Facile Synthesis of Lacunary Keggin-Type Phosphotungstates-Decorated g-C3N4 Nanosheets for Enhancing Photocatalytic H2 Generation

In this work, the lacunary Keggin-type phosphotungstates of [PW9O34]9- (PW9) clusters were loaded onto the g-C3N4 nanosheets (NSs) to synthesize the phosphotungstate clusters-decorated 2D heterojunction photocatalysts by using the electrostatic-force driven self-assembly process. The surface charge polarity of g-C3N4 NSs was changed from a negative to a positive charge through the acidizing treatment. The positively-charged g-C3N4 NSs allowed the negatively-charged PW9 clusters to be adsorbed and deposited onto the g-C3N4 NSs, forming the PW9/g-C3N4 heterojunction NSs. The as-synthesized samples were characterized by scanning electron microscopy (SEM), X-ray diffraction (XRD), Fourier-transform infrared (FTIR) spectroscopy, and UV-VIS absorption spectra, respectively. The photocatalytic activity tests indicated that, upon simulated sunlight irradiation, the photocatalytic H2-generation rate of PW9/g-C3N4 heterojunction NSs (~23.8 μmol h−1) was ~3.3 times higher than that of the pure g-C3N4 NSs (~7.3 μmol h−1). The enhanced photocatalytic activity of PW9 cluster-decorated g-C3N4 NSs could be attributed to the enhanced separation process of the photoinduced charge-carriers, due to the Z-scheme-mediate charge transfer behavior across their hetero-interface.


Introduction
The continuous burning of unsustainable fossil energy sources, which emits large amounts of harmful gases, has induced a global energy crisis and environmental contamination. The development of green renewable energy sources to replace traditional fossil energy sources has become an urgent issue [1,2]. As a kind of pollution-free sustainable fuel, hydrogen (H 2 ) has been regarded as an ideal candidate for the application of energy supply in future society [3,4]. Splitting water through the additional electric or photic energy is a common method to obtain the H 2 . Since the H 2 generation from the photo-driven water splitting over the TiO 2 semiconductor photocatalyst was first reported in 1972 [5], this photochemical reaction has been widely used in the research area of solar-to-fuels conversion, mainly owing to its low cost and low energy consumption [6][7][8]. However, because the classical photocatalyst of TiO 2 nanostructures is a wide bandgap semiconductor (3.0-3.2 eV), it can only absorb the UV light that accounts for 3-5% in the solar spectrum [9,10]. Thus, the development of narrow bandgap semiconductors as the photocatalysts for H 2 generation has emerged as a hot topic in the area of solar-to-fuels conversion.

Heterojunction NSs
The lacunary Keggin-type phosphotungstates of Na 9 [A-α-PW 9 O 34 ]·7H 2 O (named as PW 9 ) were synthesized by using the previously reported method [20]. The g-C 3 N 4 was synthesized through the traditional thermal polymerization method by using the urea as the precursor [21]. To obtain the PW 9 /g-C 3 N 4 heterojunction NSs, 100 mg of the g-C 3 N 4 bulk was ground and then dispersed into deionized water under the ultrasonic treatment for 90 min at room temperature, thereby achieving the g-C 3 N 4 nanosheets (NSs). Afterward, 1 M of HCl solution was added dropwise into the above g-C 3 N 4 nanosheets-suspended water solution to adjust the surface charge polarity of the NSs. After that, 100 mg of the PW 9 was dissolved into the above solution with vigorous stirring for 4 h. The precipitate in the solution was separated by centrifugation treatment. Finally, the obtained mixture was dried at 80 • C for 12 h, which was then under annealing at 300 • C for 2 h in vacuum. Thus, the PW 9 /g-C 3 N 4 heterojunction NSs were synthesized.

