Enhanced visible-light photocatalytic activity and photostability of Ag3PO4/Bi2WO6 heterostructures toward organic pollutant degradation and plasmonic Z-scheme mechanism

Novel Ag3PO4/Bi2WO6 heterostructured materials with enhanced visible-light catalytic performance were successfully synthesized by assembly combined with a hydrothermal treatment. The microstructures, morphologies, and optical properties of the prepared samples were characterized by multiple techniques. The irregular Ag3PO4 nanospheres dispersed on the surface of Bi2WO6 nanoflakes, and their catalytic performances were evaluated via the degradation of organic pollutants including rhodamine B (RB), methylene blue (MB), crystal violet (CV), methyl orange (MO), and phenol (Phen) under visible-light irradiation. The resulting Ag3PO4/Bi2WO6 heterostructured materials displayed higher photocatalytic activity than that of either pure Bi2WO6 or Ag3PO4. The enhanced photocatalytic activity was due to the good formation of heterostructures, which could not only broaden the spectral response range to visible light but also effectively promoted the charge separation. Meanwhile, the reasonable photoreactive plasmonic Z-scheme mechanism was carefully investigated on the basic of the reactive species scavenging tests, photoelectrochemical experiments, and photoluminescence (PL) spectrum. In addition, the excellent photostability of Ag3PO4/Bi2WO6 was obtained, which Ag formed at the early photocatalytic reaction acted as the charge transmission-bridge to restrain the further photoreduction of Ag3PO4.


Instruction
Recently, semiconductor photocatalysis has attracted more and more attention as a green, highly efficient technology for resolving the current energy and environmental problems. [1][2][3] Photocatalysts based on TiO 2 , ZnO are inactive under visible light irradiation due to their wide bandgap energy and improper band position, which requires ultraviolet irradiation that only accounts for less than 5% of the solar energy. [4][5][6] And the main reason for their poor photocatalytic activity might be due to the strong recombination of photogenerated electronhole (e À -h + ) pairs, which is a widespread phenomenon among semiconductor photocatalysts. Therefore, it is still challenging to extend the photo-response range and improve the charge separation efficiency of the semiconductor photocatalysts for satisfying the requirements of applications.
Recently, Bi-based semiconductor materials, such as Bi 2 WO 6 , BiVO 4 , Bi 2 MO 6 , Bi 2 Mo 2 O 9 , Bi 24 O 31 Br 10 , and BiOBr, have attracted widespread attention in photocatalytic application owing to the relatively narrow gap, nontoxicity, chemical inertness, stability, and sunlight utilization for wastewater treatment. [7][8][9][10][11][12][13][14] Among the Bi-based semiconductor materials, Bi 2 WO 6 , as a typical aurivillius oxide, is regarded one of the ideal photocatalyst materials because of its unique layered structure feature and relatively high visible photocatalytic activity. However, the photocatalytic activity of bare Bi 2 WO 6 is limited by the high recombination rate of photogenerated e À -h + pairs and the low absorption capacity of visible-light (less than 450 nm). 15,16 Therefore, how to improve separation efficiency of photoinduced e À -h + pairs during the photocatalytic reactions becomes a key issue for developing the Bi 2 WO 6 -based photocatalysts.
More recently, among the reported Ag-based semiconductors, especially silver orthophosphate (Ag 3 PO 4 ), which has a strong absorption (l < 530 nm), has been considered as a promising photocatalyst with outstanding visible-light catalytic activity. 17,18 Meanwhile, Ag 3 PO 4 possesses extremely oxidation power, which not only oxidizes H 2 O to produce O 2 but also degrades organic pollutants under visible-light irradiation. [19][20][21][22] However, there are some aws limit its extensive use as following: (1) Ag 3 PO 4 is slightly soluble in water due to its larger solubility product constant (K sp ¼ 1.6 Â 10 À16 ) 23 ; (2) as the photosensitive material, it can be easily corroded by light and reduced to Ag 0 in the photocatalytic process without electron acceptors; 24 (3) it suffers from the fast recombination of photogenerated charge-carries. 25 In the present study, Ag 3 PO 4 is introduced on the basis of the following points: (1) Ag 3 PO 4 is a narrow-semiconductor (E g ¼ 2.4 eV), and its introduction is benecial to extend the light response of Bi 2 WO 6 to a higher wavelength region and thereby decreasing its band gap; (2) two semiconductors possess the matching band structure, similar band gap, and the proper molar ratio, and build effective heterostructures, which promote the efficient separation of the e À -h + pairs and therefore increasing the quantum yield; (3) part Ag 3 PO 4 is reduced to metallic Ag during the initial photocatalytic reaction. It is precisely this part of the elemental silver can be used as a solidstate electron mediator to accelerate charge separation through Z-scheme system, which improve the charge separation, prevent the continuing reductive