF, -+ CsF + XeOF,
CsF + XeF, -+ CsXeF;..
+ FNO,
( 10)
(11)
New chemistry of XeOF4 and XeOzF2includes the preparation of a

series of novel pentafluoro-orthotellurate derivatives (1011. Reaction
of Xe02Fzwith B(OTeF& according to Eq. (12) has yielded the previously unknown compound OzXe(OTeF5),:
However, with excess of Xe02F2,a mixture of O,Xe(OTeF&, 0,XeF

(OTeF,) and XeOzFzwas produced.
There has been very little new information on xenic and perxenic
acids and their salts since the late 1960s. More recent information
includes study of the spectra and protolytic properties of aqueous perxenic acid in the pH range -0.2 to 13, which yielded pK values at
25C as follows:
The enthalpy of dissociation of [H&!O6]- at 25C is 0 2 5 kJ.mol-'

and that of [H,xeo612- is 3 % 2 kJ.mol-1 (102). The participation of
5d orbitals in the bonding of high-valent oxygen compounds, including
[HXeOJ and cXe0614-,has also been investigated by examination of
the Kl and I(pl X-ray emission lines in these ions (88).
68
HOLLOWAY AND HOPE
2. Reactions with Fluoride-Ion Donors and Acceptors
A major part of the chemistry of xenon fluorides and oxide fluorides

is associated with their reactions with fluoride donors and acceptors
to produce a variety of cations and anions. The known species are
summarized in Scheme 1. In addition, there are a number of more
unusual entities, notably the dixenon cation [Xezl+and [XeO3X1- (X =
F, C1, or Br). The preparations, structures, and chemistry of these
species was comprehensively reviewed in 1984 (121, so only developments since 1979 are covered in this present account.
It is important to remember that the formulation of many of these
species as salts is an oversimplification. In general, most are essentially fluorine bridged but with a high degree of ionic character. Also,
there are a number of adducts, such as XeFz.XeF4 and XeF2.1F5,
whose overall formulas are suggestive of donor-acceptor species,
which are simple molecular adducts.
a. Xenon Cations The transient green coloration observed in the
early days of noble-gas chemistry was identified in 1978 as being due
to the dixenon cation (103, 104). The cation can be produced either
by oxidation of gaseous xenon using [O,]' or by reduction of [XeFI'
with water or other reducing agents such as lead or mercury. The
green species is stable in SbF, indefinitely at room temperature under
an atmosphere of xenon gas (103-105). Now, almost 20 years after
the initial characterization of [Xez]+(103),a structural characterizahas shown that the Xe-Xe bond is weak and
tion of [Xezl+[Sb4FzJlonger, at 3.087(1) A, than any other main-group element bond observed so far (Fig. 3). It has also been shown that the presence of HF
in the SbF5 is essential (i.e., no reaction of [XeF]'[SbzFJ with Xe
XeF2
XeF,
XeOFz
XeF6
[XeF]' ; [XezF3]+
[XeF3]+
[XeOFd
[XeF5]+ ; [Xe2F,,I'
SCHEME
1. Major cationic and anionic derivatives of xenon fluorides and oxide fluorides.
69
FIG.3. The structure of [Xe,l'[Sb,F,,I-.

IReprinted with permission from (106);
copyright 1997 VCH Verlagsgesellschaft mbH. I
occurs in the presence of HF-free SbF5),and it is speculated that the

ion IXe(HF)21z+
is the active oxidizing agent in the Xe/XeFz/SbF,/HF
system (106).
One piece of recent chemistry of IXe,l+ has been its reaction with
bromide and iodide ions in the presence of xenon gas, which yields
the excited complexes XeBr": and XeI* respectively (107).
Xenon difluoride has an extensive fluoride-ion donor chemistry,
which is exemplified by its reaction with a large range of transition
metal pentafluorides to give adducts formulated as [XeFl'[MzFI,l-,
[XeFI'[MF,I-, and IXezFyl+[MF,Iin which the cations and anions are
linked by weak fluorine bridging bonds, F-Xe----F----M.The V-shaped
[Xe2F31'cation is symmetrical, two [XeFI' units being linked by a fluorine bridge. Related complexes incorporating the weaker Lewis
acids, MOF, (M = W or Mo), are similar but have less ionic character.
Essentially, no new chemistry has been derived in this area since
publication of the review by Selig and Holloway in 1984 (12). However, it has been shown that [N,FI-[AsF,I- will oxidize xenon to
[XeFl' (108). On the theoretical front, the electronic structures and
bonding in the [Xe2F31'cation and the unknown but isoelectronic and
isostructural neutral species XeIF3 (i.e., [FXeFXeF]' and FXeFIF respectively) have been calculated using ab initio molecular orbital theory with polarized split-valence basis sets (109). The [XezF31' cation
is known to be bent at the central fluorine, but the calculated structure is liqear and the bonds from the xenons to the central fluorine
are 0.09 A shorter than the experimentally determined values. The
calculated vibrational frequencies are in reasonable agreement with
70
HOLLOWAY AND HOPE
experimental values after scaling. The [XezF31fcation is a good example of a 512,6e hypervalent bond; for XeIF3,the bonding is less delocalized but the components of the six-electron (5c, 6e) hypervalent bond
are still present.
In accord with enthalpy of ionization data, xenon tetrafluoride
forms complexes with only the strongest fluoride-ion acceptors, but
XeFs has a much more extensive and complicated chemistry, forming
1: 2, 1:1, 2 : 1, and 4 : 1 adducts with a variety of tri-, tetra- and pentafluorides that incorporate [XeFJ' or [Xe2F11]+
cations. Solutions of
the hexafluoride in anhydrous HF have a higher conductivity than
those of XeFz and XeF4 because of the occurrence of both of these
cations in solution (12):
(DLeF51tF-)4+ nHF G 2[XezFlll++ [(HF),I[XezFlllt+ nHF + 2[XeF51' + [(HF),Fl-.
(13)
(14)
Work that was not covered in the 1984 review (12) or has appeared
since includes preparation of 1: 1 complexes with LnF3 (Ln = La, Pr,
or Nd), which, upon thermal decomposition, gave rise to the adducts
2LnF3 XeF, and 3LnF3* XeFs. The stabilities of the 1: 1 and 2 : 1complexes appeared to decrease with increasing atomic number of the
lanthanide (110). Reactions with tin and lead tetrafluorides have
given the adducts XeFs.4MF4,3XeFs.4MF4,4XeF6.MF4(M = Pb, Sn),
and 4XeFs.3SnF4 for the first time (111).The 1 : 4 adducts contain
[XeFJ' cations and polymeric anions, and the 4 : 1 complexes contain
[Xe2F11]+cations. The complexes with 4 : 3 and 3 : 4 stoichiometries are novel and are the first xenon hexafluoride complexes that
contain both [XezFllltand [XeFJ+ cations. More recent work from the
same group on the 4XeF6.NiF4species is the first to have shown definitively the presence of two [XezFlll+cations in the same compound.
The complex, prepared from the reaction of NiFz, KrFz,and XeFs in
(50).It has
anhydrous HF, is therefore formulated as Dre2Flll~[NiFslZalso been shown that displacement of BrF, from BrF3.AuF3with XeFs
gives [XeF,lt[AuF,l-, which is isostructural with [XeF51'[AgF41-, the
structure of which contains double layers of [XeFJ' and layers of
[AgF41-with all the layers parallel to the ab-plane. The greater basicity of [AuF4]-relative to that of [AgF41-implies that the ligand charge
in the gold anion should be higher than that in the silver salt. This
appears to be confirmed by the difference in the unit cells of the two
salts and is consistent with observed differences in the basicity and
oxidizability of the anions (51).
71
Work on species incorporating metal pentafluorides includes data

on 2XeFs-VF5,which had been reported earlier, and the novel complexes XeFs.nVFS( n = 1 or 2). This Raman study suggested that ionic
character decreases from essentially ionic to molecular in the order
2XeF6.VF,, XeFs.VF,, XeFs. 2VF5(112). More recently, the first BiF,
complex, the white diamagnetic solid [XeF,I'[BiF,I-, has been obtained. This exhibits a strong fluorine-bridge interaction between the
cation and anion, which distorts the anion from Oh symmetry (113).
Further detailed studies on the interpretation and assignment of vibrational spectra have been reported by two independent workers
(114,115).
Adducts of XeFz with [XeF,]' salts were first examined in 1971

(1161,but to obtain a firmer understanding of the factors that determine adduct formation and to investigate the possibility of lower fusion temperatures, the phase diagram for the XeF2/[XeFS]'[AsF61-system has been studied and representative structures of the
XeFz: [XeF,I'[AsF,I- products (1: 2, 1: 1, and 2 : 1) have been carried
out by X-ray single-crystal methods (117).Whether the complexes are
polymeric or nonpolymeric is correlated with the coordination number
of the [XeF,]' cations in the complexes, and this is determined by the
extent of ionization of the XeFz species.
There has been significant interest in the problems associated with
heptacoordination in recent years [see (118)references 1-111, and although [XeF71+has not been synthesized, its possible existence is of
major interest in this context. Recent ab initio calculations employing
effective core potentials and density functional theory calculations
have been used for analysis of the isoelectronic series [TeF,]-, IF7,and
[XeF71+.These show that all three possess a pentagonal bipyramidal
equilibrium geometry and that [XeF7]' is a stable structure. However,
it is conceded that the synthesis of such a species is likely to be very
difficult because its calculated oxidized strength is closely similar t o
that of [KrFl' (118).
Xenon oxide tetrafluoride is a poorer fluoride ion donor than XeFz
and XeFs, and therefore complexes only with the stronger Lewis
acids. Thus, although [XeOF31t[SbFsl- and [XeOF,I+[Sb,F,,I- have
been known for some time (121, the first single-crystal structure determination of the [XeOFJ' cation was obtained on the [XeOF,]'
[SbFJ salt more recently. As expected, the isolated cation has a geometry that is consistent with VSEPR theory. Oxygen-17 and -18 enrichment of the [XeOF3]+cation in HF and SbF5 solvents has allowed
comprehensive investigation of the 170and Iz9Xenuclear magnetic resonance of the cation (119). Xenon dioxide difluoride yields
72
HOLLOWAY AND HOPE
[Xe0~Fl'[SbF~l-and [XeOzFlt[SbzF1ll-,but these too are somewhat

unstable, the cation decomposing slowly in SbF, according to Eq.
(15) (12):
SbF,
[XeO,F]++[Xe-Fl'
rt
+ 0,.
(15)
However, the reaction of XeOzFzwith excess of AsF5 in anhydrous

HF has been shown t o yield the [XeOzFlt[AsFsI-species, which has a
dissociation pressure of 7 torr at 23C. Under vacuum or when exposed to a laser beam, this loses AsF6to give [FOzXeFXeOzFl'[AsFsl-.
This same species can also be prepared by reaction of XeOzFzwith
AsF, in a 2 : 1 molar ratio in anhydrous HF. A detailed Raman study
of the two adducts has shown that, whereas the [FO2XeFXeOzF1'
[AsFJ species has minimal interaction between the anion and cation,
the [XeOzFlt[AsF6]-species is significantly fluorine bridged ( 120).
b. Anions Derived from Xenon Fluorides and Oxide Fluorides Fluoride ion acceptor behavior among the xenon fluorides has, until the
late 1980s, been limited to XeFs. Reactions of the alkali metal fluorides with XeF, yield MzXeFs(M = Cs, Rb, K, or Na) and M'XeF7(M' =
Cs or Rb), which lose XeFs when heated. The thermal stability of the
adducts increases with increasing relative molecular mass, and the
decomposition temperature of the sodium complex, which is below
1OO"C, provides a means of separating XeFs from XeFz and XeF4 (see
Section V,A). The related NOF and NOzF adducts have also been prepared, and the NOF derivative [NOld[XeFJ- was shown some years
ago to contain a slightly distorted square antiprismatic [XeFs12-anion
(12). More recently, it has been demonstrated that XeF, also reacts
with [NF41+[HFzIin anhydrous HF to give [NF,I+[XeF,I-, which uncomdergoes laser photolysis at 488 nm t o yield the [NF41~[XeF,I2plex. Vibrational spectroscopic data on these two complexes, along
with those of the previously prepared cesium salts, and evidence for
the existence of NaXeF, have been obtained (121). The [NF41+CXeF71salt is by far the most energetic of the known [NF,]' salts, and a DSC
study of its decomposition has provided a means of calculating its
heat of formation (AH? = -490.7 kJ.mol-') (122). The most important and most recent work in this area, however, is the X-ray structural analyses of Cs'[XeF71 -, [N021+[XezF131(123), and Csz+[XeF8l2-.
4BrF5 (124). Prepared from the reaction of XeFs with CsF in BrF5 at
4"C, the anion in Cs+[XeF,] is a capped octahedron, a geometry
strictly enforced by the symmetry constraints of the cubic lattice.
73
This, of course, is at variance with most other seven-coordinate maingroup compounds, which are pentagonal bipyramidal. The bond from
xenon to the capping fluorine is long (210 pm) compared with the
average of the rest of the xenon-fluorine bonds (Xe-Foh = 195 pm).
This may be a consequence of interaction between Xe-F,,, and the
nonbonding electron pair, or it may arise from interaction with three
cesium cations (the other fluorines have contacts to only two cations).
The anion in [NOzl+[XezFJ can be considered as an adduct of an
[XeF7]-anion with a discrete XeFfimolecule (Fig. 4). This is the first
structure in which distinct XeFs as opposed to [XeFJ'F- units have
been observed. The [XeF71- adopts a capped trigonal prismatic arrangement in which the Xe-F,,, bond is the shortest and the lone pair
appears to be between the two Xe-F bonds trans to Xe-F,,, . The XeFs
part of the adduct has CZvsymmetry, with two short, two intermediate, and two long Xe-F bonds, and so is significantly distorted from
the regular octahedral structures exhibited by the isoelectronic [IF,]and [SeF,I2- anions (123).In the BrF,-solvated Csz'[XeF8I2-complex
( 124), the anion assumes the near-regular square antiprismatic structure found in (N0)2'[XeF81" (125). This structure is related to that of
[IF&, but the Xe-F bonds are longer. This can be explained in terms
I
I
I
I
Q-
F21
F2d
F23
3 F 1 5
F13
1
\
\
\
FIG. 4. The structure of [N021+[Xe2F,31-,

showing the disposition of the [XeF,I- and
discrete XeF6 molecule and the INO,]'. [Reprinted with permission from (123); copyright 1996 VCH Verlagsgesellschaft mbH.1
74
HOLLOWAY AND HOPE
of shielding by the centrosymmetric nonbonding electron pair or by

increased polarity of the Xe-F bonds due to the increased negative
charge.
The first evidence of possible fluoride-ion-acceptor behavior of XeF,
was from reports of the synthesis of Mi[XeF612-(M = Cs, Rb, K, Na)
salts from the reaction of XeF, with MF (126-128). The characterization of these species as [XeFs12-salts was based on vibrational spectra
(2281, which were later shown to be incompatible with those for an
octahedral species (129). Furthermore, the Raman spectrum attributed to the cesium salt was found to be identical to that due to a
species observed during the laser photolysis of Cs+[XeFJ and tentatively assigned to Cs~lXeFs12(1211. These discrepancies prompted an
investigation of the reactions of XeF4 with N(CH3),F, which is a powerful reagent in the synthesis of complex fluoroanions, and a reexamination of the XeF4/MF systems, which has led to the identification
and full characterization of [XeFJ for the first time (130).Direct reaction of excess of XeF, with the alkali-metal fluorides, MF (M = Cs,
Rb, K, Na) at 190C and reaction of equimolar amounts of XeF4 with
N(CH3),F in CH3N at -40C or of XeF4 with excess of FNO at 0C
gave the 1: 1 complexes. Detonation of [N(CH3)41+[XeF5]in CH3CN
when cooled with liquid nitrogen can occur, and [NOI+[XeF.J is unstable, with a dissociation pressure of 10 torr at O"C, but the alkali
metal salts are stable, as was found in the earlier work (126-128).
All of the salts were characterized by vibrational, 19F,and 129XeNMR
spectroscopy, and an X-ray single-crystal structure determination was
carried out on [N(CH3)41+[XeF51-.The structure contains planar
[XeF51- anions of D5h symmetry in which all six atoms are coplanar.
The results from the structure and from a normal coordinate analysis
indicate a greater rigidity of the XeF5 plane than that in the fluxional
IF7, which is due t o repulsion from the xenon free electron pairs
(130).A study of the mean amplitudes of vibrations of the [XeFJ ion
has also confirmed that the lone pairs exert repulsion, which forces
bond lengthening (131).
Complexes of XeOF4 with CsF, RbF, KF', and NOF were first prepared in the 1960s (12).The alkali-metal compounds were prepared
by treating the fluorides with an excess of XeOF4 and pumping to
constant weight under appropriate conditions (for example, see
Scheme 2). The structures of the 3 : 1 complexes were shown by X-ray
crystallography to contain the [F(XeOF,)J anion, which consists of
three essentially undistorted XeOF4 units solvating an F- anion
(132, 133). On the basis of vibrational spectroscopic data (1331, the
[XeOFJ anion in Cs+[XeOF,]- was assigned a low-symmetry C, struc-
75
- oocvac.
CsF + XeOF, (excess)
20OC vac.
CsF.XeOF,
-XeOF,
2CsF.XeOF4
-XeOF,
1I
-XeOF, -125OC vac

-270C vac.
3CsF.XeOF4 -3CsF.2XeOF4
-XeOF,
SCHEME
2. The reaction of XeOF4 with CsF.
ture. However, recently two independent studies have described reinvestigations of this anion with different cations. Based on vibrational
spectroscopic data and ab initio calculations, the [N(CH,),]' salt has
been shown to have a pentagonal pyramidal ((3,")structure (1341,and
a n X-ray single-crystal structure on the [NO]' salt also indicates a
similar pentagonal pyramidal arrangement for the anion with the oxygen apical and in which xenon is coordinatively unsaturated and
forms two weak interactions with adjacent anions (135).
A new synthesis of Cs+[XeO2FJ, by recrystallization from a CsF/
Cs+[XeOF51-/Xe02F2mixture, and a preparation of a n unstable adfrom the reaction of XeOzF2 with
duct, [N02]+[Xe02F3.nXe02F21-,
F N 0 2 , have been reported (100). The two oxygen atoms in [XeO,F,Iare cis and not trans to each other. The Raman spectrum previously
attributed to Cs+[Xe02F3]- has been shown to be due to
Cs' [Xe02F3.nXeF21-.
c. Fluoride I Anion Metathesis Reactions between Xenon Fluorides
and a n Anhydrous Acid Reactions involving the replacement of fluorine in xenon fluorides with other groups began with studies of the
reactions of XeFs with anhydrous oxygen-containing acids ( 3 , 4 , 9 , 1 3 ) :
XeF2 + HL+ F-Xe-L

XeF2 + 2HL+ L-Xe-L
+ HF
+ 2HF.
( 16)
(17)
In all of the earlier cases the linkage to the xenon was via the electronegative oxygen (L = OCOCH,, OCOCF,, O N O p , OPOF2, OS02F,
OSeF,, OTeF5, OC103, or OIF40). Only those containing the highly
electronegative pseudo-halogens, -OSeF, and -OTeF, , are stable
under ordinary conditions. The -0S0,F and -OIF,O derivatives decompose readily, and the rest are explosive. Recently, reaction of xenon bis(trifluor0acetate) with trifluoromethanesulfonic acid (triflic
acid) has given the novel and highly reactive unsymmetrical xenon-
76
HOLLOWAY AND HOPE
0x0 species CF3COOXeOS02CFS(136). Most new work in this area,

however, has been concerned with species containing the OTeF, or
OSeF, groups. The high electronegativity of the OTeF5 group means
that Xe(OTeF& is chemically similar to XeF2, and interesting reactions with a variety of fluoro-olefins and fluorocarbon iodides have
provided useful pathways to a variety of substituted fluorocations
(137-139). More recently, further work in this area has shown that
Xe(OTeF5)2reacts with an excess of perfluorobutadiene to yield mostly
TeF50CF2CF=CFCF20TeF5with a small trace of TeF50CF2CF
(OTeF5)CF=CF2.These are the first examples of -OTeF5-substituted
unsaturated fluorocarbons (140). The products of a similar reaction
with perfluorocyclohexene were found to be temperature dependent.
T ~ obtained
F ~ ) ~ as expected, but
At 95C a high yield of C - C ~ F ~ ~ ( Owas
at 115C the main product was the mono -OTeF5-substituted comthat, at the higher temperature, fluopound C - C ~ F ~ ~ O, indicating
T~F,
rination of one of the C = C carbons occurs (140). Reaction of
Xe(OSeF5)2with CF2=CF2,CFBCF=CF2,CF,=CFCl, and CF2=CFH
has shown that the OSeF5 group cannot be added to olefinic double
bonds as cleanly as its tellurium analog. As well as modest yield addition of two OSeF5 groups t o the double bond, higher yields of c2F5
OSeF5, n-CgF70SeF5,and SeF50CF,COC1 are also formed. In the case
of c-C5F8only monosubstitution occurred (141).
The ability of Xe(OTeF& to introduce two OTeF5 groups oxidatively
has previously been exploited in the synthesis of neutral high-oxidation-state -OTeF5 derivatives from an appropriate neutral low-oxidation-state derivative. For example, Te(OTeF& can be prepared from
Te(OTeF& (142), I(OTeF& from I(OTeF& (1431, and ORe(OTeF,),
from ORe(OTeF,), (144).This same approach has now been used for
the synthesis of -OTeF5-substituted anionic species in the form of
the weakly coordinated tetraalkylammonium pnicogen anion series,
[M(OTeF,)J (M = As, Sb, or Bi). The arsenic and bismuth anions
were prepared by the reaction of the M(OTeF& derivative (M = As
or Bi) with a small excess of [N(CH,),]+[OTeFJ in S02C1F, but the
instability of Sb(OTeF5)5made this a n inappropriate precursor for [Sb
(OTeF,),]-. Consequently, a two-step process in which [Sb(OTeF5)J
was first prepared and then oxidized to [Sb(OTeF,)J with Xe(OTeF5)2
was used (145).
An intriguing extension of the chemistry of xenon(I1) pentafluoroorthotellurates has been a study of the ability of the OTeF, group
to participate in nonredox metathetical fluorinations and to be replaced by a suitably strong protonic acid. Thus, it has been shown
that dissolution of [XeOTeF5]+[AsFs]-in BrF5 results in the formation
77
of the fluorine-bridged cations [FXeFXeOTeFJ' and [XeF2.BrOF,]'

(146).
The reaction of XeF, with bidentate ligands has been regarded as a

particularly interesting challenge, and in 1989 it was shown that it
reacts with cz~-(H0)~TeF,
to form a polymer (Xe02TeF4),1
that is stable
to around 80C. The reaction of cis-(HO)2TeF, with [Xe2F,l+IAsF6]
gives a material with the composition [FXeOTeF40Xe]+[AsF6].Recrystallization of the latter in anhydrous H F gave the HF-solvolyzed
species HF[HOTeF,OXe]'[AsF,], for which an X-ray single-crystal
structure was obtained. This shows that the complex is closest to approaching a purely ionic formulation of any xenon complex so far isolated. The Xe----F-AsF5 contact is 245.8 pm compared with 234.0 pm
in [XeFI+[Sb,F,,]- (147).
The use of B(OTeF5)to prepare -OTeF, derivatives of many binary
fluorides has inevitably impacted on xenon chemistry. In 1978
Xe(OTeF,), was reported (148), and in 1979 Xe(OTeF,), was prepared
for the first time by the reaction of XeF, with B(OTeF,), in perfluoron-pentane at -40C and the XeOF4 derivative XeO(OTeF& was prepared by a similar route, or by hydrolysis of the xenon hexakis{pentafluorooxotellurate(V1)) (149). All three were discussed in more
detail two years later (150). More recently a series of mixed fluorol
pentafluoro-orthotellurate derivatives of XeOF, and Xe02F2have been
prepared. The novel compound 02Xe(OTeF& was prepared according
to Eq. (18):
With excess of XeOzFy,a mixture of 02Xe(OTeF&, 02XeF(OTeF5),and

XeO,F, is produced (1511. The structures of all of these species and of
the related Xe(OSeF& have been of interest since they were first prepared because the ligands are almost completely unable to form bridging in either solid or liquid structures. It has been anticipated, therefore, that xenon derivatives of these ligands will exhibit the expected
geometry of the central atom without influence from weaker interactions. The structure of Xe(OTeF& , although based only on X-ray powder data, has been shown to be a true derivative of XeF, (146). More
recently, a single-crystal X-ray study of Xe(OTeF5I4has demonstrated
that this, too, is as expected, the xenon being surrounded by four
equidistant oxygen atoms in a square planar arrangement. The struc, formed according to the reaction represented by
ture of 02Xe(OTeF5)2
Eq. (18),shows a very distorted tetrahedral or pseudo-trigonal bipyra-
78
HOLLOWAY AND HOPE
midal arrangement about the xenon, which is not consistent with

VSEPR predictions (152).
Stable oxygen-bonded xenon(I1) and xenon(1V) derivatives of the
highly electronegative O=IF40 group have also been prepared. The
generally convenient route t o such species by direct interaction of the
xenon fluoride with the acid, in this case HOIOF,, was not used because it is unsafe; instead, the IOzF3dimer was used in insertion reactions of the type shown in Eqs. (19) or (20) in SOzCIF,CFC13,HF, or
BrF5 solvents, or an acid displacement reaction between Xe(OTeF&
and HOIOF, in CFC1, solvent [Eq. (21)l was used:
+ (IOZF3),+ FXeOIF40 + IOF, + &02

FXeOIF40 + (I02F3),+ Xe(OIF,O), + IOF3 + $0,
Xe(OTeF5)2+ BHOIOF, + Xe(OIF,O), + 2HOTeF5
XeF,
(19)
(20)
(21)
In solution, the products of Eqs. (19) and (20) are mixtures of cis and
trans isomers, but the solid obtained from Eq. (21) is the cis, cisXe(OIOF,), compound (153, 154). The reaction of XeF, with I0,F3 in
a 1:2 ratio in CFCI, solution produced F3XeOIOF4(153).Efforts to
bring about ammonolysis of XeF, and XeF4 have resulted in the formation of xenon, nitrogen, HF, and ammonium fluoride. Reaction of
excess of ammonia and XeF6 yields the same products, whereas reaction between ammonia and an excess of XeFs gives an explosive, white
solid (96).
The first claim for the successful synthesis of a xenon-nitrogen
bond was made in 1974 (155).Initial skepticism, based on concerns
that the species might actually involve xenon-oxygen rather than xenon-nitrogen bonds, was removed in the early 1980s when the Xe-Nbonded configuration was definitively demonstrated by a single-crystal structure analysis of FXe[N(S02F)21carried out at 218 K (Fig. 5)
(156).
In the meantime, a second Xe-N-bonded species had been proposed
(157),and another pawith the formulation [{(FS02)2NXe}2Fl+[AsFslper, which outlined further details of these two species and a new
compound Xe[N(SO,F),I2, was published (158). The species FXe"
(SO,F,),I and Xe[N(S02F)212
were prepared by the low-temperature reactions of XeF, with HN(SO,F), whereas the complex salt [{(FSO,),
NXe}2F]t[AsF61-was prepared by reaction of FXe[N(SO,F),I with AsF5
( 158).The compounds were characterized by multinuclear NMR spectroscopy (19F and 129Xe)(1581, and the nature of Xe[N(SO2F),I2was
79
FIG. 5 . Perspective drawing of FXeN(S0,F) showing bond lengths. [Reprinted with

permission from (156);copyright 1982 the American Chemical Society.]
subsequently confirmed by multinuclear magnetic resonance spectroscopy of the 15N-enrichedcompound (159).

More recently, the fluoride-ion-donor properties of FXe[N(S02F)2]
have been exploited in the synthesis of adducts with stoichiometries,
FXe[N(S02F)21.AsF5,FXe[N(S02F)21.3SbF5, and 2{FXe[N(S02F)2]}.
AsF5 (160). All of the compounds were enriched to 30% with 15N and
studied by multinuclear magnetic resonance (15N, lYF,IZ9Xe)and Raman spectroscopy, which has shown that the adducts can best be described in terms of predominantly ionic formulations, CXeN(S02F)21+
[AsFJ, [XeN(S02F)21+[Sb3F161-,
and [F(XeN(S02F)2}l+[AsFsl-.The
was determined from
crystal structure of the [XeN(S02F)21+[Sb3Fls1three-dimensional X-ray data a t - 64C and contains discrete [XeN
(S02F)21+[Sb3F161~
species (Fig. 6) in which the cation is Xe-N-bonded.
The Xe-N bond length [2.02(1) A] is somewhat shorter than that in
FXe[N(S02F)21[2.00(3) and the cation and anion are linked by a
weak covalent &-fluorine bridge bond in which the Xe----Fbond
length is 2.457(8) A (160).
Xenon-nitrogen chemistry has been further extended and a new
class of Xe-N bonded compounds obtained by the reaction of Lewis
acidic [XeFl+ with the lone pair of a series of nitriles. The first of this
new series were obtained by the reaction of the appropriate nitriles
RCGN (R = H, CH3,CH2F,C2H5,C2F5,CxF7,C6F5)with either [XeF]'
A],
80
HOLLOWAY AND HOPE
FIG.6. Perspective drawing of the [XeN(SO,F),l' [Sb3F,J structural unit. [Reprinted

with permission from (160);
copyright 1986 the American Chemical Society.]
[AsFs]- or [Xe2F31t[AsFsl-in anhydrous HF at low temperature (1611.

Subsequently, other novel xenon-nitrogen-bonded cations have been
observed by the reaction o f R&=N (RF = CF3, CzF5,n-C,F,) in a
similar way ( 3 5 )and by the reaction of equimolar quantitites of [XeFI'
[AsF,I- with the perfluoropyridines, 4-RC5F4N(R = F or CF,) in anhydrous HF at -30 to -20C or at -30C in BrF6 solutions (162). The
cations, [C5F5N-XeFl and [4-CF3C5F,N-XeFl+, are stable at - 30C
and can be isolated from BrF, solution as the [AsFsl- salts. Low-temperature Raman and multinuclear magnetic resonance (14N,I9F,129Xe)
studies suggest that the cations are planar, with the xenon atom coordinated to the aromatic ring through the nitrogen lone pair (Fig. 7).
+
FfiF
F
N
I
Xe
I
F
FIG.7
81
These nitrogen bases are all oxidatively resistant and have first adiabatic ionization potentials close to the electron affinities of [Xe-F] '
[see reference ( 3 5 ) and Section IVI. The estimated electron affinity of
KeF]' is 10.9 eV and the first ionization potential of s-trifluorotriazine, s-C3F3N3,is 11.50 eV, which led to the idea that the [s-C3F3N2NXeFl' cation should exist. Reaction of the liquid triazine with [XeF]'
[AsF,I- at room temperature, followed by removal of excess of the
s-trifluorotriazine under vacuum, left a white powder that is stable
indefinitely at room temperature, for which multinuclear magnetic
resonance (19F,lZ9Xe)and Raman spectroscopic studies indicate clearly
that it is the Xe-N-bonded cation [s-C3F3N2N-XeFl'(Fig. 8) (35).
The most recent attempt to extend xenon-nitrogen chemistry has
been an investigation of the reactions of XeF2 with HN3, NaN,, and
NaOCN in solution in HzO, anhydrous HF, or S02C1F. Although no
stable xenon-nitrogen compounds were obtained, on the basis of
product distribution, FXe(N3)and FXe(NC0) have been postulated as
intermediates and ab initio calculations have shown that both possess
stable minima (163).
Since noble-gas chemistry was initiated in 1962, there has been
interest in the possibility of making a xenon-carbon bond. Experiments in 1979 appeared t o have provided a solution. When xenon
difluoride was bled into the tail of a trifluoromethyl radical plasma, a
volatile, waxy white solid was produced and trapped a t -196C. The
reported properties of this material are consistent with the formulation, Xe(CF3)2( I 6 4 ) ,but until this is independently confirmed, some
doubt remains about its authenticity.
In 1989, two groups working independently prepared pentafluorophenylxenon borates by nucleophilic displacement of fluorine in XeFz
using B(C,FsI3 as an aryl-transfer reagent. The resulting colorless
solid, which has a stable xenon-carbon bond, was characterized in
solution by lZVXe
and 19FNMR [e.g., Eq. (22)l (165, 166):
F
Xe
F
FIG.8
82
XeFz+ B(C,jF5)3
MeCN
HOLLOWAY AND HOPE
[C,F5Xe]+[B(C6F5)3F]+ BF3.MeCN + BF;.
(22)
Reactions with Te(C,F,), and C6F51yielded [(C,F,),TeI and [(C6F5),II+

respectively (165), and the hydrolysis and thermal stability of
[C,F5Xelt were also reported (166). Very soon afterward more spectroscopic information was provided, and the Xe-C bond length was obtained in a n X-ray structural determination on [MeCN----XeC,F51t[(C,F5)zBF,I-(Fig. 9)(167).
Since then, several arylxenon derivatives have been prepared via
exchange reactions of XeF, and triarylboranes. The salts vary in thermal stability; the 2,6-difluoroaryl species are more stable than the
monofluoroaryl derivatves, and the parachlorophenyl derivative decomposes at about 0C (168).Reactions of arylxenon species, and especially [CsF5Xelt,have become a n area of intense activity. With PhX,
[C6F5Xel+[AsF61-reacts to give isomeric mixtures of biphenyls,
XCsH4CsF5(X = Me, F, CF3, NOz, CN) (169). Reaction of the same
salt with cesium pentafluorobenzoate in water [Eq. (2311 gives an unexpectedly stable covalent acyloxyxenon(I1) derivative, the X-ray crystal structure of which indicates a linear F5C6-Xe-0 arrangement
with dimerization through oxygen bridging (170):
+
[CsF5Xelf[AsFJ + Cs[0,CC,F51 + C6F5Xe02CC6F5.
(23)
The first alkenylxenon(I1) derivative has been obtained by the reaction of [C6F5Xelt[AsFsl-with xenon difluoride in anhydrous HF, the
reaction proceeding in two steps giving first the (heptafluoro-1,4-cyclo-
FIG.9. The structure of the [MeCN-Xe-C,FJ+ cation showing the Xe-ligand bond
distances. [Reprinted with permission from (167); copyright 1989 VCH Verlagsgesellschaft mbH.1
83
hexadien- 1-yl)xenon(II)derivative and then the (nonafluorocyclohexene-1-yl)xenon(II) salt (171):
Reactions of [C6F5Xe]+[AsF6](i) with halide ions in acetonitrile to

give aromatic and coupled aromatic molecules (172) and (ii) with pyridines of different basicities, where coordination weakens the Xe-C
bond and promotes C6F5radical formation, have been described (173).
As part of this work the adduct [Xe(C,F,)(NC,H,F,-2,6)l+[AsFslhas
been prepared and characterized crystallographically (I731. Xenon
difluoride in aqueous HF acts as a strong electrophilic oxygenation
system, and reaction of both [C,F,Xe]+[AsF6]- and [C6F,Xel'[AsF61with this system has yielded new examples of the transformation of
an organic moiety bonded to xenon without cleavage of the Xe-C
bond ( 174 1:
/"
Reactions of [C6F9Xe]'[AsF61-with halide ions have been shown to be

solvent dependent. With iodide and bromide, in both acetonitrile and
anhydrous HF, the xenon is displaced by the halide to give CsFJ(X =
Br or I). Chloride and fluoride do not react in anhydrous HF, but in
acetonitrile the fluoride ions initiate the formation of alkenyl radicals,
which abstract hydrogen from the solvent to give C6FSH,whereas
chloride shows intermediate behavior, giving both CGF,Hand C6F,C1
(175). The structure of [(CsH3F~-2,6)Xe]+[BF41in the solid state is an
84
HOLLOWAY AND HOPE
ion pair with an Xe----Fdistance of 279.3 pm. This is about 70 pm

shorter than the van der Waals distance (176). Anion exchange of
this salt with trimethylsilyltriflate gives the related triflate salt as
expected (177).
Perhaps the most exciting chemistry to emerge so far from investigation of arylxenon derivatives is a preliminary report of the reaction
of [C6F,Xe]+with chloride ions in acetonitrile a t below -30C, which
has led to the synthesis and structural characterization of
[(CsFSXe)2C1]'[AsF6]-and the NMR spectroscopic identification of the
more thermally labile compound C ~ F S X ~both
C ~ , containing Xe-Cl
bonds (178 1. Other interesting chemistry includes electrophilic substitution of fluorobenzenes or trifluoromethylbenzenes using the unsymmetrical xenon-oxo species mentioned earlier (136) to give the
arylxenon trifluoromethylsulfonates as colorless solids (179 ). Subsequently, CF3COOXeOS02CF3has been used in a variety of reactions
with a range of benzene derivatives containing electron-withdrawing
groups such a s -F, -CF3, -C1, and -NOz t o yield a number of arylxenon trifluoromethylsulfonates. These have been characterized
mainly by NMR, mass, and vibrational spectroscopy; however, the
structure of [Xe(2,6-FzC6H3)1
+[OS02CF31-has been determined by single-crystal X-ray diffraction methods, which have shown that the
arylxenon is weakly coordinated by one oxygen atom of the anion
(180).
In a related area, interest in the isolation of alkynyl iodonium salts
has led to the synthesis of analogous isoelectronic xenon(I1) derivatives. These have been obtained by reaction of XeFz with ButC=CLi
or RC=CSiMe3 in BF3.OEt2 (181):
R-CrC-SiMe3
+ BF3.OEt2+ XeFz+
[R-C~C-XeI+[BF,I+ Me3SiF. (27)
As well as chemical reactions, some physical measurements have

been made. I n particular, cyclic voltametric measurements have
yielded reduction potentials for a series of arylxenon(I1) tetrafluoroborate salts that indicate that these are stronger oxidizers than the
comparable aryliodonium salts (182).
3. Reactions i n Which the Xenon Fluoride Acts as a
Fluorinating Agent
As outlined in Section IV, p. 55, the xenon fluorides and their cationic derivatives are capable of oxidation andlor fluorination of many
85
other species. The oxidizinglfluorinating strength is in the sequence,

XeF2 = covalent Xe(I1) derivatives < XeF, < XeF6, and their cations
are even stronger than the neutral species. It is also known that, in
the case of XeF2, the addition of anhydrous H F enhances its fluorinating ability, and the mechanism has been discussed. Before 1977
there was already a vast literature on the fluorination of organic fluorides, especially using XeF,, and this topic will not be covered here.
However, a comprehensive and authoritative review up to 1977 [see
R. Filler in reference (511 and articles within a n excellent recent book
(183)cover this area well, and the reader is referred to these.
Xenon difluoride and the other xenon fluorides were established as
good oxidative fluorinating agents in the early days of noble-gas
chemistry (2-4, 13). Soluble in water, the difluoride was found to be
stable in neutral or acid solution and sufficiently strongly oxidizing
to convert chloride to chlorine, chromium(II1) t o chromium(IV), and
bromate, [BrOJ, to perbromate, [BrOJ (131. In direct combination
with other species or in reactions in nonaqueous solution, it reacts
with main-group and transition-metal compounds and is incorporated
into intercalation compounds. Because of the difficulty and danger
associated with the manipulation of XeF, and XeF6, the oxidative fluorination chemistry of these is less closely studied. Work carried out
up to and including 1980 has been usefully summarized (13),and only
chemistry carried out since then is covered here.
a. Reactions with Main-Group Species There has been extensive

application of XeFLas a reagent in main-group chemistry recently. It
has been used to cleave the B-0 bonds in B(OR), to give FXeOR (R =
SO,CF, or S02C,F,) and Xe(OR), (R = COCF,%or COC,F,) (184). Aryl
trimethylsilanes are readily fluorodesilylated with XeF2 in hexafluorobenzene a t room temperature to give aryl fluorides, the reaction
occurring via a radical mechanism (185), and vinyl stannanes are
fluorinated to fluoroalkenes with XeFz and silver hexafluorophosphate
or trifluoromethanesulfonate, the latter giving the faster conversion
(186). The silicon and germanium compounds C,F,MX, (M = Si, Ge;
X = C1, Br) undergo halogen exchange with XeF2 without C-M cleavage. However, if X is fluorine or alkyl, XeFz in the presence of
BFa.OEt, or with [XeFl+[NbF,l- yields the heptafluorocyclohexadienyl derivative C6F7MX3by the addition of two fluorine atoms to the
pentafluorophenyl ring (187). The surface chemistry of silicon has
also been studied by exposing it to XeF2 vapor and examining the
chemisorption of silicon at room temperature and - 150C using XPS
and AES (188).
86
HOLLOWAY AND HOPE
Studies on reactions of main group alkyl and aryl derivatives have

been going on for some years. Recent studies have shown that reaction of XeFz with MePhzP and MePhzAsin CH3CNproduces MePhzPFz
and MePhzAsFzrespectively, and with 2-MeCsH4P(NEt&, 2-MeC6H4
PF,(NEt2)zis obtained (189). Related reactions of XeF, with Me3X
(X = N, P, As, or Sb) all proceed smoothly, the rate reflecting the
basicity of the substrate, Me3X.Where X is P, As, or Sb, the difluoride
Me3XF2is formed. Where X is N, cleavage of a C-H bond is the main
reaction (190). Xenon difluoride also oxidizes phosphorus(II1) acid derivatives to give either the phosphorus(V) monomer or the ionic salt
respectively (1911,
and M(C,F,), (M = As, Sb) (192) and (2,6-FzC6H3),Bi(193) to the

respecdifluorides M(C,F,),F, (M = As, Sb) and (2,6-FzC6H3)3BiFz
tively.
Although protonation of HOF by reaction of AsF5 or SbF, with water and HF was not successful, the fluorination of water with
[XeFI+[MFJ (M = As, Sb) produced pale reddish [H20Fl+[MF61salts
almost quantitatively. These can be stored for about 2 h at room temperature without decomposition, but under vacuum release HOF, HF,
and MF5(194).
A wide variety of reactions of XeF2 and [XeF]' with low-valent sulfur species have been studied. Thus, an electrophilic fluorination reagent capable of fluorinating carbanions in moderate yield has been
observed in the reaction of XeF, with sulfur(I1) derivatives such as
MezS (195).Reactions of Me$ (and MezO)with XeF, in trichlorofluoromethane result in cleavage of C-H bonds to form CHzF derivatives
as the main reaction pathway. In the absence of a C-H bond, for
example if neat (CF3)ZS is used, oxidation to (CF3)2SF2occurs (190).
The reaction of MezS with XeF, in nonacidic medium gives rise t o a
product that is best formulated as [Me,SCHzSMel+[F(HF>,l-.In HF
solution at -23"C, however, the product is [Me2SFlt (196).A similar
reaction has also been observed in the combination of (C&?s)aSwith
[XeFl+[MFJ (M = As, Sb) in HF, which yields deep violet crystals of
An X-ray crystal structure determination of the
l(C,F,),SFl'[MFJ.
antimony derivative has shown that the cations and anions are connected via interionic fluorine bridges to produce an infinite chain. An-
87
other feature is a short S-F bond distance [158.4(3) pm] (197). The
conversion of CH3S-- to CH,FS- in methionine and methionylglycine
derivatives has also been achieved using XeFa in solution in MeCN a t
or below room temperature (198). Oxidation of MeSCN or S(CN), t o
[Me(NC>SFl+and [(NC),SFl+ respectively using [XeFI+[MFJ (M =
As, Sb) has also been reported (199). The oxidative fluorination of
(CF,),C=NSNCO by XeF, gives rise to two isomeric sulfur(IV1 compounds, (CF3),CF -N =S(F)NCO and ( CF3I2CFN=S =NC(0)F. The
third possible isomeric product of 1,3 addition, (CF,),C=N-S(F)=
NC(O)F, was not observed (200).
The oxidative fluorination of sulfur(IV) compounds such as diphenyl
sulfoxide has been shown to occur under mild conditions with XeF,
and catalytic amounts of chloride ion. In the Ph$O case the product
is Ph,S(O)F, , which is formed in essentially quantitative yield within
a few minutes. The chloride ion reacts with XeF, to produce fluoride
ion, and a mechanism is proposed that involves fluorosulfur(1V)
anions and fluorosulfur(V) radicals (2011.
Reactions of MezSe with XeF2 in trichlorofluoromethane proceed
smoothly to give MezSeFz(190). Oxidative coupling of R2Se2(R = Ph,
biphenyl, alkyl) to unsymmetrical disubstituted acetylenes and
cycloalkynes using xenon difluoride has been shown to yield the vicinal (E)-fluoroalkenyl selenides in high yield. Alternatively, the PhSeF
equivalent can be formed from PhSeSiMe:i and XeF, (202). Although
PhSeCl and PhSeBr are commercially available sources of selenic
electrophiles, PhSeF, which is too labile to be isolated, can be generated in situ from PhSeSePh with XeF, (203). This has been used to
generate the novel phosphaalkenes from phosphaacetylenes. Reaction
of XeFz with Ph,Te gives PhzTeF2(189).In related studies PhTeF3 or
PhTeTePh gave PhTeF5, PhzTeFz or PhsTeFBgave PhzTeF,, Ph3TeF
or Ph3TeCl gave Ph3TeF3 (with small amounts of Ph3TeF,Cl in the
case of Ph3TeC1),and Ph4Te gave Ph,TeF2 (204). The latter has subsequently been characterized (205). In a study on the synthesis of tetrakis(perfluoroalky1) tellurium species, Te(Rf)4,the compound Te(CF3),
was obtained from the reaction of Te(CF3),ClZwith Cd(CF,), . Reaction
of this compound with fluoride ion acceptors gives rise to the complex
cation [Te(CF:3)31i,and with certain fluorides the complex anion
[Te(CF3l4F1-is formed. However, with XeF, it appears to be oxidized
to (CF3I4TeF2(206).The related compound R,TeF, (R = CH3)was prepared similarly by the reaction of TeR4 with XeF2 (207). As part of a
study of the Lewis acid behavior of [Te(OTeF,),l toward [OTeF,l-,
higher concentrations of the [FTe(OTeF,),]- anion in solution in
88
HOLLOWAY AND HOPE
SOzCIF were generated by the interaction of XeF, with [N(CHs),l+

[Te(OTeF,),I- at -50C according to Eq. (30) (51):
2[Te(OTeF5)J
+ XeFz
S0,ClF
-50C
2[FTe(OTeF5),I-+ Xe(OTeF&. (30)
Liquid bromine reacts with [XeF]+[MF6]-(M = As, Sb) at room temperature in a complex reaction sequence that gives BrS+[MF6]-,
and its crystal structure has been determined (208).

The reaction between HI and [XeFIt[AsF6]-has been shown to give
as the final products of reaction. No experiXe, HF, and [I,12'[AsFfi1~
mental evidence for [XeIl' was found in the reaction products, and
theoretical calculations show that such a species is bound only in the
'C singlet state and, after interconversion into the unbound triplet
state, would immediately dissociate into xenon ('S) and I'(3P) (209).
In synthetic and structural studies of a number of anions the [IF,]-,
[BrFJ, and [IF4]- salts were synthesized according to Eqs. (32) to
(34) (210):
CsBr + XeF,
Cs+[BrF,]- + Xe
The chemistry of the interaction of XeFz with alkyl and aryl iodides
has been further extended by investigations of the reactions with
CF3CH21,3,5-ClzC6H31,and 2-CF3CfiH41.
In each case the alkyl(ary1)
iodine difluoride was obtained (189). In related reactions between
MeX (X = C1, Br, or I) and XeFz, the reactions proceed smoothly, the
rate of reaction reflecting the basicity of the substrate, and in the case
of MeI, MeIFz is formed. With MeCl and MeBr, however, MeF is the
product (190). Although difluoroiodo compounds have been used as
89
synthetic intermediates, none had been isolated until treatment of

4-iodotricyclene with excess of XeFz in CC1, produced the difluoro derivative [Eq. (35)]as a pale yellow, waxy solid. This is stable in air
for a few hours and indefinitely in solution under a n inert atmosphere (2111:
IF,
b. Reactions with Transition-Metal, Lanthanide, a n d Actinide Species The reaction of chromyl fluoride with XeFz a t temperatures up
to 278C is a high-yield route to CrOF3 that is superior to other routes
(52). Reaction with CoFz, CoClZ, or metallic cobalt have all been
shown to yield CoF,, and DTA data have provided information on the
temperatures of initiation of the reactions established ( 4 6 ) .The mild
fluorinating ability of XeFz in solution has been exploited in the synthesis of Ir(CO)C1z(PEt3)z(PF4),
whose I9F and
NMR spectra
clearly showed the presence of the Ir-PF, group (212):
It was also found to fluorinate the five-coordinate carbonyl complex of

Ir(I), [Ir(C0)3(PEt3)z]+,
to yield the first fluoroacyl derivative, Ir(CO)2
F(COF)(PEt,), (213). Xenon difluoride also reacts with [Pt(NH3),1C12
in the solid state to give t r a n ~ - [ P t ( N H ~ ) ~ F ~but
l C lin
~ , solution in
MeCN this is converted to t r a n ~ - [ P t ( N H ~ ) ~ C(214
l ~ ] ).F ~
The analogous
oxidation of [Pt(py)4]2+[BF41f
gives the crystallographically characterized [PtFz(py)412+[BF41;
(215).
It was established in the 1970s that XeFz in solution was a valuable
means of introducing fluorine oxidatively into transition carbonyl
compounds to produce carbonyl fluorides (216, 21 7). Recently, there
has been renewed activity in this area, supported by the improvements in multinuclear Fourier-Transform NMR techniques. Among
90
HOLLOWAY AND HOPE
the more recent papers is a report of a reaction of XeF, with

[Re2(CO),,] in perfluoro-l,3-dimethylcyclohexanethat concludes that
the product is Re(C0)3F2(218)even though the properties of the compound produced are clearly those of Re(CO)SF.ReF5 (217). Considerable progress has been made with oxidative fluorination of ruthenium
and osmium carbonyls and their derivatives. Thus, [Ru3(CO),,I with
XeF, in anhydrous H F gave, in solution, cis-[RuFz(CO),l as the main
product along with small amounts of [RuF(CO)J+, [ R U ~ F ~ ( C O ) ~ I ,
[{RuF(CO),}~(~-F)I',mer- and fuc-[RuF3(CO)J, [{RuF~(CO)~},I,
and
[{RuF,(CO)3}(p-F){Ru(C0)5}l
+,all of which have been characterized by
13C,I9F and 13C-{19F}NMR spectroscopies. Removal of the HF solvent
yielded tetrameric [{RuF,(CO)~}~]
(219). Similarly, the equivalent reaction with [ O S ~ ( C O )produced,
~~]
in solution, a range of mono- and
dinuclear osmium carbonyl fluoride species, with c i s - [ 0 ~ ( C O ) ~ Fas~ l
the major product and [Os(CO),F]+,[ O S ~ ( C O ) ~ F
and
~ I ,[ O S ~ ( C O ) ~ Fas
~]+
minor products. The main product is stable in solution, but removal of
the solvent results in CO loss to give the fluoride-bridged tetrameric
analog to the ruthenium tricarbonyl difluoride, [{OS(CO)~F,}],(220).
The mechanism of the oxidation of low-valent ruthenium and osmium complexes has also been elucidated by a study of the reaction
of XeF, with [M(CO),(PPh3),] (M = Ru, 0s) in dichloromethane (221).
Fluoroacyl complexes of ruthenium and osmium, [M(COF)(C0)2
F(PPh3)21(M = Ru, 0s) were identified as intermediates in the oxidative addition of XeF, in solution, and the mechanism, probed by addition of BF3, has been shown to involve oxidation of the metal center
by [XeFI' to give a monofluorinated cation. Subsequent nucleophilic
attack by F- at the coordinated CO with CO elimination gives, finally,
[OC-6-131[M(CO),F2(PPh3)21.
In the ruthenium case, the ligand arrangement has been confirmed by a single-crystal X-ray crystallographic study (2211.
Controlled sequential reaction of [Ir4(C0>121
with XeF2 in anhydrous
H F a t low temperature yielded, initially, fac- and mer-[Ir(CO)3FS],
and, ultimately, IrF, . Removal of the solvent at the initial stage of the
reaction results in the precipitation of f~c-[1r(CO)~F~l,
which contains
predominantly u-bonded carbonyl ligands with high CO-stretching
frequencies (222).
Xenon difluoride has been shown to react with organomercury compounds, RzHg (R = PhC=C, p-MeOC6H4,p-Me2NC6H4,p-Et02CC6H4,
and PhCH2), with cleavage of the C-Hg bond t o give RHgF or RHgFI
HgF, mixtures. When R = benzyl, the RHgF formed undergoes fast
demercurization. From the composition of the reaction mixtures it has
been established that the reactions involve free radicals. The absence
91
of fluorinated organomercury products is evidence that XeF, reacts

more readily with the C-Hg bond than with C-H or C-C bonds. With
HgX, (X = C1, Br, I) under mild conditions, HgF2, halogen gas, and
xenon are produced (223).
The xenon fluorides, XeF2, XeF,, and XeF6, have all been used in
preparations of some lanthanum tetrafluorides, LnF4 (Ln = Ce, Pr, or
Tb) and their heptafluoro complexes, Cs3LnF7(Ln = Ce, Pr, Tb, Dy,
or Tm) (224). Detailed examination of the reaction of CssLnCl6 with
XeF, a t 100-400C has shown that CsyLnF7(Ln = Ce, Pr, Tb, Nd, or
Dy) species are formed, but for E r and Yb under the same conditions
only the lanthanum(II1) species CssLnFs (Ln = E r or Yb) are produced (224).
In the actinide field, stopped-flow spectroscopy has been used to
study the oxidation of plutonium(II1) to plutonium(1V) in 1M HClO,
by XeFz. Two moles of Pu(II1)are oxidized per mole of XeF, consumed,
and there is no evidence for oxidation to PuW) or Pu(V1). Rate constants and activation parameters have been obtained and a mechanism postulated that invokes a sequence of 2 one-electron oxidation
steps (225). The oxidation of americium by a range of xenoncontaining oxidants in KHCO,, KHCO, and K2C0,, and in K&O, solutions, has been studied (226,227).Xenon difluoride and xenon trioxide convert AmlIII) to Am(IV) in solution (2261, whereas NaJXeOJ
oxidizes Am(II1) to Am(IV), A m W , and Am(V1) mixtures unIess the
perxenate is in excess, when Am(V1) is produced (226,227). The trioxide, Xe03, reacts with Am(II1) only under UV irradiation (227).
VI. Radon Chemistry
The chemistry of radon (11) was outlined by Stein in 1983 (14).

Since then, evidence for radon in a higher oxidation state (RnF, or
RnF6) and a radon oxide (RnO,) has been claimed and disputed, and
the ions, [HRnOJ', [HRnOJ, and [RnOSFl have been reported.
Soon after the initial discovery of xenon compounds in 1962, a lowvolatility fluoride was detected in the product of the combination of
radon with fluorine a t 400C (226). It was also shown that 'L22Rnis
oxidized by chlorine and bromine fluorides, IF7 (227-230) or [NiFsl"
(227-229) in H F to give stable solutions of radon fluoride, and that
radon can be collected as a solid by oxidation of the gas with liquid
BrF3 or the solid complexes [C1F2]+[SbFsl-,[BrFJ'[SbFJ,
[BrF,]'
[Sb2F113-,[IF41:'+3[SbF61-,
and [BrF2I'[BiFJ (231-232). Electromigration studies suggested that in the solution studies the radon is pres-
92
HOLLOWAY AND HOPE
ent as [RnF]' or [RnI2+(228, 229). Such ionic formulations were also

favored in relativistic quantum mechanical calculations (233). This
has been borne out further by more recent chemistry in which solutions of cationic radon have been prepared by oxidation of elemental
radon with halogen fluorides and shown t o displace hydrogen, sodium,
and potassium in solid materials such as Nafion resins (H+ and K'
forms), potassium hexafluorophosphate, and sodium hexafluoroantimonate. It has also been shown that radon in this state can be quantitatively recovered by ion exchange and eluted with bromine trifluoride in sulfuryl chloride (234). By comparison of the properties of the
products from all of these reactions with the known properties of the
fluoride and fluoride complexes of krypton and xenon, it is almost
certain that radon forms a difluoride, RnFz, and derivatives such as
[RnFI+[SbFJ, [RnF]+[TaF6]-,and [RnF]+[BiF6]-,which are direct analogs of krypton and xenon species (235). This has been further
strengthened by the observation of the reaction of radon with solid
[02]+[SbF61-a t room temperature, which yields gaseous oxygen and
an involatile radon compound (236). Comparison with xenon chemistry suggests the reaction
This was further substantiated by reactions of radon with fluoronitrogen salts and halogen-fluoride metal salts, and the development
of a method of collecting radon from air using [o21+[SbF6l-or
[IFJ+[SbFJ (237).
Hydrolysis of radon fluoride with water yields radon, oxygen, and
HF. This parallels the reactions of KrFz and XeFz and is in marked
contrast to the reactions of XeF4 and XeF6, in which some xenon is
retained in solution as Xe(V1). It also coprecipitates with XeFz from
halogen fluoride solutions and complexes with XeFzbut not XeF4(238,
239) all of which points to the fluoride being RnFz.
Claims by Russian workers that a higher fluoride of radon, RnF4 or
RnF6, can be prepared in tracer experiments by heating radon, xenon,
fluorine, bromine pentafluoride, and either sodium fluoride or nickel
fluoride, and converted to RnOs by hydrolysis (240) appeared to others
(235) to be due to the precipitation of radon as a solid complex, which
is probably [RnF]2+[NiF612-.
However, the precipitation of CsXeOBF
from aqueous solutions results in the coprecipitation of radon, and
this has been taken by the Russian group as confirmation that RnOB
is the product of hydrolysis of the fluoride formed (241). Furthermore,
93
it has been suggested that the failure of Stein to observe the highoxidation-state [RnO3F1-species was probably due t o high F- concentration (242).More recently, ultracentrifugation of hydrolyzed radon
solutions, coprecipitation studies, and kinetic data for the decomposition of the solution species have been interpreted in terms of it being
RnO, (243). Recent electromigration studies have also been used to
suggest that in acidic aqueous solutions (pH > 5) cationic, [HRnOJ,
and anionic, [HRnOJ, forms of radon are produced, the validity of
the electromigration method having been established using xenon(V1) (244).
REFERENCES
1. Bartlett, N. Proc. Chem. Soc. 1962,218.
2. Holloway, J. H. Noble-Gas Chemistry. Methuen, London, 1968.
3. Horn, H.-G.; Hein, H. In Gmelins Handbuck der Anorganischen Chemie; Buschbeck, K.-C., Lippert, W., and Slawisch, A,, Eds.; Ergiinzungswerk zur 8. Auflage
Band 1, Edelgasverbindungen, Verlag Chemie, Weinheim, 1970.
4. Bartlett, N.; Sladky, F. 0. In Comprehensive Inorganic Chemistry; TrotmanDickenson, A. F., Ed.; Vol. 1, pp. 213-330. Pergamon Press, Oxford, 1973.
5. ( a ) Hyman, H. H. Science 1964, 145, 773; tb) Holloway, J. H. In Progress in
Inorganic Chemistry; Cotton, F. A,, Ed.; Vol. 6. pp. 241-269, Interscience Publishers, New York, 1964; (c) Malm, J. G.; Selig, H.; Jortner, J.; Rice, S. A. Chem.
Rev. 1965,65,199; (d) Moody, G. J . ; Thomas, J . D. R. Rev. Pure Appl. Chern. 1966,
16, 1; (e) Sladky, F. 0. In MTP International Review of Science; Gutman, V.,
Ed.; Inorg. Chem., Ser. One, Butterworths, London, 1972; (f) Jha, N. K. R.I.C.
Reu. 1971,4,147; (g) Stein, L. Yale Scientific Magazine 1970,44,(8), 2; (h) Malm,
J . G.; Appelman, E. H. Atomic Energy Rev. 1969, 7 (3), 3; (i) Filler, R. Isr. J .
Chem. 1978,17,1.
6. (a) Hyman, H. H. (ed.)Noble Gas Compounds, University of Chicago Press, Chicago, 1963; (b) Moody, G. J.; Thomas, J . D. R. Noble Gases and their Compounds, Pergamon Press, Oxford, 1964; (c) Claassen, H. H. The Noble Gases,
D.C. Heath and Co., Boston, 1966.
7. Hawkins, D. T.; Falconer, W. E.; Bartlett, N. Noble Gas Compounds, a Bibliography, 1962-1977, Plenum Press, New York, 1978.
8. Laslo, P.; Schrobilgen, G. J. Angew. Chem., Int. Ed. Engl. 1988,27,479.
9. ( a ) Holloway, J. H. Chem.. Br. 1987,23,658; (b) Holloway, J. H. In Fluorine, the
First 100 Years; Banks, R. E., Sharp, D. W. A,, and Tatlow, J . C., Eds.; Elsevier
Sequoia, Lausanne, 1986.
10. Zemva, B. In Encyclopaedia of Inorganic Chemistry; King, R. B., Ed.; Vol. 5, pp.
2660-2680, John Wiley & Sons, Chichester, 1994.
11. Seppelt, K. Accounts Chem. Res. 1979,12, 211.
12. Selig, H.; Holloway, J. H. In Topics in Current Chemistry; Boschke, F. L., Ed.;
Vol. 124, pp. 33-90, Springer-Verlag, Berlin, 1984.
13. Seppelt, K.; Lentz. D. Prog. Inorg. Chem. 1982,29,167.
14. Stein, L. Radiochim. Acta 1983,32,163.
94
HOLLOWAY AND HOPE
15. Breckenridge, W. H.; Jouret, C.; Soep, B. In Advances in Metal and Semiconductor Clusters; Duncan, M. A., Ed.; Vol. 111, JAI Press, Greenwich, 1995.
16. Liebman, J. F.; Allen, L. C. J . Am. Chem. Soc. 1970, 92, 3539.
17. Berkowitz, J.; Chupka, W. A. Chem. Phys. Lett. 1970, 7, 447.
18. Frenking, G.; Koch, W.; Deakyne, C. A.; Liebman, J. F.; Bartlett, N. J . Am. Chem.
Sac. 1989, 111, 31.
19. Wong, M. W.; Radom, L. J. Chem. SOC.,Chem. Commun. 1989,719.
20. Allen, L. C.; Huheey, J . E. J. Inorg. Nucl. Chem. 1980,42, 1523.
21. Filippi, A.; Troiani, A.; Speranza, M. Chem. Phys. Lett. 1997,278, 202.
22. Slivnik, J.; Smalc, A.; Lutar, K.; Zemva, B.; Frlec, B. J. FZuorzne Chem. 1975,
5, 273.
23. Al-Mukhtar, M.; Holloway, J. H.; Hope, E. G. J. Fiuorine Chem. 1992, 59, 1.
24. Bezmelnitsyn, V. N.; Legasov, V. A,; Chaivanov, B. B. Dokl. Chem. (Engl. Transl.)
1977,235, 365.
25. Dang, H.; Huag, L.; Xu, R.; Yang, R.; Zhang, B.; Wang, D.; Lu, Z. Wuji Huarue
Xuebao 1996,12, 18 (Chem. Abstr. 1996,125, 24976).
26. Burbank, R. D.; Falconer, W. E.; Sunder, W. A. Science 1972, 178, 1285.
27. (a) Siegel, S.; Gebert, E. J . Am. Chem. SOC.1963,83, 240; (b) Levy, H. A.; Agron,
P. A. J. Am. Chem. SOC.1963, 85, 241; (c) Brown, G. M.; Levy, H. A. J. Physi.
Radium 1964,25, 469.
28. Gillespie, R. J.; Martin, S.; Schrobilgen, G. J. J. Chem. Soc., Dalton Trans. 1980,
1898.
29. Christe, K. 0.;Wilson, W. W.; Bougon, R. A. Inorg. Chem. 1986, 25, 2163.
30. Christe, K. 0.;Wilson, W. W.; Wilson, R. D. Znorg. Chem. 1984,23, 2058.
31. Gillespie, R. J.; Schrobilgen, G. J. Inorg. Chem. 1976, 15, 22.
32. Lutar, K.; Jesih, A,; Zemva, B. Polyhedron. 1988, 7, 1217.
33. Frlec, B.; Holloway, J . H. J. Chem. Soc., Chem. Commun. 1973, 370.
34. Sanders, J. C. P.; Schrobilgen, G. J . J. Chem. SOC.,Chem. Conzmun. 1989, 1576.
35. Schrobilgen, G. J. J. Chem. Soc., Chem. Commun. 1988,863.
36. Schrobilgen, G. J. J. Chem. Soc., Chem. Commun. 1988, 1506.
37. Hillier, I. H.; Vincent, M. A. J. Chem. Soc., Chem. Commun. 1989,30.
38. Koch, W. J. Chem. Soc., Chem. Commun. 1989, 215.
39. MacDougall, P.; Schrobilgen, G. J.; Bader, R. W. F. Inorg. Chem. 1989,28, 763.
40. Gillespie, R. J.; Schrobilgen, G. J. J. Chem. Sac., Chem. Commun. 1974, 90.
41. Gillespie, R. J.; Schrobilgen, G. J. Inorg. Chem. 1974, 13, 1230.
42. Holloway, J. H.; Schrobilgen, G. J . J . Chem. Soc., Chem. Commun. 1975,623.
43. Liebman, J. J. Fluorine Chem. 1977,9, 147.
44. Bartlett, N.; Seppelt, K, unpublished results [see reference (13)1.
45. Christe, K. 0.; Wilson, R. D. J. Fluorine Chem. 1976, 7, 356.
46. Nikulin, V. V.; Popov, A. I.; Zaitseva, I. G.; Korobov, M. V.; Kiselev, Yu. M.; Siderov, L. N. Russ. J. Inorg. Chem. (Engl. Transl.) 1984,29, 21.
47. Spitzin, V. I.; Martynenko, L. I.; Kisel, W.; Ju. M. 2. Anorg. Allg. Chem. 1982,
495, 39.
48. Kiselev, Yu. M.; Goryachenkov, S. A,; Martynenko, L. I. Russ. J. Inorg. Chem.
(Engl. Transl.) 1984,29, 38.
49. Kiselev, Yu. M.; Sokolov, V. B. Russ. J. Inorg. Chem. (Engl. Transl.)1984,29, 493.
50. Jesih, A.; Lutar, K.; Leban, I.; Zemva, B. Inorg. Chem. 1989, 2911.
51. Lutar, K.; Jesih, A.; Leban, I.; Zemva, B.; Bartlett, N. Inorg. Chern. 1989,28, 3467.
52. McHughes, M.; Willet, R. D.; Davis, H. B.; Gard, G. L. Inorg. Chem. 1986,25, 426.
95
53. Zemva, B.; Lutar, K.; Jesih, A,; Casteel Jr., W. J.; Wilkinson, A. P.; Cox, D. E.;
Van Dreele, R. B.; Borrmann, H.; Bartlett, N. J. Am. Chem. SOC.
1991, 113, 4192.
54. (a) Drobyshevskii, Yu. V.; Serik, V. G.; Sokolov, V. B. Dokl. Chem. (Engl. Transl.)
1975,225, 1079; (b) Drobyshevskii, Yu. V.; Serik, V. G.; Sokolov, V. B.; Tulskii,
M. N. Sou. Radiochem. (Engl. Transl.) 1978,20, 200.
55. Drobyshevskii, Yu. V., Prusakov, V. N.; Serik, V. F.; Sokolov, V. B. Sou. Radiochem. (Engl. Transl.) 1980,22, 591.
56. Bacher, W.; Jacob, E. Chem. -Ztg 1982, 106, 177.
57. Peacock, R. D.; Edelstein, N. J . Znorg. Nucl. Chem. 1976,38, 771.
58. Brown, D.; Berry, J. A.; Holloway, J . H. J . Actinides 1981, 11, 81.
59. Asprey, L. B.; Eller, P. G.; Kinkead, S . A. Inorg. Chem. 1986,25, 670.
60. LeBlond, N.; Schrobilgen, G. J . J . Chem. Soc., Chem. Commun. 1996,2479.
61. Christe, K. 0.;Bougon, R. J . Chem. Soc., Chem. Commun. 1992, 1056.
62. Bougon, R., J. Fluorine Chem. 1991,53, 419.
63. Artyukov, A. A.; Khoroshev, S. S. Koord. Khim. 1977,3, 1478.
64. ( a ) Sakurai, T.; Takahashi, A,; Fujisawa, G. J. Nucl. Sci. Technol. 1982, 19, 255;
(b) Sakurai, T.; Takahashi, A.; Fujisawa, G. J. Fluorine Chem. 1982,20, 683.
65. Boltalina, 0. V.; Abdul-Sada, A. K.; Taylor, R. J. Chem. SOC.,
Perkin Trans. 2.
1995,981.
66. Zemva, B.; Hagiwara, R.; Casteel, J r . , W. J.; Lutar K.; Jesih, A,; Bartlett, N. J .
Am. Chem. SOC.
1990, 112, 4846.
67. Nielsen, J . B.; Kinkead, S . A.; Purson, 3 . D.; Eller, P. G. Inorg. Chem. 1990, 29,
1779.
68. Smalc, A,; Lutar, K. In Inorganic Synthesis; Grimes, R. N., Ed.; Val. 29, pp. 1-4,
Wiley-Interscience, New York, 1992.
69. Lutar, K.; Smalc, A.; Slivnik, J . Vestn. Slov. Kem. Drus. 1979,26, 435.
70. Lutar, K.; Smalc, A,; Zemva, B. In Inorganic Synthesis; Grimes, R. N., Ed.; Val.
29, pp. 4-6, Wiley-Interscience, New York, 1992.
71. Brassington, N. J.; Edwards, H. G . M. J . Mol. Struct. 1987, 162, 69.
72. Nabiev, Sh. Sh.; Ostroukhova, I. I. Spectrochim,. Acta. 1993, 49A, 1527.
73. Nabiev, Sh. Sh.; Ostroukhova, I. I. Spectrochim. Acta. 1993,49A, 1537.
74. Gillespie, R. J. J . Chem. Ed. 1963,40, 295.
75. Pitzer, K. S.; Bernstein, L. S. J . Chem. Phys. 1975, 63, 3849.
76. Brocas, J.; Rusu, C. Int. J. Quantum Chem. 1982,22, 331.
77. Strauss, H. L. Ann. Reu. Phys. Chem. 1983, 34, 301.
78. Christe, K. 0.; Wilson, W. W. Inorg. Chem. 1989, 28, 3275.
79. Klobukowski, M.; Huzinaga, S.; Seijo, L.; Barandiaran, Z. Theo. Chim. Acta. 1987,
71, 237.
80. Cutler, J . N.; Bancroft, G. M.; Bozek, J . D.: Tan, K. H.; Schrobilgen, G. J. J. Am.
Ch,em. SOC.
1991, 113, 9127.
81. Crawford, T. D.; Springer, K. W.; Schaefer 111, H. F. J . Chem. Phys. 1995, 102,
3307.
82. Kiselev, Yu. M.; Goryachenkov, S . A.; Russ. J. Inorg. Chem., (Engl. Transl.) 1983,
28, 9.
83. Leonidov, V. Ya.; Timoleev, I. V.; Kmelev, Yu. M. Dokl. Akad. N a u k S S S R 1979,
248, 1375.
84. Goncharov, A. A.; Kozlov, Yu. N.; Purmal, A. P., Zh. Fiz. Khznz. 1979,53, 2685.
85. Tse, J . S.; Bristow, D. J.; Bancroft, G. M.; Schrobilgen, G. .J. Znorg. Chem. 1979,
18, 1766.
96
HOLLOWAY AND HOPE
86. Bancroft, G. M.; Bristow, D. J.; Tse, J . S.; Schrobilgen, G. J . Inorg. Chem. 1983,
22, 2673.
87. Bartell, M. J.; Rothman, M. J.; Ewig, C. S.; Van Wazer, J. R. J. Chem. Phys. 1980,
73, 367.
88. Scheire, L.; Phariseau, P.; Nuyts, R.; Smith, A. E., Smith, Jr., V. H. Physica A
(Amsterdam) 1980, 101, 22.
89. Molkanov, L. I.; Grushko, Yu. S.; Mishin, Ya. K.; Isupov, V. K. Zh. Eksp. Teor.
Fiz. 1980, 78, 467.
90. Bagus, P. S.; Liu, B.; Liskow, D. H.; Schaefer, 111, H. F. J . Am. Chem. SOC.1975,
97, 7216.
91. Krauss, M.; Liu, B. Chem. Phys. Lett. 1976,44, 257.
92. Hay, P. J.; Dunning, T. H. J. Chem. Phys. 1977,66, 1306; (b) Dunning, T. H.; Hay,
P. J. J. Chem. Phys., 1978, 69, 134; (c) Hay, P. J.; Dunning, T. H. J . Chem. Phys.
1978,69, 2209.
93. (a) Malli, G. L.; Styszynski, J.; da Silva, A. B. F. Int. J. Quantum Chem. 1995,55,
213; (b) Styszynski, J.; Malli, G, L. Int. J. Quantum Chem. 1996,55, 227.
94. Bancroft, G. M.; Malmquist, P.-A.; Svensson, S.; Basilier, E.; Gelius, U.; Siegbahn,
K. Inorg. Chem. 1978,17, 1595.
95. Styszynski, J.; Cao, X.; Malli, G. L.; Visscher, L. J. Comput. Chem. 1997, 18, 601.
96. Huston, J. L. Inorg. Chem. 1982,21, 685.
97. Foropoulus, Jr., J.; DesMarteau, D. D. Inorg. Chem. 1982,21, 2503.
98. Christe, K. 0.; Wilson, W. W. Inorg. Chem. 1988,27, 1296.
99. Nielsen, J. B.; Kinkead, S. A.; Eller, P. G. Inorg. Chem. 1990,29, 3621.
100. Christe, K. 0.;Wilson, W. W. Inorg. Chem. 1988,27, 3763.
101. Schumacher, G. A.; Schrobilgen, G. J. Inorg. Chem. 1984,23, 2923.
102. Klaning, U. K.; Appelman, E. H. Inorg. Chem. 1988,27, 3760.
103. Stein, L.; Norris, J. R.; Downs, A. J.; Minihan, A. R. J. Chem. SOC.,Chem., Comnun. 1978,502.
104. Stein, L.; Henderson, W. W. J . Am. Chem. SOC.1980, 102, 2856.
105. Stein, L. J. Fluorine Chem. 1982,20, 65.
106. Drews, T.; Seppelt, K. Angew. Chem., Int. Ed. Engl. 1997, 36, 273.
107. Mezyk, S. P.; Cooper, R.; Sherwell, J. J . Phys. Chem. 1993,97, 9413.
108. Christe, K. 0.;
Wilson, R. D.; Bau, R.; Sukumar, S.; Dixon, D. A. J. Am. Chem.
SOC.1991,113, 3795.
109. Dixon, D. A.; Arduengo, A. J.; Farnham, W. B. Inorg. Chem. 1989,28, 4589.
110. Misra, S. N. Indian J. Chem., Sect. A 1979, 18, 530.
111. Zemva, B.; Jesih, A. J . Fluorine Chem. 1984,24, 281.
112. Jesih, A.; Zemva, B.; Slivnik, J. J. Fluorine Chem. 1982, 19, 221.
113. Druzina, B.; Zemva, B. J. Fluorine Chem. 1988,29, 309.
114. (a) MiliCev, S. Vib. Spectrosc. 1995,8, 309; (b)MiliCev, S. Mikrochim. Acta [Suppl.]
1997, 14, 539.
115. Nabiev, Sh. Sh. Zh. Neorg. Khim. 1995, 40, 2016.
116. Bartlett, N.; Wechsberg, M. Z. 2.Anorg. Allg. Chem. 1971, 385, 5.
117. Zemva, B.; Jesih, A.; Templeton, D. H.; Zalkin, A.; Cheetham, A. K.; Bartlett, N.
J. Am. Chem. Soc. 1987,109, 7420.
118. Christe, K. 0.;Dixon, D. A.; Sanders, J . C. P.; Schrobilgen, G. J.; Wilson, W. W.
J. Am. Chem. SOC.1993,115, 9461.
119. Mercier, H. P. A.; Sanders, J . C. P.; Schrobilgen, G. J.; Tsai, S. S. Inorg. Chem.
1993,32, 386.
120. Christe, K. 0.;Wilson, W. W. Inorg. Chem. 1988,27, 2714.
97
121. Christe, K. 0 . ;Wilson, W. W. Inorg. Chem. 1982,21, 4113.

122. Christe, K. 0.;
Wilson, W. W.; Wilson, R. D.; Bougon, R.; Bui Huy, T. J . Fluorine
Ch,em. 1983,23, 399.
123. Ellern, A.; Mahjoub, A.-R.;Seppelt, K. Angew. Chem., Int. Ed. Engl. 1996,35, 1123.
124. Adam, S.; Ellern, A.; Seppelt, K. Chem. Eur. J. 1996,2, 398.
125. Peterson, S. W.; Holloway, J . H.; Coyle, B. A,; Williams, J . M. Science 1971, 173,
1238.
126. IOselev, Yu. M.; Goryachenkov, S. A , ;Martynenko, L. I.; Spitsyn, V. I. Dokl. Acud.
Nuuk SSSR 1984,278, 881.
127. Kiselev, Yu. M.; Fadeeva, N. E.; Popov, A. I.; Korobov, M. V.; Nikulin, V. V.;
Spiptsyn, V. I. Dokl. Akud. Nuuk SSSR 1987,295, 378.
128. Spitzin, V. I.; Kiselev, Yu. M.; Fadeeva, N. E.; Popov, A. I.; Tchumaevsky, N. A.
2. Anorg. Allg. Chem. 1988, 559, 171.
129. Weidlein, J.; Muller, U.; Dehnicke, K. Schwingungsspektroscopie. Geog. Thieme
Verlag, Stuttgart, 1982.

130. Christe, K. 0.; Curtis, E. C.; Dixon, D. A,; Mercier, H. P.; Sanders, J. C. P.;
Schrobilgen, G. J. J. Am. Chem. SOC.1991, 113, 3351.
131. Baran, E. J . J . Mot. Struct. 1992,271, 327.
132. Schrobilgen, G. J.; Martin-Rovet, D.; Charpin, P.; Lance, M. J. J. Chem. SOC.,
Chem. Commun. 1980, 894.
133. Holloway, J. H.; KauEiE, V., Martin-Rovet, D.; Russell, D. R.; Schrobilgen, G. J . ;
Selig, H. lnorg. Chem. 1985,24, 678.
134. Christe, K. 0 . ;Dixon, D. A.; Sanders, J . C. P.; Schrobilgen, G. J., Tsai, S. S.;
Wilson, W. W. Inorg. Chem. 1995,34. 1868.
135. Ellern, A.; Seppelt, K. Angew. Chem., Int. Ed. Engl. 1995, 34, 1586.
136. Naumann, D.; Tyrra, W.; Gnann, R.; Pfolk, D.; Gilles, T.; Tebbe, K.-F. 2. Anorg.
Allg. Chem. 1997, 623, 1821.
137. Schack, C. J.; Christe, K. 0. J. Fluorine Chem. 1984,24, 467.
138. Schack, C. J., Christe, K. 0. J. Fluorine Chem. 1984,26, 19.
139. Schack, C. J., Christe, K. 0. J. Fluorine Chem. 1985,27, 53.
140. Schack, C. J.; Christe, K. 0. J. Fluorine Chem. 1988,39, 153.
141. Schack, C . J.; Christe, K. 0. J . Fluorine Chem. 1988,39, 163.
142. Lentz, D.; Pritzkow, H.; Seppelt, K. Inorg. Chem. 1978,17, 1926.
143. Lentz, D.; Seppelt, K. 2. Anorg. Allg. Chem. 1980, 460, 5.
144. Turowsky, L.; Seppelt, K. 2. Anorg. Allg. Chem. 1990,590, 37.
145. Mercier, H. P. A,; Sanders, J. C. P.; Schrobilgen, G. J . J . Am. Chem. SOC.1994,
116, 2921.
146. Keller, N.; Schrobilgen, G. J. Inorg. Chem. 1981,20, 2118.
147. Turowsky, L.; Seppelt, K. Inorg. Chem. 1990,29, 3226.
148. Lentz, D.; Seppelt, K. Angew Chem., Int. Ed. Engl. 1978, 17, 356.
149. Lentz, D.; Seppelt, K. Aizgew. Chem., Int. Ed. Engl. 1979, 18, 66.
150. Jacob, E.; Lentz, D.; Seppelt, K.; Simon, A. 2. Anorg. Chem. 1981, 472, 7.
151. Mercier, H. P. A.; Sanders, J. C. P.; Schrobilgen, G . J. Inorg. Chem. 1995,34, 5261.
152. Turowsky, L.; Seppelt, K. 2. Anorg. Allg. Chem. 1992, 609, 153.
153. Syvret, R. G.; Schrobilgen, G. J. J . Chem. SOC.,Chem. Commun. 1985, 1529.
154. Syvret, R. G.; Schrobilgen, G. J. Inorg. Chern. 1989, 28, 1564.
155. LeBlond, R. D.; DesMarteau, D. D. J . Chem. Soc., Chem. Commun. 1974, 555.
156. Sawyer, J. F.; Schrobilgen, G . J.; Sutherland, S. J. Inorg. Chem. 1982.21, 4064.
157. DesMarteau, D. D. J . Am. Chem. Soc. 1978, 100, 6270.
98
HOLLOWAY AND HOPE
158. DesMarteau, D. D.; LeBlond, R. D.; Hossain, S. F.; Nothe, D. J. Am. Chem. SOC.
1981, lo.?. 7734.
159. SchumP(,ier, G. A.; Schrobilgen, G. J. Znorg. Chem. 1983,22, 2178.
160. Faggiani, R.; Kennepohl, D. K.; Lock, C. J. L.; Schrobilgen, G. J. Znorg. Chem.
1986,25, 563.
161. Emara, A, A. A.; Schrobilgen, G. J. J. Chem. SOC.,Chem. Commun. 1987, 1644.
162. Emara, A. A. A,; Schrobilgen, G. J . J. Chem. SOC.,Chem. Commun. 1988, 257.
163. Schulz, A.; Klapotke, T. M. Znorg. Chem. 1997,36, 1929.
164. Turbini, L. J.; Aikman, R. E.; Lagow, R. J. J. Am. Chem. SOC.1979,101, 5833.
165. Naumann, D.; Tyrra, W. J. Chem. SOC.,Chem. Commun. 1989,47.
166. Frohn, H. J.; Jakobs, St. J. Chem. SOC.,Chem. Commun. 1989,625.
167. Frohn, H. J.; Jakobs, St.; Heukel. G. Angew. Chem., Znt. Ed. Engl. 1989,28, 1506.
168. (a) Frohn, H. J.; Rossbach, Chr. 2. Anorg. Allg. Chem. 1993, 619, 1672; (b) Naumann, D.; Butler, H.; Gnann, R.; Tyrra, W. Znorg. Chem. 1993, 32, 861; (c) Naumann, D.; Tyrra, W.; Pfolk, D. 2. Anorg. Allg. Chem. 1994, 620, 987; (d) Butler,
H.; Naumann, D.; Tyrra, W. Eur. J. Solid State Chem. 1992,29, 739.
169. Frohn, H. J.; Klose, A,; Bardin, V. V. J. Fluorine Chem. 1993, 64, 201.
170. Frohn, H. J.; Klose, A.; Heukel, G. Angew. Chem., Znt. Ed. Engl. 1993, 32, 99.
171. Frohn, H. J.; Bardin, V. V. J. Chem. SOC.,Chem. Commun. 1993, 1072.
172. Frohn, H. J.; Klose, A,; Bardin, V. V.; Kruppa, A. J.; Leshina, T. V. J. Fluorine
Chem. 1995, 70, 147.
173. Frohn, H. J.; Schroer, T.; Heukel, G. 2. Naturforsch. Ted B 1995,50, 1799.
174. Frohn, H. J.; Bardin, V. V. 2. Naturforsch. Teil B 1996, 51, 1011.
175. Frohn, H. J.; Bardin, V. V. 2. Anorg. Allg. Chem. 1996, 622, 2031.
176. Gilles, T.; Gnann, R.; Naumann, D.; Tebbe, K.-F. Acta Crystallogr. Sect. C. 1994,
50, 411.
177. Naumann, D.; Gnann, R.; Padelidakis, V.; Tyrra, W. J. Fluorine Chem. 1995,
72, 79.
178. Frohn, H. J.; Schroer, T. Abstract BA(2) C-4 in Abstracts of 15th International
Symposium on Fluorine Chemistry, Vancouver, Canada, 2-7 August, 1997.
179. Naumann, D.; Tyrra, W.; Gnann, R.; Pfolk, D. J. Chem. Soc., Chem. Commun.
1994,2651.
180. Naumann, D.; Tyrra, W.; Gnann, R.; Pfolk, D.; Gilles, T.; Tebbe, K.-F. 2. Anorg.
Allg. Chem. 1997, 623, 1821.
181. Zhdankin, V. V.; Stang, P. J.; Zefirov, N. S. J. Chem. Soc., Chem. Commun.
1992,578.
182. Naumann, D.; Gnann, R.; Ignatev, N.; Datsenko, S. 2. Anorg. Allg. Chem. 1995,
621, 411.
183. Hudlicky, M.; Pavlath, A. E. Chemistry of Organic Fluorine Compounds 11.
American Chemical Society, Washington, DC, 1995.
184. Cremer-Lober, B.; Butler, H.; Naumann, D.; Tyrra, W. 2. Anorg. Allg. Chem. 1992,
607, 34.
185. Lothian, A. P.; Ramsden, C. A. Synlett 1993, 753.
186. Tius, M. A.; Kawakami, J . K. Synlett 1993, 207.
187. Bardin, V. V.; Frohn, H. J . J. Fluorine Chem. 1993,60, 141.
188. Chuang, T. J. J. Appl. Phys. 1980,51, 2614.
189. Alan, K.; Janzen, A. F. J. Fluorine Chem. 1987,36, 179.
190. Forster, A. M.; Downs, A. J. Polyhedron 1985,4, 1625.
191. Lermontov, S. A.; Popov, A. V.; Zavorin, S. I.; Sukhojenko, I. I.; Kuryleva, N. V.;
Martynov, I. V.; Zefirov, N. S.; Stang, P. J . J. Fluorine Chem. 1994,66, 233.
99
192. Naumann, D.; Nowicki, G.; Sassen, K.-J. 2. An,org. Allg. Chem. 1997, 623, 1183.
193. Lewe, T.; Naumann D.; Nowicki, G.; Schneider, H.; Tyrra, W. 2. Anorg. Allg.
Ch,em. 1997, 623, 122.
194. Minkwitz, R.; Nowicki, G. Angew Chem., Znt. Ed. Engl. 1990,29, 688.
195. Patrick, T. B.; Nadji, S. J . Fluorine Chem. 1988, 39, 415.
196. Forster, A. M.; Downs, A. J. J. Chem. Sac., Dalton Trans. 1984, 2827.
197. Minkwitz, R.; Nowicki, G.; Preut, H. 2. Anorg. Allg. Chem. 1992, 611, 23.
198. Janzen, A. F.; Wang, P. M. C . ; Lenire, A. E. J . Fluorine Chem. 1983,22, 557.
199. Minkwitz, R.; Nowicki, J.; Jahnkow, B.; Koch, M. 2. Anorg. Allg. Chem. 1991,
596, 77.
200. Steinbeisser, H.; Mews, R. J. Fluorine Chem. 1981, 17, 505.
201. Janzen, A. F.; Ou, X. J . Fluorine Chem. 1995, 71, 205.
202. Poleschner, H.; Heydenreich, M.; Spindler K.; Haufe, G. Synthesis 1994, 1043.
203. Laahi, K. K.; Fiedler, W.; Regitz, M. J. Chem. Soc., Chem. Commun. 1997, 1641.
204. Alam, K.; Janzen, A. F. J . Fluorine Chem. 1985,27, 467.
205. Minoura, M.; Sagami, T.; Akiba, K.-Y.; Modrakowski, C.; Sudau, A,; Seppelt, K.;
Wallenhauer, S. Angew. Chem., Int. Ed. Engl. 1996,35, 2660.
206. Naumann, D.; Butler, H.; Fischer, J.; Hanke, J.; Mogias, J.; Wilkes, B. 2. Anorg.
Allg. Chem. 1992, 608, 69.
207. Ahmed, L.; Morrison, J . A. J. Am. Chem. SOC.1990, 112, 7411.
208. Hartl, H.; Nowicki, J.; Minkwitz, R. Angew Chern., Int. Ed. Engl. 1991, 30, 328.
209. Klapotke, T. M. Heteroatom Chem. 1997,8, 473.
210. Zhang, X.; Seppelt, K. 2. Anorg. Allg. Chern. 1997, 623, 491.
211. Bradley, G. W.; Holloway, J . H.; Koh, H. J.; Morris, D. G.; Watson, P. J . J. Chem.
SOC.,
Perkin Trans. 1 1992,3001.
212. Ebsworth, E. A. V.; Holloway, J . H.; Pilkington, N. J.; Rankin, D. W. H. Angew.
Chem., Int. Ed. Engl. 1984,23, 630.
213. Ebsworth, E. A. V.; Robertson, N.; Yellowlees, L. J. J . Chem. SOC.,
Dalton Trans.
1993, 1031.
214. Zheligovskaya, N. N.; Kiselev, Yu. M.; Krasovskaya, E. P. Koord. Khim. 1980,
6, 1080.
215. Drews, H. H.; Preetz, W. 2. Anorg. Allg. Chem. 1997, 623, 509.
216. (a) Bruce, D. M.; Holloway, J . H.; Russell, D. R. J . Chern. SOC.,
Chem. Cornmun.
1973, 321; (b) Hewitt, A. J.; Holloway, J . H.; Peacock, R. D.; Raynor, J . B.; Wilson,
I. L. J. Chem. Soc., Dalton Trans. 1976, 579; (c) Bruce, D. M.; Holloway, J . H.;
Russell, D. R. J. Chem. Soc., Dalton Trans. 1978, 1627.
217. Bruce, D. M.; Hewitt, A. J.; Holloway, J. H., Peacock, R. D.; Wilson, I. L. J . Chem.
Sac., Dalton Trans. 1976, 2230.
218. Misra, S. N. Indian J . Chem. 1979, 178, 101.
219. Coleman, K. S.; Holloway, J . H.; Hope, E. G. J. Chern. SOC.,
Dalton Trans. 1997,
1713.
220. Brewer, S. A,; Holloway, J . H.; Hope, E. G. J . Chem. Soc., Dalton Trans. 1994,
1067.
221. Brewer, S. A,; Coleman, K. S.; Fawcett, J.; Holloway, J. H.; Hope, E. G.; Russell,
D. R.; Watson, P. G. J. Chem. Soc., Dalton Trans. 1995, 1073.
222. Brewer, S. A.; Brisdon, A. K.; Holloway, J. H.; Hope, E. G.; Peck, L. A.; Watson,
P. G. J. Chem. Soc., Dalton Trans. 1995, 2943.
223. Butin, K. P.; Kiselev, Yu. M.; Magdesieva, T. V.; Reutov, 0. A,, J . Organomet.
Chem. 1982,235, 127.
100
HOLLOWAY AND HOPE
224. Kiselev, Yu. M.; Goryachenkov, S. A.; Martynenko, L. I. Russ. J . Znorg. Chem.
(Engl. Transl.) 1984,29, 38.
225. Cook, R. L.; Woods, M.; Sullivan, J. C.; Appelman, E. H. Znorg. Chem. 1989, 28,
3349.
226. Fields, P. R.; Stein, L.; Zirin, M. H. J. Am. Chem. SOC.1962,84, 4164.
227. Stein, L. J. Am. Chem. Soc. 1969,91, 5396.
228. Stein, L. Yale Scientific Magazine 1970, 44, (8), 2.
229. Stein, L. Science 1970, 168, 362.
230. Stein, L. U.S.P. 3 660 300, 197.
231. Stein, L. Science 1972, 175, 1463.
232. Stein, L. J. Znorg. Nucl. Chem. 1973,35, 39.
233. Pitzer, K. S. J . Chem. Soc., Chem. Commun. 1975, 760.
234. Stein, L. J. Chem. Soc., Chem. Commun. 1985, 1631.
235. Stein, L. Znorg. Chem. 1984,23, 3670.
236. Stein, L. Nature (London) 1973,243, 30.
237. Stein, L.; Hohorst, F. A. Enuiron. Sci. Technol. 1982, 16, (7), 419.
238. Nefedov, V. D.; Torpova, M. A.; Avrorin, V. V.; Dudkin, B. N. Radiokhimiya 1971,
13, 916.
239. Avrorin, V. V.; Nefedov, V. D.; Toropova, M. A. Radiokhimiya 1974, 16, (21, 261.
240. Avrorin, V. V.; Krasikova, R. N.; Nefedov, V. D., Toropova, M. A. Radiokhimiya
1981, 23, (6), 879.
241. Avrorin, V. V.; Krasikova, R. N.; Nefedov, V. D.; Toropova, M. A. Radiokhimiya
1986,27, (4), 511.
1987,29, (31, 426.
1989,31, (21, 124.
1989,31, (61, 63.
ADVANCES IN INORGANIC CHEMISTRY, VOL. 46
COMING TO GRIPS WITH

REACTIVE INTERMEDIATES
ANTHONY J. DOWNS and TIMOTHY M. GREENE
Inorganic Chemistry Laboratory, University of Oxford, Oxford OX1 3QR, United Kingdom
It i s a capital mistake to theorise before one has data. Inseizsibly one begins to twist facts to suit theories, instead of theories to suit facts.
A. Conan Doyle, A Scandal in Bohemia
I. Introduction
II. Reaction Intermediates: Nerve Centers of Chemical Reactions
111. Experimental Characterization of Reaction Intermediates: Retardation
A. Gas-Phase Studies a t Low Pressure
B. Gas-Phase Studies in Supersonic Jets
C. Trapping and Matrix Isolation
D. Solution Studies a t Low Temperatures
IV. Experimental Characterization of Reaction Intermediates: Time-Resolved
Methods
V. Experimental Characterization of Reaction Intermediates: Flow and Other
Methods
VI. Conclusions
References
I . Introduction
Chemistry is the science of change. It is ironic then that, although

we have become so knowledgeable about the properties of reactants
and products-about
the beginnings and ends of chemical journeys-we remain more often than not unsure of just how those journeys are made. Our modern armory of sophisticated spectroscopic and
diffraction methods provides every expectation of a direct hit on the
fixed objects represented by reactants and products, but is seldom
101
Copyright ID1999 by Academic Presa
All rights of reproductiun in any form reserved
O R ~ X - H ~ ~$25
R Moo
~
102
DOWNS AND GREENE
equipped to target the shadowy, moving entity presented by the process of their interconversion. How the reaction rate varies with concentration, temperature, and other conditions may give vital clues but
rarely establishes with any certainty the mechanism of what may be,
at least superficially, quite a simple homogeneous reaction.
Elaborating the mechanism of a reaction is an altogether more complicated undertaking than determining the structures of long-lived reagents or products in that we are seeking ultimately a detailed description of the way in which the structure and bonding of the
reagents change with time in each of the several individual acts that
normally make up a chemical change. In general terms, a complete
account of any such mechanism is going to require knowledge of no
less than four aspects (1-3):
1. subdivision of the reaction into its individual steps or equilibria;
2. characterization of each intermediate species in terms of its composition, structure, energy states, and lifetime;
3. a description of the transition state appropriate t o each elementary reaction step, again with reference to composition, stereochemistry, and energetics; and
4. a complete specification of the processes leading to and from
each transition state in relation to the geometries and energy levels
(mainly electronic and vibrational) of the reactants, intermediates,
and products in their ground and excited states.
With reactions occurring as they usually do in solution, the role of
the solvent at each stage needs to be realistically assessed. The history of kinetic studies is littered with the wrecks of mechanistic
hypotheses that ultimately foundered on the unsuspected reefs of solvent mediation. It can come as no surprise, then, that a comprehensive mechanistic description is still well beyond the reach of present
techniques, both practical and theoretical, except in a few cases, and
those mostly confined to the simplest of systems. More often than not
we are obliged to piece together a mechanism that is consistent with
all the available facts, both kinetic and nonkinetic; very rarely do we
have compelling and unambiguous evidence of one particular mechanism.
Nothing could be much simpler, it might be thought, than the exchange that the neutral metal hexacarbonyl CI-(CO)~undergoes with
free CO in both the solution and gas phases, particularly as it proceeds in accordance with a rate law of the form
COMING TO GRIPS WITH REACTIVE INTERMEDIATES
103
The solvent having little effect on the activation parameters, it is reasonable to assume that a dissociative mechanism is a t work, with
rate-determining fission of a Cr-CO bond being followed by rapid attack a t the resulting pentacoordinated intermediate Cr(CO), ( 4 - 6 ) .
Early kinetic studies on substitution of CO by another two-electron
ligand L,
Cr(CO16+ L
Cr(C0)SL + CO
(1)
e.g., L = amine, phosphine, or phosphite,

appeared also ( 4 - 6 ) to point to a simple dissociative mechanism,
again with Cr(CO), as a common intermediate. However, investigations extending over a much wider concentration range have now
shown that a two-term rate law is commonly applicable, uzz.
The activation parameters AH$ and AS$ characterizing the kl term

are very similar to those associated with CO exchange, whereas AH2$
for the k 2 term is significantly smaller than AHl$, and AS2$, being
negative, differs not only in magnitude but also in sign from A S l $
The most plausible interpretation of these and related experiments is
that substitution occurs via two competing pathways, one involving a
dissociative mechanism closely akin to that envisaged for CO exchange (see Fig. 11, and the other a dissociative interchange (Id)
Reaction coordinate
FIG.1. Hypothetical free energy profile for the substitution reaction Cr(CO)6+ L +
Cr(CO)SL + CO on the assumption that it proceeds via a dissociative mechanism.
104
DOWNS AND GREENE
process in which there is relatively little bond-making in the transition state. How then are we to picture this Id process? Perhaps L
forms encounter complexes in which it occupies an outer-sphere site
of a solvated Cr(CO)5fragment. There are other possible explanations,
though, and we are now on the "tossing sea" of conjecture, however
well informed, rather than the firm ground of incontestable result or
inference.
Still more problematic is the apparent migratory insertion of nitric
oxide into transition metal-carbon bonds, an important reaction in
metal nitrosyl complexes and one that may be relevant to biochemical
reactions (7). On the evidence of isotopic labeling and kinetic experiments, the insertion of NO into the Co-CH3 bond of the (cyclopentadieny1)cobalt complex ( v5-C5H5)Co(NO)(CH3),
which occurs in Reaction ( 3 ) ,
is intramolecular, and the experimental results are consistent with,

but do not prove, a mechanism involving methyl migration, followed
by phosphine addition (see Scheme 1, lower path) (8).This has the
attraction of resembling the mechanism generally agreed to hold for
CO insertion in a reaction such as
MII(CO)~CH~
+ CO + Mn(CO)5[C(0)CH31
(4)
and which is a critical step in many important carbon-carbon bondforming processes mediated by homogeneous transition-metal cata-
04"
L-PR,
SCHEME
1. Alternative pathways for the migratory insertion reaction of (q5-C5H6)Co(NO)(CH3)in the presence of a phosphine (8).
105
lysts. Still, the worm of doubt may turn. For is not NO a more versatile ligand than CO, and can it not, for example, function as either a
3e- ligand or a le- ligand? Thus, addition of a phosphine molecule
accommodated by a change in the ligation of the NO to give a bent
Co-N-0 unit offers another reaction pathway, as shown by the upper
route in Scheme 1. There is persuasive evidence that such a change
affords a low-energy channel in other reactions (81,and it is not obvious why it should not do so to effect NO insertion. In the absence of
any definitive experimental information on this score, quantum chemical calculations have been carried out to explore each of the two possible pathways ( 8 ) . The geometries and energies of the reactants, intermediates, transition states, and products have been determined on
the basis of Density Functional Theory (DFT), with the results illustrated in Fig. 2. Hence it emerges that the mechanism in which NO
insertion precedes phosphine association incurs activation barriers
not exceeding 84 kJ.mol-' and is therefore favored over the mechanism involving insertion after phosphine addition, which is opposed
by appreciably higher barriers (80-218 kJ.mol-'1. This pleasing endorsement by theory of the mechanism deduced by experiment may
seem to settle the issue, but the theory has not been able to budget
for the effects of solvation. Still, we cannot be entirely sure of our
ground, and about this particular reaction the last word has surely
not been said.
The preceding examples of CO exchange and NO insertion reactions
have to do with systems that have attracted considerable experimental and theoretical attention. Accordingly, although there are fundamental questions still to be answered, some mechanistic features at
least are not in doubt. For most of the myriad chemical reactions that
FIG. 2. The calculated profiles of the potential energy surfaces (at the DFT-B3LYP
level) for the insertion reaction ( q5-C5H5)Co(NO)(CH3)
+ phosphine ($-C5H,)Co(phosphine)[N(0)CHplon the assumption (a) of an intramolecular rate-determining mechanism, or (b) of an associative mechanism (8).
--.f
106
DOWNS AND GREENE
are known to occur, though, we have neither the facts nor the theory
to argue the mechanistic case, and analogy and intuition may well be
the sole guiding principles. Hence we may be spared the great tragedy
of science-that is, the slaying of a beautiful hypothesis by an ugly
fact-but, without a secure understanding of mechanism, we are
hostages to experience, empiricism, and prejudice when it comes to
devising the most efficient, safe, and reliable way of engineering a
desired chemical change.
II. Reaction Intermediates: Nerve Centers of Chemical Reactions
Crucial to any problem of mechanism must be the part played by

any intermediate compound, for example (i) CI-(CO)~
in the exchange
reactions of Cr(CO)5L (L = CO or other ligands) and (ii) either
in Reaction (3).
(C5H5)Co[N(0)CH3]
or (C5H5)Co(phosphine)(NO)(CH3)
Indeed, were it possible to detect, identify, and characterize the relevant intermediates, the primary issue of which path is taken in Reaction (3) (see Scheme 1) could be settled beyond peradventure. More
generally, the intermediates may be open-shell molecules; they may
be radicals; they may involve elements in unusual oxidation states or
with unusual coordination geometries or partners. However, until we
know what these intermediates are and the part they play, we cannot
begin to comprehend the mechanism of any given reaction. As a start,
we need to be sure of not only the identities but also the molecular
and electronic structures of the intermediates. Such properties cannot
necessarily be deduced from the properties of stable molecules and
may well challenge conventional principles of bonding. This is all very
well, but how are we to cope with intermediates that have but a fleeting existence under normal conditions? At a vapor pressure of 1 torr
and in the presence of CO, for example, Cr(COI5is found to have a
lifetime in the order of one-millionth of a second at room temperature (9).
To counter the all too brief stay normally enjoyed by a molecule
such as Cr(CO)5,the experimenter has three main options. One way
is to slow down or inhibit the reactions disposing of the molecule.
Another is to record a spectroscopic autograph, in absorption or emission, within the lifetime of the transient and so monitor its temporal
fate (as in flash photolysis). The third is to resort to a flow method,
depending on continuous generation of the intermediate in a fluid
traveling at a uniform rate along a tube; under the right conditions,
decay of the intermediate gives rise to a steady-state concentration
107
that varies as a function of the distance traveled along the tube, and
that may be analyzed by essentially static spectroscopic measurements. Each method has its strengths and weaknesses. None is able
individually to deliver all the information we are likely to require for
a proper appreciation of how the intermediate fits into the overall
mechanism of a given reaction. If we are to realize that ambition, we
must be looking therefore to exploit not one but several methods.
To gain a better idea of the three strategies for characterizing
short-lived intermediates-namely,
retardation, time-resolved, or
flow methods-we turn next to a brief survey of how they work in
practice and of just what light they can or cannot shed on a particular
intermediate. How mechanistic studies have been advanced in this
way will be illustrated in the course of the account.
HI. Experimental Characterization of Reaction Intermediates: Retardation
There are various ways of slowing down or even suppressing totally

the reactions of a molecule. In principle, we may limit its collisions
with other molecules by confining it to the gas phase and there minimizing its pressure; we may minimize the thermal energy to which
it has access by lowering its temperature; or we may trap it with
a suitable substrate. The best policy in practice may well be either
simultaneously or in separate experiments to apply more than one of
these restraints.
A. GAS-PHASE
STUDIES
AT Low PRESSURE
The gas phase at low pressures is the best medium for detailed
structural and spectroscopic studies of very simple, robustly bound
molecules such as AIH, SiO, or C10 (10). Generated typically by
thermal or discharge reactions, these may survive a t partial pressures high enough and for times long enough to be interrogated by
their electronic emission, laser-induced fluorescence, microwave, or
infrared absorption spectra, such studies being greatly facilitated
by the development of tunable lasers and double resonance techniques. Hence an immense effort has been invested over the years in
measuring and analyzing the electronic spectra of stable and unstable
diatomic molecules, to give a rich return on information about the
bond lengths and vibrational and electronic properties that characterize not only their electronic ground states but also their excited
states.
108
DOWNS AND GREENE
A good example is the molecule SiO, a cosmic precursor to silicon

oxide chemistry as we now know it on earth (11).In more recent
times, too, this molecule has assumed considerable significance because of its relevance to oxidation reactions taking place at the surfaces of silicon wafers and to the creation of antireflection coatings on
these and other solid-state devices. Unlike its more familiar counterpart carbon monoxide, SiO is normally quick to aggregate and disproportionate [Reaction (5)] at temperatures below 1000C:
2SiO
Si + SiO,.
(5)
Stability is not a problem at high temperatures, though, and Reaction

(5) can be reversed to produce SiO vapor at a concentration sufficient
for microwave absorption measurements by heating a homogeneous
mixture of silicon and silicon dioxide to 1350C in a ceramic reaction
vessel, which also serves as the spectroscopic sample chamber (12).
Alternatively, excited SiO molecules, produced by the action of a discharge on a gaseous mixture including, say, Sic&and O,, can be characterized by their electronic emission spectrum. It has taken 10 to 20
years to unravel all the features of the electronic spectrum extending
from the near- to the vacuum-ultraviolet and assign the different
band systems, but we now have a relatively full knowledge of the
molecule in its electronic ground and numerous excited states ( 1 3 ) .
Rather easier to decipher is the microwave spectrum of the molecule
in its electronic ground state, whence a wide range of vibrational and
rotational states have been accurately detailed (111. These and other
studies give every confidence that we are dealing with a z*Si160
molecule having the following equilibrium parameters: re = 150.9739 pm,
we = 1241.56 cm-', and pe = 3.0882 D (13). Some of the microwave
lines have served as distinctive signatures by which SiO has been
recognized in astronomical sources, and SiO masers have relayed important information about both the conditions in circumstellar shells
and the mechanisms of mass loss.
Solid boron nitride is a ceramic material of some consequence that
can be formed at high temperatures by the reaction of boron atoms
with Nz or NH3. A likely intermediary in its formation is another
high-temperature molecule BN, first observed through its electronic
emission spectrum nearly 60 years ago but the properties of which
have emerged only slowly (14). Identification of the several low-lying
electronic states of the molecule is a major problem and only very
recently has it become clear that the ground state is not
(as with
the isoelectronic Cz molecule) but 311 (15).The A311,-X311itransition,
lz+
109
the equivalent of the celebrated Swann bands of C,, can be excited in

emission by the action of a microwave discharge on traces of BCls and
N2 entrained in a helium gas stream. Analysis of the well-resolved
rotational lines of the 0-0 and 1-0 bands yields the following properties for the ground state of llB1*N:re = 132.9 pm and o,= 1514.6 cm-
(13, 16).
The molecule C10 is another reactive species that has attracted still
more attention. It plays a significant role in the destruction of stratospheric ozone through catalytic cycles involving heterogeneous and
homogeneous reactions, as in Eq. (6) ( 1 7 ) :
c1 f 0:3+ c10 + 0 2
(6)
C10+0-+C1+0~.
Current wisdom can be summarized in the flow chart of Fig. 3. More

specifically, the C10 radical is involved in several possible kinetic
mechanisms linking global release of chlorofluorocarbons (CFCs) to
the Antarctic ozone loss during each austral spring ( 18).An accurate
characterization of the spectroscopic and other properties of C10 is
therefore vital, not least as a prelude to tracing the crucial correlation
between CFCs, C10, and 0,. C10 is short-lived a t the high molecular
concentrations characteristic of the condensed phases but can be generated a t low pressures in the gas phase, for example by the action of
nv
LIU
FIG. 3. Chemical cycles affecting the formation and decay of chlorine oxide trace
species in the earths atmosphere (reproduced with permission from Wayne, R. P.
Chemistry of Atmospheres, 2nd ed., p. 137, Clarendon Press: Oxford, 1991).
110
DOWNS AND GREENE
a microwave discharge or a short burst of radiation on a mixture of

Clz and 02.On the evidence of a detailed analysis of the A211i-X211i
band system as measured both in absorption and in emission at high
resolution, of the infrared spectrum also measured at high resolution,
and of the microwave spectrum, the 35Cl'60molecule has a 211 ground
state with re = 156.960 pm, we = 853.724 cm-l, pe = 1.24 D, and
D;= 265.4 k J * mol-l(I3, 19). Of its reactivity, though, the properties
give the barest hint.
Simple polyatomic molecules such as BH2 (201, BH, (211, and
HBNH (22) may be no less amenable to high-resolution spectroscopic
studies. Commonly invoked as an intermediate in reactions of boron
hydrides, the BH3 molecule is made elusive by its rapid dimerization
to diborane. It is produced, nevertheless, by the action of a dc discharge on a gas stream composed of diborane and helium flowing
through a multipass absorption cell; here it can be recognised by its
vz (a;) and v3 ( e ' ) vibrational bands as recorded with a high-resolution
FT-IR spectrometer (21). Detailed analysis of the rotational structure
associated with the bands fixes a regular planar DBhstructure for BH,
in its ground electronic state with r&B-H) = 119.001 pm. Another
discharge reaction, this time involving diborane and ammonia, generates the species HBNH with a lifetime of a few hundred ms under
normal conditions. That this is a linear molecule is made clear by the
rotational structure of the v3 band (corresponding to the BN stretching fundamental) exposed by an IR diode laser spectrometer (22);the
observed rotational constants imply a BN bond length (123.81 pm)
appreciably shorter than those in H3BNH3,HzBNHz,and even BN
itself (165.76, 140.3, and 132.9 pm, respectively).
In principle, electron diffraction is another tool capable of giving
precise information about the structure of a gaseous intermediate,
even when this is quite a complicated molecule (23). In practice,
though, there are numerous complications liable to militate against
electron diffraction as the sole agency of characterization. Chief
among these is the low resolution of the electron-scattering pattern of
a molecule. As a result it is seldom easy to discover much about the
structure-sometimes even the identity-of a specific molecule in a
gaseous mixture. To resolve an otherwise severely underdetermined
problem, we need independent testimony about the molecules making
up the mixture, for example by reference to the mass, rotational, or
vibrational spectrum of the sample or to the results of quantum chemical calculations (24). A good example of this strategy is provided by
a careful study of the thermal decomposition of trimethylamine vapor
(25).Mass spectrometric analysis of the vapor stream shows that de-
111
composition proceeds at 515C primarily in accordance with Eq. (7) to

give methane and the short-lived compound CH3-N=CH2:
Analysis of the electron-diffraction pattern of the reaction mixture

implies the following mole fractions: (CH&N, 0.63; CHI, 0.18; and
CH3-N=CH2, 0.19. The structure of the CH3-N=CH2 molecule has
then been determined by a joint analysis of the electron-diffraction
pattern and the rotational constants (derived from the microwave
spectrum), giving parameters [rZ(N=C)= 127.9 pm, rZ(N-C) = 145.8
pm, and LC-N=C = 116.6"l very close to those calcylated by ab
znitio methods. Similarly, pyrolysis of propylenimine, HNCH2CHCH3,
a t 470C results in complete decomposition to give the short-lived species syn- and antz-CH3-NH-CH=CH2 and truns-CH,-N=CHCH, (26);geometrical structures and mole fractions of all three products have been deduced by analyzing the electron diffraction pattern
of the vapor in concert with the properties implied by quantum chemical calculations. How the proportions of the transients vary with temperature sheds significant light on the reaction path taken by the
thermal rearrangement. Life is altogether simpler with only one molecular species in the vapor, and electron diffraction is well established as a primary method of characterizing high-temperature molecules such as MgC12 (27) or C,, (28). In general, however, the
optimum conditions can be met only for a relatively long-lived intermediate such as S2F4.This is formed by the dimerization of SF, on
the winding reaction path that starts with SClz and a metal fluoride
and affords the best known route to SF4, represented mainly by the
overall reaction,
3SC12 + 4MF + SF4 + S&1,
(e.g., M
+ 4MC1.
(8)
Na or K)
The labile molecule has required the combined forces of electron diffraction, microwave, vibrational, and "F NMR spectroscopy, and ab
initio calculations before yielding the secrets of its remarkable, unsymmetrical structure, I (29);this can be viewed as a trigonal bipyramid centered on one sulfur atom with a lone pair of electrons, the SF
group, and one fluorine atom occupying the equatorial sites. Electrondiffraction studies have also contributed to our understanding of how
112
DOWNS AND GREENE
gaseous organometallic compounds can, through their decomposition,

furnish solid materials with useful electronic or other properties. For
example, thermal decomposition of t-butylgallium sulfide, [(But)GaS14,
results in the growth of a new cubic phase of Gas. Not only is this
material ideal for the electronic passivation of GaAs surfaces, it has
a large band gap, which makes it suitable as the insulating gate
layer in certain transistor devices. What is striking about the structure of the gaseous precursor, 11, as determined by electron diffraction, is the presence of a distorted cubane-like Gads4core with a
t-butyl group bound to each gallium atom and T symmetry overall
(30).Irradiation of the precursor with the output of a UV excimer
laser gives rise to the photofragments (But),Ga4S4(where x = 0-31,
which can be detected by their time-of-flight mass spectra following
ionization by a second UV excimer laser. It appears, then, that fragmentation proceeds primarily by loss of the organic substituents
with retention of the Ga4S4core, clearly suggesting that growth of
cubic Gas depends on the oligomerization of preformed Ga4S4fragments.
Problems of analysis apart, experiments of this sort can at their

best provide a remarkably full picture of a gaseous molecule, at least
with regard to its physical character (10). By deliberately seeking to
113
frustrate any chemical reactions, however, we are naturally denying

ourselves any detailed chemical and, in particular, kinetic information. There are other limitations, too. The conditions are far removed
in terms of molecular concentrations, as well as environment, from
those normally prevailing in the condensed phases. For reasons of
rotational congestion, moreover, high-resolution spectroscopic measurements cannot usefully be applied to molecules containing more
than a few atoms, and for reasons of pressure and temperature such
experiments may lose sight of molecules incorporating weak bonds.
In fact, weakly bound molecules commonly feature as the initial products when two reagent molecules interact. These problems lead naturally t o the question, How can the power and precision of modern
high-resolution spectroscopic methods be brought to bear on larger
andlor weakly bound transients? Fast freezing in one form or another
is the obvious response.
B. GAS-PHASE
STUDIES
IN SUPERSONIC
JETS
When a noble gas seeded with a small amount of other molecules
expands through a nozzle into a vacuum, rapid adiabatic cooling occurs, provided that the mean free path of the molecules in the unexpanded gas is short compared with the nozzle diameter. Within only
a few nozzle diameters (usually ca. 5 mm), the molecules in the jet of
gas emerging from the nozzle achieve a very narrow distribution of
speeds in the jet direction, clustered around the supersonic value of
ca. 5 x lo4 cm . s-l, and have nearly zero relative speed in directions
transverse to this. Consequently, within 5 to 10 p s of emerging from
the nozzle into the vacuum, the molecules are in an essentially collisionless state and have a velocity distribution in the transverse direction corresponding to temperatures of a few Kelvin. A similar effective
temperature is also achieved through cooling of the rotational and
vibrational degrees of freedom. If we can arrange for the intermediate
to be formed within the noble-gas stream just before it emerges from
the nozzle, freezing times in the order of microseconds can be attained, and thereafter the intermediate is effectively isolated, having
no mechanism, either unimolecular or bimolecular, by which to react
or decompose. One experimental arrangement, illustrated in Fig. 4,
consists of two concentric, coterminal tubes (31-33). One reagent admixed with an excess of noble gas is delivered in short pulses via the
outer tube while the second reagent (typically also diluted with noble
gas) flows continuously through the central tube. The two reagents
114
DOWNS AND GREENE
Series 9 solenoid
valve
Teflon outer tube

pulse of F2-Ar from
solenoid valve
0.3 mrn 1.d.
glass capillary
FIG.4. The fast-mixing nozzle used to observe the rotational spectrum of the gaseous
complex H,N----F,[reproduced with permission from (33),p. 1121.
meet and mix in the roughly cylindrical interface between the concentric flows as they simultaneously expand. Any intermediate formed
here rapidly undergoes collisionless expansion, is then effectively frozen, and may be investigated by such techniques as microwave, infrared diode-laser, laser-induced fluorescence, or photoionization spectroscopy.
For characterizing a dipolar molecule in its electronic ground state,
few methods are more instructive than pulsed-nozzle Fourier-transform microwave spectroscopy (32). As illustrated schematically in
Fig. 5, a short pulse of microwave radiation directed at the gas pulse
excites a rotational transition in the species of interest; subsequently
the rotationally excited molecules reemit radiation, which is detected.
This technique provides a remarkably sensitive probe for transients,
the properties of which can be specified with all the precision and
detail peculiar to rotational spectroscopy only microseconds after
their production. In relation to a weakly bound adduct A----Bformed
by two molecular reagents A and B, for example, we may draw on the
rotational spectrum to determine such salient molecular properties as
symmetry, radial and angular geometry, the intermolecular stretching force constant and internal dynamics, the electric charge distribution, and the electric dipole and quadrupole moments of A----B(see
Table I).
The mechanism of the addition of molecular chlorine to an alkene
115
Pulsed nozzle
Fabry-Plrot mirrors
FIG. 5. Pulsed-nozzle FT microwave measurements. A molecule-radiation interaction occurs when the gas pulse is between mirrors forming a Fabry-Perot cavity. If the
transient molecule has a rotational transition of frequency v,,, falling within the narrow
band of frequencies carried into the cavity by a short pulse (ca. 1p s ) of monochromatic
radiation of frequency v, rotational excitation leads to a macroscopic electric polarization of the gas. This electric polarization decays only slowly (half-life T2ii: 100 p s )
compared with the relatively intense exciting pulse (half-life in the cavity T~ = 0.1 ps).
If detection is delayed until ca. 2 ps after the polarization, the exciting pulse has diminished in intensity by a factor of ca. lo6 but the spontaneous coherent emission from the
polarized gas is just beginning. This weak emission can then be detected in the absence
of background radiation with high sensitivity. For technical reasons, the molecular
emission at v, is mixed with some of the exciting radiation v and detected as a signal
proportional to the amplitude of the oscillating electric vector a t the beat frequency
0 - I),,, as a function of time, as in NMR spectroscopy; Fourier transformation leads to
the frequency spectrum [reproduced with permission from (31),p. 5631.
has been the subject of much investigation and discussion ( 3 4 ) .In a

fairly polar solvent and in the dark, the reaction has been shown to
be of first order in both components, indicating that the entity responsible for electrophilic attack on the alkene is the dihalogen molecule.
There is a general consensus that the mechanism, as represented by
Eq. (9), involves a pre-equilibrium molecular association of the alkene
and C12to give the adduct 111:
alkene
+ C12+ alkene----C12+ alkeneC1' + C1-.

I11
Iv
(9)
116
DOWNS AND GREENE
TABLE I
MEASURABLE
SPECTROSCOPIC
AND MOLECULAR
PROPERTIES
OF GASEOUS
A-B MOLECULES
(32)
ACCESSIBLE
THROUGH
GROUND-STATE
ROTATIONAL
SPECTRA
Spectroscopic parameter
Molecular property of A----B
Form of the spectrum
Symmetry
Rotational constants,
-40,
&, GI
Quantitative determination
of radial and angular geometry; nature of intermolecular binding
Quadratic force constant for
A----Bstretching, k,,
Relate to electric field
gradient qi, a t a nucleus
X having a nonzero electric quadrupole moment
Electric dipole moment of
A....B
Molecular g factor and electric quadrupole moment
of A----B
Centrifugal distortion
constant, DJor AJ
Hyperfine coupling constants, x,, and D,,
Stark effect
Zeeman effect
Comment/example
Spectral pattern varies for
linear, symmetric-top,
and asymmetric-top species, e.g., F2----NH3
is a
symmetric top (35).
(34) and
e.g., CzH4----C12
F,----NH,(35)
DJis roughly proportional

t o k,,.
qig depends on the detailed
electric charge distribution within A----B.

