Olefin cis-dihydroxylation versus epoxidation by non-heme iron catalysts: Two faces of an FeIII-OOH coin

Article abstract of DOI:10.1021/ja0120025

The oxygenation of carbon-carbon double bonds by iron enzymes generally results in the formation of epoxides, except in the case of the Rieske dioxygenases, where cis-diols are produced. Herein we report a systematic study of olefin oxidations with H₂0₂ catalyzed by a group of non-heme iron complexes, i.e., [Fe^II(BPMEN)CH₃CN)₂]²⁺ (1, BPMEN = N,N′-dimenthyl-N,N′-bis(2-pyridylmethyl)-1,2-diaminoethane) and [Fe^II(TPA)(CH₃CN)₂]²⁺ (4, TPA = tris(2-pyridlymethyl)amine) and their 6- and 5-methyl-substituted derivatives. We demonstrate that olefin epoxidation and cis-dihydroxylation are different facets of the reactivity of a common Fe^III-OOH intermediate, whose spin state can be modulated by the electronic and steric properties of the ligand environment. Highly stereoselective epoxidation is favored by catalysts with no more than one 6-methyl substituent, which give rise to low-spin Fe^III-OOH species (category A). On the other hand, cis-dihydroxylation is favored by catalysts with more than one 6-methyl substituent, which afford high-spin Fe^III-OOH species (category B). For catalysts in category A, both the epoxide and the cis-diol product incorporate ¹⁸O from H₂¹⁸O, results that implicate a cis-H¹⁸O-Fe^v=O species derived from O-O bond heterolysis of a cis-H₂¹⁸O-Fe^III-OOH intermediate. In contrast, catalysts in category B incorporate both oxygen atoms from H₂¹⁸O₂ into the dominant cis-diol product, via a putative Fe^III-η²-OOH species. Thus, a key feature of the catalysts in this family is the availability of two cis labile sites, required for peroxide activation. The olefin epoxidation and cis-dihydroxylation studies described here not only corroborate the mechanistic scheme derived from our earlier studies on alkane hydroxylation by this same family of catalysts (Chen, K.; Que, L, Jr. J. Am. Chem. Soc. 2001, 123, 6327) but also further enhance its credibility. Taken together, these reactions demonstrate the catalytic versatility of these complexes and provide a rationale for Nature's choice of ligand environments in biocatalysts that carry out olefin oxidations.

Full text of DOI:10.1021/ja0120025

Published on Web 02/28/2002
Olefin Cis-Dihydroxylation versus Epoxidation by Non-Heme
Iron Catalysts: Two Faces of an FeIII-OOH Coin
Kui Chen, Miquel Costas, Jinheung Kim, Adrianne K. Tipton, and
Lawrence Que, Jr.*
Contribution from the Department of Chemistry and Center for Metals in Biocatalysis,
UniVersity of Minnesota, Minneapolis, Minnesota 55455
Received August 20, 2001
Abstract: The oxygenation of carbon-carbon double bonds by iron enzymes generally results in the
formation of epoxides, except in the case of the Rieske dioxygenases, where cis-diols are produced. Herein
we report a systematic study of olefin oxidations with H2O2 catalyzed by a group of non-heme iron complexes,
i.e., [FeII(BPMEN)(CH3CN)2]2+ (1, BPMEN ) N,N′-dimethyl-N,N′-bis(2-pyridylmethyl)-1,2-diaminoethane)
and [FeII(TPA)(CH3CN)2]2+ (4, TPA ) tris(2-pyridylmethyl)amine) and their 6- and 5-methyl-substituted
derivatives. We demonstrate that olefin epoxidation and cis-dihydroxylation are different facets of the
reactivity of a common FeIII-OOH intermediate, whose spin state can be modulated by the electronic and
steric properties of the ligand environment. Highly stereoselective epoxidation is favored by catalysts with
no more than one 6-methyl substituent, which give rise to low-spin FeIII-OOH species (category A). On
the other hand, cis-dihydroxylation is favored by catalysts with more than one 6-methyl substituent, which
afford high-spin FeIII-OOH species (category B). For catalysts in category A, both the epoxide and the
cis-diol product incorporate 18O from H218O, results that implicate a cis-H18OsFeVdO species derived from
O-O bond heterolysis of a cis-H218O-FeIII-OOH intermediate. In contrast, catalysts in category B
incorporate both oxygen atoms from H218O2 into the dominant cis-diol product, via a putative FeIII-η2-OOH
species. Thus, a key feature of the catalysts in this family is the availability of two cis labile sites, required
for peroxide activation. The olefin epoxidation and cis-dihydroxylation studies described here not only
corroborate the mechanistic scheme derived from our earlier studies on alkane hydroxylation by this same
family of catalysts (Chen, K.; Que, L, Jr. J. Am. Chem. Soc. 2001, 123, 6327) but also further enhance its
credibility. Taken together, these reactions demonstrate the catalytic versatility of these complexes and
provide a rationale for Nature’s choice of ligand environments in biocatalysts that carry out olefin oxidations.
The oxygen activation mechanisms for hydrocarbon oxida-
tions by heme and non-heme iron catalysts have continued to
fascinate bioinorganic chemists for many years.1-4 Understand-
ing such mechanisms may lead to new insights in biocatalysis
and drug design as well as the development of new approaches
for industrial catalysis. Nature has evolved iron enzymes that
carry out alkane hydroxylation, olefin epoxidation, and olefin
cis-dihydroxylation. The first two transformations are carried
out by catalysts with either heme (e.g., cytochrome P450)1 or
non-heme iron centers (e.g., bleomycin, methane monooxy-
genase),2,3 while cis-dihydroxylation is thus far a unique activity
of the non-heme iron-containing Rieske dioxygenases, which
attack arene double bonds in the first step of the biodegradation
of arenes by soil bacteria.5
Cytochrome P450 serves as the paradigm of O2 activation
by iron enzymes;1 its mechanism of action is generally accepted
to involve a (porphyrin radical)FeIVdO species,6,7 although
recent evidence has raised the possibility that a (porphyrin)-
FeIII-OOH species can also be the oxidant in some instances.8-13
The proposed mechanisms for non-heme iron enzymes generally
follow the heme paradigm. For methane monooxygenase, an
FeIV2(µ-O)2 intermediate is proposed to be the analogue of the
(porphyrin radical)FeIVdO oxidant,14 while an FeIIIsOOH
(6) Groves, J. T.; Han, Y.-Z. In Cytochrome P450: Structure, Mechanism,
and Biochemistry, 2nd ed.; Ortiz de Montellano, P. R., Ed.; Plenum Press:
New York, 1995; pp 3-48.
(7) Schlichting, I.; Berendzen, J.; Chu, K.; Stock, A. M.; Maves, S. A.; Benson,
D. E.; Sweet, R. M.; Ringe, D.; Petsko, G. A.; Sligar, S. G. Science 2000,
287, 1615-1622.
(8) Pratt, J. M.; Ridd, T. I.; King, L. J. Chem. Commun. 1995, 2297-2298.
(9) Vaz, A. D. N.; Pernecky, S. J.; Raner, G. M.; Coon, M. J. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 4644-4648.
(1) Sono, M.; Roach, M. P.; Coulter, E. D.; Dawson, J. H. Chem. ReV. 1996,
96, 2841-2887.
(10) Lee, K. A.; Nam, W. J. Am. Chem. Soc. 1997, 119, 1916-1922.
(11) Newcomb, M.; Toy, P. H. Acc. Chem. Res. 2000, 33, 449-455.
(12) Nam, W.; Lim, M. H.; Lee, H. J.; Kim, C. J. Am. Chem. Soc. 2000, 122,
6641-6647.
(2) Wallar, B. J.; Lipscomb, J. D. Chem. ReV. 1996, 96, 2625-2658.
(3) Que, L., Jr.; Ho, R. Y. N. Chem. ReV. 1996, 96, 2607-2624.
(4) Solomon, E. I.; Brunold, T. C.; Davis, M. I.; Kemsley, J. N.; Lee, S.-K.;
Lehnert, N.; Neese, F.; Skulan, A. J.; Yang, Y.-S.; Zhou, J. Chem. ReV.
2000, 100, 235-349.
(13) Nam, W.; Lim, M. H.; Moon, S. K.; Kim, C. J. Am. Chem. Soc. 2000,
122, 10805-10809.
(5) Gibson, D. T. Microbial Degradation of Organic Compounds; Marcel
Dekker: New York, 1984.
