Dinuclear iron complexes based on parallel β-diiminato binding sites: Syntheses, structures and reaction with O2

Article abstract of DOI:10.1039/b715376g

With the background that certain natural systems utilize two Fe ^II centres in their prosthetic groups for the activation of O ₂, a ligand system containing two parallel β-diiminato binding sites linked by a xanthene backbone ([^RXanthdim]^2- with residues R = 2,3-dimethylphenyl and 2,4-difluorophenyl at the iminato units, respectively) was investigated with respect to its Fe^II coordination chemistry in order to study O₂ activation reactions. Hence, the corresponding lithium salts were treated with FeCl₂ to yield the complexes [^Me2C6H3Xanthdim]Fe₂Cl₃(Li(thf) ₃), 1 and [^F2C6H3Xanthdim]Fe₂Cl ₃(Li(thf)₃), 2, respectively, each of which comprises Cl-Fe(μ-Cl)Fe-Cl(Li(thf)₃) units. 1 and 2 indeed readily react with O₂ to give the oxides [^RXanthdim]Fe ₂Cl₂O containing Fe^III-O-Fe^III moieties. Due to the electron withdrawing F atoms 2 reacts more slowly than 1. The molecular structures of 1, 2 and [^Me2C6H3Xanthdim]Fe ₂Cl₂O, 3, as determined by single crystal X-ray diffraction are discussed, and an investigation concerning their magnetic properties revealed an antiferromagnetic coupling of the two iron centres in all complexes; naturally, the strongest coupling is observed for 3. The Royal Society of Chemistry.

Full text of DOI:10.1039/b715376g

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Dinuclear iron complexes based on parallel b-diiminato binding sites:
syntheses, structures and reaction with O2†
Maurice Frederic Pilz,a Christian Limberg,*a Serhiy Demeshko,b Franc Meyerb and Burkhard Ziemera
Received 8th October 2007, Accepted 31st January 2008
First published as an Advance Article on the web 4th March 2008
DOI: 10.1039/b715376g
With the background that certain natural systems utilize two FeII centres in their prosthetic groups for
the activation of O2, a ligand system containing two parallel b-diiminato binding sites linked by a
xanthene backbone ([RXanthdim]2− with residues R = 2,3-dimethylphenyl and 2,4-difluorophenyl at the
iminato units, respectively) was investigated with respect to its FeII coordination chemistry in order to
study O2 activation reactions. Hence, the corresponding lithium salts were treated with FeCl2 to yield
the complexes [Me2C6H3Xanthdim]Fe2Cl3(Li(thf)3), 1 and [F2C6H3Xanthdim]Fe2Cl3(Li(thf)3), 2, respectively,
each of which comprises Cl–Fe(l-Cl)Fe–Cl(Li(thf)3) units. 1 and 2 indeed readily react with O2 to give
the oxides [RXanthdim]Fe2Cl2O containing FeIII–O–FeIII moieties. Due to the electron withdrawing F
atoms 2 reacts more slowly than 1. The molecular structures of 1, 2 and [Me2C6H3Xanthdim]Fe2Cl2O, 3, as
determined by single crystal X-ray diffraction are discussed, and an investigation concerning their
magnetic properties revealed an antiferromagnetic coupling of the two iron centres in all complexes;
naturally, the strongest coupling is observed for 3.
Introduction
The b-diiminato ligand system is very versatile and has been
shown to display a very rich coordination chemistry stabilising
high as well as low oxidation states.1 This of course makes it also
ideal as ligand in transition metal complexes that activate small
substrates. For instance, b-diketiminato-ironchloride complexes
proved capable of activating dinitrogen under reducing conditions
(vide infra).2 Furthermore, b-diketiminato-zinc compounds were
shown to catalyse the copolymersation of CO2 and epoxides
which has been shown to work via a mechanism involving two b-
diketiminato-zinc units.3 Last but not least b-diiminato-copper(I)
Chart 1
complexes were synthesised that can bind dioxygen which in the
presence of a further copper(I) centre is activated sufficiently to
In the recent years b-diketiminato iron compounds have mainly
oxygenate the ligand.4 In all these three examples two metal centres
been investigated by P. L. Holland et al. The reaction of lithiated
cooperate in the activation process, a principle that is of course
b-diketimides with FeCl2(thf)1.5 leads to complexes like I8 which
found in many metalloenzymes, too. This has recently stimulated
reacts with KC8 or Na in the presence of dinitrogen to give
a lot of interest in the design of ligands containing two b-diiminato
the complex II where two FeI centres are spanning (and thus
moieties in close proximity.5 We have focussed in particular on a
activating) a dinitrogen molecule (Scheme 1).2,8,9
ligand where two b-diiminato units are oriented parallel to each
other, and recently, we have described the successful development
of a corresponding ligand precursor, [RXanthdim]H2 (Chart 1),6,7
the synthesis of zinc complexes as well as their potential as
epoxide/CO2 copolymerisation catalysts.7 Considering the liter-
ature concerning mononuclear b-diketiminato iron complexes,
further employment of [RXanthdim]2− in iron chemistry suggests
itself.
