Advances in Inorganic Chemistry - Rudi Van Eldik, Colin D. Hubbard, 2017

ADVISORY BOARD
L. H. Gade D. Darensbourg
Universität Heidelberg Texas A & M University
Germany College Station, Texas, USA
M. L. H. Green H. B. Gray
University of Oxford California Institute of Technology
Oxford, United Kingdom Pasadena, California, USA
A. E. Merbach P. A. Lay
Laboratoire de Chimie et Bioanorganique EFPL, University of Sydney
Lausanne, Switzerland Sydney, Australia
P. J. Sadler J. Reedijk
University of Warwick Leiden University
Warwick, England Leiden, The Netherlands
K. Wieghardt Y. Sasaki
Max-Planck-Institut Hokkaido University
Mülheim, Germany Sapporo, Japan
Academic Press is an imprint of Elsevier
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
525 B Street, Suite 1800, San Diego, CA 92101-4495, United States
The Boulevard, Langford Lane, Kidlington, Oxford, OX5 1GB, United Kingdom
125 London Wall, London, EC2Y 5AS, United Kingdom
First edition 2017
Copyright © 2017, Elsevier Inc. All rights reserved
No part of this publication may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, recording, or any information storage and
retrieval system, without permission in writing from the publisher. Details on how to seek
permission, further information about the Publisher’s permissions policies and our
arrangements with organizations such as the Copyright Clearance Center and the Copyright
Licensing Agency, can be found at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by
the Publisher (other than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and
experience broaden our understanding, changes in research methods, professional practices,
or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described
herein. In using such information or methods they should be mindful of their own safety and
the safety of others, including parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors,
assume any liability for any injury and/or damage to persons or property as a matter of
products liability, negligence or otherwise, or from any use or operation of any methods,
products, instructions, or ideas contained in the material herein.
ISBN: 978-0-12-812834-3
ISSN: 0898-8838
For information on all Academic Press publications

visit our website at https://www.elsevier.com/books-and-journals
Publisher: Zoe Kruze

Acquisition Editor: Poppy Garraway
Editorial Project Manager: Shellie Bryant
Production Project Manager: Vignesh Tamil
Cover Designer: Greg Harris
Typeset by SPi Global, India
DEDICATION
Volume 70 of the Series is dedicated to Professor R.G. Wilkins on the

occasion of his 90th birthday in recognition of his contributions to Inorganic
Reaction Mechanisms.
RALPH G. WILKINS
Ralph Wilkins, born in 1927, grew up in the Southampton area of England.
He attended the University of Southampton and received his BSc and PhD
degrees in Chemistry there. He conducted research at the ICI research
laboratories in Welwyn for 3 years, followed by postdoctoral research with
Arthur Adamson at the University of Southern California. Afterwards he
was appointed as a Lecturer and subsequently as a Senior Lecturer at the
University of Sheffield. In 1962 he began a year as a Visiting Professor at
the Max Planck Institute in G€ ottingen, Germany, with Manfred Eigen.
During this time results, gathered principally from reports of ultrasonic
absorption experiments in G€ ottingen, temperature-jump relaxation exper-
iments in G€ ottingen and MIT, stopped-flow experiments in Sheffield, and
NMR spectroscopy water-exchange experiments in Berkeley, led to
the formulation of what became known as the Eigen–Wilkins mechanism
for transition metal complex formation. At that time these experimental
methods were relatively novel. Ralph Wilkins was then appointed to a
Professorship of Chemistry at the State University of New York in Buffalo,
and some years later as Professor and Chairman of the Department of
Chemistry at the State University of New Mexico in Las Cruces. His
research interests expanded to include bioinorganic chemistry and many
publications in this area, including books cited below, ensued. Completed
during a Visiting Professorship at the University of Warwick, United
Kingdom, in 1991, “Kinetics and Mechanism of Reactions of Transition
Metal Complexes” was a revised version of his earlier classic work, “The
Study of Kinetics and Mechanisms of Transition Metal Complexes,” pub-
lished in 1974. Other books include “Inorganic Chemistry in Biology”
(with P.C. Wilkins), in 1997, and in 2003, “The Role of Calcium and
Comparable Cations in Animal Behaviour,” also with P.C. Wilkins. In
retirement Ralph has extended his interests, beyond the highs and lows
of the England national cricket team, to genetics, and this has culminated
in the very recent appearance, Spring 2017, of “Animal Genetics for
v
vi Dedication
Chemists,” published by the Royal Society of Chemistry. We pay tribute

to his versatile, remarkable, outstanding, enduring career in chemistry
and other science endeavors, and honor Ralph on the occasion of his
90th birthday, as one of the pioneers of Inorganic Reaction Mechanisms.
CONTRIBUTORS
Andres G. Algarra
Instituto de Biomoleculas (INBIO), Facultad de Ciencias, Universidad de Cádiz, Polı́gono
Universitario Campus Rio San Pedro, Puerto Real, Spain
Gabriel Aullón
Secció de Quı́mica Inorgànica, Facultat de Quı́mica, Universitat de Barcelona, Barcelona,
Spain
Manuel G. Basallote
Instituto de Biomoleculas (INBIO), Facultad de Ciencias, Universidad de Cádiz, Polı́gono
Universitario Campus Rio San Pedro, Puerto Real, Spain
Gábor Beller
University of Debrecen, Debrecen, Hungary
Anna Company
Grup de Quı́mica Bioinspirada, Supramolecular i Catàlisi (QBIS-CAT), Institut de Quı́mica
Computacional i Catàlisi (IQCC), Universitat de Girona, Girona, Catalonia, Spain
Miquel Costas
Margarita Crespo
Spain
Janusz M. Da˛browski
Faculty of Chemistry, Jagiellonian University, Kraków, Poland
Sam P. de Visser
Manchester Institute of Biotechnology, School of Chemical Engineering and Analytical
Science, The University of Manchester, Manchester, United Kingdom
István Fábián
Abayomi S. Faponle
Manchester Institute of Biotechnology, School of Chemical Engineering and Analytical
Science, The University of Manchester, Manchester, United Kingdom; Faculty of Basic
Medical Sciences, Obafemi Awolowo College of Health Science, Olabisi Onabanjo
University, Ogun State, Nigeria
Yuichi Himeda
Research Institute of Energy Frontier, National Institute for Advanced Industrial Science and
Technology, Tsukuba, Japan
Deogratius Jaganyi
School of Chemistry and Physics, University of KwaZulu-Natal, Scottsville,
Pietermaritzburg, South Africa
xi
xii Contributors
Jesús Jover
Spain
József Kalmár
MTA-DE Homogeneous Catalysis and Reaction Mechanisms Research Group, University
of Debrecen, Debrecen, Hungary
Hajime Kawanami
Research Institute for Chemical Process Technology, National Institute for Advanced
Industrial Science and Technology, Sendai, Japan
Gábor Laurenczy
cole Polytechnique Federale de Lausanne
Institut des Sciences et Ingenierie Chimiques, E
(EPFL), Lausanne, Switzerland
Allen Mambanda
School of Chemistry and Physics, University of KwaZulu-Natal, Scottsville,
Pietermaritzburg, South Africa
Juan P. Marcolongo
Facultad de Ciencias Exactas y Naturales and INQUIMAE, Universidad de Buenos Aires/
CONICET, Ciudad Universitaria, Buenos Aires, Argentina
Manuel Martı́nez
Spain
Jose A. Olabe
Joan Serrano-Plana
Leonardo D. Slep
Alexander B. Sorokin
Institut de Recherches sur la Catalyse et l’Environnement de Lyon IRCELYON, UMR
5256, CNRS—Universite Lyon 1, Villeurbanne cedex, France
Mária Szabó
Ari Zeida
PREFACE
The 3rd European Colloquium on Inorganic Reaction Mechanisms, held in

Krakow, Poland, in June 2016, cochaired by Rudi van Eldik and Grazyna
Stochel, provided an excellent opportunity for the Advances in Inorganic
Chemistry Series. In addition, we dedicate this volume to Ralph Wilkins,
one of the pioneers of Inorganic Reaction Mechanisms, who celebrates
his 90th birthday in 2017.
Several of the principal speakers at the 3rd ECIRM have graciously
accepted invitations to elaborate on their Colloquium presentation to
produce a comprehensive but concise chapter for this volume. The subject
matter can now reach a much wider community of inorganic chemistry
reaction mechanism investigators. Volume 70 contains an eclectic set of
contributions; however, it will be shown that within several topics a
common thread involves computational methods that support experimen-
tally derived mechanistic proposals and can provide details of mechanisms
and thus more confidence in them. The volume begins with an authoritative
account, Chapter 1, of kinetics and mechanism methods for complex redox
reactions by István Fábián. Chapter 2, by Miquel Costas, describes oxygen–
oxygen bond activation in copper- and iron-based coordination complexes.
Alexander Sorokin (Chapter 3) and Sam de Visser (Chapter 4) both describe
bioinorganic chemistry inspired by actual biological chemistry; the former
author discusses the unusual structure and interesting catalytic properties
of μ-nitrido diiron phthalocyanine and porphyrin complexes, while de
Visser explores the role of nonheme transition metal oxo, -peroxo, and
-superoxo intermediates in enzyme catalysis. These are followed by two
contributions by Manuel Martı́nez (Chapter 5) and Deogratius Jaganyi
(Chapter 6) describing various aspects of reactions of platinum(II) com-
plexes; diaryl platinum(II) scaffolds for kinetics and mechanistic studies on
the formation of platinacycles are the subject of the former, while the control
of the lability of square planar platinum(II) complexes through electronic
factors and π-conjugation is the subject of the latter. Jose Olabe provides
a fascinating account of “cross talk” between nitric oxide and hydrogen
sulfide, by referring to thionitrous acid/thionitrite and perthionitrite
intermediates (Chapter 7). Manuel Basallote, in Chapter 8, addresses,
through a lucid account, experimental and computational insights into
the reactivity at the sulfur atoms of M3S4 (M ¼ molybdenum or tungsten)
xiii
xiv Preface
and the mechanism of [3 + 2] cycloaddition with alkynes. In Chapter 9,

Janusz Da˛browski presents a detailed account of reactive oxygen species
in photodynamic therapy in which transition metal complexes can play
an important role in efficacy applications. The development of formic acid
as a source of hydrogen for energy purposes and environmental favorability
is the topic covered in Chapter 10 (Hajime Kawanami, Yuichi Himeda, and
Gábor Laurenczy); the involvement of metal-based catalysts for the conver-
sion of formic acid to hydrogen is discussed.
We are most appreciative of the efforts of the authors in their production
of such readable accounts supported by excellent figures and schemes. In
addition, we acknowledge their colleagues, who are cited as coauthors of
the corresponding authors, and who have each contributed to bring the pro-
ject to fruition. We believe this volume will be of value to research students
and many investigators in the earlier stages of their career, as well as provid-
ing a useful update regarding the current status of topics for more established
investigators within inorganic chemistry. We invite readers in other fields of
chemistry to avail themselves of the opportunity to acquaint themselves with
the expert-delivered chemistry contents herein.
COLIN D. HUBBARD
Co-Editor of this volume
Professor Emeritus of Chemistry, University of New Hampshire,
Durham, United States of America
RUDI VAN ELDIK
Editor of Advances in Inorganic Chemistry
Emeritus Professor of Inorganic Chemistry,
University of Erlangen–Nuremberg, Germany
Professor of Inorganic Chemistry,
Jagiellonian University, Krakow, Poland
April 2017
CHAPTER ONE
The Kinetics and Mechanism

of Complex Redox Reactions
in Aqueous Solution: The Tools
of the Trade
r*, József Kalmár†, István Fábián*,1
Mária Szabó*, Gábor Belle
*University of Debrecen, Debrecen, Hungary
†
MTA-DE Homogeneous Catalysis and Reaction Mechanisms Research Group, University of Debrecen,
Debrecen, Hungary
1
Corresponding author: e-mail address: ifabian@science.unideb.hu
Contents
1. Introduction 2
2. General Considerations 3
3. Selected Reactions of Oxychlorine Species 6
4. Kinetics of the Oxidation Reactions of Peroxo Compounds 22
5. The Photon as a Reactant 35
6. Selected Kinetic Studies on Heterogeneous Systems 49
7. Concluding remarks 55
Acknowledgments 57
References 57
Abstract
Redox reactions of simple inorganic species exhibit an amazingly rich variety of complex
kinetic phenomena. Typically, these reactions are interpreted on the basis of multistep
kinetic models which postulate the formation and subsequent fast reactions of reactive
intermediates. The main purpose of this chapter is to demonstrate the challenges asso-
ciated with mechanistic studies on complex redox reactions, and to offer selected exam-
ples how the complexities can be handled with currently available experimental and
computational methods. Clear arguments are presented to demonstrate that the stoi-
chiometries of these reactions are kinetically controlled. It is shown that in order to
understand the intimate details of these systems, the stoichiometry as a function of
reaction time, the final stoichiometry and the kinetic properties need to be studied
under as broad experimental conditions as possible. Furthermore, thorough character-
ization of the reactive intermediates is the key to in-depth understanding of the mech-
anism. The importance of photoinitiation and kinetic coupling between photochemical
and thermally activated reaction steps is also demonstrated in several systems. The sur-
vey of the literature results confirms that simultaneous and critical evaluation of all
Advances in Inorganic Chemistry, Volume 70 # 2017 Elsevier Inc. 1

ISSN 0898-8838 All rights reserved.
http://dx.doi.org/10.1016/bs.adioch.2017.02.004
2 Mária Szabó et al.
available experimental results is essential to validate the mechanistic conclusions.

Finally, it is shown that adapting the methodology of homogeneous reaction kinetics
for studying nonhomogeneous physicochemical processes leads to unique kinetic
information regarding the kinetics of adsorption and desorption processes.
1. INTRODUCTION
Redox reactions in aqueous solution exhibit a large variety both in
terms of kinetics and stoichiometry. The basic principles of simple electron
transfer processes are well established and there is a great number of publi-
cations which illustrate the excellent agreement between theoretical models,
most prominently the Marcus–Hush theory (1,2), and experimental data.
Complementary one-electron transfer processes are typically characterized
with straightforward second-order rate expressions which are first order
for both reactants. Simultaneous transfer of two or more electrons is less
likely, thus, complementary two-electron transfer reactions frequently
imply atom transfer. This may introduce some complications in the kinetic
features of such systems. By definition, noncomplementary redox reactions
proceed via the formation of one or more intermediates and require multi-
step kinetic models for the interpretation of the experimental data. Apart
from the redox steps, these models also include the equilibrium reactions
between the components (3). Depending on the relative rates of the indi-
vidual reaction steps simple or complex kinetic features may arise. Ulti-
mately some of these systems show nonlinear dynamic behavior under
specific experimental conditions.
Earlier, technical limitations prevented the assessment of the intimate
details of many complex redox processes. Recent developments in experi-
mental and computational methodologies opened new avenues in this field.
Quite often old but well-established principles are used; however, the per-
formance of the instruments is boosted by new technical improvements. In
other cases new approaches have been introduced. The advent of these
developments is that more reliable and larger experimental datasets, than
before, have become accessible on a routine basis. This, obviously, led to
the postulation of better defined kinetic models, and in many cases impor-
tant details of complicated reaction mechanisms have been explored. The
use of new methods is not without risks, because blind trust in sometimes
indirect pieces of information may lead to biased or wrong conclusions.
The Kinetics and Mechanism of Complex Redox Reactions 3
The main purpose of this chapter is to demonstrate the challenges asso-

ciated with mechanistic studies on complex reactions and to offer a few
examples of how the complexities can be handled with currently available
experimental and computational methods. It will also be shown that the
same kinetic approaches provide invaluable information on the mechanism
of physicochemical processes in heterogeneous systems.
2. GENERAL CONSIDERATIONS
Complex redox reactions exhibit individual properties and sometimes
the same or very similar kinetic models are not suitable for the interpretation
of the results even in strongly related reactive systems. The complexity of the
kinetic behavior is always associated with a network of individual reaction
steps, illustrated in Scheme 1.
In these systems, first the reactants produce one or more reactive inter-
mediates. At least one such a step is required to initiate the overall process.
The initiation step(s) may quickly diminish, or be operative over the whole
course of the reaction.
The reactive intermediates open a series of new reaction paths by
reacting with each other and/or the reactants. Quite often the intermediates
are regenerated in cyclic processes. For example, in catalytic systems the
active form of the catalyst reacts with one or more reactants and eventually
a sequence of reactions leads to the regeneration of the catalyst. In other
cases, a reactive intermediate accelerates its own formation and an autocat-
alytic pattern arises. The interplay of the competing reaction paths ultimately
leads to the formation of the products.
In earlier studies, typically a relatively simple and forthright approach was
used to develop kinetic models for complex redox reactions. The experi-
mental conditions were simplified by introducing various approximations
and the results were evaluated by simplified rate expressions. It was assumed
that additional reactions do not interfere with the individual reaction step
R1 l1 P1
R2 l2 P2
... ... ...
Rp lp Pp
Scheme 1 The general scheme of a complex redox reaction; R: reactant, I: intermediate,

P: product.
studied, often without testing the validity of this assumption. In other words,
kinetic coupling between the reaction steps was neglected. Consequently,
biased results were obtained for the rate constants, and the combination
of the corresponding data led to false conclusions. Kinetic studies on ozone
(O3) decomposition in aqueous solution serve as examples for this problem.
Because of its relevance in practical applications—such as water treatment
technologies, advanced oxidation processes (AOPs), preparative organic
chemistry, etc.—the kinetics and mechanism of this reaction have exten-
sively been studied. Two reasonably well-detailed kinetic models have been
developed for the interpretation of various aspects of aqueous ozone decom-
position (4–8). Most of the chain carrier radicals are the same in these models
but there are marked differences in the number of reaction steps and their
rate constants. The models were assembled using experimental data and ear-
lier literature results published on the reactions of the reactive intermediates.
Simulations on the basis of the proposed models could not reproduce the
experimental observations and led to the conclusion that none of them is
suitable for quantitative interpretation of aqueous ozone decomposition
(9). This failure was attributed to the following shortcoming of the proposed
models. The significance of several reaction steps was overestimated and the
approximations introduced during the evaluation of the individual kinetic
parameters led to a skewed set of rate constants. The calculations also dem-
onstrated the need for further experimental work and more accurate kinetic
parameters.
The fate of a complex redox reaction is determined by the competition
of parallel and subsequent reaction paths, in other words, by their relative
kinetic weights which may significantly change over the course of a reaction.
Thus, the stoichiometry of the reaction is controlled by the kinetics and it
may change as a function of reaction time. Exploring the intimate details of
these reactions requires the identification and thorough description of the
reactive intermediates. The inherent complication is that once the reactive
intermediates are formed they are typically involved in fast subsequent reac-
tion steps and present at very low concentration levels. These transient spe-
cies can be detected, if at all, by using specific experimental techniques.
Perhaps the most obvious example for this problem is the quantification
of the hydroxyl radical in solution. This species is an extremely powerful
oxidant which quickly reacts with a large variety of substrates (10). It can
be generated by different experimental methods but detected only via its fast
reaction with specific agents. However, introducing additional reactants
into a reactive system may complicate the kinetics even further. Indirect
detection of a reactive intermediate species always implies various approx-
imations with obvious uncertainties transferred into the final conclusions. In
some cases, direct observation and characterization of these species may be
feasible. This requires meticulous experimental work which pays off in bet-
ter defined kinetic models.
As an example, aqueous ozone decomposition serves as an excellent
example again. The formation and consecutive decay of the ozonide, super-
oxide, and carbonate ion radicals could be detected directly in this system
(11–13). These species have outmost importance in the overall process
and the characteristic kinetic traces were recorded under various experimen-
tal conditions. Simultaneous evaluation of all experimental data was instru-
mental in exploring the specific details of the mechanism. Discrepancies in
the literature data regarding the rate constant of the initiation step were
resolved and a comprehensive kinetic model was worked out which is suit-
able to explain the kinetic observations in the absence, as well as in the pres-
ence of hydrogen peroxide (a well-known catalyst of the reaction) and
carbonate ion (an inhibitor).
In subsequent parts of this chapter, we will cover selected topics in
mechanistic studies. The reactions of simple oxidants such as oxychlorine
species (Section 3) and peroxo compounds (Section 4) are discussed for
two reasons. First of all, these species have great practical significance in
AOPs, environmental chemistry, and biological systems. Exploring the
kinetics of the relevant reactions is of enormous importance in understand-
ing the behavior of these oxidants under a variety of conditions. On the
other hand, handling the mechanistic challenges associated with these sys-
tems has broader implications in reaction kinetics. In Section 5, it will be
demonstrated that the photon can be used as a reactant to control the for-
mation of several key intermediates in a complex reactive system. This
offers an unique experimental tool for mechanistic studies. Recent instru-
mental developments in this direction also opened new ways for studying
the kinetics of surface processes. It will be shown in Section 6 that the
methods used for mechanistic studies in homogeneous systems can effi-
ciently be adapted to explore heterogeneous reactions. Comprehensive
kinetic models for redox reactions coupled with adsorption processes
and diffusion phenomena in the confined space of high-porosity materials
lead to a new approach for the interpretation of heterogeneous catalytic
and photoinitiated processes.
3. SELECTED REACTIONS OF OXYCHLORINE SPECIES

Stable forms of chlorine may exist in aqueous solution in any oxida-
tion state between 1 and +7 with the exception of +2 and +6. Perchlorate
ion (ClO4 ) is a well-known pyrotechnic material which is inert in aqueous
solution. For this reason its salts are preferred as inert background electro-
lytes. Nevertheless, ClO4 may undergo reduction to chlorite ion
(ClO2 ) via chlorate ion as an intermediate in biological systems. This
reaction is catalyzed by perchlorate reductase enzyme which is a molybde-
num enzyme (14). Chlorite ion is further reduced to chloride ion and
dioxygen by chlorite dismutase enzyme which is a soluble heme-containing
protein (15).
Chlorate ion (ClO3 ) is also a powerful oxidant but relatively inert in
water. It is involved in various redox reactions only under very acidic con-
ditions. For example, its reaction with chloride ion (Eq. 1) was used to pro-
duce chlorine dioxide for large-scale industrial bleaching processes.
2ClO3 + 2Cl + 4H + ! 2ClO2 + Cl2 + 2H2 O (1)

This reaction exhibits complex kinetic features and the [Cl2]/[ClO2]
product ratio is strongly dependent on the initial concentration ratio of
the reactants (16). Recently, a characteristic induction period was found
in the kinetic traces of ClO2 formation (17). A multistep kinetic model
was proposed for this reaction which includes three equilibrium and one
irreversible reaction steps. The individual rate constants were selected such
that the simulated kinetic traces fit all the experimental ones reasonably well.
The kinetic model postulates the formation of HClO2, HOCl, Cl2 O3 2 ,
H2Cl2O3, and Cl2O2 as reactive intermediates. The first two compounds
are relatively stable and were characterized under appropriately selected
experimental conditions. The other three transient species could never be
detected but undoubtedly are required for the proper interpretation of
the experimental results.
Chlorine dioxide is a powerful, widely used one-electron oxidant for
various purposes. In water treatment technologies, its antibacterial and ant-
iviral properties are utilized and for the same reason it is used in che-
mosterilization and disinfection of anthrax. As an aqueous sanitizing
agent, it finds widespread applications within the food industry such as meat,
poultry, fish, vegetable processing, and other uses. It is also a preferred
whitening agent in the pulp and textile industries. Chlorine dioxide offers
several advantages over chlorine, most importantly because it does not gen-
erate harmful chlorinated byproducts (18). The main disadvantage associated
with the use of ClO2 is its instability at high pressure. This prevents the trans-
portation of ClO2 in gaseous phase and technological applications are
designed to accommodate in situ preparation of this species. Several methods
exist for generating ClO2. Industrially, the reduction of chlorate ion with
methanol or other agents in sulfuric acid is used (19). For the preparation
of pure chlorine dioxide, metalloporphyrin catalyzed oxidation of chlorite
ion was reported by Collman et al. (20). The formation of chlorine dioxide
was also found during the decomposition and oxidation reactions of ClO2 ,
as will be discussed later in this section.
Chlorine dioxide is stable under acidic conditions but undergoes decom-
position in alkaline solution. As demonstrated in Fig. 1, the experimental
3
0.3
First order
104 [CIO2] (M)
0.2
Second order
Mixed order
2 0.1
104 [CIO2] (M)
0.0
0 2000 4000
Time (s)
Second order
First order
Mixed order
0
0 1000 2000 3000 4000 5000
Time (s)
Fig. 1 Regression analysis of kinetic traces for the decomposition of ClO2 by testing the
validities of first-, second-, and mixed-order rate expressions. [ClO2] ¼ 0.37 mM;
[NaOH] ¼ 0.40 M; T ¼ 25.0°C; I ¼ 1.0 M. Reprinted with permission from Odeh, I. N.;
Francisco, J. S.; Margerum, D. W. Inorg. Chem. 2002, 41, 6500–6506. Copyright 2002,
American Chemical Society.
kinetic traces can be fitted with the mixed-order rate expression for ClO2
given in Eq. (2) (21):
½ClO2
¼ ka, obs ½ClO2 + kb, obs ½ClO2 2 (2)
dt
On the basis of the concentration dependencies of the first- and second-
order rate constants, and the final stoichiometry a detailed kinetic model was
proposed (Scheme 2).
The model postulates three parallel pathways, two of which (a and c) lead
to the formation of ClO2 and ClO3 , respectively, in a 1:1 ratio. The third
pathway (b) provides reasonable explanation for the deviation from this stoi-
chiometry by hypothesizing the formation of O2.
The catalytic decomposition and oxidation of ClO2 by hypochlorite
ion exhibit composite pH-dependent stoichiometric features (21a). The
corresponding reactions are given as follows:
2ClO2 + 2OH ¼ ClO2 + ClO3 + H2 O (3)

ClO2 + OCl ¼ ClO3 + Cl (4)
At pH > 9.0, reaction (3) is dominant. Spontaneous decomposition of
ClO2 was negligible in the pH range studied (21) and the catalytic effect
of OCl was confirmed unequivocally. The oxidative path, reaction (4),
is significant in the neutral—slightly alkaline pH range. The unique feature
a. ClO2 + OH− (HOCl(O)O)− kS1, k-S1 (S1)

(HOCl(O)O)− + ClO2 HOCl(O)O + ClO2 −
kS2, fast (S2)
HOCl(O)O + OH − ClO3− + H2O kS3, fast (S3)
b. ClO2 + OH− (OClOOH)− kS4, k-S4 (S4)

(OClOOH)− + ClO2 OClOOH + ClO2− kS5, fast (S5)
OClOOH + OH− HOClO + HOO− kS6, fast (S6)
HOO−+ 2 ClO2 + OH− 2 ClO2− + O2 + H2O kS7, fast (S7)
HOClO + OH− ClO2− + H2O kS8, fast (S8)
c. ClO2 + ClO2 Cl2O4 kS9, k-S9 (S9)

Cl2O4 + OH− HOCl(O)O + ClO2− kS10 (S10)
HOCl(O)O + OH − ClO3− + H2O kS11, fast (S11)

Scheme 2 Kinetic model for alkaline decomposition of ClO2 (21).
of this system is that regardless of the actual stoichiometry, the same rate law
applies to the entire pH range studied, i.e., the reaction is strictly first order
both in ClO2 and OCl (Eq. 5):
d½ClO2
¼ k½ClO2 ½OCl (5)
dt
These observations are consistent with a kinetic model in which the rate-
determining step is followed by two major competing reaction paths leading
to different products as a function of pH (21a). Further details of this reaction
were explored in a subsequent study (22).
Thiols play an essential role in protecting cells from oxidative damage by
reactive organic species and the removal of these reducing agents from bio-
logical systems has detrimental effects. For example, during bacteria disinfec-
tion, ClO2 penetrates into the membrane, it oxidizes thiols to disulfides and
causes the death of the bacterium. According to literature data, ClO2 is also
able to oxidize amino acids and the fastest reactions were found with cyste-
ine, tyrosine, and tryptophan (23,24). Within this group, cysteine is the most
reactive amino acid due to the highly nucleophilic thiol group (25). During
the oxidation of cysteine with ClO2 and ClO2 in acidic medium, cysteic
acid (CSO3H) was reported as the main product by Darkwa et al. (26) The
kinetic traces were simulated on the basis of a 28-step kinetic model. Con-
sidering the uncertainties associated with this kind of evaluation of the
kinetic data, the proposed rate constants should be termed as ambiguous
at best.
The oxidation reactions of ClO2 involve the formation of lower oxi-
dation state oxychlorine species which, in turn, may quickly react with
the substrates. A thorough study on the oxidation of cysteine (HCS) and
glutathione by Ison et al. (Scheme 3) serves as an excellent example for this
feature (27).
According to pH-dependent experiments, the deprotonated form of the
substrate is reactive in this system. The rate-determining step is a single elec-
tron transfer from CS to chlorine dioxide which forms chlorite ion and a
cysteinyl radical. The overall second-order reaction is first order in each
reactant and the same rate expression applies to the entire pH range,
although the stoichiometry is pH dependent. It implies again that after
the formation of a common intermediate, which is the cysteinyl–ClO2
adduct, the products are formed via two concurrent pH-dependent paths.
The low pH pathway proceeds via the formation of HOCl and produces
cysteic acid, while cysteine is the main product of the high pH pathway.
+H CO2– Keq +H CO2–

3N 3N
+ H+
pKa = 8.18
SH S–
ClO2
Cysteine (CSH)
k1 = 1.03 × 108 M–1 s–1
ClO2–
+H CO2–
3N
S•
Fast + ClO2
+H CO2–
3N
O O
S Cl
Cysteinyl-ClO2 adduct
Low pH pathway H 2O CS– High pH pathway
HOCl ClO2–
+H CO2–
3N
+H HOCl Cl– +H3N CO2–

3N CO2– S
O
OH Fast S S
S
O OH
O +H
3N CO2–
Cysteine sulfinic acid Cysteic acid (CSO3H)
(CSO2H) Cystine
(CSSC)
Scheme 3 Kinetic model for the oxidation cysteine by ClO2. Reprinted with permission
from Ison, A.; Odeh, I. N.; Margerum, D. W. Inorg. Chem. 2006, 45, 8768–8775. Copyright
2006, American Chemical Society.
It was shown that chlorite ion oxidizes HCS about six orders of magnitude
slower than ClO2. This difference in the reactivities was interpreted by
assuming that different mechanisms are operative in these reactions, i.e.,
the oxidation with ClO2 proceeds via oxygen transfer as opposed to single
electron transfer with ClO2. The oxidations of glutathione (GSH) and HCS
show close resemblance leading to the conclusion that the thiol group is the
main reactive site of GSH as well (27).
The oxidation of tryptophan (Trp) by ClO2 was investigated by Stewart
and coworkers using HPLC, UV spectrophotometric, stopped-flow, and
ESI–MS methods (28). It was confirmed that one Trp consumes two ClO2
while HOCl and ClO2 are formed in a 1:1 ratio. In spite of the relatively
simple stoichiometry regarding the reactants and inorganic products, a broad
array of organic species is formed.
Very diverse kinetic features are associated with the chemistry of chlorite
ion. This species is relatively stable under alkaline conditions but it decom-
poses erratically in neutral solution probably because of the presence of
minute amounts of impurities which may act as catalysts. Under acidic con-
ditions, ClO2 decomposes spontaneously or in catalytic reactions and pro-
duces ClO3 , ClO2, and Cl (16). The stoichiometry of this reaction is
strongly dependent on the concentration ratios of the reactants, the concen-
tration of the catalyst and the pH. It was demonstrated earlier that the com-
bination of classical, stopped-flow, and quenched stopped-flow experiments
can provide invaluable information on the iron(III) catalyzed decomposition
of chlorous acid (29,30) and the same kind of experiments may prove to be
useful in other systems, too.
Chlorite ion is a key reactant in many reactive systems exhibiting unique,
nonlinear dynamic features. Perhaps chlorite-driven reactions form the big-
gest group of the family of oscillation reactions. The first such system was the
ClO2 I reaction which has the most well-established mechanism (31).
Nowadays, such reactions are used for designing propagation fronts (32–34).
In general, very complex multistep kinetic models are proposed for the
interpretation of these phenomena but sometimes important details are
not clarified. Thus, exploring the details of the subsystems is of utmost
importance.
It is well established that HOCl is an important intermediate in the reac-
tions of chlorite ion, and the formation of ClO2 is due to the
HOCl ClO2 reaction in many cases. The most detailed kinetic study
on this subsystem was reported by Kormanyos et al. (35) In this case, the for-
mation of ClO2 was monitored by systematically changing the reactant con-
centrations and pH. A detailed kinetic model (Scheme 4) was developed by
simultaneously fitting all kinetic traces.
This model provides an excellent interpretation of the experimental
results (Fig. 2).
The reaction of chlorite ion with reducing sulfur compounds quite often
shows complexities rarely found in other systems. In the chlorite–thiosulfate
ion reaction, later called “crazy clock” reaction, Orbán and coworkers
observed periodic and aperiodic oscillations (36,37) and Maselko and Epstein
reported chaos in a continuously stirred tank reactor (38). In batch reactors,
2HOCl Cl2O + H2O Fast, Keq = 0.0115 M–1 (S1)

Cl2 + H2O HOCl + Cl− + H+ Keq = 6.1 × 10–4 M2 (S2)
HOCl + HClO2 Cl2O2 + H2O (S3)
Cl2 + HClO2 Cl2O2 + Cl− + H+ (S4)
Cl2O2 + ClO2− 2·ClO2 + Cl− (S5)
Cl2O2 + HOCl Cl2 + ClO3− + H+ (S6)
Cl2O + ClO2− Cl2 + ClO3− (S7)
ClO2− + HOCl ClO3− + Cl− + H+ (S8)
ClO2 + OCl− ·Cl2O3− (S9)
ClO2 + ·Cl2O3−+ H2O Cl− + 2H+ + 2ClO3− Fast (S10)

Scheme 4 Kinetic model for the interpretation of the chlorite ion–hypochlorous acid
reaction (35).
this process is very sensitive to the stirring rate due to fluctuations (39). The
kinetics and mechanism of the chlorite–thiosulfate ion reaction are quite
complicated because the intermediates and the products can react with
the reactants and also with each other. In other words, the main reaction
includes several subsystems, such as the redox reactions between chlorite–
hypochlorous acid, thiosulfate ion–chlorine dioxide, tetrathionate ion–
hypochlorous acid, tetrathionate ion–chlorine dioxide and the decomposition
of chlorous acid. In the following paragraphs, we give a brief overview of
these processes.
The above reactions have been studied by Horváth et al. in great detail. In
the ClO2 S2 O3 2 system, they observed sigmoidal kinetic traces in the
excess of thiosulfate ion which indicates that autocatalysis is operative in this
reaction. Tetrathionate and chlorite ions are the main products of this pro-
cess but Cl and SO4 2 are also formed. It was proposed that the initial step
is the irreversible formation of the S2 O3 ClO2 •2 radical and the reaction
proceeds via the formation of the S4 O6 •3 radical. A 10-step mechanism
was postulated and the corresponding rate constants were estimated by
0.5
0.4
Absorbance at 360 nm
0.5
0.3
0.4
0.3
0.2
0.2
0.1
0.1 0
0 10 20 30 40
0
0 50 100 150 200 250 300
Time (s)
Fig. 2 Formation of chlorine dioxide in HOCl excess as a function of [Cl]0.
[HOCl]0 ¼ 3.00 mM, ½ClO2 0 ¼ 0:568 mM, pH 5.55; [Cl]0 (mM) ¼ 0 (●), 1.0 (□), 2.0
(▲), 4.0 (♦), 8.0 (■). The inset shows the enlarged first section of the traces. Reprinted
with permission from Kormanyos, B.; Nagypal, I.; Peintler, G.; Horvath, A. K. Inorg. Chem.
2008, 47, 7914–7920. Copyright 2008, American Chemical Society.
fitting more than 130 kinetic traces simultaneously (40). Excellent agree-
ment between the measured and calculated data seemed to validate the
model (Fig. 3).
In a subsequent study, Pan and Stanbury confirmed the experimental
observations reported by Horváth and Nagypál and made an attempt to pro-
vide an alternative interpretation for the autocatalytic nature of the reaction.
These authors confirmed that L-methionine suppresses the autocatalysis.
A detailed kinetic model was proposed which postulates the formation of
HOCl and SO3 2 as chain carriers in this system. The formation of ClO
was also proposed. The formation of a Cl(II) intermediate was postulated
in many reactions of oxychlorine species earlier. In aqueous solution, such
an intermediate could not be detected directly before, however, a great
number of mechanistic studies on the redox reactions of oxychlorine species
corroborate its existence. The kinetic effect of L-methionine was interpreted
by considering that this compound is an extremely efficient and selective
scavenger of HOCl and as such terminates the chain reaction (41).
1.2
0.9
Absorbance
0.6
0.3
0.0
0.00 0.05 0.10 0.15
Time (s)
Fig. 3 Experimental and fitted kinetic traces on the basis of the kinetic model proposed

by Horváth and Nagypál (40). S2 O3 2 0 ¼ 1:55 103 M, pH 9.27; [ClO2]0
104 M ¼ 5.34 (●), 7.39 (□), 9.35 (▲), 11.0 (◊). Reprinted with permission from Horváth, A. K.;
Nagypál, I. J. Phys. Chem. A 1998, 102, 7267–7272. Copyright 1998, American Chemical
Society.
It is important to note that the two papers discussed earlier report the
same kinetic features of the ClO2 =S2 O3 2 system, yet, the proposed models
are markedly different. In a broader context, this is an inherent problem
associated with mechanistic studies on complex redox reactions. When a
large set of kinetic traces is collected by systematically changing the concen-
trations and concentration ratios of the reactants under a wide range of
experimentally accessible conditions, a multistep kinetic model may fit
the data exceptionally well. However, new experimental observations
may provide insight into further details of the mechanism and the combina-
tion of the old and new results may require substantial refinement or even
reconstruction of the kinetic model.
The chlorite–tetrathionate reaction exhibits complex kinetic patterns
leading to exotic nonlinear dynamic phenomena such as oscillation (37)
and has outstanding importance in some of the reactive systems featuring
spatiotemporal pattern formation and propagating fronts (42–45). Alkaline
decomposition of the tetrathionate ion leads to the formation of thiosulfate,
sulfite, and trithionate, thus the reactions of these ions with chlorite ion also
contribute to the overall process in the ClO2 =S4 O6 2 system. Kinetic
results on the ClO2 =S2 O3 2 reaction were reported by Nagypal and
Epstein (39), while Huff Hartz et al. studied the ClO2 =S2 O3 2 reac-
tion (46). In contrast to tetrathionate, trithionate is stable in alkaline aqueous
solution (47) and does not react with ClO2 . However, the formation of a
considerable amount of ClO2 was reported in the ClO2 =S3 O6 2 system
under slightly acidic conditions. It was also confirmed that ClO2 slowly dis-
appeared in excess trithionate (48). A systematic study on the
ClO2 =S3 O6 2 reaction revealed that the actual stoichiometry depends
on the concentration of the reactants and pH and always can be given as
the linear combination of the following two limiting stoichiometries:
5S3 O6 2 + 8ClO2 + 14H2 O ! 15SO4 2 + 8Cl + 28H + (6)

S 3 O6 2
+ 4ClO2 + 4H2 O ! 3SO4 2
+ 2Cl + 2ClO3 + 8H +
(7)
Chloride ion acts as an autocatalyst in this system, and overshoot–
undershoot kinetics were observed when it was added to the reaction mix-
ture in the presence of excess ClO2. A 13-step kinetic model was proposed
assuming that the initial step is an adduct formation between the reactants.
The kinetic model provided excellent fit of the experimental kinetic traces
(48).
In the ClO2 =S4 O6 2 system HOCl is an autocatalyst, the concentration
of which is regulated by sulfite ion (49). The kinetics of the sulfite ion–chlorine
dioxide reaction was independently studied in acidic medium (50) to explore
the significance of this reaction path in the parent chlorite–thiosulfate system.
The final stoichiometry of the reaction shows distinct concentration depen-
dency. When ClO2 is used in high excess, the limiting stoichiometry shown in
Eq. (8) prevails:
2SO3 2 + 2ClO2 + H2 O ¼ 2SO4 2 + ClO3 + Cl + 2H + (8)

The stoichiometry is gradually shifted toward the formation of sulfate
and dithionate in 1:1 ratio (Eq. 9) by increasing the [S(IV)]0/[ClO2]0 ratio.
6SO3 2 + 2ClO2 ¼ 4SO4 2 + S2 O6 2 + 2Cl (9)

Initial rate studies confirmed that the reaction order is greater than 1 for
both reactants and depends on the actual concentrations. Finally, a 9-step
kinetic model was proposed assuming that the initial step is the formation
of an adduct between the reactants, SO3 ClO2 2 , which is followed by its
rate-determining decomposition.
Thorough description of the subsystems offered a possibility to develop a
comprehensive kinetic model for the parent chlorite–thiosulfate reaction.
Accordingly, a 38-step kinetic model was proposed for this system recently.
It was shown that fast equilibria (e.g., acid–base reactions) and 12 fast reac-
tions with 30 fitted parameters are suitable to describe 367 experimental
kinetic traces with high precision (51).
The common feature of the previously discussed reactions of ClO2 and
ClO2 is that only the concentration change of chlorine dioxide could be
followed quantitatively. This species has a characteristic, relatively strong
absorbance band in the UV–vis region which overlaps other spectral effects
and practically excludes the detection of absorbance contributions from
other components. This may introduce some ambiguity in the evaluation
of the kinetics when the formation of ClO2 is followed because kinetic pro-
files of a product may not contain sufficient information on the initial part of
a complex reaction. However, this problem is greatly eliminated if the reac-
tions are studied in a broad range of initial concentrations and concentration
ratios of the reactants, as it was carried out in the cited studies. In fact, the
results unequivocally indicated the formation of various transient species
which cannot be detected directly but are required for coherent interpreta-
tion of the experimental observations.
In manganese porphyrin catalyzed alkane oxidations with ClO2 , the
existence of two pathways was confirmed which involve the formation of
O2 and a high-valent manganese(V)–oxo intermediate (52). Recently, sev-
eral other reports have indicated that various transition metal complexes may
have interesting catalytic effects on the reactions of chlorite ions. Thus, cat-
alytic formation of ClO2 was observed in the presence of manganese por-
phyrin (53) and ruthenium bisphenanthroline complexes (54).
Catalytic decomposition of chlorite ion in the presence of water-soluble
iron (55,56) and manganese porphyrins (57,58) was studied under close to
neutral pH conditions in detail. In the case of the iron complexes, the dis-
proportionation of ClO2 produces Cl and ClO3 in a 1:2 ratio.
A moderate yield of O2 was also observed in the presence of the fluorinated
[FeIII(TF4TMAP)]+ complex. DFT calculations are consistent with the for-
mation of a FeIV oxo compound and chlorine monoxide in these systems (59).
In contrast, manganese complexes assist the conversion of ClO2 into ClO2.
Kinetic models postulate the formation of higher oxidation state MnIV and
MnV oxo intermediates in these systems. The catalytic cycle is initiated by oxy-
gen transfer between ClO2 and the catalyst and subsequent reactions include
electron transfer and proton-coupled electron transfer steps (Scheme 5). The
proposed model was validated by simulating the kinetic traces.
Hypochlorous acid is the simplest oxo-acid of chlorine which is involved
in fast equilibria with OCl and also with Cl2 under acidic conditions in the
ClO3– ClO2– + H2O

k5
ClO2– ClO–
k1
OH2 + O +
k2
ClO– Cl–
MnIII MnV
ClO3–
k6 ClO2–
ClO2 k–3
k–4 ClO2
k3
k4
O
–
ClO2
ClO2 IV
Mn
+ 2H+
Scheme 5 Kinetic model for the decomposition of ClO2 in the presence of a water-
soluble manganese porphyrine. Reprinted with permission from Hicks, S. D.; Xiong, S.;
Bougher, C. J.; Medvedev, G. A.; Caruthers, J.; Abu-Omar, M. M. J. Porphyrins Phthalocya-
nines 2015, 19, 492–499. Copyright 2015, World Scientific Publishing Company.
presence of Cl. HOCl is a much stronger oxidizing agent than its conjugate
base form, therefore, the oxidation reactions of the HOCl/ClO couple
show marked pH dependencies. These reactions typically follow straightfor-
ward kinetics, and complicated kinetic patterns are rarely observed. As an
exception, the oxidation of pyruvic acid to acetic acid should be mentioned
(60). In this system, a simple 1:1 stoichiometry was confirmed for the entire
pH region, moreover, the reaction is strictly first order in both reactants
above pH 2.5. The entropy of activation is consistent with an oxygen atom
transfer mechanism which is also supported by DFT calculations. Unexpect-
edly, stopped-flow kinetic traces under pH 2.5 show two distinct first-order
phases. It was confirmed that this complexity is caused by the hydration reac-
tion of the substrate which is a fast preequilibrium to the oxidation step.
In recent years, biological relevance of HOCl has generated vast interest
in its redox chemistry. In water treatment technologies, the formation of
chlorinated products is of primary concern (18). In biological systems,
HOCl is formed by nucleophiles in the myeloperoxidase/H2O2/chloride
ion system (61,62) as part of the defense mechanism against pathogens. With
amino compounds, it produces N-chloro-amines (63–67). The formation of
these species is very fast, and the corresponding second-order rate constants
are within the range 104–108 M1 s1. Earlier studies on the formation of
N-chloramines from ammonia and amino acids revealed that the main reac-
tion path occurs between the deprotonated amine and HOCl as shown in
Eqs. (10)–(12) (62,68–73).
R NH2 + HOCl ¼ R NHCl + H2 O (10)
R NHCl + HOCl ¼ R NCl2 + H2 O (11)
NHCl2 + 2HOCl ¼ NCl3 + H2 O (12)
2NCl3 + 3H2 O ¼ N2 + 3HOCl + 3Cl + 3H + (13)
In water treatment processes, the chlorinating agent is present in excess,
thus, di- or trichlorinated amines are produced (Eqs. 11 and 12) (18). In
addition breakpoint chlorination occurs, i.e., ammonia is completely oxi-
dized leading to the formation of gaseous nitrogen (Eq. 13). In living sys-
tems, the amino acids and peptides are in excess over HOCl and only
N-monochloramines are formed (Eq. 10).
In all of these systems, HOCl is a precursor of the N-chloro species
which are able to penetrate into the cells and cause oxidative stress leading
to cell death. N-chloroamines are not stable in aqueous solution, according
to earlier studies they decompose to ammonia, carbon dioxide, chloride ion,
and carboxyl products (66,74). However, a survey of the literature reveals
contradictions in previous results and indicates the complexity of these reac-
tions (67,75–78). Recently, we have reinvestigated the kinetics of the
decomposition of N-chloroglycine (MCG) under alkaline conditions and
demonstrated that a simple kinetic approach may be misleading in this sys-
tem (79). Our spectrophotometric results were consistent with fast forma-
tion of this molecule (λmax ¼ 255 nm) followed by considerably slower
decomposition. Kinetic measurements were performed by mixing OCl
and glycine which was always used in excess. MCG formed immediately
upon mixing the reactants, and its decay could be monitored conveniently
by UV–vis spectrophotometry. While single exponential kinetic traces were
observed above 250 nm, the experiments indicated the formation of an
intermediate at lower wavelengths (Fig. 4).
The faster of the two first-order steps could be assigned to the decom-
position of MCG. The corresponding rate constant is linearly dependent on
the hydroxide ion concentration. The second step is pH independent and
corresponds to the transformation of an intermediate. The rate constants
did not change upon increasing the concentration of excess glycine, but
the final absorbance increased, indicating that glycine is involved in a side
reaction after the rate-determining step. The formation of ammonia was also
confirmed using the Nessler-test (80).
0.9
A 228 nm 0.6
0.3
0.0
0 2500 5000
t (s)
Fig. 4 Kinetic traces for the decomposition of N-chloroglycine (MCG) under alkaline
conditions and at excess glycine concentrations at 228 nm. [MCG]0 ¼ 3.00 103 M,
[OH] ¼ 0.054 M, and [Gly]0 ¼ 1.50 103 (◊), 3.00 103 (□), 1.20 102 ( ),
2.70 102 (△); I ¼ 1.0 M (NaClO4), T ¼ 25.0°C. Reprinted with permission from Szabó, M.;
Baranyai, Z.; Somsák, L.; Fábián, I. Chem. Res. Toxicol. 2015, 28, 1282–1291. Copyright
P1b GLY
P1a t (s)
P4 P2b P3b
P3a P2a
¥
7370
1804
453
I1b I1a
152
MCG
0
9.0 8.5 8.0 4.2 4.0 3.8 3.6 3.4 3.2
(ppm) (ppm)
Fig. 5 1H NMR spectra recorded at various reaction times during the decomposition of
MCG. [Gly]0 ¼ 1.00 102 M, [MCG]0 ¼ 1.00 102 M, [OH] ¼ 0.054 M, T ¼ 25.0°C.
Reprinted with permission from Szabó, M.; Baranyai, Z.; Somsák, L.; Fábián, I. Chem.
Res. Toxicol. 2015, 28, 1282–1291. Copyright 2015, American Chemical Society.
In order to explore the mechanistic details time-resolved 1H, 13C NMR

measurements, and ESI–MS experiments were made. 1H NMR spectros-
copy could be used to monitor the decay of MCG, the formation of the
products (P1, P2, P3, and P4) and the formation and decay of an intermediate
(I1) as shown in Fig. 5.
Fig. 6 shows the intensity of the 1H NMR peaks as function of time. The
concentration of the excess glycine does not change during the reaction,
MCG disappears in less than 3000 s.
A
3.2
Intensity × 10−6 (a.u.) GLY
1.6
MCG
0.0
0 2500 5000
t (s)
B
4.8
I1
Intensity × 10−5 (a.u.)
3.2
1.6
I2
0.0
0 2500 5000
t (s)
C
1.5
P1a
Intensity × 10−6 (a.u.)
1.0
P1b
0.5
0.0
0 2500 5000
t (s)
Fig. 6 The intensities of the 1H NMR peaks as a function of time during the decompo-
sition of MCG. [Gly]0 ¼ 1.00 102 M, [MCG]0 ¼ 1.00 102 M, [OH] ¼ 0.054 M,
T ¼ 25.0°C. Reprinted with permission from Szabó, M.; Baranyai, Z.; Somsák, L.; Fábián, I.
Chem. Res. Toxicol. 2015, 28, 1282–1291. Copyright 2015, American Chemical Society.
Peaks P1 and P2 are assigned to the trans and cis forms of the main prod-
uct, N-formylglycine. In separate experiments, we also demonstrated that
glyoxalic acid, a potential byproduct in this system, immediately reacts
with glycine and forms a Shiff base (P3). Peak P4 belongs to the formate
ion which appears 2 h after starting the reaction. On the basis of 1H
NMR experiments with the pure compound, the intermediate (I1) of
the reaction was identified as N-oxalylglycine. The formation of this spe-
cies is responsible for the noted increase of final absorbance of the kinetic
traces as a function of glycine excess.
Earlier, it was assumed that the decomposition of N-chloroamino
acids proceeds by the Grob-fragmentation which is a concerted process
(62,68,75,81–83). In the case of N-chloroglycine, formaldehyde was
reported as one of the main products. Our results confirm a different mech-
anism outlined in Scheme 6. The reaction is initiated with the formation of
a carbanion which is consistent with the noted OH dependence of the first
exponential step of the kinetic traces. Once the imine is formed, the reac-
tion may proceed via two parallel pathways. Pathway II is included only to
Scheme 6 The outline of the mechanism of the decomposition of MCG. Reprinted with
permission from Szabó, M.; Baranyai, Z.; Somsák, L.; Fábián, I. Chem. Res. Toxicol. 2015, 28,
1282–1291. Copyright 2015, American Chemical Society.
demonstrate that the main product could form via formaldehyde as inter-
mediate, but direct experimental evidence does not confirm its existence.
In contrast, pathway III must be operative in order to account for all the
observations in this system. The second kinetic step is assigned to slow
transformation of N-oxalylglycine to MCG. Perhaps the most exciting
aspect of these studies is that direct experimental evidence confirms the
existence of these two species, which have been shown to possess biological
activities.
4. KINETICS OF THE OXIDATION REACTIONS OF PEROXO

COMPOUNDS
Peroxo compounds are widely used for bleaching textiles and paper
and frequently utilized as oxidizing agents in organic chemistry, in the phar-
maceutical industry and in AOPs.
By far the most extensively used and most studied peroxo compound is
hydrogen peroxide (H2O2). It is utilized as a powerful oxidant and as the
main starting material in the production of other peroxo compounds.
Reviews have recently been published discussing the role of H2O2 in AOPs
as the main component of the Fenton reaction (84,85). Some of these studies
are devoted to the kinetic and mechanistic aspects of the oxidation processes.
It is widely accepted that one of the most important active intermediates in
the Fenton reaction is the hydroxyl radical (OH•). Lately, the importance of
OH• reactions has been recognized in biological systems, and in oxidative
technologies including waste treatment.
One of the main problems of using hydrogen peroxide as an oxidant is
its low thermodynamic stability. It decomposes relatively easily to form
water and oxygen, especially under alkaline conditions. The decomposi-
tion is catalyzed by various species, including transition metals (for exam-
ple, Fe2+ and Ti3+) and their compounds (e.g., manganese dioxide),
nonmetallic catalysts such as iodide ion, or by the catalase enzyme. UV
light can also initiate the decomposition of H2O2 (86). Owing to its unsta-
ble character, a relatively large excess of H2O2 may be required in some
oxidation reactions because part of the oxidizing agent is consumed in
its disproportionation reaction.
Other industrially important inorganic peroxo compounds are the salts of
peroxodisulfuric acid. The dibasic acid also known as Marshall’s acid,
H2S2O8, is unstable in aqueous solutions and eventually undergoes hydro-

lysis to give sulfuric acid and hydrogen peroxide. However, the
peroxodisulfate salts (PDS) of ammonium, sodium, and potassium ions
are stable. PDS is one of the strongest oxidizing agents known in aqueous

solution with a standard potential of E° S2 O8 2 =2HSO4 ¼ 2:12 V
(87). Despite this strong thermodynamic driving force for oxidation, it is
kinetically stable in solution and the reactions of S2 O8 2 generally are slow
at room temperature because the rate limiting homolysis of the peroxo bond
(Eq. 14) has a large activation energy (88).
S2 O8 2 ! 2SO4 • (14)
Both H2O2 and S2 O8 2 are frequently used in AOPs for generating free
radicals. However, these oxidants are rarely efficient in degrading organic
pollutants when employed on their own. Peroxomonosulfate ion (PMS)
is often used as an environmentally friendly replacement of the aforemen-
tioned peroxo species. Its main advantage over hydrogen peroxide is its eas-
ier handling and higher stability. Although PMS has a lower standard
potential, E°ðHSO5 =HSO4 Þ ¼ 1:82 V (89), than PDS, that is still suffi-
ciently high to oxidize various substrates. Many reactions of PMS proceed
faster than those of S2 O8 2 .
Peroxomonosulfate ion, primarily used in its monoprotonated form
(HSO5 ), is the anion of Caro’s acid (peroxomonosulfuric acid, H2SO5).
Potassium peroxomonosulfate is commercially available in the form of a sta-
ble triple salt (2KHSO5KHSO4K2SO4) which is called Oxone. In recent
decades, the use of PMS has increased rapidly in both industrial applications
and fundamental research. The main reasons behind its popularity are several
favorable features such as stability, simple handling, nontoxic nature, good
solubility in water, versatility of the reagent, and low cost. Its oxidation
byproducts (typically sulfate ion) do not pose a threat to aquatic life and con-
sidered environment friendly. These properties make PMS attractive for
large-scale applications.
Apart from its wide industrial and consumer utilizations (such as decol-
orizing agent in denture cleansers, microetchant in electronics, shock oxi-
dizer for swimming pools, repulping agent in papermaking, or oxidizer in
wool treatment), the use of PMS finds ever growing applications in organic
(90,91) and environmental chemistry (92). Interestingly, PMS was proven
to be readily applicable in both fields despite the fact that their objectives are
seemingly different. In organic chemistry, PMS is used for the oxidation

of various functional groups and in most cases (Scheme 7) it acts as a single
oxygen atom donor and the nonsymmetrical O–O bond is heterolytically
cleaved during the oxidation. In environmental chemistry, however,
PMS is applied in AOPs, where the major goal is the degradation of refrac-
tory organic compounds to harmless materials. For this purpose, there are
promising technologies producing radicals (typically hydroxyl and sulfate
radicals). Owing to the unsymmetrical peroxo bond, after activation and
homolytical scission, both hydroxyl and sulfate ion radical (SO4 • ) can be
generated. Therefore, PMS combines the features of H2O2 and S2 O8 2 ,
which are the common sources of OH• and SO4 • radicals, respectively.
Although PMS is thermodynamically a strong oxidant, similarly to H2O2
and S2 O8 2 , its direct reaction with most of the organic pollutants is too
slow without activation. PMS can be activated by transition metals, UV
light, heat, ultrasound, or electron conduction (92).
Although the neutral or mildly acidic solution of the pure reagent in dis-
tilled water is relatively stable, whenever the kinetics of the reactions of PMS
are studied, its spontaneous or metal-catalyzed decomposition should be
considered as a potential interfering side reaction. Both inorganic peroxo
acids such as peroxomonosulfuric acid (93) or hydrogen peroxide (94)
and organic peroxycarboxylic acids (e.g., aromatic peroxyacids (95),
O
RCHO
R OR
RCH2JOH
R3CJOH R2NOH
RCOR
R3CJH
O R 3N
Phenol Halogenation
PMS RJX
R R 3P
O
R
NO2 RSR
O R3PKO
R R R
R O
RSO2R
R R
Scheme 7 Examples of the organic oxidation reactions with PMS.
peroxyacetic and monoperoxyphthalic acid (96)) were found to decompose

to give molecular oxygen (O2) and sulfuric acid, water, or the corresponding
carboxylic acids.
The rate of the spontaneous decomposition of PMS (Eq. 15) is second
order with respect to the total peroxide concentration and does not proceed
by a free radical mechanism. The pH dependence shows that PMS is stable
below pH 6 and above pH 12. The rate of decomposition has a maximum at
the pH equal to the pKa of HSO5 , which indicates the presence of the
mononegative (HSO5 ) and the dinegative (SO5 2 ) anions in the activated
complex and the rate law can be given as follows (93):
HSO5 + SO5 2 ! 2SO4 2 + O2 + H + (15)

2
1 d½PMS k15 ½H ½PMS
+
v¼ ¼ k015 ½PMS2 ¼
2 dt ½H + 2
Ka 1 +
Ka
Analogous behavior is shown by other organic peroxy acids and the reac-
tion can generally be regarded as a nucleophilic attack of the anion upon the
protonated form.
It should also be noted that occasionally misleading citations can be
found in the literature in which the citing authors erroneously call redox
reactions of PMS with certain metal ions, metal-catalyzed decomposition
(92,97,98). In some of the original (cited) papers, the redox reactions of
PMS with Fe(II) (99) and Cu(I) (100) are referred to as peroxide decompo-
sition but the term catalysis is not used at all and no mechanistic implication
is made for any catalytic cycle. By decomposition, the authors most probably
mean that the reactions involve the cleavage of the peroxo bond of PMS. For
example, detailed kinetic and stoichiometric studies revealed that although
Fe(II) does react with PMS which involves the formation of a sulfate radical
intermediate, the metal ion does not catalyze the decomposition of the oxi-
dant (101). The same considerations apply to V(IV) and Ce(III) (101).
A thorough investigation of the Fe(III)–PMS system showed no catalytic
effect of Fe(III), either (98).
As discussed earlier, PMS may react via two distinct pathways as an oxi-
dizing agent and one can be a one-electron oxidation (via electron transfer),
and the other a two-electron (via oxygen atom transfer) oxidation. Gener-
ally, the use of activation parameters as mechanistic indicators is not without
problems. However, we have shown that the activation entropy can be a
helpful tool in distinguishing between the two pathways (101): the
one-electron oxidation reactions with PMS usually have relatively small

negative activation entropies (ΔS‡ > 50 J mol1 K1), whereas two-
electron oxidations have ΔS‡ at around 100 J mol1 K1. This seems to
be reasonable, because an oxygen atom transfer process is expected to
involve much more geometric reorganization than a one-electron rate-
determining step.
More frequently, PMS acts as a two-electron oxidant that involves the
heterolytic cleavage of the peroxo bond. This type is common and usually
preferable in organic syntheses and is typical in oxygen atom transfer
reactions, e.g., the oxidations of halide ions (101), thiocyanate ion (102),
dimethyl sulfide (103), or aromatic amines (104).
The oxidations of halide ions seem to be straightforward oxygen atom
transfer reactions, which follow simple second-order kinetics (first order
with respect to both reductant and oxidant) and the rate equation can be
written as follows (Eq. 16):
HSO5 + X ! HSO4 + OX v ¼ kapp ½HSO5 ½X (16)
where X can denote Cl, Br, or I.

However, the detailed kinetic and stoichiometric study of these reactions
revealed that the observations cannot be explained without considering
coupled proton transfer reaction (Eq. 17), the comproportionation of hypo-
halous acids and halide ions to halogens (Eq. 18) and the trihalide ion for-
mation (Eq. 19):
H + + OX Ð HOX (17)

HOX + H + X Ð X2 + H2 O
+
(18)
X2 + X Ð X3 (19)
In the case of Cl, elemental chlorine (Cl2) is the dominant product at
halide excess and HOCl in the presence of excess of PMS. When Br is used
in excess over PMS, the main products are Br2, HOBr, and OBr under
acidic, neutral, and basic conditions, in order. When the oxidant is used
in excess, Br2 is a long-lived intermediate and the final product is BrO3
which is produced in a kinetically separate step. In the case of I, I3 is
the main product at I excess between pH 0 and 10. Further oxidation
of I2 yields IO3 .
The rate of the first oxidation step of all three halides (Eq. 16) shows
characteristic pH dependence and the apparent second-order rate constant
1.0
HSO5−
kapp/kref
0.5
I−
Br−
CI−
SO52−
0.0
2.0 5.0 8.0
pH
Fig. 7 pH dependencies of the normalized apparent second-order rate constants deter-
mined in the reactions between PMS– and various halide ions. Medium: 0.10 M NaClO4
(Br, I), 1.0 M NaClO4 (Cl); T ¼ 25.0°C; kref ¼ 2.06 103 M1 s1 (Cl), 7.0 101
M1 s1 (Br), 1.41 103 M1 s1 (I). The absolute values of the kapp rate constant var-
ied by orders of magnitudes for different halides, therefore the y-axis was normalized by
dividing by kref to facilitate comparison of the three curves. Reprinted with permission
from Lente, G.; Kalmár, J.; Baranyai, Z.; Kun, A.; Kek, I.; Bajusz, D.; Takács, M.; Veres, L.;
Fábián, I. Inorg. Chem. 2009, 48, 1763–1773. Copyright 2009, American Chemical Society.
(kapp) drops sharply in the pH region 7–10 (Fig. 7). This observation can be
interpreted by two parallel oxidation pathways with the acidic and basic
forms of the oxidant which are in fast acid–base equilibrium with each other.
The pH dependence clearly shows that SO5 2 is a considerably less active
oxidant than HSO5 . A similar property was observed during the oxidation
of SCN (102) and 1,10-phenanthroline (104).
In the reaction of PMS with SCN, further oxidation of the primary
product, hypothiocyanite ion (OSCN), also takes place and the rate con-
stant of the second step is about an order of magnitude higher than that of the
first one (Eqs. (20 and 21):
HSO5 + SCN ! SO4 2 + OSCN + H + k20 ¼ 2:0 102 M1 s1

v20 ¼ k20 ½HSO5 ½SCN
(20)
1 1
HSO5 + OSCN ! SO4 + O2 SCN + H 2 +
k21 ¼ 3:3 10 M s3
v21 ¼ k21 ½HSO5 ½OSCN

(21)
The second oxidation step results in the formation of O2SCN, which

decomposes in fast reactions to yield the final mixture of stable anions:
SO4 2 , SO3 2 , CN, and OCN. Interestingly, the above reactions do
not predict a simple first-order build-up for OSCN; apparent exponential
traces were observed for this species and derivation of the rate expression on
the basis of the simultaneous differential equations confirms that [OSCN]t
can successfully be estimated with a simple first-order expression. It was
demonstrated by model calculations that multistep reaction systems may
exhibit simple first-order kinetics even when the appropriate kinetic model
a priori excludes the possibility of such behavior (105,106). In the PMS–
SCN system, the experimental kinetic dataset was fitted to the two-step
kinetic model utilizing the ZiTa software package (107). With large excess
of SCN, the PMS is consumed to form OSCN and the product stoichio-
metric ratio converges to 1:1, because the second step is suppressed.
In other types of oxidation reactions, PMS is proposed to be a one-
electron oxidant forming sulfate ion radical as an intermediate
(Eqs. 22 and 23), e.g., the oxidations of Fe(II) to Fe(III) (99) or V(IV) to
V(V) (108). These processes are analogous to the first steps of the Fenton
type reactions of hydrogen peroxide and are mostly favorable in AOPs.
HSO5 + Mn + ! SO4 • + Mðn + 1Þ + + OH (22)

• ðn + 1Þ +
SO4 n+
+M ! SO4 2
+M (23)
In earlier literature, a simple second-order reaction was reported for the
oxidation of Fe(II) with PMS. (99) However, more complex kinetic patterns
were found in later studies (98,101). Experimental kinetic traces are consis-
tent with a two-step process, the first of which is the rate-determining step of
the oxidation of Fe(II). The second step is an equilibrium process between
the products of the redox reaction, Fe(III) and SO4 2 (Eq. 24):
Fe3 + + SO4 2 Ð FeSO4+ (24)

This latter reaction is unavoidable because Oxone contains SO4 2 and it
is also formed during the redox process. The most interesting feature of this
system is that the faster step can unambiguously be assigned to reaction (24)
although it can only occur after the slower oxidation step.
During the oxidation of aqueous Fe(II), the unexpectedly complex
kinetic phenomena are partly due to the noncomplementary and inner-
sphere features of the redox process (101). In order to obtain a better
insight into the intimate nature of such reactions, further iron(II) complexes
of N-heteroaromatic ligands FeðphenÞ3 2 + , FeðbipyÞ3 2 + , and Fe(tpy)2+,

where phen ¼ 1,10-phenanthroline, bipy ¼ 2,20 -bipyridine, tpy ¼ 2,20 :60 ,200 -
terpyridine were chosen for oxidation with PMS on the basis of a pretext
that these redox reactions proceed via outer-sphere electron transfer. Sur-
prisingly, even more complicated kinetics were encountered and although
the oxidation reactions show similarities, there are striking differences in
the three systems (109–111). It was shown that the acid-assisted dissocia-
tion of the Fe(II) and Fe(III) complexes contribute to the kinetics. UV–vis
spectrophotometry was the primary source of kinetic information, but in
order to gain more profound details, the concentrations of various absorb-
ing species were also calculated by the direct linear algebraic method
reported earlier (112).
In all three reaction systems, the initial stage is very similar. The initial rates
show saturation as a function of the oxidant concentration. The findings were
interpreted with a common kinetic model shown in Scheme 8. The model
includes the acid-assisted dissociation of the Fe(II) complex (Eq. S1); the for-
mation of an intermediate, a 1:1 adduct between the reactants in a fast pre-
equilibrium step (Eq. S2); the rate-determining (rds) intramolecular
electron transfer process producing the Fe(III) complex and sulfate radical
(Eq. S3); and the fast oxidation of another Fe(II) complex by the SO4 •
(Eq. S4). In the FeðphenÞ3 2 + PMS reaction, the adduct was identified by
the ESI–MS method, in the case of the other two complexes, only indirect
(kinetic) evidence implied the presence of such an ion pair.
After the initial phase, these systems exhibit very distinct features. In the
case of the FeðphenÞ3 2 + , we have shown that after an initial loss, the con-
centration of the iron(II) complex temporarily increases on a longer time
scale (Fig. 8). This is quite an unexpected phenomenon considering that
the oxidizing agent is used in large excess.
We also demonstrated that not only the central metal ion (in a one-
electron process) but the ligand is also oxidized (in an oxygen atom transfer
reaction) and two processes occur simultaneously. The oxidation of phen
FeLn2 + (+nH+) → Fe2+ + nHL+ (S1)

–
FeLn 2+
+ HSO5 G[FeL n·HSO5]
+ (S2)
[FeLn·HSO5]+(+H+) → FeLn3+ + SO4•− + H2O rds (S3)
FeLn2+ + SO4•−→ FeLn3+ + SO42– Fast (S4)
Scheme 8 Kinetic model for the initial stage of the oxidation FeLn 2 + with PMS, where
L is an N-heteroaromatic ligand (109).
0.6
A (510 nm)
0.3
0.0
0.0 4.0 8.0
t (h)
Fig. 8 Kinetic trace detected in the reaction between PMS and FeðphenÞ3 2 + . [PMS]0 ¼
h i
11.4 mM; FeðphenÞ3 2 + ¼ 75:3 μM; [H2SO4] ¼ 1.01 102 M; T ¼ 25.0°C; path length ¼
0
1.0 cm.
yields the corresponding N-oxide (phenO) (104). This product inhibits

the oxidation of FeðphenÞ3 2 + by inducing the reduction of FeðphenÞ3 3 +
to the iron(II) complex. The reactions of the Fe(III) complex and its inter-
actions with phenO were studied thoroughly and it was shown that the
disproportionation of this species yields FeðphenÞ3 2 + and FeðphenÞ3 4 +
(Scheme 9, Eq. S1) (110). The presence of the Fe(IV) species in the reac-
tion system was confirmed by the ESI–MS method. The existence of such
an Fe(IV) complex was earlier assumed in the oxidation of FeðphenÞ3 2 + by
Ce(IV) based on kinetic observations but no experimental evidence was
shown (113). This complex decomposes to the iron(II) aqua complex, pro-
tonated ligand (Hphen+), and phenO (Eq. S2). This latter compound also
has a slight autocatalytic effect on the disproportionation reaction, which was
explained by assuming that it is oxidized by the iron(III) complex (Eq. S3) to
give an intermediate, which can in turn oxidize another FeðphenÞ3 3 +
to FeðphenÞ3 4 + (Eq. S4), completing a catalytic cycle for the N-oxide
(Scheme 9).
As shown, the overall mechanism involves several subsystems but most of
them can be studied independently. Some intermediates (Fe2+, phen,
FeðphenÞ3 3 + ) and one of the final products (phenO) influence the concen-
tration change of FeðphenÞ3 2 + and the known processes are summarized in
Scheme 10.
2Fe(phen)33+ → Fe(phen)32+ + Fe(phen)34+ (S1)

Fe(phen)3 4+
+ H2O → Fe 2+ +
+ 2phen + 2H + phenO (S2)
Fe(phen)3 3+
+ phenO → Fe(phen)3 2+
+ Int (S3)
Fe(phen)3 2+
+ Int → Fe(phen)3 4+
+ phenO (S4)
Scheme 9 Kinetic model for the disproportionation of FeðphenÞ3 3 + (S1 and S2) and the
catalytic role of phenO (S3 and S4) (110).
Scheme 10 Mechanism for the oxidation of FeðphenÞ3 2 + by PMS under acidic

conditions.
The N-oxidations of the ligands bipy and tpy are significantly slower
than those of the corresponding Fe(II) complexes and most likely this is
the reason why bipyO and tpyO play no apparent role in the redox processes
between PMS and the corresponding complexes.
The PMS FeðbipyÞ3 2 + system bears no unusual features other than the
adduct formation, whereas the PMS FeðtpyÞ2 2 + oxidation reaction shows
autocatalysis.
The independent study on the oxidation of phen by PMS presented fur-
ther unique features. For a long time, the di-N-oxide derivative of phen
(phenO2) was considered a nonexistent compound due to the rigid planar
structure of phen and the limited space in the bay area of the molecule
(114,115). However, at last, the di-N-oxide was prepared by using the mix-
ture of F2, H2O, and CH3CN (116), but it still remained the prevailing
opinion that phenO2 cannot be produced by the use of peroxo compounds.
We have shown that the stepwise oxidation of phen by PMS results in
the formation of phenO, which is oxidized further to 1,10-
phenanthroline-N,N0 -dioxide (phenO2) in neutral aqueous solution. The
overall oxidation features a complex pH dependence: the pH significantly
affects the rate of the reaction and also controls the number of feasible oxi-
dation steps. Under strongly acidic conditions, a one-step process occurs and
HphenO+ (the protonated mono-N-oxide) is the sole product of the reac-
tion, which is protected from further oxidation steps by an intramolecular
hydrogen bond (Scheme 11). The rate of the mono-N-oxidation shows a
maximum value close to the neutral pH region. Such pH dependence
can be interpreted by considering that the protonation of the substrate gives
less opportunity for the oxidative attack on the nitrogen atom under acidic
conditions. Under basic conditions, the deprotonation of HSO5 deceler-
ates the oxidation because SO5 2 is much less reactive than the mono-
protonated form, in a similar manner to the oxidation reactions of halide
ions (101) and thiocyanate ion (102).
In nearly neutral medium, the mono-N-oxide (HphenO+) undergoes
deprotonation (pKa ¼ 7.3), which opens additional paths for the oxidation.
A multiphase kinetic pattern was found under such conditions (Scheme 11).
Singular value decomposition analysis of the spectra indicated the presence
of five absorbing species (117). At appropriately selected wavelengths, four
well-defined phases can be identified in the overall reaction, which are
clearly separated by three extrema in the kinetic traces (Fig. 9). Thus, the
new absorbing components are produced in consecutive processes. When
the oxidant is used in excess, the maxima and minima of the curves practi-
cally occur at the same reaction time when the initial phen concentration is
varied. This unique feature proves that the reaction order with respect to the
limiting reactant, phen, and its derivatives produced in the consecutive steps
N N O
k1H
+ H +
N PMS N H
PMS
Hphen+ HphenO+
N k1 N k2 N
O
k3 k4
Intermediate Products
O
N PMS N O PMS N PMS PMS
phen phenO phenO2

Scheme 11 Oxidation of phen by PMS under strongly acidic (top half) and neutral
(bottom half) conditions. Reprinted with permission from Beller, G.; Szabó, M.; Lente, G.;
Fábián, I. J. Org. Chem. 2016, 81, 5345–5353. Copyright 2016, American Chemical Society.
Fig. 9 Normalized kinetic curves recorded under different initial concentration of phen.
εapp is obtained by subtracting the nearly constant contribution of PMS from the
detected absorbance and the division of the residual curves by the total concentration
of phen. [PMS]0 ¼ 9.77 mM; [phen]0 ¼ 7.62 μM; 12.7 μM; 25.4 μM; 38.1 μM; 50.8 μM and
76.2 μM; [phosphate]tot ¼ 1.01 101 M; pH 6.7; T ¼ 25.0°C, λ ¼ 243 nm; path length
1.0 cm. Reprinted with permission from Beller, G.; Szabó, M.; Lente, G.; Fábián, I. J. Org.
Chem. 2016, 81, 5345–5353. Copyright 2016, American Chemical Society.
is one. Otherwise the positions of the extrema would change with [phen]0.
As expected in such a situation, the normalized kinetic traces are practically
identical within the limits of experimental error (Fig. 9).
The kinetic model for this system is rather complex and a comprehensive
evaluation method was used which solves numerically the corresponding
differential equation system and simultaneously fits the kinetic traces by
using a nonlinear least squares algorithm (107).
Several reports are found in the literature on the syntheses of N-oxides by
the conversion of the derivatives of pyridine (118,119), pyrazine (120,121),
quinoxaline (121), or tetrazole (122). However, only limited kinetic infor-
mation is available for these reactions (104,123). As shown here, kinetic and
mechanistic information about organic transformations very often could
help in systematically fine tuning the initial conditions (concentrations of
the reactants, pH, temperature, etc.) in order to design experimental proto-
cols for synthetic purposes.
Quite interestingly, PMS is often used for modeling oxidation reactions
of H2O2, although the two oxidants show distinct features. As noted earlier,
hydrogen peroxide is the most common precursor of the OH• radical which
reacts very rapidly, typically at diffusion controlled rates with the substrates.
In contrast, the hydroxyl radical is rarely formed from PMS and the reactions
of this oxidant proceed mainly via oxygen atom transfer steps, although one-
electron radical type reactions are also possible including the formation of
SO4 • as a reactive intermediate.
The reactions of H2O2 with iron-containing complexes are frequently
used for modeling O2 activation in biochemical processes (124). In these
biomimetic studies, the iron complexes act as synthetic model compounds
of heme or nonheme enzymes. When H2O2 reacts with an iron(III)complex
(LFeIII), very often the first step of the catalytic cycle is the formation of a
hydroperoxo intermediate (LFeIII–OOH). This species can decompose
via distinct pathways to give oxidants such as OH• or OOH• radicals or
high-valent iron complexes (LFeIV¼O or LFeV¼O). On the other hand,
when PMS is used as an oxidant, the formation of hydroperoxo species is
highly unlikely and catalytic cycles more often involve the formation of
high-valent iron-oxo transient species (125–128).
The oxidation of 2,4,6-trichlorophenol (TCP) with H2O2 and PMS
catalyzed by a water-soluble iron(III) porphyrin Fe(TPPS)+, where
TPPS ¼ meso-tetra(4-sulfonatophenyl)porphine, serves as an excellent
example to demonstrate this principle (125,129). It was found that in the
Fe(TPPS)+–H2O2–TCP system, the more active form of the catalyst is
the corresponding iron(III)hydroperoxo complex (Fe(TPPS)–OOH+).
The transformation of the catalyst into a much less active form was also
observed. The latter species was proposed to be an FeV oxo complex, which
is produced in a heterolytic O–O bond cleavage. Thorough understanding
of the mechanisms of such catalytic oxidations requires detailed studies on
the reaction of the catalyst and the oxidant in the absence of the substrate.
In the absence of TCP, the oxidation of Fe(TPPS)+ by H2O2 shows
multiphase kinetics and the first detectable intermediate is the FeV oxo spe-
cies (125). Most likely the hydroperoxo complex forms, too, but it might be
a very short-lived intermediate, which undergoes fast O–O bond cleavage.
At H2O2 excess, the iron(V) complex is further oxidized to give an inter-
mediate (Int) which is likely to contain a high-valent iron-oxo center and
a hydroxyl group. The final products of the reaction are the iron(III) com-
plex of the biliverdin analog formed from TPPS and 4-sulfobenzoic acid
(P1 and P2 in Scheme 12). The further oxidation steps are responsible for
the unusual kinetic phenomena observed during the catalytic oxidation of
TCP and the degradation of the catalyst.
In the case of PMS, the formation of a hydroperoxo species is unlikely.
Although the catalytic activity of the iron(V) oxo complex toward TCP is a
O H2O2 H2O2
H2O2
Fe(TPPS)+ Fe(TPPS)–OOH+ FeV(TPPS)+ Int P 1 + P2
O OH
Cl Cl Cl Cl
+ HCl
O Cl
DCQ TCP
Scheme 12 Fe(TPPS)+ catalyzed oxidation of TCP in the presence of excess H2O2.
O
PMS PMS PMS
Fe(TPPS)+ FeV(TPPS)+ Int P1 + P2
O OH
Cl Cl Cl Cl
+ HCl
O Cl
DCQ TCP
Scheme 13 Fe(TPPS)+ catalyzed oxidation of TCP in the presence of excess PMS.
lot smaller than that of Fe(TPPS)–OOH+, the high-valent iron derivative

seems to be the catalytically active form in this system and it is also the first
detectable intermediate both in the absence and presence of TCP
(Scheme 13).
In the absence of TCP, a three-step oxidation of the catalyst was observed
yielding the same intermediate (Int) and final products (P1 and P2) as in
the case of the oxidation by H2O2. In the catalytic cycle, the active inter-
mediate O¼FeV(TPPS), is either reduced back to Fe(TTPS)+ by TCP or
oxidized to Int by the oxidant. The latter process irreversibly deactivates
the catalyst. (125)
5. THE PHOTON AS A REACTANT

The ultimate experimental challenge in studying a complex redox
reaction is to control the concentration of the reactive intermediates inde-
pendently. In regular kinetic experiments the reactions are triggered by
mixing the reactants, and the concentrations of the transient species are
determined by the kinetics of the individual reactions steps. To some extent,
the steady-state concentrations of the intermediates can be varied by chang-
ing the experimental conditions. For example, changing the concentration
ratios of the reactants will alter the relative kinetic significance of the com-
peting reaction paths and also the concentration profiles of the intermedi-
ates. Thus, there is a possibility to affect the dominant reaction paths and
to separate the kinetic roles of the intermediates. However, there are only
limited options to utilize concentration-dependent studies, because the rate
of each reaction step is altered as a function of the initial concentrations of
the reactants. Photoinitiation may be useful to overcome this problem.
Earlier we found unexpected kinetic phenomena in diode array UV–vis
spectrophotometers (130). In these instruments, a relatively high energy
undispersed light beam enters the sample. If the chemical system contains
light sensitive species, the incoming light may trigger photoinitiated pro-
cesses. As shown in Scheme 14, such an instrument can be used as a
photoreactor.
The intensity and the spectral range of the entering light beam can be
controlled mainly by putting optical filters in the light path. The progress
of the reaction can be monitored spectrophotometrically by the same instru-
ment. This experimental setup was proven to be very valuable in studying
several complex processes as shown in the following examples.
The oxidation of 2,4,6-trichlorophenol (TCP) is often used to model the
efficiency of AOPs. When hydrogen peroxide is used as an oxidizing agent,
the reaction is slow and requires the use of some catalysts. The kinetics can
conveniently be studied by following the formation of Cl as a final product
using direct potentiometry. Lente and Espenson demonstrated that even the
fluorescent room light of a laboratory can accelerate the oxidation of TCP
by H2O2 and KHSO5 using Fe(TPPS)+ as a catalyst. (131) It was shown that
Diode array
W lamp Lens Cell
D lamp Slit
Filter Dispersion unit

Scheme 14 The general scheme of a diode array spectrophotometer. Reprinted with
permission from Fabian, I.; Lente, G. Pure Appl. Chem. 2010, 82, 1957–1973. Copyright
2010, IUPAC.
the oxidation of other chlorinated phenols by H2O2/Fe(TPPS)+ and the

oxidation of TCP by H2O2 in the absence of any catalyst are also affected
by light. The photosensitivity of the system was attributed to the formation
of 2,6-dichloro-1,4-benzoquinone (DCQ) which is a reactive intermediate.
The observations were explained by the conversion of DCQ into 2,6-
dichlorohydroquinone (DCHQ) and 3,5-dichloro-1,2,4-benzenetriol
(DCBT) (Eq. 25).
Light
ð25Þ
It was concluded that catalytic oxidation of DCQ by H2O2/Fe(TPPS)+

is slow. Thus, illumination enhances the overall oxidation process because
the photoreaction converts DCQ into DCBT which is oxidized more
rapidly than DCQ. DCHQ is also oxidized relatively quickly by H2O2/
Fe(TPPS)+ and DCQ is regenerated (131).
The studies on the photoreduction of DCQ in aqueous solution helped
to refine the kinetic model for this system (132). After illuminating a DCQ
solution with a halogen lamp (500 W), NMR analysis of the spent reaction
mixture confirmed the formation of DCQ and DCBT as well as 3,5-
dichloro-2-hydroxy-1,4-benzoquinone (DCHB) in small amounts. The
formation of dioxygen (O2) was also observed. The kinetics of the reaction
were systematically studied in a diode array spectrophotometer. The obser-
vations could be described quantitatively and a detailed kinetic model was
postulated for the conversion of DCQ (Scheme 15). In this model, the first
step is the excitation of DCQ which converts into a triplet state DCQ in an
DCQ + hn → *DCQ1 (S1)
DCQ → DCQ
* 1 * 3 (S2)
DCQ + H2O → DCQw

* 3 kw (S3)
DCQw → → → DCQH + O2 kred (S4)
DCQw → DCBT kadd (S5)
DCBT + DCQ → DCHB + DCHQ (S6)
Scheme 15 Kinetic model for the photoreduction of DCQ in aqueous solution (132).
intersystem crossing step. This species reacts with water producing DCQw
which is either an adduct or a transition state for the subsequent two path-
ways, the competition of which defines the final product distribution.
Diode array spectrophotometers could also be used to explore crucial
details of the autoxidation of sulfur(IV) in aqueous solution. This reaction
is one of the main sources of acid rain formation and is also important in
industrial processes such as flue gas desulfurization. For this reason, the
kinetics and the mechanism of this reaction were extensively studied under
a variety of conditions (133). There seems to be an agreement that sulfur(IV)
is not oxidized directly by O2 because this reaction is spin forbidden. Thus,
the autoxidation is always initiated by the direct reaction between sulfur(IV)
and an oxidant which acts as a catalyst. Various mechanisms were proposed
for these reactions which include radical, nonradical, or the combination of
these two types of reactions. Several transition metal ions catalyze the autox-
idation and, in a general sense, the kinetic model found in the iron(III) cat-
alyzed autoxidation (134) applies to many other cases. This model is based
on a redox cycle which involves the metal ion (Scheme 16). The model
specifies HSO3 as the reactive form of S(IV) but it can be any other form
of S(IV) depending on the pH.
First the sulfite ion radical, SO3 • is formed which readily produces the
peroxomonosulfate ion radical, SO5 • , with O2. In fact, this reaction should
be considered as the activation step of dioxygen because SO5 • is an
extremely reactive oxidant which is involved in various subsequent reaction
steps. A radical type chain reaction commences which regenerates the
Mn+ + HSO3− = M(n−1)+ + SO3•− + H+ (S1)

SO3•− + O2 = SO5•− (S2)
SO5•− + HSO3− = SO3•− + HSO5− (S3)
•− − •− − (S4)
SO5 + HSO3 = SO4 + HSO4
•− − •− − (S5)
SO4 + HSO3 = SO3 + HSO4
− − − (S6)
HSO5 + HSO3 = 2 HSO4
•− 2− (S7)
2SO3 = S2O6
M(n−1)+ + SO5•− + H+ = Mn+ + HSO5− (S8)
(n−1)+ − + n+ •− (S9)
M + HSO5 + H = M + SO4 + H2O
(n−1)+ •− + n+ − (S10)
M + SO4 + H = M + HSO4
Scheme 16 General kinetic model for the transition metal ion catalyzed autoxidation of
S(IV) (134).
oxidized form of the catalyst and yields the final product, SO4 2 =HSO4 . As
it was demonstrated earlier, complex formation and protolytic equilibria
may complicate the kinetics of the reaction further (134). The steady-state
concentrations of the reactive intermediates in this system depend on the
experimental conditions in a very complex way and predictions regarding
the significance of the individual reaction paths are uncertain.
Spectral changes at the characteristic spectral band of sulfur(IV) confirm a
slow concentration decay under continuous illumination in a diode array
spectrophotometer in the presence of dissolved oxygen (Fig. 10) (112).
This observation clearly confirmed that the autoxidation of sulfur(IV)
can be driven by light. The following sequence of reactions was proposed
for the interpretation of the experimental results. (The formula H2OSO2
is used in the equations because it is the dominant form of sulfur(IV) under
the acidic conditions applied.)
H2 O SO2 + hν ! *H2 O SO2 (26)

*H2 O SO2 + O2 ! HSO5 + H +
(27)

HSO5 + H2 O SO2 ! 2HSO4 + H +
(28)
The formation of HSO5 was postulated in kinetic models for catalytic

autoxidation of sulfur(IV) (134) and the observations make clear that photo-
initiation is suitable to generate such a reactive intermediate in this system.
0.9
15 min
0.6
A
0.3
0.0
240 270 300 330
l (nm)
Fig. 10 Autoxidation of sulfur(IV) during continuous irradiation in a HP-8453 spectro-
photometer. [S(IV)] ¼ 2.00 mM; [O2] ¼ 0.23 mM; [H2SO4] ¼ 0.50 M; path length 1.0 cm;
V ¼ 2.00 cm3; T ¼ 25.0°C. Reprinted with permission from Kerezsi, I.; Lente, G.; Fabian, I. Dal-
ton Trans. 2006, 955–960. Copyright 2006, The Royal Society of Chemistry.
When similar experiments were made in the presence of Ce(III) and I as

catalysts, the autoxidation process became fast, demonstrating the signifi-
cance of the synergy between photoinitiation and thermally activated cata-
lytic processes (135,136). In these studies an HP-8453 diode array
spectrophotometer was used which permitted the introduction of different
illumination sequences. The reaction mixtures were alternatively illumi-
nated for a certain period of time (ti) then kept in the dark (td) and the pro-
gress of the reaction was monitored at the characteristic adsorption band of
H2OSO2, λmax ¼ 276 nm.
As shown in Fig. 11, the rate of the decay of S(VI) increases by increasing
the amount of light entering the reaction mixture in the S(IV)–Ce(III)–O2
system (135).
An interesting feature of this system is that no reaction occurs in the
dark, i.e., the overall process can fully be regulated by the photon flux. It
was shown that the main photoactive species is not sulfur(IV) in this case,
and reactions (26)–(28) have negligible contribution to the overall process.
On the basis of detailed kinetic studies by varying the reactant concentra-
tions and using different illumination protocols, a kinetic model was pro-
posed which postulates that the initiation step is the photooxidation of
Ce(III) to Ce(VI) (Eq. 29). In turn Ce(VI) rapidly oxidizes S(VI) to
SO3 • (Scheme 17).
0.4
c
A (276 nm)
a b
0.3
0.2
0 100 200
t (s)
Fig. 11 Kinetic traces measured in a diode array instrument during the photoinitiated
autoxidation of sulfite ion. [Ce3+] ¼ 0.50 mM; [S(IV)] ¼ 1.00 mM; [O2] ¼ 0.19 mM (a and b),
0.22 mM (c); [H2SO4] ¼ 0.10 M; path length 1.0 cm; V ¼ 3.00 cm3; T ¼ 25.0°C; ti ¼ 5 s;
td ¼ 0 s (a), 5 s (b), 15 s (c); λ ¼ 276 nm. Reprinted with permission from Kerezsi, I.;
Lente, G.; Fabian, I. J. Am. Chem. Soc. 2005, 127, 4785–4793. Copyright 2005, American
Chemical Society.
Ce3+ O2
SO3−
.
Initiation: SO2
SO5−
.
Ce3+ + hn Ce4+
SO2
HSO4−
SO4−
.
Ce3+ HSO4−
2−
S2O8 Termination
Scheme 17 Kinetic model for the photoinitiated Ce(III)-catalyzed autoxidation of
sulfur(IV). Reprinted with permission from Kerezsi, I.; Lente, G.; Fabian, I. J. Am. Chem.
Soc. 2005, 127, 4785–4793. Copyright 2005, American Chemical Society.
CeðIIIÞ + H + + hν ¼ CeðIVÞ + 0:5H2 (29)

Thus, photoinitiation generates a very important chain carrier radical in
this system in an easily controlled manner because the amount of entering
light can be varied as desired by using optical filters or appropriate illumina-
tion protocols. This offers a special tool for studying the kinetic role of
SO3 • . It was shown that the reaction proceeds for a certain period of time
in the dark after the illuminating light is blocked. This is not unexpected
considering that all chain carriers are present in steady state at the end of
the illumination period and their concentrations decay in thermally activated
reaction steps. In this system, the termination step, i.e., the recombination of
two SO4 • radicals into S2 O8 2 is not particularly efficient to stop the autox-
idation and a substantial fraction of the reaction occurs in the dark. Careful
analysis of the experimental data resulted in the possibility of quantifying the
fraction of the reaction occurring in the dark (Q). The proposed kinetic
model provides excellent interpretation of all the experimental observations.
As an example, fitting Q as a function of illumination time during experi-
ments with intermittent dark periods is shown in Fig. 12.
In broad terms, the kinetic features of the photoinduced autoxidation of
S(IV) in the presence of iodide ion are very similar to those found in the
S(IV)–Ce(III)–O2 system. The initial rate of the reaction was studied as a
function of reactant concentrations and pH. The kinetic model for this sys-
tem is shown in Scheme 18.
This model postulates again a radical chain reaction. The primary
photoactive species is I in this system. However, the experiments undoubt-
edly proved that somehow H2OSO2 is also involved in the initiation phase.
This species is not able to initiate a radical chain reaction in the absence of I
(112), thus, the simplest way to explain the results was to assume energy
1.0
0.5
0.0
0 5 10 15
ti (s)
Fig. 12 Fraction of the Ce(III) catalyzed autoxidation of S(IV) occurring in the dark (Q) as
a function of illumination time (ti) during experiments with intermittent dark periods.
Solid line: best fit to the model outlined in Scheme 17. [Ce3+] ¼ 0.50 mM; [S(IV)] ¼
1.00 mM; [H2SO4] 0.10 M; V ¼ 3.00 cm3; T ¼ 25.0°C; td ¼ 90 s. Reprinted with permission
from Kerezsi, I.; Lente, G.; Fabian, I. J. Am. Chem. Soc. 2005, 127, 4785–4793. Copyright
transfer between H2OSO2 and I. Such a reaction was proposed before in a
related system (112). Reactions S1–S3 (Scheme 18) are the initiation steps in
this process. They are written as nonelementary reactions because their inti-
mate details could not be explored. However, it is quite possible that these
reactions include a hydrated electron and a superoxide radical.
It was assumed that the consumption of O2 in the initiation reactions is
negligible compared to that in the chain-carrying steps, and Eq. (30) was
derived using the long-chain approach.
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

u
u aI NI + aS NS k2S4 ½I 2 ðp1 ½SðIVÞ + p1 ½I ½SðIVÞÞ2
u
v¼t
kS10 ðp1 ½SðIVÞ + p1 ½I ½SðIVÞÞ2 + kS11 kS4 ½I 2 ðp1 ½SðIVÞ + p1 ½I ½SðIVÞÞ + kS11 kS4 ½I 4
(30)
where NI and NS are the absorbed photon count per unit volume for iodide
ion and S(IV), respectively, while αI and αS are the corresponding quantum
efficiencies. Parameters p1 and p2 are given as follows:
kS5a ½H + + kS5b KS14

p1 ¼ (31)
K8 ð½H + + KS14 Þ
I−+ hn + H2O•SO2 + O2 → I• + SO4•− + H2O v1= aINI (S1)

H2O•SO2 + hn → *H2O•SO2 v2= aSNS (S2)
*H2O•SO2 + I− + O2 → I•+ SO4•−+ H2O fast (S3)
SO4•−+ I−→ I• + SO42− v4= kS4[SO4•−][I−] (S4)
I•+ H2O•SO2→ SO3•−+ I− + 2H+ v5a= kS5a[I•][H2O•SO2] (S5a)
I + HSO3 → SO3 + I + H
• − •− − +
v5b= kS5b[I ][HSO3 ]
• − (S5b)
SO3 + O2→ SO5

•− •−
v6= kS6[SO3 ][O2] •− (S6)
SO5 + H2O•SO2→ SO4 + HSO4 + H

•− •− − +
v7= kS7[SO5 ][H2O•SO2]
•− (S7)
• − •− [I •−]
I+I I2 fast, S8 = •2 − (S8)
[I ][I ]
I2•− + H2O•SO2→ SO3•− + 2I− + 2H+ v9a= kS9a[I2•−][H2O•SO2] (S9a)
I2•−+ HSO3−→ SO3•− + 2I− + H+ v9b= kS9b[I2•−][HSO3−] (S9b)
SO4•− + SO4•−→ S2O82− v10= kS10[SO4•−]2 (S10)
SO4•− + I2•−→ SO42− + I2 v11= kS11[SO4 ][I2 ] •− •−

(S11)
I2 + I2 → I3 + I
•− •− − − v12= kS12[I2•−]2 (S12)
I2 + H2O•SO2 + H2O → HSO4 + 2I + 3H − − + fast (S13)

[HSO3−][H+]
fast, S14 =
− + [H2O•SO2] (S14)
H2O•SO2 HSO3 + H
[I3−]
I2 + I− I3− fast, = (S15)
tri [I2][I−]
Scheme 18 Kinetic model for the photoinduced autoxidation of S(IV) in the presence of
iodide ion (136).
kS9a ½H + + kS9b KS14

p2 ¼ (32)
½H + + KS14
Several of these parameters are known from independent studies and
were kept constant while the rest were estimated by fitting all experimental
initial rates to Eq. (30). The eight parameters required to fit the data are listed
in Table 1.
As demonstrated by Figs. 13 and 14, the model provides excellent
description of the experimental observations.
In the previous examples, light is necessary to replenish some of the chain
carriers. When illumination is turned off the reaction terminates after all
intermediates are consumed in subsequent thermally activated steps and
the reactants are not fully consumed. In these cases, a photon should be con-
sidered as one of the reactants because an appropriate amount of it needs to
Table 1 Kinetic Parameters for the Photoinitiated Autoxidation of Sulfur(IV) in the

Presence of Iodide Ion (136)
Parameter Value Remark
α1 0.29 Refs. (137,138)
α2 0.35 Ref. (112)
1 1
k10 4.4 10 M
8
s Ref. (139)
1 1
k12 7.7 10 M
9
s Ref. (140)
4 1
p2 2.4 10 s Estimated using Ref. (141)
k4 (1.8 0.1) 108 M1 s1 Fitted
1 1
k11 (1.5 0.3) 10 M
10
s Fitted
1 1
p1 (1.4 0.2) 10 M
4
s Fitted
7.5
v0 (mM/s)
5.0
2.5
0.0
0.0 0.2 0.4 0.6
[I−] (mM)
Fig. 13 The initial rate as a function of iodide ion concentration in the photoinitiated
and iodide-catalyzed autoxidation of sulfur(IV). The solid lines show the best fit on
the basis of the proposed model (Scheme 18). [S(IV)] ¼ 3.00 (■), 2.00 (♦), 1.00 (▲),
and 0.70 mM (●); [H2SO4] ¼ 0.575 M; V ¼ 3.00 cm3; T ¼ 25.0°C. Reprinted with permission
from Kerezsi, I.; Lente, G.; Fabian, I. Inorg. Chem. 2007, 46, 4230–4238. Copyright 2007,
be consumed to complete the reaction. However, light can also be a real

initiator of a complex reaction, i.e., initial illumination triggers the reaction
which maintains itself after light is turned off.
It is well known that chlorate ion is a very sluggish oxidant in aqueous
solution. It is certainly not able to oxidize iodine in the dark. However, the
ClO3 I2 reaction goes to completion relatively quickly in a diode array
10.0
v0 (µM/s)
5.0
0.0
0.0 1.0 2.0
pH
Fig. 14 The initial rate as a function of pH in the photoinitiated and iodide-catalyzed
autoxidation of sulfur(IV). The solid line shows the best fit on the basis of the proposed
model (Scheme 18). [S(IV)]) ¼ 2.00 mM; [I] ¼ 0.20 mM; μ ¼ 1.0 M (Na,H)ClO4;
V ¼ 3.00 cm3; T ¼ 25.0°C. Reprinted with permission from Kerezsi, I.; Lente, G.; Fabian, I.
Inorg. Chem. 2007, 46, 4230–4238. Copyright 2007, American Chemical Society.
c
f
0.06
g
h
A (460 nm)
0.03
0.00
0 100 200
t (s)
Fig. 15 Kinetic curves in the photoinitiated reaction of I2 with ClO3 . Solid lines show the
best fit on the basis of the proposed kinetic model (Scheme 19). [I2] ¼ 88 μM; ½ClO3 ¼
25:1 mM (c, h, i), 16.7 mM (f ), 8.3 mM (g); [H+] ¼ 0.948 M (c, f, g), 0.237 M (h), 0.356 M (i);
continuous illumination; T ¼ 25.0°C. Reprinted with permission from Galajda, M.; Lente, G.;
Fabian, I. J. Am. Chem. Soc. 2007, 129, 7738–7739. Copyright 2007, American Chemical
Society.
spectrophotometer (142). The stoichiometry of this process is straightfor-

ward (Eq. 33), however, the kinetic traces exhibit autocatalytic features
(Fig. 15):
5ClO3 + 3I2 + 3H2 O ¼ 5Cl + 6IO3 + 6H + (33)

It was confirmed that after an initial illumination period, the reaction

goes to completion in the dark (Fig. 16).
In the initial phase of the reaction, light is essential for generating an
autocatalyst which, by definition, accelerates the overall process. Once
the reaction is triggered, the autocatalyst keeps regenerating itself via thermal
reaction steps and light is not required to complete the reaction. It was
shown that the autocatalyst is hypochlorous acid in this system. When small
amounts of HOCl were added to the reaction mixture the same type of
kinetic traces were observed in a conventional double beam spectrophotom-
eter, where photoexcitation has a negligible role compared to its role in the
diode array spectrophotometer. The kinetic model (Scheme 19) provides
excellent description of the observations as shown in Fig. 15.
In the previous examples, a photon is always used as an extra reactant.
This adds a new dimension to the kinetic studies because the experiments
provide invaluable, otherwise inaccessible or vague information on the reac-
tive system. A key issue is the kinetic coupling between the photolytic and
thermal reaction steps. This can be explored by varying the experimental
conditions and illumination protocols. As demonstrated, diode array spec-
trophotometers are very useful for studying photoinitiated reactions, how-
ever, there are inherent limitations associated with such a use of these
instruments. Since the same light source is used for photoinitiation and
0.2
A (460 nm)
Dark (40–500 s)
0.1
Continuous
illumination
0.0
0 300 600
t (s)
Fig. 16 Kinetic curve measured in the photoinitiated reaction of I2 with ClO3 . A long
dark period was inserted into the measurement to prove the fact that no light is nec-
essary to maintain the reaction in the rapidly decreasing region. [I2] ¼ 0.39 mM;
½ClO3 ¼ 25 mM; [H+] ¼ 0.948 M; T ¼ 25.0°C. Reprinted with permission from
Galajda, M.; Lente, G.; Fabian, I. J. Am. Chem. Soc. 2007, 129, 7738–7739. Copyright
I2 + ClO3− + H2O + hn → IO3− + H2OI+ + Cl− (S1)

− −
n1 = α1a[I2][ClO3 ][H ] + α1b[I2][ClO3 ]
+
HOCl + H+ + Cl− Cl2 + H2O (S2)

−
K2 = [Cl2]/([HOCl][Cl ][H ])
+
H2OI+ + ClO3− → IO3− + HOCl + H+ (S3)

v3 = kS3[H2OI+][ClO3−]/[H+]
H2OI+ + 2HOCl → IO3− + 2Cl− + 4H+ (S4)
v4 = kS4[H2OI ][HOCl]
+
I2 + Cl2 + 2H2O → 2H2OI+ + 2Cl− (S5)

v5 = kS5[I2][Cl2]
Scheme 19 Kinetic model for the photoinduced oxidation of I2 by ClO3 (142).
absorbance measurements: (i) the irradiation and detection wavelength

ranges need to be the same; (ii) the intensity of the illuminating light cannot
be varied freely; and (iii) the absorbance change cannot be followed in the
dark. For example, in the case of the ClO3 I2 system it could be
established that the reaction is completed in the dark but the kinetic profile
could not be recorded (Fig. 16).
The noted problems with diode array spectrophotometers can be cir-
cumvented by using dedicated photoreactors (143,144). Recently, we have
assembled such an instrument from commercially available Avantes modules
and optical components (Scheme 20, Fig. 17).
The centerpiece of this device is a cell holder which has four ports for
connecting optical fibers (Fig. 17). One port is used to receive the excitation
light which is produced by a high energy laser-pumped xenon lamp. In
order to achieve higher photoinitiation efficiency, the intensity of the excit-
ing light is practically doubled by placing a mirror in the opposite port. The
other two ports accommodate a fiber optic spectrophotometer with CCD
detector. The two light beams are perpendicular and the cell holder was
modified such that the vertical positions of the ports can be adjusted. This
arrangement makes possible the separation of the planes of excitation and
detection in order to avoid any interference by the exciting light during
absorbance measurements. Fluorometric detection is also possible in this
instrument when the excitation and detection planes are set to the same
height, i.e., when the two light beams cross each other. In this instrument,
the spectral range and the intensity of the exciting light can be varied flexibly
without limiting the absorbance or fluorescence measurements. The light
intensity of the spectrophotometer is about 10,000 times smaller than that
Scheme 20 The general scheme of the modular Avantes photoreactor. Reprinted with
permission from Ditroi, T.; Kalmar, J.; Pino-Chamorro, J. A.; Erdei, Z.; Lente, G.; Fabian, I. Pho-
tochem. Photobiol. Sci. 2016, 15, 589–594. Copyright 2016, The Royal Society of Chemistry
and Owner Societies.
Light from excitation lamp

Cable from Cable from
detection excitation
Light from detection lamp
Light to spectrometer
lamp light
Adjustable source
Fixed
Fixed
Sample
Adjustable
Cable to
Mirror plug Magnetic stirrer Mirror spectrometer
Flow in-and outlet (below) plug
Fig. 17 The scheme and the photo of the sample holder of the modular Avantes
photoreactor. Reprinted with permission from Ditroi, T.; Kalmar, J.; Pino-Chamorro, J. A.;
Erdei, Z.; Lente, G.; Fabian, I. Photochem. Photobiol. Sci. 2016, 15, 589–594. Copyright
2016, The Royal Society of Chemistry and Owner Societies.
of the excitation light and was not able to trigger photochemical reactions in
the photosensitive systems studied so far. This feature makes it possible to
monitor the thermal reactions during the dark periods. As an example,
the results obtained in the ClO3 I2 reaction are shown in Fig. 18.
One of the kinetic traces was obtained with continuous illumination. In
the other case, the excitation light was turned off after 60 s. Completion of
the reaction requires a longer time in the dark but the absorbance decays are
essentially parallel in the two experiments. This observation is consistent
with the kinetic model proposed for this reaction (Scheme 19). It is an inher-
ent feature of autocatalytic systems that the autocatalyst accumulates to a
B: dark
0.2
A (460 nm) A: illumination
0.1
0.0
0 100 200
t (s)
Fig. 18 Kinetic traces detected in the chlorate–iodine reaction. [I2] ¼ 0.38 mM;
½ClO3 ¼ 25 mM; [HClO4] ¼ 1.0 M; path length: 1.00 cm; T ¼ 25.0°C; V ¼ 3.00 cm3; stir-
ring: 800 rpm. Curve A: continuous illumination. Curve B: illumination switched off after
60 s. Reprinted with permission from Ditroi, T.; Kalmar, J.; Pino-Chamorro, J. A.; Erdei, Z.;
Lente, G.; Fabian, I. Photochem. Photobiol. Sci. 2016, 15, 589–594. Copyright 2016, The
Royal Society of Chemistry and Owner Societies.
certain concentration level before the reaction significantly speeds up. Con-
tinuous illumination steadily generates the autocatalyst and its “critical” con-
centration is reached earlier than in experiments when illumination is turned
off. After this point, the thermal reaction steps become the main source of
the autocatalyst and the contribution of photolytic reactions to the kinetics
can be neglected.
Testing the applicability of the photoreactor in various reactive systems
revealed that it also makes possible monitoring kinetic processes in hetero-
geneous systems. Such results are discussed in the following section.
6. SELECTED KINETIC STUDIES ON HETEROGENEOUS

SYSTEMS
The first heterogeneous system studied by adapting some of the
methods of homogeneous reaction kinetics introduced in the previous chap-
ters was the “photocatalytic” reaction of a silica–titania (SiO2–TiO2) com-
posite aerogel (145).
The term “photocatalysis” is widely used to denote photochemical pro-
cesses where light is primarily absorbed by a photosensitizer, the subsequent
photochemical reactions of which lead to the consumption of a substrate,
while the sensitizer is replenished. In this case a catalytic cycle is established,
similarly to the case of thermal catalytic reactions. However, the terminol-

ogy is not precise because essentially the photon is consumed and it is an
initiator and not a catalyst in the overall process. The photoactive substances
are usually metal oxide semiconductors, which are used as suspensions dis-
persed in the solutions of the target substances (146–148). The usual exper-
imental setup to study a “photocatalytic” reaction of a semiconductor is to
circulate the heterogeneous reaction mixture in a batch reactor which con-
tains an intensive, immersion type UV light source (146–148). The progress
of the reaction is monitored by sampling and offline analysis of the mixture
during illumination. Usually kinetic information is obtained only for the
concentration change of the substrate as a function of time. Control exper-
iments in the dark and/or under illumination but using an inert suspension
instead of the photoactive agent are usually carried out.
In the case of the composite silica–titania aerogel, the dye methylene
blue and salicylic acid were used as model compounds to be photochemi-
cally degraded. The first experiments were conducted in a conventional
immersion-type batch reactor with offline sampling. The experimental
kinetic traces proved to be unusual. The widely accepted theory to account
for the degradation of a target substance in the presence of an excited
semiconductor is that the substances either react directly with the excited
semiconductor or with secondary radicals generated in the reaction of the
excited semiconductor and water or dissolved oxygen. The target substance
is oxidized or reduced to various intermediates and final products in subse-
quent reaction steps. This mechanism is generalized in the Langmuir–
Hinshelwood model (149,150), which is widely used to approximate the
kinetic properties of “photocatalytic” reactions. However, for the interpre-
tation of the observations in the silica–titania aerogel systems, this model was
inadequate (145). In order to extract more kinetic information on the pho-
toreaction, additional experiments were made with the custom-built
Avantes photoreactor introduced in the previous chapter. Kinetic traces
could be monitored with a few seconds time resolution as demonstrated
in Fig. 19.
On the basis of meticulous experiments, the following, somewhat
unconventional conclusions were drawn.
(i) Quantitative UV–vis spectrophotometry can be conducted in suspen-
sions using the Avantes spectrophotometer. The apparent absorbance
arising from the light scattering of the dispersed aerogel particles is pro-
portional to their concentration, and this apparent absorbance is addi-
tive to the real absorbance of dissolved colored substances (144,151).
Fig. 19 Experimental kinetic curves of the photodegradation of methylene blue (MB) in

the presence of microcrystalline anatase and suspended silica–titania aerogel
(A13) recorded in the Avantes photoreactor. All kinetic curves were recorded by online
UV–vis spectrophotometry at 664 nm and only every fifth experimental point is shown
for clarity. A control experiment is also shown where A13 was mixed to MB in the dark.
Reprinted with permission from Lazar, I.; Kalmar, J.; Peter, A.; Szilagyi, A.; Gyori, E.; Ditroi, T.;
Fabian, I. Appl. Surf. Sci. 2015, 356, 521–531. Copyright 2015, Elsevier B.V.
It should be emphasized that light scattering cannot be generally quan-

tified in most of the commercial UV–vis spectrophotometers. The rea-
sons for the advantageous response of the Avantes spectrophotometer
regarding suspensions were not explored. It is most likely due to some
unique optical arrangement in this instrument.
(ii) The contributions of photochemical and other physicochemical pro-
cesses to the overall degradation rate of the substrate can be separated
on the basis of high time resolution, quantitative, online spectropho-
tometric experiments.
(iii) The adsorption of the target substrate to the surface of the aerogel is a
key step which controls the overall kinetics of the “photocatalytic”
process. The quasi-irreversible adsorption of the dissolved substance
masks the surface of the aerogel, like a built-in light filter. This reduces
the effective surface accessible for the flux of photons, thus, signifi-
cantly lowering the overall rate of photodegradation.
In order to account for the noted features of these heterogeneous systems, a
simple extension of the Langmuir–Hinshelwood model was made.
According to this model, the substrate is photodegraded in the solution

phase as well as on the surface of the heterogeneous catalyst. However,
the adsorption of the substrate on the surface lowers the effective concen-
tration of the active semiconductor (Scheme 21).
Finally, all photochemical and control kinetic experiments could be
simultaneously modeled by using the kinetic model in Scheme 21, as shown
in Fig. 20 (145).
In order to understand complicated heterogeneous catalytic processes, it
is essential to study the interaction of the reactants with the surface of the
catalyst. The rate of a heterogeneous catalytic reaction is typically deter-
mined by the rate of the adsorption of the reactants on the catalytic centers.
Thus, reliable experimental data on the kinetics of adsorption–desorption
processes are vital in exploring the mechanism of complicated redox pro-
cesses. It is important to note that even the rate of adsorption is dependent
on many factors, e.g., the morphology of the surface, or in the case of porous
supports, the effective rate of diffusion. Understanding these factors can help
to predict some key characteristics of porous supports, which will in turn
determine, for example, the overall rate of a catalytic process.
Silica (SiO2) aerogel is an ideal support for heterogeneous catalysts,
because it has a large specific surface area (c. 1000 m2/g) and open meso-
porous (dpore ¼ 5–100 nm) structure. By adopting some fundamental
methods used in homogeneous reaction kinetics, the intimate details of
the adsorption properties of silica aerogel were explored (152). In a typical
experiment, the stable suspension of silica aerogel was stirred in a spectro-
photometric cuvette and the solution of a model dye, methylene blue
(MB), was injected into it. The process was monitored online by UV–vis
SA + h products kS1 (S1)

SA + A13surface (masked surface) kS2 (S2)
SA + A13surface + h products + A13surface kS3 (S3)
Scheme 21 Kinetic model for the photodegradation of salicylic acid (SA) in the pres-
ence of suspended silica–titania aerogel (A13). The symbol hν indicates that photons
take part in a reaction step. E is the molar absorption coefficient of SA and l is the optical
path length (145).
Adsorption
Adsorption
80
90 Control Control
60
A13
A13
80 40
Extrapolation Extrapolation
20
c/c 0 (%)
c/c 0 (%)
Salicylic acid (SA) Methylene blue (MB)
70
99
98 Adsorption 99 Adsorption
97
96 98
0 50 100 150 0 50 100 150
t (min) t (min)
Fig. 20 Experimental kinetic curves of the photodegradation of salicylic acid (SA, left
panel) and methylene blue (MB, right panel) in the absence (●) and in the presence
of suspended titania–silica aerogel A13 (■). The bottom figures were magnified from
the top ones. All kinetic curves were recorded in an immersion-type batch photoreactor
by offline UV–vis quantification. Continuous lines are the results of mathematical sim-
ulation based on the extended Langmuir–Hinshelwood kinetic model (Scheme 21).
Dashed lines represent extrapolated kinetic curves calculated without taking into
account the kinetic effect of the adsorption of the substrate on the surface. Reprinted
with permission from Lazar, I.; Kalmar, J.; Peter, A.; Szilagyi, A.; Gyori, E.; Ditroi, T.;
Fabian, I. Appl. Surf. Sci. 2015, 356, 521–531. Copyright 2015, Elsevier B.V.
spectrophotometry using the Avantes CCD spectrophotometer. Quantita-

tive spectrophotometry was conducted in the suspension similarly as in the
silica–titania system detailed in the previous section. The experiments
showed that the adsorption of the dye is a fast process, complete in about
100 s, followed by a kinetically separated slower process. To some extent,
the presence of the two processes was intriguing.
The adsorption kinetics was studied as a function of the concentration of
the dye or the suspension in two sets of experiments. The kinetic curves
were evaluated by using conventional methods, such as initial rates and
nonlinear fitting the experimental data to the appropriate rate expression.
The corresponding adsorption isotherm was measured independently in

batch experiments. The same isotherm was also reconstructed from the
kinetic experiments with excellent agreement. The shape of the isotherm
and the initial rates evidently suggest that the fast adsorption process follows
a simple Langmuir-type mechanism. The second, slower kinetic process was
additionally analyzed by the decomposition of the time-resolved UV–vis
spectral traces in the 200–800 nm region. This analysis proved that the
slower second process cannot be the multilayer adsorption of the dye. This
process was confirmed to be the aggregation of the dye-covered aerogel par-
ticles, which is accompanied by a decrease in the apparent absorbance of the
suspension. A kinetic model with two reversible steps was postulated for the
interpretation of the experimental observations (Scheme 22) (152).
MB + S SMB k1 = (1.09 ± 0.03) × 103 M–1 s–1 (S1)
SMB MB + S k2 = (2.0 ± 0.1) × 10–2 s–1 (S2)
2SMB aggregate k3 = (1.4 ± 0.1) × 102 M–1 s–1 (S3)
aggregate 2SMB k4 = (3.1 ± 0.1) × 10–3 s–1 (S4)
Scheme 22 Kinetic model for the adsorption of methylene blue (MB) on silica aerogel
particles. A free adsorption site on the aerogel is symbolized by S, and an occupied site
by SMB. The first process (S1 and S2) is the reversible adsorption of the dye, and the
second process (S3 and F4) is the reversible aggregation of those aerogel particles
which are covered by MB. The time-dependent concentrations of dissolved MB, free
and covered aerogel particles, and aggregates are [MB], [S], [SMB] and [aggr], respec-
tively. The initial (total) concentrations of MB and the aerogel are cMB and cgel, respec-
tively. Time-dependent surface coverage is θ. The adsorptive capacity of the aerogel is
s ¼ 48 μmol/g. The rate constants (kS1 kS4) were determined by global kinetic data
fitting (152).
The known UV–vis spectrum of the dye and the apparent absorbance
spectrum of the initial aerogel suspension were incorporated into the model.
Eventually, a global kinetic data fitting proved that the two-step model
(Scheme 22) adequately describes the experimental traces.
To complete the mechanistic study, the morphological characterization
of the silica aerogel was performed by SEM and N2 porosimetry (152). The
suspension of silica aerogel was further characterized by NMR
cryoporometry (153) and NMR diffusiometry (154,155). These methods
gave excellent supporting information to elaborate further the mechanistic
aspects of the kinetics of adsorption. It was established that silica aerogel
retains its open mesoporous structure while dispersed in water. The self-
diffusion of water is somewhat hindered inside the pores, but the particles
are highly permeable for the solvent. Thus, it was further proved by the
structural study that the diffusion and subsequent adsorption of small mol-
ecules cannot be prolonged inside the pores of silica aerogel.
An additional interesting observation was made during the above
detailed adsorption experiments. The clean walls of the spectrophotometric
cuvette used in the study also adsorbed methylene blue. The kinetics of the
phenomenon were investigated in detail (156), and the results were found to
be in good agreement with previously published results (157).
The above examples demonstrate well the possible gains of adapting the
methodology of homogeneous reaction kinetics for studying non-
homogeneous physicochemical processes. The usefulness of this approach
lies in collecting indirect information on the mechanistic aspects of compli-
cated reactions which cannot be studied by any direct means. In this sense,
the approach described earlier can open new horizons in understanding the
mechanisms of interface processes, heterogeneous catalytic processes, elec-
trode reactions, and other elusive nonhomogeneous systems.
7. CONCLUDING REMARKS
Because of the diversity of complex redox processes, it is difficult to
suggest a bulletproof strategy for exploring their mechanisms. In general,
the stoichiometry as a function of reaction time, the final stoichiometry,
and the kinetic features need to be studied under as broad experimental con-
ditions as possible. In many cases, thorough characterization of the reactive
intermediates is the key to in-depth understanding of the mechanism. While
simplified methods may provide essential information on various details, the
pitfalls of using such techniques are obvious. Thus, simultaneous and critical
evaluation of all available experimental results is essential to validate the con-
clusions. Whenever it is feasible, the experiments should be complemented
with theoretical calculations on potential reaction paths and structures of the
intermediates and transition states.
In the last few decades, enormous developments have been achieved in
computational methods. These techniques opened new dimensions in col-
lecting, evaluating, and fitting experimental data as well as predicting energy
profiles for even complex reactive systems. There is an increasing number of
examples when calculated geometries and energies for reactants, products,
intermediates, and transition states corroborate mechanistic conclusions
obtained on the basis of experimental observations. Nevertheless, there is
plenty of room for improvement in order to obtain more precise theoret-
ical predictions. In solution phase, particularly in water, solute–solvent
interactions may have tremendous effects on the kinetic behavior of a reac-
tive system. As of yet, the arsenal of computational chemistry is not strong
enough to handle such problems with sufficiently high precision. How-
ever, this area is intensively studied and these efforts coupled with the ever
growing computational capacities are expected to bring breakthrough
results in the future.
As far as experimental methodologies are concerned, we have witnessed
massive developments in this area, too. The new devices offer new oppor-
tunities to collect larger experimental datasets with higher precision, time
resolution, and sensitivities than before. Many of the recent studies have
explored the possibilities of the nonconventional use of experimental
methods for kinetic purposes. Thus, NMR, IR, and CD spectroscopies have
been used for monitoring the progress of chemical reactions. Mass spectrom-
etry has become an invaluable tool to identify and characterize reactive
intermediates in a number of reactions. Perhaps these techniques do not
match the performance of well-established kinetic methods, but, there is
a promising tendency to construct dedicated instruments or accessories
which satisfy better the requirements of kinetic studies. As an example, spe-
cial low-field NMR spectrometers have been introduced in the market
recently with the primary objective of studying reaction kinetics.
Recent studies on complex reaction mechanisms made clear that reliable,
fairly detailed kinetic models can be postulated for these systems. The results
have clarified fundamental aspects of solution phase reaction kinetics and
provided a solid chemical background for practical applications. These
achievements will certainly inspire further research in this field.
ACKNOWLEDGMENTS
This research was supported by the Hungarian Science Foundation (OTKA: NK-105156), as
well as by the EU and cofinanced by the European Regional Development Fund under the
project GINOP-2.3.2-15-2016-00008. J.K. is indebted to the University of Debrecen (RH/
751/2015) and also to the Ministry of Human Capacities of Hungary (New National
Excellence Program) for financial support.
REFERENCES
1. Hush, N. S. Trans. Faraday Soc. 1961, 57, 557–580.
2. Marcus, R. A. Annu. Rev. Phys. Chem. 1964, 15, 155–196.
3. Quite often the reactions show marked pH dependencies associated with the acid–base
equilibria of the reactive components, i.e., the reacting species exist in different pro-
tonated forms in aqueous solution. The protolytic equilibria are typically established
at diffusion controlled rate and the corresponding reactions can be treated as fast
preequilibria. Consequently, the concentration ratios of the various protonated com-
ponents are determined by the actual pH. In this paper, we follow the commonly
accepted practice in the literature, and use only one name or formulae for a component
which actually corresponds to all protonated forms of a reactant or intermediate being in
equilibrium. Distinction between the different forms is made only when it is required
by the clarity of presentation.
4. Forni, L.; Bahnemann, D.; Hart, E. J. J. Phys. Chem. 1982, 86, 255–259.
5. Staehelin, J.; Hoigne, J. Environ. Sci. Technol. 1982, 16, 676–681.
6. B€uhler, R. E.; Staehelin, J.; Hoigne, J. J. Phys. Chem. 1984, 88, 2560–2564.
uhler, R. E.; Hoigne, J. J. Phys. Chem. 1984, 88, 5999–6004.
7. Staehelin, J.; B€
8. Tomiyasu, H.; Fukutomi, H.; Gordon, G. Inorg. Chem. 1985, 24, 2962–2966.
9. Chelkowska, K.; Grasso, D.; Fabian, I.; Gordon, G. Ozone Sci. Eng. 1992, 14, 33–49.
10. Bard, A. J.; Parsons, R.; Jordan, K. Standard Potentials in Aqueous Solution; Marcek
Dekker, Inc.: New York, 1985; p 829.
11. Nemes, A.; Fabian, I.; Gordon, G. Inorg. React. Mech. 2000, 2, 327–341.
12. Nemes, A.; Fabian, I.; Gordon, G. Ozone Sci. Eng. 2000, 22, 287–304.
13. Nemes, A.; Fabian, I.; van Eldik, R. J. Phys. Chem. A 2000, 104, 7995–8000.
14. Thorell, H. D.; Stenklo, K.; Karlsson, J.; Nilsson, T. Appl. Environ. Microbiol. 2003, 69,
5585–5592.
15. Streit, B. R.; DuBois, J. L. Biochemistry 2008, 47, 5271–5280.
16. Gordon, G.; Kieffer, R. G.; Rosenblatt, D. H. Prog. Inorg. Chem. 1972, 15, 201–286.
17. Sant’Anna, R. T. P.; Santos, C. M. P.; Silva, G. P.; Ferreira, R. J. R.; Oliveira, A. P.;
Cortes, C. E. S.; Faria, R. B. J. Braz. Chem. Soc. 2012, 23, 1543–1550.
18. White, G. C. Handbook of Chlorination and Alternative Disinfectants; Van Nostrand
Reinhold: New York, 1992.
19. Vogt, H.; Balej, J.; Bennett, J. E.; Wintzer, P.; Sheikh, S. A.; Gallone, P.;
Vasudevan, S.; Pelin, K. In: Ullmann’s Encyclopedia of Industrial Chemistry; Wiley-
VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2000.
20. Collman, J. P.; Boulatov, R.; Sunderland, C. J.; Shiryaeva, I. M.; Berg, K. E. J. Am.
Chem. Soc. 2002, 124, 10670–10671.
21. Odeh, I. N.; Francisco, J. S.; Margerum, D. W. Inorg. Chem. 2002, 41, 6500–6506.
21a. Csordas, V.; Bubnis, B.; Fabian, I.; Gordon, G. Inorg. Chem. 2001, 40, 1833–1836.
22. Wang, L.; Margerum, D. W. Inorg. Chem. 2002, 41, 6099–6105.
23. Noss, C. I.; Hauchman, F. S.; Olivieri, V. P. Water Res. 1986, 20, 351–356.
24. Tan, H. K.; Wheeler, W. B.; Wei, C. I. Mutat. Res. 1987, 188, 259–266.
25. Napolitano, M. J.; Stewart, D. J.; Margerum, D. W. Chem. Res. Toxicol. 2006, 19,
1451–1458.
26. Darkwa, J.; Olojo, R.; Chikwana, E.; Simoyi, R. H. J. Phys. Chem. A 2004, 108,
5576–5587.
27. Ison, A.; Odeh, I. N.; Margerum, D. W. Inorg. Chem. 2006, 45, 8768–8775.
28. Stewart, D. J.; Napolitano, M. J.; Bakhmutova-Albert, E. V.; Margerum, D. W. Inorg.
Chem. 2008, 47, 1639–1647.
29. Fabian, I.; Gordon, G. Inorg. Chem. 1992, 31, 2144–2150.
30. Fabian, I.; Gordon, G. Inorg. Chem. 1991, 30, 3994–3999.
31. Orban, M.; Dateo, C.; De Kepper, P.; Epstein, I. R. J. Am. Chem. Soc. 1982, 104,
5911–5918.
32. Peintler, G.; Cseko, G.; Petz, A.; Horvath, A. K. Phys. Chem. Chem. Phys. 2010, 12,
2356–2364.
33. Horváth, A. K. J. Phys. Chem. A 2005, 109, 5124–5128.
34. Liu, Y.; Zhou, W.; Zheng, T.; Zhao, Y.; Gao, Q.; Pan, C.; Horváth, A. K. J. Phys.
Chem. A 2016, 120, 2514–2520.
35. Kormanyos, B.; Nagypal, I.; Peintler, G.; Horvath, A. K. Inorg. Chem. 2008, 47,
7914–7920.
36. Orban, M.; De Kepper, P.; Epstein, I. R. J. Phys. Chem. 1982, 86, 431–433.
37. Orban, M.; Epstein, I. R. J. Phys. Chem. 1982, 86, 3907–3910.
38. Maselko, J.; Epstein, I. R. J. Chem. Phys. 1984, 80, 3175–3178.
39. Nagypal, I.; Epstein, I. R. J. Phys. Chem. 1986, 90, 6285–6292.
40. Horváth, A. K.; Nagypál, I. J. Phys. Chem. A 1998, 102, 7267–7272.
41. Pan, C. W.; Stanbury, D. M. J. Phys. Chem. A 2014, 118, 6827–6831.
42. Gauffre, F.; Labrot, V.; Boissonade, J.; De Kepper, P.; Dulos, E. J. Phys. Chem. A 2003,
107, 4452–4456.
43. Horvath, D.; Toth, A. J. Chem. Phys. 1998, 108, 1447–1451.
44. Fuentes, M.; Kuperman, M. N.; De Kepper, P. J. Phys. Chem. A 2001, 105,
6769–6774.
45. Virányi, Z.; Horváth, D.; Tóth, Á. J. Phys. Chem. A 2006, 110, 3614–3618.
46. Huff Hartz, K. E.; Nicoson, J. S.; Wang, L.; Margerum, D. W. Inorg. Chem. 2003, 42,
78–87.
47. Rolia, E.; Chakrabarti, C. L. Environ. Sci. Technol. 1982, 16, 852–857.
48. Cseko, G.; Horvath, A. K. J. Phys. Chem. A 2012, 116, 2911–2919.
49. Horváth, A. K.; Nagypál, I.; Peintler, G.; Epstein, I. R. J. Am. Chem. Soc. 2004, 126,
6246–6247.
50. Horvath, A. K.; Nagypal, I. N. J. Phys. Chem. A 2006, 110, 4753–4758.
51. Nagypal, I.; Horvath, A. K. Chaos 2015, 25, 064604.
52. Slaughter, L. M.; Collman, J. P.; Eberspacher, T. A.; Brauman, J. I. Inorg. Chem. 2004,
43, 5198–5204.
53. Umile, T. P.; Wang, D.; Groves, J. T. Inorg. Chem. 2011, 50, 10353–10362.
54. Hu, Z.; Du, H.; Man, W.-L.; Leung, C.-F.; Liang, H.; Lau, T.-C. Chem. Commun.
2012, 48, 1102–1104.
55. Zdilla, M. J.; Lee, A. Q.; Abu-Omar, M. M. Angew. Chem. Int. Ed. 2008, 47,
7697–7700.
56. Zdilla, M. J.; Lee, A. Q.; Abu-Omar, M. M. Inorg. Chem. 2009, 48, 2260–2268.
57. Hicks, S. D.; Kim, D.; Xiong, S.; Medvedev, G. A.; Caruthers, J.; Hong, S.; Nam, W.;
Abu-Omar, M. M. J. Am. Chem. Soc. 2014, 136, 3680–3686.
58. Hicks, S. D.; Xiong, S.; Bougher, C. J.; Medvedev, G. A.; Caruthers, J.; Abu-Omar, M. M.
J.PorphyrinsPhthalocyanines2015,19,492–499.
59. Keith, J. M.; Abu-Omar, M. M.; Hall, M. B. Inorg. Chem. 2011, 50, 7928–7930.
60. Galajda, M.; Fodor, T.; Purgel, M.; Fabian, I. RSC Adv. 2015, 5, 10512–10520.
61. Kettle, A. J.; Winterbourn, C. C. Redox Rep. 1997, 3, 3–15.

