Iron coordination chemistry of N2Py2 ligands substituted by carboxylic moieties and their impact on alkene oxidation catalysis

Article abstract of DOI:10.1002/ejic.201100785

A biomimetic approach based on Rieske dioxygenase mimics has been undertaken, which uses the tetradentate N₂Py₂ ligand platform that contains two pyridine moieties linked to a 1,2-diaminoethane or a trans-1,2-diaminocyclohexane backb

Full text of DOI:10.1002/ejic.201100785

FULL PAPER
DOI: 10.1002/ejic.201100785
Iron Coordination Chemistry of N2Py2 Ligands Substituted by Carboxylic
Moieties and Their Impact on Alkene Oxidation Catalysis
Frédéric Oddon,[a,b,c] Elodie Girgenti,[a,b,c] Colette Lebrun,[d]
Caroline Marchi-Delapierre,[a,b,c] Jacques Pécaut,[d] and Stéphane Ménage*[a,b,c]
Keywords: Bioinorganic chemistry / Biomimetic synthesis / Asymmetric synthesis / Homogeneous catalysis / Epoxidation /
Oxidation / Iron
A biomimetic approach based on Rieske dioxygenase mimics
has been undertaken, which uses the tetradentate N2Py2 li-
gand platform that contains two pyridine moieties linked to
tion sphere. Moreover, the two carboxylic groups of the li-
gand remain protonated and bound by its sp2 oxygen atom.
Two substitutions of the ligand with carboxylic moieties was
found to be deleterious to the reactivity of the complex dur-
ing alkene oxidation, whereas monosubstitution led to a
slight change in the reactivity. Moreover, the catalyst built
from the optically active trans-1,2-diaminocyclohexane back-
bone, L3, catalyzes the asymmetric epoxidation of trans-2-
heptene with up to 17% yield. The addition of acetic acid
leads to better selectivity, enantioselectivity (38%), and yield
of the epoxidation of nonaromatic alkenes.
a
1,2-diaminoethane or a trans-1,2-diaminocyclohexane
backbone. Here we report the impact of the incorporation of
carboxylic functionalities with the N2Py2 ligand on the cata-
lytic efficiency of its FeII complexes during epoxidation with
H2O2 as the oxidant. Five complexes have been charac-
terized in the solid state and solution. The X-ray structure of
a ferrous complex with a ligand that contains two carboxylic
acid moieties shows an unexpected N4O3-type iron coordina-
Introduction
The design of homogeneous catalysts for oxidation re-
mains a great challenge. In general, the state-of-the-art con-
sists of catalytic processes that are environmentally friendly,
use inexpensive catalysts, less energy-consuming protocols,
and the use of sustainable oxidants.[1] One strategy lies in
mimicking metalloenzymes, by tentatively reproducing the
environment of mono- or dioxygenase active sites. The tar-
geted active site of nonheme iron oxygenases, which consists
of a mononuclear iron(II) center coordinated by the so-
called 2-His-1-carboxylate facial triad, has been a great
source of inspiration (Figure 1).[2]
For instance Rieske dioxygenases, such as naphthalene
dioxygenase (NDO), catalyze a stereospecific cis-dihydrox-
ylation during the biodegradation of arenes.[2b,3] This
metal-promoted biological oxidation has been particularly
targeted for the development of bioinspired catalysts be-
cause they fulfill some of the requirements listed above and
Figure 1. Representation of the catalytic iron site of NDO from
Pseudomonas sp. NCIB 9816-4.[2b]
perform the consumption of very inert substrates. A large
number of synthetic nonheme complexes has been devel-
oped as potential oxidation catalysts. Among them, cata-
lysts based on the full reproduction of the active site of
Rieske dioxygenases were found to be poorly active with
H2O2 as the oxidant, which suggests that the presence of
carboxylic groups plays an important role on the reactiv-
ity.[4] These outcomes contrast with the promising develop-
ment of iron catalysts that contain tetradentate pyridine/
amine (N2Py2) ligands with H2O2 as the oxidant.[5] Indeed,
selective alkane functionalization has been reported with
Fe[(S,S)-PDP] [(S,S)-PDP = 2-({(S)-2-[(S)-1-(pyridin-2-yl-
methyl)pyrrolidin-2-yl]pyrrolidin-1-yl}methyl)pyridine][6]
and [Fe(S,S,R)-mcpp](CF3SO3)2 [mcpp = N,NЈ-dimethyl-
N,NЈ-bis(4,5-pineno-2-picolyl)cyclohexane-1,2-diamine].[7]
Alkene oxidation has also been extensively studied,[8] which
revealed that the product formation, epoxide[8d–8f] or cis-
diols,[8g,8h] can be controlled by Fe(BPMEN) [BPMEN =
N,NЈ-dimethyl-N,NЈ-bis(2-pyridylmethyl)ethane-1,2-di-
[a] CEA, iRTSV, Laboratoire Chimie et Biologie des Métaux,
38054 Grenoble, France
E-mail: stephane.menage@cea.fr
[b] CNRS, UMR5249,
38054 Grenoble, France
[c] Université Joseph Fourier-Grenoble I, UMR5249,
38041 Grenoble, France
[d] Laboratoire de Reconnaissance Ionique et Chimie de
Coordination, LCIB (UMR-E 3 CEA/UJF-Grenoble 1),
INAC/SCIB,
38054 Grenoble, France
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201100785.
Eur. J. Inorg. Chem. 2012, 85–96
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
85

S. Ménage et al.
FULL PAPER
amine],[8d] Fe(bispidine),[8e] Fe(terpyridine),[8f] Fe(TPA),[8g] tional group is maintained close to the iron center in order
or Fe(diazapyridinophane)[8h] as catalysts. A chiral N2Py2 to mimic the proposed essential catalytic transition state.[10]
ligand has led to enantioselective cis-dihydroxylation up to Ferrous, instead of ferric, complexes were targeted, which
99% ee, but this family of catalysts remains inefficient for enabled us to retain the acidic form of the carboxylic moie-
enantioselective epoxidation.[9] The selectivity for epoxid- ties of the ligand as the Lewis acidity of the ferrous state is
ation versus cis-dihydroxylation is partly dependent on the lower than that of the ferric state.
presence of carboxylic acids in the reaction medium, but
Herein we report the synthesis and characterization of
their effect on the enantioselectivity has not been reported. five new ferrous complexes with L1, L2, L3, and L4. The
This additive was proposed to control the active species by synthesis of the ferrous complexes was found to be
the coordination of its acid form to a η1-Fe-peroxo species, counteranion dependent, and the complexes were formed
which is a precursor of the oxidizing iron species. This with chloride or more weakly coordinating (ClO4) ligands.
helped to favor heterolytic cleavage of the O–O bond by All the complexes were characterized by various spectro-
the transfer of its proton to the distal oxygen atom of the scopic techniques in solution and in the solid state, which
hydroperoxo adduct. The resulting high-valent iron species support the description of the complexes as hexa- or hepta-
epoxidized the alkene substrate with retention of configura- coordinate with bound carboxylic moieties. The complexes
tion.[10]
were tested in epoxidation catalysis, which revealed the im-
Carboxylic functionalities have controversial effects on pact of the intrinsic carboxylic group on the reactivity and
the efficiency of iron catalysts for the epoxidation of al- enantioselectivity.
kenes. The acidic form is of great interest for selectivity but
the basic form may bind to iron, which leads to the drastic
Results and Discussion
change to the electronic properties of the iron and then to
a well documented loss of catalyst efficiency.[11] To circum-
vent this antagonistic effect, we have undertaken the design
of iron catalysts using the N2Py2 platform [BPMEN or
BPMCN, BPMCN = N,NЈ-bis(2-pyridylmethyl)-N,NЈ-di-
methyl-trans-1,2-diaminocyclohexane],[8c] to which one or
two carboxylic moieties have been introduced (Figure 2).
Three new ligands have been synthesized, which differ in
the number of carboxylic functional groups and/or the na-
ture of the strap connecting the two 2-pyridylmethylamine
units (Figure 2). Achieving the aim of this work relies on
the design of complexes in which a carboxylic acid func-
Ligands Syntheses
The synthesis of L1 has already been described,[12] but as
the direct insertion of a carboxymethyl arm to BPHMEN
gave neither satisfactory yield, purity, or degree of proton-
ation, we decided to insert the carboxylic moiety by a two-
step route: alkylation by tert-butyl 2-bromoacetate followed
by smooth, quantitative hydrolysis by trifluoroacetic acid
(Scheme 1). The sensitivity of tert-butyl 2-bromoacetate
precludes the use of strong bases. Because of its intrinsic
basicity, BPHMEN was converted into its insoluble hydro-
bromide salt, which led to a limitation of the yield of 50%
for the alkylation step. Nevertheless, BPHMEN·HBr was
recycled. This afforded a suitable synthetic pathway that
was applied to the synthesis of L2, L3, and L4. L3 and L4
incorporate the chiral trans-diaminocyclohexane backbone,
(1S,2S) for L3 and (1R,2R) for L4. The chiral diamines were
used directly for the corresponding synthesis of L3 and L4.
Scheme 1. a) tert-Butyl 2-bromoacetate, K2CO3, CH3CN, 0 °C, 3 h.
b) CF3CO2H, CH2Cl2, 0 °C/30 min, room temp./17 h.
Characterization of the Iron Complexes
New mononuclear complexes with potentially hexaden-
tate L2 and L4 and pentadentate L1 and L3 ligands have
been prepared. We tackled ferrous complexes as the Lewis
acidity of the oxidation state is lower than that of the ferric
one. Consequently, the ferric state of the complexes ex-
cludes the retention of carboxylic forms, whereas the fer-
rous state allows them to coexist in the same complex. To
Figure 2. Representation of the ligands described in this study.
86
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 85–96

