3.1 Characterization of the crystal phase and morphology
3.1.1 XRD analyses
The XRD patterns of the prepared CuFe2O4, ZnFe2O4, g-C3N4, Zn0.5Cu0.5Fe2O4 and Z0.5C0.5CN are shown in Fig. 1.
As shown in Fig. 1, XRD pattern of g-C3N4 indicates two characteristic peaks. A weak peak appeared at 13.1°, indexed as (100) crystal planes is ascribed to the inter-planar stacking peak of aromatic systems, while a strong peak at 27.4° with crystal plane (002) accords with interlayer stacking of aromatic systems. The appearance of the strong and weak peaks at 27.4° and 13.1° demonstrates the formation of nanosheets g-C3N4 [37]. For CuFe2O4 and ZnFe2O4, the recorded peaks at 30.1°, 35.5° and 62.6° with crystal planes (220), (311) and (440) respectively, validate that pure CuFe2O4 and pure ZnFe2O4 bear a cubic spinel structure and can be entirely accorded with the results reported in these literatures [38, 39]. Compared with the peaks in the XRD pattern of CuFe2O4 and ZnFe2O4, the peaks corresponding to the synthesized Zn0.5Cu0.5Fe2O4 composite also bear a cubic spinel structure due to the incorporation of Zn2+ and Cu2+ in Zn0.5Cu0.5Fe2O4 composite. This clearly indicates that Cu and Zn is successfully isomorphic in the Zn0.5Cu0.5Fe2O4 matrix. Moreover, it is worth noting that the diffraction peak of g-C3N4 at 27.4° (002) exhibits low peak intensity in Z0.5C0.5CN composites, which is owing to its low contents in the composites as compared to its pure form. The appearance of characteristic peaks of g-C3N4 and Zn0.5Cu0.5Fe2O4 validates the successful formation of Z0.5C0.5CN composites. This observation reveals that the coupling of g-C3N4 sheets with the Zn0.5Cu0.5Fe2O4 does not alter the morphology of g-C3N4 and Zn0.5Cu0.5Fe2O4, which may be beneficial for photocatalytic action of the composite [40].
3.1.2 SEM and TEM of analyses
The morphology and microstructure of g-C3N4, ZnFe2O4, CuFe2O4, Zn0.5Cu0.5Fe2O4 and the Z0.5C0.5CN-50 composites were explored by SEM. Figure 2a shows the morphology of g-C3N4 sample, which has an irregular shape made up of curly and wrinkled sheets structure. Figure 2b and 2c show the morphology of synthesized ZnFe2O4 and CuFe2O4 samples, respetively. The samples of ZnFe2O4 and CuFe2O4 present the micron particle-like structure with the diameter about 0.3 ~ 1.0 µm, which are agglomerated by much tiny nanosheets. As Fig. 2d shows, the Zn0.5Cu0.5Fe2O4 possesses hierarchical microsphere structures with porous and rough surface structure. As is reported, the surface holes are conducive to the adsorption and degradation of EBT, while the Zn0.5Cu0.5Fe2O4 is used for photocatalytic degradation. From Fig. 2e, it can be observed that the Zn0.5Cu0.5Fe2O4 microspheres are inlayed in the matrix of g-C3N4, confirming that the Z0.5C0.5CN-50 heterojunction is composed of 3D Zn0.5Cu0.5Fe2O4 microspheres and 2D g-C3N4 sheets. It is indicated that the tightly contact between the Zn0.5Cu0.5Fe2O4 and g-C3N4 was benefical for the migration of charge carriers in the interface of the Zn0.5Cu0.5Fe2O4 and g-C3N4. It is found that Zn0.5Cu0.5Fe2O4 coupling g-C3N4 leans to agglomerate together and exhibits a uniform and regular 3D microsphere structure. In addition, the structure constructed between Zn0.5Cu0.5Fe2O4 and g-C3N4 provides an appropriate electronic heterojunction, greatly shortening the carrier migration distance and thus promoting the photocatalytic reaction.
In addition, the more detailed structures of Z0.5C0.5CN-50 are investigated by TEM tests. As viewed in Fig. 3a, the Zn0.5Cu0.5Fe2O4 microspheres present the york-shell structure and are consisted of fine lamellar nanoparticles. The g-C3N4 (Fig. 3b) is consist of stacked sheets and possesses numerous pore structures. From Fig. 3c, the 3D york-shel Zn0.5Cu0.5Fe2O4 microspheres are tightly combined with the surface of 2D g-C3N4 nanosheets, in accord with the SEM observation. The lattice fringes and interface between the two components are further explored by HRTEM measurement. The evident heterojunction interface can be observed in Fig. 3c, and the lattice fringe regions are ascribed to Zn0.5Cu0.5Fe2O4, while the amorphous regions are attributed to g-C3N4, indicating that the heterojunction has been formed. The lattice spacing of 0.29 nm are corresponded to the (311) crystal plane of Zn0.5Cu0.5Fe2O4.
