Investigating the behavior of various cocatalysts on LaTaON₂ photoanode for visible light water splitting

Abstract

We performed a comparative study on the photoelectrochemical performance of LaTaON₂ loaded with NiO_x, Ni_0.7Fe_0.3O_x, CoO_x and IrO_x as cocatalysts. Ni-based oxides lead to the highest improvement on the photoelectrochemical performance, while CoO_x and IrO_x also enhance the performance though to a lower extent, but they simultaneously introduce more pseudocapacitive current thus resulting in an inefficient utilization of the photo-generated holes. Repetitive voltage cycling between 1.0 V_RHE and 1.6 V_RHE transforms the NiO_x and Ni_0.7Fe_0.3O_x into oxyhydroxides known to possess higher catalytic activities. However, these oxyhydroxides lead to lower photoelectrochemical performance compared to the as-loaded oxides, most probably due to the decay of the passivation centers at the photoelectrode–cocatalyst interface. High catalytic activities cannot be achieved without sufficient passivation of surface recombination states. Despite that the photoelectrochemical performance of LaTaON₂ can be improved by cocatalysts, the maximum achievable photocurrent density is still not comparable to that reported for other oxynitride compounds. Our study suggests that poor electronic conductivity or severe bulk recombination of the photo-generated electron–hole pairs are the main limiting factors for the photon-to-current conversion efficiency in LaTaON₂ photoanodes.

---

Introduction

The conversion of solar energy into chemical energy as a clean fuel, thus making solar energy available for use at night and meeting peak power demands, promises great potential for the development of a renewable and sustainable integrated power generation system. This could be realized by solar light driven photoelectrochemical (PEC) water splitting to produce oxygen and hydrogen. In such a system, a photoanode (n-type conductivity) drives the oxidation of water at the electrode–electrolyte interface, while a photocathode (p-type conductivity) drives the reduction of water. Oxynitrides of transition metals and rare earth metals are one class of promising semiconductor materials for solar water splitting. The width and energy positions of the band gap of some oxynitrides make it theoretically possible that both reactions of water oxidation and reduction occur within one material, resulting in the so-called overall water splitting. Since the efficiency of water splitting is mainly limited by the oxygen evolution reaction (OER),^1,2 many oxynitrides are particularly investigated as photoanodes for the water oxidation half reaction.

The PEC performance of a photoanode is usually probed using a three-electrode configuration where the immobilized photoanode is used as the working electrode, Pt as the counter electrode, and the potential of the photoanode is monitored by a reference electrode. The measured photocurrent reflects the photon-to-current conversion efficiency of the photoanode. Photoanodes, deposited by electrophoretic method (EPD) on fluorine-doped tin oxide (FTO), such as TaON,^3–5 Ta₃N₅,³ and BaTaO₂N,^6,7 show photocurrent densities for water oxidation in the range of several mA cm⁻² at 1.23 V_RHE (the water oxidation potential) under an irridation by a 300 W Xe lamp fitted with a 400 nm or 420 nm cut-off filter. Materials that are used for OER electrocatalysis such as IrO_x,^3,6 CoO_x,^4,6 Co–Pi (inorganic cobalt phosphate)⁸ and NiO_x^9,10 are often also employed as cocatalysts for photoanodes to promote the kinetics of water oxidation and suppress the recombination of photo-generated electrons and holes by passivating the surface recombination centers.¹¹ It has been reported that loading CoO_x or IrO_x cocatalysts on TaON, Ta₃N₅, BaTaO₂N, SrNbO₂N and LaTiO₂N usually result in 50–400% higher currents and more negative onset potentials for water splitting.^3,6,12–14 However, no obvious improvement in photocurrent was observed after loading cocatalysts Co₃O₄, Co(OH)₂ or Co–Pi on LaTaON₂ photoanode.¹⁵ Here we would like to take LaTaON₂ as an example and investigate whether and how proper cocatalysts can be found to further improve the PEC performance of this oxynitride.

