Carbon-doped flower-like Bi2WO6 decorated carbon nanosphere nanocomposites with enhanced visible light photocatalytic degradation of tetracycline

In search of a recyclable catalyst with synergistic adsorption and photocatalysis, unique composite photocatalysts of flower-like bismuth tungstate (Bi2WO6) and carbon nanospheres (CSs) were composited using a hydrothermal synthesis method (named CSs-Bi2WO6). Notably, based on the high visible light utilization and a reasonable band gap (2.53 eV), CSs-Bi2WO6 have good photocatalytic properties. For example, the composite with an optimized ratio (2% CSs-Bi2WO6) showed good adsorption and photocatalytic performance. Under simulated natural light conditions, the degradation rate of tetracycline (TC) by 2% CSs-Bi2WO6 was 84.6% in 60 min, which is nearly 25% higher than pure Bi2WO6. After five cycles, the observed barely decreased TC degradation rate of 2% CSs-Bi2WO6 confirmed the high cyclability and reproducibility of the photocatalyst. The total organic carbon estimation of the post-degradation reaction medium corresponded to 68.2% mineralization. Furthermore, we determined the photocatalytic reaction path by LC–MS, which confirmed that the composite catalyst could effectively degrade TC molecules into small molecules. It can be concluded that the catalyst has a broad application prospect in the field of wastewater treatment. Carbon-doped Bi2WO6 composite catalyst was used to degrade TC in wastewater efficiently.


Introduction
With the acceleration of global industrialization, environmental pollution has brought serious and inevitable questions to human life [1][2][3][4][5]. The polluted water seeps into the ground, which further leads to different degrees of pollution of soil and groundwater [6,7]. Therefore, environmental remediation is a necessary task for human beings to survive on Earth. In the past few decades, antibiotics have been diffusely used in the field of biological disease control. Among them, TC has high efficiency in the inhibition and even removal of a variety of pathogenic bacteria, ranking second in the list of global antibiotic production and consumption [8,9]. Because TC has a stable aromatic ring molecular structure, it is difficult to degrade under natural conditions, which is easy to increase the resistance of receptors, and then threaten human health [10,11]. Therefore, it is critical to develop efficient and cost-effective methods to control TC in the environment.
Among many methods of sewage treatment, adsorption is considered to be one of the key methods to alleviate water pollution due to its high efficiency and recyclability [12]. We note that the applications of carbon materials in supercapacitors, energy storage, drug delivery, and especially catalysis are widely concerned [13][14][15], since the carbon component can provide high specific surface area and a high visible light utilization, which will effectively enhance the absorption of visible light. In addition, they can also form 1 3 9 Page 2 of 13 strong interfacial electronic effects with semiconductors [16,17], which play a positive role in improving the degradation performance of catalysts. In general, the easy stacking of most carbon materials in the preparation process is an important reason that affects the material properties. Therefore, the preparation of uniform and stable carbon materials is our key research topic. On the other hand, carbon materials can only adsorb pollutants in the environment and cannot be mineralized into small molecules, thus producing secondary pollution, which limits their adhibition [18,19]. To solve the problems we mentioned, a strategy of combining adsorption and photocatalysis was proposed: the pollutants are enriched on the surface of carbon materials, and it was further decomposed and mineralized by photocatalysts into small molecules [20,21].
In recent years, research on photocatalysis technology in energy conversion and environmental remediation has been in full swing [22,23]. Many Bi-based photocatalysts have been modified by carbonaceous materials [24,25]. Bi 2 WO 6 is one of the semiconductor materials with environmental protection, high light stability, and nontoxicity. Due to its special sandwiches-like layered structure of [Bi 2 O 2 ] 2+ and [WO 6 ] 2− , Bi 2 WO 6 has a better optical property and stability than monolayer catalysts [26]. The valence band of Bi 2 WO 6 is formed by the hybridization of Bi 6 s orbital and O 2p orbital, and the W 5d orbital also constitutes the conduction band in the meantime [27][28][29]. It shows a suitable band gap value (2.75 eV) and visible light response range for photocatalytic reactions [30]. It is becoming one of the most promising photocatalysts. However, as a result of the fast recombination rate of photogenerated e − and h + , the application of pure Bi 2 WO 6 photocatalyst is greatly limited [31]. According to many reports, we have noted that carbon materials have the function of capturing and transporting photogenerated electrons, thus improving photocatalytic performance. For instance, Li et al. [32] modified the Bi 2 WO 6 material with carbon materials, and the results showed that the degradation rate of TC (20 mg/L) was 87% in 70 min. Cai et al. reported a direct S-scheme electron transfer mechanism in CZS/CDs/BWO composite with carbon dots as the electron bridge, which exhibited a wider absorption wavelength compared to the pristine Bi 2 WO 6 [33]. Similarly, other photocatalysts based on Bi 2 WO 6 composite with carbon materials, such as CDs/Cl-Bi 2 WO 6 [34], SCFs/BWO [35], BWO QDs/MCNOs [36], and Bi 2 WO 6 /rGO [37], were found to exhibit photocatalytic degradation. The above studies have well proven its excellent photocatalytic activity.
In this paper, the composite photocatalyst of CSs-Bi 2 WO 6 was prepared by hydrothermal synthesis. It is worth noting that the high dispersion of CSs avoids the aggregation and stacking of catalysts. Moreover, the photocatalytic performances of CSs-Bi 2 WO 6 with different proportions have been measured by the photodegradation of TC solution, which presented greatly enhanced photocatalytic activity compared with pure Bi 2 WO 6 . The composite catalyst has an excellent photocatalytic degradation effect. Furthermore, according to the intermediate substances produced in the degradation process, a possible photodegradation mechanism driven by visible light was proposed.

