Design of and mechanistic studies on a biomimetic iron-imidazole catalyst system for epoxidation of olefins with hydrogen peroxide

Article abstract of DOI:10.1002/chem.200802731

Novel iron catalysts, both defined and in situ generated, for the epoxidation of aromatic and aliphatic olefins with hydrogen peroxide as terminal oxidant are described. Our catalyst approach is based on bio-inspired 1-aryl-substituted imidazoles in combination with cheap and abundant iron trichloride hexahydrate. We show that the free 2-position of the imidazole ligand motif plays a key role for catalytic activity, as substitution leads to a dramatic depletion of yield and conversion. X-ray studies, UV/Vis titrations, and NMR studies were carried out to clarify the mechanism.

Full text of DOI:10.1002/chem.200802731

FULL PAPER
DOI: 10.1002/chem.200802731
Design of and Mechanistic Studies on a Biomimetic Iron–Imidazole Catalyst
System for Epoxidation of Olefins with Hydrogen Peroxide
Kristin Schrçder,[a] Stephan Enthaler,[a] Bianca Bitterlich,[a] Thomas Schulz,[a]
Anke Spannenberg,[a] Man Kin Tse,[a, b] Kathrin Junge,[a] and Matthias Beller*[a, b]
Abstract: Novel iron catalysts, both defined and in situ generated, for the epoxida-
tion of aromatic and aliphatic olefins with hydrogen peroxide as terminal oxidant
are described. Our catalyst approach is based on bio-inspired 1-aryl-substituted
Keywords: epoxidation · homoge-
imidazoles in combination with cheap and abundant iron trichloride hexahydrate.
neous catalysis · iron · N ligands ·
We show that the free 2-position of the imidazole ligand motif plays a key role for
reaction mechanisms
catalytic activity, as substitution leads to a dramatic depletion of yield and conver-
sion. X-ray studies, UV/Vis titrations, and NMR studies were carried out to clarify
the mechanism.
Introduction
larly defined transition-metal complexes provide beneficial
models for new types of oxidation catalysts. Known homo-
geneous metal complexes that make use of hydrogen perox-
ide are mainly based on ruthenium,[4] rhenium,[5] manga-
nese,[6] and more recently iron.[7] Notably, until today the de-
velopment of a generally applicable, active and selective cat-
alyst system which can epoxidize both aromatic and aliphat-
ic olefins is still a challenging goal.
Using iron as metal center of the catalyst implies many
advantages compared to precious metals. It is the second
most abundant metal in the earth crust (4.7 wt%) and rela-
tively non-toxic. A variety of iron salts and iron complexes
are currently commercially available on a large scale. It is
noteworthy that iron is involved in manifold biological sys-
tems as fundamental key element, for instance, in metallo-
proteins like methane monooxygenases for the metabolic
aerobic pathway of methane to methanol.[8] Figure 1 shows
an example of a non-heme FeII/a-ketoglutarate (aKG) de-
pendent halogenase,[9a] in which two histidine ligands and
one chloride ion are coordinated to iron in the central core.
The remaining iron coordination sites are occupied by the
co-factor aKG and a water molecule.
Catalytic epoxidation of olefins provide oxiranes, which are
of central importance in various fields of chemistry ranging
from pharmaceutical intermediates to monomers for bulk
polymers. From an industrial point of view aliphatic epox-
ides are of special interest, namely, 1,2-propylene oxide and
ethylene oxide, which are annually produced on a million-
ton scale.[1] These large scale productions are based on het-
erogeneous catalysts, for example, with Ti-substituted mo-
lecular sieve silicalite (TS-1) for the epoxidation of propyl-
ene[2] or silver catalysts supported on Al2O3 for the produc-
tion of ethylene oxide.[3] State-of-the-art heterogeneous cat-
alysts make use of benign oxidants such as hydrogen perox-
ide or molecular oxygen, but require relatively harsh
reaction conditions. Hence, there is a continuing interest in
more active and selective catalysts. In this respect molecu-
[a] K. Schrçder, Dr. S. Enthaler, Dr. B. Bitterlich, T. Schulz,
Dr. A. Spannenberg, Dr. M. K. Tse, Dr. K. Junge, Prof. Dr. M. Beller
Leibniz-Institut fꢀr Katalyse e.V. an der Universitꢁt Rostock
Albert-Einstein-Strasse 29a, 18059 Rostock (Germany)
Fax : (+49)381-1281-5000
Based on these structurally well-characterized enzymes,
many model complexes have been prepared to explore
mechanistic issues and their use in oxidation catalysis
(Scheme 1).[10] Unfortunately, each of these systems has limi-
tations such as tedious complex preparation, limited sub-
strate scope, low selectivity and the use of “non-green” or
expensive oxidants such as iodoso compounds and
peracids.[12]
E-mail: matthias.beller@catalysis.de
[b] Dr. M. K. Tse, Prof. Dr. M. Beller
Center for Life Science Automation (CELISCA)
University of Rostock
Friedrich-Barnewitz-Strasse 8, 18119 Rostock (Germany)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200802731.
Chem. Eur. J. 2009, 15, 5471 – 5481
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5471

of ligand (5-Cl-1-MeIm) with a catalyst loading of 5 mol%
FeCl3·6H2O in tert-amyl alcohol.[14] As model substrate for
catalytic testing trans-stilbene was chosen owing to its stabil-
ity and ease of handling. From the results of an optimization
process, 2–3 equivalents of 30% hydrogen peroxide are used
as oxygen source under ambient conditions (air and RT). In-
itially, various acidic and basic co-ligands were applied in
the model reaction. Other Fe-dependant reactions (e.g., oxi-
dation of benzyl alcohol to benzaldehyde) are significantly
influenced by such additives.[16] Earlier, we also discovered
the strong dependence of selectivity on pH in osmium-cata-
lyzed dihydroxylation of olefins with oxygen as terminal oxi-
dant.[17] Selected results of our investigation are shown in
Table 1. Apparently, the acidity of the reaction system plays
an important role in selectivity and conversion. Small
amounts (<5 mol%) of weak acids like acetic acid, benzoic
acid or iminodiacetic acid slightly increased or at least main-
tained conversion and selectivity (Table 1, entries 2, 8–12).
