Synthesis and Characterization of WO3/Graphene Nanocomposites for Enhanced Photocatalytic Activities by One-Step In-Situ Hydrothermal Reaction

Tungsten trioxide (WO3) nanorods are synthesized on the surface of graphene (GR) sheets by using a one-step in-situ hydrothermal method employing sodium tungstate (Na2WO4·2H2O) and graphene oxide (GO) as precursors. The resulting WO3/GR nanocomposites are characterized by X-ray diffraction, Raman spectroscopy, transmission electron microscopy, scanning electron microscopy and X-ray photoelectron spectroscopy. The results confirm that the interface between WO3 nanorod and graphene contains chemical bonds. The enhanced optical absorption properties are measured by UV-vis diffuse reflectance spectra. The photocatalytic activity of the WO3/GR nanocomposites under visible light is evaluated by the photodegradation of methylene blue, where the degradation rate of WO3/GR nanocomposites is shown to be double that of pure WO3. This is attributed to the synergistic effect of graphene and the WO3 nanorod, which greatly enhances the photocatalytic performance of the prepared sample, reduces the recombination of the photogenerated electron-hole pairs and increases the visible light absorption efficiency. Finally, the photocatalytic mechanism of the WO3/GR nanocomposites is presented. The synthesis of the prepared sample is convenient, direct and environmentally friendly. The study reports a highly efficient composite photocatalyst for the degradation of contaminants that can be applied to cleaning up the environment.


Introduction
The use of semiconductor materials for the photocatalytic decomposition of organic pollutants in the gas or liquid states has aroused much interest from researchers [1][2][3]. Metal oxides, such as tungsten trioxide (WO 3 ) and titanium dioxide (TiO 2 ), and graphene are the most promising photocatalysts [4][5][6] owing to their high chemical stability, non-toxicity, and low cost. WO 3 , with a low band gap, is an important n-type semiconductor [7,8]. It has been widely used in gas sensing [9][10][11], lithium-ion batteries [12,13], smart windows [14] and photocatalysis [15,16]. WO 3 exhibits both a high response and a high photocatalytic efficiency upon visible light irradiation [17]. However, pure WO 3 , as a photocatalyst, is not always beneficial because of the limitation of the electron-hole recombination rate and oxygen storage issues. Therefore, the following methods have previously been presented [18][19][20][21][22][23][24][25]: (1) changing the morphology of WO 3 and the particle size; (2) loading a noble metal, doping metal or non-metal in WO 3 ; and (3) combining with semiconductors. Liu et al. [26] prepared heterojunction of WO 3 /g-C 3 N 4 with well-defined morphology by in-situ liquid phase process, and start materials metal-free g-C 3 N 4 was synthesized by a thermal polycondensation method. The sample was obtained by heating in Ar atmosphere at 400 • C for 3 h. Dinari et al. [27] studied The other solvents and the chemicals were of analytical grade and used as received without further purification. Deionized-distilled water (DDW) was prepared and used exclusively in this study.

Synthesis of WO 3 /GR Nanocomposites
The WO 3 /GR nanocomposites were prepared by a one-step in-situ hydrothermal synthesis [44]. A flowchart for the WO 3 /GR nanocomposites preparation is shown in Scheme 1. Briefly, the stock GO dispersion (2 mg/mL) was pre-sonicated for 10 min and then the relevant GO dispersion (2 mL, 10 mL, and 20 mL) was dispersed in a certain amount of the deionized water, a total solution volume was 40 mL, in which the concentration of GO solution were 0.1 mg/mL, 0.5 mg/mL, and 1 mg/mL, respectively, then sonicated for 0.5 h at room temperature. Na 2 WO 4 ·2H 2 O (0.5 g), H 2 C 2 O 4 (1 g) and Na 2 SO 4 (4 g) were added into the suspension and the resulting solution was stirred for 3 h at room temperature. With vigorous stirring, the pH of the solution was adjusted to 1.5 by adding 3M HCl, and then stirring was continued for 3 h. The solution was transferred to a Teflon-lined stainless-steel autoclave and kept at 180 • C for 24 h. The final mixture was filtered and washed (three times) with deionized water and absolute ethanol by cross-centrifugation. The residual liquid was placed in a dry oven at 60 • C for 5 h. A series of WO 3 /GR nanocomposites were prepared by varying the GO content. The samples were numbered as WO 3 /GR-x (x = 0.1, 0.5, and 1), where x represents the concentration of the aqueous solution of GO. Under the same experimental conditions, pure WO 3 nanorods were prepared without GO for comparison with the obtained WO 3 /GR nanocomposites.

Materials
Sodium tungsten dihydrate (Na2WO4 2H2O), oxalic acid (H2C2O4) and anhydrous sodium sulfate (Na2SO4) were purchased from Sinopharm (Shanghai, China). A graphene oxide (GO) aqueous dispersion (2 mg/mL) was purchased from Nanjing Nano Technology Co., Ltd. (Nanjing, China). The other solvents and the chemicals were of analytical grade and used as received without further purification. Deionized-distilled water (DDW) was prepared and used exclusively in this study.

