A study of the mechanism and kinetics of cyclooctene epoxidation catalyzed by iron(III) tetrakispentafluorophenyl porphyrin

Article abstract of DOI:10.1021/ja043380n

A study has been conducted of the mechanism and kinetics of cyclooctene epoxidation by hydrogen peroxide catalyzed by iron(III) tetrakispentafluorophenyl [F₂₀TPPFe(III)] porphyrin. The formation of cyclooctene oxide, the only product, was determined by gas chromatography, and the consumption of hydrogen peroxide was determined by ¹H NMR. UV-visible spectroscopy was used to identify the state of the porphyrin as a function of solvent composition and reaction conditions and to follow the kinetics of porphyrin degradation. F₂₀TPPFe(III) was found to be inactive in the chloride-ligated form, but became active when the chloride ligand was replaced by a methoxide ligand. The methoxide-ligated form of F ₂₀TPPFe(III) reacts with hydrogen peroxide to form an iron(III) hydroperoxide species, which then undergoes both heterolytic and homolytic cleavage to form iron(IV) π-radical cations and iron(IV) oxo species, respectively. The iron(IV) π-radical cations are responsible for the epoxidation of cyclooctene, whereas the iron(IV) oxo species are responsible for hydrogen peroxide decomposition. The kinetics of cyclooctene epoxidation and hydrogen peroxide decomposition developed from the proposed mechanism describe the experimentally observed kinetics accurately. The rate parameters derived from a fit of the model to the experimental data are consistent with previous estimates of the magnitude of these parameters.

Full text of DOI:10.1021/ja043380n

Published on Web 05/26/2005
A Study of the Mechanism and Kinetics of Cyclooctene
Epoxidation Catalyzed by Iron(III) Tetrakispentafluorophenyl
Porphyrin
Ned A. Stephenson and Alexis T. Bell*
Contribution from the Chemical Sciences DiVision, Lawrence Berkeley Laboratory and
Department of Chemical Engineering, UniVersity of California, Berkeley, California 94720-1462
Received November 2, 2004; E-mail: bell@cchem.berkeley.edu
Abstract: A study has been conducted of the mechanism and kinetics of cyclooctene epoxidation by
hydrogen peroxide catalyzed by iron(III) tetrakispentafluorophenyl [F20TPPFe(III)] porphyrin. The formation
of cyclooctene oxide, the only product, was determined by gas chromatography, and the consumption of
hydrogen peroxide was determined by 1H NMR. UV-visible spectroscopy was used to identify the state of
the porphyrin as a function of solvent composition and reaction conditions and to follow the kinetics of
porphyrin degradation. F20TPPFe(III) was found to be inactive in the chloride-ligated form, but became
active when the chloride ligand was replaced by a methoxide ligand. The methoxide-ligated form of
F20TPPFe(III) reacts with hydrogen peroxide to form an iron(III) hydroperoxide species, which then undergoes
both heterolytic and homolytic cleavage to form iron(IV) π-radical cations and iron(IV) oxo species,
respectively. The iron(IV) π-radical cations are responsible for the epoxidation of cyclooctene, whereas
the iron(IV) oxo species are responsible for hydrogen peroxide decomposition. The kinetics of cyclooctene
epoxidation and hydrogen peroxide decomposition developed from the proposed mechanism describe the
experimentally observed kinetics accurately. The rate parameters derived from a fit of the model to the
experimental data are consistent with previous estimates of the magnitude of these parameters.
aimed at identifying the composition of the active species.1-32
It is suggested that the reaction of hydrogen peroxide with the
Introduction
Metalloporphyrins are effective catalysts for the selective
oxidation of hydrocarbons and other organic compounds under
mild conditions.1-3 As a consequence, considerable effort has
been devoted to investigating the effects of variables, such as
the composition of the porphyrin, the oxidant, the solvent, and
the substrate. The literature pertaining to the use of iron
porphyrins in combination with hydrogen peroxide is particularly
extensive,1-24 and a number of studies have been undertaken
native iron(III) porphyrin results in the formation of an iron-
(III) hydroperoxide species,13 which can then undergo homolytic
cleavage to form iron(IV) oxo species or heterolytic cleavage
to form iron(IV) π-radical cations, and all three of these species
have been cited as being active for the oxidation of hydrocar-
bons.5,6,13,19,21,33,34 However, notwithstanding the many studies
(15) Lee, K. A.; Nam, W. J. Am. Chem. Soc. 1997, 119, 1916-1922.
(16) Nam, W.; Lee, H. J.; Oh, S.-Y.; Kim, C.; Jang, H. G. J. Inorg. Biochem.
2000, 80, 219-225.
(1) Sheldon, R. A. Metalloporphyrins in Catalytic Oxidations; Marcel Dek-
ker: New York, 1994.
(17) Nam, W.; Lim, M. H.; Lee, H. J.; Kim, C. J. Am. Chem. Soc. 2000, 122,
6641-6647.
(2) Montanari, F.; Casella, L. Metalloporphyrins Catalyzed Oxidations; Kluwer
Academic Publishers: Boston, MA, 1994.
(18) Traylor, T. G.; Xu, F. J. Am. Chem. Soc. 1990, 112, 178-186.
(19) Bruice, T. C. Acc. Chem. Res. 1991, 24, 243-249.
(20) Traylor, T. G.; Kim, C.; Richards, J. L.; Xu, F.; Perrin, C. L. J. Am. Chem.
Soc. 1995, 117, 3468-3474.
(3) Meunier, B. Biomimetic Oxidations Catalyzed by Transition Metal Com-
plexes; Imperial College Press: London, 1999.
(4) Strukul, G. Catalytic Oxidations with Hydrogen Peroxide as Oxidant;
Kluwer Academic Publishers: Boston, MA, 1992.
(21) Traylor, T. G.; Kim, C.; Fann, W. P.; Perrin, C. L. Tetrahedron 1998, 54,
7977-7986.
(5) Traylor, T. G.; Tsuchiya, S.; Byun, Y.-S.; Kim, C. J. Am. Chem. Soc. 1993,
115, 2775-2781.
(22) Traylor, T. G.; Xu, F. J. Am. Chem. Soc. 1987, 109, 6201-6202.