3.2 FTIR spectroscopy analyses
Figure 4 shows the FTIR of the g-C3N4, ZnxCu1−xFe2O4 and Z0.5C0.5CN composite. The FTIR spectra of the synthesized ZnxCu1−xFe2O4 nanocomposites are shown in Fig. 4a, It is favor of confirming the formation of the mixed spinel structures. The broad bands around 3400 cm− 1 are due to the O–H stretching vibration of the free or absorbed water. The high surface area of the ZnxCu1−xFe2O4 materials results in the rapid adsorption of water from the atmosphere. The absorption peaks at 1380 and 1630 cm− 1 are ascribed to the N–O stretching mode of undecomposed nitrates ions and the bending mode of absorbed water molecule, respectively [41]. The absorbance peaks around 590 cm− 1 was originated from the formation of the Fe–O bonds within the tetrahedral FeO6 groups of the ferrite latticeis[41]. Distribution of significant bands in the IR spectra for ZnxCu1−xFe2O4 nanocomposites (for various Zn and Cu contents) is listed in Table 1. Small deviations in band positions are discovered for different Zn and Cu contents. It is clear from the spectra (in Table 1) that there is a minor shift in the IR active modes of different materials which is primarily due to the presence of nano-sized grains. This can be attributed to the presence of nano-sized grains.This can be explained by the fact that in the case of nanosized grains the arrangement of atoms at the grain boundaries differs from that of the bulk crystals. This gives rise to disorders in both the coordination number and bond length and this degradation in crystal symmetry is responsible for the shifting in IR active modes. The IR results are in good accord with those reported earlier in the literature [42, 43].
The FTIR spectra of the g-C3N4 (in Fig. 4b) has absorption bands at 3320, 1680, 1273 and 817 cm− 1. The broad absorption bandaround 3300 cm− 1 is caused by the NH2 or NH group at the aromatic ring defect site in the g-C3N4 and by the O–H stretching vibration of the free or absorbed water [44], the absorption band near 1640 and 1273 cm− 1 are coordinated with C–N stretching deformation and C = N stretching deformation of carbon-nitrogen ring of g-C3N4, respectively [45], and near 808 cm− 1 corresponds to the bending vibration of C–N for the triazine [46]. The occurrence of the identical peaks in Z0.5C0.5CN validates the formation of the composites. Correspondingly, a small shift in absorption peaks of Z0.5C0.5CN (approximately at 3320 cm− 1, 1650 cm− 1, 1250 and 813 cm− 1) towards the lower wavenumber is observed compared to that of pristine g-C3N4 peaks. That alteration in peaks identifies the strong chemical interaction rather than simple physical interaction between the components of the heterostructure catalyst [47].
Table 1
IR bands(cm− 1) of ZnxCu1−xFe2O4 (0.0 ≤ x ≤ 0.8)
Materials
|
Peaks of as-prepared samples (cm− 1)
|
Zn0.2Cu0.8Fe2O4
|
468, 597, 968, 1381, 1624, 3412
|
Zn0.4Cu0.6Fe2O4
|
463, 600, 968, 1383, 1631,3421
|
Zn0.5Cu0.5Fe2O4
|
464, 600, 970, 1383, 1633, 3423
|
Zn0.6Cu0.4Fe2O4
|
465, 601, 978, 1383, 1635, 3425
|
Zn0.8Cu0.2Fe2O4
|
467, 600, 978, 1383, 1637, 3429
|
3.3 Specific surface area and surface charge analysis
Figure 5 plots the N2 adsorption-desorption isotherms and the pore size distribution of all as-prepared materials. Based on the IUPAC classification, it was clearly noticed that C3N4, Zn0.5Cu0.5Fe2O4 and Z0.5C0.5CN-50 exhibited the type IV adsorption isotherms. The pure C3N4 showed the types with a type H3 hysteresis loop at a relative pressure of 0.0–1.0, demonstrating the exist of slit hole formed by the carbon nitride layers. The Zn0.5Cu0.5Fe2O4 and the Z0.5C0.5CN-50 showed with a type H2 hysteresis loop at a relative pressure of 0.0–1.0, revealing the presence of abundant mesoporous structure in the synthesized samples, mainly ascribed to the aggregation of the primary Zn0.5Cu0.5Fe2O4 crystallites. The pore sizes of pure C3N4, Zn0.5Cu0.5Fe2O4 and Z0.5C0.5CN-50 exhibit mainly a distribution from 0–20 nm, 0–12 nm, and 0–12 nm, respectively (in Fig. 5b). Correspondingly, their average pore diameters of the materials are about 29.756 nm, 1.091 nm and 6.337 nm, respectively. This implies that the samples of pure C3N4 and Z0.5C0.5CN-50 are mesoporous materials and that of Zn0.5Cu0.5Fe2O4 is a microporous material.
