The energy dissipation and Mullins effect of tough polymer/graphene oxide hybrid nanocomposite hydrogels

Ziqing Tang a, Feng Chen a, Qiang Chen *a, Lin Zhu a, Xiaoqiang Yan a, Hong Chen b, Baiping Ren b, Jia Yang a, Gang Qin a and Jie Zheng *b
aSchool of Materials Science and Engineering, Henan Polytechnic University, Jiaozuo, 454003, China. E-mail: qiangcheneric@163.com
bDepartment of Chemical and Biomolecular Engineering, The University of Akron, Akron, Ohio 44325, USA. E-mail: zhengj@uakron.edu

Received 25th June 2017 , Accepted 11th July 2017

First published on 12th July 2017


Nanocomposite hydrogels (NC gels) are considered to belong to the class of high strength hydrogels. Graphene oxide (GO), owing to its amphiphilic, mechanical, and optical properties, is widely used as a filler incorporated into different hosting materials (elastomers, plastics, and hydrogels) to improve their mechanical properties. In this work, we used in situ free radical polymerization to synthesize polyacrylamide (PAAm)/GO hybrid NC gels in the presence of GO nanosheets and a very small amount of chemical cross-linkers (N,N′-methylenebisacrylamide, MBA < 0.1 mol%). By optimizing GO and MBA concentrations, the resulting PAAm/GO gels can achieve an elastic modulus of 66 kPa, a fracture stress of 0.27 MPa, a fracture strain of 13.76 mm mm−1, deformed energy of 2.52 MJ m−3, and tearing energy of 964 J m−2. Due to the presence of physical interactions between PAAm and GO nanosheets, PAAm/GO gels demonstrate λ-dependent energy dissipation and Mullins self-recovery behaviors. The gels can rapidly recover their stiffness and toughness by 76% and 60%, respectively, after 30 min of resting at room temperature. The possible toughening mechanisms and Mullins effects of PAAm/GO gels were proposed and compared with those of filler rubbers and other high strength hydrogels. This work provides new viewpoints to develop tough hydrogels by the introduction of GO into other hydrogels with a good mechanical balance between strong chemical bonding and reversible physical bonding.


1. Introduction

Hydrogels, as soft and wet materials, contain a large amount of water in their three dimensional porous network structure, and have been extensively used in many fields including drug delivery systems, tissue scaffolds, wound healing dressings, and artificial soft tissues.1–3 However, most of the conventional hydrogels suffer from weak mechanical properties, which limit their applications in fields that demand high mechanical strength.4,5 In recent years, different strategies have been proposed to develop different types of tough hydrogels that can achieve high mechanical properties of tensile stress of >1 MPa and fracture energy of >103 J m−2.6 Unlike conventional hydrogels with either single network structures or simple polymer components, these tough hydrogels usually contain new network structures and hybrid network components, such as double network hydrogels,7–9 hydrogen bonding or dipole–dipole enhanced hydrogels,10–12 hydrophobically associated hydrogels,13–15 ionically cross-linked hydrogels,16 macromolecular microparticle composite hydrogels,17 tetra-PEG hydrogels18 and nanocomposite hydrogels,19–21 which allow effective dissipation of energy and deformation to be sustained sufficiently.

Among these tough hydrogels, the incorporation of different nanocomposites into the polymer networks of hydrogels (i.e. nanocomposite hydrogels or NC gels) is another fast-growing area of research, because nanocomposites have the advantages of chemical and structural-rich variations, easy synthesis and accessibility, high stability, large surface areas, and super electronic/thermal/mechanical properties.22 Specifically, graphene oxide (GO) is a well-known two-dimensional water-soluble nanomaterial consisting of a large number of hydrophilic oxygenated groups (e.g. hydroxyl, epoxide and carboxyl groups).23–27 Owing to a large number of hydrophilic groups in GO, GO can be easily exfoliated into monolayer nanosheets and stably dispersed in aqueous solution. Meanwhile, GO, due to its high stiffness, is considered as one of the best fillers to improve the mechanical properties of hydrogels. The introduction of stiff and hydrophilic GO as the hard phase into the hydrogels is expected to increase physical interactions between polymer chains and GO, which in turn bear larger stress and dissipate more energy and result in the improved mechanical properties of GO-based hydrogels.

A number of tough GO-based hydrogels have been developed. Ye et al. prepared GO-chemically cross-linked polyacrylamide (PAAm) and poly(acrylic acid) (PAAc) hydrogels. They found that both PAAm/GO and PAAc/GO gels exhibited better mechanical properties than pure polymer hydrogels.28,29 The fracture stress (σf) of PAAm/GO and PAAc/GO gels reached 32 kPa and 25 kPa, respectively, which was much better than that of pure PAAm and PAAc gels (<5 kPa). Yu et al. reported that GO can also be physically cross-linked with PAAm to form physical PAAm/GO gels,30 which exhibit excellent mechanical properties (σf of 385 kPa and εf of 3435%, εf present a fracture strain). Wang et al. used the peroxidized GO (GPO) as a polyfunctional initiating and cross-linking center to fabricate PAAm/GPO hydrogels.31 The PAAm/GPO gels showed high tensile strength (0.2–1.2 MPa) and strains (2000–5300%), as well as excellent self-healing properties at room temperature.32 Cong et al. fabricated a series of GO-based dual network hydrogels, consisting of a Ca2+ cross-linked GO network and a chemically-cross-linked PAAm network or a physically-cross-linked poly(N-acryloyl-6-aminocaproic acid) (PAACA) network, both of which possessed high tensile strength (60–500 kPa) and large elongation (200–1200%).33,34 Particularly, physical PAACA/GO gels also exhibited visual self-healing properties, but no tensile tests were reported on the self-healed gels.34 Shi et al. synthesized poly(N-isopropylacrylamide)/GO (PNIPAM/GO) NC gels, which showed a high strength of 700 kPa and a high extensibility of 3800%.35

All these GO-based hydrogels demonstrate that the incorporation of graphene sheets as nanoscale fillers enables enhancement of the interactions between polymer chains and GO sheets and consequently, improvement of the mechanical properties of host polymer hydrogels. On the other hand, the enhanced mechanical properties are still limited by the type and extent of cross-linkers. In general, physical crosslinks between GO sheets and polymer chains are not strong enough to largely improve the tensile strength of GO-based hydrogels, usually resulting in less than 80% mechanical improvement.31 On the other hand, the presence of excess chemical cross-linkers (>0.1 mol% of monomer) causes large stress concentrated on the shortest chains due to network heterogeneity, as evidenced by the fact that the fracture strains of PAAm/GO and PAAc/GO nanocomposite gels were all less than 700%.28,29 Thus, optimizing the combination of chemical and physical crosslinks is critical for developing tough GO-based hydrogels.36,37 Meanwhile, due to the similar network structure between polymer/GO nanocomposite gels and carbon black (CB) filled rubbers, it is expected that polymer/GO nanocomposite gels should show some behaviors of large hysteresis, mechanical recovery, and Mullins effects similar to that of CB filled rubbers. However, few studies, if any, have been reported that investigate the energy dissipation, Mullins effect and self-recovery mechanisms of GO-based NC hydrogels.

Herein, we present the formation of mechanically strong and tough hybrid PAAm/GO nanocomposite hydrogels using in situ free radical polymerization of AAm in the presence of GO nanosheets and chemical cross-linkers of N,N′-methylenebisacrylamide (MBA). The PAAm/GO hydrogels consist of a chemically cross-linked PAAm network and a physically cross-linked PAAm-GO network, where GO nanosheets are well dispersed in the PAAm network. We demonstrate that the incorporation and combination of both chemical and physical cross-linkers can effectively enhance the mechanical properties of PAAm/GO hydrogels. Under optimal conditions, PAAm/GO hybrid NC gels exhibited an elastic modulus of 66 kPa, fracture stress of 0.27 MPa, fracture strain of 13.76 mm mm−1, deformation energy of 2.52 MJ m−3, and tearing energy of 964 J m−2, similar to that of cartilage (102–103 J m−2), rubbers (102–103 J m−2) and double network hydrogels (102–104 J m−2). We further performed a series of tensile, cyclic loading-unloading, tearing tests and recovery experiments to better understand the toughening mechanisms of PAAm/GO hydrogels. It was found that when tuning MBA concentrations of <0.1 mol%, the elastic modulus and tearing energy of PAAm/GO gels increased as GO concentrations, indicating that PAAm/GO gels increase their stiffness and toughness simultaneously. More interestingly, PAAm/GO gels displayed λ-dependent energy dissipation behaviors, i.e. at λ ≤ 2, PAAm/GO gels showed excellent elasticity, but no hysteresis loop during loading–unloading cycles, while at λ > 2, the gels displayed large hysteresis to dissipate energy. The elastic modulus of PAAm/GO gels decreased at the maximum extension ratios, clearly indicating a typical “Mullins effect”, which was also found in CB-filled rubbers. However, different from CB-filled rubbers, PAAm/GO gels can recover 76% stiffness and 62% toughness after 30 min resting at room temperature. This work not only adds a new gel system (i.e. PAAm/GO hydrogel) to the GO-based hydrogel family, but also offers a better understanding of the structure–property relationship of GO-based gels, which would help to design new tough nanocomposite hydrogels.

2. Materials and methods

2.1 Materials

Acrylamide (AAm), N,N′-methylenebisacrylamide (MBA, 99%), ammonium persulfate (APS), potassium permanganate (KMnO4), sodium nitrate (NaNO3), hydrogen peroxide (H2O2, 30 wt%), concentrated sulfuric acid (H2SO4, 98%) and graphite powder (<35 μm mesh) were all purchased from Aladdin reagent Inc. (Shanghai) and used without further purification.

2.2 Synthesis of graphene oxide (GO) and PAAm/GO hybrid NC gels

Synthesis of GO. Graphite oxide was synthesized according to the Hummers method with modification.23 Briefly, 46 mL H2SO4 was added in a flask, which was immersed in ice/water mixture to maintain the temperature below 4 °C. Then, graphite (1.5 g) and NaNO3 (1.5 g) were slowly added with stirring into the flask. After 10 min, KMnO4 (9.0 g) was gradually added. The mixture was kept at 35 °C for 2 h under stirring. Subsequently, H2O (80 mL) was added into the mixture slowly, which made the temperature of the mixture rise quickly to ∼80° C. Stirring was continued for 0.5 h at this temperature. Afterwards water (200 mL) and H2O2 (6 mL) solution were added. The resulting suspension was filtered. The solid mixture was repeatedly washed with large amounts of distilled water until the solution pH reached a constant value at ∼4.0. Then, the resulting GO solution was frozen-dried by using a lyophilizer.
Synthesis of PAAm/GO hybrid NC gels. PAAm/GO hybrid NC gels were prepared by in situ free radical polymerization in the presence of GO and MBA of various concentrations. Briefly, for synthesis of NC gels with GO of 4 mg mL−1 and MBA of 0.07 mol%, 2.4 g AAm and 68 μL MBA (20 mg mL−1) were added into 10 mL GO solution (4 mg mL−1). The solution was stirred to dissolve all the reactants, and then the solution was degassed three times and protected by N2. After quickly injecting 77 μL APS (100 mg mL−1, 0.1 mol% of AAm), the solution was separated and sealed in a glass tube with a diameter of 8 mm. Upon 12 h polymerization at 60 °C in a water bath, the gel was taken out for tests. Unless otherwise stated, NC gels with GO of 4 mg mL−1 and MBA of 0.07 mol% were used for the various tests. PAAm gel with MBA of 0.07 mol% was also synthesized by the same process except for water used instead of GO solution.

2.3 Characterization

X-ray diffraction (XRD) patterns were recorded on an D8ADVANCE XRD (Bruker) with Cu-Kα radiation over a 2θ range of 5–50° with a scanning rate of 4° min−1. Raman spectra were recorded on an in Via-Laser microscopic confocal Raman spectroscope (Renishaw, UK) at an excitation wavelength of 514.5 nm with a resolution of 1 cm−1. Thermogravimetric analyses (TGA) were conducted with a PerkinElmer TGA 2050 instrument at a heating rate of 10 °C min−1 under a nitrogen atmosphere.

2.4 Mechanical measurements

Uniaxial tensile tests of gel specimens (diameter of 8 mm and length of ∼60 mm) were performed on a universal testing machine equipped with a 100 N load cell with a stretching rate of 100 mm min−1 or a different stretching rate if needed. The elastic modulus (E), fracture stress (σf), fracture strain (εf), and deformed energy (W) were all defined as previously reported. The tearing testing was conducted using the same machine. The gel specimens were cut into a trouser shape (80 mm in length, 15 mm in width, and 1 mm in thickness) with an initial notch of 33 mm. The tearing tests were done at 50 mm min−1, in which one arm of the gel was fixed, while the other one was pulled. The tearing energy (T) is defined as the work required to tear a unit area, as estimated by:38
 
image file: c7py01068k-t1.tif(1)
where Fave is the average force of peak values during steady-state tear, and w is the width of the specimen.

For hysteresis measurement, gel specimens were firstly stretched to a maximum extension ratio λmax of 3 and then unloaded. The dissipated energy (Uhys) was estimated from the area between the loading–unloading curves. For successive cyclic loading measurements, the gel specimen was firstly stretched to a maximum extension ratio λ1 and then unloaded at 100 mm min−1. The same gel specimen was reloaded and stretched to an increased extension ratio λ2 as the first loading cycle again. The loading–unloading cycles were repeatedly conducted on the same specimen with increased λ3, λ4, …, λn until the specimen fractured. λmax represents the extension ratio where the specimen deformed. The elastic modulus (E) at each λmax was estimated in the linear range from the loading curve. The dissipated energy (ΔUhys_n) during each loading cycle is calculated from the area between the nth loading–unloading curves. The total dissipated energy (Uhys_n) from the first loading to the nth loading is the sum of ΔUhys_i:

 
image file: c7py01068k-t2.tif(2)

The work required to break the gel specimen at λn, Wn, is calculated by

 
image file: c7py01068k-t3.tif(3)

For self-recovery tests, the gel specimens were firstly stretched to a maximum extension ratio λmax of 5 and then unloaded at 100 mm min−1. Then, after resting for various times (0 to 30 min) at room temperature, the same loading and unloading were performed again. The recovery rate (%) is defined as the ratio of dissipated energy (ΔUhys_t) or elastic modulus (Et) at various recovery times to that of the first loading cycle (ΔUhys_0 or E0). Cyclic loading at a maximum extension ratio λmax of 5 with a recovery time of 0 min was also conducted 6 times with the same gel specimen to investigate the Mullins effect of NC gels.

3. Results and discussion

3.1. Synthesis and characterization of PAAm/GO gels

Fig. 1 shows a general synthesis process to prepare PAAm/GO nanocomposite gels by in situ free radical polymerization in the presence of GO nanosheets as nanofillers and physical cross-linkers, a very small amount of MBA as chemical cross-linkers and APS as initiators. During the polymerization process, PAAm chains and GO nanosheets can interact with each other to form both chemically and physically linked domains. Since both GO nanosheets and PAAm chains contain many hydrophilic groups, the PAAm chains can physically adsorb on and entangle with the GO nanosheets via hydrogen bonds between the amide groups of the PAAm chains and the epoxy, hydroxyl, and carboxyl groups of the GO nanosheets. Meanwhile, due to the chain transfer of PAAm to the GO nanosheets, PAAm can also graft onto the GO nanosheets via chemical cross-linkers, followed by chain polymerization to form chemically cross-linked domains.39 Additionally, macromolecular radicals could add to the double bonds of the GO to produce GO radicals, which initiate the further PAAm polymerization on GO nanosheets. So, PAAm chains can be grafted and polymerized on GO nanosheets via chemical cross-linkers or coupling termination. As a result, GO nanosheets serve as multifunctional cross-linkers to chemically and physically interact with PAAm chains to form PAAm/GO hydrogels with both cross-linked domains.
image file: c7py01068k-f1.tif
Fig. 1 Preparation of PAAm/GO nanocomposite gel and illustration of its network structure with both chemically and physically cross-linked domains between PAAm and GO nanosheets.

As shown in Fig. 2, the GO and PAAm/GO gels were characterized by different methods. TEM images showed that the GO nanosheet was thin and flexible (Fig. 2a). Raman spectra showed that GO nanosheets exhibited a typical D band of 1350 cm−1 (an indicator of the defective structures of carbon materials) and G band of 1594 cm−1 (an indicator of the graphitization of carbon materials based on the E2g graphite mode40), while a PAAm single-network hydrogel exhibited vibrational stretching for O–H and N–H hydrogen bonding at 2920–3000 cm−1. Upon incorporation of GO nanosheets into the PAAm network, the Raman spectrum of PAAm/GO hydrogels also showed the D and G bands of GO nanosheets and O–H and N–H stretching of PAAm chains at the same locations (Fig. 2b). Moreover, both GO nanosheets and the PAAm/GO hydrogels retained a similar D/G intensity ratio (ID/IG) of 0.95, indicating that there is no clear reduction of GO during in situ polymerization. XRD analysis showed that the GO exhibited a sharp peak at 2θ = 10.2°, which corresponds to the average interlayer space of ∼0.81 nm between the stacked GO sheets (Fig. 2c). However, this XRD peak disappeared in the PAAm/GO hydrogels, and the disappearance of interlayer spacing in the GO is an indicator that upon incorporation and polymerization, the stacked GO nanosheets are disassembled and fully exfoliated in the PAAm network.


image file: c7py01068k-f2.tif
Fig. 2 Characterization of PAAm/GO hydrogels. (a) TEM of GO nanosheet (scale bar = 500 nm); (b) Raman spectra and (c) XRD curves of GO, PAAm gel, and PAAm/GO nanocomposite gel.

3.2. Mechanical properties of PAAm/GO gels

In Fig. 3, PAAm/GO nanocomposite gels demonstrate their excellent tensile properties. The gel can sustain (a) original stretching, (b) twisted stretching, and (c) knotted stretching up to 5–6 times its initial length without breaking. The concentration effects of the two cross-linkers (MBA and GO) on the mechanical properties of PAAm/GO gels were investigated. Fig. 4 and Table S1 show the tensile stress–strain curves of PAAm/GO nanocomposite gels prepared at different MBA concentrations. It can be seen that except for elastic modulus (E), with the increase in MBA concentrations from 0.01 to 0.07 mol%, PAAm/GO gels increased their tensile stress (σf) from 0.01 to 0.27 MPa, tensile strain (εf) from 6.84 to 13.76 mm mm−1, and deformed energy (W) from 0.60 to 2.52 MJ m−3, respectively. However, further increase in MBA to 0.1 mol% led to a large decrease of these values to σf of 0.19 MPa, εf of 4.28 mm mm−1, and W of 0.71 MJ m−3. Thus, there exists an optimal MBA concentration of 0.07 mol% for PAAm/GO gels, which achieved E of 66 kPa, σf of 0.27 MPa, εf of 13.76 mm mm−1, and W of 2.52 MJ m−3.
image file: c7py01068k-f3.tif
Fig. 3 (a) Original stretching, (b) twisted stretching, and (c) knotted stretching of PAAm/GO hybrid NC gels.

image file: c7py01068k-f4.tif
Fig. 4 Effect of MBA concentrations on the tensile properties of PAAm/GO nanocomposite gels. GO concentration is fixed at 4 mg mL−1.

Fig. 5a shows a similar mechanical enhancement (E, σf, εf and W) of PAAm/GO hydrogels as GO concentration increased, specifically, when GO increased from 0 to 4 mg mL−1, σf and W increased from 0.01 to 0.27 MPa and 0.01 to 2.52 MJ m−3, respectively, indicating that PAAm/GO gels are 27 times stronger and 251 times tougher than pure PAAm gels (Fig. 5b). Meanwhile, E and εf also increased from 27 to 66 kPa and 0.91 to 13.76 mm mm−1, indicating that PAAm/GO gels are also much stiffer (∼2.4 times) and more stretchable (∼15 times) than pure PAAm gels (Fig. 5c). To better analyze the stress–strain curves of PAAm/GO gels as per GO concentration, Fig. 5d shows the reduced stress (σr) of PAAm/GO gels as a function of 1/λ, where σr is defined by

image file: c7py01068k-t4.tif


image file: c7py01068k-f5.tif
Fig. 5 Effect of GO concentrations on the tensile properties of PAAm/GO nanocomposite gels. (a) Stress–strain curves of PAAm/GO gels at different GO concentrations; (b) fracture stress and deformed energy and (c) elastic modulus and fracture strain as a function of GO concentrations; (d) reduced stress of PAAm/GO gels at different GO concentrations. MBA is fixed at 0.07 mol% of AAm.

It can be seen that σr was almost independent of 1/λ at GO concentrations of <2 mg mL−1. However, at GO > 2 mg mL−1, σr was also nearly independent of 1/λ at small strains (λ < 1.67 or 1/λ > 0.6), while strain hardening became significant at large strains (λ > 1.67 or 1/λ < 0.6). This indicates the existence of weak physical interactions in PAAm/GO gels so that the uniaxial deformation can be well described by the rubber elasticity model. Since σr is nearly constant within the Hookean regime, we assume the constant σr is equal to the equilibrium shear modulus (Ge), which can be used to estimate the effective network chain density (N) in hydrogels by Ge = NRT, where R and T are the gas constant and absolute temperature, respectively. The estimated N of PAAm/GO gels increased from 5.37, 5.85, 6.13, 7.38, to 11.62 mol m−3 as GO increased from 0, 0.5, 1, 2, to 4 mg mL−1, suggesting that the crosslinking density in PAAm/GO gels increases as per GO concentrations. Swelling tests in Fig. S1 further showed that upon swelling, the pure PAAm gels increased their sizes significantly, while the PAAm/GO gels reduced their swollen size with the increase of GO concentrations. This confirms that GO nanosheets act as multifunctional cross-linkers in the PAAm/GO gels. Taken together, considering that a very small amount of GO (∼1.67 wt%) was used to prepare the PAAm/GO gels, it is clear that GO nanosheets are very effective nanofillers and cross-linkers to improve the mechanical properties of host hydrogels.

We found that PAAm/GO hybrid hydrogels achieved not only high strength but also high toughness in the presence of a very small amount of MBA (<0.1 mol%). However, if MBA = 0.1 mol%, the tensile properties of the PAAm/GO gel decreased clearly, i.e., to σf of 0.19 MPa, εf of 4.28 mm mm−1, and W of 0.71 MJ m−3, which is much weaker than that of the gel with MBA = 0.07 mol% (σf of 0.27 MPa, εf of 13.76 mm mm−1, and W of 2.52 MJ m−3). The decreased mechanical properties of PAAm/GO hybrid hydrogels as MBA concentration could be due to several factors. First, the introduction of a small amount of chemical cross-linkers (<0.1 mol%) can increase the effective elastic chains in the gels, leading to an increase of the elastic modulus (E) of the gels. As shown in Fig. 4 and Table S1, the E of the gels increased from 49.7 to 93.3 kPa as the MBA increased from 0.01 to 0.1 mol%. Second, the introduction of a small amount of chemical cross-linkers (<0.1 mol%) can also enhance the stress transfer in the network. Finally, if MBA ≥ 0.1 mol%, too many cross-linkings are introduced in the network, which will cause stress to mainly concentrate on the shortest chains. So, during the fracture process, the shortest chains will be the first, and easily, broken before the peeling-off of the PAAm chains from the GO nanosheets. As a result, the PAAm/GO gel linked by 0.1 mol% of MBA showed a weaker strength (0.19 MPa) and small deformation energy (0.71 MJ m−3).

Tearing tests were performed to quantify the toughness of PAAm/GO nanocomposite gels. As shown in Fig. 6a, the tearing energies (T) of the PAAm/GO gels were 341, 344, 577, and 964 J m−2 at GO of 0.5, 1, 2, and 4 mg mL−1, respectively, showing that the gel toughness increases with the GO concentration. In particular, for GO of 4 mg mL−1, the PAAm/GO gel exhibited high toughness (964 J m−2), comparable to cartilage (102–103 J m−2), rubber (102–103 J m−2), and double-network hydrogels (102–104 J m−2). We also examined the crack propagation of the notched PAAm/GO gel prepared with GO of 4 mg mL−1. In Fig. 6b, when PAAm/GO gel with a cut notch (∼10 mm) was stretched up to 3 times, the notch was dramatically blunted and remained stable, indicating that no stress concentrates in the front of the notch tip. It was reported that strong blunting behavior occurs at a σf/E of 2 or higher, while weak blunting behavior occurs at 0.4 < σf/E < 2.41,42 So, our PAAm/GO gel prepared with GO of 4 mg mL−1 showed a high σf/E of 4 (E of 66 kPa and σf of 0.27 MPa) and a similar blunting behavior to polyampholyte hydrogels,41 PDGI/PAAm gels43 and Agar/PAAm gels.44 Both tearing and crack propagation data confirm that PAAm/GO hybrid NC gels are also tough hydrogels.


image file: c7py01068k-f6.tif
Fig. 6 (a) Tearing energies of PAAm/GO gels as a function of GO concentrations and (b) crack propagation by stretching a notched PAAm/GO gel.

It is generally accepted that the stiffness and toughness of conventional materials are two opposite mechanical parameters. Interestingly, our PAAm/GO gels can simultaneously increase both stiffness and toughness with increasing GO concentration. This could be attributed to hybrid physical and chemical cross-linkings in the PAAm/GO gels. Since GO nanosheets act as multifunctional cross-linkers, the physical interactions between PAAm chains and GO nanosheets are enhanced with the increase of GO concentration. The increase in the elastic modulus (indicator of the stiffness) of PAAm/GO NC gels is attributed to the increase in elastically effective PAAm chains, as identified by the increase of effective network chain density (N) of NC gels from 5.37 to 11.62 mol m−3 as GO increased from 0 to 4 mg mL−1. The rupture (or peeling) of physically adsorbed PAAm chains from GO nanosheets is accompanied by energy dissipation, which consequently improves the crack resistance and tearing energies (indicator of the toughness) of PAAm/GO NC gels. In addition, as more PAAm chains are adsorbed on GO nanosheets, more energy is dissipated. Therefore, the toughness of PAAm/GO NC gels continues to increase with GO concentration. Simultaneous enhancement of stiffness and toughness has also been reported in other hybrid hydrogels. Lin et al.45 reported that poly(dimethylacrylamide)(PDMA)/silica hybrid NC gel increased its elastic modulus and tearing energies with the increase in silica content. Chen et al.44 and Li et al.46 also observed the same phenomenon in agar/PAAm hybrid double network (DN) gels and Ca2+-alginate/PAAm hybrid gels, respectively. These results indicate that the combination of physical and chemical cross-linkings in hydrogels is a feasible strategy to simultaneously achieve high stiffness and toughness of hydrogels.

3.3. Stretching-dependent tensile and energy dissipation of PAAm/GO gels

To evaluate energy dissipation during the fracture process of PAAm/GO nanocomposite gels, Fig. 7a shows the cyclic loading–unloading curves of PAAm/GO gels prepared at different GO concentrations of 0.5–4 mg mL−1. It can be seen that all PAAm/GO gels exhibited a distinct hysteresis loop at λ = 3. As GO concentrations increased from 0.5 to 4 mg mL−1, the loop area became larger and dissipated energy (Uhys) increased from 1.2, 8.2, 14.8, to 34.8 kJ m−3 (Fig. 7b). It is not surprising that at a very low GO concentration of 0.5 mg mL−1, interactions between PAAm chains and GO nanosheets were very weak, leading to the smallest hysteresis loop and energy dissipation. Furthermore, cyclic loading tests were performed on different fresh PAAm/GO gels in a series of different maximum strains (λmax). Evidently, hysteresis loop areas increased with λmax (Fig. 7c), i.e. from an almost negligible hysteresis loop at λmax = 2 to the largest loop at λmax = 10. Consequently, Uhys increased from 104.6 kJ m−3 at λmax = 4 to 943.3 kJ m−3 at λmax = 10 (Fig. 7d). These results suggest that PAAm/GO gels possess clear GO concentration-dependent and λ-dependent energy dissipation behaviors, and effective energy dissipation only occurs at the higher GO concentrations (GO > 1 mg mL−1) and higher λmax (λmax > 2).
image file: c7py01068k-f7.tif
Fig. 7 (a) Loading–unloading curves and (b) the corresponding dissipated energies of PAAm/GO nanocomposite gels prepared at different GO concentrations; (c) cyclic loading curves and (d) the corresponding dissipated energies of PAAm/GO gels (GO = 4 mg mL−1) at different maximum stretching ratios (λmax). Inset image in (c) is performed at λmax = 2.

PAAm/GO gels also showed strain-rate dependent tensile properties. In Fig. 8a, σf, εf, and W all increased as the strain rate increased from 0.006 to 0.06 s−1, however, these properties decreased at a strain rate of >0.06 s−1 (Fig. 8b and Table S2). A too fast or too slow stretching rate seems to compromise the mechanical properties of PAAm/GO gels, and only at an optimal stretching rate of 0.06 s−1 can the PAAm/GO gel achieve the best tensile properties (E of 68 kPa, σf of 0.27 MPa, εf of 17.58 mm mm−1, and W of 2.99 MJ m−3). Consistently, the PAAm/GO gels also exhibited stretching rate-dependent energy dissipation behavior. In Fig. 8c, all different strain rates led to large hysteresis loops. At 0.06 s−1, the gel showed the largest hysteresis loop and dissipated the highest energy of Uhys = 48.5 kJ m−3. However, when the strain rates were smaller or larger than 0.06, the gels dissipated the smaller energy in between 29.0–34.8 kJ m−3. The rate-dependent tensile and energy dissipation behaviors of PAAm/GO gels further indicate the existence and importance of physical interactions between PAAm and GO nanosheets.


image file: c7py01068k-f8.tif
Fig. 8 (a) Tensile stress–strain curves and (b) the corresponding fracture stress (E) and deformed energy (W) of PAAm/GO gels as a function of strain rates from 0.006 to 0.244 s−1; (c) loading–unloading curves and (b) the corresponding dissipated energy (Uhys) of PAAm/GO gels (GO = 4 mg mL−1) as a function of strain rates from 0.006 to 0.275 s−1.

3.4. Mullins effect of PAAm/GO gels

Fig. 9a shows the successive six loading–unloading cycles on the same PAAm/GO gel at λ = 5. During the loading–unloading tests, no resting time was given between any two consecutive loading cycles. The gel also exhibited a large hysteresis loop of 195.6 kJ m−3 at the 1st loading–unloading cycle. During the subsequent 5 loading cycles, the gel showed an almost negligible hysteresis loop, where the loading/unloading curves were nearly overlapped with the unloading curve of the 1st cycle. This indicates that the PAAm/GO gel lost most of its toughness, behaved like elastic materials, and experienced permanent deformation with almost no recovery. Similarly, Wang and co-workers found that both PAAm/GPO hydrogels31 and PAAm/PVP/GO gels47 also exhibited very small hysteresis loops after the 1st loading. In Fig. 9b, E decreased from 66 kPa in the 1st loading to 32 kPa in the 2nd loading, and then remained almost unchanged in the 3rd–6th cycles. In this way, the gel lost ∼52% of elastic moduli after the 1st loading. The loss of hysteresis and elastic modulus after the 1st loading cycle suggests a significant softening phenomenon in the PAAm/GO gels. Similarly a severe softening phenomenon was also observed in the chemically linked PAMPS/PAAm DN gels48 and hybrid Agar/PAAm DN gels.38
image file: c7py01068k-f9.tif
Fig. 9 (a) Successive six times loading–unloading curves of the same PAAm/GO gel at λ = 5 without resting between two cycles; (b) elastic modulus of the PAAm/GO gel at different loading cycles.

Furthermore, we continued to challenge the same PAAm/GO gel using successive loading–unloading experiments at different maximum extension ratios (λmax) without applying resting between two loadings. Different from Fig. 9, Fig. 10a showed that the PAAm/GO gels exhibited a distinct hysteresis loop at each loading cycle, particularly when λmax > 2, the pronounced hysteresis loops showed the continuously increased stress with applied strains up to 16. Additionally, E gradually decreased from 66 kPa at λmax = 2 to 20 kPa at λmax = 16, leading to the loss of ∼85% of the initial elastic modulus (Fig. 10b). Furthermore, the energy dissipation (Uhys) of the PAAm/GO gels showed an exponential increase as λmax with an R2 value of 0.9999, indicating a gradual fracture process of the gel network during deformation (Fig. 10c). Energy dissipation efficiency (Uhys/W) is often used to quantify the energy dissipation ability of tough gels. Fig. 10d shows that Uhys/W also increased as λmax and the highest Uhys/W was only ∼38%, which is smaller than that of vulcanized rubbers (59–67%),49 Agar/PAAm DN gels (50–60%)38 and PAMPS/PAAm DN gels (∼85%).49 Overall, different loading–unloading tests clearly reveal the softening phenomenon of the PAAm/GO gels, together with a significant loss of hysteresis and elastic modulus, similar to the “Mullins effect” observed in the nanofilled rubbers and some hydrogels with strong elastic characteristics.


image file: c7py01068k-f10.tif
Fig. 10 (a) Cyclic loading–unloading curves, (b) elastic modulus (inset is the successive loading curve at λmax = 2), (c) dissipated energy, and (d) energy dissipation efficiency of the same PAAm/GO hybrid gel specimen at different λmax without resting between two cycles.

3.5. Self-recovery of PAAm/GO gels

While the immediate loading–unloading tests show no or very weak self-recovery properties of PAAm/GO gels, the presence of physical interactions between PAAm and GO is expected to induce some self-recovery properties of PAAm/GO gels under appropriate conditions. To test this hypothesis, we further conducted cyclic loading–unloading tests to examine the possible self-recovery properties of PAAm/GO gels at different resting times. Two recovery ratios (%) of toughness recovery and stiffness recovery are defined, in which toughness recovery (or stiffness recovery) is calculated by a ratio of dissipated energies (or elastic modulus) at different recovery times to that of the first loading cycle at λmax = 5. In Fig. 9a and 11a, as a control, when no resting time was applied between two consecutive loadings, the hysteresis loop was almost negligible, which was accompanied by a negligible toughness recovery and 46% of stiffness recovery. This indicates that upon network deformation, the re-establishment of PAAm–GO interactions cannot occur immediately. However, when the resting time slightly increased to 3 min, a significant hysteresis loop was observed, and as expected, the loop areas became larger with the increase of resting time. Fig. 11b shows that toughness/stiffness recovery rates were 32%/59%, 46%/64%, 51%/67%, 54%/74% and 62%/76% at resting times of 3, 5, 10, 15, and 30 min, respectively. PAAm/GO hydrogels exhibited an clearly time-dependent two-stage recovery, which increased with rest time. The PAAm/GO gel experienced rapid recovery of its toughness and stiffness within a very short time resting at room temperature, i.e. the gel can recover its toughness from 0 to 46% and stiffness from 46 to 64% in 5 min. With a further increase in resting time to 30 min, the toughness and stiffness recovery gradually approached 62% and 76%, respectively.
image file: c7py01068k-f11.tif
Fig. 11 (a) Loading–unloading curves and (b) the corresponding toughness and stiffness recovery rates of the PAAm/GO gel at different resting times and at room temperature.

We also have tested the self-healing of PAAm/GO gels. Unfortunately, PAAm/GO gels did not show distinct self-healing properties. When we physically place two cut gels in contact with each other, the gels can be healed into one gel, but the healed gel cannot withstand even very gentle stretching, and will break very easily. Thus, we did not include the self-healing data in this manuscript. The lack of self-healing properties could be attributed to the covalent cross-linking in the PAAm network by a chemical cross-linker (MBA). The presence of covalent cross-linking will slow down the diffusion ability of polymer chains to the fracture interface, leading to the negligible self-healing of PAAm-GO hydrogels. Consistently, in our previous work,37 we have found that after introducing a very small amount of MBA (0.7 mol%) to PAAm/LAPONITE® NC gels, the self-healing efficiency of the NC gels significantly decreased. These results show a general observation that the presence of covalent cross-linking (even at a very small amount) would greatly compromise the self-healing of nanocomposite hydrogels.

3.6. Toughening mechanisms of PAAm/GO gels

To design strong and tough hydrogels with novel micro-/nanostructures, it is critical to better understand the toughening and energy dissipation mechanisms of our PAAm/GO nanocomposite gels. As shown in Fig. 7a and b, the PAAm/GO gel demonstrated its distinct hysteresis loops during the loading–unloading process, depending on the GO concentrations. During the deformation process, PAAm chains that are physically adsorbed on GO nanosheets are expected to be first pulled out from the GO nanosheets, followed by the continuous breaking of chemically grafted PAAm chains from GO nanosheets. So, GO concentration is critical to serve as a skeleton template to adsorb and graft PAAm chains. When the GO concentration (<0.5 mg mL−1) is low, PAAm chains, regardless whether chemically grafted or physically adsorbed, on GO nanosheets would be small, which leads to weak PAAm–GO interactions that cause small Uhys during the fracture process (Fig. 7a and b). As GO concentrations increase, more PAAm chains are likely to physically attach to GO nanosheets to form a stronger PAAm–GO network. Consequently, the facture process will cost high energy dissipation to peel off these PAAm from the GO surface. Upon deformation, the physical and chemical bonds between PAAm and GO are expected to be fractured and participate in energy dissipation via the dissociation and breaking, resulting in high toughness of PAAm/GO gels.

The PAAm/GO gels exhibit λ-dependent energy dissipation behaviors which have not been reported in the GO-based hydrogels. The possible energy dissipation mechanism was shown in Fig. 12. At λ ≤ 2, there was no hysteresis loop observed during the loading–unloading cycle, indicating that the PAAm/GO hybrid NC gels showed excellent resilience. At small strain, the adsorbed polymer chains on the GO nanosheets were only highly stretched but not peeled off. However, at λ > 2, the NC gels displayed distinct hysteresis loops, indicating that the adsorbed PAAm chains were partially peeled off from the GO nanosheets. The dissipated energies (Uhys) also increased with the increase of the maximum extension ratio (λmax), which means that the peeled off PAAm chains became more and more as the λmax increased. At a fixed λ of 5, successive loading of the same gel specimen without recovery demonstrated an obvious hysteresis loop at the first loading cycle but a negligible hysteresis loop at the following five times loading cycle (Fig. 9a). Moreover, successive loading experiments at various λmax also showed that, if there were no recovery time for NC gels, the sequential loading curve would be overlapped with the unloading curve of the former ones as the λ wasn't larger than the former λmax. At the same time, the NC gels became softer after the first loading cycle and were even much softer at larger λmax (Fig. 9b and 10b). The results elucidated that our PAAm/GO hybrid NC gels exhibited a “Mullins effect” phenomenon which is commonly detected in filled rubbers. The mechanisms of the Mullins effect in our PAAm/GO hybrid NC gels can be interpreted by a “bond rupture” model which has been proposed by Blanchard and Parkinson in filled rubbers (Fig. 12). Since GO nanosheets were acted as multi-functional cross-linkers in our PAAm/GO hybrid NC gels, the rupture of physical or chemical bonds, i.e. peeling off of PAAm chains from the surface of GO nanosheets, will cause the elastically effective chains to become smaller, which leads to a decrease in the elastic modulus after pre-strain. The absence of hysteresis loops at the second loading cycle indicated that the re-adsorption of PAAm chains onto the surface of GO nanosheets cannot occur immediately. However, after recovery at room temperature for only 5 min, the stiffness and toughness of the PAAm/GO hybrid NC gel could reach recovery rates of 64% and 46%, respectively. The time-dependent recovery of stiffness and toughness suggested that our NC gels also demonstrated self-recovery properties (i.e. Mullins recovery) at room temperature. The Mullins recovery in our NC gels can be explained by the re-adsorption process of PAAm chains. Owing to the strong interactions between the GO nanosheets and PAAm chains, the peeled off PAAm chains trend to be re-adsorbed onto the surface of GO nanosheets. As the recovery time was within 5 min, the initial re-adsorption process was rapid, however, the recovery was slowed down by the steric effect of re-adsorbed PAAm chains (Fig. 11b). Moreover, it is believed that only physical adsorption can be reformed but the ruptured chemical adsorption cannot be recovered.


image file: c7py01068k-f12.tif
Fig. 12 Illustration of λ-dependent energy dissipation behaviors and mechanisms of PAAm/GO hybrid NC gels.

3.7. Comparison with CB-filled rubbers and other hydrogels

PAAm/GO nanocomposite gels vs. CB-filled rubbers. It is worthwhile comparing PAAm/GO nanocomposite gels with CB-filled rubbers to reveal some similarities and differences for better understanding the structure–property relationship of the PAAm/GO gels. First, there are some similarities between two systems: (i) both systems contain some physical and chemical interactions between fillers and polymer chains, due to a similar hybrid network structure; (ii) both systems can improve their mechanical properties by incorporating GO nanosheets or CB particles into polymer networks; and (iii) both systems show a typical Mullins effect and Mullins recovery phenomena. However, some differences are also detected between the two systems: (i) PAAm/GO gels contain a large amount of water (up to ∼80%), while CB-filled rubbers do not contain any solvent in the polymer matrix; (ii) PAAm/GO gels demonstrate the time-dependent Mullins self-recovery of mechanical properties at room temperature, while CB-filled rubbers only exhibit the Mullins recovery upon heating to a relative high temperature (80 °C);50 and (iii) clearly, PAAm/GO gels and CB-filled rubbers show different mechanisms of energy dissipation, Mullin effect, and mechanical recovery. As shown in Fig. 12, the energy dissipation and Mullin effect of PAAm/GO gels can be well interpreted by the “bond rupture” model, while CB-filled rubbers are often explained by the “molecular slippery” model. Upon the deformation of CB-filled rubbers, the adsorbed rubber chains would slip onto the surface of CB particles. The friction caused by the molecular slipperiness of rubber chains will lead to heat loss and mechanical hysteresis.50 In this work, no data can support the slipperiness between GO nanosheets and PAAm chains.
PAAm/GO nanocomposite gels vs. other NC gels. It is also interesting to compare our PAAm/GO gels with other NC gels. It is a common strategy to use LAPONITE® nanoclay and SiO2 nanoparticles to fabricate NC gels, which are selected for comparison with our PAAm/GO NC gels. Tang et al.51 found that PNIPAM/LAPONITE® NC gels displayed hysteresis loops even at very small λmax of 1.6, and hysteresis loops increased as λmax. They proposed that PNIPAM/LAPONITE® NC gels dissipated the energy via two ways of the orientation of clay nanoplatelets and the peeling-off of polymer chains. Besides that both PAAm/GO gels and PNIPAM/LAPONITE® gels exhibited λ-dependent energy dissipation behaviors, Yang et al.37 recently reported that PAAm/LAPONITE® gels also exhibited the Mullins effect and self-recovery properties at room temperature, suggesting that both PAAm/LAPONITE® gels and PAAm/GO gels show a similar Mullins recovery mechanism. However, different from PNIPAM/LAPONITE® gels, our PAAm/GO gels showed a negligible hysteresis loop at small strains. Rose et al.52 found that PDMA/Silica NC gels dissipated more energy as strain rates and SiO2 concentration increased, while Lin et al.45 found that PDMA/Silica NC gels can fully recover their elastic modulus during successive loading–unloading cycles without any resting and such mechanical recovery was also rate-dependent.52 The fast mechanical recovery of PDMA/silica gels is attributed to reversible physical adsorption and desorption of PDMA chains on silica surfaces, leading to no permanent damage occurring in the network structure.53 A comparison of both PDMA/silica gels and PAAm/GO gels reveals that (i) both gels exhibited a similar Mullins recovery at room temperature, but the time-dependent recovery of our PAAm/GO gels was much slower than that of PDMA/silica gels and (ii) both gels exhibited rate-dependent tensile properties and energy dissipation, but differently from PDMA/silica gels whose mechanical strength and dissipated energies increase with the increase in the strain rate, PAAm/GO gels show the highest strength and the largest dissipated energies at an optimal strain rate (Fig. 8b and d).
PAAm/GO nanocomposite gels vs. double network hydrogels with hybrid cross-linkings. It is generally accepted that fully chemically cross-linked double network hydrogels (DN gels) usually show large hysteresis and Mullins effects, where energy dissipation mainly stems from the permanent breakage of a covalent bond,54 leading to irreversible network damage and lack of mechanical recovery. However, it is more appropriate to compare our PAAm/GO gels with DN gels with hybrid chemical and physical cross-linkings (hybrid DN gels). Herein, agar/PAAm hybrid DN gels are selected for comparison with PAAm/GO gels, because both gels contain physical interactions, particularly hydrogen bonds, in the networks.55 Both DN and nanocomposite gels exhibit similar high mechanical and Mullins self-recovery properties. Due to the presence of physically cross-linked networks, both gels dissipated energy via the fracture of physical interactions involved in the networks. However, agar/PAAm gels achieved mechanical recovery only at a high temperature (above the melting point of the agar gel), which was different from the self-recovery of PAAm/GO gels at room temperature. The heating-induced mechanical recovery of agar/PAAm hybrid DN gels is similar to the Mullins recovery of CB-filled rubbers, indicating that the fracture of physical interactions in PAAm/GO gels and Agar/PAAm gels is different. The fracture of agar/PAAm hybrid DN gels leads to the “chain pulling-out” of agar chains from their agar helical bundles; such a chain dissociation in the agar network cannot be recovered spontaneously at room temperature.38 In contrast, the fracture of interactions between PAAm and GO is the “peeling-off” mechanism, which can be a rapid recovery at room temperature.

4. Conclusions

In summary, we presented highly tough PAAm/GO hybrid nanocomposite gels using in situ free radical polymerization. The incorporation of GO nanosheets into PAAm gels enables simultaneous improvement of both toughness and stiffness with GO concentration. Under optimal conditions, PAAm/GO gels exhibited an elastic modulus of 66 kPa, a fracture stress of 0.27 MPa, a fracture strain of 13.76 mm mm−1, and deformation energy of 2.52 MJ m−3. The tearing energy of PAAm/GO gels can reach 964 J m−2, comparable to that of cartilage (102–103 J m−2), rubbers (102–103 J m−2) and double network hydrogels (102–104 J m−2). More interestingly, PAAm/GO gels demonstrated λ-dependent energy dissipation behaviors. During the loading–unloading tests, NC gels displayed a negligible hysteresis loop at λ ≤ 2, but large hysteresis loops at λ > 2, indicating that the deformation energy (Uhys) is dissipated more efficiently at large strains (λmax). The decrease in the elastic modulus upon cyclic loading–unloading is a typical indicator of the “Mullins effect” similar to that of CB-filled rubbers and double network hydrogels. Consequently, PAAm/GO gels exhibited fast self-recovery properties at room temperature, allowing achieving 76% and 62% of stiffness and toughness recovery after 30 min of resting at room temperature. The mechanical properties, Mullins effect, energy dissipation, and self-recovery of PAAm/GO gels can be interpreted by the “bond rupture” model. Different from the other nanocomposite elastomers and hydrogels, PAAm/GO gels provide different viewpoints to better understand the structure–property relationship of hybrid nanocomposite gels, which may be helpful for this class of hybrid hydrogels with desirable properties.

Conflict of interest

The authors declare no competing financial interest.

Acknowledgements

Q. C. is grateful for financial support, in part, from the National Nature Science Foundation of China (21504022), the Joint Fund for Fostering Talents of NSFC-Henan Province (U1304516), Henan Province (No. NSFRF1605, No. 2016GGJS-039, No. 13A430015 and No. 17HASTIT006) and Henan Polytechnic University (72105/001 and 672517/005). J. Z. thanks financial support from the NSF (DMR-1607475) and in part from the NSF (CBET-1510099).

References

  1. T. R. Hoare and D. S. Kohane, Polymer, 2008, 49, 1993–2007 CrossRef CAS.
  2. H. K. Lau and K. L. Kiick, Biomacromolecules, 2015, 16, 28–42 CrossRef CAS PubMed.
  3. K. Kabiri, H. Omidian, M. J. Zohuriaan-Mehr and S. Doroudiani, Polym. Compos., 2011, 32, 277–289 CrossRef CAS.
  4. S. Naficy, H. R. Brown, J. M. Razal, G. M. Spinks and P. G. Whittern, Aust. J. Chem., 2011, 64, 1007 CrossRef CAS.
  5. H. Chen, F. Yang, Q. Chen and J. Zheng, Adv. Mater., 2017, 29, 1606900 CrossRef PubMed.
  6. X. Zhao, Soft Matter, 2014, 10, 672–687 RSC.
  7. J. P. Gong, Y. Katsuyama, T. Kurokawa and Y. Osada, Adv. Mater., 2003, 15, 1155–1158 CrossRef CAS.
  8. Q. Chen, H. Chen, L. Zhu and J. Zheng, J. Mater. Chem. B, 2015, 3, 3654–3676 RSC.
  9. Q. Chen, H. Chen, L. Zhu and J. Zheng, Macromol. Chem. Phys., 2016, 217, 1022–1036 CrossRef CAS.
  10. J. Zhang, N. Wang, W. Liu, X. Zhao and W. Lu, Soft Matter, 2013, 9, 6331–6337 RSC.
  11. T. Bai, P. Zhang, Y. Han, Y. Liu, W. Liu, X. Zhao and W. Lu, Soft Matter, 2011, 7, 2825–2831 RSC.
  12. M. Guo, L. M. Pitet, H. M. Wyss, M. Vos, P. Y. W. Dankers and E. W. Meijer, J. Am. Chem. Soc., 2014, 136, 6969–6977 CrossRef CAS PubMed.
  13. W. Li, H. An, Y. Tan, C. Lu, C. Liu, P. Li, K. Xu and P. Wang, Soft Matter, 2012, 8, 5078–5086 RSC.
  14. S. Abdurrahmanoglu, V. Can and O. Okay, Polymer, 2009, 50, 5449–5455 CrossRef CAS.
  15. Q. Chen, L. Zhu, H. Chen, H. Yan, L. Huang, J. Yang and J. Zheng, Adv. Funct. Mater., 2015, 25, 1598–1607 CrossRef CAS.
  16. T. L. Sun, T. Kurokawa, S. Kuroda, A. B. Ihsan, T. Akasaki, K. Sato, M. A. Haque, T. Nakajima and J. P. Gong, Nat. Mater., 2013, 12, 932–937 CrossRef CAS PubMed.
  17. T. Huang, H. G. Xu, K. X. Jiao, L. P. Zhu, H. R. Brown and H. L. Wang, Adv. Mater., 2007, 19, 1622–1626 CrossRef CAS.
  18. T. Sakai, T. Matsunaga, Y. Yamamoto, C. Ito, R. Yoshida, S. Suzuki, N. Sasaki, M. Shibayama and U.-I. Chung, Macromolecules, 2008, 41, 5379–5384 CrossRef CAS.
  19. K. Haraguchi and T. Takehisa, Adv. Mater., 2002, 14, 1120–1124 CrossRef CAS.
  20. K. Xu, J. Wang, Q. Chen, Y. Yue, W. Zhang and P. Wang, J. Colloid Interface Sci., 2008, 321, 272–278 CrossRef CAS PubMed.
  21. G. Gao, G. Du, Y. Sun and J. Fu, ACS Appl. Mater. Interfaces, 2015, 7, 5029–5037 CAS.
  22. K. Haraguchi, Curr. Opin. Solid State Mater. Sci., 2007, 11, 47–54 CrossRef CAS.
  23. Y. Zhu, S. Murali, W. Cai, X. Li, J. W. Suk, J. R. Potts and R. S. Ruoff, Adv. Mater., 2010, 22, 3906–3924 CrossRef CAS PubMed.
  24. J. Gu, X. Yang, Z. Lv, N. Li, C. Liang and Q. Zhang, Int. J. Heat Mass Transfer, 2016, 92, 15–22 CrossRef CAS.
  25. C. Hu, Z. Li, Y. Wang, J. Gao, K. Dai, G. Zheng, C. Liu, C. Shen, H. Song and Z. Guo, J. Mater. Chem. C, 2017, 5, 2318–2328 RSC.
  26. H. Liu, M. Dong, W. Huang, J. Gao, K. Dai, J. Guo, G. Zheng, C. Liu, C. Shen and Z. Guo, J. Mater. Chem. C, 2017, 5, 73–83 RSC.
  27. L. Zhang, Q. Zhang, H. Xie, J. Guo, H. Lyu, Y. Li, Z. Sun, H. Wang and Z. Guo, Appl. Catal., B, 2017, 201, 470–478 CrossRef CAS.
  28. J. Shen, B. Yan, T. Li, Y. Long, N. Li and M. Ye, Composites, Part A, 2012, 43, 1476–1481 CrossRef CAS.
  29. J. Shen, B. Yan, T. Li, Y. Long, N. Li and M. Ye, Soft Matter, 2012, 8, 1831–1836 RSC.
  30. R. Liu, S. Liang, X.-Z. Tang, D. Yan, X. Li and Z.-Z. Yu, J. Mater. Chem., 2012, 22, 14160–14167 RSC.
  31. J. Liu, C. Chen, C. He, J. Zhao, X. Yang and H. Wang, ACS Nano, 2012, 6, 8194–8202 CrossRef CAS PubMed.
  32. J. Liu, G. Song, C. He and H. Wang, Macromol. Rapid Commun., 2013, 34, 1002–1007 CrossRef CAS PubMed.
  33. H.-P. Cong, P. Wang and S.-H. Yu, Small, 2014, 10, 448–453 CrossRef CAS PubMed.
  34. H.-P. Cong, P. Wang and S.-H. Yu, Chem. Mater., 2013, 25, 3357–3362 CrossRef CAS.
  35. K. Shi, Z. Liu, Y.-Y. Wei, W. Wang, X.-J. Ju, R. Xie and L.-Y. Chu, ACS Appl. Mater. Interfaces, 2015, 7, 27289–27298 CAS.
  36. P. Lin, S. Ma, X. Wang and F. Zhou, Adv. Mater., 2015, 27, 2054–2059 CrossRef CAS PubMed.
  37. J. Yang, L. Zhu, X. Yan, D. Wei, G. Qin, B. Liu, S. Liu and Q. Chen, RSC Adv., 2016, 6, 59131–59140 RSC.
  38. Q. Chen, L. Zhu, L. Huang, H. Chen, K. Xu, Y. Tan, P. Wang and J. Zheng, Macromolecules, 2014, 47, 2140–2148 CrossRef CAS.
  39. L. Kan, Z. Xu and C. Gao, Macromolecules, 2011, 44, 444–452 CrossRef CAS.
  40. J. Ma, Z. Zhu, B. Chen, M. Yang, H. Zhou, C. Li, F. Yu and J. Chen, J. Mater. Chem. A, 2013, 1, 4662 CAS.
  41. F. Luo, T. L. Sun, T. Nakajima, T. Kurokawa, Y. Zhao, A. B. Ihsan, H. L. Guo, X. F. Li and J. P. Gong, Macromolecules, 2014, 47, 6037–6046 CrossRef CAS.
  42. C. Y. Hui, A. Jagota, S. J. Bennison and J. D. Londono, Proc. R. Soc. London, Ser. A, 2003, 459, 1489 CrossRef.
  43. M. A. Haque, T. Kurokawa, G. Kamita and J. P. Gong, Macromolecules, 2011, 44, 8916–8924 CrossRef CAS.
  44. Q. Chen, D. Wei, H. Chen, L. Zhu, C. Jiao, G. Liu, L. Huang, J. Yang, L. Wang and J. Zheng, Macromolecules, 2015, 48, 8003–8010 CrossRef CAS.
  45. W.-C. Lin, W. Fan, A. Marcellan, D. Hourdet and C. Creton, Macromolecules, 2010, 43, 2554–2563 CrossRef CAS.
  46. H. Li, A. M. Koenig, P. Sloan and N. D. Leipzig, Biomaterials, 2014, 35, 9049–9057 CrossRef CAS PubMed.
  47. Z. Zhu, G. Song, J. Liu, P. G. Whitten, L. Liu and H. Wang, Langmuir, 2014, 30, 14648–14657 CrossRef CAS PubMed.
  48. J. P. Gong, Soft Matter, 2010, 6, 2583–2590 RSC.
  49. T. Nakajima, T. Kurokawa, S. Ahmed, W.-L. Wu and J. P. Gong, Soft Matter, 2013, 9, 1955–1966 RSC.
  50. J. Diani, B. Fayolle and P. Gilormini, Eur. Polym. J., 2009, 45, 601–612 CrossRef CAS.
  51. J. Tang, G. Xu, Y. Sun, Y. Pei and D. Fang, J. Appl. Phys., 2014, 116, 244901 CrossRef.
  52. S. Rose, A. Dizeux, T. Narita, D. Hourdet and A. Marcellan, Macromolecules, 2013, 46, 4095–4104 CrossRef CAS.
  53. W.-C. Lin, A. Marcellan, D. Hourdet and C. Creton, Soft Matter, 2011, 7, 6578–6582 RSC.
  54. R. E. Webber, C. Creton, H. R. Brown and J. P. Gong, Macromolecules, 2007, 40, 2919–2927 CrossRef CAS.
  55. Q. Chen, L. Zhu, C. Zhao, Q. Wang and J. Zheng, Adv. Mater., 2013, 25, 4171–4176 CrossRef CAS PubMed.

Footnotes

Electronic supplementary information (ESI) available. See DOI: 10.1039/c7py01068k
Equivalent contribution.

This journal is © The Royal Society of Chemistry 2017