Scheme 1
aHumboldt-Universita¨t zu Berlin, Institut fu¨r Chemie, Brook-Taylor-Str. 2,
So far, there is only one report of b-diketiminato Fe oxo
complexes:10 controlled hydrolysis of a hydrido complex led to
the first FeII–O–FeII compound (Chart 2).
Bearing in mind these and other previous reports11–13 on the
rich chemistry of mononuclear diiminato complexes that often
leads to dinuclear compounds, the aim of this work was to start
D-12489, Berlin, Germany
bGeorg-August-Universita¨t, Institut fu¨r Anorganische Chemie, Tammannstr.
4, D-37077, Go¨ttingen, Germany
† Electronic supplementary information (ESI) available: Additional in-
formation concerning the interpretation of the SQUID data. CCDC
reference numbers 663093–663095. For crystallographic data in CIF or
other electronic format see . See DOI: 10.1039/b715376g
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1917–1923 | 1917
©

View Article Online
Chart 2
an investigation concerning dinuclear diiminato-iron chemistry
utilising [RXanthdim]2− as a preformed scaffold. One obvious
question was, whether this ligand is capable of supporting the
cooperation of two FeII centres for the activation of O2 by analogy
to the prosthetic group of the soluble methanmonoxygenase;14
here we describe the results of a corresponding study.
Results and discussion
The complexes described above all contain 2,6-iPr2C6H3 residues
at the N atoms of the diketiminato ligands. Probably for steric
reasons it is not possible to introduce this substituent as R in
[RXanthdim]H2 (see Chart 1): reaction of the corresponding oxo
precursor with 2,6-iPr2C6H3–NH2 proceeds incompletely. Full
substitution and good yields have so far been achieved for R = 2,3-
dimethylphenyl or 2,4-difluorophenyl. We therefore set out with
a lithiation of [Me2C6H3Xanthdim]H2 in thf using LDA (lithium di-
iso-propyl amide) as the reagent, and the resulting lithium salt
was treated in-situ with anhydrous iron dichloride; as FeCl2 is
insoluble in thf, this initially yielded a suspension. However, on
stirring overnight the lithium salt extracted FeCl2 into solution,
which consequently cleared up and turned red during this process.
After appropriate work-up dark red crystals could be obtained,
and Fig. 1 shows the result of an X-ray crystal structure inves-
tigation which in combination with analytical and spectroscopic
data identified the product as [Me2C6H3Xanthdim]Fe2Cl3(Li(thf)3), 1
(Scheme 2).
Each b-dialdiminato function acts as a bidentate ligand for an
iron(II) centre that is additionally coordinated by two chloride
ligands, a bridging and a terminal one. With respect to valency
there is thus one excessive chloride ligand in the complex and
the resulting additional charge is compensated by a Li cation
interacting with one of the terminal chloride ligands—similarly
as in case of the mononuclear complex I. Both iron centres are
coordinated in a distorted tetrahedral fashion, and they are only
slightly displaced out of the plane defined by the diiminato binding
pocket into the direction of the bridging chloride ligand. Their
Fig. 1 Crystal structure of 1. All hydrogen atoms have been omitted for
clarity. Table 1 lists relevant bonds lengths and angles.
◦
˚
Table 1 Selected bond lengths [A] and angles [ ] for 1
Fe1–Fe2
Fe1–Cl2
Fe1–Cl1
Fe1–N2
N1–C25
C24–C34
N2–C35
Fe2–N3
C45–N3
C44–C43
C53–N4
4.239(5)
2.353(3)
2.265(3)
2.009(7)
1.34(1)
1.42(1)
1.44(1)
1.961(7)
1.41(1)
1.40(2)
1.34(1)
Li1–Cl1
Fe2–Cl2
Fe1–N1
C26–N1
C25–C24
C34–N2
Fe2–Cl3
Fe2–N4
N3–C44
C43–C53
N4–C54
2.31(1)
2.375(3)
1.985(7)
1.47(1)
1.38(1)
1.29(1)
2.267(3)
2.032(7)
1.35(1)
1.38(2)
1.42(1)
Fe1–Cl2–Fe2
Cl1–Fe1–Cl2
N1–Fe1–Cl2
C26–N1–C25
C25–C24–C34
C34–N2–C35
N3–Fe2–N4
N3–C44–C43
C43–C53–N4
127.5(2)
101.8(2)
116.8(3)
111.8(6)
123.1(7)
114.7(7)
92.5(3)
Li1–Cl1–Fe1
Cl2–Fe2–Cl3
N1–Fe1–N2
N1–C25–C24
C24–C34–N2
N3–Fe2–Cl2
C45–N3–C44
C44–C43–C53
C53–N4–C54
149.6(2)
102.8(2)
92.3(3)
123.9(7)
131.2(7)
107.0(2)
115.1(7)
123.3(7)
122.4(7)
127.3(7)
127.9(7)
C1–C2–C43–C44
46(2)
C11–C12–C24–C2
48(2)
˚
˚
equally long [Fe1-Cl1: 2.266(6) A, Fe2-Cl3: 2.267(5) A] and ca.
