Tuning the conversion of cyclohexane into cyclohexanol/one by molecular dioxygen, protons and reducing agents at a single non-porphyrinic iron centre and chemical versatility of the tris(2-pyridylmethyl)amine TPAFe IICl2 complex in mild oxidation chemistry

Article abstract of DOI:10.1039/c0dt00756k

We report that the oxygen sensitivity of some Fe(ii) complexes with tripodal ligands can be used, with benefit, in the oxidation of cyclohexane under mild conditions. Depending on the solvent, two very different reaction pathways are involved, which share the coordination of O₂ to the metal as the common initial step. We have synthesized a series of α-chlorinated tripods in the tris(2-pyridylmethyl)amine series Cl _nTPA (n = 1-3) and fully characterized the corresponding FeX ₂ complexes (X = Cl, CF₃SO₃). The single-crystal X-ray structure analyses of the FeCl₂ complexes are reported. In CH₃CN, the FeCl₂ complexes react smoothly with O₂, whereas the Fe(CF₃SO₃)₂ complexes are non-sensitive. In CH₃CN, the reaction of the oxygen-sensitive Cl_nTPAFeCl₂ (n = 0-3) with O₂, acetic acid and zinc amalgam, in the presence of cyclohexane, affords a mixture of cyclohexanol/one in an ≈ ol/one ratio of 3.1 and a selectivity of the C3°/C2° in the adamantane conversion that is consistent with a metal-oxo based oxidation. Limited efficiency (≈ 2 TON) was observed for the parent TPAFeCl₂ complex and Cl₁TPAFeCl₂, whereas both other complexes turned out to be poorly active. The TPAFeCl₂ complex was used to address mechanistic questions: when the reaction was carried out in pyridine, the ol/one ratio shifted to 0.15 while efficiency was improved by 7-fold. In pyridine and in the presence of a spin trap (DMPO), the radical-based character of the reaction was definitely established, by contrast with acetonitrile, where no oxygenated radicals were detected. Thus, the reactivity differences arise from involvement of two distinct active species. The dichotomous radical/biomimetic pathway is discussed to interpret these results.

Full text of DOI:10.1039/c0dt00756k

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Dalton
Transactions
Cite this: Dalton Trans., 2011, 40, 92
www.rsc.org/dalton
PAPER
Tuning the conversion of cyclohexane into cyclohexanol/one by molecular
dioxygen, protons and reducing agents at a single non-porphyrinic iron centre
II
and chemical versatility of the tris(2-pyridylmethyl)amine TPAFe Cl2
complex in mild oxidation chemistry†
a
b
a
a
Hassen Jaafar, Bertrand Vileno, Aurore Thibon and Dominique Mandon*
Received 30th June 2010, Accepted 28th September 2010
DOI: 10.1039/c0dt00756k
We report that the oxygen sensitivity of some Fe(II) complexes with tripodal ligands can be used, with
benefit, in the oxidation of cyclohexane under mild conditions. Depending on the solvent, two very
different reaction pathways are involved, which share the coordination of O
common initial step. We have synthesized a series of a-chlorinated tripods in the
tris(2-pyridylmethyl)amine series Cl TPA (n = 1–3) and fully characterized the corresponding FeX
complexes (X = Cl, CF SO ). The single-crystal X-ray structure analyses of the FeCl complexes are
reported. In CH CN, the FeCl complexes react smoothly with O , whereas the Fe(CF SO complexes
are non-sensitive. In CH CN, the reaction of the oxygen-sensitive Cl TPAFeCl (n = 0–3) with O , acetic
acid and zinc amalgam, in the presence of cyclohexane, affords a mixture of cyclohexanol/one in an ª
2
to the metal as the
n
2
3
3
2
3
2
2
3
)
3 2
3
n
2
2
◦
◦
ol/one ratio of 3.1 and a selectivity of the C3 /C2 in the adamantane conversion that is consistent
with a metal-oxo based oxidation. Limited efficiency (ª 2 TON) was observed for the parent TPAFeCl
2
complex and Cl
1
TPAFeCl
2
, whereas both other complexes turned out to be poorly active. The
TPAFeCl complex was used to address mechanistic questions: when the reaction was carried out in
2
pyridine, the ol/one ratio shifted to 0.15 while efficiency was improved by 7-fold. In pyridine and in the
presence of a spin trap (DMPO), the radical-based character of the reaction was definitely established,
by contrast with acetonitrile, where no oxygenated radicals were detected. Thus, the reactivity
differences arise from involvement of two distinct active species. The dichotomous radical/biomimetic
pathway is discussed to interpret these results.
8
,9
1
. Introduction
involving various metal centres, the biomimetic activation of
the C–H bond by iron complexes emerged as a very attractive
Setting up mild conditions in oxidation processes involving
molecular dioxygen remains, more than 25 years after the first
report of the “Gif reaction”, a real challenge in the current
economical context. In biological systems, dioxygen coordinates
to mono- and dinuclear non-heme iron-containing centres present
at active sites of proteins and is activated in such a way that major
9–11
approach.
The use of small coordination complexes is indeed
rising as a promising approach: recent developments shed light
on the ability of small iron complexes to perform site- and
stereospecific oxidation of various substrates, ranging from simple
In spite of the restricted exceptions
involving dinuclear iron-containing complexes,
oxide is generally used as the oxidant. Molecular oxygen, which is
involved in many biological processes, and is a universally available
and safe resource, would be more convenient to use. However, its
activation by cheap metal complexes is far from been scientifically
solved. Thus, there is still a wide open investigation field in this
area, from both academic and technological points of view.
1
–3
12–15
alkanes to terpenoids.
16,17
hydrogen per-
4–7
transformations can be realized. Following extensive studies
a
Laboratoire de Chimie Biomim e´ tique des M e´ taux de Transition, Institut de
Chimie, UMR CNRS no. 7177, Universit e´ de Strasbourg, 4 rue Blaise Pascal,
B.P. 1032, F-67070 Strasbourg cedex, France. E-mail: mandon@chimie.u-
strasbg.fr; Fax: 33-(0)368-851-438; Tel: 33-(0)368-851-537
b
Laboratoire des Propri e´ t e´ s Optiques & Magn e´ tiques des Architectures
◦
Mol e´ culaires, Institut de Chimie, UMR CNRS n 7177, Universit e´ de
In the course of our investigations with FeCl complexes of
tripodal ligands [the parent tris(2-pyridylmethyl)amine (TPA)
2
Strasbourg, 1 rue Blaise Pascal, B.P. 296/R8, F-67070, Strasbourg cedex,
France
†
Electronic supplementary information (ESI) available: Spectroscopic
data of all new compounds described in this study. Oxygenation curves
of compound Cl TPAFeCl and kinetic parameters. CCDC reference
ligand or some a-substituted tripods], we have already reported
that the O
2
mediated conversion of FeCl complexes into m-
2
1
2
oxo diferric stable complexes is a general reaction involving the
numbers 751029–751031. For ESI and crystallographic data in CIF or
other electronic format see DOI: 10.1039/c0dt00756k
18,19
coordination of dioxygen to the iron centre.
9
2 | Dalton Trans., 2011, 40, 92–106
This journal is © The Royal Society of Chemistry 2011

View Article Online
the metal centre, followed by two different reaction pathways
(
“biomimetic” vs. “Gif”), depending on the presence (or not) of
pyridine.
In a context where more and more studies describe the spec-
troscopic generation of potentially reactive intermediates from O
2
23,24
and protons,
our work represents one of the very few functional
examples of biomimetic C–H hydroxylation of an inert substrate,
by well-defined mononuclear non-porphyrinic iron complexes in
mild conditions. In addition, we provide a positive approach to
the long-time controversial question regarding the activation of
We demonstrated that molecular oxygen does coordinate to
the metal centre by a mechanism involving dissociation of the
19
Fe–Cl bond. Additionally, with ligands substituted by non-
innocent functional groups, in aprotic medium and in the absence
of an exogenous substrate, intramolecular reactions mimicking
the transformations encountered in biological systems (cleavage
(II)
2
2
O by the Fe /AcOH/Zn system.
19
20
21
of esters, oxidative dehydrogenation, O-demethylation ) were
found to take place in the vicinity of the metal centre.
2
. Results
Synthesis
New a-chlorosubstituted derivatives Cl1–3TPA were synthesized
following the classical pathway from 2-chloro-6-bromomethyl
18
pyridine and corresponding amines. Synthetic details and char-
acterization can be found in the experimental section and as ESI.†
These ligands were further complexed to FeCl
according to well-known procedures
2
and Fe(OTf)
2
18,25,26
Coordination chemistry
The process involved in these reactions requires activation
of dioxygen by the iron centre, leading to generation of an
active species. In our previously published articles reporting
intramolecular reactivities, we tended to favour a superoxide-
The FeCl complexes were obtained as yellow solid, thermally
stable, yet oxygen-sensitive compounds. They were characterized
by single-crystal X-ray diffraction and their molecular structure
2
is displayed in Fig. 1. Whereas Cl
1
TPAFeCl
2
and Cl
3
TPAFeCl
2
based mechanism resulting from inner sphere reduction of O
2
by
display classical distorted octahedral and trigonal bipyramidal
19,21
the metal.
We could not however rule out a metal-oxo based
geometries, respectively, the iron centre in Cl
2
TPAFeCl
2
is only
˚
pathway, in which the active species would be formed by homolytic
cleavage of the O–O bond in a putative transient dinuclear species.
If this were the case, an interesting reactivity towards exogenous
substrates might be observed. In any event, it was tempting to
address the question of whether or not any active species could
perform C–H activation on an added substrate.
moderately bound by the third pyridine, with Fe–N4 = 2.50 A.
In an ideal octahedral symmetry, the plane defined by the
tertiary amine and the two chloride atoms, <NamineCl1Cl2, should
include, in the trans position to one the chloride ions (here Cl1),
one pyridine nitrogen atom. This is, however, generally not the
case; as already stated above, the classical picture is that of
a distorted octahedral geometry, since the coordinated ligand
defines a five-membered metallacycle with sharper angles, close
The question of sensitivity of the complexes towards dioxygen
is a key point. In precedent studies, FeCl
2
complexes with a-
◦
fluorinated tris(2-pyridylmethyl)amine tripods were found to react
to 76 .
18,22
smoothly with O
2
.
We thought that increasing the steric
Cl
1
TPAFeCl
2
does not make exception to the rule, with the N3
˚
hindrance around the metal centre might increase the lifetime
of any potentially active species, allowing further chemistry to be
done and decided to prepare the series of a-chlorinated ligands
atom standing 0.79 A above the <N1Cl1Cl2 plane, the Fe atom
˚
being displaced by 0.19 A. In Cl
the trans pyridine atom from the mean plane is only 0.18 A, the Fe
being displaced by 0.10 A. This is certainly due to the elongation
of the Fe–N4 bond, which by mechanical effect decreases the
stress of the ligand, allowing a more classical arrangement around
the metal centre in Cl
in Cl TPAFeCl and Cl
structure of Cl TPAFeCl
between that of Cl TPAFeCl
In solution, the UV-vis spectrum of Cl
2
TPAFeCl
2
, the displacement of
˚
˚
Cl
of the corresponding FeCl
spectroscopic characterization of the Fe(OTf)
We then report that some of the oxygen-sensitive FeCl
n
TPA (n = 1, 2, 3). Herein we describe at first the structures
complexes and the preparation and
analogues.
2
2
2
com-
2
TPAFeCl
TPAFeCl
can be considered as being intermediate
and Cl TPAFeCl
TPAFeCl displays an
2
. Fig. 2 compares the situation
plexes can, in acetonitrile, promote the conversion of cyclohex-
ane into cylohexanol/one in mild conditions, yet with limited
1
2
2
2
. Thus, in the solid state, the
2
2
efficiency, using O
2
, a reducing agent and a source of protons.
1
2
3
2
.
Using TPAFeCl as a comparison tool, we finally demonstrate
2
1
2
the versatility of our system, the reaction profile being highly
solvent-dependent. Whereas a classical “biomimetic” reactivity
is observed in acetonitrile, pyridine promotes an oxygen-radical-
based pathway, leading to the formation of ketone as the main
product. On the basis of these results, and considering our
previously reported work, a general reaction scheme is proposed,
Fe → Py MLCT absorption at l = 392 nm, with the molecular
3 -1 2
extinction coefficient (e = 1.2·10 mmol cm ) supporting tetraden-
tate coordination of the tripod. Also, the presence of relatively
1
sharp signals in the H NMR spectra for this paramagnetic species
confirms that the geometry as observed in the solid state is retained
25
in solution. By contrast, the spectroscopic data of solutions of
involving a first common step during which O
2
coordinates to
Cl
3
TPAFeCl
2
indicates hypodentate coordination of the tripod in
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 92–106 | 93

