Synthesis of B-doped hollow carbon spheres as efficient non-metal catalyst for oxygen reduction reaction

Abstract

The oxygen reduction reaction (ORR) is one of the crucial reactions in fuel cells and metal–air batteries. Heteroatom doped carbon spheres can serve as alternative low-cost non-metal electrocatalysts for ORR. Herein, we developed an effective route to the synthesis of uniform and electrochemically active B-doped hollow carbon nanospheres (BHCSs). BHCSs were synthesized via the carbonization of a boric phenolic resin supported by SiO₂, followed by etching the SiO₂ template. The content of B, B dopant species and specific surface area were adjusted by changing the content of the B precursor and the calcination temperature. Moreover, their influence on the performance of electrocatalytic activity was explored. It was found that, among these B-doping type materials (BC₂O, BCO₂, B₄C and BC₃), B–C bonds (B₄C and BC₃) played a crucial role on improving the electrocatalytic activity. Compared with the hollow carbon nanospheres (HCSs), a 70 mV positive shift of the onset potential and 1.7 times kinetic current density could be clearly observed with BHCSs. In addition, the BHCSs revealed better stability and methanol tolerance than commercial Pt/C (HiSPEC™ 3000, 20%). Thus, the as-prepared BHCSs, as inexpensive and efficient non-metal ORR catalysts, may have a promising application in direct methanol fuel cells.

---

Introduction

The oxygen reduction reaction (ORR), a key process in metal–air batteries and fuel cells, is kinetically sluggish and thus requires a suitable catalyst.^1–5 To date, Pt-based electrocatalysts have become the popular materials for the ORR with high current density and a four-electron pathway.^6–8 However, they still suffer from poor stability and CO and methanol poisoning. Furthermore, the high cost and the limited natural resource of Pt have blocked its large-scale application.^9,10 In this respect, non-metal carbon-based catalysts have attracted great attention due to their outstanding stability, low cost, good methanol and CO tolerances and high electrocatalytic activity.^11,12

The electrons in the sp² carbon materials are too inert to be utilized directly in the ORR.¹³ On doping with heteroatoms in a carbon framework, the carbon electrons become activated; thus, O₂ molecules can be reduced on the positively charged atoms.¹⁴ When a heteroatom is bonded with the carbon framework, it can introduce a defect in the adjacent sites due to the difference in atomic size and bond length.^15,16 Moreover, the electronegativity of different atoms can make the charge distribution uneven.^17,18 Therefore, heteroatom-doped carbon has been pioneered for ORR electrocatalysts.^19–29 In particular, because C has a larger electronegativity (2.55) than B (2.04), a certain amount of positive charge is induced to the B atoms. These B atoms are favorable for capturing oxygen molecules.³⁰ Besides, once B is doped into the carbon matrix, the electrons of the C–C π* orbitals would transfer to vacant 2p_z orbitals of B. Hu et al. found that this partially filled 2p_z orbital makes up the main protruding lobe in the two highest occupied molecular orbitals (HOMO) of BHCSs.³¹ Upon adsorption, the lowest unoccupied molecular orbital (LUMO) of oxygen would overlap with the HOMO of BHCSs. As a result, a certain amount of charge transfers to oxygen, causing the stretch of the O–O bond length. The weakness of the O–O bond makes the reduction of oxygen more efficient. These B-doped carbon materials, such as graphene,²⁸ carbon nanotubes,³¹ ordered mesoporous carbons,³² and carbon rods,³³ can increase the electrocatalytic performance for the ORR.

Hollow carbon spheres (HCSs) have attracted considerable attention because of their superior properties such as low density, high specific surface area, good chemical stability and electrical conductivity. These features guarantee them to be promising in drug delivery, catalysis, energy conversion and storage.^34–36 The high specific surface area of HCSs facilitates the exposure of active sites for catalysis. To date, most synthetic routes of doped HCSs were through two main approaches: the first through a one-step pyrolysis of a heteroatom contained carbon precursor, and the second through a post-treatment of HCSs with a reactive heteroatom precursor.³⁷ However, to the best of our knowledge, there have been no studies based on a B-doped hollow carbon sphere catalyst for the ORR.