Synthesis of WO3/GR Nanocomposites
The WO3/GR nanocomposites were prepared by a one-step in-situ hydrothermal synthesis [44]. A flowchart for the WO3/GR nanocomposites preparation is shown in Scheme 1. Briefly, the stock GO dispersion (2 mg/mL) was pre-sonicated for 10 min and then the relevant GO dispersion (2 mL, 10 mL, and 20 mL) was dispersed in a certain amount of the deionized water, a total solution volume was 40 mL, in which the concentration of GO solution were 0.1 mg/mL, 0.5 mg/mL, and 1 mg/mL, respectively, then sonicated for 0.5 h at room temperature. Na2WO4 2H2O (0.5 g), H2C2O4 (1 g) and Na2SO4 (4 g) were added into the suspension and the resulting solution was stirred for 3 h at room temperature. With vigorous stirring, the pH of the solution was adjusted to 1.5 by adding 3M HCl, and then stirring was continued for 3 h. The solution was transferred to a Teflon-lined stainless-steel autoclave and kept at 180 °C for 24 h. The final mixture was filtered and washed (three times) with deionized water and absolute ethanol by cross-centrifugation. The residual liquid was placed in a dry oven at 60 °C for 5 h. A series of WO3/GR nanocomposites were prepared by varying the GO content. The samples were numbered as WO3/GR-x (x = 0.1, 0.5, and 1), where x represents the concentration of the aqueous solution of GO. Under the same experimental conditions, pure WO3 nanorods were prepared without GO for comparison with the obtained WO3/GR nanocomposites. Scheme 1. Flowchart for the preparation of WO3/GR (graphene) nanocomposites.

Characterizations of Materials
X-ray diffraction (XRD) was performed using an X'Pert PRO X-ray diffractometer (PANalytical B.V., Eindhoven, ZuidHolland, The Netherlands) with Cu Kα radiation (λ = 1.541 Å) in the 2θ range from 10° to 80° at a scan rate of 2°/min. Transmission electron microscopy (TEM) images of the WO3/GR nanocomposites were recorded using a FEI Titan Themis 200 TEM instrument (FEI, Hillsboro, OR, USA) operated at 200 kV equipped with an energy dispersive X-ray spectroscopy (EDS) detector (X-Max N TSR, Oxford Instrument, Oxford, UK). Scanning electron microscopy (SEM) was conducted on a JSM-7500F (JEOL, Tokyo, Japan) microscope to examine the WO3/GR nanocomposites.
The chemical interactions and purity of the samples were characterized by Raman spectroscopy on a Raman spectrometer (Horiba Evolution, Paris, France), which was carried out at room Scheme 1. Flowchart for the preparation of WO 3 /GR (graphene) nanocomposites.

Characterizations of Materials
X-ray diffraction (XRD) was performed using an X'Pert PRO X-ray diffractometer (PANalytical B.V., Eindhoven, ZuidHolland, The Netherlands) with Cu K α radiation (λ = 1.541 Å) in the 2θ range from 10 • to 80 • at a scan rate of 2 • /min. Transmission electron microscopy (TEM) images of the WO 3 /GR nanocomposites were recorded using a FEI Titan Themis 200 TEM instrument (FEI, Hillsboro, OR, USA) operated at 200 kV equipped with an energy dispersive X-ray spectroscopy (EDS) detector (X-Max N TSR, Oxford Instrument, Oxford, UK). Scanning electron microscopy (SEM) was conducted on a JSM-7500F (JEOL, Tokyo, Japan) microscope to examine the WO 3 /GR nanocomposites.
The chemical interactions and purity of the samples were characterized by Raman spectroscopy on a Raman spectrometer (Horiba Evolution, Paris, France), which was carried out at room temperature with a laser excitation source of 633 nm and a light spot size of 0.5 mm. X-ray photoelectron spectroscopy (XPS) was conducted on AXIS UltraDLD (Kratos, Kyoto, Japan) with an Mg K α X-ray as the excitation source to determine the chemical information from the surface and interface of the sample. The optical absorption properties of the solid samples were characterized by UV-vis diffuse reflectance spectroscopy (DRS) on a UV-vis spectrophotometer (UV-3600, Shimadzu, Kyoto, Japan) with a sphere attachment and BaSO 4 was used as the internal reflectance sample.

