Strong, flexible and thermal-resistant CNT/polyarylacetylene nanocomposite films

Abstract

This paper reports a new nanocomposite made from a combination of high char yield polyarylacetylene (PAA) resin and ductile carbon nanotube (CNT) film. Benefiting from a big specific surface area and the excellent mechanical properties of the CNT film, the resultant nanocomposite conquers the shrinkage and crack defects of neat PAA. This nanocomposite has great flexibility and tensile strength, with moduli of 303 ± 38 MPa and 22 ± 2 GPa respectively. The through-thickness thermal conductivity of the CNT film/PAA composite reaches 1.15 W (mK)⁻¹, seven times higher than that of pristine CNT film, and the electric conductivity increases to 700 S cm⁻¹. Meanwhile, a lock-up effect of PAA on the CNT network ensures good stability of the structure. TG testing demonstrates that in comparison with a CNT film/epoxy composite, the CNT film/PAA composite has a significantly high decomposition temperature with a char yield of up to 90.7%. After carbonization at 900 °C for 0.5 h, the nanocomposite retains over 66% of its tensile strength.

---

Introduction

Carbon nanotube (CNT) film is well known for its extraordinary mechanical and physical performance. CNT film, which is made of intertangled CNTs, is one of the most promising macroscopic materials. It inherits the advantages of individual carbon nanotubes, such as high strength and modulus, remarkable electrical and thermal conductivity,^1–3 low density and high specific surface area. As a multi-functional reinforcement, CNT film has been impregnated with epoxy or bismaleimide (BMI) resin to prepare structural and functional materials. Li et al.⁴ fabricated a CNT film/epoxy composite by resin solution impregnation, which showed a high tensile strength and toughness of 405 MPa and 122 J g⁻¹, as well as excellent damping properties. Wang et al.⁵ prepared an aligned CNT film/BMI composite with a high CNT fraction to achieve a tensile strength of 3.8 GPa and Young’s modulus of 293 GPa. These composites have low glass transition temperatures due to their polymer matrices. Some attempts have been made to improve heat-resistant properties by combining a phenolic resin. An et al.⁶ found that thermal decomposition temperature was enhanced greatly through dispersing 1.5 wt% CNTs into a phenolic resin with a char yield of 62.5%. Moreover, Tai et al.^7,8 reported that a CNT network produced good mechanical properties for its phenolic composites with a tensile strength of 80 MPa and a Young’s modulus of 7.5 GPa. However, the char yield of common phenolic resin (<60%) still cannot meet the requirement for higher thermal stability.^9,10

As a highly cross-linked aromatic polymer containing only carbon and hydrogen, polyarylacetylene (PAA) has excellent heat-resistant properties, and better ablation stability than phenolic resin.^11,12 The char yield of PAA can reach 80–85%, and PAA has lower shrinkage during pyrolysis and moisture absorption,^13,14 thus PAA is an ideal resin for high temperature applications. However, during PAA polymerization, a large curing exotherm and rapid reaction rate may cause flash polymerization and even explosion.¹⁵ Moreover, PAA is brittle and has low structural integrity due to its highly cross-linked structure.¹⁶ Besides, the non-polar features of PAA often cause weak interfacial adhesion with reinforcement.¹⁷ It remains a challenge to prepare defect-free PAA composites to fully utilize its heat-resistant properties. By combining CNT film and PAA resin, and taking advantage of the former’s high specific area, superb toughness and excellent mechanical and functional properties, it is promising to overcome the shortcomings of PAA and improve the thermal performance of the nanocomposite at the same time.

In this paper, a CNT film/PAA composite is successfully prepared to explore a new way to merge the strong points of a CNT network and PAA resin. The effects of CNTs on the curing characteristics and shrinkage of PAA are analysed, and the mechanical properties and thermal and electrical conductivity of the nanocomposite are investigated. Moreover, the degradation of the CNT film/PAA composite under high temperature and its influence on the nanocomposite’s properties are highlighted to reveal its heat-resistant performance.

