Regiospecific intramolecular O-demethylation of the ligand by action of molecular dioxygen on a ferrous complex: Versatile coordination chemistry of dioxygen in FeCl2 complexes with (2,3-dimethoxyphenyl) α-sinstituted tripods in the tris(2-pyri

Article abstract of DOI:10.1021/ic802359z

We report in this article one of the first examples of a reaction of O-demethylation carried out at a Fe(II) center by molecular dioxygen, in the homogeneous phase in non-porphyrinic chemistry. This reaction parallels at the intramolecular level a very im

Full text of DOI:10.1021/ic802359z

Inorg. Chem. 2009, 48,
DOI:10.1021/ic802359z
4777
Regiospecific Intramolecular O-Demethylation of the Ligand by Action of Molecular
Dioxygen on a Ferrous Complex: Versatile Coordination Chemistry of Dioxygen
in FeCl2 Complexes with (2,3-Dimethoxyphenyl) r-Substituted Tripods in the
tris(2-Pyridylmethyl)amine Series
Laila Benhamou,†,‡ Ahmed Machkour,†,‡ Olaf Rotthaus,† Mohammed Lachkar,‡ Richard Welter,§ and
Dominique Mandon*,†
†
ꢀ
ꢀ
Laboratoire de Chimie Biomimetique des Metaux de Transition, Institut de Chimie, UMR CNRS no. 7177,
‡
ꢀ
Universite Louis Pasteur, 4 rue Blaise Pascal, B.P. 1032, F-67070 Strasbourg cedex, France, Laboratoire
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
d’Ingenierie des Materiaux Organometalliques et Moleculaires, Universite Sidi Mohamed Ben Abdellah, Faculte
des Sciences, BP 1796 (Atlas), 30000 Fez, Morocco and §Laboratoire DECOMET, Institut de Chimie, UMR
ꢀ
CNRS no. 7177, Universite Louis Pasteur, 4 rue Blaise Pascal, B.P. 1032, F-67070 Strasbourg cedex, France
Received December 11, 2008
We report in this article one of the first examples of a reaction of O-demethylation carried out at a Fe(II) center by
molecular dioxygen, in the homogeneous phase in non-porphyrinic chemistry. This reaction parallels at the
intramolecular level a very important process found in biology leading to the derivatization and elimination of drugs
by oxygen-dependent enzymes that contain nonheme iron centers. To get insight into some reactivity aspects of this
reaction, we have used dioxygen and iron complexes coordinated to ligands that are substituted by methoxy groups.
We detail in this work the coordination chemistry of FeCl2 to the series of mono- (L1), di- (L2), and tris(2,3-
dimethoxyphenyl) (L3) R-substituted ligands in the tris(2-pyridylmethyl)amine series and the behavior of the complexes
upon reaction with molecular dioxygen. As main outcomes of this study, we demonstrate that the methoxy group does
not need to be coordinated to the metal center to undergo O-demethylation, but needs to be properly orientated close
to an oxygenated form of the metal. We also demonstrate the importance of the environment in the reactivity with
molecular dioxygen: whereas a regular 18-electron Fe(II) reacts with O2, a five- coordinate, 16-electron center may be
oxygen-stable, if the access of dioxygen to the reaction site is locked.
Introduction
modification of their structure.4-7 From this viewpoint,
oxidative O-dealkylation is an important and well-known
reaction in the metabolism of drugs, which is believed to
occur at the metal site center during the catalytic cycle of
heme-containing enzymes.8-10 In this context, nonheme
proteins have by far been less-studied, although similar
reactions were reported in early studies, sometimes as side-
and nondesired transformations,11,12 occasionally encoura-
ging the setup of simple synthetic functional systems.13,14
The activation of dioxygen at inexpensive metal centers is
of fundamental importance from both academic and applied
points of view in the current economical context,1,2 for many
groups expressing a renewed interest in the biomimetic
approach.3 The presence of iron centers and dioxygen turns
out to be crucial in many oxidation processes encountered in
biology, in biosynthetic pathways, of course, but also in
processes involving the metabolism of xenobiotics, further
leading to their easy removal from living organisms after
(8) Dowers, T. S.; Jones, J. P. Drug Metab. Dispos. 2006, 34, 1288
1290.
(9) Relling, M. V.; Nemec, J.; Schuetz, E. G.; Schuetz, J. D.; Gonzalez, F.
J.; Korzekwa, K. R. Mol. Pharmacol. 1994, 45, 352–358.
(10) Yu, A.; Dong, H.; Lang, D.; Haining, R. L. Drug Metab. Dispos.
2001, 29, 1362–1365.
(11) Funabiki, T.; Toyoda, T.; Ishida, H.; Tsujimoto, M.; Ozawa, S.;
Yoshida, S. J. Mol. Catal. 1990, 61, 235–246.
(12) Mathieu, D.; Bartoli, J.-F.; Battioni, P.; Mansuy, D. Tetrahedron
2004, 60, 3855–3862.
*To whom correspondence should be sent. Phone: 33-(0)390-241-537.
Fax: 33-(0)390-245-001. E-mail: mandon@chimie.u-strasbg.fr.
(1) Labinger, J. A. J. Mol. Catal. A 2004, 220, 27–35.
(2) Tong, J. H.; Li, Z.; Xia, C. G. Prog. Chem. 2005, 17, 96–110.
(3) Special issue on dioxygen activation by metalloenzymes and models:
Acc. Chem. Res., 2007, 40, 465-634.
(4) Shikama, K. Prog. Biophys. Mol. Biol. 2006, 91, 83–162.
(5) Kovacs, J. A. Science 2003, 299, 1024–1025.
(6) Costas, M.; Mehn, M. P.; Jensen, M. P.; Que, L.Jr. Chem. Rev. 2004,
(13) Katopodis, A. G.; Wilamasena, K.; Lee, J.; May, S. W. J. Am. Chem.
104, 939–986.
Soc. 1984, 106, 7928–7935.
(7) Neidig, M. L.; Solomon, E. D. Chem. Commun. 2005, 47, 6843
5863.
(14) Duprat, A. F.; Capdevielle, P.; Maumy, M. J. Mol. Catal. 1992, 77,
L7–L11.
©
2009 American Chemical Society
Published on Web 4/16/2009
pubs.acs.org/IC

4778 Inorg. Chem., Vol. 48, No. 11, 2009
Benhamou et al.
O-demethylation carried out under homogeneous conditions
upon reaction with O2 with a ferrous nonporphyrinic
complex. Additionally, we bring evidence that this reaction
involves an initial coordination of dioxygen followed by a
regiospecific process, with no need for the methoxy group to
coordinateto the metal center. This particular reaction can be
considered as a salient extension of the nonheme chemistry of
an important reaction already known in P450 chemistry.
Figure 1. Ligands used in this study.
