With the rapid development of economy, the environmental pollution issue has become increasingly prominent. For example, the phenols chemicals such as 4-chlorophenol has been massive discharged due to the industrial by-product or by chlorination, which was potentially carcinogenic and mutagenic to human [1]. The plentiful employment of antibiotics result in the excess emission as active form pharmaceutically, which aggravate biological drug resistance and affect the ecosystem. Moreover, these phenols chemicals and antibiotics were stubborn and difficult to entirely remove through the conventional treatment methods. Semiconductor photocatalysis has attracted massive attention towards environmental purification [2]. The critical point to acquire high-activity lie on the strong oxidation ability of semiconductor under light excitation. Although the semiconductor materials such as g-C₃N₄, W₁₈O₄₉, BiOX, have been widely developed [3-7], the efficiency was limited owning to the high potential of valence band (VB) which cannot endow the powerful oxidizing ability.

As one of the oxyacid salt photocatalysts, BiPO₄ displayed numerous advantages including photoelectric properties, low cost, nontoxicity, and outstanding photocatalytic efficiency [8, 9]. Due to the deeper VB position, BiPO₄ can display more attractive activity than many photocatalysts towards pollutant removal. To optimize the photocatalytic behavior of BiPO₄ materials, many strategies have been employed such as heteroatom doping [10, 11], defect adjusting [12-14], phase junction [15], surface modification [16-18] and semiconductors coupling [19-23]. Although the photocatalytic activity of the obtained BiPO₄ materials can be moderate increased, it was still of great importance and urgency to explore suitable strategy to further improve the photocatalytic performance for practical applications.

Previous studies demonstrated that engineering defects into materials can effectively improve the photocatalytic activity [24-30]. For example, Li's group found that the formation of "Bi-O" dimer vacancy pairs could result in the reduction in band gap and reduce the recombination rate of photogenerated charge carriers, which were responsible for the improved photocatalytic performance for pollutant removal [24]. Xiong's group found that the construction of oxygen defects on WO₃ nanosheets was beneficial to oxygen chemisorption and thus facilitate the formation of superoxide radicals [25]. As a result, the photocatalytic organic synthesis of amines to corresponding imines can be greatly promoted. Therefore, defects engineering seems to be a desirable method to tune the photocatalytic performance of BiPO₄. However, the defects can not only serve as separation center for photogenerated charge to boost the photocatalytic performance but also can work as recombination center to decrease the photocatalytic activity. It needs further study whether the higher defect concentration can inevitably bring about higher activity.

In this system, BiPO₄ materials with different oxygen vacancy concentration have been controlled prepared to study the affection of oxygen vacancy towards photocatalytic behavior of BiPO₄. It is worth noting that BiPO₄ materials with lower oxygen vacancy concentration can show higher photocatalytic activity for pollutant removal. The mechanism for the improved photocatalytic performance of BiPO₄ was also investigated.

X-ray diffraction (XRD) analysis was employed to study the crystallinity and phase structure of the as-synthesized samples and the result was shown in Fig. 1a. The diffraction peaks of the as-synthesized sample match well with monoclinic BiPO₄ (JCPDS card No. 15-0767), with the peaks located at 2θ = 11.9°, 21.4°, 25.4°, 27.1°, 29.1°, 31.3°, 34.5°, 37.0°, 41.9°, 42.8°, 46.4° and 52.8° can be indexed to the (011), (111), (111), (200), (120), (012), (202), (112), (130), (131), (212) and (040) crystal planes. After further vacuum activation or annealing, no distinct variation can be observed relative to that of BiPO₄, revealing the further treatment did not destroy crystal structure. The detailed structural information of the prepared BiPO₄, BiPO₄-vacuum, BiPO₄-calcine materials were further determined by Fourier transform infrared (FT-IR). It can be seen from Fig. S1 (Supporting information) that absorption peaks centered at 524 and 554 cm⁻¹ were attributed to the υ4 vibrational behavior of PO₄³⁻ ions, while the peaks of υ3 stretching modes can be seen from 900 to 1100 cm⁻¹ [31]. The similar FT-IR profiles of BiPO₄, BiPO₄-vacuum, BiPO₄-calcine materials suggested the successful preparation of BiPO₄ and main structure was well maintained.

