Complete self-assembly of discrete supramolecular dendrimers

Article abstract of DOI:10.1002/anie.200462104

Simply mixing a core (black) with defined amounts of branching units (red) and end caps (blue) leads to programmed complete self-assembly of supramolecular dendrimers through complimentary and multiple hydrogen bonding.

Full text of DOI:10.1002/anie.200462104

Communications
Self-Assembly
A total of 3 ꢀ 2nÀ3 molecules of 2 (n = generation
number) links the core 1 with end caps 3 in the supramolec-
ular dendrimer. When components 1, 2, and 3 are mixed in a
Complete Self-Assembly of Discrete
Supramolecular Dendrimers**
Alexander Franz, Walter Bauer,
and Andreas Hirsch*
The complete self-assembly of den-
drimers based on noncovalent
interactions presents a great chal-
lenge. By analogy with supramolec-
ular polymers,[1–4] the reversible and
thermodynamically controlled link-
age of the basic building blocks
leads to dynamic materials which
are able to react to external stimuli
and which are able to undergo self-
healing processes. However, the
hitherto presented concepts for
the self-assembly of discrete and
purely
organic
dendrimers
employed preassembled dendritic
systems in most cases. These enti-
ties have been linked to form
discrete superstructures, for exam-
ple, 1) by intermolecular hydrogen
Figure 1. Supramolecular assembly of discrete dendrimers by using the homotritopic core 1, heterotri-
topic AB2 block 2, and end caps 3.
bonds with each other[5–7] or with other molecules,[8,9] 2) by
electrostatic interactions between oppositely charged units,[10]
and 3) as a result of structural and hydrophobic effects in the
case of amphiphilic dendrimers.[11]
We here present the first approach for the complete self-
assembly of discrete supramolecular dendrimers where the
repetition motif is no longer based on dendritic subunits
(Figure 1). We employ the homotritopic Hamilton receptor
1,[2,12] which can form six hydrogen bonds to cyanuric and
barbituric acid derivatives. We have developed the hetero-
tritopic AB2 unit 2 as a branching element, which contains two
Hamilton receptors as well as a complimentary cyanuric acid
substrate.
The synthesis of 2 started with 4, which was converted into
the the corresponding acid chloride and treated with 5 to give
compound 6. After cleavage of the ester with KOH, treatment
with 0.5 equivalents of 8 led to ether formation in a twofold SN
reaction (Scheme 1). Steglich coupling of the cyanuric acid
derivative 11 with compound 10 (deprotected by trifluroacetic
acid) leads to the AB2 branching unit 2 in good yields.
[*] Dipl.-Chem. A. Franz, Prof. Dr. W. Bauer, Prof. Dr. A. Hirsch
Institut fꢀr Organische Chemie
der Friedrich-Alexander-Universitꢁt Erlangen-Nꢀrnberg
Henkestrasse 42, 91054 Erlangen (Germany)
Fax: (+49)9131-852-6864
E-mail: andreas.hirsch@chemie.uni-erlangen.de
[**] This work was supported by the Fonds der Chemischen Industrie.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
1564
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200462104
Angew. Chem. Int. Ed. 2005, 44, 1564 –1567

Angewandte
Chemie
Scheme 1. Synthesis of the branching element 2. Bz=benzoyl,
DCC=dicyclohexylcarbodiimide, DMAP=4-dimethylaminopyridine,
HOBt=1-hydroxybenzotriazole.
1:(3 ꢀ 2nÀ3):(3 ꢀ 2n) ratio, complete self-assembly of a supra-
molecular dendrimer of generation n can be expected,
provided that all possible hydrogen bonds are formed
(Figure 1).
For comparison, we first investigated the formation of
1(EC)3 aggregates, in which the core 1 is complexed directly
by three dendritic end caps 3a–3e (Scheme 2).[13,14] To
achieve this, end caps 3a–3e were added by titration to core
Scheme 2. Aggregates 1(EC)3 with different end caps EC=3a–e.
1
1. During this process the characteristic H NMR chemical
shifts of the NH protons to higher frequency (lower field)
were monitored.[2] From the titration curves the association
constants of complexes 1(EC)3 were derived by using a
method by Solovꢁev et al.[15] and by using the computer
program Chem-Equili (Table 1). The values obtained are in
good agreement with literature data for comparable sys-
tems.[2] The Scatchard plots[16] of these data are indicative of
pronounced cooperativity in the formation of 1(EC)3. Due to
the increase in size, the aggregation yielding dendritic
complexes 1(EC)3 results in shorter retention times tSEC
relative to those of free 1 and 3a–3e in size exclusion
chromatography experiments. Plots of the glass transition
temperatures Tg versus ne MÀ1 (ne = number of benzyl end
groups, M = molecular mass) give a linear relationship for
dendrons 3b, 3d, and 3e, as well as for the corresponding
supramolecular dendrimers 1(EC)3 (Table 1, Figure 2). This
behavior is in complete agreement with the chain end–free
volume theory for dendrimers, which is expressed as Tg =
Tg¥ÀK’(ne MÀ1). On the other hand, the observed coincidence
proves the existence of 1(EC)3 as discrete dendrimers.
Table 1: Physical properties of core 1 (C), end caps 3a–e (EC), complexes
1(EC)3, and the AB2 unit 2.
Tg [K][a] ne[b] d [nm][c] logK1
[lmolÀ1 [f]
logK2
logK3
tSEC
]
[lmolÀ1 [g]
]
[lmolÀ1 [h]
]
[min][i]
1
3a
3b
3c
3d
3e
1(3a)3
1(3c)3
418.66
322.98
310.20
324.98
317.64
321.72
–
–
2
2
4
4
8
6
0.82[d]
1.38[e]
2.14[d]
1.70[e]
0.66[d]
0.67[d]
–
–
–
–
–
–
–
1.66
4.21
3.82
4.46
3.51
–
–
–
–
–
–
–
1.29
4.71
2.60
4.46
4.51
–
–
–
–
–
–
–
9.74
10.03
10.66
10.46
8.75
–
8.05
7.22
7.96
8.18
–
26.73
7.48
8.99
6.33
4.09
–
–
12 2.66[d]
1(3b)3 312.60
6
–
1(3d)3 316.04 12 2.38[d]
1(3e)3 318.34 24
407.07
–
2
–
2.46[e]
9.73
[a] Tg =glass transition temperature. [b] ne =number of terminal benzyl
groups. [c] d=particle diameter (according to PFG NMR). [d] In CDCl3.
[e] In [D6]DMSO. [f] K1: C+ECQC(EC). [g] K2: C(EC)+ECQC(EC)2.
[h] K3: C(EC)2 +ECQC(EC)3. [i] tSEC =retention time in size-exclusion
chromatography (SEC) with MN Nucleogel GPC 500-5.
Now we focus on the systems described in Figure 1 and
formed by the stoichiometric mixing of core unit 1, branching
Angew. Chem. Int. Ed. 2005, 44, 1564 –1567
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1565

