Effects of oxygen content in the atmosphere on thermal oxidative stabilization of polyacrylonitrile fibers

Abstract

The effects of oxygen content in the atmosphere on thermal oxidative stabilization of polyacrylonitrile fibers have been studied based on the evolution of the interaction between cyclization and oxidation reactions. The results indicate that cyclization propagation was inhibited in the initial few minutes due to radical adsorption by the oxygen in the atmosphere, and then the cyclization was promoted with dehydrogenation going by oxidation, whereas the oxidation reaction was always promoted with the cyclization structure. Resultantly, a high oxygen content atmosphere didn't result in a low cyclization degree and excessive oxidation in stabilized fibers, while a low oxygen content atmosphere introduced a skin–core structure because of insufficient oxygen diffusion and rapid oxidation rate in the later stage. Finally, according to the tensile strength of carbon fibers, the optimum oxygen content in the thermal stabilization atmosphere was determined to be 15–21%.

---

1. Introduction

Polyacrylonitrile (PAN) based carbon fibers play a significant role in the aerospace area, automobile industry and so on. Therefore, great interest has been aroused for PAN fiber fabrication and its application. Recently, some nanoparticle reinforced PAN based fibers were systematically investigated to improve its magnetoresistance behavior,^1–4 and electrospinning was used for the fabrication of carbon composite fibers.^5–9 Even so the mechanisms for traditional carbon fibers are far more than clear, including thermal stabilization. Thermal oxidative stabilization is a key step for the formation of polyacrylonitrile (PAN) based carbon fibers, which makes the precursor thermostable by formation of a ladder structure and oxygen-containing groups.^10–12 During this process, various physical and chemical transformations occur, including shrinkage, relevant thermal stress, cyclization, oxidation, dehydrogenation and crosslink between molecules of PAN,^13–15 and the structure after this process largely governs the final structure and mechanical properties of the carbon fiber.¹⁶ More importantly, the oxygen-containing groups would be eliminated as water vapor during the carbonization process and make the molecules crosslink with improved mechanical properties than those only containing ladder structure produced in inert environment.^17,18 However, excessive oxygen-containing groups would be released as CO or CO₂ and leave defects in the fiber during subsequent carbonization and graphitization,¹⁹ thus reducing the fiber strength. The oxygen mal-distribution on cross-section is a main factor forming the skin–core structure in carbon fibers.²⁰ It is believed that perfect stabilized fibers tend to have the similar oxygen content along the radial direction, whereas incompletely stabilized fibers show a decreasing oxygen profile towards the center and further form skin–core structure.²¹ Therefore, to produce high-performance carbon fibers, it is significant for stabilized fibers to have maximum cyclization degree, proper oxygen content and distribution as well as a suitable crosslink structure generated by oxidation.

In order to obtain a proper oxygen content and distribution, reducing diffusion-control of the process should not be ignored. One way is to reduce the reaction rate by lowering the temperature.^22,23 The other way is to increase the oxygen content in the atmosphere. Lü²⁰ and his coworkers observed that a higher oxygen partial pressure was sufficient to prohibit skin–core structure. However, for PAN fiber, compared to pitch fiber, what makes the situation more complicated is the interaction of oxidation and cyclization. Therefore, investigations of the mechanism of stabilization reaction are significant, especially the interaction between cyclization and oxidation reactions. Since stabilization process is exothermic and can be monitored by differential scanning calorimetry (DSC), the interaction between cyclization and oxidation reactions has been studied by this method.¹¹ Fitzer et al.²⁴ observed a lower initiation temperature of stabilization reaction in air compared with in nitrogen through DSC, suggesting that oxygen promoted the initiation of the cyclization. However the consecutive cyclization was inhibited by oxygen because of the increase of the activation energy. In addition, the rings formed by cyclization reaction could promote the oxidation reaction.²⁵ The other investigation focused on thermal dynamics because the shrinkage stress is also a reflection of cyclization and oxidation crosslinking. It has been evidenced that changing oxygen contents effectively affects the shrinkage stress of the fibers.¹¹ In Ogawa's publication,¹⁰ the increase of oxygen content in the atmosphere made the fiber saturated stress lower. It was interpreted that the cyclization reaction was inhibited by the oxygen in spite that oxygen promoted the initiation of the cyclization.²⁴ Therefore the suspicion is aroused that whether the final cyclization degree will be reduced by more concentrated oxygen. If so, the carbon fiber would contain more defects due to a lower carbon yield. Moreover, how the oxygen prohibits the cyclization and the evolution of the interaction between cyclization and oxidation reactions at different stages of stabilization process and under atmosphere with oxygen content lower than air is not clear yet.