To determine the oxygen vacancy in the prepared different BiPO₄ samples, low-temperature electron paramagnetic resonance (EPR) analysis was carried out. As shown in Fig. 1b, the typical signal at g = 2.001 can be found for these three materials which powerful certify the existence of oxygen vacancies [32]. The 1, 3-butanediol could easily react with the oxygen in the BiPO₄ to produce oxygen vacancies during the synthetic process, and the similar result can also be found in other systems [33]. It is noted that the BiPO₄ displayed higher intensity than BiPO₄-calcine sample and the BiPO₄-vacuum materials exhibited the highest signal. This result indicated that the vacuum activation process can further create oxygen vacancies on BiPO₄, while annealing at atmosphere may repair the oxygen vacancies.

X-ray photoelectron spectroscopy (XPS) was employed to determine the surface composition of the obtained samples. The survey spectrum (Fig. 2a) of the obtained BiPO₄ samples were consist of Bi, P, O and C elements, in which the existence of C was ascribed to adventitious carbon species from the XPS instrument. From high-resolution Bi 4f spectra (Fig. 2b), the peaks centered at 159.7 and 164.9 eV corresponded to the Bi 4f_7/2 and Bi 4f_5/2 of Bi³⁺ in BiPO₄ materials. In high-resolution O 1s spectra (Fig. 2c), the peak located at 530.9 eV was assigned to Bi-O bonds, and the peak at 532.4 eV was derived from O-atoms in the vicinity of an O vacancy [34]. The area corresponded to O vacancy showed difference among these three materials. The BiPO₄-vacuum sample showed the largest area and BiPO₄ was larger than that of BiPO₄-calcine, indicating the oxygen vacancy can be further increased by vacuum activation and annealing decrease the oxygen vacancy. The binding energy at 133.1 eV can be assigned to P 2p peak in BiPO₄ materials (Fig. 2d). The above results suggested the BiPO₄ materials with different oxygen vacancy concentration have been controlled prepared.

To study the affection of oxygen vacancy on the light absorption ability, the UV-vis diffuse reflectance spectra (DRS) was carried out on the BiPO₄, BiPO₄-vacuum and BiPO₄-calcine samples. As shown in Fig. S2a (Supporting information), BiPO₄ materials with different oxygen vacancy concentrations displayed obvious difference of the light absorption region. The BiPO₄-calcine materials with the lowest oxygen vacancy concentration showed the photoabsorption of 200–300 nm, indicating the BiPO₄ was intrinsic UV-light response photocatalyst. With the increase of oxygen vacancy concentration, the light absorption region extended to visible and infrared region, and this phenomenon was similar to that in literatures [25, 28]. The band gap energies of these BiPO₄ samples can be thus calculated by the Kubelka-Munk function. As shown in Fig. S2b (Supporting information) the band gap of BiPO₄-calcine, BiPO₄, and BiPO₄-vacuum materials can be estimated to be 4.26, 3.89 and 3.66 eV, respectively.

The morphology and microstructure of the prepared BiPO₄ materials with different oxygen vacancy concentrations were observed by transmission electron microscopy (TEM) images. From Figs. 3a and b, it can be seen that the BiPO₄ sample showed the nanoparticle morphology with the average size about 40 nm. After further annealing (Figs. 3c and d) to decrease the oxygen vacancy or vacuum activation (Figs. 3e and f) to increase the oxygen vacancy concentration, the morphology did not show distinct variation. This result demonstrated that the morphology difference may be not the main factor to affect the photocatalytic performance in this system.

