Mechanistic study of iron(III) tetrakis(pentafluorophenyl)porphyrin triflate (F20TPP)Fe(OTf) catalyzed cyclooctene epoxidation by hydrogen peroxide

Article abstract of DOI:10.1021/ic060757c

We have recently proposed a mechanism for the epoxidation of cyclooctene by H₂O₂ catalyzed by iron(III) [tetrakis-(pentafluorophenyl) ]porphyrin chloride, (F₂₀TPP)FeCl, in solvent containing methanol [Stephenson, N. A.; Bell, A.T. Inorg. Chem. 2006, 45, 2758-2766]. In that study, we found that catalysis did not occur unless (F₂₀TPP)FeCl first dissociated, a process facilitated by the solvation of the Cl^- anion by methanol and the coordination of methanol to the (F₂₀TPP)Fe ⁺ cation. Methanol as well as other alcohols was also found to facilitate the heterolytic cleavage of the O-O bond of H₂O ₂ coordinated to the (F₂₀TPP)Fe⁺ cation via a generalized acid mechanism. In the present study, we have shown that catalytic activity of the (F₂₀TPP)Fe⁺ cation can be achieved in aprotic solvent by displacing the tightly bound chloride anion with a weakly bound triflate anion. By working in an aprotic solvent, acetonitrile, it was possible to determine the rate of heterolytic O-O bond cleavage in coordinated H₂O₂ unaffected by the interaction of the peroxide with methanol. A mechanism is proposed for this system and is shown to be valid over a range of reaction conditions. The mechanisms for cyclooctene epoxidation and H₂O₂ decomposition for the aprotic and protic solvent systems are similar with the only difference being the mechanism of proton-transfer prior to heterolytic cleavage of the oxygen-oxygen bond of coordinated hydrogen peroxide. Comparison of the rate parameters indicates that the utilization of hydrogen peroxide for cyclooctene epoxidation is higher in a protic solvent than in an aprotic solvent and results in a smaller extent of porphyrin degradation due to free radical attack. It was also shown that water can coordinate to the iron porphyrin cation in aprotic systems resulting in catalyst deactivation; this effect was not observed when methanol was present, since methanol was found to displace all of the coordinated water.

Full text of DOI:10.1021/ic060757c

Inorg. Chem. 2007, 46, 2278−2285
Mechanistic Study of Iron(III) [Tetrakis(pentafluorophenyl)Porphyrin
Triflate (F20TPP)Fe(OTf) Catalyzed Cyclooctene Epoxidation by Hydrogen
Peroxide
Ned A. Stephenson and Alexis T. Bell*
Chemical Sciences DiVision, Lawrence Berkeley Laboratory and Department of Chemical
Engineering, UniVersity of California, Berkeley, California 94720-1462
Received May 4, 2006
We have recently proposed a mechanism for the epoxidation of cyclooctene by H2O2 catalyzed by iron(III) [tetrakis-
(pentafluorophenyl)]porphyrin chloride, (F20TPP)FeCl, in solvent containing methanol [Stephenson, N. A.; Bell, A.T.
Inorg. Chem. 2006, 45, 2758−2766]. In that study, we found that catalysis did not occur unless (F20TPP)FeCl first
dissociated, a process facilitated by the solvation of the Cl- anion by methanol and the coordination of methanol
to the (F20TPP)Fe+ cation. Methanol as well as other alcohols was also found to facilitate the heterolytic cleavage
of the O−
O bond of H2O2 coordinated to the (F20TPP)Fe+ cation via a generalized acid mechanism. In the present
study, we have shown that catalytic activity of the (F20TPP)Fe+ cation can be achieved in aprotic solvent by displacing
the tightly bound chloride anion with a weakly bound triflate anion. By working in an aprotic solvent, acetonitrile,
it was possible to determine the rate of heterolytic O−O bond cleavage in coordinated H2O2 unaffected by the
interaction of the peroxide with methanol. A mechanism is proposed for this system and is shown to be valid over
a range of reaction conditions. The mechanisms for cyclooctene epoxidation and H2O2 decomposition for the aprotic
and protic solvent systems are similar with the only difference being the mechanism of proton-transfer prior to
heterolytic cleavage of the oxygen−oxygen bond of coordinated hydrogen peroxide. Comparison of the rate parameters
indicates that the utilization of hydrogen peroxide for cyclooctene epoxidation is higher in a protic solvent than in
an aprotic solvent and results in a smaller extent of porphyrin degradation due to free radical attack. It was also
shown that water can coordinate to the iron porphyrin cation in aprotic systems resulting in catalyst deactivation;
this effect was not observed when methanol was present, since methanol was found to displace all of the coordinated
water.
Introduction
the composition of the axial ligand and the composition of
the solvent in which reactions are carried out have a strong
effect on the catalytic activity of the porphyrin and the yield
The peroxidase and catalase-like activity of iron porphyrins
is well-known, and many studies have been devoted to
elucidating the mechanism of porphyrin-catalyzed oxidation
of alkanes and olefins.1-57 These studies have revealed that
(7) Grinstaff, M. W.; Hill, M. G.; Birnbaum, E. R.; Schaefer, W. P.:
Labinger, J. A.; Gray, H. B. Inorg. Chem. 1995, 34, 4896-902.
(8) Grinstaff, M. W.; Hill, M. G.; Labinger, J. A.; Gray, H. B. Science
1994, 264, 1311-1313.
(9) Mansuy, D. Coord. Chem. ReV. 1993, 125, 129-141.
(10) Balch, A. L. Inorg. Chim. Acta 1992, 197, 198-200.
(11) Meunier, B. Chem. ReV. 1992, 92, 1411-1456.
(12) Vaz, A. D. N.; Roberts, E. S.; Coon, M. J. J. Am. Chem. Soc. 1991,
113, 5886-5887.
* To whom correspondence should be addressed. E-mail: alexbell@
berkeley.edu.
