3.1 Characterization
The morphology of as-prepared MoS2/CdS was characterized by SEM and TEM images. Fig. 1a and 1b show SEM image of CdS and TEM image of 2.5 wt % MoS2/CdS, respectively. The olive-like CdS had an irregular block-shape, rough surface and obvious aggregation(Chai et al. 2018, Chen et al. 2012). For 2.5 wt% MoS2/CdS composite, the irregular flaky MoS2 was anchored at the edge of CdS(Xu &Cao 2015, Zhang et al. 2018, Zong et al. 2010), As inset in Fig. 1a and 1b, the color of bright orange CdS was transformed into deep orange after adding MoS2.
XRD analysis was used to characterize crystal structure of MoS2/CdS (Fig. 1c). The diffraction peaks at 2θ = 24.9, 26.5, 28.3, 43.8, 48 and 52º corresponded to the (100), (002), (101), (110), (103) and (112) planes of cubic CdS, respectively(Alomar et al. 2019, Jin &Li 2020). The high diffraction peak intensity of (101) plane indicated the growth of CdS along (101) crystal plane(Wu et al. 2017). The large and sharp diffraction peak intensity of 2.5 wt% MoS2/CdS showed the good crystallization. Compared to CdS, no obvious shift in the diffraction peak of MoS2/CdS composites indicated the loading of MoS2 rather than doping(He et al. 2016). However, no diffraction peak of MoS2 may be due to the small content(Li et al. 2019b). No significant change in diffraction peaks after photocatalytic reaction indicated 2.5 wt% MoS2/CdS composite with strong stability.
Figure 1d shows the FT-IR spectra of 2.5 wt% MoS2/CdS nanocomposites before and after the reaction in the frequency range of 400 ~ 4000 cm−1. The IR bands at 428 and 508 cm−1 corresponded to the stretching vibration of Mo-S and (S-S)2− bond, respectively(Habibi &Rahmati 2014). The stretching vibration of Cd-S bond and the bending vibration of H2O were observed at 630 and 1572 cm−1, respectively. The broad peak in the range of 3200 ~ 3500 cm−1 was related to the vibration of O-H bond(Khawula et al. 2016). After the reaction, no new bands and little shift indicated that main interaction mechanism of U(VI) on MoS2/CdS composite was photoreduction rather than adsorption(Yan et al. 2009, Yuan et al. 2018).
Figure S2 shows nitrogen adsorption-desorption isotherms and corresponding pore size distribution curves of CdS and 2.5 wt% MoS2/CdS composite. The IV-type adsorption isotherms showed H3 hysteresis loop, indicating the aggregation of plate-like particles on slit-like pores(Misra et al. 2009, Sing et al. 1985). The main size distribution (~ 2 and 85 nm) suggested the formation of mesoporous and microporous 2.5 wt% MoS2/CdS composite. From table S1, BET specific surface area of MoS2/CdS composite (~8.8 m2/g) was slightly higher than CdS (3.8 m2/g) due to the increase of mesopore by adding MoS2(Yu et al. 2014), which is beneficial to improve the catalytic performance due to accommodation of more surface active sites and promotion of the transport of charge carriers (Alomar et al. 2019, Di et al. 2016).
To study the light absorption characteristics, UV-vis DRS and PL spectra (at 500-700 nm) of 2.5 wt% MoS2/CdS composite were showed in Fig. 2a and 2b, respectively. For CdS, a strong absorption band edge at ~ 530 nm corresponded to the inherent band gap absorption of hexagonal CdS(Li et al. 2019b, Yin et al. 2016b). It is worth noting that the gradual increase of light absorption intensities in 580-800 nm with the increase of MoS2 may be due to the color shift from bright orange into dark orange MoS2/CdS composite(Zhang et al. 2014). According to the calculation of Tauc plots (Supporting information), the band gap of 2.5 wt% MoS2/CdS composite (2.31 eV) was significantly lower than CdS (2.43 eV, inset of Fig. 2a) owing to charge transfer from CV of MoS2 to CdS(Zong et al. 2010).
PL spectrum as an important tools provides important information of photo-generated electron-hole separation efficiency(Lin et al. 2017). The high PL intensity of CdS (Fig. 2b) indicated the high recombination centers due to trap-related states(Wu et al. 2017). The low PL intensity of 2.5 wt% MoS2/CdS composite indicated low recombination of photogenerated electrons and holes(Yuan et al. 2018), which also reduce the photo-corrosion of CdS(Zong et al. 2008). The effective transfer of photogenerated electrons from CB of CdS into CB of MoS2 indicated that the introduction of MoS2 inhibited recombination of photogenerated electrons and holes, which significantly improved the photocatalytic activity(Yin et al. 2016a, Zhang et al. 2014).
