Visible-light-driven photocatalytic degradation of rhodamine B in water by BiOCl_xI_1−x solid solutions

Heterogeneous photocatalysis has been widely accepted as an environmentally-benign technology for environmental pollution treatment and renewable energy utilization (Guo et al. 2019; Sharma et al. 2019; Iervolino et al. 2020; Koe et al. 2020; Liu et al. 2020). The rapid developments of urbanization and industrialization have led to the huge production of wastewater containing organic contaminants. Many methods have been applied for the remediation and treatment of organic wastewater including adsorption, coagulation, extraction, filtration, biodegradation, and advanced oxidation processes (AOPs) (Paździor et al. 2019; Pronk et al. 2019; Völker et al. 2019). Nevertheless, all these methods have some shortcomings as follows: adsorption requires the transportation of target molecules from the aqueous phase to the surface or porous interior of solid phase (Eskandarloo et al. 2017); biodegradation takes a long time and is limited to the biodegradable compounds; and large amounts of sludge can be produced in the coagulation process (Kumwimba & Meng 2019). Photocatalysis is considered as a promising technology for the degradation of organic pollutants. Generally, when a photon with enough energy is absorbed by the photocatalyst, the negative electrons in its valence band (VB) are excited and transferred to the conduction band (CB) and the positive holes are accordingly generated in the VB. The holes in VB can react with water (H₂O) or hydroxyl ions (OH⁻) to generate hydroxyl radicals (·OH), and the excited electrons in CB can react with oxygen (O₂) to generate superoxide radicals (·O₂⁻). These radicals subsequently take part in the redox reactions and lead to the complete mineralization of pollutant molecules by their degradation into H₂O and CO₂ (Khaki et al. 2017; MiarAlipour et al. 2018). In recent decades, various semiconductor materials and their heterojunctions, such as TiO₂ (Meng et al. 2019), TiO₂/CdS (Low et al. 2019), ZnO/CdS (Wang et al. 2019a), WO₃/g-C₃N₄ (Fu et al. 2019), MoS₂/g-C₃N₄ (Yuan et al. 2019) and TaON/V₂O₅ (Wang et al. 2019b), have been developed as effective photocatalysts triggered by UV or visible light. Furthermore, recent developments in Bi-based semiconductors have heightened the need for photocatalysis in relation to appropriate band gap for visible-light response and hybridized valence band composed of Bi 6s and O 2p orbitals (Meng & Zhang 2016; Li et al. 2018).

Bismuth oxyhalides (BiOXs, X = Cl, Br and I) have been considered as the most promising candidates as photocatalytic materials since the unique layered structure endows them with flexible band structure and superior photocatalytic activity (Li et al. 2014a; Bhachu et al. 2016). The typical layered structure of BiOXs is constructed by [Bi₂O₂]²⁺ layers intercalated with [X]⁻ between the layers. The atoms within [Bi₂O₂]²⁺ layers interact with each other via covalent bonds, whereas the stack along the c-axis of [X] layers is bonded by van der Waals force. Thereby, the sandwich-like layered structures are configured by the intralayer covalent bonding and interlayer van der Waals interaction (Li et al. 2014a; Xu et al. 2016). Moreover, the internal electric field (IEF) along c-axis can be generated in BiOX structure, attributed to the asymmetry of charge distribution between [Bi₂O₂]²⁺ and [X]⁻ layers (Li et al. 2014a). IEF can efficiently separate the photogenerated electron-hole pairs and prevent their recombinations (Li et al. 2017). However, previously published reports indicated that the photocatalytic efficiency of pristine BiOX was not high enough for practical use. Many modification attempts, such as doping by noble metals, morphology modulation, composition design and combination with other materials, have been made to improve the photocatalytic activity of pure BiOX (Bielicka-Giełdoń et al. 2019; Sharma et al. 2019; Wang et al. 2019c). Among BiOXs, the band gaps of BiOCl, BiOBr and BiOI were found to be 3.22, 2.64, and 1.77 eV, respectively (Ren et al. 2013). Therefore, BiOCl and BiOBr can be mainly excited by UV light, while BiOI can respond to visible light. Nevertheless, BiOI still exhibits poor photo-catalytic property because of its low quantum efficiency, weak photo-oxidation ability and easy recombination of photogenerated electrons and holes (Hou et al. 2017). It has been reported that, when iodine was introduced into BiOCl; or chlorine into BiOI, dramatic increases in photo-catalytic performance could be observed under visible light irradiation (Kim et al. 2014; Tian et al. 2016). Since BiOXs have the same layered structure and atomic configuration, it is very easy to realize the substitution of halogen atoms to form BiOX_xY_1−x (X, Y = Cl, Br and I) solid solutions with unlimited solubility (Keller & Krämer 2005). Previous studies have demonstrated that the formation of solid solutions could substantially modulate the band gap and improve visible-light photocatalytic activity of BiOXs (Zhang et al. 2015,, 2016; Yang et al. 2016). We have developed a facile alcoholysis route to successfully synthesize BiOCl_xBr_1−x and BiOBr_xI_1−x solid solutions and found that they exhibited higher photocatalytic performance under visible-light irradiation (Xu et al. 2017a , 2017b).

