Catalytic defluorination of perfluorinated aromatics under oxidative conditions using N-bridged diiron phthalocyanine

Article abstract of DOI:10.1021/ja505437h

Carbon-fluorine bonds are the strongest single bonds in organic chemistry, making activation and cleavage usually associated with organometallic and reductive approaches particularly difficult. We describe here an efficient defluorination of poly- and perfluorinated aromatics under oxidative conditions catalyzed by the μ-nitrido diiron phthalocyanine complex [(Pc)Fe ^III(μ-N)Fe^IV(Pc)] under mild conditions (hydrogen peroxide as the oxidant, near-ambient temperatures). The reaction proceeds via the formation of a high-valent diiron phthalocyanine radical cation complex with fluoride axial ligands, [(Pc)(F)Fe^IV(μ-N)Fe^IV(F) (Pc^+?)], which was isolated and characterized by UV-vis, EPR, ¹⁹F NMR, Fe K-edge EXAFS, XANES, and Kβ X-ray emission spectroscopy, ESI-MS, and electrochemical techniques. A wide range of per- and polyfluorinated aromatics (21 examples), including C₆F₆, C₆F₅CF₃, C₆F₅CN, and C₆F₅NO₂, were defluorinated with high conversions and high turnover numbers. [(Pc)Fe^III(μ-N)Fe ^IV(Pc)] immobilized on a carbon support showed increased catalytic activity in heterogeneous defluorination in water, providing up to 4825 C-F cleavages per catalyst molecule. The μ-nitrido diiron structure is essential for the oxidative defluorination. Intramolecular competitive reactions using C₆F₃Cl₃ and C₆F₃H ₃ probes indicated preferential transformation of C-F bonds with respect to C-Cl and C-H bonds. On the basis of the available data, mechanistic issues of this unusual reactivity are discussed and a tentative mechanism of defluorination under oxidative conditions is proposed.

Full text of DOI:10.1021/ja505437h

Article
pubs.acs.org/JACS
Catalytic Defluorination of Perfluorinated Aromatics under Oxidative
Conditions Using N‑Bridged Diiron Phthalocyanine
Cedric Colomban, Evgenij V. Kudrik, Pavel Afanasiev,* and Alexander B. Sorokin*
́
́
Institut de Recherches sur la Catalyse et l′Environnement de Lyon, IRCELYON, UMR 5256, CNRS − Universite Lyon 1, 2 av. Albert
Einstein, 69626 Villeurbanne, France
S
* Supporting Information
ABSTRACT: Carbon−fluorine bonds are the strongest single bonds in organic
chemistry, making activation and cleavage usually associated with organometallic and
reductive approaches particularly difficult. We describe here an efficient
defluorination of poly- and perfluorinated aromatics under oxidative conditions
catalyzed by the μ-nitrido diiron phthalocyanine complex [(Pc)FeIII(μ-N)FeIV(Pc)]
under mild conditions (hydrogen peroxide as the oxidant, near-ambient temper-
atures). The reaction proceeds via the formation of a high-valent diiron
phthalocyanine radical cation complex with fluoride axial ligands, [(Pc)(F)FeIV(μ-
N)FeIV(F)(Pc+•)], which was isolated and characterized by UV−vis, EPR, 19F NMR,
Fe K-edge EXAFS, XANES, and Kβ X-ray emission spectroscopy, ESI-MS, and electrochemical techniques. A wide range of per-
and polyfluorinated aromatics (21 examples), including C6F6, C6F5CF3, C6F5CN, and C6F5NO2, were defluorinated with high
conversions and high turnover numbers. [(Pc)FeIII(μ-N)FeIV(Pc)] immobilized on a carbon support showed increased catalytic
activity in heterogeneous defluorination in water, providing up to 4825 C−F cleavages per catalyst molecule. The μ-nitrido diiron
structure is essential for the oxidative defluorination. Intramolecular competitive reactions using C6F3Cl3 and C6F3H3 probes
indicated preferential transformation of C−F bonds with respect to C−Cl and C−H bonds. On the basis of the available data,
mechanistic issues of this unusual reactivity are discussed and a tentative mechanism of defluorination under oxidative conditions
is proposed.
aliphatic C−F bonds has been achieved by using carboranes
in combination with alkylsilanes under an inert atmosphere in
anhydrous solvents.3,6d These processes involving the cleavage
of C−F bonds are conceived to proceed reductively. However,
the use of sophisticated reagents and special conditions (dry
organic solvents, inert atmosphere) compromises the practical
applications of the aforementioned approaches. The develop-
ment of accessible and cheap catalysts operating under practical
conditions (air, water) would be a significant achievement to
address emerging environmental problems due to increasing
use of organofluorines and their persistence in the environ-
ment. Many biochemical aerobic processes and current
depollution methods involve oxidation steps to convert
recalcitrant chlorinated xenobiotics.7 Since fluorine is the
most electronegative element and C−F bonds are deactivated
toward electrophilic attack,4 the transformation of fluorinated
compounds under oxidative conditions is extremely difficult
and represents a fundamental challenge.
1. INTRODUCTION
Current approaches to C−F bond reactivity in organic
chemistry are based on activation at electron-rich transition-
metal complexes, reduction, and nucleophilic substitution.1−4
Up to the present, fluorine chemistry has been focused on the
development of synthetic methods for the preparation of
various fluorous compounds, which are finding wide and
increasing uses in many applications because of the unique
properties of C−F bonds (C−F bond strength, high thermal
and oxidative stability, low polarity, weak intermolecular
interactions, etc.). In particular, the annual global production
of the most important fluoroaromatics was estimated to be 35
000 tons in 2000, but the trend in the past (10 000 tons in
1994) indicates a further increase in their production and use.5
Fluorinated compounds are particularly persistent in the
environment since their biodegradation is very slow. Poly-
and perfluorinated molecules are especially difficult to degrade.
Consequently, along with the improvement in synthetic
fluorine chemistry, the development of disposal methods for
these compounds is of increasing importance. Emerging
stoichiometric and rare catalytic approaches to C−F bond
transformation are based on reduction.3,6 Catalytic hydro-
defluorination of aromatic fluorocarbons with alkylsilanes
catalyzed by ruthenium N-heterocyclic carbene complexes
under strictly anaerobic and dry conditions occurred with
turnover numbers (TONs) of 1.3−7.4 for 19.45 h in the case of
C6F6.6c Significant progress in the hydrodefluorination of
We have previously developed an iron phthalocyanine
(FePc)−H2O2 system for the oxidative degradation of
recalcitrant chlorinated phenols.8 FePc complexes are structur-
ally related to iron porphyrins and combine availability and high
reactivity in many reactions.9 Further development of FePc
catalysts resulted in the discovery of the remarkable catalytic
Received: February 12, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja505437h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
properties of μ-nitrido diiron phthalocyanines.10 The N-bridged
diiron tetra-tert-butylphthalocyanine complex [(Pc)FeIII(μ-N)-
FeIV(Pc)] (1) catalyzes the oxidation of methane by H2O2
under mild conditions in water.10a,b Using the tetraphenylpor-
phyrin (TPP) platform, we have prepared an N-bridged high-
valent diiron oxo species, [(TPP)FeIV(μ-N)FeIVO(TPP+•)],
that can also oxidize methane.11 [(TPP)FeIV(μ-N)FeIV
O(TPP+•)] is a much stronger oxidant than its mononuclear
analogue [(TPP+•)FeIVO], and therefore, the Fe(μ-N)Fe
structural feature is essential for the high catalytic activity.11
Another example of an unusual reactivity of 1 is the formation
of the high-valent [(Pc)(Cl)FeIV(μ-N)FeIV(Cl)(Pc+•)] complex
CH2Cl2 and CH2Br2, respectively.10e The UV−vis spectrum of
2 is characteristic of phthalocyanine radical cations.12 The
formation of 2 from 1 led to the disappearance of the low-spin
iron electron paramagnetic resonance (EPR) signal at g = 2.091
and the appearance of an intense symmetric and narrow signal
(8 G width at 120 K) at g = 2.0038, typical for an organic
radical (Figure 1 inset). UV−vis and EPR data suggested an
Fe(IV)Fe(IV) configuration and the presence of a radical
cation on one phthalocyanine ligand. Complex 2 was isolated in
the solid state and further characterized. The 19F NMR
spectrum of 2 showed one well-defined signal at −132 ppm
compatible with the presence of inorganic F− (Figure S1 in the
Supporting Information). Energy-dispersive X-ray fluorescence
analysis of 2 gave a 1:1 F:Fe atomic ratio, indicating the
presence of two fluoride ions per dimer molecule. The
electrospray ionization mass spectrometry (ESI-MS) peak of
2 was centered at m/z 1618.5 (Figure S2). The isotopic pattern
of the cluster corresponded to (1 + F)+ formed after the loss of
F− under positive ESI-MS conditions.