Fe Coordination Chemistry
prevent oxidation of the iron(II) complexes, they were all moieties of the ligand are protonated as FeII–O distances
synthesized in a glove box. Initially, FeCl2 was used to en- of 2.00–2.15 Å are generally reported for carboxylate li-
sure a stable coordination sphere. As the formed chlori- gands.[4a] The coordination mode of the carboxylic groups
nated complexes did not precipitate in some cases, we is rare in a discrete iron complex and has only been re-
switched to the use of Fe(ClO4)2 salts. The complexes were ported with two polynuclear species in which the carboxylic
synthesized by reacting 1 equiv. of Fe(ClO4)2 with 1 equiv. moiety connects two iron sites.[14] Finally, the Fe–O bond
of the ligand. When FeCl2 was used, 2 equiv. of iron salt length in [FeII(L4)(C3H6O)](ClO4)2 is long compared to that
were required to maximize the yield. In some cases, powder of another sp2 oxygen atom–FeII bond, which involves an
was deposited as a function of time, otherwise diethyl ether amido ligand.[15] Taken together, the ligand distances agree
was added to precipitate the product. All the complexes with a high-spin configuration of the ferrous ion in the solid
were characterized by IR, UV/Vis, and NMR spectroscopy, state. Seven-coordinate iron complexes are unusual, even if
ESI-MS, and elemental analysis. The presence of [FeCl4]2– iron complexes have been more extensively described than
as
a
counteranion for [Fe(BPMCN)(C3H6O)]2+ and their congeners.[16] Usually, examples of ferric complexes
[Fe(L1)(C3H6O)]2+ was deduced from spectroscopic studies predominate over ferrous complexes.[17]
(vide infra) and supported by satisfactory elemental analy-
ses, in which the presence of the Cl3Fe–O–FeCl3 dianion,
which is the product of the air oxidation of [FeCl4]2–, was
determined. Thanks to elemental analysis we were also able
to propose stoichiometries between FeLn and the FeCl4
counteranion. For example, in the case of the coordination
of L4 to iron, the presence of an extra chloride ligand to
complete the coordination sphere led to the presence of a
FeCl4 counteranion for two cationic species. The structures
of all the complexes were deduced by comparison of their
spectroscopic properties with those of the crystallographi-
cally characterized [FeII(L4)(C3H6O)](ClO4)2.
X-ray Crystallography of [FeII(L4)(C3H6O)](ClO4)2
Colorless crystals of [FeII(L4)(C3H6O)](ClO4)2 suitable
for X-ray analysis were obtained from the slow evaporation
Figure 3. ORTEP representation of [FeII(L4)(C3H6O)]2+. Hydrogen
of an acetone solution of the complex in a glove box. The
ORTEP representation of the cation is displayed in Fig-
ure 3. L4 wraps around the ferrous ion, which is heptacoor-
dinate. The polyhedron of the iron atom is a distorted pen-
tagonal bipyramid, in which the two nitrogen atoms from
the amine moieties, the sp2 oxygen atoms of the two carbox-
ylic moieties of the ligand, and the acetone solvent molecule
form the equatorial plane. The two pyridyl rings occupy
the axial positions and have a dihedral angle of 82.4°. This
unexpected N4O3 coordination derives from the full coordi-
nation of the ligand, of which the four nitrogen atoms are
in a “cis α” topology and the two oxygen atoms are from
the protonated form of the carboxylic ligand moieties, with
an oxygen atom from a solvent molecule (this topology was
based on a previously reported N2Py2 complex).[8c,13] The
Fe–N distances range from 2.161(3) Å, which are associated
with the pyridine nitrogen atoms, to 2.308(3) Å, which are
associated with the amino nitrogen atoms (Table 1). These
values are different to those reported for [FeII(BPMCN)]-
(CF3SO3)2 or [FeII(5-Me-BPMCN)](CF3SO3)2, as longer
Fe–Namine bonds were measured in [FeII(L4)(C3H6O)]-
atoms are omitted for clarity.
Table 1. Bond lengths [Å] and angles [°] for [FeII(L4)(C3H6O)]-
(ClO4)2.
Fe1–N1
2.308(3)
2.162(3)
2.292(2)
2.172(2)
Fe1–N2
Fe1–N4
Fe1–O20
2.300(3)
2.163(3)
2.288(2)
Fe1–N3
Fe1–O10
Fe1–O30
N1–Fe1–N2
N1–Fe1–N4
N2–Fe1–N4
N1–Fe1–O10
N1–Fe1–O30
N2–Fe1–O20
N3–Fe1–O10
N3–Fe1–O30
N4–Fe1–O20
O10–Fe1–O20
O20–Fe1–O30
76.96(10)
106.94(10)
76.50(11)
70.27(10)
142.95(10)
70.80(9)
99.00(11)
92.27(10)
93.05(10)
152.23(8)
74.90(9)
N1–Fe1–N3
N2–Fe1–N3
N3–Fe1–N4
N1–Fe1–O20
N2–Fe1–O10
N2–Fe1–O30
N3–Fe1–O20
N4–Fe1–O10
N4–Fe1–O30
75.64(11)
104.47(11)
177.41(11)
136.65(10)
133.05(10)
140.01(10)
85.06(10)
81.87(11)
85.53(11)
O10–Fe1–O30 77.48(9)
Solid-State IR Spectroscopy
All the IR spectra of the complexes displayed character-
(ClO4)2, although the variation between the different Fe–N istic features in the 1400–1750 cm–1 region (Figure 4). In
bonds (pyridine or amine functionalities) followed the same the case of [Fe(L4)(C3H6O)](ClO4)2, three resonances were
trend.[8c] The lower symmetry of [FeII(L4)(C3H6O)](ClO4)2 observed, two narrow ones at 1676 and 1607 cm–1 that
and the protonation of the ligand may be responsible for flank a broader one around 1650 cm–1 (Figure 4, A). The
the longer bonds. The Fe–O bonds are mostly long (around low energy peak is attributed to the asymmetric C=O vi-
2.39 Å) with the exception of that involving the acetone li- brations of the bound acetone by comparison with litera-
gand [2.172(2) Å]. These values indicate that the carboxylic ture values,[18] that at 1607 cm–1 to νC=N of the pyridine
Eur. J. Inorg. Chem. 2012, 85–96
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
87