(14) Shu, L.; Nesheim, J. C.; Kauffmann, K.; Mu¨nck, E.; Lipscomb, J. D.; Que,
L., Jr. Science 1997, 275, 515-518.
9
3026 VOL. 124, NO. 12, 2002 J. AM. CHEM. SOC.
10.1021/ja0120025 CCC: $22.00 © 2002 American Chemical Society

Olefin Cis-Dihydroxylation versus Epoxidation
A R T I C L E S
Chart 1. Ligands Used in This Study and the Numbering Scheme
the complexes studied here represent the first examples of iron
catalysts capable of olefin cis-dihydroxylation. In this study,
we provide strong mechanistic evidence for the involvement
of a formally non-heme iron(V)-oxo species in these reactions
and demonstrate that olefin epoxidation and cis-dihydroxylation
activities derive from a common FeIII-OOH intermediate whose
properties and reactivity can be tuned by the ligand environment.
Some aspects of this work have been published previously as
communications.25-27
for Corresponding Iron Complexes
Experimental Section
Materials and Syntheses. All reagents were purchased from Aldrich
and used as received unless noted otherwise. H218O (85 or 96.5% 18O-
enriched), H218O2 (90% 18O-enriched, 2% solution in H216O), and 18O2
(96% 18O-enriched) were obtained from ICON. H218O (88.8 or 98.6%
18O-enriched) was obtained from Isotec. Cyclooctene was purified by
vacuum distillation, while other olefins were purified by passing through
silica gel immediately before the reactions. The epoxides of cis- and
trans-2-heptenes were obtained from stereospecific epoxidation by
m-chloroperoxybenzoic acid, and the corresponding diols were from
the hydrolysis of the epoxides.
The syntheses of all complexes used in this study except for 2 and
3 have been previously reported; these complexes and corresponding
references are listed as follows: [FeII(BPMEN)(CH3CN)2](ClO4)2 (1),28
[FeII(TPA)(CH3CN)2](ClO4)2 (4),29 [FeIII2O(TPA)2(H2O)2](ClO4)4 (4a),30
[FeII(5-Me3-TPA)(CH3CN)2](ClO4)2 (5),22 [FeII(6-Men-TPA)(CH3CN)2]-
(ClO4)2 (6, n ) 1; 7, n ) 2; 8a, n ) 3),29 [FeII(6-Me3-TPA)(O2CC6H4X)]-
(ClO4) (8b, X ) 4-Me; 8c, X ) H; 8d, X ) 3-NO2),31 and
[FeII(N4Py)(CH3CN)](ClO4)2 (9).32 Caution: Perchlorate salts are
potentially explosiVe and should be handled with care.
Complexes 2 and 3 were synthesized by mixing equimolar amounts
of FeII(OTf)2‚2CH3CN (OTf ) trifluoromethanesulfonate) and the
tetradentate ligand in THF solution to afford a precipitate after a few
minutes of stirring. The precipitate was dissolved in CH2Cl2 and layered
with hexane to afford the products as yellow (2) or white (3) crystalline
solids upon standing at -20 °C for 3 days under Ar. Elemental analyses
were performed by Atlantic Microlab (Norcross, GA). [FeII(5-Me2-
BPMEN)(OTf)2] (2). Anal. Calcd (found) for C20H26F6FeN4O6S2‚
0.5H2O: C, 36.32 (36.13); H, 4.11 (3.94); N, 8.47 (8.40). [FeII(6-Me2-
BPMEN)(OTf)2] (3). Anal. Calcd (found) for C20H26F6FeN4O6S2‚
0.5H2O: C, 36.32 (36.15); H, 4.11 (4.07); N, 8.47 (8.38).
intermediate has been observed in the reaction cycle of the
antitumor drug bleomycin.15-18 These intermediates are believed
to be involved in the hydroxylation of aliphatic C-H bonds
and the epoxidation of C-C double bonds.
On the other hand, scant insight is available for the cis-
dihydroxylation of C-C double bonds by Rieske dioxygenases,
where both oxygen atoms of O2 are incorporated into the
product.5 The active site of naphthalene 1,2-dioxygenase consists
of a mononuclear iron center coordinated by two histidines and
a bidentate carboxylate group, with the remaining two cis sites
available for exogenous ligand binding.19 Evidence for an iron-
peroxo intermediate is suggested by the crystal structure of a
putative indole-dioxygen adduct bound to the mononuclear iron
center,20 but no intermediates have been detected spectroscopi-
cally in the catalytic cycle of these enzymes.21 To date,
mechanistic speculations center on the involvement of FeIII-η2-
peroxo or HOsFeVdO species.3,21
In the course of developing functional models for non-heme
iron oxygenases, we have discovered a family of non-heme iron
catalysts (Chart 1) that are capable of stereospecific hydrocarbon
oxidations with H2O2 as the oxidant. We have reported a detailed
mechanistic study of stereospecific alkane hydroxylation by this
family of catalysts and found evidence for the participation of
an FeVdO species.22 Here we present a systematic study of
olefin oxidation by these catalysts. Although examples of olefin
epoxidation with H2O2 by non-heme iron catalysts are known,23,24
Crystallographic Studies. Crystals suitable for crystallographic
analysis were obtained from CH2Cl2/hexane for 2 and from CH2Cl2/
diethyl ether for 3. Data collection and analysis were conducted on a
Siemens SMART system at the X-ray Crystallographic Laboratory of
the Chemistry Department of the University of Minnesota. Pertinent
crystallographic data and experimental conditions are summarized in
Table S1 (Supporting Information). The structures were solved by direct
methods using the SHELXTL V5.0 suite of programs. All non-hydrogen
atoms were refined anisotropically, and hydrogen atoms were placed
in ideal positions and refined as riding atoms with individual (or group
if appropriate) isotropic displacement parameters.
Instrumentation. 1H NMR spectra were recorded on a Varian Unity
300 or 500 spectrometer at ambient temperature. Chemical shifts (ppm)
(15) Stubbe, J.; Kozarich, J. W. Chem. ReV. 1987, 87, 1107-1136.
(16) Sam, J. W.; Tang, X.-J.; Peisach, J. J. Am. Chem. Soc. 1994, 116, 5250-
5256.
(17) Burger, R. M. Struct. Bonding 2000, 97, 287-303.
(18) Neese, F.; Zaleski, J. M.; Zaleski, K. L.; Solomon, E. I. J. Am. Chem. Soc.
2000, 122, 11703-11724.
(25) Kim, C.; Chen, K.; Kim, J.; Que, L., Jr. J. Am. Chem. Soc. 1997, 119,
5964-5965.
(26) Chen, K.; Que, L., Jr. Angew. Chem., Int. Ed. Engl. 1999, 38, 2227-2229.
(27) Costas, M.; Tipton, A. K.; Chen, K.; Jo, D.-H.; Que, L., Jr. J. Am. Chem.
Soc. 2001, 123, 6722-6723.
(19) Kauppi, B.; Lee, K.; Carredano, E.; Parales, R. E.; Gibson, D. T.; Eklund,
H.; Ramaswamy, S. Structure 1998, 6, 571-586.
(20) Carredano, E.; Karlsson, A.; Kauppi, B.; Choudhury, D.; Parales, R. E.;
Parales, J. V.; Lee, K.; Gibson, D. T.; Eklund, H.; Ramaswamy, S. J. Mol.
Biol. 2000, 296, 701-712.
(28) Chen, K.; Que, L., Jr. Chem. Commun. 1999, 1375-1376.
(29) Zang, Y.; Kim, J.; Dong, Y.; Wilkinson, E. C.; Appelman, E. H.; Que, L.,
Jr. J. Am. Chem. Soc. 1997, 119, 4197-4205.
(21) Wolfe, M. D.; Parales, J. V.; Gibson, D. T.; Lipscomb, J. D. J. Biol. Chem.
2001, 276, 1945-1953.
(30) Dong, Y.; Fujii, H.; Hendrich, M. P.; Leising, R. A.; Pan, G.; Randall, C.
R.; Wilkinson, E. C.; Zang, Y.; Que, L., Jr.; Fox, B. G.; Kauffmann, K.;
Mu¨nck, E. J. Am. Chem. Soc. 1995, 117, 2778-2792.
(22) Chen, K.; Que, L., Jr. J. Am. Chem. Soc. 2001, 123, 6327-6337.
(23) Nam, W.; Ho, R. Y. N.; Valentine, J. S. J. Am. Chem. Soc. 1991, 113,
7052-7054.