As is well known, the specific surface area of photocatalysts have important influence on the photocatalytic performance due to the larger specific surface area could adsorb more pollutant and active species to form a partial high concentration [35]. The BET specific surface area of BiPO₄, BiPO₄-calcine and BiPO₄-vacuum materials were investigated by nitrogen adsorption-desorption isotherms. As shown in Figs. S3–S5 (Supporting information), all the samples displayed the type Ⅳ isotherms with a H3 hysteresis loop. The BET surface areas of BiPO₄, BiPO₄-calcine and BiPO₄-vacuum samples were measured to be 28.8, 21.8 and 23.7 m²/g, respectively. The annealing and vacuum activation treatment did not have significant influence on the BET surface areas and it can be assumed that the specific surface area may be not the main activity-determining factor in this system.

It has been reported that the formed oxygen vacancies could serve as separation centers to promote the separation of photogenerated electrons and holes [12, 32]. In this system, the steady photoluminescence (PL) spectroscopy was employed to study the charge separation efficiency of the obtained BiPO₄ materials. As shown in Fig. S6a (Supporting information), the BiPO₄-vacuum samples displayed decreased PL intensity than that of BiPO₄-calcine samples, suggesting the lower recombination rate of charge carriers in BiPO₄-vacuum samples [36]. Since the high oxygen vacancy concentration in BiPO₄-vacuum materials, the oxygen vacancies could effectively facilitate the separation of charge carriers and thus ensure the decreased PL intensity. To further insight the charge transfer and recombination processes in the BiPO₄ materials, electrochemical impedance spectroscopy (EIS) was carried out and the result was shown in Fig. S6b (Supporting information). The Nyquist arc radius in the EIS analysis revealed the solid state interface layered resistance and the surface charges transfer resistance [37]. The arc radius of BiPO₄-vacuum was smaller than that of BiPO₄-calcine sample, meaning the BiPO₄-vacuum possess smaller resistance than that of BiPO₄-calcine and thus could achieve the higher charge separation efficiency.

The photocatalytic performance of the obtained BiPO₄ materials with different oxygen vacancy concentrations was evaluated by the degradation of colourless antibiotic agent ciprofloxacin (CIP). Massive emission of CIP into aquatic environments will aggravate antibiotic resistance and affect the ecological environment. It is highly desirable to found effective method to remove the CIP from aquatic environments. As shown in Fig. 4a, 78% of CIP can be removed by BiPO₄ sample under UV light irradiation for 120 min, indicating the CIP can be degraded by BiPO₄ materials. However, when the oxygen vacancies were further created in BiPO₄ by vacuum activation, the photocatalytic performance decreased for the removal of CIP. Only 67% of CIP can be degraded by BiPO₄-vacuum sample within the same time under UV light irradiation. Through annealing to further decrease the oxygen vacancy concentration, the photocatalytic degradation efficiency can be improved to 94.4% for CIP within 120 min. The result of total organic carbon experiment shows 63% of CIP can be mineralized after 4 h irradiation over BiPO₄-calcine, revealing the CIP can be effective degraded by BiPO₄-calcine materials. The result of photocatalytic degradation of CIP suggested that the lower oxygen vacancy concentration in BiPO₄ could bring about higher photocatalytic activity, which was inconsistent with the results in previous reports. Moreover, no obvious activity loss can be found after 4 times cycles, suggesting good stability of BiPO₄-calcine for photodegradation (Fig. 4b). To further prove the activity law, the photocatalytic behavior of the obtained materials was evaluated by the removal of 4-chlorophenol (4-CP). As shown in Fig. 4c, 50.4%, 68.3% and 95.3% of 4-CP can be removed by BiPO₄-vacuum, BiPO₄ and BiPO₄-calcine samples, respectively. The BiPO₄-calcine materials with the lowest oxygen vacancy concentration also displayed the optimal photocatalytic activity. In addition to photodegradation, the performance of the prepared samples was evaluated by CO₂ photoreduction (Fig. 4d). The BiPO₄-calcine showed higher photoreduction activity to yield CO, with the generation rate arrive 7.1 µmol g⁻¹ h⁻¹, higher than that of BiPO₄ (4.4 µmol g⁻¹ h⁻¹) and BiPO₄-vacuum (3.2 µmol g⁻¹ h⁻¹). Since the BiPO₄-vacuum, BiPO₄ and BiPO₄-calcine samples have similar morphology and specific surface area, the morphology and specific surface area difference may be not the main factor to determine the ultimate photocatalytic performance. Moreover, the BiPO₄-vacuum displayed higher charge separation efficiency than BiPO₄-calcine, which was also do not agree with the photocatalytic result.