To further understand the charge separation, the photocurrent responses of CdS and 2.5 wt% MoS2/CdS under several intermittent full-light irradiation was showed in Figure 2c. The transient photocurrent of all samples quickly reached a constant value under continuous illumination(Jiang et al. 2015), whereas the photocurrent reached to 0 after close of irradiation. Compared to CdS (4.76 µA/cm2), the high photocurrent density of 2.5 wt% MoS2/CdS composite (31.13 µA/cm2) indicated the effective inhibit recombination of photogenerated electrons and holes due to the formation of heterostructure(Chen et al. 2015, He et al. 2016).
As shown by Nyquist diagram of EIS spectrum in Fig. 2d, the semicircle related to charge transfer resistance (Rct) is connected in parallel with a sub-layer capacitor (CPE)(Li et al. 2019a, Yu et al. 2014). Compared to CdS, the small semicircle of 2.5 wt% MoS2/CdS indicated the fast electron transfer due to the excellent conductivity of the introduced MoS2(Qin et al. 2017). The characterization results reflected that the MoS2/CdS heterojunction improved the separation and transfer efficiency of photon-generated carriers (Xu &Cao 2015, Zhang et al. 2014).
3.2 Effect of different photocatalysts and pH
As shown in Fig. 3a, the no removal of U(VI) was observed in the absence of photocatalysts and light irradiation, which ruled out the possibility of self-photolysis of U(VI) (Wang et al. 2020). Under the dark conditions, the adsorption of U(VI) on photocatalysts slightly increased (from 11.3 to 16.7%) with increasing doping amount of MoS2, indicating no obvious difference in the adsorption capacity of photocatalysts for U(VI)(Li et al. 2019c). For CdS, the low photocatalytic efficiency of CdS (i.e., ~ 0.6% after 20 min irradiation) was mainly due to the rapid recombination of photogenerated electrons and holes(Dai et al. 2021, Qiu et al. 2021, Zhang et al. 2020b). Within 15 min irradiation, the photocatalytic efficiency of 2.5 wt% MoS2/CdS (97.51%) was remarkably higher than 10 wt% MoS2/CdS (65%) due to the reduction of active sites by adding aggregated MoS2(Chen et al. 2019, Zhang et al. 2020a). Fig. 3b shows the effect of pH on the photoreduction efficiency of U(VI) on 2.5 wt% MoS2/CdS. Almost no removal of U(VI) at pH 3.0 may be due to the competing of H+ and U(VI)(He et al. 2017, Zhang et al. 2020b). The highest removal rate of U(VI) (99.64%) was observed at pH 7.0 due to the fast transfer of generated charges of 2.5 wt% MoS2/CdS under neutral conditions(Dai et al. 2021), which limited the recombination of photogenerated electrons and holes (Li et al. 2019d).
3.3 Effect of U(VI) concentration and dosage
Figure 3c and 3d show the effect of U(VI) concentrations (10, 20 and 30 ppm) and dosage (0.5, 1.0 and 2.0 mg/mL) on U(VI) photoreduction, respectively. Within 5 min irradiation, the highest removal of U(VI) at 20 ppm was consistent with the previous studies(Lei et al. 2021, Zhang et al. 2020b). Similarly, the highest removal efficiency of U(VI) was observed at 1.0 g/L under light irradiation. The further increase of U(VI) concentration and photocatalyst dosage leads to the decrease of U(VI) photoreduction due to the occurrence of shielding effect after exceeding the optimal concentration and dosage(Jiang et al. 2021, Zhang et al. 2021).
3.4 Regeneration and quenching experiments
Figure 4a and 4b show the regeneration and quenching experiments of U(VI) photoreduction on 2.5 wt% MoS2/CdS, respectively. After 5 times of recycling, 2.5 wt% MoS2/CdS still exhibited the high photocatalytic removal efficiency (>97%, Fig. 4a). In addition, no change in microstructure and crystal structure was observed after the photocatalytic reduction process (Fig. 1c and 1d). The regeneration results indicated that 2.5 wt% MoS2/CdS had high photoreduction efficiency, good chemical stability and excellent reusability(Yu et al. 2014). The photoreduction mechanism of U(VI) on 2.5 wt% MoS2/CdS composites was investigated by quenching experiments. TBA, p- BQ and Me were used as electron acceptors to remove ·OH, ·O2− and holes, respectively(Jiang et al. 2018, Wang et al. 2019). As shown in Fig. 4b, only 20% of U(VI) was photo-reduced after adding p-BQ into the reaction system, indicating the significant inhibition of photocatalytic activity. The little effect of TBA and Me on U(VI) photoreduction was observed. This experiment confirmed that ·O2− radicals were responsible for U(VI) photocatalytic reduction (Yuan et al. 2018).
In order to further clarify the reaction mechanism, an electron spin resonance spin trap experiment was performed, which can detect the generation of ·OH and ·O2− in the light experimental system(Zhang et al. 2021). As displayed in Fig. 4c and Figure S3, no peak and characteristic peaks of DMPO-·OH and DMPO-·O2− were observed under the dark and light condition, respectively, indicating that ·OH and ·O2− were generated at λ > 420 nm (Gong et al. 2021). Compared to DMPO-·OH, the intensities of DMPO-·O2− were significantly weak after adding U(VI), which can be deduced that dissolved oxygen acts as an electron shuttle between 2.5 wt% MoS2/CdS(Le et al. 2020).
3.5 XPS analysis
XPS analysis was used to determine the photocatalytic reduction of U(VI) on CdS and 2.5 wt% MoS2/CdS composite. Apart from Cd 3d and S 2p, the appearance of low Mo 3d peak showed the success synthesis of MoS2/CdS composite (Fig. 5a). In addition, U 4f peak was also found for 2.5 wt% MoS2/CdS composite after reaction. As shown in Fig. 5b, the two peaks of Mo 3d at 226.0 eV and 232.5 eV corresponded to Mo4+ 3d5/2 and Mo4+ 3d3/2, respectively(Alomar et al. 2019, Jin &Li 2020). After the reaction, Mo4+ 3d5/2 was shifted by +0.2eV, while Mo4+ 3d3/2 had no obvious shift. Similarly, S 2p spectra at 161.8 and 163.0 eV corresponded to S 2p3/2 and S 2p1/2, respectively (Fig. 5c). Compared to CdS, the increased binding energy and slight shift (0.1-0.2 eV) of S 2p in 2.5wt% MoS2/CdS was due to addition of MoS2(Wu et al. 2017). The slight shift of Cd 3d (i.e., Cd 3d5/2 from 405.2 to 405.4 eV, Cd 3d3/2 from 411.9 to 412.1 eV, Fig. S4 of SI) was due to the change of Cd 3d orbit caused by the chemical bond(Yuan et al. 2018). For 2.5wt% MoS2/CdS, the change in binding energy of Cd 3d and S 2p may be likely to accelerate the separation of excited charges of 2.5 wt% MoS2/CdS under the light irradiation, improving the photocatalytic activity(Yu et al. 2021). U 4f peak at 381.62 and 392.50 eV were assigned to U 4f7/2 and U 4f5/2 of U(VI), respectively (Fig. 5d). U 4f7/2 can be decomposed into two peaks of 380.1 and 381.6 eV, which were attributed to U(VI) and U(IV), respectively(Zhang et al. 2020a). The similarity of U 4f5/2 peak was also observed, indicating photoreduction of adsorbed U(VI) into U(IV) by photogenerated electrons in a short time(Li et al. 2019d, Wang et al. 2019, Zhang et al. 2020b).
3.6 Mechanism analysis
Based on the above experimental results and band structure theory, the mechanism of catalytic reduction of U(VI) with 2.5 wt% MoS2/CdS catalyst under all-light is roughly as follows(Chang et al. 2015):
MoS2/CdS + hν→ e− + h+ (1)
H2O + h+ → ·OH+H+ (2)
CH3OH + h+ +·OH→ CO2+H2O (3)
O2 + e−→ O−2 (4)
U(VI)+ ·O−2 → U(IV) + 2O2 (5)
The electrons (e−) in the valence band (VB) are excited under light irradiation when the light energy is greater than or equal to the band gap of MoS2/CdS (hv ≥ Eg)(Gorshkov et al. 2006). Then photogenerated electrons were jumped into the conduction band (CB), an equal number of holes (h+) are generated in VB (Eqn. 1). The holes of VB and electrons of CB have certain oxidation and reduction ability, respectively(Yin et al. 2016b, Zong et al. 2010). h+ can react with water to form ·OH radicals (Eqn. 2), and/or h+ can oxidize organics into CO2 and H2O (Eqn. 3). The ·O−2 radicals were generated by reaction of e− and dissolved O2 (Eqn. 4). Therefore, the adsorbed U(VI) by the active site of catalyst can be reduced to U(IV) by ·O−2 (Eqn. 5)(Jiang et al. 2018). The generated O2 was quickly captured by e− again (Eqn. 4).