Effects of interfacial layers on the photoelectrochemical properties of tantalum nitride photoanodes for solar water splitting

Abstract

This work describes the effects of interlayers on the structural, crystalline and photoelectrochemical properties of Ta₃N₅ photoanodes. Nb₄N₅ interlayers with different thicknesses are formed between Ta₃N₅ films and Ta back substrates using a thin film transfer method. Zone-axis images and electron diffraction patterns of cross-sections of the resulting Ta₃N₅/Nb₄N₅/Ta/Ti electrodes acquired by transmission electron microscopy evidence the formation of a 200 nm thick oriented Ta₃N₅ grain layer in the vicinity of the interlayer. This oriented Ta₃N₅ layer promotes electron transport throughout the Ta₃N₅ film. As a result, a Ta₃N₅ photoanode incorporating a Nb₄N₅ interlayer exhibits a higher photocurrent during the oxygen evolution reaction. These results obtained using the Ta₃N₅/interlayer/substrate configuration elucidate the key role played by the buffer layer in achieving efficient water splitting with Ta₃N₅ photoelectrodes.

---

Introduction

Photoelectrochemical (PEC) water splitting is a process used to store solar light energy via the formation of hydrogen.^1–3 Ta₃N₅, an n-type semiconductor, has been widely applied as a photoanode for the PEC oxygen evolution reaction (OER). With a bandgap of 2.1 eV, Ta₃N₅ absorbs visible light up to 600 nm (ref. 4) and can potentially yield a maximum photocurrent of 12.5 mA cm⁻² under irradiation by sunlight (AM 1.5G).

Photoanodes are typically prepared as either thin films,⁵ nanostructures (such as nanorods and nanotubes)^6–8 or particulate layers.^9,10 In addition, doping, surface modification and other strategies have been applied to achieve high photocurrents and negative onset potentials during PEC water oxidation.^11–16 Theoretical studies have indicated that Ta₃N₅ is an anisotropic semiconductor material in which holes are heavier than electrons,¹⁷ resulting in low mobility of the minority charge carriers (i.e., the holes), while the porosity of the Ta₃N₅ layer in a Ta₃N₅ photoanode is believed to limit the bulk transport of the majority charge carriers (electrons).¹⁸ Controlling both the structure and crystallinity of Ta₃N₅ films is a critical aspect of improving the PEC performance of this material because these parameters significantly affect the charge separation efficiency as well as the light absorption of the Ta₃N₅.^17–20

Studies on the growth of GaN films suggest that a GaN or AlN buffer layer will generate high crystalline quality in the upper GaN film,^21,22 and a similar effect is anticipated in the case of Ta₃N₅ film growth. In order to fabricate semiconductor/conductor configurations for use as photoelectrodes, Ta₃N₅ films are generally grown on conductive substrates, such as metallic Ta foils or Pt,²³ by nitridation under a flow of NH₃ gas and at high temperatures (800–1000 °C). To date, Ta₃N₅ photoanodes with state-of-the-art PEC performance have been fabricated on Ta foil substrates.^11,24 However, an impurity interlayer consisting of TaN_x phases is inevitably formed between the Ta₃N₅ film (or nanorods and nanotubes) and the underlying Ta foil during the nitridation process.^6,20 The effects of this interlayer on electron transport in the Ta₃N₅ film and at the Ta₃N₅/TaN_x interface are expected to alter the PEC performance of Ta₃N₅ photoanodes.

In the present work, the effects of interlayers on the growth of Ta₃N₅ films and on the PEC performance of Ta₃N₅ photoanodes were assessed by introducing a niobium nitride (NbN_x) layer at the interface between the Ta₃N₅ film and the conductive substrate by a thin film transfer technique.²⁵ Inert Si wafers were used as synthesis platforms for the formation of the NbN_x and Ta₃N₅ bilayer films. After the film transfer process, the NbN_x surface layer acted as a back contact to conduct photogenerated electrons from the Ta₃N₅ film to the back conductive layer, and the introduction of a NbN_x interlayer was found to enhance the photocurrent of the Ta₃N₅ photoanodes. The growth mechanism of the Ta₃N₅ films in the presence of the NbN_x layer and the effects of the interlayer on the structural, crystalline and PEC properties of the Ta₃N₅ films are discussed based on the results of this study.

