Formation of fibril structures in polymerizable, rod-coil-oligomer-modified epoxy networks

Article abstract of DOI:10.1002/chem.200601049

This paper describes the in situ preparation of fibrils in epoxy networks in which the fibril-like structures are cured polymerizable rod-coil oligomers. The epoxy-terminated α,ω-modified PEO oligomers, which are ABA rod-coil-rod oligomers with a poly(eth

Full text of DOI:10.1002/chem.200601049

DOI: 10.1002/chem.200601049
Formation of Fibril Structures in Polymerizable, Rod–Coil-
A
Oligomer-Modified Epoxy Networks
N
Yingfeng Yu,*[a] Zongyong Gao,[a] Guozhu Zhan,[a] Liang Li,[a] Shanjun Li,*[a]
Wenjun Gan,[b] and James V. Crivello[c]
Abstract: This paper describes the in
situ preparation of fibrils in epoxy net-
works in which the fibril-like structures
are cured polymerizable rod–coil
oligomers. The epoxy-terminated a,w-
modified PEO oligomers, which are
ABA rod–coil–rod oligomers with a
poly(ethylene oxide) coil unit and two
aromatic azomethine liquid-crystalline
rod units, were synthesized and then
further blended with an epoxy precur-
sor. Uniform nanoscale columnar struc-
tures were observed in the neat rod–
coil oligomers as well as in the cross-
linked liquid-crystalline state. During
the curing of the blends, the supra-
molecular nanoscale columnar struc-
tures of the rod–coil oligomers are
transformed into polymeric fibrils
where the epoxy functional end groups
have co-reacted with epoxy precursors
to form a crosslinked network.
Keywords: epoxy networks · liquid
crystals · nanostructures · rod–coil
oligomers · self-assembly
Introduction
lenge in these methods. Furthermore, these methods also
have some drawbacks in terms of either thermal, processing,
or internal-stress properties. For example, rubber modifica-
tion is invariably accompanied by a significant drop in heat
resistance, while thermoplastic and core–shell particle addi-
tives are usually difficult to disperse in precursors and
always sharply increase the viscosity of the blends.
Thermosetting networks tend to have a characteristic low
resistance to brittle fracture, therefore the modification of
these materials with rubbers,[1] thermoplastics,[2] and core–
shell particles[3] has been a significant challenge in the past
few decades. These modifiers are initially miscible with the
uncured thermoset precursors but phase-separate to typical-
ly form spherical structures or bicontinuous structures,
which means that morphological control is a common chal-
Recently, it has been established that the use of block co-
polymers as the toughener can alleviate these processing
concerns because the morphology develops spontaneously
(self-assembles) upon blending with the uncured precursor,
which makes block copolymer modified thermosets attrac-
tive for research purposes. As has been reported by several
authors, the self-assembly characteristics of block copoly-
mers can be maintained in their blends with several homo-
polymers.[4] Numerous reports have shown that novel mor-
phologies in block copolymers, blends, and solutions can be
duplicated in crosslinked systems to give thermosets with in-
teresting properties. In this way, block copolymers are
widely used as templates for generating nanostructured
epoxy or phenolic matrixes with long-range order in both
the uncured and cured states.[5–8] For example, Bates etal.[5]
have studied methacrylic block copolymer modified epoxy
systems and have revealed that these copolymers form
nanoscale structures that are typical for dilute blends of
asymmetric block copolymers with a solvent selective for
the minority block. Like in selective solvents, as the amount
of epoxy added to the block copolymer is increased, the
blend microstructure evolves from lamellar, to gyroid, to
[a] Dr. Y. Yu, Z. Gao, G. Zhan, L. Li, Prof. S. Li
The Key Laboratory of Molecular Engineering of Polymers
Ministry of Education
and Department of Macromolecular Science
Fudan University, Shanghai, 200433 (China)
Fax : (+86)21-6564-0293
E-mail: yfyu@fudan.edu.cn
sjli@fudan.edu.cn
[b] Dr. W. Gan
Department of Macromolecular Materials and Engineering
College of Chemistry and Chemical Engineering
Shanghai University of Engineering Science
Shanghai, 200065 (China)
[c] Prof. J. V. Crivello
Department of Chemistry and Chemical Biology
Rensselaer Polytechnic Institute
1108th Street, Troy, New York 12180 (USA)
2920
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2920 – 2928

FULL PAPER
cylinders, to body-centered-cubic packed spheres, and ulti-
mately disordered micelles.
Results and Discussion
However, previous studies[5] have also shown that block
copolymers might either lower the heat resistance or in-
crease the viscosity of modified systems when the concentra-
tion of copolymers is relatively high. Since liquid-crystalline
thermosets[9] have been used to improve the toughness of
thermosets without sacrificing other properties, we hoped
that polymerizable rod–coil oligomers, which self-assemble
in solution and retain their structures during polymerization,
might be able to combine the advantages of liquid crystals
and copolymers.
It is also well known that self-assembly is a rapid pathway
from small molecules to supramolecular, nanometer-sized
objects that cannot be synthesized by conventional chemical
reactions.[10] This process involves the assembly of discrete
ensembles of molecules, supermolecules, and nanoparticles
into extended ordered arrays, such as crystals, liquid crystals,
and colloids. Block copolymers, especially rod–coil block co-
polymers, represent a unique class where various hierarchi-
cal self-assemblies with different compositions have been
observed experimentally.[11] An unprecedented self-organiza-
tion has been suggested for coil–coil–rod ABC triblock mol-
ecules,[12] which can self-assemble into mushroom-shaped
nanostructures and further organize into noncentrosymmet-
ric polar monolayer structures. Jenekhe has reported the
self-assembly behavior of rod–coil diblock copolymers con-
sisting of poly(phenylquinoline) as the rod block and poly-
styrene as the coil block.[13] Novel aggregates in the form of
hollow spheres, lamellae, hollow cylinders, and vesicles in a
solvent selective for the rod segment were observed. These
results provide potential applications for new functional ma-
terials.
