Functionalized supramolecular nanoporous arrays for surface templating

Article abstract of DOI:10.1002/chem.200800476

Controlled self-assembly and chemical tailoring of bimolecular networks on surfaces is demonstrated using structural derivatives of 3,4:9,10- perylenetetracarboxylic diimide (PTCDI) combined with melamine (1,3,5-triazine-2,4,6-triamine). Two functionalised PTCDI derivatives have been synthesised, Br₂-PTCDI and di(propylthio)-PTCDI, through attachment of chemical side groups to the perylene core. Self-assembled structures formed by these molecules on a Ag-Si(111) √3 × √3R30° surface were studied with a room-temperature scanning tunneling microscope under ultra-high vacuum conditions. It is shown that the introduction of side groups can have a significant effect upon both the structures formed, notably in the case of di(propylthio)-PTCDI which forms a previously unreported unimolecular hexagonal arrangement, and their entrapment behaviour. These results demonstrate a new route of functionalisation for network pores, opening up the possibility of designing nanostructured surface structures with chemical selectivity and applications in nanostructure templating.

Full text of DOI:10.1002/chem.200800476

DOI: 10.1002/chem.200800476
Functionalized Supramolecular NanoporousArraysfor Surface Templating
LuísM. A. Perdig¼o,[a] Alex Saywell,[a] Giselle N. Fontes,[a, b] Paul A. Staniec,[a]
Gudrun Goretzki,[c] Anna G. Phillips,[c] Neil R. Champness,*[c] and Peter H. Beton*[a]
Abstract: Controlled self-assembly and
chemical tailoring of bimolecular net-
works on surfaces is demonstrated
using structural derivatives of 3,4:9,10-
tures formed by these molecules on a
Ag–SiACTHRE(UNG 111) 3” 3R308 surface were
studied with a room-temperature scan-
ning tunneling microscope under ultra-
high vacuum conditions. It is shown
that the introduction of side groups can
have a significant effect upon both the
structures formed, notably in the case
of di(propylthio)–PTCDI which forms
a previously unreported unimolecular
hexagonal arrangement, and their en-
trapment behaviour. These results
demonstrate a new route of functional-
isation for network pores, opening up
the possibility of designing nanostruc-
tured surface structures with chemical
selectivity and applications in nano-
structure templating.
pffiffi pffiffiffi
perylenetetracarboxylic
(PTCDI) combined with melamine
(1,3,5-triazine-2,4,6-triamine). Two
diimide
functionalised PTCDI derivatives have
been synthesised, Br2–PTCDI and di-
ACHTREUNG(propylthio)–PTCDI, through attach-
ment of chemical side groups to the
perylene core. Self-assembled struc-
Keywords: hydrogen bonds · mela-
mine · perylenetetracarboxylic di-
imides · scanning probe microscopy ·
self-assembly
Introduction
molecular chemistry offers not only the possibility to control
self-assembly processes, but also, through chemical function-
alisation of the constituent species, the potential for molecu-
lar recognition by the self-assembled architecture.[4] Thus,
the utilisation of supramolecular chemistry offers a highly
flexible methodology for control of nanoscale structure.
Two-dimensional self-assembly of molecules on surfaces,
when combined with scanning probe techniques, provides
direct evidence of the great potential of this approach. In
particular, self-assembled structures that are stabilised by in-
termolecular interactions such as hydrogen bonding, metal
coordination and dipolar coupling have been demonstrated
by using a variety of molecule–substrate systems.[5–23] Fur-
thermore, we have recently demonstrated the formation of
two-dimensional bimolecular cavities stabilised by hydrogen
bonding that may be used to trap diffusing species[15–17] as
guest molecules. In related work Stepanow et al.[20] have
demonstrated that the host–guest interaction may be modi-
fied through the inclusion of additional chemical groups
within the framework molecules. Here we explore the possi-
bility of adding host–guest selectivity to these pores through
the rational design of the constituent molecules that form
the network. We demonstrate that chemical functionality be
introduced to such systems and also that functionalisation of
the nanoscale architectures can lead to new self-assembled
structures and the resultant host–guest properties of the
nanoarchitecture can be tuned.
The bottom-up approach to nanostructure fabrication offers
new routes to the systematic preparation of nanostructures
by design. One particularly promising approach is to utilise
the principles and techniques embodied in the field of
supramolecular chemistry.[1,2] In particular, the use of supra-
molecular chemistry has demonstrated that a high degree of
control can be exerted over molecular orientation by the
use of non-covalent intermolecular interactions.[3] Supra-
[a] Dr. L.M. A. Perdig¼o, A. Saywell, Dr. G. N. Fontes, Dr. P. A. Staniec,
Prof. P. H. Beton
School of Physics and Astronomy
University of Nottingham, University Park
Nottingham NG7 2RD (UK)
Fax: ( +44)115-951-5180
E-mail: Peter.Beton@nottingham.ac.uk
[b] Dr. G. N. Fontes
Departamento de Física, ICEx
Universidade Federal de Minas Gerais, Ave. Antonio Carlos
6627-Belo Horizonte, CEP 30123-970 (Brazil)
[c] Dr. G. Goretzki, A. G. Phillips, Prof. N. R. Champness
School of Chemistry, University of Nottingham
University Park, Nottingham NG7 2RD (UK)
Fax: ( +44)115-951-3563
E-mail: Neil.Champness@nottingham.ac.uk
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200800476.
7600
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 7600 – 7607

FULL PAPER
Specifically we show that bimolecular honeycomb arrays
can also be formed if the 3,4:9,10-perylenetetracarboxylic
diimide (PTCDI) from the known hexagonal PTCDI/mel-
This type of system combines the templating effect that has
been demonstrated previously for such hexagonal nanoarch-
itectures with the possibility of molecular recognition driven
by the functional group added to the original PTCDI mole-
cule. We have studied two variants of R2–PTCDI, firstly
with R being bromine, (Scheme 1, bottom left), and the
other with propylthio substituents (Scheme 1, bottom right).