Characterization
The structure and morphology of the as-synthesized samples were observed by scanning electron microscopy (Field Emission-SEM; S-4800, Hitachi, Tokyo, Japan) and transmission electron microscopy (TEM) (JEM-2100, JEOL, Tokyo, Japan). The X-ray diffraction (XRD) patterns of the as-fabricated samples were studied by X-ray diffractometer (XRD-6000, Shimadzu, Tokyo, Japan) with a Cu Kα line of 0.1541 nm and the radiation is from 10 • to 65 • at a scanning rate at 2 • /min. The Fourier-transform infrared (FTIR) spectra were recorded on Magna 560 FTIR spectrometer (Thermo Nicolet Corporation, Madison, WI, USA) with a resolution of 1 cm −1 . The UV-VIS absorption spectra of the samples were recorded on a Lambda 750 UV/VIS/NIR spectrophotometer (Perkin Elmer, Waltham, MA, USA). The specific surface areas of the as-prepared samples were measured with a Micromeritics ASAP-2020 instrument (USA), and analyzed by the Brunauer-Emmett-Teller (BET) method. The surface charge polarity of the samples was measured with a dynamic light scattering spectrophotometer and isoelectric point determination with zeta potential analysis (SZ-100, Horiba, Tokyo, Japan).

Photocatalytic H 2 Generation
Five milligrams of the as-synthesized samples were dispersed into 10 mL of a water solution containing triethanolamine (TEOA, 15 vol.%) and chloroplatinic acid (H 2 PtCl 6 , 10 µL, 12 mM). Then, the mixture solution was sealed in a quartz reactor and then ventilated with argon gas for 10 min to drive away the residual air. Afterward, the reactor was exposed under simulated sunlight (300-W Xe lamp, PLS-SXE300UV, Beijing, China, coupled with an AM 1.5 filter, the wavelength range of light was from 320-2500 nm, the light density was 100 mW cm −2 , and the photon flux of the lamp was 0.17 µmol s −1 m −2 ). The produced gas was periodically analyzed by a gas chromatograph (GC) equipped with a thermal conductivity detector (TCD) (Beifen-Ruili Analytical Instrument, SP-3420A, Beijing, China).

Results and Discussion
To synthesize the PW 9 /g-C 3 N 4 heterojunction NSs, the g-C 3 N 4 bulk was crushed into the g-C 3 N 4 nanosheets (NSs), at first, through the ultrasonic method. Then the obtained g-C 3 N 4 NSs underwent the acidizing treatment to adjust their surface charge polarity from −21 mV to +6.6 mV. Afterward, the positively-charged g-C 3 N 4 NSs were suspended into the water solution of PW 9 to enable the negatively-charged PW 9 to be self-assembled onto the g-C 3 N 4 NSs, based on the electrostatic force [22,23]. Finally, the target sample of PW 9 /g-C 3 N 4 heterojunction NSs was synthesized for further investigation. Figure 1B shows the scanning electron microscopy (SEM) image of the pure g-C 3 N 4 NSs. It can be seen that the sheet-like g-C 3 N 4 has a relatively smooth surface. After PW 9 loading, there was also no change on the surface roughness of the g-C 3 N 4 NSs in the PW 9 /g-C 3 N 4 composite NSs ( Figure 1C). In order to further investigate the as-prepared PW 9 /g-C 3 N 4 heterojunction NSs, the high-resolution transmission electron microscopy (HRTEM) and the dark-filed scanning mode TEM were carried out. Although the lattice-fringe spacing of both PW 9 and g-C 3 N 4 cannot be observed in HRTEM ( Figure 1D), further investigation, by using dark-filed scanning mode TEM (STEM), indicated that the PW 9 clusters with sizes of 3-5 nm were decorated on the surface of g-C 3 N 4 NSs in their heterojunction NSs ( Figure 1E). In addition, the specific surface areas of PW 9 /g-C 3 N 4 heterojunction NSs (~68.7 m 2 /g) was similar to that of g-C 3 N 4 (~53.1 m 2 /g), as shown in Figure S1. The phase structures of the as-synthesized samples were identified through X-ray diffraction (XRD) patterns. As shown in Figure 2, the two intense diffraction peaks around 28°~30° can be observed on the XRD pattern of the PW9, which is in accordance with the feature peaks of the PW9 reported in the literature [24]. Meanwhile, the other peaks on the diffraction pattern of the Na-PW9 were also matched with that of the reported PW9, confirming the obtained lacunary Keggin-type phosphotungstates of PW9. In the case of g-C3N4 NSs, the two feature diffraction peaks, belonging to the periodic structure of intra-planar tri-s-triazine and the interlayer stacking of conjugated aromatic structures of g-C3N4 NSs were found at 13.1° and 27.4°, respectively [25][26][27][28]. This is in agreement with the graphite structure of carbon nitride. When the modification of g-C3N4 NSs with PW9 clusters took place, the feature peaks originated from the Na-PW9 with the center at 29.3° and the g-C3N4 NSs with centers at 13.1° and 27.4° and appeared on the XRD pattern of the formed composite. This result suggests that the g-C3N4 NSs were decorated with the lacunary Keggin-type phosphotungstates of PW9 clusters. To further confirm the existence of the PW9 clusters on the g-C3N4 NSs, the Fourier-transform infrared (FTIR) spectra of the PW9/g-C3N4 heterojunction NSs along with the corresponding single Schematic diagram of the self-assembly process of the PW 9 /g-C 3 N 4 heterojunction NSs, based on the electrostatic force between these two hetero-components; SEM images of (B) g-C 3 N 4 NSs and (C) PW 9 /g-C 3 N 4 heterojunction NSs; the HRTEM image (D) and the dark-field STEM image (E) of the heterojunction NSs.
The phase structures of the as-synthesized samples were identified through X-ray diffraction (XRD) patterns. As shown in Figure 2, the two intense diffraction peaks around 28 •~3 0 • can be observed on the XRD pattern of the PW 9 , which is in accordance with the feature peaks of the PW 9 reported in the literature [24]. Meanwhile, the other peaks on the diffraction pattern of the Na-PW 9 were also matched with that of the reported PW 9 , confirming the obtained lacunary Keggin-type phosphotungstates of PW 9 . In the case of g-C 3 N 4 NSs, the two feature diffraction peaks, belonging to the periodic structure of intra-planar tri-s-triazine and the interlayer stacking of conjugated aromatic structures of g-C 3 N 4 NSs were found at 13.1 • and 27.4 • , respectively [25][26][27][28]. This is in agreement with the graphite structure of carbon nitride. When the modification of g-C 3 N 4 NSs with PW 9 clusters took place, the feature peaks originated from the Na-PW 9 with the center at 29.3 • and the g-C 3 N 4 NSs with centers at 13.1 • and 27.4 • and appeared on the XRD pattern of the formed composite. This result suggests that the g-C 3 N 4 NSs were decorated with the lacunary Keggin-type phosphotungstates of PW 9 clusters. The phase structures of the as-synthesized samples were identified through X-ray diffraction (XRD) patterns. As shown in Figure 2, the two intense diffraction peaks around 28°~30° can be observed on the XRD pattern of the PW9, which is in accordance with the feature peaks of the PW9 reported in the literature [24]. Meanwhile, the other peaks on the diffraction pattern of the Na-PW9 were also matched with that of the reported PW9, confirming the obtained lacunary Keggin-type phosphotungstates of PW9. In the case of g-C3N4 NSs, the two feature diffraction peaks, belonging to the periodic structure of intra-planar tri-s-triazine and the interlayer stacking of conjugated aromatic structures of g-C3N4 NSs were found at 13.1° and 27.4°, respectively [25][26][27][28]. This is in agreement with the graphite structure of carbon nitride. When the modification of g-C3N4 NSs with PW9 clusters took place, the feature peaks originated from the Na-PW9 with the center at 29.3° and the g-C3N4 NSs with centers at 13.1° and 27.4° and appeared on the XRD pattern of the formed composite. This result suggests that the g-C3N4 NSs were decorated with the lacunary Keggin-type phosphotungstates of PW9 clusters. To further confirm the existence of the PW9 clusters on the g-C3N4 NSs, the Fourier-transform infrared (FTIR) spectra of the PW9/g-C3N4 heterojunction NSs along with the corresponding single To further confirm the existence of the PW 9 clusters on the g-C 3 N 4 NSs, the Fourier-transform infrared (FTIR) spectra of the PW 9 /g-C 3 N 4 heterojunction NSs along with the corresponding single hetero-components were tested, as shown in Figure 3. For the pure PW 9 , the classical stretching vibrations related to the W-O-W, W-O, and P-O bonds were positioned around the 804/890, 984, and 1077 cm −1 , respectively [29,30]. It further