Liang Zhou , Yunfeng Li ,＊, Yongkang Zhang , Liewei Qiu ,＊, Yan Xing

1 College of Environmental and Chemical Engineering, Xi’an Key Laboratory of Textile Chemical Engineering Auxiliaries,Xi’an Polytechnic University, Xi’an 710048, China.

2 Jilin Provincial Key Laboratory of Advanced Energy Materials, Department of Chemistry, Northeast Normal University,Changchun 130024, China.

Abstract: With the rapid development of industrial technology, a large number of organic pollutants are routinely released into the environment, which has caused serious problems. Semiconductor photocatalysis is an environmentally-friendly and effective method to degrade and remove typical pollutants, and photocatalysts play a key role in the application of this technology. Therefore, various semiconductor materials have been tried and used in the field of pollutant removal. Graphite carbon nitride (g-C3N4) has attracted great interest because of its two-dimensional layered structure and good visible light response range. Owing to a narrow bandgap, adjustable band structure, and high physicochemical stability, g-C3N4 absorbs wavelengths up to 450 nm in the visible spectrum, leading to an opportunity for visible-light photocatalytic performance. Nevertheless, there are still some drawbacks that limit the photocatalytic efficiency of g-C3N4 in the removal of antibiotics and dyes under visible light, such as the rapid recombination of photoinduced charges and the weak oxidation capacity of holes. To advance this promising photocatalytic material, multiple methods have been tried to optimize the electronic band structure of g-C3N4, such as doping with various elements, morphology control, and functional group modification. Recently, a novel type of Step-scheme (S-scheme) heterojunction composed of two n-type semiconductor photocatalysts has been proposed, which can utilize a more positive valance band and a more negative conduction band. It was demonstrated that the formation of S-scheme heterojunctions is a valid way to increase photocatalytic activity of g-C3N4. Herein, novel 0D/2D Bi4V2O11/g-C3N4 S-scheme heterojunctions were prepared by a simple in situ solvothermal growth method. The Bi4V2O11/g-C3N4 composites displayed a high photocatalytic activity through the removal of oxytetracycline (OTC) and Reactive Red 2. In particular, the BVCN-50 composite showed the highest degradation efficiency for OTC of 74.1% and for Reactive Red 2 of 84.2% with ?O2- as the primary active species.This highly improved photocatalytic performance can be ascribed to the generation of S-scheme heterojunctions, which provides for a high redox capacity of the heterojunction system (strong oxidative ability of Bi4V2O11 and strong reductive capacity of g-C3N4) and facilitates the space separation of photo-generated charges. Moreover, the surface plasmon resonance effect of metallic Bi0 broadens the light utilization range of the heterojunction system. In addition, the possible degradation pathway and intermediates throughout the degradation process of OTC based on liquid chromatograph mass spectrometer (LC-MS) analysis were also studied. This work provides a novel tactic for the design and fabrication of g-C3N4-based S-scheme heterojunctions with enhanced photocatalytic performance.

Key Words: S-scheme; Photocatalysis; Oxytetracycline; g-C3N4

1 Introduction

With the rapid development of industry, large numbers of organic pollutants are misused in the various fields of life, which brings serious environmental problems1-4. It is well known that semiconductor photocatalytic technology is an eco-friendly and valid way to degrade and remove the typical contaminants, and photocatalyst plays a key role in the application of this technology5-8. Thus, the various semiconductor materials have been used in the field of pollutants removal. Graphitic carbon nitride (g-C3N4) has been recognized as a star material, which has attracted great interest due to its two-dimensional layered structure and well visible-light response range9-12. It is demonstrated that, owing to a narrow bandgap (～2.7 eV),adjustable band structure and high physicochemical stability, g-C3N4attains 450 nm of the whole solar spectrum, leading to an opportunity for visible-light photocatalytic performance13-15.Nevertheless, there are still some drawbacks limiting the photocatalytic efficiency of g-C3N4in the removal of antibiotics and dyes under visible light, such as the rapid combination of photoinduced charges and the weak oxidation capacity of holes,which will reduce the efficiency of surface redox reaction16-18.To advance this promising photocatalytic material, multiple tactics have been tried to optimize the electronic band structure of g-C3N4, such as doping with metal/non-metal elements19,20, morphology regulation21,22, and functional group modification23,24,etc.