These findings encouraged us to go ahead and attempt the synthesis of BiOCl_xI_1−x solid solutions through the alcoholysis method in this work. The obtained photocatalysts were characterized by X-ray diffraction (XRD), field emission scanning electron microscopy (FESEM), transmission electron microscopy (TEM), UV-vis diffuse reflectance spectroscopy (UV-vis DRS), photoluminescence emission spectroscopy (PL) and Brunauer–Emmett–Teller (BET) surface area analysis. Then, the active dye Rhodamine B (RhB) was employed as the targeted contaminant to evaluate the photocatalytic performance of these photocatalysts under simulated sunlight irradiation. Finally, the possible photocatalytic mechanism was proposed and discussed.

A facile alcoholysis method was used to prepare BiOCl_xI_1−x nanosheet solid solutions at room temperature and atmospheric pressure. All the chemical reagents were of analytical grade and were used without further purification. Bi(NO₃)₃·5H₂O, KCl and KI were employed as Bi, Cl and I sources, respectively. In a typical alcoholysis route, according to the stoichiometric value of x (x = 1.0, 0.75, 0.5, 0.25, and 0.0 for this study) in BiOCl_xI_1−x solid solutions, stoichiometric amounts of KCl and KI were first dissolved in as little deionized water as possible. After the above first step, requisite amounts of Bi(NO₃)₃·5H₂O and absolute ethyl alcohol were successively put into a 200 mL beaker and the process of alcoholysis was started by ultrasonic irradiation until Bi(NO₃)₃·5H₂O was completely dissolved in absolute ethyl alcohol. Subsequently, the KCl-KI mixed solution prepared in the first step was added to the above obtained alcoholysate and the mixture was mechanically stirred for 6 h. Then, the suspension was centrifuged to separate the solid product from solution. The resulting solid was washed by deionized water and absolute ethyl alcohol alternately until the filtrate reached neutral pH. After filtration, the resulting solid was dried at 80 °C for 6 h followed by grinding to obtain the powder of BiOCl_xI_1−x solid solutions for subsequent studies.

