Comparison of the electronic structures of imine and hydrazone side-chain functionalities with the aid of 13C and 15N NMR chemical shifts and PM3 calculations. The influence of C=N-substitution on the sensitivity to aromatic substitution

Article abstract of DOI:10.1021/jo020608l

Benzaldehyde derivatives possessing a C=N double bond in the side-chain of the aromatic ring exhibit a reverse dependence of the ¹³C NMR chemical shifts of the C=N carbon on the benzylidenic substituents X. Thus, electron-withdrawing substituents cause shielding (shift is reduced), while electron-donating ones cause deshielding. The origin of this phenomenon, which is in contrast with the idea of the generalized electronic effect, is extensively studied here by comparing the behavior of sets of benzaldehyde derivatives bearing various substitutents Y on the C=N nitrogen (Y-N= CH-C₆H₄-X). The effects of substituents X on the C=N unit change when Y is varied. Combination of the influences of the substituents X and Y gives a sensitive balance between the different resonance structures of the compounds. Our graphical treatment, where the ρ_I and ρ_R values observed for substituent X are plotted against the σ_p⁺ value of substituent Y, is a novel use of Hammett-type substituent parameters. The justification of this method and our conclusions could be verified, for instance, by the fair correlation between the ρ_I or ρ_R values and the atomic charges of the imine carbon of the unsubstitued phenyl derivatives as well as by the correlations of the relevant bond orders and/or bond lengths both with the substituent parameters and with the atomic charges.

Full text of DOI:10.1021/jo020608l

Com p a r ison of th e Electr on ic Str u ctu r es of Im in e a n d Hyd r a zon e
Sid e-Ch a in F u n ction a lities w ith th e Aid of 13C a n d N NMR
Ch em ica l Sh ifts a n d P M3 Ca lcu la tion s. Th e In flu en ce of
CdN-Su bstitu tion on th e Sen sitivity to Ar om a tic Su bstitu tion
15
,
†
‡
†
§
§
†
K. Neuvonen,* F. F u¨ l o¨ p, H. Neuvonen, A. Koch, E. Kleinpeter, and K. Pihlaja
Department of Chemistry, University of Turku, FIN-20014 Turku, Finland, Institute of Pharmaceutical
Chemistry, University of Szeged, H-6701 Szeged, POB 121, Hungary, and Department of Chemistry,
University of Potsdam, D-14415 Potsdam, POB 691553, Germany
kari.neuvonen@utu.fi
Received September 16, 2002
Benzaldehyde derivatives possessing a CdN double bond in the side-chain of the aromatic ring
exhibit a reverse dependence of the 13C NMR chemical shifts of the CdN carbon on the benzylidenic
substituents X. Thus, electron-withdrawing substituents cause shielding (shift is reduced), while
electron-donating ones cause deshielding. The origin of this phenomenon, which is in contrast with
the idea of the generalized electronic effect, is extensively studied here by comparing the behavior
of sets of benzaldehyde derivatives bearing various substitutents Y on the CdN nitrogen (Y-Nd
CH-C
of the influences of the substituents X and Y gives a sensitive balance between the different
resonance structures of the compounds. Our graphical treatment, where the F and F values
value of substituent Y, is a novel use of
Hammett-type substituent parameters. The justification of this method and our conclusions could
be verified, for instance, by the fair correlation between the F or F values and the atomic charges
6 4
H -X). The effects of substituents X on the CdN unit change when Y is varied. Combination
I
R
+
observed for substituent X are plotted against the σ
p
I
R
of the imine carbon of the unsubstitued phenyl derivatives as well as by the correlations of the
relevant bond orders and/or bond lengths both with the substituent parameters and with the atomic
charges.
In tr od u ction
dence on substitution, i.e., electron-withdrawing phenyl
substituents increase the electron density on the CdN
carbon.1 As a consequence, the imine carbon resonance
reveals a reverse behavior as concerns substitution:
electron-withdrawing substituents cause shielding. In
principle, the substituent-induced changes in the chemi-
The medicinal importance of hydrazones is well estab-
lished. However, as compared with azomethines, the
electronic effects prevailing in the -N-NdCH- func-
tionality and the sensitivity of these effects to substitu-
tion are not satisfactorily understood.
a
cal shift, SCS, depend on the inductive (σ
resonance (σ ) parameters according to eq 1:
I
or σ
F
) and
R
SCS ) F σ (or F σ ) + F σ
R R
(1)
I
I
F
F
where SCS is the 13C NMR shift for a substituted
compound relative to that for the unsubstituted one. With
imines, a better correlation is obtained with this dual-
substituent parameter (DSP) approach than with any of
+
-
the single-parameter equations tested (σ, σ , σ , σ
I
F
, σ ),
but the small values of F show that resonance factors
R
Extensive 13C NMR spectroscopic and theoretical calcula-
tion data together demonstrate that the CdN carbon of
azomethines 1 derived from substituted benzaldehydes
behaves normally and is shielded (shift is reduced) with
increasing electron density on the carbon in question.1a
In contrast, the atomic charge displays a reverse depen-
1a
are not of great importance. The F
3 to -4, while the F values range only from -0.7 to
.2.
For imines, the charge on the CdN carbon and that
I
/F
F
values range from
-
0
R
1a
on the CdN nitrogen behave inversely. This suggests
(
1) (a) Neuvonen, K.; F u¨ l o¨ p, F.; Neuvonen, H.; Koch, A.; Kleinpeter,
*
Corresponding Author. Fax: +358-2-3336700.
University of Turku.
University of Szeged.
E.; Pihlaja, K. J . Org. Chem. 2001, 66, 4132. (b) Neuvonen, H.;
Neuvonen, K. J . Chem. Soc., Perkin Trans. 2 1999, 1497. (c) Neuvonen,
H.; Neuvonen, K.; Koch, A.; Kleinpeter, E.; Pasanen, P. J . Org. Chem.
2002, 67, 6995.
†
‡
§
University of Potsdam.
1
0.1021/jo020608l CCC: $25.00 © 2003 American Chemical Society
Published on Web 02/25/2003
J . Org. Chem. 2003, 68, 2151-2160
2151

Neuvonen et al.
The reverse dependence of the CdN carbon 13C NMR
SCHEME 1
chemical shift on phenyl substitution has previously been
1
a
observed in this laboratory not only for imine structures
but also for hydrazones 7-11.6 In all cases, the CdN
shift correlated poorly with Hammett-type single-param-
eter equations but well with a DSP equation. Differing
from imines, both inductive and resonance contributions
are highly significant for hydrazones.6 In the present
work, we have focused on the characterization of the
origin of the differences in substituent dependence of the
CdN carbon resonances of imines and hydrazones. A
further goal was to investigate more generally the
influence of the substitutuent (Y) on the CdN nitrogen
,7
,7
substituent-induced polarization of the CdN unit, i.e.,
π-polarization (2) and/or inherent polarization (Scheme
1
). The concept of π-polarization has been extensively
2
-5
discussed in the literature.
