Insights into the mechanism of phenolic mixture degradation by catalytic ozonation with a mesoporous Fe₃O₄/MnO₂ composite

Abstract

A mesoporous Fe₃/MnO₂ composite was fabricated by a co-precipitation method in this paper, and it showed a much higher activity than Fe₃O₄ and MnO₂ in the catalytic ozonation of a p-cresol and p-chlorophenol mixture. The physicochemical properties of Fe₃O₄ and MnO₂ and Fe₃O₄/MnO₂ were compared using XRD, SEM, TEM and N₂ physical adsorption/desorption. pH had a significant effect on the degradation rate of the phenols and catalyst stability, and the degradation order of p-cresol and p-chlorophenol also varied in different mediums. pH 9 was found to be the optimal condition both for catalytic activity and metal leaching. Attenuated total reflection Fourier transform infrared spectra confirmed that ozone replaced chemically adsorbed water on the Fe₃O₄/MnO₂ surface and evolved into reactive radicals. Electron spin resonance and quenching experiments with different scavengers were conducted to reveal that the hydroxyl radicals were mainly responsible for the phenolic mixture degradation at pH 9, along with a little contribution of singlet oxygen and molecular ozone. A detailed mineralization pathway of the phenolic mixture was also proposed according to the gas chromatography-mass spectrometry results.

---

1. Introduction

Phenols are produced or consumed in many industries, such as in oil refineries, coking plants, chemical-synthesis factories, pharmaceuticals, plastic industries, textile manufacturing and many others. This will cause environmental pollution to the atmosphere, soil and water if they are not well governed. Among the phenols, p-cresol (Ph-CH₃) and p-chlorophenol (Ph-Cl) are the most prevalent phenolic compounds, which are used in the preparation of disinfectants, herbicides and fungicides.^1,2

The removal of phenolic compounds from industrial wastewater is currently an environmental challenge for its safe disposal and treatment,³ several physico-chemical and biological techniques have been developed, such as adsorption,⁴ extraction,⁵ biological degradation⁶ and chemical oxidation by peroxymonosulphate,⁷ permanganate,⁸ light,^9,10 H₂O₂, ozone and oxygen at high temperature and high pressure.¹¹ Among the reported chemical oxidation processes, ozonation is capable of efficient eliminating the recalcitrant phenolic compounds present in wastewater.¹²

However, this method was relatively expensive due to the cost of ozone generation, low solubility of ozone in solutions and low level of TOC removal in single ozonation.¹³ Consequently, it is often combined with ultrasound,¹⁴ UV light,¹⁵ visible light¹⁶ or other reagents¹⁷ to improve its treating efficiencies. Different catalysts were universally used in all these combined processes, and heterogeneous catalytic ozonation alone has also been recognized to be an impactful alternative for advanced treatment of wastewater, due to the generation of non-selective oxidant hydroxyl radical (·OH).¹⁸

Many transition metal oxides including MnO₂,¹⁹ Fe₃O₄,²⁰ NiO,²¹ ZnO²² and CuO²¹ led to efficient removals of organic pollutants in catalytic ozonation. They are often loaded on a support with larger surface area to get a better distribution in the solution.²³ Another strategy is to synthesize composite metal oxide material, such as Mn–Ce–O,²⁴ NiFe₂O₄,²⁵ ZnAl₂O₄ ²⁶ and NiO–CuO.²² Ma et al. believed that proper adsorption of ozone and organics on Mn–Co–Fe oxide facilitated the surface reactions on the composite,²⁷ and Lv et al. reported that addition of Mn both enhanced the interfacial electron transfer and increased the surface lewis sites on Mn–Co/γ-Fe₂O₃, which facilitate the redox process of the metal oxides and the catalytic decomposition of ozone to generate radicals.²⁸ In most of the literatures using composite material, single component solution was used, which is far from mixed composition in real wastewater. The degradation pathways of the organics and reactive radicals were also not intensively investigated.

For an easy separation of solid catalyst, magnetic material was developed in wastewater treatment. Fe₃O₄/MnO₂ with different morphologies were synthesized and applied in photocatalysis²⁹ adsorption³⁰ and activation of peroxymonosulfate.³¹ To the best of our knowledge, magnetic porous Fe₃O₄/MnO₂ has not been reported before in catalytic ozonation. In this work we synthesized a mesoporous Fe₃O₄/MnO₂ composite material and applied in catalytic ozonation for mixture of Ph-CH₃ and Ph-Cl. It was much more active than Fe₃O₄ and MnO₂ and easily magnetically separated. The competitive adsorption of O₃ with water molecular was observed on the surface of mesoporous Fe₃O₄/MnO₂ using an ATR-FTIR spectroscopy. Hydroxyl radical oxidation was mainly responsible to phenols degradation at pH 9, and detailed reaction pathway was proposed according to the intermediate products analysis.

2. Experimental

2.1 Chemicals and reagents

All the chemicals used in materials synthesis and catalytic ozonation process were analytical graded and used without further purification. FeC1₃·6H₂O (99%), C₆H₅O₇Na₃·2H₂O (≥99%), MnSO₄·H₂O (≥99%), NH₄HCO₃ (≥99%), glycerol (≥99%), p-chlorophenol (Ph-Cl, ≥98.0%), p-cresol (Ph-CH₃, ≥98%) and tert-butanol (t-BA, ≥98.0%) were purchased from Sinopharm chemical reagent Co., Ltd. KMnO₄ (≥99%), ethanol (99.7%), HNO₃ (65–68%), acetone (≥99.5%), ammonia solution (25%) and FeC1₂·4H₂O (98%) were bought from Beijing chemical plant. p-Benzoquinone (p-BQ, ≥98%) was procured from Alfa Asar. NaN₃ (99.0%) was from Tianjin Fuchen chemical reagents. Deionized water was used throughout to prepare different solutions.

