Oxidative activation of aryldiynyl-iron complexes: Regioselective dimerization

Article abstract of DOI:10.1021/om400014g

Chemical oxidations of the iron butadiynyl complexes Cp*(dppe) FeCi -CCi -CAr (Ar = Ph, 1a; Ar = C₆H₄-Me-4, 1b; Cp* = η⁵-C₅Me₅, dppe = 1,2-bis(diphenylphosphino)ethane) achieved with 1 equiv of ferroc

Full text of DOI:10.1021/om400014g

Article
pubs.acs.org/Organometallics
Oxidative Activation of Aryldiynyl−Iron Complexes: Regioselective
Dimerization
Alexandre Burgun,† Frederic Gendron,† Thierry Roisnel,† Sourisak Sinbandhit,‡ Karine Costuas,*,†
́
́
Jean-Franco̧
is Halet,*,† Michael I. Bruce,*,§ and Claude Lapinte*,†
†Institut des Sciences Chimiques de Rennes, UMR CNRS 6226, Universite
France
́
de Rennes 1, Campus de Beaulieu, F-35042 Rennes,
‡CRMPO, Universite
́
de Rennes 1, Campus de Beaulieu, F-35042 Rennes, France
§School of Chemistry & Physics, University of Adelaide, South Australia 5005, Australia
S
* Supporting Information
ABSTRACT: Chemical oxidations of the iron butadiynyl
complexes Cp*(dppe)FeCCCCAr (Ar = Ph, 1a; Ar =
C6H4-Me-4, 1b; Cp* = η5-C5Me5, dppe = 1,2-bis-
(diphenylphosphino)ethane) achieved with 1 equiv of
ferrocenium hexafluorophosphate salt in THF at −78 °C
provide the transient species 1a(PF6) and 1b(PF6), as
demonstrated by ESR experiments. Above −35 °C, intra-
molecular C−C couplings take place regiospecifically, providing the binuclear complexes [{Cp*(dppe)Fe}2C8Ar2](PF6)2 (Ar =
Ph, 2a; Ar = C6H4-Me-4, 2b), which were isolated after purification as thermally stable deep purple powders in 61 and 57%
yields, respectively. The full characterization of the new products, including X-ray analysis of 2a(PF6)2, electrochemical data,
spectroscopic properties, and electronic structures investigated at the DFT level of theory are reported. NMR investigations have
shown that the iron building blocks rotate slowly around the cyclobutene ring, allowing the observation of three rotamers in
equilibrium. The cyclic voltammetry of the dimer 2b(PF6)2 shows four redox events with the current peaks corresponding to two
one-electron-oxidation and two one-electron-reduction processes. These compounds are diamagnetic in the solid state but ESR
active in solution, indicating that the triplet excited state is readily accessible.
On the other hand, the properties and reactivity of 17e
monocationic complexes resulting from the 1e oxidation of
metal−arylalkynyl complexes have been the subject of extensive
studies.12,16−20 In some cases, selective coupling at Cβ produced
the corresponding bis-vinylidene derivatives.16,17,21,22 In the
particular example of the complex Cp(PPh3)2Ru−CCPh, it
was found that 1e oxidation enabled the formation of dimeric
species by radical coupling at sites shown to be electron rich,
particularly involving the alkynyl Cβ and phenyl Cpara atoms.23
Similar efforts on the preparation, characterization, and
reactivity of 17e polyyne−metal complexes are very limited.24
Rigaut et al. published the first example of the dimerization of a
diynyl complex activated by chemical oxidation.25 Whiteley and
co-workers have found that monoelectronic oxidation of
(C7H7)(LL)Mo(CCCCR) (LL = bpy, dppe; R = SiMe3,
H) produces molybdenum-stabilized diynyl radicals which were
characterized by extensive spectroscopic studies and by a
structural study of [(C7H7)(dppe)Mo(CCCCSiMe3)]-
(PF6).22 Further oxidation of this compound afforded a dimeric
product resulting from Cβ−Cβ coupling. Very recently, some of
us reported that chemical oxidation of Cp(dppe)Ru(CCC
CR) with [Cp2Fe](PF6) affords the binuclear cations [{Cp-
INTRODUCTION
■
Molecules comprised of two redox-active centers connected by
an all-carbon bridge provide an ideal template from which
intricate mechanistic details concerning the factors controlling
electron transfer, electron delocalization, and magnetic
exchange interactions can be extracted.1−3 The major limitation
of such systems, which have often been referred to as molecular
wires,4 comes from the chemical instability of these molecules
when unpaired electrons are present on the compounds with
chains longer than four carbon atoms.5 This chemical instability
probably arises from intermolecular carbon−carbon bond
formation, producing dimers and polymers. Few efforts have
been devoted to the “insulation” of the unsaturated linkers to
prevent bimolecular reactions.6 Incorporation of functional
groups into the bridge between the redox centers increases the
chemical stability of the oxidation states containing an odd
number of electrons, allows the control of the molecular
topology, and opens access to functional molecules with
oriented electron-transfer capabilities,7 switching abilities,8
multipath connections,9,10 optical activities,11,12 and magnetic
properties.3,10,13 Recently, it was shown that electroactive units
such as 1′,1‴-biferrocenyl moieties incorporated in the carbon
bridge linking the metal termini act as a molecular relay,
mediating electron transfer between the two ends of the
molecule by hopping processes.14,15
Received: January 9, 2013
Published: February 28, 2013
© 2013 American Chemical Society
1866
dx.doi.org/10.1021/om400014g | Organometallics 2013, 32, 1866−1875

Organometallics
Article
Scheme 1. Oxidative Dimerization of Substituted Diynyl−Ruthenium Complexes
(dppe)Ru}2C8R2](PF6)2 (R = Ph, Fc) by radical coupling at
electron-rich sites involving the inner and outer CC triple
bonds, to give cyclobutenediylidene derivatives.26 As depicted
in Scheme 1, mixtures of symmetrical and asymmetrical isomers
were obtained. DFT calculations revealed that the precursor
cationic diynyl complexes have considerable electron density on
Cβ and Cδ, suggesting that radical couplings Cδ + Cδ and Cβ +
Cδ afford the symmetrical and asymmetrical isomers,
respectively.26
Scheme 2. Oxidative Dimerization of Substituted Diynyl−
Iron Complexes
Related to this chemistry is the earlier report of the
dimerization of the butatrienylidene (dppe)2(CO)W{C
CCCH(C6H4But)} to give {(dppe)2(CO)W}{CC[
CH(C6H4But)]C[CH(C6H4But)]CC}{W(CO)(dppe)2}
by Cγ−Cγ coupling. It has been suggested that this reaction
proceeds via a 1,2-bis(ethynyl)cyclobutenyl diradical inter-
mediate which undergoes retro-electrocyclic ring opening.27
A recent experimental and theoretical investigation on iron
and ruthenium σ-polyynyl complexes has concluded that the
iron(III) radicals bearing ligands with four and six sp carbon
atoms are stable in solution for a few minutes at 20 °C and for
an unlimited time below −20 °C. It was also suggested that the
17e iron radicals may react via a smooth chemical process, since
the ESR signal disappeared cleanly, no traces of other radicals
being detected.24 Guided by these stimulating results, we have
investigated the reactivity of the two aryldiynyl−iron complexes
Cp*(dppe)FeCCCCAr (Ar = Ph, 1a; Ar = C6H4-Me-4,
1b) and we report here the regiospecific formation and the full
characterization of the binuclear cations [{Cp*(dppe)-
Fe}2C8Ar2](PF6)2 (Ar = Ph, 2a(PF6)2; Ar = C6H4-Me-4,
2b(PF6)2), including X-ray analysis, electrochemical data,
spectroscopic properties, and quantum chemical investigations
at the DFT level.
RESULTS AND DISCUSSION
■
1. Synthesis of 2a,b. Complex 1b was prepared according
to the procedure already described for complex 1a. Chemical
oxidations of the complexes 1a,b were achieved with 1 equiv of
ferrocenium hexafluorophosphate salt in THF at −78 °C
(Scheme 2). An immediate color change from orange to deep
green was observed, and this color persisted as long as the
solution was kept below −35 °C. ESR experiments established
that these transient species are the iron(III) derivatives 1a(PF6)
and 1b(PF6). Upon warming, when the temperature reached
−35 °C, the color progressively turned deep purple, indicating
that a further chemical reaction had occurred. After 1 h at 20
°C, addition of pentane precipitated colored materials. After
purification, complexes 2a(PF6)2 and 2b(PF6)2 were isolated as
thermally stable deep purple powders in 61 and 57% yields,
respectively. High resolution mass spectra of the analytically
pure sample of 2a(PF6)2 established the dimeric formula of the
dicationic complex, the structure of which was completely
elucidated by single-crystal X-ray diffraction.