According to this idea, the
substituent dipole polarizes each π-unit as a localized
system, as shown in 2. π-Polarization has been used to
(
Y-NdCH-) on the sensitivity of the -CdN carbon
6 4
resonance to the phenyl substituent X (-NdCH-C H -
X).
1
3
explain the reverse behavior of the C NMR chemical
2
shifts of sp or sp hybridized carbons in several different
systems.2
,5-7
The idea of π-polarization has also been
Resu lts
criticized, but without any explanation for the reverse
We performed 15N NMR measurements on the hydra-
8
behavior of the carbon shifts. However, we recently
1
3
zone series 7 and 10, for which the C NMR data exist
proposed an alternative mechanism to account for the
in our previous work,6 and also on the imine series 12.
,7
1
behavior of unsaturated aromatic side chains.
1
3
For the latter series, the C NMR shifts were also
recorded. Additionally, we report the PM3 atomic charges
and bond lengths for the same compounds and also for
hydrazones 13 and 14.
The substituent-sensitive inherent polarization of the
-
CdN- unit (Scheme 1) of imines was proposed as an
alternative/competitive explanation for π-polarization.
The inductive and resonance effects of substituents X on
the stabilities of the different resonance forms 3-6 were
1
a
then considered. According to this concept, electron-
withdrawing substituents decrease the contribution of
the polarized form 4 through inductive destabilization,
increasing the electron density on the CdN carbon as a
consequence, while electron-donating substituents have
an opposite effect, with negative F
F
values as a result.
Conjugative effects of the substituents seem to be
negligible.1 The small negative F
a
values were explained
R
by resonance-induced polar effects (see structure 5).
For series 13 and 14, the literature NMR data were
used.9 The NMR and atomic charge data are given in
Supporting Information in Tables S2-S6. The bond
length data are collected in Supporting Information in
Table S7 and bond lengths for some relevant compounds
are given in Table 4.
,12
Discu ssion
Table 1 includes F
hydrazones, determined in our previous works
this study, or values either found in the literature or
I
or F
F
R
and F values for imines and
1
a,6,7
or in
5
1
3
calculated by us using C NMR chemical shifts observed
earlier.6 Both the inductive and resonance correlation
,7,9
(
5) Bromilow, J .; Brownlee, R. T. C.; Craik, D. J .; Fiske, P. R.; Rowe,
J . E.; Sadek, M. J . Chem. Soc., Perkin Trans. 2 1981, 753.
6) Neuvonen, K.; F u¨ l o¨ p, F.; Neuvonen, H.; Pihlaja, K. J . Org. Chem.
994, 59, 5895.
7) Neuvonen, K.; F u¨ l o¨ p, F.; Neuvonen, H.; Simeonov, M.; Pihlaja,
K. J . Phys. Org. Chem. 1997, 10, 55.
(
1
(
(
(
8) Dahn, H.; Carrupt, P.-A. Magn. Reson. Chem. 1997, 35, 577.
9) Gordon, M. S.; Sojka, S. A.; Krause, J . G. J . Org. Chem. 1984,
(
2) Craik, D. J .; Brownlee, R. T. C. Prog. Phys. Org. Chem. 1983,
49, 97.
(10) Arrowsmith, J . E.; Cook, M. J .; Hardstone, D. J . Org. Magn.
1
4, 1.
(
(
3) Reynolds, W. F. Prog. Phys. Org. Chem. 1983, 14, 165.
4) (a) Hamer, G. K.; Peat, J . R.; Reynolds, W. F. Can. J . Chem.
Reson. 1978, 11, 160.
(11) Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165.
(12) Westerman, P. W.; Bottto R. E.; Roberts, J . D. J . Org. Chem.
1978, 43, 2590.
1
973, 51. 897. (b) Hamer, G. K.; Peat, J . R.; Reynolds, W. F. Can. J .
Chem. 1973, 51, 915.
2
152 J . Org. Chem., Vol. 68, No. 6, 2003

Influence of CdN-Substitution on Aromatic Substitution
TABLE 1. GI/GF a n d GR Va lu es Obta in ed by DSP
1
3
Tr ea tm en t of th e CdN Ca r bon C NMR Ch em ica l Sh ifts
in CDCl3 for th e pa r a -Su bstitu ted Im in e Ser ies 12 a n d
1
5-19 a n d for th e Hyd r a zon e Ser ies 7-11, 13, a n d 14
line
series
FI or FF
FR
shift range
ref
1
2
3
4
5
6
7
8
9
0
1
2
3
15
16
17
18
12
19
13
14
7
8
9
10
11
-3.7
-3.0
-3.6
-3.2
-4.0
-3.6
-4.6
-5.5
-6.2
-5.8
-4.2
-4.6
-3.8
-0.1
0.2
157.16-159.99
158.79-160.80
159.43-161.93
158.39-160.57
157.33-160.34
159.8-162.03
133.75-138.64
138.55-143.26
127.14-133.93
127.46-134.34
144.4-148.5
1a
1a
1a
1a
-0.5
-0.1
-0.7
-0.7
-3.1
-2.7
-4.9
-5.0
-2.0
-1.8
-1.6
this work
a
5
6
6
b
b
6
6
7
7
7
1
1
1
1
135.8-140.4
c
144.1-147.9
a
13C NMR data from ref 10. b 13
C NMR data from ref 9. c In
F IGURE 1. Plot of the 15N NMR chemical shifts of the CdN
DMSO-d6.
nitrogen for series 7 in CDCl
constants σ.
3
vs Hammett substituent
parameters of the hydrazones are higher than those
observed for the imines. The significantly higher contri-
bution of resonances with negative F values for the
R
hydrazones is especially noteworthy. The results indicate
the higher sensitivity of the polarization of the CdN unit
of the hydrazones to substitution as compared with that
of the imines or higher shift:charge ratios for the former
compounds. Moreover, the CdN carbon resonances of the
hydrazones appear at a field higher by ca. 15-30 ppm
than those for the imines.
