Facile synthesis of porous carbon spheres embedded with metal nanoparticles and their applications as supercapacitor electrodes

Abstract

We report a novel synthetic method to prepare porous carbon spheres embedded with metal nanoparticles (M-PCS) by hydrothermal treatment of a solution of sodium gluconate and a metal salt, followed by calcination in N₂. It is a template-free method and offers a combination of a porous structure and uniformly dispersed metal nanoparticles. As one of the potential applications, the as-prepared M-PCS possess an ultrahigh capacitance (1308.7, 1380.9 and 543.6 F g⁻¹ for Fe-PCS, Co-PCS, and Ni-PCS, respectively at a current density of 10 A g⁻¹) when applied as an electrode material for supercapacitors.

---

Introduction

Porous carbon materials are of great interest for their applications as electrode materials for batteries and supercapacitors, as absorbents for separation processes and gas storage, and as supports for catalysts.^1–3 Due to their large surface areas, good chemical stability, and high thermal and electronic conductivity, porous carbon materials are often combined with other functional materials to extend their applications. In particular, porous carbon hosts with embedded metal or metal oxides are widely studied as promising electrode materials for supercapacitors.^4–10 As an important electrical device, supercapacitors suffer from their low energy density of a value typically less than 30 W h kg⁻¹.¹¹ Since energy density (E) is related to capacitance (C) and operation voltage (V) by E = 0.5CV,^3,12 developing electrode materials with a high specific capacitance is of great interest. Porous carbon materials, such as activated carbon, carbon nanotube and graphene, have been applied as electrode materials to improve the specific capacitance.^5,13,14 However their specific capacitances are still too low. Another type of electrode materials is transition metal oxides such as MnO₂, Fe₂O₃, Co₃O₄ and NiO. They demonstrate a very high specific capacitance, but suffer from their poor electrical conductivity, which limits the charge/discharge performance for high-power applications.¹⁵ By combining the advantages of porous carbon hosts and transition metal oxides, metal or metal oxide embedded porous carbon materials are able to achieve both a good conductivity and high capacitance. With this strategy, MnO₂, NiO, Co₃O₄, Fe₂O₃, Fe, Co and Ni were encapsulated in porous carbon hosts to form hybrid electrode materials and achieve a specific capacitance above 300 F g⁻¹ and a good cycle performance.^7,10,16–20

Such metal or metal oxide embedded porous carbon materials were mainly fabricated by incorporating transition metal oxide nanoparticles into mesoporous carbon hosts prepared through a soft-template or a hard-template method.²¹ However, preparation and removal of templates are tedious and it is sometimes extremely difficult for the metal oxide nanoparticles to disperse uniformly over the carbon host. To overcome these drawbacks, metal oxides from various precursors and porous carbon from carbon sources such as glucose could be synthesized simultaneously.^22–24 Colloidal carbon spheres have been prepared from glucose solution by hydrothermal treatment.²⁵ It was proposed that the formation of such carbon spheres starts with inter-molecular dehydration of glucose molecules, followed by a polymerization step, and then the carbon spheres are formed as a result of aromatization. Based on this mechanism, Fe₃O₄ nanoparticles embedded carbon spheres were synthesized using ferrous gluconate and glucose as the iron source and carbon source, respectively.²⁶ The Fe₃O₄ nanoparticles dispersed uniformly in the carbon spheres, but no obvious porous structures of the carbon matrix could be observed. As another example, graphitic carbon microspheres embedded with cobalt monoxide nanoparticles were prepared by hydrothermal treatment of cobalt gluconate and α-cyclodextrin (α-CD) which acted as the cobalt source and carbon source, respectively.²⁷ In this case, the cobalt monoxides nanoparticles were successfully dispersed in a porous carbon matrix. However, the specific capacitance was lower than 200 F g⁻¹ due to the low loading of cobalt monoxide.

Here, inspired by the synthesis of metal oxide–carbon materials from metal gluconate, we report a facile and template-free synthetic strategy (Fig. 1) to prepare metal nanoparticles embedded porous carbon microspheres (M-PCS). In this work, Fe-PCS, Co-PCS and Ni-PCS were synthesized and tested for their potential application in supercapacitors. Briefly, metal ions in porous colloidal carbon spheres (M-CCS, where M refers to Fe, Co, or Ni) are first synthesized by a hydrothermal reaction of a solution of MCl₂ and sodium gluconate. M-PCS are then obtained by applying a calcination process to M-CCS. It is proposed that the uniform dispersion of metal nanoparticles can be secured from the extensive interaction between metal ions from metal salt and –COO⁻ from gluconate. On the other hand, such interactions can significantly hinder inter-molecular dehydration, leading to a porous structure. As such, there are several advantages of using this synthetic strategy. First, the metal ions not only promote the formation of the porous structure, but also form well-dispersed metal nanoparticles after calcination. Thus, a porous carbon matrix and uniformly dispersed metal nanoparticles can be obtained simultaneously. This is the key aspect of the present synthetic strategy. Second, the synthesis is template-free and the precursors are inexpensive, rendering the synthetic process much simpler and cheaper than most of the reported methods. Third, the combination of high surface areas and uniformly dispersed oxidized metal nanoparticles from the resultant products can achieve an ultrahigh capacitance, making M-PCS a promising electrode material, particularly suitable for supercapacitors.

Fig. 1 Schematic illustration of the synthesis strategy for M-CCS and M-PCS.

Experimental section

Preparation of M-PCS

First, 6 mmol of MCl₂ and 12 mmol of sodium gluconate were dissolved in 60 mL of deionized water to form a clear aqueous solution. Then the solution was transferred to a Teflon vessel sealed in a steel autoclave. Upon hydrothermal treatment at 180 °C for 12 h, a brown solid precipitate (M-CCS) was recovered by filtration, washed with deionized water and ethanol for three times and dried in a vacuum oven at 80 °C overnight. Finally, the M-CCS was calcined in a tube furnace at 800 °C for 3 h under N₂ protection.

Material characterization

The crystallographic structures of the materials were determined by powder X-ray diffraction (XRD) (Philips PW1830) with Cu Kα radiation at a scanning rate of 10° min⁻¹ in the 2θ range from 10° to 90°. Functional groups of the materials were identified with Fourier transform infrared (FTIR) spectra generated by a Vertex 70 Hyperion 1000 (Bruker) spectrometer in a range of 400–4000 cm⁻¹. Oxidation states of the materials were measured with X-ray photoelectron spectroscopy (XPS) using an Axis Ultra DLD spectrometer. Nitrogen adsorption/desorption isotherms were recorded by a Micromeritics ASAP 2000 Physisorption BET instrument at 77 K. Scanning electron microscopy (SEM) images with energy-dispersive X-ray spectroscopy (EDS) were obtained on a JEOL JSM-7100F. Transmission electron microscopy (TEM) was performed on a JEOL JEM 2010 operating at an acceleration voltage of 100 kV.

Electrochemical measurement

Electrodes were obtained by coating slurry onto a nickel foam current collector (1 cm × 1 cm) and dried in an oven at 60 °C for 12 h. The slurry was prepared by mixing the active material, carbon black, and polyvinylidene fluoride (PVDF) in n-methyl-2-pyrrolidone (NMP) with a mass ratio of 75 : 15 : 10. The electrochemical tests of the electrodes were performed in a three-electrode cell where platinum foil and Hg/HgO electrode were used as the counter and reference electrodes, respectively. The electrochemical measurements were carried out at room temperature in 6 M KOH aqueous electrolyte solution by a CHI 760E electrochemical workstation at ambient temperature.

The specific capacitances of the electrodes were calculated from the galvanostatic discharge curves according to the following equation:

C = IΔt/mΔV (1)

where C (F g⁻¹) is the specific capacitance, I (A) the discharge current, ΔV (V) the potential change within the discharge time Δt (s), and m (g) the weight of active material on the electrode.

Results and discussion

An obviously porous morphology of M-CCS was observed in their scanning electron microscopic (SEM) images presented in Fig. 2a–c. For comparison, colloidal carbon spheres (G-CCS) were prepared from glucose by using the same hydrothermal process as for M-CCS. Significantly different from M-CCS, G-CCS showed a smooth surface with no obvious porous structure (Fig. 2d). To further study the formation of the porous structure of M-CCS, X-ray photoelectron spectroscopy (XPS) spectra, energy-dispersive X-ray spectroscopy (EDS) element mapping and Fourier transform infrared spectroscopy (FTIR) spectra were applied to analyze the oxidation states of metal compounds, element dispersion and functional groups of M-CCS.

Fig. 2 SEM images of (a) Fe-CCS from FeCl₂ + sodium gluconate, (b) Co-CCS from CoCl₂ + sodium gluconate, (c) Ni-CCS from NiCl₂ + sodium gluconate, and (d) G-CCS from glucose.

Metal ions in M-CCS were studied by XPS spectra (Fig. S1†). In Fe-CCS Fe ions with multiple oxidation sates were detected, while in Co-CCS and Ni-CCS Co²⁺ and Ni²⁺ were observed, respectively. Specifically, Fe²⁺ and Fe³⁺ were observed at 710.7 and 711.8 eV in Fe-CCS;²⁸ Co²⁺ was found at 781.3 eV;²⁹ Ni²⁺ correlated to different compounds were detected at 855.6, 856.1, and 857.2 eV.^30–32 These metal ions, as well as carbon and oxygen compounds, disperse uniformly in M-CCS, as revealed by the EDS element mapping (Fig. S2†).

Fourier transform infrared spectroscopy (FTIR) spectra were obtained for Fe-CCS, Co-CCS, Ni-CCS and G-CCS (Fig. 3). All samples show the presence of an abundance of –OH groups with a strong and broad absorption band at 3400 cm⁻¹,³³ but there are significant differences in their spectra. On the one hand, the absorption bands at around 1700 cm⁻¹ and 2900 cm⁻¹ assigned to –COOH and bands at around 1400 cm⁻¹ and 1600 cm⁻¹ assigned to –COO⁻ were significantly stronger in Fe-CCS, Co-CCS and Ni-CCS than in G-CCS, revealing the presence of more carboxylic acid groups and carboxylate anions in M-CCS than in G-CCS. One the other hand, absorption bands at 1024 cm⁻¹ and 1210 cm⁻¹, corresponding to aromatic ethers,³⁴ were much stronger in G-CCS than in Fe-CCS, Co-CCS and Ni-CCS, indicating the formation of more aromatic compounds in G-CCS than M-CCS. These findings imply that the formation of M-CCS should follow a route different from G-CCS. It is well known that gluconate anions deionized from sodium gluconate can form metal–gluconate complexes with various metal ions in aqueous solution.³⁵ As such, the interaction between M²⁺ and –COO⁻ in M–gluconate complexes disrupts the dehydration between hydroxyl groups and promotes the formation of side branches eventually leading to a porous structure. In addition, the metal compound guarantees a uniform dispersion of metal ions in M-CCS. As a result, porous colloidal carbon spheres with uniformly dispersed metal ions can be obtained after dehydration, polymerization, and aromatization in the hydrothermal reaction.

Fig. 3 FTIR spectra of Fe-CCS, Co-CCS, Ni-CCS, and G-CCS.

Fig. 4 illustrates the proposed formation mechanism of porous M-CCS during the hydrothermal process and M-PCS generated by a subsequent calcination. First, metal ion M²⁺ and gluconate anions form metal–gluconate complexes in the solution. Second, the metal–gluconate complexes undergo inter-molecular dehydration and metal-ion-linked oligosaccharides are formed. Third, as the dehydration process proceeds, more metal-ion-linked oligosaccharides form and linked to one another, leading to the formation of branchlike oligosaccharides in the solution, which is the so-called “polymerization”. Fourth, during the subsequent aromatization step, a carbon matrix forms, which may arise from cross-linking of multiple oligosaccharides to form aromatic compounds. However, due to the presence of abundant –COO–M groups, densification through usual aromatization is inhibited, and porous structures thus form. Afterwards, M-PCS can be obtained by calcination of M-CCS under N₂ protection. During calcination, metal nanoparticles are formed from the –COO–M groups.

Fig. 4 Schematic illustration of the formation mechanism of porous M-CCS and M-PCS.

The porous spherical structure of M-PCS was confirmed by the SEM images (Fig. 5a–c), while the uniform dispersion of metal nanoparticles in the porous carbon matrix by TEM images (Fig. 5d–f). The porous M-PCS were studied by nitrogen adsorption and desorption isotherms (Fig. 6a–c). The isotherms of all three samples show type I adsorption and type IV desorption patterns, indicating the presence of both micro- and meso-pores.³⁶ For comparison, isotherms of Fe-CCS, Co-CCS, and Ni-CCS were also measured. The surface area increased significantly after calcination, indicating that the corresponding decomposition of the –COO–M complexes led to the formation of mesopores inside the carbon matrix. The corresponding pore size distributions (Fig. 6d–f) were calculated from isotherms according to the Barrett–Joyner–Halenda theory.³⁷ The BET surface areas and pore diameters were calculated (Table S1†). Fe-PCS, Co-PCS, and Ni-PCS exhibit a high surface area of 295.3, 415.8, and 250.9 m² g⁻¹, respectively. While Fe-PCS and Fe-PCS exhibit a typical bimodal size distribution, Co-PCS shows multiple peaks at the small sizes of 3–60 nm.

Fig. 5 SEM images of (a) Fe-PCS, (b) Co-PCS, (c) Ni-PCS; TEM images of (d) Fe-PCS, (e) Co-PCS, (f) Ni-PCS.

Fig. 6 BET isotherms of (a) Fe-CCS and Fe-PCS, (b) Co-CCS and Co-PCS, (c) Ni-CCS and Ni-PCS; and pore-size-distribution curves of (d) Fe-PCS, (e) Co-PCS, (f) Ni-PCS.

The composition of Fe-PCS, Co-PCS, and Ni-PCS was analyzed by X-ray diffraction (XRD) (Fig. 7) and XPS (Fig. 8). Metallic Fe (JCPDS no. 06-0696), Co (JCPDS no. 15-0806) and Ni (JCPDS no. 65-2865) were all observed in the XRD patterns. However, in the XPS analyses of Fe-PCS (Fig. 8a), only Fe²⁺ and Fe³⁺ could be detected at around 710.7 and 711.8 eV.^28,38 In the XPS spectrum of Co-PCS (Fig. 8c), though Co²⁺ (780 eV) and Co³⁺ (781.3 eV) were observed together with metallic Co (778.6 eV),^39–42 although the signal from the metallic component was relatively weak. In contrast, a strong metallic component was observed in Ni-PCS (Fig. 8e) at 853.3 eV.^31,43 To further analyze the composition, from the surface of Fe-PCS, Co-PCS and Ni-PCS, a thickness of 20 nm was removed by sputtering (Fig. 8b, d and f) with argon. After sputtering, metallic phases could be observed in the XPS spectra of all three samples and the atomic ratio of M/O increased dramatically (Table S2†). For example, from 0.67 to 2.05 for Fe-PCS. It could be concluded that the nanoparticles embedded in the carbon matrix were metallic nanoparticles with a layer of metal oxides on their surface. According to the Ellingham diagram, Fe, Co, and Ni can all be reduced to their metallic forms from their metal oxides by carbon at >700 °C.⁴⁴ Thus metallic Fe, Co, and Ni nanoparticles were formed inside the carbon matrix during calcination in inert gas. After exposure in air for 12 h, the surface of metallic nanoparticles was oxidized to form a metal oxides layer.

Fig. 7 XRD patterns of Fe-PCS, Co-PCS and Ni-PCS.

Fig. 8 XPS spectra of (a) Fe-PCS, (b) sputtered Fe-PCS, (c) Co-PCS, (d) sputtered Co-PCS, (e) Ni-PCS and (f) sputtered Ni-PCS.

The prepared M-PCS were then used as electrode materials for supercapacitors. The highly porous carbon matrix, well-dispersed metal oxide nanoparticles and the conductive metal phases are desirable electrode material attributes for supercapacitor application. Their performance as electrode materials in supercapacitors were evaluated by cyclic voltammetry (CV) and galvanostatic charge–discharge measurement in a three-electrode configuration with 6 M KOH aqueous solution as the electrolyte (Fig. 9). The CV curve of Fe-PCS (Fig. 9a) was measured within −1.2 to 0 V. A couple of well-defined redox peaks were observed with the anodic peak at −0.608 V and the cathodic counterpart at −1.04 V, corresponding to the reversible conversion of Fe²⁺ and Fe³⁺ (Fe³⁺ ↔ Fe²⁺ + e⁻).⁴⁵ In addition, an anodic peak at −0.86 V was observed, which can be attributed to the oxidation of Fe (Fe → Fe²⁺ + 2e⁻). These anodic peaks correspond to the discharge plateaus observed at around −0.7 V and −0.6 V in the galvanostatic discharge curves of Fe-PCS at different current densities (Fig. 9b). CV measurement of Co-PCS was carried out within −1.3 to 0.6 V. The CV curve clearly showed three pairs of redox peaks at −0.996/−0.669 V, −0.38/0.11 V and 0.306/0.41 V, which were attributed to the redox reactions of Co(OH)₂/Co, CoOOH/CoO, and Co(OH)₂/CoOOH, respectively, as listed in Fig. 9c.^7,27 The galvanostatic discharge curves of Co-PCS (Fig. 9d) showed three discharge plateaus at around −0.7 V, 0 V and 0.4 V, which were consistent with the anodic peaks observed on the CV curve. A relatively weak anodic peak at −0.75 V was associated with the conversion of Co to CoOH (Co + OH⁻ → CoOH + e⁻).⁴⁶ The CV curve of Ni-PCS (Fig. 9e) was measured within −1.0 to 0.6 V. Since Ni-PCS consists of a large metallic component, CV curve of Ni foam was also carried out within the same potential range; no obvious peaks were found. Within the range of −1.0 to 0 V, the CV curve of Ni-PCS showed a typical quasi-rectangular shape of carbon-based electrode in supercapacitors (Fig. S3†), while within 0–0.6 V a couple of redox peaks were observed with the anodic peak at 0.47 V and the cathodic peak at 0.297 V, corresponding to the conversion between NiO and NiOOH (NiO + OH⁻ ↔ NiOOH + e⁻).²⁰ Corresponding to this redox reaction, a discharge plateau at 0.3 V was observed in the galvanostatic discharge curves of Ni-PCS (Fig. 9f) carried out at a variety of current densities.

Fig. 9 CV curves at a scan rate of 5 mV s⁻¹ and galvanostatic discharge curves of at different current density of (a and b) Fe-PCS, (c and d) Co-PCS, and (e and f) Ni-PCS.

Specific capacitance of M-PCS was calculated from the galvanostatic discharge curves (Fig. 10a). At a discharge current density of 1 A g⁻¹, Fe-PCS, Co-PCS, and Ni-PCS exhibited ultrahigh specific capacitances of 1308.7, 1380.9, and 543.6 F g⁻¹, respectively. The ultrahigh specific capacitance of M-PCS is promoted by the porous carbon matrix and the multi-step faradaic reactions of the metal nanoparticles. The cycling performance of M-PCS was studied at a high current density of 10 A g⁻¹ (Fig. S4†). Fe-PCS exhibited a capacitance increase at the first 100 cycles and a capacitance-retention of 98.5% after 1000 cycles. The increase of capacitance was a result of the activation of Fe nanoparticles (Fe → Fe²⁺ + 2e⁻). However, the cycling performance of Co-PCS was significantly different. Its capacitance firstly decreased to 59.2% in the first 300 cycles and 51.1% were retained after 1000 cycles. Similar to Fe-PCS, the capacitance of Ni-PCS increased first and then decayed slowly, with capacitance retention of 99.2% after 2000 cycles (Fig. 10b). The capacitance decrease was attributed to the aggregation of metal oxides nanoparticles and the subsequent loss of electrochemical activity. As can be seen in Fig. 6e, Co-PCS contains pores of 3–60 nm in size, providing void space for the aggregation of cobalt nanoparticles embedded. After 300 cycles, the aggregation slowed down and a further change of porosity was minimal.

Fig. 10 (a) Specific capacitance of Fe-PCS, Co-PCS and Ni-PCS at a variety of discharge current density, and (b) cycle performance of Ni-PCS at 10 A g⁻¹ for 2000 cycles.

Conclusions

In summary, a facile, template-free synthetic method has been developed for preparing porous carbon spheres embedded with metal nanoparticles (M-PCS) from sodium gluconate and metal chloride. Fe-PCS, Co-PCS, and Ni-PCS have been successfully prepared by this synthesis strategy and applied as electrode materials for supercapacitors. The metal ions not only play an important role in the formation of the porous structure, but also form uniformly dispersed metal nanoparticles after calcination. Benefiting from the porous carbon matrix and the uniform dispersion of metal nanoparticles, they offered high capacitances, high operating voltages, and thus high energy and power densities, making them promising alternative electrode materials for supercapacitors.

Notes and references

1. W. Z. Shen and W. B. Fan, J. Mater. Chem. A, 2013, 1, 999–1013.
2. W. Q. Tian, X. Y. Wu, K. X. Wang, Y. M. Jiang, J. F. Wang and J. S. Chen, RSC Adv., 2013, 3, 10823–10827.
3. T. Wang, Z. Dong, T. Fu, Y. C. Zhao, T. Wang, Y. Z. Wang, Y. Chen, B. H. Han and W. P. Ding, Chem. Commun., 2015, 51, 17712–17715.
4. D. W. Wang, F. Li and H. M. Cheng, J. Power Sources, 2008, 185, 1563–1568.
5. C. G. Liu, Z. N. Yu, D. Neff, A. Zhamu and B. Z. Jang, Nano Lett., 2010, 10, 4863–4868.
6. X. C. Dong, H. Xu, X. W. Wang, Y. X. Huang, M. B. Chan-Park, H. Zhang, L. H. Wang, W. Huang and P. Chen, ACS Nano, 2012, 6, 3206–3213.
7. L. Wang, D. L. Wang, X. C. Hu, J. S. Zhu and X. S. Liang, Electrochim. Acta, 2012, 76, 282–287.
8. J. Yan, Z. J. Fan, W. Sun, G. Q. Ning, T. Wei, Q. Zhang, R. F. Zhang, L. J. Zhi and F. Wei, Adv. Funct. Mater., 2012, 22, 2632–2641.
9. D. Q. Liu, X. Wang, X. B. Wang, W. Tian, J. W. Liu, C. Y. Zhi, D. Y. He, Y. Bando and D. Golberg, J. Mater. Chem. A, 2013, 1, 1952–1955.
10. C. L. Long, T. Wei, J. Yan, L. L. Jiang and Z. J. Fan, ACS Nano, 2013, 7, 11325–11332.
11. J. Jiang, Y. Y. Li, J. P. Liu, X. T. Huang, C. Z. Yuan and X. W. Lou, Adv. Mater., 2012, 24, 5166–5180.
12. Z. J. Fan, J. Yan, T. Wei, L. J. Zhi, G. Q. Ning, T. Y. Li and F. Wei, Adv. Funct. Mater., 2011, 21, 2366–2375.
13. J. Gamby, P. L. Taberna, P. Simon, J. F. Fauvarque and M. Chesneau, J. Power Sources, 2001, 101, 109–116.
14. D. S. Yu and L. M. Dai, J. Phys. Chem. Lett., 2010, 1, 467–470.
15. H. Jiang, J. Ma and C. Z. Li, Adv. Mater., 2012, 24, 4197–4202.
16. H. L. Li, L. X. Jiang, Q. L. Cheng, Y. He, V. Pavlinek, P. Saha and C. Z. Li, Electrochim. Acta, 2015, 164, 252–259.
17. R. Madhu, V. Veeramani, S. M. Chen, P. Veerakumar and S. B. Liu, Chem.–Eur. J., 2015, 21, 8200–8206.
18. J. Zhi, S. Deng, Y. X. Zhang, Y. F. Wang and A. G. Hu, J. Mater. Chem. A, 2013, 1, 3171–3176.
19. J. K. Hu, M. Noked, E. Gillette, F. D. Han, Z. Gui, C. S. Wang and S. B. Lee, J. Mater. Chem. A, 2015, 3, 21501–21510.
20. X. Z. Wu, W. Xing, L. Zhang, S. P. Zhuo, J. Zhou, G. Q. Wang and S. Z. Qiao, Powder Technol., 2012, 224, 162–167.
21. C. D. Liang, Z. J. Li and S. Dai, Angew. Chem., Int. Ed., 2008, 47, 3696–3717.
22. L. Wang, Y. J. Hao, Y. Zhao, Q. Y. Lai and X. Y. Xu, J. Solid State Chem., 2010, 183, 2576–2581.
23. L. Fan, L. Tang, H. F. Gong, Z. H. Yao and R. Guo, J. Mater. Chem., 2012, 22, 16376–16381.
24. Y. Wang, S. X. Xing, E. R. Zhang, J. T. Wei, H. Suo, C. Zhao and X. Zhao, J. Mater. Sci., 2012, 47, 2182–2187.
25. X. M. Sun and Y. D. Li, Angew. Chem., Int. Ed., 2004, 43, 597–601.
26. Y. Liu, Z. Y. Ren, Y. L. Wei, B. J. Jiang, S. S. Feng, L. Y. Zhang, W. B. Zhang and H. G. Fu, J. Mater. Chem., 2010, 20, 4802–4808.
27. Z. C. Yang, C. H. Tang, Y. Zhang, H. Gong, X. Li and J. Wang, Sci. Rep., 2013, 3, 2925.
28. T. Yamashita and P. Hayes, Appl. Surf. Sci., 2008, 254, 2441–2449.
29. T. Ivanova, A. Naumkin, A. Sidorov, I. Eremenko and M. Kiskin, J. Electron Spectrosc. Relat. Phenom., 2007, 156, 200–203.
30. L. J. M. Samuel, O. Grim and W. E. Swartz Jr, J. Am. Chem. Soc., 1972, 94, 5116–5117.
31. A. N. Mansour, Surf. Sci. Spectra, 1994, 3, 231.
32. A. N. Mansour, Surf. Sci. Spectra, 1994, 3, 239–246.
33. D. Lin-Vien, The Handbook of infrared and Raman characteristic frequencies of organic molecules, Academic Press, Boston, 1991.
34. R. M. Silverstein, F. X. Webster, D. J. Kiemle and D. L. Bryce, Spectrometric identification of organic compounds, Wiley, Hoboken, NJ, 8th edn, 2015.
35. D. T. Sawyer, Chem. Rev., 1964, 64, 633–643.
36. P. Schneider, Appl. Catal., A, 1995, 129, 157–165.
37. L. G. Joyner, E. P. Barrett and R. Skold, J. Am. Chem. Soc., 1951, 73, 3155–3158.
38. T. Yamashita and P. Hayes, Appl. Surf. Sci., 2009, 255, 8194.
39. N. S. Mcintyre and M. G. Cook, Anal. Chem., 1975, 47, 2208–2213.
40. A. B. Mandale, S. Badrinarayanan, S. K. Date and A. P. B. Sinha, J. Electron Spectrosc. Relat. Phenom., 1984, 33, 61–72.
41. C. A. Strydom and H. J. Strydom, Inorg. Chim. Acta, 1989, 159, 191–195.
42. K. J. K. Beng, J. Tan and P. M. A. Sherwood, J. Am. Chem. Soc., 1991, 113, 855–861.
43. G. R. Conner, J. Vac. Sci. Technol., 1978, 15, 343–347.
44. C. E. A. Shanahan, Chem. Br., 1983, 19, 387.
45. K. Ujimine and A. Tsutsumi, J. Power Sources, 2006, 160, 1431–1435.
46. G. Jerkiewicz and P. Marcus, Proceedings of the Symposium on Electrochemical Surface Science of Hydrogen Adsorption and Absorption, Electrochemical Society, Pennington, NJ, 1997.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra20748k

---