S. Ménage et al.
FULL PAPER
rings, whereas the broad band corresponds to νCOOH of (C3H6O)](FeCl4) and [FeII(L1)(C3H6O)](FeCl4), a peak at
as
the bound carboxylic groups in the complex. This energy m/z = 160.8 was detected in the negative mode, which was
value is between that of free νCOOH (above 1750 cm–1) and attributed to [FeIICl3]–, a footprint of the unstable [FeII-
coordinated νCOO– (1700–1510 cm–1).[19] In addition, νCOOH
Cl4]2– fragment. This was replaced in the spectrum of
s
is located at 1437 cm–1, which implies a Δ(νs–νas) of [FeIII(L1)(Cl)](FeCl4) with a peak at m/z = 197.2, which cor-
213 cm–1, a value characteristic of monodentate binding responds to [FeIIICl4]–. No evidence of a Cl6Fe2O fragment
modes. In the case of [FeII(L2)](ClO4)2, the same νCOOH res- (deduced from the elemental analysis) was found.
onance was observed at 1635 cm–1 but no narrow peak was
UV/Vis Spectroscopy in Acetone
detected in this region, which indicates the absence of ace-
Two patterns of transitions were observed in the elec-
tronic spectra recorded under argon. First, only one transi-
tion in the near UV was observed for the perchlorate com-
plexes, at 342 nm for [FeII(L4)(C3H6O)](ClO4)2 and 310 nm
for [FeII(L2)](ClO4)2 with similar extinction coefficients in
the 1500–4000 cm–1 m–1 range, which is attributed to iron-
to-nitrogen metal-to-ligand charge transfer.[20] Second, the
UV/Vis spectra of [FeII(L4)(Cl)]2(FeCl4), [FeII(L3)(C3H6O)]-
(FeCl4), and [FeII(L1)(C3H6O)](FeCl4) are dominated by
two transitions at 360 and 310 nm with extinction coeffi-
cients of around 2000 cm–1 m–1, which are absent in the
tone as a ligand.
–
spectra of their ClO4 counterparts. Interestingly, these
bands were also present but more intense in the spectrum
of [FeIII(L1)(Cl)](FeCl4), which suggests that they originate
from the counteranion (Figure 5). The energy values of
these transitions were found to be identical to those found
–
in literature for the FeIIICl4 anion.[21] However, the extinc-
tion coefficients are lower in our case, which indicates that
the ferrous anion (FeCl4) was partly oxidized during the
measurements.
Figure 4. IR spectra of the iron complexes (solid lines for ligands
with a cyclohexane moiety and dashed for ligands with an ethane
moiety): A for [FeII(L4)(C3H6O)](ClO4)2 and [FeII(L2)](ClO4)2; B
for [FeII(L4)(Cl)]2(FeCl4);
C
for [FeII(L3)(C3H6O)](FeCl4) and
[FeII(L1)(C3H6O)](FeCl4); D for [FeIII(L3)(Cl)] (FeIIICl4).
In the spectrum of [FeII(L4)(Cl)]2(FeCl4), two νCOOH
bands were observed, one at 1732 cm–1, which is character-
istic of a free carboxylic acid, and the other broad one
around 1640 cm–1, which is common to these species (Fig-
ure 4, B). The latter component was also observed in the
spectra of [FeII(L3)(C3H6O)](FeCl4) and [FeII(L1)(C3H6O)]-
(FeCl4) (Figure 4, C). The change in oxidation state of the
metal in [FeIII(L3)](FeCl4) caused the slight shift of this
transition to 1690 cm–1, which indicates a different coordi-
nation mode or a different protonation state of the carbox-
ylic/carboxylate moiety (Figure 4, D) (see below).
ESI-MS in Acetone
Figure 5. UV/Vis spectra of acetonitrile solutions of [FeII(L2)]-
(ClO4)2 (– - –), [FeII(L4)(C3H6O)](ClO4)2 (– – –), [FeII(L4)(Cl)]2-
(FeIICl4) (· · ·), and [FeII(BPMCN)](FeIICl4) (- - -).
All the MS of the deoxygenated solutions of the com-
plexes displayed a [FeIIL – H]+ fragment in the positive
mode, which confirms the mononuclearity of the complex
in solution. For [FeII(L4)(C3H6O)](ClO4)2, a dicationic
This outcome was confirmed by oxygenation of the ace-
[FeL]2+ fragment was observed as well as a fragment that tonitrile solutions of the ferrous complexes (Figure 6). We
contains the ClO4 counteranion, which suggests that the observed the increase in intensity of the two transitions at
carboxylic functionalities of the ligand were totally proton- 360 and 310 nm for all the chlorinated complexes to the
ated in the starting solution. In the spectra of [FeII(L3)- level of the exact extinction coefficient of the FeIIICl4 anion
88
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 85–96

Fe Coordination Chemistry
itself. Consequently, the formation of the ferric anion arises
The 1H NMR spectrum of [FeII(L4)(Cl)]2(FeCl4) was dif-
–
from the oxidation of the FeIICl4, which does not have vis- ferent to that of its ClO4 congener (Figure 7). The reso-
ible transitions.[21–22] The absorbance measured at 360 and nance pattern looked similar, except the loss of low-field
310 nm is in accordance with a ratio of 1:1 between the iron resonances above 50 ppm, which was attributed to a change
cation and its counteranion.
in ligand topology as o-H pyridine proton resonances are
the most affected by the relative positions of the pyridine
rings.[13] We propose that the ligand conformation was cis
β, which leads to distinct pyridine positions in the complex
structure. This leads to each o-H resonance becoming dis-
tinct and merging with the baseline. Moreover, the spec-
trum looked similar to that of the ferrous BPMCN com-
plex, which was prepared by adding FeCl2 to BPMCN in
acetone. Despite the absence of X-ray structural analysis,
the topology of the BPMCN complex can be deduced from
an anion metathesis experiment to exchange the Cl– ion by
triflate (OTf). It has been reported that BPMCN isomers
do not interconvert, which allows the structural characteri-
zation of both cis-α- and cis-β-[FeII(BPMCN)](OTf)2.[13,24]
Interestingly, their syntheses involved the metathesis reac-
tion pathway with its chlorinated analog in CD3CN.[13]
Using the same procedure, the 1H NMR spectroscopic
pattern observed after metathesis was identical to that of β-
[FeII(BPMCN)(CD3CN)2]2+, which indicates a cis β top-
ology for the BPMCN complex (Figure S3).[13,24]
1
The H NMR spectrum of [FeII(L3)(C3H6O)](FeCl4) has
Figure 6. UV/Vis spectra of acetonitrile solutions of
[Fe(BPMCN)](FeCl4) Ϯ O2 (– – –); [Fe(L1)(C3H6O)](FeCl4) Ϯ O2
(· · ·); [Fe(L4)(Cl)]2(FeCl4) (- - -) Ϯ O2. Arrows indicate the absorp-
tion changes in the presence of dioxygen.
a complex pattern of resonances spanning between 160 and
–10 ppm (Figure 7), which was attributed to the superposi-
tion of two close but distinct spectra with at least 26 proton
resonances for each. It is tempting to relate this signature
to the coexistence of two related topologies of the ligand.
As both substituents on the amino nitrogen atoms of
1H NMR Spectroscopy in [D6]Acetone
1H NMR spectroscopy has been shown to be a powerful the ligand are different (Me versus carboxymethyl), two
tool for probing the structural and magnetic properties of topologies in a similar proportion are generated only for
iron complexes.[17,23] All the spectra (Figure 7) spanned be- the asymmetric geometry, i.e. cis β (opposite to trans and
tween 170 to –40 ppm, which attests that all the complexes cis α). In the case of L1, the spectrum was also very com-
contain high-spin ferrous species in solution. As the reso- plex, which indicated that several isomers were present. In
nance pattern of the NMR spectra of the complexes is remi- this case, the broadness of some peaks precluded an exten-
niscent of the ligand conformation, it was possible to pro- sive interpretation of the pattern. At least we can suggest
pose a structural topology of the ligand for most of them. the presence of cis α and β conformations in nonequivalent
First, the spectra of [FeII(L4)(C3H6O)](ClO4)2 and its ana- proportions as the trans topology has not been observed for
log [FeII(L2)](ClO4)2 consist of broad resonances above this type of ligand.
70 ppm, which are assigned to o-H of the pyridine rings
Based on a comparison of the solution and solid-state
and the methylene proton resonances (Figure 7). The two spectroscopic studies of all the complexes, we propose the
narrow resonances that flank a broader one in the 40– structures of the complexes with the new ligands as de-
60 ppm region, are attributed to m-H of the pyridine rings picted in Figure 8.
and the methylene protons. The upfield resonance at
–12 ppm is tentatively assigned to p-protons by comparison
with earlier studies.[20] The difference in the pattern of the
two spectra in the 0–110 ppm region is attributed to the
Catalytic Studies for Alkene Oxidations
presence of the extra methylene protons of the strap be-
The complexes were tested for the catalytic oxidation of
several different alkenes in acetonitrile solution with H2O2
tween the two nitrogen atoms in L2 compared to L4.
The o-protons are all equivalent and show a resonance as the oxidant with a catalyst:substrate:H2O2 ratio of
at 130 ppm, which indicates that the complexes have a high 1:200:300. The catalytic properties of the new complexes
symmetry. This is in agreement with the conservation of the were compared to their BPMCN or BPMEN analogs and
cis α topology in solution. The presence of bound acetone evaluated as a function of acetic acid or perchloric acid ad-
could not unambiguously be attributed in the spectrum of dition. The catalysis experiments performed in acidic condi-
[FeII(L4)(C3H6O)](ClO4)2, which indicates its facile ex- tions have been the subject of extensive studies as they de-
change with a solvent molecule.
termine the high selectivity for epoxides.[8d,8g]
Eur. J. Inorg. Chem. 2012, 85–96
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
89

S. Ménage et al.
FULL PAPER
1
Figure 7. H NMR spectra of iron complexes in [D6]acetone.
Figure 8. Proposed structures for the new iron complexes (S stands for acetone or solvent). A mixture of the cis α and cis β forms coexist
in [FeII(L1)(C3H6O)](FeCl4).
The experiments were performed at room temperature in end of the oxidant addition. The presence of air did not
air. The oxidant was delivered by dropwise addition over affect the yield or selectivity of the epoxidation reaction. As
30 min using a syringe pump in order to reduce H2O2 dis- most complexes contain either a chloride ligand and/or an
proportionation. The yield was measured 30 min after the Fe(II)Cl4 or Fe(III)Cl4 counteranion, five equivalents of
90
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 85–96