ACHTREUNG
amine arrangement[15,16] (melamine=1,3,5-triazine-2,4,6-tri-
ACHTREUNGamine) is replaced with similar molecules that are function-
alised through the attachment of a side group to the pery-
lene core, generically called here R2–PTCDI (Scheme 1,
top). The synthesis of perylene diimide derivatives has been
extensively examined allowing the development of such
molecules for a wide variety of applications;[24] however,
only a few examples of derivatised perylene diimides that
maintain the non-alkylated imide group have been previous-
ly described.[25] The diimide termination, present in the orig-
inal PTCDI molecule and which is essential for the forma-
tion of the triple hydrogen bond with melamine and respon-
sible for the stabilisation of the hexagonal arrangement ob-
served in the mixtures, is preserved in our structural design.
Although each of these molecules forms characteristic as-
pffiffi pffiffiffi
À
semblies on the Ag Si
A
from the arrangements of the parent compound PTCDI on
the same surface. Whereas Br2–PTCDI forms a nanoporous
hexagonal arrangement by co-self-assembly with melamine,
di(propylthio)–PTCDI forms a unimolecular self-assembled
nanoporous hexagonal arrangement, the periodicity of
which can be increased by the introduction of melamine and
subsequent bimolecular structure self-assembly.
Results and Discussion
Chemical synthesis: Br2–PTCDI was prepared by bromina-
tion of the 3,4:9,10-perylenetetracarboxylic dianhydride
(PTCDA) to afford a mixture of isomers Br2–PTCDA.[26]
The preparation of Br2–PTCDA yields a mixture of both
1,7- and 1,6-isomers in an estimated 4:1 ratio that it has not
proven possible to separate.[26b] Subsequent reaction of Br2–
PTCDA with NH4OH(aq) affords Br2–PTCDI as a highly in-
soluble red solid. Perylene diimide species, without alkyl/
aryl tail groups, are notoriously insoluble in common organ-
ic solvents and thus separation of the 1,6- and 1,7-isomers is
not feasible at this stage of the synthetic procedure. Di(pro-
pylthio)–PTCDI was prepared by two alternative routes
(Scheme 2). Introduction of the propanethio appendages is
successfully achieved by the reaction of propanethiol with
either N,N’-diACTHRE(UNG n-butyl)-1,7-dibromoperylene-3,4:9,10-tetra-
carboxylic acid diimide or N,N’-di(2,4-dimethoxy)-benzyl)-
1,7-dibromoperylene-3,4:9,10-tetracarboxylic acid diimide in
the presence of a base to give compound 1 or 4, respectively.
Compound 1 can be converted to the corresponding disub-
stituted dithioether–PTCDI by using the traditional method-
ology[26b] of converting the dialkyl perylene diimide to the
corresponding dianhydride 2 followed by subsequent con-
version to diimide 3. This synthetic procedure successfully
produces the desired product, but in a yield of only 45%
from 1 and is a relatively aggressive synthetic strategy, using
a 100-fold excess of KOH as a reagent. Thus, we have estab-
lished a new procedure for the synthesis of PTCDI mole-
cules derivatised in the so-called bay region. By using 2,4-di-
methoxybenzyl alkylated intermediates (e.g., 4) in the syn-
thetic strategy we are able to convert the 2,4-dimethoxyben-
zyl-protected perylene species to the desired PTCDI prod-
uct by a simple deprotection step, using methanesulfonic
acid, without the need to prepare an intermediary anhydride
species. This mild approach, in comparison to the conven-
tional procedure, affords the desired species 3 in 89% from
the dialkylated species 4. A similar approach of protection/
deprotection for the formation of diimide PTCDI deriva-
Scheme 1. Top: Target hexagonal arrangement with R2–PTCDI and mela-
mine stabilised by a triple hydrogen bond. The functionalisation of the
cavities is achieved through the attachment of chemical groups (repre-
sented with R and highlighted with grey circles) onto the sides of the per-
ylene core of the PTCDI molecule. Bottom left and right: Br2–PTCDI
and di(propylthio)–PTCDI, respectively.
Chem. Eur. J. 2008, 14, 7600 – 7607
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7601

N. R. Champness, P. H. Beton et al.
pffiffiffi pffiffiffi
À
Figure 1. a) STM image of the Ag Si
A
Br2–PTCDI sublimation (Vsample =À2 V, Itunnel =0.05 nA), in which the for-
mation of elongated rows following well-defined directions is evident.
b) Higher resolution image of an area with different Br2–PTCDI row
widths (Vsample =À2 V, Itunnel =0.05 nA). A single molecular row is high-
lighted with a white dashed rectangle. c) Schematic model for the single
molecular row highlighted (substrate reconstruction is represented with
the hexagonal grid, with the centres corresponding to silicon trimers) in
b), with each molecule forming a double hydrogen bond with neighbour-
Scheme 2. Reaction scheme illustrating synthetic strategy to PTDCI-de-
rivatives: i) KOH, iPrOH; ii) CH3COOH, 95%; iii) NH4OH, 47%; iv)
MeSO3H, C6H5Me, 89%.
tives has recently been reported by Würthner et al. using
the cleavage of N-methylbenzyl derivatised PTCDI-based
species using BBr3.[25a] Unfortunately we were unable to sep-
arate the 1,6- and 1,7-isomers of either 1 or 4, although the
1,7-isomer is the major component with trace amounts of
the 1,6-isomer present in either material and thus the final
product 3 is predominantly the 1,7-isomer with only traces
of the 1,6-isomer present.
¯
ing molecules and the row parallel to the substrate h112i direction. The
dimension of each STM image is given in the inserted scale bar.
lised by a double hydrogen bond between neighbouring
molecules, linked through the imide terminations.[28–31]
The arrangement proposed for single Br2–PTCDI chains
is similar to that observed for single rows of PTCDI on the
same substrate.[28] However, for PTCDI the rows are much
shorter and co-exist with large two-dimensional islands. For
Br2–PTCDI many of the molecular rows have widths corre-
sponding to one or two molecules, but we do not observe
extended two-dimensional islands (for an analysis of the dis-
tribution of row widths see the Supporting Information).
