Controlled hydrogel fiber formation: The unique case of hexaphenylbenzene-poly(ethylene glycol) amphiphiles

Article abstract of DOI:10.1002/smll.201302832

Bundles of fibers of hexaphenylbenzene-PEG derivatives in aqueous solutions are inferred from polarized and depolarized light scattering. The water uptake behavior and the self-assembled structures can be controlled by changing the length of the PEG chain

Full text of DOI:10.1002/smll.201302832

communications
Hydrogel Fibers
Controlled Hydrogel Fiber Formation: The Unique Case of
Hexaphenylbenzene-Poly(ethylene glycol) Amphiphiles
Katrin Wunderlich, Antje Larsen, John Marakis, George Fytas,* Markus Klapper,*
and Klaus Müllen
Fiber formation in nature occurs not only at surfaces but also present study, hexaphenylbenzene (HPB) was chosen as
in bulk from large molecules such as collagen. The relation the large hydrophobic section and PEG as hydrophilic part
between water and biological macromolecules is one of the of the amphiphile on the account of their large antagonistic
most important phenomena of all life processes. Connective forces.[14] Smaller amphiphiles like sodium dodecyl sulfate or
tissue, for example, depends upon the association with water lutensols with high critical micelle concentration (cmc) form
for the unique and specific mechanical properties. Here water only small aggregates (micelles) while amphiphilic block
is an important component of both the collagen fibers and the copolymers yield kinetically arrested non-equilibrium struc-
polysaccharide gel matrix which comprises the highly organ- tures.[15,16] Additionally, we favored the propeller-like HPB
ized composite structure (structural, bound and free water).[1] over the corresponding dehydrogenated disk-like hexa-peri-
Swelling yields hydrogel fibers, wherein the overall structure hexabenzocoronene (HBC) as hydrophobic building block.
is not destroyed by water uptake. This example shows that The latter case is already known for forming extended col-
nature is able to activate weak intermolecular interactions umns in solution, however, due to the strong π–π interactions
(hydrogen bonds, van der Waals forces) to control fiber forma- between the extended polyaromatic units[17] a thermody-
tion.[2] The control of the interaction strength is decisive for namic control of the aggregation process cannot be guar-
the formation of biological fibers by self-assembly processes.[3] anteed. Furthermore, the shape of the PEG functionalized
The strength should be high enough to form stable structures HPB can be easily changed by varying the number and chain
under physiological (aqueous) conditions but also sufficiently length of the solubilizing group without altering the frac-
low to yield defined thermodynamic equilibrium structures.
tion of the solubilizing PEG units. We define the shape on
In fiber formation processes, chemists use either physical Israelachvili’s packing parameter.[9] Our definition is based
methods such as extruding,[4] microfluidic processes,[5] elec- on the different shapes of the molecules with a random coil
trospinning[6] or also molecular self-assembly, i.e., formation of PEG in which we assume that the cone becomes broader
of discrete architectures from molecules via intermolecular with more arms at the same position (schematic representa-
forces, thereby competing with nature.[7–11] However, a dense tion of Figure 1).[18]
packing inside the core and the shell is always assumed and
In this concept, we anticipate that, depending on the
possible partition of water was ignored. Hydrogel fibers molecular shape and the length of the PEG chains (i.e.,
can be formed by synthetic methods as well, however, they density of the hydrophilic group in the periphery of HPB,
require a chemical crosslinking. For example, extruded fibers Figure 1), space filling will control water uptake by the
of poly(vinyl alcohol) and poly(acrylic acid) can be converted formed hydrogel fibers. In particular, we have synthe-
to bundles of hydrogel fibers after a crosslinking by ester for- sized amphiphilic compounds based on HPB with two dif-
mation upon heating.[12,13]
ferent PEG substitution patterns and investigated their
We attempt to mimic the concepts of nature to form syn- self-assembly behavior in water. The characterization of the
thetic hydrogel fibers via self-assembly by carefully adjusting resulting supramolecular assemblies imposes additional chal-
the molecular interactions of different amphiphiles. In the lenges considering their complexity (size, shape and dynamic
nature) and the inevitable extremely dilute conditions, neces-
sary to avoid interactions between them. Performing dynamic
polarized and depolarized light scattering, we assessed both
the form factor and the two transport (translation, rotation)
coefficients of the assemblies necessary to uniquely reveal
their complex structures. Based on the characteristic param-
eters (the number of molecules within the fibers NS, length
Lw, diameter d, length density M/L), the supramolecular
structure can be represented by bundles of fibers. Their size
and water content relate to the structure and amphiphilicity
of 1 and 2; both vary with the number and length of the PEG
chains (Figure 1).