62. Armesto, X. L.; Canle, M.; Garcia, M. V.; Santaballa, J. A. Chem. Soc. Rev. 1998, 27,
453–460.
63. Ingols, R. S.; Wyckoff, H. A.; Kethley, T. W.; Hodgden, H. W.; Fincher, E. L.;
Hildebrand, J. C.; Mandel, J. E. Ind. Eng. Chem. 1953, 45, 996–1000.
64. Weil, I.; Morris, J. C. J. Am. Chem. Soc. 1949, 71, 1664–1671.
65. Palin, A. T. A Study of the Chloro Derivates of Ammonia and Related Compounds, With
Special Reference to Their Formation in the Chlorination of Natural and Polluted Waters;
Vol. 54, University of London: London, 1949, 151–159, 189–200, 248–256.
66. Langheld, K. Chem. Ber. 1909, 42, 2360–2374.
67. Margerum, Dale W.; Gray, Edward T.; Huffman, Ronald P. In: Organometals and Orga-
nometalloids, Occurrence and Fate in the Environment; Brinckman, F. E., Bellama, J. M.,
Eds.; Vol. 82, American Chemical Society: Washington, DC, 1978; pp 278–291.
68. Armesto, X. L.; Canle, M.; Santaballa, J. A. Tetrahedron 1993, 49, 275–284.
69. Armesto, X. L.; Canle, M.; Garcia, M. V.; Losada, M.; Santaballa, J. A. Int. J. Chem.
Kinet. 1994, 26, 1135–1141.
70. Maitra, D.; Ali, I.; Abdulridha, R. M.; Shaeib, F.; Khan, S. N.; Saed, G. M.;
Pennathur, S.; Abu-Soud, H. M. Plos One 2014, 9, e110595. http://dx.doi.org/10.
1371/journal.pone.0110595.
71. Na, C.; Olson, T. M. Environ. Sci. Technol. 2007, 41, 3220–3225.
72. Nagy, P.; Ashby, M. T. Chem. Res. Toxicol. 2007, 20, 79–87.
73. Qiang, Z.; Adams, C. D. Environ. Sci. Technol. 2004, 38, 1435–1444.
74. Zgliczynski, J. M.; Stelmaszy nska, T.; Domanski, J.; Ostrowski, W. Biochim. Biophys.
Acta, Enzymol. 1971, 235, 419–424.
75. Fox, S. W.; Bullock, M. W. J. Am. Chem. Soc. 1951, 73, 2754–2755.
76. Dennis, W. H.; Hull, L. A.; Rosenblatt, D. H. J. Org. Chem. 1967, 32, 3783–3787.
77. Kaminski, J. J.; Bodor, N.; Higuchi, T. J. Pharm. Sci. 1976, 65, 553–557.
78. Hand, V. C.; Snyder, M. P.; Margerum, D. W. J. Am. Chem. Soc. 1983, 105,
4022–4025.
79. Szabó, M.; Baranyai, Z.; Somsák, L.; Fábián, I. Chem. Res. Toxicol. 2015, 28,
1282–1291.
80. Vogel, A. I. Quantitative Inorganic Analysis, 5th ed.; Longman: London, 1989.
81. Grob, C.; Baumann, W. Helv. Chim. Acta 1955, 38, 594–610.
82. Grob, C. A.; Schiess, P. W. Angew. Chem. Int. Ed. 1967, 6, 1–15.
83. Grob, C. A. Angew. Chem. Int. Ed. 1969, 8, 535–546.
84. Pera-Titus, M.; Garcıá-Molina, V.; Baños, M. A.; Gimenez, J.; Esplugas, S. Appl.
Catal. B 2004, 47, 219–256.
85. Pignatello, J. J.; Oliveros, E.; MacKay, A. Crit. Rev. Env. Sci. Technol. 2006, 36, 1–84.
86. Baxendale, J. H.; Wilson, J. A. Trans. Faraday Soc. 1957, 53, 344.
87. Greenwood, N. N.; Earnshaw, A. Chemistry of the Elements, 2nd ed.; Butterworth-
Heinemann: Oxford, UK, 1997; p 713.
88. Anderson, J. M.; Kochi, J. K. J. Am. Chem. Soc. 1970, 92, 1651–1659.
89. Steele, W. V.; Appelman, E. H. J. Chem. Thermodyn. 1982, 14, 337–344.
90. Epifano, F.; Marcotullio, M. C.; Curini, M. Trends Org. Chem. 2003, 10, 21–34.
91. Hussain, H.; Green, I. R.; Ahmed, I. Chem. Rev. 2013, 113, 3329–3371.
92. Ghanbari, F.; Moradi, M. Chem. Eng. J. 2017, 1(Pt. 1), 41–62.
93. Ball, D. L.; Edwards, J. O. J. Am. Chem. Soc. 1956, 78, 1125–1129.
94. Goodman, J. F.; Robson, P. J. Chem. Soc. 1963, 2871–2875.
95. Goodman, J. F.; Robson, P.; Wilson, E. R. Trans. Faraday Soc. 1962, 58,
1846–1851.
96. Evans, D. F.; Upton, M. W. J. Chem. Soc. Dalton Trans. 1985, 1151–1153.
97. Anipsitakis, G. P.; Dionysiou, D. D. Environ. Sci. Technol. 2004, 38, 3705–3712.
98. Pochtarenko, L.; Zilbermann, I.; Shamir, D.; Meyerstein, D. J. Coord. Chem. 2013, 66,
4355–4362.
99. Gilbert, B. C.; Stell, J. K. J. Chem. Soc., Perkin Trans. 2 1990, 1281–1288.
100. Gilbert, B. C.; Stell, J. K. J. Chem. Soc. Faraday Trans. 1990, 86, 3261–3266.
101. Lente, G.; Kalmár, J.; Baranyai, Z.; Kun, A.; Kek, I.; Bajusz, D.; Takács, M.; Veres, L.;
Fábián, I. Inorg. Chem. 2009, 48, 1763–1773.
102. Kalmár, J.; Lente, G.; Fábián, I. Inorg. Chem. 2013, 52, 2150–2156.
103. Betterton, E. A. Environ. Sci. Technol. 1992, 26, 527–532.
104. Beller, G.; Szabó, M.; Lente, G.; Fábián, I. J. Org. Chem. 2016, 81, 5345–5353.
105. Balogh, A.; Lente, G.; Kalmar, J.; Fabian, I. Int. J. Chem. Kinet. 2015, 47, 773–782.
106. Conrad, A. R.; Hassanin, H. A.; Tubergen, M. J.; Fabian, I.; Brasch, N. E. New J.
Chem. 2012, 36, 1408–1412.
107. Peintler, G. ZITA 5.0: A Comprehensive Program Package for Fitting Parameters of Chemical
Reaction Mechanisms; University of Szeged: Szeged, Hungary, 1999.
108. Thompson, R. C. Inorg. Chem. 1981, 20, 3745–3748.
109. Beller, G.; Bátki, G.; Lente, G.; Fábián, I. J. Coord. Chem. 2010, 63, 2586–2597.
110. Beller, G.; Lente, G.; Fábián, I. Inorg. Chem. 2010, 49, 3968–3970.
111. Beller, G.; Lente, G.; Fábián, I. Submitted to Inorg. Chem. 2017.
112. Kerezsi, I.; Lente, G.; Fabian, I. Dalton Trans. 2006, 955–960.
113. Melichercik, M.; Treindl, L. J. Phys. Chem. 1989, 93, 7652–7654.
114. Maerker, G.; Case, F. H. J. Am. Chem. Soc. 1958, 80, 2745–2748.
115. Gillard, R. D. Inorg. Chim. Acta 1981, 53, L173.
116. Rozen, S.; Dayan, S. Angew. Chem. Int. Ed. 1999, 38, 3471–3473.
117. Peintler, G.; Nagypál, I.; Jancsó, A.; Epstein, I. R.; Kustin, K. J. Phys. Chem. A 1997,
101, 8013–8020.
118. Kennedy, R. J.; Stock, A. M. J. Org. Chem. 1960, 25, 1901–1906.
119. Gallopo, A. R.; Edwards, J. O. J. Org. Chem. 1981, 46, 1684–1688.
120. McKay, S. E.; Sooter, J. A.; Bodige, S. G.; Blackstock, S. C. Heterocycl. Commun. 2001,
7, 307–312.
121. Mixan, C. E.; Pews, R. G. J. Org. Chem. 1977, 42, 1869–1871.
122. G€obel, M.; Karaghiosoff, K.; Klap€ otke, T. M.; Piercey, D. G.; Stierstorfer, J. J. Am.
Chem. Soc. 2010, 132, 17216–17226.
123. Agrawal, A.; Sailani, R.; Gupta, B.; Khandelwal, C. L.; Sharma, P. D. J. Korean Chem.
Soc. 2012, 56, 212–216.
124. Costas, M.; Chen, K.; Que, L., Jr. Coord. Chem. Rev. 2000, 200–202, 517–544.
125. Lente, G.; Fabian, I. Dalton Trans. 2007, 4268–4275.
126. Chow, T. W.-S.; Wong, E. L.-M.; Guo, Z.; Liu, Y.; Huang, J.-S.; Che, C.-M. J. Am.
Chem. Soc. 2010, 132, 13229–13239.
127. Tse, C.-W.; Chow, T. W.-S.; Guo, Z.; Lee, H. K.; Huang, J.-S.; Che, C.-M. Angew.
Chem. Int. Ed. 2014, 53, 798–803.
128. Liu, P.; Wong, E. L.-M.; Yuen, A. W.-H.; Che, C.-M. Org. Lett. 2008, 10,
3275–3278.
129. Lente, G.; Espenson, J. H. Int. J. Chem. Kinet. 2004, 36, 449–455.
130. Fabian, I.; Lente, G. Pure Appl. Chem. 2010, 82, 1957–1973.
131. Lente, G.; Espenson, J. H. Chem. Commun. 2003, 1162–1163.
132. Lente, G.; Espenson, J. H. J. Photochem. Photobiol. A 2004, 163, 249–258.
133. Brandt, C.; van Eldik, R. Chem. Rev. 1995, 95, 119–190.
134. Brandt, C.; Fabian, I.; van Eldik, R. Inorg. Chem. 1994, 33, 687–701.
135. Kerezsi, I.; Lente, G.; Fabian, I. J. Am. Chem. Soc. 2005, 127, 4785–4793.
136. Kerezsi, I.; Lente, G.; Fabian, I. Inorg. Chem. 2007, 46, 4230–4238.
137. Jortner, J.; Ottolenghi, M.; Stein, G. J. Phys. Chem. 1962, 66, 2042.
138. Jortner, J.; Ottolenghi, M.; Stein, G. J. Phys. Chem. 1964, 68, 247.
139. Huie, R. E.; Clifton, C. L. Chem. Phys. Lett. 1993, 205, 163–167.
140. Grossweiner, L. I.; Matheson, M. S. J. Phys. Chem. 1957, 61, 1089.
141. Shoute, L. C. T.; Alfassi, Z. B.; Neta, P.; Huie, R. E. J. Phys. Chem. 1991, 95,
3238–3242.
142. Galajda, M.; Lente, G.; Fabian, I. J. Am. Chem. Soc. 2007, 129, 7738–7739.
143. Gombar, M.; Jozsa, E.; Braun, M.; Osz, K. Photochem. Photobiol. Sci. 2012, 11,
1592–1595.
144. Ditroi, T.; Kalmar, J.; Pino-Chamorro, J. A.; Erdei, Z.; Lente, G.; Fabian, I. Photochem.
Photobiol. Sci. 2016, 15, 589–594.
145. Lazar, I.; Kalmar, J.; Peter, A.; Szilagyi, A.; Gyori, E.; Ditroi, T.; Fabian, I. Appl. Surf.
Sci. 2015, 356, 521–531.
146. Serpone, N.; Emeline, A. V. J. Phys. Chem. Lett. 2012, 3, 673–677.
147. Fujishima, A.; Rao, T. N.; Tryk, D. A. J. Photochem. Photobiol. C 2000, 1, 1–21.
148. Linsebigler, A. L.; Lu, G. Q.; Yates, J. T. Chem. Rev. 1995, 95, 735–758.
149. Brosillon, S.; Lhomme, L.; Vallet, C.; Bouzaza, A.; Wolbert, D. Appl. Catal. B. 2008,
78, 232–241.
150. Konstantinou, I. K.; Albanis, T. A. Appl. Catal. B 2004, 49, 1–14.
151. Veres, P.; Keri, M.; Bányai, I.; Lázár, I.; Fábián, I.; Domingoc, C.; Kalmár, J. Colloids
Surf. B Biointerfaces 2017, 152, 229–237.
152. Kalmar, J.; Keri, M.; Erdei, Z.; Banyai, I.; Lazar, I.; Lente, G.; Fabian, I. RSC Adv.
2015, 5, 107237–107246.
153. Petrov, O. V.; Furo, I. Prog. Nucl. Magn. Reson. Spectrosc. 2009, 54, 97–122.
154. Keri, M.; Peng, C.; Shi, X.; Banyai, I. J. Phys. Chem. B 2015, 119, 3312–3319.
155. Banyai, I.; Keri, M.; Nagy, Z.; Berka, M.; Balogh, L. P. Soft Matter 2013, 9, 1645–1655.
156. Kalmar, J.; Lente, G.; Fabian, I. Dyes Pigm. 2016, 127, 170–178.
157. Tsunoda, K. I.; Umemura, T.; Ueno, H.; Okuno, E.; Akaiwa, H. Appl. Spectrosc. 2003,
57, 1273–1277.
ARTICLE IN PRESS
O–O Bond Activation in Cu- and

Fe-Based Coordination
Complexes: Breaking It Makes
the Difference
Joan Serrano-Plana, Anna Company, Miquel Costas1
Grup de Quı́mica Bioinspirada, Supramolecular i Catàlisi (QBIS-CAT), Institut de Quı́mica Computacional i
Catàlisi (IQCC), Universitat de Girona, Girona, Catalonia, Spain
1
Corresponding author: e-mail address: miquel.costas@udg.edu
Contents
1. Introduction 64
1.1 O2 Activation: A Greener Alternative 64
1.2 Bioinspired Model Systems as a Common Strategy to Study Enzymes 65
2. Modeling Tyrosinase: O–O Cleavage in Dinuclear Copper Systems 67
2.1 Dicopper Model Complexes for O2 Activation 68
2.2 Model Complexes Exhibiting Tyrosinase-Like Activity. The Dilemma
of the Real Hydroxylation Agent 69
2.3 Upgrading the Challenge: Hydroxylation of Stronger C–F Bonds 72
3. O–O Cleavage in Iron-Oxygen Species 79
3.1 O–O Cleavage in Iron-Containing Enzymes 79
3.2 Mechanism of Action of Nonheme Iron Catalysts in Hydrocarbon Oxidation 83
3.3 Trapping Mononuclear Nonheme Iron-Oxygen Species Relevant to Iron
Oxygenases. O–O Cleavage in FeOOR Species: Accessing High-Valent
Compounds 84
3.4 Spectroscopically Characterized Oxoiron(V) Species 97
4. Summary 99
References 100
Abstract
Oxygenase enzymes catalyze the oxidation of hydrocarbons in a regio- and stereo-
selective manner using O2 as a sacrificial oxidant. These enzymes, which often contain
iron and/or copper in their active site, are the source of inspiration for the development
of novel methodologies to perform oxidative transformations in a more environ-
mentally friendly way, using benign oxidants such as O2 and H2O2. The ability of such
metals to attain different oxidation states is the key to facilitate the redox processes
associated with the formation and cleavage of the O–O bond. With the aim of gaining
insight into both the structure and reactivity of enzymes, bioinorganic chemists have
Advances in Inorganic Chemistry # 2017 Elsevier Inc. 63

ARTICLE IN PRESS
64 Joan Serrano-Plana et al.
devoted efforts to learn the intimate details of the processes associated with these reac-
tions. This is mostly carried out through the use of low molecular weight coordination
compounds. One of the key issues to develop further more effective bioinspired catalysts
is to understand the basis of their performance. In this vein, herein, we aim to depict
several examples for which the formation and cleavage of the oxygen–oxygen bond,
as well as the reactivity of the resulting species, have been studied through model
systems.
1. INTRODUCTION
1.1 O2 Activation: A Greener Alternative
Selective oxidations of organic molecules, especially hydrocarbons, are of
huge importance in industry (1–3). Millions of tons of alcohols, carbonyl
compounds, and epoxides are produced every year and used as reaction pre-
cursors in all areas of chemical industries. As a consequence of environmental
concerns, attention has been focused on the development of catalytic meth-
odologies (rather than stoichiometric) in order to minimize the cost of waste
disposal and avoid the use of harmful oxidants such as dichromate, perman-
ganate, and osmium tetroxide. Thus, the development of new procedures
that allows the performance of oxidation reactions under mild conditions
using green oxidants such as O2 and H2O2 is one of the biggest challenges
for modern synthetic organic chemistry (4).
The reaction of O2 with closed shell organic molecules is favorable from
the thermodynamic point of view (5). However, under ambient conditions
such oxidation reactions are very slow. This lack of reactivity results from
unfavorable kinetics: O2 has two unpaired electrons in its π antibonding
orbitals (S ¼ 1, triplet ground state), and this makes the reaction toward
organic substrates (generally S ¼ 0) spin forbidden. However, O2 can inter-
act with molecules containing unpaired electrons such as radicals (S ¼ 1/2)
and transition metal centers. Indeed, aerobic organisms have evolved to take
advantage of metal–oxygen interactions with the final aim of using their oxi-
dizing power. Interestingly, such reactions take place in a controlled fashion
in the active site of various enzymes (mainly oxidases and oxygenases).
Copper- and iron-based (heme and nonheme) proteins are commonly
found in processes where O2 activation takes place. Especially relevant
for the purpose of this review is their activity as mono- or dioxygenases; thus
the incorporation of one or two oxygen atoms from O2 into an organic sub-
strate. Upon O2 binding to the metallic center, reduced species that are more
ARTICLE IN PRESS
O–O Bond Activation in Cu- and Fe-Based Coordination Complexes 65
(Porph)FeII Mn+¯¯¯¯¯Mn+ Mn+
O2 O2 O2
III • n+ (n+1)+ •
Superoxo (Porph)Fe –OO M M
¯¯¯¯¯ –OO M(n+1)+–OO•
H+, e– H+, e–
a b 2 e−
(From cofactor)
Peroxo (Porph)FeIII–OOH M(n+1)+–OO–M(n+1)+ M(n+1)+–OO(H)
G
O–O bond H+
cleavage −H2O
O
Oxo (Porph•+)FeIV = O M(n+2)+ M(n+2)+ FeV = O CuII–O• FeIV = O
O OH
Cytochrome P450 Mn+ = FeII particulate a Mn+ = FeII Rieske oxygenases
heme peroxidases methane monooxygenase Mn+ = CuI monocopper
catalases Mn+ = CuI tyrosinase hydroxylases and amine oxidases
b Mn+ = FeII pterin or α-ketoacid-dependent

oxygenases
Scheme 1 Mechanisms of O2 activation by heme, nonheme, and copper-based

metalloenzymes. Adapted from Que, L.; Tolman, W. B. Nature 2008, 455, 333.
prone to react with organic substrates such as superoxide (O2 , S ¼ 1/2) or

peroxide (O2 2 , S ¼ 0) are formed (Scheme 1). In some cases, these species
can evolve to form high-valent species bearing oxometal moieties after
cleavage of the O–O bond (Scheme 1). Overall, unraveling the structure
and learning the intimate details of the reactivity of such molecules consti-
tute very exciting challenges for chemists, and at the same time, they rep-
resent the first step toward the development of sustainable catalytic
oxidation processes.
1.2 Bioinspired Model Systems as a Common Strategy to

Study Enzymes
The direct study of metalloproteins is often a challenge due to their very high
molecular complexity and the requirements of their isolation (often labori-
ous) and purification. In order to overcome these difficulties, much progress
has been made toward the development of bioinspired synthetic model sys-
tems. These are coordination complexes of much lower molecular weight,
composed of an organic ligand (synthesized through conventional organic
chemistry) that is able to bind metals resembling the chemical architecture
of the enzyme’s active site (Fig. 1). By tuning the nature of the ligand (donor
atom type, redox potential of the metals, coordination geometry), a wide
variety of compounds can be obtained and then examined spectroscopically
and subjected to reactivity studies. More specifically, biomimetic chemistry
ARTICLE IN PRESS
Fig. 1 Development of synthetic models to reproduce the structure and reactivity of

enzymes.
focuses on the study and development of synthetic systems that imitate the
formation, function, or structure of biologically produced substances and
materials and biological mechanisms and processes. In a long-term perspec-
tive, the preparation of compounds that reproduce enzyme functions could
provide new reagents or catalysts for practical applications. In this context,
O2 binding and activation in model systems containing transition metals
have been extensively explored, not only for the biological relevance of
this reaction but also for the potential industrial interest in performing selec-
tive oxidation catalysis in a cheaper and more environmentally friendly
manner.
Thus, the purpose of synthetic model chemistry through the synthesis of
bioinspired synthetic models is twofold: on one hand, mimicking the function
of an enzyme involved in a relevant chemical transformation and, on the
other hand, providing mechanistic and structural insight into the nature
of the biological system. The improvement and development of several
spectroscopic techniques such as electron paramagnetic resonance (EPR),
M€ ossbauer, Raman, X-ray absorption spectroscopy, and mass spectrometry
have greatly aided the trapping and characterization of several metal-oxygen
species. In some cases, these are too short lived or do not accumulate suffi-
ciently to be trapped and characterized. If this is the case, clues about the
nature of these species need to be acquired through indirect methods, such
as the study of their reactivity using specific substrates (6) and/or computa-
tional methods.
ARTICLE IN PRESS
Within the topic of synthetic model chemistry, in this chapter we

will discuss the synthesis of relevant examples of copper and iron-based
(per)oxo compounds. Special emphasis will be directed toward the descrip-
tion of systems where O–O bond cleavage of the peroxo moiety has been
observed, thus producing high-valent compounds. Indeed, acquiring the
intimate details of such reactions has led to the discovery of unprecedented
reactivities.
2. MODELING TYROSINASE: O–O CLEAVAGE IN

DINUCLEAR COPPER SYSTEMS
Tyrosinase is a copper-containing enzyme found in many living
organisms, involved in wound healing processes, immune response, or
browning processes of skin, hair, and fruit (7). More specifically, tyrosinase
catalyzes the aromatic ortho-hydroxylation and subsequent two-electron
oxidation of tyrosine (which contains a phenol moiety) to dopaquinone,
which constitutes the first step of melanin biosynthesis. Apart from the bio-
logical relevance of this reaction, it is also interesting from a chemical point
of view: the reaction is highly ortho-regioselective, a challenging goal in non-
enzymatic synthetic methodologies. Moreover, this reaction is also unique
in the sense that a full catalytic cycle involves the 4e oxidation of the sub-
strate, thus fully reducing the O2 molecule.
Many efforts have been directed during the last decades toward
unraveling the structure of the active site of tyrosinase and the oxidation state
of the metallic centers. Conclusive X-ray characterization was provided by
Matoba and coworkers in 2006, and it was confirmed that this enzyme con-
tains two copper centers, each of them coordinated by three histidine res-
idues (8,9). The mechanism of action for phenol hydroxylation by
tyrosinase is currently quite well accepted (10–16). In the absence of oxygen
(reduced or deoxy form), both metallic centers are CuI in a distorted trigonal
planar coordination geometry and they are separated by 4.6 Å. Reaction
with O2 affords the formation of the oxy form, which consists of a
μ-η2:η2-CuII2O2 species (side-on peroxo, SP) in which each copper center
is oxidized to CuII and the oxygen molecule is reduced by two electrons to
its peroxide form, and it is bound as a bridging ligand between the two
copper centers. The X-ray structure of tyrosinase shows that at this point
the structure is more contracted, with the two CuII centers disposed at
3.6 Å, each one five-coordinated in a distorted square-pyramidal geometry
(Scheme 2A). In the subsequent step, a phenolic substrate coordinates to
ARTICLE IN PRESS
A B O
O
H2O NHis NHis
CuI CuI
R NHis NHis
NHis NHis
O2
H+
R
NHis O NHis
CuII CuII
O NHis NHis
N O N NHis O NHis
CuII CuII s
P
N O N
N H N
R H+ OH
O R
N O N
CuII CuII
N O N N
N
Scheme 2 (A) X-ray structure of the substrate-bound tyrosinase (1WX2 PDB) (8,17).
(B) Mechanism of action of tyrosinase toward the hydroxylation of a monophenolic
substrate.
the Cu2O2 core. This interaction promotes the rotation of the Cu2O2
moiety and orientates one of the oxygen atoms toward the arene ring.
Then the O–O bond is cleaved and the close proximity of the substrate
facilitates an electrophilic attack of the peroxo moiety on the aromatic
ring, so that ortho-hydroxylation occurs. In the last step, the deoxy form
is regenerated after release of the ortho-quinone product and a water mol-
ecule (Scheme 2B).
2.1 Dicopper Model Complexes for O2 Activation

Copper-dioxygen chemistry has been a topic of interest in bioinorganic
chemistry since the early 1980s (13,18,19). The use of low molecular weight
transition metal complexes has resulted in several remarkable milestones
regarding understanding the mechanisms of O2 activation. In this context,
Zubieta, Karlin, and coworkers reported the first crystal structure of a
Cu2O2 compound, in this case with the O2 moiety bound in a trans-μ-
1,2-peroxo (TP) fashion (Scheme 3A), obtained from the reaction of
oxygen with a copper(I) complex bearing a nitrogen-based ligand (20).
Especially noteworthy was the report by Kitajima and coworkers who in
1989 structurally characterized the first example of a side-on bound
peroxodicopper(II) complex (SP), crucial for determining the dioxygen-
binding mode in hemocyanin and tyrosinase (21).
Another benchmark work in this field was reported by Tolman and
coworkers in 1996 demonstrating that SP and O compounds can be in equi-
librium through the reversible formation/cleavage of the O–O bond
(Scheme 3B) (22). This observation was exceptional because the formation
ARTICLE IN PRESS
A B O O
O
LCu II II
Cu L LCuII CuIIL LCuIII CuIIIL
O O O
trans-μ-1,2-peroxo μ-η2:η2 -peroxo bis(μ-oxo)
(TP) (SP) (O)
Scheme 3 Schematic representation of some Cu2O2 compounds. (A) trans-μ-1,2-peroxo
species (TP) and (B) equilibrium between SP and O species.
of the O–O bond constitutes one of the key reactions taking place during
photosynthesis, and this study constituted the first example where the forma-
tion and cleavage of the O–O bond were experimentally linked. Interestingly,
the SP/O equilibrium is highly dependent on steric factors induced by the
ligand (23), but it is also influenced by other factors such as solvent (24), elec-
tronic effects on the ligand, and counterions. As a general trend, nonbulky
ligands favor the formation of O species (shorter Cu–Cu distance) (23,24).
Furthermore, this finding opened the debate about whether the real
active species in tyrosinase, truly responsible for the biological ortho-
hydroxylation of phenols, was SP or O. Through the use of model systems,
it has been demonstrated that O is also capable of performing this reaction
(10,17,25).
2.2 Model Complexes Exhibiting Tyrosinase-Like Activity.

The Dilemma of the Real Hydroxylation Agent
Several examples of copper-dioxygen complexes that are able to perform
ortho-hydroxylation of phenolates, thus mimicking tyrosinase activity, have
been described (10,17). Karlin and coworkers reported in 1984 the first
example of a SP species that was capable of hydroxylating an aromatic C–
H bond using a dicopper(I) complex in which the two copper centers were
connected through a meta-xylyl moiety (26). After this precedent, Casella
and coworkers made important progress in the field by designing a
dicopper(I) complex with a dinucleating ligand (MeL66, Fig. 2) containing
two 1-methylimidazole arms connected through a m-xylyl linker. Exposure
of the latter to an O2 atmosphere at cryogenic temperatures afforded the for-
mation of SP species. Interestingly, no aromatic hydroxylation of the ligand
was observed in this case. However, very remarkably, this intermediate spe-
cies reacted with 4-methoxyphenol affording the corresponding catechol
product (27,28). Labeling studies showed full incorporation of 18O from
18
O2 into the final product. Thus, this represented the first example where
ARTICLE IN PRESS
N N
N N N N N N N N
N O N O O
CuII CuII D N CuII CuII N D N N CuII CuII N N
O O D O
N N N N D N N N N
N N
[CuII2(O2)(MeL66)]2+ [CuII2(O2)(LPy2Bz)2]2+ [CuII2(O2)(lm)6]2+
N
N N N N N
H H N N
N N O N O
O N CuII
II II N O N CuIII CuIII CuII N
Cu Cu
O CuIII CuIII N O N N O
N N N N N
H H O N
N N N
[CuII2(O2)(DBED)2]2+ [CuIII2(O)2(m-XYLMeAN)]2+ [CuIII2(O)2(LAG)2]2+ [CuII2(O2)(m-XYLN3N4)]2+
Fig. 2 Reported Cu2O2 systems that act as functional tyrosinase models toward external
substrates.
a well-defined copper-dioxygen adduct was able to ortho-hydroxylate exter-

nal phenolates, reproducing the activity of tyrosinase.
Since then, several Cu2O2 binding modes have been reported to perform
the ortho-hydroxylation of phenolates (Fig. 2) (17). Itoh and coworkers pro-
vided more insight regarding the capability of SP species to exhibit
tyrosinase-like activity toward phenolates. [CuII2(O2)(LPy2Bz)2]2+, gener-
ated from the reaction of two [CuI(LPy2Bz)]+ monomers with O2 in acetone
at 94°C, reacted with different phenolates to afford good yields of the cat-
echol derivatives (29).
A SP compound was also generated by Stack and coworkers by reaction
of O2 and a bidentate mononuclear copper complex [CuI(DBED)]+ (Fig. 2)
at 120°C (30). Remarkably, upon addition of 2,4-di-tert-butylphenolate
the authors detected a change spectroscopically in the nature of the
Cu2O2 core from SP to O; hence O–O cleavage occurred. This new O
compound subsequently decayed following a first-order process. Analysis
of the oxidation products showed 30% of 3,5-di-tert-butylcatechol along
with another 30% of the corresponding quinone. Thus, this example revived
the dilemma whether the real hydroxylating agent is a SP species or if this
compound is in a left-lying equilibrium with the O species, which would be
the true reactive compound (30). In a more recent example, Stack and
coworkers reported the formation of a SP species by the strikingly simple
self-assembly of imidazole rings, copper(I) and oxygen at 125°C to form
[CuII2(O2)(Im)6]2+ (Fig. 2) (31). The structure of this model compound
closely reproduced the first coordination environment in tyrosinase, and
ARTICLE IN PRESS
thus the compound could carry out the ortho-hydroxylation of various phe-
nolates yielding the catechol as the major product.
More evidence about the capability of O species to hydroxylate pheno-
lates was reported by Company et al. (25). A dicopper complex was synthe-
sized using a dinucleating ligand scaffold (m-XYLMeAN; Fig. 2) that consisted
of a meta-xylyl moiety connecting two tridentate sites. Upon reaction with
O2 at 90°C the O species [CuIII2(O)2(m-XYLMeAN)]2+ was formed and
fully characterized by several spectroscopic techniques. Interestingly, upon
addition of the phenolate a new purple intermediate species was formed.
Trapping and further characterization revealed that it corresponded to the
phenolate bound to the Cu2O2 core prior to oxidation. In the same context,
based upon a closely related study, Stack, Herres-Pawlis, and coworkers
reported in 2009 the full characterization of another O species containing
an aminoguanidine type of ligand [CuIII2(O)2(LAG)2]2+ (Fig. 2) that was able
to ortho-hydroxylate 2,6-di-tert-butylphenolate in good yields (32).
Finally, the ortho-hydroxylation of phenols to catechols was also per-
formed by an unsymmetric [CuII2(O2)(m-XYLN3N4)]2+ TP species, in
which the two copper ions have different coordination environments
(Fig. 2). One of the copper ions is bound to a triamine site, while the other
contains a tetraamine ligand site. Strikingly, [CuII2(O2)(m-XYLN3N4)]2+
exhibited different reactivity patterns compared to the symmetric TP ana-
logue [CuII2(O2)(m-XYLN4N4)]2+, both generated from the corresponding
dinuclear copper(I) complex and O2 at cryogenic temperatures. Surpris-
ingly, only the unsymmetric TP compound was able to perform ortho-
hydroxylation of external phenolates to form the catechol product, a
reaction not observed before for any TP species (33). Mechanistic studies
revealed that the unsymmetric TP was acting as an electrophilic oxidant,
contrary to the typical nucleophilicity shown by TP species (1,18). This
difference in reactivity was attributed to the unsymmetric nature of the
Cu2O2 moiety forced by the unsymmetric ligand. Coordination of the phe-
nolate substrate to the tridentate site results from the unprecedented electro-
philic reactivity of the unsymmetric end-on trans-peroxido core. Instead, the
most common TP species contain copper ions bound to tetradentate ligands,
and the copper ion is coordinatively saturated, unsuitable for phenolate bind-
ing (33). Recently, Solomon, Karlin, and coworkers performed new density
functional theory (DFT) calculations and proposed that in this system, TP is
not the actual hydroxylating species of the reaction, but instead, it is in equi-
librium with an O species that performs the hydroxylation (34). The same
authors provided experimental evidence of an equilibrium between TP
ARTICLE IN PRESS
and O species in a related system (34). However, the Solomon and Karlin
proposal of a rapid TP and O equilibrium cannot account for the lack of elec-
trophilic reactivity observed against typical substrates for O type of species
such as phosphines and sulfides, suggesting that isomerization, if it is taking
place, occurs only after phenolate binding.
Finally, it is worth mentioning that inspired by the mechanistic studies
mentioned earlier, some investigations have demonstrated and proven the
potential viability of these copper-oxygen reactions in catalytic processes
with relevance in organic synthesis (35–37).
2.3 Upgrading the Challenge: Hydroxylation of Stronger

C–F Bonds
Carbon–fluorine bonds are often considered inert functionalities owing to
their high bond-dissociation energy of 130 kcal/mol, representing the stron-
gest single bond to carbon. The presence of fluorinated organic compounds
is widespread, representing up to 30% of agrochemical products and 20% of
pharmaceuticals (38). Because of their inertness, these compounds show
great thermal stability, enhanced lipophilicity, and can suppress metabolic
detoxification, thus increasing their in vivo residence time (39). However,
the large-scale production and application of these products have been
increasing, and currently, they are the subject of debate due to their potential
toxicity, global warming potential, ozone depletion, environmental persis-
tence, and bioaccumulation. For these reasons, finding a way of degrading
fluorinated organic compounds or even more interestingly, transforming
C–F bonds into more reactive functional groups, is a topic of current inter-
est. However, such transformation constitutes a big challenge for conven-
tional organic chemistry (40).
An examination of natural systems reveals that several enzymes are capable
of cleaving C–F bonds (41) from both aromatic (e.g., fluorophenols or
fluorobenzoates) and aliphatic substrates (e.g., fluoropyruvate, fluoroacetate).
Cleavage and hydroxylation of aromatic C–F bonds normally occur in aerobic
organisms through the mediation of flavin adenine dinucleotide (FAD)-
containing phenol hydroxylases (FAD), which convert 2-fluorophenols into
catechols (42,43). Cytochrome P450 and chloroperoxidase have been
reported to oxidatively dehalogenate 4-fluorophenols (44), and methane
monooxygenase can oxidize trifluoroethylene (45). In contrast, fluoropheno-
lates act as enzyme inhibitors for tyrosinase (46). In order to challenge the
oxidizing ability of synthetic Cu2O2 species toward bonds stronger than an
ARTICLE IN PRESS
N Cl N N O N
O2
N CuI CuI N N CuII CuII N
reductant N O N
N N
r.t., 24 h H
[CuI2(XYLCl)]2+
N Br N O2 N O N
I
Cu I I
Cu r.t., 16 h CuI Cu
N N N N
O
H
[CuI2(2,6-bpb-1-Br)]2+
Scheme 4 Intramolecular oxidative dehalogenation mediated by copper(I) complexes
and O2.
aromatic C–H, we recently studied the ability of some Cu2O2 compounds to

perform ortho-hydroxylation of 2-fluorophenols.
Related to our work, Karlin and coworkers reported in the early 1990s
the oxidative intramolecular dechlorination mediated by copper(I) com-
plexes in the presence of a reducing agent and O2 albeit in moderate yields
(Scheme 4, top) (47). Remarkably, the fluoro-substituted analogue only
afforded traces of defluorinated product. In a somewhat parallel study, 1 year
later Feringa and coworkers reported intramolecular arene debromination of
the ligand 2,6-bpb-1-Br by copper-dioxygen species, yielding 80% of the
hydroxylated ligand (Scheme 4, bottom) (48).
Interestingly, during the last couple of years a few reports regarding the
biomimetic functionalization of C–F bonds have been published. As shown
later, copper- and iron-oxygen species have been found competent for the
hydroxylation of such bonds.
As detailed earlier, Company et al. showed in 2008 that high-valent
bis(μ-oxo)dicopper(III) species are capable of performing the ortho-
hydroxylation of phenolates (25). With this precedent in mind, some years
later we aimed to develop further the reactivity of this species with pheno-
lates containing fluorine substituents in the ortho position, in order to fulfill
the first biomimetic example of a defluorination reaction (49).
The addition of 3 equiv. of sodium 2,6-difluorophenolate (NaDFP)
to the preformed [CuIII2(μ-O)2(m-XYLMeAN)]2+ caused drastic color
changes in the reaction mixture. The yellow color of the O species
(Fig. 3A) immediately disappeared changing to deep purple (Fig. 3B).
ARTICLE IN PRESS
A B
21 21
18 18 N N
N N
III O
15 III O 15 N Cu CuIII N
N Cu CuIII N O
O
e (mM−1/cm)
e (mM−1/cm)
O F
N N N N
12 12 F
2(μ-O)2(m-XYL
III MeAN 2+
[Cu )] 9
9 [CuIII2(μ-O)2(OC6H3F2)(m-XYLMeAN)]+
6 6
3 3
0 0
350 550 750 950 350 550 750 950
Wavelength (nm) Wavelength (nm)
Fig. 3 Visual color changes observed during the reaction of [CuI2(m-XYLMeAN)]2+ with O2
and sodium 2,6-difluorophenolate in acetone at 90ºC along with the UV–vis spectra
recorded at 90ºC and the schematic representation of the compounds. (A) Fully
formed [CuIII2(O)2(m-XYLMeAN)]2+. (B) Sudden color change to intense purple upon addi-
tion of sodium 2,6-difluorophenolate to [CuIII2(O)2(m-XYLMeAN)]2+ caused by the coordi-
nation of the phenolate moiety to one copper center.
Accordingly, UV–vis monitoring of this reaction showed instant bleaching

of the chromophore associated with O (λmax ¼ 413 nm) and formation of
new features at λmax ¼ 520 and λmax 400 nm upon addition of Na(DFP)
(Fig. 3). Resonance Raman (rRaman) studies performed in the previous
work on a sample obtained after the addition of p-chlorophenolate to
[CuIII2(O)2(m-XYLMeAN)]2+ led to the conclusion that this purple
species corresponded to phenolate-bound O species (25).
Analysis of the reaction mixture after acidic work up once the purple
species had vanished (t1/2 40 s at 90°C) indicated the production of
3-fluorocatechol in 21% yield (with respect to the Cu2O2 species) (49).
ARTICLE IN PRESS
Importantly, labeling experiments performed by generating

[CuIII2(O)2(m-XYLMeAN)]2+ with 18O2 afforded 3-fluorocatechol with
86% 18O-label, thus confirming that indeed the Cu2O2 core is responsible
for the ortho-hydroxylation–defluorination reaction (Scheme 5).
As mentioned earlier, there is still an open debate in the field of copper-
dioxygen chemistry about the nature of the active species in the phenol
ortho-hydroxylation reaction. Even though O species have only been
observed in model systems so far, it is well known that they are usually in
a nearly degenerate equilibrium with the corresponding SP isomers (50).
Indeed, both can promote ortho-hydroxylation of phenolates as observed
in model chemistry (17). Thus, in order to study the capacity of both species
to ortho-hydroxylate-defluorinate 2-fluorophenolates, two functional
models previously described for the ortho-hydroxylation of phenolates were
selected: [CuI(DBED)]+ and [CuI(LPy2Bz)]+ (29,30,51). Both mononuclear
copper(I) complexes react with oxygen at low temperatures to form SP spe-
cies [CuII2(O)2(DBED)]2+ and [CuII2(O)2(LPy2Bz)]2+, which are capable of
performing ortho-hydroxylation of phenolates (Scheme 5). However, the
addition of the substrate caused a different effect in both species. For the lat-
ter, direct decomposition of the chromophore was observed, and afterward
via O species
2+
N N
NaDFP (3 equiv.)
N CuIII O
CuIIIN
O
N N OH
CuII O CuII F OH
O
2+ O F
NH HN F
CuII O CuII
O NaDFP (3 equiv.)
N N
H H O–O bond cleavage
via P species
2+
N N CuII O CuII
II O II O
N Cu Cu N NaDFP (3 equiv.) OH
D O D O F F OH
D N N D
F
Scheme 5 Reaction of the different Cu2O2 compounds discussed herein toward NaDFP.
ARTICLE IN PRESS
the oxidation products were quantified (29). In contrast, in the case of

[CuII2(O)2(DBED)]2+ phenolate addition caused a change in the spectro-
scopic features of the Cu2O2 species, switching from SP species to O, which
was proposed to be finally responsible for the oxidation (30). When NaDFP
was used as a substrate, more differences were observed. [CuII2(O)2(LPy2Bz)]2+
afforded the recovery of the starting material (along with <3% of catechol
product). Interestingly, product analysis after the addition of 3 equiv.
NaDFP to [CuII2(O)2(DBED)]2+ revealed formation of 3-fluorocatechol
in 23% yield (with respect to Cu2O2), thus closely resembling the result
obtained with [CuII2(O)2(m-XYLMeAN)]2+ (49). All these observations
taken together suggested that the involvement of O species was key to
achieve the ortho-hydroxylation–defluorination reaction. Control experi-
ments performed with [CuI(CH3CN)4](CF3SO3) produced only trace
amounts of the defluorinated product.
Interestingly, the reaction was found to be regioselective for the
ortho position of the phenolate: only the production of 4-fluorocatechol
was observed upon addition of sodium 4-fluorophenolate to
[CuIII2(O)2(m-XYLMeAN)]2+ (Scheme 6A). At this point, the reaction
Scheme 6 Selective ortho-hydroxylation–defluorination upon reaction of [CuIII 2(O)2

(m-XYLMeAN)]2+ at 90°C with sodium 4-fluorophenolate (A) and ortho-substituted
sodium 2-fluorophenolates (B).
ARTICLE IN PRESS
toward unsymmetric phenolates bearing a fluorine atom and a different sub-

stituent in the two ortho positions was also studied. Workup after complete
decay of the compound revealed that the reaction was selective for cleaving
the C–F bond. Thus, reaction of [CuIII2(μ-O)2(m-XYLMeAN)]2+ with
3 equiv. of sodium 2-fluorophenolate afforded 31% of 1,2-
dihydroxybenzene together with 8% of 3-fluorocatechol (Scheme 6B).
Exquisite selectivity was observed for substrates bearing another halogen in
the ortho position, such as sodium 2-choro-6-fluorophenolate and
2-bromo-6-fluorophenolate. In this case, the ortho-hydroxylated–
defluorinated product was the only one observed in 62% and 49% yields,
respectively. Thus, ortho-hydroxylation–defluorination occurred even
though presumably weaker Carene–X bonds (R ¼ Cl, Br, CH3) were pre-
sent. Interestingly, only recovery of the starting material was observed
when phenolates bearing substituents different from hydrogen and fluorine
in the ortho position were used (such as sodium 2,6-dichlorophenolate and
sodium 2,6-dimethylphenolate). This ortho-hydroxylation–defluorination
reaction is unique in the sense that whereas oxidative dehalogenation of
Carene–X (X ¼ Cl, Br) had previously been described to occur by reaction
of dicopper(I) complexes and O2 (48), no examples of defluorination by
Cu2O2 species were known.
The next goal was an attempt to improve the reaction yield, which in
most cases remained moderate (Scheme 6). The observed ortho-hydroxyl-
ation–defluorination itself does not involve a neat gain or loss of electrons,
but experimentally, it was observed that the CuIII was reduced to CuII. We
speculated that the ligand could undergo oxidative degradation to provide
these reducing equivalents, but the ligand was recovered intact after reac-
tion of [CuIII2 (O)2(m-XYL
MeAN 2+
)] with NaDFP so that an alternative
explanation was needed. It was considered that the electrons could also
originate from a reaction with a second [CuIII 2 (O)2(m-XYL
MeAN 2+
)]
unit releasing O2, thus recovering the initial CuI complex and
releasing electrons. If this was the case, the addition of an external reduc-
ing agent would provide the necessary electrons and yields would increase.
Such strategy had already been used to enhance the dechlorination yield by
a SP species (47,48). Unfortunately, and as previously reported for other
O species, the addition of sodium ascorbate, zinc dust, or a copper(I) salt
caused the immediate decomposition of [CuIII 2 (O)2(m-XYL
MeAN 2+
)] .
II 2+
However, [Cu 2(O)2(DBED)] was more tolerant to the presence of sodium
ascorbate. Very remarkably, whereas the reaction of this compound with
ARTICLE IN PRESS
sodium 2-chloro-6-fluorophenolate and sodium 2-bromo-6-fluorophenolate

at 90°C afforded 41% and 34% of the respective catechols, the presence
of sodium ascorbate (2 equiv. with respect to [CuII2(O)2(DBED)]2+) raised
both yields up to 83%.
From all the accumulated data, it was proposed that the mechanism
occurs via an electrophilic attack of O on the aromatic ring, and that a proper
orientation of the C–F bond toward the copper(III) center after an interac-
tion with an unpaired pair of electrons of the ortho-fluorine could be the rea-
son for the high selectivity observed.
Overall, this work represented the first example of an ortho-hydroxyl-
ation–defluorination of 2-fluorophenolates using O2 as sacrificial oxidant
by means of a synthetic metal complex, a reaction that finds precedent
in FAD-dependent hydroxylase enzymes in biological systems. The
involvement of O species seems to be a key issue to perform such a reaction.
This reaction constitutes a fascinating transformation from a basic chemical
point of view not only because of the inherent interest of cleaving strong
C–F bonds but also due to the rather unusual selectivity properties it
exhibits. Finally, this reaction is also interesting from the perspective of
environmental science, as it proves that the transformation of inert C–F
bonds of common persistent pollutants into a more reactive C–OH unit
can be achieved.
Very recently, three examples of iron-based model compounds capa-
ble of mediating the cleavage of C–F bonds have been reported. The
first example of an iron-based complex capable of performing inter-
molecular aromatic C–F bond cleavage was reported by Sorokin and
coworkers (52). A μ-nitrido diiron phthalocyanine (Pc) complex [(Pc)
FeIII(μ-N)FeIV(Pc)]4+ reacted with poly- and perfluorinated aromatics
under mild conditions using H2O2 or tBuOOH as oxidant to give the
hydroxylated–defluorinated product. The reaction was proposed to
occur via the generation of a high-valent diiron phthalocyanine radical
cation complex with fluoride axial ligands. Goldberg and coworkers
reported the ability of a detectable nonheme pentadentate oxoiron(IV)
to cleave an aromatic C–F bond from the ligand framework through
an electrophilic mechanism (53). In addition, Agapie and coworkers
reported 1 year later an example of a tetranuclear iron complex that could
intramolecularly cleave an aromatic C–F bond of the ligand (54), even
though in this case the high-valent intermediate species could not be
trapped.
ARTICLE IN PRESS
3. O–O CLEAVAGE IN IRON-OXYGEN SPECIES

3.1 O–O Cleavage in Iron-Containing Enzymes
Iron-containing metalloenzymes are found throughout the natural world
and participate in vital catalytic oxidative processes and biosynthetic biolog-
ical pathways involving O2 activation (5,55–57). As a result of the efforts of
several research groups over the last decades, the structure of the active site of
O2-activating iron-containing metalloenzymes has been elucidated. Fur-
thermore, in some cases, the active iron-oxygen species have been trapped
and spectroscopically characterized. According to these studies, iron
oxygenases can be classified into three main groups depending on the struc-
ture of the active site: heme (e.g., cytochrome P450), mononuclear non-
heme (e.g., taurine dioxygenase, TauD), and dinuclear nonheme
oxygenases (e.g., soluble methane monooxygenase, sMMO (58)) (Fig. 4).
Interestingly, most of the accumulated knowledge of iron enzymes involved
in dioxygen activation originates from the pioneering studies on heme
enzymes, in which the iron center is coordinated to a porphyrin cavity. Over
the last 20 years, much effort has been directed toward the study of nonheme
systems and it has been found that remarkable similarities exist in the reaction
mechanism of heme and nonheme systems (59–61).
3.1.1 O–O Cleavage in Cytochrome P450: Generation of CpdI

One of the most representative examples of heme enzymes is the cyto-
chrome P450 family (cyt-P450), which is present in nearly all living systems.
Cyt-P450s reductively activate O2 to generate species that can oxidize a
wide range of organic substrates. Their active site contains an iron center
Heme Nonheme
O Glu O
Glu O
N N O Asp FeIV O
Fe +IV Glu FeIV FeIV COO–
N N O His His
His O O His
S
Cys
Cytochrome P450 Soluble methane monooxygenase Taurine dioxygenase
(Compound I) (sMMO-Q) (TauD-J)
Fig. 4 Representative examples of spectroscopically characterized heme and nonheme
iron-oxygen species in enzymes.
ARTICLE IN PRESS
in an octahedral geometry bound to a porphyrin unit in the equatorial plane

with axial positions occupied by a water (or hydroxide) molecule and a
cysteinate ligand, which connects the active site to the protein backbone
(62). Cyt-P450s have been studied since the 1950s, and nowadays their
activity and catalytic cycle are quite well established (Scheme 7A)
(62–64). Starting from the hexacoordinate low-spin iron(III) compound,
binding of a substrate in the surroundings of the active site causes the disso-
ciation of the water molecule leaving a high-spin five-coordinate iron(III)
complex which is then reduced by one electron to a FeII center (65). At this
point, the metal binds dioxygen, leading to the formation of a superoxo unit,
which evolves to hydroperoxoiron(III) after a second one-electron reduc-
tion and protonation. Then, a proton assists the heterolytic O–O bond
cleavage to release a water molecule and form a highly reactive
oxoiron(IV) cation radical (Porph•+)FeIV(O), commonly referred to as
Compound I (CpdI, Fig. 4), which is proposed to be the real active species.
At this stage, oxidation of the substrate occurs through the so-called rebound
mechanism first proposed by Groves (66,67). This is a two-step process that
starts with the abstraction of the hydrogen atom of the substrate by the oxo
moiety so that an alkyl radical (short lived) is formed. This will then rapidly
react with the Fe–OH moiety to afford the final product (rebound, Scheme
7B). Release of the oxidized organic substrate regenerates the initial
iron(III). Even though CpdI was for a long time proposed as the final
substrate-oxidizing species, it was not until recently that Green and
coworkers could trap and characterize spectroscopically this species in a
cyt-P450 mutant (68).
A B
R-OH OH2 H C
O
R-H
FeIII
FeIV∑
H2O
N N
Fe H2O Hydrogen atom
O R-H
N N abstraction
Fe ∑ IV
Peroxide shunt R-H
S FeIII OH ∑C
Cys Cpd I NaOCl, ROOH,
H2O RCO3H, PhIO
e – FeIV
H+
HO R-H
Fe O Oxygen rebound
R-H FeII
FeIII
FeIII HO C
O2
– +
e ,H
Scheme 7 (A) Proposed catalytic cycle for cyt-P450s. (B) Postulated oxygen rebound
mechanism for C–H hydroxylation in cyt-P450s.
ARTICLE IN PRESS
3.1.2 Rieske Oxygenases

In contrast to cyt-P450s, Rieske oxygenases are less understood. The ligand
environment for these enzymes is more flexible compared to heme systems,
thus allowing a more diverse chemistry. Indeed, they are involved in a wide
range of reactions including sulfoxidation, desaturations, oxidative car-
bocyclizations (69), or O- and N-dealkylation (70–72). Even more interest-
ingly, they catalyze the selective C–H hydroxylation and stereo- and
enantioselective cis-dihydroxylation of arenes, a reaction that is difficult to
achieve through conventional organic synthesis and not observed for any
other enzyme class (73,74). Recently, it has been proposed that Rieske
oxygenases are involved in cholesterol catabolism and also in antibiotic bio-
synthetic pathways (75,76).
Naphthalene 1,2-dioxygenase (NDO) is probably the most studied
enzyme in this family, and its structure is representative for other enzymes
such as benzoate 1,2-dioxygenase, toluene 2,3-dioxygenase, benzene
dioxygenase, phthalate dioxygenase, and dicamba monooxygenase. X-ray
characterization of NDO isolated from Pseudomonas putida revealed that dif-
ferent components can be distinguished in its active site (Scheme 8A) (77).
The first one is the oxygenase component. It consists of a mononuclear non-
heme iron(II) center, which, in its resting state, is bound to two histidine
residues and one aspartate occupying the same face of the octahedron
(77,78). This structural feature is known as the 2-His-1-carboxylate facial
triad and it is characteristic of this class of enzymes (79). It leaves three coor-
dination sites on the metal available for interaction with exogenous mole-
cules such as substrates and O2. The coordination sphere of the iron in
the resting state is completed by three water molecules. The other compo-
nents of the active site are a Fe2S2 Rieske-type cluster that acts as a reductase
and mediates electron transfer from NAD(P)H (reduced form of nicotin-
amide adenine dinucleotide phosphate) to the oxygenase site, and an aspar-
tate residue that binds the oxygenase and reductase components (75).
Despite considerable efforts that have been directed toward gaining
structural and spectroscopic insight into the catalytic cycle of Rieske
oxygenases, not all reaction steps have been fully elucidated. Ramaswamy
and coworkers shed some light in 2003 by solving a valuable set of crystal
structures of NDO at different stages of its catalytic cycle: substrate bound
in the active site cavity, oxygen coordinated to the metal center (Scheme
8B), and the final oxidized product ligated to the iron (80). From these stud-
ies, a consensus has been reached about the mechanism of action of these
enzymes (Scheme 8C). In a first step, coordination of the substrate results
A Reductase H Oxygenase C
N
Cys Cys His OH2
FeIII N O
FeII Asp
S S O His OH2
His N FeII
Fe II
His OH2 His O2, H+, e–
N O
His N His OH2 O
FeII H2O
ARTICLE IN PRESS
O Asp H 2O
N O His O O
O His OH
Asp FeIII
B Asp
His O
HO via FeIII (OOH) O
HO
Asp
H+, e–, 2 H2O
O
His O
FeIII O–O cleavage
His O
O His OH via FeV (O)(OH)
O FeV
Asp His O
O
Asp
Scheme 8 (A) Schematic representation of the active site of naphthalene 1,2-dioxygenase (NDO). (B) X-ray structure of O2-bound oxygenase
component of NDO. (C) Proposed catalytic cycle of naphthalene 1,2-dioxygenase (NDO). Panel (B): Adapted from Karlsson, A.; Parales, J. V.;
Parales, R. E.; Gibson, D. T.; Eklund, H.; Ramaswamy, S. Science 2003, 299, 1039.
ARTICLE IN PRESS
in the loss of a water molecule leading to a five-coordinated ferrous com-

plex. O2 then binds to the metal and it is reduced to the peroxo state upon
protonation to yield a hydroperoxoiron(III) species, as determined by X-ray
analysis (73). Another crystallographically characterized example of
hydroperoxoiron(III) species has recently been obtained for another
Rieske-dependent oxygenase (carbazole 1,9a-dioxygenase) (81). This
FeIII(OOH) species is the last detectable species before substrate oxidation
occurs. At this point, there is an ongoing debate. This species either directly
attacks the arene double bond of the substrate or undergoes O–O cleavage,
thus isomerizing to a very reactive high-valent FeV(O)(OH) that would be
the real oxidant of the reaction. Even though a number of experimental and
theoretical studies have been carried out, a discussion about whether
FeIII(OOH) is directly responsible for substrate oxidation, or not, is still
ongoing (61,82,83). DFT calculations by Decker et al. (84) and Neese
et al. (85) showed that FeIII(OOH) could be competent for the oxidation,
while Kumar et al. (86) proposed that it is merely a precursor of the real
active species. However, direct evidence for the high-valent electromer
has not been obtained in natural systems so far. In a final step, the oxidized
substrate is uncoupled from the ferric center by protonation, regenerating
the initial ferrous resting state.
3.2 Mechanism of Action of Nonheme Iron Catalysts in

Hydrocarbon Oxidation
In a biomimetic spirit, numerous iron-based complexes have been prepared
that elicit olefin oxidation and site-selective and stereoretentive C–H bond
oxidation upon reaction with H2O2 (87–90). The selectivity and stereo-
specificity exhibited by these catalysts provide strong support in favor of
the intermediacy of metal-based oxidants akin to those involved in iron-
dependent oxygenases. Presumably because of their high reactivity, these
oxidants do not accumulate in solution, and it is challenging to establish their
nature. However, there are some examples where metastable oxo/
peroxoiron species have been trapped and their formation and reactivity
have been investigated. In this section we review some peroxoiron species
and their ability to undergo O–O cleavage to give oxoiron compounds.
Basically two mechanisms have been proposed for catalytic hydrocarbon
oxidation by synthetic nonheme iron compounds: water- and acid-assisted.
In both cases, the oxidant promotes the initial oxidation of the iron(II) pre-
cursor to iron(III)(OOR) (R ¼ H, alkyl, acyl) species. This species then
undergoes O–O cleavage to form a high-valent species, namely an
ARTICLE IN PRESS
Scheme 9 Water- and carboxylic acid-assisted mechanisms for hydrocarbon (alkanes

or alkenes) oxidation by mononuclear nonheme iron enzymes. LN4, nitrogen-based
tetradentate ligand with cis-positions available.
oxoiron(V) species, which is responsible for substrate oxidation (Scheme 9).

This cleavage can be aided by either water (water assisted) or a carboxylic
acid (acid assisted) (19,91).
Evidence for this mechanistic proposal has been gathered through the use
of specific substrates that provide certain information about the catalytic sys-
tem, the so-called mechanistic probes (6). Alternatively, despite being rela-
tively poor catalysts, iron complexes that stabilize the reactive species formed
along the catalytic cycle of nonheme iron oxygenases have provided
extremely valuable information.
3.3 Trapping Mononuclear Nonheme Iron-Oxygen Species

Relevant to Iron Oxygenases. O–O Cleavage in FeOOR
Species: Accessing High-Valent Compounds
Over the last two decades the use of cryogenic methods has yielded signif-
icant advances toward the entrapment and characterization of novel iron-
oxygen species. This has made a valuable contribution toward unraveling
the nature of the reactive species involved in oxidation reactions
(19,56,61,87,92–94). In this context, several mononuclear nonheme
iron-oxygen species with various oxidation states on the iron (and different
degrees of reduction of the oxidant) have been detected. Their relevance in
oxidation reactions has been interrogated, and significant analogies between
these compounds and iron oxygenases have also been established.
Formation of superoxoiron species is proposed to be the first step in the
activation of O2 by iron oxygenases. However, these species are generally
ARTICLE IN PRESS
regarded as pass-through points toward the generation of more oxidized

metal species. In contrast to heme systems (65), their characterization in
nonheme enzymes has been scarce (56,95,96). Similarly, very few synthetic
models of nonheme superoxoiron species are available (97–99). This initial
event is often followed by protonation, which results in the formation of
hydroperoxoiron(III) species (which have been directly detected in non-
heme oxygenases, see Section 3.1). Indeed, bioinspired synthetic chemistry
has produced a number of nonheme peroxoiron(III) compounds,
FeIII(OOR) (R ¼ H, alkyl, acyl) (92–94). In terms of reactivity, low-spin
FeIII(OOR) species are usually described as sluggish oxidants that are not
competent for the oxidation of strong C–H bonds (100–102). However,
these species can undergo homolytic or heterolytic O–O cleavage to form
high-valent species with greater oxidizing power. In the following section,
the O–O cleavage in peroxoiron(III) species to form such compounds will
be discussed.
3.3.1 Hydroperoxoiron(III) Compounds

Several examples of synthetic nonheme hydroperoxoiron(III) compounds,
FeIII(OOH), have been obtained by reaction of iron(II) complexes with
excess H2O2 at low temperatures (103–105). In very few examples, their
generation was also achieved by using O2 in the presence of a reducing agent
(106,107). Most synthetic nonheme hydroperoxoiron(III) compounds
reported so far contain N-based pentadentate ligands (104,108), but there
are also a few cases in which the ligand employed is tetradentate
(103,105,109–111). Typically, low-spin FeIII(OOH) compounds are deep
purple with a characteristic UV–vis band at λmax 550 nm corresponding to
the hydroperoxo-to-iron charge transfer transition. Excitation at 550 nm
in resonance Raman experiments gives resonance-enhanced features at
800 cm1 (downshifted 20 cm1 when 18O is used) and 600 cm1
(downshifted 50 cm1), which are assigned to O–O and Fe–O stretching
modes, respectively.
In selected examples, reaction of FeIII(OOH) with a base results in its
deprotonation yielding the conjugate base FeIII(η2-O2) in which the peroxo
moiety binds in a side-on fashion. As a representative example, [FeIII(OOH)
(N4Py)]2+ reacted with NEt3 or NH3 to form [FeIII(η2-O2)(N4Py)]+,
which presented different spectroscopic features (Scheme 10A) (100,112).
A similar deprotonation event was observed for [FeIII(OOH)(TMC)]2+.
This compound was first described by Que and coworkers in 2011 and rep-
resented the first example of a high-spin hydroperoxo species (111). By
ARTICLE IN PRESS
A B
HO O
O N O N
N N
Fe III Base FeIII
N N N Acid N N N
[FeIII(h1-OOH)(N4Py)]2+ [FeIII(h2-O2)(N4Py)]+ [FeIII(h2-O2)(TMC)]+

Low spin (S = 1/2) High spin (S = 5/2) High spin (S = 5/2)
UV–vis lmax = 548 nm UV–vis lmax = 685 nm O–O 1.46 Å
( e = 520 M−1cm−1) Fe–O 1.91 Å
( e = 1300 M−1cm−1)
rRaman: ν(O−O) = 790 cm−1 rRaman: ν(O−O) = 827 cm−1 UV–vis lmax = 750 nm
ν(Fe−O) = 632 cm−1 ν(Fe−O) = 495 cm−1 ( e = 600 M−1cm−1)
rRaman: ν(O−O) = 825 cm−1
ν(Fe−O) = 487 cm−1
Scheme 10 (A) Reversible acid–base interconversion between hydroperoxoiron(III) and

peroxoiron(III) using the ligand N4Py. (B) X-ray structure and spectroscopic features of
[FeIII(η2-O2)(TMC)]+ reported by Nam and coworkers (100).
addition of protons, this compound could be quantitatively converted to an

oxoiron(IV) species. The same year Nam and coworkers were able to crys-
tallize its conjugate base [FeIII(η2-O2)(TMC)]2+ after reaction of the iron(II)
precursor with H2O2 (5 equiv.) in the presence of triethylamine (2 equiv.) at
0°C (Scheme 10B) (100). Some differences in reactivity were observed for
the hydroperoxoiron(III) species and its conjugate base: while the
deprotonated [FeIII(η2-O2)(TMC)]2+ behaved as a sluggish reagent in
nucleophilic aldehyde deformylation and toward alkanes, the
corresponding hydroperoxo [FeIII(OOH)(TMC)]2+ species could perform
both nucleophilic deformylation of aldehydes and electrophilic oxidation
of weak C–H bonds such as xanthene or 9,10-dihydroanthracene. Indeed,
strong evidence exists that the spin state of FeIII(OOH) affects its reactivity.
Nam, Shaik, and coworkers reported that [FeIII(OOH)(TMC)]2+ was
competent to perform oxygen atom transfer reactivity toward sulfides
(113) and hydrogen atom abstraction from substrates with weak C–H bonds,
such as xanthene and 9,10-dihydroanthracene, in contrast to low-spin ana-
logues (114).
3.3.1.1 O–O Cleavage in Hydroperoxoiron(III) Compounds

The O–O cleavage in a hydroperoxoiron(III) moiety can lead to the forma-
tion of an oxoiron compound, either in oxidation state +4 after homolytic
cleavage or +5 in the case of a heterolytic cleavage, even though the spec-
troscopic characterization of the latter remains to be accomplished.
ARTICLE IN PRESS
Que and coworkers found evidence for the water-assisted heterolytic

cleavage in the [FeIII(OOH)(TPA)]2+ system (103). The ferric hydroperoxo
compound, generated at 40°C with the iron(II) precursor and excess
hydrogen peroxide, was not kinetically competent for the oxidation of
1-octene (the rate of its decay is not proportional to substrate concentration).
Instead, upon generation, it remained under a pseudosteady-state phase until
H2O2 had been consumed. However, during the lag phase linear formation
of epoxide and cis-diol products from 1-octene was observed. Product for-
mation ceased upon decay of the chromophore. This is an indication that,
even though it is not directly responsible for the activity, the ferric hydro-
peroxo species is the precursor of the real oxidizing species. Interestingly,
increasing amounts of H2O affected the steady-state duration and concom-
itantly increased the rates of product formation as a result of an increase of
the O–O cleavage rate. Relevant and related reports by Rybak-Akimova
(115) and Banse (82) suggested that low-spin FeIII(OOH) compounds
are not directly involved in substrate oxidation, as they did not observe a
correlation between the decay rate of the ferric hydroperoxo and substrate
concentration.
3.3.2 Acylperoxoiron(III) Compounds

Particularly interesting examples of iron peroxide compounds are
acylperoxoiron(III) species. These compounds are typically generated by
reaction of an iron(II) precursor and H2O2 in the presence of acetic acid
or by direct reaction with alkyl peracids. Again, the importance of these
acylperoxo compounds resides in their ability to undergo O–O cleavage
so that a high-valent electromer is reached. Indeed, acylperoxoiron(III) spe-
cies represents the precursor toward the generation of FeV(O)(OR)
(R ¼ alkyl) that has been proposed to be the active species in the catalytic
cycle of hydrocarbon oxidation by nonheme iron catalysts using H2O2 as
oxidant in the presence of carboxylic acids.
Very few examples of acylperoxoiron(III) compounds have been char-
acterized and, except in one recently reported case (116), these species are
kinetically not competent for the oxidation of organic substrates. Suzuki and
coworkers reported in 2005 the first crystallographically characterized non-
heme peracetatoiron(III) complex [FeIII(CH3CO3)(qn)2]2+ (Fig. 5A),
obtained after bubbling CO2 through a solution of the iron(III) precursor
in DMF at 60°C (117). Raman experiments using 18O2 or C18O2 showed
that indeed the peracetato compound can reversibly exchange oxygen
ARTICLE IN PRESS
A
2+
O
N
O O
FeIII O
O O
N
O
[FeIII(O2COCH3)(qn)2]2+
B R′ 2+
R R
2+
N
N N O O
N O FeIII
Fe III N O
O
O O N
O N
R R
R′
[FeIII(O2COCH3)(6Me2-BPP)]2+ R = R′ = H [FeIII(O2COCH3)(BPMEN)]2+
R = Me, R′ = OMe [FeIII(O2COCH3)(BPMEN*)]2+
R′ 2+
R R
2+
tBu
N N
N O O N O O
R FeIII FeIII
N O N O
R′ N R′′ N
R tBu
R R
R′
R = R′ = H, R′′ = C6H4Cl [FeIII(O2COC6H4Cl)(TPA)]2+ [FeIII(O2COCH3)(tBuN4)]2+

R = R′ = H, R′ = CH3 [FeIII(O2COCH3)(TPA)]2+
R = Me, R′ = OMe, R′′ = CH3 [FeIII(O2COCH3)(TPA*)]2+
R′ 2+
2+
R R O
R
N
N O O
O O
Fe III N FeIII N
N O N
N N
R R
R′
R = R′ = H [FeIII(O2COCH3)(PDP)]2+ [FeIII(O2COR)(13-TMC)]2+
R = Me, R′ = OMe [FeIII(O2COCH3)(PDP*)]2+
Fig. 5 See legend on opposite page.
ARTICLE IN PRESS
atoms, likely occurring through the formation of a side-on peroxoiron(III)

species [FeIII(O2)(qn)2].
Some years later, the combination of [Fe(6Me2-BPP)(H2O)]+ with
peracetic acid allowed the characterization of another example of a per-
acetatoiron compound that could also be crystallographically characterized
(Fig. 5B). This compound was able to epoxidize cyclooctene and to
oxidize weak C–H bonds such as those of adamantane, albeit in very low
yields (118).
As a result of the instability of acylperoxoiron(III) species, which pre-
vents their isolation, EPR spectroscopy has been commonly used as a tool
for their identification, as they are typically low spin (S ¼ 1/2). Thus, in
some cases these compounds have been detected by EPR in samples frozen
after mixing an iron(II) precursor and the desired oxidant, which can be
either H2O2 in combination with a carboxylic acid or an alkyl peracid
alone. Following this strategy, Richens and coworkers observed that
by reacting [FeII(TPA)(CH3CN)2]2+ and mCPBA (meta-chloroperbenzoic
acid) at 40°C a low spin S ¼ 1/2 species was formed, which was assigned
to an acylperoxoiron(III) species [FeIII(O2COC6H4Cl)(TPA)]2+ (Fig. 5B)
(119). Upon decomposition, the authors observed hydroxylation on the
aromatic ring of the oxidant and they rationalized this observation by invok-
ing the involvement of a high-valent electromer that would result from the
O–O cleavage of the observed acylperoxoiron(III) compound. Talsi and
coworkers studied, in 2007, by EPR and 1H NMR spectroscopies the spe-
cies generated after mixing [FeII(L)(CH3CN)2]2+ (L ¼ TPA or BPMEN)
and H2O2/AcOH or peracetic acid at 50°C in order to determine the
active species involved in nonheme iron oxidation catalysis (Fig. 5B). Their
preliminary study concluded that an oxoiron(IV), formed after decay of an
acylperoxoiron(III) precursor, was the active species in cyclohexene oxida-
tion (120). In subsequent studies performed by the same group, analyses of
samples frozen 30 s after mixing the same complexes with excess mCPBA,
peracetic acid, or H2O2/HOAc at 60°C were described (121,122). The
presence of a S ¼ 1/2 rhombic species with g-values at 2.71, 2.42, and
1.53 was detected, which accounted for 15% or 7% of the iron (using
mCPBA and peracetic acid, respectively). The decay rate of these species
Fig. 5 (A) X-ray structure of [FeIII(O2COCH3)(qn)2]2+ provided by Furutachi and

coworkers along with its schematic representation. Hydrogen atoms have been omitted
for clarity. (B) Schematic representation of spectroscopically trapped acylperoxoiron(III)
compounds.
ARTICLE IN PRESS
increased fivefold in the presence of cyclohexene compared to the self-

decay, and the formation of cyclohexene oxide was observed. At that point,
the authors assigned the gmax ¼ 2.7 species to a low-spin oxoiron(V) com-
pound. However, further spectroscopic studies in which higher amounts
of these species could be accumulated concluded that in fact these g ¼ 2.7
compounds were not oxoiron(V), but must be formulated as
acylperoxoiron(III) species (see later).
Fortunately, in some cases acylperoxoiron(III) species have been
accumulated in higher yield permitting their characterization by several
complementary techniques such as UV–vis, M€ ossbauer, rRaman, and mass
spectrometry, among others. Que and coworkers studied the formation of a
metastable brown species (λmax ¼ 460 nm) upon reaction of [FeII(TPA)
(CH3CN)2]2+ with H2O2 in the presence of acetic acid at 40°C (123).
EPR analysis revealed the presence of an S ¼ 1/2 species accounting for
20% of Fe in the sample, with similar g-values to the ones observed by Talsi
(see above) (121). In an attempt to increase the yield of this species the syn-
thesis of a more electron-rich version of the ligand achieved by introducing
methoxy and methyl substituents in the pyridine rings was envisioned
(Fig. 5B). Indeed, reaction of [FeII(TPA*)(CH3CN)2]2+ with excess H2O2
in the presence of acetic acid at 40°C afforded again the formation of a
UV–vis chromophore at λmax ¼ 460 nm (ε 4000 M1/cm) (123). EPR
analysis of a sample frozen at the time of maximum formation of this chro-
mophore revealed the presence of an S ¼ 1/2 species (g-values ¼ 2.58, 2.38,
and 1.72), in this case in much higher yield (47% of the iron). As a bonus
of this higher purity, complementary spectroscopic analysis (electrospray
mass spectrometry, M€ ossbauer, and rRaman) could be performed. All the
acquired data taken together led the authors to assign firmly the species formed
as a low-spin acylperoxoiron(III).
Using the same strategy, Akimova and coworkers used electron-rich ver-
sions of ligands BPMEN and PDP (see Fig. 5 for a display of the ligands) to
stabilize similar acylperoxoiron(III) species (124). Reaction of the
corresponding iron(II) precursors [FeII(BPMEN*)(CH3CN)2]2+ or
[FeII(PDP*)(CH3CN)2]2+ with H2O2 in the presence of acetic acid gave rise
to the formation of the S ¼ 1/2 EPR signal with gmax ¼ 2.7. Moreover, cal-
culations performed by Shaik and coworkers in 2013 on the reaction of
[FeII(PDP)(CH3CN)2]2+ and H2O2 in the presence of acetic acid pointed
toward the formation of the cyclic peracetate complex (Scheme 11), which
in a subsequent step can undergo O–O cleavage to form a high-valent elec-
tromer (125).
ARTICLE IN PRESS
H
H
O H
O O O O
H •O O
N4
L Fe III N4
L Fe IV LN4FeIII
O O O O
CH3 CH3
CH3
Scheme 11 Diagram of the mechanism proposed by Shaik, Que, and coworkers for the
formation of a peracetate intermediate
In 2013, the generation of an acylperoxoiron(III) compound after

reacting an iron(II) precursor bearing the tetradentate pyridinophane
ligand tBuN4 and peracetic acid at 35°C was presented by Mirica and
coworkers (Fig. 5B) (126). This species was characterized by UV–vis
absorption spectroscopy, which showed a chromophore at λmax ¼ 535 nm
(with a shoulder at 850 nm), and by EPR, which revealed the presence of
an S ¼ 1/2 species with g-values at 2.15, 2.10, and 1.96. Unfortunately,
information about the reactivity of this compound has not been reported
to date.
In sharp contrast to all the acylperoxoiron(III) species described earlier,
Nam and coworkers reported in 2016 the synthesis of a series of mono-
nuclear high-spin (S ¼ 5/2) acylperoxoiron(III) complexes by reaction of
[FeII(13-TMC)(CH3CN)2]2+ and 4 equiv. of different peroxycarboxylic
acids (phenylperoxyacetic acid, peroxybenzoic acid, m-chloroperbenzoic
acid, or peroxyacetic acid) (116). These species (Fig. 5) were kinetically
competent to oxidize both olefins and alkanes with strong C–H
bonds including cyclohexane (their decay rate is linearly proportional
to substrate concentration). In order to provide a reason for the observed
reactivity, the authors hypothesized that the high-spin state of this
acylperoxide makes such compounds directly responsible for substrate
oxidation (116).
3.3.2.1 O–O Cleavage in Low-Spin Acylperoxoiron(III) Species

Interestingly, the aforementioned low-spin acylperoxoiron(III) compounds
do not react with organic substrates. However, they are considered precur-
sors of powerful oxidant compounds: after undergoing O–O bond cleavage
those species can generate high-valent electromers (Scheme 12) that will be
more prone to react toward organic substrates.
A case study example for this reaction is the [FeII(TPA*)(CH3CN)2]2+
complex. As mentioned earlier, this complex reacted with hydrogen
ARTICLE IN PRESS
2+ 2+
N Heterolytic N
N O O–O cleavage N O
FeIII O FeV O
N O N O
N N
Scheme 12 Generation of high-valent oxoiron compounds after heterolytic O–O cleav-
age in acylperoxoiron(III) complexes. Please note that the formulation of the high-valent
species as an oxoiron(V) or an oxoiron(IV) coupled to an alkyl radical is still under
discussion.
peroxide in the presence of acetic acid to afford an acylperoxoiron(III) spe-

cies (123). Fascinating observations can be made upon studying the reactiv-
ity of such a compound: after full formation [FeIII(O2COCH3)(TPA*)]2+
persisted in a steady-state phase at low temperatures and its decay occurred
upon depletion of H2O2. The rate of decay was not affected by the addition
of substrates (1-octene, cyclooctene, or 2-heptene). However, oxidation
products (oxygen atom transfer to 1-octene) were formed catalytically while
the reaction progressed. This kinetic process is an indication that
[FeIII(O2COCH3)(TPA*)]2+ is not the final oxidant in the reaction, but
it is a precursor of the actual oxidant, a high-valent compound which results
from O–O cleavage of the acylperoxide. Indeed, the assignment of the
acylperoxoiron(III) species was supported by DFT calculations. The calcu-
lations showed significant barriers (28 kcal/mol) for the direct oxidation of
cyclooctene, thus making it unlikely that the acylperoxoiron(III) is the
actual oxidant. It was interesting to find that [FeIII(O2COCH3)(TPA*)]2+
could easily undergo O–O bond cleavage affording a high-valent oxoiron
compound that exhibited much lower barriers for substrate oxidation
(<10 kcal/mol).
A study with the [FeII(TPA*)(CH3CN)2]2+ system performed by Talsi
and coworkers gave further insight into the species generated in solution
when peroxides and acetic acid were added. EPR analysis of a sample frozen
just after the addition of H2O2 (6 equiv.) to [FeIII2(OH)2(TPA*)2]2+ in the
presence of acetic acid (20 equiv.) exhibited the expected S ¼ 1/2 signal
corresponding to the acylperoxoiron(III) species. However, another
S ¼ 1/2 species with g-values at 2.070, 2.005, and 1.956 was also detected,
albeit in very low yields (2%–3% of Fe content, Fig. 6) (127). This species was
highly unstable (t1/2 ¼ 5 min at 85°C) and its decay rate was accelerated by
the addition of electron-rich alkenes, giving a second-order rate constant in
the reaction with 1-octene, of 0.032 s1 (t1/2 < 0.5 min at 85°C). For
the structurally related system [FeII(PDP*)(CH3CN)2]2+, similar observations
ARTICLE IN PRESS
O 2+ O 2+
2+
N O N
N N
N V O N O
N Fe O O FeV O
N O FeV O
N N O
N O N
O O
g-values 2.07, 2.01, 1.96 g-values 2.071, 2.008, 1.960 g-values 2.070, 2.005, 1.956
(40%) (<1%) (3%)
Fig. 6 Compounds displaying the S ¼ 1/2 EPR-active species with gmax ¼ 2.07 after
reaction of the corresponding iron(II) precursor and peracetic acid or H2O2/AcOH at
low temperature along with their proposed formulation as oxoiron(V) compounds
and its accumulated yield. The high-valent compound is generated upon O–O bond
cleavage of the corresponding previously formed acylperoxoiron(III) species. Please
note that its formulation as an oxoiron(IV) radical carboxyl is still a matter of debate.
were made (127). EPR analysis of a sample, frozen following mixing the
iron complex with hydrogen peroxide (20 equiv.) in the presence of acetic
acid, afforded the acylperoxoiron(III) compound and also a minor species
(<1% of Fe) with g-values at 2.071, 2.008, and 1.960. These rhombic
species with gmax ¼ 2.07 were assigned to formally oxoiron(V) species, highly
probably generated after the heterolytic O–O bond cleavage of the
acylperoxoiron(III) compound (Fig. 6). Unfortunately, the low yield of such
species precluded further spectroscopic studies.
Recently, Talsi and coworkers published reports where the nature of the
metastable species formed upon reaction of an iron(II) precursor, H2O2, and
different carboxylic acids was analyzed by EPR (128). From their work, it is
concluded that the nature of the S ¼ 1/2 species depends on the structure of
the carboxylic acid employed. Intermediates displaying large g-factor anisot-
ropy are generated when carboxylic acids with primary and secondary
α-carbons are employed. In contrast, the use of tertiary α-carbon favors
lower g-factor anisotropy. The authors proposed an oxoiron(IV) carboxyl
radical for the latter species, while an oxoiron(V) was the chosen formulation
for the larger g-factor anisotropy (nonbranched carboxylic acids). Com-
pounds with small g-factor anisotropy exhibited higher levels of
enantioselectivity in the epoxidation of chalcone substrates. This was
explained in terms of the additional stabilization via delocalization of the
unpaired electron over the carboxylic moiety.
In very recent work, the same authors also proposed that the iron-
catalyzed oxidation of olefins might follow different pathways depending
ARTICLE IN PRESS
on the oxidant employed (129). According to their findings, the reaction of

peroxides with the iron precursor would generate an FeIII(OOR) species
that would lead to oxygen atom transfer to the olefin, presumably via for-
mation of an acyclic radical intermediate species. In contrast, when the com-
bination of H2O2/RCOOH is employed, a high-valent [FeV(O)(OCOR)
(L)]2+ would be generated and that would oxidize the olefin via an acyclic
cationic intermediate. Finally, when peracids are used, the mechanism would
occur via a concerted mechanism from the acylperoxoiron(III) species
(129). However, no conclusive spectroscopic evidence has been obtained
about the different nature of the oxidant generated in the different cases.
In 2015, Costas and coworkers contributed to the subject further by pre-
senting the synthesis and characterization of an iron(II) complex with
a novel tetradentate nitrogen-based ligand, [FeII(PyNMe3)(CH3CN)2]2+.
This compound reacted at 40°C in acetonitrile with peracetic acid
(4 equiv.) forming a metastable species α that could be monitored by
UV–vis spectroscopy (λmax ¼ 490 nm) (130). An EPR analysis of a sample
taken at the maximum formation of the chromophore showed that it con-
tained up to 40% of Fe in the high-spin region of the spectrum (g ¼ 9 and
g ¼ 4–6 regions); it was also present in its M€
ossbauer spectrum to the same
extent. Analysis of the low-spin region, which showed mainly two different
S ¼ 1/2 species, proved to be very interesting as it provided insight relative
to the study of the intriguing gmax ¼ 2.07 species observed before for related
systems (see above). This region of the spectra was dominated by the pres-
ence of an S ¼ 1/2 species with g-values at 2.07, 2.01, and 1.95 (40% of Fe)
and a smaller component with g-values at 2.20, 2.19, and 1.99 (5% of Fe).
Because of the substantially higher yield of the species with a 2.07 value
attained by using this system, it represented a valuable opportunity for per-
forming kinetic and spectroscopic studies in order to obtain more informa-
tion about its nature and reactivity.
The addition of cyclohexane accelerated the decay of α as a function of
cyclohexane concentration, exhibiting pseudo-first-order kinetics, and the
reaction afforded cyclohexanol and cyclohexanone in a 9:2 ratio. A kinetic
isotope effect (KIE) of 5 was obtained using the perdeuterated cyclohexane,
which matches the values found for nonheme iron-catalyzed cyclohexane
hydroxylations. It is interesting to note that this KIE value was obtained
both from direct measurement of the reaction rates of α with the substrate
and from the product analysis of a competitive oxidation of cyclohexane and
cyclohexane-d12 under catalytic conditions by the iron(II) precursor and
peracetic acid. α exhibited a high retention of configuration (>95%) in
ARTICLE IN PRESS
the oxidation of cis-1,2-dimethylcyclohexane and a high selectivity for the

3ary C–H bonds of adamantane (3ary/2ary ratio ¼ 29). By using 1,1-
dimethylcyclohexane the authors observed that the oxidation occurred pref-
erentially at the less sterically hindered positions. Finally, preference for the
more electron-rich 3ary C–H bonds was observed when 1-bromo-3,
7-dimethyloctane was used as substrate. All this accumulated experimental
evidence showed that indeed α was a strong oxidant capable of cleaving
unactivated strong C–H bonds with high selectivity, being able to discrim-
inate among multiple C–H bonds in the same molecule on the basis of not
only the C–H bond strength but also responding to steric and polar factors.
It was interesting to note that the 490 nm chromophore followed the
same kinetic trace for the formation and the self-decay as that of the
gmax ¼ 2.07 species observed in the low-spin region of the EPR spectrum.
More importantly, addition of excess cyclohexane yielding maximum for-
mation of α caused the faster decay of the UV–vis chomophore as well as the
species with a gmax ¼ 2.07 value (Fig. 6). This observation indicated that this
compound is either the active oxidant species or it exists in a rapid equilib-
rium with it (130).
Low-temperature mass spectrometry experiments at 40°C revealed the
presence of a peak at m/z ¼ 528.0893, consistent with the formulation of
either an iron peracetate, {[FeIII(OOAc)(PyNMe3)]}, or the oxoiron(V)
high-valent electromer that would be generated after heterolytic O–O bond
cleavage. The addition of deuterated acetic acid caused a shift of 3 mass units
in this molecular peak. This carboxylate incorporation into the molecule
might be explained by invoking the species FeV(O)(OAc), which provides
a suitable rationale for this observation after acetate exchange (Scheme 13).
Scheme 13 Proposed mechanism for the reaction of [FeII(PyNMe3)(CH3CN)2]2+ with

peracetic acid and the subsequent reaction with alkanes and alkenes.
ARTICLE IN PRESS
Further evidence of the carboxylate exchange was obtained by investigating

the reactions of [FeII(PyNMe3)]2+ with cyclooctene using peracetic acid as
oxidant. At room temperature, cyclooctene oxide was the major product (17
turnovers (TON)), but most remarkably, cis-2-acetoxycyclooctanol also was
detected (2 TON). The latter was proposed to arise from a [3 + 2] cycload-
dition of a fleeting FeV(O)(OAc) oxidant to the C]C bond of the olefin
substrate. It is important to remark that such a product does not originate
from an epoxide-ring opening, as it exclusively has syn-stereochemistry.
When the same reaction was carried out in the presence of 200 equiv.
of CD3COOD, up to 29% of deuterated cis-2-acetoxycyclooctanol was
found (130).
In subsequent work, it was showed that indeed, apart from hydroxyl-
ation reactions, α was also competent for performing oxygen atom transfer
reactivity toward olefins (131). Thus, the addition of a series of olefins to α
caused the exponential decay of its spectroscopic features, albeit in this
case the reaction exhibited exceedingly rapid reaction rates, requiring mea-
surements to be made at 60°C. Reaction rates were dependent on the
electronic properties of the substrate, as a higher degree of substitution of
the C–C double bond increased the reaction rate (probably related to the
consequent increase in electron density of the double bond), in agreement
with the electrophilic character of α. This trend was further corroborated by
using a series of para-substituted styrenes, since the presence of electron-
withdrawing substituents decreased the reaction rate.
The catalytic relevance of α was evaluated in a competition experiment
in which two different olefins were reacted at the same time under catalytic
conditions in a [FeII(PyNMe3)(CH3CN)2]2+/AcOOH system (1:10:100,
Fe/AcOOH/substrate, in a mixture of acetonitrile/acetone 1:3 at 60°C).
The relative amounts of the two epoxide products produced from the
catalytic oxidation were compared with the rate constants (k2) obtained
for the reaction of α with the individual alkenes (measured by the
stopped-flow method at 60°C). Specifically, three pairs of olefins (alkeneA
and alkeneB) were used: styrene vs 1-octene, cis-cyclooctene vs cis-2-octene,
and cis- vs trans-2-octene (Scheme 14). Blank experiments showed
that under these experimental conditions, uncatalyzed epoxidation was neg-
ligible (<3% with respect to the oxidant). Very interestingly, reactions
occurred with stereoretention and the relative amount of the two
epoxides formed after each competition reaction (epoxideA/epoxideB)
reasonably matched the ratio of the corresponding k2 values measured
individually for the reaction of α with each olefin (kA B
2 /k2 ). This result
ARTICLE IN PRESS
[FeII(PyNMe3)]2+ (1 equiv.) O O
1 2 + 3 4 +
R R R R
R1 R2 R3 R4
AlkeneA AlkeneB AcOOH (10 equiv.)
MeCN:acetone 1:3 EpoxideA EpoxideB
(100 equiv.) –60°C
AlkeneA AlkeneB Catalytic competition Reaction rate ratios

A
Epoxide
vs C6H13 = 16 ± 3 k2A / k2B = 16.9 ± 0.3
EpoxideB
EpoxideA
vs = 6.1 ± 0.4 k2A / k2B = 4.5 ± 0.2
C5H11 EpoxideB
EpoxideA
vs = 9.5 ± 0.5 k2A / k2B = 8.18 ± 0.03
C5H11 C5H11 EpoxideB
Scheme 14 Schematic representation of the intermolecular competitive epoxidation of
pairs of olefins (alkeneA and alkeneB) by the [FeII(PyNMe3)(CH3CN)2]2+/AcOOH catalytic
system. EpoxideA/epoxideB: ratio of epoxideA and epoxideB products determined by GC.
(kA2 /kB2): ratio of second-order rate constants for the reaction of α with alkeneA (kA2 ) and
alkeneB (kB2) in acetonitrile/acetone 1:3 at 60°C determined by stopped-flow (131).
provided strong evidence that α was a relevant intermediate in catalytic

epoxidation reactions (131) (Scheme 14).
3.4 Spectroscopically Characterized Oxoiron(V) Species

Oxoiron species bearing iron in the oxidation state +5 have been proposed
to be the active species in many oxidative enzymatic reactions. However,
direct characterization of such compounds in natural systems remains elu-
sive. The use of model systems has allowed the entrapment and character-
ization of a few oxoiron(V) species.
Collins and coworkers developed during the 1990s different macrocyclic
tetraamido ligands (TAML) used for the preparation of several examples of
well-defined high-valent iron compounds (132). These ligands bear four
amide groups that upon deprotonation provide exceptionally strong
σ-donor amidate moieties. In 2007, the same group reported the first exam-
ple of a well-defined oxoiron(V) by reacting the monomeric [FeIII(TAML)
(H2O)] with excess mCPBA at 60°C (Fig. 7) (133). Full characterization
by UV–vis, EPR, M€ ossbauer, X-ray absorption, and ESI-MS, as well as
complementary DFT calculations, led the authors to assign unequivocally
the compound as a d3 oxoiron(V) complex with S ¼ 1/2 ground state.
ARTICLE IN PRESS
A – – –
O O O
O O O
N O N N O N N O N
FeV FeV N FeV
N N N N N N
O O O
O O O
[FeV(O)(TAML)]– [FeV(O)(bTAML)]– [FeV(O)((Me2CNCOCMe2NCO)2CMe2)]–
B + C 2+
N O
N O N O
FeV FeIV
N N N Cl
O
N
N
O
[FeV(O)(NC(O)CH3)(TMC)]+ [FeV(O)(Cl-acac•+)(Me3tacn)]2+
Fig. 7 Spectroscopically detected oxoiron(V) compounds (A) with the TAML platforms
(B) generated from the one-electron oxidation of an oxoiron(IV) complex by Que and
coworkers and (C) generated by Che and coworkers after reaction of an iron(III) precur-
sor and Oxone®. Cl-acac ¼ 3-chloro-acetylacetonate.
By substituting a CMe2 group by an N-Me group in the TAML architec-

ture, some years later Gupta and coworkers could achieve the generation of
an FeV(O) complex [FeV(O)(bTAML)] stable at room temperature (Fig. 7A)
(134). The reactivity of [FeV(O)(TAML)] was studied simultaneously by
Nam and coworkers (135) and Collins and coworkers (136), showing
that this compound is competent for the oxidation of unactivated C–H
bonds. Gupta’s model [FeV(O)(bTAML)] could also perform C–H bond
oxidation but at slower rates compared to TAML (one order of magnitude
slower for the oxidation of cyclohexane) (134,136). The last example of an
oxoiron(V) in a TAML ligand has been reported recently by Collins and
coworkers. In this case, an entirely aliphatic scaffold allowed for the first time
the generation of an oxoiron(V) species in pure water as solvent, using
NaOCI as the oxidant (137).
Apart from TAML ligands, direct evidence of FeV(O) species has also been
obtained in other systems. Prat et al. reported in 2011 the detection of [FeV(O)
(OH)(Pytacn)]2+ (Pytacn ¼ 1-(2-pyridylmethyl)-4,7-triazacyclononane) by
means of low-temperature mass spectrometry (138,139). Such high-valent
species is most likely generated after the heterolytic O–O bond cleavage of
the ferric hydroperoxo compound formed after reacting the iron(II) precursor
and hydrogen peroxide. Low-temperature mass spectrometry was also used
by Kodera and coworkers to detect oxoiron(V) compounds during the catal-
ysis of C–H oxidation reactions by a FeIII-monoamidate complex (140).
ARTICLE IN PRESS
Similarly, Che and coworkers proposed the formation of an oxoiron(V) species

after reaction of [FeIII(Cl)2(L-N4Me2)]+ (L-N4Me2 ¼ N,N 0 -dimethyl-2,11-
diaza[3.3](2,6)pyridinophane) and Oxone® on the basis of mass spectrometry
experiments and complementary DFT computations (141).
Que and coworkers followed a different approach in order to synthesize
an FeV(O) compound. In that case, the strategy consisted of generating the
oxoiron(IV) compound [FeIV(O)(TMC)(CH3CN)]2+, and then performing
a one-electron oxidation with tBuOOH in the presence of a base. The
resulting formally [FeV(O)(NC(O)CH3)(TMC)]2+ species (Fig. 7B) was
characterized by several spectroscopic techniques including EPR,
M€ ossbauer, and rRaman (142).
In a recent report, Che and coworkers described an iron(III) compound
in which the iron was bound to a strongly chelating tridentate ligand
Me3tacn and an anionic bidentate ligand [FeIII(Cl-acac•+)(Me3tacn)]2+. This
compound mixed with Oxone® showed high activity toward the oxidation
of typical alkanes such as cyclohexane but also to the more challenging
light alkanes propane and ethane at room temperature, and it was active
for the epoxidation of several alkenes. Mass spectrometry studies revealed
the presence of a peak that was consistent with the formulation as an
oxoiron(V) compound (Fig. 7C). It showed 60% 18O incorporation
from H218O. EPR studies, together with DFT calculations, supported the
presence of an oxoiron(IV) cation radical (formally FeV) compound, which
was proposed to be the active species in the reaction (143).
4. SUMMARY
The generation of (su)peroxo species upon the reduction of O2 by
copper- and iron-based enzymes (or model systems) is a prior step toward
opening the door to an extensive scenery of opportunities. One of the most
interesting transformations is the reaction involving the cleavage of the O–O
bond. Indeed, the resulting species, namely high-valent compounds, can
present reactivity patterns not observed for the peroxo species. Herein,
we have presented examples where the O–O bond cleavage in copper
and iron-based model complexes has led to a dramatic change in the
reactivity.
O–O bond cleavage in peroxodicopper(II) species leads to the gener-
ation of high-valent bis(μ-oxo)dicopper(III) species. Both species can per-
form the ortho-hydroxylation of phenolates, mimicking the model enzyme
ARTICLE IN PRESS
tyrosinase. However, when more challenging substrates such as

2-fluorophenolates are tested, only the high-valent species is capable of
ortho-hydroxylating–defluorinating the aromatic ring. Thus, in this partic-
ular example O–O bond cleavage is crucial for the reaction to proceed.
The breakage of the O–O bond in peroxoiron species to give oxoiron
species has been a topic of interest, and many efforts have been directed
toward a better understanding of such a reaction. The knowledge acquired
through the use of model systems has allowed the development of an array of
synthetic models that has been used to reproduce the hydrocarbon oxidation
reactivity found in enzymes such as Rieske dioxygenases or the family of
Cyt-P450s. In particular, mononuclear nonheme iron complexes with
nitrogen-based ligands act as efficient catalysts in the oxidation of alkanes
and alkenes with high chemo- and stereoselectivities. This dramatic pro-
gress in the catalytic activity is, of course, directly related to a deeper knowl-
edge of mechanistic aspects. Some of the low molecular weight iron
complexes that have been designed have allowed the entrapment and char-
acterization of several iron-oxygen species formed upon reaction with O2
or peroxides, and such compounds have been studied. It prevails that low-
spin iron peroxides are mainly described as sluggish oxidants, not capable of
reacting with hydrocarbons. On the contrary, oxoiron compounds present
a wider reactivity. Indeed, it is well accepted that the heterolytic cleavage
of an iron-peroxo moiety leads to the generation of an oxoiron(V) that is
capable of carrying out the oxidation of unactivated C–H bonds as well as
olefins.
Overall, O–O bond cleavage is a key reaction to enable unprecedented
reactivity in metal-oxygen species. Learning how to control such a reaction
would provide an exceptional opportunity to develop novel methodologies
for hydrocarbon oxidation. In a more long-term goal, gaining control over
this reaction will serve to find alternatives to the traditional toxic oxidative
techniques used for the obtention of value-added products from
hydrocarbons.
REFERENCES
1. Que, L.; Tolman, W. B. Nature 2008, 455, 333.
2. Punniyamurthy, T.; Velusamy, S.; Iqbal, J. Chem. Rev. 2005, 105, 2329.
3. Crabtree, R. H. Chem. Rev. 2010, 110, 575.
4. Oloo, W. N.; Que, L., Jr. Comprehensive Inorganic Chemistry II, 2nd ed.; Elsevier:
Amsterdam, 2013; p763.
5. Kovacs, J. A. Science 2003, 299, 1024.
6. Costas, M.; Chen, K.; Que, L., Jr. Coord. Chem. Rev. 2000, 200–202, 517.
ARTICLE IN PRESS
7. Solomon, E. I.; Sundaram, U. M.; Machonkin, T. E. Chem. Rev. 1996, 96, 2563.
8. Matoba, Y.; Kumagai, T.; Yamamoto, A.; Yoshitsu, H.; Sugiyama, M. J. Biol. Chem.
2006, 281, 8981.
9. Decker, H.; Schweikardt, T.; Tuczek, F. Angew. Chem. Int. Ed. 2006, 45, 4546.
10. Rolff, M.; Schottenheim, J.; Decker, H.; Tuczek, F. Chem. Soc. Rev. 2011, 40, 4077.
11. Company, A. Ideas in Chemistry and Molecular Sciences. Wiley-VCH Verlag GmbH &
Co. KGaA: Weinheim, 2010; p265.
12. Decker, H.; Dillinger, R.; Tuczek, F. Angew. Chem. Int. Ed. 2000, 39, 1591.
13. Solomon, E. I.; Heppner, D. E.; Johnston, E. M.; Ginsbach, J. W.; Cirera, J.;
Qayyum, M.; Kieber-Emmons, M. T.; Kjaergaard, C. H.; Hadt, R. G.; Tian, L. Chem.
Rev. 2014, 114, 3659.
14. Elwell, C. E.; Gagnon, N. L.; Neisen, B. D.; Dhar, D.; Spaeth, A. D.; Yee, G. M.;
Tolman, W. B. Chem. Rev. 2017, 117, 2059–2107.
15. Citek, C.; Herres-Pawlis, S.; Stack, T. D. P. Acc. Chem. Res. 2015, 48, 2424.
16. Keown, W.; Gary, J. B.; Stack, T. D. P. J. Biol. Inorg. Chem. 2016, 1.
17. Serrano-Plana, J.; Garcia-Bosch, I.; Company, A.; Costas, M. Acc. Chem. Res. 2015,
48, 2397.
18. Mirica, L. M.; Ottenwaelder, X.; Stack, T. D. P. Chem. Rev. 2004, 104, 1013.
19. Lewis, E. A.; Tolman, W. B. Chem. Rev. 2004, 104, 1047.
20. Jacobson, R. R.; Tyeklar, Z.; Farooq, A.; Karlin, K. D.; Liu, S.; Zubieta, J. J. Am.
Chem. Soc. 1988, 110, 3690.
21. Kitajima, N.; Fujisawa, K.; Fujimoto, C.; Morooka, Y.; Hashimoto, S.; Kitagawa, T.;
Toriumi, K.; Tatsumi, K.; Nakamura, A. J. Am. Chem. Soc. 1992, 114, 1277.
22. Halfen, J. A.; Mahapatra, S.; Wilkinson, E. C.; Kaderli, S.; Young, V. G.; Que, L.;
Zuberb€ uhler, A. D.; Tolman, W. B. Science 1996, 271, 1397.
23. Mahadevan, V.; Henson, M. J.; Solomon, E. I.; Stack, T. D. P. J. Am. Chem. Soc. 2000,
122, 10249.
24. Lam, B. M. T.; Halfen, J. A.; Young, V. G.; Hagadorn, J. R.; Holland, P. L.; Lledós, A.;
Cucurull-Sánchez, L.; Novoa, J. J.; Alvarez, S.; Tolman, W. B. Inorg. Chem. 2000, 39,
4059.
25. Company, A.; Palavicini, S.; Garcia-Bosch, I.; Mas-Balleste, R.; Que, L.; Rybak-
Akimova, E. V.; Casella, L.; Ribas, X.; Costas, M. Chem. Eur. J. 2008, 14, 3535.
26. Karlin, K. D.; Hayes, J. C.; Gultneh, Y.; Cruse, R. W.; McKown, J. W.;
Hutchinson, J. P.; Zubieta, J. J. Am. Chem. Soc. 1984, 106, 2121.
27. Casella, L.; Monzani, E.; Gullotti, M.; Cavagnino, D.; Cerina, G.; Santagostini, L.;
Ugo, R. Inorg. Chem. 1996, 35, 7516.
28. Santagostini, L.; Gullotti, M.; Monzani, E.; Casella, L.; Dillinger, R.; Tuczek, F. Chem.
Eur. J. 2000, 6, 519.
29. Itoh, S.; Kumei, H.; Taki, M.; Nagatomo, S.; Kitagawa, T.; Fukuzumi, S. J. Am. Chem.
Soc. 2001, 123, 6708.
30. Mirica, L. M.; Vance, M.; Rudd, D. J.; Hedman, B.; Hodgson, K. O.; Solomon, E. I.;
Stack, T. D. P. Science 2005, 308, 1890.
31. Citek, C.; Lyons, C. T.; Wasinger, E. C.; Stack, T. D. P. Nat. Chem. 2012, 4, 317.
32. Herres-Pawlis, S.; Verma, P.; Haase, R.; Kang, P.; Lyons, C. T.; Wasinger, E. C.;
orke, U.; Henkel, G.; Stack, T. D. P. J. Am. Chem. Soc. 2009, 131, 1154.
Fl€
33. Garcia-Bosch, I.; Company, A.; Frisch, J. R.; Torrent-Sucarrat, M.; Cardellach, M.;
Gamba, I.; G€ uell, M.; Casella, L.; Que, L.; Ribas, X.; Luis, J. M.; Costas, M. Angew.
Chem. Int. Ed. 2010, 49, 2406.
34. Kieber-Emmons, M. T.; Ginsbach, J. W.; Wick, P. K.; Lucas, H. R.; Helton, M. E.;
Lucchese, B.; Suzuki, M.; Zuberb€ uhler, A. D.; Karlin, K. D.; Solomon, E. I. Angew.
Chem. Int. Ed. 2014, 53, 4935.
35. Huang, Z.; Lumb, J.-P. Angew. Chem. Int. Ed. 2016, 55, 11543.
ARTICLE IN PRESS
36. Esguerra, K. V. N.; Fall, Y.; Petitjean, L.; Lumb, J.-P. J. Am. Chem. Soc. 2014, 136,
7662.
37. Hoffmann, A.; Citek, C.; Binder, S.; Goos, A.; R€ ubhausen, M.; Troeppner, O.;
Ivanovic-Burmazovic, I.; Wasinger, E. C.; Stack, T. D. P.; Herres-Pawlis, S. Angew.
Chem. Int. Ed. 2013, 52, 5398.
38. M€uller, K.; Faeh, C.; Diederich, F. Science 2007, 317, 1881.
39. Smart, B. E. J. Fluor. Chem. 2001, 109, 3.
40. Amii, H.; Uneyama, K. Chem. Rev. 2009, 109, 2119.
41. Natarajan, R.; Azerad, R.; Badet, B.; Copin, E. J. Fluor. Chem. 2005, 126, 424.
42. Peelen, S.; Rietjens, I. M. C. M.; Boersma, M. G.; Vervoort, J. Eur. J. Biochem. 1995,
227, 284.
43. Bondar, V. S.; Boersma, M. G.; van Berkel, W. J. H.; Finkelstein, Z. I.; Golovlev, E. L.;
Baskunov, B. P.; Vervoort, J.; Golovleva, L. A.; Rietjens, I. M. C. M. FEMS Microbiol.
Lett. 1999, 181, 73.
44. Osborne, R. L.; Raner, G. M.; Hager, L. P.; Dawson, J. H. J. Am. Chem. Soc. 2006,
128, 1036.
45. Fox, B. G.; Borneman, J. G.; Wackett, L. P.; Lipscomb, J. D. Biochemistry 1990, 29,
6419.
46. Battaini, G.; Monzani, E.; Casella, L.; Lonardi, E.; Tepper, A. W. J. W.;
Canters, G. W.; Bubacco, L. J. Biol. Chem. 2002, 277, 44606.
47. Nasir, M. S.; Cohen, B. I.; Karlin, K. D. Inorg. Chim. Acta 1990, 176, 185.
48. Gelling, O. J.; Feringa, B. L. Recl. Trav. Chim. Pays-Bas 1991, 110, 89.
49. Serrano-Plana, J.; Garcia-Bosch, I.; Miyake, R.; Costas, M.; Company, A. Angew.
Chem. Int. Ed. 2014, 53, 9608.
50. Tolman, W. B. Acc. Chem. Res. 1997, 30, 227.
51. Mirica, L. M.; Vance, M.; Rudd, D. J.; Hedman, B.; Hodgson, K. O.; Solomon, E. I.;
Stack, T. D. P. J. Am. Chem. Soc. 2002, 124, 9332.
52. Colomban, C.; Kudrik, E. V.; Afanasiev, P.; Sorokin, A. B. J. Am. Chem. Soc. 2014,
136, 11321.
53. Sahu, S.; Quesne, M. G.; Davies, C. G.; D€ urr, M.; Ivanovic-Burmazovic, I.;
Siegler, M. A.; Jameson, G. N. L.; de Visser, S. P.; Goldberg, D. P. J. Am. Chem.
Soc. 2014, 136, 13542.
54. de Ruiter, G.; Thompson, N. B.; Takase, M. K.; Agapie, T. J. Am. Chem. Soc. 2016,
138, 1486.
55. Groves, J. T. J. Inorg. Biochem. 2006, 100, 434.
56. Kovaleva, E. G.; Neibergall, M. B.; Chakrabarty, S.; Lipscomb, J. D. Acc. Chem. Res.
2007, 40, 475.
57. Gamba, I.; Codolà, Z.; Lloret-Fillol, J.; Costas, M. Coord. Chem. Rev. 2017, 334, 2.
58. Banerjee, R.; Proshlyakov, Y.; Lipscomb, J. D.; Proshlyakov, D. A. Nature 2015, 518,
431.
59. Costas, M.; Que, J. L. Angew. Chem. Int. Ed. 2002, 41, 2179.
60. Bryliakov, K. P.; Talsi, E. P. Coord. Chem. Rev. 2014, 276, 73.
61. Kryatov, S. V.; Rybak-Akimova, E. V.; Schindler, S. Chem. Rev. 2005, 105, 2175.
62. Poulos, T. L. Chem. Rev. 2014, 114, 3919.
63. Denisov, I. G.; Makris, T. M.; Sligar, S. G.; Schlichting, I. Chem. Rev. 2005, 105, 2253.
64. Loew, G. H.; Harris, D. L. Chem. Rev. 2000, 100, 407.
65. Meunier, B.; de Visser, S. P.; Shaik, S. Chem. Rev. 2004, 104, 3947.
66. Groves, J. T.; McClusky, G. A.; White, R. E.; Coon, M. J. Biochem. Biophys. Res.
Commun. 1978, 81, 154.
67. Groves, J. T.; McClusky, G. A. J. Am. Chem. Soc. 1976, 98, 859.
68. Rittle, J.; Green, M. T. Science 2010, 330, 933.
ARTICLE IN PRESS
69. Sydor, P. K.; Barry, S. M.; Odulate, O. M.; Barona-Gomez, F.; Haynes, S. W.;
Corre, C.; Song, L.; Challis, G. L. Nat. Chem. 2011, 3, 388.
70. Gibson, D. T.; Resnick, S. M.; Lee, K.; Brand, J. M.; Torok, D. S.; Wackett, L. P.;
Schocken, M. J.; Haigler, B. E. J. Bacteriol. 1995, 177, 2615.
71. Mohammadi, M.; Viger, J.-F.; Kumar, P.; Barriault, D.; Bolin, J. T.; Sylvestre, M.
J. Biol. Chem. 2011, 286, 27612.
72. D’Ordine, R. L.; Rydel, T. J.; Storek, M. J.; Sturman, E. J.; Moshiri, F.; Bartlett, R. K.;
Brown, G. R.; Eilers, R. J.; Dart, C.; Qi, Y.; Flasinski, S.; Franklin, S. J. J. Mol. Biol.
2009, 392, 481.
73. Neibergall, M. B.; Stubna, A.; Mekmouche, Y.; M€ unck, E.; Lipscomb, J. D.
Biochemistry 2007, 46, 8004.
74. Ferraro, D. J.; Gakhar, L.; Ramaswamy, S. Biochem. Biophys. Res. Commun. 2005,
338, 175.
75. Barry, S. M.; Challis, G. L. ACS Catal. 2013, 3, 2362.
76. Yoshiyama-Yanagawa, T.; Enya, S.; Shimada-Niwa, Y.; Yaguchi, S.; Haramoto, Y.;
Matsuya, T.; Shiomi, K.; Sasakura, Y.; Takahashi, S.; Asashima, M.; Kataoka, H.;
Niwa, R. J. Biol. Chem. 2011, 286, 25756.
77. Kauppi, B.; Lee, K.; Carredano, E.; Parales, R. E.; Gibson, D. T.; Eklund, H.;
Ramaswamy, S. Structure 1998, 6, 571.
78. Koehntop, K. D.; Emerson, J. P.; Que, L. J. Biol. Inorg. Chem. 2005, 10, 87.
79. Bruijnincx, P. C. A.; van Koten, G.; Klein Gebbink, R. J. M. Chem. Soc. Rev. 2008, 37,
2716.
80. Karlsson, A.; Parales, J. V.; Parales, R. E.; Gibson, D. T.; Eklund, H.; Ramaswamy, S.
Science 2003, 299, 1039.
81. Ashikawa, Y.; Fujimoto, Z.; Usami, Y.; Inoue, K.; Noguchi, H.; Yamane, H.;
Nojiri, H. BMC Struct. Biol. 2012, 12, 1.
82. Thibon, A.; Jollet, V.; Ribal, C.; Senechal-David, K.; Billon, L.; Sorokin, A. B.;
Banse, F. Chem. Eur. J. 2012, 18, 2715.
83. Chakrabarty, S.; Austin, R. N.; Deng, D.; Groves, J. T.; Lipscomb, J. D. J. Am. Chem.
Soc. 2007, 129, 3514.
84. Decker, A.; Chow, M. S.; Kemsley, J. N.; Lehnert, N.; Solomon, E. I. J. Am. Chem.
Soc. 2006, 128, 4719.
85. Neese, F.; Zaleski, J. M.; Loeb Zaleski, K.; Solomon, E. I. J. Am. Chem. Soc. 2000, 122,
11703.
86. Kumar, D.; Hirao, H.; Shaik, S.; Kozlowski, P. M. J. Am. Chem. Soc. 2006, 128, 16148.
87. Oloo, W. N.; Que, L. Acc. Chem. Res. 2015, 48, 2612.
88. Olivo, G.; Cussó, O.; Costas, M. Chem. Asian J. 2016, 11, 3148.
89. Olivo, G.; Cussó, O.; Borrell, M.; Costas, M. J. Biol. Inorg. Chem. 2017, 1.
90. Lindhorst, A. C.; Haslinger, S.; Kuhn, F. E. Chem. Commun. 2015, 51, 17193.
91. Cussó, O.; Garcia-Bosch, I.; Ribas, X.; Lloret-Fillol, J.; Costas, M. J. Am. Chem. Soc.
2013, 135, 14871.
92. Costas, M.; Mehn, M. P.; Jensen, M. P.; Que, L. Chem. Rev. 2004, 104, 939.
93. Engelmann, X.; Monte-Perez, I.; Ray, K. Angew. Chem. Int. Ed. 2016, 55, 7632.
94. Ray, K.; Pfaff, F. F.; Wang, B.; Nam, W. J. Am. Chem. Soc. 2014, 136, 13942.
95. Xing, G.; Diao, Y.; Hoffart, L. M.; Barr, E. W.; Prabhu, K. S.; Arner, R. J.;
Reddy, C. C.; Krebs, C.; Bollinger, J. M. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 6130.
96. Jeoung, J.-H.; Bommer, M.; Lin, T.-Y.; Dobbek, H. Proc. Natl. Acad. Sci. U.S.A.
2013, 110, 12625.
97. Shan, X.; Que, L. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 5340.
98. Zhao, M.; Helms, B.; Slonkina, E.; Friedle, S.; Lee, D.; DuBois, J.; Hedman, B.;
Hodgson, K. O.; Frechet, J. M. J.; Lippard, S. J. J. Am. Chem. Soc. 2008, 130, 4352.
ARTICLE IN PRESS
99. Chiang, C.-W.; Kleespies, S. T.; Stout, H. D.; Meier, K. K.; Li, P.-Y.; Bominaar, E. L.;
Que, L.; M€ unck, E.; Lee, W.-Z. J. Am. Chem. Soc. 2014, 136, 10846.
100. Cho, J.; Jeon, S.; Wilson, S. A.; Liu, L. V.; Kang, E. A.; Braymer, J. J.; Lim, M. H.;
Hedman, B.; Hodgson, K. O.; Valentine, J. S.; Solomon, E. I.; Nam, W. Nature
2011, 478, 502.
101. Park, M. J.; Lee, J.; Suh, Y.; Kim, J.; Nam, W. J. Am. Chem. Soc. 2006, 128, 2630.
102. Seo, M. S.; Kamachi, T.; Kouno, T.; Murata, K.; Park, M. J.; Yoshizawa, K.; Nam, W.
Angew. Chem. Int. Ed. 2007, 46, 2291.
103. Oloo, W. N.; Fielding, A. J.; Que, L. J. Am. Chem. Soc. 2013, 135, 6438.
104. Roelfes, G.; Vrajmasu, V.; Chen, K.; Ho, R. Y. N.; Rohde, J.-U.; Zondervan, C.; la
Crois, R. M.; Schudde, E. P.; Lutz, M.; Spek, A. L.; Hage, R.; Feringa, B. L.;
M€ unck, E.; Que, L. Inorg. Chem. 2003, 42, 2639.
105. Mikhalyova, E. A.; Makhlynets, O. V.; Palluccio, T. D.; Filatov, A. S.; Rybak-
Akimova, E. V. Chem. Commun. 2012, 48, 687.
106. Thibon, A.; England, J.; Martinho, M.; Young, V. G.; Frisch, J. R.; Guillot, R.;
unck, E.; Que, L.; Banse, F. Angew. Chem. Int. Ed. 2008, 47, 7064.
Girerd, J.-J.; M€
107. Hong, S.; Lee, Y.-M.; Shin, W.; Fukuzumi, S.; Nam, W. J. Am. Chem. Soc. 2009, 131,
13910.
108. Bukowski, M. R.; Comba, P.; Limberg, C.; Merz, M.; Que, L.; Wistuba, T. Angew.
Chem. Int. Ed. 2004, 43, 1283.
109. Mekmouche, Y.; Hummel, H.; Ho, R. Y. N.; Que, J. L.; Sch€ unemann, V.;
Thomas, F.; Trautwein, A. X.; Lebrun, C.; Gorgy, K.; Leprêtre, J.-C.;
Collomb, M.-N.; Deronzier, A.; Fontecave, M.; Menage, S. Chem. Eur. J. 2002, 8,
1196.
110. He, Y.; Goldsmith, C. R. Chem. Commun. 2012, 48, 10532.
111. Li, F.; Meier, K. K.; Cranswick, M. A.; Chakrabarti, M.; Van Heuvelen, K. M.;
M€ unck, E.; Que, L. J. Am. Chem. Soc. 2011, 133, 7256.
112. Ho, Y. N.; Que, L., Jr.; Roelfes, G.; Feringa, B. L.; Hermant, R.; Hage, R. Chem.
Commun. 1999, 2161
113. Kim, Y. M.; Cho, K.-B.; Cho, J.; Wang, B.; Li, C.; Shaik, S.; Nam, W. J. Am. Chem.
Soc. 2013, 135, 8838.
114. Liu, L. V.; Hong, S.; Cho, J.; Nam, W.; Solomon, E. I. J. Am. Chem. Soc. 2013, 135,
3286.
115. Makhlynets, O. V.; Rybak-Akimova, E. V. Chem. Eur. J. 2010, 16, 13995.
116. Wang, B.; Lee, Y.-M.; Clemancey, M.; Seo, M. S.; Sarangi, R.; Latour, J.-M.;
Nam, W. J. Am. Chem. Soc. 2016, 138, 2426.
117. Furutachi, H.; Hashimoto, K.; Nagatomo, S.; Endo, T.; Fujinami, S.; Watanabe, Y.;
Kitagawa, T.; Suzuki, M. J. Am. Chem. Soc. 2005, 127, 4550.
118. Zhang, X.; Furutachi, H.; Tojo, T.; Tsugawa, T.; Fujinami, S.; Sakurai, T.; Suzuki, M.
Chem. Lett. 2011, 40, 515.
119. Guisado-Barrios, G.; Zhang, Y.; Harkins, A. M.; Richens, D. T. Inorg. Chem. Commun.
2012, 20, 81.
120. Duban, E. A.; Bryliakov, K. P.; Talsi, E. P. Eur. J. Inorg. Chem. 2007, 2007, 852.
121. Lyakin, O. Y.; Bryliakov, K. P.; Britovsek, G. J. P.; Talsi, E. P. J. Am. Chem. Soc. 2009,
131, 10798.
122. Lyakin, O. Y.; Bryliakov, K. P.; Talsi, E. P. Inorg. Chem. 2011, 50, 5526.
123. Oloo, W. N.; Meier, K. K.; Wang, Y.; Shaik, S.; M€ unck, E.; Que, L. Nat. Commun.
2014, 5, 3046.
124. Makhlynets, O. V.; Oloo, W. N.; Moroz, Y. S.; Belaya, I. G.; Palluccio, T. D.;
Filatov, A. S.; Muller, P.; Cranswick, M. A.; Que, L.; Rybak-Akimova, E. V. Chem.
Commun. 2014, 50, 645.
ARTICLE IN PRESS
125. Wang, Y.; Janardanan, D.; Usharani, D.; Han, K.; Que, L.; Shaik, S. ACS Catal. 2013,
3, 1334.
126. Khusnutdinova, J. R.; Luo, J.; Rath, N. P.; Mirica, L. M. Inorg. Chem. 2013, 52, 3920.
127. Lyakin, O. Y.; Zima, A. M.; Samsonenko, D. G.; Bryliakov, K. P.; Talsi, E. P. ACS
Catal. 2015, 5, 2702.
128. Zima, A. M.; Lyakin, O. Y.; Ottenbacher, R. V.; Bryliakov, K. P.; Talsi, E. P. ACS
Catal. 2016, 6, 5399.
129. Zima, A. M.; Lyakin, O. Y.; Ottenbacher, R. V.; Bryliakov, K. P.; Talsi, E. P. ACS
Catal. 2017, 7, 60.
130. Serrano-Plana, J.; Oloo, W. N.; Acosta-Rueda, L.; Meier, K. K.; Verdejo, B.; Garcı́a-
España, E.; Basallote, M. G.; M€ unck, E.; Que, L.; Company, A.; Costas, M. J. Am.
Chem. Soc. 2015, 137, 15833.
131. Serrano-Plana, J.; Aguinaco, A.; Belda, R.; Garcı́a-España, E.; Basallote, M. G.;
Company, A.; Costas, M. Angew. Chem. Int. Ed. 2016, 55, 6310.
132. Chanda, A.; Popescu, D.-L.; de Oliveira, F. T.; Bominaar, E. L.; Ryabov, A. D.;
M€unck, E.; Collins, T. J. J. Inorg. Biochem. 2006, 100, 606.
133. de Oliveira, F. T.; Chanda, A.; Banerjee, D.; Shan, X.; Mondal, S.; Que, L.;
Bominaar, E. L.; M€ unck, E.; Collins, T. J. Science 2007, 315, 835.
134. Ghosh, M.; Singh, K. K.; Panda, C.; Weitz, A.; Hendrich, M. P.; Collins, T. J.;
Dhar, B. B.; Sen Gupta, S. J. Am. Chem. Soc. 2014, 136, 9524.
135. Kwon, E.; Cho, K.-B.; Hong, S.; Nam, W. Chem. Commun. 2014, 50, 5572.
136. Kundu, S.; Thompson, J. V. K.; Shen, L. Q.; Mills, M. R.; Bominaar, E. L.;
Ryabov, A. D.; Collins, T. J. Chem. Eur. J. 2015, 21, 1803.
137. Mills, M. R.; Weitz, A. C.; Hendrich, M. P.; Ryabov, A. D.; Collins, T. J. J. Am.
Chem. Soc. 2016, 138, 13866.
uell, M.; Ribas, X.; Luis, J. M.; Cronin, L.; Costas, M. Nat.
138. Prat, I.; Mathieson, J. S.; G€
Chem. 2011, 3, 788.
139. McDonald, A. R.; Que, L. Nat. Chem. 2011, 3, 761.
140. Hitomi, Y.; Arakawa, K.; Funabiki, T.; Kodera, M. Angew. Chem. Int. Ed. 2012, 51,
3448.
141. Chow, T. W.-S.; Wong, E. L.-M.; Guo, Z.; Liu, Y.; Huang, J.-S.; Che, C.-M. J. Am.
Chem. Soc. 2010, 132, 13229.
142. Van Heuvelen, K. M.; Fiedler, A. T.; Shan, X.; De Hont, R. F.; Meier, K. K.;
Bominaar, E. L.; M€ unck, E.; Que, L. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 11933.
143. Tse, C.-W.; Chow, T. W.-S.; Guo, Z.; Lee, H. K.; Huang, J.-S.; Che, C.-M. Angew.
Chem. Int. Ed. 2014, 53, 798.
ARTICLE IN PRESS
μ-Nitrido Diiron Phthalocyanine

and Porphyrin Complexes:
Unusual Structures With
Interesting Catalytic Properties
Alexander B. Sorokin1
Institut de Recherches sur la Catalyse et l’Environnement de Lyon IRCELYON, UMR 5256,
CNRS—Universite Lyon 1, Villeurbanne cedex, France
1
Corresponding author: e-mail address: alexander.sorokin@ircelyon.univ-lyon1.fr
Contents
1. Introduction 108
2. Preparation and Spectroscopic Characterization of μ-Nitrido Diiron Macrocyclic
Complexes 111
2.1 Structural Determination by X-Ray Diffraction 112
2.2 Structural Determination by Extended X-Ray Absorption Fine Structure 115
2.3 Mass Spectrometry 116
2.4 Determination of the Iron Oxidation State by Mo €ssbauer, XANES, and
EPR Techniques 116
2.5 X-Ray Absorption and Emission Spectroscopies 119
2.6 Relationship Between Structure of μ-Nitrido Diiron Complexes and Their
Oxidation Properties 121
3. High-Valent Diiron-Oxo Species and Their Spectroscopic Characterization 122
3.1 Phthalocyanine Platform 122
3.2 Porphyrin Platform 124
3.3 Influence of the Macrocyclic Structure on the Formation of Diiron
Oxo Species 127
4. Catalytic Properties 128
4.1 Comparison of Oxidation Properties of Diiron and Iron Oxo Species 128
4.2 Oxidation of Methane 130
4.3 Oxidation of Ethane 134
4.4 Oxidation of Benzene 136
4.5 Reactivity of μ-Nitrido Diiron Complexes in Combination With tBuOOH 139
4.6 Oxidation of Alkylaromatic Compounds 147
4.7 Formation of C–C Bonds 149
4.8 Oxidative Dehalogenation 152
4.9 Transformation of Aromatic C–F Bonds Under Oxidative Conditions 153
4.10 Degradation of Recalcitrant Pollutants 158
Advances in Inorganic Chemistry # 2017 Elsevier Inc. 107

