Band Gap Engineering in Quadruple-Layered Sillén–Aurivillius Perovskite Oxychlorides Bi₇Fe₂Ti₂O₁₇X (X = Cl, Br, I) for Enhanced Photocatalytic Performance

Abstract

Developing efficient photocatalyst for the photoreduction of CO₂ and degradation of organic pollutants is an effective alternative to address increasingly serious energy problems and environmental pollution. Herein, the isostructural Sillén–Aurivillius oxyhalides, Bi₇Fe₂Ti₂O₁₇X (X = Cl, Br, and I; BFTOX), are fabricated for CO₂ reduction and degradation of organic pollutants for the first time. Density functional theory (DFT) calculations show that the valence band maximum (VBM) of BFTOC and BFTOB is contributed by the dispersive 2p orbitals of O-atoms, providing the narrow band gap (E_g) and possibly the stability against self-decomposition deactivation. The photocatalytic activities of BFTOX are strongly affected by the halogens (Cl, Br, and I), namely, the BFTOCl sample displays outstanding activity improvement (3.74 μmol·g⁻¹·h⁻¹) for photocatalytic performance. This is mainly attributed to the high separation of charge carriers, small optical band gap, and extended optical absorption. This work focuses on affording a reference to develop efficient and stable photocatalysts from Sillén-Aurivillius layered oxyhalide materials.

1. Introduction

With the development of industrialization, environmental problems have become a global issue of general concern in modern society, particularly for the greenhouse effect caused by carbon dioxide (CO₂) and the discharge of various organic pollutants, threatening the survival and development of mankind [1,2]. Photocatalytic technology is an appealing solution for the reduction of CO₂ and the removal of organic pollutants due to its ability to reduce them into high value-added chemicals or completely harmless molecules under sunlight or UV irradiation [3,4,5,6]. Moreover, semiconductor photocatalysis has attracted wide attention from researchers with the advantages of low cost, highly efficient, recyclable and does not lead to secondary pollution [7,8]. However, most photocatalytic materials have the problems of wide band gap, serious photogenerated carrier recombination and poor stability under light irradiation [9,10,11,12]. Thus, it is necessary to develop new efficient and stable visible light (VL) photocatalysts.

In recent years, bismuth (Bi)-based materials show the potential to be a novel series for VL responsive photocatalysts [13,14,15,16]. Consequently, large amounts of Bi-based photocatalysts have been studied, such as Bismuth trioxide (Bi₂O₃) [17,18], BiOX (X = Cl, Br, and I) [19,20,21], Bi₄Ti₃O₁₂ [22,23], and Bi₄MO₈X (M = Nb, Ta; X = Cl, Br) [24,25], etc. Among them, Sillén–Aurivillius layered oxyhalide materials, such as Bi₄MO₈X (M = Nb, Ta; X = Cl, Br), which are composed of [MO₆] and [(Bi₂O₂)₂X] blocks, have attracted considerable interest for photocatalysis with the virtues of the narrower band gap (E_g), ferroelectric polarization, layered crystal structure and stability against self-decomposition deactivation [26,27,28]. In 2016, Hironori Fujito et al. fabricated Bi₄NbO₈Cl successfully from Bi₂O₃, BiOCl, and Nb₂O₅ powders and the photocatalytic performance for oxygen (O₂) evolution was also evaluated in methanol solution [25]. Simultaneously, their density functional theory (DFT) calculations demonstrated that the valence band maximum (VBM) of Sillén–Aurivillius layered oxyhalide materials is mainly from O-2p orbitals [29,30]. This material can avoid the self-decomposition phenomenon of catalysts upon light irradiation because of the high oxidation of stable by the photo-excited holes near VBM [28,31]. Then, Xiaoping Tao et al. proved that Bi₄TaO₈Cl can be used for hydrogen (H₂) evolution under VL irradiation [32]. These results indicate that Sillén–Aurivillius layered oxyhalide materials Bi₄MO₈X (M = Nb, Ta; X = Cl, Br) should be a potential novel visible light photocatalyst material. In recent years, Sillén–Aurivillius materials were further extensively studied, such as the construction of heterojunctions [33,34], surface plasmon nanoparticle modification [26,35], element doping [27,36], etc., which proves that this materials have great development potential for photocatalysis, for example Bi₇Fe₂Ti₂O₁₇X (X = Cl, Br, I) (BFTOX). In 2022, Yan Gu et al. fabricated Bi₇Fe₂Ti₂O₁₇Cl successfully by using a one-step flux route and solid-state reaction, and its degradation efficiency of tetracycline (TC) reaches 90% within 90 min, which confirms that BFTOX should be a potential novel visible light photocatalyst material [37].