K. Wunderlich, Dr. M. Klapper, Prof. Dr. K. Müllen
Max Planck Institute for Polymer Research
Ackermannweg 10 55128, Mainz, Germany
E-mail: klapper@mpip-mainz.mpg.de
A. Larsen, J. Marakis, Prof. Dr. G. Fytas
Department of Materials Science and Technology
University of Crete and F.O.R.T.H.
P.O. Box 1527, 71110, Heraklion, Greece
E-mail: fytas@mpip-mainz.mpg.de
DOI: 10.1002/smll.201302832
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1914 wileyonlinelibrary.com
small 2014, 10, No. 10, 1914–1919

Hydrogel Fiber Formation via Hexaphenylbenzene-Poly(ethylene glycol) Amphiphiles
OCH3
Br
Br
O
(a)
(b)
OCH3
5
3
4
OH
O
O
PEG-OCH3
(c)
(d)
PEG-OCH3
OH
6
1
PEG-OCH3
O
O
O
PEG-OCH3
PEG-OCH3
PEG-OCH3
O
O
PEG-OCH3
O
PEG-OCH3
1 (PEG-750)
2 (PEG-380)
Figure 1. Synthetic scheme for 1. (a) 1,2-bis(4-bromophenyl)ethyne, diphenyl ether, microwave: 220 °C, 300 W, 12 h (45%); (b) (4-methoxyphenyl)
boronic acid, Pd(OAc)2, Cs2CO3, DABCO, DMF, 80 °C, 3d (65%); (c) BBr3, DCM, 0 °C to rt (84%); (d) CH3O-PEG-Br, KOH, THF, reflux, 18 h. Molecular
structure and the schematic representation of 1 (PEG-750) and 2 (PEG-380).
We synthesized the title compounds (Figure 1) length of worm-like micelles of sample 1 displays in the range
through
extrusion
enone
to
a
Diels-Alder reaction with subsequent CO of ∼5 and ∼350 nm with a diameter of ∼3 nm.
of 2,3,4,5-tetraphenylcyclopenta-2,4-di- The cmc values of these samples as estimated from the
and 1,2-bis(4-bromophenyl)ethyne led change of the surface tension with increasing concentra-
3′,4′,5′,6′-tetraphenyl-1,1′:2′,1′′- tion are indeed very low. Sample 2 possesses a cmc value of
(3)
4,4″-dibromo-
terphenyl (4). 4,4′′′′-Dimethoxy-3′′,4′′,5′′,6′′-tetraphenyl- 0.002 g L−1 which is even lower than for sample 1 (Figure S1,
1,1′:4′,1″:2″,1′′′:4′′′,1′′′′-quinquephenyl (5) was prepared by Supporting Information). The latter seems to display two
palladium-catalyzed Suzuki coupling reaction with 4-meth- crossover concentrations between 0.003 and 0.01 g L−1. There-
oxyphenyl boronic acid. The following deprotection of the fore, the cryo-TEM image (Figure 2), obtained at concentra-
methoxyphenyl groups and etherification of HPB con- tions (∼30 g L−1) several orders of magnitude above cmc, might
taining two hydroxyphenyl groups (6) with α-methoxy-ω- render any prediction of the structures at much lower concen-
bromo-PEG (PEG: Mn = 750 g mol−1) gave 1 (Figure 1); trations (near cmc) ambiguous. The surface tension profile
while 2 was synthesized under similar conditions.
of 2 is similar to the surface tension profile of conventional
The self-assembly behavior of hexaphenylbenzene con- ionic surfactant such as sodium dodecyl sulfate (SDS).[19] The
taining PEG molecules with different architecture (1 and measured cmc values are clearly lower than the cmc of ionic
2) was explored in aqueous solution by cryo-TEM. For the surfactants which assume values in the range of g L−1 (cmcSDS
aqueous solution of 1 at 30.7g L−1, long self-assembled struc- = 2.65 g L−1).[20] Very small cmc, comparable to the value of 2,
tures are observed with a worm-like structure (Figure 2). The occur for high molecular weight block copolymers. However,
www.small-journal.com
1915
small 2014, 10, No. 10, 1914–1919
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

communications
K. Wunderlich et al.
Figure 2. Cryo-TEM micrograph of 1 in aqueous solution at 30.7 g L−1. The white arrow shows worm-like structures and the black arrow shows a
bundle of fibers.