Herein, we report a catalyst-free and facile approach to synthesize B-doped hollow carbon spheres (BHCSs). BHCSs were prepared using SiO₂ spheres as the hard template, 4-hydroxyphenylboronic acid as the B source and resorcinol and methanol as the carbon sources. In this process, rather than simply obtaining a physical mixture, 4-hydroxyphenylboronic acid was added into phenol to create a reaction to produce boric phenyl ester, which further reacted with formaldehyde to form a boric phenolic resin.³⁸ Scheme 1 shows the schematic illustration of the formation process of the BHCSs, including a hydrothermal reaction, carbonization, and HF etching. The resulting BHCSs have superior electrochemical performance for the ORR. As metal-free catalysts, BHCSs show much better long-term durability and methanol tolerance than the commercial Pt/C (HiSPEC™ 3000, 20%) in an alkaline medium.

Scheme 1 Schematic illustration of the formation process of the B-doped hollow carbon spheres (BHCSs).

Experimental

Chemicals

Tetraethyl orthosilicate (TEOS) (99%) was purchased from Aldrich (China). Resorcinol and formaldehyde (37 wt%) were bought from Sinopharm Chemical Reagent Co. Ltd. (China). 4-Hydroxyphenylboronic acid was purchased from Ningbo Yingfa Pengna. Ltd. (China). Commercial 20 wt% Pt/C (HiSPEC™ 3000) was from Johnson Matthey. Other chemicals, such as ammonia solution (28%), potassium hydroxide and ethanol, were obtained from Sinopharm Chemical Reagent Co. Ltd. (China). All the reagents were of analytical grade and used as received. All the solutions used in the experiments were freshly prepared with ultrapure water having a resistivity of 18.2 MΩ cm⁻¹ (Nanjing Baocheng Biotechnology CO. Ltd., China).

Preparation of HCSs and BHCSs

Herein, we synthesised the HCSs using SiO₂ as the template and resorcinol–formaldehyde as the carbon source according to the literature.³⁹ In this synthesis process, silica colloidal spheres were first formed via the classical Stöber method,⁴⁰ which then acted as cores for the deposition of resorcinol–formaldehyde shells. HCSs were prepared according to the method reported by Dongyuan Zhao.³⁵ Generally, 4.25 mL of TEOS mixed with 75 mL of ethanol was added to a solution containing 50 mL of ethanol, 15 mL of ultrapure water and 10 mL aqueous solution of ammonia with stirring. After 1 h, 1 g of resorcinol and 1.4 mL of formaldehyde solution were added. The solution was stirred for 24 h at room temperature, then transferred to a 250 mL Teflon-lined stainless steel autoclave, and hydrothermally treated for 24 h at 100 °C. After cooling, the solid product was obtained by centrifugation and desiccation at 60 °C overnight. Then, as-prepared carbon spheres were thermally treated in a tube furnace at a heating rate of 5 °C min⁻¹ under a nitrogen atmosphere until a final temperature was reached, and then the samples were kept for 4 h. The samples were allowed to cool naturally to room temperature. After the composites were immersed in 10 wt% HF solutions for 36 h, the SiO₂ colloidal cores were removed and HCSs were obtained. The preparation of B-doped HCSs was performed using 4-hydroxyphenylboronic acid partly instead of only resorcinol for the preparation of HCSs. The total amount of 4-hydroxyphenylboronic acid and resorcinol was controlled to 1 g with different mass ratios. The obtained BHCSs were boiled in deionized water at 100 °C for 6 h to dissolve the B that had not been doped into the carbon framework. In this paper, we name the samples as BHCSs-m-n, in which m represents the mass percent of 4-hydroxyphenylboronic acid and n represents the different calcination temperature.

Apparatus

The surface morphologies of the samples were analyzed by field emission scanning electron microscopy (FE-SEM) using an ULTRA plus microscope (Zeiss, Germany) and a JEM-2100 field emission transmission electron microscope (FE-TEM, JEOL, Japan) with an acceleration voltage of 200 kV. Particle size distribution was obtained by dynamic light scattering analysis (DLS, Brookhaven BI-200SM, USA). Raman spectra were collected by a DXR-Microscope (Thermo Fisher, USA) with the excitation wavelength of 532 nm. Nitrogen sorption isotherms and Brunauer–Emmett–Teller (BET) surface areas were investigated by NovaWin (Quantachrom, USA). Thermogravimetric analysis (TG) was performed on the SDT-Q600 thermal analyzer (TA Instruments, USA) in an N₂ atmosphere at a heating rate of 20 °C min⁻¹ with a temperature range from 50 to 900 °C. X-ray photoelectron spectroscopy (XPS) was performed by a PHI5000 VersaProbe (Ulvac-Phi, Japan) system with monochromic Mg-Kα radiation. All the electrochemical measurements were performed on a CHI700 electrochemical workstation (CH Instruments, China) and rotating ring disk electrode (RRDE-3A, Japan) in a typical three-electrode cell equipped with gas flow systems. A glassy carbon electrode (GCE) with a diameter of 3 mm was used as the working electrode. A platinum filament and Ag/AgCl (saturated KCl) were used as the counter and reference electrode, respectively. All the electrochemical experiments were carried out at room temperature. The GCE was polished successively with 1.0 and 0.3 μm alumina powders before modification, rinsed thoroughly with ethanol/water and double distilled water, followed by drying under a stream of nitrogen. 4 mg of BHCSs was ultrasonically dispersed in 1 mL ethanol to form homogeneous BHCSs suspensions. Then, 10 μL of this suspension was cast on the GCE surface and dried at 60 °C in an oven for 15 min. Subsequently, 5 μL of Nafion (0.05 wt% in ethanol) was cast on the GCE surface and allowed to dry in a 60 °C oven for another 15 min to obtain the BHCSs modified GCE. For all the cyclic voltammograms and rotating disk electrode (RDE) experiments, the loading of catalysts was 0.28 mg cm⁻² for Pt/C and 0.56 mg cm⁻² for HCSs and BHCSs. The electrolyte was 0.1 M KOH in water, which was bubbled with O₂ or N₂ for 30 min and maintained in the same atmosphere during the measurements. Cyclic voltammetry (CV), linear sweep voltammetry (LSV) and chronoamperometry were carried out to measure the catalytic performances of the BHCSs for the ORR. All the current densities were normalized to the geometric surface area of the GCE. The capacitive current densities were subtracted from the apparent current densities by deducting the background current density obtained from the deoxygenated electrolyte.

Results and discussion

Characterization of the BHCSs

Herein, the sample BHCSs-0.3-900 is taken as an example. The colloidal SiO₂ nanospheres prepared by a classical Stöber method were smooth and monodispersed with a uniform diameter of ∼170 nm (Fig. S1a, ESI†). A further modified coating process and hydrothermal reaction lead to the formation of a B doped carbon spheres with a diameter of ∼230 nm (Fig. S1b, ESI†). The in situ polymerization was induced by the NH₄⁺ ions that were decorated on the surface of the SiO₂ spheres under basic conditions, which could not only prevent the colloidal suspension from aggregating but also accelerate the polymerization of the boric phenolic resin.⁴¹ The B doped carbon spheres retained the original core–shell structure after a heat treatment but the diameter was reduced to ∼200 nm (Fig. 1a and b), indicating an apparent structure shrinkage during the carbonization. After removing the silica cores by an etching treatment, the SEM image showed that the resulting BHCSs-0.3-900 efficiently retained the spherical morphology with no obvious collapse observed (Fig. 1c), implying that this hollow structure may possess great mechanical strength. TEM images of the BHCSs-0.3-900 clearly exhibit hollow nanostructures with a shell thickness of ∼15 nm and a void diameter of ∼170 nm (Fig. 1d), which is in agreement with the diameter of the SiO₂ cores. The DLS data further confirms that the BHCSs-0.3-900 nanospheres had a narrow size distribution (Fig. S2a, ESI†).

Fig. 1 SEM (a and c) and TEM (b and d) images of the BHCSs before (a and b) and after (c and d) the removal of colloidal SiO₂ cores. (e) Nitrogen adsorption–desorption isotherms of HCSs-900 and various BHCSs samples. (f) Raman spectra of HCSs-900 and various BHCSs samples.

Nitrogen sorption measurements showed that HCSs-900 possessed a microporous structure and that its specific surface area value was 478.31 m² g⁻¹ (Fig. 1e). The specific surface area value of the BHCSs-0.3-900 was smaller than HCSs-900, indicating that B was successfully doped into HCSs. The introduction of B into the framework of HCSs led to a decrease in the specific surface area.³² On the other hand, the samples calcined at a relatively high temperature showed a relatively low specific surface area. High annealing temperatures may lead to the collapse of the micropores. The detailed information of the samples is summarized in Table 1. According to thermogravimetric analysis under N₂ (Fig. S2b†), a three-step weight loss process could be observed for both HCSs and BHCSs-0.3-900. The majority of the chemical decomposition took place from 400–900 °C and this weight loss may be attributed to a suite of complex chemical reactions involving bond formation and crosslinking,²⁴ which results in the destruction of the main chain and the release of small gases (CO, CO₂, CH₄ and H₂).⁴³ However, there was no obvious mass loss at higher temperatures (800–900 °C) for BCHSs, indicating that the BHCSs possess higher heat resistance and carbon yield. Raman spectroscopy is the most effective technique to characterize the structure and determine the defects and disordered structures. The Raman spectra of the HCSs-900 and BHCSs-0.3-900 exhibit the presence of two strong bands at 1350 and 1590 cm⁻¹, corresponding to the D and G bands, respectively (Fig. 1e). Although there are no apparent changes in the positions of the D and G bands between BHCSs and HCSs, BHCSs (except BHCSs-0.3-600) show higher intensity ratios of I_D/I_G, due to the introduction of defects in the carbon framework by the heterogeneous B-dopants. Notably, the I_D/I_G ratio of BHCSs gradually increase upon increasing the annealing temperature, suggesting that higher annealing temperature could induce more defects.⁴²

Table 1 Properties of HCSs and BHCSs

Catalyst	Mass of B source	BET surface area (m^2 g^−1)	XPS (at%)
			C	B	O
HCS-900	0.0	478.31	96.55	—	3.45
HCS-0.2-900	0.2	395.56	91.55	0.61	7.84
HCS-0.3-900	0.3	307.27	90.94	0.91	8.15
HCS-0.4-900	0.4	242.17	89.94	1.08	8.98
HCS-0.3-1000	0.3	131.08	92.31	0.41	7.28
HCS-0.3-800	0.3	395.22	90.25	1.12	8.63
HCS-0.3-600	0.3	468.77	89.17	1.36	9.47

---

The elemental compositions and chemical bonding information of HCSs and BHCSs were investigated via XPS measurements. The XPS survey scan for BHCSs shows a clear B band for BHCSs, confirming that B was successfully doped into HCSs and the silica have been etched absolutely (Fig. 2a). The atomic contents of C, B, and O in the samples are summarized in Table 1. It was found that the content of B and O increased as the initial mass of 4-hydroxyphenylboronic acid increased, which is in agreement with the B-doped materials.³² Based on the Gauss Amp, the B 1s peak in BHCSs can be fitted into four components (Fig. 2b). The peaks centered at 187.5 eV and 189.5 eV may be assigned to B₄C and BC₃ structures, respectively. The peaks at 191.2 eV and 192.3 eV correspond to BC₂O and BCO₂, respectively.^31,32,44 Previous studies demonstrated that the variation in the amount of B–C is allegedly responsible for the ORR activity.^31,33 The B doping sites of BC₂O and BCO₂ are on the edge of the carbon framework and these two doping types, particularly BCO₂, may make negative influences on the electronic conductivity of the samples.⁴⁵ The improvement in the ORR activity by the BC₃ species contained in the BHCSs is demonstrated by the following electrochemical measurements. The potential schematic structure of the BHCSs is outlined in Fig. 2c. The content of each B species is listed in Table 2. It can be seen that, with the increase of the calcination temperature, the bands belonging to B–C (BC₃ and B₄C) became stronger, which is in agreement with the literature.⁴⁴ By increasing the B doping amount, the content of B–C bonding increased at first and then decreased. However, the reason is still unknown and requires further research.

Fig. 2 (a) XPS spectrum of HCSs-900 (black) and BHCSs-0.3-900 (red). (b) High resolution B 1s XPS spectrum of BHCSs-0.3-700, BHCSs-0.3-800 and BHCSs-0.3-900. The B 1s peak was deconvoluted into four peaks at 187.5 (B₄C, blue line), 189.5 (BC₃, pink line), 191.2 (BC₂O, green line), 192.3 (BCO₂, orange line). (c) Possible types of structure in the BHCSs derived from XPS measurements: ¹B (B₄C); ²B (BC₃); ³B (BC₂O); ⁴B (BCO₂).

Table 2 B Species content of BHCSs samples by XPS

	B species content (at%)
	B4C	BC3	BC2O	BCO2
HCS-0.2-900	19.33	30.95	32.02	17.70
HCS-0.3-900	19.23	32.14	31.59	17.11
HCS-0.4-900	17.99	28.10	32.97	20.94
HCS-0.3-1000	20.29	33.50	30.58	15.63
HCS-0.3-800	17.23	27.57	25.55	29.65
HCS-0.3-600	11.46	22.49	25.51	40.54

---

Electrochemical behavior of BHCSs

The catalytic activity of BHCSs towards the ORR was evaluated by cyclic voltammetry in an oxygen-saturated 0.1 M KOH solution. Fig. 3a portrays the electrochemical reduction of oxygen at HCSs-900 (black line) and BHCSs-0.3-900 (red line) electrodes. For comparison, the same mass of each catalyst was loaded onto the electrode and this along with the geometrical area of the electrode was used to calculate the current density. For the N₂-saturated solution, reduction peaks are negligible both for HCSs-900 (black dash line) and BHCSs-0.3-900 (red dash line). In the presence of oxygen, a remarkable reduction peak at −0.174 V is observed on the obtained BHCSs-0.3-900 electrode, showing a substantial reduction process, which is 70 mV positive than that of HCSs-900.

Fig. 3 (a) CVs for BHCSs-0.3-900 and HCSs-900 on a glassy carbon rotating disk electrode in an O₂-saturated 0.1 M KOH solution with a scan rate of 10 mV s⁻¹. (b) Rotating-disk voltammograms of HCSs-900, various BHCSs samples, Pt/C in O₂-saturated a 0.1 M KOH solution with a scan rate of 10 mV s⁻¹ and rotation rate of 1600 rpm. (c) Rotating-disk voltammograms of BHCSs-0.3-900 in an O₂-saturated 0.1 M KOH solution with a sweep rate of 10 mV s⁻¹ at different rotation rates. Inset: the corresponding Koutecky–Levich plot of J⁻¹ versus ω^−1/2 from −0.3 to −0.8 V. (d) Summary of the kinetic limiting current density and the electron-transfer number on the basis of the RDE data on various catalysts. A, HCSs-900; B, BHCSs-0.3-600; C, BHCSs-0.3-800; D, BHCSs-0.3-900; E, BHCSs-0.3-1000; F, BHCSs-0.2-900; G, BHCSs-0.4-900; H, Pt/C. (e) Chronoamperometry of BHCSs-0.3-900 and Pt/C electrodes in an O₂-saturated 0.1 M KOH solution at −0.3 V for 130 min with a rotation rate of 1600 rpm. (f) Chronoamperometric responses of BHCSs-0.3-900 and Pt/C electrodes with 3 M methanol added at 4 min with the rotation rate of 1600 rpm.

To further research the ORR process and estimate the transferred electron number (n) per oxygen molecule, a rotating disk electrode (RDE) was used to investigate the electrocatalytic activity of BHCSs for the ORR by linear sweep voltammetry (LSV). Fig. 3b shows the corresponding LSVs for HCSs and BHCSs obtained with different masses of B precursor on a Pt/C disk electrode in an oxygen-saturated 0.1 M KOH electrolyte. The HCSs-900, BHCSs-0.2-900, BHCSs-0.3-900, BHCSs-0.4-900 and Pt/C electrodes showed an ORR onset potential at around −0.173 V, −0.144 V, −0.101 V and −0.009 V, respectively. However, different mass loadings of BHCSs-0.3-900 could lead to different onset potentials (Fig. S3, ESI†). As a non-metal catalyst, the BHCSs-0.3-900 showed ORR activity with an onset potential of −0.101 V in alkaline media, which is comparable to the other heteroatom doped carbon materials (Table S1, ESI†). In addition to the obvious positive shift of the onset potential, the current density of BHCSs-0.3-900 was much stronger than that of HCSs-900. With the increase of B doping amount, the electrocatalytic activity of the BHCSs-900 ameliorates initially and then deteriorates. Compared with BHCSs-0.2-900, the enhancement in the cathodic response of BHCSs-0.3-900 is due to the increased B content. However, it should be noted that the electrocatalytic activity of BHCSs-0.4-900 did not improve as expected. This may be caused by the low surface area and low content of B–C bonding, especially the BC₃ composition.

Many studies have shown that a high temperature around 600–1000 °C is vital for a successful graphitization and B implantation with optimal ORR catalytic activity.^31,33,44 Therefore, in this paper, four different calcination temperatures were investigated. Fig. 3b shows that the onset potential and the current density enhanced obviously with the increase of temperature. The BHCSs-0.3-600, BHCSs-0.3-800, BHCSs-0.3-900 and BHCSs-0.3-1000 electrodes showed ORR onset potentials at around −0.189 V, −0.133 V, −0.101 V and −0.137 V, respectively. Although with the increase of temperature, the surface area of the samples decreased, probably due to the collapse of pores and enhanced orientation during the carbonization process, the efficient active sites (B–C bonding) and the increase of graphitization armed the BHCSs-0.3-900 with higher catalytic activity. However, when the temperature reaches 1000 °C, due to the low B content and surface area, the electrocatalytic activity of the BHCSs deteriorates.

The detailed voltammograms at different rotation rates of BHCSs-0.3-900 are shown in Fig. 3c. The measured current density (J) increased with increasing rotation rate (ω) for the shortened diffusion layer. Koutecky–Levich plots revealed good linearity between J⁻¹ and ω^−1/2, and approximately constant slopes over the potentials from −0.3 to −0.8 V were observed (Inset of Fig. 3c). The electron transfer number (n) and kinetic current density (J_k) are the two important parameters that reflect the kinetics of the ORR process; these are calculated from the slope and the reciprocal of the intercept of the K–L plots, respectively. According to the Koutecky–Levich equations (eqn (1)–(3), ESI†), the calculated n value of BHCSs-0.3-900 is 3.7 at −0.6 V (Fig. 3d), which increases to 4 at more negative potentials (Fig. S4†); this indicates that oxygen reduction primarily followed the 4 electron pathway by reducing oxygen directly to OH⁻. In contrast, the value of n for HCSs at −0.6 V is 2.4, demonstrating the oxygen reduction is mainly a two electron process with the formation of intermediate HO₂⁻ ions. The calculated kinetic current density of BHCSs-0.3-900 is 6.91 mA cm⁻², while kinetic current density of HCSs-900 is 4.06 mA cm⁻² at −0.6 V. All these results indicate that doping B into the carbon framework endows the materials with a better electrocatalytic activity for ORR than that of the pristine HCSs.

The durability of the catalysts is also of major concern in fuel cells. Another attractive feature of the BHCSs catalysts is their highly steady amperometric response to the ORR. The stability of BHCSs-0.3-900 and Pt/C catalysts in an O₂-saturated 0.1 M KOH solution was evaluated by chronoamperometry at −0.35 V with a 1600 rpm electrode rotation rate. After 8000 seconds, the commercial Pt/C catalysts exhibited a loss of 15% of the catalytic activity (Fig. 3e), while the BHCSs-0.3-900 remained stable throughout the experiment with only a 1.5% current diminution (Fig. 3e). Besides the excellent durability of the catalysts towards the ORR, BHCSs-0.3-900 also showed good methanol tolerance. As shown in Fig. 3f, the addition of 3 M methanol to the O₂ saturated electrolyte did not apparently affect the catalytic activity of BHCSs-0.3-900 towards ORR (Fig. 3f). The Pt/C catalyst curve displays a dramatic decay of relative current. This result clearly shows that BHCSs-0.3-900 selectively reduces O₂, which is beneficial to avoid the occurrence of mixed potential in methanol fuel cells.

