Covalent connection of individualized, neutral, dendronized polymers on a solid substrate using a scanning force microscope

Article abstract of DOI:10.1002/chem.200600171

The synthesis of a neutral, high-molar-mass, acrylamide-based, third-generation dendronized polymer (denpol) with a defined number of azide groups at its periphery is reported. An attach-to route is used in which a first-generation denpol is reacted with

Full text of DOI:10.1002/chem.200600171

DOI: 10.1002/chem.200600171
Covalent Connection of Individualized, Neutral, Dendronized Polymers on a
Solid Substrate Using a Scanning Force Microscope
Rabie Al-Hellani,[a] Jçrg Barner,[b] Jürgen P. Rabe,*[b] and A. Dieter Schlüter*[a]
Abstract: The synthesis of a neutral,
high-molar-mass, acrylamide-based,
third-generation dendronized polymer
(denpol) with a defined number of
azide groups at its periphery is report-
ed. An attach-to route is used in which
eration polymer and a closely related
G2 dendron. The third-generation
denpol resulting from the first dendro-
nization experiment was spin-coated as
a sub-monolayer onto highly oriented
graphite precoated with an ultrathin
layer of C12H25NH2, which was intro-
duced to provide a well-defined sub-
strate for denpol adsorption and ma-
nipulation. Scanning force microscopy
revealed single denpols, which could be
moved across the surface and “welded”
by covalent cross-linking induced by
photochemical decomposition of the
azides into highly reactive nitrenes.
The successful formation of covalent
bonds between two denpols was con-
firmed by mechanically challenging the
link with the scanning force microscope
(SFM) tip. This is the second reported
a
first-generation denpol is reacted
with second-generation (G2) dendrons.
The degree of structure perfection of
the resulting denpol is quantified as
99.8%. This value was obtained after
the introduction of a fluorescence label
at the sites that remained unaffected
by the dendronization. The high cover-
age was independently confirmed for
the dendronization of another first-gen-
case of
a move–connect–prove se-
quence using polymers and the SFM,
which for the first time employs non-
charged denpols, thus widening the ap-
plicability of this method significantly.
Keywords: dendrons · fluorescence
spectroscopy · polymers · scanning
force microscopy · single-molecule
studies
Introduction
went nonselective intra- and intermolecular bond-forming
reactions, the latter leading to covalent cross-linking be-
tween the strands. The connection was proved by mechani-
cally challenging the welding point. A critical issue for such
move–connect–prove sequences is the preparation of the in-
dividual chains on the substrate under conditions in which
the adsorption energy has just the right size. On the one
hand it should not be too high, so that the molecules can be
moved about with the SFM tip without tearing them apart.
On the other hand it should not be too low, because other-
wise the molecules diffuse on the surface until they find
each other and form islands to reduce line tension. This
subtle balance between opposing factors was achieved by
depositing the positively charged denpols on a highly orient-
ed pyrolytic graphite (HOPG) surface from organic solu-
tion. The fact that these experiments can be carried out
under ambient conditions is considered a key advantage of
this approach, compared to the ultrahigh vacuum techniques
required for related connection experiments with the scan-
ning tunneling microscope (STM).[3] It must be pointed out,
however, that the preparation of single macromolecules of
such high molar mass from solution suffers from an intrinsic
problem, namely the presence of impurities. According to
In a recent experiment, two strands of dendronized poly-
mers (denpols)[1] were covalently “welded” together by pho-
tochemical treatment after they had been brought into tight
contact on an atomically flat, solid substrate by dragging
one of them to the other and pushing them into tight con-
tact with a scanning force microscope (SFM) tip.[2] The
actual welding was then achieved by peripheral azide groups
that decomposed to nitrenes upon UV irradiation. Because
of the extremely high reactivity of the nitrenes, they under-
[a] R. Al-Hellani, Prof. A. D. Schlüter
Department of Materials, Institute of Polymers
Eidgençssische Technische Hochschule, ETH-Zürich
Wolfgang-Pauli-Str. 10, HCI J 541, 8093 Zürich (Switzerland)
Fax : (+41)446331395
E-mail: dieter.schlueter@mat.ethz.ch
[b] J. Barner, Prof. Dr. J. P. Rabe
Department of Physics
Humboldt University Berlin
Newtonstrasse 15, 12489 Berlin (Germany)
Fax : (+49)30-2093-7632
E-mail: rabe@physik.hu-berlin.de
6542
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 6542 – 6551

FULL PAPER
our current view, these impurities play an important role in
immobilizing denpols on HOPG, since they can form a soft
adsorbate layer that may slow down the diffusion of the ad-
sorbed denpols.[4] It was therefore decided to develop proce-
dures to reduce the sensitivity of the experiment toward this
problem by 1) employing a chemistry which avoids the si-
multaneous presence of side products as in the previous
case,[2] and 2) precoating the solid substrate with a soft or-
ganic monolayer that guarantees the immobilization of
single adsorbed denpols and still allows their manipula-
tion.[5] We foresee a considerable future impact of SFM ex-
periments of the kind described for the bottom-up approach
to nanosized functional devices on solid surfaces whenever
molecular objects of whatever sort need to be stabilized in
certain geometries. To create a broader basis for move–con-
nect–prove sequences, a neutral, azide-decorated denpol
was synthesized.
Herein, we describe the synthesis of a noncharged, third-
generation (G3) denpol of high molar mass, 25% of the pe-
ripheral amine groups of which carry azide functions for
cross-linking, the remaining 75% being benzyloxycarbonyl
(Cbz)-protected. The synthesis uses the attach-to proce-
dure,[1] in which appropriately functionalized G2 dendrons
are bound to the two amine anchor groups that a high-
molar-mass G1 denpol presents per repeat unit. Since this
protocol involves hundreds of such coupling steps per mac-
romolecule, the degree to which the entire conversion could
be achieved was determined quantitatively. For this purpose,
residual amine groups in the former G1 denpol that had not
reacted with a G2 dendron were amplified by treating them
with 5-dimethylaminonaphthalene-1-sulfonyl (dansyl) chlor-
ide, a potent fluorescence label commonly used in biochem-
istry.[6] The degree to which the dansyl group was covalently
incorporated into the polymer structure was determined by
fluorescence intensity measurements, the results of which
were confirmed in an independent, structurally different
system.
tion of the degree of coverage was expected to be compli-
cated, specifically if higher generations were involved. The
reasons for this are that the NMR spectra for the higher
generation systems tend to be difficult to interpret, and also
solubility problems may arise. Additionally, the coverage
process would not automatically guarantee a homogeneous
distribution of the newly bound moieties over the entire
“surface” of the macromolecule, but rather may give rise to
the formation of domains with preferred coverage. To cir-
cumvent these problems, an orthogonal synthesis strategy
was devised in which denpols would become available with
a predetermined ratio of orthogonally protected amines of
two different kinds per repeat unit. Such a synthesis goal is
rather complex and will be described in detail elsewhere.[8]
The protecting group combinations chosen for this strategy
were either Cbz/Boc or 2,7-di-tert-butyl-9-fluorenylmethoxy-
carbonyl (Fmoc*)/Boc. For the present work it was impor-
tant to have a high-molar-mass G3 denpol with 25% of all
amine groups protected with Boc and 75% with Cbz. The
main thoughts behind this selection are as follows. With
both the third-generation and the high molar mass, the den-
polꢁs mobility on HOPG was expected to be in a reasonable
range (see Introduction and earlier work[9]). Additionally, it
was known from the literature and our own work that large
numbers of Boc groups can be easily cleaved off in the pres-
ence of large numbers of Cbz groups, thereby opening the
way for the attachment of an aromatic azide through active
ester chemistry to the denpol.