these values are obtained from air/water interface measure- supramolecular assemblies in the aqueous solutions become
ments and might deviate for the bulk solution. The performed important; their structural characterization is meaningful at
dynamic light scattering experiment at different concentra- c<c*. The reduced Rayleigh ratio RVV (q→0)/c for the samples
tions utilized the photon correlation spectroscopy (PCS) 1 and 2 obtained from the extrapolation of the q-dependent
technique. This method, having the necessary spatiotemporal RVV (q) at q = 0 (Figure 3) represents the reciprocal of the
resolution, has reconciled all available structural information osmotic pressure of the solutions, and hence its concentration
of the self-assembly behavior of 1 and 2.
dependence significantly reflects interactions between sol-
We have initiated the PCS experiment with a concentra- utes.[21] This diagram identifies two regimes with a crossover
tion of 1.7 g L−1 that is between cmc and the concentration c* roughly around 0.1 g L−1 for 1 with the dilute regime, iden-
used for the cryo-TEM images. We first examined aqueous tified by RVV (q→0)∼c, below this concentration. Above c*
solutions of 1 prepared at 20 °C and also recorded the two relaxation processes are present in CVV(q,t) (Figure S2)
experimental relaxation function Cvv(q,t) also at 20 °C at a with intensity contributions (circles and squares) (Figure 3);
high scattering wave vector q = 0.03 nm−1 corresponding to a this crossover concentration is larger than the two values for
probing length of about 200 nm (open squares in Figure S2). 1 (Figure S1 and arrows in Figure 3). Above about this c*, two
Dynamics of the order of 100 µs cannot be rationalized by processes are also observed in the CVV(q,t) for 2 (Figure S3)
the small size of the structures of 1 and 2 and clearly implies but a c-independent RVV (q→0)/c is reached at much lower
formation of supramolecular assemblies. The broad shape concentrations (<0.002 g L−1) in agreement with its cmc value
of the relaxation function describing the diffusion dynamics (red arrow in Figure 3). However, the limiting RVV (q→0)/c
of the moieties present in the solution was analyzed by values for both 1 and 2 do not correspond to the molecular
ILT (Equation 1a, SI). Two populations of assemblies are species (Figure 1) but to much larger structures as indicated
revealed which unexpectedly change with annealing at 20 °C
over a period of about weeks. To speed up this equilibration
procedure and warrant time invariant structures, we have
annealed the solutions at 50 °C for about 3 h with subsequent
slow cooling to 20 °C.
The recorded Cvv(q,t) (blue solid rhombi) is different
from the relaxation function before the particular annealing
protocol was employed as both peaks of ILT shift to shorter
times with heating and the stronger impact is on the ampli-
tude and position of the slow process. This peculiar meta-
stability effect is not observed for the aqueous solutions of
2 at similar concentration (inset of Figure S2). The depicted
CVV(q,t) recorded as-dissolved (black open squares) and
after annealing at 50 °C (blue open rhombi) display experi-
mentally the same shape as it is reflected more in the cor-
responding ILT profiles. The origin of kinetically trapped
structures in relatively dilute concentrations of the present
amphiphiles due to the formation of thermodynamically
unfavorable configurations is unclear and needs further
investigation. For the elucidation of the formed structures,
however, it suffices to know the pathway towards equilibrium
of self-assembled structures.