Experimental

Materials

CNT film was synthesized using a floating catalyst chemical vapor deposition (FCCVD) method. The detailed manufacturing process of the CNT film has been reported in our previous research.^4,18,19 These fabricated CNT films have a diameter of 6–10 nm with 3–8 walls. The pristine CNTs are randomly oriented with an average thickness of 30 μm. They can be aligned via a simple mechanical stretching process using a tensile machine (Instron 5565) with a load cell of 5 kN. To achieve a higher stretching ratio, a 75 wt% ethanol and water solution was uniformly sprayed on the CNT film during stretching. The stretching speed was controlled at 2 mm min⁻¹. In this research, an aligned CNT film with a draw ratio of 35% was prepared for the fabrication of polymer nanocomposites.

PAA resin was produced by East China University of Science and Technology. The resin has a density of 1.03 g cm⁻³ at ambient temperature. Chemical purity acetone was produced by Beijing Chemical Works. Epoxy resin E51 was produced by Bluestar New Chemical Materials Co. Ltd. The E51 epoxy resin is bisphenol-A diglycidyl ether type (DGEBA) with an epoxy value of 0.48–0.54. The hardener, 2-ethyl-4-methyl imidazole, was produced by Beijing Xiangshan United Assistant Factory. The mixing mass ratio of resin to hardener was 100 : 7.

Preparation of CNT film/PAA composite

A CNT film/PAA composite was prepared using a solvent dipping process. Firstly, CNT film was cut into rectangles with dimensions of 40 × 60 cm² and then soaked in a PAA/acetone solution for 1 hour. The CNT film composites impregnated with 5 wt% and 15 wt% PAA solution were denoted as CNFP5 and CNFP15. The fully impregnated CNT film/PAA prepreg was placed in a vacuum oven at 40 °C for 0.5 hour to remove acetone. To prepare the composite, the CNT film prepreg was sandwiched between two pieces of polytetrafluoroethylene film, which were then placed in a steel mold. A hot-press method was applied to compact the CNT stack and cure the resin. The temperature schedule was 115 °C/10 h + 120 °C/4 h + 140 °C/2 h + 160 °C/2 h + 180 °C/2 h + 250 °C/4 h and the curing pressure was controlled at 10 MPa. After the CNT film/PAA composite was cured, the whole assembly was naturally cooled to room temperature.

In comparison, a CNT film composite impregnated with 5 wt% epoxy solution was also prepared using the solvent dipping process above. The temperature schedule was 80 °C/1 h + 125 °C/2 h and the curing pressure was controlled at 10 MPa.

Based on thermogravimetric testing, the pristine CNT film retained 21.9 wt% after heating to 900 °C in air, thus the content of Fe catalyst in the CNT film is calculated to be 15.3 wt%. According to the mass loss of the CNT film before 350 °C, there is about 2.6 wt% oligomers in the CNT film. Hence the content of CNTs in the pristine CNT film is about 82 wt% and the CNT content in the nanocomposites can be inferred in accordance (Table 1).

Table 1 Resin and CNT contents of different CNT film nanocomposites

Name	Sample	Resin content (%)	CNT content (%)
CNF	CNT film	0	82
CNFP5	Nanocomposite from 5 wt% PAA solution	35 ± 5	53 ± 5
CNFP15	Nanocomposite from 15 wt% PAA solution	65 ± 7	29 ± 7
CNFE5	Nanocomposite from 5 wt% epoxy solution	43 ± 6	47 ± 6

---

Characterization

The curing characteristics of PAA and the nanocomposites were characterized using a differential scanning calorimeter (Mettler DSC-1) at a heating rate of 10 °C min⁻¹. Their decomposition temperatures and char yields were measured using a thermogravimetric analyser (Mettler TGA-1) in air and N₂ atmospheres at 10 °C min⁻¹, respectively.

The surface morphologies of the CNT film and nanocomposites were observed using a JEOL JSM-7500F scanning electron microscope (SEM).

For the measurement of tensile properties, the nanocomposites were cut into 25 × 2 mm² rectangular strips. An Instron 5565 with a 500 N load cell was used. The clamping distance was 20 mm and the sample thickness was measured using a micrometer caliper. The displacement was controlled at 0.5 mm min⁻¹.