Among them, it is an effective strategy to construct the heterojunction photocatalysts that has attracted much attention because it can effectively restrain the photo-generated charges recombination25-27. Although traditional type-II heterojunction seems to inhibit electrons and holes recombination, it impairs the redox reaction capacity of photoinduced carriers28-30.Meanwhile, the interfacial carrier transfer mechanism of Zscheme heterojunctions needs to be further supplemented and improved31-33. With the aim of thoroughly overcoming the defects of conventional heterojunction, Yuet al.prepared a novel type of S-scheme heterojunction composed by twon-type semiconductor photocatalysts, which can utilize a more positive valance band and a more negative conduction band of semiconductor materials34. Researchers have made a great effort in constructing the S-scheme heterojunction, such as Ag2WO4/WO335, CeO2/PCN36, g-C3N4/Bi/BiVO437,Bi2MoO6/g-C3N438, TiO2@ZnIn2S439, NiS2/MoSe240, PDIAla/S-C3N441heterojunctions,etc.

Bismuth-based semiconductor photocatalysts can display a visible-light-driven photocatalytic performance with the valence band formedviahybridization of the oxygen 2porbitals (O 2p)and the metal orbitals (Bi 6s), which can degrade the typical organic pollutants deeply. Thus, bismuth-based photocatalysts have been intensively reported, such as Bi2MoO6, Bi2WO6,BiVO4, Bi2O3, BiOX (X = Cl, Br, I),etc42-44.Especially,Bi4V2O11has been considered as a highly promising photocatalytic material owing to its peculiar structure and optical properties. As a member of Aurivillius, Bi4V2O11has a laminar structure that involves of Bi2O2layers interleaved with vanadium oxide sheets, while the rapid recombination of photogenerated electron-hole pair for pristine Bi4V2O11is the primary impediment for effectively degrade organic contaminants45,46.Therefore, it is very necessary to improve the separation capacity of photo-generated charges to improve the photocatalytic efficiency of Bi4V2O11. For instance, Wenet al.reported a AgI/Bi4V2O11Z-scheme heterojunction photocatalystviaa facile solvothermal process combined within situcoprecipitation route. As a result, the AgI/Bi4V2O11Z-scheme photocatalytic system displayed a high photocatalytic activityviathe removal of sulfamethazine owing to the fact that the formed Z-scheme heterojunction facilitate an efficient separation of photo-generated carriers and maintain a strong redox potential of photo-generated electrons and holes47. Chenet al.reported the preparation of oxygen-induced Bi5+-selfdoped Bi4V2O11nano-tubesviaadjusting the oxygen environment throughout thermal treatment with p-n homojunctions. The Bi5+-self-doping coordination of p-n homojunction retards a rapid photo-generated charges recombination, which can improve the photocatalytic activity of Bi4V2O11for removal of Cr6+48. Therefore, if the novel Sscheme heterojunctions could be prepared by cooperating with the respective advantages of Bi4V2O11and g-C3N4, their photocatalytic activity will be greatly refined.

In this work, the novel Bi4V2O11/g-C3N4S-scheme heterojunction photocatalysts are preparedviaa facilein situsolvothermal growth method. The Bi4V2O11/g-C3N4composites displayed a high photocatalytic activityviathe removal of OTC and Reactive Red 2. Especially, BVCN-50 composite shows the highest degradation efficiency for OTC of 74.1% and Reactive Red 2 of 84.2%, respectively. The highly enhanced photocatalytic performance is ascribed to the generation of Sscheme heterojunction between Bi4V2O11and g-C3N4, which keeps a high redox capacity of heterojunction system, and facilitates the space separation of photo-generated charges.Finally, the possible degradation pathway and the possible intermediates throughout the reaction process of OTC based on LC-MS analysis are also studied.

