Glyphosate (N-(phosphonomethyl)glycine) is a broad- spectrum herbicide. In March 2015, the World Health Organization's International Agency for Research on Cancer announced that glyphosate is probably carcinogenic to humans (category 2A) [1]. At present, glyphosate wastewater is generally treated using biochemical or Fenton methods [2]; however, the significant length of the corresponding biodegradation period may create secondary pollution. Since Fujishima et al. [3] reported the decomposition of H₂O by TiO₂ electrolysis in 1972, the treatment of industrial wastewater by semiconductor photocatalysis has been widely studied [4]. The TiO₂ photocatalyst used in the photocatalytic degradation of glyphosate wastewater [4-6] only absorbs ultraviolet (UV) light; however, the solar spectrum includes a very small percentage of UV light (~ 5%). Therefore, extensive research has focused on developing visible light-active photocatalysts [7] for the degradation of wastewater containing organic pollutants [8, 9] such as glyphosate.

Bismuth vanadate (BiVO₄) is an attractive candidate due to its high stability against photocorrosion, non-toxicity, and narrow band gap [10, 11]. The band gap of monoclinic BiVO₄ is about 2.4 eV, which makes it responsive to visible light. However, pure BiVO₄ exhibits poor photocatalytic activity due to its slow charge transport and rapid recombination of photogenerated charge carriers [12]. The photocatalytic activity of a photocatalyst is commonly associated with its exposed surfaces, and in particular with the arrangement and coordination of the surface atoms on the different crystal facets. Therefore, there is great interest in the design of high-performance BiVO₄ materials with different dominant exposed facets. BiVO₄ with orientational growth along the (040) direction was shown to promote charge separation and high transfer efficiency, resulting in a marked enhancement in the photocatalytic activity of BiVO₄ [13, 14]. Chen et al. [15] showed that lamellar m-BiVO₄ samples oriented along the (040) facets, with the highest ratio between the intensities of the (040) and (121) diffraction peaks (I₀₄₀/I₁₂₁), possessed the best photocatalytic activity. Ou et al. [16] reported that the Z-scheme g-C₃N₄@Ag/BiVO₄(040) photocatalyst exhibited high photocatalytic oxidation performance, due to the exposure of photogenerated electron-rich (040) facets and the high separation rate of photogenerated charge carriers. In addition, the formation of a heterojunction between BiVO₄ and other semiconductors (such as Ag₂S [17], Bi₂WO₆ [18], g-C₃N₄ [19, 20], Ag₂CO₃ [21], BiFeWO₆ [22], or BiOI [23]) has been extensively investigated to enhance the photocatalytic activity of BiVO₄.

Metal sulfides are attracting increasing attention due to their narrow band gaps. Most metal sulfides have been synthesized by the reaction between various metal cations and inorganic sulfur sources [24, 25]. At present, the biosynthesis of nanomaterials with unique structure and excellent self-assembly properties as templates or auxiliary agents has become a key step in the preparation of materials [26]. The -SH group of the L-cysteine molecule can be used as a sulfur source to synthesize sulfides. In addition, as an amino acid, L-cysteine can form peptides by condensation reactions, and then act as a template to synthesize special morphologies [27]. Using the -SH group of L-cysteine and the strong self-assembly ability to synthesize the special structure of metal sulfides has attracted great interest and widespread attention.

Bi₂S₃ is a semiconductor with a narrow band gap of 1.19–1.5 eV; hence, it can absorb in the whole visible region [28-35]. Considering that the band gap of Bi₂S₃ is smaller than that of BiVO₄ (E_g = 2.4 eV), and that the valence band (VB) potential (E_VB) of BiVO₄ (E_VB = 2.71 eV vs. normal hydrogen electrode, NHE) is more positive than that of Bi₂S₃ (E_VB = 1.42 eV vs. NHE) [28], the energy levels of Bi₂S₃ and BiVO₄ appear to be well matched, so that a heterojunction enabling the efficient separation of photogenerated charge carriers could be formed. The formation of a heterojunction structure between BiVO₄ and Bi₂S₃ has been achieved using Na₂S [29, 30], thioacetamide [31, 32], and L-cysteine [33] as the sulfur source, resulting in enhanced photocatalytic activity due to the retarded recombination of photogenerated carriers.