Experimental section

Sample preparation

A NbN_x layer was introduced as an interlayer between Ta₃N₅ and a conductive substrate using a film transfer process, as depicted in Fig. 1. The sections of a single-crystal Si (100) substrate (1 × 1 cm, Nilaco) were cleaned by sequential ultrasonic irradiation over 10 min each in acetone, isopropanol and Milli-Q water (18.2 mΩ cm). Ta and Nb films with the desired thicknesses were sequentially deposited on the Si substrates by radio frequency (RF) magnetron sputtering (ULVAC, MNS-2000-RFG3) and the resulting multi-layered films were oxidized in a muffle furnace at 700 °C for 2 h with a temperature ramp rate of 15 °C min⁻¹. The Nb₂O₅/Ta₂O₅/Si samples fabricated in this manner were subsequently moved to a tube furnace and nitrided in a flow of NH₃ gas (100 sccm) at 900 °C for 2 h with a temperature ramp rate of 20 °C min⁻¹, forming a nitride bilayer on the Si substrates. To transfer these nitride films onto a conductive substrate, metallic Ta (150 nm) and Ti (approximately 5 μm) layers were sequentially deposited on the NbN_x/Ta₃N₅/Si sample by RF magnetron sputtering, serving as the contact and conductive layers, respectively. Because the physical contact between the Ta₃N₅ film and the Si substrate was weakened by the strain imparted upon adding the metallic Ta/Ti layers, the Ti/Ta/NbN_x/Ta₃N₅ films were readily peeled away from the underlying Si substrate using carbon tape supported with glass plates. For comparison purposes, a Ta₃N₅/Ta/Ti film without a NbN_x interlayer was synthesized in the same manner. The electrode areas exposed to the electrolyte solution were approximately 0.2 cm². Each of the Ta₃N₅ films used in this work was 630 nm thick, while the thickness of the NbN_x layer varied from 50 to 200 nm.

Fig. 1 Schematic diagram of the procedure used to synthesize a Ta₃N₅ film with a NbN_x interlayer. (i) Sputtering of Ta and Nb metallic films, (ii) oxidation and nitridation of the Ta/Nb films, (iii) sputtering of Ta/Ti metallic layers, and (iv) exfoliation of the NbN_x/Ta₃N₅ films from the Si substrate.

Characterization

Scanning electron microscopy (SEM, S-4700, Hitachi), transmission electron microscopy and scanning transmission electron microscopy (TEM, STEM, JEM-2800, JEOL) with energy dispersive X-ray fluorescence spectroscopy (EDX, EX-24055JGT, JEOL) were used to characterize the morphologies and structures of the thin films. Samples for cross-sectional SEM and STEM assessments were prepared by ion milling and focused ion beam (FIB) milling, respectively. X-ray diffraction (XRD) patterns were recorded using an X-ray diffractometer (Rigaku Ultima III) with Cu Kα radiation (λ = 0.15405 nm, 40 mV and 40 mA). The average crystallite size in the Ta₃N₅ films (n = 3, ±σ) was calculated using the Scherrer equation. X-ray fluorescence spectroscopy (XPS, JEOL, JPS-90SX) was performed using Mg Kα radiation (8 kV and 10 mA), employing the C 1s peak at 284.8 eV as an internal standard to calibrate the binding energies.

Electrochemical measurement

Prior to the PEC measurements, each Ta₃N₅ photoelectrode was deposited with a thin layer of Co(OH)_x using an impregnation method. This was accomplished by adding an aqueous NaOH solution (50 mM, 2.5 mL) to an aqueous Co(NO₃)₂ solution (6.3 mM, 10 mL) and immersing the Ta₃N₅ electrodes in the mixture for 1 h. The PEC properties of the Ta₃N₅ photoanodes were measured using a three-electrode cell with a Ag/AgCl reference electrode (in saturated aqueous KCl) and Pt wire as the counter electrode. The Ag/AgCl reference potential was converted to the reversible hydrogen electrode (RHE) potential according to the Nernst relationship

E (V vs. RHE) = E (V vs. Ag/AgCl) + 0.059 × pH + 0.197, (1)

where 0.197 represents the standard potential of the KCl-saturated Ag/AgCl electrode at 25 °C. An aqueous solution of potassium phosphate (KPi, 0.5 M, pH 13) was used as the electrolyte. During these trials, the Ta₃N₅ photoelectrodes were illuminated by using a chopped AM 1.5G solar simulator (San-EI Electric, XES-40S2-CE). Linear-sweep voltammetry (LSV) was employed, scanning from 1.5 to 0 V vs. RHE to evaluate the photocurrent density and onset potential. The photocurrent decay of each Ta₃N₅ photoelectrode was examined in the chronoamperometric mode at 1.0 V vs. RHE. Evolved gaseous H₂ and O₂ were simultaneously collected in an air-tight cell and quantified using a micro gas chromatograph (Agilent, 3000A, Micro GC) to calculate the faradaic efficiency.

