3.1 Characterization of catalysts
Fig.1 shows the XRD patterns of CdS and MnxCd1-xS with different doping ratios. As shown in Figure 2,CdS and MnxCd1-xS samples have distinct diffraction peaks at 26.5°, 30.8°, 43.9°, 52.1°, 70.3°, and these diffraction peaks correspond to the (111), (200), (220), (311), and (331) crystal planes of the cubic sphalerite structure CdS (JCPDS NO. 10-0454). The position of the diffraction peak at 26.5° was not shifted after amplification, indicating that the doping of Mn did not change the CdS crystal structure [41].
Figure.2 shows the transmission electron microscopy (TEM) images used to further demonstrate the morphological structure of CdS and Mn0.2Cd0.8S. As shown in Fig.2 (a, b), the microscopic morphology of CdS is spherical with a more uniform size of about 3-5μm. Figure.2(c, d) shows the SEM images of Mn0.2Cd0.8S, and it can be seen that there is no difference in the morphology and size of CdS and Mn0.2Cd0.8S,Mn0.2Cd0.8S microspheres have a rougher surface compared to CdS, giving them a larger specific surface area.
TEM and HRTEM images of Mn0.2Cd0.8S are shown in Figure.2(e, f). The diameter of Mn0.2Cd0.8S can be observed to be about 3μm through Fig. 2 (e, f), which is basically consistent with the SEM results. Only one lattice stripe can be observed by HRTEM, and by calculation it was found that the lattice stripe d=0.336 nm corresponds to the (111) crystallographic plane of cubic sphalerite CdS (PDF#10-0454), which suggests that Mn doping has not changed the lattice size.
Fig.2(g-i) shows the EDS elemental mapping profile of the Mn0.2Cd0.8S photocatalyst. The presence of Mn, Cd and S elements in the sample can be observed in the figure, and the three elements are uniformly distributed in Mn0.2Cd0.8S, which proves the successful doping of Mn in Mn0.2Cd0.8S.
Figure.3(a) shows the photoluminescence spectra of CdS and Mn0.2Cd0.8S. Under the excitation of light at a certain wavelength the semiconductor produces e- and h+, which in combination produce a light signal. Photogenerated carrier transfer and separation were investigated by comparing PL spectra [28, 29], which were measured at an excitation wavelength of 300 nm for CdS and Mn0.2Cd0.8S. As shown in Figure.3(a), CdS and Mn0.2Cd0.8S have obvious luminescence peaks at 375 nm, and the intensity of luminescence peaks of Mn0.2Cd0.8S is lower than that of CdS. This indicates that the doping of Mn reduces the intensity of the luminescence peaks, and the lower luminescence intensity indicates that the Mn0.2Cd0.8S photogenerated electrons and holes are not easy to be compounded, and, the separation efficiency is high.
The maximum wavelength of light absorption of the cubic sphalerite phase CdS in Figure.3(b) is 571 nm. With Mn doping, the light absorption band edge of MnxCd1-xS is red-shifted, where the light absorption band edge of Mn0.2Cd0.8S is 601 nm. For direct bandgap semiconductors, the forbidden bandwidth of the material can be calculated from hυ with [F(R)·hυ] 0.5 .[30,31] The calculated results are shown in Fig.3(c), the forbidden bandwidth of CdS in cubic fibrillar zincite phase is 2.2 eV, and the doping of Mn gradually decreases the forbidden bandwidth of CdS. Among them, Mn0.2Cd0.8S has the lowest forbidden bandwidth of 2.08 eV, which suggests that the doping of Mn improves the absorption of visible light by the photocatalyst and promotes electron leaps, which is favourable for the improvement of photogenerated carrier separation efficiency.
