Influence of electrode assembly on catalytic activation and deactivation of a Pt film immobilized H⁺ conducting solid electrolyte in electrocatalytic reduction reactions

Abstract

Symmetric (Cu–Pt|Nafion|Pt–Cu) and asymmetric (Pt|Nafion|Pt–Cu) assemblies were fabricated to study the nitrate reduction processes at the cathode. The electrocatalytic nitrate reduction reactions were performed in these assemblies in order to investigate the prerequisite for the enhanced catalytic activity, electrochemical cell durability as well as preferable product selectivity resulting from the reduction of nitrate at the cathode. It has been observed for the symmetric assembly that Cu particles were oxidized on the anode surface under an applied potential and the resulting copper ions migrated to the cathode surface through the Nafion membrane, which deposited as copper oxide on the cathode surface. The formation of this copper oxide covering layer on the Pt–Cu cathode surface is attributed as the reason for the deactivation of the cathode that governed the reduced nitrate reduction along with increasing nitrite selectivity. These problems were addressed and resolved with the asymmetric design of the electrocatalytic reactor, where enhanced hydrogen evolution activates the surface by eroding the CuO over layer as well as speeding up the slow rate determining hydrogenation reactions.

---

Introduction

The tuning of a catalytic material's properties, such as size, shape and elemental composition leads to modified reactivity and selectivity through the well known processes of electronic and geometric effects.^1–10 Besides the geometry of the elements in the catalytic materials, the geometry of the electrode materials in the electrochemical reactor are pivotal in directing the reactivity and selectivity.^11–14 In this regard, sandwich type M|conductive membrane|M reactor (M = Pt, Pd, Pt–Pd etc.) cells have been successfully employed for the selective oxidation of various organic compounds.^9,10,15–19 In these electrocatalytic cells, the conductive membrane serves as the electrolytic junction between the anode and the cathode. In a conventional reactor, ample supporting electrolyte is required to compensate for high medium resistivity towards the ion mobility through the membrane. Therefore, a supplementary process is obligatory to separate the synthesized products from the supporting electrolytes. The use of specially designed M|conductive membrane|M type reactors, which are seamlessly operational at the indefinitely high resistivity of the medium, will eliminate the need for a supporting electrolyte and therefore enables the easy separation of the synthesized products.

The nitrate reduction reaction is very important in terms of environmental pollution control as well as the synthesis of several nitrogenous compounds.^3,20–29 In the previous works, we reported the use of sandwich type reactors (M|Nafion|M–X; X = Ag,²⁶ Rh,²⁷ Cu (ref. 28 and 29)) to study the nitrate reduction reactions (NRR) in an asymmetric type reactor, where the promoter Cu metal is immobilized only on the cathode surface. To our knowledge, there is no literature precedent of investigating NRR using a symmetric type assembly of Cu–Pt|Nafion|Pt–Cu.

In this context, we investigated the comparative electrocatalytic activities obtained by a symmetric Cu–Pt|Nafion|Pt–Cu (Fig. 1a) and an asymmetric Pt|Nafion|Pt–Cu (Fig. 1b) assemblies. The catalytic comparisons have been demonstrated based on reactivity and selectivity pertaining to NRR as a model study, which in addition has led us to unveil the reasons for the catalytic deactivation of symmetric type assemblies. These interesting findings behind the deactivation of symmetric type reactors offers a way to identify the parameters that need to be addressed while designing a new electrochemical reactor. Consequently, the design of an asymmetric type reactor is shown to be effective for the electrocatalytic reduction of potential environmental pollutant nitrate ions to the valuable fertilizer precursor ammonia.

Fig. 1 Schematic representations of (a) symmetric and (b) asymmetric reactors.