(36)
e.g., HCsNN----H6Cl
e.g., Ar----HCI(37)
This intermediate then ionizes in the rate-controlling step to give a

cation, IV,which reacts rapidly with Cl-. For many years, however,
the precise nature of the pre-equilibrium complex I11 has been a matter of speculation. It is clearly desirable to establish with certainty
and precision the existence of any such complex. Mixing ethene and
molecular chlorine in an apparatus of the type described gives rise to
a new ground-state rotational spectrum attributable to the complex
CzH4----Cl2.
Analysis of the spectrum indicates that this complex has
the Czugeometry illustrated in Fig. 6(a), with the Clz molecule lying
along the C2 axis of ethene that is perpendicular to the molecular
plane. The weakness of the interaction between the molecules is made
abundantly clear by the following features (34): the rotational constant A, is only slightly larger than the constant C, of free ethene; the
inner C1 atom is relatively remote from the center of the ethene C=C
bond (estimated distance 312.8 pm); at 5.9 N m-l the intermolecular
stretching force constant, k,,,deduced from the centrifugal distortion
117
constant is quite minuscule; and the nuclear quadrupole coupling constants xgg(35Cl)
undergo quite minor changes, implying that the cylindrical symmetry of the electric charge distribution of Clz is only
slightly perturbed on complex formation. Hence, there can be little
doubt that the intermediate is formally a b7i.w complex of the weak,
outer type, according to Mulliken's classification.
More complicated and less easily controlled is the reaction between
ammonia and fluorine gases, which under normal conditions of temperature and pressure results in a spectacular flame and the formation of NF, in small yield. Despite the prodigious reactivity of molecular fluorine, it has proved possible through the device of fast mixing
and cooling of the gases in a supersonic jet to detect and characterize
the weakly bound prereactive intermediate H3N----F2on the basis
of its ground-state rotational spectrum (35). The results attest to
the C3rstructure illustrated in Fig. 6(b) with a long intermolecular
contact, r(N----F),of 270.8 pm and a stretching force constant k,,of
no more than 4.7 N.m-'. By contrast, the corresponding complex
H,N----ClFfeatures an appreciably stronger intermolecular bond, with
r(N----Cl)= 237.6 pm and k,,= 34.3 Nem-', pointing to a small but
significant transfer of charge in the sense H3N----Cl+F(38).
It is not just loosely bound complexes that can be specified with
profit by spectroscopic analysis of a supersonic jet. Of this there is no
more thrilling example than the transient organometallic radical
VCH (39),formed when vanadium atoms generated by laser ablation
are entrained in a pulse of high-pressure helium containing 5-10%
CH, prior to expansion to give a supersonic, cooled molecular beam.
High-resolution studies of the molecular fluorescence near 800 nm
excited by a tunable probe laser reveal extensive vibrational and
c'
P-------
FIG.6. Structures of the gaseous prereactive intermediates (a) C2H4----C12

and (b)
H:jN----F,.
118
DOWNS AND GREENE
rotational detail. Hyperfine patterns characteristic of a nucleus

with spin I = 7/2 confirm that the carrier contains a vanadium atom;
the subband structures are those of a linear molecule; the integer
rotational quantum numbers point to an even number of unpaired
electrons; and the presence of hydrogen is borne out by experiments
with CD4, which give rise to the corresponding bands of VCD.
The ground state of the molecule can be identified unambiguously
as 3A, and the associated rotational constants imply the following
dimensions for the substitution structure: rs(V=C) = 170.25 and
r,(C-H) = 108.0 pm. Hence, we now have substantive experimental
data for an exemplar of the simplest possible type of metal carbyne
complex, making clear inter alia that the V=C stretching mode is at
838 cm-', and not near 1300 cm-l as has been suggested for related
species.
Even more challenging are some of the neutral and charged species
that are created and destroyed within an electric discharge. Yet, by
subjecting a pulsed molecular beam to the combined actions of supersonic expansion and a corona discharge (on the downstream side of
the nozzle), it has been possible to detect and study short-lived intermediates as varied as OH, NB , and CH3(40). Optical spectroscopy has
been the traditional source of information, but major contributions to
our limited knowledge of the highly complex chemistry of plasmas
have come recently from mass, microwave, and infrared laser spectroscopies. Lifetime is not the only problem, for even within the discharge the concentrations of the relevant species are often very small.
This has stimulated the development of special modulation techniques to improve both selectivity and sensitivity (40). Use of an ac
discharge gives, for example, Doppler shifts of spectroscopic frequencies due to molecular ions, which vary with the discharge frequency
vf, and state populations of any transient species, which vary with
twice this frequency. When the spectroscopic detection system is referenced to the frequency ur or 2 u f , the resulting demodulated signal
consists only of what are velocity-modulated or population-modulated
lines, respectively.
Transients such as OH and CH3are no mere scientific curiositieswill-o'-the-wisps peculiar t o the elaborate combination of discharge,
supersonic expansion, and spectroscopic subterfuge. Some of these
molecules play vital roles in regimes seemingly quite remote from
plasma chemistry. For example, the OH radical is crucially significant
as an oxidizing initiator in combustion and in the troposphere, where
its natural concentration is liable to be seriously augmented as a re-
119
sult of pollution (17). Two lines of investigation must suffice to demonstrate the scope of modern discharge experiments.
1. The molecules HCCCS and HCCCCS have been characterized by
their microwave spectra following their generation by the action of a
pulsed discharge on a supersonic molecular beam of argon containing
traces of C2H2and CS2(411. The microwave frequency, swept in small
steps of a few MHz, is fed to a Fabry-Perot cavity, which is maintained at the resonance frequency by synchronous length adjustment
so as to retain the high Q value of the cavity. This feature, combined
with cooling t o rotational temperatures near 1-2 K, enhances greatly
the sensitivity of the experiment. The resulting high-resolution spectra are noteworthy for their disclosure that both HCCCS and
HCCCCS, in common with HCCS, are linear in their electronic
ground states (TI) and so contrast with molecules in the analogous
family HC,,O ( n = 2-41, all of which are terminated by angular
C-C-H units (41).That molecules of both series are created in discharge reactions is of more than terrestrial note, beceuse they are
potential players of some significance on the interstellar scene.
2. Silane discharges are of practical use as sources of thin-surface
films. Infrared diode laser spectroscopy with Zeeman modulation has
been exploited to detect and analyze the bending fundamental v2 of
the pyramidal SiH3 radical (42), which appears t o be a primary component of silane plasmas and is likely therefore to be a key intermediate in chemical vapor deposition processes originating in silane. Analysis of the results gives r(Si-H) = 146.8 pm and LH-Si-H = 110.5'.
Other discharge products include SiHz and the novel molecule SieHz,
which is predominant in a low-pressure silane/argon plasma on the
evidence of mass spectrometric measurements. With a low-power
plasma cooled at liquid nitrogen temperature, Si2Hecan be detected
by its submillimeter-wave rotational spectrum. The rotational lines
are unaffected by a confining magnetic field and so cannot arise from
a paramagnetic or ionic species. Moreover, they correspond to the pattern expected for a near-prolate symmetric top with ( B + C)/2 = ca.
7.2 GHz, a value close to what would be expected for a molecule containing two Si atoms, and adjacent lines exhibit a 3 : 1 intensity alternation consistent with the nuclear spin statistics associated with two
equivalent protons. Putting this and other evidence together leads to
the double hydrogen-bridged butterfly structure depicted in Fig. 7(a),
quite unlike the familiar linear configuration of CzH2( 4 3 ) .Confirmation that this is indeed the most stable conformer of SizH2comes from
120
DOWNS AND GREENE
FIG.7. Structures of the molecule SizHzpresent in a silane/argon plasma (a) i n its

ground electronic state and (b) in an excited state. Dimensions are those determined
by analysis of the high-resolution rotational spectra (43, 45).
ab initio quantum chemical calculations (441,which also predict, however, the existence of two other stable conformers, including one having the monobridged structure shown in Fig. 7(b) and lying only about
36 kJ-mol-' above the global minimum. It has been a triumph for
theory that such a conformer should now have been detected and
characterized experimentally following closer scrutiny of the millimeter- and submillimeter-wave rotational spectrum associated with the
products of the silane/argon plasma ( 4 5 ) .The rotational constants of
the normal and perdeuterated versions of the molecule afford an ro
structure with the dimensions indicated in Fig. 7(b) and in tellingly
close agreement with those predicted by the ab initio calculations.
Interestingly, the molecule, which is estimated to have a lifetime not
exceeding 10 ms, displays the shortest Si-Si bond to be observed to
date. Nor is this the end of the tale, for the heteronuclear species
SiCHz, produced by striking an electric discharge in a high-pressure
argon pulse seeded with tetramethylsilane vapor, has for its ground
state the silylidene structure Si=CH2. The molecule has been characterized not only by its microwave spectrum but also by its distinctive,
intense laser-induced fluorescence spectrum (461, which consists of
COMING TO GRIPS WITH REACTM? INTERMEDIATES
121
15-20 vibronic bands in the 300- to 342-nm region; the rotational

structure of the bands, clearly defined at high resolution, features a
central subband that is weaker than the outer ones, a necessary consequence of the nuclear statistical weights if there are two equivalent
hydrogen atoms. The rotational constants of SicHz and SiCDz have
been used to deduce the following partial substitution parameters for
the Czu ground-state structure: rs(C-H) = 109.9 and rsfC=Si) =
170.6 pm, and LH-C-H
= 114.4".
Than SizHz and SiCH2 we could have no better demonstration of
how remote open-shell reactive intermediates may be in their molecular and electronic structures from the familiar world of long-lived,
closed-shell molecules with their apparent obedience to conventional
bonding principles. Chemical studies of plasmas are still only in their
infancy, and more new and exotic intermediates will surely come to
light. Whatever we may make of their structures and bonding, they
will have to be accommodated by any mechanistic proposal that seeks
realistically to account for the reactions taking place in the discharge (47).
Then felt I like some watcher of the skies
When a new planet swims into his ken . .
C. TRAPPING
AND
MATRIX
ISOLATION
One way of catching and identifying a reactive transient is to intercept it with a reactive substrate so as to produce a known product
that has a distinctive spectroscopic signature. For example, hydrogen
atoms formed in the course of a reaction may be effectively scavenged
by carbon monoxide to form the well-authenticated radical HCO,
which can be identified unambiguously under appropriate conditions
by its infrared or ESR spectrum (48).In a similar vein, ground-state
sulfur atoms may be scavenged by dioxygen and detected indirectly
by the infrared absorption or UV emission spectrum of the SO, molecules thus generated; by contrast, sulfur atoms in the excited 'D state
insert characteristically into the C-H bonds of alkanes (48). Some
of the first evidence pointing to the formation of the trihydrides of
aluminium and gallium came similarly from trapping experiments,
this time with trimethylamine to form the known, relatively stable
adducts (Me3N),,MH3(M = Al or Ga; n = 1 or 2) ( 4 9 ) .With molecules
as with atoms, it may be possible on the basis of suitable trapping
122
DOWNS AND GREENE
experiments to determine whether the intermediate is formed in its

electronic ground state or in an excited state. Singlet dioxygen,
that is, the 0, molecule in its excited lh, electronic state, results from
the stoichiometric oxidation of hydrogen peroxide by hypochlorous
acid:
HOOH + HOCl + lo2+ H20 + HC1.
(10)
It has a lifetime of only 2 ps in aqueous solution, but it may be intercepted and thereby distinguished from ground-state dioxygen by its
facile addition to a conjugated molecule such as butadiene (50):
An alternative and more direct strategy involves matrix isolation

(51-53). Here the trick is to catch and hold the intermediate by embedding it in a rigid, inert host such as solid argon (the matrix) at
low temperatures (4-50 K). Hence, our fugitive is at once isolated
and effectively immobilized. To some extent the experiments imitate
nature, for the mineral lapis lazuli owes its beautiful blue color to the
highly reactive radical anion S , , which is entrapped in an aluminosilicate matrix. They also improve upon nature with the choice of a
transparent, weakly interacting host such as a frozen noble gas at low
temperatures, in place of the polar aluminosilicate, which is opaque to
broad regions of electromagnetic radiation.
However short-lived the transient may be under normal conditions,
its lifetime can be extended almost indefinitely by matrix isolation,
and we can appeal to a variety of spectroscopic techniques-notably
vibrational, electronic, and ESR methods-to follow what is going on
in the matrix and to detect, identify, and characterize the trapped
species. The method has great advantages. Spectroscopic measurements can be made at leisure ai., therefore in detail, and several
different methods can be applied to interrogate a particular sample.
Moreover, the infrared spectrum, which has been the principal agent
of detection and analysis of matrix samples, is usually comparatively
simple, with sharp linelike features marking the different transitions
(and reflecting the conditions of low temperature, relatively isotropic,
weakly interacting environment, and inhibition of rotation). Not only
can vibrational frequencies meaningfully be measured to -+ 0.1 cm-
123
in many cases, giving values close to those of the gaseous molecule,

but the linelike character of the bands bids fair to the study of isotopic
effects, frequently the most compelling means of determining the stoichiometry and geometry of the trapped molecule. On the other hand,
there is also a price to be paid. Thus, inhibition of rotation means that
we lose precise information about the dimensions, vibrationh-otation,
distortion, and charge distribution of the isolated molecule. In addition, various complications can arise from the weak but significant
perturbations imposed on the trapped species by the cage it inhabits,
which is likely to vary in size, geometry, and composition within a
given matrix; as a result, specific spectroscopic transitions do not always appear as sharp bands but sometimes as multiplets or relatively
broad bands. The combination of rapid quenching with the rigidity of
the resulting condensate is also liable to lead to the trapping of different conformers of a particular molecule.
In practice no less than principle, the performance of a typical matrix-isolation experiment is quite simple. As illustrated in Fig. 8, a
gas mixture is deposited under controlled, high-vacuum conditions on
a cold surface the nature of which is dictated by the type of spectroscopic measurement to be made on the sample (51-53). The species
of interest may be formed in the gas phase (by pyrolysis or in a discharge) and then quenched rapidly with a n excess of the matrix gas;
sometimes it is formed during the cocondensation of two reagents;
commonly it is generated in situ, usually by irradiation of an appropriate matrix-isolated precursor.
As noted previously, identification of a novel or unfamiliar captive
molecule is accomplished more often than not by reference to its infrared spectrum, with the following steps being typical of the way in
which the experiment might proceed.
1. First, the number of distinct species present in the matrix has
to be established; this is done by allocating the infrared absorptions
to the appropriate species on the evidence of how the spectrum varies
with changes in the experimental conditions.
2. We need then to assess the number, frequencies, and intensities
of all the absorptions associated with a given molecule before drawing
on the evidence of distinctive features due to specific groups, analogies with known compounds, or the vibrational selection rules in order to formulate its likely identity.
3. Evidence of identity is greatly strengthened if it is possible to
devise a n alternative route to the molecule in question starting from a
different precursor. This strategy is well illustrated by the 16-electron
124
DOWNS AND GREENE
Inlet for
gaseous sample
Refrigerant
Inlet for
matrix gas
(4-20 K)
High-vacuum
pumping
system
Valve
Photolysing radiation
Quartz window
Infrared
spectrometer
beam
matrix support
(C s I window)
Sample inlet
hv
125
Ar matrix
FIG.9. Different routes leading to the unsaturated intermediate tungstenocene, (q5C5H5)2W[reproduced with permission from Downs, A. J.; Hawkins, M. Adv. Infrared
and Raman Spectrosc. 1983, 10, 61.
intermediate tungstenocene, ( $-C6H5),W ( 5 4 ) , which has been

shown by its infrared spectrum to be the common product of the matrix reactions set out in Fig. 9.
4. Annealing or photolysis of the matrix may also be instructive for
the evidence of reactions such as ( 5 5 )
2BH3
Ar matnx
27
BPHB.
(12)
FIG.8. ( a ) Schematic representation of a typical matrix-isolation assembly. (b) Schematic plan of a matrix-isolation assembly suitable for infrared (transmission) measurements and for photolysis experiments [reproduced with permission from Almond,
M. J.; Downs, A. J. Adv. Spectrosc. 1989, 17, 31.
126
DOWNS AND GREENE
5. Most revealing of all is usually the response of the spectrum to

changes in the isotopic composition of the molecule. If the sample is
prepared in such a way as to produce a mixture of isotopomers, then
it is often possible to deduce unequivocally the stoichiometry and geometry of the molecule from the number, frequencies, and relative
intensities of the infrared absorptions. With sufficient information, we
may even be in a position t o estimate one or more interbond angles,
if these are not already fixed by the symmetry of the molecule. In the
case of metal carbonyls, the vibrational problem can be simplified by
concentrating on the restricted region of the infrared spectrum near
2000 cm-I, corresponding to C-0 stretching vibrations, v(C-01, which
can be successfully analyzed independently of the other fundamental
vibrations through the use of an energy-factored force field. For each
of several possible models of a particular metal carbonyl fragment
M(CO), , it is then comparatively easy to estimate the frequencies and
relative intensities of the v(C-0) infrared bands for the family of isotopomers M(12C0),(13CO),-,
(x = 0-n) and so determine which model
gives the best account of the observed spectrum. Figure 10 shows the
results of applying this procedure (i) to Fe(C0I4,the primary product
of ultraviolet photolysis of Fe(CO), (56), and (ii) to O,MO(CO)~,a
prominent intermediate in the train of reactions induced by ultraviolet photolysis of Mo(CO16 in the presence of 0, (57). Hence, there is
little doubt that Fe(C0I4 has the CZustructure V or that 02Mo(CO)4
has the tram configuration VI with a square planar Mo(CO)~
fragment.
VI
6. The greatly enhanced sophistication of modern quantum chemical calculations means that we are now able to predict with considerable confidence the equilibrium structure of a molecule, the harmonic
frequencies of its normal modes, and the intensities of the relevant
features in infrared absorption. Accordingly, it has become quite commonplace with matrix as with other studies to test the inferences
drawn from experiments against the results of appropriate ab initio
or DFT analyses. How closely theory matches experiment is shown
127

r
(i i)
A
a
a
2110
2080
cm-1
2675
2050 2000
19j5
1950
cm-1
FIG.10. (i) IR Spectra showing how the structure of matrix-isolated Fe(CO), was
established: (a) observed spectrum of Fe(CO)l with partial W O enrichment isolated in
a solid SF,, matrix a t 20 K [bands shown in dotted lines are due to residual Fe(CO),,
precursorl; tb) spectrum calculated for the C2,,geometry, V, which gives the best
agreement with the observed spectrum; (c) spectrum calculated for a C:+,geometry
which provides a poor match to the observed spectrum 156). tii) IR Spectra showing
how the structure of trans-0,Mo(CO)4was established: ( a ) observed spectrum of trans02Mo(C0)4with partial W O enrichment isolated in a solid CH1 matrix (bands shown
in dotted lines are due to photoejected '.'CO); (b) spectrum calculated for a planar
Mo (CO), fragment with Dlh symmetry, VI (57).
for the case of matrix-isolated MezAINHzformed in situ by the elimination of CH, from the adduct Me& * NH3 through irradiation with
light a t wavelengths near 210 nm (see Fig. 11)(58).
In relation to mechanistic enquiries, the principal strength of matrix isolation lies in the identification and specification of real or potential reaction intermediates. From such studies has come, for example, our first sighting of unsaturated molecules such as Fe(C01, (561,
Fe2(CO), (591,Cr(CO)5(601, (q5-C5H&W(54),and Ru(dmpeh (dmpe =
MezPCHzCH2PMe2)(61). The equilibrium structures of the ground-
128
DOWNS AND GREENE
3400
3000
1600
1000
400
FIG.11. IR Evidence for the formation of MezAINHaby W irradiation (A = 210 nm)

of the complex Me&l.NH3: (a) spectrum calculated for Me2AlNH, a t the RMPZ(fc)/
6-31G* level of theory; (b) difference IR spectrum after irradiation of Me&. NH3 (decreasing absorptions point downward; increasing absorptions point upward). x indicates CH,, y indicates NH,, and z indicates H2 0 (impurity) [reproduced with permission from (58),p. 63721.
state molecules appear to be independent of the matrix material (provided that there is no direct reaction with the matrix), and the case
histories of numerous molecules affirm that the same structures persist in the fluid phases. About the electronic and photochemical properties of such molecules there is much also to be gleaned through
matrix isolation, so there can be little doubt, for example, that
Fe(CO)*has a triplet electronic ground state or that Cr(COI5as initially generated by photolysis of Cr(CO)6is in a vibrationally or possibly electronically excited state. Some clear sign of the reactivity of
the molecule may be found. In this regard, few experiments are more
striking or more significant than those in which the visible absorption
spectrum of Cr(CO)5is found to vary dramatically according to the
nature of the matrix in which the molecule is trapped (60).Hence we
129
learn just what an extraordinarily powerful Lewis acid Cr(CO&is, for

the unsaturated molecule, which has a square pyramidal (C4")structure, evidently interacts to a n appreciable extent through the vacant,
sixth coordination site even with a matrix molecule such as CH4,Xe,
or Ar that has few claims to be regarded as a normal ligand. Experiments with mixed matrices have shown that the token ligand is interchanged by irradiating selectively into the characteristic electronic
absorption band a t lowest energy. Because the molecules are not free
to rotate, it is possible to select those in certain orientations by the
use of plane-polarized photolyzing radiation and so generate products
in orientations that can be analyzed spectroscopically, also with the
aid of polarized light. Hence, the interchange reactions are found to
occur by expulsion of one token ligand, followed by a Berry pseudorotation, and uptake of another token ligand (Fig. 12). There is still
much that we do not know about Cr(CO)5and its photochemistry, but
these matrix studies have surely opened our eyes to the nature of this
particular intermediate and the part it plays in the supposedly simple
substitution reactions of Cr(CO), .
However, matrix studies are limited ultimately by the low mobility
of the caged species, allied to the narrow temperature range over
which they can be held. The very immobilization that helps to preserve a highly labile molecule in a solid matrix becomes a n Achilles'
heel when we wish to explore its reactions. The restraining effect of
the matrix cage is much in evidence in the way photodissociation or
photoisomerization reactions are immediately arrested and in the
way bimolecular reactions induced by photolysis are normally restricted to neighboring pairs of molecules. Reaction channels that are
important in the mobile fluid phases may thus be debarred, whereas
other channels of minor importance in the fluid phases may well be
favored. For example, photolysis of the recently identified chlorine
oxide ClC10, (formally a dimer of C10) results in isomerization to
ClOClO and ClOOCl when the molecules are confined to a matrix,
but yields Clz and O2 as the only detectable products in the gas phase
(62).We have already considered the scope that exists for recognizing
and characterizing reactive intermediates in the gas phase. What options are there for extending our investigations to the liquid phase
and to solutions?
D. SOLUTION
STUDIES
AT Low TEMPERATURES
Various problems are inherent in the solution state when it comes
to tracking reactive intermediates. First, the solvent may be a far
130
DOWNS AND GREENE
Xe
FIG.12. Schematic representation of the photochemical behavior of the Cr(CO)Sfragment in a mixed Ne/Xe matrix a t low temperatures. * represents the Cq. fragment in
the excited 'E state [reproduced with permission from (60),p. 1361.
from innocent spectator; second, as with matrix samples, rotational

transitions are not a feasible source of information; third, the spectroscopic features due to a solute molecule are appreciably broader than
those of the same molecule in a solid, inert matrix, so that isotopic
effects are much harder to evaluate; and last, most solvents absorb
strongly in many regions of the electromagnetic spectrum, thereby
obscuring our vision of the solute species in these regions. Nevertheless, the simple expedient of cooling a solution may be sufficient to
extend the lifetime of an intermediate to the point at which it can be
detected by conventional spectroscopic means. A good example is the
q1-C2H3)(q2-C2H,)H
(VII). This is
vinyl hydride complex ( q5-C5H5)Ir(
13 1
VII
one of the intermediates formed by photolysis of the corresponding

bidethenel complex ($-C5H5)Ir($-C,H,), with light having wavelengths >290 nm (63).The photoinsertion into a C-H bond of a coordinated ethene molecule can be engineered simply by photolyzing a
frozen [2H81tolueneglass at 77 K, and the vinyl hydride can be recognized definitively by thawing the glass and measuring the 'H and
I3C NMR spectra of the resulting solution at ca. 230 K, raising the
temperature to 273 K results in slow decomposition, with reversion to
the starting material.
There is much to be gained from the use of a very weakly coordinating solvent such as a liquefied or even supercritical noble gas (64,651.
The noble gases may each have a very narrow liquid range a t low
pressure, but by working at pressures of 20 bars or more it is possible
to span the entire range from 77 K to ambient temperatures with
liquid argon, krypton, and xenon. I11 adapted though these liquids
may be for the dissolution of polar compounds, they are capable of
dissolving more or less freely many organic and organometallic materials as well as gases such as H2, N,, GO, and CHI, and analysis of
the resulting solutions is hugely facilitated by the absence of solvent
absorption in the infrared, visible, and normal ultraviolet regions,
with the result that much longer pathlengths are admissible than for
conventional solution studies. Accordingly, we can draw on a range of
conventional spectroscopic methods to examine the solutions. Bandwidths may be appreciably greater than for matrix spectra, but problems of matrix site effects are eliminated and the bands are regular
and symmetrical in shape. Moreover, the solutions give much flatter
spectroscopic baselines than do matrices; allied to the advantage of
the long pathlength, this increases considerably the capacity to detect
weak bands. More fundamentally still, and in contrast with matrix
studies, solution-based measurements provide the opportunity of exploring the kinetics and positions of equilibrium of reactions under
thermal control. At the same time, we lose sight of some intermediates revealed by matrix isolation or other means that do not survive
132
DOWNS AND GREENE
in solution, at least long enough to be detected by normal means.

Solution experiments may thus succeed in recording the fate of a labile, coordinatively saturated organometallic molecule such as
CI-(CO)~(N~)
but not of a coordinatively, unsaturated one such as
CI$CO)~.
Figure 13 shows the design of one such cell as used by Poliakoff
and Turner and their group at Nottingham for experiments of this
sort (64). The cell is cooled by liquid nitrogen, the flow of which is
pulsed to vary the temperature. An important feature not illustrated
in the figure is the presence within the cell of a microscale magnetic
stirrer flea; hence, even solutions that are strongly absorbing in the
ultraviolet or visible regions can be made to react photochemically
because the portion of solution immediately adjacent to the irradiation window is constantly being renewed. For photochemically generated compounds that have lifetimes in the order of seconds, it may be
FIG.13. Schematic view of a variable temperature cold cell used for infrared studies
of liquid noble-gas solutions. Cooling is achieved with a pulsed flow of liquid N2 (LN2)
controlled by the output from one of the two thermocouples T so as to stabilize the
temperature. The whole cell fits into a vacuum jacket (not illustrated), which is pumped
through the tube V in the top flange of the cell. The solution under study can be passed
through the cell from a room temperature reservoir via the two tubes marked In and
Out [reproduced with permission from (811, p. 5551.
133
necessary to record the spectrum while the solution is being irradiated.

Some idea of the compass of such studies may be gained from the
case of the dihydrogen complex Cr(CO)S(q2-H,),which for all its 18electron configuration, is too labile to be isolated or spotted by ordinary methods of attack. The complex can be detected following photolysis of Cr(CO)6 or Cr(COI5(NH3)in an H2-doped matrix, but the matrix testimony has to rely principally on v ( C - 0 ) bands in the infrared
spectrum; the broad v(H-H) feature near 3000 cm-' is some 20 to 30
times weaker than the least intense of the v(C-0) bands and, not
surprisingly, is barely visible above irregularities in the baseline.
These findings have been largely eclipsed, though, by the results of
experiments involving photolysis of Cr(CO&dissolved in liquid xenon
under a pressure of H2. Using these conditions, the Nottingham group
has succeeded in detecting a weak, broad infrared absorption fulfilling all the requirements for assignment to the v(H-H) mode of the
coordinated H2 moiety (66).The absorption comes a t 3030 cm-' when
Hz is the ligand, shifting to 2241 cm-' when Hz gives way to Dz. On
the other hand, similar experiments starting from the bis(ethene)
complex ( q5-C5H5)Rh(q2-CzH,)z
reveal the formation not of a dihydrogen complex but of the unstable dihydrido derivative (q5-C5H5)Rh($-CzHJ(H)z (67).
Ultraviolet irradiation of a liquid xenon solution of Cr(CO)6without
any other agent generates a new species that, with a half-life of
about 2 s at 175 K, is sufficiently long-lived to be detected by conventional infrared measurements. The species has d C - 0 ) absorptions
very close in frequency and intensity pattern to those assigned to the
matrix-isolated species Cr(CO)5----Xe
(64).Its identity can be checked,
first by deliberately doping the solution with a range of likely impurities in order to rule out the possibility of its being some other compound of the type Cr(CO)5L (L = Nz, H20, etc.), and second, by carrying out experiments with Cr(CO)6 dissolved in pure liquid krypton
and in liquid krypton doped with 5% xenon. Hence, the realization
is dawning that the coordination of Cr(COI5 by xenon involves not a
generalized solvation but a specific interaction with the formation of
a Cr-Xe bond of appreciable strength.
Similarly, the transient Ni(CO),(N,) has been identified following
its generation by photolysis of Ni(CO), in liquid krypton doped with
Nz a t 114 K. Ultraviolet irradiation establishes a steady-state concentration of Ni(CO),(N,); as soon as the photolysis source is switched off,
the dinitrogen complex begins to decay, reacting thermally with CO
to regenerate Ni(CO), . The kinetics of the decay reaction,
134
DOWNS AND GREENE
have been investigated in some detail over the temperature range

112-127 K (68). The rate shows a first-order dependence on the concentration of Ni(CO),(Nz)and a more complex dependence on the concentrations of CO and Nz. Hence, it appears that the reaction proceeds via two simultaneous paths, one dissociative and the other
probably associative. An estimate of about 42 t- 4 k J mol-' for AH$
for the dissociative path gives a measure of the Ni-N2 bond dissociation enthalpy in Ni(CO),(NZ).
In an extension of the technique we can now trade on the remarkable properties of supercritical fluids (64, 69). These combine a number of properties of gases and liquids; like gases they are highly compressible and are completely miscible with Hz or Nz, but like liquids
they are relatively dense and dissolve a wide range of nonionic solids.
In fact, most materials are considerably more soluble in supercritical
xenon than in liquid xenon, despite the fact that the density of the
supercritical material at the critical point is only about one-third that
of liquid xenon. Moreover, the solubility of any compound in a supercritical fluid increases with increasing pressure at a constant temperature, with the result that the solvating power of the medium can be
tuned by varying the pressure. It is the superior solubility of compounds that gives supercritical solvents a major advantage over their
normal liquid counterparts and particularly in reactions involving
gases such as Hz or Nz. Thus, for a given pressure of gas, the effective
concentrations of Hz in a supercritical solution can be nearly an order
of magnitude higher than in a liquid solution. As a result, it is altogether easier to detect, preserve, and even synthesize known or potential reaction intermediates. Previously unknown dihydrogen and dinitrogen complexes, which would normally decay rapidly at ambient
temperatures, have thus been formed in supercritical xenon, where
they are sufficiently robust at room temperature to be characterized
quite normally by their infrared spectra (69, 70). Kinetic measurements between 283 and 353 K show that (r16-C6HSMe)Cr(CO)z(772-Hz)
reacts with CO in accordance with Eq. (141, with an activation enthalpy of 70 t- 5 k J . mol-', which may correspond to the Cr-H2 bond
dissociation enthalpy :
( 776-CGHsMe)Cr(CO)z(~-H2)
CO -+ (q6-C6HsMe)Cr(CO), H2.
(14)
This conclusion finds support from more recent studies, also at high
pressure but in n-heptane solution and using photoacoustic calorime-
135
try to measure the enthalpy change of Reaction (15) (711,

M(CO)G + L + M(CO),L + CO
(15)
(M = Cr or Mo; L = n-heptane, H2or N2);

the Cr-H2 and Cr-N2 bond enthalpies are estimated in this case to
be 78 and 81 kJ . mol-', respectively.
Figure 14 illustrates the region of the infrared spectrum associated
with z4N-N) modes for a supercritical xenon solution of the rhenium
carbonyl ( T ~ - C ~ H ~ ) R ~under
( C O )a~pressure of NZ (ca. 90 bars) and
shows how this changes with time on photolysis with ultraviolet light
(70). The spectrum bears clear witness to the sequential formation of
C~H
) ,, ) R ~ ( C O )and
( N ~(q5-C5H5))~,
the species ( T ~ - C ~ H ~ ) R ~ ( C O( T) ~ (- N
Re(N2I3a t room temperature, and the last two turn out to be surprisingly stable under these conditions. In suitable circumstances,
solutions in supercritical solvents are amenable even to NMR measurements (72). Supercritical fluids are thus affording a new dimension and are giving direct access to new chemistry involving reactive
intermediates, as well as a new approach to problems of analysis,
synthesis, and materials processing.
0.30
0.15
0
2160
2130
2100
2070
v/cm-'
FIG.14. IR spectra showing the sequential formation of ( v ~ - C ~ H ~ ) R ~ ( C O

(dark
)~(N~)
peaks), (7)"-CSH5)Re(CO)(Nz)z
(medium peaks), and ( T ~ - C ~ H ~ ) R
(light
~ ( Npeaks
~ ) ~ ) from
($-C,H5)Re(CO)3 and Nz in supercritical xenon solution [reproduced with permission
from Poliakoff, M.; Howdle, S. Chem. Br. 1995, 31, 1201.
136
DOWNS AND GREENE
IV. Experimental Characterization of Reaction Intermediates:

Time-Resolved Methods
To retard or suppress chemical reactions it has been necessary to

work under rather unusual conditions. It is only natural to want to
know about the behavior of intermediates under more normal conditions. However, if the species are highly reactive, detection and monitoring of their fate demand very rapid spectroscopic measurements.
From the more or less conventional type of spectroscopic analysis that
has been the mainstay of our investigation so far we pass to methods
that can be adapted to real-time measurements and in which spectral
evolution or decay can be monitored at intervals sometimes as short
as femtoseconds. It will soon be apparent?though, that the knowledge
about the kinetic correlation between intermediates, reagents, and
products thus gained can only be at the expense of detailed knowledge
about the physical properties of the molecules-sometimes even their
very identities. The function of the gas-phase, supersonic-jet, matrixisolation, and low-temperature solution techniques described in the
preceding sections is in the provision of a spectral library of more or
less well-authenticated reaction intermediates, whether real or model
ones; this library is going to prove essential to the interpretation of
many time-resolved experiments.
The most popular procedure involves flash photolysis or, in the
terms of laser technology, pump-and-probe experiments (73-76).
Flash photolysis was developed by Norrish and Porter in Cambridge
in the 1940s and earned them the Nobel Prize in 1967. The basis of
the technique is simplicity itself. A pulse of light is used to generate
the transient, typically in a fluid sample at or near ambient temperatures, and a spectroscopic record of the sample in absorption or emission is then kept as a function oftime. The time-scales of the reactions
that can be studied are limited by the duration of the initiating light
pulse. The flash lamps used by Norrish and Porter, which give pulse
durations of several milliseconds, have now been largely superseded
by high-powered lasers giving pulse durations of nanoseconds or less.
By means of a process called mode-locking it is possible even to generate laser pulses lasting only a few picoseconds or less; the world
record is currently reported to be 4 femtoseconds. Such lasers offer a
means of tracking very fast kinetic phenomena, such as the isomerization of an electronically excited alkene, the rotational motion of
a large molecule in solution, or some of the fast processes occurring
in photosynthesis. Although these ultrafast processes (77) are not
reactions in the sense that we have discussed so far, they are none-
137
theless important in the contributions they may make to overall reactions.

An alternative technique for the rapid generation of transient species is that of pulse radiolysis triggered not by a light flash but by
a short pulse (10-9-10-6 s) of high-energy electrons (1-5 MeV); this
approach has been employed primarily for studies of liquid phase kinetics and especially of the solvated electron, but it is a very flexible
method with applications also for gas phase studies (78). Representative of this method of attack is the formation of the peroxynitrate
anion 02NOO- less than 2 ms after pulse irradiation of aerated aqueous solutions containing nitrate and formate a t relatively low concentrations (79). The anion, which can be detected and monitored by its
UV absorption near 290 nm, is a strong oxidizing agent notable for its
potential mediation in some of the reactions that make living systems
vulnerable to nitrogen dioxide, one of the most toxic components of
polluted atmospheres. It is of some consequence therefore that studies
of several oxidation reactions should reveal two distinct pathways by
which the peroxynitrate decays, one proceeding through the parent
acid OzNOOH and the other probably evolving through secondary reactive intermediates derived from the decomposition of the peroxynitrate.
Once the transient species has been formed, it has to be monitored
by some form of kinetic spectroscopy, typically with ultraviolet-visible
absorption or emission, infrared (time-resolved infrared or TRIR) (74),
or resonance Raman (time-resolved resonance Raman or TR3) (80)
methods of detection. The transient is usually tracked by a probe
beam a t a single characteristic frequency, thereby giving direct access
to the kinetic dimension. Spectra can then be built up point by point,
if necessary, with a n appropriate change of probe frequency for each
point, although improvements in the sensitivity of multichannel detectors may be expected to lead increasingly to the replacement of the
laborious point-by-point method by full two-dimensional methods of
spectroscopic assay (that is, with both spectral and kinetic dimensions).
Ultraviolet-visible absorption has traditionally been the basis of detection in flash photolysis experiments. It offers a number of advantages in sensitivity and efficiency and has certainly delivered much
vital information about the reactivity of transient species. On the
other hand, the ultraviolet-visible absorption bands characteristic of
any but the simplest molecules tend to be broad and relatively uninformative as regards identity and structure, and so we may run into
problems not only with the overlap of absorptions due to different
138
DOWNS AND GREENE
species, but also with deciding on the nature of the carrier of a particular absorption spectrum. Altogether more discriminating, albeit less
sensitive, are the TRIR methods that have been successfully developed in the past 15 years or so (74, 81). With the sort of apparatus
depicted schematically in Fig. 15, it has been possible to draw on the
spectral library built up from matrix isolation and other experiments
in the slow regime to identify transients formed in the gas phase or
in solution at ambient temperatures, as well as investigating their
kinetic behavior. Just what can be achieved in this way will be demonstrated by some examples representative of recent studies.
As a start, we return to that remarkable Lewis acid Cr(COI5.Flash
photolysis experiments have been carried out with Cr(CO)6either in
solution or in the gas phase and with either infrared or visible probes
(9,601. For example, photolysis of gaseous Cr(CO)6with the output of
an XeF laser yields a predominant transient that can be tracked by a
IR spectrometer
HeNe alignment laser
Sample cell
Oscilloscope Pulse generator
0.-
Micrmmputer
Detector
Transient digitiser
FIG.15. Schematic representation of the time-resolved infrared (TRIR) flash photolysis apparatus used a t Nottingham. The UV pulse laser generates transient species;
the continuous IR laser monitors the change in transmission a t a particular IR frequency, producing a trace showing the IR absorbance as a function of time. The experiment is repeated a t different IR frequencies so that a complete IR spectrum of the
transient can he built up [reproduced with permission from (97),p. 1031.
139
line-tunable, liquid-nitrogen-cooled CO laser, changes in intensity of

the transmitted infrared light being recorded by a high-speed InSb
detector. The v(C-0) bands clearly imply that the species is naked
Cr(CO), (freed a t last from any token ligand) with the same square
pyramidal geometry as the molecule identified previously by matrix
isolation (see Section 111,C).The identification is confirmed by the decay of the transient a t a rate matching the rate of reappearance of
Cr(CO)6.The rate constant for the reaction of Cr(CO), with CO a t 293
K comes out to be 1.5 t 0.3 X 10dm3-mol I . 5 - l . Similar studies of
the reaction of Fe(C0I4with CO give a rate constant about two orders
of magnitude smaller (56).This difference can be rationalized, a t least
in part, with the realization that Fe(CO), has a triplet ground state
whereas Cr(CO), almost certainly has a singlet ground state. Closer
examination shows that Cr(CO), is generated within 1 picosecond of
light absorption and that it is formed in vibrationally excited states
that relax over tens if not hundreds of picoseconds (60).These details
emerge not only from time-resolved ultraviolet-visible studies but also
from analogous TR3 experiments with cyclohexane solutions in which
two 5-ps pulses a t A = 266 nm are employed; the first dissociates
Cr(CO)6,and the second probes the resonance Raman spectrum of the
photoproduct Cr(CO),. Hence, it appears that the naked intermediate,
albeit vibrationally excited, can exist for a few picoseconds before taking up a solvent molecule.
Extension of these studies to supercritical fluids provides a way of
making TRIR measurements on intermediates incorporating a bond
between a metal center and a weakly coordinating ligand such as
a noble gas or CO,. Hence, for example, a series of organometallic
noble-gas compounds, W(CO),(Ar), M(CO),(Kr), and M(CO),(Xe) (M =
Cr, Mo, or W), have been detected transiently in fluid solution a t room
temperature (82).The second-order rate constants for the reaction of
M(CO15Lwith CO show that the reactivity for a given metal M varies
in the order L = Kr > Xe = C 0 2 , whereas the reactivity for a given
ligand L varies in the order M = Cr = Mo > W. The reaction of
W(CO),(Xe) with CO has an enthalpy of activation of 34 k J . mol-,
representing a lower limit for the W-Xe bond enthalpy. The bond is
therefore twice as strong as a conventional hydrogen bond, and its
mechanistic impact cannot be ignored.
The photochemistry of the binuclear iron carbonyl [(q5-C5H5)Fe( C O ) z ] 2VIII,
,
provides another outstanding example of the interplay
between time-resolved and alternative strategies (76). According to
the infrared spectrum of this compound in a suitable matrix a t low
temperatures (83), ultraviolet photolysis results in the elimination of
free CO, but the pronounced cage effect of the medium entirely sup-
140
DOWNS AND GREENE
presses the separation of (q5-C5H5)Fe(

C0lz radicals, X. Experiments
with a l3C0-labeled sample demonstrate that the iron-containing
product formed by CO loss has a binuclear framework incorporating
three CO bridges, M:
0
IX
VIII
Pulsed laser photolysis (at A = 308 nm) of VIII in an alkane solution

at room temperature can be shown by TRIR measurements to generate not one but two metal carbonyl species with quite different kinetic
properties (84).One transient displaying two infrared bands characteristic of the u(C-0) modes of terminal CO groups decays rapidly by
second-order kinetics (with kz = 5 X lo9 dm3.mo1-l. s-') to regenerate
the precursor within about 50 ps. The infrared spectrum of this species, allied to its other properties, makes clear that it is the radical
transient is the triply CO-bridged
( Q ~ - C ~ H ~ ) F ~ X.
( C The
O ) ~ second
,
complex M,easily recognizable by the close resemblance its infrared
spectrum bears to that of the matrix-isolated species. In the absence
of added CO, this has a half-life of about 0.6 s-that is, some four
orders of magnitude longer than that of the radical. IX reacts with
CO, and its decay as well as the regeneration of VIII can also be
monitored by TRIR measurements. Similar experiments with another
~ ( CLO=) CH3CN
~L
substrate L to form the product ( Q ~ - C ~ H ~ ) ~ F ~[e.g.,
or P(OMe)3, see Fig. 161 reveal the participation of both intermediates M and X through parallel reaction paths. The reactivity of the
radical X may thus be gauged by rate constants close to the diffusioncontrolled limit. For another string to our spectroscopic bow we may
also turn to Raman scattering, using a single visible laser to fulfil the
dual functions of pump and probe, and detecting the emitted light by
a suitable optical multichannel analyser. Cyclohexane solutions of
[( Q ~ - C ~ H ~ ) F ~ and
( C Oits) ~permethylated
]~
analog [ ( Q ~ - C ~ M ~ ~ ) F ~ ( C O ) ~ I
show resonance Raman behaviors correlating with the lower energy
141
el
1950
1850
1750
WO"e""lllber/Clll-'
FIG. 16. Time-resolved infrared spectrum obtained after UV flash photolysis of I($CSH5)Fe(CO)211,VIII (A), and MeCN in cyclohexane solution at 25C. The bands ar e
labeled thus: B, ($-C5H5)Fe(CO)2,X C, ($-C5H5),Fe,(p-CO)n,M;and D (r)"-CSH&FeL(CO).I(MeCN).The first three spectra correspond to the duration of the firing of the UV
flash lamp, and subsequent spectra are shown a t intervals of 10 p s . The negative peaks
in the first spectrum are due to material destroyed by t h e flash; these have been omitted from the subsequent traces to avoid undue confusion [reproduced with permission
from Dixon, A. J.; Healy, M. A,; Poliakoff, M.; Turner, J. J. J. Chem. Soc., Chem.. Commun. 1986, 9941.
electronic absorption bands, thereby shedding light on the molecular

orbitals involved in the electronic transitions (85). By the simple device of increasing the incident laser power it is possible to observe
additional rich resonance Raman spectra that must originate in the
binuclear transients (775-CSRS)zFez(~-CO)~,
where R = H (MIor Me.
142
DOWNS AND GREENE
With its different requirements, the Raman effect gives access to vibrational transitions that may well be lost to view in infrared absorption.
Time-resolved Raman spectroscopy, both spontaneous and coherent, has developed apace in the past two decades (86). As a means
of pursuing short-lived intermediates, spontaneous resonance Raman
scattering has proved particularly instructive, partly because of its
high selectivity, partly because of its enhanced sensitivity. Since studies of this sort were initiated in 1976, the time resolution has improved from microseconds to subpicoseconds. The method has really
come into its own, notably in the hands of Kitagawa and his colleagues at Okazaki (87), in the investigation of heme proteins and of
how they bind dioxygen and other ligands. For example, the reaction
of dioxygen with the reduced enzyme cytochrome c oxidase (CcO),
formed by CO photodissociation in aqueous solution, has been followed by TR3 measurements (87), which reveal within 10 ms the
growth and decay of several intermediates identifiable by emissions
due to 0-0 or Fe-0 stretching vibrations in the region 300-850 cm-*.
The origins of the bands have been traced by experiments with "0enriched samples of 02,and hence it can be shown that the O2 binds
initially to the Feassite in CcO in an end-on fashion with an Fe-0-0
bond angle close to 120". Studies with picosecond time resolution witness within 5 ps of photolysis of the CO-bound precursor an intermediate believed t o be the high-spin, five-coordinate heme Fe:; to which
a histidine (His) is ligated; this is indicated by the growth of a band
at 220 cm-' attributable to the Fe-His vibration (88).
Intense interest in how carbon-hydrogen bond activation may be
engineered has led to detailed studies of several transition-metal compounds known to compass reactions of this type, which are sometimes
initiated thermally, but more often photochemically. Two such compounds are the iron and ruthenium dihydrides M(dmpe)zHz(M = Fe
or Ru; dmpe = Me2PCH2CH2PMe2),
which react with alkanes RH under irradiation in accordance with Eq. (17) (61, 76):
143
The ruthenium compound gives an object lesson in the merits of combining matrix with time-resolved solution studies. Photochemical reductive elimination of Hz may be followed by the decay of the characteristic v(Ru-H) bands in the infrared spectrum or by the growth of
three intense visible bands at ca. 740, 540, and 460 nm. Although the
product lacks characteristic infrared features, the circumstances of
its formation and its response to matrix dopants imply that it is the
16-electron species Ru(dmpe):,, presumably with a square planar
skeleton.
Flash photolysis of a cyclohexane solution of the dihydride with a
pulsed laser yields a transient with a visible spectrum remarkably
like that of the matrix photoproduct. The transient signal decays by
second-order kinetics over ca. 80 p s as the precursor is regenerated.
Addition of H2 even in very low concentration increases the rate of
decay, causing it to become pseudo-first order; under these conditions
the reaction of the intermediate with H2 is essentially diffusion controlled, being opposed by a minimal activation barrier. The corresponding iron compound Fe(dmpe)2H, cannot be vaporized without
decomposition and so does not lend itself to matrix isolation, but flash
photolysis of a n alkane solution at room temperature g v e s a transient Fe(dmpeIz differing conspicuously from its ruthenium counterpart. For one thing, its electronic absorption spectrum shows but a
single band a t low energy, and that not in the visible but in the nearultraviolet region. The reactivities are quite different too: both intermediates add substrates to form products of the type M(dmpeIzL,
where L = CO, C2H4,or PMe3, as well as reacting with Hz, certain
hydrocarbons, or EtySiH to give insertion products of the type
M(dmpe12(X)(H)(X = H, organic group, or Eta&), with the kinetic results summarized in Fig. 17. The reactions with Hz, EtsSiH and alkenes are slower by at least two orders of magnitude for Fe(dmpeIz
than for Ru(dmpeIz,whereas alkanes and arenes find Fe(drnpe&much
the more reactive of the two. One plausible explanation of the spectral
and kinetic differences is that the intermediates have different structures, with Fe(dmpe12adopting a butterfly configuration unlike the
square planar form of its ruthenium analog, although we lack definitive evidence on this point. The reduced rate of the back-reaction with
H2 to regenerate the dihydride precursor must be one of the key factors making Fe(dmpe)z more reactive than Ru(dmpeIz toward C-H
insertion reactions, particularly with alkanes and arenes.
Similar experiments involving photolysis of the ruthenium hydride
place Ru(PP3)Hz,in which the two bidentate dmpe ligands have given
place to the tetradentate phosphine PPy= P(CHzCH2PPh2),, reveal the
144
DOWNS AND GREENE
'0
--..co
..HZ
-co
-Et&H
2cyclopantene
6.
109 k2
4 -
2 -
-benzene. max
-cyclopentene
-EtjSIH
H2
-pentane
-cyclohexane
FIG. 17. Schematic comparison of the intermediates Ru(dmpe)a and Fe(dmpe):,

contrasting their UV-visible spectra and reactivities; the
(dmpe = MeaPCH2CHaPMe2)
latter are expressed as log k a where k a is the second-order rate constant for reaction
with the substrate a t room temperature deduced by laser flash photolysis [reproduced
with permission from (611, p. 3641.
transient intermediate Ru(PP3) (89). Here the PP3 ligand obliges the
molecule to adopt not a square planar RuP4 skeleton but a pyramidal
(XI) or butterfly one (XII), and it is noteworthy that the UV-visible
XI
XI1
spectrum resembles that of Fe(dmpeI2with but a single absorption a t

395 nm. In its reactivity, moreover, Ru(PP3) is quite different from
R ~ ( d m p e )being
~,
much slower, for example, to react with H2. The
145
corresponding osmium intermediate, which has a similar structure on

the evidence of its W-visible spectrum, does not even react with H2
but inserts into the C-H bonds of alkanes with rate constants in the
order of lo5 dm mol-. s at 295 K. Hence, it appears that the enforced nonplanar geometry effectively prepares Ru(PP3),and more especially Os(PP3),for C-H activation (89).
In the first application of ultrafast TRIR spectroscopy (that is, with
time-scales of
to IO-s), reductive elimination and oxidative addition of H2 have been traced following flash photolysis of the related
ruthenium carbonyl dihydride complex Ru(PPh,),(CO)H, in benzene
solution (90).
This precursor is of particular note because it is known
to catalyze insertion of alkenes into C-H bonds a t the unsaturated
carbon center of alkenes or arenes in a position p to a carbonyl group.
The course of events has been monitored from excitation with an
ultrafast UV laser (giving pulses a t A = 304 nm with a n energy of
1-2 pJ, a full width of 4 ps a t half of maximum duration, and a repetition rate of 1.05 kHz) to 2000 ps after the laser pulse using the
v(C-0) mode as a reporter, IR absorption measurements being made
with a CO laser or with a diode laser via a detection system employing up-conversion. Photolysis results in reductive elimination of
H2 and the formation of an unsaturated transient, Ru(PPh3)&CO),
which can be identified unequivocally on the basis of its rapid buildup, the kinetics of its reaction with H2, the partial regeneration of
starting material, and the frequency of its v(C-0) band. Moreover,
the photoelimination step-involving H-H bond formation, Ru-H
bond cleavage, and any reorganization of the coordination geometry
at ruthenium-is found to be complete within 6 ps.
Nor has the speed limit yet been reached, for there are even ways
of clocking chemical events a t the femtosecond level. For example,
Ahmed Zewail and his colleagues a t the California Institute of Technology have developed a novel approach relying on ultrafast lasers
giving femtosecond pulses to study in real time the transition-state
dynamics of charge-transfer and other reactions (91, 92). The entire
molecular system is prepared on a reactive potential energy surface
and in a well-defined impact geometry. A femtosecond light pulse
induces a charge-transfer transition, and the dynamics of the ensuing events are then followed by interrogating the transition state
or the reaction products; this is done with probe femtosecond light
pulses causing photoionization and mass spectrometric sensing of
the resulting fragments. The method is well illustrated by the dissociative charge-transfer reaction of the benzene-iodine complex (91 ):
146
DOWNS AND GREENE
rcnc tniit
transition state
products
The transition state is directly accessed by excitation into the

charge-transfer absorption band of the complex formed initially by
expanding a gas mixture containing benzene, iodine, and helium
through a pulsed valve. The resulting molecular beam is skimmed
and intersected by the femtosecond laser pulses in the acceleration
region of a linear time-of-flight mass spectrometer. The pulses of
the femtosecond pump-laser (at A = 275 nm) initiate the reaction
and define t = 0. The probe pulses (at A = 304.7 nm) then detect
the free iodine atoms through a resonantly enhanced multiphoton
ionization (REMPI) process. Transient I atom growth is measured
by monitoring the It signal as a function of the pump-probe delay.
In addition, by using linearly polarized pump pulses and having
provision for varying the orientation of the plane of polarization
with respect to the axis of the mass spectrometer, it is possible to
analyze the velocity distributions of the I atoms. Hence, microscopic
elucidation of the reaction dynamics and mechanism becomes a
practical proposition. We are able to witness the evolution and
decay of the transition state and to show that dissociation with the
formation of I atoms occurs via one of two exit channels. One
(ionic) follows the ionic potential of the charge-transfer state
CsH& to produce c6Hg.1- + I; the other (neutral) involves intermolecular electron transfer through coupling of the transition state
to neutral, locally excited, iodine-repulsive states, and in this case
the products are C6Hs.I + I. With a timescale of 250 fs, the neutral
process is ultrafast, outstripping the ionic process, which proceeds
at a leisurely 800 fs or so. The kinetic energy characteristics of the
I atoms even go so far as to reveal a near-axial geometry for the
transition state. A similar approach has been adopted to show that
the two-center elimination reaction (19) is a two-step, nonconcerted
process evolving via the intermediate CzFJ (92):
147
Here the time-scales for the breaking of the two C-I bonds differ by
two orders of magnitude: the primary step takes about 200 fs,
whereas the secondary step occupies 25 ps. The speed of the first step
reflects the repulsive force in the C-I bond caused by the promotion
of a nonbonding electron from the highest occupied molecular orbital
(HOMO) to the a * lowest unoccupied molecular orbital (LUMO),
whereas the second elimination is governed by the energy redistribution attending the concerted fission of a C-I a-bond and making of a
C-C .rr-bond.
The high-resolution spectroscopic methods involving, for example,
infrared diode lasers that have been successfully applied to the detailing of gaseous transients have also been adapted to the time domain
(10). High temporal resolution cannot of course be reconciled with
high spectral resolution, but measurements made a t even millisecond
or microsecond intervals can still provide invaluable information
about many reaction systems, with the bonus of unambiguous identification of intermediate species made possible by the high spectral
definition. Reactions have been induced typically by irradiation with
a high-power source, such as a n excimer or CO, laser, or on occasion
by pulse radiolysis, and the subsequent course of events has been
monitored by a suitable high-resolution spectroscopic probe. For example, Curls group a t Rice University in Houston, Texas has investigated the reaction between the radicals NH, and NO, which appears
to proceed via more than one channel (93):
NH,
+ NO
N, t H,O
(a)
H + N,+OH
(b)
HN,(?) + OH
(c).
(20)
This is of considerable interest because of the key role of the reaction

in the thermal de-NO, process involving the catalyzed reaction of
ammonia with nitrogen oxides; hence, in principle, the gaseous emissions from industrial plants and power stations can be purged of the
148
DOWNS AND GREENE
noxious oxides, which are known to be major sources of smog and

acid rain. Reaction (20) can be initiated by excimer laser photolysis of
ammonia in the presence of NO and monitored by comparing the intensity of the infrared absorption due to NH3 with that due to NH, as
functions of time, and also with those due to OH and H,O. Hence, the
branching ratio of the OH channel or channels has been determined
at different temperatures, with results suggesting that an additional
channel may be implicated.
Replacing the probe laser of these studies with electron pulses offers a means of imaging structural changes by electron diffraction,
but observation of chemical changes in this way is nontrivial, and
only recently have the odds against the experiment been significantly
shortened (94-96). The essential experimental principles of stroboscopic electron diffraction were first devised more than a decade
ago, with a pulsed electron beam being generated by photoemission
with millisecond to picosecond time resolution. Major advances have
come with the application of online data-recording techniques, involving either photodiode array or charge-coupled device (CCD) detection. Quite apart from the low intensity of coherent molecular scattering compared with incoherent background scattering, however,
numerous problems remain to be solved. How, for example, is one to
determine in situ the zero of time in a chemical change? And how is
one to take account of molecular ensembles in nonequilibrium (that
is, non-Boltzmann) vibrational distributions? In fact, a temporal resolution of 15 ns has been achieved with a pulsed laser-driven electron
source. Hence, the structural and vibrational kinetics of CS have been
explored during the first 120 ns following the photodissociation of CS,
at A = 193 nm (94):
The observed changes of vibrational population with time have been

interpreted on the basis of inelastic collisions of CS with sulfur atoms
in the electronically excited D state. Further hazards await any attempts to probe picosecond or subpicosecond changes by electron diffraction: space-charge effects broaden the electron pulse-width, and
temporal overlap of the initiating photon pulse and the probe electron
pulse must somehow be established. For ultrafast studies it is necessary then (i> to measure the electron pulse-width, (ii) to have the
means of accurate clocking of the reaction, and (iii) to be able to detect
single electrons (so as to reduce the electron flux and minimize space-
149
charge broadening). Zewail and his group have risen to this challenge
by developing the new apparatus shown in Fig. 18 (96).This consists
of a femtosecond laser, an ultrafast electron gun, a free-jet expansion
source, and a newly designed two-dimensional single-electron detec-
Thermoelectric
I
Fibre-optic raper
FIG. 18. ( a ) Ultrafast electron diffraction apparatus consisting of a n electron gun

chamber, a diffraction chamber, and a detector chamber. Two fs laser pulses are used,
one to initiate the chemical change and the second to generate the electron pulse. (b)
Detector system: incident electrons either directly bombard a small CCD or strike a
phosphor-coated fused fiber-optic window. Light emitted from the phosphor is amplified
by an image intensifier and brought to a scientific-grade CCD. Both CCDs are thermop. 1601.
electrically cooled [reproduced with permission from (96),
150
DOWNS AND GREENE
(A)
FIG.19. Radial distribution functions f ( r )and Af(r) derived from the molecular scattering curves for gaseous CHzIza t different delay times between the photodissociating
fs laser pulse (at A = 310 nm) and the ps electron pulse (pulse-width 15 ps). The corresponding theoretical f ( r ) curve for CH212is supeorimposedon the -20-ps data set. The
changes observed are a t r = ca. 2.0 and 3.5 A corresponding to the C-I and I----I
internuclear spacings, respectively [reproduced with permission from (96), p. 1611.
tion system. Femtosecond laser pulses are created from a collidingpulse mode-locked ring dye laser, the output from which is directed
through a four-stage pulsed dye amplifier with no pulse compression
(620 nm, 2-3 mJ, 30 Hz, 300 fs). To initiate the reaction, 95% of this
beam is doubled (310 nm, ca. 250 pJ);the remainder (also doubled)
is focused on a back-illuminated, negatively biased photocathode to
generate the electron pulses. Critical to the success of the experiment
is the detector, a two-dimensional CCD operating in a direct electron
bombardment mode. Hence, a temporal resolution in the picosecond
range can be realised. The first such studies have monitored the photodissociation of CHJz a t A = 310 nm; the molecular scattering and
radial distribution functions vary significantly with time (see Fig. 19,
for example) in a manner wholly consistent with Reaction (22):
CH&&
CHzI -t I.
(22)
151
A more detailed quantitative analysis may be expected to elucidate

the structure of the CHJ intermediate. Modulation and difference detection techniques should provide a way of suppressing the background scattering from unreactive species, thereby enhancing the precision with which structure changes can be evaluated.
To understand the whole of a photochemical pathway requires a
knowledge of the excited electronic state or states (75, 97). In addition, some thermally activated processes proceed through the formation of intermediates or of products in excited states (witness, for example, the formation of singlet dioxygen). However, inorganic and
organometallic molecules in these states are generally far too shortlived to be characterized directly by conventional spectroscopic observations, and there is no physical device open to us that will significantly extend their lifetimes. Figure 20 illustrates schematically a
simple system involving just the ground and one excited state, and
Exated sfafe
2.
Resonance
Raman
Qk (distortioncoordinate)
FIG.20. Schematic representation of the potential energy curves for the ground and
an excited electronic state of a molecule and some of the spectroscopic connections that
can be made between them [reproduced with permission from (7.9, p. 1191.
152
DOWNS AND GREENE
some of the spectroscopic connections that can be made between

them. In principle, the optical absorption and excitation spectra provide a means of probing the vibrational levels of the excited state,
whereas the emission spectrum reports not only on the vibrational
levels of the ground state but also on the lifetime of the excited state.
The intensities of the vibrational components of these spectra contain
the seeds of information about the distortion the molecule experiences
with the switch from the ground to the excited state. Unfortunately,
though, relatively few molecules, and then at low temperatures, afford the sort of fine structure needed to make this a realistic proposition. The metallocene intermediates (q5-C5H5)zM
(M = W or Re) provide two cases in point, well-resolved vibrational progressions in the
optical spectra of the matrix-isolated species admitting meaningful
estimates of the changes in the metal-to-ring-centroid distance accompanying the electronic transition at lowest energy (61, 76).Resonance Raman spectra also give access to similar structural information. The case of the binuclear iron carbonyls (q5-C5R5)2Fez(C0)4
(R =
H or Me) has already been alluded to (85), and similar studies have
shown that the W-N and trans C-0 bonds of W(CO)5(pyridine)are
extended by 18 and 12 pm, respectively, on excitation to the photoactive state (75). However, none of these methods involves probing of
the excited state in real time, with the advantages this would bring
to its structural and temporal delineation. One way of observing directly the vibrational spectrum of a molecule in an excited state is to
engage TR3 measurements (801, as with W(CO),(diimine) complexes
(diimine = 2,2'-bipyridine or related species) in the tungsten-todiimine charge-transfer excited state at lowest energy (98). The totally symmetric cis-carbonyl v ( C - 0 ) mode has a frequency about
50 cm-' higher than it does in the ground state, implying some reduction of metal-carbonyl a-back-bonding in the excited state. For all its
merits, however, this approach is somewhat limited in its application,
and there are real attractions in being able to record the infrared
spectrum of a molecule in an excited electronic state. The new developments in TRIR spectroscopy now make this possible, at least in
suitable cases. Thus, fast TRIR measurements are able to detect and
monitor the lower metal-to-ligand charge-transfer state (MLCT where
L = 4-cyanopyridine) of the tungsten pentacarbonyl complex
W(C0)5(4-cyanopyridine),XI11 (99). Excitation causes the v(C-0) infrared bands to shift to higher frequency, confirming that the metal
center is oxidized in the excited state. This state can be observed to
decay, partly reverting to the ground state and partly with dissocia-
153
XI11
tion to form solvated W(CO)5and 4-cyanopyridine. The second reaction channel depends on rapid thermal equilibration between the
MLCT and a second excited state, which is metal-centered or ligand
field (LF) in nature, and it is this state that is dissociative with respect to separation of the 4-cyanopyridine ligand. confirmation of this
comes from the finding that the yield of W(CO), varies with temperature, giving an estimate of some 4000 cm-' for the energy gap separating the lowest MLCT and LF states (see Fig. 21). This example illustrates the sort of subtle details that time-resolved studies are now
able to chart with regard to excited states and the chemical processes
they beget.
About reactions occurring in the solid state there is not generally
the same urgency that distinguishes so much of the chemistry of the
more mobile phases. There may not be the same call for ultrafast
LF
FIG.21. Scheme showing the low-energy states of the photosystem W(C0),(4-CNpyr)

(4-CNpyr = 4-cyanopyridine).k , and k2 represent the sums of the radiative and nonradiative rate constants for the decay of the levels 1 and 2, respectively; kd is the rate
constant for photodissociation to W(CO), and 4-CNpyr [reproduced with permission
from (99),
p. 35461.
154
DOWNS AND CREENE
methods of detection and analysis, but the ability to follow the evolution of a solid-state transformation is no less desirable for a proper
kinetic and mechanistic understanding of the change. One such
change currently attracting much interest is intercalation whereby
mobile guest species enter a crystalline host lattice that contains an
interconnected system of empty lattice sites (100). The staging of
intercalation in Ago,,,TiS2and Hg,TiS2 has been effectively witnessed,
for example, as a function of time by high-resolution transmission
electron microscopy. There has been developed, in addition, a reaction
cell that enables intercalation reactions to be monitored in situ using real-time X-ray diffraction techniques. Exploiting the high flux,
white X-ray radiation from a suitable synchrotron radiation source
and energy-dispersive diffraction techniques permits a large energy
window of the powder X-ray diffraction spectrum to be recorded simultaneously at a single fixed detector angle, with acquisition times
in the order of seconds. Figure 22 illustrates strikingly (100) the progressive intercalation of the electron-rich metallocene (T~-C,H,),CO
into a tin disulfide host. Still more dramatic is the very recent tour
de force in protein science in which a group led by Michael Wulff and
Keith Moffat at Chicago has virtually watched a protein function
(101). Working at the European Synchrotron Radiation Facility, this
team has devised a way of collecting pulsed Laue X-ray diffraction
patterns with nanosecond time resolution. Hence, it has been possible
to trace the structural changes that accompany the process of heme
FIG.22. A stack plot showing the evolution of the energy dispersive X-ray diffraction
spectrum of SnS2 following the injection of (q5-C,H,),Co. Each spectrum took 10 s to
record. The signals a t 25 and 29 keV arise from resonances from the tin K., and & core
electrons and have been used to normalize the other signals [reproduced with permission from (IOO),p. 1831.
155
and protein relaxation brought about first by photodissociation and

then by rebinding of the carbon monoxide complex of myoglobin.
V. Experimental Characterization of Reaction Intermediates:

Flow and Other Methods
The basis of flow methods has been outlined earlier. We have to

find a way of generating our intermediate continuously in a flowing
system-sometimes by mixing gases or solutions, sometimes by the
action of a discharge-in a way that can be matched to its kinetic
properties, the flow rate, and the time-scale of the spectroscopic
method of detection. Even here it may be necessary to keep to a minimum the time taken to record the spectrum. In the past much of the
work with flowing gases relied on photographic recording of electronic
emission or absorption spectra, but only for the smallest molecules
are these likely to give detailed information about the vibrational and
other properties of any intermediates. By contrast, current laser techniques allow much more subtle interrogation of transient molecules,
both small and large, in flow systems, leading sometimes to the identification and precise characterization of these molecules ( l o ) ,as well
as reporting on the kinetics of the reactions that feed upon them.
The sort of apparatus that has been used for the study of a gaseous
reaction initiated by a discharge is shown schematically in Fig. 23
(73).The time between initiation and reactant or product detection at
some point downstream in the flow tube can be calculated if the velocity of the gas mixture is known. Detection and monitoring of the
relevant species can be effected by various spectroscopic techniques,
usually mass spectrometry, resonance fluorescence, laser-induced
fluorescence, ESR, or laser magnetic resonance; high-resolution techniques such as microwave and infrared diode laser spectroscopy are
also compatible with the analysis of reactive gas flows, although they
lack the sensitivity of the more widely used methods which are better
suited to kinetic studies. Resonance fluorescence is a well-tried
method for tracking the concentrations of atomic transients (e.g., H,
N, 0, F, C1, or Br), whereas laser-induced fluorescence (LIF) offers
greater sensitivity and applicability, especially for molecular transients (e.g., OH, CN, or CH90), at the cost of greater experimental
complexity. Laser magnetic resonance (LMR) employs a CO, laser to
produce a variety of sharp rovibrational lines to one Qf which a rovibrational line of a radical may often be tuned by the application of
156
DOWNS AND GREENE
To
r
able
ctor
Excimer laser
I
Trigger pulse
FIG.23. Discharge flow apparatus used to study reactions of OH radicals with detection by laser-induced fluorescence. The OH radicals are generated in situ by the reaction between H atoms (themselves formed by the action of a microwave discharge on
a n H2/Hemixture) and NOz. The distance between the points of initiation and detection
is vaned by moving the inlet tube [reproduced with permission from (73),p. 301.
a magnetic field. The biradical methylene is generated in its triplet

electronic ground state by the reaction of oxygen atoms with ketene:
0 + CHz=C=O + 3CH, + O=C=O.
(23)
At slightly higher energies there exists a singlet state. Whereas singlet methylene can be detected by LIF, only the triplet state responds
to LMR. Hence, it will be evident that the detection techniques are
often complementary, and that no one technique is ideal for all transients. The reaction of 3CHzwith ethene,
3CHz+ CHz=CHz+ CH3CH=CHz,
(24)
157
has been studied by monitoring the LMR signal as a function of the

distance between the mixing point and the laser detector with a range
of different ethene concentrations; on this basis the relevant kinetic
parameters have been evaluated (731.
Flow methods have found frequent application in the direct study
of C10 radicals and related species noteworthy as sources and reservoirs of the halogens in the atmosphere (17).In a typical experiment
an inert carrier gas is pumped at pressures near 1-10 torr down a
glass or metal flow tube; the first reactant may be added at a point
upstream to the carrier flow, and the second reactant is added subsequently. The evolution of concentration with distance after mixing is
then followed. Hence, for example, the reactions of C10 generated in
a discharge-flow system have been explored, and mass spectrometric
measurements have been used to determine the overall rate constant
for its decay by the second-order process [Eq. (25a)l (102);resonance
fluorescence associated with C1 atoms indicates that Reaction (25b) is
the major pathway by which C10 decays. Mass spectrometric detection has also been used to determine the relative importance of the
two channels [Eq. (25c)l followed by the reaction between C10 and the
NO3 radical (103), whereas ESR detection of the C10 provides a
means of exploring the kinetics of the reaction with HN03, which
probably proceeds according to Eq. (25d), and so allows this to be discounted as having a significant role in the stratosphere (104).
+ C10
+ C10
C10 + NO,
C10 + HNO,
C10
C10
+=---
all products
c1 + ClOO
ClOO + NO2
OClO + NO2
HCl + NO2 + 0,. (d)
<
(25)
Flow methods are not without their limitations (73). In practice

they are usually limited to reactions with time-scales no shorter than
the millisecond range. Reactions involving more than one reagent
pose particular problems through the time taken for efficient mixing
to occur-in the order of a fraction of a millisecond at a gas pressure
of about 1 torr. A further restriction on kinetic studies is the need to
retain a uniform flow velocity along the entire cross section of the
tube. In fact, these conditions of plug flow can be attained only at
low pressures, so flow experiments are generally restricted to gas
pressures below 10 torr. There are also the seeds of potential compli-
158
DOWNS AND GREENE
cations in unwanted heterogenous reactions occurring at the walls of

the flow tube that may catalyze the reaction under investigation, promote secondary reactions, or scavenge the reaction intermediates.
Most flow studies have been directed toward the kinetic elaboration
of the reactions of intermediates that have been characterized previously by independent means. However, the flow method has also
been exploited rather differently to help elicit detailed structural and
spectroscopic properties of numerous molecules and molecular ions
having but a fleeting existence under normal conditions. Hirota and
his group at the Institute for Molecular Science at Okazaki have
worked wonders with a flow reactor, using discharge, thermal, or photolysis methods to generate the transients and typically surveying the
rotational or rovibrational spectrum of the flowing vapor at high resolution by a microwave or tunable infrared diode laser probe (10).The
methods of production and sampling depend on the lifetime of the
transient-that is, whether it is generated inside or outside the cell
and how fast it must be pumped through the cell to create a workable
effective pathlength for absorption spectroscopy. Instead of probing
the flowing gas laterally for information about changes of composition
with distance, the practice now is usually to probe the gas column
longitudinally so as to intercept as many molecules as possible, and
often with multipass facilities to give optical pathlengths of tens of
meters. For example, the NO, radical-a prominent player on the atmospheric scene-has been generated in a flow system by admixing
NOz with an excess of 0, and characterized by its antisymmetric N-0
stretching mode (v,) by means of a tunable infrared diode laser spectrometer (105). Here Zeeman modulation turns out t o be invaluable
in picking out the absorption lines due to paramagnetic NO, from
among the many lines due to diamagnetic species. The results indicate that NO, is planar and conforms to DBhsymmetry in the 2Ai
ground electronic state with an N-0 bond length of 124.0 pm. On the
other hand, close scrutiny reveals some intriguing anomalies, possibly
arising from interaction between the ground state and a neighboring
2E' excited state. The transient SiClZ, an intermediate in various
chemical processes, has also been generated chemically (by passing
Sic& vapor over heated silicon powder) and characterized in a flow
reactor by both LIF and microwave studies. The microwave measurements fix the equilibrium structure of the molecule with remarkable
precision, giving r&Si-Cl) = 206.531 pm and LCl-Si-C1 = 101.324",
together with harmonic, cubic, and third-order anharmonic potential
constants (106).
159
Flow systems are not limited to the study of gaseous reactions, and
with certain alterations can be applied to the study of reactions in the
liquid phase ( 3 , 73). Because mixing is significantly slower in the liquid phase, attention has to be paid to the design of the mixing chamber, which may include, for example, an elaborate set of baffles to
promote turbulent flow and mixing of two reactant solutions. A familiar and more economical variation on the flow theme is the stoppedflow technique. As with continuous flow, the reactant solutions are
injected from syringes into a mixing cell, whence the mixture flows
into an observation tube. After a few milliseconds the flow is abruptly
arrested. The real-time evolution of the species present in solution is
then monitored; in effect, the method is exactly analogous to conventional batch-mixing kinetic studies and the stopped flow is simply a
device for the rapid mixing of two solutions. The most common detection technique is electronic absorption spectrophotometry, a relatively
blunt instrument when it comes to the identification and characterization of reaction intermediates in the condensed phases. No better
examples of the use of stopped-flow methods can be found than in the
work of Dale Margerum and his colleagues a t Purdue University, who
have developed a pulsed-accelerated-flow spectrophotometer with UVvisible detection that permits the measurement of pseudo-first-order
rate constants as large as 5 X lo5 s- (tllz = 1.4 ps) (107). Typical of
the kinetic studies carried out by this group on a wide range of classical inorganic reactions in aqueous solution are those implying
(i) that hydroxylamine is oxidized by iodine by way of the adduct I2 *
NH20H, which undergoes general-base-assisted deprotonation to give
the intermediate INHOH (108), and (ii) that bromide ions are oxidized by nitrogen trichloride through the intermediate agency of
NBrClz (109).Here we are relying on kinetic measurements mainly, if
not exclusively, for whatever inferences may be drawn about reaction
intermediates. In contrast to the gas phase, solutions labor under the
disadvantage that kinetic measurements are confined to a relatively
narrow temperature range, normally 40C or less, thereby impairing
the precision of any estimates of activation parameters. Indeed, recent years have seen a perceptible trend away from attempting any
such estimates.
From our present perspective, though, no level of confidence in reaction rates can entirely redeem the uncertainties that are apt to cloud
the identities of any transients stopped-flow and related studies may
disclose, whether directly or indirectly. For example, stopped-flow experiments have been carried out t o study the reaction occurring in
160
DOWNS AND GREENE
aqueous solution between cerium(IV) and an excess of peroxotitanium(IV), Ti(02)2+(110). This yields a transient, which, on the evidence of its ultraviolet absorption spectrum, conditions of formation,
and decay properties, was believed to be a superoxo-titanium(1V) species existing apparently in two forms that were tentatively formulated as Ti(OO)3' and Ti(OO)OH2+.A separate investigation of the
decomposition kinetics of this species suggests that it is more likely
to be TiO(H02.)2+,
that is, a complex of the aquated Ti02+ion with
the perhydroxyl radical (110). ESR measurements may shed a more
distinctive light on paramagnetic intermediates. Thus, the oxidation
of chromium(II1) to chromium(VI) by hydrogen peroxide in basic media occurs via a branching mechanism: from the initial formation of a
chromium(II1)-peroxide intermediate, one pathway leads to a chromium(IV) intermediate and a second pathway leads to a chromium(V)
intermediate, which can be identified empirically by its ESR spectrum
(111).Nevertheless, doubts assailing us about the true nature of intermediates like these contrast starkly with the conviction high-resolution studies bring to the identification and characterization of a gaseous intermediate like Sic&.
For all its immense power and versatility, NMR spectroscopy has
kept a relatively low profile in the story so far. There is no doubting
the potential of real-time NMR measurements to report directly on a
reaction and on all the components of the reaction mixture. There
are certainly ways of allying rapid mixing, stopped-flow, temperaturejump or other methods with NMR detection (3, 112),but no technical
stratagem can remedy the want of sensitivity that is inherent in the
NMR experiment. Nevertheless, the ability of the technique specifically to monitor a free ligand can be turned to good account, for example to resolve the three stages in the displacement of three dmso molecules (dmso = Me2SO)in the cation [Al(dmso),13+by one terdentate
ligand 2,2': 6'2"-terpyridine (terpy), XIV, in nitromethane solution
(113).The implication is that there are two intermediates involving
XIV
161
mono- or bidentate coordination of the terpy ligand to the metal center. Rates for the three stages have been determined using a special
stopped-flow apparatus, with H NMR measurements reporting on
the release of coordinated dmso molecules. Hence, activation parameters have been determined for the first bond formation and the two
slower chelate-ring closures that ensue. More recent experiments
have turned to the much more elaborate puzzle of protein folding,
where NMR studies are privy to unique structural insights (114). The
resulting spectra relay information not only about kinetic folding
events but also about the structures of folding intermediates, partly
folded states, peptide fragments, and unfolded or denatured proteins.
Recent developments in mass spectrometry, with electrospray ionization, provide another relatively fast device for monitoring solution
processes in real time, and one that has been applied to the teasing
problems of protein folding (115, 116). Such folding is assisted by a
host of helper proteins including the molecular chaperones, which are
thought to safeguard newly synthesized proteins from unproductive
interactions that may lead to aggregation. This raises the problem of
how to determine the conformational properties of a relatively small
substrate protein in the presence of a large chaperonin oligomer like
GroEL. One new perspective shown schematically in Fig. 24 (115)
seeks to elicit the information from hydrogen exchange protection (a
technique that has revolutionized studies of protein folding in vitro)
specifically relating to complexes relevant to folding in the cell. For
example, a partially folded protein bound within the GroEL central
cavity can be directly studied by electrospray ionization mass spectrometry (ESI-MS), allowing the visualization of both the GroEL subunits and the ligand in the same spectrum. When combined with hydrogen exchange labeling, this approach admits the study of
individual components of the complex that retain the history of the
complexed state by virtue of their deuterium content; it thus obviates
the need to quench exchange and dissociate the complex before measurement. Hence, we are able in effect to measure selectively the protection of individual components in a mixture of species and in proteins that are well beyond the molecular mass range open to detailed
NMR analysis, and also to observe states too short-lived to be characterized, say, by X-ray crystallography. In addition, differences in the
cooperativity of folding events can be detected directly by the ESI-MS
method (116).
Yet another option is being elaborated by Fraser Armstrong and his
group a t Oxford through fast-scan electrochemical techniques, which
have now advanced to the point where they are capable of probing
162
DOWNS AND GREENE

DETECTED IN
MASS
SPECTROMnER
GroEL
1-
Compkx Is diluted
1 &fold into H p solution.
Compkx dbscciates in maw

spectrometer. &and retains
bbelling. Extent of labelli k
according to iu stmctum
FIG.24. Schematic diagram outlining the experiment designed to monitor hydrogen

exchange in the GroEL: [3SS] BLA complex (BLA = bovine a-lactalbumin) by electrospray ionization mass spectrometry (ESI-MS). Before formation of the complex, all exchangeable sites in a-lactalbumin were deuterated by incubation of the apoprotein in
DzO. Hydrogen exchange is initiated by a 10-fold dilution of the complex in H,O, pH
5.0, a t 4C. Dissociation of the GroEL into monomers in the gas phase releases the
bound polypeptide ligand while maintaining the protection in the substrate protein. In
this way, the rate of hydrogen exchange of [3SS1 a-lactalbumin (and GroEL) in the
complex can be measured directly as a change in their masses a s a function of the
incubation time in HzO by ESI-MS [reproduced with permission from (115),p. 6471.
transient intermediates formed during redox reactions (117).In the

technique of protein-film voltammetry, the redox protein molecules
under investigation are adsorbed on the rotating electrode surface of
a voltammetric assembly and interrogated there electrochemically.
Cyclic voltammetry at scan rates up to 1000 V s-l may then reveal
pairs of oxidation-reduction peaks, which may serve as markers by
which the status of redox centers within the protein can be identified
and quantified. In effect, we have an interactive spectrum, albeit
one providing no structural information; for that we must turn to true
spectroscopic methods, such as ESR, to examine the species generated
in solution at the appropriate potentials. Protein-film voltammetry
has already been successfully exploited to track active-site redox
transformations over a range of conditions and so reveal the intramolecular electron relays operating in the multicentered enzyme. Large
systems like these hold out little prospect of furnishing the sort of
structurally explicit information that is now available for simple molecules like SizHzor Cr(CO),. On the other hand, what we lose in struc-
163
tural definition may well be compensated by a clearer perception of

the thermodynamic and kinetic properties that determine the reactivity of a given intermediate.
Problems of reliable identification are no strangers to studies that
concentrate on kinetics, whether they involve the stopped-flow technique, relaxation methods, or other strategies. We may know very
well the speed and order of a particular reaction; we may even be
well versed in the dependence of the kinetic parameters on various
conditions; however, we cannot always be sure of the exact nature of
the reagents and the products, still less the intermediates formed on
the way. To that extent, any mechanistic interpretation is bound to
be impaired. Indeed, the situation has something of the surreal about
it, reminiscent of the conversation between Alice and the Cheshire
Cat:
Would you tell me, please, which way I ought to go from here?
That depends a good deal on where you want to get to, said the
Cat. I dont much care where . . . , said Alice. Then it doesnt
matter which way you go, said the Cat. . . . So long as I get
somewhere, Alice added as an explanation. Oh, youre sure to do
that, said the Cat, if you only walk long enough.
Lewis Carroll, Alices Adventures in Wonderland
Vi. Conclusions
In this account we have sought to highlight reaction intermediates

and the crucial bearing they have on the way chemical processes actually take place. We have concentrated on how such intermediates can
be directly observed, identified, characterized, and tracked by their
spectroscopic properties. Frustration of the reactions t o which the intermediates are natural prey can be achieved by control of temperature, pressure, and environment, and in these circumstances it may
be perfectly possible to form a clear and detailed picture of a given
intermediate on the basis of the same sort of spectroscopic methods
that are used to study normal, long-lived species. To chart the chemistry of such an intermediate requires a closer acquaintance with the
reality of normal conditions, and a t least some concession not only to
mobility, concentration, and temperature, but also to time as a variable; this commonly involves some form of time-resolved spectroscopic
164
DOWNS AND GREENE
probe, as in flash photolysis, or the induction of reaction not in a

static but in a uniformly flowing gas or liquid.
A recurring example has been chromium pentacarbonyl, Cr(CO), ,
an intermediate most easily generated photochemically en route t o
substitution products of chromium hexacarbonyl (60, 611. Matrix-isolation experiments were the first to fix the structure of this molecule
and to alert us to its extraordinary reactivity. Experiments with fluid
samples were then needed t o elaborate on the ability of Cr(CO)5to
bind supposedly inert molecules such as Xe, H2, and CH, , and timeresolved spectroscopy has opened our eyes to the energetics of these
interactions as well as the photochemistry and kinetics of the processes involving Cr(CO)5. So strongly is CI-(CO)~solvated that it is
hard to conceive of the molecule in solution without its token ligand.
Naked Cr(CO)5plays a role, it is true, but certainly does not live long
enough t o wait for a collision in dilute solution; instead, in the space
of picoseconds, it picks up a token ligand from within its solvent cage
via the pseudorotation illustrated in Fig. 13. Hence, the orthodox notion of a simple dissociative mechanism for a substitution reaction
starting from saturated 18-electron Cr(CO), and involving the intermediate formation of an unsaturated 16-electron intermediate must
give place to an interchange model that is altogether more intricate
and subtle and in which the solvent is no mere spectator. Even now,
for all our knowledge of the chemistry of CI-(CO)~at time-scales of a
nanosecond and longer, controversy and doubts continue to surround
the very fast processes implicating the fragment.
No less striking have been the advances made in our understanding
of the substitution reactions peculiar t o the much studied binuclear
iron carbonyl, VIII (83-85). Here, too, a variety of experimental techniques has been brought to bear on the intermediate photoproductsone a binuclear species IX and the other a mononuclear radical Xwhich differ hugely in their reactivity.
The ability of a reaction intermediate such as CI-(CO)~
to bind methane may be counterintuitive but commands attention for its significance in relation to C-H bond activation, a teasing but vital issue in
the context of the chemical industry (118).Numerous complexes of
alkanes with unsaturated transition-metal fragments (including
atoms and ions) have now been detected in both the gas and condensed phases. It is not surprising that all of them are unstable at
room temperature. Experiments in which alkanes undergo oxidative
addition to, or reductive elimination from, transition-metal complexes
have revealed, nonetheless, the intermediacy of alkane complexes.
165
For example, TRIR studies of liquid xenon solutions of the rhodium

complex (v5-CSMe5)Rh(CO),
witness on photolysis the expulsion of CO
and formation of a transient characterized by its v(C-0) mode; a similar band is observed a t much the same frequency when liquid krypton is used as the solvent but belongs to a considerably more reactive
carrier (119).The transients are most likely to be (q5-CSMe5)Rh(CO)(Xe)and (v5-C5Me5)Rh(CO)(Kr),
respectively. In the presence of a
small amount of cyclohexane, however, the intermediates transform
rapidly into the insertion product (v5-C5Me,)Rh(CO)(C,H,,)(H).The kinetics of formation of this product in liquid krypton are consistent not
with a dissociative mechanism but with a reaction scheme incorporating a n equilibrium between krypton and cyclohexane complexes of (v5CSMeS)Rh(CO)prior to oxidative addition. The failure to observe a
distinct v(C-0) mode for the alkane complex ( vS-C5Me5)Rh(CO)(C6Hl2)
is probably due to overlap with the corresponding band of the krypton
complex. Unambiguous though the case for transition-metal alkane
complexes may be on the evidence of these and many other studies
(1181, the mode of coordination of the alkane has yet to be fathomed
properly. Plainly these are but the early stages of a mechanistic expedition that is likely to take many years to complete.
In very different territory we have noted how a molecule such as
C10 that is normally short-lived in the laboratory is still capable of
exercising a major influence on the fate of stratospheric ozone (17).
There is particular interest, for example, in the dimerization of C10
and subsequent photolysis of the products [see Eqs. 25(a) and 25(b)l
because these processes may have a marked impact on the ozone
depletion of the atmosphere above the polar regions. Of especial note
in this context has therefore been the first preparation of another
labile molecule with the composition Cl2OZby halogen exchange between gaseous FC10, and AICIB(62, 120); trapping in a solid noblegas matrix enables this species to be identified as chloryl chloride,
C1C102,XV, on the basis of its infrared and ultraviolet spectra. From
the vibrational frequencies of four independent isotopomers, the
known properties of related compounds, and the results of ab initio
calculations, the following geometric parameters have been estimated: r(C1-C1) = 222 pm, r(Cl=O) = 144.0 pm, LCl-Cl=O =
103.5", and ~ 0 = C 1 = 0= 116.0'. What is intriguing about the photochemistry of this new isomer of Cl2OZis that it decays under appropriate conditions not only t o the more familiar peroxo form ClOOCl,
XVI, but also to a third, photolabile isomer, chlorine chlorite, ClOClO,
XVII. At room temperature and partial and total pressures of 1 and
166
DOWNS AND GREENE
4 torr, respectively, CICIOz has a half-life of 1 min. Matrix studies

involving selective irradiation adduce the first evidence for a second
xv
XVI
XVII
photodissociation channel disposing of ClOOCl (and giving 2C10), and

also make it clear that dimerization of C10 can lead to both ClOClO
and ClOOCl. These results must surely be heeded by gas-phase kineticists and others seeking to model the atmospheric fates of the chlorine oxides.
Everythings got a moral, if you can only find it, and to this tale
there are not one but two morals. The first is that reaction intermediates, however ephemeral, can and should be explored directly. The
second is that no one experimental or theoretical technique has a monopoly of the truth we seek; to arrive at a comprehensive and realistic
picture the best policy must be to draw on as many sources as possible. It is true that the Promised Land of a proper understanding of
the mechanisms of major chemical reactions is still a great way off
and that the intermediates by which the reactions proceed are not
going to be quick to yield up their secrets. However, it is equally true
that without these secrets even the most inspired perceptions of
mechanism are either incomplete or insecure. In Robin Perutzs trenchant words (611, If you dont understand the intermediates, you
dont understand the mechanisms.
Not only is there but one way of doing things rightly, but there is
only one way of seeing them, and that is, seeing the whole of them.
J. Ruskin, The Two Paths
REFERENCES
1. Tobe, M. L. Inorganic Reaction Mechanisms; Nelson: London, 1972.
2. Jordan, R. B. Reaction Mechanisms of Inorganic and Organometallic Systems;
Oxford University Press: New York and Oxford, 1991.
3. Wilkins, R. G. Kinetics and Mechanism of Reactions of Transition Metal Complexes; 2nd ed., VCH: Weinheim and New York, 1991.
4. Darensbourg, D. J. Adu. Organomet. Chem. 1982,21, 113.
5. Howell, J.A. S.; Burkinshaw, P. M. Chem. Rev. 1983,83, 557.
COMING TO GRIPS WITH REACTrVE INTERMEDIATES
167
6. Woodward, S. In Comprehensive Organometallic Chemistry 11, Vol. 5 (J. A. Labinger and M. J . Winter, Eds.), Chap. 4, pp. 215-280, Pergamon: Oxford, 1995.
7. Richter-Addo, G. B.; Legzdins, P. Metal Nitrosyls; Oxford University Press: New
York and Oxford, 1992.
8. Niu, S.; Hall, M. B. J. Am. Chem. Soc. 1997,119, 3077 and references cited therein.
9. (a) Seder, T. A.; Church, S. P.; Ouderkirk, A. J.; Weitz, E. J. Am. Chem. Soc. 1985,
107, 1432; (b)Fletcher, T. R.; Rosenfeld, R. N. J. Am. Chem. Soc. 1985, 107, 2203.
10. (a) Hirota, E. High-Resolution Spectroscopy of Transient Molecules; Springer:
Berlin, 1985; (b) Hirota, E. Irrt. Reu. Ph.ys. Chem. 1989,8, 171; (c) Hirota, E. Annu.
Reu. Phys. Chem. 1991,42, 1; (d) Hirota, E. Chem. Rev. 1992, 92, 141; ( e )Hirota,
E. Ann. Rep. Prog. Chem., Sect. C, Phys. Chem. 1994, 91, 3.
11. Cox, P. A. The Elements: Their Origin, Abundance, and Distribution, pp. 93126, Oxford University Press: Oxford, 1989.
12. Manson, E. L., Jr.; Clark, W. W.; De Lucia, F. C.; Gordy, W. Phys. Rev. A 1977,
15, 223.
13. Huber, K. P.; Herzberg, G . Molecular Spectra and Molecular Structure. IV.Constants of Diatomic Molecules; Van Nostrand Reinhold: New York, 1979.
14. Ram, R. S.; Bernath, P. F. J . Mol. Spectrosc. 1996, 180, 414.
15. Lorenz, M.; Agreiter, J.; Smith, A. M.; Bondybey, V. E. J. Chem. Phys. 1996,
104, 3143.
16. Bredohl, H.; Dubois, I.; Houbrechts, Y.; Nzohabonayo, P. J . Mol. Spectrosc. 1985,
122, 430.
17. ( a ) Wayne, R. P. Chemistry of Atmospheres, 2nd ed., Clarendon Press: Oxford,
1991; (b) Wayne, R. P.; Poulet, G.; Biggs, P.; Burrows, J. P.; Cox, R. A,; Crutzen,
P. J.; Hayman, G. D.; Jenkin, M. E.; Le Bras, G.; Moortgat, G. K.; Platt, U.;
Schindler, R. N. Atmos. Enuiron. 1995,29, 2677.
18. Anderson, J . G.; Toohey, D. W.; Brune, W. H. Science 1991,251, 39.
19. (a) Cohen, E. A.; Pickett, H. M.; Geller, M. J. Mol. Spectrosc. 1984, 106, 430; (b)
Burkholder, J. B.; Hammer, P. D.; Howard, C. J.; Maki, A. G.; Thompson, G.;
Chackerian, C., Jr. J. Mol. Spectrosc. 1987, 124, 139; (c) McLoughlin, P. W.; Park,
C. R.; Wiesenfeld, J. R. J . Mol. Spectrosc. 1993, 162, 307.
20. Herzberg, G.; Johns, J. W. C. Proc. Roy. Soc. l967,298A, 142.
21. Kawaguchi, K. J . Chem. Phys. 1992,96, 3411; Can. J . Ph,ys. 1994, 72, 925.
22. Kawashima, Y.; Kawaguchi, K.; Hirota, E. J. Chem. Phys. 1987,87, 6331.
23. Ebsworth, E. A. V.; Rankin, D. W. H.; Cradock, S. Structural Methods in Inorganic Chemistry, 2nd ed., pp. 334-343, Blackwell Scientific Publications: Oxford, 1991.
24. See, for example, Blake, A. J.; Brain, P. T.; McNab, H.; Miller, J.; Morrison,
C. A,; Parsons, S.; Rankin, D. W. H.; Robertson, H. E.; Smart, B. A. J. Phys. Chern.
1996, 100, 12280.
25. Fujiwara, H.; Egawa, T.; Konaka, S. J . Mol. Struct. 1995, 344, 217.
26. Fujiwara, H.; Egawa, T.; Konaka, S. J . Am. Chem. Soc. 1997,119, 1346.
27. Molnar, J.; Marsden, C. J.; Hargittai, M. J . Phys. Chern. 1995, 99, 9062.
28. Hedberg, K.; Hedberg, L.; Buhl, M.; Bethune, D. S.; Brown, C. A,; Johnson, R. D.
J. Am. Chem. Soc. 1997,119, 5314.
29. Carlowitz, M. V.; Oberhammer, H.; Willner, H.; Boggs, J. E. J. Mol. Struct. 1983,
100, 161.
30. Cleaver, W. M.; Spath, M.; Hnyk, D.; McMurdo, G.; Power, M. B.; Stuke, M.;
Rankin, D. W. H.; Barron, A. R. Orgunometullzcs 1995, 14, 690.
31. Legon, A. C. Chem. Br. 1990,26, 562.
168
DOWNS AND GREENE
32. Legon, A. C. In Atomic and Molecular Beam Methods, Vol. 2 (G. Scoles, Ed.),
pp. 289-308, Oxford University Press: New York, 1992.
33. Legon, A. C. J. Chem. SOC.,Chem. Commun. 1996, 109.
34. Bloemink, H. I.; Hinds, K.; Legon, A. C.; Thorn, J . C. Chem. Eur. J. 1995, I, 17.
35. Bloemink, H. I.; Hinds, K.; Holloway, J . H.; Legon, A. C. Chem. Phys. Lett. 1995,
245, 598.
36. Campbell, E. J.; Kukolich, S. G. Chem. Phys. 1983, 76, 225.
37. Campbell, E. J.; Read, W. G. J. Chem. Phys. 1983,78, 6490.
38. Bloemink, H. I.; Evans, C. M.; Holloway, J. H.; Legon, A. C. Chem. Phys. Lett.
1996,248, 260.
39. Barnes, M.; Hajigeorgiou, P. G.; Kasrai, R.; Merer, A. J.; Metha, G. F. J. Am.
Chem. SOC.1995, 1 17, 2096.
40. Davies, P. B. Chem. SOC.Rev. 1995,24, 151.
41. Hirahara, Y.; Ohshima, Y.; Endo, Y. J . Chem. Phys. 1994,101, 7342.
42. Yamada, C.; Hirota, E. Phys. Rev. Lett. 1986, 56, 923.
43. Bogey, M.; Bolvin, H.; Demuynck, C.; Destombes, J. L. Phys. Reu. Lett. 1991,
66, 413.
44. Colegrove, B. T.; Schaefer, H. F., 111. J. Phys. Chem. 1990,94, 5593.
45. Cordonnier, M.; Bogey, M.; Demuynck, C.; Destombes, J.-L. J. Chem. Phys. 1992,
97, 7984.
46. Harper, W. W.; Ferrall, E. A.; Hilliard, R. K.; Stogner, S. M.; Grev, R. S.;
Clouthier, D. J. J . Am. Chem. SOC.1997,119, 8361.
47. See, for example, (a) Davies, P. B.; Martineau, P. M. J . Appl. Phys. 1992,71, 6125;
(b) Naito, S.; Ito, N.; Hattori, T.; Goto, T. Jpn. J. Appl. Phys., Part 1 1994, 33,
5967; (c) Fukuzawa, T.; Obata, K.; Kawasaki, H.; Shiratani, M.; Watanabe, Y. J.
Appl. Phys. 1996,80, 3202.
48. Perutz, R. N. Chem. Rev. 1985,85, 77.
49. Downs, A. J.; Pulham, C. R. Adu. Inorg. Chem. 1994,41, 171.
50. Sawyer, D. T. Oxygen Chemistry, pp. 156-158, Oxford University Press: New
York, 1991.
51. (a) Almond, M. J.; Downs, A. J . Adv. Spectrosc. 1989, 17, 1-511; (b) Andrews, L.,
and Moskovits, M., Eds., Chemistry and Physics of Matrix-Isolated Species,
North Holland: Amsterdam, 1989.
52. Downs, A. J . In Low Temperature Molecular Spectroscopy (R. Fausto, Ed.), pp.
1-93, NATO AS1 Series, Series C: Vol. 483, Kluwer: Dordrecht, 1996.
53. Bondybey, V. E.; Smith, A. M.; Agreiter, J . Chem. Rev. 1996, 96, 2113.
54. Chetwynd-Talbot, J.; Grebenik, P.; Perutz, R. N. Inorg. Chem. 1982,21, 3647.
55. Tague, T. J., Jr.; Andrews, L. J . Am. Chem. SOC.1994, 116, 4970.
56. Poliakoff, M.; Weitz, E. Acc. Chem. Res. 1987,20, 408.
57. Crayston, J. A.; Almond, M. J.; Downs, A. J.; Poliakoff, M.; Turner, J . J. Znorg.
Chem. 1984,23, 3051.
58. Muller, J. J . Am. Chem. SOC.1996, 118, 6370.
59. Fletcher, S. C . ; Poliakoff, M.; Turner, J. J. Zmorg. Chem. 1986,25, 3597.
60. Turner, J. J. In Photoprocesses in Transition Metal Complexes, Biosystems and
Other Molecules. Experiment and Theory (E. Kochanski, Ed.), pp. 125-140,
NATO AS1 Series, Series C: Vol. 376, Kluwer: Dordrecht, 1992.
61. Perutz, R. N. Chem. SOC.Reu. 1993,22, 361.
62. Jacobs, J.; Kronberg, M.; Muller, H. S. P.; Willner, H. J. Am. Chem. SOC.1994,
116, 1106.
COMING TO GRIPS WITH REACTrVE INTERMEDIATES
169
63. Bell, T. W.; Haddleton, D. M.; McCamley, A.; Partridge, M. G.; Perutz, R. N.;
Willner, H. J . Am. Chem. SOC.1990, 112, 9212.
64. Poliakoff, M.; Turner, J. J. Adu. Spectrosc. 1995, 23, 275.
65. Van der Veken, B. J. In Low Temperature Molecular Spectroscopy (R. Fausto,
Ed.), pp. 371-420, NATO AS1 Series, Series C: Vol. 483, Kluwer: Dordrecht, 1996.
66. Upmacis, R. K.; Poliakoff, M.; Turner, J. J. J. Am. Chem. SOC.
1986, 108, 3645.
67. Duckett, S. B.; Haddleton, D. M.; Jackson, S. A,; Perutz, R. N.; Poliakoff, M.;
Upmacis, R. K. Organometallics 1988, 7, 1526.
68. Turner, J. J.; Simpson, M. B.; Poliakoff, M.; Maier, W. B., 11. J. Am. Chem. SOC.
1983,105, 3898.
69. Howdle, S. M.; Healy, M. A,; Poliakoff, M. J. Am. Chem. SOC.
1990, 112, 4804.
70. (a) McHugh, M. A,; Krukonis, V. J . Supercritical Fluid Extraction: Principles
and Practice, 2nd ed., Butterworth-Heinemann: Boston, 1994; (b) Poliakoff, M.;
Howdle, S. Chem. Br. 1995, 31, 118.
71. Walsh, E. F.; Popov, V. K.; George, M. W.; Poliakoff, M. J. Phys. Chem. 1995,
99, 12016.
72. Rathke, J. W.; Klingler, R. J.; Krause, T. R. Organometallics 1991, 10, 1350.
73. Pilling, M. J.; Seakins, P. W. Reaction Kinetics; Oxford University Press: Oxford, 1995.
74. Poliakoff, M.; Weitz, E. Adu. Organomet. Chem. 1986, 25, 277.
75. Turner, J. J . In Photoprocesses in Transition Metal Complexes, Biosystems and
Other Molecules. Experiment and Theory (E. Kochanski, Ed.), pp. 113-123,
NATO AS1 Series, Series C: Vol. 376, Kluwer: Dordrecht, 1992.
76. Perutz, R. N. In Low Temperature Molecular Spectroscopy (R. Fausto, Ed.), pp.
95-124, NATO AS1 Series, Series C: Val. 483, Kluwer: Dordrecht, 1996.
77. El-Sayed, M. A,, Tanaka, I., and Molin, Y., Eds., Ultrafast Processes in Chemistry and Photobiology; Blackwell Science: Oxford, 1995.
78. Farhataziz; Rodgers, M. A. J. Radiation Chemistry-Principles
and Applications; VCH: New York, 1987.
79. Goldstein, S.; Czapski, G. Znorg. Chem. 1997, 36, 4156.
80. Tripathi, G. N. R. Adu. Spectrosc. 1989, 18, 157.
81. George, M. W.; Poliakoff, M.; Turner, J. J. Anal-yst 1994, 119, 551.
82. Sun, X.-Z.; George, M. W.; Kazarian, S. G.; Nikiforov, S. M.; Poliakoff, M.
J . Am. Chem. Sac. 1996,118, 10525.
83. (a) Hooker, R. H.; Mahmoud, K. A,; Rest, A. J. J . Chern. SOC.,Chem. Commun.
1983, 1022; (b) Hepp, A. F.; Blaha, J. P.; Lewis, C.; Wrighton, M. S. Organometallics 1984, 3, 174.
84. (a) Moore, B. D.; Simpson, M. B.; Poliakoff, M.; Turner, J. J. J . Chem. SOC.,Chem.
Commun. 1984, 972; (b) Dixon, A. J.; George, M. W.; Hughes, C.; Poliakoff, M.;
Turner, J . J . J. Am. Chem. Sac. 1992,114, 1719.
85. Vitale, M.; Lee, K. K.; Hemann, C. F.; Hille, R.; Gustafson, T. L.; Bursten, B. E.
J. Am. Chem. SOC.1995,117, 2286.
86. Hamaguchi, H.-o.; Gustafson, T. L. Annu. Reu. Phys. Chen. 1994,45, 593.
87. Kitagawa, T.; Ogura, T. Prog. Inorg. Chem. 1997,45, 431.
88. Schelvis, J. P. M.; Deinum, G.; Varotsis, C. A,; Ferguson-Miller, S.; Babcock,
G. T. J . Am. Chem. SOC.1997, 119, 8409.
89. Osman, R.; Pattison, D. I.; Perutz, R. N.; Bianchini, C.; Casares, J . A.; Peruzzini,
1997, 119, 8459.
M. J. Am. Chem. SOC.
90. Colombo, M.; George, M. W.; Moore, J. N.; Pattison, D. 1.; Perutz, R. N.; Virrels,
I. G.; Ye, T.-Q. J. Chem. SOC.,Dalton Trans., 1997, 2857.
170
DOWNS AND GREENE
91. Cheng, P. Y.; Zhong, D.; Zewail, A. H. J. Chem. Phys. 1996, 105, 6216.
92. Zhong, D.; Ahmad, S.; Zewail, A. H. J . Am. Chem. SOC.1997, 119, 5978.
93. Stephens, J. W.; Morter, C. L.; Farhat, S. K.; Glass, G. P.; Curl, R. F. J. Phys.
Chem. 1993,97, 8944.
94. (a) Ischenko, A. A,; Schafer, L.; Luo, J . Y.; Ewbank, J. D. J. Phys. Chem. 1994,
98, 8673; (b) Ischenko, A. A,; Ewbank, J. D.; Schafer, L. J. Phys. Chem. 1995,
99, 15790.
95. (a) Williamson, J. C.; Zewail, A. H. J. Phys. Chem. 1994,98, 2766; (b) Dantus, M.;
Kim, S. B.; Williamson, J . C.; Zewail, A. H. J . Phys. Chem. 1994, 98, 2782.
96. Williamson, J . C.; Cao, J.; Ihee, H.; Frey, H.; Zewail, A. H. Nature 1997,386, 159.
97. Turner, J. J.; George, M. W.; Johnson, F. P. A.; Westwell, J. R. Coord. Chem. Reu.
1993, 125, 101.
98. Perng, J.-H.; Zink, J . I. Znorg, Chem. 1990,29, 1158.
99. Glyn, P.; Johnson, F. P. A,; George, M. W.; Lees, A. J.; Turner, J. J. Znorg. Chem.
1991,30, 3543.
100. OHare, D. In Inorganic Materials, 2nd ed. (D. W. Bruce and D. OHare, Eds.),
Chap. 4, pp. 171-254, Wiley: Chichester, 1996.
101. (a) Srajer, V.; Teng, T.-y.; Ursby, T.; Pradervand, C.; Ren, Z.; Adachi, S.-i.; Schildkamp, W.; Bourgeois, D.; Wulff, M.; Moffat, K. Science 1996,274, 1726; (b) Eaton,
W. A,; Henry, E. R.; Hofrichter, J. Science 1996,274, 1631.
102. Clyne, M. A. A,; McKenney, D. J.; Watson, R. T. J. Chem. Soc., Faraday Trans. I
1975, 71, 322.
103. Kukui, A.; Jungkamp, T. P. W.; Schindler, R. N. Ber. Bunsenges. Phys. Chem.
1994,98, 1619.
104. Zagogianni, H.; Mellouki, A.; Poulet, G. C. R. Acad. Sci. Paris, Ser. IZ 1987, 304,
573.
105. (a) Ishiwata, T.; Tanaka, I.; Kawaguchi, K.; Hirota, E. J. Chem. Phys. 1985, 82,
2196; (b) Kawaguchi, K.; Hirota, E.; Ishiwata, T.; Tanaka, I. J. Chem. Phys. 1990,
93, 951.
106. Fujitake, M.; Hirota, E. Spectrochim. Acta 1994, 50A, 1345.
107. Bowers, C. P.; Fogelman, K. D.; Nagy, J. C.; Ridley, T. Y.; Wang, Y. L.; Evetts,
S. W.; Margerum, D. W. Anal. Chem. 1997,69, 431.
108. Liu, R. M.; McDonald, M. R.; Margerum, D. W. Znorg. Chem. 1995,34, 6093.
109. Gazda, M.; Kumar, K.; Margerum, D. W. Inorg. Chem. 1995,34, 3536.
120. Bourke, G. C. M.; Thompson, R. C. Inorg. Chem. 1987,26, 903.
Rotzinger, F. P.; Griitzel, M. Inorg. Chem. 1987,26, 3704.
111. Knoblowitz, M.; Morrow, J. I. Znorg. Chem. 1976, 15, 1674.
112. Moore, P. Pure Appl. Chem. 1985,57, 347.
113. Brown, A. J.; Howarth, 0. W.; Moore, P.; Parr, W. J. E. J. Chem. SOC.,Dalton
Trans. 1978, 1776.
114. Dyson, H. J.; Wright, P. E. Annu. Reu. Phys. Chem. 1996,47, 369.
115. Robinson, C. V.; Gross, M.; Eyles, S. J.; Ewbank, J . J.; Mayhew, M.; Hartl, F. U.;
Dobson, C. M.; Radford, S. E. Nature 1994,372, 646.
116. Hooke, S. D.; Eyles, S. J.; Miranker, A.; Radford, S. E.; Robinson, C. V.; Dobson,
C. M. J. Am. Chem. SOC.1995,117, 7548.
117. (a) Armstrong, F. A.; Heering, H. A,; Hirst, J. Chem. Soc. Reu. 1997, 26, 169;
(b) Heering, H. A.; Weiner, J . H.; Armstrong, F. A. J. Am. Chem. SOC.,1997,
119, 11628.
118. (a) Crabtree, R. H. Chem. Reu. 1996, 95, 987; (b) Hall, C.; Perutz, R. N. Chem.
Reu. 1996,96, 3125.
171
119. (a) Bengali, A. A,; Arndtsen, B. A.; Burger, P. M.; Schultz, R. H.; Weiller, B. H.;
Kyle, K. R.; Moore, C. B.; Bergman, R. G. Pure Appl. Chem. 1995, 67, 281; (b)
Arndtsen, B. A,; Bergman, R. G.; Mobley, T. A.; Peterson, T. H. Acc. Chem. Res.
1995,28, 154.
120. Miiller, H. S. P.; Willner, H. Znorg. Chem. 1992,31, 2527.
ADVANCES IN INOKGANI( CHEMISTKY VOL
46
TOWARD THE CONSTRUCTION OF FUNCTIONAL

SOLID-STATE SUPRAMOLECULAR METAL
COMPLEXES CONTAINING COPPER(1)
AND SILVER(1)
MEGUMU MUNAKATA, LIANG PING WU,
and TAKAYOSHI KURODA-SOWA
Department of Chemistry, Kinki University, Kowakae 3-4-1, Higashi-Osaka 577, Japan
I. Introduction
11. Helical Frameworks
A. Infinite Single-Helical Complexes
B. Double-Helical Complexes
C. Triple-Helical Complexes
111. S----SContact-Assembled Frameworks
A. dmit and the Related Ligands
B. BEDT-TTF
C. TTC,,-TTF Systems
D. C,H&
IV. Hexagonal Frameworks and Graphite-like Structures
A. Metal Cyanide-Regulated by the Metal Ion
B. Pyrazine Systems-Regulated by Substituents
C. Phenazine and Benzothiadiazole Systems-Regulated by Counteranions
V. Hydrogen-Bond-Assembled Frameworks
A. Three-Dimensional Supramolecular Cu(1) Complexes with Channels
B. Hydrogen-Bonding- and a-a-Stacking-Assembled Cu(I) Complexes
VI. a-a-Interaction-Assembled Frameworks
A. Intermolecular a--TIInteraction in Discrete Coordination Compounds
B. Inter- and Intrapolymer a-a Interaction
VII. Diamondoid Frameworks
A. Bridged by Pyridine or Pyrazine Derivatives
B. Bridged by Bisnitrile Ligands
C. Other Bridging Ligands
D. T-a Interaction in Diamondoid Frameworks
VIII. Other Frameworks Based on Covalent Bonds
A. Infinite-Chain Structures
B. Two-Dimensional Structures
C. Three-Dimensional Structures
IX. Concluding Remarks
References
173
Copyright 11 1999 hy Academic Press
N I rights of reproduction i n any form reservcd.
O ~ ~ B - B H W Y Y $za.nn
174
MUNAKATA, WU, AND KURODA-SOWA
1. Introduction
Recent years have witnessed considerable interest in the development of rational synthetic routes to supramolecular architecture from
self-assembly of component metal complexes. These solid materials
with well-defined, discrete network topologies are attractive to chemists not only for aesthetic reasons but also for their potential applications in many areas (1-13). A major difficulty in controlling such a
self-assembling process is the fact that it involves several stages, such
as recognition between the components, correct orientation so as to
allow growth, and termination of the process leading to the predesigned species ( 1 4 ) .In addition, the intermolecular forces that control
these stages are directional and weak enough to enable closely related
molecules to give widely different aggregates under slightly modified
conditions. Usually, the synthetic strategy involves selecting the coordination geometry of metal ions, the chemical structure of organic
ligands, and the favorable reaction conditions such as solvents and
counterions. In this context, a fundamental principle of the designed
chemistry of the intermolecular bond is the utilization of the intermolecular forces, which may consist of hydrogen bonding, S----Scontacts,
aromatic stackings, host-guest interactions, van der Waals forces,
and other electrostatic attractions. Unfortunately, these forces are
much less well understood than classical chemical bonds in terms of
their energetic and geometric properties. Therefore, the control of molecular assembly using supramolecular interactions is probably a new
challenge to chemists.
The copper(1) and silver(1) ions are regarded as extremely soft acids
favoring coordination to soft bases, such as ligands containing S and
unsaturated N (15, 16). Copper(1) and silver(1) complexes with these
soft ligands give rise to an interesting array of stereochemistries and
geometric configurations, with the coordination numbers of two to six
all occurring. The most common stereochemistries for both ions are
the linear two-coordinate and the tetrahedral four-coordinate geometries with some distortions of the environment, particularly in the
presence of chelating type ligands, attributable to the spherical dl0
configuration. Under suitable conditions, these simple coordination
compounds with the presence of two rodlike or four sticky sites can
be used as tectons to form the self-assembly of predictable supramolecular aggregates. On the other hand, account should be taken of
careful selection of the multifunctional organic ligands and controlling the assembly and orientation of the individual building
blocks-in other words, a combination of coordination bonds and non-
COPPERU) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
175
covalent intermolecular interactions in a mutually compatible manner. Much study has centered on the use of multifunctional ligands.
One of the simplest and most widely employed methods is to use a
bifunctional rodlike diatomic CN- anion and bidentate N,N'-donor
linking groups such as pyridine-, pyrazine-, and imidazol-based ligands with a preference for binding metals at each end in a linear
fashion, together with a metal center with a preference for a polyhedral arrangement of ligands. The recent efforts in this field have even
been extended to the novel oligopyridines, calixarenes, crown ethers,
cryptands, and tetrathiafulvalene derivatives. Such an approach
opens the possibility of rational synthesis of functional solid-state supramolecular metal complexes containing copper(1) and silver(1) with
multidimensional helical, honeycomb channel, interwoven diamondoid, and graphite frameworks among other novel structures,
Scheme 1.
rotaxane
cale na ne
honeycomb
linear chain
graphite
SCHEME
1. The self-assembly pathway for formation of copper(1) and silver(1) coordination supramolecules.
176
Polynuclear Cu(1) and Ag(1) compounds belong to very diverse stoichiometries. In general, they can be prepared either directly by a reaction between reactants or by electrochemistry. It is the aim of the
present review to give a cursory examination of recent developments
in construction of the supramolecular frameworks, which combine the
covalent bond-forming capability of the metal ion with the ligand surface capable of forming noncovalent interactions. Because of the great
volume of literature on compounds, it has been necessary to be highly
selective in the choice of material for inclusion. Emphasis throughout
is on compounds isolated in the solid state. In a review some 130
pages long one cannot expect to do justice to the depth and extent of
investigation into these systems. Some of the topics were reviewed
previously (17, 181, but the articles soon become outdated because of
the rapid growth of the fields.
II. Helical Frameworks
In this section helical complexes of copper(1) and silver(1) are reviewed as examples of self-assembly in metallosupramolecular systems. One of the most intriguing dissymmetric shapes in the natural
system is the helix. Organic and inorganic polymers existing in helical
structures are of special interest because of their structural similarities to nucleic acids. They are also intrinsically interesting for their
potential applications in the fields of supramolecular chemistry,
asymmetric catalysis, and nonlinear optical materials. Since the early
pioneering work of Lehn on double-helical copper(1) complexes with
oligopyridines ( I , 19), there has been an enormous worldwide interest in helical complexes over the past decade. Most of the work has
concentrated on the use of oligopyridines and oligophenanthrolines
to control the assembly of helical supramolecular systems. An enormous number of infinite single-helical complexes and double- and
triple-stranded helicates have appeared in the literature (see Table I).
Many comprehensive reviews have appeared in recent years, usually
covering helical structures in general without special reference to
copper(1) or silver(1) complexes (8, 20-22). In this section, the complexes with single-, double-, and triple-stranded structures are dealt
with separately. The ligands involved in this section are listed in
Fig. 1.
COPPER(1, AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
177
TABLE I
COPPER(1)AND SILVER(1)COMPLEXES WITH HELICAL
FIWEWORKS
Complex
Stereochemistry
of M
Structure
Ref.
tetrahedral
tetrahedral
linear
linear
3-coordinate
tetrahedral
linear, 3-coordinate
trigonal pyramidal
tetrahedral
infinite single strand

infinite single strang
tetrahelix chain
infinite single helical chain
3-D single helix
dinuclear double helicate
2.3
24
25
25
26
27
28
29
30, 31
tetrahedral
tetrahedral
tetrahedral
tetrahedral
tetrahedral
tetrahedral
tetrahedral
tetrahedral
octahedral,
tetrahedral
octahedral,
tetrahedral
trigonal
bipyramidal
tetrahedral
5-coordinate
linear, tetrahedral
linear, tetrahedral
tetrahedral
tetrahedral
linear
linear
2-coordinate

trefoil knot
32
33
34
35
36
37
38, 39
40
41
heterodinuclear double
helicate
42

trinuclear double helicate
trinuclear double helicate
infinite double helicate
infinite double helicate
triple helicate
43
44
45
46
47
48
50
51
61
43
A. INFINITE
SINGLE-HELICAL
COMPLEXES
The number of one-dimensional infinite single-helical copper( I) and
silver(1) complexes is rather limited. They are characterized by a single strand of ligands twisting around the helical axis defined by metal
ions [Fig. 2(a)l. It may be left-handed or right-handed. When ligand
L, reacted with two equivalents of [Cu(MeCN),IBF.,in a mixture of
178
PhzPHzC
CHzPPhz
bMe
L4
L5
L6
LlO
Ll1
L12
FIG.1. List of organic ligands in copper(1) and silver(1) helical complexes.
COPPER(1) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
L20
L19
L22
179
LP1
L23
FIG.1. (continued)
CHC& and MeCN, bright orange crystals of [Cu(LI)lBF4.CHClB

were
obtained ( 2 3 ) . In the cation each Cu atom is four-coordinated, with
two pyridylethylimine bidentate units from two different ligand molecules in a distorted tetrahedral geometry, and each L, exhibits a bis
(bidentate) fashion, bridging two metal centers forming a linear helical strain (Fig. 3). The recently synthesized ligand bisl'342-pyridyl)
pyrazol-1-yllphosphinate (L,) contains two bidentate compartments
linked by a flexible phosphinate bridge ( 2 4 ) .Its flexibility means that
it can adapt to the specific preferences of the metal ions. Its complex
180
'j
FIG.2. Schematic views of copper and silver helical arrangements: (a) infinite single
strand; (b) 3-D single helical copper(1) complex with mixed ligands; (c) dinuclear double
helicate; (d) trinuclear double helicate; (el infinite double helicate; (f) infinite chiral
double helicate of copper(I1) with arginine and rn-phthalate; (g) triple helicate of
silver(1).
with Ag', [Ag(L2)I* 2H20 contains infinite helical chains of the cation
in which each ligand donates one N,N'-bidentate arm to each of two
metals and each metal ion is four coordinated by two arms from
different ligand!. There are interligand aromatic stacking interactions (3.2-3.6 A) both within each helical strand and between
strands. The strands are further held together via a hydrogen-bonding network involving the phosphinate groups and lattice water molecules (Fig. 4).
Tetrahedral coordination of the metal ion is not an essential
requirement for formation of this type of helical structure. Twocoordinate silver(1) ion plays an important role in the formation of a
helical framework, that is, the stereoconformation of the ligand itself
is maintained on coordination to Ag(1) and arranged in a onedimensional helical chain. Reaction of the bidentate optically active
ligands (4R,5R)-and (4~,5S)-4,5-bis(2-(2-pyridyl)ethyl)-1,3-dioxolane
(R,R-L3 or S,S-L3)with silver(1) trifluoromethanesulfonate in metha-
COPPER(1) AND SILVER(I1 SUPRAMOLECULAR METAL COMPLEXES
181
FIG. 3. Structure of [Cu(LI)IBF4.CHC1,. (From Fig. 1 in Gelling, 0. J.; van Bolhuis,

F.; Feringa, B. L. J . Chern. SOC.,Chern. Comrnun. 1991, 917.)
no1 gave two isomeric complexes, [Ag(R,R-L,)IO,SCF, and [Ag(S,SL,)]OsSCF3(25).Each complex cation has an extended structure consisting of Ag' with a slightly distorted linear geometry and the bridging ligands. Projection array along the screw axis for each isomer
exhibits the left-handed helicity for the former (Fig. 5 ) , and the righthanded for the latter. Helical complexes with nonpolypyridine ligands
can give diverse structures and functions. The three-coordinate silver(1) complex with helical structure was observed in [Ag(LJ(NO,)I,
where L4 is the chiral ligand (R,R)-(4R,5R)-truns-4,5-bis[(diphenylphosphino)methyll-2,2-dimethyl-1,3-dioxalanel
(26). The structure
contains an infinite right-handed helical strand consisting of silver
atoms, each coordinated by two phosphorus atoms of two adjacent L4
ligands and an oxygen atom of the nitrate ion. The unusual iodine
tetrahelix was observed in [(C,H,),P],'[Cu,,I,I, in which tetraphenylphosphonium ions are accompanied by a helical chain of face-sharing
182
FIG.4. Structure of [Ag(L2)1.2Hz0.(From Fig. 3 in Psillakis, E.; Jeffery, J. C.;

McCleverty, J. A.; Ward, M. D. J. Chern. SOC.,Dalton Trans. 1997, 1645.)
tetrahedra formed by iodocuprate(1) anions (27). Another infinite single-helical structure was reported in a polymeric silver(1) cryptate
(28).The polymeric cation [Ag{Ag2(L5)}l
consists of dinuclear [Ag2(L,)I
units in which each of the two silver atoms located inside the cavity of
the ligand is coordinated to two imino nitrogens and one bridgehead
nitrogen in a distorted trigonal environment. The third Ag ion links
two dinuclear units by coordinating to one of the imino nitrogens
of two different L6 ligands. Therefore, the cation can be regarded as
the conjugate bis(iminobenzene1 moieties linked together by linear
N-Ag-N bridges in an alternate in- and out-conformations relative
to the benzene rings, resulting in the formation of an infinite singlehelical chain (Fig. 6). The structure determination reveals that L6 is
a unique polydentate ligand with all eight N atoms involved in coordination to the metal. To satisfy the steric requirement of the assembling process, both bond distances and bond angles of the ligand show
appreciable differences compared with those in the corresponding
cryptate compounds. At the same time, the coordination environment
around the silver ion is significantly distorted from ideal linear or
trigonal geometry as a stereochemical compromise for formation of
the helix.
COPPERII)AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
183
FIG. 5 . Structure of [Ag(R&-LdlO,SCF,. (From Fig. 1 in Suzuki, T.; Kotsuki, H.;

Isobe, K.; Moriya, N.; Nakagawa, Y.;Ochi, M. Znorg. Chem. 1995,34, 530.)
Recently an unusual type of three-dimensional single-helical framework, [Fig. 2(b)l, has been reported in the copper(1) complex [Cu(Lg)
(Me,CO), 5JBF4,where L, is a derivative of pyrazine, 2-pyrazinecarboxamide (29). In the extended structure of the cation each metal
atom is linked to two L6 ligands, forming a distorted trigonal planar
structure, and axially bridged to another metal center by one acetone
molecule with a Cu-0 distance of 2.423(9) giving rise to an infinite
helicaI structure (Fig. 7). The most remarkable feature of the complex
is that the infinite helices generate a three-dimensionally extended
hexagonal array of Cu atoms with a large cavity in which the counteranions are placed. In this structure the primary coordination involves L6 tridentately bridging two copper atoms, and the helical
184
FIG.6 . Structure of the polymeric cation [Ag{Ag2(L6)}l.(From Fig. 2 in Yu, S.-Y.;

Luo, Q.-H.; Wu, B.; Huang, X.-Y.; Sheng, T.-L.; Wu, X.-T.; Wu, D.-X. Polyhedron 1997,
16, 453.)
structure is further stabilized by secondary interactions between Cu

atoms and the bridging acetone as the continuous strand. It demonstrates that assembled helical structures of copper(1) complexes with
tridentate oligopyrazine ligands can be achieved by introducing a
suitable spacer groups such as acetone between the metal-binding
sites to match the metal ion stereochemical preference.
COPPERU) AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
185
FIG. 7. The hexagonal framework in [Cu(L6)(Me2CO)os]BF,

( a ) is generated by the
helical cyclic array (b), where only copper atoms are shown. (c) Space-filling model of
the complex, where BF; are omitted. (From Fig. 4 in Munakata, M.; Wu, L. P.; KurodaSowa, T.; Maekawa, M.; Moriwaki, K.; Kitagawa, S. Znorg. Chern. 1997,36,5416.)
B. DOUBLE-HELICAL
COMPLEXES
Double helicates constitute the most abundant species among the
copper(1) and silver(1) helical complexes. In the presence of the tetrahedral Cu(1) and AgU) ions, two oligopyridines can wrap around each
186
other in a double-helical fashion with the metal ions holding them

together [Fig. 2(c)]. Rational design of double-helical structures containing up to five metal ions and those with infinite frameworks has
been developed by different groups (8,20-22).
Among double-helical complexes, dinuclear helicates with two ligand molecular threads twisting around two metal ions predominate.
Based on the description of these complexes made by Constable (81,
they can be defined as [4 + 41, [5 + 51, and [6 + 41 dinuclear helicates,
the two numbers in the square bracket indicating the number of nitrogens to which the two metal ions are bounded. The most common
are the dinuclear [4 + 41 helicates with the tetradentate quaterpyridine derivatives. They include complexes of quaterpyridine in
(30, 31 ), the sym[CU&)ZI(PF~Z,
[AgdL7)~I(PF6),,
and [Agz(L7)21(BF4)z
metrically methylthio- and phenyl-substituted quaterpyridine deriva(32)
~ 7and
} 2 1[CU&)ZI(PF~)Z
~PF6~z
(331. To
tives in ~ ~ u 2 { 4 , 4 - ~ ~ e ~ ~ 2
probe the steric control in the self-assembly of directional helicates,
Constable has carried out systematic studies on asymmetrically alkyl( P F).~It) ~
substituted quaterpyridine derivatives such as [ C U ~ ( L ~ ) ~ I (34
is found that substituents do not control the self-assembly of quaterpyridines but are responsible for the detailed structure. All these dinuclear helicates with quaterpyridines, L7-LB,have general structure
(Fig. S), though there are some
features as exemplified by [CU~(L,)~I+
slight differences in structural parameters as result of the introduction of the substituents and changes of the metal cations and counteranions. The cation is dinuclear, with each metal ion in a distorted
four-coordinate environment, and the two quaterpyridine strands are
FIG.8. Structure of dinuclear double helicate with quaterpyridine [Cu2(L&]+.(From

Fig. 6 in Constable, E. C.; Elder, S . M.; Hannon, M. J.; Martin, A,; Raithby, P. R.;
Tocher, D. A. J . Chem. SOC.,Dalton Trans. 1996,2423.)
COPPERU) AND SILVER(I) SUPRAMOLECULAR METAL COMPLEXES
187
wrapped around the binuclear core in a double-helical array. It is also

observed that n-stacking interactions between aromatic rings play a n
important role in stabilizing the double-helical geometry. The fact
that the oligopyridine disfavors binding to only a single strand in the
system suggests that it contains sufficient molecular information for
the recognition of the metal-binding domain in self-assembling processes. Conductivity measurements and UV-visible spectra show that
the dinucleating bisbidentate ligand Ll0also forms similar double helices with copper(1) in [ C U ~ ( L ~ O ) ~ I (HC2~0O
~)~
(35).
The copper(1) complex with CHzCH2-bridgedbipyridine ligand LI1is
( C ~again
O ~ ) the
~ , two equivalent ligands
formulated as [ C U ~ ( L ~ ~ ) ~ I and
are arranged in a helical manner by complexation with two copper
ions ( 3 6 ) .Reaction of L12with Ag(1) resulted in the formation of another [4 + 41 dinuclear double-helical complex [Ag2(L12)z1(PFs)z
(37).
In this complex, the distorted four-coordinate environment of each silver is maintained, but with weak interactions with a fifth ligand. The
helical structure is stabilized by n-stacking interactions between parallel pyridyl rings. Another example worthy of comment is
[ C U ~ ( L ~ ~ ) ~ ] utilized
( B F ~ ) ~as, a precursor of the trefoil knot system in
which the two coordinating molecular threads, L13 , are well interlaced
on two copper(1) centers, forming a dinuclear double-stranded helical
complex (38).The synthesis of the first trefoil knot was reported by
the same group five years earlier, with the chelate unit, LI3,bridged
by (CH2)4 rather than a phenylene group (39). Modification of the
linker between the two 1,lO-phenanthroline chelates certainly increased the yield, but in both cases, the organic precursors are found
nicely wound and well adapted to the formation of a knot by connecting the appropriate ends of the strands (Fig. 9).
The metal-ion stereochemical preference and the compatible ligands are the first elements of helical structural control. Interestingly, L14with pyridine bridging two imidazole apartments formed a
dinuclear double-helical complex, IC U ~ ( L ~ ~ ) ~ I (similar
C ~ O ~to) ~those
,
described earlier (40).By contrast, the ligand obtained through the
replacement of the pyridine in LI4 by a benzene group gave a dinuclear nonhelical complex with Cu(I), suggesting the fundamental
role played by the spacer for the helical twist. Fully understanding
the elements that guide the preferential combination of ligands and
metal ions can surely allow the controlled assembly of multicomponent systems.
Copper(I) and silver(1) complexes derived from odd-numbered oligopyridines present a different situation. Quinquepyridine (LIE)in dinuclear helicates usually involves a rotation about a n interannual
188
FIG. 9. The dinuclear double-stranded helical complex [CU~(L,,)~I(BF,)~

is utilized as
the precursor of the trefoil knot system. (From Scheme 1 in Dietrich-Buchecher, C. D.;
Sauvage, J.-P.; Cian, A. D.; Fischer, J. J. Chem. SOC.,Chem. Commun. 1994, 2231.)
C-C bond to separate the ligand into a tridentate 2,2 :6,2-terpyridine and a 2,2-bidentate pyridine part. A double-helical array of two
L15 ligands presents a total of ten nitrogen donor atoms, which can be
arranged to create one six-coordinate [tpy + tpyl and one fourcoordinate [bpy + bpyl, or two five-coordinate (tpy + bpy) dinuclear
units (8, 21). So far copper(1) dinuclear [5 + 51 helicates are not
known because pentacoordination is not the favorable stereochemistry for Cu(1);instead, it forms a [6 + 41 dinuclear helicate with Cu(I1).
Reaction of quinquepyridine and [Cu(MeCNWPFsin air gave a brown
mixed-valence complex, [Cuz(L16)zI(PF~)3,
in which the divalent copper
is in an octahedral [tpy + tpy] and the monovalent copper is in a
tetrahedral [bpy + bpyl environment (411. The heterodinuclear double-helical complex containing Ag(1) and COW)has been reported. The
reaction of the two mononuclear cations, [ C O ( L , ~ ) ( M ~ ~ Hand
),I~+
[Ag(L16)I+,
in a 1: 1 ratio gave
(42). The helical
structure is reminiscent of that observed for the mixed-valence copper
complex [Cuz(L,6)zI(PF6)3,
in which the Cu(I1) is replaced by Co(I1).
In contrast to these observations, a dinuclear [5 + 51 helical structure has been observed in the silver(1) complex [Ag2~MezLl~~zl~C104>z
twists around
in which the ligand, 6,6-dimethyl-quinquepyridine,
the dinuclear core, with two nitrogens coordinated to one Ag and the
other three nitrogens to the second Ag, adopting the usual [2 + 31
COPPERW AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
189
mode ( 4 3 ) .As expected, both silver(1) ions are five-coordinated with a

flattened and distorted trigonal bipyramidal geometry. Interestingly,
another quinquepyridine derivative with two further phenyl groups
a t the 4 and 4positions forms a [4 + 41 helicate in [Ag{(MePh)zL15}zl
(C10&, in which (MePh)zL15actually acts as a tetradentate ligand
and the central pyridyl ring plays the role of a rigid spacer rather
than a donor ( 4 3 ) . Because spacers linking the oligopyridine units
have a notable impact on the versatility of the compounds in accommodating metal ions with different sizes and different geometries, one
of the approaches for rational design of helical structures is to use
aliphatic ethylene or aromatic 1,3-phenylene groups as spacers (3 6 ).
For example, Lls with 173-phenylenea s a spacer linking two terpyridine units exhibits essentially a pentadentate coordination mode in
[Ag2(L16)21(PF,)z
( 4 4 ) .The crystal structure reveals a [5 + 51 doublehelical bimetallic silver complex in which each Ag(1) adopts a n irregular five coordination.
Terpyridine is the simplest oligopyridine capable of forming a double-stranded helicate. Copper(1) complexes with terpyridine derivatives have been reported. In [Cuz(L17)z1(PF6)2
( 4 5 ) and [CUZ&~ZI(PF&
(461, the ligands have essentially distributed themselves to present
bidentate domains to one metal center and a single pyridine donor to
the other, to give a [4 + 21 double helicate. The two copper atoms
involve different coordination environments, one with a distorted tetrahedral geometry and the second in a approximately linear twocoordinate environment (Fig. 10).
Dinuclear double helicates are too numerous to be fully discussed
here. In contrast, tri, tetra, penta-, and polynuclear double-helical
complexes of copper(1) and silver(1) appeared only in a limited num-
FIG.10. Schematic representation of the two copper atoms involving different coordination environments in a [4 + 21 double helicate.
190
ber. A ligand with three bipyridiyl binding sites such as Llo can bind
up to three cations, giving a trinuclear helical complex. Trinuclear
double helicates [Cu3(Llo)zl(PF,)3
(47)and [Ag3(Lzo)21(CF3S03)3
(48)
were prepared by treatment of sexipyridine and the relevant metal
salts. Structure analysis and molecular modeling studies indicate
that the two ligand groups fit three metal ions inside the doublehelical array by twisting around the helical axis [Fig. 2(d)]. Each
metal center is coordinated to a pair of adjacent pyridine residues
from each ligand in a distorted tetrahedral environment. The functionalized oligopyridines of Lz0with n = 4 and 5 formed tetra- and
pentanuclear double helicates, respectively, with Ag(1) (48) and Cu(1)
(49).Their overall structural features are analogous to those of the
trinuclear double helicates, but their total length is estimated to be
as long as around 22 A and 27 A, respectively. These self-organized
nanostructures promise potential applications in the field of functional nanoscale species and molecular devices.
Discrete infinite double helices are quite rare in inorganic and coordination chemistry [Fig. 2(e)l. Ciani has reported that reaction of Lzl
with Ag' salts in the ratio 2 : 1gave completely different and noteworthy products, one of which is the infinite double-helical coordination
polymer [Ag(Lz1)ICF3SO3
(50). The structure shows the balanced
packing of left-handed and right-handed double helices of cationic
-Ag-L-Ag-Lchains (Fig. 11). The period of the helices is
and the two strands are bridged by weak Ag-Ag aurophlic
21.1
( C P ~ - ~interactions.
'~)
The copper(1) complex [ C U ~ ( L ~ ~ ) ~ Ihas
(C~O~)~
been prepared and structurally characterized (51). The basic dinuclear unit has a structure similar to that observed in [Cuz~L14~zlz+
(40). The ligands bind in a bis-monodentate coordination mode and
form the strands of the helix that twist around the helical axis on
A,
FIG.11. Space-filling model of the infinite double helix of [Ag(L,,)ICF,SO,. (From