(1) Sheldon, R. A. Metalloporphyrins in Catalytic Oxidations; Marcel
Dekker: New York, 1994.
(2) Montanari, F.; Casella, L. Metalloporphyrins Catalyzed Oxidations;
Kluwer Academic Publishers: Boston, MA, 1994.
(3) Meunier, B. Biomimetic Oxidations Catalyzed by Transition Metal
Complexes; Imperial College Press: London, 1999.
(4) Moore, K. T.; Horvath, I. T.; Therien, M. J. Inorg. Chem. 2000, 39,
3125-3139.
(5) Kalish, H. R.; Latos-Grazynski, L.; Balch, A. L. J. Am. Chem. Soc.
2000, 122, 12478-12486.
(6) Birnbaum, E. R.; Grinstaff, M. W.; Labinger, J. A.; Bercaw, J. E.;
Gray, H. B. J. Mol. Catal. A.: Chem. 1995, 104, 119-122.
(13) Bruice, T. C. Acc. Chem. Res. 1991, 24, 243-249.
(14) Traylor, T. G. Pure Appl. Chem. 1991, 63, 265-274.
(15) Gunter, M. J.; Turner, P. Coord. Chem. ReV. 1991, 108, 115-161.
(16) Ellis, P. E.; Lyons, J. E. Coord. Chem. ReV. 1990, 105, 181-193.
(17) Mansuy, D. Pure Appl. Chem. 1990, 62, 741-746.
(18) Wijesekera, T.; Matsumoto, A.; Dolphin, D.; Lexa, D. Angew. Chem.,
Int. Ed. 1990, 29, 1073-1074.
(19) Tsuchiya, S.; Seno, M. Chem. Lett. 1989, 263-266.
2278 Inorganic Chemistry, Vol. 46, No. 6, 2007
10.1021/ic060757c CCC: $37.00
© 2007 American Chemical Society
Published on Web 02/21/2007

Mechanistic Study of (F20TPP)Fe(OTf)
of oxidation products relative to the consumption of hydrogen
peroxide.58-64 A system of particular interest has been the
oxidation of olefins, such as cyclooctene, in which case the
major, and sometimes even the exclusive, product is the
epoxide. It has been shown for such systems that iron(III)
porphyrins coordinated by a weakly bound axial ligand such
as a triflate, perchlorate, hexafluoroantimonate, or nitrate
anion are catalytically active in aprotic solvents (e.g.,
acetonitrile or methylene chloride) but that porphyrins
coordinated by a strongly bound axial ligand such as a
chloride or hydroxide anion are inactive in aprotic sol-
vents.60,61 However, when porphyrins coordinated by a
strongly bound ligand are dissolved in a solvent containing
a protic component, they become active.62-66 Recent work
in our laboratory with iron(III) [tetrakis(pentafluorophenyl)]-
porphyrin chloride (F20TPP)FeCl has shown that protic
solvents enable the dissociation of (F20TPP)FeCl as a
consequence of their ability to solvate the Cl- anion and
coordinate to the (F20TPP)Fe+ cation.64,65 Protic solvents have
also been suggested to facilitate the heterolytic cleavage of
porphyrin-coordinated H2O2 via generalized acid catalysis
and thereby produce iron(IV) pi-radical cations that are active
intermediates in the epoxidation of olefins.62-65 While a
consensus appears to have emerged about the qualitative
effects of the axial ligand and solvent composition, the
magnitude of these effects expressed in terms of the rate
and equilibrium coefficients is only now beginning to be
understood.
(20) Montanari, F.; Banfi, S.; Quici, S. Pure Appl. Chem. 1989, 61, 1631-
1636.
(21) Castellino, A. J.; Briuce, T. C. J. Am. Chem. Soc. 1988, 110, 158-
162.
(22) Traylor, T. G.; Xu, F. J. Am. Chem. Soc. 1988, 110, 1953-1958.
(23) Groves, J. T.; Watanabe, Y. J. Am. Chem. Soc. 1988, 110, 8443-
8452.
(24) Tabushi, I. Coord. Chem. ReV. 1988, 86, 1-42.
(25) Ostovic, D.; Bruice, T. C. J. Am. Chem. Soc. 1988, 110, 6906-6908.
(26) Castellino, A. J.; Bruice, T. C. J. Am. Chem. Soc. 1988, 110, 1313-
1315.
The aim of the present study was to elucidate the
mechanism for epoxidation of cyclooctene using hydrogen
peroxide as the oxidant and an iron porphyrin with a weakly
coordinated ligand, (F20TPP)Fe(OTf), dissolved in an aprotic
solvent, acetonitrile, as the catalyst and to compare this
mechanism to that which we have reported for the chloride-
coordinated porphyrin in the presence of a protic solvent,
methanol.64,65 By working with a weakly coordinating ligand
in an aprotic solvent it was possible to determine the rate
parameters for the epoxidation of cyclooctene in the absence
of the effects of a protic solvent acting as either a ligand to
the iron porphyrin or as a cocatalyst in the heterolytic
cleavage of the O-O bond of coordinated H2O2. A com-
parison was also undertaken of the rate parameters for the
epoxidation of cyclooctene by H2O2 catalyzed by (F20TPP)-
Fe(OTf) and (F20TPP)FeCl in methanol, to establish whether
the catalytically active species in both cases is (F20TPP)Fe-
(CH3OH)+, as we have proposed previously.64,65
(27) Traylor, T. G.; Nakano, T.; Mikzstal, A. R.; Dunlap, B. E. J. Am.
Chem. Soc. 1987, 109, 3625-3632.
(28) Jorgensen, K. A. J. Am. Chem. Soc. 1987, 109, 698-705.
(29) Traylor, T. G.; Mikzstal, A. R. J. Am. Chem. Soc. 1987, 109, 2770-
2774.
