Direct conversion mechanism from BiOCl nanosheets to BiOF, Bi₇F₁₁O₅ and BiF₃ in the presence of a fluorine resource

Abstract

In this work, we report the facile conversion from BiOCl nanosheets to BiOF, Bi₇F₁₁O₅ and BiF₃ by a novel ion exchange approach, in which the effects of fluorine source, F/Bi molar ratio and reaction medium (ethanol/water) on the products are mainly investigated. A plausible conversion mechanism is proposed to illustrate the formation of BiOF, Bi₇F₁₁O₅ and BiF₃. Furthermore, the photocatalytic activities of the samples are also investigated. It is amazing that under ultraviolet light irradiation (λ ≤ 420 nm), the activity of the as-formed Bi₇F₁₁O₅/BiOCl sample is 3.28 times higher than that of BiOCl for the degradation of methyl orange (MO). It is demonstrated that the improved activity is mainly attributed to the formation of Bi₇F₁₁O₅/BiOCl heterojunction, which has significantly improved the separation and transfer efficiency of charges. It is important that this study provides an initiative post-synthesis strategy to develop new, efficient photocatalysts.

---

1. Introduction

Solar energy-driven photocatalysis, as a green advanced oxidation technology, has attracted much attention. However, its practical applications is still limited by low quantum efficiency, light utilization efficiency and photoactivity.^1–3 Thus, it remains a big challenge to develop inexpensive, highly efficient photocatalysts for practical applications. To date, various photocatalysts have been developed.^4–8 Recently, bismuth-based compounds have been intensively investigated due to the outstanding optical and photocatalytic properties, including Bi₂O(OH)₂SO₄,^9,10 Bi(C₂O₄)OH,¹¹ [Bi₆O₆(OH)₃](NO₃)₃·1.5H₂O,¹² BiOCl,¹³ BiPO₄,¹⁴ Bi₂O₂CO₃,¹⁵ BiOF,^16–18 BiF₃,¹⁹ Bi₇F₁₁O₅,²⁰ BiOI,²¹ BiVO₄ ^22,23 and so on. However, their photocatalytic activities are still not high enough to satisfy the practical needs. Fluorination of semiconductors has been demonstrated to be an effective method to improve the photocatalytic activities.^24–26 For example, Chen et al.²⁷ have reported that the photocatalytic activity of the fluorinated ZnWO₄ is obviously enhanced for the decomposition of rhodamine B (RhB) under ultraviolet light irradiation. On one hand, the surface fluorination can promote the dye adsorption and the interfacial charge transfer dynamics; on the other hand, the doping of fluorine in lattice can introduce special localized electronic and defect states.²⁸ These results have also been observed in fluorinated TiO₂,²⁹ Bi₂O₃,³⁰ SrTiO₃,³¹ BiPO₄,³² etc.

Inspired by the results above, a simple ion exchange approach is developed to convert BiOCl to BiOF, Bi₇F₁₁O₅ and BiF₃, in which fluoride is mainly employed as the fluorination chemical. A plausible conversion mechanism is proposed and discussed; meanwhile, we have investigated their photocatalytic activities for the degradation of MO. The results show that Bi₇F₁₁O₅/BiOCl heterojunction shows the significantly improved photocatalytic activity. To the best of our knowledge, this ion exchange approach has not been explored for the conversion of bismuth-based semiconductors so far. This work provides an initiative post-synthesis strategy to prepare the other photocatalysts.

2. Experimental

2.1. Sample preparation

All reagents were of analytical grade, purchased from Beijing Chemical Reagents Industrial Company of China, and were used without further purification.

2.1.1. Preparation of BiOCl. BiOCl sample was prepared according to previously reported method.³³ Typically, 1.5 mmol of Bi(NO₃)₃·5H₂O was dissolved in 30 mL of ethylene glycol (EG) under stirring at room temperature. After the Bi(NO₃)₃·5H₂O was completely dissolved, 10 mmol of NaCl was added to the above solution. After being stirring for 30 min, the solution was transferred into a 50 mL Teflon-lined stainless steel autoclave, and then maintained at 170 °C for 6 h. After reaction, the autoclave was cooled to room temperature naturally. The resulting sample was centrifuged and washed with ethanol and distilled water three times, and then dried at 60 °C for 3 h.

2.1.2. Ion exchange. Typically, the as-obtained BiOCl (0.2605 g) above was dispersed into a 20 mL of ethanol/water mixture containing the desired amount of NH₄F. Then, the dispersion was transferred to Teflon-lined stainless steel autoclave, and maintained at 170 °C for 6 h. The resulting precipitate was centrifuged, washed with ethanol and distilled water three times and dried at 60 °C for 3 h.

To investigate the effect of NH₄F amount added, F/Bi molar ratios (R_F) were varied from 0 to 0.5, 1, 2, 4, 6 and 8, while keeping the other conditions constant.

To investigate the effect of fluorine source, NH₄F was substituted by HF or NaF at R_F = 2, while keeping the other conditions constant.

To investigate the effect of reaction medium, V_Eth/V_w was varied from 0/20 to 2/18, 10/10 and 20/0 at R_F = 2, while keeping the other conditions constant.

2.2. Characterization

The crystal structures of the samples were determined by X-ray powder polycrystalline diffractometer (XRD, Rigaku D/max-2550VB), using graphite monochromatized Cu K_α radiation (λ = 0.154 nm), operating at 40 kV and 50 mA. The XRD patterns were obtained in the range of 10–80° (2θ) at a scanning rate of 7° min⁻¹. The samples were characterized on a scanning electron microscope (SEM, Hitachi SU-1510) with an acceleration voltage of 15 keV. The samples were coated with 5 nm-thick gold layer before observations. The fine surface structures of the samples were determined by high-resolution transmission electron microscopy (HRTEM, JEOL JEM-2100F) equipped with an electron diffraction (ED) attachment with an acceleration voltage of 200 kV. UV-vis diffused reflectance spectra of the samples were obtained using a UV-vis spectrophotometer (UV-2550, Shimadzu, Japan). BaSO₄ was used as a reflectance standard in a UV-vis diffuse reflectance experiment. Surface area was calculated by the Brunauer–Emmett–Teller (BET) method. Electrochemical impedance spectroscopy (EIS) was performed from 0.01 Hz to 100 kHz at an open circuit potential of 0.3 V and alternating current (AC) voltage amplitude of 5 mV. The photo electrodes were prepared by a dip-coating method and 0.1 M Na₂SO₄ was used as the electrolyte solution. The data were analyzed by ZSimWin software.

