Self-assembly of hydrogen-bond assisted supramolecular azatriphenylene architectures

Self-assembly of hydrogen-bond assisted supramolecular azatriphenylenearchitectures{

Matteo Palma,a Jeremy Levin,b Olivier Debever,b Yves Geerts,b Matthias Lehmann*bc and Paolo Samorı*ad

Received 4th September 2007, Accepted 23rd October 2007

First published as an Advance Article on the web 6th November 2007

DOI: 10.1039/b713570j

We report on the self-assembly of a functionalized hexaazatriphenylene into supramolecular

architectures where the single hexaazatriphenylene molecules are held together primarily through

intermolecular hydrogen bonds between amide units. Wide and small angle X-ray scattering,

polarized light microscopy, and differential scanning calorimetry revealed bulk self-organization

into columnar structures. At the surfaces, scanning force microscopy experiments showed that it is

possible to drive the self-organization from solutions of N-(2-ethylhexyl)-

hexacarboxamidohexaazatriphenylene, towards either layers on a conductive surface like graphite

or supramolecular anisotropic assemblies on an electrically insulating substrate such as muscovite

mica. The growth of this latter type of architecture is primarily driven by the physical dewetting of

the solution cast on the surface combined with intermolecular hydrogen bonds between the amide

moieties exposed in the peripheral positions that lead to the formation of the columnar stack.

Therefore, the anisotropic supramolecular azatriphenylene assemblies observed in the bulk have

been also observed in thin films on a substrate poorly interacting with the adsorbate. In view of

the interesting electronic properties of hexaazatriphenylene based architectures as n-type

semiconductors, these results might be of interest for applications in the field of organic

electronics.

Introduction

There is currently a great interest in achieving full control of

the self-organization of p-conjugated molecules into highly

ordered, anisotropic supramolecular architectures as spatially

confined electrically active nano-objects.1–4 It is indeed

well known that the order at the supramolecular level

strongly affects the electronic properties of molecular based

assemblies.5–8

In this context, columnar nanostructures made up by p–p

stacking of alkylated discotic building blocks are interesting

both as prototypes of nanowires3,9–11 and as molecular systems

forming (uniform) films with a high degree of molecular

orientation.12–15 In the discotic mesophase, the molecules self-

organize into columnar stacks;16–18 within each columnar

stack, the p-system of adjacent aromatic cores overlap

generating a one-dimensional pathway for charge transport

along the axis of the columns.19 In addition, hydrogen bonds

have been used as a tool to enhance the attractive interactions

between discotic mesogens.20–23 The resulting p-conjugated

‘‘supramolecular wires’’ are coaxially insulated from each

other by their aliphatic side-chains and it is this ability to form

self-assembling, one dimensional conducting pathways which

has aroused interest in these materials as potential charge

transport media for molecular electronic devices.9,12,13,24

Medium-size discs such as triphenylene derivatives represent

very interesting discotic molecules. The ease of substitution in

the peripheral positions grants solution processability (hence a

high level of purification), and self-organization into columnar

nanostructures that possess relatively high charge carrier

mobilities.12

Among triphenylene derivatives,25 hexaazatriphenylenes are

unique systems as potential electron carriers in view of the

peculiar nature of their conjugated core. As a matter of fact

most p-conjugated materials reported so far are electron-rich,

i.e. p-type semiconductors promoting efficient hole transport.

On the other hand, due to their high electron affinity,

azaheterocycles are promising candidates for the design

and synthesis of n-type semiconducting materials.

Hexaazatriphenylene derivatives are indeed strongly electron-

deficient heterocycles: the presence of six nitrogen atoms in the

aromatic core is expected to significantly increase the first

reduction potential, thus facilitating charge injection.26,27

Moreover, transport properties of hexaazatriphenylene stacks

are less affected by rotational disorder compared to stacks

built from triphenylene molecules.28,29

Unexpectedly, it has been shown that hexaalkylthiohexaa-

zatriphenylenes do not form columnar liquid crystalline phases

like the corresponding triphenylene derivatives, probably due

to the large negative charges on the nitrogen atoms giving rise

aNanochemistry Laboratory, Institut de Science et d’IngenierieSupramoleculaires (ISIS) – CNRS 7006, Universite Louis Pasteur, 8allee Gaspard Monge, F-67083 Strasbourg, FrancebLaboratoire de Chimie des Polymeres, CP 206/1, Universite Libre deBruxelles, Boulevard du Triomphe, B-1050 Bruxelles, Belgium,cInstitut fur Chemie Technische Universitat Chemnitz, Straße derNationen 62, D-09111 Chemnitz, Germany.E-mail: [email protected];Fax: +49-371-5311839dIstituto per la Sintesi Organica e la Fotoreattivita - Consiglio Nazionaledelle Ricerche, via Gobetti 101, I-40129 Bologna, Italy.E-mail: [email protected]; Fax: +39-051-6399844{ Electronic supplementary information (ESI) available: Synthesis andcharacterization details. See DOI: 10.1039/b713570j

