Introduction

Artificial photocatalysis for solar water splitting has been regarded as a source for clean and sustainable fuel and chemical production [1,2,3]. To achieve efficient solar energy conversion, it is indispensable that the photon energy in the visible light range should be utilized for the photocatalytic reaction [4]. It is known from previous studies that the efficient photocatalysts, including transition and typical metal oxides, oxynitrides, and nitrides, contain metal cations of Ti4+, Nb5+, Ta5+, or W6+ with d 0 electronic configuration and of Ga3+, In3+, Ge4+, Sn4+, or Sb5+ with d 10 electronic configuration [5, 6]. In recent years, transition metal oxynitride perovskites with narrower band gaps and suitable band structures (e.g., LaTiO2N [7], BaNbO2N [8], BaTaO2N [9], CaTaO2N [10], LaMg1/3Ta2/3O2N [11], and so on) are regarded as an emerging class of inorganic materials that can effectively utilize solar energy for photocatalytic reactions. However, the previously reported photocatalytic activities of transition metal oxynitride perovskites are still not sufficient to realize practically the developed solar energy systems. Therefore, it is necessary to discover novel materials having a high potential for photocatalytic water splitting and to develop new strategies to improve the photocatalytic activities of transition metal oxynitride perovskites.

In our recent work, we have explored five tungsten-based oxynitride perovskites AW(O,N)3, A = Sr, La, Pr, Nd, and Eu as novel materials for visible light-driven photocatalytic water oxidation [12]. Among the five samples, the CoO x -loaded SrW(O,N)3 exhibited higher photostability and photocatalytic O2 evolution rate of 3.3 μmol h−1 due to the presence of active sites for O2 evolution on the surfaces and absence of the intrinsic defects related to the partly filled 4f orbitals. Although the SrW(O,N)3 photocatalyst has the potential for photocatalytic O2 evolution reaction, it needs to be further improved by applying a novel strategy.

Regarding to the AB(O,N)3-type materials, it was reported that the substitution of the A-site or B-site cation can improve the photocatalytic activity of AB(O,N)3. For example, the (La,Ca)Ti(O,N)3 crystals, which were synthesized by thermal ammonolysis of A-site-deficient La1−x TiO3.5−3x/2 precursor (x = 0.05) treated in an aqueous solution containing Ca(NO3)2·4H2O, showed higher activity for photocatalytic water oxidation (O2 evolution rate: 40 μmol h−1) compared with the pure LaTiO2N crystals (O2 evolution rate: 22 μmol h−1) due to their higher crystallinity and lower concentration of intrinsic defects (Ti3+) that act as a recombination center for photo-excited electrons and holes [13]. The formation of Ti3+ was restrained by substituting La3+ with Sr2+ in (La,Sr)TiO2N because the lower-valent A-site cation may induce the oxidation of titanium [14]. On the other hand, introducing the tungsten species having 5d 1 electronic configuration into the B-site of BaTaO2N showed an exceedingly effective impact on enhancing the photocatalytic water oxidation activity [15]. Furthermore, the CoO x -loaded BaNb1−x Ta x O2N porous structures (x = 0.50) exhibited higher photocatalytic O2 evolution rate (127.24 µmol in the first 2 h) than the CoO x -loaded pure BaNbO2N porous structures (20.92 µmol in the first 2 h) owing to the high specific surface area and low density of structural defects [16]. As mentioned above, the photocatalytic activity of AB(O,N)3 can be significantly enhanced by the exquisite combination of the substitutional elements and parent materials, which still has many untouched possibilities to explore.

In the present study, we have synthesized five Ba- and/or Ta-substituted strontium tungsten oxynitride perovskites Sr1−x Ba x W1−y Ta y (O,N)3, where x,y = 0,0; 0.25,0; 0.125,0.125; 0,0.25; and 1.1, by thermal ammonolysis of their oxide precursor crystals. This study attempts to investigate the effect of partial Ba2+-to-Sr2+ and/or Ta5+-to- W4~6+ substitution on the O/N ratio, band gap, and photocatalytic water oxidation activity of SrW(O,N)3 photocatalyst.