In this work, BiVO₄ was coupled with Bi₂S₃ to form a peony-like heterojunction, which exhibited better photocatalytic activity than BiVO₄. We report the synthesis and characterization of BiVO₄ and of its composite with Bi₂S₃ (Bi₂S₃/BiVO₄). The Bi₂S₃/BiVO₄ composites were prepared by a one-pot hydrothermal method using ethylenediaminetetraacetic acid (EDTA) disodium salt as a directing agent, to grow BiVO₄ sheets oriented along (040) facets, as well as L-cysteine acting as sulfur source and soft template at the same time. The flakes of BiVO₄ and Bi₂S₃ formed Bi₂S₃/BiVO₄ structures with different morphologies, which showed an enhanced photocatalytic activity for glyphosate degradation compared to that of BiVO₄.

Bi₂S₃/BiVO₄ heterojunctions were prepared by a one-pot hydrothermal method. Typically, 0.0100 mol of Bi(NO₃)₃·5H₂O was dissolved in 5 mL of concentrated nitric acid, and deionized water was added to 20 mL of the mixture under stirring for 10 min to obtain solution A. Then, 0.0100 mol of NH₄VO₃ and 0.0068 mol of EDTA were dissolved in 20 mL of a NaOH solution (4 mol/L) with stirring for 10 min, to obtain solution B. Solution B was then added dropwise to solution A under stirring, and the pH of the resulting solution was adjusted to 5 using 2 mol/L NaOH. Afterward, a specific amount of L-cysteine (0, 0.25, 0.5, 1, 2.5, and 5 mmol) was added to the above mixture and the solution was stirred for 30 min. A 40 mL aliquot of the resulting yellow suspension was then transferred into a 50-mL Teflon-lined autoclave and heated at 180 ℃ for 24 h. The precipitates were filtered and washed with water and ethanol to remove impurities, and then dried at 65 ℃ overnight. The corresponding products were named BiVO₄, 0.25Bi₂S₃/BiVO₄, 0.5Bi₂S₃/BiVO₄, 1Bi₂S₃/BiVO₄, 2.5Bi₂S₃/BiVO₄, and 5Bi₂S₃/BiVO₄, respectively.

Powder X-ray diffraction (XRD) measurements were performed on a MSALXD-2 diffractometer with Cu K_α radiation (λ = 0.15406 nm), employing a scanning rate of 0.02°/s in the 2θ range from 10° to 80°. The microstructures of the samples were studied by field emission scanning electron microscopy (FESEM, FEI-Verios 460), transmission electron microscopy (TEM, FEI-Tecnai 12), and high-resolution transmission electron microscopy (HRTEM, JEM2100). Energy-dispersive X-ray (EDS) mappings of the as-prepared samples were obtained using a FEI-Verios 460 microscope. Diffuse reflectance spectroscopy (DRS) measurements were performed using a UV-vis spectrophotometer (UV-2550, Shimadzu), with BaSO₄ as a reflectance standard. Transient photocurrent and electrochemical impedance spectroscopy (EIS) measurements were carried out in a Na₂SO₄ solution (0.5 mol/L) using an electrochemical workstation (Epsilon-BAS). A 300-W xenon lamp equipped with a UV cut-off filter (λ > 420 nm) was used as visible light source. A film of the as-prepared sample on fluorine-doped tin oxide (FTO) served as the working electrode, while a Pt plate and Ag/AgCl (saturated KCl) were used as the counter and reference electrodes, respectively. Electron spin resonance (ESR) experiments were performed on a Bruker A300 spectrometer at room temperature.

The photocatalytic degradation of glyphosate was investigated under visible light irradiation (λ > 400 nm), using a 125-W high-pressure mercury lamp with 180 mL of 2 mol/L NaNO₂ solution as the filter liquor. The photocatalytic reaction system and the spectral energy distribution used in the experiments were reported elsewhere [20]. In a typical experiment, 0.2 g of the as-prepared photocatalyst was added to 500 mL of 0.1 mmol/L glyphosate aqueous solution under ultrasonic stirring for 1 min. The suspension was then stirred in the dark for 30 min to establish the adsorption-desorption equilibrium between the photocatalyst particles and the glyphosate molecules. At a given time after switching on the light source, 8 mL of the reaction solution was collected and centrifuged (3500 rpm, 10 min) to remove the photocatalyst particles. A 5-mL aliquot of supernatant was collected to analyze the final concentration of the PO₄^3- oxidation product by Mo-Sb-ascorbic acid colorimetry [36, 37]. The relationship between PO₄^3- concentration (x) and absorbance (A) follows the relation x = 63.093A + 0.0322, where the x units are 10^-6 mol/L. The percentage of glyphosate degradation (η) was calculated using the following equation:

(1)

where C₀ is the concentration of total organic phosphorus in the reaction solution before irradiation and C_t is the concentration of PO₄^3- produced at the irradiation time t.

