Heterometallic CoIII4FeIII2 schiff base complex: Structure, electron paramagnetic resonance, and alkane oxidation catalytic activity

Article abstract of DOI:10.1021/ic301460q

The heterometallic complex [Co₄Fe₂OSae ₈]·4DMF·H₂O (1) was synthesized by one-pot reaction of cobalt powder with iron chloride in a dimethylformamide solution of salicylidene-2-ethanolamine (H₂Sae) and characterized by single crystal X-ray diffraction analysis, magnetic measurements, high frequency electron paramagnetic resonance (HF-EPR), and Moessbauer spectroscopies. The exchange coupling in the Fe(III)-Fe(III) pair is of antiferromagnetic behavior with J/hc = -190 cm^-1. The HF-EPR spectra reveal an unusual pattern with a hardly detectable triplet signal of the Fe(III) dimer. The magnitude of D (ca. 13.9 cm^-1) was found to be much larger than in related dimers. The catalytic investigations disclosed an outstanding activity of 1 toward oxidation of cycloalkanes with hydrogen peroxide, under mild conditions. The most efficient system showed a turnover number (TON) of 3.57×10 ³ with the concomitant overall yield of 26% for cyclohexane, and 2.28×10³/46%, respectively, for cyclooctane. A remarkable turnover frequency (TOF) of 1.12×10⁴ h^-1 (the highest initial rate W₀ = 3.5×10^-4 M s^-1) was achieved in oxidation of cyclohexane. Kinetic experiments and selectivity parameters led to the conclusion that hydroxyl radicals are active (attacking C-H bonds) species. Kinetic and electrospray ionization mass spectrometry (ESI-MS) data allowed us to assume that the trinuclear heterometallic particle [Co₂Fe(Sae)₄]⁺, originated from 1 in solution, could be responsible for efficient generation of hydroxyl radicals from hydrogen peroxide.

Full text of DOI:10.1021/ic301460q

Article
pubs.acs.org/IC
Heterometallic CoIII FeIII Schiff Base Complex: Structure, Electron
4
2
Paramagnetic Resonance, and Alkane Oxidation Catalytic Activity
Dmytro S. Nesterov,† Eduard N. Chygorin,‡ Volodymyr N. Kokozay,*,‡ Volodymyr V. Bon,§
Roman Boca,∥,⊥ Yuriy N. Kozlov,⊗ Lidia S. Shul’pina,○ Julia Jezierska,# Andrew Ozarowski,▽
̌
Armando J. L. Pombeiro,*,† and Georgiy B. Shul’pin⊗
†Centro de Química Estrutural, Complexo I, Instituto Superior Tec
Lisbon, Portugal
́
nico, Technical University of Lisbon, Av. Rovisco Pais, 1049-001,
‡Department of Inorganic Chemistry, Taras Shevchenko National University of Kyiv, 64 Volodymyrska St., 01601 Kyiv, Ukraine
§Department of Chemistry of Complex Compounds, V. I. Vernadsky Institute of General and Inorganic Chemistry, National
Academy of Sciences of Ukraine, 32/34 Palladin Ave., Kyiv 03680, Ukraine
∥Institute of Inorganic Chemistry, FCHPT, Slovak University of Technology, Radlinskeho 9, 812 37 Bratislava, Slovakia
⊥Department of Chemistry, FPV, University of SS Cyril and Methodius, Trnava, Slovakia
⊗Semenov Institute of Chemical Physics, Russian Academy of Sciences, Ulitsa Kosygina, dom 4, Moscow 119991, Russia
○Nesmeyanov Institute of Organoelement Compounds, Russian Academy of Sciences, Ulitsa Vavilova, dom 28, Moscow 119991,
Russia
#Faculty of Chemistry, University of Wroclaw, 14 Joliot-Curie Str., 50-383, Wroclaw, Poland
▽National High Magnetic Field Laboratory, Florida State University, 1800 E. Paul Dirac Drive, Tallahassee, Florida 32310, United
States
S
* Supporting Information
A B S T R A C T : T h e h e t e r o m e t a l l i c c o m p l e x
[Co4Fe2OSae8]·4DMF·H2O (1) was synthesized by one-pot
reaction of cobalt powder with iron chloride in a dimethylforma-
mide solution of salicylidene-2-ethanolamine (H2Sae) and charac-
terized by single crystal X-ray diffraction analysis, magnetic
measurements, high frequency electron paramagnetic resonance
(HF-EPR), and Mossbauer spectroscopies. The exchange coupling in the Fe(III)−Fe(III) pair is of antiferromagnetic behavior
̈
with J/hc = −190 cm−1. The HF-EPR spectra reveal an unusual pattern with a hardly detectable triplet signal of the Fe(III)
dimer. The magnitude of D (ca. 13.9 cm−1) was found to be much larger than in related dimers. The catalytic investigations
disclosed an outstanding activity of 1 toward oxidation of cycloalkanes with hydrogen peroxide, under mild conditions. The most
efficient system showed a turnover number (TON) of 3.57 × 103 with the concomitant overall yield of 26% for cyclohexane, and
2.28 × 103/46%, respectively, for cyclooctane. A remarkable turnover frequency (TOF) of 1.12 × 104 h−1 (the highest initial rate
W0 = 3.5 × 10−4 M s−1) was achieved in oxidation of cyclohexane. Kinetic experiments and selectivity parameters led to the
conclusion that hydroxyl radicals are active (attacking C−H bonds) species. Kinetic and electrospray ionization mass
spectrometry (ESI-MS) data allowed us to assume that the trinuclear heterometallic particle [Co2Fe(Sae)4]+, originated from 1 in
solution, could be responsible for efficient generation of hydroxyl radicals from hydrogen peroxide.
thus suggesting the advantage in utilizing heterometallic
catalytic systems.
INTRODUCTION
■
The ongoing interest in heterometallic transition metal
complexes arises from their diverse applications.1 Many studies
in this area have been stimulated by the magnetic properties of
these materials.2 However, one of the most common
application fields for coordination compoundscatalysisis
still practically unexplored in the case of heterometallic
transition metal complexes with classical N,O-donor ligands,
in contrast to organometallic ones.3 On the other hand, the
biological metalloenzymes, which represent a peak of perfection
in the art of catalysis, often contain heterometallic active sites,4
Some of us have recently demonstrated that heterobi- and
heterotrimetallic coordination compounds of CuII/CoIII/FeIII
and CuII/CoIII/CoII metal compositions with aminoalcohols
2
can act as remarkably active and selective catalysts for oxidation
of alkanes with H2O2 under mild conditions.5,6 These results
can be associated with the synergic effect of few different metals
and, particularly, the outstanding activity of the Cu/Co/Fe
Received: July 6, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ic301460q | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
complex could point to the crucial role of iron. It was also
shown that the presence of NiII or ZnII ions instead of FeIII or
CoII, respectively,5,6 had a strong inhibitory effect on the
catalytic activity, probably because of inability of ZnII and
limited ability of NiII to participate in oxidation chemistry under
the employed conditions.
On the basis of these considerations, we focused our studies
on the search for polynuclear heterometallic M/Fe complexes
exhibiting oxidation catalytic activity. The “direct synthesis”
(DS) proved to be a powerful strategy for the self-assembly
preparation of polynuclear complexes possessing unprece-
dented structures,5,7 and it was chosen as the main synthetic
approach in the present investigation. The flexible, polydentate
N,O-donor Schiff base salicylidene-2-ethanolamine (H2Sae)
was chosen because of its recognized ability to generate
sophisticated coordination assemblies under spontaneous self-
assembly conditions.8 The usage of DS with this ligand,
prepared in situ, resulted in a hexanuclear heterometallic
complex [Co4Fe2OSae8]·4DMF·H2O (1), whose unique
structure, magnetic properties, uncommon electron para-
magnetic resonance (EPR) spectra, and unprecedented
catalytic activity are discussed here.
Figure 1. Molecular structure of 1 with the atom numberings
(hydrogen atoms from were omitted for clarity). Color codes: Co,
cyan; Fe, brown.
RESULTS AND DISCUSSION
■
type (MST, X means oxygen for the case of 1), which, as will be
demonstrated below, represents the most prominent structural
feature of 1 (Figure 2). As we showed before, according to the
Synthesis of 1. The Schiff base H2Sae was prepared in situ
as described,8c by condensation of salicylaldehyde with
ethanolamine in dimethylformamide (DMF) prior to addition
of other reagents. The one-pot reaction of cobalt powder and
FeCl2·4H2O with this solution in open air gave a red mixture,
out of which dark brown crystals of 1 precipitated after one day.
The formation of 1 (solvate DMF and water molecules are not
shown) can be accounted for by the following overall reaction
scheme:
4Co0 + 2FeCl2 + 8H2Sae + 4Et3N + 3.5O2
→ [Co4Fe2OSae8] + 4Et3N·HCl + 6H2O
The in situ formation of Schiff base at the first stage (before
1
addition of the metals) was confirmed by H NMR spectrum
(in DMSO/CCl4) of the oil formed by evaporation of solvent,
where the singlet at 8.39 ppm corresponds to the characteristic
−CHN−proton (Supporting Information, Figure S1).9 The
usage of the initial reagents in stoichiometric ratios gives a
constant yield and reproducibility of the synthesis. Other molar
ratios (for example Co: FeCl2: H2Sae =1: 1: 1) can also be used
for the preparation of 1, but with lower yield and longer
crystallization time. The details of IR spectra of 1 as well as the
thermogravimetric studies are described in the Supporting
Information.
