Substituent effects on the 13C NMR and the 1H NMR chemical shifts
1
Now that the H NMR and 13C NMR chemical shifts are all af-
the δC(CH¼N), while the σF and σR of Y group increase the
δC(CH¼N). What we want to know is how do the X and Y
groups affect the δH(CH¼N). We made discussions as follows.
Based on 129 samples of title compounds, the regression analysis
of the δH(CH¼N) values of Hammett constant σp/m of X and Y was
carried out in this paper, and Eqn (5) was obtained (Table 2). In
Eqn (5), the ration of |ρ(X)/ρ(Y)| is 2.00, which means that the effect
of X on the δH(CH¼N) is 2.00 times higher than Y on δH(CH¼N), and
the effect of Y on the δH(CH¼N) is not ignorable. On the other hand,
Eqn (5) shows that the σp/m of X group increases the δH(CH¼N),
while the σp/m of Y group decreases the δH(CH¼N). It is just opposite
with the effects of substituents on the δC(CH¼N).
fected by the electron density, and it was generally believed that
the δH(CH¼N) values should increase as the δC(CH¼N) values in-
crease in a same set of compounds. Is it true? Based on a wide set
of 129 samples of substituted benzylideneanilines (as shown in
Scheme 1), the plot of δC(CH¼N) values versus δH(CH¼N) values
was carried out (Fig. 1).
Figure 1 shows a surprising result that there is a bad correlation
between the δH(CH¼N) and δC(CH¼N) of XBAYs, the correlation co-
efficient is only 0.6347. It implies that the effects of substituents X
and Y on the δH(CH¼N) and on the δC(CH¼N) are different. What
are the reasons bringing out the aforementioned phenomenon?
In this paper, we made an investigation on this topic using the
129 samples of title compounds and attained a meaningful result.
As we know, σ = σF + σR. In the report of Cao et al.,[17] the
regression equation of the δC(CH¼N) became better when
Hammett constant σ was divided to the inductive constant σF
and conjugative constant σR; therefore, the different contribu-
tions of inductive and conjugative effects were taken into
account in the regression analysis of the δH(CH¼N) values in this
paper, and Eqn (6) was obtained (Table 2). Comparing the
correlation coefficient (R), the standard errors (S), and F value
of Eqn (6) with these of Eqn (5), the results show that Eqn (6)
has no improvement after dividing σ to σF and σR. That is to
say, the inductive and conjugative effects of substituents on δH
(CH¼N) are almost the same intensity, and they can be merged
in the correlation equation. It is different from the results of
regression analysis of the δC(CH¼N) by Cao et al.[17]
Results and discussions
The δC(CH¼N) values and δH(CH¼N) values of title compounds
were collected and listed in Table 1. Some of them were measured
in this work, and the rest were quoted from the literatures. The sub-
stituents X and Y in molecules of title compounds are of electron-
withdrawing substituents (e.g. NO2 and CN) and electron-donating
substituents (e.g. NMe2 and OMe); in addition, X and Y are of para-
substituted and meta-substituted. Taking Eqns (1 and 2) as models,
corresponding regression equations were attained [Eqns (3 and 4)]
for the 129 δC(CH¼N) values in Table 1.
It is known that the substituent specific cross-interaction effects
Δσ2 between X and Y is a necessary item in quantifying δC(CH¼N),[17]
additionally for the comparison with Eqn (4), so this item was also
added into Eqn (6) to carry out the regression analysis against the
δH(CH¼N) values, then Eqn (7) was obtained (Table 2).
δCðCH ¼ NÞ ¼ 160:25 ꢀ 4:68σFðXÞ þ 3:19σFðYÞ
(3)
ꢀ 0:86σRðXÞ þ 5:04σRðYÞ
R ¼ 0:9860; R2 ¼ 0:9722; S ¼ 0:37; F ¼ 1084:48; n ¼ 129
The results of Table 2 show that the correlation coefficient (R) and
the standard errors (S) of Eqns (5 and 7) are almost the same; but
the F value of Eqn (5) is larger than that of Eqn 7. Also, the average
absolute errors between the calculated values and experimental
values of the two equations are all equal to 0.03 ppm. In conclusion,
the Eqn (5) is enough to quantify the substituent effects on the
δH(CH¼N), and the substituent specific cross-interaction effects
Δσ2 on the δH(CH¼N) is negligible, which is different from the
effect of Δσ2 on the δC(CH¼N).[17] In the report of Cao et al.,[17]
the substituent specific cross-interaction effect on the δC(CH¼N) is
indispensable. The reason leading to the difference of substituent
effects on the δC(CH¼N) and δH(CH¼N) may be the deviation of
hydrogen atom from the conjugated main chain in the molecules
of XBAYs.
δCðCH ¼ NÞ ¼ 160:30 ꢀ 4:38σFðXÞ þ 3:07σFðYÞ
(4)
ꢀ 1:11σRðXÞ þ 4:63σRðYÞ ꢀ 0:61Δσ2
R ¼ 0:9938; R2 ¼ 0:9877; S ¼ 0:25; F ¼ 1968:77; n ¼ 129
One can observe that Eqn (4) is superior to Eqn (3) obviously. The
correlation coefficient R of Eqn (4) is larger than that of Eqn (3), and
the standard error S of Eqn (4) is smaller than that of Eqn (3).
Equation (4) shows that the σF and σR of X group decrease
In addition, the parameter effects on the δC(CH¼N) and δH(CH¼N)
are different. For that, Eqns (4 and 7) have the same parameters and
the regression results of Eqn (7) are close to that of Eqn (5). So
here the relative importance of parameters in Eqns (4 and 7) are
investigated from the relative contributions (ψγ) or fraction contri-
butions (ψf) of the corresponding parameters to the δC(CH¼N) and
δH(CH¼N).[20,21]
Scheme 1. Title compounds used in this paper.
ψγ ¼ miXi
(8)
ꢀ
ꢀ
ꢀ
ꢀ
2
R ψγðiÞ
X
ꢀ
ꢀ
ψfðiÞ ¼
ꢁ100%
(9)
ꢀ
ꢀ
ψγðiÞ
i
The mi and Xi are the coefficient and the average value of the ith
parameter in Eqn (4 or 7), respectively, and the R are the correlation
coefficients of Eqn (4 or 7). The sum is over the parameters in the
equations. The contribution results for the corresponding parame-
ters of Eqns (4 and 7) are all shown in Table 3.
Figure 1. The plot of δH(CH¼N) values versus δC(CH¼N) values of title
compounds.
Magn. Reson. Chem. 2015, 53, 520–525
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/mrc