Synthesis of Cu3P/SnO2 composites for degradation of tetracycline hydrochloride in wastewater

Antibiotic drugs have become dominating organic pollutants in water resources, and efficient removal of antibiotic drugs is the priority task to protect the water environment. Cu3P/SnO2 photocatalysts of various Cu3P loadings (10–40 wt% Cu3P) were synthesized using a combination of hydrothermal synthesis and a partial annealing method. Their photocatalytic activity was tested for tetracycline hydrochloride (TC-HCl) degradation under visible light irradiation. Cu3P/SnO2 samples were characterized by X-ray diffraction (XRD), N2-adsorption, ultraviolet-visible diffuse reflectance spectra (UV-vis DRS), scanning electron microscopy (SEM) and electrochemical impedance spectroscopy (EIS). The results showed that the p–n type heterostructure between Cu3P and SnO2 was successfully constructed, and addition of Cu3P to SnO2 could improve its photocatalytic activity at an optimized loading of 30 wt% Cu3P. In photocatalytic degradation studies, removal rates of around 80% were found in 30 minutes of dark reaction and 140 min of photodegradation. The removal rate was superior to that of Cu3P and SnO2 alone under the same experimental conditions. According to trapping experiments and electron spin resonance (ESR) measurements, photogenerated holes (h+) and superoxide radicals ˙O2− were considered as the main oxidation species in the present system. Finally, the reuse experiments showed high stability of Cu3P/SnO2. This study reports Cu3P as a cocatalyst combined with semiconductor SnO2 to form a highly efficient heterogeneous photocatalyst for the degradation of tetracycline hydrochloride for the first time.


Introduction
Recently, antibiotics have been widely misused in medical treatment as antimicrobial drugs. Due to their coupling with the food chain, they bioaccumulate at lower and higher nutrient levels, resulting in a long-term negative impact on the environment. 1,2 Tetracycline hydrochloride (TC-HCl) is a broadspectrum antibiotic of tetracycline, which is widely used medically to prevent bacterial diseases of livestock, poultry and aquatic products. 3 Its worldwide extensive misuse accelerates it accumulation in the environment. 4 Because of its obvious mutagenic and teratogenic effects, it has drawn extensive attention. 5,6 Therefore, it is particularly important to remove TC-HCl in wastewater. Compared with advanced oxidation methods, biological treatment methods, membrane ltration methods and adsorption methods, 7-12 the photocatalytic degradation method is commonly used in the treatment of refractory organic wastewater due to its advantages of high efficiency, low energy consumption, and green environmental protection. 13 The photocatalyst generates free radicals under light conditions and effectively decomposes organic chemicals in wastewater. [14][15][16][17] Many semiconductor photocatalysts have been developed for decades. For example, Shen et al. prepared hierarchical carbon nitride photocatalyst, 18 Song et al. constructed KCl/NH 4 Cl/g-C 3 N 4 composite photocatalyst, 19 Mao et al. prepared Bi 2 WO 6 / CuS composite photocatalyst, 20 these catalysts have been used to degrade TC-HCl, and have achieved good degradation effects.
