Supramolecular interaction controlled diffusion mechanism and improved mechanical behavior of hybrid hydrogel systems of zwitterions and cnt

Article abstract of DOI:10.1021/ma301366h

Hybrid hydrogels of poly(N-isopropylacrylamide) (pNIPAM) containing carboxylate carbon nanotubes (CNTs) and/or zwitterions are synthesized by free radical polymerization. The supramolecular interactions among zwitterionic monomers and CNTs influence the m

Full text of DOI:10.1021/ma301366h

Article
pubs.acs.org/Macromolecules
Supramolecular Interaction Controlled Diffusion Mechanism and
Improved Mechanical Behavior of Hybrid Hydrogel Systems of
Zwitterions and CNT
‡
†
†
Saud Hashmi, Amin GhavamiNejad, Francis O. Obiweluozor,
,§
,†
Mohammad Vatankhah-Varnoosfaderani,* and Florian J. Stadler*
†School of Semiconductor and Chemical Engineering, Chonbuk National University, Baekjero 567, Deokjin-gu, Jeonju, Jeonbuk,
61-756, Republic of Korea
5
‡
Department of Chemical Engineering, NED University of Engineering & Technology, University Road, Karachi-75270, Pakistan
§Department of Polymer Engineering, Amirkabir University of Technology, 424 Hafez Ave, P.O. Box 15875-4413, Tehran, Iran
S Supporting Information
*
ABSTRACT: Hybrid hydrogels of poly(N-isopropylacryla-
mide) (pNIPAM) containing carboxylate carbon nanotubes
(
CNTs) and/or zwitterions are synthesized by free radical
polymerization. The supramolecular interactions among
zwitterionic monomers and CNTs influence the mechanical
properties and diffusion mechanism in hybrid hydrogel
systems. These supramolecular interactions and response
behavior of hybrid hydrogels were tested mechanically and
with respect to their swelling characteristics. Hybrid hydrogels
of pNIPAM and CNT or pNIPAM, zwitterions, and CNT
follow a Fickian diffusion behavior, while adding zwitterions leads to an anomalous triple-stage swelling behavior and stiffening of
the gel due to the interactions of the zwitterions with each other, which significantly increase the viscous dissipation and change
the microscopic structure. While CNT itself stiffens the gel and slightly increases the diffusion speed, it complexes zwitterions,
which leads to a novel property profile that is both potentially antibiotic and electrically conductive. CNT affords a relaxation
process at long relaxation times, while zwitterion attachment and detachment lead to dissipation predominantly at high
frequencies. Dynamic rheological measurements were performed during swelling of these complex materials.
INTRODUCTION
Recently, hydrogels have also been used widely in a range of
areas, including biological, medical, optical, structural, mechan-
ical, material, and chemical engineering applications, because of
their outstanding biological and physical features such as
biocompatibility, transparency, impact resistance, water absorb-
■
Carbon nanotubes (CNTs) are considered to be an ideal
candidate for the manufacture of the next generation of
composite materials. In the past 10 years, research on the
application of CNTs was performed intensively due to their
unique mechanical and electrical properties. In order to harvest
the full potential of the CNT−polymer composites, researchers
have focused on attaching functional groups to the nanotube
surface.
Since CNTs were successfully solubilized in organic solvents
and in aqueous solution, their application in the biological field
has gained increasing attention. Their unique mechanical,
optical, and electrical properties
12−14
ance, and separation capability.
Considerable effort has
been made to fabricate hydrogels with enhanced mechanical
properties. Nevertheless, improving the mechanical properties
is still a challenge, particularly when using pure hydrogels cross-
1
5,16
1
linked with chemical, thermal, or irradiative stimuli.
2
Furthermore, it is important to control the mechanical
properties of hydrogels in a more targeted and systematic
manner in an effort to design a new hydrogel structure with
optimized physical properties, which would be suitable not only
for sustaining external forces from the environment but also for
biological functionality, including cell adhesion and prolifer-
3
−5
make them very attractive
6
building blocks for drug delivery and biochemical sensing.
Special CNT/gelatin hybrid hydrogels were found to have the
ability to separate proteins. Furthermore, it was found that
7
17−20
8
,9
ation.
Equilibrium water content in hydrogels is one of
CNTs had actuator properties. Thermoplastic elastomers
their basic properties. A hydrogel with high water content is
generally more advantageous for medical applications because
filled with CNTs show remote actuator and good shape-
10
recovery properties.
One application field of CNTs is in soft matter, where CNT
is not primarily used for its mechanical properties but for
enhancing its capabilities, such as by modifying the electrical
Received: July 3, 2012
Revised: November 19, 2012
1
1
properties.
©
XXXX American Chemical Society
A
dx.doi.org/10.1021/ma301366h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
2
1
of its permeability and biocompatibility. Usually, a large
swelling is accompanied by poor mechanical properties. There
are several compromise alternatives between large swelling and
good mechanical behavior. Increasing the cross-linking agent
concentration can improve the mechanical properties but
degrades the swelling. Copolymerization of hydrophilic
monomer (which favors swelling) with a less hydrophilic
monomer produces a hydrogel with good water absorbance and
improved mechanical properties in the resulting hydrogel.
Knowledge of swelling kinetics is important for designing
controlled-released devices for drugs and agriculture pesticides
based on swellable polymer matrices and for predicting the
1. Case I (when n < 0.5 Fickian): Fickian diffusion
mechanism in which the rate of diffusion is much less
than that of relaxation of the chains.
2. Case II (when n > 1): diffusion, in which the diffusion is
very rapid compared with the relaxation processes.
3. Case III (when 0.5 < n < 1 non-Fickian/anomalous):
non-Fickian diffusion, which occurs when the diffusion
and relaxation rates are comparable.
This paper reports the synthesis and characterization of
thermoresponsive poly(N-isopropylacrylamide) (p-NIPAM)-
co-zwitterionic hybrid hydrogels in filled with or without
multiwalled carbon nanotubes (MWCNTs). The aim is to
investigate the effect of supramolecular interaction among
carboxylate CNTs and zwitterionic monomer concentration on
the mechanical properties and diffusion mechanism in NIPAM-
based hydrogels and to relate these to the chemical structure as
the basis of the application of CNTs and zwitterionic
monomers in the field of biomaterials, actuator, and artificial
muscles. The swelling is conducted both in free state and in the
rheometer under small oscillatory shear. We evaluate the
structural changes during solvent diffusion in the rheometer,
allowing for the online observation of the diffusion-induced
changes in material behavior.
22
release rates of the active ingredients. The swelling and
shrinking behavior can be altered by the small change in
environment like pH, temperature, electrical field, ionic
strengths, and light.
Besides CNT, many other strategies were previously used to
modify the properties of hydrogels. One method is the
inclusion of ions in the hydrogel network, which leads to so-
called ampholytic hydrogels that can augment the range of
properties of hydrogels. There are two methods to prepare the
ampholytic hydrogels: (1) copolymerizing an anionic monomer
β-carboxyethyl acrylate) with a cationic monomer and (2)
incorporating a zwitterionic monomer into the hydrogel
network.
Such polymers (e.g., polyacrylamide copolymer compounds)
are useful for applications such as oil recovery, drug
23
(
2
4
EXPERIMENTAL SECTION
Materials. Methyl methacrylate (MMA, Merck), dibutyltin oxide
■
2
5
(
(
Merck), 3-(dimethylamino)-1-propanol, N,N′-methylenebis-
acrylamide) (MBA, Sigma-Aldrich, 99%), N,N,N′,N′-tetramethyle-
2
6
delivery, and as functional monomers used in the preparation
27
of synthetic copolymers. In these applications, both swelling
behavior and mechanical properties are essential parameters.