All synthetic steps finally leading to the target denpol 12c
are compiled in Schemes 1–3. As can be seen, the target
denpol carries 25% azide groups and 75% Cbz-protected
peripheral amines. The polymer synthesis is of the attach-to
kind (Scheme 3); the G1 denpol 11b was used as the start-
ing material (Scheme 2) to which the G2 dendrons 7b
(Scheme 1) were attached by using well-established active
ester couplings. In our laboratory, such couplings proceeded
virtually quantitatively in several cases if the conditions
were chosen carefully.[10] Most steps use conventional organ-
ic chemistry protocols and only a few comments will there-
fore be made. The starting point was the orthogonally pro-
tected dendron 1a which had already been made available
in a different context.[11] Also, the symmetrical dendron 3 of
Scheme 1 could be made according to a reported proce-
dure.[12] The selective deprotection of Cbz in 1a with hydro-
gen and Pd/C went cleanly and gave dendron 2a in quantita-
tive yield. The amide coupling of 2a and 3 was achieved
with N-hydroxybenzotriazole/N-(3-dimethylaminopropyl)-
N’-ethylcarbodiimide hydrochloride (HOBt/EDC) and gave
4 in a yield of 63%. Subsequent deprotection and addition
of dendron 1b furnished dendron 6, which carries three
Cbz-protected amines and one Boc group. To allow 6 to be
attached to a polymer, its focal point was converted into the
active ester 7b. This dendron was prepared on the 1 g scale.
Dendron 8 was synthesized according to a literature proce-
dure.[7b] To arrive at the dendronized polyacrylic amide, the
benzylic alcohol at its focal point was mesylated and then
reacted with potassium phthalimide to give 9a, which was
This characterized G3 denpol was prepared by spin coat-
ing the denpol from chloroform onto a quasi-2D network on
HOPG, to give a sub-monolayer of mostly individual den-
pols on HOPG precoated with a layer of C12H25COOH.
Single denpols were manipulated with the SFM tip in con-
tact mode for a new move–connect–prove sequence. A blind
experiment was also carried out to prove that irradiation is
an essential prerequisite for the connection to be irreversi-
ble, and that it therefore has a truly covalent nature.
Results and Discussion
Synthesis: So far only denpols with identically protected pe-
ripheral amine functions have been reported.[7] Typically, all
amine groups were either tert-butoxycarbonyl (Boc) or
trimethylsilyl(ethoxy)carbonyl (Teoc) protected. In principle
G
such denpols can be used for chemical surface modifications,
for example, the decoration with azide functions required
here. With such starting materials, however, the determina-
Chem. Eur. J. 2006, 12, 6542 – 6551
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6543

A. D. Schlüter, J. P. Rabe et al.
Scheme 2. a) 8, MsCl, CH2Cl2, Et3N, À208C, 2 h, and b) potassium phtha-
limide, DMF, 1008C, 14 h, 91% for both steps; c) N2H4·H2O, THF/EtOH,
608C, 6 h, 87%; d) MAC, CH2Cl2, Et3N, À208C, 30 min, 92%; e) DBPO,
DMF, 708C, 18 h, 70%; f) 25% HCl, THF, RT, 14 h, 96%. MsCl=mesyl
chloride, MAC=methacryloyl chloride, DBPO=dibenzoyl peroxide.
Scheme 1. a) KOH, THF/MeOH/H2O, 6 h, 55 8C, 88%; b) EtOAc/ethanol
(1:1), Pd/C, 1,4-cyclohexadiene, H2, 4 h, 97%; c) 25% HCl, THF, RT,
94.5%; d) dansyl chloride, CH2Cl2, Et3N, À308C, 20 min, 96%; e) 1. 3,
HOBt, EDC, CH2Cl2, À308C, 3 h, and 2. 2a, Et3N, CH2Cl2, MeOH,
À208C, 14 h, 63% for both steps; f) 25% HCl, THF, 08C, 3 h, 96%;
g) 1. 1b, HOBt, EDC, CH2Cl2, À308C, 3 h, and 2. 5, Et3N, CH2Cl2/
MeOH, À208C, 14 h, 78% for both steps; h) KOH, THF/MeOH/H2O,
6 h, 558C, 87%; i) HOSu, CH2Cl2, DCC, RT, 14 h, 83%. HOBt=N-hy-
droxybenzotriazole, EDC=N-(3-dimethylaminopropyl)-N’-ethylcarbodii-
mide hydrochloride, HOSu=N-hydroxysuccinimide, DCC=dicyclohexyl-
carbodiimide.
then converted into the free amine with hydrazine. Com-
pound 9b was connected to a polymerizable unit by reacting
it with freshly distilled methacrylic acid chloride. The result-
ing G1 macromonomer 10 was prepared on the 5 g scale.
Polymerization was carried out with dibenzoyl peroxide
(DBPO, 0.15 mol%) as radical initiator precursor at 708C in
a highly concentrated solution in DMF (30% w/w). It was
shown earlier that extremely high concentrations are a pre-
requisite for such polymerizations to give satisfactory molar
masses.[13] After considerable optimization work on the
small scale, one gram of monomer 10 was finally polymer-
ized to give the desired polymer 11a in 70% yield at a
molar mass of Mw =1.4”106 gmolÀ1 and polydispersity index
(PDI)=2.6 (gel permeation chromatography (GPC) univer-
sal calibration).
Scheme 3. a) 7b, CH2Cl2/MeOH, Et3N, RT, 60 h, 74 %; b) CF 3COOH,
CH2Cl2, RT, 14 h, 90%; c) N-succinimidyl-6-(4-azido-2-nitroanilino)hexa-
noate, CH2Cl2, Et3N, RT, 14 h, 85 %.
tion prior to its dendronization with G2 dendron 7b. For
that purpose, polymer 11b was dissolved in methanol which
contained 4–5 equivalents of triethylamine per ammonium
group of 11b, and a solution of 1.5 equivalents of dendron
7b per amine in methanol/methylene chloride (1:2) was
added dropwise. After stirring for two days, the solvent of
the still homogeneous solution was removed in a vacuum,
whereupon the polymeric residue was taken up in methyl-
ene chloride with some triethylamine and a solution of an-
other 0.25 equivalents of dendron 7b in the same solvent
was added. Stirring was continued for 12 hours, after which
Denpol 11a was then deprotected with aqueous 25% hy-
drochloric acid to give 11b, which was isolated by precipita-
6544
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 6542 – 6551

Dendronized Polymers
FULL PAPER
conventional workup furnished polymer 12a. For a quantita-
tive determination of the dendronization efficiency, see
below. To introduce the azide function required for the
cross-linking experiments, denpol 12a was treated with
excess trifluoroacetic acid until all Boc signals in the
concentration range 0.0274–219 mm in chloroform (Figure 1
top, Experimental Section). A linear relationship was ob-
served up to a concentration of 27 mm; this result indicated
1
700 MHz H NMR spectrum had disappeared. The resulting
25%-deprotected denpol was dissolved in chloroform and
precipitated into ethyl acetate/hexane (1:1) to give 12b. This
product was finally converted into 12c by reacting it with
the commercially available N-succinimidyl-6-(4-azido-2-ni-
troanilino)hexanoate in an excess of 1.5 equivalents per
amine of 12b. The conversion of this step was not deter-
mined. There is substantial evidence that succinimidyl active
esters react virtually quantitatively with amine groups of
denpols.[10]
Quantification of dendronization: The efficiency of succes-
sive dendronization of denpols with G1 dendrons has been
quantitatively determined all the way up from G2 to G4
denpols.[10] The attachment chemistry was based on the reac-
tion between amines and succinimidyl active esters, which is
known to proceed virtually quantitatively if the active ester
can be employed in large excess. Amine groups that hap-
pened to remain unaffected by the dendronization process
were artificially amplified by attaching the Sanger reagent
to them, which is a potent UV label for primary amines.
The resulting UV absorptions were measured to determine
the degree to which the amines had been labeled. Assuming
that all free amines had reacted with the reagent, this al-
lowed the derivation of the dendronization efficiency. Under
carefully chosen conditions efficiencies greater than 99%
were determined per growth step; this result indicates that
the attach-to procedure can lead to near-perfect, high-gener-
ation denpols. The present case differs somewhat from the
previous one in that the G3 denpol 12a was synthesized by
reaction of the starting polymer 11b not with G1 dendrons,
but rather with the G2 dendron active ester 7b. Intuitively
one would expect that in this case any remaining, non-
dendronized amine group of 11b would be more effectively
shielded than in the previous case, because it will be sur-
rounded by sterically more demanding G2 rather than just
G1 dendrons. This will hinder the diffusion of the dansyl
label to, and its reaction with, the remaining amines and
thus could result in a too low fluorescence intensity and,
therefore, falsely indicate a too high dendronization efficien-
cy. It was therefore of some concern to collect evidence that
the amines are still accessible by the dansylating reagent
(see below).
