Nitric oxide (NO) has caused extensive concerns as a major contributor to photochemical smog, acid rain, and ozone depletion [1-3]. NO released in high concentrations in the form of industrial emissions can be treated by conventional techniques including physical adsorption, biofiltration, and thermal catalysis [4-6]. However, the traditional methods are not economically feasible for NO removal at parts per billion level [7-9]. Semiconductor photocatalysis as a green technology offers the advantage of direct utilization of the natural sunlight to purify air pollutants at low concentrations under ambient conditions, thus providing a promising alternative approach to NO purification [10-14].

Visible light accounts for nearly 50% of the solar spectrum and far exceeds the proportion of ultraviolet (UV) light (4%). However, wide-band-gap semiconductors primarily respond only to UV light [15]. Although narrow-band-gap semiconductor photocatalysts exhibit excellent visible light absorbing properties, they suffer from fast recombination of electron-hole pairs and low redox ability [16-18]. Therefore, narrow- and wide-band-gap photocatalysts can display high visible light adsorptions and strong redox abilities, respectively [19-21]. Consequently, various wide- and narrow-band-gap photocatalysts have been combined to enhance the photocatalytic performance, such as TiO₂/g-C₃N₄ [22], SrTiO₃/Cu₂O [23], and BaCO₃/BiOI [24]. SrTiO₃ is a typical n-type semiconductor with a wide band gap of approximately 3.2 eV that has been considered as a promising photocatalyst because of its physicochemical stability and non-toxicity [25-27]. The p-type narrow-band-gap semiconductor BiOI has been extensively studied because of its excellent visible light absorbing properties [28]. However, the photocatalytic activity of BiOI is normally low owing to fast recombination of its photogenerated carriers [29]. Considering the properties of SrTiO₃ and BiOI, fabrication of a p-n heterostructure by combining BiOI with SrTiO₃ can potentially promote charge separation and thus enhance the photocatalytic performance under visible light irradiation, in comparison with those of the pristine catalysts [30-33].

Most importantly, most of the reported literature solely focus on the efficiency of degradation of target pollutants (such as NO in air), while attention has rarely been paid to the toxic byproducts obtained during the transformation process of the target pollutant [34, 35]. Therefore, it is desirable to explore novel photocatalysts that would simultaneously enhance the efficiency of NO purification under visible light irradiation and inhibit the generation of toxic intermediates [36-39]. Besides, the charge separation and transfer mechanism at the contacting interface of the two counterparts has not been fully revealed [40]. Exploration of the interfacial charge separation and electron transfer mechanism remains a great challenge. Therefore, the interfacial interaction of the heterojunction with the reactants requires investigation at the atomic level.

In this work, SrTiO₃/BiOI (STB) composites were successfully prepared by a facile method, through which SrTiO₃ was compactly adhered to the surface of BiOI to fabricate a heterostructure. Although SrTiO₃ or BiOI alone shows negligible photocatalytic activity, the heterostructure with matching band gaps can extend the photoabsorption range and promote charge separation, thus achieving a high photocatalytic NO removal performance under visible light irradiation. An attempt was made to combine the theoretical and experimental approaches with the aim of disclosing the formation of electron transfer channels at the interface between SrTiO₃ and BiOI. Moreover, in situ diffused reflectance infrared Fourier transform spectroscopy (DRIFTS) technology was employed to dynamically monitor the photocatalytic NO oxidation process and investigate the generation and transformation of toxic intermediates. An effective electronic transmission channel was built to realize an exceptionally high charge-separation rate. Thus, the generation of ROS was enhanced for efficient photocatalytic oxidation of NO into the end products. This work presents a simple strategy that promotes efficient and safe air purification through photocatalytic technology.

All the chemicals used in this study were of analytical grade and employed without further purification. SrTiO₃ was synthesized by hydrothermal method according to the following procedure. In a typical process, 5.0 mmol of tetrabutyl titanate was dissolved in 25 mL ethylene glycol. Then, 15.0 mL of 0.50 M Sr(NO₃)₂ solution was added to the above solution. After the two solutions were mixed, 20.0 mmol NaOH was added to the mixture solution and the resulting mixture was sealed in a 50 mL Teflon-lined stainless-steel autoclave and maintained at 180 ℃ for 12 h. After cooling to room temperature, the obtained precipitates were washed several times with ultrapure water and alcohol and dried at 70 ℃ for 12 h to realize the final samples.

