Structure and composition analysis
The crystal information of α-Bi2O3, exfoliated bentonite and Bi2SiO5/Bi2O3/EB photocatalysts was described by using X-ray diffraction (XRD). The results were displayed in Fig. 1. As shown in Fig. 1a, the characteristic diffraction peak for EB sample at 2θ = 8.20° corresponds to (001) crystal plane (Ma et al. 2016a). For the α-Bi2O3, the three distinct peaks with 2θ of 27.37°, 33.31° and 46.42° are referred to the (-121), (-202) and (041) planes of α-Bi2O3 (JSPDS: 71–0465), respectively. For BiSi-3 sample, the two diffraction peaks 2θ = 27.98° and 32.81° are correlated with the (201) and (220) planes of β-Bi2O3 (JCPDS:77-5341), separately, which indicates that Bi2O3 mainly exists in the form of β-Bi2O3 phase (Lu et al. 2018b). Meanwhile, the other three intense diffraction peaks with 2θ of 11.26°, 23.92° and 29.31° are corresponded to the (200), (310) and (311) planes of Bi2SiO5 (JCPDS: 36–0287), separately. As shown in Fig. 1d and S1, when the calcination temperature was 500 ℃, the diffraction peaks of α-Bi2O3 and β-Bi2O3 were simultaneously present, while the diffraction peak of α-Bi2O3 disappears, but β-Bi2O3 diffraction peak still exists when the temperature is raised to 600 ℃ and 700 ℃, which means that Bi2O3 mainly exists in the crystal form of β- Bi2O3 phase at the calcination temperature of 600 ℃ and 700 ℃.
The morphology and structure of the as-prepared photocatalyst as determined by scanning electron microscopy (SEM). As shown in Fig. 2a, the EB has a sheet structure with a thickness of about 1 µm. SEM image of α-Bi2O3 (Fig. 2b) demonstrates that pure α-Bi2O3 shows an agglomerated bulk structure. However, after the formation of heterojunction, BiSiB-3 andBiSiB-7 samples exhibit a fluffier scattered structure compared with pure α-Bi2O3, which is beneficial for the photocatalytic activity (Fig. 2c, 2d). Due to the fact that Bi2O3/Bi2SiO5 heterojunction was in-situ generated on EB, the sheet structure could also be observed (Fig. 2c). Fig. 3 shows the detailed microstructure of the typical heterojunction by TEM and HRTEM. It is obviously observed that the fluffy scattered structure of Bi2O3/Bi2SiO5 heterojunction is formed. As shown from the lattice spacing of BiSiB-3 in Fig. 3d, a distinct lattice fringes of 0.227 nm corresponds to the (202) plane of Bi2O3, andthe interplanar spacing of 0.381 nm can be indexed to the (400) plane of Bi2SiO5. Fig. 3f-i and Fig. S2 showed the element distribution of BiSiB-3 sample. The results indicate that the O, Si and Bi elements in the samples are evenly distributed, and their atomic ratios are 59.56%, 9.43% and 28.81%, respectively, which further expounds the as-prepared sample is homogeneous.
The detailed surface chemical composition and chemical state of EB, α-Bi2O3 and BiSiB-3 samples were studied by X-ray photoelectron spectroscopy. Figure 4a showed the survey spectrum of the EB, α-Bi2O3 and BiSiB-3 samples. The peaks of O, Al, and Si in EB, as well as the peak of Bi and O in α-Bi2O3 were detected. The peaks of O, Al, Si and Bi were observed from the surface of BiSiB-3, while the element of Si was undetected in α-Bi2O3. In Fig. 4b, the two intensive peaks of 159.1 eV and 164.4 eV were correlated with the orbitals 4f7/2 and 4f5/2 of Bi3+ in α-Bi2O3, separately (Wang et al. 2017). In addition, compared with α-Bi2O3, the binding energy of these two peaks in BiSiB-3 decreased by 0.2 eV, which might be caused by the tight binding between Bi2SiO5 and β-Bi2O3, resulting in the change of chemical state around Bi (Xie et al. 2018). As shown in Fig. 4c, the binding energy of Si 2p was 102.7 eV, which indicates that the valence state of Si in the sample is +4. Two distinct peaks of O 1s in α-Bi2O3 were located at 528.8 eV and 530.9 eV, which corresponded to O2- in α-Bi2O3 (Fig. 4d)(Zhang et al. 2017). For the BiSiB-3 sample, the peak at 530.1 eV in the O 1s energy spectrum was corresponded to Bi-O in Bi2SiO5 and Bi2O3(Di et al. 2016, Hu et al. 2015). Another stronger peak at 532.2 eV was attribute to the adsorbed surface oxygen, which increased by 1.3 eV compared with α-Bi2O3(Song et al. 2018, Yan et al. 2016). BiSiB-3 has lager specific surface, thus the surface of Bi2SiO5 can adsorb a large amount of O2.