Fig. 2 in Carlucci, L.; Ciani, G.; W. v. Gudenberg, D.; Proserpio, D. M. Inorg. Chern.
1997,36, 3812.)
COPPER(1) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
19 1
which the copper ions lie. Each Cu(1) center is nominally linear coordinated by two imidazole units, and the interactions with the two
bridging pyridines are weak. The double-helical subunits stack together in parallel columns of constant but different helicity, leading
to infinite double-helical columns.
Chiral helical complexes are of interest in connection with the manifestation of functions such as optical activity, molecular recognition,
and enantioselective catalysis (52-56). With spirobisindanol and dimethylbiphenic acid as chiral templates, Siege1 has successfully obtained bipyridine based double helicates containing up to three copper(1) ions by enanthoselective synthesis (57). The spectroscopic
results confirmed the complex as a single enantiomer of the head-tohead isomer whose stereochemistry is controlled by the single stereoelement of the template throughout a span of 20 A, demonstrating
the ability to transmit stereochemical information over nanometer
distances in the compounds. Chiral double-helical structures of copper(I1)-L- and -D-arginine complexes with aromatic dicarboxylates
have been reported, although the corresponding copper(1) species are
not known (58). These crystals reveal that the complex with L-Arg
forms a right-handed helix, whereas the complex involving D - h g
gives a left-handed helix [Fig. 2(r?l. This indicates that the handedness of the double helices in the solid state is governed by ligand
chirality. These and other synthetic routes open new ways for development of enantioselective reagents and may contribute to the understanding of spontaneous aggregation of conjugated helical molecules
in biological systems (59, 60).
C. TRIPLE-HELICAL
COMPLEXES
There is only one example of a triple helix containing Ag(I), reported recently by Williams and his co-workers (61). The colorless
crystals [Ag3(L23)51(BF4)3
were obtained by diffusion of benzene into a
solution of the compound in acetonitrile. The structure consists of an
equilateral triangle of silver ions with the ligands bridging the sides
of the triangle [Fig. 2(g)l. The silver ion is not in a strictly linear
environment, with an N-Ag-N angle of 153.3(6). Each ligand binds
to one metal from below the plane of the silver atoms and to a second
metal from above the plane. The structure may be considered as a
triple helix in which the ligands wrap around the threefold axis and
are held in place by coordination to the Ag, triangle.
192
Ill. S - - 4Contact-Assembled Frameworks
Some sulfur-rich dithiolates and the tetrathiafulvalene (TTFI-based

species involved in the following discussion are listed in Fig. 12. Following the observation that TTF (L2&forms a stable radical when
treated with chlorine (62) and the discovery that a partial charge
transfer between separately stacked donor and acceptor molecules
leads to the first organic metal, TTF * TCNQ (tetracyanoquinodimethane) (63), the use of sulfur-containing molecules as precursors for
synthesis of conductive or superconductive materials has been of unabated interest for over two decades. Because the large sulfur atomic
orbitals are capable of promoting effective intermolecular overlap in
the organic metals, partial oxidation of TTF and its substituted derivatives, when coupled with suitable electron acceptors, can give rise to
low-dimensional arrays of these donors through sulfur-sulfur molecular interactions, providing an effective pathway for electronic conduction. A tremendous number of studies have been carried out on the
L27
M e 0 2 C f ~ s
Me02C
LZ8
s-
'Xs F3c1sNcxs
S
F&
L29
S'
NC
L30
L31
.=(Ise-
Se'
0
L32
L33
L36
L34
L35
lTC,-lTF
FIG. 12. Sulfur-rich dithiolates and TTF derivatives.
L38
COPPER(I) AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
193
solid-state conductive properties of the TTF . TCNQ-like organic and

inorganic metals by modifications of TTF (64-66).
By contrast, the study of TTF-based species as building blocks in
supramolecular chemistry has attracted attention only recently (67).
With a preference for a soft donor such as the sulfur atom, both silver(1) and copper(1) readily form metal complexes with the sulfur-rich
TTF-based organic molecule, with a high possibility of generating
one- two- and three-dimensional supramolecular and macromolecular
system by the covalent bonds as well as by S----Scontacts. In general,
the S----Scontacts considered to be effective in sulfur-rich conductive
compoundso refer to those having interplanar S----S separations a t
3.35-3.75 A, close to the sum of two sulfur-atom van der Waals radii
(3.60A). This section focuses on construction of the supramolecular
frameworks through S----Scontacts based on dmit, BEDT-TTF, TTC,TTF, C5H4S,and the related species.
A. dmit
AND THE
RELATED
LIGANDS
The sulfur-rich 1,3-dithiole-2-thione-4,5-dithiolate

(dmit) (LZ5)has
received considerable attention. Using square planar coordinating
metal ions such as Ni, Pd, and Pt, many electrically conducting materials derived from planar metal complexes [M(dmit)J- have been
synthesized, and some of them have been found to undergo a superconducting transition at low temperature by applying high pressure.
The nonplanar bulky metal complexes [M(drnit),l"- (M = Fe, Re, Mo,
or W) have received equal attention. Numerous papers and reviews
have focused on the study of the role of the intermolecular and intramolecular interactions in determining the electrical properties in the
system (68-83). As a n example, Fig. 13 shows a uniform segregated
stacking arrangement of the constituent Ni(dmitI2 units with the
open-shell 7r-acceptor TTF and the closed-shell cation (CH3I4Nin the
superconductors (TTF)[Ni(drnit)& (84) and [(CH,),N][Ni(dmit)zlz(851,
respectively, where close intermolecular S----Scontacts occur between
units in adjacent stacks leading to a n infinite network.
Furthermore, extensive studies have also been carried out by using
other ligands resembling dmit, such as L26-L33 (86-92). Along the
same line, the selenium analogs (L34and L35)are also of much interest, because even more effective molecular interactions are expected
to occur owing to Se having more extended orbitals than sulfur
(93,941.
Although a number of Cu(I1) complexes with the dmit-like ligands

have been reported in which the metal ion involves a significant dis-
194
(b)
FIG. 13. Molecular packing of superconductors (TTF)[Ni(Lp)& (a) and [(CHAN]
[Ni(La&lz(b). (From Figs. 5 and 6 in Bousseau, M.; Valade, L.; Legros, J.-P.; Cassoux,
P.; Garbauskas, M.; Interrante, L. V. J. Am. Chem. SOC.1986, 108, 1908, and Figs.
1 and 4 in Kim, H.; Kobayashi, A.; Sasaki, Y.; Kato, R.; Kobayashi, H. Chem. Lett.
1987, 1799.)
COPPERW AND SILVERU) SUPRAMOLECULARMETAL COMPLEXES
195
tortion from planar to tetrahedral (73, 83, 89, 931, few nonplanar
Cu(1) or Ag(1) metal complexes of dmit are known. The most interesting of these is [mpy]z[Cu4(dmit)31,
prepared, like its Cu(I1) analog, by
a route of direct reaction of coppert11 salt and Nazdmit in the presence
of excess of [mpylI (95). The structure contains dimerized anion units,
each consisting of a tetranuclear Cuds6cluster. Each Cu(1) ion involves a distorted tetrahedral coordination comprised of four dmitsulfur atoms. The dimeic units further interact with each other
through close sulfur-sulfur contacts to form a two-dimensional molecular interaction sheet. The cations are located between the anion
sheets.
B. BEDT-TTF
BEDT-TTF (L36) is another electron donor molecule and in combination with inorganic anions has provided several air-stable superconductors at low temperature (96-99). In fact, both dmit and L36 superconductors have the characters of (i) good orbital overlap between
extended r-electron systems (S----Scontacts) and (ii) partial filling of
the conduction band through either partial oxidation or partial reduction. Here again the radical cation salts of square planar transition
metal complexes dominate, and they feature S----S interactions between planar L36'+ molecules leading to segregated columns of donors.
On the other hand, the reported copper(1) and silver(1) coordination
complexes of L36 are scarce. They include (LSt)Cu2Br3,prepared by a
redox reaction between L36 and [CuBr4I2-(100).The most unique feature of the complex is that the L36 molecules are not stacked in a
column but are coordinated to the tetrahedral copper(1) centers in a
-Cu-Br-Cu-(p-Br)z-Cuchain (Fig. 14). Due to lack of close
S----Scontacts between radical cations, the complex shows low conductivity .
FIG.14. Structure of (L,+)[Cu2Br31.
196
C. TTC,-TTF SYSTEMS
Several copper(1) and silver(1) complexes with tetrakis(alky1thio)
tetrathiafulvalene (TTC,-TTF) (L3J have appeared in the literature,
and various S----Scontact-assembled frameworks have been observed.
L37 contains a central TTF (L2J r-system, which keeps the electrondonor property. As single-component organic semiconductors with low
electrical conductivity, their physical properties have been widely investigated. They can be easily oxidized to the stable radical cation;
accordingly, their chemistry is dominated by electrotransfer processes. They react with organic electron acceptors such as TCNQ and
HCBD (hexacyanobutadiene), and inorganic oxidants to form partially or completely oxidized materials containing L37+, where p
ranges from about 0.6 to 2.0 (101-104).In most of these chargetransfer compounds the donors and acceptors are stacked in segregated columns.
Construction of two- and three-dimensional coordination polymers
using the diversity of copper halide frameworks is an area of considerable importance. Besides its electron-donor character, Lg7possesses
alkylthio groups, which have coordination ability to metal ions
through the sulfur atoms. This gives Lg7species a unique ability in
linking metal ions to form coordination polymer structures. Six Lg7
complexes with copper(1) halides, [(CUX)~(L~,)]
(where n = 1 or 2 and
X = C1-, Br-, or I-), have been structurally characterized (105-107).
The syntheses of these compounds are straightforward, usually carried out by direct reaction of copper(1)halide and Lg7in a nonaqueous
solution. The structural determination of these complexes reveals
three different copper halide frames, namely rhomb, helix, and zigzag
chain (Table 11).Although the metal ions in all cases involve a tetrahedral coordination comprising two halogen atoms and two sulfur
atoms of the Lg7moiety, with a certain degree of distortion, which in
turn acts as a bidbidentatel linker between the two metal centers
through alkyl thioether groups, the resulting compounds often exhibit
for exnovel structures dramatically changed by the halogen. In LgTa,
ample (1051,the chloride complex consists of novel two-dimensional
sheets of L37a molecules arranged between zigzag frames of {CuCl}, .
Within the two-dimensional sheet no significant S----Scontacts are observed because the long Cu----Cuseparation of 5.78 A on the same
side of the CuCl zigzag frame gives a long interplanar spacing (4.90
A) of Lg7,,. However, between the parallel sheets there are close intergiving a three-dimensional netmolecular S----Scontacts of 3.53
work (Fig. 15). By contrast, a novel framework of {CuBr},, helixes
A,
COPPER(I) AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
197
TABLE I1
BONDANGLESAND CuX FRAMEWOKKS
I N TTC,,-TTF(Ls,) COMPLEXES
OF COPPEKU)
HALIDE
01
113.4
116.0"
109.1"
112.6"
113.1"
111.8
113.6
106.5
88.8
89.9
91.1"
116.1
116.1
116.8
112.5
111.8
113.2
CuX framework
Ref.
rhomb
rhomb
rhomb
rhomb
105
107
106
87.1
86.9
helix
helix
105
106
90.9
zigzag chain
105
108
Average value.
< =c,X
-c u
rhomb
infinite
helix
zigzag
chain
which are further linked to each other by coordination of L37ato the

metal was observed in the bromide complex. The L37amolecules are
stacked together, with the dihedral angle between the neighboring
molecules being essentially zero on the stacking axis. The unique feature of the structure is that the {CuBr}, helixes orientate alternately
in the reverse direction and the L37amolecular stacks are mutually
perpendicular to each other along the c-axis. The S----S contacts of
3.68 A between the stacked L37amolecules lead to a two-dimensional
framework. The iodide complex contains an infinite chain of the bent
LS7, molecules perpendicularly bridged on the methyl thio2ther
groups by rhomboid CuzIzmoieties. The S----Sdistances of 3.75 A between the chains indicate weak interaction present, associating the
adjacent chains of -L-Cfi-Ltogether, giving a two-dimensional
network. Similar results have been observed in the corresponding LS7,,
complexes (106, 1071, in which the chloride complex involves helical
198
3.53.i
GI
12Dsheet
I
uc --uc
II
CI
I
C-ICU-
FIG.15. Schematic views of different frames of CuX in [(CuX),(L,,,)I complexes.
{CuCl}, frames whereas the bromide and the iodide both contain
CuzXzrhombs. Bond angles of copper(1) halide complexes with LSmare
summarized in Table 11. It can be seen that the main factor determining the framework of the CuX unit seems to be connected with the
Cu-X-Cu
angle (a).As a increases smoothly from 66.6 to 1 1 1 3 ,
the framework changes from rhomb to zigzag chain through infinite
helix, whereas other angles do not show such correlation. The L37a+
cation coordination complex (L37,)[CuBrz1zalso falls nicely into this
category (108).
As a comparison, it is instructive to briefly mention the corresponding LS7b+complexes of copper(I1) halide (L3&[Cu2&1 (where X = C1or Br-) prepared by oxidation of L3%with CuC12 or CuBrz (109). Unlike the copper(1) halide complexes with LS7bridging between two
metal atoms, both structures consist of two segregated stacks of L3%+
donors and CU&~- acceptors associated via short S----Scontacts (Fig.
199
FIG. 16. Packing diagram of (L,,)[Cu,CI,I showing S----Sand S----Cl.(From Fig. 1

and Fig. 2 in Wu, L. P.; Gan, X.; Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.;
Suenaga, Y. Mol. Cryst. Liq. Cryst. 1996,285, 75.)
16). The donors face to each other to assume a dimeric structure, with
close intradimeric S - - 4conducts of 3.43-3.52 A. Consistent with the
absence of significant interdimeric S----S interactions, the two compounds exhibit low electric conductivity. The diversity of copper halide frameworks, affected by several elements such as the metal ions,
coordination of the halogen atoms, and the stereofactors of the ligand,
makes it possible to design two- and three-dimensional coordination
polymers.
Copper(1) perchlorate and tetrafluoroborate salts have also been
found to form coordination compounds with the neutral ligands L3,
(n = 2 or 3), but the structures vary little due to nonparticipation of
the counteranions C10,- and BF,- for their weak coordinating ability
(110, 111). In fact, they are 1:1 metal-ligand compounds in which
200
each Cu(1) ion is tetrahedrally coordinated by four sulfur atoms from

the bridging Lg7molecules to form infinite polymeric chains separated
from each other by the anions. The shortest S----Sdistances between
suggesting negligible
the adjacent chains in all cases are over 3.8
S----Sinteractions in the system. Same conclusion can be drawn from
the Ls7, complex [Ag(L3,,)(03SCF3)I(1121, in which each Ag(1) ion is
six-coordinated by four sulfur atoms from the bridging Lg7*and two
oxygen atoms from the bidentate 03SCF3anion, resulting in a structure of two interwoven polymericochains. The closest S----Sdistance
between the two chains is 4.00 A, precluding any significant S----S
contacts present.
Apart from the varieties of interesting structures observed in the
Lg7system, it is worth noting the preparation of these complexes and
the conductivity of the Iz-doped species. As mentioned previously,
these complexes are easily prepared without elaborate synthetic procedures or particular precautions. However, successful preparation of
the single crystals of these complexes largely depends on the reaction
temperature. Again take Lg7,, for example. The preparation was carried out by allowing the mixture solution of CuX salt and L37ato stand
in a water bath at a constant temperature of 70, 55, and 23C for the
chloride, bromide, and iodide, respectively (105).The same reactions
performed at other nondesignated temperatures may lead to a mixture of precipitate and tiny single crystals unsuitable for X-ray analysis. This is also true for the preparation of the corresponding L,,,
complexes (106,107).Iodine oxidation is a viable synthetic route to
highly conducting mixed-valence materials. Although copper(1) and
<
silver(1) complexes with neutral Lg7 ligands are insulators
S . cm-') at room temperature, their iodine-doped products behave as semiconductors, with conductivities increasing in the order
[(CUC~)Z(L~~)I
< [(CuBr),(L,,)I < [ ( C U I ) ~ ( L(Fig.
~ ~ ) I17). A partial oxidation can be assumed to occur in the 12-dopedspecies and the mixedvalence ligand-ligand interactions might cause electrical conduction
pathways.
A,
4,5-Ethylenedithio-l,3-dithiole-2-thione
(C5H,S5)(Lss)is well known
as an electron donor (113, 114).It is also a derivative of dmit and has
a structure similar to half of LSs. The structure determination of its
metal complexes has demonstrated the unique coordination versatility of Lg8(Table 111). It can act as a monodentate, bidentate, or even
201
-1.5
-1.7
-2.1
%?
D
-2.5
-rn
0
lrdoping
.-c>i
.-
12-doping
c)
'0
C
-3.6
-
-3.5
orange crystal
= CIg
Bry
orange crystal
orange crystal
L37a = tetrakis(methy1thio)tetrathiafulvalene
FIG.17. Preparation scheme and conductivity of [(CuX)2(Ls~,)l.
tridentate ligand, binding one and bridging two or three metal ions
through the thiocarbonyl group alone or by both thiocarbonyl and
thioether groups.
Up to now, three silver(1) and one copper(1) complexes of L38have
been synthesized and crystallographically characterized. They display
a variety of one- to three-dimensional frameworks based on coordination bonds as well as intermolecular S----S contacts. Reaction of
AgC104.H20 and L,, in acetonitrile gave red brick crystals of
{[Ag(L38)31C104~CH3CN}2
(115).The compound has a dimeric structure
202
TABLE 111
COORDINATION
MODESOF C5H4S5(LS), THE METALCOMPLEX
STRUCTURE
AND THE
CONDUCTIVITY
OF THE I 2 - D 0 pPRODUCTS
~~
Mode"
Structure of the complex
Conductivity
S . cm-'
Ref.
[Agz(Lss)J(C104)~ I, (11)
111
[Ag(Lsd,NO,I
[ A ~ ( L s ~ ) C F ~ S O ~ IV
I
I, I11
[CuaI4(Lsa)al
dimer
infinite dinuclear chain
2-D sheet
infinite tetranuclear chain
6.9 X
3.0 x 10-5
1.5 x 10-4
2.2 x
115
116
115
117
Complex
_______
Coordination modes of C5H4S5.
n
s
sKs
S
IM
mode I
kS-"
sHs-M
A
sKs
sKs
s)-+s
sKs
JSI
M
M
mode I1
IM
mode 111
JSI
M
M
mode IV
with two nonequivalent intercoordinated quasi-trigonal [Ag(L3&1+

units, which are further assembled by very close S--------S
contacts of
3.284(4)-3.569(2) to form a one-dimensional chain structure (Fig.
18).Replacement of the anion by triflate in preparation gave a poly-
FIG.18. Schematic representation of S----Scontacts in {[Ag(LS)31C104.

CH,CN}, .
203
nuclear complex with a different stoichiometry, [Ag(L,,)CF,SO,I

( I 15). The structure contains two crystallographically independent
five-coordinate silver(1) ions, very rare stereochemistry for Ag(1). Both
the L3, group and the CF,SO, ion bridge between two Ag atoms, forming a one-dimensional chain along the c-axis. All the ligand groups
are approximately perpendicular to the chain, and between the c h a i p
there are several close S----Scontacts (the shortest one being 3.262 A),
resulting in a two-dimensional polymer structure (Fig. 19). The third
silver(1) complex belonging to this family is [Ag(L,,),NO31 (116).Each
L,, displays a bidentate fashion through both thiocarbonyl sulfur and
thioether sulfur atoms bridging two octahedral silver atoms, giving
a n infinite chain. The strong S----Sinteractions with a distance of
3.41 A between two nearby chains along the b-axis generate a twodimensional network.
The corresponding copper(1) complex [ C U ~ I ~ ( L ,was
~ ) ~prepared
I
by
reaction of L3, and CuI in acetonitrile (117). The structure consists of
stepped Cu414clusters bridged by L3, groups, and the resultant polymeric chains are further assembled by short interchain S----Scontacts
(3.257 A)to form a three-dimensional network (Fig. 20).
Unusually short S----S contacts have been noticed among theo L38
complexes; for example, 3.23 A in (TTF)[Pt(L3&I3 and 3.26 A in
(TTF)[Pd(L3d212,[Ag(L3dCF3S031, and [ C U ~ I ~ ( L(115).
~ ~ ) ~ The
I
extremely short S----Scontacts are attributable to metal-metal bonding
in the first two cases, and to the coordination bond linkage as well as
the effect of molecular packing in the rest. The copper(1) and silver(1)
complexes of L3, were partially oxidized by iodine doping, and the
electrical resistivity of the compressed pellet of the compounds as
measured by the conventional two-probe technique. Like those of L3,,
they are nonconducting a t room temperature but display semiconduc-
FIG. 19. Schematic representation of S----Scontact assembled network in IAg(Ls8)

(CF8OdI.
204
54
D: S9...S9' (3.62A)
FIG.20. Structure of rCu4I4(L8&1.
tivity after Iz doping, with conductivities of 10-4-10-5 Sacm- (Table

111).
IV. Hexagonal Frameworks and Graphite-like Structures
The scaffolding and related microporous materials are receiving

widespread interdisciplinary attention, and the borders between
chemistry, physics, and materials science are therefore vague and
COPPER^ AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
205
subjective. Several general reviews and textbooks have provided a

comprehensive overview of the chemistry and structural properties
of these materials (118-120). The literature relating to assembly of
copper(1) and silver(1) complexes into network structures with a
graphite-related lattice has reflected a general interest and advancement in our understanding of the fundamental chemistry and processing characteristics of the system and its potential applications as
molecular sieves, catalysts, and optical materials. This section is concerned with construction of hexagonal and graphite-like lattices regulated by the metal-ion stereochemistry, the ligand, and the counteranions, placing emphasis on the nitrogen-containing systems such as
pyrazine, phenazine, 2,1,3-benzothiadiazoles,and quinoxaline (quin)
(Fig. 21). Metal cyanide will be treated in a cursory manner, leaving
the readers to consult recent literature and reviews.
A. METALCYANIDE-REGULATED
BY THE METAL
ION
Metal cyanide, which has received little attention for years in many
texts concerned with inorganic structures, is now found to possess the
most outstanding feature as a building block for infinite polymeric
frameworks. Some rather elegant metal cyanide structures with extended frameworks have recently been discovered (121-128). The
synthetic strategy is based on the combination of a diatomic bifunctional rodlike CN- ion with a considerable preference for binding
metals at each end in a linear fashion together with modification of
L39
(Pz)
L40
L43
L 41
L45
L44
L46 (phz)
L42
L47
(bW
L48
FIG.21. Some nitrogen-containing ligands in copper(1) and silver(1) complexes having hexagonal frameworks.
206
the frameworks by metal ions favoring different stereochemistries.

The crystal structures of [N(CH,),I[Cu'Zn"(CN)~]and two isomorphous
compounds, Zn(CNI2 and Cd(CNI2,have shown that the tetrahedral
metal ions favor formation of interpenetrated diamondlike frameworks (127). A three-dimensional honeycomb structure containing
hexagonal channels has been observed in complexes consisting of simple cyano-cadmium frameworks in which the metal centers are of two
types: octahedral and distorted five-coordinate or octahedral and tetrahedral (122,128).By replacing CN bridges in multidimensional CNlinked octahedral cadmium complexes with linear NC-Ag-CN or
NC-Ag(CN)-Ag-CN units, Iwamoto has successfully obtained two
cadmium-silver complexes with double and triple interpenetrating
frameworks (123). An infinite three-dimensional framework related
to PtS has been observed in [NMe,l[CuPt(CN),], in which tetrahedral
Cu(1) and the square planar Pt(I1) centers are linked together by CN
rods, generating two mutually perpendicular and equivalent sets of
hexagonal channels of large cross section together with large empty
square channels perpendicular to the hexagonal channels as shown
in Fig. 22 (121).
B. PYRAZINE
SYSTEMS-REGULATED
BY SUBSTITUENTS
Pyrazine (LS9)and substituted pyrazines have long been known to
act as em-didentate ligands to linearly bridge metal ions, generating
oligomeric and polymeric metal complexes with infinite chain and
FIG. 22. Extended framework containing hexagonal and square channels observed
i n [CuPt(CN)J:-. (From Fig. 2 in Gable, R. W.; Hoskins, B. F.; Robson, R. J. Chem. SOC.
Chem. Commun. 1990, 762.)
207
pleated sheet structures, double and triple interpenetrating frameworks, and interwoven honeycomb architecture (123, 129-132 ). The
essential features of some structurally characterized copper(1)and silver(1) complexes with pyrazine and its derivatives including quinoxaline and phenazine are summarized in Table IV. Among them those
containing two-dimensional six-membered rings of metal ions linked
by LBSbridges are rather common and thus have received considerable attention recently. From the point of view of rational synthesis,
it would be possible to obtain polymeric graphite-related cationic lattices in which the three-coordinate metal ions are bonded only to
bridging LBSligands by utilizing noncoordinating anions as the counterions in the synthesis of LSs complexes. However, owing to the effects of substituents, the six-membered rings obtained differ in detail
from complex to complex in terms of distortion of planarity, which can
be grouped under four heads as shown in Fig. 23. It has been suggested that more than two substituents on a pyrazine ring would hinder four coordination of the metal ion, giving infinite zigzag chains or
sheets of the cross-linked chains as observed in disubstituted pyrazine systems (135).Therefore, controlling spatial factors plays a key
role in building infinite polymeric structures, and the modification of
linking ligands by substituents on pyrazine is a practical method for
the synthesis of designed copper(1) polymers.
Slightly distorted hexagonal frameworks defined by six LBS-bridged
copper(1)ions are observed in [CU~(L,,,)~(PF,)~I
(L40= 2,5-dimethylpyrTABLE IV
Cu(1)AND Ac(1) COMPLEXESWITH PYRAZINE
AND ITS DERIVATIVES
Compound
CN"
Frameworks
Ref.
~
[Cuz(L,o)dPFs)zl
four
three
[cuz(L41)31(c1o4)2
[Cu2(~~31(c104~,
three
[Cuz(Lsd31SiFs
three
~ C ~ Z ( L ~ , ~ ~ I ( C I O , ~ ~ ( M four,
~ ~ C five
O)~
[C~,(LS~)~(CH~CN)ZI(PF~)Z
four
four
[CUZ(L,)4 sl(ClO4)z
four
[cu(L,o)zlPF~
[Cu,(4,)31(C104)z
three, two
~cuz~L4~~31~c1o~~1CIo4 four
[Agz(La)(NO3)zI
three
[Agz(LddBFdz
three
2-D hexagonal
2-D hexagonal
2-D hexagonal
2-D hexagonal
2-D hexagonal
2-D hexagonal
2-D hexagonal, square
3-D adamantane
linear
3-D adamantane
2-D hexagonal
2-D hexagonal,
3-D interpenetrating nets
' CN, coordination numbers of the metal ions.
133
134
135
136
29
137
135
133
137
132
138
130
208
FIG.23. Cu(1) and Ag(1) complexes of pyrazine and the derivatives having hexagonal
frameworks: (a) slightly distorted hexagon; (b) distorted hexagon and square; (c) zigzag
2-D sheet; (d) ternary complex. The molecule in the box represents the ligand involved.
azine) (133), [ C U ~ ( L ~ ~ ) ~ ] (L41

) ( C ~=O2,3-dimethylpyrazine)
~)~
(1341,
[ C U ~ ( L ~ ~ ) ; J ((L42
C ~ O=~2,6-dimethylpyrazine)
)~
(135), [Cu2(L3&lSiF6
(136),and [ C U ~ ( L ~ , ) , ] ( M ~ (L43
~ C O=) ~pyrazineamide) (29).Structural
studies have established trigonal environments for the metal. These
complexes show an infinite sheet structure of macrocations with uninteracting or weakly interacting counterions. If the interaction between the metal and the anion is ignored, the Cu(1) ion in these com-
209
plexes is essentially in a three-coordinate trigonal planar geometry

and the pyrazine ligands bridge adjacent trigonal metal centers to
form a graphite-like lattice, Fig. 23(a). Thompson has successfully isolated two Cu(1) complexes from the same reaction mixture of an ethanolic solution of Cu(PF& and L 4 0 , [Cu2(L40)31(PF,)2
and [ C U ( L ~ ~ ) , I P F ~
(133).The structure of the former is made up of parallel puckered
layers of the cation, and within each layer the copper(1) ions are
linked by L40,defining hexagons. PF, anions are located between the
layers and interact very weakly with the copper(1) ions through one
F atom, leaving the Cu(1)ions essentially three coordinate. Except for
the puckering of the layers, the structure is reminiscent of that of the
intercalated graphite. By contrast, each metal ion in the latter case is
bonded to four rather than three L40 molecules, with an angularly
distorted CuN4chromophore. As a result, the structure exhibits a diamondoid framework of copper(1)ions linked by the rodlike L 4 0 ligands,
and the PF; anions are located in cavities within the lattice.
Likewise, the reaction of Cu(C104)2with L41 in an aqueous solution
yields an air-stable complex, [ C U ~ ( L ~ ~ ) ~ (whose
C ~ O ~structure
) ~ ] , is also
found to contain a similar two-dimensional graphite lattice of tricoordinate copper(1) atoms bridged by L1lligands (134).Although the exact course for reduction of copper(I1) complexes with nonchelating ligands is not clear, the most plausible explanation lies in tetrahedral
distortion of the square planar copper(I1) species by crowding effects
of the methyl groups, and consequently, such reduction process is dependent on the substituents, temperature, and solvents (233, 134 ).
On the other hand, copper(1)complexes of L39can also be prepared by
direct reaction of copper(1) salts, [CU(CH,CN)~]X,
and L3, ligands in
nonaqueous solutions, such as [CU,(L~~),]S~F,
(136).Here again the
L3, ligands bridge adjacent trigonal Cu' centers to form hexagonal
frameworks with bridged copper cations 6.936 and 6.685 A apart.
The SiFi- anions are located in channels and are hydrogen bonded to
L3, hydrogen atoms. The most significant feature of the complex is
that the hexagonal frameworks are interwoven as shown in Fig. 24.
The three-coordinate copper(1)ions can also be linked by the disubstituted Ls9on the 2- and 6-positions as observed in [ C U ~ ( L ~ ~ ) ~ , I ( C ~ O ~ ) ,
(135).The graphite-like lattice is constructed by the alternating arrangement of the two crystallographically independent three-coordinate copper atoms. The hexagons generated by the Cufi units are
slightly distorted due to compression, and the mutually confronted Lea
molecules on the sides of the hexagon are sitting out of the plane and
parallel to each other, providing a large cavity. It is surprising that
neither the anion nor the solvent molecule is included in the cavity,
2 10
.......
I................._
i cuqubcu
=
1 1
LStl C(
:i
/CU-dcu,
f c-cu*k
...........................
FIG.24. Schematic view of the interwoven honeycomb grids in [ C U ~ ( L ~ ~ ) (From

~IS~F~.
Fig. 4 in MacGillvray, L. R.; Subramanian, S.; Zaworotko, M. J. J. Chem. SOC.Chem.
Commun. 1994, 1325.)
they rather form acetone-perchlorate layers intercalated between the

copper sheets.
Formation of trigonal copper(1) polymers with dimethyl-substituted
pyrazines is presumably due to the two substituents on the L3,ring,
which hinder the formation of tetrahedral coordination of the metal.
No examples of hexagonal frameworks are found to be built up by
tetrahedral copper(1) ions. This is understandable because formation
of graphite-like sheets would require Cu6motifs on a plane with each
metal center linked to three neighbors. Therefore, trigonal, trigonal
pyramidal, and trigonal bipyramidal stereochemistries are necessary.
The recently characterized copper(1)complex [ C U Z ( L ~ ~ ) ~ ~ ( C ~ ~ ~ ) ~ (
provides the first example of mixed four- and five-coordinate copper(1)
ions composing such a hexagon (29). As illustrated in Fig. 25, the
cation contains two crystallographically independent copper(1) ions;
one involves a distorted trigonal pyramidal CuN30 core composed of
the terminal pyrazine nitrogens of three different L43molecules and
an axial caboxamide oxygen, and the second involves a distorted trigonal bipyramidal structure with two, rather than one, carboxamide
molecule exhiboxygen atoms occupying the apical positions. Each L43
2 11
cu
FIG.25. Molecular structure of [ C U ~ ( L ~ ~ ) ~ I ( C I O ~ )in~ which
( M ~ ~the
C Ofour) ~ an d fivecoordinate Cu(1) ions are bridged by the tridentate La groups (a) giving a hexagonal
channel structure with pores open to accommodate (2104 ions and acetone molecules
(b). (From Figs. 1 and 2 in Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.;
Moriwaki, K.; Kitagawa, S. Znorg. Chem. 1997,36, 5416.)
its a tridentate coordination mode, bridging two metal centers, giving

apparently hexagonal channels with pores open to accommodate the
anions and acetone molecules as guests.
The complexes falling into other groups as shown in Fig. 23(b)-(d),
are significantly less numerous. The chlorine-substituted pyrazine is
212
reported to form a coordination polymer containing both distorted

hexagonal and square frameworks. The reaction of 2-chloropyrazine
(La) with [Cu(CzH4)C1O4]
in acetone yielded orange brick crystals,
. Structural studies have shown both copper(1) ions
[CU,(L~~),,I(C~O~),
involved tetrahedral environments (135).As L3g and the dimethylsubstituted derivatives do, each L, molecule exhibits a bidentate coordination bridging two metal centers forming a square planar Cu4
framework. These square planes are linked to each other by sharing
the edges to afford one dimensional ribbons. The adjacent ribbons are
further bridged by L44,leading to two-dimensional pleated hexagonal
Cue sheets (Fig. 26). The results further demonstrate that in metal
complexes of LSgderivatives, the substituents on pyrazine can greatly
Cu(.
FIG. 26. Schematic drawing of infinite sheets in [CU~(L&](C~O~)~.

Copper atoms
and bridging Ld4molecules are denoted by open circles and solid lines, respectively.
Each pyrazine ring is distinguished by alphabetical label. (From Fig. 2 in Kitagawa,
S.; Kawata, S.; Kondo, M.; Nozaka, Y.; Munakata, M. Bull. Chem. SOC.Jpn. 1993,
66, 3387.1
COPPERU) AND SILVER(J) SUPRAMOLECULAR METAL COMPLEXES
2 13
affect the stereochemistry of the metal, which in turn plays a key role
in determining the framework of the structure.
The copper(1) complexes with pyrazine and tetramethylpyrazine
(LQ5)
may be the best examples to elucidate the steric effects imposed
on the substituted pyrazine toward rational synthesis of copper pyrazine polymeric complexes in the single-crystal phase (137).The single crystals of both [ C U ~ ( L ~ ~ ) ~ ( C H S C Nand
) ~ ~[( C
P FU~~)(~L ~ ~ ) ~ I ( C ~ O ~ )
were prepared by reaction of the corresponding ligand and the copper(1) salt in acetone. In contrast to the disubstituted pyrazine complexes, in which the metal ions in most cases are found to be in a
trigonal environments, the geometry around the copper atom in
[CU~(L,~),(CH~CN)~I(PF,),
is a distorted tetrahedron comprising three
L30 nitrogens and one bent bonding of CH,CN. Consequently, the
hexanuclear unit Cu6 present in the infinite cationic sheet is in a
chair-type cyclohexane-like framework [Fig. 23(c)l. On the other
hand, the L,,-bridged copper(1) complex [ C U ~ ( L ~ ~ ) ~ Icontains
( C ~ O ~a) ~
linear chain polymeric framework. The two copper atoms in the unit
cell have different coordination environments, one with trigonal planar and the other with perfect linear geometry. The two-coordinate
Cu(1) ions form a linear chain framework, and the trigonal ions are
attached to the chain just like a pendant (Fig. 27). The fact that a
hexagonal or cyclohexane-like framework of copper atoms is not present in the cation is due to the bulky Lq5ligands. The four methyl
groups on all the substitutable positions of pyrazine give rise to steric
constraints and, as a result, prevent formation of a hexagonal ring
unit and give a linear link as a less steric form.
y-N\
112.1"
cu\
cu
112.10
C"
C"
\cu'cu\cu~c"\cu~
5" 7" 1"
FIG. 27. Schematic views of the chairlike hexagonal framework in [Cu2(L,,),

(CH,CN),I(PF& (a) and a n infinite chain structure in [Cu2(LQ)31(C104)2
(b). (From Fig.
lb in Kitagawa, S.; Munakata, M.; Tanimura, T. Znorg. Chem. 1992,3I,1714.)
214
C. PHENAZINE
AND BENZOTHIADIAZOLE
SYSTEMS-REGULATED
BY COUNTERANIONS
Like pyrazine, phenazine (La), 2,1,3-benzothiadiazoles (L4,), and

quinoxaline (L4&are of interest in the construction of the molecular
assembly of the metal complexes because they can also act as bridging
ligands linking two metal centers (138, 139). Phenazine formed 1: 1
donor-acceptor complexes with PMDA (pyromellitic dianhydride) and
TCNQ, whose structures show alternate stackings of donor and acceptor molecules (140, 141 ). Several copper(1) and silver(1) complexes
of La have been prepared and their structures determined, and
among these silver(1) nitrate, [Ag2(L4,)(NO,),],is of particular interest
(138).The compound was prepared by reaction of AgNO, and L46 in
methanol at ambient temperature. In this complex each silver atom
is coordinated to two nitrate anions and one L46 molecule in a distorted trigonal geometry, and each L46 molecule in turn links two trigonal silver ions as expected, Fig. 23(d). The most remarkable feature
of the ternary complex is that the counteranion NO, acts as a spacer
linking two L16-bridged Ag atoms, leading to a framework of sixmembered rings of silver atoms extending in the direction of the band e-axes (Fig. 28). The Ag-0 bond distances of 2.432(6) and
2.569(7) A are in the range of 2.367 to 2.689 A for Ag(1) complexes
with nitrate, suggesting fairly strong interaction between the metal
and the nitrate ion. The six Ag atoms on a two-dimensional sheet lie
-.
I3----jJ
FIG. 28. Hexagonal framework in [Ag,(L,,)(NO,),I is constructed by six metal ions

interconnected by L, and NO;. (From Fig. 10 in Munakata, M.; Kitagawa, S.; Ujimaru,
N.; Nakamura, M.; Maekawa, M.; Matsuda, H. Inorg. Chem. 1993,32,826.)
COPPER(I) AND SILVER(I) SUPRAMOLECULAR METAL COMPLEXES
2 15
approximately on the same plane, and each sheet is weakly linked to

its neighbors by the interaction of Ag and the third oxygen atom of
the nitrate anion with an Ag-0 distance of 2.669(7) A, giving a threedimensional structure composed of the two-dimensional sheets. By
contrast, in the corresponding silver perchlorate complex of L46,
[Ag(L4&1O4, the perchlorate ions are not coordinated, and subsequently the complex is composed of infinite -Ag - -Ag - linear chains along the a-axis and a a-a interaction of the L6 aromatic rings between the chains. It is informative enough that in the
mononuclear copper(1) nitrate complex of L 4 6 , [ C U ( L ~ ) ~ ( N Othe
~)],
Cu(1) ion is coordinated to two terminal L46 molecules and chelated
by two oxygen atoms of the nitrate in a distorted tetrahedral environment. The structural differences between the copper(1) nitrate and
silver(1) nitrate complexes of L46 are obviously due to the pronounced
tendency of copper(1) ion for a tetrahedral coordination. Therefore, in
the metal-phenazine system the counteranion as well as metal ion
stereochemistry play an important role in regulation of the molecular
assembly into different frameworks and stoichiometries.
Three copper(1) complexes of L4, have been reported and it reveals
that the frameworks in the compounds are regulated by the counteranions as shown in Fig. 29 (139).Because perchlorate ion has weak
X=CI04
FIG.29. Regulation of the six-membered ring structures by anions in a n L4, system.
2 16
coordination ability, in [ C U ~ ( L ~ ~ ) ~ ( C ~one

O ~ C10;
) ] C ~isO noncoordi~
nating and the other binds weakly with only one copper ion rather
than as a interconnecting ligand. This leaves the two copper atoms
in different coordination environments, tetrahedral and trigonal geometry, respectively. As a result, the complex is composed of a
[cu&L47)G]6+framework containing a six-membered ring of copper
atoms interconnected only by L47molecules (Fig. 30). Copper atoms
in the two-dimensional sheet are arrang$d in a staggered conformation with a maximum deviation of 0.50 A from the mean plane. Replacement of copper(1) perchlorate by copper(1) hexafluorophosphate
in reaction with L47isolated another polymeric complex with the for-
FIG.30. The six-membered ring of copper atoms in [ C U ~ ( L ~ , ) ~ ( C ~ O ~is) composed

IC~O~
of [CU,(L~,)~]~+
unit. The coordination of perchlorate is not shown.
COPPERU) AND SILVERO) SUPRAMOLECULAR METAL COMPLEXES
217
mula [CU(L~~)(HPO~F)].
In this complex each Cu(1) ion is coordinated
to two nitrogen atoms of different L47ligands and two oxygen atoms
of different HP03F- ions in a distorted geometry. An important feature of the structure is that the HPO& anion acts as a n interconnecting ligand between two Cu atoms and takes part along with L47
in formation of the significantly distorted hexagonal framework composed of [C~(L47)4(HP03F)412+
units (Fig. 31). The copper atoms in the
two-dimensional sheet are arranged in a ladder pattern in the acplane and a shallow roof pattern in the ab-plane. As discussed earlier,
nitrate ion is expected to change the framework of the six-membered
rings of copper atoms because it often functions as a bridging ligand.
The copper(1) nitrate complex of L47[Cu(L4,)(N03)Iwas prepared by
reduction of copper(I1) nitrate trihydrate under an ethylene atmosphere followed by reaction with L47in THF. In the complex the tetrahedral coordination of each copper ion is achieved by two L47ligands
'1-
FIG.31. The anion HPO$ acts as an interconnecting ligand between two Cu atoms
(only the coordinating oxygen atom is shown) and takes part in formation of the hexagonal framework in ICu(L,,)(HPO,F)J.
218
FIG.32. Top (a) and side (b) views of the stacks of L4, molecules within and between
2-D sheets in rCu(L4,)(N03)1.(From Fig. 6 in Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Nakamura, M.; Akiyama, S.; Kitagawa, S. Inorg. Chem. 1994,33, 1284.)
and two nitrate anions (Fig. 32). The six-membered rings of metal
ions interconnected by L4, and NO, are extremely distorted, and are
in fact in a chair form, with the ma5imum deviation of the Cu atom
from the mean plane being 2.1-2.4 A. Examined along the c- and baxis, the copper atoms in the sheet are arranged in a ladder and a
roof pattern, respectively. In addition, the shortest interplanar spacing distanFes of L4, moeties within and between the sheets are 3.30
and 3.39 A, respectively. Consistent with the observed V-v interactions in the system, the complex in powder form shows semiconductivity, with electric conductivity cr being 10-6.3S . cm-'.
COPPERU) AND SILVERW SUPRAMOLECULAR METAL COMPLEXES
2 19
It is proposed that a combination of three- and four-coordinate

metal ions and bridging pyrazine and pyrazole molecules is highly
preferred for construction of a two-dimensional sheet structure composed of six-membered rings of metal ions interconnected by aromatic
ligands. The participation of suitable counteranions in coordination
environments of the metal ions can regulate the frame of the ring,
which governs the orientation of the aromatic ligands and the stack
of the aromatic ligand within and between the sheets (139).
V. Hydrogen-Bond-Assembled Frameworks
Self-associated intermolecular and intramolecular hydrogen-bonding assemblies have captured the attention of many research groups
involved in supramolecular chemistry, molecular recognition, and
crystal engineering because suitable matching of hydrogen bond donors and acceptors in number and orientation frequently leads to formation of new molecular aggregates (142-167). Hydrogen bonds are
of paramount importance in biochemistry for determining the secondary structure of proteins and binding substrates to enzymes, receptors, and carriers. They are weaker than conventional covalent bonds
but stronger than van der Waals interactions. Although there is no
universal agreement on the best description of the nature of the forces
in the hydrogen bond, it is generally accepted that hydrogen bonding,
X-H----Y, occurs only between a hydrogen atom bound to an electronegative atom X and another atom Y that is also highly electronegative and has one or more lone pairs, enabling it to act as a base (142).
The concept of hydrogen bonding was originally used to explain physical properties of simple organic and inorganic compounds such as abnormally high boiling points and heats of vaporization, but in recent
years it has been noted that the introduction of a hydrogen-bonding
interaction between ligands in transition-metal complexes is an indispensable tool for creating a variety of molecular architectures in a
predictable fashion via self-assembly and molecular recognition.
Several important reviews related to this field now are available
( 142-1 4 7 1.
With the combination of the covalent bond forming capability of
the metal ion and the inherent capability of the ligand for formation of complementary hydrogen bonds, a diversity of H-bonded structures has been obtained, including intramolecular hydrogen-bonded
monomers (148), intermolecular hydrogen-bonded dimers, tetramers
(149-151 ) and polymers having one-dimensional chain (152), two-
220
dimensional sheet (153-158) and three-dimensional networks (159163). The most commonly encountered hydrogen-bonding interactions
in H-bonded supramolecular frameworks are 0 -H----0 and
N -H----0.Additionally, some weak hydrogen bond interactions such
as C-H--0, C-H----Cl and even X-H----arene are also reported
(146).The interaction may involve only one hydrogen bond donor and
one acceptor, may be bifurcated or trifurcated, and may occur between atoms, molecules, or ions (142).In this respect, the bifunctional
ligands should contain the simple functional groups such as carboxyl,
amide, amino, cyano, and pyridone. Recent reports have shown that
hydrogen bonds can occur in the system between ligand functional
groups and some inorganic anions such as BF,, C104, SO$-, and halides, and even water and ammonia molecules (146, 148, 158). The
ligands involved in the following discussion of the hydrogen-bonded
complexes of Cu and Ag are listed in Fig. 33.
Mingos has shown in his recent review how the application of molecular recognition principles based on triple hydrogen bonding resulted in the crystal engineering of aggregates based on ligands that
can simultaneously form stable and inert metal-ligand bonds and
have recognition sites for complementary arrangements of hydrogenbond donors and acceptors (144),which has been excellently demonstrated by self-assembly of the copper(I1) complex [Cu(L,,), . 2 melamine] (HL4, = 5-(2-pyridylmethylene)-hydantoin)in which L4, involves simultaneous coordination with Cu and an ADA=DAD (A =
hydrogen-bond acceptor, D = hydrogen-bond donor) triple hydrogenbonding arrangement with melamine molecules (Fig. 34)(156).
Among extended polymeric structures, two-dimensional hydrogenbonded frameworks are enormously popular, and the resulting assemblies resemble practically crinkled tapes (153,154)and infinite sheets
(155-158). Smith and co-workers have structurally characterized a
b 2
HL53
L54
FIG.33. List of ligands in copper(1) and silver(1) complexes having hydrogen-bonding networks.
221
FIG. 34. Hydrogen-bonding networks in [Cu(L4&.2 melamine]. (From Fig. 2 in

Chowdhry, M. M.; Mingos, D. M. P.; White, A. J. P.; Williams, D. J. J. Chem. SOC.,
Chem. Commun. 1996,899.)
number of silver(1) carboxylates, and among these phthalate and

trimesate (LEO)
have been found to form hydrogen-bonded polymers
involving the ammonia molecule (155). The silver(1) complex
[Ag2(phthalate)(NH3)21
contains a phthalate-bridged dimeric unit in
which each Ag atom is coordinated to one ammonia molecule and a
single phthalate carboxylate oxygen in an essentially linear fashion.
All three amine hydrogens are involved in intermolecular hydrogenbonding interactions, giving a two-dimensional hydrogen-bonded
sheet structure extending across the bc-direction of the cell (Fig. 35).
The adjacent sheets are further linked by hydrogen bonds via the
uncoordinated carboxyl oxygens. In the corresponding trimesate
complex, [NH,l[Ags(L5,),(NH,)z(H20)21
. HzO, a two-dimensional sheet
framework is based on a pseudo-centrosymmetric S-type trimer unit
linked by the carboxylate groups of two independent trimesate residues, and the hydrogen bonds between the uncoordinated carboxylate
oxygens and amine and water molecules stabilize the structure. Likewise, in the copper(I1) complex of 4,4-bipyridine ( L d , the coordinating water molecule bridges between the metal center and uncoordinated LS1,affording two-dimensional rectangular grid sheets (158).
The structural determination of the Cu(1) complex of 2-cyanoguanidine (LS2),[{Cu~L51)2},~L51)][BF~l,~
MeCN, reveals that the structure
contains a dinuclear cation in which two T-shaped copper(1) centers
are bridged by L51and terminally coordinated by two L5%molecules
222
FIG. 35. Hydrogen-bonding networks in [Ag,(phthalate)(NH&I. (From Fig. 2 in

Smith, G.; Reddy, A. N.; Byriel, K. A.; Kennard, C. H. L. J. Chem. Soc., Dalton Trans.
1995, 3565.)
(157).An extensive hydrogen-bonding network between BF, ions and

LS2molecules gave a two-dimensional sheet structure as shown in Fig.
36. It is intresting to note that all four F atoms for each anion are
involved in hydrogen bonding, three of which are sitting within the
two-dimensional sheets and the fourth bridging the sheets.
The principles of molecular recognition can also be used to assemble three-dimensional structures, but additional requirements on the
number and orientation of the complementary hydrogen bonding motifs must be met (159-163). Nakasuiji has selected biimidazole, pterine,
lumazine, and glycoxime as the target ligands because their chelating
ability to a metal element and multi-H-bonding sites, and has constructed a number of three-dimensional hydrogen-bonded networks of
metal ions such as Ni(I1) and Cu(I1) (160, 161). Molecules in which
the motifs are related by Td or even S4symmetry are strong potential
candidates for the engineering of a three-dimensional network exhibiting microchannels or diamondoid structures (146). The system we
have worked with are literally double hydrogen bonding of the ligands
combined with tetrahedral copper(1) ion in an attempt to design unprecedented two-dimensional and three-dimensional molecular architectures.
A. THREE-DIMENSIONAL
SUPRAMOLECULAR
Cu(1) COMPLEXES
WITH CHANNELS
(HL53)
The bifunctional ligand 3-cyano-6-methyl-2(1H)-pyridinone
as shown in Fig. 33, possesses both a coordination group (CN) and a
223
FIG.36. Two-dimensional sheet structure of [(Cu(Lsz)2}~(L~,)11BF112

MeCN showing
hydrogen-bonding associations. (From Fig. 2 in Batsanov, A. S.;Begley, M. J.; Hubberstey, P.; Stroud, J. J . Chem. Soc., Dalton Trans. 1996, 1947.)
hydrogen-bonding site (pyridone), and it would be a suitable candidate for the formation of H-bonded metal complex supramolecules.
The second reason for choosing it in the study is its relatively small
size, which would reinforce the unidentate coordination of each ligand
to the tetrahedral copper(1) ion without steric hindrance. The reactions of copper(1) salts with HLssin acetone have isolated four coordination polymers, [Cu(HL,,),IX, where X = ClO,, BF; , PF;, and CFs
SO, (164). Although each structure contains a three-dimensional
framework of tetrahedral CuNj centers linked by intermolecular hydrogen bonds through pyridone N and 0 atoms in a head-to-tail mode,
different patterns of hydrogen bonding give rise to two types of different frameworks-namely, square channel and superadamantane networks-depending on the kinds of counteranions. In the complexes
with relatively small anions, perchlorate and tetrafluroborate, each
HL53molecule is hydrogen bonded to two adjacent others through pyridone N and 0 atoms; that is, each I C U ( H L ~ ~ ) entity
~ I X is connected
to eight neighboring counterparts as shown in Fig. 37. The dihedral
FIG.37. Part of the hydrogen-bonded structure (a), packing diagram (b), and spacefilling model (c) of [Cu(HL,),]CIO,. (From Figs. 2 and 3 in Munakata, M.; Wu,L. P.;
Yamarnoto, M.; Kuroda-Sowa, T.; Maekawa, M. J.Am. Chem. SOC.1998, 118, 3117.)
225
angle between the hydrogen-bonded HLS3groups is 83.44" and 84.98"

for the perchlorate and the tetrafluroborate, respectively. This results
in four [Cu(HL,,),IX monomers interconnected through hydrogen
bonds, forming a layer with an open cavity occupied by the counteranion. The overall structure is made of such layers, each consisting of
an infinite square array of copper atoms coordinated with HLs3groups
and bridged by hydrogen bonds, leading to the cross-linking of a given
layer with those immediately above and below it. A square channel
framework is generated with a relatively small size of cavities (ca.
12.1 A in diameter) filled with guest anions. Let us name such hydrogen bonding as A type, illustrated schematically in Fig. 38. The hydrogen bonding network in [CU(L,~)(HL,,),Iis also found to belong to
this type (165).
When the small anion is changed to the drastically larger one, PF;
or CF,SOi, the cavities in the square channel lattice for the perchlorate and tetrafluroborate cannot be effectively adapted for large guest
ions only by elongating the hydrogen bond distances. If the stoichiometry is to remain the same, one way for the system to respond
and avoid this destablizing repulsion is to rearrange the hydrogenbonding mode as B type, that is, each HLS3moiety in the monomeric
unit of [Cu(HL,,),]X is head-to-tail hydrogen bonded to only one adjacent HL,, molecule, rather than two, with a rather smaller dihedral
angle (0"-11.43"). Thus, one [CU(HL,,)~IXentity is connected to four
adjacent others, and these repeating units form a three-dimensional
adamantane architecture of metal atoms (Fig. 39). The diamondoid
frameworks are stacked with each other in such way that all the copper centers are found on lines parallel to the c-axis, which flefines
enormous linear chambers with larger open cavities (ca. 13.3 A in diameter), that just fit the large anions PF; and CF3SOi.
To show how the assembling process would respond in the system
0-1 1 '
FIG.38. Schematic presentation of two types of' hydrogen bonds in HL, complexes.
226
CU
FIG.39. Part of the hydrogen-bonded structure (a), packing diagram (b), and perspective view of diamondoid framework (c) in [Cu(HL,)JPF,.
227
when H-bonding interaction is absent, the structures of two copper(1)

complexes of L53were determined (165, 166). It reveals a completely
different stoichiometry and framework in both [c~g(L63)4]c104and
[Cu,o(L,3),](BF,)2,
in which each deprotonated LS3monoanion employs
all its three functional groups bridging three rather than two copper
atoms, forming a supramolecular channel architecture containing a
Cud cluster (Fig. 40).These findings suggest that the bifunctional ligand group m
e
3 possesses unique ability for the molecular recognition, which has decisive influence on crystal engineering of coordination polymers. Modification of the hydrogen-bonding mode and
distances as well as the direction of H-bonding sites at molecular level
can effectively facilitate transition of one framework to the other.
FIG.40. Part of the molecular structure (a) and extended framework (b) in [Cu,(L,,),I
c10,.
228
B. HYDROGEN-BONDINGAND ~-~-STACKING-~SEMBLED
Cu(1) COMPLEXES
The second bifunctional ligand selected to construct the metal architecture is 2-hydroxyquinoxaline (Lu), because it also possesses
both N-coordination sites and potential double H-bonding sites
(CN=COH or CNHC=O if it undergoes tautomerization). The reaction of copper(1) perchlorate and Ls4under an atmosphere of ethylene
gave the complex [ C U ( L ~ ~ ) ~ ( C ~ H ~
whose
) ] C ~structure
O~,
was determined by X-ray analysis (167). Its unique feature is the cooperative
effect of hydrogen-bonding and aromatic-stacking interactions simultaneously present in the system. The metal center is in a trigonal
planar geometry achieved by coordination to one ethylene and two Ls4
molecules. The cation maintains the fundamental feature of the free
ligand in the sense that two LE4molecules are linked to each other by
head-to-tail double hydrogen bonds, resulting in formation of an infinite zigzag chain (Fig. 41). The IR spectrum of the complex shows low
v(NH) stretching frequency in the region 2950-3050 cm-', consistent
with the short NH----0hydrogen bonds observed. Between the adjacent chains the H-bond assembled La4moleculeoplanes are stacked to
each other, with the average distance of 3.30 A. This gives a unique
two-dimensional cooperating structure stabilized by hydrogen bonding and r-r interactions as well as covalent bonds, reminiscent of a
proton-electron transition (PET) system. It is worth mentioning that
ethylene also plays an important role in the construction of the resultant architecture as a spacer because of its relatively small size,
which reduces the packing volume, ensuring the effective stacking of
the aromatic planes.
VI. a-a-lnteraction-Assembled Frameworks
A. INTERMOLECULAR
n-n INTERACTION
IN DISCRETE
COORDINATION
COMPOUNDS
Planar coordination compounds with aromatic ligands (Fig. 43,especially those having extended 7~ systems, show r-r interaction in
solid states. In alkene or alkyne n-bonded Cu(1) complexes, in-plane
coordination of a C =C or C = C bond to trigonal planar Cu(1) centers
often leads to planar molecular conformations (167-1 72). The infinite
n-r stacking columns are confirmed in the 2,2'-bipyridine (Lss) com-
229
FIG. 41. Packing diagram of LCu(L,)e(C2H,)IC104; side view showing H-bonded networks (a) and top view showing aromatic stacking interactions. (From Fig. 2 in Dai, J.;
Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Munakata, M. Inorg.
Chern. 1997,36,2688.)
plex [CU(L~~)(C~H,)]C~O,
(169)and the 1,lO-phenanthroline (LW)complex [ C U ( L ~ ~ ) ( C ~ H C O ~ E(1
~ )71
]C
O ~ nearest carbon-to-carbon
) ~
with
distances of 3.31 and 3.37 (172,1731,respectively, and dimer forma(171)
tion through n-rr interaction can be seen in [Cu(LW)(CZHz)IC1O4
[the nearest carbon-to-carbon distances are 3.42 ( 172)l.
Flat conformations are also observed in cationic parts of dimeric
Cu(I) and Ag(1) compounds bridged by two l&naphthyridine ligands,
(1741. The former comLS7, [ C U ( L ~ ~ ) ] ~ ( C
and
~ O[Ag(L67)12(C10,)2
~)~
pound forms a dimer through n-n interaction of the neighboring L67
230
L56
L55
ae
L59
L61
L62
Me.
Me
L63
L65
*@
\ /
L67
L68
L69
L66
ip
/
L70
Me
@
\ /
L71
L72
L73
L74
FIG.42. Ligands that appear in Sections VI and VII.
molecules; the latter shows one-dimensional infinite :tacking column

of the cationic part with interplane distance of 3.40 A (172) as shown
in Fig. 43. Another flat dimer, [Cu(L5,)l2(175), also shows a onedimensional a-a stacking column, with interplanar separation of 3.41
A (172).
Not only planar compounds but also nonplanar ones show a-a
stacking interaction. In the crystal structure of CUI(L& (176), two
23 1
3
L
-3
-c
3.40 A
Co=-o; 3 - J - 3
Ag
napy
FIG.43. A crystal-packing view of [Ag(L5,)1~(C104)2

showing r-n interaction between
planar cationic parts.
2,6-dimethylpyridines Lse and a n iodine coordinate to a Cu(1) ion in

T-shaped trigonal geometry, and the dihedral angles between two ligand planes and the CuINp plane are 87.6 and 71.2", indicating nonplanar conformation. However, one-dimensional n-n stacking can be
easily recognized in Fig. 44, with a nearest carbon-to-carbon distance
of 3.447 A. Changing the counteranion from an iodide to a perchlorate
ion gives two-coordinated copper(1) or silver(I) compounds. Whereas
planar cationic parts of [M(Lsg),1X(M = Cu, Ag; X = ClO, , NO,) (177)
with two-coordinated M(1) centers show only weak n-n interytion,
with the nearest carbon-to-carbon distances being 3.59 to 3.66 A, one
of the polymorphs of [ C U ( L ~ ~ ) ~ I(178),
C ~ Owith
~
a dihedral angle between two LS9ligands of 5 6 2 , shows a close n-n contact of 3.33 A
(172) between two Ls9ligands residing in adjacent molecules. We can
see not a one-dimensional n-n stacking column but a n-n-interactionassembled one-dimensional structure like ----Ls9
- Cu -Lse---L59- Cu-L59----.
Although a trigonal planar geometry is maintained around a Cu(1)
center in the ethylene-coordinated complex [Cu(L54)z(CzH4)IC104
(1671,
the two Ls4 molecules coordinated to the Cu(1) ion have a dihedral
angle of 102.5", giving a butterfly structure. Repetitions of double hydrogen bonds between two Ls4ligands in neighboring complexes form
an infinite zigzag chain (Fig. 41). Additionally, a n-n interaction between two LMligands residing in neighboring chains (the interplanar
separation of 3.30 A) forms a two-dimensional structure as shown in
Fig. 41. This gives a unique two-dimensional cooperating structure,
which is a fundamental characteristic of proton-electron transfer
(PET) systems (179, 180).
232
3.447 A
c sinp
FIG.44. A crystal-packing view of CuI(L6& showing r-a interaction between L60

ligands. (From Fig. 5 in Healy, P. C.; Pakawatchai, C.; White, A. H. J. Chem. SOC.,
Dalton Trans. 1983, 1917.)
The copper(1) ion in [ C U ( L ~ ) ~ ] CMeOH

~ O ~ (181) has a distorted
trigonal geometry with three nitrogen atoms of three 1-aminopyrenes
(Lm).Two of them are parallel to each other, whereas the rest lie approximately perpendicular to the other two as shown in the molecularpacking view (Fig. 45). Intra- and intermolecular n-n interactions of
the former two form a one-dimtnsional stacking column with interplane separations of 3.40-3.43 A. Moreover, intercolumn n-n interactions between LBOligands with an interplane separation of 3.52 A
result in the formation of a two-dimensional structure.
The copper(1) ion in [Cu(L&H,0)IC1O4 (138)has a distorted trigonal geometry with coordination of two nitrogen atoms of two terminal
9
COPPERU) AND SILVERW SUPRAMOLECULAR METAL COMPLEXES
233
FIG. 45. Molecular packing view of [Cu(L,)JClO,. MeOH. (From Fig. 2 in Munakata, M.; Dia, J.;Maekawa, M.; Kuroda-Sowa, T.; Fukui, J. J . Chem. SOC.,Chem. Cornmun. 1994, 2331.)
L4 molecules and a water oxygen. The coordinated L4 molecules form

an alternating r-n- st9cking column, with the nearest intermolecular
separation being 3.46 A.
Because basicity of acridine
is stronger than that of La (1821,
adding a solution of L61 in methanol to a solution containing copper(1)
and L4 gives a discrete complex, [ C U ( L ~ ~ ) , N O , IH(,L
O)~,,~ ~(183).The
copper(1) ion is trigonally coordinated by the two nitrogen atoms of L,
and one of the oxygen atoms of the NO; anion. Two coordinated L61
molecules are not parallel to each other at a dihedral angle of 63.6",
and each has a r-n- interaction with an adjacent Ll molecule coordinated to another copper atom (Fig. 46). The interplane separations of
these intermolecular r-r interactions are 3.41 and 3.49 A. A metalfree L4 is sandwiched between L61 ligands, forming an infinite columnar stacking along the a-axis with a L46-L61-]LB, repeating unit. The
closest C----Cdistance of 3.39(1)A indicates a significant n--r interaction between L4 and 4 1 .
An L46-bridgedCu(1) dimer complex, [CU,(L~,)~(M~OH),I(L~~)(PF,),
(138) shows an alternate r-r stacking of coordinated and uncoordinated La. Two copper(1) ions are crystallographically the same: a
T-shaped three-coordination of a methanol and terminal and bridging
La with a remarkably large N(termina1 L&Cu-N(bridging La) angle
of 156.1". A metal-free L4 molecule is located near the bridging L4,
with a dihedral angle of 7.6"and with a nearest carbon-to-carbon sepa-
234
FIG. 46. Molecular views of [Cu(L6&N03](La.H 2 0 ) , ,indicating r-r interaction between Lp6and La1 (a) and between two LBLligands (b). (From Fig. 4 in Kuroda-Sowa,
T.; Munakata, M.; Matsuda, H.; Akiyama, S.; Maekawa, M. J. Chern. Soc., Dalton Trans.
1995, 2201.)
ration of 3.42A, indicating a significant ?T-T interaction between them.

Molecular packing along the c-axis reveals one-dimensional T-?T stacking of bridging and metal-free La alternately.
B. INTERAND INTRAPOLYMER
r-rr INTERACTION
One-dimensional polymeric structures are observed in copper(1)
complexes with 2,9-dimethyl-l,lO-phenanthroline,
[CU(L~~)(CN)]
and
[CU(L62)(NCS)](184). Cn- or NCS- acts as a bridging ligand, and L62
chelates to a copper(1) ion to block two of four coordination sites of
the tetrahedral center, resulting in the infinite zigzag chain. In both
compounds, zigzag chains are connected through T-?T interaction between L62 molecules residing in neighboring chains, which form two-
235
dimensional polymeric structures. Similar two-dimensional network

formations are also observed in the trinuclear copper(1) complex with
2,2-biquinoline (La) [CU,(L,)~(CN),] (1851, in which tetrahedral copper(1) ions coordinated by chelating L63 ligands are bridged alternately by Cu(CN)i and CN-.
When La is used as a bridging ligand to construct coordination
polymer compounds, interpolymer T-T interactions are often observed. A one-dimensional zigzag chain composed of bridging L,, and
a tetrahedral copper(1) ion can be seen in [CU(L,~)(NO~)]
(183) (Fig.
47). The copper(1) ion is coordinated by the two nitrogen atoms of the
two L46 molecules and by the two oxygen atoms of the NO, anion.
FIG.47. Top (a) and side views (b) of the crystal packing of [Cu(L,,)(N03)]showing
the La columnar stacks. (From Fig. 2 in Kuroda-Sowa, T.; Munakata, M.; Matsuda, H.;
Akiyama, S.; Maekawa, M. J. Chem. SOC.,Dalton Trans. 1995, 2201.)
236
Two L46 molecules coordinated to the same copper atom are inclined
to each other at a dihedral angle of 82.2'. As can be seen in Fig. 47,
the L46 molecules are stackFd in the b-axis direction, with an
interplane separation of 3.47 A. Thus, the compound consists of onedimensional zigzag chains of copper atoms and L46 molecules, interconnected through L4&n-n interactions, resulting in a two-dimensional interaction. A similar zigzag chain composed of bridging 4 6
and a trigonal planar copper(1) ion is observed in [cu(L46)
(MeCN)],(L,)(PF,), (La : pyrene) (183).The L46 molecules are stacked
in the c-axis direction with an interplane separation of 3.47 A. Thus,
the compound consists of one-dimensional zigzag chains of copper
atoms and L46 molecules interconnected through L, n-n interactions,
resulting in a two-dimensional interaction.
A series of halogen-bridged ~ o p p e r ( I ) - L compounds,
~~
[CU~(~-X)~
(L46)l (X = I, Br, or C1) (1861, show one-dimensional (X = I) or twodimensional (X = Br or C1) polymer frameworks. CuzIzrhomboids in
the iodide compound are connected by bridging L, ligands through
trigonal planar Cu(I) centers, forming an infinite straight chain. The
dihedral angle of the rhomboid and L46 plane is 75.28'. The n-n interaction between L46 molecules of adjacent chains (interplanar distances of 3.46 A) gives the complex two-dimensional structure. On the
other hand, both the bromide and the chloride compounds have almost same structure: CuX infinite stairs bridged by L
, through coordination to distorted tetrahedral (Cu(1) ions, forming a two-dimensional network structure as shown in Fig. 48 for the bromide
compound. Intrasheet r-n interactions betwen LlS molecules are also
observed in both compounds, with interplanar distances of 3.40 and
3.36 A for the bromide and the chloride compounds, respectively. In
the solid-state 13C NMR spectra of these compounds, the increase of
the upfield shifts of the resonances assigned to the quaternary carbon
atoms of L46 upon coordination (-2.4, -2.9, and -3.9 ppm for the
iodide, the bromide, and the chloride compounds, respectively) are
well correlated to the decrea!e in the interplanar distances of L 4 6 molecules (3.46, 3.40, and 3.36 A, respectively).
Not zigzag but completely straight one-dimensional chains can be
seen in [Ag(L46)I(c104)
(138),in which the Ag(1) ion has a linear two
coordination of two nitrogen atoms of bridging L46 ligands. All the L 4 6
molecules are parallel, as shown in Fig. 49, which results in the formation of a one-dimensional chain structure like a flat ribbon. The
shortest intermolecular distance is 3.36
indicating significant T-n
interaction, though the overlap between them is not so large. Because
one chain interacts with four neighboring chains through n-n interac-
A,
COPPER(1)AND SILVER(1) SUPRAMOLECULAR METAL COMPLEXES
237
FIG.48. Top (a)and side views (b) of the packing arrangement of ICu2(pBr)2(p-L4B)I.
(From Fig. 4 in Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Honda, A.; Kitagawa,
S. J. Chem. SOC.,Dalton Trans. 1994, 2771.)
tions, this system has a three-dimensional network assembled by

TI-n interaction.
A two-dimensional sheetlike structure in [Ag2(L,,J(N03),I (138) is
composed of bridging L46 and bridging nitrate anions coordinated to
trigonal Ag(1) ions, forming six-membered rings of Ag(1) ions. The significant n-n interaction between two L46 molecules in the neighboring
sheets (the shortest intermolecular distance of 3.34 A) indicates the
formation of a a-n-interaction-assembled three-dimensional network
structure.
Perylene molecules &4), having an extended n-system, easily form
a-n interaction. Four peripheral C=C moieties of Lfi9coordinate to
238
FIG.49. Molecular structure (a) and a perspective view of the packing arrangement
(b) of [Ag(Le)l(CIOa).(From Figs. 7 and 8 in Munakata, M.; Kitagawa, S.;Ujimaru, N.;
Nakamura, M.; Maekawa, M.; Matsuda, H. Znorg. Chern. 1993,32,826.)
four Ag(1) ions in $-fashion in [Ag2(L6s)(C104)2]

(187).The silver(1) ion
has a distorted tetrahedral geometry with two C=C groups of two L69
molecules and two oxygen atoms of two perchlorate anions. The resulting two-dimensional sheet shows a W-type wavy conformation 5s
shown in Fig. 50, which enables effective n-n interaction (3.31 A)
between L@ molecules in adjacent sheets. Thus, the n-n interaction promotes the network dimensionality from two to three dimensions.
Chair-formed six-membered rings of Cu(1) ions in [Cu(L4,)(NO,)I
(139)form a two-dimensional sheet with a certain thickness, in which
each Cu(1) ion is bridged by two L4, and two nitrate anions. Because
L4, molecules are almost parallel to the two-dimensional sheet, both
intra- and intersheet n-n interaction can be seen, with inter-L4, separations of 3.30 and 3.39 A respectively. The latter interaction enables
the three-dimensional network structure assembled through n-n interaction.
239
(4
FIG. 50. W-type wavy conformations of two-dimensional sheets in IAg,(LBB)(C104)ZI.

(From Fig. 2 in Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga,
Y.; Sugimoto, K. Znorg. Chern. 1997,36,4903.)
240
VII. Diamondoid Frameworks
We can imagine a formation of a cubic or a hexagonal diamondrelated framework (Fig. 51) from a combination of a tetrahedral metal
ion and a rodlike bridging ligand. Although both types of frameworks
are found in minerals or polymorphs of ice (1881, so far as we know,
only the former type of framework is known in coordination compounds. Hereafter, we use the terms diamond or diamondoid
framework to mean cubic diamond-related framework. As Hoskins
and Robson have proposed (1271, a compound having a diamondoid
framework, if it can have a large cavities or channels inside, can offer
a number of features of potential interest such as molecular sieve
properties, heterogeneous catalytic properties, and mechanically
strong materials with an unusually low density. Additionally, T-T interaction between aromatic bridging ligands often controls the degree
of interpenetration of diamondoid frameworks and sometimes gives
electronic conducting material.
In this section, we will focus on coordination compounds with diamondoid frameworks, especially containing copper(1) and silver(1)
ions. These ions having d10 electronic configuration are suitable for a
tetrahedral metal center in a diamondoid framework. Table V lists
the coordination polymers having diamondoid frameworks reported so
far, together with other coordination polymers having related frameworks.
A. BRIDGED
BY PYRIDINEOR
PYRAZINE
DERIVATIVES
Pyrazine, LS8,and its derivatives are the shortest bridging ligands

(133)has a tetrahenext to cyanide. The copper(1)ion in [CU(L~~),IPF,
dral geometry, with coordination of four N atoms of Ld0.The other
end of the N atoms of each L40 coordinate to other copper(1) ions re-
%%
FIG.51. Cubic (left) and hexagonal (right) diamond-related frameworks.
COPPER^ AND SILVER^ SUPRAMOLECULAR METAL, COMPLEXES
241
TABLE V
COORDINATION
POLYMERS
HAVING
DIAMONDOID
AND RELATED
FRAMEWORKS
Compound
Bridged M~-;-M
Degree of
distance (A) interpenetration
Remarks
5.03
5.11
5.46
6.99
8.856"
8.9
5.04, 9.54
9.93
ca. 10
10.90
11.16
11.6
11.76
11.9
12.09
12.76
13.5
13.55
16.4
17.0
Ref.
127
127,204,205
127,204,205
I33
127
71-71
71-71
8-8
8-71
71-77
71-8
193
206
190
208
207
136
190
4
195, 196
194
197, 798
164
8-71
192
71-71
201,202
203
71-8
Zinc-blende structure.
', Cu----Cudistance.
sulting in the formation of a diamondoid framework (Fig. 52). The

shortness of the bridged Cu----Cuseparation of 6.99 A together with
the presence of the PF; counteranion and methyl groups in L4,,prevent the interpenetration of any other framework. Although coordination chemistry of Ag(1) and L39(130, 289) revealed several compounds
having interesting three-dimensional observed such as cw-ThSi2 or
ReOs types, no diamondoid framework has been observed in this
system.
When 4,4'-bipyridine, Leo,is used instead of L4,,fourfold interpenetrated diamondoid frameworks are obtained for both Cu(1) and Ag(1)
ions. The Cu(1) compound [Cu(L&lPF6 1136) exists as four independent concatenated diamondoid frameworks with Cu----Cuseparations
of 11.16 A. The Cu(1) centers occupy crysta!lographic 4 positions and
hence all Cu-N bonds are identical, 2.034 A. The PF; counteranions
242
FIG.52. A single diamondoid framework in [Cu(L&lPFs. For clarity, the PF, anions
are omitted.
occupy channels that are parallel to the c-axis and they sit on fourfold
(190) also
crystallographic axes. The Ag(1) analog [Ag(LBo)~CF3So3
contains four interpenetrating diamondoid frameworks (Ag----Agintraframe separations of ca. 11.6 A). The Ag(1) cations display a distorted tetrahedral geometry, with Ag-N contacts different for the two
Interconnection of
independent LWligands (mean 2.270 vs. 2.380
trigonal Cu(I) centers through LBO was achieved by hydrothermal
synthesis (191). The resultant crystalline compound [cU(L60)1.5]*
NOs(H20),,26
shows sixfold interpenetrated a-ThSi2-typeframeworks.
By lengthening the linking bipyridyl ligand via insertion of a trans
C=C double bond between the pyridyl units, the length of the resultant briding ligand, 1,2-truns-(Cpyridyl)ethene (L70), is extended
by about 2.4 A and the interpenetration in diamondoid frameworks
is also changed. The Locompound with Cu(I), [ C U ( L ~ ~ ) ~ ~ B F ~ ( O . ~ C H ~ C ~
(192),has fivefold interpenetrated diamondoid frameworks. The Cu----Cu
distances bridged by L70 range from 13.33 to 13.82 A to create large
cavities within the diamondoid framework as shown in Fig. 53. All
five independent frameworks are polycatenated with channels throughout the structure: these channels accommodate BF; counteranions
and CH2C12solvent molecules.
Among the bipyridine derivatives, L64 can also act as a linear bridging ligand if the transoid conformation is maintained. [Cu(L,&]X
(X = BF4, PF,) (1931, having a twofold interpenetrated diamondoid
framework, can be obtained as yellow-green triangular crystals. The
A).
243
FIG. 53. View of the structure of [Cu(L,,),JBF,. (From Fig. 2 in Blake, A. J.;Champness, N. R.; Chung, s. s. M.; Li, W.-s.; Schroder, M. J. Chem. Soc., Chem. Commun.
1997,1005.)
LM has surely a transoid conformation, but the rings twist out of

plane relative to each other by about 30". In addition, the Cu-N bond
is bent out of the plane of the aromatic ring by 19,leading to a closing of the Cu-L-Cu angle to 134". Bridged by this large ligand, the
resulting framework contains copper atoms separated by 8.9 A and
significant void volume. The extra space is occupied by an identical
network forming a twofold interpenetrated structure (Fig. 54).
B. BRIDGED
BY BISNITRILE
LIGANDS
Nitrile compounds are generally accepted to be weaker donors than
pyridine derivatives; still, they can coordinate to d'O metal ions. A
bisnitrile compound, if the two CN groups have opposite directions,
can act as a rodlike bridging ligand in the construction of diamondoid frameworks.
Although directions of two CN groups in alkyldicarbonitrile,
NC(CH,),CH, are not fixed a priori due to free rotation around methylene carbons, a Cu(1)-adiponitrile ( n = 4) complex, [CU(NC(CH~)~
244
FIG.54. A schematic illustration of twofold interpenetrated diamondoid frameworks

of [Cu(L&lX (X = BF,, PFB).The spheres represent the copper atoms, and the twisted
bipyridine ligands are shown as the bet cylinders linking the copper centers together.
(From Fig. 2 in Lopez, S.; Kahraman, M.; Harmata, M.; Keller, S. W. Znorg. Chem.
1997,36, 6138.)
CN),INO, (1941, shows a diamondoid framework. In the crystal, four

methylene groups of the adiponitrile form a planar zigzag chain and
the terminal CN groups align almost antiparallel, which allows the
adiponitrile to behave as a linear bridging ligand. The bridged
Cu----Cudistance of 12.09 A together with the small size of the counteranion (NOi3)enable it to show sixfold interpenetration of independent diamondoid frameworks.
Terephthalonitrile, L66, having two CN groups at the 1- and 4-positions in a benzene ring, is a good candidate for a rodlike bridging
ligand for d'" metal coordination polymers. When the aceonitrile liare substituted by L65, the crystals formed
gands of [CU(CH~CN)~]BF~
have the composition [Cu(L6,),]BF4(4). Diamondoid framgworks are
indeed formed (the bridged Cu----Cudistance of 11.76 A), but the
structure contains fivefold interpenetrated independent frameworks.
Channels of rhombic cross section are generated in which the BF,
ions are located.
Methyl substitution of La6 causes a drastic change in the interpenetration of the diamondoid frameworks. Crystals of
[Cu(L,,)21X(L66)(THF)(X = BF4, C104) (195, 196) were obtained by
solution into the corresponding copper(1)soluslow diffusion of the LBB
tion. X-ray crystal structure analysis revealed that both compounds
contain threefold interpenetrated diamondoid frameworks as shown
in Fig. 55. Because the bridged Cu----Cudistance in these compounds
(11.9 A) is almost same as that bridged by L65 in Cu(L&BF4 (41, the
lesser interpenetration in LBBcomplexes should be caused by the bulkiness of methyl groups in L66. This is also strongly correlated to the
incorporation of guest molecules, a metal-free LBB,and THF. The
metal-free dmtpn incorporated in the cu(L& lattice participates in
the formation of a T-T stacking column (discussed later).
COPPER(I) AND SILVERW SUPRAMOLECULAR METAL COMPLEXES
245
FIG. 55. A threefold interpenetrated diamondoid framework in [ C U ( L ~ ) ~ I X ( L ~ )

(THF). For clarity, the
ligands are represented by lines and the PF, molecules are
omitted.
Among the compounds having diamondoid frameworks listed in Table V, the copper complex with 2,5-dimethyl-NJV-dicyanoquinonediimine
[Cu(L7,),I (197, 1981, has a quite interesting conducting
property. Recent developments on this and related compounds show
other interesting features including metal-insulator-metal transition
(reentrant behavior) (199),three-dimensional Fermi surface character
(2001, and so on. The structure of [Cu(L,J21shows sevenfold interpenetrated diamondoid frameworks with strong n-n interaction (the interplanar distances between the neighboring L71 molecules are 3.13.2 A). It should be noted that not even-number-fold but sevenfold
interpretation in this class of complexes probably disturbs the dimer
formation between adjacent LT1molecules, which contribute to the
aforementioned conducting behavior.
A ninefold interpenetration of diamondoid frameworks, the highest
degree thus far reported, is observed in a series of Ag(1) complexes
with 4,4-biphenyldicarbonitrile,L72 (201, 202). Crystallization of L72
with AgX (X = PF6,AsF6, SbF,) (1: 1molar ratio) by heating and slow
cooling from ethanol produces yellow crystals of [Ag(L72)21X
(X = PF,,
AsF6, SbF,). All of these structures show ninefold interpenetrated diamondoid frameworks. This highest degree of interpenetration is a result of the rather large ligand used (the bridged M----Mdistance of
16.4A). The large amount of void volume created by a single network
is filled by eight identical nets (Fig. 56).These nets consist of parallel
ligands offset along the long axjs and display n-n stacking at a planeto-plane distance of 3.4to 3.6A. Columns of counterions are revealed
246
FIG. 56. A ninefold interpenetation of diamondoid frameworks of [Ag(L&IX (X =

PF,, AsFs, SbF,). For clarity, the L,, are represented by lines and the PF, are omitted.
(From Fig. 1 in Hirsch, K. A,; Wilson, S. R.; Moore, J. S. Chem. Eur. J. 1997, 3, 765.)
down the fourfold c-axis. A packing model for this system proposed by
Moore et ul. (202) explains a relationship between the degree of the
interpretation and the bridged M----Mdistance. Their model also explains the effect of the anion size on the height of adamantoid cages
in a series of compounds [Ag(L&IX (X = PF6,AsF6, SbF6).
The longest M----Mdistance in a diamondoid framework is observed
in [Ag(L,3)21(C104)H20
(203).The trunsoid conformation of 3,3'-dicyanodiphenylacetylene, L73,bridges two Ag(1) ions separated by 17.0 A.
The obtained structure shows eightfold interpenetrated diamondoid
frameworks (Fig. 57). A majority of the large amount of void space
created in a single diamondoid framework is filled through interpenetration, resulting in an eightfold diamondoid network. The remaining
space within the lattice is filled by perchlorate ions and water. Interpenetration in this structure is mediated by T-T stacking of LT3(discussed later).
C. OTHERBRIDGING
LIGANDS
One of the simplest bridging ligands in constructing diamondoid
frameworks is a cyanide, CN-. It is well known that Zn(I1) and Cd(I1)
COPPER(1jAND SILVER(1)SUPRAMOLECULAR METAL COMPLEXES
247
FIG. 57. Diamondoid framework in [Ag(L~:3~dI(C101)H20

representing 7-n stacking
along the maxis. (From Fig. 8 in Hirsch, K. A.; Wilson, S. R.; Moore, J. S. Inorg. Chem.
1997, 36, 2960.)
ions bridged by cyanide form doubly interpenetrated diamondoid

frameworks of M(CN12(127, 204, 205). A partial replacement of the
metal site by a Cu(U ion needs incorporation of countercations due to
the charge neutrality, as exemplified in the case of NMe,[CuZn(CN),I
(127). The incorporated NMe4 cation prevents the interpenetration of
another network, resulting in the formation of a single framework,
although the framework is no longer a diamondoid but a zinc blende
structure. A detailed structural analysis concluded that carbon atoms
in cyano groups coordinate to Cu(I) centers to form Cu-CN-Zn
rods. A similar zinc blende structure is also seen in
[Cu(C(CGH4CN),IBF4
.xC6H5N02 (x = 7.7) (125, 1271, in which
C6H4CN can be regarded as a bridging ligand between Cu(1) and C"
centers. The large void space created by the bridging ligand is filled
by a counteranion and solvent molecules.
A bridging ligand having mixed functional groups also affords a
diamond-related framework, 4-Cyanopyridine, LG7,having both a cyano group in one end and a pyridine nitrogen in the other, is such a
typical asymmetric bridging ligand. Cuboidal crystals of [Ag(L6,)21BF,
(190)were obtained by slow diffusion of an ethanolic solution of
AgBF, placed over LG7dissolved in THF (molar ratio 1:2). The fourfold interpenetrated diamoadoid frameworks have a Ag----Agintraframe separation of 9.93 A. The Ag(1) ions, lying on mm special
positions, exhibit markedly flattened tetrahedral geometry, with
248
N(py)-Ag-N(py), N(cn)-Ag-N(cn), and N(py)-Ag-N(cn) angles of

127.7(5), 126.5(5), and 101.45(12Y, respectively. These local distortions result in a global compression of the diamondoi! framework
along c. The Ag-N bond lengths have values of 2.270(6) A [Ag-N(py)l
and 2.350(10) A [Ag-N(cn)]. The presence of iso-oriented anisobidentate (with donor ends differing both in steric need and in basic character) ligands gives rise for [Ag(b7)JBF4to a polar axis (c). A T-T stacking column of aromatic ring! could be observed along the c-axis
(interplanar distance of 4.0 A). Le7 also forms a diamond-related
framework with CuCN. A Cu(1) ion in [CuCN(L,,)I (206) is coordinated by one carbon atom and three different types of nitrogen atoms.
The structure was first interpreted as triply interpenetrated threedimensional polymers with one-dimensional zigzag chains of CuCN
bridged by Le7 and has recently been reappraised to have diamondrelated networks (190).
2,7-Diazapyrene (LT4),having an extended T-system, also produces
a Cu(1) coordination polymer, [Cu(L7,),1PF6(207), which shows threefold interpenetrated diamondoid frameworks. The spaces within each
diamondoid framework in [Cu(L,,),IPF, are filled by two other diamondoid frameworks, and the three interpenetrating arrays are related to each other by a 90"rotation (Fig. 58). This form of interpenetration contrasts with that observed for other diamondoid frameworks
appearing in Table V, which are related t o each other by simple translation. Adjacent frameworks interact with each other via face-to-face
T-T interactions between L7, molecules, which are arranged so that
the N----Naxes of the ligand are at 90" to one another. Because of
the greater lateral steric bulk of L,,, the number of interpenetrating
frameworks of three is smaller than other bridging ligands, giving
similar M----Mdistances (Table V). This represents the identification
of another factor controlling the degree of interpenetration in diamondoid frameworks.
Although HL63 has both a CN group and a pyridone group, a dimerization of two HL53molecules through head-to-tail hydrogen bonds
between the pyridone groups brings a symmetric rodlike bridging li~IX
= PFs, CF,SOJ (164), each
gand. In the crystals of [ C U ( H L ~ ~ )(X
HL63 moiety in the monomeric unit of [ C U ( H L ~ ~ )is~ head-to-tail
IX
hydrogen-bonded to one adjacent HL53 molecule. Thus, one [Cu(HL,,),IX
entity is connected to four adjacent others, and these repeating units
form a three-dimensional diamondoid framework. The structure exists as fourfold interpenetrated diamondoid frameworks with intraframe Cu----Cuseparations of 13.5 A in both cases. The diamondoid
frameworks in both complexes are stacked with each other in such
way that all the copper centers are found on lines parallel to the
249
FIG. 58. Threefold interpenetrated diamondoid frameworks observed in [Cu(L,,),]

PF,. (From Fig. 1 in Blake, A. J.; Champness, N. R.; Khlobystov, A. N.; Lemenovskii,
D. A,; Li, W.-S.; Schroder, M. J. Chern. Soc., Chern. Cornrnun. 1997, 1339.)
c-axis, which defines enormous linear chambers. The counteranions

PF6and CF3S03occupy channels that run along the fourfold crystallographic axes.
Not a single metal ion but a tetranuclear silver cluster plays a
role of a tetrahedral center in a diamondoid framework in
[Ag2(02C(CH2)2C02)I
(208). The compound consists of planar tetrasilver clusters bridged by succinate anions as shown in Fig. 59. The
Ag, cluster exhibits pseudo-S, symmetry and is linked though its four
ligands to four neighboring clusters, resulting in the formation of
triply interpenetrated diamondoid frameworks. This is the first example of a three-dimensional coordination polymer containing metal
clusters.
D.
T-T
INTERACTION
IN DIAMONDOID
FMEWORKS
Interpenetration in diamondoid frameworks is often mediated by

interaction between aromatic bridging ligands (Table V). A pack-
T-T
250
FIG. 59. A view of the Ag, cluster and the surrounding ligands in
(CHz)zCOz)I. (From Fig. 1in Michaelides, A.; Kiritsis, V.; Skoulika, S.; Aubry,
Chem., Znt. Ed. Engl. 1993,32, 1495.)
ing model for such systems proposed by Moore et al. (202) explains a
relationship between the degree of the interpenetration and the
bridged M----Mdistance, and also explains why two polymorphs can
be obtained in [Ag(L7,),1AsFs(202).
Face-to-face n--n--stacking interactions between Fets of L70 ligands
in [ C U ( L ~ ~ ) ~ ~ B F ~ ( O(192)
.~CH
(3.659
~ C ~and
~ ) 3.855 A) control the separations between adjacent diamondoid frameworks. The a-n--stacking
interaction is clearly important in the overall control of this extended
structure, and the presence of this interaction is along the direction
in which the long-range structure is formed during crystallization.
In case of [Cu(L&IX(L,)(THF) (X = BF4, C10,) (195, 1961,metalfree and coordinated LBB
ligands stack alternately; with a nearest carbon-to-carbon separation of 3.37(2) and 3.51(2) A, to form a one-dimensional n--n-stacking column as shown in Fig. 55.
COPPERU) AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
25 1
r-a-Stacking columns in [Ag(L7,),I(C104)H,0(203)occur along both

the a- and b-directions, with a plane-to-plane distance of 3.51 A.
Along a , adjacent ligands in the stack are flipped 180" relative to one
another such that nitrile groups are not overlaid (Fig. 57). The stacking along b is unusual in that a large offset angle causes a LT3molecule to stack between two fragments each consisting of two halves of
L7, coordinated to silver(1).To allow stacking in two directions, mutually perpendicular stacks organize in a!ternating layers.
A close interplane separation of 3.38 A between adjacent HL,, molecules in [Cu(HLC3),]X(X = PFs, CF,SO,) (164) suggests the presence
of strong r-r interactions betwten pyridone rings. This separation is
slightly shorter than the 3.40 A observed in the copper(I)-LGOcomplex (186).
VIII. Other Frameworks Based on Covalent Bonds
Apart from several of the classes of intermolecular forces based in

supramolecules described earlier, coordination polymers with extended frameworks based mainly on covalent bond forces represent
another well characterized class of transition metal supramolecules.
One might anticipate that application of the strategies that have
worked so well for the preparation of copper(1) and silver(1) coordination polymers with a diversity of novel frameworks based on intermolecular forces would allow a similar easy access to formation of the
covalent-force-based frameworks by the selection of suitable ligands
and spacers and the introduction of flexibilizing groups, either in the
polymer backbone or in the side-group structure (Fig. 60). Recent publications did reveal several such notable examples of infinite chains,
two-dimensional sheets, and three-dimensional networks formed by
metal cations linked through multidentate organic ligands.
A. INFINITE-CHAIN
STRUCTURES
1. Light-Induced Crystal Oscillation
In addition to their novel frameworks, low-dimensional metalcontaining coordination polymers are of interest with respect to their
electrical, magnetic, and optical properties, and possibly their catalytic behavior. For example, the study of photochromic compounds,
which undergo thermal irreversible and fatigue-resistant photochro-
252
L75
Me
1 b
Me-A-C-C-
-Me
HZ HZ Me
c)
L7Q
L78
181
180
L82
L86
N-
N
L89
L90
191
192
FIG.60. List of ligands in copper(1) and silver(1) complexes with linear chain, 2-D
and 3-D networks.
L99
L98
A
c:v3
253
Lg7
H3
LlOl
mic reactions, is one of the key points in the current revival of interest
in designing light-triggered molecular and supramolecular devices
(209).Recent work by Munakata and co-workers has yielded a range
of interesting and well-characterized materials with intriguing prop-
254
erties (210, 211). Reaction of copper(1) perchlorate solution with cis1,2-dicyano-1,2-bis(2,4,5-trimethyl-3-thienyl)ethene

(L75)
yielded single crystals of [ C U ( L , ~ ) ~ ] C the
~ O ~structure
,
of which is composed of
noninteracting anions and macrocations in which each metal center
is coordinated with one CN group of the four dithienylethene molecules and each L,s in turn bridges two copper(1) ions with two cyano
groups, leading to an infinite network of metal cations as illustrated
in Fig. 61(a). The most striking feature of the complex is that it shows
reversible ring-closed and ring-open transformation under different
wavelengths of light when probed by both electrical and NMR spectroscopy. This cycle can be repeated many times, indicating that the
reversible cyclization reaction takes place in the crystalline state. The
light-induced crystal oscillation is suggested to be one possible pathway to account for the approach and rotation of the two thiophene
rings as shown in Fig. 61(b). Such a system may be used as an optical memory.
2. Organometallic Polymers
Several organometallic polymers containing copper(1) and silver(1)

ions have been reported that, by definition, involve unanimously direct interaction between the carbon atoms of the ligand and the metal
ions. The organic ligands studied include alkynyls, isonitriles, and
a-conjugated systems. Addition of AgC104 to truns-[Pt(C= CH)2
(PMezPh)J gave a 1 : 1adduct of [PtAg(C104)(C=CH)2(PMe2Ph)2]
(212)
in which the trans coordination of the Pt is maintained and the Ag
atom is coordinated on the same face of each Pt coordination plane,
forming a zigzag chain with a perchlorate ion in each cavity along
the chain as shown in Fig. 62. Recently Bertrand reported the first
bis(carbene1-silver polymer [Ag(L&F3S03 obtained at -30C from
an acetonitrilelether solution of 1,2,4-triazole-3,5-diylidene
(L76) and
AgCF3S03 (213). In the cation each ligand links two linear silver
atoms with C-Ag-C bond angles of 175-180",and due to the alternation in the orientation of the five-membered rings, the complex is regarded as a one-dimensional polymer in which all the rings are coplanar.
Despite the fact that silver(1)-aromatic complexes of benzene, cyclophane, indene, acenaphthene, naphthalene, and anthracene have
been reported, the corresponding organometallic polymers of pyrene
(Led and perylene (L6g) have been crystallographically characterized
only recently (187).X-ray structure determination of the complex with
pyrene [Ag2(Lm)(ClOJ21
reveals that it exists in the solid state as an
255
COPPERW AND SILVERU) SUPRAMOLECULAR METAL COMPLEXES
405nm
550nm
twisted ring-open form
coplanar ring-closed form
Light-induced crystal oscillation
FIG.61. Schematic presentation of chain structure (a) and the light-induced crystal
oscillation model (b) in [ C U ( L , ~ ) ~ ~ C ~ O ~ .
arene-linked polymer of dimers as shown in Fig. 63. Within the dimer, there are two independent Ag(1) ions coupled by one perchlorate-oxygen bridging with Ag(l)----Ag(2)
separation of 4.39
Each
pyrene moiety exhibits a tetra-q2-coordination fashion sequentially
bridging four metal centers, resulting in a polymeric W-type sandwich
A.
256
L
L
PMezPh
FIG. 62. Structure of the linear [PtAg(C104)(C-CH)2(PMe2Ph),3.(From Fig. 3 in

Yamazaki, S.; Deeming, A,; Speel, D. M.; Hibbs, D. E.; Hursthouse, M. B.; Malik,
K.M. A. J . Chem. SOC.,Chern. Commun. 1997, 177.)
of alternating aromatic and silver(1)perchlorate groups running along

the b-axis as shown in Fig. 63(b).
Several research groups have used isonitriles to prepare organometallic polymers (214,215).End-on coordination at the terminal carbons and linear geometry at the nitrogens impart to these molecular
structures the characteristic of association of metal complexes into
di-, tri-, and even polymeric species. Harvey has recently reported a
series of new organosilver polymers of 1,8-diisocyano-p-methane
(LT7),
[Ag(L,,),]X (X = BF;, NO, or ClO;), with tubular structures in the
solid state (214). Each ligand group bridges two tetrahedral silver
atoms, forming zigzag chains that crystallize side by side separated
by two rows of the counteranions (Fig. 64). Before this study two series of organosilver polymers with 2,5-dimethyl-2,5-diisocyanohexane (L78)had been reported (215). Complexes of the first series contain one diisonitrile per metal with the formula [Ag(L78)IX(X = BF;,
PF;, or NO;). The structures contain infinite chains of silver atoms
alternating with bridging ligands. The second series crystallized as
[Ag,(L78)31X2,
and again, the extended trans-conformer of L7*alternating with Ag atoms generates infinite chains. However, the unique
structural difference lies in a adjacent pairs of chains cross-linked by
extra ligand groups, leading to a ladderlike pattern.
257
FIG.63. The W-type sandwich arrangement of an arene-linked polymer of dimers in

[Agl(L,)(C104)2].(From Fig. 1 in Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa,
M.; Suenaga, Y.; Sugimoto, K. horg. Chem. 1997,36,4903.1
258
FIG. 64. Structure of the polymeric chain in [Ag(L&]NO,. (From Fig. 1 in Fortin,
D.; Drouin, M.; Turcotte, M.; Harvey, P. D. J. Am. Chem. SOC.1997,119, 531.)
3. Heterometallic Cluster Chain

Heterometallic polymeric cluster compounds have received considerable attention due to their useful properties, which are derived from
low-dimensional structures. Haushalter and co-workers recently initiated a study designed to determine the possibility of introducing unpaired electrons to closed-shell, low-dimensional Zintl-phase materials by suitable structural modifications or elemental substitutions,
with the aim of increasing their electrical conductivity (216).Figure
65 shows the structure of one such example, (Et4N)4[A~(Agl-xA~,)2
Sn2Te9],which contains a semiconductive B-A-wide, one-dimensional
chain composed of four different elements surrounded by insulating
organic material. The magnetic susceptibility of the compound [Fig.
FIG. 65. Structure of the 1-D polymeric chain [ A U ( A ~ , _ ~ ~ ) (From

~ S ~ Fig.
~ T 1
~ ~ I ~ ~ .
in Dhingra, S. S.; Seo, D.-K.; Kowach, G . R.; Kremer, R. K.; Shreeve-Keyer, J. L.; Haushalter, R. C.; Whangbo, M.-H. Angew. Chem., Znt. Ed. Engl. 1997,36,1087.)
259
66(a)l indicates that the chain is diamagnetic in nature and is nearly

independent of temperature down to 4 K. Both resistivity data obtained by the microwave cavity perturbation technique and optical
diffuse reflectance measurements as shown in Fig. 66(b) and ( c ) respectively, support the claim that the compound is a semiconductor
and the delocalized one-dimensional band structure undergoes a
Peierls distortion above room temperature. Heterometallic polymeric
cluster compounds [W4Ag5S161[M(dmf),l
(M = Nd or La) derived from
tetrathiotunstate and silver(1) have been reported (217). The onedimensional polymeric anion can be regarded as octanuclear cyclic
cluster fragments, [WAg4SI6l4-,
linked through Ag' ions as shown in
Fig. 67. The significantly short W-Ag bond length of 2.928(1) A is
observed in the analogous compound [Ag4WS41(NH4),
suggesting that
Ag- > WS, electron delocalization and stronger continuous metalmetal interactions are present in this type of heterometallic chains.
Reaction of (NEt4),MS4(M = Mo or W) with CuCl and KSCN (or
NH4SCN)afforded a set of mixed metal-sulf'ur compounds containing
infinite anionic chains Cu,(NCS),MS:- and (CuNCS),WSP- or twodimensional polymeric dianions (CuNCS),MSi- (218). Short W----Cu
distances ranging from 2.62 to 2.70 A are observed in these compounds. Incorporation of polyoxoanions into one-dimensional coordination polymer matrices, either through direct condensation to form
oxo-bridged arrays of clusters or through transition-metal coordination compounds acting as inorganic bridging ligands, provides a route
to composite organic/inorganic materials. In this respect, two composite compounds containing Cu(1) complex ions of 4,4'-bipyridine (L51),
8H20, have recently been pre[CU(L51)4MO@26]and [CU(L~,)~MO,,O~,~
pared by intercalation of molybdenum oxide clusters into the void
space in the Cu-LSl subunits (219). By comparison, in a new mixedligand C d I ) polymer containing LS1,[(PPh3)2Cu2C12(L51)1,
a different
lattice appeared (220). This structure consists of (PPh3)2Cu2C12
units
that are bridged by L51 (Fig. 68). The significance of these studies is
that members of the vast family of polyoxoanion clusters may be employed as structural motifs to fill tunable void volumes created by
extended cationic frameworks, and it would appear that this synthetic
approach provides a method for structural modification of metal oxide
and, consequently, tuning of electronic, magnetic, and optical properties of the oxide phases.
4. Mixed-Valence Copper Complex with Chain Structure
The mixed-valence copper complexes with polymeric chain structures have been described. Reaction of the cyclic thioether ligand tet-
260
R I%
COPPER(I) AND SILVER(I) SUPRAMOLECULAR METAL COMPLEXES
261
FIG. 67. Structure of the l-D polymeric anion [W&g4Sl6l4-.(From Fig. 1 in Huang,
Q.; Wu, X.; Wang, Q.; Sheng, T.; Lu, J. Angew. Chem., Int. Ed. Engl. 1996,35,868.)
rahydrothiophene (LT9)with CuClz.2H20 in acetone yielded the polymeric, mixed-valence complex [ C U ~ ~ C U ~ ~ ( (221
L , ~).) ~The
C ~structure
~I
contains two distinct types of copper atoms: The divalent copper
atoms involve an intemediate geometry between square planar and
tetrahedral, comprising four chlorine atoms, and the monovalent copper atoms have distorted tetrahedral SzC12donor sets. The magnetic
susceptibility data, ESR, and electronic reflectance spectra for the
compound indicate that no intervalence interactions occurred between
two Cu(1) and Cu(I1) sites.
5. Diradical Ligand Complex with Chain Structure

Polymeric transition metal complexes of organic radical ligands
have received recent attention in magnetic systems. The ligands concerned include the diradical 1,1',5,5'-tetramethyl-6,6'-dioxobis(verdazyl) (LBO).
Reactions of LBO
with copper(1)halides in acetonitrile or copper(I1) halides in methanol gave copper(1) complex [Cu2X2(L8,)I(2221,
whose structure is found to be a polymeric chain with the tetrahedral
copper atoms alternately bridged by two halide ions and two verdazyl
diradicals [Fig. 69(a)l. Magnetic susceptibility measurements indicate
that the spins couple in one-dimensional chains with alternating exchange parameters [Fig. 69(b)]. Although the intraligand electronic
exchange is still antiferromagnetic in the system, the triple excited
state is considerably stabilized compared with the free ligand as a
result of an increasingly important superexchange through the copFIG.66. Physical properties of [Au(Ag,-,Au,),Sn~Te,l' . (a) Temperature dependence
of the magnetic susceptibility ,y; (b) temperature-dependent resistivity p ; (c) optical
diffuse reflectance R versus wavelength A of incident light. (From Fig. 4 in Dhingra,
S. S.; Seo, D.-K.; Kowach, G . R.; Kremer, R. K.; Shreeve-Keyer, J. L.; Haushalter, R. C.;
Whangbo, M.-H. Angew. Chem., Int. Ed. Engl. 1997,36,1087.)
262
FIG.68. View of the [(PPh,)zCu,Clz(L,,)]chain. (From Fig. 2 in Lu, J.;Crisci, G.; Niu,
T.; Jacobson, A. J. Znorg. Chem. 1997,36,5140.)
Temperature IK
FIG. 69. View of the chain structure (a) and temperature dependence of magnetic
susceptibility (b) in [CuzXz(L,)I. (From Fig. 3 and Fig. 7 in Brook, D. J. R.; Lynch, V.
Conklin, B.; Fox, M. A. J. Am. Chem. SOC.1997, 119, 5155.)
263
per(1) center. Another example of coordination of diradical with copper metal is reported by Oshio in a copper(I1) complex of imino nitroxide [CuYL8JI(PF& obtained from the reaction of [Cu1(CH3CN),I
PF6 with LSl in methanol (223), where HLBl is 2-(1-oxy-4,4,5,
5-tetramethyl -4,5 - dihydro - 1H-imidazole- 2- yl)- 6 - ( 1- oxyl- 4,4,5,
5-tetramethyl-4,5-dihydro-1H-imidazole-2-yl)pyridine.
In this
complex the square planar copper(I1) ions are bridged by the iminohydroxyamino anions to form a one-dimensional helical structure.
Here again, the magnetic susceptibility data revealed antiferromagnetic interactions. From the investigation of the coordination chemistry of these unusual diradicals it can be expected that the bridging
ligands LBOand the polypyridyl type of imino nitroxide hold great
promise as building components for future design and construction of
polyradical magnetic materials.
6. Carboxylate Ligand Complexes

Although the silver(1) ion is regarded as a typical soft Lewis acid,
it forms a variety of coordination compounds with carboxylate ligands. Mak classified the coordination modes of Ag carboxylates into
four types, most of which are constructed from either dimeric units or
polymeric networks of dimeric subunits (224). The same author prepared and characterized two polymeric silver(1) complexes of betaine,
~ A g z ~ M e ~ N C H z C O O ~ z ~ H zand
O ~ z [Agz(C~H~NCH~COO),(C10~)21
~NO~~~l
(224).The two complexes are structurally similar, both consisting of
centrosymmetric carboxylato-bridged Ag,(carboxylato-O,0)2 dimers,
which extend into a stairlike polymer through the coordination of
each metal center by a carboxylate oxygen atom from an adjacent
dimer. It is worth noting that both polymers are found to involve very
short intradimer Ag----Agdistances of 2.898(1) and 2.814(2) respectively, compared with other dinuclear silver(1) carboxylates.
A,
7. Thio-Ligand Complexes
Although both copper(1) and silver(1) metal ions are expected to
readily form coordination compounds with thio ligands by covalent
bonds, few polymeric structures for thioether crown complexes have
been reported. Among these few examples, two silver(1) complexes
with thioether macrocyclic ligands are particularly interesting (225).
Reaction of [24]aneS8(Lag)with 2 molar equivalents of AgCF3S03gave
a one-dimensional polymer, [Agz(L82)(CF3S03)z(MeCN)21,
in which
each Ag atom is coordinated to four S-donors in a distorted tetrahedral geometry. As illustrated in Fig. 70, the four S-donors come from
two different ligand molecules to generate an infinite ladder polymer
264
0
0
0
0
0
0
FIG.70. View of the polymeric chain in [Ag2(L8,)I2+.
(From Fig. 1 in Blake, A. J.; Li,
W.-S.; Lippolis, V.; Schroder, M. J . Chem. Sac., Dalton Trans. 1997,1943.)
along the b-axis. A similar reaction with the crown thioether containing four S-donors, [16]aneS4(LES),gave a three-dimensional polymer, [Ag(LEs)(BF4)1,
in which each Ag(1) ion is coordinated by four
symmetry-equivalent S-donors in a tetrahedral geometry, and each
ligand group in turn bridges four different Ag atoms. The work illustrates the potential of thioether crowns as building blocks for the synthesis of inorganic architectures using their ezo-orientated S-donors.
The rodlike ligand 2,l l-dithia[3,3lparacyclophane (Lm)is found to
be a particularly versatile organic ligand for the formation of polymeric chains in the presence of bridging anions (226, 227). The arrangement of such chains in the lattice is controlled and modified by
the metal ions and coordination of a suitable anion. Two copper(1)and
one silver(1) complex of LE4have been reported (226), in one of which,
[ C U ~ B ~ ~ ( L ~ ) ( M the
~CN
two
) ~Br
] , atoms bridge pairs of Cu(1) ions to
form a rhombic CuBrCuBr ring as shown in Fig. 71 and each ligand
molecule links two separate metal cations on each side through two
sulfur atoms, resulting in a one-dimensional polymeric chain running
parallel to a diagonal axis of the triclinic cell.
Silver(1) complexes of 1,4-thioxane (L,) have been reported (228).
Both [Ag(L,)(CF,SO,)I and [Ag2(L86)(CF3S03)21
were isolated from the
mixture of Lg5 and silver triflate at ambient temperature in a 1: 1
mixture of dichloromethane and acetonitrile. The former involves silver(1) cations tetrahedrally coordinated by two sulfur atoms of two Lgg
molecules and the oxygen atoms of two bridging triflate anions, giving
in the crystal an infinite one-dimensional chain with bridging only
through the triflate moieties. Although the immediate coordination
environment around the silver(1) ion in the latter complex is similar
t o that in the former, bridging between Ag' cations occurs via the
participation of both 1,4-thioxane sulfur and triflate oxygen atoms,
leading to a two-dimensional lattice.
COPPERU) AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
265
FIG. 71. View of the infinite chains in [CuzBrz(Llcl)(MeCN),l.

(From Fig. 1 in Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Nakagawa, S.
J. Chem. Soc., Dalton Trans. 1996,1525.)
Silver(1) complexes of the nonchelating dithioether ligand have

been described (229). Crystallographic studies of [Ag(MeSCH&H,
CH2SMe)lBF4revealed a chain structure involving trigonal (p,-S),,S
ligation with one S donor bridging two adjacent Ag cations while another S donor of the dithioether ligand is nonbridging. In contrast,
revealed monodenX-ray analysis of [Ag(PhSCH2CH2CH2SPH)21BF4
tate S ligation of four distinct dithiother ligands to the Ag atom, giving an approximately tetrahedral geometry at the metal ion, with the
second S donor of each ligand linking to four other Ag' ions, hence
generating a three-dimensional infinite lattice.
Due to the biological relevance of copper-sulfur coordination compounds to metalloproteins, the copper(1) complex of thiourea,
ICU{SC(NHCH~)~)~INO~,
has been prepared from an aqueous solution
of copper(I1) nitrate and N,"-dimethylthiourea (230).The structure
contains substantially distorted CuS, tetrahedra, each being linked
by common edges with two neighbors, resulting in infinite isolated
chains parallel to the c-axis of the cell. The nitrate anions are not
coordinated to the silver(1) ions, but associate the individual chains of
edge-sharing CuS, tetrahedra together.
8. Other Nitrogen-Containing Ligand and Mixed-Ligand Complexes
Mingos and co-workers have recently reported several novel silver(1) complexes with simple polydentate acyclic nitrogen-donor ligands such as diethylenetriamine (dien) and tris(2-aminoethy1)amine
266
(LE6)(231). The dien complex [Ag(dien)lPF, is slightly light-sensitive,

but may be stored indefinitely under nitrogen at -30C in the dark.
The complex consists of two crystallographically independent silver
atoms and two nonequivalent dien ligands, which link adjacent
T-shaped silver centers to form a catena structure. The tren complex
[Ag(LBs)]PF6
also has a zigzag chain structure and possesses similar
structural features except that the two silver(1) ions involve distorted
tetrahedral geometries and tren acts essentially as a tetradentate
ligand.
Luminescent coordination compounds continue t o attract considerable attention. Zink recently reported a new mixed-ligand copper(1)
polymer that shows interesting photoluminescence (232). The complex [CuCl(L,,)Ph,P] consists of a one-dimensional chain lattice of
metal ions bridged by both C1- ions and pyrazine molecules. The compound shows conductivity of less than lo-* S cm-'. The absorption
spectrum of the complex shows a band at 495 nm, which could be
interpreted as the promotion of an electron from the valence band to
the conduction band. On the basis of resonance Raman spectra, the
lowest excited state in the polymer is assigned to the Cu(1)-topyrazine metal-to-ligand charge-transfer excited state.
In the previous section, l-cyanoguanidine (L5,) was quoted as an
example of the multifunctional ligands for the formation of H-bonded
frameworks (157). Because this molecule can act either as a monodentate ligand (nitrile N donor) or as a bidentate bridging ligand (nitrile
and amino N donor) a number of 1 : l and 1 : 2 l-cyanoguanidinecopper(1) halide adducts have been reported (233). The complex
[CuX(L5,)I,where X = C1- or Br-, is characterized by two mutually
perpendicular chains of [CuX(L,,)], and (CuXi), joined at a common
halogen. Within these two individual chains the copper atoms are
found to involve different distorted tetrahedral and trigonal planar
stereochemistries, respectively.
Among the mixed-ligand-complexed copper(1) complexes, the one
formed by pyrazine (Lso) and 3-methylpyridazine (L8,) deserves to be
mentioned (234). The compound [cuZ(L8,)Z(LS~>~1(c104)2
was prepared
by the stepwise reaction of copper(1)with two ligands. The two copper
centers involve different coordination environments, one in a threecoordinate Y-shaped form and the other in a typical four-coordinate
tetrahedral form. The structure consists of a tetracopper unit bridged
alternately by Lg, and LDBgroups (Fig. 72). In addition, the cyclic
voltammogram of the complex reveals two reversible redox couples
with Ellz = +0.22 and +0.53, which are ascribed to the Cu'Cu'/Cu'Cu"
a
4-coordinate copper
0 3-coordinate copper
267
inUaunit prz
nmpz
FIG.72. Structure of an infinite cationic chain in [Cu2(L,,),(L,,),I(C10,),. (From Fig.

1 and Fig. 2 in Kitagawa, S.; Munakata, M.; Tanimura, T. Chern. Lett. 1991, 623.)
and CurCu"/CurrCu" reactions of the closely contacting dicopper moiety, respectively.

In a n attempt to use dicopper(1) complex unit as building block for
synthesis of metal-containing polymers, Kitagawa reported his interesting results in synthesis and structural characterization of a n infinite stair-type chain compound (235).The complex of thiochrome (L88)
[CU~(L,&I(C~O,)~
was prepared by reduction of copper(I1) perchlorate
followed by reaction with the ligand. The structure contains a dinuclear cationic unit in which the two T-shaped metal ions are doubly
bridged by two thiochrome molecules with extremely short Cu----Cu
268
A.
contact of 2.476(3) The dimeric units are mutually linked by the

2-hydroxyl group of the coordinated ligand group, giving a novel stairtype chain. The cyclic voltammogram in the solid state gives a single
quasi-reversible wave, indicating that the dicopper site undergoes redox reactions.
Coordination compounds of biheteroaromatic ligands have been extensively studied (236). The nitrogen-rich ligand, 3,5-bis(2-pyridyl)
pyrazole (Lee),in which the pyrazole ring is attached to bulk nitrogen
heterocycles a t the 3-and 5-positions, belongs to this class. Polymeric
silver(I) complexes containing the uninegative pyrazolate ion had
been known for over a century when the formation of an insoluble
silver pyrazolate salt Ag(pz) was reported; however, its structure has
never been crystallographically established (237). Fortunately, the
structure of the polymeric silver(1) complex of the pyrazole derivative
ligand Leuhas been determined recently (238).Along with it, Munakata has also reported a copper(I1) and another silver(1) complex with
Lsa. The reaction of Leuand AgC10, in different solvents at various
molar ratios surprisingly leads to the isolation of two totally different
- 2MezC0, and one
silver(1) complexes, one dimer, [Ag2(L8u)41(C104)z
polymer, [Ag(L,,)]ClO,. In the polymeric structure the coordination
around the metal ion involves a distorted trigonal planar core, AgN3,
and each LSl molecule bridges two metal centers, giving an extended
zigzag chain as shown in Fig. 73. Between the chains there is no effective interaction present due t o the presence of anion column separating the cationic chains.
Like Lee,hexamethylenetetraamine (LW)is also a potentially tetradentate ligand. Two silver(1) coordination polymers with this polydentate N-donor base have been reported, and one involves a one-dimensional lattice (239). Reaction of AgSbF, and Luoin a 1: 1 molar ratio
in ethanol-CHuClz isolated the compound [Ag(Luo)lSbF,.HzO. The
structure contains a novel type of ribbon formed by hexagonal meshes
of alternate Ag' ions and Lw,molecules. The water molecules coordinated to the silver ions form O-H----N hydrogen bonds with the bo
groups of adjacent ribbons, which generate a three-dimensional network of unprecedented topology (Fig. 74).
In terms of Pearson's hard-soft acid-base principle, the soft Lewis
acid copper(1)ion is not compatible with the hard base HzO molecule,
but a recent article provided the first example of a copper(1)-water
bond (240).The novel copper(1) complex with 2,3-diphenylquinoxaline
(Ld, [Cu(LUJH~O)IC10,,is diamagnetic and indefinitely stable in air
in the solid state. Its structure consists of an infinite polymeric cationic chain in which adjacent metal centers are bridged by the aro-
COPPERU) AND SILVERII) SUPRAMOLECULAR METAL COMPLEXES
269
FIG.73. View of the chain structure in [Ag(L,)]-. The dashed line shows the contact
between the nearest silver atoms, Ag----Ag5.67 A. (From Fig. 3 in Munakata, M.; Wu,
L. P.; Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, M.; Kawata, S.; Kitagawa, S.
J . Chem. SOC.,Dalton Trans. 1995, 4099.)
matic nitroeens of the bidentate ligand of bl.The Cu-0 bond length

of 2.154(6) A is longer than the calculated distance of 1.99 A based on
the bond-valence-sum models. Therefore, the coordination of the copper is best described as distorted trigonal planar CuNIO with a weak
interaction with the water molecule.
FIG. 74. One-dimensional ribbon in [Ag(L,,)lSbFs. H,O. (From Fig. 2 in Bertelli, M.;
Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. J. Muter. Chem. 1997, 7, 1271.)
270
Several copper(1) and silver(1) complexes with trans-1,2-bis(2-pyridy1)ethylene (Lo,) have been reported (241). In [M(L,dIX, where M =
Cu or Ag and X = PF, or ClO;, a distorted linear coordination geometry completed by the nitrogen atoms of two different ligands is observed for both Cu' and Ag' ions. Both complexes have a similar chain
structure, and they only differ in the mode of polymerization, being of
rectangular and triangular wave chain types, respectively (Fig. 75).
Additionally, the pyridine ring contact between the adjacent chains
in the copper complex is 3.46 indicative of r-r interactions present.
In another copper(1) complex, [Cu(Lo,)(CO)(CH,CN)]PF,,an infinite
A,
i
(b)
FIG. 75. ORTEP views of the cationic chains in [Cu(L,)IPF6 (a) and tAg(L)1C1O4
(b). (From Fig. 1 in Powell, J.; Horvath, M. J.; Lough, A. J. Chern. SOC.,Dalton Trans.
1996,1669.)
271
one-dimensional chain based on bridging of Lg2between two metal

ions is also observed. However, with additional coordination of carbony1 and acetonitrile, each metal center involves a tetrahedral core
CUN~N'C.
Coordination compounds containing simple alkyl and aryl halides
have been structurally characterized (242, 243). Novel examples include silver(1)-iodocarbon complexes obtained from the reaction of
AgPFs and AgN03 with diiodomethane, 1,3-diiodopropane, and simple
aryl iodides (243).The compound [Ag(NO,)(l,2-12C6H4)lis composed of
spiral chains of [Ag(p-N03)I, as shown in Fig. 76, in which the nitrate
uses only one oxygen atom to bridge two silver ions. In [Ag{I(CH2)31}al
PF6,[Ag(ICH21)21PF6,
[Ag(l,2-I2C6H4),lPF6,
and [Ag(l,2-BrIC6H4)41PF~,
the structure consists of tetrahedrally coordinated Ag' ions and bridging iodocarbon ligands, giving rise to a chain polymer array.
STRUCTURES
B. TWO-DIMENSIONAL
1 . Polycatenane and Polyrotaxane Complexes
Catenanes are a group of interlocked or catenated molecules, and

rotaxanes are a class of molecules in which a dumbbell-shaped compo-
FIG.76. Structure of the spiral chains in IAg(N03)(1,2-I,C,H,)].(From Fig. 8 in Powell, J.; Horvath, M. J.; Lough, A. J. Chem. SOC.,Dalton Trans. 1996, 1669.)
272
nent is encircled by a macrocycle. The efficient synthesis of such systems fascinated and inspired many scientists in the 1960s and 1970s
(244). Due to the introduction of metal templating and supramolecular approaches for molecular threading and interlocking, this area has
experienced a renaissance in recent years (245,246).Multicomponent
polycatenane and polyrotaxane molecular systems incorporating
metal-ion templates represent a particularly interesting class of supramolecular species that might exhibit intriguing chemical topology
and display some fascinating electronic, optical, and magnetic properties (247-251 ).
Schroder and co-workers have successfully isolated a unique polycatenated undulating molecular ladder that forms interwoven twodimensional sheets (252). The reaction of [CU(M~CN)~]PF~
with 1,4bis(4-pyridy1)butadiyne (Lg3)in MeCN-CH2C12 yielded the complex
[ C U ~ ( M ~ C N ) ~ ( L ~ ~ )The
~ ] (compound
P F ~ ) ~ . exists as a network of molecular ladders in which the two independent Cu' centers are each coordinated in a tetrahedral geometry t o three LBsgroups and one MeCN
molecule. The lattices are polycatenated to give a remarkable twodimensional layer structure (Fig. 77). The sheets of interwoven molecules are separated by PF, counteranions and solvent molecules. The
structure is further stabilized by the 7r-v interactions between adjacent, symmetry-related ladders.
Robson has constructed a two-dimensional polyrotaxane sheet from
the rodlike ligand 1,4-bis(imidazol-l-yl-methyl)benzene(I&,) (253).
The compound [Ag2(L,)3(N03)21
was obtained by reaction of LB4and
silver nitrate in aqueous methanol. The coordination polymer consists
of one-dimensional chains in each of which the metal center is located
&
-l
L93
FIG. 77. View of the polycatenated sheet structure in [Cu2(MeCN),(L,),I(PF,), .