Crystal Structure. The structure of 1 features a hexanuclear
array composed of two trinuclear {Co2Fe(Sae)4} fragments
linked by a single oxygen atom (Figures 1, Supporting
Information, Figure S4; Table S2). The cobalt atom adopts
an almost regular octahedral geometry, being coordinated by
two aminoalcohol ligands with the Co−O(N) distances varying
from 1.872(3) to 1.934(3) Å. Such N2O4 coordination
environment is typical for cobalt in a trivalent oxidation state.
The coordination polyhedra around the iron atoms can be
described as a compressed square pyramid with the Fe−O axial
bond lengths of 1.7837(7) and 1.7799(10) Å for two
crystallographically independent units.
Figure 2. Ball-and-stick representation of the M6(μ-X)9 MST in 1.
Cambridge Structural Database (CSD),10 the number of
hexanuclear complexes ranks in the fourth place (ca. 1300
hits) among all other polynuclear coordination compounds
with M(μ-X)nM (n = 1−4) bridging between metal centers.11
Six metal centers can be combined in various manners and, in
contrast to tetra- or pentanuclear MSTs,6,11 the speculative
evaluation of all possible M6Xb combinations becomes difficult
because of their great diversity: a careful analysis for this case
should be based on “reciprocal” strategy, generating possible
M6Xb combinations followed by searches via CSD. General
observations suggest that the metal centers and bridging atoms
tend to pack as compact as possible and to form symmetric
MaXb assemblies.
Although one can select a few most symmetric MSTs, where
six metal centers form a line (Figure 3), rod, or wheel, these
types cover only about 20% of all the hexanuclear structures
(the wheel type is the leading one with 105 hits). Most of the
others have distorted or unclear MSTs, which can be hardly
classified. The structure of 1 is also highly symmetric, belonging
The combination of four octahedra with two square
pyramids (Figure 1) results in a M6(μ-X)9 molecular structure
B
dx.doi.org/10.1021/ic301460q | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 5. Most typical MSTs (among all structures in CSD),
possessing M(μ-X)M bridge.
Figure 3. Some typical interconnections between metal centers for the
case of hexanuclear complexes and the respective MSTs.
McXd (MST of 1 can be written as M2X4−X−M2X4),
evidencing that a single M(μ-X)M bridge between two
polynuclear MaXb fragments is quite rare.
The question arising in this context is that how is such an
uncommon structure, as that of 1, formed. To answer this, one
can propose the general formation scheme (Figure 6). The
to the S4 point group. The linkage by one oxygen atom results
in a quite rare MST M6(μ-X)9 and, as far as we are aware of, the
only known compound possessing such a structure is the
heterometallic complex [Cd2Cl{Co(aet)2(en)2}](NO3)7 (2)
(aet = 2-aminoethanethiolate; en = ethylenediamine),12 the
structure of which is depicted at Figure 4. Although the
Figure 4. Molecular structure of the complex [Cd2Cl{Co-
(aet)2(en)2}](NO3)7,25 showing the same MST as for 1.
Figure 6. Proposed formation of 1. Side reactions are marked by gray.
topologies of the M6(μ-X)9 cores in 1 and 2 are equal, the
geometric configurations are different: the planes formed by
Co2Fe triangles in 1 are perpendicular (Figure 2), while in 2 the
respective Co2Cd trinuclear fragments lie approximately in one
plane (torsion angle Co···Cd···Cd′···Co′ is 23.7°). The obvious
reason for the distortion of the structure of 1 is the sterical
hindrance of the bulky Sae ligands.
The MST of 1 and 2 is quite rare, and the M(μ-X)M bridge
seems to be the most important feature. Particularly, the initial
search within the CSD toward the hexanuclear coordination
compounds containing such a type of bridging, at least one
M(μ-X)M bridge, gave only 35 hits, and most of them had a
cyclic structure (Figure 5). In the search for polynuclear
compounds where two MaXb arrays would be connected just by
one bridging atom we extended our search to a broader range
of nuclearities (four and more metal centers were included). To
our great surprise, the new search did not reveal a great number
of desired compounds (153 hits were selected). Approximately
a half of the structures that formally fit the search conditions
were found to be based on the same M4(μ-X)6 and M4(μ-X)4
tetranuclear MSTs (Figure 5), 25 and 55 hits, respectively.
Finally, careful examination of the results revealed only 13
compounds (see ESI for refcode list) of the type MaXb−X−
coordination mode of the Schiff base ligand to a Co center
leads to the formation of a mer-[CoIII(HSae)(Sae)] unit (a)
with the octahedral environment N2O4 typical for cobalt(III),
while the participation of iron(III) in such a reaction is less
probable. At the next stage, free iron ion is coordinated to form
a binuclear [CoFe(Sae)2]2+ block (b), where the free sites in
the coordination environment of iron are occupied by solvent
molecules and chloride anions.
Although the reaction of two blocks b could lead to a
tetranuclear Co2Fe2 fragment, the excess of cobalt ions in the
reaction mixture forces the interaction of the cobalt unit a with
the binuclear unit b, resulting in a trinuclear block c
[Co2Fe(Sae)4]+. The coordination of a third block a with the
formation of a Co3Fe fragment is unlikely because of the
sterical hindrance of bulky benzene rings of the ligand. Finally,
two blocks c form a hexanuclear structure of 1 through the Fe−
O−Fe oxo-bridge. In contrast to previous steps, the last one is
most difficult to explain: why does compound c not crystallize
as a trinuclear [Co2Fe(Sae)4]Cl complex, but forms a dimer?
First, looking at the dimeric structure (Figure 7) one can notice
a self-complementary nature of the trinuclear fragments, rotated
C
dx.doi.org/10.1021/ic301460q | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
are magnetically silent. This implies that a model of the
exchange coupled Fe(III) pair could be appropriate
iso
−1
= −J(S ·SB)ℏ−2 + μ g B(SAz + SBz)ℏ
̂
H
̂
̂
A
B
iso
In the case of antiferromagnetic exchange, the ground state is
S = 0 and the excited spin multiplets are S = 1, 2, 3, 4, and 5
when the spin state of the Fe atoms is S = 5/2. A very steep
decrease of the effective magnetic moment suggests that J is
very negative. In such a case the excited triplet state is well
separated from the ground state singlet. The high-temperature
limit for the effective magnetic moment of a system of two
metal ions is μeff/μB = giso[2s(s+1)]1/2 which amounts to μeff/μB
= 4.18 giso for S = 5/2.
Inclusion of the single-ion zero-field splitting into the model
offers no improvement because its effects can only be
significant at very low temperatures at which S = 1 and higher
states of the dimer are not populated. The S = 0 ground state
implies that the magnetic susceptibility should vanish for T
approaching zero. Inspection to Figure 8 shows that this is not
the case, as the vanishing susceptibility signal from the diad is
replaced by the rising susceptibility of a paramagnetic impurity
at low temperatures. Therefore the susceptibility has been fitted
by using
Figure 7. Illustration of the conformity between two trinuclear
{Co2Fe(Sae)4} fragments (black and white), rotated to 90°, forming
the hexanuclear molecule of 1. A view along Fe−O−Fe line; only one
pair of Sae ligands are shown.
90° to each other, where the dimer is strengthened by weak C−
H···O and van der Waals interactions.
However, the compound 2 possessing the same MST (Figure
4) does not exhibit such intramolecular interactions,12 pointing
to their secondary role. The synthetic scheme of 2 includes the
elimination of excess chloride anion from the reaction mixture,
leading to a polynuclear structure. Analogously, the presence of
Et3N in the reaction mixture of 1 leads to the formation of the
salt Et3N·HCl (see reaction scheme), promoting the main
reaction (step 4 in Figure 6), while the side reaction of step 4 is
suppressed. To confirm this, we attempted to prepare complex
1 without using a base (a key component), and the desired
product was not obtained.
Magnetic Measurements. The effective magnetic mo-
ment for complex 1 at T = 300 K is μeff = 2.97 μB and on
cooling it gradually decreases to μeff = 1.10 μB at T = 50 K.
Below this temperature its decrease is less gradual to μeff = 0.95
μB at T = 7 K and then the effective magnetic moment drops
down more rapidly (Figure 8).
χ
= (1 − xPI)χdiad + 2xPICPI/(T − ΘPI) + α
sample
where the Curie constant CPI = (NAμ0μ2B/kB)gP2ISPI(SPI + 1)/3
contains the basic physical constants in their usual meaning; gPI
= 2 and SPI = 5/2 have been assumed.
The fitting procedure yielded the following set of magnetic
parameters: J/hc = −190 cm−1, giso = 2.008, the temperature-
independent term α = 2.5 × 10−9 m3 mol−1, the mole fraction
of the paramagnetic impurity xPI = 0.014, and the Weiss
constant for the PI is ΘPI = −0.88 K. The quality of the fit is
very good: the discrepancy factor R(χ) = 0.015 (Figure 8).
̈
Mossbauer Spectroscopy. One quadrupole doublet was
seen in the Mossbauer spectra (Figure 9). The small isomer
̈
shift IS = 0.335(1) (versus α-iron) and small quadrupole
splitting QS = 0.692(6) are consistent with the 6A1 state of the
iron(III) ions in 1.
The stoichiometry of the complex 1 shows the presence of
two Fe(III) centers linked by a nearly linear oxido-bridge. The
peripheral cobalt(III) atoms of the whole H-type architecture
Figure 9. Mossbauer spectrum collected at room temperature. Circles
̈
are the experimental data, the solid line is calculated.
High Frequency Electron Paramagnetic Resonance
(HF-EPR) Spectroscopy. In the HF-EPR spectra the signals
originating from a quintet spin state (S = 2) were easily
identified (Figure 10). The spectra were interpreted using the
spin Hamiltonian
Figure 8. Temperature dependence of the effective magnetic moment
for 1. Open circles, experimental data; solid line, fitted.
