In recent years, with rapid industrialization, urbanization, and population growth, environmental pollution attributed to the improper use of pesticides has attracted significant concern globally [1-3]. Much attention has been drawn to the detrimental effects of long-term residues of pesticides and their intermediate degradation products in the environment [4-7]. In this context, many scientists have devoted their efforts to the development of advanced oxidation technologies for the treatment of residual pesticides in environmental waters [8-10]. Among these technologies, photocatalysis is considered to be the most promising strategy in the area of environmental recommendation by virtue of the inexhaustible and environmentally friendly feature of solar energy [11]. Since the discovery of photocatalysis in 1972 [12], photocatalytic technology has been widely accepted for the activation of molecular oxygen by photogenerated carriers [13-18]. The contaminants are likely to react with these generated oxygen radicals, thus causing their degradation. Unfortunately, considering the difficulty in the achieving a compromise between the light absorption and the redox ability of the photocatalysts, highly efficient photocatalysis processes for molecular oxygen activation are desirable for environmental application, although combining the advantages of light absorption and redox ability remains challenging [19, 20].

Recently, defect engineering, particularly the introduction of oxygen vacancies (OVs), was certified as an efficient strategy to engineer the bandgap and tailor the electronic structures of photocatalysts, thereby appreciably boosting the photocatalytic efficiencies [21-23]. The existence of defect states can also effectively trap charge carriers, thus accelerating the separation of photogenerated electron–hole pairs. OVs with abundant trapped electrons could act as active sites for the activation of inert gas molecules; thus, significant effort has been devoted to clarifying the inherent functionality of OVs in photocatalysis at the surface molecular level [24]. Many papers have reported that native OVs could easily charge O₂ via localized electrons. This suggests that the activation of the photocatalytic molecular oxygen (O₂) over OVs is a credible alternative to overcome the spin forbidden reactions of O₂ in triplet state toward the generation of reactive oxygen species (ROS) [25-27]. The resulting ROS, such as hydroxyl radicals (^•OH), superoxide anion radicals (^•O₂⁻), and singlet oxygen (¹O₂), are generally considered to exhibit more oxidizability than O₂. Despite the numerous pioneering works centered on OV construction, the strategy for the regulation of the OV concentration on the photocatalyst is still in its infancy stage, leaving enormous room for further exploration [28-31]. Therefore, it is of great significance and importance to establish a method for controlled introduction of OVs to disclose the potential role of OVs for enhanced molecular oxygen activation [32-35].

As one of the simple members of the Aurivillius family, bismuth tungstate (Bi₂WO₆) is regarded as a promising photocatalyst by virtue of its intrinsic chemical inertness, appropriate band potentials, and unique layered structure [36-38]. Consisting of alternating stacks of [Bi₂O₂]²⁺ layers and perovskite-like [WO₄]²⁻ layers with oxygen atom sharing, Bi₂WO₆ possesses exceptional characteristics of facile OV generation under O₂-deficient thermal conditions due to the low energy of Bi–O bonds. In addition, doping of some non-metal elements is a valid route to adjust the metal–oxygen bonds in Bi₂WO₆ based on the previous report [39]. For example, the presence of F atoms was reported to remove the oxygen atom in F-doped Bi₂WO₆. Bearing these aspects in mind, we intend to choose another halogen anion as a dopant for developing a novel iodine-doped Bi₂WO₆ catalyst, where I^– is bonded with the Bi atom of Bi₂WO₆ to weaken the energy of Bi–O bonds [40]. Under the O₂-deficient thermal condition, the concentration of the OVs was regulated by the doping content of iodine.

