Construction of ZnO/Keggin Polyoxometalate Nano-heterojunction Catalyst for Efficient Removal of Rhodamine B in Aqueous Solution

Rational design of nano-heterojunction catalyst is one of the ideal ways to solve dye pollution effectively. In this study, two nano-heterojunction catalysts ZnO/Ag4SiW12O40 (ZnO/AgSiW) and ZnO/Cs3PW12O40 (ZnO/CsPW), were synthesized by the simple dissolution and precipitation method. The samples were characterized by XRD, TEM, FT-IR, Raman, DRS, XPS and N2 adsorption and desorption isotherms. The photocatalytic experiments show that the degradation rates of Rhodamine B (RhB) by ZnO/AgSiW and ZnO/CsPW within 60 min are 92.3 and 72.7%, respectively, which are 5.3 and 4.2 times higher than that by pure zinc oxide. The two catalysts still maintain excellent performance after repeated use for three times. Based on the free radical capture experiment, the active substances that play a major role in the photocatalytic process were determined. In addition, the electrochemical tests show that the construction of a nano-heterojunction system significantly increases the separation efficiency of photogenerated carriers in the heterojunction system. The possible photocatalytic degradation pathways and final products of ZnO/AgSiW and ZnO/CsPW heterojunction nanocomposite catalysts were analyzed by the liquid chromatography/mass spectrometry (LC/MS) technique. This paper provides a theoretical and experimental basis for treating organic dyes by ZnO/AgSiW and ZnO/CsPW nano-heterojunction catalysts.


Introduction
With the rapid development of industry, environmental problems are becoming increasingly serious. Among them, water pollution is a difficult problem, and the dye wastewater produced by the textile industry is one of the major sources of polluted water. Dyes are pollutants high molecular weight, complex structures and low biodegradability, while some dyes also have carcinogenicity and mutagenicity [1,2]. Therefore, it is particularly important to find a method to deal with dye pollution. At present, photocatalytic technology is considered a clean, efficient and promising method to deal with dye pollution. Compared with traditional methods, it will not cause secondary pollution and is cheap, and only photocatalyst and ubiquitous light are needed in dealing with pollutants [3,4].
In general, most semiconductor materials, such as TiO 2 [5], WO 3 [6], ZrO 2 [7], ZnO [8], g-C 3 N 4 [9] and SiO 2 [10], are considered to be ideal photocatalytic materials and are widely used in the field of photocatalysis because of their excellent properties [11]. Among these semiconductor materials, zinc oxide, as a wide-gap material, has been widely studied and applied as a kind of photocatalytic material because of its stable physical and chemical properties, low cost and non-toxicity [12,13]. However, as a homogeneous catalyst, its wide-band gap leads to inefficient utilization of visible light and a high recombination rate of photogenerated carriers, which affects its photocatalytic activity [14]. A lot of work has been done to reduce the wide-band gap of ZnO and decrease the recombination rate of photogenerated carriers, including doping ions into ZnO, deposition of precious metals, compounding ZnO with semiconductor materials and constructing heterojunction systems [15,16]. For example, Liu et al. designed Ag/Fe doped ZnO composites through the controllable integration of three strategies: layered assembly, Fe doping and Ag loading, in which Fe doping and Ag loading jointly prolonged the visible light absorption and promoted the separation of photogenerated carriers [17]. Wang et al. successfully synthesized MOF-derived C-doped ZnO/TiO 2 composites by microwave hydrothermal method and controlled calcination method, which shows greater catalytic performance than pure ZnO and TiO 2 [18]. N. Srinivasan et al. synthesized C-ZnO/BiVO 4 heterostructure nanocomposites through BiVO 4 and carbon-doped ZnO nanoparticles. The formation of the heterojunction system turned the band gap of ZnO to the visible region and inhibited the electron-hole recombination [19].
Heteropoly acid is a catalyst with rich composition, different structures and functions. It has reversible redox properties and can transfer and store electrons and protons [7,20]. As an efficient solid acid catalyst, it has been widely studied [21]. However, heteropoly acids also have deficiencies such as high solubility and difficult to recover, which limits their application as photocatalysts in removing organic pollutants in aqueous media [22]. Generally, loading heteropoly acids on recyclable carriers or using salts containing monovalent cations (such as K + , Cs + , Ag + , NH 4 + , etc.) are two methods that can effectively improve the solubility of heteropoly acids [23,24]. For example, A.M. Hussein et al. fixed the synthesized phosphomolybdic acid (PMo) and vanadium containing phosphomolybdic acid (PV 2 Mo) on HZSM-5 zeolite by the impregnation technique [25]. Alotaibi et al. proposed a heterogeneous method to improve the catalytic capacity of tungsten phosphoric acid by adding 2.5 parts of cesium salt, which significantly increases the surface area and reduces the solubility of heteropoly acids in polar solvents [26]. Zhu et al. prepared silver exchange phosphotungstic acid (AgPW) catalyst for esterification of biologically derived glycerol with acetic acid and showed excellent water resistance and stability [23]. At present, polyoxometalates such as Ag 3 PW 12 O 40 [27], Cs 3 PW 12 O 40 [28], Cs 4 SiW 12 O 40 [29], Ag 4 SiW 12 O 40 [30] are used as catalysts, but there are few studies constructing the heterojunction system with ZnO and using it to degrade organic dyes.
Therefore, to solve the application problems of zinc oxide as a kind of photocatalytic material in the photocatalytic degradation of organic dyes, ZnO/AgSiW and ZnO/CsPW nano heterojunction catalysts were prepared by a simple dissolution and precipitation method. With the introduction of Ag + and Cs + , the problem of the highwater solubility of silicotungstic acid and phosphotungstic acid was greatly improved. In addition, AgSiW and CsPW show an energy band structure matching ZnO, which can increase the separation rate of photogenerated carriers and facilitate the formation of more active species. Our work has a certain reference value for constructing the ZnO/ polyoxometalate heterojunction system and its application in the photocatalytic degradation of dyes.

