Solvothermal synthesis of facet-dependent BiVO4 photocatalyst with enhanced visible-light-driven photocatalytic degradation of organic pollutant: assessment of toxicity by zebrafish embryo

The BiVO4 photocatalyst plays a very important role in photocatalytic reactions attributed to its unique crystalline structure, size, morphology and surface area. Herein, we report a facet-dependent monoclinic scheelite BiVO4 (m-BiVO4) photocatalyst with uniform truncated square (18 sided) hexagonal bipyramidal shape synthesized by a template-free and surfactant-free solvothermal method using ethylene glycol solvent under cost-effective and mild reactions. The structural, morphological and optical properties of the m-BiVO4 photocatalyst are widely characterized. The photocatalytic activity of the m-BiVO4 photocatalyst is tested towards 20 ppm methylene blue (MB) dye aqueous solution as a pollutant model under visible light irradiation. Enhanced visible-light driven photoactivity with dye degradation efficiency of approx. 91% at a rate of 0.388 × 10−2 min−1 is obtained, presumably due to the presence of high-active (040) facets. Zebrafish embryo toxicity test of treated MB dye solution reveals the degradation and toxicity reduction of the MB dye. Moreover, the recycling experiment validates that the m-BiVO4 photocatalyst has a great structural stability with reliable performance. This work may provide a lucid and expedient strategy to synthesize highly crystalline (040) facet-dependent semiconductor photocatalyst toward dye degradation and obviously industrial wastewater remediation.

. Raman spectroscopy is more sensitive to provide structural information, degree of crystallinity, defects and disorders, particle size and electronic properties of nanomaterials. The Raman spectrum of m-BiVO 4 (Fig. 1b) excited by a red laser (632.8 nm) delineates the translational (Ex t ) and rotational (Ex r ) phonon vibration bands at 136 and 213 cm -1 , respectively. The peaks at 328 and 367 cm −1 are attributed to the typical antisymmetric (δ as ) and symmetric (δ s ) bending modes of the vanadate tetrahedral anion. The weak peak at 709 cm −1 and dominated peak at 826 cm −1 are assigned to the antisymmetric (ν as ) and symmetric (ν s ) V-O stretching vibrations in vanadate tetrahedral anion in monoclinic scheelite BiVO 4 38 . The UV-Vis absorption spectrum of m-BiVO 4 (Fig. 2a) exhibits a strong optical absorption in 420-800 nm wavelength range, indicating that m-BiVO 4 can efficiently absorb visible light and acts as solar light driven active photocatalyst for dye degradation. It is well known that semiconducting electronic structure usually plays a critical role in its photocatalytic activity. Usually, the valence band (VB) and the conduction band (CB) of BiVO 4 is poised by hybridized Bi 6s/O 2p orbitals and V 3d orbitals, respectively 30 . The band gap and absorption coefficient according to the Kubelka-Munk equation can be expressed as αhv = A(hv − Eg) 1/2 , where α, v and h represents the absorption coefficient, frequency of the light and Planck's constant, respectively. The band gap energy value can be estimated from the Tauc plot ( Fig. 2b) of (αhv) 2 versus photon energy (hv) curve. The intercept of the tangent to the X-axis is a good estimate of the band gap E g and matches with previous literature 31 . The estimated band gap E g of m-BiVO 4 is 2.5 eV, confirming its capability as visible light driven (VLD) photocatalyst.
The X-ray photoelectron spectroscopy (XPS) is used to determine the surface chemical composition of the m-BiVO 4 . The XPS is capable of differentiating the spin-orbit splitting of metal ions at two possible states, i.e. having different binding energies, and provides metal speciation information. The survey XPS spectrum of the BiVO 4 (Fig. 3a) reveals the presence of C, Bi, V and O elements in the m-BiVO 4 . The high resolution XPS  www.nature.com/scientificreports/ spectrum of the Bi4f spectra ( Fig. 3b) exhibits doublets at 158.14 and 164.14 eV that correspond to the Bi4f 7/2 and Bi4f 5/2 lines, respectively. The 6.0 eV difference between the two bonding energy suggests that Bi is in a + 3 oxidation state 39 . Whereas, the split peaks of V2p at 515.5 eV and 523.5 eV (Fig. 3c) correspond to V2p 3/2 and V2p 1/2 , respectively, suggesting the presence of V + 5 oxidation state 40 in the m-BiVO 4 . The split peaks of O1s at 528.9 eV and 531.9 eV (Fig. 3d) represent an asymmetric behaviour of the oxygen states that can be consigned to the lattice oxygen and surface hydroxyl groups in crystalline m-BiVO 4 , respectively. The field emission scanning electron microscopy (FESEM) is used to examine the morphology and elemental composition of the m-BiVO 4 under different magnification. A broad view of the m-BiVO 4 batch (Fig. 4a, 10 µm scale bar) reveals different size and shape m-BiVO 4 nanoparticles (NPs). A more detailed examination of several neighbouring m-BiVO 4 NPs (Fig. 4b, 1 µm scale bar) shows many hexagonal and few sphere NPs. A closer inspection of a single hexagonal m-BiVO 4 NP (inset in Fig. 4b, 200 nm scale bar) finds the appearance of non-planar surface. Higher magnifications (Fig. 4c, 500 nm scale bar) confirm its smooth surface structures with mostly truncated square (18 sided) hexagonal bipyramidal shape with exposed (040) facets (Fig. 4d, 200 nm scale bar). Detailed morphology information is further confirmed using high resolution transmission microscopy (HRTEM) from the top view ( Supplementary Fig. S1a,b) and side view ( Supplementary Fig. S1c,d). Photocatalysis is a surface phenomenon and the facet effect strongly relates to formation of surface-active sites. The formation of different truncated and decagonal morphologies with (040) surface facets 49 have found that facet surface energy played a very important role in determining the photocatalytic activity. The m-BiVO 4 NP (Fig. 4c) shows rough diameter approximately 50-120 nm and thickness approximately 20-50 nm, leading to higher active surface area and enhancing its photocatalytic activities.