Figure 1. Top: Concentration of the dansylated model compound 2c in
chloroform [mm] versus fluorescence intensity as obtained from signal
areas at room temperature. Bottom: Fluorescence curves of 2c (a and d)
and polymers c) 12a(80) and d) 12a(100) after reaction with dansyl chlor-
A
ide.
that there was no quenching. All subsequent measurements
were therefore performed in this concentration range al-
though, once incorporated into the denpol structure, the
dansyl units should not easily be able to undergo both intra-
and intermolecular collisional self-quenching. The latter ar-
gument holds true only for those cases in which a few
amines scattered along the backbone carry dansyl labels.
Otherwise the covalent fixation of labels to the same poly-
mer chain may go so far as to render the above concentra-
tion dependence no longer applicable.
The dendronization efficiency was analyzed by dansyla-
tion of two different samples of 12a, both of which were
prepared from an identical batch of G1 denpol 11b with
Pn =1100 and Pw =2800. In the first sample, this starting
denpol was deprotected and reacted with two equivalents of
G2 dendron 7b with the aim of complete coverage. For the
second sample, a coverage of 80% of the amines was at-
tempted by reacting denpol 11b with 0.8 equivalents of den-
dron per amine group. In the following, these samples will
First the model compound 2c, which carries two dansyl
groups, was prepared (Scheme 1). It served as a reference
point for the fluorescence measurements, and was synthe-
sized by reaction of the two unprotected amine groups of
the corresponding G1 dendron 2b with 1.2 equivalents of
dansyl chloride per amine group at À308C for 15 minutes.
Despite these rather mild reaction conditions, product 2c
was isolated in a yield of virtually 100%. Its fluorescence in-
tensity, obtained from the peak areas, was measured in the
be referred to as 12a
A
samples were dansylated under somewhat more forceful
Chem. Eur. J. 2006, 12, 6542 – 6551
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6545

A. D. Schlüter, J. P. Rabe et al.
conditions than those used for model 2c by employing a
temperature of À108C[14] and a reaction time of two hours.
One equivalent of dansyl chloride per amine of denpol 11b
dendronized amines have actually reacted with the label.
This is why, as mentioned, more forceful reaction conditions
were applied[14] and a huge excess of the labeling agent was
employed. There are two other arguments that should be
considered in this context. First, denpol 12a(80), the synthe-
sis for which the starting denpol 11b was treated with
0.8 equivalents of G2 dendron 7b, has a dansylation degree
of 18% according to the findings above. This result confirms
not only that the dendronization goes almost to completion
even if the dendron, as in this case, is not in excess, but also
that the dansylation clearly reaches completion, at least in
this sterically less hindered case. Second, denpol 13 was ob-
tained as a pure material in a yield of 91% after several pre-
cipitation/dissolution cycles, which were implemented to
remove the last traces of the G2 dendron. The dendron had
been used in an excess of two per amine group. These cycles
were associated with some losses of polymer, and therefore
a yield in this percentage range indicates a near-complete
reaction course that does not leave much free space for in-
terpretation. This experiment was carried out with denpol
was used for 12a
amine group for 12a(80). If one assumes the presence of
1% nondendronized amines in 12a(100), this corresponds to
(100) and 1.5 equivalents per expected free
AHCTREUNG
a 100-fold excess of dansyl chloride, whereas the excess em-
ployed in the case of 12a(80) was much smaller. The inten-
tion behind the more forceful reaction conditions and the
huge excess of reagent for the treatment of 12a(100) was to
U
drive the dansylation of any free amine to completion. The
excess dansyl chloride and any low-molar-mass product
formed between it and an inadvertently present nucleophile
were removed by repeated precipitation/dissolution cycles.
After each cycle the fluorescence intensity of the polymers
was measured and the process was continued until the inten-
sity remained constant. This was typically the case after
three cycles for 12a(100) and one cycle for 12a(80). Addi-
U
tionally, the polymers were checked by thin-layer chroma-
tography (TLC). If the fluorescence intensity had reached
the final value, not even a trace of a blue-fluorescing com-
ponent moved away from the polymer, irrespective of the
solvents used.
13 rather than 12a(100), because the former was more abun-
A
dantly available and the dendronization yield could there-
fore be determined with higher accuracy. Finally, a method
was recently discovered for removing all dendrons of 13 by
hydrolytically cutting them off exclusively at their former
focal point. This will in future also enable investigation of
the structural perfection of the dendrons by high-field NMR
spectroscopy, a technique that cannot be reasonably applied
for such a purpose to G3 denpols of molar mass above one
million.
Figure 1 (bottom) shows typical fluorescence curves of
dansylated 12a(100) and 12a(80) with those of model com-
G
pound 2c. From a comparison of the signal intensities,
which is described in detail in the Experimental Section,
dansylation degrees of 0.15 and 18% were obtained for 12a-
ACHTREUNG
ACHTREUNG
ments. This result was even further substantiated by a close-
ly related dendronization reaction leading to denpol 13, for
which a coverage of 99.7% was found using exactly the
same procedure as described above. In this reaction denpol
13 was synthesized from a G1 polymethacrylate and a G2
dendron active ester like 7b, except that all amine functions
carried Boc groups (both not shown).[8]
Assuming that quenching of the fluorescence intensity is
not occurring, this kind of analysis still has the intrinsic dis-
advantage of not providing any clue as to whether all non-
SFM move–connect–prove sequences: Stable SFM imaging
of 12c was not possible for preparations from relatively low
concentrations (c=0.05 mgmLÀ1); this result indicates that
single denpols are not immobilized on the basal plane of
graphite, and possible impurities do not stabilize them. At
higher concentrations (cꢀ0.5 mgmLÀ1 depending on the
polymer batch), a stable quasi-2D network of denpols is
formed (Figure 2a). This may be due to the inherent stabili-
ty of the 2D network. Alternatively, a now increased con-
centration of impurities at the surface may form an organic
monolayer, which causes a larger friction coefficient for the
adsorbed denpols than does the bare basal plane of graphite,
since the potential-energy landscape across the surface ex-
hibits increased amplitudes. To differentiate between these
cases, two strands in the network were cut in the contact-
mode SFM. Figure 2bshows the resulting loose ends, with a
length of about 80 nm each, immobilized on the surface.
This observation, in the context of individualized denpols in
the low-concentration experiment, is attributed to the in-
creased concentration of impurities causing the immobiliza-
tion. Moving two loose ends with the SFM together (Figure
2b,c) and apart again (Figure 2d) proves the reversibility of
this kind of manipulation. After UV illumination of two
denpol strands previously moved together (Figure 2e), the
same manipulation does not allow separation of the denpols
any more (Figure 2f,g). Figure 2h–j shows that the junction
6546
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 6542 – 6551

Dendronized Polymers
FULL PAPER
Figure 2. Move–connect–prove sequence of 12c adsorbed on HOPG,
probably with an immobilizing impurity layer. a) Initial conformation of
network; b) two loose ends (87 and 74 nm) were cut out of the network;
c) forming and d) opening the junction; e) denpols were moved in tight
contact again and illuminated in situ with UV light (254 nm) for 5 min;
f)–j) challenging the link mechanically until chain rupture.
Figure 3. Compound 12c adsorbed onto a monolayer of C12H25NH2 on
HOPG. a) Initial conformation of aggregates 1 and 2 of 12c; b) extracted
single strands 1 and 2 after several manipulation steps; c) and d) corre-
sponding zoom of images a) and b) before and after manipulation, re-
spectively, which display the unfolding of a claw (dashed circle) resulting
in a length release of 63 nm. The white arrows in c) indicate the tip
motion during contact-mode SFM manipulation.
point can be moved while the denpol breaks at a different
position, thus proving the strength of the new linkage. Note
that the thickness of one of the newly formed loose ends is
decreased at one position; this result indicates that it may
consist of a duplex.[15] Note also the twisted, higher-order su-
pramolecular structures.
prove clearly that individual denpols can be immobilized on
the precoated HOPG, imaged in tapping-mode SFM, and
moved across the surface with high lateral accuracy in the
contact mode. The process includes individualizing of den-
pols by dissection of denpol aggregates, which will greatly
simplify manipulation experiments at the single-molecule
level.
These single denpols were used for a move–connect–
prove sequence (Figure 4). The linkage between two den-
pols can again be induced with UV light, as proven by chal-
lenging it mechanically with the SFM tip in contact mode up
to chain rupture beside the induced linkage, thus indicating
that the linkage between the individual strands is of truly
covalent nature.