To gain further insight into the intrinsic reason of the materials with lower oxygen vacancy concentration exhibit the higher photocatalytic activity, the XPS valence spectra was performed. It can be seen from Fig. 5a that the BiPO₄-vacuum, BiPO₄ and BiPO₄-calcine samples displayed the maximum energy edge of the VB at about 2.69, 2.56 and 2.44 eV, respectively. According to the band gap obtained from DRS analysis and the formula E_CB = E_VB − E_g, the conduction band (CB) edge of BiPO₄-vacuum, BiPO₄ and BiPO₄-calcine samples can be estimated to be −0.97, −1.34 and −1.82 eV, respectively. The BiPO₄-calcine materials showed the up-shifted CB and VB edges when compared with BiPO₄ and BiPO₄-vacuum (Fig. 5b). The more negative CB position favors the CO₂ reduction to yield CO. The upshifting of the CB edge could generate more reductive photogenerated electrons to react with molecular oxygen to yield O₂^·−. The O₂^·− was one of the important active species which can effectively degrade the pollutants. In addition, the BiPO₄-calcine showed the improved VB width relative to BiPO₄ and BiPO₄-vacuum. The VB width could intrinsically dominate the mobility of the photogenerated holes. The wider VB width could bring about faster mobility of holes and thus the better photo-oxidation efficacy can be achieved [38, 39]. As a result, the BiPO₄-calcine could display the increased photocatalytic activity.

To further determine the active species produced during the photodegradation process, the ESR spin-trap technique was performed. The 5, 5-dimethyl-1-pyrroline N-oxide (DMPO) was used as radicals trapping agent to generate the stable DMPO-O₂^·− or DMPO-^·OH in methanol and water, respectively. As shown in Fig. 6a, the signal of DMPO-O₂^·− can be observed over BiPO₄-calcine materials under UV light irradiation which indicated the O₂^·− can be generated over BiPO₄. Considering the E⁰(O₂/O₂^·−) was about −0.33 eV, the electrons on the CB were negative enough to activate the oxygen via single electron transfer. At the same time, the fourfold peaks with the intensity ratio 1:2:2:1 can be seen, which was attributed to the DMPO-^·OH signal (Fig. 6b). The E⁰(^·OH/OH⁻) was 2.38 eV vs. NHE, thus the holes on the VB of BiPO₄-calcine could oxidize OH⁻ to yield ^·OH. Therefore, the excellent photocatalytic activity of BiPO₄-calcine (with the lowest oxygen vacancy concentration) was derived from the synergistic effect of O₂^·−, ^·OH and holes. Taking the above-mentioned factors into consideration, the energy band structure variation may be the main factor dominating the photocatalytic performance in this system.

In summary, BiPO₄ materials with different oxygen vacancy concentrations have been controlled prepared through 1, 3-butanediol assisted solvothermal method, vacuum chemical activation and annealing processes. The BiPO₄-calcine with the lowest oxygen vacancy concentration displayed the optimal photocatalytic performance for the degradation of CIP and 4-CP and CO₂ photoreduction to yield CO. According to different characterizations, the enhanced photocatalytic activity can be ascribed to the varied energy band structure with both the upshifting of CB and VB positions. This study can be extended to prepare other photocatalysts with controllable oxygen vacancy concentration and may provide some insight of the mechanism.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

This work was financially supported by the National Natural Science Foundation of China (No. 22002014), the Funding for scientific research startup of Jiangsu University (No. 20JDG15), Fundamental Research Funds for the Central Universities (No. 30922010302) and Start-Up Grant (No. AE89991/397) from Nanjing University of Science and Technology.

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2022.107844.