decomposition of Ag 3 PO 4 , and enhance its stability. Hence, the coupling of Bi 2 WO 6 with Ag 3 PO 4 to form heterostructures is considered to be one of the most promising methods to improve the visible-light-driven catalytic performance of photocatalysts (Bi 2 WO 6 or Ag 3 PO 4 ) because the heterostructures not only broaden the spectral response range to visible light but also promote the charge separation. This point is conrmed by the deposition of Ag 3 PO 4 onto certain semiconductor materials. Please see the Table 1 (ESI †). Table 1  Among a good deal of the heterojunction photocatalysts, the all-solid-state Z-scheme photocatalytic system possesses advantages over improving separation ratio of photo-induced e À -h + pairs and enhancing the stability of photocatalyst. In general, the all-solid-state Z-scheme photocatalytic system consists of two different semiconductor materials and an electron mediator. Notably, metallic Ag as a solid-state electron mediator would contribute to the construction of Z-scheme system. Recently, Yuan et al. successfully prepared Ag 3 PO 4 / CuBi 2 O 4 photocatalyst and formed Ag 3 PO 4 /Ag/CuBi 2 O 4 photocatalyst during the photocatalytic degradation process, which can enhance photocatalytic activity and stability of photocatalyst through the formation of Z-scheme system. 30 Liu et al. also designed a novel visible-light-driven Ag/Ag 3 PO 4 /WO 3 Zscheme heterostructures by a facile deposition-precipitation method followed by photoreducing Ag + into metallic Ag. The Ag/ Ag 3 PO 4 /WO 3 showed enhanced photocatalytic RhB efficiency, indicating the formed Z-scheme heterostructures could efficiently promote the separation and transfer of photogenerated e À -h + pairs. 31 So the combination of Bi 2 WO 6 and Ag 3 PO 4 to construct Z-scheme would be an efficient strategy.
To the best of our knowledge, there is less report on the achievement of Bi 2 WO 6 -based heterojunction by the introduction of Ag 3 PO 4 . For instance, Somchai Thongtem prepared Ag 3 PO 4 /Bi 2 WO 6 toward RB (5 ppm) degradation efficiency can reach 100% aer 80 min visible-light illumination. 32 Because it is still not sufficient for practical application. It is necessary to further improve the photocatalytic activity of Ag 3 PO 4 /Bi 2 WO 6 . In addition, there is no corresponding research reports to discuss the Ag 3 PO 4 /Bi 2 WO 6 for pollutants' removal based on the plasmonic Z-scheme mechanism. Therefore, it is meaningful to apply the SPR effect of Ag nanoparticles into the hybrid composite to design highly efficient visible-light-driven photocatalysts based on Z-scheme charge transfer mechanism.

Experiment section
2.1 Preparation of photocatalysts 2.1.1 Preparation of Bi 2 WO 6 . All the reagents used in the experiment were of analytic grades, commercially purchased and used without further purication. In a typical process, 0.972 g Bi(NO 3 ) 3 and 0.329 g Na 2 WO 4 were dissolved in glacial acetic acid (10 mL) and H 2 O (10 mL), respectively. Then the Na 2 WO 4 solution was dropwise added into Bi(NO 3 ) 3 solution. Aer being continuously stirred for 4 h, the mixture was transferred to Teon-coated autoclave and held at 150 C for 20 h followed by cooling at room temperature naturally. To remove any residue of by-products and reactants, the obtained Bi 2 WO 6 was washed with deionized water several times and dried at 80 C for 24 h.
2.1.2 Preparation of Ag 3 PO 4 /Bi 2 WO 6 . Ag 3 PO 4 /Bi 2 WO 6 composites with the different molar ratio were synthesized as follows. 1.092 g AgNO 3 and 0.762 g Na 3 PO 4 were dissolved in deionized water (10 mL) and stirred for 30 min, respectively. Then Na 3 PO 4 solution was dropwise added into the AgNO 3 solution to form yellow sediments with stirring for 2 h. Subsequently, a certain amount of as-prepared Bi 2 WO 6 was slowly added to the above mixture and stirred for 4 h followed by drying at 60 C for 24 h. Ultimately, the obtained Ag 3 PO 4 / Bi 2 WO 6 samples were collected by washing, ltration, and

Characterization of photocatalysts
The crystal structures were obtained on a Bruker-AXS (D8) X-ray diffractometer with Cu Ka radiation. X-ray photoelectron spectroscopy (XPS) characterization was carried out on an ESCALAB 250Xi spectrometer equipped with Al Ka radiation at 300 W. N 2 adsorption-desorption isotherm analysis of samples were obtained at 77 K using Micromeritics 3H-2000PS2. The morphologies of synthesized samples were analyzed using a scanning electron microscope (SEM) (Hitachi S-4300) and transmission electron microscope (TEM) and high resolution transmission electron microscope (HRTEM) (JEM-2100F). The UV-visible diffuse reectance spectra (UV-vis-DRS) were recorded using a UV-vis spectrophotometer (TU-1901) over the wavelength range of 200-800 nm using BaSO 4 as the reectance standard material. Fourier transform infrared (FT-IR) spectra were recorded using an FT-IR spectrophotometer (PE Company, America). Photoluminescence spectra (PL) were obtained by a Hitachi F-7000 spectrouorometer with an excitation wavelength of 360 nm, and all the samples were pressed into pellets in the sample holder.

