The Enhancement of H2 Evolution over Sr1−1.5xTbxWO4 Solid Solution under Ultraviolet Light Irradiation

In this work, Sr1−1.5xTbxWO4 (0 ≤ x ≤ 0.2) solid solutions were synthesized via a traditional high-temperature solid state method. Le Bail fitting on the powder X-ray diffraction (XRD) pattern showed that these solid solutions are pure phase. Scanning electron microscopy showed that the SrWO4 and Sr0.82Tb0.12WO4 samples are composed of micrometer particles and submicron crystallites, respectively. Ultraviolet–visible diffuse reflectance spectra suggested that the bandgaps of Sr1−1.5xTbxWO4 are narrower than the undoped sample. The Sr0.82Tb0.12WO4 sample, with the assistance of 1.5 wt % Ru-cocatalyst, exhibits the best performance for H2 evolution in 5 vol % aqueous triethanolamine (TEOA), which results in about 6.1 and 2.8 times efficiency improvement compared with the intrinsic SrWO4 in methanol and aqueous TEOA, respectively. All the photocatalysts recycled after the photocatalytic reactions showed no degradation when checked by powder XRD.


Introduction
Photocatalytic technology is a green way to obtain clean energy hydrogen via semiconductor from water under sunlight irradiation [1]. Over the past few decades, various semiconducting materials, including oxide [2], sulfide [3], oxynitride [4], and carbon-based material [5], have been studied extensively in photocatalytic H 2 evolution. The sulfide and oxynitride materials generally have excellent photocatalytic activities, however, both of them are not stable enough in aqueous solution under light irradiation. For instance, the Zn 1−2y Ga 1.7 In 0.3 S 4 and Pt-PdS/CdS samples have outstanding photocatalytic H 2 evolution activities in aqueous S 2− /SO 3 2− solutions [6,7]. GaN/ZnO and LaMg x Ta 1−x O 1+3x N 2−3x solid solutions require a cocatalyst or coating material for photocatalytic H 2 evolution [8,9]. On account of their distinguished electrical conductivity, carbon materials have been utilized to fabricate carbon-based composites [10][11][12]. The majority of oxides were usually only sensitized by UV light, however, if suitable crystal structure can indicate good photocatalytic application [13,14]. The photocatalytic performance of SrWO 4 has been studied extensively in the field of environmental photocatalysis. Suitable SrWO 4 crystal structures are able to degrade various dyes and organic materials under ultraviolet light [15][16][17][18][19][20][21]. For example, rhodamine B and 6G were photodegraded by star-like SrWO 4 microcrystals in water [17]. The methyl orange degradation by SrWO 4 nanoparticles was about 73% after 60 min ultraviolet light illumination [20]. Ibuprofen was photoelectrooxidized by rice-like SrWO 4 nanocrystals in Na 2 SO 4 solution [21]. In addition, there is a

Characterizations
Powder XRD data were collected on a PANalytical X'pert diffractometer equipped with a PIXcel 1D detector (Cu Kα radiation, 1.5406 Å). The operation voltage and current were 40 kV and 40 mA, respectively. Le Bail refinements were performed to obtain cell parameters using the TOPAS software package [24]. Scanning electron microscopy (SEM) was performed on an S4800 at an accelerating voltage of 10 kV. The elemental analysis was performed using SEM fitted with an INCAx-act energy dispersive spectrometer (EDS). Ultraviolet-visible diffuse reflectance spectrum (DRS) was recorded using a Shimadzu UV-3600 spectrometer equipped with an integrating sphere attachment. The analysis range was from 200 to 1200 nm, and BaSO 4 was used as the reflectance standard. Photoluminescence (PL) spectra were measured on a Hitachi F4600 fluorescence spectrometer at room temperature. The analysis range was from 450 to 600 nm, the excited wavelength was 250 nm, and the PL of the baseline was collected by using an empty glass cuvette.

Theoretical Calculations
Theoretical study on SrWO 4 was investigated via the Vienna Ab initio Simulation Package (VASP) [25]. The projector augmented-wave (PAW) method implemented in the VASP code was utilized to describe the interaction between the ionic cores and the valence electrons [26]. The generalized gradient approximation (GGA) parameterized by Perdew, Burke, and Ernzerhof (PBE) was employed to describe the exchange-correlation potential in standard density functional theory calculations [27]. For single point energy and density of states, a cutoff energy of 500 eV for the plane-wave basis and the 3 × 3 × 3 Monkhorst-Pack G-centered k-point meshes were employed.