(30) Mansuy, D. Pure Appl. Chem. 1987, 59, 759-770.
(31) Groves, J. R.; Watanabe, Y. J. Am. Chem. Soc. 1986, 108, 507-508.
(32) Collman, J. P.; Kodadek, T.; Brauman, J. I. J. Am. Chem. Soc. 1986,
108, 2588-2594.
(33) Meunier, B. Bull. Soc. Chim. Fr. 1986, 4, 578-594.
(34) Bruice, T. C. Ann. N. Y. Acad. Sci. 1986, 471, 83-98.
(35) Traylor, T. G.; Nakano, T.; Dunlap, B. E.; Traylor, P. S.; Dolphin, D.
J. Am. Chem. Soc. 1986, 108, 2782-2784.
(36) Groves, J. T.; Watanabe, Y. J. Am. Chem. Soc. 1986, 108, 507-508.
(37) Collman, J. P.; Kodadek, T.; Raybuck, S. A.; Brauman, J. I.; Papzian,
L. M. J. Am. Chem. Soc. 1985, 107, 4343-4345.
(38) Traylor, T. G.; Marsters, J. C.; Nakano, T.; Dunlap, B. E. J. Am. Chem.
Soc. 1985, 107, 5537-5539.
(39) Collman, J. P.; Brauman, J. I.; Meunier, B.; Hayashi, T.; Kodadek,
T.; Raybuck, S. A. J. Am. Chem. Soc. 1985, 107, 2000-2005.
(40) Guengerich, F. P.; MacDonald, T. L. Acc. Chem. Res. 1984, 17, 9-16.
(41) Collman, J. P.; Brauman, J. I.; Meunier, B.; Raybuck, S. A.; Kodadek,
T. Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 3245-3248.
(42) Razenberg, J. A. S. J.; Nolte, R. J. M.; Drenth, W. Tetrahedron Lett.
1984, 25, 789-792.
(43) Mansuy, D.; Battioni, P.; Renaud, J. P. J. Chem. Soc., Chem. Commun.
1984, 1255-1257.
Experimental Procedures
(44) Traylor, T. G.; Lee, W. A.; Stynes, D. V. J. Am. Chem. Soc. 1984,
106, 755-764.
Reagents. OmniSolv grade acetonitrile (99.99+%, water content
0.0022%), HPLC grade methanol (99.9+%), and ACS grade
chloroform (99.8%) were obtained from EMD Chemicals. Iron-
(III) [tetrakis(pentafluorophenyl)]porphyrin chloride, dodecane
(99+%), ACS grade hydrogen peroxide (30 wt %), and silver
trifluoromethanesulfonate (99.95+%) were obtained from Sigma
Aldrich. cis-Cyclooctene (95%) was obtained from Alfa-Aesar.
Deuterium oxide was obtained from Cambridge Isotope Labora-
tories, Inc. All chemicals were used without further purification or
(45) Traylor, P. S.; Dolphin, D.; Traylor, T. G. J. Chem. Soc., Chem.
Commun. 1984, 279-280.
(46) La Mar, G. N.; De Ropp, J. S.; Latos-Grazynski, L.; Balch, A. L.;
Johnson, R. B.; Smith, K. M.; Parish, D. W.; Cheng, R. J. J. Am.
Chem. Soc. 1983, 105, 782-787.
(47) Ortiz de Monteallano, P. R.; Mangold, B. L. K.; Wheeler, C.; Kunze,
K. L.; Reich, N. O. J. Biol. Chem. 1983, 258, 4208-4213.
(48) Collman, J. P.; Kodadke, T.; Raybuck, S. A.; Meunier, B. Proc. Natl.
Acad. Sci. U.S.A. 1983, 80, 7039-7041.
(49) Groves, J. T.; Myers, R. S. J. Am. Chem. Soc. 1983, 105, 5791-
5796.
(50) Groves, J. T.; Haushalter, R. C.; Nakamura, M.; Naemo, T. E.; Evans,
B. J. J. Am. Chem. Soc. 1981, 103, 2884-2886.
(51) Mansuy, D.; Chottard, J. C.; Lange, M.; Battioni, J. P. J. Mol. Catal.
1980, 7, 215-226.
(58) Traylor, T. G.; Popovitz-Biro, R. J. Am. Chem. Soc. 1988, 110, 239-
243.
(59) Lee, K. A.; Nam, W. J. Am. Chem. Soc. 1997, 119, 1916-1922.
(60) Nam, W.; Jin, S. W.; Lim, M. H.; Ryu, J. Y.; Kim, C. Inorg. Chem.
2002, 41, 3647-3652.
(61) 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.
(62) 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.
(63) Traylor, T. G.; Xu, F. J. Am. Chem. Soc. 1990, 112, 178-186.
(64) Stephenson, N. A.; Bell, A. T. Inorg. Chem. 2006, 45, 2758-2766.
(65) Stephenson, N. A.; Bell, A. T. J. Am. Chem. Soc. 2005, 127, 8635-
8643.
(52) Groves, J. T.; Nemo, T. E.; Myers, R. S. J. Am. Chem. Soc. 1979,
101, 1032-1033.
(53) Lowe, G. H.; Kert, C. J.; Hjelmeland, L. M.; Kirchner, R. F. J. Am.
Chem. Soc. 1977, 99, 3534-3536.
(54) Nordblum, G. D.; White, R. E.; Coon, M. J. Arch. Biochem. Biophys.
1976, 175, 524-533.
(55) Dunford, H. B.; Stillman, J. S. Coord. Chem. ReV. 1976, 19, 187-
251.
(56) Hrycay, E. G.; Gustafsson, J. A.; Ingelman-Sundberg, M.; Ernster, L.
Biochem. Biophys. Res. Commun. 1975, 66, 209-216.
(57) Lichtenberger, F.; Nastaincyzk, W.; Ullrich, V. Biochem. Biophys. Res.
Commun. 1975, 70, 939-946.
