Self-assembly of 1-D organic semiconductor nanostructures

Thuc-Quyen Nguyen,*aRichard Martel,

bMark Bushey,

cPhaedon Avouris,

d

Autumn Carlsen,eColin Nuckolls

fand Louis Brus

f

Received 13th July 2006, Accepted 29th November 2006

First published as an Advance Article on the web 4th January 2007

DOI: 10.1039/b609956d

This review focuses on the molecular design and self-assembly of a new class of crowded

aromatics that form 1-D nanostructures via hydrogen bonding and p–p interactions. These

molecules have a permanent dipole moment that sums as the subunits self assemble into

molecular stacks. The assembly of these molecular stacks can be directed with electric fields.

Depending on the nature of the side-chains, molecules can obtain the face-on or edge-on

orientation upon the deposition onto a surface via spin cast technique. Site-selective steady state

fluorescence, time-resolved fluorescence, and various types of scanning probe microscopy

measurements detail the intermolecular interactions that drive the aromatic molecules to self-

assemble in solution to form well-ordered columnar stacks. These nanostructures, formed in

solution, vary in their number, size, and structure depending on the functional groups, solvent,

and concentration used. Thus, the substituents/side-groups and the proper choice of the solvent

can be used to tune the intermolecular interactions. The 1-D stacks and their aggregates can be

easily transferred by solution casting, thus allowing a simple preparation of molecular

nanostructures on different surfaces.

1. Introduction

The self-organization of small molecules into larger functional

nanostructures is a cornerstone of biological systems and is a

powerful tool to create novel materials with emergent or

amplified properties.1 Recently, there is increasing interest in

using elementary building blocks such as atoms and molecules

to form molecular wires or 1-D nanostructures in a controlled

and predictable fashion.2 Discotic liquid crystals,3 discovered

in 1977 by Chandrasekar and coworkers,4 are examples of

such systems. This class of materials self-assembles to form

arrays of columnar stacks. Individual stacks have been com-

pared to molecular wires because the column’s interior con-

sists of co-facially aromatic cores (conducting cores) while its

exterior is surrounded by a hydrocarbon chain (an insulating

layer).5 This arrangement of the aromatic cores yields useful

and interesting electronic and optic properties.6 For tradi-

tional discotics, the intermolecular interactions between the

subunits are weak due to the poor electrostatic attraction

between the electron rich p-surfaces.7 Several approaches havebeen used to increase the affinity between the molecules within

the columnar stack including metal–ligand interactions,8 re-

cognition of polymer strands,9 electrostatic complimentarity

between p-faces,10 and hydrogen bonds.11

Understanding self-assembly processes and the preferred

molecular arrangement of functional aromatic compounds

are prerequisites for achieving predictive models that take into

consideration the composition of the molecular components.

As an example, organic and polymeric semiconductors having

better order and packing density usually exhibit better elec-

trical conduction.12 A systematic investigation on the local

order properties of a given molecular design is therefore

important for making and improving organic semiconducting

assemblies, but a detailed investigation of the resulting struc-

ture of ordered assemblies remains challenging experimentally.

Scanning probe microscopy is now among the most powerful

techniques to probe the arrangement at a surface with sub-

nanometer resolution. Scanning tunneling spectroscopy

(STM) has been used widely to study the self-assembled

processes of aromatic molecules and alkyl thiols on metal

surfaces13 and molecules at the solid-liquid interface on highly

oriented pyrolytic graphite (HOPG) substrates.14 Atomic

force microscopy (AFM) is commonly used to probe surface

topography. Electrostatic force microscopy (EFM) is another

form of scanning probe microscopy that allows the simulta-

neous mapping of surface topography and electrostatic field

gradients. EFM has been employed to study trapped charge in

SiO2 layers15 and surface charge in semiconductors16 and

organic materials.17 A combination of proximal probe techni-

ques is however required in order to gain insights on the local

structure of the assembly and on the specific interactions

driving the assembly process.

This review discusses the assembly characteristics in solu-

tion and thin film of a new class of columnar discotic liquid

crystal materials that is held together with hydrogen bonds

and p–p interactions.18 These aromatic molecules are com-

posed of a 1-D stack of an aromatic core surrounded by a

hydrocarbon sheet. The core is a hexa-substituted aromatic (1

and 2 in Fig. 1) consisting of three meta-disposed amides that

aDepartment of Chemistry and Biochemistry, University of California,Santa Barbara, CA 93106. E-mail: quyen@chem.ucsb.edu

bDepartement de Chimie, Universite de Montreal, Montreal, Quebec,Canada

c Scripps Research Institute, La Jolla, CA, USAd IBM Watson Research Center, Yorktown Heights, NY, USAePhysics Department, Albany University, Albany, NY, USAfChemistry Department, Columbia University, New York, NY, USA

This journal is �c the Owner Societies 2007 Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 | 1515

INVITED ARTICLE www.rsc.org/pccp | Physical Chemistry Chemical Physics

are flanked by substituents other than hydrogen at each of the

remaining positions. Unlike traditional discotics, these mole-

cules stack to form molecular fibers due to a synergy between

p-stacking and hydrogen bonding to produce a relatively

strong association (molecule-to–molecule cohesion) in the

stacking direction but a comparatively weaker interaction

between fibers (Fig. 1A).18 The substituent, R0 (R00 for 2a), is

a long alkyl chain to ensure the solubility in common organic

solvents whereas R0 contains an amide group, which provides

a dipole moment and hydrogen bonding that contributes to

enhance the molecular cohesion in the stacking direction (Fig.

1). The substituents in the side groups (R0 and R00) can be used

to tailor the orientation of molecules on a surface and the

intermolecular distance within a fiber through steric interac-

tions.18c Monitoring the assembly of these mesogens in mono-

layer films by scanning probe microscopy has yielded films

with two orientations. In one surface conformation a two-

dimensional sheet results that is macroscopically polar (face-

on orientation, Fig. 2A). In the other orientation on the

surface, 1-D p-stacks result that are only a few molecules wide

but microns in length (edge-on orientation, Fig. 2B). The

length of the fibers formed on surfaces depends on the solvent,

the concentration as well as the underlying substrate. In the

bulk, these materials form highly regular and well-organized

columnar assemblies.18g In solution, fluorescence spectroscopy

gave clear indication that the underlying self-assembly process

produces 1-D stacks. The length of the fibers and their number

can be varied dramatically using different solvent and con-

centration. In ultra-thin films, 1b, 1c, 2a, and 2b assembles into

dipolar columns that have their long axes and dipole moments

parallel to the surface19 that can be directed with the electric

field. The self-assembly of molecules 1 and 2 on graphite was

examined by AFM, EFM, and ultrahigh vacuum (UHV)-

STM. There are six sections below describing the self-assembly

characteristics of 1 and 2: (1) the molecular design of this new

class of discotic crystals, (2) controlling the self-assembly by

functional groups, (3) the self-assembly in solution, (4) the

effects of solvents on the aggregation/self-assembly, (5) the

effects of concentration and temperature on the self-assembly,

and (6) the effects of the surface type on the self-assembly.

2. Molecular design

Although there are other examples of benzene rings that are

held cofacially by hydrogen bonds,20 the highly substituted

nature of these subunits gives rise to new nanostructures,

Fig. 1 Crowded aromatics and their energy minimized molecular models. Side-chains and hydrogens have been removed to clarify the view (in the

model: the R0 and R00 have been removed and red = oxygen; blue = nitrogen; grey = carbon).

Fig. 2 Schematic drawings of the face-on top view (A) and edge-on side view (B) orientation of hexa-substituted aromatics on graphite substrates.

Reprinted with permission from ref. 18c. Copyright 2004 American Chemical Society.

1516 | Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 This journal is �c the Owner Societies 2007

unique polar properties, and unusual phase behavior. For the

molecules in Fig. 1, the design principle explored was how to

use the flanking alkoxy groups for 1 and alkynyl substituents

for 2 to force the amides out of the plane of the central

aromatic ring and into a conformation that is predisposed to

form three intermolecular hydrogen bonds.

Fig. 1 shows the energy-minimized dimeric models for both

1 and 2.18 The flanking alkoxyl groups for 1 and alkynl groups

for 2 force the amides out the plane of the central aromatic

ring and into a conformation that allows the formation of

three intermolecular hydrogen bonds. The size of the func-

tional groups determines the angle of twist for the amide out

of the aromatic ring plane and consequently modulates the

distance between adjacent benzene rings. From models, the

center-to-center distance between the benzene rings is ca. 3.8 A

for 1 and ca. 3.6 A for 2 reflecting the relative size of the

alkoxyl and the alkynyl groups. Additionally, each of the

subunits has a permanent dipole moment that is perpendicular

to the aromatic ring plane. The dipoles could sum as the

molecules stack yielding columns that have a macroscopic

dipole moment, similar to the moment that is seen for some

metallomesogens and conical liquid crystals.21 These polar

columns could be used as model systems to understanding

how polar properties emerge on the nanoscale as well as how

charges transport in 1-D nanostructures.

Because this class of discotic molecules was unknown before

the studies below were initiated, a large number of derivatives

were synthesized to establish the structure/property relation-

ships. The synthetic procedures were developed by the Nuck-

olls group. Bulk self-assemblies of 1 and 2 were studied by

X-ray diffraction. X-Ray diffraction studies of 1a show two

crystalline phases below 85 1C. Above 85 1C, a fluid phase

develops and at 120 1C, the diffraction pattern has reflections

that index to a rectangular lattice with parameters a = 38 A

and b = 22 A. Rectangular packing has been observed in

other columnar liquid crystals and results from a distortion of

the lattice. This distortion could arise because the side groups

are mismatched in size, which frustrates hexagonal packing, or

are bulky, which tilts or offsets the subunits.18a The diffraction

pattern of 1b at 200 1C is dominated by a single sharp peak at

low angle, signature of columnar assemblies. Up to fifth order

diffraction peak is seen that can be indexed to a hexagonal

lattice. The lateral core-to-core separation is 21 A as expected

for columns with noninterdigitating, extended side chains.18a

There is no X-ray diffraction data available for 1c.