2.2. X-ray Absorption and Emission Spectroscopies.
Core hole spectroscopies using synchrotron radiation, including
extended X-ray absorption fine structure (EXAFS), X-ray
absorption near-edge structure (XANES), and X-ray emission
spectroscopy (XES), are widely used to determine the
structures and electronic properties of enzymes and their
model complexes.13 These techniques are particularly useful for
characterization of high-valent iron species.14,15 Comparison of
the Fe K-edge EXAFS spectra of the initial complex 1 and
fluorine-containing species 2 shows the retention of the μ-
nitrido dimer topology in 2. However, the EXAFS spectrum of
2 was noticeably modified compared with that of the initial
complex 1 (Figure 2). The changes in the EXAFS spectrum
10e
t
in the presence of BuOOH via dechlorination of CH2Cl2.
Inspired by the ability of 1 to cleave C−Cl bonds in CH2Cl2,
which is usually considered as a stable solvent for performing
oxidation reactions, we checked the reactivity of other
halogenated compounds in this system. Remarkably, the
dehalogenation process can be extended to aromatics bearing
fluorine substituents. Herein we report a highly efficient
defluorination of per- and polyfluorinated aromatic compounds
in the presence of N-bridged diiron phthalocyanine complex 1
and H2O2 that can be performed either in organic solvents or in
water. A high-valent diiron intermediate involved in the
catalytic cycle has been isolated and spectroscopically
characterized. Mechanistic features of this unusual reactivity
are discussed on the basis of 19F NMR and GC−MS
characterization of intermediate and final products obtained
in the oxidation of several probe molecules.
2. RESULTS AND DISCUSSION
2.1. Preparation of the High-Valent Diiron Difluoro
Complex in a Stoichiometric Reaction and Its Character-
ization. Incubation of 1 in an equimolar C6H6/C6F6 mixture in
the presence of tBuOOH (30 equiv) at 60 °C for 6 h resulted in
the formation of a new complex 2, as indicated by the change in
the UV−vis spectrum. The intensity of the Q band at 640 nm
decreased, and new broad bands at 549, 615, and 666 nm
appeared (Figure 1).
These changes are very similar to those observed during
formation of [(Pc)(Cl)FeIV(μ-N)FeIV(Cl)(Pc+•)] and [(Pc)-
(Br)FeIV(μ-N)FeIV(Br)(Pc+•)] via oxidative dehalogenation of
Figure 2. Fourier transforms of the EXAFS spectra of the initial
complex 1 and the fluoride adduct 2.
suggest symmetrization of the coordination environment (due
to the shift of Fe atom into the plane of the macrocycle) as well
as an increase in the coordination number of the iron atom.
A high-quality fit of the EXAFS spectrum of 2 for the two
first coordination spheres (Figure S3 and Table S1) gave the
structure shown in Figure 3, in which a fluorine ligand is bound
to each iron atom at a distance of 1.99 Å, whereas other bonds
distances are slightly modified relative to the initial structure of
1, as described earlier.10e
In agreement with the energy-dispersive X-ray fluorescence
analysis, the EXAFS spectrum of 2 indicated the presence of
one F− ion per iron atom. The Fe K-edge XANES spectra attest
to a change in the oxidation state from an Fe(III)Fe(IV) core
to an Fe(IV)Fe(IV) core, as indicated by a characteristic
Figure 1. UV−vis spectrum of 2 obtained by the treatment of C6F6
t
with 1 and BuOOH (MeCN, 25 °C). The inset shows the EPR
spectrum of 2 in CH2Cl2 at 120 K (microwave frequency 9.392 GHz,
power 1.6 mW, modulation 1.0 mT/100 kHz).
B
dx.doi.org/10.1021/ja505437h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
[(Pc)FeIII(μ‐N)FeIV(Pc)]
⇄ [(Pc)FeIV(μ‐N)FeIV(Pc)] + e−
[(Pc)FeIV(μ‐N)FeIV(Pc)]
⇄ [(Pc)FeIV(μ‐N)FeIV(Pc+•)] + e−
Electrooxidation of 2 did not show any electron-transfer
process up to 1.3 V, which is consistent with the presence of
the fully oxidized species [(Pc)(F)FeIV(μ-N)FeIV(F)(Pc+•)].
The CV of 2 consists of a single reversible one-electron
reduction wave at E1/2 = −385 mV vs Fc/Fc+ (Figure S5)
involving reduction by one electron per dimer unit:
[(Pc)(F)FeIV(μ‐N)FeIV(F)(Pc+•)] + e−
⇄ [(Pc)(F)FeIV(μ‐N)FeIV(F)(Pc)]
Figure 3. EXAFS structure of [(Pc)(F)FeIV(μ-N)FeIV(F)(Pc+•)] (2)
showing the first and second coordination spheres.
Thus, the reduction of 2 containing coordinated fluoride anions
is much more difficult (by 785 mV) than that of [(Pc)FeIV(μ-
N)FeIV(Pc+•)] containing merely perchlorate anions from the
supporting tetra-n-butylammonium perchlorate electrolyte.
This reduction potential shift is consistent with the counterion
effect described for FeTPP complexes, where the reduction of
(TPP)FeX undergoes a cathodic shift of up to 710 mV as the
counterion is varied from weakly coordinating ClO4− to tightly
bound F−.18 A spectroelectrochemical study showed the
similarity of the UV−vis spectra obtained after bulk electro-
oxidation of 1 and reduction of 2 (Figure S6). Thus, all of the
experimental data confirm the formulation of 2 as [(Pc)(F)-
FeIV(μ-N)FeIV(F)Pc+•)].
splitting of the pre-edge peak12c and a significant increase of the
pre-edge maximum energy from 7114.7 eV in 1 to 7116.2 eV in
2 (Figure 4).
2.4. Homogeneous Catalytic Oxidative Defluorina-
tion. In the presence of tBuOOH, 1 is capable of defluorinating
C6F6 to form a binuclear [(Pc)(F)FeIV(μ-N)FeIV(F)(Pc+•)]
complex. This complex is stable under the reaction conditions,
and no further defluorination occurs. Clearly, 1 shows a very
t
intriguing defluorination ability, though with BuOOH the
Figure 4. Fe K-edge XANES spectra of the initial [(Pc)FeIII(μ-
N)FeIV(Pc)] complex 1 (blue) and the [(Pc)(F)FeIV(μ-N)FeIV(F)-
(Pc+•)] complex 2 (red).
process is not catalytic but instead is a stoichiometric reaction.
However, if 2 could be reduced back to 1, a fascinating
possibility would arise for the mild catalytic defluorination. Our
attempts to find a catalytic pathway led to success. Upon
interaction of 2 with H2O2, we observed the transformation of
2 to the initial 1 via a yet-unknown intermediate 3 showing a
peak at λ = 617 nm in the UV−vis spectrum. Assuming that the
defluorinations in the presence of H2O2 and tBuOOH proceed
similarly, we propose the catalytic cycle shown in Figure 5.