(23) Traylor, T. G.; Ciccone, J. P. J. Am. Chem. Soc. 1989, 111, 8413-8420.
(24) Traylor, T. G.; Fann, W.-P.; Bandyopadhyay, D. J. Am. Chem. Soc. 1989,
111, 8009-8010.
(6) Traylor, T. G.; Kim, C.; Richards, J. L.; Xu, F.; Perrin, C. L. J. Am. Chem.
Soc. 1995, 117, 3468-3474.
(7) Cunningham, I. D.; Danks, T. N.; Hay, J. N.; Hamerton, I.; Gunathilagan,
S. Tetrahedron 2001, 57, 6847-6853.
(25) Selke, M.; Sisemore, M. F.; Valentine, J. S. J. Am. Chem. Soc. 1996, 118,
2008-2012.
(8) Nam, W.; Jin, S. W.; Lim, M. H.; Ryu, J. Y.; Kim, C. Inorg. Chem. 2002,
41, 3647-3652.
(26) Nam, W.; Park, S.-E.; Lim, I. K.; Lim, M. H.; Hong, J.; Kim, J. J. Am.
Chem. Soc. 2003, 125, 14674-14675.
(9) Cunningham, I. D.; Danks, T. N.; O’Connell, K. T. A.; Scott, P. W. J.
Chem. Soc., Perkin Trans. 2 1999, 2133-2139.
(27) Lim, M. H.; Jin, S. W.; Lee, Y. J.; Jhon, G.-J.; Nam, W.; Kim, C. Bull.
Korean Chem. Soc. 2001, 22, 93-96.
(10) Nam, W.; Lim, M. H.; Oh, S.-Y.; Lee, J. H.; Lee, H. J.; Woo, S. K.; Kim,
C.; Shin, W. Angew. Chem., Int. Ed. 2000, 39, 3646-3649.
(11) Lee, K. A.; Nam, W. Bull. Korean Chem. Soc. 1996, 17, 669-671.
(12) Nam, W.; Oh, S.-Y.; Sun, Y. J.; Kim, J.; Kim, W.-K.; Woo, S. K.; Shin,
W. J. Org. Chem. 2003, 68, 7903-7906.
(28) Collman, J. P.; Chien, A. S.; Eberspacher, T. A.; Brauman, J. I. J. Am.
Chem. Soc. 2000, 122, 11098-11100.
(29) Nam, W.; Lim, M. H.; Moon, S. K.; Kim, C. J. Am. Chem. Soc. 2000,
122, 10805-10809.
(13) Nam, W.; Lim, M. H.; Lee, H. J.; Kim, C. J. Am. Chem. Soc. 2000, 122,
6641-6647.
(14) Cunningham, I. D.; Danks, T. N.; Hay, J. N.; Hamerton, I.; Gunathilagan,
S.; Janczak, C. J. Mol. Catal. A: Chem. 2002, 185, 25-31.
(30) Moore, K. T.; Horvath, I. T.; Therien, M. J. Inorg. Chem. 2000, 39, 3125-
3139.
(31) Fujii, H. Coord. Chem. ReV. 2002, 226, 51-60.
(32) Fujii, H. Chem. Lett. 1994, 1491-1494.
9
10.1021/ja043380n CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 8635-8643
8635

A R T I C L E S
Stephenson and Bell
of the mechanism and kinetics of iron porphyrin-catalyzed olefin
and paraffin oxidations, a detailed analysis of the reaction
kinetics has not been made for any porphyrin system.
decomposition, catalyzed by F20TPPFe(III). Particular attention
was devoted to identifying the roles of solvent composition and
reactant concentrations on the rates of both processes. A further
objective was to determine rate coefficients and equilibrium
constants for the elementary steps involved in cyclooctene
epoxidation and hydrogen peroxide decomposition.
Iron porphyrins halogenated on the periphery of the porphyrin
ring or containing halogenated substituents attached to the
porphyrin ring have been shown to be particularly efficient for
both epoxidation and hydroxylation reactions.35-37 Such por-
phyrins are also more resistant to degradation via free-radical
attack than those containing electron-donating substituents.1
These features have led a number of investigators to focus on
iron(III) tetrakispentafluorophenyl porphyrin (F20TPPFe(III)).
In the presence of hydrogen peroxide, F20TPPFe(III) can
catalyze the epoxidation of olefins and the hydroxylation of both
olefins and paraffins. The epoxidation of cyclooctene is
particularly notable because only a single product, cyclooctene
oxide, is produced, and for this reason, a number of studies
have been devoted to understanding the mechanism of this
process.7,13-15
Cunningham and co-workers have proposed that hydrogen
peroxide reacts with F20TPPFe(III) to produce an iron(III)
hydroperoxo species, which then undergoes heterolytic O-O
bond cleavage to form an iron(IV) π-radical cation species. The
latter species then participates in two parallel reactions: the
epoxidation of cyclooctene and the decomposition of hydrogen
peroxide. The former reaction is taken to be very fast, while
the later is taken to be fast. Both the pathways ultimately return
the catalyst to its original iron(III) state.7
On the other hand, Nam and co-workers have suggested that
the iron(IV) oxo species produced by the homolytic cleavage
of iron(III) hydroperoxide species might also be involved in
olefin epoxidation.26 However, two observations argue against
this idea. The first is that epoxidation of cis- and trans-stilbene
occurs with retention of stereospecificity.8 This result is strongly
indicative of epoxidation via an iron(IV) π-radical cation, rather
than by an iron(IV) oxo species, which would be expected to
lead to a loss of stereospecificity.5 The second observation is
that even extremely small quantities of water suppress the
epoxidation activity of iron(IV) oxo species.26 Since water is
always present together with hydrogen peroxide because reagent
H2O2 is available as 30 wt % aqueous solution, the pathway
via iron(IV) oxo species should be negligible.
