All Organic S-Scheme Heterojunction PDI-Ala/S-C3N4 Photocatalyst with Enhanced Photocatalytic Performance

Xibao Li , Jiyou Liu , Juntong Huang , Chaozheng He , Zhijun Feng , Zhi Chen , Liying Wan ,Fang Deng

1 School of Materials Science and Engineering, Nanchang Hangkong University, Nanchang 330063, China.

2 National-Local Joint Engineering Research Center of Heavy Metal Pollutants Control and Resource Utilization, Nanchang Hangkong University, Nanchang 330063, China.

3 Institute of Environmental and Energy Catalysis, School of Materials Science and Chemical Engineering, Xi'an Technological University, Xi’an 710021, China.

Abstract: Organic photocatalysts have attracted attention owing to their suitable redox band positions, low cost, high chemical stability, and good tunability of their framework and electronic structure. As a novel organic photocatalyst, PDI-Ala (N,N’-bis(propionic acid)-perylene-3,4,9,10-tetracarboxylic diimide) has strong visible-light response, low valence band position, and strong oxidation ability. However, the low photogenerated charge transfer rate and high carrier recombination rate limit its application.Due to the aromatic heterocyclic structure of g-C3N4 and large delocalized π bond in the planar structure of PDI-Ala, g-C3N4 and PDI-Ala can be tightly combined through π–π interactions and N―C bond. The band structure of sulfur-doped g-C3N4 (S-C3N4) matched well with PDI-Ala than that with g-C3N4. The electron delocalization effect, internal electric field, and newly formed chemical bond jointly promote the separation and migration of photogenerated carriers between PDI-Ala and S-C3N4. To this end, a novel step-scheme (Sscheme) heterojunction photocatalyst comprising organic semiconductor PDI-Ala and S-C3N4 was prepared by an in situ self-assembly strategy. Meanwhile, PDI-Ala was self-assembled by transverse hydrogen bonding and longitudinal π–π stacking. The crystal structure, morphology, valency, optical properties, stability, and energy band structure of the PDIAla/S-C3N4 photocatalysts were systematically analyzed and studied by various characterization methods such as X-ray diffraction, transmission electron microscopy, energy dispersive X-ray spectrometry, X-ray photoelectron spectroscopy,ultraviolet visible diffuse reflectance spectroscopy, electrochemical impedance spectroscopy, and Mott-Schottky curve. The work functions and interface coupling characteristics were determined using density functional theory. The photocatalytic activities of the synthesized photocatalyst for H2O2 production and the degradation of tetracycline (TC) and p-nitrophenol(PNP) under visible-light irradiation are discussed. The PDI-Ala/S-C3N4 S-scheme heterojunction with band matching and tight interface bonding accelerates the intermolecular electron transfer and broadens the visible-light response range of the heterojunction. In addition, in the processes of the PDI-Ala/S-C3N4 photocatalytic degradation reaction, a variety of active species (h+, ·O2?, and H2O2) were produced and accumulated. Therefore, the PDI-Ala/S-C3N4 heterojunction exhibited enhanced photocatalytic performance in the degradation of TC, PNP, and H2O2 production. Under visible-light irradiation, the optimum 30%PDI-Ala/S-C3N4 removed 90% of TC within 90 min. In addition, 30%PDI-Ala/S-C3N4 displayed the highest H2O2 evolution rate of 28.3 μmol·h?1·g?1, which was 2.9 and 1.6 times higher than those of PDI-Ala and S-C3N4,

Key Words: Visible light photocatalysis; Step-scheme heterojunction; C3N4; π–π interactions; H2O2

1 Introduction

In recent years, with the rapid development of industrialization, water pollution and energy shortage have gradually become a global problem1,2. Photocatalysis technology is widely used to solve environmental pollution problems3–5. For example, the reduction of hexavalent chromium6, the degradation of benzene7, phenol,p-nitrophenol8,organic dye9–12, antibiotic13,14, and the photocatalytic antibacterial are typical representatives15,16. In addition, in the production of pollution-free new energy, photocatalysis is used to produce hydrogen17–22, oxygen, hydrogen peroxide23,24, CO2photoreduction25–29, photocatalytic nitrogen fixation30,etc.Although photocatalysts have many advantages and are widely used, most of them have two fatal defects: low utilization of solar energy, and high recombination rate of photogenerated electron hole pairs31–33. Therefore, it is very vital to develop appropriate photocatalysts with high solar energy utilization and excellent separation efficiency of photogenerated carriers for the application of photocatalysis technology.