Similar results were obtained for monolayers of
C23H47COOH and indicate that there is some flexibility in
the choice of the organic precoating. This finding opens up
the possibility, for instance, of co-adsorbing these neutral
denpols with either anionic or cationic polyelectrolytes
(which require oppositely charged surfaces for their adsorp-
tion) to form heterojunctions with different macromolecules
including biopolymers. With the demonstration of crosses[2]
and circles (here), two fundamental 2D topologies have
been realized.
In the case described above we attribute the denpol im-
mobilization to an ultrathin layer of organic impurities, and
thus we decided to render the experiment more clear-cut by
precoating HOPG with a well-defined organic layer, that is,
a spin-coated monolayer of C12H25NH2 or C23H47COOH.[5]
In these cases denpols also adsorband immobilize from low
concentrations, which allows the observation of denpol ag-
gregates and single denpol strands. Figure 3a shows 12c ad-
sorbed onto a monolayer of C12H25NH2 on HOPG, in which
two probable aggregates 1 and 2 are visible. SFM manipula-
tion was used to prove the aggregation and to dissect the ag-
gregates into their components as a starting point for a
move–connect–prove sequence. Figure 3bdisplays two indi-
vidual strands 1 and 2 with lengths of 310 and 225 nm, re-
spectively, extracted from these aggregates by eight manipu-
lation steps. Interestingly, intramolecular clew-shaped ar-
rangements can also be opened by lateral manipulation (Fig-
ure 3c,d). The initial conformation shown in Figure 3c exhib-
its a claw (in dashed circle) with the internal structure not
being evident. A probable conformation of the polymer
strand is sketched as a black line. Moreover, to the left and
right of the claw two small denpol fragments (white
arrows 1 and 2) are adsorbed onto the strand, thus acting as
an absolute position marker along the chain at a separation
distance of 25 nm. After two steps of manipulation the claw
was opened, as revealed by an increase in the separation of
the two markers along the contour to 63 nm. These facts
Chem. Eur. J. 2006, 12, 6542 – 6551
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6547

A. D. Schlüter, J. P. Rabe et al.
Cda: concentration of dansyl units in polymer; n: number of moles of re-
peating units; nda: number of moles of dansylated repeating units; L: con-
centration of the dansylated repeat units normalized to 1 gLÀ1
.
12a
of polymer, n=3.84”10À4 mol is obtained. The values m=1.9 mg and
V=2 mL from Figure 2bgive
da =4.8”10À7 m. Normalizing this value to
1 gLÀ1 (1.9”10À3 g/(2”10À3), L=0.95 gLÀ1), it follows that
da =4.8”
(100): Assuming no free amine groups, Mru =2604 gmolÀ1. For 1 gLÀ1
C
A
n
10À7 molLÀ1/0.95 gLÀ1 =5.05”10À7 mol. From this value the ratio of
dendronized over dansylated repeating units is calculated as n/nda =3.84”
10À4/(5.05”10À7)=760. Since two dansyl groups will not be attached to
the same repeat unit, this value corresponds to, on average, one dansyl
unit at every 380th repeat unit.
12a(80): Assuming that 20% of the repeating units are dansylated gives
M
ru =2235 gmolÀ1. For 1 gLÀ1 of polymer, n=4.47”10À4 mol is obtained.
The values m=0.25 mg and V=2 mL from Figure 2bgive
10À6 m. Normalizing this value to 1 gLÀ1 (0.25”10À3 g/(2”10À3), L=
da =8.7”10À6 molLÀ1/0.125 gLÀ1 =6.96”
Cda =8.7”
AHCTREUNG
0.125 gLÀ1), it follows that
n
10À5 mol. From this value the ratio of dendronized over dansylated re-
peating units is calculated as n/nda =4.47”10À4/(6.96”10À5)=6.4, which
corresponds to, on average, one dansyl unit every 3.2 repeat units.
Figure 4. Move–connect–prove sequence of 12c adsorbed onto a mono-
layer of C12H25NH2 on HOPG. a) Initial conformation as shown in Figure
3b; b) a junction between the two strands is formed and c) opened;
d) denpols are moved back again and illuminated in situ with UV light
(254 nm) for 5 min; e)–k) challenging the link mechanically up to strand
rupture close to the junction. In images b)–d), the local topography was
refreshed only inside the dashed square to speed up the manipulation
process. The white arrows indicate the tip movement during contact-
mode SFM manipulation.
Scanning force microscopy: For SFM sample preparation, denpols were
either directly spin-coated from chloroform onto HOPG or spin-coated
onto precoated HOPG (monolayer of C12H25NH2 or C23H47COOH). For
direct deposition of denpols onto HOPG, denpols (10 mL) dissolved in
chloroform (0.5 mgmLÀ1) were spin-coated at 50 rps onto freshly cleaved
HOPG and dried under ambient conditions for 0.5 h for subsequent SFM
investigation.
For deposition of denpols onto precoated HOPG, in a first step a solution
of C12H25NH2 or C23H47COOH (10 mL, Sigma Aldrich) in chloroform
(0.1 mgmLÀ1) was spin-coated at 50 rps onto freshly cleaved HOPG to
yield a monolayer, and the sample was dried for 0.5 h under ambient con-
Experimental Section
ditions. In
a second step, denpol (10 mL) dissolved in chloroform
Syntheses: Dendrons 1a, 1b, 2b, 3, and 8 were prepared according to lit-
erature procedures.[4a, 14] Reagents were purchased from Aldrich, Across,
or Fluka. MAC was freshly distilled before use. THF and triethylamine
were refluxed over Na with benzophenone as indicator, and dichlorome-
thane (CH2Cl2) was dried by distillation over CaH2. All other reagents
and solvents were used as received. All reactions were performed under
a nitrogen atmosphere. Silica gel 60M (Macherey–Nagel, 0.04–0.063 mm,
230–400 mesh) was used as the stationary phase for column chromatogra-
phy. 1H and 13C NMR spectra were recorded on Bruker AM 300 (1H:
300 MHz, 13C: 75 MHz), AV 500 (1H: 500 MHz, 13C: 125 MHz), and
AV 700 (1H: 700 MHz) spectrometers at room temperature (if not other-
wise stated) using CDCl3 or [D4]MeOH as solvent. The 13C NMR spectra
of polymers 12b and 12c were not recorded because too long accumula-
tion times were required to obtain spectra with sufficient signal-to-noise
ratios. ESI-MS analyses were performed by the MS service of the Labo-
ratorium für Organische Chemie, ETH Zürich, on an IonSpec Ultra in-
strument. Elemental analyses were performed by the Mikrolabor of the
Laboratorium für Organische Chemie, ETH Zürich. The samples were
dried thoroughly under vacuum prior to analysis to remove strongly ad-
hering solvent molecules. GPC measurements were carried out by using
a PL-GPC 220 instrument with a 2”PL-Gel Mix-B LS column set (2”
30 cm) equipped with refractive index (RI), viscosity, and light scattering
(LS; 15 and 908 angle) detectors, and LiBr (1 gLÀ1) in DMF as eluent at
808C. Universal calibration was performed with poly(methyl methacry-
late) standards in the range of Mp =2680 to 3900000 (Polymer Laborato-
ries Ltd, UK).
(0.04 mgmLÀ1) was spin-coated at 50 rps onto the precoated HOPG. The
result was an incomplete monolayer of individual strands of denpols. The
sample was dried for 0.5 h before SFM imaging and manipulation.
A home-built SFM based on the multimode head and Nanoscope III con-
troller of Digital Instruments Inc. (Santa Barbara, CA, USA) was used
for SFM imaging and lateral manipulation, that is, dragging the molecules
across the surface. Besides basic imaging in the tapping mode, the setup
allowed movement of the SFM tip in the x–y plane along predefined
traces with additional control of the normal force,[2] in which the SFM
was gently switched from tapping to contact mode.
As every step in manipulation required verification of the resulting
object position and shape, that is, position and conformation of the
denpol, a complete global image had to be recorded; this procedure
slowed down the overall manipulation process. To bypass this problem a
fast object-tracking procedure was integrated into the SFM setup, in
which only the small area of interest (dashed white square in Figure 4)
was scanned with lower resolution than the global image, but with the
same tip velocity. A global SFM image of 512”512 pixels was recorded
in several minutes, whereas local fast object-tracking with a resolution of,
for example, 64”64 pixels gave the topographic result on a timescale of a
few seconds, thus significantly speeding up the manipulation process.
