Angewandte
Chemie
determined above (328C). The spectral changes in 5.0 ꢀ 10À3
m
solvent were not observed in the 5.0 ꢀ 10À4 m dilute solution
(Supporting Information, Figure S2).
Concentration-dependent changes in fluorescence spectra
were also determined. Figure 3b shows the normalized
spectra in 1,2-dichloroethane solution at 258C. Upon increas-
ing the concentration, bands at 411, 426, and 439 nm decay
rapidly, and in particular for concentrations higher than 1.0 ꢀ
10À4 m, which is probably due to self-absorption. At 5.0 ꢀ
10À3 m, above the CGC, a very weak band with lmax
=
551 nm remained (Fem < 0.01). The fluorescence band is
probably due to the aggregated species, in which excitation
energy migration efficiently occurs within and/or between the
fibrous assemblies, as reported in the case of the oligo(p-
phenylene vinylidene) (OPV) gels.[14] Time-resolved experi-
ments for 1 in 1,2-dichloroethane (5.0 ꢀ 10À3 m) at À58C
(Supporting Information, Figure S3) support this fact: the
lifetime t at 540 nm (1.5 ns) is shorter than that at 450 nm
(2.4 ns).
Figure 4. Crystal structures of 3. Molecular conformations and packing
diagrams for the J-type (a,b) and H-type (c,d) structures. Atoms in (a)
and (c) are drawn as thermal ellipsoids with 50% probability, except
for hydrogen atoms (sticks). Oxygen atoms are shown in dark gray.
To verify our strategy and characterize the fibrous super-
structure in more detail, DBA 1 was compared with
compounds 2 and 3. DBA 2 is the anti-isomer of 1
(synthesized according to Scheme S2 in the Supporting
Information). Naphthalene derivative 3 is formally derived
by removing two diyne groups from 1.
Noncentrosymmetric 1 has a dipole moment of 2.22–
3.58 Debye along its short axis as it has C2v symmetry,
whereas 2 in the C2h symmetric conformation has no dipole
moment (Supporting Information, Figure S4). The difference
in the dipole moment must affect the aggregation behavior.
For 1, molecular aggregation along the p-stacked direction
should be significantly accelerated by dipole–dipole interac-
tions. Such highly anisotropic growth of the aggregate should
yield no crystals, but instead yield 1D fibrous aggregates,
providing gelation. In contrast, the aggregation growth of 2 is
expected to be less anisotropic than that of 1, and results in a
3D crystalline superstructure. In fact, isomer 2 did not yield
any gel-like materials under the same conditions as 1, but did
yield crystalline precipitates by cooling of a chloroform
solution (Supporting Information, Figure S5).
Figure 5. Comparison of the PXRD patterns between a) a cryodried gel
of 1, and b) the J-type and c) the H-type crystals of 3. The patterns in
(b) and (c) are simulated on the basis of the single-crystal diffraction
data.
According to our strategy, boomerang-shaped ester 3
should also yield 1D fibrous aggregates as in the case of 1. In
contrary to the gelation found for 1, however, slow evapo-
ration of a solution of 3 dissolved in chloroform/methanol
yielded crystalline precipitates that contain two polymorphic
crystals: a majority being columnar crystals, and the remain-
der platelet crystals. Crystallographic analysis revealed that
molecule in the former, which has a down–down carbonyl
conformation (Figure 4a), arranges in a slipped p-stacked
fashion (i.e., J aggregation),[15] whereas those in the latter,
with an up–up conformation (Figure 4c), arrange in an
alternately p-stacked fashion (i.e., H aggregation)[16] such as
to cancel out their dipole moments. Thus, the up–up
conformation is a key structure for the J-type stacking.
Comparison of the powder X-ray diffraction (PXRD) pat-
terns of the cryodried gel of 1 with that of the crystals of 3
revealed that the gel has a similar pattern to that of the J-type
crystal, rather than the H-type (Figure 5). The peaks from the
gel at 2q = 14.8, 15.5, 23.5, and 27.48 (d spacing: 6.1, 5.7, 3.8,
and 3.3 ꢁ, respectively) in Figure 5a correspond to the peaks
of (010), (011), (10 Æ 1), and (11 Æ 1) planes of the J-type
crystal in Figure 5b: the former two peaks are attributed to
the periodic distances related to the molecular length along
the shorter axis, and the latter two to the periodic distance in
the p-stacking direction.[17] The peaks in the small angle
region at 2q = 4.3, 6.6, and 9.28 (d spacing: 21.0, 13.5, and
10.8 ꢁ) are related to the molecular length along the longer
axis. Therefore, 1 is expected to arrange mainly in the J-
aggregated fashion to form fibrous superstructures. The
reason that 3 did not give a gel despite its desired 1D
molecular arrangement is that exclusive 1D elongation of the
superstructure did not occur because its aromatic core is
smaller than that of 1.
Interestingly, when the methyl groups in 1 were replaced
by longer alkyl chains, such as ethyl, propyl, and butyl
groups,[18] the resulting annulenes gave no organogels but
rather crystalline precipitates. Preliminary X-ray crystallo-
Angew. Chem. Int. Ed. 2009, 48, 5465 –5469
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5467