2. Molecular Structure of 2a(PF6)2. Small thin purple
crystals of 2a(PF6)2 were obtained by slow diffusion of pentane
into a concentrated solution of the complex in 1,2-dichloro-
ethane. Although the crystals diffracted only weakly due to their
poor quality, the X-ray structure was solved (R = 0.087),
establishing definitively the connectivity between atoms. An
ORTEP view of the cation 2a2+ is shown in Figure 1, while
1867
dx.doi.org/10.1021/om400014g | Organometallics 2013, 32, 1866−1875

Organometallics
Article
selected structural parameters are collected in Table 1 with the
computed structural data (see below).
(1.784(5) Å) is shorter by ca. 0.09 Å than that in 1a, which
is consistent with the presence of an FeC double bond, as
proposed in Scheme 2. Moreover, distances in the carbon
ligand (C(1)−C(2) = 1.266(6) and C(2)−C(3) = 1.345(6) Å)
are very close to the distances found in the allenylidene
complex [Cp*(dppe)Fe(CCCPh2)](PF6) (Fe−C(1) =
1.784(4) Å, C(1)−C(2) = 1.257(7) Å, and C(2)−C(3) =
1.361(7) Å) previously reported.28 These bond lengths are
characteristic of allenylidene derivatives with the C(1)−C(2)
distances shorter than the C(2)−C(3) distances, suggesting a
contribution from the canonical form Cp*(dppe)Fe{C
CC+(R)(R′)}.29 The Fe−C(1)−C(2)−C(3) chain slightly
deviates from linearity, with angles at C(1) and C(2) being
174.4(3) and 170.3(4)°. Distances within the cyclobutene ring
are consistent with the presence of three single bonds (2 ×
C(3)−C(4) = 1.477(6) Å and C(3)−C(3′) = 1.509(8) Å) and
one CC double bond (C(4)−C(4′) = 1.417(9) Å). No short
contacts were observed in this structure; however, the two Cp*
rings are very close to each other, with distances between
hydrogen atoms of the two closest methyl groups as short as
3.7 Å.
Figure 1. ORTEP drawing of the cation 2a2+ of 2a(PF6)2. Thermal
ellipsoids are at the 30% probability level. Hydrogen atoms and solvent
molecule are omitted for clarity.
The X-ray analysis clearly shows that 2a(PF6)2 is a
symmetrical dimer with two Cp*(dppe)FeC2 fragments linked
by a C4Ph2 ligand containing a four-carbon ring. The C4 ring is
planar, and the angles are close to perfect right angles (88.0(2)
and 91.9(2)°); their sum is exactly equal to 360°. As expected,
the 18e iron atoms adopt a pseudo-octahedral geometry with
bond lengths in the Cp*(dppe)Fe fragments consistent with
the presence of a positive charge on the metal atoms. Thus, the
Fe−Cp*cent (1.769 Å) and Fe−P distances (2.233(1) and
2.243(1) Å) are clearly elongated with respect to the reference
neutral complex 1a.24 The Fe−C(1) distance in 2a(PF6)2
Starting from the X-ray structure obtained for this complex
2a(PF6)2, full geometry optimization was first performed on
compound 2a2+ at the DFT BP86/TZP level of theory.
Pertinent computed bond lengths, for both singlet and triplet
spin states of 2a2+, are given in Table 1 and are compared with
the experimental values. The optimized distances of the singlet
state compared rather well with the experimental data. The Fe−
2+
2+
Table 1. Pertinent Bond Lengths (Å) and Relative Energies (eV) Computed for Rotamers 2a2+, 2a′ , and 2a″ and for the
a
Hypothetical Asymmetrical Isomer 3a2+
2a2+
2+
2+
singlet
triplet
3a2+ (singlet)
2a′ (singlet)
2a″ (singlet)
Fe1−P(1)
Fe1−P(2)
Fe2−P(1)
Fe2−P(2)
Fe1−Cp*
Fe2−Cp*
Fe1−C1
C1−C2
C2−C3
C3−C4
C4−Ph
Fe2−C5
C5−C6
2.266 [2.2332(12)]
2.265 [2.2425(12)]
2.266 [2.2332(12)]
2.265 [2.2425(12)]
1.827 [1.769]
2.277
2.273
2.317
2.286
1.826
1.824
1.811
1.271
1.353
1.439
1.445
1.863
1.258
1.371
1.423
1.449
1.557
1.464
2.303
2.288
2.293
2.282
1.856
1.848
1.876
1.498
1.384
1.231
1.412
1.846
1.277
1.368
1.461
1.449
2.279
2.265
2.255
2.268
1.827
1.836
1.797
1.275
1.345
1.477
1.447
1.811
1.277
1.351
1.477
1.459
1.524
1.408
2.258
2.256
2.258
2.256
1.836
1.836
1.806
1.277
1.349
1.478
1.453
1.806
1.277
1.349
1.478
1.453
1.530
1.407
1.827 [1.769]
1.797 [1.784(5)]
1.274 [1.266(6)]
1.346 [1.345(6)]
1.477 [1.479(6)]
1.450 [1.450(5)]
1.797 [1.784(5)]
1.274 [1.267(6)]
1.346 [1.345(6)]
1.477 [1.477(6)]
1.450 [1.450(5]
1.513 [1.516(8)]
1.413 [1.417(8)]
C6−C7
C7−C8
C8−Ph
C3−C7
C4−C8
C1−C7
C2−C8
E (eV)
1.574
1.413
b
b
b
b
b
0
+ 0.95
+ 1.88
+0.11
+0.30
c
c
c
+solvent effect
0
+0.12
+0.35
d
d
e
d
d
+dispersion
+solv + disp
0
+ 1.01
+ 0.98
+0.21
+0.23
e
e
e
0
+0.22
+0.28
a
A third rotamer, 2a‴2+ (90° rotation of one metal end group) was also calculated. It is lying too high in energy to be found in solution (+0.978 eV).
Its characteristics are thus only reported in the Supporting Information. Experimental distances of 2a(PF6)2 are given in brackets for the sake of
comparison. Cp*: centroid of C5Me5. Relative energy after optimization. Relative energy with solvent effect (see Computational Details). Relative
energy with dispersion effects (see Computational Details). Relative energy with solvent and dispersion effects taken into account.
b
c
d
e
1868
dx.doi.org/10.1021/om400014g | Organometallics 2013, 32, 1866−1875

Organometallics
Article
C distances are computed for both metal centers at 1.797 Å and
thus are quite close to the experimental value of 1.784 Å. The
optimized C−C bond lengths also match fairly well with
experiment, especially the C−C distances of the central carbon
square, where the largest discrepancy computed is 0.004 Å.
Nevertheless, the common overestimation of the Fe−P and
Fe−Cp*centroid bond lengths is also computed in this case.24,30
These discrepancies are respectively 0.03 and 0.06 Å.
The Mulliken decomposition and energies of the first frontier
molecular orbitals of 2a2+ obtained in the singlet state are
collected in Table S1 (see the Supporting Information), and
their plots are shown in Figure 2. The energy gap between the
reminiscent of the π-MOs of the cyclobutadiene dication
[C4H4]2+.26
In contrast to the singlet state, the optimized distances of the
triplet state are not in good agreement with the corresponding
experimental values. The computed Fe−C distances (1.811 and
1.864 Å) largely overestimate the experimental bond by 0.03
and 0.08 Å, respectively. Furthermore, bond lengths of the
carbon square are far away from the experimental values.
Indeed, the C3−C4 and C7−C8 bond lengths are under-
estimated by 0.05 and 0.04 Å, respectively, whereas the C3−C7
and C4−C8 bond lengths are overestimated by 0.05 and 0.04 Å,
respectively (see Table 1). The computed deviations of the
Fe−P and Fe−Cp*centroid bond lengths are larger in this spin
state in comparison to the singlet optimized structure. The
largest deviation of the Fe−P distance is 0.08 Å, and the Fe−
Cp*centroid bond length is computed to be longer by ca. 0.05 Å.
These results are consistent with the singlet spin state being
more stable than the triplet state by roughly 0.95 eV and with
the lack of any ESR signal in measurements of solid-state
samples of 2a(PF6)2 (vide infra). Solvent effects and dispersion
corrections were added to the calculations in order to better
estimate the energy difference between singlet and triplet states
(see Table 1). Interestingly, the order of magnitude (1 eV) of
the energy difference is kept, comforting the previous
conclusions.
