Hydroxylation of aromatics with the help of a non-haem FeOOH: A mechanistic study under single-turnover and catalytic conditions

Article abstract of DOI:10.1002/chem.201102252

Ferric-hydroperoxo complexes have been identified as intermediates in the catalytic cycle of biological oxidants, but their role as key oxidants is still a matter of debate. Among the numerous synthetic low-spin Fe^III(OOH) complexes characterized to date, [(L₅²)Fe(OOH)] ²⁺ is the only one that has been isolated in the solid state at low temperature, which has provided a unique opportunity for inspecting its oxidizing properties under single-turnover conditions. In this report we show that [(L₅²)Fe(OOH)]²⁺ decays in the presence of aromatic substrates, such as anisole and benzene in acetonitrile, with first-order kinetics. In addition, the phenol products are formed from the aromatic substrates with similar first-order rate constants. Combining the kinetic data obtained at different temperatures and under different single-turnover experimental conditions with experiments performed under catalytic conditions by using the substrate [1,3,5-D₃]benzene, which showed normal kinetic isotope effects (KIE>1) and a notable hydride shift (NIH shift), has allowed us to clarify the role played by Fe^III(OOH) in aromatic oxidation. Several lines of experimental evidence in support of the previously postulated mechanism for the formation of two caged Fe ^IV(O) and OH^. species from the Fe^III(OOH) complex have been obtained for the first time. After homolytic O-O cleavage, a caged pair of oxidants [Fe^IVO+HO^.] is generated that act in unison to hydroxylate the aromatic ring: HO^. attacks the ring to give a hydroxycyclohexadienyl radical, which is further oxidized by Fe ^IVO to give a cationic intermediate that gives rise to a NIH shift upon ketonization before the final re-aromatization step. Spin-trapping experiments in the presence of 5,5-dimethyl-1-pyrroline N-oxide and GC-MS analyses of the intermediate products further support the proposed mechanism. Oxidation by Fe^III(OOH): Investigations on a genuine non-haem Fe ^III(OOH) intermediate (see figure) in the presence of either aromatic substrates or the probe substrate [1,3,5-D₃]benzene has clarified the role played by Fe^III(OOH) in aromatic oxidations. Evidence for the formation of two caged Fe^IV(O) and OH^. species from Fe^III(OOH) has been obtained for the first time. These oxidants act in unison to give phenol products with normal kinetic isotope effects and notable hydride shifts. Copyright

Full text of DOI:10.1002/chem.201102252

FULL PAPER
DOI: 10.1002/chem.201102252
Hydroxylation of Aromatics with the Help of a Non-Haem FeOOH: A
Mechanistic Study under Single-Turnover and Catalytic Conditions
Aurore Thibon,[a] Vꢀronique Jollet,[a] Caroline Ribal,[a] Katell Sꢀnꢀchal-David,[a]
Laurianne Billon,[a] Alexander B. Sorokin,*[b] and Frꢀdꢀric Banse*[a]
Abstract: Ferric–hydroperoxo com-
plexes have been identified as inter-
mediates in the catalytic cycle of bio-
logical oxidants, but their role as key
oxidants is still a matter of debate.
Among the numerous synthetic low-
strates with similar first-order rate con-
stants. Combining the kinetic data ob-
tained at different temperatures and
under different single-turnover experi-
mental conditions with experiments
performed under catalytic conditions
mechanism for the formation of two
IV
C
caged Fe (O) and OH species from
the FeIII
AHCTUNGTRENNUNG
tained for the first time. After homolyt-
ꢀ
ic O O cleavage, a caged pair of oxi-
IV
C
dants [Fe O+HO ] is generated that
spin FeIII
ized to date, [(L5 )Fe
(OOH) complexes character-
by
using
the
substrate
[1,3,5-
act in unison to hydroxylate the aro-
C
D3]benzene, which showed normal ki-
netic isotope effects (KIE>1) and a
notable hydride shift (NIH shift), has
allowed us to clarify the role played by
matic ring: HO attacks the ring to give
a hydroxycyclohexadienyl radical,
only one that has been isolated in the
solid state at low temperature, which
has provided a unique opportunity for
inspecting its oxidizing properties
under single-turnover conditions. In
which is further oxidized by FeIVO to
give a cationic intermediate that gives
rise to a NIH shift upon ketonization
before the final re-aromatization step.
Spin-trapping experiments in the pres-
ence of 5,5-dimethyl-1-pyrroline N-
oxide and GC-MS analyses of the inter-
mediate products further support the
proposed mechanism.
FeIII
ACTHNUTRGNEU(GN OOH) in aromatic oxidation. Sev-
eral lines of experimental evidence in
support of the previously postulated
2
this report we show that [(L5 )Fe-
ACHTUNGTRENNUNG
(OOH)]2+ decays in the presence of ar-
omatic substrates, such as anisole and
benzene in acetonitrile, with first-order
kinetics. In addition, the phenol prod-
ucts are formed from the aromatic sub-
Keywords: hydroxylation · iron ·
kinetics · N ligands · oxidation · re-
action mechanisms
Introduction
cin (ABLM) is 1) the active species or yielded 2) FeIV=O+
V
ꢀ
ꢀ
C
HO or 3) Fe =O+HO by homolytic or heterolytic O O
bond cleavage, respectively. Experimental and theoretical
studies have been dedicated to this issue, but have not nec-
essarily converged to the same conclusion. Decker et al.[3]
and Neese et al.[4] favored the first scenario, whereas Kumar
et al. proposed that the third case was also a realistic hy-
pothesis.[5] Concerning synthetic models of ABLM, it is gen-
In the catalytic cycles of oxygenases that use iron as a cofac-
tor, for example, haem enzymes, such as cytochro-
mes P450,[1] iron–hydroperoxo intermediates are considered
as important intermediates that break down to form high-
valent iron–oxo complexes. In the case of the anti-tumor
drug bleomycin, which oxidatively degrades DNA, the in-
volvement of such an intermediate is well documented, but
its fate is not as clear as in P450.[2] It has been frequently
discussed whether the low-spin FeIIIOOH-activated bleomy-
ꢀ
erally considered that they are activated towards O O
cleavage.[6] The similar problem of the possible reactivity
and fate of the FeIIIOOH species is also the subject of dis-
cussion in haem systems.[7] Note that the possible implica-
tion of metal–peroxo species in oxidation reactions is a fre-
quent issue not restricted to iron.[8]
[a] Dr. A. Thibon, V. Jollet, Dr. C. Ribal, Dr. K. Sꢀnꢀchal-David,
L. Billon, Prof. F. Banse
We have previously reported the generation with hydro-
gen peroxide and spectroscopic identification of mononu-
clear ferric–hydroperoxo complexes obtained with simple
Institut de Chimie Moleculaire et des Materiaux d’Orsay
Laboratoire de Chimie Inorganique, Universitꢀ Paris-Sud 11
91405 Orsay Cedex (France)
2
amine/pyridine ligands derived from our prototypical L5
Fax : (+33)169154754
(Scheme 1).[9] In addition, we have shown that the ferrous
complexes obtained with the same ligands exhibit an inter-
esting ability to catalyze aromatic hydroxylation by H2O2,
even with substrates, such as benzene and chlorobenzene.[10]
Furthermore, this reaction was shown to be controlled by
the metal center, because the oxygen atom was inserted into
the product derived from H2O2 regardless of the presence or
E-mail: frederic.banse@u-psud.fr
[b] Dr. A. B. Sorokin
Institut de Recherches sur la Catalyse et l’Environnement de Lyon
UMR 5256, CNRS-Universitꢀ Lyon 1
2 av. A. Einstein, 69626 Villeurbanne Cedex (France)
E-mail: alexander.sorokin@ircelyon.univ-lyon1.fr
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201102252.