ARTICLE IN PRESS
108 Alexander B. Sorokin
5. Conclusion and Outlook 160

Acknowledgments 162
References 162
Abstract
A novel bioinspired approach to the development of powerful catalysts for oxidation
is based on the N-bridged diiron phthalocyanine and porphyrin complexes. This scaf-
fold is particularly suited for the stabilization of FeIVFeIV entities and can therefore be
useful for the preparation of oxidizing active species. The possibility of the charge
delocalization on two iron sites, two macrocyclic ligands, and the nitrogen bridge
makes possible the activation of peroxides including H2O2. The ultra high-valent
diiron-oxo species (L)FeIV–N–FeIV(L+%)]O (L ¼ phthalocyanine, porphyrin) have been
detected at low temperatures and characterized by cryospray MS, UV–vis, EPR, X-ray
absorption spectroscopy, and Mo €ssbauer techniques. These highly electrophilic (L)
FeIV–N–FeIV(L+%)]O species show outstanding reactivity. The catalytic applications
of μ-nitrido diiron complexes include oxidation of methane and benzene, transforma-
tion of aromatic C–F bonds in oxidative conditions, oxidative dechlorination, and for-
mation of C–C bonds. Importantly, all these reactions can be carried out under clean
and mild conditions with high conversions and turnover numbers. μ-Nitrido diiron
species demonstrate similar mechanistic features (18O labeling, formation of benzene
epoxide, and NIH shift in the aromatic oxidation) (see Section 4.4 and Kudrik and
Sorokin, 2008 for the explanation) as enzymes operating via high-valent iron-oxo spe-
cies. μ-Nitrido diiron complexes transform perfluorinated aromatic compounds in oxi-
dative conditions, while the strongest oxidizing enzymes do not. Advanced
spectroscopic and reactivity studies confirm the participation of high-valent diiron-
oxo species in these catalytic reactions.
1. INTRODUCTION
The ability of cytochrome P450 (1) and methane monooxygenase
(MMO) (2,3) enzymes to perform challenging oxidations of hydrocarbons
has inspired an extensive research in bioinspired catalysis. The objective of
these studies is to understand how P450 and MMO enzymes catalyze dif-
ficult chemical reactions under very mild conditions, and, using this
knowledge, to develop efficient and practical chemical catalysts. These
enzymes activate dioxygen to form high-valent iron-oxo entities having
strong oxidizing properties. Diiron nonheme oxo species of soluble
MMO (sMMO) perform the oxidation of CH4 to CH3OH under ambient
conditions, a reaction that still remains a fundamental challenge in
ARTICLE IN PRESS
Bioinspired Oxidation 109
chemistry. Cytochrome P-450 enzymes catalyze oxidation of C–H bonds

in a variety of substrates, except methane (4). Since cytochrome P-450
chemistry involves mononuclear iron porphyrin and the active site of
sMMO contains a binuclear iron nonheme unit, the progress in this
field has been associated with the development of mononuclear meta-
lloporphyrin systems (5–7) as well as mononuclear (8,9) and binuclear
(10) iron nonheme complexes. Extensive research during the last three
decades resulted in the development of many efficient bioinspired catalysts
for different oxidation reactions (7–9,11–14). Complexes mimicking the
structural and spectroscopic properties of sMMO were obtained, but a
diiron nonheme catalyst performing methane oxidation is still to be devel-
oped (15).
Several years ago, we proposed a novel approach to the development of
bioinspired oxidation catalysts for challenging oxidations based on the diiron
macrocyclic porphyrin-like complexes (16). They reflect the structural features
of two of the most powerful oxidizing enzymes: two iron sites as in sMMO
accommodated in the porphyrinoid ligand environment as in cytochrome
P-450 (Fig. 1).
Diiron macrocyclic concept
Soluble MMO Cytochrome P-450
Binuclear nonheme site Mononuclear heme site
Phthalocyanine and porphyrin diiron complexes

Fig. 1 Schematic representation of the binuclear macrocyclic concept.
ARTICLE IN PRESS
O
IV O
LFe IV O
Fe L FeIV + .
O
or FeIV + . N
IV IV
O Fe O Fe FeIV
L L
Charge delocalization Charge delocalization Charge delocalization

at two iron sites at one iron site and at at two iron sites and
in soluble MMO one macrocycle in P-450 at two macrocycles
Fig. 2 Stabilization of different high-valent iron-oxo species by charge delocalization.
Despite the general belief in the catalytic inertness of dimeric macrocy-

clic complexes, we suggested that the binuclear construction can be advan-
tageous for oxidation catalysis. Indeed, upon formation of active oxidizing
species the iron sites of P-450 and sMMO increase their oxidation state by
two redox equivalents. The high-valent species are stabilized by charge
delocalization at the iron center and porphyrin ligand of (P+%)FeIV]O in
P-450 and at two Fe(IV) sites in sMMO (Fig. 2).
The diiron macrocyclic scaffold provides two iron sites and two
porphyrinoid macrocycles for the delocalization of the charge developed
in high-valent iron-oxo species, which should favor their formation. These
oxidizing species can be obtained using H2O2, tBuOOH, and
m-chloroperbenzoic acid (m-CPBA) oxidants. From environmental and
economic perspectives, H2O2 is the most attractive oxidant. Initially formed
iron peroxo complex (L)Fen–OOH (L ¼ ligand) can be subjected to homo-
lytic or heterolytic O–O bond cleavage to generate a poor oxidant
(L)Fen+1]O or a much stronger oxidant (L)Fen+2]O or (L+%)Fen+1]O,
respectively. The homolytic cleavage appears to be a principal route in
mononuclear iron porphyrin complexes to generate PFeIV]O species
having weak oxidizing properties (Fig. 3) (17).
In the case of the binuclear complex LFeIII–X–FeIII(L)OOH, a hetero-
lytic O–O bond cleavage can be predominant due to delocalization of the
charge on two iron sites and two macrocyclic ligands to form LFeIV–X–
FeIV(L)]O with strong oxidizing properties (Fig. 3B). The properties of
the binuclear complex and the iron oxidation state depend on the nature
of the bridging atom X. Two iron sites of the binuclear complex can be con-
nected by μ-oxo, μ-nitrido, and μ-carbido groups to form Fe(III)–O–
Fe(III), Fe(III)–N]Fe(IV), and Fe(IV)]C]Fe(IV) entities (18). Among
these single atom-bridged complexes, μ-nitrido species are particularly
interesting (Fig. 4) (19,20).
ARTICLE IN PRESS
A
Poor oxidant OH
O O O
Principal O
.
pathway - HO-
FeIV +HO FeIII FeIV + . FeV
Homolytic cleavage Heterolytic cleavage
B OH
O O O
O
FeIII Principal FeIV + . FeV

FeIV -
HO . pathway
N
N N
N - HO-
Fe IV Fe IV FeIV
FeIV
Strong oxidant
Fig. 3 Formation of the high-valent iron-oxo species from iron peroxocomplexes at
mononuclear and binuclear macrocyclic platforms.
R3
R3 R4
R1 R2
R2 R4 N
R2 N
N N
R2 N N R1 Fe
Fe N N
R1 N N R3
N R2 R4 N
R2 R4
R1 N R2 R3 N R5
R2 R2 R1 R2
R2 N
R2 N R5 N N
N R1 N
Fe Fe
R1 N N R2 N N N
R5
N
R2 R2
R2 R1
R5
R1 = Ph: (FeTPP)2 N R3 = R5 = tBu: (FePctBu4)2N

R2 = Et: (FeOEP)2 N R3 = R4 = R5 = H: (FePc)2N
R3 = SO2 Ad, R4 = R5 = H: (PcSO2R4)FeNFe(Pc)
R3 = R4 = SO2n-C6H13, R5 = H: (PcSO2R8)FeNFe(Pc)
Fig. 4 Structures of μ-nitrido diiron macrocyclic complexes.
2. PREPARATION AND SPECTROSCOPIC

CHARACTERIZATION OF μ-NITRIDO DIIRON
MACROCYCLIC COMPLEXES
Forty years ago, Summerville and Cohen published the first synthesis
of the N-bridged diiron complex supported by a tetraphenylporphyrin
ligand (21). This complex has attracted the attention of several groups
because of the particular structure and properties (22–26). An important
advance has been achieved in 1984, when Goedken and Ercolani reported
the synthesis of μ-nitrido diiron phthalocyanine (27). Further studies of
ARTICLE IN PRESS
μ-nitrido diiron homoleptic phthalocyanine (28,29), porphyrin (30), and

tetraazaporphyrin (31,32) complexes, as well as heteroleptic species
(33–36), confirmed their interesting structural and spectroscopic properties.
However, catalytic applications of these complexes have never been envis-
aged. The progress in preparation and spectroscopic characterization of
single-atom-bridged binuclear macrocyclic complexes including μ-nitrido
diiron constructions was summarized by Ercolani and coworkers (18).
μ-Nitrido macrocyclic complexes can be prepared (i) by heating of a
monomeric iron macrocyclic precursor in the presence of NaN3 and
(ii) via the initial preparation of the azido complex (L)Fe–N3 followed by
heating (18,20). In addition to porphyrin, phthalocyanine, and
porphyrazine ligands, a limited number of N-bridged diiron nonheme com-
plexes have been described (37,38). The latter complexes are usually pre-
pared by photolysis of (L)FeIII–N3 complexes. Along with homoleptic
diiron structures, heteroleptic complexes containing two iron sites in differ-
ent ligand environments can be synthesized.
The structural and electronic properties of μ-nitrido diiron species can be
studied by different spectroscopic techniques. Combination of several
methods allows characterizing the structure, the oxidation, and spin states
of the metal centers as well as the state of the macrocyclic ligands. The prop-
erties of μ-nitrido diiron complexes can be tuned using phthalocyanine
(18,20), porphyrazine (39), porphyrin (18,20), and probably other macro-
cyclic ligands, as well as by introduction of electron-donating (18,20) or
electron-withdrawing (40) substituents. Advanced spectroscopic methods
have been applied in order to obtain insights into their electronic structures
and even to study their chemical reactivity.
2.1 Structural Determination by X-Ray Diffraction

Preparation of crystals of macrocyclic complexes suitable for X-ray diffrac-
tion analysis is often challenging, in particular when phthalocyanine ligands
are involved. For this reason, a limited number of X-ray structures of
μ-nitrido diiron complexes are available (22,30,39,41–45), and some of
them are shown in Fig. 5. The most important structural parameters of all
available structures are collected in Table 1.
The Fe–Fe distances are quite short in all Fe(III)Fe(IV) structures with
different macrocycles, typically between 3.31 and 3.36 Å. This Fe–Fe dis-
tance is longer in the complexes with tetraphenylporphyrin ligands owing
to steric requirements of phenyl groups. Interestingly, the complexes in a
ARTICLE IN PRESS
N(3) N(2) N(5)

Fe(1)
N(4)
N(1)
Fe(la)
N(5a)
N(3a)
N(2a) N(4a)
(FeTPP)2N (FePzPr8)2N
C(82)
C(81) C(72)
C(71)
C(b7)
C(b8)
C(
a7
C(m1) )
C(12) C(a8) C(m4) C(62)
C(11)
N(4)
C(a1
Fe(1) N(4)
1.665(4)
C(a6) C(61)
N(1)
)
N(3) C(b6)
C(b1) Fe(1)
175.2(2) N
N(1) C(b5)
1.649(4)
C(b2)
N(3)
Fe(2) C(a2) )
a5
N(2)
C(21) N(2) C( C(51)
C(m2) C(a3) C(a4) C(52)

C(22)
C(m3)
C(b3)
C(b4)
C(31) C(41)
C(32) C(42)
(FeOEP)2N
Fig. 5 Crystal structures of (FeTPP)2N, (FePzPr8)2N, and side-on and top-down view of
(FeOEP)2N. Hydrogen atoms are omitted for clarity.
higher Fe(IV)Fe(IV) oxidation state feature shorter Fe–Fe distances,

between 3.26 and 3.30 Å. Because of phenyl substituents, out-of-plane
deviation of iron atoms is significantly larger for tetraphenylporphyrin com-
plexes (0.28–0.41 Å), leading to longer distances between macrocyclic
planes (3.93–4.00 Å). In contrast, diiron phthalocyanine complexes exhibit
a very small deviation of iron out of the macrocycle plane (0–0.08 Å),
resulting in very short interplane distances of 3.28–3.46 Å. Such short inter-
plane distances should favor a stronger π–π interaction between aromatic
macrocyclic rings in μ-nitrido diiron phthalocyanine complexes. The diiron
porphyrazine complex (FePzPr8)2N exhibits a short Fe–Fe distance of
3.28 Å and moderate out-of-plane deviation of 0.28 Å because of influence
of eight propyl groups at each porphyrazine ligand (39). While porphyrin
and phthalocyanine ligands impose Fe–μN–Fe angles close to 180°, the
Table 1 Structural Parameters of μ-Nitrido Diiron Porphyrin, Phthalocyanine, and Porphyrazine Complexes
Fe–N–Fe Fe–Nbridge Fe–Npyr Length of Fe-Axial Staggering Fe–Fe Interplane
Complex Angle Length (Å) Length (Å) Ligand (Å) Angle Distance (Å) Distance (Å)
(FeTPP)2N (22) 180.0° 1.6605(7) 1.991(3) — 31.7° 3.32 4.15
ARTICLE IN PRESS
(FeTPP)2N (41) 177.0(4)° 1.6795(4) 2.003(3) — 28.7° 3.36 3.93
(FeOEP)2N (42) 175.2(2)° 1.657(11) 2.005(5) — 23.1° 3.31 3.83
[(FeTPP)2N](SbCl6) (30) 180° 1.6280(7) 1.979(5) 30.3° 3.26 4.00
[TPPFeNFePc] [(THF) 179.0(6)° 1.625(2) 2.00(5) Eclipsed 3.28 3.51
H2O]I5 (43)
[(FePc)2N]Br2 (44) 180° 1.639(2) 1.954(9) 2.495 39° 3.28 3.28
[(FePc)2N](N3)2I3 (45) 177.4(4)° 1.650(1) 1.947(5) 2.152(7) 38.5(5)° 3.30 3.46
(FePzPr8)2N (39) 168.5(2)° 1.656(4) 1.914(4)– — 26.2 3.32 3.89
1.662(4) 1.932(4)
ARTICLE IN PRESS
porphyrazine dimer (FePzPr8)2N has a bent structure with the Fe–μN–Fe

angle of 168.5° due to a smaller size of the porphyrazine ring and a higher
flexibility of the n-propyl substituents. Except eclipsed [TPPFeNFePc]
[(THF)H2O]I5 structure, all other complexes are in staggered conformation
with the twist angles between the two macrocyclic rings from 23.1° to 39°
(Table 1).
2.2 Structural Determination by Extended X-Ray Absorption

Fine Structure
When the crystals suitable for X-ray diffraction analysis are not available, the
principal structural parameters can be acquired using the extended X-ray
absorption fine structure (EXAFS) method. For example, tetrasubstituted
phthalocyanine complexes contain four positional isomers CS, C2ν, C4h,
and D2h in a statistical ratio 4:2:1:1, given that steric properties of substituents
do not change this ratio in favor of less sterically crowded isomers. In the case
of μ-nitrido diiron tetrasubstituted phthalocyanines the situation is even
more complicated because of the presence of 10 positional isomers. They
can hardly be separated and isolated in an individual state for the preparation
of their single crystals for X-ray diffraction analysis. In this case, the principal
structural features of the first and second coordination spheres can be deter-
mined by X-ray absorption spectroscopy using synchrotron radiation.
The EXAFS structures of (FePctBu4)2N (46) and its dihalogenated deriva-
tives (F)(PctBu4)FeIV–N–FeIV(PctBu4)+%(F) (47), (Cl)(PctBu4)FeIV–N–
FeIV(PctBu4)+%(Cl) (46), and (Br)(PctBu4)FeIV–N–FeIV(PctBu4)+%(Br) (46)
show that the change of the iron oxidation state and the presence of the axial
ligands do not perturb the structure of the Fe–μN–Fe unit (Fig. 6).
The structure of (FePzPr8)2N was determined by X-ray diffraction anal-
ysis, EXAFS, and DFT optimization. All methods gave very close Fe–Nbridge
lengths: X-ray diffraction: 1.656(4) Å, 1.662(4) Å; EXAFS: 1.65(1) Å; DFT:
1.657 Å and Fe–Npyrrole distances: X-ray diffraction: 1.914(4)–1.929(4) Å,
1.915(4)–1.932(4) Å; EXAFS: 1.93(1) Å; DFT: 1.949 Å (39). Very good
agreement was obtained between EXAFS data and BP86 DFT geometry
optimization of the (FePc)2N structure in the comparative study of the
corresponding μ-oxo, μ-nitrido, and μ-carbido complexes (48). Thus, the
EXAFS method can provide valuable structural information when X-ray
diffraction cannot be applied. However, EXAFS cannot address some essen-
tial structural features, e.g., the determination of staggered or eclipsed con-
figuration of macrocycles in the dimer. This issue might be important for the
ARTICLE IN PRESS
t
t Bu
Bu
t
t
Bu
Bu
+3.5
IV +.
t
t
Bu
Bu t
Bu
t
t
t
Bu Bu
Bu t
Bu
t
Bu
IV
+3.5
t
t
Bu
Bu
t
Bu
t
Bu X = F, Cl, Br
Complex (FePctBu4)2N(Å) (FePctBu4)2N ⫻ 2F–(Å) (FePctBu4)2N ⫻ 2 Cl–(Å) (FePctBu4)2N ⫻ 2Br–(Å)
Fe–Fe distance 3.33 3.28 3.35 3.39
Fe–Nbridge 1.65 1.64 1.68 1.69
Fe–Nmeso 3.41 3.44 3.42 3.45
Fe–Npyrrole 1.94 1.94 1.94 1.93
Fe–X 1.99 2.33 2.54
Fig. 6 EXAFS structures of (FePctBu4)2N and its dihalogenated complexes (X)(PctBu4)

FeIV–N–FeIV(PctBu4)+%(X), X ¼ F, Cl, Br having quasilinear Fe–μN–Fe fragments.
electronic configuration and even might influence the catalytic properties

μ-nitrido diiron complexes.
2.3 Mass Spectrometry

μ-Nitrido diiron phthalocyanines and porphyrins exhibit strong molecular
cluster peaks with no or very minor fragmentation owing to mono-
merization. In addition to the structural identification, electrospray ioniza-
tion mass spectrometry (ESI-MS) is particularly useful because the
electrospray ion source allows keeping intact labile compounds. An ESI-
MS experiment can be performed at variable low temperatures (80°C
to 10°C). This cold spray ionization mass spectrometry (CSI-MS) allows
detecting and investigating elusive catalytic intermediates (16,19) and even
their fine structural organization using a collision-induced dissociation
MS/MS technique (CID-MS/MS) (19,49) (vide infra).
€ssbauer,
2.4 Determination of the Iron Oxidation State by Mo
XANES, and EPR Techniques
This important information can be obtained using a combination of several
spectroscopic methods. The number of distinct Fe sites in the sample and
their oxidation, spin states, and ratio can be determined by M€
ossbauer spec-
troscopy. Thus, equivalence or nonequivalence of the iron sites in μ-nitrido
diiron complexes can be probed. Diiron macrocyclic complexes can be
ARTICLE IN PRESS
analyzed using solid intermediates or frozen solutions including measure-

ments under catalytic conditions and after catalytic tests. Solid samples
can be analyzed using 2.119% natural 57Fe abundance. When the iron con-
tents of the samples are too low to obtain M€ ossbauer spectra of sufficient
57
quality within reasonable experimental time, Fe-enriched complexes with
more than 90% isotopic enrichment can be prepared and used.
The M€ ossbauer spectrum of (FeTPP)2N showed only one doublet signal
with isomer shift and quadrupole splitting values of 0.18 and 1.08 mm s1,
respectively (50). Consequently, this formally Fe(III)Fe(IV) complex is
described better as a Fe(+3.5)(μ-N)Fe(+3.5) species with equivalent iron
sites (18). The phthalocyanine dimer (FePc)2N also contains equivalent
Fe(+3.5) sites with δ ¼ 0.06 mm s1 and ΔEQ ¼ 1.76 mm s1 at 77K
(28). Interestingly, even the heteroleptic complex TPPFe(μN)FePc features
the same Fe(+3.5)(μ-N)Fe(+3.5) unit with M€ ossbauer parameters just in
between those of homoleptic (FeTPP)2N and (FePc)2N: δ ¼ 0.113 mm s1
and ΔEQ ¼ 1.467 mm s1 at 77K (33). Therefore, the efficiency of the elec-
tronic exchange along the Fe–N–Fe fragment overcomes the effects pro-
duced by the different ligand environments around the two iron ions (18).
The interesting feature of the μ-nitrido macrocyclic platform is the sta-
bilization of high iron oxidation states. First, as prepared μ-nitrido diiron
complexes are either in Fe(+3.5)(μ-N)Fe(+3.5) or in the Fe(IV)(μ-N)
Fe(IV) configuration. Electron-rich macrocyclic ligands bearing electron-
donating substituents stabilize the Fe(IV)(μ-N)Fe(IV) state (51). The steric
properties of substituents can also influence the Fe oxidation state. A series of
[FePc(SO2R)]2N (R ¼ alkyl) complexes have been prepared and studied by
M€ ossbauer, EPR, and other spectroscopic methods (40,52). While com-
plexes having bulky SO2R substituents with tert-butyl, adamantyl, and
cyclohexenyl groups are in a Fe(+3.5)(μ-N)Fe(+3.5) state, the related com-
plexes with smaller methyl, ethyl, and n-hexyl groups are Fe(IV)(μ-N)
Fe(IV) species. This influence of the size of alkylsulfonyl substituents on
the iron oxidation state is striking since their electron-withdrawing proper-
ties are similar. One can assume that geometrical conformation of μ-nitrido
dimers might depend on the size of substituents. The distances between two
phthalocyanine moieties and/or staggering angle between them might be
different for two families of complexes that could influence the iron ground
oxidation state. However, the absence of X-ray structural data for the
reasons mentioned earlier prevents us to draw this conclusion.
A number of high-valent iron complexes containing the Fe(IV)(μ-N)
Fe(IV) entity have been prepared and characterized (18). These complexes
ARTICLE IN PRESS
often contain a macrocyclic cation radical and they are isoelectronic with
Fe(IV)(μ-N)Fe(V) species. In sharp contrast to bioinspired high-valent
iron-oxo species, some of these ultra high-valent diiron complexes are stable
at room temperature. Their high oxidation state can be identified by the
M€ ossbauer method on the basis of very low and even negative values of iso-
mer shift. The Fe(IV)(μ-N)Fe(IV) porphyrin and phthalocyanine complexes
exhibit δ values between 0.03 and 0.13 mm s1.
The EPR spectra of μ-nitrido diiron complexes allow distinguishing
three electronic states. The Fe(+3.5)(μ-N)Fe(+3.5) complexes are
S ¼ 1/2 species with axial symmetry formally resulting from antiferromag-
netic coupling between Fe(IV) (S ¼ 1) and Fe(III) (S ¼ 3/2) in the Fe(III)–μ
N–Fe(IV) core. One-electron oxidized Fe(IV)Fe(IV) complexes are EPR
silent. Two-electron oxidized Fe(IV)Fe(IV) cation-radical species features
a single strong narrow symmetric signal at g 2.00 due to an organic radical.
For example, frozen (FeOEP)2N solution in THF showed low-spin signals
in the g 2 region corresponding to the S ¼ 1/2 species (Fig. 7) (51).
The signal of (FeOEP)2N contained a major axial species with g║ ¼ 2.06
and g┴ ¼ 2.01. Minor species at g ¼ 2.15 exhibited a characteristic 14N
hyperfine splitting. The EPR spectrum of (FeTPP)2N recorded at the same
conditions showed signals at g║ ¼ 2.01 and g┴ ¼ 2.15 with a well-resolved
isotropic 14N hyperfine structure, which can only be observed at a low
dimer concentration in weakly interacting solvents (53). The major axial sig-
nal in the (FeOEP)2N EPR spectrum can be formed due to an increase of
axiality owing to the interactions between the (FeOEP)2N molecules,
which should be more favorable than in the case of more hindered
(FeTPP)2N. The similar axial signal was observed for structurally related
(FePctBu4)2N in CH2Cl2 solution (54). The minor signal of (FeOEP)2N
at g ¼ 2.15 is due to naked species resembling the (FeTPP)2N complex with
200
g = 2.06 (FeOEP)2N 500 21 G (FeTPP)2N
g = 2.15
0 300
I (a.u.)
I (a .u.)
g = 2.01
100
–200
–100
g = 2.15
g = 2.01
–400 –300
3000 3100 3200 3300 3400 3500 3000 3100 3200 3300 3400
H (G) H (mT)
Fig. 7 X-band EPR spectra of (FeOEP)2N and (FeTPP)2N recorded in THF at 110K.
ARTICLE IN PRESS
unoccupied axial positions. Therefore, the EPR spectrum of μ-nitrido

diiron complexes can contain the signals corresponding to differently inter-
acting molecules. Noteworthy, porphyrin-based species are prone to strong
modifications of their EPR spectra due to solvent effects and to the second-
order magnetic interactions (55).
2.5 X-Ray Absorption and Emission Spectroscopies

The coordination of iron sites as well as iron oxidation and spin states can be
investigated employing synchrotron radiation. These techniques include
extended X-ray absorption fine edge spectroscopy and X-ray absorption
near edge spectroscopy (EXAFS/XANES), X-ray emission spectroscopy
(XES), and resonant inelastic X-ray scattering (RIXS). An introduction
to these techniques and their applications can be found elsewhere
(56,57). These methods allow the characterization of active catalytic inter-
mediates and often provide valuable data for understanding fine mechanistic
details (58–61).
The K preedge feature of iron, which originates from the excitations of
1s electrons into the lowest unoccupied 3d electronic states, is sensitive to
the oxidation state, the site symmetry, and the crystal field splitting. The
most useful parameters of the Fe K preedge are the position of its maximum
and the integrated intensity of the preedge peak. The XANES spectra of
[PcFe+3.5NFe+3.5Pc]0, [PcFeIVNFeIVPc]+PF6, and [PcFeIVNFeIV(Pc%+)]2+Br2
reveal the difference with the spectrum of monomeric PcFeCl (62). The
most oxidized [PcFeIVNFeIV(Pc%+)]2+Br2 complex revealed the preedge
maximum at 7114.7 eV.
The preedge spectra of (FePctBu4)2N, (Cl)(PctBu4)FeIV–N–
Fe (PctBu4)+%(Cl), and (Br)(PctBu4)FeIV–N–FeIV(PctBu4)+%(Br) are com-
IV
pared with the spectrum of the mononuclear FePctBu4 (Fig. 8) (46).

The XANES spectrum of FePctBu4 exhibited a small preedge feature
at 7112.8 eV expected for octahedral or square-pyramidal Fe(III) com-
plexes. Binuclear (FePctBu4)2N revealed the sharp increase in the preedge
intensity and shift of the preedge peak to higher energy 7114.1 eV. The
increase of the preedge intensity is in agreement with the Fe oxidation
since this increases the number of accessible d-orbitals resulting in the
higher probability of a 1s ! 3d electron transition. Unlike their mononu-
clear precursors, μ-nitrido dimers have very strong preedge features,
suggesting asymmetric coordination and high degree of mixing between
Fe and μN orbitals due to short Fe–μN distances. The higher energy shift
ARTICLE IN PRESS
1.55 FeIVFeIV : 7115.2eV
1.35
Absorbance
FeIIIFeIV
7114.1 eV
1.15
FeIIIPc:
7112.8 eV
0.95
7108 7112 7116
eV
Fig. 8 Preedge region of the XAS spectra of mononuclear FePctBu4 (d d d) in com-
parison with dimeric (FePctBu4)2N (dd), (Cl)(PctBu4)FeIV–N–FeIV(PctBu4)+%(Cl) (—), and
(Br)(PctBu4)FeIV–N–FeIV(PctBu4)+%(Br) (solid line) complexes.
of the preedge peak is consistent with the increase of the efficient positive
charge on the Fe sites, leading to a stronger electron bonding. Upon
formation of (Cl)(PctBu4)FeIV–N–FeIV(PctBu4)+%(Cl) and (Br)(PctBu4)
FeIV–N–FeIV(PctBu4)+%(Br), a further increase of preedge energy to
7115.2 eV was observed due to the higher Fe oxidation state. Notewor-
thy, the second lobe of the preedge peak appeared at 7113.4 eV. This
pre-peak feature with two lobes can probably be explained by a non-
equivalent character of the Fe sites owing to the presence of a cation rad-
ical on one phthalocyanine ligand. As a result, the removal of 3d orbitals
degeneration leads to multiple excited states, resulting in superposition of
two types of preedges. This two-lobe feature of the preedge peak is char-
acteristic of the high-valent FeIV–μN–FeIV complexes having a mac-
rocycle cation-radical.
The RIXS studies at Fe K-edge provide the unique possibility of observ-
ing 3d4 Fe(IV) spin systems (62). RIXS and EXAFS spectra can also be
simulated theoretically and compared with the experimental data (48).
Combined Fe K-edge XAS and the DFT simulations attribute preedge fea-
tures to the details of the electronic structure. In contrast to related μ-oxo-
dimers containing antiferromagnetically coupled high spin centers, μ-nitrido
Fe(III)Fe(IV), Fe(IV)Fe(IV), and Fe(IV)Fe(IV) cation-radical complexes are
in a low-spin (LS) state (62).
ARTICLE IN PRESS
2.6 Relationship Between Structure of μ-Nitrido Diiron

Complexes and Their Oxidation Properties
The μ-nitrido diiron construction is very stable due to the nitrogen bridge
and can stabilize high Fe oxidation states (18). μ-Nitrido complexes can for-
mally be considered as a mixed valence Fe(III)Fe(IV) system with one
unpaired electron showing typical S ¼ 1/2 axial spectra with weak or not
detectable hyperfine splitting on the bridging nitrogen. However,
M€ ossbauer spectra of these complexes with a single doublet indicate two
equivalent iron sites with intermediate +3.5 oxidation state.
A priori, the Fe+3.5Fe+3.5 oxidation state should not favor the activation
of peroxides to give oxo species. Indeed, monooxygenase enzymes and
traditional bioinspired complexes use iron sites in low oxidation states to
form active oxo species. The (L)FeII and (L)FeIII sites are transformed to
(L)FeIV]O and (L)FeV]O/(L+%)FeIV]O species having two redox
equivalents above the initial FeII or FeIII oxidation states (Fig. 9).
Obviously, it should be more difficult to form oxo entities starting with
+3.5 +3.5
Fe Fe complexes because it is more demanding to remove two elec-
trons from already oxidized species. On the other hand, if such entities can
be prepared, they might be even stronger oxidants compared to their mono-
nuclear counterparts. We have been successful in our objective of preparing
these μ-nitrido diiron-oxo complexes. These entities can be considered as
ultra high-valent species since their effective oxidation state is higher than
that in mononuclear oxo complexes. The particular electronic structure
OOH O
+
HOO–
2 redox equivalents
FeIV + .
III III
Fe Fe
above FeIII state
Cationic Neutra l Cationic
OH–
Starting Hydroperoxo Oxo

complex complex complex
OOH − O
Fe+3.5 FeIII FeIV + .

HOO– 3 redox equivalents
N N N above FeIIIFeIII state
Fe+3.5 FeIV OH– FeIV
Neutral Anionic Neutral

Fig. 9 Formation of oxo species on the mononuclear and binuclear platforms. Reprinted
with permission from Afanasiev, P.; Sorokin, A. B. Acc. Chem. Res. 2016, 49, 583–593. Copy-
right 2016 American Chemical Society.
ARTICLE IN PRESS
of μ-nitrido diiron species distinguishes them from corresponding mononu-

clear iron complexes and confers upon them unprecedented catalytic prop-
erties, which will be described later.
3. HIGH-VALENT DIIRON-OXO SPECIES AND THEIR

SPECTROSCOPIC CHARACTERIZATION
3.1 Phthalocyanine Platform
Addition of H2O2 to a solution of (FePctBu4)2N in MeCN or acetone at
25°C resulted in UV–vis and EPR spectral changes (54). The rate of the
spectral changes correlated with H2O2 concentration with saturation behav-
ior of kobs on [H2O2], indicating a two-step reaction with fast reversible
binding of H2O2 to (FePctBu4)2N with K ¼ 417 M1 followed by slow for-
mation of the hydroperoxo complex FeIV–μN–FeIII–OOH with
k1 ¼ 0.017 s1. The k1 value was more than two orders of magnitude less
than that reported for the reaction of monomeric Fe(III) porphyrazine with
H2O2 (2.3 s1, H2O, 25°C) (63). The ESI-MS analysis of a solution of
(FePctBu4)2N in MeCN in the presence of the excess of H2O2 and Et3N
confirmed the formation of hydroperoxo complex with m/z ¼ 1633.7322
and showed the formation of another species with m/z ¼ 1615.7239 attrib-
uted to the diiron-oxo species (Fig. 10).
Intens.
(%)
1599.7268 +MS, 0.8–1.3 min #(45–77)
100
(FePctBu4)2N+
H2O2–0.00018
80
60
O1–0.00279
40
OK[(FePctBu4)2N]+ HOO–[(FePct(Bu4)2N]+
20 1633.7322
1615.7239
0
1595 1600 1605 1610 1615 1620 1625 1630 1635 m/z
Fig. 10 ESI-MS spectrum (positive mode) recorded in the course of reaction of 1 μM
(FePctBu4)2N in the presence of 1000 equiv. H2O2 and 2000 equiv. Et3N.
ARTICLE IN PRESS
Importantly, ESI-MS data showed stability of the Fe–N–Fe binuclear

structure upon formation of these species. Slow formation kinetics and high
reactivity preclude direct detection of PcFeIV]N–FeIV(Pc+%)]O by low-
temperature UV–vis spectroscopy. However, an EPR study supports its for-
mation. Addition of H2O2 at 80°C resulted in the disappearance of the
iron EPR signal and appearance of the narrow signal at g ¼ 2.001 consistent
with the transient formation of PcFeIV]N–FeIV(Pc+%)]O. It was con-
firmed by using labeled H218O2 and the ESI-MS detection of a molecular
peak of labeled PcFeIV]N–FeIV(Pc+%)]18O centered at m/z ¼ 1617.9
(Fig. 11).
The product of interaction of (FePctBu4)2N with H2O2 was prepared in
the solid state with 70% yield. The M€ ossbauer spectrum revealed two quad-
rupole doublets with δ1 ¼ 0.14 mm s1, ΔEQ1 ¼ 1.57 mm s1 and
δ2 ¼ 0.10 mm s1, ΔEQ2 ¼ 2.03 mm s1, respectively. The negative δ
values are in agreement with the Fe(IV) oxidation state. Fe K-edge EXAFS
spectroscopy showed the retention of the Fe–N–Fe core and the presence of
one oxygen atom per two Fe atoms at the 1.90 Å distance. An intense
1s ! 3d preedge feature at 7114.4 eV in XANES confirmed the formation
of Fe(IV) species. On the basis of M€ ossbauer, EPR, EXAFS, and XANES
data, this species was assigned as (Pc)FeIV]N–FeIV(Pc)–OH, which could
18
16
O O
H218O2 oxidant 1617.9
H2O2 oxidant 1615.9 100
100 .
Fe IV + 1617.1 Fe IV + .
1615.0 1616.9 80
Relative abundance
80
Relative abundance
N N
1618.9
60 Fe IV 60 Fe IV
40 1617.9 40
1616.1 1619.8
1613.0 1614.0 1618.9 1622.9
20 1619.7 20
1614.1 1620.8
1612.0
0 0
1612 1614 1616 1618 1620 1615 1620
m/z m/z
1615.7 1617.7
100 1614.7 100
1616.7
Simulation
Simulation
Relative abundance
Relative abundance
80 80
60 1616.7 60 1618.7
40 40
1617.7 1619.7
20 20
1612.7 1613.7 1614.7 1615.7
1618.7 1620.7
1611.7 1619.7 1613.7 1621.7
0 0
1610 1612 1614 1616 1618 1620 1614 1616 1618 1620 1622
m/z m/z
Fig. 11 Positive ESI-MS spectra of the oxo species detected in the reaction of
(FePctBu4)2N with H216O2 and H218O2 (MeCN, 25°C).
ARTICLE IN PRESS
be formed from highly reactive PcFeIV]N–FeIV(Pc+%)]O by hydrogen

atom abstraction from a solvent molecule. Our further efforts to prepare
and to characterize PcFeIV]N–FeIV(Pc+%)]O were unsuccessful because
of the extremely high reactivity of this species.
3.2 Porphyrin Platform

Porphyrin ligands might be more suitable for generation of diiron-oxo spe-
cies. Indeed, monomeric high-valent iron-oxo ortho-substituted porphyrin
species are well-established entities (64), while the first oxo complex on a
phthalocyanine platform was reported recently (65). As a consequence of
the slow reaction rate of the diiron-oxo complex formation from H2O2
even at 25°C, m-CPBA was used as an oxygen donor. Addition of m-CPBA
to red–purple (FeTPP)2N solution in CH2Cl2 at 80°C resulted in the
immediate formation of a brown peroxo complex followed by the appear-
ance of a green species after 10 min (Fig. 12) (41).
The green species exhibits a broad Soret band at 399 nm and a very broad
Q band between 600 and 650 nm characteristics of porphyrin cation-radical
species. The CSI-MS technique revealed a transient signal centered at
Absorbance
-
1.4 O COAr
O -
O
1.2 FeIII −80°C FeIII FeIV + .
+m-CPBA 10 min
N N N
1.0 −H+ −OCOAr−
FeIV FeIV FeIV
immediately
0.8
OCOAr
385, 410, 529 nm 406, 548 nm 399, ~630 nm
0.6 t1/2 ~ 20 min at −80°C
0.4
0.2
0.0
300 400 500 600 700 800

l (nm)
Fig. 12 Formation of [(TPP)Fe(μ-N)Fe(O)(TPP)]0 from (FeTPP)2N and m-CPBA at 80°C in
CH2Cl2. Reprinted with permission from Afanasiev, P.; Sorokin, A. B. Acc. Chem. Res. 2016,
49, 583–593. Copyright 2016 American Chemical Society.
ARTICLE IN PRESS
m/z ¼ 1366.0, which rapidly disappeared regenerating the parent molecular

peak of (FeTPP)2N at m/z ¼ 1350.1. The isotopic distribution pattern of the
transient signal is close to that calculated for the target [(TPP)Fe(μ-N)Fe(O)
(TPP)]0. To provide evidence of the formation of the oxo species which can
transfer the oxygen to substrate, we added 2000 equiv. of cyclohexene to the
green solution at 80°C. The UV–vis spectrum of the initial (FeTPP)2N
was regenerated (>90%) and GC-MS analysis showed a clean formation
of cyclohexene epoxide without traces of allylic alcohol and ketone. These
results indicate two-electron oxo-transfer chemistry performed by the
diiron-oxo porphyrin cation radical [(TPP)Fe(μ-N)Fe(O)(TPP%+)]0 with
the oxo group located at one iron atom (41). A single narrow EPR signal
at g ¼ 2.001 with a peak-to-trough separation of 1.1 mT suggests a well-
isolated S ¼ 1/2 state with the unpaired electron on porphyrin without con-
tribution of metal orbitals. The labeled [(TPP)57Fe(μ-N)57Fe(O)(TPP%+)]0
complex was prepared in pure form at 90°C (41). Its M€ ossbauer spectrum
can be reproduced either with one quadrupole doublet (δ ¼ 0.001 mm s1,
ΔEQ1 ¼ +0.752 mm s1) or by two distinct quadrupole doublets of equal
area with parameters: δ1 ¼ 0.056 mm s1, ΔEQ1 ¼ +0.758 mm s1 and
δ2 ¼ 0.056 mm s1, ΔEQ2 ¼ +0.740 mm s1 (Fig. 13).
Low signal-to-noise ratio owing to unavoidable use of CH2Cl2 pre-
cluded asserting this dissymmetry. Spectra recorded under an external mag-
netic field indicated a zero spin density on FeIV(μ-N)FeIV unit behaving as a
SFeFe ¼ 0 site weakly coupled to a porphyrin radical (41). A relatively low
value of quadrupole splitting and similar or very close values of isomeric shift
of iron sites are consistent with the binding of the oxo group on Fe1 and
another ligand on the Fe2 site, most probably m-chlorobenzoic anion
derived from m-CPBA. This binding should be favored by the high Lewis
acidity of the Fe(IV) site.
EPR and M€ ossbauer data indicate that the S ¼ 1/2 spin density is not
spread over the three paramagnetic centers but is strictly localized on the por-
phyrin ligand. This can be explained by the relative strengths of the exchange
interactions (i) between a low-spin FeIV ion and a porphyrin radical and
(ii) between two low-spin FeIV ions through a μ-nitrido group. Magnetic
susceptibility measurements showed that [FeIV(μ-N)FeIV]5+ complexes were
diamagnetic even at room temperature (37), which means that the exchange
constant JFe–Fe is highly negative with JFe–Fe > 600 cm1. Therefore,
the strong antiferromagnetic interaction between the two low-spin FeIV
sites predominates within (TPP)FeIV(μ-N)FeIV(O)(TPP%+), resulting in a
SFeFe ¼ 0 diiron unit in addition to the SRad ¼ 1/2 porphyrin radical (41).
ARTICLE IN PRESS
0.0
0.2
0.0
Absorption (%)
0.2 B
0.0
0.2
C
−2 0 2
Velocity (mm/s)
Fig. 13 Mo €ssbauer spectra of the labeled [(TPP)57Fe(μ-N)57Fe(O)(TPP%+)]. Spectra were
recorded at 4.2K with a magnetic field applied parallel (A: 60 mT, B: 7 T) or perpendicular
(C: 7 T) to the γ-beam (hatched marks). Simulation 1 was obtained assuming a unique
S ¼ 0 Fe site with the following parameters: δ ¼ 0.000 mm s1, ΔEQ ¼ + 0.746 mm s1,
η ¼ 0 (fixed), and Γ fwhm ¼ 0.284 mm s1. Simulation 2 was performed assuming two dis-
tinct S ¼ 0 Fe sites in the 1:1 ratio with the following parameters: δ1 ¼ 0.056 mm s1,
ΔEQ1 ¼ + 0.758 mm s1, δ2 ¼ 0.056 mm s1, ΔEQ2 ¼ + 0.740 mm s1, η1 ¼ η2 ¼ 0 (fixed),
and Γ fwhm,1 ¼ Γ fwhm,2 ¼ 0.246 mm s1. Contributions of sites 1 and 2 are shown above
the experimental spectra as solid and dashed lines, respectively.
ARTICLE IN PRESS
3.3 Influence of the Macrocyclic Structure on the Formation

of Diiron Oxo Species
In order to obtain a deeper insight into the structural and electronic param-
eters that govern the formation of high-valent diiron-oxo species and their
catalytic activity, two heteroleptic complexes having different iron sites
were prepared (49). One iron phthalocyanine containing four or eight
electron-withdrawing alkylsulfonyl groups was connected with more
electron-rich iron unsubstituted phthalocyanine. This diiron platform gives
a unique possibility to probe the reactivity of iron sites in the intramolecular
fashion due to the formation of two isomeric Fe(μN)Fe]O species
(Fig. 14).
Can nonequivalence of the Fe centers supported by the ligands with dif-
ferent electronic properties induce discrimination in the formation of the
oxo species? To address this point, the reaction of two heteroleptic com-
plexes, (PcSO2R8)Fe(μN)Fe(Pc) and (PcSO2R4)Fe(μN)Fe(Pc), with two
oxidants, H2O2 and m-CPBA, was studied. Detailed mass spectrometric
and DFT studies showed the strong influence of the electronic properties
of the iron site on the formation of the Fe(μN)Fe]O species. The oxidant
was added to the complex solution in CH3CN at low temperature, and the
resulting mixture was introduced into the mass spectrometer using the cold
spray technique. Transient diiron-oxo species were trapped in the quadru-
pole region of the mass spectrometer and underwent a collision-induced dis-
sociation tandem MS-MS fragmentation (CID-MS/MS). Dissociation of
the molecular ion resulted in the production of monomeric fragments, some
of which contained oxo and nitrogen entities (Fig. 15).
Analysis of fragmentation patterns of oxo species obtained from two het-
eroleptic diiron complexes as well as DFT studies showed that in all cases,
Electron-poor site FeIV

SO2R O
SO2R CID-MS2 + +
SO2R
N
A N Fe + Fe
SO2R
N Fe N N N N
N N H2O2 Preferential FeIV+.
N SO2R
N
+ or
N SO2R O
SO2R m-CPBA
SO2R N
N Fe N N O
N N N B
N FeIV+.
O
CID-MS2 + +
Electron-rich site N Fe Fe
+
N N
FeIV
Fig. 14 Formation of two isomeric diiron-oxo species of heteroleptic complex

(PcSO2R8)Fe(μN)Fe(Pc) and intramolecular determination of the position of oxo group
by CID-MS/MS.
ARTICLE IN PRESS
FeIV
2351.4 N
2352.5
FeIV +. 585.1 O
Fe H+
O 568.1 Fe OH
+H+
Fe
2350 2355 2360
MS2 NH
585.1 O O
568.1 +600.1
Fe Fe
2350 2355 2360 N

600.1 Fe +
2335.5 N
472.1
500 1000 1500 2000 2500 560 570 580 590 600 610
m/z m/z
Fig. 15 CID-MS/MS spectrum of the oxo complex obtained from (PcSO2R8)Fe(μN)Fe(Pc)

and H2O2. Inset: comparison of experimental (top) and simulated (bottom) isotopic dis-
tribution of the molecular peak of oxo species. Zoom: m/z region with Fe(Pc)-derived
fragments and their assignments.
oxo ligands were situated at more electron-rich iron sites (route A, Fig. 14).
Only trace amounts of species containing an oxo entity on the electron-
deficient iron center were detected. Therefore, the nature of macrocyclic
ligands influences the formation of the active diiron-oxo species and ulti-
mately might also affect their catalytic properties.
4. CATALYTIC PROPERTIES
4.1 Comparison of Oxidation Properties of Diiron and Iron
Oxo Species
Successful preparation of high-valent iron and diiron-oxo species with the
same tetraphenylporphyrin ligands provided a rare possibility to evaluate the
influence of the dimeric structure on the catalytic activity without interfer-
ence from other factors. Both oxo species were prepared at 60°C and their
reactivity was evaluated using UV–vis decay kinetics in the presence of
ethylbenzene, adamantane, or cyclohexane (41). A direct comparison of
reactivity of (TPP)FeIV(μ-N)FeIV(O)(TPP%+) and (TPP+%)Fe]O under
single-turnover conditions revealed the diiron species to be a much stronger
oxidant (Table 2).
Remarkably, the diiron-oxo species was 26 and 130 times more active
compared to the monomeric oxo complex in the oxidation of ethylbenzene
and adamantane, respectively (Fig. 16).
ARTICLE IN PRESS
Table 2 Rate Constants for the Single-Turnover Oxidation of Alkanes by (TPP)FeIV(μ-N)

FeIV(O)(TPP%+) and (TPP+%)Fe]O Species at 60°C
k1 (M21s21)
k0 × 104
Complex (s21) Cyclohexane Adamantane PhCH2CH3
(TPP)FeIV(μ-N) 25 3 0.079 0.005 5.10 0.37 7.1 0.6
FeIV(O)(TPP%+)
(TPP+%)Fe]O 1.5 0.04 No reaction 0.039 0.002 0.270 0.007
Reaction conditions: 25 μM catalyst, 2 equiv. m-CPBA, 0.001–0.37 M substrate in MeCN–CH2Cl2
(1:1) mixture.
O 0.10 O kdimer /kmonomer ~ 130

kdimer /kmonomer ~ 26
FeIV +. FeIV +.
0.08
N N
CH3
FeIV 0.06 FeIV

kobs (S−1)
kobs (S−1)
BDEC–H = 87 kcal mol−1 5.1 M-1s-1 BDEC–H = 93 kcal mol−1

7.1 M-1s-1
0.04
O O
0.02
FeIV +. FeIV +.
0.039 M-1s-1
-1 -1
0.27 M s 0.00
0.00 0.01 0.02 0.03 0.04 0.05 0.000 0.005 0.010 0.015 0.020 0.025
[PhEt] (M) [Adamantane] (M)
Fig. 16 Comparison of the catalytic activities of (TPP)FeIV(μ-N)FeIV(O)(TPP%+) and

(TPP+%)Fe]O in the oxidation of ethylbenzene and adamantane in 1:1 MeCN–CH2Cl2
at 60°C. Reprinted with permission from Afanasiev, P.; Sorokin, A. B. Acc. Chem. Res.
2016, 49, 583–593. Copyright 2016 American Chemical Society.
The stronger the C–H bond, the larger is the difference in reactivity
between the two species. While the diiron-oxo complex was characterized
by a second-order rate constant k60°C ¼ 0.079 M1s1 in the oxidation of
cyclohexane (bond dissociation energy, BDEC–H ¼ 99 kcal mol1) at 60°
C, no reaction was observed with (TPP+%)Fe]O under the same condi-
tions. This rate constant is among the highest values for cyclohexane oxida-
tion by biomimetic complexes (41). The rate constant at 25°C estimated
from the temperature dependence was calculated to be k25°C 3.4 M1s1.
Such a high value suggests that (TPP)FeIV(μ-N)FeIV(O)(TPP%+) might be
able to oxidize stronger C–H bonds. The kH/kD value obtained from the
second-order rate constants for C6H12 and C6D12 oxidations was 3.6, indi-
cating the involvement of C–H bond cleavage in the rate-determining step
(Fig. 17).
ARTICLE IN PRESS
0.040 O
D OH D
0.035 FeIV +. H
0.030 N 25°C
+
OH
0.025 kH = 0.086
FeIV
kobs (s−1)
0.020 kH/kD = 3.6 ± 0.4 D O D D
0.015 Fe + .
IV
O
0.010 N
D FeIV + .
0.005 kD = 0.024 FeIV
kH/kD = 2.96 ± 0.10
0.000 kH/kD = 3.16 ± 0.10
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45
Concentration (M)
Fig. 17 Determination of intermolecular and intramolecular kH/kD in the oxidation of

cyclohexane and adamantane-1,3-d2, respectively.
The catalytic properties of (TPP)FeIV(μ-N)FeIV(O)(TPP%+) were stud-

ied under turnover conditions at 20°C. Using 0.01 mol% catalyst loading,
adamantane (BDE ¼ 93 kcal mol1) was oxidized with a turnover number
(TON, μmol of product/μmol of catalyst) of 430 (1 h). 1-Adamantanol
was found to be the main product and 3°:2° ratio was determined to be
14:1. Intramolecular kH/kD values were determined using an
adamantane-1,3-d2 probe (66). Similar kH/kD values for the oxidation of
the adamantane tertiary C–H bond by both oxo species indicate their similar
reaction mechanisms. The aforementioned comparative kinetic studies as
well as selective oxidation of cyclohexene to epoxide without formation
of allylic alcohol and ketone confirm μ-nitrido diiron complexes as powerful
oxidation catalysts performing the oxo-transfer chemistry.
4.2 Oxidation of Methane

Methane has particularly strong C–H bonds with dissociation energies of
104.9 kcal mol1, which can be oxidized by the most powerful oxidants.
Even cytochromes P-450 do not catalyze the oxidation of methane (4).
The strong oxidizing properties of (TPP)FeIV(μ-N)FeIV(O)(TPP%+)
prompted us to attempt the catalytic oxidation of CH4 by (FeTPP)2N-m-
CPBA in catalytic conditions. To this aim, (FeTPP)2N and FeTPPCl were
supported onto silica with specific surface of 200 m2g1 with a 20 μmol g1
complex loading. Using 1.1 μmol (FeTPP)2N–SiO2 and 105 μmol
m-CPBA, 13.7 molecules of CH4 were converted to HCOOH by one cat-
alyst molecule, thus confirming the remarkable oxidizing properties of a
μ-nitrido diiron complex. In contrast, no oxidation of CH4 was observed
in the presence of FeTPP–SiO2 under the same conditions. Although the
ARTICLE IN PRESS
yield of methane oxidation products was limited by the concurrent oxida-

tion of m-CPBA and m-chlorobenzoic acid to 2-chloro-1,4-hydroquinone,
m-chlorophenol, and benzoic acid, the yield of HCOOH, based on the oxi-
dant, achieved 43.5%.
In this context, hydrogen peroxide is the oxidant of choice for obvious
reasons. Therefore, we have studied the oxidation of CH4 by H2O2 in the
presence of μ-nitrido diiron phthalocyanine complexes, which are more
available compared to their porphyrin counterparts. Homogeneous oxida-
tion of CH4 in acetonitrile by the (FePctBu4)2N–H2O2 system furnished
HCOOH as the principal product, accompanied by minor amounts of
CH2O and CH3OH as well as CH3COOH and acetone (16,67). The detec-
tion of the latter products suggests the possible transformation of CH3CN.
Moreover, it cannot be excluded that some C1 products could also be
derived from CH3CN. Consequently, special attention should be paid to
the origin of products when the oxidation of CH4 is carried out in organic
solvents. Indeed, C–H bonds of organic solvent are prone to oxidation in the
presence of active species strong enough to oxidize methane. Labeling
experiments using deuterated solvents and labeled 13CH4 permitted the
determination of the origin of oxidation product(s) (Fig. 18) (16,67).
To distinguish between the oxidation of CH4 and possible oxidation of
acetonitrile, the reaction was performed in CD3CN. Indeed, isotopic com-
position of formic acid determined by GC-MS showed 68% HCOOH and
32% DCOOH, thus indicating parallel oxidation of CH4 and CD3CN,
respectively. The oxidation of CH4 does take place with TON ¼ 41, but
H
HCOOH + CH3OH
(FePctBu4)2N
C 68%
H H + H2O2
32 bar, 60°C, 20 h DCOOH + CD3OH
H 32%
CD3CN
44% H13COOH/56% HCOOH

165.5 ppm
(FePctBu4)2N/SiO2–H2O2
13
CH4/CH4 +
13
32 bars, 60°C, 20 h CH2(OH)2/CH2(OH)2
47% 53% 81.6 ppm
+
Labeled products: 13C NMR and GC-MS 50% 13CH3OH/50% CH3OH
49 ppm
Fig. 18 Homogeneous oxidation of methane by the (FePctBu4)2N–H2O2 system in

CD3CN and heterogeneous oxidation of 13C-labeled methane by (FePctBu4)2N/SiO2–
H2O2 system. Reprinted with permission from Afanasiev, P.; Sorokin, A. B. Acc. Chem.
Res. 2016, 49, 583–593. Copyright 2016 American Chemical Society.
ARTICLE IN PRESS
it is masked by solvent oxidation. To avoid this complication, the oxidation

of CH4 was carried out in water using SiO2-supported catalyst (20 μmol g1
(FePctBu4)2N loading, 185 m2 g1). Using labeled 13CH4 the formation of
labeled 13CH3OH, 13CH2O, and H13COOH was monitored by 13C
NMR and GC-MS techniques (Fig. 18). Acetic acid and acetone did
not contain the 13C label and were most probably formed from the oxida-
tive degradation of the complex. Formic acid was a main product produced
from the oxidation of initially formed CH3OH and HCHO present in the
aqueous solution is comparable with CH4 concentrations. Importantly,
methane can be oxidized by (FePctBu4)2N/SiO2–H2O2 in pure water
even at 25°C although with a moderate TON of 13. The optimal reaction
temperature was 40–60°C to provide 26–29 TONs. At a high tempera-
ture, the oxidation was less efficient most probably because of more rapid
catalyst degradation.
The formation of the high-valent diiron-oxo species via the heterolytic
cleavage of the O–O bond should be more favorable in acidic medium due
to the protonation of the hydroxo complex and release of a water molecule
as a better leaving group than the hydroxyl anion (Fig. 19).
The efficiency of the CH4 oxidation in diluted acidic solution showed a
bell-shaped dependence on the H2SO4 concentration (Fig. 20) (67).
The highest catalytic activity was obtained in 0.075 M H2SO4 solution at
60°C. Methane was oxidized to HCOOH with a TON up to 223 and 88%–
97% selectivity, whereas product yields on H2O2 attained 50%. Under these
conditions the catalyst stability was improved and it could be used in three
successive oxidations to achieve a total TON ¼ 492. Significantly, the cata-
lytic activity of the (FePctBu4)2N/SiO2–H2O2 system in 0.075 M H2SO4
solution at 25°C resulted in a TON of 52 compared with a TON ¼ 13 in
water.
-
OH
O O
H+
.
Fe III FeIV+
N N
Fe IV H2O Fe IV
Fig. 19 Formation of the high-valent diiron-oxo active species in the presence of acids.
ARTICLE IN PRESS
120
60°C
0.075 M H2SO4
100
120
Concentration (mM)
80 100
[HCHO]
80
60 [HCO2H]
60
40
40
20
20
0 0
0 50 65 75 85 100 200 25°C 40°C 45°C 50°C 60°C 80°C
[H2SO4] (mM)
TON: 52 74 130 207 223 160
TON: 23 65 134 223 167 74 34
Fig. 20 Dependence of the efficiency of the methane oxidation on the H2SO4 concen-
tration and temperature.
Electronic density on iron site

Electron deficient Electron rich
t
SO2R Bu
N N
N N t
Bu
SO2R
N N N N Fe N N Fe N N
Fe N N N N N
N N N t
Bu
N SO2R N
N N
SO2R N SO2R < N < t
Bu N t
Bu
N N t N
SO2R Bu N
N
N N N Fe N N N Fe N
Fe N N N N N N
N N SO2R N t
Bu
N N N
t
Bu
SO2R
Catalytic activity in methane oxidation

TON = 32.1 TON = 102.9 TON = 223.4
Fig. 21 Comparison of the catalytic activity of μ-nitrido diiron phthalocyanine com-

plexes in the oxidation of methane as a function of their electronic properties.
The efficiency of the CH4 oxidation depends also on the catalyst struc-
ture. Using heteroleptic μ-nitrido diiron phthalocyanine complexes, we
found that the formation of the oxo species occurred preferentially at an
electron-rich iron site of the dimer (Fig. 14) (49). In agreement with the
ESI-MS and DFT studies, the growth of the electron-donating
character of the phthalocyanine ligand along the series [FePc(SO2tBu)4]2
N < (FePc)2N < [FePc(tBu)4]2N led to an increase of the catalytic efficiency.
The values of TON in this row markedly increased: 32.1 ! 102.9 ! 223.4
(Fig. 21).
Importantly, the Fe(μN)Fe structural motif of (FePctBu4)2N was essen-
tial for the catalytic activity. Mononuclear FePctBu4 as well as the
corresponding diiron μ-oxo (Fe–O–Fe) and μ-carbido (Fe]C]Fe) com-
plexes were not active in the oxidation of methane.
ARTICLE IN PRESS
4.3 Oxidation of Ethane

Low-temperature direct oxidation of ethane to CH3COOH is of consider-
able interest since acetic acid is an important industrial chemical. The high
selectivity of the (FePctBu4)2N–H2O2 system in the oxidation of methane to
formic acid has prompted us to apply this system to the oxidation of ethane
(68). Indeed, (FePctBu4)2N and (FePc)2N supported onto silica, graphite, or
nafion showed a promising performance in the oxidation of C2H6 by H2O2
in water at 60°C (Table 3).
Using the (FePc)2N/SiO2 system, acetic acid was obtained with 71%
selectivity and TON of 58. The drawback of all published catalysts is the
cleavage of the C–C bond of ethane, leading to a decrease of the target prod-
uct yield. Similarly, a cleavage of the ethane C–C bond occurred using
μ-nitrido diiron complexes and HCOOH was obtained as the side product.
While the catalytic activity of the (FePctBu4)2N–H2O2 system was signifi-
cantly increased in the presence of 75 mM H2SO4, practically no improve-
ment of the ethane oxidation was observed upon replacement of water with
diluted acid. Given a lower C–H bond strength in C2H6 (101.4 kcal mol1)
with respect to that of CH4 (104.9 kcal mol1) and a higher solubility of
C2H6 in water (1.87 vs 1.42 mmol L1 or 56 vs 22.7 mg L1), one should
expect that the (FePctBu4)2N–H2O2 system should be more efficient in the
Table 3 Catalytic Properties of (FePctBu4)2N and (FePc)2N Complexes Supported Onto

Silica, Graphite, and Nafion in the Oxidation of Ethane by H2O2 at 60°Ca
Selectivity to
Catalyst Solvent TONAcOH TONHCOOH AcOH (%) Yieldb (%)
(FePctBu4)2N/SiO2 H2O 33 27 71 25
(FePc)2N/SiO2 H2O 37 33 69 34
(FePctBu4)2N/SiO2 75 mM 41 44 65 36
H2SO4
(FePc)2N/SiO2 75 mM 58 48 71 54
H2SO4
(FePctBu4)2N/C 75 mM 50 59 63 50
H2SO4
(FePctBu4)2N/N H2 O 48 84 53 65
a
Reaction conditions: 32 bar EtH, 2 mL of solvent, 1 μmol of supported catalyst (20 μmol g1, 50 mg),
60 μL of 35% H2O2 (678 μmol), 60°C, 20 h.
b
Total product yield based on H2O2, calculated assuming that 3 equiv. H2O2 are needed to form one acid
molecule.
TON, turnover number, μmol of product/μmol of catalyst.
ARTICLE IN PRESS
oxidation of C2H6 and CH4. However, the TON of the oxidation of CH4
and C2H6 in H2O was comparable: TONHCOOH ¼ 26 (16) for CH4 and
TONAcOH ¼ 33, TONHCOOH ¼ 27 for C2H6. Unexpectedly, the efficiency
of ethane oxidation in 75 mM H2SO4 was even lower than that of methane:
TONAcOH ¼ 41, TONHCOOH ¼ 44 for C2H6 vs TONHCOOH ¼ 223
for CH4.
Surprisingly, the oxidation of propane under the same reaction condi-
tions (except PPrH ¼ 20 bar at 60°C instead of 32 bar for CH4 and C2H6)
was even less efficient (69). Comparable amounts of formic
(TON ¼ 15.4), acetic (TON ¼ 5.9), and propionic acids as well as
propan-1-ol, propane-2-ol, propionaldehyde, and acetone were obtained
with an estimated total TON 60. Relative reactivity of C–H bonds in
the oxidation of CH4, C2H6, and C3H8 together with the values of their
C–H bond dissociation energy and solubility in water is given in Table 4.
The order of reactivity in the electrophilic reactions correlates with sta-
bilization of positive charges or follows the order of BDEC–H when radical
species are involved. For example, the normalized reactivity of the C–H
bonds of CH4, C2H6, and C3H8 in the oxidation by Tl(CF3COO)3–
CF3COOH system was 1:25:150 (70). The (FePctBu4)2N/SiO2–H2O2
system exhibits an unusual 1:1:1 reactivity in the oxidation of CH4,
C2H6, and C3H8 in water. The possible explanation is the formation of
an adduct between the diiron-oxo species and C2H6 or C3H8 molecules
before the oxidation event. A similar explanation for metal-oxo complexes
has recently been reported (71). Remarkably, the C–H bond in CH4 was
five times more reactive than the C–H bond in C2H6 when the oxidation
was performed in 75 mM H2SO4 (Table 4). This unusual reactivity of elec-
trophilic μ-nitrido diiron-oxo species is not yet understood. Further studies
are necessary to provide deeper insight into this reaction mechanism. The
Table 4 Comparison of the C–H Bond Dissociation Energies (BDE), Solubilities in Water
(1 bar, 20°C), and Reactivities of CH4, C2H6, and C3H8 in the Oxidation by (FePctBu4)2
N/SiO2–H2O2 System (68)
BDE Solubility in Reactivity on C–H Reactivity on C–H bond,a
Alkane (kcal mol21) H2O (mM) bond,a H2O 75 mM H2SO4
CH4 104.9 1.42 6.5 52.3
C2H6 101.4 1.87 7.7 10.5
C3H8 100.4 1.70 7.5 n.d.
a
Reactivity is defined as total turnover number normalized for the number of C–H bonds in alkane.
ARTICLE IN PRESS
nature of the support, strength, and loading of acidic groups should also be
studied to optimize activity and selectivity of these catalysts and to control
better C–C bond cleavage.
4.4 Oxidation of Benzene

The catalytic activity of μ-nitrido diiron complexes was evaluated in the oxi-
dation of aromatic compounds. The oxidation of benzene is an important
industrial process, and the oxidation of aromatic compounds is related to
essential biological functions. Several enzymes, e.g., cytochromes P-450,
nonheme iron mono- and dioxygenases, mediate hydroxylation of aro-
matics, which involves high-valent iron-oxo species (72,73). The main
mechanistic hypotheses for enzymatic benzene oxidation were discussed
by de Visser and Shaik (74). Characteristic features of the biological arene
hydroxylation are the intermediate formation of arene epoxide and the
migration of the substituent from the hydroxylation site to the adjacent posi-
tion (the so-called NIH shift arises from the National Institutes of Health
where this phenomenon was discovered). Valuable mechanistic information
was obtained in the course of the oxidation of benzene by the
(FePctBu4)2N–H2O2 system (75). Together with phenol, the principal
product obtained with a TON of 66, smaller amounts of benzene epoxide,
sym-oxepin oxide, and benzoquinone have been detected (Fig. 22).
Abundance
95,000
OH
90,000
85,000
80,000
75,000 O
70,000
65,000
60,000
55,000 O
50,000 O
45,000 O
40,000
35,000 O
30,000
25,000
20,000
15,000
10,000
5000
0
4.00 4.50 5.00 5.50 6.00 6.50 7.00 7.50 8.00 8.50 9.00 9.50
Time (min)
Fig. 22 GC-MS analysis of the oxidation of 1:1 mixture of C6H6 and C6D6 by
(FePctBu4)2N–H2O2 system. Isotopomers show slightly different retention times on
GC, deuterated compounds being eluted first. This phenomenon is well seen in the case
of 1,4-benzoquinone and phenol.
ARTICLE IN PRESS
Importantly, when the μ-oxo dimer (FePctBu4)2O or the mononuclear

FePctBu4 complex was used instead of (FePctBu4)2N, their bleaching
occurred within several minutes after the addition of H2O2, and benzene
epoxide was not formed. The observation of benzene epoxide and sym-
oxepin oxide in the (FePctBu4)2N–H2O2 system is particularly important.
Benzene epoxide is assumed to exist in rapid equilibrium with the tauto-
meric oxepin (76), which can be in turn epoxidized to sym-oxepin oxide
by P-450 monooxygenase (77). Consequently, in the first step,
(FePctBu4)2N–H2O2 system oxidizes benzene to benzene epoxide followed
by its isomerization to phenol similar to monooxygenase enzymes (Fig. 23).
The formation of a notable amount of sym-oxepin oxide derived from
the epoxidation of the oxepin, a tautomeric form of benzene epoxide, indi-
cates that the oxidation pathway involving epoxidation of benzene should
be the principal one.
Intermolecular kinetic isotope effects (kH/kD) on phenol formation in
the competitive oxidation of a 1:1 C6H6/C6D6 mixture were determined
to be 1.21 0.04 and 1.16 0.02 at 25°C and 60°C, respectively. The intra-
molecular kH/kD value obtained using benzene-1,3,5-d3 was 1.29 0.03 at
25°C. These kH/kD values are not compatible with the formation of a
σ-complex since in this case an inverse kH/kD (<1) should be observed
(72–74). According to a careful analysis of kinetic isotope effects associated
with aromatic oxidation, kH/kD 1.2 is compatible with reversible opening
of the epoxide ring (73). When labeled H218O2 (>90% isotopic enrichment)
was used as the oxidant in the presence of 16O2 and H216O, the Ph18OH
content was determined to be 93%, showing that H218O2 was the sole source
of oxygen in the product.
OH O
O
t
(FePc Bu4)2N–H2O2
MeCN (60°C)
TONPhOH = 66 O
O O O
Fig. 23 Oxidation of benzene by (FePctBu4)2N–H2O2 system via intermediate formation

of benzene epoxide.
ARTICLE IN PRESS
The formation of arene oxides in the biological systems involving high-

valent iron-oxo species is accompanied by the NIH shift. In order to find
evidence of this phenomenon, we have introduced 1,3,5-trideuterobenzene
as a mechanistic probe. This test is based on the analysis of the isotopic com-
position of 1,4-benzoquinone obtained by the oxidation of phenol as shown
in Fig. 24.
In the absence of an NIH shift, two isotopomeric phenols are formed.
Their further oxidation provides only quinone BQ-d2. In the alternative
H No NIH shift OH OH O
D D D D H H D D
+
H H H H D D H H
D D H O
PhOH-d3 PhOH-d2 BQ-d2
Epoxidation pathway with NIH shift
O H O H D
H
D D D O
D
D
H H H H H H
D D D
-D -H -D -H
K1 K2
OH OH H D
D H D D D OH D OH
H H H H H H H H
D NIH-shifted D D D NIH-shifted
PhOH-d2 PhOH-d3 PhOH-d2 PhOH-d3
O O O
D H D D D D
H H H H D H
O O O
NIH-shifted NIH-shifted
BQ-d2
BQ-d1 BQ-d3
6% 75% 19%
t
Fig. 24 Oxidation of 1,3,5-trideuterobenzene by (FePc Bu4)2N–H2O2 system involving
NIH shift. NIH-shifted products are indicated with yellow ovals.
ARTICLE IN PRESS
scenario, initially formed benzene epoxide undergoes a hydride/deuteride

shift resulting in two 2,4-cyclohexadien-1-ones K1 and K2. The enolization
of K1 and K2 affords four phenols, two of which are NIH-shifted products
with two adjacent H- or D-atoms. These NIH-shifted phenols give rise to
BQ-d1 and BQ-d3, which can only be formed according to this mecha-
nism. Analysis of the isotopic composition of 1,4-benzoquinone formed
in the (FePctBu4)2N–H2O2 system showed 19% of BQ-d1 and 6% of
BQ-d3, thus indicating the occurrence of an NIH shift.
Consequently, the mechanistic features of benzene oxidation by the
(FePctBu4)2N–H2O2 system, notably, exclusive incorporation of the
H218O2 oxygen to phenol, formation of benzene epoxide and NIH shift,
resemble those of biological oxidation (74,78). This finding confirms the
involvement of high-valent diiron-oxo-species in the oxidation reactions
mediated by μ-nitrido diiron complexes.
4.5 Reactivity of μ-Nitrido Diiron Complexes in Combination

With tBuOOH
Catalytic properties of μ-nitrido diiron complexes depend on the nature of
the oxidant. The reactivity of the (FePctBu4)2N–tBuOOH catalytic system
was investigated in the oxidation of C–H bonds in alkanes, olefins, aromatic,
and alkylaromatic compounds (79).
4.5.1 Oxidation of Cyclohexane

In the presence of 0.1 mM (FePctBu4)2N and 0.2 M tBuOOH the oxidation
of neat cyclohexane at 25°C proceeded with the turnover frequency of
3.9 min1. The monomeric complex (FePctBu4)Cl was much less efficient
in the cyclohexane oxidation giving cyclohexanol (Cy–OH) and cyclohex-
anone (Cy]O) in 1:1 ratio with the total TON of 27 and (FePctBu4)Cl
was bleached after 20 min. The dependence of the initial rates of the oxida-
tion indicates the first-order reaction with respect to the concentration of
(FePctBu4)2N. The rate of the cyclohexane oxidation did not depend on
the tBuOOH concentration between 0.05 and 0.5 M in the presence of
0.05 mM (FePctBu4)2N, indicating a zero-order with respect to tBuOOH.
However, at [tBuOOH] < 0.05 M, the reaction rate was affected by the oxi-
dant concentration.
Together with Cy–OH and Cy]O, obtained in comparable amounts as
the principal products, bicyclohexyl (Cy–Cy) was formed even in the pres-
ence of O2 and a high concentration of tBuOOH (Fig. 25).
ARTICLE IN PRESS
50
Product concentration (mM)

40
30
20
10
0
0 10 20
Reaction time (h)
Fig. 25 Accumulation of cyclohexanone (▲), cyclohexanol (●), and bicyclohexyl (■) in
the course of the oxidation of neat cyclohexane (9.26 M) at 25°C in the presence of
0.1 mM (FePctBu4)2N and 0.25 M tBuOOH.
This product of coupling of cyclohexyl radicals represented 8.0% of the

reaction products, while Cy]O and Cy–OH accounted for 50.4% and
41.6%, respectively. The formation of Cy–OH and Cy]O in a 1:1 ratio
is typically observed when Cy–O–O% radicals are involved. Two Cy–O–
O% molecules interact according to the Russell termination mechanism to
form the Cy–O–O–O–O–Cy intermediate, which generates Cy–OH,
Cy]O, and O2 upon degradation. The product composition did not
much change after the treatment of the reaction mixture with Ph3P
according to the Shul’pin method (80). It means that cyclohexyl hydroper-
oxide Cy–OOH was present in the reaction mixture only in minor
amounts. The total concentration of products was 79.2 mM after 6 h
and 98.4 mM after 24 h. The total TON attained 1063 cycles per catalyst
molecule (Cy–Cy originated from two Cy–H molecules). The product
yield based on the tBuOOH oxidant was 59%. The formation of the
Cy–Cy coupling product suggests the generation of significant amounts
of intermediate cyclohexyl radicals, which cannot be completely
converted to the oxidation products via quenching by oxidizing species.
Thus, the product composition of the oxidation of cyclohexane by the
(FePctBu4)2N–tBuOOH system is in agreement with a radical oxidation
ARTICLE IN PRESS
mechanism and involvement of t-butoxy radical tBuO% and (Pc)FeIV(μ-N)

FeIV]O(Pc) complex. These species can be formed from the homolytic
cleavage of the O–O bond of the peroxo complex (Pc)FeIV(μ-N)FeIII–
O–OtBu(Pc).
The oxidation of cyclohexane was more efficient in the presence of
0.5 M tBuOOH at 50°C. The product yield based on tBuOOH of 67%
and TON of 1877 was achieved after 14 h reaction. Cy–OH, Cy]O,
and Cy–Cy were obtained in a 38.7:54.9:6.4 ratio, and the concentrations
of Cy–OH, Cy]O, and Cy–Cy were 72.7, 103.0, and 12.0 mM, respec-
tively. Importantly, the (FePctBu4)2N complex was quite stable under these
reaction conditions. The UV–vis analysis of the reaction mixture showed
that only 10% of (FePctBu4)2N was destroyed after reaction.
A further improvement of the oxidation efficiency was achieved when
the oxidant was added in three portions, each introducing 0.25 M concen-
tration of tBuOOH in a 3-h interval. The portion method of tBuOOH
addition produced the higher product yields and TON (Table 5).
The product distribution was rather constant during the reaction. The
principal reaction products were Cy]O and Cy–OH achieving 324 and
191 mM concentrations after 23 h, respectively. Interestingly, the portion
method oxidant addition led to a decrease of the Cy–Cy amount down
to 2%–3%. The total TON progressively increased in the course of the reac-
tion to achieve almost 5300 turnovers. Remarkably, the oxidation of neat
cyclohexane has resulted in a 5.9% conversion. The yields of the oxidation
Table 5 Oxidation of Neat Cyclohexane by (FePctBu4)2N–tBuOOH Systema

Concentration (mM) Distribution (%)
Reaction Cy–OH:Cy]O:Cy– Yield on
t
time (h) Cy–OH Cy]O Cy–Cy Cy BuOOH (%)b Total TON
3 48.3 59.4 2.2 44:54:2 68 1099
6 117.8 141.9 5.0 44:54:2 81 2647
9 195.5 246.5 14.0 43:55:2 94 4560
23 191.2 323.7 14.7 36:61:3 114 5296
a
Conditions: 2 mL of cyclohexane (9.26 M), [(FePctBu4)2N] ¼ 0.1 mM, tBuOOH was added in three
portions at 0, 3, and 6 h, total [tBuOOH] ¼ 0.75 M, 50°C.
b
Yield was based on the amount of tBuOOH added before analysis taking into account that two oxidant
equivalents are consumed to produce Cy]O.
Reprinted with permission from Kudrik, E. V.; Sorokin, A. B. J. Mol. Catal. A Chem. 2017, 426,
499–505. Copyright 2017 Elsevier.
ARTICLE IN PRESS
products based on the tBuOOH amount were 68% (3 h), 81% (6 h), and
94% (9 h) achieving 114% after 23 h. This finding strongly suggests the
involvement of O2 in the cyclohexane oxidation.
4.5.2 Oxidation of Adamantane

The oxidation of 0.1 M adamantane solution in MeCN in the presence
of 0.1 mM (FePctBu4)2N and 0.25 M tBuOOH (catalyst:substrate:
oxidant ¼ 1:1000:2500) at 50°C led to a 67% conversion after 6 h. The ter-
tiary C–H bonds were much more reactive than the secondary C–H bonds:
1-adamantanol, 2-adamantanone, and 2-adamantanol were obtained with
57%, 8%, and 2% yields, respectively, with total TON of 670. The 3°/2°
parameter was determined to be 17 after correction for the number of C–
H bonds. It is generally believed that free-radical reactions show low
values of the 3°/2° parameter in adamantane oxidation (3°/2° < 6) (81)
although tBuO%-initiated reactions can exhibit the high 3°/2° values
(10) (82). The 3°/2° values greater than 13–15 are usually associated
with mechanisms involving metal-oxo species (83). Cytochrome P450
and heme complexes showed the 3°/2° adamantane regioselectivity as
high as 48:1 (84). The 3°/2° value of 17 obtained in the oxidation of
adamantane by the (FePctBu4)2N–tBuOOH system is in agreement with
simultaneous involvement of both (Pc)FeIV(μ-N)FeIV]O(Pc) and tBuO%
species.
An intramolecular kH/kD was determined in the oxidation of
adamantane-1,3-d2 (66). The analysis of the isotopic composition of the
adamantanol-1 provides a relatively high kH/kD value of 3.67 0.11. This
significant kinetic isotope effect is compatible with participation of both (Pc)
FeIV(μ-N)FeIV]O(Pc) and tBuO% species. The kH/kD values for the oxida-
tion of p-xylene and toluene by high-valent iron-oxo species of cytochrome
P-450 and by tBuO% were found to be very close: 6.4 vs 6.3 and 5.9 vs 6.0,
respectively (85).
4.5.3 Oxidation of Cyclohexene

The oxidation of neat cyclohexene afforded 2-cyclohexen-1-ol,
2-cyclohexen-1-one, 3-tert-butylperoxocyclohexene, and 3,30 -bicyclohexenyl
as principal reaction products, while the epoxide was obtained only in a
minor amount. The formation of the allylic oxidation products and 3,30 -
bicyclohexenyl due to the coupling of cyclohexenyl radicals strongly sug-
gests the participation of radical and one-electron oxidizing species in the
ARTICLE IN PRESS
OtBu
O OH O
0.2 mM (FePctBu4)2N
0.25 MtBuOOH
+ + +
60°C, 14 h
2 mL Air: 368 mM 125.2 mM 36.2 mM 6.2 mM

Ar: 38.2 mM 22 mM 107.2 mM 31.5 mM
Fig. 26 Product distribution in the oxidation of neat cyclohexene by the

(FePctBu4)2N–tBuOOH system under air and under inert atmosphere.
oxidation reaction. The yields of the products and their composition were
strongly affected by the presence or absence of dioxygen (Fig. 26).
The high total concentration of the oxidation products of 536 mM was
attained in the presence of air. In anaerobic conditions, the product concen-
tration decreased, down to 199 mM. In addition, quite different product
compositions were obtained depending on the presence of O2. Allylic alco-
hol (23.3%) and ketone (68.7%) were the dominant oxidation products in
the aerobic conditions, whereas 3-tert-butylperoxocyclohexene and 3,30 -
bicyclohexenyl accounted only for 6.8% and 1.2% of products, respectively.
In contrast, the allylic peroxide (53.9%) was the major product under an
inert atmosphere and the content of allylic alcohol and ketone decreased
to 11.1% and 19.2%, respectively. It is noteworthy that the concentration
of 2-cyclohexen-1-one dropped from 368 mM in the presence of air down
to 38.2 mM under Ar (a decrease by a factor of 9.6). The significant amount
of 3,30 -bicyclohexenyl (15.8%) arising from coupling of cyclohexenyl rad-
icals was formed (increase by a factor of 5.1 compared with aerobic condi-
tions). The total TON dropped from 2709 in air down to 1152 under Ar.
The yield of the oxidation products in the presence of air reached 361%
based on the tBuOOH amount. All these results are indicative of the signif-
icant involvement of O2 in the oxidation of cyclohexene.
4.5.4 Oxidation of Toluene

The (FePctBu4)2N–tBuOOH system carried out the oxidation of neat tol-
uene to benzyl alcohol, benzaldehyde, and benzoic acid with 10%, 52%, and
38% selectivity, respectively, and total TON of 1216. ortho- and para-cresols,
2-methylbenzoquinone, and a PhCH2–OOtBu coupling product were
detected only in very trace amounts. Thus, the (FePctBu4)2N–tBuOOH cat-
alytic system exhibits a clear preference for the aliphatic C–H bond oxida-
tion over the oxidation of aromatic C–H bonds.
ARTICLE IN PRESS
4.5.5 Oxidation of Benzene

The catalytic activity of the (FePctBu4)2N–tBuOOH system toward aro-
matic C–H bonds was found to be lower than that of (FePctBu4)2N–
H2O2. In addition, the different selectivity of the oxidation of benzene in
these systems was observed. While the (FePctBu4)2N–H2O2 system oxi-
dized benzene to phenol with the TON of 66 via benzene epoxide forma-
tion (Fig. 23) (75), the oxidation of neat benzene in the presence of 0.5 mM
(FePctBu4)2N and 0.35 M tBuOOH at 60°C afforded 14.5 mM of biphenyl
and less than 1.5 mM of phenol with the total TON of 32 turnovers. The
oxidation of a C6H6/C6D6 mixture (1:1 molar ratio) provided biphenol
with the following isotopic compositions: 47.7% of C6H5–C6H5, 45.0%
of C6H5–C6D5, and 7.3% of C6D5–C6D5, which were unchanged between
3 and 24 h (Fig. 27).
Thus, the kH/kD value in the oxidation of benzene by
(FePctBu4)2N–tBuOOH was calculated to be 2.36. This kinetic isotope
effect differs from the kH/kD value of 1.16 determined in the oxidation
of the C6H6/C6D6 mixture by the (FePctBu4)2N–H2O2 system at
60°C (75).
Again, a quite different behavior of the (FePctBu4)2N–tBuOOH system
in aromatic oxidation as compared to (FePctBu4)2N–H2O2 was observed.
While phenol was the principal product of benzene oxidation by the latter
system operating via a two-electron mechanism with the intermediate
H H H H
H H 47.7%
H H H H
+
H D H H D D
t
H H D D 0.5 mM (FePc Bu4)2N
0.35 M H2O2
+ H D 45.0%
60°C
H H D D
H D H H D D
+
D D D D
47.7 + 47.7 + 45.0
kH / kD = = 2.36 D D
45.0 + 7.3 + 7.3 7.3%
D D D D
Fig. 27 Oxidation of benzene to biphenyl by (FePctBu4)2N–tBuOOH system: determina-
tion of kinetic isotope effect.
ARTICLE IN PRESS
formation of benzene epoxide (75), biphenyl was obtained with a 91% selec-
tivity in the former case. Kinetic isotope effects in the oxidation of benzene
by tBuOOH and H2O2 catalyzed by (FePctBu4)2N were also different: 2.36
and 1.16, respectively. The tBuO% radical was claimed in a publication to be
incapable of the oxidation of benzene (86). Decomposition of tBuOOH in
benzene in the presence of the tBuOOtBu initiator resulted in the formation
of acetone, tBuOH, CO, CO2, and a small amount of O2 without products
of benzene oxidation. Consequently, the formation of the benzene oxida-
tion products in the (FePctBu4)2N–tBuOOH system suggests the involve-
ment of (Pc)FeIV(μ-N)FeIV]O(Pc) in the oxidation by abstracting of a
hydrogen atom from benzene. Phenyl radicals recombine to form the biphe-
nyl product since their recombination with alkoxyl radicals is slow.
4.5.6 Mechanism of Oxidation of Hydrocarbons by the

(FePctBu4)2N–tBuOOH System
The (FePctBu4)2N–tBuOOH system efficiently oxidizes C–H bonds of
alkanes, olefins, alkyl aromatic, and aromatic compounds in the presence
O2. Under anaerobic conditions, the same system performed the addition
of acetaldehyde to olefins, whereas the usual oxidation products were
obtained only in minor amounts (vide infra). It is noteworthy that the
(FePctBu4)2N–H2O2 system did not exhibit coupling reactivity. Both oxi-
dants are activated via initial formation of negatively charged diiron peroxo
complexes, Fe–N–Fe–OOH (A) and Fe–N–Fe–OOtBu (B), which were
detected by the ESI-MS technique (19). The different catalytic behavior
of (FePc)2N–H2O2 (strong oxidizing properties) and (FePc)2N–tBuOOH
(strong C–C forming ability in aerobic conditions) can be explained by
the formation of different species due to a different evolution of the
corresponding μ-nitrido diiron peroxo species (Fig. 28).
(Pc)FeIV(μ-N)FeIII(Pc)–O–OH undergoes heterolytic cleavage to pro-
duce a strong two-electron oxidant (Pc)FeIV(μ-N)FeIV(Pc+%)]O having
two redox equivalents above the initial Fe(+3.5)Fe(+3.5) state. This species
is competent for the methane oxidation and its reactivity is different from
that of the hydroxyl radical (16,41,67).
Contrary to the peroxo complex derived from H2O2, (Pc)FeIV(μ-N)
FeIII(Pc)–O–OtBu undergoes the homolytic O–O cleavage generating a
one-electron oxidized (Pc)FeIV(μ-N)FeIV(Pc)]O and the tBuO% radical.
The (FePc)2N–tBuOOH system exhibits a strong tendency toward the for-
mation of C–C bonds because both FeIV(μ-N)FeIV]O and tBuO% species
can abstract the hydrogen atom from C–H bonds of the substrates to
ARTICLE IN PRESS
A Heterolytic O–O cleavage

− −
OH OH
O O O O
.
Fe +3.5 Fe +3.5 Fe III Fe V Fe IV+
H2O2
N N N N N
−OH −
Fe +3.5 Fe +3.5 Fe IV Fe IV Fe IV
Strong oxidizing properties
B Homolytic O–O cleavage

− −
OtBu OtBu −
O O O
Fe +3.5 Fe +3.5 Fe III Fe IV
t
BuOOH . OtBu
N N N N +
Fe +3.5 Fe +3.5 Fe IV Fe IV
Formation of two one-electron oxidizing species

Fig. 28 Transformation of diiron peroxo complexes derived from H2O2 and tBuOOH to
different diiron-oxo species via heterolytic and homolytic O–O bond cleavage, respec-
tively. Reprinted from Kudrik, E. V.; Sorokin, A. B. J. Mol. Catal. A: Chem. 2017, 426,
499–505. Copyright 2017, with permission from Elsevier.
generate radicals, which then recombine to form coupling products. These

radicals are intercepted by O2 to form peroxo radicals, which initiate radical
oxidation reactions. The particular feature of the (FePc)2N–tBuOOH sys-
tem is the simultaneous generation of two active species capable of reacting
with C–H bonds. This increases the efficiency of the process. The proposed
reaction mechanism is shown in Fig. 29.
Formation of 3,30 -bicyclohexenyl in notable amounts even in the pres-
ence of O2 and high concentration of tBuOOH can be explained by the
proximal formation of two active species, (Pc)FeIV(μ-N)FeIV]O(Pc) and
t
BuO%, which rapidly react with surrounding cyclohexane molecules to
form two cyclohexyl radicals in close proximity to each other. These radicals
can recombine to give rise to a coupling product Cy–Cy or diffuse out of the
cage to be trapped by O2 to form a Cy–O–O% radical. The Cy% radicals can
also react with other oxidizing species. Cyclohexenyl hydroperoxide
CyOOH is an important product of this radical oxidation. In turn, it can
be activated by the μ-nitrido diiron complex to form (Pc)FeIV(μ-N)
FeIII(Pc)–O–OtCy followed by the formation of (Pc)FeIV(μ-N)
FeIV(Pc)]O and CyO% radical, which initiates further oxidation.
When 0.1 M cyclohexane solution was used, the Cy–Cy coupling prod-
uct was absent and Cy]O accounted for 72% of products despite a mod-
erate TON (87 turnovers). The predominant formation of Cy–OH can be
ARTICLE IN PRESS
t FeIVNFeIII–OH Cy–H
BuOOH
Recombination
FeIVNFeIVK O . in cage
Cy Cy–Cy
FeIVNFeIII–O–OtBu t . .
BuO Cy Cage
escape
t O2
BuOH Cy–H
IV III
Fe NFe
. .
CyOOH CyOO + OOCy
FeIVNFeIII–O–OCy .
Cy Cy–H Cy–OH + Cy K O + O2
.
Fe NFe K O
IV III CyO O2
Cy–H
CyOH
FeIVNFeIII–OH Cy–H
Fig. 29 Proposed mechanism of the oxidation of cyclohexane and other hydrocarbons
by (FePctBu4)2N–tBuOOH system. Phthalocyanine ligands are omitted for clarity. Species
reacting with C–H bonds are depicted in red and principal products are in blue color.
Reprinted with permission from Kudrik, E. V.; Sorokin, A. B. J. Mol. Catal. A: Chem. 2017,
426, 499–505. Copyright 2017 Elsevier.
expected under these conditions. In fact, the parallel formation of two adja-
cent radicals is disfavored by the lower cyclohexane concentration. Instead,
Cy–H molecule reacts with either (Pc)FeIV(μ-N)FeIV]O(Pc) or tBuO%
radical to form the Cy radical, which is then trapped by adjacent active spe-
cies to form an oxidation product.
The pronounced dependence of the yield and composition of
cyclohexene oxidation products in the presence of O2 (Fig. 26) confirms
the participation of O2 in the reaction. A sharp decrease of
2-cyclohexen-1-ol and 2-cyclohexen-1-one amounts observed in anaerobic
conditions, from 125.2 and 368 mM down to 22 and 38.2 mM, respectively,
indicates that these products should be formed through the reaction
pathways involving O2. In turn, a notable increase of 3-tert-buty-
lperoxocyclohexene and 3,30 -bicyclohexenyl contents under Ar points
out that the concurrent recombination of cyclohexenyl and tBuO% radicals
becomes a predominant pathway in the absence of O2. Thus, the
(FePctBu4)2N–tBuOOH system operates via one-electron oxidation path-
ways involving free radicals.
4.6 Oxidation of Alkylaromatic Compounds

Oxidations of alkylaromatic compounds are very important industrial pro-
cesses. For instance, terephthalic acid is produced by the oxidation of
ARTICLE IN PRESS
p-xylene at the scale of 44 million tons per year. The industrial process
involves aerobic radical oxidation in the presence of Co and Mn salts in
acetic acid at 200°C and 1.5 MPa pressure. Different catalytic approaches
have been developed in the search for more environmentally acceptable
and selective processes. Taking into account the high selectivity of μ-nitrido
diiron complexes in combination with tBuOOH to benzylic vs aromatic
oxidation, we have studied the complexes [FePc(SO2n–C6H13)4]2N and
[FePc(SO2tBu)4]2N bearing electron-withdrawing alkylsulfonyl substituents
in the oxidation of neat toluene and p-xylene at 20–60°C (40). Avoiding the
use of solvents is an important advantage from environmental and industrial
perspectives. Using 1 mM [FePc(SO2n–C6H13)4]2N solution, toluene was
oxidized to benzoic acid with 83% selectivity and a TON of 197 (Fig. 30).
PhCH2OH and PhCHO were minor products, whereas only traces of
p-cresol and 2-methyl-1,4-benzoquinone were detected. Due to the simul-
taneous generation of two active species according to the mechanism
described earlier (Fig. 29), the catalytic reaction was very efficient with only
a 0.01 mM catalyst loading affording TON ¼ 6570. It was noteworthy that
benzaldehyde was obtained as the principal product under these conditions.
The oxidation of p-xylene has afforded the products of benzylic oxida-
tion of only one methyl group with a TON of 587. When 1 mM [FePc
(SO2tBu)4]2N was used, p-toluic acid, p-tolualdehyde, and
4-methylbenzyl alcohol were obtained with 66%, 19%, and 15% selectivity,
respectively (Fig. 31) (40).
Again, the oxidation was very efficient with a low catalyst loading of
0.01 mM resulting in a TON of 13,300. Under these conditions, the selec-
tivity of the oxidation changed and p-tolualdehyde was obtained with 78%
selectivity.
CH3 CH2OH CHO COOH
+ +
[Cat] = 10−3 M 5% 12% 83%

TON = 197
[Cat] = 10−5 M 17% 65% 18%

TON = 6570
Fig. 30 Oxidation of neat toluene by [FePc(SO2n–C6H13)4]2N–tBuOOH system at 40°C
using [tBuOOH] ¼ 206 mM.
ARTICLE IN PRESS
CH3 CH2OH CHO COOH
+ +
CH3 CH3 CH3 CH3
[Cat] = 10−3 M 15% 19% 66%

TON = 587
[Cat] = 10−5 M 17% 78% 5%

TON = 13,300
Fig. 31 Oxidation of neat p-xylene by [FePc(SO2tBu)4]2N–tBuOOH system at 60°C.
It should be noted that in the presence of tBuOOH oxidant, μ-nitrido

diiron phthalocyanines exhibit the different selectivity of the oxidation of
alkylaromatic compounds compared with related μ-oxo compounds. While
μ-nitrido dimers furnish benzylic oxidation products with high selectivity,
μ-oxo diiron phthalocyanines provide high yields of quinones (87,88). The
selectivity of the alkylaromatics oxidation can also be adjusted by the choice
of the oxidant. μ-Nitrido diiron complexes catalyze either benzylic or aro-
matic cycle oxidations in the presence of tBuOOH or H2O2, respectively.
Thus, the appropriate choice of the catalyst and oxidant facilitates the
achievement of the desired selectivity of the oxidation of alkylaromatic
compounds.
4.7 Formation of C–C Bonds

The (FePctBu4)2N–tBuOOH system provides another interesting possibility
due to the efficient generation of radical species. As was shown earlier, in the
presence of O2, the main reaction products were oxidized compounds
(Fig. 26) (79). However, considerable quantities of C–C coupling products
were also obtained, even in the presence of O2. According to the proposed
mechanism (Fig. 29), a homolytic cleavage of the O–O bond in FeIVNFeIII–
O–OtBu affords two one-electron oxidizing species: FeIVNFeIV]O and
t
BuO%. They can be trapped by O2, leading to oxidation products or recom-
bine to furnish C–C coupling products. Obviously, the latter pathway
should be more efficient in anaerobic conditions.
Indeed, the (FePctBu4)2N–tBuOOH system produced a significant
amount of 3,30 -dicyclohexenyl during oxidation of cyclohexene
(Fig. 26). This C–C bond formation reactivity was extended to
ARTICLE IN PRESS
O
t CH3
0.01 mol% (FePc Bu4)2N
O 0.15 equiv. tBuOOH
CH3C + X X
H 60°C, Ar
Cyclic and linear olefins

olefins with oxidizable functions
alkenyl substituted aromatics
Fig. 32 Large scope hydroacylation of olefins with CH3CHO.
hydroacylation of unactivated cyclic or linear olefins by acetaldehyde to pro-

duce methyl ketones (Fig. 32) (89).
This is an important reaction with 100% atom efficiency: the addition of
acetaldehyde to the olefinic double bonds generates ketones containing all
atoms of both substrates. Importantly, the presence of (FePctBu4)2N is essen-
tial for the hydroacylation reactivity. When FeSO47H2O or mononuclear
FePctBu4 was used instead of (FePctBu4)2N, the products of usual radical
oxidation rather than addition products have been obtained. Optimization
of the reaction conditions led to a high selectivity to hydroacylation products
using a low catalyst loading (0.01 mol%) in the absence of organic solvent.
The scope of the (FePctBu4)2N–tBuOOH system was evaluated in the func-
tionalization of various olefins including substrates bearing functional groups
(Table 6).
The catalytic system was competent for hydroacylation of both cyclic
and linear alkenes. In the case of octene-1 and 2-cyclohexen-1-one only for-
mation of the anti-Markovnikov product was observed. Allylbenzene and
2-propen-1-ol were transformed to anti-Markovnikov and Markovnikov
products in 95/5 and 83/17 ratio, respectively. This method can also be used
for olefins having easily oxidizable positions, e.g., allylbenzene,
2-allylphenol, and 2-cyclohexen-1-one. In all cases, neither benzylic nor
allylic oxidation products were detected. No oxidation of allylphenol to
the corresponding quinone was observed. Olefins containing a readily oxi-
dizable hydroxyl group, e.g., allyl alcohol, can also be hydroacylated. Very
high TONs for formation of methyl ketones were attained
(TON 3600–5700). Unreacted acetaldehyde remaining in the reaction
mixture after reaction can easily be recycled and reused.
The remarkable reactivity of (FePctBu4)2N observed with tBuOOH and
not with H2O2 can be explained by the proposed reaction mechanism
(Fig. 33).
Homolytic cleavage of the O–O bond in (Pc)FeIV(μ-N)FeIII(Pc)–O–
O Bu provides (Pc)FeIV(μ-N)FeIV(Pc)]O and radical tBuO%, which can
t
ARTICLE IN PRESS
Table 6 Hydroacylation of Olefins by the (FePctBu4)2N–tBuOOH System Under Inert

Atmospherea
Substrate Product Conversion (%) Selectivity (%)b
O 69b 92
86 52
O
71 80
O
OH OH 71 51
O O 61 90
HO
O 62 76
HO
a
Reaction conditions: olefin (10 mmol), acetaldehyde (100 mmol), catalyst (1 μmol, 0.01 mol%),
60°C, 24 h.
b
0.1 mol% catalyst loading.
Fe+3.5NFe+3.5–O–OtBu
O
. X O
t
BuO O RC
RC FeIVNFeIV = O R H
H H
Inefficient
t O O
BuOH RC . FeIVNFeIII–O–H X O
RC.
R .
X A
Anti-Markovnikov addition
Fig. 33 Proposed mechanism for hydroacylation of olefins by the (FePctBu4)2N–tBuOOH

system.
ARTICLE IN PRESS
abstract hydrogen atom from cyclohexene and acetaldehyde to generate

cyclohexenyl radical (BDEallylic C–H bond ¼ 81.6 kcal mol1) and acyl radical
(BDEaldehyde C–H bond ¼ 86.0 kcal mol1), respectively. To limit the forma-
tion of cyclohexenyl radical leading to undesirable bicyclohexenyl, a 10-fold
excess of CH3CHO was used. Generated acyl radicals add to olefin in the
anti-Markovnikov fashion to form radical A. The hydrogen abstraction
from CH3CHO by A is inefficient. We proposed that this limiting step
can be carried out by FeIV(μ-N)FeIV–OH formed in the first step after
abstraction of the hydrogen atom from acetaldehyde. This affords the
hydroacylation product and regenerates the FeIV(μ-N)FeIV(Pc)]O oxi-
dant. μ-Nitrido diiron species accomplish two key reactions, which form
a catalytic cycle with no requirement for additional reagents. In such a
way, tBuOOH can be used in a catalytic amount (15 mol% to olefin) just
to initiate the reaction with 0.01 mol% catalyst loading to achieve 61%–
86% conversion (89). The reaction is particularly suitable for large-scale syn-
theses. This practical approach involving very small amount of nontoxic iron
catalyst is a clean alternative to traditional hydroacylation protocols
employing significant amounts of expensive Rh and Ru complexes.
4.8 Oxidative Dehalogenation

The (FePctBu4)2N–tBuOOH system is capable of reacting with CH2Cl2 and
CH2Br2 to form Cl–(Pc)FeIV(μ-N)FeIV(Pc+%)–Cl and Br–(Pc)FeIV(μ-N)
FeIV(Pc+%)–Br complexes with high isolated yields (46). According to
ESI-MS data, the formation of these species occurred via intermediate
monohalogenated complexes. This is a rare case of stable Fe(IV) complexes
bearing a macrocycle cation-radical. These complexes have two redox
equivalents above the Fe(III)Fe(IV) state and oxidize 2-mercaptoethanol
in a two-electron process (Fig. 34).
A narrow EPR signal of the phthalocyanine cation radical at g ¼ 2.0012
and one M€ ossbauer doublet (δ ¼ 0.10 mm s1, ΔEQ1 ¼ 1.64 mm s1) are
indicative of the structures with identical FeIV sites isoelectronic with high-
valent oxo-iron species. This ultra-high oxidation state was confirmed by
intense preedge energy doublet peaks in XANES spectra at 7115.2 and
7113.4 eV typical of FeIVFeIV phthalocyanine cation-radical scaffold
(48,62). The Fe K-edge EXAFS study of X–(Pc)FeIV(μ-N)FeIV(Pc+%)–X
revealed a symmetric Fe–N–Fe fragment with a Fe–Fe distance of
3.39(1) Å. The Fe–Cl and Fe–Br lengths were 2.33 and 2.54 Å, respectively
(Fig. 6). These complexes can oxidize 2-mercaptoethanol and
ARTICLE IN PRESS
2.2
2.5
2.0
1.8
635
Abs at 635 nm
1.6
2.0 1.4
1.2
1.0
0.8
1.5
Absorbance
0.6
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Molar ratio
1.0
690
0.5
531
0.0
400 500 600 700
l (nm)
Fig. 34 Spectral changes in the course of titration of 6.9 μM Cl–(Pc)FeIV(μ-N)FeIV
(Pc+%)–Cl with 2-mercaptoethanol. MeCN, 25°C. Inset: dependence of absorbance at
635 nm on the 2-mercaptoethanol/Cl–(Pc)FeIV(μ-N)FeIV(Pc+%)–Cl molar ratio.
2-mercaptophenol to corresponding disulfides as well as hydroquinone to

quinone via two-electron processes (46). The formation of Cl–(Pc)
FeIV(μ-N)FeIV(Pc+%)–Cl means that the cleavage of the C–Cl bond in
CH2Cl2 occurs; the latter is commonly considered as a very stable solvent
for performing oxidation reactions. Inspired by this unusual oxidative
dechlorination of CH2Cl2, we evaluated the reactivity of (FePctBu4)2N
toward various halogenated compounds. Quite remarkably, the rare oxida-
tive dehalogenation reactivity can be extended even to fluorinated aromatic
compounds.
4.9 Transformation of Aromatic C–F Bonds Under Oxidative

Conditions
Traditional approaches to the activation and transformation of C–F bonds
implicate electron-rich reagents, e.g., low- and zero-valent transition metal
complexes, strong nucleophiles, and reductants. Transformation of C–F
bonds under oxidative conditions using electron-deficient species is highly
challenging because fluorine is the most electronegative element.
ARTICLE IN PRESS
F 0
FeIII FeIV +
C6F6/C6H6 (1:1)
30 equiv. tBuOOH
N N
60°C, 6 h
FeIV FeIV
Isolated yield–80% F
IV IV +%
Fig. 35 Stoichiometric formation of [(Pc)(F)Fe (μ-N)Fe (F)(Pc )]. Ovals represent tetra-
tert-butylphthalocyanine.
Unexpectedly, μ-nitrido diiron complexes in combination with peroxides

are capable of reacting even with heavily fluorinated aromatic compounds
including C6F6 with BDEC–F ¼ 154 kcal mol1 (47). Incubation of
(FePctBu4)2N with 30 equiv. tBuOOH in C6F6/C6H6 mixture at 60°C
for 6 h led to the stoichiometric formation of the complex [(PctBu4)F–
FeIV(μ-N)FeIV–F(PctBu+4 %)] with 80% isolated yield (Fig. 35).
This complex was fully characterized by ESI-MS, UV–vis, EPR, 19F
NMR, EDX, EXAFS, XANES, and XES methods. Both disappearance
of the Q band in the UV–vis spectrum of the initial (FePctBu4)2N complex
at 640 nm and appearance of new bands at 549, 615, and 666 nm as well as
the change of the low-spin EPR signal at g ¼ 2.091 to the strong symmetric
narrow signal at g ¼ 2.0038 with the width of 8 G at 120K indicate the
phthalocyanine cation-radical and FeIVFeIV complex configuration.
XANES data confirmed this assignment. One signal in the 19F NMR spec-
trum is compatible with the presence of F. Energy-dispersive X-ray fluo-
rescence analysis of the complex revealed a 1:1 Fe:F ratio. EXAFS analysis
indicated coordination of fluoride anions to each iron site at 1.99 Å distance.
The ESI-MS spectrum showed a signal at m/z ¼ 1618.5 with an isotopic
cluster corresponding to [(FePctBu4)2N + F]+ ion composition formed after
the loss of fluoride anion under positive mode ESI-MS conditions.
The [(Pc)F–FeIV(μ-N)FeIV–F(Pc+%)] complex is stable under reaction
conditions and no further reaction occurs. To transform the stoichiometric
reaction to catalytic defluorination, [(Pc)F–FeIV(μ-N)FeIV–F(Pc+%)] should
be reduced back to (FePctBu4)2N. We found that the addition of H2O2 to
[(Pc)F–FeIV(μ-N)FeIV–F(Pc+%)] resulted in the regeneration of
(FePctBu4)2N. Thus, H2O2 acts as (i) oxidant to form oxo species which
mediates defluorination and (ii) reductant of the fluorinated complex to
ARTICLE IN PRESS
complete the catalytic cycle. The (FePctBu4)2N–H2O2 system transforms a

large array of fluorinated aromatics with various fluorination patterns includ-
ing perfluorinated and polyfluorinated compounds bearing different func-
tional groups with high TON (Fig. 36) (47).
Of particular importance was the finding that the (FePctBu4)2N–H2O2
system reacted with hexafluorobenzene to transform 3.12 from six fluorine
atoms to F. It should be noted that oxidative metabolism is inaccessible for
perfluorinated species via enzymatic pathways including cytochrome P-450
(90). In marked contrast to traditional organometallic and reductive
methods operating in strict inert and anhydrous conditions, the
(FePctBu4)2N–H2O2 system exhibited a high defluorination efficiency in
H2O: up to 4825 C–F bonds per catalyst molecule were cleaved with release
of F (Fig. 37) (47).
The defluorination activity in H2O was higher than that obtained in
MeCN (Figs. 36 and 37). The product composition depended on the sub-
strate:H2O2 ratio. A 16-fold oxidant excess resulted in significant mineral-
ization of fluorinated aromatics, whereas using 5 equiv. H2O2 led to a
preparative reaction. For instance, the oxidation of C6F5OH and
octafluoronaphthalene afforded difluoromaleic and tetrafluorophthalic acids
with 62% and 66% yields, respectively (47).
Conversion (TON)
(FePc)2N–H2O2 29% (226) 33% (152) 59% (341) 41% (336) 66% (351) 41% (169) 69% (240) 47% (120)
Fenton system 0% n.d. 0% 0% n.d. n.d. 0% 0%
OH CF3 NO2 CN COOH OCH3
Conversion (TON)
(FePc)2N–H2O2 82% (818) 30% (242) 45% (376) 20% (141) 31% (178) 40% (188)
Fenton system 7% (45) <2% (<15) <2% (<15) <4% (<9) <9% (70) <3% (<17)
O CH3
N
Conversion (TON) O
(FePc)2N–H2O2 99% (986) 45% (282) 15% (69) 29% (150)
Fenton system 98% (750) <7% (43) <2% (<17) 11% (78)
Fig. 36 Catalytic defluorination by (FePctBu4)2N–H2O2 system (CD3CN, 60°C, 15 h). The

catalyst:substrate:oxidant ratio was 1:250:4000. TONs were calculated as moles of Fd
per mole of catalyst.
ARTICLE IN PRESS
OH COOH SO3H F
F F F F F F F COOH
F F F F F F F COOH
Conversion (TON) F F F F
(FePc)2N – H2O2 98% (4825) 76% (2950) 53% (2140) 56% (1730)
Fig. 37 Catalytic heterogeneous defluorination by carbon-supported (FePctBu4)2N–
H2O2 system in D2O at 60°C for 15 h. The catalyst:substrate:oxidant ratio was
1:1000:26,000.
The counterintuitive defluorination of the C–F bonds via oxidation

pathway raises questions regarding the mechanistic background of this strik-
ing reactivity. Indeed, how can the highly electron-deficient (Pc)FeIV(μ-N)
FeIV(Pc+%)]O species react with the C–F bond formed by the most elec-
tronegative element? We have considered four mechanistic hypotheses:
(i) free-radical Fenton oxidation; (ii) nucleophilic substitution;
(iii) electrophilic attack, and (iv) initial epoxidation of an aromatic cycle.
(i) All fluorinated benzenes were completely stable under Fenton condi-
tions (Fig. 36). The substrates substituted with functional groups
showed conversions at the detection limit. Thus, these results are
not compatible with participation of OH% radicals.
(ii) The conversions of fluorinated benzenes became lower with the
increase of the number of fluorine substituents (Fig. 36). Other reac-
tivity trends (47) also disfavor this mechanism.
(iii) Transformation of 1,3,5-trichloro-2,4,6-trifluorobenzene furnished
phenols obtained via F and Cl elimination in a 97:3 ratio. Surprisingly,
oxidation of 1,3,5-trifluorobenzene resulted in phenols generated
from F elimination and C–H oxidation in a 98:2 ratio (Fig. 38).
Although ultra high-valent diiron-oxo species are powerful elec-
trophilic species, the observed reactivity of C–F vs C–Cl and C–H
bonds in intramolecular competition in transformation of C6F3Cl3
and C6F3H3 is not in agreement with electrophilic attack of oxo spe-
cies on the C–X bond.
(iv) Oxidation of aromatic (73,74,78) and halobenzene (91) compounds
mediated by cytochrome P450 Compound I occurs via intermediate
formation of arene epoxides accompanied by an NIH shift. The
(FePctBu4)2N–H2O2 system also performs benzene oxidation via ben-
zene epoxide with an NIH shift (75). Therefore, the initial formation
of fluoroarene epoxide seems to be a plausible route. A direct detection
of fluoroarene epoxides is hampered by their low stability at GC-MS
ARTICLE IN PRESS
F OH F
Cl Cl Cl Cl Cl Cl
+
F F F F F F
Cl Cl
OH
97% 3%
F OH F
H H
+
F F F F F F
H OH
98% 2%
Fig. 38 Relative reactivities of C–F bond vs C–Cl and C–H bonds.
OH F OH F
F OH
F F F 6% F 94%
Major product Minor NIH-shifted

product
Fig. 39 Detection of fluorine NIH shift in the oxidation of 1,4-difluorobenzene in MeCN
at 60°C after 25 min.
conditions and low ESI-MS response. Nevertheless, we were able to

detect epoxide in the oxidation of octafluoronaphthalene (47).
Although the NIH shift was reported for brominated and chlorinated
aromatics (92), the fluorine migration is far less documented (93)
because of preferential elimination of fluorine compared to its migra-
tion inducing the NIH shift (91). Among several tested fluoro-
benzenes, only 1,4-difluorobenzene resulted in detection of the
fluorinated NIH shift. Together with the p-fluorophenol main prod-
uct, the NIH-shifted 2,4-difluorophenol was identified (Fig. 39).
It is therefore conceivable that the initial step of the oxidative defluorination
is epoxidation of fluoroarene by the (Pc)FeIV(μ-N)FeIV(Pc+%)]O species
(Fig. 40).
ARTICLE IN PRESS
HF CO2 CO
Fig. 40 Proposed mechanism for the oxidative defluorination of fluorinated aromatics.