The surface area of C3N4, Z0.5C0.5CN-50 and Zn0.5Cu0.5Fe2O4 are 78.08, 92.30 and 180.42 m2⋅g− 1, respectively. Zn0.5Cu0.5Fe2O4 presented a larger specific surface area than pure C3N4 and Z0.5C0.5CN-50, which results from the smallest microscopic size of Zn0.5Cu0.5Fe2O4 among these catalysts. Obviously, the specific surface area of Z0.5C0.5CN-50 is larger than that of C3N4. This is attributed to the microscopic crystals of Zn0.5Cu0.5Fe2O4 grown on the surface of g-C3N4, resulting in a higher surface area and abundant active edges. The larger surface area of Z0.5C0.5CN-50 not only facilitates the absorption of incident light but also increases the contact of the reactants, which is beneficial to speed up the degradation reaction.
The Point of Zero Charge (PZC) analysis was applied to surface charge of photocatalyst surface as a function of pH, which in turn to investigate the mode of interaction of EBT onto photocatalyst surface. The pH impact is usually clarified via the pH of point of zero charge (pHpzc) of the photocatalyst. Consequently, at pH values near pHpzc, the interaction between the pollutants and the photocatalysts surface is tiny owing to the absence of significant electrostatic forces. However, when the pH of the solution is lower than pH pzc, the phtocatlyst surface carries positive charge and therefore it applied an electrostatic attraction towards negatively charged compounds. For the general principle, if the pH is lower than the pHpzc, the surface of the photocatalyst become positively charged and therefore it attracts anion compounds. Hence, it could be predicted to high removal of anion compound at lower pH. At pH values near the pHpzc of the photocatalyst, aggregation of the photocatalyst particles is observed resulting in photocatalyst sedimentation. The pHpzc is the pH at which the adsorbent is of a neutral charge on the surface. Figure 6 elucidated the various pH of point zero charge analysis of g-C3N4, ZnxCu1−xFe2O4 and Z0.5Cu0.5CN-50 composite and obtained as 4.94, 6.3 and 7.4, respectively. If pH of the solution is lower than pHpzc of catalyst, the catalyst surface becomes postive and attracts anion dye (EBT) from the aqueous solution.
3.4 XPS, VSM and UV–vis analysis
To elucidate the active sites of Z0.5C0.5CN-50 photocatalyst, the chemical state of elements in this photocatalyst was determinded by XPS. XPS analysis of the Z0.5C0.5CN-50 composite is shown in Fig. 7a. Figure 7b shows the C1s XPS spectrum of the Z0.5C0.5CN-50 sample, which is divided into two peaks. The peaks occur at 284.6 and 287.26 eV, the sp3 hybrid carbon (C)3-N and sp2 hybrid carbon(N–C = N), respectively [48]. The N1s spectrum (Fig. 7c) displays two peaks at 398.05 and 399.63 eV, corresponding to the aromatic tris-triazine ring sp2 hybrid nitrogen (CN = C) and tertiary nitrogen (N-C3), respectively [49], indicating the existence of g-C3N4. In the O1s spectra(Fig. 4d), the wide curve exhibits two matching peaks located at about 529.31 and 531.12 eV, attribute to the lattice oxygen (indexed as Fe–O and Zn(or Cu)–O) and surface absorbed oxygen species (such as H2O), respectively [33, 50]. The Zn 2p spectra of Z0.5C0.5CN-50 composite (Fig. 3d) show two peaks at about 1021.2 and 1044.4 eV, assigned to the Zn 2p3/2 and Zn 2p1/2 of Zn2+ [49, 50], respectively. The Fe 2p XPS spectra peak (Fig. 4e) could be deconvoluted into five fitted peaks at 709.25, 711.41, 717.89, 722.18 and 724.81 eV, respectively. The Fe 2p3/2 peak can be matched into two peaks at 709.25 and 711.41 eV, which are assigned to the Fe3+ in octahedral and tetrahedral sites. The peak at 724.81 eV is attributed to Fe 2p1/2 of spinel Zn0.5Cu0.5Fe2O4 [50]. The peak at 717.89 eV is regarded as the shakeup satellite structure [51]. The peak at 722.18 eV was attributed to the binding energies of 2p1/2 of Fe2+ and Fe3+, namely, Fe2+ and Fe3+ exist simultaneously in the composite [52]. It was reported that the co-existing of Fe2+ and Fe3+ in the spinel ZnxCu1−xFe2O4 is favor for the activation of persulfate during degradation process [53]. Figure 7f shows the XPS spectrum of Cu2p in the Z0.5C0.5CN-50 sample. The three peaks detected at 959.51 eV, 940.91 eV and 937.78 eV correspond to the satellite peaks of Cu [54, 55]. With the splitting of the 2p3/2 and 2p1/2 spin orbitals of Cu, two peaks lacated at 952.47 and 933.28 eVare assigned to Cu(II) and the other two peaks (951.08 and 931.42 eV) correspond to Cu(I) [56, 57].