Figure 1. Non-heme iron halogenase SyrB2.[9b] C gray, N blue, O red, Cl
green, Fe brown.
Addition of
a higher concentration of benzoic acid
(10 mol%) led to a decrease of nearly 50% in yield and
35% in selectivity. Similarly “strong” acids such as p-tolu-
AHCTUNGTERGeNNUN nesulfonic acid, phosphoric
acid and nitric acid as addi-
tives (Table 1, entries 13–15)
gave only poor conversion and
low selectivity.
Amines such as bipyridine
and pyrrolidine diminished sig-
nificantly yield and conversion
(Table 1, entries 3 and 4).
Probably, the higher concen-
Scheme 1. Bio-inspired ligands of iron complexes for epoxidation of olefins.[11]
tration of N-donor ligands re-
On the basis of our background in epoxidation chemistry
of aromatic olefins using hydrogen peroxide as an oxygen
source in the presence of a catalyst generated in situ from
FeCl3·6H2O, pyridine-2,6-dicarboxylic acid (H2pydic) and an
organic base like pyrrolidine,[13] we recently developed a
simplified two-component catalyst system consisting of 5-
chloro-1-methylimidazole (5-Cl-1-MeIm) and FeCl3·6H2O,
which is able to catalyze the epoxidation of various sub-
strates, such as substituted styrenes and stilbenes.[14] Further-
more, we demonstrated epoxidation of several aliphatic ole-
fins with an improved three-component system consisting of
FeCl3·6H2O, H2pydic and different N-benzylamines.[15]
Here we report a detailed study on improving our bio-
mimetic protocol involving a two-component catalyst and
demonstrate that iron-based catalysts can epoxidize both ali-
phatic and aromatic olefins with hydrogen peroxide in good
yield and with high selectivity.
duces the epoxide yield by oc-
cupying the coordination sites of the iron center or reduces
the Lewis acidity of the active catalyst species. Surprisingly,
inorganic bases like NaOH and NaHCO3 decreased the
yield and selectivity only slightly (around 10%; Table 1, en-
tries 5 and 6). Salt additives did not show any clear trend
and often slightly increased yield and conversion (Table 1,
entries 17–21). In conclusion, the results indicate a correla-
tion between acidity and catalyst activity.
Because combinations of peracetic acid and various iron
complexes are known to be efficient epoxidation cata-
lysts,[18,19] we performed several experiments to exclude for-
mation of the corresponding peracids by an iron-catalyzed
process (Scheme 2). However, experiments with peracetic
acid instead of hydrogen peroxide as oxidant[20] do not sup-
port in situ generation of peracid. Here, conversion and
yield diminished significantly to only about 5%.
Next, we examined the influence of the structure of the
imidazole ligand more closely. Unfortunately, our prior two-
component system consisting of FeCl3·6H2O and 5-Cl-1-
MeIm was able to epoxidize only aromatic olefins (e.g. sty-
Results and Discussion
Catalysis experiments: Starting with our previously de-
scribed in situ 5-Cl-1-MeIm/iron system, the aim was to in-
crease conversion and yield by introducing additives. The
original in situ two-component system consisted of 10 mol%
ACHTUNGTRENrNUNG enes, stilbenes) to the corresponding oxiranes in moderate
to excellent yields.[14]
To get a first impression of the reactivity of the shown li-
gands and to ensure comparability to the former work trans-
5472
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5471 – 5481

Biomimetic Iron–Imidazole Catalyst System
FULL PAPER
Table 1. Influence of additives on the epoxidation of trans-stilbene.[a]
Entry Additive
G
[%][b] [%][b]
[%][c]
Scheme 2. Possible formation of peracetic acid in the reaction system.
1
2
–
–
5
83
89
80
87
97
98
3
5
51
44
86
tern on the aryl ring seems to be important. Thus, methyl
and isopropyl groups or fused benzene rings enhanced the
yield significantly. Moreover, yield and conversion rose with
halogen substitution, but decreased when electron-donating
methoxy groups (in para and ortho positions) were intro-
duced. While none of these ligands gave improved results
for trans-stilbene compared to the 5-Cl-1-MeIm ligand
(Table 1, entry 1), these ligands display better performance
in the epoxidation of aliphatic olefins (see below).
Scope and limitations: To explore the scope and limita-
tions of the ligands, four 1-substituted imidazoles (3, 5, 9,
11) were applied on more challenging aromatic and aliphatic
substrates (Figure 2). In all experiments with yields greater
than 20%, good to excellent chemoselectivity (66 to
>99%) for epoxide formation was observed. In the case of
aromatic olefins the best performance was observed for sub-
strates containing mono- or di-substituted double bonds,
with yields of up to 94% (Figure 2a). Surprisingly, the only
exception was cis-stilbene, for which all catalysts showed
lower yields. On the other hand aliphatic substrates (Fig-
ure 2b) are oxidized with somewhat lower yields compared
to aromatic counterparts. However, compared to other
known iron catalysts excellent results are obtained for both
types of olefins with the imidazole-based systems.
4
5
9
7
73
5
6
NaOH
NaHO3
5
5
79
86
71
73
90
84
7[d]
5
84
83
>99
8
5
5
86
85
58
76
84
84
84
37
75
81
98
99
63
99
97
9
10
11
12
10
2.5
1
13
5
29
23
79
14
15
H3PO4
HNO3
5
5
0
69
0
55
0
81
16
5
81
76
95
For both aliphatic and aromatic substrates imidazole
ligand 5 gave the best results. Tables 3 and 4 detail the
scope and limitations of this novel iron catalyst system.