At present, the preparation of inorganic semiconductor photocatalyst becomes more and more mature. In contrast, there are few reports on organic semiconductor photocatalysts.Organic semiconductors have the advantages of rich resources,precisely designed molecular structure, adjustable electronic structure and light absorption, diverse self-assembled morphologies and functions34–36. As a classical organic semiconductor material, graphite carbon nitride (g-C3N4) has been widely concerned by researchers due to its simple preparation process, low-cost and good stability37,38. However,due to its inherent properties, g-C3N4can only absorb partial visible light39. Besides, because of the low transmission efficiency of photogenerated carriers in g-C3N4, most of the photogenerated electrons and holes are easy to recombine,resulting in reduced photocatalytic performance of g-C3N440. In order to overcome these shortcomings, a variety of strategies have emerged for g-C3N4, such as surface hybridization,heterojunction construction, manufacturing defects, morphology control,etc41,42. The strategy of constructing heterojunction is an effective method to avoid the deficiency of photocatalyst43.Qinet al.44prepared 2D/2D ZnIn2S4/g-C3N4(ZIS-S/g-C3N4)heterojunction photocatalyst by solvothermal method. It was found that the optimized 30ZIS-S/g-C3N4had the highest H2evolution rate of 6095.1 μmol·h?1·g?1. Most of the reports have focused on the combination of g-C3N4and inorganic semiconductors to construct heterostructures. However, there are few reports about the heterojunction between g-C3N4and organic semiconductors.

As a new organic semiconductor, perylene diimide (PDI)possesses high stability and relatively small band gap45. It has a strong response to visible light and even near-infrared light,which greatly enhances its utilization of solar energy46,47. Zhanget al.48synthesized Bi2WO6/PDI composite by anin situselfassembly method. It was used for photocatalytic hydrogen production and phenol degradation experiments. The results showed that the introduction of PDI greatly enhanced the photocatalytic activity of Bi2WO6/PDI. Daiet al.49prepared polyaniline/PDI all organic heterojunction photocatalyst by anin situgrowth method. The photodegradation rate of tetracycline is 17.0 times and 15.3 times higher than that of polyaniline and PDI, respectively. It has been reported that PDI molecules can be self-assembled into the organic supramolecular structure by a simple preparation process. This structure has a large delocalizedπbond, which is apt to combine with g-C3N4with heprazine ring structure, as well as the advantage of easily adjusting the performance and morphology50.

In order to obtain higher solar energy utilization, the selected photocatalysts must have a smaller band gap. However,photocatalysts with high redox capacity are required to have a higher conduction band position and lower valence band position51. Therefore, it is imperative to construct effective heterojunction. The proposed S-scheme heterojunction is more reasonable than the traditional II-scheme heterojunction, and ensures the high redox capacity of the overall photocatalyst52–56.In order to obtain efficient S-scheme heterojunction photocatalysts, it is necessary to modify the HOMO level of g-C3N4to match the orbital energy levels of PDI. Zhenget al.57synthesized sulfur-doped g-C3N4(S-C3N4) using dimethyl sulfoxide as sulfur source, which obtained a higher energy level and narrower band gap than g-C3N4so that significantly improved the photocatalytic performance. In conclusion, it is very promising to construct a S-scheme heterostructure with high redox properties by means of element doping strategy58.

Therefore, a novel all organicN,N’-bis(propionic acid)-perylene-3,4,9,10-tetracarboxylic diimide hybridised sulfurdoped carbon nitride (PDI-Ala/S-C3N4) photocatalyst with Sscheme heterojunction was designed and synthesized by anin situself-assembly method. The conduction band and valence band of g-C3N4are changed by doping with S element to form the S-scheme heterostructure. The morphology, optical property,stability and energy band structure of PDI-Ala/S-C3N4photocatalysts were systematically analyzed and studied by various characterization methods. The photocatalytic activities of the synthesized photocatalyst in H2O2production and the degradation of tetracycline and PNP under visible light irradiation were discussed in detail. In addition, the enhanced photocatalytic activity of the S-scheme heterostructure can be traced back to the separation and migration of photogenerated carriers and the formation of photocatalytic active species. The mechanism of the photocatalytic process of the S-scheme heterojunction photocatalyst and its structure-activity relationship were studied.

2 Experimental and computational sections

2.1 Preparation of photocatalyst

2.1.1 Raw materials

Urea (CH4N2O) (Shanghai Zhanyun Chemical Co. Ltd.,Shanghai, China), thioacetamide (TAA) (Tianjin Kemiou Chemical Reagent Co. Ltd., Tianjin, China), perylene-3,4,9,10-tetracarboxylic dianhydride (perylene anhydride, PA) (Shanghai Macklin Biochemical Co. Ltd., Shanghai, China), β-alanine(Ala) (Shanghai Macklin Biochemical Co., Ltd.), imidazole(Shanghai Macklin Biochemical Co., Ltd.), triethylamine (TEA)(Shantou Xilong Scientific Co. Ltd., Shantou, China). All reagents were analytically pure and used without further purification.

2.1.2 Preparation of sulfur-doped carbon nitride (S-C3N4)

15 g urea was dissolved in 20 mL distilled water, which was magnetically stirred at 50 °C for 0.5 h, and then 20 mg of TAA was added and stirred for 1 h. The resulting precursor solution was transferred to a 50 mL ceramic crucible placed in a muffle furnace. Subsequently, it was heated at 400 °C for 1 h with a heating rate of 15 °C·min?1, followed by heating at 500 °C for 2 h. After cooling down to room temperature and sieving, the obtained light yellow powder was marked as S-C3N4.