Experimental

In order to prepare a membrane-electrode assembly (MEA), platinum was first chemically deposited from H₂PtCl₆ on the H⁺ conducting Nafion-117 membrane (Du Pont) surfaces supplied by Wako Incorp. Japan. A Nafion membrane of 2 cm × 3 cm was first sand blasted, cleaned with deionized water, dried at 110 °C and immersed into 200 ml of a 7.5 mM H₂PtCl₆ solution. Next, a mixed solution of 2.0 M NaBH₄ and 4.0 M NaOH was added to the membrane containing system at a rate of 2.0 ml h⁻¹. Meanwhile, the reaction mixture was heated from 35 °C to 60 °C at a rate of 5 °C h⁻¹. The Pt plating on both sides of the Nafion membrane was completed within 12 hours. A second plating of Cu metal was deposited from 0.01 M CuSO₄ solution to construct symmetric (Cu–Pt|Nafion|Pt–Cu) type assemblies (Fig. 1a). To prepare the asymmetric type (Pt|Nafion|Pt–Cu) assembly (Fig. 1b), one surface of the Pt membrane assembly (Pt|Nafion|Pt) was sealed with masking tape so that Cu particles were deposited only on the other surface. After Cu deposition, the masking tape was removed; as a consequence an asymmetric assembly was fabricated. The as prepared MEAs were first washed with dilute HCl, then were sonicated for 30 s to remove unreacted NaBH₄ and NaOH from the surfaces, and finally dried at 110 °C in the air. The resistivity of the surfaces of the as prepared assemblies was less than 10 Ω cm⁻¹. The electrocatalytic reduction of NO₃⁻ ions, in the absence of any supporting electrolyte, was carried out using both of the symmetric and asymmetric type assemblies by installing them in the micro reactors as shown in Fig. 1. The cathode surface was separated from the anode surface by the Nafion membrane. This operation created a sandwiched type reactor where the geometric area of each surface was 6 cm². The true aloofness between the anode and the cathode was 180 μm, which corresponds to the thickness of the membrane. The cathode and the anode chambers were filled with 8 ml aqueous solution of KNO₃ and water, respectively. To accomplish electrolysis, the reactor was connected with a DC supplier. In order to evaluate the concentration changes of NO₃⁻, NO₂⁻ and NH₄⁺ with time, a portion of 20 μl was collected from the solution in the cathode compartment at a constant interval and diluted to 5 ml. The concentration changes of NO₃⁻, NO₂⁻ and NH₄⁺ were measured by an ICA-2000 model (DKK-TOA Corporation) ion chromatograph. In order to compare the reactivity differences among different types of assemblies, the rate constant (k) was evaluated on the supposition that the concentration of NO₃⁻ (C) decreases following first order kinetics, i.e., C = C₀ exp(−k₁t), where C₀ is the initial concentration and t is time (min). The metallic composition of each MEAs was determined by X-ray fluorescence analysis. The interior (inside the membrane) distribution of solid particles was analyzed with a Simudzu model electron probe micro analyzer. The electrochemical experiments were carried out using a Metrohm 797VA automatic electrochemical polarization system. The scanning electron microscope (SEM) image of the surface was taken with a JEOL model JSM-6060LV instrument. X-ray absorption near-edge structure (XANES) spectra were recorded at the BL47XU beam-line at SPring-8.

Results and discussion

Relative NRR efficiency

When a Pt film was alone used as a cathode in a membrane reactor, the reaction rate was very slow (k₁ < 2 × 10⁻³ min⁻¹). Distinct catalytic activities were noticed when Cu particles were immobilized onto the Pt surface with an enhanced NRR rates. Nitrate ions are reduced using a consecutive pathway, where nitrate ions are converted into intermediate nitrite ions, which are then converted into other products as per the following trail (1).²⁸

In order to evaluate the electrocatalytic nitrate reduction kinetics, the change in the concentrations of the reactant nitrate along with that of intermediate nitrite and product ammonia were monitored and quantified at given time intervals. Fig. 2a shows the concentration profile of the nitrate reduction reaction (0.05 M KNO₃, 0.1 A, 180 min) with the course of time while electrolysis was carried out using Pt|Nafion|Pt–Cu (16 atom% Cu deposits) asymmetric assembly (Fig. 1b). The concentration of reactant nitrate decreased exponentially with time, which is consistent with 1^st order kinetics. On the other hand, the concentration of nitrite increased fairly rapidly leading to a plateau and followed by a slow decrease. Consequently, the ammonia concentration increased slowly due to the slow conversion of intermediate nitrite into ammonia.

Fig. 2 Nitrate reduction reaction profiles at Cu–Pt catalytic assemblies. (a) Concentration profile of nitrate reduction reaction in an asymmetric reactor. (● NO₃⁻ concentration, ■ NO₂⁻ concentration, ▲ NH₄⁺ concentration) Cathode: 16 atom% Cu + Pt, 0.1 A, 180 min, 295 K, reactor capacity 8 ml, initial nitrate concentration 0.05 M. (b and c) Dependence of the rate constant and product selectivity on the repetition of experiments in Cu–Pt|Nafion|Pt–Cu symmetric and Pt|Nafion|Pt–Cu asymmetric assemblies respectively.