Recently, in the research of water electrolysis, Ni_yFe_1−yO_x has been reported to exhibit a higher catalytic capability than IrO_x in basic media for the OER,^16,17 which is attributed to the in situ formation of layered Ni_yFe_1−yOOH oxyhydroxide species upon voltage cycling with nearly every Ni atom being electrochemically active.¹⁶ It has also been suggested that the Fe ion might be the real catalytic active site in nickel iron oxides/oxyhydroxides.¹ The use of Ni(Fe)OOH cocatalyst indeed remarkably enhanced the PEC performances of hematite,⁹ BiVO₄,¹⁸ TiO₂,¹⁹etc. Depending on the ion permeability, cocatalysts have been demonstrated to form different interfaces with the photoelectrode.^19,20 Ion-permeable Ni(OH)₂ or NiOOH form an adaptive junction with the photoelectrode, where the effective Schottky barrier height changes with the oxidation level of the cocatalyst,¹⁹ while ion-impermeable cocatalysts such as dense IrO_x form a constant-barrier-height ‘buried’ junction with the photoelectrode.¹⁹ However, it has also been reported that the PEC performance of the photoelectrode–cocatalyst system is nearly independent on the activity of ion-permeable cocatalysts, while strongly influenced by the activity of ion-impermeable cocatalysts.^20,21

The comprehensive understanding of the photoelectrode–cocatalyst interface is further complicated by the existence of various chemical species of the cocatalysts characterized by different coordination environments and/or valence states. As an example, researchers have categorized NiO_x on BiVO₄ photoanode into three types: OER catalytic centers, recombination centers, and passivation centers.¹⁰ By selective removal of the first two centers while keeping the NiO_x passivation centers at the surface, a significantly increased photocurrent was obtained.¹⁰

These considerations lead to the following questions. What is the most important asset of a cocatalyst for determining a high PEC performance? Should we introduce a material with higher catalytic activity or better passivation and minimal recombination effects? This study aims at gaining insights into the roles of the cocatalysts in affecting the photoanode performance. To achieve this goal, various metal oxides cocatalysts were loaded on LaTaON₂ photoanodes and the PEC performance were measured and compared. Specifically, in order to enlarge the effects of cocatalysts on a LaTaON₂ photoanode, a very small area was illuminated in comparison to the whole size of the photoanode in this work. The dark electrocatalytic behaviors of the cocatalysts were therefore “exaggerated” and the differences between them could be easily distinguished.

Experimental

Synthesis of LaTaON₂

The oxide precursor LaTaO₄ was first prepared by a solid state reaction of La₂O₃ (Alfa Aesar puratronic, 99.993%) and Ta₂O₅ (Alfa Aesar puratronic, 99.993%) in the ratio of 1 : 1, with calcination at 1000 °C for 2 h, and then at 1400 °C for 10 h in air. Three cycles of heating were conducted with intermediate grinding until the reaction was fully completed. Afterwards, 0.5 g LaTaO₄ was mixed with a mineralizer of 0.5 g NaCl and put in an alumina boat for ammonolysis at 950 °C under an ammonia flow of 200 mL min⁻¹ for 30 h to prepare LaTaON₂.

Materials characterization

The oxide precursor LaTaO₄ and the nominal ammonolysis product LaTaON₂ were characterized by powder X-ray diffraction (XRD) on a Bruker–Siemens D500 X-ray Diffractometer. The optical band gap of LaTaON₂ was determined by measuring the diffuse reflectance spectrum in the wavelength range between 200 nm and 1000 nm with a Cary 500 Scan UV-Vis-NIR spectrophotometer using an integrating sphere. Scanning electron microscopy (SEM) images were obtained with a Zeiss Supra VP55 Scanning Electron Microscope. N₂ adsorption–desorption analysis was conducted on a BEL-Mini device, and the specific surface area was calculated according to the Brunauer–Emmett–Teller (BET) theory.

Preparation of LaTaON₂ photoanodes

LaTaON₂ photoanodes were prepared via the electrophoretic deposition method (EPD) followed by a necking treatment.³ In detail, 120 mg LaTaON₂ powders were dispersed in 50 mL acetone with 10 mg iodine and sonicated for 60 min. Iodine reacts with acetone releasing H⁺, thus the LaTaON₂ powders were positively charged. Two parallel FTO substrates (1 × 2 cm, ∼7 Ω sq⁻¹, Sigma Aldrich) were then placed in a LaTaON₂ dispersion with a distance of 1 cm under a bias of 20 V for 5 min. LaTaON₂ powders were deposited on the negative electrode with a geometric area of ∼1 × 1 cm².