Figure 5 exhibits the results of N2 adsorption-desorption analysis of α-Bi2O3 and BiSiB-3 samples. As can be seen from Fig. 5, the specific surface area of α-Bi2O3 and BiSiB-3 is 1.74 m2g− 1 and 14.72 m2g− 1, respectively. Obviously, the specific surface area of BiSiB-3 was larger than of α-Bi2O3, which was 8.5 times higher than that of α-Bi2O3. The increase of specific area, attributing to the in-situ formation of Bi2O3/Bi2SiO5 heterojunction on the bentonite surface, not only improves the adsorption capacity of the catalyst for pollutants, but also allows more O2 to be adsorbed on its surface, which can produce more O2− with intense oxidizing properties under light irradiation, thus improving the photocatalytic degradation reaction.
3.2. Optical and photoelectrical performance
To investigate the optical properties and energy band structure of the photocatalysts, UV-Vis spectroscopy and the Kubelka-Munk equation were used (Wang et al. 2013). As shown in Fig. 6a and 6b, the absorption edge of α-Bi2O3 was 454 nm, and its corresponding bandgap was 2.74 eV. Other samples with EB exhibited two absorption edges at 390 nm and 575 nm, respectively, where the absorption edge at 575 nm was corresponded to the bandgap of 2.15 eV for β-Bi2O3, and the absorption edge at 390 nm was assigned to the bandgap of 2.96 eV for β-Bi2O3/Bi2SiO5 heterojunction. The absorption edge of bare Bi2SiO5 was 360 nm and the associated bandgap was 3.58 eV. The absorption band edges of β-Bi2O3/Bi2SiO5 heterojunction were all situated between Bi2SiO5 and β-Bi2O3, which indicated that the β-Bi2O3/Bi2SiO5 heterojunction could broaden the spectral responsive range compared to bare Bi2SiO5.
Electrochemical impedance spectroscopy (EIS) and photocurrent response were used to characterize the separation and transfer efficiency of photogenerated electro-hole pairs. As we can see from Fig. 7a, Bi2SiO5/Bi2O3/EB sample exhibits a higher photocurrent than bare α-Bi2O3, which means that Bi2SiO5/Bi2O3/EB has a batter photocarrier separation efficiency. As can be seen from Fig. 7b, the arc of Bi2SiO5/Bi2O3/EB is smaller than that of α-Bi2O3, which indicates that Bi2SiO5/Bi2O3/EB has a smaller resistance than α-Bi2O3. These results indicate that the β-Bi2O3/Bi2SiO5 heterojunction formed by the combination of Bi2O3 and EB can increase the photocarrier transfer and separation speed, reduce the resistance, thus enhance its photocatalytic performance.
3.3 Photocatalytic performance of Bi2SiO5/Bi2O3/EB composites
The photocatalytic activities of exfoliated bentonite, α-Bi2O3 and Bi2SiO5/Bi2O3/EB were investigated by degrading RhB (20 mgL− 1) under simulated solar light irradiation. Fig. S3 shows the RhB degradation dynamic curve of BiSiB-3 at different temperatures. It can be seen that the sample with the calcination temperature of 600 ℃ has the best photocatalytic performance. Figure 8a shows the RhB degradation dynamic curves over different photocatalysts. Pure α-Bi2O3 exhibited extremely poor adsorption performance, but the Bi2SiO5/Bi2O3/EB composites exhibited better adsorption performance after being compounded with exfoliated bentonite. After stirring for 60 min in the dark, the adsorption rate of SiBiB-1 and SiBiB-3 reached 78.94% and 17.61%, separately. Under simulated solar light irradiation, the pure α-Bi2O3 exhibits almost no photocatalytic activity over RhB. However, when it is combined with exfoliated bentonite, it forms a Bi2O3/Bi2SiO5 heterojunction and exhibits strong degradation performance for RhB.
Among the as-prepared photocatalysts, BiSiB-3 sample shows the optimum photocatalytic activity with a removal efficiency of 96.89%. And the decomposition efficiency for the RhB is 0.0148 min-1, which is 49 folds than that of α-Bi2O3 (Fig. 8d). Furthermore, as shown in Fig. 6c, the decolourization proportion is about 94.92%, 68.35% and 61.61% for the BiSiB-1, BiSiB-5 and BiSiB-7 samples after the irradiation of 180 min, respectively. The results indicated that the combination Bi2O3/Bi2SiO5 heterojunction not only increases the specific surface area to enhance the adsorption capacity, but also improves the separation efficiency of the photogenerated carriers, thereby improving the photocatalytic property.
To further investigate the photocatalytic performance, the reaction kinetics of RhB removal were studied. The kinetic rate constant value were calculated by the pseudo-first-order model (-ln(Ct/C0) = kt) (Xu et al. 2011). C0 and Ct are the original and instantaneous concentration of RhB, respectively, k is a kinetic constant. As shown in Fig. 8b and 8d, the kinetic constants of BiSiB-1 (k = 0.0083 min− 1), BiSiB-3 (k = 0.0148 min− 1), BiSiB-5 (k = 0.006 min− 1) and BiSiB-7 (k = 0.0043 min− 1) for RhB are 27.6 times, 49.3 times, 20 times and 14.3 times higher than bare α-Bi2O3 (k = 0.0003 min− 1), respectively. It could be considered that the Bi2O3/Bi2SiO5 heterojunction could obviously improve the photocatalytic property compared with α-Bi2O3. Moreover, the decomposition kinetic constant of BiSiB-3 to RhB is 1.8, 2.5 and 3.4 times than that of BiSiB-1, BiSiB-5 and BiSiB-7 samples, which indicated that BiSiB-3 sample demonstrated the maximum photocatalytic performance toward the RhB degradation.
To investigate the stability and reusability of Bi2SiO5/Bi2O3/EB sample, the recycling experiments were performed. In this procedure, all the parameters remained the same. As represented in Fig. 9a, the photocatalytic efficiency of BiSiB-3 still achieved 91.0% after four cycles. Moreover, Fig. S4 showed the SEM image of BiSiB-3 after the removal of RhB. The morphology and the sample size did not change obviously. Additionally, after the four times photocatalytic tests, the XRD patterns of BiSiB-3 (Fig. 9d) also had no change. All the above-mentioned results indicated that the as-prepared sample had a good stability.
The capture experiments of the active species were used to prove the photocatalytic degradation of the main active species in RhB. Tert-butanol, ascorbic acid and disodium ethylenediaminetetraacetate (EDTA-2Na) were used as the sacrificial agents of hydroxyl radical (·OH), superoxide radical (·O2−) and hole (h+) in the photodegradation process of RhB, respectively. As shown in Fig. 9b and 9c, when tert-butanol, ascorbic acid and EDTA-2Na were added to the photocatalytic degradation system, the RhB degradation rate dropped from 79.29–73.30%, 37.25% and 5.05%, respectively. The results demonstrated that h+ and ·O2− were the principal active species for the photocatalytic degradation of RhB system.
In brief, the improvement of photocatalytic performance is primarily owed to the following four aspects. First, Bi2SiO5/Bi2O3/EB has a higher specific surface area than bare α-Bi2O3 and has a stronger adsorption for RhB. Thus, the active species produced by light irradiation are more likely to oxidize the contaminants adsorbed on the catalyst surface. Second, the exfoliated bentonite was added to the reactants, and as a result, not only Bi2SiO5 but also β-Bi2O3 was in-situ formed at the same time. Due to the lager light response range of β-Bi2O3, the utilization of light was improved. Third, Bi2SiO5 is a canonical n-type semiconductor and β-Bi2O3 is a p-type semiconductor. Therefore, Bi2SiO5 and β-Bi2O3 can construct a p-n heterogeneous junction, which can effectively separate the electrons and holes, and inhibit the recombination of photon-generated carriers, thereby increasing the photocatalytic efficiency. Furthermore, the negative charge electrostatic interaction on the surface of EB can accelerate the separation of electrons and holes and inhibit the recombination of electron-hole pairs.
According to the above discussion, a possible mechanism of the RhB photodegradation by Bi2SiO5/Bi2O3/EB photocatalysts was proposed. As shown in Fig S5, according to the equation: ENHE = EAg/AgCl + EθAg/AgCl (0.1976V) and EVB = ECB + Eg, the valence band positions of α-Bi2O3, β-Bi2O3 and Bi2SiO5 are estimated to be 3.07 eV, 2.62 eV and 3.52 eV, respectively. The conduction band positions of α-Bi2O3, β-Bi2O3 and Bi2SiO5 are estimated to be 0.33 eV, 0.27 eV and − 0.06 eV, respectively (He et al. 2015, Hu et al. 2015, Lu et al. 2018b, Wu et al. 2013). The band edges of β-Bi2O3 and Bi2SiO5 exhibit a typical embedded structure, which may be unfavorable for the transport of photogenerated electrons and holes. However, when the n-type Bi2SiO5 and p-type β-Bi2O3 was mutually contacted, a p-n heterogenous junction can be constituted on the interface of Bi2SiO5 and β-Bi2O3 (Han et al. 2015, Sun et al. 2017). In addition, when Bi2SiO5 and β-Bi2O3 are in contact with each other, the photogenerated carriers on the interface of the two semiconductors are redistributed to reach an equilibrium of Fermi energy (Eƒ), and Fermi levels of the two semiconductors eventually converge to be uniform (Huang et al. 2015, Huang et al. 2016, Tong et al. 2012). Therefore, an internal electric field can be established in the Bi2SiO5 and β-Bi2O3 composite, and the positive and negative charges can be accumulated on Bi2SiO5 and β-Bi2O3, separately. Simultaneously, the position of conduction band and valence band has been altered at Fermi level. As Fermi level is uniform, the band position of Bi2SiO5 moves to the positive position, while β-Bi2O3 moves to the negative position. Thus, under the solar light irradiation, the electrons on the β-Bi2O3 conduct band can be easily transferred to the CB of Bi2SiO5, and react with O2 to produce ·O2−. ·O2− is an active species with strong oxidizing ability and can react with RhB. Conversely, h+ on the Bi2SiO5 valence band can be easily transferred to the VB of β-Bi2O3, and then directly reacted with RhB, which allows the electrons and holes to separate more efficiently and reduce carrier recombination, thus enhancing the photocatalytic efficiency.