Chem. Eur. J. 2012, 18, 2715 – 2724
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2715

Scheme 1. Structures of the ligands used in this study.
absence of O2 and was not affected by the presence of la-
beled water.[10] In this paper we address the question of the
competence of FeIIIOOH intermediates to perform this de-
manding oxidation reaction. Having in hand a well-charac-
terized ferric–hydroperoxo intermediate isolated as a micro-
crystalline powder,[11] we had a unique opportunity to gain
an insight into the mechanism of this reaction. Indeed, reac-
tivity studies involving FeIII
ly been performed under catalytic conditions, because FeIII-
(OOH) species are usually obtained under steady-state con-
ACHTUNGTNER(NUNG OOH) complexes have general-
ACHTUNGTRENNUNG
2
ACTHNUTRGNEUNG
Figure 1. Kinetic monitoring of the degradation of [(L5 )FeIII(OOH)]2+ at
ditions, that is, by using an excess of H2O2. Herein this issue
78C in acetonitrile (top) and absorbance variation at 519 nm as a func-
tion of time (bottom). Experimental data ( ) and fitting by a first-order
is inspected for the first time under single-turnover condi-
^
2
tions by using the solid [(L5 )Fe
(OOH)]
N
analytical law (c).
was completed by performing catalytic experiments with
[1,3,5-D3]benzene as a convenient mechanistic probe for de-
termining kinetic isotope effect (KIE) and to evidence nota-
ble hydride shift (NIH shift).[12] Careful analyses of the in-
termediate products obtained in the presence of mechanistic
probes and an EPR trapping study provided a valuable ex-
perimental basis in support of the mechanistic conclusions.
fitted as a function of time to first-order kinetics (Figure 1).
The fit between the data and the first-order rate law was
very good and allowed us to conclude that either the
FeOOH intermediate spontaneously degraded or oxidized
the solvent. Kinetic monitoring of the degradation of
[(L5 )FeIII
C
2
(kH/kD) of 1.1, which supports the latter hypothesis.
Results and Discussion
Similar experiments were performed in the presence of a
large excess of anisole or benzene to avoid the over-oxida-
2
Single-turnover experiments: The complex [(L5 )Fe
(OOH)]-
tion of products (see the Experimental Section). The degra-
2
(PF6)2 was prepared and isolated at ꢀ608C as previously de-
dation of [(L5 )Fe
N
scribed.[11a] The hydroxylation of anisole or benzene was per-
78C is shown in Figure 2. Under these experimental condi-
tions the absorbance maximum appeared at 523 nm; this dis-
crepancy was attributed to a difference in the solvent polari-
ty with respect to pure MeCN. The data was also well fitted
formed by using the stock solution of the isolated complex
2
[(L5 )Fe
(OOH)]
(PF6)2. The kinetics of the reactions were
monitored with time by UV/Vis spectroscopy by following
the decrease of the hydroperoxo!FeIII charge-transfer band
to first-order kinetics (Figure 2). Note, in the presence of
2
2
at around 520 nm, the characteristic band of [(L5 )Fe-
anisole, the rate of disappearance of [(L5 )Fe
N
A
E
lower than in pure MeCN (with rate constants of (0.160ꢁ
2
the [(L5 )FeACHTUNGTRENNUNG
(OOH)]2+ complex were initially studied by
0.001) vs. (0.316ꢁ0.005) minꢀ1 in 2:1 (v/v) MeCN/anisole
UV/Vis spectroscopy in pure CH3CN in the absence of any
substrate. Figure 1 shows the UV/Vis spectrum of a 0.33 mm
solution of [(L5 )Fe
progressive decrease in intensity with a clear isosbestic
point. The absorbance maximum was observed at 519 nm
under these conditions. These UV/Vis changes could be
and pure MeCN, respectively). The same reaction of
2
[(L5 )Fe
N
also showed similar first-order kinetics with rate constants
of (0.518ꢁ0.002) and (1.165ꢁ0.003) minꢀ1, respectively. The
EPR spectrum of the mixture at the end of the reaction at
218C revealed the presence of a rhombic low-spin FeIII com-
2716
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2715 – 2724

Hydroxylation of Aromatics
FULL PAPER
the Supporting Information for
anisole and benzene, respec-
tively). Importantly, the rate
constants for product formation
obtained at 78C for both sub-
strates can be directly com-
pared with the rate constant for
2
the degradation of [(L5 )Fe-
AHCTUNGTRENNUNG
(OOH)]2+. In the presence of
anisole, the extinction of the
2
visible band of [(L5 )Fe-
AHCTUNGTRENNUNG
(OOH)]2+ occurred with a rate
constant
of
(0.160ꢁ
0.001) minꢀ1, whereas the for-
mation of o-methoxyphenol
was observed with kobs
=
(0.131ꢁ0.022) minꢀ1.
It
is
therefore tempting to propose
that the disappearance of the
2
complex [(L5 )Fe
N
closely linked to the formation
of the products. A similar com-
ment can be made for the oxi-
2
Figure 2. a) Kinetic monitoring of the degradation of [(L5 )FeIII
G
2
of anisole and absorbance variation at 523 nm of [(L5 )FeIII
as a function of time at b) 7, c) 14, and d) 218C. Experimental data ( ) and fitting by a first-order analytical
law (c).
G
dation of benzene because
2
[(L5 )Fe
N
plex (g=2.33, 2.15, 1.94; see the Supporting Information).
2
This spectrum is very similar to that of [(L5 )FeIII
(OMe)]2+
(g=2.33, 2.14, 1.94)[11a] and can thus be formulated as
[(L5 )FeIII(OH)]2+
.
2
Analysis of the oxidation products of anisole obtained at
7 and 218C revealed that the yields, as well as the chemo-
and regioselectivities, were very similar (Table 1).
Table 1. Yields of the oxidation products of anisole and the first-order
2
rate constants deduced from the degradation of [(L5 )Fe
G
experiments performed under single-turnover conditions.[a]
Yield [%]
k [minꢀ1
]
o-OH[b]
p-OH[b]
Phenol
2
Figure 3. Yields (vs. [(L5 )FeIII
(OOH)]2+) of PhOH ( ) and o-methoxy-
^
T=218C
T=78C
27
26
41
38
8
3
4
2
4
1.165ꢁ0.003
0.160ꢁ0.001
0.106ꢁ0.001
0.114ꢁ0.001
&
phenol ( ) as a function of time during the oxidation of benzene or ani-
9.5
20
18
2
sole, respectively, by [(L5 )FeIII
(OOH)]2+ at 78C. First-order fittings are
ACHTUNGTRENNUNG
T=78C, 30 PhSH
T=78C, 60 PhSH
2
shown as solid lines.
[a] [(L5 )FeACHTUNGTRENNUNG
(OOH)]2+/anisole/PhSH: 1/9200/0, 30, or 60, as indicated.
2
Yields given with respect to [(L5 )Fe
methoxyphenols are denoted as o-OH and p-OH, respectively.
(OOH)]2+
ACHTUNGTERNNUGN . [b] The o- and p-
k
k
obs =(0.149ꢁ0.001) minꢀ1 and the phenol was formed with
ACHTUNGTRENNUNG
obs =(0.171ꢁ0.009) minꢀ1.
The products of anisole oxidation were formed with yields
For the oxidation of anisole or benzene at 78C, the rates
of formation of o-methoxyphenol and phenol, respectively,
were determined by GC analysis (Figure 3). The vast majori-
ty of the products were formed during the first 20 min for
both substrates and reached a plateau after around 30 min.
of less than 100% (Table 1), which suggests that some of the
oxidant may be consumed in side-reaction(s). The oxidation
of the solvent could be a competing pathway in the degrada-
2
tion of [(L5 )Fe
G
.