2 Experimental section

All chemical reagents are analytically pure and used directly without further purification. Sources of chemical reagents and instruments are listed in the Supporting Information

2.1 Synthesis of 2D g-C3N4 photocatalytic materials

The g-C3N4nanosheets was preparedviaheating melamine based on a reported previously23. Typically, 5 g of melamine was heated to 823 K in a cover alumina crucible for 2 h at the rate of 2 K·min-1. After cool down naturally, a yellow g-C3N4was collected for further uses.

2.2 Synthesis of novel Bi4V2O11/g-C3N4 S-scheme heterojunctions

The Bi4V2O11/g-C3N4S-scheme heterojunction photocatalysts were synthesizedviaa facilein situsolvothermal growth method. The typical fabrication step was following: an appropriate amount of g-C3N4nanosheets were dissolved into 7 mL of ethylene glycol solutionviasuccessive sonication with 0.5 h, then 1 mmol NH4VO3and 2 mmol Bi(NO3)3·5H2O were added to 2 mL of ethylene glycol by ultrasonication with 20 min,respectively. The NH4VO3solutions and Bi(NO3)3·5H2O solutions were dropwise added into above suspension and then magnetically stirred to obtain a yellow suspension. Afterward,30 mg of urea was dispersed into the yellow suspensions.Subsequently, the pH of the mixture solution was tuned to 9.0viathe addition of ammonia (25%-28%). Afterwards, the homogeneous suspension was transferred into a 25 mL Teflonlined stainless steel autoclave by a solvothermal method at 180 °C for 12 h. The yellow precipitates were obtainedviacentrifuging and washed several times. The resultant samples were denoted as BVCN-20, BVCN-50 and BVCN-60 when the theoretical weight percentages of g-C3N4in Bi4V2O11/g-C3N4composites were 20%, 50% and 60%, respectively. The pristine Bi4V2O11powder was also synthesized by the similar approach except addition of g-C3N4nanosheets.

2.3 Photocatalytic test

The photocatalytic performance of the materials was measuredviadegradation of OTC (30 mg·L-1, 80 mL) and Reactive Red 2 (20 mg·L-1, 80 mL) under visible light. A Xe lamp was supplied as a simulated solar light source (300 W, with a 420 nm cut-off filter). Before irradiation, in order to reach adsorption-desorption equilibrium, the suspension was processed by ultrasound for 5 min and stirred without light for 30 min. At given time irradiation intervals, 3 mL reaction solution was taken out to remove the photocatalysts completely by centrifugation. The clarified solution of antibiotic OTC and Reactive Red 2 were measuredviaUV-Vis-NIR (Purkinje General, TU-1900) spectrophotometer to record their concentration. The possible intermediates of OTC throughout the degradation process were confirmed based on LC-MS analysis.

3 Results and discussion

3.1 Structure and property analysis

Fig. 1 (A) XRD patterns of the as-prepared Bi4V2O11, g-C3N4 and all the Bi4V2O11/g-C3N4 nanocomposites.(B) FT-IR spectra of the as-prepared Bi4V2O11, g-C3N4 and all the Bi4V2O11/g-C3N4 nanocomposites.

The crystalline phase purity of the different samples was measured by XRD. As depicted in Fig. 1A, the noticeable characteristic peaks of pristine Bi4V2O11sample can be well indexed to the orthorhombic Bi4V2O11phase (JCPDS No. 42-0135). The characteristic peaks of Bi4V2O11at 2θof 11.54°,23.27°, 28.57°, 32.27°, 35.11°, 37.12°, 39.78°, 45.99°, 48.43°,55.29° and 59.14° are ascribed to the (002), (111), (113), (200),(006), (115), (024), (220), (206), (313) and (226), respectively.The pristine g-C3N4displays two obvious peaks at 27.3° and 13.1°, which could be ascribed to (002) and (100), respectively.The lower diffraction peak at 13.1° is assigned to the tri-striazine unit of g-C3N4, and the broad diffraction peak at 27.3° is associated with the arising from the generation of a stack of conjugated aromatic. In addition, the characteristic diffraction peaks of g-C3N4and Bi4V2O11can be seen in the Bi4V2O11/g-C3N4composites, indicating that a co-existence of g-C3N4and Bi4V2O11. Meanwhile, all the diffraction peaks are sharp,indicating that the composites are good crystallinity without any impurity. No obvious XRD diffraction peak of Bi0can be observed due to its low content and high dispersion. The structure and component of the different composition samples were measured by FT-IR spectra (Fig. 1B). All samples show some distinct absorption peaks at 3000-3500 cm-1, which are belonged to the stretching vibration mode of N－H/O－H groups. Additionally, the peaks located at 1200-1600 cm-1originate from the stretching vibration modes of C－N and C＝N heterocycle, whilst the peaks at 810 cm-1may be registered as the breathing mode of tri-s-triazine ring unit.