Chemicals
Silver nitrate, zinc chloride and sodium hydroxide were purchased from Tianjin Kaitong Chemical Reagent Co., Ltd. Caesium chloride were purchased from Shanghai Macklin Blochemical Co., Ltd. Silicotungstic acid and Phosphotungstic acid was products of Tianjin Guangfu Fine Chemical Research Institute. RhB was obtained from Shanghai McLean Biochemical Co., Ltd. Anhydrous ethanol was supplied by Tianjin Fuchen chemical reagent factory. All reagents are analytical grade and can be used without further purification. The secondary deionized water used was prepared in the laboratory.

Preparation of ZnO/Ag 4 SiW 12 O 40 and ZnO/ Cs 3 PW 12 O 40
ZnO nanoparticles were synthesized by hydrothermal method. The specific synthesis process was as follows: 15 mL water solution with 1.36 g ZnCl 2 was added into 20 mL 1 mol/L NaOH solution and stirred evenly. Then, the mixture was transferred into a stainless steel autoclave lined with 50 mL polytetrafluoroethylene, reacted at 120 °C for 12 h, and naturally cooled to room temperature. After the precipitate was washed three times with deionized water and ethanol respectively, and then dried at 80 °C, ZnO nanoparticles were obtained. The synthesis process of nano-heterojunction photocatalyst was as follows: 0.5 g ZnO was evenly dispersed in 50 mL deionized water and the solution was ultrasonicated for 10 min. Add 0.6 mmol H 4 SiW 12 O 40 to the above solution and stir it with magnetic force for 2 h, then 2.4 mmol AgNO 3 in the mixed solution and stir it at room temperature for 24 h. The precipitate was obtained after the former mixed system was centrifuged. Then the fabricated nano-heterojunction ZnO/Ag 4 SiW 12 O 40 (ZnO/AgSiW) photocatalysts were obtained after the precipitate was alternately washed three times with water and ethanol, and was dried overnight at 80 °C. The preparation method of ZnO/Cs 3

