Hydrogen bonding versus van der waals interactions: Competitive influence of noncovalent interactions on 2d self-assembly at the liquid-solid interface

Article abstract of DOI:10.1002/chem.201001653

The structures of the self-assembled monolayers of various 4-alkoxybenzoic acids physisorbed at the liquid-solid interface were established by employing scanning tunnelling microscopy (STM). This study has been essentially undertaken to explore the competitive influence of van der Waals and hydrogen-bonding interactions on the process of two-dimensional self-assembly. These acid derivatives form hydrogen-bonded dimers as expected; however, the dimers organise themselves in the form of relatively complex lamellae. The characteristic feature of these lamellae is the presence of regular discommensurations or kinks along the lamella propagation direction. The formation of kinked lamellae is discussed in light of the registry mechanism of the alkyl chains with the underlying graphite substrate. The location of the kinks along a lamella depends on the number (odd or even) of carbon atoms in the alkyl chain. This result indicates that concerted van der Waals interactions of the alkyl chain units introduce the odd/even chain-length effect on the surface-assembled supramolecular patterns. The odd/even effects are retained even upon complexation with a hydrogen-bond acceptor. However, as the solvent is changed from 1-phenyloctane to 1-octanoic acid, the kinked lamellae as well as the odd/even effects disappear. This solvent-induced convergence of supramolecular patterns is attained by means of co-crystallisation of octanoic acid molecules in the 2D crystal lattice, which is evident from high-resolution STM images. The solvent co-adsorption phenomenon is discussed in terms of competing van der Waals and hydrogen-bonding interactions. Discontinuous but ordered: Noncovalent interactions play an exceedingly decisive role in the process of self-assembly. To gain better insight into the role of noncovalent interactions, the competitive influence of hydrogen-bonding and van der Waals interactions was explored in the monolayers formed by simple 4-alkoxybenzoic acids at the liquid-solid interface (see image). Copyright

Full text of DOI:10.1002/chem.201001653

FULL PAPER
DOI: 10.1002/chem.201001653
Hydrogen Bonding Versus van der Waals Interactions: Competitive
Influence of Noncovalent Interactions on 2D Self-Assembly at the
Liquid–Solid Interface
Kunal S. Mali,[a] Kathleen Lava,[b] Koen Binnemans,[b] and Steven De Feyter*[a]
Abstract: The structures of the self-as-
sembled monolayers of various 4-alk-
ACHTUNGTRENNUNGoxybenzoic acids physisorbed at the
the lamella propagation direction. The
formation of kinked lamellae is dis-
cussed in light of the registry mecha-
nism of the alkyl chains with the under-
lying graphite substrate. The location
of the kinks along a lamella depends
on the number (odd or even) of carbon
atoms in the alkyl chain. This result in-
dicates that concerted van der Waals
interactions of the alkyl chain units in-
troduce the odd/even chain-length
effect on the surface-assembled supra-
molecular patterns. The odd/even ef-
fects are retained even upon complexa-
tion with a hydrogen-bond acceptor.
However, as the solvent is changed
from 1-phenyloctane to 1-octanoic
acid, the kinked lamellae as well as the
odd/even effects disappear. This sol-
vent-induced convergence of supra-
molecular patterns is attained by
means of co-crystallisation of octanoic
acid molecules in the 2D crystal lattice,
which is evident from high-resolution
STM images. The solvent co-adsorption
phenomenon is discussed in terms of
competing van der Waals and hydro-
gen-bonding interactions.
liquid–solid interface were established
by employing scanning tunnelling mi-
croscopy (STM). This study has been
essentially undertaken to explore the
competitive influence of van der Waals
and hydrogen-bonding interactions on
the process of two-dimensional self-as-
sembly. These acid derivatives form hy-
drogen-bonded dimers as expected;
however, the dimers organise them-
selves in the form of relatively complex
lamellae. The characteristic feature of
these lamellae is the presence of regu-
lar discommensurations or kinks along
Keywords: liquid–solid interfaces ·
noncovalent interactions · scanning
probe microscopy · self-assembly ·
solvent effects
Introduction
blocks.[2,3] Thus, understanding the organisation of molecules
on solid surfaces is fundamental for the realisation and ensu-
ing development of surface-supported molecular nanostruc-
tures and nanodevices.[2–20] Scanning tunnelling microscopy
(STM) is one of the preferred techniques for investigating
the ordering and properties of self-assembled layers—in
general, monolayers, both under ultrahigh vacuum condi-
tions[21]—as well as the liquid–solid[22] or air–solid interface.
STM offers a unique possibility of not only submolecular
visualisation but also manipulation of these adsorbed nano-
structures. Successful formation and subsequent visualisation
of a monolayer necessitates a delicate balance between vari-
ous attractive and repulsive forces. In general, relatively
weak intermolecular interactions such as van der Waals, hy-
drogen-bonding, dipole–dipole interactions and metal coor-
dination have been exploited to form well-defined two-di-
mensional (2D) monolayers on surfaces.[23]
The phenomenon of molecular self-assembly is ubiquitous
in nature and can be observed in many chemical, physical
and biological systems.[1] Self-assembly has been recognised
as one of the powerful means of fabricating functional nano-
materials by employing well-designed molecules as building
[a] Dr. K. S. Mali, Prof. S. De Feyter
Division of Molecular and Nanomaterials
Department of Chemistry and
INPAC-Institute of Nanoscale Physics and Chemistry
Katholieke Universiteit Leuven, Celestijnenlaan 200F
P.O. Box 2404, 3001 Leuven (Belgium)
Fax : (+32)16-327990
E-mail: Steven.DeFeyter@chem.kuleuven.be
[b] K. Lava, Prof. K. Binnemans
Laboratory of Coordination Chemistry
Department of Chemistry
Katholieke Universiteit Leuven, Celestijnenlaan 200F
P.O. Box 2404, 3001 Leuven (Belgium)
Supramolecular surface motifs based on hydrogen-bond-
ing interactions are one of the most frequently encountered
self-assembled structures due to the relatively strong, selec-
tive and directional nature of hydrogen bonds. Carboxyl
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201001653.
Chem. Eur. J. 2010, 16, 14447 – 14458
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
14447

