carbene ligand to disrupt the metal–metal bond forming
process. However, utilizing more sterically demanding NHCs
resulted in no reaction with [Fe(cot)2], while using a bulky
CAAC,[22] an outer-sphere [4+1] cycloaddition was observed
(Figure 2). This unusual and slow reaction (458C, 2 d) is
attributed to the increased nucleophilicity and electrophilicity
of CAACs.[6a,b]
Figure 1. Reaction aNHC with [Fe(cot)2], and crystal structure of the
resulting [(aNHC)Fe2(cot)2] complex 2aNHC (displacement ellipsoids
drawn at the 50% probability level; protons and isopropyl groups
omitted for clarity). Red, Fe; blue, N; gray, C.
Figure 2. Reaction of a CAAC with [Fe(cot)2], and crystal structure of
the resulting outer-sphere [4+1] cycloadduct 5 (displacement ellip-
soids drawn at the 50% probability level; protons omitted for clarity).
À
Red, Fe; blue, N; gray, C. Range of Fe C distances (2.0247(10)–
2.2058(9) ꢀ) for coordinated carbons.
which over a 24 h period gives rise to large black crystals.
Regardless of the order of addition of reagents, temperature,
or equivalents of aNHC used, compound 2aNHC was the only
Notably, in contrast with the parent [Fe(cot)2], which
readily decomposes around 1008C, C6D6 solutions of the
diamagnetic iron(0) complex 5 are thermally stable for at
least 2 d at 1208C.
1
species obtained. The H NMR spectrum of the crystalline
precipitate, redissolved in [D6]benzene revealed broad reso-
nances ranging from d = + 7.4 to À1.22 ppm, suggestive of a
paramagnetic species. In solution the new compound exhibits
an effective magnetic moment of 2.64 B.M., which is close to
the spin-only value for two unpaired electrons (2.83 B.M.).
Gratifyingly, a single-crystal X-ray diffraction study[16]
confirmed that the new species was indeed an [(aNHC)Fe2-
(cot)2] complex. The metal–aNHC bond length (2.082(2) ꢁ)
is in the upper range (1.895–2.111 ꢁ) of reported values for
related iron carbenes.[17] Interestingly, the non-bridging cot
ligand of 2aNHC is h4-coordinated to the iron center, with the
Considering the previous results we sought to take a
different approach to prepare mononuclear iron–carbene
complexes related to [(L)Fe(cot)]. We predicted that if a
carbene of sufficiently small size reacted with [Fe(cot)2], a
simple addition product, [(L)Fe(cot)2], could be obtained
without dislodging a cot ligand. Owing to its unique and non-
sterically demanding shape, the carbocyclic carbene, N,N-
bis(diisopropyl)aminocyclopropenylidene (BAC)[23] seemed
like an ideal candidate. Thus, one equivalent of BAC was
carefully added at À788C over 30 min to a pentane solution of
À
other half of the non-coordinated eight-membered ring (C2
1
À
Fe2 3.085(2) ꢁ, C5 Fe2 3.262(2) ꢁ) displaying a localized
[Fe(cot)2] (Figure 3). The H NMR spectrum of the reaction
À
À
butadiene structure (C2 C3 1.342(3) ꢁ, C3 C4 1.410(3) ꢁ,
C4 C5 1.350(3) ꢁ). Such a bonding arrangement indicates
mixture revealed the complete conversion of [Fe(cot)2] into a
new complex, which exhibits resonances corresponding to two
fluxional and equivalent cot ligands, as well as a single BAC
moiety. A single-crystal X-ray diffraction study confirmed the
compoundꢀs identity as the [(BAC)Fe(cot)2] complex 6.
Although the quality of crystals precludes a detailed dis-
cussion of the bond lengths and angles, it is evident that this
diamagnetic complex adopts a square-pyramidal structure
with both cot ligands bound in an h4-fashion.
Even after several days [(BAC)Fe(cot)2] does not react
with [Fe(cot)2] to form bimetallic Fe complexes 2. This
suggests that the dissociation of the first cot ligand to form
[(L)Fe(cot)] complexes is sterically controlled, and is a key
step in the stoichiometric and catalytic aggregation of [Fe-
(cot)2]. Reinforcing this hypothesis is the fact that another
molecule of the relatively small BAC displaces a cot ligand
from [(BAC)Fe(cot)2] at room temperature to afford the
[(BAC)2Fe(cot)] species 7. A single-crystal X-ray diffraction
study of paramagnetic (meff = 2.87 B.M.) complex 7 indicates
that it adopts a tetrahedral geometry, as expected for a high-
À
that in the solid state this [(L)Fe2(cot)2] complex exists in the
non-reduced form. This observation suggests that bulky
carbene ligands, such as SIPr and aNHC, may not only
kinetically protect [(L)Fe2(cot)2] fragments, but also disfavor
electron transfer from iron to the non-bridging cot ligand
through steric disruption of the required higher hapticity of
the reduced cot ligand.[18] The other cot ligand of 2aNHC is h5-
coordinated to both Fe centers forming a very symmetrical
À
À
bridge between the two iron atoms (Fe1 C1 2.363(3) ꢁ, Fe2
C1 2.325(2) ꢁ), which indicates that the classical bis(p-allyl)
two-electron–three-center bonding model[20] of cot is appli-
À
cable in this case. The Fe1 Fe2 distance of 2.4530(5) ꢁ argues
for an iron–iron single bond,[21] which gives the two iron
centers an overall 15 and 17 electron count, respectively. With
a stable [(L)Fe2(cot)2] fragment in hand, we became inter-
ested in exploring the possibility of preparing a monomeric
[(L)Fe(cot)] complex, reminiscent of proposed intermediates
1. The obvious approach is to implement an even more bulky
Angew. Chem. Int. Ed. 2011, 50, 268 –271
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
269