View Article Online
Fig. 1 Mercury drawing of Cl1–3TPAFeCl
2
complexes. Left Cl
1
TPAFeCl
2
, Middle Cl
2
TPAFeCl
2
, Right Cl
3
TPAFeCl
2
. Details of the bond lengths are
available from the cif files provided in the ESI.† All Fe–N bond lengths lie above 2.1 A˚ .
1 2 2 2
Fig. 2 Structural arrangement around the metal in complexes Cl TPAFeCl and Cl TPAFeCl .
2
2,27
solution, as already observed with F
3
TPAFeCl
2
.
The MLCT
In the UV-vis spectra in CH
found at l = 354 nm (e
3
CN, the MLCT absorptions were
3
-1
2
absorption is almost nonexistent in UV-vis spectroscopy and
extremely broad signals are detected in the H NMR spectra. All
data are given in the ESI.† Whereas in the solid state, the geometry
1
1
= 1.3·10 mmol cm ), l
2
= 351 nm (e =
= 0.6·10 mmol cm )
2
1
3
-1
2
3
-1
2
0.6·10 mmol cm ) and l
3
= 344 nm (e
complexes. In the H NMR spectra, all
3
1
for the Cl1–3TPAFe(OTf)
2
around the metal in Cl
mediate between that found in Cl
solution studies unequivocally support hypodentate coordination
2
TPAFeCl
2
can be considered as inter-
TPAFeCl and Cl TPAFeCl ,
of the complexes displayed relatively sharp and paramagnetically
shifted signals over the +130 to +20 ppm range, in line with a
1
2
3
2
1
9
high-spin state for the iron centre. With Cl
1
TPAFe(OTf) , the F
2
of the tripod in this complex with a reduced intensity Fe → Py
NMR spectrum consisted of a sharp (Dn1/2 = 35 Hz) signal at
3
-1
2
19
MLCT (e = 0.6·10 mmol cm ) absorption at l = 373 nm and
d = -72.1 ppm. With Cl
2
TPAFe(OTf)
2
, the F NMR spectrum
1
broad resonances in the H NMR spectra, qualitatively similar to
exhibited a slightly broadened (Dn1/2 = 90 Hz) signal at d =
-74.1 ppm, whereas more consequent, but still very limited,
broadening (Dn1/2 = 230 Hz) was observed at d = -74.2 ppm in the
25
those already reported for Br
The Fe(OTf) complexes were obtained as solid materials, for
which the colour was found to be highly dependent on the substitu-
2
TPAFeCl
2
.
2
spectrum of Cl
support the presence of uncoordinated triflate ions in a CD
solution, although minor broadening may arise from a limited
3
TPAFe(OTf)
2
. All these data, shown in the ESI,†
tion of the ligand: a dark red solid was obtained for TPAFe(OTf)
2
,
3
CN
the colour becoming beige-yellow in Cl
1
TPAFe(OTf)
2
, then grey
30
in Cl TPAFe(OTf) and finally light grey in Cl
2
2
3
TPAFe(OTf)
2
.
dynamic exchange process in solution.
28
The compounds turned out to be very hygroscopic, yet they
were stable to dioxygen, even in acetonitrile solutions. They were
Reactivity
1
19
characterized by UV-vis, H and F NMR spectroscopies. All
studies were performed in CH CN, the solvent in which further
Complexes–O
2
.
The Cl
n
TPAFeCl
2
complexes react with O .
2
3
reactivity studies were carried out. It should be mentioned here
that when dissolved in the coordinating solvents, the triflate anions
are generally displaced by the solvent molecules, thus yielding
For the six-coordinate-complexes, the sensitivity towards dioxygen
can easily be quantified upon monitoring the UV-vis changes
at a single wavelength, the changes leading to the formation of
the final m-oxo species. In general, two spectroscopic steps are
26,29
dicationic species.
9
4 | Dalton Trans., 2011, 40, 92–106
This journal is © The Royal Society of Chemistry 2011

View Article Online
Table 1 Turnover numbers (product equiv./complex equiv.) observed in the H
2
O
2
reaction. Cyclohexane : H
2
O
2
: complex ratio 500 : 500 : 1. Addition
time of H
2
O
2
: 4 h 30 min. All reactions were carried out in triplicate. When the reaction is carried out under O
2
, the profile is not significantly modified
Complex
X
Cyclohexanol
Cl
Cyclohexanone
Cl
Total
Cl
A/K
OTf
OTf
OTf
Cl
OTf
TPAFeX
2
18.9
18.1
9.3
1.2
3.5
25.0
15.9
9.8
2.7
13.5
9.8
7.8
5.7
1.1
2.2
7.4
3.9
4.6
1.2
4.3
28.7
25.9
15.0
2.3
32.4
19.8
14.4
3.9
1.9
2.3
1.6
1.1
1.6
3.4
3.8
2.1
2.2
3.1
Cl
1
Cl
2
Cl
3
TPAFeX
TPAFeX
TPAFeX
2
2
2
FeX
2
5.7
17.8
1
8
observed. With Cl
0
1
TPAFeCl
2
, we obtained k
1
= 0.38 and k
2
=
Under the conditions of (i) and (ii), only trace amounts of the
oxidation products could be detected, indicating that no tractable
reaction occurred.
-
1
.052 h , i.e. in the same order of magnitude as the values
found for the TPAFeCl
18
2
and FTPAFeCl complexes. The five-
2
coordinate complexes generally exhibit transformation rates that
In the simultaneous presence of acetic acid and zinc amalgam,
i.e. under the conditions of (iii), the conversion of cyclohexane
into a mixture of cyclohexanol and cyclohexanone was observed,
as displayed in Scheme 2.
18
are estimated to go c.a. 10–20 times faster. This is the case with
Cl TPAFeCl and Cl TPAFeCl , for which total conversion was
2
2
3
2
observed over 6 and 3 h, respectively.
II
Fe triflate complexes were also studied and exhibited no, or
extremely limited, reactivity vs. O
2
when dissolved in CH CN. No
3
tractable change could be observed in the UV-vis spectra, which
remained unaffected even after several days of having dioxygen
bubbled through the solution.
Complexes–H
2
O –cylcohexane. The standard conditions for
2
Scheme 2 conversion of cyclohexane into cyclohexanol/one in the
presence of O , AcOH and Cl TPAFeCl (0<n<3).
biomimetic oxidation reactions generally involve peroxides as the
oxygen donors. We checked the activity of our compounds in the
presence of hydrogen peroxide. Also, we compared the activities
2
n
2
This conversion, however, is limited in yield: even if TPAFeCl
and Cl TPAFeCl give more than a stoichiometric transforma-
2
of the Cl0–3TPAFeCl
2
and Cl0–3TPAFe(OTf)
2
complexes, as weakly
1
2
coordinating ligands are generally used as metal counter anions in
biomimetic catalysis.
tion based on the complex, this reaction can not be described
as “catalytic” in the commonly accepted sense of the term.
As indicated in Table 2, a significant amount of substrate is
11
As can be seen in Scheme 1 and Table 1, the conversion occurred
when either type of complex was used as the catalyst. The triflate-
coordinated compounds exhibited a slightly better reactivity than
the chloride coordinated complexes. These results, which will be
discussed later in this article, can be easily compared with those
converted by the parent TPAFeCl
2
complex and Cl
1
TPAFeCl ,
2
both compounds displaying a pseudo-octahedral geometry in
solution, for which a smooth reactivity towards dioxygen was
measured in the absence of the substrate. The five-coordinate
11,31,32,45,48
described in previous reports.
complexes Cl
more reactive towards dioxygen, turned out to display a weaker
reactivity, comparable to that of FeCl , when this salt was used
2
TPAFeCl
2
and Cl
3
TPAFeCl , although kinetically
2
2
alone in blank experiments. In the reaction conditions, the stability
of these compounds turned out to be excellent and the observed
lack of reactivity cannot be due to decomposition, indeed, m-oxo
diferric compounds are generally systematically recovered upon
oxygenation of the medium (in the presence or absence of acetic
acid). Finally, a blank experiment was carried out in the absence
of the iron complex and no conversion was observed.
Scheme 1 conversion of cyclohexane into cyclohexanol/one by the
conventional H method.
2
O
2
Complexes–ZnHg–AcOH–O
TPAFeCl complexes being sensitive to dioxygen, their reactivity
was tested in the presence of cyclohexane. The already known
parent, the oxygen-sensitive TPAFeCl complex, was also studied.
.5 mg of each complex in 2 mL of dry and degassed acetonitrile
2
–cylclohexane. The
Cl0–3
-
The mass balance, the mass spectroscopy analyses and the
absence of other signals at long retention times in GC confirmed
2
Table 2 Turnover numbers (product equiv./complex equiv.) ob-
served in the O –Zn–Hg reaction. Cyclohexane : AcOH : complex ratio
2
2
1
1
100 : 300 : 1. A/K is the alcohol/ketone ratio. Reaction time: 5 h. All
was reacted with 0.35 mL of cyclohexane in the presence of
dioxygen for the duration of 5 h. Three different conditions
were investigated: (i) the medium “as is” was stirred during the
reaction time, (ii) a drop of zinc amalgam was introduced before
reactions were carried out at least five times
Complex
TPAFeCl
Cyclohexanol
Cyclohexanone
Total
A/K
2
1.6
0.8
0.4
0.3
0.2
0.5
0.3
0.2
0.1
0.1
2.1
1.1
0.5
0.4
0.3
3.2
2.7
2.0
3.0
2.0
the addition of O
amalgam were introduced before the addition of O
2
and (iii) 50 mL of acetic acid and a drop of zinc
. At the end
1
Cl TPAFeCl
2
2
2
Cl
Cl
2
TPAFeCl
TPAFeCl
2
3
of the reaction time the medium was filtered and analyzed by gas
chromatography.
FeCl
2
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 92–106 | 95