functional groups are probably the most widely exploited
synthons for hydrogen-bonded motifs since they are endow-
ed with a very interesting and unique “self-complementary”
hydrogen-bonding ability. Thus, they exhibit dual nature on
account of the presence of the oxygen atom of the carbonyl
group, which acts as a hydrogen-bond acceptor, whereas the
hydroxyl group acts as a hydrogen-bond donor. Consequent-
ly, two carboxylic groups can form a cyclic dimer intercon-
nected by two equivalent hydrogen bonds. The role of hy-
drogen bonding in the process of 2D crystallisation is per-
fectly illustrated by the behaviour of benzene carboxylic
acids, which have proven to be excellent examples of “two-
dimensional tectons” and form 2D supramolecular networks
by means of hydrogen-bonding interactions. Phthalic acid,
isophthalic acid (ISA), terephthalic acid (TA) and trimesic
acid (TMA) all contain carboxylic acid functions that can
act as hydrogen-bond donors and acceptors simultaneously.
It has been confirmed by now that phthalic acid cannot
form extended hydrogen-bonded arrays, whereas TA and
ISA form dense monolayers in which the molecular self-as-
sembly is dominated by hydrogen-bonding interactions.[24]
The self-assembly of TMA,[25–29] on the other hand, has
served as a pioneering example of a nanoporous self-assem-
bled network that is formed both in ultrahigh vacuum
(UHV) conditions[25] as well as at the liquid–solid inter-
face.[26] The well-defined and robust nanoporous networks
formed by TMA have been extensively utilised to host a va-
riety of guests.[27,28] Furthermore, TMA has proven to be a
suitable building block for a variety of multicomponent sur-
face architectures.[27–29]
on the nature of solvent and the substrate used. Since it is
one of the elegant ways to control and manipulate the self-
assembly, quite a few studies are available that deal with the
influence of solvent on the pattern formation at the liquid–
solid interface.[36–48] However, despite the multitude of ex-
amples already investigated, it is still difficult to generalise
the role of solvent molecules in the process of self-assembly.
Co-adsorption of solvent molecules has been observed in
several instances, which is partly due to enhanced substrate–
solvent interactions.[37,38] Solvent co-adsorption is also gov-
erned by the size and shape of the solvent molecules[40] as
well as the mode of interaction (for example, hydrogen-
bonding or van der Waals interactions) of the solvent mole-
cules with the adsorbate.[35] The solvent also enables adsorp-
tion–desorption dynamics, thereby affecting the mobility of
the molecules at the liquid–solid interface.[43] The two sol-
vent parameters that possibly govern this dynamics are sol-
vation energy and solvent viscosity.[44,45] The solvent-induced
polymorphism, effect of co-adsorption as well as the solvent
effects on chirality and electronic structures have been sum-
marised recently.[48]
Although the molecules that possess two or three carbox-
ylic groups have been explored extensively, investigations
that deal with simple monocarboxylic acids such as 4-alk-
AHCTUNGTREGoNNNU xybenzoic acid (4-ABAs) derivatives are astoundingly
sparse. In fact, to the best of our knowledge, the self-assem-
bly of 4-ABAs has not been explored systematically except
for a very recent report by Matzger et al.[49] In this report,
they investigated the self-assembly of 4-ABAs as well as the
corresponding amides. The acid derivatives formed a com-
pact lamellar phase irrespective of the number of carbon
atoms in the alkyl chain. On the other hand, the amide de-
rivatives gave rise to highly symmetric nanoporous networks
that contained rhombic voids. The absence of a porous
phase in the acids was attributed to the geometrical differ-
ence in the hydrogen-bonding ability of amide and acid mol-
ecules.[49]
The principles of molecular self-assembly have been ex-
ploited for more than a decade to fabricate new materials
with unprecedented properties.[2–20] However, the ability to
systematically tailor the self-assembled patterns is only
slowly evolving. This is due to the fact that the outcome of a
2D supramolecular pattern is rarely encoded by the molecu-
lar building block and it strongly depends on, among other
factors, various noncovalent interactions. Thus, the under-
standing and subsequent control of noncovalent interactions
assumes special importance in the quest for novel supra-
molecular nanostructures. In this contribution, we have sys-
tematically investigated the 2D self-assembly of various 4-
ABAs that possess different alkyl substitution at the liquid–
HOPG (highly oriented pyrolytic graphite) interface, which
is virtually an unexplored system. The main goal of this in-
vestigation is to explore the competitive influence of hydro-
gen-bonding and van der Waals interactions on the process
of 2D self-assembly at the liquid–solid interface. To this end,
we first look into the details of the self-assembly process,
which is quite unique for this class of molecules. Complexa-
In contrast to hydrogen bonds, van der Waals interactions
between alkyl chains are neither strong enough nor direc-
tional in nature. The interdigitation as well as epitaxial ad-
sorption of alkyl chains on the surface of graphite, however,
which is mainly governed by van der Waals interactions, im-
parts some kind of unconventional directionality to the mo-
lecular self-assembly. Besides this, although the typical
energy of van der Waals interactions is less than 5 kJmolÀ1,
collectively these interactions can compete with the hydro-
gen-bonding interactions. (The calculated interaction energy
for interdigitated alkyl chains is 7.88ꢁ10À21 J per methylene
group if the alkyl chain is flanked by other alkyl chains on
both sides.[30]) Typically, the energy of a twofold O H···O=
À
hydrogen bond between two carboxylic acids is approxi-
mately À60 kJmolÀ1,[31] whereas the adsorption energy of
methane on the surface of graphite is À12.2 kJmolÀ1[32] and
it increases proportionately with the length of the alkane
with an increase of around 12.1 kJmolÀ1 for each methylene
unit.[33] The influence of alkyl substituents on the 2D self-as-
sembly of alkoxylated isophthalic acid molecules has also
been investigated in the recent past. It was demonstrated
that the 2D supramolecular ordering can be controlled by
varying the location and nature of alkyl substituents on the
aromatic core in combination with the intrinsic hydrogen-
bonding properties of the ISA units.[34,35]
Apart from the functionality of the building blocks, the
fate of a 2D supramolecular pattern also depends critically
14448
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14447 – 14458

Hydrogen Bonding versus van der Waals Interactions
FULL PAPER
tion studies were undertaken with an intent to modulate the
supramolecular assembly by introducing a hydrogen-bond-
accepting component. The interplay between van der Waals
and hydrogen-bonding interactions was also probed at the
level of solvent. Studies carried out in different solvents
highlight the importance of balance between noncovalent in-
teractions such as hydrogen-bonding and van der Waals in-
teractions on the process of 2D self-assembly.
Results and Discussion
Two-dimensional supramolecular patterns formed by 4-
ABAs: The peculiarity of the 2D self-assembly of 4-ABAs,
particularly in comparison with that of di- or tricarboxylic
acids such as TA, ISA and TMA, is that one molecule of
acid can be involved in only one intermolecular hydrogen-
bonding motif.[50] This implies that once these dimers are
formed, they can be considered to be individual building
blocks of self-assembly. Deposition of 4-ABAs from a solu-
tion of 1-phenyloctane on the surface of HOPG leads to
spontaneous formation of stable physisorbed monolayers.
Figure 1a displays a large-scale STM image of 4-icosyloxy-
benzoic acid (4-OC20BA) at the 1-phenyloctane–HOPG in-
terface. The monolayer emerges with a lamellar feature with
clear distinction between regions of bright and dark con-
trast. On account of the higher electron density of the aro-
matic core, the bright regions can be attributed to the
phenyl rings of the acid molecules, whereas the regions of
darker contrast correspond to the alkoxy chains. A submo-
lecularly resolved image of the same derivative displayed in
Figure 1b allows us to discern some of the minute features
of this lamellar structure. The bright rings in Figure 1b
always show up in pairs, which is a strong indication that
these molecules form hydrogen-bonded dimers as expected.
The width of the lamella is (5.7Æ0.1) nm and it is indicated
as DL. In contrast to what has been observed for isophthalic
acid derivatives at the HOPG–liquid interface,[34,35] these
dimers form tilted rows instead of straight rows. However,
the most characteristic feature of these lamellae is the pres-
ence of regular discommensurations or kinks along the la-
mella propagation direction. The kinks appear after every
three dimers are formed (that is, every (1.8Æ0.1) nm) and
the next three dimers are shifted with respect to the previ-
ous triplet of dimers. As a consequence of these regular dis-
commensurations along the lamella axis, the monolayer
structure appears to be relatively more complex compared
to the one formed by alkylated ISA molecules. The alkoxy
chains in Figure 1b are particularly well resolved. They are
fully extended, closely packed and are aligned along one of
the symmetry axes of the underlying graphite lattice. The
orientation of the alkyl chains with respect to the lamella
axis is (57.6Æ1.2)8 and the lamellae in turn make an angle
of (5.4Æ1.0)8 with one of the graphite symmetry axes. The
alkyl chains in the neighbouring lamellae are arranged in a
tail-to-tail fashion and there is no alkyl chain interdigitation
within the adjacent lamellae. The distance between the alkyl
Figure 1. a) Large-scale and b) high-resolution STM images of 4-OC20BA
physisorbed at the 1-phenyloctane/HOPG interface. The bright ringlike
features in (b) are attributed to the phenyl rings of the acid molecules,
whereas striped features with a darker contrast represent the alkyl chains
of the molecules. The width of the lamella is indicated as DL. The unit-
cell parameters are as follows: a=(1.79Æ0.03) nm, b=(5.73Æ0.04) nm
and a=(87.9Æ0.5)8 (plane group p2). The main symmetry axes of the
underlying graphite substrate are indicated in the lower left corner of the
STM image. Imaging conditions: It =0.25 nA, Vbias =À0.78 V. c) Tentative
model depicting the molecular arrangement.
chains of two adjacent molecules within the same lamella
and perpendicular to the alkyl-chain orientation is (0.49Æ
0.02) nm, which is slightly higher than the distance observed
between long-chain alkanes (ꢀ0.45 nm) adsorbed on graph-
ite.[51] The intermolecular distance within the same lamella
and along the lamella propagation direction is (0.60Æ
0.02) nm, which correlates well with the periodicity of the
kinks.
Having resolved the microscopic features of the monolay-
er, we now focus on establishing the rationale behind the
emergence of the kinked lamellae. Figure 2, displays a
rather rough schematic of the expected (Figure 2a) and ob-
served (Figure 2b) molecular arrangement. On the basis of
hydrogen-bonding interactions and substitution pattern, one
anticipates these molecules to form dimers that are expected
to organise in a continuous lamellar structure as shown in
Figure 2a. However, we observe periodic dislocations in the
crystal lattice along the lamella propagation direction as
shown in Figure 2b. Examples of similar discontinuous la-
mellar structures can be traced back to the early STM inves-
Chem. Eur. J. 2010, 16, 14447 – 14458
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14449