Figure 3. Reduced absolute scattering intensity Rvv (q→0)/c is shown
as function of concentration, c, at 20 °C for 1 and 2. The shaded area
indicates a crossover to semidilute regime in which two relaxation
processes are resolved in Cvv(q,t). Circles and squares denote the Rvv
The second unexpected observation relates to the very low
overlap concentration c* above which interactions between the are for the CMC values from Figure S1.
intensities associated with the main and fast processes. The two arrows
www.small-journal.com
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1916
small 2014, 10, No. 10, 1914–1919

Hydrogel Fiber Formation via Hexaphenylbenzene-Poly(ethylene glycol) Amphiphiles
by the estimated molecular weight Mw = RVV (q→0)/Hc. hardly applicable and Cor(q,t) is still measurable owing to the
Using for the contrast factor H = [2πn(dn/dc)]2/(NA λ4) = optical anisotropy of the phenyl rings in the parent 1 and 2
5.5·10−7 cm2 mol g−2, with NA being the Avogadro number, Mw structures.
amounts to 1.3·108 g mol−1 and 1.8·109 g mol−1 for the assem-
Since the relaxation function CVV(q,t) is unimodal, there
blies in 1 and 2 aqueous solutions at c ∼0.001 g L−1 below cmc, is a single contribution to the static Rvv(q) which at such
respectively. The latter obtained from surface tension meas- dilute conditions is safely obtained from the total intensity
urements (Figure S1) does not signify a transition to molec- and the short amplitude of CVV(q,t). The pattern of Rvv(q) is
ular species and should therefore be considered with caution. represented well by a worm-like shape[23–25] using an average
Conversely, the variation of RVV (q→0)/c with concentration contour length L ∼ 3 µm and Kuhn segment lK ∼ 75 nm
(Figure 3) defines the dilute regime in which supramolecular (solid line in the Holtzer plot of Figure 4). Further, since
assemblies free of mutual interactions should be investigated. the thickness, d, of the chain is ignored, the applicability of
Consistently, the two different regimes are well documented this model in the light scattering experiment presumes thin
by the concentration dependence of the translation diffusion enough (qd>>1) structures; this assumption is justified by the
coefficient(s) in Figure S4.
representation of the two diffusion coefficients in Figure S5.
We seek also for information on the dynamic properties The large contour length implies large aggregation number
of these assemblies to reduce ambiguities in the structural NS, already suggested by the high molecular weight value
characterization of complex systems. Figure 4 is an example (1.3·108g mol−1) of the supramolecular structures suggesting
of such information source including the form factor (static) ∼6·104 1-monomers per assembly. Consistency between this
and transport coefficients (dynamic) for the self-assembly of number and the value of the contour length (L) implies a
1 in a dilute aqueous solution. Both translational CVV(q,t) micellar structure of the “repeat” unit.
and orientation Cor(q,t) relaxation functions are described
Complementary information on the internal structure of
by a single decay process ascribed to a single population of the supramolecular assembly stems from the length density
supramolecular assemblies whose average dimension can be M/L = qRvv(q)/(πHc) ∼6·1011 g mol−1 cm−1 obtained from the
still captured by the longest probing wavelength, as seen by quasi-plateau qRvv(q)/c value at high q’s (inset to Figure 4).
the intensity downturn in the Holtzer presentation, qRvv(q)/c For comparison, this quantity for the “monomer” 1 is about
versus q (inset to Figure 4).[22] Moreover, the limiting value of 3.8·108 g mol−1 cm−1 using 2300 g mol−1 for its molecular mass
orientation relaxation rate Γr at q = 0 is nonzero (upper inset) and 0.6 nm for its thickness.[26] Hence its repeat unit should
that allows the determination of the rotational diffusion Dr = contain approximately 16 monomer units.
Γr (q = 0)/6 in addition to the translation diffusion D = Γ(q)/
As already mentioned, uncertainties of the dynamic
q2 at q = 0.[23] We should note that at this extremely low con- parameters in the determination of self-assembled structures
centration other scattering (X-ray or neutron) techniques are are minimized if a consistent representation of the dynamic
parameters is also obtained. Both trans-
port coefficients, D0 and Dr can be rep-
resented by a worm-like model[23,27] with
three adjustable parameters, Lw, Kuhn
segment, lk (lk = 2 lp, the persistence
length), and thickness, d assuming stick
hydrodynamic boundary conditions. The
speed-up of the two transport coefficients
with flexibility L/lp for constant contour
length Lw and d, and their expected slow-
down, though with different rate, with
increasing L and d is shown (Figure S5).
A simultaneous representation of both D0
and Dr (skew and vertically hatched boxes
in Figure S5) of the 1 assembly occurs for
Lw ∼ 2.2 µm, L/lp ∼70, d ∼30 nm, i.e., lp
∼ 30 nm in agreement with the represen-
tation of the Rvv(q) pattern in Figure 4.
