Supramolecular synthons on surfaces: Controlling dimensionality and periodicity of tetraarylporphyrin assemblies by the interplay of cyano and alkoxy substituents

Article abstract of DOI:10.1002/chem.200800746

The self-assembly of three porphyrin derivatives was studied in detail on a Cu(111) substrate by means of scanning tunneling microscopy (STM). All derivatives have two 4-cyanophenyl substituents in diagonally opposed meso-positions of the porphyrin core, but differ in the nature of the other two meso-alkoxyphenyl substituents. At coverages below 0.8 monolayers, two derivatives form molecular chains, which evolve into nanoporous networks at higher coverages. The third derivative self-assembles directly into a nanoporous network without showing a one-dimensional phase. The pore-to-pore distances for the three networks depend on the size and shape of the alkoxy substituents. All observed effects are explained by 1) different steric demands of the alkoxy residues, 2) apolar (mainly dispersion) interactions between the alkoxy chains, 3) polar bonding involving both cyanophenyl and alkoxyphenyl substituents, and 4) the entropy/enthalpy balance of the network formation.

Full text of DOI:10.1002/chem.200800746

DOI: 10.1002/chem.200800746
Supramolecular Synthons on Surfaces: ControllingDimensionality and
Periodicity of Tetraarylporphyrin Assemblies by the Interplay of Cyano and
Alkoxy Substituents
Nikolai Wintjes,*[a] Jens Hornung,*[b] Jorge Lobo-Checa,[a] Tobias Voigt,[b]
Tomµsˇ Samuely,[a] Carlo Thilgen,[b] Meike Stçhr,[a] FranÅois Diederich,*[b] and
Thomas A. Jung*[c]
Abstract: The self-assembly of three
porphyrin derivatives was studied in
chains, which evolve into nanoporous
networks at higher coverages. The
third derivative self-assembles directly
into a nanoporous network without
showing a one-dimensional phase. The
pore-to-pore distances for the three
networks depend on the size and shape
of the alkoxy substituents. All observed
effects are explained by 1) different
steric demands of the alkoxy residues,
2) apolar (mainly dispersion) interac-
tions between the alkoxy chains,
3) polar bonding involving both cyano-
phenyl and alkoxyphenyl substituents,
and 4) the entropy/enthalpy balance of
the network formation.
detail on a Cu(111) substrate by means
A
of scanning tunneling microscopy
(STM). All derivatives have two 4-cya-
nophenyl substituents in diagonally op-
posed meso-positions of the porphyrin
core, but differ in the nature of the
other two meso-alkoxyphenyl substitu-
ents. At coverages below 0.8 monolay-
ers, two derivatives form molecular
Keywords: molecular recognition ·
networks on surfaces · porphyri-
noids · scanning probe microscopy ·
supramolecular chemistry
Introduction
search for many years,[3] surface and interface self-assem-
blies allow the addressing of individual units.[4,5] To control
the self-assembly process on the surface,[6] a detailed under-
standing of the molecule–surface and intermolecular interac-
tions involved is crucial. Their interplay can lead to a variety
of phases for the same coverage,[7,8] but the observed phases
can also depend on the surface coverage.[9] In the latter case,
several phases might coexist for a small coverage region,[10]
or a transition from one phase to the other occurs.[11]
Since the invention of STM in 1981, a significant number
of two-dimensional patterns at the solid–liquid and the
solid–vacuum interface has been reported.[12] Among them,
there are some striking examples of self-assemblies that are
guided, for example, by solvent interactions,[13] surface pre-
patterning,[14] molecular symmetry,[15] the use of elaborated
binding motifs as they are found in hydrogen bonding,[16] in-
teractions among polar groups,[17] metal complexation,[18] or
by exploiting the interactions of long alkyl chains with the
underlying substrate, especially on highly ordered pyrolytic
graphite (HOPG).[19]
Complex molecular layers on surfaces with engineered ar-
chitectures and properties[1] are expected to play an impor-
tant role in the development of future devices at the nano-
scale.[2] In contrast with the three-dimensional crystal pack-
ing of molecules or tectons, which has been the focus of re-
[a] Dr. N. Wintjes, Dr. J. Lobo-Checa, T. Samuely, Dr. M. Stçhr
Department of Physics, University of Basel
4056 Basel (Switzerland)
Fax : (+41)61-267-3784
E-mail: n.wintjes@unibas.ch
[b] J. Hornung, Dr. T. Voigt, Dr. C. Thilgen, Prof. Dr. F. Diederich
Laboratorium für Organische Chemie, ETH-Zürich
Hçnggerberg, HCI, 8093 Zürich (Switzerland)
Fax : (+41)44-632-1109
E-mail: hornung@org.chem.ethz.ch
diederich@org.chem.ethz.ch
[c] Dr. T. A. Jung
Laboratory for Micro- and Nanotechnology
Paul Scherrer Institute, 5232 Villigen PSI (Switzerland)
Fax : (+41)56-310-2646
Particularly interesting self-assemblies, in view of future
applications, are 1) molecular chains because of their poten-
tial to act as organic wires and 2) porous networks due to
their capability to recognize[20,21] and host molecular
E-mail: thomas.jung@psi.ch
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
5794
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 5794 – 5802

FULL PAPER
Scheme 1. Chemical structures of the three porphyrin derivatives 1–3 used in this study.
guests.[5,17,22] Previous work has shown that it is possible to
modify the pore dimensions of such networks by changing
the entire structure of the molecular building block.[8,23] So
far, however, no systematic study has focused on how the
variation of functional subunits attached to the same molec-
ular core influences the structures of self-assembled porous
layers on metal surfaces.