For imaging and lateral manipulation, Olympus etched cantilevers
(OMCL-AC240) with a nominal normal spring constant of 2 NmÀ1 and a
tip radius below 10 nm were used. Images were obtained by SFM operat-
ing in the tapping mode. During manipulation in the contact mode, the
SFM tip was moved with a velocity of 500 nmsÀ1, whereas the deflection
was kept constant at 10 nm, which corresponds to a normal force of
about 20 nN.
Fluorescence measurements: Fluorescence measurements were per-
formed in chloroform at room temperature using a Spex Fluorolog 2
spectrometer (Jobin Yvon, UK) with 1-cm quartz cells (see Figure 1).
The intensity of the fluorescence bands was determined by measuring
their area. A line shape analysis was not performed. The fluorescence in-
tensity of the non-dansylated polymer was negligible.
UV illumination was carried out with a standard spectral Ne–Hg lamp
(Pen-Ray 11SC-1, UVP, Upland, CA, USA) with maximum emission at a
wavelength of 254 nm.
In the following, the procedure is delineated according to the determina-
Ethyl-3-(3-aminopropyl)-5-(3-tert-butyloxycarbonylaminopropyl)
ben-
tion of the degrees of coverage for 12a
rescence signal; m: weighed-in mass of polymer; Mru: molecular weight
of repeating unit; V: volume of chloroform solution of labeled polymer;
A
zoate (2a): Compound 1a (1.55 g, 3.11 mmol) in EtOAc/EtOH (30 mL,
1:1), Pd/C (0.16 g, 10 wt%), and 1,4-cyclohexadiene (1.74 g, 22.0 mmol)
were added to a hydrogenation flask. The mixture was hydrogenated
6548
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 6542 – 6551

Dendronized Polymers
FULL PAPER
under H2 (3 bar) at room temperature for 4 h. The solution was then fil-
tered through celite and the solvent evaporated at room temperature to
yield 2a as a viscous oil (1.10 g, 97%). The product was used for the next
step without further purification and structural analysis.
1H NMR (CDCl3): d=1.42 (s, 9H; CCH3), 1.75 (m, 8H; CH2), 1.88 (m,
4H; CH2), 2.57 (m, 8H; CH2Ph), 2.70 (m, 4H; CH2Ph), 3.07 (t, J=
3.3 Hz, 2H; CH2NH), 3.14 (q, J=5.1 Hz, 6H; CH2NH), 3.34 (t, J=
5.1 Hz, 4H; CH2NH), 5.01 (s, 6H; OCH2Ph), 5.15 (brs, 2H; NH), 6.83 (s,
1H; ArH), 7.04 (s, 2H; ArH), 7.27 (m, 19H; ArH), 7.73 ppm (s, 2H;
ArH); 13C NMR (CDCl3): d=28.39, 31.23, 31.29, 32.58, 33.16, 39.72,
40.14, 66.79, 125.03, 127.79, 128,15, 128.24, 128.65, 131.84, 134.55, 135.59,
136.64, 141.93, 142.96, 156.87, 175.80 ppm; ESI-MS: m/z: 1198 [M+Na]+.
Ethyl-3-{3,5-bis[3-(benzyloxycarbonylamino)propyl]benzoylamino}-5-[3-
(tert-butyloxycarbonylamino)propyl] benzoate (4): N-hydroxybenzotria-
zole (HOBt; 0.30 g, 2.28 mmol) was added to a solution of 3 (1.10 g,
2.18 mmol) in CH2Cl2 (60 mL). After 10 min at À308C, N-(3-dimethyla-
minopropyl)-N’-ethylcarbodiimide
hydrochloride
(EDC;
0.46 g,
3-(3-{3,5-Bis-[3-(benzyloxycarbonylamino)propyl]benzoylamino}propyl)-
5-(3-{3-[3-(benzyloxycarbonylamino)propyl]-5-[3-(tert-butyloxycarbonyl-
amino)propyl]benzoylamino}propyl)benzoic acid 2,5-dioxopyrrolidin-1-yl
ester (7b): N-Hydroxysuccinimide (HOSu; 46 mg, 0.39 mmol) was added
to a solution of 7a (0.33 g, 0.27 mmol) in dry CH2Cl2 (10 mL) at room
temperature. After 15 min, dicyclohexylcarbodiimide (DCC; 85 mg,
0.41 mmol) was added at À208C. The resulting mixture was warmed to
room temperature and stirred overnight. The precipitate was isolated by
filtration and the solvent was evaporated at room temperature. Chroma-
tographic separation (silica gel, CH2Cl2/MeOH 4:1) yielded 7b (0.29 g,
83%) as a white foam. 1H NMR (CDCl3): d=1.37 (s, 9H; CCH3), 1.73
(m, 8H; CH2), 1.92 (m, 4H; CH2), 2.58 (m, 8H; CH2Ph), 2.69 (m, 4H;
CH2Ph), 2.85 (s, 4H; CH2), 3.00 (t, J=6.3 Hz, 2H; CH2NH), 3.08 (q, J=
8.1 Hz, 6H; CH2NH), 3.35 (t, J=6.6 Hz, 4H; CH2NH), 5.01 (s, 6H;
OCH2Ph), 7.07 (s, 2H; ArH), 7.27 (m, 15H; ArH), 7.35 (s, 3H; ArH),
7.37 (s, 2H; ArH), 7.75 ppm (s, 2H; ArH); 13C NMR (CDCl3): d=28.35,
30.66, 31.18, 32.47, 32.93, 39.49, 40.01, 66.57, 79.24, 124.90, 125.17, 127.88,
128.08, 128.51, 131.84, 134.55, 135.59, 141.93, 142.95, 157.11, 162.23,
168.77, 169.99 ppm; ESI-MS: m/z: 1295 [M+Na]+; elemental analysis
calcd (%) for C72H85N7O14 (1272.62): C 67.96, H 6.73, N 7.71; found: C
67.53, H 7.05, N 7.56.
2.38 mmol) was added. The mixture was stirred until the hydrochloride
had dissolved (ꢀ3 h), then a solution of 2 (0.59 g, 1.83 mmol) and Et3N
(1.03 mL, 7.33 mmol) in MeOH/CH2Cl2 (10 mL, 1:1) was added dropwise
at À208C. The mixture was warmed to room temperature and stirred
overnight. It was then washed with saturated NaHCO3 (50 mL) and brine
(50 mL), and dried over MgSO4. Chromatographic separation (silica gel,
1
EtOAc/hexane 2:1) yielded 4 (1.56 g, 63%) as a colorless foam. H NMR
(CDCl3): d=1.38 (t, J=7.2 Hz, 3H; CH3), 1.44 (s, 9H; CCH3), 1.86 (m,
6H; CH2), 2.00 (m, 2H; CH2), 2.64 (m, 6H; CH2Ph), 2.74 (t, J=7.8 Hz,
2H; CH2Ph), 3.13 (m, 6H; CH2NH), 3.45 (q, J=6.0 Hz, 2H; CH2NH),
4.38 (q, J=6.9 Hz, 2H; OCH2), 4.75 (brs, 1H; NH), 5.06 (s, 4H;
OCH2Ph), 5.12 (brs, 2H; NH), 6.89 (brs, 1H; NH), 7.09 (s, 1H; ArH),
7.22 (s, 1H; ArH), 7.30 (m, 10H; ArH), 7.41 (s, 2H; ArH), 7.68 (s, 1H;
ArH), 7.73 ppm (s, 1H; ArH); 13C NMR (CDCl3): d=14.36, 28.43, 31.04,
31.14, 31.59, 32.38, 32.77, 33.20, 39.75, 40.06, 60.96, 66.63, 79.22, 124.87,
127.17, 127.18, 127.99, 128.09, 128.51, 131.65, 133.20, 134.93, 136.56,
141.68, 142.04, 142.07, 156.06, 156.58, 1.66.81, 167.91 ppm; ESI-MS: m/z:
873 [M+Na]+; elemental analysis calcd (%) for C49H62N4O9 (851.04): C
69.15, H 7.34, N 6.58; found: C 69.24, H 7.49, N 6.56.
