Kinetics and mechanism of the oxidation of water soluble porphyrin Fe IIITPPS with hydrogen peroxide and the peroxomonosulfate ion

Article abstract of DOI:10.1039/b708961a

The overall six-electron oxidation of water soluble porphyrin Fe ^IIITPPS by hydrogen peroxide and peroxomonosulfate ion was studied by the stopped-flow method with UV-vis detection. A three-step consecutive reaction was observed with two intermediates: Fe^IIITPPS → Int₁ → Int₂ → products. The products were identified as the iron(iii) complex of the biliverdin analog formed from TPPS and 4-sulfobenzoic acid. All the rate constants with both oxidizing agents were determined. Intermediate Int₁ is proposed to be the species (TPPS⁺)Fe^IVO. Although no unambiguous proposal for the structure of Int₂ can be made, it is most probably the product of the four-electron oxidation of the original Fe^IIITPPS, contains an iron-oxo center and has a dissociable proton with a pK of around 3.1. In spite of the protolytic equilibria occuring in the pH region 2-4, the kinetic observations do not show pH dependence. The Royal Society of Chemistry.

Full text of DOI:10.1039/b708961a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Kinetics and mechanism of the oxidation of water soluble porphyrin
FeIIITPPS with hydrogen peroxide and the peroxomonosulfate ion†
Ga´bor Lente* and Istva´n Fa´bia´n
Received 13th June 2007, Accepted 9th July 2007
First published as an Advance Article on the web 2nd August 2007
DOI: 10.1039/b708961a
The overall six-electron oxidation of water soluble porphyrin FeIIITPPS by hydrogen peroxide and
peroxomonosulfate ion was studied by the stopped-flow method with UV-vis detection. A three-step
consecutive reaction was observed with two intermediates: FeIIITPPS → Int1 → Int2 → products. The
products were identified as the iron(III) complex of the biliverdin analog formed from TPPS and
4-sulfobenzoic acid. All the rate constants with both oxidizing agents were determined. Intermediate
•
Int1 is proposed to be the species (TPPS +)FeIV O. Although no unambiguous proposal for the
=
structure of Int2 can be made, it is most probably the product of the four-electron oxidation of the
original FeIIITPPS, contains an iron-oxo center and has a dissociable proton with a pK of around 3.1.
In spite of the protolytic equilibria occuring in the pH region 2–4, the kinetic observations do not show
pH dependence.
confirming that the stabilized and structurally studied compounds
Introduction
are the same as the intermediates appearing in actual catalytic or
biochemical processes, which usually occur in aqueous solution
at around room temperature. One possible strategy to resolve this
problem, which has been used in some earlier reports,10,11 requires
detailed kinetic studies of the corresponding processes under
conditions that are catalytically relevant. These investigations
could provide at least indirect information on the reactivities of
the intermediates involved and the kinetic role of the structurally
characterized ‘stabilized intermediates’.
Our interest in the destructive chemical oxidation of
chlorophenols12–16 by H2O2 led us to study the water soluble
porphyrin FeIIITPPS (formula is given in Chart 1), which is
an efficient catalyst in this process.14–16 A careful approach to
exploring the mechanism of the catalysis should begin with
studying independently all the possible subsystems of the three-
component chlorophenol–H2O2–catalyst system. Chlorophenols
High-valent iron compounds appearing as intermediates in
catalytic oxidation cycles have been studied very intensively
recently.1–9 These studies are usually inspired by both the
widespread practical catalytic applications of iron complexes and
the important biochemical role iron plays in heme and non-heme
enzymes. In order to characterize the structure of high-valent iron
intermediates thoroughly, it is usually necessary to stabilize them,
e.g. by lowering the temperature or clever selection of solvent
and/or ligand used. Independent efforts should be devoted to
University of Debrecen, Department of Inorganic and Analytical Chemistry,
Debrecen 10, P.O.B. 21, Hungary, H-4010. E-mail: lenteg@delfin.unideb.hu;
Fax: +36 52 489-667; Tel: +36 52 512-900/22373
† Electronic supplementary information (ESI) available: kinetic traces
of the reactions; ESI-MS of reactants and products; mathematical
background to eqn (4). See DOI: 10.1039/b708961a
Chart 1
4268 | Dalton Trans., 2007, 4268–4275
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
do not react with H2O2 directly (hence the need for a catalyst),
and our results indicated that FeIIITPPS does not interact directly
with chlorophenols, either. Therefore, the chain of catalytic events
must start with the reaction between FeIIITPPS and H2O2. This
reaction is highly important because it is in the core of the
catalytic process, but it may also result in irreversible oxidation,
and as a consequence, inactivation of the catalyst. Ultimately, the
interplay of these somewhat counteracting processes determines
the efficiency of the catalyst.
