Recently, photocatalytic degradation of harmful pollutants has become an increasingly important technology for water and air purification [1]. TiO₂ is considered to be an excellent photocatalyst because of its biological and chemical inertness, strong photo-oxidization power, cost effectiveness, and long-term stability to photo-induced and chemical corrosion [2]; however, TiO₂ has a broad bandgap (3.2 eV for anatase TiO₂) and absorbs only a small fraction of solar energy [3, 4]. Considering that UV light constitutes only 3%–5% of the solar spectrum, more efficient photocatalysts, which can be activated both by UV and visible light, are highly desired [5-7].

As one of the simplest Aurivillius oxides with a layered 2D structure, Bi₂WO₆ nanoplates are notable for their visible-light-response and photocatalytic performance [8-12]. The perovskite-like structure is defined by WO₆ units that form a layer perpendicular to the [100] direction and sandwich (Bi₂O₂)²⁺ units [13, 14]. These nanoplate materials are usually characterized by high crystallinity and well-defined chemical compositions as well as extremely high anisotropy, with ultrathin plates [8]. However, the photocatalytic activity of Bi₂WO₆ nanoplates is not yet high enough for practical applications because of the rapid recombination of photogenerated electrons and holes, and the limited visible-light-response range [15, 16]. Unmodified Bi₂WO₆ nanoplates can only adsorb UV or visible light of wavelengths shorter than 450 nm [17, 18].

To improve the photocatalytic activity of Bi₂WO₆ nanoplates and/or extend the range of light response, many strategies have been studied, including semiconductor coupling [19], metal/non-metal doping and noble metal deposition [20]. For example, Ge et al. [21] reported a red shift of the light absorption and strong absorbance of visible light region after coupling Bi₂WO₆ with g-C₃N₄. The enhancement in the photocatalytic performance of the g-C₃N₄@Bi₂WO₆ composite materials was attributed to the combination of polymeric g-C₃N₄ and Bi₂WO₆, which improved separation of photogenerated electron-hole pairs [21]. Similarly, our group [3] reported the coupling of Bi₂WO₆ with TiO₂ using a layer-by-layer deposition and calcination strategy. We found that a hybrid coated with four layers of TiO₂ (containing 20 wt% TiO₂) showed the highest photocatalytic activity, which was 10.45 and 3.20 times as high as that of pure Bi₂WO₆ and TiO₂, respectively. Zhu et al. [22] reported that, doping Bi₂WO₆ by F resulted in a two-fold increase in the photocatalytic degradation activity of MB under visible-light irradiation (λ > 420 nm). The enhanced photocatalytic activity was attributed to increased mobility of photo-excited charge carriers and a positive shift of the valance band position. Yu et al. [1] reported the fabrication of high visible photoreactive plasmonic Ag and graphene co-modified Bi₂WO₆ nanosheets, combining the surface plasmon resonance (SPR) effect of Ag and the electronic properties of graphene at the surface of Bi₂WO₆.

Bismuth sulfide (Bi₂S₃) has been reported to be a semiconductor with a narrow bandgap of approximately 1.3 eV [12, 23-25], which has been used as a sensitizer because of its ability to absorb visible and even near infrared light [26-28]. Here, we propose a structure of Bi₂WO₆ nanoplates with a surface layer of Bi₂S₃ to improve the photoreactivity of the material. Considering that strong interactions between two semiconductors is of great importance for efficient separation of carriers [29], the Bi₂S₃ was designed to grow on the surface of the Bi₂WO₆ nanoplate via an in-situ ion-exchange strategy, to form core-shell structured Bi₂WO₆@Bi₂S₃ nanoplates (Scheme 1). The effects of the amount of Bi₂S₃ loading on the structure and photocatalytic activity of the Bi₂WO₆@Bi₂S₃ nanoplates were systematically studied.