2.1.3 Preparation ofN,N’-bis(propionic acid)-perylene-3,4,9,10-tetracarboxylic diimide

N,N’-bis(propionic acid)-perylene-3,4,9,10-tetracarboxylic diimide was synthesized according to the strategy reported in the reference46. Generally, 17.700 g (0.26 mol) imidazole, 1.373 g(3.5 mmol) perylene-3,4,9,10-tetracarboxylic dianhydride and 2.495 g (28 mmol) β-alanine were added in a 100 mL threenecked flask. Under the atmosphere of argon, the precursor solution was heated and stirred at 150 °C for 6 h. After cooling to room temperature, the reaction products were dispersed and dissolved in 100 mL ethanol. Subsequently, 300 mL HCl (2 mol·L?1) was added and stirred overnight. The precipitates were separated by centrifugation, washed to neutral with deionized water, and dried in vacuum at 60 °C for 12 h. The obtained dark red solid powder was labeled PDI-Ala.

2.1.4 Synthesis of PDI-Ala/S-C3N4byin situselfassembly

The prepared S-C3N4powders (0 g, 0.80 g, 0.47 g, 0.30 g, 0.20 g, respectively) were dispersed in 50 mL deionized water by ultrasound. Then, 0.20 g PDI-Ala was added, and ultrasound was performed for 30 min. After that, 313.0 μL triethylamine (TEA)was added into the precursor solution with a micro injector to stir for 30 min. The crude product was completely dispersed and dissolved to form a dark red solution. Subsequently, 11.50 mL HCl (4 mol·L?1) was added and heated at 60 °C for 2 h, PDI-Ala wasin situself-assembled onto S-C3N4nanosheets. After standing for several hours, the supernatant was removed. The dark red solid was obtained by centrifugation and washing with deionized water for several times, and dried in vacuum at 60 °C.Finally, the solid powders prepared byin situassembly were named asx%PDI-Ala/S-C3N4(xrepresents the mass ratio of PDI to PDI-Ala/S-C3N4composites,x= 0, 20, 30, 40, 50).

2.2 Characterization

The phase composition was characterized by Bruker D8 Advance X-ray diffraction with CuKα(λ= 0.1541 nm) radiation.The microstructures were observed by the FEI Talos F200X transmission electron microscopy (TEM) with energy dispersive X-ray spectrometer (EDS). The infrared spectra including the structural composition and chemical functional groups of the sample were obtained by Nicolet NEXUS 670 spectrometer(Thermo Co. Ltd., Madison, WI, USA). The chemical states were analyzed by the Shimadzu Axis Ultra DLD X-ray photoelectron spectroscopy (XPS) with AlKαradiation. The ultraviolet visible diffuse reflectance spectroscopy (UV-Vis DRS) was detected by the Shanghai Metash UV-9000 ultraviolet visible spectrophotometer equipped with an integrating sphere.Electrochemical analyzer with three-electrode system (Chenhua CHI760E, Shanghai, China) was employed to analyze the photoelectrochemical properties. In addition, Ag/AgCl electrode, Pt sheet (20 × 20 × 0.1 mm), and carbon paper coated with photocatalyst (1 × 1.5 cm, HCP030 N, Shanghai Hesen,China) were used as the reference electrode, counter electrode and working electrode, respectively. A 300 W xenon lamp (PLSSXE 300, Beijing Perfect Light, Beijing, China) was used as visible light source and 0.5 mol·L?1Na2SO4solution as electrolyte, electrochemical impedance spectroscopy (EIS),Mott-Schottky curve (MS) and photocurrent response curve were measured on an electrochemical analyzer. The preparation method of working electrode is as follows: 4 mg sample powder is dispersed in 0.5 mL distilled water. The suspension is dropped on one side of the carbon paper and dried in the oven. The process is operated repeatedly until the surface of the carbon paper is covered with a uniform layer of the sample.

2.3 Calculation method

The work functions were calculated by Density Functional Theory (DFT). The calculations of geometric optimization and electronic properties were implemented in Dmol code of Materials Studio software using density functional theory. The exchange correlation of electrons was performed by the generalized gradient approximation (GGA) method with Perdew-Burke-Ernzerhof (PBE). The van der Waals force was corrected for long-range dispersion through Tkatchenko and Scheffler’s scheme. During geometric structural optimizations,the convergence criteria were set to 10?5Ha for the energy, 0.02 Ha?nm?1for the force, and 0.0005 nm or the displacement. The smearing value was set to 0.005 Ha to achieve self-consistent field convergence. Throughout this work, the work function (Φ)calculations are carried out using Viennaab initiosimulation package (VASP).