The incident photon-to-current conversion efficiency (IPCE) was measured under monochromatic irradiation from a Xe lamp (MAX-302, Asahi Spectra) in a 0.5 M KPi electrolyte. The wavelength-dependent IPCEs were calculated using the equation

IPCE = 1240 × (I_light − I_dark)/(λ × P_input) × 100% (2)

where λ (nm) is the wavelength of the monochromatic irradiation, I_light (mA cm⁻²) is the chronoamperometric photocurrent density, I_dark (mA cm⁻²) is the dark current density and P_input (mW cm⁻²) is the incident photon density as determined by using a Si photodiode detector.

To compare the relative electrochemically active surface areas and roughness factors of different Ta₃N₅ photoelectrodes, the scan-rate dependencies of cyclic voltammograms were estimated according to a previously reported method.²⁶ All of the electrodes were tested at different scan rates, varying from 5 to 300 mV s⁻¹, with a capacitive current estimated at −0.1 V vs. Ag/AgCl. Mott–Schottky measurements were performed in a 0.5 M KPi electrolyte at a frequency of 1000 Hz with an AC amplitude of 10 mV, using an impedance analyzer (METEK, VersaSTAT3-200). The flat band potential and donor dopant density values of the Ta₃N₅ photoanodes were determined according to the Mott–Schottky relationship

where C_SC (F cm⁻²) is the space-charge capacity per surface area, ε₀ (C V⁻¹ cm⁻¹) is the permittivity of free space, ε is the dielectric constant of Ta₃N₅, N_D (cm⁻³) is the donor density, r is the roughness factor of the Ta₃N₅ electrode, k (1.38 × 10⁻²³ J K⁻¹) is the Boltzmann constant, E (V vs. RHE) is the applied potential, and E_fb (V vs. RHE) represents the flat band potential of the Ta₃N₅ electrode. The flat band potential could be derived from the x-axis intercept, while the donor density could be obtained from the slope of a Mott–Schottky plot. Electrochemical impedance spectra (EIS) were measured from 100 kHz to 0.1 Hz using an impedance analyzer under simulated AM 1.5G light in a 0.5 M KPi electrolyte. To alleviate the photocorrosion of Ta₃N₅ photoelectrodes, all the samples were tested in the presence of sacrificial reagent 0.5 M SO₃²⁻ at 1.0 V vs. RHE. The obtained Nyquist plots were fitted from 100 kHz to 1 Hz with an equivalent circuit model to rule out the fluctuations at frequency lower than 1 Hz.

Results and discussion

Structure and crystallinity

A top-view SEM image of a Ta₃N₅/NbN_x/Ta/Ti sample after the thin film transfer is shown in Fig. 2(a). The surface of the Ta₃N₅ film exhibits roughness originating from the porous contact at the Ta₃N₅/Si interface in the parent NbN_x/Ta₃N₅/Si sample. The possible exposure of the NbN_x interlayer to the film surface is ruled out by the XPS data obtained from the Ta₃N₅/NbN_x/Ta/Ti film, as shown in Fig. S1.† In addition, the absence of Ta species on the surface of the parent NbN_x/Ta₃N₅/Si sample suggests that the Ta₃N₅ film was not in direct contact with the underlying metallic substrate. The layered structure of the Ta₃N₅/NbN_x/Ta/Ti sample is clearly observed in the cross-sectional SEM image presented in Fig. 2(b). Here, the NbN_x layer is evident as a region of darker contrast, with a thickness of approximately 100 nm. In the magnified view, a porous structure is observed throughout the Ta₃N₅/NbN_x film, due to the oxidation of the metallic film and the substitution of O²⁻ ions by N³⁻ ions during the nitridation process, just as occurs during the nitridation of Ta₂O₅ powder to Ta₃N₅. Despite its porous structure, the NbN_x layer is in intimate contact with the Ta₃N₅ film and the Ta contact layer. The EDX mapping images (Fig. 3) demonstrate that the distribution of Nb was confined to the interlayer region, so there was no appreciable doping of Nb into the Ta₃N₅ film.

Fig. 2 (a) Top-view and (b) cross-sectional SEM images of a Ta₃N₅/NbN_x/Ta/Ti film. The scale bars are (a) 5 μm, (b) 500 nm, and 100 nm for the zoomed image.

Fig. 3 (a) Cross-sectional STEM image of a Ta₃N₅/NbN_x/Ta/Ti film and (b)–(e) EDX elemental mappings of the image in (a). All scale bars are 500 nm.

