CoZnFeO4 prepared by waste ferrous sulfate as iron source: synthesis, characterization and photocatalytic degradation of methylene blue

In this study, a ferrate material with a narrow band gap and rapid electron hole separation ability is synthesized for degrading methylene blue from industrial wastewater. Hard ferrite (CoZnFeO4) with octahedral structure is synthesized by solid-phase method from waste ferrous sulfate. The structure of the catalyst was characterized by X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTPR) and scanning electron microscopy (SEM). TPR study showed that the oxidation performance of CoZnFeO4 is improved with the addition of cobalt. The total organic carbon removal rate of methylene blue over CoZnFeO4 catalyst is 90% within 60 min. The effects of reaction time, pollutant concentration, catalyst dosage and oxidant concentration on the removal efficiency were optimized. ESI–MS demonstrated the stability of the catalyst for leaching. X-ray diffractometry (XRD) and X-ray energy dispersive spectroscopy (XPS) showed that the catalyst could be reused. These findings provide a low cost and simple strategy for rational design and modulation of catalysts for the industrial degradation of organic pollutants. It not only realizes the use of waste to treat waste, but also accords with the current concept of green chemistry.


Introduction
At present, the global annual output of dyestuff exceeds 7 million tons, and the dyestuff varieties of dyestuff have more than 100,000 kinds. Nearly 15% of dyestuff is discharged into wastewater every year.
Organic dyestuff wastewater has become an urgent problem to be solved in all countries [1]. Most organic dyes are in the form of aromatic groups such as benzene, naphthalene, anthracene, and quinone, which are highly mutagenic and carcinogenic and pose a growing threat to the water environment [2,3]. Among them, methylene blue (MB) is one of the most common dye, but it is difficult to be degraded by some conventional methods, easy to cause serious environmental pollution.
At present, the treatment of organic dye wastewater including MB dye wastewater mainly includes advanced oxidation method [4], precipitation, ion exchange, flocculation, a biological method [5], and membrane filtration [6]. In the advanced oxidation method, photocatalysis is a cheap and efficient method, which efficiently convert solar energy into chemical energy, which is one of the most ideal strategies to solve environmental pollution [7]. In 1972, the report of using TiO 2 electrodes to achieve photochemical decomposition of water to produce H 2 aroused the interest and exploration of a large number of scholars [8]. Since then, a series of systems (such as TiO 2 , ZnO, ZnS) semiconductor photocatalysts have been reported to decolorize organic pollutants and harmful microorganisms. However, due to the large bandgap width of these semiconductor photocatalysts, they can only use the ultraviolet region (only 4% of the solar energy), which seriously hinders the full utilization of solar energy resources. So, to make more efficient use of the visible light region ([ 400 nm), it's important to use a visiblelight-induced photocatalyst that improves recycling efficiency.
As a kind of important magnetic material, cubic spinel ferrites have been the subject of intense research for potential application in the recovery of photocatalysts, magnetic resonance imaging, and high-density storage, magnetic diagnosis [9] and magnetic fluid hyperthermia treatment for malignant growth [10,11]. Among them, nanometer ZnFe 2 O 4 is the most studied. Because of its nanoscale effect, ZnFe 2 O 4 has remarkable super-para-magnetism, electrical properties, high stability, paramagnetism and biomedical properties [12]. Zn 2? is located in the center of the tetrahedron gap and is covalently connected with O 2-, Fe 2? is located in the center of the octahedron gap and is covalently connected with O 2-. Because there are a large number of covalently linked in the tetrahedron, nano ferric acid has stable physical and chemical properties. Spinel zinc ferrite has a narrow bandgap (2.11 eV), high sensitivity to visible light, photochemical stability, and low toxicity, and is expected to be a visible light catalytic catalyst. However, due to its wide bandgap and poor electron and hole separation ability, its photocatalytic activity is still a difficult problem. Therefore, to improve the ability of charge transfer, it is necessary to modify pure zinc ferrite materials. However, the catalytic activity of spinel ferrite depends entirely on octahedral position, so its catalytic activity is directly related to the cation on octahedron. Studies have shown that cobalt ions can occupy the position of ferrite octahedron. Sanadi et al. doped Co with ZnAlCrO 4 for photocatalytic decolorization of Rhodamine B (Rhb), and the results showed that the decolorization rate of Rhb increased from 57 to 83% after doping Co, indicating that cobalt ions can improve charge transfer, which is of great significance for photocatalysis [8].
Waste ferrous sulfate is an industrial by-product produced during the production of titanium dioxide by a sulphuric acid method. In 2018, the titanium dioxide industry produced more than 7 million tons of ferrous sulfate as a by-product [13]. With the increasing demand for titanium dioxide, the annual growth rate of by-products is more than 10%, which seriously restricts the sustainable development of the titanium dioxide industry. At present, only a small part of waste ferrous sulfate is recycled and the rest is still disposed of as industrial waste solid [14,15]. Therefore, the accumulation of waste ferrous sulfate not only causes a huge waste of sulfur and iron resources but also causes serious environmental pollution. It has become an urgent task to explore the resource utilization of waste ferrous sulfate.
In this study, CoZnFeO 4 is prepared by a one-step solid-phase method with low cost, simple process, simple operation, and easy industrialization. To reduce the synthesis cost, CoZnFeO 4 nanoparticles are prepared with waste ferrous sulfate as an iron source and pyrite as a reducing agent, which are used as photocatalyst for decolorizing MB.