2.4 Photocatalytic experiments

The photocatalytic performance was tested by the degradation of tetracycline (TC, 50 mg·L?1) orp-nitrophenol (PNP,10 mg·L?1) in a double-walled cylindrical quartz reactor under visible light irradiation. Firstly, 50 mL of prepared TC or PNP solution and 50 mg photocatalyst were added into the reactor,which were stirred for 0.5 h in the dark to achieve the physical adsorption-desorption equilibrium. Afterwards, the 300 W xenon lamp (PLS-SXE 300, 420 nm <λ< 780 nm) was used as the light source and fixed at 5 cm above the reactor. The photocatalyst samples were collected regularly, the precipitates were removed by centrifugation. For photocatalytic activity test,the concentration of TC was measured by using a high performance liquid chromatography (HPLC, LC-20A,Shimadzu, Japan). The concentration of H2O2produced by photocatalysis was measured by iodine titration. In the experiment, 50 mg photocatalyst was put into 50 mL distilled water and stirred for 1 h in darkness, then irradiated with the 300 W xenon lamp (420 nm <λ< 780 nm), 7 mL suspensions were collected and centrifuged to obtain the upper solution every hour. Next, 0.5 mL supernatant was aspirated and 1 mL of 0.1 mol·L?1potassium hydrogen phthalate and 1 mL of 0.4 mol·L?1potassium iodide were added.P-benzoquinone (p-BQ, 0.1 mmol·L?1), ammonium oxalate (AO, 10 mmol·L?1) and isopropanol (IPA, 10 mmol·L?1) were used as trapping agents for the active radicals of ·O2?, h+and ·OH, respectively. The free radical trapping experiments were carried out under visible light irradiation. The main active species produced in the process of photodegradation and the effects of active species on the performance of photocatalysts were investigated.

3 Results and discussion

3.1 The structure and morphology

The preparation process of the constructed S-scheme heterojunction is illustrated in Fig. 1. Due to the aromatic heterocyclic structure of the light yellow S-C3N4and the large delocalizedπbond of plane structure in the crude product (PDIAla), PDI-Ala wasin situbonded to the surface of S-C3N4nanoflakes due toπ–πinteraction. At the same time, PDI-Ala was self-assembled by transverse hydrogen bonding and longitudinalπ–πstacking, resulting in the formation of PDIAla/S-C3N4photocatalyst.

Fig. 1 The synthesis process of PDI-Ala/S-C3N4 S-scheme heterojunction.

The crystal structure and phase of as-prepared S-C3N4, PDIAla and 30%PDI-Ala/S-C3N4samples were evaluated by using XRD technique. The characteristic peaks of g-C3N4at 13.1° and 27.5° belong to the (100) and (002) planes (JCPDS 87-1526)59.Similarly, Fig. 2a shows two distinct peaks in the XRD pattern of S-C3N4. The peak located at 13.0° is attributed to the (100)crystal plane by the stacking of the 3-s-triazine unit in the planar structure. The strong peak located at 27.0° belongs to the (002)crystal plane of S-C3N4corresponding to the inter-layer stacking of conjugated aromatic rings60. In the XRD patterns of PDI-Ala,high-intensity diffraction peaks were found at 7.1°, 10.0°, 13.9°,20.0°, 22.9°, 24.3°, 25.9° and 29.2°, respectively, which indicated that the purity of PDI-Ala was relatively high. The diffraction peak at 25.9° can be attributed to theπ–πstacking structure50. According to the Bragg equation 2·d·sinθ=nλ, the(002) crystal plane spacing of S-C3N4is 0.33 nm, which is similar to theπ–πstacking layer spacing (0.34 nm) of PDI-Ala.This is beneficial to the formation ofπ–πinteraction between them. For 30% PDI-Ala/S-C3N4, almost all diffraction peaks of PDI-Ala are retained. Fig. 2b demonstrates that the diffraction peak of 30% PDI-Ala/S-C3N4at 27.3° has a red shift relative to PDI-Ala and S-C3N4, which means that the composite sample has a smallerπ–πstacking distance. It results in the highly overlapping of electron clouds and the promotion of charge transfer between interfaces.

In order to further confirm the strong interfacial interaction between PDI-Ala and S-C3N4, the infrared spectrum of the asprepared samples was detected. Compared with perylene anhydride, PDI-Ala synthesized by self-assembly has a new wide peak at 3300 cm?1, and three new peaks appear at 1690 cm?1, 1660 cm?1and 744 cm?1, which are attributed to the stretching vibration of the associated hydroxyl group, the stretching vibration of C ＝O and C―H and the bending vibration of O＝C―N, respectively (Fig. 3a). The bond is consistent with the DFT calculation (as shown in Fig. 3c). These indicate that PDI-Ala with high stability was successfully synthesized from perylene anhydride through the preparation process of Fig. 1. For pure S-C3N4and PDI-Ala/S-C3N4composites, the wide absorption band at 3000–3500 cm?1is mainly attributed to the stretching vibration of N―H bond and O―H bond of bound water. Moreover, two characteristic absorption bands at 1240–1650 cm?1and 808 cm?1correspond to the stretching vibration of aromatic heterocycles (including perylene ring and N―C heterocycle) and respiratory vibration of 3-s-triazine ring. Fig. 3b is the enlarged infrared spectrum of Fig. 3a in the wavelength range of 1760–1200 cm?1. It can be seen that the corresponding characteristic peak intensities gradually increase with the increase of loaded PDI-Ala. In addition, compared with S-C3N4, the stretching vibration peaks(e.g. 1263 cm?1) of aromatic heterocycles in PDI-Ala/S-C3N4composites are clearly right shifted. The peak shift phenomenon of relative position in PDI-Ala/S-C3N4composite further proves the existence ofπ–πinteraction at the interface between PDI-Ala and S-C3N4.