View Article Online
Table 3 Turnover numbers (product equiv./complex equiv.) observed in
the O –AcOH–red reaction of adamantane with TPAFeCl in CH CN
and pyridine. 3 /2 = 3 ¥ (1-adamantanol)/(2-adamantanol + 2-
adamantanone). The reactions were carried out in triplicate
Table 4 Turnover numbers (product equiv./complex equiv.) observed
2
2
3
2 2 2
in the O –AcOH–red reaction with TPAFeCl and FeCl , as a func-
◦
◦
tion of the reaction solvent (identical reaction conditions). Entry B:
Cyclohexane : AcOH : pyridine : complex ratio 1100 : 300 : 300 : 1. All reac-
tions were carried out at least five times
◦
◦
1
-adamantanol 2-adamantanol 2-adamantanone 3 /2
entry Complex
C6-ol C6-one Total A/K
CH
Pyridine 2.6
3
CN 2.5
0.7
0.7
0.6
3.5
5.8
1.8
A
B
C
D
E
TPAFeCl
TPAFeCl
TPAFeCl
2
2
2
–H
–CH
3
CN
1.6
0.5
1.6
13.1
0.1
2.8
2.1
3.1
15.1 0.15
0.3
3.3
3.2
0.9
3
CN + PyHOAc 1.4
–pyridine
CN
–pyridine
2.0
0.2
0.5
that cyclohexanol and cyclohexanone were the only reaction
products.
FeCl
FeCl
2
–CH
3
3.0
0.2
2
Focus on the TPAFeCl
2
–ZnHg–AcOH–O
2
catalytic system.
The parent TPAFeCl being the most reactive complex, additional
2
of ketone was clearly favoured in pyridine with a ketone/alcohol
ratio of above 6 (A/K = 0.15).
detailed studies were carried out with this compound. Cyclohex-
ane was replaced by cyclohexanol as the substrate and the reaction
was carried out in identical conditions. The GC analysis revealed
that no trace of cyclohexanone could be detected. The composition
of the medium remained unchanged.
In acetonitrile, the presence of an equimolar mixture of pyridine
and acetic acid in limited amounts resulted in a slight increase of
the activity, with, however, a significant change in the selectivity.
To gain an insight into the reaction mechanisms, we carried out
the oxygenations in acetonitrile and in pyridine in the presence of
5,5-dimethyl-1-pyrroline-N-oxide (DMPO), a spin trap commonly
◦
◦
Additionally, the C3 /C2 (CTertiary/CSecondary) selectivity was
tested in the oxidation of adamantane. The results are shown in
Table 3. Although these are not as high as the reported values
∑
-
∑
used in EPR to detect the presence of superoxide O
2
or OH
36–38
obtained from H
salts with related ligands,
2
O
2
as the oxygen donor and triflate or perchlorate
radicals,
as shown in Scheme 4.
11,31
◦
◦
the 3 /2 ratio = 5.8 with the O
2
–
DMPO was added to the substrate-containing solution and O
2
AcOH–ZnHg system significantly differs from that obtained when
the HO radical is used.
was subsequently added to this medium. The medium was stirred
for 5 min and then a 50 mL aliquot of the solution was introduced
into a glass capillary and immediately positioned in the EPR
cavity. The measurements were performed at room temperature.
The results are displayed in Scheme 4. No significant signal was
detected, in any case, in the absence of oxygen. Unambiguously,
no significant persistent radical species could be detected in the
acetonitrile after the introduction of oxygen. By contrast, in
pyridine an intense signal was detected, corresponding to the
reactive oxygen species trapped by the DMPO, indicating direct
∑
33–35
With TPAFeCl
2
only, another set of experiments was defined, in
IV
conditions comparable to those involved in the Gif systems, i.e.
5
1
with pyridine used as the reaction solvent. In pyridine d , the H
NMR spectrum of TPAFeCl was not significantly different than
in CH CN, confirming the high-spin state of the metal centre in
this solvent (the H-NMR and UV-vis data are given in the ESI†).
The comparison between CH CN and pyridine as the solvent is
2
3
1
3
spectacular in terms of the product distribution, as can be seen in
Scheme 3.
∑
-
∑
evidence of the presence of the superoxide anion O
2
, OH radical
3
6–40
Whereas a low efficiency was observed in CH
3
CN (see above,
and minor amounts of some thermal conversion products.
and entry A in Table 4), the use of pyridine as the solvent allows
a very significant increase in the conversion rate, switching the
turnover from 2 to 15 (entry C in Table 4). Also, the production
A low temperature EPR search was performed over a broader
field and no metal-related paramagnetic compounds, such as a
ferric hydroperoxide species for instance, were detected. Whereas
Scheme 3 Changing the solvent from acetonitrile to pyridine dramatically affects the conversion rate and distribution of the reaction products.
6 | Dalton Trans., 2011, 40, 92–106
9
This journal is © The Royal Society of Chemistry 2011

View Article Online
Scheme 4 Room temperature EPR spectra obtained when the reaction was carried out in the presence of DMPO as a spin trap. DMPO : complex ratio
5
0 : 3. Experiments were performed prior (green trace) and following (red trace) the oxygen bubbling.
we did not expect to detect the presence of such species in pyridine,
we were not surprised to miss them in acetonitrile either, since they
are known to be unstable in these conditions.
the MLCT signals. With Cl
is no doubt upon examination of the spectroscopic data that the
geometry as seen in the solid state is retained in solution.
1
TPAFeCl
2
and Cl
3
TPAFeCl , there
2
45
1
The case of Cl
2
TPAFeCl
2
is different, the H NMR and UV-
vis data are qualitatively identical to those already reported for
Br TPAFeCl , in which the TBP geometry is retained in solution.
3
. Discussion
2
2
Thus two distinct forms are found, in the solid state and in
solution, respectively, as drawn in Scheme 5.
Coordination chemistry
We have already reported the structural characterization of a
18,22
41
series of FeCl
2
complexes with fluorine-
and methyl- a-
substituted tripods in the tris(2-pyridylmethyl)amine family of
ligands. Monosubstitution at the a position generally affords six-
coordinate iron centres, whereas trisubstitution leads to hypoden-
tate coordination of the tripod, the geometry becoming trigonal
bipyramidal (TBP). The molecular structures of Cl
1
TPAFeCl
2
and Cl TPAFeCl exhibit comparable features to those derived
3
2
from complexes with mono- and trisubstituted tripods. With a-
disubstituted tripods, we reported that a progressive increase of
the size of the substituent causes, when steric hindrance remains
moderate (fluoro-substituted tripods), trans-equatorial distortion
Scheme
Cl TPAFeCl .
2
5 Solid state vs. solution coordination equilibrium in
2
18
within the coordination polyhedron. For more consequent repul-
sions (methyl-substituted tripods), there is a noticeable elongation
41
of the metal to pyridine distances. In extreme conditions,
with bulky substituents such as in Br TPAFeCl , this elongation
In the present study, the Fe(OTf)
in view of their further reaction with hydrogen peroxide, for
the purpose of comparison with our O –ZnHg system. The
preparation of such compounds does not lead to any particular
comment. We used relatively hindered substituents in the a-
position and, as expected in such cases, we obtain high-spin Fe(II)
compounds, with paramagnetic signals in the H NMR spectra
supporting pseudo-octahedral geometries. Dissolution of the
Fe(OTf) complexes with tripodal ligands in acetonitrile generally
2
complexes were prepared
2
2
becomes high enough to unhook one of the substituted pyridines
and decoordination occurs, leaving the metal centre in a TBP
2
2
5
geometry. In the solid state the Cl
2
TPAFeCl
2
complex lies in
˚
29
the intermediate situation, with Fe–N4 = 2.50 A, the pyridine
remains coordinated, yet weakly bound and more flexible. As a
mechanical effect, the equatorial Cl1, Cl2, N1 and N2 atoms are
almost coplanar.
1
2
A careful analysis of the spectroscopic data allows an estimation
of the geometry in solution. Narrow signals in the H NMR
affords dicationic species, resulting from the displacement of the
triflate ions by acetonitrile. The probe to confirm this hypothesis
is F NMR spectroscopy, a well documented technique with
1
1
9
spectra and a pronounced MLCT absorption in the UV-vis
spectroscopy generally reflect six-coordination of the tripod in
respect to the coordination of triflate ions to paramagnetic iron
2
5,41
30,42,43
a pseudo-octahedral geometry.
Switching to TBP geometry
centres.
With chemicals shifts in the range -72 to -74 ppm
TPA
1
induces a broadening of the signals in the H NMR spectra and
a noticeable (sometimes dramatic) decrease in the intensity of
and a maximum signal broadening of 230 Hz for the Cl
3
ligand, it seems obvious that the triflate ions definitely remain
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 92–106 | 97