Experimental Section
We are currently addressing the question of the iron-
dioxygen interaction in ferrous complexes in which FeCl2 is
coordinated to R-substituted ligands in the tris(2-pyridyl-
methyl)amine (TPA) series.15-17 The general formation of
μ-oxo diferric complexes is observed with two kinetically
distinguishable steps upon reaction of dry molecular oxygen
with theferrous precursors.17 Acceleration ofthe reaction has
been shown to occur upon increasing the electron deficiency
at the metal center within a series of isostructural complexes
with distorted octahedral geometries.17 Additionally and as
expected, the presence of a vacant coordination site around
the metal dramatically increases the reaction kinetics.16,17
We already reported the preparation of the bis(2-[2,3-
dimethoxy phenyl] 6-pyridylmethyl) (2-pyridylmethyl)amine
ligand L2, drawn in Figure 1, as a precursor of the bis-
catecholato R-substituted TPA for which the coordination
chemistry was investigated.18 Using L2 as a ligand in a class of
oxygen-sensitive compounds, we expected that at least one
methoxy group would lie close to the metal center, providing
insight into the behavior of this noninnocent group under
similar conditions, such as those present in biological sys-
tems. Additionnally, methoxy-substituted phenyl groups can
be considered as relatively bulky and electron-donating
substituents,19,20 allowing comparison of the reactivity versus
O2 of the corresponding complexes with that of similar
electron-deficient compounds.16,17 We thus started the pre-
paration of the series of mono-, bis-, and tris-substituted
ligands L1, L2, and L3 and decided to test the reactivity of the
corresponding FeCl2 complexes versus dioxygen. The result
is striking and is reported in this article. Whereas L1FeCl2
reacts slowly to yield a μ-oxo complex in a classical way,
L3FeCl2 reacts only very poorly with dioxygen, in spite of a
coordinatively trigonal bipyramidal unsaturated geometry.
This very low sensitivity is due to a particular environment
hampering coordination of O2 to the metal center. The case
of L2FeCl2 is unique since it reacts smoothly, yielding a
dinuclear complex in which the ligand has undergone
O-demethylation in the ortho position of the substituent.
This compound is easily reduced, and the crystal structure of
the diferrous corresponding complex is reported. The present
manuscript describes one of the very few examples of
General Considerations. Chemicals were purchased from
Aldrich Chemicals and used as received. Precursor ligands
BrTPA, Br2TPA, and Br3TPA were prepared according to
published procedures: Chuang, C. L.; Dos Santos, O.; Xu, X.;
Canary, J. W. Inorg. Chem., 1997, 36, 1967-1972. Ligand L2
was prepared following the procedure described in ref 18. All
solvents used during the metalation reactions and workup were
distilled and dried according to the procedure given in W. L. F.
Armarego, W. L. F.; Perrin, D. D. Purification of Laboratory
Chemicals, 4th ed.; Pergamon Press: Oxford, U. K., 1997.
Analytical anhydrous FeCl2 was obtained as a white powder
by reacting iron powder (ACS grade) with hydrochloric acid in
the presence of methanol under an argon atmosphere. Prepara-
tion and handling of all compounds (even L3FeCl2, which is air
stable) were performed under an argon atmosphere using
Schlenk techniques, following standard procedures. The purity
of the dry dioxygen was 99.99999% (grade 5). Zinc amalgam
was prepared as already described in ref 17. Elemental analyses
ꢀ ꢀ
were carried out by the Service d’Analyses de la Federation de
Recherche de Chimie, Universite Louis Pasteur, Strasbourg,
France, and by the Service Central d’Analyses du CNRS,
Solaize, France. Mass spectroscopy experiments were carried
out by the Service Commun de Spectrometrie de Masse de
l’Institut de Chimie de Strasbourg.
ꢀ
Physical Methods. 1H NMR data were recorded in CD3CN
for the complexes and CDCl3 for the ligands at ambient
temperature on a Bruker AC 300 spectrometer at 300.1300
MHz using the residual signal of CD2HCN (CHCl3) as a
reference for calibration. The UV-vis spectra were recorded
on a Varian Cary 05 E UV-vis NIR spectrophotometer
equipped with an Oxford instrument DN1704 cryostat, using
optically transparent Schlenck cells. Conductivity measure-
ments were carried out under argon at 20 °C with a CDM 210
Radiometer Copenhagen Conductivity Meter, using a Tacussel
CDC745-9 electrode. Cyclic voltammetry measurements were
obtained from a PAR 173A potentiostat in a 0.1 M acetonitrile
solution of TBAPF6 (supporting electrolyte), using platinum
electrodes and a saturated calomel electrode as a reference. For
each measurement, the potential was checked by the addition of
a small amount of ferrocene (Fc/Fc+ = 0.380 v/SCE) in the cell.
Synthesis of L1: Mono(2-[2,3-dimethoxy Phenyl] 6-Pyr-
idylmethyl) Bis(2-pyridylmethyl)amine. BrTPA (794 mg,
2.15 mmol) and Pd(PPh3)4 (181 mg, 0.15 mmol) were suspended
under argon in degassed toluene (100 mL). Sodium carbonate (10
mL of a 2 M aqueous solution) and commercially available 2,3-
dimethoxyphenylboronic acid (704 mg, 3.87 mmol) dissolved in
ethanol (5 mL) were added under argon. The reaction mixture
was refluxed under argon for 48 h. The orange-colored solution
was then evaporated to dryness. The remaining oil was dissolved
in methylene chloride, and this organic phase was washed with
aqueous sodium carbonate and distilled water and dried over
magnesium sulfate. After filtration and concentration by rotary
evaporation, the medium was deposited at the top of a column
filled with silica and mounted with acetone. The column was
washed with diethyl ether and the compound eluted with acetone.
A brownish oil (662 mg, 72%) was obtained. Elem anal. calcd for
C26H26N4O2, %: C, 73.24; H, 6.10. Found, %: C, 72.80; H, 6.01.
1H NMR, δ, ppm, CDCl3: 3.65 (s, 3H), 3.90 (s, 3H), 3.93 (s, 4H),
(15) Mandon, D; Machkour, A.; Goetz, S.; Welter, R. Inorg. Chem. 2002,
41, 5363–5372.
(16) Machkour, A.; Mandon, D.; Lachkar, M.; Welter, R. Inorg. Chem.
2004, 43, 1545–1550.
(17) Thallaj, N. K.; Rotthaus, O.; Benhamou, L.; Humbert, N.; Elhabiri,
M.; Lachkar, M.; Welter, R.; Albrecht-Gary, A.-M.; Mandon, D.
Chem.;Eur. J. 2008, 14, 6742–6753.
(18) Machkour, A.; Thallaj, N. K.; Benhamou, L.; Lachkar, M.; Mandon,
D. Chem.;Eur. J. 2006, 12, 6660–6668.
(19) Jensen, M. P.; Lange, S. J.; Mehn, M. P.; Que, E. L.; Que, L.Jr. J. Am.
Chem. Soc. 2003, 125, 2113–2128.
(20) He, Z.; Colbran, S. B.; Craig, D. Chem.;Eur. J. 2003, 9, 116–129. see
also this reference for a good example of flexibility of a 2-5-dimethoxyphenyl
substituted tripod in a copper complex.

Article
Inorg. Chem., Vol. 48, No. 11, 2009
4779
3.95 (s, 2H), 6.93-6.94 (d, 1H), 6.96-6.97 (d, 1H), 7.11-7.14 (m,
2H), 7.15-7.17 (t, 1H), 7.34-7.35 (d, 1H), 7.36-7.37 (d, 1H),
7.52-7.55 (t, 1H), 7.62-7.66 (m, 2H), 7.69-7.70 (d, 2H),
8.52-8.53 (d,1H), 8.54-8.55 (d, 1H). The trace is shown in the
Supporting Information. MSES+: m/z + Li+ = 433.21.
