NH3-assisted chloride flux-coating method for direct fabrication of visible-light-responsive SrNbO2N crystal layers

Kenta Kawashima ab, Mirabbos Hojamberdiev a, Oluwaniyi Mabayoje b, Bryan R. Wygant b, Kunio Yubuta c, C. Buddie Mullins b, Kazunari Domen d and Katsuya Teshima *ae
aDepartment of Environmental Science and Technology, Faculty of Engineering, Shinshu University, 4-17-1 Wakasato, Nagano 380-8553, Japan. E-mail: teshima@shinshu-u.ac.jp; Fax: +81 26 269 5550; Tel: +81 26 269 5541
bMcKetta Department of Chemical Engineering and Department of Chemistry, The University of Texas at Austin, Austin, Texas 78712, USA
cInstitute for Materials Research, Tohoku University, 2-1-1 Katahira, Aoba-ku, Sendai 980-8577, Japan
dDepartment of Chemical System Engineering, School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan
eCenter for Energy and Environmental Science, Shinshu University, 4-17-1 Wakasato, Nagano 380-8553, Japan

Received 31st March 2017 , Accepted 12th June 2017

First published on 12th June 2017


Perovskite-type SrNbO2N crystal layers were prepared on niobium substrates by using an NH3-assisted chloride flux-coating method. The optimization of synthesis parameters (holding temperature and strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio) was performed using a NaCl–KCl flux. By choosing the optimal synthesis conditions, platelet SrNbO2N crystals were grown over the entire substrate surface, and each SrNbO2N platelet has a single-crystalline structure in a cubic symmetry. The optimal crystal layer possesses a well-adhered SrNbO2N/NbNx/Nb structure and absorbed photons with wavelengths up to 680 nm. In addition, the optimal SrNbO2N/NbNx photoelectrode was used for photoelectrochemical water oxidation (2H2O → 4H+ + 4e + O2↑), as one half of the water splitting reaction. A photocurrent density of 113 μA cm−2 at 1.23 VRHE was recorded on the SrNbO2N/NbNx electrode without any additional co-catalyst loading and treatment under simulated sunlight, due to the higher crystallinity of SrNbO2N and higher interface-adhesion of the SrNbO2N/NbNx/Nb structure, which suppress the recombination of photogenerated electrons and holes at the defects and lead to an increase of the photogenerated electron collection efficiency in a niobium substrate, respectively. This study is the first to address the fabrication of quaternary oxynitride crystal layers on a conductive substrate using an NH3-assisted flux-coating method.


1. Introduction

Transition metal-based oxynitride perovskites (e.g. LaTiO2N, BaTaO2N, LaTaON2, SrNbO2N, etc.) have been studied for their dielectric, optical, and photocatalytic properties.1–8 Of note, because oxynitride materials have a narrow band-gap energy (<3 eV), they have attracted a lot of attention and have been extensively studied for their photocatalytic and photoelectrochemical (PEC) water splitting activity under solar irradiation to convert inexhaustible solar energy efficiently into chemical energy (hydrogen), which is a clean and renewable energy.5,6,8–11

Cubic perovskite-type oxynitride SrNbO2N (band gap of 1.8 eV) is a member of space group Pm[3 with combining macron]m with lattice constants a = b = c = 4.044 Å and α = β = γ = 90° and is composed of a corner-sharing cubic network of NbO6−xNx octahedral skeletal structures, with the Sr2+ ion positioned in the center of the cube, as shown in Fig. 1.12,13 This crystal structure was illustrated using VESTA 3 software.14 According to recent reports,2,15 it has been shown that tetragonal SrNbO2N, which belongs to space group I4/mcm, also exists. SrNbO2N is strongly expected to be a visible-light-driven photocatalyst for water splitting as it can absorb photons with long wavelengths (λ ≈ 700 nm) and because the band-edge potentials of SrNbO2N straddle the water oxidation and reduction potentials which allow photocatalytic and PEC water splitting without any bias for higher solar-energy conversion efficiency.16 In general, SrNbO2N oxynitride particles and layers are synthesized by thermal ammonolysis of SrNbO3, Sr5Nb4O15, and Sr2Nb2O7 oxide precursor particles and layers, respectively.16–21 It was, however, reported that the oxynitride, which was fabricated via thermal ammonolysis of the oxide precursor, contained a lot of chemical and crystal defects which may act as photogenerated carrier-recombination centers.22 Furthermore, SrNbO2N photoanodes are typically prepared by a particle-transfer and electrophoretic deposition method using SrNbO2N particles,13,16,21 likely resulting in poorly adhering interfaces between the particles and conductive substrate. For achieving high PEC performance, an effective approach involves decreasing the density of chemical and crystal defects and improving the interface-bonding quality.


image file: c7ce00614d-f1.tif
Fig. 1 Schematic illustration of the idealized crystal structure of cubic SrNbO2N.

In our research group, cube-like BaTaO2N oxynitride crystals with lower defect density and higher crystallinity were successfully grown from starting materials without forming an intermediate oxide compound by an NH3-assisted KCl flux method.23 Moreover, Suzuki et al.24 reported that binary nitride Ta3N5 crystal layers can be fabricated on tantalum substrates by using an NH3-assisted flux-coating method in which the tantalum substrate can also be used as a source of raw material and Ta3N5 crystals can be directly grown on the tantalum substrates. Thus, an NH3-assisted flux-coating method promises well-adhered interfaces between the crystal layer and conductive substrate. However, an NH3-assisted flux-coating method has not been applied in the fabrication of multi-element nitride and oxynitride crystal layers.