XRD was employed to identify the crystal phase of as-prepared photocatalysts by a PANalytical X'Pert X-ray diffractometer with Cu-K_α radiation (λ = 0.15418 nm) over the 2θ range of 5°–80°. Based on 2θ values in XRD patterns, the lattice parameters of BiOCl_xI_1−x solid solutions were calculated by UnitCell software. The interlayer spacing of (001) crystal plane was calculated by the Bragg's equation (λ = 2dsinθ). FESEM and TEM were used to observe the microscopic morphology of BiOCl_xI_1−x solid solutions, operated on FEISirion200 scanning electron microscope and JEOL JEM 2010 transmission electron microscope, respectively. The changes of optical absorption and band gap for BiOCl_xI_1−x solid solutions were investigated by UV-vis DRS spectra. They were obtained on an USB4000 UV-vis spectrometer with an integral sphere (Ocean Optics) and a standard template bought from South Africa Optics was used as the reference. PL spectra were collected by a RF-5301PC fluorescence spectrophotometer at room temperature. A 150 W Xenon lamp was used as the excitation light source (λ_ex = 304 nm). PL spectra were helpful to gain the detailed information of the recombination of photogenerated electron-hole pairs. At the liquid nitrogen temperature, BET specific surface area of BiOCl_xI_1−x photocatalysts was measured on a Sibata SA-1100 surface area analyzer based on nitrogen adsorption–desorption data.

In the photocatalytic process, a 100 mW·cm⁻² Xe lamp was used as the simulated solar light. 100 mg of BiOCl_xI_1−x photocatalyst was added into 100 mL RhB solution, whose initial concentration (C₀) was 20 mg·L⁻¹. Before turning on the Xe lamp, the suspension was magnetically stirred in the dark for 60 min to balance the adsorption and desorption. Afterwards, the Xe lamp was turned on and photocatalytic reaction started under continuous stirring. At regular 30 min intervals, about 5 mL of the suspension was extracted from the reaction liquid and the supernatant was obtained by centrifugation. The residual RhB concentration (C) in the supernatant was detected by 751II UV-vis spectrophotometer at RhB maximum absorption wavelength (553 nm). And, UV-vis absorption spectra were employed to analyze the residual RhB solutions after different photocatalytic periods.

Figure 1 provides the XRD results of BiOCl_xI_1−x solid solutions. All the sharp diffraction peaks in the figure indicated that BiOCl_xI_1−x photocatalysts prepared by the alcoholysis route were well crystallized (Lei et al. 2009). The end-members BiOCl and BiOI seem to be consistent with JCPDS 6-249 and JCPDS 2-634, respectively. It should be noted that, as the x value in BiOCl_xI_1−x decreases, all the diffraction peaks are expected to show a regular shift to the small angle region. This finding can be interpreted by the fact that, when the smaller Cl atoms (r_Cl = 0.181 nm) are replaced by larger I atoms (r_I = 0.216 nm), the interlayer spacing of the crystal structure expands. In addition, when the x value decreases, the lattice parameters a and c of BiOCl_xI_1−x solid solutions gradually increased, as listed in Table 1. These regular changes in XRD peak positions and lattice parameters strongly suggested that a group of BiOCl_xI_1−x solid solutions was obtained (Xu et al. 2017a,2017b). Moreover, no other diffraction peaks of impurity phases could be seen in the XRD patterns, suggesting that perhaps all the atoms have entered into the crystal lattice and the stoichiometric value x in BiOCl_xI_1−x can be accurately determined by the initial proportion of Cl and I sources in the preparation procedure (Yang et al. 2016). From Figure 1, it can also be observed that the (001) peak exhibits the strongest diffraction intensity in all the XRD patterns, indicating that the exposure percentage of {001} crystal plane is rather higher in BiOCl_xI_1−x solid solutions compared to pure BiOCl. This phenomenon might be attributed to the highly preferred (001) orientation (Li et al. 2014b). Therefore, it might be determined that the as-obtained photocatalysts in this study were BiOCl_xI_1−x solid solutions with high exposure of {001} crystal plane.