PAPER www.rsc.org/softmatter | Soft Matter

This journal is � The Royal Society of Chemistry 2008 Soft Matter, 2008, 4, 303–310 | 303

to a repulsion between adjacent cores.30 To overcome this

limitation, hydrogen bonds have been used to counterbalance

the Coulombic repulsion, leading to the formation of a discotic

mesophase with relatively high charge carrier mobility (up to

ca. 0.1 cm2 V21 s21), as determined by microwave conductivity

measurements.31 In particular, it has been shown that amide

hydrogen bonds can be exploited as non covalent ‘‘clamps’’ in

a hexacarboxamidohexaazatriphenylene (HAT-CONH)32 deri-

vative bearing linear alkyl chains in peripheral positions and

that the resulting p–p distance can be as small as 0.318 nm.31

Recently we reported work on the self-assembly of a HAT-

CONH into layers whose specific electronic properties have

been investigated both in dry films using Kelvin probe force

microscopy and at the solid–liquid interface by scanning

tunneling microscopy (STM) corroborated by quantum

chemical calculations.33 In view of this, it is interesting to

inquire whether such a system is able to self-organize into

anisotropic architectures, such as fiber-like objects, also in the

bulk and/or in dried thin films at surfaces. Moreover, the

multiscale correlation between self-assembly in the bulk and at

the surface of electroactive organic architectures is still

partially unexplored.

We report here on the 2D and 3D self-organization of an

alkylated HAT-CONH molecule 1 (Fig. 1) into supramole-

cular nanostructures, as obtained by balancing the interplay of

intramolecular, intermolecular, and interfacial interactions.

The self organization in the bulk has been studied by

temperature-dependent wide and small angle X-ray scattering,

while the thermotropic properties have been investigated by

differential scanning calorimetry (DSC) and polarized optical

microscopy (POM). On the other hand, scanning force

microscopy (SFM)34,35 has been employed to unravel the

self-organization at surfaces. In particular, investigations of

dried ultrathin films revealed that it is possible to drive the self-

assembly from solutions towards either layers on a conductive

surface or supramolecular anisotropic assemblies, on an

electrically insulating substrate.

Results and discussion

Among HAT-CONHs,31 we have designed and synthesized

(see ESI{) molecule 1 (Fig. 1) because of its branched alkyl

substitution is expected to provide a good solubility in organic

solvents. Similarly to other alkylated HAT-CONHs, the

peculiar design of 1 leads to the generation of different

intramolecular and intermolecular hydrogen bonds, as

depicted in Fig. 1. The intermolecular hydrogen bonding

should promote the stacking towards a co-facial configuration

that maximizes the amplitude of the intermolecular transfer

integral, as demonstrated by quantum chemical calculations.29

The intermolecular transfer integral characterizes the strength

of the electronic coupling between the molecules in a given

geometric configuration and determines the rate of charge

hopping, and hence the mobility values. Therefore the

intercolumnar stacking arrangement and its temperature-

dependence have been studied together with the assembly of

the columns in the bulk as discussed in the next section.

Thermotropic properties and bulk self-organization

The thermotropic properties of 1 have been studied by POM,

DSC and corroborated by X-ray diffraction. POM revealed a

birefringent, highly viscous phase without any specific texture.

The material becomes shearable only above 150 uC, a typical

property of thermotropic liquid crystals (LCs). Since the

material decomposes above 250 uC, we were not able to

generate an isotropic phase. DSC studies showed a first order

transition upon cooling and heating cycles at 175 uC.

Obviously, this was observed only for materials that were

not previously heated up to their decomposition temperature

(see Table 1 and Figure SI-1 in the ESI{). The hysteresis of

2.7 uC is relatively small and the transition enthalpy, DH, is

very low for the proposed columnar structure (see X-ray part

below), thus pointing to a LC–LC transition with only a small

structural rearrangement at the transition temperature. A

second small first order transition was monitored only for the

heating curve at 119.5 uC.36 Upon cooling to room tempera-

ture no further first order transition was detected. A step with

a midpoint between 70 and 80 uC was recorded which can be

ascribed to a glass transition freezing the Colob2 to a glassy

phase g(Colob), thus maintaining the mesophase order (see

below).

Wide angle X-ray scattering (WAXS) measurements have

been performed on powder and oriented samples, the latter

being obtained by extrusion at 180 uC. Fig. 2 shows the

diffractograms detected upon heating at 30 uC (A), 140 uC (B),

180 uC (C) and upon cooling at 30 uC (D). The WAXS pattern

at 30 uC exhibits five reflections on the equator perpendicular

to the fiber direction, providing evidence for the formation of

two-dimensional self-assembled columns. Small angle X-ray

Fig. 1 Chemical structure (top view and side view) of the N-(2-ethylhexyl)-substituted hexacarboxamidohexaazatriphenylene derivative (1). From

modeling data the diameter of the aromatic core including the amide units was estimated to be of ca. 1.2 nm. The additional length of the ethylhexyl

side chains was found to be roughly 0.7 nm per chain in the case where they adopt the ideal fully stretched conformation, while from the estimation

of the columnar diameter from X-ray data the length of the side chains of 0.45 nm can be derived.

304 | Soft Matter, 2008, 4, 303–310 This journal is � The Royal Society of Chemistry 2008

scattering (SAXS) measurements at room temperature (RT)

reveal twelve equatorial reflections which can be indexed

according to an oblique 2D lattice with a = 21.1 A, b = 20.2 A,

c = 114.2u (Fig. 3). Experimental and calculated reflection

positions are in good agreement (see Table SI-1 in the ESI{).