3. Conformational Analyses of 2a(PF6)2 and 2b(PF6)2
by Multinuclear NMR Spectroscopy. The 31P NMR spectra
of 2a(PF6)2 and 2b(PF6)2, display four signals at δ 89.36, 90.06,
90.45, and 91.00 for 2a(PF6)2 and at δ 89.64, 90.32, 90.76, and
91.25 for 2b(PF6)2 with relative intensities 16:58:16:10 and
14:64:14:8 (for a total intensity of 100), respectively (Figure
S1, Supporting Information). The hexafluorophosphate anions
were also observed as septuplets at δ −142.99 (1JPF = 708 Hz)
and −142.94 (1JPF = 708 Hz) for 2a(PF6)2 and 2b(PF6)2,
respectively. It was initially thought that the four signals might
result from the formation of several isomers of the dimeric
derivative, as previously observed in the ruthenium series (see
Scheme 1).26 However, this hypothesis was disproved by three
different pieces of experimental and theoretical evidence. (i)
Attempts to separate the components of the mixture were
carried out without any success. In particular, fractional
precipitations or crystallization provided samples which always
gave NMR spectra with the same number of signals with exactly
the same ratio, even when measurements were achieved from
solutions prepared from single crystals of 2a(PF6)2 used for X-
ray analyses. (ii) The spectroscopic signature of carbon−carbon
triple bond expected in asymmetrical isomers was not detected
either in the IR or 13C NMR spectra, in contrast with the
Figure 2. Plots, energies (eV), and Fe2/C8/Ph2 percentage
contributions of the first frontier molecular orbitals of 2a2+ in its
singlet spin state (isocontour value 0.03 [e/bohr3]1/2).
highest occupied molecular orbital (HOMO) and the lowest
unoccupied molecular orbital (LUMO) of ca. 1.1 eV reflects the
thermodynamic stability of 2a2+ as a singlet. The HOMO and
HOMO-1 are dπ/π-type in character, antibonding between
iron centers and adjacent carbon atoms and bonding between
carbon atoms C1 and C2 and between C5 and C6. These
HOMOs are mainly metallic in character with computed iron
percentage contributions of roughly 60%, whereas the organic
character is only 17% and 4% for the HOMO and HOMO-1,
respectively. In contrast, as previously computed for the
ruthenium analogues of formula [{Ru(η5-C5H5)-
(dppe)}2C8Ph2](PF6)2, the LUMO and LUMO+1 are fully
delocalized over the metal−carbon skeleton with an important
character computed on the carbon atoms of the central square,
Chart 1. Experimental Symmetrical (2a2+) and Hypothetical Asymmetrical (3a2+) Isomers
1869
dx.doi.org/10.1021/om400014g | Organometallics 2013, 32, 1866−1875

Organometallics
Article
Table 2. Computed Atomic Spin Densities (electron) of 2a2+(T), the Monometallic Precursor 1a+ and Related Compounds
compd
M
Cα
Cβ
0.20
Cγ
−0.05
−0.03
−0.07
−0.13
Cδ
Cortho
Cmeta
Cpara
ref
a
1a+
0.70
−0.06
0.01
0.17
0.22
0.20
0.23
0.05
0.07
−0.02
−0.03
0.07
0.10
24
26
33
31
a
[RuCp(dppe)C4Ph]+
0.33
0.22
a
[Mo(C7H7)(dppe)C4SiMe3](PF6)
0.87
−0.07
−0.17
0.18/0.01
0.21
b
[Mo(C7H7)(dppe)C4H](PF6)
0.99
0.31
2a2+(T)
0.84/0.33
0.03/0.07
−0.01/−0.02
−0.01/0.32
a
b
BP86/TZP (ADF) level of theory. B3LYP/LANL2DZ-6-31G* (Gaussian) level of theory.
ruthenium examples (Figure 1).26 (iii) As we will just see in the
following, DFT calculations on the symmetrical form 2a2+ and
asymmetrical form 3a2+ of the iron complexes (Chart 1)
isostructural with the ruthenium dimers show that the former is
thermodynamically more stable.
A large amount of spin density is found on Cβ, with quasi-
similar values of 0.20e in all examples, and on Cδ, with 0.17e,
0.22e, and 0.21e in the iron, ruthenium, and molybdenum
cases, respectively.
Thus, guided by these spin density distributions, several
routes are electronically available for the formation of binuclear
species involving radical coupling reactions between metal
centers and Cβ or/and Cδ atoms. However, in ruthenium
examples, M + M and Cβ + Cβ) radical coupling processes were
not found by Bruce et al., probably because of steric protection
by the bulky dppe ligands, and only the two Cβ + Cδ and Cδ +
Cδ radical couplings were observed, leading to the asymmetrical
and symmetrical isomers, respectively, as shown in Scheme 1.26
In the molybdenum series, dimerization of precursors only
occurs at the Cβ atom, the steric hindrance of the SiMe3 end
group protecting Cδ atoms.22 For the iron precursor 1a+, both
Cβ + Cδ and Cδ + Cδ couplings are electronically possible.
However, only the Cδ + Cδ radical coupling process, which
affords the symmetrical isomer 2a(PF6)2, was observed.
Dimerization of 1a+ via Cδ + Cδ radical coupling, as opposed
to the Cβ + Cδ coupling, might be explained using steric
considerations. The Cp* ligands used in iron cases (which are
replaced by Cp and C7H7 ligands in the ruthenium and
molybdenum series, respectively) increase the steric hindrance
about the metallic fragments which (i) sterically protect Cβ
from the radical coupling process and (ii) strongly energetically
destabilize the asymmetrical isomer. Indeed, when both
symmetrical and asymmetrical isomers were optimized using
Fe(dppe)Cp metallic fragments instead of Fe(dppe)Cp*
fragments, the energy difference computed between isomers
decreased to ca. 0.78 eV, still in favor of the symmetrical isomer
(Table S2, Supporting Information), comparable to that
computed for Ru analogues.26 These results suggest that
selectivity of the electron-rich carbon sites for radical coupling
reactions is controlled not only electronically by the spin
density distribution on the diynyl ligand but also by steric
factors. Previous works also support this conclusion. In
particular, it was found that the radical cations [Cp*(dppe)-
FeCCH](PF6) and [Cp*(dppe)FeC(OCH3)CH2](PF6)
both dimerize by formation of a carbon−carbon bond between
the β-carbon atoms of two molecules.17,34 In each case the β-
carbon atom carries either one or two hydrogen atoms.
Replacement of the hydrogen atoms by any other substituents
kinetically stabilizes the radical cations. In summary, the
stability difference between the iron and ruthenium series
probably results from steric hindrance. Note, however, that the
ruthenium complexes contain Cp ligands, whereas the iron
complexes bear Cp* ligands.
At first sight, the 31P NMR spectrum of 2a(PF6)2, shown in
Figure S1 (Supporting Information), is not very different from
that obtained for the mixture of products obtained after
oxidation of the ruthenium-based complexes Cp(dppe)Ru-
(C4R) (R = Ph, Fc).26 Indeed, the oxidative dimerization of the
monometallic ruthenium diynes leads to mixtures of sym-
metrical and asymmetrical isomers, as depicted in Scheme 1.
For the sake of comparison, the analogous iron-based
asymmetrical isomer 3a2+ (Chart 1), which is not found
experimentally, was also optimized and the main data are
reported in Table 1. The computed metal−ligand (dppe, Cp*)
distances differ strongly from those in the symmetrical isomer
2a2+. The Fe−Cp* distances lengthen from 1.827 Å for 2a2+ to
1.856 and 1.848 Å for 3a2+. The Fe−P bond lengths are also
affected, with computed values of 2.303 and 2.288 Å for Fe1−P
and 2.293 and 2.282 Å for Fe2−P in the latter, which render the
two iron fragments inequivalent and eventually distinguishable
by 31P NMR. However, this asymmetrical isomer 3a2+ is
computed to be energetically less stable than the symmetrical
isomer 2a2+ by 1.88 eV. In comparison, the energy gap between
the two analogous ruthenium isomers was also in favor of the
symmetrical dimer, but by only ca. 0.6 eV.26 This drastic
difference between the iron and ruthenium examples suggests
that the formation and existence of the iron-based asymmetrical
dimer 3a2+ is largely disfavored. This result is quite surprising at
first sight. As discussed in the Introduction, the dimerization
reaction involving radical Cβ−Cβ coupling is common for
closely related compounds such as [Mo(η-C7H7)(dppe)-
C4SiMe3],22 [Mo(η-C7H7)(dppe)C2Ph],20 and [Ru(η-C5H5)-
(dppe)C4Ph].26
With the aim of understanding the oxidative coupling
reaction affording the iron dimers, the electronic structure of
the monometallic precursor must be discussed. Precursors 1a
and [RuCp(dppe)C4Ph] were recently studied at the DFT
level.24,26 The computed atomic spin densities obtained for
these 17e species are recalled in Table 2 and are compared with
those of the analogous molybdenum precursor [Mo(η-C7H7)-
(dppe)C4H]+ recently studied by Whiteley and co-workers.31
In both cases, the largest part of the spin density is found on
the metal center. Nevertheless, the atomic spin density on the
metal atom considerably increases, from 0.33 to 0.77 and to
0.99, when the iron and molybdenum fragments successively
replace the ruthenium center. This result is consistent with the
previous studies of Ru(III)−, Fe(III)−, and Mo(III)−
arylalkynyl complexes, where localization of the unpaired
electron on the metal center increases in the same order.19,32
Surprisingly enough, the computed atomic spin densities on the
carbon-rich ligand are not strongly affected by this substitution.