A second issue is whether hydrogen peroxide decomposition
occurs exclusively via the reaction of H2O2 with iron(IV) radical
cations, as suggested by Cunningham and co-workers, or
whether it involves iron(IV) oxo species, as well. If only one
intermediate were involved in both the epoxidation of cy-
clooctene and the decomposition of hydrogen peroxide, then
one would expect that at very high cyclooctene concentrations,
the selectivity for the utilization of H2O2 for epoxidation would
approach 100% asymptotically. Since this does not occur, it
suggests that another intermediate, for example, the iron(IV)
oxo species, might be involved in the decomposition of H2O2.
The aim of this work was to develop a more comprehensive
picture of the mechanism and kinetics of cyclooctene epoxida-
tion by hydrogen peroxide and the kinetics of hydrogen peroxide
Experimental Section
Reagents. Deuterium oxide, 99.9%, was obtained from Cambridge
Isotope Laboratories, Inc. Cis-cyclooctene, 95%, was obtained from
Alfa-Aesar. F20TPPFe(III) in its chloride-ligated form and dodecane,
99+%, were obtained from Aldrich. Non-UV grade acetonitrile,
99.99%, methanol, 99.98%, ACS grade chloroform, 99.8%, and
hydrogen peroxide, 30%, were all obtained from EMD Chemicals.
Reactions. Reactions were initiated by mixing 330 µL of cy-
clooctene, 2.75 mL of solvent (see below), 250 µL of a 1 mM solution
of porphyrin dissolved in solvent, and 5.0 µL of hydrogen peroxide in
a 5 mL reaction vial. Using these conditions, cyclooctene was in excess
and hydrogen peroxide was the limiting reagent. Unless otherwise
stated, the solvent used for all reactions consisted of 25 vol % methanol
and 75 vol % acetonitrile. This choice of solvent composition was based
on the need to keep the rate of reaction at a level that could be measured
readily and yet maintain a relatively high yield of cyclooctene epoxide
based on the consumption of hydrogen peroxide. By increasing the
methanol concentration, it was also possible to decrease porphyrin
degradation. To facilitate product analysis, 10 µL of dodecane was
added as a standard to the reaction mixture. All reactions were
performed at room temperature.
Product Analysis. The concentration of cyclooctene epoxide as a
function of time was determined by gas chromatography. Analysis was
carried out using an HP 6890 series gas chromatograph equipped with
an Agilent DB Wax (30 m × 0.32 mm × 0.5 µm) capillary column
and an FID detector. The 15 min required to analyze a single sample
via gas chromatography was on the order of the time for the completion
of the reaction. Therefore, multiple reactions were performed in order
to obtain data points as a function of time. All data points were obtained
at least in triplicate, and the repeatability of the data was always within
(2%. The repeatability of measuring cyclooctene epoxide via gas
chromatography was determined to be within (1% by analyzing a
calibrated standard.
The hydrogen peroxide concentration was measured as a function
of time using 1H NMR.38 These analyses were carried out using a 400
MHz VMX spectrometer. Samples were prepared in 5 mL reaction
vials as previously described with the exception that the catalyst was
not added to the vial. A portion of this reaction mixture was transferred
to a precision NMR tube. Chloroform was used as the internal standard
for the NMR experiments. The reaction was initiated by adding a
concentrated solution of F20TPPFe(III) to the NMR tube prior to taking
scans. An NMR spectrum was taken every 32 s by averaging the data
from four scans. The area of the hydrogen peroxide peak, centered at
10.3 ppm, was then integrated and compared to the area of an internal
standard. 1H NMR experiments were run in at least triplicate to verify
the repeatability of the experiments; the experimental repeatability was
1
approximately (5%. The quantification of hydrogen peroxide via H
NMR was shown to be accurate within (2% by comparing the relative
areas of the peroxide peak and the internal standard peak for calibrated
standard solutions; multiple internal standards were used to verify the
accuracy. The purity of the internal standard was taken as stated by
the manufacturer, while the concentration of the standardized hydrogen
peroxide solution was verified by a potassium permanganate titration.
Porphyrin Analysis. Degradation of the porphyrin species was
quantified by UV-visible spectroscopy using a Varian Cary 400 Bio
UV-visible spectrometer. Scans of the reacting system were taken as
(33) Almarsson, O.; Bruice, T. C. J. Am. Chem. Soc. 1995, 117, 4533-4544.
(34) He, G.-X.; Bruice, T. C. J. Am. Chem. Soc. 1991, 113, 2747-2753.
(35) Grinstaff, M. W.; Hill, M. G.; Labinger, J. A. Science 1994, 264, 1311-
1313.
(36) Brinbaum, E. R.; Grinstaff, M. W.; Labinger, J. A.; Bercaw, J. E.; Gray,
H. B. J. Mol. Catal. A: Chem. 1995, 104, L119-L122.
(37) Gross, Z.; Simkhovich, L. Tetrahedron Lett. 1998, 39, 8171-8174.
(38) Stephenson, N. A.; Bell, A. T. Anal. Bioanal. Chem. 2005, 381, 1289-
1293.
9
8636 J. AM. CHEM. SOC. VOL. 127, NO. 24, 2005

Cyclooctene Epoxidation by Iron(III) Porphyrin
A R T I C L E S
Figure 1. The proposed mechanism for the epoxidation of cyclooctene using hydrogen peroxide as the oxidant and F20TPPFe(III) as the catalyst.
a function of reaction time. The UV-visible spectra were then
integrated to find the area under the peaks between 280 and 700 nm.
An extinction coefficient was calculated based on a linear plot of
integrated area versus porphyrin concentration. The porphyrin concen-
tration versus reaction time was then fit empirically to a first-order
exponential decay function. The repeatability of the UV-visible scans
was shown to be better than (2%. Negligible error was associated
with taking repeated scans; all error was associated with the sample
preparation.
Much of the data for this research were collected before the
1H NMR technique had been developed for measuring the
concentration of hydrogen peroxide in situ. To analyze such
data required making a number of assumptions about the
relationship of the rate of consumption of hydrogen peroxide
relative to the rate of production of cyclooctene oxide, as well
as an assumption about the concentration of catalyst during the
course of the reactions. In what follows, we give a brief
statement of the working assumptions and show how these can
be used to determine an apparent first-order rate coefficient for
the consumption of hydrogen peroxide, kobs, and the yield of
cyclooctene oxide based on the production of cyclooctene oxide.