When hydrogels are brought in contact with the water, it
diffuses into the hydrogel and leads to swelling. Such swelling
occurs due to the large-scale segmental motion, resulting in a
bigger separation between hydrogel chains. Several publications
evaluated the transport mechanism of water diffusion in the
hydrogel matrix. Results reported so far have concluded that
the diffusion mechanism depends on several factors that include
the hydrophilic and hydrophobic nature of the copolymers,
cross-linking density, and the effect of external stimuli such as a
change in temperature and pH. The complexity of the
diffusion mechanisms makes it difficult to propose a universal
model that accurately describes the diffusion mechanism and to
calculate the diffusion coefficients. One of the classical models
used to describe the diffusion process in this work is based on
Fick’s law:
thylenediamine (TMEDA, Fluka, >99%), ammonium peroxodisulfate
(
APS, Aldrich, >98%), 1,3-propane sultone (Aldrich, 98%), and 1,4-
butane sultone (Aldrich, 98%) were used as received. NIPAM
(Aldrich, 98%) was recrystallized from a 65:35 (v/v) mixture of
hexane and benzene before use. Acetone was dried over CaCl and
2
distilled before use. All aqueous solutions were prepared with ultrapure
water purified with a Milli-Q UV-Plus water purification system
1
8
2
8
(Millipore, Bedford, MA). The water had a resistivity of >10 MΩ
−1
cm . N-(3-(Dimethylamino)propyl)methacrylamide was synthesized
by the transesterification of MMA and 3-(dimethylamino)-1-propanol
29
in the presence of dibutyltin oxide as a catalyst, as previously
3
5
reported. Thin MWCNTs (purity >0.90), prepared by the chemical
vapor deposition (CVD) process, were obtained from Nano Solution
Co. Ltd. The MWCNTs had a diameter of 10 nm and a length of over
30
1
0 μm.
Synthesis. Synthesis of Zwitterionic Monomer N-(Methacrylox-
ypropyl)-N,N-dimethyl-N-(3-sulfopropyl)ammonium Betaine (Zw).
A solution of 520 mg (50.4 mmol, 4.68 mL) of 1,4-butane sultone in 5
mL of acetone was added dropwise to a chilled stirred solution of
1
.020 g (48 mmol) of N-(3-(dimethylamino)propyl)methacrylamide
2
dc
dt
d c
in 20 mL of dry acetone. The reaction mixture was stirred and
increased to 40 °C, and the reaction was continued for 5 h. The white
particles were formed gradually, and then the slurry was filtered and
washed with a small amount of dry acetone. The filtrate was dried
under dry nitrogen and kept in a refrigerator. The reaction yield was
=
D
2
(1)
dx
To enable reasonable modeling of the diffusion process, a
complementary modeling method for early time approximation
the power law approach (eq 2) is used.
diffusion (m /m ) is proportional (k) to time raised to a power
n). Whereas m and m denote the amount of solvent diffused
into the gel at time t and at infinite time, respectively, k is a
constant related to the structure of the network and n is a
3
1,32
The relative rate of
78%.
1
Figure 1 demonstrates the H NMR (400 MHz, D O) results,
2
t
∞
which prove the purity of the obtained zwitterionic monomer: 5.986 s,
1
2
(
t ∞
H, (H−C=); 5.59 S, 1 H, (H−C=); 4.12 t, 2 H, (O−CH ); 3.28 t,
2
H, (N−CH ); 3.22 t, 2 H (N−CH ); 2.94 s, 6 H (N−CH ); 2.82 t, 2
2
2
3
H, (S−CH ); 2.07 m, 2 H (CH ); 1.78 s, 3 H (C−CH ), 1.65 m, 4 H
33
2
2
3
characteristic exponent of the transport mode of the solvent.
(
−CH −).
2
Preparation of Hydrogels. Purification and Carboxylation of
mt
=
ktn
Multiwalled Carbon Nanotubes (MWCNT). 1 g of multiwalled CNTs
(MWCNTs) was added gradually to 200 mL of concentrated nitric
acid (69%, Daegeun grade) and stirred vigorously. The suspension was
refluxed for 48 h at 120 °C in a 250 mL flask to remove the catalyst
and side products of the carboxylation treatment. After cooling down
to room temperature, the reaction mixture was poured into 500 mL of
m
(2)
∞
Depending on the relative rates of diffusion and polymer chains
relaxation, three classes of diffusion mechanisms were
3
4
observed:
B
dx.doi.org/10.1021/ma301366h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
a
Table 1. Hydrogels Composition
molar fraction of
zwitterionic
molar
amount
of
NIPAM
[g]
fraction of
NIPAM
cross-
linker
[g]
CNT
[wt %]
monomer
[mol %]
sample
[mol %]
CNT0-
Zw0
100
100
100
90
2
0.0675
0.0675
0.0675
0.0675
0.0675
0.034
CNT0.5-
Zw0
0.5
1
2
CNT1-
Zw0
2
CNT0.5-
Zw10
0.5
0.5
10
20
10
20
1.8
1.6
0.9
0.8
CNT0.5-
Zw20
80
1
Figure 1. H NMR and chemical structure of Zw.
CNT0-
Zw10
90
deionized water and then ultracentrifuged to separate the MWCNTs.
Deionized water was added, and the centrifugation was repeated. This
washing operation was repeated until the pH became the same as that
of deionized water. The purified and functionalized MWCNT was
dried in a vacuum oven at 100 °C overnight.
CNT0-
Zw20
80
0.034
a
Solution concentrations of initiator and accelerator used are 1 g/mL
water. A total of 50 μL was added of both.
200 mg of the above MWCNTs was dispersed in 20 and 40 mL of
deionized water to prepare 1 and 0.5 wt % MWCNTs suspensions,
respectively. The mixture was initially sonicated for 60 min (Model
WUC-A03H, Daihan Scientific Co. Ltd.), after which a stronger IKA
sonicator (Power Sonic 505, 350 W) was used for more intensive
sonication treatment of the aqueous mixture (180 min). A microtip
experiments were performed by placing the discs in distilled water at
room temperature and measuring the amount of water absorbed. The
weight of the samples increased over of time. The discs were removed
from the water, gently wiped with a filter paper to remove any excess
water, and then weighed. The discs mass, thickness, and diameter were
measured at incremental time as the water was absorbed to monitor
the dynamics of the hydrogel swelling.
(
1/4 in. in diameter) was immersed in the suspension, and the
sonicator was run at 50% of maximum amplitude for 3 × 10 min each
at room temperature. Well-dispersed CNTs are known to have lower
viscosities than those that have agglomerated.
36
The degree of swelling at different times can be calculated from the
equation
Synthesis. NIPAM, synthesized sulfobetaine zwitterionic (ZW) as
monomers, and MBA as a cross-linker were added to a vial containing
a solution of treated MWCNTs. The mixtures were stirred for 5 min,
and the dissolved oxygen was removed by bubbling nitrogen through
the solution for 1 h. An adequate amount of APS as initiator for free
radical polymerization was added to the vial under nitrogen bubbling
at room temperature.
After the solution was homogenized, TMEDA was added as an
accelerator into the monomer solution to initiate the radical
polymerization. After mild shaking for several seconds, the solution
was transferred to the mold (PS, 43 mm inner diameter, 5 mm height).
The polymerizations were continued for 6 h. After the polymerization,
the prepared hydrogel was removed from the mold and immersed in
double-distilled water at room temperature for at least 48 h, during
which time the water was regularly refreshed in order to remove
unreacted compounds.