S. De Feyter et al.
strate in the present study is built up of layers with a honey-
comb arrangement of carbon atoms. The basal (001) plane
of graphite has a threefold symmetry and the carbon atoms
along the direction of any C3 axis display the same zigzag
extension as the zigzag skeleton of the carbon backbone of
an all-trans alkyl chain. The in-plane lattice constant of
2.46 ꢂ (the distance between two interval carbon atoms
along any C3 axis) is very close to every second neighbour
carbon atom, 2.58 ꢂ along the axis of an all-trans alkyl
chain. This fortuitous match allows the methylene groups of
an all-trans alkyl chain to rest over the voids of the hexa-
gons of the graphite lattice,[63] thereby providing an approxi-
mately commensurate packing of the alkyl chains as dis-
played in Figure 3a. The lattice match between alkyl chains
Figure 3. A molecular model depicting the packing of a) unsubstituted
and two possible arrangements of b) phenyl-substituted (4-alkoxybenzoic
acid in this case) and c) alkane derivatives on the surface of graphite. It
is evident that the presence of a bulky head group in case of the latter
prevents the close packing of the molecules in comparison with the corre-
sponding unsubstituted alkane.
and the graphite substrate has been confirmed widely.[51,61–63]
The lateral arrangement of alkyl chains on the graphite lat-
tice is also governed by the similarity in the lattice parame-
ters. Typically, it has been observed that the distance be-
tween unsubstituted alkanes within a lamella (along the la-
mella axis) is approximately 4.26 ꢂ, which is in reasonable
agreement with the distance between every other carbon
row of the graphite lattice (4.24 ꢂ). Thus the registry of the
alkyl chains with the underlying graphite lattice requires
that these chains should be placed close to (or in multiple
integer of) 4.24 ꢂ apart (Figure 3a). However, in the present
case, the bulky phenyl head group attached to the alkoxy
chain prevents the contentment of the alkyl-chain registry
condition imposed by the graphite lattice. Thus, there is a
competition between the restriction enforced by the sub-
strate and the molecular packing parameter. If the alkoxy-
benzoic acid molecules are to be placed side-by-side on the
surface of graphite, then they should be placed at least
6.0 ꢂ apart as displayed in Figure 3b. However, as already
mentioned, the registry mechanism demands a distance of
4.24 ꢂ between two alkyl chains. Thus, the phenyl cores of
the molecules are shifted with respect to each other such
that the tails of the alkoxy chains are around 4.5 ꢂ apart
(Figure 3c). According to the Groszek model,[63] the lowest
potential energy trough for the methylene units of an alkyl
Figure 2. Schematic showing a) the expected and b) observed self-assem-
bled structures of 4-alkoxybenzoic acids. Instead of forming continuous
rows as displayed in (a), the molecules form discontinuous rows as shown
in (b). In this type of arrangement, after every three dimers the mole-
cules shift with respect to the previous triplet of dimers, thus giving rise
to periodic dislocations or kinks along the lamella axis.
tigations on alkylated cyanobiphenyl molecules. The mono-
layers formed by these liquid-crystalline molecules on atom-
ically flat conductive surfaces such as HOPG[42–60] and/or
[54,59,60]
MoS2
were the centre of focus in many of the early
STM investigations. Periodic dislocations observed in the
crystal structures were explained on the basis of strong mol-
ecule–substrate interactions and the registry mechanism of
the alkyl chains with the underlying substrate.[55] Thus, on
the basis of these literature reports, we can explain the dis-
continuous lamellar features observed in the present case as
follows. A major flaw in the schematic displayed in Fig-
ure 2a is that it does not take into consideration the influ-
ence of the substrate lattice on the self-assembling pattern.
It has been well documented in the literature that the sub-
strate lattice has a strong bearing on the self-assembly of
molecules.[14,61–63] The graphite lattice that is used as a sub-
14450
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14447 – 14458

Hydrogen Bonding versus van der Waals Interactions
FULL PAPER
chain is the centre of the hexagon of the graphite lattice
with an interchain distance of 4.26 ꢂ. Considering the exper-
imentally observed distance of around 4.9 ꢂ as well as the
offset between adjacent phenyl rings, the alkyl chains of
only three molecules can be placed with an approximate
registry with the underlying graphite lattice in a given direc-
tion. The alkyl chain of the fourth molecule, however, does
not follow the lattice registry correctly, with the methylene
units located away from the centres of the graphite hexa-
gons (see the Supporting Information). Thus, the aforemen-
tioned condition results in the accumulation of internal
strain in the rows of molecules, which implies that it is ener-
getically unfavourable to pack the molecules in straight
rows along one given direction. Consequently, to follow the
substrate lattice correctly, the fourth molecule is shifted in a
direction that is different than the propagation direction of
the previous three molecules. This shift allows the alkyl
chain of the fourth molecule to register correctly with the
underlying graphite lattice and hence the kinks appear after
every three molecules. However, it must be noted that the
conjecture proposed above is rather simplistic and other fac-
tors such as the length of the molecule as well as the interac-
tions between the molecules from adjacent lamellae might
also be responsible partially for the observed molecular ar-
rangement. Besides this, the fact that 4-ABAs form a struc-
turally different monolayer on the AuACTHNUTRGNEUNG
(111) surface[64] indi-
rectly points towards the important role played by the
graphite substrate on the self-assembly in the present case.
Thus, the development of kinks or dislocations in the crystal
lattice can be viewed as an intrinsic mechanism to relax the
internal strain in the monolayer.
In addition to the registry mechanism proposed above,
one can also consider the stabilisation of the molecular as-
Figure 4. STM images of various 4-ABAs physisorbed at the 1-phenyloc-
tane/HOPG interface. The number of carbon atoms in the alkoxy chain
is indicated in each image. The inset in (e) and (f) shows the magnified
area for the sake of clarity. It is evident from these images that a de-
crease in the alkoxy-chain length has no significant influence on the mo-
lecular arrangement except for a gradual decrease in the lamella width.
Imaging conditions: It =0.31 to 0.13 nA and Vbias =1.4 to 0.65 V. (For
STM images as well as unit-cell parameters of other derivatives, please
see the Supporting Information).
À
sembly by the formation of =C H···O= hydrogen bonds be-
tween adjacent benzoic acid molecules. Such stabilisation
appears to be sterically more feasible for the displaced ar-
rangement depicted in Figure 2b than for the parallel ar-
rangement in Figure 2a. The possibility of slight tilting of
the phenyl rings to closely pack the molecules can also be
considered. Such tilting of the phenyl rings might allow the
alkoxy chains to follow the substrate lattice. However, the
high-resolution image displayed in Figure 1b rules out this
possibility because the bright features appear to be lying flat
on the graphite surface. Moreover, the diameter of the
bright spots ((0.69Æ0.02) nm) matches well with the report-
ed value for flat-lying isophthalic acid molecules on the sur-
face of HOPG.[65]
images of 4-ABAs at the 1-phenyloctane–HOPG interface.
It is clearly evident from the STM images that a decrease in
the alkoxy-chain length has no significant influence on the
packing of the acid molecules in the monolayer, apart from
the gradual decrease in the lamella width. The kinks are
present in the lamellar structure even for the shortest (4-
butoxy) derivative investigated in this study. Thus, we can
safely conclude that in case of 4-ABAs, a compromise be-
tween the registry mechanism imposed by the substrate and
phenyl-group packing is largely responsible for the observed
peculiar surface pattern.
The discussion presented in the previous paragraph illus-
trates that alkyl-chain registry and phenyl-group packing are
the two critical and competing factors that govern the mo-
lecular ordering and are responsible for the emergence of
kinked lamellae. So the obvious question is, “What happens
if the alkyl-chain length is reduced?” To clarify this point
further, a systematic study of 4-ABAs that possess shorter
alkyl chains was undertaken wherein STM images of deriva-
tives that possess alkoxy chains as short as butoxy were re-
corded. Figure 4 displays some of the representative STM
Odd/even effects: During the systematic STM investigation
of the 4-ABAs, it emerged that there is a slight difference in
the packing of the derivatives that possess an even number
of carbon atoms in the alkoxy chain compared to the ones
Chem. Eur. J. 2010, 16, 14447 – 14458
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14451