The so-called necking treatment was applied to enhance the contacts between LaTaON₂ particles. 25 μL of 10 mM TaCl₅ methanol solution was dropped on LaTaON₂ photoanodes and dried in air. This procedure was repeated for three times, followed by heating at 350 °C in air for 30 min. This temperature was chosen to avoid oxidation of the oxynitrides. For a comparison to the post-treatment in air, the as-prepared LaTaON₂ photoanodes and the samples with pre-loaded cocatalysts were also treated for 30 min under NH₃ flow at 450 °C and 400 °C, respectively. The use of FTO substrates determines an upper limit of 500 °C for the heat treatment in ammonia.³ A relatively lower temperature was used to treat cocatalysts-loaded samples under ammonia since we tried to avoid a reduction of the cocatalysts.

It is worth noting that all the cocatalysts used in this paper were supposed to be dense cocatalysts due to thermal treatments. NiO_x, Ni_0.7Fe_0.3O_x, and CoO_x cocatalysts (2 wt%, calculated as metal) were pre-loaded on LaTaON₂ powders by impregnation from their aqueous nitrate solutions, followed by drying and heating at 150 °C for 30 min in air. A mixture of Ni(NO₃)₂ and Fe(NO₃)₃ solution was used to load Ni_0.7Fe_0.3O_x. IrO_x was post-loaded on the as-prepared LaTaON₂ photoanodes by impregnation from a colloidal IrO₂ aqueous solution, which was prepared by hydrolysis of Na₂IrCl₆.²² The colloidal IrO₂ was prepared as follows: 0.008 g Na₂IrCl₆·6H₂O was dissolved in 50 mL H₂O, and the pH was adjusted to 11–12 with 1 M NaOH. The solution was heated at 80 °C for 30 min and cooled to room temperature by immersion in an ice water bath. The pH of the cooled solution was adjusted to 10 with HNO₃. Subsequent heating at 80 °C for 30 min resulted in a blue solution containing colloidal IrO₂. The photoanodes were immersed in the IrO₂ colloid for 30 min, heated at 300 °C for 30 min and then washed with distilled water before PEC measurements.

Photoelectrochemical (PEC) measurements

PEC measurements were performed in a three electrode configuration in 0.5 M NaOH (pH = 13.0) aqueous solution. The LaTaON₂ photoelectrode was used as the working electrode, while a coiled Pt wire and Ag/AgCl were used as the counter and the reference electrodes, respectively. The light source was a 405 nm laser diode (Laser 2000) with 5 mW power output and a spot size of 3 mm². PEC measurements were carried out on a Solartron 1286 electrochemical interface. Potentiodynamic measurements with a scan rate of 10 mV s⁻¹ in the potential window of 0–1.8 V_RHE were performed to investigate the PEC performance of LaTaON₂. The chopped dark-light current densities were normalized according to the illuminated area of 3 mm². Cyclic voltammetry with a scan rate of 10 mV s⁻¹ in the potential window of 1.0–1.6 V_RHE was used to study the voltage cycling effect on Ni-based cocatalysts.

Results and discussion

Characterization

After the solid state reaction at 1400 °C, the La₂O₃ and Ta₂O₅ powder mixture was completely converted to white LaTaO₄ powder, which corresponds to a mixture of orthorhombic and monoclinic polymorphs as shown in Fig. 1a.²³ The XRD pattern of the ammonolysis product corresponds to LaTaON₂ with an orthorhombic perovskite structure²⁴ with a trace of Ta₃N₅ indicated by an arrow at 28.06°. The LaTaON₂ powders have a dark red color and a specific surface area of ∼5 m² g⁻¹ as measured by the BET method.

Fig. 1 (a) XRD patterns for LaTaO₄ and LaTaON₂. (b) Kubelka–Munk absorbance for LaTaON₂ powder. Inset is the UV-Vis reflectance spectrum.