Synthesis and characterization of the rod–coil oligomers: As
rod–coil oligomers containing thermosetting mesogens have
not been studied as extensively as calamitic liquid-crystalline
block copolymers, we decided to synthesize and study a
polymerizable ABA rod–coil–rod oligomer. Previous results
have shown that a nanostructured system can be obtained
by modification of the diglycidyl ether of bisphenol A
(DGEBA) with poly(ethylene oxide)(PEO)-based block co-
G
polymers.[5] Therefore poly(ethylene oxide), which has good
solubility in epoxy resins, was chosen to be the coil unit in
our present work, while an aromatic azomethine unit with
epoxy end groups was chosen to be the rod unit. The latter
is only partly soluble in the uncured epoxy precursor and is
present in the liquid-crystalline phase at elevated tempera-
tures.
The dichloro-terminated PEO oligomers were prepared
by treatment of the corresponding oligo(ethylene glycol)s
with an excess of thionyl chloride in the presence of pyri-
dine and then with 4-hydroxybenzaldehyde to get the alde-
hyde-terminated PEO oligomers (2). As the above synthetic
method is well-known, the yield and purity of the products
were quite good. Next, the ABA-type, a,w-modified PEO
oligomers were synthesized by first coupling 4-(4-hydroxy-
benzylideneimino)aniline (1) with the aldehyde-terminated
PEO oligomers (2) and then functionalizing the resultant
epoxy (3) with a large excess of epichlorohydrin (Scheme 1).
The purity of the obtained product (4) was checked by TLC,
1H NMR spectroscopy, and GPC. Two rod–coil oligomers
were synthesized from PEG 400 and PEG 600 and denoted
LC-400 and LC-600, respectively (see Table 3 in the Experi-
mental Section); the molecular weight of their PEO struc-
tures is about 400 (n=9) and 600 (n=14), respectively. LC-
400 and LC-600 have number molecular weights of about
1130 and 1350, both with a PDI value of 1.2.
Both of the oligomers show mesomorphic phases, and
their thermal properties are listed in Table 1. LC-400 melts
into a nematic, liquid-crystalline phase at 1288C and then
transforms into an isotropic liquid at 1658C, as detected by
polarizing optical microscopy (POM). The peak melting
temperature (Tm) of the azomethine liquid-crystalline unit is
lowered with an increase of PEO unit molecular weight—
the transition temperature range is 123–1578C in LC-600.
The X-ray diffraction patterns showed only a broad halo
during the heating of LC-400 up to its melting point, which
is indicative of the change from the crystalline state to the
liquid-crystalline state shown in Figure 1.
Although block copolymers often self-assemble into nano-
fibers from well-controlled solutions, it is actually quite diffi-
cult to get a nanofiber-reinforced thermosetting network. As
reported previously, this is due to technical difficulties aris-
ing from the tendency of the growing thermosetting network
to expel any kind of inclusions or molecules initially dis-
persed or solubilized in the precursor mixture.[14]
Except for the blending of hybrid carbon nanotubes[15]
and thermotropic liquid crystalline polymers[16] with thermo-
setting systems, to the best of our knowledge, no report has
been published on the use of self-assembled fibers as rein-
forcing agents in thermosetting systems.
In the present work we report a new method for the crea-
tion of fibrils in epoxy networks in situ from an epoxy-ter-
minated ABA rod–coil–rod oligomer bearing aromatic azo-
methine liquid-crystalline structures. The rod–coil oligomer
provides low viscosity before the curing reaction, while the
epoxy function, which can be self-polymerized and co-react-
ed with epoxy precursors, combines with the liquid-crystal-
line unit to give enhanced thermal properties. During curing
of the blends, supramolecular nanoscale columnar structures
of the rod–coil oligomer are transformed into polymeric fi-
brils in the crosslinked epoxy networks.
As Griffin etal.[17] have reported previously, rod–coil–rod
oligomers show only nematic phases. These nematogens
have no rigid central core as the two rigid cores are joined
by a flexible segment, nevertheless the “center” of the struc-
ture is flexible. This flexibility is usually considered deleteri-
ous to mesogenic behavior because of the large number of
nonlinear geometries the molecule could adopt if the center
were flexible. Therefore, an increase of the length of the
PEO unit would reduce the tendency to show mesogenic be-
Chem. Eur. J. 2007, 13, 2920 – 2928
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2921

S. Li, Y. Yu et al.
columnar bundles and gives
rise to the formation of macro-
molecules with a well-defined
shape and size.
The neat oligomer can form
columnar (nanorod) structures
in both the crystalline and
liquid-crystalline states. As this
kind of structure is quite
stable, it might be possible to
preserve this structure during
blending with the epoxy pre-
cursor to give a nanostructure
in the epoxy network.
Morphology of the epoxy
blends: The morphology of
cured blends was investigated
by TM-AFM and SEM. Blend-
ing of the rod–coil oligomer
with an epoxy precursor and
imidazole was performed to
get chain-growth-polymerized
Scheme 1. Synthetic route to the a,w-modified PEO oligomers (n=9, 14). p-TSA= p-toluenesulfonic acid;
BTMA=benzyltrimethylammonium.
Table 1. Characterization of LC-400 and LC-600 by differential scanning
calorimetry (DSC) and polarized optical microscopy (POM).
Compound
Melting point [8C]
Clear point [8C]
DSC
125
118
POM
128
123
DSC
147
141
POM
165
157
LC-400
LC-600
havior, which is why LC-600 shows a lower transition tem-
perature for both Tm and the clear point than LC-400.