Preparation of CoZnFeO 4 and ZnFe 2 O 4
CoZnFeO 4 is synthesized form using ferrous sulfate as the iron source and pyrite as the reducing agent, by solid-phase reduction under the condition of nitrogen. The reaction of synthetic CoZnFeO 4 is as follows: The experimental process of preparing

Degradation experiment of methylene blue
The decolorization rates of MB at different concentrations are studied, the effect of different nanomaterials on decolorizing Photo-Fenton reaction. A certain amount of CoZnFeO 4 nanoparticles are dispersed into MB solution as a catalyst, and then the given pH value of MB solution is adjusted using an automatic pH titrator (4 mol/L NaOH or 4 mol/L HCl). The dark chamber stirred for 1 h to achieve the balance of absorption and desorption. Then, the 5 mL of H 2 O 2 at a certain concentration is added into the above-mixed solution and reacts under a certain wavelength for the light-ton reaction. The decolorization effect of MB under different reaction times is obtained under the parallel apparatuses. After taking out the parallel reaction solutions with different reaction times, 3.33 mL tert-butanol solution is added as the reaction inhibitor, the nanocatalyst is removed by centrifugation, and the concentration of MB in the supernatant is determined by spectrophotometer.

Characterization method
The crystal structure of solid powder is analyzed by X-ray diffraction (XRD, Empyrean, PANalytical). Scanning electron microscope (SEM, JSM-5900LV) and X-ray microregion analysis (EDS, JSM-5900LV) is used to analyze the sample morphology and element composition. X-ray photoelectron spectroscopy (XPS, Kratos, XSAM800) is used to analyze the surface composition of the samples. The nitrogen adsorption technique (BET, Brunner-Emmet-Teller) is used to determine the specific surface area of the sample. The surface properties of the samples are recorded by  3 Results and discussion  [16], the average grain size of the synthesized CMFNPs is estimated to be 27 nm. The infrared spectrum of the nanoparticle is shown in Fig. 1b broadband at 3445 cm -1 and, 1632 cm -1 can be considered as the stretching vibration of -OH. Besides, the absorption peak of 1376 cm -1 also indicates the presence of a large amount of -OH. Moreover, the absorption peaks of 1074 cm -1 indicate the existence H 2 O deformation vibration. The strong characteristic bands observed at 572 cm -1 can be attributed to the stretching vibration of M 3? -O 2octahedron of typical spinel ferrite, it also indicates that CoZnFeO 4 may be produced. The surface composition and chemical state of the molded sample are determined by X-ray photoelectric spectrometry, as shown in Fig. 2. The peaks of Fe, Co, Zn, and O are observed in Fig. 2a. In Fig. 2b, the two main peaks of the combined energy 780.3 eV and 795.2 eV belong to the core energy levels of Co 2p 3/2 and Co 2p 1/2 [7,17], this indicates the existence of Co 3? in nanomaterials. In Fig. 2c, the peaks of the binding energies of 711.2 eV and 724.63 eV belong to the core energy levels of Fe 2p 3/2 and Fe 2p 1/2 [18,19], this is typical of Fe(III) in inverse spinel ferrite. The synthesis of CoZnFeO 4 is demonstrated by XRD and FTIR. In Fig. 2d, O1s spectrum is divided into absorption peaks of 530.15 eV, 531.69 eV, and 532.64 eV, corresponding to lattice oxygen (O 2-) in metal oxides, the hydroxyl group (-OH) is hydroxylated on the surface and physically adsorbed the water on the sample surface [20].