Recyclability and reusability are the important aspects of any catalytic processes. In order to assess the possibility of recyclability and/or reusability of these two different types of catalytic assemblies, four repetitive experiments were conducted under same experimental condition. While the rate constant for the symmetric assembly (Fig. 1a) was decreased from 24.1 × 10⁻³ min⁻¹ to 3 × 10⁻³ min⁻¹ (see Fig. 2b), an increase of rate was observed from 22.4 × 10⁻³ min⁻¹ to 30.7 × 10⁻³ min⁻¹ (see Fig. 2c) in the fourth experiment for an asymmetric assembly (Fig. 1b). This observation suggests that the asymmetric type design of the catalytic assembly provides not only a way to overcome the limitation of deactivation associated with symmetric type reactors, but also offers increased reactivity on long term repeated use.

Having the evidence of longer time sustainability for the asymmetric assembly, we endeavoured to investigate product selectivities of the catalytic assemblies. In the case of symmetric assembly, the nitrite selectivity is seen to increase remarkably from 45% to 68% in the second experiment, which is followed by a steady increase in the subsequent experiments (see Fig. 2b). Conversely, a gradual decrease in nitrite selectivity was noticed for asymmetric assembly, numerically, from 54% in the 1^st experiment to 34% in the 4^th experiment (see Fig. 2b). However, a decreasing trend for ammonia selectivity was observed for the repeated experiments irrespective of geometry of the assembly used. The most remarkable feature for the asymmetric assembly is its consistently higher selectivity for ammonia than for the symmetric analogue, which implies that the former assembly converts the intermediate nitrite to ammonia more rapidly and efficiently. The instantaneous increase in the reaction rate and decrease of nitrite selectivity suggest that Cu–Pt surface at the asymmetric assembly was activated during the electrolysis experiments to enhance step-2 of the reactions given in eqn (1). The increase in the rate of conversion of nitrite to ammonia was balanced by the faster conversion of nitrate to nitrite and so on. This enhanced reactivity of the asymmetric assembly will offer at least two beneficial impacts. First, it will help to minimize the environmental pollution problem by effectively reducing the pollutants nitrate and nitrite. Second, it is expected to facilitate the food production by providing alternative way to synthesize the valuable fertilizer precursor ammonia.

Influence of copper oxides on NRR activities

After the recognition of the superior reactivity and preferable product selectivity for asymmetric design over the symmetric ones, we were motivated to inspect the reasons behind the activation or deactivation of the respective reactor types. Before switching into the deeper aspects of the catalytic activation–deactivation processes, it is required to discuss briefly about the catalytic nitrate reduction processes taking place on the cathode of the membrane reactor. It is established that water molecules are oxidized on the anode surface producing molecular oxygen and protons, which under the applied potential migrate towards the cathode surface through the conducting Nafion membrane. These protons are next reduced on the cathode surface to form adsorbed hydrogen.²⁶

Anode (Pt/Cu–Pt):

H₂O_ads → 2H⁺ + 1/2O₂ + 2e⁻ (2)

H⁺(anode) → H⁺(cathode) (migration) (3)

Cathode (Cu–Pt):

2H⁺ + 2e⁻ → 2H_ads →H_2,ads (4)

Considering the study reported by N. Barrabés, et al.³⁰ and our observations, the catalytic nitrate reduction reactions at the cathode surface may be summarized as shown in Scheme 1. The noble metal atoms (Pt) adsorb hydrogen atoms to form Pt–H species on the cathode surface. At the same time, metallic Cu atoms are oxidized to Cu²⁺ ions by reducing nitrate ions. The Cu²⁺ species is then reverted to its metallic state (Cu^o) by receiving electrons from the Pt–H species. In this way, catalytic cycles are completed and new cycles are started. Due to enormous stability, Cu₂O and CuO are not expected to participate in the catalytic cycle shown in Scheme 1. As a result, the presence of Cu₂O and or CuO may deactivate the catalytic performance by blocking active NRR sites.

Scheme 1 Catalytic nitrate reduction reaction mechanism at Cu modified Pt surface.

