Environmental pollution and the energy shortage have become two of the biggest challenges currently facing humans. Semiconductor photocatalysis with a primary focus on TiO₂ is one of the most promising technologies for environmental remediation and energy conversion ^{[1, 2, 3, 4]}. However, owing to its large band gap of 3.2 eV, TiO₂ can drive a catalytic reaction only under ultraviolet or near ultraviolet irradiation, which occupies less than 4% of the solar light spectrum on the earth. To make full use of solar light, therefore, the development of narrow band gap semiconductors as efficient visible-light photocatalytic materials has been of particular interest in recent years ^{[5, 6, 7, 8, 9, 10]}. Bismuth trioxide (Bi₂O₃), which exhibits a narrow band gap that varies between 2.1-2.8 eV, is a promising candidate as a visible-light photocatalyst owing to its unique characteristics, such as strong absorption in visible region and non-toxic property similar to TiO₂. It is well-known that Bi₂O₃ has six polymorphic forms, including the a (monoclinic), b (tetragonal), g (body-centered cubic), d (face-centered cubic), e (tetragonal), and w (triclinic) phases, each with unique physical properties ^[11]. Among them, a-Bi₂O₃, b-Bi₂O₃ and d-Bi₂O₃, with photocatalytic activities towards the degradation of organic pollutants, have been reported. b-Bi₂O₃, with a relatively narrow band gap (less than ~2.50 eV), has been shown to be the most active form for photocatalytic application ^{[12, 13, 14]}. Nonetheless, their efficiencies are still insufficient for practical application.

The construction of a semiconductor heterojunction by coupling with a secondary substance (noble metal, other semiconductors, and so on) has been proven to be an effective and simple strategy to improve the photocatalytic activity of photocatalysts ^{[15, 16]}. In the semiconductor heterojunction, several advantages can be achieved: (1) an improvement of the charge separation; (2) a long lifetime of the charge carriers; and (3) an enhancement of the charge transfer to the catalyst surface. All these features endow the semiconductor heterojunction with an enhanced photocatalytic performance. To date, many Bi₂O₃-based heterojunction photocatalysts have been successfully developed for the efficient degradation of organic pollutants ^{[17, 18, 19]}. In particular, the heterojunctions between a-Bi₂O₃ and bismuth oxyhalides (BiOX, X = Cl, Br, or I), formed through an in-situ treatment of a-Bi₂O₃ with HX, such as a-Bi₂O₃/BiOI and a-Bi₂O₃/BiOCl, exhibit unique advantages not only in photocatalytic performance but also in the fabrication method ^{[20, 21, 22]}. However, to the best of our knowledge, the heterojunctions between b-Bi₂O₃ and BiOX (X = Cl, Br, or I) are not yet to be reported. Among the BiOX family, bismuth oxyiodide (BiOI) possesses the smallest band gap (less than 1.8 eV) and strong absorption in the visible region [23]. With the aims of the full use of solar light, improvement in b-Bi₂O₃ photocatalysis for a practical application, and convenient fabrication features, in this work, we coupled b-Bi₂O₃ with BiOI to form b-Bi₂O₃/BiOI heterojunctions through an in-situ treatment with HI. The prepared b-Bi₂O₃-based heterojunctions have been shown to exhibit much higher photocatalytic activities than pure b-Bi₂O₃ and BiOI for the degradation of methyl orange (MO) upon visible-light irradiation. Furthermore, the mechanism of the enhanced photocatalytic performance over the heterojunctions was also investigated and discussed according to various characterization results.