˚
˚
0.1 A shorter than the bonds to the bridging chloride ligand
distance amounts to 4.239(5) A. The terminal Fe–Cl bonds are
Scheme 2
1918 | Dalton Trans., 2008, 1917–1923
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
˚
˚
the fluorine atoms of the 2,4-difluorophenyl residues apparently
undergo weak interactions with the Lewis acidic Fe centres: the
[Fe1-Cl2: 2.351(5) A, Fe2-Cl2: 2.376(6) A]. The angle Fe1-Cl2-
Fe2 amounts to 127.4(2)◦ which hints to an unstrained structure.
All C–C and C–N bonds belonging to the diiminato frameworks
are identical within the error margins, thus indicating a complete
delocalisation of the p-electrons.
˚
distances Fe1-F3 and Fe2-F7 are with 2.82(2) and 2.72(2) A far
˚
shorter than the sum of the van-der Waals radii (3.47 A).
In order to investigate the magnetic coupling between the two
ferrous ions in 1 and 2, SQUID measurements were carried out
on powdered samples at two different magnetic fields (5000 and
2000 G) in a temperature range from 295 to 2 K. No significant
field dependence was observed for any of the complexes. Diagrams
with plots of vM and vMT versus temperature are shown in Fig. 3.
The synthetic route described for
Me2C6H3Xanthdim]H2 can be transferred without any alterations
1
starting from
[
onto [F2C6H3Xanthdim]H2. After deprotonation with LDA and
addition of FeCl2 a red solution was obtained on stirring
overnight. Work-up led to crystals of the fluorinated derivative of
1 [F2C6H3Xanthdim]Fe2Cl3(Li(thf)3), 2 (Scheme 2), and Fig. 2 shows
the result of a single crystal X-ray analysis.
Fig. 2 Crystal structure of 2. All hydrogen atoms have been omitted for
clarity. Table 2 lists relevant bonds lengths and angles.
As observed for 1 the Fe2Cl3 unit adopts an M configuration
and all bond lengths and angles are very similar, apart from the
Fe–Cl–Fe angle which is somewhat more acute for 2 (116.5(2)
vs. 127.4(2)◦ in 1) and the Fe–Fe distance which in consequence
˚
is 0.2 A shorter in 2. One significant difference is the fact that
◦
˚
Table 2 Selected bond lengths [A] and angles [ ] for 2
Fig. 3 Plots of vM (solid circles) and vMT (open circles) versus temperature
for 1 (top) and 2 (bottom) at 5000 G; the solid lines represent the calculated
curve fits (see text).
Fe1–Fe2
Fe1–Cl2
Fe1–Cl1
Fe1–N1
C26–N1
C25–C24
C32–N2
Fe2–N3
C41–N3
C40–C39
C47–N4
4.030(3)
2.361(3)
2.262(3)
2.021(7)
1.43(2)
1.39(1)
1.33(1)
2.038(7)
1.43(1)
1.39(1)
1.32(1)
Li1–Cl1
Fe2–Cl2
Fe2–Cl3
Fe1–N2
N1–C25
C24–C32
N2–C33
Fe2–N4
N3–C40
C39–C47
N4–C48
2.3(3)
2.378(3)
2.274(3)
2.022(7)
1.31(1)
1.38(1)
1.41(2)
2.022(7)
1.32(1)
1.39(1)
1.43(1)
Values for the product vMT at 295 of 6.62 cm3 mol−1 K (7.28 lB)
for 1 and 6.84 cm3 mol−1 K (7.40 lB) for 2 are close to the spin-
only value expected for two uncoupled high-spin FeII ions (7.90 lB
for g = 2.28), but vMT gradually decreases upon lowering the
temperature, and the curves for vM go through a maximum around
70 (1) or 50 K (2, adumbrative). This behaviour is characteristic
for antiferromagnetic coupling of the two FeII ions and an S
= 0 ground state. Experimental data were modelled by using a
fitting procedure to the appropriate Heisenberg-Dirac-van-Vleck
(HDvV) spin Hamiltonian for two S = 2 centers with isotropic
exchange coupling and Zeeman splitting, eqn (1).15
Fe1–Cl2–Fe2
Cl1–Fe1–Cl2
N1–Fe1–Cl2
C26–N1–C25
C25–C24–C32
C32–N2–C33
N3–Fe2–N4
N3–C40–C39
C39–C47–N4
116.5(1)
104.8(2)
115.7(2)
117.0(7)
126.0(7)
116.9(7)
92.5(3)
Li1–Cl1–Fe1
Cl2–Fe2–Cl3
N1–Fe1–N2
N1–C25–C24
C24–C32–N2
N3–Fe2–Cl2
C41–N3–C40
C40–C39–C47
C47–N4–C48
125.4(3)
104.9(1)
94.1(3)
126.4(8)
128.9(8)
102.2(2)
117.9(7)
125.4(7)
117.5(7)
ˆ
ˆ ˆ
ˆ
ˆ
H = −2JS1S2 +glB(S1 + S2)B
(1)
126.9(8)
128.0(8)
A Curie-Weiss behaved paramagnetic impurity (q) that pre-
sumably causes the increase of vM at very low temperatures
and temperature-independent paramagnetism (TIP) were included
C1–C2–C39–C40
64(1)
C11–C12–C24–C25
62(1)
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1917–1923 | 1919
©

View Article Online
according to v = (1 − q)·v + q·vmono + TIP.16 The best fit gave
values of g = 2.29, J = −8.1 cm−1, q = 1.0% (fixed), TIP = 2.24 ×
10−4 cm3 mol−1 for 1 and g = 2.28, J = −5.0 cm−1, q = 1.0%