However, our data also fit with two other possible scenarios of

columnar packing: a unit cell of a = 22.5 A, b = 20.2 A, c =

121.0u and a unit cell with a = 22.5 A, b = 21.1 A, c = 124.8u(see Table SI-2 and SI-3 in the ESI{). All three possible 2D

structures can be regarded as distorted hexagonal cells. These

cells cannot be distinguished by the X-ray pattern. The 2D

organization of columns does not change significantly over the

whole investigated temperature range (Fig. 2). In addition to

the signals on the equator, the X-ray pattern at RT exposes a

number of diffuse signals on the meridian at 15.8 A, 13.8 A,

13.3 A, 12.2 A, 7.8 4.8 A, 4.4 A and 3.9 A. The first six

reflections point to an intracolumnar superstructure and will

be discussed below. The diffuse halo at 4.4 A is attributed to

liquid-like organization of aliphatic side-chains and the broad

reflections at 3.9 A are reminiscent of the average p–p stacking

between aromatic cores. The latter distance is larger than

the previously reported separation of 3.18 A for discs of a

HAT-CONH derivative bearing linear alkylated amide side-

groups.31 The signals at 3.9 A are made of four maxima,

located at left and right from the meridian, indicating that the

aromatic planes are tilted by approximately w = (26 ¡ 12)uwith respect to the columnar axis. The separation between

mesogens along the columnar axis is therefore estimated to be

4.3 A at 30 uC. From the half width of the signal, a correlation

length L of 74 A can be calculated, which corresponds to 19

correlated mesogens in the stack.37,38 This value is three times

lower than that measured for the derivative with linear dodecyl

chains, which was found to adopt a supramolecular arrange-

ment characterized by non-tilted molecules and the smallest

observed inter-disc distance within a column.31 The steric

hindrance, brought into play by the branched 2-ethylhexyl

chains, most probably induces the larger separation between

the aromatic cores, the increased disorder and the tilt of the

mesogens along the columnar stack.

On the basis of these data, a first structural model of the

columnar organization can be built from the calculated X-ray

density and the number of molecules per unit cell assuming

that the liquid crystalline character of such system results from

the nanosegregation of aliphatic chains and polar rigid

conjugated cores (see Fig. 4).17 Calculation of the X-ray

density provides evidence for the presence of one molecule per

unit cell (cf. ESI{). With dilatometry39–41 and X-ray data

the volume occupied by the core can be calculated as Vcore =

Vunit cell 2 Vchain. The core volume gives access to the second

half-axis of the elliptic cross section by using eqn (1) and (4)

given in Fig. 4. Assuming the first half axis ra to be 5.86 A,

which corresponds to the radius from the center of the

molecule to the nitrogen of the amide group, rb is estimated to

Table 1 DSC data of 1 at a heating rate of 10 uC min21

Transition temperature, T/Transition enthalpy, DH

2. Heating g(Colob) # 80 uC (Tg) Colob1 119.5 uC/0.7 kJ mol21 Colob2 175.3uC/1.5 kJ mol21Colob3

1. Cooling Colob3 172.2 uC/1.6 kJ mol21 Colob2 # 72 uC (Tg) g(Colob)

Fig. 2 Temperature-dependent wide angle X-ray scattering (WAXS)

on an extruded, oriented fiber. Panels A–C: X-ray diffraction for the

second heating of the material. Arrows highlight signals attributed to

the superstructure. Black arrows indicate the split small angle

reflections; white arrows point to the meridional reflections. At

180 uC (panel C) the superstructure has disappeared (see missing

reflections at the points of the arrows). Panel D: recovered super-

structure upon cooling at 30 uC.

Fig. 3 Integration of the SAXS X-ray diffraction pattern of an

oriented fiber. The indexation (hk) is given for a 2D oblique unit cell

(2D space group p1) with a = 21.1 A, b = 20.2 A, c = 114.2u (cf. ESI{).

Reflections at small angles, marked with asterisks (*) and not indexed,

are attributed to an intracolumnar superstructure.


be 5.05 A. From these values, the tilt of the mesogens can be

calculated to be w = arccos(rb/ra) = 30u, thus in good

agreement with the 26u determined experimentally.

As mentioned earlier, two additional sets of four weak

reflections detected at d = 15.8 A and 13.8 A, along with three

signals on the meridian at 12.2 A, 7.8 A and 4.8 A provide

evidence for another periodicity along the columnar axis.

These reflections disappear beyond the transition at 175.2 uC(see Fig. 2B,C) and are recovered reversibly upon cooling.

Before and after the transition, the columnar 2D lattice and

the tilt of the mesogens within the stacks remain unchanged.

Thus, the first order transition can be ascribed to the

formation of a superstructure along the columnar stacks,

which may be attributed to a helicoidal self-assembly of

discotic mesogens with nearest neighbours rotationally dis-

placed by 20u.36,42–44 According to the model structure in

Fig. 4, a rotational displacement, like in a helicoidal

arrangement of mesogens, leads to more favourable p-stacking

interactions than in the case of cofacially assembled aromatic

cores. The H-bonds, with an intermolecular NH–OC average

distance of 1.9 A, are still in the range of moderate hydrogen-

bonding interactions.51 Additionally, this assembly reduces the

steric hindrance of the bulky 2-ethylhexyl groups. Thus, Fig. 4

displays a model for the columnar stacking of 1, which is

obtained as a compromise between non-covalent interactions

(H-bonds, p–p interactions and van der Waals interactions

between aliphatic chains) and steric repulsion among the

branched aliphatic chains as well as Coulombic repulsion

between adjacent aromatic cores determined by the partially

negatively charged character of the nitrogen atoms. Above

175 uC, some H-bonds, which can be regarded as clamps for

the stabilization of the supramolecular order, are probably

broken to such an extent that the columnar superstructure is

lost, although the columnar 2D lattice is preserved.

Self-organization at surfaces

In view of the columnar self-organization of molecule 1 in the

bulk, it is of great interest to investigate the behaviour of such

a system on flat solid substrates, in particular to define if such

anisotropic architectures observed in the 3D of the bulk can

also be obtained in 2D on surfaces. In this context, SFM was

employed to study the growth of ultrathin dry films of 1

prepared from very dilute solutions onto ultra-flat solid

substrates. 1H-NMR and UV-Vis studies in solution indeed

provided evidence for the high tendency of 1 to form

aggregates also in solution (see ESI Fig. SI-3 and SI-4,

respectively{). The aggregation of 1 in CHCl3 is clear at

concentrations down to 5 6 1026 mol l21.

Fig. 5a shows an intermittent contact scanning force

microscopy topographical image of a film of compound 1

prepared on a muscovite mica surface by immersion in a 2.5 61025 mol l21 solution in CHCl3 at room temperature (RT).

The image displays a network of fiber-like objects which does

not cover homogeneously the whole sample surface. These

anisotropic assemblies exhibit a height (h) of 1.5 ¡ 0.6 nm and

a quite irregular width (w) of 27.0 ¡ 14.2 nm. Their height is

comparable to the molecule diameter, within the experimental

error, suggesting an edge-on packing for the molecules on the

surface. This is in agreement with the well known tendency of

conjugated molecules bearing aliphatic side-groups to pack

‘‘edge-on’’ on the basal plane of the insulating mica surface;

such an arrangement is further driven by the hydrophobic

versus hydrophilic character of the ad-molecule and of the

substrate, respectively.45,46 A smaller height compared to the

dimension of the molecule has been in some cases observed, as

highlighted by the relatively high standard deviation. This can

be due not only to the tilting of the disks with respect to the

substrate normal and/or versus the columnar main axis (26ufrom the X-ray pattern at RT), but also to a collapse of the

side chains leading to the flattening of the molecule on the

substrate.47,48 Moreover, under ambient conditions, adhesion

forces, mainly capillary, between the SFM tip and the

hydrophilic mica surface can induce a decrease in the height

of the visualized structure.49,50 These effects have already been

Fig. 4 Model of columnar organization for mesogen 1. One mesogen

is highlighted in red and the chains are reduced to a methyl group for

clarity. Panel A: top view of a column with a slightly elliptical core

cross section. Panel B: side view. Mesogens are inclined with respect to

the columnar axis. Dotted lines visualize possible intermolecular

H-bonds. Panel C: oblique unit cell with the core cross section and the

aliphatic regions obtained by micro-segregation of non-polar aliphatic

chains and polar rigid core. FEllipse and Funit cell are the cross sections of

the core and unit cell and are calculated by eqn (1) and (2). The

corresponding volumes are obtained by eqn (3) and (4).

Fig. 5 Intermittent contact SFM topographical images: (a) and (b)

from a 2.5 6 1025 mol l21 in CHCl3 solution deposited by immersion

on mica at RT; Z-scales: (a) 8.1 nm (b) 8 nm.


observed with different (macro)molecules adsorbed on sur-

faces, including dendrimers51 and single polymeric chains.52

Given the lateral size of the observed anisotropic objects

which exceeds by one order of magnitude that of a single

molecule, it is most likely that the anisotropic architectures

forming the network consist of several discotic stacks laterally

packed. This side-to-side interaction may be due to lateral

hydrogen bonding among amide units and/or van der Waals

interactions between the alkyl side chains. Moreover, the

network observed exhibits a double-fiber structural motif with

a small amount of material adsorbed in the void. This double-

fiber pattern can be ascribed to the coalescence of colliding

droplets.53 A capillary flow mechanism due to the geometrical

constraint of the free surface by a pinned contact line, squeezes

the fluid outwards to compensate for evaporative losses,

causing the molecules dissolved in the droplets to flow towards

the contact line.4,54 As the substrate by itself cannot keep the

contact line pinned, the accumulation of solid components at

the contact line perpetuates the pinning giving rise to a flow

that replenishes the lost fluid locally.55 Combined with the

propensity of 1 to from columnar aggregates in the bulk and

aggregates in solution, this process gave rise to the formation

of the network of anisotropic assemblies shown in Fig. 5.

Therefore, the anisotropic nature of the self-organization of 1

found in 3D in the bulk, as determined by X-ray diffraction, is

also observed in 2D on surfaces. Nevertheless, on surfaces

interfacial interactions play a role and have to be taken into

account: the substrate affects and plays a not negligible role in

the self-assembly of a given system.