In the present work, we have focused on the direct fabrication of quaternary oxynitride SrNbO2N crystal layers on niobium substrates using an NH3-assisted chloride flux-coating method. To optimize preparation parameters, the effects of flux, holding temperature, and the strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio on the resultant morphologies and phases of the crystal layers were investigated. Here, we also demonstrate the PEC water oxidation activity and photostability of the optimal SrNbO2N photoanode under simulated solar irradiation.

2. Experimental

2.1. Preparation of SrNbO2N crystal layers on niobium substrates

Perovskite-type oxynitride SrNbO2N crystal layers were directly fabricated by using an NH3-assisted chloride flux-coating method. Aqueous solutions (1 M) of Sr(NO3)2 (≥98%), NaCl (≥99.5%), KCl (≥99.5%), and SrCl2 (≥95%), purchased from Wako Pure Chemical Industries, Ltd., were first prepared at room temperature. Niobium substrates (12 × 17 × 0.2 mm) were press-formed into a tray-like shape in order to keep the melt flux on the niobium substrate and hydrophilized under vacuum-ultraviolet light irradiation (Xe excimer lamp, UER172-200, Ushio Inc., λ = 172 nm). The adjusted fabrication parameters are listed in Table 1. Each aqueous solution (total volume = 36 μL) was drop-coated onto the niobium substrates and dried at 100 °C for 30 min. The substrates were put in an alumina boat with quartz wool, placed in a horizontal alumina tubular furnace (tube diameter = 24 mm), heated at 800–900 °C for 1 h with a heating rate of 600 °C h−1 under an NH3 flow (100 mL min−1) and then cooled naturally to room temperature. Afterwards, the substrates were immersed in hot water to remove the remaining flux and dry-cleaned. Finally, brown-colored crystal layers were obtained on the niobium substrates.
Table 1 Adjusted parameters for the fabrication of SrNbO2N crystal layers
Chloride flux Strontium source
Run no. Flux (molar ratio) 1 M SrCl2 (aq.)/μL 1 M KCl (aq.)/μL 1 M NaCl (aq.)/μL 1 M Sr(NO3)2 (aq.)/μL Holding temp./°C
1 SrCl2 26 10 800
2 KCl 26 10 800
3 NaCl 26 10 800
4 SrCl2–KCl (1[thin space (1/6-em)]:[thin space (1/6-em)]1) 13 13 10 800
5 SrCl2–NaCl (1[thin space (1/6-em)]:[thin space (1/6-em)]1) 13 13 10 800
6 NaCl–KCl (1[thin space (1/6-em)]:[thin space (1/6-em)]1) 13 13 10 800
7 NaCl–KCl (1[thin space (1/6-em)]:[thin space (1/6-em)]1) 13 13 10 850
8 NaCl–KCl (1[thin space (1/6-em)]:[thin space (1/6-em)]1) 13 13 10 900
9 NaCl–KCl (1[thin space (1/6-em)]:[thin space (1/6-em)]1) 15.5 15.5 5 850
10 NaCl–KCl (1[thin space (1/6-em)]:[thin space (1/6-em)]1) 10.5 10.5 15 850


2.2. Characterization

The X-ray diffraction (XRD) patterns of the fabricated crystal layers were collected using an X-ray diffractometer (MiniFlex300/600, Rigaku). Morphological observation and chemical analysis of the crystal layers were carried out using a field-emission scanning electron microscope (FE-SEM, JSM-7600F, JEOL) with an energy dispersive X-ray spectrometer (EDS) and an environmental scanning electron microscope (ESEM, Quanta 650, FEI). The cross-section of an optimal crystal layer, fabricated using a NaCl–KCl flux with a strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio of 10[thin space (1/6-em)]:[thin space (1/6-em)]13 at 850 °C for 1 h, was prepared using a focused ion beam (FIB) system (JIB-4000, JEOL and Strata DB235, FEI) for SEM and EDS investigations. Transmission electron microscopy (TEM) analyses of the pieces of an optimal SrNbO2N crystal, scraped off the niobium substrate with a spatula, were performed with a high-resolution electron microscope (EM-002B, TOPCON) operated at 200 kV. X-ray photoelectron spectroscopy (XPS) core-level spectra of the optimal crystal layer were recorded using a spectrometer (JPS-9010MC, JEOL and Axis-Ultra DLD, Kratos) with a non-monochromatic Mg Kα X-ray source. The position of the C 1s peak at 284.5 eV was used to correct the recorded binding energies. The ultraviolet-visible (UV-vis) diffuse reflectance spectrum of the optimal crystal layer was measured in reference to a niobium substrate by using a spectrophotometer (V-670, JASCO).

