Progress in inorganic chemistry volume 57

Progress in
Inorganic Chemistry
Volume 57


Advisory Board
JACQUELINE K. BARTON
CALIFORNIA INSTITUTE OF TECHNOLOGY, PASADENA, CALIFORNIA
JAMES P. COLLMAN
STANFORD UNIVERSITY, STANFORD, CALIFORNIA
ALAN H. COWLEY
UNIVERSITY OF TEXAS, AUSTIN, TEXAS
RICHARD H. HOLM
HARVARD UNIVERSITY, CAMBRIDGE, MASSACHUSETTS
EIICHI KIMURA
SHIZUOKA UNIVERSITY, SHIZUOKA, JAPAN
NATHAN S. LEWIS
CALIFORNIA INSTITUTE OF TECHNOLOGY, PASADENA, CALIFORNIA
STEPHEN J. LIPPARD
MASSACHUSETTS INSTITUTE OF TECHNOLOGY, CAMBRIDGE,
MASSACHUSETTS
TOBIN J. MARKS
NORTHWESTERN UNIVERSITY, EVANSTON, ILLINOIS
KARL WIEGHARDT
¨ LHEIM, GERMANY
MAX-PLANCK-INSTITUT, MU


PROGRESS IN
INORGANIC CHEMISTRY
Edited by

KENNETH D. KARLIN
DEPARTMENT OF CHEMISTRY
JOHNS HOPKINS UNIVERSITY
BALTIMORE, MARYLAND

VOLUME 57


Copyright Ó 2012 by John Wiley & Sons, Inc. All rights reserved
Published by John Wiley & Sons, Inc., Hoboken, New Jersey
Published simultaneously in Canada
No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any
form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise,
except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without
either the prior written permission of the Publisher, or authorization through payment of the
appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers,
MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to
the Publisher for permission should be addressed to the Permissions Department, John Wiley &
Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at
/>Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best
efforts in preparing this book, they make no representations or warranties with respect to the
accuracy or completeness of the contents of this book and specifically disclaim any implied
warranties of merchantability or fitness for a particular purpose. No warranty may be created or
extended by sales representatives or written sales materials. The advice and strategies contained
herein may not be suitable for your situation. You should consult with a professional where
appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other
commercial damages, including but not limited to special, incidental, consequential, or other
damages.
For general information on our other products and services or for technical support, please

contact our Customer Care Department within the United States at (800) 762-2974, outside the
United States at (317) 572-3993 or fax (317) 572-4002.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print
may not be available in electronic formats. For more information about Wiley products, visit
our web site at www.wiley.com.
Library of Congress Catalog Card Number: 59-13035
ISBN 978-1-118-01063-1
Printed in the United States of America
oBook ISBN: 978-1-118-14823-5
ePDF ISBN: 978-1-118-14868-6
ePub ISBN: 978-1-118-14820-4
eMobi ISBN: 978-1-118-14867-9
10 9

8 7 6 5

4 3 2 1


Contents
Chapter 1

Chapter 2

Chapter 3

Chapter 4

Chapter 5

Mechanisms of Water Oxidation Catalyzed
by Ruthenium Coordination Complexes . . . . . . . . . . . . . . . . . . . .
AURORA E. CLARK and JAMES K. HURST
Biomimetic and Nonbiological Dinuclear Mxþ
Complex-Catalyzed Alcoholysis Reactions of
Phosphoryl Transfer Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . .
R. STAN BROWN

1

55

Photoactivated DNA Cleavage and Anticancer
Activity of 3d Metal Complexes . . . . . . . . . . . . . . . . . . . . . . . . .
AKHIL R. CHAKRAVARTY and MITHUN ROY

119

Design and Evolution of Artificial Metalloenzymes:
Biomimetic Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
MARC CREUS and THOMAS R. WARD

203

Functionalization of Fluorinated Aromatics by
Nickel-Mediated C – H and C – F Bond Oxidative
Addition: Prospects for the Synthesis of
Fluorine-Containing Pharmaceuticals . . . . . . . . . . . . . . . . . . . . .
SAMUEL A. JOHNSON, JILLIAN A. HATNEAN, and
MEGHAN E. DOSTER

255

Chapter 6

DNA Based Metal Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
JENS OELERICH and GERARD ROELFES

Chapter 7

Metallo-b-lactamases and Their Synthetic Mimics:
Structure, Function, and Catalytic Mechanism . . . . . . . . . . . . .
MUTHAIAH UMAYAL, A. TAMILSELVI, and
GOVINDASAMY MUGESH

v

353

395


vi

Chapter 8

CONTENTS

A New Class of Nanostructured Inorganic–Organic
Hybrid Semiconductors Based on II–VI

Binary Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
JING LI and RUIBO ZHANG

445

Oxygen Evolution Reaction Chemistry
of Oxide-Based Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
YOGESH SURENDRANATH and DANIEL G. NOCERA

505

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Cumulative Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

561
593

Chapter 9


Chapter 1, Figure 3. The DFT predicted mechanism for water oxidation catalyzed by [Ru2(OH)2(3,6Bu2Q)(btpyan)]2þ ion [3,6-Bu2Q ¼ 3,6-di-tert-butyl-1,2-benzoquinone and btpyan ¼ 1,8-bis
(2,20 :60 ,200 -terpyrid-40 -yl)anthracene]. (See text for full caption.)

Absorbance (740 nm)

Dark

(b)
(a)

0.6

0.10
0.09
0.08
0.07
0.06

0.4

300 400

Time (s)

(c)
{4,5}

0.2

NIR component
RuIII
RuII

{3,4}

0.0
300

450

600

750

Wavelength (nm)

900

EPR signal intensity

Absorbance

Light
100 200

330 340 350 360

Magnetic field (mT)

Chapter 1, Figure 9. Photocatalyzed water oxidation by the
system. (See text for full caption.)