(fixed), TIP = 1.5 × 10−4 cm3 mol−1 (fixed) for 2; calculated curve
fits are shown as solid lines in Fig. 3. As expected, g values are
somewhat higher than 2 owing to some orbital contribution that is
typical for roughly tetrahedral d6 FeII ions.17 Magnetic coupling is
found to be only moderate, which is often observed for Cl-bridged
metal ions. The more pronounced coupling in 1 (−8.1 cm−1 versus
−5.0 cm−1 in 2) is likely due to the marked differences in Fe–Cl–Fe
bond angles (127.4(2)◦ in 1 versus 116.5(2)◦ in 2). Since high-
spin FeII in a distorted tetrahedral coordination geometry may
exhibit significant zero field splitting (ZFS) that can be similar in
magnitude to the J values determined,18 data analysis was also
◦
˚
Table 3 Selected bond lengths [A] and angles [ ] for 3
Fe1–Fe2
O2–Fe2
Fe2–Cl2
Fe1–N2
N1–C25
C24–C34
N2–C35
Fe2–N4
N3–C44
C43–C53
N4–C54
3.5190(1)
1.769(2)
2.2034(4)
1.974(2)
1.328(2)
1.395(2)
1.434(2)
1.987(2)
1.316(2)
1.393(2)
1.431(2)
Fe1–O2
Fe1–Cl1
Fe1–N1
C26–N1
C25–C24
C34–N2
Fe2–N3
C45–N3
C44–C43
C53–N4
1.769(2)
2.1957(4)
1.981(2)
1.428(2)
1.399(2)
1.318(2)
1.980(2)
1.435(2)
1.404(2)
1.323(2)
Fe1–O2–Fe2
Cl2–Fe2–O2
N1–Fe1–N2
N1–C25–C24
C24–C34–N2
N3–Fe2–O2
C45–N3–C44
C44–C43–C53
C53–N4–C54
168.04(8)
116.90(4)
92.59(6)
125.6(2)
126.4(2)
110.57(6)
119.0(2)
122.6(2)
119.1(2)
Cl1–Fe1–O2
N1–Fe1–O2
C26–N1–C25
C25–C24–C34
C34–N2–C35
N3–Fe2–N4
N3–C44–C43
C43–C53–N4
116.44(4)
112.79(6)
118.8(2)
124.6(2)
118.0(2)
90.73(6)
125.1(2)
127.3(2)
probed with an additional term 2D[S2 −1/3S(S + 1)] included
z
in eqn (1) that accounts for a local zero-field splitting parameter
D. However, this was found to have only a minor effect on the J
values, while the relative error surface for varying J and D values
confirms that the latter cannot be determined reliably from the
powder susceptibility data (see Fig. S1 and S2).†
C3–C2–C24–C34
84.8(2)
C13–C12–C43–C53
98.9(2)
1 and 2 contain two FeII centres in close proximity, and they
should therefore be ideally suited for a cooperative binding and ac-
tivation of dioxygen, similarly to those of the sMMO. Hence, after
preparation of 1 according to the route described above, removing
of all volatiles and addition of acetonitrile the resulting solution
was exposed to excessive O2. This led to a rapid colour change from
red to dark red-brown (in back light a green component could be
made out, that finally vanished). After removing of all volatiles
the residual black solid was extracted with toluene, and slow
evaporation of the solvent from the resulting solution yielded black
crystals. These were investigated by means of elemental analysis,
spectroscopic methods as well as by X-ray crystallography (see
Fig. 4), which revealed [Me2C6H3Xanthdim]Fe2Cl2O, 3 as the product
(Scheme 3). 3 can be derived from 1 via elimination of Li(thf)3Cl
Scheme 3
and introduction of a l-O bridge between the two Fe centres, and
consequently, the two iron centres are in the oxidation state +III.
As in 1 the two Fe centres are located in a distorted tetrahedral
coordination environment, and they are displaced significantly
out of the planes defined by the diiminato units towards the
˚
bridging O atom; hence, the Fe–Fe distance is with 3.5190(1) A
˚
0.72 A shorter than the one in 1/2. The Fe–O bonds are equally
˚
long, and the value of 1.769(2) A is consistent with distances
found in other complexes incorporating FeIII–O–FeIII moieties.19
Typically these exhibit Fe–O–Fe angles between 114 and 180◦ and
the corresponding angle of 168.04(2)◦ in 3 is located at the upper
end.
Plots of vM and vMT versus temperature for 3 obtained from
SQUID data are shown in Fig. 5.
In this case the product vMT at 295 of 1.00 cm3 mol−1 K (2.83 lB)
is much lower than the spin-only value expected for two uncoupled
high-spin FeIII ions (8.74 cm3 mol−1 K; 8.37 lB) and rapidly tends
to zero upon lowering the temperature, indicative of a strongly
antiferromagnetically coupled S = 5/2 pair. Experimental data
were analyzed by the HDvV model described above (eqn (1)),
including a Curie-Weiss-behaved paramagnetic impurity (q) with
spin S = 5/2 that accounts for the nonzero value of vMT between
10 and 50 K, and temperature-independent paramagnetism (TIP).
