Percolation and resistivity-temperature behaviours of carbon nanotube-carbon black hybrid loaded ultrahigh molecular weight polyethylene composites with segregated structures

Abstract

An ultrahigh molecular weight polyethylene (UHMWPE) composite containing carbon nanotube–carbon black (CNT–CB) hybrid was fabricated via a facile method, i.e., mechanical mixing plus hot compaction in order to obtain low-cost conductive polymer composites with balanced electrical properties. Optical microscope and scanning electron microscope observations indicate the formation of a typical segregated structure in the CNT–CB/UHMWPE composite, with the CNT–CB hybrid selectively located at the interfaces of UHMWPE granules. Compared to the single CNT loaded UHMWPE composite, the CNT–CB/UHMWPE segregated composite with a quarter replacement of CNT with CB shows only 8% decline in electrical conductivity, with the same filler content of 4 wt%, realizing a significant reduction in the material cost. More interestingly, the CNT–CB/UHMWPE composite presents 273% higher positive temperature resistivity intensity than that of CNT/UHMWPE composites, exhibiting strong sensitivity to ambient temperature. Our work demonstrates a novel strategy to fabricate low-cost and high-performance conductive polymer composites by the combination of hybrid fillers and a segregated structure.

---

1. Introduction

The formation of a segregated structure is considered to be a highly efficient way to reduce the electrical percolation threshold in conductive polymer composites (CPCs),^1–4 due to the selective distribution of conductive fillers at the interfaces of polymer granules instead of homogeneous distribution among the whole system, which has been thoroughly reviewed in our latest article.⁵ Carbon nanotubes (CNT), with a large aspect ratio and high inherent conductivity, are predominately selected as a promising nanofiller for constructing a segregated structure in polymer matrices.^6–14 By locking CNT in the voids of polymer granules, Jurewicz et al. prepared a segregated composite with a percolation threshold four times lower than that of composite with randomly distributed CNT.¹⁵ Liu et al. reported a CNT/poly (vinylidene fluoride) (PVDF) segregated composite with a percolation threshold as low as 0.078 vol%.¹⁰ Nevertheless, it is worth noting that despite low percolation threshold for segregated CNT/polymer composites, the realization of excellent electrical conductivity still requires high CNT contents that would undoubtedly cause high cost of the final composites and inferior mechanical properties, thus greatly limiting their integration in potential industrial applications. Therefore, the achievement of high electrical conductivity for CPCs with low cost remains a daunting challenge.

Alternatively, carbon black (CB) as an economical material is the most widely used conductive additives in polymer matrices.^16–19 Whereas, the single CB loaded CPCs always exhibit weaker electrical conductivity than that of CNT ones upon the same loading because of the relatively low aspect ratio and inherent conductivity. For instance, the electrical conductivity of CB/acrylonitrile–butadiene–styrene(ABS) composite was reported to be 4000 times lower than that of CNT/ABS composite, at 5 wt% CB or CNT loading.²⁰ Interestingly, several studies have focused on the partial replacement of CNT with CB and found that the resultant CNT–CB hybrid loaded CPCs presented comparable or better electrical conductivities to that of single CNT loaded ones, which suggested a potential method to reduce the cost of CPCs without sacrificing the electrical conductivity.^21–25 Zhang et al. replaced 20% CNT with CB to form conductive networks of CNT–CB hybrid in polypropylene matrix at 10 wt% filler content, revealing comparable electrical conductivity (10 S m⁻¹) in comparison with CNT ones.²² It should be noted that the available CPCs with CNT–CB hybrid were mainly based on random conductive networks. Then, one can speculate that the combination of segregated structure and CNT–CB hybrid might realize remarkable improvement in electrical conductivity of CPCs with considerably reduced CNT loading and thus the CPCs cost.

Stimulated by the interesting but rarely reported issue, this work conducted a comprehensive research on CNT–CB/ultrahigh molecular weight polyethylene (UHMWPE) segregated composites, where UHMWPE was selected as the polymer matrix for its high melt viscosity to prevent the migration of the CNT–CB hybrid. The effect of CNT–CB ratios on the electrical and resistivity–temperature behaviours of the segregated composites were systematically investigated. By replacing a quarter of CNT with CB, the CNT–CB/UHMWPE composite exhibited an electrical conductivity of 3.33 S m⁻¹, only 8% decrease compared to that of CNT/UHMWPE composite at the same filler loading of 4 wt%. Interestingly, the replacement of CNT with CB also resulted in 273% enhancement in the positive temperature coefficient intensity compared to that of CNT/UHMWPE composite.