2.2 Material synthesis

In a typical synthesis of Fe₃O₄, 1.35 g of FeC1₃·6H₂O and 0.5 g of FeC1₂·4H₂O were dissolved in 40 mL deionized water under rapid stirring and nitrogen purging, 3 mL of concentrated ammonia solution was added when the solution was heated to 85 °C. After 30 min, 0.5 M C₆H₅O₇Na₃·2H₂O solution was added at the same condition and stirred for 60 min. The solution was precipitated by 20 mL of acetone and was rinsed with water/ethanol three times. Magnetic Fe₃O₄ nanoparticles were collected and diluted with deionized water to give a similar colloidal magnetic fluid and preserved at room temperature.

In a typical procedure (Scheme 1), certain amount of MnSO₄·H₂O was added into 100 mL of glycerol/water (1 : 1) solution, and 50 mg of Fe₃O₄ nanoparticles were added after stirring for 30 min. Another 100 mL of glycerol/water (1 : 1) solution with 0.16 M of NH₄HCO₃ was added and stirred for 60 min at 50 °C. Fe₃O₄ nanoparticles strongly interact with Mn²⁺ in solvent, and directly combine with formed MnCO₃ micro sphere.³² Fe₃O₄/MnCO₃ composite was obtained by centrifugation and rinsed several times with water and ethanol. It was dried at 90 °C for 6 h and calcinated at 400 °C for 5 h to obtain Fe₃O₄/MnO₂. In a parallel procedure, MnO₂ was synthesized in the same process in the absence of Fe₃O₄ nanoparticles.

Scheme 1 Schematic illustration for the synthesis of Fe₃O₄/MnO₂.

2.3 Material characterization

The X-ray diffraction (XRD) patterns of Fe₃O₄, MnO₂ Fe₃O₄/MnCO₃, Fe₃O₄/MnO₂ and Fe₃O₄/Mn₂O₃ were obtained on X-ray diffractometer (X'Pert-PROMPD, PAN analytical B.V). The morphology was observed by transmission electron microscopy (TEM, JEM-2100, JEOL) at an acceleration voltage of 200 kV and scanning electron microscopy (SEM, SU8000), operating at an acceleration voltage of 10 kV. Energy dispersed X-ray spectroscopy (EDX) was performed on the same SEM microscope. The surface areas and pore structures were studied by nitrogen adsorption/desorption at 77 K (Adsorb-iQ, Quantachrome). The Barrett–Joyner–Halenda (BJH) and Horváth–Kawazoe (HK) methods were used to analyze the mesoporous and microporous structures, respectively. The potential leaching of Mn and Fe were analyzed by inductively coupled plasma optical emission spectrometry (ICP-OES, ICP 6300 Radial, Thermo Scientific). Magnetization curves of the samples were examined by vibrating sample magnetometer (VSM) instrument (Lake Shore 7410, USA).

The electrochemical performance was studied by cyclic voltammetry (CV) with charge–discharge cycling. The electrochemical measurements were carried out on an Autolab 302N electrochemical workstation in a three-electrode system with a platinum foil counter electrode, calomel electrode as the reference and catalyst modified conductive glass as working electrode. The electron spin resonance (ESR) signals of hydroxyl radicals spin trapped by 5,5-dimethyl-1-pyrroline (DMPO) were detected at ambient temperature on a JEOL (JES-FA200) spectrometer. A nitrogen spin-trapping reagent 5,5-dimethyl-1-pyrolin-N-oxide (DMPO) was used in the process. DMPO and Fe₃O₄/MnO₂ suspension were added into 1 L of ozone saturated solution, which was prepared by continuous bubbling of 2.5 mg min⁻¹ ozone into water.

2.4 Catalytic ozonation and analytical procedure

Catalytic ozonation was performed in a semi-batch reactor at 298 K under stirring. The reactor was fed with 1 L aqueous solution containing a mixture of 1 mM Ph-Cl, 1 mM Ph-CH₃ and 0.2 g of catalyst. Ozone was produced by ozone generator (Anseros Ozomat GM) from oxygen with high purity. The gaseous ozone was fed through a porous diffuser at the bottom of the reactor with a flow rate of 100 mL min⁻¹ and concentration of 50 mg L⁻¹. The samples were taken at certain intervals and analyzed with high performance liquid chromatography (HPLC, Agilent Technologies Series 1200), using a mixture of methanol and water (30/70%, v/v) containing 10 mM L H₃PO₄ as the mobile phase. The separation was performed using ZORBAX SB-C₁₈ (3.5 μm, 2.1 × 150 mm) reversed-phase column at 35 °C with a flow rate of 0.3 mL min⁻¹. Sample was analyzed by a UV detector with a wavelength of 220 nm. TOC analyzer (TOC-VCPH, Shimadzu) was used to evaluate the mineralization of the target compounds during catalytic ozonation. Gas chromatography/mass spectrometry (GC/MS, QP-2010 ultra, Shimadzu) was used for the identification of foremost byproducts peaks in catalytic ozonation of phenolic mixture. Electron impact ionization with 70 eV was used for m/z fragmentation. The capillary column was Rxi®-5ms, and the column temperature was programmed to increase from 60 °C to 200 °C at 10 °C min⁻¹ and from 200 °C to 270 °C at 5 °C min⁻¹.

The adsorption of ozone on the catalyst surface was analyzed with attenuated total reflectance Fourier transformed infrared spectroscopy (ATR-FTIR). 0.1 g of Fe₃O₄/MnO₂ composite material was added to 2 mL D₂O and sonicated. After the pH of solution was adjusted to 9 with 0.05 M NaOH solution, 100 mL min⁻¹ of gaseous O₃ with concentration of 2.5 mg min⁻¹ was bubbled into the solution. The ATR-FTIR spectra were recorded by using a TENSOR 27 infrared spectrometer with a DLATGS detector and a ZnSe horizontal ATR cell. Infrared spectra in the range of 4000 to 800 cm⁻¹ were obtained by averaging 32 scans with a resolution of 4 cm⁻¹ at room temperature.