In this context, and in order to cast light onto the role

played by interfacial interactions on the self-organization of 1

at surfaces, we have extended our studies to ultrathin films

supported on a different substrate, namely an electrically

conducting solid support such as highly oriented pyrolytic

graphite (HOPG). We have prepared films by immersion of the

HOPG in a 1024 mol l21 solution of 1 in either chloroform or

in 1,2-dichlorobenzene. In both cases the obtained films

exhibit a layer structure. Films obtained from chloroform

solutions were characterized by the formation of both a

layer with a thickness of 1.3 ¡ 0.2 nm as well as globular

aggregates (see Fig. SI-4 in the ESI{). Films prepared from

1,2-dichlorobenzene solutions are displayed in Fig. 6: the

topographical image shows a layer with a height of 1.0 ¡

0.2 nm marked by a solid arrow in Fig. 6a; an over-layer was in

some cases observed, exhibiting an additional height of 0.6 ¡

0.1 nm, as highlighted by the dashed arrow in Fig. 6a. The

phase image in Fig. 6b reveals a different contrast of the two

layers, providing evidence for different viscoelastic proper-

ties56,57 This can be explained by a different molecular

packing. The first layer adjacent to the substrate is formed

by molecules assembled parallel to the HOPG basal plane,

even though a slight tilting of the molecules during growth

cannot be excluded a priori. This type of packing has been

confirmed by molecular resolved images obtained by STM

investigations at the solution–graphite interface33 (see also

Fig. SI-5 of the ESI{) and it is in line with previously reported

spectroscopic studies on a similar azatriphenylene-based

compound, i.e., hexakis(hexylthio)diquinoxalino[2,3-a:29,39-

c]phenazine, performed on dried thin films adsorbed on

HOPG.58 Such a behaviour, which differs from the edge-on

packing obtained on mica, is driven by the high affinity

between the aromatic core of 1 and the HOPG surface

determined by the overlap of the p orbitals of the two

systems33 and by the anchoring effect provided by the aliphatic

side chains. Given the interdisc distance found by SAXS for 1

(3.9 A), the separation between mesogens along the columnar

axis (estimated to be 4.3 A at RT in the 26u tilting

configuration determined by X-ray diffraction) and the height

of the first layer of 1.0 ¡ 0.2 nm, it is very likely that this first

layer consists of a stack of 2–3 molecules. On the other hand,

the over-layer has different viscoelastic properties and thick-

ness, thus suggesting a different kind of packing, i.e., molecules

arranged edge-on, in a similar way to what has been observed

on mica. Such an interpretation is further suggested by the

detailed topographical analysis of such an over-layer. Fig. 7

displays a higher magnification image of the over-layer shown

in Fig. 6, together with its topographical profile. The fiber-like

nature of the observed over-layer emerges, which can indeed be

described as a network of entangled anisotropic assemblies

exhibiting a width ranging from ca. 8 nm to ca. 20 nm, as only

roughly estimated because of the high density of the network.

The smaller height of the over-layer (y0.6 nm) if compared to

the height of the anisotropic assemblies on mica (y1.5 nm), as

determined by SFM topographical profile analysis, can be an

indication of a larger tilting of the disc with respect to the

columnar axis and/or of a partial interpenetration of the side

chains of the over-layer in the 1st self-assembled layer on

HOPG.

Therefore, the growth of 1 on HOPG from dichlorobenzene

solutions is characterized by a structural transition of the

Stransky–Krastanov type within the film, such that the

molecules of the undermost layers are oriented parallel to the

substrate surface whereas the outermost layers have a different

azimuthal order, as depicted in the cartoon shown in Fig. 8. In

these outermost layers the organization at the supramolecular

level is likely to exhibit a similar anisotropic nature as that

revealed in the bulk by X-ray diffraction, in view of the

negligible role played by the substrate–molecule interactions.

The typical growth obtained using dichlorobenzene as a

solvent differs from the one detected casting chloroform

Fig. 6 Intermittent contact SFM images of 1024 mol l21 in

dichlorobenzene (DCB) solution cast by immersion on HOPG at

RT: (a) topographical image and marked by a white line between the

two arrows; (b) phase image. The solid arrow indicates the first

underlying layer, while the dashed arrow highlights the ad-layer.

Z-Scales: (a) 7.8 nm (b) 27.7u.


solutions on HOPG (see ESI{). Such a difference is determined

by the kinetics of evaporation of the solvent. In fact chloro-

form has a much lower boiling point if compared to 1,2-

dichlorobenzene; therefore it is characterised by a faster

evaporation, a condition that affects the organization of

p-conjugated systems at surfaces59,60 freezing the molecules in

a given configuration. Apparently, as the evaporation rate is

decreased the molecules can organize in a different structural

motif at increased distance form the HOPG surface, where the

memory of the substrate is lost.

Therefore the different architectures of molecule 1 at

surfaces are generated as a result of combined self-assembly

and local dewetting, affected by the kinetics of the solvent

evaporation as well as by the different nature of the supporting

substrates. The substrate itself can in fact drive the self-

organization of the molecules in a template effect as seen on

HOPG, while columnar architectures co-planar to the

substrate are found when such an effect is not attained, either

because of the nature of the interaction with the substrate, as

in the case of mica, or when the memory of the substrate is

lost. Therefore a columnar assembly, revealed by X-ray

diffraction, can be also obtained on surfaces.