+
by σ/σ , increases. The q
(
C
(CdN) values of the hydrazones
Table 2; slopes -0.044, -0.032, -0.046, and -0.043 for
lines 15, 26, 37, and 47, respectively) possess a clearly
higher sensitivity toward substitution than do the
(CdN) values of the imines (-0.027, -0.029, and
0.026 for lines 1, 5, and 9, respectively). In contrast with
the reverse behavior of the CdN carbon, the slopes for
the correlations of the charges of the CdN nitrogen or
N2 are positive in all cases (lines 3, 7, 11, 18, 20, 29, 39,
1, 49, and 51), with the exception of the charge of N2 of
-aminobenzoylhydrazones 10. In the latter case, nega-
q
C
-
4
2
+
tive slopes are obtained for both σ and σ correlations
(lines 31 and 32), but the values are so small that they
in fact indicate that the atomic charge on N2 does not
depend on the phenyl substituent. For the other series
studied, the electron density on both nitrogens decreases
when the phenyl substituents become more electron-
withdrawing, q
tive to substitution as q
N-methylhydrazones 7, the slope of the q
0.030) is not significantly higher than the corresponding
slopes for the imines 15, 20, and 12 (0.029, 0.026, and
N
(CdN) being about 2-3 times as sensi-
(N2). For N-2-hydroxycyclohexyl-
(CdN) vs σ plot
N
N
(
0
1
.024, respectively). For the other hydrazones, 10, 13, and
4, the corresponding slopes, 0.023, 0.023, and 0.026,
respectively, are almost the same as those observed for
the imines.
15N NMR Sh ifts of Im in es a n d Hyd r a zon es. The
CdN nitrogen of the imine series 12 resonates at a field
higher by ca. 42-74 ppm than that for the reference
nitromethane. The CdN nitrogen of series 7 and 10
resonates at a field higher by ca. 78-96 ppm and 42-66
ppm, respectively, and N2 at a field higher by ca. 304-
319 ppm and 210-213 ppm, respectively, than that for
the reference nitromethane. The 15N NMR shifts for the
CdN nitrogen of 7 and 10 correlate well with the
substituent constants σ (Table 2; lines 22 and 33) with
slopes of ca. 12 and 15, respectively (cf. Figures 1 and
S1). The N2 resonance is less sensitive. For 7 and 10,
the slopes are ca. 9 (line 24) and 2.3 (line 35), respectively,
and the correlation is only moderate in both cases (cf.
Figures 2 and S2). The slope values for hydrazones, 12
and 15, are clearly smaller than the corresponding slope
of ca. 20 for the CdN nitrogen of the N-benzylidene-
anilines (12) (line 13). On the other hand, the slopes of
Atom ic Ch a r ges vs Su bstitu en t P a r a m eter s. To
assess the effect of substituents on the polarization of
the CdN unit, the variation of the atomic charges was
studied. Table 2 reports the statistical results for the
correlations of the calculated charges of the CdN carbon
and the CdN nitrogen and N2 of the hydrazone series 7,
1
0, 13, and 14, for the CdN carbon and CdN nitrogen
of the imine series 12, and for comparison, also those
calculated previously1 for the imine series 15 and 20,
a
+
with the substituent parameters σ and σ . In all cases,
+
the correlation with σ is better than that with σ . The
slopes for the CdN carbon charges are negative in all
cases. This means reverse behavior, i.e., the electron
density on the CdN carbon increases, although the
electron-withdrawing ability of the substituent, measured
J . Org. Chem, Vol. 68, No. 6, 2003 2153

Neuvonen et al.
TABLE 2. Sta tistica l Da ta for Ha m m ett-typ e Cor r ela tion s of Atom ic Ch a r ges on CdN Ca r bon s a n d CdN Nitr ogen s or
N2 for Im in e Ser ies 12, 15, a n d 20 a n d for Hyd r a zon e Ser ies 7, 10, 13, a n d 14 (Cor r ela tion s of δN(CdN) a n d δN(N2) Va lu es
w ith σ or σ for Ser ies 7, 10, 12, a n d 13 Ar e Also In clu d ed )a
+
line
series
type of compound
imine
correlation
slope ( s
r
ref
b
1
2
3
4
5
6
7
8
9
15
15
qC(CdN) vs σ
qC(CdN) vs σ
qN(CdN) vs σ
qN(CdN) vs σ
qC(CdN) vs σ
qC(CdN) vs σ
qN(CdN) vs σ
qN(CdN) vs σ
qC(CdN) vs σ
qC(CdN) vs σ
qN(CdN) vs σ
qN(CdN) vs σ
δN(CdN) vs σ
δN(CdN) vs σ
qC(CdN) vs σ
qC(CdN) vs σ
qC(CdN) vs σ
qN(CdN) vs σ
qN(CdN) vs σ
qN(N2) vs σ
-0.027 ( 0.004
-0.017 ( 0.003
0.026 ( 0.003
0.016 ( 0.003
-0.029 ( 0.003
-0.021 ( 0.003
0.029 ( 0.003
0.021 ( 0.003
-0.026 ( 0.003
-0.016 ( 0.003
0.024 ( 0.003
0.015 ( 0.003
19.6 ( 0.7
0.9389
0.9002
0.9466
0.9047
0.9599
0.9439
0.9569
0.9319
0.9357
0.8880
0.9358
0.8880
0.9945
0.9865
0.9588
0.8674
0.9040
0.9024
0.8572
0.8624
0.7640
0.9869
0.9562
0.9349
0.8655
0.9257
0.8751
0.9086
0.9657
0.9336
0.4650
0.5884
0.9924
0.9751
0.9430
0.8873
0.9605
0.9421
0.8645
0.8014
0.8657
0.7874
0.9846
0.9835
0.9855
0.9451
0.9597
0.9352
0.9334
0.8835
0.9163
0.8517
1
1
b
+
+
+
+
+
+
+
15
15
20
20
20
b
imine
imine
1
1
b
b
b
20
12
12
12
12
12
12
7
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
12.5 ( 0.7
N,N-dialkylhydrazone
-0.044 ( 0.005
-0.024 ( 0.004
-0.019 ( 0.003
0.030 ( 0.005
0.019 ( 0.004
0.011 ( 0.003
0.007 ( 0.003
11.7 ( 0.8
+
+
7
7
c
7
7
7
+
+
+
+
d
d
7
qN(N2) vs σ
7
7
7
7
δN(CdN) vs σ
δN(CdN) vs σ
δN(N2) vs σ
7.3 ( 0.9
9.4 ( 1.5
δN(N2) vs σ
5.6 ( 1.3
10
10
10
N-benzoylhydrazone
qC(CdN) vs σ
qC(CdN) vs σ
qC(CdN) vs σ
qN(CdN) vs σ
qN(CdN) vs σ
qN(N2) vs σ
-0.032 ( 0.005
-0.019 ( 0.004
-0.016 ( 0.003
0.023 ( 0.002
0.015 ( 0.002
-(9 ( 7) × 10
-(9 ( 5) × 10
14.6 ( 0.6
+
+
c
10
10
10
10
10
10
10
10
+
+
+
+
+
+
+
+
+
+
+
+
-
4
-4
qN(N2) vs σ
δN(CdN) vs σ
δN(CdN) vs σ
δN(N2) vs σ
9.2 ( 0.7
2.3 ( 0.3
δN(N2) vs σ
1.4 ( 0.3
b
13
13
N-phenylhydrazone
qC(CdN) vs σ
qC(CdN) vs σ
qN(CdN) vs σ
qN(CdN) vs σ
qN(N2) vs σ
-0.046 ( 0.005
-0.033 ( 0.004
0.023 ( 0.005
0.014 ( 0.004
0.009 ( 0.002
0.005 ( 0.001
14.6 ( 1.5
b
13
13
13
13
13
13
13
13
14
qN(N2) vs σ
e
δN(CdN) vs σ
δN(CdN) vs σ
δN(N2) vs σ
e
10.5 ( 1.1
e
8.0 ( 0.8
e
δN(N2) vs σ
5.6 ( 1.1
b
hydrazone
qC(CdN) vs σ
qC(CdN) vs σ
qN(CdN) vs σ
qN(CdN) vs σ
qN(N2) vs σ
-0.043 ( 0.005
-0.030 ( 0.004
0.026 ( 0.004
0.016 ( 0.003
0.013 ( 0.002
0.008 ( 0.002
14
14
14
14
14
qN(N2) vs σ
a
s, standard deviation; r, correlation coefficient. b p-NMe2 excluded. c p-NO2 excluded. d p-NMe2, p-Me, and p-CN excluded. e In DMSO
with p-OMe, p-Me, H, p-Cl, and p-NO2 substituents; δN values were taken from ref 12.