(31) Kim, J.; Zang, Y.; Costas, M.; Harrison, R. G.; Wilkinson, E. C.; Que, L.,
Jr. J. Biol. Inorg. Chem. 2001, 6, 276-284.
(24) Guajardo, R. J.; Hudson, S. E.; Brown, S. J.; Mascharak, P. K. J. Am.
Chem. Soc. 1993, 115, 7971-7977.
(32) Lubben, M.; Meetsma, A.; Wilkinson, E. C.; Feringa, B.; Que, L., Jr. Angew.
Chem., Int. Ed. Engl. 1995, 34, 1512-1514.
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 3027

A R T I C L E S
Chen et al.
were referenced to the residual protic solvent peaks. Electrospray
ionization mass spectral experiments were carried out on a Sciex API
III mass spectrometer (Sciex, Thornhill, ON, Canada).33 Product
analyses were performed on a Perkin-Elmer Sigma 3 gas chromatograph
(AT-1701 column, 30 m) with a flame ionization detector. GC mass
spectral analyses were performed on a HP 5898 GC (DB-5 column,
60 m) with a Finnigan MAT 95 mass detector or a HP 6890 GC (HP-5
column, 30 m) with an Agilent 5973 mass detector. A 4% NH3/CH4
mix was used as the ionization gas for chemical ionization analyses.
Reaction Conditions for Catalytic Oxidations. In a typical reaction,
0.3 mL of a 70 mM H2O2 solution (diluted from a 35 or 50% H2O2
solution) in CH3CN was delivered by syringe pump over 30 min at 25
°C in air to a vigorously stirred CH3CN solution (2.7 mL) containing
iron catalyst and olefin substrate. The final concentrations were 0.7
mM mononuclear iron catalyst (or 0.35 mM dinuclear iron catalyst
4a), 7.0 mM H2O2, and 0.70 M olefin. The solution was stirred for
another 5 min after syringe pump addition was complete. The organic
products were either directly extracted into diethyl ether solution or
esterified by 1 mL of acetic anhydride together with 0.1 mL of
1-methylimidazole34 followed by CHCl3 extraction. To the organic
extract was added an internal standard and the solution was subjected
to GC analysis. The products were identified by comparison of their
GC retention times, GC/MS, and 1H and 13C NMR spectra with those
of authentic compounds.
Isotope Labeling Studies. Similar conditions as described above
were used for isotope labeling studies except for the following details.
In experiments with H218O, 42 µL of H218O (0.70 M) was added to the
catalyst solution prior to the injection of H2O2. In experiments with
H218O2, 7.0 mM H218O2 (diluted by CH3CN from a 2% H218O2/H2O
solution) was used instead of H2O2. The organic products were subjected
to GC/CI-MS analyses. All labeling experiments were run at least in
duplicate. The data reported were the average of these reactions and
calculated on the basis of the 18O-enrichment of the reagents containing
the isotope.
Results
In a recent paper, we reported an investigation of a series of
non-heme iron(II) complexes with respect to their ability to
catalyze alkane hydroxylation with H2O2 as oxidant and found
a number of these carry out stereospecific hydroxylation.22 In
this paper, we focus on a similar series of catalysts and their
ability to oxidize olefins. All complexes in this study are six-
coordinate with a tetradentate N4 ligand based on either the
tripodal tris(2-pyridylmethyl)amine (TPA) or the linear N,N′-di-
methyl-N,N′-bis(2-pyridylmethyl)-1,2-diaminoethane (BPMEN)
framework (Chart 1). The remaining coordination sites are
occupied by labile ligands such as CH3CN, trifluoromethane-
sulfonate (OTf), or benzoate.
Physical Properties of the Catalysts. Several of the com-
plexes have been crystallographically characterized.22,28,29,31
Exemplifying the family of TPA-based complexes, 4, 5, 8a,
and 8c have two cis coordination sites for CH3CN or benzoate.
This geometry is also maintained in the BPMEN family of
complexes. Although the BPMEN ligand can in principle adopt
three possible conformations, only one of these is observed for
complexes 1-3 (Figure 1). The tetradentate ligand adopts a
cis-R conformation with the two pyridine rings coordinated to
the metal center trans to each other. While we were not able to
crystallize the bis(acetonitrile) derivatives of 2 and 3, we were
successful with the corresponding bis(triflate) complexes, and
Figure 1. ORTEP plots for [Fe(5-Me2-BPMEN)(OTf)2] (2) and [Fe(6-
Me2-BPMEN)(OTf)2] (3) showing 50% probability ellipsoids. Hydrogen
atoms omitted for clarity.
crystallographic data for these two complexes can be found in
the Supporting Information. The cis-R conformation favored by
the BPMEN family of ligands in these complexes requires the
two remaining ligands, either CH3CN or OTf, to occupy sites
that are cis to each other, as in the TPA family of complexes.
From their solid-state structures, it can be deduced that 4 and
5, representing the TPA and 5-Me3-TPA complexes, respec-
tively, have low-spin iron(II) centers, with Fe-N bond lengths
of less than 2 Å (Table 1). The parent BPMEN complex, 1, is
similar but with slightly longer Fe-N bond lengths (Table 1).
The other structurally characterized complexes are high spin,
with Fe-N bond lengths averaging 2.2 Å (Table 1). This change
in spin state is due either to the replacement of CH3CN with
the much weaker triflate ligand or to the introduction of a 6-Me
substituent on the pyridine ring. Although the presence of a
6-methyl group may be expected to make the pyridine ligand a
stronger σ-donor and favor the low-spin state, steric repulsion
between the methyl group and the metal center prevents the
formation of the shorter Fe-N bonds required for the low-spin
state and overcomes the donating effect to favor the high-spin
configuration.29,35-37
(33) Kim, J.; Dong, Y.; Larka, E.; Que, L., Jr. Inorg. Chem. 1996, 35, 2369-
2372.
(34) Elvebak, L. E., II.; Schmitt, T.; Gary, G. R. Carbohydr. Res. 1993, 246,
1-11.
9
3028 J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002

Olefin Cis-Dihydroxylation versus Epoxidation
A R T I C L E S
Table 1. Selected Bond Lengths (Å) for [FeII(BPMEN)(CH3CN)2]2+ (1), [FeII(5-Me2-BPMEN)(OTf)2] (2), [FeII(6-Me2-BPMEN)(OTf)2] (3),
[FeII(TPA)(CH3CN)2]2+ (4), [FeII(5-Me3-TPA)(CH3CN)2]2+ (5), and [FeII(6-Me3-TPA)(CH3CN)2]2+ (8a)
1a
2(A)b
2(B)b
3
4c
5d
8ae
Fe-Namine
2.055(5)
2.053(5)
2.002(4)
2.015(4)
2.214(5)
2.228(5)
2.162(4)
2.175(4)
2.210(4)
2.232(5)
2.196(4)
2.184(4)
2.2189(19)
2.2189(19)
2.2899(19)
2.2900(19)
1.99(1)
1.998(3)
2.15(1)
Fe-Npyridine
1.97(1)
1.92(1)
1.95(1)
1.92(1)
1.93(1)
1.984(3)
1.969(3)
1.985(3)
1.960(3)
1.948(3)
2.25(1)
2.18(1)
2.24(1)
2.17(1)
2.17(1)
Fe-Nacetonitrile
Fe-Otriflate
1.942(5)
1.950(5)
2.090(4)
2.127(4)
2.082(7)
2.160(4)
2.1061(15)
2.1062(15)
a Reference 28. b Two crystallographically distinct molecules in the unit cell. c Reference 29. d Reference 22. e Reference 31.
