Chemistry Reference
In-Depth Information
(Maeda et al.
2005
). This proposal was based only on the shape of the SOMO
orbital because the final La@
C
2v
-C
82
Cp product was not isolated. In a posterior
related work, 1,2,3,4,5-pentamethylcyclopentadiene (Cp*) was used as the diene in
the DA reaction, where the final product could be isolated and characterized by X-
ray crystallography indicating that the addition corresponded to the attack on bond
o
(Maeda et al.
2010
). Thus, it was surprising that relatively similar dienes (Cp and
Cp*) presented markedly different regioselectivities. Moreover, the final stabilities
of both products were found to be largely different. At 298 K, only 36 % of La@
C
2v
-
C
82
Cp* decomposes into La@
C
2v
-C
82
and Cp* after 12 h, (Maeda et al.
2010
)
while the half-life of La@
C
2v
-C
82
Cp under the same conditions is only
τ=
1.8 h
(for comparison,
1800 h for the decomposition of C
60
Cp). (Maeda et al.
2005
)
In order to analyze and to give a rational explanation for these observations,
we have performed a complete study on the regioselectivity of the process. We have
considered the thermodynamics of the DA addition between Cp and La@
C
2v
-C
82
for
all 35 nonequivalent bonds, and for the 10 most favored cases, we have also studied
the kinetics of the reaction (Garcia-Borràs et al.
2013a
). For the Cp* case, we have
considered four different additions based on the previously reported experimental
X-ray data (Maeda et al.
2010
).
In Table
4.3
, we report the electronic and Gibbs reaction energies and reaction
(retro)-barriers for the DA addition of Cp and Cp* on La@
C
2v
-C
82
. Our results
indicate that both cycloadditions present similar reactivities, with the thermodynamic
most stable product being the one corresponding to the attack over bond
o
in the Cp*
case but also when the Cp is considered. In the present case, we have not found
good correlations between the shape of the LUMO La@
C
2v
-C
82
EMF orbitals and
the final reactivity of the bonds (see Fig.
4.11
and Table
4.3
), or even with the C-C
bond distances or pyramidalization angles as expected from our previous studies (see
previous subsections).
Nevertheless, we have to mention that from the kinetic point of view, the reaction
barriers found for additions on bond
o
and bond
11
are very close in energy (Cp:
G
‡
τ=
20.2 kcal mol
−
1
and
G
‡
19.5 kcal mol
−
1
for bond
o
and
11
, respectively;
=
=
Cp*:
G
‡
10.1 kcal mol
−
1
for bond
o
and
11
, respec-
tively), presenting the addition on bond
11
with the lowest reaction barrier for the
Cp case. However, the attack over bond
11
is very endergonic and, consequently,
once product
11
is formed rapidly reverts back to the original reactants. Thus, the
reaction presents a clearly regioselective formation of the
o
products for both Cp
and Cp* cases. Our results correct the previous wrong assignment for the La@
C
2v
-
C
82
Cp adduct, indicating that there exist no regioselective differences between the
DA cycloaddition of both Cp and Cp* over La@
C
2v
-C
82
EMF.
Finally, we have investigated the different product stabilities. As mentioned previ-
ously, it was found experimentally that the stabilities of the Cp and Cp* adducts were
significantly different, with the decomposition of La@
C
2v
-C
82
Cp being one order of
magnitude faster than that of La@
C
2v
-C
82
Cp*. We have shown that electronic effects
of the methyl substituents do not play a major role in the different Cp and Cp* product
stabilities. If we compare the retro reaction barriers (
G
‡
-retro
9.8 kcal mol
−
1
and
G
‡
=
=
10.2 kcal mol
−
1
=
12.4 kcal mol
−
1
for Cp, see Table
4.3
), one could conclude
for Cp* and
G
‡
-retro
=