Evaluation of Photocatalytic Activity
The photocatalytic activity experiments on the prepared samples were studied by examining the photodegradation reaction of methylene blue (MB) as an organic pollutant. In a typical reaction, the photocatalyst (20 mg) was added to an MB solution (50 mL, 10 mg·L −1 ). The solution was magnetically stirred in the dark for 1 h to achieve adsorption equilibrium. The resulting solution was then irradiated directly with a 300 W PLS-SXE 300 xenon lamp (Beijing Perfect Light Technology Co., Ltd., Beijing, China) equipped with a λ < 400 nm cut-off filter. The purpose of the cut-off filter was to remove wavelengths below 400 nm to ensure that only visible light (420-700 nm) was irradiating the sample. The distance between the light source and the sample level is 20 cm, the irradiated surface is 3 cm × 3 cm. A sample (5 mL) of the solution was extracted from the mixture every 15 min and centrifuged to remove solids. The centrifuged liquid was poured into a cuvette and its absorbance was measured by an ultraviolet spectrophotometer (UV-1900, Haida Scientific Co., Ltd., Shanghai, China). The maximum absorbance of MB at the wavelength of 664 nm was selected to analyze the change in concentration as the MB solution was degraded. The photocatalytic activity of the samples under visible light irradiation was determined. Figure 1 shows the XRD trace of WO 3 and the WO 3 /GR nanocomposites doped with different amounts of GO. The XRD pattern of WO 3 was composed of three strong lines and broad half peaks at the diffraction angles of 13.9 • , 28.2 • and 37.1 • , which corresponded to the (100), (200) and (201) planes, respectively. This was consistent with the standard XRD pattern of hexagonal WO 3 . Hence, the sample contained hexagonal phase WO 3 . The sharpness of the main peaks indicated that the synthesized WO 3 had a good degree of crystallinity. photoelectron spectroscopy (XPS) was conducted on AXIS UltraDLD (Kratos, Kyoto, Japan) with an Mg Kα X-ray as the excitation source to determine the chemical information from the surface and interface of the sample. The optical absorption properties of the solid samples were characterized by UV-vis diffuse reflectance spectroscopy (DRS) on a UV-vis spectrophotometer (UV-3600, Shimadzu, Kyoto, Japan) with a sphere attachment and BaSO4 was used as the internal reflectance sample.

Evaluation of Photocatalytic Activity
The photocatalytic activity experiments on the prepared samples were studied by examining the photodegradation reaction of methylene blue (MB) as an organic pollutant. In a typical reaction, the photocatalyst (20 mg) was added to an MB solution (50 mL, 10 mg L −1 ). The solution was magnetically stirred in the dark for 1 h to achieve adsorption equilibrium. The resulting solution was then irradiated directly with a 300 W PLS-SXE 300 xenon lamp (Beijing Perfect Light Technology Co., Ltd., Beijing, China) equipped with a λ < 400 nm cut-off filter. The purpose of the cut-off filter was to remove wavelengths below 400 nm to ensure that only visible light (420-700 nm) was irradiating the sample. The distance between the light source and the sample level is 20 cm, the irradiated surface is 3 cm × 3 cm. A sample (5 mL) of the solution was extracted from the mixture every 15 min and centrifuged to remove solids. The centrifuged liquid was poured into a cuvette and its absorbance was measured by an ultraviolet spectrophotometer (UV-1900, Haida Scientific Co., Ltd., Shanghai, China). The maximum absorbance of MB at the wavelength of 664 nm was selected to analyze the change in concentration as the MB solution was degraded. The photocatalytic activity of the samples under visible light irradiation was determined. Figure 1 shows the XRD trace of WO3 and the WO3/GR nanocomposites doped with different amounts of GO. The XRD pattern of WO3 was composed of three strong lines and broad half peaks at the diffraction angles of 13.9°, 28.2° and 37.1°, which corresponded to the (100), (200) and (201) planes, respectively. This was consistent with the standard XRD pattern of hexagonal WO3. Hence, the sample contained hexagonal phase WO3. The sharpness of the main peaks indicated that the synthesized WO3 had a good degree of crystallinity.  The XRD patterns of the WO 3 /GR composites did not exhibit any special features compared with the XRD patterns of WO 3 . According to the literature [45][46][47], the characteristic peaks of graphene at 26 • and 42 • overlap with the WO 3 peaks for the (101) and (300) planes. Typically, the characteristic peak of GO is at 11 • . The absence of this peak confirmed that the GO had been reduced to graphene. As was mentioned above, it was again confirmed that hexagonal WO 3 was successfully grown on graphene according to the XRD patterns. The presence of graphene could be further demonstrated by Raman measurements.

Results and Discussion
A selected region of the Raman spectra of GO is shown in Figure 2a. GO showed two peaks at approximately 1336 cm −1 and 1609 cm −1 , the D-band and G-band, respectively, which was consistent with the results in the literature [48][49][50]. The G-band peak at around 1609 cm −1 is characteristic of a graphitic sheet corresponding to a well-defined sp 2 carbon-type structure, which can be attributed to the doubly degenerate zone center E 2g mode [51], whereas the D-band at approximately 1336 cm −1 can be attributed to the presence of defects or edges within the hexagonal graphite structure.