These data may be rationalized by the following simple
The first-order kinetics for product formation mirror those
reaction scheme in which FeIIIOOH inserts one oxygen
2
ꢀ
for the degradation of [(L5 )Fe
G
atom into an aromatic (ArH) C H bond to yield the phe-
2
these substrates. [(L5 )Fe
(OOH)]2+ had almost totally react-
nolic product (ArOH) [Eq. (1)] or oxidizes the solvent
MeCN. Under the experimental conditions, that is, with
ed after 30 min under the same conditions (see Figure 2 and
Chem. Eur. J. 2012, 18, 2715 – 2724
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2717

A. B. Sorokin, F. Banse et al.
much larger substrate and solvent concentrations than
2
[(L5 )Fe
tion would be given by Equations (2) and (3) in which kobs
ArHA MeCNA
[ArH]b +k
ing Information).
ACHTUNGTRENNUNG
(OOH)]2+, the rate of the pseudo-first-order reac-
=
k
G
[MeCN]c and a=a’=1 (see the Support-
FeIIIOOH þ ArH ! FeIIIOH þ ArOH
ð1Þ
a
b
a0
c
V ¼ kArH½FeIIIOOHꢂ ½ArHꢂ þ kMeCN½FeIIIOOHꢂ ½MeCNꢂ
ð2Þ
ð3Þ
V ¼ kobs½FeIIIOOHꢂ
However, this proposal is not compatible with two experi-
mental observations: 1) In the presence of an aromatic sub-
strate, the reaction rate is lower than in pure MeCN and
2) the rate constants seem not to be dependent on the aro-
matic substrate (anisole or benzene). Moreover, changing
the concentration of anisole by a factor of two had no effect
on the value of kobs, which reveals that neither the simple
mechanism written above nor the expression for kobs is cor-
rect. These surprising observations suggest that 1) a mecha-
nistic scheme with parallel reactions to account for the deg-
radation of FeIIIOOH is not correct and 2) the FeIIIOOH
should form an active species capable of aromatic oxidation
in a rate-limiting step.
Beneficial effect of a reducing agent under single-turnover
conditions: The addition of a reducing agent, such as thio-
phenol, naphthol, or hydroquinone, to the reaction mixture
has previously been shown to increase the yields of products
under catalytic conditions for different FeII complexes.[10]
Therefore, to identify the role played by these reductants,
we have performed single-turnover oxidation of anisole in
2
Figure 4. a) X-band EPR spectra recorded at 5 K of [(L5 )FeIII
(OOH)]-
AHCTUNGTRENNG(NU PF6)2 dissolved in butyronitrile, the so-called stock solution used for all
2
ACHTUNGTRENNUNG
single-turnover experiments (top) and of aliquots of [(L5 )FeIII(OOH)]2+
in the presence of 60 equiv PhSH after 3 (middle) and 8 min of reaction
(bottom) at 78C. b) High-field region displaying the resonances of the
2
[(L5 )FeIII
(OOH)]2+ intermediate. For clarification purposes, the intensi-
ties of the spectra have been arbitrarily normalized.
the presence of 30 or 60 equivalents of PhSH with respect to
2
[(L5 )Fe
G
by GC-MS. A normal KIE of 1.05 was determined, which
2
uct yields in the presence of PhSH, in accord with the re-
sults obtained under catalytic conditions (Table 1). The UV/
Vis spectrum of [(L5 )Fe
presence of PhSH and the monitoring of its reaction at 78C
showed no dependence of the rate constant on the concen-
tration of the reducing agent (see Table 1 and the Support-
ing Information). Consequently, the concentration of PhSH
does not have to be taken into account in the rate law. In
reveals the preference of [(L5 )FeCl]+ for aromatic C H
ꢀ
ꢀ
bond breaking over C D bond breaking.
To gain a deeper insight into the mechanism, FeII com-
plexes with the ligands depicted in Scheme 1, that is,
2
2
2
[(L4 )FeCl2], [(L5 )FeCl]+, [(TPEN)Fe]2+, and [(L6 4E)Fe]2+,
were used as catalysts for the oxidation of [1,3,5-D3]benzene
by H2O2 in acetonitrile. The isotopic compositions of the pri-
mary and secondary products of the reaction, that is, phenol
and benzoquinone, respectively, were analyzed by GC-MS
according to the method reported previously to determine
KIE and the occurrence of NIH shift.[12c] This method al-
lowed intramolecular kH/kD values to be determined by
comparing the intensities of the MS molecular peaks of [3,5-
D2]phenol and [2,4,6-D3]phenol obtained in the catalytic ox-
idation of [1,3,5-D3]benzene. In addition, it was possible to
see the NIH shift by analysis of the isotopic distribution of
the 1,4 benzoquinone molecular peak, which revealed the
formation of both [D1]BQ and [D3]BQ. (see Scheme 2).
The oxidation reaction was performed for 2 h for each
catalyst and the reaction mixture was analyzed every
30 min. The KIE values and isotopic compositions of the
addition, a more sensitive technique, such as EPR spectros-
copy, did not reveal any change in the [(L5 )Fe
spectrum on addition of PhSH (see Figure 4). Therefore
PhSH most likely does not act as a ligand to the metal
center. However, the rate of the reaction was reduced when
PhSH was present in the reaction mixture (rate constants of
ca. 0.11 and 0.16 minꢀ1 in the presence and absence of
PhSH, respectively). All of these observations lead to the
conclusion that the role played by the PhSH is subtle.
Kinetic isotope effects and hydride shift: The competitive
oxidation of a stoichiometric mixture of C6H6 and C6D6 by
H2O2 in acetonitrile catalyzed by [(L5 )FeCl]+ was analyzed
2718
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2715 – 2724

Hydroxylation of Aromatics
FULL PAPER
tribution of 12.9/5.9 is of the same order as those deter-
mined under catalytic conditions (see Table 2), but owing to
the low concentration of quinones (six-electron oxidation
product) under these peculiar conditions, the experimental
error is probably higher than under catalytic conditions. The
value of KIE is close to that obtained with the catalyst
[(TPEN)Fe]2+. This points to the same reaction pathway
mediated by FeIII
and single-turnover conditions. Indeed, in the presence of
H2O2, the [(TPEN)Fe
(OOH)]2+ intermediate presents a
ACHTUNGTRENUN(NG OOH) intermediates under both catalytic
AHCTUNGTRENNUNG
first coordination sphere with the hydroperoxo and five ni-
trogen ligands, with a pendant pyridine, identical to that of
2
the isolated [(L5 )Fe
G
2
with the [(L5 )FeCl]+ catalyst is somewhat different. The
presence of a chloro ligand in [(L5 )FeCl]+ may complicate
Scheme 2. Primary and secondary oxidation products of [1,3,5-
D3]benzene used to determine KIEs and NIH shift.
the reaction pattern, as has been observed for similar cata-
lysts.[15]
products were determined by GC-MS on the basis of all
scans of the product peaks according to the method previ-
ously described.[13] The results obtained (kH/kD and [D1]BQ,
[D2]BQ, and [D3]BQ distribution) were constant within ex-
perimental error throughout the reaction for all catalysts.
The mean values obtained after 30, 60, 90, and 120 min of
reaction for the four complexes are presented in Table 2.
In each case, normal KIE values (kH/kD >1) were ob-
tained. The KIE was notably higher for the complexes with
Mechanistic hypotheses for aromatic hydroxylation in the
presence of FeIII
ACTHNUGRTENUNG(OOH): Several mechanistic hypotheses can
be put forward to explain the results obtained in this study.