Fe Coordination Chemistry
Table 2. Cyclooctene oxidation catalyzed by [FeL](FeCl4) complexes under various conditions with 0.5 mol-% catalyst.[a]
Catalyst CH3CN[a]
[FeL]
L
+ 100 equiv.
CH3CN/CH3CO2H
1:2
+ 1 equiv. HClO4
CH3CO2H
Epoxide[b] [%] Conv. [%]
Epoxide[b] [%] Conv. [%]
Epoxide[b] [%] Conv. [%]
Epoxide[b] [%] Conv. [%]
None
FeCl2
0
3
4
1
1
20
18
98
97
37
98
95
1
1
–
20
–
–
–
–
–
22
68
68
37
92
78
1
BPMEN 32
95
82
3
64
59
1
92
31
4
63
80
4
98
51
37
100
99
16
44
29
–
38
43
–
85
66
–
90
91
–
L1
27
3
L2
BPMCN 56
L3
L4
37
1
[a] Ratio catalyst:substrate:H2O2 = 1:200:300, in CH3CN, oxidant delivered by a syringe pump over 30 min and 30 extra minutes of
stirring were allowed before GC injection. [b] Yield based on substrate consumption.
Ag(CF3SO3) were added prior to the addition of the oxi-
dant, which accounts for the removal of these chloride
anions from the solution as they may compete with the oxi-
dant to bind to the metal ion. This experimental require-
ment was supported by the total loss of epoxide formation
during catalysis with these complexes in the absence of sil-
ver salts. Furthermore, [FeIIICl4]–, FeIIICl3, or FeIII(ClO4)3
were found to be inactive for epoxide formation under stan-
dard conditions, although a small conversion yield was
measured. These control experiments guaranteed that the
results obtained are attributed to the starting ferrous FeLn
dication.
In the absence of either the catalyst or the oxidant, no
epoxide products were detected, although the substrate was
partially converted in some cases. This is particularly true
when acetic acid was added to the reaction medium; up to
20% conversion was measured when 100 equiv. were added
(Table 2, Column 4).
In the absence of acetic acid, [FeII(L4)(C3H6O)](ClO4)2
and [FeII(L2)](ClO4)2 were found to be quasi-inactive for
cyclooctene oxidation. Conversely, reasonable epoxide
Figure 9. Selectivity for cyclooctene epoxide depending on the li-
yields were reached with [FeII(L3)(C3H6O)](FeCl4) and
gand of the iron(II) catalyst (black), in the presence of 100 equiv.
[FeII(L1)(C3H6O)](FeCl4) (27 and 37% yield, respectively) acetic acid (grey), in a CH3CN/CH3CO2H (1:2) solution (light
grey), and in the presence of 1 equiv. of H+ (white).
under the same conditions (Table 2 and Figure 9). For com-
parison, the BPMEN and BPMCN analogs were found to
be more efficient (32[13] and 56% yield, respectively). Under [FeII(BPMEN)](FeCl4) showed a stimulation of its reactiv-
our conditions, the epoxide selectivity reached 60%, al- ity (37 to 98%) with a 100% selectivity (Figure 9).[10] These
though only a small amount of the corresponding diols results underline a difference in catalytic behavior that is
were detected (Figure 9). As the conversion ranged from related to the nature of the diamine (cyclohexanediamine
68–92% depending on the catalyst, other secondary prod- vs. ethylenediamine).
ucts were formed. 1H NMR spectroscopic studies of the
The addition of 1 equiv. of HClO4 was totally ineffective
reaction media confirmed the presence of cyclooctene oxide in stimulating the epoxidation catalysis, which implicates
and trace amounts of diols (less than 4%) as the major that protons alone are not sufficient to account for the en-
products but no other oxygenated products, such as octane- hancement of the reaction by acetic acid. This result sug-
dioic acid, were detected.
gests that acetic acid has to bind into the oxidizing species,
The presence of acetic acid increased the efficiency of the which has been proposed earlier.[10]
catalysts that contain L1 and L3, whereas those with L2 and
The complex that contains a chiral substructure, i.e.
L4 remained unreactive. For example, the epoxide and con- (S,S)-cyclohexanediamine, performed
enantioselective
version yields increased significantly (from 37 to 80% and epoxidations, and the results are shown in Table 3. Interest-
78 to 100%, respectively) as acetic acid concentration in- ingly, the enantioselectivity of the reaction depended on the
creased and reached 81% selectivity with [FeII(L3)- presence of acetic acid in the reaction medium; a large in-
(C3H6O)](FeCl4). By comparison, [FeII(BPMCN)](FeCl4) crease from 15 to 34% was observed in the case of enantio-
was unaffected under these conditions. Interestingly, meric excess for the H2O2 oxidation of the trans-2-heptene
Eur. J. Inorg. Chem. 2012, 85–96
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
91

S. Ménage et al.
FULL PAPER
Table 3. Asymmetric alkene oxidation catalyzed by [FeIIL](FeCl4) with 0.5 mol-% catalyst.[a]
Catalyst
([FeL])
L
+ 100 equiv.
CH3CO2H
Epoxide [%]
1:2
CH3CN[a]
Epoxide [%]
CH3CN/CH3CO2H
Epoxide [%]
ee [%]
ee [%]
ee [%]
trans-2-Heptene
None
1
0
–
–
1
0
BPMCN
31
18
14
17
52
29
38
30
42
40
30
34
L3
trans-β-Methylstyrene
None
3
7
0
0
3
6
0
0
L3
cis-β-Methylstyrene
None
0
13
0
13
0
21
0
18
L3
(R)-Carvone
None
0
17
0
0
27
0
L3
5[b]
15[b]
[a] Ratio catalyst:substrate:H2O2 = 1:200:300 in CH3CN, oxidant delivered by a syringe pump over 30 min and 30 extra minutes of
stirring were allowed before GC injection. Yield based on substrate. [b] A diastereomeric excess was measured by integration of the
1
relevant H NMR spectroscopic signals.
when 100 equiv. of acid was present. The presence of the which could result from Markovnikov addition of hydroxyl
carboxylic moiety on the ligand did not impact on the radicals to the double bond.[26]
enantioselectivity of this reaction as similar values were ob-
served with [FeII(L3)](FeCl4) and [FeII(BPMCN)](FeCl4). It
may suggest that the carboxylic ligand does not bind the
active catalyst or does not bind in a position that affords
Discussion
H-bonding with the peroxide adduct. This role is devoted
to the exogenous acid.
The design of complexes with carboxylic ligands was suc-
cessful, thanks to the use of ferrous salts and acetone as
The electron richness of the substrate drastically affects the solvent. The new mononuclear complexes present an
the enantiomeric excess; it was found that the chiral com- original property of a monodentate coordination mode of
plex was a poor catalyst for the oxidation of cis- or trans- a carboxylic functionality of the ligand, which was proven
β-methylstyrene but afforded retention of the configuration. by IR spectroscopy (a broad peak at 1640 cm–1) for all the
Low but remarkable ee values for complexes with the N2Py2 complexes and confirmed by the X-ray structure of [FeII-
ligand platform were measured,[8c] which are affected by the (L4)(C3H6O)](ClO4)2. The state of protonation of the car-
presence of acetic acid. Furthermore, discrimination be- boxylic group leads to a Fe–O bond, which implies the pres-
tween both diastereoisomers of the alkene by the catalyst ence of an sp2 oxygen atom. Another characteristic of the
was observed in terms of ee and yield. Although a racemic complexes was the unexpected formation of the metallic
–
mixture of the corresponding epoxide was found with trans- anion FeIICl4 , which was indirectly detected by ESI-MS
(E)-β-methylstyrene, a noticeable enantiomeric excess was and UV/Vis spectroscopy.
detected when cis-(Z)-β-methylstyrene was used as a sub-
The geometries of the six original complexes were found
strate. Finally, benzaldehyde, propiophenone, and traces of to be anion dependent in acetone. The presence of a nonco-
diol were detected in less than 5% total yield, which means ordinating anion leads to the formation of cis α complexes,
a poor selectivity for the epoxide. However, even if styrenes in which the pyridine rings are trans to each other and the
undergo competing aromatic hydroxylation reactions, carboxylic moieties are cis to each other (Figure 8). How-
which is well documented for Fe(BPMEN) and Fe(TPA),[25] ever, with a chloride anion, the major topology, deduced
such hydroxylated species were not detected here.
from NMR spectroscopic studies, seems to be cis β what-
The selective epoxidation of (R)-carvone was catalyzed ever the ligand used. We propose that the chloride ions first
by [FeII(L3)(C3H6O)](FeCl4), which proceeded by exclusive bind to FeL, which induces a preferred topology, before
attack on the electron-rich, external C=C bond. This indi- being exchanged by the carboxylic arm combined with the
cates that the oxidizing species has an electrophilic charac- presence of extra coordination sites in FeCl2 ([FeCl4]2– is
ter. Moreover, the presence of acetic acid in the medium did thermodynamically more stable than the expected FeLCl2
not enhance greatly the yield of epoxide (17 versus 27% in complex). The geometry is also related to the lower acidity
the presence of 100 equiv. of acid) but affected the dia- of carboxylic acids compared to HCl, which prevents de-
stereomeric excess (from 5 to 15% with 100 equiv. of acetic protonation of the carboxylic acid during the synthesis.
acid). Moreover, the selectivity for the epoxide was only of Such a phenomenon has been observed previously.[27] In
62% as two other oxygenated products were detected, one one report, the degree of protonation of the ligand influ-
of which was definitively identified as 2-(4-methyl-5-oxocy- enced the synthesis of the tetrachloride ferrous anion.
clohex-3-en-1-yl)propanal (see Supporting Information).
In terms of complex geometry, the propensity for this cis
The other could be the alcohol analog of the latter species, β topology remains puzzling but seems to be driven by the
92
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 85–96

Fe Coordination Chemistry
constraints imposed by the diamine part of the ligand. A rigidity of the ligand may exclude the formation of H-bond-
mixture of isomers was found with a more flexible diamine ing between the hydroperoxide and the endogenous carbox-
in an analogous complex.[24] Moreover, the cyclohexanedi- ylic ligand. Moreover, this substitution process depends on
amine guarantees the topology in solution as it is more con- the strength of the Fe–Ocarb bond. The reorganization of
strained than ethylenediamine.
the coordination sphere during the process is probably more
In [FeII(L3)(C3H6O)](FeCl4), the ligand in the sixth coor- difficult for more rigid cyclohexanediamine-based com-
dination position could not be unambiguously determined plexes compared to the ethanediamine series, which gener-
but Cl– anions are excluded. We were able to compare its ates a lower stimulating effect (Figure 9). We proposed that
O2-oxidized form with [FeIII(L3)(Cl)](FeCl4), which was the easier it is for the acetic acid to bind, the higher the
synthesized separately. The latter complex was charac- reactivity.
terized as hexacoordinate, in which one chloride anion com-
The catalyst that incorporates the optically active trans-
pleted the iron coordination sphere in addition to the an- 1,2-diaminocyclohexane backbone into its ligand frame-
ionic form of the carboxylic group. Although both com- work, i.e. complexes of L3, catalyzed the asymmetric epox-
plexes displayed similar UV spectroscopic signatures that idation of trans-2-heptene with enhanced ee values in the
arise from FeIIICl4 , their H NMR spectra (which were in presence of an external acid. Again, the presence of this
agreement with mononuclear species) were different (Figure intramolecular carboxylic group was not deleterious for the
S5). This behavior implies a different environment around enantioselective control as similar ee values were obtained
the ferric atom, which suggests that the Cl ligand is absent with [FeII(BPMCN)](FeCl4) and [FeII(L3)(C3H6O)](FeCl4).
in the ferrous complex. We propose that the solvent, in The similar value obtained in the presence of acetic acid
solution, or an acetone molecule, in the solid state, is coor- confirms that the intrinsic carboxylic group was not in-
–
1
dinated in the complex.
volved in the catalytic process (Scheme 2). The existence of
The catalytic properties of the original complexes were an asymmetric control emphasizes the existence of a metal-
then compared to those of the BPMEN and BPMCN com- based mechanism. The formation of the oxidizing species
plexes. Initially, the use of silver salt allows the exchange of and its structure is controlled by the presence of external
a chloride anion with acetonitrile in order to be under the acetic acid. It still remains to be established if a high-valent
previously reported catalytic conditions for the BPMEN iron species[10] is involved in the process. Further studies
and BPMCN complexes. The general outcome under these will be undertaken to shed light on the nature of this species
conditions is that the substitution of a methyl group by a in order to gain insight into the oxidation mechanism.
carboxylic moiety on the ligand does not alter the capacity Among the results obtained with mono- or dimeric iron
of the N2Py2 iron catalyst. However, two substitutions of catalysts for the stereoselective epoxidation of various al-
the methyl groups by carboxylic moieties in the ligand tot- kenes, this is the first time that a noticeable enantiomeric
ally poisoned the iron catalyst, which indicates that i) these excess has been obtained with a nonheme iron mononuclear
moieties still bind to the oxidizing metal species or oxidant catalyst and a nonelectron-rich substrate.[30]
activating species and ii) they hamper the ability of the oxi-
dant to bind. Similar results were obtained with catalysts
that contain dicarboxylic ligands[4b] and, to some extent,
with a system that contains one bidentate carboxylate li-
gand.[15] The existence of one labile site is required for cata-
lytic activity and accounts partly for the selectivity of the
alkene oxidation.[5] The high selectivity for epoxide over
diol is typical of pentadentate iron catalysts, as two direct
Scheme 2. Possible activation of hydrogen peroxide by the ferrous
oxygen atom transfers are precluded.[28] The catalytic stimu-
complex of L1 (and/or L3).
lation by acetic acid is more complicated to assess. In all
cases, it had a positive effect on the conversion, the selectiv-
ity, and the epoxide yields, with the exception of the com-
plex with BPMCN. Acetic acid has been proposed to assist
the heterolytic cleavage of a transient hydroperoxo ferric in-
Conclusions
termediate, which is generally accepted as the key feature
A new N4O environment for an iron complex derived
to generate the active oxidant in the cytochrome P450 from N2Py2 ligands (BPMEN or BPMCN) affords a good
mechanism.[29] In our case, proton assistance should be efficiency for epoxidation, thanks to the protonation of its
driven by the carboxylic moieties and/or acetic acid, which carboxylic group, whereas a more oxygenated N4O2 envi-
displaces the carboxylic moiety. The absence of an effect of ronment drastically affects the catalytic activity. The pres-
protons alone (HClO4) suggests that the acetic acid has to ence of an available, internal carboxylic functionality in the
bind to the complex. The observed stimulation of the catal- complex slightly competes with the coordination of external
ysis was modulated by the exchange between a carboxylic acetic acid, which is known to stimulate epoxidation cataly-
ligand and exogenous acetic acid in the iron coordination sis observed with N2Py2 catalysts such as [Fe(BPMCN)2-
sphere and also depended on the orientation of carboxylic (CH3CN)2](CF3SO3)2. Interestingly, the stimulating effect
ligand towards the hydroperoxide ligand. Accordingly, the of acetic acid combined with the weak effect of the carbox-
Eur. J. Inorg. Chem. 2012, 85–96
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
93