2.3. PEC measurements

The PEC measurements of SrNbO2N/NbNx/Nb (an optimal crystal layer) were conducted using a three-electrode electrochemical cell system under simulated solar light irradiation (150 W Xe arc lamp with an AM 1.5G filter) and the light intensity was calibrated to 100 mW cm−2 by means of an optical power meter (1916-C, Newport). An optimized SrNbO2N/NbNx/Nb sample, Ag/AgCl (saturated KCl), and Pt wire were used as the working electrode, reference electrode, and counter electrode, respectively. All PEC measurements were performed in a 1 M aqueous KOH solution (pH = 13.7) with a computer-based electrochemical workstation (660D, CH Instruments). The measured potential vs. the reversible hydrogen electrode (RHE) was calculated using the Nernst equation as ERHE = EAg/AgCl + 0.0591 × (pH) + image file: c7ce00614d-t1.tif[image file: c7ce00614d-t2.tif = 0.197 V at 25 °C].16,21,22

3. Results and discussion

3.1. Effect of flux on the fabrication of SrNbO2N crystal layers

The XRD patterns of the crystal layers fabricated separately using different chloride fluxes (SrCl2, KCl, NaCl, SrCl2–KCl, SrCl2–NaCl, and NaCl–KCl) with strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratios of 10[thin space (1/6-em)]:[thin space (1/6-em)]13 (for SrCl2, KCl, and NaCl) and 5[thin space (1/6-em)]:[thin space (1/6-em)]13 (for SrCl2–KCl, SrCl2–NaCl, and NaCl–KCl) at 800 °C for 1 h are shown in Fig. 2. The diffraction peaks of all the resultant crystal layers can be assigned to cubic SrNbO2N (ICDD PDF# 78-1458), cubic SrO (ICDD PDF# 72-2739), hexagonal NbN0.95 (ICDD PDF# 25-1361), hexagonal Nb2N (ICDD PDF# 39-1398), cubic Nb (ICDD PDF# 35-0789), and other unidentified phases. The SrO, which could not react with the niobium substrate, appeared due to the thermal decomposition of Sr(NO3)2 (∼652 °C):25
Sr(NO3)2 → SrO + NO2 + 1/2O2

image file: c7ce00614d-f2.tif
Fig. 2 XRD patterns of the crystal layers fabricated using (a) SrCl2, (b) KCl, (c) NaCl, (d) SrCl2–KCl, (e) SrCl2–NaCl, and (f) NaCl–KCl fluxes at 800 °C for 1 h.

Moreover, the niobium nitride NbN0.95 and Nb2N phases were formed by thermal ammonolysis of the niobium substrates, in which their formation sequence might be described as Nb → Nb2N → NbN0.95.26,27 The total reaction to form SrNbO2N can be generalized as the following equation: Sr(NO3)2 + Nb + NH3 → SrO + 2NO2 + 1/2O2 + Nb + N + 3/2H2 → SrNbO2N + 2O2 + N2 + 3/2H2. Here, since the NO2 formed by the decomposition of Sr(NO3)2 is a strong oxidizer due to its unpaired electron,28 the niobium metal (Nb0) might be oxidized to Nb5+ by the NO2 gas during the high-temperature reaction. Simultaneously, NH3 was thermally decomposed into hydrogen gas and atomic nitrogen,29 which acts as the nitrogen source to form SrNbO2N, Nb2N, and NbN0.95.

Fig. S1 shows digital photographs of the crystal layers fabricated using several different chloride fluxes with strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratios of 10[thin space (1/6-em)]:[thin space (1/6-em)]13 (for SrCl2, KCl, and NaCl) and 5[thin space (1/6-em)]:[thin space (1/6-em)]13 (for SrCl2–KCl, SrCl2–NaCl, and NaCl–KCl) at 800 °C for 1 h. As shown, the color of all the substrates varied from silvery-white to brown after thermal ammonolysis because the SrNbO2N crystals were grown on the substrate surface.4 Additionally, the as-obtained substrates have no irregularity in color except for the substrate using the SrCl2 flux. SrCl2 has a higher melting point (876 °C) than KCl (771 °C), NaCl (801 °C), SrCl2–KCl (619 °C), SrCl2–NaCl (566 °C), and NaCl–KCl (657 °C).30–32 As SrCl2 could not melt completely at 800 °C, the SrCl2 melt might not be attached to the substrate uniformly and a uniform crystal layer could not be formed on the substrate, resulting in the uneven color of the as-obtained substrate. To study the effects of flux on the resultant crystal morphology, the top-view SEM images of the crystal layers are shown in Fig. 3. Using the SrCl2 flux, polyhedral (50–440 nm diameter) and needle-shaped (ca. 50 × 470 nm dimensions) crystals were grown on a niobium substrate. Meanwhile, the other crystal layers were composed of platelet crystals with non-uniform sizes and small needle-shaped crystals with a size of 40–160 nm in diameter by 120–800 nm in length. Presumably, the difference in both shape and size of the crystals resulted from the difference in solute solubility of that specific chloride flux and the chemical interaction between the chloride flux and the solute (niobium substrates and SrNO3).33


image file: c7ce00614d-f3.tif
Fig. 3 Top-view SEM images of the crystal layers fabricated using (a) SrCl2, (b) KCl, (c) NaCl, (d) SrCl2–KCl, (e) SrCl2–NaCl, and (f) NaCl–KCl fluxes at 800 °C for 1 h.

