A new family of nonionic dendritic amphiphiles displaying unexpected packing parameters in micellar assemblies

Article abstract of DOI:10.1021/ja101523v

In this paper we report on the synthesis of a new family of nonionic dendritic amphiphiles that self-assemble into defined supramolecular aggregates. Our approach is based on a modular architecture consisting of different generations of hydrophilic polygl

Full text of DOI:10.1021/ja101523v

Published on Web 07/22/2010
A New Family of Nonionic Dendritic Amphiphiles Displaying
Unexpected Packing Parameters in Micellar Assemblies
Britta Trappmann,†,‡,§ Kai Ludwig,§,| Michał R. Radowski,† Anuj Shukla,‡
Andreas Mohr,† Heinz Rehage,‡ Christoph Bo¨ttcher,*,| and Rainer Haag*,†
Institut fu¨r Chemie und Biochemie, Freie UniVersita¨t Berlin, Takustrasse 3,
14195 Berlin, Germany, Physikalische Chemie II, UniVersita¨t Dortmund, Otto-Hahn-Strasse 6,
44227 Dortmund, Germany, and Forschungszentrum fu¨r Elektronenmikroskopie, Institut fu¨r
Chemie und Biochemie, Freie UniVersita¨t Berlin, Fabeckstrasse 36a, 14195 Berlin, Germany
Received March 1, 2010; E-mail: haag@chemie.fu-berlin.de; christoph.boettcher@fzem.fu-berlin.de
Abstract: In this paper we report on the synthesis of a new family of nonionic dendritic amphiphiles that
self-assemble into defined supramolecular aggregates. Our approach is based on a modular architecture
consisting of different generations of hydrophilic polyglycerol dendrons [G1-G3] connected to hydrophobic
C11 or C16 alkyl chains via mono- or biaromatic spacers, respectively. All amphiphiles complex hydrophobic
compounds as demonstrated by solubilization of Nile Red or pyrene. The structure of the supramolecular
assemblies as well as the aggregation numbers are strongly influenced by the type of the dendritic
headgroup. While the [G1] amphiphiles form different structures such as ringlike and fiberlike micelles, the
[G2] and [G3] derivatives aggregate toward spherical micelles of low polydispersity clearly proven by
transmission electron microscopy (TEM) measurements. In the case of the biaromatic [G2] derivative, the
structural persistence of the micelles allowed a three-dimensional structure determination from the TEM
data and confirmed the aggregation number obtained by static light scattering (SLS) measurements. On
the basis of these data, molecular packing geometries indicate a drastic mass deficit of alkyl chains in the
hydrophobic core volume of spherical micelles. It is noteworthy that these highly defined micelles contain
as little as 15 molecules and possess up to 74% empty space. This behavior is unexpected as it is very
different from classical detergent micelles such as sodium dodecyl sulfate (SDS), where the hydrophobic
core volume is completely filled by alkyl chains.
Introduction
remaining challenge is the generation of structurally persistent
aggregates3 that act as stable nanocarriers for hydrophobic drugs
Nonionic polymeric amphiphiles have attracted special inter-
est due to their tendency to form very stable aggregates and
are therefore useful for the efficient solubilization of active
agents, for example, in drug delivery.1 Among them are many
biocompatible poly(ethylene glycol)- (PEG-) based surfactants,
such as Cremophor, Pluronics, and Tween, which have been
extensively used to formulate hydrophobic drugs.2 Advantages
over ionic amphiphiles include their pH independency and the
generation of nontoxic nanoparticulate aggregates, which pro-
vide a great benefit for drug delivery. However, in contrast to
charged amphiphiles, which can assemble into defined aggre-
gates,3-5 nonionic surfactants have been previously reported
to lead to diverse and ill-defined micelles.1a Therefore, a
and dyes. Because of their low polydispersity and their tunable
size, dendritic structures are promising for the construction of
new amphiphiles. Grinstaff and co-workers6 have synthesized
dendritic amphiphiles from glycerol, succinic acid, and myristic
acid. By varying the hydrophobic-to-hydrophilic ratio and
including ionic headgroups, a wide range of aqueous aggregation
behavior could be studied but no defined structures were
observed. Furthermore, Percec et al. and also Lee and co-
workers7 have shown that aromatic units can drastically enhance
the stability of supramolecular assemblies. Depending on the
nature of the incorporated hydrophobic branch, hollow as-
(5) Schade, B.; Ludwig, K.; Bo¨ttcher, C.; Hartnagel, U.; Hirsch, A. Angew.