The aromatic substrates trans-stilbene and styrene
(Table 3, entries 1 and 3) gave good to excellent yields and
high selectivity similar to or improved with respect to the
parent 5-Cl-1-MeIm system.[14] p-Chlorostyrene showed a
slight enhancement in yield and selectivity (Table 3, entry 5)
compared to 5-Cl-1-MeIm, whereas the epoxidation yield of
styrene is improved by about 15%. Also, the more difficult
trisubstituted olefin a-methyl-trans-stilbene gave an im-
proved yield relative to the former system. Most notably,
the activity of the catalyst system strongly rises for aliphatic
olefins (Table 4) in comparison to the 5-Cl-1-MeIm system.
Internal olefins such as trans-2-octene and trans-5-decene
(Table 4, entries 3 and 6), show 44–49% conversion and
greater than 80% selectivity. More specifically, the yield of
trans-5-decene increases from 13 to 37% compared with 5-
Cl-1-MeIm. Moreover, the cyclic olefin cyclooctene is oxi-
dized in good yield and selectivity. Here, the yield increases
from 28% with 5-Cl-1-MeIm to 65% with imidazole 5
(Table 4, entry 1).
17
18
19
20
21
22
LiCl
LiCl
LiCl; HCl
KCl
NH4Cl
NBS
5
10
5; 0.5
89
87
86
80
87
88
87
85
83
73
85
85
98
98
98
91
97
97
5
5
5
23
5
89
86
97
[a] Reaction conditions: 0.5 mmol trans-stilbene, 5 mol% FeCl3·6H2O
and 10 mol% 5-Cl-1-MeIm, tert-amyl alcohol (9 mL), 0.44 mmol dodec-
ane as internal standard were added in sequence at room temperature in
air. To this mixture, a solution of 30% H2O2 (115 mL, 1 mmol) in tert-
amyl alcohol (885 mL) was added over 1 h at room temperature in air by
syringe pump. [b] Conversion and yield were determined by GC analysis.
[c] Selectivity refers to the chemoselectivity of epoxide from olefin.
[d] No ee detectable.
stilbene was retained as model substrate. The results are
presented in Table 2. A broad range of imidazoles lead to
moderate to good yields (Table 2, entries 3–6, 8–12). In gen-
eral, N-aryl imidazoles gave better results than N-benzyl de-
rivatives (Table 2, entries 1–3). Besides, the substitution pat-
Chem. Eur. J. 2009, 15, 5471 – 5481
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5473

M. Beller et al.
Table 2. Variation of the imidazole ligand.[a]
Table 2. (Continued)
Entry
Imidazole
Conversion Yield Selectivity
[%][b] [%][b]
[%][c]
Entry
Imidazole
Conversion Yield Selectivity
[%][b] [%][b] [%][c]
11
11
12
13
54
74
4
50
66
2
94
1
1
38
32
85
12
13
89
56
2
3
2
3
37
56
31
48
85
85
[a] Reaction conditions: 0.5 mmol trans-stilbene, 5 mol% FeCl3·6H2O
and 10 mol% imidazole derivative, tert-amyl alcohol (9 mL), 0.44 mmol
dodecane as internal standard were added in sequence at room tempera-
ture in air. To this mixture a solution of 30% H2O2 (170 mL, 1.5 mmol) in
tert-amyl alcohol (830 mL) was added over 1 h at room temperature by
syringe pump. [b] Conversion and yield were determined by GC analysis.
[c] Selectivity refers to the chemoselectivity of epoxide from olefin.
4
5
4
5
83
72
78
66
95
92
For the poorly reactive and therefore challenging terminal
aliphatic olefins, catalytic methods for epoxidation are limit-
ed.[21] The 5/iron system gave 53% yield for 2-methyl-1-hep-
tene (Table 4, entry 2) and 18% yield for 1-octene. Interest-
ingly, N-trityl-imidazole gave 25% conversion, 20% yield
and 80% selectivity. In addition, the catalyst demonstrates
good chemoselectivity for the olefin over the hydroxyl
group (Table 4, entry 4), while the 5-Cl-1-MeIm system gave
no yield at all with 1-hydroxy-3-hexene.
Mechanistic considerations: The substitution pattern of
the imidazole ligand is of crucial importance for obtaining
significant catalyst activity. As shown in Scheme 3 substitu-
tion in the 2-position of the imidazole moiety leads to a sig-
nificant depletion of yield compared to unsubstituted coun-
terparts. For instance, a catalyst containing imidazole as
ligand afforded stilbene oxide in 38% yield, while a catalyst
based on 2-methylimidazole gave stilbene oxide in less than
5% yield. This indicates that the 2-position is of main im-
portance.
One reason for the decreased reactivity of the 2-substitut-
ed imidazoles might be participation of a carbene-type
ligand in the reaction due to their s-donor strength.[22] How-
ever, formation of iron carbene complexes (in situ formation
of the carbene, see Supporting Information) with different
carbene ligands showed neither conversion nor yield
(Scheme 4). Hence, it is unlikely that carbene-type ligands
take part in this reaction.
6
7
6
7
75
18
74
13
98
74
8
8
70
76
46
65
70
41
93
92
90
9
9
10
10
Another possibility for the different reactivity could be in-
teraction between the proton in 2-position of the coordinat-
ed imidazole ligand and the intramolecular iron species
(e.g., ferric hydroperoxo species FeOOH) due to hydrogen
5474
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5471 – 5481

Biomimetic Iron–Imidazole Catalyst System
FULL PAPER
Figure 2. Epoxidation of aliphatic (a) and (b) aromatic olefins in the presence of imidazoles 3, 5, 9 and 11.