Molecules 1–3 (Scheme 1) investigated in the present
study are derived from the well-established tetrakis(meso-
phenyl)porphyrin core.[24,25] Their meso-(4-cyanophenyl)
well as the temperature-activated mobility and correspond-
ing steric requirement of the alkoxyphenyl residues are
identified as the key parameters governing the resulting
structures, that is, one-dimensional molecular chains or two-
dimensional porous layers. Furthermore, we provide evi-
dence that advanced structural control is not only ensured
by the well-known supramolecular synthons A and B shown
in Scheme 2, but also that spatial requirements and apolar
as well as polar interactions of the alkoxyphenyl substituents
induce additional assembly modes (C and D in Scheme 2).
groups were expected to undergo dipolar[26] and C H···N C
hydrogen-bonding interactions,[22] thereby providing the
main ordering and orienting ingredient for network forma-
tion on surfaces. Such structural motifs, assembled by inter-
molecular interactions between functional groups, have
been defined by Desiraju as “supramolecular synthons” and
used for the planning of supramolecular assemblies.[27,28]
Supramolecular synthons composed of two or three 4-cyano-
phenyl groups had initially been observed in 3D crystals,[29]
and their linear dimeric and cyclic trimeric interaction
motifs (A and B in Scheme 2) have been found to enforce
the formation of distinct crystal lattices. Similar structural
motifs have also recently been introduced to form defined
supramolecular assemblies on surfaces.[4,17,20,22,30] As a novel
feature, variable functional alkoxy subunits were introduced
in this work to further control the nature of the desired as-
semblies. The two (in 1) or four alkoxy groups (in 2 and 3)
have different spatial demands and undergo not only apolar
(mainly dispersion) interactions with each other, but can
also participate in polar interactions that stabilize the net-
works. Another important feature of the alkoxy residues is
their tendency to “condense” or exhibit fluidlike dynamic
behavior depending on the environmental conditions,[31] as is
well known from liquid crystals.[32]
À
ꢀ
Herein, we present a systematic study on the influence of
the different alkoxyphenyl substituents in porphyrins 1–3 on
the resulting supramolecular arrangements at the solid–
vacuum interface by using the concept of synthons for
supramolecular surface assemblies.[28] The architecture as
Scheme 2. Examples of supramolecular synthons formed by two (A) or
three (B) interacting 4-cyanophenyl groups, as well as those formed by
the interaction of the 4-cyanophenyl group with a 3,5-alkoxyphenyl
group (C and D).
Chem. Eur. J. 2008, 14, 5794 – 5802
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5795

N. Wintjes, J. Hornung, F. Diederich, T. A. Jung et al.
Results and Discussion
around 3.5 V and 50 ms, one can randomly switch between
imaging modes (see the Supporting Information). Hence, it
is plausible that the different modes are caused by subtle
changes of the tip, for example, caused by an adsorbate on
the tip apex, as also observed in other STM studies.[35]
Depending on the scanning conditions, sections of the
STM images may show a characteristic fuzzy signal regard-
less of the imaging mode (Figure 1e[36]). These streaks can
be related to parts of the molecules that move while being
passed by the scanning tip.[37] For the present systems, we at-
tribute this effect to the mobility of the alkoxy chains. The
streaks diminish with increasing proximity of tip and
sample. This can be associated with an increasing interaction
between the scanning tip and the alkoxy chains, which are
thereby pushed aside.[38]
Appearance of the molecular compounds in STM images:
The selected molecular building blocks 1–3 (Scheme 1) were
sublimed onto a Cu(111) substrate at sub-monolayer cover-
G
ages. Consistent with earlier observations,[33] the influence of
the STM tip leads, for all three derivatives, mainly to two
different imaging modes in which either the p system pro-
vides the dominant contrast (“p-imaging mode”, Figure 1a
and c), or all parts of the molecule are equally visible (“full
imaging mode”, Figure 1band d). In the p-imaging mode,
Self-assembly at coverages <0.8 monolayer: For porphyrin
1, we observed self-assembly into a nanoporous, hexagonal
network with p3 symmetry even at coverages <0.05 ML
(ML=monolayer; Figure 4a, left), after the well-known ini-
tial decoration of the step edges.[39] This network exhibits
the same internal geometry as previously published porphy-
rin assemblies[22] and will be described in more detail in the
section on tailored nanoporous networks. In contrast, mole-
cules 2 and 3 form long chains with the same characteristics
for both derivatives after decorating the step edges and for
molecular coverages up to approximately 0.8 ML (Fig-
ure 2a). The chains nucleate either at one of the molecules
decorating the step edges or at a defect on a terrace and are
stable up to at least 200 K. At room temperature, the mole-
cules are observed in a 2D mobile phase.[40]
Figure 1. STM images showing different appearances of porphyrin deriva-
Within the chains of 2 and 3, single molecules can be
clearly identified (Figure 2b, top). The porphyrin core and
the phenyl rings appear as described above for the p-imag-
ing mode. The cyano groups are not resolved. The alkoxy
chains appear as one lobe each, situated above and below
the porphyrin ring and the 4-cyanophenyl substituents. No
significant difference in the appearance of molecules 2 and
3 was found in the STM images.
A closer analysis reveals that the chains exhibit mainly
two interconnecting modes for adjacent molecules. The first
one is a “straight” connection which simply elongates the
chain (circle a in Figure 2a). The second one leads to a
“kink” (circle b), which changes the direction of the chain,
mostly by approximately 30 degrees. Also, sections are
found in which a kink towards one side is immediately fol-
lowed by a kink towards the other side, leading to a zigzag
type, overall straight section (circle c).
tive 1 on Cu(111) at coverages >0.8 ML (ML=monolayer). a) and b)
A
represent the parts of c) and d), respectively, that are enclosed by the
dashed rectangles. They show an individual molecule inside the network
of 1. Two different imaging modes could be observed depending on the
tip conditions: in a) and c) the p system of the molecules is pronounced
(c: 7.5”7.5 nm2, U=À0.8 V, I=16 pA), in b) and d) all parts of the mole-
cules provide equal contrast (d: 7.5”7.5 nm2, U=À1.3 V, I=24 pA).
