Hierarchical porous nitrogen doping activated carbon with high performance for supercapacitor electrodes

Abstract

Hierarchical porous nitrogen doping activated carbon materials were designed and prepared by carbonization of electrospun composite carbon nanofibers and subsequent chemical activation. The porous carbons were activated by potassium hydroxide (KOH) and the optimal activation weight ratio for carbons was investigated for this process. It is found that activation weight ratios played an important role on the porous structures and capacitive properties. The morphology, pore structure and surface physicochemical properties of the carbon samples were characterized by scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS) and nitrogen adsorption/desorption isotherm analysis. The sample with the activation weight ratio of 4 possessed a large 2217 m² g⁻¹ specific surface area and held a high 1.443 cm³ g⁻¹ volume value as well as a hierarchical porous structure. A symmetrical supercapacitor was fabricated using 6 mol L⁻¹ KOH as the electrolyte and the electrochemical properties were evaluated by cyclic voltammetry (CV), electrochemical impedance spectroscopy (EIS), galvanostatic charge–discharge (GCD) and life cycle measurements. The results indicate that the sample with an activation weight ratio of 4 obtained high specific capacitances of 255, 238, 230, 215 and 198 F g⁻¹ at the current densities of 0.2, 0.5, 1.0, 5.0 and 10 A g⁻¹, respectively. Furthermore, the sample materials presented good cycling behavior with respect to the specific capacitance value, which hardly decreased during 8000 cycles. This is considerably promising for this type of porous material to be far-reaching applied in the field of electrode materials for supercapacitors.

---

1 Introduction

Casting about for reliable, sustainable and efficient energy source forms free of pollution emission is a prerequisite resolvent to cope with the increasing energy shortage and concomitant environmental pressure.^1,2 Served as power storage and supply devices on account of their higher power density, faster charging–discharging rate, longer cycling life and better operation safety, supercapacitors have been deemed as promising high power energy devices.^3–5 The charge storage mechanism of supercapacitors is via electrical double layer of electrodes (EDLCs) with a high specific surface area and/or based on the pseudocapacitance associated with fast surface redox reactions.^6,7

It is known that electrode material is the leading factor that influences supercapacitor performance. To date, carbon electrode materials were regarded as the most prospective materials for future capacitors with ascendant performances due to their relatively low price, good electric conductivity and steady physicochemical properties. Currently, the highly developed surface area and porosity are integral for carbon electrodes to obtain high specific capacitance.^8–11 Activated carbon with very high surface area and pore volume has been diffusely applied as electrode materials for supercapacitors.^12–14 Excellent activated carbon has been gained by selecting appropriate precursors and activation methods; however, the pore structures were uncontrollable and the pores were chiefly micropores (<2 nm), which are not beneficial for quick electrolyte ion diffusion. The more proper porous structure for supercapacitor applications aroused our interest to search for other means. Latterly, ordered mesoporous carbon with manageable pore structure has been recognized in numerous fields. Some literature had demonstrated the electrochemical capacitive properties of mesoporous carbon as a supercapacitor electrode material.^15–18 It is well known that nitrogen functionality on a carbon surface is a good method to augment the capacitance and stability due to the pseudo faradaic effect and surface chemical properties.^19,20 In conclusion, increasing the number of appropriate pores and nitrogen doping modification on the base of activated carbons are the best choices to achieve an optimal capacitance.

After determining the optimal solution ratio and spinning conditions, we were able to prepare nanofibers easily and continuously via an electrospinning technique, which operates simply, at low cost and allows for governing the nanofibers morphology. Moreover, fiber diameter ranged between tens of nanometers and several micrometers. Thus, the electrospinning technique appears to have broad application prospects. The template method is a bargain means to prepare porous materials with effective control over the pore size distribution. In our experiments, we made some improvements based on the previous studies. Polyacrylonitrile (PAN) acted as a carbon precursor and polymethylmethacrylate (PMMA) served as the template with thermoplastic polyurethane (TPU) as the nitrogen doping material.²¹ Because the influence of nitrogen doping has been discussed in the previous published work, a detailed explanation will not be given in the following discussion. Furthermore, we adopted an available chemical activation process to obtain porous activated carbon. This chemical activation can readily proceed by KOH activation accompanied with carbonization at a high temperature. We report the electrochemical performance of carbon materials activated with difference proportions of KOH/CNF.

2 Experimental section

2.1 Materials

PAN (M_w = 150 000) and PMMA (M_w = 15 000) were obtained from Sigma-Aldrich Chemical Company Inc. (USA). N,N-Dimethylformamide (DMF) was supplied by Xi'an Xilong Chemical Reagent Factory. TPU was acquired from Zhejiang Lingxin polyurethane Company. All chemicals were used as received without further purification.

2.2 Preparation of hierarchical porous nitrogen doping activated carbon

7 wt% PAN solutions with DMF solvent DMF were prepared. The 20 wt% of PMMA template (relative to mass PAN) and 10 wt% mass fraction of TPU (relative to the PAN mass) were added into the PAN solutions synchronously. The hybrid system was stirred with magnetic stirring and heating at 70 °C for 8 h until dissolved completely. Then, the uniform copolymers solutions were electrospun into the nanofibers by electrospinning. A variable high voltage power supply (DW-P353-3 AC) was applied to offer voltages of 30 kV, a 15 cm tip-to-collector distance and a 1 mL h⁻¹ electrospinning flow rate. The as-spun fibers were calcined in an electric heat treating furnace (Changcheng Furnaces Factory). The nanofibers were stabilized in an air atmosphere at 280 °C for 3 h (at a heating rate of 2 °C min⁻¹). Subsequently, PMMA that served as a template decomposed thoroughly when the mixture was heated to 400 °C for 1 h in an air atmosphere (at a heating rate of 1 °C min⁻¹). The fibers followed by carbonization at 900 °C for 1 h (at a heating rate of 5 °C min⁻¹) under an inert argon atmosphere to yield carbonized materials. Subsequently, carbon nanomaterials acquired were immersed in diluted hydrochloric acid for several hours then washed with deionized water until the pH value of the water–carbon compounds was about 7; this was named CNF. Moreover, the CNF powder was impregnated in KOH solutions at room temperature with KOH/CNF weight ratios of 2, 3, 4 and 5, respectively.²² Furthermore, the blended slurries were pretreated at 110 °C for 24 h in a vacuum drying oven and the dried samples were shelved in tubular resistance furnace once again, as done for activation, at a 5 °C min⁻¹ heating rate from room temperature to 900 °C for 1 h in an Ar atmosphere. Ultimately, the activated materials were infused in diluted hydrochloric acid and washed with deionized water until neutral. In the light of the mixing weight ratios (KOH/CNF), the samples were denominated CNF (0/1), CNFA2 (2/1), CNFA4 (3/1), CNFA4 (4/1) and CNFA5 (5/1).

2.3 Characterization

The surface morphology of the spinning nanofibers was examined by a scanning electron microscopy (SEM, JEOL, JSM-6360). The porous structure of the carbon nanomaterials was observed with a transmission electron microscope (TEM, JEOL, JEM-2010 F). The specific surface area and pore structure of the carbon materials were determined by N₂ adsorption–desorption isotherms on a sorptiometer (Quantachrome Corp. Nova-2000e gas sorption analyzer) at 77 K. All the samples were out-gassed at 200 °C. The specific surface area and micropore volume of the activated samples were evaluated with BET and the t-plot method, respectively. The pore size distributions of these samples were calculated by the Barrett–Joyner–Halenda (BJH) method. The total pore volume (V_tot) was estimated from a single point adsorption at a relative pressure P/P₀ = 0.99. The crystallographic structures of the prepared materials were discerned by a power X-ray diffraction (XRD) system (D/max 2500) outfitted with Cu Kα radiation (λ = 0.15418 nm) and the chemical state of the surface was characterized by X-ray photoelectron spectroscopy (XPS) on a K-Alpha 1063 spectrometer with an Al Kα monochromatic source (12 kV, 6 mA).

2.4 Electrochemical performance measurements

The electrodes of the supercapacitor cells were fabricated with two symmetric electrodes separated by a filter paper, which used Ni foil as the current collector along with the carbon materials, polyvinylidenefluoride (PVDF) and acetylene black at a weight ratio of 8 : 1 : 1. By equipping coin-type cells the electrochemical performance of CNF, CNFA2, CNFA3, CNFA4 and CNFA5 samples were investigated under a two-electrode system with 6 mol L⁻¹ KOH as electrolyte. The capacitive behaviors were detected by means of potentiodynamic cyclic voltammetry (CV) (1, 10, 50, and 100 mV s⁻¹) and electrochemical impedance spectroscopy (EIS) (10⁻² to 10⁵ Hz, amplitude of signal 5 mV) using a CHI660D electrochemical testing station (Chenhua Instruments Co. Ltd. Shanghai). The galvanostatic charge–discharge curves were obtained in the potential range from 0 to 1.0 V at different current densities (0.2, 0.5, 1.0, 5.0, and 10 A g⁻¹) and multicycle charge/discharge data (in 8000 cycles) were acquired on a Neware battery test system (BTS-XWJ-6.9.27s) at a current density of 1 A g⁻¹.

3 Results and discussion

3.1 Morphology characterization

The typical SEM images of electrostatically spun nanofibers before carbonization are provided in Fig. 1a and b at low magnification. It can be observed that all the nanofibers were consecutive and formed smooth network structures with ultrafine cylindrical morphologies. This explains why the macromolecular polymers were well dissolved. This means that TPU and PAN blend solution have excellent electrospin properties. Moreover, Fig. 2a and c shows the TEM image for CNF and CNFA4 samples, respectively. Comparing the two figures, we can observe that there were few pore structures for the CNF sample. Nevertheless, the CNFA4 sample owned distinct and abundant pores. Fig. 2b and d present a stochastic distributed porous morphology for CNF and CNFA4 samples with high-resolution detection as well as the pore structure size for carbon nanofibers became overtly larger after activation. Based on these facts, three factors contributed to the observations. In the first place, the PMMA acted as a template that decomposed rapidly when the copolymers were heated to 400 °C so as to obtain a part of the porous structure. Furthermore, the molecular structure for TPU included soft and hard segments and the decomposition temperature for the soft segment was about 200 °C; however, for the hard segment it was approximately 350 °C.²³ Therefore, the TPU in the form of compounds has already partly decomposed during heating to 400 °C probably, which produced the fractional porous structure. Moreover, the KOH etched the carbon materials and immensely aggrandized porous structure in the chemical activation process.

Fig. 1 (a) and (b) Low-resolution SEM images of the electrospun nanofibers before carbonization.

Fig. 2 Low-magnification TEM images of (a) CNF and (c) CNFA4 and higher magnification image of (b) CNF and (d) CNFA4.

3.2 XRD and XPS analysis

The wide angle powder X-ray diffraction (XRD) pattern of CNF and CNFA4 samples are shown in Fig. 3a. There are two evident diffraction peaks at 2θ = 26° and 43° respectively, which correspond to diffraction from the (002) and (100) planes of graphite.²⁴ Contrasting with CNF, CNFA4 presented an obvious reduced intensity and the (002) peak was obviously broadened, which confirms that the crystalline structures were disordered in the chemical etching process. This was due to the intercalation of metal K into carbon pseudographitic layers and disruption of the crystallite.²⁵ It also indirectly proves the sample materials own high pore density. The results agreed with high resolution TEM analysis. From the CNF and CNFA4 samples XPS C 1s spectrum pattern displayed in Fig. 3b, it could be observed that the carbon content increased after activation. The inset for both samples shows the presence of nitrogen functional groups (pyridine and pyrrole) between 398 and 402 eV. Nevertheless, the nitrogen content decreased from 20.26 at% (before activation) to 13.87 at% (after activation), which shows that the KOH etching process removed nitrogen from the CNFA4 carbon materials. The relevant relative atomic concentrations for CNF and CNFA4 samples are shown in Table 1.

Fig. 3 Characterization of CNF and CNFA4 material (a) power XRD pattern; (b) XPS C 1s spectra with the N 1s region in the inset.

Table 1 Element composition by XPS of the carbon samples

Samples	Relative atomic concentration
	Cxps	Nxps	Oxps	C/N
CNF	71.13	20.26	8.61	3.51
CNFA4	84.43	13.87	1.7	6.09

---

3.3 Pore structural properties

As displayed in Fig. 4a, the N₂ adsorption–desorption isotherms of all samples are combined type I/IV isotherms showed a distinct capillary condensation step (hysteresis loop) at medium relative pressure, which indicated that mesopores were present. Furthermore, the activated samples isotherms comprised a rapid and obvious adsorption at relative pressures below 0.1, manifesting the coexistence of segmental micropores after activation. In addition, a slight increase in adsorption occurred at relative pressures near 1, which indicates that a limited fraction of macropores was also included in the samples. Moreover, homologous pore size distributions of carbon materials are exhibited in Fig. 4b. Based on the inset, contrasting all samples found the CNF sample material only contained mesopores, whereas the carbon materials after activation included both micropores and some mesopores and the total pore volume show a tendency to increase. The changes in pore size distribution were consistent with new pore generation and original pore expansion that resulted from the reaction with hydroxide/carbon, which continued up to micropore development and aggrandizement of the partial micropores. The pore structural parameters according to the BET equation and the t-plot method from the nitrogen adsorption isotherms at 77 K of all samples are included in Table 2. The CNF sample before the activation merely exhibited a specific surface area of 443.106 m² g⁻¹ and the total pore volume of 0.238 cm³ g⁻¹. Nevertheless, after treatment with KOH etching, the specific surface area was extended and the same to total pore volume, besides the CNFA4 material, obtained a maximum of 1.443 cm³ g⁻¹.

Fig. 4 (a) N₂ adsorption/desorption isotherms at 77 K for all carbon samples and (b) relevant pore size distributions (inset).

Table 2 Porosity parameters and the resistances of the electrolyte solution of all the carbon materials

Samples	SBET (m^2 g^−1)	Vmicro (cm^3 g^−1)	Vmeso (cm^3 g^−1)	VT (cm^3 g^−1)	Average pore diameter (nm)	Rs (Ω)
CNF	443.106	0.174	0.042	0.238	2.145	0.46
CNFA2	1844.401	0.481	0.409	0.988	2.144	0.41
CNFA3	1204.138	0.242	0.173	0.514	3.021	0.47
CNFA4	2216.755	0.386	0.831	1.443	2.605	0.44
CNFA5	2324.359	0.566	0.412	1.215	2.092	0.42

---

3.4 Electrochemical performance

The electrochemical behavior was examined by the cyclic voltammetry (CV) technique performed in the potential range of 0–1.0 V. With a 10 mV s⁻¹ scan rate, as shown in Fig. 5a, all the samples exhibited a nearly ideal rectangular shape, showing mirror images of the lines with respect to the zero-current line, which is characteristic of an electric double-layer capacitor (EDLC). In addition, all the CV curves possessed no evident oxidation/reduction peaks, demonstrating pseudo-capacitance forming very little contribution from the redox reaction. Due to the TPU decomposing partly in the process of high temperature carbonization, the nitrogen pseudo-capacitance effect was very weak. Other similar reports exist in the literature.²⁶ Moreover, according to the integral areas encircled by the CV curves, it could be qualitatively distinguished that the CNFA4 sample had the maximal specific capacitance compared with the four other samples, which was consistent with the physical characteristics. To detect the capacitance behaviors of the CNFA4 electrode material sample, the CV curves at different scan rates were acquired. The shapes of CV curves (show at Fig. 5b) did not change remarkably when the scan rate was increased from 1 to 100 mV s⁻¹, reflecting fast charge transfer within the electrode materials, which is rooted in the carbons having suitable pore size and seemly better ion transport pathway. Furthermore, recent studies have indicated that the heteroatom species played multiple roles in the electrochemical performance. The conductivity of the carbon can be enhanced by doping nitrogen into the carbon skeleton.²⁷ The hydrophilic nitrogen, oxygen or sulfur species facilitated wetting the carbon pore surface to the aqueous electrolyte. Thus, the addition of TPU contributed to good electrochemical performance of the carbon materials. The EIS analysis is a principal method to examine the fundamental behavior of electrode materials for supercapacitors. The Nyquist plots for the frequency ranging from 10⁻² to 10⁵ Hz with 5 mV AC amplitude for all samples are included in Fig. 6. All the Nyquist plots were similar in form, consisting of one small semicircle at high frequency and an apparent straight line in the low frequency region. The intercept of the semicircle with the real axis denotes the resistance of the electrolyte solution (R_s) in the high frequency range and the R_s values of the samples were estimated from the X-intercept of the Nyquist plots, which are also shown in Table 2. The impedance spectra analysis provided R_s values for the CNFA4 sample as low as 0.44 Ω. In addition, the diameter of the semicircle loop in the high frequency range represents the charge-transfer resistance (R_ct), which mainly includes the contact resistance between surface of electrode materials and current collector and the internal resistance for electrode materials. It can be observed from the inset, the semicircle for CNFA4 is smaller than for CNFA2, CNFA3 and CNFA5, indicating that the CNFA4 sample possessed excellent electrical conductivity and electrochemical performance. The nearly straight line for CNFA4 in the low frequency region signifies supercapacitor behavior. Thus, the CV curve analysis indicates suitability for high performance supercapacitor electrode materials.

Fig. 5 (a) Cyclic voltammograms of all samples at a 10 mV s⁻¹ scan rate; (b) cyclic voltammograms of the CNFA4 sample at different scan rates.

Fig. 6 Nyquist plot of all samples in the frequency range of 10⁻² to 10⁵ Hz.

To research the capacitive performance of the CNFA4 sample, the galvanostatic charge/discharge curves at different current densities in 6 mol L⁻¹ KOH were acquired and are depicted in Fig. 7a. It is clear to us that the charge/discharge pattern of the CNFA4 sample exhibited almost the isosceles triangle shaped curves, which indicates the assembled capacitor possessed stable and reversible electrochemical behavior. The rate capacitances of the CNFA4 sample derived from the discharging curves at different current densities were compared and are demonstrated in Fig. 7b. It can be observed that the specific capacitances decrease clearly at low current densities and this suggests that about 77% of the specific capacitances at 0.2 A g⁻¹ were retained at a current density of 10 A g⁻¹, which illustrates a good rated capability and electrochemical properties. The specific capacitances were calculated from the discharge curves at different current densities according to the equation as follows: C = 2iΔt/mΔV, where i is the constant discharge current density, Δt is the discharging time, m is the mass of active materials on a single electrode and ΔV is the potential difference. According to the abovementioned equation, the specific capacitance was 255, 238, 230, 215 and 198 F g⁻¹ at 0.2, 0.5, 1.0, 5.0 and 10 A g⁻¹ current densities, respectively. The specific capacitance of the carbon electrodes decreased with increased current density, which is very ordinary for supercapacitors and is caused by the diffusion limitation of electrolyte ions in the micropores.²⁸ Moreover, another distinct feature of the galvanostatic curves was a very small voltage drop, which is in accordance with a low equivalent series resistance of the capacitor cells. This phenomenon correlates well with the EIS test results. Long cycle life is a significant parameter for supercapacitor electrode materials, the specific capacitances of all samples as a function of cycle number are exhibited in Fig. 7c. Very stable capacitances with less than a 5% decrease after a 1 A g⁻¹ applied current density for 8000 cycles of charging and discharging, further demonstrated the long-term electrochemical stability.

Fig. 7 (a) Galvanostatic charge/discharge curves of the CNFA4 sample at different current densities; (b) specific capacitances of the CNFA4 sample at various rates and (c) the specific capacitances of all samples for 8000 cycles at 1 A g⁻¹.

4 Conclusions

The hierarchical porous carbons were successfully prepared through combining electrostatic spinning work with the nanofibers carbonization and KOH activation. Activation via different weight ratios of KOH/CNF could tremendously affect the electrochemical capacitance of activated carbon. Compared with the other samples, the CNFA4 carbon material had a large 2217 m² g⁻¹ specific surface area and high 1.443 cm³ g⁻¹ total pore volume (mainly for micropores and mesopores) as well as high specific capacitance, good conductivity, low resistance and excellent cycling stability. Thus, it can be generalized that the optimal activation weight ratio of KOH/CNF for the prepared activated carbon is affirmed to be 4. The outstanding electrochemical properties demonstrated that hierarchical porous nitrogen doped activated carbons can be very promising candidates as electrode materials for high performance supercapacitors.

Acknowledgements

The workers gratefully appreciate the financial support from the Youth Project of National Nature Science Foundation of China (Grant No. 51103124 and No. 51203131) and the Hunan Province Universities Innovation Platform of Open Fund Project (11K067).

References

1. J. R. Miller and P. Simon, Science, 2008, 321, 651.
2. H. L. Wang and H. J. Dai, Chem. Soc. Rev., 2013, 42, 3088.
3. C. Liu, F. Li, L. P. Ma and H. M. Cheng, Adv. Mater., 2010, 22, E28.
4. A. Izadi-Najafabadi, S. Yasuda, K. Kobashi, T. Yamada, D. N. Futaba, H. Hatori, M. Yumura, S. Iijima and K. Hata, Adv. Mater., 2010, 22, E235.
5. G. Lota, T. A. Centeno, E. Frackowiak and F. Stoeckli, Electrochim. Acta, 2008, 53, 2210–2216.
6. P. Ahuja, V. Sahu, S. K. Ujjain, R. K. Sharma and G. Singh, Electrochim. Acta, 2014, 146, 429.
7. D. S. Patil, S. A. Pawar, R. S. Devan, M. G. Gang, Y.-R. Ma, J. H. Kim and P. S. Patil, Electrochim. Acta, 2013, 105, 569.
8. E. Frackowiak, Phys. Chem. Chem. Phys., 2007, 9, 1774–1785.
9. H. Denisa, Y. Junya, S. Yasushi, H. Hiroaki and K. Masaya, Chem. Mater., 2005, 17, 1241–1247.
10. B. Z. Fang and L. Binder, J. Phys. Chem. B, 2006, 110, 7877–7882.
11. H. F. Li, H. A. Xi, S. M. Zhu, Z. Y. Wen and R. D. Wang, Microporous Mesoporous Mater., 2006, 96, 357–362.
12. D. Y. Qu, J. Power Sources, 2002, 109, 403–411.
13. O. Barbieri, M. Hahn, A. Herzog and R. Kötz, Carbon, 2005, 43, 1303–1310.
14. B. Kastening, M. Hahn, B. Rabanus, M. Heins and U. Z. Felde, Electrochim. Acta, 1997, 42, 2789–2799.
15. B. J. Lee, S. Yoon, S. M. Oh, C.-H. Shin and T. Hyeon, Adv. Mater., 2000, 12, 359–362.
16. H. Zhou, S. Zhu, M. Hibino and I. Honma, J. Power Sources, 2003, 122, 219–223.
17. A. B. Fuertes, F. Pico and J. M. Rojo, J. Power Sources, 2004, 133, 329–336.
18. W. Xing, S. Z. Qiao, R. G. Ding, F. Li, G. Q. Lu, Z. F. Yan and H. M. Cheng, Carbon, 2006, 44, 216–224.
19. D. Hulicova, J. Yamashita, Y. Soneda, H. Hatori and M. Kodama, Chem. Mater., 2005, 17, 1241–1247.
20. W. R. Li, D. H. Chen, Z. Li, Y. F. Shi, Y. Wan, G. Wang, Z. Y. Jiang and D. Y. Zhao, Carbon, 2007, 45, 1757–1763.
21. J. Zeng, Q. Cao, X. Y. Wang, B. Jing, X. X. Peng and X. L. Tang, J. Solid State Electrochem., 2015, 19, 1591–1597.
22. J. L. Chang, Z. Y. Gao, X. R. Wang, D. P. Wu, F. Xu, X. Wang, Y. M. Guo and K. Jiang, Electrochim. Acta, 2015, 157, 290–298.
23. P. V. Ambuken, H. A. Stretz, J. H. Koo, J. M. Messman and D. Wong, Polym. Degrad. Stab., 2014, 102, 160–169.
24. Y. Xia and R. Mokava, Adv. Mater., 2004, 16, 1553–1558.
25. S. H. Yoon, S. Lim, Y. Song, Y. Ota, W. Qiao, A. Tanaka and I. Mochida, Carbon, 2004, 42, 1723–1729.
26. G. Q. Wang, J. Zhang, S. Kuang, J. Zhou, W. Xing and S. P. Zhuo, Electrochim. Acta, 2015, 153, 273–279.
27. L. Zhao, N. Baccile, S. Gross, Y. J. Zhang, W. Wei, Y. Sun, M. Antonietti and M. M. Titirici, Carbon, 2010, 48, 3778–3787.
28. W. Xing, S. Z. Qiao, R. G. Ding, F. Li, G. Q. Lu, Z. F. Yan and H. M. Cheng, Carbon, 2006, 44, 216–224.

---