2. Experimental

2.1. Materials

UHMWPE powders, provided by Beijing no. 2 Auxiliary Agent Factory, Beijing, China, have the flowing properties: density of ∼0.94 g cm⁻³, resistivity of ∼10¹⁷ Ω cm, and an average diameter of ∼150 μm. CNT, with the diameter of 20–40 nm and length of 10–20 μm, was supplied by Chengdu Organic Chemicals Co. Ltd., Chengdu, China. CB (VXC-605) with the average particle size of 25 nm was kindly provided by Cabot Co. Ltd, Shanghai, China.

2.2. Fabrication of segregated CPC materials

The fabrication process of the CNT–CB/UHMWPE segregated composite is illustrated in Fig. 1. First, the quantified CNT, CB and UHMWPE powders were mechanically mixed for 5 min to obtain CNT–CB hybrid coated UHMWPE complex granules, at a mixing speed of 24 000 rpm. Then, the CNT–CB/UHMWPE complex granules were compression molded into 10 × 10 × 1.5 mm³ sheets at 200 °C for 5 min, with the pressure of 10 MPa after preheating for 3 min. For convenience, the segregated composites containing CNT and CB in the ratio of 3/1, 1/1, and 1/3 were labeled as CNT₃–CB₁/UHMWPE, CNT₁–CB₁/UHMWPE, and CNT₁–CB₃/UHMWPE, respectively. The segregated composites containing various CNT–CB hybrid contents from 0.1 to 4.0 wt% were prepared, regardless of CNT–CB ratios. For comparison, the single CNT loaded UHMWPE segregated composites were also fabricated under the same conditions.

Fig. 1 Schematic for the fabrication of the CNT–CB/UHMWPE segregated composites.

2.3. Characterization

Morphology observations were carried out on a field emission scanning electron microscopy (SEM, Inspect-F, FEI, Finland) and an optical microscopy (OM, Olympus BX51). The composite specimens for SEM were immersed in liquid nitrogen for 30 min, then quickly impact fractured. The freshly fractured surfaces were coated with gold and performed at an accelerating voltage of 20 kV. The OM specimens were cut into films using a microtome, with the thickness of 15 μm. The in situ OM observation was conducted at a heating rate 2 °C min⁻¹. Electrical resistivity of segregated composites were measured using two points method by a Keithley electrometer model 4200 SCS (USA). Both ends of composite samples were coated with highly conductive silver paste to reduce the contact resistance between the samples and the electrodes. For the resistivity–temperature behaviours, the samples immersed in silicone were heated from 40 to 180 °C at 2 °C min⁻¹ and kept for 3 min at this temperature. Keithley 4200 SCS was used to monitor the electrical resistivity. Thermal analysis was carried out by a TAQ200 differential scanning calorimeter (DSC). The sample was heated from 40 to 180 °C at 2 °C min⁻¹, then kept for 3 min at this temperature to match the resistivity–temperature test.

3. Results and discussion

3.1. Morphology of the segregated composites

OM images in Fig. 2 indicate the formation of typical segregated structure in the CNT–CB/UHMWPE composites, with CNT–CB hybrid mainly located at the interfacial regions between the UHMWPE granules instead of penetrating into their interior, owing to the high viscosity of UHMWPE matrix and low shear in the hot compaction process. The adjacent UHMWPE granules form very thin interfaces that hold almost the whole conductive CNT–CB, which is beneficial to develop conductive networks at an extremely low filler loading. When the CNT–CB hybrid content is only 0.1 wt% (Fig. 2a), parts of conducting pathways between UHMWPE particles are obscure, revealing that such a low CNT–CB hybrid loading is insufficient to construct interconnected conductive networks throughout the insulating polymer matrix. As the amount of CNT–CB hybrid increases to 0.5 and 2.0 wt% (Fig. 2b and c), the conductive networks become much thicker, indicative of well-defined CNT–CB conductive networks. SEM images in Fig. 3 further present the typical faceted construction of the CNT–CB/UHMWPE composites, with an ordered segregated network of CNT–CB wrapped UHMWPE polyhedrons. The evolution of such specific microstructures can be explained from the fabrication process. During the mechanical mixing, the CNT–CB hybrid gradually decorated UHMWPE granules, and then the UHMWPE granules experienced the plastic deformation to form polyhedrons by hot-pressing. Additionally, it can be seen that there are no significant difference in the CNT–CB/UHMWPE composites with various CNT–CB ratios.