Fig. 2 SEM images of Bi4V2O11 (A), g-C3N4 (B) and BVCN-50 sample (C). Elemental mapping images of C (D), N (E), O (F), V (G) and Bi (H) for BVCN-50 composite. (I) TEM and SAED images of BVCN-50 composite. (J) HRTEM of the BVCN-50 composite.

Fig. 3 XPS survey spectrum (A) and the corresponding high-resolution XPS spectra: C 1s (B), N 1s (C),V 2p (D), Bi 4f (E), O 1s (F) of BVCN-50 composite.

The SEM, TEM and HRTEM were employed to display the morphology and microstructure of Bi4V2O11, g-C3N4, and BVCN-50 composites. As shown in Fig. 2A, the pure Bi4V2O11is irregular nanoparticle accumulation. The SEM in Fig. 2B indicates that the pristine g-C3N4possesses a 2D sheet structure.For the morphology of the BVCN-50 composite (Fig. 2C,I), the 2D nanosheet of g-C3N4sample can be obviously observed,while the Bi4V2O11nanoparticles are aggregated on the surface of g-C3N4. The results of elemental mapping clearly demonstrate the existence of carbon (C), nitrogen (N), oxygen (O), vanadium(V) and bismuth (Bi) elements (Fig. 2D-H). Furthermore, it is revealed that the Bi4V2O11nanoparticles are well dispersed on the g-C3N4nanosheets. As displayed in the HRTEM pictures(Fig. 2J), the lattice fringes spacings of Bi4V2O11are about 0.156 and 0.167 nm, which are well matched with (119) and (133)crystal plane of Bi4V2O11, respectively. The results are consistent with the SAED patterns (inset in Fig. 2I). This further confirms that the heterojunction is successfully synthesized between Bi4V2O11and g-C3N4, and the lattice spacing of Bi0is about 0.119 nm, which is referred to the (208) crystallographic plane.

The element chemical valence states of the BVCN-50 composites were studiedviaXPS. The survey XPS spectrum of BVCN-50 nanocomposite is displayed in Fig. 3A, which confirms the presence of C, N, O, V and Bi elements. The C 1s in Fig. 3B can be well fitted to 280.5, 282.5, 284.8 and 288.5 eV,respectively. The peaks located at 284.8 and 288.5 eV could be ascribed to C－H and C－Bi bonds. The peaks centered at 288.5 and 284.8 eV could be related tosp2hybridization carbon (NC＝N) and graphitic carbon (C＝C) in g-C3N4, respectively. As shown in the N1s spectra (Fig. 3C), the peaks located at 395.9,398.8, and 401.1 eV are ascribed to C＝N－C bonds, N－(C)3groups and N－H structure. The two peaks in Fig. 3D are attributed to V 2p3/2and V 2p1/2in BVCN-50 sample. The two peaks of Bi 4f5/2around 164.1 eV and Bi 4f7/2around 158.8 eV,respectively, which confirms the existence of Bi3+in the BVCN-50 samples. And the two additional peaks at 156.9 and 162.3 eV demonstrate that a small amount of submetallic Bi0also exist in BVCN-50 samples (Fig. 3E). The Bi4V2O11/g-C3N4composites are prepared by a facilein situsolvothermal growth process with ethylene glycol as solvent. The ethylene glycol always shows a reducibility at high temperature and can be used as reducing agent to generate Bi0. Fig. 3F shows the O 1sXPS, which can be divided into two peaks. The peaks located at 526.1 and 529.5 eV could be ascribed to the external－OH group and the lattice oxygen in Bi－O of BVCN-50, respectively.