b-Bi₂O₃ was synthesized by using Bi₂O₂CO₃ as a precursor, as reported in previous work ^[24]. Initially, 11.64 g of Bi(NO₃)₃×5H₂O was dissolved in 60 ml aqueous solution of HNO₃ (1 mol/L). Under constant stirring, 240 ml aqueous solution of Na₂CO₃ (0.6 mol/L) was slowly added into the solution to produce a large amount of white precipitate. The suspension was further stirred for 15 min and aged at 60 °C for 12 h. Finally, the precipitate was collected and washed several times with deionized water before being dried at 60 °C for 6 h to form the Bi₂O₂CO₃ precursor. b-Bi₂O₃ was obtained by annealing the prepared Bi₂O₂CO₃ at 350 °C for 30 min. The b-Bi₂O₃/BiOI heterojunctions were prepared by an in-situ treatment of b-Bi₂O₃ with HI. In detail, 0.7 g of the freshly prepared b-Bi₂O₃ was dispersed into 20 ml of an ethanol and water mixture with the desired amount of HI, which was prepared by changing the amount of the aqueous solution of HI (45%). To obtain the b-Bi₂O₃ based heterojunctions with different amount of BiOI, the molar ratios of HI to b-Bi₂O₃ (R_HI) in the dispersions were controlled at 0.2, 0.4, and 0.8, respectively. The dispersion was stirred and sonicated for 30 min at room temperature. After the treatment with HI, the precipitate was recovered by centrifugation, washed several times with ethanol, and finally dried at 80 °C in air for 10 h. By following this procedure, the pure BiOI was synthesized in the presence of HI in excess. For comparison, the pure b-Bi₂O₃ was also treated in the same manner in the absence of HI.

The polycrystalline X-ray diffraction (XRD) patterns of the prepared samples were recorded with an X-ray diffractometer (Shimadzu, XRD-7000) using Cu K_a (l = 1.5406 Å) as X-ray radiation with an applied potential of 40 kV and electron beam current of 30 mA. Ultra violet-visible diffuse reflectance spectroscopy measurements were carried out on a UV-Vis spectrometer (Hitachi U-4100) equipped with a diffuse reflectance accessory. Field-emission scanning electron microscopy (FESEM) observations were performed with a JEOL 6701F microscope. High-resolution transmission images (HRTEM) were taken using a JEOL JEM-2010 electron microscope. X-ray photoelectron spectroscopy (XPS) measurements were performed on a XPS system (ESCALAB 250Xi) with a 300 W Al K_asource. All binding energies were referenced to the C 1s peak (284.6 eV) of surface adventitious carbon. Photoluminescence (PL) spectra were obtained on a Hitachi F-4600 spectrophotometer with an excitation wavelength at 380 nm.

MO was chosen as a model pollutant to evaluate the photocatalytic activities of the prepared pure Bi₂O₃, BiOI, and b-Bi₂O₃/BiOI heterojunctions under visible-light irradiation (l > 420 nm). In a typical photocatalytic experiment, 0.1 g of the catalyst was dispersed in 100 ml aqueous solution of MO (4×10^-5 mol/L). The light source was a 300 W Xe-arc lamp (CHF-XM150, Beijing Trusttech. Co. Ltd.) equipped with a glass filter (l > 420 nm) to remove the UV wavelength light, and positioned about 8 cm above the aqueous suspension. Prior to irradiation, the suspension was magnetically stirred in the dark for 1 h to reach an adsorption-desorption equilibrium between the catalyst surface and MO molecules. At the given irradiation time intervals, 3 ml of the suspension was sampled and separated by centrifugation before analysis using a Hitachi U-3310 spectrophotometer. The intensity change in the characteristic absorption peak of MO at 461 nm was measured to evaluate the extent of MO degradation.

To investigate the photocatalytic degradation route of MO over these catalysts, triethanolamine (TEOA, 10 mmol/L), as an effective hole scavenger, and tert-butyl alcohol (TBA, 10 mmol/L), as •OH radical scavenger, were chosen to participate in the photocatalytic degradation of MO over these catalysts. Determination of MO concentration during the photocatalytic reaction in the presence of TEOA or TBA was also conducted by measuring the absorption of MO solution at 461 nm.