Considering the presence of size polydis-
persity[28,29] and the different moments of
the distribution entering the static and
dynamic experiments, the dimensions and
conformation of the supramolecular struc-
Figure 4. Normalized field correlation functions Cvv(q,t) (open circles) for the translational
ture in dilute aqueous solutions of 1 can
diffusion dynamics and Cor(q,t) (open rhombi) for the orientation dynamics at a scattering
be adequately estimated.
wave vector q = 0.0081 nm−1 and 20 °C for 1 at 0.007 g L−1 recorded after annealing at T
= 50 °C for few hours and subsequent slow cooling to 20 °C. Inset: The single translational
Turning to the structure characteriza-
tion for 2, the experimental situation is
and rotational relaxation rate, and the Holtzer plot qRvv(q) vs q. Solid line denotes the
representation by a worm-like model with a persistence length lk = 65 10 nm.
less convenient due to the lack of access
www.small-journal.com
1917
small 2014, 10, No. 10, 1914–1919
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

communications
K. Wunderlich et al.
Structural parameters
1
2
Mw [g mol-1]
M/L [g mol-1 cm-1]
Lw [µm]
1.3 ×108
1.8 ×109
7 ×1012
6.0
11
6 ×10
2.25
70
L/ lp
150
Figure 5. Structural parameters for 1 and 2 self-assembled structures in aqueous solution obtained from dynamic and static experiment. Schematic
representation of the proposed self-assembled equilibrium structures of 1 and 2 conforming to the structural parameters of the table.
to all previous physical quantities. Figure S6, the analogue of about 0.2 g·cm3 and 0.7 g·cm3 for 1 and 2, respectively. This
Figure 4 for 1, shows that at the concentration of 2.5 10−4 g L−1 disparity can be rationalized by the difference in swelling of
the assemblies are too large to be captured by the longest light short and long (degree of polymerization N>12) PEG chains
scattering wavelength, i.e., absence of downturn in the Holtzer in water.[30] It is already reported in literature that the thick-
plot. Consistently, the rotational Dr is too low, i.e., the intercept ness and swelling behavior from monolayers of PEG chains
Γr (q = 0) is not finite (insets to Figure S6). The information is are strongly dependent on the length of the PEG chains. For
then limited to the translation coefficient D0 and the length example, an increase of the degree of polymerization of the
density M/L = qRvv(q)/(πHc) ∼7·1012 g mol−1 cm−1 obtained PEG chains from 3 to 15 leads to structural changes and the
from the quasi-plateau qRvv(q)/c value at high q’s in Figure S6 nature of the PEG chains changes from dense packing to a
which clearly exceeds the corresponding plateau in Figure 4. swollen state. We assume a similar behavior for our fibers
The large M/L along with the shape of the Holtzer plot sug- formed by the molecule 1 or 2. These amphiphiles can be
gests high molecular mass. A rough estimate of the latter can considered core-shell systems in which the core is formed by
be obtained from the RVV (q→0)/c obtained from the inter- HPB and the shell by PEG. Supported by the results obtained
cept in the Debye plot which predicts linear dependence of for layers, we deduce that the PEG shell of an isolated fiber
[RVV (q→0)]1/2 versus q2 at low q’s; Mw = RVV (q→0)/Hc is of 1, which has arms with a polymerization degree of 17, is
about 10 times higher than for 1. The description of D0 and more swollen than the shell of a fiber of 2 having only 8 eth-
Dr∼0 (skew and vertically red hatched boxes in Figure S5) by ylene glycol units per arm. This denser packing in the PEG
the worm-like model suggests larger dimensions: L ∼ 6.0 µm, shell observed for molecule 2 is further enhanced by the dou-
L/lp ∼150, d ∼30 nm, i.e., lp ∼ 40 nm which implies similar bling of the number of PEG chains per building block from 1
worm-like structure as for the self-assembly of 1.
to 2 (illustrated in Figure 5).