Besides X-ray diffraction experiments, the assembly in bulk

of 1 is deduced from a combination of experiments including

polarized light microscopy, infrared spectroscopy, and differ-

ential scanning calorimetry.18 The results from these experi-

ments show that the assembly process is dominated by the size

and polarity of the amide side groups. Solution and thin film

studies via steady state and time-resolved spectroscopy and

scanning probe microscopy also support this conclusion. For

example when the amide substituents are relatively small and

flat, such as the phenyl substituents of 1b, the material

assembles into regular cylinders that are hexagonally-packed

into millimeter-scale domains. When the phenethyl side-chain

is exchanged for the t-butylester of glycine (1a), the material is

no longer able to stack into perfect cylinders and compensates

by positioning the subunits either canted or offset. These

misalignments produce a distorted hexagonal lattice for 1a.

Further magnifying this trend, when the amide substituent is

now changed to the t-butylester of D-alanine (now made even

bulkier!), there is no discernable mesomorphism in bulk

samples.

Although there is less data, it appears that the substituents

on the side-chains of 2 have less of an influence on the

mesomorphism. The X-ray diffraction patterns for bulk sam-

ples of 2a and 2b show a columnar assembly whose primary

reflection is 18.1 and 18.7 A, and the core-to-core distances are

2.14 and 2.09 A, respectively. For both cases, higher-order

reflections allow the lattices to be indexed to a 2-D hexagonal

arrangement of columns.18e The difference in stacking pro-

pensity between 1 and 2 is likely the result of the gear-like

arrangement of side-chains in 1 that is lacking in 2 due to the

linearity of the alkynl groups. Compounds 1, 2 and other

derivatives are soluble in common organic solvents such as

methylene chloride and chloroform. Typically, thin films are

formed by spin-coating a solution between 1000–1500 rpm

onto graphite substrates.

3. Controlling the self-assembly by functional

groups

The general hypothesis about the molecular design in discotic

systems is the following: larger p-core and more compact side

groups enhance p-stacking and reduce steric interactions,

thereby reducing the intermolecular distance between mole-

cules within the stack. The studies in this section test this

assumption using a series of different molecule design (Fig. 1)

and show that the packing order and the spacing between

molecules within a columnar stack depend strongly on the

functional groups. In principle, a better molecular control on

the assembly could favorably influence charge transport along

the columnar stack. Charge transport along the columnar

stack depends strongly on the degree of p-orbital overlap,

which is influenced by the intermolecular distance and the size

of the conducting core. The former is supported by recent

simulations, which show that a molecule-to–molecule distance

of less than 4 A could lead to band-like transport as oppose to

thermal activated hopping.22 The latter has been shown

experimentally by measuring the bulk charge mobilities of

six different discotic liquid crystalline materials as a function

of the core size.23

When these molecules self-assemble on a surface, they can

either adopt a face-on or edge-on orientations depending on

the functional groups. Fig. 2 shows the schematic drawing of

the two orientations. The face-on orientation leads to the

formation of patches or islands on surface (Fig. 2A) whereas

the edge-on orientation results in the formation of columnar

stacks (Fig. 2B).18 If the molecules obtain the edge-on orienta-

tion on the surface, the height of these fibers/stacks should be

the same or within the range of the molecules’ diameters. One

feature of the assembly of derivatives of 1 and 2 that is

potentially useful has to do with their polar properties and

how they develop on the nanoscale. To investigate the assem-

bly and polarity of 1 and 2 on these length scales, conditions

were found to produce films that have less than a monolayer of

This journal is �c the Owner Societies 2007 Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 | 1517

coverage on highly ordered pyrolitic graphite (HOPG) sub-

strates through spin casting. The topography, polarity/mole-

cular orientation, and molecular packing of these films can be

measured using AFM, EFM, and STM, respectively.

3.1. Polar monolayers

When a bulky side group R00 is used such as the t-butylester of

glycine (1a), the molecules obtain the face-on orientation. This

functional group can rotate in and out of the aromatic plane.

The length of this group measured from the amide unit isB7.9

A. If this side group rotates out of the aromatic plane, it acts

as a spacer between the subunits, and hence, reducing the

intermolecular interactions between molecules substantially.

The energy minimized molecular model of 1a shows that two

out of three t-butylester groups point out of the aromatic

plane (point in the plan of the page) whereas the third one

(blue circle) is almost in the same plane with the aromatic ring

(Fig. 3, right). The model of 1b is also shown in Fig. 3 for a

comparison. The phenyl groups are aligned with the aromatic

ring. The height of the overlayer shown in Fig. 3 is ca. 0.5 nm.

This height is consistent with the distance through the aro-

matic plane (ca. 0.4 nm) but much too small for the diameter

of the molecules measured in molecular models and bulk to be

ca. 1.2 nm. Also, the monolayer height of B0.5 nm is

consistent with the functional group dimension because when

deposited onto a surface, the surface restricts the rotation of

the functional group. Thus, the functional group is aligned in

the same plane as the aromatic ring. We infer from this data

that the bulky glycine ester sidechains of 1a decrease the

intermolecular attraction relative to the attraction with the

graphite so that the molecules form 2-D, monolayer sheets

with aromatic cores parallel to the substrate. Thus, the per-

manent dipole moment is perpendicular to the sample surface.

To further confirm the orientation of the molecules on the

surface, EFM technique was used to measure the dipole

moment (surface charge density) of the molecules caused by

the change in the columnar orientation with respect to the

surface. EFM probes the long-range electrostatic attraction

between a conductive AFM cantilever and a conductive sub-

strate.24 A schematic of the EFM setup is shown along with

the results in Fig. 4. Surface charges (Q) and permanent

dipoles (P) generate interactions with the total charge on the

EFM tip through a Coulombic interactions. The attraction

between the cantilever and the substrate is proportional to the

square of the voltage difference between them. Thus, a sinu-

soidal voltage V = Vdc + Vac(sin ot) applied to the tip yields

components of the attractive force at zero, o and 2o frequen-

cies. The force at o gives a local measure of Q and P, and the

force at 2o provides information about the material dielectric

constant. The details of EFM experiment setup and theory are

described elsewhere.25 EFM is extremely sensitive being able

to measure a net charge of about 0.1 electron at a distance of

10 nm.24a Therefore, EFM could potentially detect the differ-

ence in dipole moment from columns of opposite polarity.

Due to the face-on orientation with the permanent dipole

moment is perpendicular to the sample surface, a large EFM

signal is expected compared to molecules that self-assemble

with edge-on orientation. The important finding shown in Fig.

4A and B is that the AFM and EFM images look essentially

the same and the film has a universal net negative charge. The

EFM signal is a convolution of the molecular charge due to

charge transfer from the substrate and/or a perpendicular,

permanent dipole moment. The molecules have a net dipole

Fig. 3 Left: topographic AFM image (10 � 10 mm) and the cross section of compound 1a spin-cast from methylene chloride on a HOPG

substrate. Right: energy minimized molecular models of compounds 1a and 1b. AFM image is reprinted with permission from ref. 18c. Copyright

2004 American Chemical Society.

1518 | Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 This journal is �c the Owner Societies 2007

perpendicular to the surface when the amides are forced out of

the aromatic plane. Because of the partial positive charged on

the amide N–H coupled with its out of plane conformation, a

N–H/p interaction with the graphite substrate as shown in Fig.

4C could result. Also, the reason that 1a does not adsorb with

the dipole pointed toward the surface due to electron repulsion

between the electron rich surface (graphite surface) and the

partial negative charged on the amide CQO. It is more stable

for the partial positive charged on the amide N–H pointed

toward the graphite substrate. In fact, similar interactions

have already been observed between amines and carbon

nanotube surfaces.26 Effectively, the surface could serve to

orient the amide substituents and thereby direct the molecular

dipoles.27 Electron transfer from the graphite to the mono-

layers could also be a contributor to the EFM signal.

3.2. Isolated stacks

Remarkably, the opposite surface orientation, where the

columns align parallel to the surface (edge-on orientation,

Fig. 2B), is adopted for molecules 1b, 1c, 2a, and 2b shown

in Fig. 1 and others tested that form hexagonal arrangements

in bulk. Fig. 5 shows the topography AFM images and the

cross sections of 1b, 1c, 2a, and 2b films spin-cast from

methylene chloride solution. The four molecules form long

molecular fibers on graphite at low concentration. The length

Fig. 4 (A) 2.2 � 2.2 mm AFM image (data scale: 0–3 nm), (B) 1o EFM image of the same film (data scale: 0–400 Hz), (C) Dipole and electron

transfer in thin films of 1a, and (D) EFM set-up for measuring surface charges and permanent dipoles in self-assembled columnar films. Reprinted

with permission from ref. 18c. Copyright 2004 American Chemical Society.

This journal is �c the Owner Societies 2007 Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 | 1519

of these molecular fibers is concentration dependent (see

Section 6 below).

Fig. 5A shows the topographic image and the cross section

of molecule 1c with the height of the fiber height at 0.581 �0.183 nm from the surface. The height of fibers formed by

molecule 1b, 2a, and 2b are 1.51 � 0.215 nm (Fig. 5B), 2.17 �0.152 nm (Fig. 5C), and 1.91 � 0.178 nm (Fig. 5D), respec-

tively. It is important to note that the height of molecule 1c is

much smaller than its diameter (ca. 1.2 nm from the molecular

model) but the fact that they form wire structure rules out the

possible face-on orientation. We speculate that the molecules

still achieve the edge-on orientation but they are tilted relative

to the surface at a small angle. This is probably induced by the

steric hindrance of the bulky side group with free rotation as

discussed above leading to weaker interactions between adja-

cent molecules within the columnar stack and stronger inter-

actions with the surface. From the cross sections of Fig. 5B, C,

and 5D, 1b, 2a, and 2b clearly stack with an edge-on orienta-

tion because the heights of the fibers are about the diameters

of the molecules. Both the aromatic backbone and the side-

groups contribute to the height value but the aromatic back-

bone plays a dominant role due to its rigidity whereas the side

groups are some what flexible. The height of 2b is slightly

larger than 1b and this is possibly due to the substitution of the

ether oxygen atom in R00 position for a carbon atom to form a

triple bond leading to a larger p-core in 2b. These values

correlate well with the 1.8 nm column spacing measured from

bulk synchrotron X-ray diffraction.18e For 2, substitution of

the alkyl chain by the phenyl group in R0 position increases the

diameter of the fiber further. The nanostructures observed in

the AFM images are not individual fibers. There are several

fibers packed next to each other (a sheet of fibers) due to

strong van der Waals interactions between the fibers. The

samples often comprise of individual fibers and sheet of fibers

Fig. 5 Topographic AFM images and cross sections of films spin-cast from methylene chloride on HOPG substrates for: (A) compound 1c

(1.54 � 1.54 mm), (B) compound 1b (2 � 2 mm), (C) compound 2a (3 � 3 mm), and (D) compound 2b (2 � 2 mm).