3. Results and discussion

3.1 Structural characterization of catalysts

XRD was performed to identify the phases of the synthesized composite and calcination effects on crystalline phase, the results obtained is presented in Fig. 1. All diffraction peaks are indexed to pure phases of γ-MnO₂ or Fe₃O₄. The intensive peaks at 24.1°, 31.4°, 36.4°, 42.7° and 56.5° refers to the γ phase of MnO₂, and the peaks at 30.1°, 35.6°, 42.6°, 57.2° and 63.0° were characteristic of Fe₃O₄. The pattern of initially synthesized Fe₃O₄/MnCO₃ was confirmed with low intensity peaks of Fe₃O₄ and high intensity peaks of MnCO₃ at 24.3°, 31.4°, 41.5°, 51.7° and 60.1°, suggesting the wrapping of Fe₃O₄ in MnCO₃ layers. The desired composite Fe₃O₄/MnO₂ was obtained after calcination of Fe₃O₄/MnCO₃ at 400 °C. All characteristic peaks appeared on the composite material, confirming the composition of MnO₂ and Fe₃O₄.^33,34 Diffraction peaks of Mn₂O₃ at 23.1°, 32.9°, 39.3°, 49.5°, 55.2° and low intensity peaks of Fe₃O₄ confirm the formation of composite Fe₃O₄/Mn₂O₃, after calcination of Fe₃O₄/MnCO₃ at 600 °C.³⁵

Fig. 1 XRD patterns of Fe₃O₄, MnO₂, Fe₃O₄/MnCO₃, Fe₃O₄/MnO₂ and Fe₃O₄/Mn₂O₃.

The morphologies of Fe₃O₄/MnO₂ composites were firstly elucidated by TEM. Fig. 2(A) and (B) show the TEM images of MnO₂ microsphere and Fe₃O₄ nanosphere before amalgamation, respectively. The particle sizes of them were 0.5–2 μm for MnO₂ and 8–15 nm for Fe₃O₄, respectively. Fig. 2(C) and (D) illustrated TEM images of Fe₃O₄/MnO₂ composite with different magnifications. They have sphere-like morphology, and porous surface was found in Fig. 2(D). This should be formed from calcination at high temperature. The particle size of the composite materials did not change much after the combination of Fe₃O₄ nanoparticles, within the range of 0.5–1.5 μm.

Fig. 2 TEM images of pure MnO₂ (A), Fe₃O₄ (B) and Fe₃O₄/MnO₂ composite (C and D).

The Fe₃O₄/MnO₂ was also analyzed with SEM. Fig. 3(A) and (B) with different magnifications shown that the material was uniform with microsphere morphology. Very fine particles can be found on the surface of the MnO₂ sphere in Fig. 3(A), which could be the Fe₃O₄ nanoparticles. Choosing a typical microsphere in SEM, the EDS spectrum in Fig. 3(C) showed the distribution of three elements with different colors, revealing that Fe₃O₄ and MnO₂ are very uniformly distributed as isolated nanoparticles. Fig. 3(D) showed Mn, O and Fe characteristic peaks in EDX spectrum. The peak at 6.5 keV is the overlapping signals of Fe and Mn, previously reported in Fe–Mn composite.³⁶ The insert table in Fig. 3(D) shows the weight percentage and atomic percent of the three elements in the composite microsphere. The molar contents are 24.0% for Mn, 7.3% for Fe and 68.7% for O, respectively. This indicated a low concentration of Fe₃O₄ in the composite materials.

Fig. 3 SEM images of Fe₃O₄/MnO₂ composite (A and B), elemental mapping (C) and EDS pattern (D).

The magnetic properties of Fe₃O₄/MnO₂, Fe₃O₄ and MnO₂ were studied by VSM at room temperature, and the result is shown in Fig. 4. The saturation magnetizations of Fe₃O₄/MnO₂, MnO₂ and Fe₃O₄ were 8.6, 0.9 and 26.5 emu g⁻¹, respectively. Magnetic effect in composite material Fe₃O₄/MnO₂ decreased because of coverage of Fe₃O₄ with non-magnetic MnO₂. This also confirmed with XRD peak intensities in Fig. 1. However the composite material still remains high paramagnetic nature, and it can be completely separated in about 55 min from the treated water, by applying external magnetic field.

Fig. 4 The magnetization hysteresis of Fe₃O₄/MnO₂, Fe₃O₄ and MnO₂ microspheres.

Fig. 5 showed the adsorption/desorption isotherms of MnO₂, Fe₃O₄, Fe₃O₄/MnO₂, and corresponding pore sizes distribution. All materials showed a typical adsorption/desorption curves in Fig. 5(A), indicating a mesoporous structure. MnO₂ showed the largest pore size, with average pore width of 17 nm. Fe₃O₄ showed very sharp pore size distribution and the dominant pore width was around 5 nm. The composite Fe₃O₄/MnO₂ had a larger pore size around 7 nm, and the pore volume was obviously larger than that of Fe₃O₄, and was similar to that of MnO₂. The BET surface areas of the materials were calculated from N₂ adsorption–desorption isotherms. MnO₂ obviously showed the highest particle size in Fig. 2(A), and it had the lowest surface area of 35.1 m² g⁻¹. Fe₃O₄ was much smaller than MnO₂, and its surface area increased to 66.4 m² g⁻¹. Though the particle size of Fe₃O₄/MnO₂ is close to MnO₂, it showed much higher surface area of 103.2 m² g⁻¹, for its rich porous structure, which was already revealed in Fig. 2(D).

Fig. 5 Adsorption/desorption isotherm of the materials (A) and corresponding pore size distribution (B).