To demonstrate the possibility of a catalytic cycle, we treated
0.1 M pentafluorophenol solution with H2O2 (16 equiv) in the
presence of 0.4 mol % catalyst 1 in MeCN at 60 °C in a Teflon
reactor. 19F NMR analysis of the reaction mixture after 15 h
showed 82% C6F5OH conversion and the formation of 4.0 F−
(mostly in the form of HF) per one converted C6F5OH
molecule. A high TON (i.e., moles of F− formed per mole of
catalyst) of 818 was achieved. When the reaction was
performed in a glass flask, a significant amount of
hexafluorosilicate was isolated (as shown by 19F NMR and IR
data), thus confirming the formation of HF during the reaction.
It is noteworthy that the 1−H2O2 system efficiently performs
defluorination of C6F5OH even at 20 °C, with 88% conversion
after 24 h.
Iron valence-to-core Fe Kβ XES can be used to determine
light atoms within complex multimetallic strucutures.16 For
instance, this technique allowed the identification of the unique
interstitial carbide ion in the iron molybdenum cofactor
(FeMoco) of nitrogenase.17 According to the Kβ XES data, 1
and 2 are both low-spin complexes. In agreement with the
presence of a long Fe−F distance as determined from the
EXAFS spectrum, the valence-to-core XES spectra show
negligible intensity of the characteristic crossover satellite
peak in the range 7075−7085 eV (Figure S4), which should
clearly appear for a short Fe−F bond.
2.3. Electrochemistry. The electrochemical behaviors of 1
and 2 were investigated by cyclic voltammetry (CV) and
rotating disk electrode (RDE) voltammetry in dichloro-
methane. The cyclic voltammogram for electrooxidation of 1
(Figure S5) displays two successive reversible one-electron
oxidation waves at E1/2 = 10 mV and E1/2 = 400 mV relative to
ferrocene/ferrocenium (Fc/Fc+). According to RDE voltam-
metry experiments, both of these oxidations involve the
reversible abstraction of one electron per dimer unit. These
two reversible one-electron oxidations can be assigned as
follows:
The reaction was further extended to a large scope of
substrates (Figure 6). Even hexafluorobenzene, pentafluoropyr-
idine, and octafluoronaphthalene as well as polyfluorinated
aromatic compounds containing functional groups were
C
dx.doi.org/10.1021/ja505437h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 5. Structure of N-bridged diiron tetra-tert-butylphthalocyanine
complex 1 and proposed catalytic cycle for oxidative defluorination.
The ovals stand for the tetra-tert-butylphthalocyanine ligand.
transformed with high TONs. High degrees of defluorination in
terms of the number of F− ions formed per converted substrate
molecule were achieved. From six fluorine atoms of C6F6, 3.12
were transformed to inorganic F−. The presence of strong
electron-withdrawing substituents (CN, NO2) was tolerated.
No products of transformation of the CF3 group were found in
the course of oxidation of octafluorotoluene. The oxidation of
octafluoronaphthalene afforded heptafluoronaphthols, hexa-
fluoro-1,4-naphthoquinone, tetrafluorophthalic acid, difluoro-
maleic acid, and difluorofumaric acid.
2.5. Heterogeneous Catalytic Defluorination in Water.
It should be noted that defluorination reactivity of 1 was
observed in acetonitrile, which can also be oxidized under the
reaction conditions.10a The concurrent solvent oxidation can
partially hide the real performance of 1 in the defluorination of
recalcitrant fluoroaromatics. The best way to avoid side
reactions involving organic solvents is to use water, a green
and stable solvent. To this aim, complex 1 was supported on
carbon (1-C, 300 m2/g BET surface area, 10 μmol/g), and
heterogeneous defluorination reactions were carried out in
water. The catalytic system was not deactivated in aqueous
solution, and high performance of the supported catalyst 1-C
was observed in the transformation of fluorinated compounds
(Figure 7).
In water, 4825 C−F bonds of C6F5OH per catalyst molecule
were mineralized with the release of F−. C5F5N and C6F6,
which are only sparingly soluble in water, were transformed
with 85% and 75% conversion, respectively, indicating a high
catalytic performance even in dilute solutions. This tolerance to
water is of exceptional importance since other approaches to
defluorination often need anhydrous conditions.6
Figure 6. Scope of the catalytic homogeneous dehalogenation of
polyfluorinated aromatics by the 1−H2O2 system. The solution
containing the catalyst (0.4 mM), substrate (0.1 M), and H2O2 (1.6
M) was stirred in CD3CN at 60 °C for 15 h. The catalyst:substrate:ox-
idant ratio was 1:250:4000. The substrate conversions (C) and yields
of F− were determined by 19F NMR spectroscopy. Turnover numbers
(TONs) were calculated as the molar amount of F− formed per mole
of the catalyst. The degree of defluorination (F−) is the ratio of the
amount of F− formed to the total amount of fluorine in the converted
substrate.
The efficiency of mineralization depends on the
H2O2:substrate ratio. Quasi-complete conversion of C6F5OH
with 70−80% mineralization was achieved using 10−26 equiv
of H2O2. The kinetics of the oxidative degradation of C6F5OH
at 60 °C in the presence of 1-C (0.1 mol %) and H2O2 (26
equiv) was monitored by 19F NMR spectroscopy (Figure 8).
The signals of C6F5OH at −163.3, −165.6, and −170.7 ppm
progressively disappeared, showing conversions of 37%, 51%,
and 96% after 2.5, 5, and 15 h, respectively. The main reaction
products were DF, oxalic acid, and difluoromaleic acid as well as
a small amount of fluorooxaloacetic acid [as confirmed by GC−
MS after methylation with (CH3)3SiCHN2]. The yields of the
final products of the C6F5OH transformation are shown in
Figure 9.
The total organic carbon analyses showed a 53.6% loss of the
organic carbon due to the formation of CO2 and CO, which
were identified in the gas phase in a 15:1 ratio. Importantly,
89% of the organic fluorine was transformed to inorganic
fluoride anions. These results indicate a very efficient
mineralization of the water-soluble fluorinated aromatics
using the heterogeneous catalyst in combination with clean
H2O2 oxidant in water without any additives.
2.6. Preparative Oxidative Defluorination. Smaller
excesses of H2O2 led to more selective reactions. With 5
equiv of H2O2, a 62% yield of difluoromaleic acid and a 66%
yield of tetrafluorophthalic acid were obtained during the
transformations of C6F5OH and octafluoronaphthalene,
respectively (Figure 10). Thus, the 1−H2O2 catalytic system
D
dx.doi.org/10.1021/ja505437h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 10. Preparative oxidations of pentafluorophenol and
octafluoronaphthalene using the 1−H2O2 system.
showed a low catalytic activity only toward C6F5OH. A 23%
conversion of C6F5OH, a TON of 160, and 56% mineralization
were obtained with FePctBu4(Cl), compared with 82%
conversion, a TON of 818, and 80% mineralization observed
in the presence of 1. Therefore, the dimeric strucure of 1 is
essential for the high defluorination activity. μ-Oxo and μ-
carbido diiron phthalocyanine complexes were not active
toward C6F6, indicating the essential role of the Fe(μ-N)Fe
structural unit. Previously, a pronounced axial ligand effect on
oxidation activity was observed for iron porphyrin complexes.19
The influence of the axial ligand was explained by its electron
donor effect and by the stabilization of the final catalyst state,
thus modulating the reaction free energy for epoxidation,
hydrogen atom abstraction, or demethylation. In single-atom-
bridged diiron complexes (μ-oxo, μ-nitrido, μ-carbido), the
nature of the bridging ligand seems to be even more important.
μ-Nitrido diiron species exhibit a reactivity that other
complexes do not show. Further experimental and calculation
studies are necessary to understand the reason(s) for the
unprecendented reactivity of μ-nitrido diiron complexes.