Chemicals and characterization
Detailed information on experimental reagents and characterization was provided in the supporting information.

Preparation of CSs
CSs were prepared by conventional methods [38]; 4 g of D-( +) -glucose was added to 30 mL of deionized water. After stirring for 30 min, the mixture was transferred to a 50-mL Teflon-lined steel autoclave and heated to 180 °C for 8 h. The black-brown carbonaceous material obtained by the hydrothermal reaction was washed three times by centrifugation with water and ethanol, respectively. Finally, the centrifuged products were dried in an oven at 80 °C for 6 h.

Preparation of CSs-Bi 2 WO 6
Bi 2 WO 6 were prepared according to the hydrothermal procedure as described elsewhere [39]. In short, about 0.5 mol of Bi(NO 3 ) 3 5H 2 O and 0.5 mol Na 2 WO 4 2H 2 O were dissolved in 25 mL of ethylene glycol and stirred to produce a white precipitate. The CSs were dissolved in 10 mL distilled water and then sonicated for 10 min, which was then added dropwise to the above solution. The mixture was then transferred to a Teflon-lined and heated at 140 °C for 14 h. The samples obtained were washed three times with ethanol and deionized water respectively and then centrifuged. The repeatedly washed sediment was dried at 60 ℃. The composite catalysts with different CSs contents were labeled as Bi 2 WO 6 , 0.5% CSs-Bi 2 WO 6 , 1% CSs-Bi 2 WO 6 , 2% CSs-Bi 2 WO 6 , 4% CSs-Bi 2 WO 6 , 8% CSs-Bi 2 WO 6 .

Photocatalytic TC degradation
Details of photocatalytic TC degradation are provided in the supporting information.