proves the formation of lacunary Keggin-type phosphotungstates of PW 9 . In the case of pure g-C 3 N 4 NSs, the vibration bands between 1200 and 1650 cm −1 were ascribed to the CN heterocycles [24,31]. The band at 811 cm −1 is attributed to the feature vibration mode of s-triazine ring unit [32]. These observations are consistent with the literatures. The feature stretching vibration bands of both PW 9 and g-C 3 N 4 could be found on the FTIR spectrum of the PW 9 /g-C 3 N 4 heterojunction NSs, powerfully confirming the successful decoration of the PW 9 clusters onto the g-C 3 N 4 NSs. Notably, as compared with pure g-C 3 N 4 , the feature band at 811 cm −1 was slightly shifted to 809 cm −1 , which might be the cause of the chemical interaction between PW 9 and g-C 3 N 4 produced by the electrostatic adsorption process. hetero-components were tested, as shown in Figure 3. For the pure PW9, the classical stretching vibrations related to the W-O-W, W-O, and P-O bonds were positioned around the 804/890, 984, and 1077 cm − 1 , respectively [29,30]. It further proves the formation of lacunary Keggin-type phosphotungstates of PW9. In the case of pure g-C3N4 NSs, the vibration bands between 1200 and 1650 cm − 1 were ascribed to the CN heterocycles [24,31]. The band at 811 cm − 1 is attributed to the feature vibration mode of s-triazine ring unit [32]. These observations are consistent with the literatures. The feature stretching vibration bands of both PW9 and g-C3N4 could be found on the FTIR spectrum of the PW9/g-C3N4 heterojunction NSs, powerfully confirming the successful decoration of the PW9 clusters onto the g-C3N4 NSs. Notably, as compared with pure g-C3N4, the feature band at 811 cm − 1 was slightly shifted to 809 cm − 1 , which might be the cause of the chemical interaction between PW9 and g-C3N4 produced by the electrostatic adsorption process. The light absorption behaviors of the as-synthesized samples were evaluated through UV-VIS absorption spectra. As observed in Figure 4, the absorption edge of the pure PW9 is located around 350 nm, corresponding to the forbidden gap of ~3.5 eV between HOMO and LUMO of the PW9 [15,33]. Meanwhile, the absorption edge of pure g-C3N4 NSs appears at ~480 nm, suggesting the ~2.6 eV of the bandgap of the g-C3N4 NSs [34]. The above results indicate that the PW9 and g-C3N4 NSs are the UV and visible absorbers, respectively, during the photo-excitation process. After loading the PW9 clusters onto the surface of g-C3N4 NSs, the formed PW9/g-C3N4 heterojunction NSs displayed the blue-shift behavior of the absorption edge, due to the introduction of the UV-lightactive PW9 hetero-component. By combining the results of the SEM, TEM, XRD, FTIR, and UV-VIS absorption spectra, we confirm that the PW9/g-C3N4 heterojunction NSs with a well-distribution of PW9 clusters were successfully synthesized by using the self-assembly process driven by the electrostatic force. The light absorption behaviors of the as-synthesized samples were evaluated through UV-VIS absorption spectra. As observed in Figure 4, the absorption edge of the pure PW 9 is located around 350 nm, corresponding to the forbidden gap of~3.5 eV between HOMO and LUMO of the PW 9 [15,33]. Meanwhile, the absorption edge of pure g-C 3 N 4 NSs appears at~480 nm, suggesting the~2.6 eV of the bandgap of the g-C 3 N 4 NSs [34]. The above results indicate that the PW 9 and g-C 3 N 4 NSs are the UV and visible absorbers, respectively, during the photo-excitation process. After loading the PW 9 clusters onto the surface of g-C 3 N 4 NSs, the formed PW 9 /g-C 3 N 4 heterojunction NSs displayed the blue-shift behavior of the absorption edge, due to the introduction of the UV-light-active PW 9 hetero-component. By combining the results of the SEM, TEM, XRD, FTIR, and UV-VIS absorption spectra, we confirm that the PW 9 /g-C 3 N 4 heterojunction NSs with a well-distribution of PW 9 clusters were successfully synthesized by using the self-assembly process driven by the electrostatic force.  Photocatalytic activity of the as-synthesized PW9/g-C3N4 heterojunction NSs for H2 generation was assessed under simulated sunlight excitation in the presence of the sacrificial agent of triethanolamine (TEOA) to quench the photoinduced hole. As observed in Figure 5A, there is no observable H2 generation under simulated sunlight irradiation for 2 h, when using the single PW9 as the photocatalyst. Meanwhile, the pure g-C3N4 NSs only displayed a poor photocatalytic activity for H2 generation (~14.6 μmol in 2 h-irradiation). After self-assembling the PW9 clusters onto the g-C3N4 NSs to form the PW9/g-C3N4 heterojunction NSs, the photocatalytic activity of the heterojunction NSs for H2 generation remarkably increased to ~47.6 μmol in 2 h-irradiation. The H2 generation rate of the PW9/g-C3N4 heterojunction NSs was 3.3 times higher than that of the g-C3N4 NSs, as summarized in Figure 5B. Notably, this photocatalytic activity is almost the highest one among the values reported for other g-C3N4-based photocatalytic systems (Table 1). Moreover, the solar-to-hydrogen (STH) efficiency has become an important index for evaluating the photocatalytic activity of hydrogen production. The STH efficiency under simulated sunlight irradiation can be calculated according to the following equation [35,36]: After calculating, the STH of PW9/g-C3N4 heterojunction NSs was determined to be about 0.26% at room temperature.
It should be pointed out that the LUMO potential of PW9 does not satisfy the demand of proton reduction, while the g-C3N4 possesses a suitable conduction band (CB) potential for fulfilling the H2 generation [37,38]. Thus, we could deduce that the photocatalytic sites in the PW9/g-C3N4 heterojunction NSs should be located at the surface of g-C3N4 hetero-component [39,40]. Meanwhile, the PW9 hetero-component serves as the photosensitizer to boost the separation process of the photoinduced charge-carriers of the g-C3N4 hetero-component. According to the energy band structures of the reported Keggin-type phosphotungstates and g-C3N4 NSs, the LUMO position is a little higher than the valence band (VB) position of the g-C3N4 NSs, as illustrated in the inset of Figure 5B. Thus, when the PW9/g-C3N4 heterojunction NSs is excited by simulated sunlight that contains the photon energies in both UV and visible light regions, the photoinduced electron-hole pairs are generated on the LUMO and VB of the PW9 and g-C3N4 hetero-components, respectively [41,42]. Owing to the potential difference between the LUMO of PW9 and the VB of g-C3N4, the photoinduced electrons on the LUMO of PW9 could transfer to the VB of g-C3N4 in the heterojunction NSs, thereby extending the lifetimes of the photoinduced electrons on the CB of g-C3N4 for implementing the photocatalytic H2 generation [43]. In this way, the Z-scheme Photocatalytic activity of the as-synthesized PW 9 /g-C 3 N 4 heterojunction NSs for H 2 generation was assessed under simulated sunlight excitation in the presence of the sacrificial agent of triethanolamine (TEOA) to quench the photoinduced hole. As observed in Figure 5A, there is no observable H 2 generation under simulated sunlight irradiation for 2 h, when using the single PW 9 as the photocatalyst. Meanwhile, the pure g-C 3 N 4 NSs only displayed a poor photocatalytic activity for H 2 generation (~14.6 µmol in 2 h-irradiation). After self-assembling the PW 9 clusters onto the g-C 3 N 4 NSs to form the PW 9 /g-C 3 N 4 heterojunction NSs, the photocatalytic activity of the heterojunction NSs for H 2 generation remarkably increased to~47.6 µmol in 2 h-irradiation. The H 2 generation rate of the PW 9 /g-C 3 N 4 heterojunction NSs was 3.3 times higher than that of the g-C 3 N 4 NSs, as summarized in Figure 5B. Notably, this photocatalytic activity is almost the highest one among the values reported for other g-C 3 N 4 -based photocatalytic systems (Table 1). Moreover, the solar-to-hydrogen (STH) efficiency has become an important index for evaluating the photocatalytic activity of hydrogen production. The STH efficiency under simulated sunlight irradiation can be calculated according to the following equation [35,36]: Energy of generation of hydrogen by water splitting Solar energy irradiating the reaction cell × 100% (1) After calculating, the STH of PW 9 /g-C 3 N 4 heterojunction NSs was determined to be about 0.26% at room temperature.