3.2 Catalytic ozonation of phenolic mixture

3.2.1 Influence of pH. The catalytic activity of mesoporous Fe₃O₄/MnO₂ composite was evaluated at different initial pH. The initial pH of phenolic mixture solution was adjusted with HCl or NaOH solution to cover a range from pH 5 to 11. It is already reported that ozone decomposition into hydroxyl radicals is easier in basic solution than neutral or acidic solution, because hydroxide ion may play a role of initiator of the chain reaction of ozone decomposition.³⁷ Fig. 6 showed the degradation of phenolic mixture is influenced by pH. The degradation of Ph-CH₃ followed the order pH 9 > pH 11 > pH 7 > pH5 in Fig. 6(A), and the same trend for Ph-Cl was shown in Fig. 6(B). It is believed that hydroxyl radicals are easily generated in basic solution, so the organics will be degraded faster in basic solution than in neutral and acid solution. The optimum pH for catalytic ozonation of Ph-CH₃ and Ph-Cl mixture is pH 9. At higher pH of 11, we found a little decrease in the degradation rates of phenolic mixture, similar negative effect of very strong basic medium was also observed by Ma et al.³⁸ He explained that this should be due to the interactions between hydroxyl radicals in solution or the possible scavenger of carbonate generated as byproducts during catalytic ozonation, which lowered the concentration of hydroxyl radicals in solution, and thus reduced the possibility of contact between hydroxyl radicals and contaminants.

Fig. 6 The degradation of Ph-CH₃ (A) and Ph-Cl (B) and TOC removal (C) in catalytic ozonation with Fe₃O₄/MnO₂ suspension at different pH. Reaction conditions: [Fe₃O₄/MnO₂]: 0.2 g L⁻¹, [O₃]: 2.5 mg min⁻¹, [Ph-CH₃]: 1 mM, [Ph-Cl]: 1 mM.

For the TOC removal result in Fig. 6(C), we found a similar trend to those of phenols degradation. At pH 5 the degradation rate was the lowest and basic solution benefited the phenols mineralization. The TOC removal rates were very close at pH 9 and 11, but pH 11 required more reagents to adjust the solution pH. From the results of phenols degradation and mineralization, we concluded that pH 9 was the best condition considering the catalytic activity.

The metal leaching from catalyst will cause deactivation of catalyst and second pollution of metal ions. It is also very important to examine the influence of pH on the leaching behavior of material in catalytic ozonation process. In this work, the concentrations of residual Mn and Fe ions were determined with ICP at the end of each reaction with three catalysts and the results were shown in the Table 1.

Table 1 Mn and Fe leaching from Fe₃O₄/MnO₂ at different initial pH

Catalyst	Fe (mg L^−1)	Mn (mg L^−1)	pH (initial)	pH (final)
Fe3O4/MnO2	0.60	17.7	5	2–3
Fe3O4/MnO2	0.15	1.00	7	3–4
Fe3O4/MnO2	0	0	9	4–5
Fe3O4/MnO2	0	0	11	6–7
Fe3O4/Mn2O3	0.1	3.4	9	6–7
Fe3O4/MnCO3	0	24.3	9	6–7
Fe3O4	0	0	9	4–5
MnO2	0	0	9	4–5

---

As indicated by TOC removal, mixed phenols were not completely at the end of reaction. Carboxylic acid intermediates were normally produced during organics oxidation in the solution and they are more resistant in ozonation and catalytic ozonation.³⁹ This made the solution more acidic. This was confirmed in Table 1 that final pH of the solution all decreased at the end of the reaction. It was found that both Mn and Fe were leached from the Fe₃O₄/MnO₂ catalyst in the initially acid solution; especially a serious leaching of Mn was noticed in the solution with initial pH 5. The concentrations of Fe ion and Mn ion were 0.6 mg L⁻¹ and 17.7 mg L⁻¹, respectively. The leaching was greatly decreased when initial pH increased to 7 though the final pH also became acidic. The leaching of Mn and Fe even cannot be detected after reaction in solution with initial pH 9 and 11. This told that pH is a very important parameter in the stability of the catalyst.⁴⁰

In the case of other composites, obvious leaching of Mn were observed from Fe₃O₄/Mn₂O₃ (3.4 mg L⁻¹) and significant leaching of Mn ion was detected with Fe₃O₄/MnCO₃ (24.3 mg L⁻¹) even at pH 9. This meant MnO₂ was much more stable than Mn₂O₃ and MnCO₃ in the composite. For the single component oxides, Fe₃O₄ and MnO₂ were both very stable in the solution with initial pH 9 and no metal ions were detected at the end of the reactions.

3.2.2 Catalyst activity for phenolic mixture. The activity of Fe₃O₄/Mn₂O₃, Fe₃O₄/MnO₂, Fe₃O₄/MnCO₃, Fe₃O₄ and MnO₂ were compared in degradation and mineralization of mixture of Ph-CH₃ and Ph-Cl at pH 9. Ozonation was also carried out under the same condition to compare with catalytic ozonation. It was obvious that all curves in Fig. 7(A–C) showed the same trend.

Fig. 7 Catalytic degradation of Ph-CH₃ (A) and Ph-Cl (B) and TOC removal (C) with synthesized materials, and TOC removal (D) in adsorption at pH 9. Reaction conditions: [catalyst]: 0.2 g L⁻¹, [O₃]: 2.5 mg min⁻¹, [Ph-CH₃]: 1 mM, [Ph-Cl]: 1 mM, initial pH: 9.

Fig. 7(A) showed the degradation rates of Ph-CH₃ at 20 min over Fe₃O₄/Mn₂O₃, Fe₃O₄/MnO₂, Fe₃O₄/MnCO₃, Fe₃O₄ and MnO₂ were 64.2%, 99.5%, 73.1%, 75.8% and 89.4%, respectively, while that was only 47.8% in ozonation. The degradation of Ph-Cl was slightly slower at the same condition in Fig. 7(B), and the corresponding removal rates were 56.6%, 93.9%, 59.8%, 66.9% and 79.1%, respectively. For a comparison, the Ph-Cl degradation rate was 40.3% in ozonation at 20 min. In all cases, we noticed that Ph-CH₃ was relatively more easily degraded than Ph-Cl, which is opposite to the conclusion we obtained in photocatalytic degradation of a series of phenols.⁴¹ This can be due to the different reactive species and reaction medium. Superoxide radicals was mainly produced in photocatalytic degradation with Ag⁺/TiO₂ in weak acid solution,⁴¹ while this greatly change in catalytic ozonation of basic solution, hydroxyl radicals are more possibly dominant oxidants in this process, and this would be discussed later in this paper.