In the present study, iodine-doped Bi₂WO₆ catalysts with OVs were synthesized in a reductive ethylene glycol solvent via a facile and simple solvothermal method. With the predictions afforded by theoretical calculations, various characteristic methods were employed to investigate the impacts of iodine doping on the physicochemical properties of Bi₂WO₆. The different concentrations of the OVs in Bi₂WO₆ were further detected by electron spin resonance spectroscopy (ESR) to prove the close correlation between iodine doping and OV introduction. The influence of the abundant OVs on the Bi₂WO₆ surface on its photocatalytic activity in visible light was subsequently investigated by the photocatalytic degradation of sodium pentachlorophenate (NaPCP), a widely used pesticide with high toxicity and slow biodegradation in the environment. With the molecular oxygen activation induced by the OVs, the total organic carbon (TOC) removal rate of iodine-doped Bi₂WO₆ was found to be 10.6 times higher than that of the pristine Bi₂WO₆ under visible light irradiation. In addition, the mechanism, through which the degradation of NaPCP over iodine-doped Bi₂WO₆ is enhanced, was finally speculated based on the results of radical detection and capture experiments.

All samples were synthesized by a simple solvothermal method. The specific steps are as follows. Under magnetic stirring, Bi(NO₃)₃·5H₂O (30 mL, 0.005 mmol) dissolved in ethylene glycol and sodium NaWO₄·5H₂O (Bi₂WO₆ stoichiometric ratio) dissolved in water were uniformly mixed. After the continuous addition of 20 mL of different concentrations of KI aqueous solution (0, 0.06, 0.12, 0.18, or 0.24 mol/L), the mixed liquor was stirred continuously for 20 min. The resulting reaction precursor solution was then sealed in an 80 mL PTFE lined stainless steel autoclave and heated at 160 ℃ for 12 h. Finally, the iodine-doped bismuth tungstate product was collected by centrifugation, washed severally with ethanol and distilled water repeatedly, and then dried in an oven at 60 ℃ for 12 h. To study the effect of the iodine doping concentration, in accordance with the concentration of KI added during the synthesis, the resulting products were designated as BWO, BWO-0.06, BWO-0.12, BWO-0.18, and BWO-0.24, respectively.

X-ray diffraction (XRD) patterns of the samples were recorded on a Bruker D8 Advance X-ray diffractometer with Cu K_α radiation (λ = 1.5418 nm) and irradiated at a scan rate of 5°/min in the range of 5°–80°. The X-ray tube voltage and current were set at 45 kV and 50 mA, respectively. The morphology and microstructure of the samples were characterized by scanning electron microscopy (SEM, JEOL 6700-F) and transmission electron microscopy (TEM, JEOL JEM-2010). Inductively coupled plasma analysis (ICP, Agilent 7900) was conducted to determine the iodine doping amount of the as-prepared samples that were pretreated with chloroazotic acid and diluted for subsequent analysis. The UV-vis diffuse reflectance spectra (DRS) of the prepared samples in the range of 200–800 nm were recorded in the absorption mode using a PerkinElmer lambda 650s UV-vis spectrophotometer. The surface elemental composition of the sample was analyzed by X-ray photoelectron spectroscopy (XPS) using an XSAM800 system with Mg K_αX-ray as the excitation source, and all the binding energies were referenced to the C 1s peak at 284.6 eV of the surface amorphous carbon. The Raman spectrum was collected using a confocal microscope, and the wavelength of the excitation laser was 514 nm. The adsorption and desorption isotherms, pore size distribution, and specific surface area of nitrogen were measured at –196 ℃ using a surface area and pore size analyzer. For the electron spin resonance spectroscopy (ESR) measurements, 5, 5-dimethyl-1-pyrroline-N-oxide (DMPO) was employed as a free radical trapping reagent to detect active oxygen species generated in a photocatalytic system on a Magnettech MS-5000 EPR spectrometer under visible light. The TOC content was analyzed using a Shimadzu TOC-LCSH/CPH analyzer.