0

20 40 60 80

Time (s)

S2O82À/[Ru(bpy)3]2þ/“blue

dimer”


Chapter 2, Figure 4. Molecular structure of 20:CuII2 :(HOÀ)(PhCH2O)2PO2À)(CF3SO3À)2 shown as
an ORTEP drawing at the 50% probability level. (Hydrogens and counterions are omitted for clarity.)

Chapter 4, Figure 12. Enantioselectivity of artificial-transfer hydrogenases for acetophenone reduction. In the achiral (planar trigonal) intermediate during catalytic turnover, incorporation of a hydride
from one of the two possible prochiral faces will lead to enantiomers of the three-legged d6 piano stool
complex. (See text for full caption.)


T6

N
N

Cu2+

HN
X

O

ON7

N
N

HN
X

N

ON8

Cu2+

O

O
O

1a
2

N

3a

Chapter 6, Scheme 17. Covalent approach to asymmetric DNA based catalysis.

Chapter 7, Figure 4. Mono- and binuclear structures of mbls from Bacillus cereus (BcII) of subclass B1.
(a) Panel a represents the overall protein structures of mbls, A, B and C from BcII with 2.5-, 1.85-, and 1.9˚ resolution, respectively. (b) panel b (D–F) represents the active sites of corresponding protein structures.
A
Water molecule and hydroxide ions are shown as red spheres, whereas Zn(II) ions are shown as gray
spheres. [PDB codes for structures A, B and C are 1BMC, 1BVT, and 1BC2, respectively (60–62).


Chapter 7, Figure 5. Structures of VIM-2 mbls of subclass B1 in both reduced G and I and oxidized
H and J forms. Panel a represents the overall protein structures, whereas panel b represents the active

sites of these proteins. [PDB codes for structures G and H are 1KO3 and 1KO2, respectively. Here
Ocs221 represents the cysteine sulfonic acid (66).]

Chapter 8, Figure 32. (a) A reference UV LED (360 nm) illuminating blue light (commercially
available). (b) Image of the same LED coated with a thin layer of 2D-[Cd2S2(ba)] before illumination.
(c) The illuminating image of the coated LED. (d) The illuminating image of the coated LED after Mn2þ
doping (0.1 mol%).


Mechanisms of Water Oxidation Catalyzed
by Ruthenium Coordination Complexes
AURORA E. CLARK AND JAMES K. HURST
Department of Chemistry, Washington State University. Pullman, WA
CONTENTS
I. INTRODUCTION
II. OXYGEN–OXYGEN COUPLING OF COORDINATED WATER
A. The [RuII(tpy)(H2O)]2(m-bpp)3þ Ion
B. The “Tanaka Catalyst”
III. HOMOLYTIC CLEAVAGE OF O–H BONDS: THE “BLUE DIMER”
A. Structure
B. Redox States
C. Isotopically Defined Reaction Pathways
D. Theoretical Analyses
E. “Noninnocent” Involvement of Bipyridine Ligands
IV. NUCLEOPHILIC ADDITION OF WATER TO ELECTROPHILIC RUTHENYL OXO
LIGANDS
A. General Reaction Characteristics
B. [Ru(bpm)(tpy)(H2O)]2þ and Related Ions
1. Reaction Pathways
2. Alternative Theoretical Analyses

V. EXPANSION OF THE COORDINATION SPHERE
VI. MEDIUM EFFECTS
A. Ion Pairing
B. Anation

Progress in Inorganic Chemistry Volume 57, First Edition. Edited by Kenneth D. Karlin.
Ó 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.
1


2

AURORA E. CLARK AND JAMES K. HURST
C. Influence on Catalytic Rates

VII. FUTURE DIRECTIONS
A. Tuning Reactivities through Modification of Organic Ligands
B. Electrocatalysis
C. “All-Inorganic” Molecular Catalysts
1. Reaction Characteristics
2. Theoretical Studies
VIII. CONSPECTUS
ACKNOWLEDGMENTS
ABBREVIATIONS
REFERENCES

I. INTRODUCTION
Interest in water oxidation catalyzed by transition metal ions can be traced to
studies in the early 1950s when it was suggested by Dwyer and Gyafaras (1) that tris2,20 -bipyridine (bpy) and 1, 10-phenanthroline (phen) complexes of trivalent group
8 (VIIIB) ions formed ozone (O3) and hydrogen peroxide (H2O2) during their

alkaline decomposition to the corresponding M(II) ions and the subsequent
recognition by Creutz and Sutin (2) that this instability could form the basis for
water photolysis by visible light using [Ru(bpy)3]2þ as photosensitizer. Since direct
one-electron (1eÀ) reduction of H2O to HO. is thermodynamically disallowed,
considerable attention was given to characterizing the reaction dynamics with the
intention of identifying reactive intermediates. A brief review of this early literature
can be found in (3). Speculations concerning the nature of these intermediates
ranged from species with chemically altered bpy ligands to ion aggregates containing stabilized HO. radical [e.g., HO. (HOÀ)n], and even m-oxo dinuclear bridged
ions generated in a complex sequence of reactions initiated by HO. substitution on
the metal to form seven-coordinate intermediates. This last suggestion was apparently inspired by contemporaneous research from Meyer and co-workers (4, 5)
demonstrating that [Ru(bpy)2(H2O)]2O4þ was an effective catalyst for water
oxidation in acidic solutions containing strong oxidants. Careful research on the
[M(bpy)3]3þ alkaline decomposition reactions ultimately led to the realization that
the major, if not sole, pathways for metal ion reduction involved irreversible ligand
oxidation accompanied by negligible formation of O2 (6, 7), and interest in these
ions as potential water oxidation catalysts waned. A decade later, however, in a
publication that did not receive much attention, Ledney and Dutta (8) reported that