Synthesis of L3: Tris(2-[2,3-dimethoxy Phenyl] 6-Pyr-
idylmethyl) Amine. Br3TPA (400 mg, 0.76 mmol) and Pd
(PPh3)4 (41 mg, 0.035 mmol) were suspended under argon in
degassed toluene (100 mL). Sodium carbonate (11 mL of a 2 M
aqueous solution) and commercially available 2,3-dimethoxy-
phenylboronic acid (280 mg, 1.58 mmol) dissolved in ethanol
(5 mL) were added under argon. The reaction mixture was refl-
uxed under argon for several days, until no more Br3TPA was
detected by thin-layer chromatography (several days). The or-
ange-colored solution was then evaporated to dryness. The re-
maining oil was dissolved in methylene chloride, and this organic
phase was washed with aqueous sodium carbonate and distilled
water and dried over magnesium sulfate. After filtration and
concentration by rotary evaporation, the medium was deposited
at the top of a column filled with silica and mounted with acetone.
The column was washed with diethyl ether and the compound
eluted with acetone. A brownish oil (320 mg, 61%) was obtained.
Elem anal. calcd for C42H42N4O6, %: C, 72.19; H, 6.06. Found,
%: C, 72.19; H, 5.98. 1H NMR, δ, ppm, CDCl3: 3.65 (s, 9H), 3.86
(s,9H), 4.06 (s, 6H), 6.93-7.70 (m, 18H). The trace is shown in the
Supporting Information. MSES+: m/z + H+ = 699.23.
free OH- signal in IR spectroscopy, strongly supports the pre-
sence of the [L2*Fe-OH-FeCl3]+Cl- structure. All spectro-
scopic data are shown in the Supporting Information.
Preparation of L2*FeCl3. A total of 100 mg (0.12 mmol) of
the green compound was dissolved under argon in CH3CN. The
solution was transferred into a Schlenk tube containing several
drops of 10% zinc amalgam. The medium was stirred and
became bright orange over 10 min. This solution turned out to
be extremely oxygen-sensitive, becoming dark violet in the case
of a leak into the reaction vessel. The addition of diethyl ether
allowed precipitation of an orange solid material, which was
recrystallized from CH3CN/Et2O. A total of 80 mg was
obtained, corresponding to an 83% yield. Even in the solid
state, this compound is difficult to handle. We believe that this is
the reason why a satisfactory elemental analysis could not be
obtained. Elem anal. calcd for C33H31Cl3Fe2N4O4, %: C, 51.73;
H, 4.05; Fe, 14.63; Cl, 13.91. Found, %: C, 50.21; H, 3.82; Fe,
13.17; Cl, 13.05. Spectroscopic data are detailed in the text and
shown in the Supporting Information.
Single crystals were obtained by slow diffusion of diethyl
ether in a sealed tube containing a CH3CN solution of L2FeCl2,
a CH3CN solution of L2*Fe2Cl3, and a THF solution of
L3FeCl2. Single crystals of [L1FeCl2]0 were obtained by layering
with diethyl ether a CH3CN solution of L1FeCl2 that was
previously oxygenated.
X-Ray Analysis. Single crystals of [L1FeCl2]2O CH3CN,
L2FeCl2, L3FeCl2 THF, and L2*Fe2Cl3 OEt2 were mounted
3
Preparation of the FeCl2 Complexes. Details are given for
L1FeCl2, but the following procedure applies to all complexes:
150 mg (0.35 mmol) of free L1 was dissolved in a Schlenck tube
containing 20 mL of dry and degassed THF. A total of 40 mg
(0.31 mmol) of anhydrous FeCl2 was dissolved in a second
Schlenck tube containing 10 mL of dry and degassed THF. The
solution of FeCl2 was transferred under argon in the Schlenck
containing the ligand, and the medium was stirred overnight.
The solvent was then evaporated to dryness, and the compound
was extracted with dry and degassed CH3CN, filtered under an
inert atmosphere, and concentrated. The addition of diethyl
ether afforded a yellow solid, which was washed thoroughly
with this solvent, prior to being dried under vacuum conditions.
A total of 140 mg (82%) of L1FeCl2 with good analytical and
spectroscopic data could be obtained. Spectroscopic data are
detailed in the text and shown in the Supporting Information.
Oxygenation of L2FeCl2. A total of 150 mg (0.22 mmol) of
the ferrous complex was dissolved under argon in 50 cm3 of
CH3CN in a Schlenk tube. Dry dioxygen was bubbled for 15 s,
and the medium was kept under oxygen in the stopped Schlenk
and stirred over 48 h. The solvent was evaporated. The dark-
green solid was then thoroughly washed with diethyl ether and
dried under vacuum conditions. At this stage, the diethyl ether
fractions were collected and evaporated to dryness. The 1H
NMR spectrum of this colorless remaining material (ca. 50 mg)
corresponded to pure L2 ligand (superimposable spectroscopic
NMR data). No trace of L2* (modified ligand) could be detected
by mass spectroscopy, the only signal being that of L2 + H+ at
m/z = 563.29. The dark-green material was recrystallized from
CH3CN/Et2O to afford 70 mg of a microcrystalline dark-green
solid, corresponding to a 78% yield based on the formula given
below. Spectroscopic data are detailed in the text and shown in
the Supporting Information.
3
3
on a Nonius Kappa-CCD area detector diffractometer (Mo KR,
λ = 0.71073 A). Quantitative data were obtained at 173 K for all
˚
complexes. The complete conditions of data collection (Denzo
software) and structure refinements are given in the Supporting
Information. The cell parameters were determined from reflec-
tions taken from one set of 10 frames (1.0° steps in φ angle), each
at 20 s exposure. The structures were solved using direct
methods (SIR97) and refined against F2 using the SHELXL97
software (Kappa CCD Operation Manual; Nonius B.V.: Delft,
The Netherlands, 1997. Sheldrick, G. M. SHELXL97; Univer-
::
::
sity of Gottingen: Gottingen, Germany, 1997). The absorption
was not corrected. All non-hydrogen atoms were refined aniso-
tropically. Hydrogen atoms were generated according to stereo-
chemistry and refined using a riding model in SHELXL97.
Crystal Data for [L1FeCl2]2O CH3CN. Red crystals,
monoclinic, space group C 2/c, a = 34.5912(13) A, b =
3
˚
˚
˚
12.9475(6) A, c = 12.7992(4) A, β = 103.575(2), V = 5572.2
(4) A , Dcalcd = 1.436 g cm-3, Z = 4. For 8083 unique observed
3
˚
reflections with I > 2σ(I) and 351 parameters, the discrepancy
indices are R = 0.0915 and Rw = 0.2469.
Crystal Data for L2FeCl2. Yellow crystals, monoclinic, space
˚
˚
˚
group P21/c, a = 9.004(5) A, b = 28.319(5) A, c = 12.962(5) A, β
= 104.882(5), V = 3194(2) A , Dcalcd = 1.434 g cm-3, Z = 4. For
3
˚
9282 unique observed reflections with I > 2σ(I) and 406 para-
meters, the discrepancy indices are R = 0.0445 and Rw = 0.0897.
Crystal Data for L3FeCl2 THF. Yellow crystals, orthor-
3
˚
hombic, space group Pbca, a = 15.7653 (3) A, b = 22.9257 (3)
˚
A, c = 2426.43 (4) A, V = 8769.9 (2) A , Dcalcd = 1.305 g cm-3
,
3
˚
˚
Z = 4. For 11669 unique observed reflections with I > 2σ(I) and
523 parameters, the discrepancy indices are R = 0.0584 and Rw
= 0.1778.
