126
B. Gierczyk et al.
Table 3. Calculated 15N chemical shifts and M u¨ lliken charges of nitrogen atoms
Calculated chemical shifts [ppm]
Calculated M u¨ lliken charges
N-4
Compound
N-3
N-4
NH
N-3
NH
1
1
1
1
1
1
1
2
2
2
2
2
2
a
b
c
d
e
f
g
a
b
c
d
e
f
296.8
293.1
294.0
295.8
296.6
299.2
301.7
295.8
295.7
296.6
296.9
299.2
300.4
348.4
349.1
349.1
349.1
347.3
347.4
347.5
336.6
342.0
346.4
349.1
359.2
362.3
112.8
110.0
110.9
112.6
112.6
115.2
117.4
112.6
113.5
113.5
114.4
115.2
114.7
ꢀ0.328
ꢀ0.349
ꢀ0.325
ꢀ0.328
ꢀ0.311
ꢀ0.338
ꢀ0.327
ꢀ0.253
ꢀ0.279
ꢀ0.327
ꢀ0.331
ꢀ0.338
ꢀ0.330
ꢀ0.012
0.045
0.076
0.067
0.030
ꢀ0.313
ꢀ0.384
ꢀ0.415
ꢀ0.401
ꢀ0.367
ꢀ0.318
ꢀ0.324
ꢀ0.330
ꢀ0.324
ꢀ0.320
ꢀ0.316
ꢀ0.318
ꢀ0.312
0.020
0.016
ꢀ0.096
ꢀ0.066
ꢀ0.009
ꢀ0.007
0.020
0.016
In this group of oxadiazoles, the worst correlation is
observed for N-4. It is a consequence of two effects. Firstly,
the chemical shift range for N-4 is very narrow, only
order of magnitude greater than the accuracy of the
N measurement, which introduces a serious error in the
performed for the molecule in vacuum, frozen at 0 K.
Consequently, such factors as interactions with the solvent
and other molecules, molecular dynamics and deformation
are not taken into account in the calculations. We have
not found correlations between the calculated M u¨ lliken
charges of nitrogen atoms with the experimental chemical
shifts. Only for the N-4 atom of the second series of
1
υ
1
5
statistical analysis. Secondly, because the substituent effect
1
5
on the υ N value is very small, other weak effects such
as solvation processes, hydrogen bonding, van-der-Waals
interaction as well ꢄ–ꢄ stacking affect the chemical shift.
Nevertheless, a distinct relation between the chemical shift
value of the N-4 signal of 1a–1g and the Hammett coefficients
is observed, in contrast to those for the other nitrogen atoms
in the two series of compounds. It is a result of the reversed
role of the effects influencing the chemical shifts. For all
nitrogen atoms in the second group of compounds (2a–2f)
and N-3 and NHnitrogen atoms in the first group (1a–1g), the
2
compounds (2a–2f), a correlation (r D 0.9) was established
which is not very good. As it comes out from our earlier
analysis of the relation with the Hammett parameters, this
signal is predominantly influenced by inductive effects of
the substituent. Correlations between the charge and the
1
3,14
chemical shifts were observed only for rigid molecules.
The possibility of rotation and deformation of disubstituted
1,3,4-oxadiazole system precludes the observation of a
simple relation between the calculated charge and spectral
parameters.
1
5
inductive interactions prevail in the changes of N chemical
shifts. For the N-4 atoms of oxadiazoles 1a–1g, the resonance
interactions are dominant. A similar relationship holds for
the C-2 atom of the heterocyclic ring. The resonance effects
cause polarization of the C N as well as N–N bonds and,
in consequence, a redistribution of the electron density over
the oxadiazole ring. It provokes the opposite changes in the
chemical shifts of C-2, N-4 and N-3, C-5 atom pairs effected
CONCLUSIONS
The 15N chemical shifts of substituted 2-(phenylamino)-5-
phenyl-1,3,4-oxadiazoles were correlated with values of the
Hammett parameters. The influence of the distance from
the substituent and the presence of coupled ꢄ-systems
were discussed. The experimental results were compared
with those of the calculations, obtaining a good correlation
but poor matching; calculated values are much higher.
1
1,12
by the substituent.
The calculated values of the 15N NMR chemical shifts as
well as electron densities on the nitrogen atoms are collected
in Table 3. The experimental results are much smaller than
1
5
No correlation between the N chemical shifts for the
compounds studied with the calculated M u¨ lliken charges
of nitrogen atoms was found.
1
5
the calculated values of υ N. The difference is about 50 ppm
for the nitrogen atoms of the heterocyclic ring and 30 ppm for
the amino ones. Regardless of the errors, the calculated values
correlate very well with the experimental chemical shifts.
The correlation parameters, calculated for the equation:
Acknowledgements
All calculations were performed at the Pozna n´ Supercomputing
Centre.
1
5
chemical shift D ꢃυcalc N C intercept, are presented in Table 2.
Although the relationships between these parameters are
linear, the susceptibility constant is not equal to 1. It
REFERENCES
1
5
1. Fra n´ ski R. Asian J. Chem. 2005; 17: 2063.
indicates that the errors in υ N calculations depend on
the electron-acceptor character of the substituents. Such
discrepancy between the calculated and experimental results
could be explained by several effects. The calculations were
2
3
. Schulz B, Bruma M, Brehmer L. Adv. Mater. 1997; 8: 601.
. Giebler R, Schulz B, Reiche J, Brehmer L, Wuhn M, Woll C,
Smith AP, Urquhart SG, Ade HW, Unger WES. Langmuir 1999;
15: 1291.
Copyright 2006 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2007; 45: 123–127
DOI: 10.1002/mrc