To observe the structure directly by AFM, crystalline
samples of LC-400 were prepared by slow cooling of melted
samples. As shown in Figure 2, the tapping-mode AFM
(TM-AFM) images indicate that LC-400 self-assembles into
a nanoscale, columnar structure with a length of about
10 nm.
For further observation of the mesogenic structure of the
oligomer, it was cured in the liquid-crystalline state. A
harder, imidazole C11Z was used as anionic initiator as this
could provide a quick chain-growth polymerization of the
epoxy function. The gel time when curing at 1408C is less
than three minutes, which means that the oligomer remains
in the liquid-crystalline state. As shown in Figure 2, the TM-
AFM images show organized, discrete bundles of nanoscale,
columnar structures about 30 nm in length in the liquid-crys-
talline phase. When combined with the AFM and XRD ob-
servations, these results indicate a preservation of the or-
dered symmetry and the dimensions of the discrete objects
in the LC state after polymerization. Based on the above re-
sults, it can be concluded that the polymerization proceeds
with retention of the size and shape of the self-assembled
Figure 1. a) X-ray diffraction patterns of LC-400 in different states and b)
the polarizing optical microscopy of LC-400 at 1608C.
2922
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2920 – 2928

Fibril Structures in Epoxy Networks
FULL PAPER
The morphology of the fibrils is almost unaffected by a
change of curing temperature, that is, in both the crystalline
and liquid-crystalline phase. Moreover, the fibrils are
formed no matter what weight fraction of rod–coil oligomer
is used in the blends. For example, uniform fibrils were ob-
served in the SEM images of LC-400 10% systems (Fig-
ure 3c), while they were also found in LC-400 5% and 20%
systems (Figures 3a and 3d).
The quantity and uniformity of the fibrils seems to be af-
fected only by the length of the PEO units of the rod–coil
oligomers—the higher the molecular weight of the PEO
units, the lower the volume fraction of fibrils observed in
the cured sample in the LC-600 blends.
It should be noted that the blends containing both LC-400
and LC-600 oligomers are transparent before curing, which
suggests the possibility of microphase separation and nano-
structuring. However, the curing reaction results in slightly
opaque samples despite this homogeneous dispersion of the
oligomers due to their dimensions relative to the wavelength
of visible light.
Structure formation process: Previous work has suggested
that the ability of diblock copolymers consisting of an
epoxy-miscible and an epoxy-immiscible block to form well-
defined, ordered and disordered microstructures in thermo-
setting epoxy resins is not specific to any class of copolymer
or thermoset.[5–8] This behavior appears to be highly depen-
dent upon the extent of favorable energetic interactions be-
tween the epoxy-miscible block and the cured epoxy net-
work and results in the formation of the ordered and disor-
dered morphologies typical for blends of a block copolymer
with a block-selective solvent.
Before curing, the epoxy precursor swells the PEO units,
which leads to a morphological behavior similar to those
corresponding to nano-ordered homopolymer/block copoly-
mer blends. After curing, the nanostructures are retained
and the cure-induced phase transitions occur because of the
local expulsion of PEO chains from the epoxy resin
matrix.[5–7]
In our case, the morphology of the epoxy blends was stud-
ied before and after the curing reaction. TEM images of the
rod–coil oligomer/epoxy precursor cast from THF solution
at room temperature (258C) reveal that these oligomers ex-
hibit a behavior in epoxy precursor solution that is very sim-
ilar to that shown in the previously examined rod–coil oligo-
mer in THF solution. For example, a blend of LC-400
(20 wt%) in epoxy precursor shows that the rod–coil oligo-
mer forms nanoscale columnar structures typical of its bulk
state at room temperature (Figure 4a). However, a signifi-
cant difference in the structures of the rod–coil oligomer
LC-400 was observed at higher temperatures. As shown in
Figure 4b, LC-400 displays nanoscale fiber structures in a
DGEBA/LC-400 (20%) blend above 738C, whereas some
clusters of LC-400 are also observed by TEM. Similar phe-
nomena have been explained for hydrogels as being due to
a temperature-dependent sol–gel transition in which nonpo-
lymeric hydrogels spontaneously form from small molecules
Figure 2. TM-AFM phase image (contrast mode) of LC-400 at room tem-
perature (image size: 500 nm; z-scale (contrast): 8 nm; scan rate: 1 Hz),
and LC-400 cured with C11Z (4%) at 1408C (image size: 1 mm; z-scale
(contrast): 150 nm; scan rate: 1 Hz).
thermoset systems. The samples were cured above the melt-
ing point of the oligomer (1408C) or below the melting
point of the oligomer (808C) to evaluate the effect of the
liquid-crystalline phase on the morphology of the cured net-
works.
Remarkably, highly uniform fibrils were observed in the
fully cured blend systems by SEM. The SEM images reveal
that they are actually fibril-like structures in epoxy blends.
As shown in Figure 3, fibrils with a diameter of about
500 nm and longer than 50 mm are observed in all the
blends.
Chem. Eur. J. 2007, 13, 2920 – 2928
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2923

S. Li, Y. Yu et al.
Figure 3. SEM and AFM images of nanofibers formed in DGEBA/LC-400 (5, 10, or 20%) and LC-600 (20%)/C11Z cured at 80 or 1408C for 5 h: a) LC-
400, 5%, 808C; b) LC-400, 5%, 1408C; c) LC-400, 10%, 808C; d) LC-400, 20%, 808C.
and the nanoscale fibers are also transferred from the small
molecules due to the dependence of the miscibility on the
temperature.[18]
Additionally, in a DGEBA/LC-400 (20%) blend cured at
808C in the presence of C11Z, LC-400 self-assembles into
fibril structures (Figure 4c). AFM shows the in situ forma-
tion of fibrils during the curing
process. These fibrils are in
fact bundles of aggregated,
regular, oligomeric, nanoscale
fibers dispersed in the epoxy
matrix with a nanoscopic rod
core/PEO shell-like fiber mor-
phologies.
The whole process can be
explained by the route shown
in Figure 5. As the amorphous
PEO and epoxy precursor
should be totally miscible,[5]
while the azomethine units are
insoluble in the epoxy precur-
Figure 4. TEM images of DGEBA/LC-400 (20%) at room temperature (258C) (a), heated to 808C (b), and
TM-AFM of DGEBA/LC-400 (20%)/C11Z cured at 808C for 5 h (c). Image size: 1 mm; z-scale (contrast):
150 nm; scan rate: 1 Hz.
2924
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2920 – 2928

Fibril Structures in Epoxy Networks
FULL PAPER
coat each rod, which means
that the rods become strongly
confined in the PEO phase.
The rods are physically moved
by osmotic effects resulting
from the phase separation, and
the temporal evolution of the
fluids “pushes” and “corrals”
the rods. Hence, the dynamic
coupling between preferential
adsorption and phase separa-
tion is critical for the forma-
tion of the network. Because
of the rod–rod repulsion, the
Figure 5. Schematic representation of the self-assembly of rod–coil oligomers in epoxy solution during heating
and polymerization.
sor, it can be anticipated that the epoxy precursor should
swell the PEO units in these oligomers without dissolving
the rod units and lead to an ordered composite before
curing. Upon increasing the temperature, the rod unit
begins to melt from the crystalline state, which means that
some epoxy molecules would diffuse into the ABA rod–coil
oligomers and this would enlarge the volume fraction of rod
units. The PEO unit cannot therefore maintain the stability
of the nanocolumnar structure due to its low volume frac-
tion at this stage. As a result, the nanostructures need to
combine to grow into larger structures (nanofibers) in order
to get a new balance between the PEO units and rod units
in a good solvent.
During the curing process, the polymerization of oligom-
ers themselves and their copolymerization with epoxy mole-
cules would increase the volume fraction and molecular
weight of the rod units, and the chain-polymerization of
epoxy precursors would change the blends into a ternary
system containing growing epoxy chains, epoxy precursors,
and polymerizing oligomers. These would decrease the en-
tropy of the blends and the miscibility between the growing
epoxy chain and polymerizing oligomers, therefore the
structures of rod–coil oligomers would need to aggregate
further to keep their stability, which forces the expulsion of
the polymerizing oligomers from the epoxy matrix.
high concentrations of rods have a lower free energy when
they are arranged end-to-end (beyond the range of the inter-
action) rather than side-by-side. This end-to-end arrange-
ment deforms the strongly wetting droplets, hence the rod–
rod repulsion also plays a crucial role in the formation of
the percolating network.
The extension of the rod units makes the fibril structures
grow longer and larger, while phase separation of enlarged
fibrils from the epoxy matrix further increases the dimen-
sion of the fibrils. The combination of these two effects
means that the nanoscale fibers of initial rod–coil oligomers
quickly grow into fibril structures during the curing reaction.
The imidazole-cured system quickly vitrifies to avoid fur-
ther phase separation due to polymerization. The curing
times for getting 50% curing conversion are about 13 min at
808C and about 3 min at 1408C for neat epoxy cured by imi-
dazole. At the same time, polymerization of the epoxy end-
group of the rod–coil oligomer results in anchoring of the
phase structures in the epoxy matrix, which means that
macro-phase separation is avoided due to the quick curing
process.
Thermal and mechanical properties: The thermal and me-
chanical properties of the unmodified and modified poly-
mers are listed in Table 2. As can readily be sees, the incor-
poration of LC-400 and LC-600 sharply increases the glass
transition temperature (Tg) of the modified systems com-
pared with the neat epoxy networks, although a further in-
crease of the amount of modifiers decreases Tg slightly. As a
higher modulus at the rubbery plateau was observed in the
modified systems, the increase in Tg could be attributed to
the co-reaction of rod–coil oligomers with the networks,
It has been reported from both computer simulation and
experimental results that when low volume fractions of
nanoscale rods are immersed in a binary, phase-separating
blend, the rods self-assemble into needlelike, percolating
networks.[19] The interconnected network arises through the
dynamic interplay of phase-separation between the fluids,
preferential adsorption of the minority component onto the
mobile rods, and rod–rod re-
pulsion. In our case, the nano-
Table 2. Thermal and tensile properties of selected samples.
scale fibers of ABA rod–coil
oligomers can also be consid-
ered as nanoscale rods and the
epoxy polymerization induces
Samples[a]
Glass
Tensile
Elongation
Tensile
transition
modulus [GPa]
at break [%]
strength [MPa]
temperature [8C]
neat epoxy
135
166
160
160
156
2.44
2.56
2.54
2.47
2.45
3.25
6.13
6.53
6.24
6.58
50.3
67.6
64.2
57.5
56.2
a
phase separation for the
DGEBA/LC-400 (10%)
DGEBA/LC-400 (20%)
DGEBA/LC-600 (10%)
DGEBA/LC-600 (20%)
same reasons.[19] During phase
separation, the strong prefer-
ential adsorption drives the
phase-separated PEO units to
[a] All samples were cured at 808C for 2 h and post-cured at 1508C for 4 h.
Chem. Eur. J. 2007, 13, 2920 – 2928
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2925

S. Li, Y. Yu et al.
synthesis of the rod–coil oligomers (Table 3). All the other chemicals
were provided by Sinopharm Chemical Reagent Co. (China) and purified
before use.
which might increase the ratio of propagation/termination
rates in the anionic chain polymerization and mean that less
dangling ends are formed. The PEO unit could also have
some effect on the plasticity.