Table 3. Scope and limitations of aromatic olefins.[a]
Scheme 3). Similar results were achieved with imidazole and
2-methylimidazole.[25]
Next, we correlated the results of Table 2 with the 1H
NMR chemical shift of the free ligands. It is known that
metal–ligand interactions facilitate hydrogen bonding.[26]
Figure 3 shows the correlation of yield and selectivity with
À
the chemical shift for the N=C(H) N proton. Unfortunately,
Entry
Substrate
Conversion
[%][b]
Yield
[%][b]
Selectivity
[%][c]
no clear dependency of yield and chemical shift is observa-
ble. For further understanding of the active catalysts, UV/
Vis spectroscopic investigations with FeCl3·6H2O in the
presence of the best imidazole ligand, namely, 2,6-diisoprop-
yl-N-phenylimidazole (IPrPIm), and simple imidazole were
1
2
72
32
66
25
92
77
3
4
100
34
87
23
87
67
underACTHNURTGNEUtGN aken. The experiments were performed in analogy to
the reaction conditions of the epoxidation (vide supra). Due
to the characteristic iron–imidazole bands in the UV/Vis
spectra, it is possible to determine the ligand-to-metal ratio
in solution.[27] Owing to the strong absorbance of the solvent
tert-amyl alcohol, the measurements were analyzed above
225 nm.[28] In the range of 250–400 nm typical high-energy
transitions for iron complexes are observed (Figure 4).[29] On
addition of imidazole 5 a broad shoulder appeared which is
assigned to a metal-to ligand charge-transfer band of the
emerging iron complex.[31]
5
6
92
94
88
94
96
>99
[a] Reaction conditions: 0.5 mmol trans-stilbene, 5 mol% FeCl3·6H2O
and 10 mol% imidazole 5, tert-amyl alcohol (9 mL), 0.44 mmol dodecane
were added in sequence at room temperature in air. To this mixture, a so-
lution of 30% H2O2 (170 mL, 1.5 mmol) in tert-amyl alcohol (830 mL) was
added over 1 h at room temperature by syringe pump. [b] Conversion
and yield were determined by GC analysis. [c] Selectivity refers to the
chemoselectivity of epoxide from olefin.
In the UV/Vis titration experiments, various amounts of 5
were added to a solution of FeCl3·6H2O in tert-amyl alcohol.
Addition of the ligand led to a linear increase in absorbance
at all four chosen wavelengths; the highest absorption is ob-
served at 430 nm. A plateau indicating quantitative complex
formation is reached at a ligand-to-metal ratio of 3:1
(Figure 4). Moreover, a large excess of the imidazole ligands
(10 and 40 equiv ligand to 1 equiv iron) did not show any
considerable change, and the position of the plateau was
constant. Both full concentration (Figure 5a) and half-con-
centration (Figure 5b) of the imidazole/iron system resulted
bonding, or an intermolecular interaction between free im-
ACHTUNGTRENNUNGidazole and the iron species mentioned above (Scheme 5).
Similar effects of intramolecular or intermolecular interac-
tions are proposed in other oxidation reactions.[11e,23,24]
An intermolecular interaction is excluded, since catalysts
containing a mixture of N-benzylimidazole and N-benzyl-2-
methylimidazole in a ratio of 9:1 (13.5:1.5) gave comparable
yields (7% and 5%, respectively) of stilbene oxide to those
obtained with pure N-benzyl-2-methylimidazole (5%, see
Chem. Eur. J. 2009, 15, 5471 – 5481
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5475

M. Beller et al.
Table 4. Scope and limitations of aliphatic olefins.[a]
ligands 5 and 12 (Figure 6). In
case of imidazole 5 it was even
possible to crystallize the cor-
responding iron complex di-
rectly from the tert-amyl alco-
hol solvent of the epoxidation
reaction.
The observed m-oxo dinu-
clear N4(Cl)FeOFeCl3 pattern
is known for both porphyrin
and non-heme iron-containing
enzymes.[33] Additionally, this
type of complex was postulat-
ed as intermediate in oxygen
transfer from FeIV to FeII.[34]
Recently, it was also confirmed
that the formation of tetraden-
Entry
1
Substrate
Conversion
[%][b]
Yield
[%][b]
Selectivity
[%][c]
77
65
84
2
3
67
44
53
37
82
82
4
5
6
38
24
49
28
18
42
74
75
86
tate m-oxo diironACTHNUTRGNEUGN(III) species is
based on the reaction between
iron(II) complexes and dioxy-
gen.[35] Here, the formation of
the m-oxo complex should take
place via incorporation of
water, because no change of
oxidation state is observed.
7
26
8
31
[a] Reaction conditions: 0.5 mmol trans-stilbene, 5 mol% FeCl3·6H2O and 10 mol% imidazole derivative, tert-
amyl alcohol (9 mL), 0.44 mmol dodecane, were added in sequence at room temperature in air. To this mix-
ture, a solution of 30% H2O2 (170 mL, 1.5 mmol) in tert-amyl alcohol (830 mL) was added over 1 h at room
temperature by syringe pump. [b] Conversion and yield were determined by GC analysis. [c] Selectivity refers
to the chemoselectivity of epoxide from olefin.
In general, for the unsym-
metrical diironACTHNUTRGNEUGN(III) complexes
A and B the coordination ge-
ometry at Fe1 can be de-
scribed as an octahedron,
whereas that at Fe2 is tetrahe-
dral (Figure 6). One chlorine
atom, the m-oxo bridge and
four nitrogen atoms of the im-
AHCTUNGTERGiNNUN dazole ligands coordinate to
Fe1, and three further chlorine
atoms and the oxygen atom
which bridges the metal cen-
ters to Fe2. Complexes A and
B can be formally classified
into one “cationic” part with
the Fe1 center and one “anion-
ic” part consisting of Fe2 and
Scheme 3. Some selected results of imidazole ligands in the epoxidation of trans-stilbene (reaction conditions
are similar to those reported in Table 2).
four anionic ligands. In contrast to known tetradentate poly-
imidazole[36] or tris(6-pyridylmethyl)amine ligands,[35] the
chlorido ligand (Figure 6, Cl1) is situated trans to the oxo
À
bridge. Compared to known complexes, the Fe1 Cl1 bond
length in A and B are also significantly longer.