3.2. Effect of holding temperature on the fabrication of SrNbO2N crystal layers

To investigate the effect of holding temperature on the morphology and phases of resultant crystal layers, the crystal layers were fabricated using the NaCl–KCl flux with a strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13 at different holding temperatures (800, 850, and 900 °C) for 1 h, separately. Fig. 4 shows the XRD results. The peaks in the XRD patterns of all the resultant crystal layers are identified as SrNbO2N, SrO, NbN0.95, Nb2N, Nb, and other unidentified phases. In accordance with increasing the holding temperature from 800 to 850 °C, the main 110 diffraction-peak intensity of SrNbO2N at 31.25° increased, and simultaneously, the 220 and 222 diffraction-peak intensities of SrO at 49.95 and 62.28° decreased. These happened because the formation of SrNbO2N was promoted owing to an increase in the holding temperature. On the other hand, when the holding temperature was increased from 850 to 900 °C, the 102 diffraction-peak intensities of Nb2N and NbN0.95 at 50.42 and 48.29°, respectively, increased, and the main 110 diffraction-peak intensity of SrNbO2N at 31.25° decreased. By virtue of the thermal ammonolyzability enhanced in accordance with the rise of the holding temperature, the thermal ammonolysis of niobium to form niobium nitride was likely preferential to SrNbO2N formation.
image file: c7ce00614d-f4.tif
Fig. 4 XRD patterns of the crystal layers fabricated using NaCl–KCl flux at (a) 800, (b) 850, and (c) 900 °C for 1 h.

The digital photographs of the crystal layers fabricated using the NaCl–KCl flux with a strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13 at different holding temperatures for 1 h are shown in Fig. S1. All the as-obtained substrates showed an even brown color. Fig. 5 shows the top-view SEM images of the crystal layers. As indicated, in all cases, the crystals grown on the niobium substrates have two different shapes, platelet (non-uniform sizes) and needle (20–150 nm in diameter; 320–1200 nm in length). Interestingly, when the holding temperature was elevated to 900 °C, the number of needle-shaped crystals significantly increased. Here, the temperature-dependent difference in solute solubility of the NaCl–KCl flux may be related to the difference in crystal size and shape. From the XRD results, 850 °C was found to be an optimum holding temperature for SrNbO2N crystal-layer fabrication using the NaCl–KCl flux.


image file: c7ce00614d-f5.tif
Fig. 5 Top-view SEM images of the crystal layers fabricated using NaCl–KCl flux at (a) 800, (b) 850, and (c) 900 °C for 1 h.

3.3. Effect of strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio on the fabrication of SrNbO2N crystal layers

Fig. S2 shows the XRD patterns of the crystal layers fabricated using the NaCl–KCl flux with different strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratios (5[thin space (1/6-em)]:[thin space (1/6-em)]31, 5[thin space (1/6-em)]:[thin space (1/6-em)]13, and 5[thin space (1/6-em)]:[thin space (1/6-em)]7) at 850 °C for 1 h. The XRD peaks attributable to SrNbO2N, SrO, NbN0.95, Nb2N, Nb, and other unidentified phases were observed for all the resultant crystal layers. The crystal layer fabricated with a molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]31 was found to exhibit a lower intensity for the main 110 diffraction peak of SrNbO2N at 31.25° than the crystal layer fabricated with the molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13. This is likely due to a reduced amount of Sr(NO3)2 at the lower molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]31 limiting the total amount of zSrNbO2N formed. By varying the molar ratio from 5[thin space (1/6-em)]:[thin space (1/6-em)]13 to 5[thin space (1/6-em)]:[thin space (1/6-em)]7, the diffraction-peak intensities of the target material (SrNbO2N) and by-products (SrO, NbN0.95, Nb2N) significantly decreased and increased, respectively. These observations are indicative of the fact that more by-product phases were produced relative to the desired SrNbO2N phase at the molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]7. In an NH3-assisted flux-coating method, the flux has two important roles: (i) suppression of excess ammonolysis of the solute (SrNO3 and metallic niobium from the substrate) and (ii) assistance of atomic diffusion during the reaction to form the target material (SrNbO2N). Hence, the differences in the diffraction peak intensities in Fig. S2 may originate from the smaller amount of the NaCl–KCl flux at the molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]7, which might induce excess ammonolysis of niobium and suppress the overall SrNbO2N formation.

Fig. S1 shows digital photographs of the crystal layers fabricated using the NaCl–KCl flux with different strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratios at 850 °C for 1 h. The as-obtained substrate (5[thin space (1/6-em)]:[thin space (1/6-em)]13) showed an even brown color. However, the other as-obtained substrates (5[thin space (1/6-em)]:[thin space (1/6-em)]31 and 5[thin space (1/6-em)]:[thin space (1/6-em)]7) exhibited uniform brown and grey colors because the small amounts of the Sr(NO3)2 (5[thin space (1/6-em)]:[thin space (1/6-em)]31) and NaCl–KCl flux (5[thin space (1/6-em)]:[thin space (1/6-em)]7) could not cover the entire surface of the niobium substrates, respectively. The top-view SEM images of the crystal layers are shown in Fig. S3. When 5[thin space (1/6-em)]:[thin space (1/6-em)]31 was used as the molar ratio, relatively small cuboid and platelet crystals were grown on the substrate. At the molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13, almost all the crystals have a large platelet shape. When the molar ratio was set at 5[thin space (1/6-em)]:[thin space (1/6-em)]7, the large platelet crystals were observed again and numerous small needles were also present on these crystal surfaces. Note that the difference in the size and shape of the crystals might be closely related to the molar ratio-dependent difference in solute soluble amount of the NaCl–KCl flux. Noticeably, it was found from the XRD results that 5[thin space (1/6-em)]:[thin space (1/6-em)]13 is an optimal molar ratio for SrNbO2N crystal-layer fabrication using the NaCl–KCl flux.

