Modification of carbon fiber by air plasma and its adhesion with BMI resin

Abstract

Carbon fiber (CF) was modified by air plasma at 30 Pa and 200 W with different treatment times. Interlaminar shear strength (ILSS) of CF/BMI composites suggested that air plasma-treated CFs showed better adhesion with BMI resins than the untreated resins; the maximum increment was 20.4% after plasma treatment for 900 s. Meanwhile, we investigated the surface changes of CFs using X-ray photoelectron spectroscopy (XPS) and atomic force microscopy (AFM). The results indicated that air plasma treatment roughened the CF surfaces and improved the polarity of CFs by enhancing the polar group content. These changes were believed to be the major reasons for the improvement of interfacial adhesion. The fracture morphology observed by scanning electron microscopy (SEM) indicated that the fracture failure mode changed from adhesive failure to cohesive failure. After boiling in water for 48 h, the water absorption of the composites was low while the ILSS of the composites retained about 90%, which implied that the CF/BMI composites had good humidity resistance properties.

---

1. Introduction

Carbon fiber reinforced resin composites are used in many frontier areas such as the automobile industry, aerospace and military, etc. As is well-known, the mechanical properties of this kind of material not only depends on the strength of the carbon fibers and its matrix, but is also decided by the interfacial adhesion of the composites. However, carbon fiber has a smooth surface and low polarity, as well as a lack of functional groups and has difficulties in forming strong chemical bonds with matrix; all these properties lead to weak fiber–matrix adhesion, such some solutions to improve its surface free energy and the adhesion with the matrix were necessary, such as ozone, chemical modification, electrochemically, plasma modifications, etc., and are attractive to numerous researchers.^1–10

Out of these methods, plasma treatment is popular because of its advantages over other modifications:^11–13 (1) dry and clean, plasma treatment does not produce contaminants and chemicals, and thus would be environmentally friendly; (2) effective, the time of plasma treatment is only several seconds or a few minutes; (3) low cost of energy; (4) they generate the desired results while the fiber tensile strength would be retained at a high ratio. Due to these superiorities plasma modification of polymer materials was considered as a promising method in material surface modifications.

Air plasma was employed to modify carbon fibers in our study to improve the interfacial adhesion of carbon fibers and bismaleimide resins. Bismaleimide resin is a kind of novel thermosetting resin, which has good heat resistance and mechanical properties, it is considered as an alternative to epoxy resin and is promising in fiber-reinforced composites.^14–16 Therefore, CF/BMI composites are an advanced fiber–matrix composites for use in the aerospace and military fields. Good interfacial adhesion between CF and BMI resins is important and beneficial for their applications.

Interlaminar shear strength (ILSS) of CF/BMI composites was measured to estimate the interfacial adhesion between carbon fibers and BMI resin. In order to investigate the reason for the improvement in interfacial properties, some methods were used to study the changes of CF surface treated by air plasma. The chemical components and groups of the carbon fiber surfaces were detected by X-ray photoelectron spectroscopy (XPS). The surface morphologies were studied by atomic force microscopy (AFM). Furthermore, the fracture morphology of the composites was investigated using scanning electron spectroscopy (SEM). In addition, water absorption and humidity resistance property of CF/BMI composites were studied.

2. Materials and methods

2.1 Materials

BMI matrix was used in this paper (QY8911-II), which is a novel thermosetting resin received from AVIC Beijing Aeronautical Manufacturing Technology Research Institute. The viscosity of this resin was in the range of 0.5–1.5 Pa s at 80 °C. The BMI resin was prepared by modifying 4,4′-diphenylmethane bismaleimide with polyether sulfone.

Carbon fiber (CF) was obtained from Toray Ltd.co, Japan as T700SC and washed in acetone for 48 h at room temperature before plasma treatment.