Characterization of nanomaterials
To understand the morphological characteristics of nanomaterials, FE-SEM, and EDS analysis are performed on the prepared materials (as shown in Fig. 3). In Fig. 3a, most of the CoZnFeO 4 Fig. 1 a The XRD patterns, b FTIR spectrum of copper magnesium ferrite nanoparticles nanoparticles are uniform with a narrow particle size distribution, and the average particle size is a spherical particle of 30 nm. In addition, there is a certain agglomeration of particles from the figure, which may be due to electrostatic attraction and van der Waals force, some primary particles tend to aggregate in local areas. In Fig. 3b-f, the presence of Fe, Co, O, and Zn by EDS testing indicates that the synthesized substances are made of Fe, Co, O, and Zn. The elemental atlas confirmed that the distribution of Zn, Co, Fe, and O is relatively uniform, which is beneficial to enhance the catalytic activity.
The adsorption-desorption isotherms of the samples are tested under liquid nitrogen conditions. BET equation is used to calculate specific surface area and the BJH method is used to calculate the aperture from the desorption branch of isotherm. Figure 4a shows the N 2 adsorption-desorption isotherms of the prepared CoZnFeO 4 and ZnFe 2 O 4 nanomaterials. According to the classification of IUPAC, both the CoZnFeO 4 spinel oxide and the N 2 adsorption-desorption isotherms of ZnFe 2 O 4 studied are type IV, indicating the mesoporous structure of the prepared materials [21]. Also, accompanied by the H1 hysteretic loop, such a strong hysteresis phenomenon is generally believed to be related to capillary condensation of large pore channels [22]. As shown in Fig. 4b, the pore diameter of CoZnFeO 4 nanomaterials is mainly mesoporous, the pore diameter distribution curve is a single peak,  and the pore diameter is mainly centered at 25 nm. The pore diameter distribution of ZnFe 2 O 4 nanomaterials is mainly concentrated around 8 nm. In addition, as shown in Table 3, the total specific surface area and pore volume of CoZnFeO 4 are 38.416 m 2 /g and 0.283 cm 3 /g, respectively, and the total specific surface area and pore volume of ZnFe 2 O 4 are 19.651 m 2 /g and 0.066 cm 3 /g, respectively. The surface area of the doped zinc ferrite depends on the properties of the dopant. The particle size and specific surface area of the nano-zinc ferrite particles change significantly with the substitution of cobalt. The total specific surface area ratio of CoZnFeO 4 is twice that of ZnFe 2 O 4 , which may be caused by the interaction between Co 3? and other metal ions (electrostatic effect, ion motion radius). Another reason is that the cobalt atom radius is smaller than the iron atom radius. As the iron atom is replaced, the crystal size decreases, increasing its surface area [8]. Through comparison, it is found that the surface and pore diameter of ZnFe 2 O 4 doped with Co 3? are significantly larger, which can greatly increase the light-receiving area of the catalyst, reduce the diffusion distance of photogenic carriers, and improve the absorption efficiency of the material to light [23].
To compare the redox activity of CoZnFeO 4 and ZnFe 2 O 4 , the redox activity is analyzed by the H 2 -TPR test. As shown in Fig. 5, there is only one peak of ZnFe 2 O 4 within the range of 500-700°C, which is the peak of Fe 3? reduced to Fe 2? . In contrast, there are three peaks in CoZnFeO 4 , which are Co 3? reduced to Co 2? at 350-450°C, Fe 3? reduced to Fe 3?/2? intermediate state at 450-600°C, and Fe 3?/2? converted to Fe 2? at 600-800°C [19]. Multiple reduction peaks appeared in CoZnFeO 4 , and the initial peak of the reduction peak shifted to a lower temperature. This may be due to the addition of cobalt, so that when Co 3? /Co 2? is reduced in the same range, Fe 3? /Fe 2? is also changed, leading to the interference of the reduction curve. As is known to all, reduction reactions first occur on the surface of materials, and the reduction peak moves to the low-temperature zone, which not only proves the surface occupancy of cobalt ions but also indicates that CoZnFeO 4 has a strong reduction performance.

Effects of several parameters on MB decolorization
3.2.1 Study on the decolorizing methylene blue activity of CoZnFeO 4 nanomaterials Figure 6 shows the effect of magnetic CoZnFeO 4 nanomaterials synthesized by the solid-phase

