Nitric oxide (NO) has been confirmed as a main source of atmospheric pollution, which is harmful to water sources and soil due to formation of acid rain [1, 2]. Moreover, long-term inhalation of NO can damage the respiratory tract of humans and cause serious diseases. In recent decades, the concentration of NO in the atmosphere increased considerably because of the rapid development of the modern industrial civilization.The situation is of concern and requires effective methods to remove NO or oxidize it to final harmless or non-hazardous products [3-5]. The products of NO oxidation are NO₃^- or NO₂, which are commonly converted into useful industrial materials. Recently, many researchers have demonstrated that photocatalytic technologies are efficient methods for atmospheric pollution remediation [6-9]. The photocatalytic NO removal is still an oxidation process. The core of this technology, the oxidation efficiency of the catalysts, is of significant importance for the photocatalytic NO removal.

Because of the outstanding oxidation capability under visible light, layered bismuth oxyhalide (BiOX, X=Br, I and Cl) photocatalysts displayed efficient photocatalytic activity for NO removal [10-13]. In order to further strengthen its photocatalytic property, some modified structural bismuth oxyhalide catalysts were designed and showed enhanced photocatalytic activity [14-16]. The combination of solid solution and heterojunction was suggested as effective strategy to enhance the molecular oxygen activation via exciton and carrier photocatalysis. For example, the solid solution BiOX_xY_1-x (X and Y are the different halogens) and the heterojunction BiOX/BiOY (X and Y are the different halogen) were most commonly synthesized and researched [17-19]. However, the synergistic effect of solid solution and heterojunctions is usually applied to the photocatalytic production of hydrogen, using sulphided catalysts such as Zn_1–xCd_xS/CdS [20, 21] and Fe_1-xPt_x/ZnCdS [22]. It is rarely applied for photocatalytic NO removal. In our previous work, we demonstrated that the solid solution BiOBr_0.5I_0.5 and heterojunction BiOBr/BiOI have enhanced photocatalytic activity for molecular oxygen activation due to their special structures [23]. When irradiated by visible light, the excitation initiates different photocatalytic molecular oxygen activation reactions for BiOBr_0.5I_0.5 and BiOBr/BiOI. Singlet oxygen (¹O₂) and superoxide (·O₂^-) are generated by BiOBr_0.5I_0.5 and BiOBr/BiOI via exciton and carrier photocatalysis, respectively. The enhanced molecular oxygen activation capacity supports their superior photocatalytic ability under visible light [23, 24]. However, it was uncertain whether the synergistic effect of solid solution and heterojunctions can be used for photocatalytic NO removal.

In this study, in order to improve the photocatalytic property of BiOX (X = Br and I) for NO removal, we used a one-pot solvothermal method to design and prepare the solid solution and heterojunction containing BiOBr_0.5I_0.5/BiOI catalyst. The basic structure of BiOBr_0.5I_0.5/BiOI has been confirmed by X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), and transmission electron microscopy (TEM). The molecular oxygen activation enhanced by the synergistic effect of solid solution and heterojunction resulted in the ameliorative NO removal capacity of BiOBr_0.5I_0.5/BiOI. In addition, the mechanism for photocatalytic NO removal, including the intermediate processes and products, was also explored and confirmed. To the best of our knowledge, this work should be the first to report on the synergistic effect of solid solution and heterojunction for photocatalytic NO removal. It provides a constructive thought to enhance the photocatalytic activity of BiOX (X = Br, I and Cl) to remove NO.