From the reports of photocatalytic materials in last decades, SnO 2 as a photocatalyst has drawn extensive attention due to its excellent optical properties, 21 electrical properties, 22 and stable gas-sensitive properties. 23 As an n-type semiconductor, SnO 2 has a quite wide band gap of about 3.6 eV, 24 which can only be excited by ultraviolet light with a wavelength of less than 400 nm, and its absorption efficiency is low under sunlight. 25 The recombination rate of photogenerated electron-hole is high, which restricts its direct application in photocatalysis. Most researches of SnO 2 as a photocatalytic material involved coupling SnO 2 with other narrow-band semiconductors, or adding dopants to expand the light absorption to the visible light range. 26,27 Therefore, there were multiple researches reported the photocatalytic properties of SnO 2 doped metallic elements such as cobalt-doped SnO 2 nanoparticles, Au/SnO 2 composites and Ag/SnO 2 composites 28-30 or non-metallic elements such as carbon-doped SnO 2 nanostructures and Cldoped SnO 2 photocatalysts. 31,32 Some publications reported SnO 2 coupled with other narrow-band semiconductors such as AgBr/SnO 2 composites and Co 3 O 4 /SnO 2 composites, 24,33,34 and SnO 2 coupled with wideband semiconductors such as TiO 2 (ref. 35) or ZnO. 36,37 Recently, studies have found that materials such as transition metal phosphides (MP) and their composites have excellent catalytic performance for hydrogen evolution reactions. For example, Gai et al. used Ni 2 P 38 and Cao et al. used CoP 39 to improve the hydrogen evolution efficiency of CdS. In addition, Sun et al. prepared a Cu 3 P/g-C 3 N 4 heterostructure 40 for photocatalytic hydrogen production. As a p-type semiconductor, Cu 3 P has a band gap of about 1.5 eV. 41 If the p-type semiconductor is combined with the n-type semiconductor, an internal electric eld is generated in the p-n junction area, which can efficiently facilitate the separation of photogenerated electrons and holes. 42 Ioannidi et al. synthesized Cu 3 P/BiVO 4 composite materials 43 to degrade sulfamethoxazole in aqueous media as similar study. It is worth noting that, Cu 3 P-based composite materials have not been used as a photocatalyst to degrade TC-HCl yet based on literature study. This study lled the gap into the area.
The purpose of this work is to synthesize Cu 3 P/SnO 2 heterostructure with different Cu 3 P loading in order to explore its catalysis and characteristics as a heterogeneous photocatalytic composite material in area of TC-HCl decomposition. We evaluated the photocatalytic activity of Cu 3 P/SnO 2 sample for degradation of TC-HCl under visible light irradiation. Finally, the photocatalytic activity results of Cu 3 P/SnO 2 composite material were compared with that of the self-made parent material of Cu 3 P or SnO 2 under the same experimental conditions. The results may provide an alternative method for preparation of photocatalysts, under the case of wide band gap SnO 2 semiconductors used.

Experiment section
2.1. Photocatalyst preparation 2.1.1. Synthesis of SnO 2 hollow microspheres. Hollow SnO 2 microspheres were prepared by hydrothermal method. 44 SnCl 2 $2H 2 O of 0.4890 g and a certain amount of Na 3 C 6 H 5 O 7 -$2H 2 O were dissolved in distilled water of 12.5 mL at different molar ratios of 1 : x (x from 3 to 6) to generate homogeneous solution A. NaOH of 0.30 g was dissolved in 12.5 mL distilled water to generate homogeneous solution B. Solution B was slowly added to solution A, with the mixture continuously stirred for 24 hours. The transparent solution stirred was transferred to a 50 mL reaction kettle and placed in an oven at 180 C for 12 hours. Aer cooling of the reaction kettle, the white precipitate at the bottom was alternately washed with ethanol and distilled water repeatedly. The collected samples were dried overnight in an oven at 60 C to produce white SnO 2 sample. The SnO 2 samples prepared with different ratios of raw materials were labeled as SnO 2 -1 : 3, SnO 2 -1 : 4, SnO 2 -1 : 5 and SnO 2 -1 : 6 respectively.
2.1.2. Synthesis of Cu 3 P. The 0.189 g (0.003 mol) copper powder and 0.3097 g (0.01 mol) red phosphorus were thoroughly ground evenly with an agate mortar. The mixed powder was collected in a beaker, and added with 80 mL deionized water. Aer the solution was continuously stirred for 0.5 h, the solution was transferred to a 100 mL reaction kettle and kept in an oven at 200 C for 24 h. Aer cooling of the reaction kettle, the precipitate were centrifugally washed with carbon disulde (CS 2 ), ethanol and deionized water one by one, and the resulting products were dried overnight in the oven at 60 C.