Table 2. Oxidation of Cyclooctene by H2O2 Catalyzed by Iron
complex (catalase side reaction). For example, when the 10
equiv of H2O2 were added all at once in the case of 4, the yields
of cyclooctene oxidation products decreased by 50%, from 7.4
to 3.8 turnovers (mol of product/mol of catalyst). Iodometric
titrations of the solutions at the end of the reaction indicated
that all the peroxide had been consumed. A perusal of Table 2
shows that 49-91% of the H2O2 was converted into olefin
oxidation products, depending on the catalyst used. Thus,
compared to other iron catalysts in the literature, the complexes
studied here have among the best conversion efficiencies for
the oxidation of cyclooctene with H2O2 (Table 2). Furthermore
1-3, with ligands based on an ethylenediamine backbone, are
somewhat more effective in converting H2O2 into oxidation
products than their TPA counterparts. This difference in
conversion efficiency has also been observed under similar
conditions in alkane hydroxylation reactions.22,28
Complexesa
c
cis-
conv
b
catalyst
cis-diol
oxideb diol:oxide (%)
1
2
3
4
[Fe(BPMEN)(CH3CN)2]2+
[Fe(5-Me2-BPMEN)(OTf)2]
[Fe(6-Me2-BPMEN)(OTf)2]
[Fe(TPA)(CH3CN)2]2+
0.9(2) 7.5(6) 1:8
1.3(4) 7.8(3) 1:6
6.2(1)d 1.5(1) 4:1
4.0(2) 3.4(1) 1.2:1
4.1(1) 4.0(1) 1:1
3.9(4) 2.8(1) 1.4:1
2.8(1) 2.1(2) 1.3:1
5.5(1) 1.1(2) 5:1
4.9(6) 0.7(2) 7:1
84
91
79
74
81
67
49
66
56
55
66
71
6
4a [Fe2O(TPA)2(H2O)2]4+
5
6
7
[Fe(5-Me3-TPA)(CH3CN)2]2+
[Fe(6-Me-TPA)(CH3CN)2]2+
[Fe(6-Me2-TPA)(CH3CN)2]2+
8a [Fe(6-Me3-TPA)(CH3CN)2]2+
8b [Fe(6-Me3-TPA)(O2CC6H4-4-Me)]+ 5.0(6) 0.5(1) 10:1
8c [Fe(6-Me3-TPA)(O2CC6H5)]+
6.1(4) 0.5(1) 12:1
8d [Fe(6-Me3-TPA)(O2CC6H4-3-NO2)]+ 6.7(3) 0.4(1) 17:1
9
[Fe(N4Py)(CH3CN)]2+ 26
0
0
0
0
0.6
4
0.2
8.6
10 [Fe(cyclam)(OTf)2]e 23
11 [Fe(TMP)Cl]e 39
40
2
86
12 [Fe(F20TPP)Cl]e 39
The products of cyclooctene oxidation are almost exclusively
the corresponding epoxide and cis-diol. The appearance of the
cis-diol product was unexpected in the initial stages of this work,
as cis-dihydroxylation of olefins had not previously been
observed for iron catalysts and is a reaction more commonly
associated with high-valent second and third row metal centers
with a cis-dioxometal motif. However cis-cyclooctane-1,2-diol
a See Experimental Section for reaction conditions. b Yields expressed
in moles of product per mole of iron. c Percent conversion of H2O2 into
epoxide and cis-diol products. d 0.2(1) TN of trans-diol also observed.
e Values scaled to 10 equiv of H2O2 for comparison; cyclam ) 1,4,8,11-
tetraazacyclotetradecane; TMP ) tetramesitylporphinato dianion; F20TPP
) tetrakis(pentafluorophenyl)porphinato dianion.
In general, the complexes maintain their spin states in CH3CN
solution, as determined by 1H NMR spectroscopy. Thus, 4 and
5 exhibit spectra with a compact NMR shift range of 0-12
ppm, as expected for diamagnetic complexes, while 2, 3, 6, 7,
and 8a-d exhibit spectra typical of high-spin Fe(II) complexes
with shifts that range from 120 to -50 ppm. Consistent with
1
can be unambiguously identified on the basis of its H NMR
spectrum and GC/MS profile in comparison with the authentic
compound. Control experiments demonstrated that the diol did
not come from the epoxide under the same experimental
conditions. Hydrolytic cleavage of the epoxide ring should only
give the trans-diol, and no trans-diol was found for any of the
oxidation experiments except for the small amounts observed
in the case of 3. These complexes thus represent the first
examples of iron catalysts capable of olefin cis-dihydroxylation.
Many of these catalysts are robust and maintain their ability
to catalyze olefin oxidation with more equivalents of H2O2
(Table 3). For example, for 1, the 75% conversion of H2O2 to
epoxide is maintained with the addition of 20 and 40 equiv of
H2O2, affording 15 and 29 total turnovers of epoxide products,
respectively. Thus, 1 is an excellent epoxidation catalyst, an
observation that was also reported recently by White et al.38
Similarly, for 8a and 8d, the respective 50 and 70% conversions
of H2O2 to cis-diol are maintained with the addition of 20 and
40 equiv od H2O2; these complexes are therefore excellent
catalysts for olefin cis-dihydroxylation.
1
its crystal structure, 1 exhibits a compact H NMR spectrum
expected of a diamagnetic complex, but only at -40 °C. The
shifts increase monotonically with an increase in temperature
such that a spectrum spanning 90 ppm in chemical shift is
observed at 30 °C, indicating that the iron(II) center undergoes
a transition from low spin to high spin in this temperature range.
Catalyst Efficiency. Table 2 summarizes the results of
cyclooctene oxidation experiments catalyzed by complexes 1-8.
All catalytic oxidation experiments were performed in air at
ambient temperature by delivery of 10 equiv of H2O2 via syringe
pump over a 30-min period. For 4, the yields of epoxide and
cis-diol were unaffected by the presence of air. H2O2 was
introduced by syringe pump in order to minimize dispropor-
tionation of the peroxide that may be catalyzed by the iron
Complexes 1-8 differ significantly from [Fe(N4Py)(CH3-
CN)]2+ (9), [Fe(cyclam)(OTf)2] (cyclam ) 1,4,8,11-tetraaza-
cyclotetradecane), and iron porphyrin complexes with respect
(35) Gu¨tlich, P. Struct. Bonding 1981, 44, 83-195.
(36) Gu¨tlich, P.; Hauser, A.; Spiering, H. Angew. Chem., Int. Ed. Engl. 1994,
33, 2024-2054.
(37) Constable, E. C.; Baum, G.; Bill, E.; Dyson, R.; Eldik, R. v.; Fenske, D.;
Kaderli, S.; Morris, D.; Neubrand, A.; Neuburger, M.; Smith, D. R.;
Wieghardt, K.; Zehnder, M.; Zuberbu¨hler, A. D. Chem. Eur. J. 1999, 5,
498-508.
(38) White, M. C.; Doyle, A. G.; Jacobsen, E. N. J. Am. Chem. Soc. 2001, 123,
7194-7195.
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 3029

A R T I C L E S
Chen et al.
Table 3. Oxidation of Cyclooctene by H2O2 Catalyzed by Selected
of adding 6-methyl substituents (Table 2 and Figure 2). Complex
4 affords a cis-diol/epoxide ratio of 1.2:1, a ratio that increases
only slightly with the introduction of either three 5-methyl
substituents (5) or one 6-methyl group (6). However, the addition
of a second (7) and a third 6-methyl group (8a) causes the ratio
to jump to 5 and 7, respectively. The high selectivity of 8a for
the cis-diol product is further improved when the two CH3CN
ligands are replaced with a bidentate benzoate (8b-d). In fact,
the cis-diol accounts for 95% of the overall oxidation product
when m-nitrobenzoate is used, making 8d the most selective
iron catalyst reported to date for the cis-dihydroxylation of
olefins.41
Iron Complexesa
c
catalyst
H O2 (equiv)
2
cis-diolb
oxideb
cis-diol:oxide
conv (%)
1
10
20
40
10
20
40
10
20
40
10
20
40
0.9(2)
1.4(1)
2.2(2)
4.0(2)
8.3(2)
14.9(6)
4.9(6)
10(2)
22(1)
6.7(3)
15(1)
28(1)
7.5(6)
15(1)
29(1)
1:8
84
82
78
74
70
63
56
54
59
71
79
75
1:10
1:13
1.2:1
1.4:1
1.4:1
7:1
12:1
14:1
16:1
19:1
16:1
4
3.4(2)
5.8(2)
10.3(2)
0.7(2)
0.8(1)
1.6(1)
0.4(1)
0.8(1)
1.8(1)
8a
8d
Table 4 shows that the ligand preferences for olefin epoxi-
dation versus cis-dihydroxylation observed for cyclooctene are
maintained for a variety of olefinic substrates. As noted earlier,
1 is primarily an epoxidation catalyst, while 3 and 8d strongly
favor cis-dihydroxylation, with 4 exhibiting an olefin oxidation
preference that is intermediate between these two extremes.
Thus, a small modification of the architecture of the complex,
represented in this series by the introduction of 6-methyl
substituents on the pyridine ligands, allows the exquisite tuning
of the oxidative reactivity exhibited by the iron catalyst.
