3.1. Characterization
Figure 1a shows g-C3N4 (CN) has two characteristic peaks at 13.1° and 27.4°, which are crystal plane (100) and crystal plane (002), respectively [21]. The peaks at 2θ = 12.0, 24.2, 25.9, 32.6, 33.6, 46.8 and 54.1° agreed with the (001), (002), (101), (110), (102), and (200) crystal planes of BiOCl [10], individually. All the diffraction peaks are consistent with the standard card of BiOCl (JCPDS No.06-0249), indicating that BiOCl materials can be successfully prepared by the solid grinding method. Only one BiOCl peak appears in the synthesis of the CN/BOC-X composite, which may be attributed to the weak diffraction intensity of CN and its coverage, indicating that the presence of CN still does not affect the formation of BiOCl material in the grinding process. According to the FT-IR spectrum shown in Fig. 1b, some absorption bands can be seen in the range of 1636 − 1244 cm− 1 of the CN sample, which may be accredited to the characteristic stretching vibration mode of g-C3N4 of the C = N heterocycle [22]. The distinctive peak gradually becomes more clear with an growth of CN content in the CN/BOC-X composite. In addition, a characteristic peak of CN was also found at 816 cm− 1, which was considered to be the out-of-plane respiratory vibration peak of the 3-S-triazine unit [23]. The retention peak at 521 cm− 1 is caused by the extended vibration of BI-O in BiOCl, and the retention peaks at 1632 cm− 1 and 3436 cm− 1 are related to the extended vibration modes of adsorbed water molecules and surface hydroxyl groups, respectively [24]. XRD and FT-IR results show that the CN/BOC-X composite is composed of CN and BiOCl.
The absorption edges of pure CN are at approximately 485 nm (Fig. 1c). BOC displays an absorption edge of about 375 nm, but it demonstrates an absorption tail expanding to the visible region, indicating induction of OVs [25]. Compared with BOC, the absorption wavelength of light in the CN/BOC composite gradually redshifted and the corresponding band gap also gradually decreased (Fig. S1) with the increase in CN amount. Because CN and BOC form a heterojunction structure with matched energy levels during synthesis, they have light absorption ability in the visible light range. Figure 1d shows that CN has the strongest PL intensity, while all CN/BOC-X shows a significant decrease in PL intensity and CN/BOC-5 has the lowest value, indicating that after heterojunctions are formed between CN and BOC, light-generated charge carriers are effectively separated [26]. Under the same conditions, there is only a weak PL peak of BOC at 612 nm due to the limited amount of e− and h+. Based on N2 adsorption–desorption isotherms (Fig. 1e), all synthetic materials show typical IV adsorption–desorption isotherms [27]. CN not only has the highest adsorption volume, but also has a significant hysteresis loop, indicating that, it has a larger specific surface area with a rich mesoporous structure. Conversely, the BOC has the lowest adsorption volume of all samples. With the increase of BiOCl in CN/BOC-X, the adsorption amount of N2 decreases gradually and hysteresis loops become inconspicuous, which may be attributed to the fact that BiOCl nanoparticles fill the mesoporous pores in CN. Experimental results show that the specific surface areas of CN, CN/BOC-2, CN/BOC-5, CN/BOC-10, and BOC are 20.5, 11.3, 8.1, 5.7, and 1.7 m2 g− 1, respectively. Figure 1f shows that CN and CN/BOC-X have micropores in the range of 1.2-2.0 nm, and with the increase of BOC pores on the decrease. At the same time, with the increase of BOC, the minimum mesopore in CN/BOC-X also increased from 7.1 to 9.8 nm. The pure BOC has no micropores and only has a pore size greater than 12 nm. These results indicate that the material with BOC may enter the porous CN and form an intimate contact heterogeneous interface. The CN/BOC-X with sufficient pore distribution can typically attain efficient molecular and ion transport, which can encourage more effective active sites for photocatalytic action [28].
Figure 2a XPS showed C, N, Bi, O, and Cl, with peaks of the elements appearing on the CN/BOC-5 material. In Fig. 2b, the main peak of the CN/BOC-5 composite at 288.2 eV can be attributed to the N-C = N bond [29] from CN, while the C peak at 284.7 eV should come from the adsorption of impurity carbon by BOC [30]. It can be seen from Fig. 2c that there are four peaks in the N 1S spectrum of CN at 398.7, 400.1, 401.2, and 404.5 eV, which represent the N in C-N = C, C-N-H, N-(C)3, and π-excitation, respectively [30]. After recombination with BOC, the binding energy of the N element shifted to a higher position by 0.1–0.2 eV, which may be attributed to the adsorption of Bi3+ by lone pair electrons in nitrogen [29]. As can be viewed from Fig. 2d-f, the binding energies of Bi, O, and Cl main peaks all decreased by 0.1 eV in CN/BOC-5 nanosheets compared with pure BOC. However, oxygen vacancies (OVs) appear at 531.5 eV in the composite materials, and the peak value is gradually apparent and shifted by 0.3 eV [31]. At the same time, the adsorption hydroxyl group on the peak surface at 533.0 eV also shifted by 0.3 eV. Above, XPS results indicate that heterojunctions are formed in the CN/BOC-5 composite through the strong interaction between CN and BOC [32].
Figure 3a, and d show CN is a typical aggregated lamellar structure. SEM (Fig. 3b, e) images of pure BOC show the extremely smooth surface and thick lamelae of BOC crystals. However, after BOC was introduced into the CN synthesis system, the lamellar thickness was significantly thinner, resulting in the formation of a 2D/2D thinner sheet structure of BOC/CN nanocomposites (Fig. 3c, f and Fig. S2). Compared with CN in Fig. 3d, it is confirmed that BOC nanosheets are deposited uniformly on the surface of CN nanosheets of BOC/CN (Fig. 3f), and the original form of CN is not obviously changed. According to Fig. 3g, CN/BOC shows a good contact interface between different crystal materials, and the lattice spacing of about 0.28 nm and 0.34 nm is corresponds to (110) and (101) crystal planes of BOC, respectively [33]. In addition, some lattice defects (marked by blue ovals) can be clearly seen, which may be caused by the existence of oxygen vacancies or defects on the crystal surface [34]. The EDX mapping (Fig. 3h-m) indicates that Bi and Cl have the same shape and distribution as the mapping images of C and N, and the shape presented by the O element is the superposition of the two because both of them are contained, which further confirms the presence of the nanosheet structure of BOC and CN in the sample. In addition, the uniform isolated distribution of BOC nanoparticles (Fig. 3k, I and m) with an extremely small size in the porous CN matrix (Fig. 3h, I and j). The possible mechanism is that N atoms in porous g-C3N4 nanosheets may be exercised as adsorbed Bi3+ ions [26], and then the nucleated BOC grows in situ on the surface or in the porous channel structure of g-C3N4, which corresponds to the SBET conclusion. Finally, the gradually growing BOC's small-sized nanoparticles form the structure of nanosheets. These results also indicated that CN and BOC were successfully coupled and formed a good contact interface. The 2D/2D structure of the CN/BOC heterojunction has a good interface and generates an induced electric field, which can promote the photogenerated e− and h+ separation [24–26].
3.2 Photocatalytic performance
The degradation of tetracycline by the prepared catalyst is shown in Fig. 4a. Before light irradiation, dark adsorption for 60 min made the catalyst get into the adsorption equilibrium state. The adsorption capacity of pure BOC is greater than CN. It may be attributed to the hydroxyl functional group of BOC’s interaction with these oxygen-containing groups in TC via hydrogen bonding [35, 36]. Moreover, in CN/BOC-X, the adsorption amount is obviously higher than pure CN and BOC. After visible light exposure for 120 min, the elimination rate of TC by CN/BOC-X was higher than that by CN and BOC alone. For example, the degradation rates of CN/BOC-2, CN/BOC-5, and CN/BOC-10 reached 78.3, 89.8, and 86.2%, respectively, where the CN-BOC-X is obviously higher than that of single CN (47.1%) and BOC (73.1%), respectively. In addition, the optimal degradation rate of CN/BOC-5 (0.0193 min− 1) was 4.6 and 1.5 times higher than that of single CN (0.0042 min− 1) and BOC (0.0130 min− 1), respectively (Fig. 4b). The reason may be that, compared with the single photocatalyst, the 2D/2D heterostructure and oxygen-rich vacancies can enhance the space charge separation at the interface and improve the photocatalytic activity [3, 8, 24].
In addition, the measurement of Rh.B also shows that the preferred CN/BOC-5 of photo-degradation amount reaches 88% after 60 min, much higher than 15% of CN and 67% of BOC (Fig. S3a). The value degradation rate of CN/BOC-5 is 0.0295 min− 1, which is 18.4 and 1.8 times the rate constants of CN (0.0016 min− 1) and BOC (0.0162 min− 1) (Fig. S3b). As shown in Fig. 4c, d, the samples were also tested for photocatalytic CO2 reduction by the gas-solid method. After 7 hours of light exposure, 2D/2D CN/BOC-5 heterostructure CO (2.00 µmol h− 1 g− 1) had the highest yield, which is 1.3 times and 3.8 times that of single CN (1.49 µmol h− 1 g− 1) and BOC (0.53 µmol h− 1 g− 1), respectively (Fig. 4d). It may be realized, that the CO generation rate is significantly increased after CN and BOC recombination, indicating, that 2D/2D interface contact plays a significant essential part in CO2 transformation [8, 24]. At the same time, CN/BOC-5 photocatalytic organic pollutants and carbon dioxide reduction is comparable or superior to that of BiOCl [15], porous g-C3N4 [37], even BiOCl-g-C3N4 composite system [38], etc., as listed in Tables S1and S2.
3.3 Mechanism
The 2D/2D CN/BOC heterojunction materials show excellent photocatalytic activity, and the mechanism needs to be further explored. Figure 5a shows that the ElS radius of the CN/BOC-5 composite is the smallest, indicating that the charge transfer resistance is the smallest, indicating that excellent interfacial transfer efficiency will improve the separation effect of e−-h+ pairs [23]. Figure 5b shows that the transient photo-current curves of all samples show stable photo-response characteristics, and the photo-current density of the CN/BOC-X heterojunction materials is higher than that of single CN and BOC. The photo-current density of CN/BOC-5 photocatalyst is the most significant, which indicates that appropriate heterojunction content has a positive effect on further improving the photocatalytic activity. By direct electron paramagnetic resonance (EPR) test, the peak at g = 2.003 was more intense with the addition of BOC and CN (Fig. 5c), which proved the presence of oxygen vacancies in the CN/BOC-X [39], which is consistent with the results of HRTEM and XPS. A new defect level was added by introducing OVs, which led to the enhanced photo absorption efficiency [12]. The surface OVs can act as electron trapping sites and inhibit the recombination of e−-h+, which plays an important role in the charge transmission process of the heterostructure [13].
As shown in Fig. 5d, with the addition of BQ or EDTA-2NA, the photocatalytic activity of CN/BOC-5 decreased significantly, indicating that •O2− and h + played a significant role in the photocatalytic reaction. At the same time, the addition of IPA had a moderate influence on the photocatalytic efficiency, indicating that the production of •OH radical is not sufficient. Furthermore, when DMPO was used as a spin capture agent, an ESR signal peak of •OH radical and •O2− free radicals was detected (Fig. 5e, f). Therefore, the synergistic effect of CN and BOC heterojunction materials can produce both h+, •OH and •O2−, which greatly improves the efficiency of catalytic degradation.
The possible mechanism is shown in Fig. 6. Valence band (VB) and conduction band (CB) of g-C3N4 and BiOCl were calculated as follows:
$${E}_{CB}={X-E}_{e}-0.5{E}_{g}$$
1
$${E}_{VB}={E}_{CB}+{E}_{g}$$
2
Where X (absolute electronegativity) of the g-C3N4 is 4.73 eV [40], BiOCl is 6.36 eV [15], Ee is the energy of free electrons on the hydrogen scale (4.5 eV), and Eg has a band gap of 2.3 eV for CN and 3.2 eV for BOC (Fig.S1). The calculated ECB and EVB were − 0.97 and 1.33 eV for g-C3N4, 0.26 and 3.46 eV for BiOCl, respectively. Therefore, CN has an appropriate bandgap in reaction to visible light. Meanwhile, because the defect levels caused by OVs can cause simple visible-light excitation, BOC could be photoexcited to produce e− and h+ [25]. As the trapping experiments and ESR tests demonstrate, generating •O2− (O2/•O2−, − 0.33 V vs. NHE) and •OH/OH− (1.99 eV vs. NHE) may be derived from the conduction band (VB) potential of CN and the valence band (VB) potential of BOC, respectively [10]. In the process of photo-reducing reaction of carbon dioxide, the edge potential of CB of CN is less than the standard redox potential of CO2/CO (-0.53 eV). The potential difference is sufficient for the kinetic and thermodynamic conditions [21]. In fact, in the Z-scheme structure of Fig. 6, the lower energy electrons on CB of BOC and holes on VB of CN are combined, whereas the higher energy level holes on VB of BOC and electrons on CB of BOB are retained, thus maintaining the higher oxidation and reduction ability of the composite photocatalyst. Therefore, the electrons in the CB of CN may contribute electrons to oxygen to generate •O2− radicals, and also the holes in BOC VB will further generate •OH radicals, which is justified by the test calculations as shown in Fig. 5d, f. In particular, the electronic structure is affected by the OVs on the surface of BOC and forms a new defect energy level [39], which may initiate photo-electrons to transition to OVs. As a result, OVs serve as e−-traps and slow down the recombination of e−-h + pairs, extending the lifetime of photo-induced carriers. which results in the increased lifetime of photo-induced carriers [12]. Rich in e−, OVs can be excited by being exposed to vis-light irradiation, transitioning to the CB of BiOCl and absorbing energy [40–42]. In addition, CN/BOC heterojunction forms a strong built-in electric field at the 2D/2D interface, which is favorable to the efficient separation of photogenerated electron-hole pairs [24–26, 43]. Therefore, the obtained CN/BOC-OVs heterojunction with a 2D/2D structure is endowed with synergistic effects, which can significantly improve the photocatalytic oxidation/reduction performance. It is superior to some other g-C3N4, BiOCl, and heterojunction materials (as shown in Table S1 and S2).