Catalytic wet peroxide oxidation degradation of magenta wastewater and preparation of FeOCl/montmorillonite

In recent years, due to the increase in demand for textile products, the dye and dyeing industry has developed rapidly. If the untreated dye wastewater is discharged into rivers, it will appear deep reddish-purple due to the azo dye with high biological oxygen demand (BOD) and chemical oxygen demand (COD). If the ratio of BOD/COD is between 0.2 and 0.5, the dyeing wastewater contains a large amount of non-biodegradable matter in nature (Pello et al. 2018). The dye wastewater can cause serious environmental and health problems; magenta is the main component of dye wastewater (Tang et al. 2014). Catalytic wet peroxide oxidation (CWPO) technology is an important method to degrade organic wastewater. Hydrogen peroxide (H₂O₂) is used as an oxidant in CWPO technology, and the H₂O₂ is decomposed by the catalyst to form hydroxyl radicals (HO·) (Molina et al. 2010). HO· with a high oxidation potential of 2.8 V can attack many toxic organic pollutants, and it can mineralize them into carbon dioxide, water, and harmless inorganic substances (Wang et al. 2019). CWPO reaction has been extensively studied in recent years, and the developed heterogeneous catalyst can be used and recovered in a wider pH range (pH = 3.0–9.0), which overcomes the limitation of homogeneous catalyst, thereby making the process simpler and solving the problem of secondary pollution (Molina et al. 2010).

Metal oxides that can change valence are widely distributed in nature with abundant reserves and strong oxidizing properties, and they are useful for the degradation of organic matter (Malhotra et al. 2014). Iron oxychloride (FeOCl) is an important Fenton catalyst. From Table S1, the HO· generation rate constant over FeOCl was as high as 2.93 × 10⁻⁴s⁻¹, which was 1–3 orders of magnitude higher than other iron-based materials. When the rate constant was normalized to the unit surface area, it was found that it decomposes H₂O₂ more efficiently, and its HO· generation rate was 2–4 orders of magnitude higher than traditional iron-based materials. This shows excellent performance in degrading persistent organic compounds (Yang et al. 2013; Sun et al. 2018b). FeOCl has a unique iron coordination environment, including Fe(II)–O, Fe(III)–O, Fe–Cl and Fe–Fe, and a large number of unsaturated iron atoms are exposed on the plane (020). FeOCl has excellent catalytic performance due to its unique structure. It has been proven that when Fe(III) in FeOCl is inserted or reacted with other species, it is easily reduced to Fe²⁺ (Luo et al. 2019).

Montmorillonite (MMT) is a smectic clay; every layer is composed of an Al-based octahedral plate sandwiched between two tetrahedral silicate plates. Researchers found that high mechanical strength, good thermal stability, a porous and a large specific surface area are the advantages of MMT (Chetverikova et al. 2019). Catrinescu et al. used Al-pillared clay as a high surface area carrier and prepared Fe-loaded Al—Fe–pillared clay as a catalyst for the 4-chlorophenol CWPO reaction. The degradation rate and total organic carbon removal rate of 4-chlorophenol were 100% and 52%, respectively (Tian et al. 2011).

In this paper, Al-pillared MMT loaded FeOCl was used as the catalyst. Al exists in the MMT layer in the form of Al₁₃ polycation, and it was calcined to form the Al₂O₃ ‘oxidation column’. Al-pillared MMT can be a good catalyst carrier due to its large layer spacing and specific surface area. At the same time, the FeOCl loaded on the pillared MMT will decompose H₂O₂ through Fe(III) on the polar surface to produce HO·, to accelerate the decomposition of pollutants (Tian et al. 2005).

Finding the best preparation conditions (preparation method and FeOCl loading) of the catalyst, testing the stability of the catalyst, exploring the reaction mechanism, and selecting the appropriate equation to simulate the reaction are the main goals in the paper.

1 g acid-modified MMT was added to 30 mL AlCl₃·6H₂O solution (0.4 mol·L⁻¹). The mixture was stirred vigorously at 80 °C for 10 min, and then 5 mL polyethylene glycol (PEG-400) was added to the solution. After stirring for 5 min, 75 mL NaOH solution(0.4 mol·L⁻¹) was dropped into the reaction solution through a dropping funnel at a rate of 1 mL min⁻¹ and stirred vigorously so that the molar ratio of [OH⁻]/[Al³⁺] was 2.5:1 (Guo et al. 2004). After dropping, the mixture was continuously stirred for 2 h. After ageing for 12 h at 60 °C, the precipitate was centrifuged and washed six times until it did not contain Cl⁻. The precipitate was dried for 24 h at 60 °C, then it was calcined in an air atmosphere at 400 °C for 2 h to obtain Al-pillared MMT.