1520 | Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 This journal is �c the Owner Societies 2007

with various widths. The width of the nanostructures depends

on the spin-speed and the molecule concentration: the higher

the spin-speed and concentration, the thinner the width.

The EFM images of compounds 1b and 1c shown in Fig. 6

(bottom) at sub-monolayer coverage on graphite trace the

same morphology as the topographic image (top). Similar to

molecule 1a, these films have a universal net negative charge.

Electron transfer from the graphite to the molecular column

can also contribute to the EFM signal in compound 1c (shown

in Fig. 6A), but it is unlikely the main contribution to the

signal because the hydrocarbon exterior chains minimize

charge transfer with the substrate. Overall, this EFM result

together with the AFM data confirms the orientation of

compound 1c on the surface: the molecules achieve the edge-

on orientation but they are tilted at a small angle relative to the

surface. This tilt results in a significant dipole moment per-

pendicular to the stacking direction and also efficient charge

transfer can take place. EFM measurements performed on

columnar stacks of compound 1b reveals that they have

essentially very weak or no measurable dipole or charge, which

is consistent with the column axial dipole moment being

parallel to the surface (Fig. 1B). Moreover, the hydrocarbon

exterior insulates these stacks against charge transfer from the

substrate. This is quite different compared to typically discotic

liquid crystals, which do not form isolated 1-D structures but

rather 2-D sheets with face-on orientation because molecular

cohesion from the weak p-stacking forces holding the column

together are nearly equal to the van der Waals intermolecular

forces between alkyl side chains. For 1b, 2b, and 2a, the self-

association in the stacking direction outweighs these van der

Waals interactions leading to isolated stacks with edge-on

orientation.

From the topographic AFM images, all three molecules self-

assemble to form fibers on graphite. However, we cannot

determine exactly how the molecules pack together at mole-

cular level. To investigate further the molecular packing in

these fibers, we use high-resolution UHV scanning tunneling

microscopy (STM) images such as those in Fig. 7. STM images

of molecule 1c (Fig. 7A), molecule 1b (Fig. 7B), and molecule

2b (Fig. 7C) have been acquired with atomic and molecular

resolution. These images clearly show that the three molecules

self-assemble to form very different structures at the molecular

level. The different structures result from a subtle change in

molecule–molecule and molecule–surface interactions due to

the functional groups.

STM images of molecule 1c are shown in Fig. 7A. The

images of the assembly are consistent with a packing of

molecules bonded together through hydrogen bonding to form

long molecular wires with the long alkyl chains (R00) anchored

down to the graphite lattice. The orientation of the alkyl

chains is clearly visible in Fig. 7A. The affinity of the molecules

toward the graphite surface imposes a preferential orientation

Fig. 6 Topographic (top) and 1 o EFM (bottom) images of compound 1c (A) (AFM data scale: 0–4 nm and EFM data scale: 0–15 Hz) and

compound 1b (B) (data scale: 0–2 nm for AFM and 0–200 Hz for EFM). All images are 2 � 2 mm. AFM and EFM images are reprinted with

permission from ref. 18c. Copyright 2004 American Chemical Society.

This journal is �c the Owner Societies 2007 Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 | 1521

of the wire relative to the lattice direction of graphite. The

detailed structure of the wires is somewhat intriguing. The

long fiber is composed of a periodic pattern made of large

bright spots alternated by small dark regions. Although these

images clearly resolve the lattice of the graphite and the

molecular arrangement of the molecules, they do not provide

enough contrast to provide an unambiguous structure to the

molecular packing. Details from the STM images allow us

however to add important observations about the packing

structure of molecule 1c. By counting the number of alkyl

chains in these images, we note that there are about 108 side

chains for twelve bright features (or clusters) in the left image

(Fig. 7A). Therefore, each cluster has a total of 9 alkyl chains

located on both sides from their center. Since each molecule

has three alkyl chains, each cluster units has therefore to be

composed of three molecules. Due to bulky side-groups, the

molecules cannot stack closely to each other as a result of a

steric hindrance; thus, they pack as a group of three with a

dislocation in between the groups of three molecules.

By changing the side groups with a compact unit (molecule

1b and 2b), we expect that the interactions among molecules

will be enhanced, thereby reducing the spacing between mole-

cules within the stack. Fig. 7B shows the STM images of

molecule 1b. Surprisingly, all the fibers show kinks in the same

direction and no alkyl chain perpendicular to the stacking

direction. The width of the fiber is too large (B4 nm) to be a

diameter of a single molecule. Therefore, it is possible that

three fibers bundle to form helix, and the kinks are helical

pitches. If this is the case, molecule 1b would form a sheet-like

conformation similar to b-sheet. The actual molecular packing

Fig. 7 STM images of: (A) compound 1c (left image: 25� 25 nm, right image: 10� 10 nm) (B) compound 1b (left image: 70� 70 nm, right image:

10 � 10 nm), and (C) compound 2b (left image: 15 � 15 nm, right image: 5 � 5 nm). Individual bright spots are monomers.

1522 | Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 This journal is �c the Owner Societies 2007

motif is rather compact and seems too complicated to interpret

from the STM image. It is clear that modeling will be needed

in this case.

For molecule 2b, we substitute the oxygen atom in R00 by an

alkyne to create a larger p-core with the expectation that the

larger p-core would strengthen p–p interactions within a

columnar stack. Inspection of Fig. 7C shows that molecule

2b attains straight individual fibers with similar intermolecular

distances. This result supports our hypothesis that the larger

p-core enhances p–p interactions and reduces the spacing

between molecules. Using the graphite lattice as a reference,28

we estimate the spacing between molecules is ca. 4.3 A.

Although most molecules stack to form fibers, there are still

some individual molecules (monomers) present on the surface.

The monomers appear as isolated bright spots as seen in some

areas of the film. The diameter of a molecule is ca. 2.0 nm in

agreement with the value obtained by AFM studies. Although

the exact mechanisms of the molecular packing motifs are not

conclusive, the STM results reveal that functional groups

influence the molecular packing motif and the self-assembly

process.

The results in this section show that it is possible to use the

functional side groups of the molecules as a tool to control the

molecular packing, orientation, and intermolecular spacing of

overcrowded aromatics in thin films. By further tuning the

property of this class of material, these nanostructures could

be used as model systems to study electronic transport proper-

ties in 1-D organic semiconductors.

4. Evidence of self-assembly in solution

This section focuses on the self-assembly characteristics of

hydrogen-bond enforced, crowded aromatics into 1-D

p-stacks in solution. It is not only important to master the

intermolecular assembly but also to understand how to form

and interface these assemblies with useful substrates.8,29 The

question asked is do whether the molecules self-assemble in

solution prior to deposition onto a surface. The assembly can

be monitored with optical spectroscopy due to ground and

excited states being electronically delocalized upon assembly.

In this work we utilize site-selective wavelength dependent

fluorescence spectroscopy––a valuable technique to under-

stand the self-organization in solution for molecular systems

such as proteins, peptides, and membrane-bound probes30—to

determine whether 2a and 2b self-assemble into columnar

structures in solution. Our approach is based on the different

fluorescence response from individual molecules, which we call

‘‘monomers,’’ compared to molecules stacked to form col-

umns, which we refer to ‘‘aggregates’’ (see Fig. 8B for the

schematic drawing). The terms ‘‘fibers’’ or ‘‘aggregates’’ will

be used interchangeably to refer to the 1-D nanostructures

when visualized by microscopy. Furthermore, these fibers/

aggregates form species with ground and excited electronic

properties distinct from that of the monomer. We selectively

excite the monomers or the aggregates to obtain information

about the assembly processes in solution. The excited state of

the aggregate is significantly longer lived than from the

isolated molecules and easily observed in time-resolved photo-

luminescence experiments. Aggregates for both 2a and 2b

could be detected in methylene chloride solutions as low as

10�7 M. Spectroscopy and microscopy of films cast from

solutions of 2a and 2b provide strong evidence that the

structural integrity of the aggregates in solution is preserved

during the transfer process onto the substrate.31 Thus, we can

use the solvent and concentration to control the film morphol-

ogy. Further organization of these aggregates into films with

higher order structures is determined by whether the surface is

hydrophilic or hydrophobic.

In these discotic materials, a balance must be met between

the subunits affinity for itself and for its solvent medium in a

solution and for its substrate in a film. On the one hand, when

the stacking forces between molecules are great the assemblies

grow too large to be effectively solvated and precipitate into

ill-defined superstructures. On the other hand, if the associa-

tion between the subunits is too slight (i.e. due to the steric of

the bulky side group), the p–p overlap between conjugated

cores will be diminish and interactions with the surface will

dominate resulting in a face-on orientation on the surface. The

steric bulk of the side-chains on this crowded core is one of the

determining factors of how well these molecules self-assemble.

Besides the side group and its interplay with the core size, the

solvent is also a crucial factor in the self-assembly process.

This process can be an entropy driven process when the

solvent molecules are released upon the self-assembly forma-

tion,32 but, if the solvent has very strong interaction with the

molecules, the aggregate cannot be formed. Below we monitor

the assembly of 2a and 2b, that have different groups on the

alkynes flanking their amides (either hydrocarbon for 2b or

phenyl for 2a), and probe how this assembly is crucially

affected by its environment (solvent, temperature, and con-

centration). These solution-phase aggregates pre-determine

Fig. 8 Schematic representation of possible molecular arrangements

of compound 2: (A) monomer; (B) aggregates/fibers of different sizes;

(C) a bundle of aggregates/fibers; (D) a complex, ill-defined packing

structure. Reprinted with permission from ref. 18g. Copyright 2004

American Chemical Society.

