Binary Type-II Heterojunction K7HNb6O19/g-C3N4: An Effective Photocatalyst for Hydrogen Evolution without a Co-Catalyst

The binary type-II heterojunction photocatalyst containing g-C3N4 and polyoxoniobate (PONb, K7HNb6O19) with excellent H2 production activity was synthesized by decorating via a facile hydrothermal method for the first time. The as-fabricated Nb–CN-0.4 composite displayed a maximum hydrogen evolution rate of 359.89 µmol g−1 h−1 without a co-catalyst under the irradiation of a 300 W Xenon Lamp, which is the highest among those of the binary PONb-based photocatalytic materials reported. The photophysical and photochemistry analyses indicated that the hydrogen evolution performance could be attributed to the formation of a type-II heterojunction, which could not only accelerate the transfer of photoinduced interfacial charges, but also effectively inhibit the recombination of electrons and holes. This work could provide a useful reference to develop an inexpensive and efficient photocatalytic system based on PONb towards H2 production.


Introduction
A looming global energy shortage makes it increasingly urgent to develop environmentally friendly green chemical technology for energy production to ensure the sustainable development of human society [1][2][3]. In recent decades, much effort has been devoted to developing the production of hydrogen, not merely because it is an important chemical raw material; more importantly, hydrogen energy is recognized as the most ideal and promising clean energy of the future [4]. Due to its environmental friendliness, efficient cost performance and non-polluting nature, photocatalytic technology has been widely used to produce such green energy [5]. Photocatalysis water splitting exhibits a great potential in hydrogen production and is expected to achieve industrialization [6,7]. However, the development of low-cost and highly efficient photocatalysts for hydrogen production is still a major challenge for practical application.
Polyoxometalate (POM) is a kind of metal oxide cluster based mainly on Mo, W, V, Nb and Ta elements [8], which have been widely studied due to their promising applications in fields such as water oxidation, hydrogen evolution, carbon dioxide reduction, and nitrogen reduction reaction [9]. As a special subclass of POMs with the unique electronic characteristics of high charge density and rapid electron transfer rate, polyoxoniobates (PONbs) have been employed to assemble photocatalytic composites with other semiconductor species [10]. It is worth mentioning that the Lindqvist type K 7 HNb 6 O 19 is a favorite candidate to be selected to be as a model of PONbs to investigate the construction of PONbbased hydrogen-evolution photocatalysts, for its easy preparation and high stability [11]. However, immobilization and reusability are a tremendous challenge for such materials because of the water solubility, highly alkalinity, and confined working pH region of the components of K 7 HNb 6 O 19 [12,13].
Graphitic carbon nitride (g-C 3 N 4 , CN) is reported to be a promising photocatalyst for hydrogen evolution because of its appropriate band gap and energy structure being capable of performing water reduction reaction [14,15]. However, the photocatalytic hydrogen production activity of pure g-C 3 N 4 is confined by the rapid electron-hole recombination and tardy charge mobility, leading to the low photocatalytic activity [16,17]. A large number of researchers have shown that the photocatalytic activity of g-C 3 N 4 can be enhanced by morphology regulation, element doping, dye sensitization, heterojunction construction, various noble metals loading (Pt, Au), and other methods [18][19][20][21][22]. Among them, two strategies of loading noble metals and constructing heterojunctions have been employed widely as effective ways to boost the catalytic property of g-C 3 N 4 [23]. Nevertheless, constructing heterojunctions by combining with other semiconductors is a preferred method to fabricate g-C 3 N 4 -based photocatalysts on account of the high scarcity and cost of noble metals [24]. It was proven that the combination of g-C 3 N 4 and POMs is an effective strategy to obtain heterojunctions with improved photoactivity, which not only gives them a large specific surface and a narrowed band gap, but also could accelerate the photoinduced interfacial charge transfer and inhibit effectively the recombination of the electrons and holes in the process of photocatalysis. For example, a series of type II heterojunction composites (POMs/C 3 N 4 ) were reported based on Keggin-type polyoxoanions SiW 12 [12,[29][30][31]. In this work, a series of K 7 HNb 6 O 19 /g-C 3 N 4 photocatalysts were fabricated via a facile hydrothermal method, which was used for effective H 2 generation without any co-catalyst. To the best of our knowledge, this is the first study constructing a binary K 7 HNb 6 O 19 /g-C 3 N 4 type-II heterojunction photocatalyst, and it possesses outstanding photocatalytic H 2 evolution properties under ultraviolet light driving.