In order to further advance the development of Bi-based oxyhalides semiconductor materials, a series of novel Sillén–Aurivillius layered oxyhalide materials, Bi₇Fe₂Ti₂O₁₇X (X = Cl, Br, I) (BFTOX), have been successfully synthesized via a molten-salt method by using the BiOX (X = Cl, Br, and I) precursor powders. Originally, we display the investigation of material characteristics, such as the band structure, effective mass, lattice constant, and E_g of BFTOX influenced by the halogen, Cl, Br, and I, from density functional theory (DFT) calculation. Then, material properties of BFTOX powders are determined by characterization techniques of X-ray diffraction (XRD), scanning electron microscope (SEM), energy dispersive spectroscopy (EDS), and X-ray photoelectron spectroscopy (XPS). Furthermore, the photocatalytic activities of different samples are compared by the reduction of CO₂ and the degradation of methyl orange (MO) and Rhodamine B (RhB), and BFTOX exhibits the best photocatalytic performance. In addition, the possible photocatalysis mechanisms of BFTOX have also been discussed in detail.

2. Results and Discussion

2.1. DFT Calculations

The structure of Sillén-Aurivillius materials is composed of fluorite ([Bi₂O₂]²⁺) [20], like-perovskite ([A_n−1B_nO_3n+1]³⁻) [21], and halogen ([X]) with the general formula of [Bi₂O₂]²⁺[X]^-[Bi₂O₂]²⁺[A_n−1B_nO_3n+1]³⁻ [22], where n represents the number of perovskite layers (n = 1, 2, 3…). Meanwhile, with the number of perovskite layers increasing, the difficulty of implementation will also increase, whether in calculation or experiment. After drawing on many previous works, this work put a sight on comparing three quadruple-perovskite layers semiconductor photocatalysts, Bi₇Fe₂Ti₂O₁₇X (X = Cl, Br, I). The layers are constructed in the Cmcm space group (No. 63), aiming to provide the high efficiencies of photocatalysts for the reduction of CO₂ and degradation of organic pollutants. The crystal structure of BFTOX in Figure 1a shows that the novel BFTOX is an n = 4 member of Sillén–Aurivillius type perovskite oxyhalides with alternate stacking of Aurivillius perovskite blocks ([Bi₃Ti₂Fe₂O₁₃]³⁻) and Sillén blocks ([(Bi₂O₂)₂X]³⁺).

In order to further investigate the electronic properties of BFTOX, the energy band structure plots are displayed in Figure 1b–d. We can find that all BFTOX samples belong to the indirect-gap semiconductor. The calculated E_g values of BFTOC, BFTOB, and BFTOI are 2.30 eV, 2.44 eV, and 2.65 eV, respectively. As the halogen atomic number increases from Cl to Br and I, the band gap gradually widens, which may be related to the electronegativity of the halogen atoms. Furthermore, the electronegativity of materials also influences the energy level of the halogen atoms that elements with low electronegativity have high energy level positions, and thus the halogen level (Cl 3p, Br 4p and I 5p) increases as the atomic number increases relative to the O 2p orbitals. Quite unexpectedly, the VBM of BFTOC and BFTOB is predominantly composed of O 2p orbitals, while that of BFTOI is contributed by I 5p orbitals. For BFTOI, the VBM contribution is similar to that of simple oxyhalides BiOCl and BiOBr. During photocatalytic process, it is inevitable to be oxidized by the holes excited by the light, and the stability will be poor [20]. When it comes to BFTOC and BFTOB, the VBM is mainly composed of the dispersive O 2p orbitals. Meanwhile, the calculated effective masses of all samples in

Table 1. Calculated effective masses (unit m₀) of the electrons and holes of BFTOC, BFTOB and BFTOI by different directions.

	BFTOC		BFTOB		BFTOI
Effective Mass	Electron	Hole	Electron	Hole	Electron	Hole
mxx (Γ−X)	0.023	0.181	0.013	0.131	0.053
myy (Γ−Y)	0.035	0.099	0.018	0.098	0.054
mzz (Γ−Z)	9.879	7.904	19.004	6.228	5.137
mxx (S−Y)						0.035
myy (S−X)						0.031

confirm that the lower values of BFTOC and BFTOB. Hence, we believe that the excited holes in the stable oxygen anion near the VBM of both materials can effectively oxidize water without inactivation due to self-decomposition phenomenon of catalysts upon light irradiation.

2.2. Material Properties of BFTOX (X = Cl, Br, and I)

The crystalline phases of as-prepared BFTOX powders were determined by XRD technology, and the corresponding results are displayed in Figure 2a–c. It can be observed that the diffraction peaks at 30.58°, 32.65°, 46.73°, and 56.97° can be indexed to the (116), (020), (220), and (316) crystal planes of BFTOX samples, which are all consistent with the calculated cards from DFT. These results confirm that BFTOX (X = Cl, Br, and I) materials are well-crystallized after calcination at different high temperatures in this experiment. Moreover, from the DFT results in Figure 1a, the structure of BFTOX (X = Cl, Br, and I) is the orthorhombic structure along with the space group of P21cn. In this case, the experimental lattice constants of BFTOX are calculated (BFTOC: a = 5.508 Å, b = 5.481 Å, and c = 26.593 Å; BFTOB: a = 5.501 Å, b = 5.478 Å, and c = 26.592 Å; BFTOI: a = 5.492 Å, b = 5.478 Å, and c = 26.644 Å). These results and the theoretically calculated lattice constants in Figure 2a–c are extremely similar, further indicating the successful preparation of BFTOX.

Afterwards, the microstructure of the as-prepared BFTOX samples is determined by using SEM. As depicted in Figure 2d–f, all samples possess a nanoplate-like structure, which may be caused by the weak Van der Waals’ force between the halogen layer [X] and the fluorite layer [Bi₂O₂]. Meanwhile, the size of nanoplates of BFTOX samples gradually decreased from the halogen variation in Sillén–Aurivillius material, particularly for BFTOI. One possible reason for this phenomenon is the stability of the material structure. Because the iodine element (I) is easy to sublimate during the high temperature calcination process, it is easy to destroy the crystal structure, so the crystallinity is poor and it is easy to form nano-agglomerates. In addition, the SEM-EDS mappings of BFTOB nanoplates in Figure 2g–k confirm the existence of Bi, Fe, Ti, O, and X and they are evenly distributed on the whole nanoplates. A similar phenomenon is also observed in BFTOC and BFTOI samples, seen the supporting information in Figure S1. These further demonstrate the successful synthesis of the BFTOX materials.

The elements of Bi, Fe, Ti, O and Br are found in the survey XPS spectra of BFTOBr (Figure 3a), which is in accordance with the results of SEM-EDS. The peaks at about 159.10 and 164.42 eV in Figure 3b, obtained from BFTOB, are ascribed to the Bi 4f_7/2 and Bi 4f_5/2 of Bi³⁺ [38]. In the Fe 2p region of BFTOB (Figure 3c), there were two peaks at 711.09 and 725.36 eV, referring to Fe 2p_1/2 and Fe 2p_3/2, respectively, manifesting the existence of Fe³⁺ ions [39]. There were two fitted Ti 2p peaks centered at 458.12 and 465.73 eV in Figure S2, which were attributed to Ti 2p_3/2 and Ti 2p_1/2 to the oxidation state of Ti⁴⁺ [40,41]. The O 1s peaks at binding energies of 529.50, 530.17 and 531.99 eV, which arises from the Bi-O, Fe-O and Ti-O, respectively (Figure S3). In addition, the binding energy peaks of Br 3d_5/2 and Br 3d_3/2 are located at 68.46 and 79.45 eV, respectively (Figure 2d) [42,43].