Through the FESEM images of BiOCl_0.75I_0.25, BiOCl_0.5I_0.5 and BiOCl_0.25I_0.75 (Figure 2), we can observe that all the BiOCl_xI_1−x photocatalysts exhibit two-dimension flake-like shape with the diameter of 0.2–1.0 μm. These nanoplates were stacked together irregularly without any specific architecture such as microspheres or flowers. In TEM images (Figure 3(a)–3(e)), the overlayed shadow of two-dimension BiOCl_xI_1−x nanosheets can be clearly seen. Moreover, the wrinkles of nanosheets can be obviously seen as well, which might result from the inner strain in BiOCl_xI_1−x structure. The existence of inner strain can change the electronic band structure and quicken the electron-hole separation, which could correspondingly improve the photocatalytic activity of BiOXs (Feng et al. 2015). The pattern of selected area electron diffraction (SAED) of BiOCl_0.5I_0.5 (Figure 3(f)) shows a group of diffraction rings, indicating the polycrystalline character of BiOCl_0.5I_0.5. The indices of crystal plane corresponding to diffraction rings are given in the pattern.

Figure 4(a) illustrates the UV-vis DRS spectra of BiOCl_xI_1−x solid solutions. It can be seen that the optical absorption thresholds of BiOCl and BiOI correspond to 391.7 and 672.8 nm, respectively, suggesting that BiOCl can be excited only by UV light while BiOI can respond to visible light. The optical absorption thresholds of the other three BiOCl_xI_1−x solid solutions are expected to be present between those of BiOCl and BiOI, and systematically increase with increasing I content. Furthermore, the color of BiOCl_xI_1−x solid solutions gradually changed from milky white of BiOCl to reddish brown of BiOI. The band gap (E_g) of BiOCl_xI_1−x solid solutions can be determined according to UV-vis DRS spectra, as shown in Figure 4(b) and Table 1. Interestingly, the E_g values of BiOCl_xI_1−x photocatalysts gradually decrease with I content increasing. This also indicates the formation of BiOCl_xI_1−x solid solutions. Judging from the regular change of E_g values, it can be strongly inferred that the band structure of BiOCl_xI_1−x solid solutions can be modulated by the substitution of smaller Cl ions by larger I ions. It has been proved previously that, for BiOX semiconductors, the conduction band minimum (CBM) is contributed by Bi 6p orbitals and the valence band maximum (VBM) is dominated by O 2p and X np orbitals (n = 3, 4 and 5 for X = Cl, Br and I, respectively) (Li et al. 2014a; Xu et al. 2016). Therefore, as I content increased, the contribution of I 5p orbitals dramatically increased and band gap of BiOCl_xI_1−x photocatalysts decreased.

Figure 5(a) depicts the photocatalytic degradation of RhB in aqueous solution by BiOCl_xI_1−x photocatalysts under simulated sunlight irradiation. As known, the adsorption played an important part in the heterogeneous photocatalytic reactions. In this study, when the adsorption and desorption reached equilibrium, the adsorbed amounts of RhB onto BiOCl_xI_1−x photocatalysts with different x values were almost similar. This might be attributed to the fact that the photocatalysts had no distinct difference in their surface areas (see Table 1). We can also find from this figure that BiOCl_0.75I_0.25, BiOCl_0.5I_0.5 and BiOCl_0.25I_0.75 exhibited better photocatalytic activity than BiOCl and BiOI. Only 31.5% and 81.2% of RhB could be degraded over BiOI and BiOCl, respectively, within 240 min of reaction time; nevertheless, almost all the RhB could be degraded over BiOCl_0.75I_0.25 within 150 min reaction time. Because the adsorption had little effect on photocatalytic evaluation, the enhanced photocatalytic activity of BiOCl_0.75I_0.25, BiOCl_0.5I_0.5 and BiOCl_0.25I_0.75 didn't essentially depend on adsorption. Among BiOCl_xI_1−x solid solutions, BiOCl_0.75I_0.25 had the highest photocatalytic activity, in accordance with the previous report (Ren et al. 2013).

In addition, Figure 5(b) reveals that the photocatalytic degradation of RhB in aqueous solution follows a pseudo-first-order kinetic model [ln(C/C₀) = kt]. The reaction rate constants are calculated and illustrated in the inset of Figure 5(b). The k values of BiOCl_0.75I_0.25, BiOCl_0.5I_0.5 and BiOCl_0.25I_0.75 are approximately 34.1, 7.58 and 6.24 times higher, respectively, than that of BiOI; and approximately 6.82, 1.51 and 1.24 times higher, respectively, than that of BiOCl, strongly suggesting the enhanced photocatalytic activity of BiOCl_xI_1−x solid solutions.