Early reports showed that hydroxyl radicals generated by
thermal decomposition of H2O2 can react with octafluoronaph-
thalene and heptafluoronaphthols to form dimeric coupling
products and hexafluoronaphthoquinone.20 Photo-Fenton
defluorination of C6F5OH and C6F5COOH using ferrous
sulfate or ferric oxalate in combination with H2O2 and UV-C
light (200−300 nm) has been reported.21 However, it has been
shown that C6F6 does not react with OH• radicals generated by
thermolysis of 90% H2O2 in acetonitrile even at 100 °C.22 To
verify the possible involvement of Fenton chemistry, we tested
the reactivities of 15 fluorinated substrates in the presence of
FeSO4 and H2O2 under the conditions used for the 1−H2O2
system. The only substrate that exhibited similar reactivities
with 1 and FeSO4 was fluoranil. Importantly, all of the
fluorinated benzenes from C6H4F2 to C6F6 remained
completely unchanged under Fenton conditions. The substrates
containing electron-donating (12, 20, and 21) or electron-
withdrawing substituents (14−19) showed conversions at the
detection limit. When the FeSO4 loading was increased 25
times (to 10 mol %), C6F6 (0.1 M in CD3CN) did not react
with H2O2 (1.6 M) at 60 °C for 15 h. These results indicate
that hydroxyl radicals should not be responsible for the
oxidative defluorination of fluorinated aromatic compounds
catalyzed by 1.
Figure 7. Catalytic heterogeneous dehalogenation of polyfluorinated
aromatics by the 1-C−H2O2 system. Reactions were run by stirring the
supported catalyst (20 mg, 0.2 μmol of complex), substrate [(a) 0.1
M; (b) 0.005 M for C5F5N and 0.003 M for C6F6], and H2O2 [(a) 2.6
M; (b) 0.1 M] in D2O at 60 °C for 15 h. The catalyst:substrate:oxidant
ratio was 1:1000:26000 in (a) and 1:50:1000 for C5F5N and 1:96:3200
for C6F6 in (b).
Figure 8. Kinetics of C6F5OH degradation and formation of oxidation
products. The time dependence of the 19F NMR spectra shows the
oxidative degradation of 0.1 M C6F5OH at 60 °C in D2O performed
under the conditions described in Figure 7.
Figure 9. Yields of the final products of the C6F5OH transformation.
The fluorine and carbon amounts found in the products represented
97% and 89%, respectively, based on the substrate content.
can be also adapted for the preparation of functionalized
fluorinated molecules from available poly- and perfluoroar-
omatic compounds.
2.7. Study of the Reactivity of Fluoroaromatics Using
Other Systems. The reactivities of monomeric FePctBu4(Cl)
toward C6F5OH and C6F6 were checked under the same
conditions (CD3CN, 60 °C, 15 h). Importantly, no conversion
of C6F6 was detected with FePctBu4(Cl). The FePctBu4(Cl)
To probe the stabilities of the substrates and identified
intermediates under the reaction conditions, control experi-
ments were performed without the catalyst under the same
reaction conditions in CD3CN and D2O. With all of the studied
substrates, including C6F5OH (an intermediate in the
degradation of C6F6), no reaction was observed. The only
exception was found for tetrafluoro-1,4-benzoquinone, which
was capable of reacting with H2O2 to form trifluorohydroxy-p-
E
dx.doi.org/10.1021/ja505437h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
benzoquinone.23 The treatment of fluoranil (0.1 M in CD3CN)
with H2O2 (1.6 M) for 15 h at 60 °C led to 89% substrate
conversion. However, the reaction without catalyst was slower
than catalytic reaction (29% vs 56%, respectively, after 30 min).
The noncatalyzed reaction was less efficient than the catalytic
process in terms of degree of defluorination (2.00 F− vs 4.00 F−
formed per converted substrate, respectively). All of these
results indicate the essential role of the μ-nitrido diiron
complex in defluorination of fluorinated aromatics under
oxidative conditions.
2.8. Comparison of the Reactivities of the 1−H2O2
System toward C−F, C−Cl, and C−H Bonds. The relative
reactivity of C−F bonds versus C−Cl and C−H bonds was
evaluated using an intramolecular competitive approach. The
oxidation of several substrates employed in a large excess was
monitored at short reaction times in order to determine the
initial reaction products (Figure 11). The dehalogenation of
defluorination of fluoroaromatics by 1 and provide preliminary
considerations regarding the possible mechanism. Available
EXAFS, XANES, ESI-MS, Mossbauer, EPR, labeling, and
̈
reactivity data indicate the involvement of hydroperoxo and
high-valent diiron oxo species in the 1−H2O2 system,11,12c
showing mechanistic features similar to those of biological
oxidation [e.g., intermediate formation of benzene epoxide and
the occurrence of a 1,2-shift (NIH shift) during oxidation of
benzene].10c However, no biodegradation of perfluorinated
compounds has ever been reported,24 and experimental studies
on the oxidation of highly halogenated aromatics containing
fluorine substituents are rare.25 On the basis of the available
data, we consider below several mechanistic hypotheses.
2.9.2. Nucleophilic Substitution Hypothesis. SNAr nucleo-
philic attack on C−F bonds mediated by transition-metal
complexes (Pd, Rh, Ni, etc.) or oxidative addition of C−F
bonds may result in the formation of C−X bonds (X = H, N,
O, S).26 However, all reported cases of such C−F bond
substitution occur under basic conditions and/or on electron-
rich transition-metal complexes, which is not the case for 1. In
general, the higher reactivity of the C−F bond compared with
the C−Cl bond in the transformation of C6F3Cl3 is compatible
with a nucleophilic mechanism since the reactivity of the aryl
halides decreases in the order of F > Cl > Br > I.1 However, the
conversions of fluorinated benzenes are higher for less-
fluorinated compounds: C6F6 (29%), C6F5H (33%), 1,2,4,5-
C6F4H2 (59%), 1,3,5-C6F3H3 (66%), 1,4-C6F2H4 (69%).
Similarly, the conversions of substrates with one electron-
withdrawing group are higher than those of substrates bearing
two such groups: C6F5(COOH) (52%) versus C6F5(COOH)2
(20%) and C6F5(CN) (20%) versus C6F5(CN)2 (13%). A
competition reaction between C6F5CF3 and C6F5CH3 resulted
in conversions of 31% and 49%, respectively, while C6F5CF3
and C6F5CH3 show similar reactivities in SNAr reactions.27
Although on the basis of these reactivity trends a nucleophilic
substitution mechanism cannot be definitively excluded, the
observed results seem to be better compatible with the
participation of electrophilic species.
Figure 11. Relative reactivities of C−F bonds vs C−Cl and C−H
bonds determined in intramolecular competition reactions. The ratios
of phenols were determined after 1 h. Only the major product P5 is
shown for the transformation of pentafluorobenzene.
2.9.3. Electrophilic Attack Hypothesis. High-valent iron oxo
species are potent oxidants and possess strong electrophilic
properties. However, the observed preference for cleavage of
C−F bonds versus C−H and C−Cl bonds is not compatible
with an initial direct electrophilic attack of a high-valent diiron
oxo species on the C−X bond (X = F, Cl, H) because in this
case the more electron-rich C−H or C−Cl bonds should be
preferred. On the other hand, in vitro and in vivo cytochrome
P-450-dependent biotransformations of C6F5Cl and C6F3Cl3
exhibited the preferential elimination of the p-F substituent and
formation of P1, respectively.25 Calculations predicted a higher
reactivity of the p-F position of C6F5Cl and a fluorinated
position of C6F3Cl3. A density functional theory (DFT) study
showed that activation of C6F6 proceeded by π attack of
Compound I on the aromatic ring to form hexafluorocyclohex-
adienone with a lower activation barrier compared with that for
C6Cl6.7 The intermediate formation of cyclohexadienone
species from arene epoxides has been postulated in numerous
studies on hydroxylation of aromatic compounds.28
1,3,5-trichloro-2,4,6-trifluorobenzene (C6F3Cl3) resulted in the
formation of phenol P1 (F elimination) and phenol P2 (Cl
elimination) in a ratio of ∼97:3. The transformation of 1,3,5-
trifluorobenzene yielded phenol P3 (F elimination) and phenol
P4 (C−H oxidation) in a ratio of ∼98:2. Remarkably, only
tetrafluorophenols (F elimination) were obtained in the
reaction with pentafluorobenzene. No traces of pentafluor-
ophenol resulting from C−H oxidation were detected.