Fig. 2 OM images of CNT₃–CB₁/UHMWPE segregated composites with various CNT–CB hybrid contents of 0.1 wt% (a), 0.5 wt% (b), 2.0 wt% (c), respectively.

Fig. 3 SEM images of the CNT₃–CB₁/UHMWPE (a), CNT₁–CB₁/UHMWPE (b), and CNT₁–CB₃/UHMWPE (c) segregated composites, with CNT–CB hybrid content of 0.5 wt%.

3.2. Electrical properties of composites

Fig. 4 shows the dependence of CNT–CB hybrid content on electrical conductivity of the CNT–CB/UHMWPE segregated composite. The electrical conductivity of CNT/UHMWPE composite with varying CNT content is also introduced for comparison. Regardless of using CNT or CNT–CB hybrid, the conductive filler endows the composite a considerable increase in electrical conductivity up to 12 orders of magnitude in a very narrow filler content range between 0.1 wt% and 0.3 wt%, indicating a typical percolation behaviour, i.e., the composite transforming from an insulator into a semi-conductor through the construction of preliminary conductive pathways. Nevertheless, it can be seen that although the partial replacement of CNT with CB has strong impact on the electrical conductivity of the composites at low filler content, the disparity in electrical conductivity gradually reduces along with the incremental filler content. Particularly, the CNT₃–CB₁/UHMWPE composite with 4 wt% CNT–CB hybrid exhibits an electrical conductivity of 3.33 S m⁻¹, only 8% decrease compared to that of CNT filled one (3.62 S m⁻¹), meaning that CB could almost completely replace a portion of CNT gaining identical electrical conductivity.

Fig. 4 Electrical conductivity as a function of conductive filler content for the CNT/UHMWPE and CNT–CB/UHMWPE segregated composites.

To further elaborate the effect of CB, the electrical conductivities of CNT–CB/UHMWPE composites as a function of CNT content were plotted in Fig. 5. The CNT–CB/UHMWPE composites exhibit more obvious enhancement in electrical conductivity than the CNT/UHMWPE one along with increasing CB amount at a fixed CNT content. At 1 wt% CNT loading, the addition of a small amount of CB affects the electrical conductivity of the composite slightly, because CB particles just play the role of decorating CNT conductive network. As the CB content rises from 0.33 to 1 wt%, the electrical conductivity of segregated composite increases from 0.1 to 0.36 S m⁻¹. At such CB loading, CB particles not only construct independent conductive pathways but also supplement CNT conductive networks. More significant enhancement in electrical conductivity can be observed through incorporating more CB particles into CNT/UHMWPE system. Thus, the CNT₃–CB₁/UHMWPE composite with 4 wt% CNT–CB hybrid content is endowed with high electrical conductivity thanks to CB contained in the composite enough to supplement CNT conductive networks, indicating that CNT can be partly replaced by CB without sacrificing the electrical conductivity.

Fig. 5 Electrical conductivities of composites as a function of CNT content.

To determine the percolation behaviors of the segregated composites, we employed the classical percolation theory to investigate the influence of hybrid fillers on conductive networks. Percolation theory formula is σ = σ₀(j − j_c)^t, where σ represents the electrical conductivity of composite, σ₀ is a constant related to the intrinsic conductivity of conductive filler, j is the volume fraction of the filler, j_c is the volume percolation threshold, and t is the critical exponent related to the dimensionality of the system. The resulting fitting j_c is listed in Table 1. As expected, the formation of segregated structure results in an extremely low percolation threshold of 0.07 vol% in the CNT/UHMWPE system. Additionally, CNT₃–CB₁/UHMWPE also maintains a very low percolation threshold of 0.09 vol%, though low-cost CB was utilized to replace 25% CNT.

Table 1 Electrical conductivities at 4 wt% filler content and the percolation thresholds of composites

Filler	Maximum conductivity (S m^−1)	Percolation threshold (jc)
CNT	3.62	0.149 wt% (0.07 vol%)
CNT3–CB1	3.33	0.187 wt% (0.09 vol%)
CNT1–CB1	2.63	0.293 wt% (0.145 vol%)
CNT1–CB3	2.02	0.334 wt% (0.170 vol%)

---

Meanwhile, our CNT–CB/UHMWPE segregated composites own outstanding percolation threshold and electrical conductivity compared to the reported CNT/polymer composites with randomly distributed, as listed in Table 2. The selective distribution of CNT–CB hybrid at the interfaces among UHMWPE granules increases the effective CNT–CB concentration for constructing conducting pathways, leading to the excellent electrical performances. More attractively, the resulting composites achieves comparable level electrical conductivity to other CNT/polymer segregated composites even through partial CNT replaced by low-cost CB, and the present work demonstrates that the utilization of hybrid filler in segregated structure is an effective way to decrease the cost segregated CPCs without sacrificing their electrical properties.

Table 2 Percolation threshold and electrical conductivity for CNT–CB/UHMWE segregated composites from this work, and the previously reported CNT/polymer composites with random distributed or segregated structure

Matrix	Filler	Distribution	jc	σmax (S m^−1)	Reference
a HDPE: high density polyethylene; PP: polypropylene; PANI: polyaniline; PC: polycarbonate; NR: natural rubber; phr: parts per hundred rubber.
UHMWPE	CNT	Segregated	0.07 vol%	3.62@4 wt%	This work
UHMWPE	CNT3–CB1	Segregated	0.09 vol%	3.33@4 wt%	This work
UHMWPE	CNT	Segregated	0.072 vol%	∼2@∼0.9 vol%	6
UHMWPE	CNT	Segregated	0.049 vol%	∼1 × 10^−1@1.5 vol%	7
HDPE^a	CNT	Segregated	0.32 wt%	—	8
PVDF	CNT	Segregated	0.078 vol%	∼1@0.3 vol%	10
PP^a	CNT	Random	2.4 wt%	—	22
Epoxy	CNT	Random	0.84 wt%	4.3 × 10^−3@4 wt%	26
Epoxy	CNT	Random	0.8 wt%	1.2 × 10^−1@5 wt%	27
PANI^a	CNT	Random	—	6@4 wt%	28
PVDF^a	CNT	Random	1.2 wt%	∼1 × 10^−5@2 wt%	29
PC^a	CNT	Random	—	∼7@8 vol%	30
NR^a	CNT	Random	—	1.9 × 10^−5@4 phr^a	31

---

3.3. Resistivity–temperature behaviours

Resistivity–temperature behaviours of segregated CPC materials were studied for the importance in self-regulating, overtemperature and overcurrent protection,^32,33 with an emphasis on their morphological evolution during the resistivity–temperature test. Fig. 6 presents the resistivity–temperature curve of CNT₁–CB₃/UHMWPE segregated composite during the heating progress. Though CNT–CB hybrid loading is fixed at 1 wt% far beyond the percolation threshold of the composite, the conductivity networks of CPC are still highly sensitive to the ambient temperature. The resistivity jumps from 2.24 × 10² Ω m at 40 °C to 6.93 × 10³ Ω m near the melting point of UHMWPE (corresponding to the DSC curve), which is regarded as the positive temperature coefficient (PTC) phenomenon.^2,34 Subsequently, the resistivity gradual decreases from 6.93 × 10³ to 2.42 × 10³ Ω m at 180 °C, referred to the negative temperature coefficient (NTC) phenomenon.³⁵

Fig. 6 Typical resistivity–temperature curve of the CNT₁–CB₃/UHMWPE segregated composite with 1 wt% CNT–CB hybrid. The insert is the corresponding DSC curve.

To explain the origin of resistivity–temperature behaviors of the CNT–CB/UHMWPE composite in Fig. 6, in situ OM was employed to explore the evolution of conductive networks during heating process, as shown in Fig. 7. At the beginning temperature of 40 °C, the located CNT–CB hybrid at the interfaces between UHMWPE granules forms well developed conductive network (Fig. 7a). With the increase of temperature (Fig. 7b), the volume expansion and local melt flow of UHMWPE matrix severely separate the contacted CNT–CB hybrid particles. Thus, conductive networks of segregated composite are partly damaged, giving rise to a sharp rise of resistivity, i.e., PTC effect. Further increasing temperature would cause the NTC effect, due to the reagglomeration of conductive fillers with the elevated ambient temperature (Fig. 7c).

Fig. 7 The in situ OM images of CNT–CB/UHMWPE composite with 2 wt% filler content at heating rate of 2 °C min⁻¹.

To quantitatively evaluate the PTC and NTC effects, the PTC intensity (I_PTC) and NTC intensity (I_NTC) are defined as

I_PTC = log(R_t/R₀)

I_NTC = log(R_max/R_k)

where R_t is the maximum resistivity in the heating journey, R₀ is the primary resistivity at 40 °C, R_max is the maximum resistivity in the whole heating process, and R_k is the electrical resistivity at 180 °C.

As shown in Fig. 8a, all the CNT–CB/UHMWPE composites show significantly higher I_PTC than that of CNT/UHMWPE composites, especially at low filler loadings. Due to the large aspect ratio and high inherent conductivity of CNT, the conductive networks in CNT/UHMWPE composites are very stable, which can be only slightly damaged by the expanded and melted UHMWPE matrix, resulting in little increase in the interparticle distance between adjacent CNT and thus low I_PTC.² For instance, the I_PTC value of 0.8 wt% CNT loaded CNT/UHMWPE composite is only 0.50. Nevertheless, after a certain amount of CNT is replaced by spheroidal CB particles, the I_PTC of the CNT–CB/UHMWPE composites increases by 112%, 102% and 182% with the increasing CB content, revealing higher sensitivity to temperature field. As the filler content increases, the I_PTC gradually descends because of constructing perfect conductive networks. Additionally, the CNT₁–CB₃/UHMWPE composite with 1 wt% CNT–CB hybrid achieves the highest I_PTC of 1.49, which is 273% higher than that of CNT/UHMWPE system.

Fig. 8 (a) I_PTC and (b) I_NTC of segregated composites with various filler contents.

To highlight the intensity of PTC output signals, the NTC effect is frequently expected to be attenuated and eliminated³⁶ Fig. 8b plots the I_NTC as a function of conducting filler content and I_NTC of all the composites express a downward trend with the increasing filler content. The CNT/UHMWPE composites possess very low I_NTC at same filler content, which is corresponding to other experimental results.³⁷ Likewise, CNT₁–CB₃/UHMWPE composite possesses I_NTC of only 0.45 at 1 wt% filler loading, approaching that of CNT filler filled composite. In this segregated CPC materials, conducting CNT pathways support the dominant electron transfer, achieving high electrical conductivity to guarantee a stable output signal as the application of temperature sensors. For the part of the conducting CB channels, the penetration of CB particles into UHMWPE melt renders the high resistivity–temperature effects. The high I_PTC and low I_NTC of CNT–CB/UHMWPE segregated CPCs demonstrate a novel methodology to fabricate the high-performance temperature sensors. Moreover, it is also an attractive way to obtain comparable high electrical conductivity to CNT one at low cost through using CNT–CB hybrid and segregated structure.

4. Conclusions

By introducing low-price CB particles to replace the expensive CNT, we utilized a facile dry-mixing technology to fabricate a novel low-cost CNT–CB/UHMWPE segregated composites with ultralow percolation threshold 0.09 vol% and high electrical conductivity 3.33 S m⁻¹. The morphological observation indicates that the high-performance electrical properties are ascribed to the formation of segregated CNT–CB conductive networks. The incorporation of CB into CNT composites exhibits obviously promoting PTC effect and weak NTC effect.

Acknowledgements

The authors gratefully acknowledge the financial support from the National Natural Science Foundation of China (Grant nos 51421061, 51273131, 51473102), the Innovation Team Program of Science and Technology Department of Sichuan Province (Grant no. 2014TD0002).