Influence of pH
pH is an important factor in catalytic decolorization because it determines the surface charge properties of the catalyst and the size of the aggregates it forms, which affects the formation and reaction mechanism of free radicals [27]. Figure 8a analyzes the influence of pH value on the photocatalytic decolorization rate. As shown in Fig. 8a, the decolorization rate of MB gradually decreases when pH is from 2 to 10, indicating that pH has a great influence on the degradation of MB.
If catalysis is considered to follow a first-order kinetic equation, the decomposition rate of MB can be calculated using the following equation: where, C 0 is the initial concentration of MB, and C is the concentration of MB at any time. As shown in Fig. 8b, slope constant k of the ratio of ln(C 0 /C) to time under different pH (2, 4, 6, 8, 10) and light illumination conditions. With the increase of pH, the catalytic activity of CoZnFeO 4 nanomaterials decreased gradually, leading to a gradual decrease in  Table 4, the rate constant k at pH 2 is 1.7078 min -1 , which is about ten times faster than the rate constant K at pH 10, which is 0.17147 min -1 . This indicates that the catalytic rate of CoZnFeO 4 is fast under acidic conditions and that the catalyst can still maintain high catalytic activity in a wide pH range, which is of great value for industrial applications. pH value affects the efficiency of the decolorization process and is related to the generation of positive pore, surface coverage, adsorption behavior, and ion leaching rate [28]. To explain the cause of MB decolorization, 0.1 g CoZnFeO 4 catalyst is dissolved in 50 mL MB solution, and pH is adjusted. Finally, the surface electromotive force (EMF) of the suspension is measured by ultrasonic under illumination for 20 min. As shown in Fig. 9a, under acidic conditions, the surface of the catalyst is positively charged. According to the infrared analysis of Fig. 1b and XPS analysis of Fig. 2d, it can be seen that the nano-materials synthesized by the solid-phase reduction method contain a large number of hydroxyl functional groups, indicating that the variable valence metals mainly exist in Fe-OH and Co-OH. The high concentration of H ? protonates the surface of nanomaterials, forming functional groups with a positive charge (Eqs. 3, 4). However, when pH is above 8.42, a high concentration of OHdeprotonates nanomaterials and forms negative charge functional groups on the surface(Eqs. 5, 6) [29,30]. The reaction is as follows: The relative charge on the surface of nanoparticles is a determining parameter that reflects whether the interaction between particles and the degraded material is repulsive or attractive. According to the DLVO theory, the surface charge properties of nanoparticles are sensitively dependent on their heteroelectron points. When pH is higher or lower than isoelectron (IEP), the surface charge of the particle is positive or negative, respectively. Electrostatic interactions between semiconductor surfaces lead to higher surface coverage, the farther the pH is from IEP, the greater the charge between the particles, the better the particle dispersion [31]. As shown in Fig. 9b, when pH is 2-8.43, the surface charge of nanoparticles changes from 25.93 to -0.36 mV, and IEP is predicted to be 8.3, that is, when pH \ IEP, the surface charge of nanoparticles is positive. And positively charged particles will generate positively charged holes (oxidation holes) [32], which can react with hydroxide radicals to generate hydroxyl radicals, thus oxidizing organic compounds. When pH \ IEP, the surface positive charge of particles gradually decreases with the increase of pH, that is, the attraction of positive charge holes to hydroxyl radicals decreases, and the rate of producing hydroxyl radicals slows down. Therefore, the decolorization rate of MB decreases as pH increases. Although MB has a positive charge, the negative charge on the surface of the particle at pH [ IEP will hinder the generation of oxidized holes, and the decolorization rate depends on the adsorption capacity of MB and the competitive effect of hydroxide on holes. It can be seen from Fig. 8a that the decolorization rate of MB gradually decreases with the increase of pH, indicating that the competition of hydroxide to hole is lower than that of MB, leading to a decrease in the decolorization rate of target substances. When the pH is close to IEP, the particles are severely agglomerated due to the lack of repulsive force, under which case they settle at the fastest rate. In this case, there are fewer particles in the suspension and the oxidation hole is low, so MB has the lowest decoloration rate. However, as shown in Fig. 8a, the decolorization rate is higher than that at pH 10 when pH 8, indicating that the reaction may occur not only on the particle surface but also in the solution. To verify this hypothesis, the concentration of metal ions in the suspension is tested. The content of metal ions in the solution is tested by ICP-MS, as shown in Table 5. Under different pH values (2,4,6,8,10), with the rise of pH, the leaching rate of metal ions gradually decreased, and the decolorization rate of MB gradually decreased, indicating that metal ions in the solution may have a certain role in the decolorization process of MB [33]. In photocatalysis, transition metal ions in solution capture electrons, which are easily transferred to H 2 O 2 to form hydroxyl radicals (Eqs. 7, 8). The hydroxyl radicals produced are very reactive substances that react with most organic matter. At the same time, H 2 O 2 as an electron capture agent can also capture photogenerated electrons (e -) (Eq. 9), thus enabling the separation of photogenerated electrons from photogenerated holes and improving the efficiency of photocatalysis [32]. The specific reaction is as follows: However, although pH 2 has the fastest decolorization rate of MB, it will cause excessive dissolution of metal elements, which may lead to excessive heavy metals in the water. Therefore, in the subsequent experiments, we will try to choose a more appropriate pH for the experiment.