Photocatalytic Performance Evaluation
Photocatalytic performances were tested on a gas-closed circulation system equipped with a vacuum device (LabSolar-IIIAG system, Perfect Light Ltd. Co.), a 150 mL Pyrex glass reactor, and a gas sampling port that was directly connected to a gas chromatograph (Shanghai Techcomp-GC7900, TCD detector, molecular sieve 5A, N 2 gas carrier). In a typical run, 50 mg of catalyst was dispersed by a magnetic stirrer in 50 mL of 5 vol % TEOA aqueous solution. The solution was stirred, and a 10 • C cycling water bath was applied to keep the reaction vessel at a constant temperature. The light irradiation source was generated by an external 500W Hg lamp (CEL-M500, Beijing Au Light Ltd. Co.).

Preparation of the Cocatalysts
The loading of metal cocatalyst on photocatalyst was performed by a photodeposition method using the above setup [14]. For instance, 50 mg Sr 0.82 Tb 0.12 WO 4 sample, together with 1.0 mL of RuCl 3 (0.97 mg/mL), were mixed in 10 mL of distilled water. This solution was placed in a 150 mL Pyrex glass reactor with an ultrasonic treatment for 10 min, and then the mixture, in the presence of 5 vol % of TEOA, was irradiated using a 500 W high-pressure Hg lamp for 2 h. Finally, the powder sample was collected and washed with deionized water. Figure 1 presents the whole XRD pattern of SrWO 4 with substantial doping of Tb 3+ , and the sharp peaks point to the high crystallinity of the as-prepared samples. For Sr 1−1.5x Tb x WO 4 (0.00 ≤ x ≤ 0.20), it is evident that a pure phase of solid solutions was synthesized without any impurity peaks, when compared to the simulated XRD (ICSD-155793) of SrWO 4 . The peak shift by Tb 3+ -doping is not obvious, but we can determine the change of the cell lattice parameters (a, c, and V) by Le Bail fitting of the whole powder XRD pattern [2,14]. The plot of a, c, and V, along with the increase in x, suggests a linear shrinkage (see Figure 2). These results evidently confirm that Tb 3+ has been successfully incorporated into SrWO 4 without any structural change.

Results and Discussion
Scherrer's formula works best for the nanomaterials (1-100 nm), and the average crystallite sizes of SrWO 4 and Sr 0.82 Tb 0.12 WO 4 particles were estimated as per the following equation: where t is the crystallite size of the particle (assuming particles are spherical), K = 0.9, λ is the wavelength of X-ray radiation, B is the full width at half-maximum of the diffracted peak, and θ B is the diffraction angle [28]. The estimated crystallite sizes based on the (011) peak for SrWO 4  dispersed by a magnetic stirrer in 50 mL of 5 vol % TEOA aqueous solution. The solution was stirred, and a 10 °C cycling water bath was applied to keep the reaction vessel at a constant temperature. The light irradiation source was generated by an external 500W Hg lamp (CEL-M500, Beijing Au Light Ltd. Co.).

Preparation of the Cocatalysts
The loading of metal cocatalyst on photocatalyst was performed by a photodeposition method using the above setup [14]. For instance, 50 mg Sr0.82Tb0.12WO4 sample, together with 1.0 mL of RuCl3 (0.97 mg/mL), were mixed in 10 mL of distilled water. This solution was placed in a 150 mL Pyrex glass reactor with an ultrasonic treatment for 10 min, and then the mixture, in the presence of 5 vol % of TEOA, was irradiated using a 500 W high-pressure Hg lamp for 2 h. Finally, the powder sample was collected and washed with deionized water. Figure 1 presents the whole XRD pattern of SrWO4 with substantial doping of Tb 3+ , and the sharp peaks point to the high crystallinity of the as-prepared samples. For Sr1−1.5xTbxWO4 (0.00 ≤ x ≤ 0.20), it is evident that a pure phase of solid solutions was synthesized without any impurity peaks, when compared to the simulated XRD (ICSD-155793) of SrWO4. The peak shift by Tb 3+ -doping is not obvious, but we can determine the change of the cell lattice parameters (a, c, and V) by Le Bail fitting of the whole powder XRD pattern [2,14]. The plot of a, c, and V, along with the increase in x, suggests a linear shrinkage (see Figure 2). These results evidently confirm that Tb 3+ has been successfully incorporated into SrWO4 without any structural change.