Identified intermediates and products are shown in red. Adapted with permission from
Colomban, C.; Kudrik, E. V.; Afanasiev, P.; Sorokin, A. B. J. Am. Chem. Soc. 2014, 136,
11321–11330. Copyright 2014 American Chemical Society.
The fluoroarene epoxide rearranges to phenol via a ketone intermediate

giving rise to the fluorine NIH shift. Further oxidation of phenol affords
tetrafluorobenzoquinone, which undergoes aromatic cycle cleavage via an
epoxidation–epoxide hydrolysis–HF elimination sequence to produce
difluoromaleic, fluorooxaloacetic, oxalic acids, as well as HF and CO2.
4.10 Degradation of Recalcitrant Pollutants

Chlorinated aromatic compounds are priority environmental contaminants
because of the large-scale applications such as biocides, lubricants, and sol-
vents, and their high resistance to biodegradation. In the environmental
context, fluorinated aromatic compounds have received less attention.
However, the large-scale production and application of fluorinated com-
pounds (40% of agrochemicals and 25% of drugs currently used contain
C–F bonds) (94) and their exceptional stability lead to their accumulation
in the environment.
The outstanding reactivity of the (FePctBu4)2N–H2O2 catalytic system
might be used for degradation of heavily halogenated pollutants resistant
to biodegradation and current remediation methods. Contrary to organo-
metallic and reductive methods operating in inert and anhydrous conditions,
the practical (FePctBu4)2N–H2O2 system can be advantageously used in the
presence of air and water. For example, particularly recalcitrant pen-
tafluorophenol was efficiently mineralized using supported catalyst in water.
ARTICLE IN PRESS
F
Supported catalyst COOD
F F F COOD O
0.2 µmol COOD
+ H2O2 DF + CO2 + CO + + +
D2O, 60°C, 15 h COOD
F F 5200 862 602 41 F COOD H COOD
114 mmol F
OH mmol mmol mmol mmol 33 mmol 7.4 mmol
200 mmol
Fluorine balance: 97%
96% conversion
Carbon balance: 89%
Fig. 41 Heterogeneous catalytic mineralization of C6F5OH in water. Adapted with per-
mission from Colomban, C.; Kudrik, E. V.; Afanasiev, P.; Sorokin, A. B. J. Am. Chem. Soc.
2014, 136, 11321–11330. Copyright 2014 American Chemical Society.
OH
0.5 mol% catalyst HOOC
20 equiv. H2O2 CO2 CO
Water, 60°C, 3 h
Conversion Dechlorination TOC

(FePcS)2N 92% 0.91 33%
FePcS 81% 0.81 27%
Fig. 42 Comparison of the catalytic activity of (FePcS)2N and FePcS in the degradation
of 2,6-dichlorophenol. Dechlorination is defined as a number of Cl formed per sub-
strate molecule. TOC, total organic carbon removal.
A 54% organic carbon loss due to the formation of CO2/CO and an 89%
defluorination of C–F have been achieved (Fig. 41) (47).
μ-Nitrido diiron catalysts showed promising results in the degradation of
chlorinated phenols (95). The homogeneous oxidation of 2,6-
dichlorophenol (DCP) and 2,4,6-trichlorophenol (TCP) in water using
readily accessible water-soluble μ-nitrido diiron tetrasulfophthalocyanine
(FePcS)2N (96) and H2O2 was more efficient than with mononuclear FePcS
in terms of conversion, dechlorination degree, and total organic carbon
removal (Fig. 42).
While (FePcS)2N was quite stable under reaction conditions, the blue
color of FePcS almost disappeared after 15 min. The degradation of DCP
in the presence of FePcS started only when the complex was completely
bleached after 1 h. In contrast, the reaction with (FePcS)2N was rapid
and 70%–80% DCP conversion was achieved after 1 h. The (FePcS)2N–
H2O2 system performed mineralization of DCP, and notable amounts of
CO2 and CO were found in the gas phase after reaction. The (FePcS)2N
catalyst was also more efficient in oxidation of TCP. After 6 h reaction, a
46% conversion of TCP and a 11% TOC removal were obtained at pH 7
with 2 mol% (FePcS)2N loading, while no TCP conversion (less than 1.5%)
ARTICLE IN PRESS
+ H2O
–H2O (–HCl)
FeNFe
FeNFeOO–
(OH) (OH)
(OH)
(OH)
[O]
H2O
–COx –CO2
(OH)
Fig. 43 Proposed mechanism of the degradation of chlorinated phenols by (FePcS)2N–

H2O2 system.
was obtained using FePcS. The mechanism of the degradation of chlorinated

phenols is proposed in Fig. 43.
A similar mechanism can be proposed for the degradation of the aromatic
cycle of the fluorinated compounds.
The catalytic activity of (FePcS)2N and (FePcS)2O was compared in the
degradation of Orange II in the presence of tBuOOH (97). Although the
initial catalytic activity of the μ-nitrido complex was lower than that of
its μ-oxo counterpart, the stability of (FePcS)2N was much higher and
the complete Orange II decomposition was obtained only with (FePcS)2N.
5. CONCLUSION AND OUTLOOK

The major goal of biomimetic studies in catalysis is the creation of new
efficient catalysts for clean chemical processes. We have introduced a novel
catalytic concept involving μ-nitrido diiron porphyrin-like complexes.
Despite the high Fe(III)Fe(IV) oxidation state, μ-nitrido dimers are capable
of forming ultra high-valent diiron-oxo species. Due to their particular
electronic structure, these complexes are much stronger oxidants than tra-
ditional biomimetic complexes and even exhibit reactivity not observed
ARTICLE IN PRESS
with enzymes, e.g., oxidative defluorination of perfluoroaromatics.

The high-valent diiron-oxo complexes (L)FeIV(μ-N)FeIV]O(L+%)
(L ¼ phthalocyanine or porphyrin) were obtained using H2O2 or
m-chloroperbenzoic acid at low temperatures and were characterized by
several spectroscopic techniques and isotopic labeling (16,41,49). DFT cal-
culations have also indicated particularly strong oxidizing properties of the
(L)FeIV(μ-N)FeIV]O(L+%) species and their high reactivity in the oxidation
of methane (48,49,98).
The structure of the oxidant determines the mechanism of its activation
by μ-nitrido diiron phthalocyanine, leading to different active species. Using
H2O2, a very strong two-electron oxidizing species PcFe(IV)μNFe(IV)]O
(Pc+%) is formed by heterolytic O–O bond cleavage of the peroxo complex
PcFe(IV)μNFe(IV)–OOH(Pc). In contrast, PcFe(IV)μNFe(IV)–OOtBu
(Pc) generated from tBuOOH undergoes a homolytic O–O cleavage to form
a one-electron oxidizing PcFe(IV)μNFe(IV)]O(Pc) and tBuO% radical. The
latter system exhibits a different reactivity. The (FePctBu4)2N–tBuOOH
catalytic system is efficient in the oxidation of C–H bonds of alkanes, olefins,
alkylaromatic, and aromatic compounds furnishing oxidation products with
high TONs within several hours at low catalyst loading. The efficiency of
this system can be explained by generation of two active species from one
peroxo complex molecule, which are able to react with different C–H
bonds, thus opening up interesting possibilities for catalytic oxidation.
Of particular significance is the discovery of the activation of aromatic
C–F bonds in oxidation conditions. This finding is not only of fundamental
but also of practical interest. For a long time, the principal objective of
organic fluorine chemistry has been the preparation of fluorinated com-
pounds for different applications. As a consequence of their large-scale uses
and persistence to biodegradation, fluorinated compounds have accumu-
lated in the environment. The development of remediation processes for
fluorinated emerging pollutants is becoming of primary importance. The
oxidative approach appears to be particularly promising because the
μ-nitrido diiron phthalocyanines and H2O2 constitute a tolerant to water
and air catalytic system toward a large range of substrates. In addition, this
catalytic system consists of an accessible, earth-abundant, and nontoxic
iron-based catalyst, available on a large scale and with the industrial, envi-
ronmentally compatible, H2O2 oxidant. The diiron complex can be readily
immobilized onto different supports. For all these reasons, we believe that
this novel approach might be useful for the development of novel remedi-
ation processes.
ARTICLE IN PRESS
The catalytic properties of μ-nitrido diiron complexes can be tuned by

appropriate structural modification. Higher activity of iron sites supported
by electron-donating phthalocyanine ligands was determined by combined
catalytic, spectroscopic, and DFT studies using an intramolecular approach
(49). Different reactivities of μ-nitrido diiron complexes in the presence of
H2O2 and tBuOOH points to their rich redox chemistry, which can be used
for diverse catalytic reactions including C–C bond formation. In addition to
new and versatile reactivity, iron phthalocyanines can be available in bulk
quantities that suggest potential industrial applications.
The structural simplicity and flexibility of μ-nitrido dimers make them
promising catalysts for challenging reactions. This platform has a great
potential for further development. Different macrocyclic ligands including
porphyrins, phthalocyanines, porphyrazines, corrolazines, and corroles can
be used to accommodate homometallic and heterometallic M–X–M entities
with different transition metals, and probably, connecting groups. Provided
that synthetic challenges will be overcome, a variety of homoleptic and het-
eroleptic complexes can be prepared. The structural modification should
significantly affect the electronic structure and can be used to tune the cat-
alytic properties for access to improved and tailored catalysts for a variety of
reactions. We hope that the unusual properties of μ-nitrido diiron macro-
cyclic complexes will stimulate further investigations to discover new pos-
sibilities in the application of this novel tool in catalysis.
ACKNOWLEDGMENTS
The author is grateful to all coworkers whose names are given in the references for their
valuable contributions to the development of this topic. Dr. P. Afanasiev and Dr. E. V.
Kudrik are particularly acknowledged for the fruitful collaboration and stimulating
discussions. Research support was provided by the Agence Nationale de Recherche
(ANR, France, Grants ANR-08-BLANC-0183-01 and ANR-16-CE29-0018-01), by
CNRS and RFBR (PICS project 6295, RFBR project 14-03-91054), and by Region
Rh^ one-Alpes.
REFERENCES
1. Ortiz de Montellano, P. R. Chem. Rev. 2010, 110, 932–948.
2. Tinberg, C. E.; Lippard, S. J. Acc. Chem. Res. 2011, 44, 280–288.
3. Banerjee, R.; Proshlyakov, Y.; Lipscomb, J. D.; Proshlyakov, D. A. Nature 2015, 518,
431–434.
4. Kawakami, N.; Shoji, O.; Watanabe, Y. Chem. Sci. 2013, 4, 2344–2348.
5. Meunier, B.; Robert, A.; Pratviel, G.; Bernadou, J. In: The Porphyrin Handbook;
Kadish, K. M., Smith, K. M., Guilard, R., Eds.; Vol. 4, Academic Press: New York,
2000; pp 119–187.
ARTICLE IN PRESS
6. McLain, J. L.; Lee, J.; Groves, J. T. In: Biomimetic Oxidations Catalyzed by Transition Metal
Complexes; Meunier, B. Ed.; Imperial College Press: London, 2000; pp 91–169.
7. Costas, M. Coord. Chem. Rev. 2011, 255, 2912–2932.
8. Costas, M.; Mehn, M. P.; Jensen, M. P.; Que, L., Jr. Chem. Rev. 2004, 104, 939–986.
9. Bryliakov, K. P.; Talsi, E. P. Coord. Chem. Rev. 2014, 276, 73–96.
10. Tshuva, E. Y.; Lippard, S. J. Chem. Rev. 2004, 104, 987–1012.
11. Olivo, G.; Cussό, O.; Costas, M. Chem. Asian J. 2016, 11, 3148–3158.
12. Talsi, E. P.; Bryliakov, K. P. Coord. Chem. Rev. 2012, 256, 1418–1434.
13. Lindhorst, A. C.; Haslinger, S.; K€ uhn, F. E. Chem. Commun. 2015, 51, 17193–17212.
14. Ryabov, A. D.; Collins, T. J. Adv. Inorg. Chem. 2009, 61, 471–521.
15. Friedle, S.; Reisner, E.; Lippard, S. J. Chem. Soc. Rev. 2010, 39, 2768–2779.
16. Sorokin, A. B.; Kudrik, E. V.; Bouchu, D. Chem. Commun. 2008, 2562–2564.
17. Nam, W.; Han, H. J.; Oh, S.-Y.; Lee, Y. J.; Choi, M. H.; Han, S.-Y.; Kim, C.;
Woo, S. K.; Shin, W. J. Am. Chem. Soc. 2000, 122, 8677–8684.
18. Floris, B.; Donzello, M. P.; Ercolani, C. In The Porphyrin Handbook; Kadish, K. M.,
Smith, K. M., Guilard, R., Eds.; Vol. 18, Elsevier Science: San Diego, 2003; pp 1–62.
19. Afanasiev, P.; Sorokin, A. B. Acc. Chem. Res. 2016, 49, 583–593.
€ Dumoulin, F.; Sorokin, A. B.; Ahsen, V. Turk. J. Chem. 2014, 38, 923–949.
20. İşci, U.;
21. Summerville, D. A.; Cohen, I. A. J. Am. Chem. Soc. 1976, 98, 1747–1752.
22. Scheidt, W. R.; Summerville, D. A.; Cohen, I. A. J. Am. Chem. Soc. 1976, 98,
6623–6628.
23. Kadish, K. M.; Bottomley, L. A.; Brace, J. G.; Winograd, N. J. Am. Chem. Soc. 1980,
102, 4341–4344.
24. Tatsumi, K.; Hoffmann, R. J. Am. Chem. Soc. 1981, 103, 3328–3341.
25. Schick, G. A.; Bocian, D. F. J. Am. Chem. Soc. 1983, 105, 1830–1838.
26. Bocian, D. F.; Findsen, E. W.; Hofmann, J. A.; Schick, G. A.; English, D. R.;
Hendrickson, D. N.; Suslick, K. S. Inorg. Chem. 1984, 23, 800–807.
27. Goedken, V. L.; Ercolani, C. J. Chem. Soc. Chem. Commun. 1984, 378–379.
28. Bottomley, L. A.; Gorce, J.-N.; Goedken, V. L.; Ercolani, C. Inorg. Chem. 1985, 24,
3733–3737.
29. Kennedy, B. J.; Murray, K. S.; Homborg, H.; Kalz, W. Inorg. Chim. Acta 1987, 134,
19–21.
30. Li, M.; Shang, M.; Ehlinger, N.; Schulz, C. E.; Scheidt, W. R. Inorg. Chem. 2000, 39,
580–583.
31. Stuzhin, P. A.; Latos-Grazynski, L.; Jezierski, A. Transit. Met. Chem. 1989, 14, 341–346.
32. Stuzhin, P. A.; Hamdush, M.; Homborg, H. Mendeleev Commun. 1997, 7, 196–198.
33. Ercolani, C.; Hewage, S.; Heucher, R.; Rossi, G. Inorg. Chem. 1993, 32, 2975–2977.
34. Ercolani, C.; Jubb, J.; Pennesi, G.; Russo, U.; Trigiante, G. Inorg. Chem. 1995, 34,
2535–2541.
35. Donzello, M. P.; Ercolani, C.; Kadish, K. M.; Ou, Z.; Russo, U. Inorg. Chem. 1998, 37,
3682–3688.
36. Zanotti, G.; Mattioli, G.; Notarantonio, S.; Paoletti, A. M.; Rossi, G.; Bonapasta, A. A.;
Pennesi, G. Macroheterocycles 2011, 4, 161–163.
37. J€ustel, T.; M€uller, M.; Weyherm€ uller, T.; Kressl, C.; Bill, E.; Hildebrandt, P.;
Lengen, M.; Grodzinski, M.; Trautwein, A. X.; Nuber, B.; Wieghardt, K. Chem.
Eur. J. 1999, 5, 793–810.
38. Meyer, K.; Bill, E.; Mienert, B.; Weyherm€ uller, T.; Wieghardt, K. J. Am. Chem. Soc.
1999, 5, 4859–4876.
39. Colomban, C.; Kudrik, E. V.; Tyurin, D. V.; Albrieux, F.; Nefedov, S. E.; Afanasiev, P.;
Sorokin, A. B. Dalton Trans. 2015, 44, 2240–2251.
€ Afanasiev, P.; Millet, J.-M. M.; Kudrik, E. V.; Ahsen, V.; Sorokin, A. B. Dalton
40. İşci, U.;
Trans. 2009, 7410–7420.
ARTICLE IN PRESS
41. Kudrik, E. V.; Afanasiev, P.; Alvarez, L. X.; Blondin, G.; Clemancey, M.; Latour, J.-M.;
Bouchu, D.; Albrieux, F.; Nefedov, S. E.; Sorokin, A. B. Nat. Chem. 2012, 4,
1024–1029.
42. Li, M.; Scheidt, W. R. J. Porphyr. Phthalocyanines 2014, 18, 380–384.
43. Donzello, M. P.; Ercolani, C.; Russo, U.; Chiesi-Villa, A.; Rizzoni, C. Inorg. Chem.
2001, 40, 2963–2967.
44. Moubaraki, M.; Benlian, D.; Baldy, A.; Pierrot, M. Acta Crystallogr. C. 1989, 45,
393–394.
45. Kienast, A.; Homborg, H. Z. Anorg. Allg. Chem. 1998, 624, 233–238.
46. Afanasiev, P.; Bouchu, D.; Kudrik, E. V.; Millet, J.-M. M.; Sorokin, A. B. Dalton Trans.
2009, 9828–9836.
47. Colomban, C.; Kudrik, E. V.; Afanasiev, P.; Sorokin, A. B. J. Am. Chem. Soc. 2014, 136,
11321–11330.
48. Colomban, C.; Kudrik, E. V.; Briois, V.; Shwarbrick, J. C.; Sorokin, A. B.; Afanasiev, P.
Inorg. Chem. 2014, 53, 11517–11530.
€ Faponle, A. S.; Afanasiev, P.; Albrieux, F.; Briois, V.; Ahsen, V.; Dumoulin, F.;
49. İşci, U.;
Sorokin, A. B.; de Visser, S. P. Chem. Sci. 2015, 6, 5063–5075.
50. English, D. R.; Hendrickson, D. N.; Suslick, K. S. Inorg. Chem. 1993, 32, 367–368.
51. Kudrik, E. V.; Afanasiev, P.; Bouchu, D.; Sorokin, A. B. J. Porphyr. Phthalocyanines 2011,
15, 583–591.
€ Dumoulin, F.; Ahsen, V.; Sorokin, A. B. J. Porphyr. Phthalocyanines 2010, 14,
52. İşci, U.;
324–334.
53. Schick, G. A.; Findsen, E. W.; Bocian, D. F. Inorg. Chem. 1982, 21, 2885–2887.
54. Afanasiev, P.; Kudrik, E. V.; Millet, J.-M. M.; Bouchu, D.; Sorokin, A. B. Dalton Trans.
2011, 40, 701–710.
55. Walker, F. A.; Reis, D.; Balke, V. L. J. Am. Chem. Soc. 1984, 106, 6888–6898.
56. De Groot, F. Chem. Rev. 2001, 101, 1779–1808.
57. Glatzel, P.; Bergmann, U. Coord. Chem. Rev. 2005, 249, 65–95.
58. Newcomb, M.; Halgrimson, J. A.; Horener, J. H.; Wasinger, E. C.; Chen, L. X.;
Sligar, S. G. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 8179–8184.
59. Yosca, T. H.; Rittle, J.; Krest, C. M.; Onderko, E. L.; Silakov, A.; Calixto, J. C.;
Behan, R. K.; Green, M. T. Science 2013, 342, 825–829.
60. Chandrasekaran, P.; Stieber, S. C. E.; Collins, T. J.; Que, L., Jr.; Neese, F.; DeBeer, S.
Dalton Trans. 2011, 40, 11070–11079.
61. de Oliveira, F. T.; Chanda, A.; Banerjee, D.; Shan, X.; Mondal, S.; Que, L., Jr.;
Bominaar, E. L.; M€ unck, E.; Collins, T. J. Science 2007, 315, 835–838.
62. Kudrik, E. V.; Safonova, O.; Glatzel, P.; Swarbrick, J. C.; Alvarez, L. X.; Sorokin, A. B.;
Afanasiev, P. Appl. Catal. B 2012, 113–114, 43–51.
63. Theodoridis, A.; Maigut, J.; Puchta, R.; Kudrik, E. V.; van Eldik, R. Inorg. Chem. 2008,
47, 2994–3013.
64. Nam, W. Acc. Chem. Res. 2007, 40, 522–531.
65. Afanasiev, P.; Kudrik, E. V.; Albrieux, F.; Briois, V.; Koifman, O. I.; Sorokin, A. B.
Chem. Commun. 2012, 48, 6088–6090.
66. Sorokin, A.; Robert, A.; Meunier, B. J. Am. Chem. Soc. 1993, 115, 7293–7299.
67. Sorokin, A. B.; Kudrik, E. V.; Alvarez, L. X.; Afanasiev, P.; Millet, J. M. M.; Bouchu, D.
Catal. Today 2010, 157, 149–154.
68. Alvarez, L. X.; Sorokin, A. B. J. Organomet. Chem. 2015, 793, 139–144.
69. Kudrik, E. V.; Afanasiev, P.; Bouchu, D.; Millet, J.-M. M.; Sorokin, A. B. J. Porphyr.
Phthalocyanines 2008, 12, 1078–1089.
70. Hashiguchi, B. G.; Konnick, M. M.; Bischof, S. M.; Gustafson, S. J.; Devarajan, D.;
Gunsalus, N.; Ess, D. H.; Periana, R. A. Science 2014, 343, 1232–1237.
ARTICLE IN PRESS
71. Garcia-Bosch, I.; Company, A.; Cady, C. W.; Styring, S.; Browne, W. R.; Ribas, X.;
Costas, M. Angew. Chem. Int. Ed. 2011, 50, 5648–5653.
72. Korzekwa, K. R.; Swinney, D. C.; Trager, W. F. Biochemistry 1989, 28, 9019–9027.
73. Mitchell, K. H.; Rogge, C. E.; Gierahn, T.; Fox, B. G. Proc. Natl. Acad. Sci. U.S.A.
2003, 100, 3784–3789.
74. de Visser, S. P.; Shaik, S. J. Am. Chem. Soc. 2003, 125, 7413–7424, and references
therein.
75. Kudrik, E. V.; Sorokin, A. B. Chem. Eur. J. 2008, 14, 7123–7126.
76. Boyd, D. R.; Sharma, N. D. Chem. Soc. Rev. 1996, 25, 289–296.
77. Bleasdale, C.; Cameron, R.; Edwards, C.; Golding, B. T. Chem. Res. Toxicol. 1997, 10,
1314–1318.
78. Jerina, D. M.; Daly, J. W. Science 1974, 185, 573–582.
79. Kudrik, E. V.; Sorokin, A. B. J. Mol. Catal. A Chem. 2017, 426, 499–505.
80. Shul’pin, G. B. J. Mol. Catal. A Chem. 2002, 189, 39–66.
81. Olivo, G.; Lanzalunga, O.; Di Stefano, S. Adv. Synth. Catal. 2016, 358, 843–863.
82. Barton, D. H. R.; Beck, A. H.; Taylor, D. K. Tetrahedron 1995, 51, 5245–5254.
83. Costas, M.; Chen, K.; Que, L., Jr. Coord. Chem. Rev. 2000, 200–202, 517–544.
84. Groves, J. T.; Nemo, T. E. J. Am. Chem. Soc. 1983, 105, 6243–6248.
85. Manchester, J. I.; Dinnocenzo, J. P.; Higgins, L. A.; Jones, J. P. J. Am. Chem. Soc. 1997,
119, 5069–5070.
86. Hiatt, R. R.; Mill, T.; Casteman, J. K. J. Org. Chem. 1968, 33, 1421–1428.
87. Zalomaeva, O. V.; Sorokin, A. B. New J. Chem. 2006, 30, 1768–1773.
88. Perollier, C.; Pergrale-Mejean, C.; Sorokin, A. B. New J. Chem. 2005, 29, 1400–1403.
89. Kudrik, E. V.; Alvarez, L. X.; Sorokin, A. B. Chem. Eur. J. 2011, 17, 9298–9301.
90. Hackett, J. C.; Sanan, T. T.; Hadad, C. M. Biochemistry 2007, 46, 5924–5940.
91. Rietjens, I. M.; den Besten, C.; Hanzlik, R. P.; van Bladeren, P. J. Chem. Res. Toxicol.
1997, 10, 629–635.
92. Bogaards, J. J. P.; Van Ommen, B.; Wolf, C. R.; Van Bladeren, P. J. Toxicol. Appl.
Pharmacol. 1995, 132, 44–52.
93. Rietjens, I. M. C. M.; Vervoort, J. Xenobiotica 1989, 19, 1297–1305.
94. Grushin, V. V. Acc. Chem. Res. 2010, 43, 160–171.
95. Colomban, C.; Kudrik, E. V.; Afanasiev, P.; Sorokin, A. B. Catal. Today 2014, 235,
14–19.
96. Stuzhin, P. A.; Ivanova, S. S.; Dereven’kov, I.; Makarov, S. V.; Silaghi-Dumitresku, R.;
Homborg, H. Macroheterocycles 2012, 5, 175–177.
97. Makarova, A. S.; Kudrik, E. V.; Makarov, S. V.; Koifman, O. I. J. Porphyr. Phthalocya-
nines 2014, 18, 604–613.
98. Quesne, M. G.; Senthilnathan, D.; Singh, D.; Kumar, D.; Maldivi, P.; Sorokin, A. B.;
de Visser, S. P. ACS Catal. 2016, 6, 2230–2243.
CHAPTER FOUR
The Role of Nonheme Transition

Metal-Oxo, -Peroxo, and
-Superoxo Intermediates in
Enzyme Catalysis and Reactions
of Bioinspired Complexes
Abayomi S. Faponle*,†, Sam P. de Visser*,1
*Manchester Institute of Biotechnology, School of Chemical Engineering and Analytical Science,
The University of Manchester, Manchester, United Kingdom
†
Faculty of Basic Medical Sciences, Obafemi Awolowo College of Health Science, Olabisi Onabanjo
University, Ogun State, Nigeria
1
Corresponding author: e-mail address: sam.devisser@manchester.ac.uk
Contents
1. Introduction 168
2. Metal-(Di)oxygen Intermediates 173
2.1 Short-Lived Iron-Dioxygen Complex in Cysteine Dioxygenase Enzymes 174
2.2 Iron(III)-Hydroperoxo in Heme and Nonheme Complexes 177
2.3 Spin-State Reactivity in Biomimetic Manganese(V)-Oxo Complexes 181
2.4 Aldehyde Deformylation by Manganese(III)-Peroxo Complexes 185
3. Conclusions 190
Acknowledgments 190
References 190
Abstract
Transition metals are common cofactors in enzymes and enable catalysis to take place
via reaction barriers that are accessible at room temperature. Oxygen-activating
metalloenzymes are versatile species in Nature involved in vital processes ranging from
biodegradation to biosynthesis. Since oxygen-activating intermediates are not readily
amenable to experimental study, research has started to focus on biomimetic model
systems that have the active site coordination sphere and structural features, but react
in solution. In our research group, we have been involved in computational modeling of
heme and nonheme iron dioxygenases as well as biomimetic models of these com-
plexes. In this contribution, an overview is given on recent results of the characterization
and reactivity patterns of metal-oxo, metal-peroxo, and metal-superoxo complexes. In
particular, in recent studies attempts were made to trap and characterize the short-lived
oxygen-bound intermediate in the catalytic cycle of cysteine dioxygenase. Many

168 Abayomi S. Faponle and Sam P. de Visser
suggested structures could be ruled out by theoretical considerations, yet these also
provided suggestions of possible candidates for the experimentally observed spectra.
In addition, we review recent studies on the nonheme iron(III)-hydroperoxo species
and how its reactivity patterns with arenes are dramatically different from those found
for heme iron(III)-hydroperoxo species. In the final two sections there is a description,
with illustrations, of a series of computational studies on manganese(V)-oxo and side-on
manganese(III)-peroxo moieties that identify a unique spin-state reactivity pattern with
a surprising product distribution.
1. INTRODUCTION
Oxygen activation by enzymes is one of the most common biochem-
ical reactions in Nature whereby one or both oxygen atoms of O2 are
donated into a substrate. Many of these monooxygenase and dioxygenase
enzymes employ transition metal centers to perform this catalytic reac-
tion (1–7). In monooxygenases, one oxygen atom of O2 is inserted into
the substrate, whereas the other oxygen atom is reduced to water (8–12),
while in dioxygenases both oxygen atoms are inserted into one or more sub-
strates (3,6,13,14). As molecular oxygen is in a triplet spin ground state and
organic substrates typically are in a singlet spin state, it has been hypothesized
that these enzymes enable an otherwise spin-forbidden reaction (4). These
oxidants enable this reaction due to the involvement of a transition metal
with virtual or partially occupied valence orbital subshells. However, in
recent years, several cofactor-independent dioxygenases have been trapped
and characterized that are able to transfer triplet dioxygen into a closed-shell
singlet substrate without the assistance of a transition metal center. For
instance, the enzyme 1-H-3-hydroxy-4-oxoquinaldine 2,4-dioxygenase
performs an intradiol-type dioxygenation of an aromatic substrate in the
absence of cofactor although the reaction is slow (15,16). Therefore, cofac-
tors, such as transition metals in metalloenzymes, are expected to speed up
biochemical reactions significantly by lowering the rate-determining energy
barrier in a reaction mechanism. Metalloenzymes can be mononuclear or
binuclear or contain more than two metal centers (17); however, in this
review we will focus on mononuclear iron complexes only.
A common structural motif in mononuclear nonheme iron dioxygenases is
an active site where the iron is bound to two histidine and one carboxylate
group of an amino acid side chain in a facial 2-His/1-Asp ligand feature
(7,18). In Fig. 1, the active site of the nonheme iron dioxygenase taurine/
α-ketoglutarate (α-KG)-dependent dioxygenase (TauD) is depicted and
shows its typical structural motif with the metal bound to the 2-His/1-Asp
Reactions of Bioinspired Complexes 169
Fig. 1 The active site structure of TauD with the typical 2-His/1-carboxylate ligand motif
and with α-ketoglutarate (α-KG) and taurine bound.
ligand environment, where the iron interacts with the side chains of His99,
Asp101, and His255. The structure shown in Fig. 1 is taken from the protein
data bank with identifier 1OS7 (19). In the active site, the metal iron center
is in a pentadentate coordination and thus is anchored to the two imidazolium
groups of the posttranslationally modified histidine residues and the carboxylic
group of aspartate. The fourth and fifth ligand positions of the metal are occu-
pied by the carboxylic acid and keto group of α-KG cosubstrate. TauD is
involved in the hydroxylation of taurine, which is a step in the biodegradation
of cysteine in the body.
TauD undergoes a catalytic cycle (Fig. 2) that starts off from an iron(II)
resting state (A), where the metal is bound to the 2-His/1-Asp ligand system
and three water molecules. Binding of α-KG displaces two water molecules
and subsequently dioxygen replaces the final one to form the elusive
iron(III)-superoxo complex (B). This intermediate reacts with α-KG by
decarboxylation and the formation of an iron(IV)-oxo (C), succinate, and
CO2. The iron(IV)-oxo species of TauD has been characterized with
UV–Vis, electron paramagnetic resonance (EPR), and M€ ossbauer spectro-
scopic measurements and shown to be in a quintet spin ground state (20).
Kinetic isotope effect (KIE) studies using deuterated taurine gave evidence
of a rate-determining hydrogen atom abstraction leading to an iron(III)-
hydroxo intermediate (D) that after OH rebound gives the alcohol product
complexes. Computational studies, initially with small active site models of
TauD, confirmed the catalytic cycle and showed the oxygen activation to be
Fig. 2 Catalytic cycle of TauD with key intermediates highlighted.
a strong exothermic process with the most stable structure being the
iron(IV)-oxo intermediate (21,22). Subsequently, quantum mechanics/
molecular mechanics (QM/MM) studies on nonheme iron dioxygenases
TauD and AlkB (the latter are DNA repair enzymes) were performed that
took the complete protein environment into consideration (23,24).
Although the protein environment affects the barrier heights along the
reaction mechanism dramatically, the work on TauD with QM/MM
vs active site models gave the same overall conclusions. However, AlkB
was shown to proceed with a catalytic cycle that showed a major differ-
ence to that found for TauD; namely, the iron(IV)-oxo species was found
to undergo a 90-degree rotation during the catalytic cycle prior to
the hydrogen atom abstraction step. It was hypothesized that this unusual
rotation is essential to keep the dioxygen binding and substrate binding
separate (Fig. 2).
Further computational modeling of the high-valent iron(IV)-oxo inter-
mediate for nonheme iron dioxygenases including TauD and prolyl-4-
hydroxylase was performed in our group (25–30). The studies show that
the pentacoordinate nonheme iron(IV)-oxo complex is an extremely pow-
erful oxidant that reacts with substrates with significantly lower hydrogen
atom abstraction barriers than those for heme iron(IV)-oxo cation radical
(Compound I (CpdI)) species as found in cytochrome P450.
Interestingly, in the nonheme iron dioxygenases cysteine dioxygenase

(CDO) (31,32) and ergothioneine synthase (33,34), the metal is bound
to a 3-His motif rather than 2-His/1-Asp. An extensive set of computational
studies on CDO enzymes (35–37) and model complexes were performed
(38,39). The calculations showed that the 3-His ligand system in CDO is
essential for complete dioxygenase activity and significant amount of cyste-
ine sulfoxide product is predicted in CDO with a 2-His/1-Asp ligand sys-
tem. Furthermore, the ligand environment is important for spin-state
ordering and catalysis. Thus, in TauD the ground state of the iron(III)-
superoxo is quintet spin and during the oxygen activation mechanism the
system stays in the quintet spin state. In CDO the spin-state ordering varies
strongly and is dependent on the model and method used in the calculations.
Nevertheless, the rate-determining barriers all proceed on a dominant quin-
tet spin-state surface.
In contrast to the nonheme iron dioxygenases, the heme mono-
oxygenases, such as the cytochrome P450 enzymes, have been well studied
for oxidation reactions (40–42). In the human body, the cytochromes P450
play crucial roles particularly in the liver where they are involved in the
metabolism of xenobiotics and drug molecules. Furthermore, the P450s
have been implicated in biosynthesis pathways for the synthesis of estrogen
and other hormones. As such, metalloenzymes have large differences in their
ligand environment and chemical structure and hence show a wide variety of
chemical and biological functions.
In most metal-heme-dependent enzymes, including monooxygenases
such as the P450s as well as heme peroxidases and catalases (43), the catalytic
cycles possess a high-valent iron(IV)-oxo π-cationic radical intermediate
called CpdI. In all of these, CpdI is generated in a catalytic cycle that requires
two electron reductions and two proton transfers as summarized in Eq. (1).
This is in contrast to the above-mentioned nonheme iron dioxygenases that
generate their iron(IV)-oxo species through the assistance of a cosubstrate,
such as α-KG:
ironðIIIÞ heme + O2 + 2H + + 2e

! ironðIVÞ oxo hemeð +•Þ + H2 O (1)
Structurally, the iron-heme group in the P450s is bound to the protein

through a covalent linkage of the iron with a cysteinate side chain, whereas
catalases bind a tyrosinate group and the peroxidases typically a histidine
group. It has been proposed that the nature of the axial ligand affects and
guides the reactivity patterns and gives the heme enzyme its unique chemical
properties. Indeed, computational modeling of propene activation by
models of P450 CpdI with peroxidase and catalase models of CpdI showed
higher reactivity of the P450 system by orders of magnitude (44–49). This is
due to the fact that the highest occupied molecular orbital (HOMO), the a2u
orbital, in P450 CpdI is a mixed porphyrin/axial ligand orbital and the
amount of mixing is dependent on the electron-donating properties of
the axial ligand. Therefore, a stronger electron-donating group, such as
OH or SH, yields an increase of the a2u-antibonding combination and
consequently lowers the electron affinity of the complex and makes these
complexes good oxidants. By contrast, strong electron-withdrawing groups,
such as histidine or neutral solvent molecules, lower the a2u orbitals and
increase the electron affinity of the complex and hence make it a much wea-
ker oxidant (50).
Cytochrome P450 CpdI and the iron(IV)-oxo species in nonheme iron
dioxygenases are versatile oxidants that react efficiently with substrates. Gen-
erally, these active catalytic intermediates of heme and nonheme enzymes
are able to activate strong aliphatic C–H bonds and convert them to alcohol
groups. Other common reactions catalyzed by iron(IV)-oxo complexes
include substrate epoxidation, aromatic hydroxylation, desaturation, N-
and O-dealkylation, and heteroatom transfers such as sulfoxidation reac-
tions (8–10). Although these catalytic intermediates are implicated in many
inorganic and bioinorganic reactions, spectroscopic evidence of their direct
involvement in reactions is sparse as they are short-lived due to their highly
reactive nature. However for the active species of the P450 catalytic cycle
known as CpdI, spectroscopic evidence was obtained through UV–Vis,
EPR, and M€ ossbauer spectroscopic techniques and was reported by Rittle
and Green (51). In heme and nonheme systems, the participation of a very
active metal-oxo center for chemical transformations is convincing; none-
theless, there is still a debate as to the catalytic efficiencies of alternative
oxygen-bound iron complexes, such as iron(II)-peroxo, iron(III)-superoxo,
and iron(III)-hydroperoxo species. For instance, the catalytic potency of
heme mononuclear iron(III)-hydroperoxo intermediate in P450 chemistry
has been disputed, as experimental and computational studies have shown
that it is a sluggish oxidant in substrate oxidation of P450 catalysis, and there-
fore would not be the principal oxidant in hydroxylation of organic sub-
strates (52–60). These studies contrast the findings that exist for
biomimetic nonheme iron systems whereby the nonheme iron(III)-
hydroperoxo species is catalytically efficient and can perform substrate
activation reactions (61–65). In addition, although it is very difficult to trap

side-on manganese-peroxo (MnOO) in spectroscopic studies, there is
experimental and computational evidence of its catalytic potency in the lit-
erature particularly for synthetic biomimetic models that closely resemble
enzyme intermediates (4,5,7,66).
It is clear, therefore, that nonheme and heme dioxygen-bound interme-
diates have catalytic properties that are dramatically different and result in the
nonheme iron(III)-hydroperoxo species to be catalytically active in aromatic
hydroxylation reactions, while heme iron(III)-hydroperoxo, by contrast, is
unreactive. Computational studies give insight into the origin of these reac-
tivity differences. In this review paper, we will investigate thoroughly the
catalytic and electronic properties of a variety of metal(IV)-oxo,
metal(III)-peroxo, metal(III)-superoxo, and metal(III)-hydroperoxo species
and try to understand their reactivity differences. These individual studies all
have their inspiration from enzymatic catalytic cycles and particularly to the
nonheme iron dioxygenases that efficiently perform oxygen insertion reac-
tions into substrates. It is now becoming clear that not only metal(IV)-oxo is
an active oxidant of oxygen atom transfer reactions but that in some cases
metal(III)-superoxo or metal(III)-hydroperoxo can be as efficient—if not
a better oxidant.
2. METAL-(DI)OXYGEN INTERMEDIATES
As described earlier, the catalytic cycles of heme and nonheme iron
monooxygenases and dioxygenases generally proceed through the forma-
tion of a high-valent metal(IV)-oxo intermediate. For instance, in TauD
(20–22) dioxygen binds the iron(II) center as an iron(III)-superoxo group
that attacks the keto-position of α-ketoglutarate to form succinate, CO2,
and iron(IV)-oxo species. The iron(IV)-oxo intermediate abstracts a hydro-
gen atom from the taurine substrate and after radical rebound gives hydrox-
ylated taurine, an intermediate in the cysteine metabolism pathway in the
body. In CDO, by contrast, both oxygen atoms are transferred to the sub-
strate (cysteine) in a consecutive process via cysteine sulfoxide to form the
cysteine sulfinic acid product (31,32,67). Interestingly, the nonheme iron
dioxygenase isopenicillin N synthase reacts with substrate through four con-
secutive hydrogen atom abstraction reactions (68,69). This proceeds by the
formation of an iron(III)-superoxo reactant from an iron(II) resting state and
molecular oxygen, which after hydrogen atom abstraction is converted into
an iron(II)-hydroperoxo intermediate. The latter abstracts a hydrogen atom
to form water and an iron(IV)-oxo group and leads to a ring-closure step

(desaturation) of the substrate. Subsequently, another hydrogen atom
abstraction by the iron(IV)-oxo species gives the iron(III)-hydroxo com-
plex. A final hydrogen atom abstraction by the iron(III)-hydroxo species
closes another ring in the substrate (desaturation) and returns the cycle to
the iron(II) resting state and releases product. It appears, therefore, that in
isopenicillin N synthase the iron(III)-superoxo, iron(II)-hydroperoxo,
iron(IV)-oxo, and iron(III)-hydroxo intermediates are catalytically active
oxidants that are able to react with substrates by hydrogen atom abstraction
from aliphatic groups. This is in contrast to the known reactivity patterns of
heme enzymes where iron(III)-superoxo and iron(III)-hydroperoxo are
sluggish oxidants in aliphatic hydroxylation reactions.
In bleomycin, a glycopeptide antibiotic that digests double-stranded
DNA, an iron(III)-hydroperoxo intermediate is formed in a reaction with
oxygen and iron present in the medium. The iron(III)-hydroperoxo can
attack the substrate directly or initially converts via either heterogeneous
or homogeneous cleavage of the hydroperoxo moiety to form iron(IV)-
oxo or iron(V)-oxo species (70–73). Of these intermediates that catalyze
hydrogen abstraction, iron(III)-hydroperoxo and other similar transition
metal-peroxo intermediates are the focus of this review, and in later sections,
we will discuss their reactions/catalytic potentials in substrate activation
reactions.
2.1 Short-Lived Iron-Dioxygen Complex in Cysteine

Dioxygenase Enzymes
Cysteine catabolism is a vital process for human health and its first step is
mediated by a CDO. Although cysteine is one of the amino acids that forms
building blocks of many proteins in cells, its buildup in neuronal cells is
known to be excitotoxic and increased levels of cysteine have been corre-
lated with neurodegenerative diseases (74,75). However, it was found that
increased activities of CDO result from the forced expression of its gene in
human cancer cells which is also correlated with decreased tumor cell
growths (76). As such, the catalytic activities of CDO and its regulation
ensure physiological levels of cysteine are maintained constant.
CDO catalyzes the conversion of cysteine to cysteine sulfinic acid that is
the first irreversible step in cysteine catabolism (77,78). The details of the
catalytic mechanism of CDO remain incomplete and therefore the mecha-
nism is not yet established. In particular, no experimental data on
dioxygen-bound intermediates in CDO enzymes were reported; hence,
all suggestions on short-lived intermediates came from analysis of product

distributions and computational modeling (35–39,79). Subsequent experi-
mental work and computational studies on biomimetic model complexes
gave insight into dioxygen activation and predicted a general mechanism
for CDO reaction (39,80,81).
Fig. 3 summarizes key intermediates in the catalytic cycle of CDO
enzymes. The cycle starts with the resting state complex (R), which is a non-
heme iron(II) linked to the protein through three interactions with histidine
residues, while the other three ligand positions are occupied by water mol-
ecules. The ligand coordination system distinguishes CDO from typical
nonheme iron dioxygenases that have the 2-His/1-Asp feature as discussed
earlier. The cycle starts with cysteinate entering the substrate binding pocket
and binds the iron(II) complex as a bidentate ligand with the thiolate and
amide groups to generate structure A. The last water molecule is expelled
from the iron center when dioxygen enters to form the iron(III)-superoxo
complex (B). This complex is elusive and has not been trapped and charac-
terized for any nonheme iron dioxygenase. The distal oxygen atom (labeled
Od) attacks the thiolate to give a bridging structure C. An O–O
bond-breaking step gives an iron(IV)-oxo species and cysteine sulfoxide
intermediate (D) that after the second oxygen atom transfer yields cysteine
sulfinic acid products (E). Release of product and rebinding of water returns
the active site to the resting state R and completes the catalytic cycle.
UV–Vis absorption studies identified two absorption bands at 500 and
640 nm for a species containing two oxygen atoms (82). Unfortunately,
the intermediate decays in 20 ms and attempts to characterize a
Fig. 3 Catalytic cycle of cysteine dioxygenase.

short-lived species further, failed, and no indistinguishable EPR and

M€ ossbauer spectra were obtained. Therefore, we resorted to a computa-
tional study at the time dependent-density functional theory (TD-DFT)
and complete active space—self consistent field (CASSCF) level of theory
to establish the nature of the intermediate. In particular, focus was on
two reactive intermediates, namely the iron(III)-superoxo anion (B) and
iron(III)-bicyclic-O–O ring structure (C), but also the cysteine sulfoxide
(D) and cysteine sulfinic acid products (E) were considered.
Evidence from experiments using M€ ossbauer spectroscopy established
the presence of a water-bound structure (with water in the sixth ligand
position, i.e., structure A in Fig. 3) of cysteine-bound CDO but only the
cysteine-CDO would go through the dioxygenation reaction under the
appropriate conditions (82). It was observed that when an anaerobic
cysteine-bound CDO was titrated with oxygenated buffer at 4°C in a glove
box, this led to the appearance of two absorption peaks at 500 and 640 nm.
These peaks were found to scale linearly with concentration of CDO as well
as the amount of dioxygen supplied. Since the species with absorption fea-
tures of 500 and 640 nm decayed within about 20 ms, it was proposed to be
an oxygen adduct in the catalytic cycle of CDO. A first-order rate constant
of 112 5 s1 for its decay was measured. The results were further validated
with various experimental methods such as chemical quenching and
freeze-quench, thus giving the same outcomes of dioxygenation reactions
for anaerobic cysteine-bound CDO (82).
In addition, computational modeling using density functional theory on
model complexes as well as QM/MM studies on the full enzyme calculated
high reaction barriers for the formation of the iron(III)-superoxo and the
iron(III)-bicyclic ring species, and, therefore, these calculations predicted
a finite lifetime of the two complexes B and C (35,37). On the other hand,
formation of the iron-cysteine sulfoxide and iron-cysteine sulfinic acid
products is calculated to be strongly exothermic with small barriers of for-
mation and decay. Consequently, the lifetime of these intermediates is
expected to be rather short.
Fig. 4 displays calculated absorption spectra of structures of the iron(III)-
superoxo (B), the bicyclic complex (C), the iron-cysteine sulfoxide complex
(D), and the iron-cysteine sulfinic acid product complex (E) in the quintet
spin state with TD-DFT and CASSCF methods. Although none of the
absorption spectra reproduce experiment perfectly, some features are
reproduced well. First, as predicted by the DFT and QM/MM calculations
of the reaction mechanism, and confirmed with TD-DFT and CASSCF,
5 5
TD-DFT B TD-DFT C TD-DFT D TD-DFT
5 5
E
562 353
630
CASSCF 584 CASSCF
509 543
644
Fig. 4 TD-DFT and CASSCF calculated absorption spectra of short-lived intermediates of

the catalytic cycle of CDO.
structures 5D and 5E can be ruled out as the intermediates responsible for the
absorption features in the experimental spectrum. Either no spectroscopic
signal in the correct place is seen, or no signal is seen at all for both inter-
mediates. Second, the CASSCF results implicate two clear absorption signals
in the 500–700 nm region for both 5B and 5C, whereby the intensity of the
peaks for 5C appears to match the experiment the best, although both peaks
are shifted with respect to the experimental assignment. Therefore, compu-
tational analysis implicated two possible candidates responsible for the
experimentally observed absorption maxima, namely 5B or 5C. The theory
slightly favors the 5C assignment, but the error of the calculations is too large
for a definitive conclusion. Unfortunately, theory fails to resolve the exper-
imental conundrum and assign the sole species responsible for the absorption
spectra, but the results mostly point to the bicyclic ring structure in the quin-
tet spin state as the species giving the experimental spectra. As such, further
analytical work will be necessary to distinguish between the two species,
especially using analytic tools that can detect them within a shorter time
for instance, at less than 10 ms, in order to assign the role and the nature
of the oxygen adduct intermediates in the catalytic cycle of CDO.
2.2 Iron(III)-Hydroperoxo in Heme and Nonheme Complexes

Early studies on cytochrome P450 mutants that disrupted the proton deliv-
ery pathways of P450cam, a camphor-hydroxylating P450 isozyme, impli-
cated reactivity patterns of the precursor of CpdI in the catalytic cycle,
namely Compound 0 (Cpd0) or the iron(III)-hydroperoxo heme complex
(60). These studies created controversy and suggested that Cpd0 alongside
CpdI was catalytically active and could act as a second oxidant in the catalytic
cycle of P450. Later density functional theory modeling and QM/MM
studies on double bond epoxidation and sulfide sulfoxidation (55,56,83),
however, cast doubt upon the evaluation of the experimental results and
gave sluggish reactivity patterns for Cpd0 as compared to that of CpdI. Fur-
ther experimental studies of biomimetic model complexes confirmed the
conclusions of the theory (59).
Unlike the heme iron(III)-hydroperoxo that has been shown to be a slug-
gish oxidant of substrate hydroxylation reactions, a surprising result was
found with the hexacoordinated nonheme iron(III)-hydroperoxo counter-
part. In contrast to the heme Cpd0 structure, it was proven to be a potent
oxidant of, for example, halide transfer and aromatic hydroxylation reac-
tions (61–63). This prompted us to perform a detailed computational study
of the aromatic hydroxylation of arenes by heme and nonheme iron(III)-
hydroperoxo complexes (64,65).
A detailed computational analysis was performed on an experimentally and
spectroscopically characterized pentadentate nonheme iron(III)-hydroperoxo
complex, namely, ½ðL5 2 ÞFeIII OOH with L5 2 ¼ N methyl N ,N 0 , N 0
2+
trisð2 pyridylmethylÞethane 1,2 diamine (62,84). The nonheme iron

(III)-hydroperoxo model system was carefully characterized with a broad range
of spectroscopic methods and was reacted under ideal experimental conditions
with benzene and anisole for conversion into phenol products (62). However,
the studies failed to assign unequivocally the iron(III)-hydroperoxo group as the
active oxidant in the reaction mechanism and, therefore, a computational study
was needed. Nevertheless, computational studies on the reactivity of the oxi-
dant revealed the intricate details of the aromatic hydroxylation reactions and
identified the key intermediate that performs the reactions (see Fig. 5). First, the
computational studies characterized two stable local reactant structures for the
iron(III)-hydroperoxo complex (Re1 and Re2) that have either the distal oxy-
gen atom hydrogen bonded to protons of the L5 2 ligand or the proximal oxygen
atom. These two reactant complexes are within a few kcal mol1 of each other
and both react with substrates via arene activation. In agreement with experi-
ment, the DFT calculations establish the nonheme iron(III)-hydroperoxo
complex as a suitable oxidant to activate arenes, such as benzene and anisole,
whereby low reaction barriers in solvent were obtained on a doublet spin
ground state (64). These results highlight nonheme iron(III)-hydroperoxo
complexes as potent oxidants of aromatic substrates in contrast to the heme
iron(III)-hydroperoxo complexes that are sluggish oxidants.
As shown in Fig. 5, a concerted OH transfer reaction barrier (TSCO)
2+
between ½ðL5 2 ÞFeIII OOH and benzene leads to the formation of an
iron(IV)-oxo species and C6H6OH+, which after a proton transfer in an
almost barrier-less fashion collapses to the phenol products. Also shown
2.381 H 1.950 H
O O
2.116 O1.696 1.851 O1.751
Fe Fe
2TS 2TS 2TS
CO,1
25.6 CO,1 CO,2
23.3
2TS
CO,2
2TS
PT,1
2Re
1
2I
A,1
4.4
2.9 2.9
0.0 −0.6
2Re
2 2I
A,2
2P
A
−53.7
1
Fig. 5 Free energy profiles (in kcal mol ) for arene activation by a nonheme iron(III)-
hydroperoxo complex. Free energies obtained at UB3LYP/BS2 + PCM//UB3LYP/BS1. From
Faponle, A.S.; Quesne, M.G.; Sastri, C.V.; Banse, F.; de Visser, S.P. Chem. Eur. J. 2015, 21,
1221–1236.
are optimized geometries of the rate-determining transition states (2TSCO)

starting from the two isomeric reactants. The two geometric isomers of the
complex lead to isomeric transition state structures. Nonetheless, depending
on the angle from which the substrate approaches the oxidant, the transition
states are either later or earlier on the potential energy surface and have a
longer O–O bond and a shorter C–O bond. Optimized structures of the
two transition states are given in Fig. 5.
To understand the reactivity differences of heme and nonheme iron(III)-
hydroperoxo complexes, we decided to analyze the thermochemical proper-
ties of these species and focused initially on the strength of the peroxo bond. In
particularly, we investigated the homolytic and heterolytic cleavage of the
peroxo moiety of the iron(III)-hydroperoxo complex, leading to either an
iron(IV)-oxo and an OH radical or an iron(V)-oxo and an OH anion
(Fig. 6). During the reaction mechanism shown in Fig. 5, no intermediates
that result from homolytic or heterolytic cleavage of the O–O bond were
located as local minima. Instead a concerted OH transfer takes place, so that
these intermediates after homolytic or heterolytic breaking are thermody-
namically unfeasible reactions. The initial heterolytic splitting of the peroxo
moiety in ½ðL5 2 ÞFeIII OOH costs ΔGsolv ¼ 73.9 kcal mol1 (Fig. 3) and
2+
Fig. 6 Thermochemical analysis of heterolytic and homolytic cleavage of the O–O bond
in nonheme (left) and heme (right) iron(III)-hydroperoxo complexes. Free energies in sol-
vent given in kcal mol1.
hence is an inaccessible pathway. On the other hand, the initial homolytic

2+
splitting of the O–O bond of the ½ðL5 2 ÞFeIII OOH complex appears to
be thermodynamically feasible with a small endergonicity of
ΔGsolv ¼ 13.0 kcal mol1 and hence could be a viable alternative pathway
to that shown in Fig. 5. However, the subsequent reactions between
3
2 IV 2 +
L5 Fe O and benzene encounter a reaction barrier of another
24.0 kcal mol1 on the triplet pathways (85). This implies that a total free
energy of activation of ΔGsolv ¼ 37 kcal mol1 would be required to initiate
the homolytic breakdown of the hydroperoxo group of the
2+
½ðL5 2 ÞFeIII OOH and the subsequent C–O bond formation to the substrate
on the triplet pathway. Obviously that pathway is much higher in energy than
the concerted pathway displayed in Fig. 5.
Subsequently, we calculated the homolytic and heterolytic cleavage
energies of the heme iron(III)-hydroperoxo complex, where we took a
small active site model with protoporphyrin IX scaffold (without side chains)
and thiolate to mimic the axial cysteinate residue of the iron (86–90). The
data in Fig. 6 show dramatic differences between nonheme and heme
iron(III)-hydroperoxo patterns. Thus, in heme iron(III)-hydroperoxo both
homolytic and heterolytic O–O cleavage patterns require a small amount of
energy, i.e., ΔGsolv ¼ 10.2 kcal mol1 for the heterolytic cleavage and
ΔGsolv ¼ 15.5 kcal mol1 for homolytic O–O bond cleavage. Therefore,
heterolytic cleavage and formation of CpdI are the preferred pathway for
heme iron(III)-hydroperoxo complexes, whereas homolytic cleavage of
the O–O bond and formation of a nonheme iron(IV)-oxo complex are pref-
erential in nonheme iron(III)-hydroperoxo. Note that the homolytic cleav-
age pathways are very similar for heme and nonheme iron(III)-hydroperoxo,
but major thermodynamic differences are seen for the heterolytic cleavage
reactions that are destabilized for the nonheme iron(III)-hydroperoxo spe-

cies. The reason for this may be the differences in overall charge of the
two complexes studied.
The thermodynamic differences between heme and nonheme iron(III)-
hydroperoxo provide the nonheme iron(III)-hydroperoxo system the ability
to catalyze direct aromatic hydroxylation reactions of arenes, whereas the
heme iron(III)-hydroperoxo intermediate is just a catalytic cycle intermedi-
ate en route to CpdI formation. The differences that exist between the two
iron(III)-hydroperoxo complexes lie in shape, size, and energy levels of the
molecular valence orbitals along the metal-peroxo moiety. Orbital analyses
have shown that although the valence orbitals of the two systems are char-
acterized by the mixing of metal 3d orbitals and the first sphere atoms of the
ligand architecture, there are notable orbital interaction differences of the
peroxo (O–O) group and the ligand systems (62). Specifically, in heme
Cpd0 strong orbital interaction exists between the πOO * orbital in the
xy-plane and the a2u orbital on the porphyrin ligand. Meanwhile, this inter-
action is either lacking or weaker in the nonheme iron(III)-hydroperoxo
system where the πOO * orbital can only interact with nitrogen atoms in
the equatorial plane of the ligand system. As a consequence, the nonheme
πOO * orbital becomes the HOMO rather than a ligand orbital. This implies
that the ligand cannot be oxidized in the reaction. By contrast, in P450 Cpd0
the πOO * orbital mixes with the ligand a2u orbital and becomes the HOMO
orbital which would make it susceptible to heterolytic cleavage to form Cpd
I and a hydroxide anion.
2.3 Spin-State Reactivity in Biomimetic Manganese(V)-Oxo

Complexes
Several enzymes and proteins in Nature utilize manganese in their active
center rather than iron with examples ranging from superoxide dismutase
(91,92) and photosystem II. The latter contains a cubane-type cluster with
four manganese, a calcium, and five oxygen atoms for the biosynthesis of
molecular oxygen (93). As such, understanding how oxygen atoms or mol-
ecules bind to manganese centers is important and will give insight into the
mechanism of these important enzymes in nature. In the next two sections,
we will cover biomimetic studies on two manganese complexes and their
reactivity patterns.
Manganese(V)-oxo porphyrinoid complexes are well studied particularly
with corrole and corrolazine ligand systems as they have a total charge of 2
and hence can stabilize metals in high oxidation states easily (94,95). In
particular, in collaboration with the Goldberg group, we investigated the
structure and reactivity of a manganese(V)-oxo corrolazine complex
[Mn(O)(H8Cz)X], H8Cz is corrolazine without side chains and X is the axial
ligand (X ¼ no ligand or CN/F). We initially investigated the axial ligand
effect on the dehydrogenation of 9,10-dihydroanthracene and showed that
the reaction rate was enhanced by a factor of 2100 upon the addition of F,
while it increased by a factor of 16,000 with the addition of CN (96,97).
Similar rate enhancements were seen for the use of 1,4-cyclohexadiene as a
substrate. It was established that the main cause of the rate enhancement was
increased acidity of the manganese(IV)-hydroxo product, which is formed as
a result of strong electron-donating properties of the axial ligand. These ini-
tial studies only considered the closed-shell singlet spin ground state as it is
known to be the ground state for the [Mn(O)(H8Cz)] complex. Shaik and
coworkers, however, contrasted these studies and proposed an alternative
hypothesis where the reactant has close-lying singlet and triplet spin states
and a spin-state crossing to a higher spin state and thereby would lower
the reaction rate (98).
Fig. 7 shows high-lying occupied and low-lying virtual orbitals of the
[Mn(O)(H8Cz)(CN)] complex. Thus, the molecular orbitals depend on
the manganese-oxo interactions and those on the corrolazine scaffold.
The lowest pair of orbitals shown in Fig. 7 is the bonding combination
of the 2px/2py atomic orbitals on oxygen with the 3dxz/3dyz atomic orbitals
on manganese: πxz/πyz. A little higher in energy is the nonbonding orbital
δxy that is located in the plane of the corrolazine ligand midway in between
the nitrogen atoms. Above the δxy orbital are the pair of antibonding orbitals
π * xz/π * yz and two virtual σ* orbitals for the interaction between the metal
and corrolazine nitrogen atoms (σ*x2 y2 ) and between the metal and oxo
group (σ*z2 ). A final orbital shown in Fig. 7 is the corrolazine π-orbital
labeled as a00 , which shows resemblance to the a1u orbital in porphyrins.
In the closed-shell singlet ground state of the [Mn(O)(H8Cz)] complex
the orbital occupation is [core] πxz 2 πyz 2 a00 2 δxy 2 . This orbital occupation cor-
responds to a short Mn–O distance and hence creates a triple bond between
manganese and oxygen. Higher in energy is a triplet spin state with [core]
πxz 2 πyz 2 a00 1 δxy 1 π*xz 1 π*yz 1 configuration, which becomes accessible upon
binding of, for instance, Zn2+ anions to the complex through a valence tau-
tomerism mechanism (99). In addition, there are two more possible triplet
spin states with occupation [core] πxz 2 πyz 2 a00 2 δxy 1 π*xz 1 π*yz 0 and [core]
πxz 2 πyz 1 a00 2 δxy 1 π*xz 1 π*yz 1 .
Fig. 7 Chemical structure and possible singlet and triplet spin orbital occupation of
[Mn(O)(H8Cz)(CN)]. Reprinted with permission from the American Chemical Society, Yang,
T.; Quesne, M. G.; Neu, H. M.; Cantú Reinhard, F. G.; Goldberg, D. P.; de Visser, S. P. J. Am.
Chem. Soc. 2016, 138, 12375–12386.
As the spin-state ordering of [Mn(O)(H8Cz)(CN)] was controversial,

experimental studies focused on its spectroscopic characterization, and
X-ray absorption spectroscopy characterized the structure with a short
Mn–O bond length typical for a closed-shell singlet spin state (100). Hence,
if a triplet spin state is involved in the reaction mechanism, then a spin-state
crossing from the closed-shell singlet spin state to one of the possible triplet
spin states must happen along the reaction mechanism for substrate activa-
tion and ideally before the rate-determining reaction barrier. To this end, we
investigated thioanisole sulfoxidation by [Mn(O)(H8Cz)(CN)] with sub-
strates with a range of para-substituents and measured and calculated rate
constants for the reaction (101). The studies gave a V-shaped Hammett type
plot characteristic for a change in mechanism during the reaction; however,
as the trends were not fully understood a computational study was subse-
quently performed (102).
For a large set of para-X-substituted thioanisole substrates (X ¼ N
(CH3)2, NH2, OCH3, CH3, H, Br, CN, and NO2) the sulfoxidation
mechanism by [Mn(O)(H8Cz)(CN)] was calculated with density functional
theory methods (102). Subsequently, the free energy of activation was
converted into a rate enhancement (log kx/kH) of the free energy of acti-
vation for para-X-thioanisole relative to para-H-thioanisole and the results
of the singlet and triplet spin barriers are given in Fig. 8. As can be seen, in
agreement with the experimental observation, the singlet spin plot gives a
V-shaped Hammett plot, whereas the plot is linear for the triplet spin state
of data. To confirm that no spin-state crossing from the singlet to the triplet
spin state is possible, we also calculated minimum energy crossing point
structures, but first they were high in energy (of the same level as the
sulfoxidation transition states) and second were not lying on the pathway
for sulfoxidation. Moreover, calculations of the spin–orbit coupling con-
stant gave very small values, which implicate little (if at all) spin-state cross-
ing from singlet to triplet. Consequently, the computational modeling
reveals a spin selective reactivity of [Mn(O)(H8Cz)(CN)] with thio-
anisoles and a dominant spin-state surface only. As such, a high spin state
is not essential for chemical reactivity, reactions can take place on a closed-
shell singlet spin-state surface.
It should be mentioned here that a recent benchmark study on the sub-
strate activation of para-X-substituted thioanisole sulfoxidation (X ¼ H, CH3,
OCH3, Cl) by [Fe(O)(N4Py)]2+, N4Py ¼ N,N-bis(2-pyridylmethyl)-N-bis
(2-pyridyl)methylamine, with a range of density functional theory methods
showed that computation reproduces the trends from Hammett type, linear
free energy relationship plots with almost all methods, but often gives a sys-
tematic error from the experimental enthalpy of activation (103). A parallel
outcome was found for regioselectivity patterns of bifurcation pathways
(104,105).
Singlet spin data Triplet spin data

NO2
4.0
3.5 NO2 3.0 r = 3.71
3.0
2.0
2.5 r = 3.45 CN
log(kY/kH)
1.0 H
log(kX/kH)
2.0
0.0
1.5 N(CH3)2 CN OCH3 Br
1.0 NH2 –1.0
OCH3 NH2
0.5 Br –2.0 CH3
H
r = –1.21 –3.0
0.0 CH3 N(CH3)2
–0.5 –4.0
–1 –0.5 0 0.5 1 –1.0 –0.5 0.0 0.5 1.0
σP σP
Fig. 8 Calculated Hammett plots of para-thioanisole sulfoxidation by [Mn(O)(H8Cz)

(CN)] on the singlet and triplet spin states. Reprinted with permission from the American
Chemical Society, Yang, T.; Quesne, M. G.; Neu, H. M.; Cantú Reinhard, F. G.; Goldberg, D. P.;
de Visser, S. P. J. Am. Chem. Soc. 2016, 138, 12375–12386.
2.4 Aldehyde Deformylation by Manganese(III)-Peroxo

Complexes
Another interesting chemistry property of a metal-peroxo system is exem-
plified in the reaction of a manganese(III)-peroxo complex with
2-phenylpropionaldehyde (106). Manganese-containing enzymes are
common in nature, and their reactions show similarity to those of the iron
enzymes. In particular, in a reaction with molecular oxygen, catalytic
intermediates are formed that include side-on manganese(III)-peroxo,
end-on manganese(IV)-superoxo, and the above-mentioned manganese
(V)-oxo species. Generally, these intermediates are short-lived and unfor-
tunately little experimental evidence of enzymatic side-on manganese
(III)-peroxo species exists. They are thought to be important intermedi-
ates in the catalytic cycles of biological systems such as photosystem II and
superoxide dismutase (91–93). The former catalyzes the photolysis of
water to generate molecular oxygen and the latter is involved in the bio-
degradation of toxic superoxide to hydrogen peroxide and water, respec-
tively. As such, inspired by these natural phenomena of reactions mediated
by manganese centers, synthetic biomimetic models have been developed
to understand the catalytic intermediates of these important enzymes and
proteins.
Previous reactivity studies of heme and nonheme metal–peroxo com-
plexes with aldehydes concluded that the reaction takes place with a rate-
determining nucleophilic attack on the carbonyl group of the aldehyde
(107–109). In order to explore further the catalytic ability of manganese-
peroxo intermediates with aldehydes, an experimental and computational
study was carried out for a side-on manganese(III)-peroxo complex with
a bispidine ligand system.
The work was initiated with a computational study and the mechanism is
displayed in Fig. 9. Thus, two alternative mechanisms were considered,
namely, a hydrogen atom abstraction from the α-carbon of the aldehyde
substrate and a nucleophilic attack on the carbonyl group of the substrate
(106). Interestingly, the nucleophilic attack undergoes a high-energy barrier
of ΔE{ ¼ 28.7 kcal mol1, whereas the hydrogen atom abstraction barrier
was found to be considerably lower: 23.9 kcal mol1. It appears, therefore,
that the aldehyde deformylation reaction is initiated with a rate-determining
hydrogen atom abstraction.
Experimental studies to support the computational mechanistic proposal
were performed and established a KIE of about 5 for the ratio of the rate
Fig. 9 DFT-calculated bifurcation pathways for the reaction of a bispidine-ligated

manganese(III)-peroxo complex with 2-phenylpropionaldehyde. Barriers give bond
lengths in angstroms and free energies are in kcal mol1.
constants between 2-phenylpropionaldehyde vs 2-d1-2-phenylpropional-

dehyde. These studies implicate a rate-determining hydrogen atom abstrac-
tion for aldehyde deformylation in agreement with the density functional
theory studies. Subsequent work using 2-methyl-2-phenylpropionaldehyde
as a substrate yielded no products and hence gave further evidence of a rate-
determining hydrogen atom abstraction.
The question is why does the side-on manganese(III)-peroxo complex
prefer to react through a hydrogen atom abstraction rather than the reported
nucleophilic attack on the carbonyl carbon? The reason for this originates
from the redox potential of the oxidant. Thus, the side-on manganese
(III)-peroxo has a low redox potential and can efficiently abstract an electron
from the peroxo moiety. By contrast, the electron transfer to the carbonyl
group of the substrate by the peroxo group would lead to the nucleophilic
pathway, but it is much higher in energy. To understand these differences in
electron transfer properties, a valence bond/molecular orbital analysis on the
reactants, intermediates, and products was carried out, and this is schemat-
ically depicted in Figs. 10–12.
Figs. 10 and 11 show the electron transfer and bond-breaking/forming
processes during the hydrogen atom abstraction reaction. Each dot in the
figure represents one electron and a line with two dots is a bonding orbi-
tal with two electrons. Only key orbitals are shown in this figure. Thus,
the manganese(III)-peroxo complex has an electronic configuration with
four singly occupied orbitals on the manganese in quintet spin state:
Fig. 10 Two-parabola curve-crossing diagram to rationalize nucleophilic vs electrophilic

reaction mechanisms.
C
2pC1
C
C
σCH2 2pC1
H σOH2
H 1sH1
H
O
πOO,xy2
σOO2
π*OO,xy2
O
O O πxz2 O O
πxz2 π*xz1
HAT π*xz1
Mn Mn Mn
1 1
3dxz1 3dxy1 σ*x2–y2 σ*z2 3dyz1 3dxy1 σ*x2–y2 σ*z2
1 1
3dyz1 3dxy1 σ*x2–y2 σ*z2
1 1
5[MnIII(O )(L)]2+
2 TSHA IHA
Fig. 11 Valence bond structures for the reaction mechanism of hydrogen atom abstrac-
tion by a side-on manganese-peroxo complex.
σCO2
σCO2 C O 2pO2
σCO2
C O C O
πCO2 2pC 1
2pO1 σO2-C2
πOO,xy2 O
π*OO,xy2 O
O O πxz2 O O
Nucleophilic π* 1 πxz2 π*xz1
xz
into 2pO
attack
Mn Mn Mn
1 1 1 1
3dxz1 3dxy1 σ*x2–y2 σ*z2 1
3dxy1 σ*x2–y2 σ*z2
1
3dxy σ*x2–y2 σ*z2
1
5[MnIII(O 2+
2)(L)] TSNA INA
Fig. 12 Valence bond structures for the reaction mechanism for nucleophilic addition of
aldehydes by a side-on manganese-peroxo complex.
3dxz 1 3dxy 1 σ*x2 y2 1 σ*z2 1 . In addition, there are two relevant orbitals on the
peroxo moiety, namely πOO,xy and π * OO,xy both occupied with two elec-
trons. In the hydrogen atom abstraction step, the σCH orbital for the inter-
action of the C and H atoms breaks into atomic orbitals: 2pC and 1sH. Next,
the πOO,xy and π * OO,xy convert back to atomic orbitals and one electron
pairs up with the 1sH atomic orbital to form the new σOH bond. Two elec-
trons originating from the πOO,xy/π * OO,xy orbitals pair up with the manga-
nese 3dxz electron to form a new manganese-oxo three-electron bond:
πxz 2 π*xz 1 . The last electron from the peroxo bond is promoted to the
3dyz orbital of manganese. Therefore, the hydrogen atom abstraction barrier
will be determined by the energy to break the C–H bond of the substrate
(σCH), the energy to form the O–H bond in the product (σOH), the energy
to break the π-bond of the peroxo group (πOO,xy/π * OO,xy), and finally the
energy to excite an electron from the peroxo to the 3dyz orbital on
manganese.
The contrasting nucleophilic addition pathway is shown at the bottom of
Fig. 10. Similar to the hydrogen atom abstraction step, the πOO,xy and
π * OO,xy revert to atomic orbitals and an end-on peroxo is formed with a
new set of πxz/π * xz orbitals along the manganese-oxo group. However,
no electron transfer into the manganese takes place, but instead the electron
is donated into the carbonyl group of the substrate. In particular, the π-bond
of the carbonyl (πCO) breaks into atomic orbitals (2pC and 2pO), whereby
the 2pO orbital receives a second electron from the peroxo group and the
2pC forms a covalent bond (σO2–C) with the peroxo moiety. Therefore,
the relative electron affinity of the substrate carbonyl vs the manganese atom
will determine whether the dominant pathway is hydrogen atom abstraction
or nucleophilic addition.
To further explain and rationalize the bifurcation pathways and make
predictive measurements on how the regioselectivity can be changed, we
designed a valence bond/molecular orbital scheme to understand the details
of the reaction mechanisms. These valence bond curve-crossing diagrams
were initially developed by Shaik (110) and, for instance, show predicted
hydrogen atom abstraction barriers for cytochrome P450 CpdI (111).
Recently, we designed a different model based on the crossing of two parab-
olas (112). Thus, in the two-parabola curve-crossing description the poten-
tial energy curve (y) is described as a parabola with the minimum energy
point for the reactants located at reaction coordinate xR ¼ 0, i.e., yR ¼ ax2
where a is a constant. The product geometry is located at the minimum
of a parabola at reaction coordinate xP ¼ 1 and energy function yP ¼ bx2 + cx

+ d, where b, c, and d are constants. We now assume that the transition state is
located at a reaction coordinate x ¼ 1/2 at the point where the two curves
cross. Shaik showed that the curve-crossing energy is correlated with a frac-
tion of the excitation energy between the curves in the reactant geometry
(110). Similarly, from the crossing point of the two parabolic functions,
we can derive that it is related to the values of yP(0) and yP(1), which rep-
resent the Franck–Condon energy in the reactant complex (EFC,R) and the
driving force for the reaction (ΔErp), respectively (112). As with the VB
method, the reaction barrier can be estimated by subtracting the resonance
energy from the curve-crossing energy.
By analyzing the electronic configurations of the ground and excited
state complexes in the geometry of the reactants, it can be reasoned that
the value EFC,R for hydrogen atom abstraction is dependent on the energy
to break the C–H bond of the substrate, the energy to form the O–H bond
between peroxo and incoming hydrogen atom, the breaking of the πOO
bond of the peroxo group, the formation energy of the three-electron
πMnO/π * MnO bond, and the energy to transfer an electron from peroxo
to manganese. We estimated values for each of these contributions from
the molecular orbital energies of the reactant complex and the adiabatic
C–H and O–H bond energies, which gave a predicted hydrogen atom
abstraction barrier of 31.7 kcal mol1. This value matches the DFT-
calculated hydrogen atom abstraction barrier well.
Subsequently, we predicted the nucleophilic attack barrier in a similar
way. In the nucleophilic pathway, the π-bond of the substrate carbonyl bond
is broken. In addition, the πOO bond of the peroxo moiety is broken and a
three-electron MnO bond is formed, similar to that seen in the hydrogen
atom abstraction process. Finally, in the nucleophilic addition transition state
a C–O bond is formed between carbonyl and peroxo group and an electron
is transferred from peroxo to the carbonyl oxygen atom. Based on orbital
energies, we estimate a nucleophilic addition barrier of 33.2 kcal mol1,
in good agreement with the DFT calculations.
In summary, the bispidine-ligated manganese(III)-peroxo complex
reacts with aldehydes via hydrogen atom abstraction rather than a nucleo-
philic attack on the carbonyl group. A detailed rationalization of the orbital
changes in the reaction pathways reveals that the difference between the
two pathways is mostly related to the electron transfer from the peroxo spe-
cies. Thus, in hydrogen atom abstraction, the peroxo group donates an
electron to the manganese, whereas in the nucleophilic pathway the
electron transfers to the substrate carbonyl group instead. The selectivity of

hydrogen atom abstraction vs nucleophilic addition, therefore, is deter-
mined by the relative electron affinity of the manganese with respect to
the substrate carbonyl group. In our system, the manganese is a sufficiently
effective electrophile and can abstract electrons easily and much better than
the substrate carbonyl group. This may in part be due to steric repulsions on
substrate approach.
3. CONCLUSIONS
As highlighted by several examples in this work, metal-oxo,
metal-peroxo, and metal-hydroperoxo moieties show unique reactivity pat-
terns with substrates. Often these metal complexes have close-lying spin and
electronic states and the crossing-over from one surface to the other affects
the reactivity patterns with substrates. Computational modeling has given
insight into the nature of metal-oxo, metal-peroxo, and metal-hydroperoxo
complexes. Some of these complexes are intermediates in the catalytic cycles
of heme and nonheme iron enzymes, but not all of these are reactive with
substrates. Our on-going work elucidates structure and reactivity and pre-
dicts trends.
ACKNOWLEDGMENTS
A.S.F. thanks the Tertiary Education Trust Fund Nigeria for a studentship. The Inorganic
Reaction Mechanisms Discussion group of the Dalton Division of the Royal Society of
Chemistry is thanked for support.
REFERENCES
1. Solomon, E. I.; Brunold, T. C.; Davis, M. I.; Kemsley, J. N.; Lee, S.-K.; Lehnert, N.;
Neese, F.; Skulan, A. J.; Yang, Y.-S.; Zhou, J. Chem. Rev. 2000, 100, 235–349.
2. Bugg, T. D. H. Curr. Opin. Chem. Biol. 2001, 5, 550–555.
3. Ryle, M. J.; Hausinger, R. P. Curr. Opin. Chem. Biol. 2002, 6, 193–201.
4. Costas, M.; Mehn, M. P.; Jensen, M. P.; Que, L., Jr. Chem. Rev. 2004, 104, 939–986.
5. Abu-Omar, M. M.; Loaiza, A.; Hontzeas, N. Chem. Rev. 2005, 105, 2227–2252.
6. de Visser, S. P., Kumar, D., Eds. Iron-Containing Enzymes: Versatile Catalysts of Hydrox-
ylation Reaction in Nature; RSC Publishing: Cambridge, UK, 2011.
7. Bruijnincx, P. C.; van Koten, G.; Klein Gebbink, R. J. Chem. Soc. Rev. 2008, 37,
2716–2744.
8. Sono, M.; Roach, M. P.; Coulter, E. D.; Dawson, J. H. Chem. Rev. 1996, 96,
2841–2888.
9. Ortiz de Montellano, P. R. Ed. Cytochrome P450: Structure, Mechanism and Biochemistry,
3rd ed.; Kluwer Academic/Plenum Publishers: New York, 2005.
10. Meunier, B.; de Visser, S. P.; Shaik, S. Chem. Rev. 2004, 104, 3947–3980.
11. Denisov, I. G.; Makris, T. M.; Sligar, S. G.; Schlichting, I. Chem. Rev. 2005, 105,
2253–2277.
12. Watanabe, Y.; Nakajima, H.; Ueno, T. Acc. Chem. Res. 2007, 40, 554–562.
13. Krebs, C.; Galonic Fujimori, D.; Walsh, C. T.; Bollinger, J. M., Jr. Acc. Chem. Res.
2007, 40, 484–492.
14. Hausinger, R. P. Crit. Rev. Biochem. Mol. Biol. 2004, 39, 21–68.
15. Hernández-Ortega, A.; Quesne, M. G.; Bui, S.; Heuts, D. P. H. M.; Steiner, R. A.;
Heyes, D. J.; de Visser, S. P.; Scrutton, N. S. J. Biol. Chem. 2014, 289, 8620–8632.
16. Hernández-Ortega, A.; Quesne, M. G.; Bui, S.; Heyes, D. J.; Steiner, R. A.;
Scrutton, N. S.; de Visser, S. P. J. Am. Chem. Soc. 2015, 137, 7474–7487.
17. Solomon, E. I.; Decker, A.; Lehnert, N. Proc. Natl. Acad. Sci. U. S. A. 2003, 100,
3589–3594.
18. de Visser, S. P. Coord. Chem. Rev. 2009, 253, 754–768.
19. O’Brien, J. R.; Schuller, D. J.; Yang, V. S.; Dillard, B. D.; Lanzilotta, W. N.
Biochemistry 2003, 42, 5547–5554.
20. Bollinger, J. M., Jr.; Price, J. C.; Hoffart, L. M.; Barr, E. W.; Krebs, C. Eur. J. Inorg.
Chem. 2005, 4245–4254.
21. Borowski, T.; Bassan, A.; Siegbahn, P. E. M. Chem. Eur. J. 2004, 10, 1031–1041.
22. de Visser, S. P. Chem. Commun. 2007, 171–173.
23. Godfrey, E.; Porro, C. S.; de Visser, S. P. J. Phys. Chem. A 2008, 112, 2464–2468.
24. Quesne, M. G.; Latifi, R.; Gonzalez-Ovalle, L. E.; Kumar, D.; de Visser, S. P. Chem.
Eur. J. 2014, 20, 435–446.
25. de Visser, S. P. Angew. Chem. Int. Ed. 2006, 45, 1790–1793.
26. de Visser, S. P. J. Am. Chem. Soc. 2006, 128, 9813–9824.
27. Latifi, R.; Bagherzadeh, M.; de Visser, S. P. Chem. Eur. J. 2009, 15, 6651–6662.
28. de Visser, S. P. J. Am. Chem. Soc. 2010, 132, 1087–1097.
29. Karamzadeh, B.; Kumar, D.; Sastry, G. N.; de Visser, S. P. J. Phys. Chem. A 2010, 114,
13234–13243.
30. Pratter, S. M.; Konstantinovics, C.; DiGiuro, C. L. M.; Leitner, E.; Kumar, D.; de
Visser, S. P.; Grogan, G.; Straganz, G. D. Angew. Chem. Int. Ed. 2013, 52, 9677–9681.
31. Straganz, G. D.; Nidetzky, B. ChemBioChem 2006, 7, 1536–1548.
32. Joseph, C. A.; Maroney, M. J. Chem. Commun. 2007, 32, 3338–3349.
33. Goncharenko, K. V.; Vit, A.; Blankenfeldt, W.; Seebeck, F. P. Angew. Chem. Int. Ed.
2015, 54, 2821–2824.
34. Goncharenko, K. V.; Seebeck, F. P. Chem. Commun. 2016, 52, 1945–1948.
35. Aluri, S.; de Visser, S. P. J. Am. Chem. Soc. 2007, 129, 14846–14847.
36. de Visser, S. P.; Straganz, G. D. J. Phys. Chem. A 2009, 113, 1835–1846.
37. Kumar, D.; Thiel, W.; de Visser, S. P. J. Am. Chem. Soc. 2011, 133, 3869–3882.
38. Kumar, D.; Sastry, G. N.; Goldberg, D. P.; de Visser, S. P. J. Phys. Chem. A 2012, 116,
582–591.
39. Sallmann, M.; Kumar, S.; Chernev, P.; Nehrkorn, J.; Schnegg, A.; Kumar, D.;
Dau, H.; Limberg, C.; de Visser, S. P. Chem. Eur. J. 2015, 21, 7470–7479.
40. Ortiz de Montellano, P. R. Chem. Rev. 2010, 110, 932–967.
41. Costas, M. Coord. Chem. Rev. 2011, 255, 2912–2932.
42. de Visser, S. P. Adv. Inorg. Chem. 2012, 64, 1–31.
43. de Visser, S. P.; Nam, W. In Handbook of Porphyrin Science; Kadish, K. M., Smith, K. M.,
Guilard, R., Eds.; World Scientific Publishing Co: New Jersey, 2010; pp 85–140.
Chapter 44.
44. de Visser, S. P.; Ogliaro, F.; Sharma, P. K.; Shaik, S. Angew. Chem. Int. Ed. 2002, 41,
1947–1951.
45. de Visser, S. P.; Ogliaro, F.; Sharma, P. K.; Shaik, S. J. Am. Chem. Soc. 2002, 124,
11809–11826.
46. de Visser, S. P.; Kumar, D.; Neumann, R.; Shaik, S. Angew. Chem. Int. Ed. 2004, 43,
5661–5665.
47. Kumar, D.; de Visser, S. P.; Sharma, P. K.; Derat, D.; Shaik, S. J. Biol. Inorg. Chem.
2005, 10, 181–189.
48. de Visser, S. P. J. Biol. Inorg. Chem. 2006, 11, 168–178.
49. Wang, R.; de Visser, S. P. J. Inorg. Biochem. 2007, 101, 1464–1472.
50. de Visser, S. P.; Shaik, S.; Sharma, P. K.; Kumar, D.; Thiel, W. J. Am. Chem. Soc. 2003,
125, 15779–15788.
51. Rittle, J.; Green, M. T. Science 2010, 330, 933–937.
52. Cryle, M. J.; De Voss, J. J. Angew. Chem. Int. Ed. Engl. 2006, 45, 8221–8223.
53. Fertinger, C.; Hessenauer-Ilicheva, N.; Franke, A.; van Eldik, R. Chem. Eur. J. 2009,
15, 13435–13440.
54. Franke, A.; Wolak, M.; van Eldik, R. Chem. Eur. J. 2009, 15, 10182–10198.
55. Ogliaro, F.; de Visser, S. P.; Cohen, S.; Sharma, P. K.; Shaik, S. J. Am. Chem. Soc.
2002, 124, 2806–2817.
56. Kamachi, T.; Shiota, Y.; Ohta, T.; Yoshizawa, K. Bull. Chem. Soc. Jpn. 2003, 76,
721–732.
57. Kitagishi, H.; Tamaki, M.; Ueda, T.; Hirota, S.; Ohta, T.; Naruta, Y.; Kano, K. J. Am.
Chem. Soc. 2010, 132, 16730–16732.
58. Ohta, T.; Liu, J.-G.; Naruta, Y. Coord. Chem. Rev. 2013, 257, 407–413.
59. Park, M. J.; Lee, J.; Suh, Y.; Kim, J.; Nam, W. J. Am. Chem. Soc. 2006, 128,
2630–2634.
60. Vaz, A. D.; McGinnity, D. F.; Coon, M. J. Proc. Natl. Acad. Sci. U. S. A. 1998, 95,
3555–3560.
61. Vardhaman, A. K.; Sastri, C. V.; Kumar, D.; de Visser, S. P. Chem. Commun. 2011, 47,
11044–11046.
62. Thibon, A.; Jollet, V.; Ribal, C.; Senechal-David, K.; Billon, L.; Sorokin, A. B.;
Banse, F. Chem. Eur. J. 2012, 18, 2715–2724.
63. Vardhaman, A. K.; Barman, P.; Kumar, S.; Sastri, C. V.; Kumar, D.; de Visser, S. P.
Chem. Commun. 2013, 49, 10926–10928.
64. Faponle, A. S.; Quesne, M. G.; Sastri, C. V.; Banse, F.; de Visser, S. P. Chem. Eur. J.
2015, 21, 1221–1236.
65. Faponle, A. S.; Banse, F.; de Visser, S. P. J. Biol. Inorg. Chem. 2016, 21, 453–462.
66. Kryatov, S. V.; Rybak-Akimova, E. V.; Schindler, S. Chem. Rev. 2005, 105,
2175–2226.
67. Stipanuk, M. H. Annu. Rev. Nutr. 2004, 24, 539–577.
68. Baldwin, J. E.; Bradley, M. Chem. Rev. 1990, 90, 1079–1088.
69. Lundberg, M.; Siegbahn, P. E. M.; Morokuma, K. Biochemistry 2008, 47, 1031–1042.
70. Kumar, D.; Hirao, H.; Shaik, S.; Kozlowski, P. M. J. Am. Chem. Soc. 2006, 128,
16148–16158.
71. Liu, L. V.; Bell, C. B., III; Wong, S. D.; Wilson, S. A.; Kwak, Y.; Chow, M. S.;
Zhao, J.; Hodgson, K. O.; Hedman, B.; Solomon, E. I. Proc. Natl. Acad. Sci. U. S.
A. 2010, 107, 22419–22424.
72. Stubbe, J.; Kozarich, J. W.; Wu, W.; Vanderwall, D. E. Acc. Chem. Res. 1996, 29,
322–330.
73. Burger, R. M. Chem. Rev. 1998, 98, 1153–1170.
74. Heafield, M. T.; Fearn, S.; Steventon, G. B.; Waring, R. H.; Williams, A. C.;
Sturman, S. G. Neurosci. Lett. 1990, 110, 216–224.
75. Perry, T. L.; Norman, M. G.; Yong, V. W.; Whiting, S.; Crichton, J. U.; Hansen, S.;
Kish, S. J. Ann. Neurol. 1985, 18, 482–495.
76. Brait, M.; Ling, S.; Nagpal, J. K.; Chang, X.; Park, H. L.; Lee, J.; Okamura, J.;
Yamashita, K.; Sidransky, D.; Kim, M. S. PLoS One 2012, 7, e44951.
77. Stipanuk, M. H.; Simmons, C. R.; Karplus, P. A.; Dominy, J. E., Jr. Amino Acids 2010,
41, 91–112.
78. Stipanuk, M. H.; Ueki, I. J. Inherited Metab. Dis. 2010, 34, 17–34.
79. Gonzalez-Ovalle, L. E.; Quesne, M. G.; Kumar, D.; Goldberg, D. P.; de Visser, S. P.
Org. Biomol. Chem. 2012, 10, 5401–5409.
80. McQuilken, A. C.; Goldberg, D. P. Dalton Trans. 2012, 41, 10883–10899.
81. Widger, L. R.; Davies, C. G.; Yang, T.; Siegler, M. A.; Troeppner, O.;
Jameson, G. N. L.; Ivanovic-Burmazovic, I.; Goldberg, D. P. J. Am. Chem. Soc.
2014, 136, 2699–2701.
82. Tchesnokov, E. P.; Faponle, A. S.; Davies, C. G.; Quesne, M. G.; Turner, R.;
Fellner, M.; Souness, R. J.; Wilbanks, S. M.; de Visser, S. P.; Jameson, G. N. L. Chem.
Commun. 2016, 52, 8814–8817.
83. Porro, C. S.; Sutcliffe, M. J.; de Visser, S. P. J. Phys. Chem. A 2009, 113, 11635–11642.
84. Thibon, A.; Bartoli, J.-F.; Guillot, R.; Sainton, J.; Martinho, M.; Mansuy, D.; Banse, F.
J. Mol. Catal. A 2008, 287, 115–120.
85. de Visser, S. P.; Oh, K.; Han, A.-R.; Nam, W. Inorg. Chem. 2007, 46, 4632–4641.
86. de Visser, S. P.; Tan, L. S. J. Am. Chem. Soc. 2008, 130, 12961–12974.
87. Sainna, M. A.; Kumar, S.; Kumar, D.; Fornarini, S.; Crestoni, M. E.; de Visser, S. P.
Chem. Sci. 2015, 6, 1516–1529.
88. Ji, L.; Faponle, A. S.; Quesne, M. G.; Sainna, M. A.; Zhang, J.; Franke, A.; Kumar, D.;
van Eldik, R.; Liu, W.; de Visser, S. P. Chem. Eur. J. 2015, 21, 9083–9092.
89. Faponle, A. S.; Quesne, M. G.; de Visser, S. P. Chem. Eur. J. 2016, 22, 5478–5483.
90. Quesne, M. G.; Senthilnathan, D.; Singh, D.; Kumar, D.; Maldivi, P.; Sorokin, A. B.;
de Visser, S. P. ACS Catal. 2016, 6, 2230–2243.
91. Jackson, T. A.; Brunold, T. C. Acc. Chem. Res. 2004, 37, 461–470.
92. Streit, B. R.; Blanc, B.; Lukat-Rodgers, G. S.; Lukat-Rodgers, K. R.; Dubois, J. L.
J. Am. Chem. Soc. 2010, 132, 5711–5724.
93. Siegbahn, P. E. M. Chem. Eur. J. 2008, 14, 8290–8302.
94. Neu, H.; Baglia, R. A.; Goldberg, D. P. Acc. Chem. Res. 2015, 48, 2754–2764.
95. Chen, Z.; Yin, G. Chem. Soc. Rev. 2015, 44, 1083–1100.
96. Prokop, K. A.; de Visser, S. P.; Goldberg, D. P. Angew. Chem. Int. Ed. 2010, 49,
5091–5095.
97. Prokop, K. A.; Neu, H. M.; de Visser, S. P.; Goldberg, D. P. J. Am. Chem. Soc. 2011,
133, 15874–15877.
98. Janardanan, D.; Usharani, D.; Shaik, S. Angew. Chem. Int. Ed. 2012, 51, 4421–4425.
99. Leeladee, P.; Baglia, R. A.; Prokop, K. A.; Latifi, R.; de Visser, S. P.; Goldberg, D. P.
J. Am. Chem. Soc. 2012, 134, 10397–10400.
100. Neu, H. M.; Yang, T.; Baglia, R. A.; Yosca, T. H.; Green, M. T.; Quesne, M. G.; de
Visser, S. P.; Goldberg, D. P. J. Am. Chem. Soc. 2014, 136, 13845–13852.
101. Neu, H. M.; Quesne, M. G.; Yang, T.; Prokop-Prigge, K. A.; Lancaster, K. M.;
Donohoe, J.; DeBeer, S.; de Visser, S. P.; Goldberg, D. P. Chem. Eur. J. 2014, 20,
14584–14588.
102. Yang, T.; Quesne, M. G.; Neu, H. M.; Cantú Reinhard, F. G.; Goldberg, D. P.; de
Visser, S. P. J. Am. Chem. Soc. 2016, 138, 12375–12386.
103. Cantú Reinhard, F. G.; Faponle, A. S.; de Visser, S. P. J. Phys. Chem. A 2016, 120,
9805–9814.
104. Vardhaman, A. K.; Barman, P.; Kumar, S.; Sastri, C. V.; Kumar, D.; de Visser, S. P.
Angew. Chem. Int. Ed. 2013, 52, 12288–12292.
105. Kumar, S.; Faponle, A. S.; Barman, P.; Vardhaman, A. K.; Sastri, C. V.; Kumar, D.; de
Visser, S. P. J. Am. Chem. Soc. 2014, 136, 17102–17115.
106. Barman, P.; Upadhyay, P.; Faponle, A. S.; Kumar, J.; Nag, S. S.; Kumar, D.;
Sastri, C. V.; de Visser, S. P. Angew. Chem. Int. Ed. 2016, 55, 11091–11095.
107. Nam, W. Acc. Chem. Res. 2007, 40, 522–531.

108. Coggins, M. K.; Kovacs, J. A. J. Am. Chem. Soc. 2011, 133, 12470–12473.
109. Zlatar, M.; Gruden, M.; Vassilyeva, O. Y.; Buvaylo, E. A.; Ponomarev, A. N.;
Zvyagin, S. A.; Wosnitza, J.; Krzystek, J.; Garcia-Fernandez, P.; Duboc, C. Inorg.
Chem. 2016, 55, 1192–1201.
110. Shaik, S. S. J. Am. Chem. Soc. 1981, 103, 3692–3701.
111. Shaik, S.; Kumar, D.; de Visser, S. P. J. Am. Chem. Soc. 2008, 130, 10128–10140.
112. Cantú Reinhard, F. G.; Sainna, M. A.; Upadhyay, P.; Balan, G. A.; Kumar, D.;
Fornarini, S.; Crestoni, M. E.; de Visser, S. P. Chem. Eur. J. 2016, 22, 18608–18619.
CHAPTER FIVE
Diarylplatinum(II) Scaffolds for

Kinetic and Mechanistic Studies on
the Formation of Platinacycles via
an Oxidative Addition/Reductive
Elimination/Oxidative Addition
Sequence
Gabriel Aullón, Margarita Crespo, Jesús Jover, Manuel Martínez1
Secció de Quı́mica Inorgànica, Facultat de Quı́mica, Universitat de Barcelona, Barcelona, Spain
1
Corresponding author: e-mail address: manel.martinez@qi.ub.es
Contents
1. Introduction 196
2. Compounds 198
3. Kinetic Studies 206
3.1 Spontaneous Processes With Monodentate N(imine) Ligands 209
3.2 Spontaneous Processes With Chelate N(amino)–N(imine) Ligands 215
4. Synergy With DFT Calculations 221
4.1 Five- vs Seven-Membered Platinacycle Stability 225
4.2 cis vs trans Stability for B-Type Intermediates 225
4.3 Formation of Five- or Seven-Membered Platinacycles From Bcis-Type
Intermediates 228
5. Concluding Remarks 239
Acknowledgments 240
References 240
Abstract
Oxidative addition and reductive elimination reactions are fundamental steps in pro-
cesses related to synthetic chemistry involving organometallic compounds. In these
reactions a metal in two available oxidation states (generally differing in two units) is
needed, platinum centers being a very good example. The relative inertness of diamag-
netic PtII and PtIV organometallic species (having, respectively, d8 square-planar or d6
octahedral arrangements), enables an easy monitoring of time-resolved reactivity,
including its posterior kinetic analysis. Specifically, imine ligands containing C–X bonds
have been observed to oxidatively add to {PtII(Aryl)2} moieties, which sequentially
undergo C–C reductive elimination and C–H bond activation on the new ligand formed.

196 Gabriel Aullón et al.
These new species have been found to contain mostly seven-membered metallacycles,
despite the obvious thermodynamic preference for five-membered cycles, which are
found only in some rather specific instances. The kinetic preference of the complexes
obtained has been studied from a kinetico-mechanistic perspective, that included
obtaining thermal- and pressure-derived activation parameters, and a dramatic influ-
ence on the spectator halido ligands, and the substituents on the aryl groups has been
established. To complete this kinetic and mechanism (kinetico-mechanistic) study, the-
oretical calculations have also been conducted to model the data collected and pro-
pose both the elementary steps and the factors determining the specificity of the
full process.
1. INTRODUCTION
Kinetics experiments in solution lead to the collection of kinetic data
that can be analyzed in terms of rate laws, which together with activation
parameters acquired from temperature and pressure variable experiments,
may lead to mechanistic proposals. The value of such analyses and proposals
is compromised unless due diligence regarding reactant and solvent purity is
followed, and furthermore reproducibility of measurements is assured, and
primary data and derived secondary data and parameters are subject to
appropriate error analysis.
In the last several decades the throughput of kinetics experiments has
changed dramatically by instrument improvement, automation, and vali-
dated data analysis software. Developments in chemical synthesis have
yielded a wider range of subtle reactant variations leading to wider data sets.
Consequently, provided the caveats within the experimental approach are
followed, reliable new reaction mechanisms frequently ensue.
In order to understand further the detail of reaction mechanism path-
ways, the experimentalists have, over the last two to three decades, at their
disposal, the ability to take advantage of the application of density functional
theory (DFT) computation methodology. This methodology represents a
significant and appealing addition to the overall repertoire of approaches
by investigators for a detailed explanation of the time course, energetics,
and structural aspects of their chemical reaction systems. Successes in using
this computational approach are widely reported. Indeed, it will be shown in
this contribution that a combination of kinetics and mechanism studies and
modeling by DFT computations has led to a comprehensive understanding
of the formation of platinacycles through various stages starting from
diarylplatinum(II) scaffolds.
Diarylplatinum(II) Scaffolds for Kinetic and Mechanistic Studies 197
Oxidative addition and reductive elimination are fundamental reactions

in organometallic chemistry in both stoichiometric and catalytic processes.
The inertness of platinum compounds and the availability of different oxi-
dation states make them suitable for mechanistic studies of these reactions. In
particular, although cyclopalladated complexes are much used in catalytic
processes, their equivalent platinum compounds have been often used as
model compounds for fundamental reactions due to their high stability
and inertness. The syntheses of cyclometallated compounds traditionally
involve intramolecular C–H bond activation, but a great deal of interest
is also focused in intramolecular C–X bond activation since, in these cases,
the initial oxidative addition can be followed by a subsequent reductive
elimination producing new σ bonds such as C–C, C–O, C–H, or C–X.
Among the large plethora of cycloplatinated compounds available, those
containing aryl ligands might give rise to oxidative addition/reductive elim-
ination process involving C–X and C–H bond activation as well as Caryl–
Caryl reductive elimination. Therefore, arylplatinum compounds provide
a useful platform for a thorough study of these fundamental processes.
The relative inert character of such complexes allows the monitoring of
their time-dependent reactivity in an easy and reproducible way. Further-
more, the possibility of detecting and isolating reaction intermediates pre-
cisely increases with the increase of the inert character on the reactivity
involved. Unfortunately, the possible actuation of dead-end processes also
increases along the same line, and a comprehensive monitoring of all the
processes involved in the reactivity studied is desirable. In this respect, the
use of experimental mechanistic procedures has been suffering a decrease
in use owing to the preference by some for computational methods, which
usually come into play when the situation becomes too complicated to mon-
itor experimentally, or when the number of experiments needed is exces-
sive. However, we have to keep in mind that computational methods
have to be employed wisely; although calculations may corroborate or even
guide experiments, the final conclusions have always to be ascertained by
proper real experimental situations.
Platinum chemistry is a terrain that has been explored thoroughly by
computational means. A tremendous number of publications regarding
homogeneous catalysis, surface science, or medicinal applications are
reported each year. Bond activation by platinum complexes, and its impli-
cation in homogeneous catalysts, has been explored thoroughly. In partic-
ular cycloplatination reactions are a convenient computational area of
study. Association and dissociation steps, along with well-known oxidative
addition and reductive elimination processes between closed-shell diamag-

netic PtII/PtIV species, are involved in the reactivity. Furthermore, relativ-
istic effects for platinum, which should be the most remarkable issue
in computational studies of these reactions, can be effectively tackled
nowadays by most DFT methods. Either the use of scalar relativistic
calculations or basis sets with pseudopotentials (that include the corres-
ponding corrections for the heavier atoms) represent a standard solution
to the issue.
This chapter describes a series of experimental, kinetics, mechanistic, and
computational studies on the formation of five- and seven-membered met-
allacycles obtained from diarylplatinum(II) precursors with N-donor
ligands. One aim of the material presented is to illustrate the enormous
advantage of synergic collaborations between specialized research groups
that adds great value to the studies conducted. However, by doing so, some
aspects, which are normally taken for granted, have to be discussed and
reformulated to account for the overall set of observations and results.
2. COMPOUNDS
Although the complexes [PtMe2(NN)], where NN is a diimine ligand,
such as 2,20 -bipyridine or 1,10-phenanthroline, are among the most reactive
transition-metal complexes in intermolecular oxidative addition of alkyl
halides, aryl halides fail to react with these PtII complexes. In this respect,
the first reported oxidative additions of directed aryl halides to PtII have been
achieved using ligands of formula RCH]NCH2CH2NMe2 (1–3). In these,
R is an ortho-halogenoaryl group, and the ligand coordinates via both nitro-
gen atoms to a dimethylplatinum(II) center producing a [PtMe2(RCH]
NCH2CH2NMe2)] species that ultimately leads to intramolecular oxidative
addition of the aryl-halogen bond. This strategy was successful in producing
C–X (X ¼ Br, Cl, F) bond activation leading to PtIV compounds of type
[PtMe2X(C5CH4CH]NCH2CH2NMe2)]. In the same studies C–H bond
activation was also achieved; in this case a final Me–H reductive elimination
producing methane and PtII compounds of type [PtMe(C5CH4CH]
NCH2CH2NMe2)] takes place (Scheme 1, top).
Further work in this area involved the use of ligands with the same type
of organometallic moieties, but with a single N-donor directing group, with
a general formula 2-XC6H4CH]NCH2-20 -X0 C6H4 (4–6). In this case, the
reaction of these ligands with [Pt2Me4(μ-SMe2)2] may produce different
Scheme 1 Oxidative addition reaction mechanism for ligands of general formulae RCH]NCH2CH2NMe2, 2-XC6H4CH]NCH2Ar, and ArCH]
NCH2(2-XC6H4) on [Pt2Me4(μ-SMe2)2].
types of five-membered platinacycles, either containing (endocycles) or not

(exocycles) the imine functionality. As for the amino–imino ligands indi-
cated earlier, intramolecular oxidative addition of C–X bonds (X ¼ Br,
Cl, F) produced cyclometallated PtIV compounds, while activation of
C–H bonds with loss of methane gave the corresponding cyclometallated
PtII compounds. Interestingly, formation of endo-platinacycles (Scheme 1,
bottom) is favored in all cases for these ligands. Exo-platinacycles have
only been found when imines 2,4,6-C6Me3H2CH]NCH2(2-XC6H4)
(X ¼ Br, Cl), in which formation of endo-metallacycles is precluded by
the presence of methyl substituents in the ortho positions of the benzyl ring,
are used.
Given the successful results obtained for the intramolecular C–X activa-
tion mechanisms at the {PtIIMe2} moieties, our attention was focused on
substrates having the {PtII(Aryl)2} organometallic unit. The purpose being
both to determine whether or not this type of compound could produce
cyclometallated compounds under similar mild conditions and to compare
the mechanism operating for both series of compounds. The initial work
involved preparative studies using cis-[PtPh2(SMe2)2] as substrate indicated
that intramolecular C–X (X ¼ Br or Cl) bond activation may occur for
both ligands containing one (7,8) or two nitrogen donor atoms (9,10). Anal-
ogous results were obtained using the better spectroscopically handled
[Pt2(4-MeC6H4)4(μ-SEt2)2] (11,12) or [Pt2(4-FC6H4)4(μ-SEt2)2] (13,14)
starting materials. As previously reported for [Pt2Me4(μ-SMe2)2], C–X
(X ¼ Br or Cl) bond activation at diarylplatinum(II) compounds leads to
cyclometallated PtIV compounds or PtII after reductive elimination of arene
for C–H bond activations. On the other hand, although intramolecular C–F
bond activation has been achieved upon reaction of [Pt2Me4(μ-SMe2)2]
with the ligand C6F5CH]NCH2CH2NMe2 (3) such a process has not been
observed for precursors [PtPh2(SMe2)2] or cis-[Pt(4-MeC6H4)2(μ-SEt2)]2.
This result can be related to the strength of the C–F bond and the lower
reactivity of the arylplatinum reagents when compared to methyl analogues.
A distinct feature of the mentioned reactions on bis-aryl PtII compounds
was disclosed in the preliminary study of compound cis-[PtPh2(SMe2)2]
with ligand 2-BrC6H4CH]NCH2C6H5 (7) in which the initially
formed cyclometallated PtIV compound has evolved to a novel type of
seven-membered platinacycle that was structurally characterized (Fig. 1) as
the compound [PtIIBr(CC5H4C6H4CHNCH2C6H5)(SMe2)] containing
a biaryl linkage. This seminal work was followed by the report of
Fig. 1 Molecular structure of the seven-membered cyclometalated platinum(II) com-

pound [PtIIBr(CC5H4C6H4CHNCH2Ph)(SMe2)].
further examples of seven-membered platinacycles such as [PtIIBr

(CC5H4C6H4CHNR)(SMe2)] (R ¼ Me or CH2Mes) obtained from the reac-
tion of cis-[PtPh2(SMe2)2] with the corresponding N-donor ligands (8). Further
examples based on the reaction of [Pt2(4-MeC6H4)4(μ-SEt2)2] or
[Pt2(4-FC6H4)4(μ-SEt2)2] with analogous imines containing one single
N-donor atom proved that not only initial C–Br bond activation but also
C–Cl bond activation may lead to such seven-membered platinacycles as indi-
cated in Scheme 2 (11,12). It is interesting to note that these compounds
are obtained in good yields not only when the ortho position of the imine is
blocked with a fluorine substituent (Y ¼ F) but also when Y ¼ H, where a
C–H bond activation leading to more stable five-membered PtII metallacycles
is also possible.
For these compounds, the fate of the substituent in position 4 of the aryl
group (Z ¼ Me or F), which in the final seven-membered platinacycles, is
meta to the platinum center, is consistent with the reaction sequence indi-
cated in Scheme 2. The process thus consists of: (i) initial C–X (X ¼ Br
or Cl) bond activation to produce a starting PtIV derivative, (ii) reductive
elimination coupling with formation of a Caryl–Caryl bond between one
of the aryl ligands and the metallated aryl of the imine ligand, (iii) final
cyclometallation with elimination of an arene molecule. As discussed in
detail later, combined 1H NMR and UV–vis kinetico-mechanistic studies
carried out recently for the formation of seven-membered [C,N]-
platinacycles from cyclometallated PtIV compounds (13) indicate that the
Scheme 2 Formation of seven-membered [C,N]-PtII metallacycles.
formal cyclometallation process (iii) in fact involves an isomerization step (iv,

Scheme 2), where the resulting noncyclometallated intermediate adopts the
required geometry for the final cycloplatination reaction (step v, Scheme 2).
The distinct values of J(H–Pt) for the imine trans to C (ca. 44 Hz) or trans to
X (ca. 140 Hz) allow identification and monitoring of the isomerization
reaction. NMR spectroscopic monitoring of the full reaction indicates that
the rate-determining step of the process is found to depend on the nature of
the substituent (Z ¼ H, Me, or F) at the para position of the aryl ligand; no
relevant differences in the synthesis of seven-membered [C,N]-platinacycles
were found when SMe2 or SEt2 derivatives were used or even when the
halido ligand was Br or Cl.
Interestingly, the reaction of cis-[Pt(C6F5)2(SEt2)2] with imine
2-BrC6H4CH]NCH2(4-ClC6H4) (Scheme 3) produced exclusively a
five-membered metallacycle with an exo Caryl–Caryl bond. In this case, for-
mation of a seven-membered platinacycle which requires C–F bond activa-
tion is not favored and indeed is not observed. 19F NMR monitoring allows
the detection of all proposed intermediates (15), i.e., initial formation of a
PtIV compound arising from the activation of the C–Br bond of the imine,
Caryl–Caryl reductive elimination, and final Caryl–H bond activation leading
to a five-membered cyclometallated PtII compound with concurrent elim-
ination of pentafluorobenzene.
Scheme 3 Formation of a five-membered [C,N]-PtII metallacycle.
Scheme 4 Formation of several types of [C,N]-PtII metallacycle.
In order to analyze the scope of the formation of biaryl linkages in the

coordination sphere of platinum, the reactions of [Pt2(4-MeC6H4)4
(μ-SEt2)2] with N-benzylidenebenzylamines for which formation of exo-
metallacycles is favored (as indicated earlier) were also studied. Imines con-
taining a C–Br or a C–Cl bond in an ortho position of the more flexible
benzyl ring effectively also lead to formation of biaryl linkages between
one of the 4-tolyl ligands and the benzyl group of the imine ligand. How-
ever, the biaryl linkage formed was found to be not necessarily involved in
the subsequent metalation which produced either endo-five, endo-six, or exo-
five-membered platinacycles as shown in Scheme 4 (16,17). The reactions
observed follow the general sequence indicated earlier, but with the final
cyclometallation step being more favored for Caromatic–H than for
Caliphatic–H bonds and for endo- than for exo-metallacycles, so that the latter
are formed only when the ortho positions of the benzylidene rings are
blocked with fluorine atoms.
Summarizing, all the reactions shown in Schemes 2–4 for the formation
of [C,N]-PtII metallacycles containing a biaryl linkage, included or not in
the final platinacycle, involve intramolecular C–X (X ¼ Br or Cl) bond acti-
vation followed by Caryl–Caryl reductive elimination to produce non-
cyclometallated intermediates that could be detected when the reactions
were monitored by 1H or 19F NMR spectroscopy (13,15). However, none
of these intermediates could be isolated and characterized crystallographi-
cally. Nevertheless, ligands containing two nitrogen donor atoms such as
Ar0 CHNCH2CH2NMe2 would be expected to produce more stable inter-
mediates containing a [N,N0 ]-chelate.
In contrast to the results indicated so far for [C,N] systems (Scheme 2, com-
pounds of type A), early studies carried out with [C,N,N0 ]-PtIV cyclometallated
compounds of the general formula [PtXAr2(Ar0 CHNCH2CH2NMe2],
revealed a distinct behavior for X ¼ Br or X ¼ Cl with respect to the reactivity
indicated in Schemes 2 and 3 (18). For X ¼ Cl, seven-membered [C,N,N0 ]-PtII
cyclometallated compounds of general formula [PtCl(Ar-Ar0 CH]
NCH2CH2NMe2)] are easily obtained (10–12,14,18,19) (Scheme 2), and
crystal structure determinations have been reported for representative examples
(see Fig. 2). Moreover antitumor properties have been studied for some of them
(14,19).
In contrast, a five-membered [C,N,N0 ]-PtII cyclometallated compound
containing an external biaryl linkage (Scheme 3) is formed from [PtBr(4-
MeC6H4)2(CC5H4CHNCH2CH2NMe2] under analogous conditions. In
an attempt to understand the different chemical behavior observed for
bromo or chloro-[C,N,N0 ] systems, these reactions have been thoroughly
studied (18). Careful selection of the reaction conditions extracted from
the kinetic studies indicated in Section 3 allowed the detection and charac-
terization, including by X-ray crystallography, of two isomers of the non-
cyclometallated PtII compound arising from Caryl–Caryl reductive
elimination, thus supporting the existence of the reaction intermediates
assumed for the reactivity with [C,N] systems.
In this respect and in an attempt to obtain a seven-membered [C,N,N0 ]-
PtII cyclometallated compound for X ¼ Br, the reaction of [Pt2(4-
MeC6H4)4(μ-SEt2)2] with imine 2-Br,6-FC6H3CH]NCH2CH2NMe2
Fig. 2 Molecular structure of [PtIICl(2-FCC5H3(4-MeC6H3CHNCH2CH2N(CH3)2)].
Scheme 5 Formation of five- or seven-membered [C,N,N0 ]-platinacycles.
has also been studied (11), and, effectively, the presence of an inert C–F
bond at the ortho position of the aryl ring of the imine ligand prevents the
C–H activation at the imine, and thus, the reaction is driven toward the
formation of a seven-membered platinacycle containing an internal biaryl
linkage (Scheme 5). Again, for this system two isomers of the non-
cyclometallated compound containing a biaryl ligand were isolated and crys-
tallographically characterized.
3. KINETIC STUDIES
Reactions involving platinum organometallic complexes with simple
cis-{PtIIR2} units (R ¼ Me, Ph) have been proved to be extremely well
behaved for the directed oxidative addition reaction of C–X or C–H bonds
indicated in Section 2 (Scheme 1) (4–6,9,20,21). This has allowed us to
study from a kinetico-mechanistic perspective the reaction with the ligands
indicated in Scheme 6 having monofunctional (imine) or bifunctional
(amino-imine) directing units (22).
Independently of the directing groups used, the process has been found
to be occurring via the initial formation of an unsaturated tricoordinated
species that reacts in a concerted oxidative addition fashion to produce a
pentacoordinated intermediate (23,24). The rapid coordination of the sixth
(dangling or available in the medium) ligand produces the final compound
(Scheme 1). Fig. 3A collects the isokinetic plot generated with all the data
available, indicating that the process occurs indeed via the same concerted
tricentered mechanism (22). Even the available data for systems lacking
hydrogen bonding characteristics show a very good and extended enough
ΔV{/ΔS{ correlation (25) (Fig. 3B).
Scheme 6 Ligands utilized in the preliminary studies of oxidative addition reactions on

cis-{PtIIR2} units.
A 100
50
ΔS⫽ ⲐJ K–1mol–1
–50
–100
–150
–200
–250
15 30 45 60 75 90 105 120 135
ΔH⫽ ⲐkJ mol–1
B 20
ΔV⫽ Ⲑcm3mol–1
–20
–40
–250 –200 –150 –100 –50 0 50 100
ΔS⫽ ⲐJ K–1mol–1
Fig. 3 Isokinetic (A) and ΔV{/ΔS{ (B) correlation plots for the series of concerted oxida-
tive addition C–X and C–H activation reactions of compounds indicated in Scheme 6 on
{PtIIR2} (R ¼ Me, Ph) moieties.
Interestingly, as stated in Section 2, for C–X and R ¼ Ph bond activation

reactions, the isolation of the final PtIV compound, type A, indicated in
Scheme 1 proved to be rather difficult (7,8). Both kinetics and time-resolved
NMR spectra indicate the existence of a subsequent consecutive process
(Scheme 2) that produces a seven-membered platinacycle, type 7C, as the
final crystallized species (Fig. 1). Detailed kinetic studies of the oxidative
addition of some of the ligand molecules indicated in Scheme 6 on a series
of cis-{PtII(Aryl)2} units proved, in fact, to have a rather complex and
diverse behavior. Although, in some cases well-behaved two consecutive
kinetic steps were observed, leading initially to complexes of type A and
finally to species of type 7C (13); in some other cases, the kinetic profiles
showed up to four recognizable time-resolved steps (11,18). Parallel
NMR time-resolved monitoring was thus needed in order to ascertain what

are the processes observed. Furthermore, the nature of the final platinum
species formed was also found to be extraordinarily dependent on a plethora
of variables.
The general reaction scheme that could be extracted from the data
collected is already shown in Scheme 2, where the first step, (i), corre-
sponds to the process indicated in Scheme 1 for X 6¼ H (formation of
compounds of type A) (22). The steps following this first oxidative addi-
tion reaction, (ii) and (iii), correspond to a reductive elimination coupling
(formation of compound of type B) (15), and a new oxidative addition
process of the biaryl ligand formed (formation of compound of type C)
(11,13,18). With these data at hand, and given the good time-profile
knowledge acquired for the formation of complexes of type A, isolation
of such a PtIV five-membered metallacycle has been possible in most of
the cases. Consequently, the (ii) + (iii) set of reactions (producing com-
pounds of type B and C) could be studied independently. Interestingly,
the careful time-resolved isolation/characterization of the species involved
in the process leads to the obtention of rather diverse reaction intermedi-
ates and products, which corresponds to the more detailed reaction
sequence indicated in Scheme 5. As a whole, the species shown in
Scheme 7 have been characterized, thanks to the existence of kinetically
collected time/temperature data. Clearly from the experimental
Scheme 7 Models of characterized products and intermediate species occurring during

the set of reactions indicated in Scheme 2; L ¼ SMe2, SEt2, or Me2N(chelate) from ligands in
Scheme 6.
observation of these complexes, step (iii) in Scheme 2 should effectively

include two processes: an equilibrium isomerization reaction, indicated
by (iv), plus a final metalation reaction, (v) (11,18).
3.1 Spontaneous Processes With Monodentate N(imine) Ligands