S. De Feyter et al.
that have an odd number of carbon atoms. This has been ob-
served for the derivatives that possess an alkyl-chain length
of eight or fewer carbon atoms. Thus, the molecular packing
in 4-heptyloxy (4-OC7BA) and 4-pentyloxy (4-OC5BA) de-
rivatives is different than the corresponding tetra-, hexa-
and octyloxy-substituted molecules. Figure 5 displays STM
this system is an indication that the substrate registry has a
strong influence on molecular ordering not only along the
lamella propagation direction but also perpendicular to it. A
similar phenomenon of emergence of odd/even effects in
fractured lamellae has been observed earlier in case of alky-
lated cyanobiphenyls on the surface of MoS2.[54,59,60] Lacaze
et al.[67] have built a 2D phenomenological model and inter-
preted this phenomenon in terms of commensurate–incom-
mensurate transition monitored by the molecular length.
Odd/even effects are quenched, however, as we move to
longer alkoxy chain derivatives as evident from the STM
image displayed in Figure 6. 4-Heptadecyloxybenzoic acid
Figure 6. STM image of 4-OC17BA physisorbed at the 1-phenyloctane–
HOPG interface. Note that the kinks are present after every three
dimers instead of two as observed for the derivatives that possess an odd
number of carbon atoms in the alkoxy chains. Thus the odd/even effect is
quenched upon increasing the alkoxy-chain length. Imaging conditions:
It =0.12 nA, Vbias =À1.2 V.
Figure 5. STM image of a) 4-OC5BA and b) 4-OC7BA physisorbed at the
1-phenyloctane–HOPG interface. Note that the kinks are present after
every two dimers instead of three as observed for the derivatives that
possess an even number of carbon atoms in the alkoxy chains. The unit-
cell parameters are a=(1.21Æ0.05) nm, b=(2.72Æ0.08) nm and a=
(88.3Æ1.5)8 for 4-OC5BA; and a=(1.18Æ0.04) nm, b=(3.42Æ0.07) nm
and a=(87.9Æ0.5)8 for 4-OC7BA (plane group is p2). Imaging condi-
tions: It =0.15 nA, Vbias =À0.73 V. c) Tentative model depicting the mo-
lecular arrangement in case of 4-OC7BA.
(4-OC17BA) self-assembles in an identical pattern as that of
its corresponding 4-octadecyloxy (4-OC18BA) derivative.
This is an indication that increased van der Waals interac-
tions (molecule–molecule as well as molecule–substrate)
with an increase in the alkoxy-chain length become more
important than the odd/even difference and thus quench the
odd/even effects in such a way that the molecular arrange-
ment is identical for the odd and even molecules.
There are some discrepancies, however, if we look at the
case of 4-nonyloxy (4-OC9BA) and 4-undecyloxy (4-
OC11BA) benzoic acids. It was found that 4-OC9BA forms
an entirely different pattern at the 1-phenyloctane–HOPG
interface. Figure 7a shows the STM image of 4-OC9BA that
is completely different from that observed for other deriva-
tives in the series. Considering the similarity in molecular di-
mensions, one cannot rule out the possibility of co-adsorp-
tion of 1-phenyloctane[41] in this case. On the other hand, 4-
OC11BA (Figure 7b) forms lamellae that are somewhat simi-
lar to the ones formed by the rest of the derivatives, the
only difference being the lack of periodicity in the position
of kinks along the lamella propagation direction. The loca-
tion of kinks in this case is irregular. Having already seen
that odd/even effects are quenched upon increasing the
alkyl-chain length in the case of 4-OC17BA, we can specu-
late that 4-OC11BA can be a borderline case wherein the in-
images of these derivatives at the 1-phenyloctane/HOPG in-
terface. The STM images of 4-OC5BA (Figure 5a) and 4-
OC7BA (Figure 5b) exhibit the dislocations in the lamella
after every two dimers (every (1.2Æ0.1) nm) instead of the
normally observed difference of every three dimers (every
(1.8Æ0.1) nm). Thus, this dissimilar packing in case of the
odd and even number of carbon-containing molecules gives
rise to what has been termed as “microscopic odd/even ef-
fects” in the literature.[66] The odd/even effect is a widely ob-
served phenomenon in chemistry, physics, biology as well as
material science. It encompasses the differences in the struc-
ture and/or properties depending on the odd or even
number of structural units in a molecule. Such odd/even ef-
fects observed in self-assembled organic monolayers have
been found to influence many chemical, physical and inter-
facial properties such as chemical reactivity, electronic and
electrochemical properties as well as friction behaviour.[66]
In the present case, the number of carbon atoms serves as
the odd/even unit. The appearance of odd/even effects in
14452
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14447 – 14458

Hydrogen Bonding versus van der Waals Interactions
FULL PAPER
Figure 7. STM image of a) 4-OC9BA and b) 4-OC11BA physisorbed at
the 1-phenyloctane–HOPG interface. Compound 4-OC9BA forms an en-
tirely different pattern than the remaining homologues. A possibility of
solvent co-adsorption cannot be ruled out in this case. Unit-cell parame-
ters are as follows: a=(4.13Æ0.06) nm, b=(5.12Æ0.08) nm and a=
(80.0Æ1.1)8 (plane group is p2). The monolayer formed by 4-OC11BA,
on the other hand, shows a bit of randomness in the appearance of kinks
along the lamella axis. Imaging conditions: a) It =0.2 nA, Vbias =À1.2 V;
b) It =0.13 nA, Vbias =À0.9 V.
Figure 8. STM images of a) 4-OC20BA and b) 4-OC18BA complexed with
4,4’-bipyridine. The grey arrow in (a) indicates the location of bipyridine
moieties, whereas the black arrows indicate the positions of phenyl rings
of the benzoic acid molecules. The inset in (b) shows the corresponding
high-resolution STM image. Note that the kinks are absent in this bicom-
ponent system and the monolayer structure is continuous. Imaging condi-
tions: a) It =0.075 nA, Vbias =À0.85 V; b) It =0.095 nA, Vbias =À0.95 V.
terchain van der Waals interactions start becoming more sig-
nificant than the odd/even difference. It must be noted at
this juncture that 4-decyloxy as well as 4-dodecylACTHNUTRGENUGoN xybenzoic
acids show typical periodicity in the location of kinks along
the lamella axis.
Complexation experiments carried out on the molecules
that possess an odd number of carbon atoms in the alkoxy
chain revealed an entirely different packing of the bicompo-
nent system. Figure 9 gives representative STM images of
4,4’-bipyridine complexed with 4-OC17BA. The submolecu-
larly resolved image displayed in Figure 9b clearly shows the
microscopic features of the bicomponent monolayer. It is
evident that the alkyl chains are nicely resolved, closely
Complexation studies: To modulate the supramolecular pat-
tern formed by 4-ABAs by manipulating the hydrogen-
bonding interactions, complexation studies were undertaken.
Furthermore, these experiments will also to help evaluate if
the complexation reaction can be used to control the supra-
molecular patterns in such a way that all the benzoic acid
derivatives form identical supramolecular patterns, thus
leading to a structural convergence. To this effect, 4,4’-bipyr-
idine, which acts as a hydrogen-bond acceptor, was used as
a complexing agent. A saturated solution of 4,4’-bipyridine
in 1-phenyloctane was added to a preformed monolayer of
4-ABA on the surface of HOPG. After allowing the system
to equilibrate, STM images of the bicomponent monolayer
were recorded. Figure 8 displays the STM images of 4-
OC20BA and 4-OC18BA complexed with 4,4’-bipyridine. The
central bright row in the lamellar structure corresponds to
the bipyridine moieties, whereas the adjacent rows are the
aromatic cores of the hydrogen-bonded benzoic acid mole-
cules. The alkyl chains are not well resolved; nevertheless,
there is sufficient indication that they are not interdigitated.
The width of the lamella in case of 4-OC20BA is (5.8Æ
0.2) nm, whereas it is (5.4Æ0.1) nm for the 4-OC18BA deriv-
ative. However, the most important feature of the STM
images displayed in Figure 8 is the presence of continuous
lamellar structure and complete disappearance of periodic
dislocations that were prominent in the cases of the corre-
sponding monocomponent systems. Thus, the aforemen-
tioned experiments illustrate that the incorporation of a hy-
drogen-bond-accepting agent in the 2D crystal lattice modu-
lates the monolayer structure.
Figure 9. a) Complexation between 4-OC17BA and 4,4’-bipyridine at the
1-phenyloctane–HOPG interface. b) High-resolution STM image of the
same system. The grey arrow in (b) indicates the location of the bipyri-
dine moiety, whereas the white arrows show the positions of phenyl rings
of the benzoic acid molecules. Note that the bipyridine moieties are not
arranged in straight rows as is the case in Figure 8. The white bars are
drawn as a visual aid to identify the offset in the molecular packing.
Imaging conditions: a) It =0.13 nA, Vbias =À1.2 V; b) It =0.13 nA, Vbias
=
À1.13 V.
Chem. Eur. J. 2010, 16, 14447 – 14458
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14453