Although results on many iron porphyrin catalyzed oxidation
reactions have already been published,17–23 exploring the reaction
between the catalyst and the oxidant without any third reagents
has not received the attention it deserves. A notable exception is
an earlier study on the direct reaction of FeIIITPPS with H2O2
in slightly basic solution.21 These results are not very relevant
with respect to catalytic activity because, under these conditions,
FeIIITPPS exists exclusively as an oxo-bridged dimer24,25 with
diminished reactivity.21 In addition, more data are available for the
reactions ofdifferentiron porphyrins withalkyl-peroxide oxidizing
agents in non-aqueous solvents,26,27 but the results of these studies
are not directly transferable to aqueous conditions. In this paper,
we report our detailed mechanistic studies on the FeIIITPPS–H2O2
reaction in aqueous solution and in the absence of any oxidizable
reagents under conditions where the monomeric parent complex
dominates (pH < 7). Detailed experiments with another peroxide
type oxidant, peroxomonosulfate ion (HSO5−), were also carried
out in both the absence and presence of an oxidizable chlorophenol
in order to draw more convincing mechanistic conclusions.
diode-array spectrophotometers. Fast kinetic measurements were
performed with an Applied Photophysics DX-17 MV sequential
stopped-flow apparatus using 1 cm optical path length. The dead
time of the instrument was measured to be 1.09 0.02 ms by pub-
lished methods.29,30 In some experiments, an Applied Photophysics
PD.1 Photodiode Array accessory was used as a detector of the
stopped-flow instrument. Quantitative kinetic measurements were
also performed in the HP-8543 diode-array spectrophotometer
using an RX-2000 Applied Photophysics Rapid Kinetics Acces-
sory. A YSI 5100 Dissolved Oxygen Meter and a YSI Model 5239
probe with YSI 5906 membrane cap were used for measuring the
concentration of dissolved oxygen in aqueous solutions. In some
experiments, a Weiss Research CL3005 combination chloride ion
selective electrode was used (constant ionic strength maintained
with 0.10 M NaNO3) connected to a Hanna Instruments pH302
pH-meter. A GK2401C combination glass electrode was used to
measure pH attached to radiometer PHM85 pH-meter. The pH-
meter reading was converted into log[H+].31 Mass spectrometric
analysis was performed by a Bruker micrOTOF-Q instrument with
electrospray ionization in the negative ion mode. The software
SCIENTIST,32 ZiTa,33 and Matlab34 were used in the fitting
procedures.
Results
Spectral observations
When FeIIITPPS is mixed with H2O2 or HSO5 in acidic solution,
−
the intense color of the iron complex fades. The Soret band of the
porphyrin ring disappears showing that the complex is oxidized
to a product that does not have a porphyrin ring. A series of time
resolved UV-vis spectra recorded during the reaction of FeIIITPPS
with H2O2 is shown in Fig. 1. Isosbestic points detectable at early
reaction times (at about 360, 440, 510, and 550 nm) disappear
giving unambiguous evidence that a multistep redox process is
observed.
Experimental
Materials
FeIIITPPSCl was purchased from Frontier Scientific. The iron
content of a FeIIITPPSCl stock solution (25.3 mg dm−3) was mea-
sured by ICP-AES to be 23.0 1.0 lM in acceptable agreement
with the calculated value (24.7 lM), which confirmed that the
sample was of sufficient purity. Hydrogen peroxide stock solu-
tions were standardized by permanganometry every day. H2O2-
containing samples were prepared freshly before use. Potassium
peroxomonosulfate stock solutions were prepared every day from
oxone (2KHSO5·KHSO4·K2SO4, Aldrich) and standardized by
iodometric titration. In this case, it was confirmed that the chloride
content of FeIIITPPSCl did not interfere with the measurements
as chloride ion is oxidized by HSO5−, much slower (several hours)
than the metal complex (less than a minute). Perchloric acid was
used as a strong acid to set the pH, the ionic strength was set with
NaClO4, which was prepared as described earlier.28 Extra care was
taken to make sure that the oxidant concentration is sufficiently
low to avoid the precipitation of KClO4 in the case of oxone,
which contains K+ ions. All other chemicals used in this study
were of analytical reagent grade and purchased from commercial
sources. Doubly deionized and ultrafiltered water from a Millipore
Q system was used to prepare the stock solutions and samples.
Fig. 1 Spectral observations during the reaction of FeIIITPPS with H2O2.
[FeIIITPPS] = 9.5 lM; [H2O2] = 13.2 mM; pH = 3.05; optical path length =
1.00 cm; T = 25.0 ◦C; I = 0.1 M (NaClO4); reaction times: increasing time
intervals from 0.1 s to 150 s.
The kinetic curves recorded at 520 nm at varying pH in the
range 3.7–2.1 clearly demonstrate that the observations are pH-
dependent (Fig. 2). The origin of this pH effect will be explained
later. Four of the five kinetic curves in Fig. 2 show an absorbance
decrease first, then an increase for some time, and finally a decrease
again. Therefore, both a minimum and a maximum are present in
Instrumentation and computation
UV-vis spectra and kinetic curves were recorded on Perkin Elmer
Lambda 25 or Perkin Elmer Lambda 2 S scanning, and HP-8543
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 4268–4275 | 4269
©

View Article Online
The same was not possible for H2O2, where the first phase was
not separate, but initial rates could be obtained as a function of
wavelength. After normalization, which means division of each
value with the largest absolute value in a given series of data,
the comparison of these two quantities leads to an important
conclusion (Fig. 4, Fig. S1 in the ESI† shows the vis part of the
same graph with an expanded y axis). The wavelength dependences
of the two quantities are very similar, clearly suggesting that the
same absorbing species are involved in the first phase in both
cases and the FeIIITPPS is initially oxidized to the same absorbing
−
intermediate by both H2O2 and HSO5
.
Fig. 2 Kinetic traces during the reaction of FeIIITPPS with H2O2.
[FeIIITPPS] = 9.5 lM; [H2O2] = 13.2 mM; pH = 3.68 (a), 3.05 (b), 2.76 (c),
2.37 (d), 2.16 (e); k = 520 nm; optical path length = 1.00 cm; T = 25.0 ◦C;
I = 0.1 M (NaClO4). Solid lines are fitted curves using eqn (1).
each of these curves, which unambiguously prove that the studied
reaction sequence includes at least two intermediates and at least
three consecutive reaction steps. The oxidizing agent was used
in a large (usually more than 100-fold) excess in all experiments,
therefore its concentration does not change during the process and
multiexponential curves are expected to be seen if all the reactions
are first order with respect to FeIIITPPS and the intermediates.35 In
agreement with this expectation, kinetic traces with both oxidizing
agents gave an excellent fit to a treble exponential curve:
Fig. 4 Comparison of normalized absorbance changes for the first phase
(a, with HSO5−) and normalized initial rates (b, with H2O2) as a function
of wavelength during the reaction of FeIIITPPS with HSO5 and H2O2.
−
t
t
t
obs3
A = A1e−k + A2e−k + A3e−k + E
(1)
obs1
obs2
[FeIIITPPS] = 6.9 lM (a), 9.5 lM (b); [HSO5−] = 2.95 mM (a); [H2O2] =
8.77 mM (b); pH = 1.30 (a), 3.05 (b); optical path length = 1.00 cm; T =
25.0 ◦C; I = 0.1 M (NaClO4).
The lines in Fig. 2 show the best fits for curves recorded with
H2O2 as an oxidizing agent, Fig. 3 shows a kinetic curve recorded
with the HSO5− ion and the best fit to a treble exponential function.
With H2O2, the oxidation proceeded without giving well separated
phases on the time scale of a few minutes. With HSO5−, the reaction
was faster, usually complete in about 20 s and the first phase of
the process was separate from the other two overlapping phases.
As expected, fitting the first phase only (up to about 50 ms) with
a single exponential function gave, within experimental error, the
same pseudo first order rate constants as the fastest one from the
treble exponential fits. The absorbance change associated with this
first phase could be determined in stopped-flow studies using the
PD.1 diode-array detector with a wavelength resolution of 2.5 nm.
In some experiments, a diode array spectrophotometer was
used for detection. It is well known that the light of a diode-
array instrument has relatively high intensity and can induce
photochemical reactions.13 In fact, the instrument can be used as a
photochemical device for certain purposes,36,37 and neglecting the
possible effect of light of a diode array spectrophotometer led to
seriously erroneous conclusions in some cases.38,39 As porphyrins
absorb light strongly and have widespread photochemistry,40,41
we took precautions to avoid interference from photochemical
processes. When the diode-array spectrophotometer was used for
detection, the time of illumination was the shortest possible. In
a few cases, the same experiments were carried out with longer
overall illumination times as well. Because the kinetic curves did
not depend on the time of illumination, it could be concluded
that photochemical processes do not interfere detectably with the
oxidation process under our conditions.
Quantitative kinetic results: concentration dependence
First, initial rates were measured in the reaction between FeIIITPPS
and H2O2. The initial rate showed first order dependence on the
concentrations of both H2O2 and FeIIITPPS (Fig. S2 and S3 in
ESI†).
The kinetic curves could be fully analyzed through the three
pseudo first order rate constants, kobs1, kobs2, and kobs3 determined
in each experiment. In the case of H2O2, it was not possible to
find a wavelength that was sensitive enough for all of the three
phases (i.e. at least one of them was accompanied by a minor
absorbance change at any given wavelength). However, it was
Fig. 3 Comparison of experimental (markers) and fitted treble exponen-
tial (line) kinetic traces during the reaction of FeIIITPPS with HSO5
.
−
[FeIIITPPS] = 6.9 lM; [HSO5−] = 3.9 mM; pH = 1.30; k = 395 nm; optical
path length = 1.00 cm; T = 25.0 ◦C; I = 0.1 M (NaClO4); only a selection
of measured points (∼ 7%) is shown and logarithmic time scale is used for
clarity.
4270 | Dalton Trans., 2007, 4268–4275
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
possible to extract the three pseudo first order rate constants from
multiwavelength data using global analysis.42 In this approach,
kinetic curves measured at 81 wavelengths (from 300 to 700 nm
with an interval of 5 nm) were fitted simultaneously to treble
exponential curves that could have different amplitudes for each
wavelength but had the same pseudo first order rate constants.
Fig. 5 shows the dependence of kobs1 and kobs2 on the concentration
of H2O2. It is clear that kobs1 is directly proportional to the oxidant
concentration:
was sensitive for all three phases of the reaction in this case.
Therefore, all three rate constants could be determined from treble
exponential kinetic curves measured at 395 nm. All three observed
rate constants were directly proportional to the concentration of
the oxidant (Fig. S5–S7 in ESI†):
S
−
−
−
kobs1 = k1 [HSO5
]
]
]
(6)
(7)
S
kobs2 = k2 [HSO5
kobs1 = kH[H2O2]
(2)
1
S
kobs3 = k3 [HSO5
(8)
S
The second-order rate constants k1 = (2.00
0.02) ×
104 M−1 s−1, k2 = (3.67 0.05) × 102 M−1 s−1, and k3 = (2.85
S
S
0.05) × 101 M−1 s−1 were determined by least squares fitting.
Quantitative kinetic results: pH dependence
FeIIITPPS is known to form an oxo-bridged dimer above pH 7,24,25
which was shown to be catalytically inactive in the H2O2
dependent oxidation of chlorophenols.14 In this work, it was
further confirmed that the oxo bridged dimer TPPSFeIII–O–
FeIIITPPS does not react with H2O2. In experiments carried out in
phosphate buffer at pH ∼ 7.2, no spectral changes were detected
for about 60 min upon the addition of H2O2 to a sample of
FeIIITPPS. Iodometric titrations also confirmed that no significant
amount of the oxidant is lost over a few hours in the same reaction.