To investigate the photocatalytic reaction mechanism over the Bi₂S₃/BiVO₄ composites, K₂Cr₂O₇, EDTA, isopropyl alcohol (IPA), and benzoquinone (BQ) were added in the photocatalytic degradation solution of glyphosate as active scavengers for photogenerated electrons (e^-), photogenerated holes (h⁺), hydroxyl radicals (·OH^-), and superoxide radicals (·O₂^-), respectively, under visible light irradiation for 3 h. Before irradiation, 0.2 g of photocatalyst was added to the glyphosate solution (500 ml, 0.1 mmol/L) under ultrasonic stirring for 1 min, and then the scavenger (1 mmol/L) was added. After stirring in the dark for 30 min to establish the adsorption-desorption equilibrium, the light source was turned on. After a specific time interval (30 min), 8 mL of suspension was collected and centrifuged (3500 rpm, 10 min) to remove the photocatalyst particles. An aliquot of supernatant (5 mL) was collected to analyze the final concentration of the PO₄^3- oxidation product via Mo-Sb-ascorbic acid colorimetry. In the ESR experiments, the ·OH species were detected with dimethylpyridine N-oxide (DMPO). Typically, 0.01 g of photocatalyst was completely dispersed in 25 mL of 0.1 mmol/L glyphosate aqueous solution under stirring in the dark for 5 min. Then, 100 μL of the reaction solution was removed and irradiated by a 300-W Xe lamp (equipped with a UV cut-off filter, λ > 420 nm) for 5 min, followed by injection of 10 μL of DMPO. Capillary tubes were inserted in the ESR cavity, and spectra were recorded at 10 and 20 min during irradiation. For comparison, control experiments without photocatalyst or without light irradiation were also performed.

Fig. 1 shows the XRD patterns of pure BiVO₄ and of Bi₂S₃/BiVO₄ samples with different Bi₂S₃ contents; all diffraction peaks of the pure BiVO₄ sample can be indexed to monoclinic scheelite BiVO₄ (JCPDS No. 14-0688) [38]. It is worth noting that the relative intensity of the (040) diffraction peak of pure BiVO₄ is stronger than that of the Bi₂S₃/BiVO₄ composites, suggesting the preferential crystal growth of sheet-like BiVO₄ oriented along (040) facets using EDTA as a directing agent. When the Bi₂S₃ content is in the range of 0.25 to 1 mmol, no obvious impurity peaks are detected. However, the diffraction intensity of the (040) peak of BiVO₄ decreases and the (121) diffraction intensity gradually increases with increasing Bi₂S₃ content. This suggests that the crystal growth of sheet-like BiVO₄ oriented along the (040) facets is inhibited by the formation of Bi₂S₃ on the surface of the BiVO₄ particles. When the Bi₂S₃ content is equal or greater than 2.5 mmol, the weak intensity and broadening of the diffraction peaks suggests a low crystallinity of the 2.5Bi₂S₃/BiVO₄ sample. In the case of the 5Bi₂S₃/BiVO₄ sample, the characteristic diffraction peaks of Bi₂S₃ are identified at 2θ = 22.6°, 25.4°, 28.8°, 32.0°, 33.2°, and 35.5° (JCPDS No. 17-0320) [34], matching well with the reflections of the (220), (130), (211), (221), (301), and (240) planes, respectively, which indicates that the as-prepared samples are composed of Bi₂S₃ and BiVO₄.