The specific surface area and porosity structure were investigated by using nitrogen adsorption-desorption isotherms.Fig. 4 displays isotherm of g-C3N4, Bi4V2O11and BVCN-50 samples can be categorized as a H3-type hysteresis loop and a typical IV isotherm, which is the feature of mesoporous materials. The specific surface area of g-C3N4, Bi4V2O11and BVCN-50 samples are 6.83 m2·g-1, 21.26 m2·g-1and 30.62 m2·g-1, respectively. It is well known that the specific surface area can boost the photocatalytic performance since it is closely linked to the reactive sites to adsorb and trigger the reaction substrates.

Fig. 4 N2 adsorption-desorption isotherms of g-C3N4,Bi4V2O11 and BVCN-50 samples.

Fig. 5 (A) UV-Vis DRS of the as-prepared samples. (B) Tauc plot for bandgap of of Bi4V2O11 and g-C3N4.(C) The Mott-Schottky plots with various frequencies of Bi4V2O11 and g-C3N4. (D) Electronic band structures of Bi4V2O11 and g-C3N4.

The optical properties were characterized by UV-Vis diffuse reflectance spectroscopy (DRS) spectra, as exhibited in Fig. 5A.It is worth noting that the fabricated samples have a good light absorption in the whole spectral ranges. Moreover, a clear absorption of samples the near infrared region can also be observed, which is mainly owing to the surface plasma resonance (SPR) effect of metallic Bi0. The absorption edges of Bi4V2O11and g-C3N4are estimated to be at approximately 647.4 and 467.3 nm, respectively. In particular, the light response range of the composite is significantly increased than with the pristine g-C3N4due to the existence of Bi4V2O11in the heterojunctions.Fig. 5B displays the bandgap energy (Eg) values of pristine Bi4V2O11and g-C3N4are as follows:

α(hν) = A(hν-Eg)n/2

whereαis absorption coefficient,?νis photon energy andΑis a constant, respectively. The g-C3N4is the indirect semiconductor and Bi4V2O11belongs to the direct semiconductor due to thenis 4 and 1, respectively11,46. Hence,the bandgaps of pristine Bi4V2O11and g-C3N4are calculated to be approximately 2.16 and 2.51 eV, respectively. In addition, the flat band potentials of pristine Bi4V2O11and g-C3N4are measuredviathe Mott-Schottky analysis. As shown in Fig. 5C and Fig. S1, both curves have a positive slope of the tangent line,indicating that they are typicaln-type semiconductors. The conduction band (CB) potentials of Bi4V2O11and g-C3N4are confirmed approximately at 0.46 V and -1.12 Vvs.NHE,respectively. On the basis of the measured bandgap values for Bi4V2O11and g-C3N4, their valence band (VB) potentials are about 2.62 V and 1.39 V, respectively. Moreover, the corresponding band structures of Bi4V2O11and g-C3N4samples are calculated and shown in Fig. 5D.

3.2 Photocatalytic activity

The photocatalytic activity of pristine Bi4V2O11, pure g-C3N4and Bi4V2O11/g-C3N4with different molar ratios were investigated by removal of antibiotic OTC and Reactive Red 2 under visible light. Fig. 6A shows the pristine Bi4V2O11and g-C3N4exhibit a lower photocatalytic activity about 31.5% and 34.8% of OTC being eliminated within 30 min. However, the BVCN-50 composite shows the highest OTC degradation efficiency about 74.1%, indicating that photocatalytic activity can be greatly enhanced by constructing g-C3N4/Bi4V2O11Sscheme heterojunction. In addition, the Reactive Red 2 was selected as a reactive dye to further investigate the photocatalytic performance of the prepared BVCN-50 composite. As shown in Fig. 6C, only 12.8%, 49.5% and 65.7% of Reactive Red 2 are eliminated at the same conditions for P25, g-C3N4and Bi4V2O11samples, respectively. In particular, BVCN-50 displays the highest photocatalytic performance and 84.2% of Reactive Red 2 can be removed during 45 min. In order to obtain a more intuitive contrast of the photocatalytic degradation, the corresponding degradation data of OTC and Reactive Red 2 could be fitted with a pseudo-first-order kinetic reaction process(ln(c0/ct) =kt). As shown in Fig. 6B and 6D, it is clear that the highestkvalue is obtained from BVCN-50 sample for degrading of OTC and Reactive Red 2, which is approximately 3.27 and 1.81 times higher than pristine g-C3N4and Bi4V2O11samples,respectively.