ca. 15 (line 43) for the CdN nitrogen and ca. 8 for N2
line 45), which we calculated with the values for the
nitrogen, the lower the frequency of nitrogen resonance
(at a higher field; more negative δ values). This repre-
N
sents normal behavior.
Cr oss-Cor r ela tion s. A more explicit view of the
polarization character of the CdN/CdN-N units of the
imines and hydrazones is provided by the different
(
1
2
phenylhydrazones (13) in DMSO, are close to our
hydrazone values, with the exception of N2 in 10. The
resonance of the CdN nitrogen of the hydrazones seems
to be less sensitive to phenyl substitution than that of
the CdN nitrogen of the imines. The N2 resonance also
shows a response to the substituent, but to a lesser
extent. The good or at least satisfactory fits and positive
correlations and cross-correlations of the q
(CdN), q (N2), δ , and δ values. The slopes of the
linear q (CdN) vs q (CdN) plots for series 7, 10, 13, and
C
(CdN),
q
N
N
C
N
N
C
slopes for all these correlations, together with the q
σ correlations discussed above, indicate that the δ
variations accompanying substitution are electronic in
origin and that the higher the electron density on the
N
vs
N
14 are -0.79 (Table 3, line 14; Figure 3), -0.70 (line 20),
-0.74 (line 26), and -0.77 (line 30), respectively. The
slopes for the q
N C
(N2) vs q (CdN) plots for series 7, 13,
and 14 are -0.29 (Table 3, line 15; Fig. S3), -0.27 (line
2
154 J . Org. Chem., Vol. 68, No. 6, 2003

Influence of CdN-Substitution on Aromatic Substitution
TABLE 3. Sta tistica l Da ta for Differ en t Cor r ela tion s a n d Cr oss-Cor r ela tion s of 13C a n d 15N NMR Ch em ica l Sh ifts a n d
Atom ic Ch a r ges of th e CdN Ca r bon , CdN Nitr ogen , or N2 for Im in e Ser ies 12, 15, a n d 20 a n d for Hyd r a zon e Ser ies 7, 10,
a
1
3, a n d 14
line
series
type of compound
imine
correlation
slope ( s
r
ref
1
2
3
4
5
6
7
8
9
15
15
20
12
12
12
12
7
7
7
7
7
δC(CdN) vs qC(CdN)
qN(CdN) vs qC(CdN)
qN(CdN) vs qC(CdN)
δC(CdN) vs qC(CdN)
52 ( 14
-1.06 ( 0.02
-1.00 ( 0.02
69 ( 14
0.8192
0.9990
0.9989
0.8716
0.9830
0.8357
0.9992
0.9808
0.8956
0.6416
0.9700
0.9622
0.9740
0.9988
0.9324
0.9298
0.9281
0.9682
0.7930
0.9873
0.2387
0.2730
0.9132
0.9903
0.9919
0.9997
0.9920
0.9899
0.9631
0.9999
0.9952
0.9958
1
1
1
imine
imine
b
δN(CdN) vs qN(CdN)
550 ( 40
-6 ( 2
δN(CdN) vs δC(CdN)
qN(CdN) vs qC(CdN)
δC(CdN) vs qC(CdN)c
δN(CdN) vs qN(CdN)
δN(N2) vs qN(N2)d
-0.94 ( 0.01
95 ( 9
N,N-dialkylhydrazone
325 ( 66
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
170 ( 102
-2.4 ( 0.3
-2.1 ( 0.2
1.14 ( 0.11
-0.79 ( 0.01
-0.29 ( 0.05
0.37 ( 0.06
82 ( 13
δN(CdN) vs δC(CdN)
δN(N2) vs δC(CdN)
δN(CdN) vs δN(N2)
qN(CdN) vs qC(CdN)
qN(N2) vs qC(CdN)e
qN(N2) vs qN(CdN)e
δC(CdN) vs qC(CdN)
δN(CdN) vs qN(CdN)
δN(N2) vs qN(N2)
7
7
7
7
10
10
10
10
10
10
13
13
13
13
13
13
14
14
14
14
N-benzoylhydrazone
N-phenylhydrazone
hydrazone
591 ( 54
-284 ( 89
-0.70 ( 0.04
0.011 ( 0.018
-0.02 ( 0.027
76 ( 17
qN(CdN) vs qC(CdN)
qN(N2) vs qC(CdN)
qN(N2) vs qN(CdN)
f
δC(CdN) vs qC(CdN)
g
δN(CdN) vs qN(CdN)
370 ( 40
550 ( 40
-0.74 ( 0.01
-0.27 ( 0.01
0.36 ( 0.02
99 ( 16
-0.772 ( 0.004
-0.38 ( 0.01
0.49 ( 0.02
g
δN(N2) vs qN(N2)
qN(CdN) vs qC(CdN)
qN(N2) vs qC(CdN)
qN(N2) vs qN(CdN)
f
δC(CdN) vs qC(CdN)
qN(CdN) vs qC(CdN)
qN(N2) vs qC(CdN)
qN(N2) vs qN(CdN)
a
s, standard deviation; r, correlation coefficient. p-CN and p-NMe2 excluded. c p-CN excluded. d p-NO2 and p-CN excluded. e p-CH3
b
and p-CN excluded. f δC values were taken from ref 9. g δN values were taken from ref 12.
F IGURE 2. Plot of the 15N NMR chemical shifts of N2 for
series 7 in CDCl vs Hammett substituent constants σ.