The serendipitous seminal kinetico-mechanistic work of this reactivity series
has been successfully concentration-tuned to evaluate in a separate manner,
reductive elimination coupling, and final cycloplatination reactions
(Scheme 2) (7,8). The studies have been conducted taking advantage
of the comprehensively studied substitution reactions on the initial PtIV
complexes (A, L ¼ SMe2, Aryl ¼ Ph, X ¼ Br, Scheme 8), found to react
via a pentacoordinated intermediate (iA) (26–30). This knowledge allowed
for the sequestering of the reductively coupled intermediate appearing
in the reaction mechanism (iB) with excess of SMe2 (7). For studies with-
out excess of sequestering ligand a rather different set of kinetic and activa-
tion parameters was obtained, which should, consequently, be associated
with the B ! 7C reaction steps (8). The relevant kinetic features
determined for the two steps observed for this system, including thermal-
and pressure-derived activation parameters, are collected in Table 1. The
possible isomerization reaction occurring on the complex of type 7C has
also been studied with relevance to the final stereochemistry of the species
isolated (8).
As a follow up from these studies the kinetico-mechanistic monitoring
of the reaction of the cis-{PtII(C6F5)2} unit with ligand 6, where R ¼ H
and R0 ¼ 4-Cl, was also pursued (Scheme 6) (15). The process shows a
neat two-step rate-limiting sequence perfectly associated with reactions
(ii) + (iii) in Scheme 2. That is, reaction in Scheme 9 (with L ¼ SMe2,
Aryl ¼ C6F5, X ¼ Br) applies to this process which modifies the reactivity
indicated earlier. In this case the final compound corresponds to a five-
membered platinacycle with a dangling C6F5 unit of type 5C. Interest-
ingly, the measured kinetics indicates that process (ii) is very fast, and
the initial oxidative addition of ligand 6 (with R ¼ H and R0 ¼ 4-Cl)
on cis-[PtII(C6F5)2(SEt2)2] (step (i) in Scheme 2) cannot be resolved;
i.e., the compound of type A is not observed, only species B is initially
present as an intermediate. Table 1 collects the relevant kinetic and
activation parameters collected for the resolved reaction in Scheme 9
(B → 5C).
Scheme 8 Concentration-tuned reactions for the spontaneous reductive elimination/oxidative addition sequence on complexes of type A.
Further kinetico-mechanistic studies have also been carried out on com-

plexes with a systematic variation of the PtII-attached aryl ligands in the
starting cis-{PtII(Aryl)2} moiety. Furthermore, an ortho substituent on the
initial metallated ligand has been introduced (ligand 6 in Scheme 6, with
R ¼ 2-F and R0 ¼ H) (13). The systems have been tuned from Aryl ¼ Ph
to 4-MeC6H4 and 4-FC6H4; in all cases, the final characterized compound
corresponds exclusively to an organometallic complex of type 7C. That is,
the reaction sequence indicated in Scheme 8 is operative for the spontaneous
reactivity on these complexes.
For the two systems, as for the previous cis-{PtII(C6F5)2} system, a single
step leading to the final complex is observed from the isolated complexes
of type A. Careful parallel time-resolved NMR monitoring of the reaction
samples indicates that such rate-determining step monitored corresponds
to the reaction B ! 7C shown in Scheme 8 for the Aryl ¼ 4-MeC6H4
and 4-FC6H4 complexes, while for the Aryl ¼ Ph compound the rate-
determining step corresponds to the reductive A ! B coupling. Table 1 col-
lects the relevant kinetic and thermal- and pressure-derived activation
parameters determined for these systems. Clearly the kinetic and activation
parameters of reactivity observed fall within the expected values for the
Aryl ¼ Ph compound, but two distinct sets are observed for
Aryl ¼ 4-MeC6H4 and 4-FC6H4 complexes; Fig. 4 shows clear indication
of the diverse trends observed for these systems.
A close examination of time-resolved NMR monitoring indicated that
the reductively eliminated type B complex undergoing the final formation
of the seven-membered platinacycle (compound type 7C) is not the same in
both cases. While for the Aryl ¼ 4-MeC6H4 complexes the reaction is seen
from a B species with a cis-N(imine)/Aryl stereochemistry (Bcis, Scheme 7),
for the Aryl ¼ 4-FC6H4 compounds the stereochemistry is trans (Btrans,
Scheme 7). Clearly the activation process leading to the final 7C complexes
is either very distinct for the Bcis and Btrans complexes, or one of the pro-
cesses monitored corresponds to a rate-determining isomerization reaction
between these two stereochemical entities (i.e., slow Bcis > Btrans and
fast Bcis or Btrans ! 7C). DFT studies (see Section 3.2) were conducted
to ascertain the thermodynamic stereochemical preference of compounds
of type B, and effective species of type Bcis are found to lie lower in
energy than the corresponding trans isomers (11). Consequently, the full
A–C process can be better described as in Scheme 10 (similar to
Scheme 2) with the pentacoordinated and tricoordinated iA and iB inter-
mediates maintaining the original stereochemistry of compounds A or B.
Table 1 Relevant Kinetic and Activation Parameters of the Spontaneous Reductive

Compound A A!B
ǂ
k (solvent, T) ΔH ΔSǂ ΔV ǂ
/s1 /kJ mol1 /J K1mol1 /cm3 mol1
Ph Ph
SMe2
Ph Pt N 3.6 103
93 3 29 10 12.8 0.6
(CHCl3, 298 K)
Br
Ph
SMe2
Ph Pt N 7.3 104
94 3 12 9 9.5 1.1
(CHCl3, 298 K)
Br
C6F5 4-ClC6H4
SEt2
C6F5 Pt N Fast
Br
4-MeC6H4 Ph
SEt2
4-MeC6H4 Pt N
Fast
F Br
4-MeC6H4 Ph
SMe2
4-MeC6H4 Pt N Fast
F Br
4-MeC6H4 Ph
SMe2
4-MeC6H4 Pt N Fast
Br
Ph
Ph
SMe2
Ph Pt N 10 103
98 4 22 27 2
(toluene, 340 K)
F Br
4-FC6H4 Ph
SEt2
4-FC6H4 Pt N Fast
F Br
a
Corresponds to the isomerization process indicated in Scheme 10.
Elimination/Oxidative Addition Reactions Measured on Compound of Type A (Scheme 8)

B ! 7C B ! 5C
ǂ ǂ ǂ ǂ
k (solvent, T) ΔH ΔS ΔV k (solvent, T) ΔH ΔSǂ ΔV ǂ
/s1 /kJ mol1 /J K1 mol1 /cm3 mol1 /s1 /kJ mol1 /J K1 mol1 /cm3 mol1
3.0 104
91 7 9 24 0 — — — —
(CHCl3, 298 K)
7.8 105
100 4 10 1 0 — — — —
(CHCl3, 298 K)
8.5 105
— — — — 86 8 85 22 34 4
(xylene, 361 K)
9.5 104
65 1 115 5 0 — — — —
(toluene, 340 K)
6.7 104
83 7 –65 21 0 — — — —
(toluene, 340 K)
13 104
75 3 –83 9 0 — — — —
(toluene, 340 K)
Fast — — — —
2.3 103
84 5 –71 16 –21 3 — — — —
(toluene, 340 K)a
Scheme 9 Modification of the reactivity shown in Scheme 8 for the L ¼ SEt2, Aryl ¼ C6F5, X ¼ Br system producing a five-membered final
platinacycle.
P/ atm
0 300 600 900 1200 1500 1800 2100
–7.0
–7.7
Ink
–8.4
–9.1
–9.8
–9
In(k/T )
–12
–15
–18
0.0030 0.0031 0.0032 0.0033 0.0034 0.0035 0.0036

T −1/k−1
Fig. 4 Eyring and lnk vs P plots for the rate-determining step observed in the reaction
indicated in Scheme 8 with Aryl ¼ Ph, L ¼ SMe2, X ¼ Br, ligand 4 from Scheme 6 with
Aryl ¼ Ph (circles), and ligand 4 from Scheme 6 with Aryl ¼ 4-FC6H4, Ph (squares).
The kinetic and activation parameters associated with the rate-determining

step for Aryl ¼ 4-FC6H4 thus do not correspond to a cyclometalation reac-
tion, but to an isomerization from Btrans to Bcis thus explaining the dif-
ferences observed.
3.2 Spontaneous Processes With Chelate N(amino)–N(imine)

Ligands
In view of the data shown so far the studies were extended to systems where
the relatively labile ligand on the structure (L in the previous schemes) was
substituted by the dangling arm of a chelate ligand. The purpose of such
change is the isolation of some of the reductively coupled species of type
B (Schemes 8–10), owing to the increase in stability expected by chelation.
That is, ligands of the family 1, 2, and 3 shown in Scheme 6 were used in
the scaffold for the initial oxidative addition of cis-{PtII(Aryl)2} moieties to
produce compounds of type A with L ¼ Me2 NðCH2 Þ2 . Scheme 11 col-
lects the drawing of the structures of the compounds which have been
studied and described in Section 3.1, but with the bidentate chelating
N(amino)–N(imine) ligands instead of the N(imine) plus L monodentate ligands
set (Schemes 7–10).
As indicated in Table 1, for the L, N(imine) systems, isolation of the final
compound as a 5C species was only registered for the cis-{PtII(C6F5)2}
Scheme 10 Complete reductive elimination/isomerization/oxidative addition spontaneous reactivity of cyclometallated PtIV complexes of
type [Pt(Aryl)2X(C5CH4CH]N)L].
Scheme 11 Models of the species occurring during the set of reactions indicated in
Scheme 2 for ligands of type 1, 2, or 3 from Scheme 6.
moiety (Scheme 9), possibly due to the difficulty of the activation of a C–F
bond in the species B formed as an intermediate. For the rest of the systems
studied the final compounds isolated have always the structure of complexes
of type 7C (Scheme 8). Nevertheless for the systems with the N(amino)–
N(imine) chelating ligands the outcome of the full process from the
{PtII(Aryl)2} units is more diverse, the reactivity indicated in Scheme 5 from
Section 3.1 being a clear summary. As a general rule substituting the ligands
of type 2 and 3 in Scheme 6 at the remaining ortho position leads to a com-
plex of type A that reductively couples with one of the Aryl ¼ 4-MeC6H4
ligands to produce compounds of type B that only produce final
cyclometallated complexes of type 7C as found for the systems indicated ear-
lier. Nevertheless, for the ortho unsubstituted complexes of type A the for-
mation of the more thermodynamically stable complexes of type 5C is also
observed in some cases. This fact indicates that for these chelate complexes
the formation of seven-membered metallacycles of type 7C can be associ-
ated, at least in part, with the presence of a C–Y bond that is too strong
to produce such a cyclometalation reaction.
The summary of this chemistry is indicated in Scheme 5 from

Section 3.1. Scheme 12 collects the full series of species of type 5C and
7C encountered as the final compounds produced from the spontaneous
reductive elimination/oxidative addition sequence occurring on com-
plexes of type A. The kinetic study of the process occurring on compounds
of type A proved to be much more complex than expected, with a very
diverse behavior (11,18), in line with that observed for the complexes
in the last entries of Table 1. Fig. 5 shows the Eyring plot of all the
time-resolved reaction steps observed for the spontaneous reactivity of
the bromo complex of type A with no chloro or fluoro substituents on
the initial metallated ligand.
From this plot it is evident that at least three steps can be resolved
by using carefully tuned conditions. Furthermore, careful choice of time/
temperature conditions allowed for the isolation of some of the species appe-
aring during the process. As indicated in Scheme 11 reductively coupled
intermediates have been isolated and fully characterized (Fig. 6).
Scheme 12 Series of species of type 5C and 7C encountered as the final compounds

produced from the spontaneous reductive elimination/oxidative addition sequence
occurring on the complexes of type A indicated.
A –8 First step observed at high temperature

Second step observed at high temperature
First step observed at low temperature
–10 Second step observed at low temperature
–12
Ink/T
–14
–16
–18
0.0028 0.0029 0.0030 0.0031 0.0032 0.0033
T –1 / k –1
B
–8 First step observed
Second step observed
–10
–12
Ink/T
–14
–16
–18
0.0028 0.0029 0.0030 0.0031 0.0032 0.0033
T –1 / k –1
Fig. 5 Eyring plots for reactions observed by UV–vis monitoring of the sequential spon-
taneous process occurring in toluene/xylene solutions of bromo complex with no chloro
or fluoro substituents on the initial metallated ligand (Scheme 12) of type A (A) and
Btrans (B). Fitted lines are common to both graphs.
From these intermediate compounds the formation of the final type 5C

or 7C compounds was also followed from a kinetic perspective. This pro-
cedure allowed the association of the steps observed to different single reac-
tions by careful parallel time-resolved NMR monitoring. Fig. 5B is a clear
example of this association for the simplest system; the circle points in the
plot correspond to the formation of the final 5C complex, while the triangle
points in the plot correspond to the Btrans > Bcis reaction, thus leaving the
Fig. 6 Reductively coupled species of type Btrans and Bcis appearing during
the spontaneous type A to type 5C (top, no chloro or fluoro substituents on the initial
metallated ligand) or 7C (bottom, with fluoro substituent on the initial metallated
ligand) reactions.
dashed line plot to the A ! Btrans reductive elimination reaction. Even the
isomerization between the E and Z forms of the –CH]N– bond has been
determined as for other simpler systems (18,21). As a whole the full reaction
sequence can be described by the previous Scheme 10, where most of the
different steps have been resolved as a result of careful tuning of the reaction
conditions of time and temperature. Parallel time-resolved NMR measure-
ments have confirmed these assignments, and Table 2 collects the relevant
associated data. Nevertheless for some of the systems studied the isomeriza-
tion reaction has not been observed, thus indicating that the process is fast
under the conditions studied, as already observed for some of the systems
indicated in Table 1. As a whole, when the processes on the chelate
N(amino)–N(imine) and monodentate N(imine)–L systems are compared, both
appear to be extremely complex and depending on the large number of vari-
ables involved. This is not surprising given the number of reaction steps
involved in the full reactivity, which makes any claim such as faster than or
slower than meaningless. Furthermore, the characterization of the intermedi-
ate species during the full process also brings in some news facts that have to
be considered with respect to the isolation of only the less-reactive species of
the full set.
Nevertheless, the most striking feature of the reactivity indicated in the
previous pages relates to the five- or seven-membered platinacyclic nature of
the final complex obtained. As seen in Table 2 only for one of the reactions
studied, no doubt, the most thermodynamically stable five-membered plat-
inacyclic compound is obtained, and the tuning of such specificity cannot be
related to the different reaction sequences observed, all being rather equiv-
alent. Only the detailed single-step sequence occurring in the Bcis ! 7C or
Bcis ! 5C process (Scheme 13) can explain such tuning. Further DFT cal-
culations have been conducted (see Section 4) in order to establish the pref-
erences once the empirical results are fully ascertained. A similar approach
has been conducted in other organometallic systems with rather impressive
synergic outcome (31,32).
4. SYNERGY WITH DFT CALCULATIONS

DFT calculations have, nowadays, become a widely used powerful
tool for studying reaction paths; quite a few examples are available in organ-
ometallic chemistry, catalysis, and coordination chemistry studies (33–37).
The synergy between experiment and theory has also evolved considerably,
allowing the development of both fields and producing some outcomes that
would be hardly achievable using any of the single techniques. In this
respect, the use of DFT calculations began by supporting results in a post-
experimental manner, i.e., determining reaction mechanisms, and selectivity
when the rational chemical intuition and mechanistic experimental tech-
niques have been exhausted. More recently, however, theory computations
have begun to be used in parallel to experiments, or, even prior to them. By
doing so, they have led, in some cases, to the discovery of new reactions and
chemical systems (38–40).
We have employed DFT theoretical calculations to ascertain
some relevant aspects of the cycloplatination reactions explored from a
kinetico-mechanistic perspective indicated in Section 3. The aim has been
to clarify some of the relative chemical stabilities, isomerization processes,
and kinetics observed (11). Although the studies related to the PtII com-
plexes having N(amino)–N(imine) ligands have been reported, some new
Table 2 Relevant Kinetic (Xylene Solution, 340 K) and Activation Parameters of the Spontaneous
Measured on Compound of Type A (Scheme 10)
Compound A A ! Btrans Btrans ! Bcis
105 k ΔHǂ ΔSǂ ΔV ǂ 105 k ΔHǂ ΔSǂ ΔV ǂ

4-MeC6H4
Me2N
Not Not
4-MeC6H4 Pt N 240 97 4 –13 11 150 42 5 –179 15
determined determined
Br
4-MeC6H4
Me2N
4-MeC6H4 Not
Pt N 9.1 122 5 33 13 Fast
determined
Cl
4-MeC6H4
Me2N
4-MeC6H4 N Not
Pt 3.4 118 5 13 13 Fast
determined
Cl Cl
4-MeC6H4
Me2N
4-MeC6H4 Pt N 7.1 121 2 28 7 –9 2 7.1 54 4 –169 12 –24 4
F Br
4-MeC6H4
Me2N
4-MeC6H4 Not Not
Pt N 55,000 69 15 –50 45 1.8 63 5 –154 13
determined determined
F Cl
Rate-Determining Steps Involved in the Reductive Elimination/Oxidative Addition Reactions
Bcis ! 7C Bcis ! 5C
105 k ΔH ǂ ΔSǂ ΔV ǂ 105 k ΔHǂ ΔSǂ ΔV ǂ

Not
— — — — 30 86 7 –63 20
determined
Not
0.29 141 15 60 44 — — — —
determined
29 102 5 –16 15 29 — — — —
Fast — — — —
Not
2200 106 15 32 40 — — — —
determined
Scheme 13 Detailed sequential reaction steps needed for the transformation of compounds of type Bcis to cycloplatinated compounds of
type 7C or 5C (Scheme 10).
calculations related to monodentate N(imine) ligand systems have also been

carried out and are reported here for the first time. The computational
methodology used is equivalent to that utilized before (11), and frequency
calculations were carried out to confirm stationary points and transition
states. Additional single-point calculations on the optimized geometries
were also employed to obtain improved solvated free energy values with
larger basis sets at the corresponding reaction temperatures; unless other-
wise stated all the free energy values reported correspond to those obtained
with these larger sets. As for the concentration/time kinetic models, they
have been built using Copasi software (41) using the deterministic
(LSODA) method with relative and absolute tolerance values of 106
and 1012, respectively.
4.1 Five- vs Seven-Membered Platinacycle Stability

As mentioned in the previous sections, the final five-membered platinacycle
compounds, 5C, are obviously expected to be more thermodynamically sta-
ble than the alternative seven-membered, 7C, analogues. This fact has been
confirmed in a facile manner using DFT calculations. The possible 5C and
7C products have been computed for different starting PtII complexes,
A (even including some that have not been experimentally studied); the free
energies of reaction (ΔGR) thus derived are shown in Table 3. Clearly the
five-membered platinacycles (5C) are the more stable in all cases, in agree-
ment with what is expected for this kind of motif. Consequently, the
obtention of the larger seven-membered platinacycles, from the isolated Bcis
intermediates indicated earlier, has to be due to kinetic preferences (see
below). Furthermore, the calculated free energies for compounds 5C and
7C also confirm the expected larger stabilization of the N(amino)–N(imine)
chelated PtII species; in most cases, these are found 5–20 kJ mol1 more
stable than their corresponding monodentate N(imine) analogues. Entries 9
and 10 do not include the calculated value for the corresponding 5C
complex given its nonfeasibility due to the blocking indicated in
Schemes 2, 5, and 12.
4.2 cis vs trans Stability for B-Type Intermediates

As stated in the previous sections, the Btrans > Bcis interconversion plays
an important role in the reactions studied (11,18); in some cases, the process
even becomes the rate-determining step of the general cycloplatination
reaction (13). The full reaction consists of the sequence shown in
Table 3 Computed Free Energies of Reaction for Products 5C and 7C of Type A Compounds; Toluene Solution at 70°C Unless Stated
Otherwise
ΔGR ΔGR
Entry Compound A Entry Compound A
/kJ mol21 /kJ mol21
Ph Ph Ph Ph
SMe2 SMe2
Ph Pt N 5C: 57.1 Ph Pt N 5C: 58.0
1 2
7C: 39.9 7C: 35.7
Br Cl
4-MeC6H4 Ph 4-MeC6H4 Ph
SMe2 SMe2
4-MeC6H4 Pt N 5C: 54.3 4-MeC6H4 Pt N 5C: 58.5
3 4
7C: 34.8 7C: 38.2
Br Cl
4-FC6H4 Ph 4-FC6H4 Ph
SMe2 SMe2
4-FC6H4 Pt N 5C: 52.7 4-FC6H4 Pt N 5C: 52.7
5 6
7C: 40.8 7C: 34.5
Br Cl
4-MeC6H4 4-MeC6H4
Me2N Me2N
5C: 70.3 5C: 64.5
7a 4-MeC6H4 Pt N 8a 4-MeC6H4 Pt N
7C: 38.5 7C: 31.1
Br Cl
4-MeC6H4 4-MeC6H4
Me2N Me2N
5C: — 5C: —
9b 4-MeC6H4 Pt N 10b 4-MeC6H4 Pt N
7C: 44.4 7C: 44.9
F Br F Cl
a
Xylene solution, 139°C.
b
Toluene solution, 110°C.
Scheme 14 Btrans > Bcis isomerization reaction sequence.
Scheme 14. Initially, ligand L dissociates producing a T-shaped intermediate

iBtrans, in which the imine group migrates from a trans- to a cis-aryl posi-
tion, thus generating an iBcis intermediate. From this point back coordina-
tion of the L ligand (or moiety) produces the final Bcis isomer.
In practice this isomerization process has been comprehensively com-
puted only for compounds of type A with chelating N(amino)–N(imine) ligands
with Aryl ¼ 4-MeC6H4 and X ¼ Br or Cl (complexes on entries 7 and 8 in
Table 3). The procedure used for the calculation was a linear transit potential
energy surface scan; i.e., B and iB complexes were computed normally
along with a set of structures in between, with the Aryl–Pt–N(imine) angle
kept frozen at different decreasing values. The highest energy point along
the reaction coordinate was thus associated with the isomerization transition
state (Isom_TS) of the Btrans > Bcis rearrangement. Although the free
energies shown in Fig. 7 were obtained using the smaller basis sets to simplify
the scanning procedure, the relevant values for Isom_TS were recomputed
with the bigger basis sets and found to lie at 138.4 and 140.1 kJ mol1,
respectively, for complexes in entries 7 and 8 of Table 3.
As shown, in the case of the L ¼ N(amine) initial ligand release
(Scheme 14), either from Btrans or Bcis, requires a quite high amount of
energy ca. 90–110 kJ mol1, which is obviously related to the chelating sta-
bilization introduced by the bidentate ligand, lost when forming the
corresponding T-shaped iB complexes. From this point on, the free energy
rises steadily as the Aryl–Pt–N(imine) angle decreases, reaching a maximum
around 140 degree for both complexes studied. These structures, considered
to be the transition states of the isomerization reaction (Isom_TS), have an
imaginary (ca. 100i cm1) frequency that resembles a rocking normal
mode of vibration that moves the N(imine) ligand from a trans- to a cis-aryl
geometrical position. As expected from the results collected in the previous
sections, the Bcis complexes are lower in energy than their trans counterparts
(Table 4). Interestingly for the monodentate N(imine) ligands (L ¼ SMe2) this
tendency is less pronounced.
150
Free energy (kJ mol–1)
100
50
160 150 140 130 120 110

–50 Btrans iBtrans Aryl–Pt–Nimine angle (degree) iBcis Bcis
Reaction coordinate
Fig. 7 Computed free energy profile for the Btrans > Bcis isomerization of complexes
on entries 7 (solid) and 8 (dashed) in Table 3.
4.3 Formation of Five- or Seven-Membered Platinacycles From

Bcis-Type Intermediates
4.3.1 Complexes With Monodentate N(imine) Ligands
The spontaneous reaction of three type A monodentate N(imine) complex
systems (L ¼ SMe2 and X ¼ Br, i.e., entries 1, 3, and 5 from Table 3) has been
computationally explored in order to ascertain the preferential obtention of
the seven-membered ring platinacycle product in all cases (type 7C). As
indicated in Scheme 15, two reaction pathways (I and II) from the interme-
diate Bcis compounds described earlier are possible, each one leading to the
formation of the five- or the seven-membered platinacycles, 5C and 7C,
respectively.
Both reaction sequences involve a preliminary dissociation of the SMe2
ligand from the starting Bcis species, leading to the T-shaped iBcis interme-
diate. After this step a C–H bond activation, formally an oxidative addition
process, should occur, either on the proximal (HA, I) or distal (HB, II) phe-
nyl group of the biphenyl substituent of the imine ligand. Such activation
entails the formation of the square pyramidal PtIV intermediates B-CH5
or B-CH7, where the incoming hydride ligand is placed in the apical posi-
tion. Other conformers with a different ligand arrangement were also
explored, but the calculated energies were found to be higher than the ones
shown in Scheme 15. The final platinacycles of type 5C or 7C are finally
obtained by the reductive elimination and release of the corresponding aryl
Table 4 Computed Relative Free Energies for Btrans and Bcis Isomers Derived From the Type A Compounds Shown (Scheme 10); Toluene
Solution at 70°C Unless Stated Otherwise
ΔG ΔG
Entry Compound A Entry Compound A
/kJ mol21 /kJ mol21
Btrans: 0.0 Btrans: 0.0

1 2
Bcis: 0.1 Bcis: 2.3
4-MeC6H4 Ph 4-MeC6H4 Ph
SMe2 SMe2
4-MeC6H4 Pt N Btrans: 0.0 4-MeC6H4 Pt N Btrans: 0.0
3 4
Bcis: 3.1 Bcis: 5.7
Br Cl
4-FC6H4 Ph 4-FC6H4 Ph
SMe2 SMe2
4-FC6H4 Pt N Btrans: 0.0 4-FC6H4 Pt N Btrans: 0.0
5 6
Bcis: 1.3 Bcis: 6.1
Br Cl
4-MeC6H4 4-MeC6H4
Me2N Me2N
7a 4-MeC6H4 Pt N 8a 4-MeC6H4 Pt N
Bcis: 15.8 Bcis: 32.3
Br Cl
4-MeC6H4 4-MeC6H4
Me2N Me2N
9b 4-MeC6H4 Pt N 10b 4-MeC6H4 Pt N
Bcis: 15.3 Bcis: 17.7
F Br F Cl
a
Xylene solution, 139°C.
b
Toluene solution, 110°C.
Scheme 15 Reaction pathways leading to 5C (I, top) and 7C (II, bottom) type of com-
pounds from PtII species of type Bcis with monodentate N(imine) ligands.
by-products. Thus, the formation of the five- or seven-membered products

should be dominated by the relative energy requirements of the oxidative
addition (OATS-CHx) and reductive elimination (RETS-CHx) transition
states. Experimentally, only the 7C type of platinacycles is observed for the
bromido complexes on entries 1, 3, and 5 in Table 3. Clearly the transition
states along the route leading to 7C (II) should be lower than those involved
in the formation of 5C (I). Indeed, this is what is found when the relative
free energies are computed (Table 5).
As shown in Table 5, all the computed free energy profiles are quite sim-
ilar, despite the diverse identity of the R group on the aryl ring. Nevertheless,
the fact that the rate-determining step completely changes depending on the
pathway followed by the reaction is rather interesting. For pathway I, that
leading to the five-membered platinacycle product 5C, the highest energy
demanding stage corresponds to the final C–H reductive elimination produc-
ing the PhR moiety (RETS-CH5). Contrarily, for pathway II, that leading to
the seven-membered platinacycle product 7C, the highest energy value cor-
responds to the initial C–H bond activation on the distal phenyl ring of the
imine ligand (OATS-CH7). These differences can be easily rationalized
by the careful observation of the structures of the transition states involved
in each transformation. Fig. 8 shows, as an example, the four transition states
involved for the A compound on entry 1 of Table 3, including relevant bond
distances. The analogous structures for the A compounds of entries 3 and 5 of
the same table are very similar, thus the corresponding structures are not
shown. The analysis of the bond distances in the transition states shown does
Table 5 Computed Relative Free Energies (in kJ mol1 and in Toluene Solution at 70°C)
for the Reaction Leading to PtII Compounds of Type 5C and 7C From Complexes of
Type A With Monodentate N(imine) Ligands (Entries 1, 3, and 5 of Table 3)
Reaction Entry in Table 3

Intermediate
pathway 1 3 5
Bcis 0.0 0.0 0.0
—
iBcis 30.9 34.1 37.5
OATS-CH5 75.1 77.1 78.2
B-CH5 72.6 71.6 77.4
I
RETS-CH5 112.9 110.8 117.1
5C 57.1 54.3 52.7
OATS-CH7 105.1 100.8 104.7
B-CH7 72.3 72.9 72.1
II
RETS-CH7 97.0 94.6 96.2
7C 39.9 34.8 40.8
not seem to provide a clear explanation of the observed reactivity, as they

are rather similar in all cases. Obviously, the Pt–N(imine) bond is excep-
tional, being consistently longer in the route leading to product 7C.
A more thorough examination of the structures, nevertheless, provides a
rather feasible explanation for the computed relative barrier heights. The
smallest H–Pt–(Aryl) torsion angle found in the C–H oxidative addition
and reductive elimination transition states (Table 6) is an indication of the
relative orientation of the aryl ring and the hydrogen that is being cleaved
or attached. The most favorable energy occurs when this torsion angle is
close to 90 degree, indicating that the C–H activation and reductive eli-
mination reactions proceed naturally in the plane perpendicular to the aryl
ring. Whenever this angle decreases, probably hindered by the arrange-
ment of the other substituents on the platinum, the hydride group gets
closer to one of the hydrogen atoms in the aryl ring and the free energy
increases. Consequently, the torsion angles of the species shown in Table 6
indicate that the energy requirements to get to the transition state for the
studied compounds are OATS-CH5 < RETS-CH7 < OATS-CH7
<RETS-CH5, in perfect agreement with the computed free energy values
in Table 5.
Fig. 8 Computed structures for the transition states leading to products 5C (left, path-
way I) and 7C (right, pathway II) from compounds of type Bcis originated form com-
pound of type A in entry 1 of Table 3 (distances in Å, nonrelevant H atoms have
been omitted for clarity).
Table 6 Smallest H–Pt–(Aryl) Torsion Angles (in Degrees) and Computed Free Energies
(in kJ mol1) for the Transition States Indicated in Table 5 for the C–H Activation and
Reductive Elimination Reactions Studied
Entry in Table 3
1 3 5
H–Pt–Aryl ΔG H–Pt–Aryl ΔG H–Pt–Aryl ΔG

OATS-CH5 87.8 75.1 88.0 77.1 87.9 78.2
RETS-CH5 52.5 112.9 52.3 110.8 53.8 117.1
OATS-CH7 69.0 105.1 68.6 100.8 69.5 104.7
RETS-CH7 73.6 97.0 73.8 94.6 74.0 96.2
Fig. 9 Qualitative time-resolved (arbitrary time scale) concentration profile for the com-
plexes with monodentate N(imine) and SMe2 ligands involved in entries 1, 3, and 5 of
Table 3 (solid line, compounds of type A; dashed line, compounds of type 7C; dotted line,
compounds of type 5C).
The computed relative free energy values collected in Table 5 have also
been used to build a qualitative kinetic profile simulation model to evaluate
the alternative formation of the final five- and seven-membered
platinacycles, 5C and 7C, starting from the corresponding Bcis parent inter-
mediates of entries 1, 3, and 5 in Table 3 (Fig. 9). The computed results
totally agree with the experimental observations and confirm that only prod-
ucts of type 7C are expected in the reaction from bromido complexes of
type A containing a set of monodentate N(imine) and a SMe2 ligands. The
possible amount of the analogous five-membered ring product 5C remains,
in all cases lower than 7%, and the 5C:7C ratios are 1:14, 1:30, and 1:73 for
complexes involved in entries 1, 3, and 5 of Table 3, respectively.
4.3.2 Complexes With Chelating N(amino)–N(imine) Ligands

The reactions of four different PtII complexes bearing chelate N(amino)–
N(imine) ligands, i.e., entries 7–10 in Table 3, have also been computationally
studied in the same manner as that indicated earlier for the systems con-
taining the monodentate N(imine) ligands. The experimental reactivity of
the compounds of type A shown in entries 7 and 8 of this table strongly
depends on the nature of the halido ligand within the starting material, either
bromido or chlorido. Results show a striking difference in the final products,
which correspond to a five-membered platinacycle compound, type 5C, for
X ¼ Br, but to a seven-membered analogue, type 7C, for X ¼ Cl. For the
reactions occurring on the compounds of type A of the last two entries
(9 and 10), only compounds of type 7C are produced. Obviously, this is
a result of the hydrogen by fluoride substitution on the proximal aryl ring
of the N(amino)–N(imine) ligand; nevertheless, their reactivity was also theo-

retically calculated for comparison with the experimental data available.
As stated earlier for the monodentate systems, two possible pathways
(leading to 5C or 7C) are possible for the systems on entries 7 and 8 of
Table 3, once species of type Bcis isomer are attained (Scheme 16). For
the compound on entries 9 and 10 only route II leading to platinacycles
of type 7C is possible. The reaction sequence for these mechanisms is quite
similar to that described in the previous pages for the monodentate systems.
Nevertheless, in this case the dissociation stage producing the T-shaped
intermediate ( Bcis> iBcis) involves only the dechelation of a ligand, as
the dimethylamino group still remains associated to the T-shaped species
in a dangling arrangement. As for the previous systems, the relative energy
requirements of the oxidative addition (OATS-CHx) and reductive elim-
ination (RETS-CHx) transition states govern the selective formation of the
final 5C or 7C product types.
For the bromido complex on entry 7 of Table 3, the energy requirements
for pathway I, producing the thermodynamically preferred five-membered
platinacycle 5C, should be lower in energy than those leading to 7C. Con-
versely, for the analogous chlorido system (entry 8 of Table 3), the energy
requirements should favor pathway II, being the seven-membered
platinacycle of type 7C the one generated experimentally. Indeed, the com-
puted relative free energies for these two systems confirm these trends, as
shown in Table 7.
Scheme 16 Reaction pathways leading to 5C (I, top) and 7C (II, bottom) type of com-
pounds from PtII species of type Bcis with chelate N(amino)–N(imine) ligands.
Table 7 Computed Relative Free Energies (in kJ mol1 and in Xylene Solution at 139°C)
for the Reaction Leading to PtII Compounds of Type 5C and 7C From Complexes of
Type A With Chelate N(amino)–N(imine) Ligands (Entries 7 and 8 of Table 3)
Entry in Table 3
Reaction pathway Intermediate
7 8
Bcis 0.0 0.0
—
iBcis 90.1 97.9
OATS-CH5 119.6 128.4
B-CH5 115.9 127.9
I
RETS-CH5 141.5 155.6
5C 70.3 64.5
OATS-CH7 146.4 149.8
B-CH7 126.4 136.0
II
RETS-CH7 145.1 149.1
7C 38.5 31.1
It may be noted that the dissociation of the chelating dimethylamino

group requires a rather high input of energy (90.1 and 97.9 kJ mol1 for sys-
tems in entries 7 and 8 of Table 3, respectively), which is in contrast to the
much lower requirements (typically less than 40 kJ mol1) for the complexes
having a monodentate N(imine) ligand. This difference can be easily explained
by the entropic contributions of the ligand dissociation stage. On one hand,
for the systems having monodentate N(imine) ligands, species Bcis dissociates
into two different fragments, i.e., iBcis and SMe2, thus favoring entropy
requirements. On the other hand, for the systems with chelating N(amino)–
N(imine) ligands, such dissociation produces a T-shaped intermediate
with a dangling amino group. This group still remains within the iBcis
coordination sphere, thus preventing an entropic stabilization to occur.
The computed free energies shown in Table 7 fully agree with the reac-
tivity observed with respect to the nature and size of the final platinacycles.
For the bromido complex in entry 7 of Table 3, the highest energy demands
for pathways I and II are predicted to be 141.5 (RETS-CH5) and 146.4
(OATS-CH7) kJ mol1, respectively; this fact indicates that the preferred
product should be the five-membered platinacycle of type 5C. This behav-
ior is totally contrary to that for the equivalent chlorido compound (entry
8 of Table 3), for which the highest barriers of pathways I and II are
155.6 (RETS-CH5) and 149.8 (OATS-CH7) kJ mol1, respectively. In

this case the formation of the seven-membered ring, type 7C, should be
more favorable, as experimentally observed. It should be indicated that, both
for X ¼ Br and Cl (entries 7 and 8 of Table 3), the highest transition state for
each pathway corresponds to the same process. For pathway I the rate-
determining step corresponds to the final C–H reductive elimination of
the aryl by-product, while for pathway II it is the C–H initial bond activa-
tion that determines the energy requirements. Nevertheless, for the latter the
final reductive elimination is almost as high in energy requirements.
The analysis of the bond distances of the transition state structures shows
a similar behavior to the one indicated in the previous pages for systems with
monodentate N(imine) and SMe2 ligands. Nevertheless, for the H–Pt–(Aryl)
torsion angles the trend is not as clear as before. Table 8 collects the
smallest H–Pt–(Aryl) torsion angles found in the C–H oxidative addition
and reductive elimination transition states for the computed reactivity of
the complexes of type A in entries 7 and 8 of Table 3. Nevertheless, as in
Section 4.3.1, a correlation exists between the torsion angle values and the
energy demands of the corresponding transition state (Table 7). Angles closer
to 90 degree are associated with the lower energy transition states, while
smaller angles relate to an increase of the energy requirements. However,
for the system of entry 7 in the previous tables, the reductive elimination tran-
sition state RETS-CH5 (on the way to a type 5C species) appears to be quite
low in energy when associated with the torsion angle value (Table 8).
Table 8 Smallest H–Pt–(Aryl) Torsion Angles (in Degrees) and Computed Free Energies
(in kJ mol1) for the Transition States Indicated in Table 7 for the C–H Activation and
Reductive Elimination Reactions Studied
Entry in Table 3
7 8
H–Pt–Aryl ΔG H–Pt–Aryl ΔG
OATS-CH5 89.4 119.6 89.4 128.4
RETS-CH5 58.5 141.5 53.2 155.6
OATS-CH7 66.1 146.4 66.7 149.8
RETS-CH7 70.8 145.1 71.1 149.1
Even so, a careful analysis reveals that the relatively small torsion angle is in
fact larger than those found for the analogous transition states of the reac-
tivity of the systems of entries 1, 3, 5, and 8 from Table 3 (RETS-CH5
H–Pt–(Aryl) torsion angles being 52.5, 52.3, 53.8, and 53.2 degree,
respectively). This, somehow, low energetic requirement could be directly
related to the preferential formation of compound of type 5C in this case,
thus explaining the switching in reactivity from the obtention of type 7C
complexes.
The experimental observation that the spontaneous reaction of type A
complex in entry 7 on Table 3 is much faster than that of the system of entry
8 can be immediately associated with the calculated energy demands above
(Table 8) (11). The computed free energies for the reactions, despite differ-
ences in reaction conditions and calculations, indicate that for the bromido
complex (entry 7) the energy requirement is 8 kJ mol1 less than for the chlo-
ride analogue (entry 8). Furthermore, for the systems from entries 7 and 8 of
Table 3, the reactivity shown in Scheme 16 has been found to take place after a
kinetically and spectroscopically detected Btrans > Bcis rearrangement (see
Section 3). Consequently, the highest barrier found in the preferred reaction
pathway for each compound in Scheme 16 has to be lower than that for the
isomerization reaction. Indeed, this is observed for the computed data for both
systems; even for some other systems the isomerization reaction has been
found to be rate-determining (13). For the bromido compound (entry 7 of
Table 3) the isomerization (Isom_TS) and reductive elimination (RETS-
CH5) barriers are 138.4 and 141.5 kJ mol1, respectively; for the chlorido
compound of entry 8 the isomerization transition state (Isom_TS) and the
C–H oxidative addition transition state (OATS-CH7) have values of 140.1
and 149.8 kJ mol1, respectively.
As a whole, the data collected in Table 7 allow the building of a quali-
tative kinetic simulation model for evaluating the formation, over time of
the final five- and seven-membered platinacycle compounds of types 5C
and 7C from the starting Bcis complexes of entries 7 and 8 in Table 3
(Fig. 10). A very satisfactory qualitative agreement is found, even though
the minor product is present in significant amounts in both complexes:
for the bromido compound in entry 7 the 5C:7C ratio is 87:13, while for
the chlorido complex in entry 8 the ratio diminishes to 73:24. These results
indicate that, unfortunately, the computed free energy differences between
pathways I and II are not as important as they should be from the experi-
mentally collected results. Nevertheless, the overall trend and selectivity
are successfully explained.
Fig. 10 Qualitative product concentration evolution over time for PtII complexes with
chelate N(amino)–N(imine) ligands (entries 7 and 8 in Table 3) in arbitrary time scale (solid
line, compounds of type A; dashed line, compounds of type 7C; dotted line, compounds
of type 5C).
Table 9 Computed Relative Free Energies (in kJ mol1 and in Toluene Solution at
110°C) for the Reaction Leading to PtII Compounds of Type 7C From Complexes of
Type A With Chelate N(amino)–N(imine) Ligands (Entries 9 and 10 of Table 3)
Entry in Table 3
Intermediate
9 10
Bcis 0.0 0.0
iBcis 89.3 90.8
OATS-CH7 138.1 137.1
B-CH7 121.5 119.5
RETS-CH7 137.5 136.4
7C 44.4 44.9
In order to validate the employed methodology, the mechanism indi-

cated in Scheme 16 has been also computed for the reaction of the fluori-
nated compounds on entries 9 and 10 in Table 3. As stated earlier, for these
systems the formation of the five-membered platinacycles of type 5C is not
possible, thus only pathway II has been computed (Table 9).
The calculated energy requirements for the process producing the com-
pounds of type 7C, although very similar for both complexes, are slightly
lower (ca. 1 kJ mol1) for the system containing the chlorido ligand
Fig. 11 Qualitative product concentration evolution over time for PtII complexes with
chelate N(amino)–N(imine) ligands (entries 9 and 10 in Table 3) in arbitrary time scale.
(entry 10 in Table 3). The experimental data collected, as indicated in

Section 3, show that the rate of formation of the chlorido compound of type
7C is ca. four times faster than that of the analogous bromido complex (entry
9 in Table 3) (11). The qualitative kinetic model built for these processes
totally agrees with this observation. Fig. 11 shows the concentration-time
profile for the formation of seven-membered platinacycles of type 7C from
complexes of type A in entries 9 and 10 of Table 3; as predicted, the product
ratio obtained is 1:4 in favor for the chlorido species.
5. CONCLUDING REMARKS
In this contribution, we have illustrated how a series of oxidative addi-
tion and reductive elimination reactions, occurring on platinum(II) organ-
ometallic complexes, can be resolved experimentally by the use of an
appropriate interlock between preparative and kinetic methodologies.
Namely, the reaction of bis(aryl)platinum(II) precursors with imine or
amino-imine ligands leads to the formation of platinacyclic compounds with
five- or seven-membered metallacycles via a sequence that involves C–X
bond oxidative addition plus reductive C–C coupling followed by C–H oxi-
dative addition with a final C–H reductive elimination. Given the fact that
the process includes a series of non-simple and fairly slow processes, several
intermediates have been proposed and fully characterized in the reaction
media. Nevertheless, with the precision of the approach some of these char-
acterized species are found to be completely irrelevant in the sequential
reaction pathway; that is, they are de facto false intermediates or dead-end
compounds.
The kinetico-mechanistic studies carried out on this time-resolved
reactivity has provided series of thermal- and pressure-derived activation
parameters that has allowed a fairly comprehensive description of the reac-
tion sequence involved in the full process. Nevertheless, a careful and com-
prehensive analysis has had to be applied, since the rate sequence is not
uniformly applicable for the full series of complexes. That is, the combina-
tion of two reactions having only slightly different activation energies pro-
duces dramatic changes when taken jointly, and the rate-determining step of
the full process could differ within a series of very similar reactions.
In this respect, the synergy between experiments and DFT calculations
has been found to be very important in order to improve the understanding
of both the mechanisms involved and the observed reactivity. DFT calcu-
lations have not only ascertained the elementary reaction steps of the process,
they have also explained the kinetic preference for the less thermodynam-
ically stable products found in the majority of cases. Furthermore, the pres-
ence and incidence in the overall process of dead-end false intermediate
compounds can be also assessed through calculations.
It is thus clear that computational studies, when fully supported by the
empirical results, are able to fill in the areas of the puzzle that cannot be
experimentally explored and resolved. In this respect, the set of studies pres-
ented in this contribution constitute a perfect example of how a close col-
laboration between experimentalists and theoreticians, with their specific
perspectives when observing and interpreting data, provides remarkable
value-added results to the work carried out.
ACKNOWLEDGMENTS
We thank the Spanish national research agency for its economic support. In particular project
Grant numbers CTQ2009-11501, CTQ2012-37821-C2-01, CTQ2015-65707-C2-1-P,
CTQ2015-65040-P, and CTQ2015-64579-C3-1-P from the Ministerio de Economı́a y
Competitividad/FEDER are acknowledged. Allocation of computer resources at IQTCUB
is also acknowledged.
REFERENCES
1. Anderson, C. M.; Puddephatt, R. J.; Ferguson, G.; Lough, A. J. J. Chem. Soc. Chem.
Commun. 1989, 1297–1298.
2. Anderson, C. M.; Crespo, M.; Jennings, M. C.; Lough, A. J.; Ferguson, G.;
Puddephatt, R. J. Organometallics 1991, 10, 2672–2679.
3. Anderson, C. M.; Crespo, M.; Ferguson, G.; Lough, A. J.; Puddephatt, R. J.

Organometallics 1992, 11, 1177–1181.
4. Crespo, M.; Martı́nez, M.; Sales, J.; Solans, X.; Font-Bardı́a, M. Organometallics 1992,
11, 1288–1295.
5. Crespo, M.; Martı́nez, M.; Sales, J. J. Chem. Soc. Chem. Commun. 1992, 822–823.
6. Crespo, M.; Martı́nez, M.; Sales, J. Organometallics 1993, 12, 4297–4304.
7. Font-Bardia, M.; Gallego, C.; Martı́nez, M.; Solans, X. Organometallics 2002, 21,
3305–3307.
8. Gallego, C.; Martı́nez, M.; Safont, V. S. Organometallics 2007, 26, 527–537.
9. Calvet, T.; Crespo, M.; Font-Bardı́a, M.; Jansat, S.; Martı́nez, M. Organometallics 2012,
31, 4367–4373.
10. Crespo, M.; Font-Bardia, M.; Solans, X. Organometallics 2004, 23, 1708–1713.
11. Aullón, G.; Crespo, M.; Font-Bardia, M.; Jover, J.; Martı́nez, M.; Pike, J. Dalton Trans.
2015, 44, 17968–17969.
12. Martı́n, R.; Crespo, M.; Font-Bardia, M.; Calvet, T. Organometallics 2009, 28, 587–597.
13. Crespo, M.; Font-Bardia, M.; Martı́nez, M. Dalton Trans. 2015, 44, 19543–19552.
14. Escolà, A.; Crespo, M.; Quirante, J.; Cortes, C.; Jayaraman, A.; Badı́a, J.; Baldomà, L.;
Calvet, T.; Font-Bardı́a, M.; Cascante, M. Organometallics 2014, 33, 1740–1750.
15. Calvet, T.; Crespo, M.; Font-Bardı́a, M.; Gómez, K.; González, G.; Martı́nez, M.
16. Crespo, M.; Calvet, T.; Font-Bardia, M. Dalton Trans. 2010, 39, 6936–6938.
17. Crespo, M.; Font-Bardia, M.; Calvet, T. Dalton Trans. 2011, 40, 9431–9438.
18. Bernhardt, P. V.; Calvet, T.; Crespo, M.; Font-Bardı́a, M.; Jansat, S.; Martı́nez, M. Inorg.
Chem. 2013, 52, 474–484.
19. Cortes, R.; Crespo, M.; Davin, L.; Martı́n, R.; Quirante, J.; Ruiz, D.; Messeguer, R.;
Calvis, C.; Baldomà, L.; Badia, J.; Font-Bardı́a, M.; Calvet, T.; Cascante, M. Eur. J.
Inorg. Chem. 2012, 54, 557–566.
20. Crespo, M.; Martı́nez, M.; de Pablo, E. J. Chem. Soc. Dalton Trans. 1997, 1231–1235.
21. Crespo, M.; Font-Bardı́a, M.; Granell, J.; Martı́nez, M.; Solans, X. Dalton Trans. 2003,
3763–3769.
22. Crespo, M.; Martı́nez, M.; Nabavizadeh, S. M.; Rashidi, M. Coord. Chem. Rev. 2014,
279, 115–140.
23. Tobe, M. L.; Burgess, J. Inorganic Reaction Mechanisms; Adison Wesley Longman:
Harlow, 1999.
24. Wilkins, R. G. Kinetics and Mechanisms of Reactions of Transition Metal Complexes; VCH:
Weinheim, 1991.
25. Perez-Benito, J. F.; Mulero-Raichs, M. J. Phys. Chem. A 2016, 120, 7598–7609.
26. Bernhardt, P. V.; Gallego, C.; Martı́nez, M.; Parella, T. Inorg. Chem. 2002, 41, 1747–1754.
27. Font-Bardı́a, M.; Gallego, C.; González, G.; Martı́nez, M.; Merbach, A. E.; Solans, X.
Dalton Trans. 2003, 1106–1113.
28. Esteban, J.; Font-Bardia, M.; Gallego, C.; González, G.; Martı́nez, M.; Solans, X. Inorg.
Chim. Acta 2003, 351, 269–277.
29. Bernhardt, P. V.; Gallego, C.; Martı́nez, M. Organometallics 2000, 19, 4862–4869.
30. Gallego, C.; González, G.; Martı́nez, M.; Merbach, A. E. Organometallics 2004, 23,
2434–2438.
31. Aullón, G.; Chat, R.; Favier, I.; Font-Bardı́a, M.; Gómez, M.; Granell, J.; Martı́nez, M.;
Solans, X. Dalton Trans. 2009, 8292–8300.
32. Roiban, G. D.; Serrano, E.; Soler, T.; Aullón, G.; Grosu, I.; Cativiela, C.; Martı́nez, M.;
Urriolabeitia, E. P. Inorg. Chem. 2011, 50, 8132–8143.
33. Jover, J.; Fey, N. Chem. Asian J. 2014, 9, 1714–1723.
34. Sperger, T.; Sanhueza, I. A.; Kalvet, I.; Schoenebeck, F. Chem. Rev. 2015, 115,
9532–9586.
35. Santoro, S.; Kalek, M.; Huang, G.; Himo, F. Acc. Chem. Res. 2016, 49, 1006–1018.
36. Sperger, T.; Sanhueza, I. A.; Schoenebeck, F. Acc. Chem. Res. 2016, 49, 1311–1319.
37. Zhang, X.; Chung, L. W.; Wu, Y.-D. Acc. Chem. Res. 2016, 49, 1302–1310.
38. Jover, J.; Miloserdov, F. M.; Benet-Buchholz, J.; Grushin, V. V.; Maseras, F.
39. Durr, A. B.; Yin, G.; Kalvet, I.; Napoly, F.; Schoenebeck, F. Chem. Sci. 2016, 7,
1076–1081.
40. Sperger, T.; Fisher, H. C.; Schoenebeck, F. WIREs Comput. Mol. Sci. 2016, 6, 226–242.
41. Hoops, S.; Sahle, S.; Gauges, R.; Lee, C.; Pahle, J.; Simus, N.; Singhal, M.; Xu, L.;
Mendes, P.; Kummer, U. Bioinformatics 2006, 22, 3067–3074.
CHAPTER SIX
Controlling the Lability of Square-

Planar Pt(II) Complexes Through
Electronic and π-Conjugation:
Correlation Between Kinetics and
Theoretical Parameters
Allen Mambanda, Deogratius Jaganyi1
School of Chemistry and Physics, University of KwaZulu-Natal, Scottsville, Pietermaritzburg, South Africa
1
Corresponding author: e-mail address: jaganyi@ukzn.ac.za
Contents
1. Introduction 244
1.1 π-Acceptor vs σ-Donor Effect of Nitrogen/Carbon N^C/N^N/C Tridentate
Ligands 246
2. Pull-and-Push Effects of Pt(II) Complexes With Coordinated N^C/N^N/C
π-Acceptor Nonleaving Ligands 252
2.1 Controlling the π-Acceptor Properties of N^C/N^N/C Chelates via a Donor
Atom Effect 252
2.2 Controlling the π-Acceptor Properties of Terpy via Extended Ring
Conjugation 259
2.3 Controlling the π-Acceptor Properties of Terpy and Related Chelates
Through a Substituent Effect 261
2.4 Controlling the π-Acceptor Properties of Terpy Through an Ancillary
Substituent Effect 262
2.5 Incorporating Bis(Pyrazolyl)Pyridine/Benzene; Bis(70 -Azaindolyl)Pyridine/
Benzene; 1,3-Bis(80 -Quinolyl)Pyridine/Benzene; or Bis(Iminopyridyl/
Isoquinolyl)Isoindoline as Nonleaving Ligands of Square-Planar Pt(II)
Complexes 264
2.6 Controlling the π-Acceptor Properties of Other N-donor Chelates 269
3. Conclusions 272
Acknowledgments 274
References 274
Abstract
In this chapter, we report on some of the highlights of the research work which has
been carried out in our laboratories on the role and influence of nitrogen-/carbon-
donor tridentate ligands (N^C/N^N/C) on the rate of substitution of labile ligands from

244 Allen Mambanda and Deogratius Jaganyi
model square-planar Pt(II) complexes. This work was founded on the hypothesis that
tailored kinetic control of substitution reactions through changing the electronic prop-
erties of the spectator ligand backbone of Pt(II) complexes can be an insightful avenue
toward the optimization of the efficacy of future antitumor Pt(II) drugs. Focusing on the
nitrogen-/carbon-donor tridentates (N^C/N^N/C), several Pt(II) complexes were synthe-
sized and used to study systematically the mechanisms and rates of substituting the
labile coligand from the complexes using biorelevant nucleophiles. The trends in reac-
tivity as the nonleaving ligands in the complexes were systematically changed were
used to understand the electronic and steric roles of the coordinated tridentate/
bidentates on the lability of their complexes. Our kinetic data source can be used for
the tailor-designing of future metal-based drugs with rigid nonleaving ligands, given
that the lability of such complexes in biological systems plays a role on their ultimate
in vivo pharmacokinetics (i.e., toxicity, deactivation and development of resistance, and
cytotoxicity).
ABBREVIATIONS
A associative mechanism
aaa, app… ppp coordinated nonleaving nitrogen tridentate ligands with variable number
of amine groups and pyridyl rings
aaa0 coordinated bidentate and monodentate amine ligands (aa ¼ en ¼ ethylenediamine;
a0 ¼ amine)
bpma bis(2-pyridylmethyl)amine
D dissociative mechanism
DFT density functional theory
dien diethylenetriamine
HOMO highest occupied molecular orbital
LUMO lowest unoccupied molecular orbital
(N^C/N^N/C) nonleaving tridentate ligands with nitrogen- and/or carbon-donor atoms
NBO natural nonbonding orbitals
py pyridine
terpy 20 200 :60 200 -terpyridine
tmtu 1,1,3,3-tetramethyl-2-thiourea
tu thiourea
1. INTRODUCTION
The kinetic control of the lability of low-spin d8 square-planar com-
plexes in biological systems and their ability to intercalate the planar spec-
tator backbone between the base pairs of DNA are some of the key
aspects determining the efficacy (cytotoxicity) of antitumor Pt(II)-based
drugs (1–3). An extremely inert complex will have slow rates of complex
formation with DNA nucleobases, while a very reactive complex will be
Lability of Square-Planar Pt(II) Complexes 245
deactivated easily by nontarget sulfur- and nitrogen-donor nucleophiles

before reaching the nucleus (4–8). Apart from controlling the cytotoxicity,
the lability of Pt(II) complexes and of cisplatin in particular is thought to
cause detrimental effects such as acute toxicity as well as eventual develop-
ment of resistance to the drug by tumor cells (9,10). Steric and/or electronic
properties of the nonleaving ligand are the critical factors controlling the
lability of the leaving group(s) in square-planar Pt(II) complexes (5,6). These
properties impart a pull-and-push effect on the reactivity of the complexes
toward substitution. This can be useful in the designing of future metal-
based complexes with predictable reactivity and thus of potentially better
antitumor efficacy. In the absence of significant steric hindrance to the
approach from above of the incoming ligand around the square-plane,
the pull-and-push effect on the reactivity of square-planar complexes is lim-
ited to the trans-effect and/or trans-influence of the spectator ligand (5). The
labilization of a coleaving group by a trans-coordinated ligand can be effected
through accessible σ-/π-type molecular orbitals, and hence it can be further
subdivided into the trans π-effect or trans σ-effect (5). For each subcategory,
the rate of substitution of the leaving group is dependent on the strength of
the σ-donor or π-acceptor ability of the spectator ligand (11). The
labilization can be as large as five orders of magnitude (12).
A mercurial structural property of the N^C/N^N/C tridentate
nonleaving ligand which induces high lability of the coligand of the complex
has been its π-acceptor groups (13–17). Acceleration of the rate of substitu-
tion by a strong π-acceptor ligand (18–21) is ascribed to the reduction of
electron density on the Pt(II) ion due to the back donation of electrons from
the filled d-orbitals on the metal ion into the energetically low-lying (16,18)
and empty antibonding (π*) or nonbonding molecular orbitals of the ligand.
Effectively, the π-acceptor ligand withdraws electron density from the
metal-centered orbitals, making it more electrophilic for a facile nucleo-
philic attack. This also stabilizes the electron-rich pentacoordinate transition
state. As the effective charge on the metal center increases due to the
π-acceptor effect of the N^C/N^N/C tridentate spectator ligand, the nucle-
ophilic substitution of the coligand from the square-planar complex by an
incoming nucleophile is facilitated. Computational data of Pt(II)/Pd(II)
complexes have been used to demonstrate the existence of the π-acceptor
effect in square-planar Pt(II) complexes (6). In such complexes, back dona-
tion of electron density from the metals orbitals into the empty and low-
lying energy π* orbitals of the spectator ligands increases the positive charge
on the metal or its electrophilicity which is kinetically equivalent to reducing
the activation barrier to the formation of the activated state of the d-orbitals
of the Pt(II), leading to facile substitution of the leaving group from the
complex.
1.1 π-Acceptor vs σ-Donor Effect of Nitrogen/Carbon

N^C/N^N/C Tridentate Ligands
Comprehensive studies on the trans-effect of nitrogen-donor tridentates
around the Pt(II) center first came from the studies by Jaganyi in the labo-
ratory of Professor van Eldik (14,15). Using the following cationic com-
plexes, [Pt(diethylenetriamine)H2O]2+, Pt(aaa); [Pt(bis(2-pyridylmethyl)
amine)OH2](ClO4)2, Pt(aap); [Pt{2,6-bis(aminomethyl)pyridine}H2O]2+,
Pt(apa); [Pt{bis(2-pyridylmethyl)amine}H2O]2+, Pt(pap); [Pt(bipyridine)
(NH3)(H2O)]2+, Pt(app); and [Pt(terpy)H2O]2+, terpy ¼20 200 :60 200 -
terpyridine, Pt(ppp) (refer to Fig. 1 for their structures), the aqua substitution
from the complexes by neutral and anionic nucleophiles (tu, tmtu, I) was
studied under pseudo-first-order conditions using the stopped-flow technique.
In these complexes, the coordinated amines of the reference chelate ligand,
aaa, were replaced by a variable number of pyridyl rings (denoted p).
H2N NH
H3N Pt NH2 NH2 NH2
Pt
OH2 OH2
Pt(aaa⬘) Pt(aaa)
NH N
N N
NH2 Pt N
N N N
NH2 Pt NH2 H3N Pt Pt
OH2
Pt(pap) OH2 OH2 OH2
Pt(apa) Pt(app) Pt(ppp)
NH
NH
N Pt N
N Pd N
OH2
OH2
Pt(pap) Pd(pap)
Fig. 1 Structures of cationic square-planar Pt(II) complexes used to study the effects of
chelation as well as replacing amine donors with π-accepting pyridines in tridentate
nitrogen-donor chelates (15).
The explicit role of π-acceptor groups on the reactivity of Pt(II) complexes was
reported for the first time in a quantitative manner.
The kinetic data obtained for the aqua substitution by thiourea and
iodide nucleophiles have been collated into Table 1.
As shown in Table 1, the value of k2 for the nucleophilic substitution of
the aqua leaving group from the metal complexes by tu, tmtu, and I
increased by factors of about 5.6 103, 4.8 104, and 3.3 104, respec-
tively, for changing the coordination structure of the complexes from
Pt(aaa) to the Pt(ppp) (Table 1, rate constants, entry 8 vs 2–5;7). The lability
factors calculated for Pt(ppp) were much larger than could be predicted
empirically from the combined π-effects on the rate of substitution due
to systematic structural changes introduced onto the nonleaving ligands of
its analogues having either 1 or 2 isolated pyridyl rings in their chelate ligands
(14,15). The high intrinsic reactivity of Pt(ppp) relative to Pt(aaa) arose
from the strong π-acceptor effect of the terpy ligand due to its conjugated
structure (12). The delocalized and energetically low-lying π*-molecular
orbitals of the three pyridyl rings of terpy integrate electronic communica-
tion within the entire ligand (14,15). As a consequence, the electrophilicity
of Pt(ppp) was the highest compared with its analogues. During the sub-
stitution process, the strong π-acceptor effect of terpy enhances π*-
acceptance of electron density from the metal d-orbitals. This stabilizes
the formation of a bond between the metal and the incoming nucleophile
in the transition state, leading to facile substitution of the leaving group. By
ranking the reactivity of the ternary Pt(II) complexes between analogues
with either 1 or 2 pyridyl rings at different positions, it was established that
the π-effect due to the tridentate ligand was strongest when the π*-acceptor
pyridyl ligand is coordinated trans to the leaving group (19).
Interestingly, the strength of the π-effect of chelate N-donors also
depends on the softness and hence the diffusivity of the orbitals of the donor
metal ion involved in the back donation of electron density into the chelate
ligand (5). A comparison of the aqua lability of Pt(pap), [Pt(bpma)H2O]2 +,
and Pd(pap) (23), [Pd(bpma)H2O]2 +, revealed that the latter pair is only
103 times more reactive than the former (Table 1, entry 6 vs 5). This is signif-
icantly lower than the average reactivity factor of 104–105 reported (12) for the
lability of Pd(II)/Pt(II) couples bearing amine chelate ligands on the basis of
the hard acid–base theory (7). Owing to the large diffusivity of the 5d orbitals
of Pt compared to the smaller 4d orbitals of Pd, the back donation of electron
density is stronger for Pt(pap) than it is for Pd(pap). This increases the
Table 1 Second-Order Rate Constants, k2 (Std. Error), M1 s1 at 25°C and Activation Data (Within Temperature Range: 288–308 K) for Aqua
Substitution From Cationic Amine/Pyridyl Pt(II) Complexes
Kinetic Data
2
tu tmtu I
k2 ΔH≠ ΔS≠ k2 ΔH≠ ΔS≠ k2 ΔH≠ ΔS≠
Complex pKa (M21 s21) (kJ mol ) (J mol21 K21) (M21 s21)
21
(kJ mol ) (J mol21 K21) (M21 s21)
21
(kJ mol ) (J mol21 K21) Ref.
21
Pt(aaa0 ) 8(0.1) 51(1) 56(3) (22)

Pt(aaa) 6.26 29(0.4) 44(1) 69(4) 3.2 63(1) 11(3) 6.73 54(1) 30(2) (14,15)
(0.04) (0.05)
10
28(0.1) 42(1) 77(2) (22)
Pt(aap) 5.71 1.10 42(1) 65(3) 2.93 44(1) 68(3) 2.46 50(1) 33(2) (14,15)
(0.01) (0.02) (0.01)
102 10 102
Pt(apa) 6.04 1.00 42(1) 65(2) 1.23 48(1) 62(3) 1.80 52(0.3) 29(1) (14,15)
(0.01) (0.01) (0.01)
102 10 102
Pt(pap) 5.53 3.93 40(0.4) 61(1) 1.82 43(2) 57(6) 8.54 46(1) 41(2) (14,15)
(0.02) (0.02) (0.06)
102 102 102
4.05 39(0.6) 47(3) (23)
(0.03)
102
Pd(pap) 6.67 2.12 (23)
(0.05)
105
Pt(app) 6.37 1.43 36(0.5) 63(2) 1.20 41(0.6) 48(2) 1.7 50(3) 14(0.6) (14,15)
(0.01) (0.03) (0.2)
103 103 103
Pt(ppp) 4.62 1.63 21(0.3) 73(1) 1.53 19(0.4) 82(1) 2.2 30(3) 47(4) (14,15)
(0.2) (0.3) (0.4)
105 104 104
k2(Pt(ppp))/ 5.60 103 4.78 104 3.33 103
k2(Pt(aaa))
reactivity significantly for Pt(pap) more than it does for Pd(pap), thereby
reducing the reactivity advantage of the Pt complex over its Pd analogue
by only three orders of magnitude.
The reactivity of the aqua Pt(II) complexes (14,15) also showed a pos-
itive trend with respect to the acidity of the coordinated aqua ligands (24). As
shown in Fig. 2, the Brønsted acidity of the aqua coligand of the Pt(II) com-
plexes increases with increasing number of π-acceptor rings in the
nonleaving ligand and is mirrored by an increase in reactivity of the Pt(II)
complexes. The acidity of the coordinated aqua ligand increases linearly
with increasing number of π-acceptor nonleaving ligands (24) as reflected
by a decrease in the spectrophotometrically determined pKa values of the
aqua complexes (14,15). Apart from influencing intrinsic reactivity, the
π-acceptor strength of the ligand has a profound effect upon the hydrolysis
equilibria of the complexes. This becomes an important pharmacokinetic
aspect of the Pt(II) complexes especially in the intracellular environment
of the cell where the aquation of the complex is triggered by a low concen-
tration of chloride ions (25,26).
Whether the strength of the π-effect of N-donor ligands of Pt(II)
complexes is related to the denticity or chelating property of the spectator
ligand has been explored. To determine whether denticity of a nonleaving
ligand has an intrinsic capacity to enhance the rate of substitution, two Pt(II)
amine complexes, viz., [Pt(en)(NH3)(OH2)](ClO4)2, Pt(aaa0 ) {aa ¼ en ¼
ethylenediamine; a0 ¼ amine}, and [Pt(dien)(OH2)](ClO4)2, Pt(aaa)
{aaa ¼ dien ¼ diethylenetriamine}, were synthesized and the lability of
3.0
2.5 a = Amine app

p = Pyridine
2.0
N pap
or
log(k1/k10)
1.5
N
aap
N or N Pt2+ N or N 1.0
apa
π-Acceptor ability
OH2 0.5
aaa pKa value
+ Nu, –H2O k1 0.0
Increase in k1
–0.5
0.0 0.5 1.0 1.5 2.0
s = log(ka/ka0)
Fig. 2 Correlation between the π-acceptor properties of tridentate nitrogen-donor

ligands (depicted as the Lewis acidity of the co-aqua ligand) and the reactivity of their
complexes (15).
the aqua ligand measured (22). Substitution of the aqua ligand from the
complexes by a series of neutral and anionic nucleophiles was carried
out under pseudo-first-order conditions as a function of concentration
and temperature. An increase in the extent of chelation from the bidentate
complex, Pt(aaa0 ), to the tridentate complex, Pt(aaa), increased the rate
of aqua substitution from the complexes by an average factor of 3 (Table 1,
rate constants, entry 2 vs 1). In addition, computational analysis of the com-
plexes indicated a lengthened Pt-OH2 bond for Pt(aaa) compared to the
Pt(aaa0 ). It was concluded that the thermodynamic stability associated
with chelation promoted ground-state destabilization of the Pt-OH2 bond
of the Pt(aaa) relative to that of Pt(aaa0 ), leading to increased labilization
of the leaving group.
The activation parameters which are presented in Table 1 support an
associative mode of activation for the substitution of the leaving group from
the Pt(II) complexes. The activation entropies are all negative, while the
activation enthalpies are relatively low. Characteristically negative activation
entropy and activation volume (27) for the forward reactions suggested that
in all cases nucleophilic substitution proceeded via a bimolecular activation
transition state in which bond formation was the rate-limiting step (5,28).
According to the results obtained for the substitution reactions presented
herein, the sensitivity of the rate constants to the nature (donor properties)
as well as the concentration of the entering ligands confirms the associative
character of the activated complex forming the substituted products. This is
expected since the complexes studied are all four square-planar and can
accommodate the synchronous breaking of the bond to the leaving group
and the formation of the new bond with the substituting nucleophile via
an 18-electron triangular bipyramidal transition state (7).
The important results, just described, stimulated us to carry out more
related studies on this subject aimed at enacting a systematic control of
the kinetics of the substitution reactions of the Pt(II) complexes through
changing their electronic and steric properties of the spectator ligand back-
bone. Kinetic control of the substitution behavior of the complexes is
important for optimizing their pharmacokinetic properties as well as their
efficacy as potential antitumor Pt(II) drugs. Herein, we report on some of
the highlights of the work which has been carried out in our laboratory
on the role of nitrogen-donor multidentate ligands on the rate of substitu-
tion of square-planar Pt(II) complexes in acidic aqueous or dry methanol
media.
2. PULL-AND-PUSH EFFECTS OF PT(II) COMPLEXES WITH

COORDINATED N^C/N^N/C π-ACCEPTOR
NONLEAVING LIGANDS
2.1 Controlling the π-Acceptor Properties of N^C/N^N/C
Chelates via a Donor Atom Effect
Literature is abound with d8 Pt/Pd(II) complexes of rigid tridentates of the
form N^N^N (e.g., terpy, its derivatives, structural analogues with azole/
azines and pyridyl/amine tridentates), N^C^N (1,3-bis(2-pyridyl)benzene),
or N^N^C (6-phenyl-2,20 -bipyridine). These ligands bind the Pt/Pd atom
in a tridentate fashion with the fourth position of the square-planar occupied
by the labile ligand. This is a preferred coordination geometry for monitor-
ing the ligand substitution reactions because only the coligand is replaced
from the complex. Pt/Pd(II) complexes of terpy are good references for
studying the effect of such π-acceptor chelates on the substitution reactivity
of Pt(II) complexes (6,29). This is because the terpy ligand is both a powerful
σ-donor, due to its relatively hard nitrogen lone pairs, and a strong
π-acceptor (12). The latter property is a result of the low-lying unfilled
π*-orbitals of the aromatic rings. However, the σ-effect compared to that
for unsubstituted amines is somehow inferior. Excellent reports reviewing
the reactivity relationships of several square-planar Pt/Pd(terpy) analogues
have appeared in the literature (19,30).
We undertook several case studies to understand the effect of changing
the structure of typical strong N^C/N^N/C π-acceptor ligands on the reac-
tivity of their analogous Pt(II) complexes. We synthesized several analogues
(Pt2–Pt10) of [Pt(terpy)Cl]+ (Pt1) for this purpose. Their structures are
shown in Fig. 3. In most of the case studies, Pt1 was used as a comparative
complex.
Details of the kinetic trends and conclusions drawn from them will be
elaborated in sections that follow. For all the case studies, we monitored
the rate of chloride or aqua substitution from the complexes as a function
of concentration of the incoming nucleophiles as well as the temperature
of the reaction under pseudo-first-order conditions, by providing the nucle-
ophile in at least 10-fold excess over the concentration of the complex.
A nonlinear fit of kinetic traces (see Fig. 4A for a typical example) accom-
panying the substitution process to standard exponential functions provided
values of the observed rate constant as reciprocals of the reaction lifetimes, τ.
C(CH3)3
(CH3)3C C(CH3)3
N N N
N N N N N N
Pt Pt
Pt
Cl Cl Cl
Pt1 Pt3 Pt4
me F
N N
Pt N N N N
Pt Pt
Cl
Cl Cl
Pt2 Pt2 Pt2
N N N N N
N N N N N N N N
Pt Pt Pt Pt
Cl Cl Cl Cl
Pt5 Pt6 P8
Pt7
F3C F3C
N N
N N N
Pt Pt
Cl Cl
Pt9 Pt10
Fig. 3 Structure of Pt(II) complexes with terpy and analogous chelates.
A B
TU
0,75 3,75
0,60 3,00
Rel. abs.
kobs (s–1)
0,45 2,25
0,30 1,50 DMTU
0,15 0,75 I
–
TMTU
–
SCN
0,00 0,00
0 300 600 900 1200 1500 1800 2100 0,000 0,002 0,004 0,006 0,008 0,010
Time(s) [Nucleophile] (mol dm–3)
Fig. 4 (A) Kinetic traces from the stopped-flow analyzer for the chloride substitution
from Pt4 by thiocyanate (SCN) (31). (B) Dependence of kobs on the concentration of
incoming nucleophiles. Reproduced from Reddy, D.; Jaganyi, D. Dalton Trans. 2008,
(47), 6724–6731 with permission from the RSC.
Plots of kobs vs concentration of the nucleophiles were linear, with near-

zero y-intercepts as depicted in Fig. 4B. The majority of the substitution
reactions between chelated N^C/N^N/C Pt(II) complexes (especially those
with strong π-acceptor ligands) and strong thiourea nucleophiles were irre-
versible. However (and only in rare cases), the solvent can also participate in
the substitution process and the overall substitution process can be presented
as shown in Fig. 5.
The influence of the solvent on the substitution process (as generalized
in Fig. 5) was insignificantly small and could be neglected for most of the
reactions where thiourea ligands were used as substituting ligands. The
overall rate law for substituting the coligands from the complexes could
be written as:
kobs ¼ k2 + k2 ½Nu k2 ½Nu, Nu ¼ thiourea ligands or azoles,
where k2 and k2 are the first-order rate constant for the solvolytic pathway
and the second-order rate constant for the direct substitution of the labile
2 2
R R
X
D1 D1
k2
1
R D2 Pt Cl + Nu 1
R D2 Pt Nu
N N
2
2 R
R
2
R
k–2 + S
Fast + Nu
D1
1 S
R D2 Pt
2
R
R1 = R2 = H; D1 = D2 = N : Pt1. R1 = R2 = H; D1 = C; D2 = N : Pt2.
R1 = CH3 ; R2 = H; D1 = C; D2 = N : Pt2me. R1 = F; R2 = H; D1 = C; D2 = N : Pt2F.
R1 = R2 = tBu-; D1 = D2 = N : Pt4.
phne
R1 = R2 = H; D1 = D2 = N : Pt3.
R1 = Ph-; CH3Ph-; R2 = H; D1 = D2 = N : Pt5; Pt6.
R1 = CH3Ph-; R2 = H; D1 = D2 = Npzn: Pt7. R1 = CH3Ph-; R2 = H; D1 = D2 = NQn: Pt8.

CF3
R1 = Ph-; R2 = H; D1 = D2 = N : Pt9. R1 = CF3Ph-; R2 = H; D1 = N; D2 = C : Pt10.
Fig. 5 A summary of the substitution process of the Pt(II) complexes.

ligand from the complexes, respectively. The latter was deduced from the
slope of the kobs vs concentration of nucleophiles, an example of which is
shown in Fig. 4.
In all case studies, the mechanism of substitution was limiting associative
(A) in nature, characterized by low enthalpies of activation and significantly
negative entropy of activation. For this mechanism, the final product forms
rapidly from a pentavalent-activated complex, involving a direct attack by
the nucleophile on the complex. The formation of a new bond to the nucle-
ophile is the rate-limiting step. Thus, the activated complex is characterized
by a decrease in the entropy and a collapse in the partial volumes (27) when
compared to the initial states of the reactants.
As a follow-up to our earlier studies, we studied the effect of changing
the structure of a typical strong π-acceptor ligand such as terpy on the reac-
tivity of the Pt(II) derivative complexes. Studies have shown that Pt(II) com-
plexes of terpy are bioactive and can efficiently inhibit human thioredoxin
reductase (32,33) and display antiprotozoal activity (34) as well as anticancer
potency (35). The potential applications of derivatives of terpy Pt(II) com-
plexes in oncology are rooted in their ability to link covalently to and at the
same time intercalate between the nucleobases of DNA. Although the
monofunctional DNA conjugates of these complexes are noncytotoxic, a
combination of intercalative association of the planar aromatic heterocyclic
ligands between adjacent base pairs of DNA and covalent binding is a good
recipe for effecting impairments that induce instability within the helical
structure of DNA. The instability is fundamentally related to the combined
structural effects of the monofunctional covalent binding and disruption of
the interstrand dipolar and intrastrand π/π*interactions between the com-
plementary bases and planar rings, respectively. The extent of helical insta-
bility depends also on the π-surface size of the planar backbone of the
nonleaving ligand as well as the charge on the metal complex. The des-
tabilized complex conjugate between DNA and the Pt(II) intercalator acti-
vates cellular pathways, some of which have a chemotherapeutic effect in
cancer cells. Several in vitro studies have shown that Pt(II) complexes with
planar aromatic ligands including terpy, its derivatives, and other related
ligands destabilize the helical structure of DNA via favorable intercalative
and/or stacking interactions, some of which have relevance to their
in vitro anticancer activity (35–37). For example, the monocationic Pt(II)
terpy derivative, [Pt(terpy)(2-hydroxyethanethiolate)]+, has been known
to intercalate into calf thymus DNA with a remarkably high binding con-
stant of 1.2(0.2) 106 M1 at pH 6.8 (38). A terpy derivative complex,
[Pt(II) (4-tolylterpy)]2+, exclusively coordinates the adenine nucleobases

present in the human telomeres loop of the G-quadruplex DNA (37). In
addition, the latter terpy complex has shown promising antiproliferation
(in some cases better than carboplatin) in certain cancer cell lines.
To understand more on the effect on the rate of changing the π-acceptor
properties of terpy, we synthesized and coordinated isostructurally related
nonleaving ligands to the Pt(II) ion and then studied the kinetics of their
substitution reactions. First, we studied the effect on the rate of chloride
substitution of replacing one of the N-donor atoms of terpy with a carbon
(C) atom on the trans- (Pt1 vs Pt2) or cis-position (Pt9 and Pt10) to
form cyclometalated complexes, wherein the deprotonated carbon atom
of the phenyl ring is coordinated to Pt(II) on the trans or lateral position
to the leaving ligand (39,40). For structures of Pt2, Pt9, and Pt10, refer
to Fig. 3.
When the central pyridyl ring of [Pt(terpy)Cl]+ was replaced by a
deprotonated phenyl ring to form the analogous Pt(N^C^N)Cl (Pt2) com-
plex, reactivity was enhanced by a factor of about 30 (39) when thiourea was
used as the substituting nucleophile (Table 2, rate constants, entry 3 vs 1). This
was expected on the basis of the strong trans labilization effect of the
deprotonated phenyl ring of Pt2, both in the ground and transition states
(43–45). The crystal structure of Pt2 (46) depicts a significantly longer
PtdCl bond compared to that of Pt1 as a result of the stronger trans-
influence of the PtdC bond of Pt2 on its ground-state properties. This
facilitates the substitution of the leaving group by an incoming nucleophile.
However, when one of the terminal pyridyl rings of Pt9 was replaced by a
deprotonated phenyl ring to form the analogous Pt(N^N^C)Cl (Pt10) (40),
an unexpected lower reactivity was observed when thiourea ligands were
the incoming nucleophiles (Table 2, rate constants, entry 6 vs 10). A similar
reactivity trend was observed for the Pt9/Pt10 couple when each was
reacted with azoles of different basicity and steric demand (47). As shown
in Table 2 (rate constants, entries 6 and 10), the rate of aqua substitution from
Pt10 by tu is about 7% less than for its reactive isostructural analogue, Pt9.
The same deceleration was observed for the reactivity difference of the
unsubstituted analogues, Pt1 and Pt2 (39). For Pt10, a deprotonated phenyl
ring, and thus a stronger σ-donor, is effectively placed cis to the leaving
group. The lower reactivity of Pt10 compared to Pt9 was attributed to
accumulation of electron density along the cis bond to the Pt(II) atom
due to the stronger σ-donor phenyl ring. The increase in electron density
in the dxy-orbital of Pt has an insignificant kinetic influence on the
Table 2 Second-Order Rate Constants, k2 (Std. Error), M1 s1 at 25°C for Chloride Substitution From Deprotonated Phenyl Ring in the
Tridentate Nonleaving Ligand
Kinetic Data
tu tmtu I2
Labilization Factor for tu {k2( j)/k2(Pt1 or Pt2)},
Complex k2 (M21 s21) k2 (M21 s21) k2 (M21 s21) Ref. *(%)Increase/or Decrease
Pt(N^N^N)Cl (Pt1) 1.50 9.5(0.4) 2.67 (15)
(0.01) 101 (0.01 102
103
1.49 8.2(0.4) 2.4(0.04) (31,40) 1
(0.01) 101 102 (I data)
103
Pt(N^N^C)Cl (unsubstituted 9.8(0.2) 3.2 0.64(0.02) (15) 0.065 (93.5%)
analogue of Pt10) 101 (0.03)
101
[Pt(N^C^N)Cl]+ (Pt2) 4.45 7.4(0.3) 1.46 (15) 29.7Pt1 (2880%)
(0.2) 104 103 (0.01) 103
[Pt(N^Cme^N)Cl]+ (Pt2me) 5.2(0.1) 2.3(0.1) (41) 1.17Pt2 (30%)
104 104
[Pt(N^CF^N)Cl]+ (Pt2F) 8.9(0.2) 4.6 (41) 2.00Pt2 (200%)
104 (0.07)
104
Continued
Table 2 Second-Order Rate Constants, k2 (Std. Error), M1 s1 at 25°C for Chloride Substitution From Deprotonated Phenyl Ring in the
Tridentate Nonleaving Ligand—cont’d
Kinetic Data
tu tmtu I2
Labilization Factor for tu {k2( j)/k2(Pt1 or Pt2)},
Complex k2 (M21 s21) k2 (M21 s21) k2 (M21 s21) Ref. *(%)Increase/or Decrease
[Pt(N^NPhCF3^C)Cl]+ (Pt10) 1.23 2.9 (40) 0.082Pt2 (16.5%)
(0.003) (0.05)
102 101
[Pt(tBut3terpy)Cl]+ (Pt4) 4.52 3.8 6.0(0.04) (31) 0.30Pt1 (70%)
(0.05) (0.02) 101
102 101
[Pt(Phterpy)Cl]+ (Pt5) 1.26 1.00 (40) 0.84Pt1 (16%)
(0.01) (0.003)
103 101
[Pt(CH3Phterpy)Cl]+ (Pt6) 1.22 5.4 (42) 0.81Pt1 (19%)
(0.008) (0.07)
103 101
[Pt(CF3Phterpy)Cl]+ (Pt9) 1.93 1.44 (40) 1.3Pt1 (30%)
(0.03) (0.001)
103 102
Labilization factor ¼ k2(Pt10)/k2(Pt9), when tu is the nucleophile.
(%)Increase/decrease ¼ {k2( j) k2(Pt1 or Pt2/Pt1)}/k2(Pt1 or Pt2/Pt1).
ground-state properties of the bond between the Pt atom and leaving group
as it occurs in the measurable trans-influence of a σ-strong donor in the trans
position (44). This decreases the natural nonbonding orbital (NBO) charges
and hence the Lewis acidity of the Pt atom, leading to lower rates of sub-
stitution of the leaving group from the complex. As can be noticed from
the data in Table 2 (rate constants, entry 2 vs 3), the cis-influence (39,40) is
somehow weaker than the trans-influence for square-Pt(II) complexes with
mixed σ/π groups within the coordinated chelate.
2.2 Controlling the π-Acceptor Properties of Terpy via

Extended Ring Conjugation
The complex, [Pt{2-(20 -pyridyl)-1,10-phenanthroline}Cl]Cl, (Pt3) was
synthesized and the chloride substitution reactions in methanol studied using
a series of thiourea nucleophiles (48) or azoles (49) under pseudo-first-order
conditions. Through a comparison of rate constants applicable for Pt3 with
those of Pt10 (Table 3, rate constants, entry 2 vs 1), it was established for both
case studies that increasing the π-conjugation in between the cis and the trans
position of terpy via replacing the “bipyridyl” with a “phenanthroline” moi-
ety in the extended chelate structure increases the rate of chloride substitu-
tion. In other words, increasing the π-surface connecting the coordinated
pyridyl rings effectively increased the π-acceptor property of the ligand
framework, thereby increasing the reactivity of the metal complex toward
substitution.
A similar trend (47) (Table 3, rate constants, entry 4 vs 3) was observed for
the reactions of azoles with a structurally related complex, Pt7, an analogue
with a pyrazinyl ring coordinated in the cis-position. Since pyrazine is a bet-
ter π-acceptor than pyridine, it enhances the π-back donation effect of the
resultant ligand. The Pt(II) atom of Pt7 becomes more electron deficient
compared to that of Pt6 (the reference complex), making it more prone
to nucleophilic attack by the incoming nucleophile. This is supported by
the Pt NBO charges, which are significantly more positive for Pt7 compared
to Pt6 (47).
In a related structural modification of the nonleaving ligand, the reactivity
of [Pt{40 -(2000 -CH3-phenyl)-2,20 :60 ,200 -terpyridine}Cl]+, (Pt6), and [Pt{40 -
(2000 -CH3-phenyl)-6-(300 -isoquinoyl)-2,20 -bipyridine}Cl]+, (Pt8), toward
the substitution of chloride by thiourea nucleophiles was studied under
pseudo-first-order conditions (48,49). A curious decrease in reactivity of
Pt8 relative to Pt6 (Table 3, rate constants, entry 5 vs 3) was observed and attrib-
uted to the relatively poor π-acceptor property of the cis-coordinated
Table 3 Second-Order Rate Constants, k2 (Std. Error), M1 s1 at 25°C for Chloride Substitution From Complexes With π-Conjugated Nonleaving
Ligands
Labilization Factor {k2(j)/k2(ref.) (%)},
Kinetic Data *(%)Increase/or Decrease
tu tmtu Pz
21 21 21 21
Complex (Modification) k2 (M s ) k2 (M s ) k2 (M21 s21) Ref. tu Pz
Pt(N^N^N)Cl (Pt1) 1.50(0.01) 10 3
9.5(0.4) 10 1
2.25(0.1) (47)
(reference complex for Pt3)
1.49(0.01) 103 8.2(0.4) 101 1.26(0.02)NB (40) 1 1
Pt3 (insertion of an extra 3.42(0.04) 10 3
1.4(0.01) 10 2
(48) 2.29Pt1 (129%)
pyridine ring between the NB
3.20(0.2) (49) 2.5Pt1 (154%)
central and flanking rings of terpy)
Pt6 (reference complex for Pt7/Pt8) 1.26(0.008) 103 8.4(0.1) 101 (48) 1
0.94(0.2)NB (49) 1
Pt7 (the flanking pyridine ring of Pt6 6.1(0.5) (47) 6.5Pt6 (549%)
replaced with a pyrazine)
Pt8 (a flanking pyridine ring in 3.76(0.03) 102 2.2(0.07) 101 (48) 0.297Pt6
terpy of Pt6 replaced with a quinoline) (70.3%)
0.36(0.01)NB (49) 0.38Pt6
(61.7%)
Pz ¼ pyrazole; NB ¼ reaction solvent for Pz is dry methanol. Data for other reaction were in dry methanol adjusted to an ionic strength of 0.1 M (90 mM LiCF3SO3 and 10 mM
NaCl). *(%)Increase/decrease ¼ {k2( j ) k2(Pt1 or Pt2/Pt1)}/k2(Pt1 or Pt2/Pt1).
isoquinoline moiety which reduces the π-acceptor property of the entire

ligand compared to a terpy-like analogue. As a result, the electrophilicity of
Pt8 was slightly lower than that of Pt6, resulting in the retardation of the
incoming nucleophiles toward substitution. This reasoning is well supported
by computational data in the gaseous phase, which clearly showed lower
NBO charges on the Pt atom of Pt8 and a large energy separation gap
(ΔE) between the frontier orbitals of the complex. Further, the photophysical
behavior of the free (300 -isoquinoyl)-2,20 -bipyridine ligand confirms a weaker
π-acceptor capacity of the ligand, characterized by higher energy MLCT
bands when compared to those recorded for the parent terpy analogue.
The emission bands are also blue-shifted due to the significantly higher energy
for the π∗-LUMOs of Pt8 when compared to those of Pt1 or Pt6. High ener-
gies of the LUMO, usually centered on the π-acceptor ligand, means the
capacity of the π∗-orbitals to stabilize the transition state by back donation
of electron density from the metal is weaker. A quinoline attached at the lateral
positions of the tridentate destabilizes the energy states of the LUMO, thereby
weakening the capacity of the entire ligand to accept electrons into the π∗-
LUMO. As a result, the electrophilicity of the complex is low, leading to
retarded rates of substitution. This is supported by data from geometrically
optimized calculations of the complexes.
2.3 Controlling the π-Acceptor Properties of Terpy and Related

Chelates Through a Substituent Effect
To extend our knowledge on the structure/reactivity relationships of the
terpy and its allied Pt(II) complexes, we directed our focus on studying
the effect on the rate of reaction of substituents on the terpy ligand. We syn-
thesized [Pt(tBut3terpy)Cl]+ (Pt4) (where tBut3terpy ¼ 4,40 400 -tri-tert-
butyl-20 200 :60 200 -terpyridine) (31). The crystal structure of Pt4 was already
reported in the literature (50), and for its molecular structure, see Fig. 3.
It is an analogue of Pt1, which has three strong electron-donating groups
on each ring of terpy. The kinetics of substitution of chloride from the com-
plex were studied using a series of neutral and anionic nucleophiles (31).
These were carried out under pseudo-first-order conditions, in acidic aque-
ous medium, as a function of concentration and temperature using the stan-
dard stopped-flow spectrophotometric technique. Through a comparison of
the second-order rate constants for the substitution of chloride from Pt4
(Table 2, rate constants for tu, entry 6) (31) and Pt1 (40), it was shown that
introducing strong σ-donating groups (tert-butyl) on the π-acceptor terpy
ligand in Pt4 decreased the reactivity due to the reduced capacity of the
terpy ligand to receive electron density from the d orbitals of the metal into
its π*-acceptor molecular orbitals. The tert-butyl groups donate electrons
inductively into the terpy rings, thereby reducing the π*-effect of the ligand.
This destabilized the formation of the activated complex, expected to
accommodate extra electrons from the incoming nucleophile by raising
the energies of the ligand-centered LUMOs. This resulted in retarded rates
of substitution. A similar result was reported in a comparative case study in
which these two complexes were reacted with azole nucleophiles (47). DFT
calculations of Pt1 and Pt4, and a series of other analogues containing either
electron-donating or electron-withdrawing groups in the periphery posi-
tions affirmed that the introduction of electron-donating substituents on
the π-acceptor terpy ligand decreased its propensity to receive electron den-
sity from the metal center. The positive charge on the metal center of Pt4 is
lower than on Pt1, while the energy separation of the frontier molecular
orbitals (EHOMO ELUMO) of the ground states of the complexes increased
from Pt1 to Pt4. This reflects a less reactive metal center, while the opposite
is true for electron-withdrawing groups (31).
The study of the substituent effect on the rate of substitution of chloride
was extended to two analogues of Pt2 with either a 40 -methyl (Pt2me) or a
40 -fluoro (Pt2F) group on the 1,3-bis(pyridyl)benzene (N^C^N)
nonleaving ligand (41,51). Both complexes were more reactive than Pt2
(39,41) (Table 2, rate constants, entries 4–5 vs 3) toward thiourea ligands
and azoles (51), while the 40 -fluoro derivative was slightly more reactive
than the 40 -methyl-substituted complex. In general, the 40 -substituents on
the phenyl ring of the NCN ligand increased the rate of substitution of
the complexes relative to the rate for Pt2. The electron-donating methyl
group destabilized the ground-state properties of the complex compared
to those of Pt2. The electron-withdrawing 40 -fluoro-substituent enhanced
the π-acceptor property of the ligand relative to that of the unsubstituted
ligand of Pt2. As a result, both the Pt2me and the 40 -fluoro (Pt2F) were
more reactive than Pt2, with the latter slightly edging the former.
2.4 Controlling the π-Acceptor Properties of Terpy Through

an Ancillary Substituent Effect
A comparison of analogues of Pt1 carrying substituted phenyl rings bearing
an dH (Pt5), ortho-methyl (Pt6), or an ortho-trifluoromethyl group (Pt9)
on the 40 -position of terpy (Table 2, rate constants, entries 8–10) revealed that
the push-and-pull effect on the π-acceptor property of a coordinated terpy
ligand could be effected by the 40 -Ph group by an extent modulated by its

own ancillary substituent (52). Consequently, Pt9 is more reactive than Pt1,
while both Pt5 and Pt6 are slightly less reactive. This shows clearly the effect
on the rate of substitution by the 40 -Ph group as modulated by its ancillary
substituent, despite the 40 -Ph group being ex-communication from the
terpy’s π-conjugated molecular orbitals (40,42). In two separate case studies
investigating this hypothesis, Pt(II) complexes which included Pt1, Pt6, and
Pt9 were reacted with thiourea nucleophiles (42) or azoles (47) of different
basicity and steric constraints.
An inspection of the single X-ray crystal structures of Pt6 (53) and its
substituted product Pt6pz (47) (see Fig. 6) is helpful in explaining the origins
of this ancillary substituent effect. Pt6pz was synthesized independently for
the purposes of confirming it as the nucleophilic substitution product of
chloride from Pt6 by azoles. The crystal structures of Pt6 (53) and
Pt6pz (47) show that the substituted 40 -phenyl ring is twisted at an angle
out of the plane of the terpy in order for the complex to attain a low-energy
Fig. 6 Crystal structures of Pt6 (53) and its pyrazole-substituted product, Pt6pz(47).
Reproduced from Field, J. S.; Haines, R. J.; McMillin, D. R.; Summerton, G. C. J. Chem.
Soc. Dalton Trans. 2002, 1369–1376 and Reddy, D.; Akerman, K. J.; Akerman, M. P.;
Jaganyi, D. Transition Met. Chem. 2011, 36 (6), 593–602 with permission from the IUC
and RSC, respectively.
structure in which the ortho-substituted ancillary substituents are as far apart

as possible from the plane of the terpy.
These structures clearly illustrate that extended π-delocalization of elec-
tron density which is back-donated from the metal orbitals into the terpy
cannot extend as far as the substituents on the 40 -phenyl ring due to a twisted
conformation of the phenyl ring (42). Thus, the observed difference in the
reactivity of Pt6 and Pt9 relative to Pt5 is accounted for solely by the
σ-donor strength of the phenyl ring into the terpy via the 40 -C–C bond
as modulated by its ancillary substituents (40,42). The trifluoromethyl group
withdraws electron density from the attached phenyl ring, while the methyl
group donates into the ring. Thus, Pt9 is more reactive than Pt5 due to a
reduced π-effect of the terpy by the substituents on the 40 -phenyl ring, while
the opposite is true for Pt6. Similar trends in reactivity were observed in
other case studies (40) when Pt6 and Pt9 were reacted with thiourea nucle-
ophiles. As was noted for primary substituents (Pt4 vs Pt1), an ancillary sub-
stituent on the 40 -ortho-substituted phenyl groups on the terpy affects the
reactivity of the complexes to an extent determined by its capacity to donate
inductively into the terpy ligand.
2.5 Incorporating Bis(Pyrazolyl)Pyridine/Benzene; Bis(70 -

Azaindolyl)Pyridine/Benzene; 1,3-Bis(80 -Quinolyl)Pyridine/
Benzene; or Bis(Iminopyridyl/Isoquinolyl)Isoindoline as
Nonleaving Ligands of Square-Planar Pt(II) Complexes
In separate case studies, we focused on investigating the substitution kinetics
from Pt(II) complexes incorporating the following nonleaving ligands:
2,6-bis-(pyrazolyl)pyridine (54); 1,3-bis-(pyrazolyl)benzene (55); 2,6-bis
(70 -azaindolyl)pyridine (54); 1,3-bis(70 -azaindolyl)benzene (55); 2,6-bis
(80 -quinolyl)pyridine (54)/benzene (55); 1,3-bis(80 -quinolyl)benzene (55);
and 1,3-bis(iminopyridyl/isoquinolyl)isoindolines (56) with an aim of fur-
ther tuning the push/pull properties around the Pt(II) metal center.
Through a comparison of the rate constants for substituting the labile ligands
from these complexes and the reference complexes Pt1 and Pt2, the
π-acceptor capacities of the nonleaving ligands were inferred. The schematic
structures of the complexes are shown in Fig. 7.
Chloride substitution from Pt11–Pt13 (54) or from Pt14–Pt16 (55) by
thiourea was undertaken in dry methanol solutions. A striking structural fea-
ture of these complexes is that none of the coordinated flanking aromatic
moieties, the pyrazolyl (Pt11; Pt12; Pt14), the 80 -quinolyl (Pt15), nor
the 70 -azaindolyl (Pt13/Pt16) are as efficient π-acceptors as are the flanking
N N
N
N N N N N N N
N N N N N N N Pt N
Pt Pt Pt
Cl
Cl Cl Cl
Pt12 Pt13
Pt1 Pt11
N N N N
N N N N
Pt Pt N Pt N
N Pt N
Cl Cl Cl
Cl
Pt2 Pt14 Pt15 Pt16
N N N
N N N N N N N N N
N Pt N
N Pt N N Pt N N Pt N
Cl
Cl Cl Cl
Pt17 Pt18 Pt19
Pt20
Fig. 7 Schematic structures of Pt(II) complexes with 2,6-bis(pyrazolyl)pyridine (54)/
benzene (56) or 2,6-bis(7-azaindolyl/80 -quinolyl)pyridine (54)/benzene (55) or
isoindoline (56) nonleaving ligands.
pyridyl rings of terpy (Pt1) (12,14) or those of 1,3-bis-(pyridyl)benzene

(Pt2) (12). For that reason, the electrophilicity of the metal centers for these
complexes, modulated by the π-acidity of the nonleaving ligand, is expected
to be significantly less than that of reference complexes Pt1 or Pt2, thereby
decreasing the lability of the complexes. This may partly explain the lower
reactivity toward substitution which was observed (55) for these complexes
when compared to reactivities for Pt1 or Pt2. A complete order of the
reactivity (second-order rate constants at 25°C are cited) of the analogues
with the deprotonated phenyl ring is Pt2 (4.4 104 M1 s1) ≫ Pt15
(1.7 104 M1 s1) > Pt14 (39.4 M1 s1) > Pt16 (0.55 M1 s1).
A similar trend was reported for the pyridyl derivatives (54).
DFT calculations were undertaken to explain the reactivity trend. The
DFT-optimized structures of the complexes (Pt13, Pt15, and Pt16) which
have 1,3-bis(70 -azaindolyl)benzene or 1,3-bis(80 -quinolyl)benzene nonleaving
ligands show ideal square-planar coordination of the tridentate ligands and
their coligands. Both classes of ligands form six-member chelates at the metal
center which are literally free from ring strain (57,58) when compared to
those of Pt2 (59). The bite angles for these complexes are near 180 degrees.
Such an ideal bite angle ensures inertness toward substitution due to a
stronger crystal field of the strain-free nonleaving ligand. To attain the
free-strain chelates, the ligand coordinates with a torsional twist about the
central phenyl/pyridyl-Pt plane with interplane angles in the range
19—31 degrees (55). The twisting is more pronounced for the 1,3-bis(80 -
quinolyl)benzene than for bis(70 -azaindolyl)benzene due to the difference
in the steric demands of the lateral rings in the ligands. Similar conforma-
tional distortions guaranteed a pure octahedral geometry for the Ru(II)
complex bearing the analogous ligand, 2,6-bis(80 -quinolyl)pyridine (60).
Additionally, the twisting of the coordinated ligands lowers the rate of sub-
stitution of the coligand compared to Pt2 because of the steric hindrance of
the off-plane moieties of the tridentates toward the incoming substituting
nucleophile. The out-of-plane distortions also put a bonding strain on
the linear combination of atomic orbitals forming the extended π-surface
of the frontier molecular orbitals of the planar tridentate. This weakens its
capacity to receive electron density from the d-orbitals of the metal by
raising the energies of the LUMO. This causes the substitution process to
be slower than in Pt2. Similar decelerations of the rate of substitution were
reported for Pt(II) complexes (61–63) with an isochelate (six-atom) struc-
tures. The lower reactivity of Pt16 (which has a bis(70 -azaindolyl)benzene
ligand) compared to Pt15 (with a 1,3-bis(80 -quinolyl)benzene ligand) is
attributed to the electron-rich nature of the lateral 70 -azaindolyl moieties
(57), which destabilizes the LUMO leading to a large HOMO–LUMO
energy gap (55).
The HOMOs of the complexes (55) are similar in their isodensity map-
ping. They are concentrated along the PtdCl bond, the Pt atom, and along
the Pt–C(deprotonated) of the central phenyl ring. The central phenyl ring lab-
ilizes the coligand through its strong trans-influence of the deprotonated
carbon. Additionally, some pockets of electron density are located on the
flanking aromatic rings (8-quinolyl and the 70 -azaindolyl), and this also con-
firms their weaker capacity to receive electron density through back bond-
ing. Thus, replacing the lateral pyridines of terpy with a pyrazolyl ring or
fusing 70 -azaindolyl/80 -quinolyl moieties between the pyridine rings of
terpy reduces the reactivity of their complexes towards substitutions due
to a weaker π-acceptor capacity of the coordinated ligand. The calculated
NBO charges on the Pt atoms of Pt11–Pt13 or Pt14/Pt16 were found
to be lower than those of Pt1 or Pt2, respectively. Based on the experimen-
tal data, the π-acceptor capacities of the N-donor moieties of the tridentate
nitrogen chelates (and hence the Lewis π-acidities impact at the metal cen-
ters) can be ranked as: pyridyl ≫ quinolyl > pyrazolyl > azaindolyl.
The effect of the size of chelate ring at the Pt(II) center on the rate of
substitution was broadened to include square-planar Pt(II) complexes of
the bis(iminopyridyl/isoquinolyl)isoindolines (Pt17–Pt19) (56). In general,
the neutral [bis(iminopyridyl/isoquinolyl)isoindolate)Pt(II)Cl] complexes
are significantly inert toward substitution when compared with Pt1. The
order of reactivity with tu as the substituting nucleophile is (second-order
rate constants at 25°C) Pt1 (1490 M1 s1) (40) ≫ Pt17 (155 M1 s1) >
Pt18 (30 M1 s1) > Pt19 (0.98 M1 s1).
The genesis of the inertness toward substitution of the bis(iminopyridyl/
isoquinolyl)isoindolines Pt(II) complexes arises mainly from their coordina-
tion geometry around the Pt(II) center. As shown in their optimized struc-
tures in Fig. 8 and also from literature reports on the crystal structures of
related complexes (64–66), the optimized structures are planar and their
geometry around the metal center is close to an ideal square-planar geom-
etry. The analogous ligand, 1,3-bis(2-thiazolylimino)isoindoline, also forms
near-perfect 1:1 or 1:2 octahedral complexes with Fe(II) ions (67). In
-
-
-
- -
-
Reference Pt17 Pt18 Pt119

Fig. 8 Frontier orbitals of Pt(II) complexes of bis(iminopyridyl/isoquinolyl)isoindolines
(56). Reproduced from Wekesa, I. M.; Jaganyi, D. Dalton Trans. 2014, 43(6), 2549–2558.
common with the isoindolyl square-planar Pd(II)/Pt(II) complexes are a

short central Pd(II)/Pt–Npyrrolate bond, longer Pd/Pt–Nlateral bonds, and a
quasi-tetrahedral distortion of the Pd/Pt–Cl bond (66). The near-perfect
square-planar angles about the Pt(II) ion thus induce less constraint than
would be found in the two five-membered chelating rings around the metal
center for the terpy complexes (12,53) and its derivatives (68,69), where the
bite angles are significantly smaller. This renders substitution at the fourth
site in the complexes studied slower when compared to that of Pt1, leading
to slower rates of reactions. This is despite the extended π-conjugation of the
coordinated bis(iminoaryl)isoindolates (64,65).
Another important difference in these complexes is that the central
donor Npyrrolate of the isoindolines is anionic. Thus, the chloride complexes
are neutral and are characterized by short central Pd/Pt–Npyrrolate bonds.
A lower formal charge reduces the effective charge on the Pt atom relative
to that of Pt1 in addition to canceling an increase in the electrophilicity
induced by the π-acceptor effect of the ligands. Understandably, the
bis(imino) arms of the ligands span the delocalization of electron density
in the entire ligand. On coordination, the ligand forms two planar six-atom
chelates at the metal center, where electrons are more efficiently delocalized
within the entire ligand. An expectation would be that the ligands function
as stronger π-acceptors, especially with the fusion of more phenyl rings on
the hind side due to an increased π-conjugation of the ligands. The reactivity
trend observed for these complexes as well as their reported photophysical
and electrochemical data (64,65) points otherwise. A derivative bearing a
phenyl(f )isoindoline moiety (Pt18) is even more inert toward chloride
substitution than Pt17. Why then is this the case? As shown in Fig. 8,
the LUMO orbitals of Pt18 are destabilized due to its naphthyl group con-
tributing more antibonding (π*) composition (65) of the LUMOs. This can
be tracked to an energy mismatch in the linear combination of the π*naphthyl
and the π*isoindolenic orbital characters of the naphthyl-fused isoindoline
ligand of Pt18. The same destabilization can be interpreted from the linear
combination of π*phenyl and the π*pyrrolate (reference complex) in the core iso-
indoline ligand of Pt17. As the π surface of the isoindoline is increased from
Pt17 to Pt18, the HOMO–LUMO energy gap increases as a result of the
destabilization of the LUMO orbitals. Fusing a naphthyl/phenyl moiety on
the hind side of the ligand makes back donation of electron density from the
metal d-orbitals into the LUMO more difficult, making these ligands intrin-
sically weaker π-acceptors. A consequence is a retarded rate of substitution of
the leaving ligand. An important feature when comparing the structures of
these ligands to terpy-type ligands is the bis(imino) groups which are the
gateways of the delocalization of excess electron density from the hind side
of the naphthyl/phenyl rings into the flanking π*-acceptor pyridines. The
imino bridges weaken the π-acidity of the pyridyl rings when compared
to those in terpy. This leads to slower rates of substitution of the coligand
in their complexes. Thus, the π-acceptor capacity of the N-donor moieties
of the tridentate nitrogen chelates (and hence the Lewis π-acidities impact at
the metal centers) can be ranked as: pyridyl ≫ isoindolyl > phenyl[f]
isoindolyl.
When the rate of chloride substitution from Pt19 was compared to its
isostructural analogue, Pt17, the former was found to be about two orders
of magnitude less reactive. This is possibly because the lateral isoquinoline
rings of Pt19 are intrinsically poorer π-acceptors (48,49) than the pyridines
in terpy. A weaker π-acceptor ligand at a position cis to the leaving ligand
decreases the effective charge at the Pt atom and reduces the electrophilicity
of the complex (48,49). This slows the substitution process further relative
to that of Pt17. This anomaly is supported by the trends in the DFT-
computed data. Pt19 has the largest HOMO–LUMO energy gap and the
lowest electrophilicity index among the bis(iminoaryl)isoindolate com-
plexes studied.
2.6 Controlling the π-Acceptor Properties of Other N-donor

Chelates
The effect on the rate of changing the π-acceptor property of the pap ligand
was studied. The study was aimed at quantifying the effect of changing the
trans donor atom of the pap ligand as well as checking if the anomalous effect
of replacing isolated pyridyl rings of pap with quinoline groups could be
replicated as was the case for the terpy-like ligand (49,70). We studied
the aqua and/or chloride substitution by thiourea nucleophiles from the
following complexes: [Pt(II)bis-(2-pyridylmethyl)amine]2+ (Pt20) (71);
[Pt(II)(2-pyridylmethyl)-8-quinolinylamine]2+ (Pt21) (71); [Pt(II)bis(8-
quinolinyl)amine]2+ (Pt22) (71); and [Pt(II)bis-(2-pyridylmethyl)sulfide]2+
(Pt23) (72). The structures of the complexes are shown in Fig. 9.
The lability of the chloride ligand from Pt20 and Pt23 (structural ana-
logues of Pt(pap)) using azoles as incoming nucleophiles was studied under
pseudo-first-order conditions in acidic aqueous medium (72).
The rate of reaction was found to depend strongly on the strength of the
σ-donation character of the atom trans to the leaving group. As shown by the
calculated labilization factor (rate constants, entry 4 vs 1) in Table 4, Pt23 was
N N N
N Pt N N N
Pt N Pt N
Cl Cl Cl
Pt21 Pt22 Pt23
S
N Pt N
Cl
Pt24
N
N
Cl
Cl 1 N Pt
1 N Pt R
R N
N Cl
Cl
2
2 R
R
Pt25 Pt26
Fig. 9 Structures of Pt(pap)Cl, its analogues, and two Pt complexes with
2-(pyrazolylmethyl)pyridine or 8-quinolylamine ligands.
found to be three orders of magnitude more reactive than Pt20 toward

imidazole. The thioether group of Pt23 imparts both σ-donor and
π-acceptor (73) influences on the leaving group. The sulfur atom donates
its lone pair of electrons inductively via a bond trans to the leaving group
and this destabilizes the complex in the ground state. In addition, the rela-
tively low-energy and unoccupied dπ* orbitals of the sulfur can be used for
π-back bonding of electron density from the metal. This type of back dona-
tion is not possible in Pt20, because such orbitals are not accessible on
energy grounds. Thus, the Pt atom of Pt23 is more electrophilic and more
amenable to stabilize the nucleophile in the transition state, leading to higher
reactivity.
In the second study (71), the effect on the rate of replacing isolated pyr-
idyl rings of pap with (80 -quinolyl)amine groups was investigated.
A complete trend in reactivity of the complexes (including complexes from
the first case study) followed the order: Pt23 ≫ Pt20 > Pt22 > Pt21.
Substituting isolated pyridine rings in pap with 80 -quinolyl moieties in
the nonlabile chelate framework of Pt21 and Pt22 accumulated electron
Table 4 Second-Order Rate Constants, k2 (Std. Error), M1 s1 at 25°C for Chloride
Substitution From the Complexes
Kinetic Data Labilization
Factor for
tu tmtu Im
tu/Im {k2(j)/
k2(ref.) (%)},
Complex *(%)Increase/or
(Modification) k2 (M21 s21) k2 (M21 s21) k2 (M21 s21) Ref. Decrease
1 2
(Pt20) or 5.0(0.7) 10 3.4(0.04) 10 (71) 1
Pt(bpma)Cl
(derivative of 0.45 (72) 1
Pt(pap)) (0.01)
(reference
complex)
Pt21 2.9(0.3) 101 4.9(0.1) 102 (71) 0.57Pt20
(47.2%)
Pt22 0.9(0.7) 101 2.8(0.1) 102 (71) 0.178Pt20
(82.2%)
Pt23 9.50 (72) 2108Pt20
(0.1) 102 (210.8%)
Im ¼ imidazole; *(%)increase/decrease ¼ {k2( j ) k2(Pt1 or Pt2/Pt1)}/k2(Pt1 or Pt2/Pt1).
density at the metal center, which retarded the substitution of the chloride
by the incoming nucleophile. A similar reactivity trend had been reported
before when an isoquinolyl group was incorporated in one of the lateral pyr-
idyl rings of a terpy chelate, vide supra (48,49). It is clear that despite its
extended π-surface, the tridentate ligand, bis(8-quinolinyl)amine, is a
weaker π-acceptor of electron density from the metal center compared to
terpy. This qualifies 8-quinolyl moieties as better σ-/π-donors toward the
metal center compared to pyridyl rings. The energy level (2.28 eV) of
the LUMO for Pt22 is higher than that of Pt20 Pt (3.40 eV), which sup-
ports a weaker back donation of excess electron density from metal orbitals
into the ligand-centered LUMOS.
In a more recent study (74,75), a set of Pt(II) complexes bearing either a
2-(pyrazolylmethyl)pyridine or a 2-(pyrazolylmethyl)quinoline bidentate
ligand was synthesized and the aqua substitution studied in acidic aqueous
medium under pseudo-first-order conditions using UV–visible spectro-
photometry. The structures of two of the complexes namely: [Pt(II){2-
[3,5-bis(trifluoromethyl)pyrazol-1-ylmethyl]quinoline}]2+ (Pt24) and [Pt(II)
{2-[3,5-bis(trifluoromethyl)pyrazol-1-ylmethyl]pyridine}(H2O)2]2+ (Pt25)
are shown in Fig. 9. The reactivity trend confirmed that the quinoline sub-
structure in the (pyrazolylmethyl)quinoline ligands acts as an apparent donor
of electron density toward the metal center rather than being a strong
π-acceptor of electron density from the orbitals of the metal. The phenyl ring
of the benzopyridine weakens the overall π-acceptor property of the quinoline
moiety of the bidentate ligand, leading to slower rates of substitution from the
complex compared to its pyridyl analogue. By comparing the reactivity of the
other complexes in the two case studies, electron-withdrawing substituents on
the 3,5-positions of the pyrazolyl ring of the [3,5-bis(trifluoromethyl)pyrazol-
1-ylmethyl]pyridine nonleaving ligand enhanced the π-acceptor of the
pyrazolyl ring, leading to faster rates of substitution. The opposite was true
for electron-donating groups.
3. CONCLUSIONS
From a comparative analysis of the reactivity of Pt(II) complexes
with rigid N^C/N^N/C tridentate nonleaving ligands, it can be concluded
that the substitutional lability of the leaving ligand depends to a greater
extent on the strength of the π-back bonding of the π-acceptor properties
or the π /σ-trans-effect of the N^C/N^N/C nonleaving ligand as well
as the coordination strain on the leaving group due to acute bite angles.
To a smaller extent, it depends also on the denticity of the ligand on the
platinum atom.
When the central pyridyl ring of a coordinated terpy is replaced by a
deprotonated benzene, the stronger σ-donor influence of the PtdC desta-
bilizes the ground-state properties of the bond trans to it. This facilitates the
substitution of the leaving group by an incoming nucleophile. The same
effect, for example, is observed when other strong σ-donor groups such
as the thioether (SR2) replace a secondary amine (NR2) donor in bpma,
a N^N^N tridentate bearing isolated pyridyl rings.
However, when one of the lateral pyridyl rings of the terpy ligand was
replaced by a deprotonated phenyl ring, accumulation of electron density at
the Pt atom decelerated the rate of substitution of the leaving group. A cis-
coordinated σ-donor (e.g., a phenyl ring) or a poor π-acceptor moiety
(e.g., an isoquinoline moiety) within the ligand framework of a typical
π-acceptor accumulated electron density at the metal center, leading to
lower rates of substitution.
Extending the π-surface of the terpy by replacing a bipyridyl moiety of

the ligand with a phenanthroline group strengthens the π-acceptor capacity
of the entire chelate ligand and enhances the reactivity of the complex in a
significant way. The same effect is observed when a stronger π-acceptor
moiety such as a pyrazinyl ring replaces one of the lateral pyridyl rings of
terpy. In sharp contrast, replacing the central pyridyl ring of terpy with
an effective σ-donor (e.g., a deprotonated phenyl ring) in the trans position
labilized the leaving group through the trans-influence/effect.
A terpy ligand substituted at any of its pyridyl rings with electron-
donating groups will reduce its π-acceptor properties, leading to reduced
rates of substitution of complexes. The opposite is true for electron-
withdrawing substituents. Considered together, it can thus be concluded
that square-planar Pt(II) metal complexes are activated differently by strong
π-acceptor ligands such as terpy or N^C^N when the primary and ancillary
substituents or the coordinated atoms are changed.
Adding an isoquinoline, a somehow poor π-acceptor as a replacing sub-
structure of the lateral pyridyl rings of terpy, has the same effect on the rate as
adding an effective σ-donor ring. In both cases, the rate of substitution is
reduced due to accumulation of electron density at the metal center. In con-
trast, substructures that enhance the π-acceptor properties of the ligand via
extended π-conjugation of molecular orbitals increase reactivity.
The tridentates, 2,6-bis(70 -azaindolyl)pyridine; 1,3-bis(70 -azaindolyl)
benzene; 2,6-bis(80 -quinolyl)pyridine; 1,3-bis(80 -quinolyl)benzene; and
bis(iminopyridyl/isoquinolyl)isoindolines, form two six-atom chelates at
the Pt(II) ion. Evidence of their strain-free coordination at the metal center
is provided by perfect square-planar bite angles by the ligand. However, this
ideal coordination geometry makes them kinetically inert toward substitu-
tion possibly due to their poor π-acceptor properties when compared to the
terpy-prototype ligands. The latter ligands form strained five-atom chelates
which also render the coligand in their complexes very labile.
Fusing a naphthyl/phenyl moiety on the hind side of the
bis(iminopyridyl/isoquinolyl)isoindoline nonleaving ligand makes back
donation of electron density from the Pt(II) d-orbitals into the LUMOs
more difficult, making these ligands intrinsically weaker π-acceptors.
A consequence is retarded rates of substitution of the leaving ligand.
Our results and their interpretations have amply demonstrated that the
π-acceptor properties and/or the σ-donor strength of N^C/N^N/C
ligands, the size of the chelates, and hence the bite angles at the metal ion
and the type of d8 {Pt(II)/Pd(II)} metal ion are all structural design features
which can be used to tune and optimize the reactivity of metal-based square-
planar complexes for possible maximization of the pharmacokinetics that
may be relevant to antitumor activity.
ACKNOWLEDGMENTS
The authors thank the National Research Foundation of South Africa (NRF) for funding.
D.J. thanks Prof. R. van Eldik for international collaboration. He further thanks his
former and current students, Dr. D. Reddy, Dr. P. Ongoma, Dr. A. Mambanda, Dr. I.
Wekesa, Dr. G. Kinunda, Dr. I. Shaira, Mrs. K-.L. Barry, Ms. P. Papo, Ms. S. Nkabinde,
and B. Khusi, for their contributions toward this work.
REFERENCES
1. Wong, E.; Giandomenico, C. M. Chem. Rev. 1999, 99, 2451–2466.
2. Reedijk, J. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 3611–3616.
3. Reedijk, J. Platin. Met. Rev. 2008, 52, 2–11.
4. Jansen, B. A. J.; Brouwer, J.; Reedijk, J. J. Inorg. Biochem. 2002, 89, 197–202.
5. Basolo, F.; Pearson, R. G. Mechanisms of Inorganic Reactions: A Study of Metal Complexes in
Solution, 2nd ed.; John Wiley & Sons Inc.: New York, 1967
6. Hubbard, C. D.; van Eldik, R. J. Coord. Chem. 2007, 60, 1–51.
7. Tobe, M. L.; Burgess, J. Inorganic Reaction Mechanisms. Longman: Harlow, Essex,
England; New York, 1999.
8. Bogojeski, J.; Petrovic, B.; Bugarčic, Ž. D.; Pablo, O., Eds.; Chronic Lymphocytic Leuke-
mia; InTechOpen: Europe, 2012, pp 339–366.
9. Brabec, V.; Kasparkova, J. Drug Resist. Updat. 2002, 5, 147–161.
10. Jung, Y.; Lippard, S. J. Chem. Rev. 2007, 107, 1387–1407.
11. Kukushkin, Y. N.; Ukraintsev, V. B. Zh. Neorg. Khim. 1972, 17, 2687–2689.
12. Mureinik, R. J.; Bidani, M. Inorg. Chim. Acta 1978, 29, 37–41.
13. Pitteri, B.; Marangoni, G.; Cattalini, L.; Bobbo, T. J. Chem. Soc. Dalton Trans. 1995,
3853–3859.
14. Jaganyi, D.; Hofmann, A.; van Eldik, R. Angew. Chem. Int. Ed. 2001, 40, 1680–1683.
15. Hofmann, A.; Jaganyi, D.; Munro, O. Q.; Liehr, G.; van Eldik, R. Inorg. Chem. 2003,
42, 1688–1700.
16. Czap, A.; Heinemann, F. W.; van Eldik, R. Inorg. Chem. 2004, 43, 7832–7843.
17. Pitteri, B.; Bortoluzzi, M. Polyhedron 2006, 25, 2698–2704.
18. Petrovic, B.; Bugarcic, Z. D.; Dees, A.; Ivanovic-Burmazovic, I.; Heinemann, F. W.;
Puchta, R.; Steinmann, S. N.; Corminboeuf, C.; van Eldik, R. Inorg. Chem. 2012, 51,
1516–1529.
19. Bugarcic, Z. D.; Bogojeski, J.; van Eldik, R. Coord. Chem. Rev. 2015, 292, 91–106.
20. Wilkins, R. G. Kinetics and Mechanism of Reactions of Transition Metal Complexes, 2nd ed.;
VCH: Weinheim; New York, 1991, p. 356.
21. Bugarčic, Ž. D.; Liehr, G.; van Eldik, R. J. Chem. Soc. Dalton Trans. 2002, 951.
22. Reddy, D.; Jaganyi, D. Transition Met. Chem. 2006, 31, 792–800.
23. Jaganyi, D.; Tiba, F.; Munro, O. Q.; Petrovic, B.; Bugarcic, Z. D. Dalton Trans. 2006,
2943–2949.
24. Grinberg, A. A.; Stetsenko, A. I.; Mitkinova, N. D.; Tikhonova, L. S. Zh. Neorg. Khim.
1971, 16, 264–266.
25. Reedijk, J. Chem. Commun. 1996, 801–806.
26. van Zutphen, S.; Reedijk, J. Coord. Chem. Rev. 2005, 249, 2845–2853.
27. Hubbard, C. D.; van Eldik, R. Inorg. Chim. Acta 2010, 363, 2357–2374.
28. Wilkins, R. G. Kinetics and Mechanism of Reactions of Transition Metal Complexes; VCH:
Weinheim; New York, 1991, p. 105.
29. Eryazici, I.; Moorefield, C. N.; Newkome, G. R. Chem. Rev. 2008, 108, 1834–1895.
30. Cummings, S. D. Coord. Chem. Rev. 2009, 253, 449–478.
31. Reddy, D.; Jaganyi, D. Dalton Trans. 2008, 6724–6731.
32. Becker, K.; Herold-Mende, C.; Park, J. J.; Lowe, G.; Schirmer, R. H. J. Med. Chem.
2001, 44, 2784–2792.
33. Ross, S. A.; Carr, C. A.; Briet, J. W.; Lowe, G. Anticancer Drug Des. 2000, 15, 431–439.
34. Lowe, G.; Droz, A. S.; Vilaivan, T.; Weaver, G. W.; Tweedale, L.; Pratt, J. M.;
Rock, P.; Yardley, V.; Croft, S. L. J. Med. Chem. 1999, 42, 999–1006.
35. McFadyen, W. D.; Wakelin, L. P. G.; Roos, I. A. G.; Leopold, V. A. J. Med. Chem.
1985, 28, 1113–1116.
36. Bligh, S. W. A.; Bashall, A.; Garrud, C.; McPartlin, M.; Wardle, N.; White, K.;
Padhye, S.; Barve, V.; Kundu, G. Dalton Trans. 2003, 184.
37. Bertrand, H.; Bombard, S.; Monchaud, D.; Talbot, E.; Guedin, A.; Mergny, J.-L.;
Grunert, R.; Bednarski, P. J.; Teulade-Fichou, M.-P. Org. Biomol. Chem. 2009, 7,
2864–2871.
38. Jennette, K. W.; Gill, J. T.; Sadownick, J. A.; Lippard, S. J. J. Am. Chem. Soc. 1976, 98,
6159–6168.
39. Hofmann, A.; Dahlenburg, L.; van Eldik, R. Inorg. Chem. 2003, 42, 6528–6538.
40. Jaganyi, D.; Reddy, D.; Gertenbach, J. A.; Hofmann, A.; van Eldik, R. Dalton Trans.
2004, 299–304.
41. Papo, T. R.; Jaganyi, D. J. Coord. Chem. 2015, 68, 794–807.
42. Jaganyi, D.; Boer, K.-L. D.; Gertenbach, J.; Perils, J. Int. J. Chem. Kinet. 2008, 40,
808–818.
43. Schm€ ulling, M.; Ryabov, A. D.; van Eldik, R. J. Chem. Soc. Dalton Trans. 1994,
1257–1263.
44. Schm€ ulling, M.; Grove, D. M.; van Koten, G.; van Eldik, R.; Veldman, N.; Spek, A. L.
45. Romeo, R.; Plutino, M. R.; Monsu Scolaro, L.; Stoccoro, S.; Minghetti, G. Inorg.
Chem. 2000, 39, 4749–4755.
46. Bailey, J. A.; Hill, M. G.; Marsh, R. E.; Miskowski, V. M.; Schaefer, W. P.; Gray, H. B.
Inorg. Chem. 1995, 34, 4591–4599.
47. Reddy, D.; Akerman, K. J.; Akerman, M. P.; Jaganyi, D. Transition Met. Chem. 2011, 36,
593–602.
48. Ongoma, P.; Jaganyi, D. Dalton Trans. 2012, 41, 10724–10730.
49. Shaira, A.; Reddy, D.; Jaganyi, D. Dalton Trans. 2013, 42, 8426–8436.
50. Lai, S.-W.; Chan, M. C. W.; Cheung, K.-K.; Che, C.-M. Inorg. Chem. 1999, 38,
4262–4267.
51. Papo, T. R.; Jaganyi, D. Transition Met. Chem. 2015, 40, 53–60.
52. McMillin, D. R.; Moore, J. J. Coord. Chem. Rev. 2002, 229, 113–121.
53. Field, J. S.; Haines, R. J.; McMillin, D. R.; Summerton, G. C. J. Chem. Soc. Dalton
Trans. 2002, 1369–1376.
54. Wekesa, I. M. Dissertation, University of KwaZulu-Natal, South Africa, 2014.
55. Wekesa, I. M.; Jaganyi, D. J. Coord. Chem. 2016, 69, 389–403.
56. Wekesa, I. M.; Jaganyi, D. Dalton Trans. 2014, 43, 2549–2558.
57. Garner, K. L.; Parkes, L. F.; Piper, J. D.; Williams, J. A. Inorg. Chem. 2010, 49, 476–487.
58. Hu, Y.-Z.; Wilson, M. H.; Zong, R.; Bonnefous, C.; McMillin, D. R.;
Thummel, R. P. Dalton Trans. 2005, 354–358.
59. Basolo, F.; Gray, H. B.; Pearson, R. G. J. Am. Chem. Soc. 1960, 82, 4200–4203.
€
60. Abrahamsson, M.; J€ager, M.; Osterman, T.; Eriksson, L.; Persson, P.; Becker, H.-C.;
Johansson, O.; Hammarstr€ om, L. J. Am. Chem. Soc. 2006, 128, 12616–12617.
61. Đurovic, M.; Bogojeski, J.; Petrovic, B.; Petrovic, D.; Heinemann, F. W.;
Bugarčic, Ž. D. Polyhedron 2012, 41, 70–76.
62. Mijatovic, A.; Bogojeski, J.; Petrovic, B.; Bugarcic, Z. D. Inorg. Chim. Acta 2012, 383,
300–304.
63. Selimovic, E.; Petrovic, B.; Canovic, D.; Bugarcic, Z. D.; Bogojeski, J. Aust. J. Chem.
2013, 66, 534–538.
64. Hanson, K.; Roskop, L.; Djurovich, P. I.; Zahariev, F.; Gordon, M. S.;
Thompson, M. E. J. Am. Chem. Soc. 2010, 132, 16247–16255.
65. Wen, H. M.; Wu, Y. H.; Fan, Y.; Zhang, L. Y.; Chen, C. N.; Chen, Z. N. Inorg. Chem.
2010, 49, 2210–2221.
66. Br€oring, M.; Kleeberg, C. Inorg. Chim. Acta 2007, 360, 3281–3286.
67. Martic, G.; Engle, J. T.; Ziegler, C. J. Inorg. Chem. Commun. 2011, 14, 1749–1752.
68. Yam, V. W.-W.; Tang, R. P.-L.; Wong, K. M.-C.; Cheung, K.-K. Organometallics
2001, 20, 4476–4482.
69. Zhang, H.; Zhang, B.; Li, Y.; Sun, W. Inorg. Chem. 2009, 48, 3617–3627.
70. Ongoma, P. O.; Jaganyi, D. Dalton Trans. 2013, 42, 2724–2734.
71. Kinunda, G.; Jaganyi, D. Transition Met. Chem. 2014, 39, 451–459.
72. Nkabinde, S. N. Dissertation, University of KwaZulu-Natal, South Africa, 2014.
73. Pitteri, B.; Bortoluzzi, M.; Marangoni, G. Transition Met. Chem. 2005, 30, 1008–1013.
74. Khusi, B. B.; Mambanda, A.; Jaganyi, D. Transition Met. Chem. 2016, 41, 191–203.
75. Khusi, B. B.; Mambanda, A.; Jaganyi, D. J. Coord. Chem. 2016, 69, 2121–2135.
CHAPTER SEVEN
Thionitrous Acid/Thionitrite and

Perthionitrite Intermediates in the
“Crosstalk” of NO and H2S
A. Olabe1
Juan P. Marcolongo, Ari Zeida, Leonardo D. Slep, Jose
Facultad de Ciencias Exactas y Naturales and INQUIMAE, Universidad de Buenos Aires/CONICET, Ciudad
Universitaria, Buenos Aires, Argentina
1
Corresponding author: e-mail address: olabe@qi.fcen.uba.ar
Contents
1. Introduction 278
2. S-Nitrosothiols, RSNOs, a Brief Overview on Structure and Reactivity 280
3. Thionitrous Acid HSNO and Thionitrite SNO, Elusive Aqueous Intermediates 282
4. Polysulfides and Sulfur Sols 287
5. Perthionitrite, S2NO. Identification of Iyellow 288
5.1 Available Results 288
5.2 Absorption Spectra Calculations 289
5.3 Consistency With X-ray Structural Data 291
5.4 Chemical Routes Following Transnitrosation of RSNO With H2S 292
6. Coordination Chemistry of Nitrosothiols, Thionitrous Acid, Thionitrite, and
Perthionitrite 295
6.1 Nitrosothiols 295
6.2 Thionitrous Acid/Thionitrite 298
6.3 Perthionitrite 298
6.4 The Gmelin Reaction, [Fe(CN)5(NO)]2 + HS 299
7. Conclusions 306
Acknowledgments 306
References 306
Abstract
The chemistry of aqueous NO and H2S as redox regulators of cellular and physiological
responses in cardiovascular, immune or neurological tissues has raised the question of
the overlapping pathophysiological functions often involving similar molecular targets.
The interactions of NO with H2S may functionally influence each other and focus has
been directed to new N/S hybrid species eventually determining signaling capabilities.
Besides the well-studied nitrosothiols, RSNOs, the eruption of H2S in the mechanistic
scene has stimulated increased interest in thionitrous acid, HSNO, and thionitrite, NOS,
as well as in perthionitrite (nitrosopersulfide), S2NO. We discuss the elusive chemistry of
the latter molecules as intermediates in selected reactions in aqueous solution, either as

278 Juan P. Marcolongo et al.
free species or as bound to iron metal centers. The coordination chemistry involves
mainly an updating on the “Gmelin” reaction proceeding upon mixing nitroprusside
[Fe(CN)5(NO)]2 and H2S, with controversial and still unsolved mechanistic issues related
to the onset of NO, HNO/N2O, polysulfides HSn (n ¼ 2–7), together with bound thio-
nitrous acid/thionitrite/perthionitrite and other intermediates and products.
1. INTRODUCTION
NO and H2S, two molecules frequently named “gasotransmitters”
(1,2), accomplish diverse biological functions associated with animal (3)
and plant (4,5) physiology: regulation of blood pressure, neurotransmission,
immune response, as well as plant defense responses, stomata closure, abiotic
stress, seed germination, etc. Both compounds are involved in aqueous
1-electron or multielectron redox chemistry with generation of species
whose structure-reactivity behavior needs to be elucidated for discerning
how the chemistry translates into a biological response. On the one hand,
NO is a radical molecule able to react either as an oxidant or reductant
(2,6); it is ubiquitously placed in the redox system comprising the eight-
electron interconversion between nitrates and ammonia, and affords a
versatile mechanistic chemistry covered by the activity of several enzymes:
NO—synthases, NO and NO2 —reductases, NH3—oxidases, NH2OH—
oxidoreductases, etc. (7) On the other hand H2S can only behave as a reduc-
tant; it may also produce up to 8-electron changes, with species in oxidation
states in the range 2 to +6, i.e., from sulfides to sulfates (8,9). The eventual
availability and redox reactivity of O2 may control the chemistry of each of the
intermediates, which can also be influenced by the disposal of metal-binding
coordination sites (2,10).
The NO signaling cascade has been increasingly well characterized
through the identification and chemical properties of distinct nitrosyl redox
states NO+, NO%, NO/HNO as intermediates in the oxidative or reduc-
tive cycles (2,6). Less understood are the biological–pharmacological effects
of sulfides; species other than H2S might be responsible for signaling, like
sulfane sulfur So, an uncharged species with six valence electrons having a
unique ability to attach reversibly to other sulfur atoms as in elemental sulfur

(S8), persulfides (RSSH), polysulfides (HSn , n ¼ 2–7) thiosulfate S2 O3 2 ,
and others (11–13).
There is a growing appreciation that both H2S and NO behave as mes-
sengers with connecting biochemistries (14–20). In this NO/H2S
N/S Intermediates in the “Crosstalk” of NO and H2S 279
“crosstalk” new biological mediators might be involved (21), sustaining our

focus on some early described N/S hybrid species, as advanced in Fig. 1.
One of them is thionitrous acid, described generically as HSNO (other iso-
mers have been described, see later) (22), together with its conjugated base
thionitrite SNO (23), and perthionitrite S2NO (23), the sulfur analog of
peroxynitrite O2NO. H2S plays a unique role in the generation and reac-
tivity of the latter intermediates in a significantly different way as performed
by thiols RSH, thus highlighting a hot topic in the emerging mechanistic
bioinorganic chemistry relevant to the modification of proteins by H2S
(24–27). In this context we deal with the chemistry of nitrosothiols, RSNOs
(28,29), with R ¼ glutathione (GSH), cysteine (cysSH), etc., and we discuss
the common and distinct chemical features compared to HSNO. We
expand into the coordination chemistry aspects by updating on the mech-
anistic details of the reaction of nitroprusside [Fe(CN)5(NO)]2 with H2S,
the “Gmelin” process, a fascinating sequence of reactions involving the
onset of the three relevant iron intermediates: [(NC)5Fe(NOSH)]3,
Fig. 1 Chemical structures of selected N/S hybrid species and S-nitrosothiols.

[(NC)5Fe(NOS)]4, and [(NC)5Fe(NOS2)]4, all showing a complex

mechanistic chemistry (30–34).
2. S-NITROSOTHIOLS, RSNOs, A BRIEF OVERVIEW ON

STRUCTURE AND REACTIVITY
S-nitrosothiols constitute a vast collection of compounds with prop-
erties dependent on the nature of the R group (28,29). Fig. 1 includes a
selected list with R ¼ alkyl, aryl and other substituted species with relevance
to biochemistry. RSNOs have been detected in vivo and are described
playing a role of NO carriers, sinks, and reservoirs, with potential medical
use as NO donors. Fig. 2 affords a structural description with two main res-
onance structures (I) and (II) (29,35). The dominant contribution
(I) comprises single and double S–N and N–O bonds, respectively.
A greater contribution of (II) can be achieved by increasing the electron–
donor abilities of R, influenced by the nature of substituents, or by
N-binding to transition metals. Both factors contribute strengthening the
S–N bond and weakening the N–O bond, with consequent effects on
the reactivity. A third possible resonance “ionic” structure (III) {RSNO+}
has not been included in Fig. 2 because of the minor relevance (it comprises a
nitrosonium cation with a N–O bond order of 3).
RSNOs can be prepared by reactions of thiols with oxidizing NO deriv-
atives such as NO2, N2O3, nitrites, and organic nitroso compounds. In par-
ticular cases, NO can react with thiols yielding disulfides RSSR and N2O
(28). RSNOs can be detected by using characteristic UV–vis, IR, and
NMR signatures. UV–vis absorptions comprise three bands: an intense
one at 225–261 nm (ε 104 M1 cm1), a second one at 340 nm
(ε 400–2000 M1 cm1), and a weak one at 550–600 nm (ε 60 M1
cm1). They have been assigned to allowed π ! π*, nO ! π*, and forbidden
nN ! π* transitions, respectively (29). The IR spectra exhibit two charac-
teristic peaks in the range of 1450–1530 and 610–685 cm1, both sensitive
to 15N substitution, attributed to νNO and νNS stretching, respectively. The
nitroso group in RSNOs shows N–O distances in the range of 117–120 pm,
Fig. 2 Resonance structures of S-nitrosothiols.

shorter than in NO (126 pm) and longer than in NO+ (106 pm) and NO%
(115 pm), supporting a bond order of 2 (28,29,35). Typical values for the
N-S distances are at 175 pm. The trends on the different properties of
RSNOs can be discussed under the structural framework described in
Figs. 1 and 2.
RSNOs undergo a variety of chemical reactions. A distinctive one is the
spontaneous thermal decomposition giving NO and RSSR. It has been
established that the half-life of different aqueous RSNOs varies from seconds
to hours, or even days. Most importantly, decomposition rates are currently
catalyzed by traces of metal ions, particularly by copper, a property that can
be best controlled by using chelating agents such as dipicolinic acid (dipic)
(28,29,36). For these reasons, structural correlations are difficult to establish,
a drawback that is reinforced by the influence of oxygen and light on the
decompositions (irradiation of RSNOs at 340 and 550–600 nm produces
NO% and RS% radicals) (29). The uncatalyzed aqueous decomposition reac-
tion has been proposed to be reversible, described by reaction (1):
RSNO.RS% + NO% aq (1)
The RS% radicals may combine forming RSSR (Eq. 2), and can also react
with thiolate RS to produce the very reactive RSSR% radical (Eq. 3), a
source of RSOO% in the presence of O2. These radicals can be detected by
spin-trapping techniques (37).
2RS% ! RSSR aq (2)

RS% + RS . RSSR% aq (3)
A high value of ΔG1 o ¼ + 110 kJ mol1 has been estimated for reac-
tion (1), for nitrosocysteine (38). Therefore, we would hardly expect the
uncatalyzed reaction (1) to proceed significantly to the right, unless a very
fast removal of products were onset.
Reactions of RSNOs with RS are biochemically important. They
comprise the 1,2-addition of RS at the N–O bond, followed by processes
resulting in oxidation of sulfur and NO reduction. The nature of interme-
diates and products depends on the reagent ratios; thus, N2O, NH2OH, and
NH3 are produced under moderate excess of thiolate, whereas NH3 is the
only N-containing product at a higher excess. Different mechanisms have
been proposed (39–41). RSNOs can also be reduced by alcohols, amines,
phosphines, etc. (29).
A very important reaction of RSNOs inside a cell or in a biological fluid

is the transnitrosation reaction (4):

RS + R0 S0 NO.RSNO + R0 S0 aq (4)
Reaction (4) comprises the reversible transfer of the S-nitroso functional
group from a thiolate to another, and involves the nucleophilic attack of
the thiolate anion on the nitroso nitrogen. It provides a route for the
S-nitrosation or S-denitrosation of proteins. The equilibrium position of
(3) depends on the forward- and reverse rate constants, whose values can
vary in the range of 0.1–500 M1 s1, according to the reactants. Trans-
nitrosations can be enzyme-catalyzed (viz., with thioredoxins), and in gen-
eral the kinetic/thermodynamics are important for describing the
S-nitrosations that occur when exposing cells to low-molecular-mass
RSNOs, a process for protein modifications without the involvement of
NO. The factors influencing S-nitrosation of proteins may be different from
those for other RSNOs (28).
3. THIONITROUS ACID HSNO AND THIONITRITE SNO2,

ELUSIVE AQUEOUS INTERMEDIATES
Thionitrous acid, HSNO, is frequently referred to as the “smallest”
and even the “simplest” nitrosothiol. The latter qualification seems ques-
tionable, given the availability of a mobile and ionizable H-atom in HSNO,
thus allowing for inherent specific reactivity (26,27,42). Back in 1952, four
compounds were proposed to behave in a rapid equilibrium, Eq. (5) (22):
HNSO .HOSN. HSNO. HONS (5)
The four isomers have been characterized by IR spectroscopy in an
argon matrix at 261°C (νNO, 1569 cm1 for the HSNO cis-isomer),
and were found to be light-sensitive and prone to polymerization. Looking
for the biological relevance, the isomerization reactions have been compu-
tationally explored using high-level coupled-cluster as well as density func-
tional theory (DFT) methodologies (43). Gas-phase calculations show that
the HONS tautomer and the Y-isomer SN(H)O are thermodynamically
feasible, with energy differences close to the one for HSNO by
25 kJ mol1. Notably, while the gas-phase isomerization barriers for
HSNO into HONS and SN(H)O become prohibitively high,
125–210 kJ mol1, the polar aqueous environment and water-assisted
proton shuttle decrease these barriers to 37 kJ mol1, making the latter
two isomers kinetically accessible under physiological conditions (43).
Very recently, cis- and trans-conformers of HSNO have been prepared
by mixing diluted H2S and NO that react over metallic surfaces at room
temperature (44). Identification has been achieved by Fourier-transform
microwave spectroscopy and quantum chemistry structural calculations,
yielding significantly long S–N distances for the cis- and trans-species,
183.4 and 185.2 pm, respectively (cf. with 175 pm for RSNOs). Although
changes might be expected upon hydration, the results are quite significant
for suggesting accessible homolytic/heterolytic paths for HSNO reactivity
upon biorelevant conditions, as discussed below.
There are no reports on the pKa of HSNO. A comparison with nitrous
acid HONO (3.25) allows predicting that HSNO should be more acidic,
thus converting to the anionic thionitrite form SNO in the biorelevant
aqueous solutions at pHs 7. A value of 2 has been very recently suggested
(45). By performing pulse radiolysis of anaerobic NO2 =H2 S mixtures, the
difference spectra allowed proposing a value of 340 nm for the maximum in
the UV–vis spectrum of HSNO, supported by a mass spectrometric (MS)
identification of a protonated species (24). This could be considered a ten-
tative value, casting some doubt on the putative coexistence of SNO, given
that the pH used was as high as 11. Recent modern computational work led
to calculated weak absorptions at 315 and 360 nm for HSNO in water
(Table 1) (49). A value at 315 nm was calculated for SNO in acetonitrile,
fairly close to the experimental value in the same solvent, 323 nm (23).
Care should be exercised when discussing the meaning of macroscopic
acidity constants of systems which can be protonated in nonequivalent sites.
Although each one of the conjugated weak acids has a microscopic acidity
constant (Kai), the system behaves macroscopically as if only one weak acid
in equilibrium with its conjugated weak base existed in solution. This is so
simply because the different Kai constants freeze the ratio between any pair
of protonated species (and eventually also for any two unprotonated ones)
making these values pH-independent. In such a system the apparent equi-
Xn
librium constant Kap is given by Kap 1
¼ Kai1 . In the specific situation
i¼1
where the dispersion of the individual Kai is high, the value of Kap is dom-
inated by the acid/base with the weakest microscopic Kai (50). In the present
case this would be NSOH, for which an estimation of pKa 10–11 appears
as reasonable.
Table 1 TD-DFT Electronic Spectral Calculations of N/S Hybrid and Related Species in
Water, Methanol, and Aprotic Solvents
λmax Calcd (nm)
Compound Solvent λmax Exp (nm) (Osc. Str., a.u.)
[S2NO] H2O 409 (46)–412 (26,27) 411 (0.03)a
Methanol 425 (23) 431 (0.025)a
Acetone 448b (24,25) 458 (0.03)a
Acetonitrile 450 (25) 458 (0.02)a
[O2NO] H2O 302 (47) 307 (0.03)
CH2Cl2 340 (47) 339 (0.04)
EtSNO H2O 330c (48) 280, 310, 330d
ON(SH)S Acetonitrile 358e (25) 368 (0.05)
e
HON(S)S Acetonitrile 358 (25) 351 (0.4)
HSNO Water 340f (24) 315, 360d
SNO Acetonitrile 323 (23) 315 (0.02)
a
Reported in Ref. (49), with the exception of the value for methanol.
b
Also measured at 448 nm in DMSO or DMF in Ref. (46).
c
Observed as a shoulder in the spectra of aqueous [FeII(CN)5(NOSR)]3 ions. Also measured at
330–350 nm with free thiols in organic solvents (29).
d
Low intensity absorption bands.
e
Assigned as a mixture of isomers ON(SH)S and HON(S)S.
f
Generated by pulse radiolysis of deoxygenated solutions, pH 11.
For the oscillator strengths, a damping factor γ ¼ 0.2 fs1 was used.
Low concentrations of sulfide were shown to quench NO-mediated vas-

cular responses through formation of an uncharged nitrosothiol, assumed to
be HSNO (14). An attempt to prepare and characterize HSNO in aqueous
solution at pH 7.4 has been reported by Filipovic and coworkers (24) by
studying the transnitrosation of nitrosoglutathione GSNO with H2S,
reaction (6).
GSNO + HS + H + ! HSNO + GSH (6)
The uncharged character of HSNO favors its ability for crossing mem-
branes and provides a new scenario for the modulation of the RSNO profile
in cells through the transnitrosation with cysteine residues in proteins,
reaction (7).
HSNO + P SH .PSNO + HS + H + (7)

For reaction (6), the mixture of reactants at pH 7.4 induced the decay of
the UV–vis main band of GSNO (λmax, 334 nm) with onset of a moderately
stable intermediate, Iyellow, with λmax at 412 nm (24). Fig. 3 shows a similar
picture evolving at pH 10 (51). Related reactions were also studied with
S-nitrosocysteine (CysNO), S-nitrosopenicillamine (SPEN), and S-nitroso-
N-acetylpenicillamine (SNAP) (26,51).
For GSNO, the UV–vis display (pH 7.4 or 10) did not match with the
stoichiometry of reaction (6), given that HSNO has been reported to absorb
at 330–340 nm, not at 412 nm (24,49). The band of HSNO might be hid-
den below the absorption of GSNO or it could rapidly decay in terms of the
reactivity of HSNO at room temperature. The MS positive evidence for
HSNO obtained with cryogenic experiments (24) has been complemented
by some UV–vis evidence appearing in the reaction of HS with SNAP
through the onset of a transient species with λmax at 320 nm, assigned
to SNO (26). Transnitrosation reactions such as in Eqs. (4) and (6) are
reversible processes, though under pH 7.4 conditions, the poor nucleophi-
licity of GSH could hardly ascribe a significant rate to the reverse reaction.
By measuring the decrease of [HS], a value of k6 ¼ 84 M1 s1 has been
reported (24).
The aqueous reactivity of HSNO has been under close scrutiny given its
potential ability to form a second generation of intermediates with putative
specific signaling roles (24–27). Closely related to reaction (1), reaction (8a)
implies the homolytic cleavage of the S–N bond with the production of NO
Fig. 3 Transnitrosation reaction of 103 M GSNO and HS, pH 10. Decay of GSNO at
334 nm and build-up of Iyellow at 412 nm. Adapted from Munro, A. P.; Williams, D. L. H.
J. Chem. Soc. Perkin 2000, 21, 1794–1797, with permission of The Royal Society of
Chemistry.
and reactive S% radicals (8b). Under the availability of HS, the formation
and reactivity of HS2 %2 radicals (8c) may produce NO and disulfides irre-
versibly under catalytic conditions (8d), given the strong reducing power of
HS2 %2 : (52)
HSNO. NO% + HS% ðaqÞ (8a)

% %
HS . S +H +
ðaqÞ pK8b ¼ 3:4 ðRef : 52Þ (8b)
% %2 1
S + HS .HS2 ðaqÞ K8c ¼ 9 2 10 M 3
ðRef : 52Þ: (8c)
%2 %
HS2 + HSNO ! NO + HS2 + HS ðaqÞ (8d)
Related to the influence of the bond dissociation energies (BDE) in

the rates of homolysis reactions (28,29), the comparatively long N–S bond
in HSNO is weaker than in RSNOs by 12 kJ mol1 (38), consistent
with the relative electron-withdrawing abilities of H and R. The BDE
for cis-HSNO has been calculated to be 123–127 kJ mol1 (43,44).
A second reactivity mode for HSNO has been proposed through reac-
tion (9), similarly as demonstrated for thiolates RS acting as nucleophiles
toward the S-atom in RSNOs (41). Hydrodisulfides are also direct products,
along with HNO. This reaction has also been proposed to occur in the
metal-catalyzed reactions of NO and H2S in the presence of excess of the
latter reagent (44).
HSNO + HS . HNO + HS2 (9)
Although reaction (9) has been described as endergonic (ΔG9 o ¼

+ 32 kJ mol1 ) (45), it might be plausible under conditions of fast products
removal. HNO (a precursor either for the fast generation of N2O or for sub-
sequent reduction) has been detected after mixing GSNO and H2S; this is a
remarkable fact (24), though other routes for HNO generation could onset,
namely the direct reaction of NO with HS (53), or the ensuing decompo-
sition of perthionitrite (25). A high reactivity for HS2 can also be antici-
pated, as seen later.
It should be noted at this point that neither of the above analyzed reac-
tions account for the absorption properties of Iyellow, the moderately stable
intermediate with λmax at 412 nm formed after mixing the GSNO/HS
reactants in the transnitrosation reaction (6), as can be seen in Fig. 3. The
analysis requires describing the fast reactivity of hydrodisulfides; one of these
reactions leads to polysulfides, as discussed later.
4. POLYSULFIDES AND SULFUR SOLS

Currently polysulfides can be produced through the partial oxidation
of sulfides. They have been reported to be formed after the onset of the
transnitrosation reactions such as in Eq. (6) through the subsequent dispro-
portionation reactivity of the HS2 intermediate (24). The build-up of poly-
sulfides can proceed sequentially under available oxidizing conditions by
adding a varying number of sulfur atoms to HS yielding soluble sulfane
chains, as described in reaction (10). The fast processes for HSn generation
can be followed by a slower one comprising separation as colloidal sulfur,
ending in S8 (36,54).
HS .HS2 .HS3 .… .HS9 ! S8 + HS (10)
Polysulfide formation currently arises after the exhaustive consumption

of the oxidizing substrate in reactions such as shown in Eq. (6), also observed
upon mixing peroxynitrite with HS– (55). Polysulfides are relatively stable at
pHs 7, but disproportionate under more acidic conditions, reaction (11).
HSðn + 1Þ ! n=8S8 + HS ðn ¼ 1 8Þ (11)
Colloidal sols (sulfur/polysulfide mixtures) may be formed at pHs 8,

usually in the second-time scale, depending on the medium, relative reactant
concentrations, and/or variable mixing modes that might determine a high
local concentration of a given reactant. By assuming a fast reactivity of dis-
ulfides, the colloidal sols have been proposed as responsible for the onset of
the 412 nm band assigned to Iyellow (24,25). The physical and chemical
properties of aqueous polysulfides are not clearly understood; their most
intense electronic absorptions are reported to occur at wavelengths
300 nm (56); much weaker bands have been reported for commercial
samples with maximum wavelengths <400 nm for different HSn species
(n ¼ 2–5). Our calculations with the quantum mechanics-molecular mechan-
ics (QM) (MM)/(DFT) methodologies are in agreement with the latter
observations.
Even though some disproportionation of HS2 is expected at pH 7.4, the
“in situ generation” of HS2 in a colloidal environment still allows a favor-
able competitive reactivity with other substrates (42), as will be analyzed
below seeking for a convincing explanation on the properties of Iyellow.
5. PERTHIONITRITE, S2NO2. IDENTIFICATION OF IYELLOW

5.1 Available Results
Munro and Williams already suggested in 2000 that Iyellow could be identified
as S2NO (51), supported by the X-ray structure obtained by Seel, Krebs, and
coworkers in 1985 of the [PNP][S2NO] salt (PNP+
¼Bis(triphenyl)phosphaniminium), isolated from an acetone solution after
mixing [PNP][NO2] with elementary sulfur or a PNP-polysulfide, under
anaerobic conditions (23). Although the solid was soluble only in aprotic sol-
vents with λmax at 450 nm, absorption transients at 409 nm appeared upon
mixing NaHS/Na2S2 with NO in aqueous solutions (46). Moreover, inter-
mediate values at 425 nm were measured in methanol or ethanol (46). The
picture suggests a solvatochromic behavior for the perthionitrite anion, as
recently demonstrated by QM-MM molecular dynamics (MD) simulations
combined with a real-time TD-DFT analysis (49), which we discuss later.
Filipovic and coworkers claimed providing negative spectroscopic evi-
dence (IR, 15N NMR, MS) for S2NO eventually arising subsequently
to the onset of reaction (6) (24). In fact, they reported a shift in the
IR-stretching frequency up to 1 min after mixing (1515 ! 1568 cm1),
presumably due to changes in νNO for the GSNO ! HSNO conversion;
we believe that the shift could be traced alternatively to GSNO ! S2NO,
assuming that perthionitrite is the actually observed species in the UV–vis
experiment on such a minute time scale (24). On the same grounds, the
15
N NMR signal at 322 ppm in water, traced to HSNO (24), could alter-
natively correspond to S2NO, consistent with the recently reported value
at 332 ppm, traced to its NBu4 + salt in tetrahydrofuran (THF) (57). Further
ambiguity arises if we consider that rapidly interconverting HSNO/S2NO
mixtures could exist in the solution (see later) giving a 15N NMR signal
corresponding to an averaged feature. In a crucial experiment, Feelisch
and coworkers obtained positive electrospray ionization (ESI)-
high-resolution mass spectrometry (HRMS) signals derived from aqueous
SNAP/HS mixtures at pH 7.4, consistently assigned to S2NO (27).
A very recent report states that S2NO, generated by mixing Na2S and
GSNO at a 2:1 ratio in buffered solution at pH 7.4, can be handled in anaer-
obic medium in a controlled manner, showing a slow decay of the band at
412 nm, with 80% absorbance remaining in 3 h. The decay rate was
enhanced in the presence of dioxygen (58). Finally, Grman, Jacob, and
coworkers (20) showed very recently that the reaction of GSNO with
organic polysulfanes RSxR0 (x ¼ 14), at pH 7.4 and in the presence of

cysSH or GSH, led to the build-up of Iyellow at 412 nm, identified as
S2NO. Its formation and decomposition occurred in 15–40 min under
the studied conditions, with release of NO. Mixtures of GSNO/poly-
sulfanes in the absence of cysSH or GSH were not reactive, i.e., the addition
of either of the latter reagents was required in order to generate HS,
suggesting that the reaction evolves similarly as in the transnitrosation pro-
cess initiated by reaction (6).
A claim was also raised on the intrinsic instability of S2NO in water
toward the rapid formation of HNO and sulfur, either in acidic or neutral
conditions (25). By using a freshly prepared solution of [PNP][S2NO] in
acetone, followed by dissolution either in a 10%–water/acetonitrile mixture
or in a sulfide-containing aqueous solution at pH 7.4, the authors reported
the formation of a species with λmax at 420 nm in a few seconds, which
survived for minutes along with a slow decay, yielding an intermediate at
340 nm reported as HSNO. The 420-nm peak emerged over the broad
tail of a very intense UV-band centered at 300 nm in a highly scattering
medium, with the consequent assignment of the 420 nm absorption to a col-
loidal sulfur sol. In fact, we believe that the relevant S2NO species has been
missed through a wrong interpretation, as suggested later by our computa-
tional work on the UV–vis spectra (49).
5.2 Absorption Spectra Calculations

The absorption spectra for S2NO and related species were calculated by
averaging an ensemble of instantaneous configurations sampled through
QM-MM MD simulations. Spectral line shapes were obtained by comple-
mentary real-time TD-DFT calculations, a methodology that has proved
reliable to predict the absorption properties of molecules and ions in solu-
tions and complex environments (49). Fig. 4 reinforces the assignment of the
species absorbing at 409–412 nm in water to S2NO–. Most importantly, the
experimentally observed bathochromic shifts when proceeding from water
to other less-protic solvents is faithfully reproduced by the computational
analysis. The main character of the UV–vis electronic transition is
p(S) ! π*(SNO). Our calculations show that the absorption maximum
for this molecule is strongly dependent on the geometrical parameters
explored across the MD. Thus, the specific interactions with the solvent
become extremely important to describe the spectroscopic behavior in
solution.
Fig. 4 Calculated electronic spectra for S2NO in water (black), methanol (blue), acetone
(red), and acetonitrile (green), using TD-DFT and QM-MM molecular dynamics
simulations.
We have placed the computational methodology under a control system

by performing calculations with other well characterized similar species, as
detailed in Table 1, and excellent agreement with experimentally obtained
UV–vis spectra have been found.
Most remarkably, it can be seen that the results with O2NO also
account for the experimentally observed solvatochromism (47).
Solvatochromic effects have been observed for coordination compounds
interacting with acceptor solvents. The big UV–vis and IR spectral shifts
could not be accounted for by a mere dielectric effect as produced by a con-
tinuum solvent model, and empirical donor–acceptor correlations were
employed with some success (59). Consistently, a recently reported lack
of correlation emerged by plotting the values of λmax for S2NO in different
solvents against the corresponding dielectric constants (60). However, in
that report the authors highlighted that water did not fit in a linear correla-
tion displayed by the alcohols, and interpreted that deviation as evidence for
rejecting the 412 nm maximum as characteristic of S2NO (24,25,60). For
all the used solvents, our results (Fig. 4, including alcohols) show that the
specific interactions arise both from solvent-induced structural changes
and from electrostatic solute–solvent effects, the former being dominant.
It becomes clear that the strong H-bonding interactions in water differ from
those arising in alcohols.
In a very recent work dealing with the NO/H2S crosstalk, Pluth and
coworkers reported on the reactions of diverse isolated persulfides RSSH
with NO2 in THF to produce NO via polysulfide and perthionitrite inter-

mediates (57). S2NO was identified through a strong band at 446 nm, a
feature also generated by independently mixing NO2 with S8. To the coin-
cidence of λmax in THF with the values displayed in Table 1 for the aprotic
solvents must be added the shift toward 420 nm when using mixtures of
1:1 THF/H2O. The 15N NMR signal at 332 ppm in THF, assigned to
S2 15 NO , appears to be related to the significant solvent structural influ-
ence, and suggests the best assignment of the aqueous 322 ppm signal to
S2 15 NO rather than to HS15NO, as analyzed earlier (24).
5.3 Consistency With X-ray Structural Data

Table 2 shows a comparative picture of new X-ray diffraction results (25)
(which are essentially the same as originally reported) (23), with data for
the calculated species in solution (49), presently expanded by including
methanol as solvent.
A fair agreement with the reported cis-structure for the anion can be
observed, with a trend to greater distance values in the calculated solution
spectra that can reasonably be traced to packing and environmental effects.
Table 2 Selected Distances and Angles With Standard Deviation for Solid [PNP][S2NO]
(PNP+ ¼ Bis(Triphenyl)Phosphaniminium) (25), and for the Anionic Species in Water,
Methanol, Acetone, and Acetonitrile, Calculated Through Molecular Dynamics (MD)a
MD MD MD MD
Parameter DRX (25) (H2O) (MeOH) (Me2CO) (MeCN)
d(N1-O1) (Å) 1.25 1.24 1.24 1.24 1.24
(0.01) (0.03) (0.02) (0.02) (0.04)
d(N1-S1) (Å) 1.70 1.79 1.78 1.80 1.80
(0.01) (0.01) (0.06) (0.07) (0.08)
d(S1-S2) (Å) 1.97 2.07 2.04 2.03 2.04
(0.01) (0.06) (0.04) (0.04) (0.04)
θ(O1-N1-S1) (degree) 117.8 119 119 120 120
(0.2) (5) (4) (4) (4)
θ(N1-S1-S2) (degree) 115.1 111 113 117 117
(0.2) (5) (5) (6) (5)
−
N1 S1
O1 S2
a
Reported in Ref. (49), with the exception of the value for methanol.
No significant changes can be observed by comparing the geometrical fea-

tures of S2 15 NO in the aprotic solvents, though subtle trends appear with
data for methanol and water. The onset of hydrogen bonds between the
negatively charged terminal sulfur S2 (q(S2) –0.7) and NO (q(NO)
–0.3) fragments with the solvent determine a much lower (N1-S1-S2)
angle in water with respect to acetone and acetonitrile, with an intermediate
value pertaining for methanol. This accounts for the observed spectral band
shift to higher energies observed in water. Interestingly, this description is
also in agreement with the reported small bathochromic shift when chang-
ing to alkaline solutions (27), implying a weaker H-bonding array.
5.4 Chemical Routes Following Transnitrosation of RSNO

With H2S
A third possible route for the decay of HSNO has been proposed, reac-
tion (12) (49), on the basis of the early generation of hydrodisulfides in
the medium. It is similar to a transnitrosation reaction, specifically a trans-
nitrosopersulfidation one (5).
HSNO + HS2 .S2 NO + HS + H + (12)
The interchange between the nucleophiles hydrodisulfide and
hydrosulfide at the NO group can be described as a So (sulfane)-atom trans-
fer. It appears as a kinetically favored reactivity mode of HSNO, compared
to reaction (9). We expect a greater value for the specific rate constant k12
than for k9, on the basis of a greater nucleophilic ability of the more polar-
izable HS2 over HS (see later for comparative measurements with the
bound species). This is in agreement with predicted and observed trends
for persulfides and thiolates (61). Additional routes to S2NO generation
can be imagined, however, namely the attack of HS2 on the still
unconverted GSNO, and the reaction of NO with the putatively formed
perthiyl radical, S2 % , reactions (13)–(15).
GSNO + HS2 . S2 NO + GSH (13)
S% + HS2 . HS + S2 % (14)
S2 % + NO% . S2 NO (15)
Reaction (13) implies the attack of HS2 toward the initial, still uncon-
sumed GSNO reactant. Although HSNO can be predicted to be more elec-
trophilic than GSNO, the effective variable concentrations of GSNO and
HSNO might lead to reactions (12) and (13) proceeding with similar rates.
On the other hand, although reaction (14) has not been reported, it could
proceed faster than (8c), given the expected favorable nucleophilicity of
HS2 over HS toward the S% radical. In fact, a thermochemical estima-
tion allows obtaining ΔGo 14 ¼ 62 kJ mol1 (45). Perthiyl radicals S2 %
have been characterized and described as more stable species than S%
(62). They have been proposed to participate both in the generation and
in the homolytic decomposition of S2NO during the H2S-transnitrosation
processes, through reaction (15) (27).
The above presented complex picture has been tested through simula-
tion procedures, and the details on the assumed reactions and constants
are analyzed as follows. Fig. 5 shows the traces for the decay of the initial
reactants and the build-up of different intermediates in a restricted time win-
dow, for 0.5 mM GSNO and H2S, mimicking the conditions used in Ref.
(24). We considered reaction (6) as irreversible, given the negligible nucle-
ophilic reactivity of GSH, with k6 ¼ 84 M1 s1 (24). For the equilibrium
reaction (8a), we used k8a ¼ 0.12 s1 (as measured by the [NO] build-up)
(24), and k8a 107 M1 s1 (an estimated value for a rapid radical recom-
bination). The required data for reactions (8b) and (8c) were taken from Ref.
(52). The value of K9 0.05 was estimated by assuming values for
k9 ¼ 500 M1 s1 and k9 ¼ 104 M1 s1. K9 was considered lower than
K12 (K12 100–1000, with k12 ¼ 105 M1 s1 and k12 ¼ 103 M1 s1 at
pH 7) (49), consistently with the smaller reactivity of thiolates in comparison
with persulfides (60). The consumption of HNO was traced either to reac-
tion (16) yielding N2O (63) or to reaction (17) giving HSNHOH (64) with
further timely production of NH2OH. Rate constants for the reactions of
0.5
GSH
0.4
Concentration (mM)
0.3
0.2
SSNO–
NO•
0.1 HSNHOH
HSNO
GSNO
HS–
0.0
0 20 40 60 80 100 120
Time (s)
Fig. 5 Simulated traces for selected species when mixing 0.5 mM GSNO and H2S, as
described in the text.
HS2 with different substrates were estimated as being 100/1000-fold

greater than for corresponding reactions with HS.
2HNO ! N2 O + H2 O k16 ¼ 8 106 M1 s1 ðRef : 63Þ: (16)
HS + HNO ! HS NHOH k17 ¼ 1:2 106 M1 s1 ðRef : 64Þ: (17)
The simulation results for the concentration profiles show a consistent

decay of [GSNO] (down to 90%), a consistent similar increase of [GSH],
and a 100% decay of [HS]. The latter feature indicates that the initial trans-
nitrosation step has not been completed under the proposed regime of reac-
tions and that HS displays additional reactivity with the intermediates in
order to attain a full consumption. The build-up of S2NO, NO, and
HSNHOH is onset after a short induction period and attains saturation in
30 s, while [HSNO] grows exponentially up to a maximum for 15 s
and then decays slowly. The duration of the induction period (a few sec-
onds) is reasonably accounted for in the allowed time window, if compared
with the value estimated by us from the reported absorbance traces (24). The
attained concentration of S2NO in the simulations is consistent with the
available experimental data (24), revealing a yield of 30% (based on the
reported value of ε ¼ 3125 M1 cm1 in acetone (25), which we presume
similar to the one in water). It becomes evident that the formation of
S2NO competes with the generation of other nitrogenated products
derived from the HSNO reactivity; importantly, the sum of
N-containing products (S2NO, HSNO, NO, HSNHOH) in the proposed
time window of Fig. 5 is well close to 100% with respect to initial GSNO.
On the other hand, [HNO], ½HS2 could not be established, suggesting a
negligible concentration in the steady state. Neither N2O was found in the
simulations, presumably because dimerization of HNO is less favorable that
its reactivity toward further reduction by HS (in fact, N2O was not
detected experimentally under the used equimolar concentrations of reac-
tants, it was only observed for increasing sulfide concentrations) (24).
Indeed, variations in the simulations output could be obtained by further
manipulation of the unreported rate and equilibrium constants, or consid-
ering alternative routes for NO/HNO decay (such as the possible interme-
diacy of hyponitrite radicals, HN2 O2 % , as precursors of competitive N2O/
NH2OH generation) (65). Noticeable is the previously unreported induc-
tion period (24), which is in contrast with the observations reported for the
reactions of peroxynitrite (55) and SNAP (26) with HS.
The onset of reaction (12) makes the GSNO consumption through reac-
tion (6) autocatalytic with respect to HS2 . It also provides an explanation
for the net GSNO ! S2NO conversion (334 nm ! 412 nm) and the
absence of specific UV–vis spectral features for HSNO (24). Autocatalysis
is frequently associated with the build-up of induction periods for the gen-
eration of product, showing S-shaped traces (66). In the underlying condi-
tions, a requirement of sulfur radicals (reactions 8a–8d) seems crucial for the
generation of the more reactive HS2 . The induction times have been
reported to increase with [O2] and to decrease with [HS] (29), and this fact
relates to the observed early consumption of O2 (26), a trapping agent for
S% and HS2 %2 (37). Overall, these autocatalytic features constitute favor-
able and specific evidence supporting the proposed mechanism highlighting
the role of disulfides in biochemistry.
As a conclusion on the validity of the simulation procedures, we should
remark that only a limited set of experimental conditions has been consid-
ered, namely the equimolar GSNO/HS situation (24,51). Indeed, a wider
experimental approach is needed for testing the influence of increased
[HS]/[RSNO], changing the type of used RSNOs, as well as better dis-
closing the stoichiometric results under exhaustive conditions (viz., the
yields of N2O, NH2OH, or eventually NH3 and the conditions for the onset
of polysulfides) (24).
6. COORDINATION CHEMISTRY OF NITROSOTHIOLS,

THIONITROUS ACID, THIONITRITE, AND
PERTHIONITRITE
6.1 Nitrosothiols
RSNOs are potential ligands with at least three donor atoms, N, S, and O.
Only N- and S-bonded coordination modes have been characterized.
Substituted S-nitrosothiols may contain other coordination sites as –OH,
–NH2, –COOH, and others (29).
Coordination to metal centers may change considerably the stability of
RSNOs, depending on the nature of the metal and the ligand structure.
The electron-rich metal ions Hg(II), Cu(II), Cu(I), Ag(I) prefer binding
by the S-donor atom, thus destabilizing RSNOs. On the other hand,
Fe(II) complexes usually lead to N-bonded RSNOs; they can be formed
by ligand exchange, and usually lead to more stable complexes toward
N–S bond rupture than the free RSNOs. We exemplify this with two
biorelevant cases, iron(II) dinitrosyls (DNIC) and cyano(L)ferrates(II),

reactions (18) and (19) (29):
FeII L2 aq + 2RSNO ! FeðL2 ÞðNOSRÞ2 ! FeII L2 ðNOÞ2

(18)
+RS% + RS ðL ¼ cysteine, histidine, etc:Þ
II 3 3
Fe ðCNÞ5 ðLÞ + RSNO . FeðCNÞ5 ðNOSRÞ (19)
+LðL ¼ H2 O, NH3 , etc:Þ
For the cyano complexes, the iron-bound species have a variable stabil-
ity, depending on R, and may decompose through dissociation (reverse of
reaction 19) or through redox processes (homolysis at the N–S bond) as
described by reaction (20) (67).
3 3
FeðCNÞ5 ðNOSRÞ . FeðCNÞ5 ðNO% Þ + RS% (20)
The lifetimes and yields of the decomposition products are very variable,
depending on the thiol structure, namely the inductive effects of the func-
tional groups in R. Thus the complex with cysteine has a t½ of 140 s, which
can be increased by acetylation of the NH2 substituent (t½, 450 s), whereas it
is shortened by esterification of the COOH group (t½, 70 s). The
mercaptosuccinic acid complex is remarkably stable (t½ > 3.6 h) (29).
The UV–vis spectra of the [FeII(CN)5(NOSR)]3 complexes are almost
independent of the nature of the thiol, with two bands at 525 and
320 nm, with the former three to five times stronger (6769). These
complexes have been also characterized by IR spectroscopy, with typical
νNO stretching modes at 1380 cm1 (NOSEt) (69) or 1390 cm1 (NO-
mercaptosuccinate) (70). These values are significantly smaller that those
observed with free nitrosothiols (viz., 1505 cm1 for mercaptosuccinic acid)
(67), reflecting the σ–π bonding interactions with the metal. For the
mercaptosuccinic complex, a value of 758 cm1 has been reported for
νNS, reflecting a much stronger N–S bond than in the free thiol
(νNS < 700 cm1) and consistent with the greater thermal stability in the
bound situation (70). NMR characterizations have been also published, with
δ at 607 ppm (for 15N) and at 1035 ppm (for 17O) (33,70).
Other metal centers (Ru, Rh, Ir, Os) with octaethyl- and
tetraphenylporphyrin coligands have been used to model the interactions
of RSNOs in biologically relevant iron–heme complexes (29). A valuable
mechanistic study has been provided for the reaction of RSNO:
N-acetyl-1-amino-2-methylpropyl-2-thionitrite, with a model meta-
lloporphyrin, RuII(OEP)(CO) (OEP ¼ octaethylporphyrinato dianion) in
dry toluene or benzene (71). The addition product trans-RuII(OEP)(NO)

SR has been fully characterized by X-ray and spectroscopic methods and
is stable only as a solid, decomposing under air, moisture, and light. The
mechanism involves the initial rapid equilibrium formation of an
S-bound RuII(OEP)(RSNO)(CO) intermediate, which leads to a short-
lived second species in a rate-limiting step, RuIII(OEP)(SR)(CO), presum-
ably through S–NO bond cleavage.
An unusually stable iridium complex was obtained through the reaction
of K[IrCl5(NO)] with PhCH2SH in acetonitrile solution. The trans-K[Ir
(CH3CN)N(O)SCH2Ph] salt was obtained by recrystallization from
CH3CN, and an X-ray structure could be solved (72). The complex was
characterized by UV–vis and 1H NMR in CH3CN and by FTIR spectra
in the solid state, with νNO at 1443 cm1 and νNS at 778 cm1, both sen-
sitive to 15N labeling. Note the upshift of νNO compared to the Fe(II) com-
plexes described earlier, reflecting the smaller back-bonding contribution of
Ir(III) vs Fe(II) toward the NOSR ligand. This pioneering structural work
was complemented by a comprehensive structural and spectroscopic study
with a family of stable and water-soluble IrIII–NOSR complexes, including
DFT calculations and an estimation of comparative reactivity related to the
N–S bond rupture (73).
An alternative method for preparing bound-NOSR complexes consists
of using the thiolates as nucleophiles toward the N-atom in bound nitrosyl
complexes, such as nitroprusside. Addition reactions of diverse species such
as OH, NH3, N3 , N2H4, NH2OH, and SO3 2 , on the formally consid-
ered NO+-ligand in nitroprusside have been widely studied mechanistically
(10), and this is also the case for the reactions with HS and several aliphatic
thiolates, including those derived from mercaptosuccinic acid, cysteine, and
glutathione, reaction (21):
2 3
FeðCNÞ5 ðNOÞ + RS . FeðCNÞ5 ðNOSRÞ K21 , k21 , k21
(21)
The fast kinetics for the formation/dissociation of adducts formed with

different RSH thiols have been measured using temperature-jump/stopped-
flow methods by Johnson and Wilkins (74). The second-order formation
rate constants were found to be pH dependent, showing that only the RS
species are reactive. The value of k21 varies little with the thiol type (3 103–
4 104 M1 s1) at 25°C, with similar 4H6¼ values, 33 kJ mol1. There is
a much larger variation in the adduct dissociation rate constants k21
(from 12 to 3 103 s1). The establishment of the equilibrium reaction (21)

is characterized by a marked color increase, followed by its fading in a slower
time scale described initially by reaction (20) and subsequent decomposi-
tions. It should be remarked that the equilibrium constants K21 have only
moderate values, and incomplete conversions of nitroprusside into the
[Fe(CN)5(NOSR)]3 complexes are usually achieved. Reaction (21) is
markedly accelerated in the presence of excess thiolate, leading to [Fe
(CN)5(NO%)]3 and disulfides, RSSR (67). The reaction in Eq. (21) is also
influenced by the presence of oxygen (67,75).
6.2 Thionitrous Acid/Thionitrite

A biorelevant case of HSNO generation and reactivity has been reported by
studying the reaction of NO2 with H2S at pH 7.4, catalyzed by a FeII–por-
phyrin model complex. The Fe–NOSH intermediate could be character-
ized through cryospray, time-of-flight (TOF) ESI-MS detection. HSNO
was proposed to be released from the iron center enabling the nitrosation
of added bovine serum albumin. It was also presumed as the source of
HNO generation. No evidence of S2NO has been shown in this study
(76). Strong evidence for the coordination of HSNO and NOS by the
[FeII(CN)5]3 moiety is presented in Section 6.4.
6.3 Perthionitrite
Direct reaction between freshly generated aqueous S2NO and hemoglobin
(Hb) centers, also extended to myoglobin (Mb), has been very recently
established (58). By using deoxyHbII and/or deoxygenated methemoglobin,
HbIII, the addition of free S2NO rapidly produced nitrosyl hemoglobin,
HbIINO, with additional formation of polysulfides HSx , or HS2, respec-
tively. The studies were carried out using time resolved EPR and UV–vis
methods, and also showed that heme-species without a vacant site, like
HbIICO or HbIIO2, did not produce bound NO, suggesting the necessary
previous coordination of S2NO to the metal center. The latter event has
not been demonstrated however, and further mechanistic studies are in
order. It is worth pointing out that the authors used the GSNO/H2S reac-
tion to generate free S2NO with a subsequent degassing, thus assuring the
elimination of the previously produced NO, formed through the reactivity
of HSNO (see Section 5.4). On the other hand, remarkable evidence can be
anticipated on the coordination of S2NO to [FeII(CN)5]3, as shown in
Section 6.4.
6.4 The Gmelin Reaction, [Fe(CN)5(NO)]22 + HS2

This 170-year-old reaction (77) is quite relevant for comparing the reactivity
of HS vs RS as nucleophiles toward a common electrophile, com-
plementing the analysis previously carried out for the transnitrosation reac-
tions (49). The Gmelin reaction is a complex process comprising the
generation of an intense red–purple color upon mixing the reactants (λmax,
535 nm, pH 12–13); the color develops in a less than or in a few seconds
(depending on [HS]), followed by color fading in a slower time scale
(30). In this respect, the chemistry appears similar to that discussed in
Section 6.1. The stoichiometric and mechanistic Gmelin features are
strongly dependent on the pH and on O2 availability (34). S8 is the exclusive
final product of hydrosulfide oxidation (78), not RSSR as found for RSNOs
(67,75). NH3 and N2O are the reduction products of bound nitrosyl: NH3 is
produced at all pH values but N2O only at pH > 12. The
hexacyanoferrate(II) anion, aqueous Fe(II,III), and/or Prussian-blue-type
precipitates appear as the main final sinks for iron and cyanides (30). The
overall process comprises the build-up and decay of multiple intermediates,
with particular controversial issues arising in the identification of the freshly
formed I535 (the “red” species) and its prompt successor, I575, the “blue” spe-
cies formed in seconds in a pH-dependent way. Alternative views on their
characterization and reactivity have been raised in the last 5 years by us (30)
and others (31–33). We are currently revisiting this reaction (34), and we
advance some new results here. We will summarize and present the infor-
mation from the available interpretations, while aiming to present a
balanced view.
In our previous work in 2011 (30), we showed that the absorbance max-
ima and intensities evolving just after mixing the reactants depended on the
molar ratio R ¼ [HS]/[Fe] and the pH. The values of λmax for the emerging
absorptions varied in the range of 535–570 nm in the pH range 8.5–12.5, in
anaerobic conditions. Fig. 6 shows some typical absorbance-time profiles for
R ¼ 30.
The traces exhibited a characteristic biexponential form, with a fast
increase of absorbance and a slower decrease. Lower conversions of
nitroprusside into the colored adduct were achieved at the lower pH values.
At constant pH, the following rate law has been established in the pH range
studied by means of the initial rate method (i.e., by using the rising part of
the curves): v ¼ d[Ad]/dt ¼ kad [HS][NP], with Ad representing the colored
adduct/s. The following equilibrium reactions (22–24), include the initial
Fig. 6 Absorbance at 535 nm against time for the reaction of 0.05 mM [Fe(CN)5(NO)]2
and HS, R ¼ 30. pH increases upward: 8.7, 9.5, 10.5, and 12.3. Adapted from
Quiroga, S. L.; Almaraz, A. E.; Amorebieta, V. T.; Perissinotti, L. L.; Olabe, J. A. Chem. Eur.
J. 2011, 17, 4145–4156. with permission of Wiley.
addition step and the fast isomerization and deprotonation equilibria of the
isomers:
2 3
FeðCNÞ5 ðNOÞ + HS . FeðCNÞ5 ðNOSHÞ (22)
3 3
FeðCNÞ5 ðNOSHÞ . FeðCNÞ5 ðNSOHÞ (23)
3 3 4
FeðCNÞ5 ðNOSHÞ = FeðCNÞ5 ðNSOHÞ . FeðCNÞ5 ðNOSÞ + H+
(24)
By considering a mass-balance equation for the iron-containing

species, we obtained the following parameters by fitting procedures:
k22 ¼ 190 M1 s1 and k22 ¼ 0.30 s1 (30). Finding that the addition of
HS was a reversible process, in a similar way as occurs for thiolates (74),
had not been reported previously. Not unexpectedly, k22 was significantly
smaller than k21, given the lower nucleophilicity and polarizability of HS
compared with RS. Remarkably, an unidentified intermediate with a
value of pKa of 10.5 could be discerned from the kinetic analysis (30).
We have elaborated in Section 3 on the intermediacy of NOSH/
NSOH/HNSO isomers interreacting with such a possible value of pKa.
The same would be the case here, with a mixture of rapidly reacting
bound-NOSH species as described in reactions (23) and (24).
Our analysis implied the consideration that the rising absorbance

traces only comprised the build-up of [Fe(CN)5(NOSH)]3 and [Fe(CN)5
(NOS)]4. The [Fe(CN)5(NOSH)]3 ion, predicted to be unstable by
our DFT computations (30), was tentatively (and incorrectly) assigned to
the “blue” 570 nm species. At that time, we were unable to identify the
bound [S2NO] ion, eventually being formed a few seconds after mixing.
In 2012, Filipovic and Ivanovic-Burmazovic questioned our assignments on
I535 ([Fe(CN)5(NOS)]4) and I575 ([Fe(CN)5(NOSH)]3) by working at
pH 7.4 (R ¼ 34) under anaerobic conditions (31). They reported an initial
absorption band at 535 nm (accompanied by a weaker one at 330 nm), also
extensive to all pHs. I535,330 subsequently decayed in a few seconds to a more
stable intermediate, I570. The slowness of the I535 ! I570 change precluded
assigning it to an acid–base transformation. The authors assigned I535 to
[Fe(CN)5(NOSH)]3, though I570 remained unidentified (31). In 2013
the same authors elaborated a detailed mechanism (32) by working in
oxygenated conditions at pH 7.4 and R ¼ 125. I570 was traced to
[FeII(CN)4(NCS)(H2O)]3, an ill-characterized intermediate, formed after
the release of HNO from the iron center. New colored intermediates
appeared that had not been observed under anaerobic conditions.
A detailed evaluation of that proposal (32) is not offered herein, but may
follow in time (34).
We show in Fig. 7A our results at pH 8.9, with an absorbance-rise com-
prising the transformation I535 ! I575; Fig. 7B shows the subsequent decay at
575 nm. A similar picture evolves in Fig. 8A and B, with formation of I575 in
the decay region (a greater value was achieved earlier by I535 owing to the
Fig. 7 (A) Successive UV–vis spectra after mixing 0.1 mM [Fe(CN)5(NO)]2 and HS, pH
8.9, R ¼ 30. (B) Successive UV–vis spectra, same conditions as in (A), 20 s after mixing.
8.9, R ¼ 100. (B) Successive UV–vis spectra, same conditions as in (A), 1.5 s after mixing.
12, R ¼ 100. (B) Successive UV–vis spectra, same conditions as in (A), 6 s after mixing.
faster second-order build-up of [Fe(CN)5(NOSH)]3). We briefly highlight

the onset of an early absorption at 347 nm, assigned to [Fe(CN)5(NO%)]3
(69). In sharp contrast, Fig. 9 at pH 12 shows the formation of only I535,
without any feature at 575 nm. Under these conditions, I535 was moderately
stable (t½ 60 s) and we assigned it to [Fe(CN)5(NOS)]4, quantitatively
generated according to reactions (18–20), under irreversible conditions
for reaction (20). These observations allowed an estimate of ε ¼ 6000 M1
cm1 for the thionitrite-bound complex, in agreement with results of DFT
calculations (30). The absorption features of [Fe(CN)5(NOS)]4 are consis-
tent with those for the [Fe(CN)5(NOSR)]3 complexes (68,74).
Another development in the mechanistic discussion was provided by Wu

and coworkers in 2015 (33). They used 1.2 mM [Fe(CN)5(NO)]2, a slight
excess of HS (R ¼ 3), and aerobic conditions. At pH 12, the initially formed
band with λmax at 542 nm was traced to [Fe(CN)5(NOS)]4, also described
by a 17O NMR peak at 1028 ppm. The color faded along with a shift of λmax
up to 555 nm, fully attained 6 min after mixing. Later on, the fading prog-
ressed at 555 nm for more than 1 h (cf. fig. 1c in Ref. (33)). The authors
fitted the results (fig. 1d) as a two-step process with values of t½ ¼ 1.5 min
(decay at 542 nm with assumed build-up at 570 nm), and t½ ¼ 90 min (decay
at 570 nm), suggesting that a 542 nm ! 570 nm conversion was onset dur-
ing the first rapid step. However, as clarified earlier, fig. 1c shows the inter-
mediate maximum at 555 nm, not at 570 nm. Certainly, the fitting plot does
not show that the 570 nm maximum is ever attained at all for the underlying
conditions (33).
The initial UV–vis 542 nm ! 555 nm change correlated with the decay
of the 17O peak at 1028 ppm and its replacement by a new signal emerging at
938 ppm, fully attained in 3 min. The authors proposed that the new species
rapidly formed upon [Fe(CN)5(NOS)]4 decay was [Fe(CN)5(NOS2)]4,
and added new experimental evidence: (1) By working at pH 7.4 (other
conditions as described earlier) the red–violet color (λmax, 542 nm) could
not be captured; instead the solution became “immediately” blue, with λmax
at 577 nm. (2) A further experiment in which 25 mM [Fe(CN)5(NO)]2
was mixed with HS2 at pH 11 (R ¼ 3) led also “immediately” to the blue
species. The color decayed subsequently in the minute time scale.
Given the latter results on the significant pH influence, as well as our very
recent report on the identification of S2NO as a consequence of the high
reactivity of HS2 (49), we looked at the Gmelin analog reaction by using
HS2 instead of HS. In our hands, transparent yellow solutions resulting
from the direct reaction of [Fe(CN)5(NO)]2 with aqueous Na2S2 could
only be obtained in the pH range 9–12; colloidal sulfur was always rapidly
produced by mixing at pH values <9 (34). Fig. 10 displays a set of successive
spectra, obtained at pH 12, with two well-defined regions: an initial fast
absorbance increase at 550 nm (t½, 2 s), is followed by a slower transforma-
tion (t½, 20 s) generating a band maximum centered at 575 nm. The same
pattern was observed in all of the pH range utilized. The inset shows a com-
parison of the spectral features of the final product (λmax, 575 nm) with the
one produced in the Gmelin reaction at pH 7, a remarkable coincidence.
A UV–vis spectrum of [Fe(CN)5(NOS2)]4 was obtained by using the
real-time TD-DFT/QM-MM methodology described earlier (49). Fig. 11
Fig. 10 Successive spectra for the reaction of 0.05 mM [Fe(CN)5(NO)]2 with aqueous
Na2S2, pH 12, R 1 (34). The first spectrum in black was measured 0.5 s after mixing
and corresponds to buffered HS2 . The next five spectra (red to violet) correspond to
reaction times: 1.4, 2.7, 4.0, 5.3, and 6.5 s.
Fig. 11 Real-time TD-DFT/QM-MM calculated spectrum for [Fe(CN)5(NOS2)]4 in water,

performed at the same level of theory as described in Ref. (49).
shows a spectrum that converged after several months yielding a broad band
centered at 552 nm. Overall, a preliminary consideration of our results and
those of Wu and coworkers (33) suggests that the perthionitrite complex
[Fe(CN)5(NOS2)]4 absorbs at 550 nm, and that I550 could transform
in a slower way into a polysulfide-bound species [Fe(CN)5(NOSx)]4,
I575. The conversion would require the availability of sulfane sulfur (So)
in the medium, a feasible situation in the pH range 7–9 due to the decom-
position of HS2 and ensuing formation of polysulfides, as discussed in
Section 4.
A summary of the overall mechanistic features involves identifying
[Fe(CN)5(NOSH)]3 and [Fe(CN)5(NOS)]4 as the initial intermediates
formed in a pH-dependent way. Both anions have similar band maxima
at 535–540 nm, a similar situation reported earlier for free HSNO
and NOS. The novel picture deals with the pH-dependent reactivity of
both species toward the N–S bond homolysis. On the one hand,
[Fe(CN)5(NOS)]4 is moderately stable at pH values 11, affording a decay
of the 535-nm absorption with a t½ 60 s, yielding [Fe(CN)5(NO%)]3
(detected by UV–vis and EPR), probably [Fe(CN)5(HNO)]3, and HS%
radicals that are a source of HS2 , as well as NH3 and N2O (30). The decay
at 535 nm shows a subsequent very slow step comprising the decomposition
of the [Fe(CN)5(NO)]3 intermediate (k 105 s1), also producing addi-
tional N2O (30). On the other hand, and in sharp contrast, working in the
pH range 7–10 allows I575 emerging in a few seconds after mixing, associated
with the reactivity of [Fe(CN)5(NOSH)]3 (t½ 6 s). In addition to reac-
tions (22–24), we define reaction (25) by considering similar arguments as
exposed above for the generation of free S2NO.
3 3
FeðCNÞ5 ðNOSHÞ = FeðCNÞ5 ðNSOHÞ
4 (25)
+HS2 . FeðCNÞ5 ðNOS2 Þ + HS + H +
Although the main question related to the identification of the “red” and
“blue” intermediates seems to be correctly focused by highlighting the dif-
ferent chemistries of HS and HS2 as nucleophilic reagents, solving the
ambiguities on the clear identification of I550 and I570 must be pursued.
A relevant question deals with I550 being either a true intermediate species
or an artifact appearing because of band-overlap during the
535 nm ! 575 nm transformation. Our current work (34) is focused on a
deeper mechanistic probe, also searching for the influence of O2 on the
early onset of new colored intermediates, the chemistry of HNO generation
and N2O-release, as well as on the role of the 1-electron reduced
[Fe(CN)5(NO)]3/[Fe(CN)4(NO)]2 radical-complexes in the route to

final products.
7. CONCLUSIONS
The thionitrous acid HSNO intermediate (in fact, a mixture of ther-
mally accessible tautomers in the aqueous solutions) has been reasonably well
identified by MS and UV–vis spectroscopies, aided by theoretical calcula-
tions. HSNO survives sufficiently after its generation as an initial product
of the transnitrosation reaction of GSNO with H2S. HSNO affords a rich
chemistry toward different substrates, either leading to NO, probably
HNO or eventually other transnitrosation products. It can also generate
S2NO through the reaction with early generated HS2 in the medium.
Although the biologically significant signaling ability of S2NO has not
been demonstrated yet (25,45), the identity and survival of S2NO in water
for minutes/hours are well established facts (49,58). However, the latter
assertion is still strongly questioned in the very recent reports by Filipovic
and Ivanovic-Burmazovic (60,79,80). Indeed, a better knowledge of the
O2-dependent decay routes of S2NO is needed in the time scale of
minutes/hours, giving either NO, HNO/N2O or other hybrid species,
and N-reduced products. The recently reported reactivity of S2NO in apro-
tic solvents toward CN, GSH, and H2S (60) ought to be contrasted with
measurements in aqueous media in search for a pH dependence. The coor-
dinated “Gmelin” species [FeII(CN)5(NOSH)]3 and [FeII(CN)5(NOS)]4
appear to be well-characterized reactive anions (“red” intermediates absorbing
at 535–540 nm), though the true absorption maximum of the “blue” inter-
mediate [FeII(CN)5(NOS2)]4 formed in the pH range 7–10 is still uncertain.
Unraveling the chemistry of the latter species and of other intermediates is still
a main challenge for a thorough, complete description of the Gmelin process.
ACKNOWLEDGMENTS
We thank the University of Buenos Aires and CONICET for financial support. We are
grateful to Dr. Valentı́n T. Amorebieta for his experimental contribution on the Gmelin
reaction, and to Dr. Sara E. Bari for fruitful discussions.
REFERENCES
1. Wang, R. Physiol. Rev. 2012, 92, 791–896.
2. Fukuto, J. M.; Carrington, S. J.; Tantillo, D. J.; Harrison, J. G.; Ignarro, L. J.;
Freeman, B. A.; Chen, A.; Wink, D. A. Chem. Res. Toxicol. 2012, 25, 769–793.
3. Ignarro, J. L. Nitric Oxide: Biology and Pathobiology. Academic Press: San Diego, 2000.
4. Calderwood, A.; Kopriva, S. Nitric Oxide 2014, 41, 72–78.
5. Bari, S. E.; Olabe, J. A. In Gasotransmitters in Plants. The Rise of a New Paradigm in Cell
Signalling; Lamattina, L., Garcı́a-Mata, C., Eds.; Springer International Publishing:
Switzerland, 2016; pp 289–327. ch. 14.
6. Bari, S. E.; Olabe, J. A.; Slep, L. D. Adv. Inorg. Chem. 2014, 67, 87–144.
7. Lehnert, N.; Berto, T. C.; Galinato, M. G. I.; Goodrich, L. E. In The Handbook of Por-
phyrin Science; Kadish, K. M., Smith, K. M., Guilard, R., Eds.; Vol. 14 ;World Scientific:
New Jersey, 2011; pp 1–247. 63.
8. Li, Q.; Lancaster, J. R., Jr. Nitric Oxide 2013, 35, 21–34.
9. Mishanina, T. V.; Libiad, M.; Banerjee, R. Free Rad. Biol. Med. 2015, 11, 457–464.
10. Olabe, J. A. Adv. Inorg. Chem. 2004, 55, 61–126.
11. Toohey, J. L. Anal. Biochem. 2011, 413, 1–7.
12. Ono, K.; Akaike, T.; Sawa, T.; Kumagai, Y.; Wink, D. A.; Tantillo, D. J.; Hobbs, A. J.;
Nagy, P.; Xian, M.; Lin, J.; Fukuto, J. M. Free Rad. Biol. Med. 2014, 77, 82–94.
13. Cuevasanta, E.; Moller, M. N.; Alvarez, B. Arch. Biochem. Biophys. 2017, 617, 9–25.
14. Whiteman, M.; Li, L.; Kostetski, I.; Chu, S. H.; Siau, J. L.; Bhatia, M.; Moore, P. K.
Biochem. Biophys. Res. Commun. 2006, 343, 303–310.
15. Yong, Q. C.; Cheong, J. L.; Hua, F.; Deng, L. W.; Khoo, Y. M.; Lee, H. S.; Perry, A.;
Wood, M.; Whiteman, M.; Bian, J. S. Antioxid. Redox Signal 2011, 14, 2081–2091.
16. Coletta, C.; Papapetropoulos, A.; Erdelyi, K.; Olah, G.; Modis, K.; Panopoulos, P.;
Asimakopoulou, A.; Gero, D.; Sharina, I.; Martin, E.; Szabo, C. Proc. Natl. Acad. Sci.
U.S.A 2012, 109(23), 9161–9166.
17. Fago, A.; Jensen, F. B.; Tota, B.; Feelisch, M.; Olson, K. R.; Helbo, S.; Lefevre, S.;
Mancardi, D.; Palumbo, A.; Sandvik, G. K.; Skovgaard, N. Comp. Biochem. Physiol.
A 2012, 162, 1–6.
18. Lo Faro, M. L.; Fox, B.; Whatmore, J. L.; Winyard, P. G.; Whiteman, M. Nitric Oxide
2014, 41, 38–47.
19. Berenyiova, A.; Grman, M.; Mijuskovic, A.; Stasko, A.; Misak, A.; Nagy, P.;
Ondriasova, E.; Cacanyiova, S.; Brezova, V.; Feelisch, M.; Ondrias, K. Nitric Oxide
2015, 46, 123–130.
20. Grman, M.; Jawad Nassim, M.; Leontiev, R.; Misak, A.; Jakusova, V.; Ondrias, K.;
Jacob, C. Antioxidants 2017, 6, 14.
21. King, S. B. Free Rad. Biol. Med. 2013, 55, 1–7.
22. Nonella, M.; Huber, J. R.; Ha, T. K. J. Phys. Chem. 1987, 91, 5203–5209.
23. Seel, F.; Kuhn, R.; Simon, G.; Wagner, M.; Krebs, B.; Dartmann, M. Z. Naturforsch.
1985, 40b, 1607–1617.
24. Filipovic, M. R.; Miljkovic, J. L.; Nauser, T.; Royzen, M.; Klos, K.; Shubina, T.;
Koppenol, W. H.; Lippard, S. J.; Ivanovic-Burmazovic, I. J. Am. Chem. Soc. 2012,
134, 12016–12027.
25. Wedmann, R.; Zahl, A.; Shubina, T. E.; Durr, M.; Heinemann, F. W.; Eberhard, B.;
Bugenhagen, C.; Burger, P.; Ivanovic-Burmazovic, I.; Filipovic, M. R. Inorg. Chem.
2015, 54, 9367–9380.
26. Cortese-Krott, M. M.; Fernandez, B. O.; Santos, J. L. T.; Mergia, E.; Grman, M.;
Nagy, P.; Kelm, M.; Butler, A.; Feelisch, M. Redox Biol. 2014, 2, 234–244.
27. Cortese-Krott, M. M.; Kuhnle, G. G.; Dyson, A.; Fernandez, B. O.; Grman, M.;
DuMond, J. F.; Barrow, M. P.; McLeod, G.; Nakagawa, H.; Ondrias, K.; Nagy, P.;
King, S. B.; Saavedra, J. E.; Keefer, L. K.; Singer, M.; Kelm, M.; Butler, A. R.;
Feelisch, M. Proc. Natl. Acad. Sci. U.S.A. 2015, 112, 4651–4660.
28. Broniowska, K. A.; Hogg, A. Antioxid. Redox Signal. 2012, 17, 969–980.
29. Szacilowski, K.; Stasicka, Z. Progr. React. Kin. Mech. 2001, 26, 1–58.
30. Quiroga, S. L.; Almaraz, A. E.; Amorebieta, V. T.; Perissinotti, L. L.; Olabe, J. A. Chem.
Eur. J. 2011, 17, 4145–4156.
31. Filipovic, M. R.; Ivanovic-Burmazovic, I. Chem. Eur. J. 2012, 18, 13538–13540.
32. Filipovic, M. R.; Eberhardt, M.; Prokopovic, V.; Mijuskovic, A.; Orescanin Dusic, O.;
Reeh, P. H.; Ivanovic-Burmazovic, I. J. Med. Chem. 2013, 56, 1499–1508.
33. Gao, Y.; Toubaei, A.; Kong, X.; Wu, G. Chem. Eur. J. 2015, 21, 17172–17177.
34. Olabe, J.A., 3rd European Colloquium on Inorganic Reaction Mechanisms, ECIRM
2016, June 2125, Kraków. Oral presentation, unpublished work.
35. Arulsamy, N.; Bohle, D. S.; Butt, J. A.; Irvine, G. J.; Jordan, P. A.; Sagan, E. J. Am.
Chem. Soc. 1999, 121, 7115–7123.
36. Hu, Y.; Stanbury, D. M. Inorg. Chem. 2016, 55, 7797–7803.
37. Das, T. N.; Huie, R. E.; Neta, P.; Padmaja, S. J. Phys. Chem. A 1999, 103, 5221–5226.
38. Koppenol, W. H. Inorg. Chem. 2012, 51, 5637–5641.
39. Arnelle, D. R.; Stamler, J. S. Arch. Biochem. Biophys. 1999, 318(2), 279–285.
40. Singh, S. P.; Wishnok, J. S.; Keshive, M.; Deen, W. M.; Tannenbaum, S. R. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 14428–14433.
41. Wong, P. S. Y.; Hyun, J.; Fukuto, J. M.; Shirota, F. N.; DeMaster, E. G.;
Shoeman, D. W.; Nagasawa, H. T. Biochemistry 1998, 37, 5362–5371.
42. Cortese-Krott, M. M.; Butler, A.; Woollins, J. D.; Feelisch, M. Dalton Trans. 2016, 45,
5908–5919.
43. Ivanova, L. V.; Anton, B. J.; Timerghazin, Q. K. Phys. Chem. Chem. Phys. 2014, 16,
8476–8486.
44. Nava, M.; Martin-Drummel, M. A.; Lopez, C. A.; Crabtree, K. N.; Womack, C. C.;
Nguyen, T. L.; Thorwirth, S.; Cummins, C. C.; Stanton, J. F.; McCarthy, M. C. J. Am.
Chem. Soc. 2016, 138, 11441–11444.
45. Koppenol, W. H.; Bounds, P. L. Arch. Biochem. Biophys. 2017, 617, 3–8.
46. Seel, F.; Wagner, M. Z. Anorg. Allg. Chem. 1988, 558, 189–192.
47. Bohle, D. S.; Hansert, B.; Paulson, S. C.; Smith, B. D. J. Am. Chem. Soc. 1994, 116,
7423–7424.
48. Szacilowski, K.; Chmura, A.; Stasicka, Z. Coord. Chem. Rev. 2005, 249, 2408–2436.
49. Marcolongo, J. P.; Morzan, U. N.; Zeida, A.; Scherlis, D. A.; Olabe, J. A. Phys. Chem.
Chem. Phys. 2016, 18, 30047–30052.
50. Slep, L. D.; Pollak, S.; Olabe, J. A. Inorg. Chem. 1999, 38, 4369–4371.
51. Munro, A. P.; Williams, D. L. H. J. Chem. Soc. Perkin 2000, 21, 1794–1797.
52. Armstrong, D. A.; Huie, R. E.; Koppenol, W. H.; Lymar, S. V.; Merenyi, G.; Neta, P.;
Ruscic, B.; Stanbury, D. M.; Steenken, S.; Wardman, P. Pure Appl. Chem. 2015, 87,
1139–1150.
53. Eberhard, M.; Dux, M.; Namer, B.; Miljkovic, J.; Cordasic, N.; Will, C.; Kichko, T. L.;
de la Roche, J.; Fischer, M.; Suarez, S. A.; Bikiel, D.; Dorsch, K.; Leffler, A.; Babes, A.;
Lampert, A.; Lennerz, J. K.; Jacobbi, J.; Marti, M. A.; Doctorovich, F.; Hoggestat, E. D.;
Zygmunt, P. M.; Ivanovic-Burmazovic, I.; Messlinger, K.; Reeh, P.; Filipovic, M. R.
Nat. Commun. 2014, 5, 4381.
54. Nagy, P. Methods Enzymol. 2015, 554, 3–29.
55. Cuevasanta, E.; Zeida, A.; Carballal, S.; Wedmann, R.; Morzan, U. N.; Trujillo, M.;
Radi, R.; Estrin, D. A.; Filipovic, M. R.; Alvarez, B. Free Radic. Biol. Med. 2015,
80, 93–100.
56. Giggenbach, W. Inorg. Chem. 1971, 11, 1333–1338.
57. Bailey, T. S.; Henthorn, H. A.; Pluth, M. D. Inorg. Chem. 2016, 55, 12618–12625.
58. Bolden, C.; King, S. B.; Kim-Shapiro, D. B. Free Rad. Biol. Med. 2016, 99, 418–425.
59. Estrı́n, D. A.; Baraldo, L. M.; Slep, L. D.; Barja, B. C.; Olabe, J. A. Inorg. Chem. 1996,
35, 3897–3903.
60. Wedmann, R.; Ivanovic-Burmazovic, I.; Filipovic, M. R. Interface Focus 2017, 7,

20160139.
61. Cuevasanta, E.; Lange, M.; Bonanata, J.; Coitiño, E. L.; Ferrer-Sueta, G.;
Filipovic, M. R.; Alvarez, B. J. Biol. Chem. 2015, 290, 26866–26880.
62. Everett, S. A.; Schoneich, C.; Stewart, J. H.; Asmus, K. D. J. Phys. Chem. 1992, 96,
306–314.
63. Shafirovich, V.; Lymar, S. V. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 7340–7345.
64. Smulik, R.; Debski, D.; Zielonka, J.; Michalowski, B.; Adamus, J.; Marcinek, A.;
Kalyanaraman, B.; Sikora, A. J. Biol. Chem. 2014, 289, 35570–35581.
65. Pokkrebyshev, G. A.; Shafirovich, V.; Lymar, S. V. J. Am. Chem. Soc. 2008, 112,
8295–8302.
66. Frost, A. A.; Pearson, R. G. Kinetics and Mechanism: A Study of Homogeneous Chemical
Reactions, 2nd ed.; John Wiley & Sons: New York, London, 1961.
67. Szacilowski, K.; Wanat, A.; Barbieri, A.; Wasiliewska, E.; Witko, M.; Stochel, G.;
Stasicka, Z. New J. Chem. 2002, 26, 1495–1502.
68. Szacilowski, K.; Stochel, G.; Stasicka, Z.; Kisch, H. New J. Chem. 1997, 21, 893–902.
69. Schwane, J. D.; Ashby, M. T. J. Am. Chem. Soc. 2002, 124, 6822–6823.
70. Gao, Y.; Mossing, B.; Wu, G. Dalton Trans. 2015, 44, 20338–20343.
71. Andreasen, L. V.; Lorkovic, I. M.; Richter-Addo, G. B.; Ford, P. C. Nitric Oxide 2002,
6, 228–237.
72. Perissinotti, L. L.; Estrin, D. A.; Leitus, G.; Doctorovich, F. J. Am. Chem. Soc. 2006,
128, 2512–2513.
73. Perissinotti, L. L.; Leitus, G.; Shimon, L.; Estrin, D.; Doctorovich, F. Inorg. Chem. 2008,
47, 4723–4733.
74. Johnson, M. D.; Wilkins, R. G. Inorg. Chem. 1984, 23, 231–235.
75. Morando, P. J.; Borghi, E. B.; de Schteingart, L. M.; Blesa, M. A. J. Chem. Soc. Dalton
Trans. 1981, (2), 435–440.
76. Miljkovic, J.; Kenkel, I.; Ivanovic-Burmazovic, I.; Filipovic, M. Angew. Chem. Int. Ed.
2013, 52(46), 12061–12064.
77. Playfair, L. Annalen 1850, 74, 317.
78. Butler, A. R.; Calsy-Harrison, A. M.; Glidewell, C.; Sorensen, P. E. Polyhedron 1988, 7,
1197–1202.
79. Ivanovic-Burmazovic, I. In The Chemistry and Biology of Nitroxyl (HNO);
Doctorovich, F., Farmer, P. J., Marti, M. A., Eds.; Elsevier Inc., 2017; pp 67–104,
Amsterdam, Netherlands. ISBN: 978-0-12-800934-5. ch. 5.
80. Filipovic, M. In The Chemistry and Biology of Nitroxyl (HNO); Doctorovich, F.,
Farmer, P. J., Marti, M. A., Eds.; Elsevier Inc., 2017; pp 105–126, Amsterdam,
Netherlands. ISBN: 978-0-12-800934-5. ch. 6.
CHAPTER EIGHT
Computational Insights Into the

Reactivity at the Sulfur Atoms
of M3S4 (M 5 Mo, W) Clusters:
The Mechanism of [3 + 2]
Cycloaddition With Alkynes
s G. Algarra, Manuel G. Basallote1
Andre
Instituto de Biomoleculas (INBIO), Facultad de Ciencias, Universidad de Cádiz, Polı́gono Universitario
Campus Rio San Pedro, Puerto Real, Spain
1
Corresponding author: e-mail address: manuel.basallote@uca.es
Contents
1. Introduction: The Reactivity of M3S4 (M ¼ Mo, W) Clusters 312
1.1 Metal-Centered Reactivity 313
1.2 Chalcogenide-Centered Reactivity 316
2. Experimental and Computational Methods 320
3. Kinetics and Mechanistic Studies on the [3 + 2] Cycloaddition Reaction Between
M3S4 Clusters and Alkynes 322
3.1 The Kinetics of Reaction 322
3.2 The Effect of the Alkyne 324
3.3 The Effect of the Metal 332
3.4 The Effect of the Ancillary Ligands at the Metal Centers 335
3.5 The Effect of the Solvent 336
4. Conclusions and Outlook 339
Acknowledgments 340
References 340
Abstract
Whereas the mechanistic aspects of the reactions at the metal sites of M3S4 (M ¼ Mo, W)
clusters are relatively well understood, much less is known about those occurring at
the sulfur atoms. In this chapter, we describe our recent results on the mechanism
of the [3 + 2] cycloaddition reaction between these clusters and alkynes. In all cases,
the process involves the concerted formation of two C–S bonds in a single kinetic step,
as a result of the interaction between an M(μ-S)2 moiety of the cluster and the sp
C atoms of the alkyne. The effects associated with the nature of the alkyne, the metal
centers, the ligands bound to them, and the solvent have been analyzed by using a
methodology consisting of two main steps: (a) the study of the kinetics of the reactions

312 Andres G. Algarra and Manuel G. Basallote
by using global analysis of the UV–Vis spectral changes, and the characterization of the
resulting products by means of X-ray, ESI-MS, UV–Vis, or NMR spectroscopy; (b) the com-
putational analysis of the processes by using density functional theory methods, in
some cases complemented with the use of the activation strain model and energy
decomposition analysis approaches.
ABBREVIATIONS
adc acetylene dicarboxylic acid
ASM activation strain model
btd 2-butyn-1,4-diol
CF3
PhA 1-ethynyl-3,5-bis(trifluoromethyl)benzene
dbbpy 4,40 -di-tert-butyl-2,20 -bipyridine
DFT density functional theory
diphos diphosphine ligand
dmad dimethyl acetylenedicarboxylate
EDA energy decomposition analysis
ESI-MS electrospray ionization mass spectrometry
FMO Frontier molecular orbital
F
PhA 1-ethynyl-4-fluorobenzene
HDS hydrodesulfurization
HOMO highest occupied molecular orbital
LUMO lowest unoccupied molecular orbital
NMR nuclear magnetic resonance
osa outer-sphere adduct
PES potential energy surface
PrA propargyl alcohol
pyr pyramidal
TD-DFT time-dependent density functional theory
tet tetrahedral
UV–Vis ultraviolet–visible
1. INTRODUCTION: THE REACTIVITY OF M3S4

(M 5 MO, W) CLUSTERS
Combined with chalcogenides, molybdenum and tungsten atoms
in oxidation state IV form robust trinuclear [M3(μ3-Q)(μ-Q)3]4+ clusters
(sometimes abbreviated as M3Q4, M ¼ Mo, W; Q ¼ O, S, Se) whereby each
metal center is bound to two bridging (μ-Q) and a capping (μ3-Q) chalco-
genide. These structures exhibit C3v symmetry, with the axis passing
through the capping (μ3-Q) chalcogenide and the center of the plane formed
by the three M centers (see Fig. 1, left). They are also electron precise, with
Mechanism of [3 + 2] Cycloaddition With Alkynes 313
Fig. 1 Two different representations of the core of M3Q4 clusters illustrating the C3v
symmetry (left) and the two types (c, d) of inequivalent ancillary monodentate ligands
(right).
the [M3] fragment featuring six electrons for the formation of three M–M
bonds (1). Often, three monodentate ligands complete the coordination
environment of the M centers, which feature a roughly octahedral config-
uration (Fig. 1, right), if the M–M interactions are disregarded. The reac-
tivity of [M3Q4]4+ clusters has been studied during the last decades because
they provide an excellent opportunity to detect the possible changes asso-
ciated with the proximity of close equivalent reaction sites. For the purpose
of summarizing the results of previous studies, the results have been
divided into two main categories, i.e., metal-centered and sulfur-centered
reactivity.
1.1 Metal-Centered Reactivity

1.1.1 Substitution Reactions
In water, the chemistry of [M3Q4]4+ clusters was thoroughly analyzed by the
group of Sykes in the second half of the 20th century (2). In acidic aqueous
solutions, the coordination environment of each M center is completed with
aqua ligands, leading to clusters of general formula [M3Q4(H2O)9]4+. How-
ever, some of the coordinated water molecules have acidic character, and in
some cases, this leads to the formation of hydroxo complexes with increased
lability. Comprehensive studies, specially focused on the kinetics of substi-
tution of the aqua ligands, have highlighted that the lability of these depends
on the nature of the chalcogenide in the trans position, and this results in two
types of aqua ligands, those trans to (μ3-Q) and those trans to (μ-Q). Desig-
nated by Sykes et al. as c and d, respectively (see Fig. 1, right), they showed by
means of 17O nuclear magnetic resonance (NMR) spectroscopy that aqua
ligands at the d sites are various orders of magnitude more labile than those at
the c sites (3). Moreover, they also concluded that the substitution of each
type of H2O ligand takes place with statistically controlled kinetics, i.e.,
the experimental kinetic traces can be fitted to a model with a single kinetic
step with the rate constant corresponding to the reaction at the third metal
center instead of the expected three successive steps. This phenomenon, also
observed for instance during the formation/decomposition of binuclear
metal complexes with symmetrical polyaza macrocycles (4), requires the
simultaneous fulfillment of two conditions: (a) the metal centers must
behave as independent chromophores; (b) the rate constants for the reaction
at each metal center must be in a statistical ratio (i.e., k1 ¼ 3 k3, k2 ¼ 2 k3).
From a chemical viewpoint, this occurrence is very informative as it requires
the metal centers to behave independently despite their proximity and inter-
actions through bridging ligands and metal–metal bonds. Notably, we have
revised recently the statistical behavior on various [M3Q4(H2O)9]4+ clusters
using state-of-the-art experimental and computational methods. These
studies concluded that the appearance of the phenomenon should not
be taken for granted, as for instance, while the reexamination of the substi-
tution of H2O by NCS at [W3S4(H2O)9]4+ confirmed the existence of sta-
tistical kinetics (5), a subsequent study on the reaction of [Mo3S4(H2O)9]4+
with Cl resulted in a nonstatistical behavior (6). In the latter case, density
functional theory (DFT) and time-dependent density functional theory
(TD-DFT) calculations were also employed to demonstrate the effects
associated with the presence of M–M interactions on the kinetics of the
substitutions. By computing the energy barrier for the substitution of each
of the three equivalent c and d H2O ligands, as well as the ultraviolet–visible
(UV–Vis) spectrum of reactant, intermediates, and products resulting
from these substitutions, it was concluded that none of the conditions
required to observe statistical kinetics are fulfilled during the reaction of
[Mo3S4(H2O)9]4+ with Cl.
1.1.2 Dihydrogen Release by Protonation of Coordinated Hydrides

Besides the aqueous chemistry of [M3(μ3-Q)(μ-Q)3]4+ clusters, the reactiv-
ity of a variety of diphosphino [M3Q4X3(diphosphine)3]+ derivatives
(X ¼ halide or hydride) has been studied in organic solvents (7), partly
because of the relevance of metal–phosphine complexes in homogeneous
catalysis (7). In these structures, the diphosphine ligands occupy positions
trans to (μ3-Q) and one (μ-Q) chalcogenide, in a way that the coordination
environment of the three metal centers remains equivalent. This results in
Fig. 2 Enantiomeric forms of [M3Q4X3(diphos)3]+ clusters (M ¼ Mo, W; Q ¼ S, Se,

X ¼ halide or hydride).
the formation of the two enantiomers with C3 symmetry shown in Fig. 2,

which features similar kinetic and spectroscopic properties. Here, the sub-
stitution of L ligands has been used to obtain more insights into the behavior
and properties of the clusters. Specifically, most of the recent studies deal
with the reactivity of hydrido clusters (X ¼ H), relevant to H2 activation
and storage. Interestingly, while these hydrido clusters remain inert in the
presence of anions such as Cl or Br, addition of HCl or HBr leads to
the chloro- and bromo-complexes, respectively (Eq. 1), together with the
elimination of H2 (8). Moreover, kinetic studies with the cluster
[W3S4H3(dmpe)3]+ in dichloromethane solutions have shown that the reac-
tions feature a second-order dependence on the acid concentration when
this reactant is employed in excess, whereas a first-order dependence is
obtained when the cluster is added in excess. The results thus indicate the
existence of two alternative mechanisms differing in the number of acid
molecules involved:
+ +
M3 Q4 H3 ðdiphosÞ3 + 3HX ! M3 Q4 X3 ðdiphosÞ3 + 3H2 (1)
DFT calculations for the reaction with HCl supported this hypothesis,
indicating that the interaction of the cluster with one acid molecule can
indeed result in the formation of the reaction product with a moderate
energy barrier of 13.4 kcal mol1 in CH2Cl2 (TS1 in Fig. 3). Nevertheless,
the presence of the second acid molecule slightly decreases the barrier of the
process up to 12.2 kcal mol1 via the formation of a H-bond interaction
between the two acid molecules (TS2 in Fig. 3), this pathway being there-
fore preferred under acid excess (8b). Additionally, factors such as the nature
of the metal center (9), the formation of ion pairs (8c), and the nature of the
diphosphine ligands (10) have also been analyzed recently.
2.18
2.25
1.98
2.10
1.
35
2.19
1.94 2.06
0.81 0.78
TS1 TS2
Fig. 3 Transition state structures (Å, B3LYP) for the reaction of [W3S4H3(PH3)6]+,
employed as a model for [W3S4H3(dmpe)3]+, with one (TS1) and two (TS2) HCl molecules
to generate [W3S4H2Cl(PH3)6]+ and H2 (8b).
Fig. 4 Reaction of [Mo3M0 S4(H2O)10]4+ (M0 ¼ Ni, Pd) with tet-H3PO2 to generate
[Mo3M0 (pyr-H3PO2)S4(H2O)9]4+. For simplicity, the H2O ligands coordinated to Mo cen-
ters are not shown.
1.1.3 Stabilization of the Pyramidal Forms of H3PO2 and H3PO3

It has long been known that addition of a heterometal (M0 ) to M3S4 clusters
results in a variety of cuboidal M3M0 S4 clusters (2a,b,11a,b). The cuboidal
clusters [Mo3M0 S4(H2O)10]4+ (M0 ¼ Ni, Pd) represent two of the few metal
complexes (12) known to be capable of stabilizing the pyramidal (pyr) form
of phosphinic (H3PO2) and phosphonic (H3PO3) acids. In solution, these
low-valent phosphorous acids are most stable in their tetrahedral (tet) form;
nevertheless, the presence of the clusters allows the characterization of their
pyr forms (see Fig. 4 for H3PO2). Our studies (13) have shown that the pro-
cesses take place in two consecutive steps that correspond to the coordina-
tion of the tet form of the acid at the M0 center, followed by its
rearrangement into the pyr form, with the latter step being catalyzed by
either an additional molecule of the acid or the solvent.
1.2 Chalcogenide-Centered Reactivity

The reactivity at the chalcogenide ligands of these clusters has been explored
less from a mechanistic point of view, with most of the studies focusing on
synthetic aspects. Thus, predominantly in aqueous media a variety of

[M3(μ3-Q)(μ-Q)3]4+ clusters have been used as building blocks to obtain
the previously mentioned heterometallic M3M0 Q4 (M0 ¼ transition metal)
clusters through the so-called [3 + 1] synthesis (Fig. 5) (2a,b,11a,b). Here,
the incorporated M0 center often behaves as an electron donor, and therefore
the process requires the metal to be in zero or low oxidation states (14). The
resulting heterometallic M3M0 Q4 clusters have been thoroughly studied as
models for reactions at the surface of crystalline MoS2, such as the hydro-
desulfurization of fossil fuels (HDS) (15) or the activation of H2 (16). In a
recent contribution (17), we studied the assembly and disassembly of CuCl
to the cluster [Mo3S4Cl3(dbbpy)3]Cl (dbbpy ¼ 4,40 -di-tert-butyl-2,20 -
bipyridine). Notably, the assembly process is so fast that its kinetics had to
be studied at 85°C using cryo-stopped flow techniques, whereas the dis-
assembly, which is also a fast process, was found to take place upon addition
of halide in a process that is assisted by the CuCl2 counterions.
[M3(μ3-S)(μ-S)3]4+ clusters are also known to react with alkynes leading
to products featuring new C–S bonds (Fig. 6). The reaction between
[Mo3(μ3-S)(μ-S)2(μ-X)(H2O)9]4+ (X ¼ S, O) and acetylene (Fig. 6) (18),
described by Shibahara et al. in the early 1990s, constituted the first report
on this reactivity. Although in such a case, the processes only resulted in the
formation of compounds with two C–S bonds, studies carried out by this
Fig. 5 [3 + 1] building-block synthesis of M3M0 Q4 heterometallic cubane clusters.
Fig. 6 Reaction between [Mo3(μ3-S)(μ-S)2(μ-X)(H2O)9]4+ (X ¼ S, O) and acetylene.

Fig. 7 Products observed in the reaction of alkynes with M3S4 clusters. Adapted with
permission from Pino-Chamorro, J. Á.; Algarra, A. G.; Fernández-Trujillo, M. J.;
Hernández-Molina, R.; Basallote, M. G. Inorg. Chem. 2013, 52 (24), 14334–14342. Copyright
2013 American Chemical Society.
group in the following years showed that additional products can be

obtained when the nature of the cluster or alkyne is modified (19). The pos-
sible products can be classified according to three structural types, designed
as types I, II, and III (Fig. 7) (20). Shibahara and coworkers proposed that
compounds of type I can be generated initially by the concerted formation
of two C–S bonds. Subsequent protonation of type I products generates type
II compounds, which feature a single C–S bond. Finally, type III products
are formed by addition of a second alkyne to type II compounds (19b).
Importantly, throughout this whole series of events, it is assumed that the
reaction of the cuboidal cluster with each equivalent of alkyne requires
two available (μ-S) ligands or, in other words, an M(μ-S)2 unit. M3S4 reac-
tants have in fact three available M(μ-S)2 units to form type I products, one
for each metal center. However, none remain in the latter. The cleavage of
one C–S bond via protonation of the sp2-hybridized carbon results in type II
products, which again feature an M(μ-S)2 unit able to react with alkynes.
Table 1 summarizes the bibliographic information on the structures of prod-
ucts resulting from the addition of alkynes to [M3(μ3-Q)(μ-S)3]4+ clusters.
The formation of type I products represents, undoubtedly, the most inter-
esting process from a mechanistic viewpoint. Indeed, type I compounds can
be considered as metal-dithiolene complexes obtained via the cycloaddition
reaction between an MQ2 moiety of the cluster and the sp-hybridized
–C^C– atoms of an alkyne (21). In this sense, it must be noted that a
variety of mono- and dinuclear molybdenum and tungsten chalcogenides
have been prepared using this procedure (22). The reaction presents
similarities with the OsO4-catalyzed cis-dihydroxylation of olefins, whose
Table 1 Summary of Bibliographic Information on the Structure of Cluster Reaction

Products Resulting From the Addition of Alkynes to M3Q4 Clusters
Structure
Type Reactant Alkyne (R–C^C–R0 ) References
Type I [Mo3(μ3-S)(μ-S)2(μ3-X)(H2O)9]4+ R]R0 ]H 18b
(X ¼ S, O)
[Mo3(μ3-S)(μ-S)3(Hnta)3]2 R]R0 ]H 19c
0
Type II [W3(μ3-S)(μ-S)2(μ-O)(NCS)9] 5
R]R ]H 19a
Type III [W3(μ3-S)(μ-S)3(NCS)9] 5
R]R0 ]H 19b
0
[W3(μ3-S)(μ-S)3(μ-OAc) R]R ]CO2Me 19b
(dtp)3(CH3CN)]
[Mo3(μ3-S)(μ-S)3(Hnta)3]2 R]R0 ]CO2H 19c
dtp, diethyl dithiophosphate; nta, nitrilotriacetate.
Adapted with permission from Pino-Chamorro, J. Á.; Algarra, A. G.; Fernández-Trujillo, M. J.;
Hernández-Molina, R.; Basallote, M. G. Inorg. Chem. 2013, 52 (24), 14334–14342. Copyright 2013
Fig. 8 Concerted vs stepwise mechanisms for the formation of osma-2,5-dioxolane.
mechanism has long been debated (23). Two alternative mechanisms were
postulated for such a process (Fig. 8), one involving the concerted [3 + 2]
addition of an O]Os]O moiety across the C]C bond, and the other
suggesting a stepwise process whereby the initial [2 + 2] addition of an
Os]O moiety across the C]C bond results in a four-membered
metallacyclic intermediate, which rearranges into the final osma-2,5-
dioxolane in a second step. Interestingly, by combining computational
and experimental results (23), it has been demonstrated that the [3 + 2] mech-
anism is preferred.
Continuing with our interest in the mechanistic study of the reactivity

of [M3(μ3-Q)(μ-Q)3]4+ clusters, during the last few years, we have carried
out a number of combined experimental–computational studies aimed at
analyzing the mechanistic details of the formation of type I products, as well
as the factors that affect this reaction. The clusters and alkynes included in
Fig. 9 were employed to this end, and the advances made in the area are
summarized later.
2. EXPERIMENTAL AND COMPUTATIONAL METHODS

Experimentally, the kinetics of these reactions have been studied using
both stopped-flow and conventional UV–Vis spectroscopy; changes in the
UV–Vis spectrum that accompany reactions with a half-life range from the
order of ms to days can be recorded. In this sense, it is important to note that
type I products feature an intense absorption band between 800 and
1000 nm that can be used to monitor the kinetics of reaction and to confirm
the formation of such products. During the course of our studies, the analysis
of the spectral changes has been systematically carried out using global anal-
ysis of the changes over a wide spectral range including the characteristic
absorption band of type I compounds.
Computationally, we have employed DFT methods to determine the
activation and reaction free energies associated with the formation of type I
products. Moreover, TD-DFT calculations have also been used in some cases
to compute the UV–Vis spectrum of the species involved in the reaction, thus
providing an alternative method to confirm that the experimentally observed
spectral changes truly correspond to the formation of type I products.
The activation strain model (ASM) of chemical reactivity (24) has been
used to obtain more insight into the origins of the kinetic and thermody-
namic differences resulting from changes of the nature of the alkyne and
metal center. Within this approach, each point along the potential energy
surface (PES) for the selected bimolecular reaction is divided into two com-
ponents (Eq. 2), namely, the energy required to deform both fragments (i.e.,
reactants) from their ground-state structures into their current ones (ΔEstrain,
Eq. 3), and the interaction energy (ΔEint), which accounts for the energy
associated with both covalent and noncovalent interactions between those
deformed fragments:
ΔE ¼ ΔEstrain + ΔEint (2)
ΔEstrain ¼ ΔEcluster + ΔEalkyne (3)
Fig. 9 Cuboidal clusters and alkynes employed throughout this chapter. For simplicity,
only the coordination environment of one metal center has been drawn.
In some cases, the interaction energy (ΔEint) has a major role in whether
a process takes place or not, and then it is useful to further decompose it by
means of an energy decomposition analysis (EDA) method (25). Herein, we
employed the localized molecular orbital EDA, developed by Su and Li (26),
which allows to partition the previously obtained interaction energies along
the PES into electrostatic (ΔEes), exchange (ΔEex), repulsion (ΔErep), polar-
ization (ΔEpol), and dispersion (ΔEdisp) components (Eq. 4):
ΔEint ¼ ΔEes + ΔEex + ΔErep + ΔEpol + ΔEdisp (4)
Importantly, within this scheme the polarization term (ΔEpol) corre-
sponds to the stabilizing effect caused by relaxation of the fragment orbitals
upon binding and includes empty-occupied orbital mixing within one frag-
ment due to the presence of the other (polarization), and between the two
fragments (charge transfer). Qualitatively, it can therefore be related to the
highest occupied molecular orbital (HOMO)–lowest unoccupied molecular
orbital (LUMO) interactions employed within the framework of Frontier
molecular orbital (FMO) theory to explain the reactivity and regioselectivity
of pericyclic reactions.
3. KINETICS AND MECHANISTIC STUDIES ON THE

[3 + 2] CYCLOADDITION REACTION BETWEEN
M3S4 CLUSTERS AND ALKYNES
In this section, we describe the results of our studies on the reaction of
[M3(μ3-Q)(μ-Q)3]4+ clusters with alkynes (see Fig. 9) to generate type
I products. In all cases, the studies involve the combined use of the previ-
ously mentioned experimental and computational techniques. Note that,
although a brief indication of the specific conditions and computational
methodologies is given, the reader is referred to the source titles for full
details.
3.1 The Kinetics of Reaction

During the course of our work, we have observed that the clusters [1]4+, [2]+,
and [3]+ systematically react with alkynes to generate type I reaction products
in a single kinetic step. The reactions have been studied at 25.0°C under
pseudo-first-order conditions of alkyne excess. In all cases, a first-order
dependence with respect to both the cluster and the alkyne was obtained,
and the resulting second-order rate constants are summarized in Table 2.
Table 2 Summary of the Kinetic Rate Constants (k1, M1 s1) for the Formation of Type I
Products
Cluster Solvent Alkyne k1 (M21 s21) References
[1]4+ H2O adc 45 2a 20
adc 169 7 b
20
btd 1.92 0.02 a
20
btd 3.0 0.1 b
20
[2] +
CH3CN adc 7.8 0.2 27
btd (8.1 0.1) 103 27
dmad 35 1 28
3
PrA (3.74 0.07) 10 28
PhA (3.3 0.1) 103 28
3
F
PhA (8.7 0.2) 10 28
3
CF3
PhA (13.6 0.7) 10 28
CH2Cl2 dmad (8.0 0.2) 28
3
PhA (4 1) 10 28
[3] +
CH3CN adc 6.3 0.1 29
dmad 9.0 0.1 29
2
btd (1.5 0.2) 10 29
2
PrA (4.4 0.2) 10 29
2
PhA (3.64 0.07) 10 29
a +
Mean values in Hpts/Lipts 2 M with different H concentrations ranging from 0.5 to 2.0 M.
b
Mean values in HCl/LiCl 2 M with different H+ concentrations ranging from 0.5 to 2.0 M.
The formation of type I cycloaddition products was confirmed by the

appearance of an absorption band at 850–900 nm, as well as by additional
NMR experiments in selected cases. As an illustrative example, Fig. 10 shows
the spectral changes for the reaction of [1]4+ with adc. Note that the spectral
changes following the formation of type I addition products are often
followed by slower changes corresponding to secondary processes. For the
case of [2]+ an additional kinetic step that involves the release of pyridine
could be identified, but the data concerning this process have not been
included in the present discussion.
Fig. 10 Spectral changes observed for the reaction of [1]+ with adc in 1 M HCl water
solution (T ¼ 25°C; [1]+ ¼ 8 104 M: [adc] ¼ 0.04 M; time base ¼ 5 s).
3.2 The Effect of the Alkyne

Interestingly, the reactions with adc and dmad are always faster than with
the other alkynes, thus indicating that the R and R0 substituents of the
alkyne modify significantly the rate of the reactions. Although for a given
cluster the ordering of the rate constants with the different alkynes is roughly
the same, the three clusters feature different sensitivity to the nature of the
alkyne. This sensitivity can be estimated by using the quotient between the
rate constants with adc and btd, which is 23 (pts) or 56 (Cl) for [1]4+, 420
for [3]+, and 963 for [2]+. Although it cannot be ruled out that the lower
sensitivity of [1]4+ can be caused, at least in some extent, by the change
in the nature of the solvent, the higher sensitivity of [2]+ with respect to
[3]+ in acetonitrile solutions is clear. In addition, the changes observed in
Table 2 for the reactions of [1]4+ in different supporting electrolytes indicate
that the chloro complexes react faster than the aqua and hydroxo complexes,
but the differences associated with changing the nature of the ancillary ligand
are smaller than those associated with changing the alkyne.
3.2.1 The Nature of the Transition State

To complement the experimental kinetic results, we used DFT methods to
analyze the effect of the alkyne substituents on the kinetics of the reaction.
The results are roughly similar for the three clusters. In agreement with the
kinetic results, the calculations indicated that the formation of type
I products takes place in a single step via concerted transition states (TSs) in
which both C–S bonds are formed simultaneously. Importantly, alternative
mechanisms involving the formation of intermediate structures with one C–
S bond were only located in the triplet PES and showed much higher rel-
ative energies (typically over 40 kcal mol1).
As an example, the structures computed for the TS and type I product for
the reaction of [2]+ with the symmetric alkyne btd are included in Fig. 11.
Their analysis indicates that both structures are close to the Cs symmetry
point group, with d(S1–C1) d(S2–C2) and d(Mo1–S1) (Mo1–S2), and
the C^C moiety of the alkyne approaching the cluster through the plane
formed by the S1–Mo1–S2 moiety. The structures also show the preferential
interaction between these two moieties, characterized by the shortening of
Mo1–S1 and Mo1–S2 distances in the TS (2.31 Å, cf. 2.33 Å in [2]+) and the
lengthening of Mo3–S1 and Mo2–S2 distances by ca. 0.05 Å. Mo–Mo
C1 C1
S1 S1
Mo1 C2 Mo1 C2
S2 S2
Mo3 S3 Mo3 S3
Mo2 Mo2
C1–C2 1.25 Mo2–S3 2.32 C1–C2 1.34 Mo2–S3 2.33

C1–S1 2.38 Mo1–Mo2 2.80 C1–S1 1.85 Mo1–Mo2 2.81
C2–S2 2.39 Mo1–Mo3 2.80 C2–S2 1.84 Mo1–Mo3 2.79
Mo1–S1 2.31 Mo2–Mo3 2.75 Mo1–S1 2.39 Mo2–Mo3 2.69
Mo1–S2 2.31 S1–Mo1–S2 90.9 Mo1–S2 2.38 S1–Mo1–S2 85.0
Mo2–S2 2.37 S2–Mo2–S3 93.7 Mo2–S2 2.48 S2–Mo2–S3 92.2
Mo3–S1 2.37 S1–Mo3–S3 93.6 Mo3–S1 2.49 S1–Mo3–S3 92.2
Fig. 11 Optimized structures (Å, B3LYP(PCM)) for the TS (left) and type I product (right)
of the reaction between [2]+ and btd. For simplicity, H atoms at acac and py ligands are
not drawn. Adapted from Pino-Chamorro, J. Á.; Gushchin, A. L.; Fernández-Trujillo, M. J.;
Hernández-Molina, R.; Vicent, C.; Algarra, A. G.; Basallote, M. G. Chem. Eur. J. 2015,
21 (7), 2835–2844. Copyright Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.
distances also show similar trends, with Mo1–Mo2 and Mo1–Mo3 distances
being ca. 0.05 Å longer than Mo2–Mo3.
To illustrate the similarity of the structural changes in the reaction of the
clusters [1]4+, [2]+, and [3]+ with alkynes, Table 3 includes a summary of the
changes in the most relevant distances and angles during the course of var-
ious reactions modeled computationally. The values in the table indicate that
Mo–S distances (d1 and d2) and S–Mo–S angle (α1) are similar for the three
clusters, and the same is true for the C–C bond length (d3) and R–C–C
angles (α2 and α3) at the different alkynes. The computed structures of
the reaction products also show great similarity, especially those with sym-
metrical alkynes. The structural changes in the cluster when proceeding
from reactant to type I products are small, with the values of d1 and d2
increasing only by ca. 0.05 Å and α1 decreasing by ca. 11 degrees, while
the values for the different reactions show, in general, only small deviations.
In contrast, the alkynes undergo larger structural changes, as expected from
the change in the hybridization of the carbon atoms from sp to sp2 through-
out the process. Thus, the d3 values increase by ca. 0.12 Å and the angles
change from values close to 180 degrees to ca. 125 degrees. Surprisingly,
the structures of the TSs are also quite similar for all the systems despite
the very different nature of the alkynes and the variation of the rate constants
derived from experimental kinetics results. Thus, the Mo–S distances remain
essentially constant (deviations smaller than 0.02 Å), whereas the S–Mo–S
angle decreases to ca. 92 degrees. Again, the distances and angles involving
the alkyne show larger changes; d3 increases by ca. 0.05 Å, and α2 and α3
increase to ca. 146 degrees, thus showing that the structures of the TSs
are located in all cases almost midway between those of the corresponding
reactants and products.
Despite the similarity of the structural changes, the computed activation
(ΔG#) and reaction (ΔGr) free energies show significant differences
depending on the alkyne employed, as illustrated in Table 4 for the reactions
of cluster [2]+. Thus, in agreement with the experimental observations, the
barriers for the reactions of this cluster with adc and dmad are significantly
lower than the barriers for others, and these two reactions also appear to be
thermodynamicallya more favored. The reaction with unsymmetrical alky-
nes leads to structures with higher degrees of asymmetry, although it is
important to highlight that the concerted nature of the process remains.
a
Note that the small positive ΔGr values are probably only the result of a systematic error associated with
the employed level of theory as in practice all reactions take place experimentally.
Table 3 Summary of Most Relevant Structural Changes Occurring During the

Cycloaddition Reactions of Clusters [1]+–[3]+ With Selected Alkynes
Parameter
α1 α2 α3
Reaction d1 (Å) d2 (Å) d3 (Å) (degrees) (degrees) (degrees)
Reactants [1]4+ + btd 2.297 2.297 1.210 98.3 179.3 179.4
+
[2] + btd 2.325 2.325 1.211 97.2 179.0 178.3
+
[2] + adc 2.325 2.325 1.211 97.2 176.1 176.1
[2]+ + PhA 2.325 2.325 1.211 97.2 180.0 180.0
+
[3] + btd 2.352 2.375 1.224 97.0 179.5 179.5
+
[3] + adc 2.352 2.375 1.227 97.0 176.2 176.2
[3]+ + PhA 2.352 2.375 1.228 97.0 176.2 176.2
4+
TSs [1] + btd 2.290 2.295 1.254 90.7 147.2 146.7
+
[2] + btd 2.311 2.307 1.254 90.9 145.9 145.0
[2]+ + adc 2.308 2.305 1.256 92.3 143.9 145.2
+
[2] + PhA 2.326 2.303 1.261 90.4 138.2 152.8
+
[3] + btd 2.344 2.369 1.267 91.3 142.9 145.1
+
[3] + adc 2.356 2.349 1.270 94.0 139.2 154.1
+
[3] + PhA 2.368 2.366 1.279 90.7 138.1 148.7
4+
Products [1] + btd 2.355 2.367 1.332 84.6 128.5 128.1
+
[2] + btd 2.385 2.380 1.336 85.0 126.0 126.5
+
[2] + adc 2.374 2.381 1.338 85.9 125.8 125.4
[2]+ + PhA 2.391 2.387 1.337 85.1 122.4 124.4
+
[3] + btd 2.412 2.422 1.338 86.2 125.6 127.1
+
[3] + adc 2.403 2.408 1.339 87.2 128.0 127.8
[3]+ + PhA 2.426 2.438 1.340 85.8 123.6 125.9
Table 4 Summary of Activation (ΔG#) and Reaction (ΔGr) Free Energies (kcal mol1,
B3LYP(PCM)), Computed for the Reaction of [2]+ With Alkynes, Together With the
Δr(C–S) and Δr(Mo–S) Parameters (Å) Calculated Both at the TS and at the Type
I Product Geometries
At the TS At the Type I Product
Alkyne ΔG# ΔGr Δr(C–S) Δr(Mo–S) Δr(C–S) Δr(Mo–S)
adc 13.3 0.7 0.041 0.003 0.008 0.007
dmad 11.9 2.1 0.040 0.002 0.001 0.009
btd 21.9 6.0 0.011 0.004 0.004 0.005
PrA 23.1 5.3 0.070 0.003 0.019 0.000
PhA 21.6 4.9 0.404 0.022 0.037 0.004
F
PhA 21.3 5.3 0.407 0.022 0.035 0.007
CF3
PhA 18.3 3.4 0.388 0.029 0.034 0.007
The asymmetry of the structures can be quantified using the parameters

defined in Eqs. (5) and (6) as Δr(C–S) and Δr(Mo–S). Their values, both
at the TSs and at the type I products, are also included in Table 4. These
show that phenyl substituents lead to the largest C–S bond differences at
the TS geometries; however, they are dramatically reduced when the type
I products are reached.
Δr ðC SÞ ¼ jd ðC1 S1Þ dðC2 S2Þj (5)
Δr ðMo SÞ ¼ jdðMo1 S1Þ dðMo1 S2Þj (6)
3.2.2 Activation Strain Analysis

In order to understand better the clear differences in the energy profiles for
the reactions with adc and dmad with respect to the remaining alkynes, we
used the ASM to compare those with adc and btd. The analysis was not only
limited to the TS geometries (see Table 5) but also extended to the PESs that
connect reactants and products. The latter analysis, whose results are
included in Fig. 12, was carried out using the two C–S bond-forming dis-
tances as reaction coordinate. This parameter undergoes well-defined
changes throughout the reaction, and it was computed from 4.0 (close to
noninteracting reactants) to 1.9 Å (near the reaction products). Indeed, in
both cases the early stage of the processes (ca. 4.0–2.8 Å) is characterized
by a smooth increase of ΔE that is indicative of the absence of stabilizing
Table 5 ASM Analysis (kcal mol1, B3LYP(PCM)) of the Reactions Between [2]+ and the
Alkynes adc and btd
Reaction ΔE# ΔEr ΔE#strain ([2]+) ΔE#strain (Alkyne) ΔE#strain (Total) ΔEint
[2]+ + adc 17.2 3.6 20.9 2.8 23.8 6.6
[2] + btd +
25.0 9.6 23.0 3.1 26.1 1.1
A B
60 20
ΔEint ([2]+.alkyne) (kcal mol−1)

50
0
ΔE (kcal mol−1)
40
30 −20
20
−40
10
0 −60
4.0 3.5 3.0 2.5 2.0 4.0 3.5 3.0 2.5 2.0
x (Å) x (Å)
C D
60 60
ΔEstrain (alkyne in [2]+.alkyne)
ΔEstrain ([2]+ in [2]+.alkyne)
50 50
40 40
(kcal mol−1)
(kcal mol−1)
30 30
20 20
10 10
0 0
4.0 3.5 3.0 2.5 2.0 4.0 3.5 3.0 2.5 2.0
x (Å) x (Å)
Fig. 12 Activation strain diagrams (kcal mol1, B3LYP(PCM)) for the reaction of [2]+ with
btd (open symbols) and adc (gray symbols) along the reaction coordinate x (Å) projected
onto the two forming C–S bonds. Note that ΔE ¼ ΔEint ([2]+alkyne) + ΔEstrain([2]+) +
ΔEstrain(alkyne). Adapted from Pino-Chamorro, J. Á.; Gushchin, A. L.; Fernández-Trujillo,
M. J.; Hernández-Molina, R.; Vicent, C.; Algarra, A. G.; Basallote, M. G. Chem. Eur. J.
2015, 21 (7), 2835–2844. Copyright Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.
cluster–alkyne interactions, leading to the formation of adducts. The analysis

of the ΔEint (Fig. 12B) and ΔEstrain (Fig. 12C and D) curves shows that such
destabilization originates from the addition of two unfavorable terms, i.e.,
the energy required to deform both reactants as they come closer, and
the unfavorable interaction between them that results from the steric
(Pauli) repulsion. Notably, the curves also show that ΔEstrain (cluster) < Δ
Estrain (alkyne) throughout the whole process, and this can be rationalized
by taking into account that the Mo(μ-S)2 moiety of [2]+ is already nonlinear,
and therefore, it does not require major structural changes to react with the
alkyne. In contrast, the formal change in the hybridization of the reacting
C atoms of the alkyne, from sp to sp2, produces changes in the R–C^C
angles that result in most of the total strain energy (ΔEstrain).
As shown in Table 5, the ΔE#strain values for the reactions with adc and
btd are 23.8 and 26.1 kcal mol1 at the TS geometries, respectively.
Although these values are already in line with the experiments, the same
table indicates that the differences in ΔE#int also contribute to lower the bar-
rier for the reaction with adc by 5.5 kcal mol1. Inspection of the ΔEint cur-
ves in Fig. 12B shows that in both cases, this parameter reaches its maximum
near the TS geometries (ca. 2.8–2.4 Å). Interestingly, the roughly
5 kcal mol1 difference between the two ΔE#int values is not only evident
within this region of the diagram but also seems to remain practically con-
stant up to the reaction products (ca. 1.9 Å).
Alternatively, FMO arguments can be used to explain the differences in
ΔE#int for the two reactions from a qualitative viewpoint. To do so, the ener-
gies of the HOMO and LUMO of [2]+ and alkynes are plotted in Fig. 13
both at their ground-state geometries (laterals of each plot) and at their
geometries at the TSs (central area of each plot). Similar to other
Fig. 13 Energies (eV, B3LYP(PCM)) of the HOMO and LUMO orbitals of [2]+ and alkyne
(adc (left) and btd (right)) fragments in their ground-state geometries and TSs for the
formation of type I products. The arrows represent the smallest HOMO–LUMO gap for
each TS. Adapted from Pino-Chamorro, J. Á.; Gushchin, A. L.; Fernández-Trujillo, M. J.;
Hernández-Molina, R.; Vicent, C.; Algarra, A. G.; Basallote, M. G. Chem. Eur. J. 2015,
21 (7), 2835–2844. Copyright Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.
cycloaddition processes (30), the deformation of the two fragments ([2]+ + adc
or [2]+ + btd) into their TS structures results in an increase in the energy of
the HOMOs and a decrease in those of the LUMOs. This leads to smaller
HOMO–LUMO gaps between the fragments, and therefore to more stabili-
zing interactions between them. Note that, in line with the symmetry-allowed
nature of the process according to the Woodward–Hoffmann rules (31),
the HOMO and LUMO orbitals of the cluster and of the alkyne have the
right symmetry to result in stabilizing HOMO([2]+)–LUMO(alkyne) and
HOMO(alkyne)–LUMO([2]+) interactions, the strength of these being
therefore mainly dependent on the HOMO–LUMO gaps. Fig. 14 shows a
simplified representation of these orbitals and their interactions.
In agreement with the relative small values of ΔE#strain ([2]+) for both
reactions and the negligible differences between them, [2]+ only undergoes
small changes in the energy and composition of its Frontier orbitals when
proceeding from its ground-state structure to the TS geometries (see
Fig. 13). On the contrary, clear differences are observed in the Frontier
orbitals of the alkynes, with those of adc being ca. 1–2 eV more stable than
those of btd at their both ground-state and TS geometries. Interestingly, fur-
ther analysis of these showed that the extra stabilization for adc is due to a
larger extent of charge delocalization over the carboxylic acid groups vs the
hydroxymethyl groups (27), an effect associated with the greater electron-
withdrawing character of the former. Thus, FMO theory allows an expla-
nation of the trends in ΔE#int based on the more favorable orbital interactions
between [2]+ and adc.
Fig. 14 Main orbital interactions in the reaction of [2]+ with alkynes. For simplicity, only
an Mo(μ-S)2 unit of the cluster has been represented. Adapted from Bustelo, E.;
Gushchin, A. L.; Fernández-Trujillo, M. J.; Basallote, M. G.; Algarra, A. G. Chem. Eur. J.
2015, 21, 14823–14833. Copyright Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.
Taken together the results in this section highlight the importance of the
alkyne in [3 + 2] cycloaddition reactions. The trend has similarities with
other concerted cycloadditions such as 1,3-dipolar cycloadditions or
Diels–Alder reactions, in which the FMO interactions are largely influenced
by the electron-donating or electron-withdrawing substituents on the
alkyne/alkene.
3.3 The Effect of the Metal