The mechanical properties data show that the tensile
strength of all modified samples increases while the length
of all modified samples is about twice that of the neat epoxy
samples. However, the modulus of the modified blends is
comparable to that of the neat epoxy samples. It can there-
fore be expected that these rod–coil oligomers could be ap-
plied for the reinforcement of thermosetting resins and the
development of new functional materials.
4-(4-Hydroxybenzylideneimino)aniline (1): A literature procedure was
used for the preparation of this compound.[20] Equimolar amounts of p-
phenylenediamine and 4-hydroxybenzaldehyde were dissolved in DMF
under nitrogen. A few grains of p-toluenesulfonic acid were added and
the mixture was stirred at 808C for 4 h. A yellow-orange solid was ob-
tained after precipitation in water. This solid was filtered and washed
with methanol/water (50:50, v/v), then stirred twice with fresh methanol/
water (50:50, v/v) and filtered off. The orange-colored product was
vacuum-dried at 608C for 18 h. M.p. 2418C (DSC). 1H NMR
([D6]DMSO): d=8.45 (s, 1H; CH=N), 8.0–6.6 (m, 12H; aromatic),
4.25 ppm (br.s, 2H; NH2).
Aldehyde-terminated PEG ethers (2): A procedure described for the
preparation of 1,4-dichlorobutane[21] was used for the preparation of the
PEG dichloride. The glycols HO(-CH2-CH2-O)n-H (1 mol) and dry pyri-
dine (12.4 g, 0.1048 mol) were introduced into a flask fitted with a con-
denser. Thionyl chloride (4 mol) was added dropwise to this mixture over
2.5 h whilst stirring and cooling in an ice bath. Stirring was continued for
24 h at room temperature and this was followed by refluxing for 4.5 h.
The reaction mixture was then poured into an ice–water mixture. The or-
ganic layer was separated and the aqueous layer was extracted several
times with ethyl acetate. An aqueous (10%) solution of NaHCO3 was
then added to the combined organic layers. The organic layer was sepa-
rated, washed with water, dried with anhydrous MgSO4, and filtered. The
solvent was removed from the filtrate, and the residue was subjected to
vacuum distillation to yield the a,w-dichloropoly(ethylene oxide)s (yield:
70%) as pale-yellow liquids.
Conclusion
Polymerizable, liquid-crystalline rod–coil–rod oligomers
have been self-assembled into discrete nanorod structures in
the liquid-crystalline phase. Crosslinking of the oligomer in
the 3D ordered state proceeds with retention of the ordered
structure, which leads to the formation of crosslinked ob-
jects with a well-defined shape and size. Further blending of
the oligomer with an epoxy precursor gives fibril structures
in the cured networks with lengths in the micrometer range
due to the formation of nanoscale fiber structures by the
rod–coil oligomers in the epoxy precursor solution, which
suggests that the macromolecular objects are shape-persis-
tent in epoxy precursor solutions as well as in cured net-
works. Therefore, the polymerization of reactive nanorods
within the confined space and formation of fibril structures
in epoxy networks offers a novel strategy for the creation of
shape-persistent fibril structures which could be applied for
the toughening of thermosetting resins and the development
of new functional materials.
The above dichloride (0.1 mol), K2CO3 (14.0 g, 0.25 mol), 4-hydroxyben-
zaldehyde (30.5 g, 0.25 mol), and 250 mL of dry DMSO were introduced
into a flask equipped with a mechanical stirrer and a gas inlet. The reac-
tion mixture was stirred and heated (1308C) under nitrogen for 24 h. It
was then poured into an ice–water mixture. The organic layer was sepa-
rated and the aqueous layer was extracted several times with ethyl ace-
tate. The organic layer was separated and extracted with water to remove
the water-soluble PEG from the organic layer. The solution was dried
with anhydrous MgSO4 and filtered. The solvent was then evaporated
and the crude product was purified by column chromatography (silica
gel; eluent: hexane/ethyl acetate, 3/1) to give compound 2 as a pale
yellow oil (yield 70%).
Aldehyde-terminated PEG 400 ether: 1H NMR (CDCl3): d=9.96–9.92
(m, 2H; CH=O), 7.83–7.81 (m, 4H; aromatic), 7.12–7.09 (m, 4H; aromat-
ic), 4.18 (br.s, 4H; PhO-CH2), 3.89 (br.s, 4H; PhO-CH2-CH2), 3.71–
3.63 ppm (m; O-CH2-CH2-O).
Aldehyde-terminated PEG 600 ether: 1H NMR (CDCl3): d=9.96–9.92
(m, 2H; CH=O), 7.83–7.81 (m, 4H; aromatic), 7.12–7.09 (m, 4H; aromat-
ic), 4.18 (br.s, 4H; PhO-CH2), 3.89 (br.s, 4H; PhO-CH2-CH2), 3.71–
3.63 ppm (m; O-CH2-CH2-O).
Experimental Section
Materials: DER331 resin was used as the DGEBA epoxy precursor. 2-
Undecylimidazole (trade name C11Z), a white powder with a melting
point of 70–748C, was used as anionic initiator in the epoxy polymeri-
zation. Two kinds of commercial poly(ethylene glycol)s, namely PEG 400
and PEG 600, with repeat units of 9 and 14, respectively, were used for
Table 3. Characteristics of the epoxy precursor, initiator, and PEOs used in this work.
Name
Designation
Chemical structure
Mn
Supplier
diglycidyl ether of bi-
sphenol A
186–
DGEBA
Dow Chemical Co.
