Chemistry Reference
In-Depth Information
At first it is troubling to find values of
O
,
C
that exceed the range predicted for pure
element - aryloxide bonds on the basis of organic structures. The long M-OAr distances
can, however, be rationalized by recognizing the presence of
-antibonding interac-
tions, a theory eloquently propounded by J.M. Mayer.
184
In the later transition metal
aryloxides the metal d orbitals with
symmetry are occupied and hence their interac-
tion with the oxygen p-orbitals are a repulsive, filled -filled interaction. As noted by
Mayer
“the substantial effect of
antibonding interactions on the stability and reactivity
of
donor ligands has not been widely appreciated. This lack of appreciation is due
in part to filled-filled interactions being only a subtle and uncommon effect in organic
chemistry.”
The structural data in Table 6.4 is strong evidence for a
-antibonding
situation. There is also evidence that the “pushing back” of the p electron density onto
the aryloxide ligands results in a more electron-rich (anionic) oxygen atom. Hence it
is common to find these aryloxides undergoing hydrogen bonding upon addition of
phenols. Structural studies show that the M-OAr bond length increases only slightly
upon formation of the adduct (Table 6.4). It should be noted, however, that despite the
presence of this
-antibonding situation there is evidence that the late transition metal
aryloxide bond strength is still substantial.
185
,
186
An interesting question is why there is no “additional” elongation of the
metal - aryloxide bond in the 18-electron compounds of niobium/tantalum and tungsten
above (Tables 6.2, 6.3). A possible answer lies in the fact that although electronically
saturated the electron density is stabilized in these molecules by the attached
-acceptor
carbonyl ligands, hence reducing the
antibonding effect. In the later transition
metal compounds the
electron density is essentially metal based with little ligand
stabilization.
Turning to the p-block metal aryloxides it was shown above that the M-OAr bond
lengths for derivatives of tin and germanium are as expected on the basis of known
organic structures,
i.e.
no apparent
-bonding. In these compounds the M-O-Ar angles
are all very close to 120
Ž
, again as expected. In contrast, mononuclear, bulky aryloxide
derivatives of aluminium and gallium have been shown to possess very short M-OAr
bond lengths and large M-O-Ar angles.
187
,
188
Using the parameter
O
,
C
applied to
mixed alkyl, aryloxides Barron
et al
. showed that values ranged from 0
.
22
˚
Ato
0
.
28 A, consistent with a significant
-interaction.
1i
Given the high energy of the
3-d orbitals on aluminium it was proposed that
-donation into the
Ł
orbital of
adjacent Al-X bonds was occurring. This hypothesis was given strong support by
theoretical studies as well as gas-phase photoelectron spectroscopy.
187
There have been
suggestions, however, that the bonding in three- and four-coordinate aluminium and
gallium aryloxides is best described as ionic in nature. This was argued to account for
the short M-OAr distances, lack of observed restricted rotation, and large M-O-Ar
angle (see below).
189
4.2.2
Terminal M-O-Ar Angles
Section 4.2.1 clearly shows a strong correlation between the extent of oxygen-p to
metal
-bonding and the M-OAr distance. However, the errors inherent in determining
interatomic distances by X-ray diffraction techniques means that only large differences
in M-OAr bond lengths can be safely analysed. In contrast, bond angles are typically
refined from diffraction data to an accuracy of less than a degree. Given the fact that