In the previous section, it was shown that [2]+ reacts with a number of
alkynes (see Table 4) leading to type I cycloaddition products (27,28). In
order to gain insight into the importance of the metal center on the reaction,
we tested the reactivity of the tungsten analogue of this cluster [2W]+ against
the same alkynes (28). Surprisingly, it was found that [2W]+ is inert against
all tested alkynes under conditions similar to those used for [2]+.
A computational (DFT) study was then conducted to analyze such a
striking difference in reactivity. This started with the optimization of the
TSs and type I products of the reaction between [2W]+ and all the exper-
imentally employed alkynes. Comparison of the structures optimized for the
reactions of [2]+ and [2W]+ shows in most cases bond length differences
smaller than 0.1 Å, in agreement with the similar atomic and ionic radii
of Mo and W (32). Substitution of Mo by W has nonetheless a clear effect
on the thermochemistry of the processes. Table 6 includes the activation and
Table 6 Computed Activation (ΔG#) and Reaction (ΔGr) Free Energies (kcal mol1,
B3LYP(PCM)) for the Reaction of [W3S4(Acac)3(py)3]+ ([2W]+) With Alkynes in
Acetonitrile, Together With the Differences (ΔΔG) in These Parameters Respect to
the Reactions of [2]+ (i.e., [2W]+–[2]+)
Alkyne ΔG# ΔGr ΔΔG# ΔΔGr
adc 13.8 7.9 0.5 8.6
dmad 14.3 8.4 2.5 10.5
btd 24.3 16.5 2.4 10.5
PrA 27.1 15.7 4.0 10.4
PhA 24.8 16.3 3.2 11.4
F
PhA 25.4 14.9 4.1 9.6
CF3
PhA 21.7 12.3 3.4 8.9
Adapted from Bustelo, E.; Gushchin, A. L.; Fernández-Trujillo, M. J.; Basallote, M. G.; Algarra, A. G.
Chem. Eur. J. 2015, 21, 14823–14833. Copyright Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.
reaction free energies for the reactions of [2W]+, together with their differ-
ences with respect to those of [2]+. This shows that the ΔGr values for the
reactions of [2W]+ are ca. 10 kcal mol1 larger than those of [2]+,
thus leading to clearly endergonic processes, in agreement with the
experiments. Interestingly, the computed free energy barriers only rise by
0.5–4.1 kcal mol1, resulting in values still typical of reactions at room tem-
perature. This indicates that the lack of reaction between [2W]+ and the
alkynes is only a consequence of the unfavorable thermodynamics of the
processes.
The reaction of the clusters [2W]+ and [2]+ with btd was selected for
further analysis using the ASM methodology. Fig. 15 shows the two activa-
tion strain diagrams, computed at the same level of theory and reaction coor-
dinate as in Fig. 12. In general, the plots corresponding to ΔE, ΔEstrain, and
A B
60 20
ΔEint (cluster.btd) (kcal mol–1)
50
0
ΔE (kcal mol–1)
40
30 –20
20
–40
10
0 –60
4.0 3.5 3.0 2.5 2.0 4.0 3.5 3.0 2.5 2.0
x (Å) x (Å)
C D
ΔEstrain (btd in cluster.btd) (kcal mol–1)
60 60
ΔEstrain (cluster in cluster.btd)
50 50
40 40
(kcal mol–1)
30 30
20 20
10 10
0 0
4.0 3.5 3.0 2.5 2.0 4.0 3.5 3.0 2.5 2.0
x (Å) x (Å)
Fig. 15 Activation strain diagrams (kcal mol1, B3LYP(PCM)) for the reaction of btd with
[2]+ (gray symbols) and [2W]+ (open symbols) along the reaction coordinate x (Å) pro-
jected onto the two forming C–S bonds. Note that ΔE ¼ ΔEint (cluster btd) +
ΔEstrain(cluster) + ΔEstrain(btd). Adapted from Bustelo, E.; Gushchin, A. L.; Fernández-Trujillo,
M. J.; Basallote, M. G.; Algarra, A. G. Chem. Eur. J. 2015, 21, 14823–14833. Copyright Wiley-
VCH Verlag GmbH & Co. KGaA, Weinheim.
ΔEint are relatively similar for both processes, featuring also the expected
trends previously observed. Nevertheless, the ΔE plots show that when pro-
ceeding from reactants (x ¼ 4.0) to products (x ¼ 1.9), the values for ΔE([2-
W]+btd) become gradually larger than those for ΔE([2]+btd), thus resulting
in larger differences in ΔEr than ΔE#. As expected, analysis of the strain
energies for cluster and alkyne indicates that deforming the former is less
energy-demanding. More importantly, the plots for ΔEstrain(cluster) and
ΔEstrain(btd) match almost perfectly for both reactions. This is somehow
unexpected as it implies that the nature of the metal center (Mo or W) does
not modify the energy required to deform the clusters themselves neither
that of btd. Consequently, the energetic differences for both reactions orig-
inate almost exclusively from those in interaction energy (ΔEint), i.e., the
relative differences in the plots for ΔE([2]+btd) and ΔE([2W]+btd) are
equivalent to those between ΔEint([2]+btd) and ΔEint([2W]+btd).
FMO arguments can be used again to explain the differences in ΔEint
from a qualitative viewpoint. Indeed, the computation of the HOMO
and LUMO energies of each fragment along the reaction coordinate
x shows that, while no significant differences are found for those of btd
in both reactions, the HOMO and LUMO energies of [2W]+ are system-
atically higher than those of [2]+ by 0.25 and 0.45 eV. This results in a larger
HOMO–LUMO gap for the interaction of [2W]+ with btd, thus explaining
the decrease in interaction energy. Analogous conclusions are reached when
an EDA is employed. Dissection of the ΔEint values for both reactions along
the reaction coordinate x into electrostatic, exchange, repulsion, polariza-
tion, and dispersion terms shows that most of these terms feature similar
values throughout the two PESs, with significant differences only appearing
for the stabilizing ΔEpol term, i.e., that accounting for the empty-occupied
orbital mixing between (charge transfer) and within (polarization) of each
fragment. This is shown graphically in Fig. 16, where the difference in each
term for the reactions of [2W]+ and [2]+ with btd has been represented.
Whereas most terms only show small and random changes with the distance,
significant differences in polarization energy (ΔΔEpol) start to be evident at
x 2.8 Å and continue to increase up to the reaction products. Thus, as the
ΔΔEpol increases when the two species become closer, the difference in
ΔEint (and in turn ΔE) becomes less pronounced at the TS geometries
(5 kcal mol1) than at the reaction products (12 kcal mol1).
As a summary, the results in this section demonstrate that substitution of
Mo by W in [2]+ reduces the reactivity vs alkynes of the resulting cluster.
Indeed, while the cluster [2]+ reacts with a variety of alkynes leading to
cycloaddition products, its tungsten analogue [2W]+ remains inert under
Fig. 16 Plot of the differences in interaction energy (ΔΔEint) and their LMO-EDA-derived
contributions for the reactions of [2W]+ and [2]+ with btd along the reaction coordinate
x (Å) projected onto the two forming C–S bonds. Adapted from Bustelo, E.; Gushchin, A. L.;
Fernández-Trujillo, M. J.; Basallote, M. G.; Algarra, A. G. Chem. Eur. J. 2015, 21,
14823–14833. Copyright Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.
similar conditions. The effect can be explained by using DFT methods, in

combination with the ASM and EDA approaches. These show that substi-
tution of W by Mo produces changes in the FMOs of the cluster that result
in a decrease of the extent of the interaction with the alkyne, ultimately lead-
ing to thermodynamically disfavored (endergonic) processes. Notably, the
results are also in line with the lower tendency of tungsten aqua clusters
to react with alkynes, as noted by Shibahara et al. (19a).
3.4 The Effect of the Ancillary Ligands at the Metal Centers

To obtain additional information about the effect of ancillary ligands attached
to the metal centers, the reaction between the clusters [M3S4Cl3(dbbpy)3]+
(M ¼ Mo, [3]+; W, [3W]+) and various alkynes has been studied under
conditions similar to those employed previously for [2]+ and [2W]+ (29).
As observed for the tungsten cluster [2W]+, [3W]+ remains inert in the
presence of an excess of alkyne. This observation was accounted for by
the endergonic nature of the processes, which showed computed ΔGr ener-
gies ca. 5–8 kcal mol1 larger for [3W]+ than for [3]+ (see Table 7). Similarly,
the differences in the ΔG# values calculated for the reactions of these clusters
were in all cases smaller than ca. 3 kcal/mol (all values between 6 and
14 kcal mol1), in agreement with the less pronounced effect of the metal
at the TS geometries than at the reaction products.
Table 7 Computed Activation (ΔG#) and Reaction (ΔGr) Free Energies (kcal mol1,
BP86) for the Reaction of [3]+ and [3W]+ With Alkynes in Acetonitrile, Together With the
Differences ([3W]+[3]+) in These Parameters Between Them
[3]+ [3W]+ Difference
Alkyne ΔG# ΔGr ΔG# ΔGr ΔΔG# ΔΔGr
adc 6.0 8.7 5.3 3.3 0.7 5.4
dmad 6.0 9.1 5.8 3.4 0.2 5.7
btd 12.1 4.0 13.4 2.9 1.3 6.9
PrA 10.3 5.5 12.5 2.1 2.2 7.6
PhA 8.4 6.4 10.8 0.5 2.4 6.9
Adapted from Pino-Chamorro, J. Á.; Laricheva, Y. A.; Guillamon, E.; Fernández-Trujillo, M. J.;
Bustelo, E.; Gushchin, A. L.; Shmelev, N. Y.; Abramov, P. A.; Sokolov, M. N.; Llusar, R.; Basallote,-
M. G.; Algarra, A. G. New J. Chem. 2016, 40, 7872–7880 with permission from the Centre National
de la Recherche Scientifique (CNRS) and The Royal Society of Chemistry.
The second-order rate constants obtained for the reaction of [3]+ with
alkynes in acetonitrile solutions can be compared with those of [2]+ in order
to test the importance of the ligands bound to the molybdenum centers in
the [3 + 2] cycloaddition reactions. The data, included in Table 1, show that
both [2]+ and [3]+ react with alkynes at relatively similar rates, and the reac-
tions with adc and dmad are still significantly faster than with the remaining
alkynes. Overall, this indicates that the change in the coordination environ-
ment of the Mo centers only leads to minor differences in reactivity.
An interesting point that deserves to be highlighted pertains to the ori-
entation of the dbbpy and Cl ligands at [3]+ and [3W]+. As shown in Fig. 17
for [3]+, the three possible Mo(μ-S)2 units are equivalent based on the C3
symmetry of the cluster. Nevertheless, the ligands trans to each (μ-S) center
within these units are different, and their different trans effects result in a
Δr(Mo–S) value of 0.02 Å. Although the effect is relatively small and in all
cases the processes are concerted in nature, the computations show that
the asymmetry on the coordination environment of Mo centers has an impact
on the TSs even for the reactions with symmetrical alkynes such as acetylene
(Fig. 17), which features a Δr(C–S) value of 0.03 Å for both C–S bonds.
3.5 The Effect of the Solvent

The rates of concerted cycloaddition reactions such as Diels–Alder and 1,3-
dipolar cycloaddition reactions are known to exhibit only small solvent
effects (33). In fact, the insensitivity of a reaction to changes in the solvent
2.24
2.26
25 2. 25
2.
51 2. 2.38 52 2. 2.41
2.35 2.35
2.44
2.37 2.3
7
Fig. 17 Optimized structures (Å, BP86) of [Mo3S4Cl3(bpy)3]+ (left), employed as a model

for [3]+, and the corresponding TS (right) for the type I product formation with acetylene.
For clarity, H atoms at the bipyridine ligands are not shown.
has even been claimed in some cases as an evidence of the concerted nature
of the reaction under investigation (34).
In order to ascertain whether the formation of type I cycloaddition prod-
ucts between [M3(μ3-Q)(μ-Q)3]4+ clusters and alkynes is subjected to sol-
vent effects, we studied the kinetics of reaction of [2]+ with the alkynes
dmad and PhA not only in acetonitrile solutions but also in dichloromethane
(28). As expected, the resulting second-order rate constants, also included
in Table 1, compare very well with those for the same reactions in acetoni-
trile, thus pointing toward negligible solvent kinetic effects associated with
this [3 + 2] cycloaddition process.
A relatively different scenario arises in aqueous solutions, where the clus-
ter [1]4+ also reacts with alkynes such as adc and btd to generate the
corresponding type I cycloaddition products in a single kinetic step (see
Table 2) (20). In water, the –COOH and –CH2OH groups of the alkynes
can form H-bonding interactions with free H2O molecules (solvent) or with
those acting as ligands in the cluster. As a result, the [3 + 2] cycloaddition is
expected to begin with the formation of outer-sphere cluster–alkyne
adducts. Note that the process continues to be concerted, i.e., the two
C–S bonds are formed in a single kinetic step and, as noted earlier, the reac-
tion with adc is one order of magnitude faster than with btd. The computed
free energy profile for the reaction with btd is included in Fig. 18, which
shows that various outer-sphere adducts (osas) with relatively similar stabi-
lities can be formed depending on the relative position of btd with respect
to the cluster. Among them, osa1 features the correct geometry for the
1.21
1.25
1.61 1.62 1.69 1.69
2.4
0
2.4
1
osa1 TS
TS (+2.8)
4+
[1] + btd
(0.0)
osa2 (–9.2) osa1 (–9.3)

Type I
osa3 (–9.7) product
(–12.6)
osa3 osa2
Fig. 18 Computed free energy profile (kcal mol1, B3LYP(PCM)), for the reaction
between [1]4+ and btd, together with the geometries of three possible outer-sphere
cluster-alkyne adducts. Adapted with permission from Pino-Chamorro, J. Á.;
Algarra, A. G.; Fernández-Trujillo, M. J.; Hernández-Molina, R.; Basallote, M. G. Inorg. Chem.
2013, 52 (24), 14334–14342. Copyright 2013 American Chemical Society.
formation of the type I product, which occurs with an overall barrier of

12.5 kcal mol1. The TS for this process, also included in Fig. 18, shows
how the two HOHOH interactions are maintained, while the
Mo(μ-S)2 and –C^C– units become closer to form the type I product.
In general, the possibility of formation of different osas with geometries
unsuitable for evolving into the cycloaddition product in equilibrium with
the productive osa is expected to result in a deceleration of the cycloaddition
process. The magnitude of this deceleration will depend on the relative sta-
bilities of the different osas, which are in turn strongly dependent on the
nature of cluster and alkyne.
4. CONCLUSIONS AND OUTLOOK

In this chapter, we have summarized our recent studies on the [3 + 2]
cycloaddition reaction between trinuclear [M3(μ3-S)(μ-S)3]4+ clusters
(M ¼ Mo, W; Q ¼ O, S, Se) and alkynes. These include studies on the effects
associated with the nature of the alkyne, the metal centers, the ligands bound
to them, and the solvent. The methodology employed in all cases consisted
of two main steps: (a) the study of the kinetics of the reactions by using global
analysis of the UV–Vis spectral changes, and the characterization of the
resulting products by means of X-ray, ESI-MS, UV–Vis, or NMR spectros-
copy; (b) the computational analysis of the processes by using DFT methods,
in some cases complemented with the use of ASM and EDA approaches. In
all cases, the formation of type I cycloaddition products occurs in a single
kinetic step with rate constants that show a systematic dependence upon
the nature of the alkyne. Despite the relatively different clusters, alkynes,
and solvents used, the structures of the TSs are quite similar in all cases, with
the alkynes approaching the clusters in a side-on way that allows the simul-
taneous formation of both C–S bonds.
Detailed computational studies have revealed that the differences in the
kinetics of reaction are not associated with the structural changes that both
reactants have to undergo, but with the extent of the orbital interactions
between their FMOs. Factors such as the nature of the ancillary ligands at
the M centers or the solvent have a small impact on these, thus resulting
in minimal changes in the kinetics of the cycloaddition processes. In con-
trast, our computations show that the changes on the nature of alkyne
and metal center (Mo vs W) have a significant impact on the extent of these
cluster–alkyne orbital interactions. Indeed, the reaction of a given cluster
with different alkynes has been shown to result in observed rate constants
differing by various orders of magnitude. Nevertheless, despite the fact that

the metal center does not participate directly in any bond-breaking or bond-
forming process, the effect is even more pronounced when Mo centers are
substituted by W as the resulting clusters are unable to react with alkynes.
Future work will focus on the analysis of effects not explored to date, such
as that of replacing S by O or Se, and also in exploring the effects of the for-
mation of type I products on the reactivity at other sites of the cluster, in
particular reactions at the metal centers, with the aim of developing new sys-
tems useful in stoichiometric or catalytic transformations.
ACKNOWLEDGMENTS
We express our gratitude to our coworkers Jose A. Pino-Chamorro, Marı́a Jesús Fernández-
Trujillo, Emilio Bustelo, Rita Hernández-Molina, Artem L. Gushchin, Cristian Vicent, Rosa
Llusar, and Maxim N. Sokolov. Financial support from the Spanish Ministerio de Economı́a y
Competitividad and FEDER funds from the European Union (Grants CTQ2015-
65707-C2-2-P and CTQ2015-71470-REDT) are also acknowledged.
REFERENCES
1. Feliz, M.; Llusar, R.; Andres, J.; Berski, S.; Silvi, B. New J. Chem. 2002, 26(7), 844–850.
2. (a) Hernandez-Molina, R.; Geoffrey Sykes, A. J. Chem. Soc. Dalton Trans. 1999, 18,
3137–3148; (b) Hernandez-Molina, R.; Sokolov, M. N.; Sykes, A. G. Acc. Chem.
Res. 2001, 34(3), 223–230.
3. (a) Kathirgamanathan, P.; Soares, A. B.; Richens, D. T.; Sykes, A. G. Inorg. Chem. 1985,
24(19), 2950–2954; (b) Ooi, B. L.; Sykes, A. G. Inorg. Chem. 1988, 27(2), 310–315;
(c) Ooi, B. L.; Sykes, A. G. Inorg. Chem. 1989, 28(20), 3799–3804; (d) Richens, D. T.;
Helm, L.; Pittet, P. A.; Merbach, A. E.; Nicolo, F.; Chapuis, G. Inorg. Chem. 1989,
28(7), 1394–1402; Richens, D. T.; Pittet, P.-A.; Merbach, A. E.; Humanes, M.;
Lamprecht, G. J.; Ooi, B.-L.; Sykes, A. G. J. Chem. Soc. Dalton Trans. 1993, (15),
2305–2311.
4. Basallote, M. G.; Durán, J.; Fernandez-Trujillo, M. J.; Máñez, M. A. J. Chem. Soc. Dalton
Trans. 1999, (21), 3817–3823; (b) Basallote, M. G.; Durán, J. n.; Fernández-Trujillo, M. J.;
Máñez, M. A. Polyhedron 2001, 20(1–2), 75–82; (c) Basallote, M. G.; Fernandez-Trujillo,-
M. J.; Manez, M. A. J. Chem. Soc. Dalton Trans. 2002, (19), 3691–3695.
5. Algarra, A. G.; Sokolov, M.; González-Platas, J.; Fernández-Trujillo, M. J.;
Basallote, M. G.; Hernández-Molina, R. Inorg. Chem. 2009, 48(8), 3639–3649.
6. Algarra, A. G.; Fernández-Trujillo, M. J.; Basallote, M. G. Chem. Eur. J. 2012, 18(16),
5036–5046.
7. Sorribes, I.; Wienh€ofer, G.; Vicent, C.; Junge, K.; Llusar, R.; Beller, M. Angew. Chem.
Int. Ed. 2012, 51(31), 7794–7798.
8. (a) Basallote, M. G.; Feliz, M.; Fernández-Trujillo, M. J.; Llusar, R.; Safont, V. S.;
Uriel, S. Chem. Eur. J. 2004, 10(6), 1463–1471; (b) Algarra, A. G.; Basallote, M. G.;
Feliz, M.; Fernandez-Trujillo, M. J.; Llusar, R.; Safont, V. S. Chem. Eur. J. 2006,
12(5), 1413–1426; (c) Algarra, A. G.; Basallote, M. G.; Fernández-Trujillo, M. J.;
Llusar, R.; Safont, V. S.; Vicent, C. Inorg. Chem. 2006, 45(15), 5774–5784.
9. Algarra, A. G.; Basallote, M. G.; Fernández-Trujillo, M. J.; Feliz, M.; Guillamón, E.;
Llusar, R.; Sorribes, I.; Vicent, C. Inorg. Chem. 2010, 49(13), 5935–5942.
10. Algarra, A. G.; Basallote, M. G.; Fernández-Trujillo, M. J.; Guillamón, E.; Llusar, R.;
Segarra, M. D.; Vicent, C. Inorg. Chem. 2007, 46, 7668–7677.
11. (a) Hidai, M.; Kuwata, S.; Mizobe, Y. Acc. Chem. Res. 2000, 33(1), 46–52; (b) Llusar, R.;
Uriel, S. Eur. J. Inorg. Chem. 2003, 7, 1271–1290.
12. (a) Nagaraja, C. M.; Nethaji, M.; Jagirdar, B. R. Inorg. Chem. 2005, 44(12), 4145–4147;
(b) Akbayeva, D. N.; Di Vaira, M.; Seniori Costantini, S.; Peruzzini, M.; Stoppioni, P.
Dalton Trans. 2006, (2), 389–395.
13. (a) Algarra, A. G.; Basallote, M. G.; Fernandez-Trujillo, M. J.; Hernandez-Molina, R.;
Safont, V. S. Chem. Commun. 2007, 29, 3071–3073; (b) Algarra, A. G.; Fernandez-
Trujillo, M. J.; Hernandez-Molina, R.; Basallote, M. G. Dalton Trans. 2011, 40(34),
8589–8597; (c) Algarra, A. G.; Fernandez-Trujillo, M. J.; Hernandez-Molina, R.;
Basallote, M. G. Dalton Trans. 2011, 40(34), 8589–8597.
14. (a) Mueller, A.; Fedin, V. P.; Diemann, E.; Boegge, H.; Krickemeyer, E.;
Soelter, D.; Giuliani, A. M.; Barbieri, R.; Adler, P. Inorg. Chem. 1994, 33(10),
2243–2247; (b) Hernández-Molina, R.; Kalinina, I. V.; Abramov, P. A.;
Sokolov, M. N.; Virovets, A. V.; Platas, J. G.; Llusar, R.; Polo, V.; Vicent, C.;
Fedin, V. P. Inorg. Chem. 2008, 47(1), 306–314.
15. (a) Kuwata, S.; Hidai, M. Coord. Chem. Rev. 2001, 213(1), 211–305; (b) Angelici, R. J.
In: Encyclopedia of Inorganic Chemistry; Scott, R. A. Ed.; John Wiley & Sons, Ltd.:
Hoboken, NJ, 2006; pp 1–17; (c) Infantes-Molina, A.; Romero-Perez, A.; Eliche-
Quesada, D.; Merida-Robles, J.; Jimenez-López, A.; Rodrı́guez-Castellón, E. Transition
Metal Sulfide Catalysts for Petroleum Upgrading—Hydrodesulfurization Reactions; InTech:
Rijeka, Croatia, 2012.
16. (a) Jaramillo, T. F.; Bonde, J.; Zhang, J.; Ooi, B.-L.; Andersson, K.; Ulstrup, J.;
Chorkendorff, I. J. Phys. Chem. C 2008, 112(45), 17492–17498; (b) Kibsgaard, J.;
Jaramillo, T. F.; Besenbacher, F. Nat. Chem. 2014, 6(3), 248–253.
17. Pino-Chamorro, J. Á.; Laricheva, Y. A.; Guillamón, E.; Fernández-Trujillo, M. J.;
Algarra, A. G.; Gushchin, A. L.; Abramov, P. A.; Bustelo, E.; Llusar, R.;
Sokolov, M. N.; Basallote, M. G. Inorg. Chem. 2016, 55(19), 9912–9922.
18. (a) Shibahara, T.; Akashi, H. Inorg. Synth. 1992, 29, 260–269; (b) Shibahara, T.;
Sakane, G.; Mochida, S. J. Am. Chem. Soc. 1993, 115(22), 10408–10409.
19. (a) Maeyama, M.; Sakane, G.; Pierattelli, R.; Bertini, I.; Shibahara, T. Inorg. Chem. 2001,
40(9), 2111–2119; (b) Ide, Y.; Sasaki, M.; Maeyama, M.; Shibahara, T. Inorg. Chem.
2004, 43(2), 602–612; (c) Takagi, H.; Ide, Y.; Shibahara, T. C. R. Chim. 2005,
8(6–7), 985–992; (d) Ide, Y.; Shibahara, T. Inorg. Chem. 2007, 46(2), 357–359.
20. Pino-Chamorro, J. Á.; Algarra, A. G.; Fernández-Trujillo, M. J.; Hernández-Molina, R.;
Basallote, M. G. Inorg. Chem. 2013, 52(24), 14334–14342.
21. Rauchfuss, T. B. In Dithiolene Chemistry, Synthesis, Properties and Applications; Stiefel, E. I.
Ed.; Vol. 52; John Wiley & Sons, Inc.: Hoboken, NJ, 2004; pp 1–54.
22. (a) McKenna, M.; Wright, L. L.; Miller, D. J.; Tanner, L.; Haltiwanger, R. C.;
DuBois, M. R. J. Am. Chem. Soc. 1983, 105(16), 5329–5337; (b) Coucouvanis, D.;
Hadjikyriacou, A.; Toupadakis, A.; Koo, S. M.; Ileperuma, O.; Draganjac, M.;
Salifoglou, A. Inorg. Chem. 1991, 30(4), 754–767; (c) Kawaguchi, H.; Tatsumi, K.
J. Am. Chem. Soc. 1995, 117(13), 3885–3886; (d) Mallard, A.; Simonnet-Jegat, C.;
Lavanant, H.; Marrot, J.; Secheresse, F. Transit. Met. Chem. 2008, 33(2), 143–152;
(e) Chiu, W.-H.; Zhang, Q.-F.; Williams, I. D.; Leung, W.-H. Organometallics 2010,
29(11), 2631–2633.
23. Deubel, D. V.; Frenking, G. Acc. Chem. Res. 2003, 36(9), 645–651.
24. (a) Bickelhaupt, F. M.; Baerends, E. J.; Nibbering, N. M. M.; Ziegler, T. J. Am. Chem.
Soc. 1993, 115(20), 9160–9173; (b) Bickelhaupt, F. M. J. Comput. Chem. 1999, 20(1),
114–128; (c) van Zeist, W.-J.; Bickelhaupt, F. M. Org. Biomol. Chem. 2010, 8(14),
3118–3127; (d) Wolters, L. P.; Bickelhaupt, F. M. Wiley Interdiscip. Rev. Comput.
Mol. Sci. 2015, 5(4), 324–343.
25. Hopffgarten, M. v.; Frenking, G. Wiley Interdiscip. Rev. Comput. Mol. Sci. 2012, 2(1),
43–62.
26. Su, P.; Li, H. J. Chem. Phys. 2009, 131(1), 014102.
27. Pino-Chamorro, J. Á.; Gushchin, A. L.; Fernández-Trujillo, M. J.; Hernández-Molina,-
R.; Vicent, C.; Algarra, A. G.; Basallote, M. G. Chem. Eur. J. 2015, 21(7), 2835–2844.
28. Bustelo, E.; Gushchin, A. L.; Fernández-Trujillo, M. J.; Basallote, M. G.; Algarra, A. G.
Chem. Eur. J. 2015, 21, 14823–14833.
29. Pino-Chamorro, J. Á.; Laricheva, Y. A.; Guillamon, E.; Fernández-Trujillo, M. J.;
Bustelo, E.; Gushchin, A. L.; Shmelev, N. Y.; Abramov, P. A.; Sokolov, M. N.;
Llusar, R.; Basallote, M. G.; Algarra, A. G. New J. Chem. 2016, 40, 7872–7880.
30. Ess, D. H.; Jones, G. O.; Houk, K. N. Adv. Synth. Catal. 2006, 348(16–17), 2337–2361.
31. Woodward, R. B.; Hoffmann, R. Angew. Chem. Int. Ed. Engl. 1969, 8(11), 781–853.
32. Atkins, P.; Overton, T.; Rourke, J.; Weller, M.; Armstrong, F. Shriver and Atkins’ Inor-
ganic Chemistry. 5th ed.; Oxford University Press: New York, 2009; pp 783–784.
33. Kumar, S.; Kumar, V.; Singh, S. P. Pericyclic Reactions: A Mechanistic and Problem-Solving
Approach; Elsevier Science: London, UK, 2015.
34. Sauer, J.; Sustmann, R. Angew. Chem. Int. Ed. Engl. 1980, 19(10), 779–807.
CHAPTER NINE
Reactive Oxygen Species in

Photodynamic Therapy:
Mechanisms of Their Generation
and Potentiation
Janusz M. Dąbrowski1
Faculty of Chemistry, Jagiellonian University, Kraków, Poland
1
Corresponding author: e-mail address: jdabrows@chemia.uj.edu.pl
Contents
1. Introduction 344
2. Redox Homeostasis in Cells 346
3. Photodynamic Therapy 348
4. Penetration Depth 349
5. ROS-Generating Systems for PDT 351
6. ROS Generation Mechanisms 355
6.1 Possible Deactivation Pathways (Jablonski Diagram) 355
6.2 Type II Mechanism 357
6.3 Type I Mechanism 358
6.4 Role of Ascorbate in ROS Generation 361
7. Subcellular Localization of Photosensitizers in Cells 362
8. ROS-Mediated Biological Mechanisms 366
8.1 Apoptosis 366
8.2 Necrosis 367
8.3 Autophagy 369
8.4 ROS-Mediated Vascular Occlusion 369
8.5 Antitumor Immune Response 371
9. Methods of ROS Detection in PDT 372
9.1 1270 nm Phosphorescence 372
9.2 Detection of Radical Species by Electron Spin Resonance 374
9.3 ROS Detection by Fluorescent Probes 374
10. Strategies to Enhance ROS Generation in PDT 378
10.1 Inhibition of Antioxidant Enzymes 378
10.2 Nanoformulation of Photosensitizers 379
10.3 Photosensitization of TiO2 With Tetrapyrroles 382
10.4 Addition of Inorganic Salts 383
11. Summary 384
Acknowledgments 386
References 386

344 Janusz M. Dąbrowski
Abstract
Reactive oxygen species (ROS) play key roles in cell signaling systems and homeostasis,
and they are also fundamental to photodynamic therapy (PDT). PDT efficacy can be
affected by the nature and persistence of ROS. A comprehensive understanding of
ROS generation pathways greatly facilitates the analysis of photodynamic mechanisms
and enables potentiation of PDT efficacy. Diverse methods exist to distinguish between
Type I and Type II mechanisms of ROS generation. The direct monitoring of 1O2 forma-
tion involves the detection of its phosphorescence at 1270 nm. Electron spin resonance
is also used in conjunction with appropriate spin traps for detection of oxygen-centered
radical species. Moreover, a variety of more or less specific fluorescent probes are fre-
quently used to detect both singlet oxygen and free radicals. This chapter summarizes
our recent efforts in the design and characterization of new ROS-generating systems for
PDT. Special attention is given to bacteriochlorins because they absorb in the NIR, gen-
erate ROS via both Type I and Type II mechanisms, and are very efficient in the PDT
treatment of several types of tumors including pigmented melanoma. The current sta-
tus and possible opportunities of ROS generation and potentiation in PDT are
highlighted. Particular emphasis is placed on the elucidation of the ROS-mediated pho-
tochemical and molecular mechanisms that give rise to the establishment of PDT as a
first-line systemic treatment of highly resistant diseases, especially invasive and meta-
static tumors.
1. INTRODUCTION
Reactive oxygen species (ROS) are products of partial reduction of
molecular oxygen (O2). Two unpaired electrons with parallel spins located
in two separate orbitals of molecular oxygen make it highly susceptible to
formation of radical forms (1,2). As illustrated in Fig. 1, the one-electron
reduction product of molecular oxygen is called superoxide ion, while
reduction by two electrons leads to hydrogen peroxide. H2O2 is a rather
neutral ROS, but the subsequent product, the hydroxyl radical (HO%), is
one of the strongest oxidants ever described. Its standard reduction potential
Fig. 1 Subsequent stages of molecular oxygen reduction leading to ROS generation

together with their standard reduction potentials.
Reactive Oxygen Species in Photodynamic Therapy 345
of 2.31 V ensures reactions at a very low activation energy barrier and rates
close to diffusion-controlled (3). In addition to these chemically reactive
species delivered from oxygen presented in Fig. 1, a product from an energy
transfer reaction (singlet oxygen, 1O2) (4) will also be discussed in this review
due to its importance in PDT mechanisms.
The stereotypical view of ROS is that their production is unregulated,
and their intracellular targets are rather random. ROS-dependent oxidation
of intracellular lipids, proteins, and DNA leads to the damage of biomole-
cules and implicates several pathologies, including cancer, neurodegenera-
tive diseases, and atherosclerosis (5–7). Various chemical agents such as
drugs, pesticides, some metals, and metal oxides present in smog, some phys-
ical factors such as gamma or UV radiation, high pO2, and temperature, as
well as physiological stresses (injury, aging) lead to an enhanced production
of ROS. They are also created by neutrophils, eosinophils, and macrophages
during inflammation and by-products of mitochondrial-catalyzed electron
transport reactions. As illustrated in Fig. 2, ROS generated within such pro-
cesses result in oxidative damage to proteins, lipids, and nucleic acids and
disrupt functions of organs and cells consequently leading to cell death (8,9).
Although this damaging aspect of ROS is correct, there are no doubts that
they are responsible for many beneficial effects as well. Their formation
ensures proper functioning of metabolism and several signaling pathways
in cells. Both superoxide and hydrogen peroxide have been indicated as
signaling messengers (10). However, due to a higher stability of H2O2, it seems
to play a more important role as a signaling agent. Biochemical pathways
Fig. 2 Schematic view of the influence of ROS induced by various extrinsic factors on
cell metabolism.
and regulated enzymatic systems for generating H2O2 have been described
(11). Although ROS and their targets are spatially confined within the cell,
their role in redox signaling, metabolic regulation, innate immunity, stem cell
biology, and photodynamic therapy of cancer is absolutely undeniable (12).
This chapter starts with a brief description of the redox homeostasis and
oxidative stress in cells and indicates how this overproduction of ROS can
be used in anticancer therapies. Particular attention is given to photody-
namic therapy, a treatment that employs photogenerated ROS which
destroy unwanted cells/tissue. PDT has been widely exploited as a promis-
ing, minimally invasive anticancer and antimicrobial strategy over the recent
years. Nevertheless, PDT still faces several problems namely light penetra-
tion depth, Ps delivery and localization, Ps photophysics and photochemis-
try which all together contribute to the accurate engineering of ROS
formation. Starting at the molecular level and progressing to biological out-
comes, these and other factors are discussed to create a mechanistic point of
view on the current approaches that may overcome most of the PDT chal-
lenges. Consequently, the discussion is focused on the mechanisms of cel-
lular death and vascular occlusion as well as the ability of ROS to
stimulate immune response after PDT. Finally, several examples of combi-
nations with antioxidants, inorganic salts, and nanoparticles are given to
demonstrate the opportunities of ROS potentiation and eventual enhance-
ment of the therapeutic efficacy.
2. REDOX HOMEOSTASIS IN CELLS

Cellular survival strongly depends on the redox homeostasis. In some
cases ROS may induce stress within the cell that leads to disruption of its
functions. One such situation is defined as oxidative stress and is character-
ized by enhanced production of ROS with the simultaneous impairment of
cellular defense mechanisms. To avoid or reduce this, organisms have devel-
oped effective scavenging systems. Antioxidant enzymes including three
types of superoxide dismutase (SOD1, SOD2, and SOD3), glutathione per-
oxidase, and catalase (CAT) are responsible for balancing ROS generation in
cells. Small molecules such as vitamin E, carotenoids, and vitamin
C (ascorbate) support these enzymes in maintaining the proper level of
ROS. Superoxide ion (O2 • ) is the main initial free radical species, which
can be converted to other reactive species such as H2O2 by SOD, which can
be reused to generate superoxide radicals or converted to water by catalase
and/or glutathione (GSH) (13). In the presence of transition metals (such as
Fig. 3 Schematic illustration of the redox homeostasis and oxidative stress in the cell.
Fe2+), H2O2 can be converted to hydroxyl radicals (HO%), which are highly
reactive and can cause damage to lipids, proteins, and DNA. Fig. 3 illustrates
the basic reactions in cells ensuring control of homeostasis and the situation
when this control is affected by an excess of ROS and/or by depletion of
antioxidant activity in a condition known as oxidative stress.
As illustrated in Fig. 3, mitochondria reduce molecular oxygen to water
in the process known as oxidative phosphorylation. Cytochrome c oxidase
catalyzes the transfer of four electrons to an oxygen with progressive ROS
generation. The leakage of electrons from the mitochondrial electron trans-
port chains is the source of O2 • , which then is transformed to H2O2, HO%,
and peroxynitrite (ONOO), respectively (14,15). A moderate increase of
ROS promotes cell proliferation and survival. However, when the increase
of ROS reaches a certain level, it may overcome the antioxidant capacity of
the cell and trigger cell death. Normal cells can tolerate some amount of
exogenous ROS, while cancer cells, due to an abnormal metabolic activity,
change the redox dynamics to maintain the ROS level below the toxic
threshold (16). A further increase of ROS stress in cancer cells using exog-
enous ROS-modulating agents is likely to cause elevation of ROS above the
threshold level, leading to cell death. Since cancer cells are more susceptible
to exogenous ROS, their survival is more dependent on antioxidant activity
(13). Thus, ROS-mediated therapeutic strategies may cause more damage to
malignant cells than normal cells (17). One of the medical approaches, in
which the ROS-inducing agent is activated in the unwanted lesion, is
known as photodynamic therapy (PDT) (16,18).
3. PHOTODYNAMIC THERAPY
PDT is a treatment modality for cancer (19–24), age-related mac-
ular degeneration (25–27), and localized infections (28–39) that involves
the use of three separately nontoxic components: a photosensitizing drug
or photosensitizer (Ps), a light source that emits visible and/or near-
infrared (NIR) photons, and molecular oxygen dissolved in the target
tissue (18–20,22,40–45). The proper interaction of these elements leads
to the generation of ROS that are responsible for the photodynamic
effect. Fig. 4 illustrates the basic concept and biological consequences
of PDT.
Photogenerated ROS contribute to the tumor destruction which occurs
in accordance to several direct and indirect mechanisms. ROS usually lead to
the direct cell death by apoptosis and/or necrosis. Autophagy may also cause
the cell death, but on the other hand, its cytoprotective and prosurvival
functions are well documented (46–49). If the irradiation of the tumor takes
place after relatively short DLI, ROS damage tumor-associated vessels. This
strategy is known as vascular-targeted photodynamic therapy (V-PDT)
(26,40,50–58). PDT action ends with the activation of immune response
and induction of long-term antitumor memory (20,23,59–61). Among
many factors participating in the eradication of the treated tumor and
responsible for innate and adaptive immunity, the most important ones
are listed in Fig. 4.
Fig. 4 Basic principles of photodynamic therapy (PDT): Photosensitizer (Ps) is adminis-

tered, and after a certain drug to light interval (DLI), it is activated by visible or NIR light.
Photogenerated ROS lead to tumor cell death and vascular occlusion accompanied with
induction of local inflammatory reactions and development of systemic antitumor
immunity.
PDT is often described as the cancer treatment based on the Ps accumu-

lation specifically in cancer cells (22,42–44,62–65). However, more recent
reports demonstrate that Ps can be activated without entering the cells but
can be efficiently excited to generate ROS, while bound to serum albumin,
shortly after administration (40,52,56–58,66,67). Even if the issue of selectiv-
ity is often overstated, two factors promote the selectivity of PDT. First, many
photosensitizers show affinity for tumors and their vasculature. Second, the
photosensitizer is activated only after irradiation, thus photogenerated ROS
destroy specifically the tumor irradiated area even if Ps is present in normal
tissues. Several explanations have been suggested for the preferential binding
of sensitizers to neoplastic tissues, namely, low tumor pH, elevated LDL
receptor expression, poorly developed lymphatic drainage, and abnormal
tumor neovascularization (43). Some photosensitizers are known to bind to
low density lipoproteins and are directed to malignant cells that often express
high levels of LDL receptors (18,43,68,69).
4. PENETRATION DEPTH
The wavelength of light for effective photosensitizer activation in
order to generate ROS corresponds to an electronic absorption band of
the applicable compound. The penetration depth and efficient delivery of
light are two major barriers in anticancer PDT for deep tissue treatment.
As illustrated in Fig. 5, the consequences of light interaction with tissues
Fig. 5 Light interactions with human tissues.

are not only the absorption but also the scattering, transmission, and reflec-
tion (70). Penetration depth of light through tissues depends on many
parameters such as photon energy, molar absorption coefficient, polariza-
tion, coherence, power density, time of illumination, and the tissue physi-
ology (e.g., pigmentation and hydration) (2,71). Endogenous fluorophores,
including hemoglobin and melanin, have a strong absorption in the visible
part of the spectrum below 630 nm. Therefore, an ideal Ps should have an
absorption peak above 630–700 nm to maximize tissue penetration (72).
Taking into account that water absorbs in the NIR and that the energy
decreases with increasing wavelength, the light in the range between 700
and 1000 nm is the most suitable for deep tissue penetration (73).
On the other hand, light of low energy (λ > 850 nm) is less effective in
ROS generation due to thermal effects caused by fast nonradiative transi-
tions and the narrow energy gap (45,74). Shorter wavelengths of light are
the most scattered by tissues, and their penetration is limited; therefore,
the photosensitizing agents absorbing the red or near-infrared photons
(between 650 and 850 nm) have the greatest potential because these wave-
lengths penetrate tissues the most effectively (18,45,73).
It is worth noting that the scattering and reflection of light also implicate
other challenges in PDT. These processes may lead to the situation where
the photons do not reach all parts of the irradiated area. In order to generate
ROS in whole tumor tissue, an accurate irradiation margin has to be applied.
For the tumors with thicknesses of 3–4 mm used in our studies (23,66,67),
the 3–4 mm margin should be associated with an irradiance at a depth
z ¼ 7 mm capable of producing ROS above the therapeutic threshold.
The threshold concentration for tissue necrosis by singlet oxygen was esti-
mated as [1O2] ¼ 0.9 mM for liver necrosis and [1O2] ¼ 93 mM for skin
necrosis (75,76). However, a typical threshold dose for tissue necrosis is
17 mM, and the amount of ROS produced per unit volume of tissue is given
by Eq. (1):
½ROS ¼ ΦROS ð1000λ=hcNA ÞRε ½Pslocal ¼ 460R½Pslocal (1)
where R ¼ tirrE is the radiant exposure in J/cm2, and the numeric value
results from the use of the parameters of redaporfin (23). In vascular-targeted
PDT of colon carcinoma, we have used a drug dose of 0.75 mg kg1 which
gives a Ps concentration in plasma, [Ps]plasma ¼ 13 μM and [ROS] ¼ 0.3 M at
the tumor surface when R0 ¼ 50 J cm2. At a depth of z ¼ 7 mm, where
R ¼ R0exp(z/δ), for the optimized regimen with illumination at
750 nm, we obtained [ROS] ¼ 14 mM, which should cause tissue necrosis.
Altogether, the success of the optimized treatment regimen is related to the
three-dimensional tumor margin of 3–4 mm (23).
To overcome the limited penetration depth of tissue, several strategies
have been applied including use of materials absorbing light in the NIR
and emitting in the visible. Upconverting nanoparticles and two-photon
excited nanoparticles are examples of such a methodology (77–79). More-
over, quantum dots and bioluminescence resonance energy transfer sys-
tems are also quite popular recently (2). Our approach in the PDT field is
much more simple and even more effective. We use NIR absorbing
(750 nm) photostable photosensitizers (e.g., bacteriochlorins) with ade-
quately designed treatment regimens (23). It has been estimated that light
at 750 nm can penetrate to a depth even higher than 10 mm and it is still effi-
cient in ROS generation (70,80).
5. ROS-GENERATING SYSTEMS FOR PDT

The continuously increasing number of new photosensitizers for PDT
application has led to a breakthrough in photomedicine. Currently devel-
oped PDT agents have improved most of the typical properties, although
for the most commonly used Photofrin®, several drawbacks such as low light
absorption and poor light penetration through the tissue, poor clearance
from the body, prolonged photosensitivity, and suboptimal tumor selectivity
still exist. In order to increase Ps accumulation in the cancer cells and/or
tumor microenvironment, photosensitizers are conjugated to antibodies,
serum albumins, LDL, sugars, peptides ligands, nonprotein ligands (e.g., folic
acid) and proteins exhibit special targeting. Another strategy is to use
pH-activatable agents encapsulated in liposomes, or polymeric micelles,
which respond to the higher acidity of tumor microenvironment, and
glutathione-activated Ps, since glutathione concentration is also higher in
cancer cells (64,81–92).
Among many features typical for a good Ps, the most important ones are
purity, stability, low toxicity, strong absorption in the phototherapeutic
window (630–850 nm), and ability to generate ROS with high yield
(18,45,62). Furthermore, Ps should have preferential retention in the target
tissue and favorable pharmacokinetics. In the cellular approach of PDT it is
important that Ps localizes in mitochondria or even more preferable in endo-
plasmic reticulum (ER), where ROS are capable of inducing the immuno-
genic cell death (93).
Porphyrins are the most widely studied photosensitizers for both PDT
and photodynamic inactivation of microorganisms (PDI) due to their strong
absorption in the visible region, great ability to generate ROS, particularly
singlet oxygen (94,95), and biocompatibility (96,97). They are involved in a
number of biologically important functions, namely, oxygen transport and
storage (hemoglobin and myoglobin), electron transfer reactions (cyto-
chrome c, cytochrome oxidase), and energy conversion (photosynthesis)
(97). Their versatile synthesis provides applications also for a variety of mate-
rials, particularly for photocatalysis (98–100) and optoelectronics (101,102).
Porphyrins are tetrapyrrolic macrocycles built from four pyrroles connected
with a methine bridge (97). The addition of four phenolic groups to the
macrocycle provides an amphiphilic character and gives rise to the design
of therapeutic, diagnostic, and theranostic agents (103). The reduction of
a porphyrin macrocycle to chlorin and bacteriochlorin results in profound
changes of symmetry and leads to a significant increase of absorption in
the red and near-IR regions, respectively (104–111). The reduction of
one pyrrole leads to chlorins and reduction of two pyrroles gives bacteri-
ochlorins and destabilizes the π system, increases the energy of highest occu-
pied molecular orbital (HOMO), and makes these molecules more prone to
oxidation. While energy of the lowest unoccupied molecular orbital
(LUMO) does not change significantly, a reduction of the HOMO–LUMO
energy gap and a red shift of the Q band are observed.
In collaboration with Arnaut and Pereira from the University of Coimbra
we have studied a wide range of halogenated porphyrins, chlorins, and bacte-
riochlorins as promising photosensitizers for PDT and photocatalysis. More
recently, a partnership has been started with Dumoulin and Ahsen from Gebze
Technical University in order to explore the potential biomedical application
of phthalocyanines (Fig. 6).
The structures of both porphyrin derivatives and phthalocyanines are
modified by the presence of various substituents in order to modulate their
hydrophobic/hydrophilic character and desired spectroscopic and photo-
chemical properties (45,105,108,112–114). The introduction of halogen
atoms (X ] Cl or F) in the ortho positions of the phenyl ring of the Ps
accelerates the intersystem crossing (ISC) to the triplet excited state and
maximizes the triplet quantum yield ΦT. Additionally, the steric interac-
tion between the halogen atoms and hydrogen atoms in β positions dimin-
ishes the tendency of porphyrin derivatives to aggregation and increases
their photostability (108,112,115). The presence of sulfonic, sulfoester,
Fig. 6 Scheme of possible structures of phthalocyanine- and bacteriochlorin-based compounds together with their electronic absorption
spectra.
or sulfonamide groups in the meta positions of the phenyl rings enables tai-
loring the hydrophilicity/lipophilicity of the photosensitizers and provides
an additional steric protection against oxidation. An investigation of the
effect of the polyethylene glycol substituent and its fluorinated analogue
(Fig. 6, right) on the Zn(II) phthalocyanine photodynamic properties was
also the aim of an ongoing research project (113). Fig. 6 shows the absorp-
tion spectra of the representative halogenated porphyrin, chlorin, and
bacteriochlorin as well as phthalocyanines studied in our laboratories.
Halogenated porphyrins possess a typical absorption spectrum with
an intense band around 400 nm and four other less intense bands of
lower energies (45,108,112,116,117). Various peripheral substituents
present in the phenyl ring cause only relatively small changes in the inten-
sity and energy of electronic transitions (116,118). A more pronounced
effect was observed by introducing metal ions (Zn2+, Pd2+, Co3+) into
the porphyrin ring because the symmetry of the free-base porphyrin
has changed from D2h to D4h, and the absorption spectra show only
two Q bands (119). In the series: porphyrin > chlorin > bacteriochlorin,
the HOMO–LUMO gaps are becoming progressively smaller, and the
Qy bands are shifted to longer wavelengths. This explains the red absorp-
tion of chlorins at 650 nm and the infrared absorption of the bacteri-
ochlorins at 750 nm. A less intense band around 519–529 nm,
normally labeled as Qx, is observed along the series and the Soret band
is situated at 400–420 nm for porphyrins and chlorins. For bacteri-
ochlorins, the Soret band splits into two bands with maximum absorption
wavelengths lower than 380 nm. Phthalocyanines possess characteristic
absorption with an intense π ! π* transition, referred to the Q band
(λmax 670 nm), and less intense transition in the UV (λmax 350 nm)
(120–124). Some efforts have been made to move the absorption further
into the NIR (117). Substitution of Zn(II) phthalocyanine with eight
hydroxylated sulfanyl moieties in nonperipheral positions led to an absorp-
tion at nearly 800 nm but also resulted in a decreased singlet oxygen gen-
eration. The strongest absorption of bacteriochlorins lies in the middle of
the “phototherapeutic window” (λ 750 nm), where the tissues are the
most transparent and the nontoxic photons are characterized by an energy
still sufficient to generate ROS efficiently. Thus, above listed advantages
situate this group of tetrapyrroles among the most promising photosensi-
tizers for PDT (73,125) and makes them well suited for other possible
applications (23,41,58,67,104,105,107,115,126–133).
6. ROS GENERATION MECHANISMS

6.1 Possible Deactivation Pathways (Jablonski Diagram)
Absorption of a photon leads to the formation of an “excited state” of the pho-
tosensitizing agent. The common features of the electronic excitation and
the consequent photochemical reactions may be described by involving three
electronic states: single ground state (S0), singlet excited states (S1, S2, S3), and
longer-lived triplet excited state (T1) as illustrated in Fig. 7.
Electronic absorption (A) occurs between vibrational and rotational
energy levels of the excited singlet states. Immediately after a photon absorp-
tion, several photoprocesses occur, but the most likely deactivation pathway
is a relaxation to the lowest vibrational energy level of the first excited state.
This very fast process (1015 s) is defined either as internal conversion (IC) or
vibrational relaxation (VR). An excited molecule (1Ps*) exists in the lowest
excited singlet state on the order of nanoseconds before relaxing to the
Fig. 7 Photophysical and photochemical processes illustrated by a modified Jablonski

diagram: The photosensitizer in singlet excited (S1) state may undergo intersystem
crossing (ISC) to an excited triplet state (T1) and then generate ROS.
ground state. If relaxation from S1 is accompanied by emission of a photon,

the process is known as fluorescence (F). Several other relaxation pathways
compete with the photoprocesses described earlier, and among them, the
most important one is a nonradiative process to the lowest excited triplet
state (T1) known as intersystem crossing (ISC). The latter event may result
either in emission of a photon from the triplet excited state (T1) through
spin-forbidden phosphorescence (Ph) or more likely in the photochemical
reactions. It is noteworthy that molecules in the excited states are generally
stronger oxidizing and reducing species than their analogues in the ground
states.
6.1.1 Fluorescence
Fluorescence is not a desired pathway for the PDT application but can serve
as a tool for tumor detection and imaging. Moreover, it is very convenient to
study Ps cellular uptake and intracellular localization by fluorescence and/or
confocal microscopy. Furthermore, a variety of currently available fluores-
cent probes enable exploring mechanisms of cell death in PDT and moni-
toring singlet oxygen, superoxide ion, and hydroxyl radical formation
followed by Ps illumination inside the cell (134–137).
The presence of halogen atoms (Cl or F) in the structure of photosen-
sitizers studied in our laboratories reduces the fluorescence quantum yield
(ΦF) (41,104,108,112,115,138). This significant change is explained by
the internal heavy atom effect. ΦF values determined for halogenated por-
phyrins and bacteriochlorins are remarkably lower (especially those bearing
Cl atoms) than for nonhalogenated molecules. On the other hand, fluori-
nated chlorins exhibit the most intense fluorescence (ΦF approaching
40%). Considering their relatively high values of singlet oxygen generation
(ΦF¼ 60%), these chlorin-based compounds were proposed as potent
theranostic agents that can serve in both diagnostic and therapeutic
approaches at the same time (111).
6.1.2 Generation of Triplet Excited State

The most important parameter in the design of ROS generating systems for
PDT is the triplet excited state of Ps which sensitizes the formation of ROS.
Spin–orbit coupling induced by halogen atoms present in Ps affects ISC
quantum yields and triplet state lifetimes. Laser flash photolysis was used
as a tool for detecting the Ps triplet state as well as for determining the triplet
state lifetimes. A triplet quantum yield close to unity was obtained for
dichlorinated sulfonamide porphyrins (41). Long-lived triplet states increase
the probability of interaction of 3Ps with molecular oxygen to generate

ROS. However, the presence of oxygen reduces significantly the Ps triplet
lifetimes. The quenching rate constants (kq) determined for bacteriochlorins
are generally of higher magnitude than those of the porphyrins and chlorins.
Spin statistics associated with interaction between one triplet excited state
(3Ps) and another triplet state (3O2) require kq higher than 1/9 kdiff for a pro-
cess where quenching proceeds via a singlet channel. However, if both the
singlet and the triplet channels contribute to the deactivation of the triplet
excited state by molecular oxygen, then it is expected that kq 4/9 kdiff. The
values of kq for porphyrins are 1/9 kdiff confirming a singlet deactivation
channel by energy transfer to molecular oxygen. Chlorins possess a slightly
higher than 1/9 kdiff quenching constant, and for bacteriochlorins, the
quenching constant is higher than 1/9 kdiff and lower than 4/9 kdiff. The
charge-transfer (CT) mechanism observed for bacteriochlorins enables their
participation in the photoinduced electron transfer reactions. Ps triplet
excited states have sufficiently long lifetimes to initiate bimolecular reactions
with substrate molecules, namely, by energy transfer and electron/hydrogen
transfer reactions which lead to generation of ROS according to Type II and
Type I mechanisms, respectively.
6.2 Type II Mechanism

A Type II reaction consists of the energy transfer from a Ps triplet excited
state to the ground state of molecular oxygen. Most PDT agents operate
via a Type II mechanism (Eqs. 2–4) with singlet oxygen (1O2,1Δg) gener-
ation leading to oxidation of any nearby biological species with cytotoxic
consequences (94).
Ps + hν !1 Ps∗ (2)
1
Ps∗ ! Ps∗
3
(3)
3
Ps∗ +3 O2 ! Ps +1 O2 (4)
The unoccupied π*2p orbital of singlet oxygen makes it highly reactive
toward electron-rich compounds. 1O2 oxidizes lipids to hydroperoxides
and reacts with amino acids (tryptophan, tyrosine, histidine, methionine, cys-
teine, cystine) to form hydroperoxides and endoperoxides. Endoperoxides
are also products of oxidation reaction between 1O2 and deoxyguanosine
present in DNA. These reactions are later responsible for cellular toxicity,
however cytotoxic effects induced by 1O2 is not dependent on the activity
of the antioxidant enzymes.
The generation of 1O2 (1Δg) requires an energy of a minimum of

22.5 kcal mol1. Both halogenated porphyrins and bacteriochlorins triplet
energies exceed this limit and can thus generate singlet oxygen efficiently.
The most common mechanism for 1O2 generation is quenching of the Ps
excited states by molecular oxygen. In fact singlet oxygen photosensitization
can occur by energy-transfer quenching of both excited singlet and triplet
states. While halogenated and sulfonated porphyrins follow mostly the
energy transfer channel, in the case of bacteriochlorins a quenching process
through charge transfer interactions takes place. The decrease of Ps oxida-
tion potentials results in the increase of quenching constant (kq). The CT
mechanism becomes more competitive against a non-CT process, which
is reflected by the decrease of the singlet oxygen quantum yield (ΦΔ). In
the absence of a charge transfer process (kq ¼ 1/9 kdiff), the generation of sin-
glet oxygen should be equal to unity (ΦΔ ¼ 1). When a CT process is
involved, the quenching rate is higher (kq ¼ 4/9 kdiff), and the singlet oxygen
quantum yield is reduced approximately by 25%. Therefore, porphyrins
generate essentially more singlet oxygen than chlorins. On the other hand,
the bacteriochlorins quenching constant (1/9 kdiff < kq < 4/9 kdiff) is
higher than kq obtained for chlorins, and only a part of the triplet excited
states that are quenched by molecular oxygen lead to singlet oxygen. The
rest of the bacteriochlorins triplet excited states are involved in the Type
I photochemical reactions (139).
6.3 Type I Mechanism

The first oxygen-centered radical generated according to Mechanism I is the
superoxide ion (O2 • ). It is formed when an electron is captured by one of
the π*2p orbitals of oxygen either in the presence of a reducing agent (Eqs.
5–6) or directly from an excited photosensitizer (Eq. 7).
2Ps∗ + NADðPÞH ! 2Ps• + NADP + + H + (5)

Ps• + O2 ! Ps + O2 • (6)
Ps∗ + O2 ! Ps• + + O2 • (7)
Depending on the environment, the reduction potential of superoxide

ion may change. In aqueous solution, O2 • is a weak oxidizing agent (see
Fig. 1), but it is still able to oxidize, for example, ascorbic acid (140). On
the other hand, it acts as a strong reducing agent enabling reduction of iron
complexes in cytochrome c and ferric EDTA (141).
The subsequent formation of a perhydroxyl radical proceeds according

to Eq. (8).
Ps• + + O2 • ! Ps• + HO2 • (8)
The perhydroxyl radical is a more potent oxidant than superoxide ion,
thus it is able to oxidize O2 • resulting in the H2O2 formation according to
Eq. (9).
HO2 • + O2 • + H + ! O2 + H2 O2 (9)
2O2 + 2H ! O2 + H2 O2
• +
(10)
HO2 • + HO2 • ! O2 + H2 O2 (11)
The rate of superoxide disproportionation at pH 7.4 is 2.4 105 M1 s1
(Eq. 10) (142), and the disproportionation rate of the perhydroxyl radical is
8.1 105 M1 s1 (Eq. 11) (143). Dismutation often catalyzed by SOD
(Eq. 10) or one electron reduction of O2 • , leads to H2O2 formation.
Hydrogen peroxide is characterized by a much longer half-life than other
ROS. Moreover, in contrast to other ROS, hydrogen peroxide can pass
through biological membranes and cause damage in other cellular compart-
ments (3). H2O2 is detoxified by the enzyme catalase, resulting in formation
of water and molecular oxygen, or may react further either with superoxide
ion (Eq. 12) or perhydroxyl radical (Eq. 13) and form highly reactive
hydroxyl radicals. They easily oxidize major biologically relevant molecules
such as proteins, lipids, carbohydrates, and DNA, but can also inactivate nat-
ural antioxidants (e.g., tocopherol).
H2 O2 + O2 • ! HO• + O2 + OH (12)
H2 O2 + HO2 • ! HO• + O2 + H2 O (13)
In view of possible electron transfer from the photosensitizer in the trip-
let excited state discussed earlier, the mechanism of hydroxyl radical gener-
ation through photocatalysis was proposed:
H2 O2 + Ps∗ ! Ps• + + H2 O2 • ! HO• + Ps• + H2 O (14)
Although hydrogen peroxide is a worse electron acceptor than molecular
oxygen, an electron transfer reaction from the Ps* to H2O2 may occur if the
photosensitizer triplet state lifetime is long enough (105). The intermediate
H2 O2 • , presented in Eq. (14) may be accessed by a Franck–Condon tran-
sition in a dissociative electron attachment to H2O2 which gives OH%
and OH (144). The long triplet lifetimes of bacteriochlorins enable
photoinduced electron transfer to H2O2 with the rate constant of

kH2O2 ¼ 3 107 M1 s1 (105). This is a much higher value than the rate
constants of reactions presented in Eqs. (12) and (13). The formation of
H2O2 is more favorable in acidic solutions, but only in the presence of light
hydrogen peroxide leads to the degradation of the photosensitizer. Taking
all this information together, it is possible to conclude that photocatalysis can
be a dominant mechanism for hydroxyl radical generation by bacteri-
ochlorins, even if the electron transfer to H2O2 competes with quenching
by O2, and there is a need of H2O2 to initiate this reaction. Considering the
facts that tumor microenvironment is highly hypoxic (58) and cancer cells
have a different redox status than normal cells, manifested mainly by
increased level of H2O2 (145,146) and other ROS, this mechanism should
be even more favorable in vivo.
Under biological conditions where ferrous ions are present, hydrogen
peroxide can undergo the Fenton reaction producing hydroxyl radicals
by one-electron reduction of hydrogen peroxide (Eq. 15), and the ferric
iron produced can then be reduced back to the ferrous state by superoxide
(Eq. 16). The combination of these two reactions is termed the iron-
catalyzed Haber–Weiss reaction (147).
H2 O2 + Fe2 + ! HO• + OH + Fe3 + (15)
Fe3 + + O2• ! Fe2 + + O2 (16)
Hydrogen peroxide and the sensitizer radical anion can react to form
hydroxyl radicals and hydroxide anion (Eq. 17). This is equivalent to the
one-electron reduction of hydrogen peroxide mediated by Fe2+, discussed
earlier, but in this process, Ps• acts as a reducing agent (41).
Ps• + H2 O2 ! Ps + HO• + OH (17)
The results of our completed and ongoing projects show that the pho-
todynamic effect does not correlate with singlet oxygen quantum yields, but
rather with the ability of Ps to take part in the Type I mechanism
(45,104,109,110,119). It is clear that the most desirable situation is wherein
photodynamic action is mediated by both electron transfer and the energy
transfer mechanisms. However, PDT is more effective if Mechanism I is
operative and completed with generation of highly reactive hydroxyl radi-
cals (41,58,67).
6.4 Role of Ascorbate in ROS Generation

As illustrated in Fig. 3, ascorbate (vitamin C) takes part in the cell defense
system and plays an important role in several physiological processes such
as cell division, cell death, light signaling, and pathogen responses. It acts
as an antioxidant when it reduces hydrogen peroxide or metal ions and gen-
erates ROS through the following reactions (AH is ascorbate and DHA is
dehydroascorbate):
2Fe3 + + AH ! 2Fe2 + + DHA (18)

2Fe 2+
+ 2H2 O2 ! 2Fe 3+ •
+ 2HO + 2OH (19)
The sensitizer radical cation, one product of the photoinduced electron
transfer reaction illustrated in Eq. (7), can be further reduced by ascorbate in
a polar environment as illustrated below.
Ps• + + AH ! Ps + AH (20)

Then either superoxide ion or a perhydroxyl radical reacts with ascorbic
acid or ascorbate and generates hydrogen peroxide. A rate constant deter-
mined for reaction 21 is k ¼ 3 105 M1 s1 (148).
HOO• =O2 • + AH2 =AH ! H2 O2 + A• (21)
Dismutation of superoxide ion by SOD takes place according to Eqs. (22
and 23).
Mn3 + SOD + O2 • ! Mn2 + SOD + O2 (22)

Mn 2+
SOD + O2 •
+ 2H ! Mn
+ 3+
SOD + H2 O2 (23)
The reactions described earlier are correct if the Type I mechanism
occurs. We have recently shown that ascorbate at high concentrations
(0.6 mM) efficiently quenches hydroxyl radicals in solution and improves
the stability of the ROS-generating agents. On the other hand, relatively
low concentrations of ascorbate resulted in an increased tendency to
photodegradation of the Ps studied (104). Ascorbate can also interact with
photoproducts of the Type II mechanism. In such a situation singlet oxygen
reacts with ascorbate and forms hydrogen peroxide and dehydroascorbate:
1
O2 + AH + H + ! H2 O2 + DHA (24)
The rate constant of this reaction is k ¼ 3 108 M1 s1, thus ascorbate is
a more efficient reducing agent for singlet oxygen than for superoxide ion,
Vis–NIR
O2
Antioxidant
R F
F ROS
R
F
N F
M N
1
F F
N
N O2 HO H
O O
R F HO
Living cells
F R O2 •– –
O OH
Ascorbate
SOD
Prooxidant
X H2O2 HO•
GP Dead cells
CAT
H2O
H2O + O2
Fig. 8 The role of ascorbate in PDT mediated by bacteriochlorins and NIR light.
and its antioxidant effect should be more pronounced when singlet oxygen is
responsible for the photodynamic effect (Fig. 8) (149).
In biological in vitro experiments, ascorbate as a reducing agent can
regenerate photosensitizers following the Type I mechanism and act as a
prooxidant. On the other hand, AH– is also a facile singlet oxygen quencher
and may serve as an antioxidant. The predominant role played by ascorbate
in oxidative stress depends strongly on the origin of cells and types of the
processes involved at the cellular level. For example, we have demonstrated
that the addition of ascorbate and the inhibition of SOD markedly increases
the photodynamic effect toward A549 cells, but not toward CT26 cells.
Moreover, the inhibition of catalase and the depletion of the glutathione
pool also led to a potentiation of PDT in A549 cells (16). However, in order
to develop enhanced PDT sensitizers and design efficient PDT treatment
regimens, the approach should be one of first considering tissue and intra-
cellular localization, instead of trying to maximize singlet oxygen quantum
yields in in vitro tests (150).
7. SUBCELLULAR LOCALIZATION OF
PHOTOSENSITIZERS IN CELLS
Both Type I and Type II reactions may occur simultaneously,
depending on the type of photosensitizer, its oxidation potentials, and
amount of oxygen present in the cell/tissue (42,58,104,151). Due to the
reduced lifetime in biological media, the maximal distance of hydroxyl rad-

ical cytotoxic action is estimated to be only 1 nm. The intracellular locali-
zation of the Ps remarkably determines the site of cellular damage induced
via the Type I photoreactions, because oxygen-centered radicals are rather
promiscuous and unselective species, reacting only in the organelle where
they are generated. On the other hand, the singlet oxygen-induced photo-
dynamic effect is not restricted to one cellular compartment and radius. 1O2
can explore all the cell volume, and it is a more selective oxidant than
oxygen-centered free radicals (152,153).
Lipophilicity is an important factor in predicting the ability of molecules
to cross cellular membranes and also correlates with the bioactivity of drugs.
It can be quantified by the partition coefficient between an organic solvent
and water (log Pow). We have developed a library of photosensitizers of var-
ious polarity, from very hydrophilic ones (bearing a SO3H group, log
Pow ¼ 1.7) to very hydrophobic ones (containing a sulfonamide heptyl
substituent, log Pow > 4). Amphiphilic photosensitizers can diffuse across
the plasma membrane and redistribute between the membranes of cellular
organelles. On the other hand, association of lipophilic photosensitizers with
low-density lipoproteins may facilitate cellular uptake, but it is believed that
the photosensitizer leaves the LDL particle at the plasma membrane and dif-
fuses into the cytoplasm (154). However, very lipophilic photosensitizer
must be delivered to cancer cells using an appropriate formulation/vehicle
(67,113,155–159). Depending on the lipophilicity/hydrophylicity and
charge, photosensitizers can localize in mitochondria, ER, lysosomes,
plasma membrane, and Golgi apparatus (42,160,161) and even in the
nucleus (162,163).
Our studies showed that sulfonated tetraphenyl macrocyles have a higher
affinity for an aqueous environment (log Pow < 1), while sulfonamide pho-
tosensitizers with higher log Pow tend to localize in a more lipophilic envi-
ronment. The intracellular fluorescence changes for hydrophobic and
hydrophilic porphyrins, chlorins and bacteriochlorins were followed by
fluorescence imaging. Confocal microscopy was used to determine the sub-
cellular localization of difluorinated sulfonamide porphyrin (F2PMet) by
overlapping the fluorescence image of the photosensitizer with fluorescent
probes specific for various organelles namely ER, mitochondria, and lyso-
somes. Fluorescence micrographs of A549 cells showing F2PMet intracellu-
lar distribution in various cellular organelles are presented in Fig. 9.
Fig. 9 reveals similar diffuse patterns for F2PMet and its analogous
bacteriochlorin (F2BMet). According to the topographic analysis, their
Endoplasmatic reticulum Mitochondria Lysosomes
Intensity Intensity
250 250 Intensity
200 250
200
150 200
150
100 150
100
500 100
500
0 500
0
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0
Distance (µm) Distance (µm) 0 5 10 15 20 25 30
Distance (µm)
Intensity (F2PMet) Intensity (ER-Tracker) Intensity (F2PMet) Intensity (Mito-Tracker) Intensity (F2PMet) Intensity (Lyso-Tracker)
Fig. 9 Intracellular localization of F2PMet in A549 cells. Cells were marked with dyes for endoplasmic reticulum (ER-Tracker), lysosomes (Lyso-
Tracker), and mitochondria (Mito-Tracker). The green topographic profile corresponds to the emission of the tracker and the red to the por-
phyrin emission.
subcellular localization is more consistent with the ER and mitochondria,

while overlap with the Lyso-tracker almost does not exist. Coordination
of F2PMet with Zn2+ increases the hydrophilic character of this porphyrin
and decreases the hydrophobic interaction with the membranes. Therefore,
ZnF2PMet is diffused throughout the cell and bound to membrane proteins.
On the other hand, taking into account the high affinity of F2PMet to inter-
act with cell membranes and to be localized in a lipid environment, binding
to the Golgi apparatus is not excluded. A group of bacteriochlorins-bearing
dicyano moieties possesses subcellular localization comparable to F2PMet/
F2BMet (128), while bacteriochlorins with a geminal dimethyl group local-
ize in lysosomes or in mitochondria (164). Ps localization depends not only
on its properties, but also on the type and origin of the studied cells. Our very
recent findings indicate that F2BMet possesses different localization in Lewis
lung carcinoma (LLC) cells than in A549 cells, as discussed above.
Depending on the time of incubation, Ps is found in lysosomes and ER
but not in the mitochondria of LLC cells. Photofrin®, the most commonly
used porphyrin-based Ps, enters mainly the ER/Golgi apparatus and at other
perinuclear sites of A431 cells (165,166). The commercially available
chlorin (Foscan®) after 3 h of incubation with human breast adenocarci-
noma cells (MCF-7) can be found both in the ER and in the Golgi appa-
ratus, but after 24 h, it extrudes from the Golgi and is essentially in the
ER, with only a weak distribution in the mitochondria (167). Verteporfin
also localizes in the mitochondria and in the perinuclear area, where the
nuclear membrane and ER are located (168,169). Most of the photosensi-
tizers which were confirmed to localize in both ER and mitochondria have
demonstrated excellent in vitro PDT efficacy. The cationic porphyrin,
5,10,15,20-tetrakis(N-methyl-4-pyridyl)-21H,23H porphyrin (TMPyP)
has an unusual mode of photodynamic action. It localizes in the nuclei
and irradiation leads to damage to guanine residues in DNA, triggering apo-
ptosis (170). Fig. 10 illustrates cellular targets for Ps localization and action.
Photosensitizers localized in the lysosomes have shown to induce cell
death via the release of lysosomal enzymes in the cytosol, or via the
relocalization of the sensitizer after irradiation of other targets (171). Inter-
esting comparative studies of two photosensitizers (chlorin NPe6 and
WST11), both with a tendency to localize in lysosomes, but with very dif-
ferent abilities for ROS generation, were published (49). Chlorin NPe6 pro-
duces mainly singlet oxygen upon irradiation and localizes in lysosomes.
WST11 also localizes in lysosomes (49) but produces only oxygen radicals
upon irradiation (172). The efficacy of these lysosomal photosensitizers is
TNFR
Cytoplasm
TNF
Lysosome
Caspases
Apoptosis
Cell membrane
ROS, cytochrome C, AIF, Bcl-2
Nucleus Bid
Ribosomes
Bcl-2
Necrosis
Mitochondria
Ca2+
Bax/Bak ROS↑, ATP↓, Ca2+↑
Golgi apparatus
Endoplasmatic reticulum Autophagy

Prodeath Prosurvival
Fig. 10 Cellular targets for a Ps localization and ROS generation and consequent cell
death mechanisms in PDT. The mode of cell death observed after PDT depends on
the intracellular localization of the Ps- and PDT-related damage to these organelles.
assigned to both promotion of autophagic stress and suppression of auto-

phagic prosurvival functions (49).
8. ROS-MEDIATED BIOLOGICAL MECHANISMS

Most of the photosensitizers including halogenated porphyrins,
chlorins, and bacteriochlorins, after excitation with the appropriate light
source, produce a sufficient amount of ROS to trigger destruction of cellular
and vascular components. ROS activate the expression of transcription fac-
tors and cytokines, and release a number of mediators responsible for the
process of cell death, which can occur either by apoptosis and/or necrosis.
As mentioned earlier, autophagy can also take place, but it is not classified as
a cell death mechanism. These three pathways may occur simultaneously in
the same cell population as presented in Fig. 6. However, under in vivo con-
ditions, ROS-mediated vascular shut down and activation of immune
response play a crucial role in the therapeutic outcome of PDT.
8.1 Apoptosis
Photogenerated ROS trigger a cascade of molecular events that contribute
to an apoptotic cell death. Apoptosis is an irreversible direct cell death mech-
anism, considered by some authors to play a predominant role in killing cells
after PDT (173). It is termed as “programmed cell death” because it enables
a balance between survival and death mechanisms. Apoptosis might be
either suppressed or defective, eventually leading to an uncontrolled cell
proliferation. The mechanism of apoptosis is mediated by a cascade of events

that ultimately leads to irreversible caspase activation and cell fragmentation
(174). Among several apoptotic indicators, the most characteristic signs are
changes in cellular morphology, such as cell surface blebbing, shrinkage,
condensation of chromatin and DNA fragmentation, and formation of apo-
ptotic bodies triggered by both the apoptosis inducing factor and the necrosis
inducing factor (TNF). The primary role of mitochondria in the apoptotic
signaling pathway is the regulation of release of several proapoptotic mole-
cules to cytoplasm. Among them, cytochrome c is a key protein due to its
crucial role in mitochondrial electron transport. It activates the apoptotic
protease activating factor 1 (Apaf-1) and the caspase cascade, a crucial group
of proteases triggered in the apoptotic response (175). Caspases are cysteine
proteases that are inactive until activated by an apoptotic signal. The caspases
are characterized as initiators (CASP8 and CASP9) or executioners (CASP3,
CASP6, and CASP7). The apoptotic regulating protein Bid, a natural sub-
strate of CASP8, is the nexus between extrinsic and intrinsic pathways. ROS
are able to degrade antiapoptotic protein Bcl-2 and lead to the over-
expression of Bax. The Bcl-2 family of proteins that regulates apoptosis is
responsible for controlling the permeability of the mitochondria membranes
(176). These proteins are also the targets for photosensitizers localized in
mitochondria or ER (177). Foscan® and Pc4, mitochondrial photosensi-
tizers, lead to Bcl-2 photodamage and trigger a rapid apoptotic response
(178). PDT Foscan and Pc4 result in the inhibition of respiration processes
initiated by caspase-3 and in a cleavage of the proapoptotic protein Bid that
forms a t-Bid promoting the release of cytochrome c (179). Chiu and
Olenick reported that PDT with Pc4 as photosensitzer rapidly induces apo-
ptotic cell death confirmed by release of cytochrome c already 15 min after
photodynamic treatment of mouse lymphoma cells (LY-R) (180). More-
over, Vantieghem and coworkers demonstrated the role of cytochrome c
during photodynamic action with hypericin as the photosensitizing agent
by overexpression of Bcl-2 proteins, which also can alter mitochondrial
membranes permeability and release proapoptotic factors (181). It is further
demonstrated that by adding Bcl-2 antagonists it is possible to enhance the
photodynamic effect at low PDT doses (182).
8.2 Necrosis
Necrosis occurs above the threshold of resistance of cells treated with ROS
or with other nonphysiological conditions that are responsible for the
destruction of antioxidant enzymes. Necrotic cell death is often observed

after PDT with applied high light and Ps doses (20). Necrosis has distinct
morphological features, and it is accompanied by rapid permeabilization
of the plasma membrane. It is associated with a significant decrease in the
ATP level resulting from ion imbalance due to depolarization of the mito-
chondrial membrane. The volume of the cell is increased, the membrane
integrity is lost, and consequently the passive influx of Ca2+, Na+, and water
to the cellular interior is observed. Typical necrotic changes are also a high
concentration of Ca2+ arising from their influx from the extracellular envi-
ronment and the outflow from the ER undergoing destruction. Activation
of many DNA nucleases and release of lysosomal hydrolytic enzymes con-
sequently lead to the total cell lysis (18,69). In contrast to necrosis caused by
very extreme conditions, there are many examples when this form of cell
death may be a normal physiological and programmed mechanism. Various
factors including ROS generated during PDT can trigger both apoptosis and
necrosis in the same cell population. In addition, antiapoptotic mechanisms
including Bcl-2/Bcl-x proteins are equally effective in protection against
both apoptosis and necrosis. Therefore, necrosis, along with apoptosis,
may be a specific form of the execution phase of programmed cell death,
and there are several examples of necrosis during a normal tissue renewal,
and immune response. However, the consequences of necrotic and apopto-
tic cell death for a whole organism are quite different. In the case of necrosis,
cytosolic constituents that spill into extracellular space through a damaged
plasma membrane may provoke an inflammatory response. Moreover, the
inflammatory response caused by necrosis may have obvious adaptive signif-
icance such as induction of a strong immune response (47,183).
Preliminary studies on the cell death mechanisms after PDT with
redaporfin were recently reported (184). Our experiments revealed the pres-
ence of apoptotic, late apoptotic, and necrotic cell populations after PDT.
More extensive necrosis was observed for high light doses, in contrast to cells
that were exposed to lower light doses where more apoptotic cells were
observed. Hence, the cell death pathways are modulated by the light dose.
PDT performed at low light doses tends to trigger apoptosis, while necrosis
becomes the predominant mechanism of cell damage at high light doses.
Moreover, signaling pathways, such as death receptors, kinase cascades,
and mitochondria also participate in both apoptosis and necrosis, and by
modulating these pathways, it is possible to switch between apoptosis and
necrosis (183). An interesting example of necrosis as a dominant cell death
pathway in PDT has been demonstrated in glioblastoma (185). In these
studies, the necrotic mechanism induced by PDT in human glioblastoma

cells depends on the receptor-interacting proteins (RIP3 and RIP1), but
is not correlated with the TNF-depended proteins such as caspase-8 and
flavin adenine dinucleotide.
8.3 Autophagy
Autophagy is a cellular process that involves the formation of a type of vac-
uoles called autophagosomes that devour cellular organelles and intracellular
proteins. The process ends with the fusion of the outer membranes of
autophagosomes and lysosomes. Hydrolases and lipases in the lysosomal
lumen degrade the autophagic contents, with the releasing amino acids, fatty
acids, and nucleosides proceeding into the cytosol (186). While autophagy is
usually considered as a prosurvival mechanism, if it is more excessive it may
be associated with the cell death (48). It is not, however, another cell death
mechanism because death is induced by other mechanisms that evoke an
autophagic response; therefore, it is often reported as “death with
autophagy” (187). The role of ROS as signaling molecules in starvation-
induced autophagy was recently described. It was shown that starvation
stimulates formation of ROS, specifically H2O2. These oxidative conditions
are essential for autophagy, as treatment with antioxidative agents abolished
the formation of autophagosomes and the consequent degradation of pro-
teins (188). As reported earlier, Bcl-2 damage and activation of Bid triggers
apoptosis, but Bcl-2 acts as an antiautophagic protein as well. It forms a com-
plex with the proautophagic protein Beclin-1; therefore, the loss of Bcl-2
may result in the initiation of autophagy (189). Autophagy offers protection
from PDT and inhibition of autophagy promotes PDT efficacy (190). A role
of autophagy in promoting a survival pathway was especially observed at low
light doses, inducing a moderate level of ROS (191). In contrast, a high dose
of light and a greater quantity of ROS activates prodeath mechanism (48).
The aim of our ongoing studies on LLC cells and tumors is to investigate the
infuence of the Ps localization, Ps drug dose and light dose on the prosurvival
and prodeath functions of authophagy and to correlate them with the PDT
efficacy.
8.4 ROS-Mediated Vascular Occlusion

The excess of ROS production in the vasculature implicates several changes
including oxidation of proteins, DNA damage, change of the nitric oxide
level, and induction of proinflammatory responses. Moreover, ROS are
important intracellular signaling molecules that regulate the vascular func-

tion by modulating vascular cell contraction, migration, growth/apoptosis,
and extracellular matrix protein turnover, which contribute to vascular rem-
odeling. Vascular damage and blood flow stasis are consequences of PDT,
especially performed on solid tumors. Solid tumors cannot grow larger than
1 mm3 without developing a vascular network. The number and size of
blood vessels increase proportionately to the need for nutrients and increased
concentration of oxygen. Moreover, tumor tissues develop new blood ves-
sels from the preexisting vasculature (angiogenesis) for maintaining the
blood supply. Among various angiogenic factors, VEGF (vascular endothe-
lial growth factor) plays a critical role in vascular formation. It has been first
characterized for its ability to induce vascular leakage and permeability, and
to promote vascular endothelial cell proliferation (40,50,54,55).
It is now clear that vascular damage contributes significantly to PDT effi-
cacy. We have recently demonstrated that vascular-targeted PDT led to bet-
ter long-term response than cellular-targeted PDT (C-PDT). In 44% of
animals bearing S91 melanoma cells, we observed tumor regression lasting
at least 1 year after therapy. V-PDT led to deep and chronic hypoxia,
whereas C-PDT caused transient hypoxia, that was quickly reversed (58).
Extremely low pO2 lasting for several days after vascular-targeted PDT favor
long-term tumor responses, in contrast to mild and transient hypoxia, that in
C-PDT leads to strong pO2 compensatory effects and frequent tumor
regrowths. We have suggested that the existence of a very strong hypoxia
is a good predictor of the therapy outcome. Partial dysfunction of the blood
flow does not guarantee a good long-term tumor response to PDT and
might even have a stimulatory effect on tumor growth, especially with
simultaneous increase in VEGF. To support this hypothesis, we have
selected the group of animals in which the pO2 level dropped to 0 mm
Hg. In this subset of data, the cure rate drops up to 80% of the animals treated
with V-PDT. These results clearly demonstrate the crucial role of strong
hypoxia in melanoma tumors after PDT treatment in the final therapeutic
outcome.
Even more spectacular results were obtained for V-PDT with redaporfin
encapsulated in pluronic micelles. 100% cures of mice bearing B16 F10 mel-
anoma tumors and no observable adverse effects were observed. This is the
first time that 100% of permanent cures were achieved with the very aggres-
sive and strongly pigmented B16F10 melanoma tumors (157). Redaporfin
also showed a high efficacy in the treatment of BALB/c mice with subcu-
taneously implanted colon (CT26) tumors. Vascular PDT with 1.5 mg kg1
redaporfin and a light dose of 74 J cm2 led to the complete tumor regression
in 83% of the treated mice (23,66,192). We assigned this efficacy to the fact
that redaporfin is relatively photostable and generates a high quantity of sin-
glet oxygen and oxygen-centered radicals.
8.5 Antitumor Immune Response

PDT, if designed properly, initiates production of an amount of ROS suf-
ficient to induce strong and acute inflammatory reactions that can activate
the immune system to establish the long-term antitumor memory
(18,19,193). In other words, PDT-induced oxidative stress triggers a
response that encourages the defense mechanisms associated with the
inflammatory process to participate in the eradication of the treated tumor.
Many factors of the innate and adaptive immune systems take part in the
immune response after photodynamic action (20,46,60,192,194,195).
PDT starts with neutrophils recruitment (196) and then triggers macro-
phages, leukocytes, monocytes, mast cells, and dendritic cells. These factors
are involved in primary immunomodulatory effects that activate
granulocytes and induce the generation of proinflammatory cytokines such
as IL-1β, IL-6, IL-10, and TNF-α. They are important mediators of the
photodynamic efficacy because they regulate the degree and mode of tumor
cellular death, and vascular damage. The activation of the complementary
system also induces the production of thromboxane, histamine, prostaglan-
dins, and leukotrienes that usually affect endothelial cells generating
postirradiation tumor ischemia. The progressive development of inflamma-
tion can also promote generation of adaptive immunity. A remarkable
outcome of the PDT is antitumor immunity mediated by antigen-specific
T-cells. The mechanisms of induction of adaptive immune response
are related to activation of antigen-presenting cells, such as mature
dendritic cells, after PDT. These molecules can interact with T-cells
through cross-binding between an antigen containing major histocompati-
bility complex molecules (MHC-I/MHC-II) and T-cell receptors
(19,20,60,61,193–195). Our PDT study with the difluorinated sulfon-
amide bacteriochlorin (F2BMet, redaporfin) showed a combination of local
oxidative stress in the target tissue capable of eliminating the primary tumor,
with a systemic immune response. This antitumor protection elicited by the
optimized PDT regimen is assessed by tumor rechallenges and by resistance
to the establishment of metastasis after intravenous injection of colon carci-
noma cells. The optimized treatment regimen led to the cure of 87%
BALB/c mice. Cured mice rechallenged over 3 months later with CT26
cells rejected the tumor cells in 67% of the cases. The strong immune
response is triggered by the high local inflammation after PDT, evidenced by

the recruitment of lymphocytes (23). These findings demonstrate a potential
clinical impact because such a combination of local and systemic effects
opens a new era of photoimmunotherapy of cancer.
9. METHODS OF ROS DETECTION IN PDT

9.1 1270 nm Phosphorescence
The direct method for monitoring 1O2 formation involves the detection of
1270 nm phosphorescence, assigned to the electronic transition from the
first excited singlet state of oxygen (1Δg) to the ground triplet state
(3 Σ g ). Our study of the electronic excitation of porphyrins, chlorins, and
bacteriochlorins in ethanol led exactly to emissions at 1270 nm (Fig. 11) with
lifetimes of ca. 14 μs, which is in good agreement with a 15-μs lifetime typical
for singlet oxygen phosphorescence in ethanol (197). Fig. 11 also presents
the emission spectra taken between 1250 and 1300 nm, typical for singlet
oxygen generated from photosensitizers and studied in our laboratories.
The singlet oxygen quantum yields of porphyrins, chlorins, and bacte-
riochlorins in ethanol were obtained with a procedure described in the lit-
erature (105,108,112), using phenalenone as a reference (ΦΔ ¼ 0.95 for
phenalenone in ethanol) (198). Fig. 11 (on the right) presents the laser
energy dependence of singlet oxygen emission obtained for some of our
photosensitizers (Cl2PEt is a chlorinated sulfonamide porphyrin, Cl2BEt
and F2BMet are sulfonamide bacteriochlorins either chlorinated or fluori-
nated). Fig. 11 also illustrates a kinetic mechanism of photosensitizer
quenching, together with singlet oxygen and superoxide anion generation.
As the triplet state energy increases, the rate constant of oxygen quenching
decreases and the singlet oxygen quantum yield increases. The generation of
a charge transfer (CT) state (3Sδ+ … O2δ) from quenching of the triplet
excited state of a sensitizer by oxygen can accelerate the rate of triplet
quenching by oxygen and reduce the singlet oxygen quantum yield
(139). Bacteriochlorins are characterized by significantly higher quenching
rate constants than porphyrins, suggesting the existence of a charge transfer
channel for these photosensitizers and higher probability of superoxide ion
generation. On the basis of just this small selection of our results, two general
conclusions can be drawn. First, porphyrins exhibit higher values of singlet
oxygen quantum yield than chlorins and bacteriochlorins. Second, the chlo-
rinated photosensitizers yield higher amounts of singlet oxygen than fluori-
nated photosensitizers.
0.012
0.04
l = 1270 nm 0.030 2 μs k1CT / kn = 0.02
O2 phosphorescence int.
5 μs Phenalenone
Phosphorescence int.
0.03 0.025 10 μs
0.01 Cl2PEt
15 μs
0.020 20 μs Cl2BEt
25 μs
O2 phosphorescence (a.u)
0.02
0.015 30 μs F2BMEt
0.008
k1CT / kn = 0.1
0.01 0.010
0.005 0.006
1
0.00
−5 −5 −5 −5 −5
0.000
0 1⫻10 2⫻10 3⫻10 4⫻10 5⫻10 1250 1260 1270 1280 1290 1300
Time (s) l (nm) 0.004
k1CT / kn = 1
1
k1 kCT
3
S* + O2 (3Σg–) {3S* O2 (3Σg–)} 1
{Sδ+ O2δ –} ksep S•+ + O2•−
0.002
k−1 k−CT
k2 k−et 0
0 10 20 30 40 50 60 70
S + 1O2 (1Δg) S + O2 (3Σg–) Laser pulse energy (mJ)
Fig. 11 Singlet oxygen 1270 nm phosphorescence spectra and traces recorded in the presence of halogenated porphyrins and bacteri-
ochlorins and phenalenone as the reference. In the frame at the bottom, the mechanisms of singlet oxygen and superoxide ion are illustrated.
9.2 Detection of Radical Species by Electron Spin Resonance

Spin trapping is often used for the detection and characterization of transient
radicals (41,104,109,172,199). Spin traps serve as an efficient scavenger of
reactive free radicals to produce a more stable adduct and facilitate the mea-
surement of electron paramagnetic resonance (EPR) spectra. We have used
two spin traps, DMPO (5,5-dimethyl-1-pyrroline-N-oxide) and BMPO
(5-tert-butoxycarbonyl 5-methyl-1-pyrroline-N-oxide), to detect radicals
generated by bacteriochlorins (41,58,104,105,109). The structure of
DMPO is given in Fig. 12.
Fig. 12 also presents EPR spectra in the presence of the water soluble
bacteriochlorin (ClBOH) in PBS and DMPO after illumination with a
750-nm laser (109). The line shape and the hyperfine (hf ) splitting of the
signal are typical of the spin adduct formed between DMPO and the
hydroxyl radical (top panel, red). The computer simulation of this spectrum
is shown at the same figure (top panel, blue). Irradiation of bacteriochlorins
in DMSO and in the presence of DMPO for 10 s with a diode laser, resulted
in the spectrum presented in Fig. 12 (bottom panel, red). From a computer
simulation of the line shape and the hf splitting of the signal, these EPR spec-
tra were assigned to the adduct DMPO–OOH (41,105). To confirm that
the superoxide ion is the origin of the observed signal, SOD, a known scav-
enger of the superoxide ion, and the products are oxygen and hydrogen per-
oxide, was added to the first system in which DMSO was used as solvent,
prior to irradiation. The generation of the radical adduct DMPO–OOH
was inhibited after addition of SOD, meaning that SOD is an efficient scav-
enger of O2 • and confirms that superoxide ion is responsible for the obser-
vation of the DMPO–OOH EPR spectrum for each of the photosensitizers.
Under other experimental conditions, such as in the presence of catalase,
absence of light or when the solution was saturated with nitrogen, no
EPR signal was detected.
9.3 ROS Detection by Fluorescent Probes

Many commercially available fluorescent probes are advertised to be specific
for the detection of singlet oxygen, superoxide, hydrogen peroxide, or other
ROS (135,200). APF (2-[6-(40 -amino)phenoxy-3H-xanthen-3-on-9-yl]
benzoic acid) and HPF (2-[6-(40 -hydroxy)phenoxy-3H-xanthen-3-on-9-
yl]benzoic acid) illustrated in Fig. 13 were designed and synthesized as pro-
bes for stable and selective detection of HO%. They should selectively and
dose-dependently afford a strongly fluorescent compound, fluorescein,
upon reaction with HO% and hypochlorite, but not other ROS (137).
Hydroxyl radical H3C OH
H 3C N H
O⋅
H3C
H3C N+
O–
DMPO
H C O2H
Superoxide radical 3
H3C N H
O⋅
Fig. 12 Chemical structure of DMPO and EPR spectra recorded (left) and simulated (right) of bacteriochlorins in phosphate buffer saline (PBS)
or DMSO and in the presence of DMPO after illumination with a 750-nm laser.
X
◊OH O O O–
O O O
COO– COO–
HO O O HO O O
X = OH HPF 1
Dg
X = NH2 APF Cl Cl
O O
HOOC HOOC
O◊2– COOH COOH

N N+ SOSG
H2N NH2 H2N NH2
OH
DHE
Fig. 13 Structures of commonly used fluorescent probes: APF, HPF, DHE, and SOSG and their reactions with various ROS.
It was initially reported that these probes were relatively insensitive to 1O2.
However, recent studies indicate that fluorogenic interactions of APF with
1
O2 can be quite significant (136). On the other hand, HPF seems to be
more selective to hydroxyl radicals than APF (134).
Dihydroethidium (DHE) was suggested to be selective for O2 • detec-
tion if fluorescence is monitored at 570 nm, but this probe might be oxi-
dized by other ROS; therefore, it should rather be used for detection of
superoxide ion accompanied with HPLC analysis of generated products.
The structure of DHE together with its reaction with superoxide anion is
shown in Fig. 13.
Singlet oxygen sensor green (SOSG) is a very selective probe for 1O2
detection, but it hardly penetrates cell membranes (201). Its chemical struc-
ture and reaction with singlet oxygen are presented in Fig. 13. A procedure
for promoting cellular uptake of SOSG (202) published by Ogilby was suc-
cessfully adapted in our studies (see Fig. 14).
It is reported that APF can detect hydroxyl radicals and it is suggested that
it can also detect amounts of 1O2 (136). However, we have demonstrated
Fig. 14 Fluorescence micrographs of A549 cells coincubated with redaporfin (F2BMet),

and the fluorescent probes: APF, HPF, DHE, and SOSG. Images at the top show repre-
sentative fluorescence (in red) of the Ps in A549 cells, before activation by laser light.
In the middle row representative images of the APF, HPF, DHE, and SOSG fluorescence
are shown, after activation of photosensitizers by NIR light. Images in the bottom show
mergers of fluorescence from panels above.
that the fluorescence signal from APF in comparison to SOSG was negligible
when Ps with a singlet oxygen quantum yield close to unity (ZnPMet) was
used. In view of the very high yield of 1O2 from ZnPMet (119), together with
the 200-fold greater sensitivity of APF to HO% than to 1O2 (137), it appears
that APF, HPF, and SOSG can be successfully used to distinguish between
these species generated either via Mechanism I or Mechanism II (134).
The irradiation of chlorinated bacteriochlorin (ClBOH) leads to
hydroxyl radical generation in solution as well as in cells (203). In order
to assess the generation of hydroxyl radicals by other bacteriochlorins
(F2BMet, FBMet, ClBEt, and Cl2BEt), we incubated each of them with
APF in A549 cell cultures. APF reacts with various ROS, but reaction of
APF with hydroxyl radicals is favored over that with singlet oxygen
(136). Excitation of cells at 488 nm with the laser of confocal microscopy
after the illumination of the cells with NIR light gave rise to fluorescence
at 530 nm, as illustrated in Fig. 14. The same figure demonstrates the use
of other fluorescent probes for detection of various ROS. It is clear that
our lead compound (redaporfin) generates singlet oxygen, superoxide
ion, and hydroxyl radicals at least in A549 cancer cells.
Experiments carried out with APF and HPF have demonstrated that the
highest fluorescence intensity was observed for F2BMEt, lower fluorescence
for both FBMet and Cl2BEt was noted, and only a very weak fluorescence
was observed in the case of ClBEt. Hence, the quantity of hydroxyl radical
generation follows the order ClBEt < FBMet Cl2BEt < F2BMet. It is
interesting to note that exactly the same dependence can be drawn for
the photodynamic effect results for these bacteriochlorins. F2BMet is the
most phototoxic and ClBEt possesses the lowest photodynamic activity
in vitro. Considering that among bacteriochlorins that have been studied,
F2BMet possesses the lowest yield of 1O2 production and other molecules
have similar singlet oxygen quantum yields leading to comparable amounts
of singlet oxygen, we can conclude that photodynamic efficacy is correlated
more with hydroxyl radical formation rather than with singlet oxygen gen-
eration (41,104).
10. STRATEGIES TO ENHANCE ROS GENERATION IN PDT

10.1 Inhibition of Antioxidant Enzymes
The potentiation of ROS generation by the inhibition of antioxidant
enzymes is a simple strategy to improve a PDT outcome. The inhibition
of SOD enzymes with 2-methoxyestradiol (2ME) (204), the inhibition of
catalase by 3-amino-1,2,4-triazole (3AT) (205), and the depletion of intra-

cellular glutathione using buthionine sulfoximine (BSO) (206–209) were
applied in order to potentiate ROS activity manifested in the enhanced anti-
tumor performance of PDT. The combination of PDT with the
benzoporphyrin derivative monoacid and ascorbate in leukemia cells
(HL-60 and U937) revealed a prooxidant effect that enhanced photody-
namic efficacy against HL-60 cells (149). BSO alone or in combination with
other antioxidant inhibitors, significantly potentiated PDT due to increased
ROS level and extensive apoptosis (210). Our most impressive result in this
respect is that of the combination of PDT with ascorbate and various inhib-
itors. A very strong potentiation of PDT is possible when the Type
I photochemical mechanism determines the photodynamic effect. It is more
difficult to potentiate PDT when Type II processes control PDT outcome.
Ascorbate was shown to be more toxic to CT26 cells than to A549 cells.
These results cannot be explained by the higher activity of catalase because
both cell lines are characterized by a comparable level of activity of this
enzyme. On the other hand, CT26 cells exhibit enhanced SOD activity;
hence, more H2O2 is generated in CT26 than in A549 cells. CT26 cells are
more sensitive to ascorbate because ascorbate generates an additional
amount of H2O2. It is worthwhile to hypothesize that PDT associated with
ascorbate may be more selective and effective toward tumors than to nor-
mal cells. The conclusion from these studies is that ascorbate can shift the
PDT paradigm from Type II to Type I mechanisms and that the combi-
nations of PDT with ascorbate or with 2ME may contribute to the
improvement of therapeutic efficacy especially of tumors characterized
by low SOD activities (16,211). Moreover, Fe2+ may enhance ROS gen-
eration and improve the photodynamic efficacy when it is colocated with
the photosensitizer in the same cell organelle to assist an in vivo Fenton
reaction (105).
10.2 Nanoformulation of Photosensitizers

Nanoparticles have been extensively explored for many years as effective
delivery systems for photosensitizers dedicated to PDT of cancer. However,
more recent advances in nanotechnology provide new opportunities for the
development of ROS-generating systems for PDT which can face the chal-
lenges of the current approaches in nanomedicine (2,212,213). Chlorin e6
nanoformulated with Pluronic F127® was characterized by Park and Na
(214). They obtained five times higher singlet oxygen quantum yield for this
conjugate than for chlorin alone. This increased ability to generate singlet
oxygen resulted in favorable biological activity of nanoencapsulated chlorin
e6 in colon carcinoma cells (CT26). Vilsinski and coworkers described pho-
tochemical properties and biological evaluation of aluminum chloride
phthalocyanine (AlPcCl) incorporated into Pluronic P123 and F127 dem-
onstrating also improved ROS generation and an enhanced photodynamic
effect (159). Recently, we have designed a pluronic-based nanoformulation
for redaporfin and showed that it enhances ROS generation by redaporfin.
Fig. 15 illustrates the method of preparation of pluronic-based materials,
their particle size distribution at physiological pH, and their controlled
release at slightly acidic pH, as well as the influence of the formed micelles
on the ROS generation (184).
Our results also indicate that modification of redaporfin via incorpora-
tion into Pluronic P-123 micelles led not only to an improved stability
and ROS generation but also to the increased selectivity toward tumor tis-
sue, enhanced local inflammation and significant antitumor response after its
excitation with NIR. We have also reported similar effects for a set of Zn2+-
coordinated phthalocyanines of different polarity encapsulated in F127,
P123, and L121 (113). The conclusion from our recent publications is that
our application of nanosized polymeric micelles not only increases the quan-
tity of ROS during PDT but also overcome challenges associated with the
delivery of hydrophobic photosensitizing, such as poor solubility, cellular
internalization, and tumor targeting. Pluronic micelles significantly increase
Ps stability, reduced and delayed its photobleaching, and thus enhanced
ROS generation, most notably hydroxyl radicals, leading to potentiation
of therapeutic efficacy and overcoming the resistance of highly pigmented
(B16F10) melanoma. Other authors also described the role of nanomaterials
in PDT for the same melanoma tumor model. Pheophorbide A was
conjugated via the redox-sensitive disulfide linkage to alginate. DOX/
PheoA-ALG nanoparticles which were preferentially accumulated in B16
melanoma tumors, resulting in substantial inhibition of tumor growth by
both classical chemotherapy and PDT (215). The impact of micellar formu-
lation of Ps on the PDT mechanism was also studied by Nonel and Hamblin.
They describe the modulation of Type I and II mechanisms by the micelle
core environment. Electron-rich poly(2-(diisopropylamino)ethyl methac-
rylate) micelles increased photoactivations from Type II to Type I
mechanisms, which significantly increased the generation of O2 • through
the electron transfer pathway over 1O2 production through the energy
transfer process. Again, these results demonstrate that micelles not only
Fig. 15 Influence of the nanoencapsulation of bacteriochlorin in pluronic micelles on the ROS generation, distribution, and controlled release.
improve the bioavailability of Ps but also modulate ROS generation for

improved PDT efficacy (216).
Several interesting examples of singlet oxygen-generated nanomaterials
were described very recently. The photodynamic activity of polymeric
nanofiber materials doped with classic tetraphenylporphyrin photosensitizer,
activated by visible light, was found to be dependent on the oxygen perme-
ability and the diameter of the polymeric nanofibers. This was evidenced,
based on the ability of singlet oxygen generation upon irradiation. All the
tested nanofiber materials exhibited prolonged antibacterial properties, even
in the dark after long-duration irradiation. The enhanced activity was attrib-
uted to the high quantity of generated 1O2, and the postirradiation effect was
explained by the photogeneration of H2O2 (217). In other studies, 1O2 gen-
eration from Al(III) phthalocyanine chloride tetrasulfonic acid (AlPcS4)
conjugated on an Au nanorod was optimized through a proper combination
of surface plasmon resonance and the separation distance. 1O2 enhancement
follows a different correlation with the separation distance compared to that
of AlPcS4 fluorescence enhancement (218).
10.3 Photosensitization of TiO2 With Tetrapyrroles

Photosensitization of titanium dioxide still remains a challenging task when
considering applications of these materials in PDT, photodynamic inactiva-
tion of microorganisms (PDI), or photocatalysis (219). Surface modification
with dyes absorbing visible light is one of the commonly used approaches to
improve ROS generation. Among various types of photosensitizers capable
of electron injection from their excited states to the conduction band of tita-
nia (TiO2), porphyrins deserve particular attention. New synthetic haloge-
nated and sulfonated water soluble porphyrins and their analogues, which
appeared to be active photosensitizers inducing a photodynamic effect,
can also serve as TiO2 surface-bound sensitizers. Sulfonyl and sulfonamide
groups can act as anchors improving the interaction between the semicon-
ducting support and the sensitizer. Fluorinated or chlorinated sulfonyl and
sulfonamide porphyrins, as well as their metal (Co3+ and Zn2+) complexes,
anchored on TiO2 are interesting organic/inorganic photoactive materials
(Fig. 16). Impregnation of TiO2 especially with a zinc-porphyrin derivative
improved its photocatalytic performance over that of standard TiO2, leading
to enhanced ROS generation (98). This is assigned to the photoinduced elec-
tron transfer from porphyrin to the conduction band of TiO2. The main
ROS generation pathway for unmodified TiO2 is photoinduced electron
hv
X
R X
R
R: SO3H, SO2NHCH3 X
N X
X: Cl, F M N
N
M: Zn2+ Co3+ X X N
R X
X R
O2 CB e– A
e– e–
Reduction
O⋅2– –⋅
e – A
H2O2
e– D+⋅
⋅OH Oxidation
VB h+ D
Fig. 16 The adsorption of halogenated metalloporphyrins at the surface of TiO2 results

in novel materials with a high ability to generate ROS under visible light irradiation.
transfer, while zinc-porphyrins in a homogenous system act mostly via the

Type II energy transfer mechanism with a high quantum yield of singlet
oxygen (119). These hybrid materials offer both electron and energy transfer
processes with reasonable quantum yields. This class of TiO2-based photo-
sensitizers is currently being investigated as photoactive materials for photo-
dynamic inactivation of pathogenic bacteria and fungi. However TiO2-based
materials have also a potential to induce toxicity in hypoxic tumors, because
they generate ROS through an oxygen independent electron-hole transfer to
H2O molecules that are present on the surface of TiO2.
10.4 Addition of Inorganic Salts

Recently, Hamblin and Sarna with collaborators investigated the effect of
inert inorganic salts such as KI and KBr in combination with photodynamic
treatment. The mechanism of the synergistic effects of KI and methylene
blue-mediated PDT is related to the simultaneous generation of short-lived
reactive iodine species, hydroxyl radicals produced during PDT, as well as
the HO generated via the Fenton reaction mediated by the iodide anion
(36). Potentiation by addition of iodide also applied to PDT mediated by
cationic fullerenes and UV or visible light. This strategy was highly effective
in vitro against a set of pathogenic microbes as well as during an in
vivo experiment carried out on a mouse model (220). Addition of inert
salt can also enhance the efficacy of antimicrobial photocatalysis.

Photoinactivation mediated by TiO2 and UVA irradiation indicated that
excited semiconductor nanoparticles could indeed be potentiated by addi-
tion of bromide, involving the intermediate production of hypobromite
(34,221). It was demonstrated that using iodide, it is possible to transform
the most well-known anticancer photosensitizer (Photofrin) into a very
powerful antimicrobial agent. The reaction mechanism also involves
photogeneration of molecular iodine, but in this case, it occurs via singlet
oxygen generation. Oxygen consumption was increased when Photofrin
was illuminated in the presence of KI. Moreover, hydrogen peroxide but
not superoxide ion was generated after Photofrin excitation in the presence
of KI. Finally, sodium azide completely inhibited the photodynamic effect
that originated from the irradiation of the Photofrin + KI system (222).
These findings may have clinical applications since Photofrin is commer-
cially available world-wide.
11. SUMMARY
For many years singlet oxygen was considered as a predominant ROS
involved in PDT. The role of the oxygen-centered radicals and hydrogen
peroxide in the photodynamic efficacy has been much less explored. How-
ever, in recent years Type I ROS have become discussed more often in PDT
protocols. Direct detection of singlet oxygen phosphorescence and use of an
EPR spin trap for detection of free radicals are probably the most appropriate
methods to study ROS generation mechanisms in solution. However, it is
much more difficult to adopt these techniques in biological in vitro studies.
Fluorescent ROS probes are simpler and less expensive to use, but their
selectivity toward various ROS still remains the challenge. The most useful
probes in our studies were HPF (HO% ⋙ 1O2), APF (HO% > 1O2), and
SOSG which is highly specific for 1O2. The intracellular diffusion of
ROS depends on their lifetimes. In the case of porphyrins, mostly singlet
oxygen is generated with a lifetime of a few microseconds. Thus, diffusion
between different sites within the cell is possible, and the damage can occur
not only at the generation site but also in other organelles (153,223). On the
other hand, bacteriochlorins generate ROS according to both mechanisms
(Type I and Type II). Singlet oxygen and hydrogen peroxide have longer
lifetimes and a higher diffusion range, whereas superoxide ion or hydroxyl
radicals are short-lived species and cause damage at the narrow generation
site. Thus, the localization of the photosensitizer in the cell can have a major
influence on the mechanism and efficiency of PDT. Bacteriochlorins are

particularly interesting ROS-generating molecules for PDT application
because of their strong absorption in the phototherapeutic window, and
long-lived triplet states. Bacteriochlorins are more likely to participate in less
oxygen-dependent Type I mechanism due to their HOMO-LUMO energy
gap, and lower redox potentials. Thus, in addition to singlet oxygen, they
generate significant amounts of hydroxyl radicals that are highly toxic to cells
and might enhance the efficacy of PDT performed on hypoxic tumors (104).
The use of bacteriochlorins as PDT photosensitizers on a large scale
was limited for a long time because they were unstable and quite difficult
to synthetize (106). Together with Arnaut and Pereira we have developed
a new class of photostable bacteriochlorins with excellent properties
(41,45,104,105,115). Our lead compound (redaporfin, F2BMet) is cur-
rently in phase II clinical trials for head and neck tumors. Redaporfin inter-
acts with O2 both through Type I and Type II reactions and is very efficient
in the treatment of a variety of cancer cells and tumors, such as melanoma
(58,109,110), lung adenocarcinoma (41,104), and colon carcinoma
(23,66,67). Furthermore, PDT with redaporfin led not only to primary
tumor destruction but also to a long-lasting immune memory and protection
against metastasis. Recently, new methods of synthesis of bacteriochlorins
and other NIR absorbing dyes were developed. Some of these photo-
sensitizers are very stable and highly efficient against various cancer cells
(224–229). Moreover, similar to bacteriochlorins that we investigated,
they generate ROS according to both mechanisms: energy transfer and
photoinduced electron transfer.
The increasingly better understanding of the photochemical and biolog-
ical mechanisms of PDT opens new possibilities to potentiate ROS gener-
ation and eventually therapeutic efficacy. Among several strategies described
in the literature, the most efficient ones are conjugations with various types
of nanoparticles. However, very simple approaches such as addition of inor-
ganic salts or inhibition of antioxidant enzymes gave very promising results
as well. Particularly interesting is the possibility to combine PDT with che-
motherapy and immunotherapy. In such a way, not only the primary tumors
can be eliminated, but it is possible to induce long-lasting immune memory
and control of metastasis.
Another challenge of modern medicine is the need to address the prob-
lem of multidrug resistance (MDR) and to assess the potential value of PDT
in a solution. MDR occurs when cancer cells develop mechanisms that
allow them to survive after chemotherapy or radiotherapy. It is possible that
resistance induced by one treatment can be overcome by another treatment.