Chem., Int. Ed. 2007, 46, 4393–4396.
† Institut fu¨r Chemie und Biochemie, Freie Universita¨t Berlin.
‡ Universita¨t Dortmund.
(6) (a) Luman, N. R.; Grinstaff, M. W. Org. Lett. 2005, 7, 4863–4866.
(b) Meyers, S. R.; Juhn, F. S.; Griset, A. P.; Luman, N. R.; Grinstaff,
M. W. J. Am. Chem. Soc. 2008, 130, 14444–14445.
§ B.T. and K.L. contributed equally.
| Forschungszentrum fu¨r Elektronenmikroskopie, Institut fu¨r Chemie und
Biochemie, Freie Universita¨t Berlin.
(7) (a) Percec, V.; Peterca, M.; Dulcey, A. E.; Imam, M. R.; Hudson,
S. D.; Nummelin, S.; Adelman, P.; Heiney, P. A. J. Am. Chem. Soc.
2008, 130, 13079–13094. (b) Kim, J.-K.; Lee, E.; Huang, Z.; Lee, M.
J. Am. Chem. Soc. 2006, 128, 14022–14023. (c) Percec, V.; Dulcey,
A. E.; Balagurusamy, V. S. K.; Miura, Y.; Smidrkal, J.; Peterca, M.;
Nummelin, S.; Edlund, U.; Hudson, S. D.; Heiney, P. A.; Duan, H.;
Magonov, S. N.; Vinogrador, S. A. Nature 2004, 430, 764–768. (d)
van Dongen, S. F. M.; De Hoog, H.-P. M.; Peters, R. J. R. W.; Nallani,
M.; Nolte, R. J. M.; van Hest, J. C. M. Chem. ReV. 2009, 109, 6212–
6274. (e) Rosen, B. M.; Wilson, C. J.; Wilson, D. A.; Peterca, M.;
Imam, M. R.; Percec, V. Chem. ReV. 2009, 109, 6275–6540.
(1) (a) Schick, M. J. Nonionic Surfactants: Physical Chemistry; CRC Press:
Boca Raton, FL, 1987. (b) Haag, R. Angew. Chem., Int. Ed. 2004, 43,
278–282.
(2) Kabanov, A. V.; Alakhov, V. Y. Crit. ReV. Ther. Drug Carrier Syst.
2002, 19, 1–72.
(3) Kellermann, M.; Bauer, W.; Hirsch, A.; Schade, B.; Ludwig, K.;
Bo¨ttcher, C. Angew. Chem., Int. Ed. 2004, 43, 2959–2962.
(4) Burghardt, S.; Hirsch, A.; Schade, B.; Ludwig, K.; Bo¨ttcher, C. Angew.
Chem., Int. Ed. 2005, 44, 2976–2979.
9
10.1021/ja101523v  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 11119–11124 11119

A R T I C L E S
Trappmann et al.
Scheme 1. Synthesis, Structure, and Aggregation of Aromatic Dendritic Amphiphilesa
a PG represents different generations of polyglycerol dendrons. (i) 5 mol % CuSO4, 10 mol % ascorbic acid, 30 mol % NaOH, THF/H20 1:1, rt, 2-5 days;
(ii) Lewatit K1131, MeOH, 24 h, rt. Umbrella-shaped molecular geometry of spacious G2 headgroups induces high-curvature spherical micelles, whereas
less voluminous G1 dendrons promote linear growth toward fibrous assemblies. This behavior is independent of the aromatic spacer and the alkyl chain
length used in this study (cf. Table 1).
semblies or long cylindrical micelles can be formed. Recently,
we have also observed the strong influence of aromatic units
on the solubilization of hydrophobic guest molecules within
dendritic core-shell architectures.8 Therefore, our structural
design for this study was to combine nonionic dendritic and
aromatic units with long alkyl tails in order to control the
formation of structurally defined and stable micelles with good
transport properties for hydrophobic guest molecules.
