6
518
K. N. Skala et al. / Tetrahedron Letters 51 (2010) 6516–6520
Table 2
Chemical shift changes of proton signals
LiPF or NaPF
D
3
d (ppm) of 1 in CDCl upon saturation with
6
6
.
NH
Py 3
Py 4
Py 5
Py 6
LiPF
NaPF
6
+0.78
+0.78
+0.08
+0.01
+0.15
+0.11
À0.25
À0.21
À0.38
À0.11
CH Py
2
6
6
In 4
In 5
In 6
In 7
CH
2
In
LiPF
NaPF
6
À0.11
À0.06
À0.04
À0.11
À0.09
À0.09
À0.06
+0.32
+0.14
+0.17
+0.06
À0.002
+
NH more electron-deficient. The DFT-optimization of 1ÁLi indeed
resulted in both indolyl N atoms in the coordination sphere of
+
Li , in addition to the expected pyridyl N and the amine N
(
Figure 5).
1
Ion-pair binding behavior of 1 was also examined by H NMR. In
CD
3
CN, addition of 4.5 equiv TBACl and 4.5 equiv LiPF
6
resulted in
À
+
the precipitate formation (LiCl from excess Cl and Li ), and the
resulting solution matched the spectrum of 1:1 mixture of 1 and
À
Cl from the titration experiment. Therefore 3.5 equiv of LiCl must
À
À
+
have precipitated out, and 1 equiv Cl and 1 equiv Li must have
Figure 3. Chemical shift movement over titration of 1 with Bu
4
NCl in CD
3
CN. The
remained in solution, in which the Cl ions were interacting with
pyridyl and indolyl proton signals are shown separately for clarity. See Scheme 1 for
numbering of the aromatic protons.
+
1
. But apparently Li in solution does not have any influence on
either 1 or the 1ÁCl complex. Therefore in CD
À
3
CN, there was no
À
cation binding even by 1ÁCl .
On the other hand, we detected evidences of simultaneous
À
+
3 6
binding of Cl and Li in 83:17 CDCl /DMSO-d . In Figure 6 the
chemical shift changes of proton signals of 1 upon addition of (b)
+
À
À
1
0 equiv Li , (c) 10 equiv Cl , and (d) 10 equiv Cl and 10 equiv
+
+
Li are shown. The changes from (a) to (b), upon addition of Li ,
were relatively small, with most notable changes in the two meth-
ylene peaks, CH
2
In and CH
2
Py, by +0.14 and +0.17 ppm. The large
À
changes from (a) to (c) upon addition of 10 Cl in the signals of
NH (+1.00 ppm), Py 3 (+0.41 ppm), and In 7 (+0.16 ppm) were con-
sistent with the strong (NH) and weak (Py 3 and In 7) hydrogen
À
bondings to Cl as discussed earlier. When both 10 equiv Bu
4
NCl
were added (in either order) to 1 in 83:17
, no precipitation occurred, therefore the solution
and 10 equiv LiPF
CDCl /DMSO-d
6
À
Figure 4. Two views of a DFT-optimized structure of 1ÁCl . Color scheme: Gray = C,
3
6
white = H, blue = N, scarlet = Br, and green = Cl. Some hydrogen atoms are omitted
for clarity.
+
À
contained 10 equiv each of Li and Cl . The proton spectrum of
the resulting solution (d) was quite different from 1 only, 1 plus
+
À
+
Li , and 1 plus Cl . The effect of addition of Li to the mixture of
À
1
+ 10 Cl (c vs d) is obviously larger compared to the minimal ef-
À
respectively. The former is well in the range of reported CHÁ Á ÁCl
+
À
fect of Li to free 1 (a vs b), indicating that the complex 1ÁCl has a
hydrogen bonding distances.10 Though aromatic CH bonds do not
+
+
significant interaction with Li , possibly an inclusion of Li .
However, in the presence of excess of both ions, we cannot
À
usually form hydrogen bonds, in this case the geometry of 1ÁCl
À
brings these protons in proximity of Cl ion, within a possible
À
+
ascertain the stoichiometry of 1:Cl :Li in the ternary complex.
In order to find the stoichiometry, We titrated a 1:1 mixture of 1
range of some interaction.
The interaction between 1 and alkali metal cations from MPF
M = Li, Na and K) was also examined in CD CN, 83:17 CDCl
DMSO-d , and CDCl . In the first two solvents, addition of 5 equiv
MPF caused very minimal changes (no more than 0.15 ppm,
most peaks moved less than 0.01 ppm) in the H NMR signals of
. Strong coordination of solvent molecules to the cations may be
6
and Bu
4
NCl with LiPF
6
in 83:17 CDCl
3
/DMSO-d
6
. We have already
(
3
3
/
À
established the 1:1 stoichiometry of the 1ÁCl complex based on
6
3
the titration as discussed earlier. In a 1:1 mixture of 1 and Bu
4
NCl
6
D
À
a mixture of free 1 and a complex 1ÁCl exist, but since we know
1
+
that the interaction between free 1 and Li is insignificant, any
1
+
preventing a direct interaction between 1 and M .
In CDCl , due to poor solubility of MPF , we could not carry out
a titration or quantitative addition of the cation source to 1. We did
a coarse experiment of saturating a 5 mM solution of 1 in CDCl
with MPF by adding excess solid MPF and shaking it vigorously.
The H NMR signals of 1 were affected by saturation with LiPF and
NaPF , but not by KPF . The chemical shift changes of 1 in CDCl
upon saturation with LiPF and NaPF are summarized in Table 2.
Interestingly, the NH peak had a significant downfield shift upon
3
6
3
6
6
1
6
6
6
3
6
6
addition of LiPF
6
and NaPF
6
. This is not due to the hydrogen bond-
À
ing of the NH groups to the anion PF6 , since in a separate exper-
iment, Bu
4
NPF
6
did not cause any changes in proton signals of 1.
+
Figure 5. Two views of a DFT-optimized structure of 1ÁLi . Color scheme: Gray = C,
The possible reason of the downfield shift of the NH protons is
the coordination of the indolyl N atoms to the cation, making the
white = H, blue = N, and scarlet = Br, and pink = Li. Some hydrogen atoms are
omitted for clarity.