2.3. Photocatalytic degradation reactions

The photodegradation of MO dye was carried out to evaluate the catalytic activities under ultraviolet light irradiation (λ ≤ 420 nm), using a 500 W Xe arc lamp (CEL-HXF 300) equipped with a cutoff filter (λ > 420 nm) as a light source. Typically, 0.1 g of powders was suspended in 200 mL aqueous solution of MO (12.5 mg L⁻¹) under continuous magnetic stirring. Before light-on, the suspension was stirred for 30 min to reach an adsorption–desorption equilibrium of dye molecules on the surface of the photocatalyst. The adsorption curves of the samples are shown in Fig. S1 (ESI†), which show that an adsorption–desorption equilibrium of dye molecules on the surface of the photocatalyst can be reached in 30 min. During reaction, 3 mL of suspension was collected at a given interval time and centrifuged to remove the powders. The concentration of dye was determined by using UV-vis spectrophotometer. Herein, the photodegradation of MO can be simplified as a pseudo-first-order kinetic reaction, the apparent reaction rate constants (k_a) are calculated by the formula (1) as follows.

ln(C₀/C) = k_at (1)

where C is the concentration of MO remaining in solution after irradiation, and C₀ is the initial concentration of MO before irradiation.

2.4. Theoretical calculations

The simulations of energy band structures, total and a part of densities of states (T- and P-DOS) were calculated by density functional theory (DFT) as implemented in the CASTEP. The calculations were carried out using the generalized gradient approximation (GGA) level and Perdew–Burke–Ernzerhof (PBE) formalism for combination of exchange and correlation function. The cut-off energy is chosen as 380 eV, and a density of (3 × 2 × 5) Monkhorst–Pack K-point was adopted to sample the Brillouin zone.

3. Results and discussion

3.1. Effect of F/Bi molar ratio (R_F) on the samples

3.1.1. Crystal phases, morphologies and textures of the samples. Fig. 1 shows the XRD patterns of the untreated BiOCl and treated samples prepared at different NH₄F amounts added. All the diffraction peaks of the untreated sample are in good agreement with the standard BiOCl (JCPDS no. 06-0249), indicating that the formation of phase-pure BiOCl. At R_F = 0, the crystalline phase of the sample does not change, however, the crystallinity becomes stronger than the untreated one. At R_F = 0.5, BiOF (no. 73-1595) has formed besides BiOCl, indicating that a part of BiOCl has converted into BiOF. At R_F = 1, Bi₇F₁₁O₅ (no. 50-0003) has formed, besides both BiOF and BiOCl; meanwhile, the diffraction peaks of BiOF become weak. At R_F = 2, both Bi₇F₁₁O₅ and BiOCl were observed, indicating the Bi₇F₁₁O₅/BiOCl heterojunction is formed. At R_F = 4, the phase-pure Bi₇F₁₁O₅ has formed. Herein, it should be noted that the standard no. 50-0003 only ranges in 10–60°, thus only the weak diffraction peaks in the range of 10–60° are checked. At R_F = 6, the sample consists of BiF₃ (no. 73-1988) as the main phase, besides a small amount of Bi₇F₁₁O₅. At R_F = 8, the phase-pure BiF₃ has formed. It is obvious that BiOCl can be converted into phase-pure Bi₇F₁₁O₅ and BiF₃ in the ethanol/water mixture (V_Eth/V_w = 2/18). For Bi₇F₁₁O₅/BiOCl sample (synthesized at R_F = 2), we have also investigated the influence of reaction time on the sample (Fig. S2, ESI†). By calculating the intensity ratios of the strongest peaks of Bi₇F₁₁O₅ and BiOCl phases (1–12 h), the reaction time nearly have no obvious influence on the composition in the Bi₇F₁₁O₅/BiOCl mixture.

Fig. 1 X-ray diffraction (XRD) patterns of the untreated BiOCl and treated samples by NH₄F at different molar ratios of F to Bi (R_F): (a) untreated BiOCl; (b) R_F = 0 (BiOCl); (c) R_F = 0.5 (BiOF/BiOCl); (d) R_F = 1 (BiOF/Bi₇F₁₁O₅/BiOCl); (e) R_F = 2 (Bi₇F₁₁O₅/BiOCl); (f) R_F = 4 (Bi₇F₁₁O₅); (g) R_F = 6 (Bi₇F₁₁O₅/BiF₃); (h) R_F = 8 (BiF₃). Preparation conditions: V_Eth/V_w = 18/2 (volumetric ratio of ethanol to water).

Additionally, the EDS of Bi₇F₁₁O₅/BiOCl (synthesized at R_F = 2) is shown in Fig. S3 (ESI†), which confirm that Bi, O, Cl and F elements were existed in the sample, moreover, the detail contents of Bi, O, Cl and F elements are listed in Table S1,† which demonstrated that the molar ratio of Bi₇F₁₁O₅ and BiOCl is about 35% in Bi₇F₁₁O₅/BiOCl sample.

The SEM images of the untreated BiOCl samples are shown in Fig. S4a (ESI†). The untreated BiOCl sample is composed of 4–7 μm microspheres assembled by the nanosheets. After hydrothermal treatment at R_F = 0, a part of microspheres have broken into nanosheets (Fig. S4b, ESI†). After treatment with NH₄F, the microspheres have apparently broken into nanosheets and eventually evolved into the other shapes (Fig. S4c–h, ESI†). It is obvious that NH₄F has a significant impact on the phase composition and morphology of the sample. It is interesting that BiOCl can be converted to the other new materials merely by a simple ion exchange approach.