(From Fig. 3 in Blake, A. J.; Champness, N. R.; Khlobystov, A.; Lemenovskii, D. A.; Li,
W.-S.; Schroder, M. J . Chem. SOC.,Chem. Cornmun. 1997, 2027.)
273
within the plane of the N3 donor set. The individual chains are associated t o produce the polyrotaxane sheets as shown in Fig. 78, in which
the LDpmoiety provides both the rodlike segments and two half-loops
for each of the rings.
K m and Whang have constructed a polyrotaxane containing cyclic
beads threaded on two-dimensional coordination polymer networks
(254). The compound [Agz(L9s)a~cucurbituril)31(N0,),~
40H20, where
LDs = N,N-bis(4-pyridylmethyl)-1,4-diaminobutane
dihydronitrate
and cucurbituril = C36H36N24012,
was prepared by the route shown in
Fig. 79. The compound is a polyrotaxane in which the cyclic beads of
cucurbituril are threaded on a two-dimensional coordination polymer
network. The two-dimensional polyrotaxane network forms layers
stacked on each other, which in turn fully interlock with themselves.
2. Two-Dimensional Network Comprising Copper Clusters
Copper(1) thiolate complexes show remarkable diversity of structure, and the most common structural unit in the known thiolate is a
three-coordinate copper atom bonded to p2-bridging thiolates (15).
Parish has recently reported a novel compound in which the thiolate
sulfur atom involves unusual four-way bridging (2551. The compound
with composition [ C U , ~ C ~ ~ ~HzO
( S R(R
) ~=I CHzCHzNH3)
~
was obtained
by reaction of CuCl and cysteamine hydrochloride in an aqueous solu-
FIG.78. A 2-D polyrotaxane network in [Ag,(L94),3(N03)21.

(From Insertion Fig. V and
Fig. 2 in Hoskins, B. F.; Robson, R.; Slizys, D. A. J.Am. Chem. SOC.1997, 119, 2952.)
274
FIG.79. Construction scheme for the 2-D polyrotaxane network of [Ag,(L&(cucurbi40Hz0.

t~rilhI(N0~
) ~ . (From Scheme 1in Whang, D.; Kim, K. J. Am. Chern. SOC.1997,
119, 451.)
tion. X-ray analysis revealed a polymeric structure containing a centrosymmetric Culz cubo-octahedron unit. A regular octahedron of sulfur atoms, bearing organic chains pointing outward, has each edge
bridged by a CuCl unit. There are thus three mutually perpendicular,
planar, eight-membered [(CuCl)(p2-SR)],rings intersecting at the sulfur atoms, giving a gimba-like Atlas-sphere configuration (Fig. 80).
The Cu12C112(SR)6
clusters are further linked to form two-dimensional
sheets with different linkages in the two directions.
3. Two-Dimensional Network Containing Hexagonal Meshes and a
Zeolite-like Framework
Ciani and co-workers have reported the self-assembly of two remarkable polymeric networks based on an anionic acetonyl derivative
of tetracyanoethylene (256).The attempts to obtain novel extended
frameworks of Ag ions bridged by the neutral tetracyanoethylene
275
FIG.80. Structure of the basic [Cu13C113(SR)61

unit and the layers of clusters forming
sheets. (From Fig. 1 in Parish, R. V.; Salehi, Z.; Pritchard, R. G. Angeru. Chern., Int.
Ed. Engl. 1997, 36, 251.)
have led to the unexpected formation of [Ag(L,)I, where L, = 1,1,2,2tetracyanopentan-4-one-1-ide. The structure consists of twodimensional undulated nets of hexagonal meshes formed by alternate
three-coordinate silver ions and tridentate anion LBs(Fig. 81). Two
independent nets of this type interpenetrate to give a layer structure.
The targeted synthesis of [ C U ( M ~ C N ) ~ ]with
P F ~the preformed conju-
276
FIG.81. View of the twofold interpenetrated layer of [Ag(L,)l. (From Fig. 1 in Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. Angew. Chem., Znt. Ed. Engl. 1996,
35, 1088.)
gated acid of LBBin acetone gave the corresponding copper(1) polymer

[Cu(LBB)I.
Me2C0.Although the compound holds the same stoichiometry as the silver complex, its structure consists of a three-dimensional
neutral network of four-coordinate copper(1)ions linked by anions LW.
The framework resembles a honeycomb with large channels that include guest acetone molecules. The large difference between the two
structures can be related t o the higher tendency of Cu' ions over Ag'
ions to achieve a tetrahedral coordination. Therefore, by using this
methodology it should be possible to generate other topologies with
different metal ions and/or with similar polycyano anions.
The assembly of a long-range solid-state superstructure with a zeolite-like framework has been exploited by a number of research
groups (128,227).Recently, it has proven possible to synthesize coordination networks with microporous structures that maintain crystal
integrity upon loss of guest species. Such zeolite-like behavior of a
coordination network represents a new challenge for chemists to look
for new class of microporous substances. Reaction of silver triflate
and the tritopic ligand 1,3,5-tris(3-ethynylbenzonitrile)benzene(Lu,)
yielded solid [Ag(L,,)(CF,SO,)]. 2C6H6 (257).The compound contains a
two-dimensional network in which the coordination geometry around
silver is trigonal pyramidal, with three nitriles of the network in the
basal plane and a trilate counterion in the apical position as shown
in Fig. 82. The sheets are stacked, creating a channel structure in
which solvent benzene molecules reside in the micropores. The com-

A
277
FIG. 82. Structure and coordination network in [Ag(L,,)(CF,SO,)I.(From Fig. 1 in

Venkataraman, D.; Gardner, G. B.; Lee, S.; Moore, J. S. J.Am. Chem. SOC.1996,117,
11600.)
pound shows stability toward partial loss of inclusion molecules. The

TGA trace of the compound reveals two discrete mass losses at 110
and 145C (Fig. 431, corresponding to the removal of distorted benzene
molecules. At 145C, a solid-to-solid phase transition occurs concomitant with the loss of the remaining benzene molecules. DSC and optical microscopy indicate no evidence of a phase change associated with
the first mass loss. Further observations demonstrate that the sample
can reversibly remove and reabsorb benzene molecules without collapse of the channel network analogous to that of zeolite-type materials.
4. Two-Dimensional Network Containing an Undulated Wave
Layer Structure
In the Section VIILA, Infinite-Chain Structures, we noted that
Mak et al. characterized a number of carboxylato-bridged silver(1)
complexes with chain structures (224). The same research group has
278
temperature ("C)
FIG.83. Overlay of TGA and DSC traces for [Ag(L,,)(CF,SO,)]. (From Fig. 2 in Venkataraman, D.; Gardner, G. B.; Lee, S.; Moore, J. S. J.Am. Chem. SOC.1996,117, 11600.)
also employed dicarboxylate-like ligands rneso-2,5-bis(trimethylammonio)adipate (Log)and rneso-2,5-bis(pyridinio)adipate(L,) to synthesize silver(1) complexes with two-dimensional networks (258).Structure analysis revealed that in each of the four complexes
[Agp(LQe)l(C104)z,
[Agz(LQ9~I(C104)z,
[Agz(L~g)l(NO,),,and [Ag~(Lw)l
(NO,), , a dimeric unit is involved in which each pair of adjacent metal
atoms is doubly bridged by coplanar syn-syn p-carboxylate-0,O'
groups. The dimeric subunits are extended into a step polymer
through the linkage of each metal center to a carboxylate group of an
adjacent dimer. The basic coordination environment for the metal is
completed by three carboxylato oxygen atoms in a T-shaped geometry,
if the interaction between the anion and the metal center is ignored,
as occurred in two perchlorate complexes. Because the nitrate anion
functions as a unidentate and 0,O'-bridging mode in the third and
the last complex, respectively, it leads correspondingly to a distorted
tetrahedral and unusual square pyramidal five-coordination at Ag'
ion. Figure 84 shows the structure and packing drawing of the layer
structure of the five-coordinate silver complex. All compounds exhibit
short Ag----Agseparation in the dimeric unit, ranging from 2.794(1) A
to 2.878(2) A, substantially smaller than that in metallic silver
(2.89A). It is debatable whether these distances are a result of the
bridging ligand or suggestive of strong metal-metal interactions
present.
Crown thioether can not only form a one-dimensional chain structure, it can also form two-dimensional sheet frameworks (225).The
crown thioether ligand (OH),[lGlaneS, (Llm)reacted respectively with
silver nitrate and silver acetate, giving two polymeric silver(1) com-
279
FIG.84. Structure skeleton of the asymmetric unit (a) and packing drawing of the
layer structure (b) in [Ag2(L,)(NOa)pl.(From Fig. 5 in Wu, D.-D.; Mak, T. C. W.
J. C h e n . Soc., Dalton Trans. 1995, 2671.)
plexes, [Ag(Lloo)lOzCMe
and [Ag(Lloo)lN03
(259).The two compounds
are not isostructural, but their structures and dimensionality seem to
be controlled by change of the anions. In the complex with silver acetate, each metal center is bound to four thioether sulfur atoms in a
rather distorted tetrahedral fashion. The silver atoms are in a twodimensional sheet arrangement interconnected by the macrocycle in
which two crystallographically independent Ag' ions are stacked alternately along a diagonal to the a- and c-axes of the unit cell (Fig.
85). By contrast, the complex with silver nitrate consists of a three-
280
-C
FIG.85. Packing diagram for [Ag(LlOo)lO2CMe

with dashed lines showing the hydrogen bonds. (From Fig. 4 in Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.
J. Chern. SOC.,Dalton Trans. 1995,3215.)
dimensional tetrahedral network linked by an exodentate coordination of the thioether.
5. Conductive r-Complex
Although the silver(1)-pyrene complex [Agz(Lee)(CIO&]is found to
display a chain structure as already mentioned, a similar complex
with perylene, [Agz(L6s)(C104)z],
consists of a two-dimensional framework of the metal ions bridged by the bidentate counterions and the
tetra-$-arene groups shown in Fig. 50 (187). Such two-dimensional
sheets are further connected by the superpose! intersheet aromatic
rr-r stackings at an average distance of 3.31 A, generating a threedimensional network structure. The physicochemical property measurements of two complexes show that a t room temperature no strong
ESR spectrum was observed for each case, but the light-irradiated
samples show a characteristic g value attributable to the aromatic
radicals, suggesting that upon irradiation electron transfer partially
takes place from the aromatic donor to the silver (I) ion, giving an
organic radical cation and silver(0). In addition, both compounds are
electrically nonconducting, whereas their Iz-doped samples display
semiconducting behavior at ambient temperature, which is presumed
COPPER^ AND SILVER^ SUPFLAMOLECULAR METAL COMPLEXES
281
to connect mainly with the nonintegral formal oxidation state of the

crystallized organic radicals. These conducting radical cation complexes might constitute new functional materials among lowdimensional molecular solids similar to polymetallocenes and multidecker metallacarboranes. A further example of a silver(1) n-complex
with a two-dimensional sheet framework is found in a silver(&
(243). The solid-state
iodocarbon complex, [Ag2(02PF2),(p-IC6H4Me)l
structure of the compound is composed of an infinite two-dimensional
sheet array of [Ag(O,PFz)l, in which each p-iodotoluene bonds to two
Ag' ions via one I-Ag bond and an $-arene bond involving two of the
carbons not bonded to I or Me.
6. Inorganic Polymers
The development of inorganic polymer chemistry based on incorporation of transition metals into a polymer main chain has resulted in
diverse arrays of interesting structures in recent years. These polymers are expected to constitute a new type of inorganic functional
material. It is not the aim in this section to undertake a comprehensive treatment of this field; for this, the reader is directed to the recent reviews and the references herein (260,261).Here we just list a
few examples of inorganic polymers that incorporate copper(1) or silver(1) ions with two-dimensional sheet frameworks. Pfitzner has carried out a systematic investigation on new adducts of copper(1) halides with neutral chalcogen fragments or polychalcogenide anions
(262).One of them, (CuI)&uzTeSy,was found to contain layers of the
complex thioanion [TeSJ- embedded between layers of iodide ions.
Copper atoms are distributed in the arrangement of iodine and sulfur atoms.
Kanatzidis and Zhang have utilized sulfur-rich mixed polysulfide/
telluride fluxes in combination with transition metals to synthesize
novel solid-state materials (263). Thus, six new quaternary copper(1)
and silver(1) compounds with composition AMTeS3,where A = K, Rb,
or Cs and M = Cu or Ag, were prepared. The structure consists of
anionic [MTeSJt- layers and charge-compensating alkali ions between the layers. Each layer is composed of tetrahedrally coordinated
Cu' or Ag' centers and trigonal pyramidal TeSi- units, joined together
via bridging S atoms. Similar low-dimensional quaternary compounds, KCu,AsS3 and KCu4AsS4,have also been reported, and a
unique layered framework was observed in both compounds in which
the Cu(1) ions are linked in a complex manner by a series of trigonal
ASS:- groups or ASS:- groups as well as S2-ions (264).
282
7. Inorganic Grid and Heterobimetallic Aggregates

Inspired by the idea that an n-topic ligand would form an n X n
inorganic grid based on n2metal ions and 2n ligand components, Lehn
and co-workers have reported the self-assembly of the inorganic 3 X
3 grid consisting of an array of nine silver(1) metal ions and six tritopic ligands (265). The complex [Ag,(Llol)sl(CFBSO~)9,
where Llol =
6,6-bis[2-(6-methylpyridyl)]-3,3-bipyridazine,
contains an arrangement of a 3 X 3 grid of nine Ag ions chelated by six bidentate ligand
molecules as shown in Fig. 86. Such an architecture opens the way to
a whole family of polynuclear inorganic m X n grids, which represent
a new facet in the controlled arrangement of metal ions into specific
arrays and patterns.
Among the previously discussed heterobimetallic aggregates of copper(1) with thiotungstates and thiomolybdates, those containing the
two-dimensional polymeric dianions (CuNCS),MSq- are noteworthy
(218). X-ray diffraction results show that four edges of the tetrahedral MSq- (M = Mo or W) core are coordinated by copper atoms
forming WS4Cu4 aggregates linked by eight-membered rings of
- CU(NCS)~CU-(Fig. 87).
8. Two-Dimensional Sheets Containing Cu6Hexagons
The rodlike ligand La was also found to form a polymeric complex

with copper(1) (2261, but the complex obtained by reaction of La with
1 molar equivalent of CuI in thf, [Cu2I2(La),1.thf, is surprisingly not
isostructural of the corresponding complex of the bromide. Instead, it
FIG. 86. Crystal structure of the self-assembled 3 X 3 inorganic grid [Ags(L,,,),]

(CF,SO,),. (From Fig. 3 in Baxter, P. N. W.; Lehn, J.-M.; Fischer, J.: Youinou, M.-T.
Angew. Chem., Znt. Ed. Engl. 1994,33,2284.)
283
FIG. 87. Perspective view of zig-zag anionic chains in (CuNCS),MS; . (From Fig. 2
in Manoli, J. M.; Potvin, C.; Secheresse, F.; Marzak, S. Inorg. Chem. Actu 1988, 150,
257.)
is composed of a two-dimensional framework in which a large ring

containing six copper atoms includes a thf molecule as guest inside
the open cavity (Fig. 88). As in the bromide, the rhombic CuICuI ring
formed by bridging the two iodide ioas between pairs of copper atoms
leads to a Cu----Cudistance of 3.18 A and an I----Idistance of 4.30 A.
Each L, fragment bridges two separate metal cations through two
sulfur donors, resulting in a two-dimensional sheet arrangement of
copper(1) ions. The framework contains six-membered rings of Cu6
hexagons incorporating the thf molecule in the open cavity. In contrast to the formation of the rhombic ring by copper(1)halide ions, the
one-nitrate oxygen bridging two metal ions in the silver(1) complex
[Ag(L,,)(NO,)l yielded a unique three-dimensional channel framework
of silver(1) ions (226). For these reasons, both metal ions and counteranions are considered to be the fundamental factors controlling
crystallization and variation of frameworks.
STRUCTURES
C. THREE-DIMENSIONAL
1. Construction of Channeled Frameworks
The rational design of multidimensional coordination compounds

incorporating large cavities or channels capable of hosting small mol-
2 84
(b)
FIG.88. Molecular packing diagram of complex [Cu212(L&I. thf (a) and a view of the
2-D sheet consisting of Cue hexagons (b) where only the metal centers are presented as
open circles. (From Fig. 3 in Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.;
Suenaga, Y.; Nakagawa, S. J. Chen. Soc., Dalton Trans. 1996, 1525.)
ecules is currently recognized as one of the most important issues

in the synthesis of functional materials for molecular sieves, shapeselective catalysis, ion exchangers, and many other applications ( 4 ,
266-270). The approach generally employs rigid and highly symmetric organic units as building blocks via intermolecular forces such as
COPPER^ AND SILVER^ SUPRAMOLECULAR METAL COMPLEXES
285
hydrogen bonds (159-163) or coordination to metals. Such zeolite-like

behavior of the coordination network in copper(1) and silver(1) complexes is not limited to two-dimensional structures (257). In fact, the
majority of examples are found in three-dimensional copper and silver
complexes of nitrogen ligands derived from organocyanides (121-128,
271 1, 4,4-bipyridyl, and hexamethylenetetraamine.
The simple ligand 4,4-bipyridyl
has been chosen by many
groups as a rod linking together metal centers to give extended solids
with diverse topologies (272-275). Reaction of AgNO, and L51 in
MeCN gave the silver(1) polymer [Ag(L,,)INO, (273,274j. The crystal
structure reveals a triple interpenetrated cationic network schematically shown in Fig. 89, each consisting of linear [Ag(L51)lru
chains
cross-linked by Ag-Ag bonds of 2.970(2)A. The resultant open framework contains large interconnected cavities that form stairlike microchannels of approximate cross section 6 X 23 A. Rather small channels are observed in the corresponding complex of copper(1) chloride,
[Cu(LS1)C1],
obtained by diffusing solutions of CuCl and L5,into ethylene glycol (275). The structure contains chloro-bridged Cu dimers,
which form a sheet framework with 4,4-bipyridyl molecules. The
overall structure thus formed is a neutral three-dimensional framework, in which the interpenetrating two-dimensional sheets fill yp
most of the pore space with channels of only small diameter (2 X 4 A)
remaining open, too small to accomodate acetonitrile or DMSO solvent molecules as guests in the crystal.
Paralleling the approach used in supramolecular organic chemistry
of adamantyl templates with different synthons (266-268), Ciani and
FIG.89. Schematic representation of the 3-D framework in [Ag(Lsl)lNO:,,with circles

representing Ag cations, filled bars representing L51 and empty bars representing
Ag-Ag bonds. (From Scheme 1-D in Robinson, F.; Zaworotko, M. J . Chem. Soc. Cheni.
Commun. 1995, 2413.)
286
co-workers have developed a scheme using the potentially tetradentate ligand hexamethylenetetraamine (L,) for the preparation of supertetrahedral networks with metallic synthons (239,276-278). They
employed reaction of Lm with different silver(1) salts of noncoordinating anions and isolated a number of polymeric silver complexes with
three-dimensional frameworks. The first of this series, [Ag(L,)I
PF6.HzO,was reported in 1995 (276). The compound was obtained by
slow evaporation of an ethanolic solution of AgPF6 layered on a solution of L, in CHzClz.The structure contains a molecular-based framework topologically related t o the three-dimensional, three-connected
cubic net of highest symmetry as schematically shown in Fig. 90. The
PF; anions and the guest water molecules occupy the large octagonal
channels and form an extended hydrogen-bonding network. Similar
reactions under different conditions afforded three other compounds.
The structure of [Agll(L,)6](PF6)l,.14Hz0 consists of an open threedimensional cationic frame, with all the Lm molecules acting as tetradentate ligands while nine silver cations are biconnected and two are
triconnected to the ligands per formula unit (239). The (3,4)-connected
network schematically shown in Fig. 91 is composed of triconnected
(silver ions) and tetraconnected (L,) centers in the ratio 1: 3. Such
connection generates two types of parallel channels to host the anions
and the water molecules. The compound [Ag5(LW))6](PF6)5.
3CHzC12is
- 3EtOH gives a threean oligomer, whereas [Ag,(Lso)(HzO)l(PF6)4
dimensional network with large cavities and channels of the hexagonal sections, including anions and solvent molecules as guest species
(277).
in a different molar ratio yielded two
Reaction of AgC104 and LBO
polymeric complexes (278). The crystal structure of [Ag(L,)IClO, consists of two-dimensional infinite layers of hexagonal meshes formed
by alternate triconnected silver ions and L, molecules. The structure
FIG.90. A schematic view of the 3-D network in [Ag(L,,)IPF,. HzO. (From Chart 1 in
Carlucci, L.; Ciani, G.; Proserpio,D. M.; Sironi, A. J. Am. Chem. Soc., 1995,117, 12861.)
287
FIG. 91. A schematic view of the 3-D network in [Agll(Leo)sl(PFs)ll~

14H20. (From
Fig. 7 in Bertelli, M.; Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. J. Muter. Chem.
1997, 7, 1271.)
of [Ag3(Lso)21(C104)3.
2H20 contains an open three-dimensional cationic network formed by [Ag(Lso)]hexagonal layers, which are joined
by biconnected silver ions (Fig. 92). Thus, the three-dimensional
framework is a (3,4)-connected net comprising triconnected and tetraconnected centers in the ratio 1: 1. Guest water molecules occupy the
large hexagonal channels and interact with the silver ions of the layers. Thermal analysis has shown that these water molecules can be
reversibly removed from the crystals by thermal activation.
Another highly symmetrical coordination polymer was formed in
the reaction of [CU(CH,CN)~]C~O,
and 2,4,6-tri(4-pyridyl)-1,3,5-triazine (Lloz).The compound formulated as
has a cu-
FIG.92. A schematic view of the 3-D network in [Ag3(Ld21(C10,)32Hz0. (From Fig.

2 in Carlucci, L.; Ciani, G.; v. Gudenberg, D. W.; Proserpio, D. M.; Sironi, A. J . Chem.
SOC.,Chem. Commun. 1997, 631.)
288
bic (3,4)-connectednet with large cavities (279).All the ligand groups

are equivalent and are attached to three copper centers at the corners
of an equilateral triangle, and each copper ion is coordinated by four
pyridine donors in a distorted tetrahedral arrangement. This results
in an infinite (3,4)-connected three-dimensional network containing
large octahedral chambers of diameter 18.241(4) A. Each chamber is
connected by its six copper vertices to six others whose centers are
arranged octahedrally around, generating an infinite cubic collection
of chambers (Fig. 93).
Reid has extended the investigation of channeled frameworks to the
methylene-bridged bidentate ligands MeECHzEMe (E = S, Se, or Te)
on the expectation that such ligands would disfavor chelation and
might therefore promote formation of ordered extended networks
(280).X-ray diffraction studies show that in [ C U ( M ~ S ~ C H ~ S ~ M ~ ) ~ ] P F
each copper ion is tetrahedrally coordinated to one Se donor of four
diselenoether ligands and the second Se donor of each ligand is
ligated to a different adjacent Cu' ion to give an infinite threedimensional cationic network (Fig. 94). The network contains cylindrical channels of diameter ca. 12.5 A incorporating the PF, counteranions. The silver(1) complex [Ag,(MeSeCHpSeMe)*l(BF,),contains
a similar three-dimensional lattice with rectangular-shaped channels
of cross section ca. 12.6 X 4.1 A to fill in the anions.
A large number of host-guest structures are known, but a reversible guest exchange or chemical transformation of the guest inside a
coordination polymer framework has been described for only a few
(257, 281-283). In most cases the integrity of the inclusion lattice
cannot be effectively maintained in the absence of guest molecules
L102
FIG. 93. Schematic representation of two interpenetrating, identical frameworks in

[ C U ~ ( L ~ ~ ~ ) ~ I(From
( C ~ OFig.
~ ) 3~ .in Abrahams, B. F.; Batten, S. R.; Hamit, H.; Hoskins,
B. F.; Robson, R. Angew. Chern., Znt. Ed. Engl. 1996,35, 1690.)
COPPERU) AND SILVERU) SUPRAMOLECULARMETAL COMPLEXES
289
FIG. 94. View of the 3-D network in [ C U ( M ~ S ~ C H ~ S ~ M(From

~)~]P
Fig.
F ~1. in Black,
J. R.: Champness, N. R.: Levason, W.; Reid, G. Inorg. Chem. 1996,35,4432.)
(166).By using the trigonal ligands, 1,3,5-tricyanobenzene (L103) and

1,3,5-tris(4-ethynylbenzonitrile)benzene (LlO4),to make networks
joined at the vertices by metal ions, Moore and Lee have obtained
two silver(1) complexes [Ag(Llo3)(CF,S03)I
and [Ag(Llo4)(CF3S03)I
by
mixing the constituent molecules in benzene (281, 282). The crystal
structure of the compound with L103 consists of honeycomb sheets
based on alternating LIo3and the trigonalaAg units. The honeycomb
sheets create cavities of diameter 10.03 A to accommodate CF3SOi
ions which are weakly bound to the silver at the axial position of a
trigonal pyramid. By contrast, the complex with tbe larger ligand L104
generates larger channels of diameter 15 X 22 A and consequently
involves interpenetration of lattices (Fig. 95). The three-dimensional
(3,3)-connected net is based on the end-on coordination of L104 to the
trigonal pyramidal silver cations. The interpretation of the networks
generates hexagonal channels in which the benzene molecules from
the crystallization process are located. Investigation of exchange properties within a host crystal were carried out in solution and in vapor
with a variety of guest molecules. The process was monitored by powder X-ray diffraction and thermogravimetric analyse (TGAYdifferential scanning calorimetry (DSC) studies, and the results show that
the host adduct can be repeatedly voided of and filled with guest molecules without the destruction of the channel framework.
2. Feldspar Structure Without Void Space
The term mineralomimetic chemistry was proposed to denote the
chemistry of the build-up of mineral-like structures using materials
290
FIG.95. View of [Ag(L,,)(CF,S03)1down the a-axis with plane waves corresponding

to the 002 reflection highlighted. Triflate anions are omitted. (From Fig. lc in Gardner,
G. B.; Kiang, Y.-H.; Lee, S.;Asgaonkar, A.; Venkataraman, D. J. Am. Chem. SOC.1996,
118, 6946.)
that never give stable minerals in nature (270). The topologic similarities observed between the well-defined coordination geometries and
in natural minerals prompted many chemists to mimic the structures
of simple minerals (121). The copper(1) complex of pyrimidine (LlOs),
[Cu(LlO,),IBF5,was found to have a three-dimensional structure related to that of the natural mineral feldspar (284). The complex, obtained by reaction of [Cu(CH3CN)JBF4and pyrimidine, contains two
independent copper centers in the unit cell, each coordinated in a
slightly distorted tetrahedral geometry by four nitrogen atoms from
four different ligand molecules. Each Llosin turn bridges two copper
ions, leading to a three-dimensional framework. The structure is related to that of feldspar in such that the corner-shared network consists of eight-rings of tetrahedral connected together with four-rings
to form elliptical channels (Fig. 96). Due to interpenetration of neighboring layers, such channels are almost completely constricted, leaving no void space in the structure. This phenomenon has been previously described in [Cu(Lsl)C1],where the interpenetrating of the
two-dimensional sheets resulted in small channels (275). Thus, it
would be possible to increase the volume of the channels by lengthening the spacer ligands that connect the metal centers.
291
FIG.96. Polyhedral packing diagram of [Cu(L,,,),IBF, (a) and schematic view of the
feldspar structure viewed down the a-axis (b). (From Fig. 2 in Keller, S. W. Angew.
Chem., Znt. Ed. Engl. 1997,36,247.)
3. Channel Structure Through Siluer(l)-Promoted in Situ

Ligand Synthesis
Successful construction of multidimensional frameworks largely relies on ligand design to suit different geometries and coordination
numbers of the metal ions. A recent report on isolation of an unexpected three-dimensional metal complex by in situ cyclization of the
ligand aroused tremendous interest among ligand designers (285).
Reaction of 1,2-trans-(4-pyridyl)ethene(L7J and AgBF, in MeCN and
CHzClzin the presence of light yielded [Ag(Llw)IBF4,where Llw =
1,2,3,4-tetrakis(4-pyridyl)cyclobutane.The complex exists as a threedimensional cationic polymer in which each Ag' ion is coordinated to
one pyridyl group of four different Lloagroups. Each Llw bridges four
tetrahedral metal centers leading t o a three-dimensional array of silver atoms containing channels to fill in the BF; anions and MeCN
solvent molecules (Fig. 97).
N
AgBF,
MeCN/CH,CI,
70
L 106
292
FIG.97. View of the polymeric array formed by [Ag(Llos)l+

illustrating the channels
formed. (From Fig. 3 in Blake, A. J.; Champness, N. R.: Chung, S.S.M.; Li, W.-S.;
SchrGder, M. J. Chem. SOC.Chem. Commun. 1997, 1675.)
X-ray structural determination results revealed that the ligand L,,,

had undergone a [2 + 21 cyclization to form a tetrapyridyl-substituted
cyclobutane ring. The light-induced dimerization is found to be promoted by the presence of Ag' ions. Although the detailed mechanism
of the cyclodimerization reaction is not clear, such in situ formation of
the ligand nevertheless represents a new approach to future inorganic
crystal engineering.
IX. Concluding Remarks
The design and synthesis of new functional polymeric coordination

complexes through intermolecular interactions or coordination bonds
is a continuing challenge. Our aim in this review has been to show
what this development has achieved recently in copper(1)and silver(1)
chemistry. Throughout we have employed relatively straightforward
ideas covering the latest references and reviews. The concept of supramolecular chemistry is not new, but in order to keep this review
within bounds, we have had to severely restrict the number of references. Most of these can be traced via a recent comprehensive review
(286).In the present review we have tried to emphasize three aspects
toward the construction of functional solid-state supramolecular
metal complexes, illustrating them by reference to a variety of topological systems. The first point is that the coordination number and
stereochemistry of metal ions play an important role in construction
of multidimensional networks. Most polymeric frameworks encoun-
COPPER^ AND SILVER(I) SUPRAMOLECULAR METAL COMPLEXES
293
tered, such as diamondoid, are based upon linear, trigonal, and tetrahedral metal templates. Therefore, the 4-coordinate tetrahedral copper(1) ion and the 2-4-coordination of silver(1) ion are likely to be the
best candidates to produce diverse architectures. Second, designing
ligands is an important step in making new functional solids. We
must realize that the study we are undertaking today is not just limited to appreciation of novel frameworks for the aesthetics that the
system can offer. We have already explored a number of functional
ligands such as the conductive molecule TTF and derivatives, the photochromic compound cisl,2-dicyano-1,2-bis(2,4,5-trimethyl-3-thienyl)
ethane (LT5),and the linear exodentate spacer 4,4-bipyridine. We also
have noted that in the metallic conductive compound [Cu(L,,),],
matching of metal d-orbitals with n-orbitals of ligand may be responsible for the observed extremely high electrical conductivity (197200). Thus, in striving for functional coordination materials with potential applications as inorganic devices, the search for functional
ligands constitutes a new challenge. Finally, X-ray diffraction in combination with other physical methods has proven to be the dominant
means of structure determination of inorganic polymers. To obtain
single crystals in the polymeric phase, the art of synthesis must undergo a constant modification. On the other hand, application of
newer experimental techniques for structural analysis of the powder
sample may provide further advances in the study of polymeric structures.
The structural chemistry of coordination compounds continues to
pose challenging problems, but a combination of physical methods
and revolutionized synthetic techniques should clarify the picture and
provide a real opportunity for those involved in this research. It
should be encouraging. Such a prospect provides a major impetus to
discover other coordination polymers with unexpected frameworks.
REFERENCES
Lehn, J.-M. Angew. Chem., Int. Ed. Engl. 1988,27, 89.
Lehn, J.-M. Angew. Chem., Int. Ed. Engl. 1990,29, 1304.
Vogtle, F. Supramolecular Chemistry; Wiley: Chichester, 1993.
Robson, R.; Abrahams, B. F.; Batten, S. R.; Gable, R. W.; Hoskins, B. F.; Liu,
J. In Supramolecular Architecture; Bein, T., Ed.; American Chemical Society:
Washington, DC, 1992; Chapter 19; p. 256.
5. Comprehensive Supramolecular Chemistry; Lehn, J.-M., Ed.; Pergamon Press:
Oxford, 1995; Vol. 9.
1.
2.
3.
4.
294
6. Transitional Metals in Supramolecular Chemistry; Fabbrizzi, L., and Poggi, A.,

Eds.; Kluwer: Holland, 1994.
7. Miiller, A,; Reuter, H.; Dillinger, S. Angew. Chem., Int. Ed. Engl. 1995,34, 2328.
8. Constable, E. C. In Progress in Inorganic Chemistry; Karlin, K. D., Ed.; Wiley:
New York, 1994; Vol. 42; p. 92.
9. Balzani, V. Tetrahedron 1992,48, 10443.
10. Supramolecular Chemistry; Balzani, V., and De Cola, L., Eds.; Kluwer: Holland, 1992.
11. Constable, E. C. Adu. Inorg. Chem. 1989,34, 1.
12. Supramolecular Photochemistry; Balzani, V., Ed.; Reidel: Holland, 1987.
13. Lehn, L.-M. In; Williams, A. F., Floriani, C., and Merbach, A. E., Eds.; VHCA
Basel, 1992.
14. Baxter, P.; Lehn, J.-M.; DeCian, A,; Fischer, J. Angew. Chem., Int. Ed. Engl. 1993,
32, 69.
15. Hathaway, B. J . In; Wilkinson, G., Gillard, R. D., and McCleverty, J . A,, Eds.;
Pergamon: Oxford, 1987; Vol. 5; p. 533.
16. Lancashire, R. J. In Comprehensive Coordination Chemistry; Wilkinson, G., Gillard, R. D., and McCleverty, J. A., Eds.; Pergamon: Oxford, 1987; Vol. 5; p. 775.
17. Kitagawa, S.; Munakata, M. Trends Inorg. Chem. 1993, 3.
18. Munakata, M.; Wu, L. P.; Kuroda-Sowa, T. Bull. Chem. SOC.Jpn. 1997, 70, 1727.
19. Lehn, J.-M.; Rigault, A.; Siegel, J.; Harrowfield, J.; Chevirier, B.; Moras, D. Proc.
Natl. Acad. Sci. USA 1987, 84, 2565.
20. Lehn, J.-M. Supramolecular Chemistry; VCH: Weinheim, 1995.
21. Constable, E. C. Tetrahedron 1992, 48, 10013.
22. Williams, A. Chem. Eur. J . 1997, 3, 15.
23. Gelling, 0. J.; van Bolhuis, F.; Feringa, B. L. J. Chem. SOC.,Chem. Commun.
1991,917.
24. Psillakis, E.; Kotsuki, H.; Isobe, K.; Moriya, N.; Nakagawa, Y.; Ochi, M. J . Chem.
SOC.,Dalton Trans. 1997, 1645.
25. Suzuki, T.; Kotsuki, H.; Isobe, K.; Moriya, N.; Nakagawa, Y.; Ochi, M. Inorg.
Chem. 1995,34, 530.
26. Wu, B.; Zhang, W.-J.; Yu, S.-Y.; Wu, X.-T. J. Chem. Soc., Dalton Trans. 1997,1795.
27. Hartl, H.; M.-H.-Abadi, F. Angew. Chem., Int. Ed. Engl. 1994,33, 1841.
28. Yu, S.-Y.; Luo, Q.-H.; Wu, B.; Huang, X.-Y.; Sheng, T.-L.; Wu, X.-T.; Wu, D.-X.
Polyhedron 1997, 16, 453.
29. Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Moriwaki, K.; Kitagawa, S. Inorg. Chem. 1997,36, 5416.
30. Constable, E. C.; Elder, S. M.; Hannon, M. J.; Martin, A.; Raithby, P. R.; Tocher,
D. A. J . Chem. SOC.,Dalton Trans. 1996,2423.
31. Constable, E. C.; Hannon, M. J.; Martin, A,; Raithby, P. R.; Tocher, D. A. Polyhedron 1992,11, 2967.
32. Potts, K. T.; Keshavarz-K, M.; Tham, F. S.; Abrufia, H. D.; k a n a , C. R. Inorg.
Chem. 1993,32, 4422.
33. Baxter, P. N. W.; Lehn, L.-M.; Rissanen, K. J. Chem. Soc., Chem. Commun.
1997, 1323.
34. Constable, E. C.; Heirtzler, F.; Neuburger, M.; Zehnder, M. J. Am. Chem. SOC.
1997,119, 5606.
35. Piguet, C.; Bernardinelli, G.; Bocquet, B.; Quattropani, A.; Williams, A. F. J. Am.
Chem. SOC.1992,114, 7440.
36. Youinou, M.-T.; Ziessel, R.; Lehn, J.-M. Inorg. Chem. 1991,30, 2144.
295
37. Constable, E. C.; Holmes, J . M.; Raithby, P. R. Polyhedron 1991, 10, 127.
38. Dietrich-Buchecher, C. D.; Sauvage, J.-P.; Cian, A. D.; Fischer, J. J. Chem. Soc.,
39. Dietrich-Buchecher, C. D.; Sauvage, J.-P. Angew. Chem., Int. Ed. Engl. 1989,28,
189.
40. Ruttimann, S.; Piguet, C.; Bernardinelli, G.; Bocquet, B.; Williams, A. F. J . Am.
Chem. SOC.1992,114, 4230.
41. Barley, M.; Constable, E. C.; Corr, S. A,; McQueen, R. C. S.; Nutkins, J . C.; Ward,
M. D.; Drew, M. G. B. J. Chem. Soc., Dalton Trans. 1988, 2655.
42. Constable, E. C.; Edwards. A. J . ; Raithby, P. R.; Walker, J. V. Angew. Chem., Znt.
Ed. Engl. 1993, 32, 1465.
43. Fu, Y.; Sun, J.; Li, Q.; Chen, Y.; Dai, W.; Wang, D.; Mak, T. C. W.; Tang, W.; Hu,
H. J . Chem. Soc., Dalton Trans. 1996, 2309.
44. Ho, P. K.-K.; Peng, S. M.; Wong, K.-Y.; Che, C.-M. J . Chem. Soc., Dalton Trans.
1996, 1829.
45. Constable, E. C.; Edwards, A. J.; Hannon, M. J.; Raithby, P. R. J. Chem. Soc.,
Dalton Trans. 1994, 1991.
46. Potts, K. T.; Keshavarz-K, M.; Tham, F. S.; Abruria, H. D.; Arana, C. R. Inorg.
Chem. 1993,32, 4450.
47. Constable, E. C.; Ward, M. D.; Tocher, D. A. J . Chem. Soc., Dalton Trans. 1991,
1675.
48. Garrett, T. M.; Koert, U.; Lehn, J.-M.; Rigault, A,; Meyer, D.; Fischer, J. J . Chem.
Soc., Chem. Commun. 1990,557.
49. Lehn, J.-M.; Rigault, A. Angew. Chem., Int. Ed. Engl. 1988,27, 1095.
50. Carlucci, L.; Ciani, G.; W. v. Gudenberg, D.; Proserpio, D. M. Inorg. Chem. 1997,
36, 3812.
51. Carina, R. F.; Bernardinelli, G.;Williams, A. F.Angew. Chem., Int. Ed. Engl. 1993,
32, 1463.
52. Zelikovich, L.; Libman, J.; Sbanzer, A. Nature 1995,374, 790.
53. Townsend, J . M.; Blount, J . F.; Sun, R. C.; Zawoiski, S.; Valentine, J. D. J . Org.
Chem. 1980,45, 2995.
54. Piguet, C.; Bernardinelli, G.; Biinzli, J.-C. G.; Petoud, S.; Hopfgartner, G. J. Chem.
SOC.,
Chem. Commun. 1995,2575.
55. Soghomonian, V.; Chen, Q.; Haushalter, R. C.; Zubieta, J.; OConnor, C. J. Science
1993,259, 1596.
56. Maruoka, K.; Murase, N.; Yamamoto, H. J . Org. Chem. 1993, 58, 2938.
57. Woods, C. R.; Benaglia, M.; Cozzi, F.; Siegel, J . S. Angew. Chem., Int. Ed. Engl.
1996,35, 1830.
58. Ohata, N.; Masuda, H.; Yamauchi, 0. Angew. Chem., Int. Ed. Engl. 1996,35, 531.
59. Batten, S. R.; Hoskins, B. F.; Robson, R. Angew. Chem., Int. Ed. Engl. 1997,36,636.
60. Ramos, E.; Bosch, J.; Serrano, J . L.; Sierra, T.; Veciana, J . J. Am. Chem. SOC.
1996,118, 4703.
61. Provent, C.; Hewage, S.; Brand, G.; Bernardinelli, G.; Charbonniere, L. J.; Williams, A. F. Angew. Chem., Int. Ed. Engl. 1997, 36, 1287.
62. Wudl, F.; Smith, G. M.; Hufnagel, E. J . J . Chem. SOC.,Chem. Commun. 1970,1453.
63. Ferraris, J. P.; Cowan, D. 0.;Walatka, V.; Perlstein, J . H. J . Am. Chem. SOC.1973,
95, 948.
64. Wudl, F. Acc. Chem. Res. 1984, 17, 227.
65. Williams, J . M.; Beno, M. A,; Wang, H. H.; Leung, P. C. W.; Emge, T. J.; Geiser,
U.; Carlson, K. D. Acc. Chem. Res. 1985, 18, 261.
296
66. Farraro, J. R.; Williams, J . M. Introduction to Synthetic Electrical Conductors;

Academic Press: New York, 1987.
67. J~rgensen,T.; Hansen, T. K.; Becher, J. Chem. SOC.Rev. 1994, 41.
68. Cassoux, P.; Valade, L.; Kobayashi, H.; Kobayashi, A.; Clark, R. A.; Underhill,
A. E. Coord. Chem. Rev. 1991,110, 115.
69. Olk, R.-M.; Olk, B.; Dietzsch, W.; Kirmse, R.; Hoyer, E. Coord. Chem. Reu. 1992,
117, 99.
70. Cassoux, P.; Valade, L. In Inorganic Materials; Bruce, D. W., Ed.; Wiley: ChiChester, 1992; p. 2.
71. Matsubayashi, G.; Douki, K.; Tamura, H.; Nakano, M. Inorg. Chem. 1993, 32,
5990.
72. Matsubayashi, G.; Maikawa, T.; Nakano, M. J. Chem. SOC.,Dalton Trans. 1993,
2995.
73. Matsubayashi, G.; Takahashi, K.; Tanaka, T. J . Chem. SOC.,Dalton Trans. 1988,
967.
74. Tanaka, S.; Matsubayashi, G. J. Chem. Sac., Dalton Trans. 1992,2837.
75. Nakamura, T.; Underhill, A. E.; Coomber, A. T.; Friend, R. H.; Tajima, H.; Kobayashi, A.; Kobayashi, H. Inorg. Chem. 1995,34, 870.
76. Kobayashi, A,; Kim, H.; Sasaki, Y.; Kato, R.; Kobayashi, H.; Moriyama, S.; Nishio,
Y. Chem. Lett. 1987, 1819.
77. Kobayashi, H.; Bun, K.; Naito, T.; Kato, R.; Kobayashi, A. Chem. Lett. 1992, 1909.
78. Tajima, H.; Inokuchi, M.; Kobayashi, A,; Ohta, T.; Kato, R.; Kobayashi, H.;
Kuroda, H. Chem. Lett. 1993, 1235.
79. Kobayashi, A,; Kobayashi, H.; Miyamoto, A.; Kato, R.; Clark, R. A.; Underhill,
A. E. Chem. Lett. 1991, 2163.
80. Cornelissen, J. P.; L o n , R. L.; Jansen, J.; Haasnoot, J . G.; Reedijk, J.; Horn, E.;
Spek, A. L.; Pomarede, B.; Legros, J.-P.; Reefman, D. J. Chem. SOC.,Dalton Trans.
1992,2911.
81. Pullen, A. E.; Olk, R.-M.; Zeltner, S.; Hoyer, E.; Abboud, K. A.; Reynolds, J. R.
Inorg. Chem. 1997,36, 958.
82. Faulmann, C.; Delpech, F.; Malfant, I.; Cassoux, P. J . Chem. SOC.,Dalton Trans.
1996,2261.
83. Piotraschke, J.; Pullen, A. E.; Abboud, K. A,; Reynolds, R. Inorg. Chem. 1995,
34, 4011.
84. Bousseau, M.; Valade, L.; Legros, J.-P.; Cassoux, P.; Garbauskas, M.; Interrante,
L. V. J. Am. Chem. SOC.1986,108, 1908.
85. Kim, H.; Kobayashi, A.; Sasaki, Y.; Kato, R.; Kobayashi, H. Chem. Lett. 1987,1799.
86. Faulmann, C.; Errami, A.; Donnadieu, B.; Malfant, I.; Legros, J.-P.; Cassoux, P.;
Rovira, C.; Canadell, E. Inorg. Chem. 1996,35, 3856.
87. Sun, S. Q.; Zhang, B.; Wu, P.; Zhu, D. B. J. Chem. SOC.,Dalton Trans. 1997, 277.
88. Yang, X.; Doxsee, D. D.; Rauchfuss, T. B.; Wilson, S. R. J. Chem. SOC.,Chem.
Commun. 1994,821.
89. Pullen, A. E.; Piotraschke, J.; Abbound, K. A.; Reynolds, J. R. Inorg. Chem. 1996,
35, 793.
90. Narvor, N.; Robertson, N.; Weyland, T.; Kilburn, J . D.; Underhill, A. E.; Webster,
M.; Svenstrup, N.; Becher, J. J. Chem. SOC.,Chem. Commun. 1996, 1363.
91. Narvor, N.; Robertson, N.; Wallace, E.; Kilburn, J. D.; Underhill, A. E.; Bartlett,
P.; Webster, M. J. J. Chem. SOC.,Dalton Trans. 1996, 1363.
92. Schultz, A. J.; Wang, H. H.; Soderholm, L. C.; Sifter, T. L.; Williams, J. M.; Bechgaard, K.; Whangbo, M.-H. Inorg. Chem. 1987,2626, 3757.
297
93. Pullen, A. E.; Zeltner, S.; Olk, R.-M.; Hoyer, E.; Abboud, K. A,; Reynolds, J. R.
Znorg. Chem. 1996,35, 4420.
94. Pullen, A. E.; Zeltner, S.; Olk, R.-M.; Hoyer, E.; Abboud, K. A.; Reynolds, J. R.
Inorg. Chem. 1997,36, 4163.
95. Matsubayashi, G.; Yokozawa, A. J. Chem. SOC.,Chem. Commun. 1991,68.
96. Bellitto, C.; Bonamico, M.; Fares, V.; Serino, P. Znorg. Chem. 1996,35, 4070.
97. Geiser, U.; Beno, M. A.; Kini, A. M.; Wang, H. H.; Schultz, A. J.; Gates, B. D.;
Cariss, C . S.; Carlson, K. D.; Williams, J. M. Synth. Met. 1988, 27, A235.
98. Mori, H.; Mori, T.; Kato, K.; Maruyama, Y.; Inokuchi, H.; Kirabayashi, I.; Tanaka,
S. Solid State Commun. 1992, 82, 177.
99. Horiuchi, S.; Yamochi, H.; Saito, G.; Sakaguchi, K.; Kusunoki, M. J. Am. Chem.
SOC.1996, 118, 8604.
100. Inoue, M. B.; Inoue, M.; Bruck, M. A.; Fernando, Q. J. Chern. SOC.,Chem. Commun.
1992,515.
101. Mori, T.; Wu, P.; Imaeda, K.; Enoki, T.; Inokuchi, H.; Saito, G. Synth. Met. 1987,
19, 545.
102. Katayama, C.; Honda, M.; Kumagai, H.; Tanaka, J.; Saito, G.; Inokuchi, H. Bull.
Chem. SOC.Jpn. 1985,58, 2272.
103. Imaeda, K.; Enoki, T.; Shi, Z.; Wu, P.; Okada, N.;Yamochi, H.; Saito, G.; Inokuchi,
H. Bull. Chem. SOC.Jpn. 1987, 60, 3163.
104. Honda, K.; Goto, M.; Kurahashi, M.; Anzai, H.; Tokumoto, M.; Ishiguro, T. Bull.
Chem. SOC.Jpn. 1988,61, 588.
105. Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Hirota, A,; Kitagawa, S. Znorg.
Chem. 1995,34, 2705.
106. Gan, X.; Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Misaki, Y. Polyhedron
1995, 14, 1343.
107. Gan, X.; Munakata, M.; Kuroda-Sowa, T.; Maekawa, M. Bull. Chem. SOC.Jpn.
1994,67, 3009.
108. Kuroda-Sowa, T.; Hirota, A,; Munakata, M.; Maekawa, M. Mol. Cryst. Liq. Cr.yst.
1996, 285, 69.
109. Wu, L. P.; Gan, X.; Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.
Mol. Cryst. Liq. Cryst. 1996, 285, 75.
110. Gan, X.; Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Yamamoto, M. Polyhedron 1995,14, 1647.
111. Yamamoto, M.; Gan, X.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Munakata,
M. Inorg. Chim. Acta 1997,261, 169.
112. Munakata, M.; Wu, L. P.; Gan, X.; Kuroda-Sowa, T.; Suenaga, Y . Mol. Cryst. Liq.
Cryst. 1996,284, 319.
113. Svenstrup, N.; Becher, J. Synthesis 1995, 215.
114. Dai, J. Ph.D. Thesis, Kinki University, 1997.
115. Dai, J.; Kuroda-Sowa, T.; Munakata, M.; Maekawa, M.; Suenaga, Y.; Ohno, Y.
J . Chem. SOC.,Dalton Trans. 1997, 2363.
116. Dai, J.; Munakata, M.; Kuroda-Sowa, T.; Suenaga, Y.; Wu, L. P.; Yamamoto, M.
Inorg. Chim. Acta 1997,255, 163.
117. Dai, J.;Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Suenaga, Y. Inorg. Chim. Acta
1997, 258, 65.
118. Iwamoto, T. In Inclusion Compounds; Atwood, J. L., Davies, J. E. D., MacNicol,
D. D., and Eds.; Oxford University Press: Oxford, 1991; p. 177.
119. Iwamoto, T. In Chemistry of Microporous Crystals; Inui, T., Namba, S., Tatsumi, T., and Eds.; Kodansha/Elsevier: Tokyo, 1991; p. 1.
298
120.
121.
122,
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
151.
152.
153.
Haushalter, R. C.; Mundi, L. A. Chem. Muter. 1992,4, 31.

Gable, R. W.; Hoskins, B. F.; Robson, R. J . Chem. SOC.,Chem. Commun. 1990,762.
Kim, J.; Whang, D.; Koh, Y.-S.; Kim, K. J. Chem. SOC.,Chem. Commun. 1994,637.
Soma, T.; Yuge, H.; Iwamoto, T. Angew. Chem., Int. Ed. Engl. 1994,33, 1665.
Abrahams, B. F.; Hoskins, B. F.; Michail, D. M.; Robson, R. Nature 1994,369, 727.
Hoskins, B. F.; Robson, R. J. Am. Chem. SOC.1989,111, 5962.
Abrahams, B. F.; Hoskins, B. F.; Robson, R. J. Chem. SOC., Chem. Commun.
1990, 60.
Hoskins, B. F.; Robson, R. J. Am. Chem. SOC.1990, 112, 1546.
Abrahams, B. F.; Hoskins, B. F.; Liu, J.; Robson, R. J . Am. Chem. SOC.1191,
113, 3045.
Jung, 0.-S.;Pierpont, C. G. J . Am. Chem. SOC.1994, 116, 2229.
Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. J. Am. Chem. SOC.1995, 117,
4562.
Darriet, J.; Haddad, M. S.; Duesler, E. N.; Hendrickson, D. N. Inorg. Chem. 1979,
18, 2679.
Lumme, P.; Lindroos, S.; Lidell, E. Acta Crystallogr., Sect. C 1987,43, 2053.
Otieno, T.; Rettig, S. J.; Thompson, R. C.; Trotter J. Inorg. Chem. 1993,32, 1607.
Turnbull, M. M.; Pon, G.; Willett, R. D. Polyhedron 1991, 10, 1835.
Kitagawa, S.; Kawata, S.; Kondo, M.; Nozaka, Y.; Munakata, M. Bull. Chem. SOC.
Jpn. 1993,66, 3387.
MacGillivray, L. R.; Subramanian, S.; Zaworotko, M. J. J. Chem. SOC.,Chem.
Commun. 1994, 1325.
Kitagawa, S.; Munakata, M.; Tanimura, T. Inorg. Chem. 1992,31, 1714.
Munakata, M.; Kitagawa, S.; Ujimaru, N.; Nakamura, M.; Maekawa, M.; Matsuda, H. Inorg. Chem. 1993,32, 826.
Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Nakamura, M.; Akiyama, S.-I.;
Kitagawa, S. Inorg. Chem. 1994,33, 1284.
Karl, N.; Ketterer, W.; Stezowski, J. J. Acta Crystallogr. 1982, B38, 2917.
Goldberg, I.; Shmueli, U. Acta Crystallogr. 1973, B29, 440.
Aakeroy, C. B.; Seddon, K. R. Chem. SOC.Rev. 1993,397.
Etter, M. C. Acc. Chem. Res. 1990,23, 120.
Burrows, A. D.; Chan, C.-W.; Chowdhry, M. M.; McGrady, J. E.; Mingos, D. M. P.
Chem. SOC.Rev. 1995,329.
MacDonald, J . C.; Whitesides, G. M. Chem. Rev. 1994, 94, 2383.
Subramanian, S.; Zaworotko, M. J . Coord. Chem. Rev. 1994, 137, 357.
Bernstein, J.; Davis, R. E.; Shimoni, L.; Chang, N.-L. Angew. Chem., Int. Ed. Engl.
1996,34, 1555.
Batsanov, A. S.; Hubberstey, P.; Russell, C. E. J . Chem. SOC.,Dalton Trans.
1994, 3189.
Begley, M. J.; Hubberstey, P.; Stroud, J. J. Chem. SOC.,Dalton Trans. 1996, 2323.
Burrows, A. D.; Mingos, D. M. P.; White, A. J. P.; Williams, D. J. J . Chem. SOC.,
Tadokoro, M.; Toyoda, J.; Isobe, K.; Itoh, T.; Miyazaki, A.; Enokie, T.; Nakasuji,
K. Chem. Lett. 1996, 613.
Yamanari, K.; Kida, M.; Yamamoto, M.; Fujihara, T.; Fuyuhiro, A,; Kaizaki, S. J .
Chem. SOC.,Dalton Trans. 1995, 2627.
Burrows, A. D.; Mingos, D. M. P.; White, A. J . P.; Williams, D. J . J. Chem. SOC.,
COPPERU) AND SILVERO) SUPRAMOLECULARMETAL COMPLEXES
299
154. Itoh, T.; Toyoda, J.; Tadokoro, M.; Kitagawa, H.; Mitani, T.; Nakasuji, K. Chem.
Lett. 1995, 41.
155. Smith, G.; Reddy, A. N.; Byriel, K. A.; Kennard, C. H. L. J. Chem. Soc., Dalton
Trans. 1995, 3565.
156. Chowdhry, M. M.; Mingos, D. M. P.; White, A. J. P.; Williams, D. J. J. Chem. Soc.,
157. Batsanov, A. S.; Begley, M. J.; Hubberstey, P.; Stroud, J. J. Chem. Soc., Dalton
Trans. 1996, 1947.
158. Blake, A. J.; Hill, S. J.; Hubberstey, P.; Li, W . 3 . J . Chem. Soc., Dalton Trans.
1997, 913.
159. Blake, A. J.; Fallis, I. A,; Heppeler, A,; Parsons, S.; Ross, S. A,; Schriider, M. J .
Chem. Soc., Dalton Trans. 1996, 31.
160. Nakasuji, K.; Tadokoro, M.; Toyoda, J.;Mitsumi, M.; Itoh, T.; Iijima, K. Mol. Cryst.
Liq. Cryst. 1996, 285, 241.
161. Mitumi, M.; Toyoda, J.; Nakasuji, K. Znorg. Chem. 1995, 34, 3367.
162. Wu, L. P.; Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y. Inorg.
Chim. Acta 1996,249, 183.
163. Wu, L. P.; Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, T.; Fukui, J.; Munakata,
M. In.org. Chim. Acta 1996, 239, 165.
164. Munakata, M.; Wu, L. P.; Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, M. J . Am.
Chem. SOC.1996, 118, 3117.
165. Munakata, M.; Yamamoto, M. unpublished results, 1997.
166. Wu, L. P.; Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Munakata, M. J . Chem. Soc., Dalton Trans. 1996, 2031.
167. Dai, J.; Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Munakata,
M. Inorg. Chem. 1997,36, 2688.
168. Thompson, J. S.; Whitney, J. F. Inorg. Ch,em. 1984,23, 2813.
169. Masuda, H.; Yamamoto, N.; Taga, T.; Machida, K.; Kitagawa, S.; Munakata, M.
J . Organomet. Chem. 1987,322, 121.
170. Masuda, H.; Machida, K.; Munakata, M.; Kitagawa, S.; Shimono, H. J. Chem. Soc.,
171. Munakata, M.; Kitagawa, S.; Kawada, I.; Maekawa, M.; Shimono, H. J . Chem.
Soc., Dalton Trans. 1992, 2225.
172. Not in the original paper, but we reexamined from the crystallographic data of
ref. 173.
173. Hubberstey, P., private communication.
174. Munakata, M.; Maekawa, M.; Kitagawa, S.; Adachi, M.; Masuda, H. Inorg. Chim.
Acta 1990, 167, 181.
175. Lee, S. W.; Trogler, W. C. Inorg. Chem. 1990,29, 1659.
176. Healy, P. C.; Pakawatchai, C.; White, A. H. J . Chem. Soc., Dalton Trans. 1983,
1917.
177. Engelhardt, L. M.; Pakawatchai, C.; White, A. H.; Healy, P. C. J . Chem. Soc.,
178. Munakata, M.; Kitagawa, S.; Shimono, H.; Masuda, H. Inorg. Chim. Acta 1989,
158, 217.
179. Rumpel, H.; Limbach, H. H. J. Am. Chem. Soc. 1989, 111, 5429.
180. Nakasuji, K.; Sugiura, K.; &tagawa, T.; Toyoda, J.; Okamoto, H.; Okaniwa, K.;
Mitani, T.; Yamamoto, H.; Murata, I.; Kawamoto, A,; Tanaka, J. J. Am. Ch,ena.
SOC.1991, 113, 1862.
300
181. Munakata, M.; Dai, J.; Maekawa, M.; Kuroda-Sowa, T.; Fukui, J. J . Chem. Sac.,
182. Albert, A.; Goldacre, R.; Phillips, J. J. Chem. Sac. 1948, 2240.
183. Kuroda-Sowa, T.; Munakata, M.; Matsuda, H.; Akiyama, S.; Maekawa, M. J .
184. Morpurgo, G. 0.;Fares, V.; Dessy, G. J. Chem. Sac., Dalton Trans. 1984, 785.
185. Dessy, G.; Fares, V.; Imperatori, P.; Morpurgo, G. 0.J. Chem. Sac., Dalton Trans.
1985, 1285.
186. Munakata, M.; Kuroda-Sowa, T.; Maekawa, M.; Honda, A.; Kitagawa, S. J. Chem.
Sac., Dalton Trans. 1994, 2771.
187. Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Sugimoto,
K. Inorg. Chem. 1997,36, 4903.
188. Wells, A. F. Structural Inorganic Chemistry, 5th ed.; Oxford University Press:
Oxford, 1984.
189. Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. Angew. Chem., Znt. Ed. Engl.
1995,34, 1895.
190. Carlucci, L.; Ciani, G.;Proserpio, D. M.; Sironi, A. J. Chem. Sac., Chem. Commun.
1994,2755.
191. Yaghi, 0. M.; Li, H. J. Am. Chem. Sac. 1995,117, 10404.
192. Blake, A. J.; Champness, N. R.; Chung, S. S. M.; Li, W.-S.; Schroder, M. J. Chem.
Sac., Chem. Commun. 1997, 1005.
193. Lopez, S.; Kahraman, M.; Harmata, M.; Keller, S. W. Znorg. Chem. 1997,36, 6138.
194. Kinoshita, Y.; Matsubara, I.; Saito, Y. Bull. Chem. Sac. Jpn. 1959,32, 1221.
195. Kuroda-Sowa, T.; Yamamoto, M.; Munakata, M.; Seto, M.; Maekawa, M. Chem.
Lett. 1996, 349.
196. Kuroda-Sowa, T.; Horino, T.; Yamamoto, M.; Ohno, Y.; Maekawa, M.; Munakata,
M. Znorg. Chem. 1997,36, 6382.
197. Aumuller, A,; Erk, P.; Klebe, G.; Hunig, S.; Schutz, J. U.; Werner, H.-P. Angew.
Chem., Znt. Ed. Engl. 1986,25, 740.
198. Ermer, 0.Adu. Muter. 1991,3, 608.
199. Tomic, S.; Jerome, D.; Aumiiller, A.; Erk, P.; Hunig, S.; V. Schiitz, J . U. Synth.
Met. 1988,27, B281.
200. Uji, S.; Terashima, T.; Aoki, H.; Brooks, J. S.; Kato, R.; Sawa, H.; Aonuma, S.;
Tamura, M.; Kinoshita, M. Phys. Reu. B 1994,50, 15597.
201. Hirsch, K. A.; Venkataraman, D.; Wilson, S. R.; Moore, J . S.; Lee, S. J. Chem.
SOC.,Chem. Commun. 1995, 2199.
202. Hirsch, K. A.; Wilson, S. R.; Moore, J . S. Chem. Eur. J. 1997,3, 765.
203. Hirsch, K. A.; Wilson, S. R.; Moore, J. S. Znorg. Chem. 1997, 36, 2960.
204. Zhdanow, H. C. R. Acad. Sci. URSS 1941,31, 350.
205. Shugam, E.; Zhdanow, H. Acta Physiochim. URSS 1945,20, 247.
206. Cromer, D. T.; Larson, A. C. Acta Crystallogr., Sect. B 1972, B28, 1052.
207. Blake, A. J.; Champness, N. R.; Khlobystov, A. N.; Lemenovskii, D. A,; Li, W.-S.;
Schroder, M. J. Chem. Sac., Chem. Commun. 1997, 1939.
208. Michaelides, A.; Kiritsis, V.; Skoulika, S.; Aubry, A. Angew. Chem., Znt. Ed. Engl.
1993,32, 1495.
209. Irie, M. In Photoreactive Materials for Ultrahigh Density Optical Memory; Irie,
M., Ed.; Elsevier: Amsterdam, 1994; p. 1.
210. Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Furuichi,
K. J. Am. Chem. Sac. 1996,118, 3305.
301
211. Wu, L. P.; Suenaga, Y.; Kuroda-Sowa, T.; Maekawa, M.; Furuichi, K.; Munakata,
M. Inorg. Chim. Acta 1996,248, 147.
212. Yamazalu, S.; Deeming, A.; Speel, D. M.; Hibbs, D. E.; Hursthouse, M. B.; Malik,
K. M. A. J. Chem. SOC.,Chem. Commun. 1997, 177.
213. Guerret, 0.;Sole, S.; Gornitzka, H.; Teichert, M.; Trinquier, G.; Bertrand, G. J .
Am. Chem. SOC.1997,119, 6668.
214. Fortin, D.; Drouin, M.; Turcotte, M.; Harvey, P. D. J . Am. Chem. SOC.1997, 119,
531.
215. Dartiguenave, M.; Dartiguenave, Y.; Mari, A.; Guitard, A,; Olivier, M. J . ; Beauchamp, A. L. Can. J. Chem. 1966, 66, 2386.
216. Dhingra, S. S.; Sea, D.-K.; Kowach, G. R.; Kremer, R. K.; Shreeve-Keyer, J . L.;
Haushalter, R. C.; Whangbo, M.-H. Angew. Chem., Int. Ed. Engl. 1997,36, 1087.
217. Huang, Q.; Wu, X.; Wang, Q.; Sheng, T.; Lu, J . Angew. Chem., Int. Ed. Engl. 1996,
35, 868.
218. Manoli, J. M.; Potvin, C.; Secheresse, F.; Marzak, S. Inorg. Chem. Acta 1988,
150, 257.
219. Hagrman, D.; Zubieta, C.; Rose, D. J.; Zubieta, J.; Haushalter, R. C. Angew.
Chem., Int. Ed. Engl. 1997, 36, 873.
220. Lu, J.; Crisci, G.; Niu, T.; Jacobson, A. Inorg. Chem. 1997,36, 5140.
221. Ainscough, E. W.; Brodie, A. M.; Husbands, J . M.; Gainsford, G. J.; Gabe, E. J.;
Curtis, N. F. J . Chem. SOC.,Dalton Trans. 1965, 151.
222. Brook, D. J . R.; Lynch, V.; Conklin, B.; Fox, M. A. J . Am. Chem. SOC.1997, 119,
5155.
223. Oshio, H.; Watanabe, T.; Ohto, A,; Ito, T. Inorg. Chem. 1997, 36, 1608.
224. Chen, X.-M.; Mak, T. C. W. J. Chem. SOC.,Dalton Trans. 1991, 1219.
225. Blake, A. J.; Li, W.-S.; Lippolis, V.; Schroder, M. J. Chem. SOC.,Dalton Trans.
1997, 1943.
226. Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Nakagawa, S. J. Chem. SOC.,Dalton Trans. 1996, 1525.
227. Yamamoto, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M.; Suenaga, Y.; Munakata, M. Inorg. Chim. Acta 1997,258, 87.
228. Buchholz, H. A,; Prakash, G. K. S.; Vaughan, J . F. S.; Bau, R.; Olah, G. A. Inorg.
Chem. 1996,35, 4076.
229. Blake, J. R.; Champness, N. N.;Levason, W.; Reid, G. J . Chem. SOC.,Chem. Comn u n . 1995, 1277.
230. Dubler, E.; Bensch, W. Inorg. Chim. Acta 1986, 125, 37.
231. Plappert, E. C.; Mingos, D. M.; Lawrence, S. E.; Williams, D. J . J . Chem. Soc.,
232. Henary, M.; Wootton, J . L.; Khan, S. I.; Zink, J. I. Inorg. Chem. 1997, 36, 796.
233. Begley, M. J.; Eisenstein, 0.;Hubberstey, P.; Jackson, S.; Russell, C. E.; Walton,
P. H. J . Chem. Soc., Dalton Trans. 1994, 1935.
234. Kitagawa, S.; Munakata, M.; Tanimura, T. Chem. Lett. 1991, 623.
235. Kitagawa, S.; Matsuyama, S.; Munakata, M.; Osawa, N.; Masuda, H. J . Chem.
SOC.,Dalton Trans. 1991, 1717.
236. Constable, E. C.; Steel, P. J. Coord. Chem. Reu. 1989,93, 205.
237. Buchner, E. Chem. Ber. 1669,22, 842.
238. Munakata, M.; Wu, L. P.; Yamamoto, M.; Kuroda-Sowa, T.; Maekawa, M.;
Kawata, S.; Kitagawa, S. J. Chem. Sac., Dalton Trans. 1995, 4099.
239. Bertelli, M.; Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. J . Muter. Chem.
1997, 7, 1271.
302
240. Naskar, J . P.; Hati, S.; Datta, D.; Tocher, D. A. J. Chem. SOC.,Chem. Commun.
1997, 1319.
241. Kitagawa, S.; Matsuyama, S.; Munakata, M.; Emori, T. J . Chem. SOC.,Dalton
Trans. 1991, 2869.
242. Kulawiec, R. J.; Crabtree, R. H. Coord. Chem. Reu. 1990,99, 89.
243. Powell, J.; Horvath, M. J.; tough, A. J . Chem. SOC.,Drzlton Trans. 1996, 1669.
244. Schill, G. Catenanes, Rotaxanes and Knots; Academic: New York, 1971.
245. Jager, R.; Vogtle, F. Angew. Chem., Int. Ed. Engl. 1997,36, 930.
246. Dietrich-Buchecker, C. 0.;Sauvage, J.-P. Chem. Reu. 1987,87, 795.
247. Cardenas, D. J.; Gavifia, P.; Sauvage, J.-P. J. Am. Chem. SOC.1997,119, 2656.
248. Solladie, N.; Chambron, J.-C.; Dietrich-Buchecker, C. 0.; Sauvage, J.-P. Angew.
Chem., Int. Ed. Engl. 1996, 35, 906.
249. Baxter, P. N. W.; Sleiman, H.; Lehn, J.-M.; Rissanen, K. Angew. Chem., Int. Ed.
Engl. 1997,36, 1294.
250. Piguet, C.; Bernardinelli, G.; Williams, A. F.; Bocquet, B. Angew. Chem., Int. Ed.
Engl. 1995, 34, 582.
251. Sleiman, H.; Baxter, P.; Lehn, J.-M.; Rissanen, K. J. Chem. SOC.,Chem. Commun.
1995, 715.
252. Blake, A. J.; Champness, N. R.; Khlobystov, A,; Lemenovskii, D. A.; Li, W.-S.;
Schroder, M. J. Chem. Soc., Chem. Commun. 1997,2027.
253. Hoskins, B. F.; Robson, R.; Slizys, D. A. J , Am. Chem. SOC.1897, 219, 2952.
254. Whang, D.; Kim, K. J. Am. Chem. SOC.1997,119, 451.
255. Parish, R. V.; Salehi, Z.; Pritchard, R. G. Angew. Chem., Int. Ed. Engl. 1997,
36, 251.
256. Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. Angew. Chem., Int. Ed. Engl.
1996,35, 1088.
257. Venkataraman, D.; Gardner, G. B.; Lee, S.; Moore, J. S. J. Am. Chem. SOC.1995,
117, 11600.
258. Wu, D.-D.; Mak, T. C. W. J. Chem. SOC.,Dalton Trans. 1995, 2671.
259. Munakata, M.; Wu, L. P.; Kuroda-Sowa, T.; Maekawa, M. J . Chem. SOC.,Dalton
Trans. 1995, 3215.
260. Manners, I. Angew. Chem., Int. Ed. Engl. 1996,35, 1602.
261. Schon, J . C.; Jansen, M. Angew. Chem., Int. Ed. Engl. 1996,35, 1286.
262. Pfitzner, A.; Zimmerer, S. Angew. Chem., Int. Ed. Engl. 1997,36, 982.
263. Zhang, X.; Kanatzidis, M. G. J. Am. Chem. SOC.1994, 116, 1890.
264. Jerome, J. E.; Wood, P. T.; Pennington, W. T.; Kolis, J. W . Inorg. Chem. 1994,
33, 1733.
265. Baxter, P. N. W.; Lehn, J.-M.; Fischer, J.; Youinou, M.-T. Angew. Chem., Int. Ed.
Engl. 1994, 33, 2284.
266. Desiraju, G. R. Angew. Chem., Int. Ed. Engl. 1995,34, 2311.
267. Janiak, C. Angew. Chem., Int. Ed. Engl. 1997, 36, 1431.
268. Yaghi, 0. M.; Li, H.; Groy, T. L. J . Am. Chem. SOC.1996,118, 9096.
269. Yaghi, 0. M.; Davis, C. E.; Li, G.; Li, H. J. Am. Chem. SOC.1997,119, 2861.
270. Iwamoto, T.; Nishikiori, S.; Kitazawa, T.; Yuge, H. J. Chem. SOC.,Dalton Trans.
1997,4127.
271. Hoskins, B. F.; Robson, R.; Scarlett, N. V. Y. J , Chem. SOC.,Chem. Commun.
1994,2025.
272. Fiijita, M.; Kwon, Y. J.; Washizu, S.; Ogura, K . J . Am. Chem. SOC.
1994,116, 1151.
273. Robinson, F.; Zaworotko, M. J . Chem. SOC.,Chem. Commun. 1995, 2413.
274. Yaghi, 0. M.; Li, H. J . Am. Chem. SOC.1996,118, 295.
COPPER(I) AND SILYERU) SUPRAMOLECULAR METAL COMPLEXES
303
275. Yaghi, 0. M.; Li, G. Angew. Chem., Int. Ed. Engl. 1995,34, 207.
276. Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. J . Am. Chem. Soc. 1995, 117,
1286 1.
277. Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. Znorg. Chem. 1997,36, 1736.
278. Carlucci, L.; Ciani, G.; v. Gudenberg, D. W.; Proserpio, D. M.; Sironi, A. J. Chem.
Soc., Chem. Commun. 1997,631.
279. Abrahams, B. F.; Batten, S. R.; Hamit, H.; Hoskins, B. F.; Robson, R. Angew.
Chem., Int. Ed. Engl. 1996, 35, 1690.
280. Black, J. R.; Champness, N. R.; Levason, W.; Reid, G. Znorg. Chem. 1996,35, 4432.
281. Gardner, G. B.; Venkataraman, D.; Moore, J. S.; Lee, S. Nature 1995, 374, 792.
282. Gardner, G. B.; Kiang, Y.-H; Lee, S.; Asgaonkar, A,; Venkataraman, D. J. Am.
Chem. Soc. 1996,118, 6946.
283. Yaghi, 0. M.; Li, G.; Li, H. Nature 1995, 378, 703.
284. Keller, S. W. Angew. Chem., Znt. Ed. Engl. 1997, 36, 247.
285. Blake, A. J.; Champness, N. R.; Chung, S. S. M.; Li, W . 3 ; Schroder, M. J . Chem.
Soc., Chem. Commun. 1997, 1675.
286. Philp, D.; Stoddart, J. F. Angew. Chem., Int. Ed. Engl. 1996,35, 1154.
ADVANCES IN INORGANIC CHEMISTRY, VOL. 46
MANGANESE REDOX ENZYMES AND MODEL

SYSTEMS: PROPERTIES, STRUCTURES,
AND REACTIVITY
NEIL A. LAW, M. TYLER CAUDLE, and VINCENT L. PECORARO
Department of Chemistry, The University of Michigan, Ann Arbor, Michigan 48109
I. Introduction
11. The Enzymes
A. Manganese Superoxide Dismutase, MnSOD
B. Manganese(I1) Dioxygenase
C. Manganese Peroxidase, MnP
D. Manganese Ribonucleotide Reductase, MnRR
E. Manganese Catalase
F. Binuclear Manganese Enzymes: A Comparison of Three Enzymes
G . The Oxygen-Evolving Complex, OEC
A. Mononuclear
B. Binuclear
C. Trinuclear
D. Tetranuclear
A. Electronic Spectroscopy
B. Magnetism
C . EPR Spectroscopy
D. X-Ray Absorption Spectroscopy
V. Reactivity
A. Organic Transformations
B. Enzyme Model Systems
VI. Conclusion
References
I . Introduction
Manganese is one of several first-row transition elements that have

been employed by biological systems to assist in varied metabolic and
structural roles. Manganese is used to give structural support to pro305
Copyiiglit
1999 by Academic P I F S ~
All ilghla of repiudurtion i n any form reserved
ORSR-RR JRi9Y $25 00
306
LAW, CAUDLE, AND PECORARO
teins and is a cofactor in chemical transformations that include hydrolytic and redox reactions. Perhaps the best known function, and
the one of greatest importance to aerobic life, is in the oxygen-evolving complex, which oxidizes water to dioxygen during photosynthesis.
Some representative reactions catalyzed by manganese-containing
enzymes are shown in Scheme 1.
Oxygen Evolving Complex
2 HzO
*4Hf+
4e-
O2
Manganese Catalase
Manganese Superoxide Dismutase
Arginase
arginine
ornithine+ urea
Manganese RibonucleotideReductme
RNA
DNA
SCHEME
1.
Several enzymes are known that naturally contain manganese.

Most of these employ manganese in nonredox roles (I),for example
conducting hydrolytic reactions, a Lewis acid role, or utilizing Mn" to
provide structural support to a protein so that it may maintain a specific conformation. For example, pyruvate carboxylase binds Mn"
tightly, but the Mn" is not involved in the enzyme's catalytic process,
thus playing a structural role (2). Several binuclear metallohydrolases are known that contain a binuclear Mni' site (3-5). Many of
these may be categorized as phosphohydrolases, hydrolyzing phosphate groups from a substrate molecule. Hydrolases are often activated by more than one MI1 element, although at least one is specific
for manganese.
One particularly key metallohydrolase that contains manganese is
the enzyme arginase (4-7), an enzyme for which a crystal structure
MANGANESE REDOX ENZYMES AND MODEL SYSTEMS
307
at 2.1-A resolution has recently been solved for the arginase enzyme
from rat liver (Fig. 1) (8). This enzyme catalyzes the hydrolytic conversion of arginine to urea and ornithine. Arginase is found in a great
variety of organisms for the removal of nitrogen-containing wastes
and is therefore integral to the biological cycle of nitrogen. Many of
the binuclear hydrolases may be active with metals other than Mn at
their active sites, but arginase has been shown to specifically require
manganese to be fully functional ( 5 ) .The active site contains a binuclear Mnp dimer bridged by a bidentate carboxylate and two p2-"O"
donor bridges, a monodentate carboxylate, and a solvent molecule derived hydroxide (or perhaps water a t lower pH, the crystals were
grown at pH 8.5). The catalysis of this reaction involves a bridging
p-OH unit, which is formed according to a mechanistic proposal by
loss of a proton from a water molecule upon binding of substrate. This
hydroxide is believed to be ideally situated for forming the intermediates required for arginase hydrolysis (8). Finally, arginase has recently been reported to exhibit some redox activity, catalyzing the
disproportionation of hydrogen peroxide and peroxidase activity (discussed later) (9,10).
Manganese(I1) ions may also be employed in the hydrolases that
compose the ribonuclease H domain of reverse transcriptases ( 3 , 4 ) .
Many of these enzymes, which may employ either Mn" or Mg", are
found in a variety of organisms where they may or may not be essential (e.g., Escherichia coli) ( 3 , 4 ) . This class of enzymes catalyzes the
hydrolysis of DNA-RNA hybrids. However, retroviral reverse transcriptases are critical for the replication of retroviruses, and Mn" may
be the required cofactor for these ribonuclease hydrolases to function
( 3 , 4 ) . One example of this family of hydrolase enzymes is the RNase
H enzyme from HIV-I. This enzyme has been crystallographically
characterized ( 1 1 ) . In the crystal structure at 2.4-A resolution, the
two Mn" ions are separated by a distance of about 4 A. They are
bound to carboxylate residues that are located near the surface of
the enzyme. One of these carboxylates bridges the two manganese
FIG.1. Structure of the active site of the rat liver arginase enzyme based on (8).
308
ions in a motif that is now recognized as a common structural element

in many binuclear metalloenzyme active sites (discussed later). Some
controversy with respect to the active sites of ribonuclease enzymes
does exist, with some reports suggesting that there may be only one
metal ion present (or perhaps required) in the active enzyme (12, 13).
Concanavalin A, a lectin, is another interesting nonredox manganese-containing enzyme that is found in various forms in plants and
animals (14, 15). Lectins are a class of enzymes known to bind small
carbohydrate molecules (16, 17). The active site of crystallographically characterized concanavalin A enzyme shows a six-coordinate
Mn" bridged via two carboxylate-bearing protein residues to an approximately seven-coordinate Ca" ion (Fig. 2) (14,015,18).The calcium
and manganese ions are separated by about 4.2 A. Concanavalin A is
an interesting case of a mixed metal active site in a biological system.
If the Call is replaced by another Mn" ion, an EPR spectrum consistent with a weakly coupled MnH system is observed (19, 20). Newer
crystal structures of concanavalin A with the substrate methyl-a-Dmannopyranoside have been reported. The more recent structure in
1994 has a resolution of 2.0 A. These structures show that methyl-an-mannopyranoside is bound near the metal binding site, about 8.7 A
from the calcium and about 12.8 A from manganese (21). The 1994
structure also indicates that a water bound to the calcium may be
involved in hydrogen bonding to the carboxylate-bearing protein residues, which are in turn hydrogen bond t o the methyl-a-D-mannopyranoside (18). Finally, a new crystal structure for concanavalin A at
0.94-A resolution has been solved (22).
N
FIG.2. Structure of the manganese and calcium dimer in concanavalin A, based on
( 1 4 , 15, 18,221.
309
Redox chemistry is another facet of the biological chemistry of manganese, and various enzymes employ manganese a t mono-, bi-, and
tetranuclear sites. Three mononuclear manganese enzymes are currently known that catalyze redox-based chemical transformations.
These are manganese superoxide dismutase, manganese peroxidase,
and manganese dioxygenase. Catalase and the manganese ribonucleotide reductase are two known manganese enzymes that undertake
redox transformations that utilize a binuclear manganese site, and a
tetranuclear cluster of four manganese is found in the oxygen evolving complex (OEC) of Photosystem I1 (PS 111, the site where water is
oxidized during the photosynthetic process to yield dioxygen.
The focus of this review will be on mono- and multinuclear redox
active manganese systems, and in particular, the relation of bi-, tri-,
and tetranuclear manganese complexes as models for the multinuclear manganese-containing enzymes. Aspects of the structurally
characterized compounds, spectroscopic features of synthetic compounds and magnetochemistry will also be discussed. The reactivity of complexes will also be addressed, with respect to epoxidation
and catalase models, for example. Throughout, discussion will relate our current knowledge of manganese chemistry to the broader
area of biological systems. Before addressing the synthetic complexes
and their properties in detail, a brief overview of some of the redoxactive manganese enzymes will be presented. Finally, several reviews of manganese coordination complex chemistry have appeared
in recent years. Among these are reviews by Wieghardt, Pecoraro
and co-workers, Vincent and Christou, Brudvig and Crabtree, and
Dismukes and co-workers, which have focused on a wide range of
model systems, manganese-dioxygen interactions, mechanistic considerations of reactivity studies, and single systems such as the
OEC (5, 23-50). The 1992 book Manganese Redox Enzymes (28)
also contains much relevant information about biological manganese chemistry by a variety of researchers. A 1996 issue of Chemistry Reviews focusing on bioinorganic chemistry contains several reviews that are pertinent to the biological role of manganese (3, 33,
37). In addition, several reviews aimed solely at the OEC are also
available, including reviews by Debus (42) and the recent book
Oxygenic Photosynthesis: The Light Reactions edited by Ort and
Yocum (48). Finally, the Proceedings of the Photosynthesis Conference, held periodically, presents short updates from research groups
around the world as they work to comprehend this intricate system (51).
310
II. The Enzymes
A. MANGANESE
SUPEROXIDE
DISMUTASE,
MnSOD
Manganese superoxide dismutase (MnSOD) is a redox-active manganese enzyme that employs a mononuclear manganese ion at its active site. Discovered in 1970 (52),this enzyme catalyzes the dismutation of superoxide (HOz), to dioxygen and hydrogen peroxide, as
shown in Scheme 2. Superoxide is the radical, one-electron-reduced
Mn3+
Mn2+
+
+
HOP'
HOP' + H+
Mn2+ + O2 + H'
Mn3+ + H202
SCHEME
2.
form of dioxygen. MnSOD is found in a broad range of species that

run the gamut from bacteria to humans (53).It is a member of one of
the two families of superoxide dismutase enzymes, the first of which
is the family that contains MnSOD and the analogous iron superoxide
dismutase, FeSOD. A high degree of sequence homology exists between the two types of enzymes, and the active sites of both enzymes
employ the same types and orientations of protein residues to generate a trigonal bipyramidal geometry about the active site metal (54611. The second family contains the structurally distinct copper/zinc
superoxide dismutase, CuZnSOD (62, 63). Recent reports have indicated that a new group of nickel superoxide dismutase enzymes might
also exist (64). Superoxide and superoxide dismutases have been the
subject of a variety of reviews (5,29, 53, 65).
Superoxide, H02, is produced naturally by several enzymes and
was first recognized as an intermediate of aerobic respiration in 1968
(62, 66-68). Although it may be utilized beneficially, for example by
macrophages t o fight an infection (691,it is generally not a benign byproduct. Therefore, SODS are critical to aerobic life forms, in which
they perform the important role of protecting an organism from the
deleterious effects of superoxide. Although HOz may not be the direct
cause of oxidative damage to an organism due to its relatively short
life span (self dismutation occurs at a rate of 2.0-3.2 X lo5 M-' .
s - l ) (521, it is sufficiently long-lived to allow it to engender chemical
compounds such as hydrogen peroxide or hydroxyl radical, HO.,
which readily cause oxidative damage to cellular components (70).
Another aspect of superoxide chemistry is the growing research into
peroxynitrite (71-741, which can be formed from nitrite and superoxide. Thus, superoxide has been implicated as a participant in a vari-
311
ety of destructive processes affecting DNA, cell membranes, arteriosclerosis and the damage that is caused to cells deprived of oxygen,
ischemia, during a stroke or heart attack due to an increased output
of superoxide once oxygen is restored to the affected cells and their
enzymes (69, 71, 72).
MnSOD is often found as a homodimer or homotetramer, with approximately one manganese per monomer subunit (53).A few of these
enzymes have been crystallographically characterized, including the
enzyme from human mitochondria (75) and the enzyme from the bacterium Thermus thermophilus (76).The active sites of these enzymes
have been shown to contain a single manganese bound by three histidines, one aspartate, and either a water or hydroxide in a trigonal
bipyramidal geometry as shown in Fig. 3, based on the T. thermophiZus structure (76).In this structure, two of the histidine residues and
the aspartate form the equatorial plane, and the third histidine and
waterhydroxide occupy the axial positions. Other structures of this
enzyme with azide in the active site indicate that the coordination
number of the manganese may be expanded to accommodate a sixth
ligand, for example, superoxide during turnover (58).
MnSOD catalyzes the disrnutation of HOz into dioxygen and hydrogen peroxide. As a redox enzyme, it shuttles between the MnT'and
the Mn"' oxidation states (77).This process has been studied, and the
enzyme has been shown to catalyze this reaction at a rate of 1-2.2 X
109 M-I.s-I , which is at the diffusion limit (52). A representation of
the proposed catalytic cycle is shown in Scheme 3 (78). Several studies on MnSOD kinetics have been published (78-82). Early studies on
the reaction of MnSOD from Escherichia coli and Bacillus thermophilus indicated evidence for a four-step process involving two fast and
FIG.3. Structure of the active site of manganese superoxide dismutase, based on

(76).
3 12
two slower steps (79, 80). Later studies utilizing stopped-flow techniques refined this further. Saturation kinetics were observed for the
T. thermophilus MnSOD, and the data showed that there were three
phases to this reaction: a fast burst phase of quick HOz dismutation,
a slower second phase, and a final fast phase (78).A dead-end form
of the enzyme was implicated to account for the slow phase (78). This
has been formulated as a side-on-bound MnIII-peroxo species based
on spectroscopic similarities to manganese model complexes with
side-bn-bound-peroxy groups (83). Such phases are not observed for
FeSOD (84).
H02
Mn3+SOD-Oi
Mn2+SOD
Dead-EndComplex
SCHEME
3.
B. MANGANESE(II)
DIOXYGENASE
There now exist three reported Mn(I1) dioxygenase enzymes. The
most extensively studied of these is the Mn(I1) dioxygenase from
Arthrobacter globoformis CM-2 (85).The first Mn(I1) dioxygenase was
reported in 1981 from Bacillus breuis (86).Dioxygenases catalyze the
incorporation of dioxygen into an organic substrate (87). In this case,
the substrate is an aromatic compound that contains a catechol cisdiol structure. These reactions are important in that they catalyze the
degradation of aromatic compounds to simpler carbon components for
action by other enzymes. In addition, the ability of dioxygenase enzymes to initiate the decomposition of aromatic compounds may be
useful in the bioremediation of sites contaminated with aromatic organics of this type (88).
Most of the known dioxygenases contain iron, and these enzymes
tend to be very specific in their function (87, 89). For example, the
313
Fe(II1) dioxygenases have generally been shown to be intradiol dioxygenases, meaning that they cleave the aromatic ring between the cisdiols. The Fe(I1) dioxygenases are classified as extradiol dioxygenases
and cleave to the aromatic ring to one side of the diol moiety. Substrate specificity is also exhibited to various degrees, with many dioxygenases preferentially cleaving one substrate over others (89). The
Mn(I1) dioxygenases characterized to date belong to the class of extradiol dioxygenases (86, 90).
Although none of the Mn(I1) dioxygenases have been structurally
characterized, a recent structure has appeared for an Fe(I1) dioxygenase (91).Whereas the Fe"' enzymes have phenolate and histidine ligands, the Fe" was bound to the protein by two histidines and a glutamate, with two waters completing the first coordination sphere. The
geometry about the iron atom is square pyramidal, with one of the
histidines occupying the axial position (Fig. 4).Finally, some evidence
for an Fe"/Mn" dioxygenase correspondence similar to that observed
for MnSOD/FeSOD might also exist, as evidenced by the conserved
residues that form the metal binding sites of these dioxygenases (92).
Thus, this manganese enzyme may have a structure similar to the
iron site.
The details of the reactions of the iron dioxygenases have been extensively probed for the Fe(II1) enzymes and more recently for the
Fe(I1) dioxygenases (87,891. These two groups of enzymes exhibit different catalytic processes. For both of these types of enzymes, though,
it has been shown that although the reaction catalyzed is redox in
nature, the central metal does not cycle between discrete oxidation
states during the reaction. A resonance form of an intermediate, however, has been proposed to exist briefly as Fe"' for the Fe(I1) dioxygenases. A recent proposal for the catalytic cycle of an Fe(I1) dioxygenase
is presented in Scheme 4 (89, 93). The catechol substrate is thought
to bind as a monoanion to the central metal ion followed by the coordi-
FIG.4. Structure of the active site of an extradiol iron(I1) dioxygenase, based on (91).
314
SCHEME
4. A proposed mechanism for extradiol Fe(I1)dioxygenases. [Adapted with
permission from (89).Copyright 1996 the American Chemical Society.]
nation of a dioxygen molecule to the metal, which might be formally

considered to be a superoxide bound t o an Fe. It is believed that
substrate binding activates the metal center for binding dioxygen,
based on the increased affinity for binding nitrosyl to the active site
iron in the presence of substrate (89). Attack of the superoxide
group on the carbon adjacent to the catechol moiety followed by rearrangement to give the product and the products release complete
the catalytic cycle. This mechanism contrasts with the Fe(II1) dioxygenase mechanism, wherein the iron activates the substrate for direct
attack of dioxygen on the aromatic ring of the substrate. The reaction
rate of the manganese containing A. globoformis enzyme, with its primary substrate 3,4-dihydroxyphenylacetate, has been measured,
yielding values of kcat= 1400 min-l, and K, = 7 pM (85).It is likely
315
that a similar mechanism exists between the Fe" and Mn" enzymes
(89,90).
1. Modeling Features, MnSOD and Mn(II) Dioxygenase
These two enzymes present a unique challenge for the synthesis of
small-molecule model complexes. First, the MnSOD has been shown
to have an active site with a trigonal bipyramidal geometry about the
manganese ion, whereas the Mn(I1) dioxygenase most likely contains
a square pyramidal geometry about the manganese ion in accord with
the Fe(I1) dioxygenase structure. Few complexes of manganese in a
trigonal bipyramidal geometry are known, and these are predominately of Mn", whereas the MnSOD enzyme has been crystallographically characterized with a trigonal bipyramidal geometry about the
manganese in both the Mn" and Mn"' oxidation states. In the case of
the Mn(I1) dioxygenase, the structure is probably square pyramidal
by analogy with the Fe(I1) dioxygenase structure. Square pyramidal
complexes are the more common form for five-coordinate manganese
complexes, but this geometry is most often observed for the Mnl" oxidation state due to the presence of the pseudo-Jahn-Teller distortion
axis, and not Mn".
The reactivity of model compounds is also key to gaining a better
understanding of an enzyme. The MnSOD presents a number of challenges, the first of which is catalytic dismutation of superoxide. Many
Mn compounds are known to interact (discussed later) with superoxide, but few have been shown to be catalytic in nature. The unique
characteristics of the catalytic cycle, with the "dead-end" complex,
also provides challenges for synthesis. Further, there has been some
interest in recent years in the use of manganese complexes as pharmaceutical superoxide dismutase agents to aid in reducing cellular
damage by superoxide, for example in the wake of a heart attack,
which opens questions of the biological compatibility interactions of
the complexes and their solution behavior at physiological pH values.
C . MANGANESE
PEROXIDASE,
MnP
Manganese peroxidase (MnP) is an unique enzyme in many respects. It is an extracellular enzyme that involves a heme protoporphyrin IX for the oxidation of Mn" to Mn"' (94,951. A crystal structure
at 2.06-A resolution of the manganese peroxidase from the white rot
basidiomycete Phanerocaete chrysosporium, which utilizes this enzyme to degrade lignin, appeared in 1994 (96).The active site (Fig. 5)
and the overall structure are quite similar to lignin peroxidase (Lip),
3 16
LAW, CAUDLE, AND PECOFURO
FIG.5. Structure of the active site of MnP. The left half is a side-on view of the
active site showing the protoporphyrin M, the bound proximal histidine, and the water
and histidine poised on the opposite face. The right half is a top-down view showing
the heme and the bound Mn, and the additional protein residues that act as ligands
to Mn. [Adapted with permission from (96).Copyright 1994, The American Society for
Biochemistry and Molecular Biology.]
an enzyme with which MnP shares a significant sequence homology

(96). The white rot fungi are some of the only species known that are
able to degrade lignin (97-99). Furthermore, MnP itself is one of only
two, or perhaps three, enzymes that are known to degrade lignin. The
related Lips are the other lignin-degrading enzymes that have been
known since the early 1980s (98, 99). Meanwhile, a few more recent
reports have suggested that some fungi may also utilize a laccase in
addition to MnP and Lip to facilitate lignin degradation (100, 101).
The MnP and Lip enzymes catalyze the biodegradation of lignin to
phenolic residues that can then be further decomposed. Lignin, found
predominantly in plant cell walls, is a random three-dimensional
polymeric material composed of phenolic residues that is thought to
be the second most common biopolymer after cellulose. Lignin comprises a significant portion of woody plants, with lignin content values
ranging from 15 to 36%(97-99,102). Lignin is also found in a variety
of other vascular plants. The major role of lignin in the plant is to
provide support and protection for plant tissue (98, 99). The lignindegrading peroxidases act on lignin by initiating one-electron oxidations that lead to the breakdown of the linkages that connect the
building blocks of this material. It has been shown that the decomposition of cellulose is retarded by lignin, probably as a result of lignin
317
interfering with access to the cellulose (98, 99). [As has been pointed
out, the importance of the degradation of lignin is underscored by the
extensive amounts of carbon employed in lignin and cellulose manufacture by plants and the interplay that this may have with respect
to the cycling of carbon on the globe (97-991.1 Thus, the Mn-catalyzed
degradation of lignin may be designed to allow the fungus access to
the cellulose (98, 99).
MnP functions by the oxidation of the heme by hydrogen peroxide
to form compound I, Ferv=O(P+')(Scheme 5). (95, 103-106). The oxidized heme then oxidizes a Mn" that is bound to at least one of the
oxalate
SCHEME
5.
two propionate arms of the heme group in a site located near the edge
of the enzyme, an observation based on the recently solved crystal
structure of the P. chrysosporium enzyme (96).This forms a Mn"' ion
and Compound 11, Fe(IV)=O. The Mn"' then diffuses away from the
enzyme to oxidize lignin. In this way, the enzyme can work on the
random structured polymer, where structural specificity may be a
problem for an enzyme and where Mn"' may diffuse more easily into
the substrate than a larger enzyme would be able to do. The Mn"'
initiates a one-electron oxidation, most likely forming an initial phenoxy radical cation, that triggers a cascade of oxidations that eventually break down the linkages between the lignin subunits. Compound
I1 is then reduced to the resting Fe"' oxidation state by oxidizing a
second equivalent of Mn". This step absolutely requires manganese,
preferentially a Mn"-oxalate complex (1051, unlike the first step,
which will function with substrates other than manganese (103).A
new addition to the chemistry of lignin peroxidases is the observation
that Lips are also capable of oxidizing Mn" to Mn"' in the presence of
oxalate (107). Thus, some Lips may also act as a manganese peroxidase.
Oxalate and similar small molecule chelators play a key role in the
3 18
function of MnP (104-109). Oxalate and malonate have been observed

to be excreted in active MnP-producing cultures (104), and they have
been shown to promote enzyme activity (105, 106, 108, 109). One of
the roles proposed for the small molecules is to aid in the diffusion
process of Mn"' and to stabilize the Mn"' from disproportionating to
Mn" and MntVso that it may reach the target lignin and initiate the
depolymerization of lignin (95, 103). It is not known whether oxalate
binds with manganese to the enzyme during the reduction of Compound I to Compound I1 or whether a small molecule is taken up to
assist departure of Mn"' from the enzyme. No evidence for oxalate or
other small molecules bound to manganese were observed in the crystal structure other than water. It is known, however, that without
the oxalate or a related small chelating molecule, turnover is reduced
(104-106). Furthermore, studies suggest that the manganese required for the reduction of Compound I1 must bind with the exogenous ligand oxalate (105, 106). Experiments with a-hydroxy acids
such as lactic acid or small organic acids such as malonic acid, showed
that such compounds also appeared to promote the activity of MnP
(95, 103). However, related compounds that could form stronger complexes with manganese were found to inhibit the enzyme, presumably
by out-competing the enzyme system for binding Mn". It was proposed that these compounds might facilitate the diffusion of Mn from
the enzyme and in that way lead to enhanced reactivity. Roles for
oxalate beyond the reduction of Compound I1 (e.g., assisting in the
diffusion of Mn"') remain controversial (104-107, 109). Oxalate is
known to be decarboxylated in the presence of Mn"', so it may not be
utilized in Mn"' diffusion, although that may be a role for the malonate. One other proposal for the role of oxalate was the reduction of
MnIV0 2 which
,
is known to form near these enzymes, to return the
manganese to a functional oxidation state. Some model work has addressed these points, and although unstable, such Mn'II-oxalate complexes may have lifetimes sufficient for them to reach the substrate
and initiate oxidation before they would completely decompose.
Finally, there is a functional similarity between the MnP and Lip
enzymes (110). Lip has been shown to be able to oxidize Mn" (107).
Further, it has been shown that a veratryl alcohol is produced in
conjunction with Lip, and it has been suggested that this compound
may serve Lip in a capacity similar to that served by Mn" in MnP
(109).
Functional mimic systems for the MnP enzyme will be discussed in

Section V, "Reactivity." An example of monomeric complexes prepared
with a-hydroxy acids as functional mimics is presented in Section
III.A.3 on monomeric structures.
3 19
D. MANGANESE
RIBONUCLEOTIDE
REDUCTASE,
MnRR
The conversion of ribonucleotides into deoxyribonucleotides for the
synthesis of DNA is key to the survival of any DNA-based life form.
This process is completed by one of a variety of ribonucleotide reductases that reduce the furanose ring of a ribonucleic diphosphate acid
monomer by replacing a hydroxyl functional group at the 2'-position
with a hydrogen (Scheme 6) to generate the deoxyribonucleic acid
Phosphates.
Phosphates.
*
HO
HO
OH
RNA
DNA
SCHEME
6
monomers necessary for DNA synthesis ( 111 -1 14). Ribonucleotide reductases have been identified with a wide variety of protein structures and active site compositions. Many are known that contain a
diiron core (in the oxidized form of FeRR, the active site contains a n
FeB' dimer bridged by a p-0x0 and a p-carboxylato), a stable tyrosine
radical, and potentially redox active thiols in the form of cysteines a t
the active site ( 1 1 4 , 115). The role of the diiron dimer is to generate
the stable tyrosyl radical, which in turn initiates a transient thiyl
radical at the active site (114, 115). Two other classes of ribonucleotide reductases have also been identified; one is based on adenosylcobalamin (114, 116) and the other, which is found in strictly anaerobic
microorganisms, likely utilizes a glycyl radical in conjunction with
S-adenosylmethionine to reduce RNA. These enzymes have been
grouped as Class I, Class 11, and Class I11 ribonucleotide reductases,
respectively ( 116). A fourth segment of these enzymes is represented
by the manganese-based ribonucleotide reductases, which probably
function in a manner similar to the iron ribonucleotide reductases
(discussed later) and, therefore, may fall under the Class I umbrella
(116).The MnRR enzyme is a bacterial enzyme, and is widely distributed among the Coryneform and related bacteria ( 1 1 1 , 1 1 7 , 118). Several members of this group of bacteria have been shown to absolutely
require manganese for the transformation of RNA into DNA. The
MnRR enzyme from Corynebucterium (formerly Brevibacterium) ammoniugenes has been the most studied member of this group ( 1 1 1 ) .
The FeRR system has been extensively studied (112, 114). For this
type of ribonucleotide reductase, the enzyme is comprised of two subunits, known as R1 and R2 (or occasionally B1 and B2). The R1 and
320
LAW,CAUDLE, AND PECORARO
R2 subunits are each comprised of a homodimer of the a- and p-polypeptides, respectively (112, 114), and both the R1 and R2 subunits
have been crystallographically characterized (119-122). The active
site for RNA reduction is located on the R 1 subunit. The oxidized
diiron site and the stable tyrosine radical are located in the R2 subunit. The diiron site is comprised of two Fe"' ions,obridged by an 0x0
and a carboxylate, with an Fe-Fe distance of 3.3 A. The reduction of
RNA begins with the abstraction of a hydrogen atom from the 3'position of the furanose ring by a transient thiyl radical. This is believed to be generated via an electron-transfer mechanism that allows
the tyrosyl radical to form a thiyl radical from one of the cysteines
located at the active site. In addition to that cysteine, a pair of redoxactive cysteines is located at the active site. This pair of cysteines is
then proposed to be oxidized to a disulfide-linked moiety in conjunction with substrate reduction. In the first step, a hydrogen atom is
abstracted by the thiyl radical from the 3'-position on the furanose
ring. A proton from one of the redox-active cysteines is transferred to
the hydroxyl group of the 2'-position to form a water-molecule-leaving
group. Completion of the oxidation to the disulfide and the regeneration of the initial thiyl radical by H-atom transfer to the 3'-position
of the furanose yields the DNA monomer product (Scheme 7 ) (114,
115). The disulfide linkage is then believed to be reduced by another
enzyme system to regenerate the active site (115).For several of the
MnRR systems, a thioredoxin/NADPH-dependent thioredoxin reductase system is thought to complete this phase of the process (123).
Although an X-ray crystal structure of MnRR has not been solved
to date, a structure with Mn substituted into the iron ribonucleotide
reductase has been reported (Fig. 6) (124). In this case, the two manganese(II1) ions were bridged by two carboxylates from protein residues, but without the 0x0 bridge postulated for the active Fe;" form of
the FeRR enzyme. This inactive enzyme with manganese was proposed as a model for the inactive Fe;' form of the iron enzyme and is
FIG.6. Structure of the Mn"-substituted FeRR enzyme, based on (124).
32 1
Phosphates,
DNA
Phosphates,
HO H20'
Phosphates,
O O M S E
H2O
Phosphates,
o ) W "
HO
HO
SCHEME
7. A partial proposed mechanism for the functioning of iron ribonucleotide
reductase. Manganese ribonucleotide reductase may function in a similar manner.
[Adapted with permission from (113)and ( 1 1 1 ) . Copyright 1994, Current Biology Publishers, and Copyright 1994 Marcel-Dekker, Inc., respectively.]
quite similar to the structure of the reduced Fekr active site which has
since been crystallographically determined. Based on UV-visible spectroscopy of MnRR and the similarity that this spectrum bears to that
observed for catalase and to model 0x0-, carboxylato-bridged Mn"' dimer complexes, the active site structure of MnRR is proposed to include the Mn-0-Mn structural motif ( 1 1 1 ) . This would also be similar to the active site structures observed for FeRR. This proposal is
322
further bolstered by the recent crystal structure of the catalase from

the bacterium Thermus thermophilus (125) (discussed later), which
generally corroborates the postulated catalase p-carboxylato-, p-oxobridged Mn dimer motif for the catalase active site structure. In addition, the assignment of a Mnf oxidation state is supported by the
EPR silence of this enzyme with respect to the observation of any
signal arising from a manganese dimer (111).
MnRR converts all four RNA bases into their corresponding DNA
products (111, 123), and the overall enzyme is considered to be structurally similar to FeRR ( 1 1 1 ) . For example, MnRR is comprised of R1
and R2 subunits that are analogous to those of FeRR. A mechanistic
similarity between these two systems, however, was not certain. No
radical signal equivalent to the FeRR tyrosyl radical had been observed for MnRR prior to 1996 (126).One tantalizing piece of evidence
was the hydroxyurea sensitivity of MnRR, a trait also observed for
FeRR. Hydroxyurea is known to quench the FeRR tyrosyl radical
(118). In the meantime, proposals for this lack of a radical signal had
been put forward. One proposal suggested that the coupling of a stable protein radical to the Mn of MnRR might account for the lack of
any observed radical by EPR in MnRR (111).Another suggested that
the manganese dimer might engage more directly in the RNA-to-DNA
conversion process, perhaps by abstracting the initial hydrogen atom
itself to initiate the reduction process, in a role distinctly different
from that proposed for the diiron site of FeRR (111). Proposals that
mirror FeRR, however, may now, in light of the radical signal that
has recently been observed in the MnRR R2 subunit, be reasonably
considered (126).Unlike the relatively stable FeRR enzyme radical,
though, the MnRR radical only has a half-life of 1.5 h. It was only
after developing a fast isolation technique for this subunit that
Auling and co-workers were able to observe this radical signal in the
EPR. By analogy, this radical is proposed to be located on tyrosine
residue. Finally, the presence and concentration of the radical signal
were shown to correlate with the activity of an MnRR sample (126).
Otherwise, little is known about the exact identity or proximity of this
radical to either the manganese dimer or the substrate binding sites.
E. MANGANESE
CATALASE
Catalase is yet another manganese-containing enzyme that interacts with dioxygen or its reduced forms. The role of this manganese
enzyme is to protect the organism from oxidative damage that may
be caused or initiated by the presence of hydrogen peroxide, such as
in the interaction of hydrogen peroxide with Fe in Fenton-like chem-
323
istry to generate H063. To date, manganese-containing catalases

have only been found in bacteria-for example, the enzymes from
Lactobacillus plantarum (127, 1281, Thermoleophilum album (129),
and T. thermophilus (130, 131). Several reviews with respect to the
catalase itself and the modeling chemistry of this enzyme have appeared in recent years (3,5 , 2 3 , 2 5 , 2 6 , 2 9 ,30, 38).
The catalase enzyme that utilizes manganese instead of a heme
group to catalyze the disproportionation of hydrogen peroxide t o dioxygen and water was known for several years before the presence of
manganese was confirmed in 1983 (128). A low-resolution crystal
structure in 1985 indicated that the enzyme from T. thermophilus
contained a binuclear manganese site, with a Mn-Mn separation of
about 3.6 A (130, 131). This was further supported by evidence from
optical spectroscopy that suggested that the active site contained two
manganese ions that were bridged by one 0x0, p-02-, and at least
one carboxylate ligand based on spectroscopic comparisons to 0x0- and
carboxylato-bridged manganese model complexes (30, 132, 133).
Extended X-ray absorption fine structure (EXAFS) studies of this enzyme suggested that the first coordination sphere about the manganese of the catalase consisted of O/N donors (134, 135). The presence of the dimer was further bolstered by the observation of EPR
spectra consistent with a coupled manganese dimer, for example the
MnMn EPR spectrum of the superoxidized enzyme (136).EXAFS
showed that this superoxidized catalase has a Mn-Mn vector of
2.7 A, consistent with a Mnz(p2-O2- core. Very recent reports of crystal structures at resolutions of 1.6 A for the reduced, Mnf, and 1.4 A
for the oxidized, Mn$, forms of the enzyme from T. thermophilus confirm much of what had been proposed for the structure of the catalases active site (125).A representation of these results is shown in
Fig. 7. The active site assembly contains two five- or six-coordinate
FIG. 7. Structure of the active site of the 2. fherrnophilus catalase enzyme, based
on (125).
324
manganese ions bridged by one carboxylate from a protein residue,

one water molecule, and a hydroxide ion in the reduced enzyme. The
hydroxide is suggested to be converted to an 0x0 bridge in the oxidized
form of the enzyme. One additional water molecule, carboxylates, and
imidazoles complete the first coordination spheres of the Mn ions. The
geometry about both Mn ions has been describtd in terms of square
pyramidal (125). The Mn-Mn vectors are 3.18 A in the reduced form
and 3.14 A for the oxidized form of the enzyme. Finally, this binuclear
manganese site is bound within a bundle of four a-helices. The T.
thermophilus enzyme itself is a homohexamer, composed of six identical subunits (125, 131).
During its catalytic cycle, the Mn catalases shuttle between the
MnP and the Mn$ oxidation states in 2 two-electron reactions (137).
Overall this process converts two hydrogen peroxide molecules into
one molecule of dioxygen and two water molecules. A proposal for the
catalytic cycle of catalase is presented in Scheme 8 (23). When the
, Y o
o
r
0
SCHEME
8. A proposed mechanism for the disproportionation of hydrogen peroxide
by manganese catalase, based on the mechanism proposed by Penner-Hahn in 1992,
and the 2'. thernophilus catalase crystal structure. [Adapted with permission from (24).
Copyright 1992 WILEY-VCH Verlag.]
T. thermophilus enzyme is isolated, a mixture of enzyme oxidation

states including MnH, MnIIMnIII, MniI', and Mn1"MnWis observed (138,
139). The MnIIMn"' and Mn"'MnTVoxidation states are catalytically
inactive but can be reduced to active oxidation states, Mn"Mn" or
Mn'IIMnIII, by hydroxylamine (23). The Mnl*'MnTV
oxidation state is
quite intriguing because it is thought to contain a bis-(p2-02-),core
which previously has been implicated in the oxygen-evolving complex
MANGANESE REDUX ENZYMES AND MODEL SYSTEMS
325
(23). The products of the "superoxidized" state may be explained if
one considers the possibility that occasionally during turnover an active site at Mnil' is reduced by only one electron to Mn"Mn"'. The
following two-electron oxidation step would then yield the Mn"'Mn'"
form of the enzyme, which is only slowly reduced back to an active
Mn!" form. Such chemistry was shown to be viable by the introduction
of the one-electron reductant hydroxylamine to catalase followed by
the addition of hydrogen peroxide (23). These catalases exhibit saturation kinetics and complete the disproportionation of hydrogen peroxide at a rate of 5.75 X lo5 M-l-s-' for L. plantarum (231, 1.7 X lo6
M-. s-l for T. album (129) and 3.2 X lo6 M-'. s-' for T.thermophilus
(131, 140). Finally, the catalase enzymes have been shown to be inhibited by azide, chloride, fluoride, and thiocyanate (23). In the case
of fluoride inhibition, there is significant controversy both for the
number of fluorides that are bound and whether or not the anion acts
as a terminal ligand or a bridging p2-F- species.
Numerous model complexes that model one or more traits of the
catalase enzyme have been prepared over the years (discussed later).
Several systems that mimic the function of this enzyme have been
reported, and a selection of these will be discussed in Section V, "Reactivity." Other models with respect to structures or physical properties of the catalase enzyme will be presented in the ensuing sections
of this review.
F. BINUCLEAR
MANGANESE
ENZYMES:
A COMPARISON
OF THREE
ENZYMES
One of the striking features of the two redox-active binuclear enzymes described earlier, MnRR and catalase, and the hydrolytically
active enzyme arginase is the apparent similarity of their active site
structures. Yet all three exhibit vastly different reactivities. Both the
MnRR and catalase exhibit UV-vis spectra consistent with carboxylato- and 0x0-, hydroxo-, or aquo-bridged cores. This would indicate
that both have similar Mn-Mn core structures. An EXAFS study by
Stemmler, et al. (9) compared the active sites of these three enzymes.
Those studies suggested that arginase and the iron ribonucleotide reductase substituted with Mn had similar structures that tended to be
six coordinate. Their data for catalase suggested that the manganese
of this enzyme were likely five coordinate. No crystal structure for
MnRR has been solved to allow for a specific identification of ligands
or a direct comparison of distance parameters and ligand orientations. However, it is possible to compare the arginase (8)and catalase
crystal structures in this manner (125). Other comparisons have ap-
326
peared in the literature (9, 10).These environments are quite similar,

as can be seen from Figs. 1 and 7, and the potential correlation between the two sites is strengthened by the low level of catalase activity exhibited by arginase (9, 10). The similarity of the bridging motifs
and protein residues as ligands is striking. Both have enzyme-active
sites that contain dimers of Mn ions bridged by a bidentate carboxylate and two ,uZ-"O"donor bridges: a monodentate carboxylate and a
hydroxide (crystals grown at pH) in arginase, and a water molecule
and hydroxide in the reduced catalase. The remaining ligands to the
metal ions on each side of the dimer are identical in terms of type,
carboxylates, and imidazoles, and in terms of their general positions
about the manganese dimer. In the case of arginase, one Mn ion is
five coordinate, the other is six coordinate due to a bidentate carboxylate. The catalase contains a water in the area of what would be the
corresponding open site on the five-coordinate arginase Mn, and the
carboxylate donor on the second manganese is monod$ntate instead
of bidentate. The Mn-Mn vector for arginase is 3.3 A compared to
3.14 and 3.18 for the MnH and Mnil' oxidation states of catalase.
Presumably, the native MnRR will also have a Mn-Mn vector in the
3.0- to 3.5-A range.
The broader structure about the Mn cores of these sites of the arginase and catalase enzymes, however, does vary considerably from one
enzyme to the other. Both of these enzymes are located within hydrophobic pockets. However, in the case of the active-site structure of
catalase, there is an additional glutamate that may hydrogen bond to
two of the imidazoles that are bound to the manganese and a lysine
NH3+group is also hydrogen bonded to a manganese ligand carboxylate (125). In arginase, there are only two other polar residues reported to be near the active site beyond those that are bound to manganese (8). One of these is a histidine imidazole, which may function
in a proton shuttle capacity. The other is a glutamate, which probably
salt-bridges to the guanidinium functional group of the arginine substrate. The protein structures near the active sites are also quite different. In catalase, theobinuclear active site is located within a fourhelix bundle about 18 A from the surface at the bottom of a channel.
Overall, the catalase enzyme is composed mostly of helical structural
motifs. Meanwhile, the arginase enzyme's active site is not located
within a four-helix bundle, and this enzyme contains several helix
and P-Fheet structures. The active site of arginase sits at the bottom
of 15-A-deep cleft that is located within an area of turns and loops
between helices and P-sheets.
The reactivity of catalase enzymes and rat liver arginase may also
now be directly compared. Arginase has recently been reported to dis-
327
proportionate hydrogen peroxide, albeit much less effectively than the

actual manganese catalases (in a buffered solution a t pH 9.5) (9, 10),
albeit much less effective than the actual manganese catalases, exhibiting a rate of reaction on the order of 30 s - I versus 200,000 5-l for L.
plantarum (10).It was also reported that evidence for the manganesesubstituted iron ribonucleotide reductase exhibiting a low level of catalase activity might also exist (9).No "superoxidized form of arginase
could be formed in the presence of hydrogen peroxide and hydroxylamine, but it did show hydroxylamine inhibition. Another difference
between these systems was a lowered sensitivity of arginase to the
inhibitory effects of halides (10). A final point of comparison is the
reported peroxidase activity of arginase. The catalase from T. album
has been shown to complete a peroxidase reaction on p-phenylenediamine (129). Arginase was shown to perform peroxidase activity on
this compound and o-phenylenediamine (10).Thus, as one might expect from their structural similarities, these enzymes can in some
cases catalyze similar chemistry.
Thus, despite the initial similarities of the immediate active site
dimers, a broader inspection of the enzyme begins to suggest potential
explanations for the differences in their reactivity. For example, the
existence and type of residues near the active site, with respect to
acidhase chemistry, that may attenuate the overall reactivity of the
core will play a role. The types of bridging p"0" moieties, monodenate carboxylate and hydroxide in arginase versus hydroxide and water in catalase, probably play a significant role. The hydroxide and
the proposed 0x0 (or perhaps two hydroxides) of the MniI' form of the
enzyme may be better able to support the higher oxidation state on
manganese versus the carboxylatehydroxide bridges of arginase. The
two five-coordinate manganese of catalase may play a role, too, if rearrangement were required for the activation of arginase.
Overall, these enzymes exemplify a growing theme in bioinorganic
chemistry, that of nature utilizing an active site structural motif in
more than one enzyme to catalyze quite similar or dissimilar chemistries. Such reapplication of a structural motif has been observed in
the recurring themes of binuclear iron ( 1 4 1 ) ,for example. For manganese, it appears that a general binuclear manganese active site design
has been reapplied three times in separate settings for at least three
different purposes.
G. THEOXYGEN-EVOLVING
COMPLEX,
OEC
Dioxygen is important to all aerobic life forms. It is required for
respiration and, therefore, the release of energy required to sustain
328
biological processes. Dioxygen is a by-product of photosynthesis and

is produced by the oxidation of two molecules of water to dioxygen in
conjunction with the release of four equivalents of protons and four
equivalents of electrons, which will be processed into reducing equivalents suitable for eventual carbohydrate biosynthesis. The catalytic
site for water oxidation is known as the oxygen-evolving complex
(OEC) of Photosystem I1 (PS 11),located in the chloroplasts of green
plants and algae (42, 48). The OEC, sometimes referred to as the
WOC, water oxidation complex, or water oxidase, contains a group of
four manganese ions at its active site (42).Calcium and chloride are
also required for the proper function of the OEC (35).Part of a complex system of several polypeptides and cofactors, which are membrane-bound in the thylakoid membrane of green plants, the OECs
role is to provide the reducing equivalents necessary to rereduce a
specific photoexcited chlorophyll a molecule that initiates the chain of
events leading to the production of reducing equivalents. The polypeptides of PS I1 and those surrounding the OEC serve in various roles,
many of which have been probed and elucidated ( 4 2 , 4 8 , 5 1 ) .Some of
these roles relate directly to the manganese cluster of the OEC and
the other inorganic cofactors. Based on the wealth of research in this
field related to PS 11, a gross structural similarity of green plant and
algal PS 11s is believed to exist with that of the crystallographically
characterized (142) bacterial (Rhodopseudomonas) PS 11s (42, 143)
that do not employ the OEC. The commonalities between these systems has led to a proposed orientation of the various components of
PS I1 (Fig. 8) (24).The OEC is located near the base of two polypeptides known as D1 and D2 and is surrounded and stabilized by several additional polypeptide units (42, 144).
FIG.8. A scheme of the generally proposed layout of the components of Photosystem

11. The polypeptides are labeled, too. The electron flow in PS I1 follows the arrows
shown. [Reproduced with permission from (24).Copyright 1998 Springer-Verlag.1
329
For more detailed discussions of the biochemical aspects of the PS

11, such as the role of the various polypeptides or differences between
the PS IIs, the reader is directed to one of several reviews and compilations. The OEC and PS I1 have been extensively reviewed over the
past several years. The review by Debus in 1992 sums up much of the
knowledge obtained on PS I1 to that point (42). The Photosynthetic
Congress, held every few years, has produced several volumes of material from a broad spectrum of researchers covering the whole of photosynthesis with current results to the time of that Congress (51). A
1996 book edited by Ort and Yocum (48) contains several chapters
covering various aspects of photosynthesis, from PS I1 to PS I. Manganese modeling chemistry has also been extensively covered by reviews
(24,26-31,34, 37, 45,50).
The OEC is involved in the earliest stages of the overall photosynthetic process, which produces NADPH for carbohydrate biosynthesis.
Overall, the photosynthetic system is composed of two halves, PS I1
and PS I (48). PS I1 provides reducing equivalents that are transferred to PS I and are eventually converted into the reducing energy
of NADPH. Both PS I and PS I1 absorb light energy to initiate their
respective electron flows. The protons released during these steps are
likely utilized in the generation of a proton gradient for the synthesis
of ATP, which is also required for carbohydrate biosynthesis (48).
It has been established that the OEC contains a tetranuclear cluster of manganese ions, and that it requires at least one calcium ion
(39, 145, 146) and at least one chloride ion (147, 148) to complete its
catalytic cycle. This catalytic cycle involves four successive oxidations
of the OEC (42, 48) to eventually yield a highly oxidized system that
then completes the oxidation of two molecules of water with the concomitant release of dioxygen and the return of the OEC to an overall
oxidation state four levels lower. The sequential oxidation process
starts when light energy is harvested by antennae chlorophyll molecules, which transfer this energy to the Psso chlorophyll a molecule of
the reaction center. This photoexcited P680 then reduces Pheophytin A
to give a charge-separated state of Ps+Boand Pheo-. Pheophytin A then
reduces a bound plastoquinone molecule, QA, which then reduces a
dissociable plastoquinone molecule, known as QB. Once Q B has been
doubly reduced and doubly protonated, it departs to carry the collected reducing equivalents further into the photosynthetic system
(Fig. 8). Once oxidized, Psso+ needs to be reset so that new reducing
equivalents may be initiated, and to prevent charge recombination, or
back-reactions, from occurring. This is accomplished by oxidizing the
OEC. This oxidation is mediated by a particular tyrosine residue,
known as Tyrosine Z, Y;,residue 161 of the D1 polypeptide (149,150).
330
This tyrosine is initially oxidized by P&, and then it in turn oxidizes

the OEC by one electron. In this manner, the OEC is oxidized successively four times, whereupon dioxygen release occurs, the OEC returns to its initial oxidation state, and the successive oxidation cycle
begins anew. Four protons are generated in conjunction with the electron withdrawal and dioxygen release, and these are likely utilized in
the generation of a proton gradient for ATP synthesis (42, 144). The
cycle of four progressive oxidations followed by release of dioxygen
with the concomitant reduction of the manganese center of the OEC
has been termed the S cycle, or S clock (151-153) (Scheme 9). This
2 H20
SCHEME
9.
clock is a five-stage cycle that proceeds from So to S4, with So being

the most reduced form of the enzyme whereas S4represents the most
oxidized form. Dioxygen release occurs during the S3-S4-So phase of
the cycle, presumably once S4 is reached (152, 153). S , is relatively
unstable and rapidly returns to the So state (46).The initial evidence
for the S cycle was reported by Joliot (151).Those experiments were
reproduced and reinterpreted by Kok and co-workers (153), who
showed that, for photosynthetic centers given a sequence of single
flashes of light, a burst of dioxygen occurred with the third flash of
light and a burst of dioxygen then occurred with every fourth flash
thereafter (Fig. 9). The initial burst came on the third flash, indicating that PS 11s stored in the dark will equilibrate t o the S1 state,
which is one step oxidized from the OECs lowest overall oxidation
state. Further research has tied at least some S-state progressions to
an increase in the overall oxidation of the manganese cluster of the
OEC (discussed later).
In the absence of crystallographic data, all of the structural infor-
33 1
2.4
30
FLASH NUMBER
-I
wh
1.6
0"
Q
= 0.0 -
p:
0
0 2 EVOLUTION IN FLASHING LIGHT
12
18
FLASH NUMBER
24
30
FIG.9. The Four Flash experiment. The peaks in the plot represent the dioxygen
bursts that occur with every fourth flash of light administered to a PS I1 sample. This
cycle peaks on the third flash due to the resting state of the enzyme being a t S , . The
process is eventually dampened by double hits of light leading, for example, to an inhomogenous sample with respect t o the S state. [Reproduced with permission from ( 4 6 ) .
Chapter 9. Copyright 1996 Kluwer Academic Publishers.]
mation known about the OEC has been gathered by the use of spectroscopy (24,28,33,42,46). The use of X-ray absorption spectroscopy
( U S ) has provided some particularly useful quantitative structural
information. EXAFS (extended X-ray absorption fine structure) is an
X-ray absorption spectroscopic technique that provides distance information and some idea about the nature of the first coordination
sphere of the element in question. EXAFS studies have shown that
the OEC manganese ions are present in a predominantly O/N-donor
coordination environment (24, 154). Furthermore, EXAFS data show
that there exists at least one Mn-Mn vector, and prob%bly two, of
2.7 A, and a third significant vector is at a distance of 3.3 A (155-159,
477), which may result from a Mn-Mn or Mn-Ca interaction (discussed later).
EPR spectroscopy has been another extremely important means for
exploring the OEC (24, 28, 42, 46). The S, state is readily prepared
332
by dark-adapting preparations of PS IIs, from which Sz is achieved by

flashing the sample with one pulse of light (42).The Sz state exhibits
at least two distinct EPR signals, which have been the focus of extensive research and controversy, too. The first, the multiline, is a
multiline signal centered around g = 2 (160, 161), and the second,
which depends on sample preparation technique and conditions, appears a t g = 4.1 (162-166). These EPR signals may be interconverted,
and are shown in Fig. 10. The multiline signal is indicative of magnetically coupled manganese ions, with at least 18-20 resolved lines being observed. It also indicates that at least two of the manganese are
involved in a multinuclear cluster. Based on the expected manganese
oxidation states, and more lines than the 12-16 traditionally observed for a MnlIIMnNdimer, a cluster of at least three manganese
ions is probable (46).The second EPR signal at first appeared similar
to that generated by an isolated MnN, leading to trinuclear/mononuclear proposals for the structure of the OEC. Later EPR evidence
showed that this signal did not result from a monomeric MnIVion, but
rather most likely arose from a multinuclear manganese structure
itself (167, 168). In ammonia-treated, oriented PS I1 samples, a
multiline signal was observed in the g = 4.1 EPR spectrum. Further-
'...,.,..,....(,...,...,,....,....
lo00
ZOO0
3000
Maptic Field (G)
....,'
40oo
FIG.10. The g = 4.1 EPR signal from the SS state of the OEC (left) and the g = 2
multiline signal (right). [Reproduced with permission from (46), Chapter 9. Copyright
1996 Kluwer Academic Publishers.]
333
more, another study on this signal suggested that it arose from an