In the present work, by altering oxygen contents in the atmosphere, the effects of oxidation on cyclization reaction of PAN fibers during the stabilization process were investigated. Moreover, through oxidating the fibers with different cyclization degrees and analyzing their oxygen uptake, the effects of cyclization on oxidation reaction were investigated. Finally, we attempted to find a proper range of oxygen content for stabilization atmosphere which can be used to get a stabilization fiber with proper crosslink structure, low skin–core structure and high carbon yield.

2. Experimental

2.1 Materials

The PAN co-polymer fibers (3000 filaments per tow, viscosity average molecular weight ∼ 2.537 × 10⁶ g mol⁻¹ which was determined using an ubbelohde viscometer with diameter of 0.63 mm at 30 °C) was supplied by Jilin Chemical Fiber Group Co., Ltd. The gases, with a purity of 99.999%, used in the experiment were supplied by a commercial company. The atmospheres with different oxygen contents are prepared by mixture of nitrogen and oxygen and the precision of oxygen content of the mixed gas was 99.5%.

2.2 Thermal reactions

For the effect of oxidation reaction on the cyclization reaction, the experiments were carried out as follows. The PAN fibers were subjected to the stabilization at 230 °C in the atmospheres with different oxygen contents, with a constant load of 500 cN, equivalent to a stretching stress of approximately 14.4 MPa. The residence time (0, 10, 20, 60 and 90 min) was determined by adjusting the roller speed.

For the effect of cyclization reaction on the oxidation reaction, the PAN fibers were subjected to heat treatment at 230 °C in nitrogen with residence time of 0, 10, 20, 60 and 90 min respectively firstly, and transformed into the fibers with different cyclization degree (D_c). Then the samples above were retreated at 230 °C in air for 30 min.

For finding a proper range of oxygen content in the atmosphere, the PAN fibers were heat treated from 170 to 270 °C with a heating rate of 1 °C min⁻¹ in the atmospheres with different oxygen contents. The constant load was 500 cN, equivalent to a stretching stress of approximately 14.4 MPa.

2.3 Measurement of thermal stress and strain

A dynamic mechanical analyzer (Q800, TA instruments, United States) was employed to carry out thermal stress and strain experiments with a fiber bundle containing 100 filaments. The heating rate and the free length of the specimen between the jaws were 3 °C min⁻¹ and 10.4 mm, respectively.

For the thermal shrinkage under various constant tension experiments were carried out in a batch-type oxidation furnace with different oxygen contents, under various constant loads of 2.9, 5.5, 11, 14.4 and 18.2 MPa. The fiber tows were heated from 170 to 270 °C with a heating rate of 1 °C min⁻¹.

2.4 Structure characterization

The infrared spectra (IR) of the as-received fibers was collected on the FT-IR instrument (Thermo Nicolet 6070) by collecting 64 scans at a resolution of 4 cm⁻¹ using KBr disks pressed by mixing KBr. The mass ratio of the fiber sample to KBr was 1 : 300. The degrees of cyclization (D_c) of the fibers were calculated from the following equations:¹⁷in the equation, A_C and A_N indicate the areas of the peaks at 1620–1580 cm⁻¹ (cyclized ring structure) and the area of the peaks at 2240 cm⁻¹ (–C≡N–), respectively. The subscripts 0 and ∞ indicate times are 0 and sufficient long time (in this work was 1000 min: heated from 170 to 270 °C with a heating rate of 0.1 °C min⁻¹) during the thermal oxidative stabilization. For the sake of having an exact , thickness values of both pellets of mixed known sample (treated time was 0 or 1000 min) and KBr were obtained by using digital caliper and used for the normalization of FT-IR spectra. According to calculation, the value of f was 0.29, which was consistent with others' work.^26,27

The degree of cyclization (D_c) of the fibers were also obtained from the Differential Scanning Calorimetry (DSC) (Q20 TA, Instruments, USA) at a heating rate of 10 °C min⁻¹ from 50 to 400 °C under N₂ atmosphere, then calculated by the exothermal enthalpy measured by the DSC as follows:

in the equation, H₀ is the exothermal enthalpy of the as-received fibers, H_i is the exothermal enthalpy of the stabilized fibers.

Elemental analysis was performed by Vario EL III (Germany) EA1112 elemental analysis with full automation.

Thermo-gravimetric analysis (TGA Q5000, TA Instruments, USA) was used to investigate the carbon yield of the as-received fibers at a heating rate of 10 °C min⁻¹ from 50 to 900 °C under N₂ atmosphere.

2.5 Preparation of carbon fibers

A device for continuous stabilization and carbonization was used to prepare carbon fibers with different oxygen contents. The equipment was made up of oxidation, low-temperature carbonization and high-temperature carbonization parts and every part had different temperature zones as shown in Table 1.

Table 1 The temperature setting of different part in the processes from oxidation to carbonization

Processes	Temperature zones (°C)
Oxidation	175	190	210	220	230	235	240	250	263	270	273
Low-temperature carbonization	400	450	600	850
High-temperature carbonization	1200

---

3. Results and discussion

3.1 The evolution of the interaction between cyclization and oxidation reactions during stabilization

Representative FT-IR spectra of PAN during the stabilization process in the atmosphere with 21% oxygen content at 230 °C are presented in Fig. 1. Previous investigators^11,28–30 have indicated that a typical spectrum of PAN fiber showed two characteristic peaks, one is at 2240 cm⁻¹ due to vibrations of nitrile groups (ν_C≡N), and the other is at 2940 cm⁻¹ due to vibrations of hydrocarbon backbone (ν_sC–H in CH₂). In addition, in the spectra, the peaks in the zone of 3600–3100 cm⁻¹ represents the vibrations of hydroxyl, and those at 2870, 1730 and 1450 cm⁻¹ represent the vibrations of hydrocarbon (ν_asC–H in CH₂), carboxylic acid groups (ν_C=O) and in-plane bending vibration of hydrocarbon (ν_sC–H in CH₂) respectively.

Fig. 1 FT-IR spectra of PAN fibers at 230 °C in air with different stabilization time.

It is apparent that peaks at 2240 and 1730 cm⁻¹ decrease, simultaneously, new peaks appear around 1590 cm⁻¹, indicating the transformation of nitrile groups into –C=N– groups, typically in cyclization structure. In addition, the appearance of peaks around 1660 cm⁻¹ represents the formation of ketone structure (C=O) by oxygen uptake reactions.³¹ In order to investigate the evolution of the interaction between cyclization and oxidation reactions of PAN fibers during the stabilization process, obtaining the exact area of each peak is needed. However, the FT-IR spectra in the range of 1800–900 cm⁻¹ exhibit overlapping vibration peaks. In order to overcome such a problem, we used second derivative to determine the numbers and positions of the overlapping peaks, and the curve-fitting was also used to de-convolve the spectra within a selected range. According to this method, the peaks in the 1800–900 cm⁻¹ range were deconvolved into 19 peaks as prior work.²⁶ Then the area of each peak after the deconvolution was calculated. Afterwards, the areas of the peaks at 1620–1580 cm⁻¹ (cyclized ring structure) and the area of the peak at 2240 cm⁻¹ (–C≡N–) were used to calculate the degrees of cyclization (D_c) of the fibers by eqn (1).

The evolution of oxygen contents of PAN fibers in the atmosphere with various oxygen contents are shown in Fig. 2. As we can see, the oxygen contents of the treated PAN fibers in five gases at 230 °C for the first 20 min are almost the same and even lower than that of the precursor fiber, indicating that the oxidation reaction did not largely occur and some oxygen containing groups released during heating. Afterwards, the oxygen content of the fibers increased with the oxygen content increasing in the atmospheres, indicating that a higher oxygen content in the atmosphere led to a higher rate of oxidation reaction, which is consistent with the prior work.³² When the reaction time increases beyond 60 min, oxygen contents of the fibers in low oxygen concentrated atmospheres increase faster. The result may be ascribed to two reasons, and one is the different D_c, the other is the dehydrogenation effect of the oxygen,²⁴ which will be discussed later. To investigate the effect of oxidation reaction on the cyclization reaction, D_c of the heat treated fibers in the atmosphere with various oxygen contents is calculated from eqn (1), and results are shown in Fig. 3. It is clear that the D_c of the heat treated fibers decreases with the oxygen content increasing in the atmosphere when the stabilization time is less than 10 min. However, the rate of cyclization reaction increases with the increasing oxygen content when the stabilization time is increased from 10 to 60 min. With treating time further increasing, the rates of cyclization reaction come to the same when the treating time is more than 60 min.

Fig. 2 Oxygen contents of PAN fibers treated at 230 °C in the atmosphere with different oxygen contents and stabilization time.

Fig. 3 D_c of PAN fibers treated at 230 °C in the atmosphere with different oxygen contents and stabilization time.

Combining the results in Fig. 2 and 3, we can give a comprehension about the relationship between oxidation and cyclization. When the stabilization time was less than 10 min, the oxygen uptake reactions didn't largely occur, the rate of cyclization was inhibited by the oxygen in the atmosphere which absorbed radicals.³³ Therefore, the fiber's D_c decreases with the oxygen content increasing in the atmosphere during this period. In other words, the effect of inhibition caused by the oxygen in the atmosphere played a major role during this period. As the stabilization going beyond 10 min, even the oxygen uptake increased little, dehydrogenation took place more in a more oxygen concentrated atmosphere,²⁴ which promoted the initiation of the cyclization, resulting in cyclization reaction rate increasing. Finally, the promotion by hydrogenation and the inhibition by the oxygen in the atmosphere on the cyclization reached to an equilibration when the treating time was more than 60 min.

Since oxidation reaction is indispensable during the stabilization process,²¹ it is very significant to investigate the effect of cyclization on the oxidation reaction of PAN fibers during the stabilization process. Thus, the oxygen uptakes of the fibers with different D_c were detected by elemental analysis and the results are shown in Fig. 4. As is clearly shown, oxygen uptakes of the fibers increase with the D_c increasing dramatically, indicating that the ring structure formed by cyclization reaction can promote the oxidation reaction.^25,34 Back to the assumption about Fig. 2, it is considered that on the one side, the higher D_c in a lower concentrated oxygen atmosphere made the oxidation faster, on the other side, more dehydrogenation in a more concentrated oxygen atmosphere made the oxidation going harder. So it is possible that the fibers treated in an atmosphere with lower oxygen will be oxidized at a more accelerated rate at the later stage, which means the gradient of oxygen uptake along the fiber diameter may become more serious because of the competition between the diffusion and oxidation.

Fig. 4 Oxygen uptake of the fibers with different D_c treated at 230 °C in air for 30 minutes.

From the above results, it is concluded that an atmosphere with higher oxygen content didn't cause lower cyclization degree and excessively higher oxygen content in the final oxidized fiber, while it is concerned if an atmosphere with low oxygen results in deficient oxidation.

3.2 Effect of oxygen content in the atmospheres on thermal stress and shrinkage

Cyclization and oxidation during stabilization can be reflected on the thermal stress. The thermal stress evolution with temperature rising in various atmospheres is shown in Fig. 5, which can be divided into 4 regions according to the thermal stress. In region I (50–150 °C), the thermal stress keep increasing and is not affected by the oxygen content in the atmosphere, which is generally considered to be caused by molecular conformation rearrangement that is restricted by crystalline structure.³⁵ In region II (150–230 °C), the thermal stress dramatically decease with temperature rising and is still not affected by the oxygen content, during which the force by molecular conformation rearrangement is gradually released and the crystalline structure collapse finally. In region III (230–280 °C), the stress continue to drop a little and then begin to rise, and the curves apparently go to two categories according to the heated atmosphere, one for nitrogen and another for oxygen containing atmospheres. The stress in nitrogen rises earlier and keeps higher than those in oxygen containing atmospheres. Beyond all doubt, during this stage nitrile cyclization makes the fibers shrink, and if the interaction between molecules is strong enough, the stress will go up, reversely, the stress will go down. Most literatures reported crystalline structure collapse at about 230 °C.^36–38 With temperature increasing, the crystalline and dipole–dipole force of nitriles collapsing makes the stress drop. However, if a molecular interaction takes place, the stress will go up. According to our research,³⁹ nitriles tend to crosslink in an oxygen free atmosphere. Therefore it is considered that the existence of oxygen dramatically inhibit nitrile crosslinking along with depressing cyclization, which results less stress as shown in Fig. 6. However, even either nitrile crosslinking or crosslinking by oxidation took place rarely at this stage, the amounts of oxygen containing groups are proportional to the oxygen contents in the atmospheres. Resultantly, when the temperature rise into region IV between 280 and 350 °C, nitrile crosslinking broke along with cyclization finishing, and the stress begin to decay, fastest in nitrogen, while slower in more concentrated oxygen containing atmosphere, because more concentrated oxygen groups generated more crosslinking.

Fig. 5 Development of thermal stress in the atmosphere with various oxygen contents and the heating rate were 3 °C min⁻¹.

Fig. 6 The mechanism of cyclization initiation of the irradiated PAN fibers during thermal treatment.

The results make us come back to the question: whether the final cyclization degree is reduced by more concentrated oxygen? Since cyclization corresponds to shrinkage which is also related to the stress. We checked the shrinkage of the fibers heat treated in atmospheres with various oxygen contents under different stress, as shown in Fig. 7. It is noted that when the stress is less than 11 MPa, the fibers in all the atmospheres shrink to nearly the same degree. Therefore it is considered the cyclization is not affected by the oxygen content in the atmosphere. When the stress increases to 11 MPa, the shrinkage decreases slightly with the oxygen increasing. When the stress increased to 14.4 MPa, the fibers stopped shrinking and began to extend, and the higher the oxygen in the atmosphere, the longer the fibers extended. The more extension can be ascribed to the oxygen in the atmosphere inhibiting the nitriles crosslinking at early stage. When the stress increases to 18.2 MPa, the fibers break in nitrogen for lack of crosslinking when the cyclization was completed.

Fig. 7 Changes of shrinkage of the fibers heat treated in the atmosphere with various oxygen contents.

3.3 Effect of gas atmospheres on stabilization degree of the PAN stabilized fibers

From the analysis above, it seems that the oxygen content in the atmosphere doesn't affect the cyclization degree of the final stabilized fibers, and oxygen uptake of the fibers in more oxygen concentrated atmospheres is not excessively higher. While how about the cases for full stabilized fibers? For making sure such questions, the D_c and oxygen uptake of the final stabilized fibers were investigated, where the D_c was calculated by FT-IR and the oxygen uptake was analyzed by elemental measurement instrument.

Fig. 8 shows the FT-IR spectra of the stabilized fibers. The same method was used to calculate the D_c values, and results are shown in Fig. 9. Almost no exothermic peak appears below 270 °C, and the initial reaction temperature of the peaks at higher temperature increases with the oxygen contents increasing. This suggests that more oxygen uptake make more harder cyclization happened during stabilization, leaving the un-cyclized nitrile groups harder to cyclize. It is recognized that the D_c values are not affected by oxygen contents in the atmosphere.

Fig. 8 FT-IR spectra of stabilized fibers obtained during the thermal treatment procedures with various oxygen contents.

Fig. 9 The degree of cyclization (D_c) of stabilized fibers.

The oxygen uptake of the fibers stabilized in atmospheres with different oxygen contents are shown in Fig. 10. The results indicate that oxygen uptake increases with oxygen content increasing in the atmosphere, to which the hydrogen content aliened. However, when the oxygen content in the atmosphere is above 21%, the uptake of oxygen reduced along with hydrogen reduced further more. It is considered that the higher oxygen atmosphere make the fibers dehydrogenation more and cyclized less and consequently slow down the oxidation at the successive stage. On the contrary, it means that the lower oxygen atmosphere make the fibers oxidized less and cyclized more at the early stage and consequently accelerated the oxidation at the successive stage for higher cyclization degree. Thus a high oxygen content atmosphere tend to result in uniform oxidation and to avoid excessive oxidation, while a low oxygen content atmosphere might result in insufficient or non-uniform stabilization across fiber diameter.

Fig. 10 Elemental analysis of stabilized fibers.

3.4 Effect of gas atmospheres on carbon yields and skin–core structure of the stabilized fibers

Insufficient stabilization may result in low carbon yield, which means increase of defects in the carbon fiber, so it is necessary to investigate the effect of stabilization gas atmospheres on carbon yields of the stabilized fibers. The results are depicted in Fig. 11. Between 300 and 650 °C, the rate of weight loss of the fibers treated in nitrogen is faster than the fibers treated in other atmospheres, which can be ascribed to the rapid loss of hydrogen that has been removed less for lack of oxygen during stabilization. At the temperatures above 650 °C, the fibers with none or insufficiency crosslink would be decomposed quickly. Hence, the rate of weight loss of samples treated in the atmosphere with lower oxygen content (<5%) is faster than others. It is also noted that the fibers treated in atmospheres with oxygen contents above 10% have no obvious difference on carbon yield. So an atmosphere with more that 10% oxygen content can satisfy sufficient oxidation.

Fig. 11 TGA plots (a) and final carbon yields (b) of the stabilized fibers in nitrogen at a heating rate of 10 °C min⁻¹.

Even the total stabilization is enough in the atmosphere with oxygen content of 10–21%, as analysis above, non-uniform stabilization may occur in a low oxygen content atmosphere. Therefore the micrographs across the stabilized fibers in different oxygen content atmospheres were investigated, as shown in Fig. 12. It is noted that a skin–core structure exist in the fibers treated in the atmospheres with 5% and 10% oxygen contents. When the oxygen content in the atmosphere is up to 15%, the skin–core structure becomes un-conspicuous. So even the stabilized fibers in different oxygen content atmospheres are similar in the total oxygen uptake, they are much different in stabilization degree across diameter. One reason for this is as analyzed above, low oxygen content atmosphere make the fiber oxidized slower at initial stage and faster at the successive stage with insufficient oxygen diffusion. The other reason is that a high degree of cyclization leads to a dense structure that depresses the inward diffusion of oxygen later.⁴⁰ Therefore, to guarantee the high performance of stabilized fibers, above 15% oxygen content in the atmosphere during stabilization is needed.

Fig. 12 SEM profiles of stabilized fibers obtained during the thermal treatment procedures with 5% (a), 10% (b), 15% (c), and 21% (d) oxygen contents.

3.5 Effect of gas atmospheres on mechanical properties of carbon fibers

From the analysis above, the optimum oxygen content in the atmosphere during thermal stabilization process was determined to be 15–21%. In order to verify this conclusion, the mechanical properties of corresponding carbon fibers were measured on a universal testing instrument and results were shown in Fig. 13. From the data, we can see that the tensile strength of carbon fibers firstly increases when the oxygen content of the stabilization atmosphere increases from 10% to 15%, then decreases when oxygen content increased anymore. However, the tensile strength of carbon fibers derived from stabilized fibers treated in 30% oxygen content atmosphere is still larger than that in 10%. But the Young's modulus is almost unchanged with oxygen content atmosphere. Therefore, the results indicate that best oxygen concentration is 15–21%, which is consistent with the stabilization analysis.

Fig. 13 Mechanical properties of carbon fibers.

4. Conclusions

From the analysis above, it could be concluded as follows. Although high oxygen content atmosphere inhibited the cyclization at early stage of the stabilization, it did not result in low cyclization degree and excessive oxidation in final stabilized fibers. However, low oxygen content atmosphere introduced skin–core structure for insufficient oxygen diffusion and rapid oxidation rate at later stage. Therefore, optimum oxygen content in the atmosphere during thermal stabilization process was determined to be 15–21%. And the mechanical properties of carbon fiber were also consistent with the results.

Acknowledgements

This research was supported by National Natural Science Foundation of China (No. 51372037).