2.1.3. Synthesis of Cu 3 P/SnO 2 heterostructure. 0.03 g Cu 3 P and 0.30 g SnO 2 were sufficiently ground, and the mixture was transferred to 30 mL deionized water containing 0.25 g polyvinyl pyrrolidone (PVP). Aer stirring with ultrasonic for 24 hours, the precipitate was rinsed for many times with deionized water. Then it was dried, ground, and put into an atmosphere tube furnace at 300 C rst, and then annealed for 3 hours by introducing N 2 to produce Cu 3 P/SnO 2 sample with content of 10%. In addition, by varying the loading of Cu 3 P, the Cu 3 P/SnO 2 heterostructure composite catalysts of 20%, 30% and 40% wt load were synthesized by the same preparation method, labeled as Cu 3 P/SnO 2 -1, Cu 3 P/SnO 2 -2, Cu 3 P/SnO 2 -3 and Cu 3 P/SnO 2 -4 respectively.

Characterization methods
The characterization of samples was carried out to understand their structure, surface, morphology, optical and photocatalytic properties. X-ray diffraction XRD was applied to analyze the phase composition of the prepared samples, with D/MAX-2500 diffractometer (Bruker test in Germany). The morphological characteristics were studied by scanning electron microscope (SEM), from Hitachi (S-4800). Ultraviolet-visible diffuse reectance spectrometer of Japan Shimadzu Company's UV 2600 model was used to analyze the absorption properties of the prepared samples with BaSO 4 as a reference in the wavelength range of 250 nm to 800 nm. The N 2 adsorption and desorption isotherm test uses the Quantachrome company (NOVA 2200e) pore size analyzer. The specic surface area and pore size characteristics of the sample are calculated using Brunauer-Emmett-Teller (BET) and Barret-Joyner-Halenda (BJH) theoretical models, respectively.

Electrochemical test
The electrochemical impedance spectroscopy (EIS) test was performed on the electrochemical workstation of Shanghai Chenghua Instrument Co., Ltd. (CHI660E), China. Calomel electrode and platinum electrode were used for reference electrode and auxiliary electrode respectively. 0.5 M H 2 SO 4 aqueous solution is used as the electrolyte, and copper electrode sample is used as the working electrode.
The electrode sample preparation process is as follows: 0.16 g polyvinylidene uoride (PVDF) binder were mixed with 0.02 g catalyst sample and 0.02 g carbon black into a small glass jar. Then we add 500 mL N-methyl pyrrolidone to the mixture, and stir. Aer a few hours, the paste liquid is consistently coated on a copper sheet of 10 Â 20 mm in size, and the other side of the copper sheet is pasted with insulating glue. The specic process of the electrochemical workstation was as follows: the three-electrode system and the electrochemical workstation were connected, then the open circuit potential of the electrode was tested, and later the initial voltage is set

Photocatalytic activity experiments
The photocatalytic activity of the samples was systematically evaluated by visible light-driven degradation of TC-HCl. The photocatalytic experiments were performed under a 1000 W Xenon lamp at constant temperature 25 C. The photocatalytic degradation activity of Cu 3 P/SnO 2 -1-4 were evaluated. First, 50 mg L À1 TC-HCl solution was prepared, with ultrasonic treatment for 20 minutes to ensure uniform dispersion of TC-HCl molecules. Photocatalyst Cu 3 P/SnO 2 (50 mg) was uniformly dispersed in 40 mL TC-HCl solution. Then, stirring was continued for 30 minutes in the dark to make the TC and Cu 3 P/SnO 2 sample reach the equilibrium of adsorption and desorption. Aer turning on the light source, 3 mL sample was taken every 20 minutes, and diluted with 3 mL distilled water. The diluted sample is centrifuged twice at 10 000 rpm to remove the sediment. The concentration of supernatant was determined by an UV-visible spectrophotometer at 360 nm, which is the typical absorption wavelength of TC-HCl. The photocatalytic process took for 140 minutes for several samples. 1-X microsphere exhibit a hollow structure with pore sizes ranging from 200 nm to 500 nm. The morphology of SnO 2 -1 : 5 samples (Fig. 1c) showed a hollow structure to some extent with different sizes. For the other two samples of SnO 2 -1 : 3 (Fig. 1a) and SnO 2 -1 : 6 ( Fig. 1d), the hollow structure was obviously broken. The arrangement of SnO 2 was irregular with small porosity, and the specic surface area and additional active sites increased and turned conducive to photocatalytic degradation.  As shown in Fig. S1 in ESI, † Cu 3 P is very small in size, similar in shape to cotton mass with a diameter of about 250 nm. Fig. 2a demonstrated the morphology of Cu 3 P/SnO 2 composites, SnO 2 was a sphere with much larger size than that of Cu 3 P, with a diameter of about 1 mm. Cu 3 P nanoparticles were clustered on the surface of SnO 2 microspheres, with a solid-solid interface formed between the two chemicals. These results demonstrated that Cu 3 P had been successfully installed on the surface of SnO 2 .