Scheme 4. Use of iron carbene complexes in the epoxidation of stilbene.
in the same ligand-to-metal ratio and did not show any sig-
nificant difference.[32] Hence, it is likely that the major iron
complex in solution contains three imidazole ligands.
Next, the crystallization of defined molecular iron imid-
ACHTUNGTRENNUNGazole complexes was tried, and led to the synthesis of two
Scheme 5. Possible interactions with H-2 in iron imidazole complexes. a)
Intramolecular interaction; b) intermolecular interaction.
new binuclear iron
ACHTUNGTRENNUNG(III) complexes A and B with imidazole
5476
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5471 – 5481

Biomimetic Iron–Imidazole Catalyst System
FULL PAPER
surround the iron center in the
solid state. As expected for d5
FeIII, the geometry at the metal
center is octahedral with the
chlorine atoms in a transoid ar-
rangement. Bond lengths and
angles are similar to those of
known iron complexes with
four simple imidazole ligands,
reported by Obrey et al.[37] and
later Cotton et al.[38] For com-
parison we synthesized the
known complex according to
the protocol by Cotton et al.
The catalytic reactivity towards
the model substrate trans-stil-
bene is presented in Table 5.
Remarkably, both mononu-
clear complexes gave higher
conversion and yield than the
in situ systems (Table 5). Clear-
ly, the novel mononuclear 5-
Cl-1-MeIm iron complex is a
powerful and highly selective
defined epoxidation catalyst.
À
Figure 3. Yield and selectivity of trans-stilbene oxide formation versus chemical shift of N=C(H) N proton of
aromatic N-substituted imidazoles.[30]
Figure 4. Selected UV/Vis spectra during the titration.
The catalytic reactions of the defined complexes are
shown in Schemes 6 and 7. Both complexes exhibit approxi-
mately half of the activity of the in situ system. Surprisingly,
further addition of imidazole or tetra-n-butylammonium
chloride led to considerably lower yields of the epoxide.
However, addition of imidazole hydrochloride dramatically
increased the yield and led to activity close to that of the in
situ system. When we added 5 mol% of HCl to the reaction
mixture we obtained 49% conversion and 45% yield for
complex A, and 60% conversion and 54% yield for com-
plex B. Thus, the m-oxo diiron complexes can be converted
to the active species again. Apparently, this type of complex
is a less active species which is in equilibrium with the
active catalyst.
Finally, we were able to synthesize a new mononuclear
FeIII complex C with the best in situ ligand for epoxidation
Figure 5. UV/Vis titration of FeCl3·6H2O with imidazole at full (a) and
of aromatic olefins (Figure 7). Here, four imidazole ligands
half concentration (b).
Chem. Eur. J. 2009, 15, 5471 – 5481
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5477

M. Beller et al.
Conclusions
In conclusion, we have developed and examined novel bio-
inspired iron catalysts for hydrogen peroxide based epoxida-
tion of both aromatic and aliphatic olefins. For the first time
we have shown that defined iron imidazole complexes are
capable of epoxidizing challenging aliphatic olefins. In gen-
eral, high chemoselectivity is achieved, whereas cyclooctene
and 2-methyl-1-heptene showed the best yields. Besides,
mechanistic investigations demonstrate that mononuclear
ironACTHNUGRTENNUG(III) species are in equilibrium with an m-oxo diironACHTUNGTRENNUNG(III)
complex by reaction with water. Three new defined molecu-
lar iron complexes were isolated and characterized by X-ray
diffraction. Notably, these complexes show significant cata-
lytic activity. Further work concerning transient oxygenated
species and detailed information about the catalytic cycle is
ongoing.
Experimental Section
General remarks: The imidazole ligands (Table 2, entries 1, 4–10) were
synthesized according to literature protocols.[40] All other reagents were
used as purchased from commercial suppliers (Aldrich, Fluka, Merck,
etc.) without further purification. “30%” aqueous H2O2 from Merck was
used as received. The peroxide content varied from 30 to 40%, as deter-
mined by titration. GC analyses were performed with a Hewlett Packard
HP 6890 model spectrometer. GC calibrations for alkenes and epoxides
were carried out with authentic samples and dodecane as internal stan-
dard. NMR spectra were measured on a Bruker ARX 300 or ARX 400
spectrometer. For more analytical details, see Supporting Information.
Figure 6. Molecular structure of [{N-(2,6-diisopropylphenyl)imidazo-
le}4ClFeOFeCl3] (A) and [(N-phenylimidazole)4ClFeOFeCl3] (B). Ther-
mal ellipsoids correspond to 30% probability. Hydrogen atoms are omit-
ted for clarity.
Epoxidation of olefins (general procedure): In a test tube, FeCl3·6H2O
(0.025 mmol), tert-amyl alcohol (9 mL), imidazole ligand (0.050 mmol),
olefin (0.50 mmol) and dodecane (GC internal standard, 100 mL) were
added in sequence at room temperature in air. A solution of 30% aque-
ous hydrogen peroxide (170 mL, 1.5 mmol) in tert-amyl alcohol (830 mL)
was added to this mixture over 1 h at room temperature by a syringe
pump. Conversion and yield were determined by GC analysis without
further manipulations.
In view of all these observations we propose the mecha-
nism shown in Scheme 8. In agreement with UV/Vis investi-
gations we assume equilibrium of dinuclear and mononu-
clear iron complexes, which undergo fast ligand/solvent ex-
change in the reaction mixture.[39] Free coordination sites on
the metal center may react with hydrogen peroxide to form
a highly reactive iron hydroperoxo complex. The imidazole
ligand stabilizes intramolecularly the corresponding hydro-
peroxo species by hydrogen bonding at the 2-position.
