Enhanced electrocatalytic performance of perovskite supported iron oxide nanoparticles for oxygen reduction reaction

Abstract

Perovskite oxide based materials have drawn considerable amount of attention as non-precious, non-noble cathode catalysts in oxygen reduction reactions for fuel cell and battery applications. High efficiency of perovskite catalysts often suffers due to instability and poor conductivity issues. Thus, the integration of these perovskite materials with other active catalysts is be of great interest. In this article, a series of new nanocomposites of perovskite–iron oxide with different molar ratios has been developed by hydrothermal method. We have shown that double perovskite calcium copper titanate CaCu₃Ti₄O₁₂ (CCTO) supported iron oxide (Fe₂O₃) catalysts exhibit excellent activity as cathode catalyst for oxygen reduction reactions in fuel cells. Three different composites of CaCu₃Ti₄O₁₂ (CCTO) with Fe₂O₃ were prepared and the composites were analysed by various structural and microstructural characterization methods, such as XRD and FESEM. Microstructural investigations revealed cubic shaped iron oxide nanoparticles decorated over CCTO particles. Electrochemical analysis showed that the performance of composites improved drastically with increase in the amount of CCTO. Kinetic studies by rotating ring disc electrode method revealed that the composites exhibit four electron pathway for oxygen reduction reaction and generate hydrogen peroxide as an intermediate. This work demonstrates a new type of efficient metal oxide–perovskite composite for oxygen redution catalysis.

---

Introduction

With the concerns of depletion of fossil fuels and petroleum products, there is an immediate requirement for renewable energy generation and storage resources.¹ Fuel cells are one among those promising alternative resources. The heart of any fuel cell is its electrocatalyst, which should be cost effective and energy efficient for carrying out electrochemical energy conversion processes, such as oxygen evolution reaction (OER) and oxygen reduction reaction (ORR).² Oxygen reduction reaction (ORR) is seen as one of the potential alternatives for the generation of clean energy, as it involves breaking and conversion of chemical energy stored in double bonds of oxygen molecules to generate electrical energy. The major drawbacks for the successful commercialization of fuel cells are mainly due to the facts of high cost, large over potential (≥300 mV) and slow kinetics of the state of art platinum and platinum-based alloy catalysts for ORR reactions.³ In addition, gradual degradation of Pt catalysts in alkaline medium leads to particle dissolution and aggregation due to the surface oxide formation.⁴ Hence, it is important to replace high cost platinum catalyst by low cost non platinum group transition metals and metal oxides.

In this respect, transition metals (like iron, cobalt and nickel) and metal oxides have withdrawn considerable attention as alternative catalysts because of their efficient catalytic activities, low cost, high abundance and better durability.^5–7 It is reported that a synergistic combination between carbon and oxide materials plays an important role for the enhanced catalytic behaviour of the oxide–carbon thin film electrode.⁸ It was proposed that the relatively faster electron-transfer reaction was facilitated by cooperation between the oxide and carbon support.⁹ In these cases, in the first step, oxygen gets reduced to peroxide ion by two electron pathways at carbon sites, followed by further reduction to hydroxyl ion at neighbouring oxide sites. There is also the other school of thoughts, where an outer sphere electron transfer (OSET) mechanism for ORR activity by iron oxide (Fe₂O₃) in dilute alkaline medium had been proposed.¹⁰ In case of iron oxide nanoparticles, most prevalent mechanism is likely to be the outer sphere electron-transfer mechanism due to the oxide layer passivation on electrode in aqueous alkaline medium.¹¹ By increasing the number of edges in Fe metal, it is also feasible to improve the number of active sites as oxygen gets adsorbed faster on edges.¹² The core of improvement towards oxygen reduction activity is to control the interaction of oxygen molecule with the transition metal ion on the oxide surface and also to moderate the adsorption and desorption kinetics of the different intermediates of oxygen reduction reaction.¹³ As non-precious, non-noble metal oxide, iron oxide shows good activity towards oxygen reduction reaction and is indeed a potential non-platinum group catalyst.⁵ However, the activity of iron oxide is quite low as compared to Pt. There are several attempts to improve the activity of the iron oxide based electrocatalysts either using N-doped carbon support materials⁷ or making composites of iron oxide with other metals.¹⁴ In this study, we have synthesized a novel composite of Fe₂O₃ nanoparticles with perovskite material using a simple and low cost hydrothermal method. Recently, ABO₃ perovskite materials have been received considerable attention for ORR applications.^15–17 Matsumoto et al. gave a hypothesis that formation and filling of a σ* bond between e_g orbitals of transition metals and molecular orbital of oxygen can greatly influence the ORR activity in perovskite metal oxides.¹⁸ Similarly, Suntivich et al. have proposed that the “activity descriptor” which controls the ORR activity in transition metal oxides is greatly influenced by the degree of σ*-antibonding (e_g) orbital filling of metal ions on the surface. According to this, the ORR activity will be greatly influenced by the e_g filling maximum to be 1. It has also been observed that presence of two different transition metal cations in the perovskite system often influences the ORR activity due to the mixed valence state of transition metals during the reduction–oxidation process.⁸