The photocatalytic activity of the sample was evaluated by the photocatalytic degradation of NaPCP. Photocatalytic degradation experiments were carried out using a simulated solar illumination device with a 300 W Xe lamp light source and a 400 nm cut-off filter. The constant temperature of 20 ℃ was maintained in the double glass container by circulating the condensate. A 25 mg sample of the as-prepared bismuth tungstate was dispersed in NaPCP (50 mL, 30x10^–6 mg/L). The catalyst suspension was subjected to magnetic stirring under a dark adsorption reaction condition for 60 min until adsorption–desorption equilibrium was attained. During the illumination, a 2 mL aliquot of the sample was centrifuged at 10 min intervals for the subsequent analysis. The concentration of NaPCP was analyzed by recording the signal intensity using a high-performance liquid chromatography (HPLC) instrument equipped with a UV detector at 227 nm. In the capture experiments, isopropanol (IPA), triethanolamine (TEOA), superoxide dismutase (SOD), CCl₄, and carotene were applied as scavengers for ^•OH, h⁺, ^•O₂⁻, e⁻, and ¹O₂, respectively.

All plane-wave density functional theory (DFT) calculations were performed using the Vienna Ab-initio Simulation Package (VASP) software [41, 42]. The generalized gradient approximation by Perdew, Burke, and Ernzerhof was used for the exchange-correlation energy. The ionic cores were described by projector-augmented wave potentials [43, 44]. All structural optimizations were conducted using a plane-wave expansion for the basis set with a cutoff energy of 380 eV. For the bulk Bi₂WO₆ geometry optimization, the atomic coordinates were fully relaxed with 3 × 1 × 1 Monkhorst-Pack grids. The stoichiometric surface of the (001) facet constructed by a 2 × 2 × 1 supercell (144 atoms) was used to model the surface of the experimentally prepared Bi₂WO₆, where the atomic positions of the top six layers of atoms were relaxed; however, the bottom layers were kept fixed at their bulk position. A vacuum region of 15 Å along the z-axis was used to eliminate the interactions between the neighboring cells of the slab models under periodic conditions. An iodine dopant was created on the 1 × 2 × 2 supercell by removing an O atom between two adjacent Bi atoms. For all the slab calculations, the Brillouin zones were sampled with 3 × 1 × 1 Monkhorst-Pack grids. The convergence with the force on each atom was set below 0.02 eV/Å, and the energy on each atom was within 1 × 10⁻⁵ eV for all calculations.

DFT calculations were firstly conducted to predict the impact of iodine doping on the OV formation over Bi₂WO₆ (Figure 1a). Considering the two cases of surface and bulk doping, the oxygen atoms on the surface and inside Bi₂WO₆ were replaced with iodine atoms, as the model for studying the difference in OV formation. For the Bi₂WO₆ catalyst without iodine, 2.94 eV is required to remove an oxygen atom forming the OVs. With surface iodine doping, the formation energy of the OVs in iodine-doped Bi₂WO₆ could be decreased to 1.20 eV. Although the OVs in the bulk iodine-doped Bi₂WO₆ require 2.41 eV, which is higher than that on the surface iodine-doped Bi₂WO₆, it still shows a decrease in the formation of OVs compared to the case of the pristine Bi₂WO₆, which may be attributed to the weakening effect of the Bi–O bonds owing to the introduced iodine dopant. The above results indicated that the iodine doping could decrease the formation energy of OVs, thus promoting the OV generation on Bi₂WO₆. Inspired by these calculation results, we used potassium iodide as the dopant to synthesize iodine-doped Bi₂WO₆. The crystalline phase of the as-prepared Bi₂WO₆ was first investigated by XRD studies. As shown in Figure 1b, the pure Bi₂WO₆ catalyst showed high agreement with the standard card (JCPDS No. 39-0256). It is worth noting that despite the introduction of the iodine dopant, only the diffraction peak corresponding to Bi₂WO₆ was observed on the patterns of BWO, BWO-0.06, BWO-0.12, BWO-0.18, and BWO-0.24, indicating that the iodine dopant has no effect on the formation of BiOI (JCPDS No. 11-0445). To further confirm the iodine introduction form, Raman spectroscopy was employed to study the lattice strain of the iodine-doped Bi₂WO₆. It was found that the intensity of the peak at 307 cm⁻¹, attributed to the Bi–O bending vibration, appreciably decreased with the increase in the I^– addition amount during the synthetic process, which may be ascribed to the decrease in the number of Bi–O bonds and crystallinity of Bi₂WO₆ caused by iodine doping [45].