The SrTiO₃/BiOI heterojunction was constructed by the facile chemical bath method. 0.10 g SrTiO₃ and 1.0 mmol KI were ultrasonically dispersed in 40.0 mL distilled water to form a uniform solution A. Then, 1 mmol Bi(NO₃)₃∙5H₂O was dissolved in 20 mL ethylene glycol to form solution B. Solution B was added dropwise to solution A under vigorous stirring. Afterwards, the mixture was transferred into a water bath at 80 ℃ and held for 2 h. After cooling to room temperature, the obtained precipitates were washed several times with ultrapure water and alcohol and dried at 70 ℃ for 12 h. By changing the weight of SrTiO₃ added in the synthesis process, 22.12, 36.23, and 46.01 wt% STB composites were obtained. As will be discussed later, all the three STB samples revealed similar patterns in their photocatalytic activities and characterizations. Thus, STB-36.23 (labeled as STB) was presented as a prototype, and the corresponding results of STB-22.12 and STB-46.01 were listed in Supplementary Materials. Pure BiOI was prepared according to the same procedure as detailed above, but by not adding SrTiO₃.

The crystal structures of the samples were obtained using powder X-ray diffraction (XRD). X-ray photoelectron spectroscopy (XPS) was used to analyze the chemical compositions of the samples. A UV-vis diffuse spectrophotometer was employed to analyze the optical properties of the samples with 100% BaSO₄. The morphological structures were revealed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The electron spin resonance (ESR) spectra of the radicals that were spin-trapped using 5, 5-dimethyl-1-pyrroline N-oxide (DMPO) were also obtained. More details of the characterization methods are provided in text S1.

The photocatalytic performance was evaluated by removing NO (present at ppb levels) from a continuous stream reactor (Fig. S1). Details of the photocatalytic activity tests and analyses are provided in text S2 of Supporting Information.

In situ DRIFTS measurements (Fig. S2) were performed with the help of a TENSOR Ⅱ FT-IR spectrometer (Bruker Corp., Billerica, MA, USA), an in situ diffuse-reflectance cell (Harrick Scientific Products Inc., Pleasantville, NY, USA), a gas system, a light source (MVL-210; Mejiro Genossen Inc., Tokyo, JPN), and pretreatment equipment. The reaction chamber was equipped with three windows, two for the entry and detection of IR light and the third for illumination of the photocatalyst. A Xe lamp was used as the irradiation light source. The total gas flow rate was 100 mL/min (50 mL/min NO and 50 mL/min O₂). The prepared samples were pretreated at 80 ℃ for 20 min under high-purity He gas before the measurements.

DFT calculations were performed using the "Vienna ab initio simulation package" (VASP5.4) and a generalized gradient correlation function [41-43]. A plane-wave basis set with a cut-off energy of 400 eV and the projector-augmented wave method framework were adopted. The Gaussian smearing width was set to 0.2 eV [44, 45]. The Brillouin zone was sampled with a 3 × 3 × 2 Monkhorst Pack grid. All the atoms were allowed to converge to 0.01 eV/Å. The 15 × 18 × 20 supercell includes BiOI with 16 O atoms, 12 Bi atoms, and 8 I atoms, and SrTiO₃ with 96 O atoms, 24 Sr atoms, and 24 Ti atoms. The supercells of BiOI and SrTiO₃ alone were first relaxed, before the BiOI supercell was placed on the plane of SrTiO₃ [46].

The as-prepared SrTiO₃, BiOI, and STB composites were characterized by powder XRD (Fig. 1(a)). The diffraction peaks of the as-prepared pure BiOI and SrTiO₃ can be indexed to tetragonal BiOI (JCPDS 73-2062) and cubic SrTiO₃ (JCPDS 35-0734), respectively. Some characteristic peaks of SrTiO₃ and BiOI can be observed in the patterns of STB composites that confirm successful fabrication of the heterojunction. Notably, the (110) peak of BiOI in STB (at 31.7°) is right-shifted in comparison with that of pure BiOI (30.6°), which can be attributed to the interaction of SrTiO₃ with BiOI. Besides, the peaks of SrTiO₃ in the composites are weakened, which is a result of the high diffraction intensity of BiOI [47, 48]. The other samples with different mass percentages of SrTiO₃ present similar phenomena (Fig. S3). The XPS survey spectra (Fig. 1(b)) show that all these samples contain Bi, O, I, Sr, and Ti elements, which further indicate that the STB composites have been successfully prepared [49].