The most important point is to determine the nature of the
active species competent for the oxidation of benzene and
anisole. Towards this goal, three possible mechanisms can
be considered [see Eqns. (4–6)].
FeIIIOOH þ ArH ! FeIIIOH þ ArOH
2
ð4Þ
ð5Þ
ð6Þ
the TPEN and L6 4E ligands. Experimental and theoretical
studies dedicated to the aromatic hydroxylation by cytochro-
mes P450 frequently evoke electrophilic addition of the
FeV(O) compound I to the p system of the substrate.[14] This
results in the formation of a s complex in which the target
carbon atom has changed from sp2 to sp3 hybridization. This
is typically accompanied by an inverse KIE, which is not
compatible with our results.
Importantly, [D1]BQ and [D3]BQ were detected in all the
labeling experiments, which indicates a notable hydride shift
during aromatic hydroxylation. All these observations are in
favor of a mechanism that involves the participation of an
iron complex. Note that in contrast to the N-bridged diiron–
phthalocyanine system,[12c] no benzene oxide was detected
by GC-MS.
ArH
FeIIIOOH ! FeVO þ HOꢀ
FeIIIOH þ ArOH
FeIIIOH þ ArOH
ƒƒ!
ArH
III
IV
C
Fe OOH ! Fe O þ HO
ƒƒ!
1)
Can the FeIII
U
ꢀ
omatic C H bond [Eq. (4)]. Although both haem and
non-haem FeIII
A
considered as “sluggish oxidants”,[7b,16] there are some lit-
erature data that support their oxidizing properties.[4,17]
In principle our kinetic results, which reveal similar first-
2
order rate constants for the degradation of [(L5 )Fe-
AHCTUNGTRENNUNG
(OOH)]2+ and for the formation of o-methoxyphenol
and phenol (products of the oxidation of anisole and
benzene, respectively), are in agreement with the direct
Single-turnover experiments also revealed a normal KIE
(kH/kD =1.53ꢁ0.16) and NIH shift. The [D3]BQ/[D1]BQ dis-
participation of FeIII
ACHTUNGTRENNUNG
of [(L5 )FeACTHNUTRGNEUNG
(OOH)]2+ with CH3CN (DH° =
2
Table 2. Determination of the KIE and NIH shift for the oxidation of [1,3,5-
D3]benzene by H2O2 catalyzed by FeII complexes.
81 kJmolꢀ1 and DS° =ꢀ1 JKꢀ1 molꢀ1, see the
2
2
2
Supporting Information) also support this hy-
G
C
G
ACHTUNGTRENNUNG
pothesis as they are compatible with the direct
kH/kD at 208C[a]
1.16ꢁ0.03
1.18ꢁ0.02
9.3/3.7
1.63ꢁ0.04
1.67ꢁ0.02
18.1/3.9
2
[D3]BQ/[D1]BQ[b]
kH/kD at 208C+EtPhSH
[D3]BQ/[D1]BQ[b]
7.3/n.d.[c]
17.3/5.6
oxidation of CH3CN by [(L5 )Fe
A
[d]
–
1.58ꢁ0.02
–
–
1.89ꢁ0.02
contrast, the activation entropy determined in
the presence of anisole (DH° =98 kJmolꢀ1 and
DS° =57 JKꢀ1 molꢀ1) is compatible with a disso-
–
16.3/4.5
16.6/3.5
[a] For each sample, the composition was the following: Fe/H2O2/[D3]benzene/EtPhSH
(1:20:230:ꢁ10). The intramolecular kH/kD value was determined by comparing the in-
tensities of the MS molecular peaks of [3,5-D2]phenol and [2,4,6-D3]phenol.
[b] [D3]BQ and [D1]BQ amounts are given as percentages of the total amount of BQ.
[c] Amount of [D1]BQ could not be precisely determined because of the small quanti-
ty of product obtained. [d] Not measured.
ciative transition state. However, the rate of
2
degradation of [(L5 )Fe
N
ent of both the nature of the substrate and of its
concentration (see above). The latter results
Chem. Eur. J. 2012, 18, 2715 – 2724
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2719

A. B. Sorokin, F. Banse et al.
argue against the first possible scenario and suggest that
[(L5 )Fe
agent of aromatic substrates.
In the presence of large amounts of substrate, the concen-
tration of acetonitrile and therefore the formation of
2
ACHTUNGTRENNUNG
(OOH)]2+ is probably not a direct hydroxylating
C
CH2CN would be significantly decreased. As a result, the
ꢀ
2) A second possibility is the heterolytic cleavage of the O
rate of degradation would also decrease. Along the same
line, the beneficial effect of the reducing agent PhSH (in-
crease in the yields of aromatic oxidation products) could be
V
O bond in FeIII
A
¯
[Eq. (5)]. This mechanism has recently been proposed
for the hydroxylation of aromatic rings promoted by
H2O2 and FeII complexes with related tetradentate li-
gands.[18] A key point for the observation of such a heter-
olytic cleavage was proposed to be the presence of a
labile coordination site (solvent molecule) adjacent to
C
explained by the trapping of the CH2CN radical by PhSH,
thus further reducing the consumption of FeIII
(OOH) in sec-
ondary oxidation reactions and decreasing its rate of degra-
dation. As a result, the amount of FeIII
(OOH) available for
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
ArH oxidation should be increased leading to the increase
in the yield of the ArH oxidation products, as observed (see
Table 1). The small effect of the presence of the reducing
agent on the KIE value and the retention of deuterium in
the product phenol suggest that PhSH probably does not
participate in the reaction steps leading to the reaction
products. This is in full agreement with the independence of
the rate constant on the reducing agent concentration. Note
that the absence of spectroscopic changes in the presence of
PhSH and the decrease in the overall oxidation rate are also
compatible with this assumption.
the hydroperoxo in the FeIII
ACTHNUTRGNE(UNG OOH) intermediate, which
seems to be impossible in our complexes. Taking into ac-
count the high reactivity of FeV=O species, it may not be
detectable by UV/Vis spectroscopy under our experi-
mental conditions. The rate-determining step would be
ꢀ
the cleavage of the O O bond and once formed the pu-
tative FeV=O species would react very rapidly with
MeCN or the aromatic substrate ArH. However, FeV=O
complexes with these kinds of neutral ligands have fre-
quently been postulated,[19] but their trapping has been
suggested or achieved only recently.[20] In addition, theo-
ꢀ
This third mechanistic scenario, that is, homolytic O O
IV
retical calculations are in favor of a homolytic O O
cleavage of FeIII
U
ꢀ
C
cleavage for FeIII
(OOH) complexes with pentadentate li-
pears to be the most compatible with the experimental data.
Complementary experiments were performed to check its
validity.
gands.[21] Furthermore, a decrease in solvent polarity in
the presence of ArH should not facilitate the formation
2
of two charged species, [(L5 )FeV=O]3+ and OHꢀ, under
In contrast to the kinetic monitoring of the degradation of
2
2
our experimental conditions. Moreover, if [(L5 )FeV=
[(L5 )FeIII
N
O]3+ was involved in the aromatic hydroxylation, one
could expect the intermediate formation of benzene
oxide and an inverse KIE, which is usually associated
with the intermediate formation of a s complex. All
these arguments tend to refute this hypothesis.
vealed a small normal KIE (kH/kD) of 1.1, no KIE effect
2
could be evidenced during the degradation of [(L5 )FeIII-
AHCTUNGTRENNUNG
(OOH)]2+ in these two solvents in the presence of anisole
(kH/kD =1). This observation is in agreement with a minor
contribution of solvent-derived depletion of FeIII
AHCTUNGTRENNUNG
3) Finally, the third possibility to account for the results ob-
the presence of ArH. As expected, the oxidation of anisole
2
served is the transformation of FeIII
A
catalyzed by [(L5 )FeIICl]+ showed a significant increase (by
IV
ꢀ
O O cleavage into short-living transients, Fe =O and
the OH radical [Eq. (6)]. The OH radical should be
a factor of ca. two, see the Supporting Information) in the
product yields in CD3CN with respect to CH3CN. On the
other hand, in the presence of PhSH, the yields of phenol
products increased but were almost the same in CH3CN and
CD3CN. These observations are compatible with the hypoth-
·
competent to oxidize both aromatic compounds by elec-
trophilic addition and CH3CN by abstraction of a hydro-
C
gen atom to generate the CH2CN radical. This latter
C
fleeting species can be involved in secondary oxidation
esis that 1) solvent-derived radicals ( CH2CN) react with
FeIII
(OOH) intermediates in a faster step and 2) PhSH is
able to trap these radicals avoiding their reaction with FeIII-
(OOH).
Identification of the products of PhSH transformation by
reaction(s) thus consuming further FeIII
(OOH) equiva-
ACHTUNGTRENNUNG
lents in a step faster than the initial hydrogen-atom ab-
straction. In pure CH3CN, these secondary reactions
would contribute to the principal pathway of degradation
AHCTUNGTRENNUNG
of FeIII
G
GC-MS provided further evidence in favor of this hypothe-
sis (see the Supporting Information). The main product was
PhSSPh along with a trace amount of PhSPh. Importantly,
when the oxidation was performed in CH3CN, a very small
amount of another product was detected that showed a mo-
lecular peak at m/z 149 with a principal fragmentation peak
at m/z 109. These two peaks can be tentatively assigned to
[PhSCH2CN]+ (M=149) and [MꢀCH2CN]+ (M=149ꢀ40=
109). In CD3CN, the yield of the product at the same reten-
tion time was lower, as expected. Its mass spectrum exhibits
a molecular peak at m/z 151 due to [PhSCD2CN]+ with the
principal fragmentation peak at m/z 109 assigned to
A similar value of DS° (ꢀ9 JKꢀ1 molꢀ1) was obtained for
the degradation in CH3OH of [(L5 )Fe
which L5 contains a propanediamine instead of an
ethane
N
activation enthalpy is about three times higher than the
corresponding values for the formation of FeIV=O from
the heterolytic cleavage of FeII–H2O2 adducts.[22] In addi-
tion, the activation entropy for the latter process is large-
ly negative, which suggests that the mechanism for the
2
evolution of [(L5 )Fe
N
2720
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2715 – 2724

Hydroxylation of Aromatics
FULL PAPER
[MꢀCD2CN]+ (see the Supporting Information). This find-
ing provides unambiguous proof that indeed the reducing
agent traps solvent-derived radicals, which leads to an in-
The possible assignment of the triplet signal shown in
ꢀ
Figure 5 to the DMPO (OH)2 radical was confirmed by
GC-MS (see the Supporting Information). Samples of the
reaction mixture containing 30 equivalents of DMPO were
withdrawn every 2 min, quenched in liquid nitrogen, and an-
alyzed by GC-MS. During repetitive injections, along with
DMPO (m/z 113), a product with m/z 129 ([M+O]+) was
detected and another at m/z 113 (with a fragmentation pat-
tern different to that of DMPO) that we are still unable to
assign. Importantly, a product with a molecular peak at m/z
147 ([M+2OH]+) with a longer retention time was detected
transiently. These observations can be rationalized within
the framework of the mechanism shown in Scheme 3. The
crease in the amount of Fe
ring hydroxylation.
ACHTUNGTNER(NUNG OOH) available for aromatic
The oxidation of substituted aromatic rings by H2O2 cata-
lyzed by this family of FeII complexes[10] or under single-
2
turnover conditions by [(L5 )Fe
(OOH)]2+ always shows a
large preference for ortho-hydroxylated products (Table 1).
Such regioselectivity is not compatible with the involvement
C
of freely diffusing HO radicals. Anisole was oxidized in the
C
presence of the HO trapping agent 5,5-dimethyl-1-pyrroline
2
N-oxide (DMPO, 30 equiv with respect to [(L5 )FeCl]+
under the conditions described in the Experimental Sec-
ꢀ
tion). No DMPO OH radical was detected by EPR spec-
troscopy, which reveals that no freely diffusing hydroxyl rad-
icals were formed. Instead of the EPR spectrum exhibiting
the typical doublet of triplets with a 1:2:2:1 intensity ratio
[23]
ꢀ
characteristic of the DMPO OH radical, an unusual trip-
let EPR signal (aN =13.1 G) was detected (Figure 5). This
Scheme 3. Reaction of DMPO with FeIII
ACTHNUTRGNE(UNG OOH). This proposed mecha-
nism includes the products detected by GC-MS (indicated by their m/z)
or EPR spectroscopy.
intermediate FeIII
ACTHUNTGRENNG(U OOH) yields a pair of reactive species,
2
Figure 5. X-band EPR spectra of [(L5 )FeIICl]+ +H2O2 +anisole+DMPO
IV
(1:20:3000:30) recorded at 292 K. Reaction time increasing from top to
middle spectrum. The bottom spectrum is a simulation of the spectrum
with the following parameters: g=2.0062, aN =13.1 G, linewidth=1.3 G.
ꢀ
C
Fe (O) and OH , by homolytic O O cleavage. These two
species react very rapidly with a molecule of DMPO without
C
escaping from the solvent cage: First, OH adds to DMPO
ꢀ
and then the resulting DMPO OH radical is oxidized by
signal is reminiscent of the one resulting from the double
addition of HO· to DMPO, although a value of aN of 15.3 G
has been published for such a species.[24] Because this latter
adduct is not a primary oxidation product in the Fenton re-
action, it was detected as a background EPR signal that was
FeIV(O) following hydrogen-atom abstraction. These con-
certed steps give a product with m/z 129, which explains
·
ꢀ
why the classic DMPO OH adduct was not observed in our
experiments. In the following step, addition of OH· (from
IV
C
another [Fe (O) OH ] pair) to the latter product gives a
[24]
ꢀ
dominated by DMPO OH.
In contrast, in our experi-
radical species compatible with the triplet EPR signal shown
in Figure 5. Under the experimental GC-MS conditions, this
radical species cannot be detected, but a product with m/z
147 was observed transiently. This intermediate may be
formed from the radical species after abstraction of one hy-
drogen atom in the medium. Finally, the product with m/z
147 evolves by dehydration to give a product with m/z 129.
ments we observed the unusual triplet signal as one clean
signal (see Figure 5). In our control experiments, Fenton re-
actions were carried out under conventional conditions or in
ꢀ
pure acetonitrile. In both cases, the usual DMPO OH
signal was observed without any trace of the triplet signal,
discarding the possibility that the solvent is responsible for
this unexpected signal (see the Supporting Information). In
contrast, our results support an unusual specific oxidation of
Mechanism of aromatic hydroxylation initiated by
2
2
DMPO by the [(L5 )FeCl]+/H2O2 system.
[(L5 )FeIII
(OOH)]2+ and analogous complexes: The data
Chem. Eur. J. 2012, 18, 2715 – 2724
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2721

A. B. Sorokin, F. Banse et al.
ꢀ
collected in the course of the aromatic hydroxylation experi-
ments can be rationalized within the framework of the ten-
tative mechanism shown in Scheme 4. In an initial step, the
proposal of a mechanism in which the FeO OH intermedi-
ate undergoes the so-called “somersault isomerization” to
give a hydroxyl radical hydrogen-bonded to the ferryl
oxygen.[29] The conclusions of this study are compatible with
our results.
Conclusion
Aromatic hydroxylation by H2O2 catalyzed by non-haem
FeII complexes with amine/pyridine ligands has been thor-
oughly studied and revealed the implication of FeIII
ACHTUNGTRENNUNG
IV
Scheme 4. Mechanism proposed for the aromatic hydroxylation initiated
C
[Fe (O)+ OH]. The results point towards the concerted
by non-haem FeIII
ACHTUNGTRENNUNG
action of both intermediates in the formation of the phenol
product with a normal isotope effect (KIE>1) and NIH hy-
dride shift. No free hydroxyl radical has been evidenced in
our experiments. The beneficial effect provided by the re-
ducing agent PhSH can be explained by the trapping of sol-
FeIII
ACHTUNGTRENNUNG(OOH) complex undergoes homolytic cleavage of the
IV
ꢀ
C
O O bond to form Fe =O and the OH radical in a cage.
Such an intimate pair of oxidants has already been postulat-
ed to account for aromatic hydroxylation catalyzed by iron
complexes in solution[15] or embedded in a proteic matrix.[25]
The hydroxyl radical first attacks the aromatic ring to form
a transient hydroxycyclohexadienyl radical, which is further
oxidized by the encaged FeIV=O to yield a cationic inter-
mediate. This cationic intermediate rearranges to a ketone.
The first reaction step (addition of the hydroxyl radical with
the formation of a radical s complex) should exhibit an in-
verse isotope effect (KIE<1), but ketonization, which in-
C
vent-derived radicals, such as CH2CN, which protects the
FeIII
(OOH) from being consumed in undesirable side-reac-
tions. Thus, the use of reducing agents results in a greater
amount of FeIII
(OOH) available for aromatic oxidation and
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
higher yields of the oxidation products.
Finally, one of the initial objectives of this work was to
evaluate the oxidizing properties of FeIII
ACTHNUTRGNE(NUG OOH) complexes.
Our results do not demonstrate that these species directly
hydroxylate aromatic rings. On the basis of several lines of
experimental evidence obtained for the first time, we have
ꢀ
cludes C H(D) cleavage, should exhibit a normal isotope
shown that FeIII
ACTHNGUTERN(UNG OOH) complexes are precursors of two
effect (KIE>1). The transfer of H(D) during the ketone in-
termediate formation step gives rise to a migration of the
deuterium atom resulting in an NIH shift as already report-
ed.[26] Finally, re-aromatization of the ketone intermediate,
caged oxidants that are able to perform the demanding aro-
matic oxidation in a concerted way with high efficiency.
Like cytochrome P450 compound I (i.e., formally FeV(O)),
these two one-electron oxidants are two oxidizing equiva-
lents above the ferric resting state. Designing ligands with a
hydrophobic pocket to enhance the concerted action of
these two species may lead to a significant improvement of
the oxidation reactions in terms of specificity. In addition,
water-soluble catalysts applied in an aqueous medium may
help to avoid competitive solvent oxidation.
ꢀ
which includes another C H(D) cleavage step with KIE>1,
leads to the phenol product with partial retention of the
deuterium atom. Overall, the observed KIE should be a
ꢀ
combination of the KIEs of the three steps involving C
H(D) bond transformations on a similar timescale. All the
complexes studied in this work show similar KIE values,
ranging from 1.16 to 1.67, which suggests that the same
mechanism is operating with all four iron complexes.
In summary, we propose that aromatic oxidation is ach-
ieved by the concerted action of two one-electron oxidants
rather than one two-electron oxidant as in the case of cyto-
chrome P450 compound I. The intervention of the hydroxyl
radical in the first step and the FeIV(O) intermediate in the
second step can be justified by the high reactivity of the
former and by the relative stability of the latter. The relative
stability of FeIV(O) complexes has allowed the isolation of
such species as single crystals.[27] Furthermore, it has previ-
ously been shown that non-haem FeIV(O) species are not
competent to perform this aromatic oxidation.[9f,18b,28] Inter-
estingly, in this case it has the ability to assist and control
Experimental Section
Materials: Reagents were purchased from Acros. Solvents were pur-
chased from VWR and used without further purification as it was shown
to be unnecessary in this study. The substrates were purified on alumina
before use.
2
Single-turnover reactivity: The complex [(L5 )Fe
(OOH)]
N
pared and precipitated as previously described.[11a] Due to its thermal in-
stability as a powder and for easier handling of the sample, it was washed
several times on a glass frit at ꢀ608C with cold diethyl ether and then re-
dissolved in the minimum amount of cold butyronitrile. The integrity of
2
the [(L5 )Fe
(OOH)]2+ complex in the butyronitrile solution as well as its
concentration was determined by X-band EPR spectroscopy at 5 K. It
was determined to contain 6.75 mm of FeIIIOOH by double integration of
its EPR spectrum with respect to CuII, which corresponded to an extinc-
tion coefficient of 1000mꢀ1 cmꢀ1 for the hydroperoxo!FeIII charge-trans-
fer absorption band at around 520 nm. Note that this latter value is iden-
C
the very oxidizing species HO , which results in a reaction
with a noticeable KIE and NIH shift. Note that a recent
DFT study of Fenton-like hydrocarbon oxidation led to the
2722
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2715 – 2724

Hydroxylation of Aromatics
FULL PAPER
tical to that reported for the in situ generated complex in MeOH.[30] The
Kinetics of product formation: The products of anisole and benzene oxi-
2
2
dation catalyzed by [(L5 )Fe
N
solution of [(L5 )Fe
E
T
raphy. Aliquots of the reaction mixtures were collected, frozen in liquid
nitrogen to stop the reaction, and then gently thawed just before injec-
tion in the chromatograph. For anisole, only the yield of the major prod-
uct, o-methoxyphenol, was considered for the kinetic analysis.
stock solution for the reactivity experiments. The reactivity experiments
were performed by using two different substrates, anisole and benzene,
under aerobic conditions in CH3CN at 7, 14, or 218C over 2 h. The exper-
imental conditions were set up to be similar to our previous studies per-
formed under catalytic conditions,[10] that is, with a large excess of sub-
strate with respect to the oxidant. The volume of each reaction mixture
EPR spectroscopy: EPR spectra were recorded on a Bruker Elexsys 500
spectrometer at 9.4 GHz at 5 or 100 K for the analysis of the iron com-
was 6 mL (4 mL CH3CN+2 mL substrate, which corresponds to
a
plexes, and at 100 and 292 K for the spin-trapping experiments. The con-
2
centration of the low-spin [(L5 )Fe
N
CH3CN/substrate molar ratio of 3.5 or 4.2, for benzene and anisole, re-
tion was determined by double integration with the software Xepr.[34]
A
2
spectively). The stock solution of [(L5 )Fe
N
1 mm CuSO4·5H2O stock solution containing 2m HCl and 2m Na-
ClO4·5H2O was used to prepare standards with 1, 0.75, 0.5, and 0.25 mm
CuII.
the final reactant to trigger the reaction and to reach a concentration of
0.33 mm in the reaction mixture. The volume of this added stock solution,
being less than 5% of total, was considered to be negligible. The
2
substrate/
[(L5 )Fe
N
experimental conditions of our previous studies performed under catalyt-
2
ic conditions. The decay of [(L5 )Fe
N
Acknowledgements
absorption spectroscopy at 523 nm. The products were identified and
quantified by gas chromatography (yields given in % with respect to
2
The authors thank the French program ꢃEnergie, Conception Durable
2004ꢄ (ACI ECD009 BioCatOx) for financial support, and Dr. Antoine
Tissot and Dr. Pierre Dorlet for their kind help. V.J. gratefully acknowl-
edges the Conseil Rꢀgional d’Ile-de-France for a Ph. D. grant.