Surface studies
Br2–PTCDI: Following deposition of Br2–PTCDI onto the
À
Ag Si surface, one-dimensional molecular rows are formed
as shown in the STM images in Figure 1a. A higher resolu-
À
tion STM image is shown in Figure 1b in which the Ag Si
reconstruction is visible with darker contrast regions corre-
sponding to a silicon trimer.[27] In common with PTCDA
and PTCDI[16,28] on this surface, the Br2–PTCDI molecules
have intramolecular contrast corresponding to a double lobe
that runs parallel to the long axis of the molecules. From
this identification it is possible to determine the molecular
position and orientation relative to the substrate. The self-
À
Thus the addition of the Br appendage to the PTCDI has
resulted in an enhanced stability of one-dimensional rows
relative to extended islands. In recent numerical studies of
anisotropic growth[32,33] it has been demonstrated that such
an enhancement can result either from an increase in the
hydrogen-bonding interaction or a reduction in the intermo-
lecular interactions that stabilise row–row coupling in island
¯
assembled rows adopt an orientation parallel to the h112i di-
À
rection of the substrate with a measured molecule–molecule
spacing along the row of (14.8ꢀ0.2) Š. Individual molecules
within a row can be resolved, see white dashed rectangle in
Figure 1b, and from these high-resolution images we pro-
pose a model for a row with a width corresponding to a
single molecule. This is shown in Figure 1c, which is stabi-
formation. In the case studied here the addition of the Br
appendage has a negligible effect of the strength of hydro-
gen bonding. However, the presence of this electronegative
species on the perylene core is expected to give rise to a re-
À
pulsive Br Br intermolecular interactions that leads to a re-
duction in inter-row interactions. This results in a relative
7602
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 7600 – 7607

Nanoporous Arrays
FULL PAPER
destabilisation of two-dimensional growth in favour of row
formation.
The deposition of melamine onto a sub-monolayer of
Br2–PTCDI coverage, followed by annealing at ꢁ608C, re-
sults in the formation of a bimolecular honeycomb arrange-
ment as shown in the STM image in Figure 2a. A magnified
image of the network with detail of the surface reconstruc-
tion is shown in Figure 2b. The periodicity of the network is
ꢁ35 Š, and the network lattice vectors are parallel to the
¯
substrate h110i direction. These dimensions are identical,
within experimental error, to those previously reported for a
PTCDI/melamine network.[16] This shows clearly that the
observed arrangement is a direct analogue to the previously
reported network in which Br2–PTCDI molecules adopt the
hexagonal structure illustrated in Scheme 1, top (see super-
imposed molecular model in Figure 2b).
One difference with our previous work is that we observe
a complete conversion of the Br2–PTCDI into the hexagonal
framework. For unfunctionalised PTCDI we observe a co-
existence between the bimolecular hexagonal framework
and two-dimensional PTCDI islands. The complete conver-
sion of the brominated derivative, accounted for by the de-
stabilisation of two-dimensional Br2–PTCDI islands, is po-
tentially significant for the formation of large-area, single-
phase networks and offers the possibility of much greater
homogeneity of entrapment properties. Furthermore this
difference confirms that the properties of supramolecular
arrays may be engineered through relatively minor changes
in the constituent molecules. Although the bimolecular Br2–
PTCDI···melamine honeycomb network can in principle ex-
hibit chirality (see Figure 2b) we were unable to determine
the precise positioning of individual bromine substituents
and therefore are unable to establish the precise extent of
chirality exhibited by the honeycomb framework.
Most importantly these results confirm the hypothesis
that the formation of bimolecular hexagonal assemblies be-
tween R2–PTCDI and melamine is possible. The use of Br2–
PTCDI provides the first demonstration of a route to func-
tionalise the cavities of the PTCDI/melamine regular hexag-
onal arrangement. It is worth noting that Br2–PTCDI is a
synthetic precursor to a range of R2–PTCDI molecules[25]
and as such the Br2–PTCDI/melamine array offers the po-
tential for further chemical functionalisation.
The C60-entrapment properties of this network were com-
pared with previous results on PTCDI/melamine assem-
blies.[15,16] The STM images (Figure 2c) show the formation
of C60 heptamers within the pores as observed when C60 is
deposited on the PTCDI/melamine porous network. This in-
dicates that both Br2–PTCDI/melamine and PTCDI/mela-
mine networks stabilise C60 clusters with similar geometries,
most likely as a result of the relatively short steric bulk of
the Br appendages. Interestingly we have observed that the
heptamers captured in the brominated cavities, as compared
with the parent PTCDI/melamine array, are more suscepti-
ble to interactions with the tip of the scanning probe micro-
scope providing indirect evidence that they are less strongly
bound.
Figure 2. a) Image showing hexagonal supermolecular structures formed
pffiffi pffiffi
À
with Br2–PTCDI and melamine on the Ag Si
(Vsample =+2.1 V, Itunnel =0.05 nA). b) Higher resolution image (Vsample
A
À2 V, Itunnel =0.05 nA) of the Br2–PTCDI/melamine hexagonal network,
in which the double protrusion of each Br2–PTCDI molecule is apparent.
The hexagonal mesh represents the surface reconstruction with the hexa-
¯
gon centres corresponding to silicon trimers (see text). One of the h112i
surface directions is represented with the white arrow. A molecular
model is superimposed for clarification. c) Image (Vsample =À2.8 V, Itunnel
=
0.03 nA) showing three captured C60 heptamers within the cavities of the
Br2–PTCDI/melamine network. d) Molecular model of a C60 heptamer
inside a Br2–PTCDI/melamine pore. The dimension of each STM image
is given in the inserted scale bar.
Di(propylthio)–PTCDI: Deposition of 0.4 monolayers of
À
di(propylthio)–PTCDI onto a clean Ag Si surface results in
Chem. Eur. J. 2008, 14, 7600 – 7607
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7603

N. R. Champness, P. H. Beton et al.
the formation of two-dimensional self-assembled structures,
in contrast to Br2–PTCDI. The STM image shown in Fig-
ure 3a reveals two co-existing domains labelled as a and b.
Whereas in the b region the molecules form a close-packed
array, those in the a region adopt a very interesting hexago-
nal arrangement (shown in more detail in Figure 3b; see
Supporting Information for discussion of region b). The hex-
agonal structure has a periodicity of 28.3ꢀ1 Š, and an angle
¯
of 16ꢀ28 relative to the h112i substrate direction. The mea-
sured periodicity is significantly smaller than the dimensions
observed for the PTCDI/melamine complex(Scheme 1,
top),[15,16] and, moreover, cannot arise from an arrangement
in which the molecules are linked by double hydrogen
bonds, as observed for Br2–PTCDI.
We propose a model, shown in Figure 3c in which each
vertexis formed by a junction of three molecules and stabi-
À
lised by three hydrogen bonds between the N H groups on
one molecule and one of the C=O groups on a neighbouring
molecule. This triple vertexbonding configuration is similar
to the stabilising linkage proposed for cyanuric acid on the
same surface[34] and is chiral.
It is clear that PTCDI, even in the absence of the side
groups, could adopt this hexagonal configuration for which
the overall number of hydrogen bonds per molecule is the
same as in the rows formed by Br2–PTCDI and PTCDI.
However, the hexagonal arrangement is only observed for
molecules that incorporate the propylthio side groups. We
À
hypothesise that an additional attractive C H···O hydrogen-
bonding interaction may exist between the propylthio chains
and the oxygen atom on a neighbouring molecule that is not
bound to an imide group within the trigonal vertex. The sta-
bilisation energy obtained for the vertexin this configura-
tion, calculated from density functional theory (see Support-
ing Information) is 0.86 eV. For an alternative configuration,
in which the vertexchirality is preserved but the propylthio
chains are attached to the other bay-region carbon atoms
(see arrows in Figure 3c and Supporting Information), a sta-
bilisation energy of 0.70 eV was calculated. Thus, the config-
uration shown in Figure 3c, with the propylthio chains closer
to the available oxygen atom, is calculated to be more
stable. The stabilisation energy calculated for a vertexcom-
posed with three PTCDI molecules (with no side chains) is
0.78 eV, lower than the energy obtained for the suggested
vertex, further demonstrating the contribution of the side
chains in stabilising the structure.
Figure 3. STM image (Vsample =À2.5 V, Itunnel =0.03 nA) after sublimation
pffiffiffi pffiffiffi
À
of di(propylthio)–PTCDI onto a Ag Si
(111) 3” 3R308 surface. Two
distinct domains were labelled with a for the hexagonal and b for the
close-packed arrangement. b) Magnified image (Vsample =À2.5 V, Itunnel
=
0.03 nA) of a hexagonal region (a), in which a trimer node is highlighted
with a dashed white circle and the surface reconstruction represented
with a hexagonal mesh. c) Molecular model proposed for the trimer
vertexhighlighted in the previous image. The hydrogen atoms highlight-
ed with grey circles and with arrows are alternative attachment sites for
the propylthio chains as discussed in text. d) Schematic model proposed
The additional stabilisation provided by the side chains
implies that they must adopt a specific conformation. Note
that, due to their flexibility arising from the potential for ro-
À
tation about the C C bond axes, there are several different
for the (10,4,3) molecular registry with the substrate. The arrow repre-
pffiffiffiffiffiffiffi
conformations which the molecular side chains might adopt.
In this context the stabilisation may be considered as an
adaptive effect in which the molecule adopts a specific con-
formation in response to its local environment. The link be-
tween flexibility of molecules and their potential for inter-
locking through adopting mutually energetically favourable
conformations has recently been highlighted for adsorbed
amino acids.[35]
sents the periodicity of the arrangement of 156aAg (three hexagonal
À
Si
spacings of the molecular network), and the dashed lines detail each to
the 10aAg and 4aAg periods of the surface. q is defined as the angle
À
Si
À
Si
between the pore–pore direction of the molecular network and one of
pffiffi pffiffi
¯
À
the h112i directions of the reconstructed Ag Si
ACHTRE(UNG 111) 3” 3R308 surface.
The arrangement is composed with trimers identical to that shown in
part c), but with opposite chirality for information. The dimension of
each STM image is given in the inserted scale bar.
7604
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 7600 – 7607

Nanoporous Arrays
FULL PAPER
The hexagonal arrangement is composed of an array of
these trigonal vertices as shown schematically in Figure 3d.
Based on the measured angles and periodicity of the ar-
rangement, we propose a model that is commensurate with
the substrate, with indices (l,m,n) (notation used in referen-
ces [27a–c]) for hexagonal molecular arrangements on hex-
agonally reconstructed surfaces of (3,10,4), and a periodicity
pffiffiffiffiffiffiffi
of 156aAg-Si =82.9 Š (see Figure 3d), corresponding to a
pore-to-pore distance of 27.6 Š. According to this model the
pore-to-pore direction forms an angle of ꢀ16.1 8 with the
¯
substrate h112i direction. These values are in good agree-
ment with those observed experimentally.
There are two possible angles (positive or negative) that
this molecular arrangement may adopt relative to the sub-
strate while observing the commensurability indicated with
the indices above. In the trimer configuration given in Fig-
ure 3d (positive angle), the direction of the end-to-end axis
¯
of the molecules is parallel to the h110i direction, whereas
the negative angle and opposite chirality (Figure 3c) leads to
¯
molecules being parallel to the h112i substrate direction.
Our STM images reveal that we only observe one molecular
orientation relative to the substrate which corresponds to
the long axis of the molecule (intersecting the imide groups)
¯
parallel to the h110i direction, as represented in Figure 3d.
We also note that in the proposed model the molecules are
all centred above high-symmetry positions (either Si or Ag
trimers) in the reconstructed surface. This structure offers
an alternative route to the formation of hexagonal porous
networks, such as PTCDI/melamine and Br2–PTCDI/mela-
mine, but with a smaller pore size and using a single mole-
cule species.