D
dx.doi.org/10.1021/ic301460q | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
cm−1. Dedipolar is apparently much too small to explain the
magnitude of De and the anisotropic exchange interactions must
contribute to the zero-field splitting.
ESI Mass Spectrometry. Electrospray ionization mass
spectroscopy (ESI-MS) was used as a convenient solution-
based technique to examine the behavior of 1 in solution with
the aim of screening possible transformation processes and
detecting any novel species formed. The investigation of diluted
(ca. 1 × 10−5 M) solutions of 1 in acetonitrile shows only one
major product of decomposition: [Co(HSae)2]+ at 387.0 m/z.
The negative mode spectra revealed a number of peaks of very
low intensity, whose compositions were not identified. The
particle [Co(HSae)2]+ is expected if one considers the
proposed formation scheme of 1 (Figure 6), the other
proposed intermediates either form uncharged species, or
decompose in solution. However, careful examination of the
spectra allowed to find a peak at 825.7 m/z (ca. 4% of peak
intensity) that could be attributable to a trinuclear particle
[Co2Fe(Sae)4]+, which represents one-half of the dimeric
hexanuclear structure of 1, while the molecular peak of 1 was
not detected in any spectrum.
Figure 10. HF-EPR spectra of 1. Top: experimental spectrum at 290
K, taken with 216.0 GHz. Center: Simulation assuming S = 2 with gx =
0
2.015, gy = 2.011, gz = 2.012, D = 1.029 cm−1, E = 0.008 cm−1, B4
=
−0.0014 cm−1, B44 = −0.0002 cm−1. Bottom: Resonances attributed to
the triplet spectrum (S = 1) appear at very low magnetic fields at
microwave frequencies 406−435 GHz and are better observed at
moderately lowered temperature. The estimated D parameter in the
triplet state is 13.9 cm−1. The arrows indicate a characteristic line in
the S = 2 spectrum shifting with the microwave frequency.
2
2
2
The peak intensities ratio [Co(HSae)2]+: [Co2Fe(Sae)4]+
was found to be dependent on the overall concentration of 1 in
solution (Supporting Information, Figures S5−S7). With
increasing concentration of 1, the intensity of [Co2Fe(Sae)4]+
increases, reaching about 35% of the [Co(HSae)2]+ peak
intensity for [1] = 3 × 10−5 M. This spectrum showed a
number of other peaks of lower intensity, whose compositions
were not identified, although the isotopic distribution suggested
their polynuclear nature. Since the solubility of pure 1 in
acetonitrile is limited to about 3 × 10−5 M, the further ESI-MS
tests were performed by using an acid promoter with the final
concentration of 0.04 M or lower. These conditions are related
to those employed in the catalytic experiments (see below).
With the acid, only two major peaks at 387.0 and 825.7 m/z
were detected, while all the others seem to be suppressed. The
solution with [1] = 2 × 10−5 M shows about 5% of
[Co2Fe(Sae)4]+ (compared to [Co(HSae)2]+, 100% peak in
all spectra). The increase of the concentration of 1 to 5 × 10−5
M and 1 × 10−4 M leads to 14 and 23% of [Co2Fe(Sae)4]+
peaks intensities, respectively.
̂
̂
̂
̂
̂
HS = μBB·g·S + D{Sz − S(S + 1)/3} + E(Sx − Sy )
0
4
+ B4 O 0 + B4 O 4
4
4
in which the zero-field splitting parameters are different in each
coupled-spin state S. The following parameters were found
from the simulation procedures: gx = 2.015, gy = 2.011, gz =
2.012, D = 1.029 cm−1, E = 0.008 cm−1, B4 = −0.0014 cm−1,
0
B44 = −0.0002 cm−1. The zero-field splitting parameters for S =
2 are considerably larger in this complex than in other binuclear
iron(III) complexes.13 The triplet state was difficult to detect.
At very high microwave frequencies, 406 to 435 GHz, a
resonance was seen at very low magnetic fields (Figure 10,
bottom). That resonance first moved toward lower magnetic
field when the frequency was increased from 406 GHz (lower
limit in the 400 GHz range of our system), appeared to cross
the zero magnetic field around 416 GHz, and then moved
toward higher magnetic field. This behavior allowed to assign
this resonance to the triplet state and to estimate the triplet
state D value to be about 13.9 cm−1, also much larger than in
other iron(III) dimers.13e−g
Catalytic Oxidation of Alkanes. We investigated the
catalytic potential of compound 1 for the oxidation of various
alkanes by aqueous hydrogen peroxide under mild conditions.
The reaction occurs in acetonitrile solution, and nitric acid in
low concentration is a necessary component of the reaction
mixture. Sulfuric acid has a weaker accelerating effect, whereas
oxalic, hydrochloric, and trifluoroacetic acids are absolutely
inactive. No alkane oxidation products (or only traces) were
obtained in the absence of any component of the catalytic
system: catalyst, nitric acid, or hydrogen peroxide (Scheme 1; n
= 1, 3). The reaction gives alkyl hydroperoxides which are
gradually transformed into the corresponding ketones
(aldehydes) and alcohols (the oxidation of cyclic alkanes is
presented in Scheme 1). To estimate the concentrations of the
The zero-field splitting parameters observed in various spin
states contain contribution due to the anisotropic interactions
between the two ions as well as a contribution related to the
zero-field splitting on separate high-spin Fe(III) ions. They can
be expressed as
DS = αSD + β D, ES = αSEe + β Ec
e
c
S
S
where the index S designates a coupled-spin state (1, 2,...), De
and Ee are the interaction parameters, while Dc, Ec refer to the
zero-field splitting of the S = 5/2 state of each of the two iron
ions. For a system of two S = 5/2 ions, α1 = 3.7, β1 = −6.4, 2 =
41/42, β2 = −20/21 and Dc, De can be evaluated.13,14 The EPR
data were not sufficient to determine the signs of D in the
triplet and quintet states, and two possibilities have to be
considered. Assuming that D in the triplet and quintet states
have the same sign, one obtains |Dc| = 3.6 cm−1 and |De| = 2.4
cm−1. If the signs are different then |Dc| = 6.4 cm−1 and |De| =
7.3 cm−1. De includes contributions from both the magnetic
dipole−dipole interactions and the anisotropic exchange
Scheme 1
interactions. The dipolar part is Dedipolar = −3gz2μB /R3, where
2
́
R is the inter-ion distance, R = 3.1567 Å, resulting in −0.11
E
dx.doi.org/10.1021/ic301460q | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
three products we measured the concentrations of the ketone
and alcohol twice, before and after the addition of PPh3 in
accord with the method developed earlier by some of us.15
An example of kinetic curves in the oxidation of cyclooctane
is shown in Figure 11. After 6 h the yield of oxygenates attained
An increase of cyclohexane concentration results in the
increase of the overall TON with a nearly linear dependence
(Figure 14) at the initial stage. The saturation of the reaction
mixture by cyclohexane ([CyH]0 = 0.6 M) gives the highest
overall TON of 3.57 × 103 (Table 1, Entry 6), but a further rise
of the substrate amount leads to a rapid yield decrease (16% for
[CyH] ≈ 0.8 M; Table 1, Entry 7). The usage of 50% H2O2
instead of 30% (the final concentration of H2O2 was 1.75 M
instead of 1 M, respectively) had no influence on the above
dependence.
It should be emphasized that yields up to 40% in the
oxidations of very inert alkanes can be considered as very
high.5,6,15,16 The TONs of about 1600−3600 obtained in the
case of catalysis by 1 are much greater than the typical values of
20−200. Moreover, the combination of such a high yield (up to
40%) and TON is rare and indicates an exceptional activity of
the catalytic system based on complex 1. Usually high TON
values, up to 1500, correspond to a yield lower than 5% based
on substrate.16c
Data on the oxidation of the linear alkanes n-heptane and n-
octane, and the branched cycloalkane methylcyclohexane are
summarized in Supporting Information, Tables S4, S5, and S6
(see also Supporting Information, Figure S8). Regio-
[C(1):C(2):C(3):C(4)] and bond [1°: 2°: 3°] selectivity
parameters given in Supporting Information, Tables S4 and S5
(see refs 17−19) are low which allows us to assume that the
oxidizing species in these reactions is the hydroxyl radical.
Supporting Information, Table S6 (entries 2−7) includes, for
comparison, selectivity parameters for some other systems
based on H2O2. All these systems are known to oxidize alkanes
with the participation of hydroxyl radicals.
Figure 11. Accumulation of oxygenates (cyclooctyl hydroperoxide,
curve 1; cyclooctanone, curve 2; cyclooctanol, curve 3) with time in
cyclooctane (0.25 M) oxidation with H2O2 (1.0 M, 50% aqueous)
catalyzed by complex 1 (5 × 10−5 M) in the presence of HNO3 (0.04
M, 65% aqueous) in acetonitrile (total volume of the reaction solution
was 5 mL), 50 °C.
46% based on cyclooctane and turnover number (TON,
turnover number, moles of products per mol of catalyst
precursor) was 2.28 × 103. The dependence of the product
yield after 5 h in the cyclohexane oxidation on catalyst
concentration is depicted in Figure 12. The curve exhibits a
The kinetic data of the cyclohexane oxidation presented in
Figure 15 are also in agreement with the assumption of the
participation of hydroxyl radicals in the alkane oxygenation.