The characteristic diffraction peaks of BiOBr_0.5I_0.5/BiOI, BiOBr_0.5I_0.5, BiOBr, and BiOI were determined by XRD. As is shown in Fig. 1(a), consistency was observed between the characteristic peaks of BiOBrand BiOI and the referenced values (BIOBr: PDF: 01-078-0348; BiOI: PDF: 01-073-2062) [25]. The diffraction peaks of BiOBr_0.5I_0.5 located between those of BiOI and BiOBr confirmed its solid solution structure [26, 27]. All diffraction peaks of BiOBr_0.5I_0.5 and BiOI appeared also in the BiOBr_0.5I_0.5/BiOI sample, which implied that the solid solution and heterojunctions were successfully formed [28-30]. The amplified XRD information (Fig. 1(b) and (c)) also confirmed the co-existing structure. Fig. 1(d) shows UV-Vis diffuse reflectance spectra (DRS) comparing the maximum absorption wavelengths. The maximum absorption wavelengths of BiOBr_0.5I_0.5, BiOI and BiOBr_0.5I_0.5/BiOI were 701.5, 679.5 and 691.5 nm, respectively. This indicated that these three samples all responded to visible light [31, 32] and that visible light supports NO removal. Physical nitrogen adsorption determined the specific surface area of BiOBr_0.5I_0.5/BiOI, BiOBr_0.5I_0.5 and BiOI. Figure 1(e) shows that the N₂ adsorption and desorption curves all correspond to type IV adsorption isotherms [33]. The specific surface area of BiOBr_0.5I_0.5/BiOI (35.0 m²·g^-1) was larger than that of BiOBr_0.5I_0.5 (8.2 m²·g^-1) and BiOI (14.4 m²·g^-1). Figure 1(f) shows the pore distribution curves of the samples. The peak pore diameters of the three samples were all close to 2.47 nm. The high surface area and abundant pore structure of BiOBr_0.5I_0.5/BiOI favor the photocatalytic NO removal [34].

The elemental compositions of the BiOBr_0.5I_0.5/BiOI, BiOBr_0.5I_0.5 and BiOI were confirmed by XPS. Figure 2(a) shows that the surfaces of BiOBr_0.5I_0.5/BiOI and BiOBr_0.5I_0.5 consist of the same elements Bi, O, Br, and I, while those of BiOI were Bi, O and I. The Bi 4f_5/2 binding energies of BiOBr_0.5I_0.5 and BiOI were 164.56 eV, and the Bi 4f_7/2 binding energy was 159.27 eV (Fig. 2(b)). The binding energies of BiOBr_0.5I_0.5 and BiOI at 630.73 and 619.20 eV were assigned to I 3d_3/2 and I 3d_5/2 (Fig. 2(c)), respectively. However, the Bi 4f and I 3d binding energies of BiOBr_0.5I_0.5/BiOI deviated by 0.1–0.2 eV compared to the single components which indicated that there were interactions between BiOBr_0.5I_0.5 and BiOI. The Br 3d_3/2 and Br3d_5/2 binding energies of BiOBr_0.5I_0.5/BiOI and BiOBr_0.5I_0.5 were 68.4 and 69.4 eV (Fig. 2(d)). Figure 2(e) shows the O 1s pattern whose binding energies were 530.0 and 530.8 eV for BiOBr_0.5I_0.5, 530.1 and 531.0 eV for BiOI, and 529.8 and 530.7 eV for BiOBr_0.5I_0.5/BiOI, respectively. It was crystal lattice O atoms (Bi-O) and surface hydroxyl groups that caused the two energy bands [35, 36]. The binding energy of 532.3 eV for BiOBr_0.5I_0.5/BiOI indicated adsorbed surface water on BiOBr_0.5I_0.5/BiOI [35, 36]. Based on the XPS valance band and plots of (α(hν))^1/2 versus energy (hν) of BiOBr_0.5I_0.5 and BiOI, the band structure of BiOBr_0.5I_0.5/BiOI is shown in Fig. S1. In Fig. S1(a), the valance band of BiOBr_0.5I_0.5 and BiOI were 1.37 and 1.64eV. The band gap values of BiOBr_0.5I_0.5 and BiOI were 1.90 and 1.99 eV (Fig. S1(b)). It was calculated that the conduction band values of BiOBr_0.5I_0.5 and BiOI were -0.53 and -0.35 eV. Combining the above experimental results, the band structure of BiOBr_0.5I_0.5/BiOI is shown in Fig. S1(c).