Besides the catalytic hydrogenation process, heterogeneous electron transfer (ET) reactions are also possible, which can reduce nitrate ions. Therefore, in order to investigate the influence of copper oxide on ET reactions, cyclic voltammetric experiments were performed. In this respect, Cu particles were electrochemically deposited on a Pt disk (2 mm) electrode from a 0.01 M solution of CuSO₄ under de-aerated and aerated conditions separately by cycling the electrode four times between 0 and −1 V vs. Ag/AgCl reference electrode at a scan rate of 100 mV s⁻¹. This yielded formation of Cu over layers on the Pt disk. Thus, electrode (AE) prepared under aerated condition contained higher amount of copper oxides than that of the electrode (DE) prepared under de-aerated conditions. The DE electrode was kept under an inert atmosphere so that the formation of CuO due to air oxidation could be minimized before application; whereas for the AE electrode, no precaution was taken to avoid CuO formation. Both of the electrodes were next used to record cyclic voltammograms (CVs) of 5 mM KNO₃ (in 0.5 M KCl, N₂ environment) at a scan rate of 50 mV s⁻¹ as shown in Fig. 3. In both cases, the CVs show two well-defined waves (E1 and E2) and a broad cathodic peak (E3). In a previous report,²⁸ E1 has been resolved as the reduction of surface copper oxides, E2 as the reduction of NO₃⁻ to NO₂⁻, and E3 as the formation of ammonia. The E1 wave of the DE electrode appeared at −0.52 V with a peak current of −1.52 mA in contrast to AE electrode where this wave appeared at a more negative potential (−0.64 V) with an increased peak current of −3.35 mA. This shift in reduction potential is expected as the AE electrode contained more copper oxides that required more driving potential for the reduction. Conversely, the E2 and E3 diffusive waves by the AE electrode appeared at −0.80 V (−2.64 mA) and −1.13 V (−8.21 mA), respectively. In the case of the DE electrode, these waves were generated at more negative potentials i.e., −0.83 V (−4.5 mA) and −1.21 V (−15.5 mA), respectively. The less NRR currents generated by the AE electrode at relatively positive potentials provide concrete evidence that the presence of copper oxides in the matrix of Cu–Pt electrode inhibited both of the steps (eqn (1)) on the way to complete NRR. In other words, it is reasonable to claim that the lesser is the CuO content, the higher is the NRR efficiency.

Fig. 3 Cyclic voltammograms of 5 mM KNO₃ in 0.5 M KCl using Cu–Pt electrode at a scan rate of 50 mV s⁻¹ under nitrogen atmosphere. Solid line: Cu–Pt (DE) electrode prepared under de-aerated conditions, and dotted line: Cu–Pt (AE) electrode prepared under aerated conditions.

Deactivation of the symmetric assembly

It is evident from the CV experiments as well as from the literature reports that existence of copper oxides must inhibit both of the catalytic hydrogenation reactions and ET reactions pertaining to NRR. From this point of view, we evaluated the relative abundances of Cu₂O and/or CuO in catalytic matrixes before and after use of the catalytic surfaces in the case of the symmetric assembly.

Fig. 4 exhibits the EPMA maps of symmetric MEA before and after use for NRR. These maps do not provide quantitative compositional data, but they are useful for determining the relative Cu/Pt distribution and the change in the distribution that occurred during the course of electrocatalysis. The corrosion resistant Pt particles were deposited only on the surfaces of the membrane, and it was found that the Pt particles did not penetrate the cross sectional area even after electrolysis. Moreover, copper species were distributed across the membrane during electrolysis, because Cu²⁺ ions can easily be exchanged, and they migrated deeply inside the membrane under an applied potential. After nitrate reduction reaction (0.05 M NO₃⁻, 0.1 A, 3 h), the concentration of Cu species was decreased near the anode–membrane interface, and they were populated adjacent to the membrane–cathode interface (see Fig. 4c and d).

Fig. 4 Electron probe micro analysis of membrane cross section samples from symmetric electrode-membrane assembly. (a) Sampling arrangement (b) Cu dispersion in the fresh membrane (c) Cu dispersion adjacent to the membrane–cathode interface after use, and (d) Cu dispersion near the anode–membrane interface after use. The dark areas indicate epoxy of the sample carrier.