The degree of orientation of CNTs in the aligned CNT film can be measured using polarized Raman spectroscopy (Renishaw RM2000). The intensities of the G-band were recorded when an incident laser beam (514 nm) was placed parallel and perpendicular to the alignment direction of the CNT film. The ratio of I_G‖/I_G⊥was adopted to describe the orientation degree of the CNTs.

Electrical conductivity was measured using a four-probe conductivity meter (RTS-9, 4probes tech, Inc.). Thermal conductivity was tested using a laser flash thermal analyzer (LFA447, NETZSCH, Inc.). The monolayer nanocomposite film was attached to a 0.3 mm thick aluminum sheet to meet the thickness requirement. A double-layer mode in the testing software was selected to measure the through-thickness thermal diffusivity. Thermal conductivity can be calculated using the following equation:

λ = α × C_p × ρ (1)

wherein α is the thermal diffusivity, C_p is the specific heat capacity and ρ is the density. The in-plane and through-thickness thermal diffusivities of the eight-layered CNT film/PAA composite can be measured directly by selecting a single layer mode in the testing software.

Results and discussion

Effect of the CNT network on curing shrinkage of PAA

Fig. 1 compares the morphology of the neat PAA casting, carbon fiber/PAA composite, and CNT film/PAA composite. The PAA casting is extremely difficult to prepare because its shrinkage ratio can reach 11.7%. The neat PAA often cracks into pieces after curing as shown in Fig. 1a. Even in the carbon fiber reinforced PAA composite, micro-cracks can often be observed,²⁰ as shown in Fig. 1b. These crack defects are closely related to the large curing exotherm and highly cross-linked structure of the PAA matrix. PAA is mainly composed of p-diethynylbenzene, m-diethynylbenzene and solvents. After curing, highly cross-linked cis conjugated polyene and polyphenylene structures are formed, as shown in Fig. 2. Both solvent vaporization and high cross-linking density cause shrinkage of PAA.

Fig. 1 Morphology of (a) cracked PAA casting, (b) micro-cracks in the carbon fiber/PAA composite; (c) CNFP5, with the bent nanocomposite film shown in the inset, and (d) its enlarged image.

Fig. 2 Infrared spectra of PAA before and after resin curing.

On the contrary, the CNT film reinforced PAA composite shows a neat surface (as shown in Fig. 1c and d), which indicates that the compatible combination of PAA and a CNT network forms a nanocomposite without macro- or micron-size defects. Hence, we can infer that the CNT network can effectively restrict the shrinkage of the PAA matrix. Moreover, benefiting from the super ductility of carbon nanotubes, this CNT/PAA film is very flexible. This film can be easily bent, and no cracks are observed after shape recovery.

In order to reveal the effect of CNTs on crack elimination in the PAA matrix, the morphology of the CNT film and curing characteristics of the CNT film/PAA composite were investigated. The pristine CNT film is porous with an average pore diameter of approximately 45 nm (as illustrated in Fig. 3a), which effectively separates the PAA matrix into nanoscale domains. The contact angle between a 15 wt% PAA solution and the CNT film is only 6.4°, showing the excellent wettability of the PAA resin with the CNT film, even better than epoxy resin.⁴ This suggests that the PAA resin can easily infiltrate the nanoscale pores to avoid resin enrichment on the surface. Fig. 3b further illustrates the curing characteristics of neat PAA and the PAA matrix in the CNT film. Neat PAA has a sharp exothermic peak with a heat release of 513 J g⁻¹. The PAA matrix in the CNT film shows a gentle DSC profile, suggesting a distinct stereo-hindrance effect of the CNT network to retard PAA polymerization. This phenomenon has also been reported in CNT/epoxy composites.²¹ A lower crosslinking density of PAA in the CNT film tends to result in a lower shrinkage ratio. Besides, the high thermal conductivity of the CNT network also helps to transfer reaction heat to avoid heat accumulation. All of these factors contribute to the formation of a CNT film/PAA composite without macro- or micron-size defects.

Fig. 3 (a) Porous structure of the CNT film, the inset shows the pore diameter distribution; (b) curing exothermic curves of neat PAA and different CNT film/PAA composites.