S. De Feyter et al.
packed and interdigitated. The alkyl-chain interdigitation is
in complete contrast to the bicomponent system of the cor-
responding 4-OC18BA and 4-OC20BA derivatives. Moreover,
the bipyridine moieties are not arranged in straight rows. In-
stead, after every two bipyridine molecules, there is a shift
in the propagation direction. Thus, in a sense, the kinks are
still present in the lamellar structure. The white bars in Fig-
ure 9b are drawn for identifying the offset in the molecular
packing. This peculiar molecular arrangement is surprising
in view of the fact that 4-OC17BA does not show odd/even
effect as already discussed in previous section. Thus, the
odd/even difference that is insignificant in the monocompo-
nent system appears to become prominent in the presence
of the second component, which is the bipyridine unit in the
present case. The outcome of complexation studies reveals
that, although it is an effective way of modulating the self-
assembly, it is not efficient enough to converge the struc-
tures of the various alkoxybenzoic acid derivatives on the
surface, thus highlighting the importance of van der Waals
interactions. These experiments, when extended to other
molecules that possess an odd number of carbon atoms in
the alkoxy chain (Figure S3 in the Supporting Information),
reaffirm this point.
Similar experiments have been carried out in the recent
past with an intent to tune 2D supramolecular patterns by
adding an appropriate complexing agent to the self-assem-
bled monolayer.[68–73] Qian et al.[68,69] have successfully em-
ployed 4,4’-bipyridine as a marker to study self-assembly[68]
and chirality[69] in the monolayers formed by aliphatic fatty
acids. Alkylated derivatives of 2,2’-bipyridine have been
used as complexation scaffolds for metal-ion complexes. The
use of such metal-ion complexes has been explored as a
route to novel 2D templates.[8] Recently, Kikkawa et al.[70]
have investigated the self-assembly and odd/even chain
length effect in the case of various bipyridine derivatives.
They could successfully induce a structural convergence in
different 2D structures by using metal-ion complexation.[70]
The outcome of the present investigation on complexation
appears to be somewhat different than the previous reports
in view of the fact that the odd/even effects remain promi-
nent even in the bicomponent crystal lattice.
terminal carboxyl functional group that can be involved in
hydrogen-bonding interactions with the solute molecules
and hence may have an impact on the monolayer structure.
In fact, hydrogen-bonding solvents have been known to play
a significant role in the adsorbate monolayer structure.[74–77]
Thus, in comparison with 1-phenyloctane, which is relatively
a neutral partner in the self-assembly, 1-octanoic acid can
indeed influence the molecular packing. Figure 10 gives a
Figure 10. STM image of 4-OC20BA physisorbed at the 1-octanoic acid–
HOPG interface. The width of the lamella is indicated as DL. Note that
the structural pattern is entirely different than the one observed in 1-phe-
nyloctane. The different contrast on either side of the bright spots is due
to scanning artifacts. Imaging conditions: It =0.13 nA, Vbias =À1.2 V.
representative STM image of 4-OC20BA physisorbed at the
1-octanoic acid–HOPG interface. The bright spots that cor-
respond to the phenyl rings are arranged in a zigzag fashion.
The characteristic fractured molecular arrangement ob-
served in 1-phenyloctane is absent. The width of the lamella
is (4.11Æ0.06) nm, which is smaller in comparison to that
observed in 1-phenyloctane ((5.8Æ0.1) nm) and is represent-
ed as DL in the image. The region of the lamella that lies in
between the bright spots is not well resolved, though. Never-
theless, this STM image primarily illustrates that the molec-
ular packing is entirely different in 1-octanoic acid than that
observed in 1-phenyloctane. One of the surprising aspects of
this monolayer structure is the presence of laterally placed
bright features instead of head-on. In other words, based
primarily on the molecular structure and substitution pat-
tern, one expects these molecules to form hydrogen-bonded
dimers and thus the bright features should appear opposite
each other. Thus, the appearance and rationalisation of such
features is an intriguing aspect and the role of 1-octanoic
acid may hold a key in the development of such a peculiar
structure.
Solvent-induced structural convergence: For surface assem-
blies at the liquid–solid interface, solvent molecules play a
vital role in the outcome of a given supramolecular pat-
tern.[36–48] Apart from its chemical identity, typical solvent
properties such as viscosity and polarity have an impact on
the process of self-assembly. Various factors that contribute
to the solvent dependence of self-assembly are solvent co-
adsorption, solvophilic/solvophobic effects, hydrogen-bond-
ing and van der Waals interactions.[48] The delicate balance
between all these factors makes it difficult to generalise the
role of solvent in the process of 2D self-assembly. Thus, to
comprehend the role of solvent in the formation of 2D
supramolecular patterns in the case of 4-ABAs, we extended
these investigations in 1-octanoic acid. 1-Octanoic acid is a
structurally different solvent than 1-phenyloctane. It has a
With an intent to figure out the exact mode of interaction
between solute and the solvent molecules, high-resolution
STM images were recorded for this system. Figure 11 gives
one such STM image that resolves the region between
zigzag bright spots. It is evident that opposite every bright
spot there lies one striped feature that can be attributed to
the adsorbed molecule of 1-octanoic acid. The appearance
14454
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14447 – 14458

Hydrogen Bonding versus van der Waals Interactions
FULL PAPER
hols in the monolayer of TMA. The only difference in their
case was that alcohols were added as solutes and heptanoic
acid served as solvent.[29]
Having established the role of a hydrogen-bonding sol-
vent in the process of self-assembly, it is imperative to verify
if a similar effect can be induced in case of other 4-ABA
molecules. More importantly, it would be intriguing to check
if the solvent can be used as an internal agent to converge
the structures. To this effect, STM images of a few represen-
tative 4-ABA derivatives were recorded in 1-octanoic acid.
Figure 12a–d gives STM images of four different 4-ABA de-
rivatives adsorbed at the 1-octanoic acid–HOPG interface.
It can be easily noticed from these images that they all rep-
resent identical supramolecular pattern, thus giving ample
evidence of structural convergence. However, it still remains
Figure 11. a) High-resolution STM image of 4-OC20BA physisorbed at
the 1-octanoic acid–HOPG interface. Submolecularly resolved features
indicate that 1-octanoic acid molecules are co-crystallised with benzoic
acid molecules. b) Tentative model depicting the molecular arrangement.
Arrows in (a) as well as (b) indicate the location of terminal methyl
groups of constituent molecules. The region with darker contrast opposite
each bright spot indicates the location of the hydrogen bonding between
two molecules. Unit-cell parameters are as follows: a=(1.24Æ0.03) nm,
b=(4.11Æ0.03) nm and a=(88.5Æ1.0)8 (plane group p2). Imaging condi-
tions: It =0.13 nA, Vbias =À1.2 V.
of a dark region opposite each bright spot can therefore be
explained by the fact that the carboxyl functional groups do
not appear in STM images.[61] The length of the striped fea-
ture opposite every bright spot ((1.1Æ0.1) nm) is in good
agreement with the length of 1-octanoic acid molecules
(ꢀ1.02 nm). Moreover, one can also decipher two rows of
additional features that lie in between two zigzag rows of
bright spots and within the striped features. These features
correspond to the terminal methyl groups of 1-octanoic acid
molecules, which are indicated by arrows in Figure 11.
From the discussion presented in the previous paragraph,
it is clear that 1-octanoic acid molecules have a propensity
to co-adsorb in view of their hydrogen-bonding ability. As a
consequence of such hydrogen-bonding-induced co-adsorp-
tion phenomenon, a very small amount of solvent molecules
actually become adsorbates and the rest of them remain in
the continuous phase. Co-adsorption of solvent molecules
can thus be considered as an alternative method to fabricate
hybrid supramolecular patterns.[41] Similar co-adsorption be-
haviour has been reported by Bernasek et al.[76] when they
investigated the self-assembly of 5-(octadecyloxy)isophthalic
acid at the 1-octanoic acid–HOPG interface. They observed
the formation of a molecular mesh as a result of insertion of
octanoic acid molecules in the monolayer. We believe that,
along with the co-adsorption phenomenon, the lateral hy-
drogen-bonding interactions between adjacent 5-(octadecy-
loxy)isophthalic acid molecules are equally significant in
their experiments. In the present case, however, the only lat-
eral interactions between adjacent molecules are the van
der Waals interactions between the alkoxy chains. More-
over, we do not observe any pore formation as the mole-
cules pack compactly without leaving any empty voids in
the lamella. The fact that this two-component monolayer
could be imaged with submolecular resolution demonstrates
the excellent stability of this system. Perepichka et al.[29]
have also observed co-adsorption of various aliphatic alco-
Figure 12. STM images of 4-ABAs a)–d) in 1-octanoic acid and e) in 1-
hexanoic acid. The length of the alkoxy chain is indicated in each figure.
It is evident that in the presence of a hydrogen-bonding solvent, the
monoACTHNUTRGNElNUG ayers are converged into identical zigzag structures. The inset in
(e) shows the magnified area for the sake of clarity. f) A tentative molec-
ular model displaying molecular packing of 4-OC7BA in 1-hexanoic acid.
Chem. Eur. J. 2010, 16, 14447 – 14458
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14455