W (%) = (m − m /m ) × 100
(3)
t
0
t
where m and m are the weights of the hydrogels at the predetermined
t
0
time and at time 0 (xerogel), respectively.
Rheological Measurements. Dynamic-mechanical rheological
measurements were performed on the swollen hydrogels until they
reached equilibrium. The change in the mechanical properties of the
hydrogels during swelling as a function of MWCNT and zwitterionic
concentration was measured.
For dynamic rheological measurement, a parallel plate geometry
with 20 mm diameter was used on a Malvern Kinexus rheometer in
oscillatory mode. The hydrogels, with as little water as possible to
obtain a gel flexible enough to fit in the aforementioned parallel plate
geometry, were placed in a Petri dish 60 × 15 mm fixed to the bottom
plate (containing a Peltier heater). The upper plate geometry was
pressed on to the gel with a constant normal force of 1 N (→ 3183 Pa)
to give a compressive deformation around 0.1. This normal force level
was found to be sufficient to maintain a continuous adhesion of the
sample to both plates throughout the test. The Petri dish was filled
with distilled water, and the hydrogel was allowed to swell at 25 °C.
Frequency sweep and time sweep experiments were used to analyze
both elastic (G′) and viscous (G″) material response.
The monomers and other reagent composition of the prepared
hydrogels are listed in Table 1. The designation of the samples reflects
the composition, where CNTx shows the amount of MWCNT in wt
%
of the total polymer content and Zwy stands for the molar
concentration of zwitterionic monomer (in relation to the total
monomer).
Structural Characterization and Morphological Observation.
Variable pressure field emission scanning electron microscopy
All experiments were carried out in the linear-viscoelastic regime
using a stress τ = 50 Pa, which leads to deformations γ ≈ 0.1−0.6%,
(
FESEM; Zeiss Supra 40VP) was used to observe fractured
0
0
morphologies of the as-prepared hydrogels. In order to evaluate the
morphology and uniform dispersion of MWCNTs in the NIPAM
hydrogels, dried hydrogels were fractured into fine particles and
dispersed in ethanol and sonicated for 20 min.
depending on the sample. The time-dependent data were measured at
−1
ω = 0.16 s (1 Hz). The frequency-dependent data were limited to an
−1
angular frequency ωmax of 60 s due to the low modulus and low
phase angle, which lead to geometry inertia artifacts.
Scanning electron microscopy (SEM; JEOL/JSM-6400) was used to
observe the surface morphologies and fractured surfaces of the
xerogels. Signals produced by secondary electrons, backscattered
electrons, and characteristic X-rays resulting from interactions of the
electron beam with atoms at or near the surface of the samples were
amplified, and the morphologies of the images were obtained.
Dynamic Swelling. The compositional effect of MWCNT and
zwitterions on the swelling behavior of the prepared hydrogels is the
main objective of the present study. The gravimetric method was used
to evaluate the swelling kinetics of the hydrogels. Dynamic swelling
This setup, while being able to monitor the change of properties as
a function of water uptake, suffers potentially from three experimental
artifacts.
First, the sample is significantly larger than the 20 mm plate, hence
leading to a slightly higher modulus than found in reality. This can be
compensated for by comparing the data with elongational data and
shear data with a defined sample shape and swelling state. The
correction is around 50%, which is in the range expected and can be
translated as a geometry that is apparently larger by 10%. Considering
that the sample is swelling in all three dimensions during the test, this
C
dx.doi.org/10.1021/ma301366h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
material behavior are shown best by δ; this quantity is discussed
primarily in this article.
Elongational Rheology. Samples for elongational rheology were
cut from the prepared sample disks using wetted razor blades. The
thickness of the sample disks was around 4 mm. Slices of
approximately 1.5−2.5 mm width were cut from the samples having
a length of 25−30 mm.
Elongational measurements were performed with a TA Instruments
ARES using an EVF (elongational viscosity fixture) geometry. Unlike
in the standard operation, the samples were not clamped on the
sample drums but superglued, which ensured proper adhesion and
avoided high notch stresses, which can easily lead to sample rupture
during sample loading and early stages of the experiment. The
3
8
experiments were conducted under constant Hencky strain rate ε
̇,
which allows for a physically more reasonable determination of
max
max
elongation at break ε
and stress at break σH
than classical
H
3
9
Cauchy terms.
Hencky strain rates of ε
1
̇
= 10, 1, 0.1, and 0.01 s− were conducted to
obtain information on the gel-like structure of the material, which
theoretically should be time- and, thus, strain-rate-independent.
Furthermore, the different strain rates also allow for a determination
of the modulus of the gel from the linear part of the time-dependent
Figure 2. Experimental setup of the swelling.
elongational viscosity η (t,ε ̇) , which is found for a Hencky strain ε <
e
H
artifact is unavoidable and needs to be corrected for. Fortunately, using
a sample that is significantly too large in diameter for the geometry
enables the problem to be simplified, as this leads to a constant
correction factor because the apparent increase in geometry diameter
is limited to a value of approximately the height. Hence, using a sample
that is significantly larger than the diameter + 2 × gap will lead to an
approximately constant error, which can be easily compensated for.
Second, the compressive stress on the sample, albeit relatively small,
will somewhat influence the swelling kinetics, which can be
extrapolated by comparing the swelling rates for different compressive
stresses. Third, the surface of the sample is reduced due to the two 20
mm contact points. Furthermore, an additional unknown fraction of
the lower side of the sample is stuck to the lower plate and, therefore,
needs a longer diffusion path than the sample floating freely in the
wateras performed in the normal swelling tests. The last two
artifacts lead to slower swelling kinetics.
0.2−0.5, depending on the structure of the material. The test setup has
40−43
been described elsewhere in full detail.
RESULTS
■
Electron Microscopy. Figure 3 shows the FESEM images
of the hydrogels containing significant amounts of solvent. The
CNT0-Zw10 sample (Figure 3c) clearly shows an entirely
different structure from CNT0.5-Zw10 (Figure 3a). As clusters
of CNTs were not observed in any of the samples, it is
concluded that the MWCNTs are uniformly dispersed into the
hydrogels. The CNT0.5-Zw10 sample (Figure 3a) is clearly
seen to be CNT-coated with NIPAM layers, as the thickness of
the fibrous structures is 16 nm vs the 10 nm of the MWCNTs
used for the material.
37
These artifacts have been discussed elsewhere. As the phase angle
δ is not influenced by the artifacts significantly, and the changes in the
It is concluded that the carboxyl and hydroxyl groups created
at the surface of MWCNTs during acid treatment stabilize the
Figure 3. FESEM images of (a) CNT0.5-Zw10, scale bar 20 nm; (b) CNT0.5-Zw20, scale bar 300 nm; (c) CNT0-Zw10, scale bar 1000 nm = 1 μm;
and (d) CNT1-Zw0, scale bar 200 nm.
D
dx.doi.org/10.1021/ma301366h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
Figure 4. SEM images of (a) CNT0-Zw0, scale bar 2 μm; (b) CNT0.5-Zw20, scale bar 5 μm; (c) CNT0-Zw10, scale bar 10 μm; and (d) CNT0-
Zw0, scale bar 5 μm.
MWCNT dispersions due to their strong interaction with the
zwitterion-containing matrix. As CNT1-Zw0 (Figure 3d) does
not show this increase in thickness of the CNTs, it is concluded
that this sample does not contain CNTs coated with the pure
pNIPAM matrix. This is a strong indicator that the interaction
with the functional groups, probably with the −NHSO− of the
structures at the bottom of the deepest pits, which are sized as
would be expected for polymer-covered CNTs. Hence, the
cracks in xerogels containing CNT can be stopped by CNTs
bridging the crack tips.