Fig. 4 presents the XRD patterns of NbN_x/Ta₃N₅/Si samples with different NbN_x thicknesses prior to the thin film transfer, as a means of confirming the nitridation products. A Ta₃N₅ phase was evidently formed regardless of whether a NbN_x layer was present or not. In addition, impurity phases such as TaN_x or SiN_x were absent. The diffraction pattern of the NbN_x layer deposited on a Si substrate (Fig. 4(e)) suggests that Nb₄N₅ was the main phase formed in the NbN_x/Ta₃N₅/Si sample, although the crystallinity of this specimen was low, and the diffraction peaks of the (101) and (211) planes of the Nb₄N₅ overlap with those generated by the (110) and (113) planes of the Ta₃N₅ at 24.4° and 36.0°, respectively. A peak attributable to NbN_x (x ≤ 1) was also detected at 34.0° when the thickness of the NbN_x layer was at its maximum (200 nm). Nevertheless, the XRD peaks obtained from the NbN_x/Ta₃N₅/Si primarily provide information regarding the Ta₃N₅ phase because the Ta₃N₅ peaks are much stronger than those generated by the Nb₄N₅ and other NbN_x phases. It should also be noted that Nb₃N₅ was not observed, as expected because Nb⁵⁺ is prone to reduction to Nb⁴⁺ or lower-valence nitride species.²⁷

Fig. 4 XRD patterns of (a) Ta₃N₅/Si, (b)–(d) NbN_x/Ta₃N₅/Si with NbN_x thicknesses of (b) 50, (c) 100, and (d) 200 nm, and (e) NbN_x (100 nm)/Si samples. XRD patterns of Nb₄N₅ (PDF# 74-0606) and Ta₃N₅ (PDF# 79-1533) are presented at the bottom and the top, respectively, as references.

The crystallite sizes of the Ta₃N₅ films were calculated from the (110) diffraction peak at 24.4°. The data in Table 1 demonstrate that each of the Ta₃N₅ films had almost identical average crystallite sizes, in the range of 40–45 nm. Thus, the effect of the NbN_x layer on the crystallite size in the Ta₃N₅ film was not significant. Notably, the intensity ratio (R_in) obtained from the peaks at 24.4° and 36.0°, which correspond to the (110) and (113) planes of Ta₃N₅, decreased from 2.4 to 0.5–0.8 when NbN_x layers were present on the Ta₃N₅ films.

Table 1 Average crystallite sizes in Ta₃N₅ films with different NbN_x layer thicknesses, as estimated from XRD peaks at 24.4°, as well as the ratios of the (110) and (113) diffraction intensities of the peaks at 24.4 and 36.0°

NbNx thickness (nm)	Crystallite size (nm)	Peak intensity ratio
0	42.6 ± 1.8	2.4 ± 0.2
50	44.3 ± 1.8	0.5 ± 0.1
100	41.8 ± 4.4	0.7 ± 0.2
200	42.0 ± 2.7	0.8 ± 0.1

---

The crystalline growth of each Ta₃N₅ film was examined in detail by acquiring cross-sectional TEM images of Ta₃N₅/NbN_x films (Fig. 5). These bright field (BF) zone-axis TEM images provide a clear view of the multi-layered structure of the Ta₃N₅/NbN_x/Ta samples, and the bright spots scattered within the darker contrast can provide information regarding the sizes and boundaries of individual grains. In Fig. 5(a) and (b), a high density of bright spots is observed throughout the 100 nm-thick NbN_x interlayer, suggesting that the grain size of the NbN_x phase was much less than the layer thickness of 100 nm (likely below 10 nm). This result agrees with the low-crystalline Nb₄N₅ phase indicated by the XRD patterns in Fig. 4. Notably, the Ta₃N₅ region close to the Ta₃N₅/NbN_x interface generated fewer bright spots when over 200 nm in thickness, as shown in Fig. 5(c), indicating the presence of fewer boundaries among the Ta₃N₅ grains near the interface with the NbN_x layer, while the rest of the Ta₃N₅ film exhibited a number of bright spots.

Fig. 5 (a) Zone-axis TEM images of the cross-section of a Ta₃N₅ film on a Ta/Ti substrate with a NbN_x interlayer. The region (200 nm thick) within the white dashed lines represents the oriented Ta₃N₅ layer. SAED patterns of Ta₃N₅ crystallinites near (b) the surface of the Ta₃N₅ film, and (c) the interface of the Ta₃N₅/NbN_x. The insets are diffraction patterns from the selected regions highlighted by the circles. All scale bars are 200 nm.