S. Ménage et al.
FULL PAPER
(1S,2S)-N-Methyl-N,NЈ-bis(2-pyridylmethyl)-1,2-cyclohexanedi-
amine (“BPHMCN”): Synthesized by an adaptation of a previously
published procedure.[31] (1S,2S)-N,NЈ-Bis(2-pyridylmethyl)-1,2-
cyclohexanediamine (1.6 g, 5.39 mmol) was dissolved in aqueous
formic acid (50 mL, 90 %). Formaldehyde (9.92 mL, 25 equiv.,
37%) was added and the resulting solution was stirred at 90 °C for
24 h. The mixture was cooled to 20 °C and the pH was adjusted to
12 by addition of aqueous sodium hydroxide (3 m) with constant
cooling. The aqueous layer was extracted into dichloromethane,
and the combined organic layers were dried with anhydrous
Na2SO4, filtered, and evaporated to give a brown oil. Purification
on a column of neutral aluminium oxide using a pentane/ethyl acet-
ate/triethylamine mixture (50:50:1 v/v) as eluent afforded the title
product (1.412 g, 85%) as a clear oil in addition to dimethylated
ylic moiety ligands was also observed for the first time for
the enantioselective reaction with trans-2-heptene.
The presence of carboxylic functionalities may be toler-
ated for the design of bioinspired catalysts.
Experimental Section
General: All reagents were purchased from commercial sources and
were used as received unless noted otherwise. Solvents were dried
and degassed before use.
Crystallographic Studies: Data collection of [FeII(L4)(C3H6O)]-
(ClO4)2 were conducted with a Bruker SMART CCD system at the
crystallography service of the SCIB laboratory (CEA-Grenoble).
Experimental conditions for [FeII(L4)(C3H6O)](ClO4)2: (C25H34Cl2-
FeN4O13) M = 725.31 gmol–1, yellow plate 0.26ϫ0.17ϫ0.13 mm,
orthorhombic, collection wavelength 0.71073 Å; collection tem-
perature 150(2) K, space group P212121, a = 12.1116(4) Å, b =
15.7961(6) Å, c = 16.4185(6) Å, α = 90.0°, β = 90.0°, γ = 90.0°, V
= 3141.11(19) Å3, Z = 4, Dcalc = 1.534 gcm–3; Reflections collected
= 10236 and 6361 independent reflections collected; R1 = 0.0412
[IՅ2σ(I)]. Pertinent crystallographic data are reported in Table 1.
The structure was solved by direct methods using the SHELXTL
5.03 software package. All non-hydrogen atoms were refined aniso-
tropically. Hydrogen atoms were placed in ideal positions and re-
fined as riding atoms with individual (or group) isotropic displace-
ments.
1
BPMCN (15%). H NMR (300 MHz, CDCl3): BPMCN: δ = 8.55
(d, J = 4.8 Hz, 2 H, Py), 7.63 (m, 4 H, Py), 7.17 (m, 2 H, Py), 3.98
and 3.86 (d, J = 14.7 Hz, 2ϫ 2 H, CH2), 2.71 (m, 2 H), 2.33 (s, 6
H, 2ϫCH3), 2.02 (m, 2 H), 1.81 (m, 2 H), 1.26 (m, 4 H) ppm.
BPHMCN: δ = 8.43 (d, J = 4.2 Hz, 2 H, Py), 7.56 (m, 3 H, Py),
7.23 (m, 1 H, Py), 7.07 (m, 2 H, Py), 3.97 (d, J = 14.1 Hz, 1 H,
CH2), 3.78 and 3.73 (ddd, J = 4.5, J = 4.8 Hz, 2 H, CH2), 3.54 (d,
J = 14.4 Hz, 1 H, CH2), 2.43 (m, 2 H), 2.24 (m, 1 H), 2.15 (s, 3 H,
CH3), 1.89 (m, 1 H), 1.76 (m, 1 H), 1.67 (m, 1 H), 1.17 (m, 5 H)
ppm.
(1S,2S)-N-Carboxymethyl-NЈ-methyl-N,NЈ-bis(2-pyridylmethyl)-
1,2-cyclohexanediamine (L3) (General Procedure): BPHMCN
(1.006 g, 3.26 mmol) was dissolved in a mixture of anhydrous
MeCN (20 mL) and K2CO3 (720 mg, 5.22 mmol) under an inert
atmosphere. tert-Butyl 2-bromoacetate (723 μL, 4.89 mmol) in an-
hydrous MeCN (20 mL) was added dropwise with a syringe over
about 5 min. The reaction was then left to stir at room temperature
and was monitored by TLC. Once all of the starting material had
disappeared (ca. 3 h) H2O (approx. 20 mL) was added, and the
reaction mixture was extracted into dichloromethane three times.
The organic phase was washed twice with brine. The combined
aqueous phases were re-extracted into CH2Cl2. All the CH2Cl2
fractions were dried with Na2SO4 for 30 min. The drying agent was
removed by filtration through a sintered funnel and the solvent
was evaporated under reduced pressure. (Note: Treatment of the
combined aqueous phases with NaOH followed by CH2Cl2 extrac-
tion led to the recovery of the starting amine with an average yield
of 50%.) The crude product was then purified on a column of
neutral aluminium oxide (pentane/ethyl acetate/triethylamine,
CCDC-809749 contains the supplementary 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.
1
Physical Studies: H and 13C NMR spectra were recorded with a
Bruker Avance 300 MHz spectrometer. 1H NMR chemical shifts
are reported in ppm with the solvent as the internal reference. ESI-
MS studies were performed with an LXQ-linear ion trap (Thermo
Scientific, San Jose, CA, USA) equipped with an electrospray
source in an aqueous or aqueous/acetone mixture. Electrospray full
scan spectra in the range m/z = 50–2000 were obtained by infusion
through fused silica tubing at 2–10 μLmin–1. The LXQ calibration
was achieved according to the standard calibration procedure from
the manufacturer (mixture of caffeine/MRFA and Ultramark
1621). The temperature of the heated capillary of the LXQ was set
to 150–200 °C, and the ion spray voltage was in the range 1–3 kV.
The experimental isotopic profile was compared in each case to the
theoretical one. UV/Vis spectroscopic studies were carried out with
a Shimadzu UV-1800 spectrophotometer. IR spectra of the com-
plexes as KBr discs were measured with a Perkin–Elmer Spectrum
100 FT-IR spectrometer. GC was performed with a Perkin–Elmer
Autosystem XL instrument with a FID detector, which used an SE
30 column coupled to a Perkin–Elmer Turbomass EI spectrometer.
Enantiomeric excesses were measured by GC with a Chiraldex TA
column. The values are the average of three separate samples. The
1
50:50:1) (732 mg, 53%). H NMR (75 MHz, CDCl3): δ = 8.43 (m,
2 H, Py), 7.77 (d, J = 7.8 Hz, 2 H, Py), 7.47–7.56 (m, 2 H, Py),
7.04–7.10 (m, 2 H, Py), 3.98 (AB, J = 12 Hz, J = 60 Hz, 2 H,
CH2CO), 3.71 (AB, J = 14.4 Hz, J = 32.4 Hz, 2 H, CH2), 3.37 (AB,
J = 16.5 Hz, J = 44.7 Hz, 2 H, CH2), 2.6 (br., 2 H, Cy), 2.15 (s, 3
H, Me), 1.70 (br., 2 H, Cy), 1.39 (s, 9 H, tBu), 1.10–1.16 (m, 4 H,
Cy) ppm.
The resulting oil (732 mg, 1.73 mmol) was dissolved in CH2Cl2
(3 mL) under an inert atmosphere. Trifluoroacetic acid (2.53 mL,
20 equiv.) was then added slowly at 0 °C. The reaction mixture was
stirred at 0 °C for 30 min. The reaction was allowed to warm to
room temperature and left to stir overnight. The volume of the
resulting mixture was reduced to about 4/5 under reduced pressure.
The solution was then added dropwise into diethyl ether (30 mL)
to precipitate the product. The resulting yellow paste was triturated
several times to remove all of the trifluoroacetic acid. The remain-
1
diastereoisomeric excess for carvone epoxide was measured by H
NMR spectroscopy and was based on the average of the splitting
of the epoxide peak at 2.68 ppm and the methyl peak at 1.30 ppm.
Elemental analyses were performed by the Microanalysis Service
of CNRS (Vernaison).
Synthesis of Ligands
ing ether was then evaporated and left under vacuum for 5 h. The
1
N-Methyl-N,NЈ-bis(2-pyridylmethyl)-1,2-ethanediamine reaction was quantitative (636 mg). H NMR (300 MHz, [D6]acet-
(“BPHMEN”): Synthesized according to the method reported by
one): δ = 8.65 (br., 2 H, Py), 7.92 (br., 2 H, Py), 7.61 (br., 2 H, Py),
7.46 (br., 2 H, Py), 4.89 (br., 2 H, CH2), 4.37 (br., 2 H, CH2), 3.53
Baffert et al.[12]
94
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 85–96