S. De Feyter et al.
to be verified as to what happens to the lower members of
4-ABAs in 1-octanoic acid. The experiments aimed towards
STM imaging of lower homologues in 1-octanoic acid were
not fruitful. Preferential adsorption of the solvent molecules
was observed instead of the solute molecules of interest.
Thus, to prove that a hydrogen-bonding solvent can indeed
bring about structural convergence, we carried out measure-
ments in lower aliphatic acids. Figure 12e gives a representa-
tive case of 4-OC7BA physisosorbed at the 1-hexanoic acid–
HOPG interface. Typical zigzag bright features can be dis-
cerned in this STM image, which clearly validate the struc-
tural convergence brought about by the solvent molecules.
assembly process either as a neutral (dispersant) or as an
active (counterpart) partner. 1-Phenyloctane was found to
behave as a relatively neutral partner, mainly contributing
as a dispersion medium for the solute molecules. On the
other hand, 1-octanoic acid and 1,2,4-trichlorobenzene were
found to act as counterparts in the 2D lattice through co-ad-
sorption with solute molecules.[78] The behaviour of 1-phe-
nyloctane and 1-octanoic acid observed in the present study
is somewhat in line with the results of Wan et al.[77] Concen-
tration dependence on the solvent co-adsorption phenomen-
on was also investigated. It was observed that solvent co-ad-
sorption is favoured at lower solute concentration.[78]
While considering the co-adsorption phenomenon of alky-
lated molecules, one cannot discount the role of interchain
van der Waals interactions. To evaluate the influence of
these interactions on the co-adsorption phenomenon, a
second set of STM experiments was performed on 4-
OC20BA and 4-OC7BA adsorbed at the butyric acid–HOPG
interface. The justification for using this particular solvent is
as follows. Butyric acid can also form complementary hydro-
gen bonds with benzoic acid molecules in an identical fash-
ion to that of 1-octanoic acid. However, the shorter alkyl
chain of the solvent will lead to decreased van der Waals in-
teractions (both solute–solvent as well as solvent–substrate),
thus allowing us to compare the competitive influence of
complementary hydrogen-bonding and van der Waals inter-
actions on the process of co-adsorption. However, it was ob-
served that, despite having the capability to form comple-
mentary hydrogen bonds with the solute molecules, the co-
adsorption of butyric acid molecules is not favoured in case
of both the solutes, plausibly in view of insufficient van der
Waals interactions. The molecular packing in butyric acid is
exactly the same as the one observed in 1-phenyloctane
(Figure S6 in the Supporting Information). Thus, on the
basis of the control experiments described above, we can
safely conclude that the co-adsorption phenomenon is gov-
erned by both hydrogen-bonding as well as van der Waals
interactions and solvent co-adsorption is observed only
when there is a perfect balance between these two interac-
tions.[79]
Co-adsorption-induced structural convergence: Universal
phenomenon or a specific case for 1-octanoic acid? The dis-
cussion presented in the previous paragraph suggests that
presence of a hydrogen-bonding solvent molecule can mod-
ulate the monolayer structure through co-adsorption, there-
by inducing structural convergence. It needs to be estab-
lished whether this aspect is universal or specific for the
case of 1-octanoic acid. Moreover, such a solvent co-adsorp-
tion phenomenon will allow us to test the competitive influ-
ence of hydrogen-bonding and van der Waals interactions at
the level of the solvent. As already mentioned, the under-
standing of solvent effects in the process of 2D self-assembly
is still in its infancy. Thus, to establish the parameters that
govern the solvent co-adsorption phenomenon in the case of
4-ABAs, the following control experiments were carried
out. First, the self-assembly of 4-OC20BA was attempted in
1-octanol, which is also capable of hydrogen bonding with
the solute molecules. This molecule possesses a hydroxyl
functional group instead of carboxyl and has identical mo-
lecular dimensions as that of 1-octanoic acid (effectively, the
only difference being the absence of a carbonyl group).
However, the molecular packing observed in this solvent
(Figure S5 in the Supporting Information) is exactly the
same as the one observed in case of 1-phenyloctane, thereby
indicating that 1-octanol does not undergo co-adsorption
with solute molecules. Thus, it appears that a solvent mole-
cule that can form complementary hydrogen bonds is more
favoured for co-adsorption than the one that can form only
a single hydrogen bond. It can be argued that the effect of
concentration might come into play in this case. At relative-
ly high concentration, the benzoic acid molecules will ꢃfindꢄ
each other more easily than molecules of solvent and hence
solvent co-adsorption will not be favoured. However, it
must be noted here that the concentration of benzoic acid
was maintained at almost identical levels in the experiments
that involved 1-octanoic acid as well as 1-octanol. Moreover,
no pattern change was observed when similar experiments
were carried out at a concentration that was two orders of
magnitude lower (ꢀ10À6 m). It is worthwhile to consider re-
sults by Wan et al.[77,78] in terms of concentration effects on
the solvent co-adsorption phenomenon. They recently re-
ported the solvent dependence of the 2D self-assembly of a
5-(benzyloxy)isophthalic acid derivative.[77] According to
their investigations, a solvent can contribute in the 2D self-
Conclusion
It is difficult to predict the outcome of a supramolecular
pattern on the basis of the chemical structure of the building
blocks due to the crucial role played by noncovalent interac-
tions. A better understanding of such interactions will lead
to the improved ability to envisage and control the fate of
surface-supported supramolecular patterns, which will in
turn be valuable to program newer functional surfaces. We
have systematically explored the competitive influence of
two such noncovalent interactions, namely, hydrogen-bond-
ing and van der Waals interactions, on the self-assembly of
4-alkoxybenzoic acids. The present study demonstrates that
intermolecular hydrogen-bonding and interchain as well as
molecule–substrate interactions (dominated by van der
14456
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14447 – 14458