e) An image of derivative 2 showing that mobile parts of the molecules
can lead to characteristic “streaks” in STM images (7.5”7.5 nm2, U=
À2.5 V, I=20 pA).[36]
the porphyrin ring is imaged as two opposing lobes separat-
ed by a dark line. This appearance is caused by a saddle-
shaped deformation of the porphyrin ring induced by steric
hindrance between the pyrrolic hydrogen atoms (b-hydrogen
atoms) of the core and hydrogen atoms of the meso-phenyl
rings.[34] Consequently, two pyrrole rings are tilted upwards
and provide strong contrast in the STM images, whereas the
other two rings are tilted downwards and appear as a dark
line (Figure 1a). The four phenyl rings can be recognized as
four lobes in close vicinity to the porphyrin ring, each pro-
viding nearly the same contrast as the pyrrole rings that are
tilted upwards. These features allow for a precise identifica-
tion of an individual molecule even inside a network (see
the superimposed molecular structure in Figure 1a). The
mode of imaging (p- or full imaging mode) is not influenced
by the applied voltage. However, by using voltage pulses of
Most probably, in a straight connection the cyano groups,
which are not visible in the STM images, are lying antiparal-
lel and interact through dipole–dipole interactions, and ben-
À
ꢀ
efit additionally from C H···N C interactions, as shown for
molecules 3 and 4 in Figure 2b, bottom.[41] This type of inter-
action corresponds to the dimeric supramolecular synthon A
in Scheme 2 and was theoretically described by Y. Okuno
et al. for a benzonitrile dimer on a Au
C
5796
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 5794 – 5802

Supramolecular Synthons on Surfaces
FULL PAPER
meso-aryl substituents to the porphyrin core is observed.[46]
After adsorption on the metal surface, however, the dihedral
angle between the meso-aryl and porphyrin planes is typical-
ly fixed in a position that depends on the interaction with
the substrate.[25] Consequently, two conformational isomers
corresponding to a positive or negative tilt angle of the al-
koxyphenyl substituents are observed upon adsorption. In
STM images, these can be distinguished by the relative ori-
entation (+45 or À458) of the characteristic diagonal dark
line with regard to the cyanophenyl substituents (cf. mole-
cules 2 and 3 in Figure 2b, top). The two conformational iso-
mers are found with equal probabilities for all three por-
phyrins studied herein, which is in agreement with earlier
observations.[47] Conformers exhibiting parallel rotation
around opposite aryl–porphyrin bonds have not been identi-
fied in our study. By analyzing STM images that exhibit
both good resolution of the adsorbed molecules and atomic
resolution of the substrate (Figure 2c), it becomes apparent
that the characteristic dark lines are also aligned with the
main atomic directions of the supporting substrate (green
star in Figure 2c) within an accuracy of Æ58. This observa-
tion, in combination with the two possible orientations of
the dark line with respect to the axis through the two cyano-
phenyl rings, can explain the changes in direction of 308
within a kink: Two different enantiomers, which are adja-
cent within a chain (as molecules 2 and 3 in Figure 2b, top)
and are aligned along different principal directions of the
threefold symmetric (111) substrate, exhibit a relative orien-
Figure 2. STM images showing the arrangements of porphyrins 2 (a and
b) and 3 (c) on Cu(111) at coverages <0.8 ML (see also the Supporting
A
Information). Both molecules arrange in chainlike structures and individ-
ual molecules can be clearly identified. a) STM image showing typical
chains formed by molecule 2 (41”41 nm2; U=À1.5 V, I=20 pA. The
color scale of the image was adjusted to provide good contrast at the ter-
races for highlighting the molecular features. A step edge, seen in the
image as a black line, runs vertically from top to bottom.). b) Top: A de-
tailed STM image (12”6 nm2; U=À1.6 V, I=11 pA) showing the main
types of connections between adjacent molecules in the molecular chains.
Bottom: A tentative model, generated with Spartan,[41] for the intermo-
lecular interactions within the chains. c) Simultaneous imaging (12.5”
25 nm2; U=+0.9 V, I=10 pA) of molecular chains of 3 and substrate
tation of (120À
A
tical enantiomers (as molecules 3 and 4 in Figure 2b, top),
which are aligned along the same principal directions, form
a straight connection. In the STM images, this is revealed by
the observation that in a straight connection two adjacent
molecules look like perfect copies of each other (with the
orientation of the dark lines), whereas in a kink they appear
as mirror images. Notably, this result shows that within the
chains an intermixing of the two different conformational
enantiomers occurs. For two-dimensional assemblies of simi-
lar porphyrin derivatives, long-range interactions that
extend beyond nearest neighbors have also been reported.[20]
However, no clear evidence for such interactions was found
here, since they should lead to a preferred arrangement with
either straight, curved, or zigzagged chain sections.
atoms of Cu(111). The orientation of the characteristic dark lines (see de-
R
scription in the main text) is marked by black lines, which allow for an
easy correlation with the principal directions of the substrate (green
star).
Sometimes, a small variation of this bonding motif can be
found, as seen for molecules 1 and 2 in Figure 2b, bottom.
Here, the molecules in a straight section are lying closer to-
gether and show a small vertical and lateral displacement.
Our calculations indicate that this arrangement benefits
ꢀ
À
from a weak C N···H C hydrogen bond between the cyano
group and the b-hydrogen atoms of the porphyrin ring, as
indicated by the molecular model in the lower part of Fig-
ure 2b.[41] Within the kinks, the dipole–dipole interaction is
weakened because the dipoles are not perfectly antiparallel
in such an arrangement (molecules 2 and 3 in Figure 2b,
bottom). Nevertheless, this loss in interaction energy may be
partially compensated by van der Waals interactions of the
alkoxy chains,[45] which are closer together than in the
straight sections.
Branchingof the porphyrin chains : Next to the two types of
intermolecular connection modes observed in linear chain
sections, three different types of branching arrangements
were observed (Figure 3). For the first one (Figure 3a), a
cyano group of one molecule undergoes favorable
ꢀ
À
C N···H C hydrogen bonding with a b-hydrogen atom of
the porphyrin core or a hydrogen atom in the meta position
to the nitrile group in a 4-cyanophenyl ring of an adjacent
molecule. This motif has been found for both molecules 2
and 3. In a second dimeric branching arrangement (Fig-
ure 3b), the 4-cyanophenyl group of one molecule forms a
Influence of the underlyingCu (111) surface on the porphy-
A
ꢀ
À
À
rin chain formation: In solution, a temperature dependent,
hindered rotation around the single bond connecting the
C N···H C hydrogen bond with the C H group ortho to
the two alkoxy substituents of one phenyl ring of a second
Chem. Eur. J. 2008, 14, 5794 – 5802
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5797

N. Wintjes, J. Hornung, F. Diederich, T. A. Jung et al.
The nanoporous network
formed by compound 1, which
can be observed at all sub-
monolayer coverages >0.05 ML
G
(we will refer to it as network 1;
Figure 4a, left), exhibits a pore-
to-pore distance of (30.9Æ
2.0) Š. The unit cell contains
one pore and three molecules.
The model (Figure 4a, right) re-
veals that the pores are chiral.