MECHANISMS OF WATER OXIDATION CATALYZED

3

[Ru(bpy)3]3þ encapsulated within Y-zeolite supercages decomposed in alkaline
solution with near-stoichiometric formation of O2. Transient species suggestive
of bpy ligand modification were detected by resonance Raman (RR), cryogenic
electron paramagnetic resonance (EPR), and diffuse reflectance spectroscopy,
prompting the researchers to propose a mechanism based upon HO. addition to
the ligand. This general type of mechanism involving “noninnocent” participation of coordinated nitrogen heterocyclic ligands had been previously explored
within a wider context of metal ion reactivity without any definitive supporting

evidence having been found (9–11), and had also been considered by the
Brookhaven group (2) as a potential mechanism for [Ru(bpy)3]3þ catalyzed
water oxidation. The dramatic change of reaction course attending zeolite
encapsulation was attributed to elimination of bimolecular reactions, among
which were presumably the ligand degradation pathways observed in homogeneous solution. Indeed, other research indicated that when [Ru(bpy)3]3þ was
reacted with HO. at high cage occupancies, dioxygen (O2) was not formed.
Rather, carbon dioxide (CO2) evolved in a manner that evoked the solution
reactions, indicating that extensive ligand degradation had occurred (12). Nonetheless, the study made on [Ru(bpy)3]3þ at low zeolite loadings provided the first
indication that, under suitably restrictive conditions, a coordinately saturated
single ruthenium center is capable of catalyzing water oxidation.
A second instructive point arising from the early studies was that in the presence
of certain redox metal ions [e.g., Co(II)] (6, 13, 14) and metal oxides (15–18), which
functioned as cocatalysts, O2 formation by [M(bpy)3]3þ reduction could become
nearly quantitative. Indeed, these observations formed the basis for several fairly
efficient photocycles for water oxidation by electron donors using [Ru(bpy)3]2þ as a
photosensitizer (Fig. 1). In these cases, in addition to functioning as the true catalyst,
the second metal ion most likely protected the [M(bpy)3]3þ by introducing a
competitive reduction pathway that did not involve ligand degradation.
During the 1980–1990s, the perception developed in the field that efficient
homogeneous catalysis of water oxidation required the presence of at least two
metal centers within the complex. Factors contributing to this viewpoint included the
intense focus on understanding biological water oxidation (24–26), then already
known to involve a tetranuclear Mn cluster (27–30), and the repeated demonstrations
that the ruthenium “blue dimer” (cis,cis-[Ru(bpy)2(H2O)]2O4þ) and analogous m-oxo
bridged diruthenium ions were efficient catalysts (31–35) but, in addition to
[Ru(bpy)3]3þ, monomeric complexes containing water ligands, including species
that might be considered dimer fragments (e.g., cis-[Ru(bpy)2(H2O)2]3þ) were
apparently devoid of activity ((4, 31, 36); see, however, 37). Indeed, the discovery
that only two of the four Mn centers in the oxygen-evolving complex undergo redox
cycling further heightened suspicions that dinuclear centers were somehow uniquely

associated with catalytic activity (38, 39). However, very recent discoveries have now


4

AURORA E. CLARK AND JAMES K. HURST

hν

S2O82– (or Co(NH3) 5Cl2+)

*RuL32+

RuL32+

2– 4 cycles

WOCn+4
RuL 33+

2SO 42– (or Co2++5NH4++Cl-)
Net: 2S 2O82– + 2H2O

2hν

WOCn

2H2O
O2 + 4H+

4SO42– + 4H+ + O2

4Co(NH3)5Cl2+ + 16H+ + 2H2O

4hν

4Co2+ + 20NH4+ + 4Cl– + O2

Figure 1. Generic scheme for [RuL3]2þ photocatalyzed water oxidation. Although the reactions
proceed by repetitive cycling of the photocatalyst, only the initial and final (i.e., water oxidizing) states
of the water oxidation catalyst (WOC) are shown. Recent studies have utilized monomeric (19),
dimeric (20, 21) and tetrameric (22) Ru containing WOCs, as described in the text. Photocatalysts have
included [Ru(bpy)3]3þ and analogues containing derivatized bpy ligands, specifically [Ru(dmb)3]2þ,
where dmb ¼ 4,40 -dimethyl-2,20 -bipyridine (19), and [Ru(dcb)2(bpy)]2þ, where dcb ¼ 4,40 -dicarbethoxy-2,2-bipyridine (20). The strongly oxidizing sulfate radical anion [Eo(SO4. À/2À) ¼ 2.4 V]
formed upon 1eÀ reduction of S2O82À reacts with ruthenium bipyridine complexes at near-diffusion
controlled rates (23) and participates in water oxidation by oxidizing both [RuL3]2þ and intermediary
oxidation states of the WOC.