S = 512 ground state (1691,which would be inconsistent with a Mn"
ion, but not a multinuclear array of manganese. The appearance of
the multiline EPR signal upon going from S1 to 9, also supports the
existence of metal-centered oxidation (160, 161). For a more detailed
discussion on these two signals and the conditions affecting their respective formation, the reader is again directed to one of the various
reviews and compilations (42, 46).
In 1997, new EPR data on the So state was reported by Messinger
and co-workers (170, 171). Until these reports, no EPR spectra had
been recorded for this S state. The first set of experiments took darkadapted PS 11s to Soby the addition of reductants (171). This S state
is labeled S$ to account for potential alterations to the manganese
structure. A multiline signal EPR signal was observed for this S$
state that exhibits 24-26 lines centered near g = 2 [Fig. ll(a)l. This
signal is wider than the S , multiline and exhibits more lines, many of
which do not line up with transitions observed in the S , multiline.
The authors suggest that this added breadth is a result of the lower
overall oxidation state of the manganese ions composing Soand that
this may be indicative of the presence of a Mn"Mn"' couple based on
model complex data. A second set of experiments reported by Messinger, Sauer, Klein, and co-workers (170) took dark-adapted samples
through one S-state cycle to poise them at So.This preparation also
presented a multiline signal, which did not appear to differ significantly from that observed for S$ [Fig. ll(b)]. For both of these spectra,
the presence of methanol (0.5-1.5%) in the sample was required to
observe the multiline signals. Without methanol, a broad EPR signal
is still observed for the So preparation. The authors theorize that
methanol may be binding to the manganese cluster and that this facilitates the observed spectra, perhaps by altering the coupling of the
manganese ions. Simulations conducted to ascertain what the possible oxidation states of the manganese ions in So might be have not
yet clarified this point, because the overall oxidation states Mn"Mnin,
Mn"Mn1I1MniV,or Mni*'MnIVall gave similar spectral widths consistent with the data. However, the simulations did suggest that neither
binuclear species nor a tetranuclear cluster arranged as MnI1Mnw/
MnlIIMnIVwould not yield comparable spectral widths. Also in 1997,
Styring and co-workers (172) reported a multiline signal for So.Their
study showed that by administering light flashes to PS I1 samples,
they could produce a multiline signal that they believed to arise from
Soafter 3 flashes from S z samples. This spectrum is also wider than
that observed for S2, similar to that reported earlier. A second se-
334
A
3 2.8 2.6 2.4
g value
2
22
1.8
1.6
at
B
I
250
300
350
Field (mT)
400
p=2
1900
2700
3500
4300
5100
Magnetic Field (Gauss)
FIG. 11. Sa and SoEPR signals. (a) EPR Spectrum of the Sg state. The top spectrum,
A, represents the S-l state that was formed by reducing PS I1 samples with hydrazine.
This was illuminated to yield a new multiline signal, C. B is the spectrum of the Sz
state multiline from control experiments. Details of sample preparation may be found
in the references provided. (b) A and B. Difference spectra of prepared So minus residual S1 signal. B. For comparison, the Sz multiline under the same conditions. Note that
So is a broader signal and is shifted with respect to the Sz signal. C. Sospectrum without
added methanol. D. Sg*Spectrum produced by reducing S1samples with hydroxylamine.
E. Simulated spectrum of So. [(a) and (b) reproduced with permission from (171) and
( 170), respectively. Copyright 1997 the American Chemical Society, respectively.]
quence of light flashes regenerated the So signal. They also utilized

methanol in their samples.
EPR signals have also been reported for the S, state (173). A very
intriguing article (174) has recently appeared in which Britt, Debus,
and co-workers showed, via the application of parallel polarization
EPR studies, that a multiline EPR signal can be observed for the S1
state (Fig. 12). When illuminated, the amplitude of this signal was
greatly reduced, and a conventional perpendicular polarization g =
2 multiline signal consistent with the formation of the S, state was
335
(a) Parallcl Polarization
?P
zoo
ooo
1MM
1200
800
Magnetic Field (G)
600
(b) Perpendicular Polarizatlon
Illuminated minus Dark
9#
1
2s00
Moo
3500
Magnetic Field (G)
4000
4500
FIG.12. S, multiline signal. In (a) the spectrum (parallel polarization EPR) labeled
Dark represents a dark-adapted PS I1 sample; thus, it is in the S, state. The second
trace represents the sample following illumination with light to generate S2. The top,
Dark, spectrum minus the middle, Illuminated, spectrum yields the third spectrum in
(a). This represents a well-resolved multiline signal arising from a multinuclear exchange coupled paramagnetic Mn cluster (174) in the S, state of the OEC. If the subtraction is reversed and the spectra are recorded using perpendicular polarization, then
when the Dark spectrum is subtracted from the Illuminated spectrum, an S2 multiline
signal is readily observed (b). [Reproduced with permission from (174). Copyright 1998
the American Chemical Society.]
336
observed. This S, multiline signal is consistent with the interaction of

multinuclear clusters of manganese ions, and these authors proposed
that it arises from a cluster of manganese ions. This proposal would
minimally require that a trimeric cluster of coupled manganese ions
exists, versus an arrangement invoking two distinct dimers, one a
M$' dimer and the other a Mny, an arrangement based on the assumed overall integer spin of this S state. Furthermore, because Sl is
expected to be integer spin, a trimer/monomer motif would suggest
that the monomeric manganese should be in the Mn"' oxidation state.
If that were true, then a six-line spectrum consistent with a Mn"'
should be observable under the conditions employed to observe this
new S1multiline signal. The lack of such a signal is suggestive that
this new multiline signal actually arises from a tetranuclear cluster
of manganese ions.
Application of the X-ray Absorption Spectroscopy (XAS) to the observation of Mn K-edge data (24, 32,33, 156, 1751, via X-ray Absorption Near Edge Structure (XANES), has provided insight into the
overall oxidation state of various S states. The inflection point of the
edge shifts for the various oxidation states of manganese (24, 33).
Studies of the OEC show changes in edge energies consistent with an
increase in the overall oxidation state of the manganese cluster. Thus,
at least some of the oxidations of the OEC occur on the tetranuclear
cluster of manganese ions. Numerous XAS studies in conjunction with
EPR data have lead to proposals of the oxidation levels of the OEC S
states (24, 33, 42, 46). Two of these proposals are presented in
Scheme 10. The S-state oxidation levels of the manganese are not
s-state
so
s,
s2
s,
s.4
IIInIInIIIiv
IrinIinvnv
riinvnvm
rvnvnvnv
tvnvnvN
IIIIIIIIIIIIIV
IIInIInvm
IIIIIVIIVIIV
IIIIIIIIVIIV
nmnvnv
IVIIVIIVIIVY I.
SCHEME
10.
clear-cut, and some controversy does remain ( 2 4 , 3 3 , 4 2 , 4 6 ) .Much of

this controversy revolves around defining oxidation states for the S
states from spectroscopic techniques. For example, the transitions
from So to Sl and S1 to S2 have strong evidence for manganese-based
oxidation state changes. However, the nature of the S2to S, transition
is quite controversial because there is no clear-cut alteration to the
observed XANES edge energies. Proposals to explain this have invoked oxidized protein residues or structural rearrangements of the
337
tetranuclear cluster (176). The imidazole oxidation that had been suggested, however, is unlikely. The signals that were attributed to the
oxidized imidazole have now been shown to arise from Y; (177, 178).
In the final S , to S4to So phase, little is certain due to the instability
of the S, and the transient nature of the Sqstate. Again the oxidation
state controversy is centered on whether oxidation occurs on a protein
residue, such as Y;, or on a manganese of the OEC.
The results of the XAS and EPR data, combined with knowledge
from model complexes, (discussed later) allow for the formulation of
structural models of the OEC. Many manganese compounds have
been proposed as structural models for the OEC, ranging from butterfly arrangements to adamantanes to cubanes (26,27,29-34,50). Most
of those proposals, however, are inconsistent with the known structural data for the OEC. Currently the most widely favored model is
the dimer of dimers model (Scheme 111, first suggested by Klein,
SCHEME
11.
Sauer, and co-workers (32, 154). This model proposes that two manganese dimers that interact with one another exist in the OEC. The
dimers are suggested to contain Mnz(pz-02-)2
cores and t o be bridged
to one another by 0x0 and carboxylato units. This model is supported
by several pieces of current data on the OEC. First, the 2.7-w Mn-Mn
vectors of the OEC are most consistent with crystallographically
characterized dimeric manganese model complexes that contain
Mnz(pz-02-)z
cores, and two such vectors are observed for the OEC.
Second, such a model would provide the potential for magnetically
coupled Mn-Mn interactions to occur. Bridging two Mn2(pz-OZIz cores
)
would proto one another by carboxylates and a single ( p 2 - 0 2 -unit
vide for an approximately 3.3-w distance, a formulation consistent
with known model compounds (discussed later). Furthermore, the
338
EPR data from So, S , , and S2indicate that there are coupled Mn ions.
In particular, the multiline signals are suggestive of coupled manganese structures that require clusters of greater than two manganese
ions.
The related Electron Nuclear Double Resonance (ESEEM) and
Electron Spin Echo Envelope Modulation (ENDOR) spectroscopies
have also proven useful in elucidating information about protein residues near the OEC. One of these residues is the important protein
radical Tyrosine Z, Y;, which is usually found in an oxidized radical
form (42).Study of this protein residue via ENDOR and ESEEM spectroscopies has been enlightening. The presence and protein residue
associated with this radical signal was defined several years ago, and
Y; was originally assigned the role of an electron-transfer mediator
for the electrons moving between the OEC and
(150, 179). More
recent work has been modifying this view. Tyrosine Z has been shown
by Britt and co-workers to be located in close proximity to the manganese cluster, probably within 5 A (180).Probes of the nature of this
tyrosyl radical by Babcocks group have suggested that this tyrosine
may play a more active role than solely to act in an electron-transfer
capacity (181,182).It does not bear the hallmarks of a protein residue
whose sole purpose is to participate in an electron-transfer chain-for
example, it does not retain a nearly rigid orientation with respect to
its surroundings. Other similarities between this redox-active tyrosine and other metalloenzyme systems that employ radical components suggested that this tyrosine might also act in such as role (47,
183). These observations have lead to recent proposals by Britt and
more intimately in the water oxidation mechBabcock that involve Y;,
anism (182, 184,185).Instead of merely acting as an electron transfer
agent between the OEC and P&, it is now proposed to perform a
hydrogen atom abstraction. New models for the function of the OEC
have been formulated based on this proposal (discussed later). The
role of Yz is still not settled, of course, as evidenced by a recent article
arguing against a hydrogen atom abstraction role (186).Finally, spectroscopic studies have also focused on identifying potential protein
residues that might be ligands to the OEC. Through ESEEM spectroscopy, the presence of at least one imidazole ligand to the OEC manganese cluster has been established (187).
Studies involving calcium and chloride have shown these ions to be
critical t o the proper functioning of the OEC (35, 146, 1881, although
their exact roles have yet to be elucidated. Without either of these,
dioxygen evolution is halted and the OEC cannot complete one full
cycle. It has recently been reported, for example, that the OEC can
reach Sz without chloride, but cannot be further oxidized to S , o r pro-
339
ceed from S3 to S, without the addition of chloride (188). There has

been speculation that the chlorides role may be to prevent premature
loss of oxidizing equivalents, perhaps by preventing the premature
loss of oxidation equivalents in the formation of hydrogen peroxide
(189).Another suggestion may relate chloride to substrate binding,
an idea that has received some support from small-molecule-binding
studies, which indicate that there is a site for a small molecule such
as amnonia to bind to the OEC in a process that displaces chloride
(35, 190). Calcium may be bound near the manganese and has been
proposed to be the source for the Mn-X 3.3-A distance observed by
EXAFS (191). One possible role for the calcium may be to act as a
water-binding site from which water is transferred at some point to
one of the four Mn ions of the OEC (192, 193). Various researchers
have studied the effects of substituting for these critical ions. Only
strontium has been successfully inserted in place of calcium with retention of any level of function, albeit severely attenuated (35, 194).
Substitution for chloride is less sensitive. The addition of bromide
does not appreciably alter the ability of the OEC to generate dioxygen
(35,421. The OEC is also functional, but at reduced levels in the presence of iodide and nitrate. Fluoride strongly inhibits the OEC and has
been suggested to alter the structure of the manganese cluster structure based on EXAFS data (195). Amines also replace the chloride
and inhibit the function of the OEC (35, 190). Ammonia is a particularly interesting case in this regard, in that it appears to bind at two
different sites within the OEC, suggesting that there may exist a second chloride-inaccessible site, which has been proposed to be a waterbinding site to a manganese (35, 196).
Water is of course key to this whole process. Water-binding studies
have been attempted to establish when the two water molecules that
are oxidized to dioxygen become bound to the OEC and whether they
are bound directly to manganese. Early water-binding studies suggested that the water converted to dioxygen is still exchangeable until
at least S, . From isotopic labeling studies, it appears that the waters
that are oxidized are not bound or highly restricted with respect to
exchange until the late S states (197). A recent water-binding study
suggested that there were two waters in exchange with the bulk water, but that these waters were exchanging at different rates (198).
One proposal put forward to explain this observation was the existence of an Mn=O moiety. A review based on the thermodynamics of
water oxidation was published in 1996 (37).
Pulling the data together again, a larger picture of some to the
structural and reactivity details of the OEC begins to emerge. The
new hydrogen atom abstraction role proposed for Yz coupled with the
340
dimer of dimers model has led to some intriguing new hypotheses for
the potential manner in which the OEC functions (46, 47, 177, 178,
181,182,184,185).A synthesis of these new concepts is presented in
Scheme 12 (182). If the OEC functions in the newly suggested manner, then it would join a growing class of enzymes that utilize protein
a', H
-!-
7kl+
Y',
-7Y-',
't.1"
SCHEME
12. A mechanism for dioxygen evolution by the OEC as proposed by Babcock
that employs the hydrogen atom abstraction concept. [Reproduced with permission
from (182). Copyright 1998, the American Chemical Society.]
residue radicals to complete their catalytic processes (47, 183). The

dimer of dimers structural model has been modified t o include hydrogen atom abstraction from a water bound to a manganese by Y ; ,
which is equivalent to proton loss being coupled to the stepwise oxidation of the OEC. This proposal eventually leads to the generation of a
high-valent manganese (IV)or (V) with an 0x0 bound, Mn=O, that
then combines with a second water-derived oxygen atom bound to a
nearby manganese to form dioxygen during the S4-So step. Another
feature of these new models is the retention of charge neutrality
throughout the process. Scheme 13 shows the Babcock proposal as
modified by our group. For any functional proposal to gain validity,
one needs to consider whether or not the energetics of the process
involved would be favorable. Recent studies by Pecoraro and co-workers suggest that the energies required for this process to occur are
likely to be favorable (discussed later).
+'%
2 H,O+ CI-
341
e- ,H+
YZ'
/e-.H+
s2
SCHEME
13. A mechanism for dioxygen evolution by the OEC involving hydrogen
atom abstraction a8 proposed by Pecoraro.
The OEC provides perhaps the most complicated manganese enzyme system yet known and thus is rich with possibilities for the
exploration of manganese chemistry. Overall, the OEC and PS I1 are
among the more intricate enzymatic systems that are currently being
explored. Structural models, spectroscopic models, and reactivity
model complexes are all integral to insights that have been attained
and those yet to be gained. The following paragraphs detail some of
the work in this area. Models related to water oxidation are presented
in the Section V, "Reactivity." A selection of the newer multinuclear
complexes will be discussed in Sections 1II.B to 1II.D. Again, many
reviews have dealt with model complexes for the OEC (29-31, 34,
45, 50).
Many bioinorganic chemists have prepared speculative structural
models over the past several years to gain a n insight into this system-for example, with respect to possible arrangement of manganese ions-or to provide characterized materials as reference compounds for spectroscopic examinations. The agreement between
EXAFS data and the Mn-Mn vector of a Mn&-O), core is one prime
example of structural modeling chemistry and spectroscopy combining to elucidate critical data on this system (discussed earlier). Multinuclear clusters are always of interest in this field, and several have
been prepared (28-31, 34, 45). The adamantane and cubane structures led to proposals for the possible oxidation of water to dioxygen
342
that involved rearrangement of a manganese cubane to adamantane

upon incorporation of water molecules, and another proposal invoked
a butterfly structure that was converted to a cubane during the catalytic cycle (199, 200). These proposals, among others, have been extensively reviewed, and the reader is directed to one of the referenced
sources noted earlier. New examples of dimers of alkyl-bridged
Mn2(pn-O),cores have been prepared and structurally characterized
in recent years (discussed later). A few of these also include waters
bound to the manganese. One of the more interesting tetranuclear
manganese structures in this class of compounds was reported in
1992 by Armstrong and co-workers, who then further probed its complicated magnetic couplings (201).
A variety of other complexes have been prepared that have proven
useful in the area of the spectroscopic understanding of the OEC.
Many have been used as models for XAS studies to establish what is
to be expected from the various oxidation states of Mn. Others have
been probed to gain a better understanding of the interactions between manganese nuclei-for example, most Mn"'MnTVcomplexes
that are known exhibit 16-line EPR signals (140). However, the Mn"'
Mn'" complex prepared with the 2-OHsalpn ligand (202) only exhibits
a 12-line spectrum. Another area that model complexes may help to
address is the effects of water binding to a manganese cluster. Small
molecules such as water can be bound t o the complexes derived from
2-OHsalpn, and these have been probed by ENDOR spectroscopy to
ascertain what perturbations might arise from that bound molecule
versus the complex without such an additional ligand (203).
The energetics of water oxidation is another key point to address.
This topic includes the interesting question of how the oxidizing
equivalents are stored in the OEC without being released too soon,
and how the requisite 3.6 V versus NHE for water oxidation can be
built up when Y; is thought to maximally be able t o provide oxidations at a level of 1.1 V per oxidation (41).In addition, the proposal
of hydrogen atom abstraction has led t o the exploration of how the
homolytic bond dissociation energy of a water bound to a manganese
dimer might be altered from the values observed for an individual
water molecule in solution. To abstract a hydrogen atom from water
would require more energy than is available to Y; (Table I). However,
Pecoraro and co-workers have shown that the AG value for removing
a proton from a protonated p-0x0 bridge is within a reasonable range
for the Y; to be able to abstract a hydrogen atom (Table I) (204,205).
These studies employ a concept worked out by Mayer for the oxidation
of, for example, toluene by permanganate (206). Studies of water
343
TABLE I
HYDROGEN
BONDDrssOCiArION ENERGIES"
Complex
HBDE (kcal/mol)
Reference
86
119
76
77
79
84
457
458
204
204
204
204,459,460
89
94
85
89
82
86
205
205
205
205
205
205
Tyrosine
Water
IMn'"(salpn)(p-0)I2
LMn'"(3,5-di-C1-salpn)(p-O)12
[MniV(3,5-di-(N02)salpn)(p-0)12
[Mniii'TV(bpy)(p-O)l~+
LMn$"(2-OH(3,5-diClsal )pn),(OHJl
[Mn'i'MnTY(2-OH(3,5-d~Clsal)pnJ2(OHL)1'
IMn~"(2-OHsalpn),(OH2)]
IMn'1iMniV(2-OHsalpn)2(OH,)1
LMnYi(2-OH(3,5-di-t-Bu~alJpn)~(OH~)l
~Mn"iMn"'(2-OH(3,5-di-t-Busal)pn)2(OH2)l
For complete details, including complex pK,'s and electrochemical reduction potentials, the reader is directed to (204)and (205).
bound to complexes derived from the 2-OHsalpn ligand show that hydrogen atom abstraction from these complexes is also in a reasonable
energy range, theoretically, for H-atom abstraction by Y; (205). In
addition, the normally observed increase in reduction potential of this
complex with an hydroxide bound is attenuated. This also suggests
that if H-atom abstraction is occurring in the OEC, this might favorably affect the redox potential of the manganese dimer to which that
water is bound, further stabilizing the system against early loss of its
oxidizing equivalents before water oxidation can occur and thus
allowing the potential for oxidation of the manganese cluster to remain in a range accessible to Y;. A report of quantum chemical calculations in 1997 supports these data (207).
Other topics of importance to the OEC will be discussed in greater
detail later. Water oxidation models will be addressed in Section V,
"Reactivity," and models for the catalase-like reaction of the OEC will
be addressed in Section V.B.3., "Catalase."
Manganese redox enzymes exhibit a range of structural motifs involving mononuclear or bi- and tetranuclear aggregates of manganese
ions. There has been a great deal of research in the area of structur-
344
ally modeling the active sites of manganese enzymes such as the

MnSOD, catalase, and the OEC. These small-active-site analogs have
been prepared in the hope that structural mimics will also mimic (i)
the spectroscopy and (ii) the reactivity of the biological center. Some
success has been achieved in modeling the reactivity of catalase and
SOD with these models, but there is no structural analog for the OEC
that exhibits water-oxidizing activity. The following is an overview of
additions in the field of manganese chemistry since the superior review by Wieghardt in 1989 (30).This section will provide a compendium of the structure types that have appeared in the literature over
the past eight years but is not meant to be an exhaustive recapitulation of every structure that has been generated during that time.
A. MONONUCLEAR
1. M n (II)
Mononuclear complexes of manganese are quite common and have

been prepared in all of the currently recognized biologically relevant
oxidation states of manganese, II-V. Mononuclear Mn" complexes exhibit a range of polyhedra about the metal, with coordination numbers of 5-8 dominating. The lack of coordination number specificity
is understandable, as this d6 ion does not, in theory, exhibit specific
geometric preferences. Numerous complexes have been prepared with
predominantly N,O,-donor atom sets, which are often augmented by
halide. or pseudohalide ligands and the occasional sulfur donor. In
general, this represents the known biologically relevant donor atoms
to manganese. Several of these complexes have been generated with
the target of mimicking or modeling MnSOD.
A variety of five-coordinate complexes of Mn" have been prepared.
Crystallographically characterized Mn complexes with a trigonal bipyramidal geometry are still relatively rare, however. Kitajima and
co-workers (208) reported two trigonal bipyramidal complexes utilizing hydrotris(3,5-diisopropylpyrazo1-l-yl)borate,HB-i-Pr, a facially
binding tridentate ligand (209,210)(Fig. 13).The coordination sphere
of the first complex was completed by a bidenate benzoate; the other
complex contains a monodentate benzoate and one equivalent of 3,5diisopropylpyrazole. Another variation that has recently appeared is
composed of tbp complexes of Mn" with tris(benzimidazoy1-2-methyl)amine bound t o the manganese (211,212) (Fig. 14).The tridbenzimidazoyl-2-methyllamine is a tetradentate tripodal amine ligand that
provides four N-donor atoms. The first coordination sphere of these
complexes was completed by a chloride. However, these products are
345
FIG. 13. [Mn~"~hydrotris(3,5-diisopropylpyrazo1-l-yl)borate)(3,5-diisopropylpyrazole~

(benzoate)] perchlorate. [Reproduced with permission from (208). Copyright 1993 the
American Chemical Society.1
not purely five-coordinate compounds. The crystal structure of each

bears both a five-coordinate form with one C1- bound to the central
Mn and one C1- counterion, and a six-coordinate complex with two
chlorides bound to the central Mn (Fig. 14). Another five-coordinate
complex appeared in 1990 (213) with a single bis(benzyimidazolyl-1methy1)amine ligand and two acetate molecules (Fig. 15). The amine
ligand binds meridionally, with the two acetate ions completing the
equatorial plane of the structure. A few Mn" complexes with square
pyramidal geometry were also reported. In one case, the basal plane
is occupied by two dihydrobis(pyrazo1-l-y1)borate ligands with an
axial chloride (Fig. 16) (214).In another, the axial position is occupied
by an axial THF molecule with two THF molecules and two phenoxy
donor groups in the basal plane (215).
The greatest number of complexes characterized are those that are
six coordinate. One general structure type involves multidentate li-
FIG.14. Crystal structure (left) of [MnYtris(benzimidazoyl-2-methyl)amino)C1It,

the
trigonal bipyramidal complex. Crystal structure (right) of the neutral six-coordinate
[Mn"(tris(benzimidazoyl-2-methyl)amine)C121.
[Reproduced with permission from (211)
and (2121, respectively. Copyright 1997 the American Chemical Society and copyright
1995 Elsevier Science, respectively.]
346
FIG.
mission
with per-
gands that fill four of the six available coordination sites, allowing
monodentate ligands such as halides or water to occupy the two remaining positions in the first coordination sphere. In complexes that
involve bidentate ligands such as phenanthroline, bipyridine, or acetylacetone, which often do not bind opposite to each other in the
equatorial plane of these compounds, a structural type in which the
two remaining coordination sites are cis to one another is promoted.
One example of these bis bidentate chelate complexes is the complex
(ditriflatolbis(bipyridin0) Mn (216) and bis (phenanthrolino)(dithiocyanato) Mn (Fig. 17) (217). One of the more unique examples in this
group involves a fused-ring system, which provides a tetradentate N4donor ligand that occupies three coordination sites on one face and
one addition site on the other. The two remaining cis positions were
filled by bromides and chlorides, ultimately leading to distorted octahedral structures (Fig. 18) (218).This group also includes complexes
derived from the tris(benzimidazoy1-2-methy1)amine ligand mentioned earlier (Fig. 14) (211,212).
When the chelating ligands are able to occupy the equatorial
plane of a Mn complex at the same time, the monodentate ligands
FIG. 16. Square pyramidal [Mnl(dihydorobis(pyrazol-lyl)borate)C1]. [Reproduced

with permission from (214).Copyright 1990 the Royal Society of Chemistry.]
347
S
FIG. 17. IMn"(phenanthr~line),(SCN)~].[Reproduced with permission from (217).
Copyright 1993 International Union of Crystallography.]
will bind to the central Mn" ion on the remaining two trans "axial"
positions. Variations on this theme include charged chelating ligands
with a pair of neutral ligands axial or equatorial neutral chelates,
such as a tetraazamacrocyle, for example, with charged axial ligands,
such as halide ions (Fig. 19) (219). Yet another variation found for
these complexes is one in which all of the ligands about the central
Mn" are neutral, for example the complex [bis(bis(2-pyridylmethyl)
amino)MnI1l2+(Fig. 19) (220).
A less common but rapidly growing family of Mn" complexes are
those that are seven coordinate. These complexes may be built up
from a variety of ligand systems that range from multiple ligands to
a single septadentate ligand. Some of these systems have been tested
as MnSOD mimics. Most of the seven-coordinate complexes that were
structurally characterized in recent years have adopted a pentagonal
bipyramidal geometry. One motif is based on a pentadentate macrocylic ligand with two unidentate ligands in the axial positions. One
particular family of pentagonal bipyramidal complexes was prepared
by Riley and co-workers from the macrocyclic ligand 1,4,7,10,13-pen-
FIG. 18. Mn"(2,5,8,10,13,16-hexaazapentacyclo~8.6.l,1'~5.09~'H.O'"~'7]o~tadecane)Cl~.

[Reproduced with permission from (218). Copyright 1989 Verlag der Zeitschriften fur
Naturforschung.]
348
FIG. 19. [Mn(l,4,7,10-tetraazadodecane)C12],

left, and [Mn(bis(2-pyridylmethyl)
amine)z]2+,
right. [Reproduced with permission from (219)and (2201,respectively. Copyright 1996 International Union of Crystallography and Copyright 1992 American
Chemical Society, respectively.]
FIG. 20. Two seven-coordinate Mn complexes that have been tested as MnSOD
mimics. [Mn11(1,4,7,10,13-pentaazacyclopentadecane)Cl~]
(left) and [Mn(trans-2,3eyclohexano-1,4,7,10,13-pentaazacyclopentadecane)Clz](right). [Reproduced with per mission from (221) and (222). Copyright 1994 and 1996 the American Chemical Society, respectively.]
FIG.21. [Mn(tris(ethan~l)arnine)~].
[Reproduced with permission from (217). Copyright 1993 Verlag der Zeitschriften fur Naturforschung.]
349
FIG.22. [Mn"(tris(benzimidazoyl-2-methyl)amino)(N0,)Z1.
[Reproduced with permission from (211).Copyright 1997 the American Chemical Society.]
taazacyclopentadecane (221, 222) (Fig. 20). The first coordination

sphere in these complexes is completed by two axial chlorides. This
work was specifically focused on the development of MnSOD mimics,
and these workers have generated a wide range of complexes by preparing a series of ligands in which the macrocyclic ligand was modified by additions to the ring's carbon skeleton (222). Another example
involves a modified 15-C-5 ligand with axial di-t-butylnaphthylsulfonates (223),and a related complex has been prepared with two axial
trifluoromethylsulfonates (224). Pentagonal bipyramidal complexes
do not, however, require macrocyles to form. Other pentadentate but
open-ended ligands have also proved to be well suited for the generation of seven-coordinate complexes (225-227). Figure 21 shows an example of a complex with an MnLz motif, wherein one tris(ethano1)
amine ligand binds as a tetradentate ligand, while the other binds as
a tridentate ligand with one unbound ethanol arm. Another example of
CN 7 is that of the tetradentate ligand tris(benzimidazoy1-2-methyl)
amine and Mn" (Fig. 22) (211), wherein the other three binding sites
are occupied by one monodentate nitrite and one bidentate nitrite
molecule. A final example is provided by the complex of Mn" with
quaterypyridine, one water, and a bidentate acetate (Fig. 23) (226).
FIG. 23. [Mn"(quaterpyridine)(OAc)(H,O)l+.[Reproduced with permission from

(226).Copyright 1993 Elsevier Science.]
350
FIG.24. [Mn~'(1,4,7,10,13,16,19-septaazacyclouneicosane)l~+.
[Reproduced with permission from (228). Copyright 1990 the American Chemical Society.]
Two seven-coordinate complexes were attained by utilizing single

septadentate ligands. The first involved 1,4,7,10,13,16,19-~eptaazacyclouneicosane, but no specific geometric orientation about the
Mn" was reported (Fig. 24) (228). The second structure is of a monocapped trigonal antiprism, prepared by using the septadentate ligand
formed by Schiff-base condensation from tris(2-aminoethy1)amine and
2-pyridinecarboxaldehyde t o give tris(2-aminoethy1)amine-pendant
pyridyl-2-methyl functional groups (Fig. 25) (229). This complex was
tested for MnSOD activity (discussed later).
A final structural motif for Mn" is represented by a few eight coordinate structures. One of these was a product of tetraazacyclododecane with pyrazolylyl-l-methyl groups appended to the ring nitrogens
FIG.25. [Mn"(tpaa)12+.[Reproduced with permission from (229). Copyright 1996 the

American Chemical Society.]
351
FIG.26. [Mni'~1,4,7,10-(l-pyrazolylmethyl~tetraazacyclododecane)]2+
(left) and [Mn"
(tris(2-pyridylmethyl)amine)12+(right). [Reproduced with permission from (230) and
(232).Copyright 1992 the Royal Society of Chemistry and copyright 1993 Elsevier Science, respectively.]
(Fig. 26) (230).This led to a Mn" complex having a square prismatic

geometry. The second complex is composed of a cryptate ligand that
yields a complex with a cubic symmetry (Fig. 27) (231).The third
structure is built up from two tetradentate tripodal amine ligands
about the Mn" ions (Fig. 26) (232).
One final note is of interest. Most of the structurally characterized
and otherwise isolated complexes of Mn" tend to be high spin. The
structure of at least one low-spin complex was reported during the
past few years. This is a six-coordinate complex composed of two
methyl-4,5-dihydro-5-{[imino(methylthio)methyllazo}-3,5-dimethyl1H-pyrazole-1-carboximidothioatoligands, which provide an N, first
coordination sphere via one aza-nitrogen and two carboximido-nitrogens (Fig. 28) (233).
FIG.27. [Mnii([2.2.2]cryptand)]'+
[2.2.2]cryptand = 4,7,13,16,21,24-hexaoxa-l,lO-diazabicyclo[8.8.8]hexaxosane. [Reproduced with permission from (231).Copyright 1992
Wiley-VCH Verlag.]
352
LAW, CAUDLE, AND PECOFURO
FIG.28. Low-spin [Mn"~methyl-4,5-dihydro-5-{~imino~methylt~o~methyllazo}-3,5-dimethyl-lH-pyrazole-l-carboximidothioate)1.

[Reproduced with permission from (233).
Copyright 1993 the Royal Chemical Society.]
2. Mn(III)
Five- and six-coordinate structures dominate the crystallographically characterized structures reported for Mn"'. Here the pseudoJahn-Teller axis of this d4ion plays a role in the overall structure of
the Mn"' coordination complex, so these complexes tend to adopt either a square pyramidal or octahedral geometry. Most of the five-coordinate complexes that have been structurally characterized are found
to adopt a square pyramidal geometry. A prime example of such products are the myriad of Mn"' complexes that utilize a tetradentate
Schiff-base ligand (234),for example H2-Salpn or H2Salen,with a fifth
donor in the axial position t o complete a square pyramidal structure.
Such structures are also observed in the Mn"' porphyrin complexes
with a single axial ligand (470). This structure ably supports the
pseudo-Jahn-Teller axis along the axial, or z, axis. Specific examples
of this group of complexes are the complexes utilized for exploring
epoxidation chemistry (235).One of these complexes is shown in Fig.
29 (236),with a structure similar to the other members of this type
of complex (237).For this type of complex, the Mn'" ion tends to be
located out of the equatorial plane toward the axial ligand. Another
area of Mn"' complexes are those that utilize a variety of porphyrins
as ligands. Five-coordinate complexes were not limited to the square
pyramidal geometry. At least two reports of structurally characterized trigonal bipyramidal complexes have appeared in the literature
in recent years. The first is of a complex in which the Mn"' is surrounded by an 0, first coordination sphere composed of two bidentate
2,2'-biphenoxide ligands and one monodentate 2,2'-biphenoxide to
create a distorted trigonal bipyramidal geometry (Fig. 30) (238).The
353
FIG.
29. {Mn1''[3,5-~di-t-buty1~Sa1icylidenimino-~~~)-(1,2-diaminocyclohexane)l}.
[Reproduced with permission from (236).Copyright 1996 Wiley-VCH Verlag.]
second is a very intriguing structure with a tegradentate tripodal

amine ligand (239).Three of the five donor atoms are nitrogens from
amide functional groups (Fig. 31). The unique feature of this complex
is the hydroxide bound to the Mn"' ion. In general, structurally characterized complexes of manganese with hydroxide tend to be rare.
Six-coordinate complexes of Mn"' are also quite common. Many of
these, again, were prepared with tetradentate Schiff-base ligands
that occupy the equatorial plane of the complex while two other ligands bind to the central manganese ion on the remaining trans axial
positions (234, 240). Structures with Mn"' were not, of course, exclusively prepared with tetradentate Schiff bases or similar ligands,
however (241). A unique six-coordinate structure was reported in
1992 for a MnIII-OH complex, another rare example of a structurally
characterized Md'I-OH complex (Fig. 31)(242).
FIG. 30. [Mni"(2,2'-bisphenoxide)(2,2'-bisphenoxideH)l.[Reproduced with permission from (238).Copyright 1991 the American Chemical Society.]
354
od
FIG. 31. Rare examples of Mn"'-OH-complexes: [Mn"'(tris(cyclopropylcarbanoylmethyl)amine)OH]- (left) and [Mn'11(bis(2-hydroxy-5-nitrobenzoiminopropyl)methylamine)OH]. [Reproduced with permission from (239) and (242). Copyright 1997 and
1992 the Royal Society of Chemistry, respectively.]
The structure of one rather key six-coordinate Mn"' complex in this

area was reported by Kitajima and co-workers in 1994. This complex
is composed of the hydrotris(3,5-diisopropylpyrazol-l-yl)borateligand,
one monodentate 3,5-diisopropylpyrazole,and a side-on-bound peroxide molecule (Fig. 32) (243).This structure is important with respect
to the question of the "dead-end" complex that is proposed to occur
during turnover in MnSOD. It is also a rare example of a structurally
characterized Mn-peroxide moiety. One of the key features of this
system is the two differing structures that were isolated. In the second structure (Fig. 32), the peroxide is presumed to be hydrogen
bonding to the extra pyrozole ligand, based on the shortened interaction between the hydrogen on the nitrogen of the pyrazole ring and
the oxygens of the peroxide.
3. M n ( W
A less common but growing class of complexes are the mononuclear

complexes of MnW.The structurally characterized MnW,d3,complexes
tend to adopt six-coordinate structures that are octahedral in nature.
Typically, these MnWstructures require charged ligands to support
the higher oxidation states. This is also true of the Mn" complexes. In
most of these complexes, phenolato, alkoxo, or amido donors have
been employed to stabilize these higher oxidation states. Among these
are the complexes of MnTVsalpn(acac)(PF6)(244) (Fig. 33), which is
octahedral with the salpn2- ligand in a cis-beta conformation, and a
MnW(salpn)C12
(245)complex with an equatorial salpn2- and two axial
chlorides (Fig. 33). An example of an O6 first coordination sphere is
355
FIG. 32. [Mn~"(hydrotris(3,5-diisopropylpyrazol-l-yl)borate)(3,5-diisopropylpyrazole~

peroxide]+ Two different structures were grown at low temperatures. In one structure
(bottom) there is a hydrogen-bonding interaction between the 3,5-diisopropylpyrazole
ligand and the peroxide moiety. The hydrogen-bonding interactions are indicated in the
expanded figures to the right. This is a n important complex with respect to modeling
manganese redox enzymes. [Reproduced with permission from (243). Copyright 1994
356
FIG. 33. [MnTY(salpn)(acac)](PF6)

(left) and MnN(salpn)C12(right). [Reproduced with
permission from (244)and (245).Copyright 1991 the American Chemical Society and
copyright 1995 the Royal Society of Chemistry, respectively.]
provided by the Mnw complexes with a-hydroxy acids that were prepared to model manganese peroxidases (Fig. 34) (246).A unique hexadentate ligand, based on octamethyltetraamine with salicylic acid
moieties appended to the terminal nitrogens to form amide functionalities, produces the mononuclear Mn'" complex shown in Fig. 35
(247).
The macrocycle 1,4,7-triazacyclononane (2481,tacn, has been a popular ligand in the manganese community in recent years. Two interesting complexes based on this ligand with N303first coordination
spheres have been reported. One Mn" example is the MnTVtacn
(OCH3I3PF6
complex (249),which has been suggested as able to epoxidize olefins. This complex adopts a structure with a facial array of
methoxides due to the nature of the other, facially binding tacn-based
ligand. The second is one in which the alkoxides are appended by
alkyl arms to the tacn ligand, similar to the structure in Fig. 36. That
structure is represented by the complex with an N6 first coordination
FIG. 34. [MnN(2-hydroxyethylbutyric acid)J. [Reproduced with permission from

(246).Copyright 1991 the American Chemical Society.]
357
FIG.35. ~Mn'V~l,l0-bis(salicylamido)hexamethylenetetraamine~].
[Reproduced with
permission from (247).Copyright 1992 the American Chemical Society.]
sphere and the triply deprotonated ligand 1,4,7-tris(o-aminobenzyl)1,4,7-triazacyclononane (Fig. 36) (250).
4. MnW)
Structurally characterized Mn" complexes are rarer still. The new

examples of mononuclear manganese complexes in this oxidation
FIG. 36. [Mn'v(1,4,7-tris(o-aminobenzyl~-1,4,7-triazacyclononane~l+.

[Reproduced
with permission from (250).Copyright 1995 the American Chemical Society.]
358
FIG.37. [MnV~N,N'-bis~salicylideneimino~-2,4-dimethyl-2,4-butanediamine~~N)].
state have been of nitrodo, N3-, and Mn-oxo derivatives. Most of

these adopt a square pyramidal geometry with the organic ligand occupying the axial plane while the 0x0 or nitrido group occupies an
axial site (251-253). A manganese-nitrido complex with this orientation appeared with the ligand sale+ and the same ligand with all of
the ethylene backbone's protons replaced by methyl groups, saltme+
(Fig. 37) (251). Both the MnV=O group and the MnV=N groups were
first characterized in porphyrin systems, but only the MnV=N has
been structurally characterized (Fig. 38) (254). More recently, two
MnV=O complexes have been crystallographically characterized
(255-257), for example the complex shown in Fig. 39. A unique feature of all of these MnVcomplexes is the short Mn-nitrido and Mn-0
bonds. The Mn-N bonds are on the order of 1.50-1.54 A, and the
Mn-oxo bonds have been determined to be about 1.54-1.55 A. Most
of these complexes are square pyramidal, but some variation is now
known. A recent report of a reputed trigonal bipyramidal Mnv-nitrido
complex composed of two bidentate ligands has appeared (2531,and a
six-coordinate MnVcomplex was produced by Wieghardt and co-workers in 1996 (258) (Fig. 40) that employs a trimethyl tacn ligand, one
acetylacetonate ligand, and a terminal nitrido. The Mn-N (nitrido)
distance in this complex is 1.518 A. One nitrogen donor atom from
the tacn ligand occupies the axial position opposite t o the nitrido at
FIG. 38. [MnV(OEP)(N)].[Reproduced with permission from (254). Copyright 1983

359
FIG.39. [MnV(LKO)I-.[Reproduced with permission from (256).Copyright 1990 the

an elongated distance of 2.301 8 compared to the other two N donors

from this ligand at 2.055 8 and 2.073 8. The structures of both of
these types of Mn" complexes, Mn"=O and Mn"-N, are important
structural models for epoxidation and aziridination reagents, and for
the OEC with respect to the possible formation of a Mn"=O during
turnover.
B. BINUCLEAR
The majority of model compounds synthesized for studying the biomimetic chemistry of manganese are binuclear complexes, due to the
FIG.40. [MnV(Me3tacn)(acac)N].
[Reproduced with permission from (258). Copyright
1996 the American Chemical Society.]
360
relative ease and reproducibility of their preparation when compared

with complexes of higher nuclearity. There are a number of binuclear
manganese enzymes for which binuclear complexes can function as
direct structural models (discussed earlier). Furthermore, higher nuclearity clusters in model chemistry and biological chemistry can often be described as assemblies of well-characterized binuclear units.
Within the myriad of binuclear manganese complexes synthesized to
date, a few core structural motifs appear with regularity. Each core
is distinguished by a characteristic Mn-Mn distance that is relatively
invariant with changes in the supporting ligands. As such, the
Mn-Mn distance, where it is known, is often used as evidence of a
particular core structure when a full crystal structure is not available.
This is often the case in the biological systems themselves, which
have been studied by X-ray absorption spectroscopy and from which
Mn-Mn distances are determined even in cases where no crystal
structure is available. Table I1 lists the binuclear manganese complexes whose cores have been structurally characterized along with
their respective Mn-Mn separations. Although a multitude of dimanganese complexes have been prepared within each core structure, we
will restrict the discussion in this section to those that have been
structurally characterized by X-ray crystallography or X-ray absorption spectroscopy.
TABLE I1
GENERAL
TRENDSIN MN-MN DISTANCES
OF SOME
COMMON
BRIDGING
MOTIFS
Bridging motif
Oxidation state Mn-Mn Distance (8)

2.8
2.6-2.7
2.6-2.7
3.2-3.3
3.1-3.4
3.1-3.2
3.25
3.6
3.3
3.1-3.6
2.6
3.4-3.5
3.5-3.6
References are given in text or are in (280).
36 1
The simplest binuclear structure possesses two manganese ions

bridged by a single 0x0 or hydroxo ligand. Because dioxo- and dihydroxo-bridged manganese dimers tend to be thermodynamically stable, the mono-oxo and monohydroxo complexes are isolated and characterized only with sterically demanding ligands that do not provide
two labile manganese coordination sites in a cis orientation. Only
three complexes with biologically relevant ligands have been structurally characterized, these have two manganese bridged by a single hydroxide. The only nonporphyrin manganese dimer bridged by a single
hydroxo linkage is the binuclear complex of the macrocyclic ligand
tpictn (Fig. 41) (259).The Mn-Mn separation in this highly supported
structure is 3.6 8, with a Mn-O-Mn angle of 126.8'. Mn2(tpictn)
(p-OH) is a special case, however, in that the Mn,(p-OH) core is supported by a macrocyclic ligand that spans both Mn" ions. The only
unsupported complexes having a single hydroxide bridge are the Mn"'
porphyrin complexes [Mn"'(OEP)],(p-OH) (260) and [Mn*1*(TPP)12
(p-OH) (261) (Fig. 42). These complexes exhibit Mn-O-Mn angles of
152.7' and 160.4, respectively, consistent with the bridging hydroxo
group. The more linear geometry of the TPP derivative is probably
due to steric demands of the pendant phenyl groups, resulting in a
longer Mn-Mn distance, 3.993 8, for this complex than for the OEP
derivative. [Mn"'(5-NOzsaldien)lz(p-O)
(262) has a single oxide ligand
bridging two Mn"' ions and has a Mn-Mn distance of 3.490 8 (Fig.
Fic. 41. [Mnb'(tpictn)(p-OH)l. [Reproduced with permission from (259). Copyright

1995 the Royal Society of Chemistry.1
362
Q
FIG.42. [Mn"'(OEP)],(p-OH)+ (left) and [Mn"(TPP)N3lZ(p-O)(right). [Reproduced
with permission from (261)and (265).Copyright 1996 the American Chemical Society
and copyright 1981 the Royal Society of Chemistry, respectively.]
43). The hydrotris[(3,5-diisopropyl)pyrazolyllborate derivative [Mn"'

(HB(3,5-iPr2pz)2(3-iPr0-5-iPrpz))12(p-O)
(263) (Fig. 43),has a similar
Mn-Mn distance, 3.53 A, and a more linear Mn-0-Mn geometry explained by steric demands of the hindered HB(3,5-iPrzpz)2(3-iPrO-5iPrpz) ligand. The MnTV
complex with tetraphenyl porphyrin forms an
0x0-bridged complex, [MnW(TPP)N3l2(p-O),
with a nearly linear Mn0-Mn angle (264,265) (Fig. 42).
Analogs of p-OH-bridged complexes having a single p-alkoxo ligand
are known as well. In fact, when analogous p-OH and p-OR bridged
complexes are structurally characterized, the structural parameters
are often quite similar. At this time the only structurally characterized complexes having a single p-OR ligand involve the highly supported 2-OHsalpn ligand motif (266-268) For example, [Mn1Y2-OH-
FIG. 43. [Mn1'1(5-N02saldien)]2(p-O)and [Mn"'(HB(3,5-iPr2pz)2(3-iPrO-5-iPrpz))lz

(p-0).[Reproduced with permission from (262)and (271).Copyright 1990 and 1991 the
American Chemical Society, respectively.]
363
FIG. 44. [Mn'"(2-OH-(5-Clsal)pn)],(CH30H)

and [Mn"'Mnw(2-OH-(3,5-(Cl)zsal)pn)2
(THF)]'. [Reproduced with permission from (270).Copyright 1992 the American Chemical Society.]
(5-Clsal)pn)lZ(CH30H)(Fig. 44) has a Mn-Mn distance of 3.808'

and a Mn-O-Mn angle of 129" (269), within the range expected for
Mn"12(p-OH) complexes (discussed earlier). The analogous mixedvalent [Mn"'Mn" (2-OH-(3,5-(Cl)zsal~pn)2(THF)lt
complex also has a
364
single alkoxide bridge (270) but with a shorter Mn-Mn distance of

3.65 8, and a Mn-O-Mn angle of 126.7"(Fig. 44).
Bis-p-oxo dimanganese complexes are the most prevalent structurally characterized dimanganese motif in the literature. A smaller
number of bis-p-hydroxo- and bis-p-alkoxo-bridged analogs are also
known. Because of the prevalence of similar structures differing only
in the supporting ligands, binuclear manganese bis-p-0x0, bis-p-hydroxo, and bis-p-alkoxo complexes are most conveniently considered
as discrete manganese cores in which the supporting ligands play a
stabilizing role. This view is supported by the remarkable similarity
in core structure among complexes having very different ligands. We
will begin by discussing lower valent states and continue with increasing oxidation state of the manganese in the binuclear cluster.
[Mn11z(HB(3,5-iPrzpz)3)]2(0H)z,
with a Mn-Mn distance of 3.314A, is
to date the only crystallographically characterized binuclear manganese complex with two hydroxo bridges (271) (Fig. 45). When
the bis-p-oxo complex
this complex is oxidized by Mn04- or 02,
[Mn"'tHB(3,5-iPrzpz>3)l~~O)z
is formed (Fig. 45).This oxidized complex
has a much shorter Mn-Mn distance of 2.696 8, between two fivecoordinate manganese"' ions. The six-coordinate complex [Mn"'(bispic-
b
FIG.45. [Mn~'(HB(3,5-iPr,p~)a)l~(OH)~
(top) and [Mn'1'(HB(3,5-iPrzpz)3)12(0)2
(bottom).
365
FIG.46. [MnT"(bispicMepen)l~(p-Oj.
Me2en)12(p-O)
(Fig. 461, has nearly the same Mn-Mn distance, 2.699
elongation axis common in Mn"' ions does
not include a bridging ligand (272). No bis-p-hydroxo dimanganese(II1) cores have been structurally characterized, but the analogous bis-p-methoxo dimanganese(II1) complex [Mn"'(salpn)],
(p-OCH,), (Fig. 47) has been prepared and exhibits a 3.192 A Mn-Mn
distance (244). The longer Mn-Mn distance for the alkoxide-bridged
complex reflects the Jahn-Teller elongation axis on each of the symmetry-related manganese ions, which includes a bridging alkoxide ligand. The supported [Mn"'(2-OHsalpn)12(Fig. 48) has two alkoxide
bridges as well, but with slightly longer Mn-Mn separation due to
the highly supported ligand structure (273,274).
Mixed-valent dimanganese(II1,IV) complexes with the Mn"'Mn'"
(~-0
core
)~
have structural parameters very similar to the corresponding MII?(~-O)~
complexes, with Mn-Mn distances in the 2.7-A range.
In spite of the similarity to homovalent cores, the prototypical [Mn"'
(276) comMn1v(bpy)4(p-0),13+
(275) and [Mn"1Mn'V(phen),(p-O)213+
A, because the Jahn-Teller
FIG.47. IMn"i(3,5-diC1-salpn)(p-OCH,)I,.
[Reproduced with permission from (244 j.
Copyright 1991 American Chemical Society.]
366
FIG.48. [Mn"'(2-OH(5-NOz)salpn)]z.
[Reproducedwith permission from (273).Copyright 1993 the American Chemical Society.]
plexes are valence localized in the solid state, with one manganese
exhibiting the axial elongation associated with isolated Mn"' centers
and the other having more octahedral geometry consistent with Mnrv
complexes exhibit valence localization
(Fig. 49). Other Mn111Mn1V(p-0)2
as well (272,277,278). The only example of a mixed-valent Mn1I1MnW
(2791, which
bis-p-alkoxo-bridged complex is [Mn111MnTV(2-OHsalpn)zl+
FIG. 49. [Mnr1LMnw(bpy)4(p-O)2]3t.

367
FIG. 50. [Mn'V(salpn)]z(p-O)z.[Reproduced with permission from (244). Copyright

1991 the American Chemical Society.1
has a Mn-Mn separation of 3.3 A, substantially longer than for bisp-0x0 complexes in the same oxidation state.
The higher valent Mnr(p-O), complexes have been prepared with
many supportin ligands, but all show the characteristic Mn-Mn
distance of 2.7 (34).The prototypical complex [Mn'Ysalpn)lz(p-O)~
is shown in Fig. 50 (280, 281). Each successive protonation of [Mn
[MnTV(salpn)l
,(p-O)(p-OH)' and [MnW(salpn)l,
( ~ a l p n ) ] ~ ( p - to
O )give
~
(p-OH),t was shown by X-ray absorption spectroscopy to result in an
increase in the Mn-Mn distance of 0.10 A (282, 283). The structural
effect of protonation on the Mniv(p-O)zcore is considerably less pronounced than alkylation of the corresponding MnY1(p-0)2complexes
due to rearrangement of the Jahn-Teller axis in the Mn"' case. A new
structure type for Mnbv(p-O)zSchiff-base complexes was reported for
the complex [MnlV(salen)]z(p-O)z
ligand (234, 284,285). In these complexes (Fig. 51), the salen ligand bridges the two manganese ions.
FIG.51. [Mn1v(salen)]2(~-0)2.
1998 Elsevier Science.]
368
FIG.52. ([Mn'"(tacn)ll(~-O)~(~-Ol)}
(left) and [Mn'"(ta~n)l&-O)~
(right). [Reproduced
with permission from (286)and (287).
Copyright 1988 and 1990 the American Chemical
Society, respectively.]
There are two unique examples of triply bridged Mny dimers that
deserve comment at this point. The first is the [MnTV(tacn)lz(p-O),
com lex with three bridging 0x0 ligands and a Mn-Mn distance of
2.3 (Fig. 52) the shortest Mn-Mn distance of any binuclear complex
t o date (286). The second is {[Mn'v(tacn)lz(p-O)~~p-Oz)}
(Fig. 52), which
is the only example of a binuclear Mn complex having a p-1,2 bridging peroxide ligand and a Mn-Mn distance of 2.531 A (287). These
two complexes may be considered as MniYp-O), cores with either an
additional p-0x0 or the p-1,2-peroxo added as a third ligand. Because
of the additional ligand, the MniV(p-O),core is puckered in contrast
to simple MnY(p-O):!cores. The importance of the latter structure is
that it may be the best model to date for the initial manganeseoxygen species formed immediately upon 0-0 bond formation in the
critical S4-So transition in photosynthetic oxygen evolution.
There is a large and important class of binuclear manganese complexes having at least one p-carboxylate bridge in conjunction with
one or two p-0x0 bridges. This class of p-oxo-p-carboxylato complexes
is important not just in the context of binuclear manganese enzymes
such as catalase [which has recently been shown to contain such a
core (12511, but as analogs for iron oxo-carboxylato complexes, which
are very important in the chemistry of nonheme iron redox enzymes.
The simplest example in this class of manganese compounds is represented by the p-oxo-p-carboxylato system [Mn"'(bispicen)lz(pU-O)(p~,~OAc) with a Mn-Mn separation of 3.28 A (288) (Fig. 53). This core
motif has been structurally characterized only in the dimanganese(II1) oxidation state (289). Although no examples of crystallographically characterized Mn;I(p-OH)(OAc) cores exist, the topologically similar MniII(salampn)(OAc)complex (290) has a MniII(p-OR)
369
FIG. 53. [Mn"'(bi~picen)],(~-O)(~,,~-OAc)

(left) and [Mn"'(tmima)]z(~-O)(~,,~-OAcj
(right). [Reproduced with permission from (288) and (289). Copyright 1994 and 1993
the American Chemical Society, respectively.]
(OAc) core with a Mn-Mn separation of 3.55 A. The catalase mimic

Mni1(2-OHbenzimpn)(OAc)(Fig. 541,has the same core in the lower
oxidation state (291) with a nearly identical Mn-Mn separation,
3.54 A.
The more common members of the class of p-0x0-p-carboxylato
complexes have three bridging ligands instead of two. The Mn,(p-O),
(p-RC0,) core may be considered as a Mnz(p-O), core with a carboxylate spanning the Mn-Mn vector (292-294). However, the presence of
the additional carboxylate bridge induces the Mn,(p-O), unit to be
(Fig. 55) has a
bent. For example, [Mn~~bpy)(H,0~12(p-O),~p,,~-OAc~3+
dihedral angle between MnYpO), planes of 162" and a Mn-Mn distance of 2.64 A (295),which is somewhat shorter than Mn-Mn separations in typical Mnr(p-O), cores. Reduced derivatives containing
core have similar structural parameters
the Mn"'Mnrv(p-0),(p~,~-OAc)
FIG.54. Mn~Y2-OHbenzimpn)(OAcj.[Reproduced with permission from (291). Copyright 1994 the American Chemical Society.]
370
FIG. 55. [Mn'v(bpy)(H,0)],(~-O)~(~~,3-OA~)3+.

[Reproduced with permission from
(295). Copyright 1994 the American Chemical Society.]
(296). No bis-p-hydroxo-p-carboxylato cores have been isolated.

complexes
However, several examples of Mn~1'(p-OR)2(p1,3-OAc)
have been prepared and characterized (266). Mn$'(2-OHsalpn)
(MeOH)2(p-OMe)(pl,3-OA~)
is typical, with a Mn-Mn separation of
2.9 8 (266).
The p-oxo-bis-p-carboxylato structure type, Mn,(p-O)(p-OAc),with
two syn-syn carboxylate groups and a single p-0x0 ligand has been
structurally characterized with a number of supporting ligands, but
only in the dimanganese"' oxidation state. The Mn-Mn distance in
these compounds is in the range of 3.1-3.2 8, as would be expected
for the single 0x0 bridge. This core topology is particularly interesting
because it is the only p-oxo-p-carboxylato motif to have been isolated
in three different protonation states: Mn;"(p-O)(p-OAc),, Mn"Mn"'
(p-OH)(p-OAc),, and Mnf(p-OHz)(p-OAc)z(297-299). Representative
members of this series are illustrated in Fig. 56. It is noteworthy that
each protonation is accompanied by a decrease in oxidation state
and a consistent increase in Mn-Mn separation from 3.1 to 3.4 to
3.6-3.8 8. This is consistent with the decreased donor ability of the
single-atom bridge with each successive protonation, which promotes
longer bond lengths as well as stabilizing the lower dimer oxidation
states.
Dimanganese model complexes having only carboxylate bridges are
known primarily in the lower manganese oxidation states, and complexes supported by only a single p-carboxylate unit are particularly
rare (300-302). However, there are examples of dimanganese(I1) complexes bridged by a single carboxylate. Mn;*(RCO2)core has the bridging carboxylate in a syn-anti configuration with a long Mn-Mn distance of 5.6 8, Scheme 14,and complexes in which the carboxylate
bridge is in the anti-anti configuration have exceptionally long 5.9to 6.0-8 Mn separations. Bis-p-carboxylato complexes, Mnr*(p-RC02)2,
371
FIG.56. Examples of MnZ(p-O)(p-RC02)2

cores with varying protonation of the "p-0"
bridge. Upper left: [Mn"(N,N,N',N'-tetramethyl~ethylenediamine)~IZ(0Ac)~(p-OAc),(pOHZ). Right: {IMn"Mn"'(Me,tacn)Z1(p-(CH,),COO)z(p-OH)}+Z.
Lower left: [Mn"'
(bipy)CIZlZ(p-OA~)Z(p-O).
[Reproduced with permission from (472),(298), and (473),respectively. Copyright 1992 American Chemical Society, copyright 1996 WILEY-VCH
Verlag, and copyright 1993 American Chemical Society, respectively.]
have much shorter 4.3-A Mn-Mn separations with a syn-syn configuration for both p-carboxylate groups. Addition of a third carboxylate
group, Mn11(p-RC02)3,
yields a Mn-Mn distance of 3.6-4.0 A (303).
Mn
Mn
0d,
syn-syn
Mn
Mn- O y O- Mn
0y o ' M n
R
syn-anti
R
anti-anti
SCHEME
14.
Other unsupported Mnf'(p-OR)2 cores have Mn-Mn separations of

3.0-3.2 A. Complexes having one supported p-alkoxo and one unsup-
372
ported p-methoxo ligand can have Mn-Mn distances of 3.0 A or lower

(266,267).
Alkoxo-p-carboxylato complexes are topologically similar to their
bis-p-0x0 analogs, however, the Mn-Mn distances tend to be considerably longer in the p-alkoxo complexes than in corresponding p-0x0
derivatives. For example, p-alkoxo-p-carboxylato complexes, Mnp
(p-OR)(p-RCO,), have typical Mn-Mn seearations of about 3.5 A.
This is considerably longer than the 3.3-A separation observed for
Mni1'(p-O)(p-RCOz).
Mnz(p-OR),( -RC02)complexes have Mn-Mn distances ranging from 2.9 to 3.0
which is about 0.40 longer than
the oxo-carboxylato-bridged analog. Carboxylato-bis-p-alkoxo complexes having the MniYp-OR)(p-RCO,), core have even longer 3.3- to
3.4-A separations (304)which extend to as much as 3.5 A upon oxidation to the MnllMnll'(p-OR)(p-RCOz)zoxidation level (305, 306). This
compares favorably with 3.4 A for the MniI(p-OH)(p-RCO,)(297,298)
complex recently prepared, demonstrating that alkylation of p-oxobridged dimanganese cores is a good structural analog for protonation.
p-Phenoxide-bridged dimanganese clusters are structurally similar
to their alkoxide-bridged counterparts (304-306). However, an additional factor not present in alkoxo-bridged complexes is the possibility
of rotation about the Ar-0 bond in phenoxo complexes (238, 307,
308). Because the donor ability of the phenoxide ligand is dependent
on the orientation of the metal ion with respect to the aromatic plane,
we might expect that this would be manifested in structural changes
upon Ar-0 rotation in p-phenoxide-bridged manganese dimers. In
fact, when Mn!Yp-ArO), dimers are examined, it is observed that as
the dihedral angle between the MnzOzplane is increased from nearly
zero to nearly go", the Mn-Mn distance increases from 3.26 .k (304)
to 3.42 A (307).This is consistent with the phenolate oxygen being a
weaker donor when the metal ion acceptor is oriented perpendicularly
to the aromatic plane.
There are a limited number of N-bridged or imidazole-bridged dimanganese complexes with biologically relevant supporting ligands
(3091, consisting of azide-bridged complexes. Azide ligation is often
used as a model for peroxide ligation to manganese complexes due to
the greater stability of azide complexes, which facilitates structural
characterization. h i d e may bridge either in a p-1,l fashion, yielding Mn-Mn separations of 3.4-3.5 & or as an p-1,3 bridge to give a
3.5-A separation (310, 311). With respect to imidazole-bridged complexes, a complex with the ligand DCBI, 4,5-dicarboxyimidazole, was
with a Mn-Mn
recently reported, [Mn'"(3,5-di-t-b~tylsalpn)]~(DCBI),
separation of 6.18 A (Fig. 57) (312).
373
FIG. 57. [Mnrv(3,5-di-t-busalpn)I,DCBI. [Reproduced with permission from (312).

Copyright 1997 the American Chemical Society.]
C . TRINUCLEAR
Trimanganese clusters (and clusters of higher nuclearity) can usually be considered as assemblies consisting of the binuclear units discussed in the previous section. There are no known trimanganese
enzymes. However, it has been proposed in the past that the fourmanganese active site in the OEC may actually be composed of a trinuclear and mononuclear manganese site (discussed earlier). The
only trimanganese complexes having only p-0x0 bridges between the
manganese ions are those of the Mn3(p-0I4class (292, 313, 314).
These consist of a MnZ(p-O),core that is spanned by a Mn(p-O), unit
(Fig. 58). The topology of these complexes is reminiscent of the Mnz
FIG. 58. Two examples of trimeric structures with Mn0,-bridged Mn2(p-0)2cores.

[Reproduced with permission from (292)and (313).Copyright 1992 and 1990 the American Chemical Society, respectively.]
374
N
N
N 6
N
FIG.59. The trimeric (w3-O)(p-peroxocomplex). The other ligands are acetates and
diethylenetriamine. [Reproducedwith permission from (317). Copyright 1988 the American Chemical Society.]
(p-0)2(p-RCOz)core, with the Mn(p-O), unit replacing the RCOz unit

of the latter.
There are several reported structures of the Mn3(p3-O)unit supported by additional p-carboxylate bridges (315, 316). The Mni11(p3O)(MeC02)6complex has each manganese ion bridged by two carboxylate and one p3-0x0group structurally analogous to the Mndp-0)
(p-RCO,), complexes. The Mn-Mn separation in Mni1'(p3-O)(MeC0z)6
3.35 8, is longer than in Mnz(p-O)(RCO,),and is more reminiscent of
the distance in the Mn2(p-OH)(p-RCOz),complexes. A unique member
of this class of p 3 - 0 ~ 0trimanganese complexes has a single pl,z-peroxo
ligand bridging two of the manganese ions (317) (Fig. 59). Another
has been produced with a ps-OH and a bridging catecholate (318).
Trinuclear manganese motifs are common among carboxylate comcore has found been with variplexes. The Mn~11Mn11(p-OR)z(p-RC02)4
ous Schiff-base-containing supporting ligands (269, 319, 469). These
cores are characterized by two terminal Mn" ions, which "encapsulate" a central Mn"' ion with Mn-Mn neighboring distances of 3.5 8,
which is consistent with the MnllMnlll(p-OR)(p-RCOz)zdimanganese
cores having an identical bridging motif. The p-hydroxo-bridged analog, MnH'Mn"(p-OH),(p-RCO,), , has a similar structure with a somewhat shorter 3.4-A Mn-Mn separation (Fig. 60) (320). Univalent
Mnil trimers having only bridging carboxylate units, Mni1(p-RC02)6,
exhibit a Mn-Mn bridging motif in which one of the three bridging
carboxylate undergoes a carboxylate shift in which one of the carboxylate oxygens bridges two Mn" ions (Fig. 61) (321). Therefore, these
core units are more precisely defined as Mn&1(pl,3-RC02)4(pl,l-RCO~)
375
FIG.60. A Mn"iNnli/Mn'ii trimer with a Mn~"Mn'1(p-OH)2(RCOs)4

motif. [Reproduced
with permission from (320).Copyright 1994 the American Chemical Society.]
with a characteristic Mn-Mn separation of 3.5-3.6 A. The p-phenolato-bridged trimanganese(I1) complexes exhibit a similar carboxylcore with interate shift, giving the Mn~l(ArO)a(pl,~-RCO,),(pl,,-RCO,),
manganese separations of 3.27 A (Fig. 62) (322).
FIG.61. Mn" trimer with a Mni'(RC02)tibridging motif that includes bridging p,,.Icarboxylates and pl,,-carboxylates. [Reproduced with permission from (321). Copyright
1990 Wiley-VCH Verlag.]
376
-N
FIG. 62. Mn~1(ArO)2(~Ll,s-RC02)2(CL1,~-RC02)2

core. [MnP(5-NOzsalim)2(OAc)ll.
D. TETRANUCLEAR
As with binuclear manganese complexes, several recurring tetramanganese core structures have been synthesized with different supporting ligand sets (44, 323, 324). The simplest tetranuclear manganese cluster is the cubane core, Mn404, which has tetrahedral
symmetry in the absence of distortion. A generalized scheme of a cubane core is shown in Scheme 15. This core is arranged in such a way
that two 0x0 bridges exist between any two manganese ions so that
all of the manganese ions are equivalent. A number of mixed-valent
distorted cubane core structures are known in which the manganese
distances are not all equivalent (325,326).This core was proposed in
the past as a model for the manganese cofactor in photosynthetic oxygen evolution, so it has been synthesized with a number of supporting ligand architectures. One of the 0x0 groups in the cubane core
can be replaced with a carboxylate donor to form the asymmetric
Mn403(PhCOJ complex, in which the four manganese are no longer
equivalent and Mn-Mn distances of 2.8 and 3.3 A are observed. This
complex is unique among tetramanganese cores in exhibiting pl,land
p1,3 bridging by a single carboxylate group. An 0x0 group may also be
replaced by a p3-chloride ligand to yield the distorted cubane,
MII,(O)~C~
(Fig. 63) (327,328).In this mixed-valent Mnil'MnTV
complex
the three Mn11'Mnw02
faces have Mn-Mn separations of 2.8 A, as expected for bis-p-oxo-bridged Mn"'MnW complexes. However, the MniI'
(O)(Cl) faces have an average Mn-Mn distance of 3.3 A. Alkoxidecontaining cubane analogs (329),Mn,(OR),, have also been prepared
and shown to be topologically similar to the conventional oxide-containing cores with somewhat longer Mn separations.
377
Mn40nX, Core Structures
0-
MnM
,/n0
-
0-Mn-
0
Adamantane
'Fused-Open' Cubane
Ih
Mn-
\O+Mn
0- -Mn
\I \I
Mn-
A Selection of Cubane Core Structural Types
'Butterfly" Type Cores
SCHEME
15
Addition of two oxide donors yields the adamantane complexes,

MQO,, which also have 0x0 bridges separating all of the manganese
(Scheme 15) (330). The difference between the cubane and adamantane core is the presence of two oxygen atoms, leading to the
proposition that an adamantane-to-cubane conversion in the OEC
may lead to oxygen evolution (31, 199, 200). Another Mn406cluster
tetranuclear cluster is the "fused open cubane," in which a Mn406
cluster is formed that is the equivalent of two face-sharing Mn304
cubanes, with one manganese site vacant (329) (Scheme 15). Loss of
two oxide bridges from a cubane core yields the Mn402core complexes,
commonly called butterfly clusters, which have an MnzOzcore as their
foundation and two p3-0x0 groups connecting to the two pendant Mn
ions lying above the MnzOzplane (Scheme 15) (331, 332). A centrosymmetric Mn402core has also been observed as well. In these complexes, two p3-0x0groups from MnzOzbridge to two additional manga-
378
FIG.63. A manganese cubane cluster with a corner oxygen atom replaced by a chloride. [Reproduced with permission from (327). Copyright 1993 the American Chemical
Society.]
nese ions, but they are now disposed on opposite faces of the MnzOz
plane.
There are a number of tetramanganese clusters that cannot be considered as merely distorted members of the basic classes, but that
may also be relevant in biological manganese chemistry. For example,
a highly symmetric tetramanganese complex is formed with the ligand 2-OHpicpn (333).This complex, an efficient catalyst for the catalase-like disproportionation of hydrogen peroxide, is an Mn4L4structure in which each pair of manganese ions is bridged by a single
alkoxide bridge covalently attached to the picpn unit (Fig. 64). Each
manganese is separated from its neighbor by 3.7 A. In this complex
all of the Mn-Mn linkages are identical, so it is not precisely a dimerof-dimers model. However, the complex M r ~ ~ O ~ ( t p h pcontains
n ) ~ + two
MnzOzcores cofacially oriented and linked by two alkoxide bridges
covalently bound to the tphpn ligand (334)(Fig. 64). In this case, two
different Mn-Mn linkages support the cluster, and this is accurately
considered as a dimer-of-dimers structure. A variation on this theme
is the dimer-of-dimers complex [Mn,O,( tmdp)]%+,
which also has two
MnzOzdimers, but in this case linked by methylene spacers of the
tmdp ligand. This complex has a rather short 2.6-A Mn-Mn distance
across the dioxo core, but a long 5.9-A Mn-Mn distance across the
methylene spacer connecting the two dimers (335)(Fig. 65).
379
FIG.64. Two tetramers. Left, is the [Mn,0,(tphpn)214atetramer. Right, [Mni1(2-OH

picpn)J tphpn = N,N ,N',N,-tetrakis~pyridylmethyl~-1,3-diaminopropan-2-01.
20Hpicpn
is the Schiff-base condensation product of pyridine-2-carboxaldehydewith 1,3-diaminopropan-2-01, [Reproduced with permission from (334)and (333),respectively. Copyright
1991 and 1996 the American Chemical Society, respectively.]
A number of chainlike tetramanganese complexes may also be relevant to biological manganese chemistry. The Mn,O,(bpy)i+ complex
(336) has the same core stoichiometry as the adamantane complexes
(Fig. 66). However, this complex contains four Mn" ions in a linear
configuration in which each pair of neighboring manganese ions is
bridged by two 0x0 bridges with a Mn-Mn distance of 2.7 A. Longer
Mn-Mn distances of 4.9 and 6.4 A are observed between the 1,3 and
1,4 manganese ions. A topologically similar complex, Mn4(2-OHPh
PhenI6, was prepared having six phenoxide instead of oxide bridges
(337).The Mn-l-Mn-2 distance here is 3.4 A, characteristic of dimanganese diphenoxide complexes, in which the Ar dihedral angle with
the Mn202plane is nearly 90". Mn902(tphpn)z(Hz0)2(CF~SO~~~+
has a
Mn,(0)2(RO)2 core with two manganese linked to a Mnz02 core
through alkoxo bridges (Fig. 67) (338).
Hydrated manganese in acidic aqueous solution exists in two oxidation states, Mn" and Mn"], with a potential of + 1.51 volts separating
them (339). Therefore, Mn"' in solution is a strong oxidant. This
380
FIG.65. A dimer of dimers. Two ( p - 0 2 - ) ,bridged Mn"'MnN dimers are bridged to

one another via the ligand. Water molecules are also bound to the manganese of this
complex. [Reproduced with permission from (335).Copyright 1994 the Chemical Society
of Japan.]
arises from the strong driving force to form the stable half-filled d5
subshell of Mn" from the d4 Mn"' ion. The MnWstate, which is a d3
ion, exists as the dioxide MnOz or as oligomeric forms thereof. Permanganate ion, having a do Mnwlcenter, is diamagnetic.
Mononuclear metal-ligand complexes with the Mn"-MnV oxidation
states are all known provided suitable chelating ligands are provided,
and polynuclear complexes with Mn1I-MnWare known as well. Consequently, to understand the spectroscopy of polynuclear complexes, it
is necessary to begin by examination of the electronic properties of
381
FIG.66. The first coordination sphere of the complex Mn106(bpy)Q+

showing the core
structure. Matching subscripts on the nitrogens designate that they originated from
the same bipy molecule. [Reproduced with permission from (336). Copyright 1994 the
manganese ions in these oxidation states. Mn" in an octahedral ligand field has a
ground state that has spin S = 5/2 but is an
orbital singlet. Therefore, no spin-orbit coupling arises from the
ground state of Mn", and S is a good quantum number for describing
FIG.67. The linear chain structure with the tphpn ligand. The two inner manganese
are involved in MnzOzcore and are then bridged to the outer two manganese ions by
ligand alkoxides. [Reproduced with permission from (338).Copyright 1990 the American Chemical Society.]
382
the spin levels of this ion. The same is true for the d3 MnIVion, which
has an orbital singlet 4A2ground state in an octahedral field. However, the ground state of the octahedral d4 Mn"' ion is the orbital
doublet 5E, which gives rise to substantial spin-orbit coupling, L S.
Therefore, S is not a good quantum number for describing Mn"' electronic properties, and the value L + S is used. The MnVcomplexes
thus far isolated are square pyramidal and diamagnetic. This suggests a very strong axial splitting in the t2gorbital set leading to a
singlet 'A ground state in these complexes.
A. ELECTRONIC
SPECTROSCOPY
Because the most intense bands in the visible spectroscopy of manganese compounds are charge-transfer bands involving the bound ligands (i.e., metal-to-ligand or ligand-to-metal charge transfer), the
visible spectra of polynuclear complexes tend to be very similar to
those of mononuclear complexes with the same ligands. Consequently,
little information on the structure of polynuclear complexes is contained in their electronic spectra. Exceptions to this rule are certain
localized mixed-valence complexes that exhibit intervalence charge
transfer bands (ITS)formally involving electron transfer between nonequivalent metal sites in the complex. Because this can only occur in
polynuclear complexes that are valence localized in the ground state,
valence-delocalized complexes cannot exhibit IT bands. An example
of an IT transition at 1170 nm ( E = 270 M-'.cm-l) is found in the
[Mn11Mn111salamp21complex, whose structure is shown in Fig. 68,
FIG.68. Core structure of [Mn"Mn"'(~almp)~].,

a diagram of the ligand and the binding motif for one ligand to the manganese are shown above the core structure. [Reproduced with permission from (308).Copyright 1991 the American Chemical Society.]
383
along with a representation of the ligand (308).The appearance of IT

bands is rationalized based on Marcus theory. Electron transfer between two coupled metal centers, MAand MB, is rapid compared to
molecular vibrations, resulting in a low-lying excited state in which
an electron formerly residing in an orbital of predominantly MAcharacter has been promoted to an orbital of predominantly MB character.
This excited state may relax via a nonradiative mechanism wherein
the ligand environment adjusts along a vibrational normal mode to
return to the ground state configuration. Therefore, the energy and
lifetime of IT excited states can be strongly influenced by the vibrational structure of the polynuclear complex.
UV-vis spectroscopy proved particularly enlightening with respect
to the catalase system (23, 30, 132). The spectra of model complexes
bridged by carboxylate and 0x0 bridges were quite similar to those
observed for the catalase. This led researchers to the proposal that
such a motif might also be present, a hypothesis that has been borne
out by the crystal structure of the catalase (125).
B. MAGNETISM
Temperature-dependent magnetic susceptibility is often used in the
study of polynuclear metal complexes to determine the magnitude of
spin exchange energetics between two metal centers in the same molecule (340, 341). This magnitude is called J and is a measure of the
energy difference between successively higher molecular spin states
when a molecule is placed in a magnetic field. The temperature dependence of the molar magnetic susceptibility x,,, arises from two
sources. The first is zero-field splitting, which is measured by D and
is observed in mononuclear complexes as well as polynuclear complexes. The primary factor contributing to D in manganese complexes
is coupling of the orbital angular momentum and spin angular momentum vectors, and is defined by L . S. Octahedral Mn" has a t&eE
configuration with a 6A1ground state, and MnN has a t!,e: configuration with a 4Azground state. Because both of these ground states are
orbitally singlet and have no net orbital angular momentum, D will
typically be on the order of 0.10 cm-' or less in complexes of these
ions (339).Therefore, in polynuclear complexes of these ions, S is a
good quantum number and the magnetic exchange interaction between the ions can be described by using a spin-only Hamiltonian.
However, Mn"' has a t!,e; configuration with a 5E ground state, which
is an orbital doublet that will have a substantial contribution to the
overall magnetic moment arising from the net orbital angular mo-
384
mentum. As a result, Mn"' systems may have D on the order of a few

cm-'. For such a system, S is not a good quantum number to describe
the energy of the state, and a spin-only description of the magnetic
exchange may not apply. The low-temperature susceptibility of Mn"'
complexes are often dominated by the thermal depopulation of the
higher spin states giving a diamagnetic ground state at low temperature.
The second source for the temperature dependence of xm is spinspin (S1* S z )coupling or spin exchange, which is unique to polynuclear
complexes. S1.Sz coupling arises when the spins on two metal centers, MI and M2, are correlated. The two spins may be coupled antiferromagnetically so that the ground spin state S for the system is S =
S , - Sz. On the other hand, the spins may be coupled ferromagnetically to give a ground spin state S = Sl + S z . We will discuss S1* S z
coupling in the context of binuclear systems with the understanding
that the general principles applied in the binuclear case also apply to
higher polynuclear complexes.
Characterized MnH systems are typically weakly coupled, with 5
values on the order of 10 cm-' or less (30).The orbitally single 'Al ion
is therefore a good example of a system whose temperature-dependent magnetic susceptibility is well treated by the spin-only formalism due to the lack of L .S coupling from the Mn" ground state. The
spin-only Hamiltonian is given by H = SIJSzwhere 5 is the spincoupling constant with units of energy and Sl and Sz are the spin
operators. The eigenvalues E for the system HW = EW are given by
E = J [ S ( S + 1) - Sl(Sl+ 1) - Sz(S2+ l)]and the spin energy levels
are given by 0, 2J,6J, 125,205, and 305. The temperature-dependent
magnetic susceptibility is then a result of the depopulation of higher
spin states as the temperature is decreased, according to the Boltzmann distribution of states, i.e., P, = A,exp(E,/kT)LU,exp(E,/kT),
where En and A, are the energy and degeneracy of state n and P, is
its relative population. For an antiferromagnetically coupled system,
the ground state is diamagnetic and the magnetic moment should approach zero as the temperature is lowered. Ferromagnetic coupling
gives rise to a paramagnetic S = 5 ground state so that the low-temperature magnetic moment will approach 11 BM per binuclear complex. Binuclear Mn'" complexes may also be treated by a spin-only
Hamiltonian with eigenvalues E = 0, 25, 6J, and 125. The ground
state is also a spin singlet in this case, but the high-temperature moment should approach 5.4 BMhinuclear complex due to equivalent
population of all of the spin states.
385
The Mn"' ion has an 5E ground state that is orbitally nondegenerate. Even so, the spin-only contribution to the temperature-dependent
magnetic susceptibility in strongly coupled Mn"MnT", Mn"'
2 , and
Mn"'MnIV complexes may be treated in the same way as for the Mnil
case discussed earlier. For cases in which J 4 D, D can be considered
as a perturbation on the spin-only energy levels and the spin-only
formalism is usually adequate to explain the magnetic data and to
calculate J . For systems with extremely weak coupling (i.e., J < D,J
may be considered as a perturbation on the D-split levels, and the
magnetism will be dominated by D. However, when J and D effects
are of the same order of magnitude, treatment of the temperaturedependent susceptibility is complicated by the fact that neither S nor
L is a good quantum number for describing the spin exchange. A full
treatment of magnetic exchange involving both spin and orbital contributions is quite involved and beyond the scope of this discussion
(340, 341). Great care must be exercised in the treatment of data on
weakly coupled systems containing Mn"' in order to accurately assign
which effect is responsible for the observed temperature-dependent
magnetic data. In many cases, one cannot uniquely fit the magnetic
data to a single set of J , D, and g values.
C. EPR SPECTROSCOPY
Electron paramagnetic resonance spectroscopy is one of the primary
tools in studying the electronic structure of polynuclear complexes
(341). Whereas magnetic susceptibility studies are capable of detecting electronic interactions as small as a wavenumber (discussed
earlier), the EPR spectrum of a polynuclear complex may be sensitive
to intramolecular exchange couplings as small as 0.001 cm-l even at
room temperature. Additionally, the 55Mnnucleus has a nuclear spin
( I ) of 5/2. For a mononuclear system, six hypefine lines are predicted.
However, for a multinuclear system containing n Mn ions, as many
as 6" lines may be observed. This gives rise to a rich EPR spectroscopy
among polynuclear manganese complexes, which has been used extensively in the study of biological manganese systems and model
complexes.
Mononuclear Mn" and Mn" complexes exhibit characteristic EPR
spectra that are, for the most part, insensitive to the ligand environment around the metal. However, the EPR spectra of polynuclear
complexes are strongly dependent on the metal ligation environment,
the nature of bridging ligands, and the degree to which the metal
386
ions are coupled. We will examine some of the relevant binuclear and
polynuclear cases in more detail.
1. Binuclear Complexes
Mononuclear Mn" complexes almost always exhibit a single derivative signal with a crossover at g = 2 and often having a well-defined
six-line Mn hyperfine splitting superimposed on it, as expected from
an S = 5/2 spin system with only slight spin-orbit coupling. However,
two antiferromagnetically coupled Mn" ions coupled antiferromagnetically would be expected to be EPR silent at low temperature owing
to the ground-state spin singlet. Binuclear MnP complexes often do
exhibit an EPR spectrum due to thermal population of the S = 1
through S = 5 spin states. Although these are non-Kramer's systems
with integer spin, which often do not show X-band EPR spectra in
mononuclear complexes due to zero-field effects, such effects are minimal in Mn" due to the 6A1ground state, and there are a number of
examples of EPR spectra of Mn" dimers (279, 438, 439, 474, 476).
Binuclear mixed-valent Mn'lMnlll complexes nearly always exhibit
an EPR signal because the ground state and all thermal excited spin
states are Kramer's spin systems. The form of the spectrum depends
strongly on the magnitude of J and on the zero-field splitting imparted by the Mn"' ion. Strongly coupled systems typically exhibit a
signal at g = 2 with a multiline hyperfine splitting from two Mn ions;
more weakly coupled systems show signals withg > 2 due to transitions involving zero-field split levels (305, 475).
Unlike Mnil dimeric systems, MnH' complexes almost never show
EPR spectra in the X-band region due to zero-field splitting that puts
allowed transitions out of range of X-band energy (9.5 GHz). However,
strongly-coupled ( J = 100 cm-' ) mixed-valence Mn"'Mn'' systems
show very characteristic EPR spectra at g = 2, consisting of between
12 and 19 well-resolved Mn hyperfine lines, consistent with an S =
1/2 ground state. A typical such spectrum, for M"'MnIV (p-O),L,, is
shown in Fig. 69. Although 36 lines are theoretically possible for two
coupled manganese ions, between 12 and 19 lines are usually observed in practice due to overlap of some of the hyperfine lines. The
g = 2 signal is the only one typically observed for strongly coupled
(~-0
complexes
)~
(140), but
Mn"'Mn" systems such as the Mn1I1MnW
(
J
=
10
cm-')
may
also exmore weakly coupled Mn"'MnTVsystems
a
lower-field
signal
at
higher
temperatures
that
is
associated
hibit
with thermal population of a zero-field split S = 312 excited state
(202), a feature also often observed in the EPR spectra of mononuclear MnIVcomplexes, which are also S = 3/2 ions (140).
11
2200
2600
3000
3400
3800
387
4200
G(H4
FIG. 69. A Mn"'Mn'" exchange coupled EPR spectrum, from [Mn"'Mn"(bpza)p
(p-O)2]3'.bpza = 2-(2'-pyridyl)benzimidazole.LReproduced with permission from (465).
Copyright 1994 Elsevier Science.]
Strongly coupled binuclear Mn;" complexes have never been shown

to exhibit X-band EPR spectra due to the low-energy singlet ground
state. Because stabilization of the high-valent Mn'" requires very
strong oxyanionic ligands, nearly all such MnJ" complexes have at
least one bridging 0x0 or hydroxo ligand, and they may contain additional bridges such as carboxylate ions, all of which contribute to a
generally strong coupling in such systems. Consequently, there are
few examples of weakly coupled MnF systems to make generalizations about the presence or lack of EPR spectroscopy in these systems.
However, we might expect that because ground-state spin-orbit contributions to zero-field splitting are absent in MnN ions, higher spin
triplet and quintet states might be EPR active here just as they sometimes are in the MnH complexes. A recent example of an EPR signal
arising from a weakly coupled ( J = -2.3 cm-') Mnk" dimer was observed for the complex [Mn'"(di-t-butylsalpn)lzDCBI (Fig. 70) (312).
2. Polynuclear Complexes
Several factors unique to polynuclear systems make the interpretation of their spectra more complicated than for the binuclear cases.
One such factor is that a given manganese ion may couple to two or
more other ions with different exchange magnitudes, J , and J z . The
interpretation of such spectra is often simplified by assuming that Jz
388
A 5.5K
B 50K
1000 2000 3000 4000 5000 6000 7C

Field (gauss)
FIG.70. X-band EPR spectra of the weakly coupled [Mn1"(3,5-di-t-busa1pn)l2DCBI

complex. [Reproduced with permission from (312).Copyright 1997 the American Chemical Society.]
is much smaller than J 1 and can either be neglected or treated as a

perturbation on the spin splitting by J,. On the other hand, J , may
be exactly equal to Jz (in a highly symmetric trinuclear species for
example), thus reducing the spin exchange problem to a simple manifold of states treatment as for binuclear species. However, when J1is
of the same order of magnitude as J2,then the spin splitting is complex and cannot be treated by a simple manifold of states.
A second factor complicating the EPR spectroscopy of polynuclear
species is the fact that the possible number of Mn hyperfine splittings
increases as 6" where n is the number of manganese ions contributing
to the hyperfine interaction. Furthermore, the hyperfine splitting constant A decreases as lln. Therefore, EPR spectra of polynuclear complexes may exhibit broadened signals due t o a large number of unresolved hyperfine splittings, which may not be distinguishable from
broadening from other sources such as slow tumbling or anisotropy.
3. EPR of Biological Manganese Complexes
EPR is one of the most useful techniques brought to bear in the

study of polynuclear manganese centers in biology. For example, it
was the hyperfine splitting in the CW EPR spectrum of spinach thylakoid membranes that first showed conclusively that the oxygen evolving complex in PS I1 consisted of a multinuclear cluster of coupled
manganese ions (160, 161). The EPR from the S, oxidation state of
active PS I1 centers has also been the primary signal for tracking the
redox cycling of the OEC during oxygen evolution by PS I1 (33, 42,
46). There have also been numerous attempts to establish the nuclear
389
and oxidation levels of the manganese cluster. Brudvig, for example,

used variable-temperature EPR data to support a tetranuclear cluster
(342, 343). Several other groups have variously assigned these EPR
spectra to two dimers (3441, dimers of dimers, and tetramers (345).
More recently, Dismukes has argued that the Sz spectra result from
a tetranuclear cluster composed of 3 Mn"' and 1 MnWions (346).To
obtain a good fit of the data, the Mn"' ions are predicted to be five
coordinate with an inverted order for the &orbital energies. The Sz
signal has also yielded information on the ligand environment (187,
347) through 15N labeling and observation of the 15N hyperfine splitting with ENDOR spectroscopy, and on the binding of substrate and
inhibitors (348) to the OEC and model complexes (203, 349). Very
recently, two new EPR signals have been observed. The Sosignal was
shown to be multiline in the presence of methanol (170-172). Using
parallel-mode EPR spectroscopy, Britt has shown a beautiful
multiline signal for the S, oxidation level (174). Interpretations of
these new signals hold great promise for understanding this enigmatic cluster's structure.
Enzymes containing binuclear Mn sites are especially amenable to
structural elucidation using EPR spectroscopy. The EPR spectra of
various oxidation levels in the binuclear manganese catalase were
used to assign the low-valent active form as a weakly coupled MnL)
center and the "superoxidized inactive form as a strongly coupled
MnlIIMnWcenters (7, 23, 136, 138-140, 350, 351). These assignments
were based in part on a comparison with the EPR spectra of crystallographically characterized model compounds. The arginase active site
was shown to be a binuclear Mn" center by the shape of the spectra
and the number and spacing of Mn hyperfine couplings (7). The structures proposed for both the manganese-containing catalase and arginase from these studies were essentially confirmed by subsequent
X-ray crystal structure analysis (8).
D. X-RAYA~SORPTION
SPECTROSCOPY
Although it is an unconventional spectroscopic technique, X-ray absorption spectroscopy (XAS) has been the primary tool in the elucidation of the structural features of biological multinuclear Mn centers
for which high-resolution crystallographic data are not yet available
(23, 24, 134). This is typically done by comparing features observed
in the XAS of the biological complex with similar features found in
the spectra of crystallographically characterized Mn model complexes.
We defer an extended discussion of the theory and principles of XAS
390
to several reviews on the subject of XAS in biological and model manganese complexes (24, 134). However, a few basics will suffice to understand the information that is contained in XAS of polynuclear complexes.
Unlike UV and visible electronic spectroscopy, which involve promotion of outer-shell valence electrons, X-ray absorption involves the
promotion of core electrons to higher energy using X-ray frequencies
(24,134).As such, X-ray absorption is a high-energy technique. There
are really two techniques that come under the heading of X-ray absorption. X-ray Absorption Near Edge Structure, XANES, involves
promotion of a 1s electron to the 3d level in the case of manganese.
(This is for the K shell; L edges with core electrons from the n = 2
shell are now being examined for models and enzyme.) This gives rise
to an absorption edge whose energy is very characteristic of the oxidation state of the manganese. As a general rule, manganese in
higher oxidation states exhibits absorption edges at higher energies,
as might be expected from the tighter binding of core electrons in
higher oxidation states. In fact, the edge energy seems to correlate
best to the average number and bond lengths of ligands bound to the
metal. In polynuclear complexes, shifts in the edge energy are correlated with changes in the average oxidation state of the manganese
ions. This complicates the use of XANES in higher nuclearity complexes because one-electron oxidation results in a smaller change in
the average oxidation state as complex nuclearity increases. Even so,
this technique has been used extensively in tracking the oxidation
states of the manganese cluster in PS I1 upon S-state advancements.
The progressive edge shift upon increasing oxidation state for polynuclear manganese complexes is nicely illustrated by the XANES spectra of the isostructural series Mn2(2-OHsalpn)i(n = -2 to +1)in four
successive oxidation states (Fig. 71) (467).
As the X-ray energy is increased beyond that required for promotion of the core electron, ejection of core electrons into the continuum
occurs. This ejected electron propagates from the Mn center until it
encounters another atom from which it can be back-scattered. The
interference of back-scattered waves with propagating waves leads to
an interference pattern that is manifested as an oscillation in the
X-ray absorption pattern. Fourier transformation of this oscillating
spectrum from the frequency domain to the distance domain gives a
new spectrum whose abscissa contains information on the distance
between the target atom (i.e., the Mn center) and the back-scattering
atoms. This second technique is called Extended X-ray Absorption
Fine Structure, E M S , and has been the only spectroscopic tech-
6531
6538
6545
6552
391
6559
Energy (ev)
FIG.71. XANES spectra of the family of complexes IMn(20Hsalpn)l" (n = -2, - 1,

0, +l), and the oxidation states on the manganese go from Mn"Mn" to Mnl'lMnl". Noticeable changes occur in the edge energy with each increase in the oxidation state of
the manganese. [Reproduced with permission of the author, from (4671.1
nique to yield direct direct structural information about the manganese cluster in PS 11.
The XANES of the Mn catalase provided the first definitive proof
that this enzyme cycles between the Mnil and MnkI' oxidation levels
(137, 352) The extent of catalase activity correlated with the proporenzyme; however, samples with Mn*" quantitation of Mnf or
tively showed reduced activity. The EXAFS of the Mn catalases have
been less informative because of the Mn-Mn separations in the reduced, active enzymes (135).Nevertheless, EXAF'S of the "superoxidized" enzyme demonstrated that the MnlIIMnl" enzyme has a
Mn-Mn separation of ~ 2 . 7
A, which is consistent with a di-p,-oxo
core (135).Subsequent spectroscopic analysis confirmed that a "diamond core" with a bridging syn,syn acetate formed the enzyme active
site ( 9 ) .
While interpretation of the XAS for the Mn catalase has been relatively straightforward, the application of XANES and EXAFS to the
OEC has led to several significant disagreements among workers in
this field (24,28,33,46, 48, 134, 154). Klein and Sauer first collected
XAS data on the OEC and provided the first detailed structural and
oxidation-state information about the Mn cluster (353, 354). The
XANES indicated that the S , and S , enzyme forms contained significant amounts of Mn"' and Mn'". At the same time, both S states could
be examined by EXAFS and exhibited short Mn-Mn vectors of 2.7 A
(154-159), which is consistent with a high-valent di-pa-oxocore, for
example the structure in Figs. 49 and 50. Because the S , state contains a Kramer's spin ladder (i.e., nonintegral spin system) as shown
by EPR spectroscopy, the S , state was assigned as MniI'MnY and the
392
S2 state was assigned as Mn"'Mn?. All of the groups using XAS have
agreed upon these oxidation levels and general structural details (24,
33,154, 156, 158, 159). However, there are members of the EPR community that suggest that the oxidation levels should be adjusted
downward by two oxidation states (346).The justification for this is
that low-coordination-number Mn"' ions may have aberrantly high
edge energies. If this is true, fits of the EPR spectra would be consistent with XAS data. However, investigations of many model compounds have not supported the idea that XANES edges are this sensitive to coordination number changes (24,33,134).
One area of significant controversy is whether the Mn center is oxidized upon each S-state transition (24, 33) Klein and Sauer argued
that upon forming S3,the edge energy did not increase sufficiently to
support manganese oxidation (355).However, the &-state EPR spectrum disappears upon additional oxidation to S3,which indicates that
the oxidizing equivalent must be in close proximity to the manganese
cluster (46).These observations have led some investigators to suggest that the substrate (water) is partially oxidized or that a protein
residue near the active center is oxidized. Significant excitement was
generated when Boussac and Rutherford formed a new EPR signal in
EGTA (Ca-depleted) samples (this signal was later shown to occur
with acetate addition) (176, 356). The broad radical signal was assigned to histidine radical. However, Britt has shown recently that
this signal originates from tyrosine Y;, which is the intermediate oxidized charge carrier of the enzyme (177, 178) Subsequent investigation has placed this tyrosine at <5 A from the manganese cluster.
Thus, there is no longer additional evidence suggesting a nonmanganese site of oxidation. In fact, groups have suggested that the Klein
and Sauer analysis is flawed based on the experimental conditions or
by the fitting of the edge energies (24).Unfortunately, there still is no
definitive resolution to this issue. At present, our bias is to assign
oxidation to manganese at each S-state transition; however, it is possible that our views will evolve with new data.
Another area of significant controversy has been the interpretation
of the higher scattering in the EXAF'S. Klein and Sauer originally
reported a Mn-X scatterer at ~ 3 . A3 (357).This feature was irreproducible (3551,and only when better sample preparation and detectors
became available did this scatterer appear reproducibly (24,33, 156,
158, 358). Some of the contribution to this peak comes from Mn-C
scattering; however, this interaction alone cannot explain the intensity of this feature (24).For this reason, the 3.3-A peak is attributed
to Mn-Mn or Mn-Ca scattering. There has been great discussion as
393
to the identity of the scatterer, and definitive proof of the presence of

calcium has not been published (477). However, there are recent reports at scientific meetings suggesting that Mn can be seen at 3.4 A in
Sr2+-treatedsamples. If these withstand further scrutiny, they would
provide strong evidence support for the close proximity of Ca to Mn
clusters.
The other cofactor often neglected in discussions of the OEC is chloride. Again, EXAFS data have been obtained to address whether or
not chloride is bound to the manganese cluster (33, 358, 359). These
experiments are technically very difficult. There is the possibility that
chloride (or bromide in bromide-substituted samples) may be observed in the EXAFS spectrum. However, we believe that it is unlikely that one can resolve this issue, because the signal-to-noise ratio
may not be sufficient (156)to observe only one chloride per four manganese ions. Very recently, chloride EXAFS of manganese model complexes has been reported (360).This new application of EXAFS may
be able to resolve this issue.
XAS studies of the So state have also been obtained by several
groups (24, 33). These are also technically demanding experiments
because the S, state is the dark stable-oxidation-state level of the
OEC. Chemical reductants are used to generate So,often with light
flashes required for reoxidation. Therefore, it is difficult to know precisely the amount of Sothat is formed and also whether any of the
reaction centers have lost manganese by treatment with reductants.
Given these caveats, it appears that the Sostate in some preparations
may contain Mn" (24, 33, 361). This would suggest a total oxidation
assignment of Mn"Mn"'MnF. An alternative interpretation is that
there is only Mn"' and MnIVpresent in So, giving a cluster assignment
of Mn&llMnw
(24, 362). It may also be possible that different methods
of generating So may give different relative oxidation states for the
manganese in the cluster. Perhaps evaluation of the SoEPR spectrum
will allow definitive interpretation of the XANES spectra.
V. Reactivity
The structures and properties of manganese complexes have provided much useful information on the possible structures and metalmetal or metal-ligand interactions of manganese ions in biological
systems. How enzymes carry out their respective transformations is
another integral aspect for understanding the overall whole of manganese-containing redox enzymes. Model complexes have been probed
394
over the years to ascertain how they mimic the reactivity of manganese-containing enzymes, with much of this work focused on the redox-active manganese enzymes, catalase, MnSOD, and the OEC. Further delving into such chemistry often unearths unexpected results
and may lead to beneficial new chemistries.
The following segment on the reactivity of manganese complexes
will be broken into two sections. The first will cover some of the areas
that compose research into the reactivity of manganese-based systems as applied to a specific synthetic organic transformation. The
second segment will concentrate on systems designed specifically to
model the reactivity of particular enzyme systems.
A. ORGANIC
TRANSFORMATIONS
Manganese has a rich history within the field of organic synthesis
(363,364).For example, the permanganate anion has long been used
as an oxidant to produce a variety of products (363).Manganese"' acetate also has been extensive explored over the years for the initiation
of free radical reactions that lead to carbon-carbon bond formation.
These topics have been reviewed and will not be presented further
here (363,364).Manganese chemistry, however, has made an impact
in other areas as well, notably the asymmetric epoxidation of alkenes.
1. Epoxidation
a. Salen-Based Systems (Kochi-Jacobsen-Katsuki

epoxidation
chemistry) One prime example of the application of manganese complexes t o organic synthesis is the burgeoning field of transition metalepoxidation catalysts based on the salen2-ligand (235, 365-367). Recent work in this area has emphasized the stereospecific generation
of enantiomeric epoxides, in good yields and with high enantiomeric
excesses. Such chemistry is of interest in light of the exquisite control
exhibited by nature in producing such stereocenter-bearing oxidation
products. Thus, one goal in exploring biomimetic oxidations is to
achieve the same level of regio- and stereoselectivity as can be
achieved through enzymological transformations. Furthermore, stereoselective epoxidation catalysis promises the ability to produce chiral products that are intermediates in or final products in stereospecific syntheses of use in fields such as medicinal chemistry. Since the
early 199Os, manganese salen-based systems have been designed that
have shown promise for the preparation of asymmetric epoxides. As
is the style in organic chemistry, this important transformation has
been named after key researchers in its development, and will hereaf-
395
ter be referred to as Kochi-Jacobsen-Katsuki epoxidation chemistry.

In general, previously described chemistry does not produce epoxidized products in enantiomerically pure form (235, 365-3671, especially if the epoxidized olefin's structure does not promote the formation of one enantiomer over another. Thus, many enantiomeric
epoxides are often prepared as racemic mixtures. Therefore, it was
necessary to design a system that would facilitate stereocontrol by
orienting the substrate with respect to the oxygen atom to be transferred from the epoxidizing reagent to the substrate. Preferably, this
0x0 transfer would occur with large enantiomeric excesses. This goal
has been achieved in part by the Mn-salen-based systems (235,365367).
Epoxidation chemistry with a variety of metal complexes has been
explored over the years. Historically, the Mn-salen catalysts grew out
of research focused on understanding the reactivity of cytochrome Pd5"
(368, 369). This cytochrome is employed in many enzymes that catalyze a variety of oxidations of organic substrates, including the epoxidation of alkenes (370). Usually these reactions are accomplished
with complete stereochemical control. Interest in modeling this reactivity has led to a rich literature on biomimetic oxidations (365, 367369,371-375). Because cytochrome P450is a heme enzyme, the initial
studies were conducted with iron porphyrin model complexes. Groves
showed that one could carry out epoxidations with these model complexes in the presence of an added oxidant, iodosylbenzene (368,376378). Further studies showed that manganese porphyrin systems
were even more efficient epoxidation catalysts (375, 378). Studies on
these systems also contributed to the development of the oxygen rebound mechanism for oxygen atom transfer from high-valent metal0x0 porphyrins to substrate. Since the early catalysis work, a variety
of MnII'porphyrin systems have been prepared with various modifications designed to improve the stereochemical control of the catalysts
(discussed earlier). Due to the syntheses required to obtain such complexes, simpler, more easily prepared compounds would be desirable
(235). The related Mn"'(sa1en)X family of complexes, where X is frequently a chloride anion, satisfies these requirements. A crystallographically characterized example of this type of complex is shown in
Fig. 29.
In 1986, Kochi and co-workers (379) reported on the epoxidation
capabilities of a series of Mn"(salenlC1 complexes prepared from ringsubstituted salicylaldehyde derivatives and ethylenediamine. Their
work established that these complexes were effective catalysts for the
epoxidation of alkenes. Kochi was unable to crystallographically char-
396
acterize the active oxygen atom transfer species, but based on the
product evidence and analogy to the porphyrin systems, a MnV=O
intermediate was invoked as the oxygen transfer agent. Comparison
to other metal-oxo compounds capable of epoxidizing alkenes also
lent support t o this proposal. Earlier, Kochi's group had explored
other M"+=O complexes, such as (salen)Crv=O (380, 381). This complex could be crystallographically characterized and was shown to catalytically epoxihze some alkenes (380, 381). The Mn-salen systems
showed behavior similar to that of the chromium system, in particular
with the appearance of a new UV-vis band at -530 nm in the presence of the oxidant iodosylbenzene. In the chromium system, iodosylbenzene produces the Crv=O product with the concomitant change in
the W-vis spectrum. Although the recently structurally characterized MnV=O complexes (255, 256) (Fig. 39) are not competent with
respect to epoxidation chemistry, they do lend additional support to
the existence of the MnV=0 intermediate during epoxidation.
Kochi's group used spectroscopy, kinetics, and isotope labeling to
explore these epoxidations and proposed a mechanism for the reaction
(379). Studies of the new W-vis spectrum suggested that it was
not the catalytically active compound and that it resulted from the
initial oxidized manganese product (salen)MnV=0 condensing with a
[Mn"Ysalen)l+ to give a (salen)MnTV-O-MnTV(salen)complex. This
new complex slowly disproportionates to reform [Mn"'(salen)]+ and
(salen)MnV=O. This behavior is consistent with observations made
earlier for porphyrin chemistry. A recent article by Feichtinger and
Plattner (382) provides additional proof for the existence of MnV=O
intermediates in these systems via mass spectrometry, in which they
observe mass peaks that are most consistent with the presence of
MnV=O moieties. With regard to actual epoxidation, Kochi proposed
-R intermediate, wherein a radithe formation of an Mn-0-C-C.
cal is positioned on the carbon p from the manganese, as shown in
Scheme 16. This intermediate is still invoked for the function of these
R
\=I
! -9
Mn(V)L
&
-+ [Mn(lll)L]+
MnL
SCHEME
+A
R
16.
catalysts (discussed later) and is supported by the correlation of the

nature of the substituent on the 5-position of the ligand's phenyl rings
with the reaction rates. The more electron-withdrawing groups
tended to reduce the selectivity of the complex so that it reacted at
397
nearly equivalent rates with a variety of substrates. Furthermore,

complexes with the more electron-withdrawing groups were capable
of oxidizing cyclohexane, which also suggested that a "radical-like"
intermediate forms. The presence of more electron-withdrawing
groups at this position also led to higher epoxide yields. This initial
report also included isotopic labeling studies designed to ascertain the
source of the transferred oxygen atom. When run in dry solvent with
180-labelediodosylbenzene (75% label), the norbornylene oxide product was labeled (75% label) with I S 0 . In a reaction with '*0-iodosylbenzene and 3.5 molar equivalents of added H2160-enrichedwater, the
amount of l80in the product was reduced by 32%. When the labels
are reversed, 160-labelediodosylbenzene, and 3.5 equivalents of H21S0
are added, then 26% of the product shows lS0incorporation. If the
amount of H2180is increased to 10 equivalents, then 74% of the product shows l80incorporation. In addition, under these conditions, the
oxygen of the iodosylbenzene should not be able to exchange with the
water present. These results suggest, that in the normal system run
in the presence of water, the 0x0 of the MnV=O may be exchanged.
Such exchange behavior was reported by Kochi for the Cr systems
(381). Exchange of the oxygen atom in a MnV=O complex in H21B0
has been more recently reported (256).
Other groups have since picked up on this theme and have begun
to explore the potential of these complexes to catalyze the asymmetric
epoxidation of alkenes (374).Notable among these are the Jacobsen
(383-385) and Katsuki research groups (386-388). Thus, there now
exist a wide variety of complexes based on the Mn"'(sa1en)X paradigm
from numerous research groups that have been reported to be competent epoxidation catalysts. These have been predominantly applied to
epoxidizing cis-alkene substrates (235, 365-367). The logic behind
this choice is based on unfavorable steric effects with the manganese
complex. More recently, however, some trans and tetrasubstituted alkenes have been successfully epoxidized, with good enantiomeric excesses observed (367). The broad range of complexes has yielded a
broad range of results as well, with overall yields that range from a
few percent of desired products to well over 90%, with enantiomeric
excesses that cover a comparable range (235, 365-367). Thus, with
the right catalyst, a chemist may easily control the stereochemistry
of a desired alkene and do so in high yield as well. A more recent
report from Jacobsen's laboratories relates the ability of cobalt-salen
catalysts to produce enantiomerically pure epoxides in aqueous systems with excellent stereocontrol (389).Due to the obvious breadth
of this field, with its burgeoning number of catalysts and the varied
applicability based on substrate, we will not attempt to assess the
398
effectiveness of the numerous catalysts that have been reported. Instead, we will narrow our focus to an overview of this type of reactivity. For coverage in more depth, the reader is directed to the various
reviews that have been published over the past few years (235, 365367).
To conduct epoxidation chemistry, the precursor Mn"'(sa1en)X complex must be oxidized to the MnV=O state. This can be accomplished
by a variety of reagents (235, 365-367). Much of the initial work in
this field utilized iodosylbenzene; however, the drawbacks of this reagent have spurred intense efforts to identify more soluble and
cleaner oxygen atom donors. NaOCl has emerged as the oxidant of
choice in several studies, in which the reactions are often run in a
two-phase format, with the substrate and Mn"' catalyst in an organic
layer (often CH,Cl,) and with an aqueous layer bearing the sodium
hypochlorite. Hydrogen peroxide, in large excesses, has seen limited
use, whereas rn-CPBA (rn-chloroperbenzoic acid) has also been utilized at low temperatures in purely organic phase reactions. The results in systems using hydrogen peroxide are often enhanced by the
addition of an N-alkylimidazole. Some newer work has employed atmospheric dioxygen as the oxidant, which is an appealing reagent for
any process that might be considered for industrial scale applications.
In general, the work in this field supports the notion that MnV=O
is the active reagent for the epoxidation of alkenes, with the substrate
presumably approaching in a side-on manner to the MnV=O moiety
(235,365-367). By what mechanism, however, does the actual oxygen
atom transfer occur? The side-on approach was initially proposed to
account for the observed reactivity of unfunctionalized cis alkenes
over trans alkenes with porphyrins (376,377). In the case of the salen
systems, this route of approach has been lent support by the addition
of steric bulk to these complexes and the suggested similarity to the
porphyrin systems (235,365-367). The functionalization of the 3- and
5-positions of the phenyl rings of the ligand has further refined the
proposed approach of the alkene as occurring across the catalyst's ligand backbone. The exact mechanism of the transfer of the oxygen
atom, however, has not been fully worked out (367,390).At least two
proposals have been put forward for the mechanism involved for these
reactions. Arguments for both of these mechanistic positions appeared
back-to-back in Angewandte Chemie articles in 1997 (391,392).
The first mechanistic proposal involves the side-on approach of the
alkene to the catalyst. The radical intermediate, the Mn-O-CC.-R (discussed earlier) then forms (Scheme 16). This @-carbonradical then combines with the oxygen atom, presumably leading to con-
399
comitant product release and reduction of the manganese to the Mn"'

oxidation state. Some earlier work by Jacobsen has been taken to
support the potential for a concerted process in some cases, based
on the epoxidation of trans-2-phenyl-1-vinylcyclopropane(393). The
cyclopropane ring of this substrate should be sensitive to a radical
intermediate, to give ring-opened products. Such products were not
observed. Otherwise, the epoxidation product distributions generally
support the notion of the radical intermediate proposed by Kochi
(379). Furthermore, the epoxidation product distributions are most
consistent with there having been the potential for free rotation about
a previously existing alkene C-C axis before closure of the epoxide
ring. Oxygen atom transfer via a concerted addition across the alkene
bond would not be expected to yield products consistent with rotation
about the previous alkene bond (379). A supporting experiment by
Kochi et al. showed that epoxides, for example (Z)-P-methylstyrene
oxide, added to a n active epoxidation system in the presence of another alkene substrate are not themselves isomerized (379). This result indicates that the observed isomerizations occurred during the
oxygen transfer steps, and not after the epoxide ring had formed.
An alternative viewpoint is based on the earlier work of Sharpless
and co-workers from their research on oxygen transfer reactions with
metal-oxo catalysts (394), for which the existence of metallaoxetane
rings forming during oxygen transfer was proposed. In 1995, a proposal, based on calculations using Macromodel/MM3 (3951, suggested
that such a mechanism might function in the Mn"'(sa1en) systems,
as well. This mechanistic proposal (392, 395) holds that the olefin
approaches the LMnV=O complex, followed by formation of the fourmembered metallaoxetane ring. One donor atom of the salen ligand
is displaced, presumably a phenolate donor, into the site trans to the
former 0x0 oxygen (Scheme 17). This ring then opens into a radical
M-0-C-C.
-R intermediate identical to that invoked for the
radical mechanism and closes to release epoxide and the [Mn"'LI
o L R-[Mn(lll)L]+
+A
MnL
SCHEME
17.
400
precursor complex. Again, bond rotation would be feasible for the

M -0-C -C -R intermediate, leading to the observed mixture of
products. The common factor in both of these proposals is the transient existence of M-O-C-C.-R.
Finally, exploration of nonporphyrin manganese-containing systems that are competent in epoxidizing alkenes are not limited to salen-based systems (396).Tris(acetylacetonato)Mn(III) in the presence
of pivaldehyde and dioxygen was observed to enantioselectively epoxidize cis-alkenes. The dioxygedpivaldehyde system has also been
tested with Mn"'(salen), but appears to be less efficient overall than
systems utilizing sodium hypochlorite (396). An interesting system
(249) based on a [MnTV(Me3tacn)(OMe)l(PF6)
complex has also been
reported. Water-soluble alkene-bearing substrates were readily epoxidized by this system, in which the MnN complex is proposed to activate hydrogen peroxide a t room temperature in the aqueous solution;
however, this complex is not reported to do this in an enantioselective manner.
b. Porphyrins The preparation of manganese-containing porphyrin systems for use as epoxidation catalysts has been intensively studied (372, 373, 375, 397). Research in Mn"'-porphyrin epoxidation
chemistry predated the current generation of the salen systems (discussed earlier). The Mn"'-porphyrin systems do not appear to exhibit
the same degree of broad applicability as the Mn"'(sa1en)X systems.
For porphyrin complexes to function well with respect to generating
an enantiomeric excess of one product, they often require that modifications be made to the porphyrin ligand to create pockets or other
structures that lead to the enantioselective epoxidation of a substrate
(372, 373, 375,397). Often these modifications require extensive syntheses to produce an elaborate chiral porphyrin. This is one drawback
to the porphyrin systems compared with the relatively more easily
synthesized salen ligands. A second drawback is that these complexes
may not be as robust as the salen-based system and have been reported to decompose due to oxidation of the porphyrin (235,365-367).
To combat this latter problem, olefin epoxidations with Mn-porphyrin
catalysts are often run with a large excess of substrate (235, 365).
This is less efficient and potentially quite costly when expensive substrates are to be epoxidized. Overall, the porphyrin systems give results with test substrates, for example styrene, that are on a par with
the results reported for the Mn(sa1en) systems. They have also been
shown to be functional with reagents other than iodosylbenzene such
as sodium hypochlorite and dioxygen (365, 372, 373, 375, 397). An
401
example of a crystallized MnIII-porphyrin system is shown in Fig. 72

(398). This actual complex as shown has been utilized in aziridination
reactions (discussed later), which it proved effective and catalytic.
This same porphyin with a MnlI1-C1 core, instead of Mn"'(0H)
(CH,OH), has been applied to epoxidation catalysis and has been
shown to be one of the more efficient porphyrin epoxidation catalysts
(399).
The mechanisms proposed for manganese-porphyrin-catalyzed
epoxidation of olefins are similar in several respects to that proposed
for salen systems (365, 4001,and this was utilized initially to provide
insight into the Mn-salen systems. In the case of porphyrins, a
MnV=O moiety is also formed by reaction with the secondary oxidant,
for example iodosyl benzene, and the olefin is then believed to approach in a side-on manner. Various intermediates have been proposed for the epoxidation of the olefin, including ones in which the
alkene is polarized to give "positive" and "radical" ends, and those
shown earlier for the salen systems. The concept of the metallaoxetane (371, 372) was employed to explain isomerization of some of the
olefinic substrates during conversion to the epoxide products, but now
FIG. 72. Structure of the Mn'"(P)(OH)(CH,~OH).P = [5,10,15,20-tetrakis

(1,2,3,4,5,6,7,8-octahydro-1,4
: 5,8-dimethanoanthracen-9-yl)porphyrin12.
This complex
has been tested a s a catalytic aziridination catalyst (398). This same porphyrin ligand
in a n Mn"'PC1 complex has been tested a s a n epoxidation catalyst. It is one of the more
efficient porphyrin epoxidation catalysts (399). [Reproduced with permission from (398).
Copyright 1997 the Royal Society of Chemistry.]
402
is regarded as unlikely (400). Thus, the reaction mechanism for the

porphyrin systems also remains quite controversial.
Porphyrin chemistry is not limited to epoxidation chemistry. There
has been considerable recent interest in the toxicology of peroxynitrite, ONOO- (71, 73, 74,401).This dangerous compound is known to
nitrate aromatic amino acids and has been implicated in processes
and disease states as far ranging as amyotrophic lateral sclerosis
(ALS) (402), Parkinson's disease (71, 72), and cell apoptosis. Presumably peroxynitrite is formed by the rapid condensation of superoxide
with nitric oxide. It is now known that peroxynitrite can be efficiently
destroyed by amphiphilic manganese porphyrins in liposomal assemblies (403). The process is believed to proceed by the reaction of Mn'"
(porphyrin) and ONOO- to give Mnv=O(porphyrin) and nitrite,
NO,. The nitrite can then react rapidly to give a MnI"(porphyrin) and
NO2".Subsequent electron transfer to biological reductants destroys
the NO2",and regenerates the catalyst (403,404).
2. Nitrido Group Transferlhiridination
A new entrant on the high-valent manganese reagent scene are the
manganese-nitrido complexes (251-253). Only a handful of these
have been structurally characterized to date, but their potential as a
nitrogen-transfer reagents is being explored. The initial work in this
field again resulted from experiments with porphyrin-based systems.
In 1983, Groves and co-workers generated a Mnv-nitrido compound
and showed that in the presence of trifluoroacetic acid anhydride, this
compound could transfer a nitrogen to produce an aziridine on
cyclooctene (405), the nitrogen analog of an epoxide. The nitrogen
atom of this ring bears a trifluoroacetyl moiety. One promising recent
example of a catalytic porphyrin based aziridination system was reported in 1997 (Fig. 72) (398). This compound produced aziridination
products from alkene precursors, for example styrene, with the aziridine nitrogen bearing a tosyl group from the co-oxidant (tosyliminoiodo)benzene, PhINTs. Up to 480 turnovers were reported, with reported yields ranging from 44-76% and enantiomeric excesses
between 44-66%. The newer nonporphyrin Mn-nitrido systems react
in a stoichiometric manner and are not catalytic like the MnV=O systems. Catalytic systems are now a focus of research with the newer
systems (253). Despite the current lack of catalytic turnover, MnVnitrido systems promise to yield new, useful and easier pathways to
a number of amine or aziridine products. Applications of these complexes to the amination of carbohydrates may allow these compounds
403
to be more readily synthesized in reactions that employ milder reagents than might otherwise be utilized (253).
Initial studies have shown that the transfer of the nitrido nitrogen
atom moiety to a n organic substrate from a nonporphyrin MnVnitrido complex is feasible (251). Both the complex with the underivatized salen2--ligand complex, as well as the tetramethylsalen2-,
saltmen2-, with four methyl groups appended to the two backbone
carbons of the ligand, were prepared. Figure 37 shows the crystal
structure of this complex. The latter was employed in nitrogen-transfer reactions due to its greater solubility in organic solvents and was
found to be a competent nitrogen-transfer agent to a variety of silylated enol substrates in the presence of trifluoroacetic acid anhydride.
This reaction produced a-aminoketone products, with the product
amine acetylated with a trifluoroacetyl group. Studies on these systems continue, with the ultimate goal being the production of a complex that is competent for asymmetric nitrogen transfer, presumably
in a manner similar to the Kochi-Jacobsen-Katsuki epoxidation
format.
Another interesting result in this field is the report that the nitrogen of a (sa1en)Mn"GN to another Mn"'(5-MeOsalchxn)Cl complex,
where chxn represents 1,2-diaminocyclohexane (406). This process
was shown to be reversible by the addition of the Mn"'(sa1en)Cl to a
solution of (5-MeOsalcxhn)Mnv=N, wherupon the N3- could be transferred "back" to regenerate the "original" (salen)Mnv= N complex. Related work has shown that a nitrido can be transferred from a MnV
nitrido(p0rphyrin) complex to a CrIYporphyrin) complex to yield
[Mn"'(porphyrin)l+ and Crvnitrido(porphyrin) (407).
A note on the synthesis of Mnv-nitrido complexes seems warranted
at this point. There now exist several reports of first-row transitionelement nitrido complexes (4081, of which a growing number are manganese complexes. There are also a broad range of syntheses available. Manganese-nitrido complexes have been reported via the irradiation of manganese-azido precursors in several cases, although the
yields vary (258,405,409).Other techniques employ ammonia or ammonium in the presence of a chemical oxidant (251-253). The work of
Carierra et al. (251) with Mn"Ysa1en) derivatives employed ammonium hydroxide and sodium hypochlorite to generate the nitrido complex. In newer complexes with bidentate ligands that expand this
class of compounds, the nitrido complex is prepared via a synthesis
that utilizes N-bromosuccinamide as the oxidant with the addition of
gaseous ammonia to the Mn-compledN-bromosuccinamidesolution at
low temperature (253).
404
3. Chlorination
Many routes for the chlorination of alkenes exist, and one new entry in recent years has been manganese-based chlorinating reagents.
Many of these systems have been prepared by Bellesia and co-workers
(410414). These systems in general produce trans-chlorinated alkenes. In only one system has the proposed halogenating complex
been structurally characterized (245).
Most of the systems explored for the chlorination of alkenes with
manganese have involved high-valent manganese compounds or the
known oxidant Mn"' acetate. The systems reported by Bellesia et al.
(410-412) have been tested against a substantial body of alkenecontaining substrates. One system that these researchers devised was
based on a mixture of Mn11102and Me3SiC1in THF (410, 411), and a
later system was devised that employed MnrV02,Mn"C12, and acetyl
chloride in DMF (412).This group proposed that a transient "MnIVC12
or "Mn*I1C13,,acts as the reactive halogenating reagent in some of their
systems. They also showed that such a Mn-based system could chlorinate ketones in the a-position (415)under milder conditions than earlier work (416).Two other similar systems have also been reported,
the first based on Mn"'acetate, CaC12 and acetyl chloride (417) and
the other on KMn04 and oxalyl chloride (418, 419).
A variety of mechanisms may be functioning for these different systems. In the systems of Bellesia and the Mn"' acetate system of Donnelly et al., it was proposed that these reactions occurred via a nonchain radical mechanism (411, 417). The product distributions
observed are most consistent with chlorination by a radical pathway
and not by an ionic pathway. This is best demonstrated by norbornylene (420-422). The products of chlorination for these systems were
trans-2,3-dichloronorbornaneand a much smaller amount of cis-2,3dichloronorbornane, and no other products were observed. The latter
compounds would result from the generation of a nonclassical carbocation during chlorination. The lack of such products speaks against
an ionic intermediate, and the observed products are consistent with
chlorination having occurred via a radical-based halogenation. A
mechanism involving radical character in this system is given credence by the known chemistry of Mn"' acetate, which is known to
initiate radical-based chemistry (363, 364).
In 1997, Marko and co-workers (423) followed up on their initial
report of the KMn04 and oxalyl chloride system. This new system
consists of KMn04, Me3SiC1, benzytriethylammonium chloride with
dichloromethane as the solvent. It gives good yields of dichlorinated
alkenes. The permanganate anion is brought into solution by the ben-
405
zyltriethylammonium chloride, and the addition of four molar equivalents of MesSiCl produces the halogenating reagent, which they propose to be a dichlorobis(trimethy1silol)manganesecomplex.
None of these systems presents an initial, structurally characterized halogenating agent. Recently we reported that the MnTV(salpn)C12
complex (245),which had been crystallographically characterized previously (Fig. 33), acted as a new reagent capable of halogenating alkenes in a trans manner. In this case a stoichiometry of two dichloro
complexes to one alkene was required, and the final manganese product has been identified as Mn"Ysa1pn)Cl. The products formed with
this reagent are consistent with a nonchain radical mechanism.
B. ENZYME
MODEL
SYSTEMS
1. Manganese Peroxidase
The reactivity of the manganese peroxidase (MnP) system has been
explored with model compounds. The chemistry of MnP is the oneelectron oxidation of the lignin substrate that initiates decomposition
'of the lignin structure. Some of the initial functional modeling chemistry in this field established the ability of Mn"' complexes to initiate
the decomposition of lignin analogs, for example, the decomposition of
dimeric lignin models in the presence of Mnl" and a-hydroxy acids
(95, 103).The products from these experiments involving Mn"' lactate
and a variety of other a-hydroxy acids showed that Mn"' was indeed
capable of initiating the one-electron oxidation of lignin. Another
study with the lignin analog, and test substrate for Lip and MnP,
vanillylacetone also showed that these systems were capable of such
oxidations (424).Furthermore, the lactate was proposed to stabilize
the Mn"' to prevent generation of Mn oxides before the manganese
ion could act on the lignin, yet would not produce a complex so stable
as to prevent the Mn"' lactate complex from oxidizing lignin. More
recent work has shown that lignin-degrading fungi produce oxalate,
which is required for the enzyme to function. Based on these model
studies and data, it has been proposed that oxalate may stabilize the
Mn"' to prevent loss of Mn"' by disproportion to Mn" and MnIVO2.
In these studies, the manganese complex was generated in situ
from added manganese and ligand. In 1991, Saadeh et al. reported
the preparation and isolation of complexes with a-hydroxy acids, for
example lactic acid, 2-hydroxyisobutyric acid (HIB) and 2-hydroxy2-ethylbutyric acid (HEB) (425).The latter two ligands led to Mnw
complexes with 0, first coordination spheres, whereas Mn"' complexes
were prepared and characterized with lactic acid. The crystal structure of the complex with HIB shows it to be two MnwL2units that are
406
bound to a single sodium through the facial array of hydroxy oxygens

in the complex. The monomer is shown in Fig. 34. Among the other
complexes prepared were Mn"' dimers with chelated and bridging lactates or malonates (425).
These isolated and well-characterized complexes were then utilized
to explore the ability of such a-hydroxy acid-manganese complexes to
oxidize the lignin analog vanillylacetone. A scheme representing the
Mn"' system is shown in Scheme 18 (425).It was found that the Mn'"
Vanillylacetone
2 H 2 0 + 4 Mn3+
do
HO
Pyruvaldehyde
Vanillin
4Mn2+ + 4 H +
OCHa
+
0
SCHEME
18.
complex with HIB and the Mn"'-lactate complexes were both capable
of oxidizing this organic substrate. The oxidation product, pyruvaldehyde, is consistent with the oxidation of the substrate via successive
one-electron oxidations by manganese (425).The reaction is proposed
to begin via the generation of a phenoxy radical, which then rearranges to form a ketone from the phenol alcohol and places a radical alpha to the ketone. Two successive Mn"'-initiated oxidations lead
to the formation of pyruvaldehyde. A final oxidation of the remaining
substrate radical, followed by incorporation of the second molecule of
water and a rearrangement, yields vanillin. Furthermore, Mn" was
identified by EPR spectroscopy in these reaction mixtures. This lends
additional strong support to the one-electron oxidant role for manganese in MnP.
A second area of MnP model chemistry has dealt with Mn" oxidation by Fe-porphyrin systems. An early example suggested that Mn"
in the presence of excess pyrophosphate could be catalytically oxidized to Mn"' by an Fe-porphyrin system. In these experiments, sulfonated porphyrins were utilized, and the co-oxidant was potassium
monopersulfate. Co-substrates, such as veratryl alcohol, were reported to enhance the rate of Mn" oxidation (468).
Another interesting reactivity-related system is the recent report of
an iron-containing porphyrin that has been functionalized with propi-
407
onate arms, patterned from the crystallographic data for MnP (426).
This system was designed to study electron transfer from the iron/
porphyrin to the manganese ion. The porphyrin bears pentafluorophenyl groups on three of its meso positions, and the fourth meso
position is modified by the addition of an alkyl arm that terminates
in a 6-(methylamino)-2,2'-bipyridine
functional group that is intended
to be the site of manganese binding. Among other characterization
data, EPR spectra confirm the presence of Fe"' and Mn" in the final
product, and the apparent lack of magnetic interactions between the
two ions suggests that they are not coupled t o one another. In the
presence of the oxidant pentafluoroiodosylbenzene, the Mn" EPR signal was observed to slowly disappear. The addition of 2,6-dimethylphenol, a standard test substrate for MnP, led to the formation of the
oxidation product 2,2',6,6'-tetramethyldiphenoquinone and the reappearance of the Mn" EPR signal. This appears to be fully reversible
and was shown to function in a catalytic manner. In control experiments, neither Mn'I bound to the iron-free modified porphyrin nor
Mn1Ybipy),C12 were oxidized by pentafluoroiodosylbenzene. In the
presence of a (chloro)[tetrakis(pentafluorophenyl)porphyrinliron complex and pentafluoroiodosylbenzene, MnYbipy),Cl:,was observed to be
oxidized. These results are consistent with the petafluoroiodosylbenzene oxidation of the ironlporphyrin portion of the complex with ensuing electron transfer to oxidize the bound Mn". The Mn"' thus produced would then oxidize the substrate.
2. Manganese Superoxide Dismutase
The various complexes that have been tested for their ability to
mimic MnSOD form another set of systems that have been the focus
of reactivity studies. One emphasis of research in this area has been
the potential for the development of pharmaceutical agents that are
good SOD mimics. This research is driven by the proposal that much
of the damage following a shortage of oxygen due to a stroke or heart
attack may be the result of a build-up of superoxide and its hazardous
by-products (discussed earlier). The goal, therefore, is to design
agents to scavenge superoxide and to thereby prevent an appreciable
build-up of this molecule. Much of this work has centered on manganese, because this element is less likely to be able to catalyze other
side reactions, such as Fenton-like chemistry (221,222, 427).
The manganese complexes that have been prepared to date cover a
range of structural types and ligands. Due to the self-dismutation of
superoxide (2.0-3.2 X lo5 M-I. 5-l ) (52, 631, these complexes need to
be quite efficient to qualify as catalysts of superoxide dismutation.
Several complexes have proven to be competent MnSOD mimics. Un-
408
fortunately, one of the main techniques once utilized t o establish SOD

activity suffers from potential interference, so early examples of catalytic MnSOD mimics may not actually be catalytic in nature. Thus,
from all that has been published, it is clear that there are roughly two
classes of compounds: those that are efficient and catalytic mimics for
MnSOD, and those that interact with superoxide but are not necessarily catalytic in their interaction.
In 1993 Weiss, Riley, and co-workers reported a study on purported
SOD mimics by stopped-flow UV-vis spectroscopy (428) in which they
assessed reactivity by following the decay of the superoxide absorption at 245 nm. Two of the earlier techniques that had been used to
assess SOD activity included observation by W-vis spectroscopy of
the oxidation of nitroblue tetrazolium (NBT) (68) or the oxidation of
a cytochrome c by superoxide (52). Both systems used superoxide from
an in situ generator, frequently xanthine oxidase, wherein the complex being analyzed was compared to a calibrated oxidation of the
chromophore alone and in the presence of MnSOD. The direct observation of the decrease in the superoxide signal with time by UV-vis is
also possible, and superoxide may be introduced as a solution (428) or
generated, in some cases, by pulsed radiolysis (79,801.
In these direct
observation experiments, the rate of decay of superoxide in the presence of the complex is compared to the rates of decay of superoxide
alone and in the presence of one unit of activity of MnSOD. In all
cases, the systems are usually referenced, or calibrated, against the
same set of conditions with MnSOD. Due to interactions with cytochrome c with components of assay mixtures other than superoxide,
false readings of activity were often observed for some early SOD
mimics. The NBT, stopped-flow, or pulsed radiolysis techniques have
tended to provide the more accurate answers on the ability of reputed
MnSOD mimics. To be considered active in any manner with respect
to the decay of superoxide in the stopped-flow analyses, Weiss et al.
stated that compounds based on their analyses needed to exhibit keat
values in excess of 105.6M- * s-l (428).
The complex of manganese with desferrioxamine B (429432), reported to be a MnSOD mimic in 1987, illustrates this point. In the
initial studies, the cytochrome c assay was applied, and this compound was reported to be a catalytic MnSOD mimic. Follow-up studies by Riley and co-workers by stopped-flow techniques showed that
this complex interacts at best on a stoichiometric level with superoxide, but it could not be considered to be a catalyst and was termed
inactive (428). This occurred due to interactions with cytochrome
c wherein the manganese complex interfered with the reduction of
cytochrome c , leading in effect to a false positive. Thus, although some
409
systems may no longer be considered to be catalytic, many of these

still may exhibit interactions with superoxide.
Despite earlier systems that eventually proved not to be catalytic,
catalytic MnSOD mimics have been successfully prepared with several examples appearing since 1990. One of the first appeared in 1993
by Kitajima and co-workers (208) and is based on the ligand hydrotris(3,5-diisopropylpyrazol-l-yl)borate,HB-i-Pr. The two fivecoordinate MnSOD mimics that were prepared in this report are the
Mn" complex with one HB-i-Pr ligand and one bidentate benzoate,
and another Mn" complex composed of one HB-i-Pr, one unidentate
benzoate, and one equivalent of 3,5-diisopropylpyrazole.The crystal
structure of the latter is shown in Fig. 13. These complexes yielded
IC,, values of 0.75 and 0.80 pmol-dm-3, respectively. These values
were determined by observing the inhibition of the reduction of nitroblue tetrazolium by superoxide in a calibrated system using xanthine/xanthine oxidase as a superoxide generator. The IC,, value implies that a concentration of that particular reagent exhibits activity
equivalent to 50% of that observed for one unit of actual MnSOD enzyme under the specific conditions of the experiment in question. For
comparison, a n iron complex tested by Weiss et ul. (428) that was
shown to be "active" in the presence of superoxide exhibited a k,,,
value of 2.15 X lo6 M-'. s-l, and a n ICsovalue of 2.1 pmol .dm-3.
Riley, Weiss, and co-workers have prepared a series of MnSOD
mimics using macrocyclic ligands (221,222,427).The first report with
several simple macrocycles indicated that the seven-coordinate Mnl'
complexes with the 15- and 16-member pentaaza macrocycles and
axial chlorides were efficient catalysts for the dismutation of superoxide (Fig. 20) (221). At pH 8.1, the pentaaza and hexaaza complexes
exhibited kcatvalues of 2.2 X lo7 M-' * s-l and 1 x lo6 M-'. s-l, respectively. At a physiological pH, pH = 7.4, only the complex with the
15-member ring was active. The complex with the 16-member ring
proved to be quite pH sensitive. At pH 7.4, the efficiency of the functional complex increased to a kcat value of 4.1 X lo7 M-' s-' . This
initial work was followed up by modification of the periphery of the
macrocyclic ligand to create a family of new MnSOD mimics (222).
These complexes were then screened by stopped-flow UV-vis spectroscopy. All of the complexes tested catalytically decomposed superoxide,
with kcatvalues ranging from 1.4 to 9.1 X lo7 M-' . s-' a t pH 7.4. The
best of these was dichloro(truns-2,3-cyclohexano-1,4,7,10,13-pentaazacyclopentadecane)manganese(II).
A third system is based on Mn"'-Schiff-base complexes. There are
two reports of reactivity of such systems with superoxide. The first,
by Matsushita and Shono, appeared in 1981 (433). In this study, the
410
authors observed that in the presence of superoxide, Mn"'LC1 complexes, where L is a Schiff-base ligand such as salen2-, and in KOd
complex ratios <2-3, oxygenated products were observed for complexes with Mn(IIUI1) reduction potentials of about -0.19 V or lower
vs. a Hg pool (approximately 0.21 V vs. SCE). Complexes with more
positive reduction potentials, ranging from -0.18 to 0.00 V vs. the Hg
pool, were reduced, but oxygenated products were not observed. In
1993, Baudry et al. examined Mn"'(salen)-based systems with ligands
that had been designed for the Mn"'(sa1en)Cl epoxidation systems
(434).The activity of these complexes considerably varied, from rates
that were nearly undetectable to rates indicative of fairly efficient
MnSOD mimics. One of these Mn"'(salenjC1 complexes, with the salen
ligand comprised of a 1,2-diaminocyclohexane backbone and 2-hydroxy-3-tert-butyl-5-methoxy-benzaldehyde,
gave a n ICb0value of 0.32
Fmol. dm-3, or 3200 U/mM. The activity in this system was measured
against the reduction of the nitroblue tetrazolium absorption. In 1997
it was reported that underivatized Mn"'(salen)Cl, referred to as
EUK-8, was also a competent superoxide scavenger (435).
A Mn" complex with a n all-nitrogen first-coordination sphere was
reported in 1996 to be a competent MnSOD mimic (436).In this case
the complex is composed of two meridonally binding 2,6-bis(benzimidazol-2-ylj-pyridine ligands. When tested against the NBT system,
this complex was reported t o exhibit a n LCs0value of 0.72 pmol dm-3.
This level of activity is comparable to that observed for the other systems tested by the nitroblue tetrazolium system.
Most manganese complexes that have been reported to react with
superoxide, however, do not react catalytically. Several have been reported by various groups in attempts to generate catalytic systems,
but these only show stoichiometric interactions with superoxide. Examples of this group of complexes includes the recently reported
seven-coordinate Mn" complex, with the same ligand and a structure
similar to the complex shown in Fig. 25 (229).
In conclusion, systems have been generated that do interact catalytically with superoxide and function as MnSOD mimics. The potential for errant readings of activity with the indirect assays of SOD
mimics shows the need for careful control experiments, and perhaps
the wider application of direct techniques for the determination of
MnSOD mimicry to prevent false positives and to provide readily
compared kcatvalues.
3. Catalase
Work on the catalase frontier has focused on the synthesis of binuclear manganese complexes that are capable of catalyzing the dispro-
411
portionation of hydrogen peroxide. This field has grown considerably

over the past few years, and a large number of catalase mimics are
now known. Catalase and catalase model systems have been reviewed
previously ( 3 , 5 , 2 3 , 2 5 , 2 6 , 2 9 , 3 0 , 4 3 6 In
) . general, two types of catalase chemistry may be mimicked by manganese model complexes. The
first is the reactivity of the catalase itself. The second is the catalaselike reaction thought to occur in the OEC between the Soand S2states
of that enzyme (437).Several model systems have been prepared, and
a selection of these will be described in some detail. A table of rate
data, Table 111, is presented so that these systems may be more
readily compared. A few systems not discussed here are also presented in the table.
One of the earliest functional mimics of catalase was reported by
Mathur et al. in 1987 (438). This system employed the ligand
N , N , N ,N' - tetrakis - (a- methylenebenzimidazolato)- 1,3- diaminopropane-2-01. The complex formed with this ligand is composed of one
ligand, with the alkoxide bridging the two manganese(I1) ions, with a
proposed bridging chloride and two terminal chlorides, one on each
TABLE I11
CATALASE
ACTIVITY:
ENZYMES
AND MODELSYSTEMS"
Enzyme or complex
Thermus thermophilus catalase
260,000
200,000
26,000
250
150
22
13.2
5.5
5.4
4.4
1.2
[Mn~'(20H-benzimpn)(p-OAc)(n-BuOH)I(CI04),
1.04
(BaL+,Ca2),[Mn~llMn"(p-O)(p-OH)(p-OAc)L(tapn),l
0.79
lMneLh(p-C,H,COL)Z(NCS)I
0.26
LMn"'Mn'"(bpg' )l(p-O),I(C10,)
0.013
[Mn"'(tetraphenyIporphynn)l
0.0063
[MnY H,O)&C10, )1
Lactobacillus plantarum catalase

Thermoleophilum album catalase
1Mn"'(sa1pn)(p-0)l2
[MnW-OHpicpn)J,(CIO,),
IMn"(2-OHsalpn)l,
[Mn"lMnl\(p-Oj&-OAc)(tacn)( bipy)(MeOH)1(C104)1
[ Mn"lMnl"(p-0 )2(p-OAc)(tacn
)( 0Ac)z l(CIO,)L
[Mn]" (anthracenediporphyrin)lC12
lMngLh (p-3,4-N01benzoate)1(C10,)
3.2
lo6
5.7 x 1 0 5
1.7 X lo6
1 x 10J
70
350
131, 295,
461, 462
23
129
446
333
440
294
294
443
463
439
447
441, 442
464
443
440
" T h e rates observed for these systems are for conditions which vary from study to study,
whereas the k,,, values are presumably maximal rates with saturating hydrogen peroxide.
" L = the macrocyclic 2 : 2 Schiff base condensation of 2,6-diformyl-4-methylphenol
with N,Nbis(2-aminomethyl)methylamine.
bpg = bis(N,N-(2-pyridylmethyllglycine.
1
412
manganese. This proposed structure is similar to that of the related

structure shown in Fig. 53. In the absence of crystallographic data,
EPR spectroscopy was used to suggesk the binuclear structure, with a
Mn-Mn distance on the order of 3.3 A. When reacted in a two-phase
system of complex dissolved in butanol and a hydrogen peroxidebearing water layer, catalytic dioxygen evolution was observed. Over
the course of the reaction, a Mn$ dimer built up, which was proposed
to be an oxo-bridged product. This product could be independently
prepared for comparison to the product of the catalase-like reaction.
As the reaction continued, the pH rose, and the loss of reactivity over
time was attributed to this increase in pH, which led to complex decomposition. The proposed mechanism for this catalase mimic involved the initial formation of an oxo-bridged Mn$ complex from the
Mn!j complex and the formation of one molecule of water. The second
step involved the reaction of this new Mn"'-O-Mn"' compound with
hydrogen peroxide to reduce the complex to the Mn# state, with the
concurrent formation of a second molecule of water and the release of
dioxygen. The progress of the reaction was followed by collecting the
evolved dioxygen.
In 1994, the chemistry of this system was revisited with an analogous complex (291,439),one in which the chlorides had been replaced
by a bridging acetate and two nonbinding perchlorates. One of the
manganese in this structure is five-coordinate, and the other is six
coordinate, with a butanol occupying the sixth coordination site. The
crystal structure of this complex gives credence to the structure that
had been proposed for the catalyst with chlorides (438).The exploration of this system led to a new mechanistic proposal. In this case, the
lag time for observed catalase activity was reduced by the addition of
2% water to a methanolic solution of the catalyst. They proposed that
the water displaces the bridging acetate; thus, a molecule of hydrogen
peroxide may more easily displace the terminal waters than the
bridging acetate to bridge the two manganese ions itself. The 0-0
bond of the peroxide was then proposed to be cleaved, producing two
Mn"'-OH moieties. This intermediate reacts with a second molecule
of hydrogen peroxide to release dioxygen and to regenerate the Mn"
complex. The terminal hydroxides were protonated during this second
step to form two molecules of water, thus completing the disproportionation of hydrogen peroxide (Scheme 19).
In 1992, Wieghardt and co-workers reported two manganese catalase mimics based on triazacyclononane ligands (294). These were
asymmetric complexes, with one half bearing an N,N',N"-trimethyl1,4,7-triazacyclononane molecule bound to a manganese that was
bridged via two p-0x0 bridges and one p-carboxylato bridge to either
a manganese bearing a bipyridine ligand and a terminal water or
413
2H 2 4
H202
SCHEME
19. Pronosed mechanism of hydrogen peroxide disproportionation by
[Mn"(2-OHbenzimpn)(p-OA~~l~~+.
1994 the American Chemical Society.]
methanol or one coordinated only by additional acetates. The former

complex is shown in Fig. 73. The oxidation state of these systems is
Mnl*IMnN,and presumably they cycle to Mn"MnlI1. In sodium acetatebuffered solutions, approximately 1300 turnovers were realized during the first three minutes of these reactions, with rate values of 13.2
and 5.5 s-l, respectively. Overall, both of these catalysts were five
orders of magnitude slower than the catalase enzyme.
In 1993, Gelasco and Pecoraro (273) reported initial data for a series of complexes based on the 2-OHsalpn ligand, N,N'-(salicylideniminel- 1,3-diaminopropan-2-01(Fig. 48). These complexes are effective catalase mimics. With all of the biologically relevant oxidation
states of the catalase enzymes, MnP, MnIIMnIII, Mniu, and MnlIIMnTV,
having been prepared (279). This is the first series of manganese complexes to exhibit this range of oxidation states with the same ligand.
FIG. 73. Structure of [(Me,tacn)Mnl"~p-O~~(p-OAc~Mn"'(OAc~,l.