High-angle annular dark field (HAADF) image, transmission electron microscopy (TEM), high-resolution transmission electron microscopy (HRTEM) and element mapping of BiOBr_0.5I_0.5/BiOI are presented in Fig. 3. HAADF (Fig. 3(a)) and TEM (Fig. 3(b)) images show that BiOBr_0.5I_0.5/BiOI aggregated in the form of nanosheets. In the HRTEM image of BiOBr_0.5I_0.5/BiOI (Fig. 3(c)), there was an obvious boundary at the junction, which implied the heterostructure formed by coupling BiOI with BiOBr_0.5I_0.5. In addition, two interplanar distances of 0.18 and 0.29 nm were observed. Based on the crystal structure parameters of bismuth oxyhalide [37-39], 0.18 and 0.29 nm were attributed to the {202} plane of BiOI and {012} plane of BiOBr_0.5I_0.5. It validated the formation of the heterojunction between BiOI and BiOBr_0.5I_0.5. Another HRTEM image of BiOBr_0.5I_0.5/BiOI also confirmed it. As shown in Fig. S2, the distances of 0.27 and 0.33 nm were attributed to the {111} plane of BiOI and {011} plane of BiOBr_0.5I_0.5. In addition, comparing the corresponding element mapping, it was observed that the signal of Br is relatively weak in the elliptic marked region (Fig. 3(d)). It indicated that this region corresponds to BiOI, and the other regions are related to BiOBr_0.5I_0.5. It was concluded that the binary solid-heterostructural BiOBr_0.5I_0.5/BiOI was formed successfully.

Figure 4 displays the results of the photocatalytic NO removal experiments of BiOBr_0.5I_0.5/BiOI, BiOBr_0.5I_0.5 and BiOI. Comparing their photocatalytic NO removal efficiency (Fig. 4(a)), the enhanced photocatalytic NO oxidation properties of BiOBr_0.5I_0.5/BiOI were obvious, where 36.2% NO was removed after a 20-min irradiation under visible light. The activity was higher than most of the reported results (Table 1). At the same conditions, the NO removal efficiency of BiOBr_0.5I_0.5 and BiOI was just 22.3% and 7.0%, respectively. The BET specific surface areas were also determined. The comparison of BiOX based NO removal photocatalysts is presented in Table 1. The BiOBr_0.5I_0.5/BiOI photocatalyst displayed a better activity than reported BiOX based photocatalysts. The BET areas of BiOBr_0.5I_0.5/BiOI, BiOBr_0.5I_0.5 and BiOI were 9.4, 3.5 and 2.8 m²·g^-1, respectively. Noticeably, the samples with higher specific surface area display better activity. It implied that there is a significant influence of BET specific surface area on photocatalytic activity. The stability of BiOBr_0.5I_0.5/BiOI was confirmed after repeating experiments three times (Fig. 4(b)). The results established that the considerable photocatalytic NO removal stability of BiOBr_0.5I_0.5/BiOI of approximately 36% NO remained in cyclic experiments. In addition, the recorded TEM image (Fig. S2) of BiOBr_0.5I_0.5/BiOI after three cycles did not show obvious changes. NO degradation products could affect the photocatalytic process. Figure 4(c) shows the amount of NO₂ generated from NO degradation during the three cycles. It was observed that NO degradation and NO₂ generation was stable during the cycling experiments. Moreover, the NO degradation amount was approximately equal to the generated NO₂ amount, which implied that the main product of photocatalytic NO removal with BiOBr_0.5I_0.5/BiOI was NO₂. This is also in agreement with previous reports, where NO reacted with reactive oxygen species (ROS) to yield NO₂ by the selective oxidation process [40].