The XRD patterns of the WO3/GR composites did not exhibit any special features compared with the XRD patterns of WO3. According to the literature [45][46][47], the characteristic peaks of graphene at 26° and 42° overlap with the WO3 peaks for the (101) and (300) planes. Typically, the characteristic peak of GO is at 11°. The absence of this peak confirmed that the GO had been reduced to graphene. As was mentioned above, it was again confirmed that hexagonal WO3 was successfully grown on graphene according to the XRD patterns. The presence of graphene could be further demonstrated by Raman measurements.
A selected region of the Raman spectra of GO is shown in Figure 2a. GO showed two peaks at approximately 1336 cm −1 and 1609 cm −1 , the D-band and G-band, respectively, which was consistent with the results in the literature [48][49][50]. The G-band peak at around 1609 cm −1 is characteristic of a graphitic sheet corresponding to a well-defined sp 2 carbon-type structure, which can be attributed to the doubly degenerate zone center E2g mode [51], whereas the D-band at approximately 1336 cm −1 can be attributed to the presence of defects or edges within the hexagonal graphite structure.  [15,[31][32][33][34]. Compared with pure WO3, the peak intensity of the WO3 in the WO3/GR-0.1 nanocomposite was significantly weakened and broadened owing to the strong bonds between WO3 and graphene rather than the physical bonding (which weakens the internal W-O bond) as discussed by O. Akhavan et al. [52]. Two characteristic peaks at 1331 cm −1 and 1592 cm −1 were observed in the Raman spectrum of the WO3/GR-0.1 nanocomposite. The existence of graphene in the composites was confirmed. The characteristic peak of graphene tended to migrate slightly to a lower frequency. The intensity ratio ID/IG between the band-D and band-G can be used to determine the extent of graphene material defects [15]. This value was calculated to be 1.27 in GO and 1.05 in the WO3/GR nanocomposites. The results indicated the absence of defects in the composites and many oxygen-containing functional groups were reduced during the hydrothermal process. Figure 3a,b illustrates the microstructure of the WO3 nanorod. In the microstructures of the WO3/GR nanocomposites, WO3 nanorods were shown to be distributed on the surface of graphene at low magnification (Figure 3c,e,g) and high magnification (Figure 3d,f,h). A reunion phenomenon was observed for the WO3/GR-0.1 nanocomposite. At a high magnification, the graphene nanosheets were shown to have rough surfaces and curled edges. Similarly, agglomeration was still observed in the WO3/GR-0.5 nanocomposite. However, the amount of WO3 agglomeration significantly decreased with discoid WO3 at a higher magnification. The agglomeration of the WO3/GR-0.5 nanocomposite was not observed in the WO3/GR-1 nanocomposite. In the SEM images in Figure 3, the graphene was composed of a multi-GR sheets folded structure. It was concluded that the  [15,[31][32][33][34]. Compared with pure WO 3 , the peak intensity of the WO 3 in the WO 3 /GR-0.1 nanocomposite was significantly weakened and broadened owing to the strong bonds between WO 3 and graphene rather than the physical bonding (which weakens the internal W-O bond) as discussed by O. Akhavan et al. [52]. Two characteristic peaks at 1331 cm −1 and 1592 cm −1 were observed in the Raman spectrum of the WO 3 /GR-0.1 nanocomposite. The existence of graphene in the composites was confirmed. The characteristic peak of graphene tended to migrate slightly to a lower frequency. The intensity ratio I D /I G between the band-D and band-G can be used to determine the extent of graphene material defects [15]. This value was calculated to be 1.27 in GO and 1.05 in the WO 3 /GR nanocomposites. The results indicated the absence of defects in the composites and many oxygen-containing functional groups were reduced during the hydrothermal process. Figure 3a,b illustrates the microstructure of the WO 3 nanorod. In the microstructures of the WO 3 /GR nanocomposites, WO 3 nanorods were shown to be distributed on the surface of graphene at low magnification (Figure 3c,e,g) and high magnification (Figure 3d,f,h). A reunion phenomenon was observed for the WO 3 /GR-0.1 nanocomposite. At a high magnification, the graphene nanosheets were shown to have rough surfaces and curled edges. Similarly, agglomeration was still observed in the WO 3 /GR-0.5 nanocomposite. However, the amount of WO 3 agglomeration significantly decreased with discoid WO 3 at a higher magnification. The agglomeration of the WO 3 /GR-0.5 nanocomposite was not observed in the WO 3 /GR-1 nanocomposite. In the SEM images in Figure 3, the graphene was composed of a multi-GR sheets folded structure. It was concluded that the agglomeration of WO 3 nanorods decreased with the increase of GO content. Therefore, the introduction of the appropriate amount of graphene can easily disperse the graphene sheets and inhibit the agglomeration of the WO 3 nanorods, which increased the specific surface area and the active sites of the photocatalytic process and further improved the photocatalytic activity. An amount of WO 3 nanorods was found to grow on the surface of graphene. This was further proven by the bright-field TEM images (Figure 4a) of the WO 3 /GR-0.1 nanocomposite, EDS curve (Figure 4b) through the WO 3 nanorods and the EDS element mapping images of C, O and W (Figure 4d-f) in the selected area. As shown in the typical bright-field TEM image of the WO 3 /GR-0.1 nanocomposite (Figure 4c), the WO 3 nanorods were fixed to the flocculated graphene. It was found that chemical bonds were formed between WO 3 and the graphene sheets. agglomeration of WO3 nanorods decreased with the increase of GO content. Therefore, the introduction of the appropriate amount of graphene can easily disperse the graphene sheets and inhibit the agglomeration of the WO3 nanorods, which increased the specific surface area and the active sites of the photocatalytic process and further improved the photocatalytic activity. An amount of WO3 nanorods was found to grow on the surface of graphene. This was further proven by the bright-field TEM images (Figure 4a) of the WO3/GR-0.1 nanocomposite, EDS curve (Figure 4b) through the WO3 nanorods and the EDS element mapping images of C, O and W (Figure 4d-f) in the selected area. As shown in the typical bright-field TEM image of the WO3/GR-0.1 nanocomposite (Figure 4c), the WO3 nanorods were fixed to the flocculated graphene. It was found that chemical bonds were formed between WO3 and the graphene sheets.  As a powerful surface analysis technique, XPS is often used to determine the surface elemental composition, valence states and molecular structure of as-prepared samples. Figure 5 shows the XPS results of pure WO 3 and the WO 3 /GR-0.1 nanocomposite. The elements of C, O and W could be easily identified through the characteristic peaks of C1s, O1s and W4f. The survey spectra of pure WO 3 (Figure 5a) and the WO 3 /GR-0.1 nanocomposite (Figure 5d) exhibited similar peak shapes; therefore, the sample needed narrower scanning. The high-resolution W4f spectrum for pure WO 3 could be deconvoluted into two peaks at 35.0 eV and 37.1 eV, which were assigned to the W4f7/2 and W4f5/2 spin orbital splitting photoelectrons of the W 6+ chemical state. O1s can be decomposed into two separate binding energies at 529.5 eV and 531.1 eV, corresponding to the surface lattice oxygen (O latt ) and surface adsorbed oxygen (O ads ) [53], respectively (Figure 5b,c). In the WO 3 /GR-0.1 nanocomposites, the high-resolution spectrum of W4f are located at the binding energies of 36.4 eV and 38.4 eV, corresponding to the W4f7/2 and W4f5/2 peaks of the WO 3 phase in the composite, respectively. Peak area ratio of these two peaks is obtained 0.75, which is in line with the 4f level theory of spin orbit splitting; its O1s XPS spectra can be deconvoluted into binding energy at 531 eV and 532.8 eV peaks, corresponding to the W-O bond and C-OH bond. In conclusion, the deconvolution signal of the binding energies of O1s and W4f in WO 3 /GR-0.1 composites is more intense and migrates to higher binding energy than pure WO 3 , it is probably due to the internal C, O and W elements of the interaction between the graphene and WO 3 , which is consistent with the results of Raman analysis. Figures 4 and 5g is a high-resolution C1s spectrum of WO 3 /GR-0.1 nanocomposite; the three peaks are located at 284.6 eV, 286.3 eV and 287.9 eV binding energies, attributed to C-C, C-OH, C=O and C-O-C bonds, respectively.
The representative TEM-EDS of rGO, WO 3 and the WO 3 /GR-0.1 nanocomposite, and the corresponding HRTEM (High-resolution transmission electron microscopy) images are shown in Figure 6. As shown in Figure 6a, rGO is a two-dimensional nanosheet with several crinkles. The d-spacing of the lattice plane (002) GR was approximately 0.33 nm. The typical bright-field TEM images (Figure 6b) demonstrated that WO 3 presented as irregular nanorods. The irregular WO 3 nanorods were assembled by radial multi-rods, among which the interstitial space could be observed if the content of WO 3 was sufficiently large. The d-spacing of the lattice plane (001) WO3 was approximately 0.39 nm, which was in accordance with the (001)-spacing of h-WO 3 crystals. The TEM image of WO 3 /GR nanocomposites is shown in Figure 6c, from which a hybrid nanoarchitecture with WO 3 nanorods on the surface of graphene could be observed. The nanocomposites exhibited a clear interface between the WO 3 nanorods and graphene nanosheets.  As a powerful surface analysis technique, XPS is often used to determine the surface elemental composition, valence states and molecular structure of as-prepared samples. Figure 5 shows the XPS results of pure WO3 and the WO3/GR-0.1 nanocomposite. The elements of C, O and W could be easily identified through the characteristic peaks of C1s, O1s and W4f. The survey spectra of pure WO3 (Figure 5a) and the WO3/GR-0.1 nanocomposite (Figure 5d) exhibited similar peak shapes;  The representative TEM-EDS of rGO, WO3 and the WO3/GR-0.1 nanocomposite, and the corresponding HRTEM (High-resolution transmission electron microscopy) images are shown in Figure 6. As shown in Figure 6a, rGO is a two-dimensional nanosheet with several crinkles. The d-spacing of the lattice plane (002)GR was approximately 0.33 nm. The typical bright-field TEM images (Figure 6b) demonstrated that WO3 presented as irregular nanorods. The irregular WO3 nanorods were assembled by radial multi-rods, among which the interstitial space could be observed if the content of WO3 was sufficiently large. The d-spacing of the lattice plane (001)WO3 was approximately 0.39 nm, which was in accordance with the (001)-spacing of h-WO3 crystals. The TEM image of WO3/GR nanocomposites is shown in Figure 6c, from which a hybrid nanoarchitecture with WO3 nanorods on the surface of graphene could be observed. The nanocomposites exhibited a clear interface between the WO3 nanorods and graphene nanosheets.   