Ethyl-3-{3,5-bis[3-(benzyloxycarbonylamino)propyl]benzoylamino}-5-(3-
amino)propyl benzoate (5): Aqueous 25% HCl (0.75 mL, 5 equiv) in
THF (3 mL) was added slowly to a solution of 4 (0.98 g, 1.15 mmol) in
THF (25 mL) at 08C. The mixture was stirred for 3 h. The solvent was
evaporated at room temperature to yield 5 as viscous oil (0.87 g, 96%).
The product was used for the next step without further purification and
structural analysis.
3,5-Bis(3-tert-butyloxycarbonylaminopropyl)benzyl phthalimide (9a):
Mesyl chloride (1.77 mL, 22.8 mmol) in CH2Cl2 (5 mL) was added to a
solution of 8 (8.4 g, 19.9 mmol) and Et3N (8.5 mL, 59.7 mmol) in CH2Cl2
(150 mL) at À308C. The mixture was stirred at À208C for 2 h and regu-
larly checked by TLC (EtOAc/hexane 1:1) until 8 had disappeared. The
reaction was then quenched with MeOH and the solution washed four
times with cold water and brine. The organic phase was dried over
MgSO4, the solvent was evaporated at room temperature, and the prod-
uct was dried under vacuum. The residue was then dissolved in DMF
(100 mL), potassium phthalimide (4.05 g, 21.8 mmol) was added at room
temperature, and the mixture was stirred overnight at 1008C. The DMF
was evaporated and the residue was dissolved in CH2Cl2 (70 mL), washed
with 0.5m NaOH, saturated brine, and saturated NaHCO3, and finally
dried over MgSO4. Chromatographic separation (silica gel, EtOAc/
hexane 1:1) yielded 9a (10 g, 91%) as a pale gray solid. 1H NMR
(CDCl3): d=1.44 (s, 18H; CH3), 1.78 (m, 4H; CH2), 2.58 (t, J=7.8 Hz,
4H; ArCH2), 3.14 (m, 4H; NHCH2), 4.66 (brs, 2H; NH), 4.79 (s, 2H;
CH2), 6.92 (s, 1H; ArH), 7.08 (s, 2H; ArH), 7.72 (m, 2H; ArH),
7.85 ppm (m, 2H; ArH); 13C NMR (CDCl3): d=28.43, 31.64, 32.88, 40.13,
41.57, 79.04, 94.71, 123.36, 126.38, 128.06, 132.14, 133.96, 136.54, 142.21,
156.00, 168.06 ppm; ESI-MS: m/z: 574 [M+Na]+; elemental analysis
calcd (%) for C47H60N4O8 (551.67): C 67.49, H 7.49, N 7.62; found: C
67.54, H 7.50, N 7.57.
Ethyl-3-(3-{3,5-bis[3-(benzyloxycarbonylamino)propyl]benzoylamino}-
propyl)-5-(3-{3-[3-(benzyloxycarbonylamino)propyl]-5-[3-(tert-butyloxy-
carbonylamino)propyl]benzoylamino}propyl) benzoate (6): HOBt
(90 mg, 0.66 mmol) was added to a solution of 1b (0.30 g, 0.63 mmol) in
CH2Cl2 (20 mL) at room temperature. After 10 min at À308C, EDC
(0.13 g, 0.68 mmol) was added and the mixture stirred for 3 h. Then a sol-
ution of 5 (0.40 g, 0.51 mmol) and Et3N (0.57 mL, 4.1 mmol) in MeOH/
CH2Cl2 (15 mL, 1:1) was added dropwise at À208C. The resulting mixture
was warmed to room temperature and stirred for 14 h. It was then
washed with brine (40 mL) and saturated NaHCO3 (40 mL). The organic
phase was dried over MgSO4 and the solvent removed in vacuum. Chro-
matographic separation (silica gel, EtOAc/hexane 3:1) yielded 6 (0.48 g,
78%) as a colorless foam. 1H NMR (CDCl3): d=1.37 (t, J=6.9 Hz, 3H;
CH3), 1.43 (s, 9H; CCH3), 1.75 (m, 8H; CH2), 1.97 (m, 4H; CH2), 2.58
(m, 8H; CH2Ph), 2.69 (m, 4H; CH2Ph), 3.08 (q, J=6.9 Hz, 2H;
CH2NH), 3.13 (q, J=6.1 Hz, 6H; CH2NH), 3.42 (t, J=5.8 Hz, 4H;
CH2NH), 4.33 (q, J=7.2 Hz,2H; CH2), 5.05 (s, 6H; OCH2Ph), 6.85 (brs,
2H; NH), 7.05 (s, 2H; ArH), 7.25 (s, 2H; ArH), 7.30 (m, 15H; ArH),
7.38 (s, 3H; ArH), 7.70 ppm (s, 2H; ArH); 13C NMR (CDCl3): d=14.73,
28.43, 30.66, 31.01, 32.35, 32.95, 39.42, 40.05, 66.64, 79.24, 124.86, 125.17,
127.88, 128,04, 128.11, 128.52, 136.56, 141.65, 141.99, 156.55, 162.23,
168.77 ppm; ESI-MS: m/z: 1226 [M+Na]+; elemental analysis calcd (%)
for C47H60N4O8 (1203.46): C 69.86, H 7.20, N 6.98; found: C 69.66, H
7.12, N 7.02.
3,5-Bis(3-tert-butyloxycarbonylaminopropyl)benzylamine (9b): Hydra-
zine hydrate (2.64 mL, 54.3 mmol) was added to a solution of 9a (10 g,
18.1 mmol) in THF/EtOH (80 mL, 1:1) and the mixture was stirred at
608C for 6 h. After the reaction was finished (TLC), the solvent was
evaporated. The precipitate was then dissolved in CH2Cl2 and washed
with 0.5m NaOH and saturated brine. Chromatographic separation (silica
gel, CH2Cl2/MeOH/Et3N 100:10:1) yielded 9b (6.66 g, 87%) as a yellow-
ish oil. 1H NMR (CDCl3): d=1.45 (s, 18H; CH3), 1.77 (m, 4H; CH2),
2.60 (t, J=7.5 Hz, 4H; ArCH2), 3.14 (q, J=6.6 Hz, 4H; NHCH2), 3.89 (s,
2H; CH2NH), 4.79 (brs, 2H; NH), 6.92 (s, 1H; ArH), 6.98 ppm (s, 2H;
ArH); 13C NMR (CDCl3): d=28.44, 31.67, 31.89, 32.96, 40.13, 46.41,
79.05, 124.83, 127.00, 141.96, 143.53, 156.00 ppm; ESI-MS: m/z: 422
[M+Na]+.
3-(3-{3,5-Bis[3-(benzyloxycarbonylamino)propyl]benzoylamino}propyl)-5-
(3-{3-[3-(benzyloxycarbonylamino)propyl]-5-[3-(tert-butyloxycarbonyl-
amino)propyl]benzoylamino}propyl)benzoic acid (7a): Compound
6
(0.39 g, 0.32 mmol) was heated with KOH pellets (0.07 g, 1.24 mmol) in
THF/MeOH/H2O (24 mL, 2:3:1) at 558C for 6 h. After the reaction was
finished (TLC), water (3 mL) and then acetic acid were added until pH 5
was reached. The solvent was evaporated and the G2 acid 7a extracted
with CH2Cl2. The organic phase was dried over MgSO4. After evapora-
tion of the solvent, acid 7a (0.33 g, 87%) was obtained as a white solid.