A magnetic solid catalyst apllied in the water treatment process can resolve the solid-liquid separation problem after reaction. The magnetic properties of the synthesized catalysts were determined by VSM measurement. As seen form Fig. 7h. The observed magnetic parameters of the Z0.5C0.5CN-50 such as relatively high saturation magnetization (Ms = 15.0 emu⋅g− 1) and low remnant magnetization and coercivity indicated that the Z0.5C0.5CN-50 composites exhibit soft magnetic behavior and can be easily separated by an external magnetic field. The lower Ms value in the Z0.5C0.5CN-50 hybrid materials is probably owing to the strong counteraction between the Zn0.5Cu0.5Fe2O4 composites and g-C3N4.
The photocatalytic activity is affected by the optical absorption performance. The optical absorption characteristics of the ZnxCu1−xFe2O4, g-C3N4 and Z0.5C0.5CN were investigated by UV-vis DRS and the band gap of the prepared materials was evaluated from the graph of (αhν)2 versus hv (Tauc plot) for the absorption coefficient α that is related to the band gap Eg as (αhv)2 = k(hv − Eg), where hν is the incident light energy and k is a constant. Figure 8 diplays the UV–vis spectra of the prepared catalysts and the curves of (αhv)2 versus photon energy(hv). It can be seen that the band gap absorption edge of ZnxCu1−xFe2O4 (Fig. 8a) is in the range of 700–800 nm with an energy band gap of 1.54–1.79 eV. The band gap value increases in the order: Zn0.7Cu0.3Fe2O4 (1.79eV) > Zn0.6Cu0.4Fe2O4 (1.76 eV) > Zn0.5Cu0.5Fe2O4 (1.69 eV) > Zn0.4Cu0.6Fe2O4 (1.66 eV) > Zn0.3Cu0.7Fe2O4(1.54 eV). This result is similar to the conclusion by Huang [59]. As shown in Fig. 8b, the absorption edge of bare g-C3N4 is at 458.5 nm with a estimated energy band gap of 2.7 eV, signifying its visible-light- induced photocatalytic activity mainly caused by the π-π* transition[58]. Upon the formation of heterojunctions, the Z0.5C0.5CN heterojunctions possesses increased optical absorption especially in the visible region, the absorption wavelength of all the Z0.5C0.5CN hybrid composites shows red-shift, indicating that they possess better visible light absorption and visible light catalytic capacity. As the amount of Zn0.5Cu0.5Fe2O4 increases in the hybrid composites, the energy band gaps of the hybrid composites become smaller. Among the composite materials, the Z0.5C0.5CN-50 with an energy band gap 2.45 eV is preferably activated by visible radiation in sunlight to produce more carriers (electron-hole pairs).