(66) Stephenson, N. A.; Bell, A. T. Inorg. Chem. 2006, 45, 5591-5599.
Inorganic Chemistry, Vol. 46, No. 6, 2007 2279

Stephenson and Bell
drying. As discussed below, water can coordinate to the iron(III)
porphyrin cations resulting in catalyst deactivation when only
acetonitrile is present. However, further drying of acetonitrile was
deemed unnecessary, since the acetonitrile, as received, contained
only 0.0022% water, based on Karl Fischer analysis. This concen-
tration of water is 2 orders of magnitude less than that added with
30 (w/w)% aqueous H2O2 under experimental conditions.
Synthesis of (F20TPP)Fe(OTf). Triflate-coordinated porphyrin
was synthesized based on the procedure of Nam et al.67 Silver
triflate (19.3 mg) was added to a magnetically stirred solution of
(F20TPP)FeCl (70 mg) dissolved in acetonitrile (7 mL). Conversion
of the porphyrin from the chloride-coordinated form to the triflate-
1
coordinated form was evidenced via H NMR by a shift in the
position of the â-pyrrole proton resonance from 81 to 60 ppm. The
initial addition of silver triflate (∼1 equiv based on porphyrin)
resulted in no change in coordination. Approximately four additional
equiv of silver triflate (77 mg) were then added, and the mixture
was stirred for 24 h. Triflate coordination was then evidenced by
the formation of a white precipitate (silver chloride) and the
1
appearance of a broadened H NMR peak at 60 ppm. Integration
Figure 1. Proposed mechanism for the epoxidation of cyclooctene by H2O2
of this peak indicated that approximately 75% of the chloride ligand
had been exchanged. Additional equiv of silver triflate were added
incrementally allowing 12-24 h between each addition. After the
catalyzed by (F20TPP)Fe(OTf).
respectively. The accuracy of measuring the epoxide concentration
via gas chromatography was (1% based on the analysis of a
calibrated standard. Blank reactions were carried out without
porphyrin catalyst present but with silver triflate present; these
experiments showed that the silver cations do not contribute
significantly to peroxide consumption.
1
addition of ∼10.5 total equiv, H NMR indicated that 96% of the
chloride anions had been replaced by triflate anions. The white
precipitate was then filtered from the reaction mixture, and two
additional equiv of silver triflate (34 mg) were added to the
1
porphyrin solution. The solution was filtered again, and H NMR
NMR Spectroscopy. 1H NMR was used to measure the
concentration of hydrogen peroxide in situ as a function of time
and to characterize the synthesized porphyrin. All experiments were
performed at room-temperature 22 °C; the temperature varied by
no more than (0.1 °C as measured via a thermocouple located 2
mm from the NMR sample tube. All experiments used a 400 MHz
Bruker VMX spectrometer. A detailed explanation of both proce-
dures can be found elsewhere.66,68
indicated that an additional 1% of the chloride anions had been
replaced by triflate anions. 19F NMR analysis of the porphyrin
solution revealed three resonances attributed to the fluorine sub-
stituents on the phenyl groups attached to the porphyrin ring and
one resonance attributed to the triflate anions. These features are
located at -134.6 ppm, -152.2 ppm, and -160.1 ppm for the
fluorine atoms located at the ortho-, para-, and meta-positions on
the phenyl groups and at -77.1 ppm for the fluorine atoms
associated with the triflate anions. Comparison of the areas for the
peaks attributed to phenyl fluorine atoms to the area of the peak
attributed to the fluorine atoms of the triflate anions indicated that
the concentration of the triflate anions was an order of magnitude
greater than that of the porphyrin. Since only a single 19F NMR
peak for triflate anions was observed at a position identical to that
for free triflate anions, it is believed that all of triflate anions behave
as free anions or weakly bound, outer-sphere ligands associated
with the porphyrin cations.
Reactions. To investigate the epoxidation of cyclooctene,
catalyst, solvent, cyclooctene, and dodecane, as an internal standard,
were combined in a 5-mL glass vial with magnetic stirring. The
solvent consisted of pure acetonitrile unless stated otherwise. The
cyclooctene, porphyrin, and hydrogen peroxide concentrations were
varied depending upon the experiment; however, the total reaction
volume was always 3.34 mL. Hydrogen peroxide (1-10 µL) was
added via a syringe to initiate the reactions. Reaction products were
analyzed by injecting a 0.5 µL sample of the reaction mixture into
an HP6890 series gas chromatograph fitted with an Agilent DB
Wax (30 m × 0.32 mm × 0.5 µm) capillary column and an FID
detector. The temperature program used for this analysis was as
follows: hold for 2.5 min at 50 °C, ramp at 20 °C/min to 175 °C,
and hold for 1.5 min at 175 °C. The split/splitless inlet was operated
at 200 °C and 14.2 psi with a split ratio of 5. Dodecane (an internal
standard) and cyclooctene epoxide were eluted at 5.5 and 8.2 min,
UV-Visible Spectroscopy. UV-visible spectroscopy was used
to characterize the axial coordination and oxidation state of the
porphyrin in situ under reaction conditions. Absorbance experiments
were conducted using a Varian Cary 400 Bio UV-visible spec-
trometer. The systematic error associated with UV-visible scans
was negligible. Reactions with hydrogen peroxide were studied in
situ by taking scans approximately every minute over the course
of 20-30 min after the addition of hydrogen peroxide. Hydrogen
peroxide (1 µL) was added to 3.34 mL of a reaction mixture
containing (F20TPP)Fe(OTf) (15 µM) and cyclooctene (0.14 M)
dissolved in acetonitrile. For reactions without substrate present,
additional acetonitrile was substituted for the cyclooctene to
maintain a constant reaction volume. Axial coordination of water
was studied by taking UV-visible scans after the addition of 0.5
µL aliquots of water (up to 10 µL total) to (F20TPP)Fe(OTf) (15
µM) dissolved in 3.34 mL of acetonitrile.
