Quantum chemical ab initio calculations at the MP2/6-31G(d) level of t
heory are reported for the beryllium-carbene complexes Be(CX(2))(n)(2) (X = H, F; n = 1-4), ClBe(CX(2))(n)(+) (X = H, F; n = 1-3), and Cl2B
e(CX(2))(n) (X = H, F; n = 1, 2). The complex ClBe(C(NH2)(2))(3)(+) ha
s also been calculated. Where feasible, the bond energies of some mole
cules are reported at MP4/6-311G(d)//MP2/6-31G(d). Analysis of the bon
ding situation with the help of the natural bond orbital method shows
that the carbene ligands are pure donors in the complexes. The dicatio
ns Be(CX(2))(n)(2+) (X = H, F; n = 1-4) have strong Be2+-C donor-accep
tor bonds. The bond strengths decrease clearly when the number of liga
nds increases from n = I to 4. The CH2 complexes have stronger Be-C bo
nds than the CF2 complexes. Yet, the CH2 complexes are chemically less
stable than the CF2 complexes for kinetic reasons. The carbon p(pi) o
rbital of methylene stays nearly empty in the complexes, which makes t
hem prone to nucleophilic attack. All theoretical evidence indicates t
hat the dominant factor which determines the chemical stability of car
benes and carbene complexes is the population of the carbon p(pi) orbi
tal. The chemical instability of the methylene complexes becomes obvio
us by the geometry optimizations of ClBe(CH2)(2)(+), ClBe(CH2)(3)(+),
Cl2Be(CH2), and Cl2Be(CH2)2, which lead to rearranged structures as en
ergy minimum forms. The C-H bonds and particularly the C-F bonds of th
e ligands are shorter than in free CH2 and CF2. The carbon atom of CF2
becomes electronically stabilized in the complexes via p(pi) donation
from fluorine. This finding suggests that carbene ligands, which are
unstable as free molecules, may become sufficiently stabilized to be i
solated even in complexes without metal --> carbene back-donation.