From Fig. 7(C), TOC removals were 36.6%, 63.7%, 51.1%, 49.5%, 80.6% and 48.3% in ozonation and catalytic ozonation with MnO₂, Fe₃O₄, Fe₃O₄/MnCO₃, Fe₃O₄/MnO₂ and Fe₃O₄/Mn₂O₃ suspension, respectively. Ozonation showed the lowest efficiency in phenols degradation and mineralization. The catalytic activity of the synthesized material followed the order Fe₃O₄/MnO₂ > MnO₂ > Fe₃O₄ > Fe₃O₄/MnCO₃ > Fe₃O₄/Mn₂O₃. From activity comparison, only Fe₃O₄/MnO₂ showed a positive synergetic effect, while Fe₃O₄/MnCO₃ and Fe₃O₄/Mn₂O₃ showed inferior removal efficiency. Though Fe₃O₄ showed higher surface area but its activity was lower than MnO₂. Recently analogous activity was also observed by Zhang et al. in catalytic activation of peroxymonosulfate (PMS) for degradation of methylene blue.³¹

Overall, the catalytic activities were greatly enhanced in the composite Fe₃O₄/MnO₂ material than the single component metal oxide. This revealed an obvious synergy between Fe₃O₄ and MnO₂. This synergy can be originated from multiple components, multi valence, high surface area and porous structure of the Fe₃O₄/MnO₂ microsphere.¹⁹ A synergistic effect of copper and cobalt was also observed in Cu–Co–O composite for phenol mineralization, which showed that dispersion of cobalt on copper oxide surface boosted the interaction between catalyst Lewis acid sites and ozone, which consequently increase in catalytic ozonation activity.⁴²

The degradation data of phenolic mixture in the first 20 min was fitted to be first order kinetics. The highest apparent reaction rate constant on mesoporous Fe₃O₄/MnO₂ was found to be 0.28 and 0.14 min⁻¹ for Ph-CH₃ and Ph-Cl, respectively. Meanwhile those for Ph-CH₃ were 0.11 and 0.073 min⁻¹ on MnO₂ and Fe₃O₄, and 0.078 and 0.056 min⁻¹ for Ph-Cl on MnO₂ and Fe₃O₄, respectively. We did not directly compare the activity of this composite material with others reported, as normally quite different reaction conditions were used in published papers. From the reaction rate constants, a good synergetic effect was confirmed in Fe₃O₄/MnO₂ again.

For the catalytic stability evaluation of the synthesized catalyst, Fe₃O₄/MnO₂ was magnetically separated after reaction, and washed three times with deionized water and dried at 100 °C for recycling. The stability of catalyst was evaluated by their reuses performance. As seen in Fig. 8, when the catalyst was reused at pH 9, phenolic mixture removal decreased greatly from 80.6% to 76.3%, 75%, 73.8%, 73.1% and 70.1% respectively in 6 cycles. The catalyst also slightly lost its weight during each operation of recollection and recycling, as indicated by the blue column in Fig. 8. It was noticed that the catalytic activity loss was proportional to the catalyst weight loss, they both decreased to about 90% after 6 cycles. This meant the composite Fe₃O₄/MnO₂ catalyst was very stable at pH 9, and it is also easy to separate and recycle for its magnetic property.

Fig. 8 Stability test of Fe₃O₄/MnO₂ composite in catalytic ozonation of phenolic mixture. Reaction conditions: [Fe₃O₄/MnO₂]: 0.2 g L⁻¹, [O₃]: 2.5 mg min⁻¹, [Ph-CH₃]: 1 mM, [Ph-Cl]: 1 mM, initial pH: 9.

3.3 Reaction mechanism of Fe₃O₄/MnO₂ catalytic ozonation

3.3.1 Ozone adsorption on the catalyst. It was already reported that gaseous ozone decomposition on the catalyst surface mainly consists two steps: the adsorption of ozone on the surface and desorption of the adsorbed intermediates.⁴³ Lv et al. found that ozone molecular competed with water molecular to adsorb on the surface of catalyst and converted into hydroxyl radicals later.⁴⁴ An ATR-FTIR method was firstly developed to reveal the competition of ozone adsorption with chemically adsorbed water on the surface of catalyst.²⁸ In this paper, adsorption of ozone on the catalyst Fe₃O₄/MnO₂ was also analyzed with ATR-FTIR at pH 9. During the analysis, D₂O was used as the solvent instead of H₂O to distinguish from the –OH on surface surface of the catalyst.

In Fig. 9(A), two peaks at 2262 and 1155 cm⁻¹ appeared in the systems, which were assigned to the vibrations of hydrogen-bonded D₂O on the catalyst surface. Fe₃O₄ showed weak peak at 2262 cm⁻¹ and no peak presented at 1155 cm⁻¹, and Fe₃O₄/MnO₂ showed the highest adsorption. In the system with composite Fe₃O₄/MnO₂ material, the intensities of both peaks decreased quickly with bubbling of ozone. In Fig. 9(B), the two peaks both disappeared from 5 min, and also did not present again at 10 or 15 min. This suggested that the adsorbed D₂O was replaced by ozone molecular. The experimental data verified that ozone was adsorbed on the surface of Fe₃O₄/MnO₂ while competing with water molecules. Then adsorbed ozone reacted with hydroxyl on the catalyst surface to convert to reactive radicals, and diffused into bulk solution. The addition of Mn was believed to enhance the amount of surface Lewis acid sites and thus promote the adsorption of ozone.²⁸ In this case; the introduction of Mn in the composite material also benefited the adsorption of ozone on the surface, and resulted in generation of more reactive species.