Conclusions

In summary, we have studied the self-organization of

columnar liquid crystal forming molecule 1 in the bulk and

at surfaces, as well as in solution. X-Ray diffraction studies

revealed the formation of columnar architectures, where the

single molecular discs are held together by H-bonds between

amide moieties. In these anisotropic assemblies, the optimiza-

tion of the supramolecular interactions (H-bonds, p–p stack-

ing, van der Waals interactions between aliphatic chains)

overrides the steric repulsion arising from the bulky branched

alkyl chains and the repulsion between the charges on the

nitrogen atoms, leading to the formation of columnar oblique

architectures with temperature-dependent intracolumnar

order. The intercolumnar organization remains unchanged

with temperature, suggesting the formation of H-bonds

between adjacent columns. The tendency to form such

intercolumnar cross-links was also observed for dry thin films

obtained casting ultra-diluted solutions on mica: a network of

fiber-like assemblies was formed, exhibiting a height compar-

able to the molecular diameter. An anisotropic assembly,

revealed by X-ray diffraction, can be therefore obtained also

on surfaces. In general, the self-organization at surfaces was

found to be influenced by a variety of boundary conditions

including the type of substrate and solvent, leading to growth

phenomena governed by the interplay of intramolecular,

intermolecular, and interfacial interactions as well as dewetting

phenomena. On electrically conductive substrates such as

HOPG, for example, SFM provided evidence for the forma-

tion of thin films. Such kinds of architectures, i.e., oriented

thin films potentially displaying high conductivity normal to

the surface, might be of interest in surface and materials

science in view of application in devices such as field-emission

displays. Interestingly the growth of the observed thin films

can be characterized by a structural transition of the Stransky–

Krastanov type depending on the solvent employed, such that

the outermost layers exhibit a different azimuthal order and an

anisotropic nature. The understanding of the self-organization

in 2D and 3D under various conditions of such a hexaaza-

triphenylene might be of interest for its application in organic

electronics, e.g., for the fabrication of solar cells and field-

effect transistors.

Experimental procedures

Molecule 1 was synthesized according to the previously

published procedures for similar derivatives.31,61 More infor-

mation is given in the ESI.{The WAXS measurements on aligned samples obtained by

extrusion were made by using a standard copper anode

(2.2 kW) source with pinhole collimation equipped with an

X-ray mirror (Osmic typ CMF15-sCu6) and a Bruker detector

(High-star) with 1024 6 1024 pixels. The SAXS measurement

on an aligned sample were performed by using a rotating

anode (mikromax 007, copper, Rigaku) source with

pinhole collimation equipped with a X-ray mirror (Osmic

Fig. 8 Cartoon showing the possible packing of molecule 1 on

HOPG from DCB solutions, as shown in Fig. 6.

Fig. 7 Intermittent contact topographical SFM image of 1024 mol l21

in DCB solution cast by immersion on HOPG at RT, and cross

sectional profile across the over-layer, marked by a white line.


type 140-0040 12) and a Bruker detector (High-star) with

1024 6 1024 pixels. The correlation length was determined

using the Scherrer formula and the half width and reflection

maximum obtained from the fit function. Calibration was

performed by using silver behenate.62

Dry thin films were prepared on freshly cleaved muscovite

mica and HOPG surfaces. We have chosen as solvents

chloroform and 1,2-dichlorobenzene (DCB) because, due to

their chlorination and aromaticity, they are presumed to give

good solubility.

We have processed the molecules in thin films using by

immersion of the sample in a solution for 5 min and then letting

the solvent evaporate in an air environment.

The SFM topographical imaging has been carried out using

commercial apparatus (dimension 3100 Nanoscope IV, Veeco,

S. Barbara, USA) operating in intermittent contact mode, at

room temperature in an air environment. Silicon tips with a

force constant k = 40 N m21 have been used. All widths

determined from SFM topographical profiles have been

corrected taking into account the tip broadening effect,

assuming a tip radius of 13 nm.45

Acknowledgements

We are grateful to Michael Bach and Prof. Jochen Gutmann

for their support and for providing us with the possibility to

measure X-ray diffraction at the Max Planck Institute for

Polymer Research Mainz. Financial support from the EU

through the Marie Curie EST - SUPER (MEST-CT-2004-

008128), IP-NAIMO (NMP4-CT-2004-500355), the RTNs

PRAIRIES (MRTN-CT-2006-035810) and THREADMILL

(MRTN-CT-2006-036040), the ERA-Chemistry project

SurConFold, the ESF-SONS2-SUPRAMATES project, the

Regione Emilia-Romagna PRIITT Nanofaber Net-Lab and

the Bundesministerium fur Bildung und Forschung (BMBF)

are gratefully acknowledged.

References

1 F. J. M. Hoeben, P. Jonkheijm, E. W. Meijer and A. Schenning,Chem. Rev., 2005, 105, 1491.

2 B. W. Messmore, J. F. Hulvat, E. D. Sone and S. I. Stupp, J. Am.Chem. Soc., 2004, 126, 14452.

3 P. Samorı, H. Engelkamp, P. A. J. de Witte, A. E. Rowan,R. J. M. Nolte and J. P. Rabe, Adv. Mater., 2005, 17, 1265.

4 R. van Hameren, P. Schon, A. M. van Buul, J. Hoogboom,S. V. Lazarenko, J. W. Gerritsen, H. Engelkamp, P. C. M.Christianen, H. A. Heus, J. C. Maan, T. Rasing, S. Speller,A. E. Rowan, J. A. A. W. Elemans and R. J. M. Nolte, Science,2006, 314, 1433.

5 F. Cacialli, J. S. Wilson, J. J. Michels, C. Daniel, C. Silva,R. H. Friend, N. Severin, P. Samorı, J. P. Rabe, M. J. O’Connell,P. N. Taylor and H. L. Anderson, Nat. Mater., 2002, 1, 160.

6 M. Van der Auweraer and F. C. De Schryver, Nat. Mater., 2004, 3,507.

7 A. P. H. J. Schenning and E. W. Meijer, Chem. Commun., 2005,3245.

8 V. Palermo and P. Samorı, Angew. Chem., Int. Ed., 2007, 46,4428–4432.

9 A. M. van de Craats, J. M. Warman, A. Fechtenkotter, D. J. Brand,M. A. Harbison and K. Mullen, Adv. Mater., 1999, 11, 1469.