Selected area electron diffraction (SAED) patterns taken from the regions at 100 nm to the film surface and 100 nm to the Ta₃N₅/NbN_x interface are displayed in the insets of Fig. 5(b) and (c), respectively. The Ta₃N₅ near the interface with the NbN_x generated a more ordered diffraction pattern compared to the Ta₃N₅ close to the surface, which was characterized by random diffractions. This finding indicates that the arrangement of the Ta₃N₅ grains was well oriented in the vicinity of the NbN_x interlayer. It should be noted that crystallite orientation was not observed at the interface between the Ta₃N₅ film and the Ta/Ti substrate of the Ta₃N₅/Ta/Ti film without a NbN_x interlayer, as seen in Fig. S2.†

As suggested by the SAED pattern, the formation of a highly oriented 200 nm Ta₃N₅ layer is not continuous throughout the entirety of the 630 nm Ta₃N₅ film. The growth mode of the Ta₃N₅ film in the presence of the NbN_x interlayer appears to be consistent with that observed for Ta₃N₅ films directly grown on Ta foil substrates by thermal nitridation, as reported by Pinaud et al.¹⁸ The electrochemically active surface areas (ECSAs) of the Ta₃N₅ films fabricated by Pinaud's group were found to exhibit minimal change up to a film thickness of 260 nm due to the growth of a compact Ta₃N₅ film. However, the ECSA increased by one or two orders of magnitude when the film thickness was increased from 260 to 630 nm, owing to the growth of disordered layers over top of the compact Ta₃N₅ films. The formation of an oriented Ta₃N₅ layer was also observed in the present Ta₃N₅ films grown on Ta foil substrates (Fig. S3†). Therefore, it is believed that both Ta₃N₅ films grown on Ta substrates and films nitrided in contact with a NbN_x interlayer have similar growth mechanisms, presumably involving Ta₄N₅ and Nb₄N₅ with the same crystal structures^28,29 acting as interlayers, respectively.

Photoelectrochemical properties

Current–potential curves obtained from the Ta₃N₅ film photoelectrodes deposited with a layer of Co(OH)_x as an OER catalyst are shown in Fig. 6(a). In the absence of the NbN_x interlayer, the Ta₃N₅ photoelectrode exhibited an anodic photocurrent of approximately 2.0 mA cm⁻² at 1.23 V vs. RHE. The photocurrent densities increased significantly for the Ta₃N₅ films with NbN_x interlayers 50–100 nm in thickness. However, the higher resistance of the NbN_x (Nb₄N₅) interlayer compared to that of the metallic Ta substrate lowered the electron mobility. As a result, the photocurrent decreased as the thickness of the NbN_x interlayer was increased from 100 to 200 nm. Moreover, the onset potential of the photoanodic current generated by the Ta₃N₅/NbN_x (100 nm)/Ta/Ti electrode shifted cathodically by 0.1 V from 0.9 V vs. RHE relative to the Ta₃N₅/Ta/Ti electrode. The NbN_x (Nb₄N₅ or NbN) phase acts as a conductive contact for the transport of the majority charge carriers (electrons) collected from the Ta₃N₅ film.

Fig. 6 (a) Current–potential curves obtained from Ta₃N₅/Ta/Ti and Ta₃N₅/NbN_x/Ta/Ti photoelectrodes with different NbN_x interlayer thicknesses, (b) hydrogen and oxygen gas evolution with respect to the total charge passed, as calculated from the (c) current–time curve obtained from the Co(OH)_x/Ta₃N₅/NbN_x (100 nm)/Ta/Ti photoelectrode held at 1.0 V vs. RHE. The PEC activities and gas evolution were assessed in a 0.5 M potassium phosphate solution (pH 13) under simulated AM 1.5G light.

To confirm that the enhancement of the photocurrent of Ta₃N₅/NbN_x photoanodes is associated with the PEC water oxidation performance, gaseous H₂ and O₂ evolved at the counter electrode and the photoanode at 1.0 V vs. RHE were analyzed using a micro gas chromatograph (Fig. 6(b)). It was confirmed that the amounts of H₂ and O₂ were virtually equal to the values calculated based on the total charge passed through the system. In addition, the ratio of H₂ to O₂ was close to 2 : 1. These results indicate that the faradaic efficiency during the OER process was unity when employing the Ta₃N₅/NbN_x photoanode. However, the gas evolution rates decreased in accordance with changes in the photocurrent over time, as can be seen in Fig. 6(c). The decreased photocurrent and O₂ evolution rate likely resulted from dissolution of the Co-based catalysts from the surface of the Ta₃N₅ film into the electrolyte solution.³⁰ The loss of the Co-based catalyst led to direct exposure of the Ta₃N₅ film to the electrolyte solution, and consequently to gradual degradation of the photoanode because bare Ta₃N₅ photoanodes are prone to photocorrosion.³¹ A uniform coverage of the Ta₃N₅ electrode with protective and catalytic layers is likely to stabilize the photocurrent. In addition, the Co-based catalyst can be replaced by other oxygen-evolution catalysts (e.g. NiO_x), which are electrochemically stable against the dissolution under alkaline conditions.