F IGURE 3. Cross-correlation between the atomic charge of
the CdN nitrogen and that of the CdN carbon for series 7.
3
2
7), and -0.38 (line 31), respectively, whereas the slope
carbon seems to be distributed on both nitrogen atoms.
is 0.011 only for series 10 (line 21). In accordance with
the behavior of the imines, the atomic charges on the
CdN carbon and the CdN nitrogen of the hydrazones
display an inverse trend. However, in contrast with the
In addition to the difference in the slopes of the q
vs q (CdN) plots between the imines and hydrazones, a
N
(CdN)
C
marked difference is to be seen between N,N-dialkylhy-
drazones (7) and N-alkyl-N-benzoylhydrazones (10) when
the dependence of the charges of the CdN nitrogen and
N2 on the phenyl substitutent is compared. For series 7,
the q (CdN) and q (N2) display a similar trend, although
N N
N2 exhibits a diminished change, only about 40% of that
behavior of the imines, for which ∆q
N
C
(CdN)/∆q (CdN)
is shown to be close to -1.0 (Table 3; lines 2, 3, and 7),
for the hydrazones, the charge on the CdN nitrogen is
less sensitive to substitution than that on the CdN
carbon. The counterpart of the variable charge on the
observed for the CdN (Table 3; lines 14-16). For series
J . Org. Chem, Vol. 68, No. 6, 2003 2155

Neuvonen et al.
1
0, the charge on N2 hardly varies with the substituent
Also, the CdN nitrogens of the hydrazones (7, 13, and
14) both possess more negative charges and resonate at
a lower frequncy than those of the imines (12). The
hydrazones belonging in series 10, however, behave
analogously to the imines in series 12.
(Table 3; lines 20-22). This agrees with the different
1
5
behavior of the N NMR shifts of N2 for series 7 and 10
discussed above (Table 2; cf. lines 24 and 35). In sum-
mary, while the polarizable unit in the imines is the CdN
fragment, in the hydrazones, the charge at the other end
of the dipole seems to be distributed on both nitrogen
atoms. The hydrazide character of benzoylhydrazones is
clearly indicated by the insensitivity of both the charge
In light of the present theoretical calculations and the
correlations observed, the higher dependence of the CdN
carbon 13C NMR chemical shift of the hydrazones on the
phenyl substituent as compared with the imines (cf. Table
1) can be ascribed primarily to two main reasons: (i) the
higher sensitivity of the chemical shift at the probe site
to changes in electron density, i.e., a higher shift:charge
ratio, and (ii) the higher sensitivity of the charge on the
carbon to substitution. Of these two reasons, the former
seems to be the more important. Both of these facts
indicate that the charge distribution on the CdN unit is
different for hydrazones and imines. For imines, we
recently suggested a significant contribution of the sub-
stituent-sensitive inherent polarization of the CdN unit,
the main contributing structures being 3 and 4 (Scheme
N
and δ of N2 to phenyl substitution. Obviously, the amide
resonance prevents interaction of the lone pair of N2 with
the CdN unit.
For the hydrazones, the dependence of the CdN carbon
1
3
C NMR chemical shift on the charge on the carbon is
significant. The slopes of the plots δ (CdN) vs q (CdN)
are 95 and 82 ppm/electron, respectively, for series 7 and
0 (Table 3; lines 8 and 17; cf. Figure S4) and 17. For
C
C
1
the hydrazones 14, the slope is 99 ppm/electron, and for
the N-phenylhydrazones 13, the slope is 76 ppm/electron
(
(
lines 29 and 23, respectively). In all cases, deshielding
increased shift) of the carbon is observed with decreasing
1
a
1). As has already been discussed, for imines, the F
R
electron density (less negative atomic charge) on the
carbon. The CdN carbon chemical shift of the hydrazones
seems to be clearly more sensitive to the atomic charge
on the carbon than that for the imines. For the imine
values are negative but small. This indicates a low
significance of structure 5 and structure 6, too.
Differ en ces in th e Mech a n ism of P ola r iza tion of
th e CdN Un it of Im in es a n d Hyd r a zon es. Scrutiny
of the information derived from the 13C and N NMR
chemical shifts and the PM3 atomic charges reveals
several characteristic differences between the imines and
hydrazones.
15
series 15 and 12, the slopes of the plots δ
(CdN) are only 52 and 69 ppm/electron, respectively
Table 3; lines 1 and 4).1a
For the imine series 12, a slope of 550 ppm/electron is
observed for the plot of δ (CdN) vs q (CdN) (Table 3;
C
(CdN) vs
q
C
(
N
N
(i) On average, the F
analysis of the 13C NMR chemical shifts are somewhat
larger for the hydrazones. (ii) For the F values, a
I
F
or F values derived from DSP
line 5). For the N,N-dialkylhydrazones 7, the depend-
ences of the CdN nitrogen and N2 resonances on the
electron density on the nitrogen in question are different.
Correlation coefficients of 325 and 170 ppm/electron,
respectively, are observed (lines 9 and 10). For series 10,
R
considerable difference is seen (5-10 times larger for the
hydrazones as compared with those for the imines). (iii)
For the hydrazones, both the carbon and nitrogen of the
CdN unit are shielded as compared with imines. (iv) The
atomic charges on the CdN unit are more negative and
more sensitive to effects of substituents for the hydra-
zones. (v) The mutual atomic charge behavior for the
CdN carbon and nitrogen also differs. For the imines,
N N
the slope of the plot δ (CdN) vs q (CdN) is 590 ppm/
electron (Table 3; line 18), and that for series 13 is 370
ppm/electron (line 24). The former value is close to that
observed for series 12, and the latter value is very similar
to that observed for series 7. As regards the plot of δ
vs q (N2), the slopes are -284 and 550 ppm/electron,
respectively, for series 10 and 13 (Table 3; lines 19 and
5, respectively). Both of these values differ considerably
N
(N2)
N
the slope of the plot of q
and close to -1, whereas, as an example, it is -0.79 for
the hydrazone series 7. For the plot of q (N2) vs q (CdN),
N C
(CdN) vs q (CdN) is negative
2
N
C
from the slope of 170 ppm/electron observed for series 7.
For series 13, the correlations in question are very good
for both CdN and N2, while for series 10, the correlation
is satisfactory for CdN and poor for N2. For series 7, both
the slope is only -0.29. Clearly, for the hydrazones, the
charge is distributed on both nitrogen atoms. The facts
listed above suggest a difference between the imines and
hydrazones in terms of the mechanism by which the
substituents induce changes in the charges and the NMR
chemical shifts of the nuclei of the CdN group. The
behavior of the benzoylhydrazones seems to be interme-
diate between those of the imines and the other hydra-
zones. For the imines, the substituent dependence of the
N
correlations are poor. With the exception of δ (N2) for
series 10, in all cases, the less negative the electron
density of the atom, the lower the field of nitrogen
resonance. This is according to the generalized electronic
effect.