Fig. 2 XRD patterns of as-prepared samples (a); partially enlarged view (2θ at 24°–30°) of S-C3N4, PDI-Ala, 30%PDI-Ala/S-C3N4 (b).

Fig. 3 Infrared spectra (a) and partially enlarged view (wavenumber at 1200–1760 cm?1) (b) of perylene anhydride, PDI-Ala, S-C3N4 and PDI-Ala/S-C3N4 composites with different contents of PDI-Ala; N―C bonded structure of PDI-Ala/S-C3N4 (c).

Fig. 4 TEM images of PDI-Ala/S-C3N4 (a, b); elemental mapping of C, N, O and S (c–f); micro constituency and spectrum of EDS (g, h).

The morphologies of 30%PDI-Ala/S-C3N4were observed by transmission electron microscopy (TEM), and the elemental mapping was analyzed by EDS. As shown in Fig. 4a,b, a porous two-dimensional nanosheet structure was formed, which is conducive to adsorb the pollutants23. In addition, the EDS spectrum (Fig. 4c–f) showed that there were a lot of C, N, O elements and trace S elements in the composite. The mapping in Fig. 4c–f demonstrated that all elements were uniformly distributed61. In particular, the element content of selected area#1 (Fig. 4g) was detected, and the corresponding spectrum (Fig.4h) was obtained. The results illustrated that trace S elements existed in PDI-Ala/S-C3N4composite, the content of C element was significantly higher than that of N element, which further confirmed the successful preparation of the composite. In addition, the presence of O element was detected by EDS and elemental mapping, which indicated that PDI-Ala was included in the PDI-Ala/S-C3N4composite.

3.2 Photocatalytic performance of PDI-Ala/S-C3N4 Sscheme heterojunction

The photocatalytic properties of as-prepared samples were evaluated by photodegradation of TC, PNP and H2O2production. The process of photocatalytic production of H2O2is as follows62:

As shown in Fig. 5a,c, the concentration of TC and PNP solution without photocatalyst has no obvious change under visible light irradiation, which indicates that TC and PNP have high photostability and are feasible as target pollutants. Fig. 5a shows that both self-assembled PDI-Ala and S-C3N4have excellent degradation activity, and the TC photodegradation activity of PDI-Ala/S-C3N4is significantly improved. The optimum 30%PDI-Ala/S-C3N4sample can photodegrade 90%TC within 90 min. On the contrary, the degradation rate of the powders with 30% mass ratio of PDI-Ala prepared by physical mixing for TC degradation is only 60% within 90 min, which indicates that thein situself-assembly strategy has been successfully applied. In addition, the kinetic equation diagram(Fig. 5b, wherekis the apparent rate constant of photocatalytic degradation of TC) was obtained by data linear fitting transformation. Compared withkvalue, the optimum 30%PDIAla/S-C3N4sample reached 0.02487 min?1, which was 2.64 times of PDI-Ala and 2.14 times of S-C3N4, respectively.Finally, the photodegradation rate of PDI-Ala/S-C3N4composite photocatalysts with different PDI-Ala relative content were decreased compared with the 30%PDI-Ala/S-C3N4sample. It is crucial that the best specific surface area and the most reactive sites for 30%PDI-Ala/S-C3N4are obtained.

Fig. 5 Photocatalytic degradation of TC (a) and apparent rate constant of linear fitting (b) by as-prepared samples; photocatalytic degradation of PNP (c), standard curve of the linear relationship between H2O2 concentration and absorbance in iodometry (d), generation of H2O2 (e) and summary of photocatalytic performance (f) for S-C3N4, PDI-Ala and 30%PDI-Ala/S-C3N4.