The first assumption made in the derivation of an expression
for kobs is that the percentage of iron(III) porphyrin axially
ligated by methanol does not change with time. It was also
assumed that the rate of reaction 7 is much faster than the rate
of reaction 6, which leads to the expectation that the yield of
cyclooctene oxide based on the amount of hydrogen peroxide
consumed should be time independent. On the basis of the
mechanism shown in Figure 1, the rate of epoxidation can be
written according to eq 1. Making use of the pseudo-steady-
state hypothesis for the iron(IV) π-cation intermediate and
assuming that reaction 2 is equilibrated, eq 1 can be rewritten
as eq 2. The concentration of H2O2 as a function of time can
be determined from eq 3, which follows from the assumption
that the ratio of heterolytic to homolytic cleavage is constant
throughout the entire reaction combined with the fact that
cyclooctene oxide is the only oxidation product. In this equation,
Y∞ is the ratio of the amount of epoxide generated after all of
the hydrogen peroxide has been consumed to the initial amount
Analyses of the porphyrin species were carried out using a 400 MHz
VMX spectrometer. Concentrated porphyrin solutions (∼17 mM) in
acetonitrile/methanol solutions were analyzed. The spectral width was
set to 75 188 Hz so that the â-pyrrole protons of the porphyrin could
be monitored. NMR spectra of the porphyrin were collected by taking
256 scans.
Proposed Mechanism and Method of Data Analysis
The proposed mechanism is shown in Figure 1. This scheme
is similar to that presented by Cunningham and co-workers7 in
that epoxidation results from the reaction of cyclooctene with
the π-radical cation formed via heterolytic cleavage of the
oxygen-oxygen bond of the hydroperoxo intermediate (reaction
3). Cyclooctene epoxidation occurs via reaction 7. In distinction
to the mechanism of Cunningham and co-workers, the mech-
anism shown in Figure 1 also includes a pathway for the
homolytic cleavage of the hydroperoxo species (reaction 4).
Hydrogen peroxide decomposition is envisioned to occur via
reaction 5. Reaction 6 is included to account for the conversion
of the iron(IV) π-radical cation to the iron(IV) oxo species. The
participation of hydroperoxo species as intermediates in olefin
epoxidation appears unlikely based on recent theoretical work,
which shows that the activation energy for ethene epoxidation
by hydroperoxo species is high (53 kcal/mol) compared to that
for expoxidation by iron(IV) π-radical cations (14 kcal/mol).39,40
(39) Kamachi, T.; Shiota, Y.; Ohta, T.; Yoshizama, K. Bull. Chem. Soc. Jpn.
2003, 76, 721-732.
(40) Ogliaro, F.; de Visser, S. P.; Cohen, S.; Sharma, P. K.; Shaik, S. J. Am.
Chem. Soc. 2002, 124, 2806-2817.
9
J. AM. CHEM. SOC. VOL. 127, NO. 24, 2005 8637

A R T I C L E S
Stephenson and Bell
In the next section, the experimental results are examined in
light of the proposed mechanism and the method of data analysis
presented above. This is then followed by a more detailed
analysis of the kinetics of cyclooctene epoxidation and hydrogen
peroxide decomposition in the presence and absence of cy-
clooctene. For that part of the discussion, data on the consump-
1
tion of H2O2 were obtained from H NMR analyses. It is also
shown that the rate parameters obtained through the simplified
analysis presented above and those obtained through the more
detailed analysis are very similar.
Results and Discussion
Solvent Effects. When acetonitrile or methylene chloride is
used as the solvent, the porphyrin remains in its chloride-ligated
form, and epoxidation does not occur. Epoxidation occurs once
methanol is added to the aprotic solvents. It was also observed
that in the absence of cyclooctene hardly any hydrogen peroxide
decomposition occurred without the addition of methanol to the
acetonitrile. Similar effects of solvent composition have been
reported previously and have been attributed to replacement of
the chloride ligand by an alkoxide.10-12
Figure 2. Gas chromatography data allow for the graphical extraction of
k3K2 and k4K2.
of hydrogen peroxide. Combining eqs 2 and 3, leads to eq 4, in
which kobs is defined by eq 5. As shown in eq 6, integration of
eq 4 gives the expression for hydrogen peroxide as a function
time.
UV-visible spectroscopy was used to identify the effects of
solvent composition on the nature of the axial ligand with the
F20TPPFe(III). In pure acetonitrile, a Soret peak is observed at
407 nm, which we assign to F20TPPFe-Cl. When acetonitrile
is progressively replaced by methanol, the Soret peak shifts to
388 nm and then to 404 nm. We assign the feature at 388 nm
to the methoxide-ligated porphyrin, F20TPPFe(III)-OCH3, and
the feature at 404 nm to F20TPPFe(III)-OCH3 interacting with
a second molecule of methanol. We note that Lee and Nam
have made a similar assignment for the band at 404 nm.15
Under reaction conditions, Soret bands are observed at 388
and 410 nm, but not at 404 nm, independent of the methanol
concentration. We suggest that coordination of hydrogen
peroxide at the proximal position of the porphyrin precludes
the ligation of a second equivalent of methanol. After all
hydrogen peroxide is consumed, UV-visible spectra of the
reaction mixture indicate that bismethanol ligation is again
possible, as evidenced by a Soret band at 404 nm. The intensity
of the feature at 410 nm under reaction conditions varies strongly
depending on whether cyclooctene is present or absent from
the reaction mixture. This feature is assigned to overlapping
bands due to F20TPPFe(III)-Cl and an iron(IV) oxo species.