Crystal Data for L2*Fe2Cl3 OEt2. Yellow crystals,
3
˚
orthorhombic, space group P212121, a = 12.397 (3) A, b =
Elem anal. calcd for C33H32N4O5Cl4Fe2, %: C, 48.45; H,
3.94; N, 6.85; Fe, 13.65; Cl, 17.33. Found, %: C, 48.09; H, 4.33;
N, 6.66; Fe, 13.01; Cl, 16.81.
3
˚
˚
˚
14.716(3) A, c = 20.345(5) A, V = 3711.6(15) A , Dcalcd = 1.437
g cm-3, Z = 4. For 8399 unique observed reflections with I >
2σ(I) and 431 parameters, the discrepancy indices are R =
0.0608 and Rw = 0.1477.
MSES+: m/z = 638.1325, C33H31ClFeN4O+4 . Molecular
conductimetryvalueat c = 5. 10-3 mol L-1: Λ = 86 S mol-1 cm2.
The addition of dilute Et3N to a solution of this compound in
CH3CN resulted in the loss of the ligand-to-metal charge transfer
(LMCT) absorption and a continuous absorption increase be-
tween 580 and 320 nm, in line with the occurrence of a deprotona-
tion process. This observation, together with the absence of any
Results
Ligands L1-3 (see Figure 1) were prepared by the
reaction of 2,3-dimethoxyphenylboronic acid with the

4780 Inorg. Chem., Vol. 48, No. 11, 2009
Benhamou et al.
Scheme 1. Structure of L1FeCl2 and Its Conversion into the μ-Oxo Dinuclear Complex (L1FeCl2)2O upon Oxygenation
R-bromo-substituted tripods, following the procedure al-
ready published for L2.18 The cyclic voltammetry curves of
all ligands display an irreversible anodic wave at a potential
that remains invariant at Ea ≈ 1.05 V/SCE. Another anodic
wave is observed at Ea ≈ 1.50 V/SCE with variable intensity
patterns along the L1-3 series, assigned to the oxidation of the
methoxy groups.
Structure of the Complexes. The preparation of the
complexes follows the classical procedure, involving the
reaction of 1 equiv of FeCl2 with a slight excess of the
ligand.15 The stable complexes thus obtained are char-
acterized by elemental analyses and in solution by UV-
visible spectroscopy, 1H NMR, cyclic voltammetry, and
molecular conductivity. The molecular structure of
L1FeCl2 is deduced from spectroscopic data and electro-
chemical measurements, including conductimetry. In the
solid state, X-ray diffraction structures have been
obtained for L2FeCl2 and L3FeCl2.
L1FeCl2. The presence of a distinguishable MLCT
absorption at λ = 376 nm (ε = 1.0 103 mmol-1 cm2)
together with a typical 1H NMR paramagnetic spectrum
supports tetradentate coordination of the tripod in this
complex.15,21,22 At a concentration of 4 mM, the mole-
cular conductivity value is Λ = 46 Scm2 mol-1, indicating
a neutral electrolytic behavior. In cyclic voltammetry, a
single wave is observed at E1/2 = 184 mV/SCE, with a
Figure 2. Mercury drawing of complex L2FeCl2. Principal distances (in
˚
A) and angles (in deg): FeN1, 2.1675(19); FeN2, 2.224(2); FeN3, 2.266(2);
FeCl1, 2.3248(10); FeCl2, 2.2656(10); N2FeN3, 154.65(7); N1FeCl1,
100.11(5); N1FeCl2, 122.09(5); Cl2FeCl1, 137.56(3).
with another pseudo-reversible wave at E1/2 = 340 mV/
SCE, assigned to the presence of the dissociated mono-
cation [L2FeCl(MeCN)]+, Cl- in equilibrium under the
conditions of the measurement.21
pseudoreversible pattern, ΔE = 84 mV and Ipc/Ipa
=
0.80, in line with the presence of a distorted Oh geometry.
The 1H NMR spectrum displays typical patterns (chemi-
cal shifts and mid-intensity width) of distorted Oh high-
spin FeCl2 complexes with monosubstituted tripods. The
trace is given in the Supporting Information. We thus
believe that the metal center lies in a distorted octahedral
geometry, in solution and in the solid state as well, as
drawn in Scheme 1.
The X-ray crystal structure of L2FeCl2 could be solved,
and a Mercury drawing is displayed in Figure 2. In
L2FeCl2, the tripod acts as a tridentate ligand, one of
the substituted pyridines being remote from the coordi-
nation center.
A five-coordinate metal center generally exhibits
square-pyramidal or trigonal-bipyramidal geometries.23
In L2FeCl2, the calculated trigonality index is τ = 0.28. In
this structure, all metal-to-ligand distances lie above 2.0
L2FeCl2. The 1H NMR spectrum is very similar to that
of the complex with the parent [bis(2-phenyl 6-pyridyl-
methyl) (2-pyridylmethyl)]amine].15 The main difference
lies in the presence of signals from the methoxy groups at
δ = 15 ppm. The weak MLCT absorption at λ = 368 nm
(ε = 0.75 103 mmol-1 cm2) together with the 4 mM
molecular conductivity value of Λ = 47 Scm2 mol-1
indicate a five-coordinate geometry in this complex. All
spectroscopic data are given in the Supporting Informa-
tion. The cyclic voltammogram of this complex consists
of one reversible wave at E1/2 = 8 mV/SCE standing for
L2FeCl2, with ΔE = 80 mV and Ipc/Ipa = 0.95, together
˚
A, in line with a high-spin state for the metal center.
L3FeCl2. The H NMR spectrum is not informative,
1
with a prominent signal emerging at δ ≈ 0.5 ppm over a
broad envelope lying between +10 and -5 ppm. The
absence of any well-defined MLCT absorption in the
UV-visible region is reminiscent of the spectra observed
in complexes with ligands in which all three R positions
are substituted.16,21 The 4 mM molecular conductivity
value of Λ = 23 Scm2 mol-1 indicates a neutral electro-
lytic behavior of the complex. The cyclic voltammogram
of this complex consists of only one reversible wave at E1/2
= 9 mV/SCE with ΔE = 100 mV and Ipc/Ipa = 0.90.
(21) Benhamou, L.; Lachkar, M.; Mandon, D.; Welter, R. Dalton Trans.
2008, 6996–7003.
(22) Machkour, A.; Mandon, D.; Lachkar, M.; Welter, R. Inorg. Chim.
(23) Addison, A. W.; Nageswara, R.; Reedijk, J.; Van Rijn, J.; Verschoor,
Acta 2005, 358, 839–843.
G. C. J. Chem. Soc., Dalton Trans. 1984, 1349–1355.

Article
Inorg. Chem., Vol. 48, No. 11, 2009
4781
Figure 5. Mercury view of complex [L1FeCl2]2O. Principal distances (in
˚
A) and angles (in deg): Fe1N1, 2.387(4); Fe1N2, 2.172(5); Fe1N3, 2.208
(5); Fe1Cl1, 2.3361(17); Fe1Cl2, 2.3580(17); Fe1O1, 1.7885(7);
Fe1O1Fe1, 175.9(4). Distance Fe1-Fe1: 3.575 (7). Torsion angles
Cl1Fe1Fe1Cl2 = Cl2Fe1Fe1Cl1: 124.7°.
˚
Figure 3. Mercury view of complex L3FeCl2. Principal distances (in A)
and angles (in deg): FeN1, 2.183(2); FeN2, 2.235(2); FeN3, 2.205(2);
FeCl1, 2.2930(8); FeCl2, 2.3576(7); N2FeN3, 156.89(8); N1FeCl1, 108.62
(6); N1FeCl2, 98.66(6); Cl2FeCl1, 152.61(3).