It should be pointed out that the LUMO potential of PW 9 does not satisfy the demand of proton reduction, while the g-C 3 N 4 possesses a suitable conduction band (CB) potential for fulfilling the H 2 generation [37,38]. Thus, we could deduce that the photocatalytic sites in the PW 9 /g-C 3 N 4 heterojunction NSs should be located at the surface of g-C 3 N 4 hetero-component [39,40]. Meanwhile, the PW 9 hetero-component serves as the photosensitizer to boost the separation process of the photoinduced charge-carriers of the g-C 3 N 4 hetero-component. According to the energy band structures of the reported Keggin-type phosphotungstates and g-C 3 N 4 NSs, the LUMO position is a little higher than the valence band (VB) position of the g-C 3 N 4 NSs, as illustrated in the inset of Figure 5B. Thus, when the PW 9 /g-C 3 N 4 heterojunction NSs is excited by simulated sunlight that contains the photon energies in both UV and visible light regions, the photoinduced electron-hole pairs are generated on the LUMO and VB of the PW 9 and g-C 3 N 4 hetero-components, respectively [41,42]. Owing to the potential difference between the LUMO of PW 9 and the VB of g-C 3 N 4 , the photoinduced electrons on the LUMO of PW 9 could transfer to the VB of g-C 3 N 4 in the heterojunction NSs, thereby extending the lifetimes of the photoinduced electrons on the CB of g-C 3 N 4 for implementing the photocatalytic H 2 generation [43]. In this way, the Z-scheme photocatalytic mechanism can be employed to explain the enhanced photocatalytic activity of the PW 9 /g-C 3 N 4 heterojunction system.  If the Z-scheme electron-transfer process is the main factor for enhancing the photocatalytic activity of PW9/g-C3N4 heterojunction NSs, the hetero-interface combination force between these two hetero-components should have an influence on the photocatalytic H2 generation of the NSs. In order to investigate this hypothesis, we synthesized another heterojunction NSs of g-C3N4, decorated with the normal Keggin-type phosphotungstic acid (H3PW12O40) (PW12). Because the number of the negative charges in PW12 is less than that in PW9 when dissolving them in water solution, the combination force between the PW12 and g-C3N4 should be lower than that in the PW9 and g-C3N4 system. In theory, a strong combination force between the hetero-components would build a high-quality of the transport channel for electron transfer between the hetero-components [51,52]. Thus, the interfacial electron transfer process in the PW9/g-C3N4 heterojunction NSs should be more effective for driving the Z-scheme photocatalytic H2 generation as compared to the PW12/g-C3N4 heterojunction NSs ( the insets of Figure 6) [53]. The comparison study of the photocatalytic activities of the above two heterojunction NSs indicated that the H2-generation activity of PW9/g-  If the Z-scheme electron-transfer process is the main factor for enhancing the photocatalytic activity of PW 9 /g-C 3 N 4 heterojunction NSs, the hetero-interface combination force between these two hetero-components should have an influence on the photocatalytic H 2 generation of the NSs. In order to investigate this hypothesis, we synthesized another heterojunction NSs of g-C 3 N 4 , decorated with the normal Keggin-type phosphotungstic acid (H 3 PW 12 O 40 ) (PW 12 ). Because the number of the negative charges in PW 12 is less than that in PW 9 when dissolving them in water solution, the combination force between the PW 12 and g-C 3 N 4 should be lower than that in the PW 9 and g-C 3 N 4 system. In theory, a strong combination force between the hetero-components would build a high-quality of the transport channel for electron transfer between the hetero-components [51,52]. Thus, the interfacial electron transfer process in the PW 9 /g-C 3 N 4 heterojunction NSs should be more effective for driving the Z-scheme photocatalytic H 2 generation as compared to the PW 12 /g-C 3 N 4 heterojunction NSs (the insets of Figure 6) [53]. The comparison study of the photocatalytic activities of the above two heterojunction NSs indicated that the H 2 -generation activity of PW 9 /g-C 3 N 4 heterojunction NSs (~47.6 µmol in 2 h irradiation) was~1.3 times higher than that of the PW 12 /g-C 3 N 4 heterojunction NSs (~35.3 µmol in 2 h irradiation). This evidence further confirms a fast Z-scheme charge-transfer process occurred in the intimate hetero-interface of PW 9 /g-C 3 N 4 for enhancing the photocatalytic H 2 generation. C3N4 heterojunction NSs (~47.6 μmol in 2 h irradiation) was ~1.3 times higher than that of the PW12/g-C3N4 heterojunction NSs (~35.3 μmol in 2 h irradiation). This evidence further confirms a fast Z-scheme charge-transfer process occurred in the intimate hetero-interface of PW9/g-C3N4 for enhancing the photocatalytic H2 generation.

Conclusions
In summary, the lacunary Keggin-type phosphotungstates/g-C3N4 heterojunction NSs were synthesized through self-assembling the PW9 clusters onto the g-C3N4 NSs, based on the electrostatic adsorption process. When adjusting the surface charge polarity of g-C3N4 NSs from the negative to the positive charge in the PW9-dissolved solution, it was found that the intimate heterointerface between PW9/g-C3N4 heterojunction NSs could be built for boosting the Z-scheme chargetransfer process. Upon simulated sunlight irradiation, the lifetimes of the photoinduced electrons on the g-C3N4 hetero-component were prolonged in the heterojunction NSs. Thus, a ~3.3-fold enhancement of the photocatalytic activity for H2 generation was observed over the PW9/g-C3N4 heterojunction NSs as compared to the pure g-C3N4 NSs. Our work will provide a new platform to construct the ultra-small sized polyoxometalates clusters-based Z-scheme photocatalysts for enhancing the photocatalytic solar-to-fuels conversion.

Conflicts of Interest:
The authors declare no conflict of interest. Figure 6. Time-dependent photocatalytic H 2 -generation plots over (a) PW 9 /g-C 3 N 4 heterojunction NSs and (b) PW 12 /g-C 3 N 4 heterojunction NSs upon simulated sunlight irradiation. The insets showing the interfacial adsorption and electron transfer abilities of the PW 9 /g-C 3 N 4 heterojunction and PW 12 /g-C 3 N 4 heterojunction NSs.

Conclusions
In summary, the lacunary Keggin-type phosphotungstates/g-C 3 N 4 heterojunction NSs were synthesized through self-assembling the PW 9 clusters onto the g-C 3 N 4 NSs, based on the electrostatic adsorption process. When adjusting the surface charge polarity of g-C 3 N 4 NSs from the negative to the positive charge in the PW 9 -dissolved solution, it was found that the intimate hetero-interface between PW 9 /g-C 3 N 4 heterojunction NSs could be built for boosting the Z-scheme charge-transfer process. Upon simulated sunlight irradiation, the lifetimes of the photoinduced electrons on the g-C 3 N 4 hetero-component were prolonged in the heterojunction NSs. Thus, a~3.3-fold enhancement of the photocatalytic activity for H 2 generation was observed over the PW 9 /g-C 3 N 4 heterojunction NSs as compared to the pure g-C 3 N 4 NSs. Our work will provide a new platform to construct the ultra-small sized polyoxometalates clusters-based Z-scheme photocatalysts for enhancing the photocatalytic solar-to-fuels conversion.