The presence of the latter species is indicated by the appearance
of a characteristic band at 540 nm.11 When cyclooctene is
present, the peak at 540 nm is absent, suggesting that the solution
contains almost exclusively F20TPPFe(III)-Cl and F20TPPFe-
(III)-OCH3. On the basis of this evidence, we propose that the
active species for catalysis is F20TPPFe-OCH3, and that this
species is present in equilibrium with F20TPPFe-Cl defined by
the following reaction:
d[C8 - O]
) k7[+•Fe(IV) ) O][C8]
(1)
(2)
(3)
dt
d[C8 - O]
) k3K2[Fe - OMe][H2O2]
[C8 - O]
dt
[H2O2] ) [H2O2]0 -
Y∞
d[H2O2]
) -kobs[H2O2]
(4)
(5)
dt
k3K2[Fe - OMe]
kobs
)
Y∞
[H2O2]0
[H2O2]
ln
) kobs
t
(6)
(
)
[C8 - O]
k3[MeOH]
Yield )
× 100% =
× 100%
[H2O2]0
k3[MeOH] + 2k4
(7)
If all of the assumptions involved in the analysis are valid,
then a plot of the logarithm of the ratio of the initial hydrogen
peroxide concentration to that occurring at any time should be
a linear function that passes through the origin. As seen in Figure
2, this is indeed the observed relationship. The individual
assumptions underlying the definition of kobs are substantiated
below. The value of k3K2 can be determined from the expression
for kobs and the slope of the plot shown in Figure 2. A value for
k4K2 can be estimated by making a few additional assumptions.
If peroxide decomposition occurs only via the homolytic
pathway and the ratio of the rate of homolytic to heterolytic
cleavage is constant throughout the reaction, the final yield can
be approximated by eq 7. A factor of 2 is placed before the k4
term in the yield expression since two molecules of hydrogen
peroxide are consumed in the homolytic pathway (see Figure
1).
The equilibrium constant, K1, associated with this reaction
can be calculated from the relative concentrations of F20TPPFe-
Cl and F20TPPFe-OCH3, which are determined by deconvo-
9
8638 J. AM. CHEM. SOC. VOL. 127, NO. 24, 2005

Cyclooctene Epoxidation by Iron(III) Porphyrin
A R T I C L E S
with literature,41,42 we found that the â-pyrrole protons of
F20TPPFe-Cl produce an NMR signal at 83 ppm in purely
aprotic solvents (acetonitrile and CDCl3). As methanol was
substituted progressively for acetonitrile, the peak at 83 ppm
decreased in magnitude, while a new peak appeared at 65 ppm
and increased in magnitude. Both NMR peaks are broad and
are located significantly downfield from the position of the
â-pyrrole resonance for iron-free porphyrin. These observations
are characteristic of a paramagnetic porphyrin species, suggest-
ing no change in the spin state or oxidation state of the Fe cation
occurs as methanol is substituted for acetonitrile. We attribute
the peak at 65 ppm to the â-pyrrole protons of iron porphyrins
in which methoxide anions serve as the axial ligand. This
interpretation is consistent with previously reported observations
of the effects of axial ligands on the 1H NMR shifts of
paramagnetic Fe(III) porphyrins.41 Since the methoxide anion
is less electron withdrawing than the chloride anion, the electron
density on the porphyrin ring will be higher in the former case.
Consistent with this reasoning, the â-pyrrole resonance would
be expected to shift upfield relative to that for the chloride-
ligated porphyrin, which is what is observed. The presence of
two distinct peaks at 83 and 65 ppm indicates an equilibrium
relationship between the two species. Due to the high concentra-
tion of porphyrin required for NMR analysis and the rapid
kinetics of the epoxidation reaction, the porphyrin catalyst could
not be studied by NMR under reaction conditions. However,
the above results do provide further evidence to support our
claim of methoxide for chloride ligand substitution occurring
upon progressive replacement of acetonitrile by methanol in the
solvent.
Figure 3. Nonlinear relationship between the observed rate constant and
the total porphyrin concentration.
The mechanism shown in Figure 1 suggests that methanol
participates in the heterolytic cleavage of the iron-hydroperoxo
intermediate in a manner analogous to that proposed for
hydroperoxide species in the heme-containing enzyme, cyto-
chrome P-450. In that system, the iron-hydroperoxo intermedi-
ate is converted to the ferryl porphyrin radical-cation inter-
mediate via protonation and elimination of water. The proton
is donated from a threonine unit on the enzyme; the process is
also facilitated by several other amino acid derivatives on the
enzyme.43,44 Similar acid/base processes have been reported in
model porphyrin systems, and it is thought that the acidic
character of methanol facilitates the heterolytic cleavage of the
oxygen-oxygen bond of the hydroperoxo-porphyrin interme-
diate when the solvent system contains methanol and an aprotic
solvent.18
As shown in Figure 5, both the observed rate constant and
the yield of epoxide based on hydrogen peroxide consumption
increase with increasing methanol concentration. Heterolytic
cleavage of the iron-hydroperoxo intermediate involves proton
transfer between methanol and the hydroperoxo intermediate,
while homolytic cleavage does not. If only heterolytic cleavage
occurs, as suggested by Cunningham and co-workers,7 then
changing the methanol concentration should not affect the yield.
The increase in yield and the increase in the observed rate
constant strongly indicate that the pathways involving homolytic
Figure 4. Linear relationship between the observed rate constant and the
methoxide-ligated porphyrin.
lution of the UV-visible spectrum taken under reaction
conditions. By this means, a value of K1 ) 1.3 × 10-5 is
obtained.
To test the ideas that only the methoxide-ligated porphyrin
species are active for epoxidation and that F20TPPFe-Cl and
F20TPPFe-OCH3 are in equilibrium, the value of kobs was
determined as a function of the total porphyrin concentration.
On the basis of the aforementioned equilibrium relationship and
material balances, one would expect the concentration of the
methoxide-ligated porphyrin to increase nonlinearly with the
concentration of total porphyrin. If the observed rate of reaction
is first order in the methoxide-ligated porphyrin species, then
the observed rate constant will also be nonlinear as a function
of the total porphyrin concentration. As seen in Figure 3, this
relationship is observed. However, Figure 4 shows that kobs is
a linear function of F20TPPFe-OCH3; the concentration of
which is determined by fitting the value of K1 to the kinetic
data shown in Figure 3; this fitting results in a value of K1 )
1.2 × 10-5, which is comparable to that obtained from UV-
visible spectroscopy (1.3 × 10-5), as described above. The very
close agreement of the two values of K1 obtained by independent
means supports the proposition that the addition of methanol
leads to the conversion of F20TPPFe-Cl to F20TPPFe-OCH3.
