Remarkably durable platinum cluster supported on multi-walled carbon nanotubes with high performance in an anhydrous polymer electrolyte fuel cell

Abstract

Reducing platinum (Pt) usage in the polymer electrolyte fuel cells (PEFCs) has become one of the main issues in the global commercialization of PEFCs. In this work, we describe a facile and scalable method to deposit Pt clusters (1.2 nm) on multi-walled carbon nanotubes (MWNTs) by the aid of NaOH in a reduction process. The electrocatalyst loses 50% of the initial electrochemical surface area (ECSA) after 200 000 potential cycles from 1.0 to 1.5 V vs. RHE, which is 20 times higher compared to commercial CB/Pt. The mass power density of the Pt cluster electrocatalyst measured under 120 °C without any humidification reaches 1320 mW mg_Pt⁻¹, which is 6.7 times higher compared to that of commercial CB/Pt. To the best of our knowledge, the mass power density of our electrocatalyst is one of the highest values measured in high-temperature PEFCs.

---

Introduction

The global energy crisis forces us to develop new sustainable energy sources. Polymer electrolyte fuel cells (PEFCs) attract much attention due to the zero emission and high-energy conversion efficiency when hydrogen is fed to the anode side.^1–3 Platinum is recognized as the most efficient catalyst for the oxygen reduction reaction (ORR) and hydrogen oxidation reaction (HOR),⁴ which are the cathodic and anodic reactions in the real fuel cell operation, respectively. Meanwhile, the commercialization of the PEFCs is still blocked by two main issues,⁵ namely, (i) the high Pt usage in the PEFC resulting in high cost of the PEFCs since the Pt is expensive; (ii) the low durability of the electrocatalyst shortening the life of the PEFCs.⁶

Many reports have described that combining some inexpensive metal (M) with Pt to form alloy^7–10 or core–shell^11–14 structure electrocatalysts was an effective way to decrease the dosage of the Pt, however, these metals suffer from a serious problem in the real fuel cell operation, namely, the dissolution in acidic medium.^15–19 Another way is to deposit Pt cluster on the carbon supporting materials to enlarge the electrochemical surface area (ECSA). In recent publications, the Pt cluster was synthesized by two main routes, (i) the colloidal method,^20,21 in which the protectors should be eliminated from the resulting electrocatalyst since the capping polymer poisons or decrease the catalytic activity of the electrocatalyst. This synthetic route is rather complicated and difficult,²² (ii) impregnation,^23–25 in which the Pt precursor is reduction by gas or liquid. However, this method is quite difficult to homogeneously load Pt cluster on the carbon supporting materials, especially carbon nanotubes (CNTs) and graphene due to the few binding sites for Pt. As well known, the oxidation process introduces the –COOH/–OH groups on the CNTs and graphene for anchoring Pt cluster,^22,26,27 while, the oxidation procedure damages the unique structure of CNTs and graphene resulting in lower electronic conductivity and durability.^16,28–30 Thus, loading of Pt cluster on the non-oxidized CNTs is still challenging task for fabrication of an electrocatalyst with high performance and durability to reduce the Pt usage in PEFCs.

Here, we described a facile way to deposit Pt cluster on the pristine MWNTs after wrapping with poly[2,2′-(2,6-pyridine)-5,5′-bibenzimidazole] (PyPBI) as shown in Fig. 1. Compared to our previous publications,^16,28,31 in this study, we focused on the deposition of Pt cluster on MWNT/PyPBI in order to reduce the Pt usage and possess higher durability. PyPBI is an efficient dispersant for MWNTs and provide anchor site for Pt due to the π–π interaction and Pt–N bonding, respectively.^32,33 The additional NaOH was proved to control the Pt-NP size as reported by us previously, in which we studied the Pt size effect on fuel cell performance and Pt stability with similar Pt loading on MWNT/PyPBI.³¹ The high amount of NaOH was expected to decrease Pt size to sub-nanometre. The newly fabricated electrocatalyst was compared to commercial CB/Pt in high-temperature PEFCs since the high temperature facilitates the electrochemical reaction and water management.³⁴

Fig. 1 Schematic illumination of the deposition of Pt cluster on MWNT/PyPBI.