Tensile properties of CNT film/PAA and its high temperature performance

Fig. 4 shows the tensile stress–strain curves of the CNT film/PAA composites, CNT film/epoxy composites and pristine CNT film. The stress–strain curves of CNT/PAA and CNT/epoxy are quite different. The PAA nanocomposite exhibits a much higher modulus but smaller strain at breakage than CNT/epoxy. CNFP5 shows two stages: an initial linear stage (I) and a following stage with slight inflection (II). In comparison, the CNT film impregnated with 5 wt% epoxy (CNFE5) went through three main stages before break: the first linear stage with a tensile strain less than 1%; a second stage with the occurrence of inflection (1% < ε < 5%); and a third stage with a slow stress increase until break.⁴ Moreover, the pristine CNT film shows similar tendencies to CNFE5. The higher modulus and wider linear strain of the PAA nanocomposite mean better network stability than that of CNT/epoxy. This is also the reason for the shape recovery of the PAA composite without plastic deformation of the CNT network.

Fig. 4 Tensile stress–strain curves of the pristine CNT film, PAA and epoxy nanocomposites.

As regards the mechanical properties, CNFP5 shows a tensile strength of 303 ± 38 MPa, Young’s modulus of 22 ± 2 GPa, and toughness of 5.5 J g⁻¹. Comparatively, CNFE5 has a tensile strength and Young’s modulus of 281 ± 36 MPa and 11 ± 4 GPa, respectively, and a toughness of 49 J g⁻¹. When the CNT film is soaked in a PAA solution with a higher concentration of 15 wt%, the tensile strength and Young's modulus of CNFP15 decrease slightly to 157 ± 26 MPa and 21 ± 4 GPa, as shown in Fig. 5. In order to improve the orientation of the CNTs, mechanical stretching is applied to the CNT film, and the degree of orientation, I_G‖/I_G⊥, is 9.1. The aligned CNT/PAA composite shows a completely linear relationship between tensile stress and strain as shown in Fig. 4. Correspondingly, the tensile strength and Young’s modulus of the aligned CNFP5 increased by 270% and 460% to 830 ± 58 MPa and 101 ± 6 GPa, respectively. Likewise, the aligned CNFP15 has a higher tensile strength and Young’s modulus than the random CNT film composite, which increase by 430% and 395% respectively.

Fig. 5 Tensile strengths and Young’s moduli of CNT film/PAA composites.

Fracture morphologies of different nanocomposites were also observed (Fig. 6). The CNT film/PAA composite shows random and crooked pull-out CNTs in the fracture with an average length of only about 14 μm (Fig. 6b). For the CNT film/epoxy composite, the pull-out CNTs become highly aligned with a longer length of about 20 μm (Fig. 6a), which presents a typical ductile failure mode. The fracture toughness (K_IC) of PAA and epoxy casting measured by an indentation microfracture method²² is 1.85 MPa m^1/2 and 4.58 MPa m^1/2, respectively. This indicates that the epoxy resin has a better ductility while PAA is brittle. Fig. 7 illustrates a strain-induced deformation during tensile testing of the PAA and epoxy composites. When the tensile force is loaded, the ductile epoxy nanocomposite greatly deforms and the slippage of CNTs occurs. The strain-induced alignment causes the significant yield before break. In comparison, the brittleness of PAA restrains the deformation of the CNT network, and alignment can be hardly induced. As a result, a brittle failure mode predominantly happens. Hence, the CNT film/PAA composite inherits the flexibility of the CNT film and good rigidity of the PAA matrix to achieve a flexible and strong composite film.

Fig. 6 Fracture morphologies of the (a) CNT film/epoxy and (b) CNT film/PAA composite.

Fig. 7 Schematic diagram of the ductile fracture mode of the epoxy nanocomposite and brittle fracture mode of the PAA nanocomposite.