3.5 DFT simulation
To explain the physical properties of Zn0.5Cu0.5Fe2O4 and g-C3N4, as well as to demonstrate their heterojunction structure, theoretical calculations were performed on them based on DFT, the result is shown in Fig. 9. Figure 9a shows the optimized g-C3N4 structure, from which it can be seen that there are three N atoms near each C atom, labeled N1 and N2, where N1 is connected to two C atoms and N2 is fully connected to three C atoms. Figure 4d shows the optimized structure of Zn0.5Cu0.5Fe2O4, where the Zn and Cu atoms are in the 8a (1/8, 1/8, 1/8) tetrahedral sites, the Fe atoms are in the 16d (1/2, 1/2, 1/2) octahedral sites and the O atoms are in the 32e (u, u, u) face-centered cubic structure. Figure 8b shows the energy band structure of g-C3N4, Fig. 9e-f show the energy band structure of Zn0.5Cu0.5Fe2O4, Fig. 8c and Fig. 8g show the DOS diagrams of g-C3N4 and Zn0.5Cu0.5Fe2O4, respectively. The dashed line in the diagram shows the Fermi energy level (0 eV), the upper side of the dotted line is the conduction band (CB) and the lower side is the valence band (VB). It can be observed that the valence band and conduction band of g-C3N4 are located at the top and the bottom of the Fermi energy level, respectively, and it can be seen that the valence band top (VBM) is closer to the Fermi energy level, which proves that g-C3N4 is a typical indirect band gap semiconductor with the conduction band bottom (CBM) at the K point and the VBM at the G point. Notably, the calculation band gap value of 1.27 eV for C3N4 is lower than the measured band gap value of 2.70 eV, it can be explained that the PBE calculation method underestimates the actual band gap value of the semiconductor [60], however, it is consistent with the information presented in the DOS diagram of g-C3N4. The energy band diagram for Zn0.5Cu0.5Fe2O4 has bands that intersect the Fermi energy level, it is indicated that is of distinctly metallic feature. It also exhibits clear spin-polarization characteristics, being a semiconductor with magnetic properties, distinguishing between spin-up and spin-down, while the asymmetry between the spin-up and spin-down parts can be observed from its DOS diagram, and the spin-polarized electrons are large near the Fermi energy level, providing a theoretical basis for photo-induced charge separation [61]. It is also worth mentioning that the semiconducting nature of the compound is again clearly confirmed by the multi-conducting and less-conducting spins of the compound. As can be seen from the energy band diagrams of both compounds, the conduction and valence bands of Zn0.5Cu0.5Fe2O4 and g-C3N4 are interleaved, indicating that the complexes have a heterojunction structure.
3.5 The catalytic activity
The experiments of catalyst catalytic activity on the photodegradation of EBT solution were performed at the conditions of initial concentration of EBT 50 mg/L, initial concentration of SPS 1.00 mmol/L, solution initial pH 6.0, dosage of catalyst 0.40 g/L, 35 W tungsten lamp light. The results are shown in Fig. 10.
As shown in Fig. 10a, more than 63.0% of EBT was degraded in 120 min by Z0.5C0.5CN with synergistic effect of SPS under visible light irradiation at room temperature. Direct photolysis of EBT by visible light was negligible without catalysts. Interestingly, the degradation rates of EBT by photocatalysis of Z0.5C0.5CN are higher those rates of Zn0.5Cu0.5Fe2O4 and g-C3N4. Figure 10b displays the photodegradation kinetic curve obtained from the dates in Fig. 10a, which is fitted with the pseudofirst order rate kinetics equation, and the homologous rate constant (k) is listed in Table 2. Based on the blank experiment, the self-degradation of EBT can be ignored without catalyst in presence of SPS. The degradation rates in Table 2 showed that all the hybrid composites exhibited superior photocatalytic activity compared to single-phased g-C3N4 and Zn0.5Cu0.5Fe2O4, indicating the in situ generated heterojunction in significantly enhancing the photocatalytic efficiency of the composite materials. Among the three Z0.5C0.5CN hybrid composites, the highest EBT removal rate was shown by the Z0.5C0.5CN-50 sample, equally to 63.4%, and the apparent first-order reaction rate constant (kapp) was 0.0061min− 1 and 2.91and 1.96 times to that of Zn0.5Cu0.5Fe2O4 and g-C3N4, respectively. The consequence ideal heterojunction structure is the main reasons for the superior photocatalytic performance.
In addition, the effect of different ratios of Zn to Cu in ZxC1−x CN-50 on EBTdegradation was studied (Fig. 10c). When the dosage of different catalysts was fixed, the degradation was proportional to an increase of Zn: Cu in the range of 0.2 − 0.7. The Z0.5C0.5CN-50 composite exhibited the highest photocatalytic ability for SPS and had the large amount of super paramagnetic fraction, indicating that the super paramagnetic fraction in Z5C0.5CN-50 seems to play a major role in the activation of SPS. However, Z5C0.5CN-50 possessed both a strong magnetic property and a high amount of super paramagnetic fraction, and thus it might be a better choice for the activation of SPS owing to its magnetically removable and photocatalytic capacity.