2.3. Photocatalytic CO₂ Reduction Performance for BFTOX Materials

In order to evaluate the photocatalytic activities of BFTOX affected by the halogens, Cl, Br, and I, we conducted CO₂ reduction under the light irradiation. Under the simulated sunlight illumination, all BFTOX (X = Cl, Br, and I) samples can reduce the CO₂ to carbon monoxide (CO) in a sealed gas-solid phase reaction vessel without any organic sacrificial reagent. Figure 4a–c displays the relatively steady CO yields as a function of the irradiation time generated by BFTOC, BFTOB, and BFTOI. The yield values are about 14 μmol, 12 μmol, and 10 μmol for 210 min, respectively. To compare the performance clearly, the average yields of CO in 3.5 h are calculated in Figure 4d. It can be observed that the production rate of CO for BFTOC and BFTOB are 3.74 μmol g⁻¹ h⁻¹ and 3.6 μmol g⁻¹ h⁻¹, which are approximately 1.4 times higher than that of BFTOI (2.66 μmol g⁻¹ h⁻¹). This result means that the trend of photocatalytic activities of BFTOX is BFTOC > BFTOB > BFTOI. The decrease in the CO production rate for samples may be related to the absorption of solar photons.

Subsequently, the UV–Vis diffuse absorbance spectrum (DRS) is used to evaluate the light absorption of BFTOX samples. It can be seen that all samples exhibit a strong photo-response in the range of 390–550 nm in Figure 5a, which indicates that the BFTOX (X= Cl, Br, and I) materials are visible light catalysts. It is observed also that, the order of light response intensities is BFTOC > BFTOB > BFTOI, which can be ascribed to the sub-band gap absorption of powder materials. As we know, the significant sub-band gap absorption of films is indicative of the presence of the secondary phases or the band tailing [44,45,46,47]. This observation suggests the BFTOC powder has the lowest defect concentrations and photogenerated carrier recombination compared with the two other samples, which should be the reason for the highest performance of CO₂ reduction for BFTOC sample, as is consistent with the SEM results.

In addition, the BFTOC powder has the lowest defect concentrations compared with the two other samples. This should be the reason for the highest performance of CO₂ reduction for BFTOC sample, as is consistent with the SEM results. Furthermore, as the atomic number of the halogen increases in BFTOX (X = Cl, Br, and I), the absorption edge decreases, thus widening the E_g of samples. For a clearer comparison, the E_g values are determined by applying the Kabelka-Munk function [48,49,50]:

$${\left(\alpha hv\right)}^{1/m}=A\left(hv - E_g\right)$$

(1)

where A, $\alpha$, and hv are constant, absorption coefficient, and photon energy, respectively. The constant m is connected with the type of materials: the direct-gap semiconductor or the indirect-gap semiconductor [49]. From DFT calculation results, BFTOX is an indirect-gap semiconductor. Therefore, the value of m is 2. By the linear intersection in Figure 5b, the corresponding E_g values of BFTOX (X = Cl, Br, and I) are 2.64 eV, 2.68 eV, and 2.74 eV, respectively. This result is consistent with the trend of DFT calculation, as depicted in Figure 1.

In general, the positions of the valence band maximum (VBM) and conduction band minimum (CBM) of photocatalysts determines the photocatalytic activities of CO₂ reduction under light irradiation. Here, ultraviolet photoelectron spectroscopy (UPS) is employed to clarify the VBM and CBM positions of BFTOX samples, and the results of BFTOB and BFTOI are shown in Figure 5c. The result of BFTOC can be found in our previous work [44]. It can be calculated that the values of secondary cutoff energy (E_cut-off) of BFTOX are 16.30 eV, 16.30 eV, and 16.40 eV, respectively. And the position energy of corresponding Fermi levels (E_F) can be deduced from the equation [49]:

$$\phi =hv - E_{\mathrm{cut}-\mathrm{off}}$$

(2)

where hv is about 21.22 eV for the He I electron energy. Therefore, the E_F values of BFTOC, BFTOB, and BFTOI are 4.92 eV, 4.92 eV, and 4.82 eV, respectively. In addition, the on-set energy (E_on-set) of samples can be applied to determining the gap between valence band and E_F. From Figure 5c, the E_on-set values of BFTOC, BFTOB, and BFTOI are 1.80 eV, 1.64 eV, and 1.63 eV. Therefore, the valence band positions of these samples are 6.72 eV, 6.56 eV, and 6.45 eV, respectively. As the E_g values are depicted in Figure 5b, the conduction band positions are 4.06 eV, 3.88 eV, and 3.72 eV from the equation that E_CB = E_VB − E_g. According to previous works, the relationship between the vacuum energy (E_abs) and the normal electrode potential (E^Θ) can be described as follows [51]:

$$E_{\mathrm{abs}}=-E^{\mathsf{\Theta }} - 4.44$$

(3)