3.4. Cross-sectional EDS analyses

Cross-sectional EDS analyses were conducted to determine the structure and chemical composition of the optimal crystal layer fabricated using the NaCl–KCl flux with a strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13 at 850 °C for 1 h. FIB milling was used to produce the cross-sectional sample with a deposited carbon layer. The carbon layer served to protect the selected region of the crystal-layer surface from ion damage. The EDS elemental mapping images are shown in Fig. 6. Even distributions of strontium (cyan), niobium (purple), oxygen (yellow) and nitrogen (green) elements through the top-most layer (under the carbon layer) of approximately 570 nm thickness are observed, indicating that the first layer corresponds to SrNbO2N. Surprisingly, the second layer of approximately 250 nm thickness showed high-intensity signals of both niobium and nitrogen which can be attributed to the presence of niobium nitride (NbNx). In view of the XRD results, the thin NbNx layer is composed of NbN0.95 and Nb2N. In addition, the small amount of nitrogen species is uniformly distributed over the niobium substrate, implying that nitrogen has a large diffusivity during thermal ammonolysis. All of the mapping image results are consistent with the EDS spectrum shown in Fig. 6. It is noted that gallium peaks were detected in the spectrum, which is ascribed to incorporation of gallium in the sample during FIB milling. The cross-sectional EDS results confirmed that the optimal sample possesses a well-adhered SrNbO2N/NbNx/Nb structure.
image file: c7ce00614d-f6.tif
Fig. 6 Cross-sectional EDS mapping images and spectrum of the crystal layer fabricated using NaCl–KCl flux with a strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13 at 850 °C for 1 h.

3.5. TEM studies

To investigate the morphological and crystallographic details, TEM investigations were carried out on the platelet SrNbO2N crystals grown on niobium substrates via the NaCl–KCl flux with a strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13 at 850 °C for 1 h. Prior to performing the TEM analysis, the platelet SrNbO2N crystals were scraped off the substrate, dispersed in ethanol, dropped onto a grid, and dried at room temperature. The TEM image of a typical SrNbO2N platelet viewed along the [001] zone axis is displayed in Fig. 7. The SrNbO2N platelet has lateral dimensions of 4.1 × 2.9 μm and its surface is roughened. Fig. 7 shows the selected-area electron diffraction (SAED) patterns obtained from the red-circled area. The SAED patterns revealed that the SrNbO2N platelets are single-crystal structures. In the SAED patterns obtained with the incident beam along the [001] zone axis, the bright diffraction spots are consistent with the (100) and (0−10) planes of cubic SrNbO2N, suggesting the presence of the (001) crystal plane as the in-plane surface and (100) and (0−10) crystal planes as the side surfaces. We surmise that the SrNbO2N platelets were grown along the <001> direction. In order to double-check the crystal system for the SrNbO2N platelets, the crystal was tilted along the [102] zone axis and the SAED pattern was re-obtained. The diffraction spots can be indexed as the (0−10) and (20−1) planes of cubic SrNbO2N. Accordingly, the crystal system of the as-obtained SrNbO2N crystals is cubic.
image file: c7ce00614d-f7.tif
Fig. 7 TEM image and SAED patterns of the plate-like SrNbO2N crystals fabricated using NaCl–KCl flux with a strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13 at 850 °C for 1 h.

3.6. XPS measurements

To study the surface chemical composition and states of the optimal crystal layer fabricated using the NaCl–KCl flux with a strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13 at 850 °C for 1 h, XPS was performed. Fig. 8 shows the Sr 3d, Nb 3d, O 1s, and N 1s XPS core-level spectra. The Sr 3d spectrum was deconvoluted into two doublets assignable to the Sr–N and Sr–O linkages.34 The Sr 3d5/2–3/2 doublets are at 132.3 and 134.1 eV for the Sr–N linkage and at 134.3 and 136.1 eV for the Sr–O linkage, respectively. The curve fitting for the Nb 3d spectrum required two individual Nb 3d5/2–3/2 doublets corresponding to Nb4+ (205.1 and 207.8 eV) and Nb5+ (206.6 and 209.3 eV) oxidation states,35,36 indicating that SrNb5+O2N, SrNb4+O3, and anion-defective oxynitride (SrNb(O,N)3−δ) are probably present together on the surface of the crystal layer. In the O 1s spectrum, the three components centered at around 529, 530.9, and 533 eV can be recognized as lattice oxygen species, less electron-rich oxygen species, and molecular water or carbonates adsorbed on the surface, respectively.37 The N 1s spectrum appears to consist of two peaks at 395 and 397.3 eV which are related to the lattice nitrogen species and niobium nitride, respectively.36,38
image file: c7ce00614d-f8.tif
Fig. 8 (a) Sr 3d, (b) Nb 3d, (c) O 1s, and (d) N 1s XPS core-level spectra of the crystal layer fabricated using NaCl–KCl flux with a strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13 at 850 °C for 1 h. The XPS results confirm that niobium is present in two oxidation states (Nb5+ and Nb4+).

3.7. UV-vis diffuse reflectance spectroscopy

Fig. 9 shows the UV-vis diffuse reflectance spectrum of the optimal crystal layer fabricated using the NaCl–KCl flux with a strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13 at 850 °C for 1 h. The absorption edge of the crystal layer is at approximately 680 nm. Here, the strong absorption background above the absorption onset is indicative of the presence of reduced niobium species (such as Nb+, Nb2+, Nb3+, and Nb4+) generated during the thermal ammonolysis process.16 According to the Tauc plot of the optical absorption coefficient versus photon energy (see inset of Fig. 9),39 the band-gap energy is estimated to be 1.82 eV.
image file: c7ce00614d-f9.tif
Fig. 9 UV-vis diffuse reflectance spectrum and Tauc plot (insert): optical absorption coefficient (αhν)2vs. photon energy () of the crystal layer fabricated using NaCl–KCl flux with a strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13 at 850 °C for 1 h. The absorption edge is observed at approximately 680 nm (Eg = 1.82 eV) for the SrNbO2N crystal layer.