Fe Coordination Chemistry
(m, 2 H, CH2CO), 3.01 (br., 2 H, Cy), 2.42 (m, 2 H, Cy), 1.90 (m,
2 H, Cy), 1.38 (m, 4 H, Cy) ppm.
(νC=O ) cm–1. ESI-MS (acetone): m/z = 467.4 (100) [L – H +
FeII]+,s 933.4 (83) [2(L – H) – H + 2FeII]+, 955.4 (34) [2(L – 2H) +
Na + 2FeII]+. [FeII(L4)(Cl)]2(FeIICl4): C44H56Cl6Fe3N8O (1093.24):
calcd. C 43.85, H 4.68, N 9.30; found C 43.62, H 4.97, N 9.15.
N-Carboxymethyl-NЈ-methyl-N,NЈ-bis(2-pyridylmethyl)-1,2-
ethanediamine: L1 was synthesized using the same procedure from
BPHMEN. Yield for the two steps: 50 %. 13C NMR (75 MHz,
MeOD): δ = 172.8 (CO2H), 157.8 (Py), 149.5 (Py), 147.9 (Py), 138.7
(Py), 137.7 (Py), 125.0 (Py), 124.4 (Py), 65.5 (CH2Py), 58.8
(CH2Py), 57.6 (CH2CO), 54.2 (CH2), 49.3 (CH2), 40.1 (Me) ppm.
[FeII(L4)(C3H6O)](ClO4)2: In a glove box, L4 (105.3 mg,
0.26 mmol) was dissolved in acetone (approx. 10 mL). FeII(ClO4)2
(65.1 mg, 0.26 mmol) was added, and the mixture was stirred for
30 min. Addition of diethyl ether (5 mL) to the yellow-green solu-
tion led to the precipitation of the complex The product was
washed with diethyl ether and dried under vacuum to afford a yel-
N,NЈ-Dicarboxymethyl-N,NЈ-bis(2-pyridylmethyl)-1,2-ethanedi-
amine: L2 was synthesized using the same procedure from N,NЈ-
bis(2-pyridylmethyl)-1,2-ethanediamine with 2.2 equiv. of tert-butyl
2-bromoacetate. Yield for the two steps: 41%. 13C NMR (75 MHz,
[D6]acetone): δ = 172.1 (CO2H), 152.3 (Py), 145.1 (Py), 142.8 (Py),
126.1 (Py), 125.3 (Py), 56.1 (CH2Py or CH2CO2H), 54.2 (CH2Py
or CH2CO2H), 51.1 (CH2) ppm.
1
lowish solid (43%). H NMR (300 MHz, [D6]Acetone): δ = 127.9
(Ho), 97.9 and 88.7 (CH2), 60.3, 57.7, 49.4 and 45.2 (Hm, HmЈ and
CH ), 14.8 (H ), –4.63, –8.20 and –14.7 (CH ) ppm. IR (KBr): ν =
˜
2
p
2
3400 (νOH), 1640 (ν
), 1608 (νC=N), 1436 (νC=O ), 1144, 1090
C=Oas
and 1111 (νClO) cm–1. ESI-MS (acetone): m/z (%) =s 234.2 (64) [L
+ FeII]2+/2, 467.3 (100) [L – H + FeII]+, 566.8 (7) [L + ClO4
+
FeII]+.
[FeII(L4)(C3H6O)](ClO4)2·(H2O):
C25H36Cl2FeN4O14
N,NЈ-Dicarboxymethyl-(1R,2R)-N,NЈ-bis(2-pyridylmethyl)-1,2-
cyclohexanediamine: L4 was synthesized using the same procedure
from N,NЈ-bis(2-pyridylmethyl)-1,2-cyclohexanediamine with
2.2 equiv. of tert-butyl 2-bromoacetate. Yield for the two steps:
41%. 13C NMR (75 MHz, [D6]acetone): δ = 171.0 (CO2H), 152.2
(Py), 145.5 (Py), 142.2 (Py), 125.6 (Py), 125.2 (Py), 65.2 (Cy), 62.1
(CH2CO), 51.4 (CH2Py), 24.2 (Cy) ppm.
(743.33): calcd. C 40.40, H 4.84, Fe 7.50, N 7.54; found C 40.60,
H 4.93, Fe 6.50, N 7.49.
[FeII(L2)](ClO4)2: The same procedure was used as described above;
yield 51%. 1H NMR (300 MHz, [D6]Acetone): δ = 125.9 (Ho), 98.3,
89.8 and 67.9 (CH2), 57.4, 49.1 and 45.3 (Hm, HmЈ and CH2), 13.0
(Hp), –29.7 (Hp or CH ) ppm. IR (KBr): ν = 3436 (νOH), 1622
˜
2
(νC=O ), 1611 (νC=N), 1447 (ν
), 1121 and 1108 (νClO) cm–1. ESI-
Synthesis of Metal Complexes
C=Os
as
MS (acetone): m/z (%) = 448.2 (100) [L – H + Cl + FeIII]+, 413.3
(46) [L – H + FeII]+. [FeII(L2)](ClO4)2: C18H22Cl2FeN4O12 (613.14):
calcd. C 35.26, H 3.62, Cl 11.56, Fe 9.11, N 9.14; found C 35.63,
H 3.93, Cl 10.40, Fe 9.44, N 8.93.
[FeII(BPMCN)](FeIICl4) and [FeII(BPMEN)](FeIICl4): [FeII-
(BPMCN)](FeIICl4) and [FeII(BPMEN)](FeIICl4) were synthesized
according to a literature procedure.[13,32] ESI-MS (acetone): m/z (%)
= 162.8 (100) [FeIICl3]–.
[FeIII(L4)(Cl)](FeIIICl4): Under air, L4 (63 mg, 0.15 mmol) was dis-
solved in acetone (approx. 6 mL). FeIIICl3·6H2O (41 mg,
0.15 mmol) dissolved in acetone (approx. 2 mL) was added drop-
wise to afford an orange solution. After 2 h of stirring, the complex
was precipitated by the addition of diethyl ether (approx. 10 mL).
The product was washed with diethyl ether and dried under vac-
[FeII(L3)(C3H6O)](FeIICl4) (General Procedure): In a glove box, L3
(102.9 mg, 0.28 mmol) was dissolved in acetone (approx. 10 mL).
FeIICl2 (55.9 mg, 0.28 mmol) was added, and the mixture was left
to stir for 30 min. The complex precipitated out within 5 min. More
solid appeared on addition of diethyl ether. The complex presented
as a yellow solid (60%). 1H NMR (300 MHz, [D6]acetone): δ =
153.1, 149.9, 121.0, 105.6, 93.1, 86.1, 67.6, 60.5, 59.1, 56.8, 53.4,
51.1, 49.4, 48.3, 47.4, 46.6, 44.1, 38.6, 36.9, 29.4, 27.2, 23.6, 22.7,
21.3, 15.0, 14.5, 13.8, 9.89, 6.42, 5.25, 2.20, 1.13, –0.11, –1.54,
1
uum to give a yellow solid (44%). H NMR (300 MHz, MeOD): δ
= 76.7, 58.9, 26.7 and –18.6 ppm. IR (KBr): ν = 3412 (νOH), 2936
˜
(νCH), 1643 (ν
), 1445 (νC=O ) cm–1. ESI-MS (MeOH): m/z (%)
C=Oas
s
= 466.3 (100) [L – 2H + FeIII]+, 502.2 (5) [L – H + Cl + FeIII]+.
–3.11, –3.69, –4.80, –7.88, –12.5, –15.1, –19.8 ppm. IR (KBr): ν =
˜
[FeIII(L4)(Cl)](FeIIICl4): C21H30Cl5Fe2N4O4 (691.45): calcd.
C
3442 (νOH), 2926 (νCH), 1638 (ν
), 1443 (νC=O ) cm–1. ESI-MS
C=Oas
(acetone): m/z (%) = 423.4 (100) [L – H + FeII]+,s 459.2 (24) [L +
Cl + FeII], 481.3 (83) [L – H + Cl + Na + FeII]. Because of its
sensitivity to oxygen, the FeIICl4 counterion was transformed to
the more stable dimer Cl3FeIIOHFeIICl3 during sample preparation
for elemental analysis. [FeII(L3)(C3H6O)](Cl3FeIIOHFeIICl3)2·
1H2O: C24H40Cl12Fe5N4O7 (1201.27): calcd. C 24.35, H 3.16, N
4.73; found C 24.87, H 3.58, N 4.41.
36.48, H 4.37, Fe 16.15, Cl 25.64, N 8.10; found C 36.89, H 4.13,
Fe 16.45, Cl 26.02, N 8.04.
Reaction Conditions for Catalysis Experiments: A H2O2 solution in
CH3CN (150 μL, 2.0 m, diluted from a 50% H2O2 solution) was
delivered by syringe pump over 30 min at room temperature in air
(or added all at once) to a vigorously stirred CH3CN solution
(850 μL) that contained catalyst (1 μmol), substrate (200 μmol),
and of AgCF3SO3 (4.5 μmol). The final concentrations were 1 mm
iron catalyst, 300 mm H2O2, and 200 mm substrate. The solution
was stirred for another 30 min after the H2O2 addition. These ex-
periments were carried out at room temperature with a variable
amount of acetic acid (100 equiv. of CH3COOH or a 1:2 mixture
CH3CN/CH3COOH) or with HClO4 (1 mm). A benzophenone
solution in CH2Cl2 (20 μL, 1 m) was added as an internal reference
to the samples before GC analysis. The products were identified by
comparison of their GC retention times and GC–MS with those
of authentic compounds and 1H NMR spectroscopy. For NMR
spectroscopic characterization, the solution was washed with satu-
rated NaHCO3 solution, and the organic layer was dried with
Na2SO4 then evaporated under reduced pressure before analysis.
All experiments were run at least in duplicate, and the reported
[FeII(L1)(C3H6O)](FeIICl4): The same procedure was used as de-
scribed above; yield 74%. 1H NMR (300 MHz, [D6]acetone): δ =
161.5, 152.8, 129.6, 105.4, 97.5, 94.5, 88.1, 77.3, 57.1, 56.1, 53.5,
52.6, 51.8, 50.0, 48.6, 43.7, 42.8, 39.7, 31.9, 29.5, 27.7, 16.4, 9.27,
5.46, 5.26, 5.01, 2.99, 0.56, –1.14, –2.52, –4.23, –17.4 ppm. IR
(KBr): ν = 3400 (νOH), 1620 (ν
), 1609 (νC=N), 1445 (ν
)
˜
C=Oas
C=Os
cm–1. ESI-MS (acetone): m/z (%) = 369.3 (100) [L – H + FeII]+.
[FeII(L1)(C3H6O)](Cl3FeIIOFeIICl3)·2H2O: C20H32Cl6Fe3N4O6
(804.76): calcd. C 29.85, H 4.01, Cl 26.43, Fe 20.82, N 6.96; found
C 29.22, H 3.38, Cl 28.39, Fe 19.62, N 7.95.
[FeII(L4)(Cl)]2(FeIICl4): The same procedure was used as described
above; yield 59 %. 1H NMR (300 MHz, [D6]acetone): δ = 54.8,
53.2, 51.6 and 42.3 (CH2, Hm and HmЈ), 21.5, 18.9, 17.0, 13.6, 13.2,
10.7, 10.3 and 8.14 (CH2), –3.53, –3.43, –8.08, –17.2 (CH2 and Hp) data are the average of these reactions. See Supporting Information
ppm. IR (KBr): ν = 3392 (νOH), 1647 (ν
), 1607 (νC=N), 1444
for product identification for (R)-carvone oxidation.
˜
C=Oas
Eur. J. Inorg. Chem. 2012, 85–96
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
95