The Kubelka–Munk absorbance in Fig. 1b derived from a UV-Vis diffuse reflectance spectrum (inset) shows an absorption edge at around 600 nm. Thus the optical bandgap of LaTaON₂ is around 2.1 eV. The Kubelka–Munk absorbances for cocatalyst-loaded LaTaON₂ are also shown in Fig. S1 (ESI†). In comparison with the bare LaTaON₂, only slight changes occur in the absorbance at 405 nm (light source for PEC measurement) for the cocatalysts loaded LaTaON₂.

An electrophoretic deposition method was utilized to immobilize the LaTaON₂ powders from the suspension onto FTO substrates, followed by a necking treatment using TaCl₅ methanol solution and a heat treatment to improve the particle-to-particle and particle-to-substrate contacts by the formation of Ta₂O₅ thin layers, as reported previously.³ Four materials NiO_x, Ni_0.7Fe_0.3O_x, CoO_x and IrO_x were used as cocatalysts on the oxynitrides to investigate their effects on the PEC performance.

Under visible light illumination, Ta₂O₅ does not contribute to the photocurrent due to its large band gap, which was also confirmed in our experiment. Thus, we assume that Ta₂O₅ only helps the charge transport between the LaTaON₂ particles. Fig. 2a shows a schematic illustration of the cross-section for the cocatalyst-loaded LaTaON₂ photoanode.

Fig. 2 (a) Schematic illustration of the cross-section for the cocatalyst-coated LaTaON₂ photoanode. (b–f) SEM images of bare LaTaON₂, NiO_x–LaTaON₂, Ni_0.7Fe_0.3O_x–LaTaON₂, CoO_x–LaTaON₂, and IrO_x–LaTaON₂.

The SEM image in Fig. 2b shows that the bare LaTaON₂ particles with diameters in a range of 50–200 nm, have smooth surfaces. The preloading process generated highly dispersed cocatalyst nanoparticles (NiO_x, Ni_0.7Fe_0.3O_x, and CoO_x in Fig. 2c–e) with diameters less than 10 nm on the LaTaON₂ surface. However, the post-loaded IrO_x nanoparticles (see Fig. 2f) were poorly dispersed on the LaTaON₂ surface, along with a large portion of uncovered surfaces, similar to the post-loaded CoO_x and IrO_x on TaON electrode reported previously.⁴

PEC measurement

Fig. 3 shows the PEC performances of LaTaON₂ photoelectrodes which were treated with TaCl₅ followed by heat treatment in air or NH₃ to form tantalum oxide or oxynitride to connect LaTaON₂ particles. In comparison to the air-treatment, NH₃-treated photoanodes showed an evident increase of the photocurrent due to the improved electronic conductivity after the nitridation of Ta₂O₅ layer. At 1.23 V_RHE, air-treated LaTaON₂ exhibited a photocurrent density of 18.2 μA cm⁻², while a value of 54.4 μA cm⁻² was measured for NH₃-treated LaTaON₂. At potentials higher than 1.6 V_RHE, the voltage-driven OER (dark OER) dominated in the water oxidation, which makes the light-driven photocurrent undetectable.

Fig. 3 Potentiodynamic measurements under chopped light in 0.5 M NaOH (pH = 13.0) for as-prepared LaTaON₂ after TaCl₅ necking treatment and a heat treatment in air at 350 °C (black), under NH₃ flow at 450 °C (red), pre-loaded NiO_x–LaTaON₂ (blue), Ni_0.7Fe_0.3O_x–LaTaON₂ (green), CoO_x–LaTaON₂ (olive) and post-loaded IrO_x–LaTaON₂ (wine) followed by TaCl₅ necking treatment in air at 350 °C. Photocurrent densities were normalized based on the illuminated area of 3 mm².