This journal is �c the Owner Societies 2007 Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 | 1523

(both in terms of quality and quantity) the type of nanostruc-

tured morphology manifest in thin films.

Here we show that the self-assembly process exists in

solution prior to the deposition of the supramolecular assem-

bly on the surface. Fig. 9 shows the normalized absorption and

photoluminescence (PL) of 2a and 2b in methylene chloride.

Several observations can be made from this simple compar-

ison. First, for 2a, the substitution of central ring with the

phenylethynyl groups at R00 substituent causes a red shifting of

ca. 40 nm in the absorption and ca. 20 nm in the emission

compared to 2b with an alkyl group substituting the alkyne on

the central aromatic ring at the R00 position. The red shifting is

due to the larger conjugated core of compound 2a compared

to 2b. Second, for two compounds, the PL spectra have a

bifurcated peak structure. As explained below, one peak is

from the monomers and the peak appeared at a longer

wavelength is from the aggregates. Third, the ratio of the

monomer/aggregate peak of 2a is larger, meaning that the

concentration of monomers in the solution is higher in 2a than

in 2b. Last, while the monomer emission is similarly shifted

compared to that of the absorption spectra, the aggregate

emission from compound 2a is blue shifted relative to that

from the aggregate peak of 2b. One explanation for this blue-

shift in the aggregates with the larger aromatic core is that the

freely rotating phenyl group on the ethynl-substituents causes

problems in the packing.

The PL spectrum of 1c is also included for a comparison.

Similar to 2, the PL spectrum of 1c shows two peaks but both

peaks are blue-shifted compared to 2. The monomer peak is

blue-shiftedB50 nm from the monomer peak of 2b, a result of

having smaller conjugated core. The aggregate peak is also

blue-shifted B25 nm compared to 2b. Additionally, this peak

does not shift to longer wavelength with increasing the excita-

tion wavelength as in the case of 2. Therefore, we conclude

that 1c could form dimmer and trimer in solution. Compound

1c does not form large aggregates as in 2a and 2b because of

the weak intermolecular interactions, a result of bulky func-

tional groups as discussed in the previous section.

The emission spectra can be understood by considering the

fact that the molecules 2a and 2b stack in solution into

columnar aggregates. These supramolecular structures form

helical stacks with an intermolecular distance for the co-

facially arranged aromatic rings to be within their van der

Waals radii. In PL, the excited state of the monomers is

localized on a single molecule and the emission wavelength

depends mostly on the size of the conjugated core. In contrast,

the emission of the columnar structure is red-shifted relative to

the monomer, because it has an excited state that is delocalized

over several subunits/molecules within the stack. The deloca-

lization of the excited state wave function across several

molecules lowers the energy relative to the localized excited

state wave function of a monomer resulting in a red-shifted

luminescence from the aggregate. Of course, if the aggregate

has poor overlap between the molecular subunits, the excita-

tion will be more localized and therefore at a wavelength closer

to the monomer emission. Thus, the position of the emission

peak provides a good characterization of the intermolecular

interaction in the aggregates; i.e., aggregates with stronger

interactions present emission spectra that are red shifted

compared to weakly interacting aggregates.33 The monomer

peak at about 370–390 nm is clearly resolved from the broad

structure in the emission at wavelengths between 400 nm and

600 nm, which originates from the aggregates. The broad

emission spectrum of the aggregates suggests a wide distribu-

tion of sizes or the presence of different forms of aggregates.

This point will be discussed further below.

Fig. 10 shows the PLE (Fig. 10A) and PL (Fig. 10B) spectra

of 2b in methylene chloride collected as the excitation and

emission are moved to longer wavelengths. Because both 2a

and 2b show similar results, only the results of 2b are presented

below. A large red shift is observed in both PLE and PL

spectra as the emission and excitation wavelengths increase.

Generally, this red shift is either due to the strong electronic

interactions among the chromophores or to the motion of

chromophores in a restricted condensed media such as in a

very viscous solution, in a membrane or in a micelle. 34 For the

experiments performed here, the solvents used, i.e. methylene

chloride and methanol, have similar viscosity (0.413 vs. 0.544

centipoice at 25 1C) while the viscosity of dodecane is three

times larger (1.383 centipoice at 25 1C). However, there is no

shift in the PL spectra for dodecane with increased excitation

wavelength (see Section 5 for the effect of solvent).35 We can

rule out the contribution of the red-shifted PL due to the

increase in the solution viscosity. In addition, there are many

cofacially stacked p-systems that show red-shifted absorbance

and emission upon aggregation.36

PLE spectroscopy has no interfering background so it is

very sensitive at low concentration. This is why the red-shifted

absorption/aggregate bands in Fig. 10A at low concentration

are observed only in the PLE spectra and not in the UV-visible

absorption spectra. As seen in Fig. 10, the emission from the

aggregates can be enhanced by preferentially exciting toward

longer wavelengths (400 nm–500 nm) of PLE band maxima.

Fig. 10B plots the normalized PL spectra of 2b excited at

290 nm (the peak of monomer exciton absorption) and at

longer wavelength, 360 nm–480 nm, (aggregate band). The

fluorescence spectra show that the aggregate emission shifts

continuously to the red with the excitation wavelength. This

can be explained in terms of a wide distribution in the

aggregate length and number. From the PL spectra, it is

Fig. 9 Normalized (at the monomer peak) absorption and PL spectra

of 2a and 2b in methylene chloride (concentration = 10�5 M, l =

(excitation) = 290 nm for 2b and 320 nm for 2a). Reprinted with

permission from ref. 18g. Copyright 2004 American Chemical Society.

1524 | Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 This journal is �c the Owner Societies 2007

difficult to estimate the size or number of the aggregates

because their absorption spectra consist of a superposition

of states with transition energies varying continuously. More-

over, their absorption cross-sections are unknown. Neverthe-

less, the results presented here provide direct evidence for

aggregate formation in solution. The aggregate and its size

can be visualized and estimated from the AFM images in

Section 6 below.

To further characterize the aggregates in solution, we

measure the lifetimes of the excited states at different excita-

tion and emission wavelengths. Because of the difference in the

excited states between the aggregates and the monomers, the

lifetimes are expected to be quite different. Our ability to excite

selectively one or the other allows us to test this idea; i.e. the

aggregate excited state would have a longer lifetime due to the

delocalization through several p-stacked monomer units. With

the correct choice of excitation and emission wavelengths, the

excited state dynamics of the monomer and the aggregate are

measured independently. From the PLE results in Fig. 10, we

select the excitation for the aggregate in the red portion of the

main absorption band (Z 400 nm), while the monomer is

obtained with excitation in the blue portion (r350 nm). Fig.

11 shows the fluorescent decay dynamics of compound 2b in

methanol. Methanol was used because there was a substantial

amount of the monomer present. In this experiment, the

monomer is excited at 320 nm and the emission is collected

at 380 nm. Since the aggregate emits much further to the red,

we set the excitation to 420 nm and the emission was collected

at 480 nm. The decay curve at 320 nm excitation in Fig. 11 was

fitted to a single exponential with the lifetime of 1.72 ns—

presumably dominated by the monomer lifetime. The decay at

420 nm excitation, however, fits to two lifetimes, one at 5.72 ns

and one at 1.46 ns. The longer lifetime, 5.72 ns, is clearly the

decay dynamic of the aggregates. The difference in the shorter

lifetimes (1.72 ns versus 1.46 ns) is likely due to the error in

assuming that the decay curve is comprised of two distinct

lifetimes when in reality it is likely a composite of aggregates of

different length and therefore different lifetimes. In methylene

chloride 2b also shows two distinct lifetimes—one at 7.08 ns

and the other at 1.70 ns.

5. The effects of the solvents on the aggregation

In this section, we examine the effects of solvents on the

molecular packing of molecule 2b. The aggregation process

is a result of more favorable molecule–molecule interactions

over molecule–solvent interactions.37 If the interaction be-

tween solutes and solvent molecules is stronger than molecule–

molecule and solvent–solvent interactions, there would be very

few aggregates in the solution and mostly monomers. On the

contrary, if the molecule-solvent interaction is weak, there

would be very few monomers present in solution. Fig. 12

shows the PLE and PL spectra of 2b (normalized at the

monomer emission peak) in methylene chloride, methanol, a

mixture of methylene chloride and methanol (4 : 1 by volume,

referred to as ‘‘mix’’ in Fig. 12 and 13), and in dodecane at the

same concentration (10�4 M). The aggregate absorption and

emission peak positions and intensities are used to estimate

their numbers and sizes. From Fig. 6, the aggregate peak is

red-shifted and increases in intensity as it goes from dodecane,

pure methanol, mixed solvent, to methylene chloride. This

Fig. 10 (A) PLE and (B) normalized PL spectra of compound 2b in

methylene chloride collected at different emission and excitation

wavelengths. Reprinted with permission from ref. 18g. Copyright

2004 American Chemical Society.

Fig. 11 The fluorescence decays of compound 2b in methanol: the

monomer was excited at 320 nm, and its emission was monitored at

380 nm (opened circles); the aggregates were excited at 420 nm and

their emission were monitored at 480 nm (crosses); the fit to each curve

is shown as solid line. Reprinted with permission from ref. 18g.

Copyright 2004 American Chemical Society.

This journal is �c the Owner Societies 2007 Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 | 1525

implies that there is a higher degree of aggregation in pure

methylene chloride than in the other solvents tested. In

methanol, however, the number of monomers increases due

to the additional competition for the hydrogen bonds between

the solvent and the molecules.

As we go from methylene chloride to mixed solvents and

then to pure methanol, the monomer emission increases

smoothly with the methanol concentration probably due to

the competition for the hydrogen bonding with the solvent.

This effect is clearly seen from the spectra of the mixed solvent.