The elemental mapping composition of m-BiVO 4 was further studied by the energy-dispersive X-ray spectroscopy (EDS) attached on the FESEM. The spatial distribution of bismuth (Fig. 5b), vanadium (    photocatalytic activity and stability. The photocatalytic activity of m-BiVO 4 was investigated by degradation of MB irradiated with visible light (λ > 420 nm). MB is an odourless and dark blue-black heterocyclic aromatic compound. It forms a blue coloured solution when dissolved in room temperature water. Most textile plants use MB for dyeing purpose. MB is highly soluble in water and is very difficult to dispose before discharge into environment. Trace amounts of MB are still harmful to humans as well as aquatic life 6 . Figure 6a shows the photocatalytic MB degradation efficiency in the absence and presence of m-BiVO 4 photocatalyst. Under dark condition, only negligible amounts of MB were degraded, presumably due to the adsorption of dye onto the surface of the catalyst. In contrast to the dark condition, the degradation activity was extremely enhanced and achieved about 91% degradation efficiency under light irradiation condition in the presence of m-BiVO 4 photocatalyst in 60 min, presumably due to the presence of high-active (040) facets 41,42 . The rate constant of the photocatalytic reaction (Fig. 6b) is calculated from the slope of the plot of ln(C o /C t ) versus irradiation time, where C o is the initial MB concentration and C t is the concentration of MB at the time t. The rate constant k is 0.388 × 10 −2 min −1 , indicating that the m-BiVO 4 photocatalyst improved its degradation on MB dyes when exposed to visible light. The UV-Vis spectrum of 20 ppm MB solution after different times of VLD photocatalytic reaction (Fig. 6c) shows corresponding absorbance decrease and the colour changes from initial blue to final transparent (Inset in Fig. 6c). In practical applications, the stability of the photocatalyst is very important.
Continuously recycle experiments show the reusability potential of m-BiVO 4 for MB treatment (Fig. 6d). During the first treatment, most of the surface active sites were occupied by the MB. As the treatment was continued, the number of active sites available for the MB subsequently decreased, and as a result, the degradation efficiency of MB would be decreased. After three consecutive cycles, no noticeable deactivation of m-BiVO 4 was observed.
The results of the recycle experiments (Fig. 6d) indicate that the m-BiVO 4 photocatalyst exhibits a reliability of activity and good stability with excellent degradation efficiency for MB. Under visible light irradiation, the electrons in m-BiVO 4 generated from VB to CB because of their small band gap (Eg = 2.5 eV). Moreover, due to the different energy levels in BiVO 4 (040) and BiVO 4 (110) facets, we presume that the separation of www.nature.com/scientificreports/ electron-hole pairs among the BiVO 4 (040) and BiVO 4 (110) facets may enlarge the potential differences. The reduction and the oxidation reactions may therefore preferentially happen separately on the BiVO 4 (040) and BiVO 4 (110) facets surface. The results indicate that with rational structure design of facet-dependent BiVO 4 semiconductor with small band gap can achieve faster degradation rate 43 . To assess the active species generated in the dye degradation of MB over BiVO 4 , the trapping experiment was performed. In species capturing process, various scavengers such as ethylene diamine tetraacetic acid disodium salt (EDTA-2Na), 1,4-benzoquinone (BQ) and isopropyl alcohol (IPA) was added into the reaction solution as a quenchers of holes (h + ), superoxide radicals (O 2 •− ) and hydroxyl radicals ( • OH), respectively 44,45 . The degradation efficiency of MB was significantly inhibited by the addition of EDTA-2Na and slightly suppressed by the addition of BQ and IPA ( Supplementary  Fig. S4a). The addition of scavengers resulted in decreased degradation is also evidenced by the decreased rate constant (Supplementary Fig. S4b) from k = 0.38 × 10 −2 min −1 (without scavenger) to k = 0.21 × 10 −2 min −1 (IPA), 0.14 × 10 −2 min −1 (BQ) and k = 0.02 × 10 −2 min −1 (EDTA-2Na). The results indicate that a large number of holes (h + ) in the VB of m-BiVO 4 have offers powerful oxidation ability and plays as the main active species in the MB degradation. Based on the active species trapping analysis and band energy level analysis results, we propose a schematic of photocatalytic degradation of MB by VLD m-BiVO 4 (Fig. 7), mainly due to the transfer behaviour of photo-generated electrons and holes among the m-BiVO 4 (040) facet and the follow up reactions between the generated active species and B and BiVO 4 . In-depth mechanistic insights are available elsewhere (Supplementary Discussion S1). In brief, h + VB is the key active species for the degradation of MB by VLD m-BiVO 4 and the O −2• and OH • radicles play supplementary roles during the MB degradation.