Influence of light wavelength on methylene blue decolorization
Through UV-Visible diffuse reflection absorption spectrum, the optical absorption properties of synthetic materials within the range of UV-Visible spectrum can be explored, and the bandwidth of the material can be calculated, and then the photocatalytic ability of the material can be inferred indirectly. Fig. S1 shows the bandgap width of the catalyst is 2.02 eV, which is consistent with literature reports [34], and the lower the energy required for the electron transition reaction, the easier the photocatalysts, and Lan [35] reported bandgap of ZnFe 2 O 4 is 2.11 eV. This indicates that the doping of Co, which replaces the position of Fe, forms impurity levels in the bandgap and reduces the energy required for electron transition, which is beneficial to the light absorption capacity. In this work, the optimal wavelength of MB decolorization is studied. As shown in Fig. 10, the decolorization efficiency gradually decreases with the increase of wavelength, which is consistent with the UV-Vis test results. The light wavelength reflects the energy intensity of incident light and is directly related to electron transition, which determines the concentration of hydroxyl radical produced. However, due to the presence of an appropriate concentration of H 2 O 2 , it also has a better decolorization effect under dark conditions.

Influence of hydrogen peroxide concentration on MB decolorization rate
The H 2 O 2 concentration is an important factor affecting MB decolorization in photocatalytic reactions. As shown in Fig. 11, the decolorization rate of MB increases, and decolorization time shorten with the increase of H 2 O 2 concentration. When the concentration of H 2 O 2 increases from 20 to 25 mM, the overdosing of H 2 O 2 will decrease the decolorization rate of MB. At a low H 2 O 2 concentration, the improvement of the decolorization rate may be due to the following reasons. Firstly, The H 2 O 2 can produce hydroxyl radical under light conditions due to the main rate improvement mechanism for this process (Eq. 10). Ollis [36] and Ilisz [37] suggest the possible mechanisms in the H 2 O 2 is considered to be a better electron acceptor than oxygen which will reduce the chances of electron-hole pair recombination, which may result in a hydroxyl radical (Eq. 11) rather than a weak radical (Eq. 12) [38].

Influence of CoZnFeO 4 dosage on MB decolorization rate
From the economic point of view, the amount of catalyst is one of the important parameters in decolorization MB. To choose the adding of catalysts, it is necessary to determine the optimal amount of load for the effective removal of MB dyes. The effect of the catalyst load on dye removal rate is investigated, and the results are shown in Fig. 12. As can be seen, the decolorization rate of MB gradually increases with the increase of catalyst dosage from 0.5 to 1.0 g/L. This is due to an increase in the number of active sites, and the more active sites there are, the more opportunities there are to receive light. However, when the amount of catalyst increased from 1.0 to 3.0 g/L, the decolorization rate decreased, which may be because the high concentration of catalyst led to the aggregation of catalyst particles, which led to the reduction of the number of surface-active sites and reduced the catalytic efficiency. Besides, the high concentration of catalyst will increase the opacity of suspension and the scattering of light, thus reducing the number of times that the irradiated light passes through the catalyst. In this study, economic and environmental factors are taken into account to determine the optimal catalyst usage of 1.0 g/L.

Influence of methylene blue concentration on decolorization rate
The effect of MB concentration from 200 to 500 mg/ L on decolorization rate is investigated under the optimal conditions of the dosage CoZnFeO 4 under the given wavelength. As shown in Fig. 13, the time for decoloration of MB is extended from 6 to 12 min with the initial concentration of MB increasing from 200 to 500 mg/L. Photocatalytic decolorization is completely dependent on the generation of hydroxyl radicals and positively charged holes. However, with the increase of MB concentration, the hydroxyl radical yield is decreased, resulting in the decolorization rate decreased. Moreover, with the continuous increase of MB concentration, MB molecules per unit volume increase and occupy more active sites [42,43]. Besides, lots of dye molecules hinder the utilization efficiency of light [44]. Therefore, in photocatalytic decolorization, the decolorization rate decreases with the increase of MB concentration.

CoZnFeO 4 the oxidation of different pollutants
The cyclic degradation performance of CoZnFeO 4 on different pollutants such as ciprofloxacin(CIP), methyl orange(MO) and RhB is shown in figure. Figure 6 showed that the degradation rates of MB, CIP, Mo and RhB within 30 min were 100%, 95%, 74% and 62%, respectively, K are 0.1106, 0.0825, 0.04218 and 0.0309 min -1 , Compared with the other three pollutants, the degradation rate of MB is the highest (Fig. 14).
In addition, compared with the recently reported catalysts, the performance of zinc ferrite is obviously better than that of other catalysts [45,46].

Effect of temperature on decolorization MB
Temperature is one of the important factors of photocatalysis. The effect of temperature (15°C, 25°C, 35°C, 45°C) on the decolorizing MB reaction rate constant of CoZnFeO 4 is studied. As can be seen from Fig. 15a, the reaction rate constant gradually increases with the rise of temperature, indicating that the leaching of CoZnFeO 4 metal ions may be an endothermic process [47].
The higher temperature is favorable for CoZnFeO 4 decolorization MB reaction because the higher temperature can provide energy for reactant molecules to overcome the activation energy barrier [48]. Therefore, the kinetic constants of the generation of free radicals and the regeneration of Fe 3? /Fe 2? and  Co 3? /Co 2? increase with the rise of temperature [49,50]. Miller and Sotolongo established the relationship between the reaction temperature and the decomposition rate constant of H 2 O 2 . The results show that at a certain temperature, the decomposition rate of H 2 O 2 is accelerated with the increase of temperature, thus promoting the generation of HOÁ [51].
Through experiments at different temperatures, the Arrhenius equation (Eq. 16) is used to calculate the activation energy of the reaction [52]: where A is the front factor, E is the apparent activation energy (J/mol), R is the ideal gas constant (8.314 J/mol/K), and T is the reaction temperature (K). The Arrhenius diagram of lnk and 1/T is shown in Fig. 16b, there is a good linear relationship between the natural lnK and 1/T. Activation energy E obtained from Arrhenius in Fig. 15b is 31.102 kJ/mol, and A value is 1.157 9 105 M -1 min -1 . In general,

Comparison of decolorization MB with other catalysts
As shown in Table 6, the decolorization performance of the synthesized CoZnFeO 4 catalyst for MB is compared with other catalysts reported in the literature [53][54][55][56][57][58]. It can be observed that CoZnFeO 4 decolorizes MB within 8 min under normal pH. This indicates that the nanocatalysts prepared from waste materials have better catalytic activity than other catalysts. It also proves that the method of site control has the prospect of large-scale industrialization.

Structure and stability of the catalyst
The application of catalyst in wastewater treatment requires it to be stable for metal ion leaching in the liquid phase under operating conditions. Continuous and progressive leaching not only leads to deactivation of the catalyst but also pollute water further.  Stability is one of the key factors affecting the catalyst application [59]. In this study, CoZnFeO 4 samples are continuously used five times under the same conditions to evaluate the maintenance of their photocatalytic activity. Among them, MB is photocatalysts decolorization, CoZnFeO 4 particles are recovered (such as centrifugation and magnetic adsorption), organic pollutants are removed with deionized water and ethanol, and finally, freezedrying is carried out. As can be seen from Fig. 16a, within 5 cycles, the decolorization rate of MB does not change significantly, and the complete decolorization of MB is completed within 12 min. This indicates that the CoZnFeO 4 catalyst can maintain high catalytic activity after repeated use many times. However, after 5 cycles, the decolorization rate is slightly slowed down, which may be because of the continuous operation. In addition to, the slight inactivation caused by metal leaching, it may also be due to the deposition of carbon compounds in the active part of the catalyst, resulting in the decrease of activity [60]. As shown in Fig. 16b, by comparing the XRD patterns of CoZnFeO 4 before and after the reaction, there is no obvious change. At the same time, it can be seen from Fig. 17a that the FTIR of CoZnFeO 4 catalyst did not change significantly before and after the reaction, which also indicated that the CoZnFeO 4 nanometer material maintained a good reuse performance and a long service life. Besides, as shown in Fig. 16b, compared with other reported results [59], the leaching amount of metal ions in the reaction mixture (g/L) is quite low. Among them, the leaching amount of Co 3? ions is higher than that of Fe 3? , indicating that the interaction between Co 3? ions and reactants is greater than that of Fe 3? , that is, the surface occupancy rate of Co 3? is higher. As shown in Fig. 18, the EDS of CoZnFeO 4 after reaction shows that the metal elements still maintain a relatively uniform distribution, which indicates the stability of the octahedral structure in spinel ferrite. Meanwhile, by comparing the mass fraction of metal ions in Fig. 18f, the leaching rate of Co 3? and Zn 2? after the reaction is higher than that of Fe 3? , which is consistent with the test structure in Fig. 18b, that is, the surface occupancy rate of Co 3? is higher.
It is found that in the presence of BQ, the decolorization effect of MB is not significant, indicating that O ÁÀ 2 is not the main factor determining the photocatalytic decolorization of MB by CoZnFeO 4 . After