Chemicals and Reagents
All chemicals were reagent grade and used without further purification. Niobium pentoxide (Nb 2 O 5 , AR), methyl alcohol (MeOH, AR), potassium hydroxide (KOH, AR), sodium sulfate (Na 2 SO 4 , AR), ethanol (EtOH, AR) and urea (CO(NH 2 ) 2 , AR) were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China).  [32]. Typically, 13.3 g Nb 2 O 5 and 26 g KOH were added into the nickel crucible and mixed evenly. The resultant mixture was heated in a muffle furnace at 480 • C for 50 min; then the product was collected and dissolved in 100 mL boiling water after it was cooled to room temperature. When the volume of the solution reached about 50 mL by evaporating water, the heating was stopped. The white solids were collected and dried in vacuum.

Synthesis of g-C 3 N 4
The g-C 3 N 4 was prepared through the direct calcination of urea according to the methods reported in the literature [33]. Typically, urea was placed evenly in a covered crucible and heated at 550 • C for 3 h in a tube furnace, with a heating rate of 10 • C min −1 . The resultant product was taken out when the temperature cooled to about 30 • C and then ground for use.

Synthesis of Nb-CN-X Composites
Nb-CN-X composite photocatalysts were prepared according to the following process, where X refers to the mass of g-C 3 N 4 added based on 1 g K 7 HNb 6 O 19 as a benchmark; X = 0.2, 0.4, 0.6, 0.8. Typically, taking Nb-CN-0.4 as an example, as-obtained 0.4 g g-C 3 N 4 was added to a solution containing 1 g Linquist type K 7 HNb 6 O 19 and 15 mL deionized water and stirred for 3 h to form yellow mixed suspension, and then it was transferred to a 25 mL Teflon-lined reactor and heated at 180 • C for 12 h. The resulting precipitate was collected and washed separately with high-purity water and ethanol three times and then dried in a vacuum oven at 60 • C for 12 h to obtain the white photocatalyst Nb-CN-0.4 composite. Similarly, other composites were also synthesized by adjusting the amount of g-C 3 N 4 (X = 0.2, 0.6, 0.8) according to the same process. The preparation process is shown in Figure 1. The detailed characterization methods and analytical methods are described in the Supporting Information.

Morphology and Structure Analysis
The crystalline phase of the as-prepared K 7 HNb 6 O 19 , g-C 3 N 4 and Nb-CN-X composites were characterized by X-ray diffraction (XRD). As shown in Figure 2, the diffraction peaks of pure g-C 3 N 4 are 13.3 • and 27.5 • , while the original K 7 HNb 6 O 19 has sharp diffraction peaks at 9.8 • , 26 • and 48 • , which is consistent with the reported results [30,34]. There are three distinct characteristic peaks located at the positions of 8.8 • , 27.5 • and 46 • for Nb-CN-X composites. Among them, the peak at 27.5 • matches with the (002) crystal plane of g-C 3 N 4 , indicating that g-C 3 N 4 has been successfully introduced. However, the diffraction peak corresponding to the (100) crystal plane of g-C 3 N 4 at 13.3 • cannot be observed clearly in the Nb-CN-X composites with the introduction of K 7 HNb 6 O 19 , which may be due to the low content of g-C 3 N 4 . Concurrently, the sharp XRD pattern signals from the polyoxoniobate are not visible in the binary composites, while two new wide peaks at 8.8 • and 46 • are observed, which could be attributed to the peaks of K 7 HNb 6 O 19 shifted from 9.8 • and 48 • . This phenomenon might be caused by the coupling between the two components K 7 HNb 6 O 19 and g-C 3 N 4 in the composites [35]. In addition, with the increasing of the content of g-C 3 N 4 in the Nb-CN-X composites, the peak intensity of K 7 HNb 6 O 19 decreased little by little, while that of g-C 3 N 4 increased gradually, showing the successful composite of the two components in Nb-CN-X.
To give the further information of the composite photocatalysts, the FT-IR spectra of the K 7 HNb 6 O 19 , g-C 3 N 4 and Nb-CN-0.4 composite were carried out as representatives. As shown in Figure 3, for Nb-CN-0.4 composite, the three vibrational absorption peaks observed at 853, 674 and 530 cm −1 belonged to the tensile vibration of Nb-O and Nb-O-Nb of K 7 HNb 6 O 19 [24,36]. The characteristic peaks of g-C 3 N 4 in Nb-CN-0.4 appeared at different wavebands in the absorption region. The wide peak between 3000 and 3600 cm −1 was caused by the stretching vibration of N-H and O-H of the physical absorption water molecules [37]. The peaks at 1630 cm −1 are attributed to the stretching vibration of C=N in g-C 3 N 4 , while the peaks at 1405, 1321, 1246 and 807 cm −1 belonged to a typical C-N heterocyclic stretching vibration for triazinyl units [23,38]. Combined with the results of XRD and FT-IR, the conclusion can be drawn that under hydrothermal conditions, K 7 HNb 6 O 19 was successfully embedded in g-C 3 N 4 , rather than simply mixed, to form the final hybrid material Nb-CN-0.4 composite.  The morphology and structures of photocatalysts K 7 HNb 6 O 19 , g-C 3 N 4 and Nb-CN-0.4 composite were investigated by SEM, TEM and mapping techniques. It is generally known that the Lindqvist type K 7 HNb 6 O 19 is made up of six co-planer Nb-O octahedrons to form a super octahedron, as shown in Figure S1, leading to a highly symmetric compact structure [39]. Figures 4a and S2 reveal that the pure K 7 HNb 6 O 19 showed a transparent rodlike morphology. The pure g-C 3 N 4 obtained by annealing urea exhibited a large piled sheet shape (Figure 4b), which was caused by thermal polycondensation of urea [40]. The SEM image of Nb-CN-0.4 composite (Figure 4c) presented a thin nanosheet morphology, just like that of the pure g-C 3 N 4 but with better dispersion. However, K 7 HNb 6 O 19 couldn't be observed in the SEM image, maybe due to its highly uniform distribution caused by the good solubility of K 7 HNb 6 O 19 under hydrothermal conditions. To further confirm the existence of K 7 HNb 6 O 19 , element mapping was performed. As shown in Figure 5, C, O, Nb, K and N elementals were distributed uniformly, further proving the formation of the heterogeneous structure containing K 7 HNb 6 O 19 and g-C 3 N 4 . Exposing the individual components (pristine g-C 3 N 4 and K 7 HNb 6 O 19 ) to the hydrothermal condition is of impor-tance in order to verify the reason causing morphology changes of the achieved composite catalysts in the hydrothermal process. Therefore, we treated the individual component K 7 HNb 6 O 19 at 180 • C for 12 h, and found it was still a clear and transparent aqueous solution. After evaporating, the obtained crystals showed an identical IR to that of the freshly prepared K 7 HNb 6 O 19 ( Figure S3), indicating that K 7 HNb 6 O 19 can exist stably under the hydrothermal condition. For the pristine g-C 3 N 4 , as seen in Figure S4, the hydrothermal treatment made its morphology change from an original sheet shape to the extremely irregular rodlike shape. Based on the comparison of the morphology of g-C 3 N 4 before and after hydrothermal treatment, it can be concluded that the presence of K 7 HNb 6 O 19 could positively induce the morphology change of the obtained composite, which is consistent with our previous reports [35,36]. The results of HRTEM measurement showed that the introduction of K 7 HNb 6 O 19 made the Nb-CN-0.4 composite possess a more regular ultrathin nanosheet structure compared to that of the pure g-C 3 N 4 (Figure 4d-f). Ultimately, a number of nanosheets stacked together to further form a flower-like shape unit of the composite (Figure 4f). For the photocatalyst Nb-CN-0.4 composed of ultrathin nanosheets, an improved charge transfer efficiency could be expected compared with pure K 7 HNb 6 O 19 [41], which is consistent with good photocatalytic performance.  The surface chemical status of K 7 HNb 6 O 19 , g-C 3 N 4 and Nb-CN-0.4 composite was studied by X-ray photoelectron spectroscopy (XPS). As shown in Figure S5, the full XPS spectra clearly show the existence of C, Nb, N, K and O elements in the composite sample, indicating that the hybrid material had all the characteristic elements of K 7 HNb 6 O 19 and g-C 3 N 4 . In the C 1s spectrum (Figure 6A), the main characteristic peak centered at 284.8 eV could be attributed to that of sp 2 C-C or C=C bonds, which derived from the amorphous reference carbon on the surface. For the pure g-C 3 N 4 , the peak centered at 288.3 eV corresponded to the sp 2 carbon atoms bonded to N (N=C-(N) 2 ) in the graphitic structure [42,43]. It should be noted that the C 1s peak centered at 286.2 eV in Nb-CN-0.4 showed a distinct shift toward the lower band energy direction after hydrothermal treatment compared with the pure g-C 3 N 4 , indicating the presence of strong interactions between the terminal oxygen atom of Nb=O or the bridge oxygen of Nb-O-Nb from K 7 HNb 6 O 19 and the N=C-(N) 2 group from g-C 3 N 4 . In addition, the strong electronic pull of the Linqvist anions might also have affected the position of the C 1s peak [30]. The three N 1s peaks in Nb-CN-0.4 composite were observed at the positions 399.8, 400.9 and 401.8 eV, with the intensity slightly greater than those of g-C 3 N 4 , corresponding to the peaks of sp 2 N atoms involved in tris-triazine rings (C=C-N), tertiary nitrogen (N-(C) 3 ) and N atoms bonded with H atoms (C-N-H) in the aromatic rings, respectively [44]. It can be seen from Figure 6c that three peaks, centered at 528.9, 530.6 and 531.9 eV, were observed on the K 7 HNb 6 O 19 sample, which were assigned to Nb-O-Nb, Nb-O-H and adsorbed water, respectively [24]. After hydrothermal treatment, the O 1s in Nb-CN-0.4 had two main peaks, of which the peak at 532.1 eV was adsorbed water, while the other one resulted from the binding energy of different oxygen-containing species. Compared with the pure K 7 HNb 6 O 19 , the binding energy peaks located at 206.8 and 209.5 eV corresponded to Nb 3d 5/2 and 3d 3/2 in the Nb-CN-0.4 composite, which indicates that Nb is in its highest oxidation state (+5 valence state) [24] (Figure 6d), and the shift of Nd 3d binding energy further confirmed that there was an interaction between the Lindqvist unit of the K 7 HNb 6 O 19 sample and g-C 3 N 4 sample. Figure S6 shows that K 2p binding energies did not change significantly in K 7 HNb 6 O 19 and the composite, except for their intensities. The BET tests of K 7 HNb 6 O 19 , g-C 3 N 4 and Nb-CN-0.4 composite were carried out, and the N 2 adsorption-desorption isotherms are shown in Figure S7. The calculated specific surface areas of K 7 HNb 6 O 19 and g-C 3 N 4 were about 3.15 m 2 g −1 and 52.13 m 2 g −1 , respectively, while that of Nb-CN-0.4 composite increased to 55.34 m 2 g −1 after the introduction of g-C 3 N 4 . The larger specific surface area is conducive to the enhancement of the photocatalytic activity [45], which would promote more water molecule adsorption on the surface of the catalysts. Meanwhile, it is also beneficial for the increase of catalytic active sites.