However, there is still a measurable number of aggregates in

these two solutions indicating that hydrogen bonding is not

the only force that holds the molecules together in the aggre-

gates. For example, p–p interactions and solvophobic effects

between the aromatic cores can also play an important role in

holding molecules together. As discussed below, methanol

takes an active part in the formation of a new complex

aggregate structure for 1.

Surprisingly, the aggregate emission in dodecane is very

weak and most, if not all, of the PL emission is from the

monomer. This effect is likely due to the strong solvophobic

interactions between the solvent and the long alkyl chains

attached to the monomer. The solvent stabilizes the monomer

in the solution and therefore interferes with the packing

process. This result is unexpected and coupled with the results

from the studies with methanol indicate that not only hydro-

gen bonds but also solvophobic forces from the sidechains and

the core facilitate the assembly.

Fig. 13 presents the AFM topology of films made with

compound 2b (B10�4 M) on graphite obtained by spin casting

from methylene chloride, methanol, a mixture of methylene

chloride/methanol, and dodecane at room temperature. The

film in Fig. 13A from methylene chloride is composed of short

fibers arranged in a multi-layered film ordered over small

domains of about 0.1 � 0.1 mm. In the same conditions, the

film cast obtained with the mixed solvent (methylene chloride/

methanol) in Fig. 13B has a much lower number of fibers.

Thus, the AFM results agree very well with our interpretation

of the PL data for both solutions. The electronic structure of

the films evidenced by the long wavelength emission survives

the casting process. The films obtained from the methanol

solution contain large, tangled bundles with complex packing

structures (Fig. 13C). For this solution, the PL spectra show a

lower concentration of aggregates relative to the monomer,

probably resulting from the strong competition for hydrogen

bonding with the solvent. This competitive action prevents the

formation of fibers with isolated columns (with a diameter of

1 B 1.9 nm) in favor of larger physical aggregates (with the

diameter of ca. 30–50 nm) that have no influence on the optical

properties of 2b in methanol. Presumably, the aggregate size in

this bundle is smaller and there are fewer numbers of aggre-

gates in methanol-cast film than for methylene chloride-cast

film. This is in agreement with the blue shifted peaks of the

aggregate emission in the PL emission when compared to the

fibers emission observed in methylene chloride. These obser-

vations are consistent with poor molecular packing of units

that interact weakly with each other. Thus, it is likely that the

solvent participates in this packing structure by competing for

hydrogen bonding.

The case of the dodecane solution is rather simple. It forms

a film on graphite (see Fig. 13D) with only a few aggregates/

fibers. The material deposited is mostly embedded in a fea-

tureless layer. In dodecane, although hydrogen bonds can be

formed, this system does not favor the assembly and yields

poor results. This is again consistent with the PL results

described above.

As we spin cast the solutions of 2a and 2b in methylene

chloride onto graphite (B10�4 M), high-aspect ratio aggre-

gates can be seen in the film morphology shown in Fig. 14. The

topographic images of 2b (Fig. 14A) and 2a (Fig. 14B) are of

short fibers packed closely together in random directions. The

fibers of 2b are straight and packed closely to from ordered

sheets/layers. However, this is not the case for 2a. The fibers of

compound 2a have the tendency to tangle together and form

poorly defined layers. We speculate that this disorder is related

to the freely rotating phenyl groups on the triple bonds that

introduce steric interaction in the assembled structure.

The emission spectra of these fibrous films can be correlated

with the type of solvent that is used for the casting. The

important point is not that the exact morphology of a film is

present in solution, but rather that the size and number of the

Fig. 12 (A) PLE and (B) PL spectra of compound 2b in different

solvents: methylene chloride (MeCl2, solid curve), mixture of methy-

lene chloride and methanol (4 : 1 by volume) (Mix, dashed curve),

methanol (MeOH, dashed-dot curve), and dodecane (long and two

short dashed curve). They are normalized at the monomer peak

(310 nm for PLE and 360 nm for PL). The excitation was at 320 nm

and the emission was collected at 360 nm. Reprinted with permission

from ref. 18g. Copyright 2004 American Chemical Society.

1526 | Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 This journal is �c the Owner Societies 2007

underlying aggregates in a thin film is a consequence of

concentration and the type of solvent used. Fig. 15 displays

the PLE and PL of compound 2b spin cast from methylene

chloride and methanol solutions. When the films are spin-cast

from methylene chloride the aggregate absorption and emis-

sion bands are relatively high in intensity (vide infra). For

methanol, the long wavelength aggregate emission is substan-

tially attenuated (Fig. 15) implying that the delocalized

aggregates are less in films from methanol. Again, this is

qualitatively the same behavior observed in solution.

6. The effects of concentration and temperature

To understand how temperature effected the aggregation of

2b, we collected the PL of 2b in methylene chloride at room

temperature, at 50 1C, at 80 1C, at �2 1C, and at �40 1C. We

observed no apparent change in the PL intensity or the

emission peak in the PLE and PL spectra as the solution

was cooled down. At high temperature (50 1C and 80 1C), the

PL intensity decreases slightly as excited at the monomer

band, probably, due to the monomer PL quantum yield is

lower at high temperature. There is no apparent change both

in the PL intensity and the emission peak as we excited the

aggregate band at 80 1C. Films cast from this solution show

both fibers and some monomers. Heating these films at 80 1C

does not change the orientation of the molecules with respect

to the surface (face-on or edge-on orientations) or the number

of the monomers/aggregates. We observe only phase segrega-

tion between the monomers and the aggregates. Heating the

films spin-coated from 1b and 2b solutions in methylene

Fig. 13 Topographic images of compound 2b in (A) methylene chloride (scan size: 1 � 1 mm), (B) mix (scan size: 1 � 1 mm), (C) methanol (scan

size: 1.5 � 1.5 mm), and (D) dodecane (scan size: 1 � 1 mm) spin-cast on graphite. Reprinted with permission from ref. 18g. Copyright 2004

American Chemical Society.

Fig. 14 Topographic images from methylene chloride spin-cast on graphite for: (A) compound 2b (scan size: 1.5 � 1.5 mm) and (B) compound 2a

(scan size: 2 � 2 mm). Reprinted with permission from ref. 18g. Copyright 2004 American Chemical Society.

This journal is �c the Owner Societies 2007 Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 | 1527

chloride at 245 1C for 1 min and cooling down at 5 1C min�1,

only aggregates/fibers are observed as shown in Fig. 16. These

fibers are packed closely together with the same height as

observed in individual fibers before annealing.

Next, we investigate the effects of concentration on the way

the molecules pack in solution and on the film morphology.

Fig. 17 presents the PLE and PL spectra (normalized at the

monomer emission peak) of 2b in methylene chloride at

different concentrations, B10�4 M, B10�5 M, B10�6 M.

As the concentration increases, the aggregate band is red-

shifted and increases in the intensity. This peak is at about

340 nm at low concentration and shifted to 365 nm at higher

concentration, and it is related to the fibrous structures seen

with the AFM (Fig. 5C, 13, 16, and 18). In addition, the

number of fibers/aggregates increases with concentration, as

seen in the enhancement of the red-shifted emission (aggregate

emission at 450 nm). The PL spectra are also red shifted with

an increase in concentration, which is a signature of the

formation of aggregates in solution. This can also be seen in

the film morphology as shown in Fig. 18. At high concentra-

tion, the fibers formed are short and pack closely together.

They also form a multilayered film on the graphite substrate

(Fig. 18A and Fig. 15A). As the concentration of the solution

decreases, the fiber length increases and creates isolated stacks

of fibers, oriented according to the graphite lattice (Fig. 18B).

This can be explained using basic concepts of crystal growth,

but in this case, it is a 1-D growth process. There are several

factors that influence the numbers of fibers formed and their

length. These are the nucleation sites, the nucleation rate, and

the growth rate. Generally, the nucleation sites, which deter-

mines the number of fibers formed, increase with the concen-

tration, and this is the same for the nucleation rate.38 The

growth rate is controlled by the diffusion process, and hence,

at higher concentration or higher temperature, the probability

that the molecules encounter a nucleation site increases.38

Consequently, the monomer depletion rate is much higher

for concentrated solution than for dilute solution. Thus, there

are more fibers in concentrated solution, but they are much

shorter compared to the low concentration. Therefore, con-

centration can be used to control the length and number of the

fibers formed.

Fig. 15 PLE and PL of compound 2b in methylene chloride (MeCl2,

solid curve) and in methanol (MeOH, dashed curve) spin-cast on

quartz substrates. The spectra are normalized at the monomer peak.

The PLE spectra were collected at 360 nm and the PL spectra were

excited at 320 nm. Reprinted with permission from ref. 18g. Copyright

2004 American Chemical Society.

Fig. 16 AFM images of 2b (A, scan size: 3 � 3 mm) and 1b (B, scan

size: 2 � 2 mm) films annealing at 245 1C for 1 min and the cross

section of 1b. The height of the fiber is 1.558 nm.

1528 | Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 This journal is �c the Owner Societies 2007

7. The effects of the surface type on the

self-assembly

In this section, we examine how the surface polarity influences

the self-assembled nanostructures.

Spin-casting 1b from CH2Cl2 solution (B10�6 M) onto the

basal plane of highly ordered pyrolytic graphite (HOPG) or

onto a silicon wafer having a 200 nm thermally grown silicon

oxide layer produces very thin, elongated nanostructures. A

typical atomic force micrograph (AFM) is shown in Fig. 19A

on HOPG. The fibers are only one-molecule high, a few

molecules in width, but microns in length. On graphite, the

molecules form straight fibers in registry with the graphite

lattice. When using a cleaned silicon wafer (200 nm silicon

oxide) as the substrate, the fibers were much longer and tended

to bundle together to form ropes. This aggregation may arise

from the mismatch between the polar and hydrophilic silicon

oxide and the hydrophobic columns trying to minimize con-

tact. Nonetheless, we found conditions to form very long ropes

of columns that are about 50 nm in diameter. They are shown

in the micrograph Fig. 19B where each fiber is about 25

molecules wide. In contrast to the regular arrangement of

fibers on HOPG that arrange along the graphite lattice, on

silicon oxide these ropes orient randomly on the glassy,

amorphous silicon oxide layer.