View Article Online
uncoordinated. Thus, in CH
3
CN solution, these complexes can be
Reactivity with O
2
formulated as [Cl TPAFe(CH
n
3
CN)
2
](OTf) .
2
Complexes alone. We previously reported that FeCl
plexes within the series of a-fluoro substituted tripods react with
to afford diferric m-oxo species and that subsequent recovery of
2
com-
O
2
Reactivity with H
2
O
2
the ferrous compound can be achieved by reduction with zinc
+
The reactivity of the [TPAFeCl
2
] cation vs. H
2
O
2
in the presence
amalgam, allowing several oxygenation–reduction cycles to be
4
4
18,21
of cyclooctane has already been reported. With respect to the
present work, there was a significant difference in the oxidation
state of the iron centre, which was +3 in that case, and the use
of pyridine, chosen because of its efficiency as a solvent in the
Gif systems. Even so, a lower activity than that observed in the
present study was observed. In the present study, we re-investigated
completed.
Thus, although sensitive, our complexes turned out
not to decompose under oxygen. For six-coordinate complexes, we
reported that O does coordinate to the metal centre, through
2
a mechanism involving initial dissociation of one of the Fe–
19
Cl bonds. The reactivity vs. dioxygen could be addressed by
monitoring the formation of these m-oxo dimers at a single
(
II)
18
the Fe /H
purposes with our O
catalytically active systems involving the Fe /H
2
O
2
couple in order to obtain data for comparison
wavelength. From kinetic analysis, two distinguishable steps with
2
–ZnHg–AcOH system. The reactivity of the
rate constants termed k and k , respectively, could be deduced
1
2
(II)
2
O
2
couple is very
and these were used to appreciate the oxygen-sensitivity of our
complexes. The oxygenation of Cl TPAFeCl was monitored and
10–15,45,46,48
well documented.
As stated above, in the chemistry of
1
2
TPA complexes, cyclohexane can be converted into a mixture of
cylcohexanol/one with A/K ratios of up to 11.0, using complexes
with weakly coordinating counter anions. We chose to use a larger
excess of oxidant than is usually reported to improve the number
complete conversion was observed within 36 h. The values of the
two-step kinetic constants k and k lie within a comparable order
1
2
of magnitude to those found for the F TPAFeCl and TPAFeCl
1
2
2
complexes, indicating smooth oxygen sensitivity in solution. In the
case of Cl TPAFeCl and Cl TPAFeCl , the reaction was found
47
of cycles at the metal site. We checked that the reaction profile
remained unchanged when the reaction was carried out under
dioxygen, confirming the absence of short-lived radical species
in the reaction pathway. We also noted that previous studies
reported that g-nitro substitution of the ligands attached to iron
2
2
3
2
to be complete over 6 and 3 h respectively, following a multi step
process. The faster conversion of these complexes is attributed to
their TBP (i.e. coordinatively unsaturated) geometry. We thus have
a series of complexes that display oxygen sensitivity in the order
TPAFeCl ª Cl TPAFeCl ꢀ Cl TPAFeCl ª Cl TPAFeCl .
32
complexes decreased the activity of the catalyst. We followed
2
1
2
2
2
3
2
the H procedure with the Cl TPAFeCl and Cl TPA(OTf)
2
O
2
n
2
n
2
In dry CH CN solution, the Fe(OTf) complexes turned out
3
2
series of complexes and found the same trend, i.e. a decrease
of the activity upon substitution, with a slight increase of the
A/K ratio when using a monosubstituted ligand. Finally, we
found that the efficiency was slightly better, yet not substantially
to be spectroscopically insensitive to dioxygen. Absolutely no
1
spectroscopic changes (either in UV-vis or H NMR) could be
49
detected upon long-time exposure to O2. For this reason, the
reactivity of the triflate complexes in the alkane oxidation by the
different, with triflate-coordinated complexes than with the FeCl
2
2
O –ZnHg system will not be described in the present report.
complexes. A dramatic selectivity effect when switching from
chloride ligated complexes to bis-acetonitrile dications has already
We wish to come back to the fact that the formation of
m-oxo diferric compounds from O and Fe(II) complexes is a
2
48
been reported. In the present case, this effect is weak. At this
point, it is useful to remember that the decrease in selectivity
as a direct consequence of increasing the amount of oxidant is
general reaction and that free access of O to the metal centre,
2
9,21
1
allowing its coordination to iron, is crucial.
In former reports,
we adapted the scheme published in the late 70’s by porphyrin
chemists, the reaction pathway known as “autoxidation of iron
45
well known and has already been reported. Thus, the moderate
difference in activity of both complexes certainly reflects the use
of large amounts of oxidant.
18,22,50
porphyrins”.
The four key steps in this mechanism, displayed
in Scheme 6, are: (i) the formation of the iron-dioxygen adduct,
2
2,50
Scheme 6 Adaptation of the Autoxidation Mechanism to the chemistry of O
8 | Dalton Trans., 2011, 40, 92–106
2
-sensitive TPA complexes.
9
This journal is © The Royal Society of Chemistry 2011

View Article Online
termed “oxy” intermediate A, (ii) the formation of the m-peroxo
diferric species B by reaction of the oxy intermediate with 1 equiv.
of the ferrous precursor, (iii) the cleavage of the O–O bond in the
m-peroxo diferric to afford 2 equiv. of Fe(IV) O, termed oxo ferryl
species C and (iv) the reaction of one oxo ferryl with one ferrous
precursor to afford the final m-oxo diferric compound.
is higher than that generally observed in free-radical oxidation
33–35
35
processes,
in line with a weakly selective metal oxo-process.
The fact that cyclohexanol is not oxidized under the same
experimental conditions indicates that the ketone is not formed
by oxidation of the alcohol.
In order to compare our reaction conditions with those of
1,44,52
the “Gif system”,
we replaced acetonitrile with pyridine. A
Complex TPAFeCl
protons. Among all the complexes studied, only the parent
TPAFeCl complex and Cl TPAFeCl exhibited a noticeable
reactivity (a likely explanation for this will be given below). We
thus focused our studies on the behaviour of the parent TPAFeCl
The possible involvement of a potentially reactive species
prompted us to investigate the reaction in the strict respect of
the conditions depicted in Scheme 6. Since no conversion was
observed, it appeared that this scheme does not account for the
formation of the reaction product.
2
with cyclohexane, Zn/Hg and the role of
drastic modification of the reaction course was observed, with
an increase of the activity and a selectivity now being driven
towards the formation of the ketone form (A/K = 0.15). The
same effect was observed in the oxidation of adamantane, with
a 3 /2 ratio more than three-fold less than that in acetonitrile.
This suggests the occurrence of a radical-based mechanism in
pyridine.
At this point, the question of what kind of species could be
proton-sensitive can be raised.
(i) The mesomeric form of a dioxygen adduct A can be written
as a ferric superoxide species according to Scheme 7:
The presence of a source of protons in the medium would very
2
1
2
◦
◦
2
.
The addition of acetic acid turned out to be crucial in order to
observe the conversion of cyclohexane into alcohol and ketone,
yet with a weak efficiency in CH
a similar mechanism to that postulated in the reaction of triflato
Fe(II) TPA complexes with hydrogen peroxide. The selectivity
in the oxidation of adamantane, although not as high as that
observed from triflato Fe(II) complexes and hydrogen peroxide,
3
CN. The 3.1 A/K ratio suggests
likely result in the protonation of the superoxide ion to afford the
∑
HO
2
radical, leading to radical-based chemistry.
46
ii) The transient m-peroxo species B for which mesomeric forms
can be drawn as shown in Scheme 8 is also likely to be protonated
to yield the hydroperoxide D.
46
∑
Scheme 7 The “oxy-Fe(II)” forms can also be described as “superoxo-Fe(III) species”, yielding the HO
2
radical in protic media.
Scheme 8 The m-peroxo diferric species can also be described as “peroxoferric-Fe(III) species”, yielding two species upon protonation: a hydroperoxoferric
compound D and a ferric derivative.
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 92–106 | 99

View Article Online
Ferric hydroperoxide species D are known in the literature.
They are generally obtained from H and Fe(II) or Fe(III)
complexes and can sometimes be spectroscopically detected at
with the experimental observation reported more than thirty years
ago.
54
2
O
2
11,31,45,46,51
low temperature.
Considering molecular oxygen, two
Mechanistic considerations. In the present study, the main
result concerns O-atom transfer from O to an inactive substrate
recent reports describe the spectroscopic characterization of such
2
23,24
species from O
2
, protons and external reducing agents.
Thus,
and a dramatic solvent effect on the product profile. Scheme 9
depicts the possible mechanistic pathways that are consistent with
our experimental observations.
protonation of B to yield D (and some ferric compound left over)
seems to be a credible hypothesis.
Interestingly, the mesomeric form of the ferric hydroperoxide
species D can be written as a ferric compound plus the first
deprotonated form of hydrogen peroxide according to:
1. In the absence of protons, nothing other than the formation
of the diferric m-oxo species happens (pathway I: deactivation by
formation of m-oxo), in spite of the involvement of a potentially
reactive oxo-ferryl species C.
n
(III)
n
(III) +
-)]
L Fe (O
2
H) ↔ [(L Fe ) , (HO
2
2
. We show that even if the efficiency is much lower, the reactivity
(II)
of our AcOH–O –ZnHg–Fe –acetonitrile system qualitatively
2
The chemistry of hydrogen peroxide has been the subject of
parallels that observed with hydrogen peroxide, i.e. cyclohexane
is converted into cyclohexanol and some cyclohexanone. This
strongly suggests the involvement of a similar active species D
in both conditions (II: peroxide biomimetic pathway).
53
many investigations in the past. At this stage of the discussion, it
is now essential to remember that in pyridine, not in acetonitrile,
54
the hydroperoxide anion was, more than thirty years ago, shown
∑
-
∑
to decompose into O2 and HO according to:
3
. A gradual increase in the amount of added pyridine results in
pulling the reaction towards the formation of radicals. Acetic acid
is gradually converted into pyridinium acetate as a mild source
−
pyridine
•−
•
HO2 + H O ⎯⎯⎯→ O2 +OH + H O
2
2
2
55,56
54
of protons,
providing stable radicals. As a consequence:
In this work, hydrogen peroxide was deprotonated by a strong
base.
Coming back to the present case, our reaction medium contains
some peroxide and hydroperoxide anions, and possibly some
hydrogen peroxide, all resulting from the reduction of the coor-
(1) the activity increases (by up to 7 fold when pyridine is the
solvent) and (2) the selectivity is driven towards the formation of
ketone as the main product. This reaction pathway (pathway III,
radical pathway) is obviously different from that mentioned for
acetonitrile.
dinated O
2
in a protic medium. In pyridine, it is therefore not
Finally, the radical-based character of the reaction in pyridine
could definitely be established by our spin trapping experiments.
The striking point came from the spectroscopic comparison
between the two reaction conditions: significant amounts of the
surprising to observe a radical pathway, as previously reported in
early works, leading to the formation of ketone as the main
reaction product. This means our results are perfectly in line
54
Scheme 9 Autoxidation of ferrous compounds in the absence of protons (in yellow) and the pathways leading to radical chemistry and the conversion
of cyclohexane into cyclohexanol/one in the presence of protons.
1
00 | Dalton Trans., 2011, 40, 92–106
This journal is © The Royal Society of Chemistry 2011