UV/Vis analysis: UV/Vis measurements were performed at 208C with a
Specord S600 spectral photometer. 1 cm quartz cuvettes were used with
solutions prepared as follows: In analogy to the reaction, FeCl3·6H2O
(6.76 mg, 0.05 mmol) was stirred together with x equiv imidazole in 9 mL
of tert-amyl alcohol. From the clear solution 3 mL was transferred into
Scheme 6. Activity of the synthesized ironACTHNUTRGNEUNG(III)complex with N-(2,6-diisopropyl)phenylimidazole (5).
5478
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5471 – 5481

Biomimetic Iron–Imidazole Catalyst System
FULL PAPER
Scheme 7. Activity of the synthesized iron
dazole.
ACHTUNGTRNE(NUNG III) complex with N-phenylimi-
Table 5. Comparison of the in situ catalyst system and defined iron com-
plexes.[a]
Entry Catalyst system
Conversion Yield Selectivity
[%][c] [%][c]
[%][d]
1[a]
2[a]
3[b]
4[b]
trans-[FeCl2(5-Cl-1-MeIm)4]Cl
99
91
33
21
91
84
26
18
92
93
77
85
in situ system (5-Cl-1-MeIm)
trans-[FeCl2(imidazole)4]Cl
in situ system (imidazole)
[a] Reaction conditions: 0.25 mmol trans-stilbene, 5 mol% iron complex
(e.g., 5 mol% FeCl3·6H2O and 20 mol% 5-chloro-1-methylimidazole),
tert-amyl alcohol (5 mL), 0.44 mmol dodecane as internal standard were
added in sequence at room temperature in air. To this mixture, a solution
of 30% H2O2 (85 mL, 0.75 mmol) in tert-amyl alcohol (915 mL) was added
over 1 h at room temperature by syringe pump. [b] Reaction conditions:
0.5 mmol trans-stilbene, 5 mol% iron complex (e.g., 5 mol% FeCl3·6H2O
and 20 mol% imidazole), tert-amyl alcohol (9 mL), 0.44 mmol dodecane
as internal standard were added in sequence at room temperature in air.
To this mixture, a solution of 30% H2O2 (115 mL, 1.0 mmol) in tert-amyl
alcohol (885 mL) was added over 1 h at room temperature by syringe
pump. [c] Conversion and yield were determined by GC analysis. [d] Se-
lectivity refers to the chemoselectivity of epoxide from olefin.
Figure 7. Molecular structure of trans-[FeCl2(5-chloro-N-methylimida-
zole)4]Cl (C). Thermal ellipsoids correspond to 30% probability. Hydro-
gen atoms are omitted for clarity.
ACHTUNGTRENNUNG
quartz cuvettes. Temperature-controlled measurements were recorded at
four different wavelengths and at least three times for every data point.
X-ray crystal structure analysis: Data were collected on a STOE IPDS II
diffractometer using graphite-monochromated MoKa radiation. The struc-
tures were solved by direct methods (SHELXS-97)[41] and refined by full-
matrix least-squares techniques on F2 (SHELXL-97)[40] XP (Bruker
AXS) was used for graphical representations.
¯
Compound A: C60H80Cl4Fe2N8O, Mr =1182.82, triclinic, space group P1,
a=11.8817(5), b=13.3486(5), c=22.0564(9) ꢃ, a=103.311(3), b=
Compound B: C36H32Cl4Fe2N8O, Mr =846.20, monoclinic, space group
P21/n, a=13.8321(5), b=15.7173(4), c=23.4636(10) ꢃ, b=104.677(3)8,
105.083(3), g=94.950(3)8, V=3246.9(2) ꢃ3, Z=2, 1calcd =1.210 gcmÀ3
,
m=0.654 mmÀ1, T=200 K, 52163 reflections measured, 14921 independ-
ent reflections (Rint =0.0817), of which 6593 were observed (I>2s(I)),
R1 =0.0407 (I> 2s(I)), wR2 =0.0697 (all data), 668 refined parameters.
V=4934.6(3) ꢃ3, Z=4, 1calcd =1.139 gcmÀ3 m=0.836 mmÀ1
, , T=200 K,
53587 reflections measured, 9186 independent reflections (Rint =0.0594),
of which 4872 were observed (I>2s(I)), R1 =0.0356 (I> 2s(I)), wR2 =
Scheme 8. Proposed mechanism of the epoxidation reaction with the iron/imidazole system.
Chem. Eur. J. 2009, 15, 5471 – 5481
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5479

M. Beller et al.
0.0727 (all data), 412 refined parameters. PLATON/SQUEEZE[42] was
used to remove disordered solvent.
2003, 5, 2469–2472; TACN=triazacyclononane, e) A. Company, L.
Gꢅmez, X. Fontrodona, X. Ribas, M. Costas, Chem. Eur. J. 2008, 14,
5727–5731.
Compound C: C16H25Cl7FeN8O2.5, Mr =673.44, monoclinic, space group
C2/c, a=13.6304(3), b=13.2331(3), c=30.5846(6) ꢃ, b=92.843(2)8, V=
5509.8(2) ꢃ3, Z=8, 1calcd =1.624 gcmÀ3, m=1.259 mmÀ1, T=200 K, 40596
reflections measured, 5829 independent reflections (Rint =0.0349), of
which 4264 were observed (I> 2s(I)), R1 =0.0335 (I>2s(I)), wR2 =
0.0864 (all data), 317 refined parameters.
[12] For a list of common oxidants, their active oxygen contents and
waste products, see: J.-E. Bꢁckvall, Modern Oxidation Methods,
Wiley-VCH, Weinheim 2004, p. 22.