Adsorption kinetics study
Details of adsorption kinetics are provided in the supporting information. 3 Results and discussion Figure 1a shows that Bi 2 WO 6 is a typical 'flower-like' structure with a diameter of 3-4 μm. It consists of a large number of nanosheets that cross over and come together to form a 3D structure. As we can see from Fig. 1b, CSs and Bi 2 WO 6 are interconnected. However, CSs were clustered on the surface of Bi 2 WO 6 , and the end is easier to be covered [40]. The primary reason for this special structure may be that the Bi 2 WO 6 sheet is stacked at one end and the stress distribution on the three-dimensional structure surface is not uniform. The structure and morphology of the prepared CSs are shown in Fig. 1c, from which we can find that the CSs are uniform and well-dispersed nanospheres with an average diameter of about 450 nm. The particle size distribution was shown in Fig. S1 in supporting information. The surface of CSs in Fig. 1b changed from smooth to rough, which may be affected by the growth of Bi 2 WO 6 [41]. It could not prevent the CSs from forming close contact with Bi 2 WO 6 to form a 'flower-spheres' structure and will even be more conducive to the surface adsorption of organic pollutants. However, if the content of CSs increases, it is easy to form large clusters. This will affect the interaction between CSs and Bi 2 WO 6 . The EDS results in Fig. 1d-g demonstrate the distribution of C, O, W, and Bi elements in the CSs-Bi 2 WO 6 composites, indicating the successful production of the composite catalysts.

Morphology and structures
In Fig. 1h, it can be seen that the 2% CSs-Bi 2 WO 6 composite structure exhibits the 'flower-spheres' morphology. The layered structure of Bi 2 WO 6 can be seen in the picture as consisting of a large number of nanosheets. The CSs are attached to the nanosheets. The TEM image of pure Bi 2 WO 6 and the magnified structure of the 2% CSs-Bi 2 WO 6 in Fig. 1h were shown in supporting information (Fig. S2). Figure 1i shows the HRTEM image of Bi 2 WO 6 . We observed lattice fringes of 0.27 nm and 0.32 nm matched perfectly with the (020) and (131) crystal faces of Bi 2 WO 6 , which proved the successful synthesis of Bi 2 WO 6 [42]. The HRTEM image (Fig. 1j) depicts the good crystallization of 2% CSs-Bi 2 WO 6 nanoparticles but an amorphous structure for CSs [43]. The above indicated that the composite catalyst was successfully prepared.
The XRD patterns showed the phase structure of the CSs-Bi 2 WO 6 complexes in different proportions (  [44,45]. Figure S3 further showed the XRD image of CSs. It exhibited a broad peak at around ~ 44°, corresponding to (101) reflections of carbon materials. In the XRD pattern of the composite, the diffraction peak of the amorphous structure of CSs phase was not found. This is to some extent expected because of the low crystallinity of amorphous carbon and the incorporation of trace carbon materials during the preparation [46].