Reaction Mechanism. Nam and co-workers have presented
convincing evidence that the (F20TPP)Fe(OTf) catalyzed reaction
with hydrogen peroxide as the oxidant results in the formation of
pi-radical cation species as active intermediates for epoxidation.60
Based on this information and our previous work,64,65 the reaction
mechanism shown in Figure 1 is postulated. (To facilitate com-
parison, the reaction numbering sequence used here is identical to
that used in our earlier work with (F20TPP)FeCl.64,65) In this
(67) Nam, W.; Lim, M. H.; Lee, H. J.; Kim, C. J. Am. Chem. Soc. 2000,
122, 6641-6647.
(68) Stephenson, N. A.; Bell, A. T. Anal. Bioanal. Chem. 2005, 381, 1289-
1293.
2280 Inorganic Chemistry, Vol. 46, No. 6, 2007

Mechanistic Study of (F20TPP)Fe(OTf)
mechanism, iron(III) porphyrin cations react reversibly with H2O2.
The equilibrium constant for this reaction is K2. The coordinated
H2O2 then participates in reaction 3, as suggested by Nam and co-
workers,60 to produce the pi-radical cations and water via heterolytic
cleavage of the oxygen-oxygen bond of hydrogen peroxide
coordinated to (F20TPP)Fe+. In this process, the proton on the
proximal oxygen is believed to be transferred to the distal oxygen
by intramolecular rearrangement (see below). The oxygen-oxygen
bond may also be cleaved homolytically via reaction 4 to produce
an iron(IV) porphyrin species; this pathway results exclusively in
the decomposition of hydrogen peroxide. In addition to being active
for epoxidation via reaction 7, the pi-radical cations can also
contribute to peroxide decomposition via reaction 6. However, this
competitive pathway is minimal at high substrate concentrations.64,65
Iron(IV) species generated by either reaction 4 or 6 are regenerated
to the iron(III) state by reaction with additional H2O2, reaction 5.
The hydroperoxyl radicals generated in reactions 5 and 6 react
further to produce molecular oxygen,69 while the hydroxyl radicals
generated via reaction 4 are believed to contribute to oxidative
degradation of the porphyrin ring.65,70
Figure 2. The effect of hydrogen peroxide concentration on the yield and
observed rate constant. Porphyrin and substrate concentrations were 150
µM and 0.72 M, respectively.
The observed rate constant for peroxide decomposition, kobs, can
be expressed according to eq 1. Pseudo-steady-state is assumed
for the iron(IV) pi-radical cation species based on reaction 3 being
the rate determining step. Here Y∞ is the final yield of epoxide
based on the overall consumption of H2O2, the limiting reagent,
[(F20TPP)Fe+] is the concentration of iron(III) porphyrin cations,
and k3K2 is defined according to the mechanism shown in Figure
1. Y∞ can also be expressed in terms of the rate constants for
homolytic and heterolytic cleavage, eq 2. Derivations of kobs and
Y∞ are given elsewhere.64,65
Figure 3. The effect of water concentration on the observed rate constant.
The water concentration was varied both by changing the amount of H2O2
added and by direct addition of water. Porphyrin and substrate concentrations
were 150 µM and 0.72 M, respectively. Hydrogen peroxide concentration
was 2.9 mM for the experiments in which the concentration of H2O was
varied.
k3K2[(F20TPP)Fe+ ]
kobs
)
(1)
(2)
Y∞
k3
[epoxide]
[H2O2]0
Y∞ )
=
k3 + 2k4
sible for the change in the observed rate constant, deionized
water was added to the reaction mixture prior to the addition
of hydrogen peroxide. The results of this experiment are
shown in Figure 3, where kobs is plotted as a function of the
water concentration. The diamond shaped data points con-
nected by a solid line were obtained by varying the initial
concentration of hydrogen peroxide, and the circular data
points were obtained by holding the initial hydrogen peroxide
concentration constant and adding water. The coincidence
of the two sets of data indicates that kobs is independent of
the initial concentration of H2O2 but decreases with increas-
ing concentration of water.
A possible cause for the decrease in the reaction rate with
increasing water concentration is that water binds to the iron-
(III) porphyrin cation as an axial ligand resulting in a species
that does not participate in catalysis. To investigate this
possibility, porphyrin samples with different water concen-
trations were analyzed by UV-visible spectroscopy. In the
absence of water, the iron(III) porphyrin cations were
characterized by a Soret peak at 390 nm. As shown in Figure
4, the addition of water to the porphyrin solution resulted in
a decrease in the absorbance at 390 nm, the formation of a
shoulder on the Soret peak, and the formation of a new peak
near 565 nm. Isosbestic points were seen at 405, 505, and
537 nm. We suggest that these changes in the absorption
Results and Discussion
Effects of Oxidant Concentration. In our previous studies
of the mechanism of cyclooctene epoxidation by H2O2
catalyzed by (F20TPP)Fe+, we noted that both kobs and Y∞
are independent of the initial hydrogen peroxide concentra-
tion when the concentration of cyclooctene is sufficiently
high that the rate of reaction 6 is small relative to the rate of
reaction 7 (see Figure 1).64,65 Based on the mechanism shown
in Figure 1, both of these findings were also expected for
the current porphyrin system. However, as shown in Figure
2, while Y∞ is independent of the initial hydrogen peroxide
concentration, [H2O2]0, kobs decreases monotonically with
increasing [H2O2]0. Since hydrogen peroxide was added to
the system as a 30 (wt/wt)% aqueous solution, either water
or hydrogen peroxide could be responsible for the observed
behavior of kobs with [H2O2]0. The independence of Y∞ on
the oxidant concentration suggests that the cause for the
decrease in kobs occurs prior to reactions 3 and 4. To
determine whether water or hydrogen peroxide was respon-
(69) Nam, W.; Han, H. J.; Oh, S. Y.; Lee, Y. J.; Han, S. Y.; Kim, C.;
Woo, S. K.; Shin, W. J. Am. Chem. Soc. 2000, 122, 8677-8684.