Optical and Electric Properties
Raman spectroscopy of K 7 HNb 6 O 19 , g-C 3 N 4 and Nb-CN-0.4 composite are depicted in Figure 7. There were two obvious peaks at 759 and 901 cm −1 for Nb-CN-0.4 composite, which were attributed to the terminal oxygen and bridging oxygen in K 7 HNb 6 O 19 [46]. The broad band in 400-700 cm −1 was attributed to g-C 3 N 4 in Nb-CN-0.4 composite. As for the pure g-C 3 N 4 , the Raman peaks located at 592 cm −1 , 725 cm −1 , 1055 cm −1 , 1165 cm −1 , 1296 cm −1 , 1441 cm −1 and 1550 cm −1 were related to CN heterocyclic stretching vibration, of which the last two in the high wavenumber region were attributed to D and G bands of g-C 3 N 4 of a typical graphitic structure, respectively [47,48]. Compared to the bare K 7 HNb 6 O 19 and g-C 3 N 4 , the Raman peaks of Nb-CN-0.4 composite became broader, further proving the formation of the hybrid material and matching well with the results of IR and X-ray powder diffraction analysis. The optical absorption properties of photocatalysts K 7 HNb 6 O 19 , g-C 3 N 4 and Nb-CN-X composites were characterized by UV-vis DRS ( Figure S8). As shown in Figures S6 and 8a, K 7 HNb 6 O 19 only absorbed ultraviolet light, and the absorption edge was about 310 nm, while all the composites showed a slightly broader absorption after introducing C 3 N 4 , which was undoubtedly attributed to the contribution of g-C 3 N 4 with the absorption band edge around 450 nm [49]. The band gap values of as-prepared photocatalysts were calculated according to the equation of (αhυ) 2 = A(hυ − E g ), where α, h, υ and A represent the absorption coefficient, Plank constant, light frequency and a constant, respectively [50]. As described in Figure 8b, the band gap value E g of the pure K 7 HNb 6 O 19 and g-C 3 N 4 were about 3.98 and 2.73 eV, respectively, which were consistent with literature reports [29,51]. The E g value of Nb-CN-0.4 was about 3.42 eV, slightly narrower than that of K 7 HNb 6 O 19 , indicating the presented composite is an ultraviolet light photocatalyst without activity under visible light irradiation. To study the separation efficiency of the photogenerated carriers of the samples, photoluminescence spectroscopies (PL) of g-C 3 N 4 and Nb-CN-0.4 composite were performed. The fluorescence intensity of each sample was studied at the same excitation wavelength of 300 nm. Generally, a lower PL intensity indicates a lower recombination rate of the photoexcited electrons and holes, meaning a higher photocatalytic activity [52]. As shown in Figure S9, the introduction of K 7 HNb 6 O 19 can partially quench the fluorescence of g-C 3 N 4 , which means that the introduction of K 7 HNb 6 O 19 promotes the separation of photogenerated carriers, which is consistent with its good photocatalytic hydrogen production activity.
The photocurrent transient response experiment (Figure 9a) showed that the Nb-CN-0.4 had the highest photocurrent intensity, much higher than the original g-C 3 N 4 and K 7 HNb 6 O 19 , implying that Nb-CN-0.4 had the faster transfer efficiency and more efficient separation of photogenerated charges [53]. Moreover, the electrochemical impedance spectroscopy Nyquist plot was used to analyze the charge transfer rate (Figure 9b). The smallest arc radius of Nb-CN-0.4 under the same test conditions implied that it had the highest separation and migration rate of the photo-generated carriers [54], which was in keeping with the results of photocurrent response and photocatalytic performance studies.

Photocatalytic H 2 Production
The photocatalytic performance of hydrogen production of K 7 HNb 6 O 19 , g-C 3 N 4 and Nb-CN-X composites was tested using methanol (MeOH) as a sacrificial agent under a 300 W Xenon Lamp without a co-catalyst. Herein, methanol was selected as the sacrificial agent because it could act mainly and preferentially as a hole scavenger to decrease the recombination rate of photogenerated charge carriers for enhanced photocatalytic H 2 evolution efficiency, similar to our previous work [35,55]. Typically, 50 mg photocatalyst powder was added into the mixed solution of 40 mL deionized water and 10 mL methanol, and then stirred and sonicated for 15 min to ensure the uniform dispersion of the catalyst; the concentration of the solid catalyst was 1mg/mL. As shown in Figure 10, the average hydrogen evolution rates were 2.72 µmol g −1 h −1 for K 7 HNb 6 O 19 and 9.05 µmol g −1 h −1 for g-C 3 N 4 . After combining K 7 HNb 6 O 19 with g-C 3 N 4 , Nb-CN-0.4 composite exhibited the highest hydrogen production rate, as much as 359.89 µmol g −1 h −1 , approximately 133 and 40 times of the original K 7 HNb 6 O 19 and g-C 3 N 4 , respectively. To the best of our knowledge, the maximum rate of hydrogen production reported is 207.6 µmol h −1 g −1 for such kinds of photocatalytic materials in the literature (Table S1). Therefore, Nb-CN-0.4 composite represented the binary PONb-based photocatalytic material with the highest hydrogen production rate up to now. Figure 10b showed that with the amount of g-C 3 N 4 increasing, the H 2 producing activity of Nb-CN-X composites increased gradually to a maximum value and then decreased. Although the coupling of K 7 HNb 6 O 19 and g-C 3 N 4 could effectively increase the hydrogen production efficiency and significantly improve the photocatalytic performance, the excessive g-C 3 N 4 would bring about a shielding effect and introduce more recombination centers, which was not conducive to the charge separation, leading to a low performance of photocatalytic hydrogen production. To verify the stability of the prepared photocatalyst, the recyclability performance of the optimal sample of Nb-CN-0.4 composite was evaluated under the same condition. As shown in Figure 11a, after four recycling experiments, the hydrogen evolution rate did not decrease obviously, indicative of a high stability of the sample. Moreover, the crystallinity of the catalyst was retained well after four rounds of recycling, which could be drawn from the XRD pattern in Figure 11b, further proving the good structural and catalytic stability of the prepared composite.