Thus, in contrast with the case
of the chains formed by 2 and
3, in which an intermixing of
the two conformational enan-
tiomers was observed, for nano-
porous network 1 a separation
of the two conformational en-
antiomers occurs and conse-
Figure 3. High-resolution STM images (6.4”6.4 nm2; U=À1.5 V, I=20 pA) of molecule 2 on Cu
N
ꢀ
À
cule 3. a) Branching by C N···H C hydrogen-bond formation between the cyano group of one molecule and
À
C H groups on the porphyrin core and/or the 4-cyanophenyl substituent of another molecule. b) Branching by
ꢀ
À
À
a C N···H C hydrogen bond between the cyano group of one molecule and the C H residue ortho to the
alkoxy substituents on the phenyl ring of another molecule. c) Branching induced by a trimeric cyanophenyl
array, such as that shown for the trimeric synthon B in Scheme 2.
molecule (cf. C in Scheme 2). This hydrogen bond benefits
quently two types of enantiomerically pure domains can be
found on the copper surface.[51] The pores are formed by the
alkoxy chains of six different molecules, each of which is
part of two neighboring pores. Through this, network 1 ben-
efits from dispersion interactions from the six alkoxy groups
converging into each pore. The network is furthermore sta-
bilized by a so far unknown trimeric supramolecular syn-
thon: the 4-cyanophenyl groups and b-hydrogen atoms of
three adjacent porphyrins form a cyclic arrangement with
d+
À
from the substantial polarization of the C H
residue
ortho to two electronegative oxygen atoms. Secondary re-
pulsive N···O interactions[48] are minimized because the lone
pairs of the sp2-hybridized oxygen atoms point away from
ꢀ
the negatively polarized nitrogen atom of the C N group.
This type of dimeric branching, which is again observed for
both molecules 2 and 3, clearly demonstrates that the intro-
duction of the alkoxy substituents enables new self-assem-
bling motifs, which could also be viewed as new supramolec-
ular synthons. A dimeric structure similar to C has already
been reported in crystal structures,[49] but is so far unknown
on the surface. The formation of a third association motif
(Figure 3c), a cyclic trimeric arrangement of cyanophenyl
moieties resembling the trimeric supramolecular synthon B
in Scheme 2, was observed exclusively for molecule 2.
Within this trimeric assembly, a cyanophenyl group of each
ꢀ
À
three C N···H C hydrogen bonds (blue circle in Figure 4a).
In network 1 the alkoxyphenyl and cyanophenyl groups are
spatially separated from each other: the former reside in the
pores, whereas the latter are found inside the network.
Molecule 2 forms a network (network 2; Figure 4b, left)
with a pore-to-pore distance of (33.5Æ1.2) Š, which is
slightly larger (about 8%) than that of network 1. Again,
the unit cell contains one pore and three molecules per unit
cell. According to the model (Figure 4b, right), the struc-
tures of network 2 and network 1 are similar: The pores are
chiral and surrounded by six molecules, in which each mole-
cule is shared by two pores. The alkoxy groups converge
into the pores, whereas the three cyanophenyl groups are in-
volved in a trimeric interaction motif. However, this time
the porphyrin core does not participate in the trimers, but
ꢀ
molecule acts as hydrogen-bond acceptor (C N) towards
À
one and as a donor (C H) towards the other partner, there-
by interconnecting three molecules with three hydrogen
bonds. Similar trimeric arrangements have been reported in
the literature for cyanophenyl derivatives in the gas
phase,[43] in two-dimensional assemblies,[22,30,42] and also in
3D crystal structures.[50]
ꢀ
À
instead hydrogen bonding involves C N and H C moieties
from 4-cyanophenyl substituents of three adjacent porphyr-
ins (blue circle in Figure 4b). This interaction type is the
same as that found for the branching mechanism shown in
Figure 3c and resembles the trimeric supramolecular syn-
thon B in Scheme 2. The space required for this motif is
slightly higher than that for the similar trimeric arrangement
in network 1 (cf. the blue circles in Figure 4a and b), which
explains why the pore-to-pore distance in network 2 is
slightly larger than that in network 1.
Self-assembly at coverages >0.8 monolayer: Tailored nano-
porous networks: Above
a coverage of approximately
0.8 ML, all three porphyrins 1–3 self-assemble into nanopo-
rous, hexagonal networks with p3 symmetry (Figure 4). We
were able to develop a tentative model for each of the three
networks, guided by the following aspects: First, we com-
pared the networks observed in the present study to a net-
work previously reported by our group, which is formed by
a similar porphyrin derivative.[22] Second, we took advantage
of the p-imaging mode, which allowed reliable identification
of single molecules inside the network. Finally, we related
the fuzzy parts observed in some STM images (see above)
with the alkoxy chains.
Molecule 3, which differs from molecule 2 only by the in-
creased length of the alkoxy chains, forms a network (net-
work 3; Figure 4c, left) with a pore-to-pore distance of
(48.0Æ1.4) Š, which is significantly larger than the distances
5798
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 5794 – 5802

Supramolecular Synthons on Surfaces
FULL PAPER
truly remarkable, which shows the interplay of alkoxyphenyl
and cyanophenyl rings. An asymmetric trimeric motif (blue
circle in Figure 4c, cf. D in Scheme 2) forms in which one cy-
anophenyl substituent acts both as a hydrogen-bond donor
À
ꢀ
and as an acceptor. It forms a C H···N C hydrogen bond
with the cyanophenyl substituent of a second molecule of 3
ꢀ
and, additionally, interacts through its C N group with the
À
positively polarized hydrogen of the C H moiety ortho to
the two alkoxy substituents in the phenyl ring of a third
molecule of 3 (see also green rectangle in Figure 4c). The
latter interaction had already been observed in the branch-
ing mechanism shown in Figure 3b. A trimeric structure sim-
ilar to D has not yet been reported. Additional stabilizing
interaction motifs become apparent in network 3, caused by
the imperfect separation of the two synthons used: Two
(orange rectangle) or three (red circle) alkoxy chains, which
are found outside the pores, interact through van der Waals
dispersion forces, thereby further stabilizing the network.
The hereby formed “apolar pockets” (red circle) can hardly
be seen in the STM images.