Experimental

Materials

Perchloric acid (HClO₄, 70%), N,N-dimethylacetamide (DMAc), hydrogen hexachloroplatinate hexahydrate (H₂PtCl₆·6H₂O), 85% phosphoric acid (PA), ethylene glycol (EG) and isopropanol were purchased from Sinopharm Chemical Reagent Co., Ltd. The MWNTs with ∼20 nm diameter were offered Timesnano Corp. Commercial CB/Pt electrocatalyst was purchased from Tanaka Kikinzoku Kogyo K. K. Nafion solution (5 wt%) was purchased from Sigma-Aldrich. Poly[2,2′-(2,6-phenyl)-5,5′-bibenzimidazole] (PBI) and poly[2,2′-(2,6-pyridine)-5,5′-bibenzimidazole] (PyPBI) were synthesized according to previously reported methods.³⁵

Synthesis of MWNT/PyPBI/Pt cluster

MWNTs (20 mg) was dispersed in DMAc (40 mL) by sonication for 1 h to which PyPBI (10 mg) dissolved in DMAc (10 mg) was added, then sonicated for another 2 h to obtain homogeneous solution. The solution was then filtered, washed by DMAc for several times to remove any remained PyPBI followed by drying overnight under vacuum at 80 °C. The deposition of Pt cluster was carried out by the reduction of H₂PtCl₆·6H₂O in EG aq. (EG : H₂O = 3/2, v/v) in which NaOH solution was added. First, MWNT/PyPBI (10 mg) was dispersed in 30 mL EG aq. to which H₂PtCl₆·6H₂O (24 mg) and 1.5 mL NaOH solution (15 mg mL⁻¹) were added, then refluxed at 140 °C for 6 h under stable N₂ flow. By filtration of the dispersion, the electrocatalyst was obtained and dried in oven at 80 °C to remove any remained solvent.

Characterization

The X-ray photoelectron spectroscopy (XPS) spectra were measured using AXIS-ULTRA^DLD (Shimadzu) instrument. The pressure in the XPS chamber was kept at 10⁻⁹ Pa or lower during the measurements. In the calibration process, the binding energy (BE) of the core level C_1s peak was fixed at 284.5 eV as standard. TGA curves were measured (EXSTAR 6000, Seiko Inc.) with the heating rate of 10 °C min⁻¹ under stable air flow. The TEM images were taken by JEM-2010 (JEOL, acceleration voltage of 120 kV) electron microscope.

Electrochemical measurements

The electrochemical analysis was performed using a glass carbon electrode (GCE) with a conventional three-electrode system in a vessel at 25 °C. A GCE with a geometric surface area of 0.1257 cm² (D = 4 mm) was used as the working electrode. A Pt wire and an Ag/AgCl were used as the counter and reference electrodes, respectively. The potential of the electrode was controlled by CHI-640e potentiostat. The electrocatalyst ink was prepared as follows. The synthesized electrocatalyst (1.0 mg) was dispersed in isopropanol aqueous solution (2.0 mL, v/v = 4/1) to form a homogeneous solution. A portion of the solution was casted on GCE to form a thin electrocatalyst layer. Finally, the casted GCE on the electrode was air-dried at room temperature. The cyclic voltammetry (CV) of the electrocatalyst was carried out at the scan rate of 50 mV s⁻¹ in N₂-saturated 0.1 M HClO₄ electrolyte to determine the electrochemical surface area (ECSA) value. All the potentials were transformed to the reference hydrogen electrode (RHE). Oxygen reduction reaction (ORR) measurements were performed in O₂-saturated 0.1 M HClO₄ electrolyte using AFMSRCE instruments (PHYCHEMi, CO. Ltd).

Durability testing

The carbon corrosion was examined based on the protocol from the Fuel Cell Commercialization Conference of Japan (FCCJ)^36,37 (measured in N₂-saturated 0.1 M HClO₄ electrolyte at 25 °C), in which the potential was kept at 1 V vs. RHE for 30 s and increased to 1.5 V vs. RHE with the scan rate of 0.5 V s⁻¹, then decreased to 1 V vs. RHE. After every 1000 cycles, CV was measured to determine the ECSA (see the ESI, Fig. S1†).

Gas diffusion electrode (GDE) fabrication

The electrocatalyst ink was prepared by dispersing in a 50 mL isopropanol aqueous solution (v/v = 4/1) by sonication for 1 h. The gas diffusion electrode (GDE) was obtained by filtration of resultant electrocatalyst ink on carbon gas diffusion layer (GDL). The obtained GDE was dried overnight under vacuum at room temperature to remove residual solvent.

Fabrication of PA doped-PBI membrane

PBI (200 mg) was dissolved in DMAc (10 mL) by stirring for 1 h. After complete dissolution, the solution was casted on glass plate and heated at 100 °C for 5 h to remove DMAc. The resultant PBI membrane was peeled off from the glass substrate. And the PBI membrane was dipped in 85 wt% phosphoric acid solution for 5 days before using. The membrane thickness of the obtained membrane was determined to be ∼25 μm. The doping level calculated by the weight change of the dry membrane upon doping was ∼5 H₃PO₄ molecules/repeat unit of the PBI.

Membrane electrode assembly (MEA) fabrication and FC test

The MEA was prepared by hot pressing the GDE and the PBI membrane under 2 MPa at 120 °C for 30 s. The geometric area of the MEA was 1 cm². The FC performance of the assembled MEA was evaluated at 120 °C without any external humidification using a computer-controlled fuel cell test system (Model 890e, Scribner Associate, Inc.). The in situ CV was carried out by flowing hydrogen and nitrogen at the anode and cathode, respectively. The working and reference electrodes were collected with cathode and anode, respectively. The polarization and the power density curves were measured under the atmospheric pressure by flowing dry hydrogen (flow rate; 100 mL min⁻¹) and dry air (flow rate; 200 mL min⁻¹) at the anode and the cathode, respectively.

Results and discussion

After the synthesis of the electrocatalyst, the XPS was used to detect the successful deposition of the Pt-NPs on MWNT/PyPBI (for survey scan, see ESI, Fig. S2a†). The N_1s peak coming from the PyPBI was appeared at 400 eV (Fig. 2a) and the two Pt_4f peaks were clearly observed. From the deconvoluted Pt_4f peaks, the ratio of the Pt(0) : Pt(II) : Pt(IV) was 56.5 : 17.6 : 25.9, suggesting the abundant product was Pt(0), which was similar to that of the commercial CB/Pt.³⁸ The TGA measurements were carried out to determine the Pt loading on MWNT/PyPBI shown in Fig. 3a. Before Pt loading, the TGA of MWNT/PyPBI was carried out and the weight loss started at 500 °C was assigned to be PyPBI, which was 7 wt% in MWNT/PyPBI (blue line, Fig. 3a). The weight residue of MWNT/PyPBI/Pt cluster after heating to 900 °C was assigned to the Pt amount in the electrocatalyst, which was 10 wt% (red line, Fig. 3a). Thus, the MWNT, PyPBI and Pt contents in MWNT/PyPBI/Pt cluster were 83.7 wt%, 6.3 wt% and 10 wt%, respectively. Also the commercial CB/Pt was measured by TGA to confirm the Pt amount, which was 38.2 wt%. From the TEM images shown in Fig. 3b, the Pt cluster was homogeneously deposited on the MWNT assisted by the PyPBI, which anchors the Pt-NPs via Pt–N bonding.^18,33,39 And the diameter of the Pt-NPs was calculated to be 1.2 ± 0.1 nm from the HR-TEM image as shown in Fig. 3c. The smaller Pt cluster was due to the additional OH⁻ in the Pt deposition process, which greatly affects the species and size of the Pt(IV) complexes.⁴⁰ By increasing the NaOH amount in the reduction process, almost no Pt-NP was loaded on MWNT/PyPBI shown in Fig. S3† since the high OH⁻ concentration increases the ionic strength and suppressed the Pt loading.^41,42

Fig. 2 XPS narrow scans of the commercial CB/Pt (black line) and MWNT/PyPBI/Pt cluster in N_1s region (a) and Pt_4f of MWNT/PyPBI/Pt cluster (b) in which the deconvolution curve was shown.

Fig. 3 (a) The TGA curves of commercial CB/Pt (black line), MWNT/PyPBI (blue line) and MWNT/PyPBI/Pt cluster (red line) under stable air flow (100 mL min⁻¹) with heating rate of 10 °C min⁻¹. TEM images of MWNT/PyPBI/Pt cluster with low (b) and high (c) magnifications.

The cyclic voltammetry is of importance to determine the electrochemical surface area (ECSA) of the electrocatalyst, evaluating from the hydrogen adsorption peak from 0.1 to 0.4 V vs. RHE based on the following equation:

ECSA = Q_H/210 × (Pt loading on electrode) (1)

where Q_H is the charge exchanged during the electro-adsorption of hydrogen on Pt.^43–45

Here, the commercial CB/Pt (60 m² g_Pt⁻¹) was used as control sample and the ECSA of MWNT/PyPBI/Pt cluster was 72.8 m² g_Pt⁻¹. The long-term durability was carried out continuously, which was simplified to measure in N₂-saturated HClO₄ electrolyte based on FCCJ protocol and evaluated by the ECSA loss after potential cycling from 1.0 to 1.5 V vs. RHE, in which the carbon corrosion (C + 2H₂O → CO₂ + 4H⁺ + 4e⁻, 0.207 V vs. RHE) was accelerated resulting in Pt detachment and aggregation. The durability was evaluated from the loss in ECSA during the potential cycling and terminated when the ECSA decreases to half of the initial value. From Fig. 4a, the relative hydroquinone–quinone (HQ/Q) redox peak were observed at 0.55 V vs. RHE and the ECSA lost 50% after 10 000 potential cycles for commercial CB/Pt, in sharp contrast, MWNT/PyPBI/Pt cluster showed the same loss after 200 000 potential cycles (Fig. 4b), suggesting that MWNT/PyPBI/Pt cluster possessed 20 times higher durability compared to commercial CB/Pt (Fig. 4c). In order to emphasize the advantage of MWNT, CB/PyPBI/Pt cluster was prepared according to the same procedure and measured the durability as shown in Fig. 4c. The ECSA remained 42% after 40 000 potential cycles, suggesting that the usage of MWNT as Pt cluster support obtained 5 times higher durability compared to the utilization of CB. It should be noted that MWNT/PyPBI/Pt cluster showed HQ/Q peak after durability test indicated that the MWNTs were corroded after 200 000 potential cycles evidenced by the XPS narrow scan in C_1s region shown in Fig. S2b.† However, the higher durability was evaluated by the ECSA loss even though both samples showed HQ/Q peaks. After the durability test, the electrocatalysts were collected from the electrodes by sonication in 2-propanol to check the morphology by ex situ TEM. As shown in Fig. 4d, the Pt-NPs were grown to 3.8 ± 1.2 nm due to the migration of the Pt-NPs and the Pt growth rate was 1.5 × 10⁻⁵ nm per cycle, which was much slower than commercial CB/Pt (1.1 × 10⁻⁴ nm per cycle, Fig. S4†), indicating the Pt-NPs were more stable on MWNT/PyPBI. As well known, the theoretical ECSA loss can be calculated from the Pt size change before and after durability test from the equation: theoretical surface area = 6/ρd, where ρ is the density of the Pt metal, d is the mean diameter of the Pt-NP from TEM).¹⁶ Based on the theoretical calculation, the ECSA loss should be 22% for commercial CB/Pt due to Pt aggregation, while; the experimental ECSA loss was 46% indicating that Pt aggregation and detachment contributes equally to the ECSA loss during the potential cycling from 1.0 to 1.5 V vs. RHE. The theoretical ECSA of MWNT/PyPBI/Pt cluster was 68%, while; the experimental value was 58% indicating the Pt aggregation led to all the ECSA loss and Pt detachment is negligible. Thus, the higher durability of MWNT/PyPBI/Pt cluster attributed to the smaller Pt size and usage of the pristine MWNTs, which suppressed the Pt aggregation and carbon corrosion. So the commercial CB/Pt and MWNT/PyPBI/Pt cluster showed different mechanism for ECSA degradation under carbon corrosion as schematically shown in Fig. 5. Also the XPS spectrum showed an apparent N_1s peak after durability test indicating the PyPBI was stabilized on the MWNT shown in Fig. S2.†

Fig. 4 CV curves of the commercial CB/Pt (a), MWNT/PyPBI/Pt cluster (b) and CB/PyPBI/Pt cluster (c) before (solid line) and after (dotted line) durability test measured in N₂-saturated 0.1 M HClO₄ electrolyte. (d) Normalized ECSA of the commercial CB/Pt (black line), CB/PyPBI/Pt cluster (blue line) and MWNT/PyPBI/Pt cluster (red line) as a function of the numbers of potential cycling from 1.0 to 1.5 V vs. RHE. (e) TEM image of MWNT/PyPBI/Pt cluster after durability test and the histogram was inserted.

Fig. 5 Schematic illumination of different mechanism of ECSA degradation under 1.0–1.5 V vs. RHE for commercial CB/Pt and MWNT/PyPBI/Pt cluster.

ORR is the cathodic reaction in the real fuel application and sluggish compared to the hydrogen oxidation reaction (HOR); thus, the ORR kinetic is of importance for the fuel cell performance.⁴⁶ It should be noted that the Pt loading on the electrode for MWNT/PyPBI/Pt cluster was 4 times lower than of the commercial CB/Pt due to the lower Pt amount in MWNT/PyPBI/Pt cluster (10 wt%) in order to make a same thickness of the catalyst film on the electrode since the thickness of the catalyst film affects the oxygen diffusion and Pt utilization efficiency. As shown in Fig. 6a, MWNT/PyPBI/Pt cluster showed 25 mV negative shift in the mixed kinetic-diffusion control region (0.8–0.9 V vs. RHE) compared to commercial CB/Pt, while, the onset potential was comparable to the commercial CB/Pt as shown in the insert. As shown in the Fig. 6b, the mass current density (normalized with Pt loading) of MWNT/PyPBI/Pt cluster was super high at the same applied potential and the onset potential was higher due to the very low Pt loading on the electrode. The numbers of the electron transferred during ORR (Fig. S5†) collected from Levich–koutecky plots (Fig. 6c) was 3.9 and 3.7 for MWNT/PyPBI/Pt cluster and CB/Pt, respectively, suggesting that MWNT/PyPBI/Pt cluster showed an idea ORR and lower H₂O₂ generation (0.85%, Fig. 6d), which was corrosive for the carbon supporting materials leading to lower durability. The lower H₂O₂ generation would be due to the strong binding between Pt cluster and H₂O₂ because of higher surface energy of small Pt size, which is important for further reduction to H₂O by Pt.⁴⁷ The higher ORR activity of MWNT/PyPBI/Pt cluster attributed to the smaller Pt-NP size with enough space for oxygen species because Watanabe et al. pointed out that a highly dispersed Pt-NP on the support with maintenance of a sufficient inter-particle distance is highly critical for the ORR test since the oxygen molecule is supplied to Pt-NP by spherical diffusion.⁴⁸

Fig. 6 (a) ORR curves of the commercial CB/Pt (black line) and MWNT/PyPBI/Pt cluster (red line) measured in O₂-saturated 0.1 M HClO₄ with 1600 rpm. (b) ORR curves with normalized Pt loading of the commercial CB/Pt (black line) and MWNT/PyPBI/Pt cluster (red line) measured in O₂-saturated 0.1 M HClO₄ with 1600 rpm. Levich–koutecky plots (c) and H₂O₂ generation (d) from the ORR for the commercial CB/Pt (black line) and MWNT/PyPBI/Pt cluster (red line).

Fuel cell test was carried out under 120 °C without any humidification to lower the risk of carbon monoxide (CO) poisoning, faster electrochemical kinetics and easy water management. As well known, the gas diffusion and proton conductivity are highly depended on the thickness of the catalyst layer, affecting the fuel cell performance. The MEA fabricated from commercial CB/Pt was controlled to be 0.45 mg_Pt cm⁻² and the electrocatalyst loading was calculated to be 1.18 mg cm⁻² (Pt content in commercial CB/Pt: 38.2 wt%, Fig. 3a). In order to make a comparable electrocatalyst loading, the Pt loading in MEA fabricated from MWNT/PyPBI/Pt cluster was calculated to be 0.1 mg_Pt cm⁻² since the Pt content was 10 wt% in MWNT/PyPBI/Pt cluster. From Fig. 7a, MWNT/PyPBI/Pt cluster showed 132 mW cm⁻², which was higher compared to commercial CB/Pt (89 mW cm⁻²) with almost similar resistance for two electrocatalysts during the I–V measurements shown in Fig. 7c. The mass power density of MWNT/PyPBI/Pt cluster reached 1320 mW mg_Pt⁻¹, which was 6.7 and 3.5 times higher than those of commercial CB/Pt (Fig. 7b) and MWNT/PyPBI/Pt-3 nm (Pt: 47 wt%) reported previously, respectively.⁴⁹ The mass power density was one of the highest values among the recent publications.²⁸ Also the mass power density was much higher compared to our previous reported MWNT/PyPBI/Pt-1.7 nm (Pt: 47 wt%) due to sufficient inter-particle distance facilitating the gas diffusion to active site.³¹ The in situ CV was carried out to determine the Pt utilization efficiency in the catalyst layer shown in Fig. 7d and the ECSA of MEA fabricated from MWNT/PyPBI/Pt cluster was 50.3 m² g_Pt⁻¹, which was twice higher compared to that of commercial CB/Pt (20.5 m² g_Pt⁻¹). It is should be noted that the ECSA value from single cell is normally lower compared to the value from half-cell since the gas diffusion and proton transfer significantly affect the ECSA of single cell; while, the electrolyte penetrates the electrocatalyst film easily in half-cell. And the ECSA from single cell is highly important for the evaluation of the real Pt utilization efficiency. The higher ECSA indicated that MWNT/PyPBI/Pt cluster has a higher Pt utilization efficiency resulting in the higher mass power density. The lower Pt loading amount was useful for the reducing the cost of the fuel cells.

Fig. 7 (a) I–V and power density curves of the MEA fabricated from the commercial CB/Pt (black line) and MWNT/PyPBI/Pt cluster (red line). (b) I–V and mass power density curves of the MEA fabricated from the commercial CB/Pt (black line) and MWNT/PyPBI/Pt cluster (red line) as a function of Pt mass activity. (c) Relative resistances of commercial CB/Pt (black line) and MWNT/PyPBI/Pt cluster (red line) during the I–V measurements. (d) In situ CV of commercial CB/Pt (black line) and MWNT/PyPBI/Pt cluster (red line) under 120 °C with nitrogen and hydrogen flow for cathode and anode, respectively.

Conclusions

In conclusion, we have successfully deposited Pt cluster (1.2 nm) on the MWNT/PyPBI assisted by the additional NaOH in the EG/water solution. The as-prepared electrocatalyst showed higher ORR activity and 20 times higher durability compared to the commercial CB/Pt. The fuel cell test under 120 °C showed that the mass power density of MWNT/PyPBI/Pt cluster was 6.7 times higher than that of commercial CB/Pt. The durable MWNT/PyPBI/Pt cluster with high performance is utilizable in the real fuel cell application.

Acknowledgements

The authors gratefully acknowledge support of the National Natural Science Foundation of China (21603197), Natural Science Foundation of Hubei Province of China (No. 2016CFB181) and Fundamental Research Funds for the Central University, China University of Geosciences, Wuhan (No. CUG150620).