The morphology, particle size, and microstructure of the BiVO₄ and Bi₂S₃/BiVO₄ samples were analyzed by FESEM, TEM, and HRTEM, and the results are shown in Fig. 2. The morphology of the Bi₂S₃/BiVO₄ samples shows significant variations when the Bi₂S₃ content is increased from 0 to 0.25, 0.5, 1, 2.5, and 5 mmol, while keeping all other reaction parameters fixed. Fig. 2(a) shows the SEM image of the as-prepared BiVO₄ sheets. With a Bi₂S₃ content of 0.25 mmol, a fraction of the flake-like Bi₂S₃/BiVO₄ nanoparticles grows directly into peony-like particles, but some of them exhibit disk-like aggregates (Fig. 2(b)). With increasing Bi₂S₃ content, most of the 0.5Bi₂S₃/BiVO₄ nanoparticles grow into peony-like aggregates (Fig. 2(c)), and the 1Bi₂S₃/BiVO₄ sample is almost entirely made of peony-like aggregates of flake-like nanoparticles (Fig. 2(d)). Upon increasing the Bi₂S₃ content (≥ 2.5 mmol), the average particle size of the 2.5Bi₂S₃/BiVO₄ and 5Bi₂S₃/BiVO₄ samples decreases drastically, but their peony-like morphology is maintained (Fig. 2(e) and (f)). However, some fragments are observed on the surface of the 2.5Bi₂S₃/BiVO₄ and 5Bi₂S₃/BiVO₄ samples, while a fraction of the peony-like Bi₂S₃/BiVO₄ particles aggregate. These results indicate that the Bi₂S₃ content affects the morphology of the Bi₂S₃/BiVO₄ samples, with 1Bi₂S₃/BiVO₄ showing a more regular peony-like morphology. The EDS elemental mappings of 1Bi₂S₃/BiVO₄ show that all elements (Bi, V, O, and S) are uniformly distributed throughout the scan area of the Bi₂S₃/BiVO₄ heterojunction (Fig. 2(g)). This suggests the presence of Bi₂S₃ and BiVO₄ in the composites, resulting in the formation of the Bi₂S₃/BiVO₄ heterojunction.

Consistent with the FESEM images, the TEM images show that BiVO₄ and 1Bi₂S₃/BiVO₄ exhibit a sheet-like (Fig. 2(h)) and peony-like (Fig. 2(i)) morphology, respectively. The HRTEM image of 1Bi₂S₃/BiVO₄ (Fig. 2(j)) further confirms the close interfacial contact and continuity of lattice fringes between BiVO₄ and Bi₂S₃. The fringe spacings of 0.398 and 0.272 nm correspond to the (220) and (301) planes of Bi₂S₃ (JCPDS No. 17-0320) [34], respectively, while the fringe spacing of 0.309 nm is consistent with the (121) plane of BiVO₄ (JCPDS No. 14-0688) [38]. In the hydrothermal reaction, the strong EDTA ligand not only forms a stable complex with Bi³⁺, but also acts as the directing agent, orienting pure BiVO₄ along the (121) facet to form well-defined 2D sheets. L-Cysteine was employed as both sulfur source and soft template. As a sulfur source, L-cysteine has many functional groups (-NH₂, -COOH, and -SH), and the -SH group has a strong tendency to coordinate Bi³⁺. The Bi₂S₃/BiVO₄ heterojunction was prepared from an ion-exchange reaction between -SH and BiVO₄ [33]. Moreover, as a soft template, the -NH₂ group can react with the -COOH group of another L-cysteine molecule to form a dipeptide or polypeptide, which preferentially grows along the (040) direction and forms Bi₂S₃ under hydrothermal treatment for 24 h [24]. Therefore, Bi₂S₃/BiVO₄ does not form a sheet-like morphology such as that of pure BiVO₄ after the addition of L-cysteine, but exhibits a peony-like morphology composed of flake-like nanoparticles. However, if the Bi₂S₃ content reaches or exceeds 2.5 mmol, the growth of BiVO₄ along the (040) direction is largely suppressed due to the formation of Bi₂S₃. A possible formation mechanism of the Bi₂S₃/BiVO₄ composites proposed on the basis of the experimental results discussed above is shown in Fig. 3.

The optical absorption properties of pure BiVO₄ and of Bi₂S₃/BiVO₄ samples with different Bi₂S₃ contents were measured using a UV-vis diffuse reflectance spectrometer, and the results are shown in Fig. 4. Fig. 4(a) shows that the absorption edge of BiVO₄ is located at about 510 nm, within the visible light region. Compared to BiVO₄, the optical absorption of the Bi₂S₃/BiVO₄ samples in the visible region increases because Bi₂S₃ can harvest more photons due to its small band gap, resulting in a strong absorption in the visible region [24]. The band gap (E_g) of a semiconductor can be estimated by the following equation:

(2)

where ν is the frequency of the light, h is Planck's constant, α is the absorption coefficient, and A is a constant; moreover, n is a constant depending on the transition type of the semiconductor, and takes the values of 1 and 4 for direct and indirect transitions, respectively [39]. The band gap of the as-prepared samples was determined from a plot of (αhν)² vs. the photon energy (hν) (Fig. 4b) [40]. The bandgap values of BiVO₄ and Bi₂S₃ were calculated to be 2.42 and 1.27 eV, respectively. These results are in good agreement with previous reports [15, 41]. The visible light absorption capacity of BiVO₄ is greatly enhanced after coupling with Bi₂S₃; hence, the Bi₂S₃/BiVO₄ samples are expected to show high photocatalytic activity under visible light irradiation.

To investigate the effects of the Bi₂S₃ content on the photocatalytic activity of BiVO₄, we tested the degradation of a glyphosate solution by P25, BiVO₄, and Bi₂S₃/BiVO₄ samples with different Bi₂S₃ contents under visible light irradiation. As shown in Fig. 5(a), pure BiVO₄ shows a degradation percentage of 36%, while P25 TiO₂ exhibits a lower degradation rate of 24%. The degradation rate increases with increasing Bi₂S₃ content; in particular, the 1Bi₂S₃/BiVO₄ sample shows the highest photocatalytic activity, with a degradation rate of 79% after 180 min. The enhanced photocatalytic activity of Bi₂S₃/BiVO₄ can be attributed to the formation of the Bi₂S₃/BiVO₄ heterostructure with exposed (040) facets, which results in high light absorption efficiency and efficient separation of electron-hole pairs [15, 42, 43]. However, when the Bi₂S₃ content reaches or exceeds 2.5 mmol, the glyphosate degradation rate decreases and becomes lower than that of pure BiVO₄. This might be attributed to the decrease of the contact interface (that reduces the separation efficiency of electrons and holes), the poor charge transport in Bi₂S₃ [30], and the low-lying valence band of Bi₂S₃, which cannot drive the photocatalytic oxidation reaction. This result indicates that an optimum amount of Bi₂S₃ is required to achieve the best photocatalytic performance in the Bi₂S₃/BiVO₄ heterostructures. Moreover, cycling runs were carried out for the 1Bi₂S₃/BiVO₄ sample because of its importance in practical applications. Fig. 5(b) shows that the degradation rates of glyphosate in the second and third run are 66% and 59%, corresponding to 84% and 75% photocatalytic activity retention, respectively. This suggests that 1Bi₂S₃/BiVO₄ exhibits relative stability and reusability during the photocatalytic reaction process. In addition, the corresponding XRD pattern (Fig. 5(c)) reveals that there are no changes in the crystal structure of 1Bi₂S₃/BiVO₄ after three cycles, further confirming the relative stability of the photocatalyst.

In order to confirm the possible mechanism of the enhanced photocatalytic performance under visible light irradiation, we investigated the exposure of the (040) crystal facets and carried out transient photocurrent tests as well as EIS measurements.

The effect of the Bi₂S₃ content on the ratio of the intensities of the (040) and (121) diffraction peaks (I₀₄₀/I₁₂₁) and on the photocatalytic degradation rate of glyphosate over the Bi₂S₃/BiVO₄ samples is illustrated in Table 1. The amount of Bi₂S₃ has no effect on the crystal phase of BiVO₄, because all Bi₂S₃/BiVO₄ samples are monoclinic (Fig. 1). However, the Bi₂S₃ content has a major influence on the morphology of the Bi₂S₃/BiVO₄ samples (Fig. 2), the I₀₄₀/I₁₂₁ ratio, and the photocatalytic degradation rate of glyphosate (Table 1). It is worth noting that a gradual decrease in relative intensity and sharpness is observed for the (040) diffraction peak with the increase of the Bi₂S₃ content, and the I₀₄₀/I₁₂₁ ratio decreases with increasing Bi₂S₃ content. This suggests that the growth along the (040) direction stopped due to the coupling with Bi₂S₃. Because the VB and conduction band (CB) energy levels of the (040) and (110) facets are different, the photoinduced electrons and holes, which transfer to the (040) and (110) facet, respectively, can be separated [44]. In addition, the (040) facets on the surface of monoclinic BiVO₄ are multi-electron transfer centers and serve as active sites for the photocatalytic reaction [16]. Chen et al. [15] showed that BiVO₄ samples with a greater I₀₄₀/I₁₂₁ ratio exhibit higher photocatalytic activities, because BiVO₄ sheets with a shorter time and migration distance for photogenerated carriers to reach the surface lead to a more efficient separation and transport of photogenerated electrons and holes. However, in this study the I₀₄₀/I₁₂₁ ratio is not consistent with the measured degradation rates of glyphosate over the Bi₂S₃/BiVO₄ samples, because of the coupling with Bi₂S₃.

