Photocatalytic Water Splitting for Hydrogen Production with Gd2MSbO7 (M = Fe, In, Y) Photocatalysts under Visible Light Irradiation

Novel photocatalysts Gd2FeSbO7, Gd2InSbO7 and Gd2YSbO7 were synthesized by the solid state reaction method for the first time. A comparative study about the structural and photocatalytic properties of Gd2MSbO7 (M = Fe, In, Y) was reported. The results showed that Gd2FeSbO7, Gd2InSbO7 and Gd2YSbO7 crystallized with the pyrochlore-type structure, cubic crystal system and space group Fd3m. The lattice parameter a for Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was 10.276026 Å, 10.449546 Å or 10.653651 Å. The band gap of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was estimated to be 2.151 eV, 2.897 eV or 2.396 eV. For the photocatalytic water-splitting reaction, H2 or O2 evolution was observed from pure water with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst under visible light irradiation (wavelength > 420 nm). Moreover, H2 or O2 also spilt by using Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst from CH3OH/H2O or AgNO3/H2O solutions under visible light irradiation (λ > 420 nm). Gd2FeSbO7 showed the highest activity compared with Gd2InSbO7 or Gd2YSbO7. At the same time, Gd2InSbO7 showed higher activity compared with Gd2YSbO7. The photocatalytic activities were further improved under visible light irradiation with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 being loaded by Pt, NiO or RuO2. The effect of Pt was better than that of NiO or RuO2 for improving the photocatalytic activity of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7.


Introduction
Since water splitting reaction with TiO2 as catalyst was discovered in 1972 [1], photocatalysis, as one energy-conserving and green emerging technology, has been studied massively and in detail for producing renewable hydrogen. However, early photocatalysts were applied conditionally for their ultraviolet light response. In order to make good use of exhaustless solar energy, plenty of endeavors were made to develop visible light responsive photocatalysts. By summarizing recent research achievements, there were mainly two approaches to obtain visible light responsive photocatalysts. The first approach was to modify TiO2, such as by ion doping [2][3][4] or forming heterojunction [5][6][7], which was studied most repeatedly due to its non-toxic performance, excellent stability and low cost. The second approach was to develop new photocatalysts, which had also achieved great progress [8][9][10][11][12][13][14]. Especially, the reported CdS photocatalyst [15], with Pt and PdS together as cocatalysts, realized a high quantum efficiency up to 93% in photocatalytic H2 production, which had aroused wide attention across the world. It was known that TiO2 could not be used in the visible light region and could only split water under ultraviolet light irradiation. Moreover, ultraviolet light only occupied 4% of sunlight, which was a limitative factor for photocatalytic technology with TiO2 as the catalyst. Thus, the efficient photocatalysts which could produce electron-hole pairs under visible light irradiation should be developed because visible light occupied 43% of sunlight.
Bi2MNbO7 (M = Fe 3+ , In 3+ ) [16] photocatalysts, as one remarkable representation of A2B2O7 compound family, were first found to realize H2 evolution which was obtained from pure H2O during ultraviolet light irradiation. Thereafter, our research group found that Y2GaSbO7 or Y2GdSbO7 [17] compound also successfully realized photocatalytic water splitting for hydrogen production under visible light irradiation. Therefore, we could reasonably deduce that the substitution of metallic element in A2B2O7 compounds by another suitable metallic element might provide a way to find new photocatalysts which could efficiently catalyze water splitting reaction even under visible light irradiation. In addition, the substitution might result in the lattice O 2− and O − ionosorbed on the surface, which created another way to generate hydroxyl radicals and the highly reactive atomic oxygen species and thus could enhance the photocatalytic activity of solid-solution photocatalysts [18]. Moreover, the substitution of metal ion in A2B2O7 compounds would probably affect the energy level and narrow the band gap of the A2B2O7 compound. As a result, some impurity energy levels or narrow band gaps which had low band gap energy would promote carrier concentration, and thus led to huge change in photocatalytic properties of photocatalysts [19]. Therefore, we speculated Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 catalyst might have phocatalytic potential for water splitting reaction, of which we were more convinced when we found that above novel photocatalysts successfully realized degrading rhodamine B under visible light irradiation [20,21].