The mode of dependence of the initial cyclohexane (CyH)
oxidation rate W0 on the initial cyclohexane concentration (W0
is approaching to a plateau at high concentrations of
cyclohexane, see Figure 15A) can be explained by the
competitive interaction of hydroxyl radicals both with cyclo-
hexane and acetonitrile.18 The following kinetic scheme
describes the dependence of W0 on [cyclohexane]0:
H2O2 + 1 → HO •
(rate W)
i
(1)
HO • + CyH → Cy• → → CyOOH
(constant k2)
(2)
(3)
Figure 12. Effect of the catalyst concentration on the total yields of
cyclohexanol and cyclohexanone (determined by GC after reduction
with PPh3) in the oxidation of cyclohexane with H2O2 (1.0 M, 30%
aqueous) catalyzed by complex 1 in the presence of HNO3 (0.04 M,
65% aqueous) in acetonitrile (total volume of the reaction solution
was 5 mL), room temperature, 5 h reaction time. The inset shows the
[1]0 = (0−5) × 10−5 M region in detail.
HO • + MeCN → → products
(constant k3)
where the initiation stage (1) corresponds to the catalytic
generation of hydroxyl radicals with rate Wi, the stages (2)
correspond to the sequence of transformations of CyH into
CyOOH (with kinetically rate-limiting interaction between
HO• and CyH), and the stage (3) is the reaction of HO• with
acetonitrile.
The analysis of this kinetic scheme in the quasi-stationary
approximation relative concentration of hydroxyl radicals leads
to the following equation:
steep ascent in the low concentration region (1 × 10−5 M <
[1]0 < 5 × 10−5 M), then reaching a plateau with a slight
decrease at the concentrations of 1 higher than 2 × 10−3
M
(because of the overoxidation). The highest total yield of 38%
was achieved for catalyst concentration of 1.4 × 10−4 M and
H2O2/catalyst ratio of 36: 1 (Table 1,Entry 1). The highest
overall TON of 1.6 × 103 (Figure 13) was obtained for [1]0 =
3.9 × 10−5 M, with the overall yield of 31% (Table 1, Entry 2).
d[CyH]
d[CyOOH]
W
i
k3[MeCN]
k2[CyH]
−
=
=
dt
dt
1 +
(4)
F
dx.doi.org/10.1021/ic301460q | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 1. Oxidation of Cyclohexane with Hydrogen Peroxide to the Corresponding Cyclic Alcohols and Ketones Catalyzed by
a
Complex 1.
b
yields of products obtained from CyH, %
c
entry
1
[CyH], M
n(H2O2)/n(CyH)
[1] × 105, M
ketone
alcohol
total
TON
543
0.2
0.2
0.2
0.2
0.6
0.6
5.0
14
6
32
11
29
7
38
31
31
26
26
23
16
21
d
2
5.0
14
20
2
443
3
5.0
3.9
3.9
4.4
4.4
4.4
11
1.60 × 103
1.34 × 103
3.57 × 103
3.20 × 103
2.83 × 103
722
d
4
5.0
20
3
5
1.67
1.67
1.25
2.7
23
9
d
6
15
1
e
7
0.8
15
18
f
8
0.37
3
a
Selected data; reaction conditions (unless stated otherwise): catalyst precursor, H2O2 (1.0 M, 30% aqueous), CyH (0.2 M), HNO3 (0.04 M, 65%
b
aqueous), acetonitrile (total volume of the reaction solution was 5 mL), 5 h reaction time, ambient temperature (ca. 20 °C). Moles of product/100
mols of cyclohexane, measured upon reduction of the reaction mixture by PPh3 (ROOH is transformed to the alcohol ROH). Overall TON values
(mols of products per mol of catalyst). Yields measured prior to addition of PPh3 (for comparative purposes). Total amount of cyclohexane in the
catalytic system (the real concentration is limited by solubility to ca. 0.6 M). Accumulation within 15 min under the following conditions: catalyst
c
d
e
f
precursor, H2O2 (1.0 M, 50% aqueous), CyH (0.37 M), HNO3 (0.04 M, 65% aqueous), acetonitrile (total volume of the reaction solution was 5
mL), 50 °C.
Figure 13. Effect of the catalyst concentration on the overall TON
(turnover number; mols of products/mol of catalyst). The conditions
are the same as described for Figure 12.
Figure 15. Oxidation of cyclohexane with hydrogen peroxide (1.0 M,
50% aqueous) catalyzed by compound 1 (3.2 × 10−5 M) in the
presence of HNO3 (0.04 M, 65% aqueous) in MeCN at 50 °C. Graph
A: dependence of the initial rate of oxygenate formation W0 on initial
concentration of cyclohexane. Graph B: anamorphosis of curve shown
−1
in Graph A in coordinates [cyclohexane]0 − W0−1. Concentrations
of oxygenates (sum of cyclohexanol and cyclohexanone) were
measured after reduction with PPh3.
Figure 14. Effect of the cyclohexane (CyH) concentration on the total
yields of cyclohexanol and cyclohexanone (red) and total TONs
(blue) in the oxidation of cyclohexane with H2O2 (1.0 M, 30%
aqueous) catalyzed by complex 1 (4.4 × 10−5 M) in the presence of
HNO3 (0.04 M, 65% aqueous) in acetonitrile at room temperature, 5
h reaction time. Total initial volume of the reaction solution was 5 mL,
taking into account the limited solubility of cyclohexane in acetonitrile.
For the tests with [CyH]0 > 0.4 M a complete dissolution of
cyclohexane was observed in about 30 min.
In accordance with eq 4 for the initial fragments of the
kinetic curves, the term
1
d[CyOOH]
dt
should be a linear function of 1/[CyH]0. The experimental data
are in agreement with this dependence which is shown in
G
dx.doi.org/10.1021/ic301460q | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 15B. This dependence allows us to determine the
following ratio:
of 1, and the ratio of their concentrations depend on the overall
concentration of 1 (the higher [1], the higher [Co2Fe]).
Considering these results, the assumption was made that the
trinuclear Co2Fe particle is responsible for the intensive growth
k3[MeCN]
= 8 × 10−2
M
of the initial reaction rate starting from the [1] ∼ 3 × 10−5
M
k2
(Figure 16). Let us consider that monomeric (M) cobalt (Co)
and trimeric (T, of Co2Fe type) species are formed from the
initial hexameric molecules 1 (while the whole molecule of 1
has not been detected in the solution). In accordance with this
observation we can write the following material balance
equation:
Assuming [MeCN] = 18 M in the reaction solution, we
obtain k3/k2 = 4.4 × 10−3. Values k3[MeCN]/k2 (8 × 10−2 M)
and k3/k2 (4.4 × 10−3) are close to the corresponding
parameters obtained by some of us for oxygenation of
cyclohexane in acetonitrile with H2O2 catalyzed by the systems
19a
Cp*2Os/py ({9÷20} × 10−2 M and {5÷11} × 10−3), VO3 /
PCA (15 × 10−2 M and 8.3 × 10−3),19b Fe(ClO4)3 (12 × 10−2
M and 7 × 10−3),18c a binuclear iron(III) complex with 1,4,7-
triazacyclononane/PCA (20 × 10−2 M and 11 × 10−3).18c
These parameters, measured for other systems, for example, for
the photochemical oxidation of cyclohexane with H2O2 (23.4 ×
10−2 M and 13 × 10−3)18c and radiation-chemical experiments
(21.6 × 10−2 M and 12 × 10−3),18c are comparable with those
obtained in the oxidation catalyzed by 1. For all these reactions
the hydroxyl radical is known to be the oxidant.
−
6 × [1]0 = [Co] + [Fe] + 3 × [T]
(5)
Here [1]0 is the concentration of starting hexameric complex
1, 6 × [1]0 corresponds to the maximum concentration of the
monomeric species with the cobalt and iron ions generated
from 1, [Co] and [Fe] is equal to the real concentrations of the
monomeric species with the cobalt and iron ions in the
solution, and 3 × [T] is equal to the concentration of
monomeric species formed when the trimeric complex T
completely dissociates.
It is clear that the concentration of Co ions in the solution is
two times higher than the concentration of ions Fe, that is
[Co]/[Fe] = 2. A portion of ions is present in the form of the
monomers Co and Fe, and another portion exists as the trimers
T. Let us introduce the effective equilibrium constant K for the
equilibrium in the formation of trimers from monomers:
In the cyclohexane oxidation, the dependence of the initial
reaction rate (W0 = d[CyOOH]/dt) on concentration of 1
introduced into the reaction solution (Figure 16A) testifies that
the reaction order is higher than first order relative to 1. The
ESI-MS tests (see above) showed two major particles,
[Co(HSae)2]+ and [Co2Fe(Sae)4]+, formed in the solutions
2Co + Fe ⇄ T
(K = [T]/[Co]2[Fe])
(6)
Denoting [Fe] as [M] and taking into account that [Co]/[Fe]
= 2 we can rewrite eq 6 in the form:
K = [T]/4[M]2
(6a)
The data presented in Figure 16 demonstrate that the
catalytic activity of the trimers is many times higher than that of
the monomers. Thus, let us assume that in the whole studied
interval of the hexamer concentrations the initial oxidation rate
is
W = keff[T]
(7)
0
The effective rate constant keff is in particular a function of
[H2O2]. In this case, with [H2O2] = const, it follows from eqs 6a
and 7 that
1/3
⎛
⎜
⎝
⎞
⎟
⎠
W
0
M =
4k K
(8)
eff
which means that
1/3
⎛
⎞
W
W
0
keff
0
6 × [1] =
+ 3
⎜
⎝
⎟
0
4k K
⎠
(9)
eff
Let us transform eq 9 into the following form:
[1]0
= a + bW 2/3
0
W 1/3
(10)
0
Figure 16. Oxidation of cyclohexane (0.37 M) with hydrogen peroxide
(1.0 M, 50% aqueous) catalyzed by compound 1 in the presence of
HNO3 (0.04 M, 65% aqueous) in MeCN at 50 °C. Graph A:
dependence of the initial rate of oxygenate formation W0 on initial
where
1
1 1
2 keff
a =
and b =
2(4keffK)1/3
concentration of 1. Graph B: the anamorphosis of curve shown in
1/3
Graph A in coordinates [1]0/W0
− W02/3. Concentrations of
The linear anamorphosis of the experimental dependence
(Figure 16A) calculated in accord with eq 10 is shown in Figure
oxygenates (sum of cyclohexanol and cyclohexanone) were measured
after reduction with PPh3.