The SEM and TEM images show the morphologies of the SrTiO_3, BiOI, and STB samples (Figs. 2 and S4). In Figs. 2(a) and 2(b), it is observed that the SrTiO₃ is composed of nanoparticles and the BiOI is constructed of nanosheets. The SEM images of STB (Figs. 2(c) and 2(d)) reveal that SrTiO₃ nanoparticles are dispersed on the surface of BiOI nanoplates. In Figs. 2(e) and 2(f), the lattice fringes observed with interplanar spacings of 0.36 and 0.30 nm can be indexed to the (011) and (012) planes of orthorhombic BiOI phase, respectively. The other lattice fringes with an interval of 0.275 nm are ascribed to the (110) planes of cubic SrTiO₃ perovskite. The EDX elemental maps (Fig. 2(g)) clearly show that SrTiO₃ is distributed on the surface of BiOI nanoplates. The BiOI and SrTiO₃ in the STB sample are tightly contacted with each other, which is consistent with the peak shifts observed in the XRD pattern of STB, which confirm that an interface is formed between BiOI and SrTiO₃ (Fig. 2(f)).

In order to elucidate the interfacial interaction between BiOI and SrTiO₃ in STB heterojunction at the atomic level, the theoretical and experimental approaches were combined. The high-resolution XPS pattern reveals the chemical states of the samples. Fig. 3(a) shows that the Sr 3d and Ti 2p binding energies of the STB shift to higher values, compared to those of pure SrTiO₃. On the other hand, the I 3d and Bi 4f (Fig. 3(b)) binding energies of the STB shift to lower values, in comparison with those of pure BiOI. The characteristic O 1s peaks are also offset to some extent (Fig. S5). These chemical shifts experimentally reveal that electrons are transferred from the Bi and I atoms of BiOI to the Sr and Ti atoms of SrTiO₃ in the STB compound. The shifts in the Sr 3d, Ti 2p, Bi 4f, and I 3d binding energies also confirm the strong interaction between SrTiO₃ and BiOI.

DFT was employed in calculations of the electronic structure to further confirm the nature of the interaction. The ELF indicates that covalent interactions exist at the contacting interface of STB (Fig. 3(c)), where electronic transmission channels are formed to reduce carrier recombination. The as-produced covalent bonds between SrTiO₃ and BiOI lead to the accumulation of localized excess electrons (e_ex⁻), as depicted by the red dotted frame (Fig. 3(d)). These accumulated e_ex⁻ can further transfer through the interface with the covalent interaction channel [50]. It is apparent that the calculation results are consistent with the XPS results. More importantly, since the electron transfer channel is established, intensified electron exchange is expected across the interface under visible light irradiation, which leads to more efficient activation of the reactants and generation of ROS.

The photocatalytic performances of the as-prepared samples in NO removal were investigated under visible light irradiation by using a house-customized continuous flow reaction system (Figs. 4(a) and S6). The NO purification performance on pure BiOI is negligible, although BiOI exhibits large visible light absorption. The STB displays a remarkably enhanced visible light photocatalytic activity, with a NO removal of 59.0%, compared with those of SrTiO₃ (33%) and BiOI. The UV-vis spectra reveal that the photoabsorption of the STB heterojunction is extended to the visible light range (Figs. 4(b) and S7). SrTiO₃ exhibits a decent visible light activity, which is attributed to the existence of O vacancies, as revealed by EPR (Fig. 4(c)) [51]. The DOS was subsequently calculated, and the results show that the O vacancies can form a defect energy level that lowers the photoenergy required for exciting the electrons (Fig. 4(d)) [52]. The limited NO removal activity of BiOI is a result of the rapid recombination of the photogenerated carriers and a feeble redox ability.

Typically, the photocatalytic performance is closely related to the generation of ROS. The ESR profiles were obtained to identify the changes in the ROS during the photocatalytic reaction, including those of DMPO-∙OH, DMPO-∙O₂⁻, TEMP-¹O₂, and e^-, under visible light irradiation (Fig. 5). Compared with those of pure SrTiO₃ and BiOI, the ESR signal intensities of STB are all much stronger for ∙OH, ∙O₂⁻, and ¹O₂ radicals, which indicate the superior oxidation capacity of the STB heterojunction. This fact implies that the photogenerated charge carriers are efficiently separated in STB via the electron transfer channel. The accelerated accumulation of superoxide radical signals in STB suggests that the separated photoelectrons are prone to combine with O₂ molecules to produce more ∙O₂⁻ (O₂ + e⁻ → ∙O₂⁻) under visible light irradiation (Fig. 5(c)). Holes are left behind in the BiOI when the photogenerated electrons transfer from BiOI to SrTiO₃, which can promote the generation of ∙OH radicals via the oxidation of OH⁻ (OH⁻ + h⁺ → ∙OH) [53]. Then, the singlet oxygen (¹O₂) produced by the oxidation of the superoxide radicals (∙O₂⁻ + h⁺ → ¹O₂) is stronger than those of pure BiOI and SrTiO₃ (Fig. 5(c)). The electron sacrificial agent is rapidly consumed in STB under visible light irradiation to generate ROS, but it is still maintained in BiOI, as shown in Fig. 5(d). These ROS radicals can efficiently oxidize NO molecules into the final products. Hence, the outstanding photocatalytic performance of the STB compounds should be attributed to the establishment of electrons transfer channels that accelerate charge separation and ROS generation.

In order to confirm the theoretical predictions and establish a comprehensive understanding of the photocatalytic mechanism, in situ DRIFTS was employed to dynamically track the primary reaction intermediates and products during the NO purification reaction. BiOI had been confirmed to exhibit negligible NO removal performance under visible light irradiation (Fig. 4(a)). Therefore, this section only compares SrTiO₃ with STB (Fig. S8). After the adsorption equilibrium is reached (Ads. Equ.), the DRIFTS pattern was recorded with light irradiation to explore the photocatalytic NO purification mechanism, especially for the generation and transformation of toxic intermediates. As shown in Fig. 6(a), the peak intensities of NO₂ (1616 cm⁻¹) and NO₂⁻ (1353 and 1303 cm⁻¹) on SrTiO₃ gradually increase, especially those of NO₂ [54, 55]. As NO₂ is more toxic than NO, the conversion of NO into NO₂ should be largely inhibited.

Notably, the peak of NO₂ on STB was barely detected (Fig. 6(b)). The normalized evolution of species (Fig. 6(c)) implies that the amount of NO₂ on STB is kept at a low level. In addition, the peak at 1565 cm⁻¹ was evidently detected for STB under visible light irradiation, which can be assigned to NO₃⁻ [56]. Similarly, the final product NO₂⁻ (1366 cm⁻¹) was monitored on the STB. However, the peaks of the final products (i.e., NO₂^- and NO₃⁻) cannot be obviously detected on SrTiO₃. Based on these results, it is obvious that NO was smoothly converted into the target products (NO₂⁻ and NO₃⁻) and that NO₂ did not accumulate on STB. On SrTiO₃, NO dominantly transformed to NO₂. Acceleration of the charge separation and transfer process on STB promotes the generation of ROS. Thus, the toxic intermediate NO₂ cannot be further effectively converted into the target products. After a comprehensive comparison of the DRIFTS results of STB and SrTiO₃, it can be concluded that STB can inhibit the generation of toxic intermediates and enhance the photocatalytic oxidation of NO. Based on these results, it can be summarized that the photocatalytic activity and the inhibition of toxic intermediates are simultaneously reinforced by the covalent interactions occurring at the interface of SrTiO₃ and BiOI, which can be attributed to the enhanced charge separation that increases the generation of ROS. The mechanism proposed in this work is illustrated in the form of a schematic in Fig. 7. The photoexcited electrons of BiOI can migrate to the CB of SrTiO₃ under the driving force of the energy difference between the CBs of BiOI and SrTiO₃. The electrons in the CB of SrTiO₃ can react with the O₂ to produce ∙O₂⁻ radicals. The residual holes in the VB of BiOI in the heterojunction react with the OH⁻ to produce ∙OH radicals. The SrTiO₃/BiOI heterostructure with matching band gaps can promote charge separation and enhance ROS generation, thus achieving a high photocatalytic activity under visible light irradiation.

SrTiO₃/BiOI heterojunction photocatalysts were designed and fabricated to simultaneously enhance the efficiency of NO purification in air under visible light irradiation and inhibit the generation of toxic intermediates. By combining experimental and theoretical approaches, we revealed the formation of electron delivery channels at the contacting interface of the two counterparts. The electrons photogenerated in the BiOI semiconductor can directly transfer to the SrTiO₃ surface through the preformed electron delivery channel. Therefore, the electron channels can facilitate the generation of ROS, which enhances the overall NO purification efficiency and inhibits the formation of toxic intermediates. This work presents a simple and novel strategy that promotes efficient and safe air purification through photocatalytic technology.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (21822601, 21501016, 21777011), the National R & D Program of China (2016YFC02047), the Innovative Research Team of Chongqing (CXTDG201602014), the Natural Science Foundation of Chongqing (cstc2017jcyjBX0052), and the Plan for "National Youth Talents" of the Organization Department of the Central Committee. The authors also acknowledge AM-HPC in Suzhou, China for computational support.