The experimental and theoretical data clearly show that the
dimerization of the iron(III) diynyl complexes 1a(PF6) and
1b(PF6) regioselectively provides the symmetrical complexes
2a(PF6)2 and 2b(PF6)2, respectively, as the single product of
the coupling reactions. The presence of five resonances in the
1870
dx.doi.org/10.1021/om400014g | Organometallics 2013, 32, 1866−1875

Organometallics
Article
Scheme 3. Selected Rotamers of 2a(PF6)2 and 2b(PF6)2
31P NMR spectra results from the presence of three rotamers in
equilibrium, which interconvert slowly on the NMR time scale.
1H and 13C VT experiments were carried out in d6-acetone, but
only a very small line broadening was detected at 55 °C, the
highest temperature accessible in this solvent. The existence of
found at δ 278.84, 281.05, 283.25, and 283.46 as triplets (the
two last triplets being superimposed) with 2JCP = 36 Hz in each
case. Two similar patterns of singlets were observed at δ 180.11,
180.63, 181.00, 181.95 and at δ 146.67, 147.51, 147.91, 149.42
ppm (Figure S4, Supporting Information), both including one
intense peak and two medium peaks of similar intensity
together with one small peak. These resonances were assigned
to the Cγ and Cδ carbon atoms, respectively. The pattern
corresponding to the Cβ resonances was observed around 205
ppm, but unfortunately it was overlapped by the carbonyl
resonance of the solvent d6-acetone. Chemical shifts observed
for Cα, Cβ, and Cγ of the allenylidene fragments are consistent
with values previously obtained for iron allenylidene com-
plexes.35 Additionally, resonances of the Cp* ligands were also
observed as patterns of singlets (Figure S5, Supporting
Information) at δ 102.39, 102.56, 103.09, and 103.11 for the
C5Me5 carbon atoms and at δ 10.36, 10.43, 10.64, and 10.78 for
the Me carbon atoms. Chemical shifts around 100 ppm for the
Cp* ring carbon atoms are characteristic of cationic Fe(II)
complexes.36−38
In order to give further support for the existence of the three
rotamers, a conformational study was carried out starting from
the most stable isomer 2a2+. The relative orientation between
the metallic fragments Fe(dppe)Cp* is known to be crucial.10
Relative energies and geometrical data of the most stable
species are given in Table 1. Rotamer 2a′2+, where one metallic
fragment has rotated by 180°, is computed to be energetically
less stable by 0.11 eV in comparison with 2a2+. Distances of the
carbon backbone are not strongly affected by rotation. The
largest deviation is computed for the C3−C7 bond length,
which increases from 1.513 to 1.524 Å when going from 2a2+ to
2a′2+, respectively. However, a change of the metal−ligand
distances is observed upon rotation. Indeed, from rotamer 2a2+
to rotamer 2a′2+, the Fe(2)−C5 and Fe(2)−Cp* separations
are lengthened by 0.014 and 0.009 Å, respectively. Fe−P bond
lengths are also affected. In 2a2+, Fe−P distances are computed
to be quasi-identical for both metal centers, whereas in 2a′2+,
the Fe−P bond lengths strongly differ between metallic
fragments (Fe(1)−P = 2.279/2.265 Å, Fe(2)−P = 2.255/
2.268 Å), rendering the two dppe ligands inequivalent and
distinguishable in 31P NMR. Rotamer 2a″2+ (Scheme 3), where
the two metallic fragments are rotated by 180°, is calculated to
be 0.30 eV less stable than 2a2+. Geometrical changes observed
upon the rotation of one metallic end group, from 2a2+ to 2a′2+,
are reproduced upon rotation of the second metallic fragment.
However, in 2a″2+, the computed Fe−P distances are computed
to be identical for both metallic fragments (Fe−P(1) = 2.256 Å,
Fe−P(2) = 2.258 Å), rendering the two dppe ligands
equivalent. Using this conformational study, peaks observed
1
the three rotamers was clearly validated by EXSY H NMR
experiments. The EXSY (or NOESY) spectra are symmetrical
along the diagonal, and correlation spots indicate exchange
situation (or spatial proximity). The EXSY spectrum obtained
with 2b(PF6)2 is shown in Figure S2 (Supporting Information).
The protons of Cp* groups were used as probes to observe the
exchange processes. Circled spots in Figure S1 (Supporting
Information) clearly show that exchange occurred in solution
between the Cp* signals located in the range δ 1.20−1.45.
Correlation spots observed between the methyl group of the
tolyl ligand in the aromatic region also confirm the exchange
process. The 31P NMR spectra of 2a(PF6)2 and 2b(PF6)2 are
shown in Figure S1 and are fully consistent with the rotamer
hypothesis. The major peaks at δ 90.06 in the spectrum of the
dimer 2a(PF6)2 and at δ 90.32 for 2b(PF6)2 correspond to the
thermodynamically more stable rotamers. They are identified as
2a(PF6)2 and 2b(PF6)2 in Scheme 3. In crystalline samples only
these rotamers are present and 2a(PF6)2 was identified by X-ray
analysis. In these rotamers, the Cp* ligands are on the same
side of the molecule. A rotation of 180° of one metal fragment
provides the rotamers 2a′(PF6)2/2b′(PF6)2. The two inequi-
valent dppe ligands resonate as two singlets with the same
intensity at δ 89.36 and 90.45/89.64 and 90.76, respectively.
The last peaks at δ 91.00/91.25 could correspond to rotamers
2a″(PF6)2/2b″(PF6)2, respectively. These rotamers are derived
from rotamers 2a(PF6)2 and 2b(PF6)2 by a rotation of 180° of
the two metal fragments, and the two dppe ligands are still
equivalent.
1
In the H NMR spectra, similar patterns containing three or
four different peaks with the same relative intensities were
observed for the Cp* protons at δ 1.34 (×2), 1.37, and 1.43 for
the three rotamers 2a(PF6)2, 2a′(PF6)2, and 2a″(PF6)2 and at δ
1.31, 1.36, 1.38, and 1.55 for the three rotamers 2b(PF6)2,
2b′(PF6)2, and 2b″(PF6)2. In these two examples, the dppe
CH2 protons were found between δ 2.40 and 3.65 as broad
signals. The methyl substituents of the tolyl groups from the
three rotamers 2b(PF6)2, 2b′(PF6)2, and 2b″(PF6)2 were found
between δ 2.40 and 2.50.
In the 13C NMR spectrum of dimer 2a(PF6)2 (the 13C NMR
spectrum of dimer 2b(PF6)2 is similar), patterns analogous to
the 31P NMR spectra, containing four signals of specific
intensities which are characteristic of the three rotamers, are
observed in the spectrum. The four signals corresponding to
the Cα carbons (Figure S3, Supporting Information) were
1871
dx.doi.org/10.1021/om400014g | Organometallics 2013, 32, 1866−1875

Organometallics
Article
in the 31P NMR spectrum of 2a(PF6)2 can now be assigned.
The major peak centered at δ 90.06 corresponds to the most
stable rotamer 2a2+. Then, the two equivalent signals measured
at δ 89.36 and 90.45 can be assigned to rotamer 2a′2+, where
the two dppe ligands are inequivalent. This result contrasts with
the analogous ruthenium compounds, where similar signals
were assigned to the asymmetrical isomer.26 Finally, the last
peak centered at δ 91.00 ppm corresponds to rotamer 2a″2+,
where the two metallic fragments are rotated by 180°, again
making the two dppe ligands equivalent.
Dimers 2a(PF6)2 and 2b(PF6)2 were further characterized by
IR spectroscopy. The IR spectra exhibit one strong band at
1888 cm−1 together with a medium band at 1968 cm−1, while
the P−F bands of the PF6 anions were observed at νPF 839
cm−1 in both cases. In the relevant literature, allenylidene
νCCC bands for organometallic complexes [M]{CC
C(R)R′} can be found over a large range between 1865 and
2200 cm−1, depending on the metal, the ligands attached to the
metal atom, and especially the nature of the R and R′ end
groups.29 However, the νCCC band in [Fe(CC
CPh2)(dppe)Cp*][PF6]35 was found at 1896 cm−1, suggesting
that the band at 1888 cm−1 corresponds to the νCCC band of
the dimers 2a(PF6)2 and 2b(PF6)2. The medium band at 1968
cm−1 might be due to the restricted rotation in the allenylidene
fragments, giving the different rotamers.37
0.578(8) mm/s) are exceptionally large with respect to the
values generally found for organometallic complexes of the
Cp*(dppe)Fe series evidencing FeII metal centers.2,28,37,38
The IS parameter reflects the electron density at the iron
nucleus. It decreases as the positive charge is more localized on
the metal center or less stabilized by electrostatic interactions
between cations and anions.40 In the case of 2b(PF6)2 the value
is very large and does not compare with the data obtained in
the case of iron cumulenylidene complexes, for which the IS
parameters range between 0.8 and 1.7 mm/s vs Fe.28 Values of
IS as large as that found for 2b(PF6)2 have been occasionally
found for substituted ferrocenium cations such as {1′,1‴-
[Cp*(dppe)Fe−CC]2bfc}(PF6)n (bfc = 1,1‴-biferrocenyl).15
It is interesting to note that the redox potential of this
ferrocenium derivative (E° = 0.49 V) is not far from the first
oxidation potential of 2b(PF6)2.15
In the Cp*(dppe)Fe series, the neutral iron(II) complexes
have QS values close to 2.0 mm/s while the cationic iron(III)
analogues exhibit QS values slightly below 1.0 mm/s. The
metallacumulenylidenes are characterized by QS parameters
ranging between 1.0 and 1.5 mm/s.28 The Mossbauer data
̈
strongly suggest that the C8 ligand with a four-carbon
cyclobutene strongly modifies the ligand field of the iron
nucleus and plays a role different from those of conventional
cumulene ligands. It apparently acts as a very electron
withdrawing ligand, in accord with the redox potentials.