Since photocurrents on the scale of mA cm⁻² have been reported for TaON,^3–5 Ta₃N₅,³ and BaTaO₂N^6,7 after the decoration with cocatalysts, we expect that higher photocurrent density could be achieved if appropriate cocatalysts are loaded on LaTaON₂. As mentioned above, we loaded NiO_x, Ni_0.7Fe_0.3O_x, CoO_x, and IrO_x cocatalysts on LaTaON₂ by the impregnation method. In order to best represent the buried junction scenario,²⁰ only the first scans of all the samples are shown in Fig. 3 unless otherwise noted.

Although literature has reported the use of NH₃ post-treatment for the cocatalysts loaded photoelectrodes,^3,6 we observed an adverse effect of ammonia on the PEC performance of the system (see Fig. S2, ESI†). All the samples loaded with cocatalysts were therefore only annealed in air.

The pre-loaded NiO_x–LaTaON₂ and Ni_0.7Fe_0.3O_x–LaTaON₂ (see Fig. 3) after necking in air exhibited very similar performances, both higher than the NH₃-treated LaTaON₂. At low bias, NiO_x–LaTaON₂ exhibited a slightly smaller photocurrent than Ni_0.7Fe_0.3O_x–LaTaON₂, while at potentials higher than 1.0 V_RHE, NiO_x–LaTaON₂ showed a gradually larger photocurrent. At 1.23 V_RHE, the photocurrent densities of NiO_x–LaTaON₂ and Ni_0.7Fe_0.3O_x–LaTaON₂ were 140.8 μA cm⁻² and 105.3 μA cm⁻², respectively. In addition, the loading of NiO_x and Ni_0.7Fe_0.3O_x both negatively shifted the onset of the dark OER current to ∼1.4 V_RHE in comparison to ∼1.6 V_RHE for as-prepared LaTaON₂, indicating that they both were well-loaded cocatalysts, which is also confirmed by the SEM images in Fig. 2. Although it has been reported that Fe incorporation increases the electrocatalytic activity of NiO_x,^1,9 our results show that the Ni_0.7Fe_0.3O_x does not result in higher PEC performance than NiO_x at 1.23 V_RHE. One possible explanation for this discrepancy is that the electrolysis of water is mainly limited by the OER kinetics, but other factors related to the charge separation efficiency play more important roles for OER in photoelectrochemical water splitting, such as the surface recombinations of photo-generated charge carriers.

CoO_x and IrO_x both negatively shifted the onset potential of the dark OER to ∼1.45 V_RHE, indicating that they are active OER cocatalysts (Fig. 3). At 1.23 V_RHE, the photocurrent densities of CoO_x–LaTaON₂ and IrO_x–LaTaON₂ were 80.8 μA cm⁻² and 76.6 μA cm⁻², respectively, larger than NH₃-treated LaTaON₂ but still smaller than NiO_x and Ni_0.7Fe_0.3O_x loaded LaTaON₂.

The chopped dark-light current densities were normalized according to the illuminated area. It is important to mention that in this work the light source is a laser with a spot size of 3 mm². On the other hand, not only the illuminated area contribute to the dark current, but the whole area of the photoanode in contact with the electrolyte, which is in our case 1 cm². Taking this into account, it is possible to readily distinguish the different dark behaviors from the various cocatalysts. Obviously, CoO_x and IrO_x both experienced dark oxidation of metal ions at voltages below 1.23 V_RHE, while the Ni-based oxides did not.

According to the Pourbaix diagram, Co²⁺ is oxidized to Co³⁺ between 0.6–1.4 V_RHE,²⁵ which will be further oxidized to Co⁴⁺ by photo-generated holes as evidenced by the anodic transients of CoO_x–LaTaON₂ in Fig. 3.^27,28 Hamann et al. have reported that water oxidation occurs only when a sufficient concentration of Co⁴⁺ species is present.²⁷ While Ir is oxidized to Ir³⁺ hydrous oxide between 0.4–0.8 V_RHE and further to Ir⁴⁺ hydrous oxide between 0.8–1.2 V_RHE.²⁶ Because of the oxidations of surface metal ions, large redox capacitances are observed for these samples. In other words, before the oxidation of water, part of photo-generated holes were “wasted” to oxidize surface Co or Ir species, and less holes participated in PEC water splitting, resulting in lower PEC currents.