[13] a) B. Bitterlich, G. Anilkumar, F. G. Gelalcha, B. Spilker, A. Grote-
vendt, R. Jackstell, M. K. Tse, M. Beller, Chem. Asian J. 2007, 2,
521–529; b) F. G. Gelalcha, B. Bitterlich, G. Anilkumar, M. K. Tse,
M. Beller, Angew. Chem. 2007, 119, 7431–7435; Angew. Chem. Int.
Ed. 2007, 46, 7293–7296; c) G. Anilkumar, B. Bitterlich, F. Gadissa
Gelalcha, M. K. Tse, M. Beller, Chem. Commun. 2007, 289–291.
[14] K. Schrçder, X. Tong, B. Bitterlich, M. K. Tse, F. G. Gelalcha, A.
Brꢀckner, M. Beller, Tetrahedron Lett. 2007, 48, 6339–6342.
[15] B. Bitterlich, K. Schrçder, M. K. Tse, M. Beller, Eur. J. Org. Chem.
2008, 4867–4870.
CCDC-714375, CCDC-714376, and CCDC-714377 contain the supple-
mentary crystallographic data for this paper. These data can be obtained
free of charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif
[16] F. Shi, M. K. Tse, Z. Li, M. Beller, Chem. Eur. J. 2008, 14, 8793–
8797.
Acknowledgements
[17] a) C. Dçbler, G. Mehltretter, M. Beller, Angew. Chem. 1999, 111,
3211–3212; Angew. Chem. Int. Ed. 1999, 38, 3026–3028; b) C.
Dçbler, G. Mehltretter, U. Sundermeier, M. Beller, J. Am. Chem.
Soc. 2000, 122, 10289–10297.
[18] M. Fujita, L. Que, Jr., Adv. Synth. Catal. 2004, 346, 185–190.
[19] E. A. Duban, K. P. Bryliakov, E. P. Talsi, Eur. J. Inorg. Chem. 2007,
852–857.
We thank the State of Mecklenburg-Western Pommerania, the Federal
Ministry of Education and Research (BMBF) and the Deutsche For-
schungsgemeinschaft (SPP 1118 and Leibniz prize) for financial support.
K.S. appreciates the financial support provided by the Graduiertenkolleg
1213 “Neue Methoden fꢀr Nachhaltigkeit in Katalyse und Technik” and
the Max-Buchner-Forschungsstiftung (DECHEMA). Dr. A. Surkus is ac-
knowledged for the photo in the graphical abstract.
[20] Same reaction conditions as described in Table 1 except for 35%
peracetic acid solution as oxidant.
[21] M. Colladon, A. Scarso, P. Sgarbossa, R. A. Michelin, G. Strukul, J.
Am. Chem. Soc. 2007, 129, 7680–7689.
[22] W. A. Herrmann, Angew. Chem. 2002, 114, 1342–1362; Angew.
Chem. Int. Ed. 2002, 41, 1290–1309.
[23] See also: K. Zamani, A. Mobinikhaledi, N. Foroughifar, K. Faghihi,
V. Mahdavi, Turk. J. Chem. 2003, 27, 71–75.
[24] E. Derat, S. Shaik, C. Rovira, P. Vidossich, M. Alfonso-Prieto, J.
Am. Chem. Soc. 2007, 129, 6346–6347.
[25] Also the combined catalyst containing a mixture of imidazole and 2-
methylimidazole in a ratio of 9:1 (13.5:1.5) gave comparable yields
of 8% (7%) of stilbene oxide, as obtained in the case of pure 2-
methylimidazole (4%).
[26] W. W. Henderson, R. E. Shepherd, J. Abola, Inorg. Chem. 1986, 25,
3157–3163.
[27] Examples of UV/Vis titrations: a) L. Rao, G. R. Choppin, R. J. Ber-
geron, Radiochim. Acta 2000, 88, 851–856; b) B. Strauß, W. Linert,
V. Gutmann, R. F. Jameson, Monatsh. Chem./Chem. Monthly 1992,
123, 537–546; c) H. Hofmeier, A. El-ghayoury, A. P. H. J. Schen-
ning, U. S. Schubert, Chem. Commun. 2004, 318–319.
[28] For the strong absorbance of tert-amyl alcohol relative to methanol,
ethanol, and heptane, see: H.-H. Percampus, UV-VIS Atlas of Or-
ganic Compounds, Parts 1 and 2, 2nd ed., VCH, Weinheim, 1992.
[29] M.-C. Rodriguez, F. Lambert, I. Morgenstern-Badarau, Inorg.
Chem. 1997, 36, 3525–3531.
[30] Chemical shifts were either taken from reported values or deter-
mined by NMR measurements of authentic samples. For literature,
see: a) L. Zhu, P. Guo, G. Li, J. Lan, R. Xie, J. You, J. Org. Chem.
2007, 72, 8535–8538; b) B. E. Ketz, A. P. Cole, R. M. Waymouth,
Organometallics 2004, 23, 2835–2837; c) E. Alcalde, I. Dinares, S.
Rodriguez, C. Garcia de Miguel, Eur. J. Org. Chem. 2005, 1637–
1643.
[31] Literature analogues indicate the MLCT band: a) R. Johnson, R. E.
Shepherd, Inorg. Chem. 1983, 22, 3506–3513; b) E.-G. Jꢁger, E.
Stein, F. Grꢁfe, W. Schade, Z. Anorg. Allg. Chem. 1985, 526, 15–28;
c) J.-Y. Xu, J.-L. Tian, H.-D. Bian, S.-P. Yang, D.-Z. Liao, P. Cheng,
P.-W. Shen, Appl. Organomet. Chem. 2007, 21, 129–134.
[32] See Supporting Information for the UV/Vis titration spectra of im-
[1] H. Orzesek, R. P. Schulz, U. Dingerdissen, W. F. Maier, Chem. Eng.
Technol. 1999, 22, 691–700.
[2] R. Murugavel, H. W. Roesky, Angew. Chem. 1997, 109, 491–494;
Angew. Chem. Int. Ed. Engl. 1997, 36, 477–479.