In order to evaluate the photogenerated charge separation efficiency and charge transport properties under visible light irradiation (λ > 420 nm), transient photocurrent tests and EIS measurements were carried out, and the results are shown in Fig. 6. The transient photocurrent response measurements were employed to compare the photoinduced charge separation efficiency of pure BiVO₄ and Bi₂S₃/BiVO₄ samples with different Bi₂S₃ contents. As shown in Fig. 6(a), the BiVO₄ and Bi₂S₃/BiVO₄ samples exhibit a stable photocurrent response for several on-off cycles under visible light irradiation. The photocurrent is generated immediately after the light source is switched on, due to the separation of electron-hole pairs at the electrode/electrolyte interface [20, 37, 45]. Once the light source is turned off, the photocurrent rapidly decreases to zero. With the exception of 2.5Bi₂S₃/BiVO₄, the Bi₂S₃/BiVO₄ samples exhibit a stronger transient photocurrent than BiVO₄, indicating an improved interfacial charge transfer due to the formation of the Bi₂S₃/BiVO₄ heterostructure, thus inhibiting the recombination of photogenerated charges and extending the lifetime of the photoinduced electrons and holes. The 1Bi₂S₃/BiVO₄ sample shows the highest transient photocurrent density among all photocatalysts, resulting in the best photocatalytic activity for glyphosate degradation.

Because the charge transport properties at the electrode/electrolyte interface play an important role in the photocatalytic activity, EIS measurements were performed under visible light irradiation. The Nyquist plots of pure BiVO₄ and of Bi₂S₃/BiVO₄ samples with different Bi₂S₃ contents, reflecting the charge-transfer resistance at the electrode/electrolyte interface, are shown in Fig. 6(b). The radius of the arc is related to the charge-transfer resistance occurring on the surface of the electrode, with a smaller arc radius denoting a higher charge-transfer efficiency [45]. Fig. 6(b) shows that the arc radius of the 1Bi₂S₃/BiVO₄ electrode is the smallest among all photocatalyst electrodes, confirming that the fastest charge transport takes place in the 1Bi₂S₃/BiVO₄ electrode. Furthermore, with the exception of 2.5Bi₂S₃/BiVO₄, the Bi₂S₃/BiVO₄ samples exhibit a smaller arc radius than that of the BiVO₄ electrode. These findings are similar to the transient photocurrent results. As mentioned above, a more effective charge separation and faster interfacial charge transfer occurs in 1Bi₂S₃/BiVO₄, greatly contributing to its enhanced photocatalytic performance.

In addition, active species trapping experiments and ESR measurements were performed to identify the reactive species in the photocatalytic process. Fig. 7 shows the photocatalytic degradation of glyphosate over 1Bi₂S₃/BiVO₄ in the absence and presence of scavengers (IPA, BQ, EDTA, and K₂Cr₂O₇) under visible light irradiation for 3 h. When charge carrier scavengers (EDTA for h⁺ or K₂Cr₂O₇ for e^-) were added in the reaction system [46, 47], the photocatalytic degradation of glyphosate was significantly inhibited. These results suggest that photogenerated electrons and holes are the main active species involved in the photocatalytic reaction process. Moreover, the addition of radical scavengers (IPA for ·OH^- and BQ for ·O₂^- [48-51]) resulted in a decrease in the photocatalytic reaction rate. These results indicate that the photocatalytic oxidation of glyphosate over 1Bi₂S₃/BiVO₄ is mainly driven by photogenerated electrons and holes. The presence of ·OH^- was further confirmed by the ESR analysis, as shown in Fig. 8. In comparison, no ESR signals were detected when the reaction was carried out in the dark. Similarly, no ESR signal was observed when no photocatalyst was added. These findings suggest that no ·OH radicals were formed. However, the four characteristic peaks of DMPO-·OH adducts (in 1:2:2:1 intensity ratio, consistent with previous studies of ·OH adducts [52-54]) appeared when the BiVO₄ photocatalyst was added. This suggests the formation of strong and nonselective ·OH radicals. 1Bi₂S₃/BiVO₄ exhibited stronger signals due to the higher concentration of produced ·OH radicals compared to BiVO₄, indicating that a larger amount of ·OH radicals is formed in the presence of 1Bi₂S₃/BiVO₄ than BiVO₄. Furthermore, the intensity of the DMPO-·OH signals increased with increasing irradiation time.