H
dx.doi.org/10.1021/ic301460q | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
16B. It can be seen that the experimental data are in a tolerable
agreement with eq 10. The data presented in Figure 16B yield
parameters a = 0.95 × 10−3 M2/3 s1/3 and b = 0.096 s, and,
consequently, under the conditions of our experiment, keff = 1/
(2b) = 5.2 s−1 and K ≈ 7 × 106 M−3. The simulation of the
dependence of W0 on [1]0 using eq 10 and parameters a and b
given above produced the dashed-line curve in Figure 16A
which demonstrates good agreement between calculated and
experimental values W0 at different initial concentrations of 1.
For solutions with the initial concentration of 1 equal to 1.1
× 10−4 M the value of W0 was found to be 3.44 × 10−4 M s−1.
Using the above coefficient keff = 5.2 s−1 one can find [T] = 6.6
× 10−5 M and, furthermore, the constant K ≈ 7 × 106 M−3
allows us to estimate [Co] + [Fe] = 4 × 10−4 M in the solution.
In all ESI-MS spectra we observe only one type of monomeric
species, namely, [Co(HSae)2]+, while the monomeric iron
species were not detected. Thus, the concentration of cobalt-
only particles is [Co] = 2.7 × 10−4 M. The final [T]/[Co] ratio
of 0.24 for [1] = 1.1 × 10−4 M perfectly fits to observed peak
intensities ratio in the respective mass-spectrum, where 0.23
ratio was found (Supporting Information, Figure S5). The
Figure 17. Accumulation of oxygenates (sum of cyclohexanol and
cyclohexanone) with time in cyclohexane (0.37 M) oxidation with
hydrogen peroxide (1.0 M, 50% aqueous) catalyzed by compound 1 in
the presence of HNO3 (0.04 M, 65% aqueous) in MeCN at 50 °C.
Concentrations of the products were measured after reduction with
PPh3.
analogous calculations for [1]0 = 5.0 × 10−5 and 1.2 × 10−5
M
system, since it accumulates no more than 20% of total yield
before reaching plateau (ca. 10 min; Figure 17). Hence, it is
assumed that the catalyst undergoes degradation upon attack by
hydroxyl radicals generated at a very high rate. Really, the
catalyst contains a Schiff base ligand with numerous C−H
bonds and N-atoms that can be oxidized with the subsequent
change of ligand’s structure and function. Also, the observation
of the color of solution, changing from light-brown to colorless
during the reaction, suggested the decomposition of the Schiff
base group. The ESI-MS spectra of these colorless solutions
showed the absence of both [Co(HSae)2]+ and [Co2Fe(Sae)4]+
particles and, moreover, revealed no dominant peaks in the
100−2000 m/z region.
reveals 0.11 and 0.06 theoretical [T]/[M] ratios, respectively.
The mass-spectra of the solutions shows peak intensities ratios
of 0.14 and 0.03, respectively, which are in good agreement
with predicted ones. Finally, it is quite important to notice that
successful fitting of the experimental data to the model with the
2Co + Fe ⇄ T equilibrium, as well as its correspondence with
mass-spectral data, confirm our supposition that [Co2Fe-
(Sae)4]+ is a catalytically active particle of the system based
on 1.
Further Mechanistic Discussion. At the concentration
[1]0 = 1.1 × 10−4 M the initial reaction rate shows a very high
turnover frequency (TOF) value of 1.12 × 104 h−1 (Table 1,
Entry 8). This rate is comparable to one of the highest reported
rates (2.4 × 104 h−1), observed for an osmium-based catalytic
system,15d and is much greater than those for an overwhelming
majority of catalytic systems based on coordination compounds
as precatalysts. It was shown above that such a high rate could
be associated to the heterometallic particle [Co2Fe(Sae)4]+
present in the concentrated solutions of 1. To prove that this
particle remains under the catalysis conditions, the respective
solutions were investigated by ESI-MS. The addition of nitric
acid and hydrogen peroxide (at room temperature) does not
influence the spectra significantly. After one hour the peak
intensity of Co2Fe particle dropped from 23% to 14% for the
room-temperature catalytic system with [1] = 1.1 × 10−4 M,
while the initial solution of 1 does not show significant changes
with time. This stability is also supported by the fact that the
catalytic activity of already prepared solutions of 1 keeps the
same for a long time (more than six months).
The UV/vis spectrum of 1 exhibits a smooth absorbance at
the visible region with one notable maximum at 388 nm, ε388
=
17713 M−1 cm−1 (Figure 18). In the UV region two clear peaks,
at 210 and 253 nm, are observed. The spectral pattern of 1
remains unchanged in [1] = 5.4 × 10−6−3.0 × 10−5
M
concentrations range (Supporting Information, Figure S9). A
The accumulations of products at 50 °C in the catalytic tests
with [1] equal to 1.1 × 10−4 and 9.6 × 10−5 M (highest
investigated concentrations) are shown at Figure 17. As can be
seen, the reaction rate rapidly decreases with time. The quasi-
linear regions of both dependencies are negligible, forcing the
use of exponential fitting to evaluate the correct initial reaction
rate W0. The decrease of [1] leads to the increase of the initial
quasi-linear region, which become evaluable for [1] < 4 × 10−5
M, finally reaching about 1 h period for the lowest investigated
concentrations.
Figure 18. UV/vis spectra of acetonitrile solution of 1 (5 ×10−5 M),
dashed line (---); acetonitrile solution of 1 (5 × 10−5 M) with HNO3
(0.04 M, 65% aqueous), dotted line (···); solid lines illustrate the
spectra of catalytic solutions in cyclooctane (0.25 M) oxidation with
H2O2 (1.0 M, 50% aqueous) catalyzed by complex 1 (5 × 10−5 M) in
the presence of HNO3 (0.04 M, 65% aqueous) in acetonitrile (total
volume of the reaction solution was 5 mL), 50 °C. The inset shows the
difference spectra taking 5 min spectrum as the basic one.
The observed reaction rate decay could not be explained only
by decrease of substrate or H2O2 amounts in the catalytic
I
dx.doi.org/10.1021/ic301460q | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
further increase of concentration is limited by solubility of 1 in
acetonitrile and can be enhanced only with the usage of small
amounts of acid promoter (ca. 4 × 10−3 M of HNO3 was used
for spectral investigations). However, the acid causes minor
spectral changes (Figure 18), such as a blue shift of the 388 nm
peak to 377 nm with a slightly lower molar extinction
coefficient. The UV region for concentrations of 1 higher
than 3.0 × 10−5 M can not be evaluated because of very strong
absorption. To clarify the color changes of the catalytic systems
(from light-brown to colorless), the UV/vis spectra of the
cyclooctane catalytic system, under conditions similar to those
described above (Figure 14), were monitored (catalyst
concentration was [1] = 5.1 × 10−5 M, all other parameters
remained the same). The dependence of the spectral pattern on
the time, as well as the respective UV/vis spectra of pure 1 and
1 + HNO3 in acetonitrile, are depicted at Figure 18. As can be
seen, the main changes occur in the 250−600 nm region
(Figure 18, inset). The intensity of the 388 nm peak, typical for
1, continuously decays with the time, thus confirming the
decomposition of the compound 1 under the present catalytic
conditions. In 4.5 h time the spectrum reaches a plateau that
corresponds to the colorless, catalytically inactive solution. The
ESI-MS spectra of the 50 °C catalytic test with various
concentrations of 1 revealed no considerable peaks in the 100−
2000 m/z region after the solution became colorless, pointing
to a complete decomposition of all coordination species under
elevated temperatures.
peak to some definite compound because of the great
complexity of the catalytic systemthe absorbance at about
330 nm could be due to organic species or charge transfer in
cobalt or iron coordination compounds, including peroxo- and
hydroxyperoxo species of these metals.20
Although the hydroxyl radicals were proved to be the main
attacking particles in the present catalytic systems, the question
about the catalytic mechanism in general remains open. In spite
of the fact that the catalytic reactions of transition metal
compounds with hydrogen peroxide have been investigated for
more than a century, the great complexity of these Fenton-like
processes21 still does not allow reliable predictability of such
reactions. In the classical Fenton reagent, the first stage is the
decomposition of hydrogen peroxide to hydroxyl radical and
hydroxyl anion:21d
Fe2+ + HOOH → Fe3+ + HO·+OH−
The hydroxyl radical interacts further with the C−H bond to
form the respective alkyl radical:
The first stage has a high activation barrier illustrated by the
rate constant of only 76 M−1 s−1, while, for example,
cyclohexanone reacts with the hydroxyl radical with the much
higher rate of 6 × 109 M−1 s−1.21d Thus, a huge number of the
catalytic systems, including that based on 1, differ only by
efficiency of the initiation reaction of the coordination
compound with HOOH or t-Bu-OOH.