Herein we report on a modular approach for the synthesis of
a new family of nonionic dendritic amphiphiles. This approach
is based on a modular architecture consisting of different
generations of polyglycerol dendrons [G1-G3] as hydrophilic
parts connected to C16 or C11 alkyl chain as hydrophobic units
via mono- or biaromatic spacers, respectively (Scheme 1).
The dendrimers showed self-assembly into different supramo-
lecular architectures depending on their molecular structure with
unexpected packing parameters. The aim of this study is to gain
insight into the molecular factors governing self-assembly and
to survey the transport behavior for hydrophobic guest molecules
in water. Surprisingly, we found that the dendritic headgroups
are capable of imposing the formation of structural persistent
assemblies, which is the first example to be proven for nonionic
amphiphiles.
Results and Discussion
We designed a modular approach that allows for a facile
synthesis of structurally varied compounds represented by the
amphiphiles 1-7, making use of “click” coupling as a highly
reliable step for scale-up of the process. This study is focused
on preselected dendritic amphiphiles that differ in the size of
the hydrophilic headgroup by using different generations of
polyglycerol dendrons.9
As starting materials, we chose acetal-protected [G1] to [G3]
polyglycerol dendron azides as hydrophilic building blocks and
propionic acid alkyl ester or 4-ethynylbenzoic acid alkyl ester
as hydophobic units. The synthesis of all dendritic polyglycerol
monoazides was carried out via a previously reported proce-
dure.9 The acetal protecting groups were removed in the last
step to obtain the active amphiphiles 1-7 (Scheme 1).
At first we studied the aggregation behavior of our new
dendritic nonionic amphiphiles in detail using different physical
(8) Tu¨rk, H.; Shukla, A.; Rodrigues, P. C. A.; Rehage, H.; Haag, R.
Chem.sEur. J. 2007, 13, 4187–4196.
(9) Wyszogrodzka, M.; Haag, R. Chem.sEur. J. 2008, 14, 9202–9214.
9
11120 J. AM. CHEM. SOC. VOL. 132, NO. 32, 2010

Self-Assembling Nonionic Dendritic Amphiphiles
A R T I C L E S
Table 1. Critical Micelle Concentrations, Aggregation Numbers
and Hydrodynamic Diameters,a and Diametersb of Dendritic
Amphiphiles 1-7
dynamic light scattering data and mean diameters obtained from
TEM measurements are highly consistent in the case of the [G2]
and [G3] derivatives 2, 3, 5, and 7 as we observed uniform
spherical micellar aggregates for them (for a detailed analysis
see below).
ST + SLS
aggregation DLS hydrodynamic
TEM
diameter (nm)
compd
cmc (M)
numberc
diameterd (nm)
[G1] derivatives 1, 4, and 6 behave differently, which is
already apparent by their poor solubility in water. For example,
compound 1 can only be dissolved in water at higher temper-
atures (above 70 °C) where an almost clear solution (a faint
opacity is retained) is obtained. Upon cooling to room temper-
ature, the solution becomes viscous but remains clear for at least
∼12 h before visible precipitation occurs. By repeated cryo-
TEM measurements we observed that multiple heating-cooling
cycles of the sample affect the appearance of the aggregate
structure, particularly with regard to the aggregate shape and
size (Figure 1A). From the first cycle, predominantly fibrous
micelles with a uniform diameter of ∼8 nm were obtained. The
length of the fibers varied from 15 nm up to several hundred
nanometers. In later cycles, smaller micellar aggregates of ∼14
nm diameter became dominant (Figure 1B). Tilt experiments
showing the assemblies at different angles (see Supporting
Information) confirmed the impression that the micellar orga-
nization is of ringlike geometry. Figure 1B shows both fibers
as well as ringlike micelles within one single micrograph.
However, the heterogeneity of the aggregates explains the
average aggregate size of 11.5 nm obtained from DLS data.
We therefore conclude that the scattering data and the deter-
mination of the aggregation numbers are not reliable in this case.