4. Cyclic Voltammetry of 2b(PF6)2. The initial scan in the
CV of the dimer 2b(PF6)2 from −1.4 to +1.4 V vs SCE shows
four redox events with the same current peaks corresponding to
two one-electron-oxidation and two one-electron-reduction
processes. The reversibility of the oxidation waves found at E°3
= +1.07 (PF6)2 and E°4 = +1.25 V vs SCE could not be clearly
established because of their positions close to the solvent front;
nevertheless, an ia/ic current peak ratio smaller than 1 is
suspected. In contrast, the two reduction waves centered at E°1
= −0.94 and E°2 = −0.59 V are reversible (ia/ic = 1) at the
platinum electrode. In accord with the strong metallic character
of the HOMO of these binuclear complexes (see above), the
oxidation processes are probably iron-centered and correspond
to the sequential formation of the (FeII/FeIII)3+ and (FeIII/
FeIII)4+ species. The two reduction processes are probably
carbon-bridge centered, and they more especially concern the
cyclobutene ring, due to its strong contribution in the
description of the LUMO (see above). Despite an apparent
stability on the platinum electrode, attempts to electrolyze
2b(PF6)2 revealed that the monocationic derivative 2b(PF6) is
not stable in solution, new electroactive species appearing
6. Glass ESR Spectroscopy. Powdered samples of
complexes 2a(PF6)2 and 2b(PF6)2 are not ESR active, but
surprisingly these binuclear compounds, which were fully
characterized by NMR at 20 °C, are ESR active at liquid
nitrogen temperature in frozen solution. For this reason,
2a(PF6)2 was subjected to a detailed ESR study. Optimal
resolution was obtained for samples dissolved in CH2Cl2 and
transferred into an ESR quartz tube before being cooled to 67
K. The X-band ESR spectrum shows an intense signal
characteristic of an Fe(III) complex in a pseudo-octahedral
environment (Figure S6, Supporting Information). The
characteristic three g tensors for the Δms =
1 transition
were found at g1 = 2.443, g2 = 2.035, and g3 = 1.989. The
magnetic anisotropy Δg = g1 − g3 = 0.454 is very large,
indicating the metallic character of this magnetic complex.
Interestingly, ca. 1.2 electrons out of the two unpaired electrons
are found on the iron centers in the high-spin configuration of
2a2+. The quantum chemical calculations performed for
2a2+(T) are in agreement with an important spin density on
the metal centers (0.842 and 0.327). Nevertheless, with a total
energy of roughly 1 eV higher than that of the singlet state
(solvent and dispersion effects included), this compound is
most probably not the magnetic compound observed
progressively as the reduction proceeds.
57
5. Fe Mossbauer Spectroscopy. 57Fe Mossbauer
̈
̈
spectroscopy is a very sensitive probe for establishing the
iron oxidation state and the nature of the Fe−C bond.37
Moreover, in a homogeneous series of iron compounds such as
Cp*(dppe)Fe(CCR), it has been shown that the isomeric
shift is related to the electron density at the metal center.28
experimentally. Furthermore, the Δms
=
2 transition
characteristic of the triplet state was clearly observed at g =
4.263. This interesting observation demonstrates that complex
2a(PF6)2 carries two unpaired electrons and magnetic exchange
interactions between them take place. In its high-spin
configuration this complex can be regarded as a bis-FeIII
derivative.
Additionally, Mossbauer spectroscopy constitutes a sensitive
̈
and useful means of probing the selectivity of the redox
reactions. The 57Fe Mossbauer spectrum of a microcrystalline
̈
sample of the complex 2b(PF6)2 was run at 80 K and least
squares fitted with Lorentzian line shapes.39 It displays a unique
doublet, which proves the thermal stability of the complex and
confirms the X-ray data obtained on a single crystal showing
that all the iron atoms are in the same environment, excluding
traces of unsymmetrical complexes. The isomer shift (IS =
0.538(5) mm/s vs Fe) and the quadrupole splitting (QS =
CONCLUSION
■
In this contribution, we have found that the radical cations
[Cp*(dppe)FeCCCCAr](PF6) (Ar = Ph, 1a; Ar = C6H4-
Me-4, 1b) are stable below −35 °C in solution. Above this
temperature they dimerize to provide specifically the sym-
metrical dimers [{Cp*(dppe)Fe}2C8Ar2](PF6)2 (Ar = Ph,
1872
dx.doi.org/10.1021/om400014g | Organometallics 2013, 32, 1866−1875

Organometallics
Article
H, 6.36. Found: C, 77.11; H, 6.36. IR (KBr): ν(CC) 2149, 1999
cm−1. 1H NMR (C6D6, 300 MHz): δ 1.47 (s, 15H, Cp*), 1.78 (m, 2H,
PCH2), 1.97 (s, 3H, Me), 2.60 (m, 2H, PCH23), 6.76 (d, 3JHH = 8.0 Hz,
2H, CH), 6.99−7.29 (m, 16H, Ph), 7.46 (d, JHH = 8.0 Hz, 2H, CH),
2a(PF6)2; Ar = C6H4-Me-4, 2b(PF6)2). The formation of these
new complexes results from the initial formation of a carbon−
carbon bond between atoms in Cδ with respect to the iron
nucleus of each radical cation. The iron building blocks rotate
slowly around the cyclobutene ring on the NMR time scale,
allowing the observation of three rotamers in equilibrium.
Structures for the rotamers are proposed, and quantum
chemical investigations have shown that their relative stabilities
are largely dependent on the orientations of the metal end caps.
From the redox potentials of the binuclear complex 2b(PF6)2, it
has been shown that a strong electronic interaction occurs
between the redox centers. The compound is diamagnetic in
the solid state, while the triplet excited state becomes thermally
accessible in solution. Further efforts will be devoted to the
extension of the scope of this reaction, replacing the aryl groups
by functional substituents.
3
8.01 (t, JHH = 8.7 Hz, 4H, Ph). 13C NMR (C6D6, 75 MHz, ppm): δ
8.91 (s, C5Me5), 19.86 (s, Me), 28.79−29.21 (m, dppe), 59.79 [s, C
CCC(C6H4Me-4)], 80.20 [s, CCCC(C6H4Me-4)], 87.20 (s,
C5Me5), 100.11 [s, CCCC(C6H4Me-4)], 122.77−137.88 (m,
2
Ph), 142.98 [t, JCP = 38 Hz, CCCC(C6H4Me-4)], 154.70 (s,
Ph). 31P NMR (C6D6, 121 MHz): δ 100.3 (s). ES-MS (m/z): calcd for
C47H46FeP2 728.2424, found 728.2427 [M]+.
Synthesis of [{Cp*(dppe)FeCC}2{μ-C4Ph2}](PF6)2
(2a(PF6)2). Fe(CCCCPh)(dppe)Cp* (1a; 100 mg, 0.14 mmol)
and [FeCp2](PF6) (46 mg, 0.14 mmol) were dissolved in THF (10
mL) at −78 °C, upon which the color changed immediately from
orange to deep green. After it was stirred for 1 h at −78 °C, the
solution was slowly warmed to room temperature over a period of 5 h.
When the temperature reached −35 °C, the color of the solution
changed from deep green to deep purple. After the solution was stirred
for 1 h at room temperature, pentane (60 mL) was added and the
resulting precipitate was filtered off and washed with pentane (3 × 10
mL) to afford [{Cp*(dppe)FeCC}2{μ-C4Ph2}](PF6)2
(2a(PF6)2; 74 mg, 61%) as a deep purple powder. Anal. Calcd for
C92H88F12Fe2P6·CH2Cl2: C, 61.91; H, 5.03. Found: C, 62.02; H, 5.24.