Characterization
The morphologies of the samples were characterized using transmission electron microscope (TEM) (Hitachi, H-7650, Japan). X-ray diffraction (XRD) analysis of samples was performed on X-ray diffractometer using Cu Kα radiation as X-ray source, operated at 60 kV and 80 mA (Bruker, AXS(D8), German). X-ray photoelectron spectroscopy (XPS) was performed on electron spectrometer (VG, EscaLab 250Xi, Britain) with Al Kα X-ray radiation at 300 W. All binding energies were calibrated using C 1 s peak at 284.8 eV. The Fourier Transform Infrared Spectrometer (FT-IR) spectrum was recorded employing Fourier transform infrared spectrophotometer (PE, America) at wavenumber range from 400 to 4000 cm −1 with KBr discs. Raman spectra of all samples were analyzed at room temperature by a Thermo Scientific DXR Smart Raman spectrometer (Thermo Electron Corporation, DXR2xi, America). The absorbance results were determined by the UV-Vis dual-beam spectrophotometer (Persee, TU-1900, China) manufactured by Beijing General Electric Co., Ltd. The pore diameter of catalyst was determined by N 2 absorption-desorption isotherms (Quantachrome Nova Win2, American Kang Tai Company).

Photoelectrochemical Measurements
Photocurrent measurements of conventional three electrode structure were carried out on CHI660E Electrochemical Workstation, including counter electrode, reference electrode and working electrode. The effective area of the working electrode prepared with the sample is about 1 cm 2 . Platinum foil and Ag/AgCl electrode are used as counter electrode and reference electrode respectively. A 200 W (PLS-FX300HU, Nanjing chunyinjia Biotechnology Co., Ltd) xenon lamp was used as the light source. The electrolyte was 0.1 M Na 2 SO 4 aqueous solution. In general, the preparation of working electrode is as follows: 40 mg of sample was suspended in 200 μL of prepared solution (1.8 mL of ethanol and 0.2 mL of naphthalene) and the slurry was obtained by ultrasonic treatment for 30 min. Then, 100 μL solution was evenly dropped onto 1 × 2 cm 2 FTO glass substrate. Finally, the working electrode was obtained by drying the prepared electrode at room temperature.

Photocatalytic Activity Evaluation
In this work, RhB was chosen as representative of organic pollutants. The process of photocatalytic experiment was as follows: in 50 mL quartz tube, 15 mg photocatalysts was dispersed in 50 mL RhB solution of 50 mg/L. The quartz tube was placed in a reactor with cooling water jacket. The mixture was stirred in dark for 30 min in to reach the adsorption saturation of RhB on nanophotocatalysts. Next, the reaction system was irradiated under 500 W xenon lamp. At 10 min intervals, 5 mL of the mixture was removed and centrifuged. The absorbance of supernatant was measured by UV spectrophotometer at 554 nm.

Photocatalytic Cycle Performance Test
After the single photocatalytic degradation reaction, the photocatalytic system was centrifuged and the solid catalyst was collected. Then, the catalyst was washed three times under ultrasound and dried at 60 °C for 12 h. The mass of the photocatalyst solid was weighed and supplemented to reach 15 mg with new catalyst that had not been subjected to the photocatalytic reaction. The rest of the experiment operations were same as the single reaction process.

Scavenger Experiment of Photogenerated Carriers
In order to identify the reactive species which play an important role in the photocatalytic degradation process, scavenger experiment of photogenerated carriers was carried out. For this purpose, p-benzoquinone, isopropanol and methanol were used as trapping agents for superoxide radical (·O 2 − ), hydroxyl radical (·OH) and hole (h + ) in the photocatalytic process. The rest of experiment process was same as the single photocatalytic reaction.

Crystal Structure Analysis
The phase structure of ZnO before and after compounding with AgSiW and CsPW was studied by XRD. The results are shown in Fig. 1

Morphology Analysis
The morphology of the prepared nanomaterials was determined by TEM. Figure 2 shows the TEM images of ZnO, AgSiW, CsPW, ZnO/AgSiW and ZnO/CsPW. In Fig. 2a, ZnO is a nano sheet with a size of about 100-400 nm and a thickness of about 30 nm. Figure 2b shows the TEM image of nano AgSiW. It can be seen from the image that AgSiW is massive with a size of about 200 nm, and nano dots with uniform size are distributed on the surface and edge. Figure 2c shows the TEM image of CsPW. It can be seen from the image that CsPW is a nanosphere with a diameter of 100-200 nm and has good dispersibility. As shown from Fig. 2d, when ZnO and AgSiW form a composite, the morphology of AgSiW changes, so that the massive AgSiW becomes dispersed and evenly distributed on the surface and edge of ZnO. Figure 2e shows that ZnO and CsPW are closely combined in the composite system. Figure 2 shows that ZnO/AgSiW and ZnO/CsPW composite catalysts were successfully prepared.

FTIR Analysis
To further determine the structure of the composite catalyst, the infrared spectra of the samples were analyzed.  The peak at 878 cm −1 may be due to the connection between counter ion and bridge oxygen W-O b -W [31].  (Fig. 3).

Raman Spectra Analysis
The  [27,40]. The peaks at 336, 439 and 583 cm −1 belong to the vibration modes of 2E2 (M), E2 (high) and E1 (LO) of ZnO [19]. When ZnO and AgSiW are combined, a significantly enlarged band compared with AgSiW is observed at 897 cm −1 , and there is an obvious displacement, indicating that there is a certain interaction between ZnO and AgSiW [41,42]. However, no obvious characteristic peak of ZnO was observed in the spectrum, which may be due to the weak Raman intensity of ZnO after recombination. After ZnO and CsPW were compounded, the characteristic peaks belonging to ZnO and CsPW can be found in the spectrum, but the characteristic peak of CsPW is weaker than that of pure CsPW, which may be due to the small compounding amount of CsPW or the hydrogen bond interaction between the oxygen atom of Keggin anion and the hydroxyl group on the surface of ZnO [41].

Band Gap Analysis
The optical properties of ZnO/AgSiW, ZnO/CsPW, ZnO, AgSiW and CsPW were studied by UV diffuse reflectance spectra (DRS). The results are shown in Fig. 5. The illustrated part in the figure is calculated by the formula (1). It can be estimated that the band gaps of ZnO/AgSiW, ZnO/CsPW, ZnO, AgSiW and CsPW are 2.53, 3.10, 3.15, 3.06 and 3.21 eV respectively. It can be seen that the heterojunction system constructed with ZnO, AgSiW and CsPW increases the absorption of ZnO of light, and the effect of introducing AgSiW is very obvious, which also promotes ZnO/AgSiW to make full use of sunlight. This result may also be directly related to the stronger photocatalytic degradation ability of ZnO/AgSiW than that of ZnO/CsPW.  Figure 6 shows the X-ray photoelectron spectroscopy (XPS) of the sample to further verify the elemental composition and chemical state of the synthetic samples. As shown in Fig. 6a, Zn 2p, Ag 3p, O 1 s, W 4p, Ag 3d, C 1 s, W 4d and W 4f signals are clearly displayed in the full scan measurement spectrum of ZnO/AgSiW. The low content of Zn 2p may be related to the uniform distribution of AgSiW on the surface of ZnO. The signals of Zn 2p, Cs 3d, O 1 s, W 4p, C 1 s, W 4d, Zn 3 s, Zn 3p, W 4f and Zn 3d were clearly displayed in the full scanning measurement spectrum of ZnO/CsPW. The elemental composition of ZnO/AgSiW and ZnO/CsPW was determined by full spectrum scanning. Figure 6b shows the XPS spectra of Zn 2p in ZnO/AgSiW, ZnO/CsPW and ZnO. The binding energy of Zn in ZnO/AgSiW and ZnO/ CsPW samples is shifted compared with that of Zn in pure ZnO at 1021.63 eV (Zn 2p 3/2 ) and 1044.63 eV (Zn 2p 1/2 ). Among them, the binding energy of Zn in ZnO/AgSiW samples shifted to the right, while that of Zn in ZnO/ CsPW samples shifted to the left, which indicated that there may be a certain interaction force between Zn and AgSiW and CsPW when ZnO formed a composite heterojunction system with AgSiW and CsPW. Figure 6c and d show the XPS spectra of samples Ag 3d and Cs 3d respectively. The peaks at 366.88 and 372.88 eV are attributed to the 3d 3/2 and 3d 5/2 orbits of Ag (I), while the peaks at 724.78 and 738.68 eV are attributed to the 3d 3/2 and 3d 5/2 orbits of Cs (I). Compared with the AgSiW and CsPW of monomers, the binding energy of Ag and Cs has shifted after the formation of the composite [43]. Figure 6e shows the W 4f XPS spectra of ZnO/AgSiW and ZnO/CsPW samples. In the spectrum of AgSiW, the peaks at 34.17 and 36.27 eV can be attributed to W (VI), and the peaks can be observed at 33.56 and 35.66 eV, which indicated the presence of reduced W (V) in the complex. This value may be due to the partial decomposition of H 4 SiW 12 O 40 and the formation of WO x type oxides during the preparation process, or more likely to indicate the existence of a disturbed tungstate environment caused by the interaction between AgSiW and the carrier [44]. Compared with pure AgSiW, the binding energy of W shifted significantly to the right. In the spectrum of CsPW, the peaks at 36.  [45]. The above results show that the binding energy of Zn ions and W ions changes when ZnO constructs a heterojunction system with AgSiW and CsPW. In addition, the formation of Zn-O-W chemical bond is confirmed [46,47], and the composite catalyst is successfully prepared. These results are also consistent with those of infrared and Raman spectra.

BET Analysis
The surface area and pore structure of the samples were further analyzed by the nitrogen adsorption-desorption technique. As can be seen from Fig. 7a, the isotherms of ZnO and ZnO/AgSiW samples is type III with H3 hysteresis loops [48,49], and those of ZnO/AgSiW samples is type IV isotherms with H3 hysteresis loops [50,51], which indicated that the mesoporous structures are cracked in the samples. The specific surface area of pure ZnO is 13.043 m 2 /g. When the heterojunction system was constructed with ZnO/CsPW and ZnO/AgSiW, the specific surface areas of the heterojunction catalysts were 13.032 and 8.191 m 2 /g, respectively. Figure 7b shows that the introduction of CsPW and AgSiW has a certain impact on the pore size of ZnO, and AgSiW has a significant impact on the specific surface area of the heterojunction system. The possible reason is that AgSiW is evenly dispersed on the surface of ZnO and strongly coupled with ZnO [52], which is consistent with the TEM results.

Photocatalytic Activity
The photocatalytic activities of ZnO/AgSiW and ZnO/ CsPW samples were tested with RhB as the target pollutant. Under the irradiation of simulated sunlight, the removal effect of the catalyst on dyes was evaluated under the conditions of catalyst dosage of 15 mg and dye concentration of 50 mL. According to Fig. 8a  where k app is the apparent rate constant (min −1 ), t is the irradiation time (min), C o and C t are the absorbance at t = 0 min and t = ∞ [7].
The calculated k app values of ZnO/AgSiW, ZnO/CsPW and ZnO are 0.0403, 0.0203 and 0.0028 min −1 respectively, and the reaction rate constants of ZnO/AgSiW and ZnO/ CsPW are 14.4 and 7.3 times bigger than those of ZnO, respectively. The results show that the two composite catalysts have stronger catalytic activities than that of pure ZnO.

Recyclability of Photocatalysts
In addition to the efficiency of photocatalysts, their stability and recoverability are also important indicators to evaluate whether they can be applied to practices. In order to explore the stability and recoverability of photocatalyst, the cyclic experiment is carried out. As shown in Fig. 9, under simulated sunlight, the removal rates of RhB by ZnO/AgSiW and ZnO/CsPW for the first time are 91.3% and 73.7%, respectively. After the two catalysts are recycled for three times, the removal rates of RhB are 84.8% and 65.3% respectively. They still have high photocatalytic activities, which proves the stability and recoverability of photocatalysts.

Reaction Mechanism
To confirm the effective separation of photogenerated carriers, the transient photocurrent response and  Figure 10a shows the transient photocurrent response of ZnO, ZnO/ AgSiW and ZnO/CsPW under visible light irradiation. As can be seen from the figure, when the light is switched three cycles, the generated photocurrent is stable and repeatable. Under visible light irradiation, the photocurrent density of ZnO/AgSiW and ZnO/CsPW is higher than that of ZnO. The higher photocurrent density means the lower efficiency of photoinduced charge recombination [53,54], which means the increase of charge separation rates of the two heterojunction catalysts. The Nyquist curve is shown in Fig. 10b. Generally, the arc radius on the EIS Nyquist diagram corresponds to the charge transfer resistance (Rct), while the small arc radius represents the low charge transfer resistance [55,56]. Obviously, ZnO/ AgSiW and ZnO/CsPW composite catalysts have lower charge transfer resistance than ZnO, which means that the formed composite system can effectively improve the charge transfer between interfaces. This result is consistent with that of the photocurrent response, which shows that the composite system formed by ZnO, AgSiW and CsPW can significantly improve the separation efficiency of photogenerated carriers, the charge transfer between interfaces, and the photocatalytic degradation performance.  Figure 11 shows the free radical capture experiment that determines the active substances that play a major role in the photocatalytic process. P-benzoquinone, isopropanol and methanol were used as capture agents for superoxide anion radicals, hydroxyl radicals and holes. It can be seen from the left half of Fig. 11 that the addition of p-benzoquinone has the greatest impact on the photocatalytic activity of ZnO/ AgSiW composite catalyst, which is followed by methanol. That indicates that superoxide anion free radicals and holes play the main catalytic role in the photocatalytic process.

Free Radical Capture Experiment
In addition, the addition of isopropanol has relatively little impact on the photocatalytic activity, which indicates that holes also hinder the degradation reaction, but they are not the main active substance. It can be seen from the right half of Fig. 11 that the addition of isopropanol and methanol has the greatest impact on the photocatalytic activity of ZnO/ CsPW composite catalyst. That indicates that hydroxyl radicals and holes are the main active substances in the photocatalytic reaction process and also play a certain inhibitory role after the addition of p-benzoquinone, indicating that superoxide anion radical is also produced in the photocatalytic process. On this basis, the possible mechanisms of photocatalytic degradation of RhB by ZnO/AgSiW and ZnO/CsPW were proposed. The band gaps of ZnO, AgSiW and CsPW are estimated to be 3.15, 3.06 and 3.21 eV, respectively, according to the UV Vis diffuse reflectance spectra. The measured potentials and Ag/AgCl were converted to the normal hydrogen electrode (NHE) scale by Eq. (3) [57]: Therefore, the conduction band (CB) of ZnO, AgSiW and CsPW can be estimated by Mott-Schottky spectrum shown in Fig. 12, which are − 0.34, 0.76 and − 0.01 eV, respectively. The valence band (VB) of ZnO, AgSiW and C S PW can be calculated according to formula (4) [58], which are 2.81, 3.29 and 3.09 eV, respectively.
According to the estimated CB and VB positions, the possible mechanism diagrams of photocatalytic degradation of RhB by ZnO/AgSiW and ZnO/CsPW were constructed. As shown in Fig. 13, it is assumed that ZnO and AgSiW form a Z-scheme heterojunction. Under the irradiation of simulated sunlight, electrons are excited from VB of ZnO to CB, and photogenerated holes are left in VB. Because the CB of ZnO is more negative than that of O 2 /·O 2 − potential (− 0.046 V vs. NHE) [57,59], electrons on CB of ZnO can form ·O 2 − with oxygen, and the formed ·O 2 − can directly react with RhB molecule in water.
In AgSiW, the electrons are transferred from VB to CB and combined with the holes on VB of ZnO. This electron transfer can reduce electron-hole recombination in ZnO and improve photocatalytic activity. Because the VB (+ 2.81 V vs. NHE) of ZnO is more positive than that of OH − /·OH  According to the capture experiment, only a small part of the holes will react with H 2 O and convert into ·OH, and the remaining holes will directly participate in the degradation of RhB. The other inference is inconsistent with the free radical capture experiment, because when the electrons on CB of ZnO are transferred to the CB of AgSiW and accumulated, the CB (+ 0.76 V vs. NHE) of AgSiW is positive than that of the O 2 /·O 2 − potential (-0.046 V vs. NHE). AgSiW is a deep electron trap, which makes it difficult for electrons to jump out of AgSiW, so that the conversion from O 2 to ·O 2 − is blocked [60].
According to the free radical capture experiment, holes and hydroxyl radicals play a major role in the photocatalytic process, and superoxide anion radicals also play a certain inhibitory role. As shown in Fig. 14 . Therefore, some holes on VB of CsPW can react with water to form hydroxyl radicals, and the remaining holes directly play the role of oxidative degradation of RhB. If the type II heterojunction is formed and the electrons on the CB of ZnO are transferred to the CB of CsPW whose potential is positive than the potential of O 2 /·O 2 − (− 0.046 V vs. NHE), a shallow electron trap, the electrons are not easy to react with O 2 to generate superoxide anion radicals. That is inconsistent with the experimental results of radical capture. Therefore, it is inferred that a Z-scheme heterojunction is formed between ZnO and CsPW.

Degradation Pathways and Intermediates of RhB
To better understand the intermediate products generated in the degradation process and the degradation mechanism, LC-MS was used to detect the intermediates. Then the degradation mechanism was speculated through the intermediate products. Typically, the RhB molecules in the heterojunction catalyst would adsorb on the catalyst particle surface through the positively charged diethylamine function. Under attack of reactive free radicals, the RhB structure was destroyed. Then, some unstable colorless organic intermediates or their isomers were produced, resulting in the color fading of the solution. According to previous reports, the degradation of RhB can be divided into the following processes: N-deethylation, deamination, dealkylation, decarboxylation, chromophore cracking, ring-opening and mineralization. We speculate the degradation pathway of RhB according to the test results in Fig. 15a. The degradation mechanism of ZnO/AgSiW is shown in Fig. 16. The degradation of RhB (m/z = 443) began with N-deethylation by the cleavage of C = N in a stepwise manner. At first, the removal of one ethyl group leads to two isomer intermediate products, P1 and P1-1 (m/z = 415). Subsequently, another two isomers, P2 and P2-1 (m/z = 359), are obtained via further N-deethylation of P1 and P1-1. At last, P3 (m/z = 331) is obtained with complete elimination of ethyl [61]. De-amination of the intermediate P3 could generate the intermediate P4 with m/z values of 301 [62]. At the same time, the cleavage of the chromophore structure might occur, leading to the formation of some monocyclic aromatic compounds. Afterward, the ring-opening reactions took place, causing the formation of some small molecular compounds with charge mass ratios of 158, 149, 145, 118, 116, 91 and 85 [62][63][64]. Finally, these low molecular weight organic compounds are further mineralized into inorganic products. According to the LC-MS test results, the main intermediates obtained by the two catalysts for the degradation of RHB are similar. Therefore, we speculate that the degradation mechanism of ZnO/CsPW is similar to that of ZnO/AgSiW. The degradation mechanism of Fig. 16 is also applicable to ZnO/CsPW.

Conclusions
In conclusion, flake nano-ZnO was prepared by the hydrothermal method, and ZnO/AgSiW and ZnO/CsPW nanoheterojunction composite catalysts were successfully prepared by the simple dissolution and precipitation method. Compared with pure ZnO, the prepared catalyst has a higher absorption threshold of light. Under the irradiation of simulated sunlight, the two catalysts showed excellent photocatalytic degradation performance on RhB, and the degradation process accorded with the first-order kinetic equation. The reaction rate constants of ZnO/AgSiW and ZnO/CsPW were 14.4 and 7.3 times higher than that of ZnO, respectively. In addition, the active substances that play a major role in the degradation of ZnO/AgSiW and ZnO/CsPW are ·O 2 − and h + and ·OH and h + , respectively. Transient photocurrent response and electrochemical impedance show that the heterojunction system can effectively inhibit carrier recombination. It can be seen from the above experiments that the two catalysts have high catalytic activities, good stability and recoverability. Our study shows that the construction of ZnO/Keggin polyoxometalate nano-heterojunction catalyst is feasible and has a good application prospect in the field of degradation of organic dyes.