Result and discussion
XRD patterns in Fig. 2b showed crystalline structures of the synthesized parent materials of Cu 3 P and SnO 2 , and the Cu 3 P/ XRD pattern of Cu 3 P in Fig. 2 was consistent with the standard card (JPDF no. 71-2261), and the diffraction peaks were at 36.005 , 39.078 , 41.577 , 45.091 and 46.158 , which was corresponded to (112), (202), (211), (300) and (113) lattice planes of Cu 3 P, respectively. Besides, the purity of Cu 3 P was very high since there is no extra impurity peak. It is worth noting that both diffraction peak of Cu 3 P and diffraction peak of SnO 2 were plotted in the XRD diffraction pattern (Fig. 2) of the Cu 3 P/SnO 2 heterostructure, indicating the successful combination of Cu 3 P and SnO 2 .
Compared with the XRD peaks on the (110), (101) and (211) lattice planes (Fig. 2) of SnO 2 , the peak width of the Cu 3 P/SnO 2 heterostructure became smaller, reecting bigger average crystallite size. From Table 1 The pore size and BET specic surface area were determined by N 2 adsorption of the synthesized samples (Fig. 3). Fig. 3a showed the N 2 adsorption and desorption isotherms of SnO 2 and Cu 3 P/SnO 2 -1, 2, 3, and 4. According to the classication standard, the isotherm of the composite Cu 3 P/SnO 2 at a relative pressure (p/p 0 ) of 0.5-1.0 belonged to the IV type isotherm, and a typical H 3 hysteresis loop. This veried that composite Cu 3 P/ SnO 2 heterostructure was a mesoporous material. Compared with pure SnO 2 , the BET specic surface area of the heterostructure Cu 3 P/SnO 2 -3 increased signicantly. According to the BET theoretical calculation model, Table 2 was categorized with the specic surface area of pure SnO 2 was 3.7892 m 2 g À1 , while the specic surface area of Cu 3 P/SnO 2 -3 was 53.1470 m 2 g À1 , which was about 14 times that of SnO 2 . Fig. 3b showed the pore size distribution curves of Cu 3 P/SnO 2 samples. The pore size of Cu 3 P/SnO 2 heterostructure was distributed around 5-20 nm. Based on BJH theoretical calculation model, the pore volume of SnO 2 was 0.003904 cm 3 g À1 , and the pore volume of Cu 3 P/SnO 2 -3 heterostructure was 0.046550 cm 3 g À1 , with a tremendous increase of 12 times (Table 2). From Table 2, the specic surface  area of Cu 3 P/SnO 2 -3 was the largest (53.1470 m 2 g À1 ) while the average pore size was the smallest (3.8213 nm). The increase in the surface area were attributed to the surface modication of Cu 3 P nanoclusters on SnO 2 . The larger the specic surface area of the sample, the easier of the sample to expose active sites on the surface. As a photocatalyst, it increased the photocatalytic activity of the pollutant tetracycline (TC-HCl).