The calculated curve fit that represents the best fit parameters g =
2.06, J = −104.3 cm−1, q = 3.5%, TIP = 0.6 × 10−4 cm3mol−1 is
shown as a solid line in Fig. 5. The oxo bridge in 3 obviously me-
diates strong antiferromagnetic coupling, as has previously been
Fig. 4 Crystal structure of 3. All hydrogen atoms have been omitted for
clarity. Table 3 lists relevant bonds lengths and angles.
1920 | Dalton Trans., 2008, 1917–1923
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Fig. 5 Plot of vM (solid circles) and vMT (open circles) versus temperature for 3 at 5000 G; the solid lines represent the calculated curve fit (see text).
observed for many other binuclear oxo-bridged FeIII complexes.20
While early studies had advocated an angle dependence of J,21
it has become clear that magnetic coupling in these systems is
rather insensitive to the Fe–O–Fe angle, but is mainly dependent
on the Fe–(l-O) distances.22,23 The J value of −104.3 cm−1 for
the present complex 3 that features a quite symmetric Fe–O–Fe
1 could in fact arrange itself in a way that allows taking over an
oxygen atom from a potential peroxide or superoxide intermediate,
so that we favour this mechanism. Preliminary studies showed
that it is possible to replace the bridging chloride ligands in 1
and 2 by bridging acetate units using TlOAc. While this certainly
further improves the resemblance with the active centre of sMMO,
reactions with O2 so far did not proceed homogeneously and apart
from that no marked difference could be noted in comparison to
the conversions of 1 and 2.
Future research will address the oxidation mechanisms in detail
that appear to be complicated on the basis of the preliminary
results. The orienting UV/Vis experiments point to O2 binding
as the rate-determing step, and stopped-flow techniques have
to be employed to identify any short-lived intermediate and
its behaviour. We will also try to find conditions that enable
this intermediate to transfer one of the oxygen atoms to an
organic substrate. This would represent a first step towards the
establishment of a catalytic cycle.
˚
arrangement with d(Fe–O) = 1.769(2) A is in good agreement with
expectations.22,24
Some orienting experiments were undertaken to follow the
reaction of 1 with dioxygen at −20 ◦C in acetonitrile with the
aid of UV/Vis spectroscopy. These gave no distinct hints to an
intermediate with a life-time long enough on this time scale. On
the other hand no isosbestic points occur in the spectra which is
suggestive of an non-unimolecular mechanism.
A similar reaction with O2 was performed starting from 2.
However, crystals suitable for X-ray diffraction could not be
obtained so that we cannot be sure about the outcome even though
it seems likely that the reaction proceeds analogously to the one
◦
of 1. Performing the reaction at −30 C in acetonitrile a colour
change to red-brown could be observed as before for 1. However,
an intermediate colouration to green in the back light did not
occur. Pre◦liminary qualitative UV/Vis measurements showed that
even at 0 C the reaction of 2 is slower than the reaction of 1 at
−30 ◦C (probably due to the electron-withdrawing F substituents)
and in case of 2 isosbestic points indicate a unimolecular reaction.
Even thought this was not proved by isotopic enrichment
studies, there is no doubt that the O atoms within 3 and
4 are derived from O2, and this suggests initial O2 binding
and subsequent O–O bond cleavage. The latter could proceed
intramolecularly, possibly via a Fe2O2 diamond core structure,
or via reaction with a molecule of unreacted starting material that
could take over one of the O atoms. In the former case an O atom
has to be transferred to any substrate. However, after trapping
experiments performed at r.t. involving addition of reagents like
tetramethylethylene or cyclooctene in a ten-fold excess, oxygenated
products could not be detected. The second possibility at first
sight seems counterintuitive considering the steric bulk around
the reaction pocket, but a closer look shows that a molecule of
Experimental
General considerations and physical methods
Apart from the ligand synthesis all manipulations were carried
out in a glove-box, or else by means of Schlenk-type techniques
involving the use of a dry argon atmosphere. Solvents were dried
using a Braun Solvent Purification System, degassed by a vacuum-
freezing cycle and stored above molecular sieve 4 A. The H and
1
˚
13C NMR spectra were recorded on a Bruker AV 400 NMR
spectrometer (1H 400.13 MHz) or on a Bruker DPX 300 MHz
(1H 300.13 M◦Hz, 13C 75.5 MHz) with CDCl3, C6D6 or thf-d8 as
solvent at 25 C. The H and 13C NMR spectra were calibrated
1
against the residual proton and natural abundance 13C resonances
of the deuterated solvent (CDCl3 dH 7.24 ppm, dC 77.0 ppm; C6D6
dH 7.15 ppm dC 128.0 ppm; thf-d8 dH 3.58 ppm, dC 67.7 ppm).
Microanalyses were performed on a Leco CHNS-932 elemental
analyzer. Infrared (IR) spectra were performed using samples
prepared as KBr pellets on a Shimadzu FTIR-8400S spectrometer.