[Reproduced with
permission from (294). Copyright 1992 the Royal Society of Chemistry.]
414
Furthermore, this family of complexes exhibited EPR spectra that

modeled those observed for the catalase enzymes (26, 279). The
[Mn(2-OHsalpn)12complexes disproportionate H202by shuttling between the MniI and the MnkI' oxidation levels. Under appropriate conditions in the presence of hydrogen peroxide, either the Mnf or the
Mn;" complex may be produced cleanly. The observed rate for the disproportionation ranged from 4.2 to 21.9 s-l, depending upon the derivatization of the complex's phenolate moieties. The rate-limiting step
was determined to be the reduction of the Mn;" dimer to Mnf. Thus,
complexes with the more electron-withdrawing ligands proved to be
the most efficient catalysts in this system, because this Mn;" oxidation state was less stabilized. The complexes with the more electrondonating ligands, in contrast, tended to bind peroxide better.
In reactions with hydrogen peroxide (4401, members of this family
of complexes exhibit saturation kinetics. It was observed that the nature of the substituents on the phenolates affected the kcat and K,,,
values of the complex with respect to hydrogen peroxide disproportionation. It has been proposed (440) that this system involves terminal binding of the hydroperoxide anion, HOO-, to a [Mn"'(20Hsalpn)12
dimer, in a manner similar to that observed in crystal structures of
these compounds, with, for example, methanol or THF (Fig. 44). In
this case, the hydrogen peroxide is deprotonated by a base equivalent
and is bound to the dimer. One of the alkoxide bridges shifts to become a terminal instead of a briding ligand to open a coordination
site for HOO- (Scheme 20). A second base equivalent then deprotonates this hydroperoxide monoanion unit, the dimer is reduced to
Mnf, and dioxygen forms and is released. A second equivalent of hydrogen peroxide then bridges the Mnkl dimer, followed by reoxidation
of the dimer with the concomitant reduction of this second equivalent
of hydrogen peroxide to form two equivalents of hydroxide. These hydroxide molecules are base equivalents as described earlier, so that
the two water molecules of hydrogen peroxide disproportionation are
formed. Isotope-labeling studies showed that the dioxygen evolved
arose from the same molecule of hydrogen peroxide.
Okawa and co-workers have published several papers in recent
years on binuclear catalase mimics based on the pentadentate or septadentate binucleating ligands, an example of which is shown in
Scheme 21 (441, 442). These ligands have been used to prepare a
variety of binuclear MnH complexes. Their ability to catalyze the disporportionation of hydrogen peroxide varies, but is generally low, on
the order of 0.79 s-l. A functional scheme for the disproportionation
of hydrogen peroxide has been proposed for these complexes that invokes a Mn'"=O intermediate (Scheme 21).
415
SCHEME
20. Proposed mechanism of hydrogen peroxide disproportionation by the
[Mn(S-OHsalpn)j, system. [Reproduced with permission from ( 4 4 0 ) .Copyright 1998 the
Another unique dimeric system was first reported by Naruta and

Maruyama in 1991 (443-445). In this system, two Mn"'-porphyrin
molecules are linked by an anthracene spacer instead of the o-phenylene spacer seen in Fig. 75 (discussed later). Their studies showed this
H
H\
,L
g(,,,)
.
2y
6
H202
Mn(II)Mn(II)
Complex
H202
7-O;
&,)
0
I
I
Mn(LI1) Mn(II1)
k4
("
N
'\
OH
"1
Y\
12H2
02-1
P-7
in(N)
hnw)
7
in(IV)
Mn(N)
II
H202
SCHEME
21. A proposed mechanism of hydrogen peroxide disproportionation by
Okawa, et al.'s system. [Reproduced with permission from (466). Copyright 1995 the
Royal Chemical Society.]
416
complex to be a competent catalase mimic, with a rate of 5.4 s-' observed for the disproportionation of hydrogen peroxide. It was proposed that this system functioned by the formation of two MnIV-OH
porphyrin groups or high-valent Mn-oxo porphyrin groups in the
presence of HzOzand imidazole, which likely acts as a requisite base.
Hydrogen peroxide will then react with this intermediate to yield dioxygen, two molecules of water, and the reduced form of the complex.
Isotopic labeling studies showed that isotope mixing was not observed
in 1: 1 mixtures of l60and la0hydrogen peroxide,
This was followed by an article in 1997 (445), wherein Naruta et al.
prepared a dimer with MnIvporphyrin(OMe)(OMeor OH). The analysis of this system indicated that catalase activity occurred via approach of hydrogen peroxide to the pocket of the dimer, followed by
deprotonation and concomitant reduction of the dimer and oxidation
of the hydrogen peroxide to dioxygen. Their proposed catalytic cycle
is shown in Scheme 22. Similar systems have been used to generate
dioxygen from water (discussed later).
8:
8:
O2
B:
py;+, c;
B:
I
B:
8:
Mn(IV)
H?'
B:
~ - 0
yn(lv)
H202
Mn(lV)I
B:
0:
SCHEME
Naruta, et al.'s system. [Reproduced with permission from (445).Copyright 1997
Elsevier Science.]
The most efficient catalase mimic to date is the [Mnrvsalpn(p-O)la

complex (Fig. 50) (244, 446). This complex cycles between Mn;" and
Mn"'Mn"'. The latter oxidation state is represented by two monomeric
units, as exhibited by ligand-scrambling experiments. This was determined in catalase mimicry experiments by utilizing two complexes
with differing ligand derivatization. When they were reacted with hy-
417
drogen peroxide and isolated, mass spectrometry showed that the resulting material contained manganese dimers with both derivatized
ligands in one molecule. In experiments without added hydrogen peroxide, no ligand mixing was observed. The ability to form these [MrP
salpn(p-0)I2complexes from [Mn"'LI+ precursors also supported this
supposition. The dioxygen produced during the reaction with hydrogen peroxide arises from the same molecule of hydrogen peroxide,
based upon isotope-labeling studies with labeled hydrogen peroxide.
The reoxidation step of this complex's catalytic cycle involves the incorporation of a hydrogen peroxide molecule into the compound in the
form of the new 0x0 bridges. The rate a t which this complex is capable
of disproportionating hydrogen peroxide is 250 s-l, and saturation kinetics are observed. A proposed catalytic cycle for this system is presented in Scheme 23. Although, this complex shuttles to a higher oxi-
H20
LMn(lll)L
+
HO
/Mn(l'')L
SCHEME
[Mn'V(salpn)(p-O)]p.[Reproduced with permission from (27). Copyright 1992 WILEYVCH Verlag.1
418
dation state than is proposed to exist in the catalase, its structural

similarities t o the OEC make it an excellent model for the catalaselike reaction of the OEC that occurs between the So and S2 states of
that enzyme.
Another class of catalase mimics are comprised of tetranuclear complexes, of which at least two examples are known. The first example
was reported in 1990 by Stibrany and Gorun (447). In this complex
(Fig. 741, two manganese ions are bridged by the alkoxide from
the ligand 1,3-diamino-2-hydroxypropane-N,N,N,
W-tetraacetic acid.
The tetramer is then formed when these dimeric units are bridged t o
one another by two acetates and by one 0x0 and one hydroxo bridge,
which hydrogen bond. The overall oxidation state of this compound
is Mn"Mnbl'. This complex was shown to have a rate of hydrogen peroxide disproportionation on the order of 1.04 s-l. A second tetramer
was reported by Gelasco et al. in 1995 (333). In this complex, four
FIG.74. Structure of ~ M n ~ J ' M n 1 1 ~ ~ - O ~ ~ ~ - O H ~ ~ ~anion;

- O A cL) ,=~ L1,3-diamino),12~
propan-2-ol-N,N,N',N,-tetraaceticacid. CaZ' or Ba2+were reported as counterions. [Reproduced with permission from (447).Copyright 1990 Wiley-VCH Verlag.]
419
manganese ions are bridged by the ligand N , N -( picoly1ideneimino)1,3-di-aminopropan-2-01,

2-OHpicpn. This complex appears approximately box shaped (Fig. 63), with one ligand lying along each edge of
the box. The oxidation state of each manganese in this complex is
Mn. This complex disproportionates hydrogen peroxide a t a rate of
150 s-, making it the second most efficient catalase mimic yet reported. It also exhibits saturation kinetics.
4 . Water Oxidation
Dioxygen is at the heart of many biological processes, and without

the generation of dioxygen by green plants and algae, the world be a
much different place. The need for a catalase or a n MnSOD results
from an aerobic existence, for example. The overall oxidation of two
molecules of water to form dioxygen is a four-electron process. Although the oxidation of the OEC itself occurs by four single-electron
events, it has not been determined whether the final oxidation of water to dioxygen is a four-electron process or one that occurs in smaller
steps, either by sequential one-electron oxidations or by 2 twoelectron processes. One newer and intriguing proposal has suggested
that this may occur via hydrogen atom abstraction from water molecules bound to a n OEC manganese ion (182). That would eventually
produce a higher-valent manganese/oxygen moiety such as Mn =0
prior to a final step that forms and releases dioxygen (Schemes 12
and 13). Water oxidation by a variety of inorganic systems (26) and
considerations, such as thermodynamics, for the oxidation of water to
dioxygen by the OEC have been recently reviewed (37).Model chemistry related to the viability of hydrogen atom abstraction is discussed
in the Section II.G, The Oxygen-Evolving Complex. We will focus on
well-defined manganese model complexes that have been implicated
in water oxidation as potential models for the function of the OEC.
Although the OECs tetranuclear manganese cluster is most likely
not light-activated in and of itself, with respect to oxygen evolution,
models have been prepared in which manganese complexes oxidize
water upon irradiation by light. The classical example is the photochemical decomposition of permanganate, MnOl-, a process first reported in 1913 (448). Dioxygen could be generated by irradiating either of two ligand-metal charge-transfer bands at 311 nm or 546 nm.
The 31l-nm band is the more efficient of the two. The release of dioxygen was established to be pH independent in the range from slightly
basic to slightly acidic. Isotopic labeling studies with labeled oxygen
atoms suggested that the oxygen atoms in the product dioxygen originated on the same permanganate anion. Furthermore, an intermedi-
420
ate for this decomposition was proposed that will react with oxidizable organic substrates that were stable in the presence of
permanganate prior to irradiation of the reaction mixture. In these
experiments, dioxygen evolution is curtailed, and the rate of permanganate consumption increases. These data suggested that an intermediate existed during this decomposition process. It was proposed to
bear, in part, a Mnv-peroxide moiety. This system is principally of
historical interest, in that it served as an early functional model for
the OEC. The data for the OEC suggest, however, that such a process
does not occur there.
Other manganese-containing molecules have been synthesized
since that time that have also been claimed to oxidize water upon
irradiation in solution. Two such systems have been prepared by
McAuliffe and co-workers (449-451).These complexes were prepared
with the sale+ ligand, and characterized as [Mn111L(H20)12(C10,)z
compounds wherein the crystal structure shows two planar LMn"'
moieties bridged by two water molecules. When an aqueous solution
of this complex was irradiated, these authors reported that the dioxygen concentration in the solution rose, as measured by an oxygen electrode. Furthermore, they reported this to be a stoichiometric amount
of dioxygen. This reaction was conducted in the presence of hydrogen
atom accepting quinone molecules, which were reported to be reduced
to hydroquinone products. The authors proposed that the dioxygen
produced comes from the bridging water molecules, because conducting this reaction in anhydrous ethanol did not appear to noticeably
retard the rate of dioxygen production. The productive region of light
was found to be the range of 450 to 600 nm, with a maximal rate for
dioxygen evolution at 590 nm. The final manganese complex from the
More recently,
process is proposed to be in the form [(Mn111L)201.
McAuliffe and co-workers have prepared a system in which two
Schiff-base ligands are connected via two naphthyl moieties (451).
This dimeric product also binds water and has been suggested to oxidize water to dioxygen upon irradiation, too.
Because the tetranuclear cluster of the OEC is most likely not photoactivated in and of itself during turnover, manganese systems that
oxidize water without being irradiated first are of interest with respect to understanding the function of this complex enzyme. One system that oxidizes water was prepared and studied by Matsushita and
co-workers (452,453). These complexes are mononuclear MnmL2C12
complexes, where L is a bidentate Schiff base comprised of 5-nitrosalicylaldehyde and an alkylmonoamine, with structures that should be
similar to Mnl"(salpn)Clz (Fig. 33). These researchers showed that
42 1
these complexes are capable of oxidizing water to dioxygen in a noncatalytic process. The complexes of this type exhibit high reduction
potentials, all near 1.0 V (454).Water oxidation was proposed to occur
in the manner shown in Scheme 24, wherein two molecules of water
SCHEME
24.
would be oxidized to form dioxygen, and two Mn"L2 compounds would

be produced. This evolution of dioxygen from water was confirmed by
the observation of 180-enriched dioxygen following the introduction of
isotopically labeled water into a reaction mixture. Titration by AgNO,
indicated that four equivalents of C1- were retained in solution, indicating that chloride oxidation had not occurred. The observed pH
changes in unbuffered reaction mixtures were consistent with the
evolution of four equivalents of proton. The Mn products were confirmed as Mn" complexes; however, no yield was listed. The yield of
dioxygen was only half the maximal possible yield, at neutral pH.
Naruta et al. published a report of dioxygen evolution from water
that was catalyzed by a dimeric-linked manganese-porphyrin dimer
complex (4551, in a system similar to that employed for their catalase
mimic (Fig. 75). This complex is composed of two triphenylporphyrins
linked via an o-phenylene bridge. With this series of complexes, they
were able to show oxidation of water to dioxygen. This was accomplished by preparing a solution of the dimeric porphyrin complex in
5% v/v HzO in acetonitrile in the presence of Bu,NOH. In dry acetonitrile, reversible electrochemistry was observed, but in the aqueous
acetonitrile mixture, a n irreversible current discharge was observed
a t L 1.2 V vs. Ag/Ag+.Dioxygen evolution was observed over a voltage
range of 1.2 to 2.0 V; however, in the absence of the complex, anodic
oxidation of the watedacetonitrile mixture did not evolve dioxygen up
to +2.3 V. The rate of dioxygen evolution was found to correlate with
the concentration of the manganese complex in solution, and up to 9.2
turnovers were achieved. Under the same conditions, the monomeric
analogs of these dimeric complexes did not evolve dioxygen. By mass
spectrometry, the authors observed that a 1: 1 mixture of HZl6Oand
HZI80in the reaction led to a 1: 2 : 1 mixture of lSOz,160180,
and 180z,
respectively. Finally, they reported that 3.7 electrons were involved
42 2
12+
1 Ar = 4-1BuC6H1
2 Ar = 2,4.6-Me3C6H2
3 Ar=C6FS
FIG.75. Representation of the linked porphyrin catalyst utilized for water oxidation.
Replacement of the o-phenylene linker with a n anthracene linker generates the catalase mimic reported in the catalase segment (443).The authors have noted that reactivity can be altered by the nature of the linker (444).[Reproduced with permission from
(455).Copyright 1994 Wiley-VCH Verlag.1
in this reaction overall, as determined by a rotating disk electrode

and a Levich plot.
Two mechanistic proposals could be made for this dioxygen evolution. The first invokes 2 two-electron processes. Because similar dimers had been shown to disproportionate hydrogen peroxide (443445) (discussed earlier), this proposal would rely on the formation of
hydrogen peroxide by a dimer, followed by disproportionation of that
hydrogen peroxide by one of the dimers in solution. The second proposal invokes a four-electron process. The authors found no evidence
of hydrogen peroxide evolution in their system in the absence of a
manganese dimer, either by addition of a dimer to a water/acetonitrile
solution after anodic oxidation or by analysis by a hydrogen-peroxidesensitive electrode. Thus, they proposed that a four-electron process
was operative. Their mechanism involves the coordination of two hydroxides to the manganese of the dimer on the inside of the dimer
cavity. Oxidation to MnV=O moieties is then followed by formation of
a Mnw-O-O-Mnw
compound. They proposed that coordination of
hydroxide to the manganese could then occur with loss of the p1,2peroxo bridge as dioxygen to return the dimer to the initial Mni*'oxidation state.
A more recent system in which water oxidation has been observed
via a manganese complex (456) is a manganese system based on the
ligand dipicolinate (dpa). When [Mn**'(dpa)2]reacts with oxone, PO-
423
tassium peroxymonosulfate, a rather strong oxidant, the generation

of dioxygen, and the formation of MnOl- was observed. These reactions were carried out in buffered aqueous solutions, over a pH 3-6
range, with a n increasing rate of permanganate anion production
noted for higher pH levels. The Mn"' complex appears to be a precursor to a reactive intermediate, which has been proposed to be a
bis(p-0" )-bridged dimer with a terminal water on each of the manganese, perhaps with a structure similar t o that shown in Fig. 55. This
green intermediate complex exhibited a Mnl*'MnTV
16-line EPR signal
and had a distinctive UV-vis spectrum, which converted isosbestically
to that of Mn04 during the course of the reaction. Below pH 3.5, this
system catalytically generated dioxygen. Finally, in control experiments, this system did not appear to generate dioxygen from hydrogen peroxide formed via oxone hydrolysis in the presence of other
Lewis acids, such as Ni"/sulfate. This also implied that a redox process involving the manganese complex was producing the observed dioxygen.
To follow up on these observations, some additional ligand systems
were screened, including 2,2 '-bipyridine (bpy), 1,lO-phenanthroline
(phen), picolinate (pic), and 2.2' : 6',2"-terpyridine (terpy). In the cases
with the bidentate ligands, fewer turnovers leading to dioxygen evolution compared to the dpa were observed. In the case of the tridentate
terpy, many more turnovers were achieved and no appreciable concentration of Mn04- was observed. The authors concluded that the
meridionally coordinating tridentate ligands were ideal for promoting
the reactivity that they observed, and they attributed the robust nature of the terpy system to a lesser degree of ligand dissociation versus dpa, pic, or bipy (456) because the ligand must come off of the
manganese ion before it may be converted to a permanganate.
Two mechanistic proposals for this process were suggested. The
first requires oxygen atom transfer and the production of a n MnV=O
moiety; the second is based on the hydrolysis of a manganese-bound
peroxymonosulfate anion to yield a Mn-0-0moiety. I n the first
case, MnV=O would combine with either a second equivalent of
MnV=O or a n oxone molecule in a n overall four-electron process. In
the second, a two-electron process leads to the release of dioxygen
from the Mn-0-0complex. Neither of these possibilities has yet
to be ruled out, and the authors state that they are following up with
isotope-labeling studies to ascertain what the reaction mechanism
may be. Some preliminary evidence does, however, suggest that oxygen transfer may be involved. By utilizing dimethyldioxirane, a reagent capable of oxygen atom transfer, but not the transfer of a per-
424
0x0 group, in the presence of Mn" and terpy, the authors observed the
production of dioxygen.
In conclusion, the reactivity of manganese complexes has proved to
be a fertile area for research. It has provided new insights into the
function of manganese-containing redox enzymes, and it has provided
new synthetic techniques to the field of organic chemistry.
VI. Conclusion
The biological applications of manganese are numerous and quite

varied. Although most of these biological roles utilize manganese in
structural or hydrolytic systems, some very striking redox systems
are also known. The hydrolytic roles are, of course, not to be overlooked with respect to the widespread and critical arginase enzyme,
which exhibits a strict requirement for two Mn" ions. The ribonuclease hydrolases of retroviral reverse transcriptase are another
key nonredox application-for example, the ribonuclease hydrolase
from HIV-1.
The manganese-containing redox enzymes are quite unique. Most
of these deal with dioxygen metabolism in one form or another. One
key enzyme is the OEC, which generates the dioxygen required by
aerobic life forms to exist by catalytically oxidizing water. New evidence now suggests that this enzyme may soon join the ranks of other
known metalloenzymes that employ radical protein residues to complete their catalytic cycle. Catalase disproportionates hydrogen peroxide, and the MnSOD catalyzes the dismutation of superoxide. Both
are critical for protecting organisms from oxidative damage, one of
the hazards of aerobic life. Meanwhile, the newly discovered dioxygenases operate on aromatic-bearing cis-diol functionalities and open
these rings by catalyzing the incorporation of dioxygen into them. All
of these enzymes are driven by the oxophilicity of manganese, which
is important in defining the chemistry that occurs. Another interesting feature of several of these enzymes is their iron analogs. MnSOD
and Mn-dioxygenase have direct iron enzyme analogs, which in the
case of SOD, are known to have nearly identical active sites, and in
some cases (but not all) the metal from one may be added to the other
with retention of enzyme activity. In addition, MnP bears analogy to
the Lip enzymes, which rely solely on heme, although new evidence
shows that they, too, may be capable of oxidizing manganese.
However, one notable exception stands out, and that is the OEC.
This enzyme is functional only with manganese. With so many man-
425
ganese enzymes that bear similarities to enzymes that conduct the

same or similar reaction with different metals at their active site,
why is manganese so strictly required? If the newer proposals for the
function of the OEC are correct, then an answer may emerge that
suggests that manganese allows the proton-coupled oxidations necessary for water oxidation to occur at a potential at which the oxidizing
equivalents may be easily stored. Furthermore, there are sufficient
readily available oxidation states available to allow the sequential oxidations to occur, yet despite this stability, once the high-valent
MnV=O is reached, it is not so stable as to prevent the formation
of dioxygen.
Finally, with the acknowledged similarities that currently exist between several manganese enzymes and counterparts utilizing other
metal ions, it seems quite likely that other redox-active manganese
enzymes will be discovered in the future. The discussions of data
gathering on enzyme systems and model chemistry, both structural
and functional, as presented in this article will provide a foundation
for exploring current and as yet undiscovered manganese enzymes.
ACKNOWLEDGMENTS
We wish to acknowledge the NIH for funding our research endeavors (GM 39406 to
V.L.P.). We also wish to thank Professor G. T. Babcock for providing us with a copy of
Scheme 12 and Professor V. V. Barynin for the provision of data on the crystallographic
details of the T. thermophilus enzyme.
IN PROOF
NoTE ADDED
A recent article on the T. thermophilus manganese-containing catalase enzyme presents data that indicates that the structure of the active site is affected by pH. The
experiments were conducted over the pH range 6.6 to 9.8. The authors propose that a t
high pH a Mn"12(p-O)2(p-OAc)
form predominates, while a t low pH a Mn"'2(p-O)(pOAc) form predominates. The latter form has only a single 0x0-bridge with a proposed
hydroxide and water ligand to fill the remaining coordination sites on the Mn"' ions,
one site per Mn"' (478).
REFERENCES
1. Wedler, F. C. In "Manganese in Health and Disease"; Klimis-Tavantzis, D. J., Ed.;
CRC Press, Inc., Ann Arbor, 1994, 1.
2. Scrutton, M. C. In "Manganese in Metabolism and Enzyme Function"; Schramm,
V. L., and Wedler, F. C., Eds.; Academic Press, Inc., New York, 1986, 147.
426
Dismukes, G. C. Chem. Reu. 1996,96, 2909.

Wilcox, D. Chem. Reu. 1996,96, 2435.
Christianson, D. W. Prog. Biophys. Mol. Biol. 1997, 67, 217.
Reczkowski, R. S.; Ash, D. E. J . Am. Chem. SOC.1992,114, 10992.
Khangulov, S. V.; Pessiki, P. J.; Barynin, V. V.; Ash, D. E.; Dismukes, G. C. Biochemistry 1995,34, 2015.
8. Kanyo, Z. F.; Scolnick, L. R.; Ash, D. E.; Christianson, D. W. Nature 1996, 383,
3.
4.
5.
6.
7.
554.
9. Stemmler, T. L.; Sossong, Jr., T. M.; Goldstein, J. I.; Ash, D. E.; Elgren, T. E.;
Kurtz, Jr., D. M.; Penner-Hahn, J. E. Biochemistry 1997, 36, 9847.
10. Sossong, Jr., T. M.; Khangulov, S. V.; Cavalli, R. C.; Soprano, D. R.; Dismukes,
G. C.; Ash, D. E. JBZC 1997,2, 433.
11. Davies, 11, J . F.; Hostomska, Z.; Hostomsky, Z.; Jordan, S. R.; Matthews, D. A.
Science 1991, 252, 88.
12. Cowan, J . A. JBZC 1997,2, 168.
13. Strater, N.; Lipscomb, W. N.; Klabunde, T.; Krebs, B. Angew. Chem., Znt. Ed. Engl.
1996,35, 2024.
14. Hardman, K. D.; Ainsworth, C. F. Biochemistry 1972, 11, 4910.
15. Hardman, K. D.; Agarwal, R. C.; Freiser, M. J . J. Mol. Biol. 1982, 157, 69.
16. The Lectins: Properties, Functions, and Applications in Biology and Medicine;
Liener, I. E., Sharon, N., and Goldstein, I. J., Eds.; Academic Press: New York,
1986.
17. Sharon, N.; Lis, H. Lectins; Chapman and Hall: New York, 1989.
18. Naismith, J. H.; Emmerich, C.; Habash, J.; Harrop, S. J.; Helliwell, J . R.; Hunter,
W. N.; Raftery, J.; Kalb, A. J.; Yariv, Y. Acta Cryst. 1994,050, 847.
19. Antanaitas, B. C.; Chasteen, N. D.; Freedman, J. H.; Koenig, S. H.; Lilienthal,
H. R.; Peisach, J.; Brewer, C. F. Biochemistry 1987,26, 7932.
20. Reed, G. H.; Cohn, M. J . J . Biol. Chem. 1970,245, 662.
21. Derewenda, Z.; Yariv, J.; Helliwell, J. R.; Kalb (Gilboa), J. R.; Dodson, E. J.; Papiz,
M. Z.; Wan, T.; Campell, EMBO J. 1989, 8, 2189.
22. Deacon, A.; Gleichmann, T.; Kalb (Gilboa), A. J.;Price, H.; Raftery, J.; Bradbrook,
G.; Yariv, J.; Helliwell, J. R. J . Chem. Soc., Faraday Trans. 1997, 4305.
23. Penner-Hahn, J. E. In Manganese Redox Enzymes; Pecoraro, V. L., Ed.; VCH
Publishers, Inc.: New York, 1992, 29.
24. Penner-Hahn, J. E. In Structure and Bonding; Hill, H. A. O., Sadler, P. J., and
Thomson, A. J., Eds.; Springer Verlag: Berlin, 1998,90, 1.
25. Pecoraro, V. L.; Gelasco, A.; Baldwin, M. J . In Mechanistic Bioinorganic Chemistry; Thorp, H. H., and Pecoraro, V. L., Eds.; ACS Books: Washington, DC, 1995,
p. 265.
26. Pecoraro, V. L.; Baldwin, M. J.; Gelasco, A. Chem. Reu. 1994, 94, 807.
27. Pecoraro, V. L. In Manganese Redox Enzymes; Pecoraro, V. L., Ed.; VCH Publishers: New York, 1992, 197.
28. Manganese Redox Enzymes; Pecoraro, V. L., Ed.; VCH Publishers, Inc.: New
York, 1992.
29. Vincent, J. B.; Christou, G. In Advances in Inorganic Chemistry; Sykes, A. G.,
Ed.; Academic Press, Inc.: New York, 1989,33, 197.
30. Wieghardt, K. Angew. Chem., Znt. Ed. Engl. 1989,28, 1153.
31. Wieghardt, K. Angew. Chem., Znt. Ed. Engl. 1994,33, 725.
32. Yachandra, V. K.; DeRose, V. J.; Latimer, M. J.; Mukerji, I.; Sauer, K.; Klein,
M. P. Science 1993,260, 675.
427
33. Yachandra, V. K.; Sauer, K.; Klein, M. P. Chem. Rev. 1996,96, 2927.
34. Manchanda, R.; Brudvig, G. W.; Crabtree, R. H. Caord. Chem. Rev. 1995,144, 1.
35. Yocum, C. F. In Manganese Redox Enzymes; Pecoraro, V. L., Ed.; VCH: New
York, 1992, 71.
36. Thorp, H. H.; Brudvig, G. W. New J. Chem. 1991, 15, 479.
37. Ruttinger, W. F.; Campana, C.; Dismukes, G. C. J. Am. Chem. SOC.1997, 119,
6670.
38. Pecoraro, V. L.; Gelasco, A,; Baldwin, M. J. In Bioinorganic Chemistry: An Inorganic Perspective of Life; K. D. P., Ed.; Kluwer Academic Publishers: Amsterdam,
1995,459, 287.
39. Ghanotakis, D. F.; Yocum, C. F. Annu. Rev. Plant Phys. Plant Mol. Biol. 1990,
41, 255.
40. Pecoraro, V. L. Photochem. Photobiol. 1988,48, 249.
41. Diner, B. A.; Babcock, G. T. In Oxygenic Photosynthesis: The Light Reactions;
Ort, D. R., and Yocum, C. F., Eds.; Kluwer Academic Publishers: Boston, 1996,
213.
42. Debus, R. J . Biochim. Biophys. Acta 1992,1102, 269.
43. de Paula, J. C.; Beck, W. F.; Brudvig, G. W. New J. Chem. 1987, 11, 103.
44. Christou, G. Ace. Chem. Res. 1989,22, 328.
45. Brudvig, G. W.; Crabtree, R. H. In Progress in Inorganic Chemistry; Lippard,
S. J., Ed.; John Wiley & Sons: New York, 1989, 37, 99.
46. Britt, R. D. In Oxygenic Photosynthesis: The Light Reaction; Ort, D. R., and
Yocum, C. F., Eds.; Kluwer Academic Publishers: Netherlands, 1996, 137.
47. Babcock, G. T.; Espe, M.; Hoganson, C.; Lydakis-Simantiris, N.; McCracken, J.;
Shi, W.; Styring, S.; Tommos, C.; Warncke, K. Acta Chem. Scand. 1997, 51, 533.
48. Oxygenic Photosynthesis: The Light Reactions; Ort, D. R., and Yocum. C. F.,
Eds.; Kluwer Academic Publishers: Boston, 1996.
49. Hage, R.; Iburg, J . E.; Kerschner, J.; Koek, J . H.; Lempers, E. L. M.; Martens,
R. J.; Racherla, U. S.; Russell, S. W.; Swarthoff, T.; van Vliet, M. R. P.; Warnaar,
J . B.; van der Wolf, L.; Krijnen, B. Nature 1994,269, 637.
50. Armstrong, W. H. In Manganese Redox Enzymes; Pecoraro, V. L., Ed.; VCH
Publishers: New York, 1992, 261.
51. Photosynthesis: From Light to Biosphere; Mathis, P., Ed.; Kluwer Academic
Publishers: Amsterdam, The Netherlands, 1995.
52. Keele, Jr., B. B.; McCord, J . M.; Fridovich, I. J. Biol. Chem. 1970, 245, 6176.
53. Ludwig, M. L.; Pattridge, K. A,; Stallings, W. C. In Manganese in Metabolism
and Enzyme Function; Schramm, V. L., and Wedler, F. C., Eds.; Academic Press,
Inc.: New York, 1986, 405.
54. Stallings, W. C.; Pattridge, K. A.; Strong, R. K.; Ludwig, M. L. J. Biol. Chem.
1984,259, 10695.
55. Stallings, W. C.; Pattridge, K. A.; Strong, R. K.; Ludwig, M. L. J. B i d Chem.
1985,260, 16424.
56. Stallings, W. C.; Powers, T. B.; Pattridge, K. A,; Fee, J . A,; Ludwig, M. L. Proc.
Natl. Acad. Sci. U.S.A. 1988.80, 3884.
57. Tierney, D. L.; Fee, J . A.; Ludwig, M. L.; Penner-Hahn, J . E. Biochemistry 1995,
34, 1661.
58. Lah, M. S.; Dixon, M. M.; Pattridge, K. A,; Stallings, W. C.; Fee, J . A.; Ludwig,
M. L. Biochemistry 1995,34, 1646.
59. Barra, D.; Schinina, M. E.; Bannister, W. H.; Bannister, J . V.; Bossa, F. J . B i d .
Chem. 1987,262, 1001.
428
LAW, CAUDLE, AND P E C O W O
60. Harris, J. I.; Auffret, A. D.; Northrop, F. D.; Walker, J. E. Eur. J. Biochem. 1980,
106, 297.
61. Beyer, W. F.; Fridovich, I. J. Biol. Chem. 1991,263, 303.
62. McCord, J. M.; Fridovich, I. J. Biol. Chem. 1969, 244, 6049.
63. Valentine, J, S. In Bioinorganic Chemistry; Bertini, I., Gray, H. B., Lippard,
S. J., and Valentine, J. S., Eds.; University Science Press: Mill Valley, California, 1994.
64. Youn, H.-D.; Youn, H.; Lee, J.-W.; Yim, Y.-I.; Lee, J. K.; Hah, Y. C.; Kang, 53.-0.
Arch. Biochem. Biophys. 1996,334, 341.
65. Fridovich, I. J. Biol. Chem. 1989,264, 7761.
66. McCord, J. M.; Fridovich, I. J. Biol. Chem. 1968,243, 5753.
67. Morel, F.; Doussiere, J.; Vignas, P. V. Eur. J. Biochem. 1991,201, 523.
68. Beauchamp, C.; Fridovich, I. Anal. Biochem. 1971,44, 276.
69. Davies, K. J. A. In Free Radicah and Oxidative Stress: Environment, Food and
Drug Additives; Rice-Evans, C., Halliwell, B., and Lunt, G. G., Eds.; Portland
Press: London, 1994, 61, 1.
70. Afanasev, I. B. Superoxide Ion: Chemistry and Biological Implications; CRC
Press, Inc.: Boca Raton, 1989.
71. Darley-Usmar, V.; Wiseman, H.; Halliwell, B. FEBS Lett. 1995,369, 131.
72. Jenner, P.; Olanow, C. W. Neurology 1996,47, S161.
73. Dawson, V. L.; Dawson, T. M. J. Chem. Neuroanat. 1996,10, 179.
74. Beckman, J. S.; Beckman, T. W.; Chen, J.; Marshall, P. M.; Freeman, B. A. Proc.
Natl. Acad Sci. U.S.A. 1990,87, 1620.
75. Borgstahl, G. E. 0.; Parge, H. E.; Hickey, M. J.; Beyer, W. F.; Hallewell, R. A,;
Tainer, J. A. Cell 1992, 71, 107.
76. Ludwig, M. L.; Metzger, A. L.; Pattridge, K. A.; Stallings, W. C. J. Mol. Biol. 1991,
219, 335.
77. Fee, J. A.; Shapiro, E. R.; Moss, T. H. J. Biol. Chem. 1976,251, 6157.
78. Bull, C.; Neiderhoffer, E. C.; Yoshida, T.; Fee, J. A. J. Am. Chem. SOC. 1991,
113, 4069.
79. McAdam, M. E.; Fox, R. A,; Lavelle, F.; Fielden, E. M. Biochem. J. 1977, 165, 71.
80. McAdam, M. E.; Fox, R. A,; Lavelle, F.; Fielden, E. M. Biochem. J. 1977, 165, 81.
81. Osman, R.; Basch, H. J. Am. Chem. SOC.1984, 106, 5710.
82. Pick, M.; Rabani, J.; Yost, F.; Fridovich, I. J. Am. Chem. SOC.1974, 96, 7329.
83. Valentine, J. S.; Quinn, A. E. fnorg. Chem. 1976, 15, 1997.
84. Bull, C.; Fee, J. A. J. Am. Chem. SOC.1985, 107, 3295.
85. Whiting, A. K.; Boldt, Y. R.; Hendrich, M. P.; Wackett, L. P.; Que, L. Biochemistry
1996,35, 160.
86. Que, Jr., L.; Widom, J.; Crawford, R. L. J. Biol. Chem. 1981,256, 10941.
87. Que, L. J. In Iron Carriers and Iron Proteins; Loehr, T. M., Ed.; VCH Publishers,
Inc.: New York, 1989,5, 467.
88. Lipscomb, J. D.; Orville, A. M. In Metal Ions in Biological Systems; Sigel, H.,
and Sigel, A., Eds.; Marcel Dekker, Inc.: New York, 1992,28, 243.
89. Que, Jr., L.; Ryn, H. Chem. Rev. lB96,96, 2607.
90. Boldt, Y. R.; Sadowsky, M. J.; Ellis, L. B. M.; Que, Jr., L.; Wackett, L. P. J. Bacterial. 1995, 177, 1225.
91. Han, S.; Eltis, L. D.; Timmis. K. N.; Muchmore, S. W.; Bolin, J. T. Science 1995,
270, 976.
92. Boldt, Y. R.; Whiting, A. K.; Wagner, M. L.; Sadowsky, M. J.; Que, Jr., L.; Wackett, L. P. Biochemistry 1997, 36, 2147.
429
93. Shu, L. J.; Chiou, Y. M.; Orville, A. M.; Miller, M. A,; Lipscomb, J . D.; Que, Jr.,
L. Biochemistry 1995, 34, 6649.
94. Glenn, J. K.; Gold, M. H. Arch. Biochem. Biophys. 1985,242, 329.
95. Glenn, J. K.; Akileswaran, L.; Gold, M. H. Arch. Biochem. Biophys. 1986,251, 688.
96. Sundaramoorthy, M.; Kishi, K.; Gold, M. H.; Poulos, T. L. J. Biol. Chem. 1994,
269, 32759.
97. Sarkanen, K. V.; Ludwig, C. H. Lignins: Occurrence, Formation, Structure and
Reactions; Wiley-Interscience: New York, 1971.
98. Gold, M. H.; Wariishi, H.; Valli, K. In Biocatalysis in Agricultural Biotechnology;
Whitaker, J . R., and Sonnet, P. E., Eds.; American Chemical Society: Washington,
DC, 1989, 127.
99. Crawford, R. L. Lignin Biodegradation and Transformation; John Wiley & Sons:
New York, 1981.
100. Eggert, C.; Temp, U.; Eriksson, K.-E. L. FEBS Lett. 1997, 407, 89.
101. Youn, H.-D.; Yung, C. H.; Sa-Ouk, K. FEMS Microbiol. Lett. 1995, 132, 183.
102. Hammel, K. E. In Degradation of Environmental Pollutants by Microorganisms
and Their Metalloenzymes; Sigel, H., and Sigel, A., Eds.; Marcel Dekker, Inc.:
New York, 1992.
103. Wariishi, H.; Valli, K.; Gold, M. H. Biochemsitry 1989, 28, 6017.
104. Wariishi, H.; Valli, K.; Gold, M. H. J. Biol. Chem. 1992, 267, 23688.
105. Kuan, I.-C.; Johnson, K. A,; Tien, M. J . Biol. Chem. 1993,268, 20064.
106. Kuan, I.-C.; Tien, M. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 1242.
107. Khindaria, A,; Barr, D. P.; Aust, S. D. Biochemistry 1995, 34, 7773.
108. Timofeevski, S. L.; Aust, S. D. Biochem. Biophys. Res. Commun. 1997,239, 645.
109. Zapanta, L. S.; Tien, M. J. Biotech. 1997, 53, 93.
110. Pasczcynski, A,; Huyhn, V.-B.; Crawford, R. L. Arch. Biochem. Biophys. 1986,
244, 750.
111. Auling, G.; Follman, H. In Metal Ions in Biological Systems; Sigel, H., and Sigel,
A,, Eds.; Marcel Dekker, Inc.: New York, 1994, 30, 131.
112. Stubbe, J . J. Biol. Chem. 1990, 265, 5329.
113. Sjoberg, B.-M. Structure 1994, 2, 793.
114. Sjoberg, B.-M. In Structure and Bonding; Hill, H. A. O., Sadler,
P. J., and Thomson, A. J., Eds.; Springer Verlag: Berlin, 1997, 88, 139.
115. Mao, S. S.; Holler, T. P.; Yu, G. X.; Bollinger, Jr., J . M.; Booker, S.; Johnston,
M. I.; Stubbe, J. Biochemistry 1992, 31, 9733.
116. Reichard, P. Science 1993,260, 1773.
117. Schimpf-Weiland, G.; Follman, H.; Auling, G. Biochem. Biophys. Res. Commun.
1981.102, 1276.
118. Willing, A.; Follmann, H.; Auling, G. Eur. J. Biochem. 1988, 170, 603.
119. Uhlin, U.; Eklund, H. J . Mol. Biol. 1996, 262, 358.
120. Uhlin, U.; Eklund, H. Nature 1994,370, 533.
121. Nordlund, P.; Sjoberg, B.-M.; Eklund, H. Nature 1990,345, 593.
122. Nordlund, P.; Eklund, H. J. Mol. Biol. 1993, 232, 123.
123. Willing, A,; Follmann, H.; Auling, G. Eur. J. Biochem. 1988, 175, 167.
124. Atta, M.; Nordlund, P.; Aberg, A,; Eklund, H.; Fontecave, M. J . B i d . Chem. 1992,
267, 20682.
125. Barynin, V. V.; Hempstead, P. D.; Vagin, A. A.; Antonyuk, S. V.; Melik-Adamyn,
W. R.; Lamzin, V. S.; Harrison, P. M.; Artymiuk, P. J . J. Inorg. Biochem. 1997,
67, 196.
126. Griepenburg, U.; Lapmann, G.; Auling, G. Free Rad. Res. 1996,263, 473.
430
Beyer, Jr., W. F.; Fridovich, I. Biochemistry 1985,24, 6460.

Kono, Y.; Fridovich, I. J. Biol. Chem. 1983,258, 6015.
Algood, G. S.; Perry, J. J. J. Bacteriol. 1986, 168, 563.
Barynin, V. V.; Grebenko, A. I. Dokl. Akad. Nauk. S.S.S.R 1986,286, 461.
Barynin, V. V.; Vagin, A. A.; Melik-Adamyan, V. R.; Grebenko, A. I.; Khangulov,
S. V.; Popov, A. N.; Andrianova, M. E.; Vainshtein, B. K. Dolk. Akad. Nauk.
S.S.S.R 1986,288, 877.
132. Sheats, J. E.; Czernuszewicz, R. S.; Dismukes, G. C.; Rheingold, A. L.; Petrouleas,
V.; Stubbe, J.; Armstrong, W. H.; Beer, R. H.; Lippard, S. J. J. Am. Chem. SOC.
127.
128.
129.
130.
131,
1987,109, 1435.
133. Wieghardt, K.; Bossek, U.; Ventur, D.; Weiss, J . J. Chem. SOC.,Chem. Commun.
1985,347.
134. Riggs-Gelasco, P. J.; Stemmler, T.; Penner-Hahn, J . E. Coord. Chem. Rev. 1995,
144, 245.
135. Waldo, G. S.; Yu, S. Y.; Penner-Hahn, J. P. J. Am. Chem. SOC.1992, 114, 5869.
136. Fronko, R. M.; Penner-Hahn, J . E. J.Am. Chem. SOC.1988,110, 7554.
137. Waldo, G. S.; Penner-Hahn, J. E. Biochemistry 1995,34, 1507.
138. Khangulov, S. V.; Voyevodskaya, N. V.; Varynin, V. V.; Grebenko, A. I.; MelikAdamyan, V. R. Biofizika 1987,32, 960.
139. Khangulov, S. V.; Goldfeld, M. G.; Gerasimenko, V. V.; Andreeva, N. E.; Barynin,
V. V.; Grebenko, A. I. J. Inorg. Biochem. 1990,40, 279.
140. Zheng, M.; Khangulov, S. V.; Dismukes, G. C.; Barynin, V. V. Inorg. Chem. 1994,
33, 382.
141. Kurtz, Jr., D. M. JBIC 1997,2, 159.
142. Diesenhofer, J.; Epp, 0.;Miki, K.; Huber, R.; Michel, H. Nature 1985,318, 618.
143. Michel, J.; Diesenhofer, J. Biochemistry 1988,27, 1.
144. Ort, D. R.; Yocum, C. F. In Oxygenic Photosynthesis: The Light Reactions; Ort,
D. R., and Yocum, C. F., Eds.; Kluwer Academic Publishers: Boston, 1996, 1.
145. Adelroth, P.; Lindberg, K.; Andreasson, L.-E. Biochemistry 1995,34, 9021.
146. Yocum, C. F. Biochim. Biophys. Acta 1991,1059, 1.
147. Lindberg, K.; Andreasson, L.-E. Biochemistry 1996,35, 14259.
148. Lindberg, K.; Vanngard, T.; Andrbasson, L. E. Photosynth. Res. 1993,38, 401.
149. Barry, B. A.; Babcock, G. T. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 7099.
150. Gerken, S.; Brettel, K.; Schlodder, E.; Witt, H. T. FEBS Lett. 1988,237, 69.
151. Joliot, P.; Barbieri, G.; Chabaud, R. Photochem. Photobiol. 1969, 10, 309.
152. Bioenergetics in Photosynthesis; Joliot, P., Kok, B., and Govindjee, Eds.; Academic Press: New York, 1975.
153. Kok, B.; Forbush, B.; McGloid, M. Photochem. Photobiol. 1970, 11, 457.
154. Sauer, K.; Yachandra, V. K.; Britt, R. D.; Klein, M. P. In Manganese Redox Enzymes; Pecoraro, V. L., Ed.; VCH Publishers: New York, 1992.
155. Yachandra, V. K.; Guiles, R. D.; McDermott, A. E.; Cole, J . L.; Britt, R. D.; Dexheimer, S. L.; Sauer, K.; Klein, M. P. Biochemistry 1987,26, 5974.
156. Penner-Hahn, J. E.; Fronko, R. M.; Pecoraro, V. L.; Yocum, C. F.; Betts, S. D.;
Bowlby, N. R. J.Am. Chem. SOC.1990,112, 2549.
157. MacLachlan, D. J.; Hallahan, B. J.; Ruffle, S. V.; Nugent, J. H. A,; Evans,
M. C. W.; Strange, R. W.; Hasnain, S. S. Biochem. J. 1992,285, 569.
158. George, G. N.; Prince, R. C.; Cramer, S. P. Science 1989, 10, 789.
159. DeRose, V. J.; Mukerji, I.; Latimer, M. J.; Yachandra, V. K.; Sauer, K.; Klein,
M. P. J . Am. Chem. SOC.1994, 116, 5239.
160. Dismukes, G. C.; Siderer, Y. FEBS Lett. 1980, 121, 78.
431
161. Dismukes, G. C.; Siderer, Y. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 274.
162. Cole, J.; Yachandra, V. K.; Guiles, R. D.; McDermott, A. E.; Britt, R. D.; Dexheimer, S . L.; Sauer, K.; Klein, M. P. Biochim. Biophys. Acta 1987, 890, 395.
163. Hansson, 0.;Aasa, R.; Vaenngaard, T. Biophys. J. 1987,51, 825.
164. Casey, J. L.; Sauer, K. Biochim. Biophys. Acta 1984, 767, 21.
165. Zimmermann, J.-L.; Rutherford, A. W. Biochim. Biophys. Acta 1984, 767, 160.
166. Beck, W. F.; de Paula, J . C.; Brudvig, G. W. J. Am. Chem. SOC.1986, 108, 4018.
167. Kim, D. H.; Britt, R. D.; Klein, M. P.; Sauer, K. Biochemistry 1992, 31, 541.
168. Kim, D. H.; Britt, R. D.; Klein, M. P.; Sauer. K. J. Am. Chem. SOC.1990,112, 9389.
169. Haddy, A.; Dunham, W. R.; Sands, R. H.; Aasa, R. Biochim. Biophys. Acta 1992,
1099, 25.
170. Messinger, J.; Robblee, J. H.; Yu, W. 0.; Sauer, K.; Yachandra, V. K.; Klein,
M. P. J. Am. Chem. SOC.1997, 119, 11349.
171. Messinger, J.; Nugent, J . H. A.; Evans, M. C. W. Biochemistry 1997, 36, 11055.
172. h r l i n g , K. A.; Peterson, S.; Styring, S. Biochemistry 1997, 36, 13148.
173. Dexheimer, S. L.; Klein, M. P. J . Am. Chem. SOC.
1992, 114, 2821.
274. Campbell, K. A,; Peloquin, J . M.; Pham, D. P.; Debus, R. J.; Britt, R. D. J. Am.
Chem. SOC.1998, 120, 447.
175. Yachandra, V. K.; Guiles, R. D.; McDermott, A. E.; Cole, J . L.; DeRose, V. J.;
Zimmermann, J. L.; Sauer, K.; Klein, M. P. Physica B (Amsterdam) 1989,158, 78.
176. Boussac, A,; Zimmermann, J.-L.; Rutherford, A. W.; Lavergne, J . Nature 1990,
347, 303.
177. Gilchrist, M. L.; Ball, J. A,; Randall, D. W.; Britt, R. D. Proc. Natl. Acad. Sci.
U.S.A. 1995, 92, 9545.
178. Tang, X.-S.; Randall, D. W.; Force, D. A,; Diner, B. A,; Britt, R. D. J . Am. Chem.
SOC.1996, 118, 7638.
179. Babcock, G. T.; Barry, B. A.; Debus, R. J.; Hoganson, C. W.; Atamian, M.; McIntosh, L.; Sithole, I.; Yocum, C. F. Biochemistry 1989,28, 9557.
180. Force, D. A,; Randall, D. W.; Britt, R. D. Biochemistry 1997,36, 12062.
181. Tommos, C.; Tang, X.-S.; Warncke, K.; Hoganson, C. W.; Styring, S.; McCracken,
1995, 117, 10325.
J.; Diner, B. A,; Babcock, G. T. J. Am. Chem. SOC.
182. Tommos, C.; Babcock, G. T. Acc. Chem. Res. 1998,31, 18.
183. Fontecave, M.; Pierre, J. C. Bull. SOC.Chim. Fr. 1996, 133, 653.
184. Hoganson, C. W.; Lydakis-Simantiris, N.; Tang, X.-S.; Tommos, C.; Warncke, K.;
Babcock, G . T.; Diner, B. A.; McCracken, J.; Styring, S. Photosyn. Res. 1995, 46,
177.
185. Babcock, G. T. In Photosynthesis: From Light to Biosphere; Mathis, P., Ed.;
Kluwer Academic Publishers: Netherlands, 1995,II, 209.
186. Ahlbrink, R.; Haumann, M.; Cherepanov, D.; Bogershausen, 0.;Mulkidjanian, A.;
Junge, W. Biochemistry 1998,37, 1131.
187. Tang, X.-S.; Diner, B. A.; Larsen, B. S.; Gilchrist, J.; M. L.; Lorigan, G. A,; Britt,
R. D. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 704.
188. Wincencjusz, H.; Van Gorkom, H.; Yocum, C. F. Biochemistry 1997,36, 3663.
189. Fine, P. L.; Frascb, W. D. Biochemistry 1992,31, 12204.
190. Sandusky, P. 0.;Yocum, C. F. FEBS Lett. 1983,162, 339.
191. Latimer, M. J.; DeRose, V. J.; Mukerji, I.; Yachandra, V. K.; Sauer, K.; Klein,
M. P. Biochemistry 1995,34, 10898.
192. Tso, J.; Sivaraja, M.; Dismukes, G. C. Biochemistry 1991, 30, 4734.
193. Rutherford, A. W. Trends Biochem. Sci. 1989, 14, 227.
194. Ghanotakis, D. F.; Babcock, G. T.; Yocum, C. F. FEBS Lett. 1984, 167, 127.
432
LAW, CAUDLE, AND PECORAFtO
195. DeRose, V. J.; Latimer, M. J.; Zimmerman, J. L.; Mukerji, I.; Yachandra, V. K.;
Sauer, K.; Klein, M. P. Chem. Phys. 1995,194, 443.
196. Sandusky, P. 0.;Yocum, C. F. Biochim. Biophys. Acta 1986,849, 85.
197. Radmer, R.; Ollinger, 0. FEBS Lett. 1986, 195, 285.
198. Messinger, J.; Badger, M.; Wydrzynski, T. Proc. Natl. Acad. Sci. U.S.A. 1995,
92, 3209.
199. Brudvig, G. W.; Crabtree, R. H. Proc. Natl. Acad. Sci. U.S.A. 1986,83, 4586.
200. Vincent, J. B.; Christou, G. Inorg. Chzm. Acta 1987, 136, L41.
201. Kirk, M. L.; Chan, M. K.; Armstrong, W. H.; Solomon, E. I. J . Am. Chem. SOC.
1992,114, 10432.
202. Larson, E.; Haddy, A.; Kirk, M. L.; Sands, R. H.; Hatfield, W. E.; Pecoraro, V. L.
J. Am. Chem. SOC.1992, 114, 6263.
203. Randall, D. W.; Gelasco, A,; Caudle, M. T.; Pecoraro, V. L.; Britt, R. D. J. Am.
Chem. SOC.1997,119, 4481.
204. Baldwin, M. J.; Pecoraro, V. L. J. Am. Chem. SOC.1996, 118, 11325.
205. Caudle, M. T.; Pecoraro, V. L. J. Am. Chem. SOC.1997, 119, 3415.
206. Gardner, K. A.; Mayer, J. M. Science 1995,269, 1849.
207. Blomberg, M. R. A.; Siegbahn, P. E. M.; Styring, S.; Babcock, G. T.; Akermark,
B.; Korall, P. J . Am. Chem. SOC.1997,119, 8285.
208. Kitajima, N.; Osawa, M.; Tamura, N.; Moro-oka, Y.; Hirano, T.; Masaaki, H.;
Nagano, T. Inorg. Chem. 1993,32, 1879.
209. Kitajima, N.; Tolman, W. B. In Progress in Inorganic Chemistry; Karlin, K. D.,
Ed.; John Wiley & Sons, Inc.: New York, 1995,43, 419.
210. Trofimenko, S. Chem. Reu. 1993,93, 943.
211. Lah, M. S.; Chun, H. Inorg. Chem. 1997,36, 1782.
212. Oki, A. R.; Bommarreddy, P. R.; Zhang, H.; Hosmane, N. Inorg. Chim. Acta 1996,
231, 109.
213. Adams, H.; Bailey, N. A.; Crane, J. D.; Fenton, D. E.; Latour, J.-M.; Williams,
J . M. J. Chem. Soc., Dalton Trans. 1990, 1727.
214. Di Vaira, M.; Mani, F. J. Chem. SOC.,Dalton Trans. 1990, 191.
215. Roesky, H. W.; Scholz, M.; Noltemeyer, M. Chem. Ber. 1990, 123, 2303.
216. Holleman, S. R.; Parker, 0. J.; Breneman, G. L. Acta. Cryst. 1994, C50, 867.
217. Andruh, M.; Huebner, K.; Noltemeyer, M.; Roseky, H.; W., 2. Naturforsch. 1993,
48B, 591.
218. Stolz, P.; Saak, W.; Strasdeit, H.; Pohl, S. Z. Natursforsch. 1989, 44B, 632.
219. Chen, X.; Long, G.; Willett, R. D.; Hawks, T.; Molnar, S.; Brewer, K. Acta Cryst.
1996,1252, 1924.
220. Glerup, J.; Goodson, P. A,; Hodgson, D. J.; Michelsen, K.; Nielsen, K. M.; Weihe,
H. Inorg. Chem. 1992,31, 4611.
221. Riley, D. P.; Weiss, R. H. J. Am. Chem. SOC.1994, 116, 387.
222. Riley, D. P.; Henke, S. L.; Lemon, P. J.; Weiss, R. H.; Neumann, W. L.; Rivers,
W.; Aston, J. K. W.; Sample, K. R.; Rahman, H.; Ling, C. S.; Shieh, J. J.; Busch,
D. H.; Szulbinski, W. Inorg. Chem. 1996, 35, 5213.
223. Burns, J. H.; Lumetta, G. J . Acta Cryst. 1991, C47, 2069.
224. Deng, Y.; Burns, J. H.; Moyer, B. A. Inorg. Chem. 1995,34, 209.
225. Othman, A. H.; Lee, K.-L.; Fun, H.-K.; Yip, B.-C. Acta Cryst. 1996, C52, 602.
226. Chan, C.-W.; Che, C.-M.; Peng, S.-M. Polyhedron 1993, 12, 2169.
227. Brooker, S.; McKee, V. Acta Cryst. 1993, C49, 441.
228. Bencini, A.; Biannchi, A.; Dapporto, P.; Garcia-Espafia, E.; Marcelino, V.; Micheloni, M.; Paoletti, P.; Paola, P. Inorg. Chem. 1990,29, 1716.
433
229. Deroche, A.; Morgenstern-Badarau, I.; Cesario, M.; Guilhem, J.; Keita, B.; Nadjo,
L.; Houee-Levin, C. J . Am. Chem. SOC.1996, 118, 4567.
230. Di Vaira, M.; Mani, F.; Stoppioni, P. J. Chem. SOC.,Dalton Trans. 1992, 1127.
231. Hagen, K. S. Angew. Chem., Int. Ed. Engl. 1992,31, 764.
232. Gultneh, Y.; Farooq, A.; Karlin, K.; Liu, D. S.; Zubieta, J. Inorg. Chim. Acta 1993,
211, 171.
233. Knof, U.; Weyhermueller, T.; Wolter, T.; Wieghardt, K. J . Chem. SOC.,Chem. Commun. 1993, 726.
234. Bermejo, M. R.; Castineiras, A,; Garcia-Monteagudo, J . C.; Rey, M.; Sousa, A,;
235.
236.
237.
238.
239.
240.
241.
242.
243.
244.
245.
246.
247.
248.
249.
250.
251.
252.
253.
254.
255.
256.
257.
258.
Watkinson, M.; McAuliffe, C. A.; Pritchard, R. G.; Beddoes, R. L. J . Chem. SOC.,

Dalton Trans. 1996,2935.
Jacobsen, E. N. I n Catalytic Asymmetric Synthesis; Ojima, I., Ed.; VCH Publishers, Inc.: New York, 1993, 159.
Pospisil; Jacobsen, E. N. Chem. Eur. J . 1996,2, 974.
Pecoraro, V. L.; Butler, W. M. Acta Cryst. 1986, C42, 1151.
Schake, A. R.; Schmitt, E. A,; Conti, A. J.; Streib, W. E.; Huffman, J. C.; Hendrickson, D. N.; Christou, G. Inorg. Chem. 1991, 30, 3192.
Shirin, Z.; Young, Jr., V. G.; Borovik, A. S. J . Chem. Soc., Chem. Commun.
1997, 1967.
Aurangzeb, N.; Hulme, C. E.; McAuliffe, C. A,; Pritchard, R. G.; Watkinson, M.;
Bermejo, M. R.; Sousa, A. J. Chem. SOC.,Chem. Commun. 1994, 2193.
Neves, A.; Vencato, I.; Erthal, S. M. D. Inorg. Chzm. Acta 1997,262, 77.
Eichhorn, D. M.; Armstrong, W. H. J. Chem. SOC.,Chem. Commun. 1992, 85.
Kitajima, N.; Komatsuzaki, H.; Hikichi, S.; Osawa, M.; Moro-oka, Y. J. Am. Chem.
SOC.
1994, 116, 11596.
Larson, E. J.;Pecoraro, V. L. J. Am. Chern. SOC.1991, 113, 3810.
Law, N. A,; Machonkin, T. E.; McGorman, J . P.; Larson, E. J.; Kampf, J. W.;
Pecoraro, V. L. J. Chem. SOC.,Chem. Comrnun. 1995, 2015.
Saadeh, S. M.; Lah, M. S.; Pecoraro, V. L. Inorg. Chem. 1991.30, 8 .
Chandra, S. K.; Chakravorty, A. Inorg. Chem. 1992,31, 760.
Chaudhuri, P.; Wieghardt, K. In Progress in Inorganic Chememistry; Lippard,
S. J., Ed.; Wiley Interscience, 1987,35, 329.
Quee-Smith, V. C.; DelPizzo, L.; Jureller, S. H.; Kerschner, J. L. Inorg. Chern.
1996,35, 6461.
Schlager, 0.; Wieghardt, K.; Nuber, B. Inorg. Chern. 1996,34, 6456.
Du Bois, J.; Hong, J.; Carreira, E. M.; Day, M. W. J. Am. Chem. SOC.1996, 118,
915.
Du Bois, J.; Tomooka, C. S.; Hong, J.: Carreira, E. M. Acc. Chem. Res. 1997,
30, 364.
Du Bois, J.; Tomooka, C. S.; Hong, J.; Carreira, E. M.; Day, M. W. Angew. Chem.,
Int. Ed. Engl. 1997,36, 1645.
Buchler, J.; Dreher, W. C.; Lay, K.-L.; Lee, Y. J. A.; Scheidt, W. R. Inorg. Chem.
1983, 22, 888.
Collins, T. J.;Gordon-Wylie, S. W. J . Am. Chem. SOC.
1989, 111, 4511.
Collins, T. J.; Powell, R. D.; Slebodnick, C . ; Uffelman, E. S. J . Am. Chem. SOC.
1990, 112, 899.
Fackler, N. L. P.; Zhang, S.; OHalloran, T. V. J. Am. Chenz. SOC.1996, 118, 481.
Niemann, A,; Bossek, U.; Haselhorst, G.; Wieghardt, K.; Nuber, B. Inorg. Chem.
1996, 35, 906.
434
259. Tetard, D.; Rabion, A,; Verlhac, J.-B.; Guilbem, J . J. Chem. SOC.,Chem. Commun.
1995,531.
260. Cheng, B.; Cukiernik, F.; Fries, P.; Marchon, J.-C.; Scheidt, W. R. Inorg. Chem.
1995,34, 4627.
261. Cheng, B.; Fries, P. H.; Marchon, J.-C.; Scheidt, W. R. Inorg. Chem. 1996,35, 1024.
262. Kipke, C. A,; Scott, M. J.; Ghodes, J. W.; Armstrong, W. H. Inorg. Chem. 1990,
29, 2193.
263. Kitajima, N.; Osawa, M.; Tanaka, M.; Moro-oka, Y. J. Am. Chem. SOC. 1991,
113, 8952.
264. Schardt, B. C.; Hollander, F. J.; Hill, C. L. J. Am. Chem. SOC.1982, 104, 3964.
265. Schardt, B. C.; Hollander, F. J.; Hill, C. L. J. Chem. SOC.,Chem. Commun. 1981,
765.
266. Mikuriya, M.; Yamato, Y.; Tokii, T. Bull. Chern. SOC.Jpn. 1992, 65, 2624.
267. Mikuriya, M.; Yamato, Y.; Tokii, T. Inorg. Chim. Actu 1991, 181, 1.
268. Nishida, Y.; Oshino, N.; Tokii, T. 2. Nuturforsch. 1988,43B, 472.
269. Bonadies, J. A.; Kirk, M. L.; Lah, M. S.; Kessissoglou, D. P.; Hatfield, W. E.;
Pecoraro, V. L. Inorg. Chem. 1989,28, 2037.
270. Larson, E.; Haddy, A,; Kirk, M. L.; Sands, R. H.; Hatfield, W. E.; Pecoraro, V. L.
J. Am. Chem. SOC.
1992,114, 6263.
271. Kitajima, N.; Singh, U. P.; Amagai, H.; Osawa, M.; Moro-oka, Y. J. Am. Chem.
SOC.
1991, 113, 7757.
272. Glerup, J.; Goodson, P. A.; Hazell, A.; Hazell, R.; Hodgson, D. J.; McKenzie, C. J.;
Michelsen, K.; Rychlewska, U.; Toftlund, H. Inorg. Chem. 1994,33, 4105.
273. Gelasco, A.; Pecoraro, V. L. J. Am. Chem. SOC.1993, 115, 7928.
274. Zhang, Z. Y.; Brouca-Cabarreq, C.; Hemmert, C.; Dahan, F.; Tuchagues, J. P. J.
275. Jensen, A. F.; Su, Z.; Hansen, N. K.; Larsen, F. K. Inorg. Chem. 1995,34, 4244.
276. Manchanda, R.; Brudvig, G. W.; de Gala, S.; Crabtree, R. H. Inorg. Chem. 1994,
33, 5157.
277. Frapart, Y.-M.; Boussac, A,; Albach, R.; Anxolabehere-Mallart, E.; Deiroisse, M.;
Verlhac, J.-B.; Blondin, G.; Girerd, J.-J.; Guilhem, J.; Cesario, M.; Rutherford,
A. W.; Lexa, D. J. Am. Chem. SOC.
1996,118, 2669.
278. Brewer, K. J.; Calvin, M.; Lumpkin, R. S.; Otvos, J. W.; Spreer, L. 0. Inorg. Chem.
1989,28, 4446.
279. Gelasco, A.; Kirk, M. L.; Kampf, J. W.; Pecoraro, V. L. Inorg. Chem. 1997,36, 1829.
280. Larson, E.; Lah, M. S.; Li, X.; Bonadies, J . A.; Pecoraro, V. L. Inorg. Chem. 1992,
31, 373.
281. Gohdes, J. W.; Armstrong, W. H. Inorg. Chem. 1992,31, 368.
282. Larson, E. J.; Riggs, P. J.; Penner-Hahn, J. E.; Pecoraro, V. L. J. Chem. SOC.,
283. Baldwin, M. J.; Stemmler, T. L.; Riggs-Gelasco, P. J.; Kirk, M. L.; Penner-Hahn,
J. E.; Pecoraro, V. L. J. Am. Chem. SOC.1994, 116, 11349.
284. Torayama, H.; Nishide, T.; Asada, H.; Fujiwara, M.; Matsushita, T. Polyhedron
1998, 17, 105.
285. Torayama, H.; Nishide, T.; Asada, H.; Fujiwara, M.; Matsushita, T. Chem. Lett.
1996,387.
286. Wieghardt, K.; Bossek, U.; Nuber, B.; Weiss, J.; Bonvoisin, J.; Corbella, M.; Vitols,
S. E.; Girerd, J. J. J. Am. Chem. SOC.1988, 110, 7398.
287. Bossek, U.; Weyhermiiller, T.; Wieghardt, K.; Nuber, B.; Weiss, J. J. Am. Chem.
SOC.1990, 112, 6387.
435
288. Arulsamy, N.; Glerup, J.; Hazell, A.; Hodgson, D. J.; McKenzie, C. J.; Toftlund,
H. Znorg. Chem. 1994,33, 3023.
289. Oberhausen, K. J.; 0. B. R. J.; Richardson, J. F.; Buchanan, R. M.; Costa, R.;
Latour, J.-M.; Tsai, H.-L.; Hendrickson, D. N. Znorg. Chem. 1993,32, 4561.
290. Bertoncello, K.; Fallon, G. D.; Murray, K. S. Polyhedron 1990, 9, 2867.
291. Pessiki, P. J.; Khangulov, S. V.; Ho, D. M.; Dismukes, G. C. J. Am. Chem. SOC.
1994,116, 891.
292. Pal, S.; Armstrong, W. H. Znorg. Chem. 1992, 31, 5417.
293. Wieghardt, K.; Bossek, U.; Zsoinai, L.; Huttner, G.; Blondin, G.; Girerd, J.-J.;
Babonneau, F. J. Chem. SOC.,Chem. Comm.un. 1987, 651.
294. Bossek, U.; Saher, M.; Weyhermuller, T.; Wieghardt, K. J . Chem. SOC.Chem. Commun. 1992, 1780.
295. Reddy, K. R.; Rajasekharan, M. V.; Padhye, S.; Dahan, F.; Tuchagues, J.-P. Znorg.
Chem. 1994,33, 428.
296. Bashkin, J. S.; Schake, A. R.; Vincent, J. B.; Chang, H.-R.; Li, Q.; Huffman, J . C.;
Christou, G.; Hendrickson, D. N. J. Chem. SOC.,
297. Caneschi, A,; Ferraro, F.; Gatteschi, D.; Melandria, M. C.; Rey, P.; Sessoli, R.
Angew. Chem., Znt. Ed. Engl. 1989,28, 1365.
298. Bossek, U.; Hummel, H.; Weyhermuller, T.; Wieghardt, K.; Russell, S.; van der
Wolf, L.; Kolb, U. Angew. Chem., Znt. Ed. Engl. 1996, 35, 1552.
299. Coucouvanis, D.; Reynolds, R. A,; Dunham, W. R. J. Am.. Chem. SOC.1995, 117,
7570.
300. Nepveu, F.; Gaultier, N.; Korber, N.; Jaud, J.; Castan, P. J. Chem. SOC.,
Dalton
Truns. 1995,4005.
301. Turner, P.; Gunter, M. J.; Hambley, T. W.; White, A. H.; Skelton, B. W. Znorg.
Chem. 1992,31, 2295.
302. Chen, X.-M.; Tong, Y.-X.; Xu, Z.-T.; Mak, C. W. T. J. Chem. SOC.,Dalton Trans.
1995, 4001.
303. Menage, S.; Vitols, S. E.; Bergerat, P.; Codjovi, E.; Girerd, J.-J.; Guillot, M.; Solans, x.; Calvet, T. Znorg. Chern. 1991, 30, 2666.
304. Hodgson, D. J.; Schwartz, B. J.; Sorrell, T. N. Znorg. Chem. 1989,28, 2226.
305. Chang, H.-R.; Diril, H.; Nilges, M. J.; Zhang, X.; Potenza, J . A.; Schugar, H. J.;
Hendrickson, D. N.; Isied, S. S. J. Am. Chem. SOC.1988, 110, 625.
306. Diril, H.; Chang, H.-R.; Zhang, X.; Larsen, S. J. Am. Chem. SOC.1987, 109, 6207.
307. Wesolek, M.; Meyer, D.; Osborn, J. A,; De Cian, A.; Fischer, J.; Derory, A.; Legoll,
P.; Drillon, M. Angew. Chem., Int. Ed. Engl. 1994, 33, 1592.
308. Yu, S. B.; Wang, C. P.; Day, E. P.; Holm, R. H. Znorg. Chem. 1991, 30, 4067.
309. Ishimura, Y.; Inoue, K.; Koga, N.; Iwamura, H. Chem. Left. 1994, 1693.
310. Cortes, R.; Lezama, L.; Pizzaro, J. L.; Arriotura, M. I.; Solans, X.; Rojo, T. Angew.
Chem., Znt. Ed. Engl. 1994, 33, 2488.
311. Cortes, R.; Pizzaro, J. L.; Lezama, L.; Arriotura, M. I.; Rojo, T. Inorg. Chem. 1994,
33, 2697.
312. Caudle, M. T.; Kampf, J. W.; Kirk, M. L.; Rasmussen, P. G.; Pecoraro, V. L. J .
Am. Chem. SOC.
1997,119, 9297.
313. Auger, N.; Girerd, J.-J. J. Am. Chem. SOC.1990, 112, 448.
314. Pal, S.; Chan, M. K.; Armstrong, W. H. J . Am. Chem. SOC.
1992, 114, 6398.
315. Vincent, J. B.; Chang, H.-R.; Folting, K.; Huffman, J. C.; Christou, G.; Hendrickson, D. N. J . Am. Chem. SOC.1987,109, 5703.
316. Bakie, A. R. E.; Hursthouse, M. B.; New, L.; Thomton, P.; Whit, P. G. J . Chem.
SOC.,Chem. Comnzun. 1980, 684.
436
31 7. Bhula, R.; Gainsford, G. J.;Weatherburn, D. C. J. Am. Chem. SOC.1988,110, 7550.

318. Reynolds, R. A.; Yu, W. 0.;
Dunham, W. R.; Coucouvanis, D. Inorg. Chem. 1996,
35, 2721.
319. Kessissoglou, D. P.; Kirk, M. L.; Bender, C. A,; Lah, M. S.; Pecoraro, V. L. J.
Chem. SOC.,Chem. Commun. 1989, 84.
320. Kitajima, N.; Osawa, M.; Imai, S.; Fujisawa, R; Moro-oka, Y. Znorg. Chem. 1994,
33, 4613.
321. Rardin, R. L.; Bino, A.; Poganiuch, P.; Tolman, W. B.; Liu, S.; Lippard, S. J. Angew. Chem., Znt. Ed. Engl. 1990,29, 812.
322. Baldwin, M. J.; Kampf, J. W.; Kirk, M. L.; Pecoraro, V. L. Inorg. Chem. 1995,
34, 5252.
323. Wang, S.; Tsai, H.-L.; Hagen, K. S.; Hendrickson, D. N.; Christou, G. J. Am. Chem.
SOC.1994, 116, 8376.
324. Hendrickson, D. N.; Christou, G.; Schmitt, E. A.; Libby, E.; Bashkin, J. S.; Wang,
S.; Tsai, H. L.; Vincent, J. B.; Boyd, P. D. W.; Huffman, J. C.; Folting, K.; Li,
&. Y.; Streib, W. E. J. Am. Chem. SOC.1992, 114, 2455.
325. Aubin, S. M. J.; Wemple, M. W.; Adams, D. M.; Tsai, H.-L.; Christou, G.; Hendrickson, D. N. J . Am. Chem. SOC.1996, 118, 7746.
326. Mikuriya, M.; Hashimoto, Y.; Kawamori, A. Chem. Lett. 1995, 1095.
327. Wemple, M. W.; Tsai, H.-L.; Folting, K.; Hendrickson, D. N.; Christou, G. Inorg.
Chem. 1993,32, 2025.
328. Wang, S.; Tsai, H.-L.; Libby, E.; Folting, K.; Streib, W. E.; Hendrickson, D. N.;
Christou, G. Znorg. Chem. 1996,35, 7578.
329. Brooker, S.; McKee, V.; Shepard, W. B. J. Chem. SOC.,Dalton Trans. 1987, 2555.
330. Wieghardt, K.; Bossek, U.; Gebert, W. Angew. Chem., Znt. Ed. Engl. 1983,22, 328.
331. Bashkin, J. S.; Chang, H.-R.; Streib, W. E.; Huffman, J . C.; Hendrickson, D. N.;
Christou, G. J. Am. Chem. SOC.1987, 109, 6502.
332. Bouwman, E.; Bocar, M. A.; Libby, E.; Huffman, J. C.; Folting, K.; Christou, G.
Inorg. Chem. 1992,31, 5185.
333. Gelasco, A,; Askenas, A.; Pecoraro, V. L. Inorg. Chem. 1996,35, 1419.
334. Chan, M. K.; Armstrong, W. H. J. Am. Chem. SOC.1991,113, 5055.
335. Kawasaki, H.; Kusumoki, M.; Hayashi, Y.; Suzuki, M.; Munezawa, K.; Suenaga,
M.; Senda, H.; Uehara, A. Bull. Chem. SOC.Jpn. 1994, 67, 1310.
336. Philouze, C.; Blondin, G.; Girerd, J.-J.; Guilhem, J.; Pascard, C.; Lexa, D. J. Am.
Chem. SOC.1994,116, 8557.
337. Jeffery, J. C.; Thornton, P.; Ward, M. D. Znorg. Chem. 1994,33, 3612.
338. Chan, M. K.; Armstrong, W. H. J. Am. Chem. SOC.1990,112, 4985.
339. Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry (John Wiley & Sons,
New York, 1988).
340. Weil, J. A.; Bolton, J . R.; Wertz, J. E. Electron Paramagnetic Resonance: Elementary Theory and Practical Applications (John Wiley & Sons, New York, 1994).
341. Bencini, A.; Gatteschi, D. Electron Paramagnetic Resonance of Exchange Coupled
Systems (Springer-Verlag, New York, 1989).
342. de Paula, J. C.; Beck, W. F.; Brudvig, G. W. J. Am. Chem. SOC.1986,108, 4002.
343. de Paula, J. C.; Brudvig, G. W. J . Am. Chem. SOC.1986,107, 2643.
344. h r l i n g , K.; Pace, R. J . Biophys. J. 1995, 68, 2081.
345. Bonvoisin, J.; Blondin, G.; Girerd, J. J.; Zimmermann, J. L. Biophys. J. 1992,
61, 1076.
346. Zheng, M.; Dismukes, G. C. Znorg. Chem. 1996,35, 3307.
437
347. Tang, X. S.; Gilchrist, M. L.; Lorigan, G. A,; Larson, B.; Britt, R. D.; Diner, B. A.
J. Am. Chem. SOC.1993, 115, 2382.
348. Britt, R. D.; Zimmermann, J.-L.; Sauer, K.; Klein, M. P. J. Am. Chem. SOC.1989,
111, 3522.
349. Randall, D. W.; Sturgeon, B. E.; Ball, J . A.; Lorigan, G. A,; Chan, M. K.; Klein,
M. P.; Armstrong, W. H.; Britt, R. D. J. Am. Chem. SOC.1996, 117, 11780.
350. Khangulov, S. V.; Voyevodskaya, N. V.; Varynin, V. V.; Grebenko, A. I.; MelikAdamyan, V. R. Biofizika 1987,32, 1044.
351. Khangulov, S.; Sivaraja, M.; Barynin, V. V.; Dismukes, G. C. Biochemistry 1993,
32, 4912.
352. Waldo, G. S.; Fronko, R. M.; Penner-Hahn, J . E. Biochemistry 1991, 30, 10486.
353. Kirby, J. A,; Robertson, A. S.; Smith, J . P.; Thompson, A. C.; Cooper, S. R.; Klein,
M. P. J. Am. Chem. SOC.1981, 103, 5528.
354. Kirby, J . A.; Goodin, D. B.; Wydrzynski, T.; Robertson, A. S.; Klein, M. P. J . Am.
Chem. SOC.1981,103, 5537.
355. Guiles, R. D.; Zimmermann, J.-L.; McDermott, A. E.; Yachandra, V. K.; Cole,
J . L.; Dexheimer, S. L.; Britt, R. D.; Wieghardt, K.; Bossek, U.; Sauer, K.; Klein,
M. P. Biochemistry 1990,29, 471.
356. Berthomieu, C.; Boussac, A. Biochemistry 1996,34, 1541.
357. Guiles, R. D.; Yachandra, V. K.; McDermott, A. E.; Cole, J. L.; Dexheimer, S. L.;
Britt, R. D.; Sauer, K.; Klein, M. P. Biochemistry 1990,29, 486.
358. Klein, M. P.; Sauer, K.; Yachandra, V. K. Photosyn. Res. 1993,38, 265.
359. Yachandra, V. K.; DeRose, V. J.; Latimer, M. J.; Mukerji, I.; Sauer, K.; Klein,
M. P. Photochem. Photobiol. 1991,53, 98s.
360. Rompel, A,; Andrews, J. C.; Cinco, R. M.; Wemple, M. W.; Christou, G.; Law,
N. A,; Pecoraro, V. L.; Sauer, K.; Yachandra, V. K.; Klein, M. P. J . Am. Chem.
SOC.1997, 119, 4465.
361. Roelofs, T. A.; Liang, W.; Latimer, M. J.; Cinco, R. M.; Rompel, A.; Andrews,
J. C.; Sauer, K.; Yachandra, V. K.; Klein, M. P. Proc. Natl. Acad. Sci. U.S.A. 1996,
93, 3335.
362. Riggs-Gelasco, P. J.; Mei, R.; Yocum, C. F.; Penner-Hahn, J. E. J . Am. Chem. SOC.
1996,118, 2387.
363. Amdt, D. In Manganese Compounds as Oxidizing Agents in Organic Chemistry;
Lee, D. G., Ed.; Open Court Publishing Company: La Salle, IL, 1981.
364. Snider, B. B. Chem. Rev. 1996,96, 339.
365. Jacobsen, E. N. In Comprehensive Organometallic Chemistry 11: A Review of the
Literature 1982-1994; Hegedus, L. S., Ed.; Elsevier Science, Inc.: Tarrytown,
NY, 1995,12, 1097.
366. Katsuki, T. Coord. Chem. Reu. 1995, 140, 189.
367. Katsuki, T. J. Mol. Catal. A-Chem. 1996, 113, 87.
368. Groves, J. T.; Nemo, T. E.; Myers, R. S. J . Am. Chem. SOC.1979, 101, 1032.
369. Groves, J . T.; Haushalter, R. C.; Nakamura, M.; Nemo, T. E.; E. B. J., J. Am.
Chem. SOC.1981,103, 2884.
370. Cytochrome P450: Structure, Mechanism, and Biochemistry; Ortiz de Montellano, P. R., Ed.; Plenum Press: New York, 1995.
371. Jorgensen, K. A. Chem. Rev. 1989,89, 431.
372. Jorgensen, K. A.; Schiott, B. Chem. Rev. 1990, 90, 1483.
373. Meunier, B. Chem. Rev. 1992,92, 1411.
374. Meunier, B., Ed., J . Mol. Catal. A-Chem. 1996, 113.
438
375. Stern, M. K.; Groves, J . T. In Manganese Redox Enzymes; Pecoraro, V. L., Ed.;
VCH Publishers, Inc.: New York, 1992, 233.
376. Groves, J. T.; Stern, M. K. J. A m . Chem. SOC.1983, 105, 5791.
377. Groves, J . T.; Nemo, T. E. J. Am. Chem. SOC.1983,105, 5786.
378. Groves, J. T.; Stern, M. K. J. Am. Chem. SOC.1987, 109, 3812.
379. Srinivasan, K.; Michaud, P.; Kochi, J. K. J. Am. Chem. SOC.1986, 108, 2309.
380. Samsel, E. G.; Srinivasan, K.; Kochi, J . K. J. Am. Chem. SOC.1986, 107, 7606.
381. Srinivasan, K.; Kochi, J. K. Inorg. Chem. 1985,24, 4671.
382. Feichtinger, D.; Plattner, D. A. Angew. Chem., Int. Ed. Engl. 1997,36, 1718.
383. Zhang, W.; Loebach, J . L.; Wilson, S. R.; Jacobsen, E. N. J. Am. Chem. SOC.1990,
112, 2801.
384. Jacobsen, E. N.; Zhang, W.; Muci, A. R.; Ecker, J . R.; Deng, L. J. Am. Chem. SOC.
1991,113, 7063.
385. Jacobsen, E. N.; Zhang, W.; Guler, M. L. J. Am. Chem. SOC.1991, 113, 6703.
386. hie, R.; Ito, Y.; Katsuki, T. Synlett 1991, 3, 265.
387. Irie, R.; Noda. K.; Ito, Y.; Matsumoto, N.; Katsuki, T. Tetrahedron Lett. 1990,
31, 7345.
388. Irie, R.; Noda, K.; Ito, Y.; Katsuki, T. Tetrahedron Lett. 1991, 32, 1055.
389. Tokunaga, M.; Larrow, J. F.; Fumistoshi, K.; Jacobsen, E. N. Science 1997,277,
936.
390. Linker, T. Angew. Chem., Znt. Ed. Engl. 1997,36, 2060.
391. Finney, N. S.; Pospisil, P. J.; Chang, S.; Palucki, M.; Konsler, R. G.; Hansen,
K. B.; Jacobsen, E. N. Angew. Chem., Int. Ed. Engl. 1997,36, 1720.
392. Linde, C.; Arnold, M.; Norrby, P.-0.; Akermark, B. Angew. Chem., Int. Ed. Engl.
1997,36, 1723.
393. Fu, H.; Look, G. C.; Zhang, E. N.; Jacobsen, E. N.; Wong, C.-H. J . Org. Chem.
1991,56, 6497.
394. Sharpless, K. B.; Teranish:, A. Y.; Backvall, J.-E. J. Am. Chem. SOC.1977,99, 3120.
395. Norrby, P.-0.; Linde, C.; Akermark, J . J. Am. Chem. SOC.1995, 117, 11035.
396. Yamada, T.; Imagawa, K.; Mukaiyama, T. Chem. Lett. 1992, 2109.
397. Collman, J. P.; Zhang, X.; Lee, V. J.; Uffelman, E. S.; Braumann, J. Science 1993,
261, 1404.
398. Lai, T.-S.; Kwong, H.-L.; Che, C.-M.; Peng, S.-M. J. Chem. SOC.,Chem. Commun.
1997,2373.
399. Halterman, R. L.; Jan, S.-T. J . Org. Chem. 1991, 56, 5253.
400. Groves, J . T.; Han, Y.-Z. In Cytochrome P450: Structure, Mechanism, and Biochemistry; Ortiz de Montellano, P. R., Ed.; Plenum Press: New York, 1996, 3.
401. Beckman, J. S. Chem. Res. Toxicol. 1996, 9, 836.
402. Beal, M. F. Neuroscientist 1997,3, 21.
403. Hunt, J . A.; Lee, J. B.; Groves, J. T. Chem. Biol. 1997, 4, 845.
404. Marla, S. S.; Lee, J.; Groves, J . T. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 14423.
405. Groves, J . T.; Takahashi, T.; Butler, W. M. Inorg. Chem. 1983,22, 884.
406. Chang, C. J.; Low, D. W.; Gray, H. B. Znorg. Chem. 1997,36, 270.
407. Bottomley, L. A.; Neeley, F. L. Znorg. Chem. 1997,36, 5435.
408. Dehnicke, K.; Strahle, J. Angew. Chem., Int. Ed. Engl. 1992,31, 955.
409. Groves, J . T.; Takahashi, T. J . Am. Chem. SOC.1983, 105, 2073.
410. Bellesia, F.; Ghelfi, F.; Pagnoni, U. M.; Pinetti, A. J. Chem. Res. ( S ) 1989, 360.
411. Bellesia, F.; Ghelfi, F.; Pagnoni, U. M.; Pinetti, A. J. Chem. Res. ( S )1989, 108.
412. Bellesia, F.; Ghelfi, F.; Pagnoni, U. M.; Pinetti, A. Synth. Commun. 1991,21, 489.
413. Bellesia, F.; Ghelfi, F.; Pagnoni, U. M.; Pinetti, A. J. Chem. Res. ( S )1991, 284.
439
414. Bellesia, F.; Ghelfi, F.; Pagnoni, U. M.; Pinetti, A. Gazz. Chim. Italiana 1993,
123, 289.
415. Bellesia, F.; Ghelfi, F.; Pagnoni, U. M.; Pinetti, A. J . Chem. Res. ( S )1990, 188.
416. Kurosawa, K.; Yamaguchi, K. Bull. Chem. Soc. Jpn. 1981,54, 1757.
417. Donnelly, K. D.; Fristad, W. E.; Gellerman, B. J.; Peterson, J. R.; Selle, B. J.
Tetrahedron Lett. 1984,25, 607.
418. Marko, I. E.; Richardson, P. F. Tetrahedron Lett. 1991, 32, 1831.
419. Richardson, P. F.; Marko, I. E. Synlett, 1991, 733.
420. Uemara, S.; Okazaki, H.; Onoe, A.; Okano, M. Bull. Chem. Soc. Jpn. 1978, 51,
3568.
421. Tanner, D. D.; Gidley, G. C. J . Org. Chem. 1968, 33, 38.
422. Poutsma, M. L. J. Am. Chem. Soc. 1965,87, 4285.
423. Marko, I. E.; Richardson, P. R.; Bailey, M.; Maguire, A. R.; Coughlan, N. Tetrahedron Lett. 1997, 38, 2339.
424. Huynh, V.-B.; Crawford, R. L. FEMS Microbinl. Lett. 1985,28,119.
425. Saadeh, S. M. Ph.D. Thesis, University of Michigan, 1992.
426. Policar, C.; Artaud, I.; Mansuy, D. Znorg. Chem. 1996,35, 210.
427. Riley, D. P.; Lennon, P. J.; Neumann, W. L.; Weiss, R. H. J. Am. Chem. Soc. 1997,
119, 6522.
428. Weiss, R. H.; Flickinger, A. G.; Rivers, W. J.; Hardy, M. M.; Aston, K. W.; Ryan,
U. S.; Riley, D. P. J. Biol. Chem. 1993,268, 23049.
429. Beyer, W. F. J.; Fridovich, I. Arch. Biochem. Biophys. 1989, 271, 149.
430. Beyer, W. F. J.;Fridovich, I. Methods E n q m . 1990, 186, 242.
431. Darr, D.; Zarilla, K. A,; Fridovich, 1. Arch. Biochem. Biophys. 1987,258, 351.
432. Rush, J . D.; Maskos, Z.; Koppenol, W. H. Arch. Biochem. Biophys. 1991,289, 97.
433. Matsushita,T.; Shono, T. Bull. Chem. Soc. Jpn. 1981, 54, 3743.
434. Baudry, M.; Etienne, S.; Bruce, A,; Palucki, M.; Jacobsen, E.; Malfroy, B. Biochem.
Biophys. Res. Commun. 1993, 192, 964.
435. Doctorow, S. R.; Huffman, K.; Marcus, C . B.; Musleh, W.; Bruce, A.; Baudry, M.;
Malfroy, B. Academic Press: New York, 1997,38, 247.
436. Rajan, R.; Rajaram, R.; Nair, B. U.; Ramasami, T.; Mandal, S. K. J . Chem. Soc.,
437. Frasch, W. D.; Mei, R. Biochim.. Biophys. Acta 1987, 891, 8.
438. Mathur, P.; Crowder, M.; Dismukes, G. C. J . Am. Chem. Soc. 1987,109, 5227.
439. Pessiki, P. J.; Dismukes, G. C. J. Am. Chem. Soc. 1994, 116, 898.
440. Gelasco, A,; Bensiek, S.; Pecoraro, V. L. Znorg. Chem. 1998, 37, 3301.
441. Sakiyama, H.; Okawa, H.; Isobe, R. J. Chem. Soc., Chem. Commun. 1993,882.
442. Sakiyama, H.; Tamaki, H.; Kodera, M.; Matsumoto, N.; Okawa, H. J. Chem. Sac.,
443. Naruta, Y.; Maruyama, K. J. A m . Chem. Soc. 1991, 113, 3595.
444. Naruta, Y.; Sasayama, M. J . Chem. Soc., Chem. Commun. 1994, 2667.
445. Naruta, Y.; Sasayama, M.-A.; Ichihara, K. J . Mol. Cat. A 1997, 117, 115.
446. Larson, E. J.; Pecoraro, V. L. J . Am. Chem. SOC.1991, 113, 7809.
447. Stibrany, R. T.; Gorun, S. M. Angew. Chem., Int. Ed. Engl. 1990, 29, 1156.
448. Mathews, J. H.; Dewey, L. H. J. Phys. Chem. 1913, 17, 211.
449. Ashmawy, F. M.; McAuliffe, C . A,; Parish, R. V.; Tames, J. J . Chem. Soc., Chem.
Cornmun. 1984, 14.
450. Ashmawy, F. M.; McAuliffe, C. A. J . Chenz. Soc., Dalton Trans. 1985, 1391.
451. Watkinson, M.; Whiting, A,; McAuliffe, C . A. J . Chem. Soc., Chem. Commun.
1994, 2141.
440
452.
453.
454.
455.
456.
457.
458.
459.
460.
461.
462.
463.
464.
Fujiwara, M.; Matsushita, T.; Shono, T. Polyhedron 1985,4, 1895.

Matsushita, T.; Fujiwara, M.; Shono, T. Chem. Lett. 1981, 631.
Matsushita, T.; Hirata, Y.; Shono, T. Bull. Chem. SOC.Jpn. 1982,55, 108.
Naruta, Y.; Sasayama, M.; Sasaki, T. Angew. Chem., Znt. Ed. Engl. 1994,33, 1839.
Limburg, J.; Brudvig, G. W.; Crabtree, R. H. J . Am. Chem. SOC.1997,119, 2761.
Lind, J.; Shen, X.; Ericksen, T. E.; Merenyi, G. J. Am. Chem. SOC.1990,112, 479.
McMillen, D. F.; Golden, D. M. Ann. Rev. Phys. Chem. 1982,33, 493.
Thorp, H. H.; Sarneski, J. E.; Brudvig, G. W.; Crabtree, R. H. J. Am. Chem. SOC.
1989, I l l , 9249.
Cooper, S. R.; Calvin, M. J. Am. Chem. SOC.1977,99, 6623.
Shank, M.; Barynin, V.; Dismukes, G. C. Biochemistry 1994,33, 15433.
Vainshtein, B. K.; Melik-Adamyan, W. R.; Barynin, V. V.; Vagin, A. A. In Progress in Bioorganic Chemistry and Molecular Biology; Ovchinnikov, Y. A,, Ed.;
Elsevier Science Publishers: New York, 1984, 117.
Nagata, T.; Mizukami, J. J. Chem. SOC.,Dalton Trans. 1995, 2825.
Nishida, Y.; Akamatsu, T.; Kazuyoshi, T.; Sakamoto, M. Polyhedron 1994, 13,
2251.
465. Dave, B. C.; Czernuszewicz, R. S. Znorg. Chin. Acta. 1994,227, 33.
466. Higuchi, C.; Sakiyama, H.; Okawa, H.; Fenton, D. E. J. Chem. Soc., Dalton Trans.
1995,4015.
467. Riggs-Gelasco, P. J. PbD. Thesis, University of Michigan, 1995.
468. Defrance, S.; Meunier, B.: Seris, J.-L. New J. Chem. 1992, 16, 1015,
469. Tangoulis, V.; Malamatari, D. A.: Soulti, K.; Stergiou, V.; Raptopoulou, C. P.;
Terzis, A.; Kabanos, T. A.; Kessissoglou, D. P. Znorg. Chem. 1996,35, 4974.
470. Cheng, B.; Scheidt, W. R. Acta Cryst. 1996, C52, 361.
471. Stephenson, G. R. In Advanced Asymmetric Synthesis; Stephenson, G. R., Ed.;
Chapman and Hail: New York, 1996,367.
472. Vincent, J. B.; Tsai, H.-L.; Blackman, A. G.; Wang, S.; Boyd, P. D. W.; Folting, K.;
Huffman, J. C.; Lobovsky, E. B.; Hendrickson, D. N.; Christou, G. J. Am. Chem.
SOC.1993, 115, 12353.
473. Yu, S.-B.;Lippard, S. J.; Shewky, I.; Bino, A. Znorg. Chem. 1992,31, 3502.
474. Mikuriya, M.; Fujii, T.; Tokii, T.; Kawamori, A. Bull. Chem. SOC.Jpn. 1993, 66,
1675.
475. D i d , H.; Chang, H.-R.; Nilges, M. J.; Zhang, X.; Potenza, J. A.; Schugar, H. J.;
Isied, S. S.; Hendrickson, D. N. J. Am. Chem. SOC.1989, 111, 5102.
476. Kessissoglou, D. P.; Butler, W. M.; Pecoraro, V. L. Znorg. Chem. 1987,26, 495.
477. Riggs-Gelasco, P. J.; Rui, M.; Ghanotakis, D. F.; Yocum, C. F.; Penner-Hahn,
J. E. J . Am. Chem. SOC.1996,118, 2400.
478. Michaud-Soret, I.; Jacquamet, L.; Debaecker-Petit, N.; Le Pape, L.; Barynin,
V. V.; Latour, J.-M. Znorg. Chem. 1998,37, 3874.
ADVANCES IN INORGANIC CHEMISTRY. VOL. 46
CALCIUM-6INDING PROTEINS
BRYAN E. FINN and TORBJORN DRAKENBERG
Division of Physical Chemistry 2, Lund University, S-22100 Lund, Sweden
I. Introduction
11. Intracellular EF-Hand Calcium-Binding Proteins
A. Introduction
B. Calmodulin
C. Troponin C
D. S l O O proteins
111. Calcium-Mediated Membrane-Binding Proteins
A. Introduction
B. Recoverin
C. Annexins
D. C2 Domains
E. y-Carboxyglutamic Acid Sites
IV. Extracellular Calcium-Binding Proteins
A. Introduction
B. Conantokins
C. EGF-like Domains
D. Serine Proteases
E. a-Lactalbumin and Lysozyme
F. Cadherins
G. SPARC
V. Summary
References
I. Introduction
Calcium is of widespread and fundamental importance in biochemistry, for the calcium ion functions as a second messenger, that is, one
whose signals are propagated by proteins specifically evolved for this
purpose. It is regarded as one of the most important bioinorganic elements (1). Apart from its relevance to biological solid-state materials,
Ca2' has key roles in many physiological processes. The Ca2+ion has
a radius of 100 pm and can be compared to Mg2+(72 pm) with its
441
Copyright 0 1999 by Academic Press
All rights of repioduct~onin any farm reserved
0898.8638199 $25 00
442
FINN AND D-NBERG
strict coordination requirements, having coordination numbers from

6 to 8. In its biological coordination chemistry it has a preference for
0-donor ligands from aminoacid side-chain carboxylate or hydroxyl
groups, peptide carbonyl oxygens, or HaO. The hexaaqua ion is very
labile and has an HaO exchange off rate constant of 5 10 s- (2).
This could in a simplistic way be thought of as an upper limit for
the on rate of Ca2+to its binding site. Higher on rates are possible,
however, when outer-sphere complexes are favorable and especially
when the binding site is highly negatively charged as discussed in
(3,4).
A review on calcium-binding proteins is a quite difficult task. During work on this review we have found that there are now an immense number of such proteins, making some selection necessary.
Even for those proteins selected we have no claims of completeness in
our coverage. We have therefore divided the proteins (or more correctly protein domains), into those that are (a) intracellular, (b) calcium mediated membrane bound, and (c) extracellular. The intracelM a r calcium-binding proteins are dominated by the EF-hand family
(defined later). This can be clearly seen from the recent book Guidebook to the Calcium Binding Proteins (5),in which the EF-hand proteins occupy 180 of a total of 230 pages. Likewise, in the section on
intracellular proteins in this review we have chosen to concentrate on
the EF-hand family.
In the section on calcium-mediated membrane-binding proteins we
have chosen to be more varied and discuss proteins interacting with
membranes from the inside as well as from the outside. In this section
it might be more proper to talk about protein domains than proteins.
In general it is the location of the calcium-binding domain that has
been used to define whether it is intra- or extracellular or membrane
bound. For example, the cadherins are treated as extracellular in this
review even though they are truly membrane proteins, and the various domains from the blood coagulation factors are treated as membrane binding (Gla domains) as well as extracellular (EGF and serine
protease domains). We hope this will not be too confusing for those
used to other definitions; however, in this review focused on the calcium ion we found this division natural.
We are aware that several important calcium-binding proteins have
been neglected and that our selection of proteins is clearly biased by
our own particular interests. We have also included some odd proteins
such as conantokins and neglected more important ones such as phospholipase Aa.
CALCIUM-BINDING PROTEINS
443
(I. lntracellular EF-Hand Calcium-Binding Proteins
A. INTRODUCTION
The EF hand is by far the most common motif for intracellular calcium-binding proteins. Briefly, it is an approximately 30 amino acidlong peptide chain composed of a central calcium-binding loop flanked
by two alpha helixes. The name EF-hand was coined by Kretsinger
( 6 ) and refers to the resemblance of the E and F helices of parvalbumin to a hand with the thumb and index finger extended. The calcium ion is bound by oxygen ligands contributed mostly from carboxylic acid side chains from aspartic or glutamic acid. Often one ligand
is contributed by water, and in rare cases, as discussed later, the
backbone carbonyl oxygens can contribute ligands. The sequence of a
typical EF-hand is shown in Fig. 1 and a stick figure of its calciumbinding loop is shown in Fig. 2. The ligating residues are relatively
highly conserved and have been used t o identify almost a thousand
EF-hands in over 150 proteins (7). However, of those identified, approximately a third are believed to have lost their ability to bind
calcium tightly, that is to say, with a dissociation constant of 10 ' to
lo-* M as is typical for these proteins. As expected, those that have
Helix
E
F
L
E
D
L
T
W
L
Helix
K
K
I
S
FIG.1. A schematic drawing of the second EF-hand calcium-binding subdomain from

calbindin Dgk (residues 45-74). Ligands from side-chain carboxyls are indicated in solid
lines; the ligand from the backbone carbonyl is indicated by a dashed line.
444
FINN AND DRAKENBERG
FIG.2. A stick figure of the second EF-hand binding loop from calbindin Dgk (residues
54-65). The carboxylic side chains and backbone carbonyls that ligand the calcium ion
are indicated in black.
lost the ability to bind calcium are also those whose sequences vary
the most from the canonical EF-hand sequence.
A second feature that is conserved among the EF-hand proteins is
the arrangement of hydrophobic residues on the interior faces of the
two helices and the central loop. These residues are clustered as a
hydrophobic surface and help to explain another key feature of EFhand motifs. This feature is that except on rare occasions, EF-hands
occur in pairs with their hydrophobic surfaces joined together and the
calcium loops paired in a small P-sheet interaction. The two EF-hands
in the pair share a pseudo-axis of rotation centered between the four
helices. This four-helix bundle arrangement is very stable and is observed even in non-calcium-binding proteins such as hemerythrin (8)
and cytochrome BEG2
(9).For example, the stability of calbindin Dgk,a
small calcium-binding protein composed of a single EF-hand pair, in
its calcium-loaded state is so high that it cannot be denatured by urea
or heating to 100C (10).
445
Although EF-hands tend to occur in pairs, proteins can contain anywhere from 1 to 8 EF-hand motifs with the most common number
being 4. A survey of a large number of known EF-hand calciumbinding proteins has been presented by Nakayama, et al. (7). The
functions of EF-hand proteins can be divided into two categories, signaling and buffering, and the distinction between the two groups is
determined by their structural response to calcium binding. In the
case of signaling proteins, the binding of calcium induces a conformational change, which renders the protein in its active form. Once active, the protein is able to bind a target protein or peptide to carry
out its function. The classical example of this type is calmodulin. The
activation process is tightly regulated by the concentration of calcium
in the cell, which is
M in a resting cell but can increase by several
orders of magnitude as a result of calcium intake from outside the
cell or release from internal stores.
In the case of the buffering proteins, calcium binding has little or
no significant effect on the structure of the protein and the function
of the protein is most likely limited to its calcium-binding role. The
best studied example of the latter is calbindin D9k.
The majority of the EF-hand family of calcium-binding proteins occur within the cell. However, there are several recent cases in which
EF-hand motifs have been identified in proteins expressed outside of
the cell or on the cell surface. These are discussed in Sections I11 and
IV.In this section we will concentrate our discussions on intracellular
calcium-binding proteins and give some examples of the best-known
and most studied examples of such proteins. This is by no means an
exhaustive list of EF-hand proteins, only a consideration of some examples that have been extensively studied from a structural perspective. Excellent introductions to the entire EF-hand family by Kawasaki and Kretsingar ( 1 1 ) and Celio ( 5 ) are recommended to those
interested in other members of this fascinating family of proteins.
B. CALMODULIN
That the EF-hand proteins are often referred to as the calmodulin
superfamily of calcium-binding proteins is no coincidence. Though the
term EF-hand was coined for parvalbumin (61, calmodulin is the
most well characterized of the EF-hand proteins with respect to every
aspect from molecular structure to physiological role. It serves a central role as a calcium-sensitive second messenger and is found in all
eukaryotic organisms (12). Rather than having a single target as is
446
FINN AND DRAKENBERG
common for signaling proteins, calmodulin has over 100 targets and
thus regulates a wide range of cellular function including signal
transduction, DNA synthesis, secretion, motility and cell division ( I 1,
12).It is composed of a single chain of 148 amino acids with a molecular mass of approximately 18 kDa. It contains four EF-hand calciumbinding subdomains that are arranged pairwise in two domains separated by a central linker. All four sites are active and bind calcium
with high affinity, with the carboxy-terminal EF-hand pair having
slightly higher affinity (KD=
MI than the amino-terminal pair
(KD=
M). The first structure of calcium-loaded calmodulin from
rat was determined by X-ray crystallography in 1985 (13)and was
later refined to high resolution in 1988 (14).
These structures reveal a surprising architecture for the protein in
that the two EF-hand pair domains are separated by a long, apparently rigid helix (Fig. 3). This dumbbell shape of calmodulin was
surprising because biochemical data indicated that the target binding
site was evenly distributed between the two domains, but its target
peptides were obviously too small to span the distance between the
two domains. In addition, small-angle X-ray scattering and neutron
scattering experiments indicated that the solution form of calmodulin
was much more compact than that determined by X-ray crystallography (15,161. Further X-ray structures of calmodulin from human (171,
Drosophila (18) and Paramecium (19) spp. also demonstrated the
presence of a central helix.
It was not until a high-resolution structure of calmodulin in solution was obtained by nuclear magnetic resonance (NMR) spectroscopy
that the apparent paradox of calmodulins structure was resolved. Al-
FIG.3. A ribbon diagram of active (Cazt human calmodulin (1 7). The central flexible linker separating the two EF-hand pair domains is indicated in dark gray.
447
though there was good agreement between the structures determined

by X-ray crystallography and NMR spectroscopy, the central linker
was found to be nonhelical in solution (20). Furthermore, measurements of the dynamics of the backbone using NMR relaxation data
showed that this linker was more flexible than the surrounding helices (21).
Target binding by calmodulin is an interesting case of molecular
recognition. Although calmodulin binds a large number of target
proteins, it binds with high affinity, often in the nanomolar range.
Therefore, one might have expected that they share a common calmodulin binding sequence. However, the many calmodulin recognition sequences share little sequence homology and only resemble
one another in the general properties of being 14 to 26 amino acids
long, hydrophobic, and basic (22).
Both NMR spectroscopy and X-ray crystallography played key roles
in understanding the target binding of calmodulin. In this case, the
two methods show much more agreement. Three high-resolution
structures of complexes between calmodulin and a target protein have
been published, two by X-ray crystallography on complexes with peptides with the smooth muscle light-chain kinase (smMLCK) (23)and
calmodulin-dependent protein kinase I1 (CaMKII) (241, and one by
NMR spectroscopy on the complex with a peptide from skeletal muscle light-chain kinase (skMLCK) (251. This latter structure is shown
in Fig. 4.
The smMLCK and skMLCK complex structures resemble one another, as is expected because the peptides share a degree of sequence
homology. The main features of the complex are two hydrophobic anchors a t opposite ends of the peptide that interact with hydrophobic
residues in the calmodulin-binding pocket. In addition, electrostatic
interactions between the basic residues of the peptide with glutamic
acid residues in both domains of calmodulin are observed. In the case
of CaMKII, the hydrophobic anchors are present, but are closer to one
another in the peptide sequence. Calmodulin adapts its binding to
this sequence in two ways. First, because the two domains are connected by a flexible tether, they are able to rearrange with respect to
one another and the target peptide so as best to accommodate the
target peptides hydrophobic amino acid pattern. The second adaptation occurs within the two binding clefts themselves. These clefts
contain a larger than usual proportion of methionine residues that
have a long side chain, thus giving them more degrees of freedom to
adjust their conformations as necessary to accommodate the target
peptide.
448
FINN AND DRAKENBERG
FIG.4. A ribbon diagram of the binding of skMLCK peptide to calmodulin (25).The

peptide and calcium ions are indicated in light gray and calmodulin in dark gray.
Although the structures of calmodulin bound to target peptides

have yielded much information about the complexes and their plasticity, the mechanism by which binding of calmodulin to the target proteins activates them remains to be definitively demonstrated. This
will require the structure of a calmodulin-protein complex. However,
much evidence available now allows us to infer the mechanism of calmodulin-dependent activation. The strongest hypothesis is the pseudosubstrate or autoinhibition hypothesis, which was first suggested
by Kemp et al., who noted the similarity of the substrate and autoinhibitory region of smMLCK (26). Subsequently, it was demonstrated
that removal of the autoinhibitory region by proteolysis could activate
smMLCK in the absence of calmodulin (27).A recent structure of the
autoinhibited form of calmodulin-dependent protein kinase I (CaMKI)
clearly shows that its autoinhibitory region lies in the substrate binding site and that part of the calmodulin binding segment protrudes
out, thus potentially able to bind calmodulin (28).
Though the vast majority of target proteins bind to calciumactivated calmodulin, there are cases in which calmodulin has been
449
FIG.4. (continued)
reported to bind proteins in the absence of calcium. However, in most

cases they are believed to be affected by changes in calcium concentration. One example is that of neuromodulin, which binds calmodulin
tighter in the absence of calcium at low ionic strength. The physiological significance of this binding is unclear because it is much weaker
at physiological salt concentrations. Other evidence suggests that
neuromodulin may act to bind and sequester calmodulin a t the
plasma membrane in a calcium-dependent manner (29,301.
The mechanism of calcium-induced activation of calmodulin has
been and continues to be a debated topic. Since the first X-ray crystal
structure of calcium-loaded calmodulin (13), models have been proposed for the structure of the apo form of calmodulin (32) based on
450
FINN AND DRAKENBERG
the structure of related troponin C, which was first crystallized with

only two of four calcium-binding sites loaded with calcium (32, 33).
Briefly, the model predicted that the hydrophobic binding sites of calcium-calmodulin are closed and sequestered from solvent in the apo
form. The general features of this model were confirmed ten years
after the crystal structure of the calcium-loaded from by several solution structures of apo-calmodulin determined by NMR (34-36). The
conformational change observed for the carboxy-terminal domain is
shown in Fig. 5 . Additionally, this structure demonstrated the autonomy of the two domains with respect to their conformational response
to calcium binding being independent of the linker connecting the two
domains (37).
C. TROPONIN
C
Troponin C (TnC) is a calcium-binding protein that is part of the
troponin complex localized on the thin filament of muscle fibers. The
other two components of the complex are troponin I (TnI) and troponin T (TnT), which together with TnC are arranged in a heterotrimer.
Two isoforms are known, the fast skeletal muscle (sTnC) and the cardiac and slow skeletal muscle (cTnC) forms. The structure of TnC is
similar to that of calmodulin, with four E F hands organized pairwise
in two domains. Indeed, our understanding of these two proteins has
evolved in parallel, with a good deal of what has been learned about
troponin C being applicable to calmodulin and vice versa. Like cal-
FIG.5. A space-filling drawing of the conformational change in the TRzC domain of

calmodulin (37). The hydrophobic side chains lining the half of the target binding site
in the TRzC domain are indicated in white with the backbone and remainder of the
side chains in gray.
45 1
modulin, all four calcium-binding sites in sTnC are active, with the
carboxy-terminal sites being tighter than the N-terminal pair (KD
lo-* M vs. KD
M). However, in cTnC the first site is inactive.
Unlike calmodulin, which binds to a large array of target proteins,
TnC has one very specific function, namely the cyclic binding and release of TnI in a calcium-dependent manner, which in turn regulates
muscle contraction.
The structure of sTnC was first determined by X-ray crystallography and remarkably in a half-saturated (Ca2+)2
state (32, 33). These
initial structures formed the basis for modeling of the mechanism of
calcium-induced activity of troponin C (38) as well as the calciuminduced activation of calmodulin (31).Studies of the solution structures of sTnC added the structures of the (Ca2l4 calcium-saturated
form (39)and the apo and calcium-saturated forms of the N-terminal,
regulatory, domain (40,411to the already known half-saturated form.
Together, these paint a complete picture of the structural changes
induced by calcium binding on troponin C and have led to a model for
troponin Cs role in muscle contraction. The model predicts that in
the resting state, only the C-terminal domain of sTnC is calcium
loaded and bound to TnI a t its N-terminal domain. Actin is also bound
to TnI in an inhibitory complex. Upon binding of calcium to TnCs
N-terminal domain, a hydrophobic patch is opened and binds to the
inhibitory and C-terminal domains of TnI, thus releasing actin from
TnI. The actin is thus available for binding to myosin, which leads to
the muscle contraction. Support for another part of this model was
provided by studies of the interaction of sTnC and its binding target
domain on TnI by NMR (421,which confirmed that TnI interacts with
the regulatory domain of TnC. Further details of the mechanism of
action of troponin C have also been provided by solution structural
studies, including mutants of troponin C aimed a t dissecting the role
of individual residues in the process. The final bidentate glutamate
ligand in the calcium-binding loops has been a n especially attractive
target in troponin C (43, 44) as well as calmodulin (45, 46). These
studies have demonstrated these to be key residues in the calciuminduced activation process. Whether they are the only keys or part of
an interplay between a number of calcium-ligand and other interactions is a subject of continuing study.
D. S l O O PROTEINS
The S l O O proteins make up a distinct subfamily within the EFhand family. They share structural characteristics of being small
452
FINN AND DRAKENBERG
(9-12 kDa), being acidic, containing only 2 EF-hands, and having a

homology of between 25 and 40%. They also share one feature that is
unique among calcium-binding EF-hand proteins, that is, the insertion of two residues in the first EF-hand, which alters the conformation of the calcium-binding loop such that the ligands are contributed
from the peptide backbone carbonyl oxygens instead of the side-chain
carboxyl groups as in a standard EF-hand. This variation is known
as a pseudo-EF-hand. Beyond this, they differ widely with respect
to expression, activity, oligomer state, and conformational response to
calcium binding.
1. Calbindin Dgk
Calbindin Dgk,also known as ICBP and SlOOD, is expressed highly

in the intestine, placenta, and uterus, and has been suggested to be
involved in calcium transport, though a definitive demonstration of
its function remains to be accomplished. It has been shown to interact
with the calmodulin-binding domain of the plasma membrane calcium
pump (47),though it is unclear if this interaction plays a significant
functional role. It is a unique member of the ,9100 family in that it
is monomeric whereas most members of this subfamily are homo- or
heterodimeric. From a structural perspective, calbindin Dgk is the
most extensively studied of the SlOO subfamily. It is precisely its
small size as well as its stability that makes it an attractive target
for structural studies. The first X-ray structure of calcium-loaded
calbindin Dgk was determined in 1981 (48) and was later refined
to 2.3 A (49),and then to 1.6 (50). Solution structures have also
been determined for several forms of bovine calbindin Dgkincluding
(Ca2+12(51), apo (52, 531, (Cd2+I1(54) and (Cd2+I2.In addition, the
structure of the porcine protein has been determined (55) by NMR
methods.
These studies when taken together show another unique feature
of calbindin Dgk:the relatively small effect that calcium has on the
conformation of the protein. On going from the apo to calcium-loaded
state, only a minor adjustment of the secondary structure is observed,
and this is limited to the second EF hand. The structure of the singly
loaded (Cd2+Ilstate showed that most of these changes occur upon
binding of the first ion, which binds in the second, C-terminal binding site (54). This lack of a large structural response to calcium
binding, which is typical for the calcium sensors such as calmodulin
and troponin C, indicates that calbindin Dgk)s function is probably
not one of signaling but more likely one of regulating calcium homeostasis.
453
Although ion binding has little effect on the structure of calbindin,

it has been shown to have a large effect on the dynamics and stability
of the protein. Calbindin Dgkhas served as a convenient model system
for studies of protein dynamics by NMR and other techniques. Results
from studies of hydrogen-deuterium exchange (56, 57) have shown
that the structure of calbindin can be significantly stabilized by ion
binding. The effect is greatest in the binding loops, but it is also propagated through hydrogen-bonding networks in the helices. Similar
studies on the fast (nano- to picosecond) time-scale dynamics by spinrelaxation techniques (58, 59) have shown that ion binding exerts
an effect here as well but is localized to the C-terminal binding
loop, which is more dynamic in the apo than in the calcium-loaded
form.
Structural studies of calbindin Dgkhave also yielded several unique
observations of a more general nature. The first was the observation
that the purified protein appeared heterogeneous in isoelectricfocusing experiments. This heterogeneity was traced to an isoaspartyl
linkage formed by deamidation of an Asn-Gly pair (60).A second apparent heterogeneity in the protein preparation was identified as
well. However, instead of being a configurational heterogeneity it was
found to be a conformational heterogeneity traced to the cis-trans
isomerization of a proline in the linker between the two EF hands
(61).This result, later confirmed by crystallographic studies GO), was
one of the first direct demonstrations that such isomers of proline
occur at equilibrium.
Calbindin Dgkhas also proved to be a convenient model system for
studies of calcium binding and cooperativity (62).Because it has only
two calcium ions, it presents a simpler system than those proteins
that have four or more calcium binding sites.
2. sloop
Sloop occurs most commonly as SlOOB, a pp dimer, but can form

heterodimers (ap). It is one of the better characterized members of
the S l O O family. It was the first protein given the designation S100,
which was given to Sloop when it was first isolated and means, simply, that it is partially soluble in 100% saturated ammonium sulfate
(63). Thus, though the name S l O O does not indicate any functional
aspect, it has since been used by some to designate this entire family
of proteins due to their homology to Sloop and common chromosomal
localization (64).
The exact function of Sloop is not known. However, it has been
implicated in a number of cellular functions such as cell growth,
454
FINN AND DRAKENBERG
apoptosis, and energy metabolism. It is most abundant in the nervous

system, especially the glial cells, but it can also be found in other
tissues such as the skin, testis and cartilage (65). I t has also been
shown to exhibit extracellular functions related to neuronal and glial
cell growth and proliferation (66). Sloop contains two cysteine residues (68 and 84),and the SlOOB dimer is thus able to form a covalently linked dimer. Although the disulfide linkage is not necessary
for dimerization, it has been shown to be essential for its extracellular
function and some intracellular functions (67).
Calcium binding causes a conformational change in SlOOB, which
produces an increase in hydrophobic surface area, similar to calmodulin and troponin C. However, the details of this conformational
change are only now beginning to come to light. The high-resolution
structure of SlOOB was first determined in apo form for both the bovine (68) and rat (69) proteins. The rat SlOOB structure is shown in
Fig. 6. These structures showed that the protein is composed of a
dimer of two regular EF-hand pairs as expected. The dimer interface
is an X-type four-helix bundle made up of helices 1, 4, l', and 4'.
However, the structures differed significantly from one another in
several important respects. First, the position of helix 3 in the second
EF hand in each monomer differs enormously. Second, a small extra
helix in the linker between helices 2 and 3 is reported for the bovine
structure but not observed in the rat structure. Some suggest that
these differences may be due to modeling rather than actual difference in the structures of the proteins from different species (70),but
independent evidence will be required to resolve this question satisfactorily.
Several structures of the calcium-bound form of SlOOB determined
by X-ray crystallography of the bovine protein (71), and by NMR spectroscopy of the rat (72) and human (73) proteins, have been recently
determined. The structures confirm several features predicted earlier
(69, 70), namely that the major conformational changes are in the
linker between the two EF hands, helix 3, and the second calciumbinding loop. The dimer interface is essentially unchanged.
Evidence has also been recently obtained by NMR chemical shift
analysis on the interaction of calcium-activated SlOOB and target proteins CapZ (74) and p53 (75). Here again, there are differences in the
reported binding sites on SlOOB, which may be due to differences in
the binding peptides.
3. Calcyclin
Calcyclin, also called S100A6 (64) or CaBP (761, is a dimeric protein

similar to Sloop in many respects. It is found in various tissues in-
455
FIG.6 . A ribbon diagram of the structure of SlOOB a s determined by Drohat et al.

