Nitrogen(N)-doped activated carbon materials with a narrow pore size distribution derived from coal liquefaction residues as low-cost and high-activity oxygen reduction catalysts in alkaline solution

Abstract

Coal liquefaction residues with a high content of nitrogen were used to prepare N-doped activated carbon as a catalyst of the oxygen reduction reaction (ORR). N-Doped activated carbon materials prepared at 900 °C (AC-900) exhibited a large specific surface area of 3130 m² g⁻¹ and a small average pore size of 1.91 nm. AC-900 also displayed more edges on the surface and higher N-doped contents of pyridinic-N and graphitic-N, which enhanced the conductivity and created more catalytic active sites. Electrochemical tests showed for AC-900 a half-peak current density of −2.11 mA cm⁻² at a potential of −0.17 V, which was only 63 mV lower than that of the commercial Pt/C (20%) catalyst. As much as 85.1% of the initial current density for AC-900 was maintained during the course of a stability test lasting 20 000 s compared with a value of 82.1% for the commercial Pt/C (20%) catalyst. Low-cost N-doped activated carbon should not only be an efficient non-precious catalyst for the ORR in alkaline fuel cells but also offer a porous carbon precursor for the further preparation of compound catalysts.

---

Introduction

Activated carbons are widely used for many areas of energy storage including fuel cells and metal-air batteries due to their advantages such as large specific surface areas, low cost, and easy doping.^1–3 Different pore structures of activated carbons have been shown to be caused by different activation processes and their pore structures should be optimized to enhance their catalysis of the oxygen reduction reaction (ORR).⁴ Micropores can usually electro-adsorb simple hydrated ions at a low concentration-dependent rate, and the pores with dimensions greater than 0.5 nm have recently been made electrochemically accessible^5,6 but contribute little to improving the use of active sites. In contrast, mesopores can provide a large electro-active surface area for the catalyst and facilitate the rapid transport of ions and charges, contributing to an increase in the number of active sites.^7–9 Ordered mesoporous carbons can show better catalytic performance by improving the mass transport of reactants.^10,11 Therefore, rationally controlling the pore structures of activated carbon is critical for improving the electrocatalytic performance for the ORR. In addition, N-doping is useful for improving the ORR activity of activated carbon. Since Dai et al. reported the high ORR activity of nitrogen-doped carbon nanotubes in 2009,¹² a series of N-doped carbon materials have been exploited, and have shown excellent performance compared with expensive Pt- or Pt alloy-based catalysts.^13–18

Direct coal liquefaction is a strategic technology to relieve shortages of oil and ensure energy security throughout the world. However, thirty percent of the all liquefied coal that has been produced has been treated as waste. Fortunately, high contents of nitrogen with different structures have been shown to be present in coal liquefaction residues, which is a precondition for the direct synthesis of N-doped carbon materials and facilitate the use of the coal liquefaction residues in ORR applications.

Herein, a simple method is proposed to synthesize N-doped microporous/mesoporous activated carbon, one that involves activating the coal liquefaction residues with KOH at different temperatures. The sample resulting from treatment at 900 °C, denoted as AC-900, was found to exhibit the optimal mean pore size, the largest specific surface area of the samples made, many exposed edge sites on the surface, and a high content of doped N, which enhance the catalytic activity for the ORR. In our electrochemical tests, AC-900 exhibited a half-peak current of −2.11 mA cm⁻² at −0.17 V, which was only 63 mV lower than that of the commercial Pt/C (20%) catalyst. The AC-900 also exhibited excellent stability for the ORR.

Experimental

Coal liquefaction residues (Shenhua Industry Co., Ltd. China) were crushed into a powder form by using a crushing machine (SQ2119B, Shuaijia Industry Co., Ltd. China). Subsequently, the obtained sample was mixed with KOH at a mass ratio of 1 : 6 and then heated at various temperatures (700 °C, 800 °C, and 900 °C) for 2 h in a tubular furnace at a heating rate of 5 °C min⁻¹ in an N₂ atmosphere with a flow rate of 100 mL min⁻¹. After becoming activated carbons, the samples were washed with 1 M HCl and distilled water several times, followed by drying for 12 h at 60 °C. The obtained samples were denoted as AC-700, AC-800, and AC-900, respectively.

A scanning electron microscope (SEM, HITACHI S-4800), X-ray diffraction (XRD, Rigaku D/Max 2500/PC), and a transmission electron microscope (TEM, JEM-2100HR) were used to characterize the structure and morphology of each of the samples. X-ray photoelectron spectroscopy (XPS, ESCALAB 250) was used to determine the compositions. Nitrogen sorption isotherms were acquired at 77 K with a Micromeritics Tristar 3000 analyzer (USA).

The electrochemical evaluation was carried out by using an electrochemical workstation (RST5200F, Zhengzhou RST Instrument Co., Ltd. China). Cyclic voltammetry (CV) and rotating-disk electrode (RDE) polarization curves were acquired at room temperature in a conventional three-electrode electrochemical system. A platinum wire (CHI115) and an Ag/AgCl (sat.) (CHI111) electrode were used as the counter electrode and reference electrode, respectively. A glassy carbon (GC) electrode (3 mm in diameter, 0.0707 cm², CHI104) and GC disk (4 mm in diameter, 0.1256 cm², Jiangsu Jiangfen Electroanalytical Instrument Co., Ltd. China) were used for the working electrodes to test the CV and RDE curves, respectively. Before taking the measurements, the GC electrodes were polished with gamma alumina powders (0.05 mm) until a mirror-like surface was obtained, and then washed with distilled water and dried in vacuum. Subsequently, an appropriate amount of the catalyst (600 μg cm⁻²) was administered onto the GC electrode followed by administering a drop of alcohol. After drying at 60 °C, a drop of Nafion solution (5 wt%, Dupont) was administered onto the catalyst layer to improve the adhesion on the GC surface. The CV curves were acquired using potentials ranging from −1 to 0 V vs. Ag/AgCl (sat.) in N₂-saturated and O₂-saturated 0.1 M KOH solution. RDE polarization curves were acquired using potentials ranging from −1 to 0.1 V vs. Ag/AgCl (sat.) with a scan rate of 10 mV s⁻¹ and rotation rates from 400 to 3600 rpm in an O₂-saturated 0.1 M KOH solution. Before taking the RDE and CV measurements, O₂ was bubbled into the cell for 30 min. In order to confirm the influence of the background current, CV and RDE curves were acquired in an N₂-saturated 0.1 M KOH solution. The current–time curves were acquired at a constant potential of −0.40 V for 20 000 s, during which time the O₂ was bubbled in continuously at a flow rate of 20 mL min⁻¹, and the rotation rate was 1600 rpm.

Results and discussion

Fig. 1 shows the SEM images of the activated coal liquefaction residues at various activating temperatures. All three activated carbons exhibited porous structures, and AC-900 (Fig. 1c) showed more mesopores than the AC-700 and AC-800 (Fig. 1a and b). As shown in Fig. 1d, AC-900 displayed three-dimensionally cross-linked structures, which would be expected to facilitate the rapid transport of ions and charges. Many wrinkles and edge sites on the surface of AC-900 were observed (Fig. 1e and f), and these features can improve the conductivity effectively during the process of ion transport in the ORR.

Fig. 1 (a and b) SEM images of the AC-700 and AC-800 samples, (c and d) SEM and (e and f) TEM images of the AC-900 sample.

The structures observed in the SEM images were also consistent with the XRD data. As shown in Fig. 2, as the activating temperature was increased, the degree of crystallinity of the carbon materials gradually increased as well. Generally, activated carbon has an amorphous structure and lacks long-range order in its three-dimensional structure. AC-900 yielded a prominent peak at 25.7° and a weak peak at about 42.5° in its XRD pattern, corresponding to the (002) and (100) peaks of graphite. Liu proposed the parameter “R” to evaluate the proportion of edge graphite.¹⁹ “R” is measured as the ratio of the height of the (002) Bragg peak to the background. The specific catalytic activity of the activated carbon increases with increasing values of “R”,¹⁹ indicating that the higher the percentage of the edge orientation, the higher the intrinsic catalytic activity for oxygen reduction. The “R” values of our activated carbon samples were found to increase from 1 to 1.38 to 2.67 as the activating temperature was increased. The AC-900 sample was observed in the TEM image of Fig. 1f to have a relatively large number of edges exposed on the surface, consistent with the XRD analysis.^19–21 The edge of graphite is as conductive as a metal, while the basal orientation is perpendicular to the graphene layers like an intrinsic semiconductor.²¹ The high percentage of the edge orientation acting as active sites improves the catalytic activity due to O₂ superoxide and functional groups being more easily adsorbed onto the edge plane. In addition to the effects of edges, some reports have indicated a higher carbonization temperature to also be crucial in enhancing the activity.^22,23 Apparently both the effects of the graphite edge and the activating temperature resulted in AC-900 exhibiting a higher activity than did AC-700 and AC-800.

Fig. 2 XRD patterns of the AC-700, AC-800, and AC-900 samples.

The specific surface area and the distribution of pore sizes are two key factors for the effectiveness of activated carbon for the ORR. The isotherms for AC-900 in Fig. 3a resulted from a mixture of type I and type IV isotherms. According to the IUPAC classification, the type I isotherm is associated with microporous structures and the type IV isotherm indicates a mixture of microporous and mesoporous material. The pore size distribution analysis (Fig. 3b) further proved the above conclusion. The AC-900 sample showed a specific surface area of 3130 m² g⁻¹, larger than those of AC-700 (982 m² g⁻¹) and AC-800 (1780 m² g⁻¹). Density functional theory (DFT) was used to calculate the pore size. The average pore size of AC-900 was calculated to be 1.91 nm, larger than the 1.56 nm value calculated for AC-700 and the 1.73 nm value for AC-800. Appleby et al. found that a pore diameter of 1.8 nm would be required to facilitate an effective oxygen reduction reaction.²⁴ Qu et al. assumed the surface area of pores larger than 1.5 nm to be effective at participating in the reduction reaction.⁵ Of the three samples, AC-900 showed the most effective pore size and the largest specific surface area, and showed the best ORR performance.

Fig. 3 (a) N₂ adsorption–desorption isotherms of the AC-700, AC-800 and AC-900 samples, and (b) the corresponding pore size distributions of the three samples.

The XPS spectra of the catalysts are given in Fig. 4a, and showed the presence of C, N, and O without any other elements for all three catalysts. The contents of the different elements are listed in Table 1. The carbon content gradually increased with increasing carbonization temperature, while the nitrogen content decreased from 1.65% to 0.89%. The XPS N 1s spectra of the three catalysts may each be divided into four peaks, corresponding to pyridinic-N at 398.7 ± 0.3 eV, pyrrolic-N at 400.3 ± 0.3 eV, graphitic-N at 401.4 ± 0.5 eV, and pyridinic N⁺–O⁻ at 402–405 eV.²⁵ Fig. 4b–d show the N 1s peaks of all three samples, respectively. According to the analysis of these results as shown in Table 2, the proportions of pyridinic-N and graphitic-N increased from 18.3% to 47.1% and from 9.0% to 22.8%, respectively, as the activating temperature was increased. AC-900 showed the best electrochemical performance due to its being N-doped with the most pyridinic-N and graphitic-N, which have been shown to serve as active sites for the ORR.²⁵

Fig. 4 (a) XPS spectra and (b–d) N 1s peaks of the samples carbonized at different temperatures: (b) AC-700; (c) AC-800; (d) AC-900 (fitting lines are also shown, with green for pyridinic-N, magenta for pyrrolic-N, navy for graphitic-N, and violet for pyridine-N-O⁻).

Table 1 Characteristics of the tested catalysts and the relative intensities of their XPS peaks

Catalyst	Relative intensities of different elements (at%)
	C	O	N
AC-700	91.42	6.93	1.65
AC-800	94.13	4.96	0.91
AC-900	94.91	4.20	0.89

---

Table 2 Characteristics of the tested catalysts and the relative intensities of their XPS N 1s peaks

Catalyst	Relative intensities of different elements (at%)
	Pyridinic-N	Pyrrolic-N	Graphitic-N	pyridinic N^+–O^−
AC-700	18.3	66.3	9.0	6.4
AC-800	21.3	44.8	18.5	15.4
AC-900	47.1	25.4	22.8	6.7

---

As shown in Fig. 5a–c, the CV curve of the AC-700 sample indicated almost no ORR activity. As the activation temperature was increased, however, the ORR activity was found to gradually increase. The peak position and the corresponding peak current density of AC-900 were observed at a potential of −0.168 V vs. Ag/AgCl and 1.45 mA cm⁻², respectively, in an O₂-saturated 0.1 M KOH solution (Fig. 5c). Compared with the RDE curves of AC-700, AC-800, and the Pt/C commercial catalyst at a rotation speed of 1600 rpm (Fig. 5d), the AC-900 catalyst with a scan rate of 10 mV s⁻¹ exhibited excellent electrocatalytic activity with a 6.0 mA cm⁻² peak current density, and a half-peak current density of −2.11 mA cm⁻² at −0.17 V, which was 63 mV lower than that of the commercial 20% Pt/C catalyst. At a potential of −1 V, the current density was about 6 mA cm⁻² and a plateau was observed. Fig. 5e shows the RDE voltammograms of AC-900 at different rotation speeds from 400 to 3600 rpm. The derived Koutecky–Levich (K–L) plots (Fig. 5f) showed good linearity and were nearly parallel, indicating that the ORR process catalyzed by the AC-900 electrocatalyst followed first-order kinetics in the selected potential range from −0.60 to −0.25 V.

Fig. 5 (a–c) Cyclic voltammograms (CV) curves of (a) AC-700, (b) AC-800, and (c) AC-900 (black line-N₂, red line-O₂), (d) linear sweep voltammograms of AC-700, AC-800, Ac-900, and Pt/C (20%) RDE electrodes at 1600 rpm and 10 mV s⁻¹, (e) polarization curves for the ORR in O₂-saturated 0.1 M KOH solution on AC-900 electrodes with rotation rates from 400 rpm to 3600 rpm, and (f) Koutecky–Levich plots of the AC-900 electrode at various electrode potentials and corresponding electron transfer number (n) values.

The number of transferred electrons (n) was calculated by the well-known Koutecky–Levich equation²⁶

1/J = 1/J_k + 1/(0.62nFC₀D₀^2/3v^−1/6ω^1/2) (1)

where J (shown in eqn (1) above) corresponds to the current density (mA cm⁻²), J_k is the kinetically limited current density (mA cm⁻²), ω is the angular velocity of the disk (ω = 2πN, and N is the linear rotation speed), n is specifically the overall number of transferred electrons per oxygen molecule as a result of the ORR, D₀ is the diffusion coefficient (cm s⁻¹), F is the Faraday constant (F = 96 485 C mol⁻¹), C₀ is the bulk concentration of O₂ (mol L⁻¹), v is the kinematic viscosity of the electrolyte, and k is the electron transfer rate constant, with the values of C₀, D₀ and v for the O₂-saturated 0.1 M KOH solution being 1.14 × 10⁻⁶ mol cm⁻³, 1.73 × 10⁻⁵ cm² s⁻¹, and 0.01 cm² s⁻¹, respectively. The values of n and J_k can be estimated from the slope and intercept of the K–L plots based on eqn (1). The values of n and J_k for the AC-900 catalyst were in this way estimated to be 3.32 e and 3.61 mA cm⁻², respectively, at the potential of −0.25 V in an O₂-saturated 0.1 M KOH solution. For AC-900, n did not reach 4 e between −0.60 V and −0.25 V, which is in keeping with the previously reported results of activated carbons and most N-doped carbon catalysts.²⁷ The 2-e reduction mechanism with the formation of hydrogen peroxide ion HO₂⁻ is widely accepted.²⁸ Thus, regardless of the detailed mechanism, active sites for the adsorption of oxygen species on the carbon surface are essential for oxygen reduction. For the AC-900 sample, n was found in the range from 3.00 to 3.68 e, indicative of an average electron transfer number of between two and four, and hence including both two- and four-electron transfer mechanisms.²⁷ In general, intercepts of the extrapolated K–L lines being close to zero would indicate diffusion control conditions to be in effect,²⁹ whereas non-zero intercepts would indicate the process of oxygen reduction to be under mixed kinetic–diffusion control.³⁰ As shown in Fig. 5f, with the decrease of the potential from −0.25 V to −0.60 V, the intercepts of the extrapolated lines were found to be close to zero, indicating that diffusion control conditions were in effect in our reactions.

The stabilities of AC-900 and the commercial Pt/C (20%) catalyst were evaluated by carrying out current–time chronoamperometric experiments at a constant voltage of −0.40 V and with a rotation speed of 1600 rpm (Fig. 6). During the course of testing for 20 000 s, as much as 85.1% of the initial current density of AC-900 was maintained, while only as much as 82.1% of the initial current density of commercial Pt/C (20%) catalyst was maintained.

Fig. 6 Current–time chronoamperometric response for AC-900 and Pt/C (20%) catalysts at −0.40 V vs. Ag/AgCl with a rotation speed of 1600 rpm.

Conclusions

We have shown a facile and controlled approach to prepare porous N-doped activated carbon for the ORR through one-step activation of coal liquefaction residues with KOH. The obtained AC-900 sample displayed a tunable porous pore size, a surface area as high as 3130 m² g⁻¹, a relatively large number of edges on the surface, and an optimal N-doped content of pyridinic-N (47.1%) and graphitic-N (22.8%), which endowed AC-900 with higher conductivity, more active sites, and better electrocatalytic performance for the ORR than were displayed by the other samples tested. The electrochemical investigations showed a half-peak current of −2.11 mA cm⁻² at −0.17 V for AC-900, which was only 63 mV lower than that of the commercial Pt/C (20%) catalysts. The AC-900 sample also exhibited excellent stability: as much as 85.1% of the initial current density was maintained over the course of a stability test of 20 000 s compared with a value of 82.1% for Pt/C (20%) catalyst.

There are several possible explanations for the high electrochemical performances displayed by AC-900. First, its large specific surface area of 3130 m² g⁻¹ and optimal mean pore size of 1.91 nm likely facilitated the rapid transport of ions and charges and improved the use of active sites. Second, the graphite edges of AC-900 may have enhanced the conductivity and also served as active sites, improving the catalytic activity. Third, the high contents of pyridinic-N and graphitic-N in AC-900 probably yielded many active sites on the surface of this activated carbon for the ORR. Activated carbons with large specific surface area and the rational average pore size from the liquefaction residues take a more significant result in developing the composite ORR catalysts based on activated carbon.

Acknowledgements

The authors acknowledge financial support from the National Key Basic Research Program No. 2013CB933103 funded by MOST and the Program for the Fundamental Research Supported by Shenzhen Science and Technology Innovations Council of China (Grant No. JSF201006300047A, No. JC201105201126A, and No. ZDSY20120619140933512). The authors are very grateful to the China Postdoctoral Science Foundation (No. 2015M5811087) for funding the project.

Notes and references

1. K. N. Wood, R. O'Hayre and S. Pylypenko, Energy Environ. Sci., 2014, 7, 1212–1249.
2. R. Cao, J. S. Lee and M. Liu, Adv. Energy Mater., 2012, 2, 816–819.
3. X. Yan, Y. Jia, T. Odedairo, X. Zhao, Z. Jin, Z. Zhu and X. Yao, Chem. Commun., 2016, 52, 8156–8159.
4. J. Maruyama, K. Sumino, M. Kawaguchi and I. Abe, Carbon, 2004, 42, 3115–3121.
5. D. Qu, Carbon, 2007, 45, 1296–1301.
6. A. Soffer and M. Folman, J. Electroanal. Chem., 1972, 38, 25–43.
7. G. Tao, L. Zhang, L. Chen, X. Cui, Z. Hua, M. Wang, J. Wang, Y. Chen and J. Shi, Carbon, 2015, 86, 108–117.
8. B. Fang, J. H. Kim, M. S. Kim and J. S. Yu, Acc. Chem. Res., 2013, 46, 1397–1406.
9. R. W. Fu, Z. H. Li, Y. R. Liang, F. Li, F. Xu and D. C. Wu, New Res. Carbon Mater., 2011, 26, 171–178.
10. E. Antolini, Appl. Catal., B, 2009, 88, 1–24.
11. K. Vignarooban, J. Lin, A. Arvay, S. Kolli, I. Kruusenberg, K. Tammeveski, L. Munukutla and A. M. Kannan, Chin. J. Catal., 2015, 36, 458–472.
12. K. Gong, F. Du, Z. Xia, M. Durstock and L. Dai, Science, 2009, 323, 760–764.
13. C. Guo, W. Liao, Z. Li and C. Chen, Carbon, 2015, 85, 279–288.
14. H. Jin, H. Zhang, H. Zhong and J. Zhang, Energy Environ. Sci., 2011, 4, 3389–3394.
15. M. Zhou, F. Pu, Z. Wang and S. Guan, Carbon, 2014, 68, 185–194.
16. Y. Wang and X. Jiang, ACS Appl. Mater. Interfaces, 2013, 5, 11597–11602.
17. P. Zhang, F. Sun, Z. Xiang, Z. Shen, J. Yun and D. Cao, Energy Environ. Sci., 2014, 7, 442–450.
18. G. Wu, K. L. More, C. M. Johnston and P. Zelenay, Science, 2011, 332, 443–447.
19. Y. Liu, J. S. Xue, T. Zheng and J. R. Dahn, Carbon, 1996, 34, 193–200.
20. L. Zhao, Y. Qiu, J. Yu, X. Deng, C. Dai and X. Bai, Nanoscale, 2013, 5, 4902–4909.
21. D. Qu, J. Power Sources, 2002, 109, 403–411.
22. G. Zhong, H. Wang, H. Yu and F. Peng, Electrochem. Commun., 2014, 40, 5–8.
23. I. Y. Jeon, H. J. Choi, S. M. Jung, J. M. Seo, M. J. Kim, L. M. Dai and J. B. Baek, J. Am. Chem. Soc., 2013, 135, 1386–1393.
24. A. J. Appleby and J. Marie, Electrochim. Acta, 1979, 24, 95–202.
25. D. Long, J. Zhang, J. Yang, Z. Hu, G. Cheng, X. Liu, R. Zhang, L. Zhan, W. Qiao and L. Ling, Carbon, 2008, 46, 1253–1269.
26. C. Jin, F. Lu, X. Cao, Z. Yang and R. Yang, J. Mater. Chem. A, 2013, 1, 12170–12177.
27. J. Masa, W. Xia, M. Muhler and W. Schuhmann, Angew. Chem., Int. Ed., 2015, 54, 10102–10120.
28. L. Dai, Y. Xue, L. Qu, H.-J. Choi and J.-B. Baek, Chem. Rev., 2015, 115, 4823–4892.
29. M. Vikkisk, I. Kruusenberg, S. Ratso, U. Joost, E. Shulga, I. Kink, P. Rauwel and K. Tammeveski, RSC Adv., 2015, 5, 59495–59505.
30. N. Alexeyeva, E. Shulga, V. Kisand, I. Kink and K. Tammeveski, J. Electroanal. Chem., 2010, 648, 169–175.

---