The chemical states of Bi 2 WO 6 and 2% CSs-Bi 2 WO 6 were analyzed by XPS. Figure 2b shows the existence of elements C, O, Bi, and W in the sample. The elements of Bi and W peaks were found in supporting information of Fig. S4. Figure 2c shows C 1 s peak for pure Bi 2 WO 6 and 2% CSs-Bi 2 WO 6 . All elements' binding energy was calibrated with C 1 s of aliphatic carbon at 284.60 eV. The C 1 s peak at around 288.24 eV of Bi 2 WO 6 was attributed to the adsorbed CO 2 on the surface [47]. The slight shift of C 1 s peak at 288.35 eV was speculated of 2% CSs-Bi 2 WO 6 to be caused by C = O bond generated by the introduction of BCs. Notably, 2% CSs-Bi 2 WO 6 shows a new peak at 286.2 eV, which may be due to the formation of C-O-C between Bi 2 WO 6 and CSs (O atom in Bi 2 WO 6 ). The above result indicated that most of C in 2% CSs-Bi 2 WO 6 was sp 2 hybridized. The presence of oxygen-containing functional groups in the 2% CSs-Bi 2 WO 6 system suggests that the strong interaction between Bi 2 WO 6 and carbon nanospheres is formed during the hydrothermal reaction.
As shown in Fig. 2d, the asymmetric peak centered of O 1 s at 530 eV was decomposed into two components at the binding energy of 529.74 eV and 531.33 eV for pure Bi 2 WO 6 , which were due to the surface lattice oxygen and the adsorbed oxygen species respectively [48,49]. With the addition of the CSs, the binding energy shifts from 529.75 eV, 531.33 eV (Bi 2 WO 6 ) to 531.08 eV, 527.68 eV (2% CSs-Bi 2 WO 6 ). The peaks of O 1 s of 2% CSs-Bi 2 WO 6 shifted to the lower binding energy compared to Bi 2 WO 6 , indicating the electron cloud density and electronegativity around O decrease due to the addition of carbon spheres. the interactions between Bi 2 WO 6 and BCs in the composite [50,51].
Nitrogen sorption measurements evaluated the surface area and pore structural properties in Fig. 3. In Fig. 3a, the isotherm curves belonged to type IV with H3 hysteresis loops [43,52,53]. And specific surface area of pure Bi 2 WO 6 , 0.5% CSs-Bi 2 WO 6 , 1% CSs-Bi 2 WO 6 , 2% CSs-Bi 2 WO 6 , 4% CSs-Bi 2 WO 6 , and 8% CSs-Bi 2 WO 6 was 20.76, 82.45, 84.80, 84.88 85.04, and 82.31 m 2 g −1 , respectively. Obviously, the addition of surface area was mainly caused by the surface folds of CSs, which will increase the contact area of CSs with contaminants. Interestingly, the surface area of the complex actually decreased when an excess of CSs was introduced. This may be due to the fact that with the increase of CSs, the packing of Bi 2 WO 6 became dense, the specific surface area decreased slightly, and the active site decreased [54,55]. This proves that adsorption is strongly related to the specific surface area of the composite catalyst. The pore size distribution is shown in Fig. 3b; the pore size of samples is almost distributed from 1 to 20 nm. Furthermore, the slit-shaped pore structure displayed by Bi 2 WO 6 is consistent with the sheet morphology shown by SEM results.