The IPCE spectra shown in Fig. 7 are consistent with the increases in photocurrent densities brought about by the NbN_x interlayers (see Fig. 6(a)). The onset of a photoresponse at 600 nm corresponds to the direct band gap of Ta₃N₅ at 2.1 eV. It should be noted that a shoulder is observed at 500 nm in each of the IPCE plots. This is also ascribed to the direct transition of Ta₃N₅ between the N 2p and Ta 5d states, as was suggested by earlier optical and theoretical studies of Ta₃N₅ films.^17,32 The IPCEs of Ta₃N₅/NbN_x photoelectrodes at 400–550 nm were evidently increased compared to the bare Ta₃N₅ photoelectrode. Because less than 5% of light can traverse through the Ta₃N₅ film to the NbN_x interlayer below 550 nm (see Fig. S4†), it is unlikely that the photoresponse of the Ta₃N₅/NbN_x electrode was affected by the light absorption characteristics of the NbN_x interlayer and the Ta₃N₅/NbN_x interface. Therefore, it is thought that the enhancements in the IPCE and the photocurrent density of the Ta₃N₅/NbN_x photoelectrodes are primarily due to the changes in the structural features of the Ta₃N₅ film induced by the NbN_x interlayer.

Fig. 7 IPCE spectra of Ta₃N₅/Ta/Ti and Ta₃N₅/NbN_x/Ta/Ti photoelectrodes with different NbN_x interlayer thicknesses.

Charge separation and transport

In the PEC water oxidation process, light harvesting, charge separation and transport, and hole injection are the three key processes that define the PEC activity and IPCE. The light harvesting of the Ta₃N₅ films would not be expected to have been altered by the NbN_x interlayers because the film thickness and crystallite size were unchanged. It is also believed that the hole injection process was not affected by the presence of the interlayer because of the similarities in the roughness factors, flat band potentials, and donor densities. ECSA values were estimated from the double-layer charging in the cyclic voltammograms (CVs) as a means of examining the surface roughness of the Ta₃N₅ electrodes (Fig. S5†), and the roughness factor was found to be barely changed by the presence of the NbN_x interlayer. Mott–Schottky plots (Fig. S6†) were used to estimate the flat band potentials and the carrier densities, with values of −0.06 and −0.08 V vs. RHE, respectively. Hence, charge separation and transport are evidently the major factors contributing to the enhancement of the photocurrent by the NbN_x interlayer. Electrochemical impedance spectra (EIS) of Ta₃N₅ and Ta₃N₅/NbN_x electrodes are shown in Fig. 8. The Nyquist plots with a feature of single capacitance were fitted using the Randles equivalent circuit model, similar to the case for hematite photoanodes in the presence of [Fe(CN)₆]^3−/4− redox couples.³³ This equivalent circuit consists of the series resistance (R_s) of the NbN_x and Ta/Ti layers, the capacitance of the bulk Ta₃N₅ film (CPE), and the charge transfer resistance from the valence band of Ta₃N₅ to the solution (R_CT), which may be associated with resistances representing bulk recombination and charge transfer at the electrolyte/semiconductor interface. The values of the charge transfer resistance (R_CT, Table S1†) of all the Ta₃N₅/NbN_x photoelectrodes were smaller than that of the unmodified Ta₃N₅ photoelectrode although the surface reaction was the same, oxidation of sulfite ions on the Ta₃N₅ surface. This result suggests an increase in the population of holes that survive recombination with electrons and successfully reach the semiconductor/liquid junction as a result of promotion of charge separation and transport process by the NbN_x interlayers. The R_s increased with the thickness of the NbN_x layer because of its lower conductivity. However, the contribution of R_s to the total resistance is negligible.

Fig. 8 Nyquist plots of electrochemical impedance spectra of Ta₃N₅ and Ta₃N₅/NbN_x electrodes measured under simulated AM 1.5G light in a 0.5 M KPi solution (pH 13) with a sacrificial reagent (0.5 M SO₃²⁻) at 1.0 V vs. RHE. The inset represents the Randles equivalent circuit used for fitting the Nyquist plots.