Fig. 6 Photocatalytic degradation of (A) OTC antibiotic and (C) Reactive Red 2 by the as-prepared samples under visible-light irradiation.The corresponding degradation kinetics (B) and (D) for the as-prepared samples.

Fig. 7 The specific degradation route and the main intermediates during the degradation process of OTC based on LC-MS analysis.

During the photocatalytic degradation of OTC process, the primary intermediates were identified according to LC-MS analysis, and the detailed degradation pathways were illustrated in Fig. 7. Based on main intermediates and the results reported in previous literature, possible degradation route of OTC is proposed. The product P1 is generatedviaremoval of N-methyl and amino. Then, P1 is cleaved to generate intermediate P4 by the loss of methyl, N-methyl, carbonyl and hydroxyl.Intermediate P2 and P3 are formed through loss of N-methyl and hydroxyl of OTC. Subsequently, P2 is decomposed to produce P5 productviadegradation of amino and N-methyl. Afterwards,the formed P6 product is removedviaa multistep removal process of P2, P3 and P5 products. The obtained degradation products of P4-P6 could continue to be mineralized into the intermediates with low molecular weight (P7-P10) and ultimately converted to CO2and H2O.

3.3 Possible photocatalytic mechanism

Rapid separation rate of photo-generated carriers is an imperative factor to increase photocatalytic performance49-51.Fig. 8A shows the pristine g-C3N4samples display a strong fluorescence emission peak. Nevertheless, the intensity of fluorescence emission peak significantly diminishes after the addition of Bi4V2O11nanoparticles on the surface, which indicates that the separation of photo-generated carriers in BVCN-50 composite is accelerated evidently. Fig. 8B shows that carrier lifetime of BVCN-50 is obviously shorter compared with pristine Bi4V2O11and g-C3N4. This shortened carrier lifetime indicates that a new nonradiative decay pathway may be opened owing to the generation of S-scheme heterojunction between Bi4V2O11and g-C3N4, which facilitates photo-excited electrons/holes transfer in the Bi4V2O11/g-C3N452,53. The recombination rate of photo-excited electron-hole pairs for the BVCN-50 composite is effectively restrained due to the heterojunction generated at the interface of Bi4V2O11and g-C3N4. The electrochemical impedance spectroscopy (EIS) were conducted to further illustrate the charges migration and separation process in Fig. 8C. As we all know, the size of arc radius indicates charges transfer rate at the reaction interface.The BVCN-50 has the lowest arc radius of Nyquist plot, which indicates that its photo-excited electrons/holes can be rapidly separated. The Linear sweep voltammetry (LSV) curves have also been measured in Fig. 8D. Compared with pure Bi4V2O11and g-C3N4samples, the fast photo-generated charges transmission for the BVCN-50 composite is attributed to a lower overpotential over the whole voltage range.

The electron paramagnetic resonance (EPR) spectra were also employed to determine the active radicals by using DMPO as trapping agent54-57. As depicted in Fig. 9A, both the Bi4V2O11and g-C3N4samples exhibit the signal peaks of ·OH radicals,while the signals intensity of Bi4V2O11is obviously stronger than that of g-C3N4owing to a superior oxidation capacity of Bi4V2O11. However, the signals of ·O2-radicals are more obvious(Fig. 9B) for g-C3N4sample, implying that a stronger reduction ability of g-C3N4. It is noteworthy that the BVCN-50 composite possesses the superior redox capacity for information of radicals because it indicates the both distinctive typical signal peaks of·O2-and ·OH radicals. In addition, the trapping experiments were conducted to identify the active radicals produced in photocatalytic process. As illustrated in Fig. 9C, the benzoquinone (BQ) was utilized as a sacrificial agent for superoxide radical (·O2-),tert-butyl alcohol (t-BuOH) as scavenger agent for hydroxyl radical (·OH), and methanol as sacrificial agent for photo-induced holes. The results imply that the ·O2-may perform a primary role in the photocatalytic process.The results of photo-induced charges transfer are consistent with the EPR spectroscopy and trapping experiments (Fig. 9D).

Fig. 8 (A) Steady-state PL spectra of g-C3N4, Bi4V2O11 and BVCN-50 samples. (B) Time-resolved PL spectra of g-C3N4, Bi4V2O11 and BVCN-50 samples. (C) EIS tests and (D) LSV curves of g-C3N4, Bi4V2O11 and BVCN-50 samples.

Fig. 9 EPR spectra of (A) DMPO-·OH adducts and (B) DMPO-·O2- adducts for pure g-C3N4, Bi4V2O11 and BVCN-50 samples.(C) The photo-generated carriers trapped tests during the photocatalytic degradation of OTC BVCN-50 heterojunction.(D) The tendency of photo-induced charges transfer between pure g-C3N4 and Bi4V2O11 samples.

Hence, the charges transfer pathway of S-scheme mechanism over BVCN-50 photocatalysts was proposedviathe above analysis58-61. As illustrated in Fig. 10, g-C3N4has a higher Fermi energy compared with Bi4V2O11. When the Bi4V2O11is tightly contacted with g-C3N4before light irradiation, the electrons on Fermi energy of g-C3N4automatically flow to the Fermi energy of Bi4V2O11until their Fermi energy attains equilibrium.Thereby, an internal electric field (IEF) and a band bending are formed between the g-C3N4surface and Bi4V2O11surface, which extremely promotes the reverse transfer of the photo-induced electrons on CB of Bi4V2O11to VB of g-C3N4. Moreover, the Sscheme heterojunction eliminates the poor reductive electrons in CB of Bi4V2O11and the weak oxidative holes in VB of g-C3N4.As a result, the electrons on CB of g-C3N4and the holes on VB of Bi4V2O11would be retained with the strong redox ability,which effectively promotes separation of photo-excited electrons/holes62-64. In addition, the electrons of partially reduced Bi0nanoparticles would be transferred and accumulated on the CB of Bi4V2O11viautilizing the surface plasmon resonance effect of metallic Bi0to make full use of sunlight.Then, more electrons will transfer from the CB of Bi4V2O11to the VB of g-C3N4to improve the utilization efficiency of charge carriers. The stability of the BVCN-50 photocatalyst was also investigated by the XRD and XPS in Fig. S2. The initial degradation efficiency of OTC by using BVCN-50 photocatalyst is about 74%, while it still maintains a high degradation efficiency of about 65% after five cycle experiments. We can find that the phase structure and surface valence state of BVCN-50 almost unchanged after five photocatalytic cycling experiments.

Fig. 10 The proposed charges transfer mechanism of Bi4V2O11/g-C3N4 S-scheme heterojunction photocatalysts.

4 Conclusions

In summary, the novel 0D/2D Bi4V2O11/g-C3N4S-scheme heterojunction photocatalysts have been preparedviaa facilein situsolvothermal growth method. Compared to the pristine Bi4V2O11and g-C3N4, the BVCN-50 composite displayed a high photocatalytic activityviathe removal of OTC and Reactive Red 2 with a degradation efficiency of 74.1% for OTC and 84.2% for Reactive Red 2, respectively. The possible degradation pathway according to the possible intermediates of OTC are also produced. The highly photocatalytic activity is ascribed to the generation of S-scheme heterojunction between Bi4V2O11and g-C3N4, which inhibits photo-induced carriers recombination, and maintains a high redox ability. Moreover, the SPR effects of submetallic Bi effectively extend light response range and utilization.