Thermodynamic considerations for the formation of the
chains and networks: Self-assembled structures are formed
by reversible association of molecular building blocks, there-
by representing thermodynamic minima.[52] To explain the
phenomena described in this manuscript, one has to take
into account the different energy contributions guiding the
formation of networks and chains on the surface. First of all,
depending on the strength of the molecule–surface and in-
termolecular interactions, organic adsorbates on surfaces
tend to maximize the coverage (molecules per area) and
even a rearrangement into a new phase can be preferred
over a second layer arrangement.[40,53] Next, due to the con-
siderably weaker intermolecular interaction energies in the
supramolecular aggregates, when compared with covalent
bonding, the interplay of entropy and enthalpy is of much
higher importance here.[52] Finally, on threefold symmetric
surfaces, the 4-cyanophenyl group is known to favor a di-
meric arrangement with antiparallel dipoles or a cyclic
trimer (A and B in Scheme 2).[30,42] In this work, we addi-
tionally discovered two supramolecular synthons, which in-
volve interactions between both cyanophenyl and alkoxy-
phenyl substituents (C and D in Scheme 2). Due to the com-
plexity of the presented surface assemblies, accurate numeri-
cal calculations are not straightforward. However, the ob-
served structures have to correspond to an energetically
optimized interplay of the aforementioned contributions.
Network 1 (Figure 4a) features different remarkable char-
acteristics: although it is a porous phase, the fraction of cov-
ered surface area, as determined from STM images, is
Figure 4. Left: STM images (all 15”15 nm2; a) U=À1.3 V, I=24 pA; b)
U=À1.5 V, I=20 pA; c) U=À0.4 V, I=17 pA) of the nanoporous net-
works (network 1 (a), network 2 (b), and network 3 (c)) formed by por-
phyrins 1–3 at coverages close to one monolayer. The networks exhibit
different pore-to-pore distances d. Right: Tentative models of the three
networks. The filled gray circles indicate the positions of the pores. The
colored circles and rectangles indicate different bonding arrangements
stabilizing the network (see text for details). The pores in c) exhibit char-
acteristic streaks that indicate the mobility of the alkoxy chains.
in network 1 and network 2 (by about 43 and 55%, respec-
tively). Remarkably, in this network the unit cell still con-
tains one pore, whereas the number of molecules enclosed is
doubled to six molecules. As a common characteristic, a
single pore is formed by six molecules (Figure 4c, right). But
unlike in the other two networks, each molecule belongs dis-
tinctly to one pore, which leads to the significantly larger
pore-to-pore distance. The mechanism for self-assembly is
A
stituents from the more polar cyanophenyl moieties is ob-
served. The alkoxy chains interact with each other in the
pores through van der Waals forces, whereas outside the
pores the cyano groups form energetically favored trimeric
hydrogen-bonding arrangements with the b-hydrogen atoms
of the porphyrin cores of adjacent molecules (blue circle in
Chem. Eur. J. 2008, 14, 5794 – 5802
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5799

N. Wintjes, J. Hornung, F. Diederich, T. A. Jung et al.
Figure 4a). Network 1 exhibits the densest structure ((0.36Æ
0.05) molecules per nm2) of the three investigated
(Figure 4).
molecule of 3 to participate in a symmetric trimeric arrange-
ment similar to those of network 1 and network 2. Instead, a
hydrogen bond is formed between the bis(alkoxy)phenyl
A
For molecules 2 and 3, the formation of the network is
hindered by a restriction of the access of potential binding
partners to the hydrogen atoms of the pyrrole units and the
4-cyanophenyl rings. As a consequence, molecules 2 and 3
do not assemble into a network at low coverage, but instead
form chains through antiparallel dipole–dipole interactions.
By comparing this behavior to that of molecule 1, which
readily forms trimeric structures instead of the also possible
dimers even at low coverage, we conclude that without
steric hindrance by an additional substituent, for the 4-cya-
nophenyl group, the dipole–dipole interaction is energetical-
ly less favored than the formation of trimers through hydro-
gen bonding. This is in agreement with a calculation by
Okuno et al.[42] and was also observed for a different molec-
ular structure with similar substituents.[54]
With increasing number of molecules 2 and 3 on the sur-
face, the number of branches in the chains also rises (see
the Supporting Information). Because the branching motifs
of the chains can also be found in the networks, we assume
that the growth of the network originates from the branch-
ing points once a critical ratio between the number of mole-
cules and the free surface area is reached. In fact, STM
images reveal that the transition from the chains to the net-
works starts with the formation of single pores (see the Sup-
porting Information).
ring of the third molecule and a cyano acceptor group of the
other two molecules (cf. Figure 3band D in Scheme 2). In
this new asymmetric configuration, the alkoxy chains do not
interfere with each other. The third molecule is then a start-
ing point for a second pore. Due to this uncommon asym-
metric assembly, half of the alkoxy chains of each molecule
are found outside the pores. This leads to an even stronger
restriction of their mobility, presumably amplified by an in-
creased interaction with the substrate, and a concomitant
loss of entropy, which becomes evident in the STM images
by the absence of streaks in the corresponding parts of the
network (Figure 4c, left) for those alkoxy chains situated
outside the pores (Figure 4c, right, red circle and orange rec-
tangle). As in network 1 and network 2, in network 3 the
alkoxy chains show a tendency to associate by forming
pores, in which they are still mobile, and additionally by
condensation outside the pores, thereby increasing the
amount of van der Waals interactions in the network. The
density of network 3 is (0.30Æ0.02) molecule per nm2, which
is comparable to that of network 2.
Conclusions
We have presented a systematic study on the influence of
different alkoxyphenyl substituents on the resulting surface
Similar to network 1, the alkoxy chains in network 2 (Fig-
ure 4b) are situated inside the pores and are thus separated
from the cyano groups. This arrangement is possible even
though the number of alkoxy chains per molecule of 2 in-
creased by a factor of two compared with 1. However, the
corresponding higher packing density inside the pores re-
stricts the mobility of the alkoxy chains. This results in a loss
of entropy that can be compensated for by increased van
der Waals interactions between the alkoxy chains.[55] The
cyano groups form the energetically favored trimeric
assemblies on a Cu(111) surface. By systematically varying
G
the steric bulk, we showed how an elaborated change in the
molecular architecture influences the resulting assembly on
Cu(111) at different surface coverages. Variation of the
R
alkoxy chains in size (chain length), number, and position
controls the dimensionality of the resulting structure, that is,
one-dimensional wires or two-dimensional porous networks,
as well as the pore-to-pore distance in the observed net-
works without significantly affecting the pore diameter.