The characteristic lengths summarized in the table of
Comparing the overall diameter of the bundles of 1 and
Figure 5 support the proposed supramolecular structures of 1 2, the diameter is in both cases 30 nm and therefore inde-
and 2 in dilute aqueous solutions as schematically illustrated pendent of the shape of the molecule and the PEG chain
in the same Figure. As the diameter of the worm-like struc- length. Surprisingly, the persistence length of the supramo-
tures of 1 largely exceeds the fully extended molecular size lecular structures appears to be virtually unaffected by the
(8.9 nm and 5.1 nm for 1 and 2 according to to Corey-Pauling- number of molecules per layer (much higher for 2) which
Koltun model) a columnar structure with a single molecule do, however, affect the overall fiber length. Therefore, we
per layer, in which PEG protects the hydrophobic core can conclude that the persistence length is controlled less by
be excluded. In fact, each layer contains large number of the PEG chains than by the columnar stacking of HPB
molecules accounting for the high M/L values, which is dif- (Figure 5).
ferent for the two structures. According to the hydrophilic/
In conclusion, HPB-PEG derivatives form bundles of
hydrophobic interactions and Percec’s tobacco mosaic virus hydrogel fibers in very dilute aqueous solution. Remarkably,
model the HPB-PEG derivatives can self-assemble to small the water amount in the hydrogel fibers does not depend on
fibers.[7,8]
the number of hydrophilic groups but on the length of the
Further, we assume that several fibers are bundled to a PEG chains and the substitution pattern of the precursor
single large supermolecular object (Figure 5) confirmed by molecules. Molecule 2 with the shorter PEG chains led to
cryo-TEM at high concentration (Figure 2). Superstruc- longer self-assembled structures containing less water than
tures of bundled fibers were reported for amphiphilic block the fibers formed by the longer PEG chains. We assume that
copolymers of oligo(phenylene vinylene) and poly(propylene the denser packing of the four PEG chains in 2 hampers
oxide).[13]
the water uptake, while 1 has only two PEG arms that can
The assemblies of 2 have a higher density than 1. In 2, easily swell in an aqueous phase. The hydrogel fibers with the
the shorter PEG chains can accommodate more molecules in variable water uptake resemble nature systems unlike fiber
the same cross-section. Based on the table in Figure 5, the formation processes e.g., extruding, microfluidic processes,
superstructures of 1 and 2 with a volume V = πd2L/4 and and electrospinning. Similar to natural systems, antagonistic
mass Mw/NA (NA is the Avogadro number) have a density of forces (hydrogen bonds, van der Waals forces) are responsible
www.small-journal.com
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1918
small 2014, 10, No. 10, 1914–1919

Hydrogel Fiber Formation via Hexaphenylbenzene-Poly(ethylene glycol) Amphiphiles
[6] A. Hsieh, T. Zahir, Y. Lapitsky, B. Amsden, W. Wan, M. S. Shoichet,
Soft Matter 2010, 6, 2227–2237.
[7] S. D. Hudson, H.-T. Jung, V. Percec, W.-D. Cho, G. Johansson,
G. Ungar, V. S. K. Balagurusami, Science 1997, 278, 449–452.
[8] V. Percec, J. Heck, G. Johansson, D. Tomazos, Macromol. Symp.
1994, 77, 237–265.
for the formation of well-defined objects. It is demonstrated
that in water swellable fibers can be created utilizing hydro-
phobic and hydrophilic interactions. These interactions have
to be carefully tuned to allow for a thermodynamic control
which we showed by the comparison between amphiphiles
containing HPB and HBC as hydrophobic building blocks.
Due to the strong interaction among the HBC units the
yielded self-assembly becomes kinetically controlled as only
undefined aggregates are obtained by TEM. Herein, we dem-
onstrate an analogue to the fiber formation process in nature
by using a self-assembly process to yield hydrogel fibers. Sim-
ilar to nature, our case was controlled by different molecular
interactions. However, the main difference is that nature
achieves such extended hydrogel fibers by the self-assembly
of large polymers such as polycarbohydrates or polypeptides,
while we focused on “small” two-dimensional amphiphiles.
[9] J. N. Israelachvili, Intermolecular and Surface Forces 1991,
291–292.
[10] T. Aida, E. W. Meijer, S. I. Stupp, Science 2012, 335, 813–817.
[11] P. Y. W. Dankers, T. M. Hermans, T. W. Baughman,
Y. Kamikawa, R. E. Kieltyka, M. M. C. Bastings, H. M. Janssen,
N. A. J. M. Sommerdijk, A. Larsen, Marja J. A. van Luyn,
A. W. Bosman, E. R. Popa, G. Fytas, E. W. Meijer, Adv. Mater. 2012,
24, 2703–2709.