Zebrafish embryo toxicity test of the raw and treated MB dye solutions. In photocatalytically dye degradation process, smaller degradation fragments are produced as by-products from the breaking of dye molecules which sometimes are even more toxic than the parent dyes. It is therefore obligatory to investigate the subsistence capacity of the living organism and its biological response after dye degradation 46 . We therefore investigate the ecotoxicity of the raw and photocatalytically treated MB dye solutions by Zebrafish embryo toxicity test 47,48 . On the potential for chemical stressors to affect ecosystems. Herein, the test is based on a 24 h exposure of 2, 24, 48, 72 and 96 h post fertilization (hpf) embryos in a static system. The rates of morphological changes as well as the survival capability are used as endpoints used to generate dose response curves. The Zebrafish embryo is an ideal model as its internal structures are nearly transparent and genetic structure are almost similar to humans. The rates of morphological change results show that the Zebrafish embryo was highly affected in raw MB solution, i.e. 72 hpf (Inset c in Fig. 8) and 96 hpf (Inset d in Fig. 8); whereas the VLD m-BiVO 4 www.nature.com/scientificreports/ treated MB solution (20 ppm) was little affected, i.e. up to 96 hpf (Inset d in Fig. 8). Furthermore, it was noticed that as increased MB concentration (5 to 20 ppm) along with exposure time the MB toxicity expressed as mortality, teratogenicity and survivability apparently worsened on the Zebrafish embryo (Supplementary Table S2).
The zebrafish embryo toxicity test results demonstrate that the raw (untreated) MB dye solution is toxic for aquatic life. After treatment by VLD m-BiVO 4 photocatalyst, the toxicity considerably decreased to a tolerance level. Obviously, the presence of inorganic ions such as iron, zinc, magnesium, calcium, copper, bicarbonate, nitrate, sulphate, phosphate and chlorides in natural water and polluted water might eventually be adsorbed onto the surface of photocatalyst and retard the rate of photocatalytic reaction 49,50 .  photocatalytic dye degradation. Methylene blue with a major absorption band at 668 nm was used as a model pollutant for evaluation of photocatalytic activity of m-BiVO 4 . The experiment was carried out under visible light irradiation at room temperature. Photo-irradiation was carried out using a 1000 W xenon lamp (λ ≥ 420 nm) with a UV cut-off filter to deliver only visible radiation below 420 nm. The distance between the lamp and the MB solution was 20 cm. The photocatalytic degradation of the MB (20 ppm) dye was performed in an aqueous solution. Appropriate amount of m-BiVO 4 (10 mg) was dispersed in an aqueous solution containing MB. Before starting irradiation, the reaction mixture was sonicated for half hour and further stirred for another half hour all in the darkness to reach adsorption − desorption equilibrium between the MB dye and the m-BiVO 4 photocatalyst. The absorbance of MB for adsorption reaction was measured first followed by visible light irradiation under constant magnetic stirring. An aliquot (5 mL) was drawn at regular intervals (10 min) and centrifuged (1,000 rpm) to remove photocatalyst NPs. Different dye concentrations (blank, 5, 10 and 20 ppm) were prepared and studied. The photocatalytic degradation efficiency of the dye was estimated with C t /C 0 , where C t was the concentration of dye at each irradiated time interval, and C 0 was the initial concentration of the dye, respectively. At each irradiated time interval, the UV-Vis absorbance spectrum of treated solution was measured using a UV-Vis spectrophotometer under full scan mode.

Methods
Reactive species capturing experiment. The active species generated during the photocatalytic reaction play very important role. In situ reactive species capture experiments were conducted to identify possible active species by the addition of trapping agents (scavengers). The scavengers such as disodium ethylene diamine tetraacetate (EDTA-2Na), 1,4-benzoquinone (BQ), and isopropyl alcohol (IPA) were added as a quencher for capturing the holes (h + ), superoxide radicals (O2 •− ) and hydroxyl radicals ( • OH), respectively. The EDTA (0.1 mol L −1 ), benzoquinone (0.1 mol L −1 ) and isopropyl alcohol (0.1 mol L −1 ) was separately added (10 mL) in each experiment into a MB solution 40 .
Assessment of dye toxicity on zebrafish embryo and experimental setup. The ecotoxicity of industrial effluent is important which could be verified by the corresponding water quality and survival capability of inherent aquatic animal species. We follow the Taiwan  www.nature.com/scientificreports/