3.8. PEC performance

The PEC performance of the optimal crystal layer fabricated using the NaCl–KCl flux with a strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13 at 850 °C for 1 h was investigated in a three-electrode electrochemical cell using a 1 M KOH electrolyte at pH 13.7 under simulated solar light irradiation. Fig. 10 shows the current–potential curves of the resultant SrNbO2N/NbNx photoelectrode which exhibited an anodic photoresponse and the onset potential was about 0.87 VRHE. Compared to the reported data for the SrNbO2N photoelectrode without any additional co-catalyst loading and treatment,13 a relatively larger photocurrent density of 113 μA cm−2 was recorded over the SrNbO2N/NbNx photoelectrode at 1.23 VRHE in this work owing probably to both the higher crystallinity of this material and the higher interface-adhesion of the SrNbO2N/NbNx/Nb structure accomplished by an NH3-assisted flux-coating method. Further, since compounds containing low-valent transition metal cations (such as Nb2N, NbN, and WO2) generally have metallic characteristics because of their partially filled d-orbitals, they have greater potential for acting as co-catalysts to accept the photogenerated electrons, and the resultant NbNx layer is expected to be an electron acceptor, promoting effective charge separation.40–42 A low dark current was observed in the range from 0.5 to 1.3 VRHE, corresponding to the partial oxidation of niobium nitride (see inset of Fig. 10).22 Here, the electrolyte penetrated through small crevices between the SrNbO2N platelets to oxidize the NbNx layer. As can be seen from Fig. S4, a drastically degraded photocurrent was observed on the SrNbO2N/NbNx photoelectrode for the second PEC test. After the second test, chopped current–potential curves were measured for 31 cycles on the SrNbO2N/NbNx photoelectrode to confirm its photostability. Although the photocurrent gradually decreased each time, the SrNbO2N/NbNx photoelectrode still exhibited a low photocurrent density of 3.8 μA cm−2 at 1.23 VRHE after 30 cycles. To understand the mechanism degrading the photocurrent of the SrNbO2N/NbNx photoelectrode, several analyses were conducted on the SrNbO2N/NbNx photoelectrode after testing.
image file: c7ce00614d-f10.tif
Fig. 10 Current–potential curves of the SrNbO2N/Nb2N/Nb photoanode in darkness (blue broken lines) and under AM 1.5G simulated sunlight irradiation (red solid lines). The crystal layer was fabricated using the NaCl–KCl flux with a strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio of 5[thin space (1/6-em)]:[thin space (1/6-em)]13 at 850 °C for 1 h. A photocurrent density of 113 μA cm−2 was recorded at 1.23 VRHE for the SrNbO2N/Nb2N/Nb photoanode.

According to the XRD results (see Fig. S6), the diffraction-peak intensities for SrNbO2N decreased and diffraction peaks corresponding to SrO disappeared after testing. The SrO might dissolve in the electrolyte during the PEC tests. As shown in Fig. S7 and S8, the basic morphology of the SrNbO2N/NbNx/Nb structure was maintained after testing, whereas the crystal surfaces of the SrNbO2N platelets were slightly damaged and became rougher. Post-testing XPS (Fig. S9) shows that only the Sr 3d5/2–3/2 doublet (133 and 134.7 eV) corresponding to the strontium niobium oxide was found after testing.43 The Nb4+ content obviously decreased and the Nb5+ content significantly increased after testing, indicating that the Nb4+ species were oxidized to the Nb5+ species due to the photogenerated electrons during the PEC tests. In the O 1s spectrum of the SrNbO2N/NbNx photoelectrode after testing, the deconvoluted peaks are placed into three categories: (i) the lattice oxygen species in Sr2Nb2O7 (at around 529.8 eV), (ii) hydroxyl groups or surface-adsorbed oxygen (at around 531.2 eV) and (iii) molecular water or carbonates adsorbed on the surface (at around 533 eV).37,43 Additionally, there was no peak in the N 1s region after testing. The above analysis results for the SrNbO2N/NbNx photoelectrode after testing revealed that the photo-oxidative self-corrosion of SrNbO2N occurred during the PEC tests. From the XPS results, the SrNbO2N platelets before testing possess considerable defects such as anion-defective SrNbO2N and reduced Nb4+ species on their surfaces which might provide the photogenerated carrier trap-states and become the trigger to cause their self-corrosion. During the PEC tests, the nitrogen species in the SrNbO2N lattice were released due to the photo-oxidative self-corrosion of SrNbO2N, as shown in the following chemical reaction:44

2N3− + 6h+ → N2

In conjunction with this reaction, the SrNbO2N crystal surface might be transformed to amorphous strontium niobium oxide and hydroxide which have high resistivity.45 Hence, the main reason for the degrading photocurrent of the SrNbO2N/NbNx photoelectrode is likely the formation of amorphous strontium niobium oxide and hydroxide on the SrNbO2N crystal surface. Therefore, it is anticipated that removal of the surface defects can suppress the photo-oxidative self-corrosion of SrNbO2N and then improve the PEC performance and photostability of the SrNbO2N/NbNx photoelectrode more. Further studies are under way directed by this hypothesis.