EXPERIMENTAL SECTION
■
General Data. Manipulations of air-sensitive compounds were
performed under an argon atmosphere using standard Schlenk
techniques or in an argon-filled Jacomex 532 drybox. All glassware
was oven-dried and vacuum or argon flow degassed before use. Fourier
transform infrared (FT-IR) spectra were recorded using a Bruker
IFS28 spectrophotometer (range 4000−400 cm−1) as solids dispersed
in KBr pellets. UV−visible spectra were recorded on a CARY 5000
spectrometer. 1H, 13C, and 31P NMR spectra were recorded on a
Bruker AVIII 400 NMR multinuclear spectrometer at ambient
temperature, unless otherwise noted. 29Si NMR spectra were run on
a Bruker 300-MSL spectrometer. Chemical shifts are reported in parts
1
IR (KBr): ν(CCC) 1968, 1888; ν(P−F) 839 cm−1. H NMR (d6-
acetone, 600 MHz): δ 1.34, 1.37 (2 × s, 13H, Cp*), 1.43 (s, 17H,
Cp*), 2.51, 2.94, 3.44, 3.65 (4 × s (br), 8H, PCH2), 6.93−7.69 (m,
50H, Ph). 13C NMR (d6-acetone, 150 MHz): δ 10.36, 10.43, 10.64,
10.78 (4 × s, C5Me5), 29.57−32.33 (m, dppe), 102.39, 102.56, 103.09,
103.11 (4 × s, C5Me5), 126.75−135.52 (m, Ph), 146.67, 147.51,
147.91, 149.42 (4 × s, Ph), 180.11, 180.63, 181.00, 181.95 (4 × s, Cγ),
202.59, 205.46, 205.62 (3 × s, Cβ), 278.84, 281.05, 283.25, 283.46 (4
× t, 4 × 2JCP = 36 Hz, Cα). 31P NMR (d6-acetone, 121 MHz): δ 89.36,
90.06, 90.45, 91.00 (4 × s), −142.99 (septet, 1JPF = 708 Hz, PF6). ES-
MS (m/z): calcd for C92H88P4Fe2 714.2262, found 714.2274 [M]2+;
calcd for C92H89P4Fe2 1429.4608, found 1429.4539 [M + H]+; calcd
for C92H88F6Fe2P5 1573.4172, found 1573.4097 [M + PF6]+.
per million (δ) relative to tetramethylsilane (TMS) for H and 13C
1
spectra and external 85% aqueous H3PO4 for 31P NMR spectra.
Coupling constants (J) are reported in hertz (Hz), and integrations are
reported as the number of protons. The following abbreviations are
used to describe peak patterns: br = broad, s = singlet, d = doublet, t =
triplet, q = quartet, m = multiplet. Mass spectra were run on a HP
5971/A/5890-II GC/MS coupling (HP 1 capillary column, length 25
m, diameter 0.2 mm, 0.33 μm polydimethylsiloxane). ESR spectra
were recorded on a Bruker EMX-8/2.7 (X-band) spectrometer. The
Synthesis of [{Cp*(dppe)FeCC}2{μ-C4(1,4-C6H4Me-
4)2}](PF6)2 (2b(PF6)2). Similarly, from Fe(CCCCC6H4Me-4)-
(dppe)Cp* (1b; 100 mg, 0.14 mmol) and [FeCp2](PF6) (45 mg, 0.14
mmol) was obtained [{Cp*(dppe)FeCC}2{μ-C4(C6H4Me-
4)2}](PF6)2 (2b(PF6)2; 70 mg, 57%). IR (KBr): ν(CCC) 1968,
1888; ν(P−F) 839 cm−1. 1H NMR (d6-acetone, 400 MHz): δ 1.31 (s,
4H, Cp*), 1.36 (s, 4H, Cp*), 1.38 (s, 3H, Cp*), 1.55 (s, 19H, Cp*),
2.08, 2.28, 2.38 (3 × s, 6H, C6H4Me), 2.48, 2.97, 3.45, 3.64 (4 × m,
8H, PCH2), 6.84−7.67 (m, 48H, Ph). 13C NMR (d6-acetone, 100
MHz): δ 10.23, 10.27, 10.41, 10.55 (4 × s, C5Me5), 21.91(m,
C6H4Me), 29.27−32.61 (m, dppe), 101.99, 102.14, 102.71, 102.75 (4
× s, C5Me5), 126.77−135.65 (m, Ph), 140.99, 143.32, 143.42, 143.64
57Fe Mossbauer spectra were recorded with a 2.5 × 10−2 C (9.25 × 108
̈
Bq) 57Co source using a symmetric triangular sweep mode. Computer
fitting of the Mossbauer data to Lorentzian line shapes was carried out
̈
with a previously reported computer program.39 The isomer shift
values are reported relative to iron foil at 298 K. Elemental analyses
were conducted on a Thermo-FINNIGAN Flash EA 1112 CHNS/O
́
analyzer by the Microanalytical Service of the Centre Regional de
Mesures Physiques de l’Ouest (CRMPO) at the University of Rennes
1, Rennes, France.
(4 × s, Ph), 180.37, 180.62, 181.98, 181.70 (4 × s, Cγ), 200.10, 203.10,
Materials. Reagent grade toluene, tetrahydrofuran (THF), diethyl
ether, and pentane were dried and deoxygenated by distillation from
sodium diphenylketyl. Dichloromethane was distilled under argon
from P2O5 and then from Na2CO3. Cp*(dppe)FeCCCC-Ph
(1a),24 Cp*(dppe)FeCl,41 and [FeCp2](PF6) (ferrocenium hexafluor-
ophosphate)42 were prepared following published procedures. All
other reagents are commercially available and were used as received
without further purification.
2
203.16, 203.73 (4 × s, Cβ), 278.07, 280.45, 282.58 (3 × t, 3 × JCP
=
36 Hz, Cα). 31P NMR (d6-acetone, 121 MHz): δ 89.64, 90.32, 90.76,
1
91.25 (4 × s), −142.94 (septet, JPF = 708 Hz, PF6).
X-ray Crystal Structure Determination for 2a(PF6)2. Single
crystals were mounted on a APEXII Bruker-AXS diffractometer
equipped with a CCD camera and a graphite-monochromated Mo Kα
radiation source (λ = 0.71073 Å), from the Centre de Diffractomet
́
rie
(CDFIX), Universite de Rennes 1, Rennes, France. Data were
́
Synthesis of Fe{CCCC(C6H4Me-4)}(dppe)Cp* (1b). To a
suspension of Cp*(dppe)FeCl (100 mg, 0.16 mmol) and NaBPh4 (66
mg, 0.19 mmol) in triethylamine (15 mL) was added a solution of
HCCCC(C6H4Me-4) (27 mg, 0.19 mmol) in THF (1 mL). The
mixture turned slowly from dark green to orange. After one night at
collected at 100(2) K. The structure was solved by direct methods
using the SIR-97 program and refined with a full-matrix least-squares
method on F2 using the SHELXL-97 program.43 Crystal data, data
collection, and structure refinement parameters for 2a(PF6)2:
empirical formula, C102H108Fe2P6F12·5C2H4Cl2; formula mass,
2213.90 g mol−1; crystal system, monoclinic; space group, P2/c (No.
13); a, 16.7558(10) Å; b, 12.3342(7) Å; c, 25.7292(12) Å; α, 90°; β,
95.714(2)°; γ, 90°; V, 5291.0(2) Å3; Z, 2; Dcalcd, 1.390 g cm−3; crystal
size, 0.25 × 0.13 × 0.05 mm; F(000), 2280; absorption coefficient,
0.683 mm−1; θ range, 3.40−27.46°; range h,k,l, −21 ≤ 21, −15 ≤ 15,
−33 ≤ 33; total number of reflections, 44871; number of unique
reflections, 12007; number of observed reflections (I > 2σ(I)), 8936;
t
room temperature, BuOK (excess) was added to the solution before
removal of the solvent under reduced pressure. The solid residue was
then extracted with toluene (3 × 10 mL), and the solvent was removed
under reduced pressure. The residue was extracted a second time with
diethyl ether (3 × 10 mL), and after removal of the solvent under
reduced pressure, the resulting orange powder was dried under
vacuum to afford Fe{CCCC(C6H4Me-4)}(dppe)Cp* (1b; 93
mg, 80%) as an orange powder. Anal. Calcd for C47H46FeP2: C, 77.47;
1873
dx.doi.org/10.1021/om400014g | Organometallics 2013, 32, 1866−1875

Organometallics
Article
number of data/restraints/parameters, 12007/0/588; final R, R1 =
0.0867, wR2 = 0.2057; R indices (all data), R1 = 0.112, wR2 = 0.2237;
goodness of fit/F2, 1.034; largest difference peak and hole, 2.101 and
−1.932 e Å−3.
REFERENCES
■
(1) (a) Crutchley, R. J. Adv. Inorg. Chem. 1994, 41, 273−325.
(b) Ward, M. D. Chem. Soc. Rev. 1995, 24, 121−134. (c) Ward, M. D.
Chem Ind-London 1996, 568−573. (d) Paul, F.; Lapinte, C., Magnetic
communication in binuclear organometallic complexes mediated by
carbon-rich bridges. In Unusual Structures and Physical Properties in
Organometallic Chemistry; Gielen, M., Willem, R., Wrackmeyer, B.,
Eds.; Wiley: London, 2002; pp 220−291. (e) Demandis, K. D.;
Hartshorn, C. M.; Meyer, T. J. Chem. Rev. 2001, 101, 2655−2685.