Correspondingly, the valence band positions of BFTOC, BFTOB, and BFTOI (vs. NHE) are 2.28 eV, 2.12 eV, and 2.01 eV, while the valence band positions of samples are −0.38 eV, −0.56 eV, and −0.72 eV, respectively. Based on the above discussions, the energy band diagrams of the CBM and VBM potentials for BFTOX samples are exhibited in Figure 5d. We can find that as the halogen in BFTOX is substituted from Cl to I, both the conduction and valence bands of samples are shifted upward. Meanwhiles, the E_g values are increased from 2.64 eV to 2.74, this result means that the widening of the band gap is mainly due to the shifting of the conduction band. Since the CBM of BFTOX (−0.38 eV, −0.56 eV, and −0.72 eV) are more negative than that of the CO₂/CO (−0.11 eV) redox potential, it confirms that the valence bands of BFTOX (X = Cl, Br, and I) are capable of CO₂ reduction, as the results in Figure 4.

In general, the size of photocatalyst nanoplates can directly affect the photocatalytic activity. When the nanoplate size is smaller, the number of nanoplates per unit mass is larger, and the specific surface area is larger. According to SEM results in Figure 2, the size of nanoplates of BFTOX samples gradually decreased from the halogen variation in Sillén–Aurivillius material, particularly for BFTOI. These results indicate that the BFTOI sample has the largest specific surface area, and it should have the maximum photocatalytic efficiency. However, this result is the opposite, which may be related to the generation and recombination of photogenerated charge carriers. As we know, the photocatalytic performance of BFTOX samples is affected by many factors, such as the separation and transfer of photogenerated electrons and holes. Here, photoelectrochemical characterizations, transient photocurrent densities (TPC) and electrochemical impedance spectra (EIS), are employed to investigate the charge transfer capability of the catalysts. Figure 6a shows the photocurrent response curves of BFTOX under light irradiation. It can be found that the trend of the photocurrent density of BFTOC is BFTOC > BFTOB > BFTOI. This result reveals the higher efficient separation of photo-generated electrons and holes for the BFTOC sample. Meanwhiles, a similar conclusion is also observed in the results of EIS in Figure 6b. The diameter of the arc radius of BFTOC on the EIS is much smaller that these of BFTOB and BFTOI, indicating that BFTOC sample possesses a faster transfer of the photo-generated electrons and holes. This excellent performance of separation and transfer of photogenerated carriers for BFTOC can be related to the highest quality of powder material along with the lowest defect concentrations, as exhibited in Figure 5a. In addition, the calculated effective masses of the electrons and holes in Table 1 shows that the values of BFTOB is smallest than other two samples, it suggests that the electron-hole pair recombination of BFTOB should be lowest and the BFTOB sample possesses the highest photocatalytic efficiency. However, the sample with the highest photocatalytic efficiency is BFTOC in experiment. The reason for this result is that the BFTOC sample has higher crystallization compared with the other two samples.

2.4. Photocatalytic Activity Enhancement Mechanism for the Photodegradation of Organic Pollutants

The photocatalytic activities of BFTOX samples are also measured by using MO and RhB (10 mg/L) as the simulated pollutant under light irradiation, the results are displayed in Figure 7. To get a clear contrast, the degradation efficiencies of MO and RhB by BFTOX at different concentrations are calculated by the following equation [36]:

$$\eta =(C_0 - C) / C_0 \times 100\%$$

(4)

where C and C₀ are the sampling concentration of the reaction solution at different times and the initial concentration of MO and RhB solutions before light irradiation, respectively. From Figure 7a, we can find that the trend of degradation efficiencies for MO is BFTOC > BFTOB > BFTOI, and its result is similar for that of RhB solutions. Moreover, according to the fitting of pseudo-first-order: ln(C₀/C) = kt, the degradation constants (k) are evaluated, as depicted in Figure 7b.

For MO, the BFTOC sample exhibits the highest apparent rate constant of about 0.011 min⁻¹, which is around 1.18 and 1.30 times of bare BFTOB and BFTOI samples, respectively. In addition, it can be observed that the degradation efficiencies of MO over BFTOX are higher than that of RhB, which indicates that that photocatalytic degradation of BFTOC powder is best, particularly for MO. Besides, BFTOC sample was also used to degrade the MO and RhB mixed pollutants, and its result is shown in Figure. 7c. Obviously, the degradation efficiencies of MO at the wavelength of approximately 460 nm is higher compared with that of RhB at 550 nm. This further confirms the above conclusions. More importantly, the degradation efficiency of the mixture of MO and RhB is higher than that of single MO, which might be ascribed to the electrostatic force promoted the degradation of MO [52]. Moreover, the degradation efficiency of the mixture of MO and RhB is lower than that of single RhB due to the active sites of BFTOX being encircled intermediates, decelerating the production of electro-hole pairs and restraining the formation of radicals. In addition, the degradation efficiencies of MO are also higher than that of RhB in the mixture of MO and RhB. Namely, the degradation rates of MO over the BFTOX nanoplates are always higher than those of RhB, it may be mainly attributed to the molecular structure of oxychlorides, which needs further research in the future.

Afterwards, the active species trapping experiments of BFTOX (X = Cl, Br, and I) were carried out under light irradiation to evaluated the exact photocatalytic degradation mechanism. Radical scavenging materials, such as BQ, EDTA-2Na and TBA, are employed as the scavengers of ∙OH, h⁺, and $·{\mathrm{O}}_2^{-}$, respectively. The results in Figure 7d exhibit that the degradation efficiencies of MO and RhB over BFTOC are significantly inhibited under BQ and EDTA-2Na, while TBA only slightly reduces the degradation. In addition, similar capture trends are also found for BFTOB (Figure 7e) and BFTOI (Figure 7f). Therefore, h⁺ and $·{\mathrm{O}}_2^{-}$ should be the major active species for MO and RhB degradation by BFTOC.

In fact, the stability and durability of photocatalysts is very important in practical applications. Figure 8a exhibits that the RhB degradation efficiency over BFTOC nanoplates with the same photocatalyst for five cycles. It can be found that the excellent photocatalysis performance of BFTOC is still significant after five cycles use, which confirms that the good recycling ability of BFTOC nanoplates. In addition, the stability of sample was also ascertained from XRD pattern in Figure 8b. The crystallinity of the as-prepared sample is similar to that of the used samples, it indicates that BFTOC nanoplates have the potential for the practical applications.

Based on the aforementioned discussions, the probable photocatalytic degradation mechanism of organic pollutants over BFTOX materials is proposed in Figure 9. Under the light illumination, the e⁻ and h⁺ will be produced on the CBM and VBM of BFTOX. The accumulated e⁻ on the CBM of BFTOX readily can reduce oxygen to $·{\mathrm{O}}_2^{-}$, which has more negative potentials (−0.38 eV, −0.56 eV, and −0.72 eV vs. NHE) than the O₂/${\mathrm{O}}_2^{-}$ redox potential (−0.33 eV vs. NHE). The reason why ∙OH ions do not contribute to the photocatalytic degradation process of organic pollutants is that in the acidic reaction solution, the E_VBs of BFTOX (BFTOC 2.28 eV; BFTOB 2.12 eV; BFTOI 2.01 eV) are more negative than the redox potential of H₂O/·OH (2.73 eV), and photogenerated holes cannot induce H₂O to ∙OH, as shown in Figure 5d. Finally, the free radicals $·{\mathrm{O}}_2^{-}$ and e⁻, which migrate to the surface of the powder and have good oxidation ability, directly degrade MO and RhB molecules into oxidation products. The possible photocatalytic mechanism of BFTOX powders can be described as follows:

$${\mathrm{Bi}}_7{\mathrm{Fe}}_2{\mathrm{Ti}}_2{\mathrm{O}}_{17}\mathrm{X} (\mathrm{X}=\mathrm{Cl}, \mathrm{Br}, \mathrm{and} \mathrm{I})+\mathrm{hv}+\left(\mathrm{ultrasound}\right) \to { \mathrm{e}}^{-}{+\mathrm{h}}^{+}$$

(5)

$${\mathrm{O}}_2{+\mathrm{e}}^{-} \to ·{\mathrm{O}}_2^{-}$$

(6)

$$·{\mathrm{O}}_2^{-}{+\mathrm{h}}^{+}+\mathrm{RhB}/\mathrm{MO}\to \mathrm{Intermediates} +{\mathrm{CO}}_2 + {\mathrm{H}}_2\mathrm{O}$$

(7)

3. Materials and Methods

3.1. Materials

Bismuth trioxide (Bi₂O₃) was purchased from Shanghai Aladdin Biochemical Technology CO., Ltd. Ferric oxide (Fe₂O₃) was purchased from Shanghai Shanpu Chemical Co., Ltd., Shanghai, China. Titanium dioxide (TiO₂), polyethylene glycol (EG), potassium chloride (KCl), potassium bromide (KBr), potassium iodide (KI), sodium chloride (NaCl), sodium bromide (NaBr), and sodium iodide (KI) were brought from Sinopharm Chemical Reagent Co., Ltd., Shanghai, China. All chemical reagents were analytical grade and directly used without any further treatment.

3.2. Preparation of Photocatalysts

3.2.1. Preparation of BiOX (X = Cl, Br, and I)

In this experiment, BiOX (X = Cl, Br, I) was prepared by a facile soft liquid deposition method, as reported in our previous works [27,49,50,51]. 24 mmol of Bi(NO₃)₃·5H₂O was dissolved into 240 mL of ethylene glycol solution (EG) to obtain solution A at room temperature. Then, 24 mmol of KX (X = Cl, Br, and I) was dissolved into 120 mL of deionized water to obtain solutions B, which was dropped slowly into the previous solution A. The mixed solution was magnetically stirred for 1 h at room temperature. Subsequently, the products were obtained by washing them with deionized water several times and dried at 60 °C for 12 h.

3.2.2. Preparation of Bi₇Fe₂Ti₂O₁₇X (X = Cl, Br, and I)

BFTOX powders were synthesized by a flux method using Bi₂O₃, BiOX, Fe₂O₃ and TiO₂ as the precursor materials. KCl/NaCl, KBr/NaBr, and KI/NaI with a ratio of 1:1 were served as the molten salt. The molten salt was mixed with the solute of Bi₂O₃, BiOX, Fe₂O₃ and TiO₂ at the stoichiometric molar ratio for BFTOX (3:1:1:2). The solute concentration is about 6.25 mol% (BFTOX/(BFTOX + flux)). 10 mol% excess of BiOX over the stoichiometric ratio was added to prevent volatilization. Afterwards, three different mixtures were transferred to alumina crucibles for the high-temperature calcination. The Bi₇Fe₂Ti₂O₁₇Cl (BFTOC), Bi₇Fe₂Ti₂O₁₇Br (BFTOB), and Bi₇Fe₂Ti₂O₁₇I (BFTOI) were obtained at the calcination temperature of 750 °C, 750 °C, and 700 °C for 6 h, respectively. After natural cooling, the obtained products were collected by filtration and thoroughly washed with the deionized water for several times. Finally, the products were dried at 60 °C for 12 h.

3.3. Characterization

X-ray diffraction (XRD) (X’ Pert3 Powder, PANalytical, Almelo, the Netherlands, λ = 0.15406 nm) was used to analyze the crystal structure of BFTOX samples with Cu Kα radiation. The sample morphologies were examined by field emission scanning electron microscope (FE-SEM) (FEI, Hillsboro, OR, USA). X-ray photoelectron spectroscopy (XPS) was carried out to investigate the chemical nature of the samples via a Thermo Scientific Escalab 250 Xi spectrometer (Thermo Fisher Scientific, Carlsbad, CA, USA). The UV-vis diffuse reflectance spectroscopy (UV-vis DRS) was attained by using a UV-2550 (Shimadzu, Kyoto, Japan). The UV-vis absorption spectra of the samples were performed by UV-2450 (Shimadzu, Kyoto, Japan).

3.4. Photocatalytic Experiments

Photocatalytic experiments of the samples were discussed with a 300 W Xenon lamp (CEL-HXF300 Beijing China Education Au-light Co., Ltd., Beijing, China) as the light source. 100 mg of photocatalysts were separately added into 100 mL of 10 mg/L MO and RhB, and then stirred in the dark for 30 min to reach adsorption-desorption equilibrium. 4 mL solution was withdrawn every 30 min. The above suspension was centrifuged and the residual concentration of dye was detected by UV-vis spectrophotometer. During the degradation process of MO and RhB over BFTOX, ∙OH, h⁺, and $·{\mathrm{O}}_2^{-}$ were captured by a certain proportion of tert-butyl alcohol (TBA, 1 mM), EDTA (1 mM), and benzoquinone (BQ, 0.25 mM), respectively.

The experiments on the reduction of CO₂ were performed in a gas closed-circulation system. Photocatalyst powders of 0.01 g were dispersed in the absolute ethyl alcohol. Then, the photocatalysts solution was coated on a glass slide and placed in a reaction chamber (1 L). Deionized water entered the reaction chamber in the form of steam, and 1000 ppm CO₂ was introduced. After the whole system was stabilized, a 300 W Xe-arc lamp was applied to simulate sunlight.

3.5. Electrochemical Measurement

Photocurrent density versus time curves and electrochemical impedance spectra (EIS) were performed by a CHI660D electrochemical analyzer (Chenhua Instruments Co., Ltd., Shanghai, China) with a standard three-electrode system, including a working electrode, a reference electrode (saturated Ag/AgCl), and a counter electrode (Pt wire). The working electrode was obtained: 5 mg of as-prepared photocatalysts were dispersed into 40 μL of naphthol to make the slurry, and the slurry was deposited on the fluorine-tin oxide (FTO)glass. Afterwards, the substrate was dried at 80 °C for 4 h and calcined in a muffle furnace at 120 °C for 1 h.

3.6. Theoretical Calculations

In this work, the calculation of DFT with the projector-augmented wave (PAW) was carried out by VASP code [53]. The generalized gradient approximation (GGA) for the electron exchange and correlation corrections used Perdew-Burke-Ernzerhof (PBE) [54] to optimize the geometric structure and the next procedures were performed within the Heyd-Scuseria-Ernzerhof screened hybrid functional (HSE) [55], including band structures. In order to obtain accurate results, the cut-off energy was set to 520 eV and the irreducible Brillouin zone grid was used to 5 × 5 × 1 mesh. The energy convergence criterion was 1 × 10⁻⁴ eV as well as −1 × 10⁻² eV for atomic force convergence. In addition, the magnetic moment of the Fe atom was set to 4 μB. Due to the unpredictability of the synthesized samples in the experiment, Fe and Ti alternate arrangement is used to simulate the disorder situation in this DFT calculated process.

4. Conclusions

In summary, a series of novel Sillén–Aurivillius oxygen halides, Bi₇Fe₂Ti₂O₁₇X (X = Cl, Br, and I), were successfully synthesized by molten salt method for the efficient CO₂ reduction and degradation of organic pollutants. The highly dispersive O 2p band of BFTOC and BFTOB samples contributing to the VBM not only renders the VBM essentially negative, but also explains its stability against self-decomposition deactivation from theory and experiment. For BFTOI, self-oxidation deactivation of I 5p orbitals leads to inferior CO₂ reduction in BFTOI. BFTOC and BFTOB samples possess outstanding activity for CO₂ reduction to CO, and the production rates of CO are 3.74 μmol g⁻¹ h⁻¹ and 3.6 μmol g⁻¹ h⁻¹, respectively. More importantly, the BFTOX with unique nanosheet morphology displays a superb photocatalytic performance on MO and RhB mixture degradation. The removal efficiencies of the MO and RhB for BFTOC and BFTOB are 90% and 85% higher than BFTOI, which is a result of the improved separation of photogenerated charge carriers and optical absorption. To be more specific, the radical trapping tests and possible mechanism of BFTOX for the photocatalytic degradation of MO and RhB indicated that the h⁺ and ·O₂^- were the major active species to boost photocatalytic reactions. As an attractive photocatalyst, BFTOX has promising potential for CO₂ reduction and the degradation of organic pollutants.