The CdN carbons of the hydrazones carry significantly
more negative atomic charges from -0.13 to -0.06 (cf.
Tables S1, S4, S5, and S6) than those previously observed
1
3
15
C and N NMR chemical shifts on the benzylidenic
substituent X can largely be explained by comparison of
the contributions of the resonance forms 3-6 in Scheme
1
a
for the CdN carbon of the imines (from -0.06 to
1. To date, no explicit explanation has been given for
the behavior of the hydrazones. However, resonance
structure 22 was proposed by Gordon et al. (cf. Scheme
-
0.003)1 or now observed for the imine series 12 (Table
a
1
3
S2, from -0.06 to -0.02). This is consistent with the
NMR chemical shift data, which show that the CdN
C
9
2). The basic question is that of whether the effect of Y
carbon of the hydrazones resonates at a lower frequency
is conjugative (if there is a lone pair in use, as has been
proposed), inductive, or steric.
(i.e., with reduced shift values) than that of the imines
(cf. Table 1). This difference suggests a more significant
For a better understanding of the behavior of the
contribution of the resonance form 4 (cf. Scheme 1) in
the case of the imines than in the case of the hydrazones.
hydrazones vs that of the imines, we compared the F
F
(X)
and F (X) values of the compounds studied in this work
R
2
156 J . Org. Chem., Vol. 68, No. 6, 2003

Influence of CdN-Substitution on Aromatic Substitution
TABLE 4. Effect of Su bstitu en t Y in Y-NdCH-C6H4-p-X System on GR a n d GF Va lu es Descr ibin g th e Dep en d en ce of 13C
a ,b
NMR Ch em ica l Sh ifts of th e CdN Ca r bon on th e Ben zylid en ic Su bstitu en t X
CdN
bond order
(bond length)
dN-Y
bond order
(bond length)
dC-C(Ph)
bond order
(bond length)
Y
σp+
δC(CdN)
qC(CdN)
-0.0795
FF
FR
refc
NH2
-1.30
142.77
-5.2
-4.0
1.8541
1.0697
(1.3824)
1.0473
(1.3914)
1.0130
(1.4007)
0.9810
(1.4264)
1.0130
(1.4499)
0.9720
(1.4838)
1.0330
(1.4310)
0.9892
1.0190
(1.4652)
1.0186
(1.4653)
1.0158
(1.4660)
1.0130
(1.4668)
1.0090
(1.4682)
1.0041
(1.4704)
1.0113
(1.4676)
1.0077
9
9
(
1.3020)
1.8564
1.3013)
1.8961
1.2982)
1.9064
1.2958)
1.9161
1.2919)
1.9285
1.2883)
1.8891
1.2939)
1.9217
1.2910)
1.8852
1.2939)
NHPh
HO
-1.40
-0.92
-0.78
-0.31
-0.26
-0.18
-0.28
+0.04
137.25
150.33
148.26
162.44
155.08
160.34
161.93
162.46
-0.0771
-0.0531
-0.0489
-0.0452
-0.0295
-0.0289
-0.0340
0.0042
-4.6
-4.4
-2.9
-3.4
-3.2
-4.1
-3.5
-3.6
-3.2
-0.7
-1.2
-0.30
-0.4
-0.75
-0.5
0.07
(
9, 15
(
MeO
Me
13
(
13
(
(
CH3)3C
13
(
Ph
this work
(
PhCH2
p-NO2-C6H4
10
14
(
(1.4656)
1.04570
(1.4264)
(1.4686)
1.0145
(1.4666)
(
a
δC(CdN) and qC(CdN) values and CdN, dN-Y, and dC-C(Ph) bond orders and bond lengths (Å) are given for substitution X ) H.
b
dC-C(Ph) refers to the bond between the CdN carbon and the substituted phenyl ring. F values were calculated in this work via eq 1:
1
3
SCS ) FFσF + FRσR, with the aid of the literature C NMR data, with the exception of Y ) Ph (series 12), for which the NMR data were
c
13
determined in this work. Reference for the C NMR chemical shift data used to calculate the FF and FR values.
SCHEME 2
with the corresponding values that we derived from the
literature NMR shift data (Y-NdCH-C
6
H
4
-p-X; Table
(X) for series
2 (Y ) -Ph) are based on 1 C NMR shifts recorded in
F IGURE 4. Plots of the correlation parameters F
F
(X)(O) and
4
). The correlation parameters F
F
(X) and F
R
3
F
R
(X) (b) (determined with respect to phenyl substituent X)
1
+
vs substituent constant σ
-p-X.
p
(Y) with varying Y for Y-NdCH-
3
this study. For the other series, i.e., Y ) -Me, -CMe ,
6 4
C H
-
CH
2
Ph, -C
6
H
4
-p-NO
2
, -OMe, -OH, -NH
2
(series 13),
and -NHPh (series 14), we calculated F
F
(X) and F
R
(X)
stabilize 4 (-Me, -CMe
3
, -Ph, -CH
2
Ph, -C
6
H
4
2
-p-NO ),
with the aid of 1 C NMR shift data from the literature.
The fact that the sensitivity of the electronic character
of the CdN function to electron donation/ electron with-
drawal by the phenyl substituent X is adjusted by the
electronic effects of Y is seen on inspection of Figure 4.
3
the contribution of 4 (Scheme 1) will increase as a
+
P
function of σ (Y). With electron-donating substituents
X, its contribution further increases. This will result in
a more positively charged CdN carbon (deshielding) and
negative F
F
R
and F (the latter via structure 5) values. On
Both the F
F
(X) and F
R
(X) values from Table 4 can be
the other hand, although Y is inductively electron
withdrawing but at the same time electron donating via
+
successfully correlated with the σ
P
values of group Y,
(Y). Thus, the effect of Y is electronic in nature. We
also examined correlations with σ, σ , and R, but these
+
σ
P
resonance (-NH , -NHPh, -OH, or -OMe), resonance
2
I
structure 22 (Scheme 2) is stabilized. As regards the
effect of substituent X, electron-withdrawing substituents
stabilize 22 inductively and make possible the contribu-
tion of 23 through resonance. As a consequence, the Cd
N carbon exhibits more negative charges. This results
were less successful, indicating that both inductive and
resonance effects contribute. The varying sources of the
NMR data (different sample concentrations, varying set
of substitutions) account for the fact that the goodness
of the correlation is not very high. Figure 4 can be
rationalized by inspecting the contributions of the dif-
ferent resonance forms in Schemes 1 and 2. For imines,
when Y is an electron-withdrawing group or can at least
in negative F
F
R
and F (the latter via structure 23) values.
Figure 4 clearly shows that the hydrazones exhibit
characteristics that are common for all compounds pos-
sessing an atom with a lone pair attached to the CdN
J . Org. Chem, Vol. 68, No. 6, 2003 2157

Neuvonen et al.
nitrogen, i.e., when Y is a group capable of conjugative
electron donation. The correlations in Figure 4 can be
explained by different balances between the possible
resonance structures for the imines (Scheme 1) and the
hydrazones and related compounds (Scheme 2). The
substituents Y capable of electron donation by resonance
(
Y ) -NH
more negative F
indicating an increase in the contribution of 22 (Scheme
). The fact that the F (X) values are highly negative for
2
, -NHPh, -OMe, -OH; group A) result in
+
R
p
(X) values the more negative σ (Y) is,
2
R
group A means that the resonance effects of the substit-
uents X must be resonance-induced polar effects (22 T
2
3) and not the normal resonance effects (22 T 24). The
latter case would lead to F (X) > 0.
R
It seems evident that the resonance-induced polar
effects (5, Scheme 1) are not of high importance (only
small negative F
CH Ph, -C
is that in both cases, i.e., for the compounds in groups A
and B, both the normal resonance effect [6 in Scheme 1,
R
) for the imines (Y ) -Me, -CMe
3
-Ph,
R
F IGURE 5. Plot of the correlation parameter F (X) (deter-
-
2
6
H
4
-p-NO ; group B). The other possibility
2
mined with respect to phenyl substituent X) vs the CdN
13
carbon C NMR chemical shift for the phenyl unsubstituted
6 5
derivative Y-NdCH-C H with varying Y.
2
4 in Scheme 2; leading to F
resonance-induced polar effect [5 in Scheme 1, 23 in
Scheme 2; leading to F (X) < 0] operate; however, in the
R
(X) > 0] and the inverse
R
case of the imines (group B), the two effects compensate
each other, while in the case of group A, the inverse effect
is overwhelming.
F
For the imines (group B), the F values are clearly
negative in all cases, varying within only a narrow range.
Obviously, the contribution of 4 does not vary largely
with the different Ys studied. Electron-withdrawing
phenyl substituents X inductively destabilize 4, and a
decrease in the imine carbon charge (a less positive
charge), a shielding effect, and F
F
(X) < 0 result. For the
hydrazones and related compounds (group A), an in-
+
p
creased electron donation by Y [a more negative σ (Y)]
stabilizes 22. Consequently, its contribution increases,
leading to a higher sensitivity of the electronic character
of the CdN function to phenyl substitution via the
inductive effects of X. Electron-withdrawing substituents
X inductively stabilize structure 22, and a decrease in
the charge on the imine carbon (more negative charges),
F IGURE 6. Plot of the CdN carbon 13C NMR chemical shift
in CDCl for the phenyl unsubstituted derivative Y-NdCH-
vs the atomic charge of the CdN carbon with varying Y.
3
6 5
C H
a shielding effect, and F
F
(X) < 0 result. The varying
to the electronic resonance effects of the phenyl substitu-
+
sensitivity to X is seen as a dependence of F
F
(X) on σ
p
(Y)
ent X, i.e., F (X), is a useful tool for evaluating the
R
with a positive slope.
electrophilicity of the CdN carbon.
Additional support is lent to our conclusions by an
inspection of the plot in Figure 5, where F (X) is plotted
against δ (CdN) for the unsubstituted derivatives, i.e.,
X ) H, for the series with different Ys. The more shielded
the carbon, the more sensitive its C NMR shift is to
conjugative effects by X. We consider the shielding of the
CdN carbon to reflect the fact that the contribution of
resonance structure 22 is increased. This conclusion could
be verified by the plots shown in Figures 6 and 7. In
Figure 6, it is clearly seen that shielding corresponds to
the increase in electron density at the CdN carbon.
The bond order/bond length data in Table 4 further
support the above conclusions. For compounds in group
A, the CdN bond order is lower and the bond is
lengthened, while the dN-Y and dC-C(Ph) bond orders
are higher and the bonds are shortened as compared with
compounds in group B. The imine derivatives possessing
an aromatic ring attached directly to the CdN nitrogen
R
C
1
3
(Y ) Ph or p-NO
2
6 4
-C H ) show bond orders between those
typical for group A and those typical for other members
of group B. Further, when varying Y, good or at least
-
2
fair correlations with slopes of (6.2 ( 0.7) × 10 (r )
Figure 7 shows correlation between F
C
R
and F
F
values and
0.9650) and 1.54 ( 0.10 (r ) 0.9895), respectively, were
+
q
(CdN). Both parameters turn more negative when the
observed between the CdN bond order and σ
p
(Y) and
atomic charge at the CdN carbon becomes more negative.
According to our interpretation, the inductive effects of
substituents X affect the stabilities of forms 4 and 22 with
practically the same efficiency, while the resonance
effects overwhelmingly affect via the interaction 22 T
between the CdN bond order and q
Ph and p-NO -C were excluded.
The present results show that the sensitivity of the
electronic character of the CdN function in Y-NdCH-
C
(CdN) when Y )
2
6 4
H
C
6
H
4
-p-X to electron donation/ electron withdrawal by
2
3. Thus, it can be concluded that particularly the
the phenyl substituent X is adjusted by the electronic
1
3
13
sensitivity of the C chemical shift of the CdN carbon
effects of Y. This is revealed by the behavior of the
C
2
158 J . Org. Chem., Vol. 68, No. 6, 2003

Influence of CdN-Substitution on Aromatic Substitution
by δ
order vs σ
C
(CdN) vs q
C
(CdN), F
F
/F
R
vs q
C
(CdN), the CdN bond
(CdN)
+
p
(Y), and the CdN bond order vs q
C
correlations with varying Y. According to our interpreta-
tion, the inductive effects of substituent X affect the
stabilities of forms 4/25 and 22 with practically the same
efficiency, while the resonance interaction 22 T 23 is
more efficient than the interaction 4 T 5. We conclude
that these results can be useful as concerns an explana-
tion of the reactivities of the compounds in question
toward nucleophilic agents; this property is governed by
the electrophilicity of the CdN carbon.
Exp er im en ta l Section
Gen er a l P r oced u r es. Melting points are uncorrected. All
new compounds (see Table S1) gave satisfactory microanalyses
(C, H, N).
Syn th esis of Ben zylid en e Der iva tives 12. Aniline (2
F IGURE 7. Plots of the correlation parameters F
R
F
(X) (O) and
mmol) and the appropriate aromatic aldehyde (2 mmol) were
dissolved together in ethanol (10 mL) and left to stand
overnight at room temperature. The solvent was evaporated
off, and the products were recrystallized from ethanol or
diisopropyl ether. The yields were 77-90%.