192 gequivÀ1
2-undecylimidazole
C11Z
222 gmolÀ1
Shikoku Chemical Co. (Japan)
poly(ethylene glycol)s
poly(ethylene glycol)s
PEG 400
(n=9)[a]
PEG 600
(n=14)[a]
HO
HO
U
414 gmolÀ1
634 gmolÀ1
Sinopharm Chemical Reagent
Co. (China)
Sinopharm Chemical Reagent
Co. (China)
[a] PDI (polydispersity index, Mw/Mn) value of 1.1. (as determined by GPC).
2926
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2920 – 2928

Fibril Structures in Epoxy Networks
FULL PAPER
Bis-Schiff base of PEG ethers (3): Equimolar amounts of 1 and the alde-
hyde-terminated PEG ethers 2 were dissolved in DMF with a few drops
of glacial acetic acid and the mixture stirred whilst heating at 608C for
15 min. The temperature was then raised to 908C and the stirring contin-
ued for 4 h. After cooling, the reaction mixture was poured into water,
the organic layer separated, and the aqueous layer extracted several
times with ethyl acetate. The organic solution was washed with water and
brine and dried with anhydrous MgSO4. The solvent was evaporated, and
the crude product was purified by column chromatography (silica gel;
eluent: hexane/ethyl acetate, 2/1) to give compound 3 as a yellow, wax-
like solid (yield: 90%).
Bis-Schiff base of PEG 400 ether: 1H NMR (CDCl3): d=7.63–7.58 (m,
4H; aromatic), 7.54–7.51 (m, 4H; aromatic), 7.41–6.38 (m, 8H; aromat-
ic), 6.95–6.82 (m, 8H; aromatic), 5.08 (s, 2H; PhOH), 4.19 (br.s, 4H;
PhO-CH2), 3.91 (br.s, 4H; PhO-CH2-CH2), 3.71–3.63 ppm (m; O-CH2-
CH2-O).
The dynamic mechanical properties were determined in the dual-cantile-
ver bending mode between room temperature and 4008C for rectangular
polymer samples of approximate dimensions 1”3”30 mm3 using
a
Netzsch DMA 242 apparatus. The measuring frequency was 1 Hz, the
sample displacement was 3 mm, and the heating rate was 108CminÀ1
.
The mechanical properties of the samples were recorded with an Instron-
8500 universal tester according to China State Standard GB 1040-79.
Sample preparation: Crystalline samples of rod–coil oligomer LC-400
were prepared by the slow cooling of melted samples at a rate of approx-
imately 0.18CminÀ1 on a silicon wafer to room temperature. For the poly-
merization of the ABA rod–coil oligomers, LC-400 was dissolved in THF
with 4% of 2-undecylimidazole (C11Z) and, after removal of the solvent,
the sample was cured in the liquid-crystalline state at 1408C under nitro-
gen for 5 h. Rod–coil oligomers (LC-400 or LC-600) with the epoxy pre-
cursor were prepared by dissolving the oligomer (5%, 10%, and 20% by
weight) in the epoxy precursor at elevated temperature and adding 4%
of imidazole C11Z to the blend to yield an optically homogeneous mix-
ture.
Bis-Schiff base of PEG 600 ether: 1H NMR (CDCl3): d=7.63–7.58 (m,
4H; aromatic), 7.54–7.51(m, 4H; aromatic), 7.41–6.38 (m, 8H; aromatic),
6.95–6.82 (m, 8H; aromatic), 5.08 (s, 2H; PhOH), 4.19 (br.s, 4H; PhO-
CH2), 3.91 (br.s, 4H; PhO-CH2-CH2), 3.71–3.63 ppm (m; O-CH2-CH2-O).
a,w-Modified PEO rod–coil oligomers (4): Compound 3 (0.05 mol) was
dissolved in 90 mL of epichlorohydrin(molar ratio: 40:1). After heating
at 1108C and stirring for 1 h, the solution became clear and
Acknowledgments
This work was supported by the Shanghai Leading Academic Discipline
Project (project number P1402). Y. Y. and S. L. express their sincere
thanks to Prof. Daoyong Chen (Fudan University) for many fruitful dis-
cussions and suggestions.
benzyltrimethylammonium(BTMA) was added as catalyst. The mixture
A
was then refluxed for another 2 h. Epichlorohydrin was removed by dis-
tillation and the residue was purified by column chromatography (silica
gel; eluent: hexane/ethyl acetate, 2/1) to give 4 in 95% yield as a yellow
solid.
LC-400: 1H NMR (CDCl3): d=8.42 (s, 4H; CH=N), 7.84–7.83 (m, 8H;
aromatic), 7.26–7.24 (m, 8H; aromatic), 7.00–6.98 (m, 8H; aromatic),
4.25, 4.02 (m, 4H; epoxy-CH2-O), 4.19 (br.s, 4H; PhO-CH2), 3.88 (br.s,
4H; PhO-CH2-CH2), 3.72–3.64 (m; O-CH2-CH2-O), 3.37 (s, 2H; CH in
epoxy ring), 2.85, 2.79 ppm (m, 4H; CH2 in epoxy).
LC-600: 1H NMR (CDCl3): d=8.42 (s, 4H; CH=N), 7.84–7.83 (m, 8H;
aromatic), 7.26–7.24 (m, 8H; aromatic), 7.06–6.98 (m, 8H; aromatic),
4.25, 4.02 (m, 4H; epoxy-CH2-O), 4.19 (br.s, 4H; PhO-CH2), 3.88 (br.s,
4H; PhO-CH2-CH2), 3.72–3.64 (m; O-CH2-CH2-O), 3.37 (s, 2H; CH in
epoxy ring), 2.85, 2.79 ppm (m, 4H; CH2 in epoxy).