Using this principle of enhancement of ORR activity, we have prepared the composites of iron oxide with CaCu₃Ti₄O₁₂ (CCTO) and a perovskite material and systematically studied the influence of CCTO towards the activity of iron oxide. CCTO is a cubic double perovskite material with B site occupied by Ti⁴⁺, and A site occupied by Ca²⁺ and Cu²⁺ metal ions. Rotating disk electrode (RDE) and rotating ring disk electrode (RRDE) measurements also show that iron oxide as well as perovskite material exhibit 4e⁻ pathway. In our observation, with the increase of CCTO in the composites, the ORR reactivity of the composites enhances and the over potential of oxygen reduction also decreases. We have done systematic electrochemical studies to understand the mechanistic pathways followed by the composites and the role of CCTO in the improvement of the catalysts. X-ray diffraction (XRD) and field emission scanning electron microscopy (FESEM) analysis were also carried out to study the crystal structure and morphology of the composites, respectively.

Experimental

Materials synthesis

All chemical reagents are of analytical grade purchased from Sigma Aldrich. For the synthesis of 1 : 1 molar ratio of Fe₂O₃–CaCu₃Ti₄O₁₂ composite, referred as Fe₂O₃–CCTO 1 : 1, 0.86 M (10 mL) ferric chloride hexahydrate (FeCl₃·6H₂O) solution was mixed with 10 mL triethylamine. The mixture was vigorously stirred for 20 minutes. Later, 0.86 M (10 mL) solution of CCTO was added to the reaction mixture and was stirred continuously till a homogeneous solution was prepared. The solution was transferred into a Teflon-lined stainless steel autoclave at 180 °C for 18 hours. After the autoclave was cooled naturally to room temperature, the solution was centrifuged with distilled water, washed with absolute ethanol thrice and kept for air drying at room temperature. The same procedure was followed for preparing 1 : 0.5 and 1 : 0.33 molar ratio of Fe₂O₃–CCTO composite (referred as Fe₂O₃–CCTO 1 : 0.5 and Fe₂O₃–CCTO 1 : 0.33, respectively). Bare Fe₂O₃ and CCTO were also synthesized for comparing its activity with the composite, following modified procedures reported earlier.^19,20

Material's characterization

The X-ray diffraction (XRD) study of the as-prepared samples was carried out by X'Pert diffractometer (Philips, Netherlands) using Cu Kα radiation (λ = 0.154 nm) with a range of 2θ = 10–90° and 0.02 step size to study the crystal structure and phase constitution of the samples. Field emission scanning electron microscopy (FESEM, FEI-Technai SEM Sirion) was performed to observe the morphology of the composite. Energy dispersive spectroscopy (EDS) analysis was also carried out to determine the composition and elemental mapping of the samples.