Communications
Figure 2. Plot of the glass transition temperature Tg vs. ne MÀ1, the
quotient of the number of terminal benzyl groups and the molar mass,
for the isolated end caps 3b, 3d, and 3e, and for the complexes
Figure 3. Section of the 1H PFG NMR spectrum of a mixture of 1, 2,
and 3d in the G3 stoichiometry. The appearance of several discrete
structures 3.4, 7.6, and 13.4 nm in diameter is evident. In addition,
larger particles y are found, whose size cannot be determined due to
unfavorable signal-to-noise ratio. This also holds for aggregates x.
1(EC)3, and the supramolecular dendrimer 1·2(3ꢂ2nÀ3)·3d(3ꢂ2n)
.
element 2, and end cap 3d in a ratio of 1:(3 ꢀ 2nÀ3):(3 ꢀ 2n).
DSC measurements of the AB2 unit 2 afford a high glass
transition temperature, indicative of aggregation to polymeric
structures. As a result of the addition of core 1 and end cap
3d, the polymer is disrupted, and a perfectly linear relation-
ship between the glass transition temperature Tg and the
ne MÀ1 ratio (and, thus, the generation number n) results
(Figure 2). This demonstrates that discrete supramolecular
dendrimers must be present in the solid state.
Due to the complexity of the associated exchange
processes, the 1H NMR shift of the NH proton signals
during the titration of components 1, 2, and end cap 3a–3e
may no longer be employed to determine the various
association constants. However, when one examines the
determined equilibrium constants for the 1(EC)3 complexes
(Table 1) one can assume that exchange processes will take
place in solution. Nonetheless, by using pulsed field gradient
(PFG) NMR spectroscopy we proved the existence of discrete
dendrimers in chloroform. This method can be used to
determine the diffusion constants of the dendrimers and thus,
in turn, the dendrimer sizes. We employed the intensity
decrease of the benzyl proton signals of core unit 1 as a probe.
Components 1, 2, and 3d were dissolved in CDCl3, in
stoichiometric amounts corresponding to dendrimers of the
first, third, and fifth generation. For the G1 system 1·23·3d6,
aggregates of uniform size and with a diameter of 2.0 nm were
found.[17] This particle size was confirmed by dynamic light-
scattering experiments (3.12 nm). For complete self assembly
of the dendrimers of lower generation, the required stoichio-
metric ratio of branching unit 2 and end caps 3a–3e differs
more than for higher generations (where a 1:1 limit is
approached). Thus, it can be assumed that comprehensive
discrimination will become increasingly difficult. This expect-
ation has been verified. Figure 3 shows the 1H NMR spectrum
Figure 4. AFM images of supramolecular dendrimers with G1, G3, and G5 stoichiometry adsorbed on a Si wafer. Note the uniform structures
with 4 nm (height profile)in the micrograph for G1. In addition, note the perfect globular shape of the higher generation aggregates.
1566
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2005, 44, 1564 –1567

Angewandte
Chemie
of a mixture of components 1, 2, and 3d in a stoichiometry
corresponding to the G3 dendrimer. Besides the G3 den-
drimer 1·221·3d24 itself (3.4 nm), higher dendrimers are
observed in low concentration (Figure 3). Similar observa-
tions are made for stoichiometries that correspond to even
larger values of n. Nonetheless, all NMR experiments confirm
unequivocally that discrete dendrimers are formed. Branch-
ing unit 2 exhibits very broad 1H NMR signals in chloroform.
However, these signals become sharp exactly at the point
where core 1 and end cap 3 are added in amounts that
correspond to a ratio of 1:(3 ꢀ 2nÀ3):(3 ꢀ 2n) (1:2:3), which is
the stoichiometry of a dendritic assembly. This indicates the
spontaneous formation of supramolecular dendrimers, in
complete agreement with DSC results.
The existence of defined dendrimers is further corrobo-
rated by AFM investigations. After adsorption of the G1
dendrimer 1·23·3d6 on a Si surface, uniform and widespread
distribution of completely homogeneous objects is observed
with a diameter of 4 nm (Figure 4, left ), which is in agreement
with the expected size of 1·23·3d6. Repeated sample prepara-
tions gave completely analogous pictures. In no case were
objects of different sizes observed. Similar preparations of
solutions of higher generations such as 1·221·3d24 (G3) and
1·293·3d96 (G5) showed in a reproducible way that in addition
to small discrete objects with diameters of 4, 7, and 12 nm
larger objects are formed. The vast majority of these
structures have diameters of 24, 28, and 32 nm (height
profiles). These objects are too large to be attributed to
isolated dendrimers. Presumably, these objects are clusters of
higher generation dendrimers formed during an incomplete
drying process in the course of the sample preparation. The
discontinuity in the size distribution of aggregates is also a
striking feature for the higher generation dendrimers. To
summarize, the investigations presented here prove for the
first time the programmed self-assembly of discrete supra-
molecular dendrimers from nondendritic units.
Keywords: dendrimers · hydrogen bonds · scanning probe
microscopy · self-assembly · supramolecular chemistry
.
[1] L. Brunsfeld, B. J. B. Folmer, E. W. Meijer, R. P. Sijbesma,
Chem. Rev. 2001, 101, 4071.
[2] V. Berl, M. Schmutz, M. J. Krische, R. G. Khoury, J.-M. Lehn,
Chem. Eur. J. 2002, 8, 1227.
[3] J.-M. Lehn, Supramolecular Polymers (Ed.: A. Cifferi), Marcel
Dekker, New York, 2000, p. 615.
[4] R. F. M. Lange, M. van Gurp, E. W. Meijer, J. Polym. Sci. Polym.
Chem. Ed. 1999, 37, 3657.
[5] Y. M. Ma, S. V. Kolotuchin, S. C. Zimmerman, J. Am. Chem. Soc.
2002, 124, 13757.
[6] P. S. Corbin, L. J. Lawless, Z. Li, Y. Ma, M. J. Witmer, S. C.
Zimmerman, Proc. Natl. Acad. USA 2002, 8, 5099.
[7] D. K. Smith, F. Diederich, Chem. Eur. J. 1998, 4, 1353.
[8] W. T. S. Huck, R. Hulst, P. Timmerman, F. C. J. M. van Veggel,
D. N. Reinhoudt, Angew. Chem. 1997, 109, 1046; Angew. Chem.
Int. Ed. Engl. 1997, 36, 1006.
[9] A. Dirksen, U. Hahn, F. Schwanke, M. Nieger, J. N. H. Reek, F.
Vꢂgtle, L. De Cola, Chem. Eur. J. 2004, 10, 2036.
[10] B. Zhishan, L. Zhang, Z. Wang, X. Zhang, J. Shen, Mater. Sci.
Eng. C 1999, 10, 165.
[11] M. Kellermann, W. Bauer, A. Hirsch, B. Schade, K. Ludwig, C.
Bꢂttcher, Angew. Chem. 2004, 116, 3019; Angew. Chem. Int. Ed.
2004, 43, 2959.
[12] S. K. Chang, A. D. Hamilton, J. Am. Chem. Soc. 1988, 110, 1318.
[13] Complete syntheses, characterizations, and supramolecular
properties of all newly presented units will be described in a
detailed forthcoming paper.
[14] A similar system with a non-dendritic end cap was recently
analyzed by Lehn and co-workers in context with investigations
on supramolecular polymers (cf. ref. [2]).
[15] V. P. Solovꢁev, V. E. Baulin, N. N. Strahova, V. P. Kazachenko,
V. K. Belsky, A. A. Varnek, T. A. Volkova, G. Wipff, J. Chem.
Soc. Perkin Trans. 2 1998, 1489.
[16] B. Perlmutter-Hayman, Acc. Chem. Res. 1986, 19, 90.
[17] The determination of particle sizes relies on the Stokes–Einstein
relationship for globular particles. Dendrimers of low generation
actually exist as disc-shaped structures. However, this method
determined an average molecule size over all directions which is
smaller than the true main diameter. Nonetheless, uniformity
can be demonstrated clearly.
Received: September 24, 2004
Published online: January 28, 2005
Angew. Chem. Int. Ed. 2005, 44, 1564 –1567
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1567