Chemistry Reference
In-Depth Information
Hibbs et al. [
19
]. In these cases an improved agreement between theoretical and
experimentally derived densities and Laplacians were observed. Table
5
shows a
similar trend.
The strong deviations between experiment and theory may also result from the
replacement of sterically demanding substituents, which are often replaced by
smaller groups to lower the computational cost. These substituents are important
for the kinetic stability of the substances. It is generally assumed that the electronic
structure is not influenced by these groups. For less sensitive properties this
assumption is surely justified. For the more sensitive ones such as the Laplacian
at the BCP of a polar bond, it has to be tested whether this simplification within
theory is still justified. For the present model systems, the bond topological proper-
ties were found to be weak functions of the substitution patterns, as shown in
Table
6
for the formal S
¼
N double bonds of S(NR)
3
. Compound 3 was selected
as an example since the formal double bonds showed stronger sensitivities with
respect to basis set size and method. Similar variations were obtained for all other
compounds [
50
]. Using the 6-311G(d,p) basis set, the Laplacian varies in the series
R
t
Bu from 3.54 to 3.95 e/
˚
5
. This variation is smaller than the
one resulting from changes in the basis set. Also the ratio
d
(N)/
d
(S) remains nearly
unaffected. This indicates that the differences in the experimental and theoretical
bond topological properties do not result from the different substituents. A replace-
ment of bulky substituents by smaller ones (e.g., R
¼
H, R
¼
Me, and R
¼
¼
Me) does not overly influence
the results. For the present cases, even R
¼
H seems to be a good approximation for
many cases [
50
].
The strong dependence of the bond topological data on the basis sets may be a
result of the fact that the atomic basis sets are centered at the atoms. Hence, the
regions around the nuclei are better described than the bonding region where the
BCP is located. However, as can be seen from Fig.
2
, a different problem seems
to exist in the present case. The strong dependence of the
l
3
values on the method
of calculation seems to be connected with the position of the BCP.
d
(N)/
d
(S)
values around 1.5 were computed for the two formal S
¼
N double bonds, S1-N2
¼
of compound 1 (R
Me, Table
5
) and S1-N1 of compound 3 (Table
4
), i.e., the
BCP is located considerably closer to the sulfur than to the nitrogen center.
Table 6 Parameters derived from a linear regression between theoretically and experimentally
derived bond topological properties at the BCP,
2
(
r
BCP
) vs. bond distance
d
r
(
r
BCP
), and
r
-
Calc.
r
¼
a
d
þ
b
2
r
r
¼ a d þ b
R
2
R
2
-
a
b
a
b
Experiment
-
2.72
6.38
0.91
16.78
38.47
0.11
opt
a
B3PW91/6-311G(d,p)
2.02
4.94
0.98
97.38
153.03
0.870
sp
b
B3PW91/6-311G(d,p)
2.38
5.50
0.939
28.37
45.35
0.046
sp
b
PW91/6-311++G(d,p)
2.24
5.19
0.694
91.89
144.58
0.633
For the density, eight different S-N and two S-C bonds were selected from compounds 1
4. For
the Laplacian values, only the eight different S-N bonds were taken into account
a
The optimization was performed with the methyl-substituted compounds
b
Single point (sp) calculations were performed using the experimental (solid state) geometry
(R
¼
t
Bu)