The representative TEM-EDS of rGO, WO3 and the WO3/GR-0.1 nanocomposite, and the corresponding HRTEM (High-resolution transmission electron microscopy) images are shown in Figure 6. As shown in Figure 6a, rGO is a two-dimensional nanosheet with several crinkles. The d-spacing of the lattice plane (002)GR was approximately 0.33 nm. The typical bright-field TEM images (Figure 6b) demonstrated that WO3 presented as irregular nanorods. The irregular WO3 nanorods were assembled by radial multi-rods, among which the interstitial space could be observed if the content of WO3 was sufficiently large. The d-spacing of the lattice plane (001)WO3 was approximately 0.39 nm, which was in accordance with the (001)-spacing of h-WO3 crystals. The TEM image of WO3/GR nanocomposites is shown in Figure 6c, from which a hybrid nanoarchitecture with WO3 nanorods on the surface of graphene could be observed. The nanocomposites exhibited a clear interface between the WO3 nanorods and graphene nanosheets.   Figure 7 reveals the morphology of the WO 3 nanorod and the WO 3 /GR nanocomposites. As shown in Figure 7a, it was clear that the characteristic WO 3 nanorods were successfully synthesized through a one-step in-situ hydrothermal reaction. Taking the WO 3 /GR-0.1 nanocomposite as an example, a one-dimensional WO 3 nanorod was spread on the surface of the two-dimensional graphene nanosheet. The HRTEM images of regions B and C (Figure 7b,c) show the interfaces of the WO 3 nanorod and graphene nanosheet. The results indicated that an intimate interfacial contact between them was readily synthesized by such a simple hydrothermal reaction approach. The d-spacing of the lattice plane (001) near the interface in Figure 7b became smaller as the interface was approached. Some stacking faults and dislocations were also observed at the interface. The HRTEM image of region C (Figure 7c) shows the different lattice structures around the interface between WO 3 and graphene. The d-spacing of these lattice planes was in the range of 2.13 Å to 2.26 Å. Figure 7d shows the standard lattice plane (001). The result means that the lattice structure inside WO 3 was not affected by the graphene or the interfacial energy. example, a one-dimensional WO3 nanorod was spread on the surface of the two-dimensional graphene nanosheet. The HRTEM images of regions B and C (Figure 7b,c) show the interfaces of the WO3 nanorod and graphene nanosheet. The results indicated that an intimate interfacial contact between them was readily synthesized by such a simple hydrothermal reaction approach. The d-spacing of the lattice plane (001) near the interface in Figure 7b became smaller as the interface was approached. Some stacking faults and dislocations were also observed at the interface. The HRTEM image of region C (Figure 7c) shows the different lattice structures around the interface between WO3 and graphene. The d-spacing of these lattice planes was in the range of 2.13 Å to 2.26 Å. Figure  7d shows the standard lattice plane (001). The result means that the lattice structure inside WO3 was not affected by the graphene or the interfacial energy.  Figure 8 shows the UV-vis diffuse reflectance electronic spectra of pure WO3, the WO3/GR nanocomposites with different contents of GO and the effect of hv on the corresponding (αhv) 1/2 for the as-prepared samples. Compared with the pure WO3 nanorods, the WO3/GR nanocomposites showed a stronger photoabsorbance in the visible region from λ = 400 to 800 nm. This arose from the introduction of graphene. The solar light harvesting could be ascribed to the formation of W-O-C bonds between the WO3 and graphene [54]. In addition, the absorption edges were at 440 nm, 420 nm, 400 nm, and 255 nm for the pure WO3 nanorods, and the WO3/GR-0.1, WO3/GR-0.5, and WO3/GR-1 nanocomposites, respectively. It was clear that the absorption edge shifted to shorter  Figure 8 shows the UV-vis diffuse reflectance electronic spectra of pure WO 3 , the WO 3 /GR nanocomposites with different contents of GO and the effect of hv on the corresponding (αhv) 1/2 for the as-prepared samples. Compared with the pure WO 3 nanorods, the WO 3 /GR nanocomposites showed a stronger photoabsorbance in the visible region from λ = 400 to 800 nm. This arose from the introduction of graphene. The solar light harvesting could be ascribed to the formation of W-O-C bonds between the WO 3 and graphene [54]. In addition, the absorption edges were at 440 nm, 420 nm, 400 nm, and 255 nm for the pure WO 3 nanorods, and the WO 3 /GR-0.1, WO 3 /GR-0.5, and WO 3 /GR-1 nanocomposites, respectively. It was clear that the absorption edge shifted to shorter wavelength for the WO 3 /GR nanocomposites with the increase of GO content. According to the following formula for the semiconductor band gap, the band gap of the sample can be calculated: where K = b × c, b is the thickness of the cuvette or film sample, and c is the concentration. K has no effect on Eg, α is the absorbance, hv is the energy of the exciton, and Eg is the band gap energy.
wavelength for the WO3/GR nanocomposites with the increase of GO content. According to the following formula for the semiconductor band gap, the band gap of the sample can be calculated: where K = b × c, b is the thickness of the cuvette or film sample, and c is the concentration. K has no effect on Eg, α is the absorbance, hv is the energy of the exciton, and Eg is the band gap energy. Therefore, the band gap energies were 2.60, 2.25, 1.65 and 1.17 for pure WO3, and the WO3/GR-0.1, WO3/GR-0.5, and WO3/GR-1 nanocomposites, respectively. As shown in Figure 9, the initial concentration of MB exhibited almost no changes in the catalyst-free and pure WO3 solutions, whereas a decrease in the solution containing WO3/GR nanocomposites arose from the introduction of graphene. Part of the dye molecule adsorbed onto the surface of the WO3/GR nanocomposites, which resulted in a decrease of the MB concentration after the adsorption equilibrium. This was also beneficial for enhancing the photocatalytic performance of the WO3/GR nanocomposites. As shown in Figure 9, the initial concentration of MB exhibited almost no changes in the catalyst-free and pure WO 3 solutions, whereas a decrease in the solution containing WO 3 /GR nanocomposites arose from the introduction of graphene. Part of the dye molecule adsorbed onto the surface of the WO 3 /GR nanocomposites, which resulted in a decrease of the MB concentration after the adsorption equilibrium. This was also beneficial for enhancing the photocatalytic performance of the WO 3 /GR nanocomposites. Figure 10 displays the curve of the normalized concentration (C/C 0 ) of MB versus time under visible light, where C 0 is the initial concentration of MB in solution, and C is the concentration of MB remaining in the solution at each illumination interval. After 70 min of illumination, the normalized concentrations of MB in solution were 92%, 59%, 46%, 26% and 17%, respectively, for the catalyst-free, pure WO 3 , and the WO 3 /G-0.1, WO 3 /G-0.5 and WO 3 /G-1 nanocomposites. Therefore, the WO 3 /GR composites exhibited a higher photocatalytic efficiency than pure WO 3 . With the increase of graphene oxide content, the photocatalytic efficiency of WO 3 /GR was also increased, which arose from the introduction of graphene. The introduction of graphene increased the optical absorption efficiency of the composite materials, which was consistent with the DRS results, and inhibited the recombination rate of photogenerated electron-hole pairs, which thus improved the photocatalytic efficiency. The maximum degradation rate was achieved by WO 3 /G-1 (83%), which was double that of pure WO 3 (41%).  were 92%, 59%, 46%, 26% and 17%, respectively, for the catalyst-free, pure WO3, and the WO3/G-0.1, WO3/G-0.5 and WO3/G-1 nanocomposites. Therefore, the WO3/GR composites exhibited a higher photocatalytic efficiency than pure WO3. With the increase of graphene oxide content, the photocatalytic efficiency of WO3/GR was also increased, which arose from the introduction of graphene. The introduction of graphene increased the optical absorption efficiency of the composite materials, which was consistent with the DRS results, and inhibited the recombination rate of photogenerated electron-hole pairs, which thus improved the photocatalytic efficiency. The maximum degradation rate was achieved by WO3/G-1 (83%), which was double that of pure WO3 (41%). As mentioned above, some WO3-based and graphene-based composites had been prepared for the degradation of organic pollutants [26,27,42,43,55]. As shown in Table 1, their photocatalytic efficiency is usually 65-95% within 56-150 min. Compared with the previously reported composites   After 70 min of illumination, the normalized concentrations of MB in solution were 92%, 59%, 46%, 26% and 17%, respectively, for the catalyst-free, pure WO3, and the WO3/G-0.1, WO3/G-0.5 and WO3/G-1 nanocomposites. Therefore, the WO3/GR composites exhibited a higher photocatalytic efficiency than pure WO3. With the increase of graphene oxide content, the photocatalytic efficiency of WO3/GR was also increased, which arose from the introduction of graphene. The introduction of graphene increased the optical absorption efficiency of the composite materials, which was consistent with the DRS results, and inhibited the recombination rate of photogenerated electron-hole pairs, which thus improved the photocatalytic efficiency. The maximum degradation rate was achieved by WO3/G-1 (83%), which was double that of pure WO3 (41%). As mentioned above, some WO3-based and graphene-based composites had been prepared for the degradation of organic pollutants [26,27,42,43,55]. As shown in Table 1, their photocatalytic efficiency is usually 65-95% within 56-150 min. Compared with the previously reported composites used in the field of photocatalytic degradation of