(f) Brunschwig, B. S.; Creutz, C.; Sutin, N. Chem. Soc. Rev. 2002, 31,
168−184. (g) Szafert, S.; Gladysz, J. A. Chem. Rev. 2003, 103, 4175−
4205. (h) Szafert, S.; Gladysz, J. A. Chem. Rev. 2006, 106, PR1−PR33.
(i) Akita, M.; Koike, T. Dalton Trans. 2008, 3523−3530. (j) Halet, J.-
F.; Lapinte, C. Coord. Chem. Rev. 2013, DOI: 10.1016/
j.ccr.2012.09.007. (k) Low, P. J. Coord. Chem. Rev. 2013,
DOI: 10.1016/j.ccr.2012.08.008.
Computational Details. DFT calculations were carried out with
the Amsterdam Density Functional package (ADF 2010)44 without
any symmetry constraints on the bimetallic dimers 2a2+, 3a2+, 2a′2+,
and 2a″2+. Electron correlation was treated within the local density
approximation (LDA) in the Vosko−Wilk−Nusair parametrization.45
The nonlocal corrections of Becke and Perdew were added to the
exchange and correlation energies, respectively.46 The analytical
gradient method implemented by Versluis and Ziegler was used.47
The standard ADF TZP basis set was used: i.e., a triple-ζ STO basis set
for the valence core augmented with a 2p polarization function for H, a
3d polarization function for C, P, and Si, and a 4p polarization function
for Fe. Orbitals up to 1s, 2p, and 3p were kept frozen for C, P, Si, and
Fe, respectively. Geometry optimization convergence criteria were
more drastic than default criteria (energy change <0.0005 hartree and
atomic position displacement <0.005 Å). Single-point calculations
were performed on the optimized geometries in order to partially
evaluate solvent dispersion effects. To address solvation effects, the
conductor-like screening model (COSMO)48 was used with a
dielectric constant simulating dichloromethane solvent. The geo-
metries were not relaxed because of the prohibitive computational
cost. For the dispersion effects, DFT-D3 corrections proposed by
Grimme were added to the total bonding energy, gradient, and second
derivatives.49 Molecular orbitals were plotted with the ADF-GUI
package.50
(2) Paul, F.; Lapinte, C. Coord. Chem. Rev. 1998, 178−180, 427−505.
(3) Lapinte, C. J. Organomet. Chem. 2008, 693, 793−801.
(4) Lehn, J.-M., Supramolecular Chemistry-Concepts and Perspectives;
Wiley-VCH: Weinheim, Germany, 1995; p 271.
(5) (a) Coat, F.; Lapinte, C. Organometallics 1996, 15, 477−480.
(b) Szafert, S.; Paul, F.; Meyer, W. E.; Gladysz, J. A.; Lapinte, C. C. R.
Chim. 2008, 11, 693−701.
(6) Weisbach, N.; Baranova, Z.; Gauthier, S.; Reibenspies, J. H.;
Gladysz, J. A. Chem. Commun. 2012, 48, 7562−7564.
(7) (a) Tsiperman, E.; Becker, J. Y.; Khodorkovsky, V.; Shames, A.;
Shapiro, L. Angew. Chem. 2005, 44, 4015−4018. (b) Coat, F.;
Guillevic, M.-A.; Toupet, L.; Paul, F.; Lapinte, C. Organometallics
1997, 16, 5988−5998. (c) Paul, F.; Meyer, W. E.; Toupet, L.; Jiao, H.;
Gladysz, J. A.; Lapinte, C. J. Am. Chem. Soc. 2000, 122, 9405−9414.
(d) Bruce, M. I.; Costuas, K.; Davin, T.; Ellis, B. E.; Halet, J.-F.;
Lapinte, C.; Low, P. J.; Smith, K. M.; Skelton, B. W.; Toupet, L.;
White, A. H. Organometallics 2005, 24, 3864−3881.
(8) Tanaka, Y.; Ishisaka, T.; Inagaki, A.; Koike, T.; Lapinte, C.; Akita,
M. Chem. Eur. J. 2010, 16, 4762−4776.
(9) Ami, S.; Joachim, C. Phys. Rev. B 2002, 65, 155419−155425.
(10) Costuas, K.; Cador, O.; Justaud, F.; Le Stang, S.; Paul, F.;
Monari, A.; Evangelisti, S.; Toupet, L.; Lapinte, C.; Halet, J.-F. Inorg.
Chem. 2011, 50, 12601−12622.
ASSOCIATED CONTENT
■
S
* Supporting Information
Mulliken decomposition and energies of the first molecular
frontier orbitals of 2a2+ obtained in the singlet state (Table S1),
energy and composition of the first electronic excitation energy
for 2a2+ (Table S2), Cartesian coordinates of the optimized
geometries of 2a2+ (Table S3), 31P NMR spectra of 2a(PF6)2
and 2b(PF6)2 (Figure S1), 1H EXSY NMR spectrum of
2b(PF6)2 (Figure S2), 13C NMR spectrum of 2a(PF6)2 (Cα
pattern; Figure S3), 13C NMR spectrum of 2a(PF6)2 (Cγ and
quaternary aromatic carbon patterns; Figure S4), 13C NMR
spectrum of 2a(PF6)2 (C5Me5 patterns; Figure S5), ESR
spectrum of the dimer 2a(PF6)2 in CH2Cl2 at 67 K (Figure S6),
and a CIF file for 2a(PF6)2. This material is available free of
charge via the Internet at http://pubs.acs.org. Full details of the
structure determination have also been deposited with the
Cambridge Crystallographic Data Centre as CCDC 916647.
Copies of this information may be obtained free of charge from
The Director, CCDC, 12 Union Street, Cambridge CB2 1EZ,
U.K. (fax, +44-1223-336-033; e-mail, deposit@ccdc.cam.ac.uk).
(11) (a) Weyland, T.; Ledoux, I.; Brasselet, S.; Zyss, J.; Lapinte, C.
Organometallics 2000, 19, 5235−5237. (b) Samoc, M.; Gauthier, N.;
Cifuentes, M. P.; Paul, F.; Lapinte, C.; Humphrey, M. G. Angew.
Chem., Int. Ed. 2006, 45, 7376−7379. (c) Wong, K. M.-C.; Lam, S. C.-
́
F.; Ko, C.-C.; Zhu, N.; Yam, V. W.-W.; Roue, S.; Lapinte, C.;
Fathallah, S.; Costuas, K.; Kahlal, S.; Halet, J.-F. Inorg. Chem. 2003, 42,
7086−7097.
(12) Paul, F.; Costuas, K.; Ledoux, I.; Deveau, S.; Zyss, J.; Halet, J.-F.;
Lapinte, C. Organometallics 2002, 21, 5229−5235.
(13) (a) Le Narvor, N.; Lapinte, C. C. R. Acad. Sci., Paris, t. 1, Ser. IIc
1998, 745−749. (b) Hamon, P.; Justaud, F.; Cador, O.; Hapiot, P.;
Rigaut, S.; Toupet, L.; Ouahab, L.; Stueger, H.; Hamon, J.-R.; Lapinte,
C. J. Am. Chem. Soc. 2008, 130, 17372−17383.
(14) (a) Lohan, M.; Ecorchard, P.; Ruffer, T.; Justaud, F.; Lapinte,
̈
C.; Lang, H. Organometallics 2009, 28, 1878−1890. (b) Lohan, M.;
Justaud, F.; Lang, H.; Lapinte, C. Organometallics 2012, 31, 3565−
3574. (c) Xu, G. L.; Crutchley, R. J.; DeRosa, M. C.; Pan, Q.-J.; Zhang,
H.-X.; Wang, X.; Ren, T. J. Am. Chem. Soc. 2005, 127, 13354−13365.
(d) Semenov, S. N.; Blacque, O.; Fox, T.; Venkatesan, K.; Berke, H. J.
Am. Chem. Soc. 2010, 132, 3115−3127. (e) Bruce, M. I.; Le Guennic,
B.; Scoleri, N.; Zaitseva, N. N.; Halet, J.-F. Organometallics 2012, 31,
4701−4706. (f) Hervault, Y.-M.; Ndiaye, C. M.; Norel, L.; Lagrost, C.;
Rigaut, S. Org. Lett. 2012, 14, 4454−4457.
AUTHOR INFORMATION
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The MNERT (Ph.D. grants to A.B. and F.G.) and the
Universite Europeenne de Bretagne (UEB, travel grant to A.B.)
́ ́
(15) Lohan, M.; Justaud, F.; Roisnel, T.; Ecorchard, P.; Lang, H.;
Lapinte, C. Organometallics 2010, 29, 4804−4817.
are acknowledged for financial support. We thank the Centre
National de la Recherche Scientifique (CNRS) and the
Australian Research Council (ARC) for support of this work.
K.C., J.-F.H., and F.G. thank the French GENCI-CINES and
GENCI-IDRIS centers for high-performance computing
resources (Grant 2012-80649).
(16) Iyer, R. S.; Selegue, J. P. J. Am. Chem. Soc. 1987, 109, 910−911.