Although these phenols are the intermediate products and
these ratios might slightly deviate from the ratios of the intrinsic
reactivities of C−F versus C−Cl and C−H bonds, the results
obtained clearly show an unusually strong preference of the
catalytic system for the transformation of C−F bonds with
respect to C−Cl and C−H bonds. Such unusual selectivity
gives an indirect indication that the initial reaction step involves
a transformation of a C−F bond.
2.9. Mechanistic Considerations. 2.9.1. General Re-
marks. The counterintuitive defluorination of aromatic C−F
bonds under oxidative conditions and preferential defluorina-
tion versus dechlorination and oxidation of C−H bonds
demonstrated by 1 raise questions about the possible
mechanistic background of this striking reactivity. In this
section, we discuss several possible scenarios of oxidative
2.9.4. Initial Epoxidation Step. Aromatic oxidation
performed by cytochrome P-450 involving high-valent iron
oxo species leads to intermediate formation of arene epoxides
accompanied by NIH shifts.28 Epoxide intermediates were also
proposed to be the primary metabolites in cytochrome P450-
catalyzed oxidation of halobenzene derivatives.29 In turn, the
F
dx.doi.org/10.1021/ja505437h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
oxidation of benzene by the 1−H2O2 system was shown to
occur via the initial formation of benzene epoxide.10c
Consideration of the shapes and energies of the lowest
unoccupied molecular orbital (LUMO) and highest occupied
MO (HOMO) of C6F5H suggests that initial epoxidation
should occur on the CFCF bond rather than on the CH
CF bond, in agreement with the observed selectivity. Moreover,
the observed higher reactivity of C6F5H versus C6F6 also agrees
with the relative energy of the HOMOs (see Figure S7 and the
attached comment). Thus, the hypothesis of the initial
formation of the fluoroarene epoxide seems to be plausible.
Up to now, our attempts to detect the putative hexafluor-
obenzene epoxide have been unsuccessful, probably because of
its low stability under the high temperatures of GC−MS
analysis. The low response of fluorinated compounds under
ESI-MS conditions also prevents the application of this
technique for detection of epoxides of fluoroarenes. Never-
theless, when octafluoronaphthalene was used as the substrate,
GC−MS analysis of the reaction mixture showed the presence
of a minor product with a signal at m/z 288 corresponding to
C10F8O formulation, which can only be attributed to epoxide.
The fragmentation pattern of this product is also compatible
with an octafluoronaphthalene epoxide structure (Figures S8−
S13).
The initial epoxide formation can be also proposed upon the
detection of the NIH shift (migration of the substituent to an
adjacent position).10c,28,30 The NIH shift was evidenced for
chlorinated and brominated aromatics,31 but the migration of
fluorine is far less documented.32 The fluorine substituent is
preferentially eliminated from an aromatic molecule rather than
giving rise to an NIH shift.29
Several fluorobenzenes have been tested in the search of the
fluorine NIH shift. The occurrence of this phenomenon in P-
450-catalyzed oxidation was evidenced using 1,4-difluoroben-
zene.30 We applied the same approach to probe the fluorine
NIH shift in our system. Along with p-fluorophenol as a major
product, the oxidation of 1,4-difluorobenzene using the 1−
H2O2 system did lead to the formation of the NIH-shifted 2,4-
difluorophenol along with 2,5-difluorophenol (the normal
oxidation product) in a ratio of ∼1:15 (Figure 12). The
also be competent to attack fluoroaromatics. Alternatively, the
previously identified hydroperoxo complex [(Pc)FeIV(μ-N)-
FeIII(Pc)(OOH)] formed from 1 and H2O2 in the initial step of
H2O2 activation10e could also be responsible for the
epoxidation of fluorinated aromatics. The epoxidation of
electron-deficient olefins by nucleophilic monoferric porphyrin
peroxo complexes has been reported,33 providing support for
this scenario.
A reaction pathway for the oxidative defluorination of C6F6 is
proposed in Figure 13. The epoxide of the fluoroarene
undergoes a transformation to phenol via the formation of
the ketone intermediate. This mechanism giving rise to the
NIH shift is well-established for the oxidation of aromatic
compounds by high-valent iron oxo species.28 The fluorine
NIH shift detected in the oxidation of p-C6H4F2 strongly
suggests a similar mechanism for the transformation of
fluoroarenes. The further oxidation of C6F5OH results in the
formation of fluoranil, which was identified by GC−MS and 19
F
NMR analyses. The well-documented oxidation of halogenated
phenols to quinones catalyzed by cytochrome P-4507,25,28 and
phthalocyanine complexes8 supports this mechanism. The high
degrees of defluorination observed in the transformations of
fluoroarenes imply cleavage of the aromatic cycle with further
formation of F−. We assume that fluoranil can be further
epoxidized, with subsequent hydrolysis of the epoxide and
elimination of HF according to the mechanism for degradation
of chlorinated quinones proposed in ref 8b. The fluorinated
quinone can also undergo a nucleophilic attack by the peroxo
complex at the electron-deficient carbonyl position, resulting in
cleavage of the aromatic cycle and formation of oxalic and
maleic acids similar to the mechanism proposed for the
degradation of chlorinated phenols.8 The degradation of
fluoranil in the presence of nucleophilic H2O2 is in agreement
with this mechanism. Further epoxidation and C−C bond
cleavage steps lead to the formation of HF and CO2.
Difluoromaleic, fluorooxaloacetic, and oxalic acids as well as
CO2 and CO have been identified by GC−MS and 19F NMR
techniques (Figures S14−S17).
The proposed reaction pathway where numerous inter-
mediate and final products have been identified (shown in red
in Figure 13) is compatible with all of the available
experimental data. However, further experimental and theoreti-
cal studies are still necessary to gain deeper insights into the
mechanism of defluorination of aromatic compounds under
oxidative conditions.
3. CONCLUSION
Although there are a variety of efficient systems for C−F
activation based on electron-rich organometallic complexes,
reduction, and nucleophilic substitution, the oxidative trans-
formations of poly- and perfluorinated compounds are
practically unknown in chemistry and biology. In this context,
the discovery and development of a novel route for the
activation of C−F bonds would be of the utmost importance.
We have found that a μ-nitrido diiron phthalocyanine complex
Figure 12. Initial products of the 1,4-difluorobenzene transformation
in the presence of 1 and H2O2. Conditions: MeCN, 60 °C. The ratios
of phenols were determined after 25 min.
detection of the fluorine NIH shift is in agreement with the
involvement of the epoxidation as an initial step of the oxidative
defluorination of poly- and perfluorinated aromatics.
t
2.9.5. Tentative Mechanism of Defluorination under
Oxidative Conditions. It is thus conceivable that the first
step of defluorination could involve epoxidation of a fluorinated
aromatic molecule by [FeIV(μ-N)FeIVO phthalocyanine
radical cation]0 species. RI-PBE and B3LYP DFT calculations
showed the feasibility of this mechanism for cytochrome P450
Compound I.7 In our case, [FeIV(μ-N)FeIVO phthalocyanine
radical cation]0 species capable of oxidizing methane10a might
reacts with C6F6 in the presence of peroxides. Using BuOOH
we have prepared the high-valent diiron complex [(Pc)(F)-
FeIV(μ-N)FeIV(F)(Pc+•)] with two fluoride ligands originating
from the cleavage of aromatic C−F bonds. This complex was
characterized by UV−vis, EPR, 19F NMR, Fe K-edge EXAFS,
XANES, and Kβ X-ray emission spectroscopy, ESI-MS, and
electrochemical techniques. This stoichiometric chemistry was
further developed into a catalytic version. The first catalytic
G
dx.doi.org/10.1021/ja505437h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 13. Proposed reaction pathway for oxidative defluorination of hexafluorobenzene in the presence of 1. Identified intermediates and final
products are shown in red.