Conclusions

In conclusion, BHCSs were synthesized from polymerized SiO₂ supported boric phenolic resin by the carbonization of the latter, followed by etching away the SiO₂ template. BHCSs exhibited outstanding catalytic activity for the ORR with better long-term stability and higher selectivity than commercial Pt/C catalysts in alkaline electrolytes. Furthermore, we discussed the role of B content, active sites, specific surface areas and calcination temperature in the oxygen reduction process. The results demonstrate that BHCSs is a promising metal-free catalyst for the ORR in fuel cells and other catalytic applications.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (nos 21035002, 21121091 and 21005016) and the Natural Science Foundation of Jiangsu Province (no. BK2011591).

Notes and references

1. H. A. Gasteiger and N. M. Markovic, Science, 2009, 324, 48–49.
2. M. K. Debe, Nature, 2012, 486, 43–51.
3. N. Daems, X. Sheng, I. F. J. Vankelecom and P. P. Pescarmona, J. Mater. Chem. A, 2014, 2, 4085–4110.
4. R. G. Cao, J. S. Lee, M. L. Liu and J. Cho, Adv. Energy Mater., 2012, 2, 816–829.
5. Y. Y. Liang, Y. G. Li, H. L. Wang and H. J. Dai, J. Am. Chem. Soc., 2013, 135, 2013–2036.
6. Y. H. Bing, H. S. Liu, L. Zhang, D. Ghosh and J. J. Zhang, Chem. Soc. Rev., 2010, 39, 2184–2202.
7. Y. J. Li, Y. J. Li, E. B. Zhu, T. McLouth, C. Y. Chiu, X. Q. Huang and Y. Huang, J. Am. Chem. Soc., 2012, 134, 12326–12329.
8. Y. H. Bing, H. S. Liu, L. Zhang, D. Ghosh and J. J. Zhang, Chem. Soc. Rev., 2010, 39, 2184–2202.
9. Z. L. Liu, X. H. Lin, J. Y. Lee, W. D. Zhang, M. Han and L. M. Gan, Langmuir, 2002, 18, 4054–4060.
10. A. Morozan, B. Jousselme and S. Palacin, Energy Environ. Sci., 2011, 4, 1238–1254.
11. H. X. Chang and H. K. Wu, Energy Environ. Sci., 2013, 6, 3483–3507.
12. A. Ambrosi, C. K. Chua, A. Bonanni and M. Pumera, Chem. Rev., 2014, 114, 7150–7188.
13. Y. Zhao, L. J. Yang, S. Chen, X. Z. Wang, Y. W. Ma, Q. Wu, Y. F. Jiang, W. J. Qian and Z. Hu, J. Am. Chem. Soc., 2013, 135, 1201–1204.
14. E. J. Yoo, J. J. Nakamura and H. S. Zhou, Energy Environ. Sci., 2012, 5, 6928–6932.
15. Y. M. Liu, S. Chen, X. Quan, H. M. Zhao, H. T. Yu and B. Y. Zhang, J. Mater. Chem. A, 2013, 1, 14706–14712.
16. D. W. Kim, O. L. Li and N. Saito, Phys. Chem. Chem. Phys., 2015, 17, 407–413.
17. Y. M. Liu, S. Chen, X. Quan, H. T. Yu, H. M. Zhao, Y. B. Zhang and G. H. Chen, J. Phys. Chem. C, 2013, 117, 14992–14998.
18. L. Wang, P. Yu, L. Zhao, C. G. Tian, D. Y. Zhao, W. Zhou, J. Yin, R. H. Wang and H. G. Fu, Sci. Rep., 2014, 4, 5184–5192.
19. Y. J. Zhang, K. Fugane, T. Mori, L. Niu and J. H. Ye, J. Mater. Chem., 2012, 22, 6575–6580.
20. B. Zheng, J. Wang, F. B. Wang and X. H. Xia, J. Mater. Chem. A, 2014, 2, 9079–9084.
21. W. J. Jiang, J. S. Hu, X. Zhang, Y. Jiang, B. B. Yu, Z. D. Wei and L. J. Wan, J. Mater. Chem. A, 2014, 2, 10154–10160.
22. X. Xu, S. J. Jiang, Z. Hu and S. Q. Liu, ACS Nano, 2010, 4, 4292–4298.
23. J. Wu, Z. R. Yang, Q. J. Sun, X. W. Li, P. Strasser and R. Z. Yang, Electrochim. Acta, 2014, 127, 53–60.
24. Y. Li, T. T. Li, M. Yao and S. Q. Liu, J. Mater. Chem., 2012, 22, 10911–10917.
25. J. L. Shui, F. Du, C. M. Xue, Q. Li and L. M. Dai, ACS Nano, 2014, 8, 3015–3022.
26. D. S. Yang, D. Bhattacharjya, S. Inamdar, J. Park and J. S. Yu, J. Am. Chem. Soc., 2012, 134, 16127–16130.
27. M. Seredych and T. J. Bandosz, Carbon, 2014, 66, 227–233.
28. Z. H. Sheng, H. L. Gao, W. J. Bao, F. B. Wang and X. H. Xia, J. Mater. Chem., 2012, 22, 390–395.
29. H. Shi, Y. F. Shen, F. He, Y. Li, A. R. Liu, S. Q. Liu and Y. J. Zhang, J. Mater. Chem. A, 2014, 2, 15704–15716.
30. Y. Zheng, Y. Jiao, M. Jaroniec, Y. G. Jin and S. Z. Qiao, Small, 2012, 8, 3550–3566.
31. L. J. Yang, S. J. Jiang, Y. Zhao, L. Zhu, S. Chen, X. Z. Wang, Q. Wu, J. Ma, Y. W. Ma and Z. Hu, Angew. Chem., Int. Ed., 2011, 50, 7132–7135.
32. X. J. Bo and L. P. Guo, Phys. Chem. Chem. Phys., 2013, 15, 2459–2465.
33. G. Jo and S. Shanmugam, Electrochem. Commun., 2012, 25, 101–104.
34. P. Rijiravanich, M. Somasundrum and W. Surareungchai, Anal. Chem., 2008, 80, 3904–3909.
35. S. S. Feng, W. Li, Q. Shi, Y. H. Li, J. C. Chen, Y. Ling, A. M. Asiri and D. Y. Zhao, Chem. Commun., 2014, 50, 329–331.
36. Y. Li, M. Yao, T. T. Li, Y. Y. Song, Y. J. Zhang and S. Q. Liu, Anal. Methods, 2013, 5, 3635–3638.
37. T. Y. Yang, J. Liu, R. F. Zhou, Z. G. Chen, H. Y. Xu, S. Z. Qiao and M. J. Monteiro, J. Mater. Chem. A, 2014, 2, 18139–18146.
38. G. P. Yin, Y. Z. Gao, P. F Shi, X. Q. Cheng and A. Aramata, Mater. Chem. Phys., 2003, 80, 94–101.
39. N. Li, Q. Zhang, J. Liu, J. Joo, A. Lee, Y. Gan and Y. D. Yin, Chem. Commun., 2013, 49, 5135–5137.
40. W. Stober, A. Fink and E. Bohn, J. Colloid Interface Sci., 1968, 26, 62–69.
41. J. Liu, S. Z. Qiao, H. Liu, J. Chen, A. Orpe, D. Y. Zhao and G. Q. Lu, Angew. Chem., Int. Ed., 2011, 50, 5947–5951.
42. J. T. Jin, X. G. Fu, Q. Liu, Y. R. Liu, Z. Y. Wei, K. X. Niu and J. Y. Zhang, ACS Nano, 2013, 7, 4764–4773.
43. R. Balgis, S. Sago, G. M. Anilkumar, T. Ogi and K. Okuyama, ACS Appl. Mater. Interfaces, 2013, 5, 11944–11950.
44. J. T. Jin, F. P. Pan, L. H. Jiang, X. G. Fu, A. M. Liang, Z. Y. Wei, J. Y. Zhang and G. Q. Sun, ACS Nano, 2014, 8, 3313–3321.
45. C. Wang, Z. Y. Guo, W. Shen, Q. J. Xu, H. M. Liu and Y. G. Wang, Adv. Funct. Mater., 2014, 24, 5511–5521.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra07909h

---