All reagents were of analytical purity and were used as received from the Shanghai Chem. Co., China. The Bi₂WO₆@Bi₂S₃ nanoplate hybridized photocatalyst was synthesized by an ion exchange strategy using Bi₂WO₆ nanoplates as a precursor, which were synthesized by a simple hydrothermal process according to the literature [8]. To synthesize the Bi₂WO₆@Bi₂S₃ nanoplates, 1.3 g of square Bi₂WO₆ precursor was added to a beaker containing 40 mL of Na₂S∙9H₂O solution under magnetic stirring. This mixture was then evaporated to dryness in an oven at 60 ℃ for 4 h. The resulting powder was further calcined at 350 ℃ for 2 h in muffle furnace at a heating rate of 5 ℃ min⁻¹. Afterwards, the obtained samples were rinsed with anhydrous ethanol followed by drying in an oven at 80 ℃ for 2 h. The prepared hybridized photocatalysts are denoted as S_x, where x indicates the amount of Na₂S∙9H₂O used (0–3.0 g, Table 1). Thus, the S0 sample represents the 350 ℃-calcined Bi₂WO₆ precursor without treatment by the Na₂S solution.

X-ray diffraction (XRD) patterns were obtained on a D8-advance X-ray diffractometer (German Bruker) using Cu K_α radiation at a scan rate of 0.02° s⁻¹. The accelerated voltage and applied current were 15 kV and 20 mA, respectively. The morphology of the catalysts was characterized with a field emission scanning electron microscope (FESEM, S-4800, Hitachi, Japan) operating at an accelerating voltage of 10 kV and a transmission electron microscopy (TEM: Tecnai G² 20, USA) with an accelerating voltage of 200 kV. The optical absorption properties of the samples were investigated through diffuse reflectance spectrum (DRS) with a UV-Vis spectrophotometer (Lambda, Bio 35, PE Co., USA). Powder photoluminescence (PL) spectrum were measured at room temperature on a Fluorescence Spectrophotometer (F-7000, Hitachi, Japan). The excitation wavelength was 240 nm with a scanning speed of 1200 nm min⁻¹ and a PMT voltage of 400 V. The excitation and emission slits were both set to a width of 10.0 nm. The surface photovoltage spectroscopy (SPS) measurements of the samples were performed with a home-built apparatus. The SPS signals reflect the change of the potential barrier of the tested electrode surface in the presence of light and that in the dark. X-ray photoelectron spectroscopy (XPS) measurements were performed using a Multilab 2000 XPS system with a monochromatic Mg K_α source and a charge neutralizer. All the binding energies were referenced to the C 1s peak at 284.4 eV from adventitious carbon. The Brunauer-Emmett-Teller surface area (A_BET) of the samples was analyzed by nitrogen adsorption in a Micromeritics ASAP 2020 nitrogen adsorption apparatus (USA).

The photocatalytic activity of the as-prepared photocatalyst was evaluated by decomposition of Brilliant Red X-3B dye (X3B) under visible light irradiation at room temperature. The photocatalytic experiments were performed with the sample powder (50 mg) suspended in X3B solution (50 mL, 1.0×10⁻⁴ mol L⁻¹) with constant stirring. Prior to illumination, the suspensions were shaken at a constant rate in the dark for 12 h to achieve adsorption-desorption equilibrium. The suspension was then irradiated under a 350 W Xe lamp with a 420-nm cutoff filter. The reaction was performed at room temperature to prevent any thermal catalytic effects. At certain time intervals, 3.0 mL of the suspension was taken and centrifuged to remove the photocatalyst particles. The concentration of dye was monitored by UV-Vis spectroscopy at λ = 530 nm.

The photocurrents and electrochemical impedance spectroscopy (EIS) of the film electrode were measured on an electrochemical system (CHI-760e, Shanghai, China) with the use of a standard three-electrode cell. The working electrode was composed of a 2 cm × 1.5 cm ITO substrate onto which the sample was deposited, a Ag/AgCl electrode was used as a reference electrode and a platinum wire was used as the counter electrode. The ITO/sample working electrode was prepared by a dip-coating method: 50 mg of the photocatalyst was suspended in 1.0 mL of ethanol, which was then dip-coated onto the ITO glass electrode. The electrolyte was 0.1 mol L⁻¹ Na₂SO₄. The photoelectric response of the sample was measured with switching of an LED lamp (3 W) emitting mainly at 420±10 nm. The EIS measurements were performed in the dark at open circuit voltage, and a sinusoidal ac perturbation of −0.3 V was applied to the electrode over the frequency range of 0.1–10000 Hz.