(69).The two members of the dimer are shaded differently to distinguish them. Helixes
3 and 3', those believed to be involved in conformational change and target binding,
are located a t the top. Helixes 4 and 4' lie a t the dimer interface perpendicular to the
plane of the figure; helixes 1 and 1' lie under these parallel to the plane of the figure.
cluding lung, heart, platelets, and muscle. It undergoes a conformational change upon calcium binding and exposes a hydrophobic patch,
which is most probably the target binding site. As was the case for
calmodulin and troponin C, the structural characterizations of calcyclin and SlOOB have been closely linked. A solution structure of
apo-calcyclin was determined by NMR spectroscopy (771, and the protein was found to be a symmetrical dimer whose interface was composed of helices 1, 4, l', and 4',as was later found for S100B. However, there are differences between calcyclin and the two reported
apo-S100B structures (68, 691, especially in the region of helix 3.
These differences may be genuine, but helix 3 is the least well determined segment of the apo-calcyclin structure, and a higher-resolution
structure will be needed to settle the question of how different they
really are. Characterization of the calcium-loaded form of calcyclin
456
FINN AND DRAKENBERG
has also been done by NMR methods (78). As opposed to SlOOB, only
subtle changes in the conformation of calcyclin are observed upon calcium binding, and these are similar to those observed in calbindin
Dgk.This result is surprising and in apparent conflict with calcyclins
role as a calcium sensor and data that indicate that a hydrophobic
surface is opened upon calcium binding (76, 79, 80). The functional
role and interactions with target proteins remain an area of intense
study. One target protein has been identified and is expressed predominantly in the brain, similar to calcyclin itself (81,82). This suggests that calcyclin may act as a calcium-dependent neuronal signaling protein.
111. Calcium-Mediated Membrane-Binding Proteins
A. INTRODUCTION
Calcium-mediated membrane binding of proteins occurs both inside
cells and outside. At first one might think that these events should
be very different in these two different environments, with calcium
concentrations in pM and mM, respectively. It is, however, not necessarily so.
Two extreme models for the calcium-mediated membrane binding
may be envisaged. The first one, exemplified by recoverin, is when
calcium binding to the protein causes a structural change that, somewhere else, exposes a hydrophobic patch that may bind to the membrane without any direct involvement of the calcium ion. In the case
of recoverin the hydrophobic patch is in fact a myristoyl chain
attached to the N terminus of the protein. The other extreme case
would be where the calcium ions form a link between the protein and
membrane surface. This has for some time been a working model for
the binding of Gla-rich domains from blood coagulation proteins to
phosphatidyl serine-containing membranes. More recently, however,
another, more specific, model for this interaction has emerged, as will
be discussed later. In our opinion the mode with the calcium ion as a
glue between two negatively charged surfaces will probably not be
found because there will be no specificity in such an interaction. However, there are certainly examples in which the calcium ion is coordinated to a protein as well as to a phospholipid head group. We will
now discuss a few cases of calcium-mediated protein binding to membranes.
457
B. RECOVERIN
Recoverin is a protein of the EF-hand type described in Section 11.
However, as opposed to most EF-hand proteins, it is a membraneassociated protein that binds the intracellular face of membranes in a
calcium-dependent manner. The protein is composed of four E F hands
arranged pairwise, with only two of the four sites able to bind calcium
(83). It is a member of the neuron-specific calcium sensor (NCS) family of proteins, which includes closely homologous proteins such as Smodulin and visinin, and less homologous proteins such as the vilip
proteins or frequenin. The membrane binding is mediated by a n acyl
group, usually myristoyl, covalently attached to the amino terminus.
In the apo form this myristoyl group is sequestered from solvent and
released upon the binding of two equivalents of calcium (84). The
structure of the unmyristoylated recoverin (85) was determined by
X-ray crystallography and was followed by structures of the myristoylated apo form (86) and myristoylated calcium form (87). Together
these structures paint the full pictures of the molecular transition
between the inactive apo form and the active calcium form (Fig. 7).
Briefly, in the apo state, the myristoyl group is buried in a deep hydrophobic pocket predominantly located in the amino-terminal EFhand pair but with some contacts to the first E F hand in the carboxyterminal domain. Upon calcium binding the hydrophobic pocket is
closed and the myristoyl group is forced out. This is accompanied by
a 45 degree rotation of the two EF-hand pair domains relative to one
another. This type of calcium-induced conformational change is
faintly reminiscent of that observed in calmodulin and troponin C but
of opposite direction, (i.e., calcium binding closes a hydrophobic binding site in recoverin whereas it opens one in calmodulin).
Functionally, recoverin acts by binding to rhodopsin kinase and inhibits rhodopsin phosphorylation (88).It has also been observed to
bind rod outer segment (ROS) membranes (89). The interaction with
rhodopsin kinase does not require amino-terminal acylation, but binding to ROS membranes is acylation dependent, suggesting two distinct binding sites a n d o r binding modes on recoverin.
C. ANNEXINS
The annexins form a group of mainly intracellular proteins with a
Ca2+-dependentbinding to phospholipids. No clear physiological role
of the annexins has so far been defined, but a wide range of biological
functions have been suggested (90-97). Amino acid sequence analysis
459
indicates that the Ca2+coordination in the annexin sites differs from

that of the EF hand, even though both structures are dominated by
an antiparallel four-helix bundle. The crystal structures have been
determined for annexin V from humans (98-1011, chickens (102),and
rats (103, 104) as well as for annexin I from humans (105). These
structures all contain four domains with multiple Ca-binding sites.
The loop between helixes A and B contains a double site (AB and
AB),and the loop between helixes C and D contains a single site
(CD). In the structure described by Seaton and co-workers (103, 104)
the four domains have very similar AB loops, RMSD = 0.39 A, and
the Ca2+coordination is identical for all domains. The AB site carries
a fully coordinated Ca2+whereas the Ca2 in the AB site has fewer
ligands and only half the occupancy of the AB site. It therefore appears clear that the AB site is the stronger one of these two sites. The
fact that the AB site is only partially occupied in the crystal is in
agreement with Ca2+-bindingstudies that have shown that the Cabinding is weak in the absence of phospholipid membranes compared
to what it is in the presence of membranes, mM and pM, respectively
(106-111). In the presence of C a Z 4annexin
,
V binds to phosphatidylserine (PS) and phosphatidylethanolamine (PE)-containing membranes, and shows a clear preference for PS (112). Annexin V,
however, binds very poorly t o phosphatidylcholine (PC) or
phosphatidylinositol (PI), so charge is not the only determinant for
this binding, because PE and PC are neutral whereas PS and PI are
negatively charged at physiological pH. In a recent time-resolved fluorescence study (113)it was shown that the environment of the sole
tryptophan in annexin V changes upon Ca2+binding in the presence
of phospholipid vesicles. The changes were interpreted to indicate
that the tryptophan side chain moved from a hydrophobic to a more
hydrophilic environment (i.e., from the hydrophobic core of the protein to the interface between the protein and the membrane). This
interpretation is in good agreement with crystal structures of annexin
V in the presence of glycerophosphoserine (GPS) or glycerophosphoethanolamine (GPE). These crystals were obtained from soaking
annexin crystals with either GPS or GPE solutions and were found to
bind only one GPS/GPE per annexin, namely to domain 3 (104). Due
to the high degree of similarity of the four domains and the observa-
FIG.7. A schematic diagram of the conformational change and extrusion of the myristoyl group from the N-terminal domain of recoverin upon calcium binding followed by
insertion into the membrane.
460
FINN AND DRAKENBERG
tion by high-resolution electron microscopy that the annexin molecules are flattened on the membrane surface (114, 115),it has been
assumed that in solution all four domains will bind to the surface in
a similar way (104). In both complexes a phosphoryl oxygen coordinates to the CaZt on the Al3 site (Fig. 8) replacing a water molecule.
This will place the tryptophan side chain clearly inside the polar head
group of the lipid and in level with the bond connecting the hydrocarbon chains with the glycerol moiety. These crystal structures will also
explain the preference of PS over PE for annexin binding. The serine
carboxylate coordinates the AEi CaZt and the serine amino group
forms hydrogen bonds to both the side chain of Thr187 and the main
chain of Glu226. None of these interactions are present in the PE
complex, and for the bulky PI and PC, which bind annexin poorly,
repulsive steric interaction may be important. The crystal structure
of annexin V/GPS clearly indicates that two Ca2+in each domain are
simultaneously coordinated to the protein and the lipid head group.
This is the only direct evidence so far for this kind of interaction that
has also been discussed for the membrane binding of C2 and Gla domains (discussed later). It is, however, also clear that there are many
interactions that contribute to the binding (e.g., all four domains have
a hydrophobic residue in the position analogous to the tryptophane in
domain 3).
D. C~DOMAINS
The Ca2+-bindingC2 domain, first identified in various isoforms of
mammalian Ca2+-dependentprotein kinase C (PKC) (116-118) is ca.
130 amino acids long. Because it was found that PKCs lacking the
second domain did not show any Ca2+regulation, it was proposed that
the C2 domain was responsible for the CaZt regulation of PKC. Similar domains have been identified in various proteins such as synaptotagmin (119-123), cytosolic phospholipase Az (cPLAz)(124-126), and
phosphoinositide-specific phospholipase C, IP-PLC (127-135), among
others. [For a recent review see (136).]
It has been shown that a single C2 domain from synaptotagmin will
bind to phospholipid membranes in a CaZt-dependent manner (119,
120). It has also been shown that this binding depends on the phospholipid composition (121). C2 domains from all the studied synaptotagmins (II-VI) bound to negatively charged phospholipids
(phosphatidylserine, PS, and phosphatidylinositol, PI) in a CaZtdependent manner. All but synaptotagmin IV were also found to bind
to vesicles with a 1: 1 mixture of negative and neutral phospholipids.
46 1
FIG. 8. Stereo view of the phospholipid binding site in the third domain of rat
annexin V, including calcium ions in sites AB and AB, ( a ) with bound glycerophosphoserine (GPS), and (b) with bound glycerophosphoethanolamine (GPE). GPS and GPE
a s well as the calcium ions are drawn dark. The oxygens in the protein fragment are
drawn gray. (Redrawn from Swairjo et al. 1995).
462
FINN AND DRAKENBERG
None bound to only neutral phospholipid vesicles. The difference between IV and the other was localized to an Asp-to-Ser substitution
and verified by mutant studies. Falke and co-workers (137) have
shown that an isolated C2 domain from cPLA2binds Ca2' and binds
to phosphatidylcholine vesicles in a Ca2+-dependentmanner. The domain binds two Ca2+in the absence as well as in the presence of
phospholipid vesicles; however, the binding affinity increases from
KD = 24 p M in the absence of vesicles to KD = 3 pM in the presence
of vesicles.
In both instances the binding appears to be cooperative, resulting
in a steep response to changes in Ca2' concentration. The binding of
Ca2' to the C2 domain also increased the tryptophan fluorescence
from Trp71, showing a conformation change caused by Ca2+binding.
The fluorescence increase was twice as large in the presence of vesicles as in the absence, showing that the tryptophan environment
changes upon membrane binding. Falke and co-workers also studied
the kinetics of the Ca2+binding. It was thus found that in the phospholipid free state there is a single off rate for the two Ca2+.This
shows that either the two off rates are the same or the off rate for
C2"Ca is much faster than that for C2*Ca2. The latter could well be
imagined for a system with strong cooperative Ca2+binding and has
been observed for EF-hand proteins (62, 138).In the presence of vesicles, however, the Ca2+off rate is much slower and there are clearly
two off rates differing by a factor of 10 (Fig. 9). Very similar rates
were obtained by measuring either the Ca2' off-rate directly, the conformational change caused by the Ca2+release, or the release of the
C2 domain from the phospholipid surface. This therefore shows that
the C2 domain with a single Ca2+is still bound to the membrane
surface, at least transiently. The equilibrium measurements, however, indicate that under conditions where, on average, one Ca2' is
bound per C2 domain, about half of the C2 domains will bind two
Ca2+and the other half will be in the apo form. It would be interesting
to confirm this with a stopped-flow experiment designed in such a
way that the final state will have a Ca2+/C2domain ratio of 1: 1.
The Ca2'-dependent binding of protein kinase C to phospholipid
membranes has been shown to be specific for negatively charged phospholipids, and it has also been shown that Ca2+binding depends on
the presence of negatively charged phospholipids (116). The binding
of PKC to phospholipid membrane depends not only on Ca2+but also
on the presence of diacylglycerol (117). This effect is very specific for
PS-containing micelles. Even though PKC binds to phosphatidic acidcontaining micelles, this binding is not affected by diacylglycerol. Fur-
463
-0.5
-0.25
~
Lu
K
1.25
C
1.00
Time (sec)
F ~ G9.. Kinetics of CaL+dissociation from the cPLA, C2 domain in the absence and
presence of phosphatidylcholine vesicles. (a) Stopped-flow measurement of the Cad+off
rate using the fluorescent calcium indicator Quin-2. (b) Conformational changes triggered by Ca2+removal from the C2 domain monitored using the intrinsic fluorescence
of Trp71. (c) Membrane release of the C2 domain followed by protein-to-membrane
FRET. (From Nalefski et al. 1997 with permission.)
464
FINN AND DRAKENBERG
thermore, only phosphatidyl-L-serine, and not phosphatidyl-D-serine,

will cause the high-affinity binding of PKC to diacylglycerolcontaining membranes (118).
The structure of the N-terminal C2 domain of synaptotagmin has
been solved to a resolution of 1.9 A for both the apo and Ca2+forms
(139). The structure is described as an eight-stranded @-sandwich
with a conserved four-stranded motif named a C, key. The Ca2+-binding site is found at the top of the @-sheetwith ligands from four Asps
in the loops connecting @2/@3and @6/@7(Fig. 10). A second possible
but not occupied Ca2+-bindingsite was identified in the other end of
the molecule. Ca2+binding to this domain caused only minor structural changes, indicating that the site is preformed. The structure
of a C2 domain in an intact protein has also been determined for
phosphoinositide-specific phospholipase C-8, (133-135). The structure
of PI-PLCS1 comprises an N-terminal PH domain, a catalytic core,
and a C-terminal C2 domain. Only the C2 domain will be further
FIG. 10. Structure of the synaptotagmin C2 domain showing the calcium binding
site a t the top and ligands for a potential calcium binding site a t the bottom. (From
Sutton et al. 1995 with permission.)
465
discussed here. From a comparison of the structure of C2 domains

from synaptotagmin and PI-PLCG1, it is apparent that there are a t
least two different topological variants of the C2 domains (135).It is
possible to go from the synaptotagmin topology to the PI-PLC topology by connecting the C terminus of /3-strand 8 with the N terminus
of /3-strand 1 and cutting the loop between /3-strands 1 and 2 open.
This, however, appears to have no effect on the calcium-binding regions that are in the opposite end of the structure. Using lanthanum
as a calcium analog showed that three loops form the metal-ion-binding sites with two adjacent metal ions. For another crystal form of
PI-PLC, with samarium as a replacement for calcium, the same two
sites were found (134).A comparison of the apo and lanthanide structures revealed only minor changes in the metal-ion-binding loops;
however, the maximum distance between metal-ion-binding loops 1
and 2 changed from 12 A in the apo form to 21 A in the Sm3+form
(Fig. 11).Even though the functional role of the C2 domain in PI-PLC
is still unknown, mutational studies have shown that it is essential
for catalytic activity (130, 140). The calcium-mediated binding of
C2 domains to lipid membranes is still not well understood. The
calcium-bridging model is to us very unattractive because it does
not explain the specificity for PS. More studies are certainly needed,
and our expectations are that a more specific membrane-binding
model will eventually emerge, as it has for the annexins and the
GLA domain in the vitamin K-dependent serum proteins (discussed
later).
N649
tc..
FIG.11. Stereo view of C2 jaws of the apo and Sm forms of PLC-81. The apo form
is shown with dark backbone and light side chains, and the Sm+form with light backbone and dark side chains. (Redrawn with permission from Grobler et al. 1996.)
466
FINN AND DRAKENBERG
E. y-CARBOXYGLUTAMIC
ACIDSITES
The amino acid y-carboxyglutamic acid (Gla) is not one coded by
DNA, but glutamic acid is posttranslationally carboxylated. Gla contains a malonic acid moiety with an affinity for Ca2 of Kb = 30 M-.
This is too weak to be of any biological relevance by itself. However,
many of the modified amino acids often appear together, though not
necessarily in sequence, as in the case of the Gla module of blood
coagulation factors, in which there are ca. 10. Robertson et al. (141)
have shown, by using 43CaNMR, that calcium binding to a Gla-Gla
moiety is sufficiently strong to be relevant under physiological conditions (KO = 0.6 mM). Likewise Colpitts and Castellino (142) have
found a KD = 1.6 mM for calcium binding to the peptide CysIleGlaGlaIleCys but with a stoichiometry of two peptides per calcium ion.
Later, we will discuss the Gla domain from the blood coagulation system at some length and, to a minor degree in Chapter 4, the Glacontaining conotoxins, whereas for the bone Gla and matrix Gla proteins (143)the reader is referred to older reviews (144).
The Gla-containing proteins of the blood coagulation system are all
modular with the Gla-domain N terminal. Factors VII, IX, and X and
protein C form a group with the same domain structure. The Gla
domain is followed by two EGF-like domains, of which the first one
also binds calcium (see Section IV.C), and a serine protease domain,
also with a calcium-binding site (see Section 1V.D). Prothrombin
and protein S have somewhat different domain structures (145, 146).
I n prothrombin the Gla domain is followed by a hexapeptide with
a disulfide loop, two kringle domains, and the C-terminal serine
protease domain. In protein S the Gla domain is followed by the
thrombin-sensitive loop, four EGF-like domains, and the C-terminal
domain that is homologous to plasma steroid hormone-binding proteins.
The DNA sequence for the Gla domain show that there should be
9-12 Glu residues. These Glu residues are all carboxylated to Gla by
a vitamin K-dependent carboxylase. This modification is a prerequisite for calcium binding and biological activity. The first attempts to
characterize the calcium binding properties of Gla domains were
made in the early 1970s (147-149). These studies were made before
the Gla residue had been identified (250, 151). Since these early
works, the calcium-binding properties of various Gla domains have
been reported on several occasions. In most cases equilibrium dialysis
experiments with 45Cahave been used, but spectroscopic techniques
as well as calorimetry and ion-selective electrodes have also been em-
467
ployed (152-163). Equilibrium studies have indicated that all vitamin

K-dependent plasma proteins bind 6 to 12 calcium ions (159). The
low affinity of these calcium binding sites (KDbetween 0.1 and 2 mM)
and their large number have made the determinations difficult. In
two recent studies (164, 165) on synthetic peptides with sequences
corresponding to the N-terminal domain(s1 of factor IX and protein C
it has been shown that it is not sufficient with only the Gla domain
[factor IX(1-42) and protein C(1-3811 for full calcium binding. However, including the amino acids from the hydrophobic stack [factor
IX(1-471, protein C(1-48)I seems to restore both calcium and lipid
binding up to the level in intact proteins. Even though the sequences
of these peptides are 50% identical, there seem to be some important
differences regarding calcium and lipid binding of the truncated peptides. Removing the five C-terminal residues from factor IX(1-47)
drastically decreases calcium binding as well as the binding to phospholipid vesicles, whereas for protein C( 1-38) the calcium binding
has been reduced merely by a factor of two to three, and full binding
to phospholipid vesicles can be obtained by a modest increase in the
calcium concentration. It is a t present not clear what causes this difference, but it seems to reside in the first 30 amino acids, because a
peptide with the first 30 amino acids from the human protein C sequence and the remaining eight residues from the factor VII sequence
showed a calcium-binding pattern similar to the 38-mer with the protein C sequence (142). Culpitts et al. (1661, using synthetic protein C
peptides selectively 13Clabeled on individual Gla residues, have been
able to show that Gla 6, 16, 25, and 26 (factor X numbering) participate in the stronger calcium binding.
Even though it seems clear that it is sufficient with the Gla domain
and the hydrophobic stack to restore full calcium-binding activity, the
apo form of these peptides appears essentially structureless (1671,
whereas the fragment from factor X (also comprising the N-terminal
EGF domain) has a more ordered structure (168).It has also been
shown recently that the dynamic properties of Trp42 in the apo form
of factor X peptides critically depend on the length of the peptide
(Drakenberg, unpublished results). The correlation time of Trp42 is
lengthened ca. threefold (from 1.5 t o 4.5 ns) by extending the
46-amino acid peptide with either the EGF domain or five amino
acids from the C-terminal helix from calbindin Dgk.It is thus clear
that even though the Gla domains of factor IX and protein C do not
need the EGF domain for full calcium-binding properties, the apo-Gla
domain of factor X needs the EGF domain to stabilize the structure.
The structure of the apo form of the Gla domain from factor X, in the
468
FINN AND DRAKENBERG
GlaEGF fragment, has all the secondary structure found in the calcium form (discussed later); however, the relative orientations of the
structural elements differ, especially in the N terminus. In this structure the Gla residues are all, as could be expected, exposed to the
solvent, and the hydrophobic residues Phe4, Leu5, and Val8 form a
cluster facing the interior of the Gla domain [Fig. 12(a)l.
Pheh
-7
Leu5
Val8
IY
Gla7
&
Gla14
aPO
Ca2+
FIG.12. Comparison of the energy-minimized average NMR structure of the Gla
domain from factor X with model-built calcium-loaded Gla domain (based on the X-ray
structure of prothrombin). (a) Location of residues Phe4, Leu5 and Val 8. (b) Location
of Gla residues 6,7, 16, 19,20,26, and 29.The essential residues 7, 16, 20,26,and 29
are shaded dark.
469
The structure of calcium-saturated Gla domains has been determined both in the crystal state by X-ray crystallography for prothrombin fragment 1 (169)and for factor VII in complex with tissue factor
(170)and in solution by NMR for the N-terminal47 amino acids from
factor IX (171, 172). These three structures are essentially the same
and are shown in Fig. 12(b). The main secondary structure elements
of the structures are three helixes (14-17, 25-31, and 36-47) and a n
N-terminal loop. Seven Ca2* are identified in the structure, and they
have coordination numbers ranging from three to seven. Five of these
Ca2+ are chelated between Gla6, Gla7, Gla16, Gla25, Gla26, and
Gla29, rendering four of them inaccessible to water, In the NMR
structure determination there is no direct information on the position
of the Ca2+;however, in a refinement they used a genetic algorithm
to identify the locations of the Ca2+(172).The identified Ca2+positions
agreed well with the crystallographically determined positions and
were subsequently used in the refinement of the solution structure.
The two crystal structures of the Gla domains are very similar, even
though they are in completely different settings, with a n RMSD of
0.63 A for 40 a-carbons. A similar comparison of the solution structure of factor !X and the prothrombin crystal structure reveals an
RMSD of 1.3 A for residues 12-47. The backbone conformation for
the 11 N-terminal residues differs significantly between the solution
and crystal structures. All Gla domains except that in human factor
IX have a n Ala as the N terminus. In factor IX it is a Tyr.This larger
residue cannot be accommodated within the crystal structure, which
may explain the difference between the crystal and solution structures in this region. However, the unusual solvent mixture with 3 M
urea and 2.5 M guanidine hydrochloride used in the NMR study may
also cause some structural changes. Based on these structures and
the structure of the apo-Gla domain from factor X determined by
NMR (1731, we can now speculate about how the Gla domains are
bound to phospholipid membranes (122, 174, 175). It can be assumed
that the conserved hydrophobic residues, Phe4, Leu5, and Val8 (human factor X numbering) are important for the membrane binding
because they are exposed on calcium binding (Fig. 12). Also, based on
site-specific mutagenesis studies of Leu5, it has been suggested that
this side chain is bound into the phospholipid membrane. The hydrophobic binding is certainly not sufficient to explain all experimental data on the membrane binding. For optimal binding the
membrane must contain negatively charged head groups, phosphatidylserine. This clearly indicates that also electrostatic effects are important for the membrane binding, and Li et al. (122)argue that some
470
FINN AND DFtAKENBERG
of the peripheral calcium ions in the Gla domain as well as the side
chains of Lys3, Arg9, LyslO, and Argl5 in human prothrombin may
be directly involved in the interaction with the membrane. However,
only for residues 9 and 15 is the charge conserved in the Gla domains,
so we may argue that they are strong candidates for direct interaction
with the negatively charged lipid head groups.
IV. Extracellular Calcium-Binding Proteins
A. INTRODUCTION
The total concentration of calcium in the blood plasma is ca. 2 mM,
and about half of it is bound to proteins, mainly serum albumin. This
high calcium concentration is typical for all extracellular fluids, in
stark contrast to the very low free calcium concentration in resting
cells. One might therefore be led to believe that there is no specific
function of calcium in the extracellular fluids. This is certainly not
true. It has been known for almost a century that calcium is critical
for blood coagulation, and it is also well known that calcium is a major component in our skeleton. It is, however, also obvious that the
requirement for a protein to be calcium binding in a milieu with free
calcium at 1 mM is very different from one with calcium at p M levels.
B. CONANTOKINS
Cone snails produce biologically active peptides to paralyze their
prey. The peptides are known as conotoxins (176-1 78). Most of them
are small and stabilized by disulfide bonds. Two unusual members of
the conotoxin family are conantokin G from Conus geographica (179)
and conantokin T from Conus tulipa (180),which lack disulfide bonds
but are rich in y-carboxyglutamic acid residues:
con-G: GlyGluGlaGlaLeuGlnGlaAsnGlnGlaLeuIle ArgGlaLysSerAsn
con-T: G l y G l u G l a G l a T y r G l n L y s M e t L e u G l a A s n L e u A r g a
The role of the Gla residues is not yet fully understood, but a synthetic conantokin peptide with Glu instead of Gla has been shown to
be inactive. However, it is still not clear whether Ca2+is essential
for biological functioning of the conantokin peptides (181, 182).There
appears to be no thorough study of the calcium ion binding to these
peptides, though CD spectroscopy has been used to study the change
in helical content as a function of calcium concentration (183). The
47 1
two peptides behave quite differently. con-G had almost no helical

structure in the calcium-free state but it increased to ca. 50% in the
presence of 9 mM Ca. con-T on the other hand already had roughly
50% helix without Ca2+added, and this did not change upon addition
of calcium. The same effect was found with Cu2- whereas Mg2+and
Zn2 were found to introduce almost 70% helix in con-G (Fig. 13).
Another interesting observation is that ca. half of the calcium-induced
helix in con-G remains in the presence of excess EDTA. The authors
have interpreted this as being due to a very strong calcium ion binding to con-G; however, we also have to consider that EDTA has been
shown to interact with various proteins in a calcium-dependent manner (184). We therefore find a more likely explanation that the Ca2 EDTA binds t o con-G and introduces some helical structure. Skjaerbaek et al. (183)have also determined the solution structure of con-G
in the presence of calcium, and it appears to be more helical than in
the absence of calcium. However, the resolution is not sufficient to
define binding sites for calcium, making the presence of a site strong
enough to compete with EDTA for calcium less likely. Rigby et al.
have determined the solution structure of con-G to high resolution in
the absence (185)as well as in the presence of CaL+(186).The peptide
is well structured under both conditions, with backbone RMSDs of 0.8
and 0.6 A, respectively. As shown in Fig. 14, there is a pronounced
structural change upon CaL+binding, with the formation of a straight
a-helix from Gla3 to Lysl5. In this structure Gla residues 3, 7, 10,
and 14 become aligned in a linear array on one side of the helix. The
Ca, which cannot be seen by NMR, were localized with a genetic
algorithm (172).The best model was obtained with 4 Callpeptide. All
the Ca2 were found to have 3 or 4 ligands from the Gla side chains.
This arrangement has some similarity to the Ca binding to the Gla
domains in blood coagulation proteins (170, 1871, in which there also
is a linear array of Ca2- coordinated to Gla side chains. In both cases
the calcium binding will also expose some hydrophobic amino acids t o
become solvent accessible.
C. EGF-LIKEDOMAINS
The epidermal growth factor (EGFblike domains are approximately 45 amino acids long and contain six cysteine residues that are
paired in a characteristic manner, 1 to 3, 2 to 4, and 5 to 6, with a
double-stranded 0-sheet as the main structural feature. The EGF domain has been found in a wide variety of proteins including these
involved in blood coagulation, fibrinolysis, neuronal development, and
472
FINN AND DRAKENBERG

4 10
3 10-
l
i
0
+-.
---i)-
-+-
No Cac2
76 UM CaC12
0.38 mM CaC12
1.5 rnM CnC12

3.0 rnM CaCIz
-74.5 mM CaC12
-6.5
mM CaCn
+9.2 mM Cam
-u-
E 210
a
-&-
Q, -1 ld
-210
Eu
-4 10
180
210
220
230
240
250
200
210
220
230
240
250
6 lo4
200
4 10
-0 2
10
a,
Q, -2
104
-4 10
190
FIG. 13. CD spectra of con-G and con-T under different conditions. (a) CD spectra
of con-G a t different CaC1, concentrations. (b) CD spectra of con-T a t different CaCl,
concentrations. ( c ) CD spectra of con-G with different metal ions. (d) CD spectra of conT with different metal ions. (e) CD spectra of con-G with fixed CaClz concentration and
different EDTA concentrations. (From Skjaerbaek et al. 1997 with permission.)
473
- 1
-210
I90
200
210
220
230
240
250
3 lo
-8
210.
t 10
--
4 o mM caaz + 3.0 mM EDTA

4 0 mM Cam + 20.0 nw mTA
E
0
2
a
v
~
+ 05 mM EDTA
iro4
-2 l @
180
200
210
220
230
240
250
Wavelength (nm)
FIG.13 (continued)
cell adhesion (188).A subset of these domains has a hydroxylated

aspartic acid (Hya) or hydroxylated asparagine (Hyn), most prominently found in vitamin K-dependent plasma proteins with a n identified consensus sequence of Cys-X-Hya/Hyn-X-X-X-X-TyrPhe-X-Cys-XCys (189, 190). The consensus sequence for Asp/Asn hydroxylation
has also been found in EGF domains from, for example, fibrillin,
which has 46 EGF domains whereof 43 have the consensus sequence
for hydroxylation (191-193); fibulin-1, in which four of nine EGF domains have the consensus sequence (194-197); and the Notch receptor, in which 21 of 36 EGF domains have the consensus sequence
for hydroxylation (198).[For a review see (1881.1 The presence of the
hydroxylated amino acid has also been correlated with the presence
of a Ca2+-bindingsite with a KDin the range of 10 to 100 FM (199204). Studies of Ca-binding properties of a protein fragment from
protein C containing the two EGF domains demonstrated that this
fragment contains one Ca2+-bindingsite (205-207). This site was fi-
474
FINN AND DRAKENBERG
C
Gla 14
Gla 10
Ca2+
N
N
FIG.14. Calcium-induced transition in c o n 4 with the apo form to the left and the
calcium form to the right. (From Rigby et al. 1997 with permission.)
nally localized to the N-terminal EGF domain in protein C as well as

in factors IX and X (206,208,209).
The isolated EGF domains have been shown to bind Ca2+with low
affinity (KD, 0.5-5 mM) (206,208-212); however, these domains usually appear a t least in pairs and often several in tandem or as in the
vitamin K-dependent serum proteins with a Gla domain bound
N-terminally. It has been shown that when the N-terminal domain is
intact, the Ca2+affinity to the next domain has increased by approximately 10-fold, to KD= 0.1 mM, making these sites essentially saturated under physiological conditions (162,213,214).Similarly, Handford and co-workers (211, 215) have found that a recombinant
fragment of fibrillin comprising a pair of EGF domains has a high
affinity (KD= 0.35 mM) and a low affinity (KD= 9.2 mM) site. The
increased affinity to the C-terminal domain is explained as caused by
the donation of a ligand from the N-terminal EGF domain to the site
of the C-terminal EGF domain. The Ca2+binding to the high-affinity
site is in good agreement with a value of 0.25 mM reported as the
average affinity to a proteolytic fragment of fibrillin containing seven
intact EGF domains (293). Protein S has four EGF domains in tandem, and it has been shown that the Ca2+affinity to this protein is
much stronger (KD,
to
MI than for other EGF-containing
475
proteins studied so far (216).There is no obvious sequence variation

to explain this extremely strong Ca2+binding, and the isolated EGF
domains from protein S do not display any extreme Ca2+affinity (217).
However, fragments containing various numbers of EGF domains
from protein S showed a very interesting Ca2+affinity (218). Extremely strong Ca2+binding was thus found for fragments containing
EGF1-4 and EGF2-4 (Kr, <
in physiological salt) whereas hundredfold higher Kd)s where found for fragments not containing EGF4.
Another noteworthy observation is that EGF3 has a 350-fold higher
affinity for calcium in fragment EGF2-4 than in fragment EGF1-3.
It has been shown on several occasions that the Ca2+affinity to an
EGF domain is reduced significantly if the protein segment N-terminal to it is removed. This may be an effect of stabilization of the site,
either by having a calcium ligand in this part of the protein or just
by stabilizing the conformation of the calcium site. The data reported
on the site in EGF3 of protein S demonstrates an influence on Ca2+
binding from a domain on the C-terminal side. This indicates that the
three-dimensional structures of these fragments are such that there
are direct interactions between nonadjacent domains and not a linear
array of domains. Another protein displaying an unusually high calcium binding to EGF domains is Notch-l, a fundamental receptor for
cell fate decisions (219). This protein has 36 EGF domains, of which
21 have the consensus sequence for calcium binding. Calcium binding
studies to the recombinant fragments comprising EGFll-12 and
EGF10-13 have shown that this protein has an calcium affinity in
the low pM range and thus somewhere in between protein S and the
other calcium-binding EGF domains.
Although the role of AsplAsn hydroxylation on calcium binding has
been a matter of debate, there have been very few attempts to resolve
this question. Morita and Kisiel (200) compared the calcium affinity
of bovine and human factor IX. The calcium affinities were shown to
be similar even though the degree of hydroxylation is different for
these two proteins. The conclusion was uncertain, however, due to
interference with other calcium-binding sites in the proteins. Later
Selander Sunnerhagen et al. (210) showed, using single EGF domains,
that even though there is a small effect due to hydroxylation, it is not
significant, so what might be the role of the hydroxylation is still an
open question.
The structures of several isolated EGF domains have been determined by NMR, the first one more than 10 years ago (220-232). These
structures are all quite similar, revealing a domain structure consisting of two relatively independent subdomains. The N terminal
476
FINN AND DRAKENBERG
constitutes about two thirds of the domain, and its major structural
motif is a &sheet with two or three strands. The C-terminal third of
the domain has been described as a small antiparallel double-hairpin
structure (231) or as two loops connected by a short p-sheet and a
flexible C terminus (224).The three-dimensional structures of EGF
domains have also, more recently, been determined by X-ray crystallography (170, 233-236). Only a few of these studies deal with EGF
domains of the Ca2+-bindingtype (170,235,236).In the following we
will restrict the discussion to the EGF domains of the Ca-binding
type (cbEGF).
The solution structure of the N-terminal cbEGF domains from factor IX (fIX-EGFN)and factor X (fX-EGFN), both containing Hya, in
the absence of calcium, has been determined by means of 2D NMR
(227-229). The structure of fX-EGFNhas also been determined in the
presence of calcium (230). Even though the calcium ion cannot be
seen by NMR, the calcium site can be readily identified (Fig. 15). In
the calcium form of ~X-EGFN
there is a cavity of appropriate size for
a calcium ion that is lined with oxygens oriented toward the center.
There are five well-defined ligands, two side-chain carboxylate groups
(Asp46 and Hya631, one side-chain carbonyl group (Gln49), and two
45
64
50
70
FIG.15. Stereo view of the calcium-binding site of fX-EGFN. Amino acids unambiguously assigned as calcium ligands are labeled and the coordinating oxygens are marked
with filled symbols (Gly47, Gln49, Hya63, and Gly64). Asp46, a potential ligand, is
marked with an open circle.
477
backbone carbonyls (Gly47 and Gly64). This was later confirmed in

two crystallographic studies on the N-terminal EGF domain of factor
IX (235) and for the N-terminal EGF domain in factor VII in complex
with tissue factor (170). Furthermore, the side chain of the Hya residue is rotated in such a way that the hydroxyl group cannot be a
ligand to the calcium ion (230). The changes in the NMR spectra
caused by calcium binding are localized to residues 46-51 and 62-68,
and are not very dramatic. A comparison of the structures of the apo
and calcium forms of fX-EGFN shows no major differences. There are,
however, some minor differences. The N terminus is more well defined in the calcium form and has moved slightly toward the main
p-sheet. The turn connecting the two strands in this p-sheet is bent
somewhat toward the N terminus. These conclusions are confirmed
by the X-ray crystal structure of fIX-EGFN with calcium (2351, and
the backbone RMS deviation between the two structures is 1.2 A
(237).
There are now also available structures on the Gla-EGF pair from
factor X (168)and on a n EGF pair from fibrillin (2371,both in solution
and solved by NMR in the presence of calcium. Information on the
function of calcium in this kind of structure is still very limited. Only
for the Gla-EGF pair is information available for both the apo and
calcium-loaded states. Even though the structures of the individual
domains are reasonably well defined, the relative orientation, which
is probably more interesting, is not. It is, however, from NMR as well
as SAXS data (1681, clear that the relative orientation of the two domains is calcium dependent. In the apo state there is a more or less
linear arrangement of the two domains, whereas in the calcium form
there is a more compact structure, with the two domains oriented
almost perpendicular to each other (Fig. 16). This orientation, however, does not agree with what was found in the crystal state for the
complex between factor VII and tissue factor (170),where the Gla and
the N-terminal EGF domain are arranged in a more or less linear
manner. We have to consider, though, that these are different proteins, though quite similar, each with its own specific interactions.
Based on the crystal structure of the isolated cbEGF domain of factor
IX,it was speculated that an N-terminal domain might donate a ligand to the calcium site in the following domain (235).In none of the
structures determined so far with two domains and a calcium-binding
site in the C-terminal domain has such a ligand donation been observed. More recently Downing et al. (237)have found from sequence
analysis of domain pairs indications that they can be divided into two
classes based solely on the number of residues linking the two do-
478
FINN AND DRAKENBERG
(b)
FIG.16. Stereo representation of the secondary structure elements and relative orientation of the domains in the Gla-EGF domain pair. (a) with 1 equivalent of CaZ
added and (b) in the absence of Ca. (From Sunnerhagen et al. 1996 with permission.)
mains. Sequence alignment of the two classes of pairs was applied to

define the consensus sequences as shown in Fig. 17. They also suggest
that the Class I1 pairs are likely to adopt a different conformation
and that a conserved carboxylatelcarboxamide side chain may donate
a ligand to the following domain, as is observed in the crystal structure of fIX-EGFN. The confirmation of this hypothesis has to await
the structure determination of such a domain pair. This hypotheClass I
EGF1: Dx D/N ECxxxxxxCxxxxxCxNxxGSY/F xCxCxxG Y/F xxxxxxxxC
EGFB: xD IN D/N ECxxxxxxCxxxxxCxNxxGSYIF xCxCxxG Y/F xxxxxxxxC
Class I1
EGF1: xx DM xCxxxPCxNG G/A xCxxxxxx Y/F xCxCxxG Y/F XGXXC
EGFS: xx D/N W D/N E D CxxxPCxNG GIA xCx DM x IN xx Y/F xCxCxxG Y/F xGxxC
FIG. 17. Consensus sequences for class I an? class I1 EGF-cbEGF and cbEGFcbEGF pairs. x indicates that there is no preferred amino acid and letters separated
by / indicates that both amino acids are commonly present.
479
sis may explain the high calcium affinity in the fragments from
Notch studied by Rand et al. (219)with KDin the low pM range even
in the presence of physiological salt concentration. For protein S on
the other hand, this does not offer any explanation for the high affinity because the protein S pairs with high affinity belong to Class
I(218).
D. SERINE
PROTEASES
It has long been known that some serine proteases contain one or
two calcium-binding sites. It has thus been shown that trypsinogen
has two sites, one of low affinity in the activation peptide and one
with a higher affinity (KD= 1.6 .
MI that is also present in the
active enzyme (238, 239). The structure of trypsi? with one bound
calcium has been determined to a resolution of 1.8 A, and the calcium
ligands could be identified (Fig. 18) (239). The calcium ion is coordinated by six ligands a t the edges of an octahedron. Four ligands are
from a loop of the protein, Glu70-Glu80, where Glu70 and Glu80 are
coordinating with their side-chain carboxylate groups and Am72 and
Val75 are coordinating with the backbone carbonyls. The two remaining ligands are water molecules that are also hydrogen bonded
to Glu70 and Glu77, respectively. Sites similar to this archetype can
be inferred from the sequences of chymotrypsin; blood coagulation factors VII, IX, and X; and protein C (184,240-243). For trypsin, chymotrypsin, and their zymogenes it has been shown, using 43CaNMR and
stopped-flow experiments, that the Ca2 exchange is fairly slow even
though the binding affinity is modest (Fig. 19) (184). The on rate of
the calcium ion was thus found to be 105-10b s two to three orders
GLU70
ASN72
GLU70
ASN72
GLUE0
GLU80
FIG.18. Stereo view of the calcium-binding loop in trypsin including the calcium ion
and internal water molecules.
480
FINN AND DRAKENBERG
40
20
-20
FIG.19. '"a NMR spectra of (a) 2 mM tosyl-trypsin and 2.5 mM Ca2' at pH 6.5 and
24OC; (b) 1 mM tosyl-trypsinigen and 1.3 mM Ca2+at pH 6.2 and 24C.
of magnitude slower than what could be expected for a diffusionlimited process. Also, the off rate from chymotrypsin(0gen) is at least
one order of magnitude faster than that from trypsin(ogen). These
findings are in agreement with the reported higher flexibility in the
calcium-binding site in chymotrypsin apparent in the crystal structure. However, the crystal structures do not offer any explanation of
the slow calcium-ion exchange because the binding site is close to the
surface and there are two water ligands.
The calcium-binding sites in the protease domains of coagulation
factors VII, IX, and X and protein C all have calcium affinities comparable to that in trypsin (241, 244-246). In the crystal structures of
factors IX and X, no calcium ions could be identified, most likely due
to the crystallization conditions because the calcium-binding site is
present though not occupied (234). The crystal structure of factor VII
in complex with tissue factor shows the presence of calcium in the
48 1
serine protease domain with a structure very similar to the one in

trypsin. Calcium binding to the serine protease domain of protein C
has a pronounced effect on the activation of protein C by thrombin.
This activation is very slow in solution and is reduced even further in
the presence of calcium. However, if protein C is activated by the
thrombin-thrombomodulin complex, the activation is rapid and is increased in the presence of calcium (234). Calcium binding to the serine protease domain of coagulation factor VII is important both for
amidolytic activity and for interaction with the tissue factor (247).
The subtilisins all have a calcium-binding site whose occupation
protects the enzyme against autolysis (248). This site is very different
from the one in trypsin, however. The ligands come from three different parts of the protein, and the calcium binding seems to bring the
N terminus close to the 75-81 and 40-42 loops with coordination to
the side chains of Gln2 and Asp41, and to the peptide carbonyl of
Leu75, Asn77, Gly79, and Val81 (249-251).
E. Q-LACTALBUMIN
AND LYSOZYMES
a-Lactalbumins (aLACs) are milk proteins that play a n important
role in the biosynthesis of the milk sugar lactose. By binding to the
enzyme galactosyltransferase (GT), the lactose synthase complex is
formed. This binding is reversible, and aLAC functions as a regulatory subunit. GT alone is unable to catalyze the synthesis of lactose
a t physiological glucose concentrations due to its low affinity for glucose. In the lactose synthase complex the affinity for glucose has increased 1000-fold and lactose can be produced (252-254). On the
other hand are the lysozymes (LZs), lytic enzymes that catalyze the
degradation of peptidoglycans. From the high degree of homology between aLAC and LZ as revealed by amino acid sequences (255,2561,
intron-exon arrangements (257, 2581, and three-dimensional structure (259, 2601, it has been postulated that these functionally different proteins have evolved from a common ancestral molecule. It is
recognized that all known aLACs possess one high-affinity Ca2+-binding site, and crystal structures have revealed that the Ca2+-binding
site in aLAC is formed by the side-chain carboxylate groups from
three aspartate residues (Asp82, Asp87, and Asp%), which are conserved in all aLACs. Also, the backbone carbonyls from Lys79 and
Asp84 contribute to the ligation of Ca2+as well as two water molecules (259, 260). Ca2+ binding to lysozymes from horse and pigeon
with the conserved aspartate residues required for Ca2' binding in
aLAC was initially less well defined. Even though both lysozymes
482
FINN AND DRAKENBERG
FIG. 20. Stereo view of a n overlay of the structures of aLAC from human, guineapig, and goat. (From Pike et al. 1996 with permission.)
were shown to be Ca2'-binding proteins (261-263) the crystal structure of the same proteins showed no electron density for Ca2+(264).
On the other hand a mutant of human LZ in which the two missing
coordinating Asp residues were introduced showed both Ca2+binding
and a Ca2' in the crystal, with a geometry analogous to that in human
and baboon aLAC (265). In agreement with the Ca2'-binding studies
of LZ from horse and pigeon (261-263) a 43CaNMR study of CaZt
binding to the two lysozymes and bovine and human aLAC showed
very similar behavior, strong Ca2+binding for one site, and the same
symmetry of the site as revealed by the quadrupole coupling constant
of the bound 43Ca2t.Furthermore, the off rate of the bound Ca2+was
shown t o be slow for all four proteins (266, 267).
The Ca2+-bindingsite in a-lactalbumin has some similarity with the
EF-hand binding sites in that it consists of a helix-loop-helix motif,
with all protein ligands contained in the loop. However, the pattern
of ligands is not the same, and the site in aLAC has been named an
"elbow" to differentiate it from the EF hand (260,268).Another difference between these two kinds of Ca2+-bindingsites is that the elbow
is a single site whereas the EF hands appear in pairs (discussed earlier). Crystal structures of several aLACs are available (260,269,270)
and are all very similar. The structure is divided into a large domain
( a domain) and a small domain ( p domain) by a cleft (Fig. 20) (270).
483
The Ca2+-bindingsite is in the interface between the two domains.

The only variation between the various structures occurs for residues
101-110. In the structure from human and recombinant bovine
aLAC, residues 105-1 10 adopt a distorted a-helical conformation,
whereas in goat and guinea-pig aLAC these residues form a loop
(270).The difference may be due to different crystallization conditions
and may indicate that this region is flexible in solution (271). Mutations in the flexible loop indicate that this is important for the interaction with GT.
Recently there have also appeared studies on the thermodynamics
of the Ca2+-loadedpartially unfolded state (2721, the importance of
the disulfide bonds (2731, and the functional role of calcium-binding
residues (274).Furthermore, aLAC has been a model for protein folding because it forms a molten globule state during folding (275, 276),
which will not be discussed here.
F. CADHERINS
The cadherins form a family of cell-cell adhesion receptors (277,
278). They are transmembrane glycoproteins with typically five tandemly repeated extracellular domains (CAD repeats) (279, 280). The
adhesive action of cadherins depends on the presence of Ca2+(278). A
detailed study of Ca2+binding to E-cadherin domains has shown that
two CAD repeats are needed for Ca2+binding (2811. It was shown that
the domain pair comprising the two N-terminal domains, ECAD12,
bind Ca2' strongly compared to the extracellular Ca2+concentration,
whereas individual domains showed no or very weak Ca2+affinity. A
crystal structure solved for the N-terminal domain of N cadherin
shows that the polypeptide fold includes seven &strands arranged in
two @sheets with the N and C termini a t opposite ends of the domain,
making a linear arrangement of the five domains possible. A metalion binding site occupied by an Ybs' has been identified with two bidentate ligands from G l u l l and Glu69 (282).The solution structure
of the N-terminal domain of E cadherin in the presence of Ca2+has
been determined by NMR. Even though the Ca2+cannot be "seen" by
NMR, Ca2+ligands were inferred from Ca2+-dependentchemical shift
effects. In this way G l u l l , Glu69, and Asp100 were suggested as ligands (280, 2831, in agreement with the crystal data (282) in which
two water molecules could also be identified as ligands, still resulting
in an incomplete coordination of the metal ion. Recently the crystal
structure of the two N-terminal domains of E cadherin was solved
(284). The general location of the Ca2+-bindingsite suggested in the
484
FINN AND DRAKENBERG
earlier studies is confirmed; however, it is now shown that in each

domain-domain interface there are three Ca2+.Furthermore, the
structure reveals a twofold symmetric dimer. The dimeric structure
is in good agreement with solution state studies, which have shown a
Ca2+-dependentdimerization of the same protein fragment (2811. A
model based on the assumption that the relative orientation of the
two domains in ECAD12 is also used for the other linking regions in
a 240-A-long extracellular region. This is in good agreement with a
value of 220 A obtained by electron microscopy (285) in the presence
of calcium.
G. SPARC
The protein SPARC (Secreted Protein Acidic and Rich in Cysteine)
is also known as BM-40 and osteocalcin (286-288). It was originally
isolated from bone but is found in other matrix-producing tissues such
as ligaments. It is also found in corticosteroid-secreting tissues such
as the adrenal cortex; in epithelia in, for example, the gut and skin,
and in platelets (289). It is a multidomain protein that is unique
among extracellular proteins in that it possesses an EF-hand pair
motif in the last of its three domains. The first two domains consist
of a variable acidic domain and a follistatin-like module. It shares
this domain organization with other proteins such as the brain protein SC1 (290) and the retinal protein QR1 (291). In addition to calcium, it has been shown to bind a number of extracellular proteins in
uitro, including collagen, plasminogen, and albumin (289). It is believed to be a modulator of cell-matrix interactions.
The structure of the calcium-binding domain of SPARC was determined by X-ray crystallography to a resolution of 2.0 A (292). The
structure of the EF-hand pair resembled those found in intracellular
proteins but had several unique features. In the first E F hand, there
was a single residue insertion, which alters the fold of the calciumbinding loop such that a cis-peptide bond is formed and a backbone
carbonyl acts as a calcium ligand. In the second EF hand, a disulfide
bond is found, which may act to stabilize the calcium-binding loop.
Finally, rather than having the hydrophobic interior of the EF-hand
pair available for binding of target proteins, an amphiphilic aminoterminal helix binds in the cleft. The structure of the calcium-binding
domain together with the second follistatin-like domain has also been
determined (293).
The effect of calcium binding on the structure of SPARC has not
been determined at molecular resolution. However, what is known
485
about the calcium binding is that the EF-hand pair contains one highaffinity and one low-affinity site (294).Calcium binding induces a conformational change resulting in an increase in helical structure as
monitored by circular dichroism (295). Curiously, this apparent increase in regular structure does not alter the stability of the protein
as measured by chemical denaturation.
V. Summary
In this review of calcium-binding proteins, we have not attempted

to give an exhaustive summary of all calcium-binding proteins, but
have considered some of the best examples, especially those for which
high-resolution structural data are available. However, even this limited subset of the hundreds of calcium-binding proteins identified to
date demonstrates the myriad ways in which nature uses calcium to
regulate biological processes. It also demonstrates how finely tuned
each protein is with respect to its environmental parameters such as
calcium concentration as well as rates and magnitudes of changes in
thia concentration. It also shows the many ways in which the calciumbinding proteins are adapted to the particular target, whether it be a
single protein, many proteins, membranes, or simply calcium itself.
Despite the wealth of information available to us, much is left to learn
about the functions and mechanisms of the calcium-binding proteins.
Readers interested in further information are encouraged to explore
the articles and reviews listed in the references.
REFERENCES
1. Forsen, S.; Kordel, J . In Bioinorganic Chemistry; Bertini, H. B. C . I., Lippard,
J. S., and Valentine, J. S., Eds.; University Science Books, 1994.
2. Lincoln, S. F.; Merbach, A. E. In Advances in Inorganic Chemistry; Sykes,
G. A,, Ed.; Academic Press: New York, 1995, 1.
3. Johansson, C.; Brodin, P.; Grundstrom, T.; Forsen, S.; Drakenberg, T. Eur. J .
Biochem. 1991,202,1283.
4. Johansson, C. Thesis; Lund University, Lund, 1993.
5. Guidebook to the Calcium Binding Proteins; Celio, M. R., Ed.; Oxford University
Press: New York, 1996.
6. Kretsinger, R. M.; Nockolds, C . B. J. B i d . Chem. 1973, 248, 3313.
7. Nakayama, S.; Moncrief, N. D.; Kretsinger, R. H. J. Mol. Euol. 1992,34, 416.
8. Holmes, M. A,; Stenkamp, R. E. J. Mol. Blol. 1991, 220, 723.
9. Mathews, F. S.; Bethge, P. H.; Czenvinski, E. W. J. Biol. Chem. 1979,254, 1699.
486
FINN AND DRAKENBERG
10. Wendt, B.; Hofmann, T.; Martin, S. R.; Bayley, P.; Brodin, P.; Grundstrom, T.;
Thulin, E.; Linse, S.; Forsen, S. Eur. J . Biochem. 1988, 175, 439.
11. Kawasaki, H.; Kretsinger, R. H. Calcium-Binding Proteins. Academic Press:
New York, 1994.
12. Cohen, P.; Klee, C. B. In Molecular Aspects of Cellular Regulation; Klee, C. B.,
Ed.; Elsevier: Amsterdam, 1988.
13. Babu, Y. S.; Sack, J. S.; Greenbough, T. C.; Bugg, C. E.; Means, A. R.; Cook,
W. J. Nature 1985,315, 37.
14. Babu, Y. S.; Bugg, C. E.; Cook, W. J. J. Mol. Biol. 1988,204, 191.
15. Seaton, B. A,; Head, J . F.; Richards, F. M. Biochemistry 1985,24, 6740.
16. Heidorn, D. B.; Trewhella, J . Biochemistry 1988,27, 909.
17. Chattopadhyaya, R.; Meador, W. E.; Means, A. R.; Quiocho, F. A. J. Mol. Biol.
1992,228, 1177.
18. Taylor, D. A.; Sack, 3. S.; Maune, J . F.; Beckingham, K.; Quiocho, F. A. J. Biol.
Chem. 1991,266, 21375.
19. Rao, S. T.; Wu, S.; Satyshur, K. A.; Ling, K. Y.; Kung, C.; Sundaralingam, M.
Protein Sci. 1993, 2, 436.
20. Ikura, M.; Spera, S.; Barbato, G.; Kay, L. E.; Krinks, M.; Bax, A. Biochemistry
1991,30, 9216.
21. Barbato, G.; Ikura, M.; Kay, L. E.; Pastor, R. W.; Bax, A. Biochemistry 1992,
31, 5269.
22. ONeil, K. T.; DeGrado, W. F. Trends Biochem. Sci. 1990, 15, 59.
23. Meador, W. E.; Means, A. R.; Quiocho, F. A. Science 1992,257, 1251.
24. Meador, W. E.; Means, A. R.; Quiocho, F. A. Science 1993,262, 1718.
25. Ikura, M.; Clore, G. M.; Gronenborn, A. M.; Zhu, G.; Klee, C. B.; Bax, A. Science
1992,256, 632.
26. Kemp, B. E.; Pearson, R. B.; Guerriero, J. V.; Bagchi, I. C.; Means, A. R. J . Biol.
Chem. 1987,262, 2542.
27. Pearson, R. B.; Wettenhall, R. E. H.; Means, A. R.; Hartshorne, D. J.; Kemp,
B. E. Science 1988,241, 970.
28. Goldberg, J.; Nairn, A. C.; Kuriyan, J. Cell 1996, 84, 875.
29. Liu, Y.; Storm, D. R. Trends Pharmacol. Sci. 1990, 11, 107.
30. Gamby, C.; Waage, M. C.; Allen, R. G.; Baizer, L. J. Biol. Chem. 1996,271, 26698.
31. Strynadka, N. C. J.; James, M. N. G. Proteins: Struct. Func. Genet. 1988,3, 1.
32. Herzberg, 0.;James, M. N. G. Nature 1985,313, 653.
33. Herzberg, 0.;James, M. N. G. J. Mol. Biol. 1988, 203, 761.
34. Zhang, M.; Tanaka, T.; Ikura, M. Nut. Struct. Biol. 1995,2, 758.
35. Kubinowa, H.; Tjandra, N.; Grzesiek, S.; Ren, S.; Klee, C. B.; Bax, A. Nut. Struct.
Biol. 1995, 2, 768.
36. Finn, B. E.; Forsen, S. Structure 1995,3, 7.
37. Finn, B. E.; Evenas, J.; Drakenberg, T.; Waltho, J. P.; Thulin, E.; Forsen, S. Nut.
Struct. Biol. 1996,2, 777.
38. Herzberg, 0.;Mvult, J.; James, N. G. J. B i d . Chem. 1986,261, 2638.
39. Slupsky, C. M.; Sykes, B. D. Biochemistry 1995,34, 15953.
40. Findlay, W. A.; Sonnichsen, F. D.; Sykes, B. D. J . Biol. Chem. 1994,269, 6773.
41. Gagne, S. M.; Tsuda, S.; Li, M. X.; Smillie, L. B.; Sykes, B. D. Nut. Struct. Biol.
1995, 2, 784.
42. McKay, R. T.; Tripet, B. P.; Hodges, R. S.; Sykes, B. D. J. Biol. Chem. 1997,
272, 28494.
487
43. Li, M. X.; Gagne, S. M.; Spyracopoulos, L.; Kloks, C. P.; Audette, G.; Chandra, M.;
Solaro, R. J.; Smillie, L. B.; Sykes, B. D. Biochemistry 1997, 36, 12519.
44. Gagne, S.; Li, M.; Sykes, B. D. Bioch,emistry 1997, 36, 4386.
45. Beckingham, K. J. Biol. Chem. 1991,266, 6027.
46. Evanas, J.; Thulin, E.; Malmendal, A,; Forsen, S.; Carlstrom, G. Biochemistry
1997,36, 3448.
47. James, P.; Vorherr, T.; Thulin, E.; Forsen, S.; Carafoli, E. FEBS Lett. 1991,278,
155.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
Szebenyi, D. M. E.; Obendorf, S. K.; Moffat, K. Nature 1981,294, 327.

Szebenyi, D. M. E.; Moffat, K. J. Biol. Chem. 1986,261, 8761.
Svensson, L. A.; Thulin, E.; Forsen, S. J. Mol. Biol. 1992, 223, 601.
Kordel, J.; Skelton, N. J.; Akke, M.; Chazin, W. J . J . Mol. Biol. 1993, 231, 711.
Skelton, N. J.; Kordel, J.; Akke, M.; Forsen, S.; Chazin, W. J . Nut. Struct. Biol.
1994, 1, 239.
Skelton, J . N.; Kordel, J.; Chazin, W. J. J . Mol. Biol. 1995,249, 441.
Akke, M.; Forsen, S.; Chazin, W. J . J. Mol. Biol., in press (1995).
Akke, M.; Drakenberg, T.; Chazin, W. J . Biochemistry 1992,31, 1011.
Linse, S.; Teleman, 0.;Drakenberg, T. Biochemistry 1990,29, 5925.
Skelton, N. J.; Kordel, J.; Akke, M.; Chazin, W. J . J. Mol. Biol. 1992, 227, 1100.
Akke, M.; Skelton, N. J.; Kordel, J.; Palmer, A. D.; Chazin, W. J . Biochemistry
1993,32, 9832.
Kordel, J.; Skelton, N. J.; Akke, M.; Palmer, A. D.; Chazin, W. J. Biochemistry
1992,31, 4856.
Chazin, W. J.; Kordel, J.; Thulin, E.; Hofmann, T.; Drakenberg, T.; Forsen, S.
Biochemistry 1989,28, 8646.
Chazin, W. E.; Kordel, J.; Drakenberg, T.; Thulin, E.; Brodin, P.; Grundstrom, T.;
Forsen, S. Proc. Nutl. Acud. Sci. U.S.A. 1989, 86, 2195.
Linse, S.; Forsen, S. Adu. Second Messenger Phosphoprotein Res. 1995, 30, 89.
Moore, B. W. Biochem. Biophys. Res. Commun. 1965, 19, 739.
Schafer, B. W.; Wicki, R.; Engelkamp, D.; Mattei, M. G.; Heizmann, C. W. Genomics 1995,25, 638.
Zimmer, D. B.; Cornwall, E. H.; Landar, A,; Song, W. Bruin Res. Bull. 1995, 37,
417.
66. Barger, S. W.; Wolchok, S. R.; Van Eldik, L. J. Biochim. Biophys. Actu 1992,
1160, 105.
67. Landar, A,; Hall, T. L.; Cornwall, E. H.; Correia, J . J.; Drohat, A. C.; Weber,
D. J.; Zimmer, D. B. Biochim. Biophys. Actu 1997, 1343, 117.
68. Kilby, P. M.; Van Eldik, L. J.; Roberts, G. C. K. Structure 1996,4, 1041.
69. Drohat, A. C.; Amburgey, J . C.; Abildgaard, F.; Stanch, M. R.; Baldisseri, D.;
Weber, D. J. Biochemistry 1996,35, 11577.
70. Groves, P.; Finn, B. E.; Kuznicki, J.; Forsen, S. FEBS Lett. 1998,421, 175.
71. Matsumura, H.; Shiba, T.; Inoue, T.; Harada, S.; Kai, Y. Structure 1998, 6, 233.
72. Drohat, A. C.; Baldisseri, D. M.; Rustandi, R. R.; Weber, D. J. Biochemistry 1998,
37, 2729.
73. Smith, S.; Shaw, G. S. Structure 1998, 6, 211.
74. Kilby, P. M.; Van Eldik, L. J.; Roberts, G. C. K. Prot. Sci. 1997, 6, 2494.
75. Rustandi, R. R.; Drohat, A. C.; Baldisseri, D. M.; Wilder, P. T.; Weber, D. J. Biochemistry 1998,37, 1951.
76. Kuznicki, J ; Filipek, A. Biochem. J. 1987, 247, 663.
488
FINN AND DRAKENBERG
77. Potts, B. C.; Smith, J.; Akke, M.; Macke, T. J.; Okazaki, K.; Hidaka, H.; Case,
D. A,; Chazin, W. J. Nut. Struct. Biol. 1996,2, 790.
78. Sastry, M.; Ketchum, R. R.; Crescenzi, 0.;
Weber, C.; Lubienski, M. J.; Hikada,
H., Chazin, W. J. Structure 1998, 6, 223.
79. Filipek, A,; Kuznicki, J. Actu Biochim. Pol. 1993,40, 171.
80. Pedrocchi, M.; Schafer, B. W.; Durussel, I.; Cox, J. A.; Heizmann, C. W. Biochemistry 1994,33, 6732.
81. Filipek, A,; Wojda, U.Biochem. J. 1993,320, 585.
82. Filipek, A,; Kuznicki, J. J . Neurochem. 1998, 70, 1793.
83. Ames, J. B.; Porumb, T.; Tanaka, T.; Ikura, M.; Stryer, L. J. Biol. Chem. 1995,
270, 4526.
84. Ames, J. B.; Tanaka, T.; Ikura, M.; Stryer, L. J. Biol. Chem. 1996,270, 30909.
85. Flaherty, K. M.; Zozulya, S.; Stryer, L.; McKay, D. B. Cell 1993, 75.
86. Tanaka, T.; Ames, J. B.; Harvey, T. S.; Stryer, L.; Ikura, M. Nature 1995,376, 444.
87. Ames, J. B.; Ishima, R.; Tanaka, T.; Gordon, J. I.; Stryer, L.; Ikura, M. Nature
1997,389, 198.
88. Chen, C.-K.; Inglese, J.; Lefkowitz, R. J.; Hurley, J. B. J. Biol. Chem. 1994,
2780, 18060.
89. Dizhoor, A. M.; Chen, C.-K.; Olshevskaya, E.; Sinelnikova, V. V.; Phillipov, P.;
Hurley, J. B. Science 1993, 259, 829.
90. Liemann, S.; Lewit-Bentley, A. Structure 1995, 3, 233.
91. Swairjo, M. A.; Roberts, M. F.; Campos, M.-B.; Dedman, J. R.; Seaton, B. A. Biochemistry 1994,33, 10944.
92. Raynal, P.; Pollard, H. B. Biophys. Biochim. Actu 1994, 1197, 63.
93. Davidson, F. F.; Lister, M. D.; Dennis, E. A. J. Biol. Chem. 1990,265, 5602.
94. Russo Mane, F. In The Annexins; Moss, S. E., Ed.; Portland Press: London,
1992,35-46.
95. Andree, H. A. M.; Stuart, M. C. A.; Hermens, W. T.; Reutelingsperger, C. P. M.;
Hemker, H. C.; Frederick, P. M.; Willems, G. M. J. Biol. Chem. 1992,267, 17907.
96. Pollard, H. B.; Guy, H. R.; Arispe, N.; de la Fuente, M.; Lee, G.; Rojas, E. M.;
Pollard, J. R.; Srivastava, M.; Zhang Keck, Z. Y.; Merezhinskaya, N.; Caohuy, H.;
Burns, A. L.; Rojas, E. Biophys. J. 1992,62, 15.
97. Demange, P.; Voges, D.; Benz, J.; Liemann, S.; Gottig, P.; Berendes, R.; Burger,
A.; Huber, R. Trends Biochem. 1994, 19, 272.
98. Huber, R.; Romisch, J.; Pbques, E.-P. EMBO J . 1990,9, 3867.
99. Huber, R.; Schneider, M.; Mayr, I.; Romisch, J.; Pbques, E.-P. FEBS Lett. 1990,
275, 15.
100. Lewit-Bentley, A.; Morere, S.; Huber, R.; Bodo, G. Eur. J. Biochem. 1992,210, 73.
101. Sopkova, J.; Renouard, M.; Lewit-Bentley, A. J. Mol. Biol. 1993,234, 816.
102. Bewley, M. C.; Boustead, C. M.; Walker, J. H.; Waller, D. A,; Huber, R. Biochernistry 1993,32, 3923.
103. Concha, N. 0.; Head, J. F.; Kaetzel, M. A,; Dedman, J. R.; Seaton, B. A. Science
1993,261, 1321.
104. Swairjo, M. A.; Concha, N. 0.;
Kaetzel, M. A.; Dedman, J. R.; Seaton, B. A. Nut.
Struct. Biol. 1995,2, 968.
205. Weng, X.; Luecke, H.; Song, I. S.; Kang, D. K.; Kim, S.-H.; Huber, R. Prot. Sci.
1993,2, 448.
106. Shadle, P. J.; Gerke, V.; Weber, K. J. Biol. Chem. 1986, 260, 16354.
107. Schlaepfer, D. D.; Mehlman, T.; Burgess, W. H.; Haigler, H. T. Proc. Natl. Acad.
Sci. U.S.A. 1987, 84, 6078.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
489
Schlaepfer, D. D.; Haigler, H. T. J . Biol. Chem. 1987,262, 6931.

Haigler, H. T.; Schlaepfer, D. D.; Burgess, W. T. J . Biol. Chem. 1987,262, 6921.
Bazzi, M. D.; Nelsestuen, G. L. Biochemistry 1991, 30, 971.
Tait, J. F.; Gibson, D.; Fujikawa, K. J . Biol. Chem. 1989, 264, 7944.
Blackwood, R. A.; Ernst, J. D. Biochem. J . 1990,266, 195.
Follenius-Wund, A.; Piemont, E.; Freyssinet, J.-M.; Gerard, D.; Pigault, C. Biochem. Biophys. Res. Commun. 1997,234, 111.
Voges, D.; Berendes, R.; Burger, A.; Demange, P.; Baumeister, W.; Huber, R. J.
Mol. Biol. 1994,238, 199.
Pigault, C.; Follenius-Wund, A,; Schmutz, M.; Freyssinet, J.-M.; Brisson, A. J.
Mol. Biol. 1994,236, 199.
Bazzi, M. D.; Nelsestuen, G. L. Biochemistry 1990,29, 7624.
Orr,J. W.; Newton, A. C. Biochemistry 1992,31, 4660.
Newton, A. C.; Keranen, L. M. Biochemistry 1994,33, 6651.
Davletov, B. A,; Suhof, T. A. J . Biol. Chem. 1993, 268, 26386.
Chapman, E. R.; Jahn, R. J. Biol. Chem. 1994,269, 5735.
Fukuda, M.; Kojima, T.; Mikoshiba, K. J. Biol. Chem. 1996,271, 8430.
Li, L.; Darden, T.; Foley, C.; Hiskey, R.; Pedersen, L. Prot. Sci. 1996,4, 2341.
Ullrich, B.; Li, C.; Zhang, J. Z.; McMahon, H.; Andersson, R. G. W.; Geppert, M.;
Sudhof, T. C. Neuron 1994,13, 1281.
Clark, J . D.; Lin, L.-L.; Kriz, R. W.; Ramesha, C. S.; Sultzman, L. A,; Lin, A. Y.;
Milona, N.; Knopf, J. L. Cell 1991, 65, 1043.
Nafelski, E. A,; Sultzman, L. A,; Martin, D. M.; Kriz, R. W.; Towler, P. S.; Knopf,
J. L.; Clark, J. D. J. Biol. Chem. 1994,269, 18239.
Kramer, R. M.; Sharp, J. D. FEBS Lett. 1997,410, 49.
Takenawa, T.; Homma, Y.; Emori, Y. Methods Enzymol. 1991, 197.
Rhee, S. G.; Choi, D. J. J. Biol. Chem. 1993, 267, 12393.
Rebecchi, M.; Boguslavsky, V.; Boguslavsky, L.; McLaughlin, S. Biochemistry
1992,31, 12748.
130. Ellis, M. V.; Carne, A,; Katan, M. Eur. J . Biochem. 1993,213, 339.
131. Yagisawa, H.; Hirata, M.; Kanematsu, T.; Watanabe, Y.; Ozaki, S.; Sakuma, K.;
Tanaka, H.; Yabutan, N.; Kamata, H.; Hirata, H.; Nojima, H. J. Biol. Chem. 1994,
269, 20179.
132. Cheng, H.-F.; Jiang, M.-J.; Chen, C.-L.; Liu, S.-M.; Wong, L.-P.; Lomasney, J. W.;
Kmg, K. J. Biol. Chem. 1996, 270, 5495.
133. Essen, L.-0.; Perisic, 0.; Cheung, R.; Katan, M.; Williams, R. L. Nature 1996,
380, 595.
134. Grobler, J . A,; Essen, L.-0.; Williams, R. L.; Hurley, J. H.; Nut. S t r u t . Biol. 1996,
3, 788.
135. Essen, L.-0.; Perisic, 0.; Lynch, D. E.; Katan, M.; Williams, R. L. Biochemistry
1997.36, 2753.
136. Nalefski E. A,; Falke, J. J . Prot. Sci. 1996, 5, 2375.
137. Nalefski, E. A,; Slazas, M. M.; Falke, J. J. Biochemistry 1997, 36, 12011.
138. Falke, J. J.; Drake, S. K.; Hazard, A. L.; Peerse, 0. B. Q.Reu. Biophys. 1994,
27, 219.
139. Sutton, R. B.; Davletov, B. A,; Berghuis, A. M.; Sudhof, T. C.; Sprang, S. R. Cell
1996, 80, 929.
140. Cifuentes, M. E.; Honkanen, L.; Rebecchi, M. J . J . Biol. Chem. 1993,268, 1158611593.
141. Robertson, P. J.; Hiskey, R. G.; Koehler, K. A. J . Biol. Chem. 1978,253, 5880.
490
FINN AND DRAKENBERG
142. Colpitts T. L.; Castellino, F. J. Znt. J. Peptide Protein Res. 1993, 41, 567.
143. Atkinson, R. A,; Evans, J. S.; Hauschka, P. V.; Levine, B. A,; Meats, R.; Triffitt,
J. T.; Virdi, A. S.; Williams, R. J. P. Eur. J. Biochem. 1995,232, 515.
144. Hauschka, P. V.; Lian, J. B.; Cole, D. E.; Gundberg, C. M. Physiol. Reu. 1989,
69, 990.
145. Furie, B.; Furie, B. C. Cell 1988,53, 505.
146. Stenffo,J.; Dahlback, B. In The Molecular Basis of Blood Diseases; Stamatoyan-
147.
148.
149.
150.
nopoulus, G., Nienhuis, A. W., Majerus, P. W., and Varmus, H., Eds.; Saunders:
Philadelphia, 1994.
Stenflo, J.; Garnot, P. J. Biol. Chem. 1972, 247, 8160.
Stenflo, J. J . Biol. Chem. 1973, 248, 6325.
Nelsestuen, G. L.; Suttie, J. W. Biochemistry 1972, 20, 351.
Stenflo, J.; Fernlund, P.; Egan, W.; Roepstorff, P. Proc. Natl. Acad. Sci. U.S.A.
1974, 71, 2730.
151.
152.
153.
154.
155.
156.
157.
158.
159.
160.
161.
Nelsestuen, G. L.; Zytkonics, T. H.; Howard, J. B. J. Biol. Chem. 1974,249, 6347.

Henriksen, R. A.; Jackson, C . M. Arch. Biochen. Biophys. 1975,170, 149.
Nelsestuen, G. L. J. Biol. Chem. 1976,25, 5648.
Prendergast, F. G.; Mann, K. G. J . Biol. Chem. 1977,252, 840.
Marsh, H. C.; Robertson, P. J.; Scott, M. E.; Koehler, K. A.; Hiskey, R. G. J . Biol.
Chem. 1979,254, 10268.
Strickland, D.; Castellino, F. J. Arch. Biochem. Biophys. 1980, 199, 61.
Amphlett, G. W.; Kisiel, W.; Catellino, F. J. Biochemistry 1981,20, 2156.
Bajaj, P. S. J. Biol. Chem. 1982, 257, 4127.
Jackson, C. M. In Current Advances in Vitamin K Research; Suttie, J. W., Ed.;
Elsevier: New York, 1988.
Ohlin, A.-K.; Bjork, I.; Stenflo, J . Biochemistry 1990, 29, 644.
Monroe, D. M.; Deerfield, D. W. I.; Olson, D. L.; Stewart, T. N.; Treanor, R. E.;
Roberts, H. R.; Hiskey, R. G.; Pedersen, L. G. Blood Coagul. Fibrinolysis 1990,
1, 633.
162. Persson, E.; Bjork, I.; Stenflo, J. J. Biol. Chem. 1991,266, 2444.
163. Sabharwal, A. K.; Padmanabhan, K.; Tulinsky, A.; Mathur, A.; Gorka, J.; Bajaj,
P. J. Biol. Chem. 1997,272, 22037.
164. Jacobs, M.; Freedman, S. J.; Furie, B. C.; Furie, B. J . Biol. Chem. 1994,269, 25494.
165. Colpitts, T. L.; Castellino, F. J. Biochemistry 1994, 33, 3501.
166. Colpitts, T. L.; Prorok, M.; Castellino, F. J. Biochemistry 1995,34, 2424.
167. Freedman, S. J.; Furie, B. C.; Furie, B.; Baleja, J. D. J. Mol. Biol. 1996,270, 7980.
168. Sunnerhagen, M.; Olah, G.; Stenflo, J.; Drakenberg, T.; Trewhella, J . Biochemistry
1996,35, 1547.
169. Soriano-Garcia, M.; Park, C. H.; Tulinsky, A.; Ravichandran, K. G.; SkrzypczakJankun, E. Biochemistry 1989,28, 6805.
170. Banner, D. W.; DArcy, A.; Chene, C.; Winkler, F. K.; Guha, A,; Konigsberg,
W. H.; Nemerson, Y.; Kirchhofer, D. Nature 1996,380, 41.
171. Freedman, S. J.;Furie, B. C . ; Furie, B.; Baleja, J. D. Biochemistry 1995,34, 12126.
172. Li, L.; Darden, T. A.; Freedman, S. J.; Furie, B. C.; Furie, B.; Baleja, J. D.; Smith,
H.; Hiskey, R. G.; Pedersen, L. G. Biochemistry 1997,36, 2132.
173. Sunnerhagen, M.; Forsen, S.; Hoffren, A.-M.; Drakenberg, T.; Teleman, 0.;Stenflo, J. Nut. Struct. Biol. 1996, 2, 504.
174. Arni, R. K.; Padmanabhan, K.; Padmanabhan, K. P.; Wu, T. P.; Tylinsli, A. Chem.
Phys. Lipid 1994, 67/68, 59.
175. Zhang L.; Castellino, F. J. J. Bid. Chem. 1994,269, 3590.
CALCIUM-BINDINGPROTEINS
49 1
176. Olivera, B. M.; Rivier, J.; Clark, C.; Ramilo, C.; Corpuz, G.; Bogadie, F.; Woodward, s.;Hillyard, D.; Cruz, L. J . Science 1990, 249, 257.
177. Olivera, B. M.; Rivier, J.; Scott, J. K.; Hillyard, D. R.; Cruz, L. J. J . Biol. Chem.
1991,266, 22067.
178. Cruz, L. J.; Ramilo, C. A.; Corpuz, G. P.; Olivera, B. M. Biol. Bull. 1992,183, 159.
179. McIntosh, J . M.; Olivera, B. M.; Cruz, L. J.; Gray, W. R. J . Biol. Chem. 1984,
259, 14343.
180. Haak, J . A.; Rivier, J.; Parks, T. N.; Mena, E. E.; Cruz, L. J.; Olivera, B. M. J .
Biol. Chem. 1990, 265, 6025.
181. Chandler, P.; Pennington, M.; Maccacchini, M.-L.; Nashed, N. T.; Skolnick, P. J.
Biol. Chem. 1993,268, 17173.
182. Zhou, L.-M.; Szendrei, G. I.; Fossom, L. H.; Maccecchini, M.-L.; Skolnick, P.;
Otvos, L. J. J . Neurochem. 1996, 66, 620.
183. Skjaerbaek, N.; Nielsen, K. J.; Lewis, R. J.; Alewood, P.; Craik, D. J . J. Biol. Chem.
1997,272, 2291.
184. Chiancone, E.; Drakenberg, T.; Teleman, 0.; Forsen, S. J. Mol. Biol. 1985, 185,
201.
185. Rigby, A. C.; Baleja, J. D.; Furie, B. C.; Furie, B. Biochemistry 1997,36, 6906.
186. Rigby, A. C.; Baleja, J. D.; Li, L.; Pedersen, L. G.; Furie, B. C.; Furie, B. Biochemistry 1997, 36, 15677.
187. Soriano-Garcia, M.; Padamanbhan, K.; deVos, A. M.; Tulinsky, A. Biochemistry
1992,31, 2554.
188. Campbell, I. D.; Bork, P. Curr. Opin. Struct. Biol. 1993, 3, 385.
189. Stenflo, J.; Lundwall, A.; Dahlback, B. Proc. Nutl. Acad. Sci. U.S.A. 1987,84, 368.
190. Gronke, R. S.; Van Dusen, W. J.; Garsky, V. M.; Jacobs, J . W.; Sardana, M. K.;
Stern, A. M.; Friedman, P. A. Proc. Nutl. Acud. Sci. U.S.A. 1989,86, 3609.
191. Pereira, L.; DAlessio, M.; Ramirez, F.; Lynch, J. R.; Sykes, B.; Pangilinan, T.;
Bonadio, J . Hum. Mol. Genet. 1993,2, 961.
192. Corson, G. M.; Chalberg, S. C.; Dietz, H. C.; Charbonneau, N. L.; Sakai, L. Y.
Genomics 1993, 17, 476.
193. Glanville, R. W.; Qian, R.-Q.; McClure, D. W.; Maslen, C. L. J . Biol. Chem. 1994,
269, 26630.
194. Argraves, W. S.; Tran, H.; Burgess, W. H.; Dickerson, K. Cell Biol. 1990, I l l , 3155.
195. Pan, T. C.; Sasaki, T.; Zhabg, R. Z.; Fassler, R.; Timpl, R.; Chu, M. L. J. Cell. Biol.
1993,123, 1269.
196. Tran, M. J.; Mattei, M.; Godyna, S.; Argraves, W. S. Matrix Biol. 1997, 15, 479.
197. Tran, H.; Van Dusen, W. J.; Argraves, W. S. J . Biol. Chem. 1997,272, 22600.
198. Wharton, K. A,; Johansen, K. M.; Xu, T.; Artavanis-Tsakonas, S. Cell 1985, 43,
567.
199. Morita, T.; Isaacs, B. S.; Esmon, C. T.; Johnson, A. E. J. Biol. Chem. 1984, 259,
5698.
200. Morita, T.; Kisiel, W. Biochenz. Biophys. Res. Commun. 1985, 130, 841.
201. Sugo, T.; Bjork, I.; Holmgren, A,; Stenflo, J . J. Biol. Chem. 1984,259, 5705.
202. Sugo, T.; Dahlback, B.; Stenflo, J. J. Biol. Chem. 1986, 261, 5116.
203. Esmon, N. L.; Debault, L. E.; Esmon, C. T. J . Biol. Chem. 1983,258, 5548.
204. Viliers, C. L.; Arlaud, G . J.; Painter, R. H.; Colomb, M. G. FEBS Lett. 1980, 117,
289.
205. Ohlin A.-K.; Stenflo, J. J. Biol. Chem. 1987, 262, 13798.
206. Ohlin, A.-K.; Linse, S.; Stenflo, J . J . Biol. Chem. 1988, 263, 7411.
492
FINN AND DRAKENBERG
207. Ohlin, A.-K.; Laudes, G.; Bourdon, P.; Oppenheimer, C.; Wydro, R.; Stenflo, J. J.
Biol. Chem. 1988,263, 19240.
208. Persson, E.; Selander, M.; Linse, S.; Drakenberg, T.; Stenflo, J. J. Biol. Chem.
1989,264, 16897.
209. Handford, P. A,; Baron, M.; Mayhew, M.; Willis, A.; Beesley, T.; Brownlee, G. G.;
Campbell, I. D. EMBO J. 1990, 9, 475.
210. Selander Sunnerhagen, M.; Persson, E.; Dahlqvist, I.; Drakenberg, T.; Stenflo, J.;
Mayhew, M.; Robin, M.; Hanford, P.; Tilley, J. W.; Campbell, I. D.; Brownlee,
G. G. J. Biol. Chem. 1993,268, 23339.
211. Handford, P.; Downing, A. K.; Rao, Z.; Hewett, D. R.; Sykes, B. C.; Kielty, C. M.
J. Biol. Chem. 1995,270, 6751.
212. Wu, Y.-S.; Bevilacqua, V. L. H.; Berg, J. M. Chem. Biol. 1995,2, 91.
213. Astermark, J.; Bjork, I.; Ohlin, A.-K.; Stenflo, J. J. Biol. Chem. 1991,266, 2430.
214. Valcarce, C.; Selander Sunnerhagen, M.; Tamlitz, A.-M.; Drakenberg, T.; Bjork,
I.; Stenflo, J. J. Biol. Chem. 1993, 268, 26673.
215. Knott, V.; Downing, A. K.; Cardy, C. M.; Handford, P. J. Mol. Biol. 1996,255, 22.
216. Dahlback, B.; Hildebrand, B.; Linse, S. J. Biol. Chem. 1990,265, 18481.
21 7. Stenberg, Y.; Linse, S.; Drakenberg, T.; Stenflo, J . J . B i d . Chem. 1997,272, 23255.
218. Stenberg, Y.; Julenius, K.; Dahlqvist, I.; Drakenberg, T.; Stenflo, J. Eur. J. Biochem. 1997,248, 163.
219. Rand, M. D.; Lindblom, A.; Carlson, J.; Villoutreix, B. 0.; Stenflo, J . Prot. Sci.
1997, 6, 2059.
220. Cooke, R. M.; Wilkinson, A. J.; Baron, M.; Pastore, A.; Tappin, M. J.; Campbell,
I. D.; Gregory, H.; Sheard, B. Nature 1987,327, 339.
221. Kohda D.; Inagaki, F. J. Biochem. 1988,103, 554.
222. Campbell, I. D.; Cooke, R. M.; Baron, M.; Harvey, T. S.; Tappin, M. J. Prog.
Growth Fac. Res. 1989, 1, 13.
223. Tappin, M. J.; Cooke, R. M.; Fitton, J . E.; Campbell, I. D. Eur. J. Biochem. 1989,
179, 629.
224. Kline, T. P.; Brown, F. K.; Brown, S. C.; Jeffs, P. W.; Kopple, K. D.; Mueller, L.
225. Cooke, R. M.; Tappin, M. J.; Campbell, I. D.; Dohda, D.; Miyake, T.; Fuwa, T.;
Miyazawa, T.; Inagaki, F. Eur. J . Biochem. 1990,193, 807.
226. Harvey, T. S.; Wilkinson, A. J.;Tappin, M. J.; Cooke, R. M.; Campbell, I. D. Eur.
J. Biochem. 1991, 198, 555.
227. Huang, L. H.; Cheng, H.; Pardi, A.; Tam, J . P.; Sweeney, W. V. Biochemistry 1991,
30, 7402.
228. Baron, M.; Norman, D. G.; Harvey, T. S.; Handford, P. A.; Mayhew, M.; Tse,
A. G. D.; Brownlee, G . G.; Campbell, I. D. Prot. Sci. 1992, 1, 81.
229. Ullner, M.; Selander, M.; Persson, E.; Stenflo, J.; Drakenberg, T.; Teleman, J.
230. Selander Sunnerhagen, M.; Ullner, M.; Persson, E.; Teleman, 0.; Stenflo, J.;
Drakenberg, T. J. Biol. Chem. 1992,267, 19642.
231. Moy, F. J.; Li, Y.-C.; Rauenbuchler, P.; Winkler, M. E.; Scheraga, H. A.; Montelione, G. T. Biochemistry 1993,32, 7334.
232. Freedman, S. J.; Sanford, D. G.; Bachovchin, W. W.; Furie, B. C.; Baleja, J. D.;
Furie, B. Biochemistry 1996,35, 13733.
233. Graves, B. J.; Crowther, R. L.; Chandran, C.; Rumberger, J. M.; Li, S.; Huang,
K.-S.; Presky, D. H.; Familletti, P. C.; Wolitzky, B. A.; Burns, D. K. Nature 1994,
367, 532.
493
234. Padmanabhan, K.; Padmanabhan, K. P.; Tulinsky, A,; Park, C. H.; Bode, W.;
Huber, R.; Blankenship, D. T.; Cardin, A. D.; Ksiel, W. J . Mol. Biol. 1993,232,947.
235. Rao, Z.; Handford, P.; Mayhew, M.; Knott, V.; Brownlee, G. G.; Stuart, D. Cell
1995, 82, 131.
236. Banner, D. W. Thromb. Haemostasis 1997, 78, 512.
237. Downing, A. K.; Knott, V.; Werner, J. M.; Cardy, C. M.; Campbell, I. D.; Handford,
P. A. Cell, 1996, 85, 597.
238. Bode, W.; Schwager, P. FEBS Lett. 1975,56, 139.
239. Bode, W.; Schwager, P. J. Biol. Chem. 1975, 98, 693.
240. Persson, E.; Hogg, P. J.; Stenflo, J. J. Biol. Chem. 1993, 268, 22531.
241. Rezaie, A. R.; Neuenschwander, P. F.; Morrissey, J. H.; Esmon, C. T. J . Biol.
Chem. 1993,268, 8176.
242. Rezaie, A. R.; Mather, T.; Sussman, F.; Esmon, C. T. J. Bid. Chem. 1994, 269,
2151.
243. Wildgoose, P.; Foster, D.; Schiodt, J.; Wiberg, F. C.; Birktoft, J. J.; Petersen, L. C.
Biochemistry 1993, 32, 114.
244. Bajaj, S. P.; Sabhanval, A. K.; Gorka, J.; Birktoft, J. J. Proc. Natl. Acad. Sci.
U.S.A. 1992.89, 152.
245. Rezaie, A. R.; Esmon, N. L.; Esmon, C. T. J. Biol. Chem. 1992,267, 11701.
246. Sabharwal, A. K.; Birktoft, J. J.; Gorka, J.; Wildgoose, P.; Petersen, L. C.; Bajaj,
S. P. J. B i d . Chem. 1996,270, 15523.
247. Freskghrd, P.-0.; Olsen, 0. H.; Persson, E. Prot. Sci. 1996,5, 1531.
248. Brittain, H. G.; Richardson, F. S.; Martin, R. B. J. Chem. Soc. (London) 1976,
98, 8255.
249. Neidhart D. L.; Petsko, G. A. Prot. Eng. 1988, 2, 271.
250. Bode, W.; Papamokos, E.; Musil, D. Eur. J. Biochem. 1987,166, 673.
251. Bott, R.; Ultsch, M.; Kossiakoff, A.; Graycar, T.; Katz, B.; Power, S. J . B i d . Chent.
1988,263, 7895.
252. Khatra, B. S,;Herries, D. G.; Brew, K. Eur. J. Biochem. 1974,44, 537.
253. Bell, J. E.; Beyer, T. A,; Hill, R. L. J. Biol. Chem. 1976,251, 3003.
254. Brew, K.; Grobler, J. A. In Advanced Dairy Chemistry; Fox, P., Ed.; Elsevier
Press: London, 1992, 1, 191.
255. Brew, K.; Vanamna, T. C.; Hill, R. L. J. Bid. Chem. 1967,242, 3747.
256. Brew, K.; Castellino, F. J.; Vanaman, T. C . ; Hill, R. L. J . Bid. Chem. 1979, 245,
4570.
257. Qasba, P. K.; Safaya, S. K. Nature 1984, 308, 377.
258. Hall, L.; Craig, R. K.; Edbrooke, M. R.; Campbell, P. N. Nucleic Acids Res. 1982,
10, 3503.
259. McKenzie, H. A,; Whitw, F. H. Adu. Protein Chem. 1991,41, 173.
260. Acharya, K. R.; Stuart, D. I.; Walker, N. P. C.; Lewis, M.; Philips, D. C. J. Mol.
B i d . 1989,208, 99.
261. Nitta, K.; Tsuge, H.; Sugai, S.; Shimazaki, K. FEBS Lett. 1987,223, 405.
262. Nitta, K.; Tsuge, H.; Shimazaki, K.; Sugai, S. Biol. Chem.. Hoppe-Seyler 1988,
369, 671.
263. Sugai, S.; Nitta, K.; Tsuge, H. Colloq. INSERM 1989, 179, 591.
264. Tsuge, H.; Ago, H.; Noma, M.; Nitta, K.; Sugai, S.; Miyano, M. J. Biochem. (Tokyo)
1992,111, 141.
265. Inaka, K.; Kuroki, R.; Kikuchi, M.; Matsushima, M. J . Biol. Chem. 1991, 266,
20666.
494
FINN AND DRAKENBERG
266. Aramini, J . M.; Drakenberg, T.; Hiraoki, T.; Ke, H.; Nitta, K.; Vogel, H. J. Biochemistry 1992, 31, 6761.
267. Aramini, J . M.; Hiraoki, T.; Grace, M. R.; Swaddle, T. W.; Chiancone, E.; Vogel,
H. J . Biochim. Biophys. Acta 1996,1293, 72.
268. Stuart, D. I.; Acharya, K. R.; Walker, N. P. C.; Smoth, S. G.; Lewis, M.; Philips,
D. C. Nature 1986,324, 84.
269. Acharya, K. R.; Ren, J.; Stuart, D. J.; Philips, D. C.; Fenna, R. E. J . Mol. Biol.
1991,221, 571.
270. Pike, A. C.; Brew, K.; Acharaya, K. R. Structure 1996,4, 691.
271. Harata, K.; Muraki, M. J. Biol. Chem. 1992,267, 1419.
272. Vanderheeren, G.; Hanssens, I.; Meijberg, W.; Van Aerschot, A. Biochemistry
1996,35, 16753.
273. Wu, L. C.; Schulman, B. A.; Peng, Z.-Y.; Kim, P. S. Biochemistry 1996,35, 859.
274. Anderson, P. J.; Brooks, C. L.; Berliner, L. J. Biochemistry 1997, 36, 11648.
275. Ptitsyn, 0. B. Curr. Opin. Struct. Biol. 1995, 5, 74.
276. Privalov, P. L. J. Mol. Biol. 1996, 258, 707.
277. Chothia C.; Jones, E. Y. Annu. Rev. Biochem. 1997,66, 823.
278. Takeichi, M. Annu. Rev. Biochem. 1990,59, 237.
279, Hatta, K.; Nose, A,; Nagafuchi, A.; Tackeichi, M. J. Cell Biol. 1988, 106, 873.
280. Overduin, M.; Harvey, T. S.; Bagby, S.; Tong, K. I.; Yau, P.; Takeichi, M.; Ikura,
M. Science 1995,267, 386.
281. Koch, A. W.; Pokutta, S.; Lustig, A,; Engel, J . Biochemistry 1997, 36, 7697.
282. Shapiro, L.; Fannon, A. M.; Kwong, P. D.; Thompson, A.; Lehman, M. S.; Grubel,
G.; Legrand, J.-F.; Als-Nielsen, J.; Colman, D. R.; Hendrickson, W. A. Nature

1996,374, 327.
283. Overduin, M.; Tong, K. I.; Kay, C. M.; Ikura, M. J . Biomol. NMR 1996, 7, 173.
284. Nagar, B.; Overduin, M.; Ikura, M.; Rini, J. M. Nature 1996,380, 360.
285. Pokutta, S.; Herrenknecht, K.; Kemler, R.; Engel, J . Eur. J. Biochem. 1994, 223,
1019.
286. Termine, J . D.; Kleinman, H. K.; Whitson, S. W.; Conn, K. M.; McGarvey, M. L.;
Martin, G. R. Cell 1981,26, 99.
287. Mason, I. J.; Taylor, A,; Williams, J. G.; Sage, H.; Hogan, B. L. M. EMBO J. 1986,
5, 1465.
288. Dziadek, M.; Paulsson, M.; Aumailley, M.; Timpl, R. Eur. J. Biochem. 1986, 161,
455.
289. Lane, T. F.; Sage, H. FASEB J. 1994,8, 163.
290. Johnston, I. G.; Paladino, T.; Gurd, J. W.; Brown, I. R. Neuron 1990,2, 165.
291. Guermah, M.; Crisanti, P.; Langier, D.; Dezelee, P.; Bidou, L.; Pessac, B.; Calothy,
G. Proc. NatL. Acad. Sci. U.S.A. 1991, 88, 4503.
292. Hohenester, E.; Maurer, P.; Hohenadi, C.; Timpl, R.; Jansonius, J. H.; Engel, J .
Nut. Struct. Biol. 1996, 3, 67.
293. Hohenester, E.; Maurer, P.; Timpl, R. EMBO J. 1997, 16, 3778.
294. Pottgiesser, J.; Maurer, P.; Mayer, U.; Nischt, R.; Mann, K.; Timpl, R.; Krieg, T.;
Engel, J. J. Mol. Biol. 1994, 238, 563.
295. Maurer, P.; Mayer, U.; Bruch, M.; Jeno, P.; Mann, K.; Landwehr, R.; Engel, J.;
Timpl, R. Eur. J . Biochem. 1992,205, 233.
ADVANCES IN INORGANIC CHEMISTRY, VOL
46
LEGHEMOGLOBIN: PROPERTIES AND REACTIONS

MICHAEL J. DAVIES,* CHRISTEL MATHIEUt, and ALAIN PUPPOt
The Heart Research Institute, Carnperdown, Sydney, New South Wales 2050, Australia
t Laboratoire de Biologie Vegetale et Microbiologie,CNRS URA 1114,
Universite de NiceSophia Antipolis, 06018 Nice Cedex 2, France
I. Introduction
11. Structure
111. Biological Localization
IV. Reactions with Different Molecules
A. Diatomic Gases
B. Other Molecules
V. Oxidation of Fe(I1) Leghemoglobin
VI. Oxidation of Fe(II1) Leghemoglobin
VII. Reduction of F e W ) = 0 Leghemoglobin
A. Reaction with H200
B. Reaction with Ascorbate
C. Reaction with Glutathione, Other Thiols, and Related Species
D. Reaction with Other Reductants
VIII. Reduction of Fe(II1) Leghemoglobin
A. Enzymatic
B. Nonenzymatic
IX. Reaction of Globin-Derived Radicals
A. Reactions with the Heme Ring
B. Reactions with Low-Molecular-Weight Species
C. Reactions with Other Leghemoglobin Molecules
D. Reactions with Membranes
References
I. Introduction
Plants release oxygen during photosynthesis and use it as the terminal electron acceptor in the respiratory chain of mitochondria.
There is now evidence that they contain a monomeric oxygen-binding
protein to coordinate and transport oxygen (I). The occurrence of
these myoglobin-like proteins in plants was first described by Kubo in
1939, who demonstrated that the nitrogen-fixing root nodules of soy495
Copyright 'c 1999 by Academic Press.
All Fights of reproduction in any form reserved.
0898-883R/99 $25.00
496
DAVIES, MATHIEU, AND PUPPO
bean and other plants contained a red-colored soluble heme protein

that could reversibly bind oxygen (2). The protein was named leghemoglobin (Lb) at an early stage by analogy to the mammalian protein.
The nodules from which the protein is extracted arise from the interaction between Rhizobium spp., a soil bacteria, and leguminous
plants (Fig. 1).More recent studies have demonstrated the presence
of similar structures in the nonleguminous plant Parasponia spp. (3).
The origin of the symbiosis in legumes has been reviewed (4),
and the
processes that occur during the development of the relationship are
reasonably well characterized (5-7). Hemoglobins and their parent
genes have also been recently identified in a number of nonnodulating plants and in phylogenetically diverse plant genera including monocots (8, 9). Moreover, a new hemoglobin gene has been
isolated recently from soybean (10)that is expressed in different
parts of the plant (including the nodules). This contrasts with the
expression of Lbs, which are active only in the infected cells of the
nodule. Thus, two different types of hemoglobin are known to be
present in plants, (i) a nonsymbiotic type that is widely distributed
among species and (ii) a symbiotic type that is induced only on
nodulation in (at least) two plant families (1). The genes that code
for these two types of hemoglobins are separated into four exons
by three introns: the first and third introns are in positions homologous to those of the two introns found in vertebrate a-globin and
/I-globin and myoglobin genes. It has been proposed, and seems
likely, that there is a common ancestral gene for plant and animal
hemoglobins (1, 11-13). This review will concentrate on the properties and reactions of the symbiotic type of plant protein, the leghemoglobins.
Until now, Lb has been the most extensively studied oxygen-carrying hemoprotein from plants. The main role of Lb in uiuo is now
widely established (though for many years this was hotly disputed)
as being to facilitate oxygen transport t o the vigorously respiring,
phosphorylating, nitrogen-fixing Rhizobium microsymbionts, at a low
and stable oxygen tension (around 10 nM in soybean nodules) (7, 12,
14, 15). Terminal oxidases having a K,,, for oxygen as low as 7 nM are
present in the microsymbionts and are able to use the Lb-delivered oxygen to support their respiration under the oxygen-limiting
conditions of the nodule (16).This low oxygen tension appears to be a
compromise between the requirement for oxygen by the respiring bacteria and the damage that oxygen can inflict on the nitrogen-fixing
complex; the latter is very readily inactivated at high oxygen levels
(12,15).
497
FIG.1. Nitrogen-fixing root nodules on a soybean plant.
II. Structure
Early studies by Ellfolk demonstrated that there were several Lb

isomers in soybean nodules (17). Ammonium sulfate fractionation of
nodule extracts followed by anion exchange chromatography in ace-
498
tate buffers allowed the separation and purification of two major and
two minor species of soybean Lb. The major species were named Lb,
and Lb,, and the minor components Lbb and Lbd. Subsequent work
using isoelectric focusing, which has a greater resolving power, has
shown that there are four major components (subscripts a , c 1 , c2, and
c 3 ) rather than the two originally separated, and four minor ones (b,
d l , d2, and d3) (18).The latter minor components are N-terminal
acetylation products of the major components (19).All of these species
are present in single nodules, with the relative concentrations of the
different components varying with nodule development (20). This pattern of multiple gene products is widely established in other legume
and nonlegume symbioses (12, 18,211. All of the multiple Lb proteins
from soybean have been sequenced, and it is now known that the
differences in amino acid sequences between different Lb components
from a particular plant species are relatively slight. Despite discrepancies between the published sequences obtained in early studies, it
has been shown in a large study of both 69 domesticated cultivars of
soybean and 18 wild species that there are few (if any) differences
between the same type of leghemoglobin obtained from a number of
different cultivars of a single species. The published discrepancies
have been ascribed to problems in the sequencing procedures (22).In
all cases the proteins (molecular weight ca. 16,000) consist of a single
polypeptide chain of 144-155 amino acids, and the iron protoporphyrin IX prosthetic group. The proteins are acidic with PI values in the
range 5.0-5.5.
The amino acid sequences of the soybean proteins, as well as many
others, have a very high degree of amino acid sequence homology with
vertebrate myoglobins and hemoglobins, and an even higher degree
of homology of amino acid type. The sequence of soybean Lb,, which
is one of the most extensively studied species, contains two histidine
residues (His61 and His92) (171, which are analogous to the proximal
and distal histidines of mammalian hemoglobins and myoglobins. The
spacing of these residues in the sequence is also very similar. Crystal
structure work on both the soybean [(23);Ellis and Freeman, unpublished data] and lupin (24-30) forms of the protein have emphasized
the similar key features of the three-dimensional protein structures
of the plant and mammalian proteins. There are also a number of key
differences, which appear to have quite dramatic functional consequences and determine both the very high oxygen-binding affinity of
this protein (discussed later) and the ability of the plant proteins to
bind larger ligands at the heme site. The evolution of the plant globin
gene family has been the subject of a number of studies [see, for ex-
499
ample, (3111 and it is now believed that the plant and animal gene
families diverged a t a very early stage. This is estimated to be 900 to
1.4 billion years ago, though the derivation of these numbers requires
certain assumptions about the rate of evolutionary change (9, 13,311.
A number of studies have shown that the iron protoporphyrin IX
group of Lb can be replaced, in uitro, by a variety of other species
including mesohemin IX, deuterohemin IX, hematohemin IX, diacetyldeuterohemin IX, and other analogs. As expected, a number of
these substitutions cause significant changes in the heme environment, and this is reflected in both the optical absorption and EPR
spectra of these species (32,33). The apoprotein can also be reconstituted with a number of noniron porphyrins. Thus, holoproteins have
been produced with both cobalt (disulfophthalocyanine and protoporphyrin IX) and zinc (protoporphyrin 1x1among others (18, 34-36); in
the case of the cobalt protoporphyrin protein, information has also
been obtained on the oxygen-binding capabilities of this species (18).
111. Biological Localization
After considerable controversy in the literature over the last thirty

years it is now widely accepted that Lb only occurs in the cytoplasm
of the host cell, and not in the peribacteroid space that surrounds
the bacteroid (bounded by the peribacteroid membrane, Fig. 2) or the
bacteroid itself (7, 15, 37). The nature and development of the membrane that surrounds the bacteroids have been examined in some detail because this membrane regulates the transport of materials to
and from the bacteroids (e.g., dicarboxylates, metal ions, etc.) (38-41 ).
The local concentration of Lb can be high (up to 2-3 mM) in very
active nodules, with an average concentration over the entire nodule
of ca. 300 pM (42).Although it is well established [see reviews (12,
I8)l that the apoprotein is produced by the plant (43)and that the Lb
assembly must occur in the host cell or on the outside the peribacteroid membrane, the origin of the protoporphyrin moiety remains unclear. It is generally proposed that heme is made by the Rhizobium
symbiont (44, 451, but the possibility remains that plant organelles
are a t least partly responsible for this process (46).The mechanism
of control of the synthesis of the apoprotein is disputed. The observation that heme synthesis in isolated Bradyrhizobium japonicum is
greatly stimulated by microaerobic conditions (45)suggests that heme
export from the nascent bacteroids to the plant cell cytoplasm might
be a major stimulus for apoprotein synthesis. However, the observa-
500
FIG.2. Transmission electron micrograph (16,OOOX enlargement) through an infected root cell of a soybean plant. The subcellular organelles of the host cell [mitochondria and plasts; (a)] are present at; the periphery of the cell adjacent to the host cell
wall (b). The nitrogen-fixing bacteroids (c) are kept apart from the host cell cytoplasm
(the location of leghemoglobin) by the peribacteroid space (d) and the peribacteroid
membrane (e), which regulates transport of materials to and from the bacteroids.
tion that ineffective nodules in some mutant forms of Bradyrhizobium

japonicum deficient in heme synthesis can contain almost normal levels of apoprotein (47, 48) suggests that synthesis of the protein may
not be wholly heme regulated. The assembly of holoprotein appears
to precede and be independent of the presence of functional nitrogenase (37,43,49). However, the synthesis of the latter appears to depend on the former, possibly as a result of the control exerted by intact Lb on oxygen tension, which prevents damage to the
nitrogenase (37).
The turnover time of intact Lb is relatively short [reviewed in (37)],
cf: the long half-lives of myoglobin and hemoglobin in mammals, with
the major Lb components of pea nodules having turnover times of
about two days (49). This is considerably less than the lifetime of
the nitrogenase (50).There must therefore be continual synthesis and
degradation of Lb within the host cell cytoplasm, and there is evidence for dramatic changes (as mentioned earlier) in relative concen-
50 1
trations of different Lb isoforms during this period. Thus, it is known

that Lb I predominates over Lb IV in young but not old pea nodules
(51), and Lbrpand Lb, predominate in young soybean nodules (20).
The destruction of Lb is believed to occur by multiple pathways (12,
37), with the relative contributions of these different routes strongly
dependent on the environmental situation; some of these processes
are discussed further later. Whatever the process, there is a strong
correlation between nitrogen-fixing ability and the presence of intact
functional Lb, and considerable work is under way to both understand
and control these factors.
IV. Reactions with Different Molecules
A. DIATOMIC
GASES
The extremely high oxygen-binding activity of the Fe(1I) Lb is critical for its biological function, and considerable work has been undertaken to understand the factors that determine this feature of the
protein. It is now well established that the high affinity arises from
larger on (association) rate constants and smaller off (dissociation)
rate constants (see Table I), which results in KDvalues in the range
of 36 to 78 nM depending on the source of the Lb and the pH (15,21,
37,51,52). The 0, off rate constants for soybean Lb decrease by about
TABLE I
KINETICS AND
EQUILIBRIUM
CONSTANTS FOR BINDING
OF GAsEOLJS LIGANDS
TO LEGHEMOGLOBINS
AND OTHER
MONOMERIC
HEMOGLOBINS
(21, 55, 56, 166, 167)
Oxygen
Protein
Spermwhale Mb
Spermwhale Mbd
Human a-chain Hb
Soybean Lb
Lupin Lb I
Lupin Lb 11
ka
kb
12
14
140 1600
50
120
540
320
Carbon monoxide
28
5.6
20
25
Kc
0.51
860
11400
5.8
560
4.0
48 13
36 42
78 52
Association rate constant; pM-s
0.019
0.038
0.013
0.0078
0.014
0.015
37
6.6
3.25
0.6
0.33
0.015
Dissociation rate constant; s-l.

Equilibrium constant, kik; nM.
Mutant spermwhale myoglobin with E7 His
Major component of heterogeneous reaction.
kh
Gly mutation.
Nitric oxide
k!e
kh
17
150
30
120
0.00012
0.00015
0.000046
0.00002
0.007
0.001
0.0015
0.0001
-
502
fivefold as the pH is lowered from 7 to 4,with a pK, of 5.46. This has

been ascribed to the formation of a hydrogen bond between the bound
oxygen and distal histidine (52). The KD values are 11 to 24 times
smaller than those for sperm whale myoglobin, which is mainly due
to the higher on rate constant (21, 29). Thermodynamic quantities
(AG, AH, AS) have been determined for the binding of a number of
ligands including O2 and CO to soybean Lb (53), and the KD values
for all eight components of soybean Lb have been determined. Though
there are significant differences in these values, the variation in
terms of free 0, concentration within nodules has been calculated to
be a maximum of 3% (53).Thus, an earlier suggestion (54)that the
large number of different forms of Lb in particular species is due to
differing oxygen requirements, hence requiring Lb species of different
oxygen affinities, is not strongly supported by this data.
High on rate constants have also been determined in studies on a
sperm whale myoglobin mutant in which the E7 distal His was mutated to a Gly. The on rate constant for this mutant is approximately
the same as that of soybean Lb, though it is still smaller than for two
lupin forms (21, 55, 56).This type of alteration to the heme pocket is
not, however, responsible for the high affinity of Lb, as the off rate
constant of this mutant is orders of magnitude greater than those for
the soybean and lupin forms (55, 56). There must therefore be other
factors that contribute to the very high affinity of Lb molecules for
oxygen.
Detailed X-ray crystallographic studies of a number of different
states of lupin Lb have been reported (24-30,57),and similar studies
have been undertaken on the soybean protein (23).The structures of
the oxy, deoxy, CO, and NO forms of Fe(I1) lupin Lb, determined to
1.7- to 1.8-A resolution (29, 301, have allowed Fe-ligand geometries
and structural changes on ligand binding to be examined. It has been
found that the Lb protein has a unique feature as compared to other
hemoglobins in that the proximal histidine has a remarkable rotational freedom. Thus, in the deoxyform of the protein, the imidazole
is able to oscillate between two different orientations, eclipsing either
a line between the N-1 and N-3 pyrrole nitrogens of the porphyrin or
a line between N-2 and N-4 (29).In the oxy form, however, it is fixed
in a staggered orientation. As with other heme proteins, the iron atom
moves (by ca. 0.3 A) from an out-of-plane position to an in-plane position on oxygen binding, with the Fe-N bond distances remaining constant (29).The Fe-0-0 bond angle is 152", very close to that seen in
the human hemoglobin a-chain (153'1, and the oxygen is hydrogen
bonded to the distal histidine. In both forms of Lb the heme is ruffled
503
due to rotation of the pyrrole groups around the N-Fe-N bonds, in

accord with earlier suggestions based on resonance Raman studies
(58),though this suggestion was later disputed (59). This ruffling results in the methine bridges being out of plane by up to 0.32 A (29).
It has been suggested from the crystal structures that the reason
for the high on rate constant is the mobility of both the proximal and
distal histidines in the deoxy form of the protein (29). The eclipsed
nature of the proximal histidine would maximize steric hindrance
with the pyrrole nitrogens in the deoxy form and hence minimizer
T --$ p electron donation, whereas the staggered position in the oxy
form would result in the opposite effect. These two factors in combination might result in a reduction of the activation energy for oxygen
binding (29). Furthermore, though the distal histidine appears fixed
in the crystal structure, it has been postulated that this can swing
out of the heme pocket a t a rapid rate to allow ready access of the
oxygen molecule (15,23,60).
The kinetics of CO and NO binding to Lb have been studied, and
rate constants for both association and dissociation determined under
a number of different conditions (e.g., pH). The values obtained are
compared with those for Mb and Hb in Table I (61). The off rate constants for CO are, unlike those for 02,independent of pH (521. In
particular, such studies have been directed to determine the mode of
binding of the ligand and whether this has a bent geometry as has
been observed for some hemoglobins. Such an orientation is believed
to result from a steric interaction with the globin and is a major factor
in reducing the affinity for these poisons (30).The large flexible heme
site of Lb and the rapid movement of the distal histidine out of the
heme pocket might be expected to result in large on rate constants
for reaction with both NO and CO. The crystal data a t 1.8-A resolution has been interpreted in terms of a bent orientation of both CO
and NO with an Fe-CO angle of 160" and an Fe-NO angle of 147"
(30).The former angle is in conflict with spectroscopic data obtained
with myoglobin, where a linear geometry was determined (62, 631,
but in accord with the angles obtained from myoglobin X-ray crystal
structures (64, 65) and picosecond time-resolved infrared measurements (66).The NO conformation is in accord with the expected bent
geometry, with a strong Fe-NO bond (and hence low off rate constant), and the (predicted) lengthened (2.2-A) iron-proximal histidine
bond (30).
The binding of NO to the Fe(I1) form of Lb (LbNO) may be of considerable biological relevance as a result of the slow off rate constant
(21,30).This will result in the formation of Lb, which cannot fulfill its
504
role as an oxygen carrier. Such reactions have long been recognized

as having a potential role in nitrate-induced stress (67). It has been
demonstrated that Lb extracted from plants grown in the presence of
nitrate contains significant amounts of LbNO; thus, an early study
reported that up to 27% of the Lb present in soybean nodules was in
this form of the protein (68).The species can be readily recognized on
the basis of its characteristic UV-vis absorption features (69, 70) or its
intense EPR signal in the g ca. 2 region, which has highly distinctive
hyperfine couplings arising from the nitrogen nucleus (68). Although
it is possible that some of the NO may arise as an artifact of the
isolation procedure, it may also arise from the reduction of nitrate to
NO by either enzymatic or nonenzymic reactions (67, 68, 71). The
former process may involve a nitrate reductase from either the plant
or bacteroid. Such reactions have been characterized previously (71).
It is known that a variety of reducing agents such as ascorbate or
reduced flavins can also bring about this reduction. Such reactions
are believed to be responsible for the formation of nitrosylmyoglobin
in mammalian systems (72). Recent careful studies have confirmed
that the root nodules of nitrate-fed soybean plants do indeed contain
LbNO and that this species is not an artifact of the extraction procedure (69, 70). Thus, extraction of Lb from root nodules in the presence
of high concentrations (100 nmol/mg protein) of nitrate added to the
nodules at the time of harvest did not result in amounts of LbNO
higher than those seen in plants grown in the absence of nitrate. It
has also been demonstrated that the formation of LbNO requires both
the presence of the Fe(I1) protein and a reducing agent. Incubation
with monoiodoacetic acid, a thiol (-SH) blocking agent, repressed the
formation of this species, suggesting that an enzymatic process may
be involved (69). The levels of LbNO in root nodules appear to be
enhanced by culturing the plants in higher concentrations of nitrate
[cf: the detection of 27% LbNO by Maskell et al. for plants grown with
0.5 mM nitrate, and of 60% by Kanayama and Yamamoto using 10
mM nitrate (68, 6913 or for increasing lengths of time [cf. 60% and
86% LbNO for culture with 10 mM nitrate for 12 and 24 h respectively (6911. The formation of such LbNO complexes may have profound physiological significance, particularly at these very high nitrate levels, as the levels of LbNO in root nodules have been shown
to be inversely correlated with the acetylene (ethylenebreducing activity of the nodules (a commonly used marker of their reductive capacity), and hence their ability to fix nitrogen (69).
Although these experiments have determined that LbNO can arise
from the growth of plants in the presence of nitrate, the situation
505
with plants grown in the absence of nitrate is less well established.