The porphyrin-based building blocks (1–3) presented in
this study contain, in addition to the previously used 4-cya-
nophenyl rings, mono- (1) and dialkoxy-substituted (2 and
3) phenyl rings. These additional functional groups induce
new assembly motifs: besides the well-established dimeric
and trimeric supramolecular synthons involving the interac-
tion between two or three cyanophenyl groups (A and B in
Scheme 2), we observed in network 3 a new trimeric asym-
metric interaction motif (D in Scheme 2). Here, one of two
cyanophenyl groups of neighboring porphyrins forms a
ꢀ
À
C N···H C hydrogen-bonded arrangement (blue circle in
Figure 4b). However, the b-hydrogen atoms of the porphyrin
core cannot act as hydrogen-bond donors, presumably be-
cause the increased steric demand of the alkoxy chains pre-
vents adjacent molecules from being as close as in the case
of network 1. Therefore, the cyano group forms a hydrogen
bond with the hydrogen atoms of the 4-cyanophenyl group.
This results in a larger pore-to-pore distance and a de-
creased density of network 2 ((0.31Æ0.02) molecules per
nm2) compared with network 1.
ꢀ
À
À
As described above, we assume that the formation of net-
work 3 nucleates from chain intersections that form a single
pore. However, it appears that the even higher steric
demand of the larger alkoxy chains of 3 in comparison with
2 does not allow for the formation of a 4-cyanophenyl-based
trimeric interaction motif, which is characteristic for net-
work 1 and network 2. In network 3 only two molecules as-
C N···H C hydrogen bond with the polarized H C residue
ortho to the electron-withdrawing alkoxy groups of a third
neighbor (cf. blue circle in Figure 4c). This hydrogen bond is
also clearly revealed in a similar, but dimeric arrangement
(C in Scheme 2), observed at branching points of the molec-
ular chains formed by 2 and 3 at low coverages (Figure 3b).
The different assemblies are discussed in terms of the
compensation of entropic losses, due to increased restriction
of the flexibility of the alkoxy chains with increasing size, by
ꢀ
À
sociate through C N···H C hydrogen bonding. Tentatively,
this behavior can be explained by a lack of space for a third
5800
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 5794 – 5802

Supramolecular Synthons on Surfaces
FULL PAPER
[2] http://www.itrs.net/reports.html
enthalpic interaction energies. This phenomenon is of gener-
al importance in the self-assembly of thermodynamically
controlled systems, such as the folding of biomolecules in a
cellular fluid.[56] It should be noted that for biomolecular
self-assembly, solvation processes considerably affect the
thermodynamics,[57] which is not the case for the dry, in
vacuo assemblies studied herein.
[3] a) G. Desiraju, Crystal Engineering: The Design of Organic Solids,
Elsevier, New York, 1989; b) M. W. Hosseini, Acc. Chem. Res. 2005,
38, 313.
[4] D. Bonifazi, H. Spillmann, A. Kiebele, M. de Wild, P. Seiler, F. Y.
Cheng, H. J. Güntherodt, T. Jung, F. Diederich, Angew. Chem. 2004,
116, 4863; Angew. Chem. Int. Ed. 2004, 43, 4759.
[5] M. Wahl, M. Stçhr, H. Spillmann, T. A. Jung, L. H. Gade, Chem.
Commun. 2007, 1349.
Future work aims to exploit other supramolecular syn-
thons known from supramolecular chemistry, directed to-
wards supramolecular surface assemblies. In particular, the
controlled change of the pore diameter, which might be pos-
sible by a sophisticated variation of the two different inter-
action groups (cyanophenyl and alkoxyphenyl) used in this
study would be the next step to build more complex, ad-
dressable supramolecular structures with well-defined prop-
erties far beyond the currently established toolbox.
[6] J.-M. Lehn, Supramolecular Chemistry, VCH, Weinheim, 1995.
[7] a) K. Morgenstern, S. W. Hla, K. H. Rieder, Surf. Sci. 2003, 523, 141;
b) J. Y. Grand, T. Kunstmann, D. Hoffmann, A. Haas, M. Dietsche,
J. Seifritz, R. Mçller, Surf. Sci. 1996, 366, 403.
[8] S. Stepanow, N. Lin, J. V. Barth, K. Kern, J. Phys. Chem. B 2006,
110, 23472.
[9] a) M. Stçhr, M. Wahl, C. H. Galka, T. Riehm, T. A. Jung, L. H.
Gade, Angew. Chem. 2005, 117, 7560; Angew. Chem. Int. Ed. 2005,
44, 7394; b) G. E. Poirier, Langmuir 1999, 15, 1167; c) B. Xu, C. G.
Tao, E. D. Williams, J. E. Reutt-Robey, J. Am. Chem. Soc. 2006, 128,
8493.
[10] C. B. France, P. G. Schrçder, J. C. Forsythe, B. A. Parkinson, Lang-
muir 2003, 19, 1274.
Experimental section
[11] J. V. Barth, J. Weckesser, G. Trimarchi, M. Vladimirova, A. De Vita,
C. Z. Cai, H. Brune, P. Günter, K. Kern, J. Am. Chem. Soc. 2002,
124, 7991.
[12] a) T. Fritz, M. Hara, W. Knoll, H. Sasabe, Mol. Cryst. Liq. Cryst.
Sci. Technol. Sect. A 1994, 252, 561; b) C. Ludwig, B. Gompf, J. Pe-
tersen, R. Strohmaier, W. Eisenmenger, Z. Phys. B 1994, 93, 365;
c) E. I. Altman, R. J. Colton, Phys. Rev. B 1993, 48, 18244; d) M. Ku-
nitake, N. Batina, K. Itaya, Langmuir 1995, 11, 2337.
[13] a) M. Lackinger, S. Griessl, W. A. Heckl, M. Hietschold, G. W.
Flynn, Langmuir 2005, 21, 4984; b) F. Tao, S. L. Bernasek, Langmuir
2007, 23, 3513.
[14] a) F. J. Himpsel, T. Jung, A. Kirakosian, J.-L. Lin, D. Y. Petrovykh,
H. Rauscher, J. Viernow, MRS Bull. 1999, 24, 20; b) K. Aït-Man-
sour, P. Ruffieux, W. Xiao, P. Grçning, R. Fasel, O. Grçning, Phys.