[1] a) H. Zhang, L. A. Berglund, Polym. Eng. Sci. 1993, 33, 100–107;
b) H. Hsich, Polym. Eng. Sci. 1990, 30, 493–510; c) H. S. Lee, T.
Kyu, Macromolecules 1990, 23, 459–464.
[2] a) C. B. Bucknall, C. M. Gomez, I. Quintard, Polymer 1994, 35, 353–
359; b) B. G. Min, J. H. Hodgkin, Z. H. Stachurski, J. Appl. Polym.
Sci. 1993, 50, 1065–1073; c) G. S. Bennett, R. J. Farris, S. A. Thomp-
son, Polymer 1991, 32, 1633–1641; d) C. C. Riccardi, J. Borrajo,
R. J. J. Williams, E. Girard-Reydet, H. Sautereau, J. P. Pascault, J.
Polym. Sci., Part B: Polym. Phys. 1996, 34, 349–356; e) A. Bonnet,
J. P. Pascault, H. Sautereau, M. Taha, Y. Camberlin, Macromolecules
1999, 32, 8517–8523; f) Y. F. Yu, M. H. Wang, W. J. Gan, S. J. Li, J.
Phys. Chem. B 2004, 108, 6208–6215.
1
Measurements: H NMR spectra were recorded with a Bruker DMX 500
spectrometer (500 MHz). The polydispersities of the samples were mea-
sured with a Perkin–Elmer S-250 gel permeation chromatograph (GPC)
equipped with a Perkin–Elmer LC-235 UV detector set at 245 nm and an
LC-30 RI refractive index detector, and using three Waters Styragel col-
umns (HR 2, HR 4, and HR 5E), whose molecular weight detection
range is 500–20000, 5000–500000 and 2000–4000000, respectively.
[3] a) G. Levita, A. Marchetti, A. Lazzeri, Makromol. Chem. Macro-
mol. Symp. 1991, 41, 179–194; b) H. J. Sue, J. Mater. Sci. 1992, 27,
3098–3107; c) J. Choi, A. F. Yee, R. M. Laine, Macromolecules 2004,
37, 3267–3276.
[4] a) M. W. Matsen, Macromolecules 1995, 28, 5765–5773; b) Polymer
Blends (Eds.: D. R. Paul, S. Newman), Academic Press, New York,
1978, Vol. 1; c) M. Xanthos, S. S. Dagli, Polym. Eng. Sci. 1991, 31,
929–935; d) C. Auschra, R. Stadler, Macromolecules 1993, 26, 6364–
6367; e) C. Koning, M. V. Duin, C. Pagnoulle, R. Jerome, Prog.
Polym. Sci. 1998, 23, 707–757.
[5] a) R. B. Grubbs, J. M. Dean, F. S. Bates, Macromolecules 2001, 34,
8593–8595; b) M. A. Hillmyer, P. M. Lipic, D. A. Hadjuk, K.
Almdal, F. S. Bates, J. Am. Chem. Soc. 1997, 119, 2749–2750;
c) P. M. Lipic, F. S. Bates, M. A. Hillmyer, J. Am. Chem. Soc. 1998,
120, 8963–8970; d) J. M. Dean, P. M. Lipic, R. B. Grubbs, R. F.
Cook, F. S. Bates, J. Polym. Sci., Part B: Polym. Phys. 2001, 39,
2996–3010; e) R. B. Grubbs, J. M. Dean, M. E. Broz, F. S. Bates,
Macromolecules 2000, 33, 9522–9534.
XRD experiments were performed with a Philips X’pert MPD Pro X-ray
diffractometer. The thermal properties of the polymer samples were de-
termined with a Perkin–Elmer Pyris 1 differential scanning calorimeter
equipped with a liquid nitrogen cooling accessory. A 3–6-mg sample of
the purified material was placed on an aluminum pan and measured
against another empty pan as a reference.
Polarized light microscopy (PLM) experiments were performed with an
Olympus BX51P microscope equipped with an Instec HCS410 hot stage.
Atomic force microscopy (AFM) was performed with a Nanoscope IV
scanning microscope (controller, Veeco Metrology Group). The experi-
ments were performed in the soft tapping mode and the height and
phase images were obtained under ambient conditions in dry air (RH
ꢀ20%)with typical scan speeds of 0.5–1 lines per second.
[6] J. Mijovic, M. Shen, J. W. Sy, I. Mondragon, Macromolecules 2000,
33, 5235–5244.
Transmission electronic microscopy (TEM) was performed with a JEM-
1200EX TEM; samples of rod–coil oligomers and uncured blends were
prepared in THF solution. Both AFM and TEM samples of cured blends
were prepared with an ultramicrotome (Leica Ultracut R) equipped with
a diamond knife. The morphologies of the isothermally cured blends
were observed under a scanning electron microscope (Tescan TS 5163
MM) for samples fractured in liquid nitrogen. All samples were coated
with gold and mounted on copper mounts.
[7] a) Q. Guo, R. Thomann, W. Gronski, T. Thurn-Albrecht, Macromo-
lecules 2002, 35, 3133–3144; b) Q. Guo, R. Thomann, W. Gronski,
R. Staneva, R. Ivanova, B. Stuhn, Macromolecules 2003, 36, 3635–
3645.
[8] a) S. Ritzenthaler, F. Court, L. David, E. Girard-Reydet, L. Leibler,
J. P. Pascault, Macromolecules 2002, 35, 6245–6254; b) S. Ritzenthal-
er, F. Court, L. David, E. Girard-Reydet, L. Leibler, J. P. Pascault,
Macromolecules 2003, 36, 118–126; c) V. Rebizant, V. Abetz, F.