Having the evidence of migration of Cu(II) species, we attempted to evaluate the relative abundance of copper oxides in the Cu–Pt matrix of the cathode assembly before and after use. Because every single element has its distinct core level energy, X-ray Absorption Near Edge Structure (XANES) spectroscopy permits the extraction of the signal from a surface monolayer or even a single buried layer, such as dopants below an electrode surface in the presence of huge background signals. The edge position, defined as the maximum point of the first derivative function in the rapidly rising edge step of the absorbance vs. energy plot, have been found to be 8978.6, and 8993.0 eV for Cu₂O, and CuO, respectively as shown in Fig. 5a. This shows the increase in edge energy as the oxidation number increases, which is a well known Phenomenon of XANES that allowed its uses to detect oxidation state and chemical environment of an element. In order to quantify the relative existence of Cu₂O and CuO, XANES spectra of fresh and used Cu–Pt cathodes were taken as shown in Fig. 5b. It has been noticed that the existence of Cu₂O was almost removed from the cathode matrix after 3 h use and the absorbance vs. energy peak area between 8989.0 eV and 9000.0 eV was enlarged by 11.5%. These observations clearly indicate the increase of CuO content in the cathode matrix after electrolysis.

Fig. 5 XANES spectra of (a) Cu foil and CuO, and (b) Cu species sampled from the cathode of symmetric assembly before and after 3 h use.

Consequently, oxidation of metallic Cu at the anode surface followed by the subsequent appearance of Cu(II) species at the cathode surface entails us to speculate the migration of copper (as Cu²⁺) from the anode to the cathode surface where they deposited as oxides. Parts of the Cu²⁺ ions are also assumed to be trapped within the membrane pores. This might have caused a significant increase of anode to cathode impedance during electrolysis causing an inhibition towards the NRR catalytic process as shown in Fig. 2b.

Activation of the asymmetric assembly

Because the copper oxides inhibit the catalytic process, the catalytic activation (as shown in Fig. 2c) is speculated to be possible by minimizing the concentration of the copper oxides. On the cathode surface of the asymmetric assembly (Pt|Nafion|Pt–Cu), the only possible source of oxides is the air oxidation of metallic copper atoms because copper metal is slowly oxidized to CuO in presence of air according to eqn (5).³⁰

2Cu (s) + O₂ (g) → 2CuO (s) (5)

In order to estimate the relative abundance of the oxides in the Cu–Pt matrix of the asymmetric assembly before and after long term electrolysis, voltammetric experiments were performed in the presence of 0.5 M KCl, and at a scan rate of 50 mV s⁻¹ as shown in Fig. 6. In this case, a circular (2 mm in diameter) area of a fresh and a used (after electrolysis) Cu–Pt/Nafion surface specimens from Pt|Nafion|Pt–Cu assembly were used as working electrodes. It was found that the area under the Cu(II) reduction peak was decreased by ca. 40% in the case of the used electrode assembly. These observations therefore suggests that the reduction and/or removal of CuO content from the cathode surface of Pt|Nafion|Pt–Cu assembly during the period of electrolysis increased NRR efficiency as shown in Fig. 2c.

Fig. 6 Cu(II) reduction waves of fresh (solid line), and used (after 9 h, dotted line) Cu–Pt/Nafion in 0.5 M KCl at a scan rate of 50 mV s⁻¹. The circular working electrode (ϕ = 2 mm) was collected from the cathode of the Pt|Nafion|Pt–Cu assembly.

Now, the question arises why CuO content differed on the cathode surface of Pt|Nafion|Pt–Cu assembly before and after use. Because the E1 peak in Fig. 3 denotes the reduction of CuO to Cu, the content of the oxides might be decreased due to its possible Cu(II) reduction during electrolysis. Considering this assumption, repetitive CVs (20 cycles) of 5 mM KNO₃ in 0.5 M KCl using Cu–Pt (AE) electrode was recorded consecutively as shown in Fig. 7. From this figure, it is seen that the shapes and sizes of all of the three E1, E2 and E3 peaks remained unaltered irrespective of cycle numbers. This observation suggests that the electron transfer reactions could not alter the Cu to CuO balance. Consequently, it can be ascribed that the Cu–Pt surface in the Pt|Nafion|Pt–Cu assembly was not activated due to direct electrochemical activities. The most acceptable process of reduction of copper oxides into metallic copper is heating the oxides under H₂ atmosphere at 280 °C.³¹

CuO + H₂ → Cu + H₂O (6)

Fig. 7 Repetitive cyclic voltammograms (20 cycles) of 5 mM KNO₃ in 0.5 M KCl using a Cu–Pt (AE) electrode at a scan rate of 50 mV s⁻¹ under nitrogen atmosphere.