Photocatalytic tests
Photocatalytic activities of the Ag 3 PO 4 /Bi 2 WO 6 heterostructures were studied by monitoring the degradation behaviors of organic contaminants, including RB, Phen, MO, CV, and MB.
The photocatalytic experiments were carried out in a hollow cylindrical photoreactor equipped with a water jacket. 400 W Xe lamp (l > 410.0 nm; moreover, the inner sleeve was made of No. 11 glasses to lter out ultraviolet from the Xe lamp) was used as the visible-light source. The Xe lamp was positioned within the inner part of the photoreactor and cooling water was circulated through a pyrex jacket surrounding the lamp to keep room temperature. In a typical experimental procedure, 200 mg of catalyst was placed into 220 mL of dye (50 ppm) or Phen (25 ppm) solution via ultrasonication for 10 min, and magnetically stirred in the dark for 1 h to establish the adsorption-desorption equilibrium between the catalyst and organic contaminant. Then the suspension was exposed to visible light irradiation under magnetic stirring. 4 mL of the suspension was collected at a regular time interval and analyzed aer centrifugation. The dye concentration was analyzed by UV-vis spectrophotometer (TU-1901) at the maximum absorption spectra. Changes of Phen concentrations were monitored by a Yilite P230II HPLC: C 18 column, UV detector (l ¼ 270 nm), methanol/water (50/50, v/v), and 1 mL min À1 .

Photoelectrochemical experiments
Electrochemical measurements were carried out in a traditional three-electrode system (CHI660E, China). Indium-tin oxide (ITO) glass electrode (1 cm 2 ), saturated calomel electrode (SCE), and Pt sheet were used as the working electrode, reference electrode, and the counter electrode, respectively. The sample mixed with Naon ionomer was dissolved ethanol aqueous solution to obtain 5 mg L À1 suspension. And then the suspension was uniformly drop-coated onto the clean ITO electrode surface and dried in air. A Xe lamp was used as the light source, and aqueous Na 2 SO 4 solution (0.01 mol L À1 ) served as the electrolyte. All the experiments were performed at room temperature (about 25 C). The photocurrent was measured under light illumination from a 400 W Xe lamp. The electrochemical impedance study was carried out over a frequency domain of 1 Hz to 100 kHz with a sinusoidal perturbation potential of 5 mV.

Characteristics of microstructures of Ag 3 PO 4 /Bi 2 WO 6
XRD is used to investigate the crystal structure of the samples. As displayed in Fig. 1 increases with increasing the Bi 2 WO 6 loading, while the peak intensity of Ag 3 PO 4 lowers simultaneously. However, the peak position of Ag 3 PO 4 does not signicantly change, which indicates Bi 2 WO 6 is not incorporated into Ag 3 PO 4 lattice. In addition, no diffraction peaks of Ag or other impurities are found. X-ray photoelectron spectroscopy (XPS) as surface analytic technique is employed to detect the electronic structures of Ag 3 PO 4 , Bi 2 WO 6 , and Ag 3 PO 4 /Bi 2 WO 6 composites. Fig. 2a shows the high-resolution XPS spectra of Ag of the Ag 3 PO 4 and Ag 3 PO 4 / Bi 2 WO 6 composite catalysts. The typical peaks of Ag 3d at 368.0 eV (Ag 3d 5/2 ) and 374.0 eV (Ag 3d 3/2 ) are ascribed to Ag + in Ag 3 PO 4 . 33 Aer the Bi 2 WO 6 introduction, the binding energy of spin-obit Ag 3d is divided into two peaks at 368.3 and 374.3 eV (ref. 24 and 34), which is 0.3 eV higher than those of Ag 3 PO 4 . The characteristic peaks of Ag nanocrystals existing on the surface of Ag 3 PO 4 and Ag 3 PO 4 /Bi 2 WO 6 are not found, suggesting no Ag 0 was formed during the preparation of the catalysts. Compared with pure Ag 3 PO 4 , the binding energy of P 2p of Ag 3 PO 4 /Bi 2 WO 6 is turned from 132.8 eV to a higher value of 133.8 eV (Fig. 2b). As shown in Fig. 2c, the peaks located at 159.3 eV, 164.6 eV, and 159.9 eV, 165.2 eV correspond to the Bi 4f 7/2 and Bi 4f 5/2 of Bi 2 WO 6 and Ag 3 PO 4 /Bi 2 WO 6 , respectively, implying the bismuth species in the composite is Bi 3+ cations. The peaks for W 4f 7/2 (35.98 eV) and W 4f 5/2 (37.98 eV) can be attributed to a six-valent oxidation state for W 6+ in Ag 3 PO 4 / Bi 2 WO 6 , meaning 0.2 and 0.1 eV deviation of 4f 7/2 and 4f 5/2 relative to the values in pure Bi 2 WO 6 (Fig. 2d).
The O 1s high-resolution XPS spectra of samples are provided in Fig. S1 (ESI †). The O 1s XPS spectra of Bi 2 WO 6 (Ag 3 PO 4 ) show three individual peaks with the binding energies of 530.2 eV (530.2 eV, denoted as O 1s (1)), 531.6 eV (531.3 eV, denoted as O 1s (2)), and 533.1 eV (533.1 eV, denoted as O 1s (3)), which can be attributed to the lattice oxygen in Bi 2 WO 6 (Ag 3 PO 4 ), the external hydroxyl groups, and adsorbed oxygen species at the surface of the composite. [35][36][37] For Ag 3 PO 4 /Bi 2 WO 6 -0.3, the binding energies of O 1s orbits are 529.9 eV, 531.4 eV, and 533.1 eV, respectively. And they also display more or less lower energy shi compared with individual counterparts. The variations of the binding energies of Ag 3d, P 2p, Bi 4f, W 4f, and O 1s are attributed to chemical interactions between Bi 2 WO 6 and Ag 3 PO 4 , which proves the formation of heterostructures and facilitates interfacial charge transfer, leading to the improved photocatalytic activity of Ag 3 PO 4 /Bi 2 WO 6 nanocomposites. Beyond that, the tight chemical interactions can improve the structural stability of photocatalyst. The similar results were also reported by another group. 38,39 FT-IR spectra were carried out to investigate the presence of Ag 3 PO 4 and Bi 2 WO 6 in the Ag 3 PO 4 /Bi 2 WO 6 composite. As shown in Fig. 3, the typical peaks located at 577. 8  The typical morphologies of Ag 3 PO 4 , Bi 2 WO 6 , and Ag 3 PO 4 / Bi 2 WO 6 composites were characterized by SEM images. From  Fig. 4a, Ag 3 PO 4 shows irregular spherical morphology with a diameter of 100-180 nm. While Bi 2 WO 6 exhibits a typical structure of nanoakes consist of nanoparticles with the side length of 50-250 nm (Fig. 4b). Fig. 4c illustrates the typical SEM image of Ag 3 PO 4 /Bi 2 WO 6 -0.3, where irregular spherical Ag 3 PO 4 nanoparticles disperse on the surface of Bi 2 WO 6 nanoakes. These tiny particles intertwine with each other, suggesting that Ag 3 PO 4 nanoparticles can restrain the agglomeration of Bi 2 WO 6 nanoakes.
The morphological and microstructural details of Ag 3 PO 4 / Bi 2 WO 6 -0.3 are obtained by TEM and HRTEM technique. As   displayed in the inset of Fig. 4d, the Ag 3 PO 4 irregular spheres disperse over the surface of Bi 2 WO 6 nanoakes, which is coincided with the aforementioned SEM observations. From the HRTEM image of Ag 3 PO 4 /Bi 2 WO 6 -0.3 (Fig. 4d), the lattice distances for the Bi 2 WO 6 (131) and Ag 3 PO 4 (211) facets are measured to be 0.315 and 0.245 nm, respectively, which is in well accordance with the XRD results. Furthermore, to further prove the formation of heterostructure, the energy dispersive spectroscopy (EDS) elemental mapping of Ag 3 PO 4 /Bi 2 WO 6 -0.3 was performed. As shown in Fig. S2(b-f) (ESI †), Ag 3 PO 4 and Bi 2 WO 6 are relatively uniform dispersion and well connected, which is in accordance with that of SEM and TEM images. Therefore, according to the above results, it gives solid evidence for the formation of heterostructure between Bi 2 WO 6 and Ag 3 PO 4 .
The BET specic surface area and the porous structure of the samples were analyzed by nitrogen adsorption-desorption isotherms, the results shown in Table 1. As shown in Fig. 5a, the N 2 isotherms of samples show characteristic type IV isotherms with H3 hysteresis loops, implying the presence of mesopores in the size of 2-50 nm. This result can be further proved by the pore size distribution analysis (Fig. 5b). And the specic surface area of pure Ag 3 PO 4 , Ag 3 PO 4 /Bi 2 WO 6 -0.3, Ag 3 PO 4 /Bi 2 WO 6 -0.5, and Bi 2 WO 6 is 8.06, 14.50, 18.21, and 43.55 m 2 g À1 , respectively. The BJH absorption cumulative pore volume increases from 0.038 to 0.070 cm 3 g À1 aer incorporation with Bi 2 WO 6 . The results imply the specic surface area and pore volume of Ag 3 PO 4 /Bi 2 WO 6 slight enhance with increasing Bi 2 WO 6 loading contrast with Ag 3 PO 4 . But the specic surface area and pore volume have no obvious change among these samples, which play a minor role in the enhanced photocatalytic activity of Ag 3 PO 4 /Bi 2 WO 6 composite.

Optical absorption properties
The optical absorbance properties and gap energies of the Ag 3 PO 4 , Bi 2 WO 6 , and Ag 3 PO 4 /Bi 2 WO 6 composites were determined using the UV-vis-DRS, and the results were displayed in Fig. 6. From Fig. 6a, the pure Bi 2 WO 6 exhibits strong absorbance in the wavelengths shorter than 450 nm owing to the intrinsic band-gap transition, 43 and Ag 3 PO 4 has the absorption edge at about 530 nm, which is consistent with the reported result. 44 Compared with Bi 2 WO 6 , the wavelength regions of the Ag 3 PO 4 /Bi 2 WO 6 composite are extended towards the visiblelight region. Moreover, the more obvious red shi phenomenon is observed with increasing the Ag 3 PO 4 loading, which implies more visible light is absorbed by the composite, and produce more e À -h + pairs, further improve the photocatalytic activity.
The band gap energies of the photocatalysts are calculated according to the following equation: In this equation, A, a, h, hn, and E g are constant, absorption coefficient, Planck constant, the energy of the incident photon, and band gap, respectively. And n is 1 and 4 for a direct and indirect band gap semiconductor, respectively. The value of n for Bi 2 WO 6 and Ag 3 PO 4 is 1 (ref. 20 and 45). By calculating, the band gaps of Bi 2 WO 6 and Ag 3 PO 4 are 2.90 and 2.45 eV, respectively. The band edge positions of photocatalysts can be determined by the empirical equations: where E VB is the valence band edge potential, E CB is the conduction band edge potential, X is the electronegativity of the semiconductor obtained from the geometric mean of the electronegativity for the component atoms (6.21 eV for Bi 2 WO 6 and 5.965 eV for Ag 3 PO 4 ), E e is the energy of free electrons on the hydrogen scale (4.5 eV), E g is the band gap energy of the semiconductor. Thus, E VB and E CB are calculated to be 3.34 eV and 0.44 eV for Bi 2 WO 6 and 2.69 eV and 0.24 eV for Ag 3 PO 4 , respectively.

Photocatalytic tests
Based on the above results, the photocatalytic activities of prepared photocatalysts were evaluated by degradation of RB under visible-light irradiation. As shown in Fig. 7a Fig. 7b shows the absorption spectra of RB (l max ¼ 553 nm) under visible-light irradiating the photoactive Ag 3 PO 4 /Bi 2 WO 6 -0.3. The absorption peak intensity of RB located at 553 nm reduces rapidly with prolonged irradiation time, and it is hardly observed aer 120 min visible-light illumination. The results are consistent with Fig. 7a. In addition, we nd the obvious blue shi of the absorption peak at 553 nm, which corresponds to the de-ethylation process. Therefore, the cleavage of the whole conjugated chromophore structure of RB leads to rapidly reducing the absorption peak intensity of RB, which indicates intermediate products were formed and then degraded to small molecules or CO 2 during the RB degradation process. 15,46 The pseudo-rst-order model based on the Langmuir-Hinshelwood (LH) kinetic model is mainly used to estimate the kinetics of the photocatalytic degradation of RB, as shown in the following equation: where k is the apparent rate constant of degradation; C t and C 0 are the initial and instantaneous concentration of RB at the time t 0 and t, respectively. Plots of Àln(C t /C 0 ) vs. time for photocatalysts are shown in Fig. 7c. All these photocatalysts display good linear relation meeting a pseudo-rst-order reaction, and the kinetic constants calculated are 0 min À1 (no photocatalyst), 0.00515 min À1 (Bi 2 WO 6 ), 0.00646 min À1 (Ag 3 PO 4 ), 0.0295 min À1 (Ag 3 PO 4 /Bi 2 WO 6 -0.3), and 0.0164 min À1 (Ag 3 PO 4 /Bi 2 WO 6 -0.5), respectively. During all photocatalysts, Ag 3 PO 4 /Bi 2 WO 6 -0.3 shows the highest photocatalytic activity and decomposition rate, which are 5.7 and 4.6 times than that of Bi 2 WO 6 and Ag 3 PO 4 , respectively. In short, Ag 3 PO 4 /Bi 2 WO 6 -0.3 exhibits the dramatical enhancement on photocatalytic activity of identical conditions than pure Bi 2 WO 6 and Ag 3 PO 4 . In order to eliminate the inuence of dye sensitization, Phen as the colorless compound is selected as a model molecule to further study the photocatalytic performance of Ag 3 PO 4 /Bi 2 WO 6 under visible-light irradiation. Because that Phen (l max ¼ 270 nm) has no absorption and no photosensitization in visiblelight region. Fig. 7d and e show the relative concentration (C t / C 0 ) and Àln(C t /C 0 ) vs. time curves of Phen in the presence of Bi 2 WO 6 , Ag 3 PO 4 , and Ag 3 PO 4 /Bi 2 WO 6 -0.3. As displayed in Fig. 7d, the direct photolysis of Phen is neglected during the whole visible-light irradiation, indicating Phen is also a stable pollutant. However, the degradation efficiency of Phen is about 7.1%, 63.5%, and 83.1% in the presence of Bi 2 WO 6 , Ag 3 PO 4 , and Ag 3 PO 4 /Bi 2 WO 6 -0.3, respectively. Their corresponding rate constants are obtained from Fig. 7e to be 0.00033 min À1 , 0.00634 min À1 , and 0.00999 min À1 , respectively. The results are consistent with those obtained for RB degradation. Meanwhile, the above results conrm that photocatalytic performance of Ag 3 PO 4 /Bi 2 WO 6 -0.3 is due to the excitation of the photocatalyst rather than the sensitization mechanism.
In order to test the extensive adaptability of the Ag 3 PO 4 / Bi 2 WO 6 photocatalyst, anionic dye (MO) and cationic dyes (RB, MB, and CV) are also employed as target pollutants. As depicted in Fig. 7f, MO, RB, MB, and CV are effectively eliminated aer visible-light irradiation, which implies that the Ag 3 PO 4 /Bi 2 WO 6 photocatalyst is highly efficient in the visible-light degradation of pollutants, especially cationic dyes. The excellent photocatalytic activity of Ag 3 PO 4 /Bi 2 WO 6 towards to organic pollutants is due to the good formation of heterostructures, which not only broaden the spectral response range to visible light but also effectively promote the charge separation. In addition, the excessive Ag 3 PO 4 may act as a recombination center, and cover the active sites on the Bi 2 WO 6 surface, leading to reducing the separation efficiency of the photogenerated charge carriers. It can be conrmed from the following PL results (Fig. S4, ESI †). That is why Ag 3 PO 4 /Bi 2 WO 6 -0.3 shows higher photocatalytic activity than Ag 3 PO 4 /Bi 2 WO 6 -0.5.
The photostability of a photocatalyst is essential for practical application. To research the reusability of photocatalysts, recycled photocatalytic degradation RB test was performed over Ag 3 PO 4 /Bi 2 WO 6 -0.3 composite. As shown in Fig. 8a, the photodegradation percentage of RB aer visible light irradiation for 90 min reduced from the original 92.9-86.5%, 75.8%, 71.5%, 68.5% and 68.1% aer six cycles. Although the degradation efficiency of RB in the reactive system went on 24.8%, the photodegradation percentage of RB was almost the same aer the ve and six cycles, indicating that the catalyst had a certain stability. The slight decline of photocatalytic efficiency is attributed to the inevitable loss of photocatalysts during the recycle runs. In addition, the color of RB-photocatalysts suspension had been changed from pink to black, indicating metallic Ag was formed on the surface of the catalyst during the photocatalytic process. This phenomenon is conrmed by the comparative XRD patterns of Ag 3 PO 4 /Bi 2 WO 6 before and aer the photocatalytic experiments in Fig. 8b. There is one weak peak located at 38.1 for Ag 3 PO 4 /Bi 2 WO 6 -0.3 aer cycles, which can be classied as the characteristic peak of metallic silver. 47 This part Ag formed at the early photocatalytic reaction. Meanwhile, UV-vis/DRS, the high-resolution XPS spectra of Ag 3d, and SEM image of Ag 3 PO 4 /Bi 2 WO 6 -0.3 aer reaction are also shown in Fig. S3 (ESI †). As displayed in Fig. S3(a), † Ag 3 PO 4 / Bi 2 WO 6 -0.3 displays enhanced photo-absorption in the visiblelight region aer the cycle degradation experiments. In particular, a broad prominent absorption in the visible-light region of 450-800 nm is observed, owing to the surface plasmon resonance (SPR) effect of Ag nanoparticles. 47 Fig. S3(b) † shows the Ag XPS spectra of Ag 3 PO 4 /Bi 2 WO 6 -0.3 aer cycles. The peaks located at 368.8 and 374.9 eV are assigned to Ag 0 , and the typical peaks located at 367.8 and 373.8 eV are ascribed to Ag + . According to the XPS results, the content of Ag 0 is about 30% aer cycles. As shown in Fig. S3(c), † although the aggregation phenomenon becomes more obvious aer the photocatalytic reaction, Ag 3 PO 4 /Bi 2 WO 6 -0.3 still maintain their morphologies. These results further prove that the photocorrosion resistance and stability of Ag 3 PO 4 were improved by introducing Bi 2 WO 6 to construct Ag 3 PO 4 -Ag-Bi 2 WO 6 Z-scheme heterostructures.

Photocatalytic mechanism discussion
3.4.1 Free radical and hole scavenging experiments. According to the photocatalytic literature, 48,49 the movement of the charge carriers such as e À and h + is the crucial parameter to generate reactive species such as cO 2À , cOH, and h + . Thus, active species generated during the photodegradation process of RB over the Ag 3 PO 4 /Bi 2 WO 6 -0.3 are identied by free radical and hole trapping experiments in the presence of various scavengers such as tert-butyl alcohol (t-BuOH, cOH scavenger, 1 mM), ethylenediaminetetraacetic acid disodium salt (EDTA-2Na, h + scavenger, 1 mM), and benzoquinone(BQ, cO 2À scavenger, 1 mM). Fig. 9 depicts the effects of various scavengers on the degradation of RB. The addition of t-BuOH can hardly inhibit RB degradation, which demonstrates that hydroxyl radicals cOH have little inuence in the photocatalytic process. However, when the EDTA-2Na and BQ were added into the reaction system, the photocatalytic activities of Ag 3 PO 4 /Bi 2 WO 6 -0.3 are obviously decelerated. Furthermore, the photodegradation percentage of RB is reduced from the original 100-6.6% and 45.1%, respectively. The above results suggested that h + and cO 2À are main active species during the degradation process.

Photoelectrochemical experiments.
Photoelectrochemical experiments as a useful technique to monitor the generation of photoelectrons and holes as well as their transfer and efficient separation in the Ag 3 PO 4 /Bi 2 WO 6 system were performed. As shown in Fig. 10, the transient photocurrent responses of Bi 2 WO 6 , Ag 3 PO 4 , and Ag 3 PO 4 /Bi 2 WO 6 -0.3 are recorded for several on-off cycles under visible-light irradiation at a bias of 1 V. It is found that the photocurrent responses of the tested sample working electrodes decrease to zero as soon as the lamp is turned off, and a rapid increase appears when the light is on. Meanwhile, photocurrent responses regain a reproducible value when the lamp is turned on again during on-off intermittent irradiation cycles. In addition, the Ag 3 PO 4 /Bi 2 WO 6 -0.3 exhibits the enhanced the photocurrent compared with Bi 2 WO 6 and Ag 3 PO 4 . The sequence of the photocurrent is in accordance with the order of the photocatalytic activity of photocatalysts. This suggests a smaller recombination and a more efficient separation of photogenerated e À -h + pairs occurs across the interface between Bi 2 WO 6 and Ag 3 PO 4 in the Ag 3 PO 4 /Bi 2 WO 6 composite.   Electrochemical impedance spectroscopy (EIS) is also employed to investigate the charge transfer resistance and the separation of photogenerated e À -h + pairs at solid/electrolyte interfaces in the photocatalyst. 50 Fig. 11 displays the EIS Nyquist plots of Bi 2 WO 6 , Ag 3 PO 4 , and Ag 3 PO 4 /Bi 2 WO 6 -0.3 under visible-light irradiation. It is clearly seen that the smallest arc radius of the EIS Nyquist plot of Ag 3 PO 4 /Bi 2 WO 6 -0.3 implies that it has the fastest interfacial electron transfer and more separation of photogenerated e À -h + pairs when compared to those of Bi 2 WO 6 and Ag 3 PO 4 . It is the reason for the Ag 3 PO 4 /Bi 2 WO 6 -0.3 composites exhibit the highest photocatalytic activity.