The structural parameters of the catalysts were determined by N₂ adsorption/desorption isotherms measured on the surface area and pore size analyzer (NOVE 2200e, Quantachrome). The setting test conditions were: the samples were pretreated at 160 °C for 18 h and then tested at −196 °C. The crystal structures of the samples were analyzed using an X-ray diffractometer (SmartLab SE, Rigaku). The test conditions were: Cu-Kα radiation at 40 kV and 30 mA, scanning speed 0.1°/s, scanning range 5–80°. The microscopic morphology and structure of the sample were analyzed using scanning electron microscope (SEM) ZEISS SIGMA, Carl Zeiss). A small amount of the sample was dispersed on the surface of the conductive tape, then scanned and observed after spraying gold. The molecular group of the sample was measured by Fourier infrared spectroscopy (IRTffinity-1S, Shimadzu). The test conditions were: no tablet forming was required, but the sample was fully dried; the sample scan wavenumber was 400–4,000 cm⁻¹. The catalysts were analyzed using an UV–Vis spectrophotometer (UV-2600, Shimadzu). Using barium sulfate as the reference, the wavelength scanning range was 200–800 nm.

0.05 g catalyst was put in to a three-necked flask. Then, 100 mL magenta-simulated dye wastewater (200 mg·L⁻¹) was poured into the flask, pH was adjusted to 3 and the mixture was stirred for 30 min to reach adsorption equilibrium. The H₂O₂/magenta ratio was controlled to 60:1; the reaction time was controlled to 210 min; samples at different times were taken to test and analyze the COD removal rate and the magenta removal rate in the water sample.

After that, the absorbance of the magenta solution was measured at different times during the reaction, and the concentration of the magenta solution was calculated by the fitting formula.

The N₂ adsorption/desorption isotherms of samples prepared with unpillared MMT and different loading methods are shown in Figure 1. From Figure 1, it can be seen that FeOCl/MMT(g) samples, FeOCl/MMT(s) samples, and the original MMT all belong to type IV isotherms. All of them have the characteristic performance of mesopores. It shows that the mesoporous structure of the original MMT of FeOCl/MMT(g) and FeOCl/MMT(s) samples were not destroyed during the pillaring and loading process. At low relative pressure (P/P₀ < 0.4), N₂ adsorption on the surface of the sample forms a monolayer. At moderate relative pressure (0.4 < P/P₀ < 0.5), N₂ continued to adsorb to form a multi-molecular layer. At high relative pressure (P/P₀ > 0.5), the capillary was condensed (Sun et al. 2014). The adsorption process of mesoporous materials usually shows adsorption—desorption hysteresis, which will lead to the formation of H3 hysteresis loops, indicating that the accumulation of nanosheets in the sample had formed slit holes (Gao et al. 2014).

Also, the relevant sample structure parameters are listed in Table 1. The specific surface area, pore size, and pore volume of FeOCl/MMT(g) and FeOCl/MMT(s) samples supported by metal ions and loaded are significantly increased, and the d₀₀₁ value increased from 1.235 nm of unpillared MMT to 1.441 nm and 1.524 nm, respectively. On the one hand, it shows that Al₂O₃ successfully pillared the layered structure of MMT. On the other hand, it shows that a small amount of active component FeOCl exists on the surface or between layers of the sample in oligomers, forming a new pore structure (Li et al. 2020). It shows that FeOCl can be used as an active component in the MMT layer, and it can also stabilize the carrier structure. Among them FeOCl/MMT(g) has a larger specific surface area (S_BET = 141 m²·g⁻¹) compared with FeOCl/MMT(s), increasing by 6%. The pore diameter (D_med) increased from 3.02 nm to 3.41 nm, and the total pore volume (V_total) increased from 0.137 cm³/g to 0.156 cm³/g. FeOCl/MMT(g) may therefore exhibit higher catalytic performance.

The elemental analysis of samples prepared with MMT and different loading methods are shown in Table 2. The element content of FeOCl/MMT(g) and FeOCl/MMT(s) samples are different from unpillared MMT. On the one hand, this can be attributed to the fact that cation exchange during the pillaring process causes a large number of metal Al ions to enter the MMT layer, resulting in the decrease of Na, Mg, K, Ca, and Si contents in the sample, indicating the success of pillared MMT (Wu et al. 2020). On the other hand, the content of Fe increased in the subsequent loading. Moreover, FeOCl/MMT(g) has a higher Fe content than FeOCl/MMT(s), indicating that the solid melting method may be more beneficial to load FeOCl on MMT (Kong et al. 2010).