Investigation of Water Diffusion in Hydrogel. The
swelling properties were evaluated by free swelling experiments,
which were evaluated using eq 2. The exponents n obtained are
listed in Table 2. The maximum degree of swelling normalized
by the polymer weight m /m shows that the addition of Zw or
44
zwitterions, is responsible for the encapsulation of the CNTs.
Figure 3b shows large fibrous structures with a thickness of
several hundred micrometers in CNT0.5-Zw20, which seem to
be made up of weakly bundled CNTs complexed by the
zwitterions, presumably leading to the significantly different
properties of the hydrogels containing CNT and zwitterions
compared to the previously characterized pNIPAM hydrogels.
Figure 4 shows SEM images of dried hydrogels (xerogels)
predominantly without CNT, which cannot be characterized by
FESEM as well as the CNT-containing hydrogels, due to the
local heat produced by the electron beam. When comparing the
samples containing zwitterions (Figure 4b,c) to the pure
pNIPAM-hydrogel (Figure 4a,d), it is clear that zwitterions
cause a flatter surface structure on a small range and leaf, arc, or
flat on a length scale of ∼100 μm. This can be attributed to the
high interaction between the zwitterionic moieties, which,
especially at low water contents, hold the hydrogel together
strongly and, thus, prevent ductile-like cracks, leading to flat
breaking surfaces, which, in turn, due to the elasticity of the
samples (cf. Figure 12) can form some daughter cracks, leading
to the arcs and lamellae described.
∞
0
CNT increases the maximum water uptake.
Table 2. Summary of Diffusion Data
diffusion
m /m
∞
sample
characteristic exponent n
mechanism m /m
(solid)
∞
0
CNT0-
Zw0
0.33
Fickian
Fickian
Fickian
Fickian
Fickian
4.4
5.7
7.2
5.9
8.3
7.9
8.2
10.36
CNT0.5-
Zw0
0.42
0.45
0.47
0.45
11.22
10.74
11.58
12.8
CNT1-
Zw0
CNT0.5-
Zw10
CNT0.5-
Zw20
CNT0-
Zw10
0.66 (t ≤ 140 min) 0.35
(t ≥ 140 min)
non-Fickian
Fickian
10.5
CNT0-
Zw20
0.66 (t ≤ 140 min) 0.38
non-Fickian
Fickian
13.8
(t ≥ 140 min)
CNTs (Figure 4b) lead to finer fracture surface structures, as
the encapsulated CNTs lead to local zwitterion and,
consequently, polymer density fluctuations, which open up
paths for the cracks to penetrate into the sample. Furthermore,
the sharp ends of CNTs having an end radius of about 5 nm are
very strong stress raisers, which can induce the fracture in the
polymer-depleted vicinity of the CNT ends. At the same time,
the CNTs have the ability to bridge cracks, as evidenced in the
Figure 5 shows the swelling data of the different hydrogels,
which were evaluated by eq 2, with the first and last data points
omitted. For most samples, however, the first data point
roughly lies on the extrapolated regression line, indicating that
the diffusion mechanism is already close to its equilibrium at
low times (30 min).
E
dx.doi.org/10.1021/ma301366h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
When considering the gels containing zwitterions (Figure
5
b), the diffusion changes its characteristic around m /m = 0.6
t ∞
for both Zw samples (CNT0-Zw10 and CNT0-Zw20). At
lower swelling ratios/lower times (t ≤ 140 min), the slope n ≈
0
.66, indicating an anomalous diffusion, while for higher
swelling ratios/longer times, the exponent is almost the same as
that for CNT0-Zw0 (n = 0.35/0.38), indicating the action of
normal Fickian swelling in this regime. Furthermore, the
significantly slower slope that is apparent at short times (t ≤ 50
min) must, however, include the omitted data points and, thus,
the startup effects. Hence, this cannot be quantified. Explaining
this change in swelling behavior necessitates a discussion of the
rheological properties, which is presented in the Conclusions
section.
Figure 5c shows the effect of the CNT and Zw addition to
the hydrogel. The behaviors of CNT0.5-Zw10 and CNT0.5-
Zw20 are very similar to that of CNT0.5-Zw0, whose slope is
slightly lower. Unlike the sample containing only Zw and no
CNT, the illustrated behavior does not show two clear regimes.
Nevertheless, a small change of slope remains present for these
materials. Hence, the effect of the zwitterions on the diffusion
behavior is almost negligible in the presence of CNT.
To elucidate the processes happening in the hydrogels
containing zwitterions, the water content in relation to the
solids content should be investigated, for which the solids
content m = 0.6 g (0.603 and 0.606 g for the CNT-containing
0
samples) is used to normalize the weight of the sample m at
time t. Figure 6 shows the data for the samples containing
Figure 6. Ratio of total weight to solids weight as a function of time.
zwitterions only (CNT0-Zw10 and CNT0-Zw20) and for
CNT0-Zw0 and CNT0.5-Zw10. Beyond the superficially
similar curves, closer inspection reveals that CNT0-Zw10 and
CNT0-Zw20 are significantly below CNT0-Zw0 for t < 100
min. At 150 min, the slope of CNT0-Zw10 and CNT0-Zw20
changes, which is at the same time as the mass of the Zw-
containing samples reaches the curve of CNT0-Zw0; i.e., the
swelling of CNT0-Zw10 and CNT0-Zw20 is significantly
slower in the beginning and then speeds up and slows down at
the point where it meets the water uptake of CNT0-Zw0.
These results provide an explanation for the behavior.
Initially, the hydrogel’s zwitterions in the chains are only
separated by very little water, which allows for strong
zwitterionic interactions, as the hydrostatic forces are too
weak to temporarily separate them. As a consequence, the
Figure 5. Swelling as a function of time: (a) hydrogels with CNT, (b)
hydrogels with Zw, and (c) hydrogels with CNT and Zw.
The pure pNIPAM hydrogel (CNT0-Zw0, Figure 5a) shows
a diffusion coefficient n of 0.33, meaning that the diffusion is
significantly slower than the chain relaxation, which is further
clarified by the rheological data that are presented below.
Adding CNT increases the slope n to 0.42 and 0.45 for 0.5%
and 1% CNT, respectively, indicating that the relaxation is
slower than the diffusion. Furthermore, the diffusion is slowed
down in terms of m /m (t), as the values lie lower than for
t
∞
CNT0-Zw0.
F
dx.doi.org/10.1021/ma301366h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
swelling is hindered because the zwitterions’ interactions are
45
strong and thus do not make space for the water to diffuse in.
This leads to a less swelling at short times, which, however,
cannot be quantified due to the aforementioned startup effects.
At a critical water concentration, the interactions between the
zwitterions become so weak that they no longer hinder the
water uptake. At this time, the osmotic pressure to dilute the
(
zwitter)ions sets in, which significantly increases the swelling
rate. Furthermore, the loose interactions between the
zwitterions slow down the relaxation of the increasingly free
chain segments, which increases the diffusion exponent n. This
fast diffusion is slowed down when the swelling degree of
CNT0-Zw0 is reached, which is caused by the twin facts that
the diffusion rate is now limited by the NIPAM-network and
that the zwitterions are so greatly diluted that they no longer
have a significant osmotic contribution.
Shear Rheology. Figure 7 shows the rheological data as a
function of swelling time. In comparison to the free swelling
Figure 8. Time-resolved G′(ω) and G′′(ω) of (a) CNT0-Zw0 and (b)
CNT0-Zw 20.
Figure 7. Storage modulus G′, loss modulus G′′, phase angle δ, and
gap as a function of swelling time for sample CNT1-Zw0.
significantly higher G′′(ω) and a significant change with time.
Also, CNT0-Zw20 shows a significant decrease of G′(ω), while
G′(ω) of CNT0-Zw0 is almost constant. The other plots are
given in the Supporting Information, as in the following δ(ω)
will be discussed, as this quantity shows the changes in a way
easier to comprehend.
experiments, the samples were loaded with significantly more
water, as the less swollen samples were too stiff for loading
them. While the storage modulus G′ only decreases slightly at
longer times, due to the softening of the sample, the gap
continues increasing, suggesting that the swelling is not yet in
equilibrium. Loss modulus G′′ and phase angle δ continue
decreasing for almost 20 h until reaching steady state.
The reduction of the damping of the sample, i.e., the
decrease of δ, is due to the stretching of the chains by the water
diffusing in. As the chain segments between two cross-links are
stretched further, less energy can be dissipated by the dangling
of the segments. At equilibrium swelling, consequently, δ is very
low, indicating that only very little dissipation is possible. This
also suggests that very few free chain ends or free chains are
present in the sample. The latter point is also important for
applications, as the absence of free chains prevents the washing
out of any polymeric component of the hydrogel, thus making
it safer for medical applications.
The time-resolved plots of G′(ω) and G′′(ω) are shown in
Figure 8. For the different samples, G′(ω) shows a more or less
constant value, while G′′(ω) is lower by a factor 10−1000. The
swelling time is color coded starting from red and going to blue.
Figure 8a shows CNT0-Zw0, the standard PNIPAM-hydrogel,
whose properties show almost no change as a function of time
and G′′(ω) is very low, while CNT0-Zw20 (Figure 8b) shows a
In Figure 9, the frequency phase angle δ(ω) is discussed at
different times of water uptake at 25 °C. The filled symbols
(low) were taken in the beginning of the swelling, when the
swelling ratio was around 3. The open symbols were taken at
the end of the swelling with equilibrium water content (high).
In general, δ(ω) decreases with increasing swelling due to the
increasing stretch of the chain segments that results from the
increased water incorporation (cf. Figure 7). However, that the
change between “low” and “high” is lowest for the pure
pNIPAM hydrogels (CNT0-Zw0), followed by the Zw
hydrogels (CNT0-Zw10 and CNT0-Zw20). Pure pNIPAM
hydrogels (CNT0-Zw0) have an almost frequency-independent
phase angle δ(ω) of ≈0.8°, which means that almost no energy
dissipation is found in the frequency range. This indicates the
absence of any relaxation processes within the frequency range
of the experiment. As the samples are true covalent hydrogels, it
can be concluded that no further relaxation occurs at lower
frequencies, which was also confirmed by the finding that the
sample shape does not change under gravity after initial
sagging. As the relaxation process typically is detectable in
terms of δ(ω) at least 2 orders of magnitude away from the
G
dx.doi.org/10.1021/ma301366h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
deswelling, we conclude that both species must exist. The
dangling chain ends exist at a constant concentration
independent of the swelling state, but the stretching of the
network segments depends on the amount of water absorbed.
The difference in behavior for the CNT0-Zw0 hydrogel is very
small at low and high water content, indicating that the
unstretched network segments must only have a minor effect
on the behavior, as otherwise the observed change of energy
dissipation would be greater. Hence, the dangling chain ends
account for most of the observed damping δ(ω). This is
confirmation of the time-dependent data (Figure 7).
Figure 9a shows that the addition of CNT leads to a high
phase angle at low frequencies, which means that some slow
processes can occur. This is in agreement with the higher
exponent n of the swelling for these systems, which indicates
that the swellinga very slow process at very low frequency
occurs on roughly the same time scale as the relaxation. Hence,
it is logical that the swelling is delayed in comparison to the
pure pNIPAM hydrogel (CNT0-Zw0, Figures 5 and 6) and
that the exponent is close to anomalous diffusion. Considering
the composition of these samples, it is clear that the dissipation
originates from the interaction of the pNIPAM chains with the
CNT. With further water penetration, this energy dissipating
process becomes weaker and slower due to the depletion and
dilution of the CNTs and the surrounding pNIPAM chains.
The zwitterionic samples show a much higher damping at
high frequencies (Figure 9b) due to the interaction of the
zwitterions with each other. This interaction weakens with
increasing water content, as the concentration of zwitterions
decreases proportional to the water content. However, when
assuming that δ should be proportional to the Zw
concentration, approximately the right difference is found
between CNT0-Zw0, CNT0-Zw10, and CNT0-Zw20 (≈3° per
1
0% Zw), but the swelling dependence is much weaker than
that expected from a simple dilution, which is δ ≈ 1.5° (at 1
Hz) for swollen CNT0-Zw20 and not 4.5°. CNT0-Zw10
exhibits almost no swelling dependence.
Hence, it is concluded that the swelling does not separate the
zwitterionic groups so much that they can no longer interact.
Instead, they form local temporary concentration fluctuations,
which move with the Brownian motions of the chains. The
reduction of δ(ω) is then due to a weakening of these local
clusters of interaction. These conclusions agree with the
47
45
findings of Lee and Chen and Lowe et al.
The hydrogels containing CNT and zwitterions show yet
another behavior (Figure 9c). CNT leads to higher δ at low ω
and Zw to a higher δ at high ω. CNT0.5-Zw10 shows a
constant low δ(ω), which is almost identical to CNT0-Zw0 in
the swollen state. A small upturn at low ω is found only in the
“
low” swollen state, which indicates that a weaker version of the
processes, observed for CNT0.5-Zw0, also occurs here. The
CNT0.5-Zw20 sample shows a weaker version of the behavior
of CNT0-Zw10 in the “low” swelling state and basically the
same behavior as CNT0-Zw0 in the “high” swollen state.
On the basis of these findings and the swelling data, it can be
concluded that CNT adsorbs zwitterions on its surface, which
has two consequences. First, the anomalous swelling induced
by zwitterions without CNT cannot occur, as the zwitterions
cannot initially hinder diffusion. This lowers δ(ω) to 1°−2°.
Interestingly, CNT0.5-Zw10 and CNT0.5-Zw20 differ dis-
tinctly in their “low” swelling state behavior, which can be
attributed to the balance between CNT and zwitterion
concentration. 10% Zw is insufficient to prevent the relaxation
Figure 9. Phase angle δ(ω) as a function of frequency at different
swelling times: (a) hydrogels with CNT, (b) hydrogels with Zw, and
(
c) hydrogels with CNT and Zw. Temperature T = 25 °C.
4
3,46
relaxation time, at which it occurs,
the fast relaxation times
have to be in the sub-millisecond range.
This offers an interesting insight into the molecular structure
of the CNT0-Zw0 hydrogel. When considering the structure of
this hydrogel and the changes occurring during the swelling, it
is clear that free chain ends as well as long nonstretched
segments between two cross-linking points are candidates for
energy dissipation. From the reaction mechanism and the prior
H
dx.doi.org/10.1021/ma301366h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
process of CNT completely, leading to the small upturn at low
ω. 20% Zw cannot be absorbed on the CNT completely,
leaving some zwitterions unabsorbed and, thus, increasing δ(ω)
by about 1°. These results explicitly show that the swelling
behavior of these gels is affected by the internal structure and
the content of the zwitterions in the polymeric gels.
where G is the shear modulus, i.e., G = 3 × E . Figure 10b
0
shows the stress as a function of strain σ (ε ). While the data
H
H
for ε
̇
= 10 s− slightly deviates, the lower elongational rates
1
1
roughly agree. The disagreement of ε
̇
= 10 s− is due to the
high speed of the experiment, which lasts only 0.055 s until
rupture and which indicates that the inertias in the torque
transducer and the sample itself play a role. Nevertheless, the
similarity of the shape and measured values indicates that these
data can be considered valid as well. The prediction of σ (ε )
Elongational Rheology. Figure 10 shows the elongational
data of the hydrogel CNT0.5-Zw0, which is representative of
H
H
assuming a constant modulus E is given in Figure 10b as the
0
thick line, which describes the data well up to εH ≈ 0.25.
Following this, a power law increase is found due to the strain
max
hardening of the system. At σH ≈ 78 kPa, the tensile strength
of the network is reached, which is a relatively high value for a
NIPAM-based hydrogel. Some hydrogels with significantly
max
lower stiffness, higher elongation at break εH , and ionic
max 42
tenside surfmer micelles show a higher σH
.
When
comparing the data with the predictions from neo-Hookean
model (eq 5), it becomes obvious from Figure 10b that the
neo-Hookean model (eq 5) predicts a higher stress at high εH
than the classical Hookean model but that the experimentally
found stresses are still higher than neo-Hookean model. Hence,
it is concluded that while part of the strain-hardening can be
explained by the 3D-stress tensor, the material is still strain
hardening.
Figure 11 compares the hydrogels containing different
amounts of CNT. The modulus is given as the average of the
Figure 10. Elongational rheological data of the hydrogel CNT0.5-Zw0
Figure 11. Modulus and maximum elongation εHmax as a function of
CNT content. Lines are added to guide the eye.
with 0.5% CNT: (a) time-dependent elongational viscosity η (t,ε ̇) and
e
(
b) Hencky stress as a function of Hencky strain.
the other hydrogels. Figure 10a demonstrates that for ε < 0.25
value in elongation E and the triple value of the corrected
H
0
the elongational viscosity η (t,ε
log−log scaling, which is typical for ideal rubbers and can be
̇
shear modulus G0 in the swelling test, which assuming
incompressible fluids and simple elongation should be
e
49
fitted with
identical.
An increase of CNT content affords a small but clear increase
in strain at break εH , which is not obeyed by the samples
E = η (t, ε)/t
̇
0
e
(4)
max
which in this case yields E = 48 000 Pa, a relatively high value
for a hydrogel.
with zwitterions (red dots). The modulus goes through a
maximum around 0.5 wt % CNT. At 1 wt % CNT, the stiffness
of the hydrogel is significantly lower, which can only mean that
the network is weaker than for the sample with 0.5% CNT.
This behavior can probably be attributed to the adsorption of
some NIPAM on the surface of the CNTs (cf. Figure 3a),
which depletes the network and thus lowers the stiffness of the
hydrogel. The increased stretchability of these materials
underlines these findings and indicates that the network is
also depleted to some degree for CNT0.5-Zw0, but less than
0
For larger Hencky strains ε , the viscosity exceeds this
H
threshold, as the elongational deformation leads to a structural
orientation that stiffens the network. The strain hardening
42
found is related to the structure of the molecules. One might
argue that this strain hardening is caused by the 3D-tensorial
48
behavior, which can be modeled by a neo-Hookean model:
2
σ = G(λ + 1/λ) with ε = ln λ
(5)
E
H
I
dx.doi.org/10.1021/ma301366h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
that for the CNT1-Zw0 hydrogel. Hence, CNT0.5-Zw0 has a
higher modulus. This behavior is similar to the adsorption of
42
NIPAM in surfmer micelles reported earlier.
When comparing the samples with and without zwitterions
Figure 12), the functional group adds about 8 kPa in terms of
(
Figure 13. Schematic of the structure and diffusion mechanism.
relaxation mode at slightly higher frequencies than measured.
Additionally, this relaxation mode persists from almost
unswollen to equilibrium swollen state, suggesting an
interaction that does not depend significantly on the water
content. The most logical explanation is the interaction of the
zwitterionic groups with each other, significantly slowing down
the relaxation of the dangling or random chain segment.
In the elongational tests, this can be observed as a
significantly higher modulus for 10% Zw, which, however,
reverses for the CNT0-Zw20 sample. Further, the CNT0-Zw20
Figure 12. Modulus and maximum elongation ε max as a function of
Zw content. Lines are added to guide the eye.
H
stiffness for 0 and 10 wt % Zw, which is equal to a decrease of
the cross-linking molar mass from 17 000 to 14 000 g/mol and
which indicates that 10 wt % Zw is as effective as 0.44 wt %
cross-linker. However, both stiffness and elongation at break are
significantly decreased for 20 wt % Zw, probably due to the
zwitterions slightly disturbing the formation of cross-links
max
sample is very brittle, which shows in εH ≈ 0.45, a rather low
value for a hydrogel. This suggests that the zwitterions disturb
the cross-linking-reaction, leading to a gel with a significant
amount of errors in the cross-linked network and, hence, a
brittle sample.
during polymerization and, thus, giving a lower modulus than
max
expected. The lower εH
can be understood as a consequence
of the zwitterions creating clusters of interaction, which then
induce local stress risers that make the sample fail prematuremlyax.
The swelling behavior of the Zw-containing hydrogels shows
a clear change of slope from non-Fickian (n ≈ 0.65) to Fickian
for swelling times longer than 140 min. This can be explained
by the increasing dilution of the zwitterions, the content of
which is reduced from 10% or 20% of the sample mass in the
beginning to less than 1% in equilibrium swelling state. Hence,
the zwitterions stick together so strongly at low swelling state
that they slow down the dilution by water, which leads to a low
slope in the swelling curve (Figure 5). Once a certain threshold
in the water content is exceeded, the swelling speeds up
significantly as the half-diluted zwitterions “want to be diluted”.
Once the water content reaches that of the pure pNIPAM gel
(CNT0-Zw0), the water uptake rate of the pNIPAM network
dominates, leading to a slower uptake rate that is closely
correlated to a normal pNIPAM hydrogel.
In terms of rheology, the second regime has a high phase
angle δ at high ω, which is lower at equilibrium swelling. The
increasing dilution of the zwitterions introduces two effects.
First, fewer zwitterions are present in the sample measured (a
round section of the sample with a diameter of around 22 mm)
than at a lower swelling state. Second, the higher swelling state
also separates the zwitterions more, when no interaction
between them is assumed. The interactions will draw them
closer together, which, however, induces tensile forces due to
the entropic spring nature of the chain between two covalent
cross-links. As the swelling lengthens the distance between the
cross-linkers, while the length of the chain stays constant, the
entropic spring becomes stiffer; i.e., the zwitterions need to be
pulled out of their position toward the other zwitterions with a
Interestingly, adding CNT leads to a minimum in εH
followed by this quantity returning to approximately its original
value. This can be understood as the consequence of the fact
that Zw encapsulate CNTs. It is shown in Figure 9 that a little
more than 10% Zw fully encapsulate the CNTs, which means
that the sample CNT0.5-Zw20 has some free Zw, which, just
like for CNT0-Zw10, leads to a stabilization of cracks due to
zwitterionic interactions before leading to too big “holes” in the
network (CNT0-Zw20).
CONCLUSIONS
■
The results show some significant differences between
hydrogels containing only NIPAM, NIPAM and CNT, and
NIPAM and Zw and CNT on one hand and NIPAM and Zw
on the other. The swelling properties suggest a Fickian behavior
for the first group and a non-Fickian behavior for the second
group, which is highly anomalous, as it is divided into three
different stages. Figure 13 shows this behavior schematically.
A non-Fickian behavior indicates that the chain relaxation
and diffusion are comparable. When looking at the NIPAM and
Zw hydrogels in terms of their rheological properties, it is
obvious that their frequency-dependent phase angle δ(ω) is
significantly higher (3°−7° in the range between ω = 60 and
−
1
0
.1 s ) than for the other hydrogels (0.5°−2°). At the same
time G′(ω) decreases significantly for decreasing frequency,
which is logical considering the connection between δ(ω) and
d log |G*|/d log ω, as can be derived from the Kramers−Kronig
5
0
relations. This is clearly associated with the presence of a
J
dx.doi.org/10.1021/ma301366h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
higher force. This effectively leads to less zwitterion−zwitterion
interactions.
The hydrogel containing only pNIPAM self-evidently does
not show these additional interactions and, hence, shows a
Fickian diffusion behavior. In the absence of any relaxation
within the frequency range of the experiment, the phase angle
significant fraction of the Zw groups. For this reason, their
influence on the mechanical behavior is comparable to that of
fiber composites, while for the sample without Zw, the CNTs
are much less strongly bound to the network, which renders the
gel softer and more flexible. The FESEM images confirm this
finding by showing significantly higher CNT thicknesses than
the initial thickness of the unreacted CNTs, which indicates
that the CNT surface is covered by zwitterions and the adjacent
pNIPAM chains.
δ(ω) is very low (≈1°) and almost independent of swelling
max
time. The modulus and the elongation at break ε
typical range of pNIPAM hydrogels.
are in the
H
39−42
When considering the hydrogels containing pNIPAM and
CNT, the stiffness is distinctly increased for 0.5% CNT but
This is further evidence that CNTs can be trapped within the
hydrogel by zwitterions in the polymeric network, which
reduces the likelihood of CNTs being washed out of the
hydrogel and potentially harming the environment.
decreased for 1%. At the same time, the elongation at break
max
εH
increases slightly. This behavior can be attributed to
CNTs bridging the microcracks and, thus, delaying crack
propagation. Furthermore, the FESEM images show a
thickening of the CNTs, which indicates that a significant
amount of polymer is adsorbed on the CNTs. This adsorption
depletes the network and, hence, effectively lowers the
concentration of chains in the material, which are rheologically
active; i.e., they contribute to the polymeric network. The
dynamic-mechanical data show a slow relaxation process at low
frequency, which is related to the adsorbed pNIPAM
interacting with the CNT. The most likely explanation is a
motion of chain segments on the CNT surface.
SUMMARY
■
The diffusion mechanisms in pNIPAM-based hydrogels were
proven to be tailorable by the addition of zwitterions and
CNTs. Supramolecularly cross-linked zwitterions lead to a very
anomalous three-stage behavior: delayed initial diffusion, rapid
anomalous non-Fickian diffusion, followed by normal Fickian
diffusion. While the first two stages are dominated by the
zwitterions first delaying but then speeding up the diffusion, the
diffusion in the final stage is determined by the pNIPAM
network.
While CNT has almost no influence on the diffusion itself, its
capability to adsorb Zw on the surface changes the diffusion
mechanism in the hydrogels that contain both components.
The change of the properties of the hydrogels in terms of
diffusion mechanism, and shear and elongational rheology, as
well as the structure observed by FESEM, confirms the
supramolecular interaction between CNT and Zw, which
affords a new property profile that can be used in various fields
The swelling of these samples is unchanged in relation to the
normal pNIPAM hydrogel in terms of mechanism. The
maximum swelling, however, is significantly increased.
The addition of Zw to the hydrogels containing pNIPAM
and CNT significantly increases the stiffness, while simulta-
max
neously significantly decreasing the elongation at break. εH of
CNT0.5-Zw10 is smaller than that of both Zw10 and CNT0.5.
When comparing the shear rheological data with the other
samples, the phase angle δ(ω) is lower than that of the other
samples (CNT0.5-Zw0 and CNT0-Zw10) and is almost time
independent, indicating that although the water content
increases by a factor of 7% throughout the experiment, the
structure remains almost unchanged. This greatly decreased
phase compared to the hydrogels containing either CNT or Zw
suggests an interaction between these two components.
53,54
such as sensor, actuator, and biomedical applications.
ASSOCIATED CONTENT
■
*
S
Supporting Information
Figures S1 and S2. This material is available free of charge via
the Internet at http://pubs.acs.org.
51,52
Previous results indicate
that the zwitterions can form a
stable shell around CNT, which means that in the NIPAM,
CNT, and Zw hydrogels the zwitterions “cross-link supra-
molecularly” the CNT with the polymeric network. As this
interaction is rather stable, the phase angle δ is reduced to
almost the level of the pure pNIPAM-gel (CNT0-Zw0). For
CNT0.5-Zw10, δ(ω) is slightly increased at the lowest
frequencies, indicating that the zwitterion content is too low
to fully suppress the CNT relaxation at low ω, while CNT0.5-
Zw20 has an upturn of δ(ω) at high ω, which can be attributed
to an excessively high zwitterion content in relation to the CNT
content; i.e., the surface of the CNT is too small for the amount
of Zw in the sample.
AUTHOR INFORMATION
■
Corresponding Author
*(F.J.S.) E-mail fjstadler@jbnu.ac.kr; Ph +82-63-270-4039; Fax
+82-63-270-2306. (M.V.) E-mail: mvatankhah@aut.ac.ir.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors acknowledge the financial aid from the National
Research Foundation of Korea (“Synthesis and Character-
ization of Intelligent Self-Healing Double-Network Hydrogels”)
and “Human Resource Development (Advanced track for Si-
based solar cell materials and devices, project number
201040100660)” of the Korean Institute of Energy Technology
Evaluation and Planning (KETEP) grant funded by the Korea
government Ministry of Knowledge Economy. F.O. would like
to thank Chonbuk National University for the financial support.
The authors thank the staff of the CBNU central lab for help
with the FESEM images. The authors also thank Mr. Hyun
Geun Ock and Prof. Dr. Ahn Kyung-hyun for help with SER
measurements.
The mechanism of the interaction between Zw and CNT
(“encapsulation”) also rationalizes the finding that the diffusion
mechanism in NIPAM, CNT, and Zw hydrogels are Fickian, as
these two groups form a stable complex, which does not change
its properties, throughout the swelling.
The fact that in both dynamic-mechanical properties and in
swelling properties, the Zw become “invisible” by complexing
the CNTs as they attach themselves to the surface also explains
why the stiffness is significantly higher and the elongation at
max
break εH
is very low. The CNTs are mechanically well
integrated into the hydrogel network by their attraction of a
K
dx.doi.org/10.1021/ma301366h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
(
37) Hashmi, S.; Obiweluozor, F.; GhavamiNejad, A.; Vatankhah-
REFERENCES
■
Varnoosfaderani, M.; Stadler, F. J. Korea-Australia Rheol. J. 2012, 24
(
1) Chen, J.; Hamon, M. A.; Hu, H.; Chen, Y. S.; Rao, A. M.; Eklund,
P. C.; Haddon, R. C. Science 1998, 282 (5386), 95−98.
2) O’Connell, M. J.; Boul, P.; Ericson, L. M.; Huffman, C.; Wang, Y.
H.; Haroz, E.; Kuper, C.; Tour, J.; Ausman, K. D.; Smalley, R. E. Chem.
Phys. Lett. 2001, 342 (3−4), 265−271.
3) Yu, M. F.; Lourie, O.; Dyer, M. J.; Moloni, K.; Kelly, T. F.; Ruoff,
R. S. Science 2000, 287 (5453), 637−640.
4) Xia, Y.; Yin, X. C.; Burke, N. A. D.; Stover, H. D. H.
Macromolecules 2005, 38 (14), 5937−5943.
5) Douglas, T. A.; Tamburro, D.; Fredolini, C.; Espina, B. H.;
(
3), 191−198.
(
(
38) Hencky, H. Z. Tech. Phys. 1928, 9, 215−220.
39) Stadler, F. J.; Friedrich, T.; Tieke, B.; Bailly, C. Rheol. Acta 2013,
(
submitted.
40) Friedrich, T.; Tieke, B.; Stadler, F. J.; Bailly, C. Langmuir 2011,
7, 2997−3005.
41) Friedrich, T.; Tieke, B.; Stadler, F. J.; Bailly, C. Macromolecules
(
2
(
(
(
2
(
010, 43 (23), 9964−9971.
42) Friedrich, T.; Tieke, B.; Stadler, F. J.; Bailly, C. Soft Matter 2011,
(
7
(
(14), 6590−6597.
Lepene, B. S.; Ilag, L.; Espina, V.; Petricoin, E. F.; Liotta, L. A.;
Luchini, A. Biomaterials 2011, 32 (4), 1157−1166.
6) Wang, D.; Liu, T.; Yin, J.; Liu, S. Y. Macromolecules 2011, 44 (7),
43) Stadler, F. J. Rheol. Acta 2010, 49 (10), 1041−1057.
44) Chattopadhyay, D.; Lastella, S.; Kim, S.; Papadimitrakopoulos,
(
(
2
(
F. J. Am. Chem. Soc. 2002, 124 (5), 728−729.
(
282−2290.
7) Saleem, Q.; Liu, B. X.; Gradinaru, C. C.; Macdonald, P. M.
Biomacromolecules 2011, 12 (6), 2364−2374.
8) Ma, X. M.; Huang, X. B.; Zhu, L.; Zhao, X.; Tang, X. Z. Polym.
Int. 2005, 54 (1), 83−89.
9) Gholap, S. G.; Badiger, M. V. J. Appl. Polym. Sci. 2004, 93 (3),
454−1461.
10) Lynch, I.; Dawson, K. A. J. Phys. Chem. B 2004, 108 (30),
0893−10898.
11) Shi, J. H.; Guo, Z. X.; Zhan, B. H.; Luo, H. X.; Li, Y. F.; Zhu, D.
45) Lowe, A. B.; McCormick, C. L. Chem. Rev. 2002, 102 (11),
4
(
177−4189.
46) Stadler, F. J.; Bailly, C. Rheol. Acta 2009, 48 (1), 33−49.
47) Lee, W. F.; Chen, C. F. J. Appl. Polym. Sci. 1998, 69 (10), 2021−
(
(
2
034.
(
1
(
1
(
48) Macosko, C. W. Rheology: Principles, Measurements and
Applications; Wiley/VCH: Poughkeepsie, NY, 1994.
(
(
49) Trouton, F. T. Proc. R. Soc. London 1906, 77, 426.
50) Stadler, F. J.; Munstedt, H. Macromol. Mater. Eng. 2009, 294
̈
(
(
1), 25−34.
51) Soll, S.; Antonietti, M.; Yuan, J. Y. ACS Macro Lett. 2012, 1 (1),
4−87.
52) Zhang, W.; Swager, T. M. J. Am. Chem. Soc. 2007, 129 (25),
714−+..
53) Chang, Y.; Yandi, W.; Chen, W. Y.; Shih, Y. J.; Yang, C. C.;
Chang, Y.; Ling, Q. D.; Higuchi, A. Biomacromolecules 2010, 11 (4),
101−1110.
54) Chaterji, S.; Kwon, I. K.; Park, K. Prog. Polym. Sci. 2007, 32 (8−
), 1083−1122.
B. J. Phys. Chem. B 2005, 109 (31), 14789−14791.
(
8
(
7
(
(
12) Isik, B. Adv. Polym. Technol. 2003, 22 (3), 246−251.
13) Annaka, M.; Matsuura, T.; Kasai, M.; Nakahira, T.; Hara, Y.;
Okano, T. Biomacromolecules 2003, 4 (2), 395−403.
14) Haraguchi, K.; Takada, T. Macromol. Chem. Phys. 2005, 206
15), 1530−1540.
15) Kuckling, D.; Harmon, M. E.; Frank, C. W. Macromolecules
002, 35 (16), 6377−6383.
16) Zha, L. S.; Hu, J. H.; Wang, C. C.; Fu, S. K.; Luo, M. F. Colloid
Polym. Sci. 2002, 280 (12), 1116−1121.
17) Collins, M. N.; Birkinshaw, C. J. Mater. Sci.: Mater. Med. 2008,
9 (11), 3335−43.
18) Yildiz, B.; Isik, B.; Kis, M. React. Funct. Polym. 2002, 52 (1), 3−
0.
19) Garcia-Salinas, M. J.; Romero-Cano, M. S.; de las Nieves, F. J. J.
Colloid Interface Sci. 2001, 241 (1), 280−285.
20) Yildiz, B.; Isik, B.; Kis, M. Eur. Polym. J. 2002, 38 (7), 1343−
347.
21) Gan, D. J.; Lyon, L. A. J. Am. Chem. Soc. 2001, 123 (34), 8203−
209.
22) Jones, C. D.; Lyon, L. A. Macromolecules 2000, 33 (22), 8301−
306.
23) Xiao, X. C.; Chu, L. Y.; Chen, W. M.; Wang, S.; Xie, R.
(
(
(
(
2
1
(
9
(
(
1
(
1
(
(
1
(
8
(
8
(
Langmuir 2004, 20 (13), 5247−5253.
(24) Lee, W.-F.; Yeh, P.-L. J. Appl. Polym. Sci. 1999, 74, 2170−2180.
(25) Isik, B.; Gunay, Y. Colloid Polym. Sci. 2004, 282 (7), 693−698.
(26) Akdemir, Z. S.; Kayaman-Apohan, N. Polym. Adv. Technol. 2007,
1
(
(
8 (11), 932−939.
27) Zhang, L. M.; Hu, Z. H. Starch/Starke 2002, 54 (7), 290−295.
28) Paris, R.; Barrales-Rienda, J. M.; Quijada-Garrido, I. Polymer
2
009, 50 (9), 2065−2074.
29) Patel, M. P.; Johnstone, M. B.; Hughes, F. J.; Braden, M.
Biomaterials 2001, 22 (1), 81−86.
30) Cai, W. S.; Gupta, R. B. J. Appl. Polym. Sci. 2003, 88 (8), 2032−
037.
(
(
2
(
(
(
31) Ritger, P. L.; Peppas, N. A. J. Controlled Release 1987, 5, 37−42.
32) Ritger, P. L.; Peppas, N. A. J. Controlled Release 1987, 5, 23−36.
33) Mullarney, M. P.; Seery, T. A. P.; Weiss, R. A. Polymer 2006, 47
(
11), 3845−3855.
(
(
34) Frisch, H. L. Polym. Eng. Sci. 1980, 20 (1), 2−13.
35) Wasaki, K.; Kajihara, T.; Kawashima, T.; Nagano, H.;
Nakashima, S. Method for production of alkylamino(Meth)acrylate
and apparatus therefor. US Patent 6417392, 2002.
(
36) Grunlan, J. C.; Liu, L.; Kim, Y. S. Nano Lett. 2006, 6 (5), 911−
915.
L
dx.doi.org/10.1021/ma301366h | Macromolecules XXXX, XXX, XXX−XXX