Hydrogen Bonding versus van der Waals Interactions
FULL PAPER
by allowing a more accurate unit-cell determination. The unit-cell param-
eters were determined by examining at least 10 images and only the aver-
age values are reported. After the determination of the unit cell from the
acquired STM images, a molecular model of the observed monolayer was
constructed using the HyperChem program. First, a molecular model for
a single molecule was built, and then this model was duplicated. The
model of the entire monolayer was constructed by placing the molecules
in accordance with the intermolecular distances and angles obtained
from the analysis of the calibrated STM images. The imaging parameters
are indicated in the figure captions: tunnelling current (It), and sample
bias (Vt).
Waals interactions) direct the molecules to form a peculiar
self-assembled pattern with periodic dislocations in the crys-
tal lattice. A closer look at the molecular structure and sub-
strate lattice reveals that a competition between phenyl-
group packing and alkyl-chain registry with the substrate is
responsible for the development of periodic dislocations in
the lamellae. These regular dislocations serve as an intrinsic
mechanism to relieve the strain developed in the monolayer
as a consequence of incommensurability with the underlying
substrate. A systematic study on various 4-ABAs revealed
that the periodicity of the kinks along the lamella axis de-
pends on the number of carbon atoms in the alkoxy chain,
thereby giving rise to microscopic odd/even effects promi-
nent in the case of shorter alkoxy derivatives. The odd/even
effect tends to vanish with increasing alkoxy-chain length.
The emergence of odd/even effects further confirms the im-
portance of van der Waals interactions (both interchain and
molecule–substrate) in the 2D self-assembly. Insertion of a
complexing agent that acts as a hydrogen-bond acceptor
alters the supramolecular patterns. However, the complexa-
tion reaction fails to induce a structural convergence. The
odd/even effects as well as kinked lamellae disappear when
the molecules self-assemble at 1-octanoic acid–HOPG inter-
face instead of 1-phenyloctane–HOPG interface. Thus, the
solvent plays a significant role by being able to interact spe-
cifically with the solute molecules through complementary
hydrogen-bonding interactions. High-resolution STM images
illustrate that the solvent-induced structural convergence is
effected by the coadsorption of solvent molecules in the 2D
crystal lattice. The solvent-induced polymorphism probed by
using solvents possessing different hydrogen bonding ability
and alkyl chain lengths reveals that a perfect balance of
these two factors is necessary so that the solvent molecules
can undergo coadsorption.
For the synthesis of the 4-alkoxybenzoic acids that were employed in this
study, please see the Supporting Information.
Acknowledgements
This work is supported by the Fund of Scientific Research-Flanders
(FWO), K. U. Leuven (GOA 2006/2), the Belgian Federal Science Policy
Office (IAP-6/2) and the Institute of Nanoscale Physics and Chemistry
(INPAC). K.S.M. thanks INPAC for a visiting postdoctoral fellowship
and K.L. thanks the IWT-Flanders for financial support.
[1] Comprehensive Supramolecular Chemistry, Vol. 6, (Eds.: J. L.
Atwood, J. E. D. Davies, D. D. MacNicol, F. Vçgtle), Pergamon
Press, New York, 1996.
[2] J.-M. Lehn, Proc. Natl. Acad. Sci. USA 2002, 99, 4763–4768.
[3] G. M. Whitesides, M. Boncheva, Proc. Natl. Acad. Sci. USA 2002,
99, 4769–4774.
[4] A. Semenov, J. P. Spatz, M. Mçller, J.-M. Lehn, B. Sell, D. Schubert,
C. H. Weidl, U. S. Schubert, Angew. Chem. 1999, 111, 2701–2705;
Angew. Chem. Int. Ed. 1999, 38, 2547–2550.
[5] J. V. Barth, J. Weckesser, C. Cai, P. Gꢅnter, L. Bꢅrgi, O. Jeandupeux,
K. Kern, Angew. Chem. 2000, 112, 1285–1288; Angew. Chem. Int.
Ed. 2000, 39, 1230–1234.
[6] K. W. Hipps, L. Scudiero, D. E. Barlow, M. P. Cooke, Jr., J. Am.
Chem. Soc. 2002, 124, 2126–2127.
[7] G. B. Pan, J.-M. Liu, H.–M. Zhang, L.-J. Wan, Q.-Y. Zheng, C.-L
Bai, Angew. Chem. 2003, 115, 2853–2857; Angew. Chem. Int. Ed.
2003, 42, 2747.
[8] S. De Feyter, M. M. S. Abdel-Mottaleb, N. Schuurmans, J. V. Ver-
kuijl, J. H. van Esch, B. L. Feringa, F. C. De Schryver, Chem. Eur. J.
2004, 10, 1124–1132.
Experimental Section
Materials and methods: All experiments were performed at room tem-
perature (20–228C) using a PicoSPM (Agilent) operating in constant-cur-
rent mode with the tip immersed in the supernatant liquid. STM tips
were prepared by mechanical cutting from Pt/Ir wire (80%:20%, diame-
ter 0.2 mm). Prior to imaging, the 4-alkoxybenzoic acid molecules were
dissolved (concentration=10À3 m to 10À4 m) in requisite solvents and a
drop of this solution was applied onto a freshly cleaved surface of highly
oriented pyrolytic graphite (HOPG, grade ZYB, Advanced Ceramics
Inc., Cleveland, USA). The solvents used were 1-phenyloctane (Aldrich,
98%), 1-octanoic acid (Sigma, 99%), 1-hexanoic acid (Acros Organics
98%), butyric acid (Acros Organics 99%), 1-octanol (Sigma–Aldrich
99%) and tetradecane (99% Acros Organics). 4,4’-Bipyridine was ob-
tained from Aldrich (98%) and was used without further purification.
All complexation reactions were performed in situ. Thus, a drop of satu-
rated solution of 4,4’-bipyridine in 1-phenyloctane was applied to a pre-
formed monolayer of benzoic acid on the surface of HOPG. The experi-
ments were repeated in several sessions using different tips to check for
reproducibility and to avoid experimental artefacts, if any. For analysis
purposes, recording of a monolayer image was followed by imaging the
graphite substrate underneath it under the same experimental conditions,
except for lowering the bias. The images were corrected for drift by using
Scanning Probe Image Processor (SPIP) software (Image Metrology
ApS), using the recorded graphite images for calibration purposes, there-
[9] Y. Liu, L. Mu, B. Liu, J. Kong, Chem. Eur. J. 2005, 11, 2622–2631.
[10] D.-G. Choi, S. Kim, E. Lee, S.-M. Yang, J. Am. Chem. Soc. 2005,
127, 1636–1637.
[11] M. Lackinger, S. Griessl, W. Heckl, M. Hietschold, G. W. Flynn,
Langmuir 2005, 21, 4984–4988.
[12] J. V. Barth, G. Costantini, K. Kern, Nature 2005, 437, 671–679.
[13] M. Ruben, Angew. Chem. 2005, 117, 1620–1623; Angew. Chem. Int.
Ed. 2005, 44, 1594–1596.
[14] J. A. A. W. Elemans, S. De Feyter, Soft Matter 2009, 5, 721–735 and
references therein.
[15] T. Kudernac, S. Lei, J. A. A. W. Elemans, S. De Feyter, Chem. Soc.
Rev. 2009, 38, 402–421 and references therein.
[16] L. Piot, C. Marie, X. Feng, K. Mꢅllen, D. Fichou, Adv. Mater. 2008,
20, 3854–3858.
[17] J. M. Mativetsky, M. Kastler, R. C. Savage, D. Gentilini, M. Palma,
W. Pisula, K. Mꢅllen, P. Samorꢆ, Adv. Funct. Mater. 2009, 19, 2486–
2494.
[18] A. Cnossen, D. Pijper, T. Kudernac, M. M. Pollard, N. Katsonis,
B. L. Feringa, Chem. Eur. J. 2009, 15, 2768–2772.
[19] A. G. Phillips, L. M. A. Perdigao, P. H. Beton, N. R. Champness,
Chem. Commun. 2010, 46, 2775–2777.
[20] K. Mꢅllen, J. P. Rabe, Acc. Chem. Res. 2008, 41, 511–520.
[21] S. Chiang, Chem. Rev. 1997, 97, 1083–1096.
Chem. Eur. J. 2010, 16, 14447 – 14458
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14457