X-ray diffraction (XRD) spectra of samples prepared with MMT and different loading methods are shown in Figure 2; the d₀₀₁ values were listed in Table 1. From Figure 2, the spectra of FeOCl/MMT(g) and FeOCl/MMT(s) catalysts were similar to the unpillared MMT, indicating that the layered structure of the original MMT was not destroyed.

The original montmorillonite (001) crystal plane diffraction peak signal was at 2θ = 9.4°. The characteristic diffraction peaks of the FeOCl/MMT(s) and FeOCl/MMT(g) catalysts moved at 2θ = 6.2°, indicating that Al ions had successfully entered between the MMT layers, which confirmed that the carrier MMT was successfully pillared. Also, the XRD peaks of FeOCl/MMT(s) and FeOCl/MMT(g) catalysts at 2θ = 37.9°, 35.6°, 25.9°, and 10.6° corresponding to {111}, {021}, {110}, {010} crystal planes are consistent with pure FeOCl crystal (JCPDS No.72-0619), indicating that both loading methods can convert FeCl₃·6H₂O into FeOCl and load it on MMT. However, due to the higher peak intensity of FeOCl/MMT(g) here, it exhibits higher crystallinity and better catalytic activity. According to Scherrer's equation, calculating the size of FeOCl crystals in FeOCl/MMT(g) is about 50 nm (Schnleber et al. 2015).

Figure 3 shows scanning electron micropscopy (SEM) spectra of samples prepared with unsupported MMT and different loading methods. It was found that unpillared MMT has a curled layered structure from Figure 3(a), which was made up of irregular agglomeration of plate-like particles formation. Figure 3(b) shows that pure FeOCl is a coral-like structure formed by random stacking of plate-shaped nanosheets. Figure 3(c) and 3(d) shows that the structure of FeOCl/MMT(g) and FeOCl/MMT(s) samples is similar to the original MMT structure, indicating that the layered structure of MMT was not destroyed during the pillar support and load process (Zhang et al. 2019). Also, in the FeOCl/MMT(g) catalyst in Figure 3(c), flat FeOCl crystals were observed on the surface of the carrier MMT. However, there were no obvious FeOCl crystals observed in the FeOCl/MMT(s) catalyst in Figure 3(d).

The catalysts prepared by different loading methods were treated with the same CWPO method to treat magenta-simulated dye wastewater. The COD removal rate of magenta wastewater changes with time, as shown in Fig. S2. It can be found that FeOCl/MMT(g) has a positive effect on COD compared with FeOCl/MMT(s); the removal rate of COD increased by 13.3%. The COD removal rate and magenta removal rate of the two catalysts are shown in Table 3, the catalytic effect of FeOCl/MMT(g) is better than that of FeOCl/MMT(s). The removal efficiencies expressed as COD are smaller than those obtained for magenta, confirming that magenta oxidation occurs in multiple steps and results in several by-products, which are more resistant to oxidation than the initial pollutant (Catrinescu et al. 2011).

According to the above analysis, the solid melting method is more suitable for removing magenta in wastewater.

In the CWPO reaction, FeOCl formed by FeCl₃·6H₂O conversion was the active component. Loading the proper amount of FeCl₃·6H₂O on the carrier formed more FeOCl to enhance the catalytic effect. Therefore, a series of x-FeOCl/MMT(g) samples with different loadings of FeOCl were prepared, where x represents the mass ratio of FeOCl to unsupported MMT, which were 1%, 3%, and 5%, respectively. The optimal FeOCl loading was selected through characterization analysis.