The phase structure of the prepared photocatalyst was investigated by XRD. Fig. 1 shows that the XRD patterns of precursor (S0 sample) could be indexed to the russellite phase Bi₂WO₆ (PDF#26-1044) [3]. No impurity diffraction peaks were observed, indicating the high purity of the Bi₂WO₆ phase. After the treatment with Na₂S, the peak intensity corresponding to Bi₂WO₆ decreased, while some unknown impurity phase was found in the S0.5 and S1.0 samples. These results suggest that a reaction occurred between the Bi₂WO₆ and Na₂S. For the S1.5 sample, we identified the formation of Bi₂S₃. The reaction between Bi₂WO₆ and Na₂S to produce Bi₂S₃ proceeded through ion exchange (Eq. (1)):

(1)

This ion exchange reaction resulted in the transformation of the surface of the Bi₂WO₆ nanoplates to Bi₂S₃ to form a core-shell structure of Bi₂WO₆@Bi₂S₃. SEM (Fig. 2) and TEM (Fig. 3) images show the plate-like structure of the Bi₂WO₆@Bi₂S₃ sample (S1.5). The Bi₂S₃ layer was produced by in-situ growth at the surface of the Bi₂WO₆ nanoplate, thus, interactions between the Bi₂WO₆ core and Bi₂S₃ shell should be strong. The close proximity of the two semiconductors should be beneficial for efficient separation of photogenerated electron-hole pairs and improve photocatalytic activity [30].

The TEM image of the S1.5 sample indicated that some holes formed in Bi₂WO₆@Bi₂S₃ nanoplate, which could be attributed to the attack of S²⁻, which resulted in the dissolution of Bi₂WO₆ units (Eq. (1)). Further increases of the amount of Na₂S∙9H₂O to 3.0 g resulted in a sharp diffraction peak emerging at 31° in the XRD patterns of S3 sample, which was identified as NaBiS₂. This result indicates that some Bi₂S₃ was further transformed into NaBiS₂ in the high concentration Na₂S solution (Eq. (2)), and the S3 sample was a ternary composite of NaBiS₂/Bi₂S₃/ Bi₂WO₆.

(2)

Fig. 2(e) and (f) show SEM images of the S3 sample, which indicate the formation of a porous structure. The TEM image (Fig. 3(c)) suggests that the S3 sample was an aggregate made up of smaller nanoplates, which resulted in the formation of porous coral-like structures.

Energy-dispersive X-ray spectroscopy (EDS) line scans and mapping analysis were used to study the composition of the photocatalyst. The elemental mapping images (Fig. 4) show that the content of S in the photocatalyst increased in the order: S0 < S1.5 < S3. The EDS elemental mapping results also show that the content of W decreased as the Na₂S content of the treatment was increased. These results further indicate that the reactions described by Eqs. (1) and (2) occurred during the preparation of the photocatalyst [31, 32].

The UV-vis diffuse reflection spectroscopy (DRS) results of the as-prepared samples are compared in Fig. 5. Consistent with previous reports [33], the onset of the DRS for the pristine Bi₂WO₆ sample (S0) started at approximately 450 nm, corresponding to a bandgap of 2.76 eV. This reflects the poor visible light harvesting ability of the material. However, the light absorption of Bi₂WO₆ was greatly enhanced in the visible light region after the sulfur treatment of the surface and even extended to the NIR region (Fig. 5(b)-(f)). This can be attributed to the small bandgap and large absorption coefficient of Bi₂S₃ [34]. Therefore, the in-situ formed Bi₂S₃ on the surface of the Bi₂WO₆ nanoplate can be used as a photosensitizer to excite Bi₂WO₆, which might enable full use of the solar spectrum.

PL analysis is commonly used to analyze the recombination rate of photogenerated electrons and holes in a photocatalyst [35]. Herein, we conducted PL measurements for Bi₂WO₃ before and after treatment with Na₂S. The PL spectra shown in Fig. 6, indicate that the pristine Bi₂WO₆ (S0) sample possessed the highest light emission intensity among the three photocatalysts. This result also suggests the material has poor photocatalytic activity owing to the rapid recombination rate of carriers. The S1.5 sample exhibited the lowest light emission intensity, which suggests that the recombination of electron-hole pairs was substantially suppressed after deposition of Bi₂S₃ on the surface of the Bi₂WO₆ nanoplates. This result is attributed to the formation of a heterojunction between the two semiconductors (Bi₂WO₆ and Bi₂S₃), which facilitated efficient separation of photogenerated electrons and holes, retarding the recombination [36]. However, if the concentration of Na₂S was too high, the PL intensity started to increase, as for the S3 sample.