Electrochemical measurement

Electrochemical measurements for ORR were carried out on an Autolab electrochemical workstation (Metrohm) at room temperature in a three-electrode system using 0.1 M KOH as electrolyte. Rotating glassy carbon disk electrode (RDE) with diameter of 5 mm was used as working electrode, Ag/AgCl as reference electrode and platinum wire as counter electrode. To prepare the ink for working electrode, 5 mg CCTO–Fe₂O₃ and 10 mg Vulcan XC-72 were dispersed in 2.5 mL distilled water and 2.5 mL isopropyl alcohol. 300 μL Nafion solution (5 wt%) was also added. The entire solution was ultra-sonicated for an hour and 20 μL of the suspension was drop-casted over the glassy carbon substrate. Rotating ring disk electrode (RRDE) measurements were also done at room temperature on Pine Instruments setup connected to Autolab bi-potentiostat.

Results and discussion

The synthesis of iron oxide nanoparticles was carried out using wet chemical route, where triethylamine acted as reducing agent as well as capping agent. The proposed mechanism for this reaction is given below:¹⁹

FeCl₃ + 3H₂O + 3(CH₃–CH₂)₃–N → Fe(OH)₃ + 3[(CH₃–CH₂)₃–NH⁺]Cl⁻ (1)

Ferric hydroxide (Fe(OH)₃) generated in the intermediate step got transformed first into α-FeOOH in hydrothermal environment and then into α-Fe₂O₃ through thermal decomposition

2Fe(OH)₃ → 2α-FeOOH + 2H₂O → α-Fe₂O₃ + H₂O (2)

Different compositions of iron oxide–CCTO nanocomposites were prepared by mixing the calculated amount of iron precursors with already synthesized CCTO and triethylamine by solvothermal method, described in the materials synthesis section.

Fig. 1 shows the XRD data of bare CCTO, iron oxide and three different iron oxide–CCTO composites. Fig. 1(a) shows the XRD pattern of bare CCTO confirming pure phase of CCTO. From the XRD pattern shown in Fig. 1(b), the product can be identified as hematite Fe₂O₃ (JCPDS no. 033-0664, rhombohedral crystal system). No other peak of impurity was observed. In hematite rhombohedral (hexagonal) structure, oxygen anions exhibit the hexagonal close packing and Fe³⁺ ions occupy two-third of octahedral sites between oxygen. Hence, each iron atom is surrounded by six oxygen atoms, which share a face with each other in the layer above or below. Fig. 1(c)–(e) show the XRD pattern of Fe₂O₃–CCTO composites with different molar ratios (1 : 1, 1 : 0.5 and 1 : 0.33, respectively). The peaks match with the peaks of both bare CCTO (Fig. 1(a)) and bare Fe₂O₃ (Fig. 1(b)). The XRD data clearly show the variation in the intensity of iron oxide and CCTO peaks with an increase in the amount of CCTO in the composites. In Fig. 1(d), Fe₂O₃–CCTO 1 : 1, where the highest amount CCTO is present, some of the iron oxide peaks were of very small intensity.

Fig. 1 X-ray diffraction patterns for (a) CCTO; (b) Fe₂O₃; (c) Fe₂O₃–CCTO 1 : 0.5; (d) Fe₂O₃–CCTO 1 : 1; (e) Fe₂O₃–CCTO 1 : 0.3.

