by silica gel chromatography. Oxidative coupling8 of 8
afforded dimer 9 which was treated with potassium carbonate
to give partially deprotected 10 (45%), thoroughly depro-
tected 11 (38%), and recovered 9 (12%). Oxidative coupling8
of 10 afforded linear tetramer 13, and removal of the TMS
group gave 14. Intramolecular oxidative coupling of 14 under
high dilution conditions by Eglinton’s method9 yielded cyclic
tetramer 4 in 50% yield. Palladium-catalyzed heterocou-
pling10 of dibromide 12, prepared by bromination of 11 with
NBS, with 2 equiv of 10 gave linear hexamer 15. Depro-
tection of 13 gave unstable hexamer 16 which was subjected
to intramolecular coupling to furnish cyclic hexamer 5 in
29% yield.
having the same macrocyclic ring size as shown in Figure
2. The chemical shift change of 4 shown in Figure 2 was
Pyridinophanes 4 and 5 did not show concentration
1
dependence in the H NMR spectra in CDCl3 (10-2-10-4
M), indicating that these did not tend to self-associate. This
is probably due to the electrostatic repulsion between the
nitrogen atoms,4b as indicated by the electrostatic potential
plot (Figure 1) based on the AM1 method for the respective
Figure 2. 1H NMR titration of 4 and 5 with metacyclophanes 1
and 2 in CDCl3 at 30 °C following the chemical shift change of
the aromatic protons of 4 and 5: 2, 4 and 2; 4, 4 and 1 with a
theoretical curve obtained by nonlinear least-squares analysis; b,
5 and 1; O, 5 and 2.
analyzed by assuming the formation of homodimer of 1 with
a known dimerization constant of K1 (26 M-1 at 30 °C)4
and of heterodimer 1‚4 with an association constant of K2.
The nonlinear least-squares approximation yielded K2 of 72
M-1. In the case of heteroassociation between hexamers 2
and 5, however, similar treatment of the experimental data
did not give a good fit. This means that the tetramers 1 and
4 form a dimer 1‚4 mainly whereas the hexamers 2 and 5
aggregate to form dimer 2‚5 and higher aggregates.12
A
similar tendency to form higher aggregates was observed in
the hetetoaggregation between hexamer 2 and its cyano
derivative 3.4b By contrast, the chemical shifts of the aromatic
protons of 4 and 5 did not change upon addition of 2 and 1,
respectively, having different macrocyclic ring size as shown
in Figure 2. We assume that the driving force for the
heteroaggregate formation is dipole-dipole interaction be-
tween the pyridine rings of 4 and 5 and the benzene rings of
1 and 2.13 Since the dipoles of the individual aromatic ring
are arranged along the periphery of the planar macrocycle,
the macrocycles have no or, if any, little dipole moment.
Accordingly, the overall interaction between pyridinophanes
4 and 5 and cyclophanes 1 and 2 is similar to the
quadrupole-quadrupole interaction between the aromatic
rings substituted by electron-donating and -accepting groups.14
Such interaction should be sensitive to the shape of the
interacting molecules because the mutual overlap of the rings
Figure 1. Electrostatic potentials between -19 (red) to 20 (blue)
kcal/mol on the van der Waals molecular surface of the model
compounds 17 and 18 according to the AM1 calculations.
model compounds 17 and 18 for the tetramers 1 and 4.11 On
the other hand, the aromatic protons of 4 and 5 moved upfield
on addition of the corresponding metacyclophanes 1 and 2
(7) For a review, see: Sonogashira, K. ComprehensiVe Organic Synthesis;
Trost, B. M., Fleming, I., Eds.; Pergamon: Oxford, 1991; Vol. 3, pp 521-
549.
(8) Hay, A. S. J. Org. Chem. 1962, 27, 3320.
(9) (a) Eglinton, G.; Galbraith, A. R. Proc. R. Chem. Soc. 1957, 350.
(b) Behr, O. M.; Eglinton, G.; Galbraith, A. R.; Raphael, R. A. J. Chem.
Soc. 1960, 3614.
(10) Cai, C.; Vasella, A. HelV. Chim. Acta 1995, 78, 2053.
(11) The AM1 calculations were performed using the SPARTAN ver.
5.0 program package; Wavefunction, Inc., Irvine, CA.
(12) An attemped nonlinear least-squares analysis of the data assuming
the formation of a heterodimer (with K2) and a heterotrimer (with K3) did
not give a unique solution.
(13) The calculated (AM1) dipole moment of methyl 2,6-diethynyl-
pyridine-4-carboxylate (1.26 D) is pointing from C(4) to the nitrogen atom,
whereas that of methyl 3,5-diethynylbenzoate (0.85 D) is pointing to the
opposite direction, i.e., to the ester group.
Org. Lett., Vol. 2, No. 21, 2000
3267