Experiment

Growth of Sr1−x Ba x W1−y TayO4−(1/2)y and Ba5Ta4O15 crystals

The Sr1−x Ba x W1−y Ta y O4−(1/2)y , where x,y = 0,0; 0.25,0; 0,0.25; and 0.125,0.125, and Ba5Ta4O15 crystals were synthesized by a solid-state reaction and KCl flux method with 10 mol % solute concentration, respectively. Reagent-grade SrCO3 (≥ 95%, Wako), BaCO3 (99.9%, Wako), WO3 (≥ 95%, Wako), Ta2O5 (99.9%, Wako) and KCl (99.5%, Wako) were dry mixed manually in stoichiometric ratio using an agate mortar and a pestle. Continuously, each mixture (10 g) was placed in an alumina crucible and closed loosely with an alumina lid. The mixture-containing alumina crucible was heated at 1100 °C for 3 h at a heating rate of 100 °C h−1 for Sr1−x Ba x W1−y Ta y O4−(1/2)y or at 1000 °C for 10 h at a heating rate of 50 °C h−1 for Ba5Ta4O15, cooled to 500 °C at a cooling rate of 150 °C h−1, and then naturally cooled to room temperature. As for Ba5Ta4O15, the flux-grown crystals were washed with deionized water to remove the remaining flux and then dried at 100 °C for 12 h in a drying oven.

Synthesis of Sr1−x Ba x W1−y Ta y (O,N)3 and BaTa(O,N)3 crystal structures

To obtain the Sr1−x Ba x W1−y Ta y (O,N)3 and BaTa(O,N)3 oxynitride crystal structures, 1.0 g of the as-synthesized Sr1−x Ba x W1−y Ta y O4−(1/2)y crystals were placed on an alumina plate and ammonolyzed at 900 °C for 10 h for SrWO4 and Sr0.75Ba0.25WO4, 30 h for Sr0.875Ba0.125W0.875Ta0.125O3.9375, and 40 h for SrW0.75Ta0.25O3.975, while 1.0 g of the flux-grown Ba5Ta4O15 crystals were ammonolyzed at 950 °C for 40 h at a heating rate of 600 °C h−1 under an NH3 flow (200 mL min−1) in a vertical tubular furnace.

Characterization

X-ray powder diffraction (XRD) patterns were recorded on a Rigaku MiniflexII X-ray diffractometer (Cu Kα radiation (λ = 1.5406 Å)). The recorded XRD peak positions were calibrated using an unstressed Si powder standard with a 111 diffraction peak at 28.44°. The high-resolution scanning electron microscopy (SEM) observations were performed using a JSM-7600F scanning electron microscope (JEOL) attached with an energy-dispersive X-ray spectrometer (EDS). The ultraviolet–visible (UV–Vis) diffuse reflectance spectra were measured by a V-670 spectrophotometer (JASCO). The nitrogen and oxygen contents were determined by hot-gas extraction using a LECO TC-436 N/O analyzer (LECO Co.). The specific surface areas of the oxynitride samples were determined by the Brunauer, Emmett, and Teller (BET) method from the linear portion of the nitrogen gas adsorption isotherms measured at −196 °C using an Autosorb-1A (Quantachrome Instruments). The samples were preheated at 150 °C for 24 h under vacuum before the measurements. The surface chemical compositions and chemical states of elements of the synthesized oxynitrides were investigated using a JPS-9010MC X-ray photoelectron spectroscope (XPS, JEOL) with a non-monochromatic Mg Kα-excitation X-ray source. The C 1 s peak at 284.5 eV was used to calibrate the spectra.

Photocatalytic water oxidation activity test

The photocatalytic O2 gas evolution reactions were performed in a glass cell attached to a closed gas circulation system connected to a gas chromatograph (GC-8A, TCD, Ar gas carrier, Shimadzu) equipped with a vacuum line. First, 100 mg of CoO x -loaded oxynitride sample and 200 mg of La2O3 (as a pH buffer) were suspended in 300 mL of 10 mM AgNO3 (> 99%, Wako) aqueous solution (as a sacrificial electron scavenger) in the glass cell under magnetic stirring. A 300 W Xe lamp with a cutoff filter (λ > 420 nm) was used as a visible light source and its light intensity was 11 mW cm−2 at 360 nm. The evolved O2 and N2 gases were detected by a GC. The CoO x (2 wt % Co) nanoparticles as a cocatalyst for O2 gas evolution were loaded by immersing the oxynitride samples in Co(NO3)2·6H2O (> 99%, Wako) aqueous solution and heat treatment at 700 °C for 1 h under an NH3 flow (200 mL min−1) followed by heat treatment at 200 °C for 1 h in air.