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1917–1923 | 1921
©

View Article Online
Magnetic data were measured with a Quantum-Design MPMS-
5S SQUID magnetometer equipped with a 5 T magnet. The
powdered samples were contained in a gel bucket and fixed in
a nonmagnetic sample holder. Each raw data file for the measured
magnetic moment was corrected for the diamagnetic contribution
from the sample holder and the gel bucket as well as the sample.
Anal.
Calcd
for
[
C6H3Me2Xanthdim]Fe2Cl3(Li(thf)3),
[C73H92Cl3Fe2LiN4O4]: C, 66.70; H, 7.05; N, 4.26; Cl, 8.09.
Found: C, 66.14; H, 7.17; N, 4.19; Cl, 7.99.
[
F2C6H3Xanthdim]Fe2Cl3(Li(thf)3), 2. The synthesis was per-
formed in close analogy to the one reported for 1 employing 0.300 g
(0.330 mmol) of [F2C6H3Xanthdim]H2 and 0.084 g (0.662 mmol) of
FeCl2. After similar workup the final addition of hexane (20 mL)
to a thf (4 mL) solution of the crude product yielded 2 in form of
large red crystals within 2 weeks (0.171 g, 0.123 mmol, 37%).
IR (KBr): 3061 (w), 2962 (w), 2906 (w), 2870 (w), 1701 (w), 1617
(w), 1590 (m), 1502 (s), 1466 (vs), 1444 (s), 1366 (w), 1317 (s), 1264
(s), 1195 (m), 1141 (m), 1098 (m), 1042 (w), 967 (w), 848 (m), 810
X-Ray crystallography for 1, 2 and 3. Single crystals of the
compounds were selected in a cold stream of nitrogen. They
were then mounted on top of a fiber and quickly frozen on the
diffractometer. Centered reflections were used by least-squares
calculations to indicate the unit cells. Unit cells and some impor-
tant parameters for data collection and structure refinement are
listed in Table 4. Diffraction data were collected in the appropriate
hemispheres under conditions specified also in Table 4.
The structures were solved by direct methods (program:
SHELXS-97 or SIR2004)25 and refined versus F2 (program:
SHELXL-97)26 with anisotropic temperature factors for all non-
hydrogen atoms at fully occupied positions, but three atoms in
the structure of 1 could only be refined isotropically. H atoms are
refined as riding on their carbon atoms throughout isotropically.
(w) cm−1.
1
Anal. Calcd for [C6H3F2Xanthdim]Fe2Cl3(Li(thf)3) ·
hexane,
2
1
2
[C65H68Cl3F8Fe2LiN4O4 · C6H14]: C, 58.78; H, 5.44; N, 4.03;
Cl, 7.65. Found: C, 58.47; H, 5.69; N, 4.10; Cl, 8.00.
[
Me2C6H3Xanthdim]Fe2Cl2O, 3. According to the procedure
described above 1 was synthesised (starting from 0.410 g
(0.468 mmol) [Me2C6H3Xanthdim]H2) and all volatiles were removed
in vacuum. Under a dry argon atmosphere the residual orange-
red solid was dissolved in 5 mL acetonitrile, and the resulting
solution treated with 12 mL (0.491 mmol) of O2 via a syringe.
Immediately the colour changed to dark red, and after 30 min.
of stirring all volatiles were removed again. The remaining solid
was extracted with toluene, and after filtration slow evaporation
of the solvent led to black single crystals of 3, that were suitable
for X-ray diffraction (0.440 g, 0.410 mmol, 87%).
IR (KBr): 3062 (w), 2961 (m), 2924 (w), 2866 (w), 1576 (m),
1458 (s), 1431 (vs), 1363 (m), 1291 (vs), 1225 (m), 1189 (w), 1081
(w), 974 (w), 863 (s), 781 (m), 730 (w) cm−1.
Anal. Calcd for [C6H3Me2Xanthdim](Fe2Cl2)-l-O · 1.5 tol.,
[C71.5H80Cl2Fe2N4O2]: C, 70.97; H, 6.66; N, 4.63; Cl, 5.86. Found:
C, 70.64; H, 6.70; N, 4.50; Cl, 6.49.
[
Me2C6H3Xanthdim]Fe2Cl3(Li(thf)3), 1. 0.400 g (0.457 mmol) of
Me2C6H3Xanthdim]H2 were dissolved in thf (5 mL) in a dry argon
[
atmosphere and treated with 0.098 g (0.916 mmol) of LDA. 0.120 g
(0.947 mmol) FeCl2 were added and the resulting suspension was
stirred for 15 h. Subsequently, all volatiles were removed under
vacuum, and the residual red-brown solid was dissolved in thf
(5 mL). Addition of hexane (20 mL) led to the precipitation of 1
(0.440 g, 0.324 mmol, 71%) in form of small red crystals. Somewhat
larger (single) crystals can be obtained by slow diffusion of hexane
into a thf solution of 1 within 3 weeks.
IR (KBr): 3064 (w), 2961 (s), 2927 (m), 2868 (w), 1575 (m), 1459
(vs), 1436 (vs), 1365 (w), 1299 (s), 1260 (w), 1224 (w), 1043 (s), 805
(w), 783 (w) cm−1.