10 A. S. Drager, R. A. P. Zangmeister, N. R. Armstrong andD. F. O’Brien, J. Am. Chem. Soc., 2001, 123, 3595.

11 A. C. Grimsdale and K. Mullen, Angew. Chem., Int. Ed., 2005, 44,5592.

12 D. Adam, P. Schuhmacher, J. Simmerer, L. Haussling,K. Siemensmeyer, K. H. Etzbach, H. Ringsdorf and D. Haarer,Nature, 1994, 371, 141.

13 C. D. Simpson, J. Wu, M. D. Watson and K. Mullen, J. Mater.Chem., 2004, 14, 494.

14 J. Hoogboom, P. M. L. Garcia, M. B. J. Otten, J. A. A. W.Elemans, J. Sly, S. V. Lazarenko, T. Rasing, A. E. Rowan andR. J. M. Nolte, J. Am. Chem. Soc., 2005, 127, 11047.

15 R. Friedlein, X. Crispin, C. D. Simpson, M. D. Watson, F. Jackel,W. Osikowicz, S. Marciniak, M. P. de Jong, P. Samorı,S. K. M. Jonsson, M. Fahlman, K. Mullen, J. P. Rabe andW. R. Salaneck, Phys. Rev. B, 2003, 68, 195414.

16 R. J. Bushby and O. R. Lozman, Curr. Opin. Colloid Interface Sci.,2002, 7, 343.

17 S. Chandrasekhar, in Handbook of Liquid Crystals, Vol. 2B, ed.D. Demus, J. Goodby, G. W. Gray, H.-W. Spiess and V. Vill,Wiley-VCH, Weinheim, Germany, 1998.

18 T. Kato, N. Mizoshita and K. Kishimoto, Angew. Chem., Int. Ed.,2006, 45, 38.

19 W. Pisula, A. Menon, M. Stepputat, I. Lieberwirth, U. Kolb,A. Tracz, H. Sirringhaus, T. Pakula and K. Mullen, Adv. Mater.,2005, 17, 684.

20 L. Brunsfeld, B. J. B. Folmer, E. W. Meijer and R. P. Sijbesma,Chem. Rev., 2001, 101, 4071.

21 P. A. Albouy, D. Guillon, B. Heinrich, A. M. Levelut andJ. Malthete, J. Phys. II, 1995, 5, 1617.

22 T.-Q. Nguyen, M. L. Bushey, L. E. Brus and C. Nuckolls, J. Am.Chem. Soc., 2002, 124, 15051.

23 T.-Q. Nguyen, R. Martel, M. Bushey, P. Avouris, A. Carlsen,C. Nuckolls and L. Brus, Phys. Chem. Chem. Phys., 2007, 9,1515–1532.

24 N. Boden, R. J. Bushby, J. Clements and B. Movaghar, J. Mater.Chem., 1999, 9, 2081.

25 S. Kumar, Liq. Cryst., 2004, 31, 1037.26 G. Kestemont, V. de Halleux, M. Lehmann, D. A. Ivanov,

M. Watson and Y. H. Geerts, Chem. Commun., 2001, 2074.27 M. Lehmann, G. Kestemont, R. G. Aspe, C. Buess-Herman,

M. H. J. Koch, M. G. Debije, J. Piris, M. P. de Haas,J. M. Warman, M. D. Watson, V. Lemaur, J. Cornil,Y. H. Geerts, R. Gearba and D. A. Ivanov, Chem.–Eur. J., 2005,11, 3349.

28 J. Cornil, V. Lemaur, J.-P. Calbert and J. L. Bredas, Adv. Mater.,2002, 14, 726.

29 V. Lemaur, A. da Silva Filho, V. Coropceanu, M. Lehmann,Y. H. Geerts, J. Piris, M. G. Debije, A. M. van de Craats,K. Senthilkumar, L. A. Siebbeles, J. M. Warman, J. Bredas andJ. Cornil, J. Am. Chem. Soc., 2004, 126, 3271.

30 O. Roussel, G. Kestemont, J. Tant, V. de Halleux, R. Gomez Aspe,J. Levin, A. Remacle, D. Ivanov, R. I. Gearba, M. Lehmann andY. H. Geerts, Mol. Cryst. Liq. Cryst., 2003, 396, 35.

31 R. I. Gearba, M. Lehmann, J. Levin, D. Ivanov, M. H. J. Koch,J. Barbera, M. G. Debije, J. Piris and Y. H. Geerts, Adv. Mater.,2003, 15, 1614.

32 J. C. Beeson, L. J. Fitzejerald, J. C. Gallucci, R. E. Gerkin,J. T. Rademacher and A. W. Czarnik, J. Am. Chem. Soc., 1994,116, 4621.

33 M. Palma, J. Levin, V. Lemaur, A. Liscio, V. Palermo, J. Cornil,Y. Geerts, M. Lehmann and P. Samorı, Adv. Mater., 2006, 18,3313.