In order to study the modication effect of mesoporous Cu 3 P nanoclusters on SnO 2 microspheres to improve its photocatalytic performance, UV-vis DRS was applied to analyze absorption properties of SnO 2 , Cu 3 P and Cu 3 P/SnO 2 -1-4 heterostructure. Fig. 4a showed the UV-vis absorption spectrum of SnO 2 , Cu 3 P and heterostructure Cu 3 P/SnO 2 -1, 2, 3, and 4 samples. Cu 3 P has strong absorption strength in the whole wavelength range, especially in the visible light range. The SnO 2 has a strong absorption intensity in range of 200-360 nm (UV), while there was negligible absorption intensity above 400 nm (visible). This indicated that SnO 2 can only be excited by ultraviolet photons to undergo electronic transitions, without any optical response in the visible wavelength range. Compared with bare SnO 2 , the absorption intensity of Cu 3 P/SnO 2 heterostructure was quite different, with a higher absorption intensity in the visible range of 400-800 nm, which was related to the modication effect of mesoporous Cu 3 P nanoclusters on the surface of SnO 2 nanosheets. This indicated a strong interaction between Cu 3 P and SnO 2 , which was benecial to improving the migration efficiency of electron-hole pairs and inhibited their rapid recombination. Furthermore, the band gap values were estimated by Tauc plots and plotted in Fig. 4b. From the Fig. 4b the E g for Cu 3 P and SnO 2 samples are about 1.31 eV and 3.54 eV, respectively. Fig. 5 showed the EIS measurements of pure SnO 2 and Cu 3 P/ SnO 2 samples. The arc size of the Nyquist plot of SnO 2 were large due to its large energy gap. In the EIS Nyquist diagram, the   smaller the arc size, the smaller the charge transfer resistance and the higher the efficiency of electron-hole separation. Based on previous publications of SnO 2 materials, 44 different studies reported different results, from 20 to 1250 ohms. 45,46 From Fig. 5 the arc diameter of SnO 2 was about 120, reecting the charge transfer resistance Rct of 120 ohms. The Rct of Cu 3 P/SnO 2 -3 sample was the smallest (40 ohms), followed by the order of Cu 3 P/SnO 2 -1 (80 ohm) > Cu 3 P/SnO 2 -2 (60 ohm) and Cu 3 P/SnO 2 -4 (50 ohm), respectively. The results indicated that the optimal Cu 3 P loading is 30% wt. The tight solid-solid interface generated between the two semiconductors of Cu 3 P and SnO 2 turned into a channel for charge transfer, which reduced the charge transfer resistance in interface and improved the migration rate of electric load current carrier.

Adsorption and photocatalytic activity evaluation
From the photocatalytic experiment, the photocatalytic activity was explored of different samples under visible light radiation. Fig. 6a showed the dynamic curves of SnO 2 , Cu 3 P and composite material heterostructure Cu 3 P/SnO 2 -1, 2, 3 and 4 in the photocatalytic degradation of tetracycline. The samples underwent dark reaction adsorption for 30 minutes, and then was irradiated under visible light for 140 minutes. The results showed that aer adsorption equilibrium was achieved in the dark, the maximum adsorption capacity was 10% of the heterostructure Cu 3 P/SnO 2 -3, and the adsorption capacity of SnO 2 sample was 5%, which indicated that the increase of physical adsorption capacity. Such difference was due to the morphological structure of Cu 3 P/SnO 2 composite system increased their specic surface area.