Surprisingly, the absorbance at 337 nm was found to be in
nonlinear dependence on the reaction time (Figures 18, 19). In
In the present case, the combination of high yield and TON
for cyclohexane (26% and 3.57 × 103, respectively, for Table 1,
Entry 5), as well as the high TOF of 1.12 × 104 h−1 (Table 1,
Entry 8), was associated to the trimetallic particle [Co2Fe-
(Sae)4]+ which, probably, is the highly efficient initiator of the
HO• generation. It was shown earlier15,18c,22 that the presence
of polydentate ligands has a prominent influence on the
catalytic activity of iron ions under conditions close to that for
1, enhancing the reaction rate up to 103 times.15,18c,22 The
Co(Sae)2 blocks also could be viewed as a bidentate ligand,
participating in square-pyramidal coordination environment
around iron atom in 1, as evidenced by single crystal X-ray
analysis (Figure 1). However, according to ESI-MS data the
axial oxygen atom decoordinates in solution, and it is most
probable that in the [Co2Fe(Sae)4]+ particle the Co(Sae)2
blocks form a tetracoordinated environment. The geometry of
such environment could range from nearly square planar
(square pyramid without axial bonding) to tetrahedral. The
coordination of the third Co(Sae)2 block is less probable
because of the steric hindrance of the bulky Sae ligands; also
this contrasts to ESI-MS data. Such a coordination environment
is unsaturated and in this way the iron center could actively
participate in catalytic processes. Since the activation of
hydrogen peroxide by a metal ion is the limiting stage, one
could propose a few possibilities that make the [Co2Fe(Sae)4]+
particle especially active in this reaction. First, a stable,
unsaturated coordination geometry could stipulate specific
coordination of hydrogen peroxide to the iron metal center.
This may include additional hydrogen bonding of HOOH with
the deprotonated hydroxyl groups of the aminoalcohol ligand,
located in close proximity. Furthermore, these ligands could
serve as a proton-transfer device, forcing the abstraction of the
proton from the hydrogen peroxide, as was observed previously
for some N-donor ligands.22 One can notice that the nitrate
Figure 19. Variation of absorbance (A) at 337 nm with the time for a
catalytic solution of 1 (the conditions are the same as described for
Figure 18). The first point belongs to a pure solution of 1 (5 × 10−5
M) in acetonitrile.
the period from 5 to 20 min from the beginning the peak
reaches a maximum intensity, being stable for about 10 min,
and then quickly falls to the background absorption level
(Figure 19). Moreover, one could notice some conformity
between the intensity of the 337 nm peak and the reaction rate:
the period showing the highest reaction rate (ca. until 30 min)
corresponds to the highest intensity of this peak, while the
further rate decay is in agreement with the respective decay of
the 337 nm peak intensity. Although the initial reaction rate
was always observed to be the highest one for all the reaction
conditions (this contrasts with the increase of the 337 nm peak
in the 0−20 min range), one can suppose that this absorption
corresponds to an intermediate compound, a component of the
reaction solution. Unfortunately, it is not possible to assign this
J
dx.doi.org/10.1021/ic301460q | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
a
Table 2. Selectivity Parameters in Oxidations of Alkanes with m-CPBA Catalyzed by Complex 1
C(1):C(2):C(3):C(4)
1°:2°:3°
trans/cis
entry
[HNO3], M
n-heptane
MCH
cis-1,2-DMCH
trans-1,2-DMCH
1
2
0
1:88:61:105
1:180:860
0.7
1.0
8.5
0.02
1:190:230:220
1:130:2130
0.06
a
Conditions: [1]0 = 5 × 10−5 M, solvent MeCN, 23 °C, 22 h. Selectivity parameters were determined on the basis of concentrations of alcohol and
ketone isomers after reduction with PPh3.
anion could also play a special role, since the catalytic system
based on 1 possesses weak or no activity in the absence of nitric
acid. Although the above scheme considers the cobalt Co(Sae)2
block only as a bulky ligand, one can suppose at least few
possibilities of CoIII ions to participate in the oxidation reaction
mechanism. This may include Ox-Red reactions between cobalt
and iron ions (when one of the ions is in the reduced state) or
interaction of cobalt and peroxo-species by decoordinating one
of the bridging μ-O atoms from the deprotonated amino-
alcohol.
We have also found that cyclohexane (0.25 M) in the
presence of complex 1 (5 × 10−5 M) and HNO3 (0.02 M) in
acetonitrile solution (total volume 5 mL) can be oxidized by m-
chloroperoxybenzoic acid (m-CPBA, 45 mg) to afford (after
reduction with PPh3) cyclohexanone (0.01 M) and cyclo-
hexanol (0.009 M). The reaction in the absence of nitric acid
was much less efficient and gave cyclohexanone (0.003 M) and
cyclohexanol (0.002 M) (Table 2, Entry 1). The selectivity
parameters for the oxidations by these systems of normal and
branched alkanes are summarized in Table 2. The oxidation of
isomeric 1,2-DMCH (DMCH = dimethylcyclohexane) ex-
hibited a very remarkable peculiarity: the reaction is highly
stereoselective, but only in the presence of nitric acid. The
selectivity parameters dramatically differ from those obtained
for the reactions with H2O2 (compare with the data of
Supporting Information, Table S6) which indicates that in the
case of m-CPBA the reaction proceeds without the participation
of free radicals. As in the oxidation with H2O2 nitrate anions are
required for the creation of efficient catalytically active species.
The oxidation of benzene with H2O2 catalyzed by 1 in the
presence of HNO3 at 50 °C (2 h) was not efficient, and phenol
was formed with TON of only 27.
spectra showed uncommon parameters for an iron dimer.
These investigation methods, together with the Mossbauer
̈
spectroscopy and charge balance calculations, unambiguously
define the metal oxidation states in 1 as Co(III) and Fe(III).
The catalytic data revealed an exciting activity of 1 in the
mild oxidation of alkanes by hydrogen peroxide. Complex 1 is,
probably, the most active coordination compound among those
with N,O-donor ligands, as evidenced by achieved combina-
tions of the overall yield(%)/TON of 31/1.60 × 103 and 26/
3.57 × 103 (for oxidation of cyclohexane), whereas related
catalytic systems typically could show, as most, either a high
yield or a high TON. The advantages of the catalyst 1 also lie in
the great TOF values, with the highest observed value of 1.12×
104 h−1 (3.1 s−1). We tried to understand the nature of active
attacking species and found that hydroxyl radicals react with
alkanes to form the respective alkyl radicals. Thus, the catalytic
system belongs to the class of Fenton-like systems where the
reaction of a coordination compound (solvated metal ion in the
simplest case) with oxidizer (peroxide) is the initial, principal
stage. By employing ESI-MS techniques together with kinetic
experimental data we were able to assume that the [Co2Fe-
(Sae)4]+ particles, present in concentrated solutions of 1, are
the catalytically active species, responsible for the quite fast
generation of hydroxyl radicals in the catalytic system.
The proposed catalytic system provides a rare example of
application of a heterometallic coordination compound in
homogeneous catalysis and, to the best of our knowledge, is the
first case when the catalytic activity is directly associated to a
heterometallic species. Although the detailed mechanism of the
present system has yet to be established, we suppose that the
high activity of the [Co2Fe(Sae)4]+ particle could be due to the
specific coordination environment of a tetracoordinated iron
center, as well as to other effects such as hydrogen-bonded
assistance and Co−Fe Ox-Red interactions. We expect that the
results obtained within the present research would improve the
understanding of oxidation processes and inspire further
development of highly efficient catalysts for mild hydrocarbon
functionalization.
CONCLUSIONS
■
We have shown that application of the “Direct Synthesis”
method together with the use of a polydentate Schiff base
ligand, such as salicylidene-2-ethanolamine, represents an
excellent strategy for the synthesis of polynuclear heterometallic
complexes of uncommon structure. As evidenced by the
searches in the CSD, the coordination core of 1 is an extremely
rare example of the M6(μ-X)9 MST, where two trinuclear
M3(μ-X)4 species are linked by single M(μ-X)M bonding into
the hexanuclear assembly.
EXPERIMENTAL SECTION
■
All chemicals were of reagent grade and used as received. All
experiments were carried out in air. Elemental analyses for CHNS
were performed by the Microanalytical Service of the Instituto
To understand why such an unusual structure is formed, we
explored possible reaction pathways and proposed a speculative
mechanism for the synthesis of 1 in solution. Unfortunately,
nowadays such synthetic systems usually are not subject of
investigation, and even tentative discussions of how the
polynuclear species are formed from multicomponent systems
are quite rare.2c In the present case we introduced a respective
scheme and, moreover, two particles present from this scheme
were observed in ESI-MS spectra of the solutions of 1.
The magnetic properties of 1 revealed an antiferromagnetic
coupling between Fe(III)−Fe(III) centers. The HF-EPR
́
Superior Tecnico, and the elemental analyses for metals were
performed by atomic absorption spectroscopy by the Department of
Chemistry, Taras Shevchenko National University of Kyiv. Infrared
spectra (4000−400 cm−1) were recorded on a BX-FT IR “Perkin
Elmer” instrument in KBr pellets. UV−vis spectra were recorded using
Perkin-Elmer Lambda 35 spectrometer in a 200−800 nm spectral
range with 2 nm spectral resolution. Pure acetonitrile was measured as
background sample. ESI-MS(+) spectra were run on a 500-MS LC Ion
Trap instrument (Varian Inc., Alto Palo, CA, U.S.A.) equipped with an
electrospray (ESI) ion source, using 10−3−10−5 M solutions of 1 in
methanol, acetonitrile, or DMF. A Perkin-Elmer STA-6000 model
thermogravimetric analyzer was used for determination of the thermal
K
dx.doi.org/10.1021/ic301460q | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
stability of complex 1. Samples weighing 5−30 mg were heated under
a dinitrogen flow of 30 mL min−1 from 30 to 1000 °C at a heating rate
of 10 °C min−1.
organic compounds at elevated temperatures may be explosive.). The
reactions were stopped by cooling, and analyzed twice, that is, before
and after the addition of an excess of solid PPh3. This method was
developed and used previously by some of us15 for the analysis of the
reaction mixtures obtained from various alkane oxidations. Applying
this method in the present work for the oxidation of cyclooctane and
cyclohexane, we demonstrate that the reaction affords predominantly
the alkyl hydroperoxide as a primary product which slowly
decomposes to form cyclohexanol and cyclohexanone. In our kinetic
studies for precise determination of oxygenate concentrations only
data obtained after reduction of the reaction sample with PPh3 were
used. A Fisons Instruments GC 8000 gas chromatograph with a DB-
WAX capillary column (J&W), Perkin-Elmer Clarus 500 gas
chromatograph with a BP-20 capillary column (SGE) and Perkin-
Elmer Clarus 600 gas chromatograph, equipped with Perkin-Elmer
Clarus 600 C mass-spectrometer (electron impact), with a BPX5
capillary column (SGE) were used for quantitative and qualitative
(MS) analyses of the reaction mixtures. The parameters of all the
columns are 30 m × 0.32 mm × 25 μm; helium was used as the carrier
gas; the internal standards were nitromethane for kinetic tests and
cycloheptanone for the other ones.
Synthesis of [Co4Fe2OSae8]·4DMF·H2O (1). Salicylaldehyde
(0.53 mL, 5 mmol), 2-aminoethanol (0.3 mL, 5 mmol), and
triethylamine (0.7 mL, 5 mmol) were dissolved in DMF (25 mL),
forming a yellow solution, and magnetically stirred at 50−60 °C (10
min). Then, cobalt powder (0.15 g, 2.5 mmol) and FeCl2·4H2O (0.25
g, 1.25 mmol) were added to the hot yellow solution with subsequent
stirring for 6 h, until total dissolution of cobalt was observed and a
dark red solution formed. Dark red crystals suitable for X-ray analysis
were isolated after 1 day. Yield: 0.53 g, 42.1% (per cobalt). Elemental
analysis for C84H102Co4Fe2N12O22 (Mr = 1979.24). Calcd: C, 50.98; N,
8.49; H, 5.19; Fe, 5.64; Co, 11.91. Found: C, 50.8; N, 8.3; H, 5.2; Fe,
5.7; Co, 11.9. The compound is sparingly soluble in DMSO and DMF,
insoluble in water, and it is stable in air.
X-ray Structure Determination. The diffraction images from the
single crystal of 1 were collected on Bruker APEX-II CCD three-
circled diffractometer equipped with graphite monochromated Mo−
Kα radiation (λ = 0.71073 Å) using ω- and φ-scans technique. The set
of reflection intensities was obtained by integration of image frames
using SAINT.23 The intensities were corrected for polarization and
Lorentz effects as well as for absorption (semiempirical method)
utilizing the SADABS program.23 The structure was solved by direct
methods and refined by full matrix least-squares on F2 in anisotropic
approximation for non-H atoms using SHELXTL.23 All hydrogen
atoms were treated geometrically depending on hybridization of the
parent atom and refined using the “riding model” with Uiso(H) =
1.5Uiso(C) for CH3 and Uiso(H) = 1.2Uiso(C) for other groups. At the
end of the structure refining, several electron density peaks, located
out of the main residue, with values from 1.81 to 1.00 e Å−3 remained.
All attempts to resolve the disorder of the mentioned water molecules
were unsuccessful. On this ground, the SQUEEZE routine from
PLATON was applied to modify reflection intensities corresponding
to disordered water molecules.24
ASSOCIATED CONTENT
* Supporting Information
■
S
Details of NMR, IR, TGA, ESI-MS studies; selected bond
lengths and angles for crystal structure of 1; list of CSD
refcodes for MaXb−X−McXd structures; selectivity parameters
in the oxidations of alkanes; discussion of the overoxidation
problem in the catalysis. This material is available free of charge
via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: pombeiro@ist.utl.pt (A.J.L.P.), kokozay@univ.kiev.ua
(V.N.K.).
■
C84H102Co Fe N O , M = 1979.24, tetragonal, I4, a = b =
̅
4
2
12 22
20.9562(16), c = 22.4758(18) Å, V = 9870.5(13) Å3, T = 173(2) K, Z
= 4, D(calc) = 1.332 Mg m−3, μ = 1.013 mm−1, F(000) = 4104.0, Θ =
1.37−26.40°, 15403 reflections collected, 9127 reflections unique (Rint
= 0.0301), 7521 reflections observed [I > 2σ(I)], R = 0.0393, wR =
0.1032, GoF = 1.005, largest difference peak and hole: 0.312 /−0.286 e
Å−3. CCDC reference number 781917.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Magnetic Measurements. The magnetic data were taken using a
SQUID magnetometer (Quantum Design MPMS-XL7) operating in
the RSO mode at B = 0.1 T. Raw magnetic susceptibility has been
corrected for the underlying diamagnetism using an estimate of χdia/
(10−9 m3 mol−1) = −5Mr [kg mol−1]. The susceptibility data were
converted to the effective magnetic moment μeff.
HF-EPR Spectroscopy. HF-EPR spectra were recorded with a
home-built spectrometer at the EMR facility of NHMF.25 The
instrument is a transmission-type device in which waves are
propagated in cylindrical light-pipes. The microwaves were generated
by a phase-locked oscillator (Virginia Diodes) operating at a frequency
of 13 1 GHz and generating its harmonics, of which the 4th, 8th,
16th, 24th, and 32nd were available. A superconducting magnet
(Oxford Instruments) capable of reaching a field of 17 T was
employed.
This work has been partially supported by the Foundation for
Science and Technology (FCT), Portugal, its PPCDT (FEDER
funded) program, and projects PTDC/QUI-QUI/102150/
2008 and PEst-OE/QUI/UI0100/2011; the Russian Founda-
tion for Basic Research (Grant 12-03-00084-a). HF-EPR and
Mossbauer spectra were taken at the NHMF which is funded
̈
by the NSF through the Cooperative Agreement No. DMR-
0654118, the State of Florida, and the DOE. The Mossbauer
̈
instrument was purchased using the User Collaboration Grant
Program UCGP 5064 funds awarded to A.O. Slovak grant
agency VEGA (projects 1/0052/11 and 1/0233/12) is
acknowledged for the financial support. Thanks are also due
to Dr. M. Candida Vaz (IST) and Dr. M. Conceica̧ o Oliveira
̂
̃
(IST) for the elemental and ESI-MS analyses, respectively.
̈
Mossbauer Spectroscopy. Mossbauer spectra were collected in
̈
constant acceleration mode using a spectrometer manufactured by
SEE Co Inc., Edina, Minnesota, which was equipped with a 57Co/Rh
gamma source purchased from Cyclotron Instruments, Mainz,
Germany. A Janis (Wilmington, Massachusetts) cryostat was employed
for low-temperature measurements.
Catalytic Oxidation of Alkanes. The catalyst precursor 1 and the
cocatalyst HNO3 were used in the form of stock solutions in
acetonitrile. Aliquots of these solutions were added to the reaction
mixtures in the oxidations of alkanes. The oxidation reactions were
typically carried out in air in thermostatted Pyrex cylindrical vessels
with vigorous stirring; total volume of the reaction solution was 5 mL.
(Caution! The combination of air or molecular oxygen and H2O2 with
ABBREVIATIONS
■
DMF, dimethylformamide; H2Sae, salicylidene-2-ethanolamine;
DMCH, dimethylcyclohexane; m-CPBA, m-chloroperoxyben-
zoic acid; TON, turnover number (mols of products/mol of
catalyst); TOF, turnover frequency (mols of products/mol of
catalyst per second or hour)
REFERENCES
■
(1) See, for example: (a) Barron, P. M.; Son, H.-T.; Hu, C.; Choe,
W. Cryst. Growth Des. 2009, 9, 1960. (b) Halper, S. R.; Do, L.; Stork, J.
L
dx.doi.org/10.1021/ic301460q | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
R.; Cohen, S. M. J. Am. Chem. Soc. 2006, 128, 15255. (c) Zhang, H.;
Yang, J.-H.; Shpanchenko, R. V.; Abakumov, A. M.; Hadermann, J.;
Clerac, C. R.; Dikarev, E. V. Inorg. Chem. 2009, 48, 8480.
(2) See, for reviews: (a) Andruh, M.; Costes, J.-P.; Diaz, C.; Gao, S.
Inorg. Chem. 2009, 48, 3342. (b) Tanase, S.; Reedijk, J. Coord. Chem.
Rev. 2006, 250, 2501. (c) McInnes, E. J. L.; Piligkos, S.; Timco, G. A.;
Winpenny, R. E. P. Coord. Chem. Rev. 2005, 249, 2577.
(3) See, for reviews: (a) Ritleng, V.; Chetcuti, M. J. Chem. Rev. 2007,
107, 797. (b) Wheatley, N.; Kalck, P. Chem. Rev. 1999, 99, 3379.
(c) Mukherjee, A.; Nembenna, S.; Sen, T. K.; Sarish, S. P.; Ghorai, P.
Kr.; Ott, H.; Stalke, D.; Mandal, S. K.; Roesky, H. W. Angew. Chem.,
Int. Ed. 2011, 50, 3968. (d) Mandal, S. K.; Roesky, H. W. Acc. Chem.
Res. 2010, 43, 248.
(4) (a) Que, L.; Tolman, W. B. Compr. Coord. Chem. II 2004, 8, 1.
(b) Doukov, T. I.; Iverson, T. M.; Seravalli, J.; Ragsdale, S. W.;
Drennan, C. L. Science 2002, 298, 567.
(5) (a) Nesterov, D. S.; Kokozay, V. N.; Dyakonenko, V. V.;
Shishkin, S. O.; Jezierska, J.; Ozarowski, A.; Kirillov, A. M.;
Kopylovich, M. N.; Pombeiro, A. J. L. Chem. Commun. 2006, 4605.
(b) Pombeiro, A. J. L.; Kirillov, A. M.; Kopylovich, M. N.; Kokozay, V.
N.; Nesterov, D. S. Patent PT 103526, 2007.
(6) Nesterov, D. S.; Kokozay, V. N.; Jezierska, J.; Pavlyuk, O. V.;
Boca, R.; Pombeiro, A. J. L. Inorg. Chem. 2011, 50, 4401.
(7) (a) Wang, W.-G.; Zhou, A.-J.; Zhang, W.-X.; Tong, M.-L.; Chen,
X.-M.; Nakano, M.; Beedle, C. C.; Hendrickson, D. N. J. Am. Chem.
Soc. 2007, 129, 1014. (b) Semenaka, V. V.; Nesterova, O. V.; Kokozay,
V. N.; Zybatyuk, R. I.; Shishkin, O. V.; Boca, R.; Shevchenko, D. V.;
Huang, P.; Styring, S. Dalton. Trans. 2010, 39, 2344.
(8) (a) Boskovic, C.; Bircher, R.; Tregenna-Piggott, P. L. W.; Gudel,
H. U.; Paulsen, C.; Wernsdorfer, W.; Barra, A.-L.; Khatsko, E.; Neels,
A.; Stoeckli-Evans, H. J. Am. Chem. Soc. 2003, 125, 14046. (b) Oshio,
H.; Hoshino, N.; Ito, T. J. Am. Chem. Soc. 2000, 122, 12602. (c) Oshio,
H.; Hoshino, N.; Ito, T.; Nakano, M. J. Am. Chem. Soc. 2004, 126,
8805. (d) Li, Y.; Wu, Q.; Lecren, L.; Clerac, R. J. Mol. Struct. 2008,
890, 339.
Y. Y.; Martins, L. M. D. R. S.; Monari, M.; Pandolfo, L.; Pettinari, C.;
Pombeiro, A. J. L. Eur. J. Inorg. Chem. 2009, 666. (d) Fernandes, R. R.;
Lasri, J.; Kirillov, A. M.; da Silva, M. F. C. G.; da Silva, J. A. L.; da Silva,
J. J. R. F.; Pombeiro, A. J. L. Eur. J. Inorg. Chem. 2011, 3781.
(e) Fernandes, R. R.; Lasri, J.; da Silva, M. F. C. G.; da Silva, J. A. L.; da
Silva, J. J. R. F.; Pombeiro, A. J. L. Appl. Catal., A 2011, 402, 110.
(17) (a) Shul’pin, G. B. Org. Biomol. Chem. 2010, 8, 4217. (b) Nizova,
G. V.; Krebs, B.; Suss-Fink, G.; Schindler, S.; Westerheide, L.;
̈
Gonzalez Cuervo, L.; Shul’pin, G. B. Tetrahedron 2002, 58, 9231.
(c) Kirillova, M. V.; Kuznetsov, M. L.; Romakh, V. B.; Shul’pina, L. S.;
Frausto da Silva, J. J. R.; Pombeiro, A. J. L.; Shul’pin, G. B. J. Catal.
́
2009, 267, 140. (d) Shul’pina, L. S.; Kirillova, M. V.; Pombeiro, A. J.
L.; Shul’pin, G. B. Tetrahedron 2009, 65, 2424. (e) Mandelli, D.;
Chiacchio, K. C.; Kozlov, Y. N.; Shul’pin, G. B. Tetrahedron Lett. 2008,
49, 6693.
(18) (a) Suss-Fink, G.; Nizova, G. V.; Stanislas, S.; Shul’pin, G. B. J.
̈
Mol. Catal. A: Chem. 1998, 130, 163. (b) Shul’pin, G. B.; Nizova, G.
V.; Kozlov, Y. N.; Gonzalez Cuervo, L.; Suss-Fink, G. Adv. Synth.
̈
Catal. 2004, 346, 317. (c) Kozlov, Y. N.; Nizova, G. V.; Shul’pin, G. B.
J. Mol. Catal. A: Chem. 2005, 227, 247.
(19) (a) Shul’pin, G. B.; Kirillova, M. V.; Kozlov, Y. N.; Shul’pina, L.
S.; Kudinov, A. R.; Pombeiro, A. J. L. J. Catal. 2011, 277, 164.
(b) Shul’pin, G. B.; Kozlov, Y. N.; Nizova, G. V.; Suss-Fink, G.;
̈
Stanislas, S.; Kitaygorodskiy, A.; Kulikova, V. S. J. Chem. Soc., Perkin
Trans. 2 2001, 1351.
(20) (a) Mirza, S. A.; Bocquet, B.; Robyr, C.; Thomi, S.; Williams, A.
F. Inorg. Chem. 1996, 35, 1332. (b) Rush, J. D.; Koppenol, W. H. J.
Am. Chem. Soc. 1988, 110, 4957. (c) Bull, C.; McClune, G. J.; Fee, J. A.
J. Am. Chem. Soc. 1983, 105, 5290. (d) Shinohara, N.; Matsufuji, S.;
Okubo, W. Polyhedron 1991, 10, 107.
(21) (a) Fokin, A. A.; Schreiner, P. R. Chem. Rev. 2002, 102, 1551.
(b) Newhouse, T.; Baran, P. S. Angew. Chem., Int. Ed. 2011, 50, 3362.
(c) Buda, F.; Ensing, B.; Gribnau, M. C. M.; Baerends, E. J. Chem.
Eur. J. 2001, 7, 2775. (d) Gozzo, F. J. Mol. Catal. A: Chem. 2001, 171,
1.
(22) Shul’pin, G. B.; Golfeto, C. C.; Suss-Fink, G.; Shul’pina, L. S.;
̈
(9) Spectral Database for Organic Compounds (SDBSWeb): http://
riodb01.ibase.aist.go.jp/sdbs/; National Institute of Advanced Indus-
trial Science and Technology (AIST), Japan.
(10) CSD (version 5.32; November 2011): Allen, F. H. Acta
Crystallogr. 2002, B58, 380.
(11) Nesterov, D. S.; Kokozay, V. N.; Skelton, B. W. Eur. J. Inorg.
Chem. 2009, 5469.
(12) Konno, T.; Tokuda, K.; Sakurai, J.; Okamoto, K. Bull. Chem. Soc.
Mandelli, D. Tetrahedron Lett. 2005, 46, 4563.
(23) SADABS, SAINT, SHELXTL and SMART; Bruker AXS Inc.:
Madison, WI, 2003.
(24) (a) Spek, A. L. Acta Crystallogr. 1990, A46, C34. (b) Spek, A. L.
J. Appl. Crystallogr. 2003, 36, 7. (c) Spek, A. L. PLATON, A
Multipurpose Crystallographic Tool; Utrecht University: Utrecht, The
Netherlands, 2000.
(25) Hassan, A. K.; Pardi, L. A.; Krzystek, J.; Sienkiewicz, A.; Goy, P.;
Rohrer, M.; Brunel, L.-C. J. Magn. Reson. 2000, 142, 300.
Jpn. 2000, 73, 2767.
(13) (a) Owen, J. J. Appl. Phys. 1961, 32, 213S. (b) Abragam, A.;
Bleaney, B. In Electron Spin Resonance of Transition Ions; Clarendon
Press: London, U.K., 1970. (c) Bencini, A.; Gatteschi, D. In EPR of
Exchange Coupled Systems; Springer Verlag: Berlin, Germany, 1990.
(d) Okamura, M. Y.; Hoffman, B. M. J. Chem. Phys. 1969, 51, 3128.
(e) Ozarowski, A.; McGarvey, B. R.; Drake, J. E. Inorg. Chem. 1995, 34,
5558. (f) Chygorin, E. N.; Nesterova, O. V.; Rusanova, J. A.; Kokozay,
V. N.; Bon, V. V.; Boca, R.; Ozarowski, A. Inorg. Chem. 2012, 51 (1),
386. (g) Ozarowski A. et al., in preparation. (h) ter Heerdt, P.; Stefan,
M.; Goovaerts, E.; Caneschi, A.; Cornia, A. J. Magn. Reson. 2006, 179,
29.
(14) Semenaka, V. V.; Nesterova, O. V.; Kokozay, V. N.;
Dyakonenko, V. V.; Zubatyuk, R. I.; Shishkin, O. V.; Boca, R.;
Jezierska, J.; Ozarowski, A. Inorg. Chem. 2010, 49, 5460.
(15) (a) Shul’pin, G. B. J. Mol. Catal. A: Chem. 2002, 189, 39.
(b) Shul’pin, G. B. C. R. Chim. 2003, 6, 163. (c) Shul’pin, G. B. Mini-
Rev. Org. Chem. 2009, 6, 95. (d) Shul’pin, G. B.; Kozlov, Y. N.;
Shul’pina, L. S.; Kudinov, A. R.; Mandelli, D. Inorg. Chem. 2009, 48,
10480. (e) Shul’pin, G. B.; Kozlov, Y. N.; Shul’pina, L. S.; Petrovskiy,
P. V. Appl. Organomet. Chem. 2010, 24, 464.
(16) See, for example: (a) Kirillov, A. M.; Kopylovich, M. N.;
Kirillova, M. V.; Haukka, M.; da Silva, M. F. C. G.; Pombeiro, A. J. L.
Angew. Chem., Int. Ed. 2005, 44, 4345. (b) Trettenhahn, G.; Nagl, M.;
Neuwirth, N.; Arion, V. B.; Jary, W.; Pochlauer, P.; Schmid, W. Angew.
Chem., Int. Ed. 2006, 45, 2794. (c) Di Nicola, C.; Garau, F.; Karabach,
M
dx.doi.org/10.1021/ic301460q | Inorg. Chem. XXXX, XXX, XXX−XXX