Sdarticle 010

METAL ION CATALYZED AUTOXIDATION
REACTIONS: KINETICS AND MECHANISMS
ISTVÁN FÁBIÁN and VIKTOR CSORDÁS

University of Debrecen, Department of Inorganic and Analytical Chemistry,
Debrecen, Egyetem tér 1, H-4010, Hungary
I. Introduction
II. General Considerations
III. Autoxidation of L -Ascorbic Acid
IV. Autoxidation of Catechols and Related Compounds
V. Autoxidation of Cysteine
VI. Autoxidation of Sulfur(IV)
VII. Autoxidation of Miscellaneous Substrates
VIII. Exotic Kinetic Phenomena
IX. Perspectives
References
I. Introduction
Indisputable biological signi¢cance, industrial applications and envi-

ronmental issues have combined to generate considerable interest in
the redox chemistry of dioxygen which o¡ers an inexhaustible pool of
scienti¢c problems and has challenged researchers for well over
a century. Indeed, considerable e¡ort has been invested into understand-
ing both the basic principles and the ¢ne details of autoxidation
processes. The most fundamental properties are discussed extensively
in inorganic chemistry text books, and a recent monograph by Sawyer
(1) provides a deep insight into oxygen chemistry. Numerous publications
deal with preparative, kinetic, mechanistic, theoretical, biochemical,
analytical and other aspects of autoxidation reactions and thorough
reviews have been compiled from time to time to cover speci¢c areas.
Due to the enormous amount of information reported on these reactive
395
ADVANCES IN INORGANIC CHEMISTRY ß 2003 Elsevier Science (USA)

VOLUME 54 ISSN 0898-8838 All rights reserved.
396 I. FA¤BIA¤N and V. CSORDA¤S
systems, such compilations usually re£ect the interest of the authors and
were often criticized as somewhat biased. This paper is no exception as
we will limit our discussion to recent results on the kinetics and mecha-
nism of metal ion1 catalyzed autoxidation reactions. While we do not
intend to cover older studies in detail, in some cases the most important
results from earlier publications will be surveyed in order to put the
new studies in context.
Metal ions play an important role as catalysts in many autoxidation
reactions and have been considered instrumental in regulating natural
as well as industrial processes. In these reactive systems, in particular
when the reactions occur under environmental or in vivo biochemical
conditions, the metal ions are involved in complicated interactions
with the substrate(s) and dioxygen, and the properties of the actual
matrix as well as the transport processes also have a pronounced
impact on the overall reactions. In most cases, handling and analyzing
such a complexity is beyond the capacity of currently available
experimental, computational and theoretical methods, and researchers
in this ¢eld are obliged to use simpli¢ed sub-systems to mimic the com-
plex phenomena. When the simpli¢ed conditions are properly chosen,
these studies provide surprisingly accurate predictions for the ‘real’
systems. In this paper we review the results obtained in kinetic and
mechanistic studies on the model systems, but we do not discuss their
broad biological or environmental implications.
A brief overview on why most of the autoxidation reactions develop
complicated kinetic patterns is given in Section II. A preliminary
survey of the literature revealed that the majority of autoxidation
studies were published on a small number of substrates such as L -ascor-
bic acid, catechols, cysteine and sul¢te ions. The results for each of
these substrates will be discussed in a separate section. Results on
other metal ion mediated autoxidation reactions are collected in
Section VII. In recent years, non-linear kinetic features were discovered
in some systems containing dioxygen. These reactions form the basis of
a new exciting domain of autoxidation chemistry and will be covered in
Section VIII.
1
The term ‘‘metal ion’’ will be used in a general sense throughout this paper.
Distinction between metal ions and complexes as well as between di¡erent iso-
mers of the complexes will be made only when it is required for the clarity of
the presentation. Furthermore, the accessible coordination sites of the metal
ions are always assumed to be occupied by the corresponding solvent molecules
which will be shown only when they are directly involved in a given redox
reaction.
METAL ION CATALYZED AUTOXIDATION REACTIONS 397
II. General Considerations
Dioxygen is a formidable oxidizing agent with relatively high standard

redox potential under acidic conditions (2):
O2 þ 4Hþ þ 4e )

*
2H2 O "0 ¼ 1:229 V ð1Þ
The redox potential is strongly pH dependent and it decreases consid-

erably in alkaline solution:
O2 þ 2H2 O þ 4e )
4OH
*
"0 ¼ 0:401 V ð2Þ
The thermodynamic properties are also a¡ected by the solvent and the
composition of the reaction mixture, for example the corresponding
values in acetonitrile are +1.79 V (1 M HClO4) and 0.53 V (1 M
Bu4NOH), respectively (3).
The reduction of dioxygen to its fully reduced form, H2O, requires the
transfer of 4 electrons, and the transfer may proceed via a series of inter-
mediate oxidation states, such as O
2 / HOO , HOO / HOOH, O / OH .

These reduced forms of oxygen exhibit di¡erent redox properties and

in the presence of substrate(s) and/or catalyst(s) may open di¡erent
reaction paths for the electron transfer process. Fast proton transfer
reactions between the corresponding acid-base pairs can introduce
composite pH dependencies into the kinetic and stoichiometric charac-
teristics of these systems.
Oxidation reactions by dioxygen are often kinetically hindered and do
not occur even if the reaction is favored thermodynamically.The simplest
explanation for this can be given by assuming that the electron transfer
to produce lower oxidation state intermediates occurs step by step. The
redox potentials for the formation of HOO and O 2 are 0.05 V (2) and
0.325 V (4), respectively. It follows that only relatively strong reducing
agents are capable of reducing dioxygen to O
2 = HO2 and, in terms of
thermodynamics, the limiting step is the formation of the superoxide
radical or its acidic form in any stepwise electron transfer reaction of
O2. The superoxide radical is a relatively strong oxidant and it can read-
ily be reduced to hydrogen peroxide. Thus, once O
2 (HO2 ) is formed,
thermodynamic constraints are not expected to a¡ect the overall
kinetics.
For a more in-depth interpretation of the inertness of dioxygen, the fact
that O2 is a triplet state bi-radical, i.e. it has two unpaired electrons in
the 2pg orbitals, needs to be considered. It follows that the oxidation of
singlet state substrates by the triplet O2 to form singlet products is
spin-forbidden and, as a consequence, relatively slow.
The main role of the metal catalysts is to enhance the O2 + e ! O2

transition or mediate the electron transfer between oxygen and substrate
(S) via an alternative path. The following basic reaction schemes can
be envisioned:
Mnþ þ O2 ! Mðnþ1Þþ þ O
2 ð3Þ
O
2 þ S ! I ! P ð4Þ
Mðnþ1Þþ þ I ! Mnþ þ P ð5Þ

SCHEME 1.
where I and P stand for one or more intermediates and products,

respectively.
In this case, direct electron transfer between the catalyst and O2
produces the superoxide radical (or other reduced forms of oxygen)
which can be involved in a series of subsequent redox reactions. If these
reactions are relatively fast, the rate determining step is Eq. (3) and the
overall process can be interpreted in terms of relatively simple rate laws.
It should be noted that the metal ion may enhance the oxidation rate of
the substrate without being a real catalyst. In this case the oxidation of
the substrate becomes faster because of the formation of reactive species
in Eqs. (3) and (4), but Eq. (5) does not occur and the metal ion is not
recycled to its original oxidation state. Consequently, the ‘catalyst’
quickly loses its activity over the course of the reaction.
Mnþ þ S ! MSnþ ð6Þ
MSnþ þ O2 ! MS Onþ
2 ð7Þ
MS O2 ! P þ Mnþ ð8Þ
SCHEME 2.
In this case, the actual redox step is preceded by the formation of an

adduct or a complex between the catalyst, the substrate and dioxygen.
The order of these reaction steps is irrelevant as long as the rate
determining step is Eq. (8). If Eqs. (6) and (7) are rapidly established
pre-equilibria the reaction rate depends on the concentrations of all
reactants. In some instances, the rate determining step is the formation
of the MSn+ complex and the reaction rate is independent of the
concentration of dioxygen.
It should be added that MSO2 is not necessarily a mono-nuclear

complex. It could be shown in a few cases that the catalytic activity of
the metal ion is due to the formation of dinuclear metal-substrate
complexes. Presumably in these species each oxygen atom of dioxygen
coordinates to a di¡erent metal center. Such systems were extensively
used to model the reactivity patterns of various enzymes containing a
bimetallic active center.
Mnþ þ S1 ! Mðn1Þþ þ I ð9Þ
I þ O2 ! I0 ð10Þ
Mðn1Þþ þ O2 ! Mnþ þ O

2 ð11Þ
I0 ; O
2 þ S2 ! P ð12Þ
SCHEME 3.
This scheme implies that dioxygen is activated only in a secondary step

by an intermediate formed in the initial electron transfer reaction
between the metal ion and a co-substrate, S1. The reduced form of the
metal ion is re-oxidized to its original oxidation state by O2, but such a
reaction with a secondary intermediate cannot be excluded. If the rate
determining step is Eq. (9), the overall reaction is again zeroth order
with respect to dioxygen.
The common element of Schemes 1^3 is that they each postulate direct
interaction between the metal center and dioxygen. Although it is not
stated explicitly, Eqs. (3) and (11) most likely proceed via an inner-
sphere mechanism. Thus, the metal^dioxygen interaction implies spin
pairing between the reactants when the metal ion is paramagnetic. As a
consequence, the formation of the MÔ2 type intermediates circumvents
the restriction posed by the triplet to singlet transition which seems to
be the major kinetic barrier of autoxidation reactions (5).
It should be emphasized that clear-cut situations described in
Schemes 1^3 are uncommon and typically the combination of these
models needs to be considered for kinetic and mechanistic description
of a real system. However, even when one of the limiting cases prevails,
each of these models may predict very di¡erent formal kinetic patterns
depending on where the rate determining step is located. For the same
reason, di¡erent schemes may be consistent with the same experimental
rate law, i.e. thorough formal kinetic description of a reaction and the
analysis of the rate law may not be conclusive with respect to the
mechanism of the autoxidation process.
The general features discussed so far can explain the complexity of

these reactions alone. However, thermodynamic and kinetic couplings
between the redox steps, the complex equilibria of the metal ion and/or
the proton transfer reactions of the substrate(s) lead to further complica-
tions and composite concentration dependencies of the reaction rate.
The speciation in these systems is determined by the absolute concen-
trations and the concentration ratios of the reactants as well as by the
pH which is often controlled separately using appropriately selected
bu¡ers. Perhaps, the most intriguing task is to identify the active form
of the catalyst which can be a minor, undetectable species. When the
protolytic and complex-formation reactions are relatively fast, they can
be handled as rapidly established pre-equilibria (thermodynamic
coupling), but in any other case kinetic coupling between the redox
reactions and other steps needs to be considered in the interpretation
of the kinetics and mechanism of the autoxidation process. This may
require the use of comprehensive evaluation techniques.
The model shown in Scheme 2 indicates that a change in the formal
oxidation state of the metal is not necessarily required during the cata-
lytic reaction. This raises a fundamental question. Does the metal ion
have to possess speci¢c redox properties in order to be an e⁄cient cata-
lyst? A de¢nite answer to this question cannot be given. Nevertheless,
catalytic autoxidation reactions have been reported almost exclusively
with metal ions which are susceptible to redox reactions under ambient
conditions.This is a strong indication that intramolecular electron trans-
fer occurs within the MSn+ and/or MSO2 precursor complexes. Partial
oxidation or reduction of the metal center obviously alters the electronic
structure of the substrate and/or dioxygen. In a few cases, direct
spectroscopic or other evidence was reported to prove such an internal
charge transfer process. This electronic distortion is most likely neces-
sary to activate the substrate and/or dioxygen before the actual electron
transfer takes place. For a few systems where deviations from this pattern
were found, the presence of trace amounts of catalytically active
impurities are suspected to be the cause. In other words, the catalytic
e¡ect is due to the impurity and not to the bulk metal ion in these cases.
Examples of how the general principles apply to speci¢c reactions will
be discussed in the following Sections.
III. Autoxidation of L-Ascorbic Acid
Autoxidation reactions of L -ascorbic acid (H2A) have been the subject

of intensive studies for decades. It was shown that some of the most
common transition metal ions are very active catalysts for this reaction
(6). The majority of earlier studies focused on copper catalysis because
of the biological importance of the reactions of ascorbic acid with
oxidase enzymes containing copper binding sites. Under slightly acidic
or neutral conditions the copper catalyzed reaction was con¢rmed to
produce dehydroascorbic acid (A) and hydrogen peroxide in the initial
phase (7):
H2 A þ O2 ! A þ H2 O2 ð13Þ
The oxidation of dehydroascorbic acid to a higher oxidation product is
relatively slow and could be neglected for all practical purposes in
kinetic and stoichiometric studies.
Earlier results regarding the kinetics and mechanism of the copper(II)
catalysis are controversial. Reaction orders for [O2], [CuII], [H2A] and
[H+] were reported in the following respective ranges: 0.5 to 1, 0.5 to 1, 0
to 1, and 2 to +1 (8). It is also disputed whether the redox cycling of
the catalyst includes oxidation states +1 and +2 or +2 and +3.The discre-
pancies are too marked to be explained only by the di¡erences in the
experimental conditions applied.
The ¢rst thorough study on the Cu(II) and Fe(III) catalyzed autoxida-
tion of ascorbic acid was reported by Taqui Khan and Martell (6). These
authors found evidence for a slow, overall second-order reaction between
the HA form of ascorbic acid and dioxygen in the absence of added cata-
lyst. The corresponding rate constant was reported to be 0.57 M1 s1
(aqueous solution, 25 C, = 0.1 M KNO3). Later studies also supported
the existence of such a reaction path (9), which was negligible under cata-
lytic conditions.
In the presence of Cu(II) or Fe(III), a two-term rate law was con¢rmed
in which the two terms correspond to the reaction paths via the mono-
and diprotonated forms of ascorbic acid. The reaction was found to be
¢rst-order with respect to [H2A]tot, [O2] and [Mn+] with both catalysts:
d½H2 A=dt ¼ ðka14 ½HA þ kb14 ½H2 AÞ ½Mnþ ½O2 ð14Þ
The third-order rate constants for the Cu(II) and Fe(III) catalyzed
reactions are: ka14 ¼ 6:0 107 and 2.4 107 M2 s1, kb14 ¼ 3:8 105 and
4.0 105 M2 s1, respectively.
The possibility that the metal ion is reduced by ascorbate ion in a rate
determing step could be excluded because in that case the rate would be
independent of [O2]. Thus, the following mechanism was proposed for
the catalytic oxidation of HA:

H2 A ) HA þ Hþ
* ð15Þ
HA þ Mnþ )
½MðHAÞðn1Þþ
*
ð16Þ
½MðHAÞðn1Þþ þ O2 )

*
½MðHAÞðO2 Þ
ðn1Þþ
ð17Þ
½MðHAÞðO2 Þðn1Þþ ! ½MðHAÞþ ðO

2 Þ
ðn1Þþ
ð18Þ
½MðHAÞþ ðO
2 Þ
ðn1Þþ
! A þ Mnþ þ HO2 ð19Þ
A þ Mnþ ! A þ Mðn1Þþ ð20Þ
A þ O2 ! A þ O
2 ð21Þ
Mðn1Þþ þ HO2 þ Hþ ! Mnþ þ H2 O2 ð22Þ
O
2 þ H
þ *
) HO2 ð23Þ
In this mechanism, the rate-determining step is the intramolecular
electron transfer from the ligand to dioxygen, Eq. (18), via the metal
center of the [M(HA)(O2)](n1)+ complex for which the following struc-
ture was proposed:
The intramolecular electron transfer leads to fast formation of semi-

quinone and the lower oxidation state metal ion. The catalytic cycle is
completed by fast reoxidation of the metal ion. Signi¢cant deviations
from this model were observed at low dioxygen concentrations and it
was suggested that another oxidation path becomes operative under
such conditions. Although earlier they had been proposed to participate
(10), side reactions with dehydroascorbic acid could be excluded.
The same model was applied to the oxidation of the H2A form of ascor-
bic acid. In this case iron(III) was found to be a somewhat more active
catalyst than copper(II). The di¡erence could be explained by assuming
that Fe(III) forms a more stable complex with H2A than does Cu(II)
because of the higher charge of the metal ion.
The formation of the [M(HA)](n1)+ complex was con¢rmed in indepen-

dent pH-metric experiments in the case of copper(II). These studies also
provided evidence that ascorbic acid is coordinated to the metal center
in its monoprotonated form. Because of relatively fast redox reactions
between iron(III) and ascorbic acid, similar studies to con¢rm the
formation of [Fe(HA)]2+ were not feasible. However, indirect kinetic
evidence also supported the formation of the [M(HA)](n1)+ complex in
both systems (6).
Electron transfer within the [M(HA)(O2)](n1)+ complex was envi-
sioned as a two-stage process in which ¢rst, a 2p electron of the
ascorbate oxygen is transferred to a t2g non-bonding or an eg antibonding
orbital of the metal ion.The subsequent step is the transfer of an electron
to the py 2p or pz 2p orbital of the oxygen molecule.
Although some of the results reported by Taqui Khan and Martell (6)
were challenged in subsequent papers, several elements of their
model were included in the interpretation of the results obtained in
related studies.
The kinetic results reported by Jameson and Blackburn (11,12) for
the copper catalyzed autoxidation of ascorbic acid are substantially
di¡erent from those of Taqui Khan and Martell (6). The former could
not reproduce the spontaneous oxidation in the absence of added cata-
lysts when they used extremely pure reagents. These results imply that
ascorbic acid is inert toward oxidation by dioxygen and earlier reports
on spontaneous oxidation are artifacts due to catalytic impurities. In
support of these considerations, it is worthwhile noting that trace
amounts of transition metal ions, in particular Cu(II), may cause
irreproducibilities in experimental work with ascorbic acid (13). While
this problem can be eliminated by masking the metal ion(s), the masking
agent needs to be selected carefully since it could become involved in
side reactions in a given system.
Jameson and Blackburn con¢rmed the following rate law in 0.1 M
KNO3 and at 25 C (11):
d½O2 =dt ¼ k24 ½CuðIIÞ½HA ½O2 1=2 ð24Þ
with k24 = 2.2 103 M3/2 s1.

The di¡erence in the rate expressions shown in Eqs. (14) and (24)
probably re£ects the di¡erences in the experimental conditions of the
corresponding studies. It should be noted that the validity of Eq. (14)
was con¢rmed at relatively high [O2] (6), and the deviations from the
rate law at lower oxygen concentrations found in that study seem to be
consistent with Eq. (24). The half-order dependence on [O2] led to the
conclusion that a chain mechanism is operative in this reaction, and the

following kinetic model was proposed (11):
Cu2þ þ HA )

*
½CuðHAÞ
þ
K25 ð25Þ
2þ
2½CuðHAÞþ )

*
½Cu2 ðHAÞ2 K26 ð26Þ
½Cu2 ðHAÞ2 2þ þ O2 )

½Cu2 ðHAÞ2 ðO2 Þ2þ
*
K27 ð27Þ
½Cu2 ðHAÞ2 ðO2 Þ2þ ! ½CuAðO2 HÞ þ Cu2þ þ A þ Hþ k28 ð28Þ
½Cu2 ðHAÞ2 ðO2 Þ2þþ A! ½CuAðO2 HÞ þ½CuðHAÞþ þA k29 ð29Þ
½CuAðO2 HÞ ! Cu2þ þ A þ HO

2 k30 ð30Þ
2½CuAðO2 HÞ ! 2Cu2þ þ HA þ A þ 2O2

2 k31 ð31Þ
Equilibrium studies under anaerobic conditions con¢rmed that
[Cu(HA)]+ is the major species in the Cu(II)âscorbic acid system.
However, the existence of minor polymeric, presumably dimeric, species
could also be proven. This lends support to the above kinetic model.
Provided that the catalytically active complex is the dimer produced in
reaction (26), the chain reaction is initiated by the formation and subse-
quent decomposition of [Cu2(HA)2(O2)]2+ into [CuA(O2H)] and A.
The chain carrier is the semi-quinone radical which is consumed and
regenerated in the propagation steps, Eqs. (29) and (30). The chain is ter-
minated in Eq. (31). Applying the steady-state approximation to the con-
centrations of the radicals, yields a rate law which is fully consistent
with the experimental observations:
k24 ¼ k30 ðk28 =k31 Þ1=2 ðK25 K26 K27 Þ1=2 ð32Þ
Chloride ion has a profound e¡ect on the kinetics because the rate is
half-order in both ascorbic acid and dioxygen, but ¢rst-order in Cu(II)
in 0.1 M KCl (12). It was assumed that in this case the catalytically
active species is a [Cu2A]2+ type complex which somehow also incorpo-
rates chloride ion; however, the exact composition of the complex was
not clari¢ed. The role of this dimer is very similar to that of
[Cu2(HA)2(O2)]2+ in the absence of Cl, i.e. after coordinating dioxygen
it generates [CuA(O2)] (the conjugate base form of [CuA(O2H)]) and
A. These species are involved in propagation steps analogous to those
in Eqs. (29) and (30). An additional reaction sequence was also included
in the model which postulates the formation of Cu(III):
½Cu2 AðO2 Þ2þ ! ½CuAðO2 Þ þ Cu3þ ð33Þ
Cu3þ þ H2 O )
Cu2þ þ OH þ Hþ
*
ð34Þ
OH þ HA ! A þ H2 O ð35Þ

According to these models, oxygen is activated via the formation of the
dinuclear oxygen complexes. It was hypothesized that the binding of
oxygen is perpendicular to the Cu^Cu bond and the px and py orbitals of
dioxygen interact with the dxz and dyz metal orbitals.Thus, the reduction
of O2 may occur in a concerted two-electron process. This assumption is
in strict contrast with the stepwise one-electron mechanism proposed
by Taqui Khan and Martell (6). A further di¡erence is that the formation
of a Cu(I) intermediate was excluded from the latter models.
Features of the UV^Vis spectra obtained in the absence of oxygen indi-
cate ligand to metal charge transfer in the Cu(II)âscorbate complexes
(11). This would suggest that the formation of complexes, formally
copper(I), is favored in this reaction even if the oxidation of ascorbic
acid by copper(II), i.e. the complete electron transfer is thermodynami-
cally unfavorable (14). First, Jameson and Blackburn presented con-
vincing arguments that the electronic structure of the ascorbate
complexes in the absence of oxygen is irrelevant with respect to the
autoxidation reaction and the catalytic reaction proceeds via a Cu(II)^
Cu(III) cycle (11,12). They proposed that the two-electron reduction of
oxygen in [Cu2(HA)2(O2)]2+ yields two non-equivalent copper(III)âscor-
bate units. While one of them immediately dissociates into Cu2+ and the
semi-quinone radical, copper(III) remains bound to ascorbate and
peroxide in the other species. It was noted that the oxidation state of
copper cannot be established ¢rmly in the latter species and a
copper(II)^semiquinone^peroxide arrangement can also be feasible. In
any case, the formation of copper(I) was clearly excluded, and a Cu(III)
intermediate was explicitely included in the kinetic model in the
presence of Cl.
In a follow-up study, Jameson and Blackburn studied the e¡ect of
chloride ion in more detail and reported the following concentration
dependence of the pseudo-half-order rate constant for the dioxygen
decay (15):
A þ B ½Cl
kobs ¼ ð36Þ
C þ D ½Cl þ E ½Cl 2
where A, B, C, D and E are experimentally determined parameters.