The microstructures of the pure Bi₂WO₆ (BWO) and iodine-doped Bi₂WO₆ (BWO-0.18) catalysts were further investigated by SEM and TEM. As shown in Figures 2a and 2b, the SEM images showed that BWO and BWO-0.18 with stereoscopic three-dimensional structures were assembled by thick nanosheets. BWO is in the shape of a dense persimmon, while BWO-0.18 exhibits flower-like structures with more loose cavities. This loose structure of BWO-0.18 may be caused by the intervening iodine dopant introduced during the solvothermal reaction. The microstructure of the as-prepared material was also analyzed by HRTEM images (Figures 2c and 2d). Under HRTEM, both BWO and BWO-0.18 were observed to have legible orthorhombic lattice fringes with spaces of 0.272 and 0.273 nm (Figures 2c and 2d), respectively, ascribed to the (020) and (200) planes of orthorhombic Bi₂WO₆, thereby suggesting that both of them are indeed exposed with the {001} facets[46, 47]. To further confirm the presence of iodine, the elemental mapping and energy dispersive X-ray (EDX) spectroscopic analysis of BWO-0.18 were also carried out (Figures 2e and 2f). The distribution of the individual elements, Bi, W, O, and I, can be clearly observed, revealing that iodine was uniformly distributed in the BWO-0.18 structure. This was further confirmed by the result of ICP analysis displayed in Table 1, from which the iodine doping content was observed to gradually increase with the addition of iodine during synthesis.

The chemical environments of BWO and BWO-0.18 characterized by XPS are displayed in Figure 3a. The survey XPS implies the presence of Bi, W, O, and C elements in the two samples as well as an enhanced I 3d peak in BWO-0.18. To obtain the direct evidence of iodine doping, the I 3d time-dependent XPS analysis of BWO-01.8 was conducted. The peaks of I 3d located at 630.5 and 618.5 eV show no shift during the Ar⁺ sputtering, which demonstrated that the iodine dopants were homogeneously incorporated into the lattices of Bi₂WO₆ through this synthesis strategy. The high-resolution XPS spectra of Bi 4f (Figure 3c) exhibited two main peaks arising from 4f_5/2 at 164.5 eV and 4f_7/2 at 159.2 eV, indicating the characteristic +3 oxidation state of Bi [28]. There are two additional peaks that can be distinguished in the BWO-0.18 spectrum with lower binding energies at 162.5 and 157.2 eV, which can be ascribed to the relatively low valence states of oxygen vacancy-connected Bi atoms. The OV generation on iodine-doped Bi₂WO₆ was also confirmed via EPR measurement (Figure 3d). As expected, characteristic OV signals at g = 2.0001 were observed for all the solvothermal synthesized iodine-doped Bi₂WO₆ catalysts, which showed an enhanced OV signal with the increase in the iodine doping concentration, thereby signifying the successful introduction of OVs on the Bi₂WO₆ surface. From the N₂ adsorption-desorption isotherm measurement, the surface areas of BWO and BWO-0.18 were measured as 32.1 and 40.4 m²/g, respectively, by the Brunauer-Emmett-Teller specific surface area analysis (Figure 3e). Although the materials did not show much difference in physical adsorption, the chemisorption of BWO-0.18 presented a significant enhancement in O₂ chemisorption in comparison to that presented by BWO (Figure 3f), which is widely believed to has a catalytic effect on the molecular oxygen activation [48]. According to the above results, we can infer that OVs were successfully introduced into Bi₂WO₆ via this facile solvothermal method. In addition, the amount of OVs in Bi₂WO₆ increased significantly with the introducing of iodine dopants, which could provide a more convenient environment for molecular oxygen activation.