According to our previous work, heterojunction and solid solution structures efficiently improve molecular oxygen activation carrier and exciton photocatalytic process [20, 21]. In order to confirm that the synergistic effect of solid solution and heterojunction of BiOBr_0.5I_0.5/BiOI leads to molecular oxygen activation enhancement, we verified the carrier and exciton photocatalytic processes separately. The separation efficiency of holes and photogenerated electrons should be improved in order to obtain a better photocatalysis of carriers from heterostructures. Photoluminescence (PL) is a method to assess the photo-generated electron-hole pair efficiency. Figure S5 shows PL spectra of BiOBr_0.5I_0.5/BiOI with a weaker intensity than those of BiOBr_0.5I_0.5 and BiOI, which indicated that the carrier photocatalysis could be enhanced at a lower electron-hole recombination rate by BiOBr_0.5I_0.5/BiOI [46]. Compared with BiOBr_0.5I_0.5 and BiOI, the higher photocurrent intensity (Fig. 5(a)) and smaller diameter semicircle arc (Fig. 5(b)) of BiOBr_0.5I_0.5/BiOI was consistent with the enhanced carrier photocatalysis [47, 48]. It was also in agreement with the PL results.

The typical ROS measurements were the most direct evidence to confirm the molecular oxygen activation in the photocatalytic process. It was reported that there are two typical main products in the processes of exciton and carrier photocatalysis, singlet oxygen (¹O₂) and the superoxide radical (·O₂^-) [49, 50]. A higher generation of ¹O₂ and ·O₂^- implies the enhanced ability of molecular oxygen activation. ESR spectroscopy together with the compounds, 3', 3', 5, 5'-tetramethylbenzidine (TMB) and nitrobluetetrazolium (NBT) were used to evaluate the generation of ¹O₂ and ·O₂^- [51, 52]. As is shown in Fig. 5(c), for ·O₂^-, BiOBr_0.5I_0.5/BiOI displayed stronger ESR signals than BiOBr_0.5I_0.5 and BiOI. It indicated that BiOBr_0.5I_0.5/BiOI produced more ·O₂^- under visible light irradiation via enhanced carrier photocatalysis due to the heterostructure. With regards to the ¹O₂ generation, Fig. 5(d) reveals a stronger ESR signal for BiOBr_0.5I_0.5/BiOI than for BiOBr_0.5I_0.5 and BiOI. The experimental results suggested that the solid solution structure leads to more ¹O₂ produced by BiOBr_0.5I_0.5/BiOI under visible light via the enhanced exciton photocatalysis. In addition, the ROS quantitative experiments also support these results. Figs. 5(e) and S6 show that BiOBr_0.5I_0.5/BiOBr/BiOI decreased the NBT (53%) in the absorption spectra more than BiOBr_0.5I_0.5 (28%) or BiOI (13%). Figures 5(f) and S7 reveal a higher TMB oxidation rate for BiOBr_0.5I_0.5/BiOI (65%) compared to BiOBr_0.5I_0.5 (27%) and BiOI (10%). This suggested that the solid solution structure leads to more ¹O₂ produced by BiOBr_0.5I_0.5/BiOI under visible light via the enhanced exciton photocatalysis. The ROS quantitative experiments further supported that the synergistic effect of heterojunction and solid solution via exciton and carrier photocatalysis leads to an enhancement of the molecular oxygen activation of BiOBr_0.5I_0.5/BiOI. Therefore, the photocatalytic NO removal efficiency of BiOBr_0.5I_0.5/BiOI was enhanced by the improved molecular oxygen activation capacity due to the synergistic effect of solid solution and heterojunction.