The [G1] derivatives 4 and 6 show a similar behavior even
though the solubility is slightly better due to the shorter
hydrophobic building block. Fibers and spherical micelles can
also be found here (Figure 1A,C).
Most interestingly, spherical micelles formed by amphiphile
2 showed a well-resolved internal structure upon closer inspec-
tion of the electron microscopic data (Figure 1, bottom row). It
was supposed that the differences in the ultrastructural pattern
represent different spatial views of the assemblies and might
be used to determine their three-dimensional organization, as
has been demonstrated earlier for structurally persistent micelles
of different ionic calixarene and fullerene dendrimers.3-5 By
use of the cryo-negative staining sample preparation procedure,11
the contrast of these structural details can be enhanced (Figure
1C). Image processing techniques were applied and alignment
and classification algorithms were used to reconstruct the three-
dimensional volume of the micelle. The obtained three-
dimensional structure determined for [G2] micelles is shown
in a surface representation in Figure 2. The micelle is character-
ized by D3 symmetry and shows a specific arrangement of 20
equal high-density patches distributed over the outer micellar
corona. The interior core volume thought to accommodate the
hydrophobic alkyl chains appears hollow. This is an expected
phenomenon as the reconstruction is based on averaging
methods that consider only structurally identical aggregate
features. The alkyl chains, however, are known to be highly
flexible (fluidlike) and therefore do not contribute structural
information to the reconstruction.
1 [G1] 1.5 × 10-5
2 [G2] 1.1 × 10-5
3 [G3] 7.1 × 10-6
4 [G1] 3.7 × 10-5
5 [G2] 9.0 × 10-6
6 [G1] 1.4 × 10-4
7 [G2] 6.4 × 10-4
a From light scattering data. b From TEM data. c Error values were
based on linear regression analysis and determined in the context of the
Excel RGP function. d Hydrodynamic diameters of spherical micelles
(values of intensity mean peak distribution). Values of error result from
three consecutive measurements of the same sample, and the
corresponding standard deviations are given. Deviations between two
measurements performed over a time lag of 6 months were within the
given error margin. e Due to the formation of fiber aggregates of
different length, measurement of the hydrodynamic diameter and
determination of the aggregation numbers for [G1] derivatives became
nonapplicable.
nae
60 ( 2
21 ( 2
nae
nae
8.1 ( 0.4
7.2 ( 0.4
nae
8.0 ( 1.0 (fibers)
8.5 ( 1.0
6.5 ( 1.0
7.7 ( 1.0 (fibers)
6.8 ( 2.0
5.7 ( 1.0 (fibers)
5.2 ( 2.0
42 ( 2
7.3 ( 0.4
nae
nae
15 ( 2
5.3 ( 0.3
characterization methods such as surface tension measurements,
static and dynamic light scattering (SLS and DLS), cryogenic
transmission electron microscopy (cryo-TEM), and image
processing techniques. Subsequently the transport capacities of
the amphiphilic aggregates were tested.
Surface tension measurements were performed in a pendant
drop apparatus. Surface tension data were plotted as a loga-
rithmic function of the surfactant concentration, where a break
in the curve occurs at the critical micelle concentration (cmc).
Thus, the influence of the PG generations, the size of the
aromatic spacer, and the length of the alkyl chains were
compared (Scheme 1, Table 1).
The cmc for the biaromatic compounds 1, 2, and 3 differing
in the size of the dendritic headgroup are 1.5 × 10-5, 1.1 ×
10-5, and 7.1 × 10-6 M, respectively. This trend shows that a
larger degree of hydrophilicity leads to lower cmc values. A
similar observation has already been made by Eisenberg and
co-workers10 for amphiphilic block copolymers composed of
styrene and sodium acrylate, where the cmc decreases with
increasing size of the hydrophilic unit. The shortening of the
alkyl chain length from C16 to C11 chains also shows a significant
impact on the cmc if we compare, for example, the correspond-
ing values of 5 (9.0 × 10-6 M) and 7 (6.4 × 10-4 M),
respectively. Here the larger degree of hydrophobicity leads to
a significantly lower cmc. However, the shortening of the aryl
spacer (although considered to be an integral part of the
hydrophobic segment) has obviously no significant effect on
the cmc if we compare the [G2] derivatives 2 and 5 with bi-
and monoaromatic spacers, respectively. Here the impact of the
alkyl chain might be dominating.