Stereoselectivity. A more detailed study of the oxidation of
the isomeric cis- and trans-2-heptenes reveals a significant
ligand dependence on the stereoselectivity of the reactions (Table
5). With respect to epoxidation, all the complexes convert trans-
2-heptene stereospecifically to the corresponding trans-epoxide
with RC values of greater than 99% (RC ) 100(cis - trans)/
(cis + trans) or 100(trans - cis)/(cis + trans) for cis- or trans-
olefins, respectively). More discriminating is the epoxidation
of cis-2-heptene. The parent BPMEN and TPA complexes (1
and 4, respectively) afford the corresponding cis-epoxide with
a high degree of stereoretention (RC values of 92 and 80%,
respectively), and the results are not affected by 5-Me substitu-
tion. On the other hand, the stereoselectivity of cis-2-heptene
epoxidation decreases dramatically with the introduction of
6-methyl substituents. For the 6-Me3-TPA complexes, where
the epoxide represents less than 10% of the total oxidized
product, the two epoxide isomers are formed in comparable
amounts. In sharp contrast, the cis-dihydroxylation of the
isomeric heptenes occurs with a high degree of stereoretention
for all catalysts studied. These differences show that epoxidation
and cis-dihydroxylation, although related, must proceed by
distinct mechanisms.
a See Experimental Section for reaction conditions. b Yields expressed
in moles of product per mole of iron. c Percent conversion of H2O2 into
epoxide and cis-diol products.
Figure 2. Percent conversion of H2O2 into epoxide and cis-diol products
derived from cyclooctene oxidation catalyzed by complexes 1-12 based
on data listed in Table 2. See Chart 1 for structures of complexes 1-9; 10
) [Fe(cyclam)(OTf)2]; 11 ) [Fe(TPP)Cl]; 12 ) [Fe(F20TPP)Cl].
to their ability to catalyze olefin oxidations with H2O2 as oxidant
(Table 1). Complex 9 has a pentadentate ligand that differs from
TPA in having an additional pendant pyridine, but unlike 4, it
is a rather poor catalyst for olefin oxidation.26 On the other hand,
complexes of tetradentate macrocyclic ligands such as cyclam23
and porphyrins12,39,40 catalyze olefin epoxidation but not cis-
dihydroxylation. A structural element that distinguishes 1-8
from the other complexes is the fact that these complexes all
have two cis labile coordination sites. As in our earlier study
of stereospecific alkane hydroxylation by this family of com-
plexes,22 we will argue that this feature plays a key role in the
olefin oxidation catalysis observed.
Ligand Tuning of the cis-Diol-to-Epoxide Ratio. As shown
in Table 2 and illustrated in Figure 2, the cis-diol/epoxide ratio
depends dramatically on the nature of the iron catalyst employed.
Complexes 1 and 2 afford predominantly epoxide product, with
diol/epoxide ratios of 1:8 and 1:6, respectively. In contrast, the
corresponding 6-methyl-substituted analogue 3 converts only
15% of the H2O2 added to epoxide and 62% to diol for a diol/
epoxide ratio of 4:1. Thus, the introduction of 6-methyl groups
on the pyridine rings results in a 32-fold change in the selectivity
of the oxidation from epoxidation to cis-dihydroxylation.
The TPA series 4-8 exhibits a less dramatic change in
product ratios but nevertheless emphasizes the significant effect
The loss of stereoselectivity in cis-2-heptene epoxidation
suggests the participation of an oxidized olefin species capable
of fast epimerization to the more stable trans isomer prior to
epoxide ring formation, such as a radical cation intermediate.
While not very important for the parent BPMEN and TPA
catalysts, this channel becomes more predominant as the number
of 6-methyl substituents on the catalyst increases. The involve-
ment of a radical cation is also supported by the results of
styrene oxidation (Table 4). Indeed PhCHO is a significant
byproduct of the oxidation, which derives from O2 trapping of
(41) The somewhat better conversion efficiency of 8d compared to 8a (67 vs
49%) is not understood. However, we note that the best catalyst among
the 6-Me3-TPA complexes is 8d, the one with the least basic carboxylate
ligand. We speculate that the carboxylate is displaced as the acid in the
generation of the FeIII-OOH intermediate (vide infra). At the end of the
catalytic cycle, this acid protonates the bound diolate product to enhance
its displacement from the metal coordination sphere and initiate another
catalytic cycle.
(39) Traylor, T. G.; Tsuchiya, S.; Byun, Y.-S.; Kim, C. J. Am. Chem. Soc. 1993,
115, 2775-2781.
(40) Bartoli, J.-F.; Le Barch, K.; Palacio, M.; Battioni, P.; Mansuy, D. Chem.
Commun. 2001, 1718-1719.
9
3030 J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002

Olefin Cis-Dihydroxylation versus Epoxidation
A R T I C L E S
Table 4. Oxidation of Various Olefins by Selected Non-Heme Iron Catalystsa
1
3
4
8d
olefin
diolb
epoxideb
diolb
epoxideb
diolb
epoxideb
diolb
epoxideb
cyclooctene
cis-2-heptene
trans-2-heptene
1-octene
vinylcyclohexane
styrene
0.9(2)
0.9(1)
0.6(1)
1.3(1)
1.5(2)
2.0(1)
7.5(6)
5.5(3)
5.3(1)
4.1(1)
5.0(2)
1.1(1)
6.4(1)c
6.4(1)
4.6(2)
2.6(2)
5.3(1)
5.0(2)
4.8(2)
1.5(1)
2.0(1)
1.5(3)
0.7(1)
2.2(2)
1.6(1)
2.2(4)c
4.0(2)
3.0(3)
3.3(3)
5.3(3)
5.3(3)
2.4(1)
3.4(1)
1.9(1)
1.5(1)
2.2(1)
2.7(1)
1.7(1)
1.6(1)c
6.7(3)
4.5(2)
5.3(1)
5.5(2)
5.6(2)
5.2(1)
0.4(1)
0.4(1)
0.2(1)
0.3(2)
0.5(1)
0.6(1)
1.9(1)c
a See Experimental Section for reaction conditions. b Yields expressed in moles of product per mole of iron. c PhCHO byproduct observed due to autoxidation.
Table 5. Oxidation of cis- and trans-2-Heptenesa
cis−2-heptene
trans−2-heptene
catalyst
ligand
BPMEN
5-Me2-BPMEN
6-Me2-BPMEN
TPA
5-Me3-TPA
6-Me-TPA
6-Me2-TPA
6-Me3-TPA
6-Me3-TPA,
O2CC6H4-3-NO2
diolb
RCc
epoxideb
RCc
diolb
RCc
epoxideb
RCc
1
2
3
4
5
6
7
8a
8d
0.9(1)
0.9(1)
4.6(2)
3.0(3)
4.4(1)
4.2(1)
3.9(5)
4.1(4)
4.5(2)
79
79
78
96
94
95
84
93
97
5.5(3)
5.2(1)
2.0(1)
1.9(1)
2.3(1)
1.6(1)
1.3(1)
0.4(1)
0.4(1)
92
90
60
80
83
69
40
35
0
0.6(1)
1.1(1)
2.6(2)
3.3(3)
5.2(1)
4.1(1)
4.8(1)
3.8(6)
5.3(1)
88
92
95
96
>99
94
95
96
>99
5.3(1)
5.1(1)
1.5(3)
1.5(1)
1.6(1)
1.1(1)
0.6(1)
0.3(1)
0.2(1)
>99
>99
>99
>99
>99
>99
>99
>99
>99
a See Experimental Section for reaction conditions. b Yields expressed in moles of product per mole of iron. c RC, the percentage of retention of configuration
in the dihydroxylation and in the epoxidation of 2-heptenes, expressed as the ratio of the diols or the epoxides: 100(cis - trans)/(cis + trans) for cis-olefin
oxidation and 100(trans - cis)/(cis + trans) for trans-olefin oxidation.
