Chemistry Reference
In-Depth Information
4
-C
{CH
2
}
3
)-(CO)
3
) are close to structural instability; that is, new structures may
occur even for small geometry changes. By decreasing the Fe-C
a
-C
b
angle from
the optimized equilibrium value of 76.6
down to 73
, while retaining the original
C
3v
symmetry, three new bond paths linking iron to the C
b
carbons appear [
72
].
Despite this evident structural change - namely from a
The experimental and the gas-phase ab initio geometries of
(Fe(
4
complex,
according to the bond path criterion - the (Fe,C
a
) and the (Fe,C
b
) delocalization
indices are hardly changed, as they decrease to 0.351 and increase to 0.609,
respectively. Analogously, no significant change is observed in the SF pattern
when the rp is taken at the Fe-C
b
bcp in the distorted geometry, rather than at the
Fe-C
b
mp in the minimum energy structure (Fig.
11c
). Overall, due to the proxim-
ity of the optimized geometry to structural catastrophe, the SF contributions look
almost the same, whether a Fe-C
b
bcp is present or lacking. In both cases, the
contributions from the C
b
atoms to the density at the rp are greater than observed
from the C
a
atom. It is thus gratifying that the delocalization indices and the SF
provide pictures of bonding in this complex, which nicely fit within each other and
also with the other available experimental evidences. Clearly the occurrence or lack
of a bond path is also a physical outcome, and the bond path criterion should not be
ignored because of its apparently contrasting conclusions. Simply, the nonlocal
descriptors seem more appropriate for describing an inherently nonlocal interaction
such as that of TMs with the delocalized
1
to a
-hydrocarbyl ligands, as pointed out
earlier for other interactions characterized by flat energy surfaces and electron
densities.
Similar conclusions were drawn by Farrugia et al. in their combined experimen-
tal and theoretical charge density study [
83
] of three different metallocenes (
p
5
-
5
-
C
5
H
5
)
2
Fe
2
. For the sake of conciseness, we focus here only on the first of such
compounds whose experimental molecular graph and SF percentage sources for a
number of representative rps are shown in Fig.
12
. The molecular graph (Fig.
12a
)
shows only four Mn-C bond paths, one less than expected on the basis of formal
hapticity. Curiously enough, the bond path is missing for the shortest (Mn-C7)
rather than for the longest Mn-C
ring
distance.
The Mn-C5 bond path is rather curved, and its bcp is 0.02
˚
close to the
C4-C5-Mn rcp, indicating that these two CPs are proximal to coalesce to yield a
structural evolution. Indeed, this bond path was even not recovered with some of
the investigated multipole models, whereas the theoretical density for the gas-
phase-optimized structure exhibits the number of bond paths expected from the
formal hapticity of the complex [
83
]. Similar discrepancies were found also for the
other studied systems, which illustrate once more the difficulties encountered when
using a bond path approach to characterize the delocalized metal-(
6
-C
6
H
6
)Cr(CO)
3
, and (
E
)-{(
5
-C
5
H
4
)CF
5
-C
5
H
4
)}(
C
5
H
5
)Mn(CO)
3
,(
¼
CF(
p
-hydrocarbyl
ligands) bonding.
The experimental SF patterns shown in Fig.
12b-f
are instead very similar to
those obtained from the ab initio density (Fig. 9 of [
83
]) and visibly distinguish the
quite different types of chemical interactions present in the complex. The SF
contributions to Mn-C(O) bcp (Fig.
12b
) are typical for such interaction and very