The Fourier-transform infrared spectroscopy (FT-IR) spectra of pure FeOCl and samples with different loadings are shown in Figure 4. In Figure 4, we can see similar unpillared MMT characteristic peaks that were observed in the FT–IR spectra of samples prepared with different loadings. The absorption peak at 3,622 cm⁻¹ corresponds to the –OH stretching vibration of the bound water and adsorbed water between the MMT layers. Its deformation vibration appeared at wavelength 1,746 cm⁻¹. The peak at 1,006 cm⁻¹ is the peak of deformation vibration of MMT Si–O–Si. The peak at 434 cm⁻¹ is the peak of deformation vibration of Si–O–Al formed by the Al ion entering the MMT layer (Pratiwi et al. 2017). The –OH was removed from the MMT carrier to form Si–O–Si and Si–O–Al bonds during the pillared heating process, indicating that the MMT is successfully pillared (Zhang et al. 2018). At 3,622 cm⁻¹ and 1,746 cm⁻¹ in the FT-IR spectrum of x-FeOCl/MMT(g) sample, the peaks of the –OH stretching vibration and deformation vibration at the position were blue-shifted. Also, the faint vibration peaks of Fe–O and Fe–Cl were observed from the FT-IR spectrum of x-FeOCl/MMT(g) sample at 670 cm⁻¹ and 1,596 cm⁻¹, which means that the active component FeOCl was produced in different loadings. Moreover, the peak intensities were different due to different loadings (Scully et al. 2010).

The UV–Vis spectra of pure FeOCl and samples with different loadings are shown in Figure 5. The transition absorption band of t₁ → t₂ and t₁ → e⁻ of Fe³⁺ were observed at λ = 230 nm and λ = 270 nm at the spectral lines. A weak absorption peak belonging to the polymeric iron oxide cluster (FeO)_n appeared at an external wavelength of 550 nm. The intensity of this absorption peak continued to increase with the increase of FeOCl loading (Merli et al. 2015). The existential form of iron depended on the amount of FeOCl loading, so when the loading was increased to 5%, high loading increases the formation of oligomeric iron, and affects the catalytic activity of the sample (Abd-El-Aziz et al. 1995).

The x-FeOCl/MMT(g) samples prepared with different FeOCl loadings were used as catalysts to degrade the magenta solution. Figure 6 shows the COD removal rate of the magenta solution over time, and Table 4 shows the COD removal rate and magenta removal rate after the reaction. When only MMT was used to degrade magenta solution, the COD removal rate was only 21.2%. The COD removal rate when only MMT was used is close to the COD removal rate when only H₂O₂ was used (Figure 7), indicating that MMT does not have a catalytic effect in this experiment. When FeOCl was loaded on MMT, the COD removal rate increases. This may be because MMT provides more active sites and makes FeOCl more evenly distributed. All catalysts (except MMT) show a high removal rate of magenta (Sun et al. 2014). When the loading was 3%, it can provide enough active sites to degrade dye wastewater well. However, when the loading exceeded 3%, the COD removal rate dropped slightly. This may because the formation of too much FeOCl caused an agglomeration behavior, which is not conducive to the dispersion of active components on the surface of the carrier. Some FeOCl may even enter the pores of the carrier and block some pore channels, resulting in a decrease in the molecules diffusion rate of reactants and products (Biddinger et al. 2011). The data demonstrates that 3%-FeOCl/MMT(g) was not only the most active for magenta but also the most efficient catalyst for COD removal.

Table 5 shows the COD removal rate of the magenta solution was only reduced by 12.4% after the catalyst was reused six times, suggesting that the catalyst maintains a high activity. However, it also indicated that a large amount of accumulation of intermediate products in the catalyst might block the pores, or it was due to the recovery and excessive heating treatments so that the active component FeOCl in the catalyst was polymerized (Tagliabue et al. 2004). Also, according to the dissolution data of total Fe ions, it can be found that as the reaction proceeds successively, the amount of total Fe ions dissolved gradually decreases. It showed that the active metal content on the catalyst surface was reduced, and the FeOCl nanoparticles existing between the catalyst layers were more stable. The quality of total Fe ions leaching showed that the active components of the catalyst were less lost (Bailón-García et al. 2013).

In the process of CWPO degradation of the magenta solution, benzoquinone (BQ) and tert-butanol (TBA) were added as free radical scavengers to inhibit the main oxidation groups · and HO· in the process of free radical reaction (Buc-Calderon & Roberfroid 1988). Potassium iodide (KI) was used to inhibit HO· and · (Ali et al. 2021). We explored the active groups that play a major role in the CWPO reaction system and set up six different reaction systems: (1) catalyst + H₂O₂; (2) catalyst + H₂O₂ + BQ; (3) catalyst + H₂O₂ + TBA; (4) catalyst + H₂O₂ + KI; (5) catalyst; (6) H₂O₂.

Figure 7 shows the effect of different active group trapping agents on the COD removal rate of the CWPO reaction system. There were significant differences in the COD removal rate of the magenta solution in different reaction systems.