Generally, the absorption of light is the first step in photocatalytic reactions. The iodine-doped Bi₂WO₆ catalyst exhibited a trailing absorption in comparison to that of BWO around 450 nm to 700 nm, as shown in the UV-Vis absorption spectra (Figure 4a), which could be attributed to the sub-band excitation from defect states to the conduction band (CB) induced by the introduction of OVs and iodine dopant. The bandgap energies of BWO and BWO-0.18 were calculated to be 2.87 and 2.74 eV, respectively (inset of Figure 4a), implying an enhancement in the adsorption of BWO-0.18 in the visible light region. The degradation process of NaPCP requires the involvement of a large amount of ROS; thus, it was selected as the degradation target to evaluate the photocatalytic molecular oxygen activation ability of the as-prepared materials. As shown in Figure 4b, we evaluated and compared the photocatalytic activities of the as-prepared Bi₂WO₆ materials. Notably, all of the iodine-doped Bi₂WO₆ catalysts exhibited excellent photocatalytic performance for NaPCP degradation in comparison to that of the pure Bi₂WO₆. Among these iodine-doped Bi₂WO₆ samples, BWO-0.18 possessed the highest photocatalytic activity, as it could almost completely remove NaPCP within 60 min. It should be pointed out that a high number of OVs can be introduced into Bi₂WO₆ with the continual addition of the iodine dopant, which may serve as recombination centers and hinder the separation of photogenerated carriers. Therefore, the photocatalytic performance of iodine-doped Bi₂WO₆ first gradually increased with the iodine dopant addition and then decreased, thereby implying that BWO-0.18 possesses the optimal doping ratio. The photocatalytic kinetic simulation indicates that the rate constant of the NaPCP degradation over BWO-0.18 was 268 times higher than over BWO, indicating that the visible light photocatalytic activity of BWO could be notably enhanced by iodine doping induced OVs (Figure 4c). In addition, the stability of BWO-0.18 was also evaluated by the photocatalytic degradation of NaPCP under visible light. Within five consecutive runs, the BWO-0.18 catalyst still exhibited a relatively high removal rate, indicating the good stability of BWO-0.18 (Figure 4d). The TOC removal rate is an important indicator to investigate the complete oxidation of NaPCP. As shown in Figure 4e, the TOC content removal rates in the solution over BWO, BWO-0.18, and BWO-0.24 were evaluated. Consistent with the degradation efficiency, BWO-0.18 exhibited a TOC remove rate almost 12 times that of BWO, and the removal rate exceeded 90% in 2 h, thereby demonstrating the deep oxidation of NaPCP with our iodine-doped Bi₂WO₆ employed as photocatalysts. Subsequently, we also evaluated the role of each radical in the process of degradation by adding scavengers. As shown in Figure 4f, the scavenger of triethanolamine (TEOA) for photogenerated holes completely inhibited the photodegradation of NaPCP, while the scavenger of isopropanol (IPA) for hydroxyl radicals slightly affected its photodegradation. The other scavengers for electrons, ^•O₂⁻ and ¹O₂, displayed relative high inhibition of the photodegradation, thereby indicating that the holes, ^•O₂⁻ and ¹O₂, were the main oxidative species in the photocatalytic degradation process.