Figure 6. Spectral absorption changes versus time in the course of
oxygenation of complex L2FeCl2. One scan every 60 min over 36 h. Inset:
the 450-820 nm range, showing the appearance of the LMCT absorption
at λ = 637 nm.
coordinates in the tridentate mode, the substituted pyr-
idine being remote from the metal center. The geometry
around the metal is pseudo-octahedral, with two chloride
atoms and one oxo bridge completing the coordination
sphere. As a consequence, (L1FeCl2)2O is a neutral
molecule. A mercury diagram is displayed in Figure 5.
The structural parameters are close to those already
reported for the isostructural (FTPAFeCl2)2-O,17 with d
Figure 4. Spectral absorption changes versus time in the course of
oxygenation of complex L1FeCl2 over the first 16 h (1 scan/h). From
the 16th hour, the reaction goeson, but precipitation occurs in the cuvette,
yielding poorly tractable data, not shown on this graph.
Single crystals could be obtained for L3FeCl2, and the
Mercury diagram is displayed in Figure 3.
Here again, the tripod coordinates in the tridentate
mode with one dangling substituent. The distortion
versus ideal trigonal-bipyramidal is important, with
τ = 0.07, although the observed geometry is not planar,
as seen from Figure 3 (inset). This point will be discussed
later. In this complex also, all metal-to-ligand distances lie
˚
(Fe1-Fe2) = 3.57 A and <Fe1OFe2 = 175.9°.
L3FeCl2. Solutions of this complex in acetonitrile or in
THF could be handled under atmospheric conditions for
several hours without alteration. Bubbling dry dioxygen
in the solution resulted in the progressive appearance of
brownish green shades that became distinguishable only
over one day of stirring. Monitoring the absorbance by
UV-visible spectroscopy over 60 h allowed for the
detection of a 10-15% increase in absorbance around
350 nm. No spectroscopic modification was observed in
the 400-820 nm range. The data are shown in the
Supporting Information.
L2FeCl2. Upon exposure to dry dioxygen, the yellow
solution of the complex turned deep green. UV-visible
spectroscopy allowed progressive detection of a broad
absorption at λ = 637 nm (ε = 0.41 103 mmol-1 cm2,
based on the initial concentration of L2FeCl2), similar to
the LMCT ArO f Fe(III) band observed in a catecholato
complex,18 as shown in Figure 6. Noteworthy is the fact
that no tractable absorption could be detected within the
˚
above 2.0 A, in line with a high-spin state for the metal
center.
Reactivity with O2. L1FeCl2. When exposed to dry
dioxygen, the color of this compound gradually turned
to brownish green. Monitoring the oxygenation by UV-
vis allowed for the detection of signals at λ = 341 and 384
nm, typical of μ-oxo diferric complexes. The reaction goes
slowly, and a poorly soluble compound is obtained. At a
preparative scale, the reaction is complete over 40-48 h
after the addition of dioxygen. Figure 4 shows the spectral
absorption changes versus time in the course of oxygena-
tion of complex L1FeCl2.
Single crystals of (L1FeCl2)2O could be obtained. As
expected from the UV-vis measurements, the com-
pound displays the μ-oxo diferric structure. The ligand

4782 Inorg. Chem., Vol. 48, No. 11, 2009
Benhamou et al.
300-400 nm range. In contrast with L3FeCl2, the change
in color became obvious over 1 to 2 h. The reaction was
considered as complete over 30-36 h.
Cyclic voltammetry experiments on the green com-
pound afforded two reduction waves: a reversible one
was observed at E1/2 = -80 mV/SCE with ΔE = 80 mV
and Ipc/Ipa = 1.0; the second irreversible wave was
detected at Ec = -500 mV/SCE.
Full reduction of the green complex with zinc amalgam
under anaerobic conditions in a THF or CH3CN solution
afforded, over 10 min, a bright orange solution, which
turned out to be extremely oxygen-sensitive.
This reaction was carried out on a preparative scale,
and a bright orange solid could be isolated after treatment
with an 83% yield. The 1H NMR spectrum of this
complex displays paramagnetically shifted signals within
the -40 to +150 ppm range (the spectrum is displayed in
the Supporting Information). We later could obtain
single crystals of this compound and checked that the
data obtained from the bulk synthesis were identical to
those recorded from hand-picked single crystals. Thus,
this reduction reaction turned out to be clean. The strik-
ing point is the extreme oxygen sensitivity of this com-
pound which, even in the solid state, becomes dark within
a few seconds.
The UV-visible spectrum of this new species, shown in
Figure 7, displayed no characteristic feature in the 500-
800 nm region. A new absorption at λ = 377 nm (ε = 1.39
 103 mmol-1 cm2) is present, reminiscent of that ob-
served in the catecholato complex already mentioned
above.18
The color, spectroscopic features, and extreme sensi-
tivity to dioxygen strongly suggested coordination of a
phenolate derivative to Fe(II).
This was confirmed by the molecular structure of this
yellow-orange compound, as single crystals of the com-
pound termed L2*Fe2Cl3 could be obtained. A Mercury
diagram is displayed in Figure 8.
This new complex is a dinuclear di-Fe(II) compound,
with an asymmetric structure: only one ligand is present
for two iron atoms. As expected, the L2 ligand in the
original ferrous complex has undergone O-demethyla-
tion, becoming L2*. In L2*Fe2Cl3, one iron atom lies in
a pseudo-octahedral geometry, being coordinated by the
four nitrogen atoms from the ligand, one bridging
Monitoring the oxygenation reaction by 1H NMR
resulted in a progressive loss of the signal of the Fe(II)
complex, no new tractable species being detected in the
paramagnetic region. Release of the free ligand, however,
was observed in the diamagnetic area. In order to make
sure that these spectroscopic modifications were not the
result of any partial degradation process, we ran this
reaction on the preparative scale. After workup, a dark
green microcrystalline solid could be isolated, with a 78%
yield based on the C33H32Cl4Fe2N4O5 formulation (vide
infra). On the other hand, almost 50% (based on the
amount of starting complex) of free and unmodified
ligand L2 was recovered in the washing phases. Mass
spectroscopy analysis of the green compound allowed
detection of the [L2*FeCl]+ cation at m/z = 638.15 as the
main signal, with L2* being a modified ligand in which
one methyl group has been lost. However, the description
of this green complex as a ferric mononuclear compound
does not correspond to the results of the elemental
analysis. The molar element ratios indicate the presence
of a dinuclear asymmetric complex (L2*/Fe/Cl = 1:2:4).
The found elemental percentages (%), C = 48.09, H =
4.33, N = 6.66, Fe = 13.01, and Cl = 16.8, match those
obtained from the C33H32Cl4Fe2N4O5 formulation with
the calculated values (%) C = 48.45, H = 3.94, N = 6.85,
Fe = 13.65, and Cl = 17.33. The molecular conductivity
indicates that the compound behaves as a monocharged
species with Λ = 86 S mol-1 cm2 at a concentration of
5 Â 10-3 mol L-1. Thus, two structural hypotheses can be
drawn, [L2*Fe-Cl-FeCl3]OH (A) or [L2*Fe-OH-
FeCl3]Cl (B). The IR spectrum of the green compound
(shown in the Supporting Information) points out the
lack of vibration assigned to “free” OH-. By contrast, a
very weak signal is observed at 3650 cm-1 which might fit
with the presence of a bridging OH group. Furthermore,
our green compound is sensitive to basic media, reacting
with Et3N. Considered together, these observations sup-
port structure B, [L2*Fe-OH-FeCl3]Cl, as the most
likely hypothesis.