In this experiment, we synthesized aforementioned photocatalysts Gd2MSbO7 (M = Fe, In, Y) again by a solid-state reaction method. Different from previous studies, Gd2MSbO7 (M = Fe, In, Y), herein, served as photocatalysts for splitting water into hydrogen under visible light irradiation. Meanwhile, the structural, photophysical and photocatalytic properties of photocatalysts Gd2MSbO7 (M = Fe, In, Y) were also analyzed more comprehensively.

Experimental Section
The novel photocatalysts were synthesized by a solid-state reaction method. Gd2O3, In2O3, Y2O3, Fe2O3 and Sb2O5 with purity of 99.99% (Sinopharm Group Chemical Reagent Co., Ltd., Shanghai, China) were used as starting materials. All powders were dried at 200 °C for 4 h before synthesis. In order to synthesize Gd2YSbO7, the precursors were stoichiometrically mixed, then pressed into small columns and put into an alumina crucible (Shenyang Crucible Co., Ltd., Shenyang, China). Finally, calcination was carried out at 1320 °C for 50 h in an electric furnace (KSL 1700X, Hefei Kejing Materials Technology Co., Ltd., Hefei, China). For the sake of preparing Gd2FeSbO7, the precursors were stoichiometrically churned up, subsequently pressed into small columns and put into an alumina crucible. Eventually, calcination was performed at 1250 °C for 65 h in an electric furnace. Similarly, Gd2InSbO7 was prepared by calcination at 1320 °C for 65 h. The heating rate of calcination was 0.24 °C/s. The crystal structure of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was analyzed by the powder X-ray diffraction method (D/MAX-RB, Rigaku Corporation, Tokyo, Japan) with CuKα radiation (λ = 1.54056 Å). The voltage was 40.0 kV and current was 30.0 mA. The data were collected at 295 K with a step-scan procedure in the range of 2θ = 10°-100°. The step interval was 0.02° for Gd2FeSbO7 or Gd2YSbO7 and the time per step was 1.2 s. The step interval was 0.01° for Gd2InSbO7 and the time per step was 1.0 s. The chemical composition of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was determined by scanning electron microscope-X-ray energy dispersion spectrum (SEM-EDS, LEO 1530VP, LEO Corporation, Krefeld, Germany). The scanning accelerating voltage was 20 kV and linked with an Oxford Instruments X-ray analysis system (Oxford, UK) and X-ray fluorescence spectrometer (XFS, ARL-9800, ARL Corporation, Geneva, Switzerland). The diffuse reflectance spectra of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was analyzed with an UV-visible spectrophotometer (Lambda 40, Perkin-Elmer Corporation, Waltham, MA, USA) in a UV-Vis diffuse reflectance experiment by the dry-pressed disk samples and BaSO4 was used as the reference material. The surface area of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was measured by the Brunauer-Emmett-Teller (BET) method (MS-21, Quantachrome Instruments Corporation, Boynton Beach, FL, USA) with N2 adsorption at liquid nitrogen temperature. All the samples were degassed at 180 °C for 8 h prior to nitrogen adsorption measurements. The BET surface area was determined by a multipoint BET method using the adsorption data in the relative pressure (P/P0) range of 0.05-0.3. A desorption isotherm was used to determine the pore size distribution by the Barret-Joyner-Halender (BJH) method, assuming a cylindrical pore model. The nitrogen adsorption volume at the relative pressure (P/P0) of 0.994 was used to determine the pore volume and average pore size.
The photocatalytic water splitting was carried out under visible light irradiation in a gas closed circulation system with an inner-irradiation type reactor (quartz cell). A light source (300 W Xe arc lamp, Beijing Dongsheng Glass Light Source Factory, Beijing, China) with the incident photon flux I0 of 0.056176 µmol cm −2 s −1 or 0.078245 µmol cm −2 s −1 was focused through a shutter window and a 420 nm or 390 nm cut-off filter onto the window face of the cell. The incident photon flux I0 was determined by a Ray virtual radiation actinometer (FU 100, silicon ray detector, Thorlabs Corporation, Newton, NJ, USA). According to the measured I0 values with 420 nm and 390 nm cut-off filter, the percentage of UV light passing 390 nm cut-off filter was estimated to be 28.2%. The gas which evolved was determined with a thermal conductivity detector TCD-based gas chromatograph (6890 N, Agilent Technologies, Tempe, AZ, USA), which was connected to the gas closed circulation system. One gram of catalyst was suspended in 300 mL H2O under stirrer. Before reaction, the closed gas circulation system and the reaction cell were degassed until O2 and N2 could not be detected. Then, about 35 Torr of argon was charged into the system. H2 evolution reaction was carried out in CH3OH/H2O solution (50 mL CH3OH, 300 mL H2O) with Pt, NiO or RuO2-loaded powder as the catalyst.
For H2 evolution reaction, Pt, NiO or RuO2 which was loaded on the surface of the catalysts, was prepared. Pt was loaded on the catalyst surface by an in situ photodeposition method by using aqueous H2PtCl6 solution (Shanghai Chemical Reagent Research Institute, Shanghai, China) as the Pt source. A typical synthesis procedure was as follows: 1 g catalyst powder and a calculated amount of (0.2 wt%) H2PtCl6 solution were mixed in 150 mL deionized water, and the suspension was then irradiated by a 300 W Xe lamp (λ > 420 nm) under continuous stirring. After 5 h photo-deposition, the suspension was filtered, washed with de-ionized water for 4 times, and finally dried in vacuum at 60 °C for 12 h. NiO or RuO2 which was loaded on the surface of the catalysts, was prepared by the impregnation method by using Ni(NO3)2 or RuCl3 solution (Sinopharm Group Chemical Reagent Co., Ltd., Shanghai, China), separately. Normally, 1 g catalyst powder was ultrasonically dispersed in 20 mL deionized water, and then a calculated amount of metal precursors (1.0 wt%) Ni(NO3)2 or RuCl3 solution was added to the catalyst powder dispersed water. After magnetically stirring for 1 h, the mixed solution was boiled at 100 °C, and next dried at 60 °C for 12 h. Lastly, the as obtained sample was put into an electric furnace to calcine for 4 h at 450 °C in moving air. Figure 1 shows the X-ray powder diffraction patterns of Gd2FeSbO7, Gd2InSbO7 and Gd2YSbO7. It could be seen from Figure 1 that Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was a single phase. The calculations of lattice parameters were performed with the program of Cambridge serial total energy package (CASTEP) and first-principles simulation. The CASTEP package was provided by Materials Studio software and the CASTEP calculation was composed of the plane-wave pseudopotential total energy method according to the density functional theory. Thus, our calculations were based on the plane-wave-based density functional theory (DFT) in generalized gradient approximations (GGA) with Perdew-Burke-Ernzerh of (PBE) exchange-correlation potential.

Characterization
In order to obtain the crystal lattice parameters, Rietveld refinement from X-Ray Diffraction (XRD) data was performed with DBWS software, experimental XRD data and simulation XRD data. The uncertainty of the refined lattice parameters lay in the estimated standard deviation (e.s.d.), calculated by the full pattern fitting program. However, e.s.d. was a measure of precision rather than of accuracy, and these two terms must not be confused. For a sound estimation of the measurement uncertainty of lattice parameters that were refined from XRD data, more information was needed than just the e.s.d. that was provided by the Rietveld refinement of the diffraction pattern of the sample. The outcome of refinements for Gd2InSbO7 generated the unweighted R factors, Rp = 12.13% with space group Fd3m. As for Gd2YSbO7, Rp was 12.16% with space group Fd3m. As for Gd2FeSbO7, Rp was 16.20% with space group Fd3m. According to the Rietveld analysis, Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 had the pyrochlore-type structure and a cubic crystal system which owned a space group Fd3m. The atomic coordinates and structural parameters of Gd2FeSbO7, Gd2InSbO7 and Gd2YSbO7 are listed in Tables 1-3, respectively. The lattice parameter a for Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was 10.276026 Å, 10.449546 Å or 10.653651 Å. Moreover, the XRD results showed that two theta angles of each reflection of Gd2FeSbO7 changed with Fe 3+ being substituted by In 3+ or Y 3+ . The lattice parameter α increased from α = 10.276026 Å for Gd2FeSbO7 to α = 10.449546 Å for Gd2InSbO7, which indicated a decrease in the lattice parameter of the photocatalyst with a decrease of the M ionic radii, Fe 3+ (0.78 Å) < In 3+ (0.92 Å). The lattice parameter α also increased from α = 10.276026 Å for Gd2FeSbO7 to α = 10.653651 Å for Gd2YSbO7, which indicated a decrease in lattice parameter of the photocatalyst with decrease of the M ionic radii, Fe 3+ (0.78 Å) < Y 3+ (1.019 Å). Meanwhile, The lattice parameter α also increased from α = 10.449546 Å for Gd2InSbO7 to α = 10.653651 Å for Gd2YSbO7, which indicated a decrease in the lattice parameter of the photocatalyst with a decrease of the M ionic radii, In 3+ (0.92 Å) < Y 3+ (1.019 Å).      Figure 2 represents the diffuse reflection spectra of Gd2FeSbO7, Gd2InSbO7 and Gd2YSbO7. Compared with well-known photocatalyst TiO2 whose absorption edge was only 380 nm, the absorption band edge of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was found to be 586 nm, 428 nm or 479 nm. Clearly, the obvious absorption did not result from reflection and scattering. Consequently, the apparent absorbance at sub-band gap wavelengths (600-800 nm for Gd2FeSbO7, and 425-800 nm for Gd2InSbO7, and 490-700 nm for Gd2YSbO7) was higher than zero. For a crystalline semiconductor, the optical absorption near the band edge followed the equation: αhν = A (hν − Eg) n [22,23]. Here, A, α, Eg and ν were proportional constant, absorption coefficient, band gap and light frequency, respectively. Eg and n could be calculated by the following steps: (i) plotting ln(αhν) vs. ln(hν − Eg) by assuming an approximate value of Eg; (ii) deducing the value of n according to the slope in this graph; (iii) refining the value of Eg by plotting (αhν) 1/n vs. hν and extrapolating the plot to (αhν) 1/n = 0. According to the above method, the band gap of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was estimated to be 2.151 eV, 2.897 eV or 2.396 eV.

Photocatalytic Activity of Gd2FeSbO7, Gd2InSbO7 and Gd2YSbO7
Generally speaking, the semiconductor photocatalysis started from the direct absorption of supra-band gap photons and the generation of electron-hole pairs in the semiconductor particles. Subsequently, the diffusion of the charge carriers to the surface of the semiconductor particle was followed. Under visible light irradiation, we measured H2 or O2 evolution rate by using Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as photocatalyst from CH3OH/H2O or AgNO3/H2O solution, respectively. The wavelength (λ) dependence on the photocatalytic activity under light irradiation from full arc up to λ = 420 nm was measured by using different cut-off filters. Figure 3a shows the photocatalytic H2 evolution from pure water with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst under visible light irradiation (λ > 420 nm, 0.5 g powder sample, 250 mL pure water). It could be found from Figure 3a that under visible light irradiation, the rate of H2 evolution in the first 28 h with Gd2FeSbO7 as catalyst was 6.329 μmol h −1 g −1 , and that with Gd2InSbO7 as catalyst was 5.157 μmol h −1 g −1 , and that with Gd2YSbO7 as catalyst was 4.314 μmol h −1 g −1 . Besides, under dark condition, no H2 evolution was detected from pure water with above three catalysts, which reflected the phocatalytic H2 evolution activities from pure water of three synthesized catalysts. The reason that water could be split for H2 evolution from pure water with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst under visible light irradiation (λ > 420 nm) was as following: Water could be split at a wavelength higher than 420 nm. However, the wavelength was not cut in exactly at 420 nm, in fact, the wavelength was cut by +50 or −50 nm, which meant that the wavelength up to 370 nm was probably absorbed by Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7, which could split water to provide tiny amounts of hydrogen generation in our experiment. The recycling experiments were performed three times with the same experimental conditions of Figure 3a, and the results were almost the same as the above results in Figure 3a. It could be seen that the photocatalysts we had produced had good stability and were thus desirably recyclable. Figure 3b shows the photocatalytic O2 evolution from pure water with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst under visible light irradiation (λ > 420 nm, 0.5 g powder sample, 250 mL pure water). It could be found from Figure 3b that under visible light irradiation, the rate of O2 evolution in the first 28 h with Gd2FeSbO7 as catalyst was 3.158 μmol h −1 g −1 , and that with Gd2InSbO7 as catalyst was 2.574 μmol h −1 g −1 , and that with Gd2YSbO7 as catalyst was 2.149 μmol h −1 g −1 . Figure 3c shows the photocatalytic H2 evolution from aqueous methanol solution with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst under visible light irradiation (λ > 420 nm, 0.5 g 0.1 wt% Pt-loaded powder sample, 50 mL methanol solution, 200 mL pure water). It could be found from Figure 3c that under visible light irradiation, the rate of H2 evolution in the first 28 h with Gd2FeSbO7 as catalyst was 18.271 μmol h −1 g −1 , and that with Gd2InSbO7 as catalyst was 12.479 μmol h −1 g −1 , and that with Gd2YSbO7 as catalyst was 10.986 μmol h −1 g −1 , indicating that the photocatalytic activity of Gd2FeSbO7 was much higher than that of Gd2InSbO7 or Gd2YSbO7. We would estimate apparent quantum yield in this paper because scattering effects were assumed to be the same for all the photocatalysts and our system was a suspension rather than a homogeneous solution. The apparent quantum yield for hydrogen evolution at 420 nm with Gd2FeSbO7 as catalyst was 0.446%, and that with Gd2InSbO7 as catalyst was 0.305% and that with Gd2YSbO7 as catalyst was 0.268% under visible light irradiation. Moreover, Gd2InSbO7 showed higher photocatalytic activity than Gd2YSbO7. This also proved that the conduction band level of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was more negative than the reduction potential of H2O for forming H2. The formation rate of H2 increased with decreasing the M ionic radii within Gd2MSbO7 (M = Fe, In, Y), Fe 3+ (0.78 Å) < In 3+ (0.92 Å) < Y 3+ (1.019 Å). The reason was that the surface area of the photocatalyst increased with decreasing the M ionic radii, and the creation of more active sites was realized. As a result, the hydrogen generation rate increased. Moreover, the decrease of the M ionic radii would result in a decrease for the migration distance of photogenerated electrons and holes to reach the reaction site on the photocatalyst surface. Thus, the photogenerated electrons and holes could get to the photocatalyst surface more quickly. Above factors would suppress the electron-hole recombination, therefore, the photocatalytic activity would be enhanced. Such results were in good agreement with the optical absorption property of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 (see Figure 2). The rate of H2 evolution also increased with increasing illumination time. The photocatalytic activity of Gd2FeSbO7 increased by about 166% than that of Gd2YSbO7. Figure 4 shows the photocatalytic O2 evolution from AgNO3 solution with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst under visible light irradiation (λ > 420 nm, 0.5 g photocatalyst, 1 mmol AgNO3, 270 mL pure water). It could be seen from Figure 4 that under visible light irradiation, the rate of O2 evolution in the first 28 h with Gd2FeSbO7 as catalyst was 34.329 μmol h −1 g −1 , and that with Gd2InSbO7 as catalyst was 23.264 μmol h −1 g −1 , and that with Gd2YSbO7 as catalyst was 17.200 μmol h −1 g −1 , indicating that the valence band level of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was more positive than the oxidation potential of H2O for forming O2. The formation rate of O2 increased with decreasing the M ionic radii within Gd2MSbO7 (M = Fe, In, Y), Fe 3+ (0.78 Å) < In 3+ (0.92 Å) < Y 3+ (1.019 Å). The apparent quantum yield for the oxygen evolution at 420 nm with Gd2FeSbO7 as catalyst was 1.677%, and that with Gd2InSbO7 as catalyst was 1.136%, and that with Gd2YSbO7 as catalyst is 0.840% under visible light irradiation.  Figure 4. Photocatalytic O2 evolution from AgNO3 solution with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst under visible light irradiation (λ > 420 nm, 0.5 g photocatalyst, 1 mmol AgNO3, 270 mL pure water). Light source: 300 W Xe lamp. Figure 5 shows the photocatalytic H2 evolution from aqueous methanol solution with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst under light irradiation (390 nm cut-off filter, 0.5 g 0.1 wt% Pt-loaded powder sample, 50 mL CH3OH, 200 mL pure water). It was depicted in Figure 5 that under light irradiation (390 nm cut-off filter), the rate of H2 evolution in the first 28 h with Gd2FeSbO7 as catalyst was 49.707 μmol h −1 g −1 , and that with Gd2InSbO7 as catalyst was 35.900 μmol h −1 g −1 , and that with Gd2YSbO7 as catalyst was 29.457 μmol h −1 g −1 , indicating that the effect of wavelength (λ) dependence on the photocatalytic activity was very important. The apparent quantum yield for hydrogen evolution at 390 nm with Gd2FeSbO7 as catalyst was 0.871%, and that with Gd2InSbO7 as catalyst was 0.629% and that with Gd2YSbO7 as catalyst was 0.516% under light irradiation (390 nm cut-off filter).
The photocatalytic H2 evolution from aqueous methanol solution with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst under light irradiation (No cut-off filter, 0.5 g 0.1 wt% Pt-loaded powder sample, 50 mL CH3OH, 200 mL pure water) are shown in Figure 6. It could be found from Figure 6 that under light irradiation without using any filters, the rate of H2 evolution in the first 28 h with Gd2FeSbO7 as catalyst was 94.614 μmol h −1 g −1 , and that with Gd2InSbO7 as catalyst was 70.893 μmol h −1 g −1 , and that with Gd2YSbO7 as catalyst was 57.100 μmol h −1 g −1 , indicating that Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 shows high photocatalytic activity under full arc irradiation. The apparent quantum yield for hydrogen evolution at 420 nm with Gd2FeSbO7 as catalyst was 2.311%, and that with Gd2InSbO7 as catalyst was 1.731%, and that with Gd2YSbO7 as catalyst was 1.394% under light irradiation without using any filters. The photocatalytic activity decreased with increasing incident wavelength λ. As to Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7, the turnover number-the ratio of total amount of gas evolves to catalyst-exceeded 1 for Gd2FeSbO7 after 46 h reaction time, exceeded 1 for Gd2InSbO7 after 57 h reaction time, and exceeded 1 for Gd2YSbO7 after 66 h reaction time under visible light irradiation (λ > 420 nm). Under the condition of full arc irradiation, after 28 h of reaction time, the turnover number exceeded 1.60 for Gd2FeSbO7, and the turnover number exceeded 1.32 for Gd2InSbO7, and the turnover number exceeded 1.02 as to Gd2YSbO7. Above results were enough to prove that the reaction occurred catalytically. The reaction stopped when the light was turned off in this experiment, showing the obvious light response. It was known that TiO2 has very high photocatalytic activity under ultraviolet light irradiation. By contrast, the photocatalytic activity was not obtained with Pt/TiO2 as catalyst under visible light irradiation (λ > 420 nm), while an obvious photocatalytic activity was observed with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst, showing that Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 could respond to visible light irradiation. The formation rate of H2 evolution with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst was much larger than that with TiO2 as catalyst under visible light irradiation. This indicated that the photocatalytic activity of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 for decomposing CH3OH/H2O solution was higher than that of TiO2. The structure of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 after photocatalytic reaction was also checked by using X-ray diffraction method, and no change in their structures was observed during this reaction, which indicated that the H2 evolution was induced from the photocatalytic reaction of H2O. Figure 7 shows the effect of Pt, NiO and RuO2 co-catalysts on the photoactivity of Gd2FeSbO7 under visible light irradiation (λ > 420 nm, 0.5 g powder sample, 50 mL methanol solution, 200 mL pure water). In principle, the photoinduced electrons preferentially enriched on the surface of co-catalyst particles and the recombination of the photoinduced electrons with the photoinduced holes was therefore markedly suppressed. It could be found from Figure 7 that in the first 28 h under visible light irradiation, the rate of H2 evolution was estimated to be 41.471 μmol h −1 g −1 with 0.2 wt%-Pt/Gd2FeSbO7 as catalyst, and that was estimated to be 32.064 μmol h −1 g −1 with 1.0 wt%-NiO/Gd2FeSbO7 as catalyst, and that was estimated to be 24.114 μmol h −1 g −1 with 1.0 wt%-RuO2/Gd2FeSbO7 as catalyst, indicating that the photocatalytic activities could be further improved under visible light irradiation with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 being loaded by Pt, NiO or RuO2. The apparent quantum yield for hydrogen evolution at 420 nm with 0.2 wt%-Pt/Gd2FeSbO7 as catalyst was 1.013%, and that with 1.0 wt%-NiO/Gd2FeSbO7 as catalyst was 0.783%, and that with 1.0 wt%-RuO2/Gd2FeSbO7 as catalyst was 0.589% under visible light irradiation (λ > 420 nm). The effect of Pt was better than that of NiO or RuO2 for improving the photocatalytic activity of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7. It was known that the process for photocatalysis of semiconductors was the direct absorption of photon by band gap of the materials and generated electron-hole pairs in the semiconductor particles, and the excitation of an electron from the valence band to the conduction band was initiated by light absorption with energy equal to or greater than the band gap of the semiconductor. Upon excitation of photon, the separated electron and hole could follow the solid surface. This suggested that the narrow band gap could more easily excite an electron from the valence band to the conduction band. If the conduction band potential level of the semiconductor was more negative than that of H2 evolution, and the valence band potential level was more positive than that of O2 evolution, decomposition of water could occur even without applying electric power [1]. According to the above analysis, the photon absorption of Gd2FeSbO7 was much easier than that of the Gd2InSbO7 or Gd2YSbO7, which resulted in higher photocatalytic activity of Gd2FeSbO7.
The information about the specific surface area, pore volume, average pore size of Gd2FeSbO7, Gd2InSbO7 and Gd2YSbO7 are presented in Table 4, from which we could see that the specific surface area of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was measured to be 4.12 m 2 g −1 , 3.26 m 2 g −1 or 1.28 m 2 g −1 , which was significantly smaller than that of TiO2 photocatalyst (53.8 m 2 g −1 ). Above results indicated much higher potential efficiency of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7. Although the surface area of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was smaller than that of TiO2, but Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 showed higher photocatalytic activity for H2 evolution under visible light irradiation, which indicated that the high photocatalytic activity of the Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was not owing to a big surface area, but rather due to the narrow band gap. It was obvious that further increase in photocatalytic activity might be prospected from increasing the surface area of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7. Since an efficient photocatalytic reaction process occurred on the photocatalyst surface, the increase of the surface area for the photocatalysts might lead to the increase of their photocatalytic activity.

Conclusions
In the present work, we prepared single phase of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 by solid-state reaction method and studied the structural, optical and photocatalytic properties of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7. Rietveld structure refinement results revealed that Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 crystallized with the pyrochlore-type structure, cubic crystal system and space group Fd3m. The lattice parameter a for Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was 10.276026 Å, 10.449546 Å or 10.653651 Å. The band gap of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 was estimated to be 2.151 eV, 2.897 eV or 2.396 eV. Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 showed optical absorption in the visible light region, indicating that above photocatalysts had the ability to respond to the wavelength of visible light region. For the photocatalytic water-splitting reaction, H2 or O2 evolution was observed from pure water with Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst under visible light irradiation (λ > 420 nm). In addition, under visible light irradiation (λ > 420 nm), H2 or O2 was also produced by using Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 as catalyst from CH3OH/H2O or AgNO3/H2O solutions. Gd2FeSbO7 showed the highest activity compared with Gd2InSbO7 or Gd2YSbO7. At the same time, Gd2InSbO7 showed higher activity compared with Gd2YSbO7. The photocatalytic activities were further improved under visible light irradiation with Gd2FeSbO7 being loaded by Pt, NiO or RuO2. The effect of Pt was better than that of NiO or RuO2 for improving the photocatalytic activity of Gd2FeSbO7. Moreover, the synthesis of Gd2FeSbO7, Gd2InSbO7 or Gd2YSbO7 offered some useful insights for the design of new photocatalysts for the photocatalytic evolution of H2 or O2.