Photoluminescence (PL) spectrum is also an effective tool to reveal the migration, transfer and recombination processes of the photogenerated e À -h + pairs in semiconductors. Usually, a lower PL intensity indicates lower recombination rate of the charge carriers and higher photocatalytic activities of the photocatalysts. Fig. S4 (ESI †) shows the PL spectra of as-prepared Ag 3 PO 4 , Bi 2 WO 6 , and Ag 3 PO 4 /Bi 2 WO 6 composites with an excitation wavelength of 360 nm. As for the Bi 2 WO 6 -based materials, their PL intensities follow the order Bi 2 WO 6 > Ag 3 PO 4 / Bi 2 WO 6 -0.5 > Ag 3 PO 4 /Bi 2 WO 6 -0.3 > Ag 3 PO 4 . Compared with pure Bi 2 WO 6 , the PL intensities of Ag 3 PO 4 /Bi 2 WO 6 composites present an obvious decrease, implying a lower recombination feasibility of free charges in the Ag 3 PO 4 /Bi 2 WO 6 heterostructures. The Ag 3 PO 4 /Bi 2 WO 6 nanocomposite displays a higher PL intensity than Ag 3 PO 4 , implying that the migration pathways of the photoexcited e À -h + in the Ag 3 PO 4 /Bi 2 WO 6 nanocomposite is plasmonic Z-scheme theory (Scheme 1a) not as heterojunction energy-band theory (Scheme 1b). 51,52 From the Scheme 1a, under visible-light irradiation, both Bi 2 WO 6 and Ag 3 PO 4 can be excited to produce e À and h + simultaneously owing to their visible-light response. On the one hand, the part of Ag 3 PO 4 can be photoreduced to Ag 0 during the initial photocatalytic process. Ag nanoparticles (NPs) can absorb visible light and induce e À -h + pairs on account of the dipolar character and the SPR effect of the Ag NPs. The photogenerated electrons in the CB of Ag 3 PO 4 shi to the photogenerated holes produced by plasmonic absorption in the Ag NPs, which results in higher PL intensity. On other the hand, the plasmon hot electrons generated by the localized SPR oscillations of Ag NPs can capture the dissolved O 2 in water to form cO 2À (ref. 53). Meanwhile, the plasmon hot electrons directly transfer from the Ag NPs to the CB of Bi 2 WO 6 . While the photogenerated holes are still in the VB of Ag 3 PO 4 and Bi 2 WO 6 . Therefore, the photo-generated charge carriers are efficiently separated in space, which retards the photocorrosion of Ag 3 PO 4 . With the assistance of h + and cO 2À the two main active species, the organic pollutant is effectively degraded in aqueous solution. But from the heterojunction energy-band theory described in Scheme 1b, the VB potential (2.68 eV vs. NHE) of Ag 3 PO 4 is more negative than that of Bi 2 WO 6 (3.08 eV vs. NHE), which leads to the migration of the h + from Bi 2 WO 6 to Ag 3 PO 4 . Since the CB potential (0.24 eV vs. NHE) of Ag 3 PO 4 is more negative than that of Bi 2 WO 6 (0.44 eV vs. NHE), photo-induced electrons from Ag 3 PO 4 migrate to the CB of Bi 2 WO 6 and then transfer to Ag 0 . Therefore, the photo-generated electrons cannot reduce O 2 to produce cO 2À , due to the CB potential of Bi 2 WO 6 being more positive than the redox potential of cO 2À formation (O 2 /cO 2À ¼ À0.33 eV, NHE). 51 Thus this phenomenon cannot explain the stronger effect of cO 2À on the degradation of RB. In conclusion, the plasmonic Z-scheme theory for the photocatalysis of Ag 3 PO 4 /Bi 2 WO 6 is much more reasonable.

Conclusions
Novel Ag 3 PO 4 /Bi 2 WO 6 heterostructured materials were successfully synthesized by assembling Ag 3 PO 4 irregular nanospheres on the surface of Bi 2 WO 6 nanoakes. Ag 3 PO 4 /Bi 2 WO 6 -0.3 showed obviously superior visible-light catalytic activity toward degradation of organic pollutants. Free radical and hole scavenging experiments suggested h + and cO 2À are two main active species through the degradation process. In brief, the enhanced photocatalytic activity was due to the good formation of heterostructures, which could not only broaden the spectral response range to visible light but also effectively promoted the charge separation. These results were solidly conrmed by photocurrent responses, electrochemical impedance spectroscopy, and photoluminescence spectrum. In addition, Ag 3 PO 4 / Bi 2 WO 6 -0.3 exhibited relative higher photostability toward RB degradation, which was explained by the reasonable photoreactive mechanism. This work provides the potential application of Bi 2 WO 6 -based heterostructures as efficient visible light responsive catalysts for environmental remediation.

Conflicts of interest
There are no conicts to declare.