Polycrystalline XRD was used to identify the phase structures and constitutions of the prepared samples. The XRD patterns of the prepared samples are illustrated in Fig. 1. It can be clearly observed that the prepared b-Bi₂O₃ exhibits a single tetragonal phase of well-crystalline b-Bi₂O₃ (JCPDS 27-0050). After the treatment with excess HI, the crystalline b-Bi₂O₃ was entirely transformed to the pure tetragonal BiOI phase (JCPDS 73-2062). In addition, with the increase in R_HI values in the b-Bi₂O₃/BiOI heterojunctions, the intensity of the characteristic peaks for tetragonal BiOI are gradually increased, while those for tetragonal b-Bi₂O₃ are decreased simultaneously. These results reveal that the prepared b-Bi₂O₃/BiOI heterojunctions are composed of b-Bi₂O₃ and BiOI phases, and that b-Bi₂O₃ is the major phase.

Fig. 1. XRD patterns of b-Bi2O3, BiOI, and b-Bi2O3/BiOI heterojunctions with different RHI values.

The detailed morphology and microstructure of the prepared samples were inspected by FESEM and HRTEM. Fig. 2 presents the FESEM images of the b-Bi₂O₃, BiOI, and b-Bi₂O₃/BiOI heterojunctions (R_HI = 0.4). It is evident that the prepared pure b-Bi₂O₃ consists of aggregated short nanorods with lengths ranging from 200 to 300 nm and diameters of approximately 100 nm (Fig. 2(a)). The morphology of pure BOI (Fig. 2(b)) appears to be irregular sheets with a thickness of less than 10 nm, which is similar to that reported in previous reports ^{[25, 26]}. As for the b-Bi₂O₃/BiOI heterojunction (R_HI = 0.4) sample (Fig. 2(c)), it can be found that the short nanorods became small, accompanying the formation of nanosheets after the treatment of b-Bi₂O₃ with HI. Fig. 3 displays the typical HRTEM images of pure b-Bi₂O₃, BiOI, and b-Bi₂O₃/BiOI heterojunction (R_HI = 0.4). All prepared samples were highly crystallized. The clear and uniform fringes with the measured lattice spacing of 0.34 nm (Fig. 3(a)) and 0.29 nm (Fig. 3(b)) matched well with the crystallographic planes of tetragonal b-Bi₂O₃ phase (210) and tetragonal BiOI phase (122), respectively ^{[27, 28]}. Both of the fringes with the lattice spacing of 0.34 and 0.29 nm can also be clearly observed in Fig. 3(c), confirming the co-existence of the b-Bi₂O₃ and BiOI phases in the b-Bi₂O₃/BiOI heterojunctions. The HRTEM observations are in good agreement with the XRD results. Moreover, it can also be seen in Fig. 3(c) that there is a well-contacted interface between the b-Bi₂O₃ and BiOI phases. The well-contacted interface is expected to be favorable for charge transfer between the two phases ^[15].

Fig. 2. FESEM images of b-Bi2O3 (a), BiOI (b), and b-Bi2O3/BiOI heterojunctions (RHI = 0.4) (c).

Fig. 3. HRTEM images of b-Bi2O3 (a), BiOI (b), and b-Bi2O3/BiOI heterojunctions (RHI = 0.4) (c).

The optical properties of the prepared samples were studied by UV-Vis diffuse reflectance spectroscopy, as illustrated in Fig. 4. All samples exhibited strong absorption in the visible light region. The pure b-Bi₂O₃ has an intense absorption starting at the wavelength near 535 nm, whereas the BiOI presents an absorption extending up to approximately 675 nm. Additionally, the absorption band edges of the b-Bi₂O₃/BiOI heterojunctions are between those of b-Bi₂O₃ and BiOI, and gradually shift to that of pure BiOI with the increase in the amount of BiOI (R_HI).

Fig. 4. UV-diffuse reflectance spectra of b-Bi2O3, BiOI, and b-Bi2O3/BiOI heterojunctions with different RHI values.