[12] J. Fei, Z. Zhang, L. Zhong, L. Gu, J. Appl. Polym. Sci. 2002, 85,
2423–2430.
[13] H. Wang, W. You, P. Jiang, L. Yu, H. H. Wang, Chem. Eur. J. 2004,
10, 986–993.
[14] M. Muthukumar, C. K. Ober, E. L. Thomas, Science 1997, 277,
1225–1232.
[15] G. Riess, Prog. Polym. Sci. 2003, 28, 1107–1170.
[16] L.-I. Atanase, G. Riess, Polym. Int. 2011, 60, 1563–1573.
[17] L. Schmidt-Mende, A. Fechtenkötter, K. Müllen, E. Moons,
R. H. Friend, J. D. MacKenzie, Science 2001, 293, 1119–1122.
[18] M. Johnsson, P. Hansson, K. Edwards, J. Phys. Chem. B 2001, 105,
8420–8430.
[19] B. Jönsson, B. Lindman, K. Holmberg, B. Kronberg, Surfactants
and Polymers in Aqueous Solution, Wiley: New York 1998, 255.
[20] H. Kölbel, P. Kurzendörfer, Konstitution und Eigenschaften von
Tensiden. Fortschritte der chemischen Forschung Bd. 12, 2, 1969,
Springer Verlag, Berlin.
[21] M. Bockstaller, W. Koehler, G. Wegner, G. Fytas, Macromolecules
2001, 34, 6353–5358.
[22] A. Holtzer, J. Polym. Sci. 1955, 17, 432–434.
[23] A. Kroeger, J. Belack, A. Larsen, G. Fytas, G. Wegner, Macromol-
ecules 2006, 39, 7098–7106.
[24] R. J. Koyama, J. Phys. Soc. Jpn. 1973, 34, 1029–1038.
[25] A. Kroeger, V. Deimende, J. Belack, I. Lieberwirth, G. Fytas,
G. Wegner, Macromolecules 2007, 40, 105–115.
[26] a) L. Jiménez-García, A. Kaltbeitzel, W. Pisula, J. S. Gutmann,
M. Klapper, K Müllen, Angew. Chem. 2009, 121, 10135–10138;
b) J. García, A. Kaltbeitzel, W. Pisula, J. S. Gutmann, M. Klapper,
K. Müllen, Angew. Chem. Int. Ed. 2009, 48, 9951–9953.
[27] H. Yamakawa, M. Fujii, Macromolecules 1973, 6, 407–415.
[28] S. Bantle, M. Schmidt, W. Burchard, Macromolecules 1982, 15,
1604–1609.
Supporting Information
Supporting Information is available from the Wiley Online Library
or from the author.
Acknowledgements
The authors acknowledge financial support from the European
Commission under the Seventh Framework Program by means of
the grant agreement for the Integrated Infrastructure Initiative N.
262348 European Soft Matter Infrastructure (ESMI) and GSRT in
the framework of the program THALIS(Metaassembly). The authors
thank the Stiftung Stipendien-Fonds des Verbandes der Chemis-
chen Industrie e.V. for funding.
[1] S. Nomura, A. Hiltner, J. B. Lando, E. Baer, Biopolymers 1977, 16,
231–246.
[2] E. L. Bakota, O. Sensoy, B. Ozgur, M. Sayar, J. D. Hartgerink, Bio-
macromolecules 2013, 14, 1370–1378.
[3] S. Zhang, Biotechnol. Adv. 2002, 20, 321–339.
[4] F. Jianqi, G. Lixia, Eur. Polym. J. 2002, 38, 1653–1658.
[5] H. Onoe, R. Gojo, K. Kuribayashi-Shigetomi, S. Takeuchi, in 25th
IEEE International Conference on Micro Electro Mechanical Sys-
tems (MEMS) 2012, pp. 1065–1068.
[29] S. R. Aragon, R. Pecora, J. Chem. Phys. 1976, 64, 2395–2404.
[30] G. Mathe, C. Gege, K. R. Neumaier, R. R. Schmidt, E. Sackmann,
Langmuir 2000, 16, 3835–3845.
Received: August 30, 2013
Revised: December 6, 2013
Published online: February 25, 2014
www.small-journal.com
1919
small 2014, 10, No. 10, 1914–1919
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim