Synthesis of Dual Z-Scheme CuBi 2 O 4 /Bi 2 Sn 2 O 7 /Sn 3 O 4 Photocatalysts with Enhanced Photocatalytic Performance for the Degradation of Tetracycline under Visible Light Irradiation

: The dual Z-scheme heterojunction CuBi 2 O 4 /Bi 2 Sn 2 O 7 /Sn 3 O 4 (CBS) was successfully constructed through in situ growth methods, and its photocatalytic performance was evaluated via degradation of tetracycline hydrochloride under visible light. Out of all samples, CBS-2 exhibited the highest photocatalytic activity, with an apparent rate constant of 2.34, 20.16, and 44.17 times that of Bi 2 Sn 2 O 7 , CuBi 2 O 4 , and Sn 3 O 4 , respectively. Even after four cycles, the photocatalytic efﬁciency remained above 85%. The improvement can be attributed to the construction of the Z-scheme hetero-junction, which effectively promotes the separation and migration of photogenerated carriers. The possible photocatalytic degradation mechanism of dual Z-scheme heterojunction CBS was deduced based on the theory of free radical capture and energy band.


Introduction
In recent years, antibiotics have been widely used in many fields, but at the same time, they are difficult for the human body to completely metabolize, and a large quantity of them enter the water environment through the digestive system, causing increasing water pollution. Furthermore, they are reabsorbed by the human body through the water environment, which is harmful to human health [1][2][3][4]. Therefore, antibiotic pollution has become a major problem in water pollution control. Among them, tetracycline, as a class of common broad-spectrum antibiotics, is widely used in animal husbandry and animal medicine, but due to years of overuse, its residues exist in the water environment for a long time and are difficult to degrade [5]. Therefore, safe and effective removal methods have received extensive attention.
Commonly used methods for the removal of antibiotics are adsorption [6,7] and biodegradation [8,9], but their shortcomings are also obvious: the enrichment transfer process producing secondary pollution, the low removal efficiency, the high price, the more toxic by-products, and the large energy consumption. Photocatalysis can be used as a safe, effective, and environmentally friendly method to degrade tetracycline [10][11][12][13]. Mukhtar et al. prepared dual Z-scheme core-shell PANI-CeO 2 -Fe 2 O 3 -NiO heterojunctions to remove MB and MO dyes, with removal rates of 99% and 98%. Increased photocatalytic activity can be attributed to dual Z-scheme formation and the core-shell morphology of PANl, which effectively promote the separation of photogenerated carriers [14]. Munawar et al. prepared a dual Z-scheme Sm 2 O 3 -WO 3 -La 2 O 3 nanocomposite to remove MB and MO dyes, with removal rates of 99% and 96% [15]. However, its photocatalytic stability, absorption spectrum range, and degradation capability limit its further application.

Characterization
The phase of the BSO, CBO, SNO, and CBS-2 composites was characterized via XRD.  . There are no extra peaks from the SnO 2 or SnO phase appearing in the XRD pattern of SNO, implying that the as-synthesized sample is pure SNO. The sharp feature peaks of the pure sample represent high crystallinity, and no extra impurity peaks indicate its high purity. Because of the low content and overlapping characteristic peaks, the main diffraction peaks of the composite samples are similar to BSO; however, they are consistent with the (111) crystal plane of SNO at 27 • and the (202) crystal plane of CBO at 37.5 • , which proves the coexistence of BSO, CBO, and SNO phases in the CBS composites. In summary, the double Z-scheme CBS-2 was successfully obtained. (JCPDS NO. . There are no extra peaks from the SnO2 or SnO phase appearing in the XRD pattern of SNO, implying that the as-synthesized sample is pure SNO. The sharp feature peaks of the pure sample represent high crystallinity, and no extra impurity peaks indicate its high purity. Because of the low content and overlapping characteristic peaks, the main diffraction peaks of the composite samples are similar to BSO; however, they are consistent with the (111) crystal plane of SNO at 27° and the (202) crystal plane of CBO at 37.5°, which proves the coexistence of BSO, CBO, and SNO phases in the CBS composites. In summary, the double Z-scheme CBS-2 was successfully obtained. TEM and HRTEM indicate the morphology and crystalline structure of the as-prepared photocatalyst. As shown in Figure 2a-c, pure BSO is composed of many microspheres, CBO presented rod-like morphology, and SNO exhibited the flower structure. Figure S1 shows the TEM image of composite CBS-2, in which CBO and SNO can be identified; parts of BSO microspheres were successfully distributed on the surface of SNO and CBO due to the in situ generation method. In Figure 2d, we can display the 0.31 nm crystal lattice corresponding to the (222) crystal plane of BSO, the 0.24 nm crystal lattice corresponding to the (211) crystal plane of CBO, and the 0.32 nm crystal lattice corresponding to the (111) crystal plane of SNO, respectively. BSO, CBO, and SNO are TEM and HRTEM indicate the morphology and crystalline structure of the as-prepared photocatalyst. As shown in Figure 2a-c, pure BSO is composed of many microspheres, CBO presented rod-like morphology, and SNO exhibited the flower structure. Figure S1 shows the TEM image of composite CBS-2, in which CBO and SNO can be identified; parts of BSO microspheres were successfully distributed on the surface of SNO and CBO due to the in situ generation method. In Figure 2d, we can display the 0.31 nm crystal lattice corresponding to the (222) crystal plane of BSO, the 0.24 nm crystal lattice corresponding to the (211) crystal plane of CBO, and the 0.32 nm crystal lattice corresponding to the (111) crystal plane of SNO, respectively. BSO, CBO, and SNO are observed to be closely combined. Therefore, the TEM and HRTEM images prove the existence of CBO and SNO in the composite material and they are in close connection with the basal BSO material and jointly construct the heterojunction structure. observed to be closely combined. Therefore, the TEM and HRTEM images prove the existence of CBO and SNO in the composite material and they are in close connection with the basal BSO material and jointly construct the heterojunction structure. The micro-strain variation is linked with the crystallite size changes, and δ (δ = 1/D 2 ) has a direct impact on lattice imperfection. We can calculate the average crystalline size from the XRD results using the Debye-Scherrer formula [34]: The average crystalline size of BSO, CBO, SNO, and CBS-2 is about 17.5 nm, 11.6 nm, 10.3 nm, and 17.2 nm, respectively, which also supported TEM results. XPS testing was used to further confirm the elemental composition and valence state of the prepared materials. The XPS survey spectra of BSO and CBS-2 are reproduced in Figure S2. The BSO contains Bi, Sn, and O elements, while the composite samples contain Bi, Sn, O, and Cu elements, which suggest the coexistence of CBO in CBS-2. In Figure 3a, the Bi 4f high-resolution spectrum, the peaks near 164.7 eV and 159.4 eV correspond to  The micro-strain variation is linked with the crystallite size changes, and δ (δ = 1/D 2 ) has a direct impact on lattice imperfection. We can calculate the average crystalline size from the XRD results using the Debye-Scherrer formula [34]: The average crystalline size of BSO, CBO, SNO, and CBS-2 is about 17.5 nm, 11.6 nm, 10.3 nm, and 17.2 nm, respectively, which also supported TEM results.
XPS testing was used to further confirm the elemental composition and valence state of the prepared materials. The XPS survey spectra of BSO and CBS-2 are reproduced in Figure S2. The BSO contains Bi, Sn, and O elements, while the composite samples contain Bi, Sn, O, and Cu elements, which suggest the coexistence of CBO in CBS-2. In Figure 3a, the Bi 4f high-resolution spectrum, the peaks near 164.7 eV and 159.4 eV correspond to the orbital binding energies of Bi 4f 5/2 and Bi 4f 7/2 , respectively, which proves that in BSO and composites, Bi exists in the form of Bi 3+ [35,36]. In Figure 3b, the Sn 3d high-resolution spectrum of BSO, the orbital binding energies of 3d 3/2 and 3d 5/2 correspond to the peaks near 486.9 eV and 495.6 eV, respectively, which proves that it exists in the form of Sn 4+ [37]. In the Sn 3d high-resolution spectrum of the complex, 3d 3/2 and 3d 5/2 can be deconvoluted into four peaks, indicating that there is more than one valence state of Sn in the sample. The peaks near 486.8 eV and 495.5 eV correspond to Sn 4+ , and the peak at 486.3 eV and 494.8 eV corresponds to Sn 2+ [38]. In Figure 3c, the Cu 2p high-resolution spectrum, the binding energies of Cu 2p 3/2 and Cu 2p 1/2 peaks are 934.7 eV and 954.8 eV, representing Cu 2+ [39]. In Figure 3d, the O 1s high-resolution spectrum, the binding energy of 529.8 eV corresponds to the Bi-O bond, 531.7 eV corresponds to the Sn-O bond, and 532.2 eV corresponds to the surface adsorbed oxygen [40]. similar to that of BSO, its average particle size is smaller, and the particles are more dispersed, which means that it has a larger specific surface area and more active sites. The light absorption range of a photocatalyst is closely related to its photocatalytic capacity. The as-prepared samples were characterized via UV-Vis diffuse reflectance spectroscopy to investigate their optical absorption characteristics. According to Figure  4a, the absorption edge range of BSO is about 450 nm-500 nm, CBO is about 750 nm-800 nm, and that of SNO is about 500-550 nm. The composite photocatalyst CBS-2 is slightly red-shifted on the base of pure BSO, which confirms the visible-light response ability of CBS-2. Their band gaps can be obtained from the following formula [41]: We measured the average particle size using dynamic light scattering. The average particle size of BSO, CBO, and SNO is 1142 nm, 463.1 nm, and 287.6 nm, CBO/BSO is 522.1 nm, and CBS-1, CBS-2, and CBS-3 is 207.3 nm, 208.6 nm, and 211.5 nm, respectively, which is similar to the TEM images. Although the average crystalline size of CBS-2 is similar to that of BSO, its average particle size is smaller, and the particles are more dispersed, which means that it has a larger specific surface area and more active sites.
The light absorption range of a photocatalyst is closely related to its photocatalytic capacity. The as-prepared samples were characterized via UV-Vis diffuse reflectance spectroscopy to investigate their optical absorption characteristics. According to Figure 4a, the absorption edge range of BSO is about 450 nm-500 nm, CBO is about 750 nm-800 nm, and that of SNO is about 500-550 nm. The composite photocatalyst CBS-2 is slightly red-shifted on the base of pure BSO, which confirms the visible-light response ability of CBS-2. Their band gaps can be obtained from the following formula [41]: where α, h, ν, and Eg represent proportionality constant, Planck constant, and optical frequency, and n is related to semiconductor type. According to the literature, BSO and SNO are indirect semiconductors, n = 4, and CBO is a direct semiconductor, n = 1 [35,39,42]. Therefore, it can be roughly deduced that the Eg of BSO is 2.76 eV, the Eg of CBO is 1.55 eV, and the Eg of SNO is 2.39 eV. The calculation results are shown in Figure 4b,c, which is consistent with the previous literature. The minimum value of the n-type conduction band (CB) semiconductor is about 0.1 eV higher than the flat-band potential, and the maximum value of the p-type semiconductor valence band (VB) is about 0.1 eV lower than the flat-band potential. Therefore, the ECB values of BSO and SNO are calculated as 0.40 eV and −0.75 eV, and the EVB values of CBO are calculated as 0.93 eV. Then, we calculate their corresponding potential according to the following formula:

EVB = ECB + Eg
Therefore, it can be concluded that the EVB values of BSO and SNO are calculated as 3.16 eV and 1.64 eV, and the EVB values are calculated as −0.92 eV [30,40,44].

Photocatalytic Activity
We tested the photocatalytic performance of the photocatalyst via the degradation efficiency of the model pollutant, tetracycline solution, under visible light. Results are as shown in Figure 5a,b: pure samples reveal poor photocatalytic activities, the photocata- In addition, the band structure of photocatalysts is also one of the important factors affecting the photocatalytic ability of photocatalysts. We can implement the Mott-Schottky (M-S) curves to measure its flat-band potential, then estimate the position of the energy band of the photocatalytic materials. As illustrated in Figure 4d-f, the slopes of BSO and SNO tangents are positive, indicating that both BSO and SNO are n-type semiconductors. The slope of the CBO tangent is negative, indicating that CBO is a p-type semiconductor. The tangents of BSO, CBO, and SNO are extended, and the flat band potential (Efb) of BSO, CBO, and SNO is about 0.3 eV, 0.63 eV, and −0.85 eV, respectively (Efb| Ag/AgCl ). We can obtain standard hydrogen electrode values by using the following equation [43]: The minimum value of the n-type conduction band (CB) semiconductor is about 0.1 eV higher than the flat-band potential, and the maximum value of the p-type semiconductor valence band (VB) is about 0.1 eV lower than the flat-band potential. Therefore, the E CB values of BSO and SNO are calculated as 0.40 eV and −0.75 eV, and the E VB values of CBO are calculated as 0.93 eV. Then, we calculate their corresponding potential according to the following formula:

Photocatalytic Activity
We tested the photocatalytic performance of the photocatalyst via the degradation efficiency of the model pollutant, tetracycline solution, under visible light. Results are as shown in Figure 5a,b: pure samples reveal poor photocatalytic activities, the photocatalytic activities of BSO, CBO, and SNO are only 56.2%, 10.3%, and 5%, and the CWO/BSO heterojunction material showed higher photocatalytic activity than pure samples, but lower than the CBS Z-scheme heterojunction under the same conditions. The best photocatalytic activity of CBS-2 is 89.7% under the same conditions. This may be because, due to the increase in CBO and SNO content, more heterojunction structures were generated, thus improving the photocatalytic activity. However, with the further increase in CBO and SNO content, the photocatalytic activity of the composite decreases, which may be because the decrease in active reaction sites reduces the optical adsorption capacity of the composite catalyst [45,46]. four photocatalytic cycling experiments. The decrease in activity may be due to the adsorption of TC by the active site of the catalyst. In short, the composite material can be reused.

Photocatalytic Mechanism
To investigate the transfer and separation efficiency of photogenerated carriers, photocurrent cycle experiments were carried out on samples. A high photocurrent intensity reflects high separation efficiency [54]. As shown in Figure 6a, the photocurrent intensity of CBS-2 is higher than that of the pure samples, which indicates that the CBS-2 composite material has the strongest carrier transfer and separation efficiency.
The relatively weak PL intensity represents a high separation efficiency [55]. In Figure 6b, the high PL emission peaks of BSO, CBO, and SNO imply a high photogenerated carrier recombination rate, while CBS-2 shows a lower emission peak. Constructing a double Z-scheme heterostructure between BSO, CBO, and SNO, the carrier separation efficiency is higher than that of the pure sample, and the recombination of electron-hole pairs can be suppressed effectively so that the composite material exhibits enhanced photocatalytic activity.
To confirm the difference in the charge transfer separation efficiency and impedance between different samples, electrochemical impedance spectroscopy (EIS) tests were The comparison of the tetracycline degradation activity of CBS-2 and other photocatalytic materials is shown in Table 1. −ln(C t /C 0 ) = kt C 0 represents the concentration of the tetracycline solution before illumination starts after dark treatment, and C t represents a concentration of the tetracycline solution at t min. As illustrated in Figure 5c  In addition to photocatalytic activity, the stability and reusability of photocatalysts are also key considerations restricting their applications [53]. XRD tests were carried out before and after the photocatalytic cycling experiment, respectively. Figure 5e shows that the crystal structure of the photocatalyst has no obvious change after four cycles of photocatalysis. In Figure 5f, the degradation efficiency remained stable after four cycles, the degradation rates of the four repeated experiments were 88.9%, 87.2%, 86.1%, and 85.6%, and the photocatalytic activity of CBS-2 was only slightly decreased by about 4.3% after four photocatalytic cycling experiments. The decrease in activity may be due to the adsorption of TC by the active site of the catalyst. In short, the composite material can be reused.

Photocatalytic Mechanism
To investigate the transfer and separation efficiency of photogenerated carriers, photocurrent cycle experiments were carried out on samples. A high photocurrent intensity reflects high separation efficiency [54]. As shown in Figure 6a, the photocurrent intensity of CBS-2 is higher than that of the pure samples, which indicates that the CBS-2 composite material has the strongest carrier transfer and separation efficiency. Figure 6c shows the relative radius of the pure and composite photocatalyst, sorted by size as CBO > SNO > BSO > CBS-2.
According to characterization results, due to the successful construction of the heterojunction, the average particle size was reduced, the optical absorption range of CBS was widened, the carrier separation efficiency was improved, and the photogenerated carrier recombination rate was reduced; therefore, the photocatalytic activity was significantly improved.
To inspect the main active radicals in the photocatalytic process of the CBS-2 composite, we performed a radical trapping experiment. The experimental results are shown in Figure 6d: the addition of IPA has almost no effect on the degradation efficiency, but the degradation efficiency is significantly inhibited by the addition of BQ, and the addition of AO also has a certain inhibitory effect on the degradation efficiency. The degradation efficiency dropped slightly from 88.9% to 32.8% and 68.8% when BQ and AO were added. These results indicate that •O2 − and h + act synergistically in the photocatalytic process of CBS-2, of which •O2 − is more important.
To elucidate the possible mechanism, ESR characterization and trapping experiments were used to detect the main active radicals. From Figure 6e, •O2 − characteristic signals of pure BSO cannot be obtained under visible light or in the dark, while that of CBS-2 can, which illustrates the presence of •O2 − and its participation in the photocatalytic process. As shown in Figure 6f, both BSO and CBS-2 can receive •OH characteristic signals after irradiation. Trapping experiments also illustrate that h + and •O2 − play critical roles in the degradation of TC. The relatively weak PL intensity represents a high separation efficiency [55]. In Figure 6b, the high PL emission peaks of BSO, CBO, and SNO imply a high photogenerated carrier recombination rate, while CBS-2 shows a lower emission peak. Constructing a double Z-scheme heterostructure between BSO, CBO, and SNO, the carrier separation efficiency is higher than that of the pure sample, and the recombination of electron-hole pairs can be suppressed effectively so that the composite material exhibits enhanced photocatalytic activity.
To confirm the difference in the charge transfer separation efficiency and impedance between different samples, electrochemical impedance spectroscopy (EIS) tests were performed on BSO, CBO, SNO, and the composite material. The smaller radius of the circle in the diagram means lower resistance, and faster electron migration rate [56]. Figure 6c shows the relative radius of the pure and composite photocatalyst, sorted by size as CBO > SNO > BSO > CBS-2.
According to characterization results, due to the successful construction of the heterojunction, the average particle size was reduced, the optical absorption range of CBS was widened, the carrier separation efficiency was improved, and the photogenerated carrier recombination rate was reduced; therefore, the photocatalytic activity was significantly improved.
To inspect the main active radicals in the photocatalytic process of the CBS-2 composite, we performed a radical trapping experiment. The experimental results are shown in Figure 6d: the addition of IPA has almost no effect on the degradation efficiency, but the degradation efficiency is significantly inhibited by the addition of BQ, and the addition of AO also has a certain inhibitory effect on the degradation efficiency. The degradation efficiency dropped slightly from 88.9% to 32.8% and 68.8% when BQ and AO were added.  Figure 7 illustrates the possible mechanism of the photodegradation of TC. The double Z-scheme photocatalytic is constructed with BSO, CBO, and SNO. Under visible light, they are all motivated to generate photoproduced electron-hole pairs. According to the classic type II heterojunction transfer mechanism, the CB of BSO (0.4 eV) is more positive, and the electrons on the CB of CBO and SNO will transfer to the CB of BSO. Meanwhile, the h + shift from BSO to CBO and SNO occurs because of the more positive VB of BSO. The CB of BSO is more positive than O 2 /•O 2 − (−0.33 eV), and the VB of CBO and SNO is more negative than H 2 O/•OH (2.27 eV), so it is impossible to generate •O 2 − thermodynamically. However, that inference contrasts with the results of trapping experiments and ESR analysis. It can be inferred that the photogenerated carrier migration follows a Z-scheme mechanism. When BSO, CBO, and SNO are motivated to generate holes and electrons, the electrons on the CB of BSO recombine with the holes on the VB of CBO and SNO, which not only retains the e − of CBO and SNO with stronger reduction and the h + of BSO with oxidation ability in the CBS-2 system but also maintains the separation of charge carriers. Some holes on the VB of BSO can directly degrade the TC solution, and others can react with H 2 O to generate •OH to further degrade the TC solution. In addition, the electrons on the CB of CBO and SBO can convert the adsorbed oxygen into •O 2 − , thereby further oxidatively degrading the TC solution. The construction of the double Z-scheme mechanism improves the efficiency of carrier separation and migration while retaining the strong redox ability of the catalyst, thereby enhancing the photocatalytic performance of the composite. lution, and others can react with H2O to generate •OH to further degrade the TC solu tion. In addition, the electrons on the CB of CBO and SBO can convert the adsorbed ox ygen into •O2 − , thereby further oxidatively degrading the TC solution. The constructio of the double Z-scheme mechanism improves the efficiency of carrier separation an migration while retaining the strong redox ability of the catalyst, thereby enhancing th photocatalytic performance of the composite.

Synthesis of BSO
Based on previous reports, Bi2Sn2O7 was synthesized using a simple hydrotherma method [40]. Generally, SnCl4·5H2O (2.80 g) was dissolved in 80 mL H2O, Bi(NO3)3·5H2O (3.88 g) was added slowly to it under rapid stirring, and 80 mL 2 M NaOH solution wa gradually dropped into it afterward. Then, the suspension was transferred to a 200 m stainless steel hydrothermal kettle and kept at 180 °C for 24 h, filtered and cleaned with large amount of water and absolute ethanol after the reaction to separate the powde

Synthesis of BSO
Based on previous reports, Bi 2 Sn 2 O 7 was synthesized using a simple hydrothermal method [40]. Generally, SnCl 4 ·5H 2 O (2.80 g) was dissolved in 80 mL H 2 O, Bi(NO 3 ) 3 ·5H 2 O (3.88 g) was added slowly to it under rapid stirring, and 80 mL 2 M NaOH solution was gradually dropped into it afterward. Then, the suspension was transferred to a 200 mL stainless steel hydrothermal kettle and kept at 180 • C for 24 h, filtered and cleaned with a large amount of water and absolute ethanol after the reaction to separate the powder from the liquid, and finally dried at 60 • C to collect the yellow powder. The material is recorded as BSO.

Synthesis of CBO/BSO
Firstly, Bi(NO 3 ) 3 ·5H 2 O (2.42 g), Cu(NO 3 ) 2 ·3H 2 O (0.60 g) and NaOH (0.87 g) were distracted into 80 mL water and stirred for 3 h. Then, a certain amount of BSO was added and stirred for 1 h. Finally, the mixed solution was transferred to a 100 mL stainless steel hydrothermal kettle and kept at 180 • C for 24 h, filtered and cleaned with a large amount of water and absolute ethanol after the reaction to separate the powder from the liquid, and finally dried at 60 • C for use. The mass ratio between CBO and BSO in the prepared samples was 1%, which was denoted as CBO/BSO. Pure CBO samples were prepared through the same protocol without the addition of BSO [30].

Synthesis of CBS
We fully dissolved Na 3 C 6 H 5 O 7 ·2H 2 O (2.94 g) and SnCl 2 ·2H 2 O (0.90 g) in 10 mL of water, gradually dropped 10 mL of 0.2 M NaOH into it under continuous stirring, added a certain amount of CBO/BSO after stirring for half an hour, and stirred for another hour. The suspension was put into a 50 mL hydrothermal reactor, and reacted at 180 • C for 12 h; after cooling, the product was filtered and washed with a large amount of water and absolute ethanol, and finally dried at 60 • C for later use. The sample with a mass of CBO accounting for 1% of the total mass and a mass of SNO accounting for 1% of the total mass was labeled as CBS-1, the sample with a the mass of CBO accounting for 1% of the total mass and a mass of SNO accounting for 2% of the total mass was labeled as CBS-2, and the sample with a mass of CBO accounting for 2% of the total mass and a mass of SNO accounting for 5% of the total mass was labeled as CBS-3. Pure SNO samples were prepared through the same protocol without the addition of CBO/BSO [44].