Figures.4(a-d) show the electrochemical test patterns of CdS and Mn0.2Cd0.8S. The potentials measured by the electrochemical workstation were calibrated to the reversible hydrogen electrode (RHE) using the following formulae, and a series of electrochemical test patterns were then obtained:
The linear voltammetric scan (LSV) curves of the cubic sphalerite phases CdS and Mn0.2Cd0.8S are shown in Figure.4(a). In the figure, the overpotentials of CdS and Mn0.2Cd0.8S are 1146 mV and 881 mV, respectively, when the current density is 10 mA/cm2, and Mn0.2Cd0.8S has lower overpotentials. The doping of Mn decreases the overpotential of CdS, which is able to inhibit the complexation of photogenerated electrons and holes and increase the electron transfer[32,33], thus improving the photocatalytic activity. The transient photocurrent response (i-t) curves of the cubic sphalerite phases CdS and Mn0.2Cd0.8S are shown in Figure.4(b). Both CdS and Mn0.2Cd0.8S showed fast transient photocurrent response under visible light, indicating that the samples have high sensitivity to visible light. Higher peak intensities indicate higher separation efficiency of photogenerated carriers [34], and Mn0.2Cd0.8S has higher peak intensities compared to CdS, which suggests that Mn0.2Cd0.8S has better photogenerated carrier separation efficiency. Figure.4(c) shows the electrochemical impedance (EIS) profile of CdS and Mn0.2Cd0.8S. The interfacial resistance during photogenerated electron migration is proportional to the radius of the curve in the Nyquist plot; the smaller the radius, the lower the interfacial resistance and the greater the carrier migration and transport [35]. Compared to CdS, Mn0.2Cd0.8S has a smaller curve radius, which implies that the interfacial resistance during photogenerated electron migration in Mn0.2Cd0.8S samples is smaller. Therefore, it is favourable for the separation of photogenerated carriers, so that Mn0.2Cd0.8S has a better photo-charge separation efficiency. The flat-band potentials of CdS and Mn0.2Cd0.8S photocatalysts were calculated by the Mott-Schottky equation, measured at 1000 Hz, as shown in Figure.4(d). Both CdS and Mn0.2Cd0.8S have positive slopes, indicating that CdS and Mn0.2Cd0.8S have n-type semiconductor properties.[36] The value of the intersection of the tangent to the M-S curve with the x-axis is the flat-band potential, yielding flat-band potentials of -0.61 eV and -0.48 eV for CdS and Mn0.2Cd0.8S, respectively (vs. NHE). For n-type semiconductors, the conduction band is about 0.1 eV-0.3 eV more negative than the flat band [37], so the conduction band positions can be roughly calculated to be roughly -0.71 eV and -0.58 eV for CdS and Mn0.2Cd0.8S.
3.2 Photocatalytic performance and mechanism
As shown in Fig.5(a), the adsorption rate of CdS on MB was close to 0 in the light-avoidance (0-60min) phase, and the adsorption rate of MnxCd1-xS was significantly higher than that of the cubic sphalerite phase CdS. This is due to the fact that the doping of Mn roughens the surface of the MnxCd1-xS microspheres and increases the specific surface area, which leads to an increase in the adsorption rate of MnxCd1-xS. The photocatalytic degradation efficiency of MB by CdS was the lowest in the photocatalytic phase of 60-240 min, which was only 40%. The photocatalytic performance of MnxCd1-xS was improved by Mn doping, and the photocatalytic degradation efficiencies of MnxCd1-xS were all better than those of CdS. Among them, Mn0.2Cd0.8S showed the highest photocatalytic activity of 90 % at 240 min, which was 50 % higher compared to CdS. This is due to the fact that Mn0.2Cd0.8S has the lowest forbidden band width, and the doping of Mn improves the visible light absorption ability of the photocatalyst, promotes electron leaps, accelerates electron transfer and reduces the complexation of photogenerated electrons and holes, so the photocatalytic performance is significantly improved. It has been shown that most photocatalytic degradation systems can be simulated by the Langmuir-Hinshelwood first-order kinetic model. Therefore, the photocatalytic processes of CdS and Mn0.2Cd0.8S were subjected to first-order kinetic fitting to analyse the effect of Mn elemental doping on the reaction rate [38]. The simulation formula is shown below:
In the formula, C0 is the concentration of the original MB sample; Ct is the concentration measured after t samples, and k1 is the quasi-primary reaction rate constant (min-1).
The kinetic fitting analyses for CdS and Mn0.2Cd0.8S are shown in Figure.6.(b). Both curves have a high degree of fit (R2 > 0.96), indicating that the photocatalytic degradation process fits well into the first-order kinetic model. The reaction rate constant for the cubic sphalerite phase CdS can be calculated by fitting a first-order kinetic curve to be 0.00254 min-1 (R2 =0.981), but that for Mn0.2Cd0.8S to be 0.00982 min-1 (R2 = 0.962), which is 3.86 times higher than that for CdS.