To obtain more solid evidence for the advantages in enhancing the molecular oxygen activation over BWO-0.18, the EPR spectra were further employed as the detector for generated ROS in the presence of a spin trapping agent. As shown in Figure 5a, the enhanced characteristic peaks of DMPO-^•OH can be observed in the solution with BWO as the catalyst after irradiation, while under the same condition of irradiation, BWO-0.18 did not show noticeably enhanced signals of DMPO-^•OH. For DMPO-^•O₂⁻ (Figure 5b), both of the catalysts evidently possess relatively strong DMPO-^•O₂⁻ signals under visible light irradiation, although the enhancement over BWO-0.18 was more pronounced. As illustrated in Figure 5c, visible light irradiated BWO-0.18 showed an enhanced triplet signal, which can be indexed to ¹O₂. Compared to the sample without irradiation, the enhancement of the ESR intensity in BWO-0.18 was much higher than that in BMO, which is attributed to the stronger ability of ¹O₂ generation. Generally, ¹O₂ could be converted via two different pathways: from the charge transfer process, in which ¹O₂ generates from the ^•O₂⁻ species oxidized by holes, and from the energy transfer, in which O₂ absorbs the energy generated by electron–hole recombination and transforms into ¹O₂ directly. Therefore, the contributions of the charge and energy transfer processes for ¹O₂ generation in BWO-0.18 were subsequently distinguished by selecting superoxide dismutase (SOD) and nitrogen (Figure 5d). With the addition of SOD or in nitrogen atmosphere, BWO-0.18 did not change the ¹O₂ concentration in response to irradiation, and both of them held a relatively low concentration of ¹O₂, indicating that the ¹O₂ species is mainly produced through the electron transfer pathways for BWO-0.18. Furthermore, using the VB-XPS and Mott-Schottky plot (Figure 5e), the valence band (VB) positions of BWO and BWO-0.18 could be directly obtained by the equation based on our previously reported work, and the CB positions of BWO and BWO-0.18 could also be determined combined with the obtained bandgap value [20]. Therefore, the VB positions of BWO and BWO-0.18 were calculated to be 2.03 and 1.66 eV, respectively. Accordingly, the CB positions of BWO and BWO-0.18 were calculated to be –0.84 and –1.07 eV, respectively. As shown in Figure 5f, the VB of BWO is positive enough for ^•OH generation, while the VB of BWO-0.18 does not exhibit enough potential to produce ^•OH. In addition, the CB values of BWO and BWO-0.18 are both negative enough to transfer the electron on them to the absorbed oxygen, thereby achieving the molecular oxygen activation. Moreover, BWO-0.18 possessed a highly negative CB position and relatively large number of OVs; therefore, the molecular oxygen activation of BWO-0.18 is much stronger than that of BWO. These are highly consistent with the detection of radicals by EPR [49].

Based on the above results and discussion, a reasonable mechanism of the action of the iodine doped Bi₂WO₆ is proposed in Scheme 1. First, the iodine dopants combined with the Bi atoms, thus decreasing the energy of the Bi–O bonds, which promoted the generation of OVs on the Bi₂WO₆ surface. The introduced OVs in iodine-doped Bi₂WO₆ could not only localize photogenerated electrons but also act as active sites for molecular oxygen activation. With the coordination of OVs and iodine doping, the optimized iodine-doped Bi₂WO₆ (BWO-0.18) exhibited enhanced molecular oxygen activation, producing a large number of ROS under light irradiation. With the enhanced molecular oxygen activation induced by OVs, the TOC removal rate of BWO-0.18 was 10.6 times higher than that of BWO under visible light irradiation.

A novel iodine-doped Bi₂WO₆ catalyst bearing OVs was synthesized via a simple solvothermal method. By adjusting the iodine dopant concentration, a high density of oxygen vacancies was successfully introduced by homogeneous iodine doping. The as-prepared Bi₂WO₆ catalyst was characterized by theoretical calculations and various characterization techniques, from which the OVs in Bi₂WO₆ were confirmed to be intensely correlated with the iodine dopant due to its weakening effect on lattice oxygen bonds. The high activity and TOC removal efficiency in the photocatalytic degradation of NaPCP were attributed to the molecular oxygen activation induced by the OVs. Moreover, the mechanism through which the degradation of NaPCP over iodine-doped Bi₂WO₆ is enhanced, was speculated by ROS detection. This work provides a new perspective for the enhanced photocatalytic degradation of organochlorine pesticides via OV-induced molecular oxygen activation over iodine-doped Bi₂WO₆ catalysts.