3,5-Bis(3-tert-butyloxycarbonylaminopropyl)benzylmethacrylamide (10):
Methacryloyl chloride (2.05 mL, 21.2 mmol) in CH2Cl2 (4 mL) was added
dropwise to a solution of 9b (7.45 g, 17.7 mmol) and Et3N (3.72 mL,
Chem. Eur. J. 2006, 12, 6542 – 6551
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6549

A. D. Schlüter, J. P. Rabe et al.
26.5 mmol) in CH2Cl2 (60 mL) at À308C. The mixture was stirred at
À208C for 30 min and monitored by TLC (EtOAc/hexane 1:1) until 9b
had disappeared. The reaction was then quenched with MeOH. The solu-
tion was washed with saturated brine and saturated NaHCO3. The organ-
ic phase was dried over MgSO4 and the solvent evaporated at room tem-
perature. Chromatographic separation (silica gel, EtOAc/hexane 1:1) was
carried out twice to yield 10 (7.95 g, 92%) as a white solid. 1H NMR
(CDCl3): d=1.45 (s, 18H; CH3), 1.77 (m, 4H; CH2), 1.98 (s, 3H; CH2),
2.60 (t, J=8.1 Hz, 4H; ArCH2), 3.14 (q, J=6.3 Hz, 4H; NHCH2), 4.45
(d, J=6.2 Hz, 2H; NH), 4.69 (brs, 2H; NH), 5.35 (s, 1H; CH), 5.75 (s,
1H; CH), 6.26 (t, J=6.1 Hz, 1H; NH), 6.93 (s, 1H; ArH), 6.94 ppm (s,
2H; ArH); 13C NMR (CDCl3): d=18.76, 28.43, 31.62, 32.85, 40.02, 43.74,
79.09, 119.67, 125.73, 127.76, 138.45, 139.95, 142.21, 156.00, 168.24 ppm;
ESI-MS: m/z: 512 [M+Na]+; elemental analysis calcd (%) for
C27H43N3O5 (489.65): C 66.23, H 8.85, N 8.58; found: C 65.95, H 8.87, N
8.49.
solving it in CHCl3 followed by precipitation into EtOAc/hexane (1:2) to
yield 12b (45 mg, 90%) as a slightly yellow solid. 1H NMR (CDCl3/
CD3OD): d=1.65 (br; CH2), 2.40 (br; ArCH2), 2.94 (br; NHCH2), 4.85
(br; CH2), 6.94 (br; ArH), 7.11 (br; ArH), 7.28 ppm (br; ArH).
Poly
N
N
propyl}-5-{3-[3-(3-{benzyloxyamino}propyl)-5-(3-{6-(4-azido-2-nitroanili-
no)hexanoylamino}propyl)benzoylamino]propyl}benzoylamino)propyl]-
benzylmethacrylamide} (12c): A solution of N-succinimidyl-6-(4-azido-2-
nitroanilino)hexanoate (7.5 mg, 3 equiv per repeating unit) in CH2Cl2
(2 mL) was added dropwise to a solution of 12b (17 mg, 0.0065 mmol)
and Et3N (0.01 mL) in CH2Cl2 (2 mL) at À208C in the dark. The mixture
was stirred for 14 h and the solvent evaporated. The polymer was dis-
solved in CHCl3 and precipitated into EtOAc/hexane (1:2) to yield 12c
(16 mg, 85%) as a slightly red solid. 1H NMR (CDCl3/CD3OD): d=1.55
(br; CH2), 1.75 (br; CH2), 2.01 (br; CH2), 2.40 (br; ArCH2), 2.55 (br;
CH2), 2.94 (br; NHCH2), 3.05 (br; NHCH2), 4.85 (br; CH2), 6.7 (br;
ArH), 6.95 (br; ArH), 7.11 (br; ArH), 7.55 ppm (br; ArH); IR: n˜ =
Poly[3,5-bis(3-tert-butyloxycarbonylaminopropyl)benzylmethacrylamide]
G
2118 cmÀ1
Dansylated model compound 2c and polymers 12a(80) and 12a(100)
.
(11a): A solution of DBPO (0.47 mg, 0.16 mol%) in DMF (0.2 mL) was
added to a Schlenk tube containing monomer 10 (0.60 g, 1.22 mmol). If
the monomer did not completely dissolve, more DMF was added until a
homogeneous solution was achieved, which was then concentrated up to
approximately the initial amount of DMF by evacuation with visible in-
spection. An exact concentration cannot, therefore, be given. The result-
ing mixture was degassed several times by freeze–pump–thaw cycles,
then kept at 708C for 18 h. After polymerization, the polymer was dis-
solved in CH2Cl2 and purified by column chromatography (silica gel,
C
Ethyl-3,5-bis[3-(5-dimethylamino-1-naphthalenesulfonamide)propyl] ben-
zoate (2c): Dansyl chloride (0.13 g, 0.48 mmol) in CH2Cl2 (4 mL) was
added dropwise to
a solution of 2b (68 mg, 0.20 mmol) and Et3N
(0.34 mL, 12 equiv) in CH2Cl2/MeOH (15 mL, 2:1) at À308C. The mix-
ture was stirred at À308C for 20 min and monitored by TLC (CH2Cl2/
MeOH/Et3N 100:10:1) until 2b had disappeared. The reaction was then
quenched with MeOH. The solution was washed with saturated brine
and saturated NaHCO3. The organic phase was dried over MgSO4 and
the solvent evaporated. Chromatographic separation (silica gel, EtOAc/
hexane 1:2) was carried out to yield 2c (0.14 g, 96%) as a yellowish solid.
1H NMR (CDCl3): d=1.38 (t, J=7.2 Hz, 3H; CH3), 1.68 (m, 4H; CH2),
2.48 (t, J=7.8 Hz, 4H; ArCH2), 2.88 (s, 12H; CH3), 2.92 (q, J=6.8 Hz,
4H; NHCH2), 4.35 (q, J=7.2 Hz, 2H; CH2), 5.26 (t, J=6.0 Hz, 2H;
NHCH2), 6.85 (s, 1H; ArH), 7.16 (s, 1H; ArH), 7.18 (s, 1H; ArH), 7.52
(m, 12H; ArH), 8.24 (d, J=6.3 Hz, 2H; ArH), 8.35 (d, J=8.7 Hz, 2H;
ArH), 8.52 ppm (d, J=6.3 Hz, 2H; ArH); 13C NMR (CDCl3): d=14.23,
31.05, 32.25, 42.61, 45.42, 60.95, 115.22, 118.79, 123.24, 127.01, 128.46,
129.61, 129.64, 129.86, 130.42, 130.65, 133.16, 134.80, 141.42, 152.00,
166.68 ppm; ESI-MS: m/z: 752.8 [M+Na]+; elemental analysis calcd (%)
for C39H46N4O6S2 (730.81): C 64.09, H 6.34, N 7.66; found: C 64.05, H
6.49, N 7.51.
CH2Cl2 as eluent) to yield 11a (0.42 g, 70%) as
a colorless foam.
1H NMR (CDCl3): d=0.85 (br; CH3), 1.45 (br; CH3), 1.85 (br; CH2), 2.55
(br; ArCH2), 3.11 (br; NHCH2), 4.05 (br; CH2CH3), 5.35 (br; NH),
6.94 ppm (br; ArH); 13C NMR (CDCl3): d=28.45, 31.62, 32.85, 40.02,
41.24, 79.09, 125.73, 127.76, 138.45, 139.95, 142.21, 158.00, 168.24 ppm; el-
emental analysis calcd (%) for (C27H43N3O5)n (489)n: C 66.23, H 8.85, N
8.58; found: C 65.63, H 9.01, N 8.33.
Poly[3,5-bis(3-aminopropyl)benzylmethacrylamide·2HCl] (11b): Aque-
G
ous 25% HCl (0.95 mL, 6 equiv per Boc group) in THF (2 mL) at 08C
was added dropwise to a solution of 10a (0.30 g, 0.61 mmol) in THF
(6 mL) and the mixture was stirred for 14 h. Evaporation of the solvent
at room temperature yielded 11b (0.21 g, 96%) as a colorless solid.
1H NMR (CD3OD): d=0.85 (br; CH3), 1.85 (br; CH2), 2.55 (br; ArCH2),
3.11 (br; NHCH2), 4.05 (br; CH2CH3), 6.94 ppm (br; ArH); 13C NMR
(CD3OD): d=19.18, 32.55, 39.61, 126.60, 127.61, 139.18, 141.38,
168.24 ppm.
12a(80): Dansyl chloride (1.50 mg, 5.00 mmol) in CH2Cl2 (2 mL) was
added dropwise to a solution of 12a(80) (15 mg, 6.71 mmol) and Et3N
(21 mL, 12 equiv) in CH2Cl2 (4 mL) at À308C. The mixture was stirred at
À108C for 2 h. The reaction was then quenched with MeOH. The solvent
was evaporated and the polymer was purified by dissolving it in CHCl3
followed by precipitation into EtOAc/hexane (1:2), and also by dissolving
it in DMSO followed by precipitation into ethyl ether. This procedure
was carried out several times until the bluish color on the TLC plate had
disappeared. Compound 12a(80) (11 mg) was produced as a greenish
solid.