made it abundantly clear that this general assumption is invalid. Examples of efficient
catalysis by mononuclear, dinuclear, and tetranuclear Ru complexes, as well as
similar complexes containing other metal centers, have now surfaced; moreover, this
body of emerging work is transformative in that one no longer seeks to unlock the
mystery of how the O–O bond could possibly form, but rather how to distinguish
among the many demonstrated and proposed pathways that are revealed in these
reactions and to understand how structural factors dictate the expression of one
pathway over another.
This chapter reviews the current state of knowledge concerning water oxidation
as revealed by reactions involving heterocyclic Ru coordination complexes. These
ions possess spectroscopic signatures that make them particularly suited to
mechanistic studies and often accumulate intermediary species during turnover

that can provide important clues to reaction mechanisms. Moreover, advanced
computational analyses based upon density functional theory (DFT), as well as
multiconfigurational self-consistent field (MCSCF) and perturbation theories have
been utilized, which are extremely helpful in evaluating the plausibility of
proposed mechanisms. Although application of DFT and wave function based
methods is now widespread within this field (40–52), it is perhaps worthwhile to
emphasize that, although important as validatory tools, their full predictive power
has not yet been realized. As a recent report suggests (40), difficulties in reliably


MECHANISMS OF WATER OXIDATION CATALYZED

5

predicting mechanisms may be due to limitations in the chemical model that is
studied rather than the computational method that is employed. Indeed, most
theoretical studies do not consider the role of extended explicit solvation of the
complex during the myriad transformations that occur along the reactive
potential energy surface, thus ignoring a key facet of the experimental reaction
conditions. A pertinent case in point is the study of water oxidation catalyzed
by [RuII(tpy)(H2O)]2(m-bpp)3þ (bpp ¼ 2,6-bis(pyridyl)pyrazolate anion and
tpy ¼ 2,20 :60 .200 terpyridine) ion, which is discussed in detail in Section II.
This complex contains two hexacoordinate Ru ions templated within a heterocyclic bridging bpp (Fig. 2). The coordination environment enforces a geometry
in which the water ligands are facially oriented with an OÁ Á ÁO separation distance
˚ . Four-electron (4eÀ) oxidation to the corresponding [RuIV(tpy)
of only $2.09 A
3þ
(O)]2(m-bpp) ion leads to O2 evolution by a unimolecular pathway (54);
18
O-isotopic labeling studies indicate that both O atoms are obtained from the

coordination sphere of the complex ion (53). These data strongly implicate a
mechanism involving coupling between two adjacent RuIV¼O atoms, followed
by reductive elimination of O2 and regeneration of [RuII(tpy)(H2O)]2(m-bpp)3þ,
as illustrated in Fig. 2. However, a DFT computational analysis made prior to the
definitive isotope-labeling study predicted the existence of a prohibitively high
activation energy barrier for this reaction pathway (44). In this study, it was found
that a 1,2-peroxo-bridged intermediate readily formed from [RuIV(tpy)(O)]2(mbpp)3þ, but that decomposition of this intermediate was energetically very
demanding. Thus, by this analysis, the peroxo-bridged complex was identified
as a dead-end species. An alternative low-energy pathway was found that
involved protonation of one of the ruthenyl oxo atoms, causing electron density
to be withdrawn from the adjacent ruthenyl group. This electronic polarization
rendered the ruthenyl oxygen atom sufficiently electrophilic to undergo nucleophilic attack by a solvent molecule with formation of a hydroperoxo–hydroxo
intermediate. Internal electronic rearrangement then led to release of O2 with
regeneration of the catalyst in its original form (Fig. 2). However plausible this
mechanism may be, the subsequently published 18 O labeling studies clearly
show it is not operative under the reaction conditions investigated. Specifically,
this mechanism requires that one O atom be obtained from solvent and the other
from the coordination sphere of the catalyst, which is clearly not the case (53).
This set of studies constitutes an example of the subtlety of forces at play that can
determine which of several potential pathways for water oxidation are expressed,
as well as the extreme challenge this presents to theorists in accurately predicting
activation barriers. Correspondingly, this chapter first focuses attention upon
catalysts for which experimental evidence has given some indication of the actual
reaction pathways and then enumerates other catalytic systems where experimental evidence on proposed reaction pathways is less definitive.


6

AURORA E. CLARK AND JAMES K. HURST

O3
1.854A

N

O4

N
N

N

Ru2

Ru1
N6

N5

N

N

N

{2,2}
pathway a
3+

H 2O

(tpy)RuIII-L-Ru III(tpy)

(tpy)Ru III-L-Ru III(tpy)
H 2O

OO

OH2

O O
3+

O=O

II

4Ce 4+

OH2

H 2O

O

{2,2}

H 2O
3+

(tpy)RuIII-L-RuIII(tpy)

(tpy)Ru III-L-Ru III(tpy)
OO

O
{4,4}

4Ce3+ + 4H+

3+
H 2O

3+

(tpy)Ru IV-L-RuIV(tpy)

II

(tpy)Ru -L-Ru (tpy)

O=O

3+

OH2

HOO

OH

pathway b

Figure 2. Optimized calculated structure of [RuII(tpy)2(H2O)]2(m-bpp)3þ and alternative proposed
pathways for catalyzed water oxidation. For pathway a, both O atoms are derived from the coordination
sphere, whereas for pathway b, one atom is from the coordination sphere and the other is from the
solvent (as identified by the solid circle). [Adapted from (53).]