The turnover numbers (TONs) for 5 h photocatalytic O2 gas evolution reactions were estimated as the total moles of substrates (H2O) consumed over the reaction per moles of active catalytic species (CoO x ) [17, 18]:

$${\text{TON}} = \frac{{{\text{number}}\;{\text{of}}\;{\text{moles}}\;{\text{of}}\;{\text{H}}_{2} {\text{O}}}}{{{\text{number}}\;{\text{of}}\;{\text{moles}}\;{\text{of}}\;{\text{CoO}}_{x} }}$$
(1)

It is known that the loaded cobalt oxide particles are mainly in the forms of CoO and Co3O4 [3]. For the calculation, assuming that the molar amount of CoO is the same as that of Co3O4 in the loaded cobalt oxide particles, CoO x was accordingly regarded as Co2O2.5. The turnover frequencies (TOFs), the number of TONs per unit time of reaction, were determined using the following equation [18]:

$${\text{TOF}} = \frac{\text{TON}}{{{\text{unit}}\;{\text{time}}\;{\text{of}}\;{\text{reaction}}}}$$
(2)

Results and discussion

Characterization of the synthesized oxide precursors and oxynitride perovskites

Figure 1 shows the XRD patterns of Sr1−x Ba x W1−y Ta y O4−(1/2)y and Ba5Ta4O15 precursor crystals synthesized by a solid-state reaction. As shown in Fig. 1a, e, the XRD patterns of the samples are identical to that of a single-phase SrWO4 (ICDD PDF# 08-0490) and Ba5Ta4O15 (ICDD PDF# 72-0631) with tetragonal and hexagonal structures, respectively. The XRD patterns of the Sr1−x Ba x W1−y Ta y O4−(1/2)y (x,y = 0.25,0; 0,0.25; and 0.125,0.125) crystals mostly correspond to the tetragonal SrWO4 phase, and the presence of minor Sr2Ta2O7 (ICDD PDF# 72-0921) and some unidentified impurity phases were also detected (Fig. 1b–d). The formation of the Sr2Ta2O7 phase evidences that the Ta5+-substitution levels of 12.5 and 25% have already exceeded the substitution limit. On increasing the Ba2+-substitution level from 0 to 25%, the main 112 diffraction peak of SrWO4 at 27.65° in the recorded XRD patterns slightly shifted toward a lower 2θ angle because the unit cell volumes of the Sr1−x Ba x W1−y Ta y O4−(1/2)y crystals increased as a result of the increase of the ratio of larger Ba2+ ions (r ionic = 1.42 Å: CN = 8) to smaller Sr2+ ions (r ionic = 1.26 Å: CN = 8), which is consistent with the Vegard’s law for solid solutions [19, 20]. In contrast, for the Ta5+-substitution levels of 0 and 25%, although the W6+ may be partially substituted with Ta5+, there is no significant difference between the positions of the main 112 diffraction peaks of the XRD patterns of the synthesized crystals because the ionic radius of W6+ is quite similar to that of Ta5+ [19]. Nevertheless, an entire amount of tantalum could not be incorporated in the SrWO4 owing to the partial separation of the Sr2Ta2O7 phase during the crystal growth process. The lattice constants and unit cell volumes of the Sr1−x Ba x W1−y Ta y O4−(1/2)y crystals estimated using the XRD results are listed in Table 1.

Fig. 1
figure 1

XRD patterns of Sr1−x Ba x W1−y Ta y O4−(1/2)y : (a) x,y = 0,0; (b) x,y = 0.25,0; (c) x,y = 0,0.25; (d) x,y = 0.125,0.125; and (e) Ba5Ta4O15 crystals grown by a solid-state reaction and KCl flux method

Full size image
Table 1 Unit cell parameters of tetragonal Sr1−x Ba x W1−y Ta y O4−(1/2)y (x,y = 0,0; 0.25,0; 0,0.25; and 0.125,0.125) crystals grown by a solid-state reaction
Full size table