system for the efficient defluorination of poly- and perfluori-
nated aromatics under oxidative conditions is based on the μ-
nitrido diiron phthalocyanine complex and H2O2. A variety of
poly- and perfluorinated aromatic compounds (21 examples)
were defluorinated with high turnover numbers, showing the
large substrate scope of this approach. The mechanistic basis of
this unprecedented reactivity is of great interest since the
oxidative transformation of heavily fluorinated compounds has
been considered as almost impossible. A tentative mechanism
has been proposed on the basis of available data, but further
experimental and theoretical studies are necessary to provide
deeper insights. We believe that this novel approach to the
transformation of C−F bonds has fundamental character with a
great potential for further development. In addition, this
unexpected reactivity might be used in practical applications,
such as the disposal of fluorinated compounds. In fact, the
large-scale production and application of fluorinated xeno-
biotics (40% of agrochemicals and 20% of pharmaceuticals
currently used contain C−F bonds2) coupled with their
exceptional stability lead to their accumulation in the
environment with the risks of contamination of the food
chain and ground and drinking water.34 Poly- and perhalo-
genated aromatics are among the most recalcitrant pollutants.
The development of practical disposal methods for these
contaminants of emerging concern is therefore of great
importance.5 Reductive hydrodefluorination of organofluorines
as proposed in the literature is usually performed under
anaerobic and dry conditions.6 The only exception is the
preparative reductive hydrodefluorination using Zn in con-
centrated ammonia solutions.35 However, any practical
applications of these reductive methods for remediation can
hardly be envisaged. In contrast, the N-bridged diiron
phthalocyanine complex in combination with H2O2 reacts
with aromatic C−F bonds in water under near-ambient
conditions. In this context, the proposed oxidative approach
seems to be particularly promising. Indeed, (i) the iron catalyst
is nontoxic and cheap and can be available on a large scale; (ii)
H2O2 is a green and cheap industrial oxidant; (iii) the catalytic
system shows a large substrate scope and is tolerant to water
and air; (iv) the process can be performed in concentrated
solutions (typically the case of industrial wastes) or in dilute
aqueous solutions (treatment of contaminated water). For all of
these reasons, this novel fundamental approach to the
activation of C−F bonds can be developed for the degradation
of fluorinated pollutants that are resistant to biodegradation
and traditional remediation methods.
4. EXPERIMENTAL SECTION
4.1. Equipment and Methods. Mass spectra were acquired on a
ThermoFinnigan LCQ Advantage ion trap instrument, detecting
positive ions (+) or negative ions (−) in the ESI mode. Liquid-state
1H and 19F NMR spectra were obtained using a Bruker AM 250
spectrometer. Solid-state 19F NMR spectra were acquired on a Bruker
DSX 400 spectrometer. The UV−vis spectra of solutions were
obtained with an Agilent 8453 diode-array spectrophotometer. The
reaction products were identified by GC−MS [Hewlett-Packard 5973/
6890 system; electron impact ionization at 70 eV, He carrier gas, 30 m
× 0.25 mm HP-INNOWax capillary column with poly(ethylene
glycol) (0.25 μm coating) or DB-5MS 50 m capillary column (0.250
mm × 0.25 μm)]. The reaction products were analyzed using The
NIST Mass Spectral Search Program of NIST/EPA/NIH Mass
Spectral Library (version 2.0f, July 23, 2008). In order to analyze
organic acids, reaction mixtures were treated with a 2.0 M solution of
(trimethylsilyl)diazomethane in diethyl ether to transform caboxylic
groups to the methyl esters. Response factors of phenol products were
determined using authentic compounds. When authentic compounds
were not available, phenol products were analyzed by assuming that
they have the similar response factors as their available isomers. Total
organic carbon analyses were performed on a Shimadzu TOC-5050A
spectrometer. Infrared spectra were recorded on a Bruker Vector 22
FTIR spectrometer using KBr pellets. EPR spectra were recorded on a
Bruker Elexsys e500 spectrometer (conditions: 25 °C, microwave
frequency 9.415 GHz, power 0.4 mW, modulation 1.0 mT/100 kHz).
X-ray emission spectra were measured at beamline ID 26 at the
European Synchrotron Radiation Facility (ESRF) (Grenoble, France).
The electron energy was 6.0 GeV, and the ring current was varied from
50 to 90 mA. Two u35 undulators were used to perform the
measurements. To detect the (resonant) inelastic X-ray scattering, the
sample pellet was aligned at an angle of 45° with respect to the X-ray
beam. The incident X-ray energy was selected by a pair of Si crystals
cut in the (220) orientation. The beam was focused in a small spot
(350 μm × 60 μm) on the sample. The scattered X-rays were
monochromatized by the (531) Bragg planes of a spherical bent Si
crystal and focused on an avalanche photodiode (APD). When the
energy of the scattered X-rays was scanned, the APD detector and the
spherical bent Si crystal were moved concertedly in order to keep the
beam spot on the sample, the bent crystal, and the detector on a
Rowland circle. A He bag was fixed between the sample, analyzer
crystal, and detector in order to minimize the absorption of the X-rays
by air. In order to minimize radiation damage, the samples were cooled
to a temperature in the range 20−30 K, and the position of the beam
on the sample was changed regularly between scans. In order to be
H
dx.doi.org/10.1021/ja505437h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
sure that the X-ray beam did not destroy the iron species under study,
we checked the absence of radiation damage during the exposure time.
To do this, prior to every measurement, 30 scans were performed (30
s each), and XANES profiles were detected. The absence of evolution
within the series of subsequent XANES spectra was applied as a
criterion of stability under the beam.
dried in vacuum at room temperature for 1 h and at 60 °C for 6 h. The
complex loading was 10 μmol/g.
4.5. Preparation of the μ-Nitrido Diiron(IV) Tetra-tert-
butylphthalocyanine Difluoride Complex. μ-Nitrido diiron tert-
tetrabutylphthalocyanine (16 mg) was dissolved in 10 mL of a 1:1
C6H6/C6F6 mixture. After addition of a 70% aqueous solution of
tBuOOH (40 μL, 30 equiv with respect to the complex), the reaction
mixture was stirred at 60 °C for 6 h. The solvent was removed, and the
isolated product was washed with water and dried at 60 °C in vacuum.
The yield of the dark-blue powder was 13 mg (80%).
4.6. Typical Procedures for Oxidative Defluorination.
4.6.1. Homogeneous Oxidation in Acetonitrile. A 3 mL Teflon
reactor was charged with 2 mL of CD3CN containing polyfluorinated
substrate (0.1 M), H2O2 (typically 1.6 M), and catalyst (0.4 mM). The
reactor was kept at the desired temperature (20 or 60 °C) with
magnetic stirring for 1, 3, 6, or 15 h. After cooling to room
temperature, the reaction mixture was analyzed by GC−MS and 19F
NMR methods.
4.6.2. Homogeneous Oxidation under Fenton Conditions. A 3
mL Teflon reactor was charged with 2 mL of CD3CN or D2O
containing fluorinated substrate (0.1 M), H2O2 (1.6 M), and a 4 mM
solution of iron(II) sulfate dihydrate in water (0.4 mM). The reactor
was kept at 60 °C with magnetic stirring for 15 h. After cooling to
room temperature, the reaction mixture was analyzed by the 19F NMR
method.
4.6.3. Heterogeneous Oxidation in Water. A 3 mL Teflon reactor
was charged with 2 mL of D2O containing polyfluorinated substrate
(0.1 M in the case of water-soluble pentafluorophenol, pentafluor-
obenzoic acid, tetrafluorophthalic acid, and pentafluorosulfonic acid;
0.005 and 0.003 M for pentafluoropyridine and hexafluorobenzene,
respectively), H2O2 (0.1, 0.8, or 2.6 M), and carbon-supported
(FePctBu4)2N catalyst [20 mg containing 0.2 μmol of (FePctBu4)2N].