2.2 Plasma treatment for CFs

CFs were rolled on glass frames and treated by inductively coupled plasma (ICP) with a 1 kW radio frequency (13.56 MHz) energy source. Air was fed into the vacuum chamber as a carrier gas at flow rates of 10–12 sccm to maintain the pressure at 30 Pa; meanwhile, the plasma power was controlled at 200 W and the plasma treatment time was varied from 300 s to 1500 s.

2.3 Characterization

2.3.1 Composites preparation. After air plasma treatment, carbon fibers were removed from the plasma chamber and immediately impregnated through BMI/acetone (40 wt%) solution. Then, they were dried at 40 °C for 1 h in a vacuum oven to remove the acetone and the prepregs were obtained. These were then tailored into specific dimensions and put into a mould, and then the CF/BMI composites were prepared by a compression molding process at 130 °C for 1 h, 190 °C for 3 h, and 230 °C for 3 h.

2.3.2 Characterizations of composites. According to ASTM D 2344, the three-point, short-beam bending test method was used to measure the ILSS of CF/BMI composites on a Shimadzu universal testing machine (dimension was 25 mm × 6 mm × 2 mm). A constant cross-head movement rate of 2 mm min⁻¹ was used. More than five successful samples were tested for every ILSS and the average value was calculated. Eqn (1) was applied to obtain the ILSS:where τ is ILSS, P_b represents the maximum load, b and h are the width and height of the specimen, respectively.

2.3.3 Chemical analysis of carbon fibers. Carbon fibers were subjected to X-ray photoelectron spectroscopy (XPS; ESCALAB 250, Thermo) to study the surface chemical components. An Al Kα (hν = 1486.6 eV) X-ray monochromatic source (15 kV, 150 W) was employed to excite the photoelectrons. The XPS spectra were obtained when a take-off angle was 90° relative to the sample's surface. The pass energy was 20 eV while the step energy was 0.1 eV.

2.3.4 Surface morphology and roughness of carbon fibers. Atomic force microscopy (AFM; Picoplus II, Agilent) using a tapping mode was applied to investigate the surface roughness and morphology of the carbon fibers. The scan area was 6 μm × 6 μm. The arithmetic and root mean roughness (R_a and R_q) were calculated by the instrument software to evaluate the surface roughness.

2.3.5 Water absorption and humid resistance. The CF/BMI composites were boiled in distilled water for 48 h to measure the water absorption of the composites. The weights of the composites were measured at fixed time interval and accurate to 0.1 mg. The rate of water absorption was calculated according to eqn (2):where m_i (g) is the weight of boiled samples and m₀ (g) is the initial weight of the dry specimen.

After the composites were boiled in water for 48 h, the retained ILSS was measured to investigate the humidity resistance properties of the composites.

2.3.6 Fracture morphology of CF/BMI composites. Scanning electron microscopy (SEM; QUANTA 450, FEI) was applied to observe the interlaminar shear rupture morphologies of CF/BMI composites. The magnification was set at 2000×.

3. Results and discussions

3.1 ILSS Changes of CF/BMI composites

Fig. 1 shows the ILSS of the CF/BMI composites versus the plasma treatment time. It can be observed from the figure that the original samples of CF/BMI composites had an ILSS of about 108.0 MPa. After the plasma modification, the ILSS of the composites improved as the plasma treatment duration was extended; the ILSS reached 112.5 MPa and 123.8 MPa at 300 s and 600 s plasma treatment, respectively. Then, the ILSS achieved its highest value 130.6 MPa at 900 s plasma treatment and had an increment of 20.4%. These results indicated that the interfacial adhesion of the CFs and BMI resins could be improved effectively by air plasma modification, and we could assume that the surface properties of CFs were altered by plasma; therefore, the newly-created chemical and physical properties of the fibers were able to combine better with BMI resins and enhance the adhesive strength between them. However, when the plasma treatment time exceeded 900 s, the ILSS of the composites decreased to 122.1 MPa and 111.3 MPa for 1200 s and 1500 s, respectively. This was attributed to overdo plasma treatment, excessive bombardment of plasma destroyed the newly-formed surfaces, so the interfacial properties of the composites declined as well.

Fig. 1 ILSS of CF/BMI composites versus air plasma treatment time.

Plasma treatment was proven to be an effective method to improve the ILSS of the CF/BMI composites. However, 900 s was the best treatment time and above this the results were reduced; therefore, the plasma treatment time should be controlled in a proper way to obtain better results.

3.2 Chemical changes of CFs

XPS was applied to study the chemical compositions of CFs both before and after air plasma treatment. The CFs contain three kinds of atoms: C, O, and N, and the relative contents of each atom was listed in Table 1. As can be seen, the C, O and N atoms of the untreated CFs were 82.06%, 16.91% and 1.02%, respectively. After treatment by air plasma for 900 s, the C content decreased to 78.82% while the O content increased to 20.34%; it was assumed that this was because the high-energy particles in air plasma, such as O⁺ and O˙, reacted with the groups on the CF surfaces and some oxygen-containing groups were introduced onto CF surfaces, such as –C–O– and –C=O, etc. However, the N content of the CFs was low and decreased along with the plasma treatment time, indicating that the air plasma cannot effectively introduce nitrogen-containing groups onto CFs. Fig. 2 shows the changes of the atom ratios for CFs. The O/C and O + N/C both increased after treatment by air plasma and reached their highest values at 900 s plasma treatment but the N/C stayed nearly constant for all the samples; therefore, the introduction of O atoms onto the CFs played an important role in improving the polarity of the CFs.

Table 1 Atom contents of CFs

Samples	C	O	N
Untreated	82.06	16.91	1.02
300 s	81.91	17.16	0.93
900 s	78.82	20.34	0.84
1500 s	83.01	16.57	0.42

---

Fig. 2 Atomic ratio of CFs.

Further studies were carried out to investigate the changes in functional groups by C1s spectra of CFs both before and after air plasma treatment. In Fig. 3(a), the original CFs contained five functional groups: –C–C–, –C–N–, –C–O–, –O–C=N– and –O–C=O, which were fitted at 284.5 eV, 285.4 eV, 286.6 eV, 288.2 eV and 289.0 eV, respectively.^17–19 After air plasma treatment, the types of functional groups were the same as the untreated sample and there were no newly-created groups but the area of each group changed with the plasma treatment time. The relative contents of all the chemical groups were calculated from the areas of the C1s spectra and are listed in Table 2. The untreated CF had a content of 64.12%, 17.24%, 11.15%, 3.03% and 4.46% for –C–C–, –C–N–, –C–O–, –O–C=N– and –O–C=O, respectively, and the ratio of polar to nonpolar groups was 0.56. However, after CFs were treated by air plasma for 300 s, the –C–C– content was 56.13% while the polar groups increased to 43.87%; meanwhile, the polar groups/nonpolar groups ratio increased to 0.78. When the plasma treatment time was 900 s, the polar groups/nonpolar groups ratio reached its highest value of 0.96. In summary, the functional groups of CFs were changed by air plasma treatment, the polarity of CFs improved, the chemical reactivity between CFs and BMI resin was activated, and the interfacial adhesion of CF/BMI was also enhanced as well. Otherwise, when the plasma treatment time was 1500 s, the polar groups of the CF surface were destroyed and the polar groups/nonpolar groups ratio decreased, so the polarity of CF reduced.

Fig. 3 C1s spectra of CFs: (a) untreated, (b) treated for 300 s, (c) 900 s and (d) 1500 s.

Table 2 Concentrations of functional groups of CFs

Samples	–C–C–	–C–N–	–C–O–	–O–C=N–	–O–C=O	Polar/nonpolar
Untreated	64.12	17.24	11.15	3.03	4.46	0.56
300 s	56.13	25.58	10.83	3.25	4.21	0.78
900 s	51.01	23.80	16.15	5.46	3.58	0.96
1500 s	55.52	24.66	10.88	4.90	4.05	0.80

---

3.3 Morphology and surface roughness of CFs

AFM is an effective and frequently-used method to investigate the surface morphology of polymer materials at the nano-scale.^20,21 Fig. 4 shows the AFM images of CF both untreated and plasma-treated images. The untreated CF has a relatively smooth surface with only some small stripes, which could be attributed to the production processes such as stretch and sizing, etc. However, after the air plasma treatment, the CF had obviously complex surfaces. For 300 s treatment, some embossments were generated onto the CF surfaces but the changes were little; when the treatment time was extended to 1500 s, the number of bulges increased and the size of some became larger, and the fiber surfaces changed dramatically. The bombardment of the high-energy particles in the plasma on the fiber surfaces may be the primary reason for these phenomena.²² The sputtering of the plasma may destroy the fiber chains and some atoms would be knocked off; however, when the plasma treatment time was prolonged, these atoms and chains may accumulate together, forming embossments.^23,24

Fig. 4 AFM images of CFs: (a) untreated, (b) treated for 300 s, (c) 900 s and (d) 1500 s.

According to the 3D images of AFM, the surface roughness of CF was calculated and listed in Table 3. The same trend with the morphology of the fibers could be seen in the values of the fiber roughness. R_q and R_a were 238.9 nm and 199.5 nm for untreated CF, respectively. For 300 s plasma treatment, the R_q and R_a increased to 246.1 nm and 202.6 nm, respectively, indicating the complicated morphology of the plasma-treated CF. Then, the surface roughness of the CFs increased as the treatment time increased and reached 301.0 nm for R_q and 250.1 nm for R_a at 1500 s treatment; the values verified the role the plasma in modifying the CF surfaces.

Table 3 R_a (arithmetic mean roughness) and R_q (root mean roughness) of CFs

Treatment time (s)	Rq (nm)	Ra (nm)
0 (Untreated)	238.9	199.5
300	246.1	202.6
900	286.6	231.5
1500	301.0	250.1

---

The complicated surfaces of the CF could provide sufficient contact areas for the BMI resin to wet the fiber surfaces, and then the mechanical interlocking between the CF and the BMI resins became stronger; consequently, the ILSS of the CF/BMI composites improved.

3.4 Water absorption and humidity resistance of CF/BMI composites

Water absorption of the fiber–matrix composites was another method to measure the interfacial adhesion of the composites. The mechanism of the moisture permeating into the composites was complex and different compared with the pure matrix. The moisture infiltrated into the composites not only from the resin and reinforcement fibers but also from the interface; therefore, the water absorption of the composites can reflect the interfacial adhesion of the CF/BMI composites because the fiber volume of the composites was constant in all the samples.^25–28

The water absorption and humidity resistance property of the CF composites are shown in Table 4. The untreated CF/BMI composites had water absorption of 1.32% after boiling for 48 h in distilled water, while the value for the plasma-treated samples was 1.18% (CF was treated by air plasma at 30 Pa, 200 W, 900 s). The results were consistent with the ILSS values, indicating that the plasma treated composites had better interfacial adhesion and can effectively resist water permeation.²⁸ After moisture permeated into the fiber-reinforced composites, it can produce irreversible effects on composites such as chemical degradation of matrix, attack on the interface, and micro-cracking of internal voids.²⁹ Because of these phenomena, the interfacial adhesion of the composites would be destroyed, and their ILSS indicates the humidity resistance property of the CF/BMI composites. The results in Table 4 showed that the untreated and plasma-treated composites retained ILSS ratio of 91.2% and 89.5%, respectively. First, both of the composites had good humidity resistance properties. Second, the treated CF/BMI composites had a relatively lower retained ILSS ratio, which may be attributed to the more reactive chemical groups on the CFs after plasma treatment; therefore the water molecules may influence the chemical groups more seriously and lead to a lower retained ILSS ratio.²⁴

Table 4 Water absorption and retained ILSS ratio

Samples	Water absorption (%)	Retained ILSS ratio (%)
Untreated	1.32	91.2
Air plasma treated (900 s)	1.18	89.5

---

3.5 Fracture morphology of CF/BMI composites

The interfacial adhesion of the fiber–matrix composites influenced the load transfer of the composites, and the fracture morphology of the composites was also influenced by the adhesive properties of the composites. In general, when loads were applied to the composites, the stress distribution on the fiber and matrix was in accordance with the Kelly–Tyson theory;^24,30,31 when the adhesion of the composites was weak, the reinforcement fibers were pulled out from the resin, the fiber is smooth and adhered with little resin. On the contrary, when the adhesive strength of the composites was sufficiently enough, failure occurred in the matrix.

The fracture morphology of the CF/BMI composites were scanned by SEM, as shown in Fig. 5. The differences between the two images can be clearly observed. In Fig. 5(a), the fibers in the composites was untreated; after destruction of the composites, the fibers were clean and, as we pointed out in the images, little resin adhered to them. The failure of the composites occurred on the interface between the fiber and the matrix. The loads were not well transferred from the matrix to the fiber, thus the load-capacity of the composites was low and led to a low ILSS value. However, after the CFs were treated by plasma, the morphology of the composites after being fractured by the loads is shown in Fig. 5(b); there was more resin stuck to the fibers and numerous fibers were embedded into the resin, which are also indicated by arrows, indicating that the primary failure occurred in the matrix, which was due to the good transfer of the load. In the failure process, this failure mode consumed more energy and gave a higher ILSS value.

Fig. 5 Fracture morphology of CF/BMI composites: (a) untreated and (b) treated for 900 s.

4. Conclusions

Air plasma was confirmed to be an effective method to improve the interfacial adhesion between CFs and BMI resin. The ILSS of CF/BMI composites increased by 20.4% from 108.0 MPa to 130.6 MPa after CFs were treated by air plasma for 900 s at 30 Pa and 200 W. The O atom content increased after air plasma treatment while the N atom content decreased slightly, indicating that air plasma could not incorporate N onto CFs but the O + N/C ratio was enhanced after plasma treatment. Furthermore, the nonpolar group –C–C– content decreased and the content of polar groups such as –C–N–, –O–C=N– improved; the polar groups/nonpolar groups ratio also improved. These changes increased the polarity and reactivity of CFs, which was good for the ILSS of the CF/BMI composites. The AFM revealed that the physical morphology was roughened by air plasma and the surface roughness of CFs also increased significantly. The physical interlocking between CFs and BMI resins was strengthened and also contributed to the ILSS of CF/BMI composites. The water absorption of the plasma-treated CF/BMI composites was lower than the untreated sample; both of them had good humidity resistance properties because they both retained an ILSS ratio of about 90%. The place where the composites cracked was removed from the interface with the matrix after the CF was treated with air plasma, which also indicated the improved interfacial adhesion.

Acknowledgements

This work was supported by the National Defense 12th 5-year program Foundational Research Program (no. A3520110001), Liaoning Excellent Talents in University (no. LR2013003), the National Natural Science Foundation of China (no. 51303106) and Key Laboratory of Materials Modification by Laser, Ion and Electron Beams of Ministry of Education (DP1051204). The authors give special thanks to Mr. Xinglin Li for his XPS analysis.

References

1. J. P. Boudou, J. I. Paredes, A. Cuesta, A. Martínez-Alonso and J. M. D. Tascón, Carbon, 2001, 41, 41.
2. S. J. Park, Y. H. Chang, C. W. Moon, D. H. Suh, S. S. Im and Y. C. Kim, Bull. Korean Chem. Soc., 2010, 31, 335.
3. J. F. Xie, D. W. Xin, H. Y. Cao, C. T. Wang, Y. Zhao, L. Yao, F. Ji and Y. P. Qiu, Surf. Coat. Technol., 2011, 206, 191.
4. K. M. Ma, P. Chen, B. C. Wang, G. L. Cui and X. M. Xu, J. Appl. Polym. Sci., 2010, 118, 1606.
5. J. Zheng, Z. Q. Zhang and L. H. Meng, Mater. Chem. Phys., 2006, 97, 167.
6. F. Severin, L. Formaro and P. Mario, et al., Carbon, 2002, 40, 735.
7. V. K. Kaushik, Polym. Test., 2000, 19, 17.
8. Z. R. Yue, W. Jiang and L. Wang, et al., Carbon, 1999, 37, 1785.
9. R. J. Zaldivar, J. P. Nokes, D. N. Patel, B. A. Morgan, G. Steckel and H. I. Kim, J. Appl. Polym. Sci., 2012, 125, 2510.
10. H. Li, H. Liang, F. He, Y. Huang and Y. Z. Wan, Surf. Coat. Technol., 2009, 203, 1317.
11. R. Z. Li, L. Ye and Y. W. Mai, Composites, Part A, 1997, 28A, 73.
12. D. Papaksontantinou, E. Amanatides, D. Mataras, V. Ioannidis and P. Nikolopoulos, Plasma Processes Polym., 2007, 4, s1057.
13. Z. Liu, P. Chen, D. B. Han, F. Lu, Q. Yu and Z. F. Ding, Vacuum, 2013, 92, 13.
14. Z. R. Rong, F. W. Huang, X. N. Shen and F. R. Huang, Polym. Compos., 2008, 483.
15. H. Y. Tang, N. H. Song, Z. H. Gao, X. F. Chen, X. H. Fan, Q. Xiang and Q. F. Zhou, Eur. Polym. J., 2007, 43, 1313.
16. X. H. Xiong, P. Chen, N. B. Zhu, Q. Yu, J. X. Zhang and B. C. Wang, Polym. Eng. Sci., 2011, 1659.
17. Y. Ren, C. X. Wang and Y. P. Qiu, Surf. Coat. Technol., 2008, 202, 2670.
18. C. Ronning, H. Feldermann, R. Merk and H. Hofsäss, Phys. Rev. B: Condens. Matter Mater. Phys., 1998, 58, 2207.
19. Q. Wang, P. Chen, C. X. Jia, M. X. Chen and B. Li, Appl. Surf. Sci., 2011, 258, 513.
20. Z. Liu, P. Chen, X. L. Zhang, Q. Yu, K. M. Ma and Z. F. Ding, Appl. Surf. Sci., 2013, 283, 38.
21. B. Song, L. H. Meng and Y. D. Huang, Appl. Surf. Sci., 2012, 258, 5505.
22. D. Liu, P. Chen, J. J. Mu, Q. Yu and C. Lu, Appl. Surf. Sci., 2011, 257, 6935.
23. W. Chen, C. Jie-rong and L. Ru, Appl. Surf. Sci., 2008, 254, 2882.
24. D. Liu, P. Chen, M. X. Chen, Q. Yu and C. Lu, Appl. Surf. Sci., 2013, 257, 10239.
25. X. Y. Zhang, P. Chen, D. B. Han, Q. Yu, Z. F. Ding and X. L. Zhu, Appl. Surf. Sci., 2013, 266, 110.
26. A. J. Cervenka, D. J. Bannister and R. J. Young, Composites, Part A, 1998, 29, 1137.
27. R. Selzer and K. Friedrich, Composites, 1997, 28, 595.
28. J. E. Lundgren and P. Gudmundson, Compos. Sci. Technol., 1999, 59, 1983.
29. P. J. Herrenra-Franco and L. T. Drzal, Composites, 1992, 23, 2.
30. D. P. N. Vlasveld, P. P. Parlevliet, H. E. N. Bersee and S. J. Picken, Composites, Part A, 2005, 36, 1.
31. P. Chen, C. Lu, Q. Yu, Y. Gao, J. F. Li and X. L. Li, J. Appl. Polym. Sci., 2006, 102, 2544.

---