II. OXYGEN–OXYGEN COUPLING OF COORDINATED WATER
A. The [RuII(tpy)(H2O)]2(m-bpp)3þ Ion
This bis-(pyridyl)pyrazolate-bridged dimer is particularly amenable to analysis
of water oxidation because each of the oxidation steps is thermodynamically
and kinetically resolved and each of the oxidation states has a distinct optical
spectroscopic signature (45, 53). Moreover, following oxidation to the highest
accessible state ([RuIV(tpy)(O)]2(m-bpp)3þ), a transient species accumulates
whose first-order decay parallels O2 release. Consequently, this species could be
a bona fide reaction intermediate in the O2 forming cycle; its accumulation presents


MECHANISMS OF WATER OXIDATION CATALYZED

7

a unique opportunity for structural characterization that is lacking in other catalytic
systems. The cyclic voltammogram (CV) of [RuII(tpy)(H2O)]2(m-bpp)3þ displays
three quasireversible (1eÀ) waves in acidic aqueous solutions; a fourth irreversible
oxidation can be detected at potentials approaching catalytic water oxidation.
These data indicate a regular progression in thermodynamic stabilities that follow
the order: {2,2} ! {2,3} ! {3,3} ! {3,4} ! {4,4} (where the notation given is
meant to indicate only the overall oxidation state of the complex based upon

assignment of formal charges, i.e., not the actual electronic distribution). Oxidation is accompanied by release of protons, as dictated by the increasing acidities of
the higher oxidation states so that, upon complete oxidation to {4,4}, the coordinated aquo ligands are completely deprotonated to give ruthenyl oxo atoms. Rate
constants for stepwise oxidation by Ce4þ progressively decrease with increasing
oxidation state, so that each of the intermediary oxidation states can be isolated and
physically characterized. Upon oxidation to {4,4}, however, spontaneous O2
evolution occurs in a reaction that is associated with first-order formation
and decay of a spectroscopically distinct reaction transient. The visible spectra
of both {4,4} and the transient species (I) have been obtained by global kinetic
analysis.
Species I is suggested to be a 1,2-m-peroxo-bridging intermediate formed by
coupling of the two juxtaposed oxo radicaloid atoms on the adjacent Ru atoms of
{4,4}. Due to the close energetic spacing of the various electronic states of I, the
theoretically predicted ground state is dependent on the exact density functional
used within DFT (43, 44). However, complete active space self-consistent field
calculations with second- order M€
oller–Plesset perturbation theory (CASPT2)
generally agrees quite well with the M06-L DFT implementation, predicting that
each low-spin Ru(III) couples as a triplet with its respective O. À, with the two
triplet RuIII–O. À units coupling as a net S ¼ 2 configuration; these calculations also
indicate that the low-lying S ¼ 0 state lies within 4 kJ molÀ1. From a computational
perspective, the reaction energetics of I are somewhat sensitive to the specific
density functional used. Yet the chemical model employed to mimic both I and its
solvation environment is significant and may be more important. The direct O–O
coupling pathway (Fig. 2) is predicted by both B3LYP and M06-L functionals to
have a reasonable activation barrier for formation of the first intermediate, a cyclic
1,2-peroxo bridged Ru–O–O–Ru3þ{3,3} ion. However, discrepancies exist over
the appropriate treatment of the second transition state to form the {2,2}3þ
protocatalyst. Irrespective of whether the calculation is performed in the gas
phase or utilizing a solvent continuum model to mimic the effects of the bulk
dielectric, it is apparent that the activation barrier is much too high unless the

chemical model is expanded to include more of the explicit solvation environment
surrounding the Ru–O–O–Ru3þ{3,3} intermediate. The approach of Yang and
Baik (44) was to take into account the effects of acidity present in the experimental
solution by examining formation of {2,2}3þ from protonated Ru–O–O–Ru3þ{3,3}.


8

AURORA E. CLARK AND JAMES K. HURST

This approach did not yield significantly improved energetics and, as such, this
reaction pathway was dismissed as a viable mechanism for I. Instead, Yang and
Baik (44) proposed that an alternate pathway consisting of coupling of the
terminal oxo and water oxygen atoms (Fig. 2) would be energetically more
favorable. However, improvement in the microsolvation environment around
Ru–O–O–Ru3þ{3,3} through addition of two waters of hydration yielded a
calculated activation barrier for formation of {2,2}3þ (45) that agreed within
9 kJ molÀ1 with the experimental value.
Although the experimental and theoretical results present a self-consistent
and intuitively reasonable model for catalyzed water oxidation, the reaction
itself presents some unexplained anomalies. The rate laws for oxidation of {2,2}
to {3,3} are first order in both Ce4þ and the dimer. However, the rate law
for oxidation of {3,4} shows apparent saturation of the dependence upon Ce4þ
concentration. Potential causes are discussed below in Section VI on medium
effects. More strikingly, the global kinetic analyses for reactions made at ambient
temperatures indicate that, following a single turnover, the {2,2} product undergoes apparent sequential conversion to two new species that have markedly
altered optical absorption spectra (45). These are suggested to be anated species
that may be similar to Ru2–bpp complexes that have been isolated containing
bridging ClÀ, MeCOOÀ, and CF3SO3À anions in place of the coordinated water
molecules (53, 54). However, the optical changes are considerably greater than

have been reported for m-oxo bridged Ru dimers, where SO42À substitution
occurs (32, 33) and where ClO4À and CF3SO3À anation has been proposed based
upon kinetic effects (55) (Section VI.A). In those cases, the modified catalysts
exhibit optical spectra that are almost indistinguishable from the corresponding
catalytically active diaquo forms. Under conditions where Ce4þ is in large excess,
[RuII(tpy)(H2O)]2(m-bpp)3þ is reported to catalyze water oxidation through as
many as $500 cycles prior to deactivation, so it appears that either the structural
changes implied by the optical spectra occurring after a single cycle are reversible or
the chemically modified complexes are also capable of catalyzing water oxidation.
It was also reported that “exhaustive” electrochemical oxidation led to formation of
a small amount of dinuclear complex containing an oxidized bpp ligand.
B. The “Tanaka Catalyst”
A long-lived diruthenium catalyst for water oxidation containing a binucleating
anthracene-linked pair of terpyridyl groups with redox-active benzoquinone and
hydroxide ions as additional ligands (Fig. 3) was first reported in 2000 (57).
Athough this complex, isolated as [Ru2(OH)2(3,6-Bu2Q)(btpyan)](SbF6)2; structure given in (Fig. 3), is water insoluble, Tanaka and co-workers (68) were able to
demonstrate limited electrocatalytic activity by constant potential electrolysis