Two charge generation and transport scenarios can be considered, depending on whether photons are absorbed near the Ta₃N₅ film surface (case i) or the Ta₃N₅/NbN_x interface (case ii), as depicted in Fig. 9. Long-distance electron transport and short-distance hole transport occur in the former case (i) while long-distance hole transport and short-distance electron transport take place in the latter case (ii). The analysis of the integrated photocurrent calculated on the basis of the standard AM 1.5G (ASTM G173-03) spectrum and the IPCE spectra (Fig. S7†) indicates that the short-wavelength light absorbed near the Ta₃N₅/NbN_x electrode surface showed a larger contribution to the enhancement in photocurrent than the long-wavelength light absorbed near the Ta₃N₅/NbN_x interface. The formation of an oriented Ta₃N₅ layer (200 nm thick) by the NbN_x interlayer probably promotes the electron mobility in the Ta₃N₅ film near the Ta₃N₅/NbN_x interface. Therefore, photogenerated charges near the surface may contribute to the PEC water oxidation more effectively in the presence of a NbN_x interlayer, given that the migration distance required for holes to reach the surface is shorter in a disordered Ta₃N₅ layer. Nevertheless, it should be pointed out that the photons absorbed near the Ta₃N₅/NbN_x interface are not necessarily utilized efficiently even in the presence of a NbN_x interlayer because holes generated deeper in the Ta₃N₅ film have a higher probability of recombining with electrons before reaching the surface, since the hole mobility is lower than the electron mobility.¹⁷ As a consequence, the photocurrent observed in the present study was still lower than the maximum theoretical value for Ta₃N₅, so it is most likely necessary to tune the thicknesses of the oriented and disordered Ta₃N₅ layers.

Fig. 9 Schematic depicting the separation and transport of photogenerated electrons and holes in (a) Ta₃N₅/Ta and (b) Ta₃N₅/NbN_x/Ta films.

A photoanode configuration consisting of a lower oriented Ta₃N₅ layer and an upper disordered Ta₃N₅ layer represents a combination of a short migration distance for photoexcited holes and efficient electron transport toward the back conductor layer. The introduction of the NbN_x interlayer effectively realizes this ideal photoanode configuration and thus enhances the PEC performance. This result suggests that the structural and PEC properties of Ta₃N₅ photoanodes can be improved by designing appropriate interlayers, regardless of the nitridation of the Ta₃N₅ film and the particular back conductor material employed. The film transfer method developed in this study presents a means of studying such aspects owing to the controllability of the thickness and the stacking order of multilayers.

Conclusions

NbN_x layers principally consisting of Nb₄N₅ were introduced as interlayers between Ta₃N₅ films and the back substrate, using the film transfer method. The Nb₄N₅ interlayer did not change the phase purities and the crystallite sizes of the Ta₃N₅ films significantly, although the preference in the crystalline orientation of the Ta₃N₅ film was enhanced by the Nb₄N₅ phase. The photocurrents obtained from Ta₃N₅ photoanodes during the PEC water oxidation reaction increased significantly in the presence of the NbN_x interlayer. The results of zone-axis imaging and electron diffraction from cross-sectional TEM revealed that ordered growth of the Ta₃N₅ film occurred and that a 200 nm layer of highly grain-oriented Ta₃N₅ was formed near the Ta₃N₅/NbN_x interface. The formation of an oriented Ta₃N₅ layer did not affect the ECSA, flat band potential or carrier density of a 630 nm Ta₃N₅ film. The increased photocurrent associated with the presence of a NbN_x interlayer is primarily attributed to the increased long distance electron mobility in the Ta₃N₅ film. The film growth on the Nb₄N₅ layer observed in this work demonstrates the key roles of a buffer layer in the design and synthesis of efficient Ta₃N₅ photoelectrodes.

Acknowledgements

This work was financially supported by Grants-in-Aid for Specially Promoted Research (No. 23000009) and for Young Scientists (A) (No. 15H05494) and Young Scientists (B) (No. 5K17895) of Japan Society for the Promotion of Science (JSPS). This work was partly supported by the Artificial Photosynthesis Project of the New Energy and Industrial Technology Development Organization (NEDO) and Companhia Brasileira de Metalurgia e Mineração (CBMM). Part of this work was conducted at the Research Hub for Advanced Nano Characterization at the University of Tokyo, under the support of the “Nanotechnology Platform” (project No. 12024046) of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. The authors thank Dr Qian Wang, Dr Miao Zhong, Dr Taro Yamada and Mr Yutaka Sasaki of the University of Tokyo for SEM characterisation and discussion.