[3] S. Linic, H. Piao, K. Adib, M. A. Barteau, Angew. Chem. 2004, 116,
2978–2981; Angew. Chem. Int. Ed. 2004, 43, 2918–2921.
[4] For our recent work on Ru-based oxidations, see: a) M. K. Tse, C.
Dçbler, S. Bhor, M. Klawonn, W. Mꢁgerlein, H. Hugl, M. Beller,
Angew. Chem. 2004, 116, 5367–5372; Angew. Chem. Int. Ed. 2004,
43, 5255–5260; b) M. K. Tse, M. Klawonn, S. Bhor, C. Dçbler, G.
Anilkumar, H. Hugl, W. Mꢁgerlein, M. Beller, Org. Lett. 2005, 7,
987–990; c) S. Bhor, G. Anilkumar, M. K. Tse, M. Klawonn, C.
Dçbler, B. Bitterlich, A. Grotevendt, M. Beller, Org. Lett. 2005, 7,
3393–3396; d) F. Shi, M. K. Tse, M. Beller, J. Mol. Catal. 2007, 270,
68–75; e) F. Shi, M. K. Tse, M. Beller, Adv. Synth. Catal. 2007, 349,
303–308.
[5] W. A. Herrmann, R. W. Fischer, D. W. Marz, Angew. Chem. 1991,
103, 1706–1709; Angew. Chem. Int. Ed. Engl. 1991, 30, 1638–1641.
[6] E. M. McGarrigle, D. G. Gilheany, Chem. Rev. 2005, 105, 1563–
1602.
[7] For an excellent review on iron-based catalysts, see: a) C. Bolm, J.
Legros, J. Le Paih, L. Zani, Chem. Rev. 2004, 104, 6217–6254;
b) B. S. Lane, K. Burgess, Chem. Rev. 2003, 103, 2457–2473; c) S.
Enthaler, K. Junge, M. Beller, Angew. Chem. 2008, 120, 3363–3367;
Angew. Chem. Int. Ed. 2008, 47, 3317–3321.
[8] M. Merkx, D. A. Kopp, M. H. Sazinsky, J. L. Blazyk, J. Mꢀller, S. J.
Lippard, Angew. Chem. 2001, 113, 2860–2888; Angew. Chem. Int.
Ed. 2001, 40, 2782–2807.
[9] a) L. C. Blasiak, F. H. Vaillancourt, C. T. Walsh, C. L. Drennan,
Nature 2006, 440, 368–371; b) PDB ID: 2fct; Data were taken from
Research Collaboratory for Structural Bioinformatics PDB (http://
www.pdb.org/) and arranged with MBT Protein workshop. As men-
tioned in the text, a water molecule is omitted in Figure 1.
[10] M. Costas, M. P. Mehn, M. P. Jensen, L. Que, Jr., Chem. Rev. 2004,
104, 939–986.
[11] mep=N,Nꢄ-dimethyl-N,N-bis(2-pyridylmethyl)ethane-1,2-diamine,
a) M. C. White, A. G. Doyle, E. N. Jacobsen, J. Am. Chem. Soc.
2001, 123, 7194–7195; b) K. Chen, M. Costas, J. Kim, A. K. Tipton,
L. Que, Jr., J. Am. Chem. Soc. 2002, 124, 3026–3035; TPA=tris(2-
pyridylmethyl)amine,c) J. Y. Ryu, J. Kim, M. Costas, K. Chen, W.
Nam, L. Que, Jr., Chem. Commun. 2002, 1288–1289; phen=
phenanthroline, d) G. Dubois, A. Murphy, T. D. P. Stack, Org. Lett.
AHCTUNGTRENNiGUN dazole 5.
[33] a) The Porphyrin Handbook, Vol. 2 (Eds.: K. M. Kadish, K. M.
Smith, R. Guilard), Academic Press, New York, 2000; b) R. M.
Theisen, J. Shearer, W. Kaminsky, J. A. Kovacs, Inorg. Chem. 2004,
43, 1545–1550.
5480
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5471 – 5481

Biomimetic Iron–Imidazole Catalyst System
FULL PAPER
[34] C. V. Sastri, K. Oh, Y. J. Lee, M. S. Seo, W. Nam, Angew. Chem.
2006, 118, 4096–4099; Angew. Chem. Int. Ed. 2006, 45, 3992–3995.
[35] a) A. Machkour, D. Mandon, M. Lachkar, R. Welter, Inorg. Chem.
2004, 43, 1545–1550; b) N. K. Thallaj, O. Rotthaus, L. Benhamou,
N. Humbert, M. Elhabiri, M. Lachkar, R. Welter, A.-M. Albrecht-
Gary, D. Mandon, Chem. Eur. J. 2008, 14, 6742–6753.
[36] J. Wang, M. S. Mashuta, Z. Sun, J. F. Richardson, D. N. Hendrickson,
R. M. Buchanan, Inorg. Chem. 1996, 35, 6642–6643.
[39] There are no differences in yield and conversion between ligand:Fe
ratios of 2:1 and 3:1 in the in situ system. This implies that iron spe-
cies with lower coordination numbers in solution may be involved in
the catalytic cycle and/or that further ligand/solvent exchange
occurs.
[40] For detailed information including NMR, MS and HRMS data, see
the Supporting Information.
[41] G. M. Sheldrick, Acta Crystallogr. Sect. A 2008, 64, 112–122.
[42] A. L. Spek, PLATON:
A Multipurpose Crystallographic Tool,
[37] S. J. Obrey, S. G. Bott, A. R. Barron, J. Chem. Crystallogr. 2000, 30,
61–63.
Utrecht University, Utrecht, The Netherlands, 2005.
[38] S. A. Cotton, V. Franckevicius, F. Fawcett, Polyhedron 2002, 21,
2055–2061.
Received: December 26, 2008
Published online: April 15, 2009
Chem. Eur. J. 2009, 15, 5471 – 5481
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5481