Notes and references

1. J. Zhu, J. Mater. Chem. C, 2014, 2, 715–722.
2. J. Zhu, Polymer, 2011, 52, 2947–2955.
3. D. Zhang, J. Phys. Chem. C, 2010, 114, 212–219.
4. D. Zhang, Polymer, 2009, 50, 4189–4198.
5. Y. Zhao, Phys. Chem. Chem. Phys., 2015, 17(2), 1498–1502.
6. J. Liu, Macromol. Mater. Eng., 2015, 300(3), 358–368.
7. S. Liu, Mater. Lett., 2016, 172, 149–152.
8. J. Huang, Macromol. Mater. Eng., 2016, 10, 1002.
9. H. Lyu, J. Mater. Chem. A, 2016, 4, 9881–9889.
10. J. Liu, P. H. Wang and R. Y. Li, J. Appl. Polym. Sci., 1994, 52, 945–950.
11. M. S. A. Rahaman, A. F. Ismail and A. Mustafa, Polym. Degrad. Stab., 2007, 92, 1421–1432.
12. P. Rangarajan, V. A. Bhanu, D. Godshall, G. L. Wilkes, J. E. McGrath and D. G. Baird, Polymer, 2002, 43, 2699–2709.
13. H. Ogawa and K. Saito, Carbon, 1995, 33, 783–788.
14. Y. Liu, H. G. Chae and S. Kumar, Carbon, 2011, 49, 4477–4486.
15. N. Grassie and J. N. Hay, J. Polym. Sci., 1962, 56, 189–202.
16. I. Shimada, T. Takahagi, M. Fukuhara, K. Morita and A. Ishitani, J. Polym. Sci., Part A: Polym. Chem., 1986, 24, 1989–1995.
17. A. K. Gupta, D. K. Paliwal and P. Bajaj, J. Macromol. Sci., Part C, 1991, 31, 1–89.
18. L. M. Manocha and O. P. Bahl, Fibre Sci. Technol., 1980, 13, 199–212.
19. O. P. Bahl and L. M. Manocha, Carbon, 1974, 12, 417–423.
20. Y. Lü, D. Wu, Q. Zha, L. Liü and C. Yang, Carbon, 1998, 36, 1719–1724.
21. C. Blanco, S. Lu, S. P. Appleyard and B. Rand, Carbon, 2003, 4, 165–171.
22. M. S. Morales and A. A. Ogale, J. Appl. Polym. Sci., 2013, 128, 2081–2088.
23. W. Liu, M. Wang, Z. Xing, Y. Qi and G. Wu, Radiat. Phys. Chem., 2012, 81, 622–627.
24. E. Fitzer and D. J. Müller, Carbon, 1975, 13, 63–69.
25. W. D. A. S. Potter, Nature, Phys. Sci., 1972, 236, 30–32.
26. N. U. Nguyen-Thai and S. C. Hong, Carbon Mater., 2013, 46, 5882–5889.
27. G. L. Collins, N. W. Thomas and G. E. Williams, Carbon, 1988, 26, 671–679.
28. Q. Ouyang, L. Cheng, H. Wang and K. Li, Polym. Degrad. Stab., 2008, 93, 1415–1421.
29. Z. Wangxi, L. Jie and W. Gang, Carbon, 2003, 41, 2805–2812.
30. W. Geng, T. Nakajima, H. Takanashi and A. Ohki, Fuel, 2009, 88, 139–144.
31. J. Mittal, O. P. Bahl, R. B. Mathur and N. K. Sandle, Carbon, 1994, 32, 1133–1136.
32. T. Sun, Y. Hou and H. Wang, J. Macromol. Sci., Part A: Pure Appl.Chem., 2009, 46, 807–815.
33. W. Liu, M. Wang, Z. Xing and G. Wu, Radiat. Phys. Chem., 2012, 81, 835–839.
34. W. Johnson and W. WATT, Nature, 1975, 5523, 210–212.
35. I. Harrison and A. Gupta, Carbon, 1996, 34, 1427–1455.
36. A. Gupta and I. R. Harrison, Carbon, 1996, 34, 1427–1445.
37. B. Wang, S. Xiao, W. Cao, X. Shi and L. Xu, J. Appl. Polym. Sci., 2012, 124, 3413–3418.
38. G. Wu, C. Lu, L. Ling, A. Hao and F. He, J. Appl. Polym. Sci., 2005, 96, 1029–1034.
39. L. Zhou, et al., Polym. Degrad. Stab., 2016, 128, 149–157.
40. M. Yu, C. Wang, Y. Bai, Y. Xu and B. Zhu, J. Appl. Polym. Sci., 2008, 107, 1939–1945.

---