34 P. Samorı, Chem. Soc. Rev., 2005, 34, 551.35 G. Binnig, C. F. Quate and C. Gerber, Phys. Rev. Lett., 1986, 56,

930.36 Noteworthy, no indication of a structural change has been

monitored for the transition at 119 uC.37 R. Jenkins and R. L. Snyder, in Introduction to X-Ray Powder

Diffractometry, ed. J. D. Winefordner, Wiley, Weinheim, 1996.38 P. Scherrer, Nachr. Ges. Wiss. Goettingen, Math.–Phys. Kl., 1918,

2, 96.39 E. Fontes, P. A. Heiney and W. H. Jeu, Phys. Rev. Lett., 1988, 61,

1202.40 B. Donnio, B. Heinrich, H. Allouchi, J. Kain, S. Diele, D. Guillon

and D. W. Bruce, J. Am. Chem. Soc., 2004, 126, 15258.41 M. Marcos, R. Gimenez, J. L. Serrano, B. Donnio, B. Heinrich and

D. Guillon, Chem.–Eur. J., 2001, 7, 1006.


42 If a helical order is assumed, the projection of the reflections at 2h =5.6u and 6.4u (d = 15.8 A and 13.8 A) on the meridian, whichcorresponds to a distance of 24.8 A, can be attributed to a helicalpitch of 74.4 A, due to the C3-symmetry of the molecule. Signals atthe layer lines at (74.4 A)/m (m = 3, 6, 9, 12, 15 etc.) should then beexpected. As a matter of fact, reflections at 12.2 A ((74.4 A)/6 =12.4 A), 7.8 A ((74.4 A)/9 = 8.3 A) and 4.8 A ((74.4 A)/15 = 5.0 A)were observed, which are close to the predicted values. However,all these signals are meridional reflections at angular positionssmaller than those corresponding to the axial separation of themesogens (4.3 A, 2h = 20.6u). Although such reflections are usuallynot observed for helix forming molecules, like DNA or proteins, asomewhat similar situation has been reported for triphenylenederivatives with chiral aliphatic chains and has been attributed tothe formation of columns with helical order. Assuming a helicalpitch of 74.4 A, nearest neighbors are rotationally displaced by 20u.

43 C. Knupp and J. M. Squire, J. Appl. Crystallogr., 2004, 37, 832.44 A. M. Levelut, P. Oswald and A. Ghanem, J. Phys., 1984, 45, 745.45 P. Samorı, V. Francke, K. Mullen and J. P. Rabe, Chem.–Eur. J.,

1999, 5, 2312.46 P. Samorı, I. Sikharulidze, V. Francke, K. Mullen and J. P. Rabe,

Nanotechnology, 1999, 10, 77.47 S. A. Prokhorova, S. S. Sheiko, A. Mourran, R. Azumi, U. Beginn,

G. Zipp, C. H. Ahn, M. N. Holerca, V. Percec and M. Moller,Langmuir, 2000, 16, 6862.

48 J. Tang, J. W. Li, C. Wang and C. L. Bai, J. Vac. Sci. Technol., B,2000, 18, 1858.

49 V. S. J. T. Van Noort, K. O. Van der Werf, B. G. De Goorth,N. F. Van Hults and J. Greve, Ultramicroscopy, 1997, 69, 117.

50 W. Zhuang, C. Ecker, G. A. Metselaar, A. E. Rowan, R. J. M.Nolte, P. Samorı and J. P. Rabe, Macromolecules, 2005, 38, 473.

51 A. Hierlemann, J. K. Campbell, L. A. Baker, R. M. Crooks andA. J. Ricco, J. Am. Chem. Soc., 1998, 120, 5323.

52 P. Samorı, C. Ecker, I. Gossl, P. A. J. de Witte, J. J. L. M.Cornelissen, G. A. Metselaar, M. B. J. Otten, A. E. Rowan,R. J. M. Nolte and J. P. Rabe, Macromolecules, 2002, 35, 5290.

53 A. P. H. J. Schenning, F. B. G. Benneker, H. P. M. Geurts,X. Y. Liu and R. J. M. Nolte, J. Am. Chem. Soc., 1996, 118, 8549.

54 R. D. Deegan, O. Bakajin, T. F. Dupont, G. Huber, S. R. Nageland T. A. Witten, Nature, 1997, 389, 827.

55 R. D. Deegan, Phys. Rev. E, 2000, 61, 475.56 P. Leclere, R. Lazzaroni, J. L. Bredas, J. M. Yu, P. Dubois and

R. Jerome, Langmuir, 1996, 12, 4317.57 M. O. Finot and M. T. McDermott, J. Am. Chem. Soc., 1997, 119,

8564.58 X. Crispin, J. Cornil, R. Friedlein, K. K. Okudaira, V. Lemaur,

A. Crispin, G. Kestemont, M. Lehmann, M. Fahlman,R. Lazzaroni, Y. Geerts, G. Wendin, N. Ueno, J. L. Bredas andW. R. Salaneck, J. Am. Chem. Soc., 2004, 126, 11889.

59 P. Samorı, V. Francke, K. Mullen and J. P. Rabe, Thin Solid Films,1998, 336, 13.

60 P. Samorı, M. Keil, R. Friedlein, J. Birgerson, M. Watson,M. Mullen, W. R. Salaneck and J. P. Rabe, J. Phys. Chem. B, 2001,105, 11114.

61 K. Kanakarajan and A. W. Czarnik, J. Heterocycl. Chem., 1988,25, 1869.

62 T. C. Huang, H. Toraya, T. N. Blanton and J. Wu, J. Appl.Crystallogr., 1993, 26, 180.


Self-assembly of hydrogen-bond assisted supramolecular azatriphenylene architectures

Documents