This lag in product formation is rationalized by the prior
oxidation of the FeII complex to the catalytically active FeIII
species, which would consume only 0.5 equiv of H2O2. Once
formed, the active FeIII catalyst reacts with more H2O2 to
generate the olefin oxidation products. Consistent with this
required preoxidation for the iron(II) catalyst, no lag phase is
observed in product formation when [FeIII2O(TPA)2(H2O)2]-
(ClO4)4 (4a) is used in place of 4. Complex 4a exhibits a
catalytic efficiency comparable to that of 4 and, with 10 equiv
of H2O2, affords 4.1(1) TN cis-diol and 4.0(1) TN epoxide under
the same conditions (Table 2). Thus, both 4 and 4a act as
precursors to the active oxidant for olefin oxidation. As shown
in electrospray ionization mass spectrometry (ESI-MS) studies
Figure 3. Yields of epoxide and cis-diol products as a function of the
amount of H2O2 added for the oxidation of cyclooctene by 4.
discussed below, both generate an intermediate FeIII-OH
species,44 which in turn reacts with more H2O2 to form the active
FeIII-OOH oxidant (Scheme 1).45
a radical cation species,42,43 and the amount of PhCHO formed
decreases significantly when styrene oxidation is carried out
We earlier proposed that a low-spin FeIII-OOH intermediate
is involved in catalytic alkane hydroxylation by the TPA and
BPMEN families of catalysts.22 Such species have recently been
trapped and spectroscopically characterized in a number of
related non-heme iron systems,46 although pentadentate ligands
were necessary in many cases in order to stabilize the
intermediates sufficiently for observation. The low-spin FeIII-
under Ar.
Formation of an FeIII-OOH Intermediate. More detailed
studies of olefin oxidation catalyzed by 4 provide insights into
the nature of the metal species involved. Figure 3 shows the
oxidation of cyclooctene monitored as a function of added H2O2.
For 4, the amounts of cis-diol and oxide products increase
(44) The importance of an intermediate FeIII-OH species raises the question of
linearly with increasing amounts of H2O2, and the diol/epoxide
why iron(II) complexes are generally used as starting points for catalysis
ratio is maintained. These results support the notion that cis-
diol and epoxide are produced by related but independent
pathways, the partitioning between which is controlled by the
nature of the tetradentate ligand. However, no organic products
are detected upon the addition of the first 0.5 equiv of H2O2, as
also noted in our study of alkane hydroxylation by 4/H2O2.22
in this study instead of corresponding iron(III) complexes. Both 4 and 4a
are effective catalysts because they have ready access to the FeIII-OH
species that allows efficient formation of the key FeIII-OOH intermediate.
However, iron(III) complexes such as 4a with an easily displaced water
ligand are not readily available. Often, the high Lewis acidity of the iron(III)
center leads to the formation of stable complexes with less labile ligands
that inhibit the ligand exchange needed to rapidly form the FeIII-OOH
intermediate and catalyze H2O2 disproportionation instead. Such ligand
differences have also been shown to affect the course of alkane hydroxyl-
ation by non-heme iron complexes.22,45
(42) Heimbrook, D. C.; Carr, S. A.; Mentzer, M. A.; Long, E. C.; Hecht, S. M.
Inorg. Chem. 1987, 26, 3835-3836.
(43) Groves, J. T.; Gross, Z.; Stern, M. K. Inorg. Chem. 1994, 33, 5065-5072.
(45) Mekmouche, Y.; Me´nage, S.; Toia-Duboc, C.; Fontecave, M.; Galey, J.-
B.; Lebrun, C.; Pecaut, J. Angew. Chem., Int. Ed. 2001, 40, 949-952.
(46) Girerd, J.-J.; Banse, F.; Simaan, A. J. Struct. Bonding 2000, 97, 143-177.
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 3031

A R T I C L E S
Chen et al.
Scheme 1. Formation of an FeIII -OOH Intermediate
OOH intermediate for 4, albeit less stable than its pentadentate
ligand counterparts, has been observed at -40 °C and found to
have a ν(O-O) near 800 cm-1 47
, a value comparable to those
observed for other low-spin FeIII-OOH intermediates but
significantly lower than those observed for high-spin FeIII-
peroxo complexes.29,46,48,49 Thus, the low-spin iron(III) center
weakens the O-O bond and activates it for cleavage.25,47 Since
the conditions for alkane hydroxylation and olefin oxidation are
identical save for the nature of the substrates, it seems reasonable
to postulate the involvement of this intermediate in olefin
oxidation.
Figure 4. Electrospray ionization mass spectra for the reaction of 4 with
ESI-MS has proven to be an excellent tool for establishing
the molecular compositions of transient species in iron-catalyzed
oxidations.33 Its utility is illustrated for the reactions of the TPA
catalyst. When the effluent derived from combining precooled
solutions of 4 or 4a and H2O2 in a mixing tee is introduced
directly into the mass spectrometer, two signals are observed
in the positive ion mass spectrum at m/z 462 and 478 (Figure
4A). The m/z 462 signal has a mass and isotope distribution
pattern that corresponds to {[FeIII(TPA)(OH)](ClO4)}+, which
is the precursor to the species that gives rise to the prom-
inent signal at m/z 478. The latter has a mass value and an
isotope distribution pattern consistent with its formulation as
{[FeIII(TPA)(OOH)](ClO4)}+. MS/MS experiments show that
the fragmentation of the m/z 478 ion gives rise to a prominent
daughter ion at m/z 445 due to the loss of OOH forming
{[FeII(TPA)](ClO4)}+ and a much weaker daughter ion at m/z
461, corresponding to the loss of OH forming {[FeIV(TPA)-
(O)](ClO4)}+. Last there is also a molecular ion signal in the
negative ion mass spectrum at m/z 676 corresponding to the
ion {[FeIII(TPA)(OOH)](ClO4)3}- (data not shown).
H2O2 in CH3CN at -40 °C in the absence (A) and in the presence of styrene
(B).
Chart 2. Proposed Structures Corresponding to the ESI-MS Ions
of the Species Generated in the Reaction of 4 and H2O2 in the
Presence of Styrene
at m/z 482, 500, 516, and 582, in addition to the original features
at m/z 462 and 478. The most intense of these signals at m/z
582 has a mass and an isotope distribution pattern that
correspond to a combination of the FeIII-OOH intermediate ion
and styrene (m/z 478 + 104). This combination can be construed
as the {[(TPA)FeIII-OOH]‚styrene‚ClO4}+ adduct, {[(TPA)-
FeIII-OH]‚styrene oxide‚ClO4}+, or {[(TPA)FeIII-styrene di-
olate]ClO4}+. The MS/MS experiment on this ion shows that
its fragmentation generates only one major daughter ion at m/z
482 due to the loss of HClO4 (Figure S1c, Supporting Informa-
tion). The fact that there is no daughter ion due to the loss of
styrene alone suggests that the styrene and OOH components
have already combined chemically. Since there is also no
daughter ion due to the loss of styrene oxide alone, it is
reasonable then to formulate the m/z 582 ion as the complex of
the diol product (structure D in Chart 2). The stability of this
ion is not surprising, given the expected bidentate coordination
of the diol to the iron center and the oxophilicity of iron(III).
The assignment of the m/z 482 ion to structure A in Chart 2
can be rationalized on the same basis. The ions at m/z 500 and
516 both fragment by loss of diol and respective losses of H2O
and OOH (Figure S1a and b, Supporting Information) and are
thus formulated as structures B and C in Chart 2, respectively.
Notably, the molecular ion at m/z 478 of the Fe-OOH species
is not affected by the presence of H218O. This result shows that
the peroxo oxygen atoms in this ion do not exchange with water
and that the O-O bond remains intact at this stage of the
characterization, as expected. The mass spectral data un-
fortunately do not provide any insight as to the identity of the
sixth ligand, which would very likely be required to achieve
the low-spin iron(III) state indicated by the EPR spectrum of
the intermediate.25 Considering the readily dissociable nature
of this ligand, we presume that the sixth coordination site
is occupied by a labile solvent molecule such as CH3CN or
H2O.
New species are observed in the ESI-MS spectra of the FeIII-
OOH intermediate generated from 4 and H2O2 in the presence
of styrene. Figure 4B shows the appearance of four new signals
(47) Ho, R. Y. N.; Roelfes, G.; Feringa, B. L.; Que, L., Jr. J. Am. Chem. Soc.
1999, 121, 264-265.
(48) Lehnert, N.; Ho, R. Y. N.; Que, L., Jr.; Solomon, E. I. J. Am. Chem. Soc.
2001, 123, 8271-8290.
(49) Lehnert, N.; Ho, R. Y. N.; Que, L., Jr.; Solomon, E. I. J. Am. Chem. Soc.
2001, 123, 12802-12816.