1H NMR was used to further substantiate the hypothesis that
a change in the nature of the axial ligand on the porphyrin occurs
as the percentage of methanol in the solvent is varied. Consistent
(41) Birnbaum, E. R.; Hodge, J. A.; Grinstaff, M. W.; Schaefer, W. P.; Henling,
L.; Labinger, J. A.; Bercaw, J. E.; Gray, H. B. Inorg. Chem. 1995, 34,
3625-3632.
(42) Song, B.; Park, B.; Han, C. Bull. Korean Chem. Soc. 2002, 23, 119-121.
(43) Meunier, B.; Bernadou, J. Struct. Bonding 2000, 97, 1-35.
(44) de Montellano, O. Cytochrome P450: Structure, Mechanism, and Bio-
chemistry, 2nd ed.; Plenum Press: New York, 1995.
9
J. AM. CHEM. SOC. VOL. 127, NO. 24, 2005 8639

A R T I C L E S
Stephenson and Bell
Figure 5. Observed rate constant and epoxide yield as a function of
methanol concentration.
Figure 6. Yield and observed rate constant as a function of substrate
concentration.
and heterolytic cleavage of the O-O bond in iron(III) hydro-
peroxo species compete. Even though the dependence on
methanol concentration disappears from the expression for kobs
when the pseudo-steady-state hypothesis is applied, the observed
rate constant continues to be a function of methanol concentra-
tion because of the dependence on the concentration of
methoxide-ligated porphyrin.
hydrogen peroxide may be a complex function of the interactions
between the solvent system and the hydrogen peroxide mol-
ecules. A process that would be consistent with the observed
phenomenon is the interaction of two methanol molecules with
a single molecule of hydrogen peroxide via hydrogen bonding.
Such an interaction could result in weakening the oxygen-
hydrogen bonds in the hydrogen peroxide molecule, thus
facilitating cleavage of the O-O bond of the hydroperoxide
ligand. Consistent with this hypothesis, we observed that the
position of the proton NMR peak for hydrogen peroxide shifts
downfield as the amount of methanol in the reaction mixture is
increased; this trend is indicative of an increase in the extent of
hydrogen bonding.
Substrate Concentration. If the mechanism proposed in
Figure 1 is valid, both the yield of epoxide and kobs should be
independent of the initial cyclooctene concentration at high
substrate concentrations. Figure 6 shows that both expectations
are borne out for cyclooctene concentrations between 0.4 and
1.4 M. This indicates that over this range of cyclooctene
concentrations the rate of reaction 7 is much greater than the
rate of reaction 6, and that reaction 3 is irreversible. The fact
that the observed rate of epoxide formation does not change
with increasing cyclooctene concentration confirms the applica-
tion of the pseudo-steady-state hypothesis to the intermediate
formed via heterolytic cleavage of the iron(III) hydroperoxide
species formed upon the initial interaction of H2O2 with the
catalyst.
To obtain more information about the relative values of k6
and k7, the initial cyclooctene concentration was decreased until
the rates of reactions 6 and 7 became comparable, as evidenced
by a reduction in the epoxide yield. Figure 6 shows that the
yield of epoxide decreased noticeably once the cyclooctene
concentration was decreased below 0.1 M. On the basis of the
model of the reaction kinetics, described below, the value of k6
is determined to be 0.3 of that of k7.
Oxidant Concentration. In the proposed mechanism shown
in Figure 1, it is assumed that the rate of cyclooctene epoxidation
is first order in hydrogen peroxide concentration, in agreement
with what has been reported previously.7,18 Several experiments
were carried out to verify this assumption. As seen in Figure 7,
neither the observed rate constant nor the final yield of epoxide
relative to hydrogen peroxide consumed changes as the initial
concentration of hydrogen peroxide is varied. Therefore,
consistent with the proposed mechanism and the underlying
The proposed mechanism (see Figure 1) should also explain
the sigmoidal behavior of the observed rate constant as the
methanol concentration increases. This type of relationship has
been reported before for F20TPPFe(III) in protic solvents,
although not interpreted.12 As would be expected based on the
proposed mechanism, the observed rate constant plateaus at
higher methanol concentrations. The proposed equilibrium
relationship between the chloride and methoxide-ligated species
predicts a hyperbolic increase in the amount of methoxide-
ligated porphyrin present. The simultaneous saturation of the
concentration of the methoxide-ligated porphyrin species and
the observed rate constant at high methanol concentrations
support the idea that only the methoxide-ligated species are
active for catalysis. However, below 32 vol % methanol, the
observed rate constant does not increase hyperbolically as would
be expected based only on the amount of methoxide-ligated
porphyrin present in the system. Instead, kobs appears to increase
as the square of the methanol concentration. This suggests that
methanol may play some other role in the equilibrium reaction
between the methoxide-ligated porphyrin or in the cleavage of
the O-O bond of the iron(III) hydroperoxo species.
Several factors can be proposed to account for the observed
reaction rate being lower than would be predicted based on the
amount of methoxide-ligated porphyrin present. First, at lower
methanol concentrations, a higher percentage of the porphyrin
will be in the iron(IV) oxo form at steady-state reaction
conditions. If the amount of porphyrin in the iron(IV) oxo state
became significant relative to the total porphyrin concentration,
then less porphyrin would be in the methoxide-ligated state
under reaction conditions, thus lowering the observed rate of
reaction. Second, porphyrin degradation increases as the metha-
nol concentration decreases. If a significant amount of porphyrin
is degraded during the initial few minutes of the reaction, then
the assumption that the concentration of methoxide-ligated
porphyrin is constant over the interval in which the rate data
are collected will be invalid, resulting in a lower observed rate
constant. Finally, the cleavage of the oxygen-oxygen bond in
9
8640 J. AM. CHEM. SOC. VOL. 127, NO. 24, 2005

Cyclooctene Epoxidation by Iron(III) Porphyrin
A R T I C L E S
While the decomposition of F20TPPFe(III) during the oxida-
tion of hydrocarbons by hydrogen peroxide has been reported
previously,7 a clear picture of the degradation mechanism is
not available. Cunningham and co-workers propose that por-
phyrin degradation occurs via attack of the porphyrin ring by
hydroxyl radicals.7 This idea is supported by both theoretical45
and experimental work46 indicating that the porphyrin ring in
the biological heme oxygenase system degrades via a radical
mechanism. In our work, the participation of radicals is
supported by the observation that porphyrin degradation is
reduced in the presence of acetone and dimethyl sulfoxide, both
of which are known to be radical traps.47-49 For example, when
the reaction was carried out in an acetonitrile/methanol mixture
containing less than 10 vol % methanol, all of the porphyrin
was destroyed by the end of the experiment, even though
hydrogen peroxide remained. When the acetonitrile was replaced
with acetone, the porphyrin was protected from complete
degradation, and the reaction stopped upon complete consump-
tion of the hydrogen peroxide.