Figure 2 shows the UV–Vis diffuse reflectance spectra of the as-synthesized SrWO4, Sr0.75Ba0.25WO4, SrW0.75Ta0.25O3.975, Sr0.875Ba0.125W0.875Ta0.125O3.9375, and Ba5Ta4O15 precursor crystals. The SrWO4, Sr0.75Ba0.25WO4, SrW0.75Ta0.25O3.975, Sr0.875Ba0.125W0.875Ta0.125O3.9375, and Ba5Ta4O15 crystals have clear absorption edges at about 265, 264, 287, 323, and 281 nm, respectively. The band-gap energies (E g), estimated by the equation: E g = hc/λ, where h is the Planck’s constant, c is the speed of light (m s−1), and λ is the wavelength (nm) [12], are about 4.68 eV for SrWO4, 4.7 eV for Sr0.75Ba0.25WO4, 4.32 eV for SrW0.75Ta0.25O3.975, 3.84 eV for Sr0.875Ba0.125W0.875Ta0.125O3.9375, and 4.41 eV for Ba5Ta4O15. The Sr0.75Ba0.25WO4 crystals show a higher band-gap energy compared with the SrWO4 crystals because the electrons belonging to the A-site of AWO4 contribute to the valence (O2− 2p states) and conduction bands (W6+ 5d states) [21]. Lacomba-Perales et al. [21] pointed out that the larger the ionic radius of the A 2+ cation is, the higher the band-gap energy of the AWO4 crystals is, which is in good agreement with our results and the previous experimental and theoretical reports on Ba1−x Sr x WO4 [22, 23]. In comparison with other samples, the SrW0.75Ta0.25O3.975 and Sr0.875Ba0.125W0.875Ta0.125O3.9375 crystals show lower band-gap energies, although it was previously reported that the band-gap energy of tantalum-doped hex-WO3 was higher than that of pure hex-WO3 due to the Ta5+ ions with d 0 electronic configuration [24]. The redshift of the band-gap energies can be explained by two possibilities: (i) the insertion of the tantalum species (Ta5+) into the Sr2+W6+O4 crystal lattice generated oxygen vacancies that reduced the band-gap energy of Sr1−x Ba x W1−y Ta y O4−(1/2)y [25] and (ii) the W-doped Sr2Ta2O7 phase with a narrow band gap was formed during the crystal growth process [26]. The narrow band gap of the W-doped Sr 2+2 Ta2 5+O7 was observed because the tungsten (W5+ with 5d 1 electronic configuration) doping increased its electronic density near the conduction band edge and formed the intermediary energy levels in the forbidden band. Interestingly, the Sr0.875Ba0.125W0.875Ta0.125O3.9375 crystals show the lowest band-gap energy among all samples, possibly owing to the formation of the Ba-doped Sr2Ta2O7 impurity phase that also possesses a lower band-gap energy [27].

Fig. 2
figure 2

UV–Vis diffuse reflectance spectra of Sr1−x Ba x W1−y Ta y O4−(1/2)y : blue x,y = 0,0; green x,y = 0.25,0; orange x,y = 0,0.25; purple x,y = 0.125,0.125; and red Ba5Ta4O15 crystals grown by a solid-state reaction and KCl flux method

Full size image

The as-synthesized precursor oxides were successfully transformed into the cubic perovskite-type oxynitrides. Figure 3 shows the XRD patterns of the Sr1−x Ba x W1−y Ta y O4−(1/2)y and Ba5Ta4O15 crystals ammonolyzed at 900 and 950 °C for different holding times (10–40 h) under an NH3 flow. The diffraction lines of the ammonolyzed Sr1−x Ba x W1−y Ta y O4−(1/2)y and Ba5Ta4O15 crystals can be assigned to the cubic perovskite-type SrWO2.05N0.95 (ICDD PDF# 88-0828) and BaTaO2N with space group of Pm-3 m (ICDD PDF# 84-1748), respectively. During the Ba5Ta4O15-to-BaTaO2N conversion, the BaO is typically formed. However, the formation of BaO was not observed here because the volatilization of barium took place during the long thermal ammonolysis:

$${\text{Ba}}_{5} {\text{Ta}}_{4} {\text{O}}_{15} + 4{\text{NH}}_{3} \to 4{\text{BaTaO}}_{2} {\text{N}} + {\text{BaO}} \uparrow +\, 6{\text{H}}_{2} {\text{O}} \uparrow$$
(3)