[(L5 )Fe
N
as described below, but by using concentrated samples (in a stream of
argon without heating) and a splitless mode because of the very small
concentrations of the products. The substrate/oxidant ratio of the sam-
2
ples, [1,3,5-D3]benzene/
[(L5 )Fe
N
lution was introduced as two equivalent portions at t=0 and t=60 min,
to favor the formation of the benzoquinones.
[1] a) B. Meunier, S. P. de Visser, S. Shaik, Chem. Rev. 2004, 104, 3947–
3980; b) P. R. Ortiz de Montellano, Chem. Rev. 2010, 110, 205.
[2] a) R. M. Burger, Struct. Bonding 2000, 97, 287–303; b) M. Pitie, G.
Pratviel, Chem. Rev. 2010, 110, 1018–1059.
[3] A. Decker, M. S. Chow, J. N. Kemsley, N. Lehnert, E. I. Solomon, J.
Am. Chem. Soc. 2006, 128, 4719–4733.
[4] F. Neese, J. M. Zaleski, K. L. Zaleski, E. I. Solomon, J. Am. Chem.
Soc. 2000, 122, 11703–11724.
[5] D. Kumar, H. Hirao, S. Shaik, P. M. Kozlowski, J. Am. Chem. Soc.
2006, 128, 16148–16158.
[6] S. V. Kryatov, E. V. Rybak Akimova, S. Schindler, Chem. Rev. 2005,
105, 2175–2226.
[7] a) S. P. de Visser, A. M. Valentine, W. Nam, Angew. Chem. 2010,
122, 2143–2146; Angew. Chem. Int. Ed. 2010, 49, 2099–2101; b) E.
Derat, D. Kumar, H. Hirao, S. Shaik, J. Am. Chem. Soc. 2006, 128,
473–484.
[8] a) G. Yin, A. M. Danby, D. Kitko, J. D. Carter, W. M. Scheper, D. H.
Busch, Inorg. Chem. 2007, 46, 2173–2180; b) G. C. Yin, M. Buchalo-
va, A. M. Danby, C. M. Perkins, D. Kitko, J. D. Carter, W. M. Schep-
er, D. H. Busch, Inorg. Chem. 2006, 45, 3467–3474; c) A. Bakac,
Inorg. Chem. 2010, 49, 3584–3593; d) A. Kunishita, J. Teraoka, J. D.
Scanlon, T. Matsumoto, M. Suzuki, C. J. Cramer, S. Itoh, J. Am.
Chem. Soc. 2007, 129, 7248–7249.
Reactivity under catalytic conditions: Experiments were carried out in
acetonitrile at 208C with [1,3,5-D3]benzene (Sigma–Aldrich, 98%) as a
probe substrate for KIE (kH/kD) and NIH shift determination and the
2
2
complexes [L4 FeCl2],[31] [(L5 )FeCl]
G
(PF6)2,[10b] or
ACHTUNGTRENNUNG
2
[10b]
[(L6 4E)Fe]ACHTUNGTRENNUNG(PF6)2
the reaction, that is, the phenols and benzoquinone, respectively, were
analyzed by GC-MS to determine the KIEs and occurrence of the NIH
shift following the method previously reported by some of us.[12c] The
composition of the reaction mixtures were Fe/H2O2/ACHTNUTRGNE[UNG 1,3,5-D3]benzene in
a ration of 1:20:230 with a concentration of 1 mm of the iron complex.
The substrate/H2O2 ratio was less than in our previous catalytic experi-
ments to allow the formation and detection of the secondary oxidation
products benzoquinones, which provide evidence for the NIH shift. The
experiments in the presence of 5,5-dimethyl-1-pyrroline N-oxide
(DMPO) as spin trap were carried out by using 30 equiv DMPO with re-
spect to the iron complex, all other conditions being similar to our previ-
2
ous studies,[10] that is, [(L5 )FeCl]+/H2O2/anisole/ꢁPhSH in a ratio of
ꢀ
1:20:3000:ꢁ20. Control experiments for the detection of the DMPO
-OH radical were carried out by using FeSO4·7H2O by the Fenton reac-
tion following a known procedure[32] or under similar conditions but by
using acetonitrile as solvent instead of water.
[9] a) V. Balland, F. Banse, E. Anxolabehere Mallart, M. Nierlich, J. J.
Girerd, Eur. J. Inorg. Chem. 2003, 2529–2535; b) V. Balland, F.
Banse, E. Anxolabehere-Mallart, M. Ghiladi, T. A. Mattioli, C. Phil-
ouze, G. Blondin, J. J. Girerd, Inorg. Chem. 2003, 42, 2470–2477;
c) M. Martinho, F. Banse, J. Sainton, C. Philouze, R. Guillot, G.
Blain, P. Dorlet, S. Lecomte, J. J. Girerd, Inorg. Chem. 2007, 46,
1709–1717; d) A. J. Simaan, F. Banse, J. J. Girerd, K. Wieghardt, E.
Bill, Inorg. Chem. 2001, 40, 6538–6540; e) A. J. Simaan, S. Dçpner,
F. Banse, S. Bourcier, G. Bouchoux, A. Boussac, P. Hildebrandt, J. J.
Girerd, Eur. J. Inorg. Chem. 2000, 1627–1633; f) A. Thibon, J.-F.
Bartoli, S. Bourcier, F. Banse, Dalton Trans. 2009, 9587–9594.
[10] a) V. Balland, D. Mathieu, Y. M. N. Pons, J. F. Bartoli, F. Banse, P.
Battioni, J. J. Girerd, D. Mansuy, J. Mol. Catal. A Chem. 2004, 215,
81–87; b) A. Thibon, J.-F. Bartoli, R. Guillot, J. Sainton, M. Martin-
ho, D. Mansuy, F. Banse, J. Mol. Catal. A Chem. 2008, 287, 115–120.
[11] a) M. Martinho, P. Dorlet, E. Riviere, A. Thibon, C. Ribal, F. Banse,
J.-J. Girerd, Chem. Eur. J. 2008, 14, 3182–3188; b) A high spin (S=
5/2) synthetic FeIIIOOH was previously prepared as a powder, but
the solid was not characterized in detail, H. Wada, S. Ogo, S. Naga-
tomo, T. Kitagawa, Y. Watanabe, K. Jitsukawa, H. Masuda, Inorg.
Chem. 2002, 41, 616–618.
Analysis of the oxidation products: The products of the single-turnover
reactions were analyzed by GC after 2 h on an Agilent 6890 instrument
with an BP20 column. The products of the catalytic reactions were ana-
lyzed by GC-MS using a Hewlett-Packard 5973/6890 instrument by elec-
tron impact ionization at 70 eV and with helium carrier gas. All analyses
were performed by using an Agilent DB-5MS column (50 mꢂ0.25 mmꢂ
0.25 mm). The reactions in the presence of DMPO were analyzed with a
HP-INNOWAX column (30 mꢂ0.25 mmꢂ0.25 mm) using a new septum
and injector insert. Reaction samples were withdrawn from the reaction
mixture after 1, 2, 4, 6, 8, and 10 min. A small amount of product 3 was
found after 2 and 8 min of reaction. A maximum amount of the product
3 was observed after 4 and 6 min of reaction.
2
2
Kinetics of the decay of [(L5 )Fe
(OOH)]2+ was monitored at its absorbance maximum with a Varian
(OOH)]2+: The decay of [(L5 )Fe-
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
CARY 50 spectrophotometer equipped with a Hellma immersion probe.