Fig. 2 shows the FESEM images of all the samples and elemental mapping of the Fe₂O₃–CCTO 1 : 1 composite. Fig. 2(a) displays the image of CCTO particles, which are elliptical in shape with average dimensions 312 nm length and 235 nm width. Fig. 2(b) shows SEM micrograph of bare Fe₂O₃ nanoparticles. The particles are cubic in shape with an average dimension of 60 nm. The reagent triethylamine caps particular faces and hinders growth along those faces of the particles, and therefore rhombohedral α-Fe₂O₃ particles are obtained. Fig. 2(c)–(e) show CCTO–Fe₂O₃ composite. It can be observed that the larger CCTO particles are fully decorated with smaller Fe₂O₃ particles. Moreover, the increase in the amount of Fe₂O₃ is also clearly evident in the images. It was also observed that the increasing ratio of CCTO did not affect the particle shape and size of Fe₂O₃. Fig. 2(f) shows the EDAX elemental mapping of the Fe₂O₃–CCTO 1 : 1 composite. It displays the presence of all metals – Ca, Cu, Ti, Fe and O.

Fig. 2 The SEM micrograph of the nanocomposites; (a) shows the morphology of as-prepared bare CCTO, (b) shows the bare iron oxide nanoparticles, (c–e) shows the morphology of Fe₂O₃–CCTO 1 : 1; (d) Fe₂O₃–CCTO 1 : 0.5; (e) Fe₂O₃–CCTO 1 : 0.3. (f) The elemental mapping of Fe₂O₃–CCTO 1 : 1 composites indicating the presence of all the components in the composites.

Electrochemical studies

The electrocatalytic activity of CCTO–Fe₂O₃ composites was studied for oxygen reduction reaction in an alkaline medium. Rotating ring disk electrode (RRDE) measurements were performed from 0.2 V to −1.5 V at a scan rate of 50 mV s⁻¹ and giving rotation rate of disk electrode from 100 to 1600 rpm. Prior to these measurements, the electrolyte solution was saturated with nitrogen gas, followed by oxygen gas. For comparison, we have also studied the ORR activity of Fe₂O₃ nanoparticles alone (without CCTO). The same trend of ORR activities has been observed for all the catalysts, where the current density is found to enhance with an increase in the rotating speed.

Fig. 3 is the polarization curve, which shows that with increased amount of CCTO, the ORR activity of the composite increases. There is a gradual shift of onset potential of ORR towards more positive potential with the increment of CCTO in the composites. The onset potential of ORR activity for Fe₂O₃ nanoparticles has been shifted towards more positive potential in the nanocomposites following the order (Fig. 3) – Fe₂O₃ < Fe₂O₃–CCTO 1 : 0.33 < Fe₂O₃–CCTO 1 : 0.5 < Fe₂O₃–CCTO 1 : 1.

Fig. 3 RDE plots (a) Fe₂O₃; (b) Fe₂O₃–CCTO 1 : 0.3; (c) Fe₂O₃–CCTO 1 : 0.5; (d) Fe₂O₃–CCTO 1 : 1.

Fig. 4 shows the decrease in the over potential with the increase in concentration of CCTO in the catalysts. This clearly indicates that CCTO has a positive impact on the ORR activity of iron oxide in the composites.

Fig. 4 Plot showing decreasing over potential of samples with increasing concentration of CCTO.

To understand the number of electrons transferred (n), which are involved in the ORR process, by these catalysts we have used Koutecky–Levich (K–L) eqn (3) as mentioned below

1/I = 1/I_k + 1/I_lev (3)

where, I is the disk current density, I_k is the kinetic current density and I_lev is the Levich current density. I_k also can be expressed as eqn (4)

I_k = nFAk_O2C_O2Γ_catalyst (4)

I_lev = 0.62nFAC_O2D_O2^2/3v^−1/6ω^1/2 (5)

where n is the number of overall electron transfer in the oxygen reduction reaction, F is the faradaic constant (96 485 C mol⁻¹), k_O2 is the rate constant for ORR (m s⁻¹), A is the area of electrode (0.19 cm²), C_O2 is the concentration of dissolved oxygen (1.21 mol m⁻³ in 0.1 M KOH) and Γ_catalyst is the surface concentration of the catalysts (20 μg cm⁻²). In eqn (5), D_O2 is the diffusion coefficient of O₂ in solution (1.87 × 10⁻⁹ m² s⁻¹ in 0.1 M KOH), v is the kinematic viscosity (1 × 10⁻⁶ m² s⁻¹ in 0.1 M KOH) and ω is the angular frequency of rotation (rad s⁻¹). Levich current is directly proportional to the square root of the rate of rotation of the electrode. Fig. 5 shows the corresponding Koutecky–Levich plots (K–L plot; I⁻¹ vs. ω^−1/2), which also have been used to calculate the electron transfer number (n). The linear nature of the K–L plots (Fig. 5) of all the catalysts indicates the first order reaction kinetics towards ORR, as reported earlier by iron oxide nanoparticles.⁵ However, careful observation of the K–L plots indicates that at higher positive potential (0.75 V and 0.7 V), K–L plot (Fig. 5) is different in bare Fe₂O₃ from the other Fe₂O₃–CCTO composites. Based on the intercept of K–L plot, kinetic limiting current (I_k) has been calculated.

Fig. 5 Koutecky Levich plot: (a) Fe₂O₃; (b) Fe₂O₃ : CCTO = 1 : 1; (c) Fe₂O₃ : CCTO = 1 : 0.5; (d) Fe₂O₃ : CCTO = 1 : 0.3.

To study the mechanistic pathways of iron oxide–CCTO nanocomposites, we have performed RRDE experiments and calculated the percentage of peroxide produced during the ORR (Fig. 6). RDDE technique is very useful to identify the nature of various intermediates formed in the different steps of ORR reaction and study the kinetics of ORR by a given catalyst. For performing RRDE, glassy carbon disc electrode and gold ring electrode were used (collection efficiency 37.5%). We have used the following formula for calculating the number of electron transfer in this process

n = 4I_D/[I_D + I_R/N] (6)

% H₂O₂ = 2 × [I_R/N]/[I_D + I_R/N] × 100% (7)

where I_D is faradaic current at the disk; I_R is faradaic current at the ring and N is the collection efficiency (0.37 in this case). The ring current follows the order: Fe₂O₃ < Fe₂O₃–CCTO 1 : 0.33 < Fe₂O₃–CCTO 1 : 0.5 < Fe₂O₃–CCTO 1 : 1. From eqn (6) and (7), it is found that n remains close to 4 in all cases (Fig. 6) and the percentage of peroxide evolution increases with an increase in rotation per minute. The average electron transfer number was ∼3.97 from the potential 0.0 V to 0.6 V, where 3.99 for bare Fe₂O₃ and 3.89 for Fe₂O₃ : CCTO 1 : 1 composites. Table (1)† shows the oxygen reduction performance of other perovskite-based catalysts and Table (2)† shows the performance of Fe₂O₃–CCTO towards hydrogen peroxide generation. If we compare with the literature data, Fe₂O₃–CCTO composites show comparatively better ORR activity than most of the reported perovskite materials. However, it also generates a significant amount of hydrogen peroxide in comparison to others. The amount of hydrogen peroxide generation essentially increases linearly with the increment of CCTO in the composites.

Fig. 6 Ring current plots of (a) Fe₂O₃; (b) Fe₂O₃–CCTO 1 : 0.3; (c) Fe₂O₃–CCTO 1 : 0.5; (d) Fe₂O₃–CCTO 1 : 1, indicating the liberation of hydrogen peroxide.