To determine the origin of the high photocatalytic activity of the Bi₂S₃/BiVO₄ samples, we investigated the energy level alignment of the photocatalysts. The conduction (E_CB) and valence band potentials (E_VB) of Bi₂S₃ and BiVO₄ were calculated empirically using the following equations:

(3)

(4)

where E_g is the band gap of the semiconductor, E_e is the energy of free electrons on the hydrogen scale (4.5 eV), and χ is the absolute electronegativity of the semiconductor. The χ value of a compound semiconductor is calculated as the geometric mean of the absolute electronegativity of the constituent atoms. Taking the absolute electronegativity values for Bi, V, S, and O as 4.69, 3.6, 6.22, and 7.54, respectively, the calculated χ values of Bi₂S₃ and BiVO₄ are 5.56 and 6.16 eV, respectively. The calculated VB potential of BiVO₄ (E_VB = 2.87 eV vs. NHE) is more positive than that of Bi₂S₃ (E_VB = 1.69 eV vs. NHE), while the calculated CB potential of Bi₂S₃ (E_VB = 0.42 eV vs. NHE) is more negative than that of BiVO₄ (E_VB = 0.45 eV vs. NHE) [29, 32]. Therefore, the energy levels of Bi₂S₃ and BiVO₄ are well matched and the Bi₂S₃/BiVO₄ heterojunction could be formed, favoring the separation of photogenerated charge carriers. On the basis of the above results, a possible mechanism for the enhanced photocatalytic activity of Bi₂S₃/BiVO₄ is schematically illustrated in Fig. 9. Upon irradiation with visible light, the electrons of both BiVO₄ and Bi₂S₃ are promoted from the VB into the CB, leaving holes behind. As the visible light absorption capacity of Bi₂S₃/BiVO₄ is better than that of BiVO₄, a higher number of electron-hole pairs are generated over Bi₂S₃/BiVO₄. The photogenerated electrons in the CB of Bi₂S₃ transfer to BiVO₄ [29], then react with adsorbed O₂ to yield ·O₂^-, and subsequent redox reactions may occur. The photogenerated holes formed in the VB of BiVO₄ react with the adsorbed H₂O/OH^- ions to produce ·OH or directly oxidize glyphosate. Therefore, the photogenerated electrons and holes can be efficiently separated at the interfaces of the Bi₂S₃/BiVO₄ heterostructures. Eventually, the ·O₂^- and ·OH radicals also drive the glyphosate oxidation reaction to form PO₄^3- and other ions; in this work, the PO₄^3- concentration was determined to assess the photocatalytic degradation percentage of glyphosate. As a result, the photocatalytic activity of the Bi₂S₃/BiVO₄ heterostructures was greatly enhanced due to the formation of the heterostructure and the improved absorption of visible light.

Bi₂S₃/BiVO₄ (040) heterojunctions with various morphologies were successfully synthesized by a one-pot hydrothermal method using L-cysteine as sulfur source/soft template and EDTA as directing agent. BiVO₄ showed an enhanced visible light absorption capacity after coupling with Bi₂S₃, owing to the formation of a Bi₂S₃/BiVO₄ heterojunction. The coupling with Bi₂S₃ had a significant influence on the growth of (040) and (121) facets, as well as on the morphology of the Bi₂S₃/BiVO₄ samples. When the amount of added L-cysteine reached 1 mmol, 1Bi₂S₃/BiVO₄ assumed a peony-like morphology. Photoelectrochemical performance tests highlighted the enhanced charge separation and efficient transport of photogenerated charge carriers in Bi₂S₃/BiVO₄. The enhanced photocatalytic activity of Bi₂S₃/BiVO₄ could be attributed to the enhanced visible light absorption and the formation of the heterojunction structure. In addition, Bi₂S₃/BiVO₄ exhibited relative stability and reusability. The simple method presented in this work could be used to fabricate composite photocatalysts with high activity for different applications, such as photocatalytic degradation of organic pollutants, photocatalytic splitting of water, and photocatalytic reduction of carbon dioxide.