Chemistry Reference
In-Depth Information
close to the metals to reduce the strain energy of the cage, for the Sc
3
N@
D
3h
-C
78
the most reactive bonds are far from the metals. Bond distances, pyramidalization
angles or LUMOs shapes are also useful to analyze the DA regioselectivity of the
fullerenes and EMFs. Nevertheless, our results clearly demonstrate that none of them
alone predicts systematically the bond for which the attack is most favorable.
The enormous effect of the metal cluster on the DA regioselectivity is also clearly
shown in the comparison of exohedral reactivity of Sc
3
N, Lu
3
N, and Gd
3
N encapsu-
lated in both the
D
5h
-C
80
and the
I
h
-C
80
fullerenic cage. This theoretical investigation
presented the difficult handicap of taking into account the free rotation of the clusters
inside the cage. Our calculations clearly show that the preference for the [6, 6] addi-
tion increases with the size of the metallic cluster. The origin of this regioselectivity
change is found in the decrease of the DA [6, 6] activation barrier induced by the
strain energy of the Gd
3
N@
I
h
-C
80
fullerenic cage. Another very important result
of this study is the correction of the wrong experimental assignation based on
1
H
NMR spectra of
D
5h
-
5
as the most stable thermodynamic Prato adduct. Both, the
theoretical reaction energies and simulations of
1
H NMR spectra clearly show that
the most thermodynamically stable Prato cycloaddition product is the one obtained
through attack to the [5, 6]
b
bond.
In order to have a complete guide on the influence of metal clusters on the DA
regioselectivity of
I
h
-C
80
EMFs we have studied the DA cycloaddition between s-
cis
-1,3-butadiene and Sc
3
N, Lu
3
N, Y
3
N, La
2
,Y
3
,Sc
3
C
2
,Sc
4
C
2
,Sc
3
CH, Sc
3
NC,
Sc
4
O
2
, and Sc
4
O
3
I
h
-C
80
EMFs. Taking into account the free rotation of the clus-
ters, a rigorous study of the regioselectivity of these eleven EMFs was only possible
because we used the Frozen Cage Model (FCM). In the first phase, the FCM deter-
mines the most stable adducts at a very low computational cost. In the second phase
a number of selected adducts are optimized without any restriction. Our systematic
study confirmed that the HOMO- LUMO gaps, the charge transfer, and the cluster
volume are three key factors that rule the DA regioselectivity. The reactivity decrease
when the HOMO-LUMO gaps and charge transfer increases. And the increase of the
cluster volume decreases the regioselectivity, increasing the preference for the [6, 6]
adducts. Nevertheless, as pointed above, none of these parameters has a general pre-
dictive power since there are always exceptions. For instance, opposing the general
trend, the metallic oxide EMFs increase their regioselectivity when the cluster size
increases.
A second main goal of this chapter was to remark the essential role of dispersion
interactions to accurately reproduce the experimental reactivity of the EMFs. The
inclusion of dispersion effects in the DFT simulations decrease the errors in the
reaction enthalpy from 13.2 kcal mol
−
1
to only 2.2 kcal mol
−
1
. The dispersion
corrections were also the key factor to correct the wrong assignation of the DA
addition of cyclopentadiene on bond
19
of La@
C
2v
-C
82
. Our results show that for
both the cyclopentadiene and the 1, 2, 3, 4, 5-pentamethylcyclopentadiene, the most
stable DA adduct is the attack on the
o
bond. Furthermore, our DFT study including
the dispersion effects explain that the origin of the major stability of the Cp
∗
adduct
is due to the larger dispersion interactions present in Cp* as compared to Cp.