Poly
N
N
propyl}-5-{3-[3-(3-{benzyloxyamino}propyl)-5-(3-{tert-butyloxyamino}pro-
pyl)benzoylamino]propyl}benzoylamino)propyl]benzylmethacrylamide}
(12a): A solution of 7b (0.25 g, 1.5 equiv per amine group) in CH2Cl2
(6 mL) at À208C was added dropwise to a solution of 11b (20 mg,
0.055 mmol) and Et3N (0.05 mL) in MeOH/CH2Cl2 (5 mL, 2:1). The mix-
ture was stirred at room temperature for 48 h. The solvent was evaporat-
ed, another portion of 7b (25 mg) and Et3N (0.05 mL) in CH2Cl2 (3 mL)
was added, and the mixture was stirred for a further 12 h. The polymeric
product was washed with brine, the solvent was evaporated, and the resi-
due was purified three times by dissolving it in CHCl3 followed by pre-
cipitation into EtOAc/hexane (1:1) until excess 7b was completely re-
moved. This furnished 12a (0.13 g, 74%) as a colorless solid. 1H NMR
(CDCl3): d=1.35 (br; CH3), 1.65 (br; CH2), 2.35 (br; ArCH2), 2.89 (br;
NHCH2), 4.85 (br; CH2), 6.21 (br; NH), 6.94 (br; ArH), 7.11 (br; ArH),
7.28 ppm (br; ArH); 13C NMR (CDCl3): d=28.23, 31.42, 32.95, 40.58,
66.69, 79.29, 125.27, 128.03, 128.25, 128.73, 132.08, 135.05, 137.15, 142.41,
157.06, 157.50, 168.86 ppm; GPC (DMF with 1% LiBr, 808C, versus pol-
ystyrene standard): Mn =3.9”106, Mw =12.6”106, PDI=3.2.
12a(100): Dansyl chloride (2.20 mg, 7.68 mmol) in CH2Cl2 (2 mL) was
T
added dropwise to a solution of 12a (10 mg, 3.84 mmol) and Et3N (10 mL,
excess) in CH2Cl2 (2 mL) at À308C. The mixture was stirred at À108C
for 2 h. The reaction was then quenched with MeOH. The solvent was
evaporated and the polymer was purified by dissolving it in CHCl3 fol-
lowed by precipitation into EtOAc/hexane (1:2), and also by dissolving it
in DMSO followed by precipitation into ethyl ether. This procedure was
carried out several times until the bluish color on the TLC plate had dis-
appeared. Compound 12a(100) (8 mg) was produced as a slightly gray
G
solid.
Poly
N
N
propyl}-5-{3-[3-(3-{benzyloxyamino}propyl)-5-(3-{amino}propyl)benzoyl-
amino]propyl}benzoylamino)propyl]benzylmethacrylamide·2CF3COOH}
(12b): A solution of CF3COOH (0.11 mL, excess) in CH2Cl2 (2 mL) at
08C was slowly added to a solution of 12a (50 mg, 0.019 mmol) in
CH2Cl2 (3 mL). The mixture was stirred for 14 h and then the solvent
was evaporated at room temperature. The polymer was purified by dis-
Acknowledgements
This work was financially supported by the German Science Foundation
(DFG, grant number Sfb 448, TP A11) and ETH startup funds. We cor-
6550
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 6542 – 6551

Dendronized Polymers
FULL PAPER
dially thank Prof. Afang Zhang, Zhenzhou University and ETH Zürich,
for his continuous interest in and support of this work, Dr. Heinz Rüeg-
ger for competent help with the NMR analyses, Prof. Peter Walde for
help with the fluorescence measurements, and Martin Colussi for the
molar mass determinations.
[5] N. Severin, J. Barner, A. A. Kalachev, J. P. Rabe, Nano Lett. 2004, 4,
577–579.
[6] a) S. Fuchs, H. Otto, S. Jehle, P. Henklein, A. D. Schlüter, Chem.
Commun. 2005, 1830–1832; b) S. Fuchs, T. Kapp, H. Otto, T. Schç-
neberg, P. Franke, R. Gust, A. D. Schlüter, Chem. Eur. J. 2004, 10,
1167–1192, and references therein.
[7] For example, see: a) N. Ouali, S. MØry, A. Skoulios, Macromolecules
2000, 33, 6185–6193; b) A. Zhang, B. Zhang, E. Wächtersbach, M.
Schmidt, A. D. Schlüter, Chem. Eur. J. 2003, 9, 6083–6092; c) M.
Malkoch, A. Carlmark, A. Woldegiorgis, A. Hult, E. E. Malmstrçm,
Macromolecules 2004, 37, 322–329; d) A. Luebert, T. Q. Nguyen, F.
Sun, S. S. Sheiko, H.-A. Klok, Macromolecules 2005, 38, 2064–2071;
e) J. L. Mynar, T.-L. Choi, M. Yoshida, V. Kim, C. J. Hawker, J. M. J.
FrØchet, Chem. Commun. 2005, 5169–5171.
[8] R. Al-Hellani, A. D. Schlüter, unpublished results.
[9] L. Shu, A. D. Schlüter, C. Ecker, N. Severin, J. P. Rabe, Angew.
Chem. 2001, 113, 4802–4805; Angew. Chem. Int. Ed. 2001, 40, 4666–
4669.
[1] a) A. D. Schlüter, J. P. Rabe, Angew. Chem. 2000, 112, 860–881;
Angew. Chem. Int. Ed. 2000, 39, 864–883; b) A. Zhang, L. Shu, Z.
Bo, A. D. Schlüter, Macromol. Chem. Phys. 2003, 204, 328–339;
c) A. D. Schlüter, Top. Curr. Chem. 2005, 245, 151–191; d) H.
Frauenrath, Prog. Polym. Sci. 2005, 30, 325–384.
[2] J. Barner, F. Mallwitz, L. Shu, A. D. Schlüter, J. P. Rabe, Angew.
Chem. 2003, 115, 1976–1979; Angew. Chem. Int. Ed. 2003, 42, 1932–
1935.
[3] S. W. Hla, G. Meyer, K.-H. Rieder, ChemPhysChem 2001, 2, 361–
366.
[4] We assume that in previously described experiments (refs. [2,9]) im-
purities also formed an ultrathin organic coating on HOPG. Given
the complexity and the unspecific host capability of higher genera-
tion denpols, the presence of impurities stemming from solvents and
reagents used during synthesis cannot be excluded even if several
purification steps are applied. Carefully dried denpols give correct
data in combustion analysis. This, however, is insufficient evidence
to propose that there are no impurities still being entrapped in the
branched layer around the backbone. The amount of material re-
quired to form a monolayer on a solid substrate is so small that it
cannot possibly be detected by combustion analysis. The experi-
ments described in references [2] and [9] were carried out with den-
pols prepared on the solid substrate directly from the reaction mix-
ture. It is therefore necessary to assume that products from the final
azidification step have interfered in one way or another with both
the adsorption to and lateral diffusion of the denpols on the surface.
A main advantage of the experiments described here is the fact that
they can be carried out under ambient conditions. The price to be
paid for this, and thus a further complication, is the unclear situation
with adsorptions of compounds contained in the surrounding air
prior to the application of the denpols.
[10] L. Shu, I. Gçssl, J. P. Rabe, A. D. Schlüter, Macromol. Chem. Phys.
2002, 203, 2540–2550.
[11] The yield for this compound by mistake was given as 95%: S.
Müller, A. D. Schlüter, Chem. Eur. J. 2005, 11, 5589–5610. The fully
reproducible value is 65%, which was obtained in runs with up to
25 g of product.
[12] R. Klopsch, S. Koch, A. D. Schlüter, Eur. J. Org. Chem. 1998, 1275–
1283.
[13] For example, see: L. Shu, A. D. Schlüter, Macromol. Chem. Phys.
2000, 201, 239–245, and references therein.
[14] During the synthesis of model compound 2c it was observed that if
the reaction was performed at 08C, the amines were doubly dansy-
lated to a small extent (NMR spectroscopy). This is why, under the
more forceful conditions, the temperature was not elevated above
À108C.
[15] C. Bçttcher, B. Schade, C. Ecker, J. P. Rabe, L. Shu, A. D. Schlüter,
Chem. Eur. J. 2005, 11, 2923–2928.
Received: February 7, 2006
Published online: June 26, 2006
Chem. Eur. J. 2006, 12, 6542 – 6551
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6551