When BiOCl_0.75I_0.25 was used as the photocatalyst, UV-vis absorption spectra of RhB solution after different photodegradation periods were analyzed. Figure 6(a) demonstrates that, as the photocatalytic reaction progressed, the UV-vis absorption peaks of RhB solution gradually weakened and even disappeared. This result infers that almost all the RhB molecules were decomposed into the substances with small molecular weight. Furthermore, the stability is equally important for a photocatalyst. We also use BiOCl_0.75I_0.25 as the representative sample to evaluate the recyclability of BiOCl_xI_1−x solid solution photocatalysts. Figure 6(b) reveals that, even after four runs, BiOCl_0.75I_0.25 still has high photocatalytic activity. This suggests that BiOCl_xI_1−x solid solutions not only exhibited high photocatalytic activity but also showed excellent photocatalytic stability.

In our previous studies, it has been found that the photocatalytic properties of BiOXs solid solutions were closely dependent on the IEF, electronic band structure and interlayer spacing of (001) crystal plane (d₀₀₁) (Xu et al. 2017a,2017b). Stronger IEF could promote electron-hole pairs transfer and inhibit their recombination, accordingly resulting in higher photocatalytic activity of {001} facet-dominant BiOX nanosheets (Li et al. 2014a). On the other hand, the increased d₀₀₁ might augment the transfer distance of photogenerated charge driven by IEF, suppressing the effective electron-hole separation and thereby decreasing the photocatalytic ability of BiOCl_xI_1−x photocatalysts. Based on XRD results and Bragg's equation (λ = 2dsinθ), the interlayer spacing d₀₀₁ of BiOCl_xI_1−x samples was calculated and listed in Table 1. It can be obviously seen that the (001) interlayer spacing gradually increases with the increase of I content in BiOCl_xI_1−x photocatalysts. This means that the electron-hole separation efficiency might be suppressed when I content increases. PL spectrum was employed to analyze the separation efficiency of photogenerated electron-hole pairs in BiOCl_xI_1−x photocatalysts and the results are presented in Figure 7. We can find that BiOCl has the best separation efficiency of electron-hole pairs among these photocatalysts. Whereas, the PL peak intensity of BiOI is the strongest, implying that the electron-hole pairs in BiOI tend to recombine (Jiang et al. 2012). For the other three BiOCl_xI_1−x photocatalysts, the electron-hole separation efficiency seems not in positive correlation with the interlayer spacing of (001) crystal plane. This might be ascribed to the difference in IEF magnitudes of these three photocatalysts. Hence, it can be inferred that the separation efficiency of electron-hole pairs in BiOCl_xI_1−x is determined by the synergy between (001) interlayer spacing and IEF magnitude.