Rev. B 2006, 74, 195418.
Synthesis: Three different zinc porphyrin derivatives (1–3) were synthe-
sized by condensation of 5-(4-cyanophenyl)dipyrromethane with the cor-
responding alkoxybenzaldehydes and subsequent insertion of Zn2+ (see
the Supporting Information).[58]
Testingthe thermal stability of molecules 1–3 duringsublimation : Large
organic molecules often decompose under the conditions for ultra-high
vacuum (UHV) sublimation. To verify the thermal stability of molecules
1–3, sublimation tests were performed prior to the STM experiments at
10À5 mbar in a tailor-made sublimation apparatus, including a turbomo-
lecular vacuum pump and a liquid tin heating bath (see the Supporting
Information). In a typical experiment, the molecular compound (1 mg)
was sublimed at bath temperatures between 380 and 4208C within 10–
30 min. 1H NMR spectroscopic and mass spectrometric analysis of the
sublimed compounds confirmed that porphyrin derivatives 1–3 sublime
without fragmentation or decomposition.
[15] 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.
STM experiments: All experiments were performed in a two-chamber
UHV system (base pressure of 1”10À10 mbar). As a substrate for the mo-
lecular films, a (111)-oriented Cu single-crystal was used, which was
cleaned by cycles of sputtering with Ar+ ions and subsequent annealing
at 800 K. By this procedure, flat terraces of about 100 nm in width sepa-
rated by monoatomic steps are obtained. The molecular compounds were
deposited by thermal evaporation from a commercial Knudsen-cell-type
evaporator[59] onto the Cu substrate held at room temperature. In this
study, the surface coverage is defined by the fraction of the surface area
occupied by the deposited molecules. This can be directly determined
from STM images due to the lack of a coexisting phase or a gas phase.
All STM measurements were done at a sample temperature between 77
and 200 K. Temperatures >77 K were reached by PID controlled coun-
ter-heating of the sample.
[16] a) M. Ruiz-OsØs, N. Gonzalez-Lakunza, I. Silanes, A. Gourdon, A.
Arnau, J. E. Ortega, J. Phys. Chem. B 2006, 110, 25573; b) M. E.
CaÇas-Ventura, W. Xiao, D. Wasserfallen, K. Müllen, H. Brune, J. V.
Barth, R. Fasel, Angew. Chem. 2007, 119, 1846; Angew. Chem. Int.
Ed. 2007, 46, 1814; c) J. A. Theobald, N. S. Oxtoby, M. A. Phillips,
N. R. Champness, P. H. Beton, Nature 2003, 424, 1029.
[17] H. Spillmann, A. Kiebele, M. Stçhr, T. A. Jung, D. Bonifazi, F. Y.
Cheng, F. Diederich, Adv. Mater. 2006, 18, 275.
[18] a) S. Stepanow, M. Lingenfelder, A. Dmitriev, H. Spillmann, E.
Delvigne, N. Lin, X. B. Deng, C. Z. Cai, J. V. Barth, K. Kern, Nat.
Mater. 2004, 3, 229; b) A. Dmitriev, H. Spillmann, N. Lin, J. V.
Barth, K. Kern, Angew. Chem. 2003, 115, 2774; Angew. Chem. Int.
Ed. 2003, 42, 2670.
[19] a) S. Buchholz, J. P. Rabe, J. Vac. Sci. Technol. B 1991, 9, 1126; b) S.
De Feyter, F. De Schryver, Top. Curr. Chem. 2005, 258, 205.
[20] A. Kiebele, D. Bonifazi, F. Y. Cheng, M. Stçhr, F. Diederich, T.
Jung, H. Spillmann, ChemPhysChem 2006, 7, 1462.
[21] M. Stçhr, M. Wahl, H. Spillmann, L. H. Gade, T. A. Jung, Small
2007, 3, 1336.
[22] N. Wintjes, D. Bonifazi, F. Cheng, A. Kiebele, M. Stçhr, T. Jung, H.
Spillmann, F. Diederich, Angew. Chem. 2007, 119, 4167; Angew.
Chem. Int. Ed. 2007, 46, 4089.
[23] S. Stepanow, N. Lin, D. Payer, U. Schlickum, F. Klappenberger, G.
Zoppellaro, M. Ruben, H. Brune, J. V. Barth, K. Kern, Angew.
Chem. 2007, 119, 724; Angew. Chem. Int. Ed. 2007, 46, 710.
[24] a) T. A. Jung, R. R. Schlittler, J. K. Gimzewski, H. Tang, C. Joachim,
Science 1996, 271, 181; b) F. Moresco, G. Meyer, K. H. Rieder, H.
Tang, A. Gourdon, C. Joachim, Appl. Phys. Lett. 2001, 78, 306; c) C.
Loppacher, M. Guggisberg, O. Pfeiffer, E. Meyer, M. Bammerlin, R.
Acknowledgements
This work was supported by the European Union through the Marie-
Curie Research Training Network PRAIRIES, contract MRTN-CT-2006-
035810, the Swiss National Science Foundation, and the NCCR “Nano-
scale Science”. We also thank the Swiss Federal Commission for Technol-
ogy and Innovation, KTI, and Nanonis for the fruitful collaboration on
the data acquisition system.
[1] J. V. Barth, Annu. Rev. Phys. Chem. 2007, 58, 375.
Chem. Eur. J. 2008, 14, 5794 – 5802
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5801

N. Wintjes, J. Hornung, F. Diederich, T. A. Jung et al.
Lüthi, R. Schlittler, J. K. Gimzewski, H. Tang, C. Joachim, Phys.
Rev. Lett. 2003, 90, 066107; d) F. Moresco, G. Meyer, K. H. Rieder,
H. Ping, H. Tang, C. Joachim, Surf. Sci. 2002, 499, 94.
[41] The model is derived from Spartan calculations (Spartanꢁ06, Wave-
function, Irvine, CA (USA)). Approximate arrangements of the two
interacting porphyrins were taken from the surface data and simu-
lated in the gas phase by using both PM3 and AM1 methods.
[42] Y. Okuno, T. Yokoyama, S. Yokoyama, T. Kamikado, S. Mashiko, J.
Am. Chem. Soc. 2002, 124, 7218.
[43] M. Itoh, M. Takamatsu, N. Kizu, Y. Fujiwara, J. Phys. Chem. 1991,
95, 9682.
[44] T. Michinobu, C. Boudon, J.-P. Gisselbrecht, P. Seiler, B. Frank,
N. N. P. Moonen, M. Gross, F. Diederich, Chem. Eur. J. 2006, 12,
1889.
[25] T. A. Jung, R. R. Schlittler, J. K. Gimzewski, Nature 1997, 386, 696.
[26] a) S. Lee, A. B. Mallik, D. C. Fredrickson, Cryst. Growth Des. 2004,
4, 279; b) R. Paulini, K. Müller, F. Diederich, Angew. Chem. 2005,
117, 1820; Angew. Chem. Int. Ed. 2005, 44, 1788.
[27] a) G. R. Desiraju, Angew. Chem. 1995, 107, 2541; Angew. Chem. Int.
Ed. Engl. 1995, 34, 2311; b) V. R. Thalladi, B. S. Goud, V. J. Hoy,
F. H. Allen, J. A. K. Howard, G. R. Desiraju, Chem. Commun. 1996,
401.
[28] For the application of the concept of supramolecular synthons to
self-assemblies on surfaces, see: L. Scudiero, K. W. Hipps, D. E.
Barlow, J. Phys. Chem. B 2003, 107, 2903.
[45] F. Tao, S. L. Bernasek, J. Am. Chem. Soc. 2005, 127, 12750.
[46] M. J. Crossley, L. D. Field, A. J. Forster, M. M. Harding, S. Sternhell,
J. Am. Chem. Soc. 1987, 109, 341.
[29] a) M. S. K. Dhurjati, J. Sarma, G. R. Desiraju, J. Chem. Soc. Chem.
Commun. 1991, 1702; b) D. S. Reddy, B. S. Goud, K. Panneerselvam,
G. R. Desiraju, J. Chem. Soc. Chem. Commun. 1993, 663; c) P. M.
Ivanov, J. Mol. Struct. 1998, 440, 121.
[30] T. Yokoyama, S. Yokoyama, T. Kamikado, Y. Okuno, S. Mashiko,
Nature 2001, 413, 619.
[31] a) D. Vollhardt, J. Phys. Chem. C 2007, 111, 6805; b) F. Charra, J.
Cousty, Phys. Rev. Lett. 1998, 80, 1682.
[32] M. I. Boamfa, M. W. Kim, J. C. Maan, T. Rasing, Nature 2003, 421,
149.
[33] S. Weigelt, C. Busse, L. Petersen, E. Rauls, B. Hammer, K. V. Goth-
elf, F. Besenbacher, T. R. Linderoth, Nat. Mater. 2006, 5, 112.
[34] T. Yokoyama, S. Yokoyama, T. Kamikado, S. Mashiko, J. Chem.
Phys. 2001, 115, 3814.
[35] M. Schmid, H. Stadler, P. Varga, Phys. Rev. Lett. 1993, 70, 1441; L.
Bartels, G. Meyer, K. H. Rieder, Surf. Sci. 1999, 432, L621.
[36] We show an image of the network formed by porphyrin 2 here
[47] M. Bçhringer, W. D. Schneider, R. Berndt, Angew. Chem. 2000, 112,
821; Angew. Chem. Int. Ed. 2000, 39, 792.
[48] For secondary electrostatic interactions in hydrogen-bonding arrays,
see: T. J. Murray, S. C. Zimmerman, J. Am. Chem. Soc. 1992, 114,
4010.
[49] Y. Kçysal, S. Isy´k, N. Akdemir, E. Agar, C. Kantar, Acta Crystallogr.
E 2004, 60, O930.
[50] a) D. Britton, J. Chem. Crystallogr. 1991, 2127, 405; b) U. Drück, A.
Kutoglu, Acta Crystallogr. C 1983, 39, 638.
[51] S. De Feyter, F. C. De Schryver, Chem. Soc. Rev. 2003, 32, 139.
[52] G. M. Whitesides, J. P. Mathias, C. T. Seto, Science 1991, 254, 1312.
[53] L. Wang, D. Qi, L. Liu, S. Chen, X. Gao, A. T. S. Wee, J. Phys.
Chem. C 2007, 111, 3454.
[54] L. Gross, K. H. Rieder, A. Gourdon, C. Joachim, F. Moresco, Chem-
PhysChem 2007, 8, 245.
[55] L. Merz, H. J. Güntherodt, L. J. Scherer, E. C. Constable, C. E.
Housecroft, M. Neuburger, B. A. Hermann, Chem. Eur. J. 2005, 11,
2307.
ACHTREUNG
ACHTREUNG
[56] K. K. Frederick, M. S. Marlow, K. G. Valentine, A. J. Wand, Nature
2007, 448, 325; R. S. Spolar, M. T. Record, Science 1994, 263, 777.
[57] J. M. Sturtevant, Proc. Natl. Acad. Sci. USA 1977, 74, 2236.
[58] a) C. H. Lee, J. S. Lindsey, Tetrahedron 1994, 50, 11427; b) B. J. Lit-
tler, Y. Ciringh, J. S. Lindsey, J. Org. Chem. 1999, 64, 2864; c) L. H.
Yu, J. S. Lindsey, J. Org. Chem. 2001, 66, 7402.
ACHTREUNG
[37] a) M. Wahl, M. von Arx, T. A. Jung, A. Baiker, J. Phys. Chem. B
2006, 110, 21777; b) S. de Feyter, A. Gesqui›re, P. C. M. Grim, F. C.
de Schryver, Langmuir 1999, 15, 2817.
[38] T. Belser, M. Stçhr, A. Pfaltz, J. Am. Chem. Soc. 2005, 127, 8720.
[39] a) M. T. Cuberes, R. R. Schlittler, J. K. Gimzewski, Appl. Phys. A
1998, 66, S669; b) T. Kamikado, T. Sekiguchi, S. Yokoyama, Y. Wa-
kayama, S. Mashiko, Thin Solid Films 2006, 499, 329.
[59] See www.kentax.de.
[40] S. Berner, M. Brunner, L. Ramoino, H. Suzuki, H. J. Güntherodt,
T. A. Jung, Chem. Phys. Lett. 2001, 348, 175.
Received: April 18, 2008
Published online: May 30, 2008
5802
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 5794 – 5802