Fig. 9 ATR-FTIR spectra of different materials (A) and different time (B) in Fe₃O₄/MnO₂ suspension in ozonation.

It was already reported that the mechanism of gaseous ozone decomposition consists mainly of redox steps: composites of transition metal oxides are useful for the catalyst to undergo oxidation and reduction, enhancing the decomposition of ozone.⁴⁴ Therefore, the results suggest that the same procedure might occur in the catalyst aqueous suspension with ozone. The coexistence of Fe²⁺/Fe³⁺ (ref. 45) and Mn⁴⁺/Mn³⁺ (ref. 46) in Fe₃O₄/MnO₂ increased the oxidation and reduction involved in the decomposition of ozone. The rate of ozone decomposition was obviously enhanced in the catalyst suspensions (Fig. 6), similarly to TOC removal in these systems.

Furthermore, the cyclic voltammetry (CV) behaviors of Fe₃O₄/MnO₂, Fe₃O₄, and MnO₂ electrodes were investigated in pH 9 solution under an air or ozone. As shown in Fig. 10, the amount of current was observed at electrode as Fe₃O₄/MnO₂ > MnO₂ > Fe₃O₄. The results indicate surface charge on Fe₃O₄/MnO₂, MnO₂, and Fe₃O₄. With the addition of ozone, a reduce current was observed with the Fe₃O₄/MnO₂, MnO₂, and Fe₃O₄ electrodes, verifying that the reduction reaction occurred on the surface of catalyst with the decomposition of O₃.⁴⁴ In general conditions ozone can oxidize Mn³⁺ to Mn⁴⁺ in Fe₃O₄/MnO₂, considering the standard reduction potential E(Mn⁴⁺/Mn³⁺) = 0.95 V and standard reduction potential of E(O₃/O₂) = 2.07 V. Then the Fe²⁺ in composite can reduce the Mn⁴⁺ to its initial state Mn³⁺, E(Fe³⁺/Fe²⁺) = 0.77 V. Therefore, the efficient regeneration of Mn³⁺ surface species by this process would be responsible for the remarkable increase on the activity of organic oxidation.

Fig. 10 CV scans results of (A) Fe₃O₄, (B) MnO₂ and (C) Fe₃O₄/MnO₂ electrodes at pH 9, before and after 10 min of ozonation. Scan rate: 30 mV s⁻¹.

3.3.2 Quenching experiments. To reveal the efficient degradation mechanism of phenolic mixture, different kinds of radical scavengers were used in the oxidation processes. Fig. 11(A) showed that the presence of 5 mM t-BA in the phenolic mixture solution caused a 31.2% reduction of removal rate of Ph-CH₃ at 30 min, and similarly Ph-Cl removal rate was decreased by 28.3% in Fig. 11(B). The quenching effect was more obvious when increasing the concentration of t-BA. The degradation rates of Ph-CH₃ and Ph-Cl decreased by 66.5% and 55.3% with 10 mM t-BA, but this did not vary when more t-BA was added. This indicated the predominant function of hydroxyl radical in catalytic ozonation of phenols in basic solution. In our previous work we report superoxide radical to be the major reactive species in catalytic ozonation of organic contaminants. Superoxide radical was also described to be the main reactive species in the photocatalytic oxidation of toluene.^19,47 In this work, 5 mM p-BQ, a widely recognized scavenger of superoxide radical, was used to evaluate its role in our system. The negligible inhibition effect indicated that superoxide radical did not contribute to the phenols degradation. Furthermore, 5 mM NaN₃ was introduced to determine the possible role of singlet oxygen in catalytic ozonation in basic solution. The results in Fig. 11(A) and (B) showed that both phenols degradation were partially inhibited. The suppressions were 18.2% for Ph-CH₃ and 2% for Ph-CH₃, which indicated that singlet oxygen played a more important role in Ph-CH₃ degradation. Higher concentrations of NaN₃ were also tried, but the suppression effect was almost the same.

Fig. 11 The influence of t-BA, p-BQ and NaN₃ in catalytic ozonation of Ph-CH₃ (A) and Ph-Cl (B) and (C) ESR spectrum at different pH. Reaction conditions: [catalyst]: 0.2 g L⁻¹, [O₃]: 2.5 mg min⁻¹, [Ph-CH₃]: 1 mM, [Ph-Cl]: 1 mM.

To verify the speculation of hydroxyl radical production, ESR experiments were performed to directly detect it. The result in Fig. 11(C) clearly showed the existence of ·OH radical at pH 9 in the system with Fe₃O₄/MnO₂ suspension, and the peak intensity decreased at pH 11. This was because –OH in basic solution benefited the decomposition of ozone, but higher pH might also cause invalid annihilation between hydroxyl radicals.³⁸ The degradation at high pH was comparatively lower in photocatalytic degradation of creatinine and phenol, showing similar decreasing effect at high pH.⁴⁸ This can well explain the degradation trends of phenols in Fig. 6(A) and (B), that phenols degradation was accelerated in basic solution and pH 9 is the best choice. The signal of hydroxyl radical did not appear in the suspension of Fe₃O₄/MnO₂ at pH 5 or 7. Previously, we reported that phenols degradation in MnO₂ catalytic ozonation in weak acid solution was not due to the hydroxyl radical oxidation, but superoxide radical and other species.¹⁹

The ESR result obtained in this paper was consistent with that conclusion. Moreover, only very weak peak of hydroxyl radical presented in the ozonation at pH 9, this could possibly be due to two reasons. Firstly, phenols were easily attacked by ozone molecular; ozone was partially consumed by phenol before transformation into hydroxyl radicals. Secondly, to achieve a quick analysis of ESR, a smaller catalytic ozonation system was built and 0.25 mg min⁻¹ of ozone was used. This much lower dosage of ozone input will also decrease the signal intensity in ESR. Based on the quenching experiments and ESR result, we concluded that hydroxyl radicals were primarily generated from catalytic decomposition of ozone in basic solution and attack phenols. Besides hydroxyl radical, singlet oxygen and ozone molecular also partially contributed to the phenols degradation. This is easy to understand because ozone molecular was also very powerful to attack phenols directly.