View Article Online
∑
-
∑
∑
superoxide anion O
2
(as its protonated HO
2
form) and OH
cyclohexane in the presence of protons and a reducing agent,
in a similar fashion to that reported when hydroperoxides and
iron complexes are reacted together. Although very limited in
efficiency, this transformation represents one of the very few
functional examples of C–H hydroxylation of an inert substrate
by a well-defined mononuclear non-porphyrinic iron complex in
mild conditions.
radical could be trapped by DMPO in pyridine only. On the
other hand, the use of acetonitrile as a solvent did not enable
the detection of any paramagnetic species, neither oxygenated
nor ferric hydroperoxide. The lack of the signature of this latter
product was attributed to the instability of such species in our
45
reaction conditions.
In light of these considerations, we believe that we observe here a
versatile acid–base chemistry involving two forms of oxygen-active
species. The doubly reduced form of dioxygen within the m-peroxo
transient is protonated. In acetonitrile, the main process follows
a metal–oxo based pathway. In pyridine, radicals are produced,
leading to a different reaction course. This interpretation is
illustrated in Scheme 10.
Keeping in mind the Gif chemistry, we then decided to use the
TPAFeCl complex as a tool to compare our approach with that
2
previously reported in the former abundant literature and worked
58
IV
in similar conditions as those reported in the Gif chemistry.
Following these conditions, we found a completely different
reactivity, with enhanced efficiency but opposite selectivity, the
ketone being obtained as the main product within a radical-based
reaction mechanism.
57
The radical process or biomimetic pathway dichotomy has
been the subject of many controversies in the past. The first out-
come of the present work is to demonstrate that the coordination
of O is a pre-requisite to promote such conversions. The second
2
one is the illustration of the chemical versatility of active sites,
in the present case an oxygen-coordinated metal centre, in an a
priori simple process. Future development of this work will involve
optimization of the balance between the efficiency in reactivity and
ease of use, in other words the “operational stability”, of simple
ferrous complexes.
5
. Experimental
General considerations
Scheme 10 Schematic interpretation of the experimental results observed
Chemicals were purchased from Aldrich Chemicals and used
as received. The DPA bis(2-pyridylmethyl)amine and TPA lig-
ands were prepared according to published procedures: G. An-
dregg; F. Wenk, Helv. Chim. Acta, 1967, 50, 2330; M. S. Nel-
son, J. Rodgers, J. Chem. Soc. A, 1968, 272; M. M. Da
Mota, J. Rodgers, S. M. Nelson, J. Chem. Soc. A, 1969, 2036–
2044. All the solvents used during the metallation reactions,
work-up and reactivity were distilled and dried according to
W. L. F. Armarego.; D. D. Perrin, Purification of Laboratory
Chemicals, 4th ed.; Pergamon Press, Oxford, 1997.
in the present study.
Comparison within the series of Cl
ing back to Table 2, it is obvious that the efficiency of the O
ZnHg–AcOH system in acetonitrile remains weak. This might
be due to the fact that formation of the m-oxo species is the
thermodynamically favoured reaction (the main favoured process
in Scheme 10). This also explains why the most oxygen-sensitive
n
TPAFeCl complexes. Com-
2
2
–
five-coordinated complexes, the Cl
2
TPAFeCl
2
and Cl
3
TPAFeCl
2
complexes, exhibit a poor reactivity in the presence of a substrate,
they simply convert too fast into the m-oxo species. Thus, opti-
mization of the system would require decreasing the ability of the
ferrous compounds to form m-oxo compounds, while maintaining
a good activity vs. dioxygen. Efforts are ongoing in this direction
in our laboratory.
Analytical anhydrous FeCl was obtained as a white powder by
2
reacting iron powder (ACS grade) with hydrochloric acid in the
presence of methanol under an argon atmosphere. Preparation
and handling of all the compounds was performed under an
argon atmosphere using Schlenk technique, following standard
procedures. The purity of the dry dioxygen was 99.999% (grade
5
). Zinc amalgam was prepared as follows: a mixture of 200 g of
3
mercury, 2 g of zinc powder and 100 cm of 25% aqueous sulfuric
acid were stirred until no zinc powder was visible. The acidic
phase was removed and the amalgam was thoroughly washed
4
. Conclusion
Our first idea when we started this work was to exploit the oxygen
sensitivity of the FeCl complexes within a series of a-chloro-
substituted TPA-type ligands, aiming to study the feasibility
3
2
with distilled water until a neutral pH was reached. 5 cm of
fresh mercury was added. The resulting amalgam was filtered
on a paper skimmer and washed twice with dry THF. It was
then left under primary vacuum overnight. Elemental analyses
were carried out by the Service de Microanalyse de l’Institut de
Chimie de Strasbourg, Universit e´ de Strasbourg, France. Mass
spectroscopyexperiments were carriedout bythe Service Commun
de Spectrometrie de Masse de l’Institut de Chimie de Strasbourg.
The structural determinations were carried out by the Service
of oxygen-transfer reactions on inactive substrates. The full
characterization of all complexes with the generic Cl
n
TPAFeCl
2
formula has now become available. As expected, we found that
the effect of a-chlorination was to increase the oxygen sensitivity
with a maximum effect for complexes with trigonal bipyramidal
geometry in solution. We then observed that the complexes with
a moderate sensitivity allow oxygen atom transfer from O to
2
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 92–106 | 101

View Article Online
1
de Diffraction des Rayons X, Institut de Chimie, Universit e´ de
Strasbourg.
H NMR, CDCl
3
, d, ppm: 8.54 (m, a-CHarom, 2H); 7.60 (m,
CHarom, 6H); -7.15 (m, CHarom, 3 H); 3.89, (s, CH
CH , 2H).
2
, 4H); 3.87 (s,
2
1
3
C NMR, CDCl
3
, d, ppm: 160.6 (ipso Carom substPy); 159.0 (ipso
Physical methods
C
arom unsubstPy); 150.6 (ipso ClCaromPy); 149.1, 136.4, 123.0 and
1
122.0 (CHarom unsubstPy); 139.0, 122.4, 121.2 (CHarom substPy);
H NMR data were recorded in CD
for the ligands at ambient temperature on a Bruker
AC 300 spectrometer at 1300 MHz using the residual signal
of CD HCN (CHCl ) as a reference for calibration. The UV-
3
CN for the complexes and
◦
6
0.1 (CH
Elemental analysis for C18
7.2. Found (%): C 66.0, H 5.4, N 17.0.
2
unsubstPy); 59.5 (CH
2
substPy).
CDCl
3
H
17
N
4
Cl: calc. (%): C 66.5, H 5.2, N
1
2
3
vis spectra were recorded on a Varian Cary 05 E UV-VIS
NIR spectrophotometer equipped with an Oxford instrument
DN1704 cryostat, using optically transparent schlenck cells. The
kinetic data were analysed with the commercial software IGOR
pro version 4.0.8.0, Wavemetrics inc., USA, 2003. Conductivity
Cl TPA = bis (2-chloro-6-pyridylmethyl)(2-pyridylmethyl)amine.
2
To a solution of 1.5 g (7.3 mmol) of 2-chloro-6-bromomethyl
3
pyridine in 200 cm of EtOH were added 0.39 g (3.6 mmol) of
picolylamine and 1.5 g (14.15 mmol) of Na
2
CO . The reaction
3
mixture was refluxed over 16 h. Afterward, the solvent was
evaporated to dryness. After water was added, the mixture was
◦
measurements were carried out under argon at 20 C with a
CDM 210 Radiometer Copenhagen Conductivity Meter, using a
Tacussel CDC745-9 electrode. Cyclic voltammetry measurements
were obtained from a PAR 173A potentiostat in a 0.1 M
extracted several times with CH
phases washed with water then dried over MgSO
2
Cl
2
, and the combined organic
. Addition of
4
pentane to the concentrated organic phases yielded 0.59 g (40%)
acetonitrile solution of TBAPF (supporting electrolyte), using
6
of a white solid.
platinum electrodes and saturated calomel electrode as reference.
For each measurement, the potential was checked by addition of
1
H NMR, CDCl
3
, d, ppm: 8.54 (m, a-CHarom, 1H); 7.68–7.50
, 2H); 3.87 (s,
(
m, CHarom, 6H); 7.17 (m, CHarom, 3 H); 3.89, (s, CH
2
+
a small amount of ferrocene (Fc/Fc = 0.380 v/SCE) in the cell.
CH , 4H).
2
Gas-phase chromatography was performed on a Varian GC3900
apparatus, equipped with a capillary column (CP-Wax 52 CB, 15 m
13
C NMR, CDCl
ipso Carom unsubstPy); 150.6 (ipso ClCarom Py); 149.2, 136.4, 123.1
and 122.1 (CHarom unsubstPy); 139.0, 122.5, 121 (CHarom substPy);
3
, d, ppm: 160.3 (ipso Carom substPy); 158.8
(
¥
0.25 mm ¥ 0.39 mm), using a Flame Ionization Detector and
helium as the carrier gas. A continuous-wave EMXplus X-band
spectrometer (Bruker Biospin GmbH, Germany) equipped with
a high sensitivity resonator (4119HS-W1, Bruker) was used for
acquiring the EPR spectra.
6
0.1 (CH
Elemental analysis for C18
5.6. Found (%): C 59.6, H 4.9, N 15.3.
2
unsubstPy); 59.5 (CH
2
substPy).
H
16
N
4
2
Cl : calc. (%): C 60.1, H 4.4, N
1
Cl
mmol) of 2-chloro-6-bromomethyl pyridine were suspended with
NH Cl (145 mg, 2.43 mmol) in a mixture of THF and H
0 : 10 (150 mL). The pH was adjusted to 10 by addition of
aqueous NaOH. The resulting medium was stirred for 4 d at room
temperature in a tightly closed flask. After THF evaporation, the
3
TPA = tris (2-chloro-6-pyridylmethyl)amine. 1.5 g (7.3
Synthesis of ligands
4
2
O
9
2
-chloro-6-bromomethyl pyridine. To a solution of 4 g, (30
mmol) of 2-chloro-6-methyl pyridine in 200 mL of CCl were
4
added 6 g (30 mmol) of N-bromosuccinimide and 300 mg (1.25
mmol) of dibenzoyl peroxide. The medium was refluxed for 5 h,
then cooled to room temperature. The solvent was evaporated,
and the residue extracted with toluene. After filtration, the
reaction mixture was then poured into CH
2
2
Cl and the organic
phase separated, washed with water and then dried by MgSO
4
.
Addition of cold hexane to the concentrated organic phases
yielded 0.22 g (23%) of a white solid.
toluene solution was washed with water and dried on MgSO .
4
1
H NMR, CDCl
3
, d, ppm: 7.62 (t, g-CHarom, 3H); 7.50 (d, b-
, 6H).
, d, ppm: 160.0 (ipso Carom substPy); 147.2
ipso ClCarom Py); 140.2, 125.0, 122.1 (CHarom substPy); 50.6 (CH
substPy).
Mass spectroscopy: ES , m/z: 393.0347 (Cl
The concentrated toluene phase was deposited at the top of a
chromatography column mounted with silica gel and toluene.
Elution using toluene afforded the desired compound as the third
fraction. 2.22 g (10 mmol) of a pale white solid were obtained,
CHarom, 3H); 7.19 (d, b-CHarom, 3 H); 3.87 (s, CH
2
1
3
C NMR, CDCl
3
(
2
corresponding to a 34% yield of 2-chloro 6-bromomethyl pyridine.
+
+
TPAH ), 395.0318
TPAH ), corresponding to the two isotopes of Cl of
1
3
H NMR, CDCl
3
, d, ppm: 7.66 (t, g-CH, 1H), 7.38, (d, b-CH,
, 2H)
+
(Cl
3
1
H), 7.25 (d, b-CH, 1H); 4.48, (s, CH
2
+
C
18
H
15Cl
3
4
N H .
Elemental analysis for C18
4.2. Found (%): C 54.3, H 4.1, N 14.6.
H
15
N
4
Cl : calc. (%): C 54.9, H 3.8, N
3
Cl TPA = (2-chloro-6-pyridylmethyl)bis(2-pyridylmethyl)amine.
1
1
To a solution of 0.5 g (2.42 mmol) of 2-chloro-6-bromomethyl
3
pyridine in 200 cm of EtOH, were added 0.48 g (2.41 mmol) of
Preparation of the complexes
Preparation of the FeCl complexes. Details are given for
bis(2-pyridylmethyl)amine and 0.85 g (0.01 mmol) of NaHCO .
3
The reaction mixture was refluxed over 16 h. Afterwards, the
solvent was evaporated to dryness. After water was added,
2
Cl
1
TPAFeCl , but the following procedure applies to all com-
plexes: 150 mg (0.46 mmol) of free Cl TPA were dissolved in a
schlenk tube containing 20 mL of dry and degassed THF. 55 mg
(0.43 mmol) of anhydrous FeCl was dissolved in a second schlenck
tube containing 10 mL of dry and degassed THF. The solution of
2
the mixture was extracted several times with CH
combined organic phases washed with water and then dried over
MgSO . Addition of pentane to the concentrated organic phases
2
Cl
2
, and the
1
4
2
allowed precipitation of a white solid, which was filtered and dried
under vacuum. The yield was 0.34 g (43.5%).
FeCl
2
was transferred under argon in the schlenck containing the
1
02 | Dalton Trans., 2011, 40, 92–106
This journal is © The Royal Society of Chemistry 2011