In order to verify whether the PDI-Ala/S-C3N4S-scheme heterojunction has high photocatalytic performance, the comparative experiments of photodegradation of PNP and photocatalytic production of H2O2were carried out for S-C3N4,PDI-Ala and 30%PDI-Ala/S-C3N4(Fig. 5c,e). As shown in Fig.5c, compared with PDI-Ala and S-C3N4, the photocatalytic activity of 30%PDI-Ala/S-C3N4heterojunction for degradation of PNP was significantly improved, which indicates that the heterojunction plays a very important role. Fig. 5d shows the standard curve of linear correlation between H2O2concentration determined by iodine titration and its absorbance. It can be seen from Fig. 5e that 30%PDI-Ala/S-C3N4exhibits excellent photocatalytic activity for H2O2production. Under visible light irradiation for 180 min, the H2O2concentration in the reactor reaches 28.3 μmol·h?1·g?1, which is 2.9 times as high as that of PDI-Ala and 1.6 times as high as that of S-C3N4, respectively.This may be due to the fact that porous S-C3N4and 30%PDIAla/S-C3N4are favorable for the capture of dissolved O2and the further H2O2production. The apparent photocatalytic rates of SC3N4, PDI-Ala and 30%PDI-Ala/S-C3N4are summarized in Fig.5f. It is clear from the histogram that the photocatalytic performance of 30%PDI-Ala/S-C3N4has been greatly improved in the different experiments. The construction strategy of Sscheme heterojunction successfully improves the separation efficiency and surface charge transfer capability of photogenerated carriers.

An excellent photocatalyst should have good photostability and high recycling efficiency. Therefore, the cycle stability test of 30%PDI-Ala/S-C3N4with the best photocatalytic performance was carried out. As shown in Fig. 6a, the photocatalytic degradation rate of TC by 30%PDI-Ala/S-C3N4slightly decreases after three cycles, which may be caused by the slight corrosion of S-C3N4and the sample loss during the process of the separation and recovery. However, the degradation rate of TC retains 83% by 30%PDI-Ala/S-C3N4after three cycles. XRD(Fig. 6b), UV-Vis DRS (Fig. 6c) and FT-IR (Fig. 6d)characterizations of 30%PDI-Ala/S-C3N4were detected before and after the cycle stability test. It is found that the cyclic test had little effect on the crystal structure, light response range and functional group structure of the samples. In other words,30%PDI-Ala/S-C3N4synthesized byin situassembly maintains stable photocatalytic performance and structural stability during the photocatalytic process.

Fig. 6 Cycle experiment of photodegradation of TC by 30%PDI-Ala/S-C3N4 (a); XRD (b); UV-Vis DRS (c) and FT-IR (d)of 30%PDI-Ala/S-C3N4 photocatalyst before and after the cycle stability experiment.

Fig. 7 Transient photocurrent responses (a) and electrochemical impedance spectroscopy (b) of S-C3N4, PDI-Ala,30%PDI-Ala/S-C3N4 under visible light irradiation.

Fig. 8 Mott-Schottky curves of S-C3N4 (a) and PDI-Ala (b); UV-Vis DRS of as-prepared samples (c); Tauc plots of S-C3N4 and PDI-Ala (d).

Under visible light irradiation, the related electrochemical experiments of S-C3N4, PDI-Ala and 30%PDI-Ala/S-C3N4were carried out. The effect of heterojunction on the photocatalytic activity of 30%PDI-Ala/S-C3N4photocatalyst was further discussed from the perspective of carrier separation. Under visible light irradiation, the photocatalyst with high separation and mobility will transfer more electrons and holes to the surface. The photocurrent detected by the workstation is shown in Fig. 7a. It can be found that the curves appear jitter after xenon lamp on, which can be attributed to the interference of background light and the rotation of the rotor63. The photocurrent of pristine S-C3N4and PDI-Ala under visible light irradiation is weak, while 30%PDI-Ala/S-C3N4shows enhanced photocurrent response. The construction of S-scheme heterojunction strategy makes a contribution to promote carriers’separation migration of 30%PDI-Ala/S-C3N4photocatalyst.Similar results can be obtained from Fig. 7b (electrochemical impedance spectroscopy, EIS). The arc curve in the high frequency region corresponds to the electron transfer process,and the linear part in the low frequency region corresponds to the mass transfer process64. In the EIS, the smaller the radius of the arc or the slope of the linear part, the smaller the impedance of the material. A material with smaller impedance has smaller charge transfer resistance and faster charge transfer rate inside or on the surface of the substance. 30%PDI-Ala/S-C3N4shows the minimum charge transfer resistance, which indicates that the internal electric field and electron delocalization effect between PDI-Ala and S-C3N4are conducive to the separation and transfer of photogenerated carriers. Finally, the excellent photoelectrochemical properties of 30%PDI-Ala/S-C3N4heterojunction are revealed.

3.3 Energy band structure of photocatalysts

Mott-Schottky curves (Fig. 8a,b) of S-C3N4and PDI-Ala were also obtained to determine their semiconductor type and flat band potential (Ef). At 0.5, 1.0 and 1.5 kHz, theEfvalues of SC3N4and PDI-Ala are ?2.23 V and ?0.32 V (vs.Ag/AgCl, pH =7), respectively. According to the general hydrogen electrodeE(NHE)=E(Ag/AgCl)+ 0.197, the calculatedEfof S-C3N4and PDIAla are ?2.03 V and ?0.12 V (vs.NHE, pH = 7), respectively. In addition, the straight line part of Mott-Schottky curves of S-C3N4and PDI-Ala has a positive slope, which means that both S-C3N4and PDI-Ala belong ton-type semiconductors. The bottom value of conduction band ofn-type semiconductor is approximately equal toEf. The lowest unoccupied molecular orbitals energy levelE(LUMO)of S-C3N4and PDI-Ala are ?2.03 eV and ?0.12 eV,respectively.