As further validations of the experimental results, we also used sputtered metallic Ni and non-catalytic CuO_x as cocatalysts for LaTaON₂ (Fig. S3, ESI†). As expected, metallic Ni also introduced a large pseudocapacitive current, due to the oxidation of Ni to Ni²⁺, while the non-catalytic CuO_x had no detectable impact on the photocurrent.

As mentioned before, the enhancement in PEC performance can be attributed to both the passivation of recombination centers and catalytic improvement of the water oxidation from the cocatalysts. The fact that Ni-based oxides outperform CoO_x and IrO_x can be explained by the higher redox capacitance of the CoO_x and IrO_x, which results in an extra consumption of photo-generated holes. What is not yet clear is whether the catalytic or passivating effect plays a more important role.

H₂O₂ is considered as an efficient hole scavenger in electrolytes to suppress surface recombinations.²⁹ The addition of H₂O₂ into the electrolyte could therefore help to maximize the photocurrent. Thus we also performed potentiodynamic measurements of cocatalysts loaded LaTaON₂ in the presence of H₂O₂, as shown in Fig. S4 (ESI†). Unfortunately with Ni-based oxides, we did not observe any significant improvement in photocurrents comparing measurements with and without H₂O₂ in the electrolyte, indicating that the amount of holes reaching the surface is probably the rate determining factor. This implies that pronounced recombination of photo-generated charge carriers or the low electronic conductivity of the LaTaON₂ bulk could be the bottleneck for limited PEC performances. It is worthwhile developing perovskite oxynitrides with a low number of crystalline defects, e.g., single crystalline layer, to eliminate bulk recombination and get the highest possible PEC performance. In addition, introducing Mo⁶⁺ or W⁶⁺ as donors to the Ta⁵⁺ site can increase the charge carrier concentration and enhance the electronic conductivity to further improve the PEC performance.⁷ With Co- and Ir-oxides, however, the light-driven photocurrent was undetectable due to the very high voltage-driven catalytic activity (dark reaction).^29–32

Voltage cycling effect

Repetitive voltage cycling transforms NiO_x and Ni_0.7Fe_0.3O_x into layered oxyhydroxides, which usually have higher OER catalytic activities than oxides.^1,33 Since NiO_x and Ni_0.7Fe_0.3O_x exhibit very similar behaviors, here we only show the results for Ni_0.7Fe_0.3O_x–LaTaON₂. The transformation is confirmed by the anodic peak at ∼1.44 V_RHE corresponding to the oxidation of α-Ni(OH)₂ to γ-NiOOH, as shown in Fig. 4a.³³ Prior to PEC tests, 50 scans of cyclic voltammetry (CV) were performed between 1.0 V_RHE and 1.6 V_RHE to convert the oxides into oxyhydroxides. A drastic change of the CV behavior for Ni_0.7Fe_0.3O_x–LaTaON₂ occurred during the first 10 scans, but only a slight difference was observed between the 30th and 40th scan, and the 50th scan was basically identical to the 40th, indicating that the oxides have been completely converted into oxyhydroxides.

Fig. 4 (a) Cyclic voltammetry of Ni_0.7Fe_0.3O_x–LaTaON₂ in 0.5 M NaOH (pH = 13.0); comparison of potentiodynamic measurements before and after CV for (b) Ni_0.7Fe_0.3O_x–LaTaON₂ and (c) CoO_x–LaTaON₂. Both dark current and photocurrent densities were normalized based on the illuminated area of 3 mm².

After the CV measurements, however, the system showed a decreased PEC performance, as shown in Fig. 4b. Assuming that the catalytic activity should be higher for oxyhydroxides, a possible explanation could be the change in the surface passivation centers of LaTaON₂. NiO_x and Ni_0.7Fe_0.3O_x at passivation centers have been transformed into oxyhydroxides, destroying the passivation and resulting therefore in a lower PEC performance. We can now conclude that high performance cocatalysts will not work without a good passivation as a prerequisite, i.e., the interface between the photoelectrode and cocatalysts should be stable. It should be noted that although in literature Ni-based oxides have been often used as highly catalytic cocatalysts, the passivation properties and the resulting stability were rarely discussed. Our study shows that it is actually not appropriate to directly contact Ni-based oxides with both the photoelectrode and electrolyte at the same time. An alternative way would be inserting a robust layer with suitable band alignments in-between the photoelectrode and cocatalysts, for instance, a thin layer of TiO₂ or Al₂O₃ deposited by atomic layer deposition, to ensure a stable passivation.^34,35