Results and Discussion
Scherrer's formula works best for the nanomaterials (1-100 nm), and the average crystallite sizes of SrWO4 and Sr0.82Tb0.12WO4 particles were estimated as per the following equation: where t is the crystallite size of the particle (assuming particles are spherical), K = 0.9, λ is the wavelength of X-ray radiation, B is the full width at half-maximum of the diffracted peak, and θB is the diffraction angle [28]. The estimated crystallite sizes based on the (011) peak for SrWO4 and Sr0.82Tb0.12WO4 are approximately 89.4 and 50.3 nm, respectively.  The as-prepared solid solutions were observed by electron microscopy. Figure 3 shows the SEM images of nondoped SrWO4 and Sr0.82Tb0.12WO4 powders prepared under identical conditions. Both particles were well-crystallized, which is consistent with the XRD analysis. The particle size of Sr0.82Tb0.12WO4 powder, 0.1-1 μm, was remarkably smaller than that of nondoped SrWO4 powder, 1-2.4 μm. Since the SrWO4 and Sr0.82Tb0.12WO4 powders are bulk samples, the particle size determined by SEM is different from the particle size calculated by Scherrer's formula from XRD, but the relative size of the SrWO4 and Sr0.82Tb0.12WO4 powders were not changed. A sharp surface edge was observed for the La-doped SrWO4 powder, whereas the surface of nondoped SrWO4 was flat. The insert in Figure 3b shows an elemental analysis performed on the Sr0.82Tb0.12WO4 sample, which gave an average atomic ratio of Sr/Tb/W/O = 0.84:0.12:1.00:4.02.  Figure 4a shows that the absorption band of SrWO4 lies mainly in the UV region, and there is a steep edge which indicates that the absorption band is not due to the transition from impurity energy levels but from the bandgap transition [29]. For most semiconductors, the dependence of the absorption coefficient α on the bandgap energy Eg can be expressed by the following equation: αhv = A(hv − Eg) n/2 , where h, v, and A are the Planck constant, light frequency, and proportionality, respectively, and n is determined on the basis of the transition type (i.e., n = 1 for direct transition, n = 4 for indirect transition) [23]. The best fit of (αhv) 2 vs. Eg was obtained only when n is 1, suggesting that direct transition across the energy bandgap of SrWO4 is allowed (see Figure 4b). The extrapolated value of hv at α = 0 gives an absorption edge energy corresponding to Eg, which is 4.73 eV for SrWO4. Note that the reported value of the bandgap is in the range of 3.2-4.96 eV in the literature [16,[18][19][20][21], where the difference comes from the different morphology of these samples. The as-prepared solid solutions were observed by electron microscopy. Figure 3 shows the SEM images of nondoped SrWO 4 and Sr 0.82 Tb 0.12 WO 4 powders prepared under identical conditions. Both particles were well-crystallized, which is consistent with the XRD analysis. The particle size of Sr 0.82 Tb 0.12 WO 4 powder, 0.1-1 µm, was remarkably smaller than that of nondoped SrWO 4 powder, 1-2.4 µm. Since the SrWO 4 and Sr 0.82 Tb 0.12 WO 4 powders are bulk samples, the particle size determined by SEM is different from the particle size calculated by Scherrer's formula from XRD, but the relative size of the SrWO 4 and Sr 0.82 Tb 0.12 WO 4 powders were not changed. A sharp surface edge was observed for the La-doped SrWO 4 powder, whereas the surface of nondoped SrWO 4 was flat. The insert in Figure 3b shows an elemental analysis performed on the Sr 0.82 Tb 0.12 WO 4 sample, which gave an average atomic ratio of Sr/Tb/W/O = 0.84:0.12:1.00:4.02. The as-prepared solid solutions were observed by electron microscopy. Figure 3 shows the SEM images of nondoped SrWO4 and Sr0.82Tb0.12WO4 powders prepared under identical conditions. Both particles were well-crystallized, which is consistent with the XRD analysis. The particle size of Sr0.82Tb0.12WO4 powder, 0.1-1 μm, was remarkably smaller than that of nondoped SrWO4 powder, 1-2.4 μm. Since the SrWO4 and Sr0.82Tb0.12WO4 powders are bulk samples, the particle size determined by SEM is different from the particle size calculated by Scherrer's formula from XRD, but the relative size of the SrWO4 and Sr0.82Tb0.12WO4 powders were not changed. A sharp surface edge was observed for the La-doped SrWO4 powder, whereas the surface of nondoped SrWO4 was flat. The insert in Figure 3b shows an elemental analysis performed on the Sr0.82Tb0.12WO4 sample, which gave an average atomic ratio of Sr/Tb/W/O = 0.84:0.12:1.00:4.02.  Figure 4a shows that the absorption band of SrWO4 lies mainly in the UV region, and there is a steep edge which indicates that the absorption band is not due to the transition from impurity energy levels but from the bandgap transition [29]. For most semiconductors, the dependence of the absorption coefficient α on the bandgap energy Eg can be expressed by the following equation: αhv = A(hv − Eg) n/2 , where h, v, and A are the Planck constant, light frequency, and proportionality, respectively, and n is determined on the basis of the transition type (i.e., n = 1 for direct transition, n = 4 for indirect transition) [23]. The best fit of (αhv) 2 vs. Eg was obtained only when n is 1, suggesting that direct transition across the energy bandgap of SrWO4 is allowed (see Figure 4b). The extrapolated value of hv at α = 0 gives an absorption edge energy corresponding to Eg, which is 4.73 eV for SrWO4. Note that the reported value of the bandgap is in the range of 3.2-4.96 eV in the literature [16,[18][19][20][21], where the difference comes from the different morphology of these samples.  Figure 4a shows that the absorption band of SrWO 4 lies mainly in the UV region, and there is a steep edge which indicates that the absorption band is not due to the transition from impurity energy levels but from the bandgap transition [29]. For most semiconductors, the dependence of the absorption coefficient α on the bandgap energy E g can be expressed by the following equation: αhv = A(hv − E g ) n/2 , where h, v, and A are the Planck constant, light frequency, and proportionality, respectively, and n is determined on the basis of the transition type (i.e., n = 1 for direct transition, n = 4 for indirect transition) [23]. The best fit of (αhv) 2 vs. E g was obtained only when n is 1, suggesting that direct transition across the energy bandgap of SrWO 4 is allowed (see Figure 4b). The extrapolated value of hv at α = 0 gives an absorption edge energy corresponding to E g , which is 4.73 eV for SrWO 4 . Note that the reported value of the bandgap is in the range of 3.2-4.96 eV in the literature [16,[18][19][20][21], where the difference comes from the different morphology of these samples.  As a result of the fluorescence phenomenon of Sr1−1.5xTbxWO4 solid solutions, the DRS patterns were interrupted at a range of 200~270 nm (see Figure 4a), in agreement with the excitation wavelength of excitation spectra patterns (see Figure 5a). However, we can conclude that the bandgaps of the Tb-doped samples are smaller than the nondoped SrWO4, because the light absorbance of the Tb-doped samples are obviously bigger than the nondoped samples at the wavelengths of 275~400 nm. Sometimes, the weak absorbance of photocatalyst could lead to a significant change of photocatalytic activity, such as Ni-doped InTaO4 [2], and carbon dot-decorated C3N4 [30]. Nevertheless, in our work, the doped sample did not respond to simulated sunlight with 5 vol % TEOA solution. Figure 5a shows the photoluminescence emission spectra of the solid solutions by monitoring the Tb 3+ emission at 250 nm. The change trends of emission intensity appeared as "volcano" types by increasing the Tb 3+ content (see Figure 5b). The band-band PL phenomenon was monitored with the light energy approximately equal to the bandgap energy of the sample [31]. Usually, the recombination of photocarriers and PL intensity are positively correlated in photocatalytic studies [32,33], and the photocatalytic activity related to the recombination of photogenerated electron and hole [34]. In other words, the weak PL intensity means that photoexcitation is difficult. The optimal sample regarding photoluminescence is Sr0.82Tb0.12WO4, which means that the photogenerated electrons are easily excited from valence band to conduction band.  As a result of the fluorescence phenomenon of Sr 1−1.5x Tb x WO 4 solid solutions, the DRS patterns were interrupted at a range of 200~270 nm (see Figure 4a), in agreement with the excitation wavelength of excitation spectra patterns (see Figure 5a). However, we can conclude that the bandgaps of the Tb-doped samples are smaller than the nondoped SrWO 4 , because the light absorbance of the Tb-doped samples are obviously bigger than the nondoped samples at the wavelengths of 275~400 nm. Sometimes, the weak absorbance of photocatalyst could lead to a significant change of photocatalytic activity, such as Ni-doped InTaO 4 [2], and carbon dot-decorated C 3 N 4 [30]. Nevertheless, in our work, the doped sample did not respond to simulated sunlight with 5 vol % TEOA solution. Figure 5a shows the photoluminescence emission spectra of the solid solutions by monitoring the Tb 3+ emission at 250 nm. The change trends of emission intensity appeared as "volcano" types by increasing the Tb 3+ content (see Figure 5b). The band-band PL phenomenon was monitored with the light energy approximately equal to the bandgap energy of the sample [31]. Usually, the recombination of photocarriers and PL intensity are positively correlated in photocatalytic studies [32,33], and the photocatalytic activity related to the recombination of photogenerated electron and hole [34]. In other words, the weak PL intensity means that photoexcitation is difficult. The optimal sample regarding photoluminescence is Sr 0.82 Tb 0.12 WO 4 , which means that the photogenerated electrons are easily excited from valence band to conduction band.  As a result of the fluorescence phenomenon of Sr1−1.5xTbxWO4 solid solutions, the DRS patterns were interrupted at a range of 200~270 nm (see Figure 4a), in agreement with the excitation wavelength of excitation spectra patterns (see Figure 5a). However, we can conclude that the bandgaps of the Tb-doped samples are smaller than the nondoped SrWO4, because the light absorbance of the Tb-doped samples are obviously bigger than the nondoped samples at the wavelengths of 275~400 nm. Sometimes, the weak absorbance of photocatalyst could lead to a significant change of photocatalytic activity, such as Ni-doped InTaO4 [2], and carbon dot-decorated C3N4 [30]. Nevertheless, in our work, the doped sample did not respond to simulated sunlight with 5 vol % TEOA solution. Figure 5a shows the photoluminescence emission spectra of the solid solutions by monitoring the Tb 3+ emission at 250 nm. The change trends of emission intensity appeared as "volcano" types by increasing the Tb 3+ content (see Figure 5b). The band-band PL phenomenon was monitored with the light energy approximately equal to the bandgap energy of the sample [31]. Usually, the recombination of photocarriers and PL intensity are positively correlated in photocatalytic studies [32,33], and the photocatalytic activity related to the recombination of photogenerated electron and hole [34]. In other words, the weak PL intensity means that photoexcitation is difficult. The optimal sample regarding photoluminescence is Sr0.82Tb0.12WO4, which means that the photogenerated electrons are easily excited from valence band to conduction band.  To study the photocatalytic activity of SrWO 4 , the aqueous methanol and TEOA solutions were utilized as electron-donating sacrificial agents for H 2 evolution in our photocatalytic reactions. The H 2 evolution rate of SrWO 4 in 5 vol % aqueous TEOA is 2.1 times higher than it is in the 20 vol % aqueous methanol (see Figure 6). Furthermore, compared with the literature [22], the H 2 evolution rate of SrWO 4 is almost the same value as in aqueous methanol under UV light. To study the photocatalytic activity of SrWO4, the aqueous methanol and TEOA solutions were utilized as electron-donating sacrificial agents for H2 evolution in our photocatalytic reactions. The H2 evolution rate of SrWO4 in 5 vol % aqueous TEOA is 2.1 times higher than it is in the 20 vol % aqueous methanol (see Figure 6). Furthermore, compared with the literature [22], the H2 evolution rate of SrWO4 is almost the same value as in aqueous methanol under UV light. The photocatalytic activity of the Sr1−1.5xTbxWO4 solid solutions toward the H2 production at 5 h reaction times, in the presence of 5 vol % TEOA as electron-donating sacrificial agent under UV light, is shown in Figure 7. The H2 production rates were changed by increasing the Tb 3+ content in a trend of "volcano" type, which is in agreement with the photoluminescence emission intensity. Figure 7a shows that the optimum sample is Sr0.82Tb0.12WO4, and that its H2 evolution rate is 128.6 μmol/h/g. Compared with the undoped SrWO4, the photocatalytic activity of Sr0.82Tb0.12WO4 is improved about 1.3 times, which can be explained as follows: (1) The light absorption of the solid solutions are enhanced via the increasing of Tb 3+ , leading to the improvement of photocatalytic activity [35]; (2) The non-equivalent doping of Tb 3+ to replace Sr 2+ in SrWO4 can fabricate a metal vacancy in the bulk and surface of the material, which may constitute a recombination center [36]. Figure 7b shows that the H2 evolution of the solid solutions is linear against time.  The photocatalytic activity of the Sr 1−1.5x Tb x WO 4 solid solutions toward the H 2 production at 5 h reaction times, in the presence of 5 vol % TEOA as electron-donating sacrificial agent under UV light, is shown in Figure 7. The H 2 production rates were changed by increasing the Tb 3+ content in a trend of "volcano" type, which is in agreement with the photoluminescence emission intensity. Figure 7a shows that the optimum sample is Sr 0.82 Tb 0.12 WO 4 , and that its H 2 evolution rate is 128.6 µmol/h/g. Compared with the undoped SrWO 4 , the photocatalytic activity of Sr 0.82 Tb 0.12 WO 4 is improved about 1.3 times, which can be explained as follows: (1) The light absorption of the solid solutions are enhanced via the increasing of Tb 3+ , leading to the improvement of photocatalytic activity [35]; (2) The non-equivalent doping of Tb 3+ to replace Sr 2+ in SrWO 4 can fabricate a metal vacancy in the bulk and surface of the material, which may constitute a recombination center [36]. Figure 7b shows that the H 2 evolution of the solid solutions is linear against time. To study the photocatalytic activity of SrWO4, the aqueous methanol and TEOA solutions were utilized as electron-donating sacrificial agents for H2 evolution in our photocatalytic reactions. The H2 evolution rate of SrWO4 in 5 vol % aqueous TEOA is 2.1 times higher than it is in the 20 vol % aqueous methanol (see Figure 6). Furthermore, compared with the literature [22], the H2 evolution rate of SrWO4 is almost the same value as in aqueous methanol under UV light. The photocatalytic activity of the Sr1−1.5xTbxWO4 solid solutions toward the H2 production at 5 h reaction times, in the presence of 5 vol % TEOA as electron-donating sacrificial agent under UV light, is shown in Figure 7. The H2 production rates were changed by increasing the Tb 3+ content in a trend of "volcano" type, which is in agreement with the photoluminescence emission intensity. Figure 7a shows that the optimum sample is Sr0.82Tb0.12WO4, and that its H2 evolution rate is 128.6 μmol/h/g. Compared with the undoped SrWO4, the photocatalytic activity of Sr0.82Tb0.12WO4 is improved about 1.3 times, which can be explained as follows: (1) The light absorption of the solid solutions are enhanced via the increasing of Tb 3+ , leading to the improvement of photocatalytic activity [35]; (2) The non-equivalent doping of Tb 3+ to replace Sr 2+ in SrWO4 can fabricate a metal vacancy in the bulk and surface of the material, which may constitute a recombination center [36]. Figure 7b shows that the H2 evolution of the solid solutions is linear against time.  The H 2 evolution rates of SrWO 4 micro/nanostructures were obviously improved by Pt-and Ru-cocatalysts [22]. Usually, metal cocatalysts offer photocatalytic water reduction sites, which means that photogenerated electrons prefer to move to metal cocatalysts [37,38]. In our experiment, the various cocatalysts, including Cu, Ag, Au, Pt, Ni, and Ru, were loaded onto the Sr 0.82 Tb 0.12 WO 4 sample to promote photocatalytic activity (Figure 8a). The Ru-cocatalyst loaded onto the photocatalysts had the optimum H 2 production rate (128.6 µmol/h/g), which is 1.3 times that of the nonloaded SrWO 4 . Figure 8b shows that the H 2 evolution of the cocatalyst-loaded Sr 0.82 Tb 0.12 WO 4 samples is linear against time. The H2 evolution rates of SrWO4 micro/nanostructures were obviously improved by Pt-and Ru-cocatalysts [22]. Usually, metal cocatalysts offer photocatalytic water reduction sites, which means that photogenerated electrons prefer to move to metal cocatalysts [37,38]. In our experiment, the various cocatalysts, including Cu, Ag, Au, Pt, Ni, and Ru, were loaded onto the Sr0.82Tb0.12WO4 sample to promote photocatalytic activity (Figure 8a). The Ru-cocatalyst loaded onto the photocatalysts had the optimum H2 production rate (128.6 μmol/h/g), which is 1.3 times that of the nonloaded SrWO4. Figure 8b shows that the H2 evolution of the cocatalyst-loaded Sr0.82Tb0.12WO4 samples is linear against time. The optimum usage of Ru-cocatalyst for the Sr0.82Tb0.12WO4 sample was observed to be 1.5 wt % (Figure 9a). Its H2 evolution rate was 281.9 μmol/h/g, which is 2.8 times that of the nonloaded SrWO4. The photocatalytic activity of the optimum sample is stable after 30 h UV light irradiation, as shown in Figure 9b. In our work, the control experiments were performed without any sacrificial donor. However, the SrWO4, Sr0.82Tb0.12, and 1.5 wt % Ru/Sr0.82Tb0.12WO4 samples had no detectable photocatalytic activity in H2 evolution. The XRD patterns of the as-prepared photocatalysts showed no obvious change before and after photocatalytic reaction (see Figure 10).  The optimum usage of Ru-cocatalyst for the Sr 0.82 Tb 0.12 WO 4 sample was observed to be 1.5 wt % (Figure 9a). Its H 2 evolution rate was 281.9 µmol/h/g, which is 2.8 times that of the nonloaded SrWO 4 . The photocatalytic activity of the optimum sample is stable after 30 h UV light irradiation, as shown in Figure 9b. In our work, the control experiments were performed without any sacrificial donor. However, the SrWO 4 , Sr 0.82 Tb 0.12 , and 1.5 wt % Ru/Sr 0.82 Tb 0.12 WO 4 samples had no detectable photocatalytic activity in H 2 evolution. The XRD patterns of the as-prepared photocatalysts showed no obvious change before and after photocatalytic reaction (see Figure 10). The H2 evolution rates of SrWO4 micro/nanostructures were obviously improved by Pt-and Ru-cocatalysts [22]. Usually, metal cocatalysts offer photocatalytic water reduction sites, which means that photogenerated electrons prefer to move to metal cocatalysts [37,38]. In our experiment, the various cocatalysts, including Cu, Ag, Au, Pt, Ni, and Ru, were loaded onto the Sr0.82Tb0.12WO4 sample to promote photocatalytic activity (Figure 8a). The Ru-cocatalyst loaded onto the photocatalysts had the optimum H2 production rate (128.6 μmol/h/g), which is 1.3 times that of the nonloaded SrWO4. Figure 8b shows that the H2 evolution of the cocatalyst-loaded Sr0.82Tb0.12WO4 samples is linear against time. The optimum usage of Ru-cocatalyst for the Sr0.82Tb0.12WO4 sample was observed to be 1.5 wt % (Figure 9a). Its H2 evolution rate was 281.9 μmol/h/g, which is 2.8 times that of the nonloaded SrWO4. The photocatalytic activity of the optimum sample is stable after 30 h UV light irradiation, as shown in Figure 9b. In our work, the control experiments were performed without any sacrificial donor. However, the SrWO4, Sr0.82Tb0.12, and 1.5 wt % Ru/Sr0.82Tb0.12WO4 samples had no detectable photocatalytic activity in H2 evolution. The XRD patterns of the as-prepared photocatalysts showed no obvious change before and after photocatalytic reaction (see Figure 10).  In order to determine that the enhancement is really due to Tb or Ru for the optimum 1.5 wt % Ru-loaded Sr0.82Tb0.12WO4 sample, we compared the photocatalytic H2 evolution of SrWO4, 1.5 wt % Ru-loaded SrWO4, Sr0.82Tb0.12WO4, and 1.5 wt % Ru-loaded Sr0.82Tb0.12WO4 (see Table 1). Indeed, the value of [3]: [1] is larger than the value of [2]: [1], which means the doped Tb is better than the loaded Ru for H2 evolution. The value of [4]: [2] is larger than [4]: [3], which indicates that the doped Tb is also better. Therefore, we can conclude that Tb is relatively more important for the H2 evolution of SrWO4 in our experiment. To obtain further insight into the photocatalytic activity of SrWO4, the electronic structure of SrWO4 was investigated by VASP calculations. Figure 11 shows the energy band dispersion and density of states (DOS). Although the bandgap (4.284 eV) from VASP calculations is usually underestimated, it nonetheless often provides important insight into the physicochemical behavior of the investigated materials [6,14]. The material is a direct bandgap semiconductor, as has been revealed from the DRS pattern. As shown in Figure 11a, both the valence band maximum and the conduction band minimum of SrWO4 are located at the S point of the Brillouin zone, which explains the features of the measured DRS pattern, as does the reported literature [22]. This is in agreement with the inference from the DRS pattern. Both the lowly dispersive valence bands and the conduction bands should not be beneficial for the transport of the photoexcited electrons and holes. Their lowly dispersive bands, in turn, are likely to result in high recombination of electron hole pairs and, thus, account for the low photocatalytic activity of SrWO4. The total DOS shows that the bands of SrWO4 can be classified into two parts (see Figure 11b). The top of the valence band is completely dominated by the O 2p orbital, while the bottom of the conduction band is mainly constituted by the W 5d orbital and the Sr 4d orbital. In addition, the band structure indicates that In order to determine that the enhancement is really due to Tb or Ru for the optimum 1.5 wt % Ru-loaded Sr 0.82 Tb 0.12 WO 4 sample, we compared the photocatalytic H 2 evolution of SrWO 4 , 1.5 wt % Ru-loaded SrWO 4 , Sr 0.82 Tb 0.12 WO 4 , and 1.5 wt % Ru-loaded Sr 0.82 Tb 0.12 WO 4 (see Table 1). Indeed, the value of [3]: [1] is larger than the value of [2]: [1], which means the doped Tb is better than the loaded Ru for H 2 evolution. The value of [4]: [2] is larger than [4]: [3], which indicates that the doped Tb is also better. Therefore, we can conclude that Tb is relatively more important for the H 2 evolution of SrWO 4 in our experiment. To obtain further insight into the photocatalytic activity of SrWO 4 , the electronic structure of SrWO 4 was investigated by VASP calculations. Figure 11 shows the energy band dispersion and density of states (DOS). Although the bandgap (4.284 eV) from VASP calculations is usually underestimated, it nonetheless often provides important insight into the physicochemical behavior of the investigated materials [6,14]. The material is a direct bandgap semiconductor, as has been revealed from the DRS pattern. As shown in Figure 11a, both the valence band maximum and the conduction band minimum of SrWO 4 are located at the S point of the Brillouin zone, which explains the features of the measured DRS pattern, as does the reported literature [22]. This is in agreement with the inference from the DRS pattern. Both the lowly dispersive valence bands and the conduction bands should not be beneficial for the transport of the photoexcited electrons and holes. Their lowly dispersive bands, in turn, are likely to result in high recombination of electron hole pairs and, thus, account for the low photocatalytic activity of SrWO 4 . The total DOS shows that the bands of SrWO 4 can be classified into two parts (see Figure 11b). The top of the valence band is completely dominated by the O 2p orbital, while the bottom of the conduction band is mainly constituted by the W 5d orbital and the Sr 4d orbital. In addition, the band structure indicates that charge transfer upon photoexcitation occurs from the O 2p orbital to the empty W 5d orbital and Sr 4d orbital. charge transfer upon photoexcitation occurs from the O 2p orbital to the empty W 5d orbital and Sr 4d orbital.

Conclusions
We prepared Sr1−1.5xTbxWO4 solid solutions by high-temperature solid state reaction. Le Bail fitting on powder XRD verified high purity and crystallinity of the as-prepared samples, and the SEM images showed that the SrWO4 and Sr0.82La0.12WO4 were composed of micron and submicron crystallites, respectively. DRS and theoretical calculations of SrWO4 samples suggested a wide bandgap characteristic, which was 4.73 eV, assuming the direct semiconductor model. The valence band maximum was composed of O 2p orbitals, and the conduction band minimum was composed of both W 5d and Sr 4d orbitals. The H2 evolution rates of SrWO4 samples were 46.4 μmol/h/g and 99.3 μmol/h/g, with the 20 vol % methanol and 5 vol % TEOA aqueous solutions, respectively. The photocatalytic activity of SrWO4 was improved by doping Tb 3+ and then loading cocatalysts. The 1.5 wt % Ru-cocatalyst loaded onto Sr0.82Tb0.12WO4 improved the H2 evolution rate to 281.7 μmol/h/g. This photocatalyst remained stable and active even after five cycles (25 h in total). The improvement of photocatalytic performance for 1.5 wt % Ru/Sr0.82Tb0.12WO4 was mainly due to the doped Tb in the crystal structure. These samples did not respond to simulated sunlight with the 5 vol % TEOA, and had no detectable photocatalytic activity in water under UV light. Our study demonstrates the photocatalytic H2 evolution of Sr1−1.5xTbxWO4, and our preliminary attempt of Tb 3+ -doping and loading of various cocatalysts did, indeed, improve the activity. The next steps involve extending the studies to other rare earth ions to produce materials that absorb more of the visible region of the spectrum, or to complete the water-splitting reaction.