The improvement of ORR activity for iron oxide–CCTO composites on addition of CCTO can be attributed to the fact that CCTO provides plenty of active sites in the composites. In CCTO, Cu 3d⁹ act as a major charge carrier and displays a direct movement of charge from Cu 3d–O 2p to Ti 3d band. Cu 3d and O 2p form the valence band of CCTO and the conduction band is comprised of the Ti 3d band. Thus, CCTO shows good photo-electrocatalytic behavior under the illumination of visible light as Cu²⁺ and Ti⁴⁺ get excited to Cu³⁺ and Ti³⁺, respectively, reported earlier by us.²⁰ A similar report¹⁶ on another iron-infused copper-based perovskite material (Fe⁴⁺ based quadruple perovskite CaCu₃Fe₄O₁₂) shows very good OER activity, where covalent bonding network incorporating both Cu²⁺ and Fe⁴⁺ enhances the structural stability of the catalysts, showing a high OER activity in comparison to the standard RuO₂ catalysts.¹³ The RRDE experiments carried out with Fe₂O₃–CCTO composites, help us to identify the activity descriptor responsible for enhancement of the ORR activity in the composites. In the case of ORR on platinum surfaces, 4e⁻ process takes place via two different routes: (1) “direct” pathway, where 4e⁻ transfer takes place without any hydrogen peroxide formation, and (2) “serial” pathway involving sequential transfer of 2 electrons to adsorbed molecular oxygen resulting the formation of peroxide species, and ultimately breaking down to water by another 2e⁻ transfer process.^21,22

RRDE data (Fig. 6) indicate (from the data in Table (2)†) that with the increment of CCTO in composites, the amount of hydrogen peroxide generation increases and on addition of CCTO the electron transfer of ORR shifts from 3.99 to 3.89. Thus, from the RRDE measurement, we can conclude that this oxygen reduction reaction follows a 4e-series pathway involving the generation of hydrogen peroxide.¹³ The presence of more oxide sites on the surface of composites in comparison to that on bare iron oxide is mainly responsible for more hydrogen peroxide generation in composites.²³ The contribution of oxide sites from CCTO accounts for the weak adsorption of peroxide intermediate on the surfaces.²⁴ This had been reported earlier for other perovskite complexes, where higher hydrogen peroxide generation was observed due to higher concentration of oxide sites.²⁵ The increment of that CCTO in composites helps the adduct formation between oxygen and the metal center of CCTO. From Table (2),† it is clearly seen that the hydrogen peroxide yield changes from 1.28% (bare Fe₂O₃) to 4.31% (Fe₂O₃–CCTO 1 : 1 composites) at 0.4 V. The oxygen vacancy in perovskite also leads to better electron transfer and ultimately more hydrogen peroxide formation. The higher yield of peroxide along with the number of electron transfers, during the ORR process, indicates the reduction of oxygen to hydroxyl (OH⁻) ion through 4e⁻ or 2e⁻ + 2e⁻ routes and greatly influenced by CCTO concentration in the composites.

Conclusions

In summary, a new type of composite of metal oxide–perovskite (Fe₂O₃–CCTO) was synthesized in different molar ratios (1 : 0.3; 1 : 0.5 and 1 : 1) via a hydrothermal method. The microstructural characterization techniques reveal that the composite contains phases of both CCTO and α-Fe₂O₃, and that ∼60 nm cubic shaped α-Fe₂O₃ are decorated over CCTO particles. Electrochemical analysis shows that the composites have better electrocatalytic activity towards the ORR process in comparison to bare Fe₂O₃. RRDE study indicates that with the increment of CCTO in the composites more hydrogen peroxide generation occurs due to the presence of more active oxide sites contributed by CCTO and the ORR activity increases in the order: α-Fe₂O₃ < Fe₂O₃–CCTO 1 : 0.3 < Fe₂O₃–CCTO 1 : 0.5 < Fe₂O₃–CCTO 1 : 1. The RRDE analysis also shows that the electron transfer number in ORR is close to 4, indicating that the composites exhibit four electron transfer pathway method. Thus, this new type of Fe₂O₃–CCTO nanocomposite has the potential to be an effective non-precious, non-noble metal electrocatalyst for fuel cells and batteries.

Acknowledgements

The authors acknowledge MHRD for the financial aid and Advanced Material Research Centre (AMRC), IIT Mandi for providing user facilities. We also thank Department of Science and Technology (DST), India for providing financial support under Young Scientist Project Scheme.

References

1. H. B. Gray, Nat. Chem., 2009, 1, 7.
2. N. Markovic, T. Schmidt, V. Stamenkovic and P. Ross, Fuel Cells, 2001, 1, 105–116.
3. H. A. Gasteiger, S. S. Kocha, B. Sompalli and F. T. Wagner, Appl. Catal., B, 2005, 56, 9–35.
4. P. Mathew, J. P. Meyers, R. Srivastava and P. Strasser, J. Electrochem. Soc., 2012, 159, B554–B563.
5. H. Zhu, S. Zhang, Y.-X. Huang, L. Wu and S. Sun, Nano Lett., 2013, 13, 2947–2951.
6. Y. Liang, Y. Li, H. Wang, J. Zhou, J. Wang, T. Regier and H. Dai, Nat. Mater., 2011, 10, 780–786.
7. Z. S. Wu, S. Yang, Y. Sun, K. Parvez, X. Feng and K. Mullen, J. Am. Chem. Soc., 2012, 134, 9082–9085.
8. J. Sunarso, A. A. Torriero, W. Zhou, P. C. Howlett and M. Forsyth, J. Phys. Chem. C, 2012, 116, 5827–5834.
9. X. Li, W. Qu, J. Zhang and H. Wang, J. Electrochem. Soc., 2011, 158, A597–A604.
10. N. Ramaswamy, U. Tylus, Q. Jia and S. Mukerjee, J. Am. Chem. Soc., 2013, 135, 15443–15449.
11. N. Ramaswamy and S. Mukerjee, J. Phys. Chem. C, 2011, 115, 18015–18026.
12. P. H. Matter, L. Zhang and U. S. Ozkan, J. Catal., 2006, 239, 83–96.
13. N. Anastasijević, V. Vesović and R. Adžić, J. Electroanal. Chem. Interfacial Electrochem., 1987, 229, 305–316.
14. A. Mendoza-Garcia, H. Zhu, Y. Yu, Q. Li, L. Zhou, D. Su, M. J. Kramer and S. Sun, Angew. Chem., 2015, 54, 9642–9645.
15. Y. Miyahara, K. Miyazaki, T. Fukutsuka and T. Abe, J. Electrochem. Soc., 2014, 161, F694–F697.
16. S. Yagi, I. Yamada, H. Tsukasaki, A. Seno, M. Murakami, H. Fujii, H. Chen, N. Umezawa, H. Abe and N. Nishiyama, Nat. Commun., 2015, 6, 8249.
17. S. Gupta, W. Kellogg, H. Xu, X. Liu, J. Cho and G. Wu, Chem.–Asian J., 2016, 11, 10–21.
18. Y. Matsumoto, H. Yoneyama and H. Tamura, J. Electroanal. Chem. Interfacial Electrochem., 1977, 83, 237–243.
19. Z. Li, X. Lai, H. Wang, D. Mao, C. Xing and D. Wang, Nanotechnology, 2009, 20, 245603.
20. H. Kushwaha, N. A. Madhar, B. Ilahi, P. Thomas, A. Halder and R. Vaish, Sci. Rep., 2016, 6, 18557.
21. A. Damjanovic, J. Chem. Phys., 1966, 45, 4057.
22. H. S. Wroblowa and G. Razumney, J. Electroanal. Chem. Interfacial Electrochem., 1976, 69, 195–201.
23. E. Yeager, M. Razaq, D. Gervasio, A. Razaq and D. Tryk, The electrolyte factor in O₂ reduction electrocatalysis, DTIC Document, 1993.
24. V. S. Murthi, R. C. Urian and S. Mukerjee, J. Phys. Chem. B, 2004, 108, 11011–11023.
25. C.-F. Chen, G. King, R. M. Dickerson, P. A. Papin, S. Gupta, W. R. Kellogg and G. Wu, Nano Energy, 2015, 13, 423–432.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra20002h

---