In view of these findings, the present study was limited to the
acidic pH range. Concern about possible interference of buffers16
(complexation or oxidation) led us to avoid using them at all.
Therefore, the pH was set with strong acid (usually perchloric
acid) only and no experiments were carried out in the 4.5–6 pH
range. The initial rate was measured in the reaction with H2O2 and
it was independent of pH in the entire pH region studied (1.8–3.8,
Fig. S8 in ESI†). Similarly to the spectral comparison shown in
Fig. 4, the initial rates measured at several different wavelengths
were also compared between experiments at pH = 2.16 and 3.05
(Fig. S9 in ESI†). The same curve was obtained, therefore the pH
did not influence the formation rate or the identity of the product
in the first phase. The pseudo first order rate constants determined
in a series of experiments from the analysis of full curves did not
show any pH-dependence, either. At first sight, this may seem to
contradict the fact that different kinetic traces were measured at
different pH values as shown in Fig. 2. However, the curves in
Fig. 2 give identical pseudo first order rate constants and only the
amplitudes A2 and A3 are significantly different. This implies that
the spectrum and most probably the identity of an intermediate
changes in this pH range (1.8–3.8), but the rate constants remain
the same. This unexpected phenomenon will be discussed in more
detail in the next sections. In the case of HSO5−, experiments were
carried out in the pH range 0.5–2, and the kinetic curves did not
depend on the pH in this range (Fig. S10 in ESI†). The absence of
a kinetic pH effect can be rationalized easily: no significant acid–
base equilibria involving any of the reagents, H2O2, HSO5−, and
FeIIITPPS occur in the pH range of the present study. It follows
that if the reactions occur between the dominant species, no pH-
dependence arises.
Fig. 5 Dependence of pseudo first order rate constants kobs1 and kobs2 on
the oxidant concentration during the reaction of FeIIITPPS with H2O2.
[FeIIITPPS] = 9.5 lM; pH = 3.05; T = 25.0 ◦C; I = 0.1 M (NaClO4).
Linear least squares fitting yielded the second-order rate
constant k1 = 15.1 0.1 M−1 s−1. Fig. 5 shows that kobs2 depends
H
on the oxidant concentration linearly, but there is a significant
intercept. Therefore, the corresponding equation is
kobs2 = kH [H2O2] + kH
(3)
2a
2b
H
H
Parameters k2a = 2.23
0.09 M−1 s−1 and k2b = 0.060
0.002 s−1 were resolved by the least squares fitting procedure.
It should be noted that kobs1 and kobs2 are very close at the lowest
H2O2 concentration shown in Fig. 5. In fact, it was not possible
to fit the experimental kinetic curves measured at this point to
a treble exponential function. This is not very surprising as the
analytical solution of the linear, first order differential equation
for the special case kobs1 = kobs2 is not a treble exponential function
as given in eqn (1). The solution contains a term which is composed
of a polynomial of the independent variable (t) multiplied with an
exponential function. The correct mathematical form is (see more
detailed mathematical background on pages S11–S12 of ESI†):
t
t
obs3
A = (A1 + A2t)e−k
+ A3e−k + E
(4)
obs12
The value of kobs12 = kobs1 = kobs2 was fitted using eqn (4) for this
special case.
The third pseudo first order rate constant, kobs3 of the process
depended linearly on the oxidant concentration (Fig. S4 in ESI†):
H
kobs3 = k3aH[H2O2] + k3b
(5)
H
H
The values k3a = 0.30 0.03 M−1 s−1 and k3b = (7.6 0.6) ×
10−3 s−1 were determined.
As already noted, the first phase of the reaction was separate
from the others during the oxidation with HSO5−. Quantitative
measurements were carried out at a single wavelength by the
stopped-flow method. The 395 nm Soret band of FeIIITPPS
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 4268–4275 | 4271
©

View Article Online
Stoichiometric observations
It is well known that the biochemical pathway of porphyrin
degradation yields biliverdin and carbon monoxide.43 We note
that this oxidation scheme is an overall six electron oxidation
or three consecutive oxygen transfer steps. This suits nicely our
kinetic observations, which showed that a three-step process is
operative. Thus, it was assumed that the products of the oxidation
of FeIIITPPS are the iron(III) complex of the biliverdin analog
formed from TPPS (P1 in Chart 1) and a carboxylic acid (P2 in
Chart 1), which is the analog of carbon monoxide when the carbon
atom in the porphyrin ring has a substituent.
It was noted that some acid is formed in the oxidation reaction
in unbuffered solution. Experiments were designed to study the
stoichiometry of this acid production: the pH of an unbuffered
solution was followed after the addition of H2O2 to FeIIITPPS.
Fig. 6 shows the pH as a function of time. In the beginning,
the sample is slightly acidic because FeIIITPPS has four strongly
acidic sulfonic acid groups which are ionized completely. The pH
decreases as the reaction progresses, and the concentration of acid
formed can be calculated for each point from the measured data.