The aggregation of the amphiphiles was confirmed by light
scattering data, which allowed determination of the hydrody-
namic diameters as well as the aggregation numbers (Table 1).
In order to complement the aggregation phenomena by direct
structural data, cryo-TEM was performed for aqueous sample
solutions at a general concentration of 1.0 × 10-3 M, which is
well above the compounds’ corresponding cmc values. It can
clearly be seen that hydrodynamic diameters determined from
If the reconstructed micelle is viewed along the 3- and 2-fold
axes, respectively, the slightly oblate geometry becomes clearly
visible. Here, the shorter dimension (7.8 nm) of the micelle
corresponds perfectly to a bilayer arrangement of the am-
(10) Astafieva, I.; Zhong, X. F.; Eisenberg, A. Macromolecules 1993, 26,
(11) De Carlo, S.; El-Bez, C.; Alvarez-Rua, C.; Borge, J. J. Struct. Biol.
2002, 138, 216–226.
7339–7352.
9
J. AM. CHEM. SOC. VOL. 132, NO. 32, 2010 11121

A R T I C L E S
Trappmann et al.
Figure 1. (A) Cryo-negative stain and (B) cryo-TEM image of an aqueous solution of the dendritic amphiphile 1 ([G1] derivative). Fiberlike and ringlike
micelles in different orientations are clearly visible. Tilting experiments confirm the ringlike character of the micelles (see Supporting Information). Fiber
assemblies have a diameter of about 8 nm, which roughly corresponds to a molecular bilayer, whereas diameters of the ring micelles are on the order of up
to 14 nm, depending on the micelles’ spatial orientation. (C) Cryo-negative stain and (D) cryo-TEM micrograph showing the vitrified sample of amphiphile
2 at a concentration of 2.5 mg mL-1. Spherical micelles of 8-9 nm in diameter are clearly visible. Differences in shape and size are due to different spatial
orientation of the micelles in the frozen sample volume. (Bottom row) Ultrastructural details of the micelles can be extracted by using the cryo-negative
staining data in combination with image processing procedures. The gallery shows a selection of class sum images representing different spatial orientations
of the micelles.
phiphiles. The maximum extension of the micelle, however,
amounts to 8.8 nm. The resolution value of 1.4 nm that we
achieved by evaluating a data set of ∼8000 individual particles
of the [G2] derivative raises the question of what does the data
tell us on the molecular level. With the aggregation number of
60 molecules per micelle from the scattering data, we should
accommodate three molecules in each of the 20 high-density
patches constituting the reconstructed micelle. Figure 2B shows
an arrangement of three [G2] amphiphiles in close association
with a corresponding surface representation limited at a resolu-
tion value of 1.4 nm (to meet the information content of the
reconstruction) by applying adequate filter parameters (Gaussian
filtered by use of SITUS software). It is obvious that the three
dendritic headgroups generate a triangular cone, although the
typical features of the dendritic peculiarity are blurred. It is
noticeable (and in accordance with the reconstruction) that the
hydrophobic alkyl chains are barely visible at this resolution
level. However, the triangular cones fit perfectly into the three-
dimensional reconstruction (see Figure 2C), so that we end up
with 20 × 3 ) 60 molecules forming the reconstructed micelle.