In the 140 minutes visible light irradiation stage, the degradation effect on tetracycline was obvious. The degradation effects were different of heterostructure materials Cu 3 P/SnO 2 -1, 2, 3 and 4 with different composite ratios. The degradation  efficiency SnO 2 samples was only 35%, since SnO 2 cannot fully absorb visible light, and ultraviolet light accounts for a small proportion of visible light. When Cu 3 P was introduced into the SnO 2 material and Cu 3 P/SnO 2 was generated, the photocatalytic degradation effect was signicantly enhanced. The photocatalytic degradation effect of Cu 3 P/SnO 2 -3 became the best among rest samples and the degradation efficiency reaches 80%. From Fig. 6a, the order of degradation efficiency was: Cu 3 P/SnO 2 -3 (80%) > Cu 3 P/SnO 2 -4 (70%) > Cu 3 P/SnO 2 -2 (64%) > Cu 3 P/SnO 2 -1 (53%) > Cu 3 P(45%) > SnO 2 (34%), the higher the better. Fig. 6b showed the kinetic curves of Fig. 6a, which tted to the rst-order kinetic model ln(C 0 /C) ¼ kt. The photo degradation process was rst order reaction. Fig. 6c was a histogram of the rate constant K of degradation of tetracycline by SnO 2 , Cu 3 P and composite heterostructure Cu 3 P/SnO 2 -1, 2, 3 and 4. The rate constant k of pure SnO 2 and Cu 3 P were 0.00228 and 0.00353 min À1 , while the rate constant k of the sample Cu 3 P/ SnO 2 -3 was the maximum (k ¼ 0.00940 min À1 ), which was 4.1 times and 2.7 times than that of SnO 2 (k ¼ 0.00228 min À1 ) and Cu 3 P (k ¼ 0.00353 min À1 ), indicating the Cu 3 P/SnO 2 effectively enhanced the photocatalytic degradation efficiency. The rate constant k of the other samples Cu 3 P/SnO 2 -1, 2 and 4 were 0.00429, 0.00570 and 0.00697 min À1 , respectively. These results veried that the optimal composite percent was 30% wt of Cu 3 P/ SnO 2 composite, with the best photocatalytic performance for tetracycline. In addition, Fig. 6d showed the absorption spectra of Cu 3 P/SnO 2 -3 for tetracycline at different times. With the gradual extension time of visible light radiation, the characteristic absorption wavelength of 360 nm of tetracycline absorption showed a signicant downward trend. This indirectly conrmed that with increased period of visible light radiation, tetracycline gradually was decomposed and opens the ring to breakdown into small molecules or ions. Table 3 categorized several photocatalysts that had been reported in literature. The results showed that Cu 3 P/SnO 2 -3 contained 80% degradation rates, which was comparable with different reported catalysts, such as FeOOH/FeS 2 of 90%, 47 51 Besides, the dosage (W cat /TC) of Cu 3 P/SnO 2 -3 was 25 mg mg À1 , which was the minimum around the rest 5 types of catalysts around 50-100 mg mg À1 . The amount of catalysts Cu 3 P/SnO 2 -3 used in photocatalytic experiment was much less with less pollution and lower cost. This indicates that the catalyst is more favorable to treat the wastewater containing relatively high tetracycline.

Stability and degradation mechanism studies 6tb
To evaluate the stability and reusability of the Cu 3 P/SnO 2 heterostructure as a photocatalyst, we conducted ve cycles of photodegradation experiments on the Cu 3 P/SnO 2 -3 catalyst. Aer completing a cycle of photodegradation experiment, the catalyst was collected by simple centrifugation, and treated with rinse, drying and other operations. The collected catalyst was prepared for the next cycle experiment. Fig. 7a and b showed the degradation curves and removal efficiency of tetracycline aer ve cycles. Aer multiple photodegradation cycles, the degradation efficiency of tetracycline changed a little, from 80% to 75%. Aer 5 consecutive cycles, it only dropped by 5%. These results proved stability of Cu 3 P/SnO 2 -3 as a photocatalyst, and it can be recycled with treatment, which effectively reduced the cost of catalyst production.