9
3032 J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002

Olefin Cis-Dihydroxylation versus Epoxidation
A R T I C L E S
Table 6. Isotope Labeling Results from the Oxidation of
Cyclooctene by H2O2 Catalyzed by Iron Complexesa
cis-cyclooctane-1,2-diol
H 18 H 18O2
cyclooctene oxide-18O%
H 18 H 18O2 18O2b 16O18O 18O18O 16O18O 18O18O
O
2
2
ligand
TPA
5-Me3-TPA 2(1) 93(6)
6-Me-TPA 7(1) 71(9)
6-Me2-TPA 3(1) 61(8)
O
2
2
4
5
6
7
9(1) 90(8)
1
5
22
36
43
86(5) 1(1) 97(3)
83(9) 1(1) 97(3)
80(1) 2(1) 94(4)
3(1)
2(2)
6(1)
3(1) 1(1)
1(1)
7(1) 93(3)
4(1) 96(1)
8a 6-Me3-TPA 3(1) 54(2)
0
a See Experimental Section for reaction conditions. b Calculated consider-
ing mass balance of the oxygen atom.
Thus, the product diol coordinates to the metal center in the
course of a turnover and is presumably displaced by H2O2 to
initiate the next round of catalysis.
Figure 5. Linear correlation between the RC value for cis-2-heptene
epoxidation and the amount of O2 incorporation into cyclooctene oxide for
the TPA series of catalysts (r2 ) 0.94). Data obtained from Tables 5 and
6.
18O-Labeling Experiments. The subsequent fate of the bound
hydroperoxide has been probed by 18O-labeling experiments for
the TPA series of catalysts, and these studies provide key
insights into the nature of the oxidants in the mechanisms for
olefin oxidation (Table 6). As for the previously reported alkane
hydroxylation experiments,22,28 oxidations were carried out in
the presence of added water to determine the source(s) of the
oxygen atom(s) incorporated into the epoxide and cis-diol
products. The presence of the added water does not significantly
affect the product yields observed for the various catalysts in
this study.
Quite distinct from the isotope labeling results for epoxidation
are those for cis-diol formation (Table 6), emphasizing the
notion that cis-dihydroxylation operates by a parallel but
independent mechanism. Both oxygen atoms of the cis-diol
obtained from 7- and 8a-catalyzed oxidations derive essentially
from H2O2, with no incorporation of isotope label from the
addition of 1000 equiv of H218O. As previously reported,26
a
labeling experiment with a 60:40 mixture of H216O2 and H218O2
in the oxidation of cyclooctene by 8a afforded a 60:40 ratio of
unlabeled and doubly labeled cis-diol and a negligible amount
of the 16O18O isotopomer. This experiment demonstrates that
the diol oxygens in 8a-catalyzed cis-dihydroxylation derive from
one molecule of H2O2.
The epoxidation of cyclooctene with H218O2 in the presence
of 1000 equiv of H216O catalyzed by 4 affords the corresponding
epoxide with 90% incorporation of the 18O label. When H216O2
is used in the presence of 1000 equiv of H218O, the comple-
mentary result is obtained with 9% of the epoxide 18O-labeled.
Thus, all the oxygen present in the epoxide derives either from
H2O2 or from the H2O present in the solution, with essentially
no incorporation of O2 from air into the epoxide due to
autoxidation. The same results are obtained for 5. Thus, the
incorporation of 18O from H218O into the epoxide products
requires the participation of an oxidant capable of solvent water
exchange in the epoxidation mechanism.
The extent of 18O incorporation into the epoxide product from
various oxygen sources is affected by the number of 6-Me
substituents on the pyridine ligands (Table 6). For 6 with one
6-Me substituent, the epoxide product derives its oxygen atom
71% from H2O2 and 7% from H2O, with the remaining 22%
presumably from air. No 18O from H218O is incorporated into
epoxide products of 7 and 8a, complexes with two or three 6-Me
substituents, and the percentage incorporation from H2O2 also
decreases to ∼60%, so the presumed contribution from air
increases to ∼40% (Table 6). In fact, the increase in percent O
from O2 in cyclooctene epoxidation correlates well with the
decrease in RC value in the epoxidation of cis-2-heptene, as
shown in Figure 5. This linear correlation suggests an increased
involvement of a radical cation intermediate leading to epimer-
ization and autoxidation in the mechanism of olefin epoxidation
as the number of 6-methyl substituents increases. This trend
parallels the observations we reported previously for alkane
hydroxylation by this family of catalysts, where increasing the
number of 6-methyl substituents causes a decrease in the
stereospecificity in the hydroxylation of cis-1,2-dimethylcyclo-
hexane and an increase in the percent O derived from O2 into
the alcohol product.22
Dramatically different isotopic labeling results are obtained
for cis-dihydroxylation by 4, 5, and 6, catalysts with either no
6-Me substituent or only one 6-Me substituent (Table 6). In
these cases, only one of the oxygen atoms of the cis-diol
becomes isotopically labeled either from experiments carried
out with H216O2 in the presence of 1000 equiv of H218O or from
the complementary reactions with H218O2 and 1000 equiv of
H216O. Thus, one oxygen atom of the cis-diol derives from H2O2,
while the other derives from H2O. These observations require
a mechanism involving an oxidizing species that can fully
equilibrate with solvent water.
18O incorporation from H218O into the epoxide and cis-diol
products of 4 display saturation behavior (Figure 6), as shown
previously for 5-catalyzed hydroxylation of cyclohexane with
H2O2.22 This behavior indicates a preequilibrium binding of
water to the metal center during turnover. Double-reciprocal
plots of the data (Figure 6 insets) in fact show straight lines
that give estimated association constants of about 20 and 30
M-1, respectively, values that are comparable to the 16 M-1
value obtained in the cyclohexane hydroxylation study.22 This
congruence in the isotope labeling results suggests that the same
water binding equilibrium affects the oxidizing species involved
in Fe(TPA)-catalyzed alkane hydroxylation, olefin epoxidation,
and olefin cis-dihydroxylation.
Mechanistic Considerations. Scheme 2 shows the proposed
mechanism for olefin epoxidation and cis-dihydroxylation
catalyzed by the family of non-heme iron complexes discussed
here. These olefin oxidation studies build on the foundation
provided by our earlier work on alkane hydroxylation by this
same family of catalysts22 and further validate key features of
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 3033

A R T I C L E S
Chen et al.
as precedented by the [FeIII(6-Mex-TPA)(OOtBu)]2+ series (x
) 0-3).29
Category A complexes (1, 2, 4, 5, 6) afford low-spin FeIII-
OOH intermediates, whose O-O bonds are weakened by the
1
presence of the S ) /2 iron center.47 The fact that 9, with the
closely related but pentadentate N4Py ligand, also forms a
similar low-spin FeIII-OOH intermediate47 but is not a catalyst
for olefin oxidation26 suggests that the presence of the low-
spin iron(III) center is insufficient to account for the reactivity
observed. The adjacent second labile site on the catalyst must
thus serve a critical role in activating the bound peroxide.
In the right branch of Scheme 2, we propose that peroxide
activation occurs by preequilibrium water binding to the adjacent
labile site of the FeIII-OOH intermediate. A five-membered
ring species is formed that promotes O-O bond heterolysis to
generate the cis-H18OsFeVdO oxidant. The observation of 18
O
incorporation from H218O into the epoxide and cis-diol products
of these reactions strongly supports this mechanistic hypothesis,
as it requires another intermediate that is capable of solvent
water exchange. Partial labeling of the epoxide product with
18O from H218O would occur by oxo-hydroxo tautomerism of
the cis-H18OsFeVdO oxidant, as precedented in heme sys-
tems,50,51 which would introduce 18O into the terminal iron-
oxo unit before oxygen atom transfer to the olefinic substrate.
Cis-dihydroxylation, on the other hand, entails the transfer of
two oxygen atoms from oxidant to substrate. Thus, the observa-
tion that one oxygen atom of the cis-diol product derives from
H2O while the other comes from H2O2 strongly supports the
involvement of the proposed cis-H18OsFeVdO species; it in
fact represents the strongest evidence that the cis-H18OsFeVdO
species is the oxidant for category A catalysts in the presence
of excess water.
Figure 6. Fraction of 18O-labeled cis-diol (top) and epoxide (bottom)
obtained in the oxidation of cyclooctene catalyzed by 4/H2O2 as a function
of the concentration of H218O ([H218O]). Insets: the double-reciprocal plots.