Table 2
Band gap, EBT removal rate, and rate constant kapp
Sample
|
Bandgap (eV)
|
EBT removal rate (%)
|
kapp (min –1 )
|
Zn0.5Cu0.5Fe2O4
|
1.69
|
38.4
|
0.0021
|
ZCCN-30
|
2.47
|
58.4
|
0.0056
|
ZCCN-50
|
2.45
|
63.4
|
0.0061
|
ZCCN-30-70
|
2.41
|
51.7
|
0.0035
|
C3N4
|
2.70
|
47.0
|
0.0031
|
3.6 Effect of operating parameters on the degradation of EBT
The effects of catalyst loading, initial concentration of persulfate, initial pH and initial concentration of EBT on the degradation of EBT were investigated separately, the results are shown in Fig. 11.
From Fig. 11a, the degradation efficiency improves steadily as the catalyst dosage increases from 0.40 to 0.80 g/L, meanwhile, the observed kinetic rate constant increased rapidly with the rise of catalyst loading from 0.40 to 0.80 g⋅L− 1 before slowing (the rate of change slowed at higher concentrations). In the range of 0.40–0.80 mg⋅L− 1, the rise in amount of catalyst can provide more surface area and available active sites, resulting in higher degradation of EBT. However, in the range of 1.00–1.20 g⋅L− 1, the excessive catalyst not only blocked the diffusion of SPS and EBT to the solid surface of catalysts, but also restrained the penetration of visible light, which weakens light penetration and harvest in the reactive system. Hence, the desired catalyst dosage for the following experiment is 0.80 g/L.
The effect of SPS adding amount on degradation efficiency is shown in Fig. 11b. The degradation efficiencies gradually enhance as the SPS amount increases, most likely because the higher SPS amount may decompose to generate more free radical (for example, ⋅O2− or ⋅SO4−)[57]. When the addition concentration is increased from 1.00 to 1.50 mmol/L, however, the degradation efficiency does not significantly improve. Under the current conditions, the ideal SPS addition amount is 1.00 mmol/L.
The initial pH of reaction solutions is a crucial parameter for the degradation of EBT due to its influence on the active sites of catalyst. Initial pH values of 4.0, 6.0, 8.0, 10.0 and 12.0 were chosen in the visible light/Z0.5C0.5CN-50/SPS system. Firstly, the surface charge of catalysts is subject to the pH due to the protonation and deprotonation of surface –OH groups. The surface–OH groups were characterized by the zeta potential at varied pH values. According to the performed point of zero charge analysis, the pHpzc of Z0.5C0.5CN-50 was found to be 6.3, indicating that the surface charge is positive at pH < 6.3 and negative when pH > 6.3. The impact of on adsorption and degradation by Z0.5C0.5CN-50 was investigated by varying the solutin pH within the range of 4–12. As observed in Fig. 11c, the EBT removal (%) was reduced with a rise in solution pH. Firstly, the EBT adsorption equilibrium was achieved after 30 min the adsorption process. At acidic pH ranging from 4 to 6, a high removal was observed owing to the electrical attraction between anions (such as EBT anion or persulfate ion) and the postively charged sites Z0.5C0.5CN-50, resulting in generating more free radical (such as and) and therefore in obtaining high degradation rate with over 98.89%. Upon increasing pH (8.0–12.0), the EBT adsorption removal decreased, due to the increase in the number of negatively charged sites, which give rise to the existence of an electrostatic force of repulsion between anions (such as EBT anion or persulfate ion) and the negatively charged sites Z0.5C0.5CN-50, and while the degradation rate is lower than that rate at acidic pH.
Figure 11c shows the effect of initial concentration of EBT on the degradation efficiency of occurred Z0.5C0.5CN-50 composite. In all concentration cases, an initially high rate of adsorption arose because the EBT concentration provided the driving force for the rapid attachment of EBT onto the adsorbent surface. As the adsorption proceed, the ratio of EBT molecules available adsorption sites decreased, which led to a decrease in the adsorption rate until adsorption equilibrium achieved. The concentration of EBT is a significant factor that affects the efficiency and the kinetics of degradation. For general principle, at low concentration of contaminants, the rate of degradation increses at low substrate concentration. However, beyond the optimal concentration, the removal rate decreases owing to the inadequate number of reactive radicals [60]. The adsorption of EBT molecules on catalyst (Z0.5C0.5CN-50) was investigated by varying the EBT concentration (25.0–75.0 mg/L). As seen from the Fig. 11d, the maxium degradation eficiency of EBT was found to 100% at 25 mg/L of EBT concentration. When the concentration of EBT increases, more reactant molecules are adsorbed on the surface of Z0.5C0.5CN-50 and as a result reducing the generation of radicals (such as ⋅O2− and ⋅SO4−) since there are fewer aitive sites for the adsorption of S2O82− anions.