Here, the differences in both the XRD peak shifts and the unit cell parameters (Table 2) of the Sr1−x Ba x W1−y Ta y O4−(1/2)y crystal structures can be interpreted by Vegard’s law [20]. As shown in Fig. 3a, b, d, e, the 110 diffraction peak of the Sr1−x Ba x W1−y Ta y (O,N)3 crystal structures at 30.72–31.77° in the XRD patterns shifted slightly to a lower 2θ angle compared with that of the SrW(O,N)3 crystal structures on increasing the level of barium substitution because of an increase in the lattice volume. As the radius of Ba2+ (r ionic = 1.61 Å: CN = 12) is larger than that of Sr2+ (r ionic = 1.44 Å: CN = 12), the increase in the unit cell volume indicates that the strontium in cubic perovskite-type SrW(O,N)3 is partially substituted with barium [19]. In spite of the substitution of smaller W6+ (r ionic = 0.6 Å: CN = 6) and W5+ (r ionic = 0.62 Å: CN = 6) by larger Ta5+ (r ionic = 0.64 Å: CN = 6), the SrW0.75Ta0.25(O,N)3 crystal structures have a smaller unit cell volume in comparison with the SrW(O,N)3 crystal structures because the content of N3− (r ionic = 1.46 Å: CN = 4), which is larger than that of O2− (r ionic = 1.36 Å: CN = 4), was gradually decreased as a consequence of the charge compensation [19]. Based on the obtained results, the effect of the cation and anion substitution on unit cell volume of Sr1−x Ba x W1−y Ta y (O,N)3 can be placed in the following order: Sr2+-to-Ba2+ > N3−-to-O2− > Ta5+-to-W6+ and W5+.

Fig. 3
figure 3

XRD patterns of Sr1−x Ba x W1−y Ta y (O,N)3: (a) x,y = 0,0; (b) x,y = 0.25,0; (c) x,y = 0,0.25; (d) x,y = 0.125,0.125; and (e) BaTa(O,N)3 crystal structures synthesized by thermal ammonolysis of oxide precursors at 900–950 °C for 10–40 h under an NH3 flow

Full size image
Table 2 Unit cell parameters of cubic Sr1−x Ba x W1−y Ta y (O,N)3 (x,y = 0,0; 0.25,0; 0,0.25; and 0.125,0.125) crystal structures synthesized by thermal ammonolysis of oxide precursors at 900–950 °C for 10–40 h under an NH3 flow
Full size table

Figure 4 shows the UV–Vis diffuse reflectance spectra of the Sr1−x Ba x W1−y Ta y (O,N)3 and BaTa(O,N)3 crystal structures. As the valence band edge of oxynitride is more positive than that of oxide due to the contribution of N 2p orbitals having a more positive potential energy than the O 2p orbitals [12, 28], all the oxynitride crystal structures exhibit visible light absorption. The SrW(O,N)3 and Sr0.75Ba0.25W(O,N)3 crystal structures have an upward shift in absorption in the visible light region (λ = 270–800 nm) with unclear absorption edges, and their color changed to dark taupe, which is attributed to the defects such as oxygen and nitrogen vacancies and reduced tungsten species (W5+ and W4+) [12, 13]. According to the theoretical prediction of the formability of the transition metal oxynitrides based on the tolerance and octahedral factors [29], the AW(O,N)3 crystal structures would be formed with the formal compositions of A 2+W6+ON2, A 2+W5+O2N, and A2+W 6+1−x W 5+ x O1+x N2−x . Furthermore, Yashima et al. [30] reported that the Ca x Sr1−x WO2N could have a black color and metallic characteristics because of the presence of the W5+ ion with 5d 1 electronic configuration. The results obtained in this study are consistent with the previously reported theoretical and experimental data. As shown in Fig. 4, the absorption edges of the dark olive-colored SrW0.75Ta0.25(O,N)3, dark viridian-colored Sr0.875Ba0.125W0.875Ta0.125(O,N)3, and brown-colored BaTa(O,N)3 crystal structures are confirmed to be at 557, 581, and 645 nm, respectively. On the contrary to SrW(O,N)3 and Sr0.75Ba0.25W(O,N)3, the light absorption of SrW0.75Ta0.25(O,N)3 and Sr0.875Ba0.125W0.875Ta0.125(O,N)3 beyond 560 and 580 nm is weakened. This indicates that the lower absorption in the near-infrared region stems from the suppression of the generation of the reduced W5+ and W4+ ions with 5d 1 and 5d 2 electronic configuration during thermal ammonolysis due to the partial substitution of B-site with Ta5+ having 5d 0 electronic configuration. Here, the band gap shifted from 2.23 eV for SrW0.75Ta0.25(O,N)3 to 2.13 eV for Sr0.875Ba0.125W0.875Ta0.125(O,N)3 and 1.92 eV for BaTa(O,N)3. The O and N were quantitatively analyzed and found to be in the following stoichiometric compositions in the samples: SrWO2.15N0.85, Sr0.75Ba0.25WO2.37N0.63, SrW0.75Ta0.25O2.52N0.48, Sr0.875Ba0.125W0.875Ta0.125O2.31N0.69, and BaTaO2.45N0.55. The discrepancy between the band-gap energies of SrW0.75Ta0.25(O,N)3, Sr0.875Ba0.125W0.875Ta0.125(O,N)3, and BaTa(O,N)3 can be explained by both the effects of A- and B-site cation change and variation in the oxygen/nitrogen ratio among the synthesized oxynitrides. In previous reports on ABO2N (A = Ca, Sr, Ba; B = Ta, Nb), it was proposed that the band-gap energies of ABO2N can be narrowed by increasing the ionic radius of A 2+ (Ca2+ < Sr2+ < Ba2+) [31,32,33]. Moreover, Ikeda et al. [34] proposed that since the potential of W 5d orbitals is located at a more negative potential than that of Ta 5d orbitals, the conduction band energy level of tantalum-containing tungsten-based materials is more positive compared to tungsten-based materials. The Sr0.875Ba0.125W0.875Ta0.125(O,N)3 crystal structures exhibited a narrower band-gap energy than the SrW0.75Ta0.25(O,N)3 crystal structures due to the partial replacement of smaller Sr2+ by larger Ba2+, lower Ta5+-substitution level, and higher nitrogen concentration in Sr0.875Ba0.125W0.875Ta0.125(O,N)3. Although the Ta-based BaTa(O,N)3 crystal structures have a lower concentration of nitrogen than the Sr0.875Ba0.125W0.875Ta0.125(O,N)3 crystal structures, their band-gap energy was found to be the smallest one among the five samples, resulting from the effect of Ba2+.

Fig. 4
figure 4

UV–Vis diffuse reflectance spectra of Sr1−x Ba x W1−y Ta y (O,N)3: blue x,y = 0,0; gree x,y = 0.25,0; orange x,y = 0,0.25; purple x,y = 0.125,0.125; and red BaTa(O,N)3 crystal structures synthesized by thermal ammonolysis of oxide precursors at 900–950 °C for 10–40 h under an NH3 flow

Full size image

The EDS element mapping images are shown in Fig. 5. The strontium (cyan), barium (purple), oxygen (yellow), and nitrogen (green) elements are uniformly distributed over the SrW(O,N)3, Sr0.75Ba0.25W(O,N)3, SrW0.75Ta0.25(O,N)3, Sr0.875Ba0.125W0.875Ta0.125(O,N)3, and BaTa(O,N)3 crystal structures, implying that the strontium and oxygen were partially substituted by barium and nitrogen in the SrW(O,N)3 lattice. However, the lower intensity of tungsten (blue) signal and higher intensity of tantalum (red) signal are observed in the regions of a few microns in the SrW0.75Ta0.25(O,N)3 and Sr0.875Ba0.125W0.875Ta0.125(O,N)3 crystal structures which are probably related to the ammonolyzed Sr2Ta2O7 crystals (SrTaO2N). All the results of the EDS element mapping images are consistent with the EDS spectra shown in the Supporting Information (Fig. S3).

Fig. 5
figure 5

EDS element mapping images of Sr1−x Ba x W1−y Ta y (O,N)3: (a) x,y = 0,0; (b) x,y = 0.25,0; (c) x,y = 0,0.25; (d) x,y = 0.125,0.125; and (e) BaTa(O,N)3 crystal structures obtained by thermal ammonolysis of oxide precursors at 900–950 °C for 10–40 h under an NH3 flow

Full size image

Photocatalytic water oxidation activities of Sr1−xBaxW1−y Ta y (O,N)3 crystal structures

The as-synthesized oxynitrides were tested as visible light-active photocatalysts for water oxidation half-reaction in the presence of Ag+ as a sacrificial electron acceptor that prevents the H2 evolution reaction, avoids the H2O-forming back reaction (H2 + 1/2O2 → H2O), and enhances the O2 evolution reaction [35]. Figure 6 shows the time course curves of O2 and N2 evolution over the as-synthesized oxynitrides loaded with CoO x (2 wt % Co) nanoparticles and suspended in aqueous AgNO3 solution under visible light irradiation (λ > 420 nm). The loaded CoO x nanoparticles as a cocatalyst for photocatalytic water oxidation can trap photo-generated holes to separate effectively the photo-excited charges using a cocatalyst/photocatalyst interface and to improve the photocatalytic performance. The corresponding photocatalytic reactions are schematically illustrated in the insets of Fig. 6 [7, 36,37,38]. The detected O2 gas was evolved as a result of the oxidation of H2O to O2 by photo-generated holes on CoO x nanoparticles:

$$4{\text{AgNO}}_{3} + 2{\text{H}}_{2} {\text{O}} \to 4{\text{Ag}} + 4{\text{HNO}}_{3} + {\text{O}}_{2} \uparrow$$
(4)

The TONs and TOFs estimated for 5 h photocatalytic O2 evolution reactions are given in Table 3. Generally, a TON of less than one does not confirm the existence of catalysis [39]. Since the resultant TONs of all the samples exceed one, the observed O2 gas evolution reactions are catalytic. Here, the N2 gas evolution was also noticed because the nitrogen anions (N3−) in the as-synthesized oxynitride crystal structures are self-oxidized to N2 by photo-generated holes:

$$2{\text{N}}^{3 - } + 6{\text{h}}^{ + } \to {\text{N}}_{2} \uparrow$$
(5)

Simultaneously, the highly resistive oxides (Sr1−x Ba x W1−y Ta y O4−(1/2)y ) were also formed on the oxynitride surface due to the self-oxidation, hindering the transport of the photo-generated charge carriers from the inside of the crystals to the interface between the crystal and reaction solution [40, 41]. As shown in Fig. 6, all the O2 evolution rates drastically decreased over time, especially after 0.5 h, because of two possible reasons: (i) the self-oxidative formation of the highly resistive oxides and (ii) photo-deposition of metallic silver nanoparticles on the oxynitride crystal surfaces [42]. Here, the metallic silver nanoparticles disturb light absorption and cover surface active sites.

Fig. 6
figure 6

O2 and N2 evolution of CoO x -loaded Sr1−x Ba x W1−y Ta y (O,N)3: (a) x,y = 0,0; (b) x,y = 0.25,0; (c) x,y = 0,0.25; (d) x,y = 0.125,0.125; and (e) BaTa(O,N)3 crystal structures obtained by thermal ammonolysis of oxide precursors at 900–950 °C for 10–40 h under an NH3 flow in the presence of a sacrificial reagent (AgNO3)

Full size image
Table 3 Photocatalytic performance of CoO x -loaded Sr1−x Ba x W1−y Ta y (O,N)3 (x,y = 0,0; 0.25,0; 0,0.25; and 0.125,0.125) on water oxidation and specific surface areas of Sr1−x Ba x W1−y Ta y (O,N)3 crystal structures
Full size table

The CoO x -loaded SrW(O,N)3 (CoO x -SWON), Sr0.75Ba0.25W(O,N)3 (CoO x -SBWON), SrW0.75Ta0.25(O,N)3 (CoO x -SWTON), Sr0.875Ba0.125W0.875Ta0.125(O,N)3 (CoO x -SBWTON), and BaTa(O,N)3 (CoO x -BTON) crystal structures exhibited the total amounts of evolved gases: 17.7, 11.6, 20.5, 11.8, and 15.4 μmol for O2 gas and 19.8, 11.8, 36.1, 9.6, and 11.3 μmol for N2 gas, respectively. CoO x -SWTON demonstrated nearly a 1.2-fold higher amount of evolved O2 gas compared to the CoO x -SWON because of its higher specific surface area and lower concentration of intrinsic defects (W5+) owing to the partial substitution of Ta5+ for W6+ prompting the oxidation of tungsten. However, the total amounts of evolved O2 gas over CoO x -SBWON, CoO x -SBWTON, and CoO x -BTON were much lower than that of CoO x -SWON. According to previous reports on nitrogen-doped tungsten oxide [43, 44], the doping of nitrogen onto the WO6 octahedra in WO3 changes the bond overlap population of W–O leading to the lattice distortion. In addition, Wang et al. [45] revealed that a large lattice distortion of the corner-linked BO6 octahedra in the A 2 B 2O7-type photocatalyst induces the high photocatalytic activity. Therefore, when the nitrogen content of the as-synthesized AB(O,N)3-type oxynitrides is higher, the distortion of the BO6 octahedra (B = W and/or Ta) is expected to become larger, leading to the higher photocatalytic activity. As the nitrogen content of SrW(O,N)3 was much higher than that of SrW0.75Ta0.25(O,N)3, Sr0.875Ba0.125W0.875Ta0.125(O,N)3, and BaTa(O,N)3, the CoO x -SWON showed the highest water oxidation activity among the four samples. From the EDS and XPS results (see Fig. S4 in the Supporting Information), it was found that the SrW0.75Ta0.25(O,N)3 and Sr0.875Ba0.125W0.875Ta0.125(O,N)3 crystal structures contain strontium tantalum oxynitride as a secondary phase. As is known, the strontium tantalum oxynitride has the photocatalytic activity for water oxidation half-reaction under visible light irradiation [46]. Possibly, the secondary phase can provide a positive effect on the O2 evolution reaction. Additionally, the XPS results (see Fig. S4 in the Supporting Information) show that the crystal surfaces of all the synthesized oxynitrides may be partially covered with the metallic tungsten and compounds containing low-valent transition metal cations (e.g., W, WO2, W2N, TaO2, etc.). In general, the compounds with low-valent transition metal cations have metallic characteristics due to the partially filled d orbitals [30, 47,48,49]. Furthermore, the compounds with the fully/partially filled d orbitals are expected to act as a cocatalyst accepting the photo-generated electrons [50]. The total amount of evolved O2 and N2 increases in the following order: CoO x -SWTON > CoO x -SWON > CoO x -BTON > CoO x -SBWON > CoO x -SBWTON, giving a close agreement with the order of the TOF and initial O2 evolution rate shown in Table 3. Noticeably, both O2 and N2 gas evolution amounts certainly depend on the number of photo-generated carriers arriving near the crystal surfaces. Hence, it is concluded that the partial Ta5+-to-W4~6+ substitution is most effective in enhancing the photocatalytic activity of SrW(O,N)3. The photocatalytic water oxidation activities of the Sr1−x Ba x W1−y Ta y (O,N)3 (x,y = 0,0; 0.25,0; 0.125,0.125; 0,0.25; and 1,1) crystal structures can be further improved by enhancing their photostability during the photocatalytic reaction.

Conclusion

In summary, SrW(O,N)3 crystal structures with fully/partially substituted Ba2+ and/or Ta5+ were synthesized by thermal ammonolysis of their corresponding oxide precursors Sr1−x Ba x W1−y Ta y O4−(1/2)y (x,y = 0,0; 0.25,0; 0,0.25; and 0.125,0.125) and Ba5Ta4O15. The synthesized oxynitrides possess porous structures with higher specific surface areas. In the photocatalytic water oxidation activity test, the estimated TONs of the oxynitride samples exceeded one, evidencing that the observed O2 gas evolution reactions were catalytic. The CoO x -loaded SrW0.75Ta0.25(O,N)3 showed a comparatively higher O2 gas evolution activity compared to CoO x -loaded SrW(O,N)3 because of its relatively higher specific surface area and lower concentration of intrinsic defects. Therefore, the partial Ta5+-to-W4~6+ substitution is most effective for enhancing the activity of SrWO2N photocatalyst. After the initial 0.5 h reaction, the O2 gas evolution rate decreased dramatically due to the self-oxidation of the surfaces of the oxynitride crystals, forming the highly resistive oxides, and the photo-deposition of metallic silver nanoparticles on the oxynitride crystal surfaces, disturbing light absorption and cover surface active sites. Further improvement of the photocatalytic activity for water oxidation of the Sr1−x Ba x W1−y Ta y (O,N)3 is in process.