Characterization Methods
The composition and crystallinity characteristics of the as-prepared photocatalysts were measured with an X-ray powder diffractometer (Shimaduz-6100, Kyoto, Japan) at a scan rate of 10 • /min from 10 • to 80 • . Transmission electron microscope (TEM) images and high-resolution transmission electron microscope (HRTEM) images were used to observe the size and microstructure of the photocatalytic materials. X-ray photoelectron spectroscopy (XPS) was carried out on a Thermo (Waltham, USA) ESCALAB 250Xi spectrometer, which was devoted to analyzing the elemental composition and its valence state in the photocatalysts. The particle size of the characterized photocatalysts was analyzed with a Malvern Zetasizer Nano ZS90. Taking BaSO 4 as the reflection standard to test the absorption edge of the sample, the UV-Vis diffuse reflectance spectra (UV-DRS) were performed on a Shimadzu UV-3600 spectrophotometer in the range 200 nm to 800 nm. Electrochemical tests were obtained with a CHI-660D electrochemical workstation. We used 0.1 M sodium sulfate solution as an electrolyte solution in photoelectric tests. The sample electrode played the role of the working electrode. In the beginning, 10 mg of photocatalysts and microscale Nafion were uniformly dispersed in a minute quantity of ethanol via ultrasonic treatment. Then, the mixture was spread evenly on one square centimeter area of FTO glass. Finally, it was dried at 120 • C for 2 h. Photoluminescence (PL) spectra were characterized using a fluorescence spectrometer (Cary Eclipse, Agilent, Santa Clara, CA, USA) at an excitation wavelength of 520 nm. The electron paramagnetic resonance spectra of the as-prepared photocatalysts were characterized with the JES FA200 EPR Spectrometer. Radicals were measured via electron spin resonance (ESR, EOL JES-FA200 EER) using 5,5-dimethyl-1-pyrroline oxynitride (DMPO) as a spin trap.

Photocatalytic Evaluation
The photocatalytic performance of the samples was evaluated by calculating the degradation rate of tetracycline under 300 W xenon lamp irradiation with a filter (λ > 400 nm). Firstly, an aqueous solution containing tetracycline (100 mL, 5 ppm) and 50 mg photocatalyst was stirred uniformly in the dark for half an hour to reach adsorption-desorption equilibrium. Then, it was placed in a 250 mL cylindrical quartz reactor cooled with cold water circulating outside the reactor. The stirrer was used to ensure that the suspension stayed homogeneous during the photocatalytic reaction, removing the supernatant from the reactor at intervals. Finally, after sampling and centrifugation, the supernatant was filtered through a 0.22 m filter membrane and then quantitatively analyzed via high-performance liquid chromatography. The mobile phase was methanol, acetonitrile, and 0.01 M oxalic acid in a 1:2:7 volume ratio.

Conclusions
In summary, we synthesized a double Z-scheme CBS-2 photocatalyst using the in situ generation method and found that it exhibited stronger photocatalytic performance than the constituent parts. The best sample, CBS-2 composite, achieved 89.7% degradation efficiency of TC solution under visible light irradiation. The value of k was about 2.34, 20.16, and 44.17 times those of BSO, CBO, and SNO. The active species •O 2 − and h + all participated in the removal of TC, in which •O 2 − plays a major role. Considering the results of the capture experiment and ESR characterization, the electron transfer pattern follows the Z-scheme heterojunction. The Z-scheme heterojunction suppressed the recombination of photogenerated carriers and retained the strong redox ability of the catalyst. Furthermore, it possessed good stability after cycle experiments. Hence, the dual Z-scheme CBS-2 photocatalyst has a huge application prospect for removing wastewater under visible light. In a word, we advance the study of Z-heterojunction photocatalysis, then provide a new method to improve the photocatalytic activity of BSO, and finally find a photocatalytic material that can degrade tetracycline efficiently.