The interaction between the molecule and the surface is

important for guiding the patterns formed by the film, such as

the orientation of the fibers and their size or number on

the surface. For example, compound 1b deposited on graphite

at low concentration (B10�6 M) gives fibers that are

spread out to form a monolayer on graphite, due to the strong

van der Waals interaction between graphite and the molecules.

The fibers are straight, packed parallel and in registry with

the graphite lattice at either 601 or 1201 angles. On Si/SiO2

(see Fig. 19B), the fibers have the tendency to form

bundles (7–50 nm in diameter) and orient randomly. This is

a result of a less favorable interaction between this hydrophilic

substrate and the hydrophobic aggregates. Thus, the fibers

bundle up together to minimize the interactions with the

surface and optimize the van der Waals interaction among

the fibers.

It is possible to use the electric field between vertical

(between two ITO-coated glass plates) and lateral electrodes

(between two gold electrodes patterned on silicon wafers) to

direct the self-assembly of these 1-D nanostructures. For the

lateral electrodes, there are several physical parameters influ-

enced the alignment of the molecules under an external electric

field: the width of the electrodes, the height/thickness of the

electrode, the gap size, the strength of the electric field, and the

number of molecules between the gap area.

Fig. 17 (A) PLE and (B) PL spectra of compound 2b in methylene

chloride at different concentration: 9.45 � 10�4 M (dot-dashed curve),

9.45 � 10�5 M (dashed curve), 4.74 � 10�5 M (dot curve), and 4.74 �10�6 M (solid curve). The PLE (collected at 420 nm) and PL (excited at

270 nm) spectra are normalized at the monomer peak. Reprinted with

permission from ref. 18g. Copyright 2004 American Chemical Society.

Fig. 18 AFM images of compound 2b in methylene chloride at different concentration spin-cast on graphite: (A) 9.45 � 10�5 M (scan size: 1.5 �1.5 mm) and (B) 9.45 � 10�6 M (scan size: 3 � 3 mm). Reprinted with permission from ref. 18g. Copyright 2004 American Chemical Society.

This journal is �c the Owner Societies 2007 Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 | 1529

8. Summary and future work

This review details the design, synthesis and the self-assembly

in solution and in thin film of hexasubstituted aromatics that

form columnar nanostructures via hydrogen bonds and p–pinteractions. The structures of 1 and 2 are unique synthetic

targets due to the highly substituted central benzene ring. The

functional side groups can be used as a tool to control the

molecular packing, orientation, and intermolecular spacing of

overcrowded aromatics in thin films. When using compact side

groups as in molecule 1b, the molecules stack to form fiber and

these fibers bundle together to form helices. Using compact

side groups and a larger p-core as in 2b leads to the formation

of straight and isolated fibers with a reducing of the inter-

molecular spacing within a columnar stack. The assembly of

these subunits produces polar stacks in solution and can be

transferred onto surfaces. Spin casting films of 1a produces

polar monolayers (face-on orientation) whereas in 1b, 1c, 2a,

and 2b, molecules self-assemble forming long fibers that can be

visualized with AFM, EFM, and STM. The exact mechanisms

of molecular packing motifs in 1b and 1c are unclear at the

moment and appears to be quite complex. Thus, modeling is

needed to further understand these molecular packing motifs.

The processing conditions such as solvent, concentration, and

type of substrate used provide control on the size, orientation

and number of the aggregates formed. Finally, it is possible to

align 1-D organic nanostructures using DC voltage, and that

the alignment depends on the strength of the electric field, the

thickness of the electrodes, the gap size, and the number of

molecules within the gap area. Using this method with more

electroactive molecules represents an untested method to

assemble and wire 1-D organic semiconductors in devices.

By the proper tuning of the chemical functionality, these

columnar structures can be used as model systems for inves-

tigating the charge transport in 1D organic semiconducting

nanostructures.

Acknowledgements

This work was supported by the Office of Basic Energy

Sciences, US D.O.E. (#DE-FG02-01ER15264) and US

National Science Foundation CAREER award (#DMR-02-

37860). TQN thanks the UCSB setup fund for financial

support. We thank Dr Arnold Tamayo for the energy mini-

mized molecular models of compounds 1a and 1b shown

in Fig. 3.

References

1 (a) J.-M. Lehn, Supramolecular Chemistry, VCH, Weinheim,1995; (b) J.-M. Lehn, Struct. Bonding, 2000, 96, 3–29; (c) G. M.Whitesides and B. Grzybowski, Science, 2002, 295, 2418–2421; (d)D. S. Lawrence, T. Jiang and M. Levett, Chem. Rev., 1995, 95,2229–2260; (e) M. M. Conn and J. Rebek, Chem. Rev., 1997, 97,1647–1668; (f) D. Philp and J. F. Stoddart,Angew. Chem., Int. Ed.Engl., 1996, 35, 1155–1196; (g) M. Muthukumar, C. K. Ober andE. L. Thomas, Science, 1997, 277, 1225–1232; (h) L. J. Prins, D.N. Reinhoudt and P. Timmerman, Angew. Chem., Int. Ed., 2001,40, 2382–2426.

2 (a) J. Wu, A. C. Grimsdale and K. L. Muellen, J. Mater. Chem.,2005, 15, 41–52; (b) T. Sekiguchi, S. Yoshida and K. M. Itoh,Phys. Rev. Lett., 2005, 95, 106101–106104; (c) G. H. Woehrle, M.G. Warner and J. E. Hutchison, Langmuir, 2004, 20, 5982–5988;(d) A. Gesquiere, M. M. S. Abdel-Mottaleb, S. De Feyter, F. C.De Schryver, F. Schoonbeek, J. van Esch, R. M. Kellogg, B. L.Feringa, A. Calderone, R. Lazzaroni and J.-L. Bredas, Langmuir,2000, 16, 10385–10391; (e) J. V. Barth, J. Weckesser, C. Cai, P.Gunter, L. Burgi, O. Jeandupeux and K. Kern, Angew. Chem.,Int. Ed., 2000, 39, 1230–1234; (f) B. Wen, C. Liu and Y. Liu,Chem. Lett., 2005, 34, 396–397; (g) B. W. Messmore, J. F. Hulvat,E. D. Sone and S. I. Stupp, J. Am. Chem. Soc., 2004, 126,14452–14458; (h) A. Datar, R. Oitker and L. Zang, Chem.Commun., 2006, 15, 1649–1651; (i) S. Lin, M. Li, E. Dujardin,C. Girard and S. Mann, Adv. Mater., 2005, 17, 2553–2559; (j) Z.Deng, Y. Tian, S.-H. Lee, A. E. Ribbe and C. Mao, Angew.Chem., Int. Ed., 2005, 44, 3582–3585; (k) J. Bae, J.-H. Choi, Y.-S.Yoo, N.-K. Oh, B.-S. Kim and M. Lee, J. Am. Chem. Soc., 2005,127, 9668–9669.

3 (a) S. Chandrasekhar and G. S. Ranganath, Discotic liquidcrystals, Rep. Prog. Phys., 1990, 53, 57–84; (b) C. Destrade, P.Foucher, H. Gasparoux, H. T. Nguyen, A. M. Levelut and J.Malthete, Disk-like mesogen polymorphism, Mol. Cryst. Liq.Cryst., 1984, 106, 121–46; (c) S. Chandrasekhar, S. Prasad andS. Krishna, ‘‘Recent developments in discotic liquid crystals’’,Contemp. Phys., 1999, 40, 237–245.

4 (a) S. Chandrasekhar, B. K. Sadashiva and K. A. Suresh, Liquidcrystals of disc-like molecules, Pramana, 1977, 9, 471–80; (b) S.Chandrasekhar, B. K. Sadashiva, K. A. Suresh, N. V. Madhu-sudana, S. Kumar, R. Shashidhar and G. Venkatesh, Disk-likemesogens, J. Phys. Colloq., 1979, 3, 120–4.

5 S. Chandrasekhar, Columnar, discotic nematic and lamellarliquid crystals their structures and physical properties, Handb.Liq. Cryst., 1998, 2B, 749–780.

Fig. 19 Topographic images of 1b on HOPG (A, scan size: 2 � 2 mm) and on a cleaned silicon wafer with 200 nm of thermally grown silicon

dioxide (B, scan size: 3 � 3 mm).

1530 | Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 This journal is �c the Owner Societies 2007

6 (a) N. Boden and B. Movaghar, Handb. Liq. Cryst. Res., 1998,2B, 781–798; (b) N. Boden, R. J. Bushby, J. Clements and B.Movaghar, J. Mater. Chem., 1999, 9, 2081–2086; (c) A. M. Van deCraats, J. M. Warman, A. FechtenkKtter, J. D. Brand, M. A.Harbison and K. Mullen, Adv. Mater., 1999, 11, 1469–1472; (d)V. Percec, M. Glodde, T. K. Bera, Y. Miura, I. Shiyanovskaya,K. D. Singer, V. S. Balagurusamy, P. A. Heiney, I. Schnell, A.Rapp, H.-W. Spiess, S. D. Hudson and H. Duan, Nature, 2002,419, 384–387; (e) V. Percec, G. Johansson, J. Heck, G. Ungar andS. V. Batty, J. Chem. Soc., Perkin Trans. 1, 1993, 1411–1420; (f) J.Simon and C. Sirlin, Pure Appl. Chem., 1989, 61, 1625–1629; (g)O. E. Sielcken, L. A. van de Kuil, W. Drenth, J. Schoonman andR. J. M. Nolte, J. Am. Chem. Soc., 1990, 112, 3086–3093; (h) T.Christ, B. Gluesen, A. Greiner, A. Kettner, R. Sander, V.Stuempflen, V. Tsukruk and J. H. Wendorff, Adv. Mater., 1997,9, 48–52; (i) D. Adam, P. Schuhmacher, J. Simmerer, L.HNnssling, K. Siemensmeyer, K. H. Etzbach, H. Ringsdorf andD. Haarer, Nature, 1994, 371, 141–143; (j) L. Schmidt-Mende, A.FechtenkKtter, K. Mullen, E. Moons, R. H. Friend and J. D.MacKenzie, Science, 2001, 293, 1119–1122.