3.4 Reaction pathway of phenolic mixture

The intermediates of phenolic mixture mineralization with porous Fe₃O₄/MnO₂ were illustrated through GC-MS to explore the reaction pathways.

Fig. 12 shows the evolution of organic components in the solution at different time, and the corresponding components are confirmed with MS and shown in Table 2. All of the identified compounds were unequivocally identified using the NIST98 library database. Result in Fig. 6(C) that phenolic mixtures were nearly completely mineralized.

Fig. 12 The organic components in solution during Fe₃O₄/MnO₂ catalytic ozonation of phenol mixture at different time [I: p-cresol; II: p-chlorophenol; III: hydroquinone; IV: muconic acid; V: p-benzoquinone; VI: malic acid; VII: hydroxyhydroquinone; VIII: chloromaleic acid; IX: p-chlorocatechol; X: maleic acid; XI: oxalic acid, XII: formic acid].

Table 2 Parent compounds and intermediates in the catalytic ozonation of phenolic mixture^a

	Compounds	R.T (min)	m/z(R.A%)
a “R.T” means retention time, “R.A” means relative abundance.
I	p-Cresol	5.6	107(100), 108(91.7), 77(25.9), 79(21.3)
II	p-Chlorophenol	7.4	128(100), 127(31.1), 65(30.3), 64(13.7)
III	Hydroquinone	3.9	110(100), 39(62.3), 55(21.2), 81(17.4)
IV	Muconic acid	4.1	97(100), 96(75.1), 41(63.9), 51(40.4)
V	p-Benzoquinone	6.2	108(100), 82(59.3), 109(44.1), 81(27.7)
VI	Malic acid	6.7	133(100), 151(69.5), 32(39.2), 44(24.5)
VII	Hydroxyhydroquinone	8.9	126(100), 52(41.2), 80(24.3), 39(20.11)
VIII	Chloromaleic acid	11	149(100), 32(80), 44(60), 177(25)
IX	p-Chlorocatechol	14.9	144(100), 146(38.1), 63(20.1), 147(14.8)
X	Maleic acid	4.5	72(100), 45(57.0), 55(30.7), 43(17.0)
XI	Oxalic acid	7.6	44(100), 57(81.3), 43(43.5), 56(18.5)
XII	Formic acid	9.4	46(100), 45(62.4), 60(29.8), 68(17.4)

---

Initially Ph-CH₃ and Ph-Cl separately showed their characteristic peaks. After 5 min reaction, the intensities of their peaks both decreased and seven new peaks appeared. These components with these components with low concentration were identified as hydroquinone, muconic acid, p-benzoquinone, malic acid, hydroxyhydroquinone, chloromaleic acid and p-chlorocatechol. It meant that a small amount of phenols opened their aromatic rings and small molecular organic acids were formed. With the reaction proceeding, more phenolic mixture was degraded and some intermediate products were further oxidized. Subsequently parent compounds of phenolic mixture completely disappeared after 30 min, and a very sharp peak of p-benzoquinone presented. Hydroquinone, malic acid, hydroxyhydroquinone, p-chlorocatechol, chloromaleic acid, p-chlorocatechol also disappeared and peaks for maleic acid, oxalic acid and formic acid was presented. From 5 to 10 and 30 min, the peak intensity for muconic acid continuously increased. At the end of the reaction, only little amount of formic acid and oxalic acid remained in the solution, which were due to their relatively higher resistance to oxidizing species and low concentration in solution. This trend was in consistent with the TOC result in Fig. 6(C) that phenolic mixtures were nearly completely mineralized.

According to the intermediates evolution in solution, we proposed the phenolic mixture mineralization pathway in mesoporous Fe₃O₄/MnO₂ catalytic ozonation in Scheme 2.

Scheme 2 Mineralization pathway for Ph-CH₃ and Ph-Cl mixture in Fe₃O₄/MnO₂ catalytic ozonation at pH 9.

4. Conclusions

Mesoporous Fe₃O₄/MnO₂ composite exhibited much higher activity in catalytic ozonation and mineralization of cresol and chlorophenol mixture than Fe₃O₄ and MnO₂, due to its high surface area, rich porosity and multi components, which provide more reaction sites on the surface of catalyst. pH 9 was found to be the optimal condition for the least metal leaching and highest catalytic activity. The catalyst is easily magnetically separated and recycled, and the activity almost kept unchanged in 6 cycles of catalyst reuse, excluding the influence of weight loss of catalyst. Ozone competed with water molecular to adsorb on the surface of Fe₃O₄/MnO₂ and evolved into hydroxyl radical and other species. The degradation of cresol and chlorophenol was both mainly attributed to hydroxyl radical oxidation, along with little contribution of singlet oxygen and ozone molecular. The reaction pathway in mineralization of the phenolic mixture was proposed according to the GC-MS results obtained at different intervals.

Acknowledgements

The authors greatly appreciate the financial support from the National Natural Science Foundation of China (No. 21207133) and the National Science Fund for Distinguished Young Scholars of China (No. 51425405).