According to the absorption spectra, the band gap energy (E_g) of a semiconductor can be evaluated from the following formula ^[29]:

αhv = A(hv - E_g)^n/2(1)

where a, hv, E_g, and A are the optical absorption coefficient, photo energy, band gap energy, and a constant, respectively. The value of n depends on the characteristic of the transition in a semiconductor (i.e., n = 1 for direct transition and n = 4 for indirect transition). Here, n is 1 for b-Bi₂O₃ and is 4 for BiOI. In the plots of (ahv)² versus hv for b-Bi₂O₃ and (ahv)^1/2 versus hv for BiOI, as shown in Fig. 5, the band gap energies of b-Bi₂O₃ and BiOI are estimated from the tangent lines to be 2.28 and 1.77 eV, respectively. The band gap values for b-Bi₂O₃ and BiOI determined here are very similar to the values reported previously in the literature ^{[25, 30, 31]}.

Fig. 5. (a) Plots of (ahv)2 versus hv for b-Bi2O3. (b) Plots of (ahv)1/2 versus hv for BiOI.

XPS was used to examine the surface chemical states and compositions of the elements in the samples. The XPS survey spectra, shown in Fig. 6(a), reveal no peaks of elements other than Bi, O, and C in the b-Bi₂O₃, and Bi, O, I, and C in both BiOI and b-Bi₂O₃/BiOI heterojunction (R_HI = 0.4) samples, evidencing the purity of our prepared samples. Fig. 6(b) displays the high resolution XPS spectra of Bi 4f regions of the pure b-Bi₂O₃ and BiOI samples. As observed in Fig. 6(b), the Bi 4f spectra of both samples consist of two individual symmetric peaks. In the Bi 4f core-level spectrum of the b-Bi₂O₃, two peaks appearing at the binding energies of 158.53 and 163.83 eV, with a separation of 5.3 eV, are assigned to Bi 4f_7/2 and Bi 4f_5/2, respectively, which is a characteristic of Bi³⁺ in b-Bi₂O₃ according to a previous report ^[11]. The positions and separation of two peaks in the Bi 4f core-level spectrum of the pure BiOI also support the presence of Bi³⁺ in the tetragonal BiOI phase ^[32].

Fig. 6. (a) XPS survey spectra of b-Bi2O3, BiOI, and b-Bi2O3/BiOI heterojunctions (RHI = 0.4). (b) High-resolution Bi 4f XPS spectra of b-Bi2O3 and BiOI. (c) High-resolution I 5d XPS spectra of BiOI and b-Bi2O3/BiOI heterojunctions (RHI = 0.4).

Figure 6(c) presents the high resolution XPS spectra of I 5d regions of the pure BiOI and b-Bi₂O₃/BiOI heterojunctions (R_HI = 0.4). There are two individual symmetric peaks in the I 5d region. The two symmetric peaks of I 5d located at the binding energies of 619.08 and 630.58 eV in the I 5d core-level spectrum of the pure BiOI with a peak difference of 11.5 eV are ascribed to I 3d_5/2 and I 3d_3/2 of the I^- species in BiOI, respectively ^{[25, 33]}. Notably, the binding energies of I 3d_5/2 and I 3d_3/2 obviously shift to lower values after the formation of b-Bi₂O₃/BiOI heterojunctions, whereas the peak separation value is still the same as that of the corresponding pure BiOI. This result suggests an interaction between b-Bi₂O₃ and BiOI in the heterojunctions. Earlier investigations indicated that when two semiconductors were in contact, electron migration between them occurs until a Fermi level equilibration of the composite is established ^[34]. In the present case, the observed shift of the binding energies of I 5d in BiOI to lower values suggests that the Fermi level of b-Bi₂O₃ is more negative than that of BiOI, so electron transfer from b-Bi₂O₃ to BiOI occurs when b-Bi₂O₃ and BiOI are in contact to form heterojunctions. The observed shift in binding energy also reflects the presence of a well-contacted interface between BiOI and b-Bi₂O₃ in the b-Bi₂O₃/BiOI heterojunctions, further confirming the above HRTEM observations.

To investigate the photocatalytic efficiencies of the prepared b-Bi₂O₃/BiOI heterojunctions, the photocatalytic activities for the degradation of MO over pure b-Bi₂O₃, BiOI, and b-Bi₂O₃/BiOI heterojunctions with different R_HI values were evaluated under visible-light irradiation (l > 420 nm), as shown in Fig. 7. The control experiment shows that the self-degradation of MO is negligible under visible-light irradiation in the presence of no catalysts. It is clear that all b-Bi₂O₃/BiOI heterojunctions with different R_HI values exhibit much higher photocatalytic activities than b-Bi₂O₃ or BiOI, demonstrating that the construction of the heterojunctions between b-Bi₂O₃ and BiOI is an efficient approach for enhancing the photocatalysis of b-Bi₂O₃ and BiOI. Moreover, the photocatalytic activity also depends on the amount of BiOI (R_HI) in the b-Bi₂O₃-based heterojunctions. The highest activity is reached over the b-Bi₂O₃/BiOI heterojunctions (R_HI = 0.4). Approximately 95% of MO was degraded over the heterojunctions under visible-light irradiation for 2 h. The presence of excess BiOI (R_HI ≥ 0.8) appears to be detrimental to the photocatalytic performance of the b-Bi₂O₃/BiOI heterojunctions. Clearly, changing the content of BiOI in the heterojunction has a significant influence on the heterojunction interface, b-Bi₂O₃ surface coverage, and surface morphologies of both components. Based on the results shown in Fig. 7, it could be concluded that an optimum combination of these factors has been achieved over the b-Bi₂O₃/BiOI heterojunctions (R_HI = 0.4), exhibiting the highest photocatalytic activity.

Fig. 7. Photocatalytic degradation of MO over b-Bi2O3, BiOI, and b-Bi2O3/BiOI heterojunctions with different RHI values under visible-light irradiation for 2 h.

Figure 8 presents the degradation of MO over pure b-Bi₂O₃ and b-Bi₂O₃/BiOI heterojunctions (R_HI = 0.4) in the presence of different scavengers under visible-light irradiation. Almost no degradation of MO in two samples occurs when TEOA, as an effective hole scavenger, is added into the reaction solution, while the addition of TBA as •OH scavenger slightly influences the degradation of MO. These results reveal that the photogenerated hole is almost the sole active species involved in the degradation of MO over both samples. The hydroxyl radical •OH is regarded to be a very powerful oxidizing agent which can degrade most organic compounds. Almost no change in the degradation of MO in the presence of TBA infers the production of •OH is negligible in both systems. The effective production yield of the photogenerated holes in a semiconductor photocatalyst strongly depends on the separation degree of the photogenerated electron-hole pairs. According to the results shown in Fig. 7, therefore, an efficient separation of the photogenerated electron-hole pairs is achieved after b-Bi₂O₃ combines with BiOI to form the heterojunction, which is also supported by the PL spectra of pure b-Bi₂O₃ and b-Bi₂O₃/BiOI heterojunctions (R_HI = 0.4), as shown in Fig. 9. It can clearly be seen that a strong emission peak around 530 nm appears for the pure b-Bi₂O₃, which originates from the direct electron-hole recombination. The intensity of the PL emission peak of b-Bi₂O₃/BiOI heterojunctions (R_HI = 0.4) is dramatically decreased compared with that of pure b-Bi₂O₃, indicating an increase in the separation of the photogenerated electron-hole pairs after the formation of b-Bi₂O₃/BiOI heterojunctions. Moreover, the observed broadening and red-shift of the PL emission peak of b-Bi₂O₃/BiOI heterojunctions (R_HI = 0.4) are probably caused by the direct electron-hole recombination of both b-Bi₂O₃ and BiOI band transitions.

Fig. 8. Photocatalytic degradation of MO over b-Bi2O3 (a) and b-Bi2O3/BiOI heterojunctions (RHI = 0.4) (b) in the presence of different scavengers under visible-light irradiation for 2 h.

Fig. 9. PL spectra of pure b-Bi2O3 and b-Bi2O3/BiOI heterojunctions (RHI = 0.4).

To understand the effects of the b-Bi₂O₃/BiOI heterojunctions on the separation of photo-generated electron-hole pairs, the band structures of b-Bi₂O₃ and BiOI before and after the formation of their heterojunctions were studied. The band positions of a semiconductor at the point of zero charge can be calculated through the following equations:

E_VB = χ-E^e + 0.5E_g (2)

E_CB = E_VB E_g (3)

where E_VB and E_CB are the valence band (VB) and conduction band (CB) potentials, respectively, c is the absolute electronegativity of the semiconductor, and E^e is the energy of the free electron on the hydrogen scale (about 4.5 eV), and E_g is the band gap energy. The c values for b-Bi₂O₃ and BiOI are approximately 5.95 and 5.94 eV, respectively. Given the band gap values from Fig. 5 and the equations above, the E_VB and E_CB of b-Bi₂O₃ can be calculated to be 2.59 and 0.31 V, respectively. Accordingly, the E_VB and E_CB of BiOI are 2.33 and 0.56 V, respectively.

According to the obtained band position data above, the b-Bi₂O₃ and BiOI appear to be in the nested architecture in their band structure before their contact to form the heterojunction, which seems to not be favorable for an efficient separation of photogenerated electron-hole pairs. Nevertheless, such nested architecture is completely changed after their contact. As has already been reported ^{[35, 36, 37]}, b-Bi₂O₃ is a n-type semiconductor with the Fermi level close to its conduction band, whereas BiOI is a p-type semiconductor with the Fermi level close to its valence band. When b-Bi₂O₃ and BiOI are in contact to form a heterojunction, the electron migration from b-Bi₂O₃ to BiOI occurs, as evidenced by the XPS results in Fig. 6(c), until the Fermi levels of the two semiconductors reach equilibrium. As a result, the energy bands of b-Bi₂O₃ and BiOI shift downward and upward, respectively, and an electronic field is built at the interface of the heterojunction. Under visible-light irradiation, both b-Bi₂O₃ and BiOI are excited to generate electron-hole pairs. Driven by the electronic field at the interface, the electrons are transferred to the reformed CB of b-Bi₂O₃, probably to form H₂O₂ through two-electron reduction process according to the potential of the reformed CB of b-Bi₂O₃ ^[19], and the holes are transferred to the reformed VB of BiOI to degrade MO in the present case. Hence, the efficient separation of the photogenerated electron-hole pairs is successfully achieved across the interface of the heterojunction between b-Bi₂O₃ and BiOI, resulting in enhanced photocatalysis over the heterojunction. The above understanding of the enhanced photocatalysis over the b-Bi₂O₃/BiOI heterojunctions is schematically illustrated in Scheme 1.

Scheme 1. Schematic diagram of the photogenerated electron-hole pair separation and transfer over b-Bi2O3/BiOI heterojunctions under visible-light irradiation.

We have successfully synthesized b-Bi₂O₃/BiOI heterojunctions through a precipitation method, followed by an in-situ treatment with HI. Compared with pure b-Bi₂O₃ and BiOI, the prepared b-Bi₂O₃/BiOI heterojunctions, especially with the molar ratio of HI to b-Bi₂O₃ at 0.4, exhibit much higher photocatalytic activities for the degradation of MO under visible-light irradiation. The enhanced photocatalysis could be attributed to the efficient separation of the photogenerated electron-hole pairs achieved across the interface of the heterojunction between b-Bi₂O₃ and BiOI.