(70) Cunningham, I. D.; Danks, T. N.; Hay, J. N.; Hamerton, I.; Gunathila-
gan, S. Tetrahedron 2001, 57, 6847-6853.
Inorganic Chemistry, Vol. 46, No. 6, 2007 2281

Stephenson and Bell
Figure 4. The effect of water on the UV-visible absorption spectrum.
The porphyrin concentration was 15 µM, while the concentration of added
water was varied from 0 to 166 mM.
Figure 6. Variation in [H2O2] with time in the presence of cyclooctene.
The data points are [H2O2] measured via 1H NMR, and the solid line is
[H2O2] predicted by the proposed mechanism. The total porphyrin concen-
tration was ∼150 µM, and the cyclooctene concentration was ∼0.72 M.
shown, between the observed rate constant and the concen-
tration of free iron(III) porphyrin cations. Based on the slope
obtained from this analysis, k3K2 was found to be 82 ( 8
s-1 M-1, and k4K2 was equal to 8 ( 1 s-1 M-1 based on eq
2. The final yield of epoxide was not affected significantly
by the porphyrin concentration.
The mechanism by which water inhibits the catalytic
activity of (F20TPP)Fe+ is not understood, and experiments
to gather further insight into this phenomenon were not
undertaken as a part of the present study. One possibility is
that the water-coordinated porphyrin species form µ-oxo
dimers, which are well-known to be catalytically inactive
for the oxidation of organic substrates.1,2 This interpretation
may also explain why water deactivates the porphyrin
catalyst while methanol does not. For two porphyrin species
to dimerize, the individual iron cations must be able to
approach each other to react. The alkyl group of a methanol
molecule coordinated to the iron(III) cation may provide
sufficient steric hindrance to inhibit the formation of µ-oxo
dimers.
Figure 5. The observed rate constant versus [(F20TPP)Fe+] determined
as a function of water concentration (12-116 mM).
spectrum were due to the coordination of water to the iron-
(III) porphyrin cations.
Water is thought to interact with the porphyrin cation as
shown in eq 3. The porphyrin cation is likely coordinated
by acetonitrile prior to coordination by water, and the
resulting species after water coordination is likely six-
coordinated with either one water molecule and one aceto-
nitrile or two water molecules occupying the axial positions.
However, the analysis discussed below is unaffected by initial
coordination by acetonitrile, since the concentration of
acetonitrile is large and essentially constant.
Consumption of Hydrogen Peroxide. The value of kobs
and the values of k3K2 and k4K2 reported above are based
solely on the measured rate of cyclooctene epoxide forma-
tion. As a further test of the reaction kinetics, the actual rate
1
of hydrogen peroxide consumption was measured by H
Kw
[(F20TPP)Fe+] + H2O { } [(F20TPP)Fe(H2O)+] (3)
\
NMR. As shown in Figure 6, the temporal variation in the
concentration of H2O2 is described accurately by the mech-
anism shown in Figure 1, and the rate parameters were
determined from measurements of the initial rate of cy-
clooctene epoxidation. A value of k5 ) 225 M-1 s-1 was
assumed, based upon previous studies of the kinetics for (F20-
TPP)FeCl.64,65 The validity of this assumption is discussed
below.
The rate of peroxide consumption in the absence of
cyclooctene was also determined by 1H NMR. As shown in
Figure 7, the consumption of hydrogen peroxide proceeds
as predicted by the proposed mechanism and the rate
parameters derived from the previous experiments with
cyclooctene present. Furthermore, the predicted rate of
peroxide decomposition in the absence of cyclooctene is very
sensitive to the rate parameter k5. This is due to the fact that
Based on eq 3 and the change in absorbance at 565 nm,
the value of Kw was determined to be 33 ( 6 M-1. To
determine whether eq 3 fully explains the nonlinear decrease
in reaction rate, the concentrations of water-coordinated and
free porphyrin cations were calculated, based on the above
relationship, for the reaction conditions shown in Figure 2.
As shown in Figure 5, the observed rate of reaction increases
linearly with the concentration of free porphyrin cations,
supporting the hypothesis that water-coordinated species are
catalytically inactive. From the slope of the line in Figure 5,
k3K2 is determined to be 90 ( 10 s-1 M-1, and k4K2 is
determined to be 9 ( 1 s-1 M-1 from eq 2. The relationship
shown in eq 3 was also tested by varying the total porphyrin
concentration. This also resulted in a linear correlation, not
2282 Inorganic Chemistry, Vol. 46, No. 6, 2007

Mechanistic Study of (F20TPP)Fe(OTf)
Figure 7. Variation in [H2O2] with time in the absence of cyclooctene.
Data points are [H2O2] measured via 1H NMR, and the solid line is [H2O2]
predicted by the proposed mechanism. The total porphyrin concentration
was ∼75 µM.
more than one-third of the peroxide is consumed via reaction
5 when cyclooctene is not present. The excellent agreement
between the NMR data and the kinetic model validates the
choice of 225 M-1 s-1 for k5 and indicates that the rate
coefficient for this reaction step is independent of the
composition of the axial ligand associated with the porphyrin.