These experiments showed that approximately one equivalent acid
is formed per FeIIITPPS, which means that the final amount of
extra acid equals the initial amount of FeIIITPPS. The sulfonic acid
groups are conserved without modification during the reaction
and remain ionized. Therefore, they cannot be responsible for
acid production. The source of acidification is sulfobenzoic acid
P2, in which the carboxylic group is newly formed as a result of
the oxidation reaction. Independent literature data show that this
carboxylic acid has a pK of 3.65,44 and it is mostly ionized at the
final pH of the experiment shown in Fig. 6. These considerations
are in agreement with the production of a single equivalent of
acid in the reaction. Similar pH-metric experiments were not
Fig. 7 ESI mass spectra of the products formed in the reaction of
FeIIITPPS with HSO5−. Large graphs are the measured spectra, insets
are simulated spectra for the ion with the composition shown. The spectra
are consistent with the formula of P1 and P2 (Chart 1).
3−
characteristic of the ions C37FeH20N4O11S3 (the dominant form
−
possible with the HSO5 ion because the reagent used to prepare
of P1 at pH = 1) and C7FeH5O5S− (the dominant form of P2 at
pH = 1) were both clearly recognizable. This lends strong support
to our assumed stoichiometry, and together with the quantitative
pH measurements and the clean kinetic observations, confirms
that no other significant products are formed during the reaction.
When the oxidants were used in larger excess, slow further
reactions were also observed. These may have involved further
oxidation of P1 or P2, but were too slow to interfere with our
kinetic measurements.
−
the solutions, oxone, also contains HSO4 ions, the dissociation
of which produces further acid, and makes reliable quantitative
stoichiometric studies difficult to perform.
Spectrophotometric titration was also attempted to establish the
oxidant/FeIIITPPS stoichiometric ratio experimentally. However,
these experiments remained inconclusive. In the case of the reac-
tion with H2O2, clear evidence was obtained that decomposition
of H2O2 to water and oxygen also occurs. This was confirmed
by experiments with a dissolved oxygen electrode, which showed
a slight increase in oxygen concentration during the reaction.
Therefore, H2O2 is consumed in two different reactions and
it is not possible to characterize the stoichiometry of one of
them alone. The loss of H2O2 was negligible compared to the
initial concentration during the kinetic runs. Thus, pseudo first
order conditions were maintained in all kinetic experiments. The
dissolved oxygen electrode experiments also proved that O2 does
not have a significant role in the process as an oxidant because
its concentration did not decrease. Evidence for some catalytic
oxidant decomposition was also obtained in the experiments
Fig. 6 The change in pH (left axis) and the amount of acid produced
(right axis) during the reaction of FeIIITPPS with H2O2. [FeIIITPPS] =
17.3 lM; [H2O2] = 25 mM; T = 25.0 ◦C; I = 0.2 M (KCl).
Further identification of reaction products was attempted with
ESI mass spectroscopy in negative ion mode. The oxidation
reaction was driven to completion by adding about 5 equivalents
of HSO5 to a solution of FeIIITPPS at about pH 1, and the final
−
−
solution was analyzed. As shown in Fig. 7, the sets of MS peaks
with HSO5 as the spectral observations during titrations were
4272 | Dalton Trans., 2007, 4268–4275
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
dependent on the increment size of oxidant addition and not only
on the amount of oxidant totally added.
It is much more difficult to make a reasonable proposal for
the structure of Int2. This species is the product of the two
electron oxidation of Int1. Analysis of spectral data showed that the
absorbance contribution of Int2 is rather significant above 500 nm.
Therefore, once the assignment of rate constants to processes was
successful, the UV-vis spectrum of Int2 could be resolved with
acceptable precision in this wavelength region. Fig. 8 shows the
resolved molar absorption spectra of Int2 at 6 different pH values.
The spectrum of Int2 seems to depend on pH quite strongly. This
explains the earlier observation that the amplitudes of the fitted
treble exponential functions depend on the pH whereas the rate
constants do not (see Fig. 2). The pH variation of spectra shown
in Fig. 8 can be readily interpreted by assuming that Int2 contains
a dissociable proton. Data measured at 15 different wavelengths
were simultaneously fitted to the following equation:
Discussion
As shown in the previous sections, FeIIITPPS is oxidized in a
consecutive three-step process which can be represented by the
simple series of reactions shown in Scheme 1. The fact that the
detected curves are treble exponential proves that the first, second,
and third reactions are first order with respect to FeIIITPPS, Int1,
and Int2, in order. It is well known that the fastest pseudo first
order rate constant obtained from the treble exponential fit does
not necessarily correspond to the first reaction step in Scheme 1.35
The assignment of rate constants to reactions should always be
based on independent considerations. Usually, the values of the
molar absorbances required by a particular assignment provide a
good way to do this because assignments requiring unreasonably
high molar absorbances can be ruled out.35
HInt
+
eIint K + ei [H ]
2
2
ei =
(9)
K + [H+]
Scheme 1
For HSO5−, it was confirmed that the fast pseudo first order
rate constant belongs to the first step, the middle rate constant
to the second step, the slow rate constant to the third step.
Any other assignment led to impossibly high (>>108) molar
absorbances for one or both of the intermediates Int1 and Int2.
Similar considerations showed that the slow rate constant must
correspond to the third step in the case of H2O2, but the first
two could not be assigned on the basis of similar considerations
(Fig. 5 shows that kobs1 and kobs2 are usually close to each other). A
differentmethod was used to clarifythis problem:the second-order
rate constant calculated from the initial rates was compared with
Fig. 8 pH dependence of the absorption spectrum of intermediate Int2
formed during the reaction of FeIIITPPS with H2O2. pH = 3.68 (a), 3.05
(b), 2.76 (c), 2.37 (d), 2.16 (e), 1.95 (f); T = 25.0 ◦C; I = 0.1 M (NaClO4).