pollutants, WO3/GR prepared in this work shows a high photocatalytic efficiency within short time, which can be easily found in Table 1. In addition, the method of the composite is also simpler and more feasible. As mentioned above, some WO 3 -based and graphene-based composites had been prepared for the degradation of organic pollutants [26,27,42,43,55]. As shown in Table 1, their photocatalytic efficiency is usually 65-95% within 56-150 min. Compared with the previously reported composites used in the field of photocatalytic degradation of pollutants, WO 3 /GR prepared in this work shows a high photocatalytic efficiency within short time, which can be easily found in Table 1. In addition, the method of the composite is also simpler and more feasible.  Figure 11 illustrates the mechanism of the photocatalytic degradation of MB by the WO 3 /GR nanocomposites under visible light irradiation. In the solar spectrum, the energy distribution in the ultraviolet spectrum almost disappears (UV light below 0.29 m is almost all absorbed), and only approximately 3% remains. The infrared spectrum occupies 53%, while the visible spectrum accounts for 44%, which is nearly half of the solar energy. Under visible light irradiation, hv (vis) ≥ Eg. The electron in WO 3 acquires enough energy to generate electron-hole pairs; consequently, the WO 3 electron in the valence band is transferred into the conduction band of graphene. These electron-hole pairs undergo reduction-oxidation reactions, where the valence band hole is a good oxidant and the electron in the conduction band is a good reductant. Owing to the high carrier mobility of graphene (20,000 cm 2 V −1 s −1 ) and its 2D conjugated structure, the introduction of graphene behaves as an electron acceptor. Therefore, the efficiency of the electronic separation is enhanced. The rate of recombination of the photogenerated electron-hole pairs is inhibited. This results in an improvement of the photocatalytic efficiency for the WO 3 /GR nanocomposites. The process can be written as: OH * + MB → CO 2 + H 2 O Materials 2018, 11, x FOR PEER REVIEW 13 of 16 Figure 11 illustrates the mechanism of the photocatalytic degradation of MB by the WO3/GR nanocomposites under visible light irradiation. In the solar spectrum, the energy distribution in the ultraviolet spectrum almost disappears (UV light below 0.29 m is almost all absorbed), and only approximately 3% remains. The infrared spectrum occupies 53%, while the visible spectrum accounts for 44%, which is nearly half of the solar energy. Under visible light irradiation, hv (vis) ≥ Eg. The electron in WO3 acquires enough energy to generate electron-hole pairs; consequently, the WO3 electron in the valence band is transferred into the conduction band of graphene. These electron-hole pairs undergo reduction-oxidation reactions, where the valence band hole is a good oxidant and the electron in the conduction band is a good reductant. Owing to the high carrier mobility of graphene (20,000 cm 2 V −1 s −1 ) and its 2D conjugated structure, the introduction of graphene behaves as an electron acceptor. Therefore, the efficiency of the electronic separation is enhanced. The rate of recombination of the photogenerated electron-hole pairs is inhibited. This results in an improvement of the photocatalytic efficiency for the WO3/GR nanocomposites. The process can be written as:

Conclusions
WO3/GR nanocomposites, which are visible light photocatalysts, were successfully synthesized by a one-step in-situ hydrothermal method. This method could directly reduce GO to graphene without any reductant, and this was accompanied by the attachment to WO3. The formation of WO3/GR composites was confirmed by SEM, TEM, EDS, Raman and XPS. The presence of graphene in the composites promoted the electron transfer and optical absorption properties. The results showed that the WO3/GR composites exhibited an enhanced photocatalysis efficiency in visible light, which was double that of pure WO3. In addition, the synthesis method of the graphene-based

Conclusions
WO 3 /GR nanocomposites, which are visible light photocatalysts, were successfully synthesized by a one-step in-situ hydrothermal method. This method could directly reduce GO to graphene without any reductant, and this was accompanied by the attachment to WO 3 . The formation of WO 3 /GR composites was confirmed by SEM, TEM, EDS, Raman and XPS. The presence of graphene in the composites promoted the electron transfer and optical absorption properties. The results showed that the WO 3 /GR composites exhibited an enhanced photocatalysis efficiency in visible light, which was double that of pure WO 3 . In addition, the synthesis method of the graphene-based composite material can be extended to other functional material synthesis fields to realize promising applications in photocatalysis.