species (O ÁÀ
2 , HOÁ) cannot be generated because conductive electrons (e -) cannot be reduced to oxygen [61]. Therefore, the decolorization of dyes is not significantly inhibited under visible light irradiation, that is, eplay a secondary role in this process [62,63]. At the same time, it can be seen in the figure that compared to the ideal state, the presence of N 2 has been decolorized by about 97% MB. This verifies that the whole reaction is not formed, that N 2 is not inhibited in the reaction [64]. Therefore, we conclude that the presence of O 2 in the reaction has less effect on the process.
To further analyze the reaction mechanism of the system, it is realized by photoluminescence characterization. The intensity of PL emission reflects the recombination rate of photogenic electron-hole pairs [65]. In general, lower intensity indicates rapid transfer or annihilation of excited electrons, while higher intensity indicates a high rate of recombination of excited electrons [66]. Figure 20 shows the CoZnFeO 4 and ZnFeO 4 systems excited by 350 nm. It can be seen that the emission intensity of CoZnFeO 4 samples is significantly lower than that of ZnFeO 4 , which indicates that the charge transfer and separation capability of CoZnFeO 4 are relatively strong. Among them, the peaks at 415.09 nm and 435.26 nm correspond to characteristic near-band-Edge (NBE) blue emission, which is mainly attributed to the characteristic NBE blue emission, namely intermittent zinc deficiency, which is caused by the recombination of electrons in oxygen vacancies and photogenic holes generated at tetrahedral and octahedral positions [66][67][68]. The peaks at 627.13 nm and 569.40 nm are due to the transition of Fe 3? from 3d 5 to 3d 4 4s. The conduction band electrons from the 3d 5 state balance the 4S orbital of Fe 3p [69]. Compared with CoZnFeO 4 , the lower PL emission intensity is because doping Co ions can produce a large number of defects, reduce the luminescence intensity of zinc ferrite, and lead to the reduction of photoelectronhole pair recombination rate. In the process of MB photocatalysis, when the visible light energy larger than the energy gap irradiates the catalyst, the electrons move to the conduction band, and the valence band thus produces photogenic holes. Valence band electrons (e -) are excited to the conduction band of the CoZnFeO 4 catalyst. Then the valence band generates photogenic holes (h ? ), oxidizes H 2 O into HOÁ, and forms (-OH) hydroxyl group on the surface of the CoZnFeO 4 nanometer sheet after adsorption. The photogenerated electron has high electron mobility, which can decompose O 2 into ÁOH. The separation of holes and photogenerated electrons leads to a decrease in the recombination rate between them and leads to the migration of more charge carriers to the surface of the CoZnFeO 4 nanosheet [37][38][39]. The unpaired photogenerated electrons in the conduction band can generate hydroxyl radicals with H 2 O 2 , while the holes in the valence band can also react  with H 2 O to generate hydroxyl radicals, and then decolorize MB dyes [70].
To determine the chemical state changes and reaction mechanism of CoZnFeO 4 before and after decolorizing MB, XPS is used to analyze the samples. As shown in Fig. 21a, the peak position before and after the reaction did not deviate significantly, indicating that the CoZnFeO 4 octahedral structure is stable and no lattice distortion occurred. At the same time, there is an extra N spectrum after the reaction, which may be caused by the residue of small molecules broken in MB on CoZnFeO 4 . Figure 21b, c shows the high-resolution XPS spectrum of Fe 2p 3/2 between 705 and 720 eV before and after the catalytic reaction. Peaks of Fe(III) and Fe(II) ions correspond to 712.6 eV and 710.5 eV, respectively [71,72]. After the reaction, the proportion of Fe(II) increased to 32.93% and the proportion of Fe(III) decreased to 62.07%, indicating that part of Fe(III) is converted into Fe(II) ions. Figure 21d, e also shows the XPS spectra of Co 2p 3/2 at 770-810 eV [73]. After the reaction, the proportion of Co(II) increased to 43.63% and decreased to 56.37% compared with Co(III), indicating that part of Co(III) changed to Co(II) [74,75]. In Fig. S2, the values of the conduction band and valence band are calculated theoretically. In Fig. 21f, g the maximum binding energy of Zn 2p 3/2 is about 1022.5 eV [76,77], indicating that the valence of Zn 2? did not change before and after the reaction. These results showed that both Fe 3? /Fe 2? and Co 3? /Co 2? are involved in catalytic reactions, and the amount of cobalt valence transformation is greater than that of iron, indicating that Co 3? /Co 2? have a prominent influence in the whole reaction.
According to the free radical capture experiment and XPS spectral analysis, the decolorization of MB is mainly carried out through the decomposition of H 2 O 2 into hydroxyl radicals and the h ? generated by the catalyst.
Pathway 1: As shown in Fig. 22 [79,80]. And then Co 2? reacts with H 2 O 2 to form Co 3? and HOÁ, to complete the recycling of Co 3? over Co 2? , complete the recycling of Co 3? / Co 2? . Next, HOÁ will react with MB to generate carbon dioxide and water to complete the degradation process [81]. The specific reaction steps are shown below: Pathway 2: When visible light irradiates the semiconductor, the band gap (\ 3.147 eV) is larger than that of the catalyst (2.02 eV). Some of the electrons in the valence band will be excited and jump from the valence band over the forbidden band to the conduction band, thus generating photogenic electrons and h ? , respectively, in the conduction band and the valence band [82]. The conduction band gap (0.3975 eV) of CoZnFeO 4 is not lower than that of CoZnFeO 4 (-0.33 eV); therefore, electrons in the conduction band cannot react with oxygen molecules to form superoxide radicals [59]. On the contrary, the photogenerated electrons in the CoZnFeO 4 conduction band react with H 2 O 2 to produce more, which leads to the oxidation reaction. Moreover, the hole (h ? ) has a high oxidation potential, which can lead to direct oxidation of MB. At the same time, the valence band of CoZnFeO 4 (2.41 eV) has a higher potential than that of (1.99 eV), so that the hole can directly participate in the oxidation of MB, and can also react with hydroxide ions to generate hydroxyl radical. So under the action of hydroxyl radicals and holes, it leads to the decolorization of MB. The decolorization mechanism is shown as follows: Fig. 21 The XPS spectra of a wide-scan, b Co 2p c Fe 2p d O 1s of copper magnesium ferrite nanoparticles