The e¡ect of chloride ion was interpreted in terms of the formation of
various ternary complexes between Cu(II)âscorbate and Cl. It was
demonstrated that the involvement of copper(I) is feasible and a corre-
sponding mechanism was presented as an alternative to the Cu(II)/
Cu(III) model.
In the absence of chloride ion, the Cu(I)/Cu(II) catalytic cycle could be
initiated by an intramolecular redox reaction of the [Cu2(HA)2]2+ dimer
to yield unsymmetrical products:
½Cu2 ðHAÞ2 2þ ! ½CuðHAÞ þ Cuþ þ A þ Hþ ð37Þ
Cuþ þ HA ! ½CuðHAÞ ð38Þ

In the following steps dioxygen coordinates to the Cu(I) complex and is
subsequently reduced to hydrogen peroxide:
½CuðHAÞ þ O2 ! ½CuðHAÞO2 ð39Þ
½CuðHAÞO2 þ ½CuðHAÞ ! ½Cu2 ðHAÞ2 2þ þ O2

2 ð40Þ
O2 þ
2 þ 2H ! H2 O2 ð41Þ

The composite e¡ect of Cl was interpreted by assuming that the for-
mation of Cu(I) intermediates is enhanced via the [Cu2ACl2] complex.
However, the relative concentration of this species decreases with
increasing [Cl] because of the formation of the [Cu2ACl4]2 complex,
and ultimately this leads to chloride ion inhibition at higher concentra-
tions of Cl.
Reports by Li and Zuberbu«hler were in support of the formation of
Cu(I) as an intermediate (16). It was con¢rmed that Cu(I) and Cu(II)
show the same catalytic activity and the reaction is ¢rst-order in [Cu(I)
or (II)] and [O2] in the presence of 0.6^1.5 M acetonitrile and above
pH 2.2. The oxygen consumption deviated from the strictly ¢rst-order
pattern at lower pH and the corresponding kinetic traces were excluded
from the evaluation of the data. The rate law was found to be identical
with the one obtained for the autoxidation of Cu(I) in the absence of
Cu(II) under similar conditions (17). Thus, the proposed kinetic model is
centered around the reduction of Cu(II) by ascorbic acid and reoxidation
of Cu(I) to Cu(II) by dioxygen:
2Cu2þ þ HA ! 2Cuþ þ A þ Hþ ð42Þ
2Cuþ þ O2 þ 2Hþ ! 2Cu2þ þ H2 O2 ð43Þ

The reaction features a complex pH dependence which was not

resolved. Nevertheless, it was suggested that the variation of the reaction
rate as a function of pH was consistent with an additional term in the
rate law which was proportional to [HA]. The kinetic role of [HA] was
interpreted in terms of its reaction with CuOþ 2 , which is presumably
formed as an intermediate during the autoxidation of Cu(I). In the
presence of acetonitrile, Cu(I) can be stabilized as [Cu(MeCN)2]+ and
[Cu(MeCN)3]+. As a consequence, the oxidation of Cu(I) becomes rate
determining and the overall rate of the catalytic reaction becomes
independent of [HA].
The feasibility of the above model rests on the formation of Cu(I) in the
copper(II)âscorbic acid system. A recent study ¢rmly established that
Cu(I) can indeed accumulate in the presence of a stabilizing ligand,
Cl, and in the absence of O2 (14).The actual form of the rate law is deter-
mined by the relative rates of Cu(I) formation and consumption, and
further studies should clarify how the stability and reactivity of
copper(I) are a¡ected by the presence of various components and the
conditions applied.
A non-participating ligand (L) in the coordination sphere of Cu(II) may
signi¢cantly alter the reactivity patterns discussed above. When
bis(histidine)copper(II),2 CuL2, is used as a catalyst the initial rate
shows saturation as a function of both [HA] and [O2] at physiological
pH (18). This observation was interpreted by postulating the formation
of a ternary complex which undergoes an intramolecular redox reaction
and produces A and O 2 in a rate determining step. These species are
involved in fast subsequent reactions:3
½HA ”CuL2 ”O2 ! CuL2 þ A þ O

2 þH
þ
ð44Þ
HA þ O þ
2 þ H ! A

þ H2 O2 ð45Þ
A þ O þ
2 þ 2H ! A þ H2 O2 ð46Þ
According to ESR measurements, the semiquinone radical forms at

nM concentration levels and its steady-state concentration was
reported to increase by increasing the total concentration of ascorbic
acid. The kinetic role of O
2 was con¢rmed by the inhibitory e¡ect of
2
The bis-complex, CuL2, was marked as a double positive ion in the original
paper. However, the carboxylic group of histidine is expected to be deproto-
nated and we prefer to consider this species as a neutral complex.
3
For sake of simplicity, the recombination steps of A and O
2 are not shown.
Cu,Zn superoxide dismutase. Essentially, these observations support a

stepwise one-electron model again. Interestingly, the oxidation state of
copper does not change during the catalytic reaction, i.e. the sole
kinetic role of the histidine coordinated metal center is to alter the
electronic structures of the substrate and O2 in order to facilitate the
electron transfer process between them.
The coordination mode and structure of a given copper(II) complex are
important characteristics in determining how the complex is involved
in the autoxidation reaction. For example, the dinuclear [Cu2(BPY)2
(m-OX)]X2 complex (BPY = 2,20 -bipyridine, OX ¼ C2 O2
4 , X ¼ ClO4 ; PF6 )
catalyzes the autoxidation of ascorbic acid via a Cu(I)/Cu(II) cycle. In
this case, ascorbic acid is presumably oxidized in a two-electron transfer
process after it cleaves the dimer by protonating the bridging ligand
(19). This reaction was used to model the functional properties of dopa-
mine b-hydroxylase and was shown to convert benzylamine to benzalde-
hyde e¡ectively. The Cu(I)/Cu(II) cycle was also proposed in a series of
reactions catalyzed by dinuclear copper(II) complexes of Schi¡ bases
(L), Cu2LCl3. However, detailed kinetic analysis was not performed in
these systems (20).
Iron(III)-catalyzed autoxidation of ascorbic acid has received consider-
ably less attention than the comparable reactions with copper species.
Anaerobic studies con¢rmed that Fe(III) can easily oxidize ascorbic
acid to dehydroascorbic acid. Xu and Jordan reported two-stage kinetics
for this system in the presence of an excess of the metal ion, and
suggested the fast formation of iron(III)âscorbate complexes which
undergo reversible electron transfer steps (21). However, Ba«nsch and co-
workers did not ¢nd spectral evidence for the formation of ascorbate
complexes in excess ascorbic acid (22). On the basis of a combined pH,
temperature and pressure dependence study these authors con¢rmed
that the oxidation by FeðH2 OÞ3þ 6 proceeds via an outer-sphere mechan-
ism, while the reaction with Fe(H2O)5OH2+ is substitution-controlled
and follows an inner-sphere electron transfer path. To some extent,
these results may contradict with the model proposed by Taqui Khan
and Martell (6), because the oxidation by the metal ion may take place
before the ternary oxygen complex is actually formed in Eq. (17).
An interesting, pH-dependent mechanistic changeover was reported
in the H2AÔ2^[FeIII(TPPS)] (TPPS = 5,10,15,20 -tetrakis(p-sulfonatophe-
nyl)porphyrinate) system in aqueous solution (23). This water-soluble
metalloporphyrin exists as a monomer under slightly acidic conditions
and reacts with ascorbic acid on the time-scale of several hours to pro-
duce [FeII(TPPS)] in the absence of dioxygen. The formation of the
[FeIII(TPPS+)] radical was also reported at pH 5, but it is not clear how
this species could form when no oxidizing agent was present. In the pre-
sence of oxygen, the decay of the characteristic Soret band of
[FeIII(TPPS)] becomes considerably faster. The kinetic model for this
reaction assumes a series of one-electron steps which are initiated
by the oxidation of HA to A by [FeIII(TPPS)]. The reactions of the
semi-quinone and [FeII(TPPS)] products with O2 generate the O 2 radical
which in turn oxidizes the catalyst to [FeIII(TPPS+)].The model predicts
redox cycling, but the experimental observations clearly indicate that a
substantial amount of the catalyst is lost during the reaction.
Apparently, an e⁄cient reaction channel is not available to reduce the
[FeIII(TPPS+)] radical back to the original complex. The model o¡ers
reasonable qualitative interpretation of the experimental data.
However, it requires further re¢nement in order to explain why the
initial HA^[FeIII(TPPS)] electron transfer step, which controls the over-
all reaction, becomes faster when oxygen is added.
In alkaline solution (pH 11), the complex is present as a m-oxo dimer
and ascorbic acid is fully deprotonated. In the absence of oxygen, kinetic
traces show the reduction of Fe(III) to Fe(II) with a reaction time on the
order of an hour at [H2A] = 5103 M. The product [FeII(TPPS)] is very
sensitive to oxidation and is quickly transformed to Fe(III) when O2 is
added. This leads to a speci¢c induction period in the kinetic traces
which increases with increasing [O2]. The net result of the induction
period is the catalytic two-electron autoxidation of ascorbic acid in
accordance with the following kinetic model (23):
½FeIII ðTPPSÞ2 O þ A2 ! 2½FeII ðTPPSÞ þ A þ O2 ð47Þ
2½FeII ðTPPSÞ þ O2 ! 2 ½FeIII ðTPPSÞ þ O2

2 ð48Þ
½FeIII ðTPPSÞ þ 2OH ! ½FeIII ðTPPSÞ2 O þ H2 O ð49Þ

This reaction sequence is completed by protolytic reactions of the
dianionic radicals.
The kinetic consequence of the non-participating ligand was also
noticed in the autoxidation reactions catalyzed by Ru(III) ion,
Ru(EDTA) (1 : 1) and Ru(IMDA) (1 : 1) (EDTA = ethylenediaminetetraace-
tate, IMDA = iminodiacetate) (24,25). Each reaction was found to be ¢rst
order in ascorbic acid and the catalysts and, owing to the protolytic equi-
librium between HA/H2A, an inverse concentration dependence was
con¢rmed for [H+]. Only the oxygen dependencies were di¡erent as the
Ru(III)-catalyzed reaction was half-order in [O2], whereas the rates of
the Ru(III)-chelate-catalyzed reactions were independent of [O2]. In the
latter cases, the rate constants were in good agreement with those
obtained for the oxidation of H2A by the same Ru(III) complexes under
anaerobic conditions.
The kinetic models for these reactions postulate fast complex-forma-
tion equilibria between the HA form of ascorbic acid and the catalysts.
The noted di¡erence in the rate laws was rationalized by considering
that some of the coordination sites remain unoccupied in the
[Ru(HA)Cl2] complex. Thus, O2 can form a m-peroxo bridge between two
monomer complexes: [Cl2(HA)RuÔÔ^Ru(HA)Cl2]. The rate determin-
ing step is probably the decomposition of this species in an overall four-
electron transfer process into A and H2O2. Again, this model does not
postulate any change in the formal oxidation state of the catalyst
during the reaction.
In the case of the Ru(III)-chelates, the crowded coordination sphere
around the metal center prevents the coordination of O2.Thus, the corre-
sponding kinetic model postulates that ascorbic acid is oxidized by
Ru(III) in two subsequent redox steps and Ru(II) is reoxidized by O2:
½RuIII Lðn3Þ þ HA )
½RuIII LðHAÞðn2Þ þ H2 O
*
K50 ð50Þ
½RuIII LðHAÞðn2Þ ! ½RuII Lðn2Þ þ A þ Hþ k51 ð51Þ
½RuIII Lðn3Þ þ A ! ½RuII Lðn2Þ þ A fast ð52Þ
2½RuII Lðn2Þ þ O2 þ 2Hþ ! ½RuIII Lðn3Þ þ H2 O2 fast ð53Þ

where n = 2 (IMDA) or 4 (EDTA). Reported values for log K50 and
k51 are respectively 4.22 M1 and 6.7 104 s1 (IMDA) and 5.23 M1
and 4.4 104 s1(EDTA) at 25 C and = 0.1 M KNO3. On the basis of
these results, it is not obvious why the Ru(III)>Ru(IMDA)>Ru(EDTA)
reactivity order was proposed for the catalysts (24).
The similarities in the rate constants and activation parameters (25)
for the reactions with the Ru(III) chelates strongly suggest that the corre-
sponding reactions proceed via the same mechanism. Peculiarly, the
EDTA complex binds HA about an order of magnitude stronger than
the IMDA complex. This trend is in contrast to plausible considerations,
because the number of available coordination sites and electrostatic
interactions would favor stronger coordination to the IMDA complex. A
clear-cut explanation is not readily available for this ¢nding.
An interesting aspect of the results is that catalysis by the IMDA
complex proceeds via the one-electron instead of the four-electron
path. The three empty coordination sites in this complex would allow
simultaneous coordination of HA and O2 and the formation of a
dinuclear complex similar to the one formed with the Ru(III) ion.
This may indicate that factors other than the availability of empty
coordination sites may also have important mechanistic implications.
In the autoxidation kinetics of a series of Fe(II)-aminopolycarboxylato
complexes the signi¢cance of steric e¡ects was unequivocally con¢rmed
(26). These results may also bear some relevance with respect to the
Ru(III)-catalyzed reactions.
The Ru(III)^H2AÔ2 systems were also used for the oxidation and
epoxidation of various organic substrates. These reactions will be
discussed in Section VII.
IV. Autoxidation of Catechols and Related Compounds
The kinetics and mechanisms of autoxidation reactions of catechols

(H2C) are similar to those of ascorbic acid in many ways. This is due to
the common HO^C^CÔH motif in these molecules, which are trans-
formed into the corresponding di-ketone species in the oxidation process.
In addition, most of these reactions proceed via an inner-sphere mecha-
nism and the two neighboring OH groups are the key structural factors
in the formation of the precursor complex between the catalyst and the
substrate. Beside the similarities, these reactions may also show signi¢-
cant di¡erences in the product distributions. The oxidation of ascorbic
acid always leads to the formation of dehydroascorbic acid in the ¢rst
stage of the reaction. This compound is reasonably stable and resistant
to further oxidation which may occur only in very slow subsequent
reaction steps. In contrast, the formation of the di-ketonic quinone (Q)
is not necessarily preferential over the formation of other products in
the oxidation of catechols. These reactions typically yield more than one
product which can form either directly from the substrate in parallel
reaction steps or through a common intermediate. Metal ions catalyze
not only the quinone formation but also intra- and extradiol oxidative
cleavage of catechols leading to a variety of aliphatic products such as
muconic acid or furanone derivatives, (27^31) and various polymeric
species (32,33).
In non-aqueous solution, the copper catalyzed autoxidation of catechol
was interpreted in terms of a Cu(I)/Cu(II) redox cycle (34). It was
assumed that the formation of a dinuclear copper(II)-catecholate inter-
mediate is followed by an intramolecular two-electron step. The product
Cu(I) is quickly reoxidized by dioxygen to Cu(II). A somewhat di¡erent
model postulated the reversible formation of a substrate-catalyst-dioxy-
gen ternary complex for the Mn(II) and Co(II) catalyzed autoxidations
in protic media (35).
In a thorough study by Balla and coworkers, it was shown that in

slightly acidic aqueous solution (25 C, = 1.0 KNO3), the mechanism is
very similar to that followed in aprotic media (36). Parallel measurement
of the O2 consumption and quinone formation con¢rmed the following
stoichiometry up to 35% conversion:
ð54Þ
At longer reaction times, the formation of an acidic product, probably

cis,cis-muconic acid, and a copper containing precipitate were observed.
This latter could be a polynuclear product of quinone and semiquinone
fragments (37,38). In agreement with this stoichiometry, about 50% of
O2 was regenerated by adding a small amount of catalase to the reaction
mixture at relatively short reaction times.
The stability constants of the mono- and bis-complexes between Cu(II)
and catecholate were determined under anaerobic conditions and were
found to be the same as reported earlier, i.e. log 1 = 13.64 (CuC) and
log 2 = 24.92 ðCuC2þ
2 Þ (36,39). A comparison of the speciation and oxida-
tion rate as a function of pH clearly indicated that the mono-catechol
complex is the main catalytic species, though the e¡ect of other com-
plexes could not be fully excluded. The rate of the oxidation reaction
was half-order in [O2] and showed rather complex concentration depen-
dencies in [H2C]0, [Cu(II)]0 and pH. The experimental data were consis-
tent with the following rate equation:
d½O2 =dt ¼ ðka55 ½HC þ kb55 ½CuC þ kc55 ½CuC2 2þ
2 Þ ð½Cu ½CuC½O2 Þ
1=2
ð55Þ
with ka55 ¼ ð3:2 0:2Þ 103 M3=2 s1 , kb55 ¼ 15:5 0:7 M3=2 s1 and kc55 ¼
ð3:0 0:3Þ 102 M3=2 s1 .
The possibility of a free-radical type mechanism was excluded on the
basis that neither superoxide dismutase (which would remove the super-
oxide radical, O 2 ) nor acrylonitrile (which is a very e⁄cient radical
scavenger) had a signi¢cant e¡ect on the reaction rate. In contrast, the
reaction rate went through a maximum upon adding acetonitrile in
increasing concentrations. A similar e¡ect was also observed in
copper(II) catalyzed autoxidation of ascorbic acid by Shtamm et al. (40).
This observation was interpreted by assuming that the formation of a
ternary Cu(II)âcetonitrile-substrate complex enhances the rate of
intramolecular electron transfer, but at higher concentrations acetoni-
trile stabilizes Cu(I) against reoxidation. Though this interpretation is
not con¢rmed thoroughly, the kinetic e¡ect of acetonitrile strongly
suggests the formation of copper(I) as an intermediate.This conclusion is

also supported by the results reported by Speier for copper(I)-catalyzed
autoxidation of 3,5 -di-tert-butylcatechol (H2DTBC) in non-aqueous
solvents (41). In that reaction, the formation of a dimeric species,
[Cu2O2]2+, was postulated in a rate determining step.
The following non-radical chain mechanism was proposed for the
reaction in aqueous solution (for sake of simplicity, fast protolytic and
complex-formation reactions are not shown) (36):
Initiation:
Cu2þ þ CuC )
2Cuþ þ Q
*
k56 ; k56 ð56Þ
Chain propagation:
Cuþ þ O2 )

*

þ
CuO2 K57 ð57Þ
CuOþ þ
2 þ HC ! Cu þ HO2 þ Q k58 ð58Þ
2 2þ
CuOþ þ
2 þ CuC ! Cu þ O2 þ Cu þQ k59 ð59Þ
2 2
CuOþ þ
2 þ CuC2 ! Cu þ O2 þ CuA þ Q k60 ð60Þ
Termination:
2þ
CuOþ þ
2 þ Cu ! 2Cu þ O2
2 k61 ð61Þ
It was suggested that initiation proceeds via the same dimer that was
proposed in aprotic media (34). Furthermore, the reverse step of this
reaction was considered to be very slow. The CuOþ 2 intermediate is
formed in a reversible step in this model, which is in agreement with
the results reported for autoxidation of Cu(I) (17). It should be added
that Shtamm et al. assumed that this reaction is irreversible (40).
Provided that the rates of the initiation and termination are
equal, the model predicts the experimental rate law, and it can be
shown that ka55 ¼ k58 ðk56 K57 =k61 Þ1=2 , kb55 ¼ k59 ðk56 K57 =k61 Þ1=2 and kc55 ¼
k60 ðk56 K57 =k61 Þ1=2 .
Results from subsequent studies were consistent with the model pro-
posed by Balla et al. (36) in that copper accelerates the autoxidation of
catechins4 through the Cu(I)/Cu(II) redox cycle and complex formation
4
The term catechin is often used for the derivatives of catechol (1,2 dihydroxy
benzene).
of the substrate with Cu(II) is an important part of the overall reaction

(42^44). Nevertheless, in some cases one-electron oxidation and the
formation of a semiquinone intermediate was proposed instead of the
two-electron model.
Catalytic autoxidation of four di¡erent catechins occurring
naturally in green tea, i.e. ()-epicatechin, ()-epicatechin gallate,
()-epigallocatechin and ()-epigallocatechin gallate, was studied by
Mochizuki et al. by monitoring the reaction with amperometry, spectro-
photometry and a Clark-type oxygen sensing electrochemical system
(44). Special attention was paid to ()-epigallocatechin gallate which
is particularly prone to oxidation. Typical kinetic traces showed
an induction period (Fig. 1) which was presumably due to the initial
accumulation of a reactive intermediate.
After the addition of Cu(II), absorbance changes were identical under
both anaerobic and aerobic conditions. This was taken as evidence for
semi-quinone (SQ) formation, though it is not clear why the oxidation
would stop at the SQ stage in the presence of excess dioxygen. It should
be noted that the two-electron oxidation would also lead to the same
spectral changes in the presence and absence of dioxygen provided that
the spectrum of the oxidized species corresponds to that of the quinone.
The dependence of the initial steady-state autoxidation rate (de¢ned as
FIG. 1. Amperometric monitoring of the autoxidation of epigallocatechin gallete in

the presence of (A) 0, (B) 2.0, (C) 5.0, (D) 10, (E) 20, and (F) 50 mm CuCl2. The measure-
ments were performed in 0.1 M Tris buffer (pH 9.0) with a Clark type oxygen electrode
at 28 C. The epigallocatechin gallate concentration was fixed at 50 mm. The catechin
stock solution was injected into the test solution at t = 0. The inset shows the (initial)
steady-state autoxidation rate as a function of Cu 2+ concentration. Reprinted
from Biochimica et Biophysica Acta, vol. 1569, Mochizuki, M.; Yamazaki, S.; Kano, K.;
Ikeda,T., Kinetic analysis and mechanistic aspects of autoxidation of catechins, p. 35,
Copyright (2002), with permission from Elsevier Science.
the rate just after the induction period) on [Cu2+] (cf. Fig. 1, inset) indi-
cated some sort of complex formation between the substrate and catalyst,
but this e¡ect was not further evaluated. A one-electron oxidation
model, which postulates the formation of the O 2 intermediate, was
proposed:
H2 C”R þ Cu2þ ! SQ”R þ Cuþ þ 2Hþ ð62Þ
Cuþ þ O2 ! Cu2þ þ O

2 ð63Þ
However, direct evidence was not presented for the formation of the
superoxide radical in the presence of Cu(II) and, as indicated above, the
reported observations can be interpreted in terms of the two-electron
oxidation model equally well.
It was established that Cu(II) promotes the formation of hydrogen per-
oxide. However, most likely this e¡ect was partly o¡set by a Fenton-type
side reaction with Cu(I):
H2 O2 þ Cuþ ! OH þ OH þ Cu2þ ð64Þ
In agreement with literature results (45,46), independent experiments
con¢rmed that the addition of an aliquot of a CuCl acetonitrile solution
to an H2O2 solution induced the immediate decay of H2O2 at pH 9.0 (44).
Most likely, the OH radical produced is involved in fast oxidation of
the H2C^R, SQ^R or even the Q^R forms of the substrate and the
stoichiometry shown in Eq. (54) is not valid anymore. The formation of
free radicals was excluded under acidic conditions (36) implying that
the Fenton-type decomposition of H2O2 may gain signi¢cance only in
alkaline solution.
Mochizuki et al. also demonstrated that the uncatalyzed autoxidation
of catechins is suppressed in the presence of a su⁄cient amount of
borate bu¡er (44). This was interpreted by considering the adduct forma-
tion between borate ion and the catechins. The oxidation resumes on
addition of Cu(II) in the presence of the borate ions. This may indicate
that Cu(II) forms more stable complex(es) with the catechins than the
borate ion, but direct oxidation of the catechin^borate complex by
Cu(II) cannot be excluded either.
The autoxidation of 3,5 -di-tert-butylcatechol (H2DTBC) was frequently
used to test the catalytic activity of various metal complexes. Speier stud-
ied the reaction with [Cu(PY)Cl] (PY = pyridine) in CH2Cl2 and CHCl3,
and reported second-, ¢rst- and zeroth-order dependence with respect
to Cu(I), O2 and substrate concentrations, respectively (41). The results
are consistent with a kinetic model in which the rate determining
oxidation of Cu(I) is followed by fast reduction of Cu(II) by H2DTBC.
In spite of the signi¢cant di¡erences in the catalysts and conditions

applied, essentially the same kinetic model was proposed for the cata-
lytic reactions with two pyrazolate-bridged dicopper(II) complexes,
[Cu2(LEP)2]2+ and [Cu2(BLEP)(OH)]2+ in 1 : 1 methanol^water mixture
(47). The product was con¢rmed to be 3,5 -di-tert-butyl-1,2-benzoquinone
(DTBQ).
In the case of these catalysts, the two metallic redox centers are
provided by the same molecule. Therefore, the reaction is ¢rst-order in
the complex and dioxygen concentrations:
d½O2 =dt ¼ k65 ½Cu2 L½O2 ð65Þ
The metal centers are separated by 370 and 365 pm in the solid state in
[Cu2(LEP)2](ClO4)2 and [Cu2(BLEP)Cl(H2O)]Cl, respectively. In both
complexes, two square pyramidal copper ions are connected in the
equatorial plane with the apical sites located on the same side of the
complex. This cofacial geometry is favorable for simultaneous binding
the catechol molecule to the two metal ions. In the case of the reduced
forms, O2 may bind to one apical site but coordination in a bridge posi-
tion is also feasible. Additional coordination modes can also be envi-
sioned in the case of [Cu2(BLEP)(OH)]2+ by assuming that the OH
group is displaced. Thus, the substrate may bind to the Cu(II) Cu(II)
complex asymmetrically by replacing the OH group with one oxygen
and coordinating to one of the apical sites with the other. In this case,
the reduced form of the catalyst may coordinate dioxygen as a m-peroxo
bridge between the two Cu(I) centers.
Alternative kinetic models were considered for this reaction. Both of
them predict rapid reduction of Cu(II) Cu(II) to Cu(I) Cu(I)
by H2DTBC and subsequent formation of the CuðIIÞðO2 2 ÞCuðIIÞ
intermediate in the reaction of the reduced form with O2. The ¢rst
model assumes that the rate determining formation of the intermediate
is followed by a fast, acid assisted dissociation into the oxidized form of
the catalyst and H2O2. In the other model, the rate determining step is
the oxidation of H2DTBC by the intermediate. The two models predict
the same rate law and cannot be distinguished. However, it is very likely
that the m-peroxo-bridged intermediate decomposes before any intermo-
lecular redox reaction takes place, such that the second model can be
rejected.
Autoxidation became slower when the pH was decreased. Because the
substrate is not involved in the rate-determining steps, its protolytic
equilibria cannot account for this observation. A straightforward
explanation was not readily available for this observation.
At pH 6.40 and 25 C, the values of k65 for [Cu2(LEP)2]2+ and
[Cu2(BLEP)(OH)]2+ are 77.5 M1 s1 and 6.98 103 M1 s1, respectively
(47). The di¡erence in the rate constants indicates that the interaction
between the reduced form of the catalyst and O2 is stronger with the
latter complex. This may re£ect that the BLEP ligand maintains the
dimer structure in the reduced oxidation state whereas the LEP complex
may dissociate into monomers. An additional factor could be that the
replacement of the OH bridge allows the formation of a m-peroxo bridge
in an equatorial position with the BLEP complex, while O2 can coordi-
nate only in the apical position(s) in the other case.The catalytic activity
of [Cu2(BDPDZ)]Cl2 (BDPDZ = 3,6 -bis-(di-2 -pyridylmethyl)pyridazine)
in the autoxidation of H2DTBC was also attributed to the adduct
formation between dioxygen and the two Cu(I) centers (48).
A study with a series of [Cu(R-TMED)X2] complexes (R-TMED =
N,N,N 0-trimethyl-N 0-alkylethylenediamine, X = Cl, Br, ACAC = acetyl-
acetone, R = CH3^, C12H25^, C16H33^) demonstrated that a micellized
environment enhances the catalytic activity of copper(II) in the autoxida-
tion of catechols (49). The observations, in the presence of Cu(II) surfac-
tants above their critical micelle concentrations (CMC), can be
summarized as follows: (i) in large excess of O2 and at constant catalyst
concentration the reactions were ¢rst-order in the substrate concentra-
tion; (ii) when the same head group and counter ion were used the surfac-
tant with longer alkyl chain was more active; (iii) when the head group
and the alkyl chain were the same, the bromide surfactant was a better
catalyst than the chloride analogue; (iv) the [Cu(R-TMED)(ACAC)X]
complex was less e⁄cient than the [Cu(R-TMED)X2] complexes
(X = halogenide ion); (v) the reactivity order of the catechols was found
to be 3,5 -di-tert-butylcatechol>catechol>D- or L -dopamine.
The results were interpreted in terms of the model proposed by Balla
and co-workers (36). It is reasonable to assume that the micelle formation
produces a somewhat organized pattern of the metal centers and, due to
the shortened distance between the copper(II) containing head groups,
the coordination of catechol to two metal centers may increase the stabi-
lity of the catalyst^substrate complex. Perhaps, the same principles
apply to the interaction between the reduced form of the catalyst and O2,
and the reaction may also feature the formation of an intermediate with
a m-peroxo bridge between two copper centers. While these conclusions
require further con¢rmation, the e¡ect of added salt on the kinetics
could be rationalized on the basis of the above model. It was assumed
that the addition of NaCl at low concentration levels increased the oxida-
tion rate because it favored the formation of larger micelles with shorter
distances between the metal centers (50). The reaction rate goes through
a maximum and ¢nally levels o¡ as a function of [NaCl]. The somewhat
smaller rate at higher electrolyte concentration is probably due to the
saturation of the cationic head group layer of the micelles with anions
which obstruct the complexation with the catechols.
The catalytic activities of Cu(II), Co(II) and Mn(II) are considerably
enhanced by sodium dodecyl sulfate (SDS) in the autoxidation of
H2DTBC (51). The maximum catalytic activity was found in the CMC
region. It was assumed that the micelles incorporate the catalysts and
the short metal^metal distances increase the activity in accordance
with the kinetic model discussed above.The concentration of the micelles
increases at higher SDS concentrations. Thus, the concentrations of the
catalyst and the substrate decrease in the micellar region and, as a
consequence, the catalytic reaction becomes slower again.
In the presence of cobalt complexes as catalysts, the following
stoichiometry was reported for the autoxidation of H2DTBC in chloro-
form (52):
H2 DTBC þ 1=2O2 ! DTBQ þ H2 O ð66Þ
The yield of quinone was practically 100% and H2O2 accumulated in

negligible concentration.
A comparison of the initial rates obtained with various cobalt com-
plexes (Table I) reveals that the chelate complexes of Co(II) are more e⁄-
cient than the simple salts, the catalytic activity of Co(III) is lower than
that of Co(II) and the reaction becomes slower by increasing the
number of N atoms in the coordination spheres in both oxidation states.
In general, the addition of amine derivatives increased the activity of
the catalysts.
The reaction with CoII(ACAC)2 was studied in more detail and the rate
law was established.The reaction was found to be ¢rst-order with respect
to the substrate and the catalyst concentrations, and the partial pressure
of O2. The corresponding kinetic model postulates reversible formation
of a H2DTBC^CoII(ACAC)2Ô2 adduct which undergoes redox decompo-
sition in the rate-determining step. Hydrogen peroxide is also a primary
TABLE I
CATALYTIC ACTIVITY OF Co(II) AND Co(III) COMPLEXES IN AUTOXIDATION OF
a
DI-tert-BUTYLCATECHOL (52)
Catalyst v0 (M s1)
bis(acetylacetonato)cobalt(II) 1.6103
bis(benzoylacetonato)cobalt(II) 1.6103
bis(ethylbenzoylacetonato)cobalt(II) 1.8103
bis(N-cyclohexylsalicylaldiminato)cobalt(II) 1.2103
bis(N-phenylsalicylaldiminato)cobalt(II) 8.6104
cobaloxime(II)-bis(pyridine) 2.1104
tris(acetylacetonato)cobalt(III) 1.4104
chloro(pyridine)cobaloxime(III) b 0
cobalt(II) acetate 4.0105
cobalt(II) nitrate 0
cobalt(II) nitrate-pyridine c 2.8106
a
[H2DTBC]=100 mM, [catalyst]=1.0 mM, 27 C, in CHCl3.
b
[catalyst]=3.2 mM.
c
[PY]=250 mM.
product of this reaction, but it cannot accumulate because of fast

decomposition in a cobalt(II) catalyzed reaction:
H2 O2 ! H2 O þ 1=2O2 ð67Þ
The kinetic role of amines was interpreted by assuming that they are
capable of promoting the formation of the precursor complex by partially
deprotonating the substrate via hydrogen bonding with the OH group.
Semi-quinones are formed as reactive intermediates in the autoxida-
tion of catechols and they are present either bonded to a metal center or
as free radicals. Recently, Sima¤ndi and Sima¤ndi reported ESR spectra
for each type of these species in the triphenylphosphine-bis-(dimethyl-
glyoximato)cobalt(II)-catalyzed autoxidation of H2DTBC (catalyst =
[CoII(HDMG)2(Ph3P)]) (53). An 8 -line spectrum with g = 2.0017, aCo = 10.4
G and coupling constant of 3.3 G is characteristic of DTBSQ when it
is coordinated to the catalyst as a unidentate ligand (Fig. 2a). At higher
resolution, the 18 -line spectrum of the free radical was observed at
the center of the 8 -line signal with g = 2.0046 and a2H = 1.7 G, a18H 0.35 G
(Fig. 2b).
Detailed kinetic studies con¢rmed a two-stage reaction for the
cobaloxime(II)-catalyzed autoxidation of this system in methanol
(54,55). First, within about 30 s, the reaction reached steady-state
conditions via reversible oxygenation of Co(II) to the corresponding
FIG. 2. ESR spectra of free-radical intermediates during the cobaloxime(II) cata-

lyzed oxidation of 3,5 -CatH2 by O2 in benzene. Reprinted with permission from
Sima¤ndi, T. L.; Sima¤ndi, L. I. React. Kin. Catal. Let. 1998, 65, 301.
superoxo complex to which the substrate is bonded via a hydrogen

bond (for sake of simplicity, CoII and CoIII correspond to the respective
[CoII(HDMG)2(Ph3P)] and [CoIII(HDMG)2(Ph3P)]+ moieties in the follow-
ing equations):
CoII þ O2 )

*

III
Co O2 K68 ð68Þ
CoIII O2 þ H2 DTBC )
½CoIII O2 H2 DTBC
*
K69 ð69Þ
It was assumed that the hydrogen bond forms between the superoxo
moiety and one of the OH-groups of the substrate. The rate-determining
step is the redox decomposition of the hydrogen bonded adduct into
hydroperoxocobaloxime and semiquinone. The latter participates in an
equilibrium reaction with Co(II):
½CoIII O2 H2 DTBC ! CoIII O2 H þ DTBSQ þ Hþ k70 ð70Þ
CoII þ DTBSQ þ Hþ )
CoIII ðHDTBCÞ
*
K71 ð71Þ
III
The major cobaloxime species, Co (HDTBC), was present at an
approximately constant concentration level in the slow stage, in which
the following additional reaction steps were considered:
2CoIII O2 H ! 2CoIII OH þ O2 ð72Þ
CoIII OH þ DTBSQ þ Hþ ! CoII þ DTBQ þ H2 O ð73Þ
CoIII OH þ CoIII ðHDTBCÞ ! CoII þ CoIII DTBSQ þ H2 O ð74Þ

II
CoIII DTBSQ )

*
Co þ DTBQ ð75Þ
In agreement with the experimental results, this model predicts the
following expression for the initial rate of oxygen uptake after steady-
state is reached:
aK68 K69 ½O2 ½H2 DTBC0 ½Co0
vi ¼ k70 ð76Þ
1 þ K68 K69 ½O2 ½H2 DTBC0
where a is the ratio of free cobaloxime(II) compared to the total
amount of the catalyst. An analysis of the concentration dependencies
of the initial rate gave K68K69 = 6.9 103 and 5.1 104 M2 as well as
k70 = 7.0 102 and 3.8 102 s1 in benzene and methanol, respectively.
It should be noted that the dioxygen complex formed in Eq. (68) would
ultimately be converted into cobaloxime(III) which is catalytically inac-
tive. Thus, in order to maintain the catalytic cycle, the cobaloxime(II)
species needs to be regenerated relatively quickly.
A very similar kinetic model was proposed for the ferroxime(II)-cata-
lyzed autoxidation of H2DTBC in methanol (55). The catalyst was added
to the reaction mixture in the form of [Fe(HDMG)2(MeIM)2] (MeIM = N-
methylimidazole). Upon dissolution in methanol, this complex undergoes
solvolysis and one of the imidazole ligands is replaced by a solvent mole-
cule. The solvolysis occurs at the same rate under a N2 or O2 atmosphere.
Again, the reactions between the catalyst, substrate and dioxygen
leads to the formation of a hydrogen bonded precursor complex which
undergoes relatively slow redox decomposition into FeIIIO2H and
semiquinone. (FeII and FeIII represent the [Fe(HDMG)2(MeIM)] and
[Fe(HDMG)2(MeIM)]+complexes, respectively.) Subsequent fast reaction
steps are:
FeIII O2 H þ DTBSQ þ Hþ ! ½FeIV O þ DTBQ þ H2 O ð77Þ
½FeIV O þ H2 DTBC ! FeII þ DTBQ þ H2 O ð78Þ

The intense blue color of the reaction mixture was assigned to the
paramagnetic [Fe(HDMG)2(MeIM)(DTBSQ)]+ complex which is char-
acterized by a broad spectral band at lmax 680 nm and a distinct doublet
with g = 2.00425 and a1H = 3.135 G in the visible and ESR spectra, respec-
tively. This iron(II) species is not involved in a direct redox step and
acts only as a reservoir for the semi-quinone radical.
The sole product of the reaction is DTBQ with the dimethylglyoximato
complexes. The high chemoselectivity was rationalized by considering
that the cleavage of the substrate may occur when the semi-quinone
radical is directly attacked by O2 (presumably this is a slow reaction
path) or by a superoxide radical coordinated to a metal center. The most

e⁄cient path assumes adjacent coordination of the semi-quinone and
superoxide radicals. However, such a coordination mode is not feasible
in the presence of HDMG because of the very strong di-equatorial
coordination which resembles that of a macrocycle. Consequently, the
ring scission is negligible in these reactions if it occurs at all.
An Ir(III) catecholate complex [(TRIPHOS)Ir(DTBC)]+ was reported
to catalyze chemoselective oxidation of H2DTBC to DTBQ via the forma-
tion of a dioxygen adduct [(triphos)Ir(m-O2)(DTBSQ)]+ (TRIPHOS =
CH3C(CH2PPh2)3) (56). This oxygen complex was prepared by bubbling
O2 through a CH2Cl2 solution of the catecholate complex and charac-
terized by using UV/Vis, IR, 1H and 31P NMR spectroscopies, as well as
X-ray analysis (57).
An interesting feature of the reaction is that it produces exclusively
DTBQ and H2O2 with the Ir(III) catalyst (Scheme 4). Deviations from
this pattern were reported only at relatively high partial pressures of
O2. In contrast, other organic products also formed when the rhodium
analogue of the catecholate complex is used as a catalyst, although the
Ir and Rh complexes were reported to possess the same structure
(58,59). This clearly indicates that factors other than the geometry and
the overall charge of the catalyst may have an important role in the
mechanism.
The rate of the Ir(III) catalyzed reaction was found to be ¢rst-order in
[Ir] and [H2DTBC], but independent of O2 concentration in chloroform
(56). The mechanism proposed for the reaction (Scheme 4) postulates
that the protonation of the hydroperoxo a-oxygen by the hydroxy group
of the bonded catechol in Int 1 leads to the formation of H2O2. The o-qui-
none ligand of Int 2 is replaced by the partially coordinated catechol in
the next step. In order to comply with the experimental rate law, the
rate-determining step needs to be the reaction of the oxygen adduct (B)
with catechol.
The di¡erences between the iridium and rhodium systems were inter-
preted by considering that Int 1 may open an oxygenation path via intra-
molecular rearrangement. In this case, the scission of a MÔ bond
between the semi-quinone and metal center would be followed by intra-
diol insertion of an oxygen and ultimately by the formation of muconic
acid and water. The results indicate partial preference of the rhodium
complex toward the oxygenation path.
The cleavage of catechols with the incorporation of oxygen is clearly
favored in the presence of some of the iron(III) complexes as catalysts.
Que and co-workers proposed a substrate activation mechanism for
these reactions, wherein the delocalization of the unpaired spin density
SCHEME 4. Reprinted with permission from Barbaro, P.; Bianchini, C.; Frediani, P.;
Meli, A.; Vizza, F. Inorg. Chem. 1992, 31, 1523. Copyright (2002) American Chemical
Society.
from the metal center onto the catecholate in the corresponding complex
makes the substrate more susceptible to dioxygen attack (60). The non-
participating ligand was shown to have a profound e¡ect on the Lewis
acidity of the metal center and, as a consequence, on the ligand to metal
charge transfer. This e¡ect also manifests itself in the kinetics and
product distribution of the autoxidation reactions (28). It was shown
that oxidative cleavage of catechols becomes preferred over extradiol
oxidation to quinone when the Lewis acidity of the ferric center is
increased.
In a systematic study, the reaction of [FeL(DTBC)] with dioxygen was
studied in various non-aqueous solvents (29). In the ternary catecholate
complex, L represents the following (RCH2)2N(CH2R0 ) type tripodal

ligands: NTA, R = R0 = COO; PDA, R = COO, R0 = 2 -pyridyl; BPG, R =
2 -pyridyl, R0 = COO; HDP, R = 2 -pyridyl, R0 = 2 -hydroxy-3,5 -dimethyl-
phenyl. These ligands induce relatively pronounced Lewis acidity on
the metal center in the following order: BPG>PDA>NTA>HDP. It
should be noted that the isotropic shift of the 1H NMR signal of the
coordinated catecholate was considered to be the best spectroscopic
parameter as indicator of the Lewis acidity of the ferric center. The
oxidation reaction yielded almost exclusively the intradiol scission
product in all cases (Table II).
The kinetic experiments were not performed under true catalytic con-
ditions, i.e. the pre-prepared [FeL(DTBC)] complexes were introduced
into the reaction mixtures as reactants and excess substrate was not
used. Nevertheless, the results are important in exploring the intimate
details of the activation mechanisms of the metal ion catalyzed autoxida-
tion reactions of catechols. In excess oxygen the reaction was ¢rst-order
in the complex concentration and the ¢rst-order dependence in
dioxygen concentration was also con¢rmed with the BPG complex. As
shown in Table II, the rate constants clearly correlate with the Lewis
character of the complex, i.e. the rate of the oxidation reaction increases
by increasing the Lewis acidity of the metal center.
The scission product from the diamagnetic [Ga(BPG)(DTBC)] complex
is considerably less (4%) than that from the corresponding Fe complex
(97%) (29). Gallium(III) is a diamagnetic d10 ion with an ionic radius
very similar to that of iron(III) and expected to possess similar Lewis
properties to that of Fe(III). Thus, the results clearly demonstrate that
the ability of the metal center to transfer paramagnetic spin density to
TABLE II
KINETIC PARAMETERS FOR THE REACTIONS OF [FeL(DTBC)] COMPLEXES WITH
DIOXYGEN IN DIFFERENT SOLVENTS (29) a
Ligand % 102k 102k 102k Hz b Sz b

Scission b,c (DMF) (CH3OH) (CH2Cl2) (kJ/mol) J/(mol K)
HDP 91 0.33 0.49 0.81

NTA 84 3.7 1.0 53.5 92
PDA 95 4.3 5.0 3.7 47.7 113
BPG 97 18.3 42.6 36.6 42.6 117
a
Rate constants are in M1 s1 at 25 C.
b
In DMF.
c
Under 1 bar O2 in the presence of excess piperidine.
SCHEME 5. Reprinted with permission from Cox, D. D.; Que, L. Jr. J. Am. Chem. Soc.
1988, 110, 8085. Copyright (2002) American Chemical Society.
the substrate is, besides the Lewis acidity, an essential requirement for
the scission reaction.
The mechanism shown in Scheme 5 postulates the formation of a
Fe(II)-semi-quinone intermediate. The attack of O2 on the substrate gen-
erates a peroxy radical which is reduced by the Fe(II) center to produce
the Fe(III) peroxide complex. The semi-quinone character of the
[FeL(DTBC)] complexes is clearly determined by the covalency of the
iron(III)^catechol bond which is enhanced by increasing the Lewis acid-
ity of the metal center.Thus, ultimately the non-participating ligand con-
trols the extent of the Fe(II) ^ semi-quinone formation and the rate of
the reaction provided that the rate-determining step is the reaction of
O2 with the semiquinone intermediate. In the ¢nal stage, the substrate
is oxygenated simultaneously with the release of the FeIIIL complex. An
alternative model, in which O2 attacks the Fe(II) center instead of the
semi-quinone, cannot be excluded either.
According to a recent study with iron(III) complexes of tripodal
ligands, systematic variation of one ligand arm strongly a¡ects the
steric shielding of the iron(III) center and the bonding of catechol sub-
strates (61). It was shown that the dioxygenation reactions of catechols
depend on the redox potentials of the complex and substrate as well as

the steric properties of the tripodal ligand. As the redox potential of
the catechol was increased the oxidation became slower. It was con-
cluded that the ideal non-participating tripodal ligand needs to induce
high Lewis acidity on the iron(III) center and to consist of non-bulky
coordinating groups.
The oxidation of adrenaline (H2C^RH+: R ¼ ”CHðOHÞCH2 NHþ 2 CH3 )
to andrenochrome is e⁄ciently catalyzed by VO2+ in aqueous solution
(62). The following experimental rate law was reported for the oxygen
consumption ( = 0.1 M KNO3, 25 C):

d½O2 A½VOT ½LT E
¼ þ þ F ½O2 þ p ð79Þ
dt B½Hþ 2 þ C½LT þ D½LT ½Hþ
where A, B, C, D, F, p are experimentally determined parameters which
could be expressed by combining characteristic equilibrium and rate
constants for this system; [VO]T and [L]T are the total concentrations
of the catalyst and substrate, respectively.
The experimental observations were interpreted by assuming that the
redox cycle starts with the formation of a complex between the catalyst
and the substrate. This species undergoes intramolecular two-electron
transfer and produces vanadium(II) and the quinone form of adrenaline.
The organic intermediate rearranges into leucoadrenochrome which is
oxidized to the ¢nal product also in a two-electron redox step. The +2
oxidation state of vanadium is stabilized by complex formation with the
substrate. Subsequent reactions include the autoxidation of the V(II)
complex to the product as well as the formation of a VOV4+ intermediate
which is reoxidized to VO2+ by dioxygen. These reactions also produce
H2O2. The model also takes into account the rapidly established equi-
libria between di¡erent vanadium-substrate complexes which react
with O2 at di¡erent rates. The concentration and pH dependencies of
the reaction rate provided evidence for the formation of a VðC”RHÞþ 3
complex in which the formal oxidation state of vanadium is +4.
V. Autoxidation of Cysteine
Earlier studies demonstrated a rich variety of oxidation states,

geometries and compositions of the intermediates and products formed
in the autoxidation reactions of cysteine (RSH). Owing to the complexity
of these systems, only a limited number of detailed kinetic papers were
published on this subject and, not surprisingly, some of the results are
contradictory. The majority of available data was reported on copper

catalysis because of the signi¢cance of copper^sulfur interactions in
enzymatic electron transfer, oxygen transport and oxygenation
reactions (63).
Zwart and co-workers con¢rmed that the catalytic autoxidation
produces cystine (RSSR) with 100% selectivity and the actual stoichio-
metry can be given by the following equation in alkaline solution
([NaOH] = 0.25 M) (64):
4RS þ ð1 þ aÞO2 þ 2H2 O ! 2RSSR þ 2ð2 aÞOH þ 2aHO
2 ð80Þ
which is the linear combination of Eqs. (81) and (82)
4RS þ O2 þ 2H2 O ! 2RSSR þ 4OH ð81Þ
4RS þ 2O2 þ 2H2 O ! 2RSSR þ 2OH þ 2HO

2 ð82Þ
The stoichiometry strongly suggests that HO2 (the basic form of H2O2)
also oxidizes the substrate.
As shown in Fig. 3, the kinetic traces feature a speci¢c break-point
when cysteine is oxidized by excess H2O2 under anaerobic conditions,
and in the presence of Cu(II). The break-point occurs when all of the
FIG. 3. Decay of the H2O2 concentration versus time during the anaerobic oxidation
reaction with cysteine in the presence of CuSO4. First stage of constant rate (first-
order in [Cu]) during the period of oxidation, second stage of increasing rate
after completion of the oxidation of cysteine to cystine. Reprinted from Journal
of Molecular catalysis, vol. 11, Zwart, J.; van Wolput, J. H. M. C.; van der Cammen,
J. C. J. M.; Koningsberger, D. C. Accumulation and Reactions of H2O2 During the
Copper Ion Catalyzed Autoxidation of Cysteine in Alkaline Medium, p. 69,
Copyright (2002), with permission from Elsevier Science.
substrate is oxidized. In this part of the reaction, the rate of hydrogen

peroxide consumption is practically constant, independent of [H2O2]
and [RSH] but shows ¢rst-order dependence on the concentration of the
catalyst. Once cysteine is used up, the reaction starts to generate
oxygen in the copper catalyzed decomposition of H2O2.
A coherent qualitative interpretation can be given for the observations
by assuming that the primary product of the catalytic RSHÔ2 reaction
is H2O2 which also oxidizes the substrate. The latter reaction needs to
be considerably slower than the ¢rst one because accumulation of signi¢-
cant amounts of H2O2 was observed under aerobic conditions. In the
absence of added oxygen, catalytic decomposition of H2O2 also needs to
be considered. The product of this reaction is O2 which in turn regener-
ates hydrogen peroxide via the oxidation of the substrate.
Quantitative ESR measurements con¢rmed that almost all of the total
quantity of copper is present as [Cu(RS)] complex during the reaction
(65). The kinetic data were consistent with a rate law which is zeroth-
order in cysteine concentration:
d½O2 =dt ¼ ka83 ½O2 1=2 ½CuðIIÞ þ kb83 ½O2 1=2 ½CuðIIÞ2 ð83Þ
ka83 ¼ 2:2 M1=2 s1 ; kb83 ¼ 1:2 104 M3=2 s1
The half-order of the rate with respect to [O2] and the two-term
rate law were taken as evidence for a chain mechanism which involves
one-electron transfer steps and proceeds via two di¡erent reaction
paths. The formation of the dimer [(RS)2Cu(m-O2)Cu(RS)2] complex in
the initiation phase is the core of the model, as asymmetric dissociation
of this species produces two chain carriers. Earlier literature results
were contested by rejecting the feasibility of a free-radical mechanism
which would imply a redox shuttle between Cu(II) and Cu(I). It was
assumed that the substrate remains bonded to the metal center through-
out the whole process and the free thiyl radical, RS, does not form
during the reaction. It was argued that if free RS radicals formed they
would certainly be involved in an almost di¡usion-controlled reaction
with dioxygen, and the intermediate peroxo species would open alterna-
tive reaction paths to generate products other than cystine. This would
clearly contradict the noted high selectivity of the autoxidation reaction.
Kinetic studies in the physiological pH range (6.5 to 7.8) provided con-
sistency with the above results in that the accumulation of H2O2 was
also observed and the stoichiometry of the reaction depended on the con-
ditions applied (66^69). However, a simple 2 : 1 stoichiometry was con-
¢rmed between cysteine consumed and hydrogen peroxide formed in
dilute solution. The reaction followed Michaelis^Menten kinetics with
respect to [O2] which is in clear contradiction with the half-order [O2]

dependence in alkaline solution. The results were interpreted by assum-
ing the formation and re-oxidation of Cu(I) in the catalytic cycle,
though the mechanism was not explained in detail.
Ehrenberg and co-workers found Michaelis-Menten type kinetics not
only with respect to [O2] but also with respect to the substrate concentra-
tion at pH 7.2 (67).They pointed out that the independence of the reaction
rate upon the substrate concentration at high [RSH] and low [Cu(II)] is
consistent with zeroth-order kinetics in [RSH]. The reported kinetic
trace under such conditions features a slight induction period, but this
observation was not addressed in that paper. For comparison, the cata-
lytic e¡ect of iron(III) was also studied. It was concluded that the cata-
lytic activities of the two metal ions are fundamentally di¡erent. Copper
e⁄ciently enhanced the rate of H2O2 formation but practically had no
e¡ect on further reduction of hydrogen peroxide. In contrast, H2O2 was
not detected in the presence of iron(III). The reaction was ¢rst-order in
[RSH] with freshly prepared catalyst solution; however, conditions favor-
ing the formation of colloidal Fe(III) hydroxide(s) shifted the kinetics
toward zeroth-order. This observation seems to be consistent with a
mechanistic changeover to surface catalysis in the presence of Fe(III)sol.
Interestingly, the catalytic e¡ects of the two metal ions were not addi-
tive; the copper(II) catalyzed reaction became signi¢cantly slower on
the addition of Fe(III) to the reaction mixture.This observation was inter-
preted in terms of the model proposed by Zwart and co-workers (64) by
assuming that one of the two copper centers is replaced by Fe in the
catalytically active dimer (67).
The e¡ect of non-participating ligands on the copper catalyzed
autoxidation of cysteine was studied in the presence of glycylglycine-
phosphate and catecholamines, (2 -R^)H2C, (epinephrine, R = CH(OH)^
CH2^NHCH3; norepinephrine, R = CH(OH)^CH2^NH2; dopamine, R =
CH2^CH2^NH2; dopa, R = CH2^CH(COOH)^NH2) by Hanaki and co-wor-
kers (68,69). Typically, these reactions followed Michaelis-Menten
kinetics and the autoxidation rate displayed a bell-shaped curve as a
function of pH. The catecholamines had no kinetic e¡ects under anaero-
bic conditions, but catalyzed the autoxidation of cysteine in the following
order of e⁄ciency: epinephrine = norepinephrine
dopamine > dopa.
The concentration and pH dependencies of the reaction rate were inter-
preted by assuming that the redox active species is the [L^CuII(RS)]
ternary complex which is formed in a very fast reaction between CuIIL
and cysteine. Thus, the autoxidation occurs at maximum rate when the
conditions are optimal for the formation of this species. At relatively
low pH, the ternary complex does not form in su⁄cient concentration.
Upon increasing the pH, ¢rst this species is present in increasing con-
centrations, but eventually the non-participating ligand is replaced
with another cysteine molecule and the redox inactive Cu(RS)2 is
formed. Consequently, the reaction rate decreases sharply. The following
reaction sequence was proposed for the redox reaction:
L”CuII ðRS Þ ! L”CuI þ RS ð84Þ
I
L”CuI þ RS )

*

Cu ðRS Þ þ L ð85Þ
I
CuI ðRS Þ þ O2 )

*

Cu ðRS Þ O2 ð86Þ
CuI ðRS Þ”O2 þ RS þ 2Hþ ! CuII ðRS Þ þ RS þ H2 O2 ð87Þ

In the presence of the catecholamines the following two steps also need
to be considered:
CuI ðRS Þ”O2 þ ð2-R”ÞHC þ 2Hþ
! CuII ðRS Þ þ ð2-R”ÞHSQ þ H2 O2 ð88Þ
RS þ ð2-R”ÞHSQ ! RS þ ð2-R”ÞHC ð89Þ

The product cystine is presumably formed in the recombination of two
thiyl radicals. This free-radical model is suitable for formal treatment of
the kinetic data; however, it does not account for all possible reactions
of the RS radical (68).The rate constants for the reactions of this species
with RS, O2 and CuILn (n = 2, 3) are comparable, and on the order of
109^1010 M1 s1 (70^72). Because all of these reaction partners are pre-
sent in relatively high and competitive concentrations, the recombina-
tion of the thiyl radical must be a relatively minor reaction compared to
the other reaction paths even though it has a di¡usion controlled rate
constant. It follows that the RS radical is most likely involved in a
series of side reactions producing various intermediates. In order to
comply with the noted chemoselectivity, at some point these transient
species should produce a common intermediate leading to the formation
of cystine.
In a recent study, spectral evidence was found for the formation of a
[RS^CuI^RS] dimer when cysteine was added to Cu(II) under anaerobic
conditions and at pH 7.4 (73). On exposure to air, the dimer slowly
decayed after the cysteine excess was consumed. This may indicate that
indeed all the reaction proceeds via an intra-molecular mechanism.
Kachur and co-workers found evidence for a two-stage autoxidation
reaction in an excess of dioxygen (74). In the ¢rst stage, the formation of
hydroxyl radicals could not be detected and 0.25 mole of O2 was con-
sumed by one mole of cysteine in agreement with the stoichiometry
shown in Eq. (81). Further oxygen consumption occurred in the second
stage of the reaction in a hydrogen peroxide-mediated Fenton-type pro-
cess with sulfonate as product. Signi¢cant O/OH production was also
observed. Copper was proposed to catalyze the reaction via the formation
of a cysteine complex and the corresponding mechanism postulates
simultaneous superoxide- and peroxide-dependent paths for the ¢rst
stage and a peroxide-dependent path for the second.
Copper(II) retains its catalytic activity when attached to a polyethyle-
nepolyamine base.The synthetic conditions of the sorbents have a signi¢-
cant e¡ect on the catalytic reaction which most likely is related to a
change in the coordination sphere around the metal center (75).
Metal ion catalyzed autoxidation reactions of glutathione were found
to be very similar to that of cysteine (76,77). In a systematic study, cata-
lytic activity was found with Cu(II), Fe(II) and to a much lesser extent
with Cu(I) and Ni(I). The reaction produces hydrogen peroxide, the
amount of which strongly depends on the presence of various chelating
molecules. It was noted that the catalysis requires some sort of complex
formation between the catalyst and substrate. The formation of a radical
intermediate was not ruled out, but a radical initiated chain mechanism
was not necessary for the interpretation of the results (76).
VI. Autoxidation of Sulfur(IV)
A review by Brandt and van Eldik provides insight into the basic
kinetic features and mechanistic details of transition metal-catalyzed
autoxidation reactions of sulfur(IV) species on the basis of literature
data reported up to the early 1990s (78). Earlier results con¢rmed that
these reactions may occur via non-radical, radical and combinations of
non-radical and radical mechanisms. More recent studies have shown
evidence mainly for the radical mechanisms, although a non-radical,
two-electron decomposition was reported for the HgSO3 complex
recently (79). The possiblity of various redox paths combined with
protolytic and complex-formation reactions are the sources of manifest
complexity in the kinetic characteristics of these systems. Nevertheless,
the predominant sulfur containing product is always the sulfate ion.
In spite of extensive studies on this topic for well over a century, impor-
tant aspects of the mechanisms remain to be clari¢ed and the interpreta-
tion of some of the reactions is still controversial. Recent studies were
FIG. 4. Absorbance-time traces for the iron(III) catalyzed autoxidation of sulfur(IV)

oxides: (a) [O 2] = 0 m; (b) [O 2] = 7.5104 M. Experimental conditions: [Fe(III)] =
5.0 10 6 M; [S(IV)] = 5.0 103 M; ionic strength = 0.5 M; T = 25 C; pH = 2.5; =
390 nm; absorbance scale is inV (10 V = 1 absorbance unit). Reprinted with permission
from Brandt, C.; Fa¤bia¤n, I.; van Eldik, R. Inorg. Chem. 1994, 33, 687. Copyright (2002)
American Chemical Society.
designed to improve previous kinetic models and to understand

how speci¢c conditions and environmental factors a¡ect the overall
reactions.
Due to the environmental and industrial signi¢cance of the reaction,
the most thorough studies were reported on the iron(III) catalyzed oxida-
tion of sulfur(IV). As shown in Fig. 4, the kinetic traces are distinctively
di¡erent in the presence and absence of dioxygen.
Brandt and co-workers proposed a detailed mechanism for this reac-
tion which was validated using kinetic modeling and the most viable
values of the rate constants were estimated on the basis of sensitivity
analysis (80). In this model, the absorbance increase observed at the
beginning of the kinetic traces was assigned to the formation of a

generalized iron(III) monosul¢to complex which undergoes redox
decomposition in the rate determining step:
FeIII SIV ! FeðIIÞ þ SO
3 ð90Þ
Fast reaction of the sul¢te ion radical with a second Fe(III) concludes
the overall reaction and yields the main ¢nal products which are Fe(II),
SO2 2
4 . Small amounts of S2 O6 are also formed in the reaction. In the pre-
sence of dioxygen, a characteristic break point (at time tbp) was observed
in the kinetic traces indicating total consumption of O2. After the break
point, the absorbance decay was the same as in the absence of oxygen.
An increase in [O2] increased tbp while variation of [Fe(III)] had the oppo-
site e¡ect. The formation and decay of the iron(III)-sul¢to complex
became slower when iron(II) was added initially to the reaction mixture,
but tbp was only slightly a¡ected. Hydroquinone, a well known radical
scavenger, had no e¡ect on the decomposition rate of the sul¢to complex
under an argon atmosphere, although it inhibited the oxygen-induced
step.This is a clear indication that a free-radical mechanism is operative
in the autoxidation stage.
In the presence of oxygen, SO 3 generates the peroxomonosulfate
anion radical (Eq. (91)) in a reaction step with a rate constant close
to the di¡usion controlled limiting value on the order of 1.0 109 to
2.5 109 M1 s1 (81,82):
SO
3 þ O2 ! SO5 ð91Þ
The SO
5 radical, being a reasonably powerful and stronger oxidant
than dioxygen ("0 = 1.24 V for the SO
5 =HSO5 couple at pH 2.0 (83)), has
a central role in the mechanism, because it can easily oxidize either sul-
¢te ion or Fe(II). Reoxidation of Fe(II) to Fe(III) is essential to establish
the catalytic redox cycle, and parallel reactions of SO 5 with HSO3

develop a branch of the reaction chain via the formation of additional

reactive intermediates such as SO
4 and HSO5 . The possible reactions
of various sulfur oxyanions and oxyanion radicals in the uncatalyzed
autoxidation of HSO 3 were discussed by Connick and co-workers in
detail (84), and the rate constants for some of them were reported by
Das recently (85). These reactions occur after the rate-determining step,
i.e. Eq. (90), of the catalytic cycle, and only the ratios of their rates is rele-
vant with respect to the kinetic model, because the product distribution
is determined by the relative kinetic weight of the competing reaction
paths. The formation of S2 O2 6 at relatively low concentration levels com-
pared to SO2
4 con¢rms that the recombination of the SO3 radical is of
marginal kinetic importance in the catalytic reaction. Warneck and
Ziajka proposed essentially the same model on the basis of kinetic stud-
ies in which benzene was used as a radical scavenger (86). They used a
steady-state approach for evaluation of the data and reported the follow-
ing ratios of the second-order rate constants for the reactions of a given
intermediate with di¡erent reactants: SO
4 reacting with HSO3 and
2+
benzene: 0.31; SO5 reacting with Fe and HSO3 : 54.1; HSO5 reacting
with Fe2+ and HSO
3 : 2.2 and the reaction of SO5 with HSO3 leading to

SO3 and SO4 : 0.04. Kinetic studies under hydrometallurgical condi-
tions at 80 C were also consistent with the considerations discussed
above (87).
The model proposed by Brandt et al. is consistent with the experimen-
tal observations, reproduces the peculiar shape of the kinetic curves in
the absence and presence of dioxygen reasonably well, and predicts the
same trends in the concentration dependencies of tbp that were observed
experimentally (80). It was concluded that there is no need to assume
the participation of oxo-complexes in the mechanism as it has been
proposed in the literature (88^90). However, the model provides only a
semi-quantitative description of the reaction because it was developed
at constant pH by neglecting the acid-base equilibria of the sul¢te ion
and the reactive intermediates, as well as the possible complex-formation
equilibria between various iron(III) species. In spite of the obvious
constraints introduced by the simpli¢cations, the results shed light on
the general mechanistic features of the reaction and could be used to
identify the main tasks for further model development.
The most controversial issue is the number and exact stoichiometries
of the iron(III)-sul¢to complexes formed under di¡erent experimental
conditions. Earlier, van Eldik and co-workers reported the formation
of a series of [Fe(SO3)n]32n (n = 1 to 3) complexes and the [Fe(SO3)(OH)]
complex (89,91,92). The stability constants of these species were deter-
mined by evaluating time resolved rapid-scan spectra obtained from the
sub-second to several minutes time domain. The cis^trans isomerization
of the complexes was also considered, under feasible circumstances. In
contrast, Betterton interpreted his results assuming the formation and
linkage isomerization of a single complex, [Fe(SO3)]+ (93). In agreement
with the latter results, Conklin and Ho¡mann also found evidence
only for the formation of a mono-complex (94). However, their results
were criticized on the basis that the experiments were made in 1.0 M
formic acid/formate bu¡er where iron(III) existed mainly as formato
complex(es). Although these reactions could interfere with the formation
of the sul¢to complex, they were not considered in the evaluation of
the results (95). Finally, van Eldik and co-workers re-examined the com-
plex-formation reactions and presented additional data in support of
their earlier results (96^98). The inherent di⁄culty of obtaining reliable

experimental evidence for the complexes formed is the strong overlap
between concurrent complex-formation reactions and kinetic coupling
between the equilibrium and redox reactions. Nevertheless, the complex-
ity of the spectroscopic data seems to be consistent with the existence
of more than one sul¢to complex.
A kinetic study was designed by Lente and Fa¤bia¤n to explore the
properties of the mono-sul¢to complex in the presence of a substantial
metal ion excess over sul¢te ion concentration. The observations
indicated surprisingly complex, multi-stage kinetic behavior (95). In
order to separate the overlapping kinetic e¡ects and impede the
complex-formation steps, the initial part of the reaction was studied at
10 C where the overall process lasted for over 20 minutes. Typical
stopped-£ow traces showed an absorbance increase within the
dead-time of the instrument ( 1 ms) indicating an instantaneous reac-
tion. It could be shown that this absorbance change corresponds to the
formation of the [Fe(SO3)]+ complex which is much faster than expected
on the basis of the rates of other ligand substitution reactions of
iron(III). This may indicate that the complex-formation occurs via an
oxygen exchange mechanism instead of the usual ligand-exchange
path. Such a mechanism was proposed earlier for the formation of other
sul¢to complexes, (99,100) and supported by the extremely fast complex
formation between sul¢te ion and the otherwise substitution inert
Cr(III) (101,102).
Simple ¢rst-order kinetics were observed in the next, well-separated
stage which was complete within 200 ms. The main absorbing species at
340 nm is [Fe2(OH)2]4+ and the absorbance decay at this wavelength
indicated that the hydroxo dimer is directly involved in the fast process.
An analysis of the concentration dependence of the absorbance change
con¢rmed the formation of a unique dinuclear complex presumably
with the following composition: [(H2O)4Fe(m-OH)(m-SO3)Fe(H2O)4]4+. In
follow-up studies the formation of this species was modeled without the
interference of redox steps, and the equilibrium and kinetic features of
a series of novel di- and tetranuclear iron(III) complexes with simple
inorganic oxoanions were described (103^105). The results seem to con-
¢rm that the formation of a dinuclear complex from [Fe2(OH)2]4+ is
always feasible when the O^XÔ moiety of the entering ligand has the
appropriate geometry to replace one of the OH bridges of the hydroxo
dimer. These complexes should be regarded as reactive intermediates as
they dissociate into the corresponding mono-complexes at a rate compa-
rable with that of the reactions between the monomer and dimeric
hydroxo complexes. The combination of these reactions typically yields
composite kinetic traces, and the pH dependencies of the individual

reaction steps complicate the situation even further.
The kinetic models developed for the formation of dinuclear iron(III)
complexes served as a basis for the evaluation of the sul¢te ion reaction.
Matrix rank analysis of the time-resolved spectra did not indicate the
formation of additional absorbing species at longer reaction times, and
the common complex-formation model was modi¢ed by considering
redox reactions only with the monosul¢to and the dinuclear complex
(106). A series of model calculations assuming di¡erent redox paths
revealed that the dinuclear species is not involved in direct inner- or
outer-sphere electron transfer reactions. Finally, the kinetic model
given in Eqs. (92)^(101) was veri¢ed by simultaneous evaluation of kinetic
traces recorded at 340 and 430 nm under di¡erent experimental condi-
tions.Thus, the equilibrium and rate constants were estimated by ¢tting
the kinetic model directly to about 105 time-dependent absorbance
values. These parameters are given at 25 C and = 1.0 M NaClO4 in the
following equations:
Fe3þ )

*

2þ
FeðOHÞ þ H
þ
K92 ¼ 1:91 103 M; fast ð92Þ
Fe3þ þ HSO
3 ) FeðSO3 Þþ þ Hþ
* K93 ¼ 1:35; fast ð93Þ
þ

H2 O SO2 ) *
HSO3 þ H K94 ¼ 1:82 102 M; fast ð94Þ
2
HSO
3 ) SO3 þ Hþ
* K95 ¼ 4:57 107 M; fast ð95Þ
Fe2 ðOHÞ4þ
2 þHSO3 )
*
Fe2 ðOHÞðSO3 Þ
3þ
K96 ¼ 5:37 103 M
k96 ¼ 3:9 105 M1 s1 ð96Þ
2
Fe2 ðOHÞ4þ Fe2 ðOHÞðSO3 Þ3þ k97 4 109 M1 s1
þ *
2 þSO3 ðþH Þ) ð97Þ
Fe2 ðOHÞ4þ
2

) *
2FeðOHÞ
2þ ka98 ¼ 0:35 M1 s1
kb98 ¼ 3:5 s1 ð98Þ
kc98 ¼ 3:6 10 3
Ms 1
Fe2 ðOHÞðSO3 Þ3þ )

*

2þ
FeðOHÞ þFeðSO3 Þ
þ
k99 ¼ 3:6 s1 ð99Þ
FeðSO3 Þþ ! Fe2þ þ SO

3 k100 ¼ 0:19 s1 ð100Þ
Fe3þ þ SO
3 þ H2 O ! Fe
2þ
þ SO2
4 þ 2H
þ
fast ð101Þ
The three rate constants for Eq. (98) correspond to the acid-catalyzed,
the acid-independent and the hydrolytic paths of the dimer-monomer
equilibrium, respectively, and were evaluated independently (107). The
results clearly demonstrate that the complexity of the kinetic processes
is due to the interplay of the hydrolytic and the complex-formation
steps and is not a consequence of electron transfer reactions. In fact,
the ¢rst-order decomposition of the FeSOþ 3 complex is the only redox
step which contributes to the overall kinetic pro¢les, because subsequent
reactions with the sul¢te ion radical and other intermediates are consid-
erably faster. The presence of dioxygen did not a¡ect the kinetic traces
when a large excess of the metal ion is present, con¢rming that either
the formation of the SO 5 radical (Eq. (91)) is suppressed by reaction
(101), or the reactions of Fe(II) with SO
5 and HSO5 are preferred over

those of HSO3 as was predicted by Warneck and Ziajka (86). Recently,
¢rst-order formation of iron(II) was con¢rmed in this system (108),
which supports the ¢rst possibility cited, though the other alternative
can also be feasible under certain circumstances.
Although the above model was developed under non-catalytic
conditions, some of the results may bear signi¢cance under natural con-
ditions or in the presence of excess sul¢te ions. Thus, the decomposition
of the mono-sul¢to complex was considered to be the rate-determining
step in the catalytic cycle, but only estimates could be given for the rate
constant in earlier studies. The comprehensive data treatment used by
Lente and Fa¤bia¤n yielded a well established value for this parameter
(106), which can then be used to improve previous kinetic models.
Furthermore, the participation of reactions of the [Fe2(OH)(SO3)]3+ com-
plex was never considered in kinetic studies where excess sul¢te ion
was used over low iron(III) concentration in mildly acidic solution (pH
2.5^3.0). The above model predicts that in some cases the formation of
the dimeric sul¢to complex could make a substantial contribution to
the spectral changes and omission of this species could lead to biased
conclusions. Reevaluation of data sets reported earlier by including the
reactions of [Fe2(OH)(SO3)]3+ may resolve some of the controversies
found in literature results.
Recent studies demonstrated that the composition of the reaction
mixture, and in particular the pH have signi¢cant e¡ects on the kinetics
of iron(III)-catalyzed autoxidation of sulfur(IV) oxides. When the reac-
tion was triggered at pH 6.1, the typical pH pro¢le as a function of time
exhibited a distinct induction period after which the pH sharply
decreased (98).The S-shaped kinetic traces were interpreted by assuming
2
that the bu¡er capacity of the HSO 3 = SO3 system e⁄ciently reduces
the acidifying e¡ect of the oxidation process. The activity of the
catalyst was signi¢cantly reduced by increasing the pH or aging the

reagent iron(III) solutions. These observations were rationalized by
considering the formation of polymeric iron(III) hydroxo species which
are catalytically less active than the aqua or the monohydroxo iron(III)
complex.
In order to model chemical features of aerosols, the e¡ect of oxalate
ion on the iron(III) catalyzed autoxidation of sulfur(IV) was reported
in related papers by Grgič et al. (109,110) and Wolf et al. (111). Oxalate
strongly inhibits the catalytic reaction, and when the catalyst was
added as Fe(II) the kinetic traces showed autocatalysis with an induction
period that increased upon increasing the oxalate concentrations. The
observations were interpreted on the basis of the general kinetic model
proposed for iron(III) catalysis. It was assumed that the inhibiting
e¡ect is due to the relatively low free iron(III) concentrations arising
from its consumption in the formation of complexes with the organic
component. According to Grgič and co-workers, the induction period in
the presence of Fe(II) is required to reach an equilibrium between
Fe(III)-oxalato complexes and free Fe(III) (110).
Manganese catalyzed autoxidation of sulfur(IV) has been studied
in di¡erent pH regions and under a large variety of experimental
conditions. The results show spectacular diversity with respect to the
rate equations which may contain one, two or even three terms.
Reaction orders of sul¢te ion and manganese concentrations that vary
between 0 and 2 and from less than 1 to 2, respectively, have been
reported. At the same time, there seems to be agreement in that the
rate of the catalytic cycle is independent of the oxygen concentration
and most of the studies support a free-radical chain mechanism.
Recent studies clari¢ed important details of manganese catalysis, but
veri¢cation of a generally applicable mechanism still appears to be a
remote goal (112^115).
Connick and Zhang reported a three-term rate law for the oxygen
consumption at constant pH 4.5, 25 C and = 0.05 M (112):
d½O2 =dt ¼ ka102 ½HSO 2 b c

3 þ k102 ½HSO3 ½MnðIIÞ þ k102 ½MnðIIÞ
2
ð102Þ
with ka102 ¼ 3:6 103 M1 s1 , kb102 ¼ 1:23 M1 s1 ; kc102 ¼ 98:6 M1 s1 .
The ¢rst term was assigned to the uncatalyzed path which includes
the formation of the same reactive radical and non-radical intermediates
that were reported in related studies. A clear distinction from other
kinetic models is that accumulation of the disulfate radical was also
postulated during the autoxidation phase (84).The continued production
of H+ after the total consumption of dioxygen was assigned to the rela-

tively slow decomposition of this species with k104 0.013 s1:
2
HSO
5 þ HSO3 ! S2 O7 þ H2 O ð103Þ
2
S2 O2
7 þ H2 O ! 2SO4 þ 2H
þ
ð104Þ
The obvious challenge in the interpretation of the data is to ¢nd a suit-

able explanation for the independence of the third term of the rate law,
Eq. (102), on the concentrations of HSO 3 and O2. The rate expression
determined experimentally could be modeled quantitatively by combin-
ing the following propagation steps with the uncatalyzed reaction
mechanism:
MnðIIÞ þ SO
5 ! MnðIIIÞ þ HSO5 ð105Þ
MnðIIIÞ þ HSO
3 ! MnðIIÞ þ SO3 ð106Þ
The above equations are not balanced for proton budget, and it was
noted that a change in the pH may have signi¢cant kinetic consequences.
Provided that manganese is present predominantly as manganous ion,
the kinetic model predicts a second-order dependence on [Mn(II)]
and zeroth-order in the concentration of the other components at high
catalyst concentrations, as well as ¢rst-order in both [Mn(II)] and
[HSO
3 ] when the [Mn(II)]/[HSO3 ] ratio is increased and [HSO3 ] is

decreased. Furthermore, it provides a coherent interpretation of

other kinetic e¡ects such as the inhibition by methanol or catalysis by
S2 O2
8 . On the basis of the results, the formation of stable sul¢to
complex(es) could be excluded. This conclusion is in contrast to an ear-
lier study which reported the formation of the stable [Mn(HSO3 Þ]+
complex (113).
Berglund and co-workers (113) observed an induction period in the
manganese(II) catalyzed autoxidation of sulfur(IV) followed by strictly
¢rst-order decay of HSO 3 at pH = 2.4 and at a considerably lower concen-
tration of sulfur(IV) than used by Connick and Zhang. In the presence
of added Mn(III) the induction period was eliminated and the reaction
became faster.The validity of the following experimental rate expression
was con¢rmed:
1 þ B½MnðIIIÞ0
kobs ¼ k107 ½MnðIIÞ ð107Þ
A þ ½MnðIIÞ
where A and B are constants and [Mn(III)]0 represents the concentration
of Mn(III) initially added.
First, the observations were explained on the basis of a kinetic model

which was centered around the formation of a stable hydrogensul¢to
complex with Mn(II). However, in light of the results reported by
Connnick and Zhang (112), the same authors reevaluated their data and
presented evidence for the formation of a mixed-valence dimanganese
(II,III) complex, (OH)MnIIIOMnII (114). An attack by HSO 3 on the
Mn(II) moiety and the formation of a precursor complex were hypothe-
sized and within the latter a bridged electron transfer occurs from
S(IV) to Mn(III). The mixed-valence complex was expected to be a more
active oxidant than Mn(III) alone. The catalytic cycle is completed by
subsequent reactions of the products, Mn(II) and SO 3 . It was concluded
that signi¢cantly di¡erent rate expressions may correspond to the same
catalytic cycle, and the di¡erences can be rationalized by considering
relocation of the rate-determining step when the experimental condi-
tions are altered.
Connick and Zhang did not specify how exactly the catalytic cycle was
initiated, though it can be inferred that the presence of some impurity
is critical in triggering the reaction (112). Fronaeus and co-workers pre-
sented strong arguments that trace amounts of iron(III) are su⁄cient to
initiate the catalytic reaction due to synergism between the two metal
ions (114). When iron(III) is added to the Mn(II)^sul¢te ion system, an
equilibrium between the Mn(II)/Mn(III) and Fe(II)/Fe(III) couples is
rapidly established.The dimanganese(II,III) complex is capable of oxidiz-
ing HSO 3 much faster than by iron(III); thus, the main kinetic role of
Fe(III) is to reoxidize Mn(II) to Mn(III). This reaction produces Fe(II)
which is quickly reoxidized to Fe(III) by SO 5 formed as a secondary
radical in the oxidation of HSO 3 . This qualitative model is supported
by the results presented by Grgič and Berčič (115). However, other
conclusions of these authors are inconsistent with previous results as
they propose a simpli¢ed kinetic model which postulates the formation
of a high stability sul¢to complex with manganese(II). As in the case of
iron(III) catalysis, the speciation of Mn(II) and Mn(III) is crucial in
the mechanism, and further studies should be directed toward
exploring the nature of complexes formed under di¡erent experimental
conditions.
While many important details of the iron- and manganese-catalyzed
reactions are yet to be explored, the common features of the correspond-
ing mechanisms are well established and also applicable in the presence
of other catalysts. Thus, the formation of the SO
5 , SO4 and HSO5

intermediates was reported in all of the free-radical type reactions.

These species are very reactive oxidants and this explains the apparent
paradox of sul¢te autoxidation reactions that various substrates

undergo oxidation with high yields, in the presence of the strongly
reducing sulfur(IV).
The oxidation of two water-soluble Ni(II) complexes with a macrocyclic
(CR = 2,12 -dimethyl-3,7,11,17-tetraazabicyclo-[11.3.1]heptadeca-1(17),2,11,
13,15 -pentaene) and a polydentate (KGH-CONH2 = lysylglycylhistidine-
carboxamide) ligand has been reported (116). The reaction with
[NiCR]2+ shows typical auto-catalytic behavior with an induction
period which decreases by increasing the sul¢te ion concentration. In
this case, the reaction was complete within 200 s in an oxygen-saturated
solution ([[NiCR]2+] = 1.25104 M, [S(IV)] = 1.0^4.0103 M, pH = 6.5)
and the Ni(III) complex product is stable in solution over an extended
period of time. The oxidation of Ni(II) was found to be faster when
peroxomonosulfate was used instead of sul¢te / O2. The oxidation of
the other complex is also auto-catalytic with SO2 3 / O2 and considerably
faster with both oxidants than the corresponding reactions
with [NiCR]2+. The formation of the Ni(III) product was observed only
in a limited sul¢te ion concentration range and analysis of kinetic
traces provided evidence for a back reaction to Ni(II) at longer reaction
times. The kinetic model for these reactions postulates a Ni(II) / Ni(III)
redox cycle and a series of redox reactions with various oxy-sulfur
radicals.
Most of the kinetic models predict that the sul¢te ion radical is
easily oxidized by O2 and/or the oxidized form of the catalyst, but
this species was rarely considered as a potential oxidant. In a recent
pulse radiolysis study, the oxidation of Ni(II and I) and Cu(II and I)
macrocyclic complexes by SO 3 was studied under anaerobic conditions
(117). In the reactions with Ni(I) and Cu(I) complexes intermediates
could not be detected, and the electron transfer was interpreted in
terms of a simple outer-sphere mechanism. In contrast, time resolved
spectra con¢rmed the formation of intermediates with a ligand-
radical nature in the reactions of the M(II) ions. The formation of a
product with a sulfonated macrocycle and another with an addi-
tional double bond in the macrocycle were isolated in the reaction
with [NiCR]2+. These results may require the re¢nement of the
kinetic model proposed by Lepentsiotis for the [NiCR]2+^SO2 3 Ô2
system (116).
The oxidizing power of the catalytic sul¢te ion/O2 systems was utilized
in oxidative cleavage of DNA (118^121), in an analytical application
for the determination of sulfur dioxide in air (122) and in developing a
luminescent probe for measuring oxygen uptake (123).
VII. Autoxidation of Miscellaneous Substrates
Copper complexes were used as e⁄cient catalysts for selective

autoxidations of £avonols (HFLA) to the corresponding o-benzoyl
salicylic acid (o-BSH) and CO in non-aqueous solvents and at elevated
temperatures (124^128). The oxidative cleavage of the pyrazone ring is
also catalyzed by some cobalt complexes (129^131).
In the [CuII(FLA)2]-catalyzed reaction, at the beginning of the kinetic
traces, a relatively slow induction period was observed; this was assigned
to the conversion of the complex into a catalytically active species (127).
The actual oxygenation of £avonol was believed to occur in the consider-
ably faster second phase of the reaction which was con¢rmed to be ¢rst-
order in the concentrations of the catalyst, the substrate and dioxygen
with k = 2.0 103 M2 s1 at 120 C in DMF. The kinetic model was
developed by assuming that a third £avonolate weakly coordinates to
the [CuII(FLA)2] complex. It should be noted that the reactive form of
the catalyst was not speci¢ed, although it must be di¡erent from the ori-
ginal bis-£avonalate complex because of the slow transformation process
in the induction period.
The third ligand was assumed to be coordinated to the metal center via
the deprotonated 3 -hydroxy and 4 -carbonyl groups. This coordination
mode allows delocalization of the electronic structure and intermolecu-
lar electron transfer from the ligand to Cu(II). The Cu(I)-£avonoxy radi-
cal is in equilibrium with the precursor complex and formed at
relatively low concentration levels. This species is attacked by dioxygen
presumably at the C2 carbon atom of the £avonoxyl ligand. In principle,
such an attack may also occur at the Cu(I) center, but because of the
crowded coordination sphere of the metal ion it seems to be less favour-
able. The reaction is completed by the formation and fast rearrangement
of a trioxametallocycle.
Essentially the same mechanism was proposed for the [CuI(FLA)
(PPh3)] (127) and [CuII(FLA)(IDPA)]+ (IDPA = 3,30 -iminobis(N,N-di-
methylpropylamine) (128) catalyzed reactions. In the former case, compo-
site kinetic features were found in the initial phase of the reaction
indicating a slow transformation of the complex presumably to the same
catalytically active species which was postulated in the Cu(FLA)2 system.
For the reaction of [CuII(FLA)(IDPA)]+ straightforward kinetics
were reported, as an induction period was not observed at the beginning
of the reaction, and the rate was ¢rst-order in the catalyst and O2
concentrations but independent of the substrate concentration, with
k = 4.2 102 M1 s1 at 130 C in DMF. In the absence of £avonol,
the catalyst is also oxidized in an overall second-order process and the
SCHEME 6. Reprinted from Inorganica Chimica Acta, vol. 320, Barha¤cs, L.; Kaizer, J.;
Pap, J.; Speier, G. Kinetics and mechanism of the stoichiometric oxygenation of (CuII
(fla)(idpa)]ClO4 [fla=flavanolate, idpa=3,30 -imino-bis(N,N-dimethylpropylamine)] and
the (CuII(fla)(idpa)]ClO4 -catalyzed oxygenation of flavanol, p. 83. Copyright (2002),
with permission from Elsevier Science.
corresponding rate constant, k = 6.1 103 M1 s1 (100 C), 2.9 102
M1 s1 (130 C), is in reasonable agreement with that of the catalytic
reaction under the same conditions.This suggests that the rate-determin-
ing step in the catalytic cycle is the oxidation of the coordinated FLA
ligand of the catalyst which is rapidly regenerated as long as the
substrate is present in excess. The details of the mechanism proposed
for this reaction are shown in Scheme 6.
A comparison of the rate constants for the [CuII(FLA)(IDPA)]+-cata-
lyzed autoxidation of 40 -substituted derivatives of £avonol revealed a
linear free energy relationship (Hammett) between the rate constants
and the electronic e¡ects of the para-substituents of the substrate (128).
The logarithm of the rate constants linearly decreased with increasing
Hammett values, i.e. a higher electron density on the copper center
yields a faster oxidation rate.
Catalytic activation of dioxygen by Co(ClO4)26H2O, cobalt(II) phthalo-
cyanines (PC = tetrakis(3,5 -di-tert-butyl- 4 -hydroxyphenyl)dodecachlor-
ophthalocyaninate) and cobaloxime(II) ([CoII(HDMG)2(L)2; H2DMG =
dimethylglyoxime, L = Ph3P, Ph3As, Ph3Sb) has been studied in oxidation
reactions of phenol, thiophenol, and aniline derivatives, primarily, at
ambient temperature and in non-aqueous solvents by Sima¤ndi and co-
workers (132^141). A common kinetic feature of these reactions is the
accumulation of reactive intermediates in a relatively fast initial phase
which persisted for several minutes under typical experimental condi-
tions. The rate laws were evaluated by analyzing the concentration
dependencies of the oxygen uptake rates measured after the initial

stage, i.e. when the catalytic systems reached steady states. The
observations were consistent with a mechanism very similar to that
postulated for cobaloxime(II)-catalyzed oxidation of H2DTBC (54,55)
(cf. Section IV). Thus, fast reversible formation of a superoxocobalt(III)
complex, which is bonded reversibly to the substrate via a hydrogen
bond, was proposed. The rate-determining step is the dissociation of
this adduct into a hydroperoxocobalt complex and a radical derived
from the substrate. Subsequent steps include the recombination of the
free radical, its reaction with cobalt(III) and decomposition of the hydro-
peroxocobalt species. These ¢ndings were supported by results from
ESR measurements which were used to identify some of the intermediate
radical species (135,139,141).
With some modi¢cations, the above mechanism is also operative when
the substrate is suitable for oxidative attack at more than one position.
For example, in the cobaloxime(II)-catalyzed oxidation of 3,30,5,50 -tetra-
tert-butyl- 4,40 -dihydroxystilbene (H2StA), oxidative dehydrogenation
yielded stilbenequinone (StQ), and simultaneous cleavage of the double
bond gave 2,6 -di-tert-butyl- 4 -hydroxybenzaldehyde (141). When the
system reached steady-state, the formation of the two products pro-
ceeded at comparable rates presumably via the common StQ anion
radical. This species can be involved in a further hydrogen abstraction
step with CoIIIOH (CoIII represents the [CoIII(HDMG)2(Ph3P)]+ moiety)
or attacked in its mesomeric form at one of the ole¢nic carbon atoms by
the superoxocobalt(III) species forming an alkylperoxocobaloxime(III)
intermediate. The aldehyde product is formed in subsequent fast steps.
Such a reaction does not occur in the absence of the catalyst and this
reaction serves as a rare example of the ability of cobaloxime(II) to
catalyze the oxidative cleavage of an ole¢nic double bond.
In a broad sense, the model developed for the cobaloxime(II)-catalyzed
reactions seems to be valid also for the autoxidation of the alkyl
mercaptan to disul¢des in the presence of cobalt(II) phthalocyanine
tetra-sodium sulfonate in reverse micelles (142). It was assumed that the
rate-determining electron transfer within the catalyst-substrate-dioxy-
gen complex leads to the formation of the ¢nal products via the RS and
O
2 radicals. The yield of the disul¢de product was higher in waterôil
microemulsions prepared from a cationic surfactant than in the
presence of an anionic surfactant. This di¡erence is probably due to
the stabilization of the monomeric form of the catalyst in the former
environment.
Ruthenium(III) was shown to be a potent catalyst for the autoxidation
of various organic substrates by Taqui Khan and co-workers (143^148).
TABLE III
CONCENTRATION DEPENDENCIES OF THE EXPERIMENTAL RATE LAW FOR RUTHENIUM
CATALYZED AUTOXIDATION OF ORGANIC SUBSTRATES a,b
Substrate Catalyst H2A c Ph nO Reference
cyclohexene Ru(III) no 2.0 1 (146)

cyclohexene RuIII(EDTA) no 1.0^3.0 1/2 (145)
cyclohexene Ru(III) yes 2.0 1 (146)
cyclohexene RuIII(EDTA) yes 2.5 1 (148)
cyclohexane d RuIII(EDTA) yes 2.5 1 (148)
cyclohexanol Ru(III) no 1.5^3.0 1 (143)
cyclohexanol RuIII(EDTA) no 1.5^3.0 0 (143)
cyclohexanol RuIII(EDTA) yes 2.5 1 (148)
allyl alcohol Ru(III) no 1.0^2.0 0 (144)
no 2.0^3.0 1/2 (144)
allyl alcohol RuIII(EDTA) no 1.0^2.0 0 (147)
no 2.0^3.0 1 (147)
a
Typical experimental conditions: T=288^313 K, =0.1 M KNO3 or KCl, in water, 1 : 1 waterêthanol
or water-1,4 -dioxane.
b
nO is the reaction order in O2. The reaction was first-order in the catalyst and the substrate con-
centration.
c
The reaction was first-order in ascorbic acid (H2A) concentration when it was used as a co-sub-
strate.
d
Fractional-order was reported for the substrate.
Some of the details of these reactions are summarized in Table III. The
corresponding kinetic models were centered around the formation of a
complex which included all the reactants, i.e. the catalyst, the substrate,
in most cases dioxygen and also ascorbic acid when it was used as a
co-substrate.This precursor complex was considered to undergo an inter-
nal redox process in a rate-determining step. According to this general
model, the rate expression can take di¡erent forms depending on how
many components are assembled in the precursor complex and where
the rate determining-step is located in the reaction sequence.
The oxidation of allyl alcohol shows an interesting pH dependence
(144,147). When the pH<2.0, the main product is acrylaldehyde and
the rate is independent of, or inversely proportional to [H+], with the
catalysts Ru(III) and Ru(EDTA), respectively.5 In this case, hydride
5
In these studies Ru(III) and Ru(EDTA) represent [RuCl2(H2O)4]+ and a gener-
alized EDTA complex of Ru(III), respectively. The coordination mode and com-
position, in particular with respect to possible protonated complexes, were not
speci¢ed for the EDTA complex in the cited references.
abstraction from the substrate to the metal center within the catalyst-
substrate complex was considered to be the rate-determining step
which is followed by fast oxidation of the hydrido complex. Above pH
2.0, the reaction becomes independent of pH with both catalysts, and
epoxidation of the double bond occurs simultaneously with the
oxidation of the alcoholic group. In the case of Ru(III), the formation of
a [S^RuIII^(m-O2)^RuIII^S] (S stands for substrate) dinuclear complex,
which decomposes via the cleavage of the OÔ bond in a rate determin-
ing step, was postulated. The reaction supposedly proceeds via mono-
nuclear complexes when Ru(EDTA) is used as catalyst.
The above models imply that the proton loss of the OH group of
the coordinated substrate shifts the mechanism from oxidation to
epoxidation with Ru(III). Such a straightforward interpretation of the
pH e¡ect was not presented for reactions of the other substrates, i.e.
the protolytic reaction, which would act as a switch between the two
mechanisms, cannot be identi¢ed.
The models proposed for the Ru(III) and Ru(EDTA) catalyzed epoxida-
tion of allyl alcohol replace each other in the corresponding reactions of
cyclohexene. Thus, mono-nuclear and di-nuclear paths were reported for
the Ru(III) and Ru(EDTA) catalysis, respectively (145,146).
The presence of ascorbic acid as a co-substrate enhanced the rate
of the Ru(EDTA)-catalyzed autoxidation in the order cyclohexane<
cyclohexanol<cyclohexene (148). The reactions were always ¢rst-
order in [H2A]. It was concluded that these reactions occur via a
Ru(EDTA)(H2A)(S)(O2) adduct, in which ascorbic acid promotes the
cleavage of the O2 unit and, as a consequence, O-transfer to the substrate.
While the model seems to be consistent with the experimental
observations, it leaves open some very intriguing questions. According
to earlier results from the same laboratory (24,25), the Ru(EDTA)
catalyzed autoxidation of ascorbic acid occurs at a comparable or even
a faster rate than the reactions listed in Table III. It follows, that
the interference from this side reaction should not be neglected in
the detailed kinetic model, in particular because ascorbic acid may be
completely consumed before the oxidation of the other substrate
takes place.
It is not clear either how the Ru center can accommodate four ligands
simultaneously. The crowded coordination sphere around the metal
center in the Ru(EDTA)âscorbate complex is expected to hinder the
coordination of other ligands as was proposed earlier (24,25). The
contradiction between the two sets of results reported in Refs. (24,25)
and (148) is obvious.While the Ru(EDTA)(H2A)(O2) complex was not con-
sidered in the kinetic model proposed for the oxidation of ascorbic acid,
a relatively high equilibrium constant, log K = 3.18, was reported for

the formation of the same species from Ru(EDTA)(H2A) and O2 in the
latter study. Because di¡erences in the experimental conditions cannot
account for the above discrepancies, the need for a coherent re-interpre-
tation of the kinetic data seems to be imperative and inevitable.
The enhanced chemiluminescence associated with the autoxidation
of luminol (5 -amino-2,3 -dihydro-1,4 -phthalazinedione) in the presence
of trace amounts of iron(II) is being used extensively for selective
determination of Fe(II) under natural conditions (149^152). The speci¢-
city of the reaction is that iron(II) induces chemiluminescence with O2,
but not with H2O2, which was utilized as an oxidizing agent in the deter-
mination of other trace metals. The oxidation of luminol by O2 is often
referred to as an iron(II)-catalyzed process but it is not a catalytic reac-
tion in reality because iron(II) is not involved in a redox cycle, rather it
is oxidized to iron(III). In other words, the lower oxidation state metal
ion should be regarded as a co-substrate in this system. Nevertheless,
the reaction deserves attention because it is one of the few cases where
a metal ion signi¢cantly a¡ects the autoxidation kinetics of a substrate
without actually forming a complex with it.
The source of chemiluminescence in the oxidation of luminol was
explored by Mere¤nyi and co-workers in detail (153).The oxidation of lumi-
nol yields aminophthalate as a ¢nal product and the reaction proceeds
via a series of electron transfer steps. The primary oxidation product is
the luminol radical which is transformed into either diazaquinone or
the a-hydroxide-hydroperoxide intermediate (a-HHP). The latter oxida-
tion step occurs between the deprotonated form of the luminol radical
and O2 . The chemiluminescence is due to the decomposition of the
mono-anionic form of a-HHP into the ¢nal products:
ð108Þ
An alternative path includes oxidation, in the absence of light, of the

diazaquinone with weak chemiluminescence (154). The effect of iron(II)
on the luminescent intensity was interpreted by considering that it
can efficiently generate the O
2 radical in a reaction with O2 and, as
a consequence, increase the importance of reaction (108) in the overall
process (155).
Trace amounts of Cu(II) were reported to catalyze the oxidation of I

to I2 (156) and the phosphinate ion ðH2 PO 2 Þ to peroxodiphosphate ion
(PDP), which could be present as P2 O4 8 , HP2 O3
8 or H2 P2 O2
8 (157).
Individual kinetic traces showed some unusual patterns in these reac-
tions, such as the variation between ¢rst- and zeroth-order kinetics
with respect to the formation of I2 under very similar conditions, or an
autocatalytic feature in the concentration pro¢les of PDP, but these
events were not studied in detail. The catalytic e¡ect was interpreted in
terms of a Cu(II) / Cu(I) redox cycle and the superoxide ion radical,
O
2 = HO2 , was considered as a reactive intermediate in both cases. The
addition of radical scavengers strongly retarded the oxidation of the
phosphinate ion con¢rming the radical type mechanism. It was also
demonstrated that the reaction ceased when the catalyst was masked
with EDTA.
Iron(III) chelate complexes were utilized as catalysts for oxidative
removal of H2S from natural gas.The application of FeIII(NTA) is plagued
by the degradation of the ligand to weaker chelating agents such as
iminodiacetic acid, glycine, oxalate and eventually CO2 is generated.
According to Chen et al., the degradation occurred on the time-scale of
3^4 days when air and H2S were included in a reaction mixture originally
containing only the catalyst (158). Substantial acidi¢cation was observed
in unbu¡ered reaction mixtures, and the pH dropped from 8.6 to 3.5 in a
typical experiment. This combined with the fact that the reaction
became much slower at constant pH 8.5 indicates the autocatalytic
feature of the overall reaction.
It was proposed that the degradation of the catalyst occurs simulta-
neously with the following catalytic cycle:
2FeðNTAÞ þ H2 S ! 2FeðNTAÞ þ 1=8S8 þ 2Hþ ð109Þ
2FeðNTAÞ þ 1=2O2 þ H2 O ! 2FeðNTAÞ þ 2OH ð110Þ
A free radical mechanism was considered for the reoxidation of the

Fe(II) complex involving the formation of H2O2. The decomposition of
hydrogen peroxide generates the primary OH and the secondary O 2
radicals. While O2 is not a su⁄ciently strong oxidant to attack NTA, the
free radicals are reactive enough to induce oxidative degradation of the
ligand and byproducts. The formation of H2O2 was con¢rmed in
experiments with the enzyme catalase and the reaction was inhibited
by addition of Na2S2O3, a well known free radical scavenger. These
observations fully support the proposed model.
VIII. Exotic Kinetic Phenomena
A number of autoxidation reactions exhibit exotic kinetic phenomena

under speci¢c experimental conditions. One of the most widely studied
systems is the peroxidase-oxidase (PO) oscillator which is the only
enzyme reaction showing oscillation in vitro in homogeneous stirred
solution. The net reaction is the oxidation of nicotinamide adenine dinu-
cleotide (NADH), a biologically vital coenzyme, by dioxygen in a horse-
radish peroxidase enzyme (HRP) catalyzed process:
NADH þ O2 ! 2NADþ þ 2H2 O ð111Þ
Since the ¢rst report of oscillation in 1965 (159), a variety of other non-
linear kinetic phenomena have been observed in this reaction, such
as bi-stability, bi-rhythmicity, complex oscillations, quasi-periodicity,
stochastic resonance, period-adding and period-doubling to chaos.
Recently, the details and sub-systems of the PO reaction were surveyed
and a critical assessment of earlier experiments was given by Scheeline
and co-workers (160). This reaction is beyond the scope of this
chapter and therefore, the mechanistic details will not be discussed
here. Nevertheless, it is worthwhile to mention that many studies
were designed to explore non-linear autoxidation phenomena in less
complicated systems with an ultimate goal of understanding the PO
reaction better.
The main features of the copper catalyzed autoxidation of ascorbic
acid were summarized in detail in Section III. Recently, Strizhak
and coworkers demonstrated that in a continuously stirred tank
reactor (CSTR) as well as in a batch reactor, the reaction shows various
non-linear phenomena, such as bi-stability, oscillations and stochastic
resonance (161). The results from the batch experiments can be
suitably illustrated with a two-dimensional parameter diagram shown
in Fig. 5.
At high [Cu(II)] and low [H2A] initial concentrations, the Pt electrode
potential, used to follow the chemical process, increased monotonously.
When both species were present at high initial concentrations, a monoto-
nous decrease was observed. Various non-monotonic transient regimes
were found at approximate initial concentrations of [Cu(II)] 104 M
and [H2A] 104 M. Thus, the batch experiments properly illustrate
that the system is sensitive to variations of the initial concentrations of
ascorbic acid and copper(II) ion, and the observations can be indicative
of a transient bi-stability.
F IG. 5. Two-dimensional parametric diagram of system response at different

initial concentrations of reagents in batch: n, monotonic growth of Pt potential
[Fig. 1(a)]; ,, monotonic decrease of Pt potential [Fig. 1(b)]; s, Pt electrode potential
first decreases and then increases in time [Fig. 1(c)]; , various nonmonotonic
transient regimes [Fig. 1(d^f)]. Strizhak, P. E.; Basylchuk, A. B.; Demjanchyk, I.;
Fecher, F.; Shcneider, F. W.; Munster, A. F. Phys. Chem. Chem. Phys. 2000, 2, 4721.
Reproduced by permission of The Royal Society of Chemistry on behalf of the PCCP
Owner Societies.
The experiments were performed at a constant in£ow concentration

of ascorbic acid ([H2A]) in the CSTR. Oscillations were found by
changing the £ow rate and the in£ow concentration of the copper(II)
ion systematically. At constant Cu(II) in£ow concentration, the electrode
potential measured on the Pt electrode showed hysteresis between two
stable steady-states when ¢rst the £ow-rate was increased, and then
decreased to its original starting value. The results of the CSTR
experiments were summarized in a phase diagram (Fig. 6).
In Fig. 6, separate regions of bi-stability, oscillations and single stable
steady-states can be noticed. This ‘‘cross-shaped’’ phase diagram is
common for many non-linear chemical systems containing autocatalytic
steps, and this was used as an argument to suggest that the Cu(II) ion
catalyzed autoxidation of the ascorbic acid is also autocatalytic. The
FIG. 6. Dynamical phase diagram of the ascorbic acid/copper(II)/oxygen system in

a CSTR in the kf [Cu2+ ]0 plane. Fixed reactor concentrations: [H2 Asc] 0 =
5.0 104 M; [H2SO4]0 = 6.0 105 M; [Na2SO4]0 = 0.04 M. Symbols: s, steady state;
, oscillations; r, bistability. The asterisk (*) marks the Takens-Bogdanov point.
Strizhak, P. E.; Basylchuk, A. B.; Demjanchyk, I.; Fecher, F.; Shcneider,
F. W.; Munster, A. F. Phys. Chem. Chem. Phys. 2000, 2, 4721. Reproduced by permis-
sion of The Royal Society of Chemistry on behalf of the PCCP Owner Societies.
following reaction scheme was proposed for the interpretation of the

results:
H2 A þ Cu2þ ! HA þ Cuþ þ Hþ ð112Þ
HA þ O2 ! HO2 þ A ð113Þ
HO2 þ H2 A ! HA þ H2 O2 ð114Þ
Cuþ þ H2 O2 þ Hþ ! OH þ H2 O þ Cu2þ ð115Þ
OH þ HA ! HA þ H2 O ð116Þ

This model is consistent with the results discussed in Section III in

that it also postulates the Cu(II)/Cu(I) redox cycle. Otherwise, there are
signi¢cant di¡erences, such as the removal of H2O2 as in Eq. (115), or
the formation of the hydroxyl radical, which would require justi¢cation.
Since calculations were not presented to support the kinetic model, it is
open to question as to how well the observations would be reproduced
by reactions (112^116) or, for that matter, by the alternative models.
The oxidation of benzaldehyde with dioxygen in the presence of Co(II)
and bromide ion also shows non-linear kinetic phenomena. The net
reaction of the oxidation process is given as follows:
2PhCHO þ O2 ðsolnÞ ! 2PhCO2 H ð117Þ
Jensen was the ¢rst to report in 1983 that the color of the solution
oscillated between pink and dark brown in the presence of cobalt(II)
and bromide ions when the reaction was carried out in a 90/10 (w/w)
acetic acid/water mixture (162). This color change was accompanied by
a change in the redox potential and the oscillations were observed for
over 16 h and 800 cycles. Presumably, the pink color corresponds to a low
Co(III)/Co(II) ratio, the dark brownish black to a high Co(III)/Co(II)
ratio or to a Co(III)Br complex in this reaction.
In later work, Roelofs and co-workers discovered further details of
the reaction by investigating the sub-systems, and they suggested a 21
step chemical model (RWJ model) to explain the observed non-linear
kinetic patterns (163). According to the experimental observations, the
oscillation process can be divided into two distinct alternating stages,
the stoichiometries of which can be approximated as follows:
PhCHO þ O2 ðsolnÞ þ 2CoðIIÞ þ 2Hþ ! PhCO2 H þ 2CoðIIIÞ þ H2 O
ð118Þ
PhCHO þ 2CoðIIIÞ þ H2 O ! PhCO2 H þ 2CoðIIÞ þ 2Hþ ð119Þ

At the beginning of stage I the concentration of dissolved oxygen is
high, and that of Co(III) is low. Equation (118) represents an autocatalytic
process, and as the concentration of the product Co(III) increases the
rate accelerates, until it reaches a point where the rate of depletion of
dissolved oxygen becomes higher than the rate of oxygen transfer
from the gas phase to the solution. As a consequence, the concentration
of dissolved oxygen falls, and at the end of Stage I dioxygen is totally
depleted in the liquid phase. At this point reaction (119) becomes
operative, i.e. the Co(III) formed in stage I is consumed in a rapid reac-
tion with benzaldehyde. The model postulates that the reactive Co(III)
species is a dimer complex, and a number of radicals (PhCO, PhCOO,
PhCOOO, Co(II)Br) were assumed to be important intermediates in the

reaction. The presence of some of the free radicals was con¢rmed by
EPR spectroscopy, providing additional support for the validity of the
RWJ model (164).
The model proposed on the basis of the experimental work of Colussi et
al. (CGYN) is not consistent with the RWJ model, in some details (165).
The most important di¡erence lies in the interpretation of the role of
bromide ion and of bromine that could be formed from bromide ion.
Roelofs and co-workers assumed that Br is important only in the
complex-formation and they did not take into consideration the forma-
tion and further reactions of Br2 in their oscillatory model (164). Colussi
et al. suggested that Br2 may form at low concentration levels and veri¢ed
experimentally that bromine added to the system during the oscillation
reduces the concentration of Co(III), and makes the periods shorter.
A detailed molecular mechanism was proposed in agreement with the
observations on the overall reaction and sub-systems (165).
The oxidation of benzaldehyde by molecular oxygen is an autocatalytic
process which may feature propagating reaction fronts. Boga et al.
investigated this phenomenon in the absence of bromide ion by using a
mixture of Co(II) acetate and benzaldehyde solutions saturated
with an O2^N2 mixture (166). The wave was initiated in a capillary tube
by addition of perbenzoic acid, an intermediate of the reaction. The
velocity of front propagation as a function of Co(II), PhCHO, and O2
concentrations was explored. At low concentration levels, Co(II) is a
catalyst, but at high concentrations it acts as a scavenger of the
PhCOOO radical and inhibits the autoxidation of benzaldehyde,
Eq. (124).The authors proposed the following skeleton model for interpre-
tation of the results:
PhCOOOH þ CoðIIÞ ! PhCOO þ CoðIIIÞ þ OH

k120 ¼ 1:9 105 M1 s1 ð120Þ
CoðIIIÞ þ PhCHO ! CoðIIÞ þ PhCO þ Hþ

k121 ¼ 1:6 104 M1 s1 ð121Þ
PhCO þ O2 ! PhCOOO k122 ¼ 5:0 108 M1 s1 ð122Þ
PhCOOO þ PhCHO ! PhCO þ PhCOOOH

k123 ¼ 3:9 103 M1 s1 ð123Þ
PhCOOO þ CoðIIÞ ! CoðIIIÞ þ PhCOOO

k124 ¼ 8:0 106 M1 s1 ð124Þ
PhCOO þ PhCHO ! PhCO þ PhCOOH

k125 ¼ 1:0 107 M1 s1 ð125Þ
It was shown that an appropriate combination of the above steps, i.e.
(120) + (121) + 2(122) + 2(123) + (125), predicts autocatalytic accumulation
of the perbenzoic acid:
PhCOOOH þ 2PhCHO þ 2O2 þ 2CoðIIÞ
! 2PhCOOOH þ 2CoðIIIÞ þ PhCOO þ OH ð126Þ
Guslander and co-workers developed another simple skeleton
model on the basis of the CGYN model, named Cobaltolator, which
consists of only four steps (167). Numerical simulations with this
model showed that it can reproduce the main features of the oscillation
reaction, and it can also simulate the front propagation found by
Boga et al.
Searching for other oscillatory autoxidation reactions led Druliner
and Wasserman to use cyclohexanone as a substrate instead of benzalde-
hyde (168). Unlike the simple stoichiometry found for the benzaldehyde
reaction, the ketone gives at least six or more products, and the relative
amounts of these vary substantially with the experimental conditions
(Scheme 7).
The overall appearance of the oscillation in the cyclohexanone
autoxidation is di¡erent from that in the corresponding reaction of
benzaldehyde, indicating signi¢cant di¡erences in the mechanisms. The
oscillation can be divided into three stages. In the ¢rst stage, organic
intermediates, mainly KO2H and HOK, are formed in a slow autoxida-
tion of cyclohexanone (cf. Scheme 7). During this phase the electrode
potential of the system is practically constant, which indicates that the
Co(III)/Co(II) ratio is not altered signi¢cantly. When enough peroxide is
generated the second stage begins, where ¢rst HOK reacts with Co(III)
to give RCO, and subsequent steps follow as shown in Scheme 8.
In this stage, the conversion of Co(II) to Co(III) is indicated by the
rapid rise in the electrode potential, and fast generation of Co(III)
accelerates the formation of HOK. Finally, in the third stage the accumu-
lated Co(III) oxidizes the organic substrates to radical intermediates,
while itself is reduced back to Co(II). Because of the noted complexity,
the oscillation model for the cyclohexanone reaction was not elaborated
upon in detail.
SCHEME 7. Reprinted with permission from Druliner, J. D.; Wasserman, E. J. Am.

Chem. Soc. 1988, 110, 5270. Copyright (2002) American Chemical Society.
SCHEME 8.
IX. Perspectives
The results surveyed in this chapter demonstrate the composite nature

of autoxidation reactions and the potential problems involved in explor-
ing the intimate details of the appropriate mechanisms. The kinetic
observatiuons have been interpreted to di¡erent depths, but there is
still plenty of room for improving available kinetic models. Most
certainly, a great number of future studies will be designed to character-

ize better the reactive intermediates and to identify the dominant
reaction paths in these systems. Recent developments in reaction
kinetics o¡er sophisticated technical and computational tools to accom-
plish these objectives. New fast kinetic devices allow the recording of
time resolved spectra on the millisecond time-scale with a reliability
comparable to that of high quality spectrophotometers; the sensitivity,
detection limits and response times of electrochemical instruments
have been improved by orders of magnitude using new technologies. The
availability of micro and nano devices has opened new territories for
time-dependent studies.
The quantity and quality of experimental information determined by
the new techniques call for the use of comprehensive data treatment
and evaluation methods. In earlier literature, quite often kinetic studies
were simpli¢ed by using pseudo-¢rst-order conditions, the steady-state
approach or initial rate methods. In some cases, these simpli¢cations
were fully justi¢ed but sometimes the approximations led to distorted
results. Autoxidation reactions are particularly vulnerable to this
problem because of strong kinetic coupling between the individual
steps and feed-back reactions. It was demonstrated in many cases, that
these reactions are very sensitive to the conditions applied and their
kinetic pro¢les and stoichiometries may be signi¢cantly altered by
changing the pH, the absolute concentrations and concentration ratios
of the reactants, and also by the presence of trace amounts of impurities
which may act either as catalysts and/or inhibitors.
By lifting the simplifying restrictions, the kinetic observations can be
examined in more detail over much wider concentration ranges of the
reactants than those relevant to pseudo-¢rst-order conditions. It should
be added that sometimes a composite kinetic trace is more revealing
with respect to the mechanism than the conventional concentration
and pH dependencies of the pseudo-¢rst-order rate constants.
Simultaneous evaluation of the kinetic curves obtained with di¡erent
experimental methods, and recorded under di¡erent conditions, is
based on ¢tting the proposed kinetic models directly to the primary
data. This method yields more accurate estimates for the rate constants
than conventional procedures. Such an approach has been used sporadi-
cally in previous studies, but it is expected to be applied more widely
and gain signi¢cance in the near future.
Reliable mechanisms can serve as the basis for the design of e⁄cient
new catalysts for autoxidation reactions. A systematic analysis of the
e¡ects of the non-participating ligands on the kinetics of the overall
reaction and on the catalytic activity of the metal center(s) could be a
driving force to prepare novel complexes with special functionalities.

Furthermore, the transfer of the catalyst to a non-traditional environ-
ment may be an exciting ¢eld for mechanistic studies. Already there are
a few examples where certain catalytic properties are retained when
the metal centers are attached to polymeric surfaces; however, the
porosity and other properties of the carrier also a¡ect the reactions.
Non-conventional solvents, such as supercritical £uids, or biphasic
systems can also be used to tune the activity of some of the catalysts.
Finally, kinetic and mechanistic studies of autoxidation reactions
will not only lead to a better understanding of these essential reactions
but also may trigger extensive studies on various areas of dioxygen
chemistry.
ACKNOWLEDGEMENTS
This work was supported by the Hungarian National Research Foundation under grant
Nos. OTKA T 029568 and 042755.
REFERENCES
1. Sawyer, D. T. ‘‘Oxygen Chemistry’’; Oxford University Press: New York, 1991.

2. ‘‘Standard Potentials in Aqueous Solution’’; Eds. Bard, A. J.; Parson, R.; Jordan, J.;
Marcel Dekker: New York, 1985.
3. Cofre¤, P.; Sawyer, D. T. Inorg. Chem. 1986, 25, 2089.
4. Czapski, G. Annu. Rev. Phys. Chem. 1971, 22, 171.
5. Miller, D. M.; Buettner, G. R.; Aust, S. D. Free Radic. Biol. Med. 1990, 8, 95.
6. Taqui Khan, M. M.; Martell, A. E. J. Am. Chem. Soc. 1967, 89, 4176.
7. Gampp, H.; Zuberbu«hler, A. D. ‘‘Metal Ions in Biological Systems’’, vol. 12; Ed. Sigel,
H.; Marcel Dekker: New York, 1981, pp. 133^189.
8. Zuberbu«hler, A. D. ‘‘Copper Coordination Chemistry: Biochemical and Inorganic
Perspectives’’; Eds. Karlin, K. D.; Zubieta, J.; Adenine, New York, 1983; pp. 237^258.
9. Shtamm, E. B.; Purmal, A. B.; Skurlatov, I. Zhur. Fiz. Khim. 1974, 48, 2229.
10. Weissberger, A.; Luvalle, J. E.; Thomas, D. S., Jr. J. Am. Chem. Soc. 1943, 65, 1934.
11. Jameson, R. F.; Blackburn, N. J. J. Chem. Soc., Dalton Trans. 1976, 534.
13. Hadady-Kova¤cs, K.; Fa¤bia¤n, I. J. Pharm. Biomed. Anal. 1996, 14, 1479.
14. Sisley, M. J.; Jordan, R. B. J. Chem. Soc., Dalton Trans. 1997, 3883.
16. Mi, L.; Zuberbu«hler, A. D. Helv. Chim. Acta 1992, 75, 1547.
17. Mi, L.; Zuberbu«hler, A. D. Helv. Chim. Acta 1991, 74, 1679.
18. Scarpa, M.; Vianello, F.; Signor, L.; Zennaro, L.; Rigo, A. Inorg. Chem. 1996, 35, 5201.
19. Thomas, A. M.; Mandal, G. C.; Tiwary, S. K.; Rath, R. K.; Chakravarty, A. R. J. Chem.
Soc., Dalton Trans. 2000, 1395.
20. Srinivas, B.; Arulsamy, N.; Zacharias, P. S. Polyhedron 1991, 10, 731.
21. Xu, J. H.; Jordan, R. B. Inorg. Chem. 1990, 29, 4180.
22. Ba«nsch, B.; Martinez, P.; Uribe, D.; Zuluaga, J.; van Eldik, R. Inorg. Chem. 1991, 30, 4555.
23. Lepentsiotis, V.; van Eldik, R. J. Chem. Soc., Dalton Trans. 1998, 999.
24. Taqui Khan, M. M.; Shukla, R. S. J. Mol. Catal. 1986, 34, 19.
25. Taqui Khan, M. M.; Shukla, R. S. J. Mol. Catal. 1986, 37, 269.
26. Seibig, S.; van Eldik, R. Inorg. React. Mech. 1999, 1, 91.
27. Speier, G.; Tyekla¤r, Z. J. Chem. Soc., Perkin Trans. 2 1981, 1176.
28. Que, L., Jr.; Kolanczyk, R. C.; White, L. S. J. Am. Chem. Soc. 1987, 109, 5373.
29. Cox, D. D.; Que, L., Jr. J. Am. Chem. Soc. 1988, 110, 8085.
30. Que, L., Jr.; Ho, R. Y. N. Chem. Rev. 1996, 96, 2607.
31. Funabiki, T. ‘‘Oxygenases and Model Systems’’; Ed. Funabiki, T.; Kluwer: Boston,
1997, pp. 19^156.
32. Demmin, T. R.; Swerdlo¡, M. D.; Rogic, M. M. J. Am. Chem. Soc. 1981, 103, 5795.
33. Bolus, D.; Vigee, G. S. Inorg. Chim. Acta 1982, 67, 19.
34. Rogic, M. M.; Swerlo¡, M. D.; Demmim, T. R. ‘‘Copper Coordination Chemistry:
Biochemical and Inorganic Perspectives’’; Eds. Karlin, K. D., Zubieta, J.; Adenine,
New York, 1983; p. 259.
35. Tyson, C. A.; Martell, A. E. J. Am. Chem. Soc. 1972, 94, 939.
36. Balla, J.; Kiss, T.; Jameson, R. F. Inorg. Chem. 1992, 31, 58.
37. Abakumov, G. A.; Lobanov, A. V.; Cherkasov, V. K.; Razuvaev, G. A. Inorg. Chem.
Acta 1981, 49, 135.
38. Pierpont, C. G.; Buchanan, R. M. Coord. Chem. Rev. 1981, 38, 45.
39. Tyson, C. A.; Martell, A. E. J. Am. Chem. Soc. 1968, 90, 3379.
40. Shtamm, E. V.; Purmal, A. P.; Skurlatov, Y. I. Int. J. Chem. Kinet. 1979, 11, 461.
41. Speier, G. J. Mol. Catal. 1986, 37, 259.
42. Li, Y. B.; Trush, M. A. Arch. Biochem. Biophys. 1993, 300, 346.
43. Zhang, L. P.; Bandy, B.; Davison, A. J. Free Radic. Biol. Med. 1996, 20, 495.
44. Mochizuki, M.; Yamazaki, S.; Kano, K.; Ikeda, T. Biochim. Biophys. Acta 2002, 1569, 35.
45. Hayakawa, F.; Kimura, T.; Maeda, T.; Fujita, M.; Sohmiya, H.; Fujii, M.; Ando, T.
Biochim. Biophys. Acta 1997, 1336, 123.
46. Hayakawa, F.; Kimura, T.; Hoshino, N.; Ando, T. Biosci. Biotechnol. Biochem. 1999, 63,
1654.
47. Chyn, J. P.; Urbach, F. L. Inorg. Chim. Acta 1991, 189, 157.
48. Manzur, J.; Garcia, A. M.; Cordova, C.; Pizzaro, O.; Acuna, V.; Spodine, E. Polyhedron
2002, 21, 181.
49. Lim, Y. Y.; Tan, E. H. L.; Foong, P. C. J. Mol. Catal. 1993, 85, 173.
50. Mazer, N. A.; Benedek, G. B.; Carey, M. C. J. Phys. Chem. 1976, 80, 1075.
51. Lim, Y. Y.; Fong, H. Y.; Tan, L. H. J. Mol. Catal. A 1995, 97, L1.
52. Tsuruya, S.; Yanai, S.; Masai, M. Inorg. Chem. 1986, 25, 141.
53. Sima¤ndi, T. L.; Sima¤ndi, L. I. React. Kin. Catal. Let. 1998, 65, 301.
54. Sima¤ndi, L. I.; Sima¤ndi, T. L. J. Chem. Soc., Dalton Trans. 1998, 3275.
55. Sima¤ndi, T. L.; Sima¤ndi, L. I. J. Chem. Soc., Dalton Trans. 1999, 4529.
56. Barbaro, P.; Bianchini, C.; Frediani, P.; Meli, A.; Vizza, F. Inorg. Chem. 1992, 31, 1523.
57. Barbaro, P.; Bianchini, C.; Mealli, C.; Meli, A. J. Am. Chem. Soc. 1991, 113, 3181.
58. Bianchini, C.; Frediani, P.; Laschi, F.; Meli, A.; Vizza, F.; Zanello, P. Inorg. Chem.
1990, 29, 3402.
59. Barbaro, P.; Bianchini, C.; Linn, K.; Mealli, C.; Meli, A.; Vizza, F. Inorg. Chim. Acta
1992, 200, 31.
60. Que, L., Jr.; Lau¡er, R. B.; Lynch, J. B.; Murch, B. P.; Pyrz, J. W. J. Am. Chem. Soc.
1987, 109, 5381.
61. Pascaly, M.; Duda, M.; Schweppe, F.; Zurlinden, K.; Mu«ller, F. K.; Krebs, B. J. Chem.
Soc., Dalton Trans. 2001, 828.
62. Jameson, R. F.; Kiss, T. J. Chem. Soc., Dalton Trans. 1986, 1833.
63. Karlin, K. D.; Tyekla¤r, Z. ‘‘Advances in Inorganic Biochemistry’’, vol. 9; Eds.

Eichhorn, G. L.; Marzilli, L. G.; Prentice Hall: New Jersey, 1994, p. 123.
64. Zwart, J.; van Wolput, J. H. M. C.; van der Cammen, J. C. J. M.; Koningsberger, D. C.
J. Mol. Catal. 1981, 11, 69.
65. Zwart, J.; van Wolput, J. H. M. C.; Koningsberger, D. C. J. Mol. Catal. 1981, 12, 85.
66. Hanaki, A.; Kamide, H. Bull. Chem. Soc. Jpn. 1983, 56, 2065.
67. Ehrenberg, L.; Harms-Ringdahl, M.; Fedorcsa¤k, I.; Granath, F. Acta Chem. Scand.
1989, 43, 177.
68. Hanaki, A. Bull. Chem. Soc. Jpn. 1995, 68, 831.
69. Hanaki, A.; Kinoshita, T.; Ozawa, T. Bull. Chem. Soc. Jpn. 2000, 73, 73.
70. Barton, J.; Packer, J. E. Int. J. Radiat. Phys. Chem. 1970, 2, 159.
71. Ho¡man, M. Z.; Hayon, E. J. Am. Chem. Soc. 1972, 94, 7950.
72. Mezyk, S. P.; Armstrong, D. A. Can. J. Chem. 1989, 67, 736.
73. Pecci, L.; Montefoschi, G.; Musci, G.; Cavallini, D. Amino Acids 1997, 13, 355.
74. Kachur, A. V.; Koch, C. J.; Biaglow, J. E. Free Rad. Res. 1999, 31, 23.
75. Menshikov, S. Y.; Vurasko, A. V.; Molochnikov, L. S.; Kovalyova, E. G.; Efendiev,
A. A. J. Mol. Catal. A 2000, 158, 447.
76. Albro, P. W.; Corbett, J. T.; Shcroeder, J. L. J. Inorg. Biochem. 1986, 27, 191.
77. Kachur, A. V.; Koch, C. J.; Biaglow, J. E. Free Rad. Res. 1998, 28, 259.
78. Brandt, C.; van Eldik, R. Chem. Rev. 1995, 95, 119.
79. Loon, L. V.; Mader, E.; Scott, S. L. J. Phys. Chem. A 2000, 104, 1621.
80. Brandt, C.; Fa¤bia¤n, I.; van Eldik, R. Inorg. Chem. 1994, 33, 687.
81. Huie, R. E.; Neta, P. Atmos. Eviron. 1987, 21, 1743.
82. Buxton, G. V.; McGowan, S.; Salmon, G. A.; Williams, J. E.; Wood, N. D. Atmos.
Environ. 1996, 30, 2483.
83. Das, T. N.; Huie, R. E.; Neta, P. J. Phys. Chem. A 1999, 103, 3581.
84. Connick,R.E.;Zhang,Y.X.;Lee,S.Y.;Adamic,R.;Chieng,P.Inorg.Chem.1995,34,4543.
85. Das, T. N. J. Phys. Chem. A 2001, 105, 9142.
86. Warneck, P.; Ziajka, J. Ber. Bunsenges. Phys. Chem. 1995, 99, 59.
87. Zhang, W. S.; Singh, P.; Muir, D. Hydrometallurgy 2000, 55, 229.
88. Boyce, S. D.; Ho¡mann, M. R.; Hong, P. A.; Molberly, L. M. Environ. Sci. Technol.
1983, 17, 602.
89. Kraft, J.; van Eldik, R. Inorg. Chem. 1989, 28, 2306.
90. Pasiuk-Bronikowska, W.; Bronikowski, T. Chem. Eng. Sci. 1989, 44, 1361.
91. Kraft, J.; van Eldik, R. Atmos. Environ. 1989, 23, 2709.
92. Kraft, J.; van Eldik, R. Inorg. Chem. 1989, 28, 2297.
93. Betterton, E. A. J. Atmos. Chem. 1993, 17, 307.
94. Conklin, M. H.; Ho¡mann, M. R. Environ. Sci. Technol. 1988, 22, 899.
95. Lente, G.; Fa¤bia¤n, I. Inorg. Chem. 1998, 37, 4204.
96. Lepentsiotis, V.; Prinsloo, F. F.; van Eldik, R.; Gutberlet, H. J. Chem. Soc., Dalton.
Trans. 1996, 2135.
97. Prinsloo, F. F.; Brandt, C.; Lepentsiotis, V.; Pienaar, J. J.; van Eldik, R. Inorg. Chem.
1997, 36, 119.
98. Brandt, C.; van Eldik, R. Transition Met. Chem. 1998, 23, 667.
99. Carlyle, D. W.; King, E. L. Inorg. Chem. 1970, 9, 2333.
100. van Eldik, R.; von Jouanne, J.; Kelm, H. Inorg. Chem. 1982, 21, 2818.
101. Bazsa, G.; Diebler, H. React. Kin. Catal. Lett. 1975, 3, 217.
102. Magalha‹es, M.E.A.; Lente, G.; Fa¤bia¤n, I. ‘‘Abstracts, 34th Int. Conf. Coord. Chem.’’;
Edinburgh, UK, 2000; p. 277.
103. Lente, G.; Magalha‹es, M. E. A.; Fa¤bia¤n, I. Inorg. Chem. 2000, 39, 1950.
104. Lente, G.; Fa¤bia¤n, I. React. Kin. Catal. Lett. 2001, 73, 117.
106. Lente, G.; Fa¤bia¤n, I. J. Chem. Soc., Dalton Trans. 2002, 778.
108. Hadady, Zs.; Lente, G.; Fa¤bia¤n, I. unpublished results.
109. Grgič, I.; Dovžan, A.; Berčič, G.; Hudnik, V. J. Atmos. Chem. 1998, 29, 315.
110. Grgič, I.; Poznič, M.; Bizjak, M. J. Atmos. Chem. 1999, 33, 89.
111. Wolf, A.; Deutsch, F.; Ho¡mann, P.; Ortner, H. M. J. Atmos. Chem. 2000, 37, 125.
112. Connick, R. E.; Zhang, Y. X. Inorg. Chem. 1996, 35, 4613.
113. Berglund, J.; Fronaeus, S.; Elding, L. I. Inorg. Chem. 1993, 32, 4527.
114. Fronaeus, S.; Berglund, J.; Elding, L. I. Inorg. Chem. 1998, 37, 4939.
115. Grgič, I.; Berčič, G. J. Atmos. Chem. 2001, 39, 155.
116. Lepentsiotis, V.; Domagala, J.; Grgič, I.; van Eldik, R.; Muller, J. G.; Burrows, C. J.
Inorg. Chem. 1999, 38, 3500.
117. Dutta, S. K.; Ferraudi, G. J. Phys. Chem. A 2001, 105, 4241.
118. Song, Y. H.; Yang, C. M..; Kluger, R. J. Am. Chem. Soc. 1993, 115, 4365.
119. Muller, J. G.; Hickerson, R. P.; Perez, R. J.; Burrows, C. J. J. Am. Chem. Soc. 1997,
119, 1501.
120. Hickerson, R. P.; Watkins-Sims, C. D.; Burrows, C. J.; Atkins, J. F.; Gesteland, R. F.;
Felden, B. J. Mol. Biol. 1998, 279, 577.
121. Wietzerbin, K.; Muller, J. G.; Jameton, R. A.; Pratviel, G.; Bernadou, J.; Meunier, B.;
Burrows, C. J. Inorg. Chem. 1999, 38, 4123.
122. Neves, E. A.; Valdes, J.; Klockow, D. Fresenius J. Anal. Chem. 1995, 351, 544.
123. Lima, S.; Bonifa¤cio, R. L.; Azzellini, G. C.; Coichev, N. Talanta 2002, 56, 547.
124. Utaka, M.; Hojo, M.; Fujii, Y.; Takeda, A. Chem. Lett. 1984, 635.
125. Utaka, M.; Takeda, A. J. Chem. Soc., Chem. Comm. 1985, 1824.
126. Balogh-Hergovich, E¤.; Speier, G. J. Mol. Catal. 1992, 71, 1.
127. Balogh-Hergovich, E¤.; Kaizer, J.; Speier, G. J. Mol. Catal. 2000, 159, 215.
128. Barha¤cs, L.; Kaizer, J.; Pap, J.; Speier, G. Inorg. Chim. Acta 2001, 320, 83.
129. Nishinaga, A.; Tojo, T.; Matsuura, T. J. Chem. Soc., Chem. Comm. 1974, 896.
130. Nishinaga, A.; Numada, N.; Maruyama, K. Tetrahedron Lett. 1989, 30, 2257.
131. Nishinaga, A.; Kuwashige, T.; Tsutsui, T.; Mashino, T.; Maruyama, K. J. Chem. Soc.,
Dalton Trans. 1994, 805.
132. Sima¤ndi, L. I.; Ne¤meth, S.; Rumelis, N. J. Mol. Catal. 1987, 42, 357.
133. Sima¤ndi, L. I.; Fu«lep-Poszmik, A.; Ne¤meth, S. J. Mol. Catal. 1988, 48, 265.
134. Szevere¤nyi, Z.; Sima¤ndi, L. I. J. Mol. Catal. 1989, 51, 155.
135. Milaeva, E. R.; Szevere¤nyi, Z.; Sima¤ndi, L. I. Inorg. Chim. Acta 1990, 167, 139.
136. Szevere¤nyi, Z.; Sima¤ndi, L. I.; Iwanejko, R. J. Mol. Catal. 1991, 64, L15.
137. Szevere¤nyi, Z.; Milaeva, E. R.; Sima¤ndi, L. I. J. Mol. Catal. 1991, 67, 251.
138. Sima¤ndi, L. I.; Barna, T.; Szevere¤nyi, Z.; Ne¤meth, S. Pure. Appl. Chem. 1992, 64, 1511.
139. Sima¤ndi, L. I.; Barna, T.; Korecz, L.; Rockenbauer, A. Tetrahedron Lett. 1993, 34, 717.
140. Sima¤ndi, L. I.; Barna, T.; Ne¤meth, S. J. Chem. Soc., Dalton Trans. 1996, 473.
141. Sima¤ndi, L. I.; Sima¤ndi, T. L. J. Mol. Catal. A 1997, 117, 299.
142. Chauhan, S. M. S.; Gulati, A.; Sahay, A.; Nizar, P. N. H. J. Mol. Catal. A 1996, 105, 159.
143. Bajaj, H. C.; Taqui Khan, M. M. React. Kin. Catal. Lett. 1985, 28, 339.
144. Taqui Khan, M. M.; Rao, A. P. J. Mol. Catal. 1986, 35, 237.
145. Taqui Khan, M. M.; Rao, A. P. J. Mol. Catal. 1987, 39, 331.
146. Taqui Khan, M. M.; Rao, A. P.; Shukla, R. S. J. Mol. Catal. 1989, 49, 299.
147. Taqui Khan, M. M.; Rao, A. P.; Shukla, R. S. J. Mol. Catal. 1989, 50, 45.
148. Taqui Khan, M. M.; Shukla, R. S.; Rao, A. P. Inorg. Chem. 1989, 28, 452.
149. O’Sullivan, D. W.; Hanson, A. K., Jr.; Kester, D. R. Mar. Chem. 1995, 49, 65.
150. King, D. W.; Lounsbury, H. A.; Millero, F. J. Environ. Sci. Technol. 1995, 29, 818.
151. Emmenegger, L.; King, D. W.; Sigg, L.; Sulzberger, B. Environ. Sci. Technol. 1998, 32,
2990.
152. King, D. W. Environ. Sci. Technol. 1998, 32, 2997.
153. Mere¤nyi, G.; Lind, J.; Eriksen, T. E. J. Biolumin. Chemilumin. 1990, 5, 53.
154. Mere¤nyi, G.; Lind, J.; Eriksen, T. E. J. J. Am. Chem. Soc. 1986, 108, 7716.
155. Rose, A. L.; Waite, D. T. Anal. Chem. 2001, 73, 5909.
156. Kimura, M.; Tokuda, M.; Tsukahara, K. Bull. Chem. Soc. Jpn. 1994, 67, 2731.
157. Kimura, M.; Seki, K.; Horie, H.; Tsukahara, K. Bull. Chem. Soc. Jpn. 1996, 69, 613.
158. Chen, D. A.; Motekaitis, R. J.; Martell, A. E.; McManus, D. Can. J. Chem. 1993, 71,
1524.
159. Yokota, K.; Yamazaki, I. Biochim. Biophys. Acta 1965, 105, 301.
160. Scheeline, A.; Olson, D. L.; Williksen, E. P.; Horras, G. A.; Klein, M. L.; Larter, R.
Chem. Rev. 1997, 97, 739, and references therein.
161. Strizhak, P. E.; Basylchuk, A. B.; Demjanchyk, I.; Fecher, F.; Schneider, F. W.;
Munster, A. F. Phys. Chem. Chem Phys. 2000, 2, 4721.
162. Jensen, J. H. J. Am. Chem. Soc. 1983, 105, 2639.
163. Roelofs, M. G.; Wasserman, E.; Jensen, J. H. J. Am. Chem. Soc. 1987, 109, 4207.
164. Roelofs, M. G.; Jensen, J. H. J. Phys. Chem. 1987, 91, 3380.
165. Colussi, A. J.; Ghibaudi, E.; Yuan, Z.; Noyes, R. M. J. Am. Chem. Soc. 1990, 112, 8660.
166. Boga, E.; Peintler, G.; Nagypa¤l, I. J. Am. Chem. Soc. 1990, 112, 151.
167. Guslander, J.; Noyes, R. M.; Colussi, A. J. J. Phys. Chem. 1991, 95, 4387.
168. Druliner, J. D.; Wasserman, E. J. Am. Chem. Soc. 1988, 110, 5270.

Sdarticle 010

Uploaded by

Copyright:

Available Formats

Sdarticle 010

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Sdarticle 010

Uploaded by

Copyright:

Available Formats

METAL ION CATALYZED AUTOXIDATION

REACTIONS: KINETICS AND MECHANISMS

ISTVÁN FÁBIÁN and VIKTOR CSORDÁS

Indisputable biological signi¢cance, industrial applications and envi-

ADVANCES IN INORGANIC CHEMISTRY ß 2003 Elsevier Science (USA)

II. General Considerations

Dioxygen is a formidable oxidizing agent with relatively high standard

The redox potential is strongly pH dependent and it decreases consid-

These reduced forms of oxygen exhibit di¡erent redox properties and

The main role of the metal catalysts is to enhance the O2 + e ! O2

Mðnþ1Þþ þ I ! Mnþ þ P ð5Þ

where I and P stand for one or more intermediates and products,

In this case, the actual redox step is preceded by the formation of an

It should be added that MSO2 is not necessarily a mono-nuclear

Mðn1Þþ þ O2 ! Mnþ þ O

This scheme implies that dioxygen is activated only in a secondary step

The general features discussed so far can explain the complexity of

III. Autoxidation of L-Ascorbic Acid

Autoxidation reactions of L -ascorbic acid (H2A) have been the subject

½MðHAÞðO2 Þðn1Þþ ! ½MðHAÞþ ðO

A þ Mnþ ! A þ Mðn1Þþ ð20Þ

Mðn1Þþ þ HO2 þ Hþ ! Mnþ þ H2 O2 ð22Þ

The intramolecular electron transfer leads to fast formation of semi-

The formation of the [M(HA)](n1)+ complex was con¢rmed in indepen-

d½O2 =dt ¼ k24 ½CuðIIÞ½HA ½O2 1=2 ð24Þ

with k24 = 2.2  103 M3/2 s1.

conclusion that a chain mechanism is operative in this reaction, and the

½Cu2 ðHAÞ2 2þ þ O2 )

½CuAðO2 HÞ ! Cu2þ þ A þ HO

2½CuAðO2 HÞ ! 2Cu2þ þ HA þ A þ 2O2

OH þ HA ! A þ H2 O ð35Þ

where A, B, C, D and E are experimentally determined parameters.

Cuþ þ HA ! ½CuðHAÞ ð38Þ

½CuðHAÞO2  þ ½CuðHAÞ ! ½Cu2 ðHAÞ2 2þ þ O2

2Cuþ þ O2 þ 2Hþ ! 2Cu2þ þ H2 O2 ð43Þ

The reaction features a complex pH dependence which was not

½HA ”CuL2 ”O2  ! CuL2 þ A þ O

According to ESR measurements, the semiquinone radical forms at

Cu,Zn superoxide dismutase. Essentially, these observations support a

2½FeII ðTPPSÞ þ O2 ! 2 ½FeIII ðTPPSÞ þ O2

½FeIII ðTPPSÞ þ 2OH ! ½FeIII ðTPPSÞ2 O þ H2 O ð49Þ

½RuIII LðHAÞðn2Þ ! ½RuII Lðn2Þ þ A þ Hþ k51 ð51Þ

½RuIII Lðn3Þ þ A ! ½RuII Lðn2Þ þ A fast ð52Þ

2½RuII Lðn2Þ þ O2 þ 2Hþ ! ½RuIII Lðn3Þ þ H2 O2 fast ð53Þ

IV. Autoxidation of Catechols and Related Compounds

The kinetics and mechanisms of autoxidation reactions of catechols

In a thorough study by Balla and coworkers, it was shown that in

At longer reaction times, the formation of an acidic product, probably

suggests the formation of copper(I) as an intermediate.This conclusion is

of the substrate with Cu(II) is an important part of the overall reaction

FIG. 1. Amperometric monitoring of the autoxidation of epigallocatechin gallete in

Cuþ þ O2 ! Cu2þ þ O

In spite of the signi¢cant di¡erences in the catalysts and conditions

H2 DTBC þ 1=2O2 ! DTBQ þ H2 O ð66Þ

The yield of quinone was practically 100% and H2O2 accumulated in

product of this reaction, but it cannot accumulate because of fast

FIG. 2. ESR spectra of free-radical intermediates during the cobaloxime(II) cata-

superoxo complex to which the substrate is bonded via a hydrogen

CoIII OH þ DTBSQ þ Hþ ! CoII þ DTBQ þ H2 O ð73Þ

½MðHAÞðO2 Þðn1Þþ ! ½MðHAÞþ ðO

d½O2 =dt ¼ k24 ½CuðIIÞ½HA ½O2 1=2 ð24Þ

with k24 = 2.2 103 M3/2 s1.

½Cu2 ðHAÞ2 2þ þ O2 )

½CuAðO2 HÞ ! Cu2þ þ A þ HO

2½CuAðO2 HÞ ! 2Cu2þ þ HA þ A þ 2O2

Cuþ þ HA ! ½CuðHAÞ ð38Þ

½CuðHAÞO2 þ ½CuðHAÞ ! ½Cu2 ðHAÞ2 2þ þ O2

½HA ”CuL2 ”O2 ! CuL2 þ A þ O

2½FeII ðTPPSÞ þ O2 ! 2 ½FeIII ðTPPSÞ þ O2

½FeIII ðTPPSÞ þ 2OH ! ½FeIII ðTPPSÞ2 O þ H2 O ð49Þ

½RuIII LðHAÞðn2Þ ! ½RuII Lðn2Þ þ A þ Hþ k51 ð51Þ

½RuIII Lðn3Þ þ A ! ½RuII Lðn2Þ þ A fast ð52Þ

2½RuII Lðn2Þ þ O2 þ 2Hþ ! ½RuIII Lðn3Þ þ H2 O2 fast ð53Þ

½FeIV O þ H2 DTBC ! FeII þ DTBQ þ H2 O ð78Þ

Ligand % 102k 102k 102k Hz b Sz b

k96 ¼ 3:9 105 M1 s1 ð96Þ

d½O2 =dt ¼ ka102 ½HSO 2 b c

PhCO þ O2 ! PhCOOO k122 ¼ 5:0 108 M1 s1 ð122Þ