Optical and electronic properties
It is well known that the utilization of visible light mainly depends on the band gap (Eg) and energy level of the semiconductor photocatalyst. The optical absorption abilities and bandgap of prepared samples were measured by DRS in Fig. 4. Figure 4a shows the comparison of UV-VIS DRS optical absorption characteristics between pure Bi 2 WO 6 and 2% CSs-Bi 2 WO 6 . It can be observed that the optical absorption capacity of 2% CSs-Bi 2 WO 6 at 300-800-nm wavelength is higher than that of Bi 2 WO 6 . In addition, after the introduction of CSs, the light absorption capacity of the sample has undergone a significant redshift. It shows that the composite photocatalyst has higher light utilization efficiency, promotes the transfer of charge, and improves photocatalytic performance. The band gaps energies (E g) were calculated as the following formula [41]: where A represents a constant, and α, ν, and E g represent the adsorption coefficient, optical frequency, and bandgap, respectively [56]. The electronic transition type of the semiconductor determines the value of n [57]. Referring to previous studies, Bi 2 WO 6 is an indirect bandgap semiconductor with an N value of 4 [58].
In Fig. 4b, the band gap of the pure Bi 2 WO 6 was estimated to be 2.75 eV, which was consistent with the reported experimental values [59]. The band gap of the 2% CSs-Bi 2 WO 6 (1) ahv = A(hv − Eg) n 2 sample was about 2.53 eV, which was lower than that of Bi 2 WO 6 .
Photoluminescence (PL) spectra is considered to be an important experiment to discuss the efficiency of e − and h + pair separation. Figure 5a showed the strong emission peaks of the samples at 425 nm under excitation at 378 nm. In general, the emission intensity of CSs-Bi 2 WO 6 was lower than that of pure Bi 2 WO 6 , and the 2% CSs-Bi 2 WO 6 showed the lowest emission intensity, which proved its better photocatalytic activity. This trend suggested that the separation efficiency of photogenerated e − and h + can be effectively improved by introducing CSs, which promotes molecular diffusion and transfer.
In Fig. 5b, the flat-band (C FB ) potential of the composite photocatalyst is derived from the X-intercept of the linear region. And the obtained results can be further transformed into NHE potential [29,60,61]. The straight upward curves demonstrated that Bi 2 WO 6 was an n-type semiconductor. The C FB potentials of both the Bi 2 WO 6 and 2% CSs-Bi 2 WO 6 composites were − 0.55 V vs. Ag/AgCl (− 0.35 eV vs. NHE). The E CB for both the Bi 2 WO 6 and 2% CSs-Bi 2 WO 6 Schottky contacts was approximately − 0.45 eV. Combined with the E g values obtained from Tauc plots in Fig. 4b, the valence band (VB) potentials of the 2% CSs-Bi 2 WO 6 sample were calculated to be 2.08 eV [62].
The photocurrent responses of the pure Bi 2 WO 6 and 2% CSs-Bi 2 WO 6 were tested, and the results are displayed in Fig. 5c. Obviously, both the photoelectrodes of all samples show stable photocurrent responses over several switching cycles. All samples immediately produce a photocurrent when the electrode is momentarily exposed to visible light. On account of the higher recombination rate of photogenerated e − and h + in the Bi 2 WO 6 crystal, it can be found that Bi 2 WO 6 had lower photocurrent responses. The photocurrent response of the composite photocatalyst 2% CSs-Bi 2 WO 6 was about 9 times higher than pure Bi 2 WO 6 , which effectively indicated more efficient photo-induced charge separation and faster electron transport. Figure 5d shows the EIS Nyquist plots for Bi 2 WO 6 and 2% CSs-Bi 2 WO 6 under different light conditions, respectively. The graph demonstrated that the arc radius of Bi 2 WO 6 was larger than 2% CSs-Bi 2 WO 6 , indicating that the introduction of CSs enhanced the charge migration of Bi 2 WO 6 and reduces the reaction resistance at the semiconductor interface. Furthermore, the EIS semicircle radius of 2% CSs-Bi 2 WO 6 under light is smaller than its under dark condition, which indicates that light can excite more charge carriers and enhance photocatalytic activity [63].