References

1. H. Pang, Y. Y. Piao, C. H. Cui, Y. Bao, J. Lei, G. P. Yuan and C. L. Zhang, J. Polym. Res., 2013, 20, 304.
2. H. Pang, Q. Y. Chen, Y. Bao, D. X. Yan, Y. C. Zhang, J. B. Chen and Z. M. Li, Plast., Rubber Compos., 2013, 42, 59–65.
3. H. Pang, T. Chen, G. Zhang, B. Zeng and Z. M. Li, Mater. Lett., 2010, 64, 2226–2229.
4. P. G. Ren, Y. Y. Di, Q. Zhang, L. Li, H. Pang and Z. M. Li, Macromol. Mater. Eng., 2012, 297, 437–443.
5. H. Pang, L. Xu, D. X. Yan and Z. M. Li, Prog. Polym. Sci., 2014, 39, 1908–1933.
6. J. F. Gao, Z. M. Li, Q. j. Meng and Q. Yang, Mater. Lett., 2008, 62, 3530–3532.
7. H. Pang, C. Chen, Y. Bao, J. Chen, X. Ji, J. Lei and Z. M. Li, Mater. Lett., 2012, 79, 96–99.
8. J. Du, L. Zhao, Y. Zeng, L. Zhang, F. Li, P. Liu and C. Liu, Carbon, 2011, 49, 1094–1100.
9. Y. Mamunya, A. Boudenne, N. Lebovka, L. Ibos, Y. Candau and M. Lisunova, Compos. Sci. Technol., 2008, 68, 1981–1988.
10. Q. Liu, J. Tu, X. Wang, W. Yu, W. Zheng and Z. Zhao, Carbon, 2012, 50, 339–341.
11. M. O. Lisunova, Y. P. Mamunya, N. I. Lebovka and A. V. Melezhyk, Eur. Polym. J., 2007, 43, 949–958.
12. X. Hao, G. Gai, Y. Yang, Y. Zhang and C. W. Nan, Mater. Chem. Phys., 2008, 109, 15–19.
13. M. H. Al-Saleh, S. A. Jawad and H. M. El Ghanem, High Perform. Polym., 2013, 26, 205–211.
14. S. Zhang, H. Deng, Q. Zhang and Q. Fu, ACS Appl. Mater. Interfaces, 2014, 6, 6835–6844.
15. I. Jurewicz, P. Worajittiphon, A. A. King, P. J. Sellin, J. L. Keddie and A. B. Dalton, J. Phys. Chem. B, 2011, 115, 6395–6400.
16. F. Gubbels, S. Blacher, E. Vanlathem, R. Jerome, R. Deltour, F. Brouers and P. Teyssie, Macromolecules, 1995, 28, 1559–1566.
17. K. Dai, X. B. Xu and Z. M. Li, Polymer, 2007, 48, 849–859.
18. Z. Xu, C. Zhao, A. Gu and Z. Fang, J. Appl. Polym. Sci., 2007, 103, 1042–1047.
19. C. H. Chen, H. C. Li, C. C. Teng and C. H. Yang, J. Appl. Polym. Sci., 2006, 99, 2167–2173.
20. M. H. Al-Saleh and W. H. Saadeh, Mater. Des., 2013, 52, 1071–1076.
21. R. Socher, B. Krause, S. Hermasch, R. Wursche and P. Pötschke, Compos. Sci. Technol., 2011, 71, 1053–1059.
22. S. M. Zhang, eXPRESS Polym. Lett., 2011, 6, 159–168.
23. J. Sumfleth, X. C. Adroher and K. Schulte, J. Mater. Sci., 2009, 44, 3241–3247.
24. P. C. Ma, M. Y. Liu, H. Zhang, S. Q. Wang, R. Wang, K. Wang, Y. K. Wong, B. Z. Tang, S. H. Hong, K. W. Paik and J. K. Kim, ACS Appl. Mater. Interfaces, 2009, 1, 1090–1096.
25. M. Wen, X. Sun, L. Su, J. Shen, J. Li and S. Guo, Polymer, 2012, 53, 1602–1610.
26. L. Yue, G. Pircheraghi, S. A. Monemian and I. Manas-Zloczower, Carbon, 2014, 78, 268–278.
27. S. Gong, Z. H. Zhu, J. Li and S. A. Meguid, J. Appl. Phys., 2014, 116, 194306.
28. L. C. Mariano, R. V. Salvatierra, C. E. Cava, M. Koehler, A. J. G. Zarbin and L. S. Roman, J. Phys. Chem. C, 2014, 118, 24811–24818.
29. S. H. Wang, Y. Wan, B. Sun, L. Z. Liu and W. Xu, Nanoscale Res. Lett., 2014, 9, 522.
30. X. Zhi, H. B. Zhang, Y. F. Liao, Q. H. Hu, C. X. Gui and Z. Z. Yu, Carbon, 2015, 82, 195–204.
31. P. Thaptong, C. Sirisinha, U. Thepsuwan and P. Sae-Oui, Polym.-Plast. Technol. Eng., 2014, 53, 818–823.
32. H. Pang, Y. C. Zhang, T. Chen, B. Q. Zeng and Z. M. Li, Appl. Phys. Lett., 2010, 96, 251907.
33. H. Tang, X. G. Chen and Y. X. Luo, Eur. Polym. J., 1997, 33, 1383–1386.
34. M. Hindermann-Bischoff and F. Ehrburger-Dolle, Carbon, 2001, 39, 375–382.
35. H. F. Xie, P. Y. Deng, L. S. Dong and J. Z. Sun, J. Appl. Polym. Sci., 2002, 85, 2742–2749.
36. S. L. Jiang, Y. Yu, J. J. Xie, L. P. Wang, Y. K. Zeng, M. Fu and T. Li, J. Appl. Polym. Sci., 2010, 116, 838–842.
37. X. J. He, J. H. Du, Z. Ying, H. M. Cheng and X. J. He, Appl. Phys. Lett., 2005, 86, 062112.

---