In all cases where the hydrodynamic diameters were deter-
mined for spherical micelles by dynamic light scattering
measurements, we observed very good agreement with the mean
diameters obtained from the TEM data. A general trend is
indicated in the examined series of compounds. By increasing
the spatial demand of the headgroup, the aggregation number
and hence the micellar diameter decreases. This trend endures
if the aromatic spacer (4 and 5) or the alkyl chains (6 and 7)
are shortened. Spherical micelles of 5 [again the [G1] derivative
4 forms fiberlike aggregates (Figure 1a in Supporting Informa-
tion) like 1] have a smaller diameter, which can be expected
because of the shortening of the spacer. Here, the higher
curvature permits a smaller aggregation number of about 42
amphiphiles. If we assume analogous packing conditions as were
observed in the case of the biaromatic compound 2, we would
expect accurately this value. The aggregates of the biaromatic
[G2] amphiphile 2, however, turned out to be a unique case
where the structural persistence allowed for a consistent three-
dimensional reconstruction. Structural identity of all assemblies
(8700 randomly selected assemblies were used in the reported
case here) is a precondition for a successful application of the
approach, which method includes a multivariate statistical
analysis of the data set (see also Supporting Information for
details). We suspect that the dimensions of alkyl chains, spacer,
and headgroup geometry constitute a unique balance in the case
of 2, which made the formation of structurally persistent3
aggregates possible. In all other cases (compounds 3, 5, and
7), this criterion is not fulfilled, which became apparent from
the multivariate statistical analysis and classification involved
in the image processing procedure showing that the assemblies
9
11122 J. AM. CHEM. SOC. VOL. 132, NO. 32, 2010

Self-Assembling Nonionic Dendritic Amphiphiles
A R T I C L E S
analysis is needed to provide a clear picture. Larger aggregates
in the 30-150 nm range, however, are also detectable by TEM
and appear as unstructured agglomerates (see Figure 8 in
Supporting Information). Their number is also here very limited
and in the parts per million (ppm) range. Due to the minor
importance of the large particles in the context of our observa-
tions, we abandon them from a detailed discussion.
The general trend in terms of micellar dimensions, however,
is also clearly in line with expectations if the alkyl chain length
is shortened from C16 to C11. Apart from the fiber formation of
the [G1] derivative 6 (note that the diameter of the fibers
decreases in accordance with the shortening of the molecular
length), spherical micelles (proven by TEM) of the [G2]
derivative 7 possess a significantly smaller diameter of about
5.4 nm if compared to the long-chain derivative 5. Consequently,
a lower aggregation number of only 15 amphiphiles is found
by SLS due to the high curvature of the small aggregates.
Irrespective of the exact supramolecular organization of the
amphiphiles in a micelle, it is obvious from the molecular
dimensions alone that there is a clear difference in the spatial
demand between the voluminous dendrons and the linear
hydrophobic chains, which creates a more or less umbrellalike
molecular geometry, particularly in the case of the [G3]
amphiphile (see Scheme 1). The aggregation of such molecules
has been theoretically rationalized by the concept of the
molecular packing parameter introduced by Israelachvili et al.12
This concept is based on geometrical considerations and makes
it possible to predict the shape and size of equilibrium associates.
Therefore, an increase in the headgroup area of amphiphiles
(wedge angle of the dendron) leads to a decrease in the
aggregation number, whereby particle size remains almost
constant. The general trend predicted by the molecular packing
parameter is also valid for our new dendritic amphiphiles,
although a number of structural aspects are surprising. The
shielding of the corresponding aromatic spacer by the head-
groups in the aggregates and thus their spatial separation is
definitely supported by corresponding absorption spectra. We
found, for example, no indication for π-π interactions of the
biaromatic spacer in the case of 2, below and above the cmc,
neither in freshly prepared nor in months old solutions. However,
it appeared very obvious from the umbrella-shaped geometry
of the molecules that the contribution to the hydrophobic volume
(e.g., of 60 molecules in the micelles of 2) would be very
limited, particularly if the molecules are arranged in a spherical
assembly. By simulating the fitting of 60 complete molecules
(headgroups and alkyl chains) into the reconstructed volume,
we notice that the hydrophobic parts do not entirely fill out the
interior volume of the micelle. Volume calculations give a
noticeable low volumetric efficiency of only 24%. This is in
utter difference to “classic” detergent micelles (i.e., SDS). We
have measured an SDS solution and determined an aggregation
number of 58 ( 8, which is in good agreement with the
literature. Sixty SDS molecules at a micelle diameter of 3.8
nm were reported by Turro and Yekta.13 However, volume
calculations reveal that the hydrophobic chains completely
occupy the interior core volume of the SDS micelle.