Although vigorous H₂ is evolved on the cathode surface of the reactor (see Fig. 1) under the present experimental conditions (25 °C), the existence of copper oxides in the cathode matrix should remain unaltered because the reduction of copper oxide to metallic copper by evolved H₂ at this ambient temperature is not possible. Thus, finally we assumed that the erosion of CuO due to continuous flow of H₂ gas evolved at the Cu–Pt cathode surface was the way of decreasing CuO from the Cu–Pt surface. Therefore, we measured copper species in the catholyte by means of Inductively Coupled Plasma (ICP) technique. After each of the four experiments reported in Fig. 2c, we determined 62, 40, 21 and 18 ppb of copper species in the catholyte. This successive removal trend of CuO is consistent with the upward trend of the rate constant in the repetitive experiments as shown in Fig. 2c.

The removal of CuO from the cathode surface increased the metallic luster and decreased the sizes of the deposited Cu particles, which is apparent from the SEM images (Fig. 8) of fresh and used surfaces. The decrease of particle size increased the specific surface area of the used cathode, which played vital roles of surface activation of Pt|Nafion|Pt–Cu assembly concerning nitrate reduction reactions. In addition, as the cathode surface was negatively charged and H₂ evolution reaction creates a reduction environment, further slow air oxidation of Cu was returded. This phenomenon ensured the stability of the cathode surface for long term use.

Fig. 8 SEM Images of Cu–Pt surface. (a) As prepared, and (b) after extensive electrolysis (Pt|Nafion|Pt–Cu).

Influence of H⁺ migration on activation–deactivation processes

Finally, knowing the beneficial impacts of hydrogen evolution on the catalytic performances, we compared the relative hydrogen evolution efficiency of the symmetric and asymmetric reactors with a view to constructing the rationale for deactivation of symmetric and activation of asymmetric assemblies. Here, the results have been presented in terms of volume of the H₂ gas collected from the cathode chamber at certain time intervals as shown in Fig. 9. With the passage of time, the volume of the evolved hydrogen gas was increased linearly while 0.1 A current was passed through the asymmetric reactor due to its considerable stability. On the other hand, the volume vs. time curve with respect to the symmetric reactor gradually deviated negatively from linearity implying that the gas evolution rate decreased with the passage of time. This observation clearly supports that H⁺ migration through the membrane (eqn (3)) was hindered due to presence of Cu(II) species (trapped inside the membrane) and reduction of H⁺ ions to molecular hydrogen (eqn (4)) were inhibited as a consequence. Overall, the electrochemical redox reactions of an ideal electrolysis cell was perturbed, hence, the performance of the symmetric reactor was lost as reported in Fig. 2b.

Fig. 9 Cathodic hydrogen evolution due to water hydrolysis. (●) Cu–Pt|Nafion|Pt–Cu, and (○) Pt|Nafion|Pt–Cu electrode assemblies. Current passed: 0.1 A.

Conclusions

Presence of oxidizable metallic species on the anode surface deactivates the catalytic performance of the cathode surface of the Cu–Pt|Nafion|Pt–Cu type symmetric assemblies. The asymmetric design of the assembly as in Pt|Nafion|Pt–Cu not only alleviates the problem of deactivation, but also increases the nitrate reduction rate and provides preferable product (ammonia) selectivity. More precisely, vigorous hydrogen evolution reaction in the first place improves NRR reactivity of a cathode surface of Pt|Nafion|Pt–Cu assembly by removing CuO particles; second, it drives faster reduction by facilitating the rates of hydrogenation of the nitrate and nitrite species; third, it protects the further formation of CuO by creating reducing environment; and finally, when the content of CuO is minimized on the cathode surface, the cathode assembly exhibited enhanced performance with respect to nitrate reduction reactions.

Acknowledgements

The authors acknowledge Prof. Masato Machida, Department of Nano Science and Technology, Kumamoto University, Japan for providing required materials and experimental setups. Shahjalal University of Science and Technology research center and UGC of Bangladesh is acknowledged for financial supports (2014–15).