MB intermediate and degradation pathway
At present, the research on photocatalytic degradation of dyes mainly focuses on the preparation of photocatalysts, the optimization of photocatalytic conditions, and reaction kinetics. While less attention is paid to the photocatalytic degradation mechanism of dye molecules [82]. Spadaro et al. [7] proposed that the photocatalytic degradation of dyes is mainly carried out by adding HOÁ to the carbon-nitrogen bond, and a series of unstable intermediates are formed and decomposed during the structural decomposition of dye chromophore [83]. Some intermediates may be more toxic than the parent molecule. Therefore, it is of great significance to determine the intermediates and whether there are intermediates that are difficult to be degraded. MB is a kind of dye with a complex molecular structure. Its color is usually determined by the chromophobe group. This means that the decolorization of MB is caused by the destruction of chromogenic groups, but often the dye has not been completely degraded, there are still a large number of toxic intermediates. To determine the degradation pathways and intermediates of MB, ion chromatography and ESI-MS are used for analysis (Fig. S3). We performed TOC tests on methylene blue at 200 mg/L and methylene blue degraded for 25 min. The TOC of the original methylene blue is 413.34 mg/L, after degradation for 25 min, TOC is 40.26 mg/L. The TOC removal rate is 90.25%. As shown in Fig. 23, the anions in the solution are determined by ion chromatography, and the MB produced Cl -, NO 3 and SO 4 2during the degradation process. The results showed that C-N, S-Cl, and C-S bonds are broken during the degradation of MB, thus Cl -, NO 3 and SO 4 2-are generated in the solution. To further determine the degradation pathway of MB, the degradation of products at different reaction times are identified by ESI-MS. As shown in Fig. 23, MB first deionized, S-Cl bond broke and Cl -(m/z = 284) is lost, and then HOÁ attacked C-S=C to generate sulfoxide (C-S(=O)-C) and -NH 2 fracture (m/z = 303). The rest of the structure degrades in two ways, with the -CH 3 bond having the lowest energy and breaking first (m/z = 270, 256, 281), Or -CH 3 is oxidized to HCHO on the surface of nucleophilic oxygen by chemical adsorption (m/z = 332). Sulfoxide then breaks under the hydroxyl radical attack. The involvement of the two pathways in MB degradation depends on the competition between surface chemisorption and direct hydroxylation of MB [84].
Finally, under the action of hydrogen peroxide, MB is broken down into small molecules such as carbon dioxide and water (Fig. 24).

Conclusions
In this study, a CoZnFeO 4 magnetic material with narrow band gap and rapid electron hole separation is designed and used for photocatalytic degradation of methylene blue in wastewater. CoZnFeO 4 with octahedral structure is synthesized from waste Under the optimal photocatalytic degradation conditions, green and efficient degradation of methylene blue is achieved. Meanwhile, the synthesis mechanism of CoZnFeO 4 nanomaterial and its photocatalytic degradation of methylene blue are explored. This study provided an effective solution for degrading dye wastewater by CoZnFeO 4 and exhibits great potential for the industrial application of ferrous sulfate waste.

Acknowledgements
This study is financially supported by Sichuan University-Panzhihua City Science and Technology Corporation Special Fund Project for Titanium White by-product Ferrous Sulfate Preparation 500 tons/year Nanometer iron Red Pigment and Co-production Sulfuric Acid Pilot Study Project (No. 2018CDPZH-5), and Sichuan Science and Technology Planning Project (No. 2019YFH0149).

Declarations
Conflict of interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.