4. Conclusions

In summary, quaternary oxynitride SrNbO2N crystal layers were fabricated on niobium substrates using SrCl2, KCl, NaCl, SrCl2–KCl, SrCl2–NaCl, and NaCl–KCl fluxes under an NH3 flow. In particular, the optimal synthesis parameters such as holding temperature and strontium source[thin space (1/6-em)]:[thin space (1/6-em)]flux molar ratio were investigated for the synthesis conditions using the NaCl–KCl flux. The optimal crystal layer consisted of platelet SrNbO2N crystals which uniformly covered the substrate. From the cross-sectional SEM and EDX analysis results, a well-adhered SrNbO2N/NbNx/Nb structure was observed for the optimal crystal layer on a niobium substrate. According to the TEM results, the crystal symmetry of the as-obtained SrNbO2N platelets was cubic and each platelet had single-crystalline nature. The absorption edge of the optimal crystal layer was approximately 680 nm and the band gap energy was estimated to be approximately 1.82 eV. Regarding the PEC performance, a photocurrent density of 113 μA cm−2 at 1.23 VRHE was exhibited on the SrNbO2N/NbNx photoelectrode without any additional co-catalyst loading and treatment under simulated sunlight. The XPS results indicate considerable lattice defects, which cause trapping and recombination of photogenerated carriers, exist on the SrNbO2N crystal surface, and may limit the PEC performance and trigger the photo-oxidative self-corrosion of SrNbO2N. Our results suggest that eliminating surface defects further improves the PEC performance and photostability of the SrNbO2N/NbNx photoelectrode. Further studies to test this hypothesis are under way.

Acknowledgements

The authors would like to thank Dr. Hugo Celio and Ms. Reiko Shiozawa for their kind assistance in XPS measurements. FIB assistance from Dr. Raluca Gearba is gratefully acknowledged. K. K., O. M., B. R. W., and C. B. M. gratefully acknowledge the U. S. Department of Energy Basic Energy Sciences Grant DE-FG02-09ER16119 and Welch Foundation Grant F-1436. This research was supported in part by the Japan Technological Research Association of Artificial Photosynthetic Chemical Process (ARPChem).