View Article Online
1
ligand, and the medium was stirred overnight. The solvent was
then evaporated to dryness, and the compound was extracted with
Cl
CHD
3
FeCl
2
.
H NMR spectrum in CD
3
CN, R.T. d, ppm
2
CN: extremely broad signals over 70–30 ppm, in line with
dry and degassed CH
concentrated. Addition of diethyl ether afforded a yellow solid,
which was washed thoroughly with this solvent, prior to be dried
under vacuum. 180 mg (86%) of Cl
and spectroscopic data could be obtained.
3
CN, filtered under inert atmosphere and
a high-spin Fe(II) complex in trigonal bipyramidal geometry (ref.
14 and 18)
1
Cl
CHD
and 0 ppm.
3
Fe(OTf)
2
.
H NMR spectrum in CD CN, R.T. d, ppm
3
1
TPAFeCl with good analytical
2
2
CN: 40, 60 and broad and poorly defined signals between 5
1
9
Preparation of the Fe(OTf)
2
complexes. Details are given
F NMR, d: -74.2 ppm (Dn1/2 = 230 Hz).
for TPAFe(OTf) , but the following procedure applies to all
2
complexes: 150 mg (0.51 mmol) of free TPA were dissolved in
a schlenk tube containing 20 mL of dry and degassed THF.
Single-crystal X-ray data
Crystallization of Cl TPAFeCl
1
2
.
A solution of Cl
1
TPAFeCl
2
2
04 mg (0.46 mmol) of anhydrous Fe(OTf)
a second schlenck tube containing 10 mL of dry and degassed
THF. The solution of Fe(OTf) was transferred under argon in
2
was dissolved in
in dry CH Cl was layered with dry and degassed diethyl ether.
2
2
Orange single crystals were easily obtained over 2–3 d of diffusion
time under strict anaerobic conditions.
2
the schlenck containing the ligand, and the medium was stirred
overnight. The solvent was then evaporated to dryness, and the
Crystal data for Cl
.35 mm . C18 H17 Cl3 Fe N4, CH2Cl2, M = 536.48 g mol .
1
TPAFeCl
2
. Orange crystals, 0.40 ¥ 0.38 ¥
3
-1
0
compound was extracted with dry and degassed CH CN, filtered
3
Monoclinic, space group P2
1
/c, a = 10.2750(4), b = 13.5295(5), c =
under inert atmosphere and concentrated. Addition of diethyl
ether afforded a brownish-red solid, which was washed thoroughly
with this solvent, prior to be dried under vacuum. When dry,
◦
3
-3
˚
˚
16.2447(4) A, b = 94.919(2) , V = 2249.95(13) A , r
c
= 1.584 g cm ,
Z = 4, 0.998 < q < 30.034. Of 6556 total reflections, 4551 were
considered to be observed [I > 2s(I)], with 262 parameters. Final
the compound appears as a dark-red solid. 280 mg (84%) of
results: R = 0.0486 and R
w
= 0.1437; GOF = 1.081.
1
TPAFe(OTf)
2
with a H NMR trace identical to that reported
in: A. Diebold; K. S. Hagen, Inorg. Chem., 1998, 37, 215–223.
Elemental analysis for Cl TPAFe(OTf)
·H
ClF Fe: calc. (%): C 34.4, H 2.7, N 8.0. Found (%):
C 34.3, H 3.0, N 8.1.
Elemental analysis
Cl 10Fe: calc. (%): C 30.6, H 3.1, N 7.1.
Found (%): C 30.7, H 3.1, N 7.0
Elemental analysis for Cl TPAFe(OTf)
hygroscopic and different samples of the same batch gave non re-
producible values. Best result obtained for Cl TPAFe(OTf) O,
·8H
Cl 14Fe: calc. (%): C 27.1, H 3.5, N 6.3. Found (%):
C 26.8, H 2.9, N 5.9
In complexes with Cl1–3TPA ligands, paramagnetic H and
Crystalization of Cl TPAFeCl . A solution of Cl TPAFeCl
2
2
2
2
1
2
2
O,
in dry CH CN was layered with dry and degassed diethyl ether:
3
C
20
H
19
N
4
6
S
2
O
7
yellow single crystals were easily obtained over 2–3 d of diffusion
time under strict anaerobic conditions.
for
Cl
2
TPAFe(OTf)
2
·4H
2
O,
Crystal data for Cl TPAFeCl : Yellow crystals, 0.40 ¥ 0.35 ¥
2
2
3
-1
C
20
H
24
N
4
2
F
6
S
2
O
0.30 mm . C18 H16 Cl4 Fe N4, M = 486.00 gmol . Monoclinic,
space group Cc, a = 9.1920(5), b = 15.5454(10), c = 15.1189(8) A˚ ,
b = 106.520(3) , V = 2071.2(2) A˚ , r = 1.559 g cm , Z = 4, 0.998
◦
3
-3
3
2
: this compound is very
c
< q < 30.034. Of 4056 total reflections, 3368 were considered to
be observed [I > 2s(I)], with 244 parameters. Final results: R =
2
2
2
C
20
H
31
N
4
3
F
6
S
2
O
0.0387 and R = 0.0956; GOF = 1.017.
w
1
19
Crystalization of Cl
3
TPAFeCl
2
.
A solution of Cl
3
TPAFeCl
2
in
F
dry CH NO was layered with dry and degassed diethyl ether:
3
2
NMR data were obtained (all traces are given in the ESI†). In the
1
yellow single crystals were easily obtained over 2–3 d of diffusion
time under strict anaerobic conditions.
H NMR spectra, all the complexes display broadened and shifted
signals in line with a high-spin state for the iron centre. For the
1
9
Crystal data for Cl
3
TPAFeCl
2
: Yellow crystals, 0.24 ¥ 0.11 ¥
triflate complexes, all the F NMR data support the presence of
uncoordinated triflate ions in a CD CN solution, although minor
3
-1
0
.10 mm . C18 H15 Cl5 Fe N4, M = 520.44 gmol . Triclinic,
3
¯
◦
˚
space group P1, a = 8.6446(5), b = 11.4985(9), c = 13.3214(9) A,
broadening may arise from a limited dynamic exchange process in
solution.
◦
◦
a = 86.774(3) , b = 75.367(4) , g = 78.051(4) , V = 1253.42(15)
3
-3
˚
A , r
c
= 1.379 g cm , Z = 2, 0.998 < q < 27.485. Of 5716 total
1
Cl
1
FeCl
2
.
H NMR spectrum in CD
3
CN, R.T. d, ppm
reflections, 4062 were considered to be observed [I > 2s(I)], with
CHD
2
CN: 125: a; 63, 55, 21: CH
2
; 52, 44: b,b¢ unsubst. Py; 45,
253 parameters. Final results: R = 0.0650 and R = 0.1967; GOF =
w
31: b,b¢subst. Py; 17: g, unsubst. Py; 14: g, subst. Py.
1.038.
Crystallographic data have been deposited with the CCDC,
1
Cl
CHD
6: b,b¢subst. Py; -4: g, unsubst. Py; -2: g, subst. Py.
1
Fe(OTf)
2
.
H NMR spectrum in CD
3
CN, R.T. d, ppm
7
(
51031 (Cl
Cl TPAFeCl
3
1
TPAFeCl
).
2
), 751030 (Cl
2
TPAFeCl ) and 751029
2
2
CN: 105: a; 88, 83, 37: CH
2
; 52, 54: b,b¢ unsubst. Py; 56,
2
4
1
9
F NMR, d: -72.1 ppm (Dn1/2 = 35 Hz).
Reactivity studies
Reaction
1
Cl
2
FeCl
2
.
H NMR spectrum in CD
; 52, 44: b,b¢ unsubst. Py; 45, 31: b,b¢subst. Py;
7: g, unsubst. Py; 14: g, subst. Py.
3
CN, R.T. d, ppm
1.
conditions
/Zn/Hg/CH
for
CO
the
H in CH
conversions
CN. All reactions
with
CHD
2
CN: CH
2
Cl
n
TPAFeCl /O
2
2
3
2
3
1
were carried out at least five times, yielding reproducible results.
Average values of TONs are reported with estimated uncertainties
± 5–8%.
Typical reaction. In a finger-size schlenk tube containing one or
two drops of zinc amalgam under argon was transferred a solution
1
Cl
CHD
b,b¢unsubst. Py; 0: g, subst. Py; -4: g, unsubst. Py.
2
Fe(OTf)
2
.
H NMR spectrum in CD
3
CN, R.T. d, ppm
2
CN: 99, 80, 58: CH
2
; 64, 44: b,b¢ subst. Py; 57, 55:
1
9
F NMR, d: -74.1 ppm (Dn1/2 = 90 Hz).
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 92–106 | 103