S. Ménage et al.
FULL PAPER
[13] M. Costas, L. Que, Angew. Chem. 2002, 114, 2283; Angew.
Chem. Int. Ed. 2002, 41, 2179–2181.
[14] a) Y. Xu, R. H. Wang, B. Y. Lou, L. Han, M. C. Hong, Acta
Crystallogr., Sect. C: Cryst. Struct. Commun. 2004, 60, M296–
M298; b) F. Marchetti, B. Melai, G. Pampaloni, S. Zacchini,
Inorg. Chem. 2007, 46, 3378–3384.
Supporting Information (see footnote on the first page of this arti-
cle): Complex characterization (ESI-MS, NMR spectra in the pres-
ence of O2 or triflate), characterization of the products of (R)-car-
vone oxidation.
[15] P. D. Oldenburg, A. A. Shteinman, L. Que, J. Am. Chem. Soc.
2005, 127, 15672–15673.
[16] D. Casanova, P. Alemany, J. M. Bofill, S. Alvarez, Chem. Eur.
J. 2003, 9, 1281–1295.
Acknowledgments
This research was supported by the Agence Nationale pour la Re-
cherche (ANR) (H:AMAC Project ANR-08-CP2D-12) and by the
Région Rhône-Alpes Cible.
[17] D. G. Lonnon, G. E. Ball, I. Taylor, D. C. Craig, S. B. Colbran,
Inorg. Chem. 2009, 48, 4863–4872.
[18] I. Perelygin, M. Klimchuk, J. Appl. Spectrosc. 1974, 20, 687–
688.
[19] K. Nakamoto, Infrared and Raman Spectroscopy of Inorganic
and Coordination Compounds, 4th ed., Wiley, New York, 1986,
p. 231–233.
[20] Y. Mekmouche, S. Menage, J. Pecaut, C. Lebrun, L. Reilly, V.
Schuenemann, A. Trautwein, M. Fontecave, Eur. J. Inorg.
Chem. 2004, 3163–3171.
[21] A. P. Ginsberg, M. B. Robin, Inorg. Chem. 1963, 2, 817–822.
[22] P. A. Shapley, W. S. Bigham, M. T. Hay, Inorg. Chim. Acta
2003, 345, 255–260.
[23] S. Menage, J. M. Vincent, C. Lambeaux, G. Chottard, A.
Grand, M. Fontecave, Inorg. Chem. 1993, 32, 4766–4773.
[24] R. Mas-Balleste, M. Costas, T. van den Berg, L. Que, Chem.
Eur. J. 2006, 12, 7489–7500.
[25] a) O. V. Makhlynets, P. Das, S. Taktak, M. Flook, R. Mas-
Balleste, E. V. Rybak-Akimova, L. Que, Chem. Eur. J. 2009,
15, 13171–13180; b) N. Y. Oh, M. S. Seo, M. H. Lim, M. B.
Consugar, M. J. Park, J. U. Rohde, J. H. Han, K. M. Kim, J.
Kim, L. Que, W. Nam, Chem. Commun. 2005, 5644–5646; c)
S. Taktak, M. Flook, B. M. Foxman, L. Que, E. V. Rybak-Aki-
mova, R. Akimova, Chem. Commun. 2005, 5301–5303.
[26] R. Rinaldi, H. F. N. de Oliveira, H. Schumann, U. Schuchardt,
J. Mol. Catal. A 2009, 307, 1–8.
[27] a) P. Pelagatti, A. Bacchi, M. Balordi, A. Caneschi, M. Gian-
netto, C. Pelizzi, L. Gonsalvi, M. Peruzzini, F. Ugozzoli, Eur.
J. Inorg. Chem. 2007, 162–171; b) M. Kim, Y. U. Kim, J. H.
Han, Polyhedron 2007, 26, 4003–4008.
[28] K. Chen, M. Costas, J. L. Que, J. Chem. Soc., Dalton Trans.
2002, 672–679.
[29] a) B. S. Lane, K. Burgess, Chem. Rev. 2003, 103, 2457–2473; b)
S. Taktak, W. H. Ye, A. M. Herrera, E. V. Rybak-Akimova, In-
org. Chem. 2007, 46, 2929–2942.
[30] a) Q. H. Xia, H. Q. Ge, C. P. Ye, Z. M. Liu, K. X. Su, Chem.
Rev. 2005, 105, 1603–1662; b) H. L. Yeung, K. C. Sham, C. S.
Tsang, T. C. Lau, H. L. Kwong, Chem. Commun. 2008, 3801–
3803; c) Y. Nishikawa, H. Yamamoto, J. Am. Chem. Soc. 2011,
133, 8432–8435; d) C. Marchi-Delapierre, A. Jorge-Robin, A.
Thibon, S. Menage, Chem. Commun. 2007, 1166–1168; e) F. G.
Gelalcha, B. Bitterlich, G. Anilkumar, M. K. Tse, M. Beller,
Angew. Chem. 2007, 119, 7431; Angew. Chem. Int. Ed. 2007,
46, 7293–7296.
[1]
[2]
P. T. Anastas, M. M. Kirchhoff, Acc. Chem. Res. 2002, 35, 686–
694.
a) E. G. Kovaleva, J. D. Lipscomb, Science 2007, 316, 453–457;
b) B. Kauppi, K. Lee, E. Carredano, R. E. Parales, D. T. Gib-
son, H. Eklund, S. Ramaswamy, Structure 1998, 6, 571–586.
S. M. Resnick, K. Lee, D. T. Gibson, J. Ind. Microbiol. Biotech-
nol. 1996, 17, 438–457.
a) P. D. Oldenburg, C. Y. Ke, A. A. Tipton, A. A. Shteinman,
L. Que, Angew. Chem. 2006, 118, 8143; Angew. Chem. Int. Ed.
2006, 45, 7975–7978; b) P. C. A. Bruijnincx, I. L. C. Buurmans,
S. Gosiewska, M. A. H. Moelands, M. Lutz, A. L. Spek, G.
van Koten, R. J. M. K. Gebbink, Chem. Eur. J. 2008, 14, 1228–
1237; c) P. D. Oldenburg, Y. Feng, I. Pryjomska-Ray, D. Ness,
L. Que, J. Am. Chem. Soc. 2010, 132, 17713–17723.
L. Que, W. B. Tolman, Nature 2008, 455, 333–340.
M. S. Chen, M. C. White, Science 2007, 318, 783–787.
[3]
[4]
[5]
[6]
[7] L. Gomez, I. Garcia-Bosch, A. Company, J. Benet-Buchholz,
A. Polo, X. Sala, X. Ribas, M. Costas, Angew. Chem. 2009,
121, 5830; Angew. Chem. Int. Ed. 2009, 48, 5720–5723.
[8] a) M. Klopstra, G. Roelfes, R. Hage, R. M. Kellogg, B. L. Fer-
inga, Eur. J. Inorg. Chem. 2004, 846–856; b) A. Company, L.
Gomez, X. Fontrodona, X. Ribas, M. Costas, Chem. Eur. J.
2008, 14, 5727–5731; c) M. Costas, A. K. Tipton, K. Chen,
D. H. Jo, L. Que, J. Am. Chem. Soc. 2001, 123, 6722–6723; d)
M. C. White, A. G. Doyle, E. N. Jacobsen, J. Am. Chem. Soc.
2001, 123, 7194–7195; e) M. R. Bukowski, P. Comba, A. Li-
enke, C. Limberg, C. L. de Laorden, R. Mas-Balleste, M.
Merz, L. Que, Angew. Chem. 2006, 118, 3524; Angew. Chem.
Int. Ed. 2006, 45, 3446–3449; f) P. Liu, E. L. M. Wong,
A. W. H. Yuen, C. M. Che, Org. Lett. 2008, 10, 3275–3278; g)
K. Chen, M. Costas, J. H. Kim, A. K. Tipton, L. Que, J. Am.
Chem. Soc. 2002, 124, 3026–3035; h) T. W. S. Chow, E. L. M.
Wong, Z. Guo, Y. G. Liu, J. S. Huang, C. M. Che, J. Am. Chem.
Soc. 2010, 132, 13229–13239.
[9] K. Suzuki, P. D. Oldenburg, L. Que, Angew. Chem. 2008, 120,
1913; Angew. Chem. Int. Ed. 2008, 47, 1887–1889.
[10] R. Mas-Balleste, L. Que, J. Am. Chem. Soc. 2007, 129, 15964–
15972.
[11] M. Costas, K. Chen, L. Que, Coord. Chem. Rev. 2000, 200,
517–544.
[12] C. Baffert, M. N. Collomb, A. Deronzier, S. Kjaergaard-
Knudsen, J. M. Latour, K. H. Lund, C. J. McKenzie, M. Mort-
ensen, L. Nielsen, N. Thorup, Dalton Trans. 2003, 1765–1772.
[31] T. Soundiressane, S. Selvakumar, S. Menage, O. Hamelin, M.
Fontecave, A. P. Singh, J. Mol. Catal. A 2007, 270, 132–143.
[32] K. Chen, L. Que, Chem. Commun. 1999, 1375–1376.
Received: July 27, 2011
Published Online: December 2, 2011
96
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 85–96