Figure 2. (A) Three-dimensional reconstruction (surface representation)
in top (left) and side (right) views obtained from the structural data of ∼8000
micelles formed by 2. Bar is 2 nm. The reconstruction is orientated with
respect to the 3-and 2-fold symmetry axes of the inherent D3 point group
symmetry. (B) Trimeric arrangement of [G2] amphiphiles together with a
low-resolution depiction in side- and top-view orientation thought to
represent the building blocks of the micellar arrangement. (C) Arrangement
of the trimeric building blocks shown in panel B within the reconstructed
micellar volume. Altogether 60 molecules can be accommodated within
the micelle, which is in agreement with the aggregation number obtained
from scattering data.
are not strictly monodisperse. Differences in the aggregate size
and structural organization of the molecules prevent the calcula-
tion of a consistent three-dimensional volume, as structural
identity is a precondition for the application of the approach
(see above). We therefore conclude that the aggregates are
polydisperse to some extent.
These findings, however, have to be discussed in the context
of the scattering data. The intensity and volume size distribution
data and plots from micellar aggregates are shown in the
Supporting Information. In the case of compounds 3, 5, and 7,
a relatively broad size distribution is due to the polydispersity
of the assemblies proven by TEM.
From the 3D reconstruction of 2, however, we know that the
micelles are not ideal spherical particles but are flattened in one
direction (compare top and side views in Figure 2). Such
differences in the aggregate geometry obviously result in a slight
broadening of the corresponding size distribution, though the
micelles are highly monodisperse as proven by the image data
processing. One should note that all intensity plots indicate a
second peak at a particle size >100 nm. For small, isotropic
particles the scattering intensity from a spherical particle is
proportional to the size to the sixth power. Therefore, there is
a strong overemphasis of bigger particles in the intensity
distribution.
However, from the volume distributions (Figure 1b in
Supporting Information), it can be inferred that their absolute
numbers must be small. As the polydispersity index (PDI) values
(in the range of 0.28-0.69) appear to represent the average of
a mixture of several species, complementary microscopic
We conclude that the interaction of space-demanding dendritic
polyglycerol headgroups dominates the assembly process very
clearly. Therefore, the special umbrella-shaped geometry of the
(12) (a) Israelachvili, J. N.; Mitchell, D. J.; Ninham, B. W. J. Chem. Soc.,
Faraday Trans. 2 1976, 72, 1525–1568. (b) Kratzat, K.; Finkelmann,
H. Langmuir 1996, 12, 1765–1770.
(13) Turro, N. J.; Yekta, A. J. Am. Chem. Soc. 1978, 100, 5951.
9
J. AM. CHEM. SOC. VOL. 132, NO. 32, 2010 11123

A R T I C L E S
Trappmann et al.
Table 2. Transport Capacities and Efficiencies of Dendritic
Micelles for Solubilizing the Hydrophobic Compounds Nile Red
and Pyrene
on the transport capacity, but the transport efficiency (here
defined as milligrams of guest molecule per gram of amphiphile)
decreases with the increasing molecular weight of the growing
dendron headgroups [G1]-[G3]. Amphiphiles 6 and 7 with
shorter (C11) hydrophobic chains transport much less than all
other amphiphiles with longer (C16) chains. It is noteworthy that
the pure aromatic guest pyrene is solubilized much better (up
to 5 dye molecules per [G2] micelle) than the heterocyclic dye
Nile Red (only up to 2 dye molecules).
As we observed a prominent alteration of the micellar
structure upon uptake of hydrophobic guest molecules in the
case of ionic dendrimers,16 we also analyzed the loaded nonionic
systems by TEM. However, the data revealed no significant size
change upon guest loading, neither with Nile Red nor with
pyrene. This finding might be another indication that the sparsely
populated hydrophobic micellar core offers sufficient space for
an uptake of guest molecules without alteration of the overall
structure.
dendritic molecules does not principally allow for a complete
filling of the interior micelle core. This conclusion is supported
by both characterization methods used (TEM and DLS), as the
micelle diameter (as well as the aggregation number for 2) was
confirmed independently by both techniques. Interestingly, the
above observations are confirmed for all dendrimer generations
in which spherical micelles are formed. For example, if 21
molecules of [G3] amphiphile 3 are accommodated in a spherical
micelle of 7 nm diameter, a similar low volumetric efficiency
of 26% is obtained. Compounds 5 and 7 show similar behavior
with 45% and 32% efficiency, respectively.