(17) Le Narvor, N.; Toupet, L.; Lapinte, C. J. Am. Chem. Soc. 1995,
117, 7129−7138.
(18) (a) Denis, R.; Toupet, L.; Paul, F.; Lapinte, C. Organometallics
2000, 19, 4240−4251. (b) Paul, F.; Mevellec, J.-Y.; Lapinte, C. Dalton
1874
dx.doi.org/10.1021/om400014g | Organometallics 2013, 32, 1866−1875

Organometallics
Article
Trans. 2002, 1783−1790. (c) Costuas, K.; Paul, F.; Toupet, L.; Halet,
J.-F.; Lapinte, C. Organometallics 2004, 23, 2053−2068. (d) Paul, F.;
da Costa, G.; Bondon, A.; Gauthier, N.; Sinbandhit, S.; Toupet, L.;
Costuas, K.; Halet, J.-F.; Lapinte, C. Organometallics 2007, 26, 874−
896. (e) Paul, F.; Ellis, B. E.; Bruce, M. I.; Toupet, L.; Roisnel, T.;
Costuas, K.; Halet, J.-F.; Lapinte, C. Organometallics 2006, 25, 649−
665. (f) Fox, M. A.; Roberts, R. L.; Khairul, W. M.; Hartl, F.; Low, P. J.
J. Organomet. Chem. 2007, 692, 3277−3290. (g) Grime, R. W.;
Helliwell, M.; Hussain, Z. I.; Lancashire, H. N.; Mason, C. R.;
McDouall, J. J. W.; Mydlowski, C. M.; Whiteley, M. W. Organometallics
2008, 27, 857. (h) Roberts, H. N.; Brown, N. J.; Edge, R.; Lewin, R.;
Collison, D.; Low, P. J.; Whiteley, M. W. Organometallics 2011, 30,
3763. (i) Schauer, P. A.; Low, P. J. Eur. J. Inorg. Chem. 2012, 390.
(39) Varret, F.; Mariot, J.-P.; Hamon, J.-R.; Astruc, D. Hyperfine
Interact. 1988, 39, 67−81.
(40) Gutlich, P.; Link, R.; Steinhauser, H. G. Inorg. Chem. 1978, 17,
̈
̈
2509−2514.
̂
(41) Roger, C.; Hamon, P.; Toupet, L.; Rabaa, H.; Saillard, J.-Y.;
Hamon, J.-R.; Lapinte, C. Organometallics 1991, 10, 1045−1054.
(42) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877−910.
(43) Sheldrick, G. M. Acta Crystallogr. 2008, A64, 112.
(44) (a) ADF2010.02, SCM; Theoretical Chemistry, Vrije
Universiteit, Amsterdam, The Netherlands, 2010; http://www.scm.
com. (b) te Velde, G.; Bickelhaupt, F. M.; van Gisbergen, S. J. A.;
Fonseca Guerra, C.; Baerends, E. J.; Snijders, J. G.; Ziegler, T. J.
Comput. Chem. 2001, 22, 931−967. (c) Fonseca Guerra, C.; Snijders, J.
G.; te Velde, G.; Baerends, E. J. J. Theor. Chem. Acc. 1998, 99, 391−
403.
́
(19) (a) Paul, F.; Toupet, L.; Thepot, J.-Y.; Costuas, K.; Halet, J.-F.;
Lapinte, C. Organometallics 2005, 24, 5464−5478. (b) Brown, N. J.;
Collinson, D.; Edge, R.; Fitzgerald, E. C.; Helliwell, M.; Howard, J. A.
K.; Lancashire, H. N.; Low, P. J.; McDouall, J. J. W.; Raftery, J.; Smith,
C. A.; Yufit, D. S.; Whiteley, M. W. Organometallics 2010, 29, 1261−
1276.
(45) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 1200−
1211.
(46) Becke, A. D. Phys. Rev. 1988, A38, 3098−3200. Perdew, J. P.
Phys. Rev. B 1986, 33, 8822−8824.
(47) Verluis, L.; Ziegler, T. J. Chem. Phys. 1988, 88, 322−328.
(48) (a) Pye, C. C.; Ziegler, T. Theor. Chem. Acc. 1999, 101, 396−
408. (b) Klamt, A. J. Phys. Chem. 1995, 99, 2224−2235.
(49) Grimme, S.; Anthony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys.
2010, 132, 154104.
(20) Beddoes, R. L.; Bitcon, C.; Ricalton, A.; Whiteley, M. W. J.
Organomet. Chem. 1989, 367, C21−C24.
(21) (a) Bruce, M. I.; Low, P. J.; Costuas, K.; Halet, J.-F.; Best, S. P.;
Heath, G. A. J. Am. Chem. Soc. 2000, 122, 1949−1962. (b) Lancashire,
N. N.; Brown, N. J.; Carthy, L.; Collison, D.; Fitzgerald, E. C.; Edge,
R.; Helliwell, M.; Holden, M.; Low, P. J.; McDouall, J. J. W.; Whiteley,
M. W. Dalton Trans. 2011, 40, 1267.
(50) ADF-GUI 2010, SCM; Theoretical Chemistry, Vrije Universiteit,
Amsterdam, The Netherlands, 2010; http://www.scm.com.
(22) Brown, N. J.; Collison, D.; Edge, R.; Fitzgerald, E. C.; Low, P. J.;
Helliwell, M.; Ta, Y. T.; Whiteley, M. W. Chem. Commun. 2010, 46,
2253−2255.
(23) Bruce, M. I.; Burgun, A.; Gendron, F.; Grelaud, G.; Halet, J.-F.;
Skelton, B. W. Organometallics 2011, 30, 2861−2868.
(24) Gendron, F.; Burgun, A.; Skelton, B. W.; White, A. H.; Roisnel,
T.; Bruce, M. I.; Halet, J.-F.; Lapinte, C.; Costuas, K. Organometallics
2012, 31, 6796−6811.
(25) Rigaut, S.; Le Pichon, L.; Daran, J.-C.; Touchard, D.; Dixneuf, P.
H. Chem. Commun. 2001, 1206−1207.
(26) Bruce, M. I.; Costuas, K.; Gendron, F.; Halet, J.-F.; Jevric, M.;
Skelton, B. W. Organometallics 2012, 31, 6555−6566.
(27) Semenov, S. N.; Blacque, O.; Fox, T.; Venkatesan, K.; Berke, H.
Angew. Chem., Int. Ed. 2009, 48, 5203.
(28) Argouarch, G.; Thominot, P.; Paul, F.; Toupet, L.; Lapinte, C. C.
R. Chim. 2003, 6, 209−222.
(29) Bruce, M. I. Chem. Rev. 1998, 98, 2797−2858.
(30) (a) Bonniard, L.; Kahlal, S.; Diallo, A. K.; Ornelas, C.; Roisnel,
T.; Manca, G.; Rodrigues, J.; Ruiz, J.; Astruc, D.; Saillard, J.-Y. Inorg.
Chem. 2011, 50, 114−124. (b) Fitzgerald, E. C.; Ladjarafi, A.; Brown,
N. J.; Collison, D.; Costuas, K.; Edge, R.; Halet, J.-F.; Justaud, F.; Low,
P. J.; Meghezzi, H.; Roisnel, T.; Whiteley, M. W.; Lapinte, C.
Organometallics 2011, 30, 4180−4195.
(31) Roberts, H. R.; Brown, N. J.; Edge, R.; Fitzgerald, E. C.; Ta, Y.
T.; Collison, D.; Low, P. J.; Whiteley, M. W. Organometallics 2012, 31,
6322−6335.
(32) Gauthier, N.; Tchouar, N.; Justaud, F.; Argouarch, G.; Cifuentes,
M. P.; Toupet, L.; Touchard, D.; Halet, J.-F.; Rigaut, S.; Humphrey, M.
G.; Costuas, K.; Paul, F. Organometallics 2009, 28, 2253−2266.
(33) Gendron, F.; Costuas, K.; Halet, J.-F. Unpublished results.
(34) Mahias, V.; Cron, S.; Toupet, L.; Lapinte, C. Organometallics
1996, 15, 5399−5408.
(35) Argouarch, G.; Thominot, P.; Paul, F.; Toupet, L.; Lapinte, C. C.
R. Chim. 2003, 6, 209−222.
(36) (a) Coat, F.; Guillemot, M.; Paul, F.; Lapinte, C. J. Organomet.
Chem. 1999, 578, 76−84. (b) Guerchais, V.; Lapinte, C. J. Chem. Soc.,
Chem. Commun. 1986, 663−664. (c) Roger, C.; Toupet, L.; Lapinte,
C. J. Chem. Soc., Chem. Commun. 1988, 713−715.
(37) Guillaume, V.; Thominot, P.; Coat, F.; Mari, A.; Lapinte, C. J.
Organomet. Chem. 1998, 565, 75−80.
(38) Guillaume, V.; Mahias, V.; Mari, V.; Lapinte, C. Organometallics
2000, 19, 1422−1426.
1875
dx.doi.org/10.1021/om400014g | Organometallics 2013, 32, 1866−1875