SPS was also performed to confirm the separation efficiency of the photogenerated carriers of the photocatalyst. The SPS signal of pristine Bi₂WO₆ was negligible (Fig. 7), while the S1.5 sample exhibited the strongest SPS response, which was consistent with the PL characterization results (Fig 6).

XPS was used to investigate the chemical states and surface composition of the Bi₂WO₆@Bi₂S₃ composite. The XPS survey spectra of the unmodified (S0) and the S-modified Bi₂WO₆ nanoplate (S1.5) are compared in Fig. 8(a). The unmodified Bi₂WO₆ nanoplates contained the elements Bi, W, O, and C. The carbon peak was attributed to residual carbon from the precursor solution and adventitious hydrocarbons from the XPS instrument itself [37]. Peaks with binding energies of 164.43 and 159.11 eV corresponded to photoelectron emission from Bi 4f_7/2 and Bi 4f_5/2 of Bi (Ⅲ), respectively (Fig. 8(b)), while the peaks with binding energies of 37.6 and 35.5 eV were attributed to W 4f_5/2 and W 4f_7/2 of W (Ⅵ), respectively, (Fig. 8(c)) [38].

After the Na₂S treatment, the signal from W decreased (Fig. 8(c)), and a weak signal from S was observed in the high resolution XPS spectrum (Fig. 8(d)). The peaks centered at 161.7 and 162.7 eV were attributed to the S 2p_3/2 and S 2p_1/2 of S²⁻ [39]. In addition, the binding energy of the spin-orbit Bi 4f could be deconvoluted into four peaks after the sulfur treatment. Two new weak peaks centered at 157.6 and 162.7 eV were related to the formation of Bi₂S₃ in the Bi₂WO₆@Bi₂S₃ nanoplates. These results were consistent with the XRD results (Fig. 1), and further confirmed successful substitution of S²⁻ into WO₆⁶⁻ through an ion exchange reaction (Eq. (1)).

To determine the BET specific surface area and microstructure of the photocatalyst, nitrogen adsorption-desorption isotherms were measured (Fig. 9). The nitrogen adsorption-desorption isotherms were similar and all possessed type IV isotherms (BDDT classification) with a hysteresis loop at high relative pressures, indicating the mesoporous character of the surface. The shape of the hysteresis loop was of H3 type, suggesting the presence of narrow slit-shaped pores, which are generally associated with plate-like particles [2], agreeing well with the nanoplate morphology of the particles (Figs. 2 and 3).

The adsorption isotherm of the S1.5 sample was shifted upward compared with those of the S0 and S3.0 samples, indicating its larger BET surface area. The BET surface areas of the S0, S1.5 and S3.0 samples were measured to be 18.8, 42.6, and 17.8 m² g⁻¹, respectively (Table 1). A larger BET specific area suggests that more active sites are available for adsorption and photocatalytic reactions, which might contribute to enhanced photocatalytic activity.

The hysteresis loop of the S1.5 sample begins at a relative pressure of 0.45, while those of S0 and S3 samples begin at higher relative pressures of approximately 0.9, indicating that the S1.5 sample possesses smaller mesopores than those of the S0 and S3 samples. This result was confirmed by the corresponding pore size distribution curves, as shown in Fig. 9(b). The S1.5 sample possesses a larger pore volume (PV) and a smaller average pore size (APS) than those of the S0 and S3.0 samples. The larger PV and smaller APS of the Bi₂WO₆@Bi₂S₃ nanoplates can be attributed to the etching of Bi₂WO₆ by the S²⁻ to form a porous structure (Eq. (1) and Fig. 3).

We evaluated the photocatalytic degradation of X3B dye under visible light irradiation. Fig. 10(a) shows that the degradation of X3B was negligible under light irradiation without the addition of photocatalyst (control experiment). The photocatalytic activity of pristine Bi₂WO₆ (S0 sample) was very poor, owing to its limited response to visible light (Fig. 5) and rapid recombination of photogenerated electron-hole pairs (Fig. 6 and 7). As the amount of Na₂S∙9H₂O was increased from 0 to 1.5 g, the X3B removal efficiency steadily improved, and the S1.5 sample showed the highest photocatalytic activity among all the photocatalysts. A further increase in the content of Na₂S∙9H₂O caused the reactivity of the photocatalyst to decrease, likely because of the formation of a photo-inactive NaBiS₂ surface layer (Fig. 1), which prevented reactions between the substrate and photocatalyst.