According to the change of COD removal rate of systems (1)–(3), it can be seen that HO· plays an important role in the reaction. Comparing system (3) and system (4), the contribution of free HO· was dominant in the CWPO reaction system. However, the surface HO· also contributed. When only the catalyst was added in system (5), the COD removal rate was the lowest. From the aforementioned characterization and analysis, it can be seen that the catalyst has a large specific surface area and can adsorb organic pollutants. In system (6), only H₂O₂ was added, and the COD removal rate was about 20%. This is because H₂O₂ has an oxidizing effect and can degrade a small amount of magenta (Yang et al. 2019).

Kwan & Voelker proved that the decomposition of H₂O₂ and the degradation of organic compounds mainly occur on the surface of Fe₂O₃ (Kwan & Voelker 2003). Voinov et al. found that specific catalytic sites on the surface of nano-scale γ-Fe₂O₃ are at least 50 times more efficient in the production of HO· than Fe ions (Voinov et al. 2011). The polar surface (especially the 020 plane) in the basic structure of FeOCl is filled with a large number of unsaturated Fe atoms. These Fe atoms are uniformly arranged on the plane in the O–Fe–Cl configuration so that the polarization plane has a greater affinity for H₂O₂ molecules (Sun et al. 2018a). Therefore, the excellent catalytic effect of FeOCl/MMT in the reaction is attributed to the fact that FeOCl in the catalyst can decompose H₂O₂ in a short time to generate a large number of hydroxyl radicals. First, when H₂O₂ and organic pollutants are adsorbed, Fe on the polarized surface of FeOCl acts as the reaction center for the activation and decomposition of H₂O₂. The electrophilic Cl and O may increase the reduction potential of Fe, and the self-oxidation–reduction potential of FeOCl can promote and pass the hydrogen oxide reactions to produce ·HO₂. Moreover, Fe³⁺ is rapidly reduced to Fe²⁺. The resulting Fe²⁺ reacts with H₂O₂ to form HO·, the rate of generation of HO· on Fe²⁺ is 3–4 orders of magnitude higher than that on Fe³⁺. The HO· then attacks organic compounds. Finally, the redox cycle of the Fe²⁺/Fe³⁺ pair is completed, and the redox cycle is considered to be the essence of the Fenton reaction (Sun et al. 2018b; Luo et al. 2019). Figure S3 shows this process.

Figure 8 shows the degradation process of magenta at different temperatures. C/C₀ is the ratio of the concentration of magenta at time t to the initial concentration. The relevant data are listed in Table 6. It can be seen that the fitting curves of 60 °C and 75 °C had the same changing trend; the curve was a lying S shape, the magenta degradation efficiency was close to 100% after 120 min of reaction, and the linear correlation values (R²) were above 0.99. Both of them had the basic conditions for fitting with the Fermi equation. However, when the temperature was low (45 °C and 30 °C), it was not suitable to use the Fermi equation to simulate them. This is because the induction period of the reaction was too long when the temperature was low (Peleg 2010).

When the reaction temperature was high, this reaction was suitable for simulation with the Fermi function. Moreover, the reaction rate constants were 0.1726 min⁻¹ for 60 °C and 0.2987 min⁻¹ for 75 °C. At high temperatures, the catalyst performance was excellent. When there was a reaction induction period at the beginning of the reaction, the reaction speed was slow, and the catalyst surface was activated, but it was not uniform. Only a small amount of HO· was generated. After the reaction induction period, the reaction entered a pseudo-steady state stage when the concentration of H₂O₂ continuously decreased and the decomposition to produce HO· accelerated, speeding up the oxidation rate (Gordon & Marsh 2009; Sang et al. 2011).

FeOCl/MMT was used as catalyst, H₂O₂ as an oxidant, and a heterogeneous CWPO reaction was used to treat magenta wastewater. The reaction occurred easily and, it could be carried out at 60 °C. The highly active and stable FeOCl/MMT catalyst is the key to affecting the CWPO reaction rate. The 3%-FeOCl/MMT(g) catalyst was selected after optimizing the preparation conditions to degrade the magenta solution. The specific surface area of this catalyst is 141 m² g⁻¹, which has a relatively high specific surface area. The catalyst still had a high catalytic activity after six uses. The synthesis method of the FeOCl catalyst is simple, and it can effectively promote the generation of HO·. The fitting results show that the Fermi equation can describe the degradation process of magenta at high temperatures. Therefore, the catalyst provides a new opportunity for the CWPO method to degrade magenta water pollutants.

This work was supported by the key research and development program of Shaanxi, China (2018GY-067).

All relevant data are included in the paper or its Supplementary Information.