Figure.5.(c) shows the absorption spectra of Mn0.2Cd0.8S degraded MB. In the experiment of photocatalytic degradation of MB by Mn0.2Cd0.8S the liquid was taken every 30 min for the absorption spectrum test, and the absorption spectrum of MB was obtained. In this A0-A8 are samples of MB solution taken at 30 min intervals, where A0-A2 is the light-avoidance phase. From Figure.5.(c), it can be seen that the MB solution has distinct characteristic peaks at 293 nm and 664 nm. At light avoidance (A0-A2), a significant decrease in the intensity of the two characteristic peaks was observed. In the light phase (A2-A8), the intensity of the characteristic peak at 293 nm was decreasing with increasing reaction time, which indicates the degradation of the branched chains and their intermediates in the MB molecule . The characteristic peak at 664nm was blueshifted and the intensity of the characteristic peak decreased continuously from the A3 stage, and finally the characteristic peak was blueshifted to 619 nm in the A7 stage. This suggests that demethylation of the MB molecule occurs and the unsaturated conjugated structure in the MB molecule is broken, giving rise to intermediates such as asphaltene A . In addition, the decreasing intensity of these characteristic peaks also indicates that these intermediates are also being decomposed, eventually to CO2 and H2O.
Figures.5(d,e) show the experimental diagrams of photocatalytic cycling of CdS and Mn0.2Cd0.8S and the XRD diagrams of Mn0.2Cd0.8S before and after cycling. Cycling stability is an important index for evaluating the performance of photocatalysts. The cubic sphalerite phase CdS and Mn0.2Cd0.8S, which has the highest photocatalytic activity in the photocatalytic experiments, were selected to perform three photocatalytic cycling experiments, respectively, and the results are shown in Figure.5(d). The photocatalytic decolourisation rate of CdS in the cubic fibrous zinc ore phase after three cycles of experiments was only 18.9 %, which was lower than 1/2 of the first photocatalytic decolourisation rate. While the decolourisation rate of Mn0.2Cd0.8S was 73.6 % after three cycle tests, which was 16.4 % lower than that of the first photocatalytic experiment. The Mn0.2Cd0.8S photostability is improved compared to the cubic fibrous zincite phase CdS, but the photocatalytic performance still shows a decrease. On the one hand, this is due to a certain amount of loss when recovering the photocatalyst in the cyclic test, and on the other hand, a certain amount of photocorrosion may still be present. In addition, we compared the XRD of the Mn0.2Cd0.8S photocatalysts before and after the cycling test. As shown in Fig.5(e), Mn0.2Cd0.8S did not show any other diffraction peaks and the chemical properties were able to remain stable. In summary, the doping of Mn can alleviate the photocorrosion of CdS to a certain extent and improve the repeatability of the photocatalyst.
To elucidate the photocatalytic degradation mechanism of MnxCd1-xS, 0.2 mM of TBA, BQ, and EDTA-2Na were each added to the catalytic system prior to the photocatalytic degradation of MB experiments with Mn0.2Cd0.8S. Detection of active groups involved in the photocatalytic reaction, as shown in Fig.5(f) for the trap test curve of Mn0.2Cd0.8S. As can be seen from the figure, the photocatalytic degradation MB curves did not differ significantly when TBA was added to capture·OH. It indicates that·OH does not play a role in the photocatalytic degradation of MB in the Mn0.2Cd0.8S photocatalytic degradation experiment. When EDTA-2Na was added to the system of photocatalytic degradation of MB by Mn0.2Cd0.8S to capture holes (h+), the rate of photocatalytic decolourisation decreased by only 6.5 % at 240 min, which suggests that h+ is not the main active substance in the reaction system. The addition of BQ-captured superoxide radicals (O2-·) to the system of photocatalytic degradation of MB by Mn0.2Cd0.8S decreased the rate of photocatalytic decolourisation to 56 % for 240 min, a decrease of 34 %. This indicates that the active group playing a major role in the system of photocatalytic degradation of MB by Mn0.2Cd0.8S is the superoxide radical (O2-·).
In summary, we can deduce the mechanism of photocatalytic degradation of MB by Mn0.2Cd0.8S, as shown in Fig.6. A combination of UV-vis analyses and Mott-Schottky tests yielded Eg and ECB of 2.08eV and -0.58eV each for Mn0.2Cd0.8S, so the EVB can be calculated to be 1.46 eV. The redox potential E0(O2/O2-·) of O2/O2-·with respect to NHE is -0.33 eV [39, 40], and the conduction band position of Co0.08Cd0.92S (-0.58 eV) is more negative than E0(O2/O2-·). Thus the photogenerated electrons (e-) at the bottom of the CB can react with the surrounding O2 to generate O2-·as the active substance in the degradation process, and ultimately the O2-·degrades the MB solution to H2O, CO2 and other products, as evidenced by the MB absorption spectrum. The valence band position of Co0.08Cd0.92S (1.5 eV) is more negative than that of E0(·OH/OH- ) (+2.38 eV vs. NHE) [34], and the photogenerated holes are unable to oxidise the ambient OH- to produce·OH, which is also consistent with the results of the trapper test. Therefore, the following reaction equation can be deduced for the photocatalytic process: