Enhanced Photocatalytic Hydrogen Production of the Polyoxoniobate Modified with RGO and PPy

The development of high-efficiency, recyclable, and inexpensive photocatalysts for water splitting for hydrogen production is of great significance to the application of solar energy. Herein, a series of graphene-decorated polyoxoniobate photocatalysts Nb6/PPy-RGO (Nb6 = K7HNb6O19, RGO = reduced graphene oxide, PPy = polypyrrole), with the bridging effect of polypyrrole were prepared through a simple one-step solvothermal method, which is the first example of polyoxoniobate-graphene-based nanocomposites. The as-fabricated photocatalyst showed a photocatalytic H2 evolution activity without any co-catalyst. The rate of 1038 µmol g−1 in 5 h under optimal condition is almost 43 times higher than that of pure K7HNb6O19·13H2O. The influencing factors for photocatalysts in photocatalytic hydrogen production under simulated sunlight were studied in detail and the feasible mechanism is presented in this paper. These results demonstrate that Nb6O19 acts as the main catalyst and electron donor, RGO provides active sites, and PPy acted as an electronic bridge to extend the lifetime of photo-generated carriers, which are crucial factors for photocatalytic H2 production.


Introduction
With the widespread energy consumption and environmental problems, it is necessary and urgent to search for sources of green and clean energy to replace traditional fossil fuels [1][2][3][4]. Since the first discovery of photoelectrochemical water splitting into H 2 via the semiconductor TiO 2 by Fujishima and Honda in 1972, photocatalytic water splitting to hydrogen production has long been viewed as a promising and attractive strategy [5]. In recent years, many semiconductor materials have been reported for photocatalytic hydrogen evolution, in which the band gap, the position of the conduction band, the separation of electrons and holes and the lifetime of carriers are the key factors affecting photocatalysis [6,7]. Nonetheless, the exploitation of highly effective semiconductor materials for photocatalytic hydrogen evolution is still in the infant stage [8][9][10].

Characterization
The morphology and microstructure of the synthetic samples were characterized using a field emission scanning electron microscopy (FESEM, JSM-7610F, Electronics Co., LTD, Tokyo, Japan) at 10 kV and a transmission electron microscopy (TEM, JEM-2100F, Electronics Co., LTD, Tokyo, Japan) at 200 kV). Powder X-ray diffraction (XRD, Bruker optics Instruments company, Karlsruhe, Germany) was measured from 5° to 80° at room temperature on a Bruker D8 Advance diffractometer with Cu-Kα radiation and the Fourier transform-infrared spectra (FT-IR, KBr pellets, Bruker optics Instruments company, Karlsruhe, Germany) were recorded on a Nicolet 170 SXFT-IR spectrophotometer ranging from 4000 to 500 cm −1 . X-ray photoelectron spectroscopy (XPS, Thermo Fisher Scientific, Waltham, MA, USA) was recorded on an ESCALAB 250Xi X-ray photoelectron spectrometer with an Al-Kα (hv = 1486.6 eV) monochromatic radiation source. The binding energy peak position of each sample was calibrated by the C 1s peak at 284.8 eV. UV-visible spectra were achieved on a Shimadzu UV-2600 UV/Vis spectrophotometer (Shimadzu Instrument Co., LTD, Kyoto-fu, Japan) using BaSO4 as a reference. Raman spectroscopy (Renishaw company, London, UK) was prepared using a Renishaw in Via Raman spectrometer with excitation by 325 nm Photoluminescence (PL) spectra were obtained at excitation wavelength 385 nm with TU-1900 (PuXi Company, Beijing, China).

Photocatalytic Hydrogen Production
Photocatalytic hydrogen producing experiments were performed using a method similar to the literature [6]. Photocatalyst (50 mg) was added to a solution of deionized water (40.0 mL) and CH3OH (10.0 mL, pH = 7) in a quartz vessel. The reaction system was irradiated by a 300 W Xe lamp (Perfect light PLS-SXE300, Perfectlight Technology Co., Ltd, Beijing, China), and the temperature was maintained at about 5 °C. The product H2 was measured by gas chromatograph (FL-9790) on line. After each photocatalytic test, photocatalysts were collected by centrifugation without reactivated before each cycle, washed with H2O and C2H5OH to remove impurities, then used after drying.

Characterization
The morphology and microstructure of the synthetic samples were characterized using a field emission scanning electron microscopy (FESEM, JSM-7610F, Electronics Co., LTD., Tokyo, Japan) at 10 kV and a transmission electron microscopy (TEM, JEM-2100F, Electronics Co., LTD., Tokyo, Japan) at 200 kV). Powder X-ray diffraction (XRD, Bruker optics Instruments company, Karlsruhe, Germany) was measured from 5 • to 80 • at room temperature on a Bruker D8 Advance diffractometer with Cu-Kα radiation and the Fourier transform-infrared spectra (FT-IR, KBr pellets, Bruker optics Instruments company, Karlsruhe, Germany) were recorded on a Nicolet 170 SXFT-IR spectrophotometer ranging from 4000 to 500 cm −1 . X-ray photoelectron spectroscopy (XPS, Thermo Fisher Scientific, Waltham, MA, USA) was recorded on an ESCALAB 250Xi X-ray photoelectron spectrometer with an Al-Kα (hv = 1486.6 eV) monochromatic radiation source. The binding energy peak position of each sample was calibrated by the C 1s peak at 284.8 eV. UV-visible spectra were achieved on a Shimadzu UV-2600 UV/Vis spectrophotometer (Shimadzu Instrument Co., LTD, Kyoto-fu, Japan) using BaSO 4 as a reference. Raman spectroscopy (Renishaw company, London, UK) was prepared using a Renishaw in Via Raman spectrometer with excitation by 325 nm Photoluminescence (PL) spectra were obtained at excitation wavelength 385 nm with TU-1900 (PuXi Company, Beijing, China).

Photocatalytic Hydrogen Production
Photocatalytic hydrogen producing experiments were performed using a method similar to the literature [6]. Photocatalyst (50 mg) was added to a solution of deionized water (40.0 mL) and CH 3 OH (10.0 mL, pH = 7) in a quartz vessel. The reaction system was irradiated by a 300 W Xe lamp (Perfect light PLS-SXE300, Perfectlight Technology Co., Ltd, Beijing, China), and the temperature was maintained at about 5 • C. The product H 2 was measured by gas chromatograph (FL-9790) on line. After each photocatalytic test, photocatalysts were collected by centrifugation without reactivated before each cycle, washed with H 2 O and C 2 H 5 OH to remove impurities, then used after drying.

Photocatalytic Measurement
The Mott-Schottky curves were tested on an AMETEK Princeton Applied Research (Versa STAT 4, Princeton, NJ, USA) electrochemical workstation (FTO = fluorine-doped tin oxide substrate, 1 cm × 1.5 cm). The electrochemical impedance spectroscopy (EIS) and photocurrent response were recorded on an electrochemical workstation (CHI660, Chenhua, Shanghai, China) equipped with a three-electrode system with a complex/FTO as the working electrode, platinum foil as the counter electrode and Ag/AgCl as the reference electrode in 0.2 M sodium sulfate solution (Na 2 SO 4 ). The working electrode complex/FTO was prepared by dropwise adding 50 µL of sample suspensions containing Nb 6 /PPy-RGO composites (3.0 mg), ethanol (1.0 mL), and nafion (20 µL) onto a FTO substrate (1 cm × 1.5 cm).

Structural Characterizations
The XRD patterns of GO, Nb 6 , and Nb 6 /PPy-RGO composites were investigated. As shown in Figure 1 and Figure S1, pristine GO and RGO show a single peak near 12 • and 26 • , respectively, while pristine K 7 HNb 6 O 19 has sharp diffraction peaks at 9.7 • , 26 • and 48 • , respectively, which is consistent with the literature reports [46][47][48]. The broad peaks centered on 27.5 • of Nb 6 /PPy-RGO composites can be attributed to the combined action of PPy and RGO, while the characteristic peaks at 46 • are attributed to the superposition of Nb 6 and PPy [49][50][51]. In Nb 6 /PPy-RGO composite materials, although it is not obvious in Nb 6 /PPy-RGO-0.325 and Nb 6 /PPy-RGO-0.25, the characteristic peaks of raw materials can still be found. No additional diffraction peaks were observed, confirming the formation of the Nb 6 /PPy-RGO composites. The crystallite sizes of the catalysts calculated from the XRD patterns were shown in Table 1.

Structural Characterizations
The XRD patterns of GO, Nb6, and Nb6/PPy-RGO composites were investigated. As shown in Figure 1 and Figure S1, pristine GO and RGO show a single peak near 12° and 26°, respectively, while pristine K7HNb6O19 has sharp diffraction peaks at 9.7°, 26° and 48°, respectively, which is consistent with the literature reports [46][47][48]. The broad peaks centered on 27.5° of Nb6/PPy-RGO composites can be attributed to the combined action of PPy and RGO, while the characteristic peaks at 46° are attributed to the superposition of Nb6 and PPy [49][50][51]. In Nb6/PPy-RGO composite materials, although it is not obvious in Nb6/PPy-RGO-0.325 and Nb6/PPy-RGO-0.25, the characteristic peaks of raw materials can still be found. No additional diffraction peaks were observed, confirming the formation of the Nb6/PPy-RGO composites. The crystallite sizes of the catalysts calculated from the XRD patterns were shown in Table 1.  The structure of Nb6/PPy-RGO-0.25 composite has also been evidenced by Fourier transforminfrared (FT-IR) spectroscopy ( Figure 2). The peak at 1398 cm −1 is assigned to the C-O stretching vibration due to π-π interaction between PPy and RGO. The peak at 1637 cm −1 is assigned to H-O-H stretching vibration and the peak at 1258 cm −1 was assigned to C-H and C-N in-plane deformation vibration, further implying the combination of GO with PPy [52,53]. The characteristic peaks at 840 and 887 cm −1 could be attributed to the terminal Nb-O stretching vibration, and the peaks appearing at 532 and 597 cm −1 correspond to Nb-O-Nb bonds [54]. The result shows that Nb6 was successfully inserted into the final composite.  The structure of Nb 6 /PPy-RGO-0.25 composite has also been evidenced by Fourier transform-infrared (FT-IR) spectroscopy ( Figure 2). The peak at 1398 cm −1 is assigned to the C-O stretching vibration due to π-π interaction between PPy and RGO. The peak at 1637 cm −1 is assigned to H-O-H stretching vibration and the peak at 1258 cm −1 was assigned to C-H and C-N in-plane deformation vibration, further implying the combination of GO with PPy [52,53]. The characteristic peaks at 840 and 887 cm −1 could be attributed to the terminal Nb-O stretching vibration, and the peaks appearing at 532 and 597 cm −1 correspond to Nb-O-Nb bonds [54]. The result shows that Nb 6 was successfully inserted into the final composite. From the Raman spectra shown in Figure 3, the higher wavenumber bands at 1334 and 1584 cm −1 are ascribed to D band and G band of GO. It is worth noting that the peak intensity ratio (ID/IG) is a popular method to evaluate the disorder and reduction of graphene materials. The ID/IG ratio of the pure GO is about 0.76, while that of the Nb6/PPy-RGO-0.25 composite is about 0.99. The numeric addition indicated that the oxygen-containing functional groups were partially reduced, meaning the transition of GO to RGO in the process of solvothermal reaction [54]. Compared with the characteristic peaks of Nb6 (541, 827 and 875 cm −1 ) [55,56], the redshift could be found for Nb6/PPy-RGO-0.25 (878 and 620 cm −1 ), which may be assigned to the strong interaction among the Nb6, PPy, and RGO of the composites. In order to further identify the structure of composites, the XPS spectra were tested for Nb6/PPy-RGO-0.25 composites and GO, as well as Nb6 ( Figure 4). As shown in Figure 4A, the full spectrum clearly shows the presence of C, O, N, K, and Nb elements in the composite sample, which is consistent with the chemical composition of the photocatalyst. As shown in Figure 4B, the binding energies of C 1s, including 284.05, 285.60, 286.47 and 288.37 eV, could be attributed to carbon of the From the Raman spectra shown in Figure 3, the higher wavenumber bands at 1334 and 1584 cm −1 are ascribed to D band and G band of GO. It is worth noting that the peak intensity ratio (I D /I G ) is a popular method to evaluate the disorder and reduction of graphene materials. The I D /I G ratio of the pure GO is about 0.76, while that of the Nb 6 /PPy-RGO-0.25 composite is about 0.99. The numeric addition indicated that the oxygen-containing functional groups were partially reduced, meaning the transition of GO to RGO in the process of solvothermal reaction [54]. Compared with the characteristic peaks of Nb 6 (541, 827 and 875 cm −1 ) [55,56], the redshift could be found for Nb 6 /PPy-RGO-0.25 (878 and 620 cm −1 ), which may be assigned to the strong interaction among the Nb 6 , PPy, and RGO of the composites. From the Raman spectra shown in Figure 3, the higher wavenumber bands at 1334 and 1584 cm −1 are ascribed to D band and G band of GO. It is worth noting that the peak intensity ratio (ID/IG) is a popular method to evaluate the disorder and reduction of graphene materials. The ID/IG ratio of the pure GO is about 0.76, while that of the Nb6/PPy-RGO-0.25 composite is about 0.99. The numeric addition indicated that the oxygen-containing functional groups were partially reduced, meaning the transition of GO to RGO in the process of solvothermal reaction [54]. Compared with the characteristic peaks of Nb6 (541, 827 and 875 cm −1 ) [55,56], the redshift could be found for Nb6/PPy-RGO-0.25 (878 and 620 cm −1 ), which may be assigned to the strong interaction among the Nb6, PPy, and RGO of the composites. In order to further identify the structure of composites, the XPS spectra were tested for Nb6/PPy-RGO-0.25 composites and GO, as well as Nb6 ( Figure 4). As shown in Figure 4A, the full spectrum clearly shows the presence of C, O, N, K, and Nb elements in the composite sample, which is In order to further identify the structure of composites, the XPS spectra were tested for Nb 6 /PPy-RGO-0.25 composites and GO, as well as Nb 6 ( Figure 4). As shown in Figure 4A, the full spectrum clearly shows the presence of C, O, N, K, and Nb elements in the composite sample, which is consistent with the chemical composition of the photocatalyst. As shown in Figure 4B, the binding energies of C 1s, including 284.05, 285.60, 286.47 and 288.37 eV, could be attributed to carbon of the non-oxygenated ring, C-OH, C-O and C=O in Nb 6 /PPy-RGO-0.25, respectively, which are lower and weaker than that of pure GO (284.74, 285.97, 287.07 and 289.08 eV). These results demonstrate that the GO has been successfully reduced in composites, which is in agreement with Raman spectra [57][58][59][60].  Figure 4D illustrates a comprehensive Nb 3d XPS analysis of the Nb 6 /PPy-RGO-0.25 and Nb 6 . The binding energies of 206.49 and 209.26 eV could be assigned to Nb 3d 5/2 and Nb 3d 3/2 of Nb 5+ (highest oxidation state) in Nb 6 /PPy-RGO-0.25. Compared with Nb 6 , the peaks of Nb 3d shifted to higher binding energy by~1 eV, which may be caused by electron transfer from Nb 6 to RGO in composites [54]. As shown in Figure 4E, the N 1s peaks centered at 398.01 and 399.55 eV are vested to the pyridinic and pyrrolic of N atom in polymer. In addition, there is a weak peak at the center of 394.11 eV, corresponding to the Nb-N bond, indicating the formation of chemical bonds between PPy and Nb 6 [62].  Figure 4C) [61]. Figure 4D illustrates a comprehensive Nb 3d XPS analysis of the Nb6/PPy-RGO-0. 25 [54]. As shown in Figure 4E, the N 1s peaks centered at 398.01 and 399.55 eV are vested to the pyridinic and pyrrolic of N atom in polymer. In addition, there is a weak peak at the center of 394.11 eV, corresponding to the Nb-N bond, indicating the formation of chemical bonds between PPy and Nb6 [62].

Nanosphere Morphologies
As shown in Figure 5, the results show that the morphology of Nb6 is rodlike. However, when fabricated with PPy and RGO, Nb6/PPy-RGO-X (X = 0.125, 0.25, 0.375) display the flower-like

Nanosphere Morphologies
As shown in Figure 5, the results show that the morphology of Nb 6 is rodlike. However, when fabricated with PPy and RGO, Nb 6 /PPy-RGO-X (X = 0.125, 0.25, 0.375) display the flower-like methodology. To deeply understand the formation of flower-like methodology, the controlled experiments including solvothermal reaction temperatures and the amount of Py have been respectively conducted using Nb 6 /PPy-RGO-0.25 as a representative. As shown in Figures S2 and S3, the samples probably agglomerated when changing the conditions of 160 • C or Py (130 µL). However, the lower amount of Py, with 90 µL, or higher temperature of 190 • C may reduce the surface area and could not form morphology of flowers for Nb 6 /PPy-RGO-0.25. These results demonstrate that the flower-like morphology for Nb 6 /PPy-RGO-0.25 largely depends on the temperature and the amounts of Py in the reaction. The corresponding EDS elemental mapping of Nb 6 /PPy-RGO-0.25 indicated the elements of C, N, O, K and Nb were uniformly distributed throughout the whole composite ( Figure 5E).  Figure  5E).

Photochemistry and Electrochemistry
The UV-Vis diffuse reflectance spectra of Nb6/PPy-RGO composites were recorded at room temperature to discuss their optical properties, together with Nb6 for comparison. As shown in Figure  6, Nb6/PPy-RGO composites exhibit broad absorptions in the range of 200-340 nm. Compared with the absorption of Nb6 (200-315 nm), the light absorption of Nb6/PPy-RGO samples was obviously widened, which is related to the introduction of PPy and RGO. Due to the zero band gap for RGO, with the increasing amounts of Nb6, the absorbance intensity for Nb6/PPy-RGO composites become weaker (Figure 6a). Simultaneously, the color becomes lighter and the ability to capture light is decreased ( Figure S4), hinting that the enhanced absorption is mainly attributed to the introduction of RGO. The band-gap values of Nb6 and Nb6/PPy-RGO composites (from 0.125 to 0.375) obtained by the Kubelka-Munk function via the UV-Vis diffuse reflectance are 4.14, 3.66, 3.71, and 3.76 eV, respectively, indicating the semiconductor characters of Nb6/PPy-RGO composites (Figure 6b). Compared to the band gap value of Nb6, those of Nb6/PPy-RGO composites are obviously narrowed through the incorporation of RGO and PPy.
Furthermore, the potentials of conduction band (CB) of Nb6/PPy-RGO composites could be obtained from electrochemical Mott-Schottky plots ( Figure S5). The plots of C −2 vs. potential exhibited positive slopes, indicating that we should consider the composites as typical n-type semiconductors [57]. As shown in Figure S6, compared with other composites, the Nb6/PPy-RGO-0.25 composite displays the weakest PL intensity at 378 nm, and the amount of pyrrole and the solvothermal

Photochemistry and Electrochemistry
The UV-Vis diffuse reflectance spectra of Nb 6 /PPy-RGO composites were recorded at room temperature to discuss their optical properties, together with Nb 6 for comparison. As shown in Figure 6, Nb 6 /PPy-RGO composites exhibit broad absorptions in the range of 200-340 nm. Compared with the absorption of Nb 6 (200-315 nm), the light absorption of Nb 6 /PPy-RGO samples was obviously widened, which is related to the introduction of PPy and RGO. Due to the zero band gap for RGO, with the increasing amounts of Nb 6 , the absorbance intensity for Nb 6 /PPy-RGO composites become weaker (Figure 6a). Simultaneously, the color becomes lighter and the ability to capture light is decreased ( Figure S4), hinting that the enhanced absorption is mainly attributed to the introduction of RGO. The band-gap values of Nb 6 and Nb 6 /PPy-RGO composites (from 0.125 to 0.375) obtained by the Kubelka-Munk function via the UV-Vis diffuse reflectance are 4.14, 3.66, 3.71, and 3.76 eV, respectively, indicating the semiconductor characters of Nb 6 /PPy-RGO composites (Figure 6b). Compared to the band gap value of Nb 6 , those of Nb 6 /PPy-RGO composites are obviously narrowed through the incorporation of RGO and PPy.
Furthermore, the potentials of conduction band (CB) of Nb 6 /PPy-RGO composites could be obtained from electrochemical Mott-Schottky plots ( Figure S5). The plots of C −2 vs. potential exhibited positive slopes, indicating that we should consider the composites as typical n-type semiconductors [57].  [63], indicating the title composites should be used as photocatalysts for hydrogen evolution via water splitting.
As shown in Figure S6, compared with other composites, the Nb 6 /PPy-RGO-0.25 composite displays the weakest PL intensity at 378 nm, and the amount of pyrrole and the solvothermal temperature had little effect on the separation of photo-excited charge and holes, indicative of the lowest combination of electron-hole pairs. Because K 7 HNb 6 O 19 has good water solubility, electrochemical measurement cannot be performed under the same conditions. As shown in Figure 7a, the highest photocurrent response of Nb 6 /PPy-RGO-0.25 shows better separation ability and lower recombination of photogenerated electron-hole pairs [42]. As depicted in Figure 7b, the smallest radius of Nb 6 /PPy-RGO-0.25 indicates lower charge transfer resistance. This result confirms that more photo-generated charge carriers can be separated by constructing Nb 6 /PPy-RGO-0.25, which is consistent with the result of PL analysis and photocurrent response. lowest combination of electron-hole pairs. Because K7HNb6O19 has good water solubility, electrochemical measurement cannot be performed under the same conditions. As shown in Figure  7a, the highest photocurrent response of Nb6/PPy-RGO-0.25 shows better separation ability and lower recombination of photogenerated electron-hole pairs [42]. As depicted in Figure 7b, the smallest radius of Nb6/PPy-RGO-0.25 indicates lower charge transfer resistance. This result confirms that more photo-generated charge carriers can be separated by constructing Nb6/PPy-RGO-0.25, which is consistent with the result of PL analysis and photocurrent response.

Photocatalytic Hydrogen Evolution
Photocatalytic activity of Nb6/PPy-RGO composites for hydrogen evolution via water splitting was evaluated in aqueous solution with methanol (CH3OH) as a sacrificial electron donor. As shown in Figure 8, the H2 production for Nb6/PPy-RGO composites was 351.5 (Nb6/PPy-RGO-0.125), 1038 (Nb6/PPy-RGO-0.25) and 690.5 µmol g −1 (Nb6/PPy-RGO-0.375) in 5 h, respectively. Under the same conditions, the H2 production was 25 µmol g −1 for the Nb6 catalyst. Apparently, the composites exhibit lowest combination of electron-hole pairs. Because K7HNb6O19 has good water solubility, electrochemical measurement cannot be performed under the same conditions. As shown in Figure  7a, the highest photocurrent response of Nb6/PPy-RGO-0.25 shows better separation ability and lower recombination of photogenerated electron-hole pairs [42]. As depicted in Figure 7b, the smallest radius of Nb6/PPy-RGO-0.25 indicates lower charge transfer resistance. This result confirms that more photo-generated charge carriers can be separated by constructing Nb6/PPy-RGO-0.25, which is consistent with the result of PL analysis and photocurrent response.

Photocatalytic Hydrogen Evolution
Photocatalytic activity of Nb6/PPy-RGO composites for hydrogen evolution via water splitting was evaluated in aqueous solution with methanol (CH3OH) as a sacrificial electron donor. As shown in Figure 8
Turnover frequency (TOF) was based on Nb 6 , which has been widely used to evaluate the property of catalysts, especially for POM-based catalysts. The details are as follows: Taking Nb 6 /PPy-RGO-0.25 as an example: Turnover frequency (TOF) was based on Nb6, which has been widely used to evaluate the property of catalysts, especially for POM-based catalysts. The details are as follows:   The photocatalytic system has also been optimized by changing the amount or species of sacrificial electron donors by using the optimal as Nb6/PPy-RGO-0.25 photocatalyst. As shown in Figure S7a, when the concentration of CH3OH increased from 0 to 20% (v:v), the hydrogen evolution increased from 0 to 1038 µmol g −1 , but decreased to 274 and 94 µmol g −1 with higher concentrations of 25% (v:v) and 30% (v:v), respectively. These results imply that the amount of CH3OH played an important role in photocatalytic hydrogen evolution. The more sacrificial electron donors may react with holes of photocatalyst to release more photo-excited electrons to produce more hydrogen. However, with an excess of CH3OH, it should also decrease the concentrations of substrate, leading to lower photocatalytic efficiency. Furthermore, the species of sacrificial electron donor was optimized by controlling the ratio of sacrificial agent/H2O (v:v = 10:40) ( Figure S7b). The title photocatalyst exhibited the optimal H2 production of 1038 µmol g −1 with CH3OH as the sacrificial The photocatalytic system has also been optimized by changing the amount or species of sacrificial electron donors by using the optimal as Nb 6 /PPy-RGO-0.25 photocatalyst. As shown in Figure S7a, when the concentration of CH 3 OH increased from 0 to 20% (v:v), the hydrogen evolution increased from 0 to 1038 µmol g −1 , but decreased to 274 and 94 µmol g −1 with higher concentrations of 25% (v:v) and 30% (v:v), respectively. These results imply that the amount of CH 3 OH played an important role in photocatalytic hydrogen evolution. The more sacrificial electron donors may react with holes of photocatalyst to release more photo-excited electrons to produce more hydrogen. However, with an excess of CH 3 OH, it should also decrease the concentrations of substrate, leading to lower photocatalytic efficiency. Furthermore, the species of sacrificial electron donor was optimized by controlling the ratio of sacrificial agent/H 2 O (v:v = 10:40) ( Figure S7b). The title photocatalyst exhibited the optimal H 2 production of 1038 µmol g −1 with CH 3 OH as the sacrificial agent, but varied in the range of 249.5-834 µmol g −1 with other sacrificial agents. As shown in Figure S8, the best sample, Nb 6 /PPy-RGO-0.25, exhibits the highest photocatalytic H 2 production, indicating the optimal reaction conditions with pyrrole dosage of 110 µL and reaction temperature of 180 • C.
In order to prove the reusability of Nb 6 /PPy-RGO-0.25, the recycling experiment was carried out (Figure 9). There was no noticeable deactivation observed for the catalyst and the yields of H 2 evolution for the four recycling experiments all remained at 1000 µmol g −1 in 5 h. In addition, the XRD pattern ( Figure S9), Raman spectra ( Figure S10) and SEM image ( Figure S11) of the recovered catalysts were alike to those of the as-prepared sample, further demonstrating the recyclability and stability of Nb 6 /PPy-RGO-0.25 in the photocatalysis reaction. In order to prove the reusability of Nb6/PPy-RGO-0.25, the recycling experiment was carried out (Figure 9). There was no noticeable deactivation observed for the catalyst and the yields of H2 evolution for the four recycling experiments all remained at 1000 µmol g −1 in 5 h. In addition, the XRD pattern ( Figure S9), Raman spectra ( Figure S10) and SEM image ( Figure S11) of the recovered catalysts were alike to those of the as-prepared sample, further demonstrating the recyclability and stability of Nb6/PPy-RGO-0.25 in the photocatalysis reaction.

Proposed Photocatalytic Mechanism
According to the Mott-Schottky plot shown in Figure S5, the conduction band minimum (CBM) of the composites could be estimated to explain the enhanced photocatalytic activity mechanism, which is about −1.34 eV for the sample Nb6/PPy-RGO-0.25 possessing the best H2 evolution performance. Meanwhile, valence band maximum (VBM) potential of Nb6/PPy-RGO-0.25 can be determined based on the formula EVBM = ECBM + Eg, which is about 2.37 eV [64]. As displayed in Figure  10, the pure Nb6 exhibits negligible photocatalytic activity, which may be due to the fast recombination of electrons and holes. Higher photocatalytic efficiency can be achieved for Nb6/PPy-RGO composites by the modification with PPy and RGO. It has been reported that conductive polymers can be used as photocatalysts for H2 production from water splitting [65]. However, in such Nb6/PPy-RGO-0.25 composites, pure PPy and RGO cannot exhibit photocatalytic activity due to their combination of electrons and holes. On the other hand, they can prolong the lifetime of charge carriers to promote the separation efficiency of electron-hole pairs in Nb6/PPy-RGO-0.25 composites due to their excellent π conjugated chain structure and electrical conductivity. In addition, RGO can also serve as active sites to produce H2, avoiding the use of noble metals, while PPy acted as conductive polymers, promoting the transfer of the electrons [52,53,[66][67][68][69]. In particular, the π-π interaction between RGO and PPy further strengthens the electrical conductivity. Based on the above results, a probable photocatalytic mechanism for composites could be proposed as follows: under light irradiation, the electrons in the valence band (VB) of Nb6/PPy-RGO-0.25 can be excited to the CB, causing holes in the VB. On one hand, the holes in the VB could be quenched by accepting electrons

Proposed Photocatalytic Mechanism
According to the Mott-Schottky plot shown in Figure S5, the conduction band minimum (CBM) of the composites could be estimated to explain the enhanced photocatalytic activity mechanism, which is about −1.34 eV for the sample Nb 6 /PPy-RGO-0.25 possessing the best H 2 evolution performance. Meanwhile, valence band maximum (VBM) potential of Nb 6 /PPy-RGO-0.25 can be determined based on the formula E VBM = E CBM + E g , which is about 2.37 eV [64]. As displayed in Figure 10, the pure Nb 6 exhibits negligible photocatalytic activity, which may be due to the fast recombination of electrons and holes. Higher photocatalytic efficiency can be achieved for Nb 6 /PPy-RGO composites by the modification with PPy and RGO. It has been reported that conductive polymers can be used as photocatalysts for H 2 production from water splitting [65]. However, in such Nb 6 /PPy-RGO-0.25 composites, pure PPy and RGO cannot exhibit photocatalytic activity due to their combination of electrons and holes. On the other hand, they can prolong the lifetime of charge carriers to promote the separation efficiency of electron-hole pairs in Nb 6 /PPy-RGO-0.25 composites due to their excellent π conjugated chain structure and electrical conductivity. In addition, RGO can also serve as active sites to produce H 2 , avoiding the use of noble metals, while PPy acted as conductive polymers, promoting the transfer of the electrons [52,53,[66][67][68][69]. In particular, the π-π interaction between RGO and PPy further strengthens the electrical conductivity. Based on the above results, a probable photocatalytic mechanism for composites could be proposed as follows: under light irradiation, the electrons in the valence band (VB) of Nb 6 /PPy-RGO-0.25 can be excited to the CB, causing holes in the VB. On one hand, the holes in the VB could be quenched by accepting electrons from CH 3 OH. On the other hand, the photoelectron in the CB of Nb 6 /PPy-RGO-0.25 could be transferred to RGO through the bridge of PPy. Acting as photoactive sites, the RGO accepting electrons would react with H 2 O to produce H 2 .

Conclusions
In summary, a series of unprecedented polyoxoniobate-graphene nanocomposites, Nb6/PPy-RGO, have been successfully synthesized through a simple one-step solvothermal method. This is the first example of polyoxoniobate-graphene nanocomposites for photocatalytic H2 production. This work presents a new idea for the design and synthesis of high-performance polyoxometalatesgraphene photocatalysts in photocatalytic H2 production. Nb6/PPy-RGO was optimized by adjusting the amount of Nb6, volume of pyrrole and temperature, displaying superb photocatalytic H2 production of 1038 µmol g −1 in 5 h without a co-catalyst, almost 43 times more than pure Nb6, in which the collaborations among Nb6, PPy and RGO play vital roles in such a photocatalytic system. In further work, we will focus on studying more fascinating polyoxoniobate-graphene nanocomposites with new interesting properties.

Conclusions
In summary, a series of unprecedented polyoxoniobate-graphene nanocomposites, Nb 6 /PPy-RGO, have been successfully synthesized through a simple one-step solvothermal method. This is the first example of polyoxoniobate-graphene nanocomposites for photocatalytic H 2 production. This work presents a new idea for the design and synthesis of high-performance polyoxometalates-graphene photocatalysts in photocatalytic H 2 production. Nb 6 /PPy-RGO was optimized by adjusting the amount of Nb 6 , volume of pyrrole and temperature, displaying superb photocatalytic H 2 production of 1038 µmol g −1 in 5 h without a co-catalyst, almost 43 times more than pure Nb 6 , in which the collaborations among Nb 6 , PPy and RGO play vital roles in such a photocatalytic system. In further work, we will focus on studying more fascinating polyoxoniobate-graphene nanocomposites with new interesting properties.
Supplementary Materials: The following are available online at http://www.mdpi.com/2079-4991/10/12/2449/ s1, Figure S1: X-ray diffraction (XRD) patterns for series of concentration of pyrrole and temperature and the corresponding starting materials. Figure Figure S4: The color change of samples. Figure S5: Mott-Schottky plots of (a) Nb 6 /PPy-RGO-0.125; (b) Nb 6 /PPy-RGO-0.25; (c) Nb 6 /PPy-RGO-0.375, in 0.2 M Na 2 SO 4 aqueous solution with pH = 7. Figure S6: Photoluminescence (PL) spectra recorded at room temperature in the range of 325-450 nm with an excitation wavelength of 378 nm for adjust (a) molar ratio of Nb 6 ; (b) volume of pyrrole; (c) temperature. Figure S7: (a) Rate of H 2 evolution as a function of CH 3 OH concentration, (b) Rate of H 2 evolution as a function of type of sacrificial agents. Figure S8: Photocatalytic hydrogen production in concentration of pyrrole (a) and temperature (b) in aqueous solution with MeOH 20%. Figure S9: XRD patterns of Nb 6 /PPy-RGO-0.25 before and after photocatalytic H 2 evolution reaction. Figure S10: Raman spectra of Nb 6 /PPy-RGO-0.25 before and after photocatalytic H 2 evolution reaction. Figure