Notes and references

1. F. F. Tao, S. Zhang, L. Nguyen and X. Zhang, Chem. Soc. Rev., 2012, 41, 7980–7993.
2. M. Cao, D. Wu and R. Cao, ChemCatChem, 2014, 6, 26–45.
3. W. Siriwatcharapiboon, Y. Kwon, J. Yang, R. L. Chantry, Z. Li, S. L. Horswell and M. T. M. Koper, ChemElectroChem, 2014, 1, 172–179.
4. C. H. Cui and S. H. Yu, Acc. Chem. Res., 2013, 46, 142–1437.
5. J. Lai, L. Zhang, W. Qi, J. Zhao, M. Xu, W. Gao and G. Xu, ChemCatChem, 2014, 6, 2253–2257.
6. Y. Nakayama and M. Oyama, Chem. Commun., 2013, 49, 5228–5230.
7. M. Shand and J. A. Anderson, Catal. Sci. Technol., 2013, 3, 879–899.
8. D. Reyter, D. Bélanger and L. Roué, Electrochim. Acta, 2008, 53, 5977–5984.
9. Z. Zhang, L. Xin and W. Li, Appl. Catal., B, 2012, 119–120, 40–48.
10. S. K. Green, G. A. Tompsett, H. Ju Kim, W. Bae Kim and G. W. Huber, ChemSusChem, 2012, 5, 2410–2420.
11. S. P. Jiang, J. Appl. Electrochem., 2004, 34, 1045–1055.
12. L. Y. Woo, R. S. Glass, R. F. Novak and J. H. Visserb, J. Electrochem. Soc., 2010, 157, J81–J87.
13. C. Setevich, F. Prado and D. Z. de Florio, Int. J. Hydrogen Energy, 2014, 39, 8738–8743.
14. E. A. B. Ramirez and S. P. McGarry, Org. Electron., 2014, 15, 1131–1137.
15. Z. Ogumi, S. Ohashi and Z. Takehara, Electrochim. Acta, 1985, 30, 121–124.
16. Z. Ogumi, H. Yamashita, K. Nishio, Z. I. Takehara and S. Yoshizawa, Electrochim. Acta, 1983, 28, 1687–1693.
17. K. Otuska and I. Yamanaka, Appl. Catal., B, 1986, 26, 401–404.
18. G. R. Dieckmann and S. H. Langer, J. Appl. Electrochem., 1997, 27, 1–8.
19. G. R. Dieckmann and S. H. Langer, Electrochim. Acta, 1998, 44, 437–444.
20. Y. Y. Birdja, J. Yang and M. T. M. Koper, Electrochim. Acta, 2014, 140, 518–524.
21. A. Anastasopoulos, L. Hannah and B. E. Hayden, J. Catal., 2013, 305, 27–35.
22. J. Yang, P. Sebastian, M. Duca, T. Hoogenboom and M. T. M. Koper, Chem. Commun., 2014, 50, 2148–2151.
23. A. Miyazaki, T. Asakawa, Y. Nakano and I. Balint, Chem. Commun., 2005, 3730–3732.
24. S. Kuwabata, S. Uezumi, K. Tanaka and T. Tanaka, Inorg. Chem., 1986, 25, 3018–3022.
25. Y.-M. Fang, J.-J. Sun, A.-H. Wu, X.-L. Su and G.-N. Chen, Langmuir, 2009, 25, 555–560.
26. M. A. Hasnat, M. Saiful Alam, M. H. Mahbub-ul Karim, M. A. Rashed and M. Machida, Appl. Catal., B, 2011, 107, 294–301.
27. M. A. Hasnat, M. A. Islam, S. M. Borhanuddin, M. R. U. Chowdhury and M. Machida, J. Mol. Catal. A: Chem., 2010, 317, 61–67.
28. M. A. Hasnat, S. Ben Aoun, S. M. Nizam Uddin, M. Mahbubul Alam, P. P. Koay, S. Amertharaj, M. A. Rashed, M. M. Rahman and N. Mohamed, Appl. Catal., A, 2014, 478, 259–266.
29. M. A. Hasnat, M. A. Rashed, S. Ben Aoun, S. M. Nizam Uddin, M. Saiful Alam, S. Amertharaj, R. K. Majumder and N. Mohamed, J. Mol. Catal. A: Chem., 2014, 383–384, 243–248.
30. N. Barrabés, J. Just, A. Dafinov, F. Medina, J. L. G. Fierro, J. E. Sueiras, P. Salagre and Y. Cesteros, Appl. Catal., B, 2006, 62, 77–85.
31. J. Y. Kim, J. A. Rodriguez, J. C. Hanson, A. I. Frenkel and P. L. Lee, J. Am. Chem. Soc., 2003, 125, 10684–10692.

---