In addition, at higher MB concentrations, the photons are absorbed by the contaminants before they can reach the catalyst surface. Therefore, the light absorption by the Z0.5C0.5CN-50 surface decreases, which led to lower the photocatalytic efficiency. The high concentration of substrate may also result in the deactivation of photocatalyst.
3.7 Recycling and stability of catalysts
The reusability and chemical stability of as-prepared Z0.5C0.5CN -50 were examined for its application in wastewater treatment. As seen from Fig. 12a, Z0.5C0.5CN-50 still retained good catalytic activity with EBT removal of 88.04% after the fourth repeat, indicating its high stability in the reaction system. The concentrations of leached metal ions, such as Zn2+, Cu2+, and Fe3+, from Z0.5C0.5CN−50 were measured during photocatalytic process at the optimum operation parameters as well (i.e., initial pH = 6, loading of catalyst = 0.80 g⋅L− 1, and initial concentration of SPS = 1.00 mmol/L), and the results were shown in Fig. 12b. As seen from Fig. 12b, the concentrations of Zn2+, Cu2+ and Fe 3+ were as low as 0.08, 0.07 and 0.91 mg⋅L− 1, respectively. The permissible limits for Zn 2+, Cu 2+ and Fe3+ in reuse water for long-term irrigation are 2.0, 0.2, and 5.0 mg⋅L− 1, respectively [62], indicating the safety of the proposed treatment process.
3.8 Photocatalytic mechanism
To ascertain the active species responsible for EBT photodegradation reaction, the holes and free radicals scavenging experiments were performed on the Z0.5C0.5CN-50 composite. The primary active species were trapped using scavengers including sodium oxalate as hole (h+) scavenger, isopropyl alcohol as ⋅OH radical scavenger, methanol as sulfate free radical negative ion (‧SO4−) radical scavenger, p-benzoquinone as superoxide radical (O2-) scavenger, potassium dichromate as electron (e-) scavenger. The results from the scavenging experiments are given in Fig. 13. After the addition of sodium oxalate as h+ scavenger to the reaction system, the photocatalytic activity of Z0.5C0.5CN-50 towards EBT dye was drastically inhibited down to 25.1%. Whereas when methanol was added to the reaction system as ⋅SO4- radical scavenger, the EBT degradation was gone down to 51.32% as compared to the degradation efficiency of the composite without scavenger, this showed that ⋅SO4- radical play a considerable role in the EBT degradation. In a similar fashion, EBT degradation percentage was highly decreased to 52.63% after the addition of p-benzoquinone to the photocatalytic reaction system, demonstrating that the superoxide (O2-) radicals also have an important role as the active species in the EBT degradation system. Similarly, EBT degradation percentage was reduced to 55.26% after the addition of potassium dichromate to the photocatalytic reaction system, indicating that the e- also has a significant role as the active species in the EBT degradation system. However here's the difference that EBT degradation percentage was slightly reduced by 96.47% after the addition of isopropyl alcohol to the photocatalytic reaction system, it is indicated that the hydroxide radicals almost never acted as the active species in the EBT degradation system. From these scavenging experiments, it was revealed that the photogenerated holes and electrons, superoxide radicals and ⋅SO4- radicals were all the active species in the EBT photocatalytic degradation reaction. Figure 13 clearly reveals that the photodegradation rate of EBT is remarkably inhibited after the addition of sodium oxalate, methanol, p-benzoquinone and potassium dichromate compared with no scavenger and the inhibition performance follows the order: sodium oxalate > methanol > p-benzoquinone > potassium dichromate. Thus, it can draw a reasonable conclusion that h+, ⋅O2− and ‧SO4− as oxidation species were indeed photogenerated on catalyst surfaces and are responsible for the photocatalytic degradation. In addition, photogenerated holes and electrons are more significant than the other radicals for EBT degradation, suggesting the importance of photocatalysis induced by Z0.5C0.5CN.
To understand the electrons-holes separation mechanism, the valence band (VB) and conduction band (CB) potentials of as-prepared g-C3N4 and Zn0.5Cu0.5Fe2O4 were determined. The VB and CB edge potentials of a semiconductor were calculated using Mulliken electronegativity equation[63].