7 C. A. Hunter and J. K. M. Sanders, J. Am. Chem. Soc., 1990, 112,5525–5534.

8 (a) Metallomesogens, ed. J. L. Serrano, VCH, New York, 1996;(b) J. Simon, P. Bassoul, in Phthalocyanines Properties andApplications, Vol. 2, ed. C. C. Leznoff and A. B. P. Lever,VCH, New York, 1989, ch. 6; (c) A. G. Serrette, C. K. Lai andT. M. Swager, Chem. Mater., 1994, 6, 2252–2268.

9 V. Percec, Handb. Liq. Cryst. Res., 1997, 2B, 259–346.10 (a) M. Muller, C. Kubel and K. Mullen, Chem.–Eur. J., 1998, 4,

2099–2109; (b) H. Bengs, M. Ebert, O. Karthaus, B. Kohne, K.Praefcke, H. Ringsdorf, J. H. Wendorff and R. Wuestefeld, Adv.Mater., 1990, 2, 141–144; (c) M. Weck, A. R. Dunn, K. Matsu-moto, G. W. Coates, E. B. Lobkovsky and R. H. Grubbs, Angew.Chem., Int. Ed., 1999, 38, 2741–2745.

11 Hydrogen bonds used to stabilize p stacks: (a) Y. Matsunaga, N.Miyajima, Y. Nakayasu, S. Sakai and M. Yonenaga, Bull.Chem.Soc. Jpn., 1988, 61, 207–210; (b) L. Brunsveld, H. Zhang, M.Glasbeek, J. A. J. M. Vekemans and E. W. Meijer, J. Am. Chem.Soc., 2000, 122, 6175–6182, and references therein; (c) Y. Yasuda,E. Iishi, H. Inada and Y. Shirota, Chem. Lett., 1996, 7, 575–576;(d) M. P. Lightfoot, F. S. Mair, R. G. Pritchard and J. E. Warren,Chem. Commun., 1999, 19, 1945–1946; (e) E. Fan, J. Yang, S. J.Geib, T. C. Stoner, M. D. Hopkins and A. D. Hamilton, J. Chem.Soc., Chem. Commun., 1995, 12, 1251–1252; (f) D. Ranganathan,S. Kurur, R. Gilardi and I. L. Karle, Biopolymers, 2000, 54,289–295; (g) C. M. Paleos and D. Tsiourvas, Angew. Chem., Int.Ed. Engl., 1995, 34, 1696–1711, and references therein; (h) M. J.Brienne, J. Gabard, J.-M. Lehn and I. Stibor, J. Chem. Soc.,Chem. Commun., 1989, 24, 1868–1870; (i) D. Goldmann, R.Dietel, D. Janietz, C. Schmidt and J. H. Wendorff, Liq. Cryst.,1998, 24, 407–411; (j) G. Ungar, D. Abramic, V. Percec and J. A.Heck, Liq. Cryst., 1996, 21, 73–86; (k) V. Percec, C.-H. Ahn, T.K. Bera, G. Ungar and D. J. P. Yeardley, Chem.–Eur. J., 1999, 5,1070–1083; (l) J. Malthete, A. M. Levelut and L. Liebert, Adv.Mater., 1992, 4, 37–41; (m) D. Pucci, M. Veber and J. Malthete,Liq. Cryst., 1996, 21, 153–155.

12 (a) T.-Q. Nguyen, J. Wu, V. Doan, S. H. Tolbert and B. J.Schwartz, Science, 2000, 288, 652–656; (b) M. Funahashi and J.-I.Hanna, Phys. Rev. Lett., 1997, 78, 2184–2187. Polymer assembly;(c) L. Schmidt-Mende, A. Fechtenkotter, K. Mullen, E. Moons,R. H. Friend and J. D. MacKenzie, Science, 2001, 293,1119–1122; (d) H. Sirringhaus, P. J. Brown, R. H. Friend, M.M. Nielsen, K. Bechgaard, B. M. W. Langeveld-Voss, A. J. H.Spiering, R. A. J. Janssen, E. W. Meijer, P. Herwig and D. M. deLeeuw, Nature, 1999, 401, 685–688. Discotic assembly; (e) V.Percec, M. Glodde, T. K. Bera, Y. Miura, I. Shiyanovskaya, K.D. Singer, V. S. K. Balagurusamy, P. A. Heiney, I. Schnell, A.Rapp, H. W. Spiess, S. D. Hudson and H. Duan, Nature, 2002,419, 384–387.

13 (a) L. Mueller-Meskamp, B. Luessem, S. Karthaeuser, S. Prikho-dovski, M. Homberger, U. Simon and R. Waser, Phys. StatusSolidi A, 2006, 203, 1448–1452; (b) A. Ogunrinde, K. W. Hippsand L. Scudiero, Langmuir, 2006, 22, 5697–5701; (c) H. Inoue, G.Yoshikawa and K. Saiki, Jpn. J. Appl. Phys. P. 1, 2006, 45,1794–1796; (d) B. Luessem, L. Mueller-Meskamp, S. Karthaeu-

ser, R. Waser, M. Homberger and U. Simon, Langmuir, 2006, 22,3021–3027; (e) A. A. Dameron, L. F. Charles and P. S. Weiss, J.Am. Chem. Soc., 2005, 127, 8697–8704; (f) G. M. Florio, T. L.Werblowsky, T. Mueller, B. J. Berne and G. W. Flynn, J. Phys.Chem. B, 2005, 109, 4520–4532; (g) V. Arima, E. Fabiano, R. I. R.Blyth, F. Della Sala, F. Matino, J. Thompson, R. Cingolani andR. Rinaldi, J. Am. Chem. Soc., 2004, 126, 16951–16958; (h) E.Marx, K. Walzer, R. J. Less, P. R. Raithby, K. Stokbro and N. C.Greenham,Org. Electronics, 2004, 5, 315–320; (i) A. A. Dameron,J. W. Ciszek, J. M. Tour and P. S. Weiss, J. Phys. Chem. B, 2004,108, 16761–16767; (j) G. V. Nazin, S. W. Wu and W. Ho, Proc.Natl. Acad. Sci. U. S. A., 2005, 102, 8832–8837.

14 (a) W. Mamdouh, H. Uji-i, J. S. Ladislaw, A. E. Dulcey, V.Percec, F. C. De Schryver and S. De Feyter, J. Am. Chem. Soc.,2006, 128, 317–325; (b) H. Fukumura, H. D. I-i, H. Uji-I, S.Nishio, H. Sakai and A. Ohuchi, ChemPhysChem, 2005, 6,2383–2388; (c) S. De Feyter and F. C. De Schryver, J. Phys.Chem. B, 2005, 109, 4290–4302; (d) J. A. A. W. Elemans, M. C.Lensen, J. W. Gerritsen, H. van Kempen, S. Speller, R. J. M.Nolte and A. E. Rowan, Adv. Mater., 2003, 15, 2070–2073; (e) M.D. Watson, F. Jaeckel, N. Severin, J. P. Rabe and K. Muellen, J.Am. Chem. Soc., 2004, 126, 1402–1407; (f) D. G. Yablon, H.Fang, L. C. Giancarlo and G. W. Flynn, ACS Symp. Ser., 2002,810, 205–224.

15 (a) G. H. Buh, H. J. Chung and Y. Kuk, Appl. Phys. Lett., 2001,79, 2010; (b) C. Y. Ng, T. P. Chen, H. W. Lau, Y. Liu, M. S. Tse,O. K. Tan and V. S. W. Lim, Appl. Phys. Lett., 2004, 85, 2941; (c)H. Dongmoa, J. F. Carlottia, G. Bruguierb, C. Guascha, J.Bonneta and J. Gasiot, Appl. Surf. Sci., 2003, 212–213, 607.

16 (a) T. Uchihashi, T. Okusako, T. Tsuyuguchi, Y. Sugawara, M.Igarashi, R. Kaneko and S. Morita, J. Appl. Phys. Pt. I, 1994, 33,5573–5576; (b) T. D. Krauss and L. E. Brus, Phys. Rev. Lett.,1999, 83, 4840–4843; (c) T. D. Krauss, S. O’Brien and L. E. Brus,J. Phys. Chem. B, 2001, 105, 1725; (d) O. Cherniavskaya, L. Chenand L. Brus, J. Phys. Chem. B, 2004, 108, 4946; (e) P. M. Bridger,Z. Z. Bandic, E. C. Piquette and T. C. McGill, Appl. Phys. Lett.,1999, 74, 3522.

17 (a) H. Yamada, T. Fukuma, K. Umeda, K. Kobayashi and K.Matsushige, Appl. Surf. Sci., 2002, 188, 391–398; (b) J. N. Barisci,R. Stella, G. M. Spinks and G. G. Wallace, Electrochim. Acta,2000, 46, 519; (c) H. Takano and M. D. Porter, J. Am. Chem.Soc., 2001, 123, 8412; (d) S. Howell, D. Kuila, B. Kasibhatla, C.P. Kubiak, D. Janes and R. Reifenberger, Langmuir, 2002, 18,5120; (e) J. W. Hong, Sang-il Park and Z. G. Khim, Rev. Sci.Instrum., 1999, 70, 1375.

18 (a) M. L. Bushey, A. Hwang, P. W. Stephens and C. Nuckolls, J.Am. Chem. Soc., 2001, 123, 8157–8158; (b) M. L. Bushey, A.Hwang, P. W. Stephens and C. Nuckolls, Angew. Chem., Int. Ed.,2002, 41, 2828–2831; (c) T.-Q. Nguyen, M. L. Bushey, L. E. Brusand C. Nuckolls, J. Am. Chem. Soc., 2002, 124, 15051–15054; (d)W. Zhang, D. Horoszewski, J. Decatur and C. Nuckolls, J. Am.Chem. Soc., 2003, 125, 4870–4873; (e) M. L. Bushey, T.-Q.Nguyen and C. Nuckolls, J. Am. Chem. Soc., 2003, 125,8264–8269; (f) M. L. Bushey, T.-Q. Nguyen, W. Zhang, D.Horoszewski and C. Nuckolls, Angew. Chem., Int. Ed., 2004,43, 5446–5453; (g) T.-Q. Nguyen, R. Martel, P. Avouris, M.Bushey, C. Nuckolls and L. E. Brus, J. Am. Chem. Soc., 2004,126, 5234–5242.