It is known that PAA has a high char yield over 80%. High temperature treatment can further densify the PAA matrix. Thermogravimetric (TG) characteristics of the nanocomposites were analysed as shown in Fig. 8. The pristine CNT film decomposes slightly at 239 °C in a N₂ atmosphere due to the pyrolysis of oligomers produced by a catalyst side reaction in the synthesis process (Fig. 8a). After the introduction of PAA, the decomposition temperature at 5% mass loss of CNFP5 is 514 °C with a char yield of 90.7% (see Fig. 8b). Comparatively, the decomposition temperature of the CNT film/epoxy composite (CNFE5) is 383 °C with a char yield of only 58.1%. It is because neat PAA has good heat-resistance with a decomposition temperature over 500 °C. The all carbon structure of CNT leads to a char yield over 94.5% of the pristine CNT film, and thus the char yield of CNFP5 is higher than that of neat PAA. The high char yield ensures minimal shrinkage of the PAA nanocomposite during pyrolysis. As a result, effects of both the PAA and CNT film contribute to the excellent heat-resistant performance of the PAA nanocomposite.

Fig. 8 (a) TG curves and (b) the decomposition temperature and char yield of the CNT film, PAA resin, and CNT film nanocomposites in a N₂ atmosphere; (c) TG and DTG curves of CNT film/PAA composites in air.

In an air atmosphere (Fig. 8c), the temperatures at the maximum degradation rate of CNFP5 and CNFP15 are 605 °C and 546 °C respectively, higher than for the epoxy/CNT and BMI/CNT composites.²³ The multi-walled CNTs in the network have multiple layers to be burned through, and thus the pristine CNT film is less exposed to oxidative damage at a degradation temperature over 600 °C. Hence the CNT network provides the composite with a good thermo-oxidation stability.

In order to further reveal the potential of the PAA nanocomposite for ablative materials, its retention rate of tensile properties was measured after carbonization treatment. Thus, a heat treatment at 900 °C in N₂ for 0.5 h was adopted to make PAA fully carbonized.²⁴ In Fig. 9a, carbonized CNFP5 shows a tensile strength of 210 MPa with a retention rate of 66%. The Young’s modulus of carbonized CNFP5 decreases slightly to 19.7 GPa from 22.3 GPa. The stress–stain curves in Fig. 9a show no obvious yield before failure and elongation decreases significantly after carbonization. In Fig. 9b, the fracture of the carbonized nanocomposite is neat and smooth with minimal long pulling-out CNTs, which also suggests a typical brittle failure. The decrease of tensile strength and Young’s modulus of CNFP5 after carbonization is due to the quality loss and carbonization shrinkage of the PAA matrix and a certain degree of structural damage of the CNT network during the carbonization process at 900 °C. Likewise, the aligned CNFP5 after carbonization presents a tensile strength of 653 MPa, with a retention rate of 79%. However, the Young’s modulus of the carbonized aligned-CNFP5 increases to 156 GPa, 54% higher than that of the uncarbonized one. The tensile strength decrease of aligned CNFP5 may also be caused by quality loss of PAA and structural damage of CNTs during carbonization. As regards the increase of the Young’s modulus, this may be ascribed to the highly orientated CNTs and CNT bundles in aligned CNFP5, which may induce the formation of a certain amount of graphite microcrystals from the benzo structures of PAA molecules after carbonization.^25,26

Fig. 9 (a) Tensile stress–strain curves of the CNT film/PAA composites before and after carbonization; (b) fracture morphology of the carbonized nanocomposite.

Thermal and electrical conductivities of CNT film/PAA

Fig. 10a shows the through-thickness thermal conductivities of the CNT film/PAA composite and pristine CNT film. Though individual CNTs have excellent thermal conductivity, the through-thickness thermal conductivity of the pristine CNT film is only 0.15 W (mK)⁻¹ due to its high porosity and weak interaction between CNTs. After the introduction of PAA, CNFP15 shows a through-thickness thermal conductivity of 1.15 W (mK)⁻¹, almost seven times higher than that of the pristine CNT film. Moreover, it decreases with reduced PAA content. The thermal conductivity of CNFP5 is only 0.49 W (mK)⁻¹. After impregnation of PAA resin, the thickness of the nanocomposite film decreased by 50% to 15 μm. The result implies that PAA densifies the stacking of CNTs and fills up the pores. Additionally, the conjugated polyene structure of PAA increases the degree of crystallinity and orientation. Thus it enhanced the thermal and electrical properties of the composite,²⁷ and these effects lead to the increase of thermal conductivity of the nanocomposite.

Fig. 10 Thermal conductivity of (a) monolayer CNT film/PAA composites and (b) CNFP15 laminate; (c) surface morphology of CNFP15 after carbonization; (d) electric conductivity of CNT film and CNT film/PAA composites.

Fig. 10b presents the thermal conductivity of the eight-layered CNT film/PAA composite laminate. By comparison with the monolayer sample, the laminated composite shows a lower thermal conductivity due to the introduction of an inter-laminar interface. The through-thickness thermal conductivity of laminated CNFP15 decreases to 0.36 W (mK)⁻¹. Furthermore, an obvious anisotropic feature of thermal conductivity can be also observed. The in-plane thermal conductivity of the CNFP15 laminate is 35.36 W (mK)⁻¹. This feature indicates the significant effect of CNT orientation on thermal conductivity. After carbonization, through-thickness thermal conductivity of the laminate increases significantly to 0.97 W (mK)⁻¹, almost two times higher than that of the original CNFP15 laminate. The density of the carbonized nanocomposite remains 97% of 1.47 g cm⁻³ due to the high char yield of PAA; nevertheless it shows a slightly porous structure due to the shrinkage and densification of PAA (Fig. 10c). With an increase of pyrolytic temperature, the ratio of H/C decreases while the carbon content and degree of graphitization of PAA increase.²⁵ This indicates that the initial insulated polymer structure gradually transforms into a conductive graphite crystalline structure during carbonization. As a result, through-thickness interface thermal resistance is reduced effectively. Hence the thermal conductivity is greatly improved after carbonization.

Fig. 10d shows the electric conductivity of the CNT film and CNT film/PAA composites. The random lap joints among CNTs can form a conductive path, which provides the CNT film with a conductivity of 500 S cm⁻¹. After the introduction of PAA, electrical conductivity increases to 700 S cm⁻¹ for CNFP5 and CNFP15. Additionally, carbonization of PAA allows the conductivity of the composite to reach up to 820 S cm⁻¹.

Conclusions

This paper develops a flexible and heat-resistant nanocomposite through the combination of a high char yield PAA resin and ductile CNT film. The CNT network in the nanocomposite results in steric hindrance to retard the polymerization of the PAA matrix, which effectively reduces the shrinkage and crack defects, and achieves flexibility of the PAA composite. The CNT film/PAA (CNFP5) shows a tensile strength and modulus of 303 ± 38 MPa and 22 ± 2 GPa, respectively. Stretching induced alignment of CNTs can further increase the tensile strength and modulus of the nanocomposite up to 830 ± 58 MPa and 101 ± 6 GPa. CNT slippage and network deformation can be hardly induced by tensile stress in the CNT film/PAA composite, indicating a lock-up effect of PAA on the CNT network to ensure structure stability. Additionally, after carbonization at 900 °C for 0.5 h, the nanocomposite has a retention rate of tensile strength of over 66%. Thermogravimetric analysis shows that the PAA nanocomposite has a decomposition temperature over 500 °C, which is much higher than that of the epoxy or BMI matrix nanocomposites. The char yield after carbonization reaches 90.7%. These features contribute to the excellent heat-resistant performance of the CNT film/PAA composite. The through-thickness thermal conductivity reaches 1.15 W (mK)⁻¹, almost seven times higher than that of pristine CNT film (i.e. 0.15 W (mK)⁻¹). Densification of CNT stacking and the conjugated polyene structure of PAA contribute to the increase of thermal conductivity. Furthermore, after carbonization a certain amount of graphite microcrystals may form from the benzo structures of PAA molecules increasing the thermal conductivity of the nanocomposite, which is two times higher than that of the uncarbonized one. The electrical conductivity of the CNT film/PAA composite increases to 700 S cm⁻¹, which is 40% higher than that of the pristine CNT film. These performance advantages give the CNT film/PAA composite great potential for multifunctional applications in high-heat environments.

Acknowledgements

The authors thank the National Natural Science Foundation of China (Grant No. 51103003, 51273007 and 51403009).

Notes and references

1. M. F. L. de Volder, S. H. Tawfick, R. H. Baughman and A. J. Hart, Science, 2013, 339, 535–539.
2. J. Di, X. Wang, Y. Xing, Y. Zhang, X. Zhang, W. Lu, Q. Li and Y. T. Zhu, Small, 2014, 10, 4606–4625.
3. A. G. Nasibulin, A. Kaskela, K. Mustonen, A. S. Anisimov, V. Ruiz, S. Kivistö, S. Rackauskas, M. Y. Timmermans, M. Pudas, B. Aitchison, M. Kauppinen, D. P. Brown, O. G. Okhotnikov and E. I. Kauppinen, ACS Nano, 2011, 5, 3214–3221.
4. M. Li, Z. Wang, Q. Liu, S. Wang, Y. Gu, Y. Li and Z. Zhang, Polym. Compos., 2015 DOI:10.1002/pc.23617.
5. X. Wang, Z. Z. Yong, Q. W. Li, P. D. Bradford, W. Liu, D. S. Tucker, W. Cai, H. Wang, F. G. Yuan and Y. T. Zhu, Res. Lett. Mater. Sci., 2013, 1, 19–25.
6. Z. An, H. Wei, Y. Li and Y. Gao, Eng. Plast. Appl., 2006, 34, 19–22.
7. N.-H. Tai, M.-K. Yeh and T.-H. Peng, Composites, Part B, 2008, 39, 926–932.
8. N.-H. Tai, M.-K. Yeh and J.-H. Liu, Carbon, 2004, 42, 2774–2777.
9. M. H. Choi, I. J. Chung and J. D. Lee, Chem. Mater., 2000, 12, 2977–2983.
10. Y. J. Liu, H. P. Teng and Y. L. Guo, Thermosetting Resin, 2008, 23, 5–8.
11. H. Qi, G. Pan, L. Yin, Y. Zhuang, F. Huang and L. Du, J. Appl. Polym. Sci., 2009, 114, 3026–3033.
12. H. Guo, Y. D. Huang and L. Liu, Chem. Adhes., 2008, 30, 37–41.
13. M.-C. Wang and T. Zhao, J. Appl. Polym. Sci., 2007, 105, 2939–2946.
14. S. S. Oishi, E. C. Botelho, C. K. Luscombe and M. C. Rezende, Polimeros, 2014, 24, 541–546.
15. W.-C. Tseng, Y. Chen and G.-W. Chang, Polym. Degrad. Stab., 2009, 94, 2149–2156.
16. M. Liu, L. Gan, F. Zhao, X. Fan, H. Xu, F. Wu, Z. Xu, Z. Hao and L. Chen, Carbon, 2007, 45, 3055–3057.
17. X. Zhang, Y. Huang, T. Wang and L. Liu, Composites, Part A, 2007, 38, 936–944.
18. Q. Liu, M. Li, Y. Gu, Y. Zhang, S. Wang, Q. Li and Z. Zhang, Nanoscale, 2014, 6, 4338–4344.
19. Q. Liu, M. Li, Y. Gu, S. Wang, Y. Zhang, Q. Li, L. Gao and Z. Zhang, Carbon, 2015, 86, 46–53.
20. C. Miao, L. I. Min, S. Wang, J. Yang and G. U. Yizhuo, Acta Mater. Compositae Sin., 2010, 27, 75–80.
21. A. Moisala, Q. Li, I. A. Kinloch and A. H. Windle, Compos. Sci. Technol., 2006, 66, 1285–1288.
22. G. R. Anstis, P. Chantikul, B. R. Lawn and D. B. Marshall, J. Am. Ceram. Soc., 1981, 64, 533–538.
23. Q. Wu, J. Bao, C. Zhang, R. Liang and B. Wang, J. Therm. Anal. Calorim., 2011, 103, 237–242.
24. X. Ding, Fiber Compos., 2005, 22, 8–11.
25. X. W. Ding, D. Chen, J. H. Cao, Y. Q. Zhuang, J. Wang and Y. S. Jiao, Polym. Mater.: Sci. Eng., 2002, 18, 127–130.
26. R. J. Zaldivar, R. W. Kobayashi, G. S. Rellick and J. M. Yang, Carbon, 1991, 29, 1145–1153.
27. D. Zhang and Y. Zhang, J.Solid Rocket Technol., 2001, 24, 53–59.

---