E CB = X −Ee −0.5E g (1)
where ECB is the CB edge potential; EVB is the VB edge potential; X is the semiconductor electronegativity; Ee is the energy of free electrons on the hydrogen scale (about 4.5 eV); Eg is the semiconductor band gap energy. X is the electronegativity of the prepared semiconductor, which is calculated using the following formula:
X =[x(A) a x(B) b x(C) c]1/(a+b+c) (2)
where x is the electronegativity of each individual element, A, B, and C are the elements, a, b, and c are the number of atoms in the compounds. The obtained values Eg from Fig. 8b are used to calculate the energy potential of VB (EVB ) according to the following equation:
EVB = ECB +E g (3)
For Zn0.5Cu0.5Fe2O4 the calculated VB and CB edge positions were 2.22 and 0.51 eV vs. NHE, respectively. While HOMO and LUMO edge positions for g-C3N4 nanosheets were 1.57 and − 1.13 eV vs. NHE, respectively. According to the calculations above, before the reaction Zn0.5Cu0.5Fe2O4 NPs and g-C3N 4, when solar light strikes the semiconductors, the electrons (e−) in the VB will be excited to the CB leaving holes (h+) in the VB, this is in the case of each sample individually, the e−/h+ recombination occurs, that is the major drawback, though, As combining these semiconductors, a new suggested mechanism might have occurred, which explains the photodegradation of EBT based on separating the charge carriers (e−/h+) as listed in Table 3.
Table 3
Electronegativity (X), band gap (Eg), and conduction band (ECB) position and valance band (EVB) of the photocatalysts
Parameters
|
Photocatalysts
|
g-C3N4
|
Zn0.5Cu0.5Fe2O4
|
Eg (eV)
|
2.70
|
1.69
|
X (eV)
|
4.730
|
5.862
|
CB(eV)
|
-1.12
|
+ 0.51
|
VB(eV)
|
+ 1.58
|
+ 2.22
|
The proposed schematic mechanism of photocatalytic degradation of EBT on the Z0.5C0.5CN-50 composite was shown in Fig. 14. The early literature [64] have reported that persulfate can be decomposed to produce oxygen under optical radiation (Eq. 4). Photogeneration electron-hole (e-_h+) pairs are generated by both Zn0.5Cu0.5Fe2O4 and g-C3N4 of the Zn0.5Cu0.5Fe2O4-loaded g-C3N4 after irradiation by sufficient photon energy. The ECB of Zn0.5Cu0.5Fe2O4 (+ 0.513 eV) is less negative than that of the g-C3N4 (− 1.12 eV), and the EVB of Zn0.5Cu0.5Fe2O4 (+ 2.22 eV) is more positive than the E VB of g-C3 N4 (+ 1.58 eV). In the case of type II heterojunction phase way, the e- in the ECB of g-C3N4 is migrated to the ECB of Zn0.5Cu0.5Fe2O4 and the h+ in the EVB of Zn0.5Cu0.5Fe2O4 are migrated to EVB of the g-C3N4. These interfacial e–h+ transfers inhibit the charge carrier recombination (Eq. 5) and prolong the lifetime of separated electron − hole pairs, resulting inthe improvement of photocatalytic degradation efficienc. The photogenerated electrons in CB of g-C3N4 was negative enough to reduce O2 and S2O82-on the surface of photocatalysts to generate ⋅O2− (O2/⋅O2− = −0.33eV) [65] and ⋅SO4- (S2O82-/⋅SO4- = −0.50eV) [66–68] for the degradation of organic compounds, This satisfied the thermodynamic requirements for the productions of ‧O2− and ‧SO4 − radicals via the activation of O2 and S2O8 2- (Eqs. (6) -(7)), respectively. While the VB holes of g-C3N4 can directly oxidize EBT. Moreover, the VB potential of g-C3N4 (+ 1.495 eV) is not positive enough compared with the standard reduction potential of ⋅OH/H2O (2.27 eV) and ⋅SO4-(2.6 eV), indicating that the most of h+ on the surface of g-C3N4 cannot oxidize H2O or OH− into ⋅OH radicals[62,68] and S2O82- into ⋅SO4-radicals [66].
In addition, previous literatures [69–71] have reported that Fe(Ⅱ) and Cu (I) ions in the surface latice of spinel ferrite (such as Zn0.5Cu0.5Fe2O4, CuFe2O4) could also activate persulfate to generate generate sulfate (⋅SO4- ) radicals ( Eqs. (8)–(9) ). The standard redox potential of S2O82-/⋅SO4-, Cu(II)/Cu(I) and Fe(Ⅲ)/Fe(Ⅱ) ions are − 0.50, − 0.33, 0.16 and 0.77 V vs. normal hydrogen electrode (NHE), respectively. Thus, the formed Fe(Ⅲ) and Cu(I) can be turned into Fe(II) and Cu(II) by their redox reaction through the catalytic recycles ( Eq. (10) ). In summary, ⋅SO42-, ⋅O2− and h + are the main oxidative matters that are participated in the decomposition of EBT ( Eq. (11) ).