References

  1. A. Fuertes, J. Mater. Chem., 2012, 22, 3293–3299 RSC .
  2. Y. I. Kim, P. M. Woodward, K. Z. Baba-Kishi and C. W. Tai, Chem. Mater., 2004, 16, 1267–1276 CrossRef CAS .
  3. M. Jansen and H. P. Letschert, Nature, 2000, 404, 980–982 CrossRef CAS PubMed .
  4. R. Aguiar, D. Logvinovich, A. Weidenkaff, A. Rachel, A. Reller and S. G. Ebbinghaus, Dyes Pigm., 2008, 76, 70–75 CrossRef .
  5. S. Pokrant, A. E. Maegli, G. L. Chiarello and A. Weidenkaff, Chimia, 2013, 67, 162–167 CrossRef CAS PubMed .
  6. T. Takata, C. Pan and K. Domen, ChemElectroChem, 2016, 3, 31–37 CrossRef CAS .
  7. F. Oehler and S. G. Ebbinghaus, Solid State Sci., 2016, 54, 43–48 CrossRef CAS .
  8. C. Pan, T. Takata, K. Kumamoto, S. S. K. Ma, K. Ueda, T. Minegishi, M. Nakabayashi, T. Matsumoto, N. Shibata, Y. Ikuhara and K. Domen, J. Mater. Chem. A, 2016, 4, 4544–4552 CAS .
  9. R. Abe, J. Photochem. Photobiol., C, 2010, 11, 179–209 CrossRef CAS .
  10. T. Minegishi, N. Nishimura, J. Kubota and K. Domen, Chem. Sci., 2013, 4, 1120–1124 RSC .
  11. K. Ueda, T. Minegishi, J. Clune, M. Nakabayashi, T. Hisatomi, H. Nishiyama, M. Katayama, N. Shibata, J. Kubota and K. Domen, J. Am. Chem. Soc., 2015, 137, 2227–2230 CrossRef CAS PubMed .
  12. R. Marchand, F. Pors and Y. Laurent, Rev. Int. Hautes Temp. Refract., 1986, 23, 11–15 CAS .
  13. K. Maeda, M. Higashi, B. Siritanaratkul, R. Abe and K. Domen, J. Am. Chem. Soc., 2011, 133, 12334–12337 CrossRef CAS PubMed .
  14. K. Momma and F. Izumi, J. Appl. Crystallogr., 2011, 44, 1272–1276 CrossRef CAS .
  15. S. G. Ebbinghaus, A. Weidenkaff, A. Rachel and A. Reller, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 2004, 60, i91–i93 Search PubMed .
  16. J. Seo, Y. Moriya, M. Kodera, T. Hisatomi, T. Minegishi, M. Katayama and K. Domen, Chem. Mater., 2016, 28, 6869–6876 CrossRef CAS .
  17. A. Rachel, S. G. Ebbinghaus, M. Güngerich, P. J. Klar, J. Hanss, A. Weidenkaff and A. Reller, Thermochim. Acta, 2005, 438, 134–143 CrossRef CAS .
  18. R. Aguiar, D. Logvinovich, A. Weidenkaff, A. Reller and S. G. Ebbinghaus, Thermochim. Acta, 2008, 471, 55–60 CrossRef CAS .
  19. S. G. Ebbinghaus, R. Aguiar, A. Weidenkaff, S. Gsell and A. Reller, Solid State Sci., 2008, 10, 709–716 CrossRef CAS .
  20. S. H. Porter, S. Hwang, V. Amarasinghe, E. Taghaddos, V. Manichev, M. Li, G. Gardner, A. Safari, E. Garfunkel, M. Greenblatt and G. C. Dismukes, ECS Trans., 2016, 72, 1–19 CrossRef .
  21. M. Kodera, H. Urabe, M. Katayama, T. Hisatomi, T. Minegishi and K. Domen, J. Mater. Chem. A, 2016, 4, 7658–7664 CAS .
  22. S. Pokrant, M. C. Cheynet, S. Irsen, A. E. Maegli and R. Erni, J. Phys. Chem. C, 2014, 118, 20940–20947 CAS .
  23. M. Hojamberdiev, K. Yubuta, J. J. M. Vequizo, A. Yamakata, S. Oishi, K. Domen and K. Teshima, Cryst. Growth Des., 2015, 15, 4663–4671 CAS .
  24. S. Suzuki, H. Wagata, M. Komatsus, T. Minegishi, K. Domen, S. Oishi and K. Teshima, J. Mater. Chem. A, 2015, 3, 13946–13952 CAS .
  25. S. Yuvaraj, L. Fan-Yuan, C. Tsong-Huei and Y. Chuin-Tih, J. Phys. Chem. B, 2003, 107, 1044–1047 CrossRef CAS .
  26. C. Angelkort, H. Lewalter, P. Warbichler, F. Hofer, W. Bock and B. O. Kolbesen, Spectrochim. Acta, Part A, 2011, 57, 2077–2089 CrossRef .
  27. O. Brunkahl, R. Mertens, W. Bock, K. Thoma and B. O. Kolbesen, Microchim. Acta, 2006, 156, 39–51 CrossRef CAS .
  28. L. Valentini, I. Armentano, J. M. Kenny, C. Cantalini, L. Lozzi and S. Santucci, Appl. Phys. Lett., 2003, 82, 961–963 CrossRef CAS .
  29. M. Mesrine, N. Grandjean and J. Massies, Appl. Phys. Lett., 1998, 72, 350–352 CrossRef CAS .
  30. J. Sangster and A. D. Pelton, J. Phys. Chem. Ref. Data, 1987, 16, 509–561 CrossRef CAS .
  31. E. Vortisch, Neues Jahrb. Mineral., Geol. Palaeontol., Beilageband, 1914, 38, 182–272 Search PubMed .
  32. H. Fink and H. J. Seifert, Z. Anorg. Allg. Chem., 1980, 466, 87–96 CrossRef CAS .
  33. M. Kodera, M. Katayama, T. Hisatomi, T. Minegishi and K. Domen, CrystEngComm, 2016, 18, 3186–3190 RSC .
  34. J. Zhu, L. Wang, T. Zhou, Y. Cho, T. Suehiro, T. Takeda, M. Lu, T. Sekiguchi, N. Hirosaki and R. J. Xie, J. Mater. Chem. C, 2015, 3, 3181–3188 RSC .
  35. J. L. Cao, L. T. Li, Y. L. Wang and Z. L. Gui, J. Electrochem. Soc., 2001, 148, F180–F183 CrossRef CAS .
  36. J. Wang, X. Wang, B. Liu, X. Li and M. Cao, Mater. Lett., 2015, 152, 131–134 CrossRef CAS .
  37. W. Zhou, M. Zhao, F. Liang, S. C. Smith and Z. Zhu, Mater. Horiz., 2015, 2, 495–501 RSC .
  38. M. Nagai, R. Nakauchi, Y. Ono and S. Omi, Catal. Today, 2000, 57, 297–304 CrossRef CAS .
  39. E. A. Davis and N. Mott, Philos. Mag., 1970, 22, 0903–0922 CrossRef CAS .
  40. D. S. Katzer, N. Nepal, D. J. Meyer, B. P. Downey, V. D. Wheeler, D. F. Storm and M. T. Hardy, Appl. Phys. Express, 2015, 8, 085501 CrossRef .
  41. R. Chicault and J. C. Villegier, Phys. Rev. B: Solid State, 1987, 36, 5215 CrossRef CAS .
  42. S. L. Wang, Y. L. Mak, S. Wang, J. Chai, F. Pan, M. L. Foo, W. Chen, K. Wu and G. Q. Xu, Langmuir, 2016, 32, 13046–13053 CrossRef CAS PubMed .
  43. Z. Gao, H. Ning, C. Chen, R. Wilson, B. Shi, H. Ye, H. Yan and M. J. Reece, J. Am. Ceram. Soc., 2013, 96, 1163–1170 CrossRef CAS .
  44. K. Kawashima, M. Hojamberdiev, S. Chen, K. Yubuta, H. Wagata, K. Domen and K. Teshima, J. Mol. Catal. A: Chem., 2017, 432, 250–258 CrossRef .
  45. H. Ning, H. Yan and M. J. Reece, J. Am. Ceram. Soc., 2010, 93, 1409–1413 CAS .

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ce00614d

This journal is © The Royal Society of Chemistry 2017