MECHANISMS OF WATER OXIDATION CATALYZED

9

Figure 3. The DFT predicted mechanism for water oxidation catalyzed by [Ru2(OH)2(3,6-Bu2Q)
(btpyan)]2þ ion [3,6-Bu2Q ¼ 3,6-di-tert-butyl-1,2-benzoquinone and btpyan ¼ 1,8-bis(2,20 :60 ,200 -terpyrid-40 -yl)anthracene]. Two proton-coupled electron transfer (PCET) steps on the resting form of the
catalyst (top) lead to oxidation of juxtaposed hydroxo ligands, which couple to form a bridging superoxo
ion (bottom), with the additional electron being distributed over the quinone ligands. Further PCET
reoxidizes the quinones, leading to incorporation of solvent into the coordination sphere (left); at this
point, the superoxo ligand is terminally coordinated. The final PCEToxidizes the superoxide and returns
the catalyst to its original form. The RIMP2 calculated geometric structure of the complex ion

containing 3,5-dimethyl-substituted quinone ligands (in place of tert-butyl substituents) is shown
within the catalytic cycle. [Adapted from (56).] (See the color version of this figure in Color Plates
section.)

(CPE) in trifluoroethanol containing 10% water. When the complex was deposited
as a solid on an indium–tin oxide (ITO) electrode, remarkably efficient electrocatalyzed water oxidation could be achieved in aqueous media, with O2 evolution
turnover numbers per catalyst molecule exceeding 33,000 being measured.
However, the catalytic rate constant was very low. Several structurally similar
complexes containing modifications within the bridging group (xanthene for
anthracene) of the templating macrocyclic ligand (59) or different substituted
quinones (46) have been prepared in efforts to improve catalytic rates within this
class of compounds. However, to date, none of these complexes have been found to
exhibit detectable electrocatalytic activity.


10

AURORA E. CLARK AND JAMES K. HURST

The aqueous insolubility and the “noninnocent” nature of the quinone ligands
present formidable challenges to characterization of the “Tanaka complex”, as it is
now known, in its various accessible oxidation states. In particular, the complex
is representative of a large class of Ru–NIL (NIL ¼ noninnocent ligand) complexes whose ligand and metal orbitals are extensively mixed, giving rise to
apparent noninteger oxidation states and nearly isoenergetic electronic states with
differing spin multiplicities (56, 60), so that even ground-state configurations are
difficult to assign. Despite the challenges, mechanistic analyses of this reaction
have been carried forward with considerable success by the Tanaka and Fujita/
Muckerman groups using a combination of experimental and theoretical
approaches. These efforts have been aided by the availability of a model of the
“half-molecule”, (i.e., [Ru(H2O)(3,5-Bu2Q)(tpy)]2þ) (61). Although apparently

not capable of oxidizing water itself (46), this ion is more amenable to computational and physical analyses than the dimer. Controversies concerning
the ground-state representation of this ion, prevalent in the earlier literature
(46, 61), appear to have been recently resolved through in-depth electrochemical,
spectroscopic, and computational analyses (47, 56).
The computational studies utilized a combination of DFT, time-dependent DFT
(TD-DFT) (using the B3LYP functional) and CASSCF (complete active space
self-consistent field) methodologies to probe the relative energies of the various
available spin states of the reaction intermediates. Despite the relative simplicity of
the monomer relative to the dimer, significant computational difficulty was
encountered. Although the authors utilized the broken-spin broken-symmetry
(BS/BS) method (62–64) to obtain open-shell singlet states, a wide variety of hS2i
values were observed, indicating spin contamination from alternative S states with
the same Ms values. Indeed, spin contamination was even observed for the openshell triplet states using DFT. Interestingly, the authors avoided using the
Noodleman’s spin projection correction to the BS/BS singlet-state energy within
their calculations, perhaps due to the large amount of spin contamination observed
in the open-shell singlet states. To further test the relative energies of the various
spin states, the authors utilized TD-DFT to examine which spin states were higher
than the predicted ground state. Unfortunately, many of the excited states
encountered were charge transfer (CT) in nature, bringing into question the
reliability of the calculations, as DFT (specifically density functionals without
long-range corrections) is known to perform very poorly for CT excitations (65).
The results for the “half-molecule” most relevant to the catalytic activity of
the binuclear ion are that the best description of the formal oxidation state of
the aquo complex is [RuII(H2O)(Q)(tpy)]2þ, rather than the initially proposed
[RuIII(H2O)(SQ. À)(tpy)]2þ (SQ. À ¼ 3,5-di-tert-butylbenzosemiquinone) (61),
and that sequential deprotonation leads to [RuII(OH)(Q)(tpy)]þ and [RuII(O. À)
(SQ. À)(tpy)]0. The doubly deprotonated molecule is unique in possessing an oxyl
radical ligand, formed by internal transfer of an electron to the quinone.