Scheme 2. Proposed Mechanism for Catalysis
The mechanism shown on the right branch of Scheme 2 may
apply as well for category A catalysts in the absence of excess
water, but we cannot exclude the mechanism shown on the left
branch. In this case, the FeIII-η1-OOH intermediate would be
activated by isomerization to an η2-OOH species, which may
either react directly with the olefin or convert to the high-valent
species prior to attack of the olefin. In such a pathway, the highly
stereoselective epoxide and cis-diol products derive their oxygen
atoms exclusively from H2O2. The less than 10% incorporation
of 18O from H218O into the epoxide may argue for the
involvement of both branches of Scheme 2 for category A
catalysts, but the high level of 18O incorporation into the cis-
diol shows that the right branch of Scheme 2 dominates for
cis-dihydroxylation.
Category B catalysts (3, 7, 8a-d), on the other hand, very
likely give rise to high-spin FeIII-OOH species due to the steric
effects of the 6-methyl substituents.29 The lack of 18O incorpora-
tion from H218O into epoxide and cis-diol products excludes
the right branch of Scheme 2 as the mechanism for peroxide
activation. Consistent with the alternative mechanism on the
left branch, both diol oxygens of the cis-diol products derive
almost exclusively from H2O2. At present, it is not clear why
category B catalysts should strongly favor olefin cis-dihydroxyl-
ation, but our data unequivocally establish this preference.
Furthermore, unlike for category A catalysts, category B
catalysts yield minor epoxide products with significant loss of
a mechanism that posits the participation of an FeIIIsOOH
intermediate and the corresponding FeVdO derivative. The
proposed mechanism takes into account the following observa-
tions:
1. Only complexes with two cis labile sites are effective
catalysts for olefin epoxidation and cis-dihydroxylation.
2. Category A complexes, with ligands having either no
6-methyl substituent (1, 2, 4, 5) or only one 6-methyl substituent
(6), catalyze highly stereoselective olefin epoxidation and olefin
cis-dihydroxylation. Epoxide and cis-diol products incorporate
18O from H218O via an [H218O]-dependent preequilibrium.
3. Category B complexes, with ligands having more than one
6-methyl substituent (3, 7, 8a-d), catalyze olefin cis-dihy-
droxylation predominantly, affording cis-diols with both oxygen
atoms derived from H2O2.
As proposed in our alkane hydroxylation studies,22 what
distinguishes category A catalysts from those in category B is
the spin state of the FeIII-OOH intermediate, which can be tuned
by the number of 6-methyl substituents on the tripodal ligand,
(50) Bernadou, J.; Meunier, B. Chem. Commun. 1998, 2167-2173.
(51) Meunier, B.; Bernadou, J. Struct. Bonding 2000, 97, 1-35.
9
3034 J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002

Olefin Cis-Dihydroxylation versus Epoxidation
A R T I C L E S
stereochemistry and incorporation of oxygen from O2. Instead
of a mechanism involving direct oxygen atom transfer, these
observations strongly suggest a stepwise epoxidation mechanism
involving the initial oxidation of olefin to a radical cation
intermediate. Thus, the reactivity of high-spin FeIII-OOH
species differs significantly from that of their low-spin coun-
terparts.
bleomycin; their metal centers have only one available site with
which to form low-spin FeIII-OOH intermediates.16,52,53 Without
an adjacent labile site, such intermediates can only give rise to
olefin epoxidation.
On the other hand, cis-dihydroxylation of C-C double bonds
is catalyzed only by bacterial non-heme iron Rieske dioxy-
genases.5 These enzymes have a mononuclear iron site that is
coordinated to two histidine imidazoles and a bidentate car-
boxylate from aspartate19 with two cis sites available for
exogenous ligand binding. Our model studies suggest that olefin
cis-dihydroxylation is favored by a high-spin iron center with
two cis labile sites. Given the weak crystal field expected from
the 2-His-1-carboxylate facial triad,54 it seems likely that the
proposed but as yet unobserved iron(III)-hydroperoxo inter-
mediate in the catalytic cycle of these enzymes will have a high-
spin state. Furthermore, there are two cis labile sites available
for activating the bound hydroperoxide to the putative cis-HOs
FeVdO species. Evidence is unavailable for the participation
of such a high-valent species in cis-dihydroxylation by the
enzymes, since 18O-labeling studies show that both atoms of
O2 are incorporated into the cis-diol product. However, 18O from
H218O is found in the alcohol product of the highly enantio-
selective hydroxylation of indane by toluene and naphthalene
dioxygenases,55,56 so an FeVdO species can form in the Rieske
dioxygenase active site and may thus be involved in cis-
dihydroxylation as well. The insights we have gained in the
foregoing study illustrate the utility of biomimetic investigations
to shed light on biological mechanisms for oxygen activation.
Scheme 2 represents our best working model to date for our
family of catalysts. There is strong evidence for FeIII-OOH
intermediates as well as the participation of a cis-HOsFeVdO
oxidant in category A reactions, but not for the intermediates
shown in brackets; these various reactive species, however
plausible, are at present all speculations based on detailed
product analysis and 18O labeling. Kinetic studies may also
contribute to our understanding of the mechanism. However,
the syringe pump conditions used in this study are not conducive
for the demonstration of a particular rate law, since H2O2
concentrations are maintained at a level substoichiometric
relative to catalyst in order to minimize deleterious side reactions
and maximize H2O2 conversion into the desired products.
Summary and Perspectives. In this paper, we have char-
acterized a family of non-heme iron catalysts with two cis labile
sites, which are capable of stereospecific olefin epoxidation and
cis-dihydroxylation using H2O2 as oxidant. Ligand tuning of
the spin state of the FeIII-OOH species formed in the course
of catalysis results in the range of reactivities observed. A
dramatic illustration of this effect can be seen in the BPMEN
series, where a 32-fold change in the cis-diol/epoxide ratio is
observed on going from 1, which catalyzes epoxidation almost
exclusively (89% of the total products), to 3, which favors cis-
dihydroxylation (80% of the total products).
The complexes in this series are in fact the first examples of
iron catalysts capable of olefin cis-dihydroxylation, thus provid-
ing a unique opportunity to gain insight into its novel mecha-
nism. Category A catalysts give rise to low-spin FeIII-OOH
intermediates and afford cis-diol products that incorporate one
oxygen atom from H2O2 and the other from H2O. Taken together
with the observation that alkane hydroxylation and olefin
epoxidation products also incorporate 18O from H218O, these
results provide compelling evidence for a cis-H18OsFeVdO
species as the common oxidant in these reactions (Scheme 2
right). Category B catalysts, on the other hand, give rise to high-
spin FeIII-OOH intermediates and afford cis-diol products with
both oxygen atoms from H2O2. Activation of the O-O bond is
proposed to occur via an FeIII-η2-OOH species that may
isomerize into a cis-HOsFeVdO species in the course of
catalysis (Scheme 2 left).
Acknowledgment. This work was supported by the National
Institutes of Health (GM-33162). K.C. is grateful to a thesis
fellowship from the Department of Chemistry of the University
of Minnesota. M.C. acknowledges the Fundacio La Caixa for a
fellowship. The authors thank Dr. Victor G. Young, Jr., and
Dr. Maren Pink (X-ray Crystallographic Laboratory, University
of Minnesota) for solving the crystal structures of complexes 2
and 3.
Supporting Information Available: Table S1 of the crystal-
lographic data of 2 and 3, Figure S1 of MS/MS spectra of ions
obtained from the precooled mixture of 4 or 4a and H2O2 in
the presence of styrene, and X-ray crystallographic data for 2
and 3 (CIF). This material is available free of charge via the
Internet at http://pubs.acs.org.
JA0120025
(52) Davydov, R.; Makris, T. M.; Kofman, V.; Werst, D. E.; Sligar, S. G.;
Hoffman, B. M. J. Am. Chem. Soc. 2001, 123, 1403-1415.
(53) Burger, R. M.; Kent, T. A.; Horwitz, S. B.; Mu¨nck, E.; Peisach, J. J. Biol.
Chem. 1983, 258, 1559-1564.
(54) Hegg, E. L.; Que, L., Jr. Eur. J. Biochem. 1997, 250, 625-629.
(55) Wackett, L. P.; Kwart, L. D.; Gibson, D. T. Biochemistry 1988, 27, 1360-
1367.
Our studies provide a basis with which to rationalize Nature’s
choice of ligand environments for carrying out iron-catalyzed
alkene oxidations. Olefin epoxidation is carried out by cyto-
chrome P450 and the non-heme iron-requiring antitumor drug
(56) Gibson, D. T.; Resnick, S. M.; Lee, K.; Brand, J. M.; Torok, D. S.; Wackett,
L. P.; Schocken, M. J.; Haigler, B. E. J. Bacteriol. 1995, 177, 2615-2621.
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 3035