The voltage cycling effect of CoO_x is shown in Fig. 4c. After the CV measurements, the dark pseudocapacitive current is largely reduced, implying that the oxidation states of surface Co ions have been probably saturated. Compared to the first scan, we also observed a slight decrease in the photocurrent of CoO_x–LaTaON₂, but not as pronounced as for Ni_0.7Fe_0.3O_x–LaTaON₂. In this sense, ion-impermeable CoO_x is a more stable cocatalyst for the passivation of the photoelectrode–cocatalyst interface.

In summary, a good surface passivation is required for photoelectrodes to exhibit high PEC performance. The high catalytic activity of cocatalysts does not improve the PEC performance unless the surface recombination centers of the photoelectrodes are passivated. It is therefore recommended to first passivate the surface with an inert thin layer of e.g. TiO₂ or Al₂O₃ and then load the cocatalysts to catalyze water oxidation.

Conclusions

We have investigated the different behaviors of various cocatalysts on LaTaON₂ photoelectrodes. NiO_x and Ni_0.7Fe_0.3O_x result in higher photocurrents than CoO_x and IrO_x, due to the fact that the latter consume more photo-generated holes for the oxidation of metal ions. Repetitive voltage cycling transforms the NiO_x and Ni_0.7Fe_0.3O_x into oxyhydroxides with higher catalytic activities, however, resulting in lower photocurrents. This is ascribed to the degradation of the passivation centers at LaTaON₂–cocatalyst interface during the formation of oxyhydroxides. A robust and stable passivation layer should therefore be inserted between the photoelectrode and cocatalysts especially when Ni-based oxides are used. The limited improvement on the PEC performance observed using various cocatalysts suggests that the most important overall rate determining process in LaTaON₂ are a strong bulk recombination and low electronic conductivity. To further improve the PEC performance of perovskite oxynitrides, perovskite structure with a low number of defects and high electronic conductivity should be selected.

Acknowledgements

W. Si would like to thank colleague Markus Pichler and Christof Schneider for discussions and technical help. This research was supported by the Paul Scherrer Institute and the NCCR MARVEL, funded by the Swiss National Science Foundation.