A porphyrin degradation mechanism involving free radicals
would explain the results presented in Table 1. Radical species
(e.g., •OOH, •OH) formed during reaction can follow several
pathways. Two of the major possibilities are attack on the
porphyrin ring and radical-radical recombination. As the
porphyrin catalyst concentration is increased, the rate at which
radicals are produced will increase, as well. If porphyrin
degradation is first order in radical species and radical recom-
bination is second order in radical species, then the rate of radical
recombination will increase faster than the observed rate of
porphyrin decomposition as the initial porphyrin concentration
is increased. Less porphyrin is consumed as the porphyrin
concentration is increased because a higher radical concentration
favors second-order radical recombination as opposed to first-
order porphyrin decomposition. The apparent time constant, τ,
characterizes the duration over which radicals are produced in
the system. Since radicals are derived from hydrogen peroxide,
τ characterizes the time constant for the consumption of
hydrogen peroxide. As expected for a catalyzed reaction, the
time required to consume all of the peroxide decreases as the
porphyrin catalyst concentration increases. Therefore, the time
constant, τ, should decrease as the porphyrin catalyst concentra-
tion increases. This explanation would also predict that τ would
be independent of the initial hydrogen peroxide concentration
since the oxidation reactions shown in Figure 1 are first order
in hydrogen peroxide. As shown in Table 1, τ behaves as
predicted by the proposed radical mechanism. In addition, the
amount of porphyrin degraded should increase with an increase
in the initial concentration of hydrogen peroxide since more
radicals will be produced. The results shown in Table 1 are
consistent with this expectation.
Figure 7. Yield and observed rate constant as a function of H2O2.
Table 1. Fitted Parameters for Porphyrin Decomposition Kinetics
[H2O2]0
(mM)
y0
M)
yf
M)
τ
(s)
(
µ
(
µ
2.9
2.9
2.9
2.9
2.9
2.9
2.9
2.9
4.4
5.9
7.3
8.9
3.5
7.4
0.44
0.35
0.27
0.20
0.16
0.13
0.11
0.27
0.26
0.24
0.27
11.9
14.9
17.9
20.9
23.9
26.9
14.9
14.9
14.9
14.9
11.0
14.5
17.9
21.1
24.3
11.0
10.0
9.3
7.5
assumptions, the production of cyclooctene epoxide is first order
in hydrogen peroxide.
Porphyrin Degradation. Quantification of the porphyrin
before and after the reaction using UV-visible spectroscopy
showed that only 10% of the porphyrin was destroyed during
the reaction, confirming that the total porphyrin concentration
is approximately constant, as was assumed in order to determine
kobs from the gas chromatographic data.
To develop a better representation of the kinetics of cy-
clooctene epoxidation, the kinetics of porphyrin decomposition
were determined. UV-visible spectroscopy was used to measure
the concentration of porphyrin as a function of time for different
initial concentrations of porphyrin and hydrogen peroxide. Good
fits to the data could be obtained using the expression
y - yf
y0 - yf
) exp(- t/τ)
(8)
where y0, y, and yf are the porphyrin concentrations at t ) 0, at
time t, and at the point when all of the hydrogen peroxide has
been consumed, respectively. The parameter τ in eq 8 is the
apparent time constant for porphyrin degradation. Values of yf
and τ are shown in Table 1 for different values of the initial
concentrations of porphyrin and hydrogen peroxide. It is evident
that for a fixed initial concentration of hydrogen peroxide, the
extent of porphyrin degradation decreases as the initial con-
centration of porphyrin increases, and that the time constant
for porphyrin degradation decreases. On the other hand, for a
fixed initial concentration of porphyrin, the extent of porphyrin
degradation increases with an increase in the initial concentration
of hydrogen peroxide, but the time constant for porphyrin
degradation remains essentially constant.
Kinetic Modeling
Results. A detailed analysis of the kinetics of cyclooctene
epoxidation and hydrogen peroxide decomposition was carried
(45) Sharma, P. K.; Kevorkiants, R.; de Visser, S. P.; Kumar, D.; Shaik, S.
Angew. Chem., Int. Ed. 2004, 43, 1129-1132.
(46) Wilks, A.; Torpey, J.; Ortiz de Montellano, P. R. J. Biol. Chem. 1994,
269, 29553-29556.
(47) Scaduto, R. C. Free Rad. Biol. Med. 1995, 18, 271-277.
(48) Roelfes, G.; Lubben, M.; Hage, R.; Que, L.; Feringa, B. L. Chem.sEur. J.
2000, 6, 2152-2159.
(49) Walling, C.; El-Taliawi, G. M. J. Am. Chem. Soc. 1973, 95, 844-847.
9
J. AM. CHEM. SOC. VOL. 127, NO. 24, 2005 8641

A R T I C L E S
Stephenson and Bell
Figure 8. Kinetic model for the case where cyclooctene is present.
Figure 10. UV-visible spectra under reaction conditions with cyclooctene
present.
proximity of the Soret peaks for each porphyrin species.
However, the UV-visible spectroscopy can be used to dem-
onstrate qualitative agreement between the mechanism shown
in Figure 1 and experimental observation. It should be noted
that the same ratio of substrate, catalyst, and oxidant were used
for the UV-visible experiments as were used for the reaction
experiments; however, the concentrations of each species were
reduced by a factor of 5 to prevent saturation of the detector.