View Article Online
containing 1.2–1.7 mg of complex (1 equiv.) in 2 mL of dry and
degassed acetonitrile. 0.35 mL (1100 equiv.) of cyclohexane and
containing 1.2 mg of TPAFeCl
2
(1 equiv.) in 2 mL of degassed
pyridine. 0.35 mL (1100 equiv.) of cyclohexane and 0.05 mL (300
equiv.) of acetic acid were added, and the medium was vigorously
stirred. Dioxygen was then bubbled in the medium over 5 s, and
the reaction was kept under an O
by connection of the schlenk with an O
0
.05 mL (300 equiv.) of acetic acid were added and the medium was
vigorously stirred. Dioxygen was then bubbled in the medium over
s and the reaction was kept under O atmosphere for five hours,
by connection of the schlenk with a O -filled rubber balloon. From
5
2
2
atmosphere for five hours,
-filled rubber balloon.
2
2
yellow at the beginning of the reaction, the colour gradually turned
to orange and brown. At the end of the reaction time, the medium
consisted of a suspension of a grey solid in a pale brown solution.
At this point, a measured volume of acetophenone was added as
an external standard for quantitative calibration, magnet stirring
was stopped and the solid deposited down to the bottom of the
schlenck. A sample of the liquid suspension was extracted and
injected into the GC apparatus.
From orange at the beginning of the reaction, the colour gradually
turned to yellow. At the end of the reaction time the medium
consisted of a suspension of a grey solid in a pale yellow solution.
At this point, the same analytical procedure as that described
in reactivity study 1 was carried out.
Attempts to detect cyclohexyl hydroperoxide: the procedure
59
described by Shul’pin et al. was carried out. We found that the
reaction profile was not modified upon the addition of PPh to
3
[
Mass balance: the estimated loss of cyclohexane was around 2%
the reaction medium prior to injection into the GC. In fact, this
procedure was originally setup for hydroperoxide chemistry. In
the present case we work in the presence of molecular oxygen and
a reducing agent, which might simply react with any cyclohexyl
hydroperoxide if any present.
upon bubbling O and workup (1080 equiv. of recovered material
from 1100 equiv. of reagent)]. Only cyclohexanol (retention time:
min at 96 C) and cyclohexanone (retention time: 4.5 min at
8 C) were detected (GC [total run time of 20 min] and mass
2
◦
6
7
◦
spectroscopy).
4
.
Reaction conditions for the adamantane conversion with
Attempts to detect cyclohexyl hydroperoxide: the procedure
described by Shul’pin et al. was carried out. We found that the
reaction profile was not modified upon the addition of PPh to
5
9
TPAFeCl /O /Zn/Hg/CH CO H/in CH CN or pyridine. This
reaction was carried out three times yielding reproducible results.
Average values of TONs are reported with estimated uncertainties
2
2
3
2
3
3
the reaction medium prior to injection into the GC. In fact, this
procedure was originally setup for hydroperoxide chemistry. In
the present case we work in the presence of molecular oxygen and
a reducing agent, which might simply react with any cyclohexyl
hydroperoxide if any present.
The reaction was carried out in the absence of zinc amalgam
and/or acetic acid, following the same procedure. Only trace
amounts of the oxidation products from cyclohexane could then
be detected.
±
5–8%.
Typical reaction. In a finger-size schlenck tube containing one
or two drops of zinc amalgam was transferred under argon a
solution containing 1.2 mg of TPAFeCl (1 equiv.) in 2 mL of
degassed CH CN or pyridine. 0.45 g (1100 equiv.) of adamantane
and 0.05 mL (300 equiv.) of acetic acid were added, and the
medium was vigorously stirred. Dioxygen was then bubbled in the
2
3
medium over 5 s, and the reaction was kept under O
for five h, by connection of the schlenk with a O -filled rubber
2
atmosphere
2
2
.
Reaction
conditions
CO
for
the
conversion
with
CN.
balloon. From yellow at the beginning of the reaction, the colour
gradually turned to pink-beige. At the end of the reaction time
the medium consisted of a suspension of a grey solid in a clear
pink-beige solution.
At this point, the same analytical procedure as that described
in reactivity study 1 was carried out.
TPAFeCl /O /Zn/Hg/CH
2
2
3
2
H/pyridine 1 : 1 in CH
3
This reaction was carried out three times yielding reproducible
results. Average values of TONs are reported with estimated
uncertainties ± 5–8%.
Typical reaction. In a finger-size schlenk tube containing one
or two drops of zinc amalgam under argon, was transferred a
solution containing 1.2 mg of TPAFeCl
and degassed acetonitrile. 0.35 mL (1100 equiv.) of cyclohexane,
.05 mL (300 equiv.) of acetic acid and 0.07 mL (300 equiv.) of
pyridine were added, and the medium was vigorously stirred.
Dioxygen was then bubbled in the medium over 5 s, and the
2
(1 equiv.) in 2 mL of dry
5. Attempt to convert cyclohexanol into cyclohexanone with
TPAFeCl
2
/O
2
/Zn/Hg/CH
3
CO
2
H in CH CN. This reaction
3
0
was carried out three times yielding reproducible results. Average
values of TONs are reported with estimated uncertainties ± 5–8%.
The same procedure as that reported with cyclohexane (reac-
tivity study 1) was used, except that cyclohexane was replaced
by cyclohexanol. Mass balance and GC analysis revealed that no
conversion had occurred.
reaction was kept over O
2
atmosphere for five hours, by connection
of the schlenk with a O -filled rubber balloon. From yellow at the
2
beginning of the reaction, the colour gradually turned to greenish
yellow. At the end of the reaction time, medium consisted of a
suspension of grey solid into a greenish yellow solution.
At this point, the same analytical procedure as that described
in reactivity study 1 was carried out.
6
.
Reaction conditions for the conversion with H
2
2
O . All re-
actions were carried out at least three times, yielding reproducible
results. Average values of TONs are reported with estimated
uncertainties ± 5–8%.
3
.
Reaction
conditions
CO
for
the
conversion
with
1. Cl TPAFeCl complexes: Typical reaction. 2.0 ml of a 2.5 M
n
2
TPAFeCl /O /Zn/Hg/CH
2
2
3
2
H in pyridine. This reaction
(5 mmol) H O solution (diluted from a 33% aqueous solution)
2
2
was carried out at least five times, yielding reproducible results.
Average values of TONs are reported with estimated uncertainties
in CH CN was delivered by a syringe pump over 4 h 30 min
at 25 C under argon, to a vigorously stirred solution (5 mL)
3
◦
±
5–8%.
containing Cl0–3TPAFeCl , (10 mmols) and cyclohexane (5 mmol).
2
Typical reaction. In a finger-size schlenk tube containing one or
At this point, a measured volume of acetophenone was added as
two drops of zinc amalgam was transferred under argon a solution
04 | Dalton Trans., 2011, 40, 92–106
an external standard for quantitative calibration, magnet stirring
1
This journal is © The Royal Society of Chemistry 2011

