The colored organic dyes released from the textile and dye industries into water bodies are hardly biodegradable and can severely damage aquatic ecosystems and aquatic organisms. A recent report has shown that the global annual production of synthetic dyes and pigments has reached over 10⁶ tons. In addition, 5%–15% of the synthetic dyes will usually be lost during the process of manufacturing and processing, and these waste dyes may be toxic or carcinogenic to aquatic organisms or systems in water [1]. Commonly, synthetic organic dyes used in textile industries (e.g., MB, C₁₆H₁₈N₃SCl, Reactive Orange-16 Dye and C₂₀H₁₇N₃Na₂O₁₁S₃) are considered as allergens, and some are mutagenic and carcinogenic. Moreover, the discharge of industrial wastewater containing synthetic dyes can cause serious contamination at the dumping site [2, 3].

So far, a variety of treatment methods, such as chemical precipitation/contaminant separation, coagulation, ozone oxidation, hypochlorite oxidation, electrochemical methods and adsorption, have been used to degrade organic dyes [4-9].

Recently, the use of antibiotics in medicine has considerably improved the treatment efficiency for infectious diseases and has also been widely used in agricultural production. However, with the increasing and extensive use of antibiotics, wastewater containing antibiotics are released into water bodies in large quantities, which has resulted in a serious environmental threat to aquatic and land ecological systems [10]. As one of the most typical antibiotics, tetracycline is widely used in medicine, agriculture and aquaculture. Unfortunately, tetracycline cannot be eliminated by natural environment or biological wastewater treatment technologies [11].

In previous studies, various strategies have been adopted to solve the problem of antibiotic wastewater. Biological removal processes, such as biodegradation by bacteria and fungi, and non-biological processes, including adsorption, hydrolysis, oxidation and reduction, have aroused widespread interest from the research community [12-17]. However, these methods have some inherent defects, including high cost, poor stability and low efficiency, which has limited the application of these technologies. Hence, the development of new technologies or materials to eliminate antibiotic pollution remains as a challenge in the field of environmental chemistry.

The photocatalytic degradation of organic pollutants under visible light or ultraviolet light is an important method, as it can quickly and thoroughly eliminate pollutants without leaving any harmful residue [18]. Semiconductors, which are extensively applied to electronic devices and integrated circuits, have recently been shown to be important photocatalysts in wastewater treatment technology and have been widely studied in the past few decades [19].

Wide band gap semiconductors, such as TiO₂ and ZnO, have been shown to be the most promising photocatalysts of all the semiconductor materials. It has been demonstrated that ZnO is a better photocatalyst than TiO₂, owing to its wide band gap (3.2 eV) and strong excitation binding energy (~ 60 meV). These characteristics have resulted in a lot of interest from many scholars for their use in electronics, light emitting diodes, gas sensing and photocatalysis [20].

However, ZnO has some inherent flaws, including a rapid recombination of photo-generated electrons and holes, low quantum efficiency, susceptibility to light corrosion, poor light stability, and so forth. These flaws greatly limit the photocatalytic potential and reduce the photocatalytic activity, which therefore limit the application of ZnO in environmental remediation.

Moreover, owing to its wide band gap, ZnO can only be excited by UV light, which only accounts for approximately 5% of the solar spectrum. This results in an ineffective absorption of visible light and low usage rate for sunlight [21]. Therefore, it is a challenging task to obtain a ZnO photocatalyst with both favorable light stability and high photocatalytic activity under visible light by doping or modifying methods, but this has now become a frontier subject in international research [22].

The modification of ZnO with other semiconductors, such as CdS [23], In₂S₃ [24], NiO [25], CdSe [26], PbS [27], Ag₂S [28], Mn₃O₄ [29], Bi₂O₃ [30], Cr₂O₃ [31], Bi₂S₃ [32] and Bi₂WO₆, is an effective method to generate a large number of electron–hole pairs and enhance the photocatalytic activity of ZnO. The research on Bi-based photocatalysts has long been a hot topic in the field of photocatalysis and many important candidates have been reported, including BiOCl [33], Bi₁₂GeO₂₀ [34], Bi₂SiO₅ [35] and BiOBr [36]. Among these semiconductors, as inorganic semiconductor nanocrystals, Bi₂WO₆ has received widespread consideration for its nontoxicity, suitable band gap and excellent photocatalytic performance [37-39].

As an efficient material to improve the catalytic properties of photocatalysts, heterojunction photocatalysts can be used to facilitate the separation of photo-excited electron–hole pairs. In general, when p-type and n-type semiconductors are in contact, they form a p-n junction with a space charge region between the interface of the two semiconductors owing to the diffusion of electrons and holes. Therefore, the construction of a semiconductor heterojunction has attracted a lot of attention owing to its effectiveness in improving the photocatalytic activity. Some new findings have been also reported on the semiconductor heterojunction photocatalysts, such as the p-n junctions of BiOI/Bi₁₂O₁₇Cl₂ [40], BiVO₄/BiOI [41] and BiOBr/BiOI [42]. These results provide an important theoretical foundation for this work.

In the present work, a heterojunction catalyst of Bi₂WO₆/ZnO was synthesized by a two-step hydrothermal method. The basic performance and photocatalytic properties were investigated through the degradation of MB and tetracycline. The aim was to improve the photocatalytic efficiency and enhance the conversion for solar energy through reducing the band gap energy of ZnO, increasing the effective separation of electron-hole pairs and promoting the transfer of charge carriers. Moreover, the degradation mechanism of a flower-like Bi₂WO₆/ZnO composite heterojunction photocatalyst was also proposed.

Bismuth nitrate (99.0%), citric acid (99.5%), zinc acetate dihydrate, sodium tungstate, NaOH (96%), HNO₃ (65%–68%), Na₂C₂O₄, 4-hydroxy-2, 2, 6, 6-tetramethylpiperidinyloxy (TEMPOL), methylene blue (≥ 92%) and tetracycline (≥ 98%) were of analytical grade.

The Bi₂WO₆/ZnO heterojunction catalyst was prepared by the following two-step method [43]. The concentration and pH value of a NaOH solution can affect the morphology of as-prepared ZnO. When the concentration and pH of NaOH were 10 mol/L and 13, respectively, the morphology of the synthesized ZnO microspheres were flower-like. Therefore, the synthesis of flower-like ZnO was carried out as follows: zinc acetate (0.02 mol) and citric acid (0.014 mol) were dissolved in ethanol solution (80 mL, 20%, v/v). The mixed solution was stirred violently and then 10 M of NaOH solution was added to adjust the pH to 13.

The above solution was placed into a Teflon-lined autoclave at 150 ℃ for 15 h. After cooling, the obtained solid was filtered, washed and then dried at 105 ℃. Finally, the production was calcified at approximately 500 ℃ for 2 h to obtain ZnO nanoparticles.

Na₂WO₄·2H₂O (0.620 g) was dispersed into deionized water (20 mL) and then Bi(NO₃)₃·5H₂O (1.820 g) and 0.2 mol/L HNO₃ (15 mL) were added dropwise into the solution. The above prepared ZnO samples were added with varied molar ratios of Bi₂WO₆:Zn = 0.1:1, 0.2:1, 0.3:1 and 1:1 with sonication for 30 min. The above suspension was maintained at 140 ℃ for 20 h in a Teflon-lined autoclave. After cooling, the obtained materials were filtered, washed and then dried at 60 ℃ for 12 h to obtain Bi₂WO₆/ZnO samples, marked as (0.1:1) Bi₂WO₆/ZnO, (0.2:1) Bi₂WO₆/ZnO, (0.3:1) Bi₂WO₆/ZnO and (1:1) Bi₂WO₆/ZnO based on the various molar ratios of Bi₂WO₆ and Zn.

The characterization of photocatalysts was detected by scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), UV–vis diffuse reflectance spectroscopy (UV-DRS) and N₂ adsorption-desorption isotherms. SEM (Phenom, USA) and TEM (JEM-2100, Japan) were employed to examine the morphology. The XRD patterns were detected by a Bruker D8 Advance diffractometer. The chemical composition of Bi₂WO₆/ZnO was detected by XPS spectra (Escalab 250Xi, USA). The BET (Brunauer-Emmett-Teller) value was obtained from the N₂ adsorption-desorption isotherms with a BET analyzer (Micromeritic TriStarII 3020, USA). The UV–DRS was measured with a UV–vis spectrophotometer (Shimadzu UV-3600, Japan) from 300 to 800 nm.

The performance of Bi₂WO₆/ZnO was measured through degrading MB and tetracycline under visible light. Typically, 50 mg of Bi₂WO₆/ZnO were dissolved in 100 mL of MB solution (10 mg/L) and tetracycline solution (20 mg/L), separately. The solution containing the catalysts was stirred in the dark for 1 h until the adsorption equilibrium was reached. A 500W Xe lamp was applied as the visible light source and equipped with a 420-nm cut-off filter. The solution was then exposed to visible light irradiation and the solution was placed 15 cm from the lamp.

Samples were taken at certain intervals and filtered to remove any suspended substance. Then the supernatant was analyzed by UV-DRS to obtain the residual concentration of MB (664 nm) and tetracycline (350 nm). The degradation efficiency of the photocatalysts was calculated based on the concentration of pollutants at initial equilibrium (C₀, mg/L) and the residual concentration (C_t, mg/L) at any time t (min).

To probe the photocatalytic mechanism, several active radicals were investigated. Generally speaking, the active species include superoxide radicals [O₂]^-, holes (h⁺) and hydroxyl radicals (•OH) [44-46]. In the test, 1.0 mM isopropanol (IPA), Na₂C₂O₄ and TEMPOL were added to capture hydroxyl radicals (•OH), holes (h⁺) and superoxide radicals [O₂]^-, respectively. Several kinds of scavengers were added into the MB solution and other conditions were the same as for the degradation experiments.

The morphology and structure of the (0.3:1) Bi₂WO₆/ZnO heterojunction were researched by SEM and high resolution TEM (HR-TEM) [47].

Fig. 1(a) and (b) show the SEM images of as-prepared ZnO and (0.3:1) Bi₂WO₆/ZnO, respectively, which clearly display that the samples consisted of a flower-like structure with a diameter of 4.0 μm.

To further learn about the detailed structure of Bi₂WO₆/ZnO, HR-TEM was employed. The HR-TEM images of (0.3:1) Bi₂WO₆/ZnO clearly indicated the successful synthesis of Bi₂WO₆/ZnO, as shown in Fig. 1(c) and (d). In general, when different kinds of semiconductor materials are in close contact, they form p-n junctions at the interface with a space charge region owing to the diffusion of electrons and holes [48].

Fig. 1 shows that Bi₂WO₆ was closely attached onto ZnO to assemble the Bi₂WO₆/ZnO heterojunction. The HR-TEM images show that the lattice spaces were approximately 0.33 and 0.26 nm, which corresponded to the crystal faces of (131) and (002) for Bi₂WO₆ and ZnO, respectively. Energy dispersive spectroscopy (EDS) analysis (Fig. 1(e)) revealed that the Bi₂WO₆/ZnO sample only consisted of Bi, W, O and Zn elements.

As shown in Fig. 2, all the diffraction peaks of the photocatalysts were pure Bi₂WO₆ and ZnO, which corresponded to homogeneous orthogonal Bi₂WO₆ (JCPDS 39-256) [49] and ZnO (JCPDS 36-1451) [50], respectively. It also shows that no impurity peaks were detected. The main characteristic peaks of Bi₂WO₆ appeared at 28.3°, 32.8°, 47.1° and 56.0°, which corresponded to the (131), (200), (202) and (133) crystal faces of Bi₂WO₆, respectively [49].

The main characteristic peaks of ZnO appeared at 31.8°, 34.4°, 36.3°, 47.7°, 56.7° and 63.0°, which corresponded to the (100), (002), (101), (102), (110) and (103) crystal faces of ZnO, respectively [50]. Hence, it was clear that the composites of Bi₂WO₆/ZnO were successfully synthesized by the hydrothermal method.

Compared with pure ZnO and Bi₂WO₆, with an increasing content of Bi₂WO₆, the relative peak intensities were changed, which indicated that Bi₂WO₆ could be assimilated into the ZnO lattice to form a Bi₂WO₆/ZnO heterojunction. The peak intensities of the Bi₂WO₆/ZnO heterojunction were clearly related to the molar ratios of Bi₂WO₆ and ZnO.

Fig. 3(a) shows that the general survey spectra of (0.3:1) Bi₂WO₆/ZnO consisted of the various binding energies of Bi 4f, W 4f, O 1s, Zn 2p and C 1s states. From Fig. 3(b), the two characteristic peaks of Bi 4f with binding energies of 158.8 and 164.1 eV were consistent with Bi 4f_7/2 and Bi 4f_5/2, respectively [51]. The characteristic peaks located at 37.1 eV and 35.0 eV in Fig. 3(c) corresponded to the +6 oxidation state for W 4f_5/2 and W 4f_7/2, respectively [51]. The O 1s characteristic peaks at 529.6 eV (Fig. 3(d)) corresponded to the O_2^- in Bi₂WO₆ crystals [52].

Fig. 3(e) shows the Zn 2p spectra with two binding energies of 1044.3 and 1021.2 eV, which correspond to the 2p_1/2 and 2p_3/2 peaks in ZnO, respectively [53]. The XPS data indicated a successful synthesis of the Bi₂WO₆/ZnO photocatalyst, which was in good agreement with the XRD and HR-TEM results.

The UV-DRS patterns are presented in Fig. 4. Compared with the pure ZnO, the absorption wavelengths of Bi₂WO₆/ZnO composites showed a red shift and the absorption wavelengths extended to the visible region from 400 to 500 nm with the increasing content of Bi₂WO₆, as shown in Fig. 4(a). The as-prepared ZnO particles only exhibited a basic absorption band in the UV light region, while Bi₂WO₆/ZnO showed a slight shift to the visible light region owing to the narrower band gap of ZnO compared with that of Bi₂WO₆.

By comparison, the adsorption of the as-prepared catalysts of Bi₂WO₆/ZnO in the visible light region exhibited a significant increase owing to the addition of Bi₂WO₆ with a visible light response [54]. The values of the band gap can be calculated using Eq. (1):

(1)

where α, A and h are the absorption coefficient, constant and Planck's constant, respectively, and E_g, ν and n are the band energy, incident light frequency and constant related to the type of transition, respectively.

Fig. 4(b) shows the Kubelka–Munk transformed reflectance spectra, where the slopes are the values of the band gap energy. The obtained band gap energies of ZnO particles, (0.1:1) Bi₂WO₆/ZnO, (0.2:1) Bi₂WO₆/ZnO, (0.3:1) Bi₂WO₆/ZnO and (1:1) Bi₂WO₆/ZnO solid solutions were 3.2, 2.9, 2.8, 2.6 and 2.9 eV, respectively.

From Fig. 5, all the isotherms conformed to the H3 hysteresis loop in a type Ⅳ isotherm [55]. The surface areas (A_BET) were estimated by the N₂ isotherms to be 3.81, 195.44 and 18.98 m²/g for the catalysts of ZnO, Bi₂WO₆ and (0.3:1) Bi₂WO₆/ZnO, respectively. The results show that the A_BET of (0.3:1) Bi₂WO₆/ZnO was 4.98 times larger than ZnO, which could be attributed to the three-dimensional flower-like structures of Bi₂WO₆/ZnO and large BET value of Bi₂WO₆.

The increase of the A_BET could accelerate the enrichment and degradation rate of the target pollutants in the solution. Then, the pore size distributions of the corresponding catalysts were calculated by the BJH method (inset in Fig. 5), which indicated that the average diameters were approximately 5.55 and 18.87 nm for the catalysts of ZnO and (0.3:1) Bi₂WO₆/ZnO, respectively, and these pores were widely distributed. The inter-nanoflake/nanosheet spaces in the as-prepared catalysts may produce the mesoporous feature with widely distributed pores.

To explore the photocatalytic activity of the Bi₂WO₆/ZnO heterojunction, the photo-degradations of MB and tetracycline were performed. The activity was first evaluated through the degradation of MB. The concentration changes of MB over time during photo-degradation are shown in Fig. 6(a). The adsorption tests showed that no more than 15% of the MB molecules was adsorbed by the Bi₂WO₆/ZnO heterojunction when the equilibrium adsorption state was reached under dark conditions.

Under the irradiation of visible light, Fig. 6(a) shows the gradual drop of the MB concentration over the Bi₂WO₆/ZnO heterojunction, even with less Bi₂WO₆ ((0.1:1) Bi₂WO₆/ZnO). In addition, the (0.3:1) Bi₂WO₆/ZnO catalyst possessed the highest photocatalytic efficiency for the degradation of MB.

Tetracycline is a widely used antibiotic substance in planting and breeding areas. However, a large concentration of tetracycline remains in the surface water, groundwater and soil, owing to its limited biodegradation and the inefficiency of traditional water treatment technology, which results in a potential hazard for environment ecological systems [56]. (0.3:1) Bi₂WO₆/ZnO catalyst was used for tetracycline degradation and the results are presented in Fig. 6(b). It was clear that the (0.3:1) Bi₂WO₆/ZnO hybrid material showed the highest photocatalytic activity, compared with the other as-prepared catalysts.

The kinetic behavior of the degradation process can be explained by the kinetic models described by the following equations. Here, the Langmuir-Hinshelwood model was applied under the condition of a millimol concentration of pollutants [57, 58]:

(2)

where k is the rate constant. The plots of –ln(C_t/C₀) versus irradiation time of t are shown in Fig. 7. All catalysts were well fitted and had a very high correlation coefficient (R²> 0.94), as shown in Fig. 7.

By a linear fitting of these plots, it could be observed that ZnO had almost no photocatalytic capacity. However, the photo-degradation efficiency and rate of the as-prepared (0.3:1) Bi₂WO₆/ZnO for MB were approximately 246 and 120 times higher than those of ZnO, and for tetracycline were approximately 4500 and 200 times higher than those of ZnO, respectively, which indicated a successful recombination of Bi₂WO₆ and ZnO. The corresponding values for the different catalysts are listed in Tables 1 and 2.

The results indicated that the improved activity of the as-prepared Bi₂WO₆/ZnO could be attributed to the effective separation of the electron–hole pair between Bi₂WO₆ and ZnO.

PL is a reliable and effective method for evaluating the recombination efficiency of charge carriers [59]. The lower the luminous intensity of photoluminescence is, the lower the recombination rate of photo-generated carriers is, which implies a high efficiency of photocatalysis [60].

Fig. 8 shows that pure ZnO exhibits a strong photoluminescence at 384 nm under an excitation wavelength of 320 nm and at room temperature. Pure Bi₂WO₆ also showed a strong photoluminescence at 427 nm under the same conditions. However, the strength of the emission band of ZnO was significantly reduced after being combined with Bi₂WO₆. This situation showed that the recombination of Bi₂WO₆ and ZnO could improve the separation efficiency for photoelectron–hole pairs, thereby reducing the recombination rate of electron–hole pairs and facilitating the photocatalytic reaction.

The electrical performance of the as-obtained composite photocatalysts was explored under visible light by determining the photocurrent, where the larger the photocurrent, the higher the separation efficiency of the electron–hole pair [61]. Fig. 9 indicates the photocurrent transient response of ZnO, Bi₂WO6 and (0.3:1) Bi₂WO₆/ZnO heterojunction electrodes.

The photocurrent increased sharply and reached a steady state after 160 s upon light illumination. The (0.3:1) Bi₂WO₆/ZnO heterostructures exhibited a higher transient photocurrent density of approximately 4.5 μA compared with Bi₂WO₆ and ZnO, which indicated an enhanced separation efficiency of the photoelectron–hole pairs.

To better understand the mechanism of photocatalysis, detection of the main active radical species in the photocatalytic experiments was performed [62]. From Fig. 10(a), TEMPOL, IPA and Na₂C₂O₄ exhibited a considerable inhibition for the photo-degradation of MB. The contributions of IPA and Na₂C₂O₄ to the degradation of MB were 20.7% and 11.2%, respectively. However, the contribution of TEMPOL reached 64.8%, which was much higher than those of IPA and Na₂C₂O₄. The values of the rate constant k for the photo-degradation of MB (Fig. 10(b)) were 0.036, 0.003, 0.012 and 0.017 min^-1 for no scavenger, TEMPOL, IPA and Na₂C₂O₄, respectively.

Based on the above data, it was shown that the photo-degradation of MB over the (0.3:1) Bi₂WO₆/ZnO photocatalyst was mainly dominated by [O₂]^– radicals (formed by the generation of photo-generated electrons and dissolved oxygen), followed by •OH and h⁺.

To explore the mechanism of the photo-degradation activity of the Bi₂WO₆/ZnO heterojunction, the relative band positions of Bi₂WO₆ and ZnO were explored. The conduction band (CB) values (E_CB) were estimated by Eq. (3) [63].

(3)

where E_g, E₀, and X are the band gap energy of the catalyst, energy of free electrons on hydrogen and electronegativity of a semiconductor, respectively [64]. X values of ZnO and Bi₂WO₆ were obtained to be 5.79 and 6.36 eV and their band gap energies were 3.2 and 2.68 eV [65], respectively.

The E_CB values of ZnO and Bi₂WO₆ were -0.31 and -0.84 eV, and the valence band (VB) values (E_VB) of ZnO and Bi₂WO₆ were 2.89 and 1.89 eV, respectively. Therefore, it was clear that the values of E_CB and E_VB of ZnO were all higher than those of Bi₂WO₆. The calculated results indicated that the modification of ZnO with Bi₂WO₆ benefited the separation of the photo-generated charge carriers.

As shown in Scheme 1, photo-induced electrons and holes were produced when Bi₂WO₆ was excited by visible light. Photo-generated electrons were transferred from the CB of Bi₂WO₆ to the CB of ZnO owing to mutual contact between the two catalysts. Thus, the photo-generated electrons and holes were effectively separated, which resulted in an inhibition for the recombination of electron and holes. Moreover, a large number of electron-hole pairs and other oxidative species were produced during the process of the photocatalytic reactions. Therefore, the photocatalytic efficiency was improved and the pollutants were completely degraded. Thus, the catalysts of Bi₂WO₆/ZnO exhibited an enhanced capability compared with bare ZnO and Bi₂WO₆ [48].

Moreover, as the CB potential of Bi₂WO₆ was more negative than O₂/•O₂^- (-0.33 eV), superoxide radical anions ([O₂]^–) were produced from the reaction of dissolved oxygen molecules and electrons (e^-) on the surface of ZnO. Then, the as-produced [O₂]^– was converted into an hydroxyl radical (•OH) through protonation [66, 67]. In addition, the CB potential of ZnO was more positive than O₂/•O₂^- (-0.33 eV) but more negative than that of the reaction e_CB^- + O₂ + H⁺ → H₂O₂ (+0.682 V vs. NHE). Ultimately, the organic pollutants of MB and tetracycline were decomposed effectively by the produced strong oxidizing agent of •OH. The above photo-degradation mechanism could be proposed as follows [68]:

(4)

(5)

(6)

(7)

(8)

where [e^–] and [h⁺] refer to the electrons on the CB and holes in the VB, respectively.

The efficient heterogeneous junction photocatalyst Bi₂WO₆/ZnO was successfully synthesized through a simple hydrothermal method. After modification, the band gap energy of the ZnO-based photocatalyst was reduced to 2.6 eV from 3.2 eV. Under visible light, the Bi₂WO₆/ZnO composite showed an excellent photocatalytic activity for the degradation of MB and tetracycline. The photo-degradation efficiencies and rates for MB and tetracycline with (0.3:1) Bi₂WO₆/ZnO were hundreds of times higher than those of bare ZnO. The as-prepared (0.3:1) Bi₂WO₆/ZnO photocatalyst exhibited a higher transient photocurrent density of approximately 4.5 μA compared with bare Bi₂WO₆ and ZnO nanoparticles. The successful recombination of Bi₂WO₆ and ZnO significantly increased the effective separation of electron–hole pairs and correspondingly improved the photocatalytic activity of ZnO. Moreover, the major active species of [O₂]^– played a key role during the process of photo-degradation. This study suggested that the Bi₂WO₆/ZnO composite was a novel and effective catalyst for contaminated water restoration.