Figure 7. Spectral absorption changes upon Zn/Hg reduction of a 1.4 Â 10-3 M solution of the green compound over 10 min.

Article
Inorg. Chem., Vol. 48, No. 11, 2009
4783
as those used for electrochemical measurements.21 The
presence of two waves with L2FeCl2 is probably due to an
equilibrium between the L2FeCl2 form (low positive
potential, E1/2 = 8 mV/SCE)) and the dissociated mono-
cation [L2FeCl(MeCN)]+, Cl- form (high positive po-
tential, E1/2 = 340 mV/SCE). The single wave observed
for L3FeCl2 at a low positive potential (E1/2 = 9 mV/
SCE) strongly suggests that only one species is present
and that the structure observed in the solid state is
completely retained in the electrochemical solution. In
these complexes, the reason for a possible equilibrium
(L2FeCl2) or not (L3FeCl2) might be related to the effect
of aromatic susbtituents close to the metal (vide infra).
Anyway, the low positive potential values are consistent
with a higher electron density at the metal center, in line
with hypodentate coordination of the ligand. Thus, even
indirectly, electrochemical measurements can be useful
tools for geometry predictions.21
Structures of L2FeCl2 and L3FeCl2. Hypoden-
tate24 coordination of various tripods within the TPA
family25,26 to iron is now well-known, leading to com-
plexes in which the metal lies in a trigonal-bipyramidal
environment,15,16,27,28 or square-pyramidal in a particu-
lar case.30 Tridentate coordination of the ligand is found
in L2FeCl2 and L3FeCl2. The values of the τ index suggest
square-pyramidal geometries rather than trigonal-bipyr-
amidal. However, the trans-pyridyl angles (axial direction
for determination of τ) open in the wrong direction, with
relatively invariant values typically in the 152-157° range
within the series of isotructural characterized com-
plexes.15,16,21 This obviously moderates the significance
of the τ parameter in our case. The value of the <Cl1FeCl2
angle seems to be more indicative of distortions.
L2FeCl2. In this complex, <Cl1FeCl2 = 137.6°, that is,
the average of the values found within the series of
isostructural complexes.15,16,21,22 It is noteworthy that a
repulsion occurs between the Cl2 chloride and the o-OMe
group of the substituent, as can be seen in Figure 9 (left).
As a consequence, the structure is open, providing full
lateral access to the metal center.
L3FeCl2. In this complex, the value of <Cl1FeCl2 =
152.6° corresponds to the most open angle found for such
complexes. A careful examination of the structure pro-
vides rationale for this: both chloride ligands lie in a
position where they interact with hydrogen atoms from
the methoxy groups and methylene, as shown in Figure 9
(right).
Figure 8. Mercury view of complex L2*Fe2Cl3. Principal distances (in
˚
A) and angles (in deg): Fe1N1, 2.238(5); Fe1N2, 2.189(5); Fe1N3, 2.265
(5);Fe1N4, 2.139(5);Fe1Cl3, 2.4546(17); Fe1O4, 1.983(4); Fe2Cl3, 2.4778
(17); Fe2O4, 2.074(4); Fe2O3, 2.230(4); Fe2Cl2, 2.313(2); Fe2Cl1, 2.296
(2); N4Fe1Cl3, 174.45(14); O4Fe1Cl3, 85.97(11); Fe1Cl3Fe2, 83.09(5);
Cl3Fe2O4, 83.47(11); Fe2O4Fe1, 107.45(17). Distance Fe1-Fe2: 3.274
(7).
chloride atom, and a bridging phenolato oxygen atom
that originates from the bis(methoxy)phenyl arm of the
˚
ligand. All Fe-N distances lie above 2.1 A; the phenolate
˚
is tightly bound with d(Fe1-O4) = 1.983 A. The bridging
˚
chloride lies at d(Fe1-Cl3) = 2.455 A. The second iron
atom is coordinated by the bridging chloride with d(Fe2-
Cl3) = 2.478 A, the bridging phenolato oxygen atom with
˚
˚
d(Fe2-O4) = 2.074 A, one oxygen atom from a methoxy
˚
group with d(Fe2-O3) = 2.229 A, and two terminal
˚
chloride ligands with d(Fe2-Cl2) = 2.313 A and d(Fe2-
Cl1) = 2.296 A. This five-coordinated environment
˚
provides the metal center with a severely distorted
square-pyramidal environment, with τ = 0.31. The
˚
metal-to-metal distance is 3.271 A.
Discussion
Structures of the Complexes. L1FeCl2. The spectro-
scopic features that characterize distorted octahedral
geometry for similar complexes in solution are now
well-known. For this complex, all spectroscopic data
support the occurrence of such a geometry: pronounced
MLCT absorption, 1H NMR data with sharp paramag-
netic signals, and neutral electrolytic behavior in solu-
tion.15,17,21,22 Obviously, in the present case, as in other
similar complexes,19 the steric hindrance provided by a
single aromatic substitution remains weak, not being
strong enough to induce decoordination of the substi-
tuted pyridine. Thus, the metal center lies in the classical
geometry usually encountered for Fe(II) complexes.
Moreover, the E1/2 = 184 mV value found in electro-
chemistry lies, as expected, slightly below that measured
from the distorted Oh parent TPAFeCl2 complex.21
Solution Studies of L2FeCl2 and L3FeCl2. All spectro-
scopic data support the presence of a five-coordinate
species in solution.15,16,21 The 1H NMR spectrum of
L2FeCl2 is very similar to that obtained from
[(C6H5)2TPA]FeCl2 that exhibits a trigonal bipyramidal
geometry in the solid state and in solution.15 We recently
observed that the FeCl2 complex with the Me3TPA ligand
is cleanly converted into the bis-acetonitrile dication
when dissolved in a strongly dissociative medium, such
Correlatively, the two mean planes of the adjacent bis
(methoxy)phenyl substituents are not coplanar, since the
corresponding angle value is 62.3°. These aromatic arms
define a narrow interstice leading to the metal, the short-
est distance between these two aromatic substituents
˚
being d(C25-C11) = 4.26 A. Under these conditions,
and with a 5.59 A intercentroid distance, only a weak
˚
(24) Blackman, A. G. C. R. Chim. 2004, 8, 107–119.
(25) Blackman, A. G. Polyhedron 2005, 24, 1–39.
(26) Blackman, A. G. Eur. J. Inorg. Chem. 2008, 2633–2647.
(27) Diebold, A.; Hagen, K. S. Inorg. Chem. 1998, 37, 215–223.
(28) Jo, D.-H.; Chiou, Y.-M.; Que, L.Jr. Inorg. Chem. 2001, 40, 3181–
3190.
(29) Schueneman, V.; Jung, C.; Trautwein, A. X.; Weiss, R.; Mandon, D.
FEBS Lett. 2000, 479, 149–154.
(30) Thallaj, N. K.; Machkour, A.; Mandon, D.; Welter, R. New J. Chem.
2005, 29, 1555–1558.

4784 Inorg. Chem., Vol. 48, No. 11, 2009
Benhamou et al.