View Article Online
was stopped and 40 ml of the crude medium were diluted into 1 mL
of CH CN. A sample of the resulting solution was injected into
the GC apparatus.
5 S. V. Kryatov, E. V. Rybak-Akimova and S. Schindler, Chem. Rev.,
2
005, 105, 2175–2226.
3
6
M. Abu-Omar, A. Loaiza and N. Hontzeas, Chem. Rev., 2005, 105,
2
227–2252.
Another set of experiments was carried out under an atmo-
7 E. G. Kovaleva and J. D. Lipscomb, Nat. Chem. Biol., 2008, 4, 186–193.
8 T. Punniyamurthy, S. Velusamy and J. Iqbal, Chem. Rev., 2005, 105,
sphere of dioxygen: O
of H started. The results were not significantly different.
. Cl
.5 M (5 mmol) H
2
was bubbled during 5 s before the addition
2
329–2363.
2
O
2
9
J. A. Labinger, J. Mol. Catal. A: Chem., 2004, 220, 27–35.
2
n
TPAFe(OTf)
2
complexes: Typical reaction. 2.0 ml of a
solution (diluted from a 33% aqueous
1
1
0 M. Costas, K. Chen and L. Que, Jr., Coord. Chem. Rev., 2000, 200–202,
2
2
O
2
517–544.
1 M. Costas, M. P. Mehn, M. P. Jensen and L. Que, Jr., Chem. Rev., 2004,
104, 939–986.
solution) in CH
3
CN was delivered by a syringe pump over 4h
◦
3
0 min at 25 C under argon, to a vigorously stirred solution
1
1
2 M. S. Chen and M. C. White, Science, 2007, 318, 783–787.
3 L. Gomez, I. Garcia – Bosch, A. Company, J. Benet-Buchholtz, A.
Polo, X. Sala, X. Ribas and M. Costas, Angew. Chem., Int. Ed., 2009,
(
5 mL) containing Cl0–3TPAFe(OTf)
2
, (10 mmol) and cyclohexane
(
5 mmol). At this point, a measured volume of acetophenone was
4
8, 5720–5723.
added as an external standard for quantitative calibration, magnet
stirring was stopped and 40 mL of the crude medium were diluted
1
1
4 M. S. Chen and M. C. White, Science, 2010, 327, 566–571.
5 V. B. Romakh, B. Therrien, G. Suss-Fink and G. B. Shul’pin, Inorg.
Chem., 2007, 46, 3166–3175.
into 1 mL of CH
3
CN. A sample of the resulting solution was
injected into the GC apparatus.
16 N. Kitajima, M. Ito, H. Fukui and Y. Moro-Oka, J. Chem. Soc., Chem.
Commun., 1991, 102–104.
1
7 N. Raffard-Pons, Y. Moll, F. Banse, K. Miki, M. Nierlich and J.-J.
7
.
Spin trapping experiments. In a finger-size schlenck tube
containing one or two drops of zinc amalgam was transferred,
under argon, a solution containing 1.2 mg of TPAFeCl (1 equiv.)
in 2 mL of dry and degassed solvent (acetonitrile or pyridine).
.35 mL (1100 equiv.) of cyclohexane and 0.05 mL (300 equiv.) of
acetic acid were added, and the medium was vigorously stirred.
5 mL of a 2 M solution of 5,5-dimethyl-1-pyrroline-N-oxide
DMPO, purchased from Sigma Aldrich) in distilled H O was
Girerd, Eur. J. Inorg. Chem., 2002, 1941–1944.
18 N. K. Thallaj, O. Rotthaus, L. Benhamou, N. Humbert, M. Elhabiri,
M. Lachkar, R. Welter, A.-M. Albrecht-Gary and D. Mandon, Chem.–
Eur. J., 2008, 14, 6742–6753.
2
1
9 A. Wane, N. K. Thallaj and D. Mandon, Chem.–Eur. J., 2009, 15,
0593–10602.
0
1
20 A. Machkour, D. Mandon, M. Lachkar and R. Welter, Eur. J. Inorg.
Chem., 2005, 158–161.
1 L. Benhamou, A. Machkour, O. Rotthaus, M. Lachkar, R. Welter and
D. Mandon, Inorg. Chem., 2009, 48, 4777–4886.
2 A. Machkour, D. Mandon, M. Lachkar and R. Welter, Inorg. Chem.,
2004, 43, 1545–1550.
3 A. Thibon, J. England, M. Martinho, V. G. Young, J. R. Frisch, R.
Guillot, J.-J. Girerd, E. M u¨ nck, L. Que, Jr. and F. Banse, Angew. Chem.,
Int. Ed., 2008, 47, 7064–7067.
2
(
2
2
2
2
added, without further purification. Dioxygen was then bubbled
in the medium over 5 s, and the medium was stirred for five min.
A glass capillary was then filled with 50 mL of solution, centred in
a regular 4 mm OD EPR quartz tube and immediately introduced
into the EPR cavity. The principal experimental parameters were:
gain 40 dB, modulation amplitude 0.5 to 1 G, microwave power
24 S. Hong, Y.-M. Lee, W. Shin, S. Fukuzumi and W. Nam, J. Am. Chem.
Soc., 2009, 131, 13910–13911.
2
5 D. Mandon, A. Machkour, S. Goetz and R. Welter, Inorg. Chem., 2002,
1, 5364–5372.
6 A. Diebold and K. S. Hagen, Inorg. Chem., 1998, 37, 215–223.
1
.8 mW, time constant 20.48 ms, conversion time 30 to 60.1 ms.
4
The centre field was set to 3532 G, 200 G were swept in 60 s per
2
scan. All experiments were performed at room temperature (295 ±
27 A. Machkour, D. Mandon, M. Lachkar and R. Welter, Inorg. Chim.
Acta, 2005, 358, 839–843.
1
K).
2
8 We even found that upon exposure to diethyl ether, this solvent
coordinated our complexes and could only be removed by several
prolonged washings of the complex in N-pentane under ultrasonic
conditions.
Attempt to trap any ferric hydroperoxide species: the same
experimental procedure was applied, the medium being trans-
ferred into a 4 mm EPR quartz tube (instead of a capillary)
and immediately frozen. The data acquisition were performed at
2
3
3
3
9 Y. Zang, J. Kim, Y. Dong, E. C. Wilkinson, E. H. Appelman and L.
Que, Jr., J. Am. Chem. Soc., 1997, 119, 4197–4205.
0 G. J. P. Britovsek, J. England and A. J. White, Inorg. Chem., 2005, 44,
8125–8134.
1
00 K. No paramagnetic species could be observed under these
conditions.
1 A. Company, L. Gomez, X. Fontrodona, X. Ribas and M. Costas,
Chem.–Eur. J., 2008, 14, 5727–5731.
2 Y. Hitomi, S. Furukawa, M. Higushi, T. Shishido and T. Tanaka, J. Mol.
Catal. A: Chem., 2008, 288, 83–86.
Acknowledgements
3
3
3 A. F. Trotman-Dickenson, Adv. Free Radical Chem., 1965, 1, 1–38.
4 G. V. Buxton, C. L. Greenstock, W. P. Helman and A. B. Ross, J. Phys.
Chem. Ref. Data, 1988, 17, 513–886.
The CNRS and UdS are gratefully acknowledged. This work was
in part realized when Prof. R e´ my Louis was Head Manager of
the Institut de Chimie de Strasbourg. We wish to address special
thanks to him for his constant support and encouragement. The
Conseil Scientifique de l’ULP (now UdS) is acknowledged for
specific support no. AO CS ULP 2006.
35 S. Miyajima and O. Simamura, Bull. Chem. Soc. Jpn., 1975, 48, 526–
30.
6 E. Finkelstein, G. M. Rosen and E. J. Rauckman, J. Am. Chem. Soc.,
980, 102, 4994–4999.
37 Y. Xia, A. L. Tsai, V. Berka and J. L. Zweier, J. Biol. Chem., 1998, 273,
5804–25808.
5
3
1
2
3
8 L. C. Wilms, J. C. S. Kleinjans, E. J. C. Moonen and J. J. Briede, Toxicol.
in Vitro, 2008, 22, 301–307.
References
3
9 M. Spulber and S. Schlick, J. Phys. Chem. A, 2010, 114, 6217–
6225.
1
D. H. R. Barton, M. J. Gastiger and W. B. Motherwell, J. Chem. Soc.,
Chem. Commun., 1983, 41–43.
40 B. Vileno, PhD Thesis manuscript No. 3501,2006, Ecole Polytechnique
2
P. Stavropoulos, R. Celenligil-Cetin and A. E. Tapper, Acc. Chem. Res.,
F e´ d e´ rale de Lausanne, Switzerland.
2
001, 34, 745–752.
41 L. Benhamou, M. Lachkar, D. Mandon and R. Welter, Dalton Trans.,
2008, 6996–7003.
42 J. England, C. R. Davies, M. Banaru, A. J. P. White and G. J. P.
Britovsek, Adv. Synth. Catal., 2008, 350, 883–897.
3
4
S. Tanase and E. Bouwman, Adv. Inorg. Chem., 2006, 58, 29–75.
J. Du Bois, T. J. Mizogushi and S. Lipperd, Coord. Chem. Rev., 2000,
2
00–202, 443–485.
This journal is © The Royal Society of Chemistry 2011
Dalton Trans., 2011, 40, 92–106 | 105

View Article Online
4
4
3 M. P. Jensen, S. J. Lange, M. P. Mehn, E. L. Que and L. Que, Jr., J. Am.
Chem. Soc., 2003, 125, 2113–2128.
52 D. H. R. Barton, Tetrahedron, 1998, 54, 5805–5817.
53 See for instance: G. Strukul, Ed., Catalytic Oxidations with Hydrogen
Peroxide as Oxidant. Catal. Met. Complexes, 1992, 9, Dordrecht, The
Netherlands.
4 D. H. R. Barton, A. H. Beck and D. K. Taylor, Tetrahedron, 1995,
5
1, 5245–5254 Previous reactivity studies in acetonitrile with the same
complex, TPAFeCl
2
(ClO
4
), showed that tertbutyl hydroperoxide was
54 J. L. Roberts, Jr., M. M. Morrison and D. T. Sawyer, J. Am. Chem.
efficient in promoting the stoichiometric formation of haloalkanes and
some cyclohexanol/one when an excess of oxidant was used. See: T.
Kojima, R. A. Leising, S. Yan and L. Que, Jr., J. Am. Chem. Soc., 1993,
Soc., 1978, 100, 329–330.
55 S. Venkataraman, Journal of the Indian Chemical Society, 1940, 17, 297–
303 Ref. 46 and 47: a 1 : 1 mixture of pyridine and acetic acid provides a
poorly-defined compound termed pyridinium acetate, which in solution
equilibrates with its conjugated base/acid.
1
15, 11328–11335.
5 C. Kim, K. Chen, J. Kim and L. Que, Jr., J. Am. Chem. Soc., 1997, 119,
964–5965.
4
5
56 J. Hurwic, J. Michalczyk and K. Pogorzelska, Roczniki Chemii, 1957,
4
4
6 K. Chen and L. Que, Jr., J. Am. Chem. Soc., 2001, 123, 6327–6337.
7 The obvious consequence is to decrease the yield: in general, in
biomimetic catalysis involving peroxides, the yield is calculated on the
consumption of oxidant.
8 Y. Mekmouche, S. M e´ nage, C. Toia-Duboc, M. Fontecave, J.-B. Galey,
C. Lebrun and J. Pecaut, Angew. Chem., Int. Ed., 2001, 40, 949–952.
2
9 This is one of the a posteriori justifications for the choice of FeCl as
the source of metal for all complexes already reported by our group.
0 D. H. Chin, J. Del Gaudio, G. La Mar and A. Balch, J. Am. Chem.
Soc., 1977, 99, 5486–5488.
1 A. Wada, S. Ogo, S. Nagatomo, T. Kitagawa, Y. Watanabe, K. Jitsukawa
and H. Masuda, Inorg. Chem., 2002, 41, 616–618.
31, 265–275.
57 We prefer the term “radical process” to the “Fenton process”, which is
sometimes used, because historically, the Fenton reaction corresponds
to the reaction of hydrogen peroxide, not molecular dioxygen, with
metal salts.
4
4
5
5
58 The main reaction is the formation of the alcohol. The fact that
some ketone is formed and particularly not by oxidation of the
2
alcohol, suggests that some oxy form [Fe(O )] might be protonated
by acetic acid to generate radicals, converting cyclohexane into
cyclohexanone.
59 G. B. Shul’pin and G. S u¨ ss-Fink, J. Chem. Soc. Perkin Trans., 1995,
1459–1463.
1
06 | Dalton Trans., 2011, 40, 92–106
This journal is © The Royal Society of Chemistry 2011