MECHANISMS OF WATER OXIDATION CATALYZED

11

This radical is expected to be highly reactive and, in experimental systems, appears
to abstract a hydrogen atom (from unspecified sources) to give [RuII(OH)(SQ. À)
(tpy)]0 as the final product. The calculated electronic spin states for these three
protonation states are difficult to assign using DFT, as spin contamination is
observed for the varying states. As such, the hS2i values were interpreted in terms
of a simple generalized valence bond configuration interaction (GVB-CI) within a
(2,2)CAS type model as in stretched H2. This interpretation suggests that low-lying
singlet, open-shell singlet, and triplet spin multiplicities can exist that contain Ru
in formal oxidation states ranging from Ru(II) to Ru(IV) (47).
The water-oxidizing capacity of the dinuclear catalyst is attributed to formation
of intermediates similar to [RuII(O. À)(SQ. À)(tpy)]0, in which the templating
btpyan ligand juxtaposes the coordinated oxyl groups to direct O–O bond
formation via radical coupling (Fig. 3). These researchers originally proposed a
mechanism based upon DFT computational results in which sequential deprotonation of the resting form of the catalyst ([(RuII)2(OH)2(Q)2(btpyan)]2þ) led to an
intermediate containing a bridging superoxide anion with electron density shifting
to the quinone ligands (i.e., best described as [(RuII)2(O2À)(QÀ1.5)2(btpyan)]0),
following which net 4eÀ oxidation led to release of O2 with regeneration of the
resting form of the catalyst (46). More recently, this mechanism has been modified
so that the overall cycle contains a series of four PCET steps (Fig. 3) (56). Here, the
resting form of the catalyst is indicated as an asymmetrically hydrogen-bonded
pair of coordinated hydroxo ligands. The intermediate formed following the first
PCET step contains an oxyl anion that is stabilized by hydrogen-bonding to the
adjacent hydroxyl ligand. Loss of this proton in the second oxidation step then
allows O–O bond formation, in which the 1,2-bridging O2 group is formulated as
superoxo anion with the additional electron density shifting to the quinone ligands.
Subsequent PCET then leads to formation of a terminally coordinated superoxo

anion via addition of solvent and, in the final step, oxidation of the coordinated
O2. À releases O2, closing the catalytic cycle.
One remarkable feature of this reaction as written is that the Ru ions do not
change their formal oxidation states throughout the cycle. Instead, redox changes
occur primarily through complementary changes in electron density in orbitals that
are centered in the oxo and quinone ligands and reflect the highly delocalized
character of the frontier orbitals in this coordination complex. Nonetheless, the
complex nature of the wave functions observed here and elsewhere, as well as the
small energy differences between spin states, call for more thorough computational studies. In particular, note that few benchmarking calculations have been
performed on Ru catalysts so as to understand more broadly the performance of
various density functionals and how that performance changes with varying
systems. While it is becoming more commonplace for CASSCF and CASPT2
methods to be used in conjunction with DFT, this needs to become standard
practice and researchers must ensure that the size of the active space in which the


12

AURORA E. CLARK AND JAMES K. HURST

electronic excitations are allowed to occur is sufficiently large to capture the
essential aspects of the wave function. Moreover, in both DFT and wave function
based methods, benchmarking of the basis sets used to describe the metal and
ligands must be performed. To our knowledge, no studies have examined the basis
set dependence of the reaction energetics and spin state distributions, nor have any
attempts been made to extrapolate any type of basis set or Kohn–Sham limit for any
methodology employed. Similarly, no examples exist that have benchmarked
the performance of varying continuum approximations and their effects upon the
reaction energetics.

III. HOMOLYTIC CLEAVAGE OF O–H BONDS:
THE “BLUE DIMER”
A. Structure
The water oxidizing capacity of the m-oxo bridged cis,cis-[RuIII(bpy)2
(H2O)]2O4þ “blue dimer” (hereafter identified as {3,3}) was originally reported
by Meyer’s group (4) in 1982. For the ensuing $20 years, this ion and structural
analogues bearing substituted bipyridine ligands were the only known homogeneous catalysts for water oxidation whose reactivity could be reproducibly
demonstrated (5, 31–35). Correspondingly, they are the ions whose physical
properties and reactivities have been most extensively investigated. X-ray
crystallographic analyses of {3,3} (5) and {3,4} (as the dihydroxy-ligated
[Ru(bpy)2(OH)]2O3þ ion) (66) reveal a nearly linear oxo-bridge and torsional
dislocation about the Ru–O–Ru bond that places the O atoms of the adjacently
˚ . The DFT calculations
coordinated H2O or OH ligands at a distance of $4.5 A
indicate that this same general orientation is maintained in the chemically unstable,
catalytically relevant higher oxidation states of the complex (Fig. 4) (67), and the
near-linear bridging character of the Ru–O–Ru bond over the entire range of
accessible oxidation states ({3,3} to {5,5}) has been experimentally confirmed by
resonance Raman (RR) measurements of the 18 O isotope-dependent frequency
shifts occurring in the ns(Ru–O–Ru) symmetric stretching vibrational modes (68).
CASSCF methods have characterized the electronic ground state of {3,3} as
a weakly antiferromagnetically coupled singlet (43). In the computed structures,
progressive oxidation of the metal centers leads primarily to modest shortening
of the metal–ligand bonds throughout the complex accompanied by an increase
in the torsional angle between the adjacently coordinated terminal oxo ligands,
the net effect being that their critical OÁ Á ÁO distances do not change appreciably
upon oxidation (67). Consequently, although compositionally similar to
the bis(pyridyl)pyrazolate-bridged diruthenium complex recently described by

MECHANISMS OF WATER OXIDATION CATALYZED

13

Figure 4. The B3LYP/6-31GÃ /LANL2DZ high-spin ferromagnetically coupled optimized conformation of the “blue dimer” in its catalytically active {5,5} ([Ru(bpy)2(O)]2O4þ) oxidation state.