As shown in Fig. 8c, the spectral absorption range of the prepared sample was analyzed by UV-Vis DRS. With the increase of PDI-Ala content, red shift of the absorption edge of PDI-Ala/S-C3N4composites increases gradually, indicating that the heterojunction is successfully formed65. It has a strong response to almost full-range visible light, which greatly improves the solar energy utilization rate of the composite photocatalyst, and effectively generates more photogenerated carriers. According to UV-Vis DRS, Tauc diagrams of S-C3N4and PDI-Ala (Fig. 8d) can be obtained by transformation. The straight line part of Tauc curve is extended to the intersection point of X-axis (band gap,Eg), corresponding to the band gap values of S-C3N4and PDI-Ala are 2.55 eV and 1.56 eV,respectively. According to this formulaEg=E(LUMO)?E(HOMO),the highest molecular occupied orbitals energy levelE(HOMO)of S-C3N4and PDI-Ala can be calculated as 0.52 eV and 1.44 eV,respectively. Therefore, it is feasible to construct S-scheme heterojunction between S-C3N4and PDI-Ala. The cross band structure of S-scheme heterojunction is able to generate an internal electric field on the non-uniform interface of PDI-Ala/SC3N4, which provides an effective internal driving force for the charge separation.

3.4 Photocatalytic mechanism analysis

The work function (Φ) of the PDI-Ala, S-C3N4and PDIAla/S-C3N4monolayer are determined by using the following relation:

WhereEfermiis the Fermi energy of the monolayer surface andV∞is the electrostatic potential. The Fermi energy is determined self-consistently while performing the DFT based electronic structure calculations of the PDI-Ala, S-C3N4and PDI-Ala/SC3N4monolayer and the corresponding vacuum potentialV∞is calculated as the limiting value of electrostatic potential in the direction perpendicular to the surface at a distance of around 4.5 nm of vacuum66. The work function values for PDI-Ala, S-C3N4and PDI-Ala/S-C3N4monolayer are represented in Fig. 9b. As showed in Fig. 9a, the work function of PDI-Ala and S-C3N4is 5.77 eV and 3.33 eV, respectively. After the contact of PDI-Ala and S-C3N4, the electrons on the S-C3N4surface trend to transfer to PDI-Ala due to the relatively high work function of PDI-Ala.Finally, when the Fermi level is flattened to reach equilibrium,the work function between the two interfaces reaches 3.93 eV.Due to the electron transfer, the surface charge distribution of PDI-Ala and S-C3N4is different, which leads to the appearance of internal electric field and band bending. It is worth noting that the photogenerated electrons on the LUMO of PDI-Ala tend to recombine with the holes on the HOMO of S-C3N4under the visible light irradiation because of the effects of the internal electric field and band bending. In addition, it can be seen from the differential charge diagram (Fig. 9c) that the charge density of PDI-Ala in the composite is significantly lower than that of SC3N4. It also directly reflects the charge transfer and distribution between the PDI-Ala/S-C3N4heterojunction.

Fig. 9 The S-scheme charge transfer mechanism of PDI-Ala/S-C3N4 (a); the work functions of PDI-Ala, S-C3N4 and PDI-Ala/S-C3N4 (b);differential charge diagram (the yellow area indicates that the charge density increases and the blue part decreases) of PDI-Ala/S-C3N4 (c).

In situXPS technique can not only detect the relative element content of the material by semi-quantitative analysis, but also can infer the transfer direction of photogenerated electron in the heterojunction by changing the binding energy. The binding energy of 30%PDI-Ala/S-C3N4was studied by using a 300 W xenon lamp as visible light source and AlKαas a radiation source before and after illumination. As shown in Fig. 10a, the characteristic peaks of C, N, O elements can be found in the full spectrum XPS diagram. However, the characteristic peaks of S element are not detected, which is due to the small doping amount of the S element. Fig. 10b shows the characteristic peaks of C 1s. The five peaks corresponding to binding energies at 289.1 eV, 288.2 eV, 286.5 eV, 285.7 eV and 284.8 eV are attributed to O＝C―N, N―C＝N, C＝O, N―C, C―C bond in 30%PDI-Ala/S-C3N4, respectively67. In the XPS spectra of N 1s(Fig. 10c), the peaks at 401.3 eV, 400.2 eV and 398.8 eV are attributed to the N ofsp2hybrid of N―Hx, N―(C)3, and C―N＝C bond, respectively. For the XPS spectra of O 1sin the PDIAla/S-C3N4composites (Fig. 10d), a majority of them come from the contribution of O element in PDI-Ala, and a minority of them is the contribution of the adsorbed bound water. The peaks at 533.5 eV, 532.1 eV and 531.2 eV belong to O―H,O―C and O＝C bond, respectively67. As the quantity of S element is too small, only one rough S 2pspectra (Fig. 10e) can be obtained. There are two response peaks at 168.5 eV and 164.0 eV, indicating the successful implementation of S doping. In addition, the peaks of C 1s, N 1s, O 1sand S 2pin thein situXPS spectra are significantly shifted before and after illumination, which indicates the electronic density of the elements in the complex changes. In the O 1sdiagram, the binding energy of C―O bond existing in PDI-Ala increases from 532.1 eV to 532.3 eV under visible light irradiation,indicating that the electron density of PDI-Ala decreases.However, in the S 2pdiagram, the binding energy of S―C bond which only exists in S-C3N4decreases by 0.2 eV, indicating that the electron density of S-C3N4increases. In other words,photogenerated electrons in 30% PDI-Ala/S-C3N4are transferred from PDI-Ala to S-C3N4under visible light irradiation, which provides a direct evidence for the transmission direction of photogenerated electrons at the heterojunction interface.

