Full Paper
ternating alignment. Therefore this system can be regarded as
a 2D alternating molecular network (type I, Figure 1).
pyridinophane–tropylium complexes was generally observed in
equimolar mixtures of the complex and cyclophane, except in
the cases of 1a·3+2a and 1a·3+2c, for which both compo-
nents appeared on the surface in a 1:1 ratio. For example, the
preferential adsorption of 1a·3 was observed for the mixture
of 1a·3 and 2b even at a molar ratio of 1:100 (see Figure S3 in
the Supporting Information), which indicates a significantly
greater adsorption affinity of 1a·3 for graphite. The higher ad-
sorption probability of positively charged molecules is ascribed
on the one hand to the electrostatic interaction between the
positively charged ion and graphite to which a negative sub-
strate bias is applied,[25] and on the other hand to ion–dipole
interactions arising from the induced dipole generated on the
graphite surface by the positive charge as described above.[24]
When there is no favourable inter-component interaction,
each component assembles in individual domains resulting in
phase segregation (Table 1). For example, a mixture of complex
1c·3 and cyclophane 2c forms distinct domains because each
component favours a different pattern, the non-porous B pat-
tern for 1c·3 (D2h symmetry) and the porous pattern for 2c (C4h
symmetry, see Figure S4 in the Supporting Information), de-
spite both molecules having the same substituent (R=
CH2OC8H17).
Network patterns formed by single-component macrocycles
and pyridinophane–tropylium ion complexes
The macrocycles adopt either a C4h or D2h conformation with
all four alkyl chains adsorbed onto the surface or with only
two alkyl chains adsorbed and the other two dissolved in the
solution phase. To evaluate the intrinsic geometries of the
macrocycles, we performed DFT calculations at the B3LYP/6-
31g(d) level of theory. Both C4h and D2h geometries were opti-
mised (Figure 5, see the Supporting Information for details).
Figure 5. a) C4h and b) D2h symmetric molecular models of pyridinophane
1a.
In contrast, random mixing occurs when the molecular struc-
tures of the components are preserved in the monolayers. For
instance, random mixing was observed for a mixture of 1a·3
and 2a because both components form the same non-porous
A pattern. In a small number of cases, the random mixing of
1c·3 into domains of the porous pattern of 2c (C4h symmetry)
was also observed because 1c·3 can also adopt a C4h symmet-
ric geometry (see above).
The energy difference between the two geometries is small
(see Figure S6 in the Supporting Information), which suggests
that the macrocycles can adopt either conformation depend-
ing on other factors such as intermolecular interactions. It
should be pointed out that in both structures, the orientation
of the alkyl chain with respect to the side of the p core is
nearly perpendicular in 1a, 1c, 2a and 2c, whereas the angle
is around 1138 in 2b and 2d. Moreover, complexation with the
tropylium cation does not influence the orientation of the alkyl
chains. Therefore, in the case of ester-substituted macrocycles
1a and 2a, the hydrogen bonds between the oxygen atoms of
the ester groups and the hydrogen atoms attached to the aro-
matic rings lead to the formation of the non-porous A pattern.
The ether-substituted macrocycles form porous patterns prob-
ably to minimise the dipole–dipole repulsive energies between
the ether groups. In addition, co-adsorption of the solvent
molecule could stabilise this pattern.
Co-crystallisation occurs when a co-crystal is more stabilised
by favourable inter-component interactions than the networks
of each component. In the case of the mixture of 1c·3 and 2b,
both components adopt C4h symmetric geometries forming
the porous pattern. Although cyclophane 2b retains its intrin-
sic network structure, the non-porous B pattern of complex
1c·3 is altered. In the alternating network, the ether groups of
1c·3 and 2b are located close to each other (Figure 4f), gener-
ating attractive dipole–dipole interactions due to the favoura-
ble orientation of the ether dipoles (Figure 6a). Therefore the
main driving force for the formation of this 2D alternating mo-
lecular network of type II (Figure 1) with the porous network
pattern is attributed to the attractive dipole–dipole interac-
tions between the side-chain ether groups of all adjacent mol-
ecules. This result contrasts that of 1c·3+2c for which phase
separation and random mixing were observed because favour-
able inter-component interactions do not occur.
Although pristine 1c forms the porous pattern, the non-
porous B pattern is formed by complex 1c·3. No structural
change was detected in a control experiment performed on
1c by using a mixture of TCB/CH3CN/CHCl3(20:9:1) as solvent
(5.1ꢁ10ꢀ4 m), which excludes a solvent effect. A plausible
reason for this structural change is the enthalpically favourable
increased electrostatic interactions between the positive
charge of the complex and the induced dipole of the graphite
substrate.[24]
Although macrocycle 2d does not form a stable monolayer,
a mixture of complex 1c·3 and cyclophane 2d would form
similar dipole–dipole interactions between the side-chain ether
groups because of their similar relative positions and orienta-
tions. Indeed, the mixture forms an alternating porous pattern;
again, the network structure of complex 1c·3 is changed from
its original form (the non-porous B pattern). Although both
components adopt C4h symmetric geometries, the favourable
inter-component interaction operates in only one direction
Self-assembling behaviour of bimolecular mixtures
When one of the components has a greater affinity to the sur-
face than the other, the preferential adsorption of a single
component occurs. Indeed, the preferential adsorption of the
&
&
Chem. Eur. J. 2015, 21, 1 – 12
8
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!