Under reaction conditions with cyclooctene present, the model
predicts that 1% of the total porphyrin should be present as the
iron(IV) oxo species. As seen in Figure 10, the lack of a
significant contribution to the Soret peak at 410 nm suggests
that the concentration of the iron(IV) oxo species is low when
reaction occurs in the presence of cyclooctene. The shoulder
on the Soret peak near 407 nm is attributed to the chloride-
ligated porphyrin.
According to the mechanism presented in Figure 1, a higher
concentration of iron(IV) oxo species should be present when
cyclooctene is absent than when it is present since the iron(IV)
oxo intermediate will also be produced via reaction 6, and
reaction 7 will not consume the iron(IV) π-radical species. Since
the rate of heterolytic cleavage is about 15 times greater than
that of homolytic cleavage, a significant amount of the iron-
(IV) oxo species should be generated. The model of the reaction
kinetics predicts that about 17% of the total porphyrin present
will be in the iron(IV) oxo state. Figure 11 shows that there is
a noticeable contribution to the spectrum around 410 nm,
indicating that the iron(IV) oxo species are indeed present in a
significant concentration under reaction conditions. Further
confirmation of this interpretation is the clear evidence of a band
at 540 nm. The chloride-ligated porphyrin species are also
expected to contribute to this peak at 410 nm (see above). The
kinetic model predicts that the latter species should correspond
to 33% of the total porphyrin concentration. Thus, one would
expect the Soret peaks at 388 and 410 nm to have similar
magnitudes, assuming the extinction coefficients of all species
to be similar. As seen in Figure 11, the UV-visible spectrum
observed in the absence of cyclooctene is qualitatively consistent
with the predictions of the model.
Figure 9. Kinetic modeling for peroxide decomposition by F20TPPFe(III).
out based on the mechanism presented in Figure 1. Differential
equations representing the kinetics of cyclooctene epoxide
formation and hydrogen peroxide consumption were written
based on the proposed mechanism, and these equations were
solved numerically using a fourth-order Runge-Kutta algorithm.
To extract the parameters k3K2, k4K2, and k5, the changes with
time in the concentrations of cyclooctene and hydrogen peroxide
predicted by the model were fitted to the experimental data taken
in the presence and absence of cyclooctene. The method of least
squares was used to fit the rate parameters to the data. The
concentration of the methoxide-ligated porphyrin species was
determined from the previously described equilibrium relation-
ship. The total porphyrin concentration was represented using
eq 8. The time constant for the porphyrin degradation was
assumed to be identical to the time constant for peroxide
consumption for both cases with and without cyclooctene
present, and the amount of porphyrin remaining after the reaction
was determined by UV-visible spectroscopy. Upon fitting the
data, the best-fit values obtained were k3K2 ) 73 M-1 s-1
,
k4K2 ) 32 M-1 s-1, and k5 ) 37 M-2 s-1. Figures 8 and 9
show that the model does an excellent job of describing the
data. It is also noted that the values of k3K2 and k5 are consistent
with those estimated by Cunningham and co-workers, which
are on the order of 100 M-1 s-1 12
. Note, k5 should be multiplied
by the methanol concentration (6.2 M) to be consistent with
the observed rate constant predicted by Cunningham.
Qualitative Comparison to UV-Visible Data. The iron-
(IV) oxo intermediate formed by homolytic cleavage of the
oxygen-oxygen bond of the iron(III) hydroperoxo species can
be identified by a Soret peak at 410 nm and a second peak at
540 nm.11 Deconvolution of the UV-visible data taken under
reaction conditions to determine the relative amounts of each
porphyrin species present is not feasible due to the close
Comparison of Techniques. The rate coefficients extracted
from the gas chromatographic data and from fitting the
predictions of the kinetic model to both the gas chromatography
and NMR data are in excellent agreement. On the basis of
chromatographic data for cyclooctene oxide generation,
9
8642 J. AM. CHEM. SOC. VOL. 127, NO. 24, 2005

Cyclooctene Epoxidation by Iron(III) Porphyrin
A R T I C L E S
used in the model to predict the concentration of methoxide-
ligated porphyrin as a function of time. The model of the
reaction kinetics predicts that approximately 61% of the
porphyrin species is in the methoxide-ligated form over the time
scale of the reaction; this is in agreement with the assumption
that was used in the analysis of the gas chromatography data,
namely, that the amount of porphyrin that is methoxide-ligated
does not change significantly with time. On the basis of UV-
visible data under reaction conditions, 62% of the porphyrin is
in the methoxide-ligated state when the solvent mixture contains
25 vol % methanol.
Conclusions
A mechanism has been proposed and verified for the
epoxidation of cyclooctene by hydrogen peroxide using F20TPPFe-
(III) as the catalyst. Epoxidation is found to involve an iron-
(IV) π-radical cation, but hydrogen peroxide decomposition
involves an iron(IV) oxo species. To become active, the
chloride-ligated form of the catalyst must first be converted into
a methoxide-ligated form. A model of the reaction kinetics based
on the mechanism presented in Figure 1 was used to determine
the rate coefficients for three of the significant reactions steps;
the equilibrium constant between the chloride-ligated and
methoxide-ligated forms of porphyrin was also determined.
Figure 11. UV-visible spectra under reaction conditions without cy-
clooctene present.
k3K2 ) 80 ( 8 M-1 s-1 and k4K2 ) 34 ( 4 M-1 s-1
respectively, and on the basis of the model of the kinetics,
k3K2 ) 73 ( 5 M-1 s-1 and k4K2 ) 32 ( 3 M-1 s-1
,
,
respectively. This agreement validates the assumptions required
to use the gas chromatography data alone for the determination
of rate parameters and, most importantly, the assumption that
the ratio of heterolytic to homolytic cleavage is constant
throughout the course of the reaction and can be determined
from the final yield. However, NMR data must be used if the
value of k5 is to be determined.
Acknowledgment. This work was supported by the Director,
Office of Basic Energy Sciences, Chemical Sciences Division
of the U.S. Department of Energy under Contract DE-AC03-
76SF00098.
Finally, rather than assume that a constant percentage of the
porphyrin is methoxide-ligated, the equilibrium relationship was
JA043380N
9
J. AM. CHEM. SOC. VOL. 127, NO. 24, 2005 8643