Facile Synthesis of Heterojunctioned ZnO / Bi 2 S 3 Nanocomposites for Enhanced Photocatalytic Reduction of Aqueous Cr(VI) under Visible-Light Irradiation

: Heterojunctioned ZnO / Bi 2 S 3 nanocomposites were prepared via a facile solvothermal method. The obtained photocatalysts were characterized by X-ray powder di ﬀ raction (XRD), Scanning electron microscopy (SEM), High resolution transmission electron microscopy (HRTEM), X-ray photoelectron spectroscopy (XPS), UV-Vis di ﬀ use reﬂectance spectroscopy (DRS), and Photoelectrochemical and Photoluminescence spectroscopy (PL), respectively. The results showed that ZnO / Bi 2 S 3 composites exhibited the sandwiched-like structure, where ZnO nanoparticles were randomly embedded between Bi 2 S 3 nanoﬂakes. The performance of photocatalytic Cr(VI) reduction under visible light indicated that ZnO / Bi 2 S 3 composites exhibited high-e ﬃ ciency photocatalytic activity in comparison with either Bi 2 S 3 or ZnO. The 5%-ZnO / Bi 2 S 3 photocatalyst removed 96% of Cr(VI) within 120 min at 20 mg / L initial concentration of Cr(VI). The enhanced performance of ZnO / Bi 2 S 3 photocatalysts could be ascribed to the increased light harvesting and the e ﬀ ective separation and transfer of the photogenerated charge carriers across the heterojunction interface of the ZnO / Bi 2 S 3 composite. This work could pave the way for the design of new hetero-structured materials and has great potential in environmental remediation.


Introduction
Hexavalent chromium (Cr(VI)) is a common oxidation state of chromium and is considered to be of acute toxicity, carcinogenicity and mutagen, which causes significant pollution to water and soil [1,2]. Currently, methods for disposing aqueous Cr(VI) include adsorption [3], reverse osmosis [4], ion exchange [5], etc. However, those means are commonly costly and nonrenewable. Lots of recent studies have shown that photocatalytic technology is regarded as a promising technique toward removing aqueous Cr(VI) because of the merits of cheap cost, high efficiency, simple fabrication, environment-friendly, and without discharge of undesirable chemicals [6][7][8]. As a consequence, the fabrication of efficient photocatalysts is inseparable to develop a high-performance photocatalytic process [9,10].
The well-known photocatalyst TiO 2 (E g = 3.2 eV) is only sensitive to ultraviolet light and requires a mass of illumination time for the reduction of Cr(VI), which limits its practical application [11,12]. However, ultraviolet light is very little and only about 5% of the complete solar spectrum. Thence, it is desirable to manufacture innovative visible light-responsive photocatalysts to effectively enhance the removal of Cr(VI). Bi 2 S 3 with a narrow band gap (E g = 1.2~1.7 eV) has attracted great attention thanks to its strong visible-light absorption, no pollution and low cost [13,14]. Arpita Sarkar [15] et al. synthesized morphologically-tuned Bi 2 S 3 NPs by a simple solvothermal method, which emerged

Characterization of ZnO/Bi 2 S 3 Photocatalysts Composites
XRD is utilized to analyze the purity and crystallographic structure of the as-prepared samples. Figure 1 displays the XRD patterns of pure ZnO, Bi 2 S 3 and ZnO/Bi 2 S 3 composites with different ZnO loadings. For pure ZnO, the typical peaks could be indexed to hexagonal wurtzite ZnO (JCPDS No.   [30]. The diffraction peaks at 2θ = 31.  [31]. However, due to its low loading, the (110) peaks of ZnO in the composites could not be clearly detected in the XRD pattern. All diffraction peaks of the as-prepared composites matched exactly with those of pure ZnO and Bi 2 S 3 phases and no other impurity peaks were detected, unambiguously indicating that ZnO/Bi 2 S 3 composites were only composed of Bi 2 S 3 and ZnO phases. The morphologies of ZnO, Bi2S3 and 5%-ZnO/Bi2S3 samples are investigated by SEM. Pure ZnO exhibited a regular nanoparticles aggregate with an average particle size of ca. 60-80 nm (Figure 2a). Pure Bi2S3 presented sword-shaped thin nanoplates with lengths of ca. 9-15 um, thicknesses of ca. 1-1.8 um and some tiny irregular nanoparticles with an average particle size of ca. 30 nm (Figure 2b). Only Bi and S elements were detected from element mapping images ( Figure S1) and their atomic ratio was 2:3, exactly equal to the theoretic formula of pure Bi2S3. 5%-ZnO/Bi2S3 sample exhibited the sandwiched-like structure, where some tiny nanoparticles were randomly embedded between the nanoflakes (Figure 2c). From Figure 2a,b, these nanoparticles would probably be either ZnO or Bi2S3. However, the elemental mapping showed ( Figure S2) that Bi, S, Zn, and O could be observed throughout the prepared sample. This proved that sandwiched ZnO nanoparticles were randomly embedded, which demonstrated the existence of ZnO/Bi2S3 heterojunctions [32]. Besides, Compared with Figure 2b, the sheets of Bi2S3 became thinner and their lengths were reduced to 3~5 um. The sandwiched ZnO nanoparticles effectively prevented the re-stacking of lamellar Bi2S3 nanoplates. Additionally, from magnified Figure 2d, the ZnO nanoparticles were observed to be attached tightly to Bi2S3 nanoplates, thereby contributing to the fabrication of a heterojunction between ZnO and Bi2S3 within the composites. The morphologies of ZnO, Bi 2 S 3 and 5%-ZnO/Bi 2 S 3 samples are investigated by SEM. Pure ZnO exhibited a regular nanoparticles aggregate with an average particle size of ca. 60-80 nm ( Figure 2a). Pure Bi 2 S 3 presented sword-shaped thin nanoplates with lengths of ca. 9-15 um, thicknesses of ca. 1-1.8 um and some tiny irregular nanoparticles with an average particle size of ca. 30 nm (Figure 2b). Only Bi and S elements were detected from element mapping images ( Figure S1) and their atomic ratio was 2:3, exactly equal to the theoretic formula of pure Bi 2 S 3 . 5%-ZnO/Bi 2 S 3 sample exhibited the sandwiched-like structure, where some tiny nanoparticles were randomly embedded between the nanoflakes (Figure 2c). From Figure 2a,b, these nanoparticles would probably be either ZnO or Bi 2 S 3 . However, the elemental mapping showed ( Figure S2) that Bi, S, Zn, and O could be observed throughout the prepared sample. This proved that sandwiched ZnO nanoparticles were randomly embedded, which demonstrated the existence of ZnO/Bi 2 S 3 heterojunctions [32]. Besides, Compared with Figure 2b, the sheets of Bi 2 S 3 became thinner and their lengths were reduced to 3~5 um. The sandwiched ZnO nanoparticles effectively prevented the re-stacking of lamellar Bi 2 S 3 nanoplates. Additionally, from magnified Figure 2d, the ZnO nanoparticles were observed to be attached tightly to Bi 2 S 3 nanoplates, thereby contributing to the fabrication of a heterojunction between ZnO and Bi 2 S 3 within the composites. The morphologies of ZnO, Bi2S3 and 5%-ZnO/Bi2S3 samples are investigated by SEM. Pure ZnO exhibited a regular nanoparticles aggregate with an average particle size of ca. 60-80 nm ( Figure 2a). Pure Bi2S3 presented sword-shaped thin nanoplates with lengths of ca. 9-15 um, thicknesses of ca. 1-1.8 um and some tiny irregular nanoparticles with an average particle size of ca. 30 nm (Figure 2b). Only Bi and S elements were detected from element mapping images ( Figure S1) and their atomic ratio was 2:3, exactly equal to the theoretic formula of pure Bi2S3. 5%-ZnO/Bi2S3 sample exhibited the sandwiched-like structure, where some tiny nanoparticles were randomly embedded between the nanoflakes (Figure 2c). From Figure 2a,b, these nanoparticles would probably be either ZnO or Bi2S3. However, the elemental mapping showed ( Figure S2) that Bi, S, Zn, and O could be observed throughout the prepared sample. This proved that sandwiched ZnO nanoparticles were randomly embedded, which demonstrated the existence of ZnO/Bi2S3 heterojunctions [32]. Besides, Compared with Figure 2b, the sheets of Bi2S3 became thinner and their lengths were reduced to 3~5 um. The sandwiched ZnO nanoparticles effectively prevented the re-stacking of lamellar Bi2S3 nanoplates. Additionally, from magnified Figure 2d, the ZnO nanoparticles were observed to be attached tightly to Bi2S3 nanoplates, thereby contributing to the fabrication of a heterojunction between ZnO and Bi2S3 within the composites. More detailed structures of the 5%-ZnO/Bi2S3 sample are investigated by HRTEM, in which the interface of ZnO and Bi2S3 can be observed. Figure 3a manifests a large number of irregular nanosheets ranging in size from 200 to 500 nm. Furthermore, it could be seen that some tiny nanocrystals marked by a red circle in Figure 3a were conglutinated tightly to the nanosheets. Figure  3b exhibits a lattice image from the (100) planes with the interplaner spacing d (100) = 0.280 nm of ZnO [33] and that from the (130) planes with the interplaner spacing d (130) = 0.360 nm of Bi2S3 [34]. Moreover, intimate interface was clearly detected in Figure 3b and the heterojunction was somewhat preferably formed during the solvothermal fabrication of the 5%-ZnO/Bi2S3 composite [35,36]. The heterojunction structure within the prepared composite would favor the effective transfer of photoinduced carriers across the interface upon exposure to light and improve the photocatalytic performance of the resulting photocatalysts. XPS is employed for the evaluation of chemical states of the elements in 5%-ZnO/Bi2S3 composites. The survey spectrum (Figure 4a) revealed that Bi, S, Zn, and O elements existed in this composite and Zn, Bi, and O elements were further analyzed from the spectra of Zn 2p, Bi 4f and O 1s. The main peaks at 1022.3 and 1044.9 eV were allocated to Zn 2p3/2 and Zn 2p1/2 (Figure 4b), respectively, verifying the existence of Zn 2+ in the composite [37]. Figure 4c shows two strong peaks at 164.1 and 158.6 eV denoted as Bi4f5/2 and Bi4f7/2, respectively, revealing the existence of Bi 3+ in the More detailed structures of the 5%-ZnO/Bi 2 S 3 sample are investigated by HRTEM, in which the interface of ZnO and Bi 2 S 3 can be observed. Figure 3a manifests a large number of irregular nanosheets ranging in size from 200 to 500 nm. Furthermore, it could be seen that some tiny nanocrystals marked by a red circle in Figure 3a were conglutinated tightly to the nanosheets. Figure 3b exhibits a lattice image from the (100) planes with the interplaner spacing d (100) = 0.280 nm of ZnO [33] and that from the (130) planes with the interplaner spacing d (130) = 0.360 nm of Bi 2 S 3 [34]. Moreover, intimate interface was clearly detected in Figure 3b and the heterojunction was somewhat preferably formed during the solvothermal fabrication of the 5%-ZnO/Bi 2 S 3 composite [35,36]. The heterojunction structure within the prepared composite would favor the effective transfer of photo-induced carriers across the interface upon exposure to light and improve the photocatalytic performance of the resulting photocatalysts. More detailed structures of the 5%-ZnO/Bi2S3 sample are investigated by HRTEM, in which the interface of ZnO and Bi2S3 can be observed. Figure 3a manifests a large number of irregular nanosheets ranging in size from 200 to 500 nm. Furthermore, it could be seen that some tiny nanocrystals marked by a red circle in Figure 3a were conglutinated tightly to the nanosheets. Figure  3b exhibits a lattice image from the (100) planes with the interplaner spacing d (100) = 0.280 nm of ZnO [33] and that from the (130) planes with the interplaner spacing d (130) = 0.360 nm of Bi2S3 [34]. Moreover, intimate interface was clearly detected in Figure 3b and the heterojunction was somewhat preferably formed during the solvothermal fabrication of the 5%-ZnO/Bi2S3 composite [35,36]. The heterojunction structure within the prepared composite would favor the effective transfer of photoinduced carriers across the interface upon exposure to light and improve the photocatalytic performance of the resulting photocatalysts. XPS is employed for the evaluation of chemical states of the elements in 5%-ZnO/Bi2S3 composites. The survey spectrum (Figure 4a) revealed that Bi, S, Zn, and O elements existed in this composite and Zn, Bi, and O elements were further analyzed from the spectra of Zn 2p, Bi 4f and O 1s. The main peaks at 1022.3 and 1044.9 eV were allocated to Zn 2p3/2 and Zn 2p1/2 (Figure 4b), respectively, verifying the existence of Zn 2+ in the composite [37]. Figure 4c shows two strong peaks at 164.1 and 158.6 eV denoted as Bi4f5/2 and Bi4f7/2, respectively, revealing the existence of Bi 3+ in the  (Figure 4b), respectively, verifying the existence of Zn 2+ in the composite [37]. Figure 4c shows two strong peaks at 164.1 and 158.6 eV denoted as Bi4f 5/2 and Bi4f 7/2 , respectively, revealing the existence of Bi 3+ in the composite, which was consistent with other XPS results in Bi 2 S 3 [26]. Also, as seen in Figure 4d, the O1s profile was asymmetric and could be fitted into two symmetrical peaks at 530.0 and 531.8 eV, manifesting two different types of O species in the composite. The two peaks should be connected with the lattice oxygen (O-L) of ZnO and the chemisorption oxygen (O-H) caused by surface hydroxyl, respectively [38].
Catalysts 2019, 9, x FOR PEER REVIEW 5 of 14 composite, which was consistent with other XPS results in Bi2S3 [26]. Also, as seen in Figure 4d, the O1s profile was asymmetric and could be fitted into two symmetrical peaks at 530.0 and 531.8 eV, manifesting two different types of O species in the composite. The two peaks should be connected with the lattice oxygen (O-L) of ZnO and the chemisorption oxygen (O-H) caused by surface hydroxyl, respectively [38]. The DRS spectra of the as-synthesized 5%-ZnO/Bi2S3 nanocomposite as well as ZnO and Bi2S3 for comparison are shown in Figure 5a. It could be clearly observed that bare ZnO possessed strong UV light absorption, whereas it hardly absorbed light of the visible region. Bi2S3 and 5%-ZnO/Bi2S3 composites exhibited a similar light response profile and both showed strong absorption to the entire wavelength range. However, after the modification of ZnO, the absorbance of the composite to visible light was significantly enhanced, probably due to the synergistic absorption effect of the ZnO and Bi2S3 phase of the composite. The improved light response of ZnO/Bi2S3 unambiguously favored the photocatalytic performance of the prepared composite. Furthermore, using the Kubelka-Munk function (Equation (1)), one can count the band gap energy of bare ZnO, Bi2S3 and 5%-ZnO/Bi2S3 composite: where Eg, v, A, h, and k are band gap energy, light frequency, a constant, planck constant, and the absorption coefficient. It had been reported that ZnO [39] and Bi2S3 [40] had n values of 1. The Eg value of ZnO, Bi2S3, and 5%-ZnO/Bi2S3 were estimated to be 3.20 eV, 1.60 eV, and 1.52 eV, respectively ( Figure 4b). The bandgap of Bi2S3 was very close to the Sun's report [18]. In addition, the band gap energy was reported to be dependent on the size of crystal, but the conduction band (CB) potential is hardly affected by the size [15,16]. Obviously, the narrowed band gap of the ZnO/Bi2S3 nanocomposite was definitely helpful to the enhancement of the photocatalytic activity of the resulting composites. The DRS spectra of the as-synthesized 5%-ZnO/Bi 2 S 3 nanocomposite as well as ZnO and Bi 2 S 3 for comparison are shown in Figure 5a. It could be clearly observed that bare ZnO possessed strong UV light absorption, whereas it hardly absorbed light of the visible region. Bi 2 S 3 and 5%-ZnO/Bi 2 S 3 composites exhibited a similar light response profile and both showed strong absorption to the entire wavelength range. However, after the modification of ZnO, the absorbance of the composite to visible light was significantly enhanced, probably due to the synergistic absorption effect of the ZnO and Bi 2 S 3 phase of the composite. The improved light response of ZnO/Bi 2 S 3 unambiguously favored the photocatalytic performance of the prepared composite. Furthermore, using the Kubelka-Munk function (Equation (1)), one can count the band gap energy of bare ZnO, Bi 2 S 3 and 5%-ZnO/Bi 2 S 3 composite: where E g , v, A, h, and k are band gap energy, light frequency, a constant, planck constant, and the absorption coefficient. It had been reported that ZnO [39] and Bi 2 S 3 [40] had n values of 1. The E g value of ZnO, Bi 2 S 3 , and 5%-ZnO/Bi 2 S 3 were estimated to be 3.20 eV, 1.60 eV, and 1.52 eV, respectively ( Figure 4b). The bandgap of Bi 2 S 3 was very close to the Sun's report [18]. In addition, the band gap energy was reported to be dependent on the size of crystal, but the conduction band (CB) potential is hardly affected by the size [15,16]. Obviously, the narrowed band gap of the ZnO/Bi 2 S 3 nanocomposite was definitely helpful to the enhancement of the photocatalytic activity of the resulting composites. Photocurrent measurement can also provide evidence for the separation rate of photo-generated carriers when the photocatalyst is excited by light. Figure 6 presents the transient photocurrent responses for pure ZnO, Bi2S3 and 5%-ZnO/Bi2S3 composite under visible-light irradiation. The current density of 5%-ZnO/Bi2S3 composite was dramatically increased, which was 10 times and 1.6 times that of bare ZnO and Bi2S3, indicating that photoelectron-hole pairs excited over 5%-ZnO/Bi2S3 composite were effectively separated and transferred across the heterojunction interface between ZnO and Bi2S3 within the composite.
The PL spectrum detects the separation efficiency of the electron-hole pairs for samples. It could be found that all the catalysts exhibited broad emission peaks around 450-500 nm ( Figure 7). The smaller the PL intensity, the less recombination of the photogenerated charge carriers involved in the overall reaction [41]. Obviously, 5%-ZnO/Bi2S3 nanocomposite showed reduced PL intensity as compared to pure ZnO and Bi2S3, suggesting that the photo-generated charge carries for this composite were effectively separated and their lifetime was prolonged, which was largely ascribed to the heterojunction structure within the ZnO/Bi2S3 composite.  Photocurrent measurement can also provide evidence for the separation rate of photo-generated carriers when the photocatalyst is excited by light. Figure 6 presents the transient photocurrent responses for pure ZnO, Bi 2 S 3 and 5%-ZnO/Bi 2 S 3 composite under visible-light irradiation. The current density of 5%-ZnO/Bi 2 S 3 composite was dramatically increased, which was 10 times and 1.6 times that of bare ZnO and Bi 2 S 3 , indicating that photoelectron-hole pairs excited over 5%-ZnO/Bi 2 S 3 composite were effectively separated and transferred across the heterojunction interface between ZnO and Bi 2 S 3 within the composite. Photocurrent measurement can also provide evidence for the separation rate of photo-generated carriers when the photocatalyst is excited by light. Figure 6 presents the transient photocurrent responses for pure ZnO, Bi2S3 and 5%-ZnO/Bi2S3 composite under visible-light irradiation. The current density of 5%-ZnO/Bi2S3 composite was dramatically increased, which was 10 times and 1.6 times that of bare ZnO and Bi2S3, indicating that photoelectron-hole pairs excited over 5%-ZnO/Bi2S3 composite were effectively separated and transferred across the heterojunction interface between ZnO and Bi2S3 within the composite.
The PL spectrum detects the separation efficiency of the electron-hole pairs for samples. It could be found that all the catalysts exhibited broad emission peaks around 450-500 nm (Figure 7). The smaller the PL intensity, the less recombination of the photogenerated charge carriers involved in the overall reaction [41]. Obviously, 5%-ZnO/Bi2S3 nanocomposite showed reduced PL intensity as compared to pure ZnO and Bi2S3, suggesting that the photo-generated charge carries for this composite were effectively separated and their lifetime was prolonged, which was largely ascribed to the heterojunction structure within the ZnO/Bi2S3 composite.  The PL spectrum detects the separation efficiency of the electron-hole pairs for samples. It could be found that all the catalysts exhibited broad emission peaks around 450-500 nm (Figure 7). The smaller the PL intensity, the less recombination of the photogenerated charge carriers involved in the overall reaction [41]. Obviously, 5%-ZnO/Bi 2 S 3 nanocomposite showed reduced PL intensity as compared to pure ZnO and Bi 2 S 3 , suggesting that the photo-generated charge carries for this composite were effectively separated and their lifetime was prolonged, which was largely ascribed to the heterojunction structure within the ZnO/Bi 2 S 3 composite.

Photoreduction of Aqueous Cr(VI) Under Visible-Light Irradiation
Photocatalytic performances of the as-prepared photocatalysts were measured through the removal of Cr(VI) of 20 mg/L and catalyst dosage of 0.05 g under visible-light irradiation. After 1 h's dark adsorption, ZnO showed hardly adsorption ability towards aqueous Cr(VI). The adsorption removal ratio of 3%-ZnO/Bi2S3, 5%-ZnO/Bi2S3, 10%-ZnO/Bi2S3, and Bi2S3 for Cr(VI) were 43%, 33%, 45%, and 57%, respectively. Figure 8a reveals the variation of Cr(VI) concentrations with irradiation time catalyzed by the photocatalysts under visible light. The blank test (without photocatalyst) showed little photolysis under 2 h's visible light exposure, indicating that aqueous Cr(VI) was quite stable under visible-light irradiation. Upon irradiation, the solution containing Cr(VI) ions gradually lost its originally yellow color and the solution turned pale green. So the product of photocatalytic reduction of Cr(VI) would be Cr(III) ions [35]. The photoreduction rate toward Cr(VI) over pure ZnO and Bi2S3 was 10% and 77%, respectively. After modification of ZnO, the prepared composites had a higher photocatalytic performance than bare ZnO and pure Bi2S3. Compared with bare ZnO and Bi2S3, the catalytic activities of ZnO/Bi2S3 composites were significantly improved. The activity strongly depended on the ZnO content of the composites. When the content of Zno was increased, the photocatalytic activity of the composite was also increased, and 5% ZnO/Bi2S3 composite exhibited a maximum removal rate of 95%. However, further increasing the content of ZnO, the photocatalytic performance of the composite was reduced. To further compare the photocatalytic Cr(VI) reduction performance of the different samples, the pseudo-first-order reaction kinetic model was used to determine the photoreduction apparent reaction rate constants (k), where -ln(C/C0) versus t were delineated (Figure 8b). The k values of different samples were in the following order: 5%-ZnO/Bi2S3 > 3%-ZnO/Bi2S3 > 10%-ZnO/ Bi2S3 > Bi2S3 > ZnO. The k value of 5%-ZnO/Bi2S3 was 30-fold that of virgin ZnO and 2.27-fold that of bare Bi2S3, further indicating that the photocatalytic Cr(VI) reduction activity of ZnO/Bi2S3 was highly dependent on the loading of ZnO, and the photocatalytic performance of Bi2S3 toward Cr(VI) removal was greatly promoted by incorporating the ZnO component. Therefore, the 5%-ZnO/Bi2S3 composite was employed to further investigate its photocatalytic activity in the following study.

Photoreduction of Aqueous Cr(VI) Under Visible-Light Irradiation
Photocatalytic performances of the as-prepared photocatalysts were measured through the removal of Cr(VI) of 20 mg/L and catalyst dosage of 0.05 g under visible-light irradiation. After 1 h's dark adsorption, ZnO showed hardly adsorption ability towards aqueous Cr(VI). The adsorption removal ratio of 3%-ZnO/Bi 2 S 3 , 5%-ZnO/Bi 2 S 3 , 10%-ZnO/Bi 2 S 3 , and Bi 2 S 3 for Cr(VI) were 43%, 33%, 45%, and 57%, respectively. Figure 8a reveals the variation of Cr(VI) concentrations with irradiation time catalyzed by the photocatalysts under visible light. The blank test (without photocatalyst) showed little photolysis under 2 h's visible light exposure, indicating that aqueous Cr(VI) was quite stable under visible-light irradiation. Upon irradiation, the solution containing Cr(VI) ions gradually lost its originally yellow color and the solution turned pale green. So the product of photocatalytic reduction of Cr(VI) would be Cr(III) ions [35]. The photoreduction rate toward Cr(VI) over pure ZnO and Bi 2 S 3 was 10% and 77%, respectively. After modification of ZnO, the prepared composites had a higher photocatalytic performance than bare ZnO and pure Bi 2 S 3 . Compared with bare ZnO and Bi 2 S 3 , the catalytic activities of ZnO/Bi 2 S 3 composites were significantly improved. The activity strongly depended on the ZnO content of the composites. When the content of Zno was increased, the photocatalytic activity of the composite was also increased, and 5% ZnO/Bi 2 S 3 composite exhibited a maximum removal rate of 95%. However, further increasing the content of ZnO, the photocatalytic performance of the composite was reduced. To further compare the photocatalytic Cr(VI) reduction performance of the different samples, the pseudo-first-order reaction kinetic model was used to determine the photoreduction apparent reaction rate constants (k), where -ln(C/C 0 ) versus t were delineated (Figure 8b). The k values of different samples were in the following order: 5%-ZnO/Bi 2 S 3 > 3%-ZnO/Bi 2 S 3 > 10%-ZnO/ Bi 2 S 3 > Bi 2 S 3 > ZnO. The k value of 5%-ZnO/Bi 2 S 3 was 30-fold that of virgin ZnO and 2.27-fold that of bare Bi 2 S 3 , further indicating that the photocatalytic Cr(VI) reduction activity of ZnO/Bi 2 S 3 was highly dependent on the loading of ZnO, and the photocatalytic performance of Bi 2 S 3 toward Cr(VI) removal was greatly promoted by incorporating the ZnO component. Therefore, the 5%-ZnO/Bi 2 S 3 composite was employed to further investigate its photocatalytic activity in the following study.  Figure 8c displays the effect of pH(adjusted by aqueous HCl or NaOH solution) on the removal efficiency of 5%-ZnO/Bi2S3 (20 mg/L aqueous Cr(VI)). When the initial solution pH was 3.0, 5%-ZnO/Bi2S3 removed 75.4% of aqueous Cr(VI) within 60 min. With the increasing of the solution pH (5.25, 8.0, and 11.0), the photoreduction efficiency of Cr(VI) decreased noticeably from 69.3%, 52.3% to 33.3%. On the one hand, the protonation of the catalyst surface will attract HCrO4 − and Cr2O7 2− in acidic solution, which will promote the reduction of Cr (VI) [42]. On the other hand, when the solution is alkaline, the catalyst surface adsorbs a large amount of ·OH and the produced Cr(OH)3 precipitates  acidic solution, which will promote the reduction of Cr (VI) [42]. On the other hand, when the solution is alkaline, the catalyst surface adsorbs a large amount of ·OH and the produced Cr(OH) 3 precipitates on its surface, which reduces the light harvesting, thus lowering the reduction removal of chromium ions [36]. Figure 8d depicts the effect of different initial Cr(VI) concentration for photoreduction removal using a 5%-ZnO/Bi 2 S 3 photocatalyst. The degradation rate of Cr(VI) 20 mg/L was 95.9% within 2 h of visible light exposure. The degradation rates of Cr(VI) with incipient concentrations of 40 mg/L and 60 mg/L were 52% and 28% at the 6 h interval, respectively. A reason for this is the concentration of Chromium ion in the initial solution increases while the amount of catalyst feeding remains unchanged, which leads to more Cr(VI) ions adsorbing onto the catalyst solid, causing the decrease of adsorption sites on the photocatalyst, and thus decreasing the photoreduction efficiency of 5%-ZnO/Bi 2 S 3 photocatalyst [43]. Figure 8e presents the effect of catalyst dosage for photocatalytic Cr(VI) reduction activity of 5%-ZnO/Bi 2 S 3 composite. When 0.10 g of 5%-ZnO/Bi 2 S 3 was used, the 100% Cr(VI) photoreduction removal was observed within 1 h's irradiation. With the decrease of the photocatalyst dosage to 0.02 g, the Cr(VI) reduction rate was dramatically reduced to only 42% in 5 h. This behavior could be attributed to a decrease in adsorption sites on the photocatalyst and a subsequent drop of photo-induced free electrons in the conduction band during photocatalysis.
To assess the stability and reusability of 5%-ZnO/Bi 2 S 3 composite in the circulating tests for the photocatalytic reduction of Cr(VI) under visible-light irradiation, the photocatalyst solid was collected by centrifugation and washed with DI water prior to the addition of fresh Cr(VI) solution for the next run. From Figure 8f, 5%-ZnO/Bi 2 S 3 photocatalyst exhibited no obvious change of the photocatalytic activity after four cycles with the removal rate still reaching 92%, suggesting that no significant damage of the structure of the composites took place and the as-prepared composite photocatalyst had excellent photo-stability and long-term reusability. Compared with bare ZnO with photocorrosion effect [44], the prepared photocatalyst had great potential for water remediation

Mechanism for Photocatalytic Cr(VI) Reduction
To make clear the possible reactive species involved in the ZnO/Bi 2 S 3 catalyzed reduction of aqueous Cr(VI) under visible light irradiation, the species trapping experiments were conducted, in which BQ, IPA and CA were added as scavengers of·O 2 − ,·OH and h + into Cr(VI) solution prior to the addition of the catalyst, respectively, and their final concentrations in the system were 0.1, 2.0, and 20.0 mmol/L, respectively. The trapping results are presented in Figure 9.
Catalysts 2019, 9, x FOR PEER REVIEW 9 of 14 on its surface, which reduces the light harvesting, thus lowering the reduction removal of chromium ions [36]. Figure 8d depicts the effect of different initial Cr(VI) concentration for photoreduction removal using a 5%-ZnO/Bi2S3 photocatalyst. The degradation rate of Cr(VI) 20 mg/L was 95.9% within 2 h of visible light exposure. The degradation rates of Cr(VI) with incipient concentrations of 40 mg/L and 60 mg/L were 52% and 28% at the 6 h interval, respectively. A reason for this is the concentration of Chromium ion in the initial solution increases while the amount of catalyst feeding remains unchanged, which leads to more Cr(VI) ions adsorbing onto the catalyst solid, causing the decrease of adsorption sites on the photocatalyst, and thus decreasing the photoreduction efficiency of 5%-ZnO/Bi2S3 photocatalyst [43]. Figure 8e presents the effect of catalyst dosage for photocatalytic Cr(VI) reduction activity of 5%-ZnO/Bi2S3 composite. When 0.10 g of 5%-ZnO/Bi2S3 was used, the 100% Cr(VI) photoreduction removal was observed within 1 h's irradiation. With the decrease of the photocatalyst dosage to 0.02 g, the Cr(VI) reduction rate was dramatically reduced to only 42% in 5 h. This behavior could be attributed to a decrease in adsorption sites on the photocatalyst and a subsequent drop of photoinduced free electrons in the conduction band during photocatalysis.
To assess the stability and reusability of 5%-ZnO/Bi2S3 composite in the circulating tests for the photocatalytic reduction of Cr(VI) under visible-light irradiation, the photocatalyst solid was collected by centrifugation and washed with DI water prior to the addition of fresh Cr(VI) solution for the next run. From Figure 8f, 5%-ZnO/Bi2S3 photocatalyst exhibited no obvious change of the photocatalytic activity after four cycles with the removal rate still reaching 92%, suggesting that no significant damage of the structure of the composites took place and the as-prepared composite photocatalyst had excellent photo-stability and long-term reusability. Compared with bare ZnO with photocorrosion effect [44], the prepared photocatalyst had great potential for water remediation

Mechanism for Photocatalytic Cr(VI) Reduction
To make clear the possible reactive species involved in the ZnO/Bi2S3 catalyzed reduction of aqueous Cr(VI) under visible light irradiation, the species trapping experiments were conducted, in which BQ, IPA and CA were added as scavengers of·O2 − ,·OH and h + into Cr(VI) solution prior to the addition of the catalyst , respectively, and their final concentrations in the system were 0.1, 2.0, and 20.0 mmol/L, respectively. The trapping results are presented in Figure 9. The photocatalytic reduction of Cr(VI) was not affected in the presence of IPA, indicating that ·OH was not an active species in this photocatalytic system. The degradation rate of Cr(VI) was significantly reduced after the addition of BQ with only 60% removal rate in 2 h. The reason why BQ reduced the photocatalytic rate of Cr(VI) was that BQ captured electron-generated O2 − , which reduced the amount of H2O2 according to the classical photo-catalysis mechanism, thereby decreasing the photoreduction The photocatalytic reduction of Cr(VI) was not affected in the presence of IPA, indicating that ·OH was not an active species in this photocatalytic system. The degradation rate of Cr(VI) was significantly reduced after the addition of BQ with only 60% removal rate in 2 h. The reason why BQ reduced the photocatalytic rate of Cr(VI) was that BQ captured electron-generated O 2 − , which reduced the amount of H 2 O 2 according to the classical photo-catalysis mechanism, thereby decreasing the photoreduction removal rate of Cr(VI). Therefore,·O 2 − was the main active species in this photocatalytic reaction process. Meanwhile, purging nitrogen gases dramatically reduced the photocatalytic rate of Cr(VI) from 95% to 56%. When O 2 was replaced by N 2 , the reaction of successive species formation could not proceed and this phenomena was in good agreement with that of BQ scavenger, further suggesting that O 2 − played an important role in our photoreduction process. Upon the addition of CA, the photocatalytic activity of the 5%-ZnO/Bi 2 S 3 composite toward Cr(VI) photoreduction was boosted and Cr(VI) was completely removed after 1 h of illumination. The consumption of h + by CA-trapping decreased the recombination of hole-electron pairs so that there were more electrons to participate in the photoreduction reaction. In brief, photo-excited e − played a decisive role in our phtotoreduction of Cr(VI). Under air conditions, the intermediate H 2 O 2 also contributed to the ZnO/Bi 2 S 3 -catalyzed reduction of acidic aqueous Cr(VI) under visible light irradiation. Based on the above experimental results, a rational mechanism of ZnO/Bi 2 S 3 photocatalyst for the improved photoreduction of Cr(VI) was raised ( Figure 10). Bi 2 S 3 could directly absorb the light of 400-700 nm to generate electrons (e − ) and holes (h + ), and electrons were excited from the valence band (VB) to CB, whereas ZnO (E g = 3.2 eV) cannot be excited under visible light irradiation because of its broad energy gap. Because the CB (−0.33 eV) position of Bi 2 S 3 [18] is more negative than that of ZnO [35] (−0.31 eV), the electrons of CB in Bi 2 S 3 are rapidly transferred to ZnO and leave holes on the Bi 2 S 3 VB. This is feasible in the laws of thermodynamics and the transfer is preferable due to the heterojunction of ZnO/Bi 2 S 3 composites. Therefore, the photoinduced electrons and holes are separated efficiently. In the case of ZnO, its CB is more negative than E θ (Cr 2 O 7 2− /Cr 3+ ) = +1.33 eV (vs. NHE). The photo-induced electrons are accumulated to reduce Cr 6+ to Cr 3+ (Equation (2) (3)). In addition, a portion of electrons located on the surface of the ZnO react with the adsorbed oxygen to produce ·O 2 − , which ultimately produces hydrogen peroxide (H 2 O 2 ) (Equations (4) and (5)). Cr 6+ is reduced to Cr 3+ in acidic solution due to the strong reducibility of H 2 O 2 [45]. The overall reaction process is illustrated by the following equations: Catalysts 2019, 9, x FOR PEER REVIEW 10 of 14 removal rate of Cr(VI). Therefore,·O2 − was the main active species in this photocatalytic reaction process. Meanwhile, purging nitrogen gases dramatically reduced the photocatalytic rate of Cr(VI) from 95% to 56%. When O2 was replaced by N2, the reaction of successive species formation could not proceed and this phenomena was in good agreement with that of BQ scavenger, further suggesting that O2 − played an important role in our photoreduction process. Upon the addition of CA, the photocatalytic activity of the 5%-ZnO/Bi2S3 composite toward Cr(VI) photoreduction was boosted and Cr(VI) was completely removed after 1 h of illumination. The consumption of h + by CA-trapping decreased the recombination of hole-electron pairs so that there were more electrons to participate in the photoreduction reaction. In brief, photo-excited e − played a decisive role in our phtotoreduction of Cr(VI). Under air conditions, the intermediate H2O2 also contributed to the ZnO/Bi2S3-catalyzed reduction of acidic aqueous Cr(VI) under visible light irradiation. Based on the above experimental results, a rational mechanism of ZnO/Bi2S3 photocatalyst for the improved photoreduction of Cr(VI) was raised ( Figure 10). Bi2S3 could directly absorb the light of 400-700 nm to generate electrons (e − ) and holes (h + ), and electrons were excited from the valence band (VB) to CB, whereas ZnO (Eg = 3.2 eV) cannot be excited under visible light irradiation because of its broad energy gap. Because the CB (−0.33 eV) position of Bi2S3 [18] is more negative than that of ZnO [35] (−0.31 eV), the electrons of CB in Bi2S3 are rapidly transferred to ZnO and leave holes on the Bi2S3 VB. This is feasible in the laws of thermodynamics and the transfer is preferable due to the heterojunction of ZnO/Bi2S3 composites. Therefore, the photoinduced electrons and holes are separated efficiently. In the case of ZnO, its CB is more negative than E θ (Cr2O7 2− /Cr 3+ ) = +1.33 eV (vs. NHE). The photo-induced electrons are accumulated to reduce Cr 6+ to Cr 3+ (Equation (2)). Meanwhile, due to the more positive VB (1.27 eV) position of Bi2S3 than the oxidation potential of E θ (H2O/O2) = 1.23 eV (vs. NHE), the holes left on the VB of Bi2S3 oxidize H2O to O2 (Equation (3)). In addition, a portion of electrons located on the surface of the ZnO react with the adsorbed oxygen to produce ·O2 − , which ultimately produces hydrogen peroxide (H2O2) (Equations (4) and (5)). Cr 6+ is reduced to Cr 3+ in acidic solution due to the strong reducibility of H2O2 [45]. The overall reaction process is illustrated by the following equations:

Preparation of the Photocatalysts
All chemicals were of analytical grade purity in the experiments and without further purification. Zn(CH 3 COO) 2 ·2H 2 O (1.98 g) was dissolved in 50 mL of deionized (DI) water and vigorously stirred at 60 • C for 30 min. Forty milliliters of NaOH (0.3 mol/L) solution was added dropwise to adjust the solution to pH = 8. The mixture was further stirred for 60 min and then transferred into a 100 mL Teflon-lined autoclave and maintained at 160 • C for 12 h. After being cooled naturally to room temperature, the precipitate was filtered, washed successively with DI water and ethanol, and finally dried at 60 • C for 12 h. Pure ZnO powder was obtained by calcining this solid at 400 • C for 3 h.
Zero point zero six grams (0.06 g) of ZnO and 1.2 g of Bi(NO 3 ) 3 ·5H 2 O were dissolved in ethylene glycol under the aid of sonication, and then 0.71 g of thioacetamide (TAA) was added to the solution and stirred for another 30 min. The mixture solution was poured into a 100 mL Teflon-lined stainless-steel autoclave and maintained at 160 • C for 18 h. The solids were centrifuged and washed with DI water and ethanol several times and dried at 80 • C for 5 h in air to obtain ZnO/Bi 2 S 3 composite photocatalysts. By changing the amount of ZnO used in the procedure, three ZnO/Bi 2 S 3 composites with 3%, 5% and 10% mass ratios of ZnO to Bi 2 S 3 were synthesized, which were named as 3%-ZnO/Bi 2 S 3 , 5%-ZnO/Bi 2 S 3 and 10%-ZnO/Bi 2 S 3 , respectively. For comparison, pure Bi 2 S 3 was prepared with the same procedure without the use of ZnO.

Characterization
XRD was performed on a PANalytical X'pert Pro powder diffractometer (PANalytical, Almelo, The Netherlands) using Cu-Ka radiation (λ = 1.5418 Å) with a scan step of 0.013 • . SEM images were generated with a Hitachi SU8020 (Hitachi Ltd., Tokyo, Japan) scanning electron microscope with an acceleration voltage of 20KV. HRTEM was performed on a FEI Tecnai G2 F20 field-emission transmission electron microscopy (FEI Inc." Hillsboro, OR, USA) at an acceleration voltage of 200 KV. XPS measurements were performed on a Thermo Fisher ESCALAB 250Xi photoelectron spectrometer (Thermo Fisher Scientific Inc., Waltham, MA, USA) using a monochromatic Al KαX-ray source (hν = 1486.6 eV). DRS spectrum was recorded on a Hitachi U-4100 UV-vis spectrophotometer (Hitachi, Tokyo, Japan) using BaSO 4 as the reference sample. Photoelectrochemical measurements were conducted in a three electrodes quartz cell with 0.1 M Na 2 SO 4 solution on the electrochemical system (CHI-760E, Shanghai Chenhua Instruments, Shanghai, China). PL spectra were recorded in a Hitachi F-7000 fluorescence spectrophotometer (Hitachi, Tokyo, Japan).

Photocatalytic Tests
The photocatalytic Cr(VI) reduction performance of ZnO/Bi 2 S 3 samples under visible light was conducted by a 500 W Xe lamp (BILON-CHX-V, Shanghai photoreactor, BiLon, Shanghai, China) with maximum wavelength emission at 470 nm. The lamp was placed in a trap with running water circulating through a jacket in order to maintain a constant temperature in the reaction system, and the distance between the light source and the tube containing the reaction mixture was set to be 15 cm. Stock solution (100 mg/L) of Cr(VI) was prepared by dissolving K 2 Cr 2 O 7 (analytical grade, Kelong Reagent Factory, Chengdu, China) into distilled water. In a typical procedure, 0.05 g of the photocatalyst was added to 50 mL of Cr(VI) solution (20 mg/L), and the suspension was stirred in the dark for 60 min to reach the adsorption-desorption equilibrium. Then, the solution was exposed to light irradiation under magnetic stirring. During the irradiation, 4 mL of the reaction solution was withdrawn at certain time intervals and centrifuged to remove the photocatalyst from the solution. The Cr(VI) concentration in the supernatant solution was determined spectrophotometrically at 540 nm using the diphenylcarbazide method [46] by UV-1000 spectrophotometer (AOE, Shanghai, China).
The initial solution pH was adjusted by 0.01 mg/L aqueous HCl or 0.01 mg/L NaOH solutions. For comparison, the blank experiments to test the Cr(VI) stability under irradiation without any photocatalysts and the dark experiments to test the physical adsorption capacity of the photocatalysts were conducted with no irradiation under the identical conditions. For the radical species trapping tests, Benzoquinone (BQ), Iso-propyl alcohol (IPA) and Citric Acid (CA) were introduced into Cr(VI) solution prior to the addition of the catalyst as the scavengers of ·O 2 − ,·OH and h + , respectively [47].

Conclusions
Briefly, ZnO/Bi 2 S 3 nanocomposites were prepared via a facile solvothermal method. Detailed characterization showed the introduction of ZnO played a key role in the formation of the heterojunction structure of sandwiched-like ZnO/Bi 2 S 3 , where some tiny ZnO nanoparticles were randomly embedded between Bi 2 S 3 nanoflakes. The enhanced photocatalytic reduction activity towards aqueous Cr(VI) revealed that there was a synergistic effect between the two components of the composites. The mechanism strongly suggested that the enhanced visible-light-driven photoreduction performance of ZnO/Bi 2 S 3 photocatalyst was ascribed to the increased light harvesting and the effective separation and transfer of the photogenerated charge carriers across the heterojunction interface of the ZnO/Bi 2 S 3 composite. The photocatalysts prepared herein in the study were evaluated by photodegrading other wastewater pollutants and had great potential in environmental remediation.

Conflicts of Interest:
The authors declare no conflict of interest.