Formation of Iron(IV) Species. As shown above, the
proposed model for the reaction kinetics accurately predicts
the rates of hydrogen peroxide consumption both in the
presence and absence of cyclooctene. The predictions of the
model also agree qualitatively with observations made by
UV-visible spectroscopy. The porphyrin reaction mixture
was analyzed as a function of time after the addition of
hydrogen peroxide both with and without cyclooctene present
in order to establish evidence for the formation of iron(IV)
species (see Figure 1). The proposed kinetics predict that
when cyclooctene is present approximately ∼1% of the
porphyrin will be present as iron(IV) species; approximately
10% of the porphyrin is expected to be present as iron(IV)
species when cyclooctene is absent. Iron (IV) species can
be identified by peaks at 410 and 540 nm in the UV-visible
spectroscopy.71 As seen in Figure 8a, there is no significant
change in the shape of the UV-visible spectrum taken at
different reaction times when cyclooctene is present. The
decrease in the intensity of the spectra as a function of time
is due to porphyrin degradation. However, as shown in Figure
8b, a shoulder near 410 nm appears on the Soret peak when
cyclooctene is absent. In addition, as shown in the inset of
Figure 8b, there is a small increase in the absorbance near
540 nm after the addition of hydrogen peroxide. As predicted
by the model of the reaction kinetics, iron(III) species
predominate but there is an observable contribution to the
UV-visible spectra from iron(IV) species in the absence of
cyclooctene.
Figure 8. In situ UV-visible spectra taken as a function of time: (a)
with cyclooctene present and (b) without cyclooctene present. The initial
concentrations of porphyrin, peroxide, and cyclooctene were 15 µM, 2.5
mM, and 0.14 M, respectively.
Figure 9. The yield of cyclooctene epoxide as a function of cyclooctene
concentration. The concentration of H2O2 and porphyrin were 14.7 mM
and 150 µM, respectively, for all experiments. The data points are for
epoxide yield determined by gas chromatography, and the solid line
represents the yield predicted by the proposed model.
pete with reaction 7 at high substrate concentrations. The
ratio k6/k7 was determined by fitting the predicted to the
observed variation in the yield of cyclooctene epoxide as a
function of the concentration of cyclooctene. As seen in Fig-
ure 9, an excellent fit is achieved for a value of k6/k7 ) 0.6.
Heterolytic Cleavage of the O-O Bond of Coordinated
H2O2. Nam and co-workers have presented convincing
evidence that heterolytic cleavage of the O-O bond in
coordinated H2O2 occurs during the epoxidation of cy-
Effects of Substrate Concentration. The substrate con-
centration was varied to determine the ratio of the rate con-
stants for reactions 6 and 7 (see Figure 1). As seen in Figure
9, the final yield is independent of the substrate concentration
above about 0.4 M, indicating that reaction 6 does not com-
(71) Lee, K. A.; Nam, W. Bull. Korean Chem. Soc. 1996, 17, 669-671.
Inorganic Chemistry, Vol. 46, No. 6, 2007 2283

Stephenson and Bell
Figure 10. Mechanism suggested by Sawyer and co-workers for the
activation of H2O2 by Fe3+ in acetonitrile.
Figure 11. The effect of [H2O2] on kobs and epoxide yield in a (1:1) mixture
of acetonitrile and methanol. Porphyrin and substrate concentrations were
75 µM and 0.72 M for all experiments.
clooctene by hydrogen peroxide in acetonitrile catalyzed by
(F20TPP)Fe(OTf).60 Two scenarios can be envisioned to
explain how a hydrogen atom is transferred from the
proximal to the distal oxygen of coordinated H2O2 either
during or prior to heterolytic cleavage of the O-O bond in
an aprotic solvent. The first explanation is that proton transfer
is facilitated by some species in the reaction mixture. While
water and hydrogen peroxide are obvious candidates for this
role, we have shown above that water deactivates the catalyst
rather than promoting catalytic activity and that the observed
kinetics are independent of the hydrogen peroxide concentra-
tion. Experiments in which the concentration of free triflate
was varied by the addition of silver triflate also did not affect
the observed kinetics or final yield, indicating that triflate
anions do not affect the cleavage of hydrogen peroxide.
The second possibility is that the hydrogen is transferred
intramolecularly within the coordinated H2O2. Although such
a scheme has not been proposed previously for porphyrins,
Sawyer and co-workers have proposed such a process for
iron salts dissolved in acetonitrile.72-76 FeCl3 was hypoth-
esized to form a Lewis acid-Lewis base adduct equilibrated
between a side-on and end-on form of the iron-hydrogen
peroxide intermediate, as shown in Figure 10, inducing either
a biradical (homolytic cleavage) or an oxene (heterolytic
cleavage) character to the coordinated peroxide molecule.73-75
Sawyer and co-workers later proposed that the oxene-like
intermediate leads to the formation of an iron(V) oxo species
which exhibits chemical characteristics and reactivities
similar to those of iron(IV) pi-radical cation porphyrin
species found in nature.72 We suggest that a similar mech-
anism may account for the heterolytic cleavage of the O-O
bond in H2O2 coordinated to (F20TPP)Fe+ in acetonitrile and
that the iron(IV) hydroxo species may be formed via
homolytic cleavage of the biradical species followed by a
radical rebound step.
acetonitrile was used.64,65 To investigate the effect of water
in the presence of methanol, (F20TPP)Fe(OTf) was dissolved
in a 1:1 mixture of acetonitrile and methanol. As shown in
Figure 11, neither Y∞ nor kobs are affected by the initial
concentration of H2O2 (i.e., the water content) added to the
reaction mixture; this in complete agreement with our earlier
observations.64,65
The extent of methanol coordination was investigated
using UV-visible spectroscopy. The addition of microliter
aliquots of methanol to a 15 µM solution of (F20TPP)Fe-
(OTf) dissolved in acetonitrile caused a shift in the position
of the Soret peak from 390 to 388 nm. This shift occurred
at low methanol concentrations, and complete conversion
to the methanol-coordinated species was evidenced by no
further change in the spectrum for methanol concentrations
above ∼1 vol %. This observation suggests that the equi-
librium constant for coordination of methanol to the iron-
(III) porphyrin cation is of the same order of magnitude as
that found for coordination by water. Thus, when methanol
is present in large excess relative to water, methanol will
displace all of the water present in the coordination sphere
of the iron(III) porphyrin cation. Coordination of methanol
to the iron(III) porphyrin cation was also verified by 1H NMR
spectroscopy, as indicated by a shift in the â-pyrrole
resonance from 60 to 65 ppm66 upon the addition of
methanol.