One can at present only speculate about the nature of the
micellar core. From plausible reasons we exclude vacuum as
an explanation. An additional filling with water molecules might
be an alternative suggestion, which has been discussed for
classical detergent micelles since the mid-1980s.14 Evidence for
the presence of structured water within the hydrophobic core
of membranes was, for example, reported by Meier et al.15 At
least an alternative molecular organization such as a vesicle can
be ruled out. Although vesicle formation might explain an
explicit water-filled interior volume, there are no structural
indications such as a bilayered membrane. Moreover, when the
voluminous headgroup geometry is taken into account, vesicle
formation is barely possible within the boundary of 5-9 nm
spheres.
In addition to structural studies, the transport properties of
all amphiphiles were investigated by uptake of hydrophobic dyes
such as Nile Red and pyrene in water. To establish a
structure-transport relationship, the quantities of solubilized
Nile Red and pyrene were compared (Table 2). It turns out that
the amphiphiles 1, 2, and 3 with a biaromatic spacer show
relatively high transport capacities (here defined as millimoles
of guest molecule per mole of amphiphile). This is in line with
our previous report on dendritic core-shell architectures based
on hyperbranched polyglycerol-containing biphenyl units for the
efficient transport of hydrophobic compounds.8 In fact, a
comparison of amphiphiles 2 and 5 demonstrates increased
transport capacities with increased number of aromatic units,
with a more than 2-fold increase in solubilized dye when
comparing one and two aromatic units. On the other hand, the
size of the hydrophilic polyglycerol headgroup has little effect
Conclusions
We have developed a new family of dendritic amphiphiles
that undergo controlled aggregation at very low critical micelle
concentrations. All amphiphiles were tested as potential nano-
carriers for hydrophobic compounds and were shown to entrap
the dyes Nile Red and pyrene. Different structural parameters
influencing the transport and aggregation behavior were inves-
tigated. It was shown that the aromatic spacers of the amphiphile
play an important role for the transport of aromatic guest
molecules. Furthermore, the aggregation number is strongly
influenced by the dendritic headgroup, which leads to structur-
ally persistent and highly defined spherical micelles for [G2]
dendrons. Also, the structurally more defined [G2] micelle (60
amphiphiles) shows much better transport efficiencies due to
the higher order compared to other structurally less defined
assemblies.
This is the first example for structurally highly defined
micelles based on nonionic surfactants consisting of as little as
15 dendritic amphiphiles. The most exciting and novel aspect
of our study is the unquestionable alkyl chain volume deficit in
the micelles’ hydrophobic core, which has been proven by two
independent methods. This finding spurs on further studies
characterizing and explaining this supramolecular phenomenon
as well as the synthesis of new dendritic supramolecular
assemblies with unique properties.
Acknowledgment. This work was supported by the SFB 765
from the Deutsche Forschungsgemeinschaft. Also, the Studienstif-
tung des Deutschen Volkes is kindly acknowledged for a scholarship
to B.T. We also thank A. Wiedekind for repititive DLS measure-
ments and W. Mu¨nch for HPLC purification of dendritic amphiphiles.
Supporting Information Available: Experimental Section and
additional text, 9 figures, and one table describing purification
and characterization details of dendritic amphiphiles 1-7, light
scattering data, determination of the transport capacity for Nile
Red, and additional TEM data. This material is available free
of charge via the Internet at http://pubs.acs.org.
(14) (a) Dill, K. A.; Koppel, D. E.; Cantor, R. S.; Dill, J. D.; Bendedouch,
D.; Chen, S. H. Nature 1984, 309, 42–45. (b) Menger, F. M.; Doll,
D. W. J. Am. Chem. Soc. 1984, 106, 1109–1113.
JA101523V
(15) Meier, E. M.; Schummer, D.; Sandhoff, K. Chem. Phys. Lipids 1990,
55 (2), 103–113.
(16) Ja¨ger, C. M.; Hirsch, A.; Schade, B.; Ludwig, K.; Bo¨ttcher, C.; Clark,
T. Langmuir 2010, 26 (13), 10460–10466.
9
11124 J. AM. CHEM. SOC. VOL. 132, NO. 32, 2010