The reactor was kept at 60 °C with magnetic stirring for 15 h. After
the mixture was cooled to room temperature, the solid catalyst was
separated by filtration and the reaction mixture was analyzed by GC−
MS and 19F NMR methods. The total organic carbon content was
measured in several experiments.
The EXAFS and XANES spectra were recorded on the SAMBA
beamline at the SOLEIL synchrotron (Gif-sur-Yvette, France)
operating at 300 mA and 2.75 GeV. Spectra were collected in
transmission mode at the Fe K edge with a sagittal focusing Si(220)
double crystal and focusing mirrors graded at 5 mrad to remove the
harmonics. The beam spot was defocused to prevent beam damage to
the sample. To compare the pre-edge energies, a metallic Fe foil
reference was applied. The first inflection point of metallic Fe was
observed at 7111.6 eV. The data were treated with the FEFF36 and
VIPER37 programs, and then the edge background was extracted using
Bayesian smoothing with a variable number of knots. The curve fitting
was done alternatively in the R and k spaces, and the fit was accepted
only in the case of simultaneous convergence in k and R spaces
(absolute and imaginary parts for the latter). The coordination
numbers (CN), interatomic distances (R), Debye−Waller (DW)
parameters (σ2), and energy shifts (ΔE0) were used as fitting variables.
Constraints were introduced on the variables such as DW factors and
energy shifts to get values lying in physically reasonable intervals.
Constraints imposed by the molecular structure were also introduced.
The quality of the fit was evaluated using the values of variance and
goodness.
4.2. Determination of Substrate Conversions and Product
Yields by 19F NMR Spectroscopy. The disappearance of fluorinated
substrates and the appearance of fluorinated products were determined
by 19F NMR spectroscopy using a hexafluoroisopropanol solution in
CD3CN as an external standard in a sealed capillary placed in an NMR
tube or CF3COOH as an internal standard. Prior to analyses of the
products of catalytic reactions, calibration was performed with
solutions of KF in D2O and solutions of hexafluorobenzene in
CD3CN. In both cases, the integration values linearly depended on the
KF and C6F6 concentrations. The signal intensities of inorganic and
organic fluorine atoms were identical. In some experiments performed
in water, CF3COOH was used as a standard. The total mass balance
for fluorine was >90% by 19F NMR spectroscopy. Hydrofluoric acid is
a weak acid and is present in dilute aqueous solutions in nonionized
and ionized forms showing two 19F NMR signals. On the basis of
literature data, a signal at −129.9 ppm was assigned to F−. This
assignment was confirmed by measurement of the 19F NMR spectrum
of a KF solution in D2O, which showed the signal with the same
chemical shift. This signal was observed only using D2O solvent. When
the reactions were performed in CD3CN, this signal was absent. A
broad peak at approximately −160 ppm was assigned to hydrated
nonionized HF. Concentrations of fluoride ion were additionally
determined by a spectrophotometric method according to a published
protocol.38 The fluoride ion concentrations determined by the two
methods were the same within experimental errors of 5−10%.
4.3. Determination of Gas-Phase Products by the Micro-
GC−MS Method. Typical amounts of the supported catalyst,
pentafluorophenol, and H2O2 were placed in a Schlenk tube under
an argon atmosphere, and the gas phase was immediately analyzed by
the micro-GC−MS method. CO was not detected, whereas the
background content of CO2 was equal to 0.15%. The reaction mixture
was stirred under argon at 60 °C for 15 h. The analysis of the gas
phase after reaction showed 48.18% CO2 content and 3.27% CO
content.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental details, characterization data, and details of
catalysis studies. This material is available free of charge via
the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Authors
alexander.sorokin@ircelyon.univ-lyon1.fr
pavel.afanasiev@ircelyon.univ-lyon1.fr
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Research support was provided by the Agence Nationale de
Recherche (ANR, France, Grant ANR-08-BLANC-0183-01).
́ ̂
C.C. is grateful to the Region Rhone-Alpes for Ph.D. fellowship.
We thank Dr. D. Bouchu for the help with ESI-MS. We
acknowledge the SOLEIL (Gif-sur-Yvette, France) and the
ESRF (Grenoble, France) for provision of time on the SAMBA
and ID26 beamlines, respectively. We thank Dr. V. Briois
(SOLEIL), Dr. J. C. Swarbrick (ESRF), and Dr. M. Rovezzi
(ESRF) for technical assistance.
4 . 4 . P r e p a r a t i o n o f C a t a l y s t s . T e t r a - t e r t -
butylphthalocyaninatoiron(II) was synthesized and purified according
to a published protocol.39 μ-Nitridobis(tetra-tert-butylphthalocyanina-
toiron), (FePctBu4)2N, was prepared as described previously.10a The
supported catalyst was prepared as follows: (FePctBu4)2N (80 mg, 50
μmol) was dissolved in 200 mL of CH2Cl2, and 5 g of carbon (HSAG
300, specific surface area 300 m2/g) was added to the solution. The
resulting mixture was stirred for 6 h at room temperature, and the
solvent was evaporated under reduced pressure. The solid material was
REFERENCES
(1) Amii, H.; Uneyama, K. Chem. Rev. 2009, 109, 2119.
(2) Grushin, V. V. Acc. Chem. Res. 2010, 43, 160.
■
I
dx.doi.org/10.1021/ja505437h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(3) (a) Douvris, C.; Ozerov, O. V. Science 2008, 321, 1188.
(b) Perutz, R. N. Science 2008, 321, 1168.
(4) Clot, E.; Eisenstein, O.; Jasim, N.; Macgregor, S. A.; McGrady, J.
E.; Perutz, R. N. Acc. Chem. Res. 2011, 44, 333.
(5) Baumgartner, R.; McNeill, K. Environ. Sci. Technol. 2012, 46,
10199.
(6) (a) Lv, H.; Cai, Y.-B.; Zhang, J.-L. Angew. Chem., Int. Ed. 2013,
52, 3203. (b) Yow, S.; Gates, S. J.; White, A. J. P.; Crimmin, M. R.
Angew. Chem., Int. Ed. 2012, 51, 12559. (c) Reade, S. P.; Mahon, M. F.;
Whittlesey, M. K. J. Am. Chem. Soc. 2009, 131, 1847. (d) Douvris, C.;
Nagaraja, C. M.; Chen, C.-H.; Foxman, B. M.; Ozerov, O. V. J. Am.
Chem. Soc. 2010, 132, 4946.
(22) Bogachev, A. A.; Kobrina, L. S.; Yakobson, G. G. Russ. J. Org.
Chem. 1986, 22, 2313.
(23) Osman, A. M.; Posthumus, M. A.; Veeger, C.; van Bladeren, P.
J.; Laane, C.; Rietjens, I. M. C. M. Chem. Res. Toxicol. 1998, 11, 1319.
(24) Murphy, C. D. Biotechnol. Lett. 2009, 32, 351.
(25) Rietjens, I. M.; Vervoort, J. Chem. Res. Toxicol. 1992, 5, 10.
(26) (a) Kuehnel, M. F.; Lentz, D.; Braun, T. Angew. Chem., Int. Ed.
́ ́
2013, 52, 3328. (b) Nova, A.; Mas-Balleste, R.; Lledos, A.
Organometallics 2012, 31, 1245. (c) Sun, L.; Rong, M.; Kong, D.;
Bai, Z.; Yuan, Y.; Weng, Z. J. Fluorine Chem. 2013, 150, 117.
(d) Arisawa, M.; Suzuki, T.; Ishikawa, T.; Yamaguchi, M. J. Am. Chem.
Soc. 2008, 130, 12214.
(27) Liu, C.; Cao, L.; Yin, X.; Xu, H.; Zhang, B. J. Fluorine Chem.
2013, 156, 51.
(28) An NIH shift is the migration of a substituent at a hydroxylation
site to an adjacent position. See: (a) Jerina, D. M.; Daly, J. W. Science
1974, 185, 573. (b) de Visser, S. P.; Shaik, S. J. Am. Chem. Soc. 2003,
125, 7413. (c) Mitchell, K. H.; Rogge, C. E.; Gierahn, T.; Fox, B. G.
Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 3784.
(29) Rietjens, I. M.; den Besten, C.; Hanzlik, R. P.; van Bladeren, P. J.
Chem. Res. Toxicol. 1997, 10, 629.
(30) Koerts, J.; Soffers, A. E.; Vervoort, J.; De Jager, A.; Rietjens, I. M.
Chem. Res. Toxicol. 1998, 11, 503.
(31) (a) Daly, J. W.; Jerina, D. M.; Witkop, B. Experientia 1972, 28,
1129. (b) Schwartz, H.; Chu, I.; Villeneuve, D. C.; Benoit, F. M. J.
Toxicol. Environ. Health 1987, 22, 341. (c) Bogaards, J. J. P.; Van
Ommen, B.; Wolf, C. R.; Van Bladeren, P. J. Toxicol. Appl. Pharmacol.
1995, 132, 44.
(32) Rietjens, I. M. C. M.; Vervoort, J. Xenobiotica 1989, 19, 1297.
(33) Sisemore, M. F.; Selke, M.; Burstyn, J. N.; Valentine, J. S. Inorg.
Chem. 1997, 36, 979.
(34) (a) Adams, D. E. C.; Halden, R. U. ACS Symp. Ser. 2010, 1048,
539−560. (b) Murphy, M. B.; Loi, E. I. H.; Kwork, K. Y.; Lam, P. K. S.
In Fluorous Chemistry; Horvath, I., Ed.; Springer: Berlin, 2011; pp
339−364.
(35) Shteingarts, V. D. J. Fluorine Chem. 2007, 128, 797 and
references therein.
(7) Hackett, J. C.; Sanan, T. T.; Hadad, C. M. Biochemistry 2007, 46,
5924.
́
(8) (a) Sorokin, A.; Seris, J.-L.; Meunier, B. Science 1995, 268, 1163.
(b) Sorokin, A.; Meunier, B. Chem.Eur. J. 1996, 2, 1308. (c) Sorokin,
A.; De Suzzoni-Dezard, S.; Poullain, D.; Noel, J.-P.; Meunier, B. J. Am.
̈
Chem. Soc. 1996, 118, 7410.
(9) Sorokin, A. B. Chem. Rev. 2013, 113, 8152.
(10) (a) Sorokin, A. B.; Kudrik, E. V.; Bouchu, D. Chem. Commun.
2008, 2562. (b) Sorokin, A. B.; Kudrik, E. V.; Alvarez, L. X.; Afanasiev,
P.; Millet, J.-M. M.; Bouchu, D. Catal. Today 2010, 157, 149.
(c) Kudrik, E. V.; Sorokin, A. B. Chem.Eur. J. 2008, 14, 7123.
̇ ̈
(d) Isç i, U.; Afanasiev, P.; Millet, J.-M. M.; Kudrik, E. V.; Ahsen, V.;
Sorokin, A. B. Dalton Trans. 2009, 7410. (e) Afanasiev, P.; Bouchu, D.;
Kudrik, E. V.; Millet, J.-M. M.; Sorokin, A. B. Dalton Trans. 2009,
9828. (f) Alvarez, L. X.; Kudrik, E. V.; Sorokin, A. B. Chem.Eur. J.
2011, 17, 9298.
(11) Kudrik, E. V.; Afanasiev, P.; Alvarez, L. X.; Dubourdeaux, P.;
Clemancey, M.; Latour, J.-M.; Blondin, G.; Bouchu, D.; Albrieux, F.;
́
Nefedov, S. E.; Sorokin, A. B. Nat. Chem. 2012, 4, 1024.
(12) (a) Nyokong, T.; Gasyna, Z.; Stillman, M. J. Inorg. Chem. 1987,
26, 548. (b) Ough, E.; Gasyna, Z.; Stillman, M. J. Inorg. Chem. 1991,
30, 2301. (c) Afanasiev, P.; Kudrik, E. V.; Millet, J.-M. M.; Bouchu, D.;
Sorokin, A. B. Dalton Trans. 2011, 40, 701.
(13) Glatzel, P.; Bergmann, U. Coord. Chem. Rev. 2005, 249, 65.
(14) (a) Chandrasekaran, P.; Stieber, S. C. E.; Collins, T. J.; Que, L.,
Jr.; Neese, F.; DeBeer, S. Dalton Trans. 2011, 40, 11070. (b) Jackson,
T. A.; Rohde, J.-U.; Seo, M. S.; Sastri, C. V.; DeHont, R.; Stubna, A.;
(36) Ankudinov, A. L.; Bouldin, C. E.; Rehr, J. J.; Sims, J.; Hung, H.
Phys. Rev. B 2002, 65, 104.
(37) Klementiev, K. V. J. Synchrotron Radiat. 2001, 8, 270.
(38) Bellack, E.; Schouboe, P. J. Anal. Chem. 1958, 30, 2032.
(39) Metz, J.; Schneider, O.; Hanack, M. Inorg. Chem. 1984, 23, 1065.
Ohta, T.; Kitagawa, T.; Munck, E.; Nam, W.; Que, L., Jr. J. Am. Chem.
̈
Soc. 2008, 130, 12394. (c) Aliaga-Alcalde, N.; Mienert, B.; Bill, E.;
Wieghardt, K.; DeBeer George, S.; Neese, F. Angew. Chem., Int. Ed.
2005, 44, 2908. (d) Berry, J. F.; Bill, E.; Bothe, E.; DeBeer George, S.;
Mienert, B.; Neese, F.; Wieghardt, K. Science 2006, 312, 1937. (e) de
Oliveira, F. T.; Chanda, A.; Banerjee, D.; Shan, X.; Mondal, S.; Que, L.,
Jr.; Bominaar, E. L.; Munck, E.; Collins, T. J. Science 2007, 315, 835.
̈
(15) Newcomb, M.; Halgrimson, J. A.; Horener, J. H.; Wasinger, E.
C.; Chen, L. X.; Sligar, S. G. Proc. Natl. Acad. Sci. U.S.A. 2008, 105,
8179.
(16) (a) Delgado-Jaime, M. U.; Dible, B. R.; Chiang, K. P.;
Brennessel, W. W.; Bergmann, U.; Holland, P. L.; DeBeer, S. Inorg.
Chem. 2011, 50, 10709. (b) Pollock, C. J.; DeBeer, S. J. Am. Chem. Soc.
2011, 133, 5594.
(17) Lancaster, K. M.; Hu, Y.; Bergmann, U.; Ribbe, M. W.; DeBeer,
S. J. Am. Chem. Soc. 2013, 135, 610.
(18) (a) Kadish, K. M.; Rhodes, R. K. Inorg. Chem. 1983, 22, 1090.
(b) Kadish, K. M.; Bottomley, L. A. Inorg. Chem. 1981, 20, 1348.
(19) (a) Takahashi, A.; Yamaki, D.; Ikemura, K.; Kurahashi, T.;
Ogura, T.; Hada, M.; Fujii, H. Inorg. Chem. 2012, 51, 7296. (b) Gross,
Z. J. Biol. Inorg. Chem. 1996, 1, 368. (c) de Visser, S. P.; Tahsini, L.;
Nam, W. Chem.Eur. J. 2009, 15, 5577. (d) Kang, Y.; Chen, H.;
Jeong, Y. J.; Lai, W.; Bae, E. H.; Shaik, S.; Nam, W. Chem.Eur. J.
2009, 15, 10039.
(20) Bogachev, A. A.; Kobrina, L. S.; Yakobson, G. G. Russ. J. Org.
Chem. 1986, 22, 2307.
(21) (a) Ravichandran, L.; Selvam, K.; Swaminathan, M. J.
Photochem. Photobiol., A 2007, 188, 392. (b) Ravichandran, L.;
Selvam, K.; Swaminathan, M. Desalination 2010, 260, 18.
J
dx.doi.org/10.1021/ja505437h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX