Synthesis of Bi₂WO₆ photoanode on transparent conducting oxide substrate with low onset potential for solar water splitting

Abstract

To enable water splitting in photoelectrochemical cells, the conduction-band (CB) and valence-band edges must straddle the hydrogen-reduction and water-oxidation potentials; however, the CB edge potential of many photoanodic semiconductors is insufficient. Here, we demonstrate the nanostructured Bi₂WO₆ for photoanodic application, which has a high, flat-band potential of 0.15 V vs. RHE. Single-phase, orthorhombic Bi₂WO₆ nano-structures were successfully grown on an FTO substrate through a two-step hydrothermal synthesis with subsequent thermal treatment at 600 °C. The synthesized Bi₂WO₆ photoanode shows a lower onset potential and improved photocurrents which were enhanced further by a factor of three by coating the surface with Co-Pi. The low onset potential of the Bi₂WO₆/Co-Pi photoanode results in a higher operating photocurrent in a p/n photodiode cell combined with a p-Si/n-Si/Pt photocathode.

---

1. Introduction

The economic, political, and environmental issues related to fossil fuels are expediting the global interest regarding the conversion of solar energy to fuel to aid in the development of renewable energy resources.^1,2 In this method of fuel-production, solar energy is converted and stored in chemical bonds by transferring photo-excited charge carriers to chemical oxidation/reduction reactions. For example, in a photoelectrochemical (PEC) water-splitting system, a semiconductor material absorbs sunlight, generates charge carriers, and then catalyzes hydrogen/oxygen evolution reactions using the excited charge carriers. Because the semiconductors play multiple roles, ideal semiconductors must satisfy several requirements,^3,4 and the preparation of the photoanodes composed of the ideal semiconductors is critical to maximizing their photocatalytic performances.⁵

Metal-oxide semiconductors, such as TiO₂,⁶ WO₃,^7,8 and Fe₂O₃,^9–11 have been intensively studied for PEC water-splitting due to their high chemical stability and synthetic feasibility. However, the low conduction-band (CB) edge positions of metal oxides relative to the hydrogen reduction potential are disadvantageous for water splitting. For example, the valence band (VB) of WO₃ has sufficiently low energy (∼3.0 V vs. NHE) to oxidize water, but its low CB edge (∼0.3 V vs. NHE) thermodynamically prohibits spontaneous hydrogen reduction.^12,13 This drawback encourages research into the development of new types of metal-oxide semiconductors with higher CB edges.

Bi₂WO₆, which is a ternary oxide, has been reported to have a CB that is closer to the hydrogen reduction potential and higher than that of WO₃.¹⁴ The bandgap of Bi₂WO₆ is ∼2.7–2.8 eV, which is slightly larger than that of WO₃ (∼2.6–2.7 eV); this larger bandgap boosts the CB potential.^15,16 Therefore, a Bi₂WO₆ photoanode can split water without or with a small applied bias potential, and its combination with another p-type semiconductor, such as a p/n photodiode system, enables overall water splitting. The earlier PEC studies regarding Bi₂WO₆ focused on the photodegradation of environmental pollutants and powder-type approaches that did not involve the fabrication of an electrode.^17–23 Only few recent studies regarding PEC water-splitting using an electrode-based approach,^14,24 which can measure improved PEC properties, including the onset potential for water oxidation and a flat-band potential related to the CB position, have been reported.

In this report, we describe the hydrothermal synthesis of nanostructured Bi₂WO₆ directly onto an FTO substrate for photoanode applications. The intermediate WO₃·0.33H₂O nanorod film was converted to Bi₂WO₆ during the second hydrothermal reaction. The photocurrents and onset potentials of the anodic photocurrent are discussed with respect to the annealing conditions and deposition of cobalt phosphate (Co-Pi) as an oxygen evolution catalyst. After loading Co-Pi onto the Bi₂WO₆ electrode surface, the water oxidation dynamics improved and the onset potential (0.15 V vs. RHE) decreased down to a flat-band potential, which was measured using the Mott–Schottky measurement method.

2. Experimental section

A two-step hydrothermal reaction was applied to synthesize Bi₂WO₆. Firstly, tungsten oxide nanorods were prepared directly on an FTO substrate⁸ and then converted to Bi₂WO₆ in the presence of a Bi precursor. Ammonium paratungstate pentahydrate ((NH₄)₁₀W₁₂O₄₁·5H₂O; 1 g) (Alfa Aesar, 99.999%) was dissolved in distilled water (93 mL) and stirred for 1 h before the addition of HCl (3 mL, Sigma Aldrich, 37.5%). After stirring for 30 min, 4 mL of H₂O₂ (Junsei, 30%) was added and the mixture was stirred for an additional 1.5 h. This tungsten precursor solution was transferred to a Teflon liner and a clean FTO substrate was placed face-down in the solution for the duration of the hydrothermal reaction at 160 °C (4 h). The reaction resulted in the growth of WO₃·0.33H₂O nanorods on the FTO side, which was then rinsed with water. The second hydrothermal reaction was conducted in a 0.1 mM Bi(NO₃)₃·5H₂O (Sigma Aldrich, 98%) aqueous solution at 160 °C for 24 h with the as-prepared WO₃·0.33H₂O film or with the 500 °C for 2 h annealed film. After the second reaction, the film was rinsed and annealed at 400–800 °C for 4 h. Scheme 1 shows the overall synthesis.

Scheme 1 Bi₂WO₆ synthetic strategy with crystal structures obtained at each step.

Co-Pi was deposited on the Bi₂WO₆ surface by electrodeposition using a three-electrode system.^25,26 The electrolytic bath comprised 0.1 mM CoCl₂·6H₂O (Sigma Aldrich, 98%) in a 0.1 M potassium phosphate buffer solution, and 1.1 V vs. Ag/AgCl was applied for 15 min using a potentiostat (Iviumstat, Ivium Technology).

PEC performances were measured in a 0.5 M Na₂SO₄ aqueous solution using linear sweep voltammetry at a rate of 20 mV s⁻¹ under simulated one-sun illumination (100 mW cm⁻²), which was generated from a solar simulator (ABET, Sun 2000) equipped with 300 W Xe lamp and 1.5 air mass filter. The measurements were carried out using a three-electrode system with an Ag/AgCl reference electrode, a Pt counter electrode, and the synthesized films as the working electrode. Mott–Schottky measurements were performed to measure the flat-band potential of the synthesized Bi₂WO₆ electrode in a 0.5 M Na₂SO₄ aqueous solution. The potential was scanned from 0.0 to 0.6 V vs. RHE (potential amplitude of 50 mV) with two different frequencies, i.e., 0.5 and 1.0 kHz, to minimize the effects of the surface states.

An operating current of the overall water splitting was measured by linking the Bi₂WO₆ photoanode with a silicon photocathode under simulated sunlight illumination. p-Si photocathode was prepared from Si (111) substrate (boron doped, 0.1–0.3 Ω cm resistivity), and Pt was deposited on the Si electrode by electrochemical deposition under illumination with 0.04 M H₂PtCl₆·5.6H₂O (Kojima Chemicals Co. Ltd., 98%) precursor solution.²⁷ To shift the onset potential of Si photocathode toward anodic direction, p-Si/n-Si junction was formed by doping with phosphorous on the surface of the p-Si substrate as Boettcher et al. reported.²⁸ Spin on phosphorous dopant (Filmtronics, P509) solution was coated on the HF cleaned p-Si substrate by spin coating at 2000 rpm followed by baking at 150 °C, and annealing at 950 °C for 1 h. After the annealing, spin on dopant was removed by etching in buffered HF solution (Aldrich) before Pt electrodeposition. I–V curves of the p-Si/Pt and p-Si/n-Si/Pt photocathodes were measured similarly to Bi₂WO₆ photoanode. The only difference was that linear sweep voltammetry was applied to cathodic direction (from 0.8 V to −0.5 V vs. RHE).

Scanning electron microscopy (SEM; S-4100, Hitachi) was used to investigate the surface morphology of the hydrothermally synthesized films, and the crystal structures were characterized using X-ray diffraction (XRD, Schimadzu X). The absorbance was measured using UV-vis spectrometer (Varian, Cary 100).

3. Result and discussion

The morphologies of the nanostructured films varied during each step of hydrothermal growth and annealing according to the SEM images (Fig. 1). Rod-like nanostructures were observed after the first hydrothermal growth (Fig. 1a); these transformed to a flower-like morphology during the second hydrothermal reaction with the Bi precursor and annealing at 400 °C (Fig. 1b). A smoother surface was obtained with annealing at 600 °C (Fig. 1c and d) and EDS elemental analysis confirmed a Bi/W atomic ratio of 2 : 1. (Fig. S1†) On the other hand, annealing the as-prepared film in the first hydrothermal growth (Fig. 1a) at 500 °C prior to the second reaction results in smoother rod-like structures (Fig. 1e); these nanostructures break down to smaller sizes after the second hydrothermal reaction and subsequent annealing at 400 °C (Fig. 1f).

Fig. 1 SEM images of (a) as-grown nanorod arrays after the first hydrothermal reaction (orthorhombic WO₃·0.33H₂O) and samples annealed (b) at 400 °C, and (c) and (d) at 600 °C after the second hydrothermal synthesis (orthorhombic Bi₂WO₆) ((d) shows a lower magnification). SEM images of samples annealed (e) at 500 °C after the first hydrothermal reaction (monoclinic WO₃) and (f) at 400 °C (mixture of monoclinic WO₃ and orthorhombic Bi₂WO₆) after the second hydrothermal synthesis using sample (e).

Although the morphologies of the films are different, the films adhere stably to the FTO substrates even after the second hydrothermal growth and subsequent annealing. During the first hydrothermal reaction, the in situ formation of a seed layer on the FTO substrate is evident initially, i.e., after less than 1 h, and the rod-like structures grow on top of the seed layer, which results in good adhesion.⁸

An attempt to grow Bi₂WO₆ via a one-pot hydrothermal reaction failed to result in the formation of only orthorhombic WO₃ H₂O film on the FTO substrate (ESI Fig. S3 and S4†) due to the unbalanced reactivity of the W and Bi precursors. SEM-EDS showed that a negligible amount of Bi is included in the final film although the morphology and crystal structure of the hydrated tungsten oxide are affected by the presence of the Bi precursor. This unsuccessful synthesis of Bi₂WO₆ guided us to design a two-step hydrothermal synthesis to introduce Bi into the crystal structure of tungsten oxide during a second step of growth at 160 °C for 24 h.

An orthorhombic Bi₂WO₆ nanostructured film can be synthesized directly onto the FTO substrate using a two-step hydrothermal approach in which an orthorhombic WO₃·0.33H₂O nanorod film is successfully converted, as evident from the XRD spectra (Fig. 2). After the first hydrothermal growth with only the tungsten precursor, the sample has an orthorhombic WO₃·0.33H₂O crystal structure⁸ (Fig. 2a), which converts to a monoclinic WO₃ structure upon annealing at 500 °C in air (Fig. 2b). Complete conversion to Bi₂WO₆ was only observed (Fig. 2d) when the second hydrothermal reaction was performed using the as-prepared sample; the crystal structure of the orthorhombic Bi₂WO₆ did not change even with the post-annealing (Fig. S5†). Meanwhile, a mixture of WO₃ and Bi₂WO₆ was obtained (Fig. 2c) when the second hydrothermal reaction was carried out using the annealed monoclinic WO₃ nanorod film, which indicates the partial inclusion of Bi into the tungsten oxide crystal structure. The relatively weak interactions between adjacent layers in the orthorhombic WO₃·0.33H₂O crystal structure make it more susceptible to attack by Bi³⁺ than the thermodynamically stable monoclinic WO₃.⁸ We used pure samples of Bi₂WO₆ resulting from the two-step hydrothermal synthesis for the following photocatalytic analyses.

Fig. 2 XRD spectra of the prepared sample at each step showing complete conversion of WO₃·0.33H₂O to Bi₂WO₆ and partial conversion of monoclinic WO₃ to WO₃/Bi₂WO₆ during the second hydrothermal reaction.

The photocurrent of the Bi₂WO₆ photoanodes strongly depends on the post-annealing conditions. The sample annealed at 600 °C has the highest photocurrent (Fig. 3a). The prepared Bi₂WO₆ photoanodes were applied to photo-oxidation of water in a 0.5 M Na₂SO₄ aqueous solution post annealing at 500, 600, and 800 °C, respectively. The sample was annealed at 800 °C for only 10 min because of the thermal stability of the FTO substrate, while the others were annealed for 1 h in air. The solid lines in Fig. 3a represent the photocurrent under simulated sunlight conditions, while the dotted lines show the current in the dark. The photocurrents (at 1.23 V vs. RHE) of the Bi₂WO₆ photoanodes annealed at 500, 600, and 800 °C are 5.55, 11.3, and 8.5 μA cm⁻², respectively; thus, the photocurrent of the sample annealed at 600 °C is almost twice that of the sample annealed at 500 °C. In addition, a dramatic change in the slope of the photocurrent versus the bias potential (Fig. 3a) is observed when Bi₂WO₆ is annealed above 600 °C. The morphology changes or the enhanced crystalline structures may contribute to the higher activities of the high temperature-annealed samples. Although the FTO substrate is known to be stable up to 650 °C,²⁹ the effect of the thermal stability of FTO on the reproducibility of the photocurrent is a concern after annealing at high temperatures such as 800 °C, and decreased electrical conductivity of the FTO substrate can cause a lower photocurrent despite a lower onset potential.

Fig. 3 (a) Photoanodic current densities vs. RHE with varying annealing temperatures (solid line: under simulated sunlight illumination, dotted line: in dark), and (b) enhanced photocurrent densities upon Co-Pi deposition. (Red line shows photocurrent under the chopped illumination).

The photocurrent of the Bi₂WO₆ electrode was enhanced by a factor of three (∼30 μA cm⁻²) upon deposition of Co-Pi, which is known to be an efficient electro-catalyst for the oxygen evolution reaction (OER). In particular, the photocurrent was improved under positively biased conditions with a high fill-factor, as shown in the I–V curves. (Fig. 3b) A previous study regarding Bi₂WO₆ also reported a poor photocurrent level in the order of few tens of μA cm⁻²,^15,16,24 which is low considering the reported bandgap of ∼2.7–2.9 eV and our UV-vis absorption data (Fig. 4b). Therefore, the low photocurrent levels are suspected to be due to poor charge collection efficiency which can be caused by insufficient charge transport within the semiconductor material and/or insufficient charge transfer across the semiconductor and electrolyte interface. The former is related to the material properties, such as charge-carrier mobility, depletion width, etc., while the latter is also related to the catalytic activity on the surface of the semiconductor. The influence of the OER catalyst loading on the photocurrent is significant if the oxidation activity on the semiconductor is poor. Co-Pi has been demonstrated to be a good OER catalyst material on a BiVO₄ photoanode in which poor water-oxidation activities obstruct charge separation at the semiconductor surface.^26,30 Here, we demonstrate that a Co-Pi layer can also enhance water oxidation activity of the Bi₂WO₆ photoanode, meaning the poor water oxidation activity is one of the reasons to show low level of photocurrent with Bi₂WO₆. The electrodeposited Co-Pi layer is expected to be effectively coated onto the nanostructured Bi₂WO₆ surface and can facilitate charge transfer at the Bi₂WO₆/electrolyte interface.

Fig. 4 (a) Characterization of a flat-band potential of the Bi₂WO₆ photoanode by Mott–Schottky plots, and (b) UV-vis absorption spectra of various photoanode samples.

The onset potential of the Co-Pi/Bi₂WO₆ photoanode is around +0.15 V vs. RHE (Fig. 3b), which is much lower than those of Fe₂O₃ (+0.8 V vs. RHE) and WO₃ (+0.5 V vs. RHE).³ Its flat-band potential (E_fb) was determined to be 0.15 V vs. RHE through Mott–Schottky measurements (Fig. 4a) at 0.5 and 1.0 kHz. Using the following equation, E_fb can be determined from the x-intercept value:

where C_sc is the space charge capacitance, ε is the dielectric constant of the semiconductor, ε₀ is the permittivity of free space, N_D is the carrier concentration, A is the surface area of the electrode, E is the applied potential, k is the Boltzmann constant, and T is the temperature. The onset potential is well matched with the observed flat-band potential when Co-Pi is deposited. The bandgap of the prepared Bi₂WO₆ is ∼2.7 eV according to UV-vis measurements (Fig. 4b), which is slightly wider than that of monoclinic WO₃ (∼2.6 eV); these results are consistent with those of previous studies¹⁴ and suggest that the CB positions of Bi₂WO₆ are at a more positive potential than those of monoclinic WO₃. The shift direction is preferable for overall water splitting.

Notably, lowering the onset potential of the Bi₂WO₆ photoanode is promising for the realization of spontaneous water splitting in a composite PEC system, such as a p/n photodiode PEC cell or photovoltage cell-combined tandem PEC cell.^31,32 In these PEC systems, the operating photocurrent is determined from the overlap of the I–V curves, which means that the currents (or charge flows) must be balanced on the cathodic and anodic sides. Therefore, lowering the onset potential is important to obtain high operating photocurrents because most of the onset potentials of n-type semiconductors are generally too high to cross the I–V curves of p-type semiconductors. For example, the onset potential of a p-type Si photocathode,³³ which is one of the popular semiconductors for hydrogen producing photo-cathodes, is known to be +0.3 V vs. RHE. Its onset potential can shift anodically up to 0.55 V vs. RHE by forming p/n⁺ Si junction,²⁸ but it is too cathodic to couple effectively with photoanodes such as WO₃ and Fe₂O₃. The overlap of the IV curves either does not occur or occurs at a negligible current level. Fig. 5 shows the I–V curves of p-Si/Pt and p-Si/n-Si/Pt along with Bi₂WO₆/Co-Pi and WO₃ photoanode. We confirmed that the measured operating photocurrent using the combination of a photocathode and a photoanode was well matched with the cross point in the I–V curves.

Fig. 5 (a) I–V curves of photocathodes and photoanodes which show lowering the onset potential of the photoanode contributes the enhanced operating current with the combination of a silicon photocathode. (b) Schematic diagram of a p/n photodiode PEC cell composed of a silicon photocathode and a Bi₂WO₆ photoanode illustrating the overall water splitting.

When Bi₂WO₆/Co-Pi was combined with p-Si/Pt, the operating current was 3 μA cm⁻² without any external bias potentials while no photocurrent was observed when WO₃ was combined. Again, when p-Si/n-Si/Pt was linked with Bi₂WO₆/Co-Pi, the photocurrent was enhanced to 20 μA cm⁻² which is a higher value compared to the case using WO₃. Bi₂WO₆/Co-Pi photoanode in the present work resulted in the higher operating current although its maximum photocurrent level is much lower than that of WO₃ photoanode,⁸ which shows the importance of controlling onset potentials of the photoelectrodes in p/n PEC cells. Here, lowering the onset potential of Bi₂WO₆ provides the opportunity for efficient spontaneous water splitting in a composite PEC system. The poor photocurrent is expected to increase upon improving the electrical properties of Bi₂WO₆.

4. Conclusions

We demonstrated the growth of Bi₂WO₆ nanostructures on an FTO substrate via a two-step hydrothermal synthesis during which WO₃ 0.33 H₂O nanorods on the FTO substrate convert to orthorhombic Bi₂WO₆. The effect of the annealing temperatures on the photoanodic activities of Bi₂WO₆ was monitored, and 600 °C is the optimum annealing temperature. A three-fold improvement in the photocurrent occurs upon coating the surface of the nanostructured Bi₂WO₆ photoanode with an OER catalyst, i.e., Co-Pi. In the present work, hydrothermally prepared Bi₂WO₆ has a lower onset potential (0.15 V vs. RHE) than other metal-oxide semiconductors, such as WO₃ and Fe₂O₃ as a result of shifting the CB to a higher level. Although the photocurrent level of Bi₂WO₆ requires further improvement, its low onset potential is promising for the realization of spontaneous water splitting in composite systems, such as p/n photodiodes or PV cell-combined tandem cells.

Acknowledgements

This research was funded by the institutional program of Korea Institute of Science and Technology (KIST).

Notes and references

1. T. A. Faunce, W. Lubitz, A. W. Rutherford, D. MacFarlane, G. F. Moore, P. Yang, D. G. Nocera, T. A. Moore, D. H. Gregory, S. Fukuzumi, K. B. Yoon, F. A. Armstrong, M. R. Wasielewski and S. Styring, Energy Environ. Sci., 2013, 6, 695.
2. Solar Fuels and Artificial Photosynthesis: Science and innovation to change our future energy options, 1st UK Solar to Fuels Symposium, London, UK, 2012.
3. M. Gratzel, Nature, 2001, 414, 338.
4. Z. Chen, T. F. Jaramillo, T. G. Deutsch, A. Kleiman-Shwarsctein, A. J. Forman, N. Gaillard, R. Garland, K. Takanabe, C. Heske, M. Sunkara, E. W. McFarland, K. Domen, E. L. Miller, J. A. Turner and H. N. Dinh, J. Mater. Res., 2010, 25, 3.
5. I. S. Cho, Z. Chen, A. J. Forman, D. R. Kim, P. M. Rao, T. F. Jaramillo and X. Zheng, Nano Lett., 2011, 11, 4978.
6. M. Ni, M. K. H. Leung, D. Y. C. Leung and K. Sumathy, Renewable Sustainable Energy Rev., 2007, 11, 401.
7. M. Higashi, R. Abe, T. Takata and K. Domen, Chem. Mater., 2009, 21, 1543.
8. S. S. Kalanur, Y. J. Hwang, S. Y. Chae and O. S. Joo, J. Mater. Chem. A, 2013, 1, 3479.
9. S. C. Warren, K. Voïtchovsky, H. Dotan, C. M. Leroy, M. Cornuz, F. Stellacci, C. Hébert, A. Rothschild and M. Grätzel, Nat. Mater., 2013, 12, 842.
10. K. Sivula, F. Le Formal and M. Grätzel, ChemSusChem, 2011, 4, 432.
11. K. Sivula, F. L. Formal and M. Grätzel, Chem. Mater., 2009, 21, 2862.
12. K. Maeda and K. Domen, J. Phys. Chem. Lett., 2010, 1, 2655.
13. B. D. Alexander, P. J. Kulesza, I. Rutkowska, R. Solarska and J. Augustynski, J. Mater. Chem., 2008, 18, 2298.
14. J. C. Hill and K.-S. Choi, J. Mater. Chem. A, 2013, 1, 5006.
15. L. Zhang, C. Baumanis, L. Robben, T. Kandiel and D. Bahnemann, Small, 2011, 7, 2714.
16. L. Zhang and D. Bahnemann, ChemSusChem, 2013, 6, 283.
17. Y. Zhang, N. Zhang, Z.-R. Tang and Y.-J. Xu, Chem. Sci., 2013, 4, 1820.
18. J. Tian, Y. Sang, G. Yu, H. Jiang, X. Mu and H. Liu, Adv. Mater., 2013, 25, 5075.
19. A. Kudo and S. Hijii, Chem. Lett., 1999, 28, 1103.
20. X. Ding, Z. H. Ai and L. Z. Zhang, J. Hazard. Mater., 2012, 239, 233.
21. L. W. Zhang, Y. J. Wang, H. Y. Cheng, W. Q. Yao and Y. F. Zhu, Adv. Mater., 2009, 21, 1286.
22. S. M. Sun, W. Z. Wang and L. Zhang, J. Phys. Chem. C, 2012, 116, 19413.
23. D. K. Ma, S. M. Huang, W. X. Chen, S. W. Hu, F. F. Shi and K. L. Fan, J. Phys. Chem. C, 2009, 113, 4369.
24. C. Ng, A. Iwase, Y. H. Ng and R. Amal, J. Phys. Chem. Lett., 2012, 3, 913.
25. M. W. Kanan and D. G. Nocera, Science, 2008, 321, 1072.
26. F. F. Abdi and R. van de Krol, J. Phys. Chem. C, 2012, 116, 9398.
27. Y. L. Kawamura, T. Sakka and Y. H. Ogata, J. Electrochem. Soc., 2005, 152, C701.
28. S. W. Boettcher, E. L. Warren, M. C. Putnam, E. A. Santori, D. Turner-Evans, M. D. Kelzenberg, M. G. Walter, J. R. McKone, B. S. Brunschwig, H. A. Atwater and N. S. Lewis, J. Am. Chem. Soc., 2011, 133, 1216.
29. K. Goto, T. Kawashima and N. Tanabe, Sol. Energy Mater. Sol. Cells, 2006, 90, 3251.
30. F. F. Abdi, L. Han, A. H. M. Smets, M. Zeman, B. Dam and R. van de Krol, Nat. Commun., 2013, 4, 1.
31. M. G. Walter, E. L. Warren, J. R. McKone, S. W. Boettcher, Q. Mi, E. A. Santori and N. S. Lewis, Chem. Rev., 2010, 110, 6446.
32. M. F. Weber and M. J. Dignam, J. Electrochem. Soc., 1984, 131, 1258.
33. Y. Hou, B. L. Abrams, P. C. K. Vesborg, M. E. Björketun, K. Herbst, L. Bech, A. M. Setti, C. D. Damsgaard, T. Pedersen, O. Hansen, J. Rossmeisl, S. Dahl, J. K. Nørskov and I. Chorkendorff, Nat. Mater., 2011, 10, 434.

---

Footnote

† Electronic supplementary information (ESI) available: These might include comments relevant to but not central to the matter under discussion, limited experimental and spectral data, and crystallographic data. See DOI: 10.1039/c4ra02868f

---