Figure 9. (Left)Structure of L2FeCl2: lateral view along the Cl1FeCl2 plane, showing access tothe metal center. This structure is muchmore open thenthat
˚
of L3FeCl2, providing easy access to the metal center. dCl2-H2 = 3.50 A; dCl2-H1 = 3.49 A; dCl1-H22 = 2.65 A; dCl1-H9 = 3.21 A; dCl1-H16 = 3.11 A. (Right)
˚
˚
˚
˚
Structure of L3FeCl2: lateral view along the Cl1FeCl2 plane, showing access to the metal center. Short contacts between the chloride ligands and neighboring
˚
˚
˚
˚
˚
H atoms in L3FeCl2: dCl1-H29B = 3.33 A; dCl1-H29A = 3.13 A; dCl1-H27A = 2.76 A; dCl1-H13A = 2.93 A; dCl2-H1B = 3.07 A; dCl2-H15A = 2.93 A; dCl2-H12
˚
˚
= 2.65 A; dCl2-H26 = 2.95 A.
˚
potential π-stacking effect can be suspected.31 As drawn
in Figure 9 (right), the global picture is that of a tweezer
with two converging arms kept in position by hydrogen
bonds, with the net effect of opening the <Cl1FeCl2
angle, while limiting the access to the metal center. This
is in striking contrast with the situation in L2FeCl2.
Reactivity versus Dioxygen: General Considerations.
L1FeCl2. FeCl2 complexes react with dioxygen much
better, as the metal center exhibits Lewis acidity character,
that is, the ligands are electrodeficient.17 The bis(methoxy)
phenyl substituted ligands presented in this study display
anodic potential that is relatively invariant, typically Ea ≈
1.05 mV/SCE, slightly below the reported value of Ea =
1.12 mV/SCE found for the parent TPA ligand.17 The E1/2
oxygenation reaction takes place via coordination of O2
to the metal center.17 The fact that a five-coordinate
environment accelerates the reaction is well-known and
was experimentally demonstrated.16,17 We thus have here
an exception. However, comparison of the two five-
coordinate structures provides the clue to the unexpected
lack of sensitivity of L3FeCl2. From Figure 9 (left), it
seems obvious that the access of exogenous ligands to the
metal center is open in L2FeCl2. The reaction goes slowly,
but it works. Figure 9 (right) displays the lateral view,
showing the access to the iron center in L3FeCl2. As
already stated above, the particular orientation of the
two bis(methoxy)phenyl groups is not due to chance but
results from stabilization because of methoxy/chloride
interaction. The result is that the converging aromatic
arms delineate a sterically hindered nonpolar environ-
ment. As a consequence, access to the iron center is
dramatically restricted, even for a small molecule such
as O2. In biological systems, controlling access of small
molecules to active sites by the proteic environment is an
obvious fact. In large synthetic systems, an elegant de-
monstration has recently been reported: a reduction in
activity versus O2 of synthetic non-heme iron sites is
observed when those are buried in dendrimers.32 Far
below in size and structure, the case of L3FeCl2 illustrates
such a principle. We postulate that the poor reactivity of
this complex is simply due to overprotected access to the
metal center by its particular aromatic environment.
O-Demethylation in L2FeCl2. μ-Oxo diferric com-
pounds generally display a brownish color. Thus, the
appearance of a dark green color upon oxygenation of
L2FeCl2 was surprising and strongly suggested the pre-
sence of a coordinated phenolate at the ferric site. We
obviously checked that what was detected in spectroscopy
corresponded indeed to a major reaction easily being
carried out on the preparative scale. In order to elucidate
the structure of the dark green reaction product, we tried
for a long time to grow single crystals, without success.
= 186 mV value found for L1FeCl2 neighbors the E1/2
=
195 mV found in the parent TPAFeCl2 complex. There-
fore, the observed smooth oxygenation process is not
surprising, although careful UV-visible monitoring is
not possible because the end product is poorly soluble.
We estimate however that the reaction is complete over
40-48 h, a reaction time slightly longer than that pre-
viously reported for the parent TPAFeCl2 complex. The
μ-oxo diferric complex obtained displays a structure simi-
lar to that obtained upon oxygenation of the complex with
a monofluoro-substituted ligand in which the substituted
arm of the tripod is uncoordinated.17
L2FeCl2. This complex reacts smoothly with O2, the
conversion rate reaching 90% over 30 h. This is consider-
ably slower than what is observed with the isostructural
F3TPAFeCl2 complex for which immediate conversion
was reported.16 This might reflect the differences in the
electronic properties of the corresponding ligands, since
Ea = 1.28 mV/SCE for F3TPA, that is, ≈250 mV higher
than L2. Thus, L2FeCl2, being more electron-rich than the
isostructural complex with a fluorinated ligand, reacts
slower than the latter.
L3FeCl2. The surprise came from the very low sensitiv-
ity of this complex, for which a reactivity similar to that of
L2FeCl2 was expected. We already postulated that the
(32) Zhao, M; Helms, B.; Slonkina, E.; Friedle, S.; Lee, D.; DuBois, J.;
ꢀ
Hedman, B.; Hodgson, K. O.; Frechet, J. M. J.; Lippard, S. J. Am. Chem.
(31) Spek, A. L. J. Appl. Cryst. 2003, 36, 7–13.
Soc. 2008, 130, 4352–4363.

Article
Inorg. Chem., Vol. 48, No. 11, 2009
4785
a
Scheme 2. The Two-Step Conversion of L2FeCl2 into L2*FeCl3
a The total yield of the reaction is 64% relative to the starting Fe(II) complex.
1
The H NMR spectrum being featureless, we ran mass
spectroscopic methods, and keeping in mind the sensiti-
vity of the compound toward bases, it seems that the
structure [L2*Fe-OH-FeCl3]Cl represents the mostlikely
hypothesis. The unquestionable point remains that
O-demethylation of the ligand has occurred upon the
reaction of dioxygen with the metal center in L2FeCl2.
In the chemistry of cytochromes P-450, O-demethyla-
tion is proposed to occur via reaction of the methoxy
group with the oxygen-active species (a formal Fe(V)dO
center, in fact an Fe(IV)dO/protein radical transient29),
although the way demethylation occurs is not known
yet.8-14 In the present case, oxygenated species (an oxy
form such as Fe(II)O2, and perhaps an oxo form such as
Fe(IV)dO) of iron complexes have to be present in the
medium.15,17 If one considers the regiospecificity of the
reaction, that is, the fact that only the ortho OMe group
of the substituent is O-demethylated, the occurrence of a
process involving radical scrambling in solution seems
very unlikely. By contrast, the presence of the chloride-
methoxy interaction in the crystal structure of L3FeCl2
validates the hypothesis of an interaction between the
methoxy group and any oxygenated form of the metal
center. In the present case, the oxygen-adduct FeII(O2) S
spectroscopy analyses and could obtain a signal corre-
sponding to the monomeric form [L2*FeCl]+, L2* being
the O-demethylated form of L2. Ferric mononuclear
complexes of similar structure exist: we reported in the
past the structure of an analogue compound in which the
iron center is bound to a catecholate moiety.18 In the
present case, however, the elemental analysis was in
contradiction with the occurrence of a mononuclear
structure. The results strongly suggested that the
observed [L2*FeCl]+ in mass spectrometry was the result
of a fragmentation. Indeed, the elemental analysis data
supported the presence of a dinuclear asymmetric
complex (one ligand for two metals), as it is generally
the case for oxygenation products with sterically hindered
ligands.16,17 Additionally, the fact that around 50% of the
total ligand was released as a free tripod upon oxygena-
tion also brought credit to the presence of an asymmetric
complex.