Photocatalytic tests
The photodegradation result of samples is shown in Fig. 6. TC was selected as the target pollutant to test its photocatalytic performance. Under visible light, 50 mg photocatalyst was added to 100 mL 50 mg/L TC solution for photocatalytic degradation. Figure 6a is a diagram of the removal rate of TC by the catalyst. The composite photocatalyst reached adsorption-desorption equilibrium within 40 min of the dark reaction (adsorption kinetics data in Fig. S5). The degradation rate of TC by 2% CSs-Bi 2 WO 6 was 84.6%, which is nearly 25% higher than pure Bi 2 WO 6 . This was because 2% CSs-Bi 2 WO 6 had a large specific surface area, especially the folded structure of CSs, which effectively increased the adsorption of TC by the catalyst and promoted photoreaction. Notably, we found that the degradation effect of 0.5% CSs-Bi 2 WO 6 was slightly higher than that of 1% CSs-Bi 2 WO 6 . In order to further explore its regularity, we added the photocatalytic degradation data of TC by 0.2% CSs-Bi 2 WO 6 and 0.8% CSs-Bi 2 WO 6 in the supporting information (Fig. S6) [64,65]. However, even with the addition of a trace amount of CSs, the photocatalytic performance of the composite photocatalyst is still higher than that of pure BW, which indicates that CSs can effectively enhance the utilization of visible light and form a strong interfacial electronic effect with semiconductor and then improve the photocatalytic efficiency. In the Supporting Information of Table S2, we compared recent literature reports on the degradation of TC by carbon materials. Figure 6b shows the quasi-first-order degradation kinetics of the catalyst composite with a linear relationship between irradiation time and ln (C 0 /C t ). The degradation rate constant was worked from the slope of the kinetic curve. The rate constants for Bi 2 WO 6 , 0.5% CSs-Bi 2 WO 6 , 1% CSs-Bi 2 WO 6 , 2% CSs-Bi 2 WO 6 , 4% CSs-Bi 2 WO 6 , and the 8% CSs-Bi 2 WO 6 were 0.01502 min −1 , 0.00257 min −1 , 0.02141 min −1 , 0.03929 min −1 , 0.02951 min −1 , and 0.0703 min −1 , respectively. In particular, 2% CSs-Bi 2 WO 6 had the highest photocatalytic degradation rate, which should be put down to more efficient charge separation by adding CSs. The stability and reproducibility of photocatalysts are of great research value in practical applications. Figure 6c shows the photocatalytic cycling test plots of the prepared catalyst samples. After the 5th cycle, the final degradation rate was about 80.5%, which displayed better photocatalytic stability of 2% CSs-Bi 2 WO 6 .
We used different capture scavengers under visible light to explore the main active substances. In the present study: dimethyl sulfoxide (DMSO), ammonium oxalate (AMO), 1,4-benzoquinone (BQ), and isopropyl alcohol (IPA) were used as electron (e − ), hole (h + ), superoxide radical (·O 2 − ), and hydroxyl radical (·OH − ) scavengers, respectively [66]. In Fig. 6d. the presence of BQ and AMO significantly inhibited the photodegradation of TC, indicating that ·O 2 − was the main active species. Photodegradation was also somewhat inhibited when AMO was added into the system, indicating that the h + was more active in the composite photocatalyst. In addition, the addition of IPA and DMSO had little effect on the photocatalytic performance. In short, the order of action of the active species in the 2% CSs-Bi 2 WO 6 photocatalytic degradation TC was ‧O 2 − > h + > e − > ‧OH − In Fig. 6e, a further identification test of the free radical O 2 − was carried out using electron spin resonance (ESR). No ESR signal was observed in the dark condition. When 2% CSs-Bi 2 WO 6 and DMPO were mixed under light for 5 and 10 min, a set of characteristic peaks attributed to ·O 2 − was found. This indicates that O 2 − was produced under light conditions in the catalytic system. The signal intensity of ·O 2 − increased with the increase of irradiation time. Relevant studies have shown that 2% CSs-Bi 2 WO 6 photocatalytic TC not only has a higher degradation rate, but also a higher degree of organic matter mineralization [9,67,68]. Therefore, we studied the removal efficiency of total organic carbon (TOC), and the results are shown in Fig. 6f. It seems that in the 90-min degradation process, the reduction curve of TOC is mainly based on the degradation curve, and the degradation rate is 68.17%. TOC ratio was slightly higher than residual antibiotic concentration ratio, indicating that some intermediates were not completely degraded or degraded for a long time.