Investigation of Photocatalytic Mechanism
The valence band XPS spectra test was carried out to study the photocatalytic mechanism ( Figure S10). The estimated valence band (VB) positions of K 7 HNb 6 O 19 and g-C 3 N 4 were 3.61 and 1.67 eV vs. NHE, respectively, which were consistent with the reported values [36]. The conductance band (CB) positions (E CB ) of K 7 HNb 6 O 19 and g-C 3 N 4 were calculated to be −0.37 and −1.06 eV vs. NHE based on the empirical formula of E g = E VB − E CB . Based on the above analysis and the photocatalytic experimental results, we proposed a possible mechanism of photocatalytic hydrogen production, as seen in Figure 12. On the basis of the definition of a heterojunction [56], a type-II heterojunction photocatalytic system could be formed between K 7 HNb 6 O 19 and g-C 3 N 4 due to a staggered band gap. Under the full spectral irradiation, the electrons were excited from VB to CB of both g-C 3 N 4 and K 7 HNb 6 O 19 . Because the CB potential of g-C 3 N 4 was lower than that of K 7 HNb 6 O 19 , the photo-induced electrons on the CB of g-C 3 N 4 could migrate more easily to the CB of K 7 HNb 6 O 19 , while the photo-induced h + were concentrated on the VB of g-C 3 N 4 simultaneously. For the photocatalytic H 2 production, the electrons accumulated on the CB of K 7 HNb 6 O 19 could be captured by H + to produce H 2 , and the holes in the VB of g-C 3 N 4 were consumed by the sacrificial agent CH 3 OH. The mechanism of photocatalytic hydrogen evolution is shown in Figure 12, and the pathways can be illustrated as follows: Nb-CN-X + hυ → Nb-CN-X + e − + h + (1) h + + CH 3 OH → CH 3 OH + (3)

Conclusions
In summary, a series of binary type-II heterojunction Nb-CN-X composites were successfully prepared via a one-step hydrothermal method. The obtained optimal photocatalyst Nb-CN-0.4 composite displayed superior photocatalytic H 2 generation activity, with a H 2 generation rate of 359.89 µmol g −1 h −1 . The binary type-II heterojunction photocatalyst containing g-C 3 N 4 and polyoxoniobate (PONb, K 7 HNb 6 O 19 ) with excellent H 2 production activity under ultraviolet light driving was synthesized by decorating via a facile hydrothermal method for the first time and without any co-catalyst, which was approximately 133 and 40 times than that of bare K 7 HNb 6 O 19 and g-C 3 N 4 . The enhanced photocatalytic hydrogen evolution was ascribed to a good heterojunction formed between K 7 HNb 6 O 19 and g-C 3 N 4 , which significantly accelerated the charge carries transfer rate on the interface and improved the separation of photogenerated electrons and holes, as demonstrated by the result of photochemical tests. It is worth mentioning that Nb-CN-X composites are the first example of binary type-II heterojunction K 7 HNb 6 O 19 /g-C 3 N 4 photocatalyst to H 2 evolution. This work could provide reference for the preparation and modification of PONb-based photocatalysts with low cost and high performance, as well as their further application in clean energy.

Conflicts of Interest:
The authors declare no conflict of interest.