F
(X) (b) (determined with respect of phenyl substituent X)
6 5
vs the atomic charge of the CdN carbon for Y-NdCH-C H
with varying Y.
or 15N NMR chemical shifts. Related results have recently
been obtained by Liu et al. in theoretical bond dissocia-
NMR Mea su r em en ts. NMR spectra were recorded at 27
1
3
°C on an NMR spectrometer operating at 125.78 MHz for
C
and 50.688 MHz for 15N on 0.4 M solutions in CDCl
. C NMR
13
tion energy studies concerning para-substituted aromatic
3
silanes17 and para-substituted anilines.18 For instance,
spectra were referenced internally to tetramethylsilane (0.00
ppm), while 15N NMR spectra were referenced externally to
for their 4-Y-C
Cl, or Li; Y ) NO
6
H
4
-Z-X system (Z ) SiH
2
; X ) H, F,
, CONH
CH
3 2 3 2
NO (0.00 ppm) containing 10% w/w CD NO for locking
2
, CHO, CN, CF , COCH
3
3
2
,
purposes. The signal of the deuterium of the solvent was used
COOH, F, Cl, H, Me, SMe, OMe), they state that “the
direction and magnitude of the effect of substituents Y
13
13
as a lock signal for C NMR spectra. C NMR spectra were
1
1
acquired with H broad-band decoupling and NOE H nonde-
coupling techniques. 13C NMR spectra were acquired with the
following conditions: spectral width ) 30 kHz, 32 K data
have some important dependence on the polarity of the
Z-X bond undergoing homolysis.”17 Their results as well
1
1
points ( H decoupled)/64 K data points ( H coupled), digital
as ours indicate that there prevails an interesting
interaction over the aromatic and side-chain systems. The
item is under continuous interest.
1
1
resolution ) 0.92 Hz/point ( H decoupled)/0.46 Hz/point ( H
coupled), pulse width ) 4.35 µs (45°), acquisition time ) 1.09
1
1
s ( H decoupled)/2.18 s ( H coupled), number of transients )
1
1
1
000-12000, pulse delay ) 3 s ( H decoupled)/5 s ( H coupled),
1 1
pulse sequence ) (J EOL) SGBCM ( H decoupled)/SGNOE ( H
Con clu sion s
coupled). Exponential windowing with a line-broadening term
1
1
of 2 Hz ( H decoupled)/1 Hz ( H coupled) was applied prior to
Fourier transformation. N NMR spectra were acquired with
refocused INEPT technique optimized on 8 Hz. N NMR
The extensive data relating to the azomethines, hy-
drazones, and related compounds derived from substi-
tuted benzaldehydes have shown that the substituent
effects on the CdN 13C and/or CdN N NMR chemical
shifts can be explained by considering the different
resonance structures depicted in Schemes 1 and 2. For
these compounds possessing a CdN double bond, the
15
15
spectra were acquired with the following conditions: 90° flip
angle, pulse recycle time ) 5.1 s, spectral width ) 25 kHz
consisting of 64 K data points (digital resolution 0.39 Hz/point),
pulse sequence ) (J EOL) INPTR. Exponential windowing with
a line-broadening term of 1 Hz was applied prior to Fourier
transformation.
15
R
magnitude of the negative F derived from calculations
Sta tistica l Ca lcu la tion s. The sources of the substituent
involving the 13C NMR chemical shifts of the CdN carbon
with respect to substituent X is a useful tool for evaluat-
ing the detailed electronic character of the CdN double
bond. The strength of the effect of X varies characteristi-
+
constants are as follows: σ, σ , ref 11; σ , σ , ref 16, except σ
F
R
F
R
and σ for Br, ref 11.
P M3 Ca lcu la t ion s. Three-dimensional structures of the
compounds were composed by using the program SYBYL,
1
9
and the structures thus obtained were optimized with the
Tripos force field. The resulting conformations were used as
starting structures for the semiempirical calculations of
conformations, heats of formation, and atomic charges. These
calculations were performed by using the semiempirical
cally as a function of substituent Y. In Figure 4, the F
and F values observed with substituent X are plotted
against the σ
R
F
+
p
value of substituent Y. These plots
illustrate a novel way of using the Hammett-type pa-
rameters. The justification of this treatment was verified
2
0
quantum mechanical method PM3 of the program MOPAC
2
1
7
.0. All structures were optimized without any restriction.
The quantum chemical calculations were processed on the SGI
Octane” (2 × R 12 000) computer of the University of
Potsdam.
(
13) J ennings, W. B.; Wilson, V. E.; Boyd, D. R.; Couther, P. B. Org.
Magn. Reso. 1983, 21, 279.
14) Akaba, R.; Sakuragi, H.; Tokumaru, K. Bull. Chem. Soc. J pn.
985, 58, 1186.
15) Danoff, A.; Franzen-Sieveking, M.; Lichter, R. L.; Fanso-Free,
S. N. Y. Org. Magn. Reson. 1979, 12, 83.
“
(
1
(
(19) SYBYL 6.7, Molecular Modelling Software; TRIPOS, Inc.: St.
Louis, MO, 2001.
(
(
16) Taft, R. W.; Topsom, R. D. Prog. Phys. Org. Chem. 1987, 16, 1.
17) Cheng, Y.-H.; Zhao, X.; Song, K.-S.; Liu, L.; Guo, Q.-X. J . Org.
(20) Stewart, J . J . P. J . Comput. Chem. 1989, 10, 221.
(21) Stewart, J . J . P. MOPAC 7.0, A Public-Domain General Mo-
lecular Orbital Package; Frank J . Seiler Research Laboratory, US Air
Force Academy: Colorado Springs, CO, 1993.
Chem. 2002, 67, 6638.
18) Song, K.-S.; Liu, L.; Guo, Q.-X. J . Org. Chem. 2003, 68, 262.
(
J . Org. Chem, Vol. 68, No. 6, 2003 2159

Neuvonen et al.
Ack n ow led gm en t. Two of the authors (F.F. and
K.P.) wish to express their gratitude to the Research
Council for Natural Sciences, The Academy of Finland,
and to the National Scientific Research Grant, Hungary
PM3 atomic charges for series 7, 10, 12-14 (Tables S2-S6);
bond lengths in Table S7; δ
0 (Figure S1); δ
S2); q (N2) vs q
and δ (CdN) vs q
N
(CdN) vs σ correlation for series
1
N
(N2) vs σ correlation for series 10 (Figure
(CdN) correlation for series 7 (Figure S3);
(CdN) correlation for series 7 (Figure S4).
N
C
C
C
(T 4466), for financial support.
This material is available free of charge via the Internet at
http://pubs.acs.org.
Su p p or tin g In for m a tion Ava ila ble: Analytical data on
1
3
15
compounds of series 12 (Table S1); C and N NMR data and
J O020608L
2
160 J . Org. Chem., Vol. 68, No. 6, 2003