Fig. 10 In situ XPS survey spectra of full-spectrum, C 1s, N 1s, O 1s, S 2p, respectively (a–e); free radical capture experiments of 30%PDI-Ala/S-C3N4 (f).

As shown in Fig. 10f, a free radical capture experiment with 30%PDI-Ala/S-C3N4as photocatalyst and TC as a target pollutant was performed to confirm the main active species in the photocatalytic reaction process. AO, IPA and p-BQ were used as collectors for h+, ·OH and ·O2?, respectively. After visible light irradiation for 90 min, the removal rate of TC reached 90% in blank group, while the removal rate of TC in the experimental groups (p-BQ or AO was added before reaction)was only 43% and 52%, respectively. The photocatalytic performance of the experimental groups was obviously inhibited. These results indicate that the active species of ·O2?and h+play an important role in the photocatalytic process. The experimental group that IPA was added before reaction had almost no effect on the removal of TC, which indicates that ·OH is not the main active component of the photocatalytic degradation reaction. This may be the reason thatE(OH?/·OH)(the redox potential of ·OH) is greater than the HOMO of 30%PDI-Ala/S-C3N4photocatalyst so that ·OH can not be generated by h+. The above results confirm that the high effictive photocatalytic activity of 30%PDI-Ala/S-C3N4is mainly contributed to the h+and ·O2?free radical.

Fig. 11 Photocatalytic mechanism of PDI-Ala/S-C3N4 S-scheme heterojunction.

To sum up, a possible mechanism of photocatalytic reaction for 30%PDI-Ala/S-C3N4S-scheme heterojunction is proposed(Fig. 11). The N―C bond andπ–πconjugate interaction are formed at the interface between PDI-Ala and S-C3N4. The DFT andin situXPS results both show that the photogenerated electrons in the composites flow from PDI-Ala to S-C3N4.Therefore, it is inferred that the direction of internal electric field is from S-C3N4to PDI-Ala. As shown in Fig. 11, the band bending phenomenon is produced, which is due to the difference of Fermi level between PDI-Ala and S-C3N4. Theπ-πinteraction between self-assembled PDI-Ala and S-C3N4leads to the electron delocalization, which is conducive to the migration of electrons along theπ–πstacking direction. Due to the effects of band bending and internal electric field, the photogenerated electrons on the LUMO of PDI-Ala trend to recombine with the holes on the HOMO of S-C3N4at theπ-πconjugate interface,which effectively retained the strong oxidizing holes of PDI-Ala and the strong reducing electrons of S-C3N4. Subsequently,dissolved O2in contact with S-C3N4surface is reduced to the strong oxidizing ·O2?. The strong oxidizing h+and ·O2?on the surface reactive sites of the composites rapidly degrade TC or PNP.

4 Conclusions

In conclusion, all organic PDI-Ala/S-C3N4S-scheme heterojunction photocatalyst was synthesized by anin situselfassembly method, and its photocatalytic performance and mechanism were explored. In the experiment of H2O2production and the degradation of TC and PNP, the photocatalytic activity of the S-scheme heterojunction was significantly improved. The main reasons are as follows: firstly, aπ–πconjugation effect has arisen, which makes them closely combined so that their density of electron cloud increases even overlaps and the charges between them are transferred quickly. Secondly, the cross band structure of S-scheme heterojunction can generate an internal electric field on the inhomogeneous interface of PDI-Ala/SC3N4, which provides an internal driving force for charge separation. Thirdly, the band gap structure of S-C3N4is adjusted by doping S element, which is beneficial to the construction of the S-scheme heterojunction. Fourthly, PDI-Ala has a strong response in almost full-range visible light, which greatly improves the utilization of solar energy. This study provides a new idea for constructing and designing effective S-scheme heterojunction photocatalysts with all organic semiconductors.