Notes and references

1. B. E. Hayden, Acc. Chem. Res., 2013, 46, 1858–1866.
2. T. Fujigaya and N. Nakashima, Adv. Mater., 2013, 25, 1666–1681.
3. R. Borup, J. Meyers, B. Pivovar, Y. S. Kim, R. Mukundan, N. Garland, D. Myers, M. Wilson, F. Garzon, D. Wood, P. Zelenay, K. More, K. Stroh, T. Zawodzinski, J. Boncella, J. E. McGrath, M. Inaba, K. Miyatake, M. Hori, K. Ota, Z. Ogumi, S. Miyata, A. Nishikata, Z. Siroma, Y. Uchimoto, K. Yasuda, K.-i. Kimijima and N. Iwashita, Chem. Rev., 2007, 107, 3904–3951.
4. A. Chen and P. Holt-Hindle, Chem. Rev., 2010, 110, 3767–3804.
5. A. Iiyama, Y. Tabuchi, A. Ohma, S. Sugawara and K. Shinohara, ECS Trans., 2014, 64, 11–17.
6. Z. Zhang, J. Liu, J. Gu, L. Su and L. Cheng, Energy Environ. Sci., 2014, 7, 2535–2558.
7. I. E. L. Stephens, A. S. Bondarenko, U. Grønbjerg, J. Rossmeisl and I. Chorkendorff, Energy Environ. Sci., 2012, 5, 6744–6762.
8. X. Li, Q. Huang, Z. Zou, B. Xia and H. Yang, Electrochim. Acta, 2008, 53, 6662–6667.
9. H. Gharibi, F. Golmohammadi and M. Kheirmand, Electrochim. Acta, 2013, 89, 212–221.
10. M. E. Scofield, C. Koenigsmann, L. Wang, H. Liu and S. S. Wong, Energy Environ. Sci., 2015, 8, 350–363.
11. R. Choi, S.-I. Choi, C. H. Choi, K. M. Nam, S. I. Woo, J. T. Park and S. W. Han, Chem.–Eur. J., 2013, 19, 8190–8198.
12. D. Wang, H. L. Xin, H. Wang, Y. Yu, E. Rus, D. A. Muller, F. J. DiSalvo and H. D. Abruña, Chem. Mater., 2012, 24, 2274–2281.
13. T. Cochell, W. Li and A. Manthiram, J. Phys. Chem. C, 2013, 117, 3865–3873.
14. X. Sun, D. Li, Y. Ding, W. Zhu, S. Guo, Z. L. Wang and S. Sun, J. Am. Chem. Soc., 2014, 136, 5745–5749.
15. Z. Yang, M. R. Berber and N. Nakashima, J. Mater. Chem. A, 2014, 2, 18875–18880.
16. S. I. Gorelsky, A. B. P. Lever and M. Ebadi, Coord. Chem. Rev., 2002, 230, 97–105.
17. Z. Yang, I. H. Hafez, M. R. Berber and N. Nakashima, ChemCatChem, 2015, 7, 808–813.
18. Z. Yang and N. Nakashima, ChemCatChem, 2016, 8, 600–606.
19. Z. Yang, C. Kim, S. Hirata, T. Fujigaya and N. Nakashima, ACS Appl. Mater. Interfaces, 2015, 7, 15885–15891.
20. J. Chang, L. Feng, C. Liu and W. Xing, ChemSusChem, 2015, 8, 3340–3347.
21. X. Jin, B. He, J. Miao, J. yuan, Q. Zhang and L. Niu, Carbon, 2012, 50, 3083–3091.
22. C. Zhou, F. Peng, H. Wang, H. Yu, C. Peng and J. Yang, Electrochem. Commun., 2010, 12, 1210–1213.
23. Y. Cheng, P. K. Shen, M. Saunders and S. P. Jiang, Electrochim. Acta, 2015, 177, 217–226.
24. A. Santasalo-Aarnio, M. Borghei, I. V. Anoshkin, A. G. Nasibulin, E. I. Kauppinen, V. Ruiz and T. Kallio, Int. J. Hydrogen Energy, 2012, 37, 3415–3424.
25. W. Chen, J. Ji, X. Feng, X. Duan, G. Qian, P. Li, X. Zhou, D. Chen and W. Yuan, J. Am. Chem. Soc., 2014, 136, 16736–16739.
26. S.-H. Liu, W.-Y. Yu, C.-H. Chen, A.-Y. Lo, B.-J. Hwang, S.-H. Chien and S.-B. Liu, Chem. Mater., 2008, 20, 1622–1628.
27. R. Q. Long and R. T. Yang, J. Am. Chem. Soc., 2001, 123, 2058–2059.
28. Z. Yang, T. Fujigaya and N. Nakashima, J. Mater. Chem. A, 2015, 3, 14318–14324.
29. H. Wang, S. Lu, Y. Zhang, F. Lan, X. Lu and Y. Xiang, J. Mater. Chem. A, 2015, 3, 6282–6285.
30. H. Wang, S. Lu, Y. Zhang, F. Lan, J. Shang and Y. Xiang, ChemPlusChem, 2016, 81, 172–175.
31. V. Selvaraj and M. Alagar, Electrochem. Commun., 2007, 9, 1145–1153.
32. M. Okamoto, T. Fujigaya and N. Nakashima, Small, 2009, 5, 735–740.
33. T. Fujigaya, M. Okamoto and N. Nakashima, Carbon, 2009, 47, 3227–3232.
34. Q. Li, J. O. Jensen, R. F. Savinell and N. J. Bjerrum, Prog. Polym. Sci., 2009, 34, 449–477.
35. L. Xiao, H. Zhang, T. Jana, E. Scanlon, R. Chen, E. W. Choe, L. S. Ramanathan, S. Yu and B. C. Benicewicz, Fuel Cells, 2005, 5, 287–295.
36. A. Ohma, K. Shinohara, A. Iiyama, T. Yoshida and A. Daimaru, ECS Trans., 2011, 41, 775–784.
37. J. Speder, A. Zana, I. Spanos, J. J. K. Kirkensgaard, K. Mortensen, M. Hanzlik and M. Arenz, J. Power Sources, 2014, 261, 14–22.
38. T. Maiyalagan, Int. J. Hydrogen Energy, 2009, 34, 2874–2879.
39. L. J. Grove, J. M. Rennekamp, H. Jude and W. B. Connick, J. Am. Chem. Soc., 2004, 126, 1594–1595.
40. B. Fang, N. K. Chaudhari, M. S. Kim, H. K. Jung and J. S. Yu, J. Am. Chem. Soc., 2009, 131, 15330–15338.
41. Y. Xu, X. Xie, J. Guo, S. Wang, Y. Wang and V. K. Mathur, J. Power Sources, 2006, 162, 132–140.
42. S. T. Briskeby, M. Tsypkin, R. Tunold and S. Sunde, RSC Adv., 2014, 4, 44185–44192.
43. Z. Yang and N. Nakashima, Sci. Rep., 2015, 5, 12236.
44. D. Wang, H. L. Xin, R. Hovden, H. Wang, Y. Yu, D. A. Muller, F. J. Disalvo and H. D. Abruña, Nat. Mater., 2013, 12, 81–87.
45. D. M. Han, Z. P. Guo, Z. W. Zhao, R. Zeng, Y. Z. Meng, D. Shu and H. K. Liu, J. Power Sources, 2008, 184, 361–369.
46. S. Yu and B. C. Benicewicz, Macromolecules, 2009, 42, 8640–8648.
47. E. A. Franceschini, M. M. Bruno, F. A. Viva, F. J. Williams, M. Jobbágy and H. R. Corti, Electrochim. Acta, 2012, 71, 173–180.
48. M. Lee, M. Uchida, K. Okaya, H. Uchida and M. Watanabe, Electrochemistry, 2011, 79, 381–387.
49. E. Gyenge, M. Atwan and D. Northwood, J. Electrochem. Soc., 2006, 153, A150–A158.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra19487g

---