In the work of Kanayama and Yamamoto, levels of LbNO obtained
from plants grown in the absence of nitrate were reported as being
less than lo%, suggesting that reduction (either enzymatic or direct)
of nitrate was a major, if not the only, source of NO (69).However,
very recent reports have suggested that this may not be the case.
Thus, it has been reported that root nodules contain an NO synthase
activity that generates NO from arginine in a manner akin to mammalian NO synthases (73), and a recent EPR study has reported the
detection of signals from LbNO in frozen intact root nodules from
plants grown in the complete absence of nitrate (74). The levels of
this complex were found to be highest in young nodules, decreased in
mature nodules, and almost completely absent in old and senescent
species. These reports, if they are confirmed, together with earlier
studies that have demonstrated that NO can play a role in regulating
gene transcription via a sensory system involving proteins of the
family (75-77), suggest that the formation and levels of NO may play
a vital role in determining the nitrogen-fixing activity of root nodules.
The determination of the crystal structures of these different forms
of Lb have also allowed previous suggestions that the heme pocket of
Lb is larger and more flexible than those of other heme proteins (15,
23, 37, 781, to be examined in detail. Superposition of the backbone
carbon trace of the E and F helices of the Lb structure with that
obtained from HbCO and MbCO provides direct confirmation of the
larger size of the heme pocket and the altered orientation of the distal
histidine residue for both lupin (30) and soybean proteins (15, 23).
This greater size and flexibility also explains a further unusual property of Lb: a t slightly acidic pH values the distal His of Fe(II1) (met)
Lb can move close enough to the iron atom to become a ligand (151,
whereas at higher (neutral) pH values it can swing right out of the
heme pocket to allow entrance of large ligands such as fatty acids,
nicotinic acid, and isoquinoline to occur (30). In contrast to this larger
pocket, the Lb structure shows a somewhat narrower entrance7with
the a-carbon of the E7 (distal) His nearer to the plane of the heme
ring by ca. 1.5 A compared t o hemoglobin and myoglobin, but with
the imidazole ring at approximately 90 to the angle found in these
other proteins (30).
Early suggestions [e.g., (79)1that Lb might act to facilitate nitrogen
fixation by acting as an ancillary nitrogen carrier have been shown to
be unlikely. Thus, although it has been reported that N, complexes of
both Fe(II1) and Fe(I1) Lb can form, these only exist to any significant
extent when nitrogen gas is employed at high pressure. Thus, 1.2 at-
506
mospheres of Nz were required to give half formation of the Fe(II1)

adduct and 1.7 atmospheres for half formation of the Fe(I1) form (80).
These species are unlikely to be of biological significance (37).
B. OTHERMOLECULES
It has been established in a number of studies that Lb(I1) behaves
in a somewhat similar manner to myoglobin and hemoglobin with respect to its binding of a number of small inorganic and organic ligands (37). Thus, it is known that Lb can form stable complexes with
toxic groups such as azide, cyanide, alkylisocyanides, and fluoride in
a manner similar to other heme proteins (81). The kinetics of many
of these binding reactions have been examined using fast laser pulses
to monitor reactions after photodissociation of the complexes, and
these values have been compared with those for other heme proteins
(55).In many cases the pattern of high on and slow off rate constants
seen with diatomic molecules is mirrored with these organic ligands
[see, for example, (8111. The formation of the fluoro species has been
suggested (37) to be of use in the detection of Fe(II1) Lb in nodule
slices or extracts. When used in large excess [concentrations up t o
0.1 M have been employed (2, 8211, the fluoride will displace other
ligands and give rise to a characteristic (and diagnostic) W-vis absorption band at 608-610 nm (83). This complex is in a pure highspin state, which has proved useful in the detection of Lb by use of
EPR spectroscopy (37). The EPR parameters of a number of other
Lb-ligand complexes have been studied in some detail (68).
However, unlike most mammalian heme proteins, Lb(II1) is also
known to readily form complexes with a number of small carboxylic
acids including acetate, propionate, butyrate, and valerate (84).
The
reaction with acetate has been shown to occur in a proton-dependent
reaction, pK, 4.8 (84),
to form a high-spin complex (85).This complex
is only slowly reduced by powerful reductants such as sodium dithionite (83). The pK, coincides with that of acetic acid, and it was suggested (84)that undissociated acetic acid binds to the heme iron.
However, it has also been suggested that this the pK, may arise from
another ionizable group on the protein and that this group may need
to be protonated before the acetate ion can ligate to the iron ion (37).
These observations and those with other ligands (discussed later),
form the basis of an electrostatic gate model of binding of ligands to
Lb at high pH values (86, 871, with electronic interactions between
anionic ligands and the ionized residue(s) on the protein restricting
access (23). The exact nature of the residues that give rise to this gate
507
and its significance in viva requires further study. The early observation (84)that larger carboxylates bind to Lb(II1) almost as readily as
acetate led to the suggestion that the heme pocket of Lb must be
larger and more open than that of myoglobin, in which little or no
binding is observed. This suggestion was subsequently borne out by
X-ray crystallographic studies (23, 29).
The pocket is large enough to accommodate ligands such as imidazoles (though these appear to be somewhat different complexes from
those seen with myoglobin and hemoglobin), nicotinic acid (pyridine3-carboxylic acid) and various derivatives (e.g., the 5-bromo and 5fluoro species), pyridine and some substituted species, a wide variety
of amines, including some long-alkyl-chain species, and isoquinoline
(37, 88-90). The binding of nicotinic acid to Lb has been the subject
of several studies as a result of the observation that this ligand is
often bound to Lb extracted from root nodules (37, 88, 89, 92). It is
known that a number of nicotinic acid derivatives will not react with
Fe(II1) Lb, suggesting that the ligand makes a number of close contacts with the surrounding globin protein. Thus, amide and N-methyl
derivatives of nicotinic acid do not readily bind. The binding of nicotinic acid, like that of acetate, is pH dependent, pK, 4.9,and the dissociation constants for the Fe(I1) and Fe(II1) complexes are 33 p M and
1.3 pM respectively (89, 90). The pH dependence and lack of reaction
with the N-methyl derivative have been interpreted in terms of reaction of the un-ionized ring nitrogen with the heme iron and interaction of the ionized carboxyl group with a protonated residue on the
apoprotein. These predictions have been borne out by the crystal
structure of the soybean Lb-nicotinate complex, with the species interacting with the carboxyl substituent being either the imidazole nitrogen of the distal histidine or possibly the phenolic -OH group of
Tyr30 (23).
V. Oxidation of Fe(ll) Leghemoglobin
A number of species that are known to be or may be present in

root nodules can oxidize Fe(1I) Lb. These species include nitrite (92),
superoxide (93), and peroxides (94). The process would be expected to
be relatively facile, as the reduction potential for the Fe(II1) LbFe(I1)
Lb couple has been reported to be +0.22 V a t pH 7 and +0.27 V a t
pH 6 (95). Even in cases in which considerable care has been taken
during the extraction process, most extracts from root nodules contain
a t least traces of the Fe(II1) form of the protein. Whether this arises
508
during the extraction process or is present in the intact root nodule

has been the subject of some debate [e.g., (12, 3713. Detection of the
Fe(II1) form of the protein by optical spectroscopy in intact root nodules is fraught with difficulty, due to the light scattering produced by
the nodules and an incomplete knowledge of the different ligand species that might be present and their respective contributions. Early
studies suggested that Fe(II1) Lb could not be detected in soybean
root nodules (96). However, in a review article (97) Klucas and Becana
have recently reported the results of experiments carried out on old
intact nodules attached to roots and maintained under an atmosphere
of either air or 100% oxygen, which have suggested that low concentrations of Fe(II1) Lb can be present. These nodules were exposed t o
fluoride (presumably at high concentrations), and the absorbance at
the 625-nm parent peak and at 610 nm for the fluoride complex monitored. The low initial absorbance at the former wavelength is rapidly
lost, and an increase in absorbance at 610 nm is observed. These
changes have been ascribed to the formation of the Fe(II1) fluoride
complex (97). A similar treatment with nicotinate also resulted in a
decrease in absorbance at 625 nm and an increase in absorbance at
557 and 526 nm, which correspond to the absorption maxima of the
Fe(II1) nicotinate complex. Both experiments have been interpreted
in terms of the prior presence of low levels of Fe(II1) Lb in the intact
nodules (97). However, it is impossible to rule out the possibility that
the high concentrations of these two ions, which have a high affinity
for Fe(II1) Lb, are enhancing formation of this species through alteration of the redox potential. In contrast, recent EPR studies on intact
nodules from soybean plants of different ages that were rapidly frozen
immediately after harvesting provided no evidence for high-spin complexes of Fe(II1) Lb, which have very distinct g values of ca. 6 and 2
(74). Whether low-spin forms of the Fe(II1) protein were present was
more difficult to discern due to the relatively weak nature of these
features and the presence of strong absorption bands from the NO
complex of Lb(I1) (also discussed later). Thus, it is still not unequivocally established whether Fe(II1) species of Lb are present at any significant levels in young and mature nodules. The situation with senescent nodules is somewhat different, and it is likely that intact
Fe(II1) Lb and/or partially degraded Fe(II1) Lb are present (74, 97).
Reaction of Fe(I1) Lb with superoxide radicals has been suggested
as an important route to the formation of Fe(II1) Lb via Eq. (1)(93).
The precise nature of the reaction remains to be fully established; it
has been reported that the process is inhibited in uitro by the presence
of the enzyme superoxide dismutase (SOD), which removes the super-
509
oxide radical [Eq. (211 (93). It should be noted, however, that 02.also
appears to be able to act as a reductant for the Fe(II1) Lb, reducing it
to the Fe(I1) form [Eq. (311, and that the process is ameliorated by the
presence of SOD (93).The latter reaction [Fe(III) to Fe(1I)I appears to
be considerably slower than the former. The significance of these 0,. dependent processes will depend on the form of the protein present
initially. Current evidence is that the major form in functional nodules is the Fe(I1) form, and it is likely that the major effect of 02.generation in uiuo will be oxidative [generating inactive Fe(II1) Lbl
rather than reductive (which might be protective).
'
202.
+ 2H' 2H,O,
-t O2
Fe(II1)+ 0 2 ,-+Fe(I1)+ 0,.
(2)
(31
The LbO, adduct is known to undergo a slow autoxidation to the

Fe(II1) form with release of O,.- (93). This process is usually slow, but
the rate may be enhanced under certain conditions (e.g., pH, temperature, concentration of Lb, and the presence of certain metal ions, chelators, and anions) (93, 97). Thus, the high levels of Lb02 present in
intact nodules and the drop in pH observed in old and senescent nodules [pH values of ca. 6.4 have been reported for young nodules, and
ca. 5.5 in old or stressed species (98)] would be expected to enhance
formation of the Fe(II1) form, with release of O,--.
Incubation of soybean bacteroid preparations with nitrite (0.4 mM)
has been shown to markedly inhibit ethylene reduction and concomitantly be capable of inducing the oxidation of LbOz to Fe(II1) Lb (92).
The inhibition of nitrogen fixation in these bacteroids has been ascribed to both a direct inactivation of the nitrogenase and the effect
of nitrite in converting LbOz into an inactive form (92). In contrast,
direct evidence for (uncatalyzed) autoxidation processes in uiuo is
lacking because, despite numerous efforts, significant amounts of the
Fe(III)Lb have not been detected in intact root nodules. This is probably due to efficient reduction of the Fe(II1) form to active Fe(I1) Lb.
In contrast t o the autoxidation of LbOz, LbNO is resistant to both
autoxidation and oxidation by added oxidants such as [Fe(CN1613(68-70); it cannot therefore be readily converted back to the oxygenbinding form of Lb even by oxidationheduction cycles.
510
Reaction of deoxy Fe(I1) Lb with small excesses of Hz02results in

rapid changes in the UV-vis spectrum of the heme center, with the
appearance of new absorption bands a t 416 nm ( E 97.2 mM-'. cm-')
and 543 nm ( E 10.4 mM cm-'1, with a shoulder a t 575 nm (94). The
second-order rate constant for this reaction is 2.24 X lo4M-' s-' (94).
These changes have been ascribed to the oxidation of the Fe(I1) center
by the peroxide to give a water molecule and a species that has optical
properties similar to Compound 11, the Fe(IV)=O species in peroxidases:
Fe(I1) + H,O,
+ Fe(IV)=O
+ HzO.
(5)
Approximately 1.5 equivalents of hydrogen peroxide are required

per mole of heme to obtain maximum yields of this species (94). This
species is remarkably stable, with little decay over 12 h at 25C in
the absence of reducing agents (94). The corresponding species formed
in the case of myoglobin and hemoglobin decay much more rapidly.
The number of oxidizing equivalents present in this species has
been examined by titration with reducing agents such as sodium
dithionite. Two equivalents of the latter are required per mole of oxidized Lb, whereas only one equivalent was required to reduce Fe(II1)
Lb (94). When similar reactions were carried out with larger excesses
of HzOz, a more rapid conversion of Fe(I1) to F e W ) was observed.
From the absorbance changes at 543 nm due to the Fe(1V) species,
the yield is somewhat reduced with excess hydrogen peroxide above
1 :4 (Lb :H20z).Under such conditions the Fe(1V) absorbance is gradually lost due to slow reactions resulting in destruction of the heme
chromophore (94). Such a n interpretation is in line with studies carried out using chelating ligands with a high affinity for iron. In these
studies no significant release of iron was observed until a greater
than 4 : 1 excess of hydrogen peroxide was employed (99). It should be
noted that the release of iron by this method may not accurately reflect the release of iron in other situations, as such high-affinity chelating ligands may remove iron from the protein, where it would otherwise have remained bound. Thus, the assessment of heme
degradation by W - v i s absorbance measurements, which reflect early
stages of heme degradation, do not necessarily correlate with the release of iron as measured by other methods.
When similar reactions were carried out with LbO,, the absorbance
changes were much slower. Early studies with mammalian heme proteins suggested that such reactions might give rise to free hydroxyl
radicals via one-electron reduction of the hydrogen peroxide ( l o o ) ,but
511
it is now believed that this reaction proceeds via the formation of a

high-oxidation-state species similar to that described earlier, a t least
when low excess peroxide levels are employed. No evidence for free
hydroxyl radical formation has been obtained under these conditions
by product analysis (99) or EPR spin trapping (101).When higher
concentrations of hydrogen peroxide were employed, which can result
in heme degradation with release of iron [typically a t peroxide : heme
ratios > 6 : 1; LbO, appears to be less sensitive than the Fe(I1) form],
some hydroxyl radicals may then be generated by the iron released
(99).With low concentrations of peroxide and intact LbOz, the reaction presumably involves displacement of the dioxygen by the hydrogen peroxide and reaction as described earlier. Such changes are
readily monitored using spectrophotometry by the absorption decrease a t 574 nm or the increase a t 625 nm. As with the species
formed from the deoxy form of the protein, this species is very stable
in the absence of reducing agents. Oxidation of ligated forms of Lb
can also occur; thus, it has been shown that the Lb(I1) nicotinate can
be oxidized to the Lb(II1) form by hydrogen peroxide. Further oxidation to higher oxidation states does not occur, however (94).The nature of this oxidation, and in particular the fate of the second oxidizing equivalent from the hydrogen peroxide, remains to be established.
VI. Oxidation of Fe(lll) Leghemoglobin
Reaction of H202with Fe(1II) Lb gives lower yields of Fe(IV)= 0 Lb

than those seen with Fe(I1) (94, 102). The absorbance changes observed during reaction with low excesses of the peroxide were ascribed to the formation of both the Fe(IV)=O species and an additional (green) compound (102).The latter is characterized by UV-vis
absorbance bands a t 410 and 630 nm (Fig. 3). This species has since
been shown to be very stable and has been further isolated and characterized (see Section IX,A) (103).The nicotinate complex of Fe(II1)
Lb does not undergo these reactions (94).
The formation of Fe(IV)=O Lb suggests that the Fe(II1) center is
oxidized by only one of the two oxidizing equivalents from the peroxide. The fate of the second equivalent has been the subject of much
study. In early work, in which the ability of Fe(II1) Lb/HzOamixtures
to initiate lipid peroxidation and oxidation of organic molecules was
examined, it was postulated that the second oxidizing equivalent was
released as a hydroxyl radical:
512
Wavelength
(nm)
FIG.3. Formation of the additional (green) compound on reaction of soybean Fe(II1)

Lb (50 pM)with HzOz(100 pM) in 25 mM KH2P04/KOHbuffer, pH 7.4, containing 0.1
mM DTPA. Visible range spectrum recorded 120 s after addition of HzOz a t 25C.
Fe(II1) + H202+ Fe(IV) = 0 + HO..
(6)
Further studies on this system and related work with the analogous
mammalian heme proteins myoglobin and hemoglobin have shown
that free hydroxyl radicals are not released, at least with low excesses
of hydrogen peroxide, when oxidative damage to the heme and release
of iron is minimal (101). These reactions therefore appear to occur in
a similar manner to other peroxidase enzymes in which the initial
oxidation by the peroxide involves a two-electron transfer from the
heme iron to the peroxide. However, unlike peroxidases, the second
oxidizing equivalent does not appear to be retained by the heme center for any significant length of time, and the second equivalent
is rapidly dissipated into the surrounding protein (globin) matrix
[Eq. (7); (101)l.It appears therefore that the initial reaction gives an
Fe(IV)=O species and a heme (porphyrin) radical cation. The radical
cation, unlike those present in classical peroxidases, reacts with the
surrounding protein by what appears to be an electron-transfer process to oxidize one or more globin residues with concomitant reformation of the intact heme ring (101).Thus, within a very short period
the reaction gives an Fe(IV)=0 center identical t o that formed with
Fe(I1) Lb, and one or more globin radicals. This situation is similar to
what occurs with myoglobin (104-110):
Fe(1V)=0 (heme+.)(globin)+ Fe(IV)=0 (hemeKglobin' .).
(7)
5 13
Though there is now some considerable evidence that globin radicals are formed in these reactions, it is still not absolutely clear where
the oxidizing equivalent resides and why. In initial (direct, room temperature) EPR studies a multiplet signal was detected on reaction of
the Fe(II1) form of the protein with low excesses of hydrogen peroxide
(Fig. 4) ( 1 0 1 ) .This signal was only detected shortly after initiation of
the reaction. Monitoring of the intensity of the EPR signal by carrying
out time-course experiments in which the magnetic field was set to
that corresponding to the position of one of the absorption lines, allowed the half-life of the species to be determined as approximately
40 seconds (Fig. 5). This globin-derived species was observed with all
three types of Fe(II1) Lbs tested (a, e l , and c3) and was also detected
with a number of other hydroperoxides and two-electron oxidants.
The former observation suggests that the minor changes in sequence
and structure between the isoforms are unimportant in determining
the course of the reaction; the latter observation suggests that a common mechanism occurs, with similar globin-derived radicals generated in all cases. The latter strongly suggests that the globin radical
does not contain any part of the original oxidant (e.g., is not an adduct
formed by addition of the hydroperoxide to an amino acid). In further
experiments in which the Lb was iodinated before treatment with hydrogen peroxide (IOI),the EPR signal was not observed. The condi-
FIG. 4. EPR spectrum observed immediately after mixing soybean FeUII) Lb (250
pM) at pH 7.4. Reaction studied using a two-way stopped-flow
mixing system inserted into the cavity of the EPR spectrometer. The signal is assigned
to a sterically constrained tyrosine phenoxyl radical formed at position 133 (reproduced
with permission from Davies, M. J.; Puppo, A. Biochem. J . 1992,281,197-201).
pM) with HaOz (250
514
Time after mixing (s)
FIG.5. Build-up and decay with time of the EPR signal (see Fig. 4) from the tyrosine
phenoxyl radical formed a t position 133 of the soybean protein on reaction of Fe(II1) Lb
(250 pM) with H202 (250 p M ) a t pH 7.4 (reproduced with permission from Davies,
M. J.; Puppo, A. Biochern. J. 1992,281, 197-201).
tions under which this iodination was carried out were such that only
the Tyr residues in the protein ought to be affected, suggesting that
the Tyr residues present in the globin are either the site of the radical
species or involved in the generation of the radical.
Determination of the g value of the observed signal (2.0044) and
analysis of the hyperfine coupling constants suggest that the signal is
a Tyr-derived species, with both the g value and the coupling pattern
consistent with the presence of a Tyr phenoxyl radical (101). This
species does not, however, have coupling constants identical to those
of a free Tyr phenoxyl radical [see, for example, ( I l l ) ] , particularly
with respect to the coupling of the hydrogens of the methylene
(-CH,-) group, which attaches the aromatic ring to the backbone.
The variation between these values is consistent with the surrounding globin structure forcing the radical to adopt a fixed conformation with little free rotation round the methylene-to-ring bond
(112). This species also has somewhat different coupling constants
with regard to the methylene hydrogens than those reported for analogous Tyr phenoxyl radicals in other proteins (113-1241, suggesting
that the conformation of such radicals is not universal in proteins.
In the soybean form of the protein, which was used in these experiments, there are three Tyr residues, with one of these, that at position
133, near the heme ring (Fig. 6). This was suggested to be the site of
the radical on account of its proximity to the heme center at which
the initial oxidation must be occurring (101). Further studies with a
variety of scavenging agents demonstrated that the radical can undergo a number of reactions. Thus, its signal was lost when high per-
515
FKi. 6 . Position of the Tyr133 residue in soybean metLb in relation to the heme
ring. Computer-generated molecular model using crystal coordinates ( 2 3 ) (Ellis and
Freeman, unpublished data).
oxide concentrations were employed, a process assigned to reaction

between the radical and excess hydrogen peroxide. The signal was
also diminished in intensity, or lost completely, when various hemebinding agents (cyanide, azide, nicotinic acid) were used or reducing
agents were added (see Section IX,B). The latter was attributed to
one-electron reduction of the radical (102).The observation that the
globin-derived radical reacts with agents expected to be present solely
in the aqueous phase, and not expected to penetrate far into the protein structure, suggests that the radical must be present on the protein surface, or not far removed from it.
516
Further studies with one of the lupin forms of Lb have helped elucidate the site of the radical (112). This protein contains only two Tyr
residues, with only one of them conserved when compared to the soybean structure. Thus, it was hypothesized that if the radical had identical parameters and similar kinetics for formation and decay, then it
could be assigned to the single conserved residue (Tyr133 in soybean
and Tyr138 in lupin). This proved to be the case, with a very similar
EPR signal observed with the Fe(II1) form of the lupin protein on
treatment with small excesses of hydrogen peroxide (112).The signals
obtained in experiments with the lupin protein were better resolved
than those with the soybean protein, and this allowed a detailed analysis of the splitting pattern and coupling constants to be carried out.
The spectra were computer simulated using this data, adding further
support to the theory that the globin radical is indeed a Tyr-derived
phenoxyl radical (112).
The mechanism of formation has been postulated to involve rapid
one-electron oxidation of the aromatic ring by the heme radical cation
to give a Tyr radical cation [Eq. (8); (101, 112)l. Using model phenols
such species undergo rapid deprotonation to give a phenoxyl radical
with rate constants ca. 1OO dm3.mol-l- 5-l (125, 126). Thus, these
model studies are consistent with the formation of a radical within
the globin polypeptide:
Fe(1V)=0 (heme.)(globin-Tyr-OH)4 Fe(IV)=O (heme)(globin-Tyr-OH)
+ Fe(IV)= 0 (heme)(globin-Tyr-0)+ H i .
(8)
In the experiments reported in these studies, little attempt was

made to quantify the extent of formation of the radical. Such studies
are technically challenging, especially when the kinetics of formation
of the radical are poorly defined. However, from the intensity of the
signals the steady-state concentration of the radical is considerably
less than that expected from the concentrations of protein and oxidant
employed. Thus, it appears that the radical is not formed in a very
efficient manner. This suggests that there are other routes to the loss
of the second oxidizing equivalent from the heme center or there are
other pathways that result in the loss of the Tyr radical.
In an effort t o determine whether other radical species are generated on the globin, EPR spin-trapping experiments were carried out
using the spin trap DMPO (5,5-dimethyl-l-pyrroline-N-oxide).
In
these experiments, unlike those carried out with myoglobin, no significant signals were observed (101).At the time the interpretation
517
was in terms of nonformation of additional radicals, but later work

has shown this to be incorrect. The incorrect conclusion probably
arose from the fact that the spin trap chosen was not ideal, with the
trapping rate constants rather slow (i.e., other species formed were
not trapped and hence not observed). These negative experiments
did, however, allow one possible mechanism of formation of the
observed radical to be ruled out. The trap reacts rapidly with HO.
(K ca. 2 x 1 0 9 M - I. s - 11 to give a spin adduct with distinctive EPR
parameters and a long half-life (127, 128). The absence of this wellcharacterized product suggests that free HO. are not formed (101).
Similarly, the absence of alkoxyl radicals (and related species),
which also give long-lived and well-characterized spin adducts, in
experiments in which hydroperoxides (and similar compounds) were
used to generate the radicals, rules out the involvement of these
radicals (101).
Subsequent studies using alternative spin traps have demonstrated
that other radical species are generated on the globin when Fe(II1)
Lb is incubated with low excesses of hydrogen peroxide (129). Thus,
incubation in the presence of the spin trap MNP (2-methyl-2-nitrosopropane) has resulted in the detection of broad, partially anisotropic
EPR signals. The partial anisotropy of these signals is believed to be
due to the trapping of relatively large, slowly tumbling radical adducts. These are believed to be trapped globin-derived radicals (129).
Similar experiments carried out with other spin traps, such as the
nitrones PBN (a-phenyl-N-t-butylnitrone)
and POBN [a-(pyridine-1oxide)-N-t-butylnitrone],
also give EPR signals that have been assigned to globin-derived radicals (129).The formation of these species
does not appear to be affected by the presence of dioxygen in the reaction mixture. This suggests that the radicals are not formed from an
oxygen-derived species (such as a peroxyl radical) and that they do
not react rapidly with oxygen. There are, however, two important caveats to these suggestions. First, it is very difficult to make such reactions completely anoxic, as oxygen may be generated from the decomposition of hydrogen peroxide. Second, the high concentrations of spin
traps used in these experiments (to obtain reasonable signal-to-noise
ratios) may result in reaction with dioxygen being uncompetitive. In
only one case, with the nitroso spin trap DBNBS (3,5-dibromo-4-nitrosobenzene sulfonic acid), were signals observed in the absence of hydrogen peroxide. In the latter case the signals have been assigned to
a novel compound arising from a chemical reaction of the heme moiety with the trap and not spin trapping. Subsequent oxidation of the
initial product gives the observed signal (129).
5 18
In the case of POBN at least two adducts were detected, and the
EPR signals were almost completely isotropic in nature, in contrast
with those detected with MNP and PBN. This suggests that the radicals present have considerable mobility (129). The trapping with
POBN, which is a highly hydrophilic trap and partitions very poorly
into hydrophobic solvents, suggests that the radicals detected are
present on the outer surface of the globin. The situation with the
other traps is less clear, as PBN and MNP partition readily into hydrophobic media. Unfortunately, the spectra obtained do not give extensive information as to the nature of the radicals formed. Thus, the
spectra observed with PBN and MNP give little information due to
their anisotropic nature, whereas the sharper lined spectra observed
with POBN do not allow much information to be obtained because the
adducts formed do not show a dramatic change in hyperfine coupling
constants (128). However, the signals observed with MNP suggest
that the added radical is carbon-centered in nature.
To obtain further information about the species trapped in these
experiments, enzymatic digestion experiments were performed on the
spin adducts (129). This technique, which has been successfully employed to obtain information on the nature of radicals trapped from
other proteins, DNA, RNA, and complex carbohydrates (130, 1311, relies on the fact that many of the trapped species are stable for considerable periods of time. The macromolecular radical adduct can therefore be treated with either nonspecific o r specific proteases that will
release smaller fragments, at least some of which should still have
the spin trap attached. This process may be aided by the fact that
many proteolytic enzymes degrade damaged proteins more readily
than the corresponding undamaged parent macromolecule (126, 132).
Such fragments would be expected to tumble rapidly in solution and
hence give isotropic spectra, from which further information may be
obtained. When such experiments were carried out with the Lb adducts detected on treatment of Fe(II1) Lb with hydrogen peroxide in
the presence of MNP, a conversion of the anisotropic signal to more
isotropic features was observed. The spectra observed at the end of
this process consisted of a single triplet, which is consistent with the
trapping of at least one tertiary carbon-centered radical (129).Though
these experiments do not allow complete identification of the protein
radical, they suggest that the radicals are formed at stabilized sites
k e . , tertiary rather than secondary or primary).
Recent studies on the analogous myoglobin reactions have suggested that a highly resonance-stabilized radical that is trapped via
the C-3 position on the indole ring (108, 1101, can be formed on the
5 19
side chain of Trp residues. This is a tertiary site and would be expected to be formed readily due to the ease of oxidation of such Trp
side chains. Thus, it is possible that a similar species may form and
become trapped in the Lb experiments described earlier, and that
such a radical may be the source of the tertiary nitroxide signals seen
with MNP (and possibly the signal observed with PBN and one of
those detected with POBN). The detection of two distinct species with
POBN as the trap suggests that, even if this speculation is correct,
other types of radical must also be present. Recent studies have
shown that phenoxyl species can be trapped under certain circumstances, and in particular when other decay routes (such as dimerization) are inhibited by, for example, steric factors (133).Thus, some,
but not all, of these spin-adduct signals may arise from the trapping
of Tyr-derived species. The additional radical species may arise via
two (or more) competing damage-transfer pathways within the protein or may be species formed along a single pathway. Evidence from
the analogous myoglobin experiments suggests that more than one
pathway exists, with both tyrosine and tryptophan-derived species
observed (104-110).Recent studies have also implicated histidinederived radicals (log),though it should be noted that some of this
evidence has been obtained from the study of mutant proteins (from
site-directed mutagenesis) from which key amino acids (e.g., the Tyr
residues) have been removed. Such data might be misleading, for removal of a key residue may merely switch damage transfer to another
pathway and hence other final sites that are not normally damaged
to any great extent. Considerable work therefore still needs to be carried out in order to obtain a complete picture of the radical species
that are formed in these reactions.
VII. Reduction of Fe(lV)=O Leghemoglobin
As mentioned earlier, Fe(IV)=O Lb is particularly stable, as least

when compared to the analogous species detected with other heme
protein systems (94).
This species remains observable 12 h after its
formation under favorable conditions and can be isolated from other
forms of the protein on isoelectric focusing gels. However, in the presence of reducing agents or readily oxidizable compounds it undergoes
further reaction, resulting in its conversion to a lower oxidation state.
Some of these reactions are biologically significant, as they repair the
oxidized species.
520
A. REACTION
WITH H,O,
Reaction with excess hydrogen peroxide results in the oxidation
of the heme ring and release of iron, though the exact mechanism
of this type of reaction is unknown. It may involve oxidation of a
further molecule of hydrogen peroxide (101) with consequent formation of the HOO. radical, which may then react with the heme ring,
possibly at the methylene bridge sites. The latter is consistent with
known degradation reactions of the heme ring with loss of the iron
atom (99).
B. REACTION
WITH ASCORBATE
It has been reported recently that addition of ascorbate to Fe(1II)
Lb immediately prior to the addition of small excesses of hydrogen
peroxide results in transient formation of Fe(IV)=O Lb as monitored
by UV-vis spectrophotometry (134).The situation when ascorbate is
added after formation of Fe(IV)=O Lb is more complex due to the
formation of additional species during the reaction of Fe(II1) Lb with
the hydrogen peroxide. In the former case, with 1 mM ascorbate and
50 p M Lb, the conversion back to Fe(II1) is complete in about 2.5
minutes (Fig. 7) (134).Under these conditions a further slow reduction of the Fe(II1) species is also observed (discussed later). The observation of isobestic points is consistent with the presence of only the
Fe(1V)= 0 and Fe(II1) species in the reaction system. As expected, the
rate of the Fe(IV)=O to Fe(II1) conversion was found to depend on
the ascorbate concentration and is associated with oxidation of the
ascorbate. The stoichiometry of the ascorbate oxidized to hydrogen
peroxide added is 1: 1. This ratio implies that both oxidizing equivalents in the peroxide are responsible for ascorbate oxidation and
hence that the protein-derived radicals are also removed. The reduction of the Fe(IV)=O form by ascorbate has been shown to be a oneelectron transfer reaction by the use of EPR spectroscopy. In these
experiments intense signals due to the ascorbyl radical were observed
(134).This process may occur via reaction of ascorbate at the heme
edge, which is known to be exposed to agents present in bulk solution
(Fig. 8).
Similar results were obtained when Fe(IV)=O Lb was generated
from Fe(I1) Lb, though in this case both oxidizing equivalents from
the hydrogen peroxide are believed to remain at the heme center before being removed by reaction with ascorbate.
52 1
1
0.9
0.6
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.9
0.8
0.7
0.6
0.5
0.4
4
0.3
0.i
0.:
1
Wavelength
(nm)
FIG.7. Spectrophotometric changes during the reduction of soybean Fe(IV)=O Lb

by ascorbate. Fe(1II) Lb (50 p M ) was mixed with H L 0 2 (100 /*MI immediately after
addition of 1 mM ascorbate a t pH 7.4 ( 2 5 mM KH,PO,/KOH buffer containing 0.1 mM
DTPA). Reactions were run a t 25"C, and repetitive scans were recorded a t 30-s intervals ( a ) and then 30-min intervals (b). Data are representative of experiments carried
out in triplicate [reprinted from Phytochemistq, 39, Moreau, S.; Puppo, A.; Davies,
M. J.; The reactivity of ascorbate with different redox states of leghaemoglobin, pp.
1281-1286, copyright (19951, with kind permission of Elsevier Science Ltd, The Boulevard, Langford Lane, Kdlington, OX5 IGB, UKI.
c. REACTION
WITH GLUTATHIONE,
O T H E R THIOIS,AND RELATED SPECIES
As in the case of ascorbate reactions, addition of glutathione (GSH;
1 mM) to Fe(II1) Lb (50 pM)before reaction with hydrogen peroxide
results in a decreased lifetime of the Fe(IV)=O species (102).The loss
522
Frc:. 8. The heme edge (dark shaded residues) in soybean Fe(II1) Lb is partially
exposed on the surface of the protein. Computer generated molecular model using crystal coordinates (23)(Ellis and Freeman, unpublished data).
of the UV-vis absorption bands of Fe(IV)=O Lb is accompanied by

the reappearance of those for Fe(II1) Lb, with good isobestic points
confirming that no other species is involved (Fig. 9). The reduction to
Fe(II1) Lb is accompanied by formation of oxidized glutathione [glutathione disulfide (GSSG), assayed using NADPH in the presence of
GSSG reductase] in approximately stoichiometric amounts (H,O,
added :GSH formed), again suggesting that both oxidizing equivalents from the hydrogen peroxide are utilized. The conversion of GSH
into GSSG is also believed to occur via a one-electron (radical) process, because glutathione thiyl radicals have been detected by EPR
spectroscopy. Similar behavior is observed with Fe(I1) Lb treated with
small excesses of hydrogen peroxide (102).
As might be predicted on the basis of these data, a number of other
thiol compounds have been shown to reduce Fe(IV)=O Lb to the
Fe(II1) state with concomitant formation of thiyl radicals (detected by
use of EPR spin trapping using 5,5-dimethyl-l-pyrrolineN-oxide)
(135). In some cases, however, other species are formed, and these
have been identified from their UV-vis absorption spectra as novel
sulfur species formed by nucleophilic attack on the tetrapyrrole ring
by the thiol group (135).The ability of a thiol t o undergo two such
competing reactions is dictated by steric and electronic characteristics
523

0.60
0)
0.30
n
450
750
600
Wavelength
(nm)
FIG.9. Visible range spectral changes during the reduction of soybean Fe(IV)=O Lb
by GSH. Fe(II1) Lb (50 pM) reacted with HzOz (100p M ) at pH 7.4 (25 mM KH,PO,/
KOH buffer containing 0.1 mM DTPA). Repetitive scans were recorded 1 min after
addition of H202and then a t 3-min intervals (reproduced with permission from Puppo,
A.; Monny, C.; Davies, M. J. Biochem. J . 1993,289, 435-438).
of the thiol. Addition of cysteine results in reduction to the Fe(II1)

[and subsequently to Fe(I1) Lbl and formation of the Fe(II1) and Fe(I1)
sulfur derivatives via nucleophilic reactions. The Fe(I1) sulfur derivative appears in this system to be formed from the corresponding
Fe(II1) sulfur compound, suggesting that nucleophilic addition can be
followed by heme reduction. The difference in behavior between cysteine and GSH may arise from the greater steric bulk of the latter,
which may inhibit reaction a t the heme edge. This hypothesis is supported by the observation that the behavior of N-acetylcysteine is intermediate between the two.
Cysteamine (2-mercaptoethylamine) gives extremely rapid heme reduction with formation of Fe(I1) Lb and little formation of the sulfur
derivatives of Lb. 3-Mercaptopropionic acid on the other hand gives
only slow reduction of Fe(IV)=O Lb, even slower reduction to Fe(I1)
Lb, and some sulfur-Lb formation. This suggests that the overall
charge on the reducing agent can have a major influence on the kinetics and the extent of competition between the two pathways (135).
These differences have been rationalized in terms of the negatively
charged gate in the globin, which gives access to the heme. Electronic
interactions between the negative charge on the protein and the reductant may slow down the reduction with 3-mercaptopropionic acid
and give an enhanced rate with positively charged cysteamine. In contrast the formation of sulfur-Lb derivatives always appears to be rel-
524
DAWES, MATHIEU, AND PUPPO
atively slow and is only therefore a significant competing reaction

when reduction at the heme center is slow. The exposed nature of
the heme edge in Lb (23)(Fig. 8) may also explain the much greater
prevalence of nucleophilic processes as compared to myoglobin (1361,
where such reactions appear to occur much less readily.
Reduction of Fe(IV)=O Lb by disulfides does not occur to any significant extent, except in the case of oxidized lipoic acid, which brings
about slow reduction of Fe(IV)= 0 Lb to Fe(II1) Lb (135). No radical
species were detected by EPR spectroscopy during the reaction,
though such species are very likely generated by one-electron oxidation of the disulfide. No reaction between Fe(1V)= 0 Lb and hydrogen
sulfide has been observed, unlike the situation with myoglobin (137).
D. REACTION
WITH OTHERREDUCTANTS
Other chemical reducing agents can bring about loss of Fe(IV)=O
Lb. Thus, ferrocyanide can reduce Fe(IV)=O Lb to Fe(II1) Lb, though
this reaction is relatively slow and complex [as judged by changes in
the absorption spectra (94)l. In contrast, reaction with dithionite appears to be quantitative and has been used to estimate the concentration of Fe(IV)=O Lb (94).The Fe(IV)=O Lb does not undergo reaction with dioxygen, carbon monoxide, cyanide, or nicotinic acid (94).
Similarly, it does not appear to undergo reaction with Fe(I1) Lb (i.e.,
synproportionation to give two molecules of Fe(II1) Lb does not occur)
(138).The potential reaction of Fe(IV)=O Lb with lipids and membranes is discussed in Section IX,D.
VIII. Reduction of Fe(ll1) Leghemoglobin
A. ENZYMATIC
There is considerable evidence that functional root nodules possess
efficient means of reducing Fe(II1) Lb back to Fe(I1) Lb. It has been
shown that this occurs via a number of different mechanisms, including both enzymatic and nonenzymatic pathways [reviewed in (97,
139)l.
Early studies by Appleby (140) showed that bacteroids slowly reduce Fe(II1) Lb to Fe(I1) Lb under anaerobic conditions. It was suggested that a n enzymatic process was involved. Later reports [e.g.,
(141, 142)l confirmed that a specific enzymatic Fe(II1) Lb reduction
system is present in root nodules, and the nature of this enzymatic
525
action has been characterized in some detail (143). The protein responsible for the activity was initially characterized in lupin nodules
but has since been shown to be present in other nodules (e.g., soybean). It is a FAD-containing species that also contains active thiol
groups, but no catalytic metals (139, 143-147). The enzyme has been
purified in a number of cases. There are significant differences in size
and other properties of the enzyme from different species (e.g., the
molecular mass varies from 60 kDa for the lupin form to 100 kDa for
the soybean). In lupin nodules there are positive correlations between
activity of the enzyme and symbiotic performance as measured by the
rate of nitrogen fixation and the Lb content (148).The protein from
soybean root nodules is present primarily in the plant fraction of the
nodules (91% of the enzymatic activity), with only low levels present
in the bacteroids (143). The isolated protein appears to reduce all
eight forms of soybean Lb equally effectively. Several different forms
of this species have been identified in soybean nodules, suggesting
that there are a number of modified or different forms of the protein
present (149). The activity of this enzyme requires a reductant
(NADH or NADPH) and dioxygen but does not require an intermediate electron carrier. Activity is decreased if the Fe(II1) Lb is complexed with a number of small ligands such as acetate, nicotinate,
and nitrate, and is also inhibited by the presence of exogenous catalase, but not by superoxide dismutase. The latter observation suggests that a peroxide-type intermediate may be formed during the
catalytic cycle (97, 139). The rate of reduction does not appear to depend on the species that finally ligates Fe(I1) Lb; thus, identical rates
of reduction are observed with O2 or CO ligated in the axial position.
As with nonenzymatic reduction (discussed later) the enzymatic process is pH dependent. Rate constants are constant over the pH range
7.6 to 6.5, but increase at lower pH values, with a threefold increase
at pH 5 . 2 (143).Further studies and partial sequencing of the enzyme
have shown that this protein is similar to myoglobin reductase, and
the NADH : cytochrome b5 reductase from erythrocytes (147).
B. NONENZYMATIC
There are a number of low-molecular-weight species present in root
nodules that might act as reductants for Fe(1II) Lb. In mature root
nodules there are significant levels of NADH and NADPH (150-200
pM), free cysteine (200 pM), reduced glutathione (40-150 pM) and
reduced ascorbate (1-2 mM). The exact concentrations depend on the
age of the nodules and the method by which the extraction and quan-
526
tification is carried out (97, 139, 150). The reduction catalyzed by a

number of these species is pH dependent. Rate constants for the reduction of Fe(II1) Lb by high concentrations (1 mM) of NADH or
NADPH were found to increase with change in pH from pH 7.0 to 5.2,
with maximum rates observed at the lowest value. The reduction was
O2 dependent and was inhibited by superoxide dismutase and catalase (93, 151). These results implicate superoxide radicals andor hydrogen peroxide as the reductants. However, at higher pH and physiological levels of reductant (400 pM), the rate constants are very small
(151). Furthermore, only ascorbate and cysteine of these reductants
appear to be able to reduce Fe(II1) Lb a t significant rates (139). A
number of nonphysiological thiols also appear to be able to bring
about reduction. In a number of cases, including cysteine, competing
processes generating sulfur-Lb species have been shown to occur
(135).If such sulfur-Lb species could be detected in intact root nodules, it might give an indication of the occurrence and importance of
cysteine-mediated reduction of Fe(II1) [and Fe(1V)= 01 Lb in uiuo.
Disulfides do not appear to be capable of reducing Fe(II1) Lb to
Fe(I1) Lb (135). Reaction with sulfide itself appears to be different
and gives rise t o a Fe(II1) Lb-sulfide complex rather than reduction
or nucleophilic attack (135). The complex is, however, susceptible to
reduction by strong reducing agents such as dithionite and ascorbic
acid.
It has also been reported (152)that incubation of Fe(II1) Lb with a
large excess of ascorbic acid (10 mM) can result in the formation of
LbOz together with heme degradation products due to the generation
of radicals (from autoxidation of excess ascorbate). More recent work
has cast doubt on the significance of this type of reaction (134).
Rate constants for many of these reduction processes have been
shown to depend on a number of factors including pH, the presence
of metal ions, and cofactors. Thus, it has been reported that the reduction of Fe(II1) Lb by NADH can be stimulated by addition of Mn(I1)
ions (139).This is thought to be due to the formation of superoxide
radicals, which then reduce the Fe(II1) center. Flavins have been suggested as stimulators of Fe(II1) Lb reduction (139).Thus, in the presence of NADH or NADPH, free flavins can act as efficient reducing
agents in a process that appears to be independent of mediators such
as superoxide radicals (no effect of SOD). This direct reaction may be
aided by the exposed nature of the heme edge of Lb (Fig. 8) (23, 29).
This type of process does not require dioxygen, and may be a significant pathway in uiuo due t o the high flavin content of root nodules
(139).
527
IX. Reactions of Globin-Derived Radicals
A number of reactions of the globin radicals have been elucidated,

and these can be readily grouped into various categories.
A. REACTIONS WITH THE HEMERING

During the reaction of Fe(II1) Lb with hydrogen peroxide, the yield
of Fe(1V)= 0 Lb is relatively low compared t o reaction with the Fe(I1)
state. However a further species is formed, which can be detected by
the UV-vis absorption bands a t 410 and 630 nm (94, 102, 103). The
observation that this species is produced only under conditions in
which globin radicals are formed suggested that it results from subsequent reactions of the globin radical (153).The stability of this additional (green) species has allowed it to be isolated from reaction mixtures by isoelectric focusing using immobiline gels with a 5.7 to 4.7
pH gradient. The green compound has an isoelectric point of 5.45.
Electroelution has allowed this species to be further characterized,
and UV-vis and fluorescence spectra have been determined (153).The
visible spectrum is completely different from Fe1III) Lb in that the
Soret band 1410 nm) is much less intense and broader, and the
charge-transfer bands are dramatically altered, with a broad absorption a t 630 nm. The product reacts with strong reducing agents such
as dithionite or ascorbate (Fig. lo), which results in the formation of
a product with UV-vis bands similar to those of LbOL.This reduction
is slow and occurs over approximately 18 h (153).
The pyridine hemochromogen spectrum of the product provides
strong evidence for alterations to the heme ring but with no loss of
iron (153). The UV-vis spectrum is very different from other heme
degradation products such as biliverdin. SDS-PAGE analysis gives a
single band with similar R f to Fe(II1) Lb, suggesting that the protein
is not a fragment or dimer. The suggestion that it contains a modified
heme ring (and iron), has been confirmed by EPR spectra, which demonstrate the presence of high-spin Fe(II1) in a distorted environment.
The usual iron absorption a t g = 6 was not observed, but a sharp
peak a t g ca. 4.3 was detected together with a broad background signal. This may be due to an envelope of signals from completely free
or partially bound Fe(II1) (153).
The possibility that addition of a globin radical to the heme ring
gives an altered heme moiety has been confirmed by heme extraction experiments (using acidic 2-butanone) following treatment of soybean Fe(II1) Lb with a small excess of hydrogen peroxide. Measure-
528
Wavelength (nm)
FIG.10. Visible range spectral changes during the reduction of the additional (green)
species formed on reaction of Fe(I1I) Lb (50 pM) with H2O2(100 pM) at pH 7.4 (25 mM
KH,PO,/KOH buffer containing 0.1 mM DTPA) a t 25"C, with subsequent addition of 1
mM ascorbate. Scans were recorded 0, 1, 2, 6, and 18 h after the addition of ascorbate;
spectral changes are indicated by arrows. Data are representative of experiments carried out in triplicate [reprinted from Phytochernistry, 39, Moreau, s.;Puppo, A.; Davies,
M. J.; The reactivity of ascorbate with different redox states of leghaemoglobin, pp.
1281-1286, copyright (1995), with kind permission of Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington, OX5 lGB, UK].
ment of the ratio of the absorption of Soret band to that at 280 nm

was used to monitor heme removal from the holoprotein. With untreated Fe(II1) Lb, heme retention was of the order of 4.4% (i.e., nearquantitative removal), whereas with the green compound heme retention was calculated as ca. 63%. The green compound is therefore
believed to be a crosslinked species arising from reaction of a globin
radical with the heme (153).Similar species have been demonstrated
on reaction of myoglobin with hydrogen peroxide (154-156). Though
the exact nature remains to be confirmed, it has been suggested that
in the case of soybean the species arises via a mechanism involving
reaction of the Tyr133 phenoxyl radical with either one of the vinyl
groups of the heme moiety or a methylene bridge site; further reaction
of the initial adduct gives the stable green compound (153).Examination of molecular models of both the lupin and soybean forms of Lb
from the crystal coordinates [(23,29); (Ellis and Freeman, personal
529
communication)l show that the Tyr residue (138 and 133 respectively)
is aligned in such a way that the phenolate oxygen atom is in close
proximity to a vinyl group and methylene bridge (Fig. 6). This suggestion therefore seems very reasonable. Proteolytic digestion of the
green compound and sequence analysis of the fragments, in a manner
similar to that for the corresponding myoglobin heme-protein crosslinked species (154, 155), should allow the nature of this link to be
confirmed.
The heme-to-globin cross-linking process seems to be an inherent
feature of proteins in which a radical center is generated in the vicinity of the heme. The formation of such species has been suggested as
a marker of exposure to reaction with hydrogen peroxide in uivo (155,
156). It is interesting to note that it has long been known that exposure of functional root nodules to stress (e.g., absence of light) results
in the formation of green nodules (157). The green coloration may
be a result of the formation of hydrogen peroxide and its subsequent
reaction with Lb. It has been reported that at least two green pigments are formed on oxidative damage to Lb both in uitro and in uiuo
(12, 67, 157, 158). One of these, called cholegobin, originates from
oxidative attack on the ring without loss of iron (158), whereas biliverdins arise from heme degradation with concomitant loss of both
the iron and carbon monoxide (152, 159). Pigments that have these
characteristics have been partially purified from senescent pea nodules (12), but no further characterization has been reported. It has
also been reported that nodules from aging gram plants (Vicer arietinum) retain their pink color longer, and do not develop a green coloration until they are much older, when they are kept in a nutrient
medium containing ascorbic acid (37, 160). This result is in accord
with the protective effects of ascorbic acid against oxidative damage,
and with the suggestion that oxidative radical-mediated damage is of
importance in the aging of nodules in uiuo.
Further in uitro studies have been carried out, and it has been
shown that Fe(II1) Lb obtained from common bean (but not soybean)
can be reduced by glycine at alkaline pH with the formation of glyoxylate and hydrogen peroxide (159).The LbO, produced during this reaction is rapidly degraded to a green product that still contains iron
and has an absorption maximum at 697 nm. Although the significance of this process in uiuo remains to be established, it has obvious
parallels to the radical processes already described. It is also interesting to note that the formation of the green pigment can be stimulated
by pretreatment of the Lb with a proteolytic enzyme (carboxypeptidase) and is inhibited by superoxide dismutase and catalase (159).
530
The latter observations are in accord with the formation of the pigment from a hydrogen peroxide-mediated radical process. A similar
formation of green products has been observed during the reaction of
Fe(II1) Lb with ascorbate (10 mM). This process again results in the
transient formation of LbOz and rapid breakdown to biliverdin-like
materials (152). Analysis of the products has shown that in the case
of the pea and vetch Lbs, only a single major biliverdin isomer (the b
form) was detected, whereas with soybean and the common bean protein, mixtures of products were obtained [ca. 30% a isomer, 50% b,
and 20% d; (152)l.The difference between these Lb forms is of interest. The lack of specificity is indicative of random damage (as might
be envisaged from a free radical type of reaction, i.e., a low-molecular-weight rapidly diffusing molecule), whereas the highly specific
damage seen with pea and vetch is more suggestive of site-specific
damage induced by a bound oxidant (i.e., a radical centered on a particular residue). These suggestions relating to the type of radical (either free or protein bound) are in accord with the observation that
the formation of green products is inhibited by superoxide dismutase
and catalase (152).
B. REACTIONS
WITH LOW-MOLECULAR-WEIGHT
SPECIES
Spectroscopic evidence has been obtained for the reaction of the Tyr
phenoxyl radical derived from soybean Lb with a number of low-molecular-weight species. Thus, it has been demonstrated by EPR that
the globin radicals can react with ascorbic acid, cysteine, glutathione,
Trolox C (a water-soluble analog of vitamin El, thiourea, salicylate,
and desferal (desferrioxamine) (101,102,134).Other compounds such
as 2-deoxyribose and mannitol either do not react or react slowly. In
a number of cases radicals from these reactions could be detected,
for example the ascorbate radical, the cysteine and glutathione thiyl
radicals, the Trolox C phenoxyl radical, and the desferal nitroxide
species (101, 102, 134). In the case of glutathione, complete loss of
the phenoxyl radical was not observed, whereas with an equivalent
concentration of cysteine it was. This suggests that access to the radical by species present in bulk solution might be limited (102). The
accessibility problem is strongly supported by molecular modeling
(from the crystal coordinates) of the soybean protein, which shows
that the Tyr133 phenoxyl radical is only partially exposed on the protein surface (Fig. 11) (112).
It should be noted that with a number of these compounds, reaction
occurs with both the Fe(IV)=O heme center and the Tyr radical. Rel-
531
Fir;. 11. The Tyr133 residue (dark-shaded residues) in soybean Fe(II1) Lb is only
partially exposed on the protein surface. Computer-generated molecular model using
the crystal coordinates obtained ( 2 3 )(Ellis and Freeman, unpublished data).
atively little is known about the reactions of the other globin-derived

radicals (129).However, the stoichiometric conversion of ascorbate to
dehydroascorbate (1341 and reduced glutathione to GSSG (102) with
respect to the amount of hydrogen peroxide added suggests either
that the other globin-derived radicals are present a t low concentrations or, more probably, also react rapidly.
C. REACTIONS
WITH OTHER
LEGHEMOGLOBIN
MOLECULES
Examination by SDS-PAGE of reaction mixtures in which Fet 111)
Lb was incubated with a two-fold excess of hydrogen peroxide has
provided evidence for the formation of dimeric species (Fig. 12) (103).
It has been shown that the native protein migrates abnormally fast
and appears a t the bottom of the gel (103, 159). However a distinct
band a t ca. 32 kDa is observed, with little evidence for other species.
No dimerization was observed with apolb, confirming the role of the
heme group in this process. Confirmation of the dimerization process
was obtained using size-exclusion HPLC on columns previously calibrated with horse myoglobin ( M , ca. 16 kDa) and carbonic anhydrase
( M , 31.5 kDa). In contrast to untreated samples, in which a single
band with a retention time very similar to that observed with horse
532
FIG.12. Dimerization of soybean Lb and its inhibition by ascorbate, glutathione, and

tyrosine. Aliquots (30 p g of protein each) of incubations containing untreated 1 mM
soybean Lb (lane B), or Lb incubated for 2 h in the presence of 2 mM HzOzwithout
(lane C) or with 5 mM ascorbate (lane E), 5 mM glutathione (lane D) or 5 mM
p-tyrosine (lane F) were analyzed by SDS-PAGE. Molecular weight markers were in
lane A. The abnormally fast migration of monomeric Lb was evidenced by Western
blotting with polyclonal antibodies raised against purified soybean Lb, in the absence
and a blank was performed with egg
(lane G) or in the presence (lane H) of Haon,
lysozyme (MW 14.4 kDa) (lane I). The position of the monomeric Lb is indicated by the
arrow and that of the dimeric Lb by the double arrow [reprinted from Biochin. Biophys.
Actu, 1251, Moreau, S.; Davies, M. J.; Puppo, A.; Reaction of ferric leghemoglobin with
HzOz:
formation of heme-protein cross-links and dimeric species, pp. 17-22, copyright
(1995), with kind permission from Elsevier Science-NL, Sara Burgerhartstraat 25,
1055 KV Amsterdam, The Netherlands].
myoglobin was observed, the treated samples (two-fold excess of hydrogen peroxide over Fe(II1) Lb) gave two bands, a major band corresponding to the parent material and a weaker band that eluted with
approximately the same retention time a s carbonic anhydrase, consistent with the presence of a dimer (Fig. 13) (103).Integration of the
peak areas suggests that the dimer represents approximately 6% of
the initial Lb. Sufficient material could be obtained from such chromatographic separations to allow further studies to be carried out on
nature of the cross-linking. Unlike the situation with sperm whale
myoglobin, in which dimer formation is also observed (109, 161), no
evidence was obtained for the presence of dityrosine after acid hydrolysis of the dimeric material (103).Thus, unlike myoglobin the crosslinking does not appear to occur via the dimerization of two tyrosine
phenoxyl radicals. This is not unexpected because the phenoxyl radical site is only slightly exposed on the surface of the protein (Fig. 11)
(112),unlike the case of sperm whale myoglobin in which the Tyrl51
residue is very accessible (161).
The mechanism of formation of this dimer was further investigated
by examining the effect of various putative radical scavengers on the
intensity of the dimer band on SDS-PAGE gels (Fig. 12). These stud-
533
I li
1
10
20
30
40
Time (min)
FIG.13. High-pressure liquid chromatography of HzOz-treated Lb. Aliquots (100 pL)

of reaction mixtures containing 1 mM soybean Fe(II1) Lb incubated for 2 h in the preswere chromatographed. The trace marked (-)
shows the aromatic
ence of 2 mM H202
amino acid absorbance a t 280 nm; the peak marked with a n arrow is that of the dimer.
Control experiments were performed with Lb alone
[reprinted from Biochim. Biophys. Actu, 1251, Moreau, S.; Davies, M. J.; Puppo, A.; Reaction of Fe(II1) leghemoglobin
with H202:formation of heme-protein cross-links and dimeric species, pp. 17-22, copyright (1995),with kind permission from Elsevier Science-NL, Sara Burgerhartstraat
25, 1055 KV Amsterdam, The Netherlands].
ies have demonstrated that ascorbate completely inhibits dimer formation, and glutathione and para-tyrosine also significantly inhibit
its generation. In contrast, ortho- and rneta-tyrosine have little effect.
The reason is not immediately apparent, but may be due to steric
factors. The results are in accord with dimerization being a radical
process, but not via dimerization of two tyrosine-derived phenoxyl
radicals (103).The cross-linking may involve dimerization of the reactive radicals detected in spin-trapping experiments (129).
D. REACTIONS
WITH MEMBRANES
A number of studies have examined the possible role of both
Fe(1V)=0 and globin radicals in the initiation of membrane peroxidation. Thus, it has been reported that although reaction of Fe(II1) Lb
with hydrogen peroxide (two-fold excess) initiates peroxidation of peribacteroid membranes from French beans (as measured by the yield
of malondialdehyde), similar reactions with Fe(I1) Lb and hydrogen
peroxide did not give a similar effect (162). Little oxidation was ob-
534
served in the absence of hydrogen peroxide, whereas much higher levels of oxidative breakdown products are observed with a 5- or 10-fold
excess. In the latter, however, it is known that significant release of
iron occurs, and it is possible that the lipid oxidation is due to the
reaction of the iron with excess peroxide (99). The observation that
with Fe(II1) Lb and a low excess of peroxide, damage is initiated,
whereas none is observed with Fe(I1) Lb [which is a more efficient
method of generating F e W ) =01, suggests that Fe(IV)=0 is not able
to initiate damage to membrane fractions (162). This is not entirely
unexpected because Fe(1V)=0 would not be normally expected t o
have ready access to membrane lipids (or other membrane components) due to the surrounding protein. Recent studies with symbiosomes from French beans have supported the suggestion that
Fe(1V)=0 alone cannot initiate lipid oxidation (163).Isolated symbiosomes exposed to Fe(II1) Lb that had been preincubated with a twofold excess of hydrogen peroxide did not appear to suffer damage as
monitored by the uptake of succinate (a useful marker of symbiosome
membrane integrity). Enhanced uptake of succinate was observed,
however, with six-fold excesses of hydrogen peroxide when iron release is likely to have occurred (163).
A number of groups have suggested that metal ions derived from
Lb and/or trace transition-metal ions can play a role in initiating or
propagating damage (162-1641, but the mechanism by which this occurs is poorly understood. It is possible that the release of iron from
the heme protein (possibly as a result of oxidative damage) or from a
build-up of trace-metal ions in old nodules by other routes, may be
responsible for the catalysis of radical generation, though the details
remain uncertain. Recent reports have demonstrated that low-molecular-weight chelates of transition-metal ions are present in old nodules (162, 1641, and it has been suggested that hydroxyl radicals may
be a key species (164).It should be noted, however, that specific products of hydroxyl radical attack were not identified, and it is possible
that the oxidation of dimethyl sulfoxide observed may be due to the
formation of other radicals. Other studies have also reported that hydroxyl radicals are not involved (162).Previous studies using fluorescence quenching (165) have suggested that Lb can interact with peribacteroid membranes, possibly via specific binding sites. The presence
of such receptors might be expected to facilitate damage induced by
either protein radicals or high-oxidation-state species.
In the light of the evidence reviewed here that at least some of the
globin radicals formed on reaction of Fe(II1) Lb with hydrogen peroxide are present on the surface of the protein and that these species
535
can react both with low-molecular-weight species and other protein

molecules, it was of interest to determine whether these species could
transfer damage to a membrane (i.e., initiate lipid oxidation). Of particular relevance is whether the Lb-derived radicals could initiate
damage to the peribacteroid membrane that surrounds the microsymbiont in uiuo (Fig. 2) and whether such reactions play a role in the
loss of nitrogen-fixing ability of nodules and nodule senescence. The
latter membrane appears to be more sensitive to oxidation than others (e.g., microsomes prepared from identical nodules) (1621, possibly
as a result of the different compositions of these structures (e.g., the
high lipid : protein ratio of the peribacteroid membrane).
Incubation of purified peribacteroid membranes (PBMs) with soybean Fe(II1) Lbhydrogen peroxide and analysis of the extent of Lb
dimerization by SDS-PAGE have demonstrated that the presence of
the PBMs decreases (in a dose-dependent manner) the extent of protein dimerization (129) (Fig. 14). This has been interpreted as evidence for interaction of the radicals involved in the dimerization reaction with the PBMs. Further evidence for such a process has been
obtained from examining the intensity of the signal (which is directly
proportional to the radical concentration) of the Tyr-derived phenoxyl
radical generated from soybean Fe(II1) Lb and Hz02in the presence
of added PBMs by stopped-flow EPR spectroscopy. Inclusion of the
PBMs (>0.3 mg proteidml) resulted in a statistically significant decrease in concentration of the phenoxyl radical, consistent with scavenging by membrane components (129). Such a process would be expected to result in the formation of one or more radical species in the
membrane fractions.
FIG. 14. Quenching by peribacteroid membranes (PBMs) of the dimerization process

of Fe(II1) Lb in the presence of H p 0 2 .Fe(II1) Lb (210 pM) was incubated with H,O,
(420 pM) for 1 h in the absence (lane 2) or in the presence of PBMs at the following
protein concentrations: 0.23 mg/mL (lane 3), 1.16 mg/mL (lane 4) and 2.33 mg/mL (lane
5). Lane 1 represents Fe(II1) Lb alone and lane 6 and 7 PBMs plus H2O2and PBMs
alone, respectively (reprinted with permission from: Moreau, S.; Davies, M. J.; Mathieu,
C.; Herouart, D.; Puppo, A. J. Biol. Chern., 1996,271, 32557-32562).
536
Inclusion of PBM fractions in Fe(II1) Lbhydrogen peroxide spintrapping experiments has also provided evidence for such interactions. Thus when PBMs (ca. 3 mg proteidml) were added to an equimolar amount of soybean Fe(II1) Lb and hydrogen peroxide in the
presence of the spin trap PBN, significant changes were observed in
the EPR spectra as compared to control spectra in the absence of the
membrane fractions (Fig. 15) (129). The additional features present
are consistent with the presence of large, slowly tumbling radical adduct species and have been assigned to membrane-derived radicals.
FIG. 15. EPR spectra observed on reaction of soybean Fe(II1) Lb (500 pM) with HaOa
(1 mM) and the spin trap PBN (46 mM) in the absence or presence of peribacteroid
membranes (PBMs). (a) Incubation carried out for 2 h at 29C in the absence of added
membrane fractions; signal assigned to one (or more) large protein-derived radical adducts to the spin trap. (b) Incubation carried out for 2 h at 29C in the presence of
added membrane fractions (2.95 mg proteidml); signal assigned to one (or more) large
protein-derived radical adducts to the spin trap together with a further species (arrowed features) believed to arise from a radical generated from the membrane fraction.
(c) As (b) but spectrum of the organic layer obtained after extraction of the membrane
fractions with an equal volume of Nz-degassed toluene; signal assigned to a t least one
lipid-soluble radical adduct arising from radical induced damage to the membrane fraction (reprinted with permission from: Moreau, S.; Davies, M. J.; Mathieu, C.; Herouart,
D.; Puppo, A. J. Biol. Chem., 1996,271,32557-32562).
537
To obtain further information on the nature of these intermediates,

solvent extraction experiments were carried out. When toluene was
added, the radical adducts were observed only in the organic phase.
This is consistent with the presence of lipid-derived radicals (or other
hydrophobic species). The signals observed from the toluene extracts
are considerably more mobile than those seen in the aqueous system,
suggesting that the slow tumbling is due to the formation of a lipid
radical in the membrane, which is prevented from rotating rapidly by
the membrane. The sharp (isotropic) nature of the signals in these
organic extracts has allowed further fine structure of the EPR signals
to be observed. The hyperfine coupling constants are consistent with
a lipid-derived species (129).
There is now considerable evidence for an interaction of the globinderived radicals with the peribacteroid membrane resulting in a
transfer of the radical into the membrane (129). The membranederived radicals would be expected to participate in chain reactions
in the presence of oxygen, and this would result in peroxidative damage to the membrane. Lipid peroxidation reactions are known to result in widespread alterations to membrane structure and can lead to
irreversible loss of function. Such reactions may play a role in the
disruption of the peribacteroid membrane during nodule senescence,
when conditions for peroxide formation and subsequent radical reactions would be expected to be favorable. An obvious corollary is that
species that protect against the formation of Fe(III1, F e W ) =0 and
globin radicals might inhibit the reactions and hence nodule degradation. These may include nonenzymic reactions (e.g., with ascorbate,
cysteine, and glutathione), all of which are present in nodules at high
concentrations, and enzymatic reactions brought about by superoxide
dismutase, catalase, Fe(II1) Lb reductase, and possibly other enzymes. Manipulation of the levels of activity of these species is an
obvious potential strategy for increasing the lifetime and activity of
the nitrogen-fixing processes. Some evidence suggests that this hypothesis is indeed correct-for example, the observation that the nodules from aging gram plants remain pink for longer periods in the
presence of added ascorbate in the growth medium (160).Although
significant protection by addition of exogenous materials such as
ascorbate and thiols is unlikely to be of any great practical utility,
the manipulation by genetic means of the enzymes that regulate the
production of these low-molecular-weight materials and the concentration and activity of the protective enzymes within the host plant
and bacteroid may be realistic and achieveable goals. Work toward
these targets is already under way.
538
ACKNOWLEDGMENTS
Work reported in this review has been supported by the European Union Human
Capital and Mobility Program (contract CHRX-CT94-0605),the Alliance Program, and
the Australian Research Council. We are also extremely grateful to all our co-workers
for their very positive contributions, to Prof. H. C. Freeman and Dr. P. J . Ellis (School
of Chemistry, University of Sydney, Australia) for providing the coordinates of soybean
Lb, prior to publication, to Dr. C. A. Appleby for stimulating advice and comments, and
to Mrs. G. Van der Sype for help with the photography.
REFERENCES
1. Hardison, R. C. Proc. Natl. Acad. Sci. U.S.A. 1996,93, 5675.
2. Kubo, H. Acta Phytochim. (Tokyo) 1939,11, 195.
3. Appleby, C. A.; Tjepkema, J. D.; Trinick, M. J. Science 1983,220, 951.
4 . Sharifi, E. Biosystems 1983, 16, 269.
5. Beringer, J. E.; Brewin, N.; Johnston, A. W.; Schulman, H. M.; Hopwood, D. A.
Proc. Roy. SOC.Lond. B Biol. Sci. 1979,204, 219.
6. Legocki, R. P.; Verma, D. P. Cell 1980,20, 153.
7. Bergersen, F. J. Root Nodules of Legumes: Structure and Functions; Research
Studies Press, John Wiley and Sons Ltd.: London, 1982.
8. Tjepkema, J. D. Can. J. Bot. 1983, 63, 2924.
9. Appleby, C. A.; Bogusz, D.; Dennis, E. S.; Peacock, W. J. Plant Cell Enuiron. 1988,
11, 359.
10. Anderson, C. R.; Jensen, E. 0.;
Llewellynm, D. J.; Dennis, E. S.; Peacock, W. J.
Proc. Natl. Acad. Sci. U.S.A. 1996,93, 5682.
11. Hyldig-Nielsen, J.; Jensen, E. 0.;Paludan, K.; Wiborg, 0.;Garrett, R.; Jorgensen,
P.; Marcker, K. A. Nucleic Acid Res. 1982, 10, 689.
12. Appleby, C. A. Annu. Rev. Plant Physiol. 1984, 35, 443.
13. Appleby, C. A.; Dennis, E. S.; Peacock, W. J . Aust. J. Syst. Biol. 1990,3, 81.
14. Wittenberg, J. B. J . Biol. Chem. 1974,249, 4057.
15. Appleby, C. A. Sci. Prog. 1992, 76, 365.
16. Zufferey, R.; Preisig, 0.;
Hennecke, H.; Thony-Meyer, L. J. Biol. Chem. 1996,
271, 9114.
17. Ellfolk, N.; Sievers, G. Acta. Chem. Scand. 1972,26, 1155.
18. Fuchsman, W. H. Adu. Comp. Environ. Physiol. 1992, 13, 23.
19. Whittaker, R. G.; Lennox, S.; Appleby, C. A. Biochem. Int. 1981,3, 117.
20. Fuchsman, W. H.; Appleby, C. A. Biochim. Biophys. Acta. 1979,579, 314.
21. Gibson, Q.H.; Wittenberg, J. B.; Wittenberg, B. A,; Bogusz, D.; Appleby, C. A. J.
Biol. Chem. 1989,264, 100.
22. Fuchsman, W. H. Arch. Biochem. Biophys. 1986,243, 454.
23. Ollis, D. L.; Appleby, C. A,; Colman, P. M.; Cutten, A. E.; Mitchell Guss, J.; Venkatappa, M. P.; Freeman, H. C. Aust. J . Chem. 1983,36, 451.
24. Vainshtein, B. K.; Harutyunyan, E. H.; Kuranova, I. P.; Borisov, V. V.; Sosfenov,
N. I.; Pavlovsky, A. G.; Grebenko, A. I.; Konareva, N. V. Nature 1976,254, 163.
25. Vainshtein, B. K.; Andreeva, N. S. Mol. Biol. Mosk. 1977, 1 1 , 1258.
26. Arutiunian, E. G . Mol. Biol. Mosk. 1981, 15, 27.
539
27. Arutyunyan, E. G.; Deisenhofer, I.; Teplyakov, A. V.; Kuranova, I. P.; Obmolova,
G. V.; Vainshtein, B. K. Dokl. Akad. Nauk. Ser. Biokhim. 1983,270, 732.
28. Arutyunyan, E. G.; Safanova, T. N.; Obmolova, G. V.; Teplyakov, A. V.; Popov,
A. N . ; Rusakov, A. A.; Rubinsky, S. V.; Kuranova, I. P.; Vainshtein, B. K. Bioorg.
Khim. 1990, 16, 293.
29. Harutyunyan, E. H.; Safonova, T. N.; Kuranova, I. P.; Popov, A. N.; Teplyakov,
A. V.; Obmolova, G. V.; Rusakov, A. A,; Vainshtein, B. K.; Dodson, G. G.; Wilson,
J. C.; Perutz, M. F. J. Mol. Biol. 1996,251, 104.
30. Harutyunyan, E. H.; Safonova, T. N.; Kuranova, I. P.; Popov, A. N.; Teplyakov,
A. V.; Obmolova, G . V.; Vainshtein, B. K.; Dodson, G. G.; Wilson, J. C. J. Mol.
B i d . 1996,264, 152.
31. Brown, G. G.; Lee, J. S.; Brisson, N.; Verma, D. P. J . Mol. Euol. 1984,21, 19.
32. Ellfolk, N.; Sievers, G. Acta. Chem. Scand. 1965, 19, 2409.
33. Perttila, U.; Sievers, G. Biochim. Biophys. Acta. 1980, 624, 316.
34. Kuranova, I. P.; Lukianova, E. V. Dokl. Akad. Nauk SSSR 1983,269, 1503.
35. Dalvit, C.; Tennant, L.; Wright, P. E. J . Inorg. Biochem. 1986,28, 303.
36. Ikeda-Saito, M.; Hori, H.; Inubushi, T.; Yonetani, T. J . Biol. Chem. 1981, 256,
10267.
37. Appleby, C. A. The Biology of Nitrogen Fixation; Quispel, A., Ed.; North-Holland
Pub. Co.: Amsterdam & Oxford, 1974, 522-554.
38. Verma, D. P.; Kazazian, V.; Zogbi, V.; Bal, A. K. J. Cell. B i d . 1978, 78, 919.
39. Udvari, M. K.; Price, G. D.; Gresshoff, P. M.; Day, D. A. FEBS Lett. 1988,231, 36.
40. Herrada, G.; Puppo, A.; Rigaud, J. J. Gen. Microbiol. 1989, 135, 3165.
41. Moreau, S.; Meyer, J.-M.; Puppo, A. FEBS Lett. 1995,361, 225.
42. Bergersen, F. J.; Appleby, C. A. Planta 1981, 152, 534.
43. Verma, D. P.; Ball, S.; Guerin, C.; Wanamaker, L. Biochemistry 1979, 18, 476.
44. OBrian, M. R.; Kirshbom, P. M.; Maier, R. J. Proc. Natl. Acad. Sci. U.S.A. 1987,
84, 8390.
45. Keithley, J. H.; Nadler, K. D. J. Bacteriol. 1983, 154, 838.
46. Dimitrijevic, L.; Puppo, A.; Trinchant, J. C.; Rigaud, J. J . Plant Physiol. 1989,
134, 642.
47. Noel, K. D.; Stacey, G.; Tandon, S. R.; Silver, L. E.; Brill, W. J. J . Bacteriol. 1982,
152, 485.
48. OBrian, M. R.; Kirshbom, P. M.; Maier, R. J. J . Bacteriol. 1987, 169, 1089.
49. Bisseling, T.; van, S. J.; Houwaard, F. Biochim. Biophys. Acta. 1980, 610, 360.
50. Kijne, J. W. Physiol. Plant Pathol. 1976, 7, 17.
51. Uheda, E.; Syono, K. Plant Cell Physiol. 1982,23, 75.
52. Appleby, C. A.; Bradbury, J. H.; Morris, R. J.;Wittenberg, B. A.; Wittenberg, J. B.;
Wright, P. E. J. B i d . Chem. 1983,258, 2254.
53. Martin, K. D.; Saari, L.; Wang, G. X.; Wang, T.; Parkhurst, L. J.; Klucas, R. V. J.
Bid. Chem. 1990,265, 19588.
54. Fuchsman, W. H.; Barton, C. R.; Stein, M. M.; Thompson, J. T.; Willett, R. M.
Biochem. Biophys. Res. Commun. 1976, 68, 387.
55. Rohlfs, R. J.; Olson, J. S.; Gibson, Q. H. J. Bid. Chem. 1988, 263, 1803.
56. Rohlfs, R. J.;Matthews, A. J.; Carver, T. E.; Olson, J. S.; Springer, B. A.; Egeberg,
K. D.; Sligar, S. G. J . Biol. Chem. 1990,265, 3168.
57. Arutiunian, E. G. Dokl. Akad. Nauk. SSSR 1980,252, 1264.
58. Irwin, M. J.;Armstrong, R. S.; Wright, P. E. FEBS Lett. 1981, 133, 239.
59. Rousseau, D. L.; Ondrias, M. R.; LaMar, G. B.; Smith, K. M. J. Biol. Chem. 1983,
258, 1740.
540
60. Ollis, D. L.; Wright, P. E.; Pope, J. M.; Appleby, C. A. Biochemistry 1981,20,587.
Appleby, C. A. Biochemistry 1979,18,1309.
61. Fuchsman, W. H.;
62. Ray, G. B.; Li, X.-Y.; Ibers, J. A.; Sessler, J. L.; Spiro, T. G. J. Am. Chem. SOC.
1994,116,162.
63. Lim, M.; Jackson, T. A.; Anfinrud, P. A. Science 1995,269,962.
64. Quillin, M. L.; Arduini, R. M.; Olson, J. S.; Phillips, G. N. J. J . Mol. Biol. 1993,
234,140.
65. Quillin, M. L.; Tiangsheng, L.; Olson, J. S.; Phillips, G. N. J.; Yi, D.; Ikeda-Saito,
M.; Regen, R.; Carlson, M.; Gibson, Q. H.; Haiying, L.; Elber, R. J. Mol. Biol. 1995,
245,416.
66. Moore, J. N.; Hansen, P. A.; Hochstrasser, R. M. Proc. Natl. Acad. Sci. U.S.A.
1988,85,5062.
67. Virtanen, A. I. Biol. Rev. 1947,22, 239.
68. Maskall, C . S.;Gibson, J. F.; Dart, P. J. Biochem. J. 1977,167,435.
69. Kanayama, Y.; Yamamoto, Y. Plant Cell Ph.ysio1. 1990,31,207.
70. Kanayama, Y.; Watanabe, I.; Yamamoto, Y. Plant Cell Physiol. 1990,31,341.
71. Dean, J. V.;Harper, J. E. Plant Physiol. 1988,82,718.
72. Fox, J. B. J.; Thomson, J. S. Biochemistry 1983,2,465.
73. Cueto, M.;Hernandez-Perera, 0.;Martin, R.; Bentura, M.; Rodrigo, J.; Lamas, S.;
Golvano, M. P. FEBS. Lett. 1996,398,159.
74. Mathieu, C.; Moreau, S.; Frendo, P.; Puppo, A.; Davies, M. J . Free Rad. Biol. Med.
1998,24,1242.
75. Cebolla, A.; Palomares, A. J. Microbiologia 1994,10,371.
76. Gilles-Gonzalez, M. A.; Gonzalez, G.; Perutz, M. F.; Kiger, L.; Marden, M. C.;
Poyart, C.Biochemistry 1994,33,8067.
77. Winkler, W. C.; Gonzalez, G.; Wittenberg, J. B.; Hille, R.; Dakappagari, N.; Jacob,
J.; Gonzalez, L. A.; Gilles-Gonzalez, M. A. Chem. Biol. 1996,3,841.
78. Pieve, Y.V.; Atanasov, B. P.; Zhiznevskaya, G. Y.; Krasnobaeve, N. N. Dokl. Akad.
Nauk. SSSR 1972,202,965.
79. Abel, K.Phytochem. 1983,2,429.
80. Ewing, G. J.; Ionescu, L. G. J. Phys. Chem. 1972,76,591.
81. Stetzkowski, F.; Cassoly, R.; Banerjee, R. J . Biol. Chem. 1979,254,11351.
82. Keilin, D.; Smith, J. D. Nature 1947,159,692.
83. Wittenberg, B. A.; Wittenberg, J. B.; Appleby, C. A. J. Biol. Chem. 1973,248,3178.
84. Ellfolk, N.Acta Chem. Scand. 1961, 15, 975.
85. Ehrenberg, A.;Ellfolk, N. Acta Chen. Scand. 1983, 17, S343.
86. Trewhella, J.; Wright, P. E. Biochem. Biophys. Res. Commun. 1979,88,713.
87. Job, D.; Zeba, B.; Puppo, A.; Rigaud, J. Eur. J. Biochem. 1980, 107,491.
88. Appleby, C. A.; Wittenberg, B. A,; Wittenberg, J. B. Proc. Natl. Acad. Sci. U.S.A.
1973,70,564.
89. Appleby, C . A.; Wittenberg, B. A.; Wittenberg, J. B. J. Biol. Chem. 1973,248,3183.
90. Appleby, C. A.; Blumberg, W. E.; Peisach, J.; Wittenberg, B. A.; Wittenberg, J. B.
J. Biol. Chem. 1976,251,6090.
91. Appleby, C. A. Biochim. Biophys. Acta. 1969,189,267.
92. Rigaud, J.; Puppo, A. Biochim. Biophys. Acta. 1977,497,702.
93. Puppo, A,; Rigaud, J.; Job, D. Plant Sci. Lett. 1981,22,353.
94. Aviram, I.; Wittenberg, A.; Wittenberg, J. B. J . Biol. Chem. 1978,253, 5685.
95. Henderson, R. W.; Appleby, C. A. Biochim. Biophys. Acta. 1972,283,187.
96. Nash, D. T.; Schulman, H. M. Biochem. Biophys. Res. Commun. 1976,68,781.
97. Becana, M.; Klucas, R. V. Plant Physiol. 1992,98,1217.
541
98. Pladys, D.; Barthe, P.; Rigaud, J. Plant Sci. 1988,56, 99.
99. Puppo, A.; Halliwell, B. Planta 1988, 173, 405.
100. Sadrzadeh, S. M.; Graf, E.; Panter, S. S.; Hallaway, P. E.; Eaton, J. W. J. Biol.
Chem. 1984,259,14354.
101. Davies, M. J.; Puppo, A. Biochem. J. 1992,281, 197.
102. Puppo, A,; Monny, C.; Davies, M. J. Biochem. J. 1993,289, 435.
103. Moreau, S.; Davies, M. J.; Puppo, A. Biochim. Biophys. Acta. 1995, 1251, 17.
104. George, P.; Irvine, D. H. Biochem. J . 1952,52, 511.
105. King, N. K.; Looney, F. D.; Winfield, M. E. Biochim. Biophys. Actu 1967,133, 65.
106. Yonetani, T.; Schleyer, H. J . Biol. Chem. 1967, 242, 1974.
107. Davies, M. J. Biochim. Biophys. Acta 1991, 1077, 86.
108. Gunther, M. R.; Kelman, D. J.; Corbett, J. T.; Mason, R. P. J. Biol. Chem. 1995,
270, 16075.
109. Tschirret-Guth, R. A.; Ortiz de Montellano, P. R. Arch. Biochem. Biophys. 1996,
335, 93.
110. DeGray, J. A,; Gunther, M. R.; Tschirret-Guth, R.; Ortiz de Montellano, P. R.;
Mason, R. P. J. Biol. Chem. 1997,272, 2359.
111. Sealy, R. C.; Harman, L.; West, P. R.; Mason, R. P. J . Am. Chem. SOC.1985,
107, 3401.
112. Davies, M. J.; Puppo, A. Biochim. Biophys. Acta 1993, 1202, 182.
113. Larsson, A.; Sjoberg, B. M. EMBO J. 1986, 5, 2037.
114. Innes, J. B.; Brudvig, G. W. Biochemistry 1989,28, 1116.
115. Fontecave, M.; Gerez, C.; Atta, M.; Jeunet, A. Biochem. Biophys. Res. Chem. 1990,
168, 659.
116. Kulmacz, R. J.; Ren, Y.; Tsai, A. L.; Palmer, G. Biochemistry 1990,29, 8760.
117. Nordlund, P.; Sjoberg, B. M.; Eklund, H. Nature 1990, 345, 593.
118. Whittaker, M. M.; Whittaker, J. W. J . Biol. Chem. 1990,265, 9610.
119. Hoganson, C. W.; Babcock, G. T. Biochemistry 1992,31, 11874.
120. Kulmacz, R. J.; Palmer, G.; Tsai, A. L. J. Lipid Mediat. 1993, 6, 145.
121. Nordlund, P.; Eklund, H. J . Mol. Biol. 1993, 232, 123.
122. Pedersen, J. 2.; Finazzi, A. A. FEBS Lett. 1999,325, 53.
123. Ormo, M.; Regnstrom, K.; Wang, Z.; Que, L. J.;Sahlin, M.; Sjoberg, B. M. J. B i d .
Chem. 1995,270,6570.
124. Sun, X.; Ollagnier, S.; Schmidt, P. P.; Atta, M.; Mulliez, E.; Lepape, L.; Eliasson,
R.; Graslund, A,; Fontecave, M.; Reichard, P.; Sjoberg, B. M. J. Biol. Chem. 1996,
271, 6827.
125. von Sonntag, C. The Chemical Basis of Radiation Biology; Taylor and Francis:
London, 1987.
126. Davies, M. J.; Dean, R. T. Radical-Mediated Protein Oxidation: From Chemistry
to Medicine; Oxford University Press: Oxford, 1997.
127. Marriott, P. R.; Perkins, M. J.; Griller, D. Can. J . Chem. 1980,58, 803.
128. Buettner, G. R. Free Rad. Biol. Med. 1987,3, 259.
129. Moreau, S.; Davies, M. J.; Mathieu, C.; Herouart, D.; Puppo, A. J. Biol. Chem.
1996,271, 32557.
130. Davies, M. J.; Gilbert, B. C.; Haywood, R. M. Free Rad. Res. Commun. 1991,
15, 111.
131. Davies, M. J . Res. Chem. Intermed. 1993, 19, 669.
132. Dean, R. T.; Fu, S.; Stocker, R.; Davies, M. J. Biochem. J . 1997,324, 1.
133. Barr, D. P.; Gunther, M. R.; Deterding, L. J.; Tomer, K. B.; Mason, R. P. J. Biol.
Chem. 1996,271, 15498.
542
134.
135.
136.
137.
138.
Moreau, S.; Puppo, A,; Davies, M. J. Phytochem. 1996,39, 1281.

Puppo, A.; Davies, M. J. Biochim. Biophys. Acta 1996, 1246, 74.
Romero, F. J.; Ordonez, I.; Arduini, A.; Cadenas, E. J. Biol. Chem. 1992,267,1680.
Berzofsky, J. A,; Peisach, J.; Blumberg, W. E. J. Biol. Chem. 1971,246, 3367.
Mathieu, C.; Swaraj, K.; Davies, M. J.; Trinchant, J.-C.; Puppo, A. Free Rad. Res.
1997,27, 165.
139. Becana, M.; Klucas, R. V. Proc. Natl. Acad. Sci. U.S.A. 1990,87, 7295.
140. Appleby, C. A, Biochim. Biophys. Acta 1969, 188, 222.
141. Kretovich, V. L.; Melik-Sarkissyan, S. S.; Bashirova, N. F.; Topunov, A. F. J. Appl.
Biochem. 1982,4, 209.
142. Topunov, A. F.; Melik, S. S.; Lysenko, L. A.; Iarpilenko, G. P.; Kretovich, V. L.
Biokhimiia 1980,45, 2053.
143. Saari, L. L.; Klucas, R. V. Arch. Biochem. Biophys. 1984,231, 102.
144. Puppo, A,; Rigaud, J.; Job, D.; Ricard, J.; Zeba, B. Biochim. Biophys. Acta 1980,
614,303.
145. Golubeva, L. I.; Topunov, A. F.; Talyzin, V. V.; Kretovich, V. L. Dokl. Akad. Nauk.
S S S R 1987,294, 178 (Engl. Trans.).
146. Golubeva, L. I.; Topunov, A. F.; Goncharova, S. S.; Aseeva, K. B.; Kretovich, V. L.
Biokhimioa 1988, 53,~1478(Engl. Trans.).
147. Ji, L.; Wood, S.; Becana, M.; Klucas, R. V. Plant Physiol. 1991,96, 32.
148. Swaraj, K.; Topunov, A. F.; Golubeva, L. I.; Kretovich, V. L. Fiziol. Rust. 1986,
33, 87.
149. Puppo, A.; Rigaud, J . FEBS. Lett. 1979, 108, 124.
150. Dalton, D. A,; Russell, S. A.; Hanus, F. J.; Pascoe, G. A.; Evans, H. J . Proc. Natl.
Acad. Sci. U.S.A. 1986, 83, 3811.
151. Saari, L. L.; Klucas, R. V. Biochim. Biophys. Acta 1987,912, 198.
152. Lehtovaara, P.; Perttila, U. Biochem. J. 1978, 176, 359.
153. Moreau, S.; Davies, M. J.; Puppo, A. Biochim. Biophys. Acta 1995, 1251, 17.
154. Catalano, C. E.; Choe, Y. S.; Ortiz de Montellano, P. R. J. Biol. Chem. 1989,
264, 10534.
155. Osawa, Y.; Williams, M. S. Free. Rad. Biol. Med. 1996,21, 35.
156. Sugiyama, K.; Highet, R. J.; Woods, A.; Cotter, R. J.; Osawa, Y. Proc. Natl. Acad.
Sci. U.S.A. 1997,94, 796.
157. Virtanen, A. I. Nature 1946,155, 747.
158. Lemberg, R.; Legge, J. W. Hematin Compounds and Bile Pigments; Interscience
Publishers Inc.: New York, 1949.
159. Lehtovaara, P. Biochem. J. 1978, 176, 351.
160. Swaraj, K.; Garg, 0. P. Plant Physiol. 1969,23, 889.
161. Tew, D.; Ortiz de Montellano, P. R. J. Biol. Chem. 1988,263, 17880.
162. Puppo, A,; Herrada, G.; Rigaud, J. Plant Physiol. 1991, 96, 826.
163. Herrada, G.; Puppo, A.; Moreau, S.; Day, D. A,; Rigaud, J. FEBS. Lett. 1993,
326, 33.
164. Becana, M.; Klucas, R. V. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 8958.
165. Zhiznevskaya, G. Y.; Borodenko, L. I.; Mekler, V. I.; Belonogova, 0. V.; Kotelnikov,
A. I.; Izmailov, S. F. Russ. J. Plant Physiol. 1994,41, 69.
166. Gibson, Q. H.; Olson, J. S.; McKinnie, R. E.; Rohlfs, R. J. J. Biol. Chem. 1986,
261, 10228.
167. Olson, J. S.; Rohlfs, R. J.; Gibson, Q. H. J. Biol. Chem. 1987, 262, 12930.
INDEX
Actinides, xenon fluoride reactions, 91

Alcohol oxidation, molybdenum chloride
clusters, 19
Alkenes
chlorination, 115-117
chlorination, manganese complex applications, 403-405
epoxidation, manganese complex applications, 395-401
Aluminum trihydride, 121-122
Americium, xenon fluoride reactions, 91
Ammonia, fluorine gas reaction, 117
Amyotrophic lateral sclerosis (ALS),peroxynitrite implicated, 402
Annexins, calcium binding in, 457
Antimony complexes, xenon fluoride reactions, 86, 88
Apo-calcyclin, 455
Apo-calmodulin, 449-450
Arginase, 306-307
binuclearity, 325-327
EPR spectroscopy, 389
hydrogen peroxide disproportionation,
326-327
manganese requirement, strict, 307,
424
Argon chemistry, 52, 54
Aromatic compounds, dioxygenase decomposition, 312-324
Arsenic complexes, xenon fluoride reactions, 86, 88
Arteriosclerosis, peroxynitrite implicated,
311
Ascorbate, leghemoglobin reaction,
Fe(IV)=O, 520-521
Aziridination, manganese complex applications, 402-403
Bacteria
manganese catalase, 323-325
manganese(I1) dioxygenase, 312, 314
manganese ribonucleotide reductase,
319, 321
manganese superoxide dismutase,
3 10-3 12
BEDT-TTF-based supramolecular complexes, 195
Benzene-iodine complex, charge-transfer
reaction, 146- 148
Benzothiadiazole-bridged supramolecular
complexes, 214-219
Biliverdins, 529
Binuclear manganese redox enzymes, see
Manganese redox enzymes, binuclear
Bisnitrile-bridged supramolecular complexes, 243-246
Blood coagulation factors, 442, 466-470,
474-477,479-481
Blood plasma, calcium concentration, 470
RM-40 ISPARC), 484
Boron hydrides, reactive intermediate
studies, 110
Boron nitride, reactive intermediate studies, 108-109
Bradyrhizobium .japonicum, leghemoglobin in, 499-500
t-Butylgallium sulfide, thermal decomposition, 112-113
C
CaBP (Calcyclin), 454-456
Cadherins, calcium binding in, 483-484
Calbindin D9k, 452-453
stability, 444
543
544
INDEX
Calcium
biological coordination chemistry, 442
biological roles, 441-442
mixed-metal active site, concanavalin
A, 308
oxygen-evolving complex requirement,
328, 338-339
resting cell concentration, 470
Calcium-binding domains, 442
Calcium-binding proteins, 441-442; see
also specific proteins
EF hand domains, 442,443-445
EGF-like domains, 471-479
extracellular, 442, 470-485
intracellular, 442, 443-456
ligand preferences, 442
Calcium2+-dependentprotein kinase C,
460,462-463
Calcium-mediated membrane-binding
proteins, 442, 456-457; see also specific proteins
y-carboxyglutamic acid sites, 465-470
C2 domains, 460-465
Calcyclin, 454-456
Calixarenes, 175
Calmodulin, calcium binding, 444,
445-450
Calmodulin-dependent protein kinase I,
448
Calmodulin-dependent protein kinase 11,
447
Carbon disulfide, photodissociation,
149-150
Carbon-hydrogen bond activation, 143
Carbonyls, see Metal carbonyls; specific
carbonyl complexes
y-Carboxyglutamic acid sites, calciummediated membrane-binding proteins, 465-470
Cellulose decomposition, 316-3 17
CeriumW), peroxotitanium reaction,
160-161
Cesium niobium iodide complexes, 34
Cesium pseudohalide complexes, 42
Chain structure supramolecular complexes, 251-263
Channels, supramolecular complexes,
222-227,283-289, 291-292
Chemical reaction intermediates, see Reactive intermediates
Chemical reaction mechanisms, 102

Chemical reaction rate, 101-102
Chevral phases, 3, 14
Chloride, oxygen-evolving complex requirement, 328, 329, 338-339, 393
Chlorine oxide, 109-110, 158
Chlorofluorocarbons, 109
Chlorophyll a, 328-329
Chloroplasts, 328
Cholegobin, 529
Chromium(III), oxidation to chromium(V), hydrogen peroxide, 161
Chromium complexes, xenon fluoride reactions, 89
Chromium fluorides, 59
Chromium hexacarbonyl, reactive intermediates, 102-104, 134
Chromium pentacarbonyl, 106, 164-165
flash photolysis, 139-140
matrix isolation, 128-130
Chromium pentacarbonyl dihydrogen
complex, 133-134, 135
Chromium-salen complexes, 396
Chymotrypsin, calcium binding, 479
Clathrates, noble gases in, 51, 54
Coagulation factors, 442, 466-470, 474477, 479-481
Cobalt complexes
intercalation, 155
nitric oxide insertion, 104
xenon fluoride reactions, 89
Cobalt trifluoride, 59
Conantokins, calcium binding, 470-471
Concanavalin A, 308
Conotoxins, 470
CopperU), 174-175
Copper(1) supramolecular complexes,
174- 175
channels, 222-227,283-289,291-292
diamondoid frameworks, 240-251
graphite frameworks, 204-219
helical frameworks, 176-192
hexagonal frameworks, 204-219
hydrogen bond-assembled frameworks,
219-228
infinite-chain structures, 251-271
n-n interaction-assembled frameworks,
228-239
self-assembly pathway, 175
stereochemistry, 174-175
545
INDEX
stoichiometries, 176
sulfur-sulfur contact-assembled frameworks, 192-204
two-dimensional structures, 271-283
Copperlzinc superoxide dismutase, 310
Crown ethers, 175
Cryptands, 175
Cysteamine, leghemoglobin reaction,
523-524
Cysteine, leghemoglobin reaction, 523,
526
Cystolic phospholipase A2, 460
Cytochrome Bss2,four-helix bundle arrangement, 444
Cytochrome c oxidase, dioxygen reaction,
143
Cytochrome P,eo,395
D
Diamondoid framework supramolecular
complexes, 240-251
Dioxygen
cytochrome c oxidase reaction, 143
manganese catalase, 322
manganese dioxygenase, 312
oxygen-evolving complex, 327-328,
330,419
Electron diffraction, reactive intermediates, 110-113

time-resolved, 149
Electronic absorption spectrophotometry,
160
Electrospray ionization mass spectrometry, 162-163
Epidermal growth factor (EGF)-likedomains, calcium-binding proteins,
471-479
Epoxidation, manganese complex applications, 394-402
4,5-Ethylenedithio-l,3-dithiole-2-thionebased supramolecular complexes,
200-204
Extended X-ray absorption fine structure
(EXAF'S), 390-391
Extracellular calcium-binding proteins,
442,470-485
F
Factor VII, calcium binding, 466,467,
479-480,481
Factor IX, calcium binding, 466,467,
469,474-477,479-480
Factor X, calcium binding, 466,467-468,
474-477,479-480
singlet, 122
Dioxygenases, 312;see also Manganese(I1) dioxygenase
Fast-scan electrochemical techniques, 163

Feldspar-like supramolecular complexes,
1,3-Dithiole-2-thione-4,5-dithiolate-based
supramolecular complexes, 193-194
Dixenon cation, 68
DNA
damage, peroxynitrite implicated, 310
synthesis, 319
Femtosecond pulses, reactive intermediate characterization, 146-148

Ferrocyanide, leghemoglobin reaction,
E
Early transition metal halide clusters,
see Group 5 metal halide clusters;
Group 6 metal halide clusters
E cadherin, calcium binding in, 483-484
EF hand domains, calcium-binding proteins, 442,443-445
Electric discharge, reactive intermediates, 118-121
Electrochemical techniques, fast-scan,
163
289-290
524
Fibrillin, calcium binding, 473,474,477
Fibulin-1, calcium binding, 473
Flash photolysis, 106,137,139-140
Flow characterization methods, reactive
intermediates, 156-164
Fluorescence, 156
Fourier-transform microwave spectroscopy, pulsed-nozzle, 114-115
G
Gallium trihydride, 121-122
Gas-phase reactive intermediates
low pressure, 107-113
supersonic jets, 113-121
546
INDEX
Germanium complexes, xenon fluoride reactions, 85

Globin-derived radicals, leghemoglobin
heme-globin crosslinking, 527-530
leghemoglobin reactions, other molecules, 531-533
low-molecular weight species reactions,
530-531
membrane reactions, 533-537
Glutamic acid, posttranslational carboxylation, 466-470
Glutathione, leghemoglobin reaction,
Fe(IV)=O, 521-522
Gold, oxidative fluorination, 59
Graphite framework supramolecular complexes, 204-219
GroEL, 162-163
Group 5 metal halide clusters, 1-3, 24
charge-transfer salt complexes, 40-41
chemically rpodified surfaces, 43-44
electronic structure, 31
extended solids, 42-43, 43
halide bridges, 1-2
intercluster bridging, weak, 3
ligand substitution, 27-31
molecular structure, 31-33
oxidation states, 24
redox chemistry, 25-27
supported cluster materials, 39
synthesis, 25
Group 6 metal halide clusters, 1-4
electronic structure, 19-20
halide bridges, 1-2
halide-chalcogenide clusters, 14-15
higher nuclearity clusters, 35-37
intercluster bridging, weak, 3
ligand chemistry, axial, 5-13
ligand chemistry, inner, 13-15
mixed halo clusters, 10
photophysics, 17-19
supported cluster materials, 37-39
synthesis, 4-5
H
Helical framework supramolecular complexes, 204-219
Hemerythrin, four-helix bundle arrangement, 444

Hemoglobin, see also Leghemoglobin;
Myoglobin
ligand binding, gases, 501-506
plant analogs, 496
Heterometallic polymeric cluster compounds, 258-259
Hexagonal framework supramolecular
complexes, 204-219
HIV-1, RNase H enzyme, 307
Hydrogen atoms, scavenging, 121
Hydrogen bonding
supramolecular copper(I)/silver(I) complexes, 219-228
transition metal halide clusters, 43
Hydrogen peroxide
arginase disproportionation, 326-327
chromium(II1) oxidation, 161
leghemoglobin reaction, Fe(II),
510-511
leghemoglobin reaction, Fe(III), 511517,527
leghemoglobin reaction, Fe(IV)=O, 520
manganese catalase reaction, 322, 414
metLB reaction, 531-533, 534-535
Hydrolases, 306
Hydroxyl radical, 119
I
ICP (Calbindin Dgk),444, 452-453
Infinite-chain supramolecular complexes,
251-271
Infrared diode laser spectroscopy, 119,
148
Infrared spectroscopy, time-resolved, 138,
139
ultrafast, 146
Inorganic polymers, 281
Intracellular calcium-binding proteins,
442,443-456
Iridium complexes, xenon fluoride reactions, 89
Iridium vinyl hydride complex, 131-132
Iron carbonyl complexes
flash photolysis, 140-142
Iron dihydrides, reactive intermediates,
143-146
547
INDEX
Iron dioxygenases, 312-314

Iron protoporphyrin IX group
leghemoglobin, 498, 499
manganese peroxidase, 315-318
Iron ribonucleotide reductases, 319-322,
327; see also Manganese ribonucleotide reductase
Iron superoxide dismutase, 310; see also
Manganese superoxide dismutase
K
Ketone chlorination, manganese complex
applications, 403-405
Kochi-Jacobsen-Katsuki epoxidation, 395
Krypton-carbon bonds, 52
Krypton chemistry, 52, 55-61
Krypton difluoride, 55-61
Krypton-fluorine bonds, 52, 55-61
Krypton-nitrogen bonds, 52, 57-58
Krypton-oxygen bonds, 52, 55, 57
KryptonR5recovery, 61
L
a-Lactalbumin, calcium binding in,
481-483
Lactose synthase complex, 481
Lanthanides, xenon fluoride reactions, 91
Lanthanum tetrafluorides, 59,91
Lapis lazuli, 122, 123
Laser-induced fluorescence, 156
Laser magnetic resonance, 156-157
Lectins, 308
Leghemoglobin, 496
amino acid sequence homology, 498
carbon monoxide binding, 501-503
globin-derived radicals
heme-globin crosslinking, 527-530
leghemoglobin reactions, other molecules, 531-533
low-molecular weight species reactions, 530-531
isomers, 497-498
localization, 500-501
nitric oxide binding, 503-505
reactions, 501-507
structure, 497-499
turnover time, 500-501
Leghemoglobin, Fe(I1)
diatomic gas binding, 501-505
nitrogen complexes, 505-506
oxidation, 507-511
small ligands, 506
Leghemoglobin, Fe(II1)
NADH reaction, 526
nitrogen complexes, 505-506
oxidation, 511-519
reduction, enzymatic, 524-525
reduction, nonenzymatic, 525-526
small ligands, 506-507
Leghemoglobin, Fe(IV)=0
ascorbate reaction, 520-521
glutathione reaction, 521-522
hydrogen peroxide reaction, 520
reduction, 519-524
thiols reaction, 522-524
Leguminous plants, leghemoglobin in,
496; see also Leghemoglobin
Light-induced crystal oscillation,
251-254
Lignin peroxidase, 315-318
Liquid-phase reactive intermediates, see
Solution-phase reactive intermediates
Lupin leghemoglobin, 498, 502, 516
Lysozymes, calcium binding in, 481-482
M
Magnetic susceptibility, 383-384
Main group metals, xenon fluoride reactions, 85-89
Manganese
biological role, 305-306, 424-425
coordination chemistry, 309
electronic properties, 380-382
nonredox roles, 306-308
oxidation states, aqueous solution,
379-380
synthetic applications
chlorination, 403-405
epoxidation, 394-402
nitrido group transfedaziridination,
402-403
Manganese catalase, 306, 309, 322-325
binuclear structure, 325-327
mimics, 410-419
548
INDEX
modeling, 344
model system reactivity, 410-419
uv-vis spectroscopy, 383
x-ray spectroscopy, 391
Manganese(I1) dioxygenase, 309,
312-314
active site, 315
modeling, 315
Manganese-nitrido complexes, 358-359,
402-403
Manganese peroxidase, 309, 315-318
Manganese-porphyrin complexes,
400-402
Manganese redox enzymes, 305-309; see
also Arginase; Oxygen-evolving complex; specific enzymes
binuclear, 309, 325-327
catalase mimics, 414-415
EPR spectroscopy, 386-387
manganese-manganese distances,
360
mixed-valent complexes, 365-367
structural models, 359-372
triple bridged, 368
electronic spectroscopy, 382-383
magnetism, 383-385
metalloenzyme analogs, 424-425
model systems, 405-424
mononuclear, Mn(11)
eight-coordinate, 350-351
five-coordinate, 344-345
low spin complex, 351
seven-coordinate, 347-350
six-coordinate, 345-347
mononuclear, Mn(II1)
five-coordinate, 352-353
Mn-peroxide moiety, 354
six-coordinate, 353-354
mononuclear, Mn(IV), 354-357
mononuclear, MnW), 357-359
physical properties, 379-393
polynuclear, 309, 373-379
electronic spectroscopy, 382
magnetism, 383-385
reactivity, 405-424
structural geometry, 315
tetranuclear, 376-379
mixed-valent complexes, 376-377
trinuclear, 373-375
x-ray absorption spectroscopy, 389-393
Manganese ribonucleotide reductase,
306, 309, 319-322
binuclearity, 325-327
hydroxyurea sensitivity, 322
Manganese-salen complexes, 394-400
Manganese superoxide dismutase, 306,
309, 310-312
active site, 315
modeling, 315, 344
monomolecular structure, 344, 347351, 354
pharmaceutical application, 315
Manganese tetrafluoride, 59
Mass spectrometry, electrospray ionization, 162-163
Matrix isolation methods, 122-131
Membrane-binding proteins, calcium-mediated, see Calcium-mediated membrane-binding proteins
Membranes, leghemoglobin reactions,
533-537
Mercury complexes, xenon fluoride reactions, 90-91
Mercury titanium sulfide clusters, intercalation, 155
Metal carbonyls
fluorides, 89-90
Metal cyanides, 205-206
Metals
noble gas combinations, 51
oxidative fluorination, 59
self-assembly, 174-175
Methyl radical, 119
Molecular recognition, calmodulin, 447
Molybdenum chloride, 4-24, 35-44
Molybdenum halide clusters, 1-4
extended solids, 42-43
halide-chalcogenide clusters, 14-15
higher nuclearity clusters, 35-37
549
INDEX

photophysics, 17-19
supported cluster materials, 37-39
synthesis, 4-5
Molybdenum hexacarbonyl, 4
matrix isolation, 128
Mononuclear manganese redox enzymes,
see Manganese redox enzymes, mononuclear
Myoglobin, see also Hemoglobin; Leghemoglobin
function observation, 156
hydrogen peroxide reaction, 518-519
ligand binding, gases, 501-506
plant analogs, 495-496
MGY, transition metal halide clusters, see
Group 6 metal halide clusters
M6Y12transition metal halide clusters,
see Group 5 metal halide clusters
N
Network structure supramolecular complexes, 273-280
Neuromodulin, 449
Neuron-specific calcium sensor (NCS) proteins, 457
Nickel carbonyl complexes, reactive intermediates, 134-135
Nickel hexafluoride, 59
Nickel superoxide dismutase enzymes,
310
Niobium chloride clusters, 24-33
Niobium halide clusters, 1-3, 24
chemically modified surfaces, 43-44
electronic structure, 31
extended solids, 43
supported cluster materials, 39
synthesis, 25
Niobium iodide clusters, 2, 33-35
Nitrido group transfer, manganese complex applications, 402-403
Nitrogen, noble gas combinations, 51

Nitrogen dioxide, 138
Nitrogen-fixing root nodules, leghemoglobin in, 495-501, 529-530
Nitrogen oxides, 159
thermal de-NO,, 148-149
NMR spectroscopy, 161-162
Noble-gas chemistry
advances in, 51-54
argon chemistry, 52, 54
krypton chemistry, 52, 55-61
radon chemistry, 52-53, 54, 91-93
review literature, 53-54
xenon chemistry, see Xenon chemistry
Noble gases, 51-54
matrix isolation application, 122
reactive intermediate solution study application, 132
Notch receptor proteins, 473, 475
0
Oligophenanthrolines, 176
Oligopyridines, 175, 176
Organometallic polymers, 254-258
Osmium complexes, xenon fluoride reactions, 90
Osteocakin (SPARC), 484
Oxygen, noble gas combinations, 51
Oxygen-binding proteins, 495-496
Oxygen-evolving complex, 306, 309,
327-343
calcium requirement, 328, 338-339
catalase-like reaction, 411-419
chloride requirement, 328, 329, 338339,393
manganese requirement, strict, 425
modeling, 344
oxidation states, 330-339
photosynthesis, 327-330
spectroscopy, 331-339, 342
structure, 327-330, 337-342
Tyrosine Z residue, 330, 338
water oxidation, 339, 342-343
water oxidation model system,
419-424
x-ray spectroscopy, 391-393
Ozone depletion, 109-110
550
INDEX
P
Parasponia spp., leghemoglobin in, 496
Parkinsons disease, peroxynitrite implicated, 402
Parvalbumin, 443, 445
Peribacteroid membranes, 533-537
Peroxotitanium, cerium(1V) reaction,
160-161
Peroxynitrite
diseases implicated in, 310-311, 402
pulse studies, 138
Perxenate ion, 65
Perxenic acid, 67
Phenazine-bridged supramolecular complexes, 214-219
Phosphohydrolases, 306
Phosphoinositide-specific phospholipase
C, 460, 464-465
Phospholipase A,, 442
Phosphorus complexes, xenon fluoride reactions, 86
Photosynthesis, 327-330, 495; see also
Oxygen-evolving complex
Photosystem I, 329
Photosystem 11, 328
8-8 Interaction framework supramolecular complexes, 228-239
Plant globin gene family, 498-499
Platinum, xenon fluoride reactions, 91
Polycatenane complexes, 271-273
Polynuclear manganese redox enzymes,
see Manganese redox enzymes, polynuclear
Polyrotaxene complexes, 271-273
Porphyrin-based manganese complexes,
400-402
Potassium pseudohalide clusters, 42
Products, of reaction, 101
Protein C, calcium binding, 466, 467,
473, 479-481
Protein-film voltammetry, 163
Protein kinase C, calcium2+-dependent,
460, 462-463
Protein kinase I, calmodulin-dependent,
448
Protein kinase 11, calmodulin-dependent,
447
Proteins
calcium-binding, see Calcium-binding
proteins
electrospray ionization mass spectrometry, 162-163

function observation, 155-156
manganese structural support,
305-306
oxygen-binding, 495-496
residue radicals, enzymes, 340
Protein S, calcium binding, 466, 474475, 479
Prothrombin, calcium binding, 466, 469
Prussian Blue, 41
Pulsed laser photolysis, 141
Pulsed-nozzle Fourier-transform microwave spectroscopy, 114-115
Pulse radiolysis, 137-138
Pump-and-probe techniques, 137
Pyrazine-bridged supramolecular complexes, 206-213, 240-243
Pyridine-bridged supramolecular complexes, 240-243
Pyruvate carboxylase, 306
Q
QR1, 484
R
Radon chemistry, 52-53, 54.91-93
Radon fluorides, 52, 91-93
Radon oxide, 91
Raman spectroscopy, time-resolved resonance, 138, 139, 143
Rare earth halide clusters, 2-3
Reactants, 101
Reaction rate, 101-102
Reactive intermediates, 101-107, 164
flow characterization methods,
156-164
gas-phase studies, 107-121
lifetimes, 106
retardation characterization methods,
107
gas-phase studies, low pressure,
107-113
gas-phase studies, supersonic jets,
113-121
solution studies, low temperature,
131-136
trapping, 121-122
INDEX
solid state studies, 154-156

solution-phase studies, 131-136,
159-160
time-resolved characterization methods, 136-156
Recoverin, calcium binding in, 457
Redox enzymes, see Manganese redox enzymes
Resonance fluorescence, 156
Resonantly enhanced multiphoton ionization (REMPI), 147
Retardation reactive intermediate characterization, see Reactive intermediates, retardation characterization
methods
Retroviral reverse transcriptases,
307-308
Reverse transcriptases, 307-308
Rhenium complexes
reactive intermediates, 136
Rhczobiurn spp., leghemoglobin in, 496,
499-500; see also Leghemoglobin
Rhodium carbonyl complexes, reactive intermediates, 165-166
Ribonuclease enzymes, 307-308
Ribonucleotide reductases, 319; see also
Manganese ribonucleotide reductase
RNase H, HW-1, 307
Root nodules, leghemoglobin in, 495-501,
529-530
Rubidium pseudohalide clusters, 42
Ruthenium complexes
Ruthenium dihydrides, reactive intermediates. 143-146
S
So state, oxygen-evolving complex, 330,
333-334, 336-338
x-ray spectroscopy, 393
S, state, oxygen-evolving complex, 330,
331-334, 336-338
S2state, oxygen-evolving complex, 330,
332-333,336-338
336-338
55 1

336-338
S100A6,454-456
SlOOB, 453-454,455
SlOOD, 444,452-453
sloop, 453-454
S l O O proteins, calcium binding, 451-456
Salen-based manganese complexes,
394-400
SC1, 484
S cycle, oxygen-evolving complex,
330-339
Secreted Protein Acidic and Rich in Cysteine (SPARC), 484-485
Selenium complexes, xenon fluoride reactions, 87
Self-assembling metal complexes, 174175; see also Supramolecular copper( I)/silver(I)complexes
Serine proteases, calcium binding in,
479-481
Serum albumin, 470
Silane discharges, 119-121
Silicon complexes, xenon fluoride reactions, 85
Silicon dichloride, 159
Silicon oxide, reactive intermediates,
108
Silver(I), 174-175
SilverU) supramolecular complexes,
174-175
channels, 222-227, 283-289,291-292
219-228
n-n interaction-assembled frameworks,
228-239
three-dimensional structures, 283-292
Silver titanium sulfide clusters, intercalation, 155
Silver trifluoride, 59-60
552
INDEX
Smooth muscle light-chain kinase,

447-448
Solid-state reactive intermediates,

154-156
Solid-state supramolecular complexes,

see Supramolecular copper(I)/silver(I)
complexes
Solution-phase reactive intermediates
flow systems, 159-160
low temperature, 131-136
Solvent effects, reactive mechanisms
and, 102
Soybeans, leghemoglobin in, 495-501; see
also Leghemoglobin
SPARC (Secreted Protein Acidic and
Rich in Cysteine), 484-485
S l O O proteins, calcium binding, 451-456
Subtilisins, calcium binding, 481
Sulfur
scavenging, 121
xenon fluoride reactions, 86-87
Sulfur-sulfur contact framework supramolecular complexes, 192-204
Supercritical fluids, reactive intermediate characterization, 135, 140
Superoxide, 310
destructive effects, 310-311
leghemoglobin reaction, 508-509
Superoxide dismutase enzymes, 310; see
also Manganese superoxide dismutase
Supramolecular copper(I)/silver(I) complexes, 174-176, 292-293
channels, 222-227,283-289,291-292
219-228

a-8interaction-assembled frameworks,
228-239

three-dimensional structures, 283-292
Synaptotagmin, 460, 464-465
Tantalum bromide clusters, 26

Tantalum chloride clusters, 24-33
Tantalum halide clusters, 1-3, 24
chemically modified surfaces, 44
electronic structure, 3 1
extended solids, 43
synthesis, 25
Tellurium complexes, xenon fluoride reactions, 87-88
Tetrakis(a1kylthio)-tetrathiafulvanebased supramolecular complexes,
196-200
Tetranuclear manganese redox enzymes,

376-379
Tetrathiafulvane derivatives, 175, 192

Thiols, leghemoglobin reaction,
Fe(IV)=O, 522-524
Time-resolved characterization methods,
reactive intermediates, 136-156
Time-resolved infrared spectroscopy, 138,
139, 146
Time-resolved resonance Raman spectroscopy, 138, 139, 143

Titanium, peroxotitanium-cerium(1V) reaction, 160-161
Transition metals
halide clusters, see Group 5 metal halide clusters; Group 6 metal halide
clusters
root nodule damage initiator, 534
xenon fluoride reactions, 89-91
Transition states, 102
Trapping, reactive intermediates,
121-122
Trimethyleamine, thermal decomposition, 110-112

Trinuclear manganese redox enzymes,
373-375
Trolox C, 530
Troponin C, 450-451
Troponin I, 450
Troponin T, 450
Trypsin, calcium binding, 479
Trypsinogen, calcium binding, 479
TTC,-TTF-based supramolecular complexes, 196-200
553
INDEX
Tungsten carbonyl complexes, 4,

152-154
Tungsten chloride, 4
Tungsten complexes
reactive intermediates, 152-154
Tungsten halide clusters, 1-4
halide-chalcogenide clusters, 15
photophysics, 17-19
synthesis, 4-5
Tungsten hexacarbonyl, 4
Tungstenocene, 125, 126
U
Ultrafast time-resolved infrared spectroscopy, 146
Ultraviolet-visible absorption, 138-140
Ultraviolet-visible emission, 138
V
Vanadium carbon hydrogen radical, 118
W
Water oxidation complex, 328, 339,
342-343
model system, 419-424
White rot fungi, 315-316
X
Xenate ion, 65
Xenic acid, 65, 67
Xenon-boron bonds, 54
Xenon-carbon bonds, 54, 81-84
Xenon cations, 68-72
Xenon chemistry, 52
advances in, 61-64
literature review, 53
Xenon difluoride, 58-59, 61-64

ammonolysis, 78
fluoride-ion donor chemistry, 69-70
iodide reactions, 88-89
metathesis reactions, anhydrous acids,
75-76, 77
physical properties, 64
Xenon fluorides, 52, 61-64
anions from, 72-75
cations from, 68-72
fluoride-ion donor/acceptor reactions,
68-72
fluorinating agents, 84-91
hydrolysis reactions, 65-67
metathesis reactions, anhydrous acids,
75-84
oxidizing/fluorinating strengths, 85
Xenon hexafluoride, 61, 62-64
ammonolysis, 78
fluoride-ion acceptor chemistry, 72-74
Xenon-nitrogen bonds, 54, 78-81
Xenon oxide fluorides, 66-67
fluoride-ion acceptor chemistry, 74-75
Xenon oxides, 65-66
Xenon(I1) pentafluoroorthotellurates,
76-78
Xenon-platinum hexafluoride, 5 1-52
Xenon tetrafluoride, 61-62
ammonolysis, 78
fluoride-ion acceptor chemistry, 74
fluoride-ion donor chemistry, 70
Xenon tetraoxide, 66
Xenon trioxide, 65-66
X-ray absorption near edge spectroscopy
(XANES), 390
2
Zeolite-like supramolecular complexes,

274-277
Zeolites
krypton difluoride decomposition, 61
noble gases in, 51
Zirconium halide clusters. 3
CONTENTS OF PREVIOUS VOLUMES
VOLUME 36
Inorganic Chemistry and Drug Design
Peter J. Sadler
Lithium and Medicine: Inorganic
Pharmacology
N . J . Birch and J. D. Phillips
The Mo-, V-, and Fe-Based Nitrogenase
Systems of Azobacter
Robert R. Eady
The Extraction of Metals from Ores
Using Bacteria
D. Keith Ewart and Martin N . Hughes
Solid-state Bioinorganic Chemistry:
Mechanisms and Models of
Biomineralization
Stephen Mann and Carole C. Perry
Magnetic Circular Dichroism of
Hemoproteins
M. R. Cheesman, C. Greenwood, and
A. J. Thomson
The Uptake, Storage, and Mobilization of

Iron and Aluminum in Biology
S. Jemil, A. Fatemi, Fahrni H. A.
Kadir, David J. Williamson, and
Geoffrey R. Moore
Probing Structure-Function Relations in
Ferritin and Bacteriofenitin
P. M. Harrison, S. C. Andres, P. J.
Artymiuk, G. C. Ford, J . R. Guest, J .
Hirzmann, D. M. Lawson, J . C.
Livingstone, J. M. A. Smith, A.
Treffrey, and S. J. Yewdall
INDEX
VOLUME 37
On the Coordination Number of the
Metal Crystalline
HalogenocupratedI) and
Hal ogenoargentates(1)
Susan Jagner and Garan Helgesson
Flavocytochrome b,
Stephen K. Chapman, Scott A. White,
and Graeme A. Reid
Structures of Organonitrogen-Lithium
Compounds: Recent Patterns and
Perspectives in Organolithium
Chemistry
Karina Gregory*Paul von Rague
Schlqyer, and Ronald Snaith
X-Ray Absorption Spectroscopy and the

Structures of Transition Metal
Centers in Proteins
C. David Garner
Cubane and Incomplete Cubane-Type

Molybdenum and Tungsten 0x01
Sulfido Clusters
Takashi Shibahara
Direct Electrochemistry of Proteins and

Enzymes
Liang-Hong Guo and H. Allen 0. Hill
Interactions of Platinum Amine

Compounds with Sulfur-Containing
Biomolecules and DNA Fragments
Edwin L. M. Lempers and J a n Reedijk
Active-Site Properties of the Blue Copper

Proteins
A. G. Sykes
Recent Advances in Osmium Chemistry

Peter A. Lay and W. Dean Harnian
555
556
Oxidation of Coordinated Diimine

Ligands in Basic Solutions of
Tris(diimine)iron(III),ruthenium(III), and -osmium(III)
0. M0nsted and G. Nord
INDEX
VOLUME 38
Trinuclear Cuboidal and Heterometallic
Cubane-Type Iron-Sulfur Clusters:
New Structural and Reacticity
Themes in Chemistry and Biology
R. H. Holm
Replacement of Sulfur by Selenium in
Iron Sulfur Proteins
Jacques Meyer, Jean-Marc Moulis,
Jacques Gaillard, and Marc Lutz
Dynamic Electrochemistry of Iron-Sulfur
Proteins
Fraser A. Armstrong
EPR Spectroscopy of Iron-Sulfur Proteins
Wilfred R. Hagen
VOLUME 39
Synthetic Approach to the Structure and
Function of Copper Proteins
Nobumasa Kitcy'ima
Transition Metal and Organic RedoxActive Macrocycles Designed to
Electrochemically Recognize Charged
and Neutral Guest Species
Paul D. Beer
Structure of Complexes in Solution
Derived from X-Ray Diffraction
Measurements
Georg Johansson
High-Valent Complexes of Ruthenium
and Osmium
Chi-Ming Che and Vivian
Wing-Wah Y a m
Heteronuclear Gold Cluster Compounds
D. Michael P. Mingos and Michael J.
Watson
Structural and Functional Diversity of

Ferredoxins and Related Proteins
Hiroshi Matsubara and Kazuhiko Saeki
Molecular Aspects on the Dissolution and

Nucleation of Ionic Crystals in
Water
Hitoshi Ohtaki
Iron-Sulfur Clusters in Enzymes:

Themes and Variations
Richard Cammack
INDEX
Aconitase: An Iron-Sulfur Enzyme

Mary Claire Kennedy and C. David
Stout
Novel Iron-Sulfur Centers in
Metalloenzymes and Redox Proteins
from Extremely Thermophilic
Bacteria
Michael W. W. Adams
Evolution of Hydrogenase Genes
Gerrit Voordoww
Density-Functional Theory of Spin
Polarization and Spin Coupling in
Iron-Sulfur Clusters
Louis Noodleman and David A. Case
INDEX
VOLUME 40
Bioinorganic Chemistry of PterinContaining Molybdenum and
Tungsten Enzymes
John H. Enemark and Charles G.
Young
Structure and Function of Nitrogenase
Douglas C. Rees, Michael K. Chan, and
Jongsun Kim
Blue Copper Oxidases
A. Messerschmidt
Quadruply Bridged Dinuclear Complexes
of Platinum, Palladium, and Nickel
Keisuke Umakoshi and Yoichi Sasaki
Octacyano and 0x0- and

Nitridotetracyano Complexes of
Second and Third Series Early
Transition Metals
Johann G. Leipoldt, Stephen S. Basson,
and Andreas Roodt
Macrocyclic Complexes as Models for
Nonporphine Metalloproteins
Vickie McKee
Complexes of Sterically Hindered
Thiolate Ligands
J. R. Dilworth and J . H u
INDEX
VOLUME 41
The Coordination Chemistry of
Technetium
John Baldas
Chemistry of Pentafluorosulfanyl
Compounds
R. D. Verma, Robert L. Kirchmeier, and
Jeanne M. Shreeve
The Hunting of the Gallium Hydrides
Anthony J. Downs and Colin R.
Pulham
The Structures of the Group 15
Element(II1) Halides and
Halogenoanions
George A. Fisher and Nicholas C.
Norman
Intervalence Charge Transfer and
Electron Exchange Studies of
Dinuclear Ruthenium Complexes
Robert J. Crutchley
Recent Synthetic, Structral,
Spectroscopic, and Theoretical
Studies on Molecular Phosphorus
Oxides and Oxide Sulfides
J . Clade, F. Frick, and M. Jansen
Structure and Reactivity of Transferrins
E. N . Baker
INDEX
557
VOLUME 42
Substitution Reactions of Solvated Metal
Ions
Stephen F. Lincoln and Andre E.
Merbach
Lewis Acid-Base Behavior in Aqueous
Solution: Some Implications for
Metal Ions in Biology
Robert D. Hancock and Arthur E.
Martell
The Synthesis and Structure of
Organosilanols
Paul D. Lickiss
Studies of the Soluble Methane
Monooxygenase Protein System:
Structure, Component Interactions,
and Hydroxylation Mechanism
Katherine E. Liu and Stephen J.
Lippard
Alkyl, Hydride, and Hydroxide
Derivatives in the s- and p-Block
Elements Supported by
Poly(pyrazoly1)borato Ligation:
Models for Carbonic Anhydrase,
Receptors for Anions, and the Study
of Controlled Crystallographic
Disorder
Gerard Parkin
INDEX
VOLUME 43
Advances in Thallium Aqueous Solution
Chemistry
Julius Glaser
Catalytic Structure-Function
Relationships in Heme Peroxidases
A n n M. English and George Tsaprailis
Electron-, Energy-, and Atom-Transfer
Reactions between Metal Complexes
and DNA
H. Holden Thorp
Magnetism of Heterobimetallics: Toward
Molecular-Based Magnets
Olivier Kahn
558
The Magnetochemistry of Homo- and

Hetero-Tetranuclear First-Row
d-Block Complexes
Keith S. Murray
Diiron-Oxygen Proteins
K. Kristoffer Andersson
and Astrid Graslund
Carbon Dioxide Fixation Catalyzed by
Metal Complexes
Koji Tanaka
INDEX
VOLUME 44
Recent Developments in Chromium

Chemistry
Donald A. House
INDEX
VOLUME 45
Syntheses, Structures, and Reactions of
Binary and Tertiary Thiomolybdate
Complexes Containing the (O)Mo(S,)
and (S)Mo(S,) Functional Groups
(x = 1, 2, 4)
Dimitri Coucouvanis
Organometallic Complexes of Fullerenes

Adam H . H. Stephens
and Malcolm L. H. Green
The Transition Metal Ion Chemistry of

Linked Macrocyclic Ligands
Leonard F. Lindoy
Group 6 Metal Chalcogenide Cluster

Complexes and Their Relationships
to Solid-state Cluster Compounds
Taro Saito
Structure and Properties of Copper-Zinc

Superoxide Dismutases
Ivano Bertini, Stefan0 Mangani, and
Maria Silvia Viezzoli
Macrocyclic Chemistry of Nickel

Myunghyun Paik Suh
Arsenic and Marine Organisms
Kevin A. Francesconi
and John S. Edmonds
DNA and RNA Cleavage by Metal

Complexes
Geneufve Pratviel, Jean Bernadou,
and Bernard Meunier
The Biochemical Action of Arsonic Acids

Especially as Phosphate Analogues
Henry B. F. Dixon
Structure-Function Correlations in High

Potential Iron Problems
J . A. Cowan and Siu Man Lui
Intrinsic Properties of Zinc(I1) Ion

Pertinent to Zinc Enzymes
Eiichi Kimura and Tohru Koike
The Methylamine Dehydrogenase

Electron Transfer Chain
C. Dennison, G. W. Canters,
S. de Vries, E. Vijgenboom,
and R. J. van Spanning
Activation of Dioxygen by Cobalt Group

Metal Complexes
Claudio Bianchini
and Robert W. Zoellner
INDEX
I S B N O-L2-023646-X

Advances in Inorganic Chemistry Volume46

Caricato da

Informazioni sul documento

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Advances in Inorganic Chemistry Volume46

Caricato da

Copyright:

Formati disponibili

Advances in

Universita degli Stud; di Firenze

lmperial College of Science,

San Diego London Boston

This book is printed on acid-free paper.

Copyright 0 1999 by ACADEMIC PRESS

525 B Street, Suite 1900, San Diego, California 92101-4495, USA

Recent Advances in Noble-Gas Chemistry

Recent Review Literature

Coming to Grips with Reactive Intermediates

Toward the Construction of Functional Solid-State Supramolecular

Manganese Redox Enzymes and Model Systems: Properties,

Leghemoglobin: Properties and Reactions

This Page Intention ally Left Blank

ADVANCES IN INORGANIC CHEMISTRY, VO1.

THE OCTAHEDRAL M6Y8 AND M6Y12 CLUSTERS OF

The early transition metal halide clusters (MfiY8}"+(M = Mo, W;

PROKOPUK AND SHRIVER

of the cluster. In addition, the discrete clusters provide building

FIG. 1. Structures of [MQY&l' A and [M6Y1&10B.

OCTAHEDRAL CLUSTERS OF TRANSITION METALS

Recent interest in halide clusters containing the {M,Y8}"+ unit

PROKOPUK AND SHRIVER

chemistry. Molybdenum(I1) chloride was discovered over 100 years

Near quantitative yields are achieved for the molybdenum halides,

OCTAHEDRAL. CLUSTERS OF TRANSITION METALS

following reaction sequence was proposed by Franolic, Long, and

The lower nuclearity clusters [ M O ~ C I ~were

FIG.3. Structure of Mo,C1,C12Cl,, . Reprinted with permission from Ref. (130).

PROKOPUK AND SHRIVER

1. Mechanism of Ligand Exchange

were performed by Sheldon with 36C1-labeled[Mo6C18C1612in aqueous

OCTAHEDRAL CLUSTERS OF TRANSITION METALS

distribution of all possible compounds, inclusive of the ratios for the

Various types of neutral ligands, including nitrogen donors (12,23,

For instance, addition of PPh3 to MosCllz yields M o , C ~ , C ~ , ( P Pat

PROKOPUK AND SHRIVER

FIG.4. Structure of tralzs-MofiClsC1,tP(n-C4Hs)r12.

to the cis isomer was obtained (36).Alkylation of the trans phosphine

FIG.5. Structure of cis-MofiC1,C1,[P(C,H,),l,.

OCTAHEDRAL CLUSTERS OF TRANSITION METALS

FIG.6. Structure of all-trans-Mo~Cl,CI,(C2H,),[P(C,H,),I,.

FIG.7. Structure of trans-[M0,CI,C1,(0CH,,),]~

Mixed halo clusters (H30)2M6C&Y6

OCTAHEDRAL CLUSTERS OF TRANSITION METALS

PROKOPUK AND SHRrVER

with x < 6 persist in the presence of excess of the phosphines

OCTAHEDRAL CLUSTERS OF TRANSITION METALS

Substitution of the inner ligands of {MO&!~~}~+

2. Inner Ligand Exchange

As implied in the previous discussion, the inner ligands have lower

PROKOPUK AND SHRTVER

OCTAHEDRAL CLUSTERS OF TRANSITION METALS

Lower temperature methods for introducing chalcogenides into the

D. Redox Chemistry and Photophysics of the Group 6 Metal

PROKOPUK AND SHRIVER

OCTAHEDRAL CLUSTERS OF TRANSITION METALS

group 6 metal halide clusters adopting the {MHY1l}lt

FIG. 11. Absorption (full) and emission (dashed) spectra of IBu,NIL[Mo,C1&l,,I in

PROKOPUK AND SHRIVER

Values are reported with acetonitrile solutions of cluster.

OCTAHEDRAL CLUSTERS OF TRANSITION METALS

quenched clusters should be useful for converting solar energy to

PROKOPUK AND SHRIWR