The temperature was maintained with a Thermo Haake CT90L cryostat.
Rate constants were determined by least-squares fitting of the experi-
mental data to first-order kinetics. The errors in the rate constants are
those given by the software used for the fitting procedure (Kaleidagraph
4.1, Synergy Software).[33]
Chem. Eur. J. 2012, 18, 2715 – 2724
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2723

A. B. Sorokin, F. Banse et al.
[12] a) D. R. Boyd, J. W. Daly, D. M. Jerina, Biochemistry 1972, 11,
1961–1966; b) E. V. Kudrik, P. Afanasiev, D. Bouchu, J. M. M.
Millet, A. B. Sorokin, J. Porphyrins Phthalocyanines 2008, 12, 1078–
1089; c) E. V. Kudrik, A. B. Sorokin, Chem. Eur. J. 2008, 14, 7123–
7126.
L. Que Jr, Angew. Chem. 2003, 115, 3799–3801; Angew. Chem. Int.
Ed. 2003, 42, 3671–3673.
[23] a) E. Finkelstein, G. M. Rosen, E. J. Rauckman, J. Am. Chem. Soc.
1980, 102, 4994–4999; b) K. Makino, T. Hagiwara, A. Hagi, M.
Nishi, A. Murakami, Biochem. Biophys. Res. Commun. 1990, 172,
1073–1080; c) L. X. Li, Y. Abe, K. Kanagawa, N. Usui, K. Imai, T.
Mashino, M. Mochizuki, N. Miyata, Anal. Chim. Acta 2004, 512,
121–124.
[24] K. Makino, A. Hagi, H. Ide, A. Murakami, M. Nishi, Can. J. Chem.
1992, 70, 2818–2827.
[25] C. Cavazza, C. Bochot, P. Rousselot-Pailley, P. Carpentier, M. Cher-
rier, L. Martin, C. Marchi-Delapierre, J. C. Fontecilla-Camps, S.
Mꢀnage, Nat. Chem. 2010, 2, 1069–1076.
[13] A. Sorokin, A. Robert, B. Meunier, J. Am. Chem. Soc. 1993, 115,
7293–7299.
[14] a) C. M. Bathelt, L. Ridder, A. J. Mulholland, J. N. Harvey, J. Am.
Chem. Soc. 2003, 125, 15004–15005; b) S. P. de Visser, S. Shaik, J.
Am. Chem. Soc. 2003, 125, 7413–7424; c) K. R. Korzekwa, D. C.
Swinney, W. F. Trager, Biochemistry 1989, 28, 9019–9027; d) S. P. de
Visser, Chem. Eur. J. 2006, 12, 8168–8177.
[15] Y. Mekmouche, S. Mꢀnage, C. Toia-Duboc, M. Fontecave, J.-B.
Galey, C. Lebrun, J. Pꢀcaut, Angew. Chem. 2001, 113, 975–978;
Angew. Chem. Int. Ed. 2001, 40, 949–952.
[26] C. R. E. Jefcoate, J. R. Lindsay Smith, R. O. C. Norman, J. Chem.
Soc. B 1969, 1013–1018.
[16] M. J. Park, J. Lee, Y. Suh, J. Kim, W. Nam, J. Am. Chem. Soc. 2006,
128, 2630–2634.
[17] a) M. Newcomb, R. E. P. Chandrasena, Biochem. Biophys. Res.
Commun. 2005, 338, 394–403; b) T. Matsui, M. Unno, M. Ikeda-
Saito, Acc. Chem. Res. 2009, 42, 240–247.
[18] a) O. V. Makhlynets, P. Das, S. Taktak, M. Flook, R. Mas-Balleste,
E. V. Rybak-Akimova, L. Que, Chem. Eur. J. 2009, 15, 13171–
13180; b) O. V. Makhlynets, E. V. Rybak-Akimova, Chem. Eur. J.
2010, 16, 13995–14006.
[19] a) K. Chen, M. Costas, L. Que Jr, J. Chem. Soc. Dalton Trans. 2002,
672–679; b) K. Chen, L. Que Jr, J. Am. Chem. Soc. 2001, 123,
6327–6337; c) A. Company, L. Gomez, X. Fontrodona, X. Ribas, M.
Costas, Chem. Eur. J. 2008, 14, 5727–5731; d) A. Company, L.
Gomez, M. Guell, X. Ribas, J. M. Luis, L. Que, M. Costas, J. Am.
Chem. Soc. 2007, 129, 15766.
[27] a) E. J. Klinker, J. Kaizer, W. W. Brennessel, N. L. Woodrum, C. J.
Cramer, L. Que Jr, Angew. Chem. 2005, 117, 3756–3760; Angew.
Chem. Int. Ed. 2005, 44, 3690–3694; b) J. U. Rohde, J. H. In, M. H.
Lim, W. W. Brennessel, M. R. Bukowski, A. Stubna, E. Munck, W.
Nam, L. Que Jr, Science 2003, 299, 1037–1039; c) A. Thibon, J. Eng-
land, M. Martinho, V. G. Young, J. R. Frisch, R. Guillot, J. J. Girerd,
E. Munck, L. Que, F. Banse, Angew. Chem. 2008, 120, 7172–7175;
Angew. Chem. Int. Ed. 2008, 47, 7064–7067.
[28] M. Martinho, F. Banse, J.-F. Bartoli, T. A. Mattioli, P. Battioni, O.
Horner, S. Bourcier, J.-J. Girerd, Inorg. Chem. 2005, 44, 9592–9596.
[29] R. D. Bach, O. Dmitrenko, J. Org. Chem. 2010, 75, 3705–3714.
[30] P. Mialane, A. Novorojkine, G. Pratviel, L. Azꢀma, M. Slany, F.
Godde, A. Simaan, F. Banse, T. Kargar-Grisel, G. Bouchoux, J. Sain-
ton, O. Horner, J. Guilhem, L. Tchertanova, B. Meunier, J. J. Girerd,
Inorg. Chem. 1999, 38, 1085–1092.
[20] a) O. Y. Lyakin, K. P. Bryliakov, G. J. P. Britovsek, E. P. Talsi, J. Am.
Chem. Soc. 2009, 131, 10798–10799; b) O. Y. Lyakin, K. P. Brylia-
kov, E. P. Talsi, Inorg. Chem. 2011, 50, 5526–5538; c) I. Prat, J. S.
Mathieson, M. Guell, X. Ribas, J. M. Luis, L. Cronin, M. Costas,
Nat. Chem. 2011, 3, 788–793.
[31] N. Raffard, V. Balland, J. Simaan, S. Letard, M. Nierlich, K. Miki, F.
Banse, E. Anxolabehere Mallart, J. J. Girerd, C. R. Chim. 2002, 5,
99–109.
[32] A. G. Krainev, T. D. Williams, D. J. Bigelow, J. Magn. Reson. Ser. B
1996, 111, 272–280.
[21] N. Lehnert, F. Neese, R. Y. N. Ho, L. Que Jr, E. I. Solomon, J. Am.
Chem. Soc. 2002, 124, 10810–10822.
[33] Kaleidagraph 4.1.1, Synergy Software, 1986–2009.
[34] Xepr 2.3, Bruker Biospin, 2003.
[22] a) J. Bautz, M. R. Bukowski, M. Kerscher, A. Stubna, P. Comba, A.
Lienke, E. Munck, L. Que Jr, Angew. Chem. 2006, 118, 5810–5813;
Angew. Chem. Int. Ed. 2006, 45, 5681–5684; b) J. Kaizer, M. Costas,
Received: July 21, 2011
Published online: January 30, 2012
2724
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2715 – 2724