Figure 4a shows STM images obtained after deposition of
melamine on the sample discussed above (deposition time
ꢁ4 h with sample held at 60–808C). These images show the
formation of hexagonal networks with periodicity and orien-
tation relative to the substrate that is identical to the
PTCDI/melamine and Br2–PTCDI/melamine bimolecular
structures (Figure 4b). This indicates that the unimolecular
di(propylthio)–PTCDI hexagonal array has been converted
into a bimolecular di(propylthio)–PTCDI/melamine hexago-
nal phase. The changes in network dimension through addi-
tion of melamine demonstrate an interesting, and we believe
first demonstration of, sequential development of self-as-
sembled structures with increasing porosity.
Deposition of C60 onto the hexagonal network, Figure 4c
gives rise to a very different entrapment behaviour relative
to the Br2–PTCDI/melamine and PTCDI/melamine net-
works;[16] we see no evidence for heptamers or hexamers
within the pores. It is possible to identify some fullerenes
over some of the constituent molecules of the network as
well as within the pores, but in an irregular fashion. A close
inspection of the images shows that the underlying hexago-
nal structure is preserved demonstrating that the presence
of the propylthio chains in the cavities inhibits the entrap-
ment of fullerene molecules. This is most likely due to a
steric hindering effect of the C60–surface interaction as a
Figure 4. a) STM image (Vsample =À2.5 V, Itunnel =0.03 nA) showing the bi-
molecular hexagonal network formed after sublimation of melamine
onto adsorbed di(propylthio)–PTCDI. b) Molecular model of the ar-
rangement highlighted in (a). c) STM image (Vsample =À2.3 V, Itunnel
=
0.03 nA) after deposition of C60 onto the bimolecular honeycomb net-
work. The dimension of each STM image is given in the inserted scale
bar.
result of the presence of the propylthio chains providing a
partial coverage of alkane chains within the pore.
Conclusion
We have demonstrated that the self-assembly and entrap-
ment properties of surface-based supramolecular arrays may
be controlled through the addition of simple functional
groups to molecular building blocks. The incorporation of
bromine in the bay region of PTCDI leads to a reduction in
inter-row coupling and a preference for one-dimensional
growth, as opposed to the formation of close-packed two-di-
mensional islands. The self-assembly of a unimolecular hex-
agonal arrangement in di(propylthio)–PTCDI is particularly
noteworthy, illustrating the stabilising effect on introducing
short-chain side groups to the PTCDI scaffold. Moreover
we have demonstrated that the chemical functionalisation of
the bay region of PTCDI provides a route for the rational
introduction of chemically active groups on the perimeter of
the pores formed in a two-dimensional supramolecular net-
work. The modification of the unimolecular di(propylthio)–
PTCDI hexagonal array to the bimolecular di(propylthio)–
PTCDI/melamine array, resulting in a stepwise increase in
Chem. Eur. J. 2008, 14, 7600 – 7607
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7605

N. R. Champness, P. H. Beton et al.
procedure was used as for the synthesis of 1, but with N,N’-di[(2,4-di-
network periodicity, represents an example of dimensional
control of porous-surface-based self-assembled structures
through the exploitation of an additional supramolecular
synthon. The contrasting entrapment behaviour observed
with the di(propylthio)–PTCDI/melamine hexagonal net-
works relative to PTCDI/melamine and Br2–PTCDI/mela-
mine systems demonstrates the potential for control,
through selection of side chains, of the properties of these
cavities. Since Br2–PTCDI is the precursor to many other
functionalised R2–PTCDI derivatives, with potentially more
reactive attached groups, our work represents important
progress towards the objective of combining chemically se-
lectivity with spatial organisation within two dimensional
nanoporous networks.
AHCTREmGNU ethoxy)benzyl]-1,7-dibromoperylene-3,4:9,10-tetracarboxylic acid di-
1
imide as a starting material. H NMR (CDCl3): d=8.87 (d, J=8 Hz, 2H),
8.81 (s, 2H), 8.73(d, J=8 Hz, 2H), 7.12 (d, J=9 Hz, 2H), 6.50 (s, 2H),
6.39 (d, J=9 Hz, 2H), 5.43 (s, 2H), 3.90, 3.43 (2s, 6H each), 3.20 (t, J=
7 Hz, 4H; SCH2), 1.72 (m, 4H), 1.05 ppm (m, 6H; CH3); 13C NMR
(CDCl3): d=163.84, 163.77, 160.47, 158.61, 140.31, 138.96, 133.14, 132.95,
131.42, 129.45, 129.34, 128.86, 128.69, 125.84, 122.46, 121.99, 117.77,
104.52, 99.02, 56.02, 55.75, 39.242, 38.43, 22.38, 13.95 ppm; MS (MALDI-
TOF): m/z: 838.8 [M+].
1,7-Dipropylsulfaneperylene-3,4:9,10-tetracarboxydiimide (mixture of
1,7- and 1,6- isomers) (3)
Synthesis from compound 2: Compound 2 (mixture of isomers; 160 mg,
0.3 mmol) was suspended in NH4OH (20 mL) and heated under reflux
overnight. The reaction mixture was cooled to room temperature and
poured into glacial acetic acid in ice. The precipitate was filtered by grav-
ity and dried in an oven overnight to yield a blue powder (75 mg, 47%),
which exhibited very low solubility in common organic solvents.
Synthesis from compound 4: Compound 4 (250 mg, 0.3 mmol) was dis-
solved in toluene (20 mL), and methanesulfonic acid (2 mL) was added.
The reaction mixture was heated under reflux for 4 h, over which time all
perylene-containing material precipitated leaving a clear solution. The
solvent was removed and the residue titurated with water, filtered and
dried to yield 3 as a blue powder. Yield 89%; 1H NMR (CDCl3 + tri-
fluoroacetic acid): d=9.08 (d, J=8 Hz, 2H), 8.88 (s, 2H), 8.81 (d, J=
8 Hz, 2H), 3.27 (t, J=7 Hz, 4H; SCH2), 1.79 (m, 4H), 1.05 ppm (m, 6H;
CH3); MS (MALDI-TOF): m/z: 538.0 [M+].
Experimental Section
Perylene-3,4:9,10-tetracarboxylic acid (Aldrich) was brominated accord-
ing to a literature procedure[26] and converted to N,N’-bis
ACHTREUNG
ACHTREUNG
tanol/H2O.[26] All reactions were carried out under an atmosphere of ni-
trogen. Column chromatography was performed on silica gel (Merck
silica gel 60, 0.2–0.5 mm, 50–130 mesh). The 1H NMR (300 MHz) and
13C NMR (75 MHz) spectra were obtained on a Bruker DPX 300 spec-
trometer. Microanalyses were performed by Stephen Boyer, London
Metropolitan University. MS spectra (MALDI-TOF-MS) were deter-
mined on a Voyager-DE-STR mass spectrometer.
Surface studies: The R2–PTCDI and melamine deposition were per-
formed under ultra-high vacuum (UHV) conditions (base pressure
pffiffi pffiffi
<10À10 Torr). The Ag Si
N
À
À
prepared under UHV by outgassing a 3 mm”7 mm piece of a Si
substrate for >12 h at 600–7008C followed by flash annealing at 1150–
12008C for approximately 60 s. This procedure results in the formation of
N,N’-Di
A
ACHTREUNG
À
a Si
A
of Ag with the sample held at 5008C.[27] The evaporation rates for each
R2–PTCDI were measured with a quartz crystal microbalance and, for
deposition of submonolayer coverages, crucible temperatures of 4088C
for Br2–PTCDI and 3408C for di(propylthio)–PTCDI were used. During
R2–PTCDI deposition, the sample was held at room temperature. Mel-
moperylene-3,4:9,10-tetracarboxylic acid diimide[26] (660 mg, 1 mmol),
NaOH (100 mg, 2.5 mmol) and propanethiol (0.19 mL, 3 mmol) were sus-
pended in pyridine (30 mL) and heated under refluxfor 3 h. The progress
of the reaction was followed by thin-layer chromatography with CH2Cl2
as eluent. After cooling to room temperature, the mixture was poured
into 10% HCl and extracted with CH2Cl2. The organic phase was dried
over Na2SO4, filtered and concentrated. The crude product was pure
enough at this stage for further reaction procedures. For samples for
analysis, the compound can be purified by column chromatography
(silica) with CH2Cl2/hexane (5:1) as eluent allowing partial separation of
the isomers. Yield 93%; 1H NMR (CDCl3): d=8.78 (d, J=8 Hz, 2H),
8.73 (s, 2H), 8.65 (d, J=8 Hz, 2H), 4.24 (t, J=7 Hz, 4H; NCH2), 3.20 (t,
J=7 Hz, 4H; SCH2), 1.73 (m, 8H), 1.46 (m, 4H), 1.05 ppm (m, 12H;
CH3); 13C NMR (CDCl3): d=163.86, 163.75, 138.86, 133.03, 132.87,
131.23, 129.30, 129.25, 128.75, 125.69, 122.37, 121.89, 40.86, 38.30, 31.33,
30.64, 22.38, 20.82, 14.26, 13.99 ppm; MS (MALDI-TOF): m/z: 650.6 [M+
].
AHCTREaUNG mine was evaporated from a cell placed in a separate chamber to avoid
cross-contamination. Images of the surface were acquired using a scan-
ning tunneling microscope (STM) operating in constant current mode at
room temperature. Electrochemically etched tungsten tips, cleaned prior
to use by electron beam heating, were used throughout.
Density functional calculations: Density functional calculations were per-
formed using established methodologies[36] within the DMol3 package.
See Supporting Information for more details.
Dipropylthioperylene-3,4:9,10-tetracarboxydianhydride (mixture of 1,7-
and 1,6- isomers) (2): Compound 1 (650 mg, 1 mmol; mixture of isomers)
was dissolved in iPrOH (20 mL) and finely ground KOH (5.7 g,
100 mmol) was added. The suspension was sonicated for 15 min and sub-
sequently heated under refluxfor 4 h. The progress of the reaction was
followed by thin-layer chromatography with CH2Cl2 as eluent. After
cooling to room temperature the mixture was poured into 50% acetic
acid and stirred for 1 h. The precipitate was filtered, redissolved in
CH2Cl2 and washed with water. The organic phase was dried over
Na2SO4, filtered and concentrated to yield a red powder (525 mg, 95%
Acknowledgements
This work was supported by the UK Engineering and Physical Sciences
Research Council (EPSRC) under Grants Nos. GR/S97521/01 and EP/
D048761/1 and the European Union Framework
MESH” (NMP4-CT-2004-013817). G.N.F. is grateful for the financial sup-
port from CAPES (Brazil).
6 Grant “NANO-
[1] K. Ariga, T. Kunitake, Supramolecular Chemistry—Fundamentals
and Applications, Springer, Berlin, 2006.
1
yield). H NMR (CDCl3): d=8.93 (d, J=8 Hz, 2H), 8.88 (s, 2H), 8.74 (d,
J=8 Hz, 2H), 3.23 (t, J=7 Hz, 4H; SCH2), 1.72 (m, 4H), 1.05 ppm (m,
6H; CH3); 13C NMR (CDCl3): d=160.43, 159.97, 140.62, 134.03, 133.52,
132.96, 131.39, 129.88, 129.37, 127.92, 118.76, 118.45, 38.37, 22.25,
13.91 ppm; MS (MALDI-TOF): m/z: 540.3 [M+].
[2] J.V. Barth, G. Costantini, K. Kern, Nature 2005, 437, 671–679.
[3] G. M. Whitesides, B. Grzybowski, Science 2002, 295, 2418–2421.
[4] a) S. Hasegawa, S. Horike, R. Matsuda, S. Furukawa, K. Mochizuki,
Y. Kinoshita, S. Kitagawa, J. Am. Chem. Soc. 2007, 129, 2607–2614;
b) O. Crespo-Biel, C. W. Lim, B. J. Ravoo, D. N. Reinhoudt, J. Husk-
ens, J. Am. Chem. Soc. 2006, 128, 17024–17032; c) V.A. Azov, A.
N,N’-Di[(2,4-dimethoxy)benzyl]dipropylthioperylene-3,4:9,10-tetracar-
boxylic acid diimide (mixture of 1,7- and 1,6- isomers) (4): An identical
7606
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 7600 – 7607

Nanoporous Arrays
FULL PAPER
Beeby, M. Cacciarini, A. G. Cheetham, F. Diederich, M. Frei, J.K.
Gimzewski, V. Gramlich, B. Hecht, B. Jaun, T. Latychevskaia, A.
Lieb, Y. Lill, F. Marotti, A. Schlegel, R. R. Schlittler, P. J. Skinner, P.
Seiler, Y. Yamakoshi, Adv. Funct. Mater. 2006, 16, 147–156; d) M.
Pittelkow, C. B. Nielsen A. C. Broeren, J. L. J. van Dongen, M. H. P.
van Genderen, E. W. Meijer, J. B. Christensen, Chem. Eur. J. 2005,
11, 5126–5135; e) C. A. Nijhuis, K.A. Dolatowska, B. J. Ravoo, J.
Huskens, D. N. Reinhoudt, Chem. Eur. J. 2007, 13, 69–80; f) T. K.
Maji, R. Matsuda, S. Kitagawa, Nat. Mater. 2007, 6, 142–148;
g) M. P. Schramm, R. J. Hooley, J. Rebek, J. Am. Chem. Soc. 2007,
129, 9773–9779.
Horowitz, F. Kouki, P. Spearman, D. Fichou, C. Nogeus, X. Pan, F.
Garnier, Adv. Mater. 1996, 8, 224; f) C. W. Tang, Appl. Phys. Lett.
1986, 48, 183; g) J. Q. Qu, C. Kohl, M. Pottek, K. Mullen, Angew.
Chem. 2004, 116, 1554; Angew. Chem. Int. Ed. 2004, 43, 1528; h) F.
Würthner, C. Thalacker, S. Diele, C. Tschierske, Chem. Eur. J. 2001,
7, 2245.
[25] a) T. E. Kaiser, H. Wang, V. Stepanenko, F. Würthner, Angew.
Chem. 2007, 119, 5637–5640; Angew. Chem. Int. Ed. 2007, 46, 5541–
5544; b) E. H. A. Beckers, Z. Chen, S. C. J. Meskers, P. Jonkheijm,
A. P. H. J. Schenning, X.-Q. Li, P. Osswald, F. Würthner, R. A. J.
Janssen, J. Phys. Chem. B 2006, 110, 16967–16978; c) Y. Liu, J.
Zhuang, H. Liu, Y. Li, F. Lu, H. Gan, T. Jiu, N. Wang, X. He, D.
Zhu, ChemPhysChem 2004, 5, 1210–1215; d) F. Würthner, Z. Chen,
F. J. M. Hoeben, P. Osswald, C.-C. You, P. Jonkheijm, J. von Herri-
khuyzen, A. P. H. J. Schenning, P. P. A. M. van der Schoot, E. W.
Meijer, E. H. A. Beckers, S. C. J. Meskers, R. A. J. Janssen, J. Am.
Chem. Soc. 2004, 126, 10611–10618; e) Y. Liu, S. Xiao, H. Li, Y. Li,
H. Liu, F. Lu, J. Zhuang, D. Zhu, J. Phys. Chem. B 2004, 108, 6256–
6260; f) A. P. H. J. Schenning, J. van Herrikhuyzen, P. Jonkheijm, Z.
Chen, F. Würthner, E. W. Meijer, J. Am. Chem. Soc. 2002, 124,
10252–10253; g) F. Würthner, C. Thalacker, A. Sautter, W. Schartl,
W. Ibach, O. Hollricher, Chem. Eur. J. 2000, 6, 3871–3886; h) F.
Würthner, C. Thalacker, A. Sautter, Adv. Mater. 1999, 11, 754–758.
[26] a) C.-C. Chao, M.-K. Leung, Y. O. Su, K.-Y. Chiu, T.-H. Lin, S.-J.
Shieh, S.-C. Lin, J. Org. Chem. 2005, 70, 4323–4331; b) F. Würthner,
V. Stepanenko, Z. Chen, C. R. Saha-Mçller, N. Kocher, D. Stalke, J.
Org. Chem. 2004, 69, 7933–7939.
[5] T. Yokoyama, S. Yokoyama, T. Kamikado, Y. Okuno, S. Mashiko,
Nature 2001, 413, 619–621.
[6] J.V. Barth, J. Weckesser, C. Cai, P. Günter, L. Bürgi, O. Jeandu-
peaux, K. Kern, Angew. Chem. 2000, 112, 1285–1288; Angew.
Chem. Int. Ed. 2000, 39, 1230–1234.
[7] H. Spillmann, A. Dmitriev, N. Lin, P. Messina, J.V. Barth, K. Kern,
J. Am. Chem. Soc. 2003, 125, 10725–10728.
[8] S. De Feyter, F. C. De Schryver, Chem. Soc. Rev. 2003, 32, 139–150.
[9] S. Griessl, M. Lackinger, M. Edelwirth, M. Hietschold, W. M. Heckl,
Single Mol. 2002, 3, 25–31; M. Blunt, X. Lin, M. C. Gimenez-Lopez,
M. Schrçder, N. R. Champness, P. H. Beton, Chem. Commun. 2008,
2304–2306.
[10] J. Lu, S.-B. Lei, Q.-D. Zeng, S.-Z. Kang, C. Wang, L.-J Wan, C. Bai,
J. Phys. Chem. B 2004, 108, 5161–5165.
[11] R. Otero, M. Schçck, L. M. Molina, E. Lœgsgaard, I. Stensgaard, B.
Hammer, F. Besenbacher, Angew. Chem. 2005, 117, 2310–2315;
Angew. Chem. Int. Ed. 2005, 44, 2270–2275.
[12] C. Safarowsky, A. Rang, C. A. Schalley, K. Wandelt, P. Broekmann,
Electrochim. Acta 2005, 50, 4257–4268.
[13] M. Stçhr, M. Wahl, C. H. Galka, T. Riehm, T. A. Jung, L. H. Gade,
Angew. Chem. 2005, 117, 7560–7564; Angew. Chem. Int. Ed. 2005,
44, 7394–7398.
[27] a) M. D. Upward, P. Moriarty, P. H. Beton, Phys. Rev. B 1997, 56,
R1704–R1707; b) M. J. Butcher, J. W. Nolan, M. R. C. Hunt, P. H.
Beton, L. Dunsch, P. Kuran, P. Georgi, T. J. S. Dennis, Phys. Rev. B
2001, 64, 195401; c) M. D. Upward, P. H. Beton, P. Moriarty, P. Surf.
Sci. 1999, 441, 21–25; d) T. Takahashi, S. Nakatani, N. Okamoto, T.
Ichikawa, S. Kikuta, Surf. Sci. 1991, 242, 54–58; e) M. Katayama,
R. S. Williams, M. Kato, E. Nomura, M. Aono, Phys. Rev. Lett. 1991,
66, 2762–2765.
[14] G. Pawin, K. L. Wong, K.-Y. Kwon, L.A. Bartels, Science 2006, 313,
961–962.
[15] J.A. Theobald, N. S. Oxtoby, M.A. Phillips, N. R. Champness, P. H.
Beton, Nature 2003, 424, 1029–1031.
[16] L. M.A. Perdig¼o, E. W. Perkins, J. Ma, P. A. Staniec, B. L. Rogers,
N. R. Champness, P. H. Beton, J. Phys. Chem. B 2006, 110, 12539–
12542.
[17] P. A. Staniec, L. M.A. Perdig¼o, A. Saywell, N. R. Champness, P. H.
Beton, ChemPhysChem 2007, 8, 2177–2181.
[18] B. Xu, C. Tao, W. G. Cullen, J. E. Reutt-Robey, E. D. Williams,
Nano Lett. 2005, 5, 2207–2211.
[28] J. C. Swarbrick, J. Ma, J.A. Theobald, N. S. Oxtoby, J. N. OꢁShea,
N. R. Champness, P. H. Beton, J. Phys. Chem. B 2005, 109, 12167–
12174.
[29] M. E. CaÇas-Ventura, W. Xiao, D. Wasserfallen, K. Müllen, H.
Brune, J.V. Barth, R. Fasel, Angew. Chem. 2007, 119, 1846–1850;
Angew. Chem. Int. Ed. 2007, 46, 1814–1818.
[30] M. Ruiz-Oses, N. Gonzalez-Lakunza, I. Silanes, A. Gourdon, A.
Arnau, J. E. Ortega, J. Phys. Chem. B 2006, 110, 25573–25577.
[31] D. L. Keeling, N. S. Oxtoby, C. Wilson, M. J. Humphry, N. R.
Champness, P. H. Beton, Nano Lett. 2003, 3, 9–12.
[32] J. B. Taylor, P. H. Beton, Phys. Rev. Lett. 2006, 97, 236102.
[33] U. K. Weber, V. M. Burlakov, L. M.A. Perdig¼o, R. H. J. Fawcett,
P. H. Beton, N. R. Champness, J. H. Jefferson, G. A. D. Briggs, D. G.
Pettifor, Phys. Rev. Lett. 2008, 100, 156101.
[34] a) L. M.A. Perdig¼o, N. R. Champness, P. H. Beton, Chem.
Commun., 2006, 538–540; b) P. A. Staniec, L. M.A. Perdig¼o, B. L.
Rogers, N. R. Champness, P. H. Beton, J. Phys. Chem. C 2007, 111,
886–893.
[35] F. Klappenberger, M. E. Canas-Ventura, S. Clair, S. Pons, U. Schlick-
um, Z. R. Qu, H. Brune, K. Kern, T. Strunskus, C. Wçll, A. Comis-
sa, A. de Vita, M. Ruben, J.V. Barth, ChemPhysChem 2007 8 1782–
1786.
[19] H. Spillmann, A. Kiebele, M. Stçhr, T. A. Jung, O. Bonifazi, F.
Cheng, F. Diederich, Adv. Mater. 2006, 18, 275–279.
[20] a) S. Stepanow, M. Lingenfelder, A. Dmitriev, H. Spillmann, E.
Delvigne, N. Lin, X. Deng, C. Cai, J.V. Barth, K. Kern, Nat. Mater.
2004, 3, 229–233; b) 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, 724–727; Angew. Chem. Int. Ed.
2007, 46, 710–713. M. Ruben, D. Payer, A. Landa, A. Comisso, C.
Gattinoni, N. Lin, J.-P. Collin, J.-P. Sauvage, A. De Vita, K. Kern, J.
Am. Chem. Soc. 2006, 128, 15644–15651.
[21] M. Bçhringer, K. Morgenstern, W. D. Schneider, R. Berndt, Angew.
Chem. 1999, 111, 832; Angew. Chem. Int. Ed. 1999, 38, 821.
[22] H. L. Zhang, W. Chen, L. Chen, H. Huang X. S. Wang, J. Yuhara,
A. T. S. Wee, Small 2007, 3, 2015–2018.
[23] Q. Chen, D. J. Frankel, N. V. Richardson Langmuir 2002, 18, 3219.
[24] a) F. Würthner, Chem. Commun. 2004, 1564; b) H. Langhals, Het-
erocycles 1995, 40, 477; c) B. Yoo, T. Jung, D. Basu, A. Dodabalapur,
B. A. Jones, A. Faccetti, M. R. Wasielewski, T. J. Marks, Appl. Phys.
Lett. 2006, 88, 082104; d) K.-Y. Law, Chem. Rev. 1993, 93, 449; e) G.
[36] a) B. Delley, J. Chem. Phys. 1990, 92, 508–517; b) B. Delley, J.
Chem. Phys. 2000, 113, 7756–7764; c) J. P. Perdew, K. Burke, M.
Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865–3868.
Received: March 14, 2008
Published online: July 30, 2008
Chem. Eur. J. 2008, 14, 7600 – 7607
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7607