The reaction mechanism of tetracycline breakdown was studied by adding active oxidant species capture agent in the photocatalytic reaction stage. Isopropyl alcohol (IPA), benzoquinone (BQ) and ethylene diamine tetra acetic acid disodium salt (EDTA-2Na) were used to capture hydroxyl radical (cOH), photogenerated holes (h + ) and superoxide radicals (cO 2 À ) of active oxidizing species, respectively. From the kinetic degradation curve (Fig. 7c) and the histogram (Fig. 7d) of degradation efficiency, the degradation of tetracycline by Cu 3 P/SnO 2 -3 catalysts was greatly different with different capture reagents. Results showed that the addition of active species capture agent did take into effect. Aer adding EDTA-2Na and BQ as capturing reagents for h + and cO 2 À to the photocatalytic reaction pathway, the degradation efficiency of TC-HCl is dropped down to 19% and 24%, much lower than the blank test of non-capturing reagents (80%). This strong inhibitory effect indicated that h + and cO 2 À played a key role in the degradation of tetracycline, and the effect of h + was slightly higher than that of cO 2 À . However, if IPA was used as a capture agent for cOH, the degradation efficiency of tetracycline drops from 80% to 57%, indicating that cOH plays a minor role in the degradation of TC-HCl. In summary, in the process of tetracycline degradation by Cu 3 P/SnO 2 composite, the role of h + , cO 2 À and cOH follows h + > cO 2 À > cOH of the degradation reaction.
The ESR measurements were further used to identify cO 2 À in Cu 3 P/SnO 2 -3 composite catalysts. As shown in Fig. S2 in ESI, † no signal was detected under dark conditions. Under visible light irradiation, cO 2 À appeared obviously, which proved that the production of the active substance was consistent with the free radical trapping experiment. Therefore, Cu 3 P/SnO 2 photocatalyst showed good performance with the help of reactive oxygen species. Fig. 8 showed the possible mechanism of the degradation of TC by heterostructure Cu 3 P/SnO 2 catalyst under visible light. In the previous calculation, the band gap values of Cu 3 P and SnO 2 were 1.31 eV and 3.54 eV, respectively. Under the visible light radiation, SnO 2 and Cu 3 P are respectively excited, and the electrons in the valence band (VB) migrate to the conduction band (CB), leaving the hole h + in the VB. Part of the electrons in the CB of Cu 3 P were transferred to the CB of SnO 2 through the heterostructure interface. H 2 O molecules trap holes h + in the VB and convert them into cOH, and part of the oxygen O 2 molecules adsorbed on the catalyst surface trap electrons e À in the CB and converted them into superoxide radical cO 2 À . Finally, the holes h + , cO 2 À and cOH directly oxidize the tetracycline TC molecules.
This was the general degradation principles of tetracycline TC by the studied Cu 3 P/SnO 2 heterostructure. The specic and Fig. 7 The cyclic degradation curve of tetracycline over heterogeneous Cu 3 P/SnO 2 -3 composite catalyst (a); column comparison of degradation rates (b); degradation dynamic curve after the addition of capturing agent (c); histogram of degradation efficiency of different capture agents (d). detailed degradation mechanism of tetracycline molecules await future research.

Conclusion
Cu 3 P/SnO 2 , the p-n type heterostructure, was proved to be an efficient photocatalyst for the degradation of antibiotic drug pollutants like TC-HCl. Among different Cu 3 P loaded samples, Cu 3 P/SnO 2 -3 (30% load) aer 30 minutes of dark reaction adsorption and 140 minutes of photodegradation, the removal efficiency reached 80%, which is 2.3 times of the removal efficiency of SnO 2 . The improved photocatalytic activity was beneted from the heterogeneous interface between SnO 2 and Cu 3 P, which effectively enhances the efficient charge transfer and retarders the recombination of electron hole pairs. Furthermore, the cyclic photodegradation experiments veried that the heterojunction Cu 3 P/SnO 2 -3 was stable and the efficiency dropped 5% aer 5 runs. From mechanistic study, the active oxidant species as h + , cO 2 À , and cOH were all involved in photocatalytic system. It seems that h + and cO 2 À played the major role in the photocatalytic degradation of tetracycline by Cu 3 P/SnO 2 composites. Further work may involve the improvement of catalyst with better degradation activity.

Conflicts of interest
There are no conicts to declare.