Table 4 Crystal data for compounds 1, 2 and 3
1
2
3
Formula
Formula weight
C73H92Cl3Fe2LiN4O4
1314.50
P1
C136H150Cl6F16Fe4Li2N8O8
2778.62
P1
C143H160Cl4Fe4N8O4
2420.00
P1
¯
¯
¯
Space group
˚
a/A
11.644(3)
13.398(3)
24.910(5)
90.31(2)
94.28(2)
112.63(2)
3574(2)
2
14.284(5)
15.375(8)
16.471(8)
85.24(4)
76.33(4)
89.89(4)
3502(3)
1
12.7523(5)
14.5416(6)
18.6438(8)
89.961(3)
107.582(3)
104.126(3)
3186.1(2)
1
˚
b/A
˚
c/A
a/◦
b/◦
c /◦
3
˚
V/A
Z
Density/g cm−3
Radiation
1.221
1.317
1.261
˚
˚
˚
Mo Ka (k = 0.71073 A)
Mo Ka (k = 0.71073 A)
Mo Ka (k = 0.71073 A)
l/mm−1
0.566
0.598
0.587
Abs. corr.
Numerical integration
45190
Empirical (psi-scan)
11732
Numerical integration
58833
No. of reflns collected
No. of independent reflns
Ddata; parameter
9315 [R(int) = 0.3793]c
9729 [R(int) = 0.0889]
17124 [R(int) = 0.0473]
9315; 785
0.1411, 0.3207
9729; 816
17124; 763
b
R1a, wR2
0.0874, 0.2635
0.0380, 0.0918
ꢀ
ꢀ
ꢀ
a R1 =
ꢀFo| − |Fcꢀ/R|Fo|. b wR2 = { [w(Fo − Fc2)2]/ [w(Fo2)2]}
.
c The high R(int) results from low scattering power of the crystal
2
1/2
(cf. R(sigma) = 0.3346).
1922 | Dalton Trans., 2008, 1917–1923
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
G. W. Coates, J. Organomet. Chem., 2003, 683, 137 for an alternative
approach, involving, however, oxo imido units.
Conclusions
6 M. F. Pilz, C. Limberg and B. Ziemer, J. Org. Chem., 2006, 71, 4559.
7 M. F. Pilz, C. Limberg, B. B. Lazarov, K. C. Hultzsch and B. Ziemer,
Organometallics, 2007, 26, 3668.
8 J. M. Smith, R. J. Lachicotte and P. L. Holland, Chem. Commun., 2001,
1542.
9 S. A. Stoian, J. Vela, J. M. Smith, A. R. Sadique, P. L. Holland, E.
Mu¨nck and E. L. Bominaar, J. Am. Chem. Soc., 2006, 128, 10181; J. M.
Smith, A. R. Sadique, T. R. Cundari, K. R. Rodgers, G. Lukat-Rodgers,
R. J. Lachicotte, C. J. Flaschenriem, J. Vela and P. L. Holland, J. Am.
Chem. Soc., 2006, 128, 756.
10 N. A. Eckert, S. A. Stoian, J. M. Smith, E. L. Bominaar, E. Mu¨nck and
P. L. Holland, J. Am. Chem. Soc., 2005, 127, 9344.
A novel ligand system has been used to prepare dinuclear
versions of previously known b-diiminato iron complexes for
the activation of O2. Two ligand versions were employed—one
with electron donating methyl substituents at the N-aryl donor
functions and one containing electron withdrawing F atoms—
to produce two complexes [Me2C6H3Xanthdim]Fe2Cl3(Li(thf)3), 1
and [F2C6H3Xanthdim]Fe2Cl3(Li(thf)3), 2 each of which comprises
Cl–Fe(l-Cl)Fe–Cl(Li(thf)3) units. Their iron centres are coupled
(weakly) antiferromagnetically and both complexes indeed readily
react with O2 to give complexes [RXanthdim]Fe2Cl2O containing
FeIII–O–FeIII moieties with strongly coupled iron centres. Due to
the electron withdrawing fluorine substituents at the ligand, 2
reacts more slowly than 1. Future research will now focus on
elucidating the mechanism of O2 activation and its utilisation for
the oxygenation of organic substrates.
11 N. A. Eckert, J. M. Smith, R. J. Lachicotte and P. L. Holland, Inorg.
Chem., 2004, 43, 3306.
12 P. L. Holland, T. R. Cundari, L. L. Perez, N. A. Eckert and R. J.
Lachicotte, J. Am. Chem. Soc., 2002, 124, 14416; J. Vela, J. M. Smith,
R. J. Lachicotte and P. L. Holland, Chem. Commun., 2002, 2886; J. Vela,
S. Vaddadi, T. R. Cundari, J. M. Smith, E. A. Gregory, R. J. Lachicotte,
C. J. Flaschenriem and P. L. Holland, Organometallics, 2004, 23,
5226.
13 S. A. Stoian, Y. Yu, J. M. Smith, P. L. Holland, E. L. Bominaar and
E. Mu¨nck, Inorg. Chem., 2005, 44, 4915; Y. Yu, J. M. Smith, C. J.
Flaschenriem and P. L. Holland, Inorg. Chem., 2006, 45, 5742.