Photocatalytic mechanism
The intermediates of 2% CSs-Bi 2 WO 6 photodegradation of TC were identified by LC-MS technology, and the degradation pathway of TC was further described. Twelve intermediates with m/z values of 481,475,459,445,415,388,344,318,300,274,250,246,242, and 165 were generated under visible light irradiation. According to Fig. 6d, ·O 2 − was the main active substance in the degradation process. Therefore, the photodegradation pathway of TC can be proposed (Figs. 7 and S7) [69,70].
After visible light irradiation, TC was invaded from three possible sites. Pathway 1: The main reaction pathway of TC degradation is the reaction of superoxide radical attacking phenol group. Under light conditions, it was further oxidized to form a macromolecular intermediate with m/z = 475. Then, the product with m/z = 475 was converted to the product with m/z = 459 by attacking the ketone group and changing the hydroxyl group [71]. Finally, the unstable intermediate cleaved due to the loss of the hydroxyl group, resulting in the formation of small molecules with m/z of 344, 300, and 242, respectively. They are eventually mineralized into H 2 O and CO 2 . Pathway 2: The formation of m/z = 481 is due to the attack of the double bond of TC, which further introduces hydroxyl and ketone groups [72]. In the next period, the product of m/z = 246 and 165 appears due to the opening of the benzene ring by h + attack. Eventually, TC mineralized into small molecules [73]. Pathway 3: the product of m/z = 338 was attributed to the loss of the N-methyl group. The product with m/z = 318 is formed due to further cleavage of the C-ring and loss of the hydroxyl group. Furthermore, the small molecule with m/z = 318 was transformed into a product with m/z = 274 by removing the carboxyl group. Finally, the product of m/z = 250 was obtained after undergoing a different process of detachment of hydroxyl and aldehyde groups [74]. With increasing reaction time, oxidative decomposition and ring-opening reactions further occur, and the above intermediate products are finally decomposed into H 2 O, CO 2 , and NH 4 + , etc. According to the above experiments, Fig. 8 shows the mechanism of photocatalytic degradation of TC by 2% CSs-Bi 2 WO 6 under visible light irradiation. The CB of 2% CSs-Bi 2 WO 6 was − 0.45 eV, which was more negative than the redox potential (E O2/‧O2 − = − 0.33 eV). Thus, the main reason for TC degradation was that the photogenerated electrons were transferred to O 2 and further converted to ‧O 2 − . In addition, combined with the E g energy of 2% CSs-Bi 2 WO 6 , the VB potential of 2% CSs-Bi 2 WO 6 can be deduced as + 2.08 eV according to the equation (E CB = E VB − E g) [75]. Specific data are described in Fig. 4. The VB level of 2% CSs-Bi 2 WO 6 was lower than E ‧OH/OH − = + 2.31 eV. Therefore, ·OH − did not play a major role in the photocatalytic reaction. Under visible light irradiation, the semiconductor photocatalyst generates electron-hole pairs. The photogenerated electrons reacted with O 2 (from the air) on the photocatalyst surface to form the ‧O 2 − . TC is eventually mineralized into small molecule compounds, such as water and carbon dioxide [38,76]. To sum up, the 2% CSs-Bi 2 WO 6 photocatalyst has higher electron-hole separation efficiency than the bare Bi 2 WO 6 photocatalyst. Further, the photodegradation reaction process of TC has been proposed as Eqs. (2)-(6): During the dark reaction, TC was adsorbed on the surface of CSs due to irregular folds on the surface and strong adsorption. It facilitated the transfer of pollutants from the environment to the surface of the catalyst, forming a high concentration TC environment on the surface. And finally adsorption-dissolution equilibrium was achieved. In addition, CSs are more efficient at harvesting light by multiple scattering and utilizing photons. The composite can effectively separate the e − /h + pairs when photogenerated electrons in bismuth tungstate migrate to the CSs, thus improving the photocatalytic performance.

Conclusion
In summary, the composite photocatalyst of CSs-Bi 2 WO 6 was synthesized by a simple hydrothermal method. The experimental results showed that the composite showed good photocatalytic performance (84.6%) and stability for h + + TC → products tetracycline degradation under the optimal CSs loading amount (2 wt%). The degradation kinetics of TC in CSs modified Bi 2 WO 6 was 2.6 times that of pure Bi 2 WO 6 . This may be because CSs can enhance the absorption of visible light by the composite catalyst and effectively enhance the separation efficiency of e − /h + . It is worth noting that the TOC estimation of the post-degradation reaction medium corresponded to 68.2% mineralization. Based on the capture experiment and ESR measurement results, the free radical ‧O 2 − plays a leading role in the photocatalytic reaction. Further, the photocatalytic reaction mechanism and degradation pathway of 2% CSs-Bi 2 WO 6 were revealed in detail by using LC-MS determined reaction intermediates. This study provides a new idea for photocatalytic degradation of organic pollutants through the synergistic action of carbon and semiconductor materials.

Competing interests
The authors declare no competing interests.