Emerging evidence suggests that the unique mechanisms offered by PDT
can be utilized to overcome drug resistance and even to resensitize resistant
cells to standard therapies. It has been observed that PDT can promote
responses in malignant lesions that are resistant to drugs and ionizing radi-
ation. PDT can bypass drug resistance pathways because ROS generated
in photodynamic action trigger the late stages of apoptosis, avoiding defec-
tive signaling pathways often found in neoplastic cells (24). Moreover, with
the current development of nanotechnology, it is possible that light activa-
tion may be used not only to potentiate ROS generation and enhance ther-
apeutic efficacy but also to enable controlled drug release to inhibit escape
pathways that may lead to resistance or cell proliferation (24).
ACKNOWLEDGMENTS
The work was supported by Grants no 2013/11/D/ST5/02995 and 2016/22/E/
NZ7/00420 (National Science Center, NCN) and no 0085/IP3/2015/73 (Ministry of
Science and Higher Education, MNiSW) awarded to J.M.D.
REFERENCES
1. Evans, J. L.; Maddux, B. A.; Goldfine, I. D. Antioxid. Redox Signal 2005, 7, 1040–1052.
2. Zhou, Z.; Song, J.; Nie, L.; Chen, X. Chem. Soc. Rev. 2016, 45, 6597–6626.
3. Oszajca, M.; Brindell, M.; Orzeł, Ł.; Da˛browski, J. M.; Śpiewak, K.; Łabuz, P.; Pacia, M.;
Stochel-Gaudyn, A.; Macyk, W.; van Eldik, R.; Stochel, G. Coord. Chem. Rev. 2016,
327–328, 143–165.
4. Ogilby, P. R. Chem. Soc. Rev. 2010, 39, 3181–3209.
5. Simonian, N.; Coyle, J. Ann. Rev. Pharm. Toxicol. 1996, 36, 83–106.
6. Niedzielska, E.; Smaga, I.; Gawlik, M.; Moniczewski, A.; Stankowicz, P.; Pera, J.;
Filip, M. Mol. Neurobiol. 2016, 53, 4094–4125.
7. Lin, M. T.; Beal, M. F. Nature 2006, 443, 787–795.
8. Parke, D.; Sapota, A. Int. J. Occup. Med. Environ. Health 1995, 9, 331–340.
9. Fuchs, J. Environmental Stressors in Health and Disease. CRC Press, New York, 2001;
Vol. 7.
10. Sundaresan, M.; Yu, Z.-X.; Ferrans, V. J.; Irani, K.; Finkel, T. Science 1995, 270, 296.
11. Stone, J. R.; Yang, S. Antioxidants Redox Signal 2006, 8, 243–270.
12. Woo, H. A.; Yim, S. H.; Shin, D. H.; Kang, D.; Yu, D.-Y.; Rhee, S. G. Cell 2010,
140, 517–528.
13. Trachootham, D.; Alexandre, J.; Huang, P. Nat. Rev. Drug Discov. 2009, 8, 579–591.
14. Vatansever, F.; de Melo, W. C.; Avci, P.; Vecchio, D.; Sadasivam, M.; Gupta, A.;
Chandran, R.; Karimi, M.; Parizotto, N. A.; Yin, R.; Tegos, G. P.;
Hamblin, M. R. FEMS Microbiol. Rev. 2013, 37, 955–989.
15. Fridovich, I. Aging Cell 2004, 3, 13–16.
16. Soares, H. T.; Campos, J. R. S.; Gomes-da-Silva, L. C.; Schaberle, F. A.;
Dabrowski, J. M.; Arnaut, L. G. Chembiochem 2016, 17, 836–842.
17. Pelicano, H.; Carney, D.; Huang, P. Drug Resist. Updat. 2004, 7, 97–110.
18. Dabrowski, J. M.; Arnaut, L. G. Photochem. Photobiol. Sci. 2015, 14, 1765–1780.
19. Agostinis, P.; Berg, K.; Cengel, K. A.; Foster, T. H.; Girotti, A. W.; Gollnick, S. O.;
Hahn, S. M.; Hamblin, M. R.; Juzeniene, A.; Kessel, D.; Korbelik, M.; Moan, J.;
Mroz, P.; Nowis, D.; Piette, J.; Wilson, B. C.; Golab, J. CA Cancer J. Clin. 2011,
61, 250–281.
20. Castano, A. P.; Mroz, P.; Hamblin, M. R. Nat. Rev. Cancer 2006, 6, 535–545.
21. Gonçalves, N. P. F.; Simões, A. V. C.; Abreu, A. R.; Abrunhosa, A. J.;
Da˛browski, J. M.; Pereira, M. M. J. Porphyr. Phthalocya. 2015, 19, 946–955.
22. Dolmans, D. E. J. G. J.; Fukumura, D.; Jain, R. K. Nat. Rev. Cancer 2003, 3, 380–387.
23. Rocha, L. B.; Gomes-da-Silva, L. C.; Da˛browski, J. M.; Arnaut, L. G. Eur. J. Cancer
2015, 51, 1822–1830.
24. Spring, B. Q.; Rizvi, I.; Xu, N.; Hasan, T. Photochem. Photobiol. 2015, 14, 1476–1491.
25. Crabb, J. W.; Miyagi, M.; Gu, X.; Shadrach, K.; West, K. A.; Sakaguchi, H.;
Kamei, M.; Hasan, A.; Yan, L.; Rayborn, M. E. Proc. Natl. Acad. Sci. U.S.A. 2002,
99, 14682–14687.
26. Kawczyk-Krupka, A.; Bugaj, A. M.; Potempa, M.; Wasilewska, K.; Latos, W.;
Sieron, A. Photodiagnosis Photodyn. Ther. 2015, 12, 161–175.
27. Schmidt-Erfurth, U.; Hasan, T. Surv. Ophthalmol. 2000, 45, 195–214.
28. Dai, T.; Huang, Y.-Y.; Hamblin, M. R. Photodiagnosis Photodyn. Ther. 2009, 6, 170–188.
29. Gad, F.; Zahra, T.; Francis, K. P.; Hasan, T.; Hamblin, M. R. Photochem. Photobiol. Sci.
2004, 3, 451–458.
30. Geilich, B. M.; van de Ven, A. L.; Singleton, G. L.; Sepulveda, L. J.; Sridhar, S.;
Webster, T. J. Nanoscale 2015, 7, 3511–3519.
31. Hamblin, M. R. Curr. Opin. Microbiol. 2016, 33, 67–73.
32. Hamblin, M. R.; Hasan, T. Photochem. Photobiol. Sci. 2004, 3, 436–450.
33. Huang, L.; Zhiyentayev, T.; Xuan, Y.; Azhibek, D.; Kharkwal, G. B.; Hamblin, M. R.
Lasers Surg. Med. 2011, 43, 313–323.
34. Huang, Y.-Y.; Choi, H.; Kushida, Y.; Bhayana, B.; Wang, Y.; Hamblin, M. R. Anti-
microb. Agents Chemother. 2016, 60, 5445–5453.
35. Vecchio, D.; Dai, T.; Huang, L.; Fantetti, L.; Roncucci, G.; Hamblin, M. R.
J. Biophotonics 2013, 6, 733–742.
36. Vecchio, D.; Gupta, A.; Huang, L.; Landi, G.; Avci, P.; Rodas, A.; Hamblin, M. R.
Antimicrob. Agents Chemother. 2015, 59, 5203–5212.
37. Vera, D. M. A.; Haynes, M. H.; Ball, A. R.; Dai, T.; Astrakas, C.; Kelso, M. J.;
Hamblin, M. R.; Tegos, G. P. Photochem. Photobiol. 2012, 88, 499–511.
38. Yin, R.; Agrawal, T.; Khan, U.; Gupta, G. K.; Rai, V.; Huang, Y.-Y.; Hamblin, M. R.
Nanomedicine 2015, 10, 2379–2404.
39. Yin, R.; Dai, T.; Avci, P.; Jorge, A. E. S.; de Melo, W. C.; Vecchio, D.; Huang, Y.-Y.;
Gupta, A.; Hamblin, M. R. Curr. Opin. Pharmacol. 2013, 13, 731–762.
40. Abels, C. Photochem. Photobiol. Sci. 2004, 3, 765–771.
41. Arnaut, L. G.; Pereira, M. M.; Da˛browski, J. M.; Silva, E. F.; Schaberle, F. A.;
Abreu, A. R.; Rocha, L. B.; Barsan, M. M.; Urba nska, K.; Stochel, G. Chem.
A Eur. J. 2014, 20, 5346–5357.
42. Castano, A. P.; Demidova, T. N.; Hamblin, M. R. Photodiagnosis Photodyn. Ther. 2004,
1, 279–293.
43. Dougherty, T. J.; Gomer, C. J.; Henderson, B. W.; Jori, G.; Kessel, D.; Korbelik, M.;
Moan, J.; Peng, Q. J. Nat. Cancer Inst. 1998, 90, 889–905.
44. Ethirajan, M.; Chen, Y.; Joshi, P.; Pandey, R. K. Chem. Soc. Rev. 2011, 40, 340–362.
45. Da˛browski, J. M.; Pucelik, B.; Regiel-Futyra, A.; Brindell, M.; Mazuryk, O.;
Kyzioł, A.; Stochel, G.; Macyk, W.; Arnaut, L. G. Coord. Chem. Rev. 2016, 325,
67–101.
46. Garg, A. D.; Maes, H.; Romano, E.; Agostinis, P. Photochem. Photobiol. Sci. 2015, 14,
1410–1424.
47. Edinger, A. L.; Thompson, C. B. Curr. Opinion Cell Biol. 2004, 16, 663–669.
48. Kessel, D.; Oleinick, N. L. Method Enzymol. 2009, 453, 1–16.
49. Kessel, D. H.; Price, M.; Reiners, J.; John, J. Autophagy 2012, 8, 1333–1341.
50. Schmidt-Erfurth, U.; Hasan, T.; Gragoudas, E.; Michaud, N.; Flotte, T. J.;
Birngruber, R. Ophthalmology 1994, 101, 1953–1961.
51. Dolmans, D. E. J. G. J.; Kadambi, A.; Hill, J. S.; Waters, C. A.; Robinson, B. C.;
Walker, J. P.; Fukumura, D.; Jain, R. K. Cancer Res. 2002, 62, 2151–2156.
52. Azzouzi, A.-R.; Lebdai, S.; Benzaghou, F.; Stief, C. World J. Urol. 2015, 33, 937–944.
53. Chen, B.; Pogue, B. W.; Hoopes, P. J.; Hasan, T. Int. J. Radiat. Oncol. Biol. Phys. 2005,
61, 1216–1226.
54. Krammer, B. Anticancer Res 2001, 21, 4271–4277.
55. Maeda, H.; Wu, J.; Sawa, T.; Matsumura, Y.; Hori, K. J. Control. Release 2000, 65,
271–284.
56. Preise, D.; Scherz, A.; Salomon, Y. Photochem. Photobiol. Sci. 2011, 10, 681–688.
57. Zilberstein, J.; Schreiber, S.; Bloemers, M. C. W. M.; Bendel, P.; Neeman, M.;
Schechtman, E.; Kohen, F.; Scherz, A.; Salomon, Y. Photochem. Photobiol. 2001, 73,
257–266.
58. Krzykawska-Serda, M.; Da˛browski, J. M.; Arnaut, L. G.; Szczygieł, M.; Urba nska, K.;
Stochel, G.; Elas, M. Free Radic. Biol. Med. 2014, 73, 239–251.
59. Mroz, P.; Szokalska, A.; Wu, M. X.; Hamblin, M. R. PLoS One 2010, 5, e15194.
60. Gollnick, S. O.; Brackett, C. M. Immunol. Res. 2010, 46, 216–226.
61. Garg, A. D.; Nowis, D.; Golab, J.; Agostinis, P. Apoptosis 2010, 15, 1050–1071.
62. Allison, R. R.; Downie, G. H.; Cuenca, R.; Hu, X.-H.; Childs, C. J.; Sibata, C. H.
Photodiagnosis Photodyn. Ther. 2004, 1, 27–42.
63. Davia, K.; King, D.; Hong, Y. L.; Swavey, S. Inorg. Chem. Commun. 2008, 11,
584–586.
64. Derycke, A. S. L.; de Witte, P. A. M. Adv. Drug Deliv. Rev. 2004, 56, 17–30.
65. Ethirajan, M.; Saenz, C.; Gupta, A.; Dobhal, M.; Pandey, R. Advances in Photodynamic
Therapy: Basic, Translational, and Clinical. Artech House: Boston, 2008;13–39.
66. Rocha, L. B.; Schaberle, F.; Da˛browski, J. M.; Simões, S.; Arnaut, L. G. Int. J. Mol. Sci.
2015, 16, 29236–29249.
67. Saavedra, R.; Rocha, L. B.; Da˛browski, J. M.; Arnaut, L. G. ChemMedChem 2014, 9,
390–398.
68. Hamblin, M. R.; Newman, E. L. J. Photochem. Photobiol. B 1994, 23, 3–8.
69. Castano, A. P.; Demidova, T. N.; Hamblin, M. R. Photodiagnosis Photodyn. Ther. 2005,
2, 91–106.
70. Szaciłowski, K.; Macyk, W.; Drzewiecka-Matuszek, A.; Brindell, M.; Stochel, G.
Chem. Rev. 2005, 105, 2647–2694.
71. Stochel, G.; Brindell, M.; Macyk, W.; Stasicka, Z.; Szaciłowski, K. Bioinorganic Photo-
chemistry. John Wiley & Sons Ltd: Chichester, 2009, 382.
72. Hu, J.; Tang, Y. a.; Elmenoufy, A. H.; Xu, H.; Cheng, Z.; Yang, X. Small 2015, 11,
5860–5887.
73. Smith, A. M.; Mancini, M. C.; Nie, S. Nat. Nanotechnol. 2009, 4, 710.
74. Luo, S.; Zhang, E.; Su, Y.; Cheng, T.; Shi, C. Biomaterials 2011, 32, 7127–7138.
75. Farrell, T. J.; Wilson, B. C.; Patterson, M. S.; Chow, R. Optics, Electro-Optics, and
Laser Applications in Science and Engineering. SPIE (International Society for Optics
and Electronics): Bellingham, USA, 1991, pp 146–155.
76. Niedre, M. J.; Secord, A. J.; Patterson, M. S.; Wilson, B. C. Cancer Res. 2003, 63,
7986–7994.
77. Idris, N. M.; Jayakumar, M. K. G.; Bansal, A.; Zhang, Y. Chem. Soc. Rev. 2015, 44,
1449–1478.
78. Zhou, J.; Liu, Q.; Feng, W.; Sun, Y.; Li, F. Chem. Rev. 2015, 115, 395–465.
79. Chen, G.; Qiu, H.; Prasad, P. N.; Chen, X. Chem. Rev. 2014, 114, 5161–5214.
80. Brancaleon, L.; Moseley, H. Lasers Med. Sci. 2002, 17, 173–186.
81. Mitsunaga, M.; Ogawa, M.; Kosaka, N.; Rosenblum, L. T.; Choyke, P. L.;
Kobayashi, H. Nat. Med. 2011, 17, 1685–1691.
82. Damoiseau, X.; Schuitmaker, H. J.; Lagerberg, J. W. M.; Hoebeke, M. J. Photochem.
Photobiol. B 2001, 60, 50–60.
83. Kong, G.; Anyarambhatla, G.; Petros, W. P.; Braun, R. D.; Colvin, O. M.;
Needham, D.; Dewhirst, M. W. Cancer Res. 2000, 60, 6950–6957.
84. Łapok, Ł.; Cyza, M.; Gut, A.; Ke˛pczy nski, M.; Szewczyk, G.; Sarna, T.;
Nowakowska, M. J. Photochem. Photobiol. A 2014, 286, 55–63.
85. Morgan, J.; Lottman, H.; Abbou, C. C.; Chopin, D. K. Photochem. Photobiol. 1994, 60,
486–496.
86. Namiki, Y.; Namiki, T.; Date, M.; Yanagihara, K.; Yashiro, M.; Takahashi, H.
Pharmacol. Res. 2004, 50, 65–76.
87. Nunes, S. M. T.; Sguilla, F. S.; Tedesco, A. C. Braz. J. Med. Biol. Res. 2004, 37,
273–284.
88. Puri, A. Pharmaceutics 2014, 6, 1–25.
89. Reddi, E. J. Photochem. Photobiol. B. 37, 189–195.
90. Reddi, E.; Zhou, C.; Biolo, R.; Menegaldo, E.; Jori, G. Br. J. Cancer 1990, 61,
407–411.
91. Sadasivam, M.; Avci, P.; Gupta, G. K.; Lakshmanan, S.; Chandran, R.; Huang, Y.-Y.;
Kumar, R.; Hamblin, M. R. Eur. J. Nanomed. 2013, 5. http://dx.doi.org/10.1515/
ejnm-2013-0010.
92. Thompson, D. H.; Gerasimov, O. V.; Wheeler, J. J.; Rui, Y.; Anderson, V. C. Biochim.
Biophys. Acta—Biomembranes 1996, 1279, 25–34.
93. Krysko, D. V.; Garg, A. D.; Kaczmarek, A.; Krysko, O.; Agostinis, P.;
Vandenabeele, P. Nat. Rev. Cancer 2012, 12, 860–875.
94. DeRosa, M. C.; Crutchley, R. J. Coord. Chem. Rev. 2002, 233–234, 351–371.
95. Yang, S. I.; Seth, J.; Strachan, J.-P.; Gentemann, S.; Kim, D.; Holten, D.; Lindsey, J. S.;
Bocian, D. F. J. Porphyr. Phthalocya. 1999, 03, 117–147.
96. Hamblin, M. R.; Huang, Y. Y. Handbook of Photomedicine. CRC Press: Boca Raton,
2014, p 854.
97. Kadish, K. M.; Smith, K. M.; Guilard, R. The Porphyrin Handbook: Inorganic, Organo-
metallic and Coordination Chemistry; Vol. 3. Elsevier: San Diego, 2000; Vol. 3.
98. Da˛browski, J. M.; Pucelik, B.; Pereira, M. M.; Arnaut, L. G.; Macyk, W.; Stochel, G.
RSC Adv. 2015, 5, 93252–93261.
99. Silva, M.; Azenha, M.; Pereira, M.; Burrows, H.; Sarakha, M.; Forano, C.;
Ribeiro, M.; Fernandes, A. Appl Catal B 2010, 100, 1–9.
100. Pinto, S. M.; Henriques, C. A.; Tome, V. A.; Vinagreiro, C. S.; Calvete, M. J.;
Da˛browski, J. M.; Piñeiro, M.; Arnaut, L. G.; Pereira, M. M. J. Porphyr. Phthalocya.
2016, 20, 45–60.
101. Giuntini, F.; Dumoulin, F.; Daly, R.; Ahsen, V.; Scanlan, E. M.; Lavado, A. S. P.;
Aylott, J. W.; Rosser, G. A.; Beeby, A.; Boyle, R. W. Nanoscale 2012, 4, 2034–2045.
102. Jurow, M.; Schuckman, A. E.; Batteas, J. D.; Drain, C. M. Coord. Chem. Rev. 2010,
254, 2297–2310.
103. Berenbaum, M.; Akande, S.; Bonnett, R.; Kaur, H.; Ioannou, S.; White, R.;
Winfield, U. Br. J. Cancer 1986, 54, 717.
104. Da˛browski, J. M.; Arnaut, L. G.; Pereira, M. M.; Urbanska, K.; Simoes, S.; Stochel, G.;
Cortes, L. Free Radic. Biol. Med. 2012, 52, 1188–1200.
105. Silva, E. F. F.; Serpa, C.; Da˛browski, J. M.; Monteiro, C. J. P.; Formosinho, S. J.;
Stochel, G.; Urba nska, K.; Simões, S.; Pereira, M. M.; Arnaut, L. G. Chem. A Eur.
J. 2010, 16, 9273–9286.
106. Pineiro, M.; Pereira, M. M.; Formosinho, S. J.; Arnaut, L. G. J. Phys. Chem. A 2002,
106, 3787–3795.
107. Yang, E.; Diers, J. R.; Huang, Y. Y.; Hamblin, M. R.; Lindsey, J. S.; Bocian, D. F.;
Holten, D. Photochem. Photobiol. 2013, 89, 605–618.
108. Da˛browski, J. M.; Arnaut, L. G.; Pereira, M. M.; Monteiro, C. J.; Urba nska, K.;
Simões, S.; Stochel, G. ChemMedChem 2010, 5, 1770–1780.
109. Da˛browski, J. M.; Urbanska, K.; Arnaut, L. G.; Pereira, M. M.; Abreu, A. R.;
Simões, S.; Stochel, G. ChemMedChem 2011, 6, 465–475.
110. Da˛browski, J. M.; Krzykawska, M.; Arnaut, L. G.; Pereira, M. M.; Monteiro, C. J.;
Simões, S.; Urbanska, K.; Stochel, G. ChemMedChem 2011, 6, 1715–1726.
111. Silva, E. F. F.; Schaberle, F. A.; Monteiro, C. J. P.; Da˛browski, J. M.; Arnaut, L. G.
Photochem. Photobiol. Sci. 2013, 12, 1187–1192.
112. Da˛browski, J. M.; Pereira, M. M.; Arnaut, L. G.; Monteiro, C. J. P.; Peixoto, A. F.;
Karocki, A.; Urba nska, K.; Stochel, G. Photochem. Photobiol. 2007, 83, 897–903.
113. Pucelik, B.; G€ urol, I.; Ahsen, V.; Dumoulin, F.; Da˛browski, J. M. Eur. J. Med. Chem.
2016, 124, 284–298.
114. Topal, S. Z.; Isci, U.; Kumru, U.; Atilla, D.; Gurek, A. G.; Hirel, C.; Durmus, M.;
Tommasino, J.-B.; Luneau, D.; Berber, S.; Dumoulin, F.; Ahsen, V. Dalton Trans.
2014, 43, 6897–6908.
115. Pereira, M. M.; Monteiro, C. J. P.; Simões, A. V. C.; Pinto, S. M. A.; Abreu, A. R.;
Sá, G. F. F.; Silva, E. F. F.; Rocha, L. B.; Da˛browski, J. M.; Formosinho, S. J.;
Simões, S.; Arnaut, L. G. Tetrahedron. 2010, 66, 9545–9551.
116. Arnaut, L. G. In Design of porphyrin-based photosensitizers for photodynamic therapy; van
Eldik, R., Stochel, G., Eds.; Advances in Inorganic Chemistry, Vol. 63, Elsevier
Academic Press: San Diego, 2011; pp 187–233.
117. Aydın Tekdaş, D.; Kumru, U.; G€ urek, A. G.; Durmuş, M.; Ahsen, V.; Dumoulin, F.
Tetrahedron Lett. 2012, 53, 5227–5230.
118. Simões, A. V. C.; Adamowicz, A.; Da˛browski, J. M.; Calvete, M. J. F.; Abreu, A. R.;
Stochel, G.; Arnaut, L. G.; Pereira, M. M. Tetrahedron 2012, 68, 8767–8772.
119. Da˛browski, J. M.; Pucelik, B.; Pereira, M. M.; Arnaut, L. G.; Stochel, G. J. Coord.
Chem. 2015, 68, 3116–3134.
120. Aydın Tekdaş, D.; Garifullin, R.; Şent€ urk, B.; Zorlu, Y.; Gundogdu, U.; Atalar, E.;
Tekinay, A. B.; Chernonosov, A. A.; Yerli, Y.; Dumoulin, F.; Guler, M. O.;
Ahsen, V.; G€ urek, A. G. Photochem. Photobiol. 2014, 90, 1376–1386.
121. Dumoulin, F.; Ali, H.; Ahsen, V.; van Lier, J. E. Tetrahedron Lett. 2011, 52, 4395–4397.
122. Dumoulin, F.; Durmuş, M.; Ahsen, V.; Nyokong, T. Coord. Chem. Rev. 2010, 254,
2792–2847.
123. Zorlu, Y.; Dumoulin, F.; Bouchu, D.; Ahsen, V.; Lafont, D. Tetrahedron Lett. 2010, 51,
6615–6618.
124. Zorlu, Y.; Dumoulin, F.; Durmuş, M.; Ahsen, V. Tetrahedron 2010, 66, 3248–3258.
125. Hong, G.; Diao, S.; Chang, J.; Antaris, A. L.; Chen, C.; Zhang, B.; Zhao, S.;
Atochin, D. N.; Huang, P. L.; Andreasson, K. I. Nat. Photonics 2014, 8, 723–730.
126. Chen, C.-Y.; Sun, E.; Fan, D.; Taniguchi, M.; McDowell, B. E.; Yang, E.; Diers, J. R.;
Bocian, D. F.; Holten, D.; Lindsey, J. S. Inorg. Chem. 2012, 51, 9443–9464.
127. Huang, L.; Krayer, M.; Roubil, J. G.; Huang, Y.-Y.; Holten, D.; Lindsey, J. S.;
Hamblin, M. R. J. Photochem. Photobiol. B. 2014, 141, 119–127.
128. Huang, Y. Y.; Balasubramanian, T.; Yang, E.; Luo, D.; Diers, J. R.; Bocian, D. F.;
Lindsey, J. S.; Holten, D.; Hamblin, M. R. ChemMedChem 2012, 7, 2155–2167.
129. Jiang, J.; Yang, E.; Reddy, K. R.; Niedzwiedzki, D. M.; Kirmaier, C.; Bocian, D. F.;
Holten, D.; Lindsey, J. S. New J. Chem. 2015, 39, 5694–5714.
130. Krayer, M.; Yang, E.; Kim, H.-J.; Kee, H. L.; Deans, R. M.; Sluder, C. E.; Diers, J. R.;
Kirmaier, C.; Bocian, D. F.; Holten, D.; Lindsey, J. S. Inorg. Chem. 2011, 50, 4607–4618.
131. Huang, Y.-Y.; Mroz, P.; Zhiyentayev, T.; Sharma, S. K.; Balasubramanian, T.; Ruzie, C.;
Krayer, M.; Fan, D.; Borbas, K. E.; Yang, E. J. Med. Chem. 2010, 53, 4018–4027.
132. Dabrowski, J. M.; Arnaut, L. G.; Pereira, M. M.; Urbanska, K.; Stochel, G.
Medchemcomm 2012, 3, 502–505.
133. Fukuzumi, S.; Ohkubo, K.; Zheng, X.; Chen, Y.; Pandey, R. K.; Zhan, R.;
Kadish, K. M. J. Phys. Chem. B 2008, 112, 2738–2746.
134. Garcia-Diaz, M.; Huang, Y.-Y.; Hamblin, M. R. Methods 2016, 109, 158–166.
135. Price, M.; Kessel, D. J. Biomed. Opt. 2010, 15, 051605.
136. Price, M.; Reiners, J. J.; Santiago, A. M.; Kessel, D. Photochem. Photobiol. 2009, 85,
1177–1181.
137. Setsukinai, K.-I.; Urano, Y.; Kakinuma, K.; Majima, H. J.; Nagano, T. J. Biol. Chem.
2003, 278, 3170–3175.
138. Azenha, E.; Serra, A. C.; Pineiro, M.; Pereira, M. M.; Seixas de Melo, J.; Arnaut, L. G.;
Formosinho, S. J.; Rocha Gonsalves, A. M. d. A. Chem. Phys 2002, 280, 177–190.
139. Wilkinson, F.; Abdel-Shafi, A. J. Phys. Chem. A 1997, 101, 5509–5516.
140. Boyer, R. F.; McCleary, C. J. Free Radic. Biol. Med. 1987, 3, 389–395.
141. Gutteridge, J. M.; Maidt, L.; Poyer, L. Biochem. J. 1990, 269, 169–174.
142. Rizzi, C.; Samouilov, A.; Kutala, V. K.; Parinandi, N. L.; Zweier, J. L.; Kuppusamy, P.
Free Radic. Biol. Med. 2003, 35, 1608–1618.
143. Ershov, B.; Gordeev, A. Radiat. Phys. Chem. 2008, 77, 928–935.
144. Nandi, D.; Krishnakumar, E.; Rosa, A.; Schmidt, W.-F.; Illenberger, E. Chem. Phys.
Lett. 2003, 373, 454–459.
145. Evans, M. K.; Tovmasyan, A.; Batinic-Haberle, I.; Devi, G. R. Free Radic. Biol. Med.
2014, 68, 302–314.
146. Chen, Q.; Espey, M. G.; Krishna, M. C.; Mitchell, J. B.; Corpe, C. P.; Buettner, G. R.;
Shacter, E.; Levine, M. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 13604–13609.
147. Kehrer, J. P. Toxicology 2000, 149, 43–50.
148. Bielski, B. H. J.; Cabelli, D. E.; Arudi, R. L.; Ross, A. B. J. Phys. Chem. Ref. Data 1985,
14, 1041–1100.
149. Kramarenko, G. G.; Wilke, W. W.; Dayal, D.; Buettner, G. R.; Schafer, F. Q. Free
Radic. Biol. Med. 2006, 40, 1615–1627.
150. Bacellar, I. O.; Tsubone, T. M.; Pavani, C.; Baptista, M. S. Int. J. Mol. Sci. 2015, 16,
20523–20559.
151. Plaetzer, K.; Krammer, B.; Berlanda, J.; Berr, F.; Kiesslich, T. Lasers Med. Sci. 2009, 24,
259–268.
152. da Silva, E. F. F.; Pimenta, F. M.; Pedersen, B. W.; Blaikie, F. H.; Bosio, G. N.;
Breitenbach, T.; Westberg, M.; Bregnhoj, M.; Etzerodt, M.; Arnaut, L. G.;
Ogilby, P. R. Integr. Biol. 2016, 8, 177–193.
153. da Silva, E. F.; Pedersen, B. W.; Breitenbach, T.; Toftegaard, R.; Kuimova, M. K.;
Arnaut, L. G.; Ogilby, P. R. J. Phys. Chem. B 2011, 116, 445–461.
154. Berg, K.; Moan, J. Photochem. Photobiol. 1997, 65, 403–409.
155. Chowdhary, R. K.; Chansarkar, N.; Sharif, I.; Hioka, N.; Dolphin, D. Photochem. Pho-
tobiol. 2003, 77, 299–303.
156. Gelderblom, H.; Verweij, J.; Nooter, K.; Sparreboom, A. Eur. J. Cancer 2001, 37,
1590–1598.
157. Pucelik, B.; Paczy nski, R.; Dubin, G.; Pereira, M. M.; Arnaut, L. G.; Da˛browski, J. M.
J. Photobiol. Photochem. B., 2017, submitted.
158. Tije, A. J.; Verweij, J.; Loos, W. J.; Sparreboom, A. Clin. Pharmacokinet. 2012, 42,
665–685.
159. Vilsinski, B. H.; Gerola, A. P.; Enumo, J. A.; Campanholi, K. D. S. S.;
Pereira, P. C. D. S.; Braga, G.; Hioka, N.; Kimura, E.; Tessaro, A. L.; Caetano, W.
Photochem. Photobiol. 2015, 91, 518–525.
160. Jori, G. In CRC Handbook of Organic Photochemistry and Photobiology, Horspool, W. M.,
Lenci, F., Eds. 2nd ed.; CRC Press: Boca Raton (Florida), 2004, pp 146.1–14610.
161. Mojzisova, H.; Bonneau, S.; Vever-Bizet, C.; Brault, D. Biochim. Biophys. Acta—
Biomembranes 1768, 2007, 2748–2756.
162. Juarranz, A.; Villanueva, A.; Dı́az, V.; Cañete, M. J. Photochem. Photobiol. B, 1995, 27,
47–53.
163. Kuimova, M. K.; Balaz, M.; Anderson, H. L.; Ogilby, P. R. J. Am. Chem. Soc. 2009,
131, 7948–7949.
164. Mroz, P.; Huang, Y.-Y.; Szokalska, A.; Zhiyentayev, T.; Janjua, S.; Nifli, A.-P.;
Sherwood, M. E.; Ruzie, C.; Borbas, K. E.; Fan, D.; Krayer, M.;
Balasubramanian, T.; Yang, E.; Kee, H. L.; Kirmaier, C.; Diers, J. R.; Bocian, D. F.;
Holten, D.; Lindsey, J. S.; Hamblin, M. R. FASEB J. 2010, 24, 3160–3170.
165. Hsieh, Y.-J.; Wu, C. C.; Chang, C. J.; Yu, J. S. J. Cell. Physiol. 2003, 194, 363–375.
166. Hsieh, Y.-J.; Yu, J.-S.; Lyu, P.-C. J. Cell. Biochem. 2010, 111, 821–833.
167. Marchal, S.; François, A.; Dumas, D.; Guillemin, F.; Bezdetnaya, L. Br. J. Cancer 2007,
96, 944–951.
168. Runnels, J. M.; Chen, N.; Ortel, B.; Kato, D.; Hasan, T. Br. J. Cancer 1999, 80,
946–953.
169. Osaki, T.; Takagi, S.; Hoshino, Y.; Okumura, M.; Fujinaga, T. Cancer Lett. 2006, 243,
281–292.
170. Tada-Oikawa, S.; Oikawa, S.; Hirayama, J.; Hirakawa, K.; Kawanishi, S. Photochem.
Photobiol. 2009, 85, 1391–1399.
171. Moor, A. C. J. Photochem. Photobiol. B, 2000, 57, 1–13.
172. Ashur, I.; Goldschmidt, R.; Pinkas, I.; Salomon, Y.; Szewczyk, G.; Sarna, T.;
Scherz, A. J. Phys. Chem. A 2009, 113, 8027–8037.
173. Oleinick, N. L.; Morris, R. L.; Belichenko, I. Photochem. Photobiol. Sci. 2002, 1, 1–21.
174. Mroz, P.; Yaroslavsky, A.; Kharkwal, G. B.; Hamblin, M. R. Cancer 2011, 3, 2516.
175. Martinou, J.-C.; Desagher, S.; Antonsson, B. Nat. Cell Biol. 2000, 2, E41–E43.
176. Varnes, M. E.; Chiu, S.-M.; Xue, L.-Y.; Oleinick, N. L. Biochem. Biophys. Res.
Commun. 1999, 255, 673–679.
177. Xue, L. Y.; Chiu, S. M.; Oleinick, N. L. Exp. Cell Res. 2001, 263, 145–155.
178. Teiten, M. H.; Marchal, S.; D’Hallewin, M.; Guillemin, F.; Bezdetnaya, L. Photochem.
Photobiol. 2003, 78, 9–14.
179. Reiners, J., Jr.; Caruso, J.; Mathieu, P.; Chelladurai, B.; Yin, X.; Kessel, D. Cell Death
Differ. 2002, 9, 934.
180. Chiu, S.; Oleinick, N. Br. J. Cancer 2001, 84, 1099.
181. Vantieghem, A.; Xu, Y.; Declercq, W.; Vandenabeele, P.; Denecker, G.;
Vandenheede, J. R.; Merlevede, W.; De Witte, P. A.; Agostinis, P. Photochem. Photo-
biol. 2001, 74, 133–142.
182. Kessel, D. Photochem. Photobiol. 2008, 84, 809–814.
183. Proskuryakov, S. Y.; Konoplyannikov, A. G.; Gabai, V. L. Exp. Cell Res. 2003, 283,
1–16.
184. Pucelik, B.; Arnaut, L. G.; Stochel, G.; Da˛browski, J. M. ACS Appl. Mater. Interfaces
2016, 8, 22039–22055.
185. Coupienne, I.; Fettweis, G.; Rubio, N.; Agostinis, P.; Piette, J. Photochem. Photobiol.
Sci. 2011, 10, 1868–1878.
186. Wirawan, E.; Berghe, T. V.; Lippens, S.; Agostinis, P.; Vandenabeele, P. Cell Res.
2012, 22, 43–61.
187. Kroemer, G.; Levine, B. Nat. Rev. Mol. Cell Biol. 2008, 9, 1004–1010.
188. Scherz-Shouval, R.; Shvets, E.; Fass, E.; Shorer, H.; Gil, L.; Elazar, Z. EMBO J. 2007,
26, 1749–1760.
189. Pattingre, S.; Levine, B. Cancer Res. 2006, 66, 2885–2888.
190. Xue, L. Y.; Chiu, S. M.; Azizuddin, K.; Joseph, S.; Oleinick, N. L. Autophagy 2008, 4,
125–127.
191. Andrzejak, M.; Price, M.; Kessel, D. H. Autophagy 2011, 7, 979–984.
192. Krzykawska, M.; Dabrowski, J. M.; Szczygiel, M.; Stochel, G.; Arnaut, L. G.;
Pereira, M. M.; Urbanska, K.; Elas, M. Eur. J. Cancer 2012, 48, S193.
193. Reginato, E.; Wolf, P.; Hamblin, M. R. World J. Immunol. 2014, 4, 1–11.
194. Nowis, D.; Stokłosa, T.; Legat, M.; Issat, T.; Jakóbisiak, M.; Goła˛b, J. Photodiagnosis
Photodyn. Ther. 2005, 2, 283–298.
195. Anzengruber, F.; Avci, P.; de Freitas, L. F.; Hamblin, M. R. Photochem. Photobiol. Sci.
2015, 14, 1492–1509.
196. Kousis, P. C.; Henderson, B. W.; Maier, P. G.; Gollnick, S. O. Cancer Res. 2007, 67,
10501–10510.
197. Bonnett, R.; McGarvey, D. J.; Harriman, A.; Land, E. J.; Truscott, T. G.;
Winfield, V. J. Photochem. Photobiol. 1988, 48, 271.
198. Schmidt, R.; Tanielian, C.; Dunsbach, R.; Wolff, C. J. Photochem. Photobiol. A 1994,
79, 11–17.
199. Vakrat-Haglili, Y.; Weiner, L.; Brumfeld, V.; Brandis, A.; Salomon, Y.; McLlroy, B.;
Wilson, B. C.; Pawlak, A.; Rozanowska, M.; Sarna, T.; Scherz, A. J. Am. Chem. Soc.
2005, 127, 6487–6497.
200. Zielonka, J.; Kalyanaraman, B. Free Radic. Biol. Med. 2010, 48, 983–1001.
201. Flors, C.; Fryer, M. J.; Waring, J.; Reeder, B.; Bechtold, U.; Mullineaux, P. M.;
Nonell, S.; Wilson, M. T.; Baker, N. R. J. Exp. Bot. 2006, 57, 1725–1734.
202. Gollmer, A.; Arnbjerg, J.; Blaikie, F. H.; Pedersen, B. W.; Breitenbach, T.;
Daasbjerg, K.; Glasius, M.; Ogilby, P. R. Photochem. Photobiol. 2011, 87, 671–679.
203. Dabrowski, J. M.; Silva, E.; Krzykawski, M.; Rocha, L.; Urbanska, K.; Pereira, M. M.;
Stochel, G.; Arnaut, L. G. Free Radic. Biol. Med. 2012, 53, S18.
204. Golab, J.; Nowis, D.; Skrycki, M.; Czocsot, H.; Baranczyk-Kuzma, A.;
Wilczynski, G. M.; Makowski, M.; Mroz, P.; Kozar, K.; Kaminski, R.; Jalili, A.;
Kopec, M.; Grzela, T.; Jakobisiak, M. J. Biol. Chem. 2003, 278, 407–414.
205. Price, M.; Terlecky, S. R.; Kessel, D. Photochem. Photobiol. 2009, 85, 1491–1496.
206. Miller, A. C.; Henderson, B. W. Radiat. Res. 1986, 107, 83–94.
207. Bachor, R.; Scholz, M.; Shea, C. R.; Hasan, T. Cancer Res. 1991, 51, 4410–4414.
208. Jiang, F.; Robin, A. M.; Katakowski, M.; Tong, L.; Espiritu, M.; Singh, G.; Chopp, M.
Lasers Med. Sci. 2003, 18, 128–133.
209. Oberdanner, C. B.; Plaetzer, K.; Kiesslich, T.; Krammer, B. Photochem. Photobiol. 2005,
81, 609–613.
210. Kimani, S. G.; Phillips, J. B.; Bruce, J. I.; MacRobert, A. J.; Golding, J. P. Photochem.
Photobiol. 2012, 88, 175–187.
211. Huang, P.; Feng, L.; Oldham, E. A.; Keating, M. J.; Plunkett, W. Nature 2000, 407,
390–395.
212. Bechet, D.; Couleaud, P.; Frochot, C.; Viriot, M.-L.; Guillemin, F.; Barberi-Heyob,
M. Trends Biotechnol. 26, 612–621.
213. Chatterjee, D. K.; Fong, L. S.; Zhang, Y. Adv. Drug Deliv. Rev. 2008, 60, 1627–1637.
214. Park, H.; Na, K. Biomaterials 2013, 34, 6992–7000.
215. Zhang, C.; Shi, G.; Zhang, J.; Niu, J.; Huang, P.; Wang, Y.; Wang, W.; Li, C.;
Kong, D. Nanoscale 2017, 9, 3304–3314.
216. Ding, H.; Yu, H.; Dong, Y.; Tian, R.; Huang, G.; Boothman, D. A.; Sumer, B. D.;
Gao, J. J. Control. Release 2011, 156, 276–280.
217. Jesenská, S.; Plı́štil, L.; Kubat, P.; Lang, K.; Brožová, L.; Popelka, Š.; Szatmáry, L.;
Mosinger, J. J. Biomed. Mater. Res. A 2011, 99, 676–683.
218. Hu, Y.; Kanka, J.; Liu, K.; Yang, Y.; Wang, H.; Du, H. RSC Adv. 2016, 6,
104819–104826.
219. Lacombe, S.; Pigot, T. Cat. Sci. Technol. 2016, 6, 1571–1592.

220. Zhang, Y.; Dai, T.; Wang, M.; Vecchio, D.; Chiang, L. Y.; Hamblin, M. R.
Nanomedicine 2015, 10, 603–614.
221. Wu, X.; Huang, Y.-Y.; Kushida, Y.; Bhayana, B.; Hamblin, M. R. Free Radic. Biol.
Med. 2016, 95, 74–81.
222. Huang, L.; Szewczyk, G.; Sarna, T.; Hamblin, M. R. ACS Infect. Dis. 2017, 3,
320–328.
223. Kuimova, M. K.; Yahioglu, G.; Ogilby, P. R. J. Am. Chem. Soc. 2008, 131, 332–340.
224. Agazzi, M. L.; Ballatore, M. B.; Reynoso, E.; Quiroga, E. D.; Durantini, E. N. Eur. J.
Med. Chem. 2017, 126, 110–121.
225. Lai, Y. C.; Chang, C. C. J. Mater. Chem. B 2014, 2, 1576–1583.
226. Hyland, M. A.; Hewage, N.; Panther, K.; Nimthong-Roldán, A.; Zeller, M.;
Samaraweera, M.; Gascon, J. A.; Br€ uckner, C. J. Org. Chem. 2016, 81, 3603–3618.
227. Roxin, A.; Chen, J.; Paton, A. S.; Bender, T. P.; Zheng, G. J. Med. Chem. 2014, 57,
223–237.
228. Roxin, A.; MacDonald, T. D.; Zheng, G. J. Porphyr. Phthalocya. 2014, 18, 188–199.
229. Esemoto, N. N.; Yu, Z.; Wiratan, L.; Satraitis, A.; Ptaszek, M. Org. Lett. 2016, 18,
4590–4593.
CHAPTER TEN
Formic Acid as a Hydrogen Carrier

for Fuel Cells Toward a Sustainable
Energy System
Hajime Kawanami*,1, Yuichi Himeda†, Gábor Laurenczy‡
*Research Institute for Chemical Process Technology, National Institute for Advanced Industrial Science
and Technology, Sendai, Japan
†
Research Institute of Energy Frontier, National Institute for Advanced Industrial Science and Technology,
Tsukuba, Japan
‡ cole Polytechnique Federale de Lausanne (EPFL), Lausanne,
Institut des Sciences et Ingenierie Chimiques, E
Switzerland
1
Corresponding author: e-mail address: h-kawanami@aist.go.jp
Contents
1. Introduction 396
2. Liquid Organic Hydrogen Carriers 397
3. Recent Organic Materials for LOHC 398
4. Formic Acid for LOHC 401
5. Homogeneous Catalytic Dehydrogenation of Formic Acid 404
6. Cp* With Iridium Complex for H2 Generation From Formic Acid 408
7. High-Pressure H2 Generation 417
8. Application for Fuel Cell Batteries 421
9. Conclusion 424
References 425
Abstract
Formic acid is considered as one of the promising organic liquid hydrogen carriers for
the next generation; it can offer a viable method for safe hydrogen transport. In this
chapter, we introduce the potential of formic acid in terms of thermodynamics and
mechanism as described in earlier work in this area, as well as homogeneous catalysts
providing a viable method for the production of molecular hydrogen as a sustainable
fuel source through dehydrogenation. In addition, pentamethylcyclopentadienyl irid-
ium (Cp*Ir) catalysts are also focused upon for this reaction and shown as a strategy
to improve catalyst activity by introducing hydroxyl groups to increase turnover num-
bers. One of the major advantages of using formic acid as a hydrogen source is the
regeneration of formic acid through the interaction with carbon dioxide, thus
maintaining a continuous cycle, and offers a possibility for high energy output appli-
cations. The developed catalyst, Cp*Ir has potential to produce hydrogen gas with
very high pressure, 120 MPa, without facing the problem of decomposition. The gen-
erated gas pressure is sufficient for feeding a fuel cell vehicle, which requires 75 MPa,

396 Hajime Kawanami et al.
according to the present standard of a hydrogen gas station. Furthermore, even

though the generated gas consists of hydrogen and carbon dioxide with the ratio
of 1:1, hydrogen can be separated easily and purified from the generated gas under
supercritical conditions, by simply cooling to change the gas–liquid state of the sys-
tem while maintaining the pressure. Finally, we introduce some applications of this
gas generation system in fuel cells, and also for the production of electric power. It
is worth mentioning that commercialization of the developed process for hydrogen
generation via transformation of formic acid may be achievable in the near future.
1. INTRODUCTION
The increasing demand of energy especially in the transportation sec-
tor is diminishing fossil fuel supplies, and there are escalating environmental
concerns such as global warming (1). Recently renewable energy resources,
such as solar and wind power, geothermal energy, biomass energy, and
ocean energy, are receiving considerable attention in order to develop a sus-
tainable system (2). However, many researchers are focusing on these
renewable energy systems as there is not yet any widely applicable, practical
resolution. Nowadays, hydrogen gas (H2) can be considered as one of the
promising alternative clean fuels to replace conventional fossil fuels, and it
can be produced from any primary energy source. As a fuel, H2 can be used
either through direct introduction to an internal combustion engine or to a
fuel cell, and produces only water as a by-product (3,4). Thus, hydrogen can
be considered as one contender for zero-emission technology, which would
improve air quality especially of urbanized areas (5).
In Europe, a Strategic Research Agenda by the European Hydrogen and
Fuel Cell Technology Platform was published in 2004 for the development
of the necessity of hydrogen technologies in production, storage, transport,
and application in stationary and mobile systems (6). A development strategy
was also reported in 2004 for the technical, socioeconomic, and political
challenges of deploying world-class, competitive hydrogen technology
and fuel cell applications, and recommended courses of action (7).
In the United States, the Department of Energy published a hydrogen
and fuel cell program plan in September, 2011 (8), and clear technology
development targets are set, and progress is frequently assessed. In Califor-
nia, demonstration fleets of fuel cell vehicles are in use on the roads, and
state legislation on emissions represents a strong driving force for clean
vehicles.
Formic Acid as a Hydrogen Carrier for Fuel Cells 397
In Japan, after the Great East Japan Earthquake and the accident at the
Tokyo Electric Power Company (TEPCO)’s Fukushima Daiichi Nuclear
Power Plants in 2011, the energy situation changed drastically, both domes-
tically and abroad. Then in 2014, METI (the Ministry of Economy, Trade
and Industry, Japan) produced the fourth Strategic Energy Plan for Japan’s
new direction of energy policy. Within the plan a strategic road map for
hydrogen and fuel cells was published in June, 2014. This was to enable
the rapid expansion of hydrogen utilization such as hydrogen power gener-
ation, establishment of a large-scale hydrogen supply system, and a totally
carbon dioxide-free hydrogen supply system to form a “hydrogen society.”
2. LIQUID ORGANIC HYDROGEN CARRIERS

Every year, 60 million tons of hydrogen are produced in chemical
industry which mainly comes from natural gas (48% of total hydrogen pro-
duction), 30% from heavy oil and naphtha, 18% from coal, and 4% through
electrolysis (3,9). Steam reforming of natural gas and light hydrocarbons
(methane, ethane, etc.) is a commonly used technique of hydrogen produc-
tion. Basic reactions of steam reforming to produce hydrogen are shown as
Eqs. (1) and (2)
Cn Hm + nH2 O ! nCO + ðn + m=2ÞH2 (1)
CO + H2 O ! CO2 + H2 (2)
Generally, the reaction was carried out at very high temperatures of
700–850°C using a nickel catalyst. The disadvantages associated with this
process are the formation of large amounts of CO2 and the use of fossil fuels
for heating.
Similarly, the production of hydrogen from coal gasification (reaction
scheme shown in Eq. 3) also generates CO2 along with small amounts of
CO and methane
C + 0:2H2 + 0:6O2 + 0:8H2 O ! CO2 + H2 (3)
Therefore, to remove CO2 from H2 an additional step was required,
which involves washing with monoethanolamine or potassium hydroxide
to obtain 97%–98% of H2. Recently, technological development on the
production of hydrogen as a long-term energy fuel with little or no pollution
is receiving considerable attention, especially for safe and cost-effective stor-
age and transportation. However, application of H2 as an alternative fuel is
still in its infancy because of the associated difficulties, particularly, its low
volumetric energy density and gaseous properties (10,11). To overcome
these difficulties, several methods were developed considering: (1) hydrogen
compression under high-pressure conditions, (2) hydrogen liquefaction at
low temperature, (3) hydrogen adsorption in metal hydrides, (4) cryogenic
storage with hydrogen adsorbing materials, and (5) hydrogen storage in liq-
uid organic hydrides.
Among these methods, hydrogen storage in liquid organic hydrides has
several beneficial aspects in terms of environmental, economical, technical,
and social usage. The concept of using liquid organic hydrides is preferred
because of the advantage of the capability of catalytic hydrogenation and
dehydrogenation in a cyclic manner. This concept has been investigated
in the Euro-Quebec Hydro Hydrogen Project regarding liquid hydrogen
and methyl cyclohexane; in addition various other liquid organic hydrides
were developed (12–14).
3. RECENT ORGANIC MATERIALS FOR LOHC

Various reagents such as ammonia borane, N-ethylhydrocarbazole,
methyl cyclohexane, hydrazine, decalin, methanol, formic acid, and
ammonia have also become attractive, gradually as Liquid Organic Hydro-
gen Carrier (LOHC) as they provide significant advantages in terms
of availability, recharging, and safety (Table 1) (10,11,15,16). Well-
known LOHC candidates are cyclic aliphatic hydrocarbons, such as
methylcyclohexane (MCH), cyclohexane, decalin, and bicyclohexane,
which have a gravimetric hydrogen capacity in the range of 6–8 wt%
and a volumetric hydrogen density of 60–62 g L1 (17–29). MCH, which
can generate H2, then transforms to toluene (MTH cycle), is one of the
most important candidate as a LOHC. Many reports were published about
the cyclohexane, cyclohexane–benzene–hydrogen (CBH) cycle. Although,
better reversibility and selectivity than that of the MTH cycle were
obtained, CBH cannot be considered as a preferred material as an LOHC
because it is a carcinogenic system. Recently, benzyltoluene (HOBT)/
perhydrobenzyltoluene (H12-BT) and dibenzyltoluene (HODBT)/
perhydrodibenzyltoluene (H18-DBT) LOHC systems have been investi-
gated and demonstrated; they exhibit higher volumetric storage density,
easier hydrogen purification and reduced toxicity, properties that yield a
significant advantage over the TOL/MCH system (30).
Table 1 Hydrogen Storage Properties of Various LOHC Including H2 and Water
Volumetric
Liquid Organic Boiling Density Energy Gravimetric Content Reaction
Hydrogen Carrier Point (°C) (g mL21) (kJ molH221) Content (wt%) (kgH2 m23) Temperature (°C)
Liquid H2 252.8 0.071 0.9 100 70 —
Water 100 1.00 286 11.2 111 ca. 1000
Methyl cyclohexane 100.4 0.77 68 6.2 47 >350
Cyclohexane 80.7 0.78 65.3 7.2 56 >300
Decaline 193/185 0.90/0.87 63.2/66.1 7.3/7.3 32.4/32.4 >200
(cis/trans) (cis/trans) (cis/trans) (cis/trans) (cis/trans)
Bicyclohexane 227 0.86 66.6 7.3 32
Ammonia 33.4 0.73 31 17.8 121 >400
Methanol 64.6 0.79 44 12.1 100 >250
Formic acid 100.8 1.22 31 4.3 53 <100
The simplest alcohol, methanol (MeOH), can also be applicable as an

LOHC. Methanol can be stored and handled very easily at ambient condi-
tions. Industrially, methanol can be produced from various sources such as
low-cost biomass on a large scale, or by the reduction of CO2 with H2 in the
presence of an appropriate catalyst, which in turn is converted to hydrogen
via steam reforming. The general process of hydrogen production from
methanol reforming, in the presence of a heterogeneous catalyst, involves
high temperature (200–300°C), high pressure (2.5–3 MPa), and steam
(H2O) as shown in Eq. (4)
CH3 OH + H2 O ! CO2 + 3H2 (4)

CH3 OH ! CO + 2H2 (5)
CO + H2 O ! CO2 + H2 (6)
A disadvantageous aspect is the generation of impurities within CO

through methanol decomposition as described in Eq. (5); this poisons the
fuel cell catalysts and should be minimized below 10 ppm. For example,
in the case of a successful commercialized fuel cell system, the ElectraGen™
power generator, CO is transformed by the water-gas shift reaction (Eq. 6)
to CO2, then purified H2 can be obtained (31–33).
Dehydrogenation at low temperature, catalyzed by organometallic cat-
alysts, can only promote the liberation of hydrogen from alcohol along with
the corresponding aldehydes, ketones, and esters including polyesters.
ð7Þ
ð8Þ
ð9Þ
ð10Þ
ð11Þ
ð12Þ
An early example of dehydrogenation without acceptors was reported by

Robinson et al. They used perfluorocarboxylate complexes (M(OCO-
RF)2(CO)(PPh3)2: M ¼ Ru, Os; RF ¼ CF3, C2F5, or C6F5) and achieved
an initial turnover frequency (TOF) of 8172 h1 for the dehydrogenation
of benzyl alcohol (34,35). Although the catalyst shows high TOFs, methanol
cannot be dehydrogenated, and the result is attributed to the slightly stronger
α-CH bond in methyl groups compared to larger alkyl groups. Recently,
hydrogen production from methanol at low temperature in the presence
of a homogeneous catalyst was extensively investigated and the process pro-
moted (36–39). One of the examples is the aqueous phase dehydrogenation
of methanol developed by Beller and coworkers in 2013 at low temperature
(<100°C) and ambient pressure in the presence of [Ru(H)(Cl)(CO)(HN
(C2H4PiPr2)2)] (iPr ¼ CH(CH3)2) as the precatalyst and under strongly alka-
line conditions (8.0 M KOH) (40). Due to the limited availability and high
cost of precious metals, Fe complexes with pincer ligands were investigated
for participation in the reaction. With the catalyst that was studied, a high-
TOF of over 700 h1 in the first hour and a good TON of almost 10,000 in
46 h were achieved (41). Furthermore, a manganese complex that largely
resembles the Fe complex with the pincer ligands was also developed for
the dehydrogenation of methanol, whereupon the TON was reduced to
68 during the same time. Interestingly, this manganese catalyst can be appli-
cable for the dehydrogenation of different substrates, such as ethanol, para-
formaldehyde, and formic acid (42). By using various organometallic
catalysts, the hydrothermal dehydration of primary alcohols was also
reported to accelerate the dehydration of alcohols into their corresponding
aldehydes (41–43). Nevertheless, the TOF and TON obtained were still too
low for large-scale production and the requirement of a strongly basic con-
dition for activation is still a concern.
4. FORMIC ACID FOR LOHC

Formic acid can be utilized commercially in various fields, such as in
industry, grass silage, leather tanning, antiicing, textile dyeing, finishing,
food additives, natural rubber, drilling fluids, and various chemical processes.
The worldwide production of formic acid was about 621,000 t/a in 2012.
One of the main industrial routes is the carbonylation of methanol (Eq. 12)
and subsequent hydrolysis of methyl formate (Eq. 13). BASF (Germany,
BASF process), Kemira (Finland, Kemira-Leonard Process), Feicheng Acid
Chemicals (China), and Luxi Chemical Group (China) produce formic acid
by this route
CH3 OH + CO ! HCOOMe (12)

HCOOMe + H2 O ! HCOOH + CH3 OH (13)
Formic acid can also be produced by other methods such as (i) from for-
mates (usually sulfuric acid); (ii) by-products from pentaerythritol,
trimethylolpropane, and 2,2-dimethyl-1,3-propanediol; and (iii) from bio-
mass (Eq. 14)
ð14Þ
The hydrogenation reaction of carbon dioxide to formic acid is receiving

much attention again, and there are many published reports. The early tech-
nology of formic acid production from CO2 in the presence of catalysts
(especially soluble Ru complexes are preferred) was first introduced by
BP Chemicals in 1980s and then developed by BASF. In their method
the reaction takes place in a mixture of CO2, tertiary amine, and hydrogen
in alcohol at 10–12 MPa to form a formic acid–amine complex, followed by
thermal dissociation at 150–185°C. This process faces issues involving the
use of an expensive catalyst, the desirability of maintaining its activity,
and therefore its reuse, and further, whether the catalyst can be separated
from formic acid to avoid its decomposition.
The physical properties of formic acid are also favorable for the use as an
LOHC. Formic acid is a low flammable, biodegradable, and stable liquid
under ambient conditions, with a boiling point of 101°C and a freezing
point of 8.3°C. It is a colorless, clear, and corrosive liquid with a pungent
odor (44). Even though its acidity is pKa 3.74, which is the strongest among
Fig. 1 A photo of a plastic bottle for formic acid, commercially available in Japan.
unsubstituted alkyl carboxylic acids, formic acid is extremely stable. Thus,

transportation, handling, and storage of formic acid are feasible under typical
infrastructure conditions. Furthermore, dilute formic acid (less than 78%) is
not designated as a hazardous material under Fire Defense Laws in Japan
(Fig. 1) and is listed on the US Food and Drug Administration’s list of food
additives in the United States.
Ready availability of formic acid through large-scale production,
in combination with unique properties, makes it a most promising material
among the LOHCs. Considering the criteria for LOHC, formic acid has a
4.4 wt% of hydrogen content, corresponding to a volumetric capacity of
53 g L1; these values are not so high as water (11.2 wt%, 111 kg m3), liq-
uid NH3 (17.8 wt%, 0.1 kg m3), and methanol (12.6 wt%, 100 kg m3),
but almost equal with 100 MPa of compressed H2 or MCH (6.2 wt% of
hydrogen content and 47 kg m3 of volumetric capacity) (Fig. 2) (45).
Dehydrogenation of formic acid has a low-reaction enthalpy among
those of the other H2 storage chemicals as shown in Table 2 (46–50); thus,
H2 can be produced from formic acid at moderate temperatures (lower
than 100°C) that match with targets for releasing chemically stored H2,
according to the DOE (Department of Energy) (51). Therefore, less
Fig. 2 Volumetric vs gravimetric hydrogen densities.
energy is required for H2 production from formic acid and could, there-
fore, be a more attractive H2 storage material. Moreover, carbon dioxide
(CO2), which is the coproduct of formic acid dehydrogenation, can be
allowed to hydrogenate back to formic acid in water or organic solvents
on the catalyst surface or in the presence of specific homogenous catalysts
(52–55). Therefore, formic acid can be shown to be a renewable chemical
for H2 storage (56–65) (Fig. 3).
5. HOMOGENEOUS CATALYTIC DEHYDROGENATION

OF FORMIC ACID
Formic acid has a potential to be used as a valuable LOHC as outlined
earlier. Beside the dehydrogenation process, there is also a possibility of the
formation of water and carbon monoxide from dehydration of formic acid.
The generation of CO is fatal to fuel cells, because of facile catalyst poisoning
(51). In the presence of water, formic acid undergoes dehydration and dehy-
drogenation reactions (56,59), but both of the reactions remain unselective
in the aqueous phase because the difference between the enthalpies of forma-
tion of CO and CO2 (Eqs. 14 and 15) is very small. Thus, to avoid the dehy-
dration reaction, it is necessary to use an effective catalyst to dehydrogenate
formic acid selectively under moderate temperatures (lower than 100°C)
Table 2 Thermodynamic properties of the dehydrogenation reactions
ΔrH0/(kJmol21) ΔrG0/(kJmol21)
HCO2H(I) ! H2(g) + CO2(g) +31.2 33.0
HCO2H(I) ! H2O(I) + CO(g) +28.4 13.0
HCO2H (aq) ! H2(g) + CO2(g) +32.0 43.2
HCO2NH4 (aq) ! H2(g) + CO2(g) + NH3(g) +84.3 +9.5
CH4(g) + H2O(g) ! 3H2(g) + CO(g) +206.1 +140.9
CO(g) + H2O(g) ! H2(g) + CO2(g) 41.2 28.6
H2O(I) ! H2(g) + 1/2O2(g) +285.8 +238.2
NH3(g) ! 3/2H2(g) + 1/2N2(g) +46.1 +16.5
C6H11CH3(I) ! 3H2(g) + C6H5CH3(I) +202.2 +93.3
CH3OH(I) + H2O(I) ! 3H2(g) + CO2(g) +131.0 +9.0
(I), (aq) and (g) are represented as liquid phase, aqueous solution phase and gas phase of substances, respectively.
Fig. 3 Schematic of a cycle for sustainable hydrogen generation and storage with
formic acid.
HCO2 HðlÞ ! H2 ðgÞ + CO2 ðgÞ ΔG ¼ 32:9kJ=mol,

ΔH ¼ 31:2kJ=mol (15)
HCO2 HðlÞ ! H2 OðgÞ + COðgÞ ΔG ¼ 12:4 kJ=mol,
ΔH ¼ 28:7kJ=mol (16)
Early studies of the decomposition and dehydrogenation from formic

acid in the presence of a homogeneous catalyst were reported in the
1960s by Coffey (66), Otsuka (67), Strauss (68), focusing on Pt- (69),
Rh-, Ir-, and Pd-based molecular catalysts (70–73). Among these catalysts,
[IrH3(PPh3)3] was the most rapid, with the initial TOF of 8890 at 118°C,
under refluxing conditions. Even though no CO was detected by this reac-
tion, a metal carbonyl species was formed. Then, further studies progres-
sed gradually, and since then many catalysts have been applied in attempts
to generate H2 with high efficiency from the decomposition of formic acid
(57,58,60–64). Among the different metals used, the catalytic activity of a
Rh complex has been studied by Beller and Laurenczy and coworkers. They
independently screened the catalytic activity with various types of Ru com-
plexes. Moreover, homogeneous catalysts containing nonnoble metals such
as Fe were also tested in efforts to develop a cost-effective method (62,74–78).
Beller and coworkers reported a high activity of Fe(BF4)2 6 H2O with the
P(CH2CH2PPh2)3 ligand (abbreviated as PP3 in Table 3). Recently, Hazari
and Schneider and coworkers obtained the highest TOF of 196,728 h1
and turnover numbers up to 983,642 within 9.5 h. Studies on the dehydro-
genation of formic acid using homogeneous catalysts are summarized in
Table 3.
Table 3 Homogeneous Catalysts for the Dehydrogenation of Formic Acid
No. Catalyst Solvent Temp. (°C) TON TOF (h21) Conversion Remarks References
1 [IrH3(Ph3)3] Acetic acid 118 >11,000 8890 — No CO (66)
2 [Pt(2-Pr3P)3] Acetone/water 20 25 100 100 — (67)
3 Rhl/Nal Water 100 — 4.4 — Significant (73)
amount of CO
4 RhCI33H2O/NaNO2 Water 90 — 126 12.5 — (70)
5 [Fe3(CO)12]/PPh3/tpy DMF 40 200 — — With NEt3 (74)
6 [Fe3(CO)12]/PBn3/tpy DMF 40 1266 — — With NEt3 (75)
7 [Fe(BF4)2]6H2O/PP3 Propylene 80 92,417 9425 — (62)
carbonate
9 FeCl2/PP3TS Water 80 1000 133 100 (76)
10 [(tBu-PNP)Fe(H)2(CO)] Dioxane 40 >999 653 — With NEt3 (77)
11 [(PNP)Fe(CO)H(HCO2)]/LiBF4 Dioxane 80 983,642 196,728 — (78)
6. CP* WITH IRIDIUM COMPLEX FOR H2 GENERATION

FROM FORMIC ACID
Himeda et al. studied a series of iridium catalysts containing penta-
methyl cyclopentadienyl (Cp*) iridium with various bipyridine and azol
ligands, without any phosphine derivatives. At first, they investigated the
dehydrogenation in the presence of SO4 salts of the Cp*Ir complex with
the 4,40 -dihydroxy-2,20 -bipyridine (4DHBP) ligand for solvation into the
aqueous formic acid solution (79). In 2 M aqueous formic acid (FA) solu-
tion, a gas mixture of H2 + CO2 was evolved with high TOF values at
40–90°C, and the maximum of 14,000 h1 was achieved at 90°C
(Fig. 4). Almost all the FA was decomposed after the reaction, and no
CO formation was detected by a GC-FID equipped mechanizer. An Arrhe-
nius plot permitted calculation of the activation energy for the decomposi-
tion of FA to be 76 kJ mol1 in the case of the Cp*Ir complex catalyst,
whereas the activation energy was 87 kJ mol1 when [Cp*Rh(bpy)
(H2O)]2+ was used.
Fig. 4 Time course of evolved gas volume from the catalysis of formic acid conversion.
Fig. 5 The dependence of the dehydration of FA at various pH values.
An important experimental parameter with respect to the catalytic pro-

cess is the pH (Fig. 5). When the pH is adjusted by the addition of sodium
formate (SF) from 1.8 to 7.3, an increasing pH value caused a decrease of
the initial TOF values as well as the conversion of FA. Once above pH 4.5,
the reaction does not proceed and no gas evolution was detected.
According to the results, the structure of the iridium catalyst would be
changed under the various pH conditions. When the 4DHBP ligand is dis-
solved in an aqueous solution of formic acid, it assumes the bipyridinol
form (A) in the lower pH range. At higher pH conditions, hydroxyl groups
are reversibly deprotonated to generate an oxyanion group (O), which is
a Lewis base with a stronger electron donating ability than that of its con-
jugate acid. The deprotonation of hydroxy groups can also cause signifi-
cant changes in the electronic properties and water solubility of the
complex (80,81). Consequently, the ligand changes to the bipyridinolate
form (B and C, Scheme 1) (82). It prevails that an important factor in
improvement of the activity of the Ir catalyst is functionalization of the
bipyridine ligand.
Previously Himeda et al. also studied the effects of substitution on the
bipyridine ligand on the catalytic activity by using the hydrogenation of
2-cyclohexene-1-one as a substrate. They evaluated its activity in terms
of a Hammett type correlation with a substituent constant (σ +p ) at pH 2.6
and 7.3, respectively (83). A similar tendency was also observed in the
Scheme 1 The acid–base equilibrium between the hydroxy and oxyanion forms, and
the resonance structures of the oxyanion form.
Fig. 6 Hammett type plot of log(TOFR/TOFH) vs σ p+ values of substituent (R) in the cat-
alyst at 60°C in 1 M aqueous formic acid solution (10 mL).
dehydrogenation of formic acid in the presence of the Cp*Ir catalyst

(54,79,84). The substitution effect was prominent on the initial TOF of
dehydrogenation. In the acidic conditions, the initial TOF of the hydroxyl
catalyst was approximately 90 times and 2 times higher than that of the
unsubstituted bipyridine and methoxy catalysts, respectively (Fig. 6).
Fig. 7 Cp*Ir complex with 4,7-dihydroxy-1,10-phenanthroline ligand.
Fig. 8 The pH dependence of the solubility of Cp* catalyst with PHDP ligand and 4DHBP
ligand. Solid squares represent DHBP and open circles represent PHDP.
An interesting phenomenon regarding the solubility of the Cp*Ir catalyst

with 4,7-dihydroxy-1,10-phenanthroline (DHPH; Fig. 7) can be seen in
Fig. 8. The catalyst solubility changed drastically with pH, and the complex
easily precipitates. In the pH range from 4.5 to 5, the DHPH Ir catalyst has
very low solubility, around 0.1 ppm (85). The dehydrogenation of formic
acid is processed following the change in the pH of the system, which
increased from acidic to neutral (around 7) and eventually, the catalyst
started to precipitate and stopped functioning as a catalyst. Thus, it can be
separated easily without any decomposition and ready for recycling (86).
The effect of the position of the substituent on the catalytic activity was
also investigated (Fig. 9) (87). The catalytic activity was evaluated using
Fig. 9 Structure of the Cp*Ir complex with BPY ligands.
Table 4 Results of Dehydrogenation of HCO2H Catalyzed by [Cp*Ir(L)(OH2)]SO4

Concentration Initial TOF
L of Catalyst [mM] [h21] TON References
1 bpy 1.5 18 280 (87)
2 3DHBP 0.2 440 780 (87)
3 4DHBP 0.2 1800 5000 (87)
4 4DHBP 0.2 2400 5000 (54)
5 5DHBP 1.0 32 280 (87)
6 6DHBP 0.2 2200 5000 (87)
7 6DHBP 1.0 2450 10,000 (88)
transfer hydrogenation, dehydrogenation, and hydrogenation of CO2. The

Cp*Ir complex having diols at 4,40 -(4DHBP) and 6,60 -positions (6DHBP)
exhibited high catalytic activity which is attributed to the electron donating
capability of the hydroxy groups at ortho- (6-) and para-(4-) position
(Table 4) (54,87,88).
The reaction mechanism of the dehydrogenation of formic acid cata-
lyzed by various catalysts can be explained, as shown in Fig. 10. It is consid-
ered to consist of three steps: (step I) formation of the formato complex
B with formic acid, then (step II) release of CO2 by β-hydrogen elimination
to generate the iridium hydride complex C (89), and (step III) production of
H2 from the reaction of [Ir]-H and H+.
The principal difference between complexes containing 4DHBP and
6DHBP groups is the ortho and para positions of the OH groups. Experimen-
tal and computational methods reveal that hydroxyl groups at ortho positions
lead to a proton relay incorporating a molecule of H2O by a “pendant base”
Fig. 10 Reaction mechanism for the dehydrogenation of formic acid catalyzed by Cp*Ir
catalyst.
Scheme 2 H2 generation from formic acid enhanced by the pendant base effect
through a proton relay (proton transfer) in step (B).
effect that facilitates the heterolysis of H2 in CO2 hydrogenation (90).

According to the calculations, however, H2 addition through heterolytic
cleavage is rate-limiting, H2 heterolysis with the “proton relay” by the pen-
dant base is preferred as shown in Scheme 2 (88). Thus, in the reverse reac-
tion, the enhancement by the 6DHBP ligand hydroxyl groups at the ortho
position was also observed in the case of formic acid formation from
CO2 and H2. Namely, when the reaction condition is acidic, a water mol-
ecule in the form of a hydronium ion, and a hydroxyl group at the ortho posi-
tion can also form a proton relay and assist the reaction of [Ir]-H with a
proton (Scheme 2). The proton relay stabilizes the [Ir]-H2 transition state
and lowers the energy barrier for generating H2.
To understand the rate-determining step of the dehydrogenation of
formic acid, deuterium kinetic isotope effect (KIE) studies were performed
in the presence of Cp*Ir-4DHBP and -6DHBP. The results are summarized
in Table 5.
Table 5 Deuterium Kinetic Isotope Effect (KIE)

4DHBP 6DHBP 6DHBP
21 21
Substrate /Solvent TOF (h ) KIE TOF (h ) KIE TOF (h21) KIE References
1 HCOOH/H2O 2400 — 2160 — 5400 — (88)
2 HCOOH/D2O 1140 2.1 1610 1.3 1130 1.2 (88)
3 DCOOD/H2O 1660 1.4 1100 2.0 1340 1.8 (89)
4 DCOOD/D2O 940 2.6 905 2.4 2560 2.1 (88)
When the substrates or solvents were replaced with deuterated materials,

the reaction rate decreased considerably. In the case of 6DHBP, TOF values
were similar for reactions at the different pH values of 1.8 (1 M HCO2H)
and 3.5 (1 M HCO2H/HCO2Na (1:1)), and the result indicates that the
rate-determining step remained unaffected by pH values. Then, when
D2O was used instead of H2O, the KIE of the 4DHBP complex was 2.1,
whereas that of the 6DHBP complex was 1.2 (Table 5, entry 2), and when
the substrate (HCO2H) was replaced by DCO2D, the KIE values in the case
of the 4DHBP and 6DHBP complexes were obtained as 1.4 and 2.0, respec-
tively (Table 5, entry 3). These KIE experiments suggest that the deuterated
solvent (D2O) influences the reaction rate to a greater extent than the deu-
terated substrate (DCO2D) in the case of the 4DHBP complex. Accord-
ingly, when 4HHBP was used, the H2 (HD) release step for the reaction
of [Ir]-H and H+ (D+) (step III, Fig. 10) is the rate-determining step rather
than the formation of [Ir]-H from the formato complex. Thus, a high proton
concentration (low pH) will lead to high reaction rates, whereas the reaction
rate will decrease with increasing pH of the reaction solution, which is con-
sistent with the pH dependence in the case of the 4DHBP complex (Fig. 5).
On the other hand, in the case of the 6DHBP complex, KIE values were
1.3 and 2.0 with deuterated solvent (D2O) and substrate (DCO2D), respec-
tively (Table 5, entries 2 and 3). Accordingly, when 6DHBP was used as a
ligand, the generation of [Ir]-H from the formato complex (step II, Fig. 10)
should be the rate-determining step rather than H2 release from the reaction
of [Ir]-H with H+. This is consistent with the DFT calculations previously
reported (90).
According to these results, the more favorable ligand for the Cp*Ir com-
plex was developed by extending the number of OH groups, and the pen-
dant base effect. The examples of developed catalysts (THBP, TH4BPM,
THBPM) are shown in Fig. 11, and their corresponding catalytic activities
Fig. 11 Structure of the Cp*Ir complex with THBP, TH4BPM, and THBPM ligands.
Table 6 Results of Dehydrogenation of HCO2H Catalyzed by Cp*Ir Catalysts

Concentration Initial TOF
L of Catalyst [mM] [h21] TON References
1 4DHBP 0.2 2400 5000 (79)
2 THBP 0.2 3890 7650 (88)
3 TH4BPM 0.05 39,500 11,000 (88)
4 THBPM 0.0015 158,000 308,000 (54)
5 THBPM 0.0031 228,000 165,000 (54)
are shown in Table 6. Introducing four hydroxy groups at ortho and para
positions, the catalytic activity of THBP was improved with the high
TOF (3890 h1) and TON (7650) values compared to these of 4DHBP
and 6DHBP containing species (Table 6, entry 2). Interestingly, when a
pyrimidine-based ligand was introduced, cited as TH4BPM in Fig. 11,
TOF and TON were drastically improved over 10 times (39,500 h1)
and 2 times, respectively, than the values for the 4DHBP complex. In addi-
tion, a dinuclear complex, (Cp*Ir)2(THBPM), showed further higher activ-
ity with the TOF of 228,000 h1 at 90°C and TON of 308,000 at 80°C
under the optimal reaction conditions (54).
The Cp*Ir(THBPM) catalyst has four OH groups at the ortho and para
positions to activate the catalyst. When THBPM was used as a ligand, much
faster gas generation compared to the case for 4DHBP as ligand was
observed, as shown in Fig. 12. For the optimization of the catalyst, the
pH dependence of the reaction was investigated, and the maximum can
be seen at pH 3.5 with TOF of 31,600 h1 (Fig. 13), which is close to
the pKa of the catalyst (pKa 3.8) and to the pKa of formic acid (pKa
3.75). These data suggest the OH groups on the THBPM ligand play
not only the “pendant base” role, but also other critical roles in the
dehydrogenation.
Fig. 12 Time course of the gas evolution from formic acid.
Fig. 13 Optimization of pH conditions for the dehydrogenation of formic acid.

7. HIGH-PRESSURE H2 GENERATION
Formic acid dehydrogenation is thermodynamically favorable, so that
high-pressure H2 is generated easily from formic acid rather than other H2
storage chemicals. Thus, there is the possibility of high-pressure gas gener-
ation by the decomposition of formic acid to the mixture of gas with H2 and
CO2. The first example was demonstrated by Laurenczy et al., in aqueous
solution using hydrophilic ruthenium-based catalysts, generated from the
highly water-soluble ligand meta-trisulfonated triphenylphosphine with
either [Ru(H2O)6]2+ or, more conveniently, commercially available RuCl3.
The generated H2/CO2 pressure was typically between 1 and 220 bar, but
no inhibition of catalytic activity was observed up to a pressure of 75 MPa
(Fig. 14). The total conversion did not reach 100% because 10% of SF added
for the activation of the catalyst remained unconverted; however, all the
formic acid was consumed. Notably, the decomposition of formic acid
under high-pressure conditions prevents the generation of CO and H2O
as confirmed by the analysis of a gas sample using FTIR spectroscopy (detec-
tion limit of 3 ppm). The continuous evolution of gases from formic acid
was also evaluated under high-pressure conditions (typically 5–25 MPa),
which was systematically verified after prolonged addition of formic acid.
The maximum gas out flow produced was nearly 600 mL min1 at 120°C
Fig. 14 Kinetic trace of formic acid decomposition in a closed system with a pressure
increase to 75 MPa.
with [Ru(H2O)6]2+ (1.5 mmol) as the precatalyst. The catalyst’s life time
is over 1 month with the TOF of 230 5 h1, and the TON exceeded
40,000 cycles without any deactivation (53,57).
Seven years after the first report on high-pressure gas generation from
formic acid, Iguchi and Kawanami et al. reported very high-pressure gas
generation, over 120 MPa, from the decomposition of formic acid in the
presence of a water-soluble iridium catalyst at a moderate temperature of less
than 80°C (86,91). They used the Cp*Ir complex bearing 4DHBP as the
catalyst and obtained 123 MPa maximally (Fig. 15), but thermodynamic cal-
culations predicted the possibility of the generation of high-pressure gas at
225 MPa, using the method that had been developed. The TON was
37,000–38,000 in one batch with 92–93 mol% of high conversion at
40 MPa. The TOF value was decreased with the generated pressure from
9100 h1 at 0.1 MPa to 5700 h1 (2/3) at 10 MPa and 2500 h1 (1/4) at
40 MPa, respectively. Despite the successful generation of high-pressure
gas, the Cp*Ir catalyst was gradually decomposed. The catalyst undergoes
partial hydrogenolysis due to the presence of high-pressure H2 in the system,
resulting in an insoluble compound, which then precipitates after the reac-
tion. The deactivation mechanism predicted that the BPY ligand might be
Fig. 15 Time course of the pressure generated by the decomposition of FA at various

initial concentrations. The initial concentration of FA is as follows: 4 mol L1 (square),
10 mol L1 (triangle), 15 mol L1 (cross), 20 mol L1 (circle). Reaction conditions: 80°C,
aqueous solution of FA (4–20 mol L1, 13 mL), [Cp*Ir(4DHBP)(H2O)][SO4] (2.0 mmol L1,
26 mmol).
changing from its chelating conformation to another conformation under

the high-pressure H2 conditions. To resolve this problem, another catalyst
was developed; it contained a 1,10-phenantroline skeleton ligand (Fig. 7) to
prevent the rotation at the bond between the pyridine moieties of BPY by
bridging, and its potential was investigated in terms of catalytic activity,
durability, and reusability. The catalyst bearing a 1,10-phenantroline skele-
ton was successful in generating a much longer life time (more than 3
months) and the TON exceeded 5,000,000. The catalyst was also recyclable
after reaction. It could be separated by filtration more than 10 times without
any deactivation (Fig. 16).
From the view point of the catalyst cost, a noble metal complex cheaper
than iridium, viz. Ru, was employed for the generation of high-pressure gas.
Recently, Guan and Huang et al. used a Ru complex bearing a 2,20 -
biimidazoline ligand for the generation of high-pressure gas from formic
acid. The Ru catalyst can generate up to 24.0 MPa of high-pressure gas suc-
cessfully with 1.8 MPa of He (total 25.8 MPa), and no CO formation was
detected by GC-TCD (92) (Fig. 17). In the case of the Ru catalyst, the pH of
the system was maintained at 3.5. The reaction mixture consisted of formic
acid and sodium formate (FA:SF ¼ 1:1). The SF was not converted into H2
and CO2; thus, the maximum pressure obtained was lower compared to a
comparable reaction with the iridium complex.
A disadvantage of the method of production of H2 by the decomposition
of formic acid is that the generated gas requires purification for the use in
FCVs such as cars, buses, and forklifts. In the present system, FCVs equipped
PEFC (polymer electrolyte fuel cell), 99.999% of pure H2 at 35 or 70 MPa is
Fig. 16 Images of the reactant (catalyst 2.8 mmol; water, 3 mL; FA (100%), 1 mL) during
the reaction at different stages: (A) before the reaction at RT (20°C), pH 6.8;
(B) dissolution of the catalyst at the initial stage in aqueous FA solution at 50°C under
high pressure (22 MPa), pH 0.9; (C) during reaction at 50°C under high pressure
(22 MPa), pH 0.9; (D) after the reaction; and (E) the precipitation of catalyst after cooling
down to RT (20°C), pH 1.9.
30
25.8 MPa
25
20
Pressure (MPa)
15
10
H2:CO2 = 1:1
5
CO (<6 vol ppm)
0
0 1 2 3 4 5 6
Time (h)
Fig. 17 Time-dependent gas evolution through FA decomposition in the presence of Ru
catalyst bearing 2,20 -biimidazoline ligand. The reaction was carried out at 80°C in an
autoclave (internal volume is 7.0 mL) with 2 MPa of He gas, FA aqueous solution
(6.5 mol L1, 4.0 mL), and catalyst (8.0 mol, 2.0 mmol L1).
injected at the hydrogen station. For the preparation of high-pressure puri-

fied H2 gas, generally H2 was purified from the natural gas under atmo-
spheric pressure, then the purified H2 was compressed to the desired high
pressure. Although already in use at a H2 station, the separation and purifi-
cation process consumes a large amount of energy resulting in a high cost for
high-pressure H2, about 1000–1100 Yen kg1 in Japan.
One simple purification technique, that of separating H2 from the gener-
ated gas mixture (H2 and CO2), was reported to be via a gas–liquid phase sep-
aration method, simply by changing the physical state of the fluid (Fig. 18 and
Table 7). The high-pressure gas at >7.4 MPa, which is generated from FA as a
mixture of H2 and CO2 at 80°C, has a lower critical point than CO2 itself
(31.1°C, 7.4 MPa). Therefore, the generated gas is in the supercritical phase.
Thus, to purify H2 gas from the gas mixture, the gas separator was simply cooled
down to a temperature below the critical temperature in order to change the
generated gas from the supercritical state to the gas–liquid state without depres-
surization. When the generated gas entered the gas separator at 80°C and
30 MPa of pressure, the gas separator was set at 35°C, which is the supercritical
condition at 30 MPa. The equimolar mixture of H2 and CO2 gases was
obtained from the back-pressure regulator attached to the separator.
Fig. 18 View of the phase separation.
Table 7 Gas Contents of the Separated Gas Generated From the Decomposition of
FA at Various Temperatures of the Separator at 30 MPaa
Initial Gas Initial H2
Separator XH2 XCO Flow Rate Production
Entry Temp. (°C) (mol%) (mol%) (L h21)b Rate (h21)c
1 35 51 n.d.d 0.93 2560
2 0 58 n.d. 0.86 2620
3 15 69 n.d. 0.75 2790
4 40 80 n.d. 0.73 3030
5 51 85 n.d. 0.69 3050
a
Gas generation condition: 80°C, 30 MPa. Gas separation condition: 51°C to 35°C, 30 MPa. Aqueous
solution of FA: 8 mol L1, 40 mL, catalyst ([Cp*lr(4DHBP)(H2O)][HSO4]): 0.2 mmol L1, 7–8 μmol.
b
Average gas rate for initial 1 h.
c
Average rate of H2 gas per mole of the catalyst.
d
Not detected (less than 6 vol ppm).
8. APPLICATION FOR FUEL CELL BATTERIES

Formic acid is one of the promising choices as the hydrogen carrier for
fuel cell batteries that may be developed in the near future. There are pub-
lished reviews on formic acid as a hydrogen carrier, but applications as a fuel
cell using a homogeneous catalyst are limited. In order to establish an eco-
nomically feasible system for initial commercialization, a significant reduc-
tion of the cost of the catalyst is required. Furthermore the system must be
able to operate under mild conditions and without sacrificing the activity
and selectivity of the catalyst for hydrogen generation. In addition there
is a need for advances in the formic acid synthesis and CO2 capture opera-
tions. However, the identification of the main technological obstacles on the
route to efficient use of FA as a hydrogen source and storage material remains

to be established. A hydrogen generator based on the decomposition of FA
was designed and built in early 2009. An early report of its applications was
published by Boddien and Beller et al. The system combined the hydrogen
generation unit with a H2/O2 polymer electrolyte membrane fuel cell
(PEMFC) as shown in Fig. 19 (93). They developed the simple power gen-
eration (Fig. 20) through the dihydrogen production system, which involves
HCO2H/NEt3 and a Ru catalyst. The H2 and CO2 gas generated was con-
taminated by a small amount of the volatile organic amine (NEt3); charcoal
(CarboTex) was used as a gas absorber to remove the contaminant; other-
wise, it can deactivate the membrane electrode of the fuel cell. At the initial
stage, the cell power was 48 mW and then it decreased to 26 mW for 42 h,
and finally 14 mW was obtained after 69 h.
Grasemann and Laurenczy also developed a hydrogen generator based
on the decomposition of FA over a homogeneous Ru(II) catalyst
(Fig. 21) (94); the general concept had been published earlier. The hydrogen
generator successfully met the target power output of 1 kW or roughly
30 L min1 of H2/CO2 assuming 50% fuel cell efficiency.
The exact reason for the decreasing cell voltage is unknown, but there
are some possibilities for deactivation of the cell. One of the reasons could
be contamination by formic acid in H2 gas. Zhang et al. performed
long-term (100 h) tests to detect contaminants and showed that 100 ppm
H2/CO2
Air
humidification
(H2O)
Fuel cell
Cleaning unit
(charcoal)
Reaction vessel
Fig. 19 The hydrogen generation unit with a polymer electrolyte membrane fuel cell
(PEMFC).
50
40
30
P (mW)
20
10
0
4
6
17
19
21
23
25 t (h)
27
29
46
52
0 69
200 400 600 800 1000
U (mV)
Fig. 20 Power output as a function of time.
Fig. 21 Industrial prototype for 1 kW power output.
HCOOH in the H2 stream significantly degrades the electrode performance

and can significantly affect the performance of the PEMFC (95).
Recently Czaun et al. fed the mixture of H2 and CO2 obtained by
decomposition of FA, using an IrCl3/1,3-bis(2ʹ-pyridyl-imino)-isoindoline
(IndH) catalyst, into a hydrogen-air PEMFC (96). At first, they set the cell
voltage at a standard current (I ¼ 1.0 A) using ultrahigh-purity hydrogen,
then introduced H2/CO2 from FA into the cell. The cell voltage was stable
(0.85 V at I ¼ 1.0 A) for the time of the measurement, using either O2 or air
as cathode feed gas. The integrated FA decomposition showed no difference
in performance under the given experimental conditions compared with the
1.0 100 mL min−1 H2/CO2 (from FA decomposition),

240 mL min−1 Air
1.0
0.8
0.9
0.6
Voltage (V)
H2 (UHP, cylinder)/O2
0.8
H2 (UHP, cylinder)/air
0.4
0.7
0.2
0.6
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Time (h)
Fig. 22 Fuel cell performance comparison (H2/CO2 from formic acid and H2/air).
H2/O2 or H2/air fuel cell (Fig. 22). The durability of the system was assessed
in a longer duration experiment; the fuel cell maintained its voltage (0.85 V)
at a current value of 1.0 A, for the entire 14 h.
A group of students in Eindhoven “Team FAST” built successfully a
400 W model car that can carry 45 kg at approximately 8 km h1, and fur-
ther developed buses powered by formic acid. According to their website,
they will start to run a bus in the city for test purposes in 2017 (http://www.
teamfast.nl/).
9. CONCLUSION
In 2014, the Toyota Motor Cooperation started to sell the fuel cell
vehicle, Mirai, globally, followed by Honda selling the FCV, Clarity Fuel
Cell, in March, 2016. Even though FCVs are available commercially, the
overall program is still in its infancy considering the technologies as well
as the infrastructure to utilize FCVs; issues relating to the production, trans-
portation, storage, and feeding of hydrogen, especially at high-pressure, over
35 MPa up to 70 MPa, remain to be solved. As described in this chapter,
formic acid as a hydrogen storage material has immense potential that offers
many benefits to develop a sustainable society. We believe that formic acid
will be one of the promising hydrogen carriers for the next generation
throughout the world.
REFERENCES
1. Schlapbach, L. Nature 2009, 460, 3.
2. Dresselhaus, M. S.; Thomas, I. L. Nature 2001, 414, 6.
3. Nikolaidis, P.; Poullikkas, A. Renew. Sustain. Energy Rev. 2017, 67, 597–611.
4. Marbán, G.; Valdes-Solı́s, T. Int. J. Hydrogen Energy 2007, 32, 1625–1637.
5. Schlapbach, L.; Z€ uttel, A. Nature 2001, 414, 6.
6. European Hydrogen & Fuel Cell Technology Platform, Development Strategy, August
2005.
7. Barrett, S. Fuel Cells Bull. 2005, 2005, 12–19.
8. Office of Energy Efficiency and Renewable Energy, U.S. Department Energy, Hydro-
gen and Fuel Cells Program Plan, September 2011.
9. Kothari, R.; Buddhi, D.; Sawhney, R. L. Renew. Sustain. Energy Rev. 2008, 12,
553–563.
10. Armaroli, N.; Balzani, V. ChemSusChem 2011, 4, 21–36.
11. Teichmann, D.; Arlt, W.; Wasserscheid, P.; Freymann, R. Energ. Environ. Sci. 2011, 4,
2767–2773.
12. Scherer, G. W. H.; Newson, E. Int. J. Hydrogen Energy 1998, 23, 19–25.
13. Zhao, H. Y.; Oyama, S. T.; Naeemi, E. D. Catal. Today 2010, 149, 172–184.
14. Giacomazzi, G.; Gretz, J. Cryogenics 1993, 33, 767–771.
15. Eberle, U.; Felderhoff, M.; Sch€ uth, F. Angew. Chem. Int. Ed. 2009, 48, 6608–6630.
16. Dalebrook, A. F.; Gan, W.; Grasemann, M.; Moret, S.; Laurenczy, G. Chem. Commun.
2013, 49, 8735–8751.
17. Biniwale, R.; Rayalu, S.; Devotta, S.; Ichikawa, M. Int. J. Hydrogen Energy 2008, 33,
360–365.
18. Pradhan, A. U.; Shukla, A.; Pande, J. V.; Karmarkar, S.; Biniwale, R. B. Int. J. Hydrogen
Energy 2011, 36, 680–688.
19. Shukla, A.; Karmakar, S.; Biniwale, R. B. Int. J. Hydrogen Energy 2012, 37, 3719–3726.
20. Alhumaidan, F.; Cresswell, D.; Garforth, A. Energy Fuel 2011, 25, 4217–4234.
21. Zhu, Q.-L.; Xu, Q. Energ. Environ. Sci. 2015, 8, 478–512.
22. Makowski, P.; Thomas, A.; Kuhn, P.; Goettmann, F. Energ. Environ. Sci. 2009, 2,
480–490.
23. Chaouki, J.; Klavana, D. Chem. Eng. Sci. 1994, 49, 4639–4646.
24. Shukla, A. A.; Gosavi, P. V.; Pande, J. V.; Kumar, V. P.; Chary, K. V. R.;
Biniwale, R. B. Int. J. Hydrogen Energy 2010, 35, 4020–4026.
25. Okada, Y.; Sasaki, E.; Watanabe, E.; Hyodo, S.; Nishijima, H. Int. J. Hydrogen Energy
2006, 31, 1348–1356.
26. Kariya, N.; Fukuoka, A.; Utagawa, T.; Sakuramoto, M.; Goto, Y.; Ichikawa, M. Appl.
Catal. Gen. 2003, 247, 247–259.
27. Hodoshima, S. Int. J. Hydrogen Energy 2003, 28, 1255–1262.
28. Kustov, L. M.; Tarasov, A. L.; Tarasov, B. P. Int. J. Hydrogen Energy 2013, 38,
5713–5716.
29. Lázaro, M. P.; Garcı́a-Bordeje, E.; Sebastián, D.; Lázaro, M. J.; Moliner, R. Catal. Today
2008, 138, 203–209.
30. Preuster, P.; Papp, C.; Wasserscheid, P. Acc. Chem. Res. 2017, 50, 74–85.
31. Liu, D.; Men, Y.; Wang, J.; Kolb, G.; Liu, X.; Wang, Y.; Sun, Q. Int. J. Hydrogen Energy
2016, 41, 21990–21999.
32. Sá, S.; Silva, H.; Brandão, L.; Sousa, J. M.; Mendes, A. Appl. Catal. Environ. 2010, 99,
43–57.
33. Palo, D. R.; Dagle, R. A.; Holladay, J. D. Chem. Rev. 2007, 107, 3992–4021.
34. Dobscn, A.; Robinson, S. D. J. Organomet. Chem. 1975, 87, C52–C53.
35. Dobson, A.; Robinson, S. D. Inorg. Chem. 1977, 16, 137–142.
36. Rodrı́guez-Lugo, R. E.; Trincado, M.; Vogt, M.; Tewes, F.; Santiso-Quinones, G.;
Gr€utzmacher, H. Nat. Chem. 2013, 5, 342–347.
37. Stephan, D. W. Nature 2013, 495, 54–55.
38. Verendel, J. J.; Diner, P. ChemCatChem 2013, 5, 2795–2797.
39. Yang, X. ACS Catal. 2014, 4, 1129–1133.
40. Nielsen, M.; Alberico, E.; Baumann, W.; Drexler, H. J.; Junge, H.; Gladiali, S.;
Beller, M. Nature 2013, 495, 85–89.
41. Alberico, E.; Sponholz, P.; Cordes, C.; Nielsen, M.; Drexler, H. J.; Baumann, W.;
Junge, H.; Beller, M. Angew. Chem. Int. Ed. Engl. 2013, 52, 14162–14166.
42. Anderez-Fernandez, M.; Vogt, L. K.; Fischer, S.; Zhou, W.; Jiao, H.; Garbe, M.;
Elangovan, S.; Junge, K.; Junge, H.; Ludwig, R.; Beller, M. Angew. Chem. Int. Ed.
2017, 56, 559–562.
43. Alberico, E.; Lennox, A. J.; Vogt, L. K.; Jiao, H.; Baumann, W.; Drexler, H. J.;
Nielsen, M.; Spannenberg, A.; Checinski, M. P.; Junge, H.; Beller, M. J. Am. Chem.
Soc. 2016, 138, 14890–14904.
44. Gibson, H. W. Chem. Rev. 1969, 69, 673–692.
45. Hietala, J.; Vuori, A.; Johnsson, P.; Pollari, I.; Reutemann, W.; Kieczka, H. Formic Acid.
Wiley-VCH Verlag GmbH & Co: KGaA, Weinheim, 2016.
46. Wagman, D. D.; Evans, W. H.; Parker, V. B.; Schumm, R. H.; Halow, I.; Bailey, S. M.;
Churney, K. L.; Nuttall, R. L. J. Phys. Chem. Ref. Data Monogr. 1982, 11, 2–84.
47. Scott, D. W.; Guthrie, G. B.; Messerly, J. F.; Todd, S. S.; Berg, W. T.; Hossenlopp, I. A.;
McCullough, J. P. J. Phys. Chem. 1962, 66, 911–914.
48. Roux, M. V.; Temprado, M.; Chickos, J. S.; Nagano, Y. J. Phys. Chem. Ref. Data
Monogr. 2008, 37, 1855–1996.
49. Prosen, E. J.; Johnson, W. H.; Rossini, F. D. J. Res. Natl. Bur. Stand. 1946, 37, 51–56.
50. Douslin, D. R.; Huffman, H. M. J. Am. Ceram. Soc. 1946, 68, 173–176.
51. Trimm, D. L. Appl. Catal. A. 2005, 296, 1–11.
52. Sordakis, K.; Beller, M.; Laurenczy, G. ChemCatChem 2014, 6, 96–99.
53. Fellay, C.; Yan, N.; Dyson, P. J.; Laurenczy, G. Chem. A Eur. J. 2009, 15, 3752–3760.
54. Hull, J. F.; Himeda, Y.; Wang, W.-H.; Hashiguchi, B.; Periana, R.; Szalda, D. J.;
Muckerman, J. T.; Fujita, E. Nat. Chem. 2012, 4, 383–388.
55. Inoue, Y.; Izumida, H.; Sasaki, Y.; Hashimoto, H. Chem. Lett. 1976, 1976, 863–864.
56. Chen, H.-T.; Chang, J.-G.; Chen, H.-L. J. Phys. Chem. A 2008, 112, 8093–8099.
57. Fellay, C.; Dyson, P. J.; Laurenczy, G. Angew. Chem. Int. Ed. 2008, 47, 3966–3968.
58. Fukuzumi, S.; Kobayashi, T.; Suenobu, T. ChemSusChem 2008, 1, 827–834.
59. Inaba, S. J. Phys. Chem. A 2014, 118, 3026–3038.
60. Loges, B.; Boddien, A.; Junge, H.; Beller, M. Angew. Chem. Int. Ed. Engl. 2008, 47,
3962–3965.
61. Ojeda, M.; Iglesia, E. Angew. Chem. Int. Ed. Engl. 2009, 48, 4800–4803.
62. Boddien, A.; Mellmann, D.; G€artner, F.; Jackstell, R.; Junge, H.; Dyson, P. J.;
Laurenczy, G.; Ludwig, R.; Beller, M. Science 2011, 333, 1733–1736.
63. Tedsree, K.; Li, T.; Jones, S.; Chan, C. W. A.; Yu, K. M. K.; Bagot, P. A. J.;
Marquis, E. A.; Smith, G. D. W.; Tsang, S. C. E. Nat. Nanotechnol. 2011, 6, 302–307.
64. Bi, Q. Y.; Du, X. L.; Liu, Y. M.; Cao, Y.; He, H. Y.; Fan, K. N. J. Am. Chem. Soc. 2012,
134, 8926–8933.
65. Moret, S.; Dyson, P. J.; Laurenczy, G. Nat. Commun. 2014, 5, http://dx.doi.org/
10.1038/ncomms5017.
66. Coffey, R. S. Chem. Commun. 1967, 923–924.
67. Yoshida, T.; Ueda, Y.; Otsuka, S. J. Am. Chem. Soc. 1978, 100, 3941–3942.
68. Strauss, S. H.; Whitmire, K. H.; Shriver, D. F. J. Organomet. Chem. 1979, 174,
C59–C62.
69. Paonessa, R. S.; Trogler, W. C. J. Am. Chem. Soc. 1982, 104, 3529–3530.
70. King, R. B.; Bhattacharyya, N. K. Inorg. Chim. Acta 1995, 237, 65–69.
71. Gao, Y.; Kuncheria, J.; Yapb, G. P. A.; Puddephatt, R. J. Chem. Commun. 1998,
2365–2366.
72. Gao, Y.; Kuncheria, J. K.; Jenkins, H. A.; Puddephatt, R. J.; Yap, G. P. A. J. Chem. Soc.
Dalton Trans. 2000, 3212–3217.
73. Forster, D.; Deck, G. J. Chem. Soc. D 1971, 18, 1072.
74. Boddien, A.; Loges, B.; Gartner, F.; Torborg, C.; Fumino, K.; Junge, H.; Ludwig, R.;
Beller, M. J. Am. Chem. Soc. 2010, 132, 8924–8934.
75. Boddien, A.; Gartner, F.; Jackstell, R.; Junge, H.; Spannenberg, A.; Baumann, W.;
Ludwig, R.; Beller, M. Angew. Chem. Int. Ed. Engl. 2010, 49, 8993–8996.
76. Montandon-Clerc, M.; Dalebrook, A. F.; Laurenczy, G. J. Catal. 2016, 343, 62–67.
77. Zell, T.; Butschke, B.; Ben-David, Y.; Milstein, D. Chemistry 2013, 19, 8068–8072.
78. Bielinski, E. A.; Lagaditis, P. O.; Zhang, Y.; Mercado, B. Q.; W€ urtele, C.;
Bernskoetter, W. H.; Hazari, N.; Schneider, S. J. Am. Chem. Soc. 2014, 136,
10234–10237.
79. Himeda, Y. Green Chem. 2009, 11, 2018–2022.
80. Deng, J.; Wang, Y.; Pan, T.; Xu, Q.; Guo, Q. X.; Fu, Y. ChemSusChem 2013, 6,
1163–1167.
81. Tomas-Mendivil, E.; Dıez, J.; Cadierno, V. Cat. Sci. Technol. 2011, 1, 1605.
82. Klein, S.; Dougherty, W. G.; Kassel, W. S.; Dudley, T. J.; Paul, J. J. Inorg. Chem. 2011,
50, 2754–2763.
83. Himeda, Y.; Onozawa-Komatsuzaki, N.; Miyazawa, S.; Sugihara, H.; Hirose, T.;
Kasuga, K. Chem. A Eur. J. 2008, 14, 11076–11081.
84. Himeda, Y.; Miyazawa, S.; Hirose, T. ChemSusChem 2011, 4, 487–493.
85. Himeda, Y.; Onozawa-Komatsuzaki, N.; Sugihara, H.; Kasuga, K. J. Am. Chem. Soc.
2005, 127, 13118–13119.
86. Iguchi, M.; Himeda, Y.; Manaka, Y.; Kawanami, H. ChemSusChem 2016, 9,
2749–2753.
87. Suna, Y.; Ertem, M. Z.; Wang, W.-H.; Kambayashi, H.; Manaka, Y.; Muckerman, J. T.;
Fujita, E.; Himeda, Y. Organometallics 2014, 33, 6519–6530.
88. Wang, W.-H.; Xu, S.; Manaka, Y.; Suna, Y.; Kambayashi, H.; Muckerman, J. T.;
Fujita, E.; Himeda, Y. ChemSusChem 2014, 7, 1976–1983.
89. Wang, W.-H.; Hull, J. F.; Muckerman, J. T.; Fujita, E.; Hirose, T.; Himeda, Y. Chem.
A Eur. J. 2012, 18, 9397–9404.
90. Wang, W.-H.; Hull, J. F.; Muckerman, J. T.; Fujita, E.; Himeda, Y. Energ. Environ. Sci.
2012, 5, 7923–7926.
91. Iguchi, M.; Himeda, Y.; Manaka, Y.; Matsuoka, K.; Kawanami, H. ChemCatChem
2016, 8, 886–890.
92. Guan, C.; Zhang, D. D.; Pan, Y.; Iguchi, M.; Ajitha, M. J.; Hu, J.; Li, H.; Yao, C.;
Huang, M. H.; Min, S.; Zheng, J.; Himeda, Y.; Kawanami, H.; Huang, K. W. Inorg.
Chem. 2017, 56, 438–445.
93. Boddien, A.; Loges, B.; Junge, H.; Beller, M. ChemSusChem 2008, 1, 751–758.
94. Grasemann, M.; Laurenczy, G. Energ. Environ. Sci. 2012, 5, 8171–8181.
95. Zhang, X.; Galindo, H. M.; Garces, H. F.; Baker, P.; Wang, X.; Pasaogullari, U.;
Suib, S. L.; Molter, T. J. Electrochem. Soc. 2010, 157, B409.
96. Czaun, M.; Kothandaraman, J.; Goeppert, A.; Yang, B.; Greenberg, S.; May, R. B.;
Olah, G. A.; Prakash, G. K. S. ACS Catal. 2016, 6, 7475–7484.
INDEX
Note: Page numbers followed by “f ” indicate figures, “t” indicate tables, and “s” indicate
schemes.
A (2-[6-(40 -Amino)phenoxy-3H-xanthen-
π-Acceptor effect 3-on-9-yl]benzoic acid) (APF),
ancillary substituent effect, 262–264 374–378, 376f
cationic square-planar Pt(II) complexes, Ancillary ligands, 335–336
246f Antitumor immune response, 371–372
and donor atom effect, 252–259 Apoptosis, 366–367
N-donor chelates, 269–272 Aromatic C–F bonds, 153–158
vs. σ-donor effect, 246–251 Arylplatinum compounds, 197
extended ring conjugation, 259–261 Ascorbate, 361–362
Pt(II) complexes, 254f, 255–256 ASM. See Activation strain model (ASM)
second-order rate constants, 248–249t Autophagy, 369
in square-planar Pt(II) complexes, Avantes spectrophotometer, 50
245–246
substituent effect, 261–262
terpy and analogous chelates, 253f B
tridentate nitrogen-donor ligands, 250f Bacteriochlorins, in cells, 362–366
Activation strain model (ASM) Bcis-type intermediates
adc and dmad reactions, 328–330, 329t C-H activation, 232t
analysis, 328–332, 329f chelating N(amino)-N(imine) ligands,
of chemical reactivity, 320–322 233–239
Acylperoxoiron(III) compounds, 87–97 monodentate N(imine) ligands, 228–233
Adamantane, oxidation, 129f, 142 stability, 225–227
Adsorption kinetics, 53–54 Benzene, oxidation, 136–139, 144–145,
Advanced oxidation processes (AOPs), 3–4 144f
Aldehyde deformylation, 185–190, 186f Bioinorganic chemistry, 68
Alkanes Bioinspired catalysis, 108–109
C–H bonds of, 91, 145 Bioinspired model systems, 65–67
single-turnover oxidation of, 129t Biomimetic manganese(V)-oxo complexes,
AlkB, 169–170 181–184
Alkylaromatic compounds, 147–149 Biomimetic models, 172–175,
Alkyne 177–178
in acetonitrile, 336t Bis(iminopyridyl/isoquinolyl)isoindoline,
activation strain analysis, 328–332, 329f 264–269
ASM analysis, 329t Bis(pyrazolyl)pyridine/benzene,
effect, 324–332 incorporating, 264–269
free energies, activation and reaction, Bleomycin, 174
328t Bond dissociation energies
HOMO and LUMO orbitals, 330–331, (BDE), 286
330f Btrans-type intermediates, 225–227
structural changes, 327t Buthionine sulfoximine (BSO),
transition state, 324–328, 325f 378–379
429
430 Index
C of propene activation, 171–172

Carbon dioxide (CO2) spin selective reactivity, 183–184
formic acid dehydrogenation, using density functional theory, 176
403–404 Copasi software, 221–225
hydrogenation reaction of, 402 Copper-dioxygen chemistry, 68–69
volumetric vs. gravimetric hydrogen CpdI, 171–173
densities, 404f Cyclic aliphatic hydrocarbons, 398
Carbon–fluorine bonds, 72–78 [3+2] Cycloaddition reaction
Catalytic oxidation alkyne effect, 324–332
adamantane, 142 ancillary ligands effect at metal centers,
alkylaromatic compounds, 147–149 335–336
aromatic C–F bonds, 153–158 computed free energy profile, 338f
benzene, 136–139, 144–145 kinetic rate constants, 323t
C–C bonds formation, 149–152 kinetics of reaction, 322–323
cyclohexane, 139–142 metal effect, 332–335
cyclohexene, 142–143 solvent effect, 336–339
dehalogenation, 152–153 Cyclohexane, 139–142
diiron and iron oxo species, 128–130 Cyclohexene, 142–143
ethane, 134–136 Cyclometallated compounds, 197
hydrocarbons, 145–147 Cyclopentadienyl (Cp*) iridium
methane, 130–133 with BPY ligands, 412f
recalcitrant pollutants degradation, with 4,7-dihydroxy-1,10-phenanthroline
158–160 ligand, 411, 411f
toluene, 143 formic acid dehydrogenation catalyzed
C–C bonds formation, 149–152 by, 413f
Chalcogenide-centered reactivity, HCO2H catalyzed by, 412t, 415t
316–320 for H2 generation, 408–416
Chelating N(amino)-N(imine) ligands, with THBP, TH4BPM, and THBPM
233–239, 238f ligands, 415f
Chlorate ion, 6, 44–46 water-soluble iridium catalyst, 418–419
Chlorinated bacteriochlorin Cycloplatinated compounds, 197
(ClBOH), 378 Cysteine catabolism, 174
Chlorine dioxide Cysteine dioxygenase (CDO), 171
alkaline decomposition, 7–8, 7f, 8s catalytic cycle of, 175f
formation of, 13f iron-dioxygen complex in, 174–177
oxidation cysteine by, 10s M€ ossbauer spectroscopy, 176
oxidation reactions, 9 Cytochrome c oxidase, 347
water treatment technologies, 6–7 Cytochrome P450, 108–109, 171
Chlorite CpdI, 172–173
alkaline decomposition, 14–15 O–O cleavage in, 79–80
catalytic decomposition, 16
redox reactions, 11–12 D
N-Chloroamino acids, 21–22 Defluorination, 154–159, 155–156f
N-Chloroglycine, 19f Dehalogenation, 152–153
Colloidal sulfur, 287 Density functional theory (DFT), 176
Complete active space-self consistent field activation and reaction free
(CASSCF), 175–177, 177f energies, 320
Computational modeling bifurcation pathways, 186f
iron(iv)-oxo intermediate, 170 calculations, 221–239
Index 431
computation methodology, 196 Diode array spectrophotometer, 36s, 38

for HCl reaction, 315 1,3-Dipolar cycloaddition reactions,
metal effect, 332–333 336–337
substitution reactions, 313–314 Donor atom effect, 252–259
Department of Energy (United States), σ-Donor effect, 246–251
396 Drug-to-light interval (DLI),
Deuterium kinetic isotope effect (KIE), 348, 348f
413–414, 414t
DFT. See Density functional theory (DFT) E
Diarylplatinum(II) scaffolds EDA method. See Energy decomposition
cis vs. trans stability, B-type intermediates, analysis (EDA) method
225–227 Electron-donating properties, 171–172,
compounds, 198–205 181–182
DFT calculations, 221–239 Electronic absorption, 355–356
kinetic studies, 206–221 Electron paramagnetic resonance (EPR)
platinacycle stability, 225 spectroscopy
single-point calculations, 221–225 in acylperoxoiron(III) species, 89–90
spontaneous processes chemical structure, 374
with chelate N(amino)-N(imine) ligands, iron oxidation state determination,
215–221 116–119
concentration-tuned reactions, 210s in water soluble bacteriochlorin, 374
cyclometallated PtIV complexes, 216s Electron spin resonance spin
modification of reactivity, 214s trapping, 374
with monodentate N(imine) ligands, Electrospray ionization mass spectrometry
209–215 (ESI-MS), 116
rate-determining step, 215f Emission spectroscopy, 119–120
relevant kinetic and activation Endoplasmic reticulum (ER), 351
parameters, 212t Energy decomposition analysis (EDA)
UV-vis monitoring, 219f method, 322
2,6-Dichloro-1,4-benzoquinone (DCQ), EPR spectroscopy. See Electron
36–37 paramagnetic resonance (EPR)
photoreduction, 37–38, 37s spectroscopy
3,5-Dichloro-2-hydroxy-1,4- Ergothioneine synthase, 171
benzoquinone (DCHB), Ethane, 134–136
37–38 Ethylbenzene, 129f
Dicopper model complexes, 68–69 Extended X-ray absorption fine
Diels–Alder reactions, 336–337 structure (EXAFS) method,
Dihydroethidium (DHE), 377 115–116, 116f
Dihydrogen, 314–315
4,7-Dihydroxy-1,10-phenanthroline F
(DHPH), 411, 411f Fenton reaction, 22, 28
Diiron macrocyclic porphyrin-like FeOOR species, O–O cleavage in,
complexes, 109 84–97
Diiron macrocyclic scaffold, 110 (FePctBu4)2N–tBuOOH catalytic system,
Dinuclear copper systems, O–O cleavage in, 139–147
67–78 Fluorescence, 355–356
C–F bonds, hydroxylation, 72–78 Fluorinated aromatic compounds, 152–155,
dicopper model complexes, 68–69 158f
tyrosinase-like activity, 69–72 Formaldehyde, 21–22
432 Index
Formic acid High-valent diiron-oxo species

Cp* iridium complex for H2 generation, CID-MS/MS spectrum, 128f
408–416 ESI-MS spectrum, 122–123f, 123
decomposition, 406, 417f formation, 127f
dehydration, 409f low signal-to-noise ratio, 125
dehydrogenation, 403–404, 406 macrocyclic structure on formation,
evolved gas volume, 408f 127–128
fuel cell batteries, application for, M€ ossbauer spectroscopy, 126f
421–424 oxidation, 128–130
gas contents of separated gas, 421t phthalocyanine platform, 122–124
gas evolution, time course, 416f porphyrin platform, 124–126
homogeneous catalytic dehydrogenation, Histidine, 171–172
1
404–407, 407t H NMR spectroscopy, MCG
for LOHC, 401–404 decomposition, 19, 19–20f
pH condition for dehydrogenation, 416f Homogeneous catalytic dehydrogenation,
physical properties, 402–403 404–407
pressure, time course, 418f HP-8453 diode array spectrophotometer,
sustainable hydrogen generation and 39f, 40
storage, 406f H3PO2/H3PO3, 316
thermodynamic properties of Hydrocarbon oxidation, 83–84, 145–147
dehydrogenation, 405t Hydrodisulfides (H2S), 286, 292
time-dependent gas evolution, 420f crosstalk, 278–280
volumetric vs. gravimetric hydrogen Gmelin process, 278–280
densities, 404f RSNOs transnitrosation, 292–295
N-Formylglycine, 21 signaling cascade, 278
Franck-Condon energy, 188–189 Hydrogen gas (H2), 396
Fuel cell batteries cyclopentadienyl iridium for, 408–416
application for, 421–424 from formic acid, 413s
performance comparison, 424f high-pressure H2 generation, 417–420
power output Hydrogen peroxide (H2O2), 22, 34, 110
as function of time, 423f Hydroperoxoiron(III) compounds, 85–87
industrial prototype for, 423f Hydroxy form, acid–base equilibrium, 410s
Hydroxylation
G agent, 69–72
Gasotransmitters, 278 carbon–fluorine bonds, 72–78
Glutathione (GSH), 9–10 Hydroxyl radical, 22, 344–347, 359, 378
Gmelin reaction, 299–306 Hypochlorous acid (HOCl), 16–18
Hypohalous acid, 26
H
Halide ions, 26 I
Hammett type correlation, 409–410, 410f Inorganic reactions, 172–173
Heme enzyme, 171–172 Inorganic salts, 383–384
Heme iron(III)-hydroperoxo, 177–181 Interaction energy, 320–322
Heme monooxygenases, 171 Iodide ion concentration, 44f
Heterogeneous systems, 49–55 Iron-dioxygen complex, 174–177
Highest occupied molecular orbital Iron(III)-hydroperoxo, 174, 177–181
(HOMO), 171–172 Iron oxidation state determination,
High-pressure H2 generation, 417–420 116–119
Index 433
Iron(IV)-oxo intermediate, 170, 173–174 Metal-centered reactivity

Iron-oxygen species dihydrogen release, 314–315
O–O cleavage in H3PO2 and H3PO3, 316
acylperoxoiron(III) compounds, 87–97 substitution reactions, 313–314
cytochrome P450, 79–80 transition state structures, 316f
FeOOR species, 84–97 Metal-(di)oxygen intermediates,
hydroperoxoiron(III) compounds, 173–190
85–87 Metal-heme-dependent enzymes, 171
iron-containing enzymes, 79–83 Metalloenzymes, 168
nonheme iron catalysts, 83–84 Metalloproteins, 65–66
Rieske oxygenases, 81–83 Methane monooxygenase (MMO),
spectroscopically characterized 108–109
oxoiron(V) species, 97–99 Methane, oxidation, 130–133, 131f
oxidation, 128–130 Methanol, 400
Iyellow identification, 288–295 L-Methionine, 13
Methylene blue (MB), 52–53, 54s
J Modular Avantes photoreactor, 48f, 48s
Jablonski diagram, 355–357, 355f Monodentate N(imine) ligands, 228–233
Mono-N-oxidation, 31–33
K Mononuclear nonheme iron-oxygen
α-Ketoglutarate, 168–169, 169f species, 84–97
Kinetic isotope effect (KIE), 413–414, 414t M€ossbauer spectroscopy
Kinetico-mechanistic studies, 201–202, 211, CDO, 176
240 iron oxidation state determination,
116–119
L M3S4 clusters
Langmuir–Hinshelwood model, 51–52 alkynes reaction with, 318f, 319t, 321f
Lipophilicity, 363 [3+1] building-block synthesis, 317f
Liquid organic hydrogen carriers (LOHC), chalcogenide-centered reactivity,
397–398 316–320
dehydrogenation, 400–401 cuboidal clusters, 321f
formic acid for, 401–404 [3+2] cycloaddition reaction
fuel cell batteries, application for, alkyne effect, 324–332
421–424 ancillary ligands effect at metal centers,
high-pressure H2 generation, 417–420 335–336
hydrogen storage properties, 399t computed free energy profile, 338f
organic materials for, 398–401 kinetics of reaction, 322–323
Low-spin acylperoxoiron(III) species, metal effect, 332–335
91–97 rate constants, 323t
solvent effect, 336–339
M experimental and computational
Manganese-peroxo (MnOO), 172–173 methods, 320–322
Manganese(III)-peroxo complexes, metal-centered reactivity
185–190, 186f dihydrogen release, 314–315
Manganese porphyrin, 16 H3PO2 and H3PO3, 316
Marcus–Hush theory, 2 substitution reactions, 313–314
Marshall’s acid, 22–23 osma-2,5-dioxolane, 319f
MCG decomposition, 18, 19–20f, 21s reactivity of, 312–320
434 Index
N O
Nanoformulation, of photosensitizers, O2 activation, 64–65
379–382 copper- and iron-based proteins in,
Necrosis, 367–369 64–65
Nitric oxide (NO) dicopper model complexes, 68–69
crosstalk, 278–280 by metalloenzymes, 65s
signaling cascade, 278 Olefins, hydroacylation of, 151f, 151t
μ-Nitrido diiron macrocyclic complexes O–O cleavage
emission spectroscopies, 119–120 in dinuclear copper systems, 67–78
EXAFS method, 115–116 C–F bonds, hydroxylation, 72–78
Fe–Fe distances, 112–115 dicopper model complexes, 68–69
iron oxidation state determination, tyrosinase-like activity, 69–72
116–119 in iron-oxygen species
mass spectrometry, 116 acylperoxoiron(III) compounds,
oxidation properties, 121–122 87–97
oxo species, mononuclear and binuclear cytochrome P450, 79–80
platforms, 121f FeOOR species, 84–97
preparation and spectroscopic hydroperoxoiron(III) compounds,
characterization, 111–122 85–87
structural parameters, 114t iron-containing enzymes, 79–83
structures, 111f nonheme iron catalysts, 83–84
X-ray absorption, 119–120 Rieske oxygenases, 81–83
X-ray diffraction, 112–115 spectroscopically characterized
Nitrogen-/carbon-donor tridentate ligands oxoiron(V) species, 97–99
(N^C/N^N/C) Oxidative addition, 197
π-acceptor vs. σ-donor effect, concentration-tuned reactions, 210s
246–251 cyclometallated PtIV complexes, 216s
chloride substitution from deprotonated relevant kinetic and activation parameters,
phenyl ring, 257–258t 212t
mercurial structural property, Oxidative stress, in cells, 347f
245–246 Oxoiron(V) species, 97–99
Pt(II) complexes, pull-and-push effects, Oxone, 23
252–272 Oxyanion, 410s
S-Nitrosothiols (RSNOs) Oxychlorine species, 6–22
chemical structures, 279f
coordination chemistry, P
295–298 PDT. See Photodynamic therapy (PDT)
resonance structures of, 280f Peroxo compounds, 22–35
structure and reactivity, 280–282 Peroxodisulfate salts (PDS), 22–23
transnitrosation, 292–295 Peroxomonosulfate ion (PMS), 23–24
Nitroxyl (HNO), 278, 286, pathways, 25–26
289, 294 peroxo bond, 25
Nonheme complex, iron(III)-hydroperoxo, phen oxidation by, 31–32, 32s
177–181 redox reactions, 25
Nonheme iron catalysts, in hydrocarbon second-order reaction, 28
oxidation, 83–84 spontaneous decomposition, 25
Nucleophilic attack, on carbonyl group, two-electron oxidant, 26
185–186, 189–190 Peroxomonosulfate ion radical, 38–39
Index 435
Perthionitrite (S2NO–) Q
absorption spectra calculations, 289–291 Quantitative UV–vis spectrophotometry, 50
available results, 288–289
calculated spectra, 290f
coordination chemistry, 298 R
mechanistic generation routes, 292–295 Reactive oxygen species (ROS), 344–345
standard deviation for, 291t antitumor immune response, 371–372
X-ray structural data, 291–292 apoptosis, 366–367
Pheophorbide A, 380–382 ascorbate in, 361–362
Photocatalysis, 49–51 autophagy, 369
Photochemical mechanism, 355f, 378–379 biological mechanisms, 366–372
Photodynamic therapy (PDT), 348–349, cellular targets for, 366f
348f damaging aspect, 345–346
cellular approach, 351 detection in PDT, 372–378
photosensitizers for, 351–354 fluorescence, 356
ROS detection generation in PDT, 378–384
fluorescent probes, 374–378 light interactions with human tissues, 349f
1270 nm phosphorescence, molecular oxygen reduction, 344f
372–373 necrosis, 367–369
radical species, 374 penetration depth, 349–351
ROS generation photodynamic therapy, 348–349
antioxidant enzymes inhibition, photosensitizers for PDT, 351–354
378–379 phthalocyanine- and bacteriochlorin-
inorganic salts addition, 383–384 based compounds, 353f
photosensitizers nanoformulation, porphyrins, subcellular localization,
379–382 362–366
singlet oxygen-generated possible deactivation pathways, 355–357
nanomaterials, 382 redox homeostasis, 346–347
TiO2 with tetrapyrroles, 382–383 stereotypic view, 345
Photofrin®, 351, 363–365 triplet excited state, 356–357
Photoinitiated reactions, 45–46f, 46–47 Type I mechanism, 358–360
Photon, as reactant, 35–49 Type II mechanism, 357–358
Photosensitization, of titanium dioxide, vascular occlusion, 369–371
382–383 Recalcitrant pollutants, 158–160
Photosensitizer (Ps), 348f, 365–366, Redox homeostasis, in cells, 346–347, 347f
379–382 Redox reactions
Phthalocyanine, 122–124, 352–354 in aqueous solution, 2
Platinacycle, five/seven-membered chlorate–iodine reaction, 49f
from Bcis-type intermediates, 228–239 dye-covered aerogel particles, 53–54
stability, 225 general scheme of, 3s
Platinum chemistry, 197–198 heterogeneous systems, 49–55
Platinum(II) complexes, 252–272 kinetic behavior complexity, 3
Polymer electrolyte membrane fuel cell long-chain approach, 42–43
(PEMFC), 421–422, 422f oxychlorine species, 6–22
Polysulfides, 287 ozone decomposition, 3–5
Porphyrin, 124–126, 352–354, 362–366 peroxo compounds, 22–35
Potassium peroxomonosulfate, 23 photon as reactant, 35–49
Potential energy surface (PES), 320–322 stoichiometry, 4–5, 15
436 Index
Reductive elimination, 197 T

concentration-tuned reactions, 210s Taurine dioxygenase (TauD), 168–169
cyclometallated PtIV complexes, 216s active site structure, 169f
relevant kinetic and activation parameters, catalytic cycle, 169–170, 170f
212t QM/MM, 169–170
Renewable energy systems, 396 Tetraamido ligands (TAML), 97–98
Resonant inelastic X-ray scattering (RIXS), Tetrapyrroles, 382–383
120 Tetrathionate, 14–15
Rieske oxygenases, 81–83 Thioanisole sulfoxidation, 183–184, 184f
Thionitrite (SNO–)
S aqueous intermediates, 282–286
Salicylic acid (SA), photodegradation, coordination chemistry, 298
52s, 53f Thionitrous acid (HSNO)
Silica aerogel, 52–53 aqueous intermediates, 282–286
Silica–titania aerogel systems, 50 aqueous reactivity, 285–286
Singlet oxygen, 344–345 coordination chemistry, 298
APF reacts with, 378 second reactivity mode, 286
generation, 356 transnitrosation reaction, 285f
irreversibly, 358 uncharged character, 284–285
1270 nm phosphorescence spectra, 373f in water, 283, 284t
and porphyrins, 352 Time dependent-density functional theory
quantum yields, 372, 379–380 (TD-DFT), 175–177, 177f, 313–314
tissue necrosis by, 350 Titanium dioxide (TiO2), 382–383
Type II mechanism, 357 Toluene, oxidation, 143
Singlet oxygen sensor green Transition states (TSs), 324–328, 325f
(SOSG), 377 Transnitrosation, of RSNOs, 292–295
sMMO. See Soluble MMO (sMMO) 2,4,6-Trichlorophenol (TCP)
Sodium formate (SF), 409 Fe(TPPS)+ catalyzed oxidation, 35s
Soluble MMO (sMMO), 108–109 oxidation, 34, 36–37
binuclear macrocyclic concept, 109, 109f Triplet excited state, in ROS, 356–357
high-valent iron-oxo species, 110, 110–111f Tryptophan, oxidation, 10–11
Soret band, 352–354 Turnover frequency (TOF), 401
Spectroscopically characterized Tyrosinase, 67–78, 68s
oxoiron(V) species, 97–99
Spin-trapping techniques, 281, 374
U
Square-planar Pt(II) complexes, 264–269
UV-Vis absorption, 175–176
Strategic Energy Plan for Japan (Japan), 397
Strategic Research Agenda (Europe), 396
Sulfite ion radical, 38–39 V
Sulfur(IV) Vascular occlusion, 369–371
iodide-catalyzed autoxidation, 45f Verteporfin, 363–365
photoinitiated autoxidation of, 44t, 45f V-shaped Hammett plot, 183–184
Sulfur sols, 287
Superoxide ion, 344–345 X
dismutation by SOD, 361 XANES spectroscopy, 116–119
redox homeostasis, 346–347 X-ray absorption, 119–120
singlet oxygen and, 361–362, 373f X-ray diffraction, 112–115
Type I mechanism, 358 X-ray structural data, 291–292

Advances in Inorganic Chemistry - Rudi Van Eldik, Colin D. Hubbard, 2017

Hochgeladen von

Dokumentinformationen

Copyright

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Advances in Inorganic Chemistry - Rudi Van Eldik, Colin D. Hubbard, 2017

Hochgeladen von

Copyright:

ADVISORY BOARD

First edition 2017

Copyright © 2017, Elsevier Inc. All rights reserved

For information on all Academic Press publications

Publisher: Zoe Kruze

Volume 70 of the Series is dedicated to Professor R.G. Wilkins on the

Chemists,” published by the Royal Society of Chemistry. We pay tribute

The 3rd European Colloquium on Inorganic Reaction Mechanisms, held in

and the mechanism of [3 + 2] cycloaddition with alkynes. In Chapter 9,

The Kinetics and Mechanism

Advances in Inorganic Chemistry, Volume 70 # 2017 Elsevier Inc. 1

available experimental results is essential to validate the mechanistic conclusions.

The main purpose of this chapter is to demonstrate the challenges asso-

... ... ...

Scheme 1 The general scheme of a complex redox reaction; R: reactant, I: intermediate,

3. SELECTED REACTIONS OF OXYCHLORINE SPECIES

2ClO3 + 2Cl + 4H + ! 2ClO2 + Cl2 + 2H2 O (1)

2ClO2 + 2OH ¼ ClO2 + ClO3 + H2 O (3)

a. ClO2 + OH− (HOCl(O)O)− kS1, k-S1 (S1)

HOCl(O)O + OH − ClO3− + H2O kS3, fast (S3)

b. ClO2 + OH− (OClOOH)− kS4, k-S4 (S4)

OClOOH + OH− HOClO + HOO− kS6, fast (S6)

HOO−+ 2 ClO2 + OH− 2 ClO2− + O2 + H2O kS7, fast (S7)

HOClO + OH− ClO2− + H2O kS8, fast (S8)

c. ClO2 + ClO2 Cl2O4 kS9, k-S9 (S9)

HOCl(O)O + OH − ClO3− + H2O kS11, fast (S11)

+H CO2– Keq +H CO2–

+H HOCl Cl– +H3N CO2–

2HOCl Cl2O + H2O Fast, Keq = 0.0115 M–1 (S1)

Cl2 + HClO2 Cl2O2 + Cl− + H+ (S4)

Cl2O2 + ClO2− 2·ClO2 + Cl− (S5)

Cl2O2 + HOCl Cl2 + ClO3− + H+ (S6)

Cl2O + ClO2− Cl2 + ClO3− (S7)

ClO2− + HOCl ClO3− + Cl− + H+ (S8)

ClO2 + OCl− ·Cl2O3− (S9)

ClO2 + ·Cl2O3−+ H2O Cl− + 2H+ + 2ClO3− Fast (S10)

5S3 O6 2 + 8ClO2 + 14H2 O ! 15SO4 2 + 8Cl + 28H + (6)

2SO3 2 + 2ClO2 + H2 O ¼ 2SO4 2 + ClO3 + Cl + 2H + (8)

6SO3 2 + 2ClO2 ¼ 4SO4 2 + S2 O6 2 + 2Cl (9)

ClO3– ClO2– + H2O

In order to explore the mechanistic details time-resolved 1H, 13C NMR

Intensity × 10−6 (a.u.) GLY

4. KINETICS OF THE OXIDATION REACTIONS OF PEROXO

H2S2O8, is unstable in aqueous solutions and eventually undergoes hydro-

seemingly different. In organic chemistry, PMS is used for the oxidation

peroxyacetic and monoperoxyphthalic acid (96)) were found to decompose

HSO5 + SO5 2 ! 2SO4 2 + O2 + H + (15)

one-electron oxidation reactions with PMS usually have relatively small

HSO5 + X ! HSO4 + OX v ¼ kapp ½HSO5 ½X (16)

where X can denote Cl, Br, or I.

H + + OX Ð HOX (17)

HSO5 + SCN ! SO4 2 + OSCN + H + k20 ¼ 2:0 102 M1 s1

v21 ¼ k21 ½HSO5 ½OSCN

The second oxidation step results in the formation of O2SCN, which

HSO5 + Mn + ! SO4 • + Mðn + 1Þ + + OH (22)

Fe3 + + SO4 2 Ð FeSO4+ (24)

of N-heteroaromatic ligands FeðphenÞ3 2 + , FeðbipyÞ3 2 + , and Fe(tpy)2+,

FeLn2 + (+nH+) → Fe2+ + nHL+ (S1)

yields the corresponding N-oxide (phenO) (104). This product inhibits

2Fe(phen)33+ → Fe(phen)32+ + Fe(phen)34+ (S1)

Scheme 10 Mechanism for the oxidation of FeðphenÞ3 2 + by PMS under acidic

phen phenO phenO2

lot smaller than that of Fe(TPPS)–OOH+, the high-valent iron derivative

5. THE PHOTON AS A REACTANT

W lamp Lens Cell

Filter Dispersion unit

the oxidation of other chlorinated phenols by H2O2/Fe(TPPS)+ and the

H2 O SO2 + hν ! *H2 O SO2 (26)