In addition, different molar ratios of Cl and I in BiOCl_xI_1−x solid solutions would generate different energy levels of hybrid orbitals, which means that the electronic band structure of BiOCl_xI_1−x can be modulated by the alteration of x value in BiOCl_xI_1−x. As shown in Figure 4(b) and Table 1, the E_g values of BiOCl_xI_1−x photocatalysts successively decrease with the increase of I content. Hence, as I content increased, BiOCl_xI_1−x photocatalysts were more easily excited by visible light and exhibited higher photocatalytic activity. As discussed above (Figure 5(a)), BiOCl_0.75I_0.25, BiOCl_0.5I_0.5 and BiOCl_0.25I_0.75 showed better photocatalytic activity than BiOCl and BiOI. Because BiOCl had the widest band gap (3.28 eV) among BiOCl_xI_1−x photocatalysts, it could not be effectively triggered by visible light. While, for BiOI, its separation efficiency of electron-hole pairs was the worst among these photocatalysts. Therefore, the visible-light-driven photocatalytic activities of BiOCl and BiOI were much lower than those of BiOCl_0.75I_0.25, BiOCl_0.5I_0.5 and BiOCl_0.25I_0.75. Based on the above discussions, the possible mechanism of RhB photodegradation over BiOCl_xI_1−x solid solutions triggered by visible light was proposed. As shown in Figure 8, when BiOCl_xI_1−x semiconductor is irradiated by simulated solar light, the electrons (e⁻) are excited to the conduction band and the holes (h⁺) are generated in the valence band. However, the electron-hole pairs tend to recombine to reduce total energy of the system, thereby resulting in the decrease of photocatalytic ability. In BiOCl_xI_1−x semiconductor, IEF can drive the electrons and holes to [Bi₂O₂]²⁺ and [X]⁻ layers, respectively, realizing their effective separation. But, larger interlayer spacing of (001) crystal plane can lead to a longer distance between [Bi₂O₂]²⁺ and [X]⁻ layers, which would be not helpful to the separation of photogenerated charges. Therefore, the photocatalytic performance of BiOCl_xI_1−x solid solutions is simultaneously affected by these structural factors. In addition, it has been verified that there existed a strong interaction between RhB and BiOX nanosheets, which could facilitate the electron injection from excited RhB* into conduction band of BiOX nanosheets (Hu et al. 2014). The matching energy levels of excited RhB* and BiOX nanosheets also played a crucial role in the photosensitization process (Zhang et al. 2015; Chen et al. 2016). The photosensitization of RhB helps with its photodegradation.

A facile alcoholysis method was successfully used to prepare BiOCl_xI_1−x solid solutions at room temperature and atmospheric pressure. The regular changes in XRD peak positions and cell parameters strongly suggested that the obtained products belong to a group of solid solutions. The strong diffraction intensity of (001) peak indicated the high exposure percentage of {001} crystal facets in BiOCl_xI_1−x solid solutions. Two-dimensional BiOCl_xI_1−x nanosheets could be clearly observed by FESEM and TEM. The optical absorption threshold of BiOCl_xI_1−x solid solutions proportionately increased with increasing I content, accordingly their band gaps gradually decreased, suggesting that the formation of solid solution could modify the optical absorption property and band structure of BiOX. BiOCl_0.75I_0.25, BiOCl_0.5I_0.5 and BiOCl_0.25I_0.75 exhibited higher photocatalytic activity than BiOCl and BiOI. Among all the BiOCl_xI_1−x photocatalysts, BiOCl_0.75I_0.25 had the highest photocatalytic activity. Moreover, the RhB photodegradation processes over BiOCl_xI_1−x followed a pseudo-first-order kinetic equation. BiOCl_xI_1−x solid solutions not only had high photocatalytic activity but also showed excellent photocatalytic stability. The IEF, electronic band structure and interlayer spacing of (001) crystal plane were the important structural factors that affected the photocatalytic performance of BiOCl_xI_1−x solid solutions. Furthermore, the photosensitization of RhB also played a crucial role in its photodegradation. Towards practical applications of BiOCl_xI_1−x photocatalysts, future research should be put on the photocatalyst separation and solar-reactor designs. Many methods such as centrifugation, flocculation, flotation, membrane technology or their combinations seem to show great potentials for the separation of nanoscale photocatalysts from their aqueous suspensions. Moreover, magnetic separation is also considered to be an alternative strategy when a core-shell heterojunction is fabricated using a magnetic material like Fe₃O₄ and BiOCl_xI_1−x nanosheets as the core and shell, respectively. On the other hand, the solar-to-chemical efficiency might be the most important indicator for the photocatalytic reactor designs. To gain higher efficiency, light absorption, heat and mass transfer, and reaction kinetics must be further investigated. In addition, the optimizations of operating condition and geometric configuration are equally indispensable for the design of a high-efficiency reactor.