S. De Feyter et al.
[22] S De Feyter, F. C. De Schryver, J. Phys. Chem. B 2005, 109, 4290–
4302.
[23] A. Ulman, Chem. Rev. 1996, 96, 1533–1554 and references therein.
[24] M. Lackinger, S. Griessl, T. Markert, F. Jamitzky, W. M. Heckl, J.
Phys. Chem. B 2004, 108, 13652–13655.
[25] A. Dmitriev, N. Lin, J. Weckesser, J. V. Barth, K. Kern, J. Phys.
Chem. B 2002, 106, 6907–6912.
[26] S. Griessl, M. Lackinger, M. Edelwirth, M. Hietschold, W. M. Heckl,
Single Mol. 2002, 3, 25–31.
[50] They also differ by their pKa. Benzoic acid: pKa =4.19; isophthalic
acid: pKa =3.54, 4.60; terephthalic acid: pKa =3.51, 4.82; CRC
Handbook of Chemistry and Physics, 55th edition (Ed.: R. C.
Weast), CRC Press, 1974–1975.
[51] B. Ilan, G. M. Florio, M. S. Hybertsen, B. J. Berne, G. W. Flynn,
Nano Lett. 2008, 8, 3160–3165.
[52] J. Foster, J. E. Frommer, Nature 1988, 333, 542–545.
[53] D. P. E. Smith, J. K. H. Hçrber, C. Gerber, G. Binnig, Science 1989,
245, 43–45.
[54] M. Hara, Y. Iwakabe, K. Tochigi, H. Sasabe, A. F. Garito, A.
Yamada, Nature 1990, 344, 228–230.
[55] D. P. E. Smith, J. K. H. Hçrber, G. Binnig, H. Nejoh, Nature 1990,
344, 641–644.
[56] T. J. McMaster, H. Carr, M. J. Miles, P. Cairns, V. J. Morris, J. Vac.
Sci. Technol. A 1990, 8, 672–674.
[57] H. Nejoh, D. P. E. Smith, M. Aono, Nanotechnology 1991, 2, 214–
220.
[58] T. Kadotani, S. Taki, H. Okabe, S. Kai, Jpn. J. Appl. Phys. 1996, 35,
L1345–L1347.
[59] S. Taki, T. Kadatoni, S. Kai, J. Phys. Soc. Jpn. 1999, 68, 1286–1291.
[60] S. Taki, H. Okabe, S. Kai, Jpn. J. Appl. Phys., Part 1 2003, 42, 7053–
7056.
[61] D. M. Cyr, B. Venkataraman, G. W. Flynn, Chem. Mater. 1996, 8,
1600–1615.
[62] C. Arrigoni, G. Schull, D. Blꢈger, L. Douillard, C. Fiorini-Debuis-
schert, F. Mathevet, D. Kreher, A.-J. Attias, F. Charra, J. Phys.
Chem. Lett. 2010, 1, 190–194.
[63] A. J. Groszek, Proc. R. Soc. London Ser. A 1970, 314, 473–498.
[64] N. Katsonis (private communication).
[65] S. Lei, M. Surin, K. Tahara, J. Adisoejoso, R. Lazzaroni, Y. Tobe, S.
De Feyter, Nano Lett. 2008, 8, 2541.
[66] F. Tao, S. L. Bernasek, Chem. Rev. 2007, 107, 1408–1453 and refer-
ences therein.
[67] E. Lacaze, P. Barois, R. Lacaze, J. Phys. I France 1997, 7, 1645–
1664.
[68] P. Qian, H. Nanjo, T. Yokoyama, T. M. Suzuki, Chem. Commun.
1999, 1197–1198.
[69] P. Qian, H. Nanjo, T. Yokoyama, T. M. Suzuki, K. Akasakab, H.
Orhui, Chem. Commun. 2000, 2021–2022.
[70] Y. Kikkawa, E. Koyama, S. Tsuzuki, K. Fujiwara, K. Miyake, H. To-
kuhisa M. Kanesato, Chem. Commun. 2007, 1343–1345.
[71] D. G. Kurth, N. Severin, J. P. Rabe, Angew. Chem. 2002, 114, 3833–
3835; Angew. Chem. Int. Ed. 2002, 41, 3681–3683.
[72] A. Dmitriev, H. Spillmann, N. Lin, J. V. Barth, K. Kern, Angew.
Chem. 2003, 115, 2774; Angew. Chem. Int. Ed. 2003, 42, 2670—2673.
[73] W. Mamdouh, M. Dong, S. Xu, E. Rauls, F. Besenbacher, J. Am.
Chem. Soc. 2006, 128, 13305–13311.
[27] S. J. H. Griessl, M. Lackinger, F. Jamitzky, T. Markert, M. Hietsc-
hold, W. M. Heckl, J. Phys. Chem. B 2004, 108, 11556–11560.
[28] S. J. H. Griessl, M. Lackinger, F. Jamitzky, T. Markert, M. Hietsc-
hold, W. M. Heckl, Langmuir 2004, 20, 9403–9407.
[29] K. G. Nath, O. Ivasenko, J. A. Miwa, H. Dang, J. D. Wuest, A.
Nanci, D. F. Perepichka, F. Rosei, J. Am. Chem. Soc. 2006, 128,
4212–4213.
[30] S. Lei, K. Tahara, F. C. De Schryver, M. Van der Auweraer, Y. Tobe,
S. De Feyter, Angew. Chem. 2008, 120, 3006–3010; Angew. Chem.
Int. Ed. 2008, 47, 2964–2968.
[31] T. Neuheuser, B. A. Hess, C. Reutel, E. Weber, J. Phys. Chem. 1994,
98, 6459–6467.
[32] G. Vidali, G. Ihm, H.-Y. Kim, M. W. Cole, Surf. Sci. Rep. 1991, 12,
135.
[33] S. Yin, C. Wang, X. Qiu, B. Xu, C. Bai, Surf. Interface Anal. 2001,
32, 248–252.
[34] S. De Feyter, A. Gesquiꢇre, M. Klapper, K. Mꢅllen, F. C. De Schryv-
er, Nano Lett. 2003, 3, 1485–1488.
[35] S. De Feyter, A. Gesquiꢇre, M. M. Abdel-Mottaleb, P. C. M. Grim,
F. C. De Schryver, C. Meiners, M. Sieffert, S. Valiyaveetyil, K.
Mꢅllen, Acc. Chem. Res. 2000, 33, 520–531.
[36] B. Venkataraman, J. J. Breen, G. W. Flynn, J. Phys. Chem. 1995, 99,
6608–6619.
[37] P. Vanoppen, P. C. M. Grim, M. Rꢅcker, S. De Feyter, G. Moessner,
S. Valiyaveetil, K. Mꢅllen, F. C. De Schryver, J. Phys. Chem. 1996,
100, 19636–19641.
[38] D. G. Yablon, D. Wintgens, G. W. Flynn, J. Phys. Chem. B 2002, 106,
5470–5475.
[39] C.-J. Li, Q.-D. Zeng, C. Wang, L.-J. Wan, S.-L. Xu, C.-R. Wang, C.-
L. Bai, J. Phys. Chem. B 2003, 107, 747–750.
[40] W. Mamdouh, H. Uji-i, A. E. Dulcey, V. Percec, S. De Feyter, F. C.
De Schryver, Langmuir 2004, 20, 7678–7685.
[41] W. Mamdouh, H. Uji-i, J. S. Ladislaw, A. E. Dulcey, V. Percec, F. C.
De Schryver, S. De Feyter, J. Am. Chem. Soc. 2006, 128, 317–325.
[42] K. Tahara, S. Furukawa, H. Uji-i, T. Uchino, T. Ichikawa, J. Zhang,
W. Mamdouh, M. Sonoda, F. C. De Schryver, S. De Feyter, Y. Tobe,
J. Am. Chem. Soc. 2006, 128, 16613–16625.
[43] H. Xu, A. Minoia, Z. Tomovic, R. Lazzaroni, E. W. Meijer,
A. P. H. J. Schenning, S. De Feyter, ACS Nano 2009, 3, 1016–1024.
[44] L. Kampschulte, M. Lackinger, A.–K. Maier, R. S. K. Kishore, S.
Griessl, M. Schmittel, W. M. Heckl, J. Phys. Chem. B 2006, 110,
10829–10836.
[45] X. Shao, X. Luo, X. Hu, K. Wu, J. Phys. Chem. B 2006, 110, 1288–
1293.
[46] Y. B. Li, G. C. Qi, Y. L. Yang, Q. D. Zeng, X. L. Fan, C. Wang, J.
Phys. Chem. C 2008, 112, 8649–8653.
[74] B. J. Gyarfas, B. Wiggins, M. Zosel, K. W. Hipps, Langmuir 2005, 21,
919–923.
[75] W. A. English, K. W. Hipps, J. Phys. Chem. C 2008, 112, 2026–2031.
[76] F. Tao, S. L. Bernasek, J. Am. Chem. Soc. 2005, 127, 12750–12751.
[77] X. Zhang, Q. Chen, G.-J. Deng, Q.-H. Fan, L.-J. Wan, J. Phys.
Chem. C 2009, 113, 16193–16198.
[78] X. Zhang, T. Chen, Q. Chen, G.-J. Deng, Q.-H. Fan, L.-J. Wan,
Chem. Eur. J. 2009, 15, 9669–9673.
[79] The self-assembly of 4-ABAs was also studied in tetradecane. The
2D patterns in tetradecane are identical to those formed in 1-phe-
nyloctane. For STM data in tetradecane, please refer to Figure S7 in
the Supporting Information.
[47] F. Tao, S. L. Bernasek, Langmuir 2007, 23, 3513–3522.
[48] Y. Yang, C. Wang, Curr. Opin. Colloid Interface Sci. 2009, 14, 135–
147 and references therein.
[49] S. Ahn, C. N. Morrison, A. J. Matzger, J. Am. Chem. Soc. 2009, 131,
7946–7947.
Received: June 11, 2010
Published online: November 9, 2010
14458
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14447 – 14458