In eqn (9), K is the acid dissociation constant of Int2. The fitting
H
k1 calculated from the dependence of kobs1 on the concentration
gave pK = 3.08
0.02, and showed that the protonated form
of H2O2. The two values agreed within experimental error showing
that kobs1 corresponds to the first step. It follows that kobs2 belongs
to the second reaction step in the sequence. It should be noted
that this order (fastest phase is first, middle phase is second) can
change at low oxidant concentration, as also apparent from Fig. 5.
It was shown in a previous section that the identity of the
of Int2, which dominates in more acidic solution, has stronger
absorption above 500 nm. The value of the fitted pK also explains
why the shape of the kinetic curves did not depend on pH in the
case of HSO5−, where the pH range was narrower (0.5–2). It is
interesting to note that although this protonation/deprotonation
influences the spectral properties of Int2 significantly, there is no
detectable effect on the reactivity, i.e. the two forms seem to react
with the oxidizing agent at basically the same rates. This might
imply that the dissociable proton could be far from the site where
the oxidant attacks Int2.
−
intermediate Int1 does not depend on whether HSO5 or H2O2
is used as an oxidant. This fact has important consequences
for proposing a structure for Int1. When H2O2 is the oxidant, it
would be natural to assume that an end-on hydroperoxo complex,
TPPSFeIII–OOH,5 forms at first, the decomposition of which
can give rise to a species with a higher oxidation state metal
center. However, formation of the hydroperoxo complex seems
As noted previously, the pseudo first order rate constants
−
measured in the HSO5 system were directly proportional to
the oxidant concentration. The same was not true for the H2O2
system, where the dependences of kobs2 and kobs3 on the oxidant
concentration were linear but showed rather significant intercepts.
These intercepts imply that the oxidation of Int1 to Int2 and the
oxidation of Int2 to P1 and P2 can also proceed via a pathway where
H2O2 only attacks after the rate determining steps. These pathways
are simple first-order reactions of Int1 and Int2, and consequently
−
rather unlikely when HSO5 is used as the oxidant. Since the two
processes have the same detectable intermediates, the hydroperoxo
complex cannot be a spectroscopically detectable intermediate
in the reaction with H2O2, although it may still be a steady-
state intermediate whose concentration is not high enough to
be detected. Thus, Int1 is most likely a formally iron(V) oxo
species, which is often postulated to form in catalytic redox cycles
containing porphyrins.17–23,45–48 The actual structure of this species
is thought to be iron(IV) oxo with one electron removed from the
−
must be present in the HSO5 system as well. However, their
contribution to the overall rate is negligible because the direct
reaction with the oxidant is much faster here. In the case of Int1,
it is easy to imagine a process in which the oxo group on the iron
•
45,46
porphyrin ring as well, (TPPS +)FeIV O.
=
This journal is The Royal Society of Chemistry 2007
Dalton Trans., 2007, 4268–4275 | 4273
©

View Article Online
center is transferred to some part of the porphyrin ring without
the assistance of an oxidant molecule, and the resulting species is
oxidized by H2O2 quite rapidly to give Int2. A similar sequence of
steps is only possible in the oxidation of Int2 if it contains an iron-
oxo center. This may suggest that the structure of Int2 contains
a high-valent iron-oxo center and a hydroxyl group, which might
be responsible for the pK observed. A possible, though clearly
speculative structure that interprets all the listed features for Int2 is
shown in Chart 2. We note that further structural characterization
of Int2 does not seem to be feasible as it has a fast first-order
decay (0.5 min−1) even when no oxidizing agent is present and
always coexists with Int1 and the products. Compounds similar
to a-meso-hydroxyheme, which are often implied as intermediates
in heme metabolism,47,48 cannot be intermediates in this system
because there are no hydrogens in FeIIITPPS in a suitable position
to form them.
The experimental kinetic traces could not be fitted to any of
the explicit rate equations commonly used in reaction kinetics.
The set of differential equations representing the kinetic model
given in eqn (10)–(12) was solved numerically and a non-linear
least-square fitting algorithm was used to find the best values of
the fitted parameters.33,49 In these calculations, k1 and k2 were
included with fixed values reported in the previous section of this
paper, while rate constant kO and the molar absorbances of Int1
and DCQB were fitted. The following values were obtained for
these parameters: kO = (6.4 0.5) × 106 M−1 s−1, e(Int1, 395 nm) =
(1.4 0.5) × 104 M−1 cm−1, e(DCBQ, 395 nm) = (4.4 0.5) ×
103 M−1 cm−1. Fig. 9 shows two of the measured curves along with
the best fit obtained. The excellent agreement of the experimental
and calculated data demonstrates the validity of the model and
lends strong support to the conclusions reached in this study.
S
S
Chart 2
To conclude this study, control experiments were designed to
test the relevance of the results in FeIIITPPS catalyzed oxidation
of 2,4,6-trichlorophenol (TCP) with HSO5−. The process was first
followed using a chloride ion selective electrode. The reaction
produced one equivalent of chloride ion within a minute (Fig. S11
in ESI†) and further production of Cl− was not observed at
longer reaction times. UV-vis spectra confirmed that the primary
oxidation product with HSO5−, similar to the analogous H2O2
system,14 is 2,6-dichloro-1,4-benzoquinone (DCBQ). In a series
of stopped-flow experiments, the reaction was followed at the
characteristic 395 nm Soret band of the catalyst. The reaction was
complete well within 1 min but the disappearance of FeIIITPPS
became slower when the concentration of TCP was increased.
This indicates that the substrate either inhibits the oxidation of
the catalyst or regenerates it. The observations can be rationalized
in terms of the following simple kinetic model
Fig. 9 Comparison of experimental (markers) and fitted (lines) kinetic
traces during the reaction oxidation of 2,4,6-trichlorophenol with HSO5
catalyzed by FeIIITPPS. [FeIIITPPS] = 6.9 lM; [HSO5−] = 4.9 mM;
[2,4,6-trichlorophenol] = 43 lM (a), 86 lM (b); pH = 1.30; k = 395 nm;
optical path length = 1.00 cm; T = 25.0 ◦C; I = 0.1 M (NaClO4); only a
selection of measured points (∼ 7%) is shown for clarity.
Earlier, we studied the FeIIITPPS-catalyzed oxidation of TCP
using H2O2 as the oxidant and proposed that the substrate is
oxidized by an iron-containing steady-state intermediate (Catꢀ),
which was assumed to be the hydroperoxo complex, TPPSFeIII–
OOH.14 The comparison of these data14 with present results reveals
an interesting point. The FeIIITPPS-catalyzed oxidations of the
−
same substrate by H2O2 and HSO5 proceed via slightly different
mechanisms because Catꢀ cannot be the same as Int1 identified
in this work. First, steady-state kinetics clearly implied that the
decomposition of Catꢀ is a first-order process that does not depend
on the concentration of H2O2,14 whereas Int1 is oxidized through a
dominantly H2O2-dependent pathway (k2aH and k2bH in Scheme 1).
Second, the rate constants for both the oxidation of Int1 (k2aH and
k2bH) and the reaction between Int1 and TCP (kO) determined in
this work seem too low for similar reactions of Catꢀ. For example,
model calculations showed that k2bH should be greater than 10 s−1
for steady-state conditions to apply in the H2O2 oxidation if Catꢀ
and Int1 were the same. It seems likely that the catalytically most
active intermediate in the H2O2 system (Catꢀ) is TPPSFeIII–OOH
as postulated in our earlier work.14 This species is expected to
from as an immediate product in the reaction of FeIIITPPS with
H2O2 (but not with HSO5−) and can undergo fast heterolytic O–
O bond cleavage to give Int1, which is a detectable intermediate.
Int1 is not very important in the catalytic cycle in the presence of
−
FeIIITPPS + HSO5 → Int1
(10)
(11)
(12)
v10 = k1 [FeIIITPPS][HSO5
]
S
−
Int1 + TCP → FeIIITPPS + DCBQ
v11 = kO[Int1][TCP]
−
Int1 + HSO5 → Int2
S
−
v12 = k2 [Int1][HSO5
]
•
Intermediate Int1, (TPPS +)FeIV O, is either reduced back to
FeIIITTPS by TCP or oxidized to Int2 by the oxidant. The latter
process irreversibly deactivates the catalyst.
=
4274 | Dalton Trans., 2007, 4268–4275
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
H2O2 because it is a lot less reactive toward TCP than Catꢀ, and
its concentration is never as high as in the HSO5 system.
16 G. Lente and J. H. Espenson, Green Chem., 2005, 7, 28–34.
−
17 T. G. Traylor and F. Xu, J. Am. Chem. Soc., 1987, 109, 6201–6202.
18 T. C. Bruice, P. N. Balasubramanian, R. W. Lee and J. R. L. Smith,
J. Am. Chem. Soc., 1988, 110, 7890–7892.
19 G. Labat, J. L. Se´ris and B. Meunier, Angew. Chem., Int. Ed. Engl.,
1990, 29, 1471–1473.
Conclusion
20 W. Nam, M. H. Lim, H. J. Lee and C. Kim, J. Am. Chem. Soc., 2000,
122, 6641–6647.
21 V. Lepentsiotis, R. van Eldik, F. F. Prinsloo and J. J. Pienaar, J. Chem.
Soc., Dalton Trans., 1999, 2759–2767.
22 T. K. Saha, S. Karmaker and K. Tamagake, Luminescence, 2003, 18,
259–267.
23 N. A. Stephenson and A. T. Bell, Inorg. Chem., 2007, 46, 2278–2285.
24 E. B. Fleischer, J. M. Palmer, T. S. Srivastava and A. Chatterjee, J. Am.
Chem. Soc., 1971, 93, 3162–3167.
25 A. A. El-Awady, P. C. Wilkins and R. G. Wilkins, Inorg. Chem., 1985,
24, 2053–2057.
26 J. T. Groves and Y. Watanabe, J. Am. Chem. Soc., 1988, 110, 8443–8452.
27 J. T. Groves, Z. Gross and M. K. Stern, Inorg. Chem., 1994, 33, 5065–
5072.
28 G. Gordon and P. Tewari, J. Phys. Chem., 1966, 60, 200–204.
29 B. Tonomura, H. Nakatani, M. Ohnishi, J. Yamaguchi-Ito and K.
Hiromi, Anal. Biochem., 1978, 84, 370–383.
This study shows that studying the kinetics of the reaction between
an oxidant and a catalyst without any oxidizable substrate is
very important even when the primary objective is to explore the
catalytic oxidation reaction. In the FeIIITPPS–H2O2–chlorophenol
system, the most active form of the catalyst is clearly not the same
as the first spectroscopically detectable intermediate of the reaction
in the absence of chlorophenol, which gives the general warning
that isolated and spectroscopically or structurally characterized
intermediates of a reaction may in fact have little relevance in
catalytic cycles. Reactive species should not be identified solely
based on spectral or structural studies, kinetic data are also
essential.
30 G. Peintler, A. Nagy, A. K. Horva´th, T. Ko¨rtve´lyesi and I. Nagypa´l,
Phys. Chem. Chem. Phys., 2000, 2, 2575–2586.
31 H. M. Irving, M. G. Miles and L. D. Pettitt, Anal. Chim. Acta, 1967,
38, 475–488.
Acknowledgements
The authors thank the Hungarian Science Foundation for finan-
cial support under grant No. F049498. Dr Ja´nos To¨ro¨k and
Mr Lajos Nagy are gratefully acknowledged for assistance in
the ESI-MS measurements. GL wishes to thank the Hungarian
Academy of Sciences for a Bolyai Ja´nos Research Fellowship.
32 SCIENTIST, version 2.0, Micromath Software, Salt Lake City, UT,
USA, 1995.
33 G. Peintler, ZiTa, version 4.1, Attila Jo´zsef University, Szeged, Hungary,
1997.
34 MatLab for Windows, version 4.2c1, The Mathworks, Inc., Natick, MA,
USA, 1994.
35 J. H. Espenson, in Chemical Kinetics and Reaction Mechanisms,
McGraw-Hill, New York, 2nd edn, 1995.
References
36 I. Kerezsi, G. Lente and I. Fa´bia´n, J. Am. Chem. Soc., 2005, 127, 4785–
1 A. Sorokin, J. L. Se´ris and B. Meunier, Science, 1995, 268, 1163–1166.
2 T. J. Collins, Science, 2001, 291, 48–49.
4793.
37 I. Kerezsi, G. Lente and I. Fa´bia´n, Dalton Trans., 2006, 955–960.
38 D. M. Stanbury and J. N. Figlar, Coord. Chem. Rev., 1999, 187, 223–
232.
3 T. J. Collins, Acc. Chem. Res., 2002, 35, 782–790.
4 S. Sen Gupta, M. Stadler, C. A. Noser, A. Ghosh, B. Steinhoff, D.
Lenoir, C. P. Horwitz, K.-W. Schramm and T. J. Collins, Science, 2002,
296, 326–328.
39 M. Galajda, G. Lente and I. Fa´bia´n, J. Am. Chem. Soc., 2007, 129,
7738–7739.
5 J. A. Kovacs, Science, 2003, 299, 1024–1025.
40 I. Hamachi, S. Tsukiji, S. Shinkai and S. Oishi, J. Am. Chem. Soc., 1999,
121, 5500–5506.
6 J. U. Rohde, J. H. In, M. H. Lim, W. W. Brennessel, M. R. Bukowski,
A. Stubna, E. Mu¨nck, W. Nam and L. Que, Jr., Science, 2003, 299,
1037–1039.
41 J. Premkumar and R. Ramaraj, J. Mol. Catal. A: Chem., 1999, 142,
153–162.
7 J. T. Groves, J. Inorg. Biochem., 2006, 100, 434–447.
8 T. M. Makris, K. von Koenig, I. Schlichting and S. G. Sligar, J. Inorg.
Biochem., 2006, 100, 507–518.
9 F. T. de Oliveira, A. Chanda, D. Banerjee, X. P. Shan, S. Mondal, L.
Que, Jr., E. Bominaar, E. Mu¨nck and T. J. Collins, Science, 2007, 315,
835–838.
10 B. J. Brazeau, R. N. Austin, C. Tarr, J. T. Groves and J. D. Lipscomb,
J. Am. Chem. Soc., 2001, 123, 11831–11837.
11 L. G. Beauvais and S. J. Lippard, J. Am. Chem. Soc., 2005, 127, 7370–
7378.
12 G. Lente and J. H. Espenson, Chem. Commun., 2003, 1162–1163.
13 G. Lente and J. H. Espenson, J. Photochem. Photobiol., A, 2004, 163,
249–258.
14 G. Lente and J. H. Espenson, Int. J. Chem. Kinet., 2004, 36, 449–455.
15 G. Lente and J. H. Espenson, New J. Chem., 2004, 28, 847–852.
42 M. Weitzer, M. Schatz, F. Hampel, F. W. Heinemann and S. Schindler,
J. Chem. Soc., Dalton Trans., 2002, 686–694.
43 L. Stryer, in Biochemistry, W. H. Freeman & Co., New York, 4th edn,
1995.
44 E. L. Wehry and L. B. Rogers, J. Am. Chem. Soc., 1966, 88, 351–354.
45 J. Terner, V. Palaniappan, A. Gold, R. Weiss, M. M. Fitzgerald, A. M.
Sullivan and C. M. Hosten, J. Inorg. Biochem., 2006, 100, 480–501.
46 Z. Z. Pan, R. Zhang and M. Newcomb, J. Inorg. Biochem., 2006, 100,
524–532.
47 D. Kumar, S. P. de Visser and S. Shaik, J. Am. Chem. Soc., 2005, 127,
8204–8213.
48 T. Kamachi and K. Yoshizawa, J. Am. Chem. Soc., 2005, 127, 10686–
10692.
49 G. Lente and I. Fa´bia´n, J. Chem. Soc., Dalton Trans., 2002, 778–784.
This journal is
The Royal Society of Chemistry 2007
Dalton Trans., 2007, 4268–4275 | 4275
©