14 M. Merkx, D. A. Kopp, M. H. Sazinsky, J. L. Blazyk, J. Mu¨ller and
S. J. Lippard, Angew. Chem., 2001, 113, 2861; M. Merkx, D. A. Kopp,
M. H. Sazinsky, J. L. Blazyk, J. Mu¨ller and S. J. Lippard, Angew. Chem.,
Int. Ed., 2001, 40, 2782.
15 O. Kahn, Molecular Magnetism, Wiley-VCH, Publishers Inc., 1993.
16 Simulation of the experimental magnetic data with a full-matrix diag-
onalization of exchange coupling and Zeeman splitting was performed
with the julX program: E. Bill, Max-Planck Institute for Bioinorganic
Chemistry, Mu¨lheim/Ruhr, Germany.
Acknowledgements
We are grateful to the Fonds der Chemischen Industrie, the BMBF
and the Dr Otto Ro¨hm Geda¨chtnisstiftung for financial support as
well as to Bayer Services GmbH & Co. OHG, BASF AG and Sasol
GmbH for the supply of chemicals. We also would like to thank
P. Neubauer and Elke Hoppe for crystal structure analyses and
Silke Hinze for the preparation of starting materials. Finally, we
acknowledge inspiring discussions within the Cluster of Excellence
“Unifying Concepts in Catalysis” funded by the DFG.
17 F. E. Mabbs, D. J. Machin, Magnetism and Transition Metal Complexes,
Chapman and Hall, London, England, 1973.
18 (a) B. W. Dockum and W. M. Reiff, Inorg. Chim. Acta, 1986, 120, 61–
76; (b) R. T. Stibrany, M. V. Lobanov, H. J. Schugar and J. A. Potenza,
Inorg. Chem., 2004, 43, 1472–1480.
19 R. H. Beer, W. B. Tolman, S. G. Bott and S. J. Lippard, Inorg. Chem.,
1991, 30, 2082; K. Takahashi, Y. Nishida, Y. Maeda and S. Kida,
J. Chem. Soc., Dalton Trans., 1985, 2375; M. Kodera, H. Shimakoshi,
M. Nishimura, H. Okawa, S. Iijima and K. Kano, Inorg. Chem., 1996,
35, 4967.
References
1 L. Bourget-Merle, M. F. Lappert and J. R. Severn, Chem. Rev., 2002,
102, 3031.
2 J. M. Smith, R. L. Lachicotte, K. A. Pittard, T. R. Cundari, G. Lukat-
Rodgers, K. R. Rodgers and P. L. Holland, J. Am. Chem. Soc., 2001,
123, 9222.
3 G. W. Coates and D. R. Moore, Angew. Chem., Int. Ed., 2004, 43, 6618.
4 A. M. Reynolds, E. A. Lewis, N. W. Aboelella and W. B. Tolman, Chem.
Commun., 2005, 2014.
20 K. S. Murray, Coord. Chem. Rev., 1974, 12, 1–35.
21 R. N. Mukherjee, T. D. P. Stack and R. H. Holm, J. Am. Chem. Soc.,
1988, 110, 1850–1861.
5 L. Bourget-Merle, P. B. Hitchcock and M. F. Lappert, J. Organomet.
Chem., 2004, 689, 4357; D. V. Vitanova, F. Hampel and K. C. Hultzsch,
J. Organomet. Chem., 2005, 690, 5182; N. F. Curtis, Comprehensive
Coordination Chemistry II, ed. J. A. McCleverty and T. J. Meyer and
A. B. P. Lever, Elsevier, Oxford, vol. 1, 2004, ch. 1.20; S. Y. Lee, S. J.
Na, H. Y. Kwon, B. Y. Lee and S. O. Kang, Organometallics, 2004, 23,
5382; B. Y. Lee, S. J. Na, H. Y. Kwon, S. Y. Lee, S. J. Na, S. Han, H.
Yun, H. Lee and Y.-W. Park, J. Am. Chem. Soc., 2005, 127, 3031; J.
Vela, L. Zhu, C. J. Flaschenriem, W. W. Brennessel, R. J. Lachicotte
and P. L. Holland, Organometallics, 2007, 26, 3416; T. J. Hebden, W. W.
Brennessel, C. J. Flaschenriem and P. L. Holland, Dalton Trans., 2006,
3588; also compare; S. D. Allen, D. R. Moore, E. B. Lobkovsky and
22 (a) S. M. Gorun and S. J. Lippard, Inorg. Chem., 1991, 30, 1625–1630;
(b) G. Haselhorst, K. Wieghardt, S. Keller and B. Schrader, Inorg.
Chem., 1993, 32, 520–525.
23 D. M. Kurtz, Jr., Chem. Rev., 1990, 90, 585–606.
24 A. Hazell, K. B. Jensen, C. J. McKenzie and H. Toftlund, Inorg. Chem.,
1994, 33, 3127–3134.
25 G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467–473, from the
program SHELXS; SIR2004; M. C. Burla, R. Caliandro, M. Camalli,
B. Carrozzini, G. L. Cascarano, L. De, Caro, C. Giacovazzo, G. Polidori
and R. Spanga, J. Appl. Crystallogr., 2005, 38, 381–388.
26 G. M. Sheldrick, SHELXL-97, Program for refinement of crystal
structures, University of Go¨ttingen, Germany, 1997.
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 1917–1923 | 1923
©