Comparison of Protic and Aprotic Solvent Systems. As
shown here and in our previous work,64-66 the mechanism
for (F20TPP)Fe-catalyzed epoxidation of cyclooctene by
hydrogen peroxide is similar for both aprotic and protic
solvent systems. For homolytic cleavage of the O-O bond
in coordinated H2O2, the value of k4K2 is smaller in methanol
than in acetonitrile, 5 s-1 M-1 versus 9 s-1 M-1. And for
heterolytic cleavage of the O-O bond, the value of k3K2[CH3-
OH] ) 13 [CH3OH] M-1 s-1, which for pure methanol gives
a value of 290 M-1 s-1, whereas k3K2 ) 86 M-1 s-1 for an
acetonitrile solution. The faster rate of heterolytic cleavage
and slower rate of homolytic cleavage for the protic system
results in a value of Y∞ as high as 96%, while Y∞ is only
83% in the aprotic system. The iron(IV) pi-radical cations
also exhibit a higher reactivity toward cyclooctene relative
to H2O2 in protic versus aprotic solvents (k6/k7 ) 0.30 vs
0.60). By contrast, though, the reactivity of iron(IV) hydroxo
Role of Methanol Addition. The finding that water
coordinated to the iron(III) porphyrin cation makes this
species catalytically inactive raises the question as to whether
this phenomenon affected any of our previous work in which
a solvent system containing a mixture of alcohol and
(72) Tung, H. C.; Kang, C.; Sawyer, D. T. J. Am. Chem. Soc. 1992, 114,
3445-3455.
(73) Sugimoto, H.; Sawyer, D. T. J. Org. Chem. 1985, 50, 1784-1786.
(74) Sawyer, D. T. CHEMTECH 1988, 369-375.
(75) Strukul, G. Catalytic Oxidations with Hydrogen Peroxide as Oxidant;
Kluwer Academic Publishers: Boston, MA, 1992.
(76) Bruice, T. C. Acc. Chem. Res. 1991, 24, 243-249.
2284 Inorganic Chemistry, Vol. 46, No. 6, 2007

Mechanistic Study of (F20TPP)Fe(OTf)
cations is not affected by the nature of the solvent (k5 )
225 M-1 s-1 for both systems). Finally, a greater extent of
porphyrin degradation is observed for the aprotic solvent
system than for the protic solvent system. This relationship
is consistent with the observation of a higher value of k4K2
for the aprotic solvent system, since porphyrin degradation
is believed to result from the attack of free radicals (e.g.,
OH•) on the porphyrin ring, which are produced via reactions
4 and 6.64,65,70 Decreased porphyrin degradation in methanol
may also result from scavenging of the hydroxyl radical by
methanol.
Experiments were also conducted to study the kinetics of
the triflate-coordinated species in methanol (20 M). Under
these conditions, the observed rate constant was 1.02 min-1,
and the yield of epoxide was 96%. It is believed that in this
case the catalyst operates according to the same mechanism
as that proposed previously for (F20TPP)FeCl dissolved in
methanol and that, independent of the initial porphyrin salt
used, [(F20TPP)Fe(CH3OH)]+ is the precursor to the pi-
radical cation species. In support of this hypothesis, the value
of kobs was determined to be 1.02 min-1, and Y∞ was
determined to be 96%, corresponding to values of k3K2 )
11 [MeOH] M-1 s-1 and k4K2 ) 5 M-1 s-1 for (F20TPP)-
Fe(OTf) dissolved in 20 M methanol. The corresponding
values are k3K2 ) 13 [MeOH] M-1 s-1 and k4K2 ) 5 M-1
s-1 for (F20TPP)FeCl dissolved in methanol.10
of a protic solvent.64,65 In both systems, only iron(III)
porphyrin species that are not coordinated by a strongly
bound axial ligand are catalytically active. The main differ-
ence between the two solvent systems is that the mechanism
for cleavage of the oxygen-oxygen bond of coordinated
hydrogen peroxide appears to be intramolecular in an aprotic
solvent and acid-catalyzed in protic solvent. The rate
parameters obtained in this study of iron porphyrin-catalyzed
reaction occurring in an aprotic solvent can be compared
directly with the rate parameters for the same reaction
occurring in a protic solvent. In an aprotic solvent, hydrogen
peroxide is used less efficiently for epoxidation of cy-
clooctene than in a protic solvent, and porphyrin degradation
occurs much more rapidly than in a protic solvent. The higher
rates of hydrogen peroxide decomposition and porphyrin
degradation are both a consequence of the higher rate of free
radical generation via homolytic cleavage of coordinated
H2O2 occurring in an aprotic solvent. It was also observed
that water can deactivate the catalyst in aprotic solvent by
coordinating to the iron porphyrin. Since water can be
displaced from the coordination sphere of the iron(III)
porphyrin by methanol, the water present in a hydrogen
peroxide solution does not affect the catalytic activity of (F20-
TPP)Fe+ dissolved in methanol. This is a significant finding,
suggesting that future studies of iron porphyrins in aprotic
solvents can be affected by small amounts of water.
Conclusions
Acknowledgment. This work was supported by the
Director, Office of Basic Energy Sciences, Chemical Sci-
ences Division of the U.S. Department of Energy under
Contract DE-AC02-05CH11231.
A mechanism has been presented for (F20TPP)Fe(OTf)-
catalyzed epoxidation of cyclooctene in acetonitrile using
hydrogen peroxide as the oxidant and has been shown to be
consistent with all experimental evidence. This mechanism
is similar to that presented for (F20TPP)FeCl in the presence
IC060757C
Inorganic Chemistry, Vol. 46, No. 6, 2007 2285