The degradation profiles of X3B followed pseudo-first-order kinetics, and the related degradation rate constants are compared in Fig. 10(b). The rate constant of the S1.5 sample (6.60×10⁻³ min⁻¹) was 16.5 times as high as that of pristine Bi₂WO₆ (0.40×10⁻³ min⁻¹).

To account for the photocatalytic degradation mechanism of X3B on the core-shell structured Bi₂WO₆@Bi₂S₃ hybridized photocatalyst, we performed radical trapping experiments to detect the reactive oxidative species (ROSs) by introducing isopropyl alcohol (IPA), triethonoamine (TEOA), and 1, 4-benzoquinone (BQ) into the degradation system to scavenge •OH, h⁺, and O₂•⁻, respectively. After addition of IPA, TEOA, and BQ, the degradation rate constant of X3B decreased from 6.60×10⁻³ to 3.50×10⁻³, 1.75×10⁻³ and 1.03×10⁻³ min⁻¹, respectively (Fig. 11). Thus, the quenching effect of TEOA and BQ on the degradation of X3B was more pronounced than that of IPA. This result indicates that both h⁺ and O₂^•− are the main ROSs responsible for the oxidization of X3B.

According to previous reports, the potentials for the valence band (E_VB) and conduction band (E_CB) of Bi₂WO₆ are 3.38 and 0.40 V [40], while those for Bi₂S₃ are 1.42 and 0.12 V, respectively [33]. Therefore, a heterojunction between Bi₂WO₆ and Bi₂S₃ is formed owing to their staggered energy potentials (Scheme 1). Under visible light irradiation, photogenerated electrons in the CB of Bi₂S₃ can transfer to the CB of Bi₂WO₆, while holes on Bi₂WO₆ can migrate to the VB of Bi₂S₃. This semiconductor coupling effect can efficiently improve the separation rate of photogenerated electron-hole pairs to enhance the photocatalytic activity [41].

Fig. 12 compares the photocurrents [42, 43] and EIS responses of the S0, S1.5, and S3 samples. The S1.5 sample featured the largest photocurrent (Fig. 12(a)) and the smallest arc radius in the EIS Nyquist plots (Fig. 12(b)), which further confirmed the efficient separation of photogenerated carriers at the Bi₂WO₆@Bi₂S₃ nanoplate owing to the formation of a heterojunction. The enlarged BET surface area and improved light harvesting ability owing to Bi₂S₃ sensitization should also benefit the photoreactivity of the Bi₂WO₆@Bi₂S₃ nanoplate.

The stability of the photocatalyst is of great importance from the viewpoint of practical applications. Therefore, we studied the reusability of Bi₂S₃@Bi₂WO₆ in the photocatalytic degradation of X3B over four consecutive cycles under visible light irradiation, and the results are shown in Fig. 13. It can be seen that the degradation profiles of X3B in the presence of the S1.5 photocatalyst remained almost unchanged from the first to the fourth run, reflecting the good stability of the Bi₂S₃@Bi₂WO₆ composite. The high stability of the as-prepared Bi₂S₃@Bi₂WO₆ hybrid photocatalyst suggests promise for practical applications.

Bi₂WO₆@Bi₂S₃ nanoplates were synthesized by ion exchange between Bi₂WO₆ and Na₂S. The photocatalytic activity of the Bi₂WO₆@Bi₂S₃ nanoplates was 16.5 times as high as that of the pristine Bi₂WO₆. The improved photocatalytic activity of the Bi₂WO₆@Bi₂S₃ nanoplates was attributed to the formation of a heterojunction between Bi₂WO₆ and Bi₂S₃, which facilitated efficient separation of photogenerated electrons and holes and retarded their recombination. The enlarged BET surface area and photosensitization of Bi₂S₃ also contributed to the enhancement of photocatalytic activity for the Bi₂WO₆@Bi₂S₃ nanoplate hybridized photocatalyst. The present study provides an alternative approach to the design of highly efficient and ultra-stable hybrid photocatalysts.