Llobet and associates (53, 54), the conformational differences suggest that a
significant activation barrier to intramolecular coupling of oxo atoms may exist
in the “blue dimer” arising from the molecular distortions required to bring these
groups into close contact. Indeed, the 18 O isotope labeling studies described below
reveal that these two dimers catalyze water oxidation by distinct mechanisms.
B. Redox States
Extensive mechanistic investigations have been undertaken by two groups, who
have generally used alternative approaches of analysis (67, 69). Although this has
led to somewhat different viewpoints, particularly concerning the nature of reaction
intermediates, the groups concur that the oxygen-evolving form of the catalyst is
{5,5}, an oxidation state in which the coordinated water molecules have been fully
deprotonated to generate ruthenyl oxo atoms, (i.e., [RuV(bpy)2(O)]2O4þ). The
identity of this species was first inferred by Meyer and co-workers (5) using
electrochemical analyses and later confirmed by redox titrations in our laboratory,
which made use of a columnar flow-through carbon fiber electrode for fast CPE (70).
Resonance Raman spectroscopy clearly identified Ru¼O stretching vibrational
modes in the {5,5} ion at $800 cmÀ1 (Fig. 5) (70, 71); furthermore, {5,5} underwent
first-order decay with a rate constant that was equal to the rate constant for O2
evolution measured under steady-state catalytic conditions (70, 72, 73).
Under most experimental conditions, CVs of the “blue dimer” in water exhibit two
well-defined oxidation waves above {3,3} whose relative amplitudes indicate that
they are {3,4} and {4,5}, as well as an additional wave that just precedes the onset of

14

AURORA E. CLARK AND JAMES K. HURST

Figure 5. Resonance Raman spectroelectrochemical titration of the “blue dimer” {3,4} ion in 0.5 M
CF3SO3H. The inset shows the low-frequency spectra of the various detectable oxidation states. Bands
highlighted in light gray are the Ru–O–Ru symmetric stretching frequency and its first overtone; the
band highlighted in dark gray (lowest trace) is the stretching frequency of the terminal Ru¼O bond.
[Adapted from (70).]

solvent breakdown (5, 23). Based upon this behavior, one can assign the following
sequence of accumulating redox states: {3,3} ! {3,4} ! {4,5} ! {5,5}. These
potentials are pH dependent, reflecting the different states of protonation of the
coordinated aquo ligands under varying medium conditions. Below pH 2, the two
more anodic waves coalesce, so that the voltammograms appear as two waves with
relative amplitudes of 1:3, indicating that the higher oxidation step appears as the
three-electron (3eÀ) process: {3,3} ! {3,4} ! {5,5} (5). However, an intermediate
species can still be detected when more sensitive methods are used. For example,
redox spectrometric titrations utilizing the flow CPE cell described above with RR
detection clearly demonstrate the accumulation of an intermediary oxidation state at
potentials slightly lower than those required to oxidize the complex to {5,5} (Fig. 5);
furthermore, decay of flow CPE prepared {5,5} is biphasic, with the first step
proceeding to an intermediary species that only slowly converts to {3,4}, the highest
stable oxidation state (70). The identity of this intermediate has been controversial.
Based primarily upon titrimetric and transient kinetic studies using Ce4þ as oxidant


MECHANISMS OF WATER OXIDATION CATALYZED

15

and employing global kinetic analysis for spectral deconvolution, Meyer and coworkers assigned this oxidation state as {4,5} (55, 74); their kinetic analyses identified
{4,4} as an unstable transient species whose concentration levels were vanishingly
small. However, several different titrimetric measurements made in our laboratory
using flow CPE prepared solutions in various oxidation states (70), as well as direct
titration with Ce4þ (71) indicate that the accumulating intermediary oxidation state is
actually {4,4}. Recent RR and optical spectroscopic measurements have confirmed
this assignment. Specifically, as anticipated from the CV analyses (5), {4,5} contains
ruthenyl bonds, which are readily detected in the RR spectrum by their isotopesensitive Ru¼O stretching modes (23). These bands are not observed in the
intermediate that accumulates in acidic solutions, however (Fig. 5) (70). Furthermore,
the optical spectrum of {4,4} determined in neutral solutions by pulse radiolysis is
unlike that of {4,5}, but identical to the spectrum of the accumulating intermediate in
acidic solutions (23). Assignment as {4,4} is also supported by pH jump experiments
in which solutions of {4,5} are rapidly acidified. One observes by X-band EPR
spectroscopy the immediate formation of {5,5}, but no {3,4}, the inference being that
the other accumulating oxidation state is {4,4}, which is EPR silent. Upon standing,
the EPR signal of {3,4} slowly appears as the signal associated with {5,5} disappears at
a rate characteristic of water oxidation; that is, the following reaction sequence:
2{4,5} ! {4,4} þ {5,5} ! ! (redox decay to {3,4} and O2) (Fig. 6) (23).
Collectively, this body of evidence forms overwhelming support for the reaction
sequence (Scheme 1), in which {4,4} is the accumulating intermediary state in acidic
solutions, but {4,5} is the accumulating state under more alkaline conditions:
(a)

(b)

{3,4}

1.30 V
1.35 V

1.50 V

2800

{5,5}
3200

3600

Magnetic field (G)

4000

340

360

380

Magnetic field (mT)

Figure 6. The X-band cryogenic EPR spectra of paramagnetic “blue dimer” oxidation states in 0.5 M
CF3SO3H. Panel a: spectra of {3,4} and {5,5} formed by flow CPE at the indicated potentials (vs NHE);
panel b: spectral changes accompanying a pH jump of {4,5} from pH 7 to 0.3. Formation of {5,5} is
immediate and its subsequent decay is accompanied by slow accumulation of {3,4}, consistent with the
reaction sequence: 2{4,5} ! {4,4} þ {5,5} ! ! ! {3,4}. [Adapted from (70 and 23).]