Chemistry Reference
In-Depth Information
polarization set is taken as an outlier. Additionally, variations of more than 20%
were found for topological data obtained with different functionals or the MP2
method. Similar trends were seen for all bonds. A somewhat better convergence
was observed for the S-C bond but also in this case the deviations from the
experimental data are quite large.
Having this strong dependency in mind, it is not astonishing that the topological
data of experimental EDs deviate significantly from the theoretical ones. Indeed, as
mentioned before, the Laplacians are very sensitive with respect to the theoretical
level. However, the same is also observed for the ED values which usually show
much weaker dependencies on basis set or method. In all cases, the experimentally
derived densities at the BCPs were found to lie above the theoretical ones. This
indicates that in the bonding region the computed density is considerably smaller
than the experimental one. It is interesting to note that a similar behavior is found at
the position of the nuclei [
50
]. Nevertheless, the slow convergence indicates
problems in the computations [
24
,
46
-
49
].
The slow convergence of the bond topological data may be due to the compli-
cated electronic structures of the investigated molecules. However, this may not be
the case, as other molecular parameters such as the computed bond distances
converge as expected. The difference between experiment and theory could stem
from the influence of the crystal leading, for example, to slightly different bond
distances. To check this possibility, the bond topological data for the experimental
geometries were computed. The values are given in Table
5
for the S
N double
bond of compound 4. Also in this case, a very slow convergence with respect to the
basis set was found. It is important to note that the deviation from the experiment is
even larger if the experimental geometry is used. The difference in the position of
the BCPs is quite remarkable. While the experimental BCPs are located almost
in the middle of the bond,
d
(N)/
d
(S)
¼
1, theory predicts them to be much closer to
the sulfur center,
d
(N)/
d
(S)
1.5. In a similar case, Gatti and Bianchi improved the
agreement between experiment and theory by computing the properties at the same
location (e.g., at the experimental BCP) [
86
]. Similar results were obtained by
Table 5 Bond topological properties at the BCP of the formal S
¼
N double bond of compound 4
with R
¼
t
Bu (S(N
t
Bu)
0
) computed at the experimental geometry
2
r
r
r
l
l
l
e
d
N
ðÞ
d
S
ðÞ
d
N
ðÞ
d ðÞ
1
2
3
STO-3G
1.49
21.67
5.83
3.53
31.04
0.65
0.93
0.59
1.58
SV
1.69
5.75
7.86
6.52
20.13
0.21
0.90
0.61
1.47
6-31G(d,p)
1.88
10.61
11.00
7.30
28.92
0.51
0.92
0.59
1.57
6-311G(d,p)
1.90
8.09
11.15
7.51
26.75
0.49
0.92
0.59
1.56
6-311G(2d,p)
1.94
7.45
12.09
8.20
12.84
0.47
0.89
0.62
1.43
1.06
12.09
8.18
6-311G(3d,p)
1.93
19.21
0.48
0.90
0.61
1.47
6-311++G(d,p)
1.90
8.64
11.76
7.54
27.36
0.56
0.92
0.59
1.57
Exp.
2.27
10.56
14.40
11.83
15.69
0.22
0.78
0.74
1.05
6-311++G(d,p)
a
1.95
14.28
11.26
7.64
4.48
0.47
0.78
0.74
1.05
Distances are given in (
˚
), densities in (e/
˚
3
), and second derivatives in (e/
˚
5
)
a
Bond topological values at the position of the experimental BCP