References

1. M. S. Burke, L. J. Enman, A. S. Batchellor, S. Zou and S. W. Boettcher, Chem. Mater., 2015, 27, 7549–7558.
2. C. C. L. McCrory, S. Jung, J. C. Peters and T. F. Jaramillo, J. Am. Chem. Soc., 2013, 135, 16977–16987.
3. M. Higashi, K. Domen and R. Abe, Energy Environ. Sci., 2011, 4, 4138–4147.
4. M. Higashi, K. Domen and R. Abe, J. Am. Chem. Soc., 2012, 134, 6968–6971.
5. E. S. Kim, N. Nishimura, G. Magesh, J. Y. Kim, J.-W. Jang, H. Jun, J. Kubota, K. Domen and J. S. Lee, J. Am. Chem. Soc., 2013, 135, 5375–5383.
6. M. Higashi, K. Domen and R. Abe, J. Am. Chem. Soc., 2013, 135, 10238–10241.
7. M. Higashi, Y. Yamanaka, O. Tomita and R. Abe, APL Mater., 2015, 3, 104418.
8. C. Wang, T. Hisatomi, T. Minegishi, Q. Wang, M. Zhong, M. Katayama, J. Kubota and K. Domen, J. Phys. Chem. C, 2016, 120, 15758–15764.
9. C. G. Morales-Guio, M. T. Mayer, A. Yella, S. D. Tilley, M. Grätzel and X. Hu, J. Am. Chem. Soc., 2015, 137, 9927–9936.
10. Y. Liang and J. Messinger, Phys. Chem. Chem. Phys., 2014, 16, 12014–12020.
11. J. Yang, D. Wang, H. Han and C. Li, Acc. Chem. Res., 2012, 46, 1900–1909.
12. R. Abe, M. Higashi and K. Domen, J. Am. Chem. Soc., 2010, 132, 11828–11829.
13. K. Maeda, M. Higashi, B. Siritanaratkul, R. Abe and K. Domen, J. Am. Chem. Soc., 2011, 133, 12334–12337.
14. S. Landsmann, A. E. Maegli, M. Trottmann, C. Battaglia, A. Weidenkaff and S. Pokrant, ChemSusChem, 2015, 8, 3451–3458.
15. L. Zhang, Y. Song, J. Feng, T. Fang, Y. Zhong, Z. Li and Z. Zou, Int. J. Hydrogen Energy, 2014, 39, 7697–7704.
16. L. Trotochaud, J. K. Ranney, K. N. Williams and S. W. Boettcher, J. Am. Chem. Soc., 2012, 134, 17253–17261.
17. L. Trotochaud, S. L. Young, J. K. Ranney and S. W. Boettcher, J. Am. Chem. Soc., 2014, 136, 6744–6753.
18. T. W. Kim and K.-S. Choi, Science, 2014, 343, 990–994.
19. F. Lin and S. W. Boettcher, Nat. Mater., 2014, 13, 81–86.
20. F. Lin, B. F. Bachman and S. W. Boettcher, J. Phys. Chem. Lett., 2015, 6, 2427–2433.
21. T. J. Mills, F. Lin and S. W. Boettcher, Phys. Rev. Lett., 2014, 112, 148304.
22. Y. Zhao, E. A. Hernandez-Pagan, N. M. Vargas-Barbosa, J. L. Dysart and T. E. Mallouk, J. Phys. Chem. Lett., 2011, 2, 402–406.
23. N.-Y. Park and Y.-I. Kim, J. Mater. Sci., 2012, 47, 5333–5340.
24. S. H. Porter, Z. Huang and P. M. Woodward, Cryst. Growth Des., 2014, 14, 117–125.
25. E. M. Garcia, J. S. Santos, E. C. Pereira and M. B. J. G. Freitas, J. Power Sources, 2008, 185, 549–553.
26. J. Juodkazytė, B. Šebeka, I. Valsiunas and K. Juodkazis, Electroanalysis, 2005, 17, 947–952.
27. B. Klahr, S. Gimenez, F. Fabregat-Santiago, J. Bisquert and T. W. Hamann, J. Am. Chem. Soc., 2012, 134, 16693–16700.
28. B. Klahr, S. Gimenez, F. Fabregat-Santiago, J. Bisquert and T. W. Hamann, Energy Environ. Sci., 2012, 5, 7626–7636.
29. H. Dotan, K. Sivula, M. Gratzel, A. Rothschild and S. C. Warren, Energy Environ. Sci., 2011, 4, 958–964.
30. B. P. Minks, G. Oskam, D. Vanmaekelbergh and J. J. Kelly, J. Electroanal. Chem. Interfacial Electrochem., 1989, 273, 119–131.
31. A. Theuwis, I. E. Vermeir and W. P. Gomes, J. Electroanal. Chem., 1996, 410, 31–42.
32. F. F. Abdi and R. van de Krol, J. Phys. Chem. C, 2012, 116, 9398–9404.
33. L.-A. Stern and X. Hu, Faraday Discuss., 2014, 176, 363–379.
34. F. Le Formal, N. Tetreault, M. Cornuz, T. Moehl, M. Gratzel and K. Sivula, Chem. Sci., 2011, 2, 737–743.
35. T. Wang, Z. Luo, C. Li and J. Gong, Chem. Soc. Rev., 2014, 43, 7469–7484.

---

Footnote

† Electronic supplementary information (ESI) available: Potentiodynamic measurements for NiO_x–LaTaON₂ after treatment in air and under NH₃, potentiodynamic measurements for sputtered Ni–LaTaON₂ and CuO_x–LaTaON₂, PEC measurements for all samples in 0.5 M NaOH with the presence of 0.1 M H₂O₂. See DOI: 10.1039/c6cp07253d

---