Chem. Eur. J. 2007, 13, 2920 – 2928
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2927

S. Li, Y. Yu et al.
Tournilhac, F. Court, L. Leibler, Macromolecules 2003, 36, 9889–
9896; d) V. Rebizant, A. S. Venet, F. Tournilhac, E. Girard-Reydet,
C. Navarro, J. P. Pascault, L. Leibler, Macromolecules 2004, 37,
8017–8027.
[15] a) L. S. Schadler, S. C. Giannaris, P. M. Ajayan, Appl. Phys. Lett.
1998, 73, 3842–3844; b) K. T. Lau, S. Q. Shi, Carbon 2002, 40, 2965–
2968; c) S. R. Wang, Z. Y. Liang, T. Liu, B. Wang, C. Zhang, Nano-
technology 2006, 17, 1551–1557.
[9] a) G. G. Barclay, C. K. Ober, Prog. Polym. Sci. 1993, 18, 899–945;
b) S. Atsushi, K. O. Christopher, Prog. Polym. Sci. 1997, 22, 975–
1000; c) A. P. Melissaris, M. H. Litt, Macromolecules 1994, 27, 2675–
2684; d) S. Cho, E. P. Douglas, Macromolecules 2002, 35, 4550–
4552; e) C. Carfagna, E. Amendola, M. Giamberini, Comput. Struct.
1994, 27, 37–43; f) C. B. Tan, H. Sun, B. M. Fung, B. P. Grady, Mac-
romolecules 2000, 33, 6249–6254.
[10] a) S. I. Stupp, V. LeBonheur, K. Walker, L. S. Li, K. E. Huggins, M.
Keser, A. Amstutz, Science 1997, 276, 384–389; b) S. C. Zimmer-
man, F. Zeng, D. E. C. Reichert, S. V. Kolotuchin, Science 1996, 271,
1095–1098; c) J. C. M. van Hest, D. A. P. Delnoye, M. W. P. L. Baars,
M. H. P. van Genderen, E. W. Meijer, Science 1995, 268, 1592–1595;
d) V. Percec, G. Johansson, G. Ungar, J. Zhou, J. Am. Chem. Soc.
1996, 118, 9855–9866.
[11] a) M. Lee, B. K. Cho, W. C. Zin, Chem. Rev. 2001, 101, 3869–3892;
b) C. Y. Li, K. K. Tenneti, D. Zhang, H. L. Zhang, X. H. Wan, E. Q.
Chen, Q. F. Zhou, A. O. Carlos, S. Igos, B. S. Hsiao, Macromolecules
2004, 37, 2854–2860; c) S. V. Arehart, C. Pugh, J. Am. Chem. Soc.
1997, 119, 3027–3037.
[12] a) M. U. Pralle, K. Urayama, G. N. Tew, D. Neher, G. Wegner, S. I.
Stupp, Angew. Chem. 2000, 112, 1546–1549; Angew. Chem. Int. Ed.
2000, 39, 1486–1489; b) E. R. Zubarev, M. U. Pralle, L. Li, S. I.
Stupp, Science 1999, 283, 523–526.
[16] a) S. Barrau, P. Demont, E. Perez, A. Peigney, C. Laurent, C. Laca-
banne, Macromolecules 2003, 36, 9678–9680; b) Y. Breton, G. Dꢁ-
sarmot, J. P. Salvetat, S. Delpeux, C. Sinturel, F. Bꢁguin, S. Bonna-
my, Carbon 2004, 42, 1027–1030.
[17] A. C. Griffin, T. R. Britt, J. Am. Chem. Soc. 1981, 103, 4957–4959.
[18] a) J.-H. Fuhrhop, C. Boettcher, J. Am. Chem. Soc. 1990, 112, 1768–
1776; b) M. Jokic, J. Makarevic’, M. Zꢂinic, J. Chem. Soc. Chem.
Commun. 1995, 1723–1725; c) L. A. Estroff, A. D. Hamilton,
Angew. Chem. 2000, 112, 3589–3592; Angew. Chem. Int. Ed. 2000,
39, 3447–3450; d) C. Marmillon, F. Gauffre, T. Gulik-Krzywicki, C.
Loup, A.-M. Caminade, E. Rump, Angew. Chem. 2001, 113, 2696–
2699; Angew. Chem. Int. Ed. 2001, 40, 2626–2629;.
[19] a) G. W. Peng, F. Qiu, V. V. Ginzburg, D. Jasnow, A. C. Balazs, Sci-
ence 2000, 288, 1802–1804; b) K. H. Wu, S. Y. Lu, Macromol. Rapid
Commun. 2006, 27, 424–429; c) I. Szleifer, R. Yerushalmi-Rozen,
Polymer 2005, 46, 7803–7818; d) G. J. van Maanen, S. L. Seeley,
M. D. Capracotta, S. A. White, R. R. Bukovnik, J. Hartmann, J. D.
Martin, R. J. Spontak, Langmuir 2005, 21, 3106–3115.
[20] V. Cozan, E. Avram, Eur. Polym. J. 2003, 39, 107–114.
[21] a) Vogelꢀs Textbook of Practical Organic Chemistry, Longman, New
York, 1978, p. 385; b) B.-A. Feit, B. Halak, J. Polym. Sci. Part A:
Polym. Chem. 2002, 40, 2171–2183.
[13] a) S. A. Jenekhe, X. L. Chen, Science 1998, 279, 1903–1907; b) X. L.
Chen, S. A. Jenekhe, Langmuir 1999, 15, 8007–8017.
Received: July 20, 2006
Revised: October 9, 2006
[14] A.-V. Ruzette, L. Leibler, Nat. Mater. 2005, 4, 19–31.
Published online: December 21, 2006
2928
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2920 – 2928