References

1. A. J. Nozik, Annu. Rev. Phys. Chem., 1978, 29, 189–222.
2. M. Grätzel, Nature, 2001, 414, 338–344.
3. A. Fujishima and K. Honda, Nature, 1972, 238, 37–38.
4. W. J. Chun, A. Ishikawa, H. Fujisawa, T. Takata, J. N. Kondo, M. Hara, M. Kawai, Y. Matsumoto and K. Domen, J. Phys. Chem. B, 2003, 107, 1798–1803.
5. T. W. Kim and K.-S. Choi, Science, 2014, 343, 990–994.
6. Y. Li, T. Takata, D. Cha, K. Takanabe, T. Minegishi, J. Kubota and K. Domen, Adv. Mater., 2013, 25, 125–131.
7. M. Zhong, Y. Ma, P. Oleynikov, K. Domen and J.-J. Delaunay, Energy Environ. Sci., 2014, 7, 1693–1699.
8. H. X. Dang, N. T. Hahn, H. S. Park, A. J. Bard and C. B. Mullins, J. Phys. Chem. C, 2012, 116, 19225–19232.
9. T. Minegishi, N. Nishimura, J. Kubota and K. Domen, Chem. Sci., 2013, 4, 1120–1124.
10. K. Ueda, T. Minegishi, J. Clune, M. Nakabayashi, T. Hisatomi, H. Nishiyama, M. Katayama, N. Shibata, J. Kubota, T. Yamada and K. Domen, J. Am. Chem. Soc., 2015, 137, 2227–2230.
11. Y. Li, L. Zhang, A. Torres-Pardo, J. M. Gonzalez-Calbet, Y. Ma, P. Oleynikov, O. Terasaki, S. Asahina, M. Shima, D. Cha, L. Zhao, K. Takanabe, J. Kubota and K. Domen, Nat. Commun., 2013, 4, 2566.
12. S. Hu, M. R. Shaner, J. A. Beardslee, M. Lichterman, B. S. Brunschwig and N. S. Lewis, Science, 2014, 344, 1005–1009.
13. L. Chen, J. Yang, S. Klaus, L. J. Lee, R. Woods-Robinson, J. Ma, Y. Lum, J. K. Cooper, F. M. Toma, L. W. Wang, I. D. Sharp, A. T. Bell and J. W. Ager, J. Am. Chem. Soc., 2015, 137, 9595–9603.
14. G. Liu, J. Shi, F. Zhang, Z. Chen, J. Han, C. Ding, S. Chen, Z. Wang, H. Han and C. Li, Angew. Chem., Int. Ed., 2014, 53, 7295–7299.
15. G. Liu, P. Fu, L. Zhou, P. Yan, C. Ding, J. Shi and C. Li, Chem.–Eur. J., 2015, 21, 9624–9628.
16. J. Seo, T. Takata, M. Nakabayashi, T. Hisatomi, N. Shibata, T. Minegishi and K. Domen, J. Am. Chem. Soc., 2015, 137, 12780–12783.
17. J. M. Morbec, I. Narkeviciute, T. F. Jaramillo and G. Galli, Phys. Rev. B, 2014, 90, 155204.
18. B. A. Pinaud, P. C. K. Vesborg and T. F. Jaramillo, J. Phys. Chem. C, 2012, 116, 15918–15924.
19. M. X. Li, W. J. Luo, D. P. Cao, X. Zhao, Z. S. Li, T. Yu and Z. G. Zou, Angew. Chem., Int. Ed., 2013, 52, 11016–11020.
20. B. A. Pinaud, A. Vailionis and T. F. Jaramillo, Chem. Mater., 2014, 26, 1576–1582.
21. S. Nakamura, Jpn. J. Appl. Phys., 1991, 30, 1705–1707.
22. F. A. Ponce and D. P. Bour, Nature, 1997, 386, 351–359.
23. A. Dabirian and R. van de Krol, Chem. Mater., 2015, 27, 708–715.
24. G. Liu, S. Ye, P. Yan, F. Xiong, P. Fu, Z. Wang, Z. Chen, J. Shi and C. Li, Energy Environ. Sci., 2016, 9, 1327–1334.
25. C. Wang, T. Hisatomi, T. Minegishi, M. Nakabayashi, N. Shibata, M. Katayama and K. Domen, Chem. Sci., 2016, 7, 5821–5826.
26. C. C. L. McCrory, S. Jung, J. C. Peters and T. F. Jaramillo, J. Am. Chem. Soc., 2013, 135, 16977–16987.
27. R. Fix, R. G. Gordon and D. M. Hoffman, Chem. Mater., 1993, 5, 614–619.
28. T. Nobuzo, Jpn. J. Appl. Phys., 1971, 10, 248–259.
29. G. Oya and Y. Onodera, J. Appl. Phys., 1974, 45, 1389–1397.
30. G. M. Carroll, D. K. Zhong and D. R. Gamelin, Energy Environ. Sci., 2015, 8, 577–584.
31. S. Chen and L.-W. Wang, Chem. Mater., 2012, 24, 3659–3666.
32. A. H. Reshak, Comput. Mater. Sci., 2014, 89, 45–51.
33. B. Klahr, S. Gimenez, F. Fabregat-Santiago, J. Bisquert and T. W. Hamann, Energy Environ. Sci., 2012, 5, 7626–7636.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ta04363a

---