The surface areas of the samples are shown in Table 1. It is obvious that the sample treated at R_F = 0 has a smaller surface area (2.4 m² g⁻¹) than the untreated BiOCl sample (16.5 m² g⁻¹). Fig. S4a and b (ESI†) shows that the untreated sample is composed of uniform, dense microspheres assembled by numerous nanosheets, leading to the formation of more pores;³³ for the treated samples, the microspheres have been broken into the grains, the increased grains may result in the reduced surface area.

Table 1 BET areas of the untreated BiOCl and treated samples by NH₄F at different R_F at V_Eth/V_w = 18/2 at 170 °C for 6 h^a

Sample	BET area^b (m^2 g^−1)	Sample	BET area (m^2 g^−1)
a RF: the molar ratios of F to Bi; VEth/Vw: volumetric ratio of ethanol to water.b Calculated by the Brunauer–Emmett–Teller (BET) method.
Untreated BiOCl	16.5	RF = 2 (Bi7F11O5/BiOCl)	5.0
RF = 0 (BiOCl)	2.4	RF = 4 (Bi7F11O5)	3.9
RF = 0.5 (BiOF/BiOCl)	8.1	RF = 6 (Bi7F11O5/BiF3)	5.8
RF = 1 (BiOF/Bi7F11O5/BiOCl)	4.4	RF = 8 (BiF3)	7.2

---

Typically, the Bi₇F₁₁O₅/BiOCl sample obtained at R_F = 2 is characterized by the high-resolution transmission electron microscopy (HRTEM) and selected-area electron diffraction (SAED) patterns. It is clear that Bi₇F₁₁O₅ nanosheets distribute on BiOCl nanosheets (Fig. 2a). In Fig. 2b, the clear diffraction spots correspond to the (200) and (110) Bragg reflections of BiOCl, and the diffraction rings correspond to (112) and (004) Bragg reflections of Bi₇F₁₁O₅, respectively. The SAED patterns reveal the single crystalline nature of BiOCl and the polycrystalline nature of Bi₇F₁₁O₅. Moreover, Fig. 2c shows the lattice fringe images of the Bi₇F₁₁O₅/BiOCl sample. The lattice spacing of 0.278 nm is well indexed to (110) crystal plane of BiOCl, and the lattice spacings of 0.230 nm and 0.335 nm match (004) and (112) planes of Bi₇F₁₁O₅, respectively. The EDS analysis of Bi₇F₁₁O₅/BiOCl sample is shown in Fig. S3 and Table S1 (ESI†), which is also given as follows. The Bi, O, Cl and F are detected, which can also confirm the phase evolution and the formation of Bi₇F₁₁O₅/BiOCl heterojunction. The results demonstrate that a heterojunction has formed between BiOCl and Bi₇F₁₁O₅ due to their matched energy bands, which will be further discussed in the latter sections.

Fig. 2 High-resolution transmission electron microscopy (HRTEM) images and selected-area electron diffraction (SAED) patterns of Bi₇F₁₁O₅/BiOCl sample prepared at R_F = 2 at V_Eth/V_w = 18/2: (a) TEM; (b) SAED; (c) lattice fringe images.

3.1.2. Light absorptions and electrical conductivities of the samples. Fig. 3a displays the UV-visible diffuse reflectance spectra (UV-DRS) of the samples. All the samples exhibit the light absorption in UV light ranges. Compared with the untreated BiOCl sample, all the treated samples show the obvious blue-shifts, indicating the band gap increase, which is mainly due to their phase compositions. It can be determined that the absorption edges of BiOCl, Bi₇F₁₁O₅ and BiF₃ are 374, 317 and 300 nm, respectively.

Fig. 3 (a) UV-visible diffuse reflectance spectra (UV-DRS) and (b) electrochemical impedance spectra (EIS) of the untreated BiOCl and treated samples by NH₄F at different R_F: V_Eth/V_w = 18/2.

Besides, the electrochemical impedance spectroscopy (EIS) spectra of the samples are shown in Fig. 3b. It is well known that the smaller arc radius indicates a higher electrical conductivity that will favor for an efficient electron transfer.^34,35 The arc radius of the samples follow the order as follows: R_F = 0 < BiOCl < R_F = 0.5 < R_F = 1 < R_F = 2 < R_F = 4 < R_F = 6 < R_F = 8. It is obvious that with the increase of R_F, the conductivities of the treated samples decrease, which is relative to the phase composition. But, the treated BiOCl sample at R_F = 0 has a higher conductivity than the untreated one, which can be attributed to the improved crystallinity of the former one.

3.1.3. Photocatalytic activities of the samples. Fig. 4 shows the degradation curves and reaction kinetic curves of methyl orange (MO) dye under UV light irradiation (λ ≤ 420 nm). The degradation activities of the samples follow the order as follows: R_F = 2 > R_F = 1 > R_F = 0 > R_F = 0.5 > untreated BiOCl > R_F = 4 > R_F = 6 > R_F = 8. It is obvious that the Bi₇F₁₁O₅/BiOCl sample (R_F = 2) shows the highest activity amongst.

Fig. 4 Degradation curves (a) and reaction kinetic curves (b) of MO over the untreated BiOCl and treated samples by NH₄F at different R_F under ultraviolet light irradiation (λ ≤ 420 nm): V_Eth/V_w = 18/2.

First, after irradiation 30 min, 80% of MO has been degraded by Bi₇F₁₁O₅/BiOCl (R_F = 2), but only 17% and 21% of MO are degraded by Bi₇F₁₁O₅ (R_F = 4) and untreated BiOCl, respectively. The degradation rate constants (k_a) are given in Table 2. The degradation rate constant (k_a = 0.0642 min⁻¹) of Bi₇F₁₁O₅/BiOCl sample is 6.42 and 3.28 times higher than those of Bi₇F₁₁O₅ (k_a = 0.0100 min⁻¹) and the untreated BiOCl (k_a = 0.0196 min⁻¹), respectively. From Table 1 and Fig. 3, the untreated BiOCl has a higher BET area (16.5 m² g⁻¹), a stronger light absorption and a higher conductivity than Bi₇F₁₁O₅/BiOCl sample, but its activity is low. This suggests that the BET area, light absorbance and conductivity may not be the main affecting factors for the photocatalytic activity. It is well known that photocatalytic activity is strongly dependent on the separation and transfer of photogenerated charges.^36,37 We could assume that the high activity of Bi₇F₁₁O₅/BiOCl may be relative to the formation of a heterojunction between BiOCl and Bi₇F₁₁O₅. The conduct band (CB) and valence band (VB) are 0.49 and 3.81 eV for BiOCl, whereas 1.12 and 5.1 eV for Bi₇F₁₁O₅, respectively. Their energy bands are well matched, the efficient Bi₇F₁₁O₅/BiOCl heterojunctions can form, at which an interfacial electrical field can form. Under UV light irradiation, the photogenerated electrons at the CB of BiOCl would transfer to the CB of Bi₇F₁₁O₅; meanwhile the generated holes at the VB of Bi₇F₁₁O₅ would migrate to the VB of BiOCl. Driven by the interfacial electrical field, the photogenerated electron holes can be effectively separated.

Table 2 The apparent rate constants (k_a) over the untreated BiOCl and treated samples by NH₄F at different R_F for the degradation of MO^a

Sample	ka (min^−1)	Sample	ka (min^−1)
a VEth/Vw = 18/2; fluorine source: NH4Cl; 170 °C/6 h.
Untreated BiOCl	0.0196	RF = 2	0.0642
RF = 0	0.0415	RF = 4	0.0100
RF = 0.5	0.0234	RF = 6	0.0079
RF = 1	0.0532	RF = 8	0.0072

---

Second, for the untreated and treated BiOCl (R_F = 0) samples, the degradation rate (k_a = 0.0415 min⁻¹) of the treated BiOCl sample is obviously higher than that (k_a = 0.0196 min⁻¹) of the former sample, although the BET area (2.4 m² g⁻¹) of the latter sample is greatly smaller than that (16.5 m² g⁻¹) of the former sample (Table 1). This suggests the improved degradation efficiency cannot be ascribed to the BET area. Since the BiOCl sample treated at R_F = 0 has a higher conductivity (Fig. 3b) and a higher crystallinity (Fig. 1) than the untreated BiOCl, so the high degradation efficiency of the former sample may be attributed to its high charge transportation rate and high crystallinity.

Third, the degradation efficiencies of the phase-pure BiF₃, Bi₇F₁₁O₅ and untreated BiOCl samples follow the order as follows: BiF₃ < Bi₇F₁₁O₅ < untreated BiOCl; but their BET areas follow the order as follows: Bi₇F₁₁O₅ (3.9 m² g⁻¹) < BiF₃ (7.2 m² g⁻¹) < the untreated BiOCl (16.5 m² g⁻¹). Therefore, the BET area may not be the main affecting factor for the photocatalytic activity. Compared with BiOCl, the blue shifts of Bi₇F₁₁O₅ and BiF₃ at absorption thresholds are obvious (Fig. 3a). Thus, their different photocatalytic activities can be attributed to their different light absorptions.

In addition, the degradation curve of MO over P25 also have been tested shown in Fig. S5a (ESI†). The results show that about 90% of MO can be degraded by P25 after irradiated for 25 min, which was little higher than Bi₇F₁₁O₅/BiOCl sample. What is more, in order to avoiding dye photosensitization, we have also tested the photodegradation activity of 2-nitrophenol by the Bi₇F₁₁O₅/BiOCl sample. Fig. S5b (ESI†) shows the UV-vis absorption spectra of 2-nitrophenol (200 mL, 10 mg L⁻¹) over Bi₇F₁₁O₅/BiOCl sample at different reaction time under UV light irradiation (λ ≤ 420 nm). It is clear that the absorbance (at about 277 nm) are decreased gradually with the increasing of irradiation time. After been irradiated for 40 min, nearly 90% of 2-nitrophenol can be degraded by Bi₇F₁₁O₅/BiOCl sample.

3.2. Effect of fluorine source on the samples

To investigate the effect of fluorine source on the sample, the samples are prepared by using different fluorine sources (NH₄F, NaF and HF) at R_F = 2 and V_Eth/V_w = 18/2, while keeping the other condition same. Fig. 5a shows the XRD patterns of the samples. With NH₄F, the sample is mainly composed of Bi₇F₁₁O₅ and a small amount of BiOCl; with NaF, nevertheless, the crystal phase has not changed (i.e. BiOCl); with HF, the sample mainly consists of BiOCl and a small amount of Bi₇F₁₁O₅. Further, Fig. S6 (ESI†) shows the SEM images of the samples synthesized with different fluorine sources. When using NH₄F, the formed sample consists of nanosheets (Fig. S6a, ESI†). When NaF is added, however, the sample remains the original microspheres (Fig. S6b, ESI†). While HF is used, the sample is composed of loose microspheres and a few of nanosheets (Fig. S6c, ESI†). It is obvious that the fluorine source has a significant impact on the phase composition and morphology of the sample. The alkaline condition (NaF) has little impact on the sample, while the neutral (NH₄F) and acid (HF) conditions have a significantly influences on the conversion process. As shown in Table 3, the BET areas of the samples with different fluorine sources follow the order as follows: NaF (14.8 m² g⁻¹) > HF (6.8 m² g⁻¹) > NH₄F (5.0 m² g⁻¹). Compared with the untreated BiOCl sample (16.5 m² g⁻¹), the BET area of the sample treated by NaF almost does not change, which may be attributed to the almost unchanged microspheres morphology. However, the BET areas of the samples treated by NH₄F and HF decrease obviously, which can be mainly ascribed to the morphology change from porous microspheres to nanosheets.

Fig. 5 XRD patterns (a) and UV-DRS (b) of the samples treated at V_Eth/V_w = 18/2 using different fluorine sources (R_F = 2): NH₄F: Bi₇F₁₁O₅/BiOCl; HF: Bi₇F₁₁O₅/BiOCl; NaF: BiOCl.

Table 3 BET areas and the apparent rate constants (k_a) for the degradation of MO of the treated samples with different fluorine sources^a

Fluorine source	NH4F	NaF	HF
a RF = 2; VEth/Vw = 18/2; 170 °C/6 h.
Crystal phase	Bi7F11O5/BiOCl	BiOCl	Bi7F11O5/BiOCl
BET area (m^2 g^−1)	5.0	14.8	6.8
ka (min^−1)	0.0642	0.0264	0.0672

---

Fig. 5b displays UV-DRS of the samples treated with different fluorine sources. The three samples exhibit almost the same absorption. Compared with the NaF-treated sample (BiOCl), the samples with NH₄F and HF show a slight blue shift, which is attributed to the formation of Bi₇F₁₁O₅. In the sample with HF, a small amount of Bi₇F₁₁O₅ is contained; but a large amount of Bi₇F₁₁O₅ is contained in the sample with NH₄F. As determined in Section 3.1.2, Bi₇F₁₁O₅ has a wider band gap than BiOCl.

Fig. 6 shows the degradation curves and reaction kinetic curves of the samples, and the calculated rate constants (k_a) are summarized in Table 3. Their degradation rates follow the order as follows: HF (k_a = 0.0672 min⁻¹) ≈ NH₄F (k_a = 0.0642 min⁻¹) > NaF (k_a = 0.0264 min⁻¹). The photocatalytic activities of the samples treated with NH₄F (Bi₇F₁₁O₅/BiOCl) and HF (Bi₇F₁₁O₅/BiOCl) are higher than that of NaF (BiOCl), although the sample treated by NaF has the larger BET area (Table 3). Herein, we hold that the improved activities of the former samples are mainly due to the formation of heterojunction between BiOCl and Bi₇F₁₁O₅, which has efficiently improve the charge separation.

Fig. 6 Degradation curves (a) and reaction kinetic curves (b) of the samples at V_Eth/V_w = 18/2 with different fluorine sources (R_F = 2) under ultraviolet light irradiation (λ ≤ 420 nm).

3.3. Effect of reaction medium composition on the samples

In order to investigate the effect of reaction medium composition, ethanol/water volumetric ratio (V_Eth/V_w) is varied from 0/20 to 2/18, 10/10, 18/2 and 20/0 while keeping the other preparations parameters constant (NH₄F, R_F = 2). The XRD patterns of the samples are shown in Fig. 7a. At V_Eth/V_w = 0/20, 2/18 and 10/10, all products are phase-pure BiOF. However, Bi₇F₁₁O₅/BiOCl heterojunctions have formed at V_Eth/V_w = 18/2, and BiOCl has converted into BiF₃/BiOCl at V_Eth/V_w = 20/0. In brief, at lower V_Eth/V_w ratios, BiOCl can completely convert to BiOF; on the contrary, Bi₇F₁₁O₅/BiOCl and BiF₃/BiOCl can form at higher V_Eth/V_w ratios (18/2 and 20/0). Fig. S7 (ESI†) shows their SEM images. At V_Eth/V_w = 0/20, 2/18 and 10/10, the part of microspheres have broken and the microspheres become loose (Fig. S7a–c, ESI†). At V_Eth/V_w = 18/2, all the microspheres have broken into nanosheets (Fig. S7d, ESI†). At V_Eth/V_w = 20/0, however, some of irregular particles form, besides microspheres (Fig. S7e, ESI†). It is clearly observed that the V_Eth/V_w ratio has an important influence on the phase composition and morphology of the sample.

Fig. 7 XRD patterns (a) and UV-DRS (b) of the samples prepared with NH₄F (R_F = 2) at different V_Eth/V_w ratios: BiOF (0/20, 2/18, 10/10); Bi₇F₁₁O₅/BiOCl (18/2); BiF₃/BiOCl (20/0).

The BET areas of the samples are shown in Table 4. At V_Eth/V_w = 0/20–10/10, the BET areas of the phase-pure BiOF samples decrease gradually, which may be attributed to the broken microspheres that reduces the porosity. The Bi₇F₁₁O₅/BiOCl sample obtained at V_Eth/V_w = 18 : 2 has a smaller BET area (5.0 m² g⁻¹) than that (8.2 m² g⁻¹) of the BiF₃/BiOCl sample at V_Eth/V_w = 20/0, which is due to their different morphology. According to the SEM results, the former sample is composed of nanosheets resultant from completely broken porous microspheres, but the latter sample consists of the porous microsphere and a small number of irregular particles (Fig. S7e, ESI†). Fig. 7b shows the UV-DRS of the samples. With increasing V_Eth/V_w ratio from 0/20 to 2/18 and 10/10, the light absorption intensities of the BiOF samples decrease gradually, which can be attributed to their decreased BET areas (Table 4). Compared with BiOF, however, the Bi₇F₁₁O₅/BiOCl (V_Eth/V_w = 18/2) and BiF₃/BiOCl (V_Eth/V_w = 20/0) samples show a slight red shift at absorption edges, which can be attributed to the different phase compositions from BiOF. As shown in Table 5, the band gaps are 3.32, 3.98, 4.01 and 4.15 eV for BiOCl, Bi₇F₁₁O₅, BiOF and BiF₃, respectively.

Table 4 BET areas and the apparent rate constants (k_a) for the degradation of MO of the treated samples at different V_Eth/V_w ratios^a

VEth/Vw	0/20	2/18	10/10	18/2	20/0
a RF = 2; fluorine source: NH4F; 170 °C/6 h.
Crystal phase	BiOF	BiOF	BiOF	Bi7F11O5/BiOCl	BiF3/BiOCl
BET area (m^2 g^−1)	4.0	3.8	2.2	5.0	8.2
ka (min^−1)	0.0075	0.0064	0.0066	0.0642	0.0789

---

Table 5 The band gap (E_g), VBM (E_VB) and CBM (E_CB) of the samples^a

Chemicals	χ (eV)	Eg (eV)	EVB (eV)	ECB (eV)
a χ, the geometric mean of Mulliken's electronegativities; VBM, valence band maximum; CBM, conduction band minimum.
BiOCl	6.65	3.32	3.81	0.49
BiOF	7.17	4.01	4.68	0.67
Bi7F11O5	7.61	3.98	5.1	1.12
BiF3	8.53	4.15	6.11	1.96

---

Fig. 8 shows their degradation curves and reaction kinetic curves for degradation of MO under UV irradiation (λ ≤ 420 nm), and the calculated rate constants (k_a) are given in Table 4. Their degradation rates follow the order as follows: V_Eth/V_w = 20/0 (BiF₃/BiOCl, k_a = 0.0789 min⁻¹) > V_Eth/V_w = 18/2 (Bi₇F₁₁O₅/BiOCl, k_a = 0.0642 min⁻¹) ≫ V_Eth/V_w = 0/20 (BiOF, k_a = 0.0075 min⁻¹) > V_Eth/V_w = 10/10 (BiOF, k_a = 0.0066 min⁻¹) ≅ V_Eth/V_w = 2/18 (BiOF, k_a = 0.0064 min⁻¹). Among them, the BiF₃/BiOCl sample prepared at V_Eth/V_w = 20/0 shows the highest photocatalytic activity. Fig. 7b and Table 4 demonstrate that the orders of light absorption and BET area are consistent with the photocatalytic activity order. Moreover, the BiF₃/BiOCl and Bi₇F₁₁O₅/BiOCl heterojunctions also contribute to their obviously improved activities, compared with BiOF.

Fig. 8 Degradation curves (a) and reaction kinetic curves (b) of MO over the samples prepared with NH₄F (R_F = 2) at different V_Eth/V_w ratios under ultraviolet light irradiation (λ ≤ 420 nm).

3.4. Theory calculations

Further, we have mainly compared UV-DRS spectra of the phase-pure BiOCl, BiOF, Bi₇F₁₁O₅ and BiF₃ samples (Fig. 9). For BiOCl, BiOF, Bi₇F₁₁O₅ and BiF₃, with the gradual increase of fluorine atoms, the samples show an obvious blue-shift in the absorption thresholds (Fig. 9a). The band gap energy (E_g) of a semiconductor can be calculated by eqn (2) as follows.³⁸

αhv = A(hv − E_g)^n/2 (2)

where α, h, v, E_g and A represent absorption coefficient, Plank constant, light frequency, band gap energy and a constant, respectively. Among them, n is determined by the optical transition type of a semiconductor (n = 1 for a direct optical transition and n = 4 for an indirect optical transition).³⁹ The values of n are 4, 1, 1 and 1 for BiOCl, BiOF, Bi₇F₁₁O₅ and BiF₃, respectively.^19,20,40 Fig. 9b further shows their Tauc plots. The bandgap energies are determined to be 3.32, 4.01, 3.98 and 4.15 eV for BiOCl, BiOF, Bi₇F₁₁O₅ and BiF₃, respectively (Fig. 9b, Table 5).

Fig. 9 UV-DRS spectra (a) Tauc plots, (b) energy bands and (c) energy level diagram of the samples.

Further, the conduct band and valence band positions of a semiconductor can be calculated by the following empirical formula:⁴¹

E_VB = χ − E^e + 0.5E_g (3)

E_CB = E_VB − E_g (4)

where χ is the absolute electronegativity of a semiconductor, which is defined as the geometric mean of the absolute electronegativity of constituent atoms; E^e is the energy of free electrons on the hydrogen scale (ca. 4.5 eV); E_CB and E_VB are the conduct and valence band edge potentials, and E_g is band gap. Herein, the calculated E_CB and E_VB values of the samples are summarized in Table 5. It is obvious that with increase of F, their CB and VB down shift and their band gap become wider (Fig. 9c).

In our work, the energy band and electronic structures of BiOCl, BiOF and BiF₃ are calculated by Density Functional Theory (DFT) with the generalized-gradient approximation (GGA) and the exchange–correlation functional of Perdew–Burke–Ernzerhof of (PBE) by using the Materials Studios software. The adopted CASTEP energy cutoff is 340 eV (Fig. 10). The calculated band gaps of BiOCl, BiOF and BiF₃ are 2.61, 3.34 and 3.70 eV, respectively. The band gap of Bi₇F₁₁O₅ has been calculated to be 3.81 eV by Hu et al.²⁰ It can be observed that the calculated results by DFT are slightly smaller than the experimental results, which is a common feature of DFT calculations.⁴² The DFT calculations also demonstrate that with the increase of fluorine atoms, the band gaps of the BiOCl, BiOF, Bi₇F₁₁O₅ and BiF₃ samples increases. From Fig. 10d–f, we can see that the VB top of BiOCl is mainly composed of the O 2p and Cl 3p orbits, and its CB bottom is mainly composed of Bi 6p and a part of O 2p orbits (Fig. 10d). The VB top of BiOF is found to be mainly composed of both O 2p and F 2p orbits, while the CB bottom is dominantly mainly composed of Bi 6p and a little of O 2p orbits (Fig. 10e). The results show that O 2p orbit contributes to both CB and VB of both BiOCl and BiOF. Hu et al. have reported that F 2p and O 2p orbits mainly contribute to the VB of Bi₇F₁₁O₅, and Bi 6p orbit mainly contributes to the CB.²⁰ Meanwhile, the VB top of BiF₃ is mainly attributed by the F 2p orbital, and the CB bottom is mainly attributed by the Bi 6p, as well as a little of F 2p orbits (Fig. 10f). Comparing BiF₃ with BiOF, it is interesting that F 2p orbit contributes to both CB and VB of BiF₃, which is different from BiOF. This further confirms that with the increase of fluorine, the band gaps of BiOCl, BiOF, Bi₇F₁₁O₅ and BiF₃ increase, similar to the experimental result above.

Fig. 10 Energy band structures (a–c) and total and a part of densities of states (TDOS and PDOS) (d–f) of the samples: (a and d) BiOCl; (b and e) BiOF and (c and f) BiF₃.

3.5. Conversion mechanism

It is well known that BiOCl, BiOF have tetragonal structures and Bi₇F₁₁O₅, BiF₃ have monoclinic, cubic structures; respectively. Due to the layer structure of BiOCl, Cl, instead of O, is easily substituted by F. Thus, BiOCl can be converted to BiOF, Bi₇F₁₁O₅ and BiF₃. In the presence of NH₄F at V_Eth/V_w = 18/2, BiOCl can be facilely converted to BiOF, Bi₇F₁₁O₅ and BiF₃. At V_Eth/V_w = 18/2, however, we cannot achieve the complete conversion from BiOCl to phase-pure BiOF while only changing the amount of NH₄F added. Nevertheless, BiOCl can completely convert to phase-pure BiOF at low V_Eth/V_w ratios (0/20, 2/18, 10/10) at R_F = 2. According to the experiment results, the schematic conversion process is proposed and shown in Fig. 11, and two main conclusions are made as follows.

Fig. 11 Schematic conversion process.

First, generally, a higher dielectric constant results in a higher solubility of metal oxide,⁴³ which favors for the nucleation and growth processes. With increasing V_Eth/V_w ratio, the dielectric constant of reaction medium decreases (ε_Eth = 24.5, ε_w = 78.5). Thus we assume that the medium composition may mainly affect the crystal nucleation and growth in crystal conversion.

Second, NH₄⁺ as a Lewis acid can destroyed the [Bi₂O₂] layers obviously. At R_F = 4 (NH₄F), the [Bi₂O₂] slabs may have been destroyed partially, where the partial O of [Bi₂O₂] slabs will be replaced by F. Thus, tetragonal BiOCl has completely converted to monoclinic Bi₇F₁₁O₅. To further increase the amount of NH₄F to R_F = 8, the original [Bi₂O₂] slabs will be destroyed completely, where the O atoms of [Bi₂O₂] slabs may be eventually replaced by F. As a result, tetragonal BiOCl has completely converted into cubic BiF₃. When HF is used as the fluorine source, on the other hand, BiOCl convert to hybrid phase Bi₇F₁₁O₅/BiOCl, similar to the result by using NH₄F as the fluorine source. However, BiOCl cannot be changed while using NaF as the fluorine source, suggesting that the alkaline condition has no effect on [Bi₂O₂] slabs of BiOCl. Herein, we could assume that the crystal conversion is mainly controlled through altering the [Bi₂O₂] slabs. However, the real conversion mechanism needs further studying.

4. Conclusions

A facile conversion from BiOCl nanosheets to BiOF, Bi₇F₁₁O₅ and BiF₃ can be achieved through a simple ion exchange method. The added amounts of NH₄F, fluorine source and medium have significant influences on the composition and morphology of the samples. The photocatalytic activity of Bi₇F₁₁O₅/BiOCl heterojunction (R_F = 2) is 3.28 times higher than that of BiOCl for the degradation of methyl orange (MO). The as-formed Bi₇F₁₁O₅/BiOCl heterojunction benefits the separation and transfer of charges, leading to an improved photocatalytic activity. This work provides an initiative post-synthesis strategy to prepare the other photocatalysts.

Acknowledgements

This work is financially supported by National Science Foundation of China (21377060), Scientific Research Foundation for the Returned Overseas Chinese Scholars of State Education Ministry (20121707), Six Talent Climax Foundation of Jiangsu (20100292), the Key Project of Environmental Protection Program of Jiangsu (2013005), “333” Outstanding Youth Scientist Foundation of Jiangsu (20112015), the Project Funded by the Science and Technology Infrastructure Program of Jiangsu (BM201380277), and A Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD) sponsored by SRF for ROCS, SEM (2013S002).

References

1. K. L. Zhang, C. M. Liu, F. Q. Huang, C. Zheng and W. D. Wang, Appl. Catal., B, 2006, 68, 125–129.
2. H. Gnayem and Y. Sasson, ACS Catal., 2013, 3, 186–191.
3. J. Cao, B. Y. Xu, H. L. Lin, B. D. Luo and S. F. Chen, Catal. Commun., 2012, 26, 204–208.
4. M. L. Guan, C. Xiao, J. Zhang, S. J. Fan, R. An, Q. M. Cheng, J. F. Xie, M. Zhou, B. J. Ye and Y. Xie, J. Am. Chem. Soc., 2013, 135, 10411–10417.
5. K. Zhao, L. Z. Zhang, J. J. Wang, Q. X. Li, W. W. He and J. J. Yin, J. Am. Chem. Soc., 2013, 135, 15750–15753.
6. C. J. Huang, J. L. Hu, S. Cong, Z. G. Zhao and X. Q. Qiu, Appl. Catal., B, 2015, 174, 105–112.
7. C. L. Yu, F. F. Cao, G. Li, R. F. Wei, J. C. Yu, R. C. Jin, Q. Z. Fan and C. Y. Wang, Sep. Purif. Technol., 2013, 120, 110–122.
8. H. F. Cheng, B. B. Huang, X. Y. Qin, X. Y. Zhang and Y. Dai, Chem. Commun., 2012, 48, 97–99.
9. J. Xu, F. Teng, Y. X. Zhao, Y. D. Kan, L. M. Yang, Y. Yang, W. Q. Yao and Y. F. Zhu, Dalton Trans., 2016, 45, 6866–6877.
10. H. Tian, F. Teng, J. Xu, S. Q. Lou, N. Li, Y. X. Zhao and M. D. Chen, Sci. Rep., 2015, 5, 7770–7778.
11. J. Xu, F. Teng, W. Q. Yao and Y. F. Zhu, RSC Adv., 2016, 6, 23537–23549.
12. S. R. Wang, F. Teng and Y. X. Zhao, RSC Adv., 2015, 5, 76588–76598.
13. N. Li, X. Hua, K. Wang, Y. J. Jin, J. Xu, M. D. Chen and F. Teng, J. Mol. Catal. A: Chem., 2014, 395, 428–433.
14. Q. Y. Zhang, H. Tian, N. Li, M. D. Chen and F. Teng, CrystEngComm, 2014, 16, 8334–8339.
15. P. Sun, Y. J. Jin, Y. X. Zhao, J. Xu, M. D. Chen, W. Q. Yao, Y. F. Zhu and F. Teng, Nano, 2014, 9, 1450094–1450104.
16. H. B. Kang, S. T. Myung, K. Amine, S. M. Lee and Y. K. Sun, J. Power Sources, 2010, 195, 2023–2028.
17. W. Y. Su, J. Wang, Y. X. Huang, W. J. Wang, L. Wu, X. X. Wang and P. Liu, Scr. Mater., 2010, 62, 345–348.
18. S. J. Zou, F. Teng, C. Chang, Z. L. Liu and S. R. Wang, RSC Adv., 2015, 5, 88936–88942.
19. C. K. Feng, F. Teng, Z. L. Liu, C. Chang, Y. X. Zhao, S. R. Wang, M. D. Chen, W. Q. Yao and Y. F. Zhu, J. Mol. Catal. A: Chem., 2015, 401, 35–40.
20. W. Hu, P. Shan, T. Q. Sun, H. D. Liu, J. M. Zhang, X. W. Liu, Y. F. Kong and J. J. Xu, J. Alloys Compd., 2016, 658, 788–794.
21. Q. C. Liu, D. K. Ma, Y. Y. Hu, Y. W. Zeng and S. M. Huang, ACS Appl. Mater. Interfaces, 2013, 5, 11927–11934.
22. D. K. Ma, M. L. Guan, S. S. Liu, Y. Q. Zhang, C. W. Zhang, Y. X. He and S. M. Huang, Dalton Trans., 2012, 41, 5581–5586.
23. P. Cai, S. M. Zhou, D. K. Ma, S. N. Liu, W. Chen and S. M. Huang, Nano-Micro Lett., 2015, 7, 183–193.
24. J. C. Yu, J. G. Yu, W. K. Ho, Z. T. Jiang and L. Z. Zhang, Chem. Mater., 2002, 14, 3808–3816.
25. S. W. Liu, K. Yin, W. S. Ren, B. Cheng and J. G. Yu, J. Mater. Chem., 2012, 22, 17759–17767.
26. H. M. Wu, J. Z. Ma, Y. B. Li, C. B. Zhang and H. He, Appl. Catal., B, 2014, 152, 82–87.
27. S. H. Chen, S. X. Sun, H. G. Sun, W. L. Fan, X. Zhao and X. Sun, J. Phys. Chem. C, 2010, 114, 7680–7688.
28. S. W. Liu, J. G. Yu, B. Cheng and M. Jaroniec, Adv. Colloid Interface Sci., 2012, 173, 35–53.
29. J. G. Yu, W. G. Wang, B. Cheng and B. L. Su, J. Phys. Chem. C, 2009, 113, 6743–6750.
30. H. Y. Jiang, J. J. Liu, K. Cheng, W. B. Sun and J. Lin, J. Phys. Chem. C, 2013, 117, 20029–20036.
31. J. S. Wang, S. Yin, Q. W. Zhang, F. Saito and T. Sato, J. Mater. Sci., 2004, 39, 715–717.
32. Y. F. Liu, Y. H. Lv, Y. Y. Zhu, D. Liu, R. L. Zong and Y. F. Zhu, Appl. Catal., B, 2014, 147, 851–857.
33. N. Li, X. Hua, K. Wang, Y. J. Jin, J. J. Xu, M. D. Chen and F. Teng, Dalton Trans., 2014, 43, 13742–13750.
34. Y. Y. Zhu, Q. Ling, Y. F. Liu, H. Wang and Y. F. Zhu, Phys. Chem. Chem. Phys., 2015, 17, 933–940.
35. Y. D. Hou, A. B. Laursen, J. S. Zhang, G. G. Zhang, Y. S. Zhu, X. C. Wang, S. Dahi and I. Chorkendorff, Angew. Chem., Int. Ed., 2013, 52, 3621–3625.
36. C. S. Pan, J. Xu, Y. J. Wang, D. Li and Y. F. Zhu, Adv. Funct. Mater., 2012, 22, 1518–1524.
37. M. Guerrero, A. Altube, E. Garcia-Lecina, E. Rossinyol, M. D. Baro, E. Pellicer and J. Sort, ACS Appl. Mater. Interfaces, 2014, 6, 13994–14000.
38. J. Zeng, H. Wang, Y. C. Zhang, M. K. Zhu and H. Yan, J. Phys. Chem. C, 2007, 111, 11879–11887.
39. S. M. Sun, W. Z. Wang, L. Zhang, L. Zhou, W. Z. Yin and M. Shang, Environ. Sci. Technol., 2009, 43, 2005–2010.
40. Y. Z. Zhao and W. W. Dai, Inorg. Chem., 2014, 53, 13001–13011.
41. N. Zhang, S. Q. Liu, X. Z. Fu and Y. J. Xu, J. Phys. Chem. C, 2011, 115, 9136–9145.
42. A. Kudo, H. Kato and S. Nakagawa, J. Phys. Chem. B, 2000, 104, 571–575.
43. N. Garti and H. Zour, J. Cryst. Growth, 1997, 172, 486–498.

---

Footnote

† Electronic supplementary information (ESI) available: This material, SEM images of the samples with NH₄F at different R_F and fluorine sources and V_Eth/V_w. See DOI: 10.1039/c6ra11877a

---