We found in the past that using simple redox reactions
can lead to indirect characterization of complexes.18 At
this point, we examined the cyclic voltammetry data of
the green compound, exhibiting two waves at low and
moderate negative potentials, the first being fully rever-
sible. With hope that the structure would not be con-
siderably modified upon full reduction (the reduction
potential of Zn/Hg lies far below the second irreversible
reduction wave of our complex33), we reacted the dark
green compound with zinc amalgam. This quickly af-
forded a bright orange complex with an excellent yield,
which was revealed to be extremely sensitive to dioxygen.
Its UV spectrum strongly suggested the presence of a
phenolate coordinated to a ferrous center.18 The evidence
for one O-demethylation of the ligand came from exam-
ination of the crystal structure, which also confirmed the
presence of an asymmetric dinuclear species. The two-
step transformation of L2FeCl2 into L2*FeCl3 is dis-
played in Scheme 2.
FeIII-O-O might promote hydroxylation of the
3
adjacent methoxy group with a release of formalde-
hyde and concomitant formation of an FeIVdO
species. Upon reaction with the starting complex, this
high-valent species would form an asymmetric FeIII-O-
FeIII with a pendant phenol arm, and there would be a
release of the free unmodified ligand. Subsequent rear-
rangement would yield a FeIII(PhO)(OH)FeIII asym-
metric complex with phenolato and hydroxo bridging
ligands.
Why such a reaction with L2FeCl2 and not L1FeCl2, for
which the formation of a standard μ-oxo species is
observed? It seems likely that, in the case of L2FeCl2, at
least one substituted pyridine remains coordinated during
the oxygenation process, forcing the methoxy group to lie
in close proximity to the active metal center. This is
actually a prerequisite for the reaction to occur. This is
not necessarily the case with a single-substituted ligand in
which some flexibility might lead to temporary decoordi-
nation of the more bulky substituted arm.20 Again, this
finding supports a metal-centered process rather than a
free radical-based reaction.
We recently showed that, upon reduction with Zn/Hg,
μ-oxo di-ferric complexes convert back to the ferrous
mononuclear complexes.17 Thus, if any μ-oxo (or
μ-hydroxo) bridge was present in the dark green com-
pound, it would probably have been removed upon reduc-
tion. This is the reason why we cannot directly extrapo-
late the structure of L2*Fe2Cl3 to describe that of the green
precursor. However, on the basis of analytical and
Concluding Remarks
(33) Bard, A. J.; Faulkner, L. R. In Electrochemical Methods. Funda-
mentals and Applications; John Wiley and Sons, Inc. Publishers: New York,
1980.
The formation of μ-oxo diferric complexes from molecular
dioxygen and mononuclear Fe(II) complexes is currently

4786 Inorg. Chem., Vol. 48, No. 11, 2009
Benhamou et al.
emerging as a general process within nonporphyrinic che-
16-electron complex with a five-coordinate environment
might turn out to react slowly (L2FeCl2) or very poorly
(L3FeCl2).
16,17,34-36
mistry.
Kinetic and electrochemical studies
strongly support the formation of some dioxygen-metal
adduct, although the structure of such a species at a single
metal site has never been reported in nonporphyrinic chem-
istry.3,17,34,37,38 There are a few reports of intramolecular
transformations of a ligand upon oxygenation (rearrange-
ments, N-dealkylation, oxidative dehydrogenation, or intra-
molecular hydroxylations to cite a few examples39-45);
however, O-demethylation is in general poorly mentioned,
and when it is, it is as a side and nondesirable reaction.11,12
The present study brings to light two particular comments.
The first message of the present work is to demonstrate
that, even when the geometry at the metal site seems
extremely favorable to dioxygen coordination, the process
is controlled by electronic and structural factors: thus, a
The second outcome is the occurrence of a regiospecific
intramolecular O-demethylation of the ligand that occurs
when methoxy groups are kept in place, close to the metal
center. This is one of the first examples known in homo-
geneous nonheme chemistry, which parallels a well-known
reaction in P-450 chemistry. Does it mean that an oxo-ferryl
FedO moiety with strong oxidation abilities is formed upon
oxygenation of our complex? Although unlikely, we cannot
rule out this hypothesis. Or inversely does it mean that, even
in the case of the P-450s, O-demethylation of substrates;
kept in place at the distal side before the oxygenation;does
not necessary need the presence of a high-valent oxo-ferryl
FedO center but simply an oxygenated form (whatever its
nature may be) of the metal center? Even if we tend to favor
this hypothesis inthe present case, the question remains open.
In any event, experimental evidence leads to the conclusion
that coordination of dioxygen close to a judiciously oriented
methoxy group is a minimum requirement.
(34) Korendovytch, I. V.; Kryatova, O. P.; Reiff, W. M.; Rybak-Akimova,
E. V. Inorg. Chem. 2007, 46, 4197–4211.
(35) Zart, M. K.; Powell, D.; Borovik, A. S. Inorg. Chim. Acta 2007, 360,
2397–2402.
::
(36) Dutta, S. K.; Ghosh, M.; Biswas, P.; Florke, U.; Saal, C.; Haase, X.
X.; Nag, K. New J. Chem. 2007, 31, 93–101.
(37) Carvalho, N. M. F.; Antunes, O. A. C.; Horn, A.Jr. Dalton Trans.
Acknowledgment. This work was supported as a
collaborative action between CNRS (project no.
18552;France) and CNRST (project no. chimie 08/06
and 08/07;Morocco). Both organizations are gratefully
acknowledged for their support. We also wish to thank
Dr. Remy Louis, head of the Institut de Chimie in
Strasbourg, for constant support and encouragement,
and the CNRS and ULP. The Conseil Scientifique de
l’ULP is acknowledged for specific support, no. AO CS
ULP 2006.
2007, 1023–1027.
(38) May, Z.; Simandi, L. I.; Vertes, A. J. Mol. Catal. A 2007, 266, 239–
248.
(39) Nehru, K.; Seo, M. S.; Kim, J.; Nam, W. Inorg. Chem. 2007, 46, 293–
298.
(40) Celenligil-Cetin, R.; Paraskevopoulou, P.; Dinda, R.; Staples, R. J.;
Sinn, E.; Rath, N. P.; Stavropoulos, P. Inorg. Chem. 2008, 47, 1165–1172.
(41) Celenligil-Cetin, R.; Paraskevopoulou, P.; Dinda, R.; Lalioti, N.;
Sanakis, Y.; Rawashdeh, A. M.; Staples, R. J.; Sinn, E.; Stavropoulos, P.
Eur. J. Inorg. Chem. 2008, 673–677.
(42) Machkour, A.; Mandon, D.; Lachkar, M.; Welter, R. Eur. J. Inorg.
Chem. 2005, 158–161.
(43) Menage, S.; Galey, J. B.; Dumats, J.; Hussler, G.; Seite, M.; Gautier-
Luneau, I.; Chottard, G.; Fontecave, M. J. Am. Chem. Soc. 1998, 120,
13370–13382.
(44) Lange, S. J.; Miyake, H.; Que, L.Jr. J. Am. Chem. Soc. 1999, 121,
6330–6331.
(45) Yamashita, M.; Fururtachi, H.; Tosha, T.; Fujinami, S.; Saito, W.;
Maeda, Y.; Takahashi, K.; Tanaka, K.; Kitagawa, T.; Suzuki, M. J. Am.
Chem. Soc. 2007, 129, 2–3.
Supporting Information Available: All synthetic details in-
cluding the preparation and characterization of all compounds
mentioned. For the four reported structures, the crystallo-
graphic files in CIF format have been deposited. This material
is available free of charge via the internet at http://pubs.acs.org.