Trapping experiments with different trapping agents during the photocatalytic NO removal process was were conducted to explore the photocatalytic NO removal process of BiOBr_0.5I_0.5/BiOI further. As is shown in Fig. 6(a) and (b), compared to the NO removal efficiency (36.2%) without trapping agents, adding triethanolamine (13.2%), 2'2'6'6'-tetramethylpiperidine nitrogen oxide (TEMPO, 10.2%) and benzoquinone (BQ, 7.1%) reduced the efficiency [53]. It indicated that the holes (h⁺), ¹O₂ and ·O₂^- generated by BiOBr_0.5I_0.5/BiOI were all efficient ROS generated by the photocatalytic NO removal process. Based on the molecular orbital of NO (Fig. 6(c)), there was an unpaired electron present in the anti-bonding molecular orbital, which could be easily removed due to its instability [54]. We determined that the heterojunction structure of BiOBr_0.5I_0.5/BiOI resulted in more h⁺ to capture unpaired NO electrons (Fig. 6(d)), which transformed NO into NO⁺. Then ROS (¹O₂ and ·O₂^-) oxidized NO⁺ to NO₂.

In order to elucidate the process of photocatalytic NO removal further, in situ FT-IR spectra were recorded, which are shown in Fig. 7, demonstrating the photocatalytic NO oxidation on the BiOBr_0.5I_0.5/BiOI surface with irradiation of visible light. After 30 min of adsorption, it was observed that NO (1245 cm^-1), NO + O₂ (1533 and 1614 cm^-1), and NO⁺ (1373 cm^-1) peaks appeared [55]. When the light was switched on, the curves changed obviously. There was a trend of increasing intensity of the characteristic NO peaks. At longer irradiation, the NO⁺ peaks increased while the NO₂ peaks increased initially and then decreased. The NO₂ peaks increased at first because NO₂ was produced through the photocatalytic oxidation of NO by BiOBr_0.5I_0.5/BiOI. With the continuous increase in NO₂, NO₂ was partly desorbed, which agrees with the above photocatalytic results. The in-situ FT-IR results further confirmed the selective NO removal process of BiOBr_0.5I_0.5/BiOI.

Based on the experimental results, we proposed the photocatalytic NO removal process that is shown in Scheme 1. The first step, the transfer of photogenerated electrons and holes via carrier photocatalysis is promoted by the electric field caused by the heterojunction. Then, NO⁺ was generated as an intermediate product by the reaction between h⁺ and NO. Moreover, O₂ combined with e^- and ·O₂^- was produced. The BiOBr_0.5I_0.5 solid solution had a unique layered structure, exhibiting an intensive electron-hole interaction, which promoted ¹O₂ generation via an energy transfer process. Furthermore, ·O₂^- and ¹O₂ supported the superior photocatalytic oxidation activity of BiOBr_0.5I_0.5/BiOI to oxidize NO to NO₂. Finally, NO₂ reacted with water to produce nitric and nitrous acid. According to our findings, the synergistic effect of the solid solution and heterojunction induced the enhanced photocatalytic NO removal process of BiOBr_0.5I_0.5/BiOI and is presented as follows.

(1) BiOBr_0.5I_0.5/BiOI + visible light → h⁺ + e^-

(2) NO + h⁺ → NO⁺

(3) e^- + O₂ → ·O₂^-

(4) O₂ + visible light → ¹O₂

(5) NO⁺ + ·O₂^- → NO₂; NO⁺ + ¹O₂ → NO₂

(6) 2NO₂ + H₂O → HNO₃ + HNO₂

In this study, the solid solution and heterojunction containing photocatalyst BiOBr_0.5I_0.5/BiOI was prepared and characterized. The molecular oxygen activation capacity was enhanced by the heterojunction and solid solution structures and induced the exciton and carrier photocatalysis, which was confirmed by ·O₂^- and ¹O₂ generation. The superior photocatalytic oxidation activity of BiOBr_0.5I_0.5/BiOI promoted the effective NO oxidation to NO₂ under visible light irradiation. The photocatalytic mechanism for NO removal was revealed by trapping experiments and in situ infrared spectroscopy. The intermediate product NO⁺, generated from NO reacting with h⁺, was oxidized to NO₂ by ·O₂^- and ¹O₂. The NO₂ was finally converted into NO₃^- and NO₂^- by the reaction with water.

Bandgap structure, TEM images, PL, TMB and NBT adsorption spectra were described in the supplementary information.