Plasma oxidation of electrospun carbon nanofibers as supercapacitor electrodes

Abstract

The effect of oxygen-plasma treatment on the capacitive behavior of electrospun carbon nanofibers is investigated. The plasma-modified fibers are enriched in oxygen functionalities, resulting in increased fiber wettability. Moreover, the plasma-treated nanofibers exhibit a specific capacitance of 377.0 F g⁻¹, which is more than twice that of the nontreated nanofibers.

---

Electrical double-layer capacitors (EDLCs) have received considerable attention because of the increasing demands for new electrical-energy storage systems. EDLCs store electrical energy by an electric double layer formed at the interface between the electrode surface and electrolyte.¹ Ions of opposite charges are attracted by electrostatic force, which is free of fast faradic reaction and exhibits no pseudocapacitance effects. The conventional electrode materials for EDLCs are carbonaceous materials, including carbon nanotubes,^2–4 graphenes,^5–7 carbon nanofibers,⁸ and carbon spheres.⁹ Carbon nanotubes and graphenes exhibit high accessible specific surface areas with superior electrical conductivity. The high surface area provides high energy density, whereas a high electrical conductivity is essential for high rate capacity. However, carbon nanotubes and graphene have relatively high cost. In contrast, electrospun carbon nanofibers exhibit a favorable specific surface area, low cost, and easy processing, and are potential electrode materials for EDLCs. Substantial effort has been expended to improve the electrochemical performance of electrospun carbon nanofibers through modifying the structure of the nanofibers by, for example, preparing interconnected,¹⁰ hollow,¹¹ and porous carbon nanofibers.^12,13 When used as supercapacitor electrodes, interconnected carbon nanofibers increase charge-transfer efficiency. Moreover, the porous and hollow carbon nanofibers facilitate the diffusion of electrolyte ions into the fibers and provide a favorable contact between the electrode and electrolyte. These effects are essential for developing supercapacitors that demonstrate satisfactory performance.

Instead of manipulating the structure of carbon nanofibers, in this study, we tailored nanofibers through surface modification using oxygen plasma. Because the pristine carbon surface is hydrophobic, the high resistance between the fiber surface and aqueous electrolyte hinders the accessibility of the electrolyte ions into the fibers. Through oxygen-plasma treatment, we found a substantial increase in the wettability of the carbon nanofibers, which allowed the aqueous electrolyte ions to easily transfer to the fiber surface. Consequently, the specific capacitance of the modified fibers was enhanced remarkably compared with that of the nontreated fibers.

Polyacrylonitrile (PAN) with a molecular weight of 150 000 g mol⁻¹ was purchased from Sigma-Aldrich and N,N-dimethylformamide, as a solvent, was provided by J. T. Baker. The PAN solution with a concentration of 7 wt% was stirred continuously for 24 h at room temperature. Electrospinning was conducted using a high-voltage power supply (You-Shang Technical Corporation), operated at an applied voltage of 15 kV. After electrospinning, pyrolysis was conducted on the electrospun fibers at 250 °C for 4 h in air and at 800 °C for 1 h in nitrogen. Oxygen-plasma treatment was conducted on the carbon nanofibers using a PDC-001 plasma cleaner (Harrick Plasma). The plasma power was 30 W, and the oxygen flow rate was 50 mL min⁻¹. The resulting carbon nanofibers were designated as CNF, the nontreated carbon nanofiber, and CNF-3, CNF-6, and CNF-9 for the carbon nanofibers treated by oxygen plasma for 3, 6, and 9 min, respectively.

The morphological features of the carbon nanofibers were observed using a Hitachi (SU8010) scanning electron microscope (SEM). X-ray photoelectron spectroscopy (XPS, PHI 5000 VersaProbe) was used to determine the oxygen content of the carbon nanofibers in the chemical bonds of the oxygenated groups. Contact-angle measurements of the carbon nanofibers were performed using a sessile-drop method with an FTA 1000B system (First Ten Angstroms). The nitrogen sorption isotherms were conducted at 77 K (ASAP-2020, Micromeritics). The specific surface area and pore size distribution of the carbon nanofibers were evaluated using the Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda (BJH) methods, respectively. The electrochemical performance of the carbon nanofibers was determined using a conventional Teflon electrochemistry cell with a three-electrode cell and potentiostat (CHI 627D, CH Instruments). A Ag/AgCl electrode was used as the reference electrode in 2 M KOH, and a platinum foil was used as the counter electrode. The carbon-nanofiber mat was placed on a nickel foil, which was used as the binder-free working electrode. The cyclic voltammetry (CV) curves were collected with a potential window from 0 to −1.0 V at scan rates ranging from 2 to 200 mV s⁻¹. Galvanostatic charge–discharge measurements were conducted at a current density of 1 A g⁻¹. The electrochemical impedance spectroscopy (EIS) measurements were carried out in a frequency range from 0.01 to 10⁵ Hz using an electrochemical impedance analyzer (CHI 6273A, CH Instruments). The cycling stability measurement consisted of 2000 cycles of CV tests at a constant scan rate of 50 mV s⁻¹.

Fig. 1 shows the morphology of the plasma-treated carbon nanofibers. The carbon nanofibers derived from the pristine PAN were bead-free, straight, and randomly arranged (Fig. 1(a)). Plasma treatment reduced the fiber diameter from 203 ± 64 nm for the CNF to 198 ± 54 nm, 151 ± 43 nm, and 141 ± 35 nm for CNF-3, CNF-6, and CNF-9, respectively. In addition, plasma treatment increased the surface roughness and caused a fracture, particularly for CNF-9 (Fig. 1(d)). The change of fiber morphology was attributed to the etching effect caused by an ion bombardment that removed surface atoms and eventually caused a fracture of the fibers. The XRD pattern of the electrospun carbon nanofibers shows two broad reflections at 2θ values of approximately 25° and 45°, which correspond to the (002) graphitic layers and the (100) turbostratic carbon plane, respectively (Fig. S1†). Carbon nanofibers treated by oxygen plasma presented similar characteristic patterns without any remarkable peak shift, indicating that plasma treatment did not cause a structural rearrangement of the fibers.

Fig. 1 SEM images of the plasma-treated carbon nanofibers: (a) CNF, (b) CNF-3, (c) CNF-6, and (d) CNF-9; the insets are high-magnification images of the fibers.

Fig. 2 shows the O1s XPS spectra of the plasma-treated carbon nanofibers. The obtained peak was deconvoluted according to the binding energies of C=O at 531.6 eV, COH at 532.9 eV, COOH at 534.2 eV, C–O at 535.0 eV, and H₂O at 538.3 eV.¹⁴ Detailed information about the oxygen functional groups on the fiber surface is listed in Table 1. During plasma treatment, C–C and C=C bonds are broken and free radicals are produced. The free radicals interact with oxygen atoms, forming oxygen-containing groups such as carbonyl and carboxyl groups. As shown in Table 1, most of the functional groups were carbonyl groups because the carbonyl groups exhibited lower interaction energy in the chemical-addition process.¹⁵ In addition, COH in CNF-9 was larger than that in CNF-6, whereas COOH in CNF-9 was considerably smaller than that in CNF-6. Because COOH exhibited a lower interaction energy than COH;¹⁵ the formation of COOH was predominant. However, prolonging the reaction time resulted in the a decomposition of COOH, producing COH.¹⁶ Therefore, the COOH content increased in a short reaction time, but decreased over a long reaction time. This in turn led to an increase in COH when the reaction time increased. A similar change of the COOH content with plasma treatment time was also obtained with the plasma-treated carbon nanotubes.¹⁵ The O/C ratio increased from 2.9% for CNF to 15.9%, 17.5%, and 17.6% for CNF-3, CNF-6, and CNF-9, respectively. In contrast, the N/C ratio decreased considerably to 2.0% for CNF-9. The increasing O/C ratio suggested that oxygen plasma enables the introduction of oxygen functional groups on the fiber surface, which can increase surface polarity. The surface functional groups of the plasma-treated carbon nanofibers were also examined with Fourier transform infrared spectroscopy (FTIR). As shown in Fig. S2,† an increase in peak intensities at 1712 cm⁻¹ (C=O stretching) and 1259 cm⁻¹ (C–O stretching) after plasma treatment could be seen, confirming the effective producing oxygen groups under plasma oxidative environment. A water contact-angle measurement was performed on the carbon nanofibers to determine the change in the wettability of the fibers in relation to the oxygen-plasma treatment (Fig. 3). The contact angle of the nontreated fibers was approximately 129°, indicating the hydrophobic characteristics of the fibers. In contrast, the fibers became completely wet after oxygen-plasma treatment. This behaviour suggested that the oxygen-functionalized fibers became highly hydrophilic, which could be expected to reduce the resistance between the aqueous electrolyte and the fiber surface when the fibers are used as electrode materials.

Fig. 2 O1s XPS of the plasma-treated carbon nanofibers: (a) CNF, (b) CNF-3, (c) CNF-6, and (d) CNF-9.

Table 1 Relative percentages of the oxygenated groups on the plasma-treated fibers

Sample	-C=O (%)	–COH (%)	–COOH (%)	–C–O (%)	H2O (%)	O/C (%)	N/C (%)
CNF	20	77.1	—	2.9	—	6.0	13.6
CNF-3	70.6	6.35	0.75	20.1	2.2	15.9	12.0
CNF-6	65.1	8	16.6	0.55	9.75	17.5	13.4
CNF-9	69.3	16.5	1.3	12.9	—	17.6	2.0

---

Fig. 3 Contact angle of the carbon nanofibers: (a) CNF and (b) CNF-6.

We performed N₂ adsorption isotherms on the carbon nanofibers to determine their structural properties (Fig. S3†). The specific surface area calculated using the BET method and the porosity parameters analyzed using the BJH approach are summarized in Table 2. The specific surface area of the carbon nanofibers increased slightly with increasing reaction time, which was attributed to an ion bombardment, resulting in an increase in the surface roughness of the fibers. On the other hand, the mesopore volume and average pore width were considerably comparable, varying less than 10%. These results suggested that plasma treatment did not modify the microporous texture of the fibers, but mainly formed functional groups on the fiber surface.

Table 2 Structural properties of the plasma-treated carbon fibers^a

Sample	SA (m^2 g^−1)	Vt (cm^3 g^−1)	Vmicro (cm^3 g^−1)	Vmeso (cm^3 g^−1)	Dp (nm)
a SA: specific surface area. Vt: total pore volume. Vmeso: mesopore volume (1.7–300 nm). Vmicro: micropore volume (<1.7 nm), calculated by Vt − Vmeso. Dp: adsorption average pore width (nm).
CNF	247	0.167	0.085	0.082	2.7
CNF-3	265	0.184	0.105	0.079	2.9
CNF-6	274	0.181	0.099	0.082	2.7
CNF-9	301	0.197	0.112	0.085	2.6

---

The electrochemical performance of the carbon nanofibers was determined using CV, galvanostatic charge–discharge, EIS, and cycle life tests. As shown in Fig. 4(a), the CV curves of all of the carbon nanofibers exhibited nearly a rectangular shape, which correlated with the ideal electrical double-layer capacitor, suggesting that the plasma-treated carbon nanofibers were suitable for use as EDLC electrode materials. The nontreated carbon nanofibers exhibited a specific capacitance of 167.6 F g⁻¹ at a scan rate of 2 mV s⁻¹. After plasma treatment, the specific capacitance of the carbon nanofibers increased to 185.0, 377.0, and 276.6 F g⁻¹ for CNF-3, CNF-6, and CNF-9, respectively. Regarding the structural parameters that could influence the electrochemical performance of the fibers, the pore size effect could be ignored because all the fibers exhibited nearly the same mesopore volume and pore size. Hulicova-Jurcakova et al. suggested that the specific capacitance of activated carbon increased linearly with the oxygen and nitrogen contents on the carbon surface.¹⁷ In this study, the sum of the oxygen and nitrogen contents increased for CNF-3 and CNF-6, but the total ratio of the oxygen and nitrogen for CNF-9 was identical to that of CNF. The variation of the oxygen and nitrogen contents with respect to the oxygen-plasma treatment was not consistent with the specific capacitance of the fibers, indicating that the pseudocapacitance effect contributed by the oxygen- and nitrogen-containing functional groups was not the major effect on the specific capacitance of the plasma-treated carbon nanofibers. The increased specific surface area of the plasma-treated carbon nanofibers provided the easy accessibility of electrolytes into the fibers and had the potential to improve the electrochemical performance. However, the plasma-treated carbon nanofibers exhibited only a 7–22% increase in the specific surface area when compared to the nontreated carbon nanofibers, suggesting that the specific surface area played a minor role on the enhancement of the electrochemical performance of the plasma-treated fibers. Therefore, the increased specific capacitance was mainly attributed to the increased fiber hydrophilicity, which facilitated favorable contact between the fiber surface and electrolytes, increasing the ion transport. Among these samples, CNF-6 exhibited the highest specific capacitance. A prolonged reaction time caused a loss of specific capacitance because of the occurrence of fractured fibers, causing a reduction in electron transport and an increase in resistance. In addition, the presence of saturated functional groups on the fiber surface could cause the electrical conductivity of the electrode to deteriorate.¹⁴

Fig. 4 Electrochemical performance of the plasma-treated carbon nanofibers: (a) CV at a scan rate of 2 mV s⁻¹, (b) the specific capacitance at different scan rates, (c) galvanostatic charge–discharge curves at a current density of 1 A g⁻¹, (d) Nyquist plots, and (e) cyclic performance at a scan rate of 50 mV s⁻¹.

Fig. 4(b) shows the specific capacitance of the carbon nanofibers measured at different scan rates. The specific capacitance increased as the scan rate decreased. For example, the specific capacitance of CNF increased from 44.5 to 167.6 F g⁻¹ when the scan rate decreased from 200 to 2 mV s⁻¹. CNF-6 exhibited the highest specific capacitance among all the other samples with a specific capacitance ranging between 78.0 and 377.0 F g⁻¹. In addition to CNF-6, the specific capacitance of the plasma-treated carbon nanofibers was consistently higher than that of the nontreated carbon nanofibers at each scan rate. The galvanostatic charge–discharge profiles of the carbon nanofibers are presented in Fig. 4(c). The IR drop, which indicates internal electrode resistance, was apparently smaller for all the plasma-treated carbon nanofibers than that for the nontreated fibers. The reduced internal resistance of the plasma-treated fibers was consistent with the increase in fiber polarity, which provides a favorable contact between the fibers and aqueous electrolyte. Fig. 4(d) shows the EIS curves in the Nyquist plots for the carbon nanofibers. All the spectra were composed of a semicircle in the high-frequency region and a linear portion in the low-frequency region. At very high frequencies, the intercept of the semicircle at the real axis is the sum of the intrinsic resistance of the electrode, the resistance of electrolyte, and the contact resistance at the electrode/current collector interface. This value was almost identical for all the fibers. A major difference was the appearance of a semicircle at lower frequencies, which corresponded to charge-transfer resistance. Charge-transfer resistance is related to the electroactive surface area of the electrode materials. Similar to the IR drop, the plasma-treated carbon nanofibers exhibited smaller charge-transfer resistances than the nontreated fibers, and the carbon nanofibers treated by plasma for 6 min had the smallest charge-transfer resistance. The cycle lives of the oxygen-modified carbon nanofibers are shown in Fig. 4(e). The specific capacitance retention of CNF and CNF-3 remained nearly 100% after 2000 cycles, indicating that the electrospun carbon nanofibers exhibited excellent cycle life and electrochemical reversibility. In contrast, the capacitance retentions of CNF-6 and CNF-9 increased to approximately 120%, which was presumably attributed to the improved wetting of the pores after charging and discharging for several times.¹⁸ Our results indicate that the oxygen-plasma treatment on the electrospun carbon nanofibers could considerably improve the electrochemical performance of the fibers. To the best of our knowledge and following a literature review, the specific capacitance of CNF-6 is superior to that of most carbonaceous materials as EDLC electrodes. The technique of employing oxygen plasma for surface modification could also be used to modify carbonaceous materials or composite nanofibers into supercapacitor electrodes.

Using oxygen plasma, we performed surface modification on electrospun carbon nanofibers for their use as supercapacitor electrodes. After the plasma treatment, the fibers were grafted with various oxygen functional groups, with the carbonyl groups being predominant. The functional groups simultaneously resulted in the change of fiber wettability from hydrophobic to hydrophilic. The improved wettability provided for a more favorable contact between the electrodes and aqueous electrolyte, which facilitated the diffusion of electrolyte ions into the nanofibers. The electrochemical performance showed that the specific capacitance of the fibers was in close relation to the plasma treatment. The carbon nanofibers treated by oxygen plasma for 6 min exhibited a specific capacitance of 377.0 F g⁻¹ at a scan rate of 2 mV s⁻¹, which was more than twice that of the nontreated carbon nanofibers. The oxygen-functionalized carbon nanofibers also exhibited approximately 120% capacitance retention after 2000 cycles, indicating excellent cycle stability.

Acknowledgements

This work is financially supported by the Ministry of Science and Technology in Taiwan under Grant no. 102-2221-E-006-018-MY3 and the Technology Development Program for Academia no. 103-EC-17-A-08-S1-204 by Ministry of Economic Affairs in Taiwan.

Notes and references

1. Y. Zhu, S. Murali, M. D. Stoller, K. J. Ganesh, W. Cai, P. J. Ferreira, A. Pirkle, R. M. Wallace, K. A. Cychosz, M. Thommes, D. Su, E. A. Stach and R. S. Ruoff, Science, 2011, 332, 1537.
2. E. Frackowiak, K. Metenier, V. Bertagna and F. Beguin, Appl. Phys. Lett., 2000, 77, 2421.
3. D. N. Futaba, K. Hata, T. Yamada, T. Hiraoka, Y. Hayamizu, Y. Kakudate, O. Tanaike, H. Hatori, M. Yumura and S. Iijima, Nat. Mater., 2006, 5, 987.
4. J. Liu, L. Zhang, H. B. Wu, J. Lin, Z. Shen and X. W. Lou, Energy Environ. Sci., 2014, 7, 3709.
5. C. Liu, Z. Yu, D. Neff, A. Zhamu and B. Z. Jang, N. Lett., 2010, 10, 4863.
6. J. Liu, J. Sun and L. Gao, J. Phys. Chem. C, 2010, 114, 19614.
7. M. M. Hantel, T. Kaspar, R. Nesper, A. Wokaun and R. Kötz, Electrochem. Commun., 2011, 13, 90.
8. D. H. Shin, J. S. Lee, J. Jun and J. Jang, J. Mater. Chem. A, 2014, 2, 3364.
9. D. Bhattacharjya, M. S. Kim, T. S. Bae and J. S. Yu, J. Power Sources, 2013, 244, 799.
10. Y. H. Hsu, C. C. Lai, C. L. Ho and C. T. Lo, Electrochim. Acta, 2014, 127, 369.
11. A. V. Bazilevsky, A. L. Yarin and C. M. Megaridis, Langmuir, 2007, 23, 2311.
12. T. Chau and V. Kalra, J. Power Sources, 2013, 235, 289.
13. B. H. Kim, K. S. Yang, Y. A. Kim, Y. J. Kim, B. An and K. Oshida, J. Power Sources, 2011, 196, 10496.
14. G. Lota, J. Tyczkowski, R. Kapica, K. Lota and E. Frackowiak, J. Power Sources, 2010, 195, 7535.
15. A. Felten, C. Bittencourt, J. J. Pireaux, G. V. Lier and J. C. Charlier, J. Appl. Phys., 2005, 98, 74308.
16. X. Zhang, Y. Huang, T. Wang and L. Liu, Composites, Part A, 2007, 38, 936.
17. D. Hulicova-Jurcakova, M. Seredych, G. Q. Lu and T. J. Bandosz, Adv. Funct. Mater., 2009, 19, 438.
18. C. W. Huang, C. T. Hsieh, P. L. Kuo and H. Teng, J. Mater. Chem., 2012, 22, 7314.

---

Footnote

† Electronic supplementary information (ESI) available: X-ray diffraction patterns, Fourier transform infrared spectra, and N₂ adsorption isotherms of plasma-treated carbon nanofibers. See DOI: 10.1039/c5ra04284d

---