19 T.-Q. Nguyen, R. Martel, A. Carlsen, P. Avouris, M. Bushey andC. Nuckolls, Nano. Lett., unpublished work.

20 Hydrogen bonds used to stabilize p stacks: (a) Y. Matsunaga, N.Miyajima, Y. Nakayasu, S. Sakai and M. Yonenaga, Bull. Chem.Soc. Jpn., 1988, 61, 207–210; (b) L. Brunsveld, H. Zhang, M.Glasbeek, J. A. J. M. Vekemans and E. W. Meijer, J. Am. Chem.Soc., 2000, 122, 6175–6182, and references therein; (c) Y. Yasuda,E. Iishi, H. Inada and Y. Shirota, Chem. Lett., 1996, 7, 575–576;(d) M. P. Lightfoot, F. S. Mair, R. G. Pritchard and J. E. Warren,Chem. Commun., 1999, 19, 1945–1946; (e) E. Fan, J. Yang, S. J.Geib, T. C. Stoner, M. D. Hopkins and A. D. Hamilton, J. Chem.Soc., Chem. Commun., 1995, 12, 1251–1252; (f) D. Ranganathan,S. Kurur, R. Gilardi and I. L. Karle, Biopolymers, 2000, 54,289–295; (g) C. M. Paleos and D. Tsiourvas, Angew. Chem., Int.Ed. Engl., 1995, 34, 1696–1711, and references therein; (h) M. J.Brienne, J. Gabard, J.-M. Lehn and I. Stibor, J. Chem. Soc.,Chem. Commun., 1989, 24, 1868–1870; (i) D. Goldmann,

This journal is �c the Owner Societies 2007 Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 | 1531

R. Dietel, D. Janietz, C. Schmidt and J. H. Wendorff, Liq. Cryst.,1998, 24, 407–411; (j) G. Ungar, D. Abramic, V. Percec and J. A.Heck, Liq. Cryst., 1996, 21, 73–86; (k) V. Percec, C.-H. Ahn, T.K. Bera, G. Ungar and D. J. P. Yeardley, Chem.–Eur. J., 1999, 5,1070–1083; (l) J. Malthete, A. M. Levelut and L. Liebert, Adv.Mater., 1992, 4, 37–41; (m) D. Pucci, M. Veber and J. Malthete,Liq. Cryst., 1996, 21, 153–155.

21 (a) D. Kilian, D. Knawby, M. A. Athanassopoulou, S. T.Trzaska, T. M. Swager, S. Wrobel and W. Haase, Liq. Cryst.,2000, 27, 509–521; (b) H. Zimmermann, R. Poupko, Z. Luz and J.Billard, Z. Naturforsch., A: Phys. Phys. Chem. Kosmophys., 1985,40, 149–160; (c) J. Malthete and A. Collet, J. Am. Chem. Soc.,1987, 109, 7544–7545.

22 (a) A. Rochefort and R. Martel, Ph. Avouris. Nano Lett., 2002, 2,877–880; (b) E. Baillard, S. Hamel and A. Rochefort, OrganicElectronics, 2006, 7, 144–154.

23 A. M. Van de Craats and J. M. Warman, Adv. Mater., 2001, 13,130–133.

24 (a) B. D. Terris, J. E. Stern, D. Rugar and H. J. Mamin, Phys.Rev. Lett., 1989, 63, 2669; (b) T. Uchihashi, T. Okusako, T.Tsuyuguchi, Y. Sugawara, M. Igarashi, R. Kaneko and S.Morita, Jrn. J. Appl. Phys. Pt. I, 1994, 33, 5573; (c) M. P. deSanto, R. Barberi, L. M. Blinov, S. P. Palto and S. G. Yudin,Mol. Mater., 2000, 12, 329–345.

25 (a) T. D. Krauss, S. O’Brien and L. E. Brus, J. Phys. Chem. B,2001, 105, 1725; (b) J. Jiang, T. D. Krauss and L. E. Brus, J. Phys.Chem. B, 2000, 104, 11936; (c) O. Cherniavskaya, L. Chen, V.Chen and L. E. Brus, J. Phys. Chem. B, 2004, 108, 4946–4961; (d)C. H. Ben-Porat, O. Cherniavskaya, L. E. Brus, K.-S. Cho and C.B. Murray, J. Phys. Chem. A, 2004, 108, 7814–7819; (e) L. Chen,R. Ludeke, X. Cui, A. G. Schrott, C. R. Kagan and L. E. Brus, J.Phys. Chem., 2005, 109, 1834–1838; (f) L. Chen, O. Cherniavs-kaya, A. Shalek and L. E. Brus, Nano Lett., 2005, 5, 2241–2245.

26 (a) J. Kong and H. Dai, J. Chem. Phys. B, 2001, 105, 2890–2893; (b)M. Shim, A. Javey, N. W. S. Kam and H. Dai, J. Am. Chem. Soc.,2001, 123, 11512; (c) K. Bradley, M. Briman, A. Star and G.Gruner,Nano Lett., 2004, 4, 253–256; (d) J. Liu, M. J. Casavant, M.Cox, D. A.Walters, P. Boul, W. Lu, A. J. Rimberg, K. A. Smith, G.T. Colbert and R. E. Smalley, Chem. Phys. Lett., 1999, 303,125–129; (e) M. J. O’Connell, P. Boul, L. M. Ericson, C. Huffman,Y. Wang, E. Haroz, C. Kuper, J. Tour, K. D. Ausman and R. E.Smalley, Chem. Phys. Lett., 2001, 332, 461–466; (f) M. Sano, A.Kamino, J. Okamura and S. Shinkai, Nano Lett., 2002, 2, 531–533.

27 Surface poling has been observed locally in polymer films: X. Q.Chen, H. Yamada, Y. Terai, T. Horiuchi, K. Matsushige and P.S. Weiss, Thin Solid Films, 1999, 353, 259–263.

28 We measured the C–C distance of our graphite area and com-pared to the known value of C–C bond in graphite. We used the

C–C distance of 1.42 A in graphite lattice as the reference. TheC–C distance in our STM experiments was larger than for theknown value due to the piezo drift, a common problem encoun-ters in STM measurements. Often, we measure the referencesample everyday because the piezo drift changes from day today. So our calculated value has been corrected for the piezo drift.

29 (a) J. J. van Gorp, J. A. J. M. Vekemans and E. W. Meijer, J. Am.Chem. Soc., 2002, 124, 14759–14769; (b) P. Johnkheijm, F. J. M.Hoeben, R. Kleppinger, J. van Herrikhuyzen, A. P. H. J. Schen-ning and E. W. Meijer, J. Am. Chem. Soc., 2003, 125,15941–15949.

30 (a) A. Chattopadhyay, S. Mukherjee and H. Raghuraman, J.Phys. Chem. B, 2002, 106, 13002–13009; (b) A. Chattopadhyayand S. Mukherjee, Langmuir, 1999, 15, 2142–2148; (c) S. S.Rawat, S. Mukherjee and A. Chattopadhyay, J. Phys. Chem. B,1997, 101, 1922–1929.

31 Similar to what was observed for conjugated polymers, see: T.-Q.Nguyen, I. B. Martini, I. J. Liu and B. J. Schwartz, J. Phys. Chem.B, 2000, 104, 237–255.

32 S. T. Yau, D. N. Petsev, B. R. Thomas and P. G. Vekilov, J. Mol.Biol., 2000, 303, 667–678.

33 K. Shirai, M. Matsuoka and K. Fukunishi, Dyes Pigm., 1999, 42,95–101.

34 (a) R. B. Martin, Chem. Rev., 1996, 96, 3043–3064; (b) W. Wang,L.-S. Li, G. Helms, H.-H. Zhou and A. D. Q. Li, J. Am. Chem.Soc., 2003, 125, 1120–1121; (c) W. Wang, J. J. Han, L.-Q. Wang,L.-S. Li, W. J. Shaw and A. D. Q. Li, Nano Lett., 2003, 3,455–458.

35 CRC Handbook of Chemistry and Physics, 76th edn, CRC Press,Boca Raton, FL, 1995–1996, pp. 6–245, and 6–249.

36 (a) E. Y. Sheu, K. S. Liang and L. Y. Chiang, J. Phys. (Paris),1989, 50, 1279–1295; (b) O. Braitbart, R. Sasson and A. Weinreb,Mol. Cryst. Liq. Cryst., 1988, 159, 233–242; (c) D. Markovitsi, D.A. Germain, P. Millie, P. Lecuyer, L. Gallos, P. Argyrakis, H.Bengs and H. Ringsdorf, J. Phys. Chem., 1995, 99, 1005–1017; (d)K. E. S. Phillips, T. J. Katz, S. Jockusch, A. J. Lovinger and N. J.Turro, J. Am. Chem. Soc., 2001, 123, 11899–11907; (e) C. Nuck-olls, T. J. Katz, G. Katz, P. T. Collings and L. Castellanos, J. Am.Chem. Soc., 1999, 121, 79–88; (f) C. Nuckolls and T. J. Katz, J.Am. Chem. Soc., 1998, 120, 9541–9544; (g) U. Rohr, P. Schilicht-ing, A. Bohm, M. Gross, K. Meerholz, C. Brauchle andK. Mullen, Angew. Chem., Int. Ed., 1998, 37, 1434–1437; (h)Ref. 29b,c.

37 P. Attard, Mol. Phys., 1996, 89, 691–709.38 (a) O. Galkin and P. G. Vekilov, J. Am. Chem. Soc., 2000, 122,

156–163; (b) J. Lin, J. Zhu and D. Zhou, Eur. Polym. J., 1999, 36,309–314; (c) D. N. Petsev, K. Chen, O. Gliko and P. G. Vekilov,Proc. Natl. Acad. Sci. U. S. A., 2003, 100, 792–796.

1532 | Phys. Chem. Chem. Phys., 2007, 9, 1515–1532 This journal is �c the Owner Societies 2007