Notes and references

1. O. Botalova, J. Schwarzbauer, T. Frauenrath and L. Dsikowitzky, Water Res., 2009, 43, 3797–3812.
2. A. Mantilla, G. Jacome-Acatitla, G. Morales-Mendoza, F. Tzompantzi and R. Gomez, Ind. Eng. Chem. Res., 2011, 50, 2762–2767.
3. C. C. Chang, S. K. Tseng, C. C. Chang and C. M. Ho, Bioresour. Technol., 2003, 90, 323–328.
4. M. Kragulj, J. Trickovic, A. Kukovecz, B. Jovic, J. Molnar, S. Roncevic, Z. Konya and B. Dalmacija, RSC Adv., 2015, 5, 24920–24929.
5. C. F. Yang, Q. A. Yu, L. J. Zhang and J. Z. Feng, Chem. Eng. J., 2006, 117, 179–185.
6. Z. L. Li, J. Nan, J. Q. Yang, X. Jin, A. Katayama and A. J. Wang, RSC Adv., 2015, 5, 89157–89163.
7. M. Chowdhury, O. Oputu, M. Kebede, F. Cummings, O. Cespedes, A. Maelsand and V. Fester, RSC Adv., 2015, 5, 104991–105002.
8. Y. Song, J. Jiang, J. Ma, S. Y. Pang, Y. Z. Liu, Y. Yang, C. W. Luo, J. Q. Zhang, J. Gu and W. Qin, Environ. Sci. Technol., 2015, 49, 11764–11771.
9. H. Sun, G. Zhou, Y. Wang, A. Suvorova and S. Wang, ACS Appl. Mater. Interfaces, 2014, 6, 16745–16754.
10. X. Quan, X. Ruan, H. Zhao, S. Chen and Y. Zhao, Environ. Pollut., 2007, 147, 409–414.
11. Y. Tu, Y. Xiong, S. Tian, L. Kong and C. Descorme, J. Hazard. Mater., 2014, 276, 88–96.
12. R. C. Martins, R. J. G. Lopes and R. M. Quinta-Ferreira, Chem. Eng. J., 2010, 165, 678–685.
13. K. Ikehata and M. G. El-Din, Ozone: Sci. Eng., 2005, 27, 83–114.
14. L. Zhao, J. Ma and X. Zhai, Environ. Sci. Technol., 2009, 43, 5094–5099.
15. C. A. Orge, M. F. R. Pereira and J. L. Faria, Appl. Catal., B, 2015, 174, 113–119.
16. J. D. Xiao, Y. B. Xie, F. Nawaz, S. Jin, F. Duan, M. J. Li and H. B. Cao, Appl. Catal., B, 2016, 181, 420–428.
17. Y. X. Wang, Y. B. Xie, H. Q. Sun, J. D. Xiao, H. B. Cao and S. B. Wang, J. Hazard. Mater., 2016, 301, 56–64.
18. M. Sui, S. Xing, L. Sheng, S. Huang and H. Guo, J. Hazard. Mater., 2012, 227–228, 227–236.
19. F. Nawaz, Y. B. Xie, H. B. Cao, J. D. Xiao, Y. Q. Wang, X. H. Zhang, M. J. Li and F. Duan, Catal. Today, 2015, 258, 595–601.
20. R. Yin, W. Guo, X. Zhou, H. Zheng, J. Du, Q. Wu, J. Chang and N. Ren, RSC Adv., 2016, 6, 19265–19270.
21. T. Zhang and J. Ma, J. Mol. Catal. A: Chem., 2008, 279, 82–89.
22. W. Qin, X. Li and J. Qi, Langmuir, 2009, 25, 8001–8011.
23. X. Li, Q. Zhang, L. Tang, P. Lu, F. Sun and L. Li, J. Hazard. Mater., 2009, 163, 115–120.
24. S. T. Xing, X. Y. Lu, L. M. Ren and Z. C. Ma, RSC Adv., 2015, 5, 60279–60285.
25. H. Zhao, Y. M. Dong, P. P. Jiang, G. L. Wang, J. J. Zhang and K. Li, Catal.: Sci. Technol., 2014, 4, 494–501.
26. H. Zhao, Y. M. Dong, P. P. Jiang, G. L. Wang, J. J. Zhang and C. Zhang, Chem. Eng. J., 2015, 260, 623–630.
27. Z. C. Ma, L. Zhu, X. Y. Lu, S. T. Xing, Y. S. Wu and Y. Z. Gao, Sep. Purif. Technol., 2014, 133, 357–364.
28. A. H. Lv, C. Hu, Y. L. Nie and J. H. Qu, Appl. Catal., B, 2010, 100, 62–67.
29. L. Zhang, J. Lian, L. Wu, Z. Duan, J. Jiang and L. Zhao, Langmuir, 2014, 30, 7006–7013.
30. E. J. Kim, C. S. Lee, Y. Y. Chang and Y. S. Chang, ACS Appl. Mater. Interfaces, 2013, 5, 9628–9634.
31. S. W. Zhang, Q. H. Fan, H. H. Gao, Y. S. Huang, X. Liu, J. X. Li, X. J. Xu and X. K. Wang, J. Mater. Chem. A, 2016, 4, 1414–1422.
32. J. Peng, L. N. Feng, K. Zhang, J. J. Li, L. P. Jiang and J. J. Zhu, Chem.–Eur. J., 2011, 17, 10916–10923.
33. J. Z. Zhao, Z. L. Tao, J. Liang and J. Chen, Cryst. Growth Des., 2008, 8, 2799–2805.
34. C. Hui, C. M. Shen, T. Z. Yang, L. H. Bao, J. F. Tian, H. Ding, C. Li and H. J. Gao, J. Phys. Chem. C, 2008, 112, 11336–11339.
35. H. B. Lin, J. N. Hu, H. B. Rong, Y. M. Zhang, S. W. Mai, L. D. Xing, M. Q. Xu, X. P. Li and W. S. Li, J. Mater. Chem. A, 2014, 2, 9272–9279.
36. D. P. Dubal, W. B. Kim and C. D. Lokhande, J. Phys. Chem. Solids, 2012, 73, 18–24.
37. J. Staehelin and J. Hoigne, Environ. Sci. Technol., 1982, 16, 676–681.
38. J. Ma, M. Sui, T. Zhang and C. Guan, Water Res., 2005, 39, 779–786.
39. H. Einaga and S. Futamura, Appl. Catal., B, 2005, 60, 49–55.
40. R. C. Martins and R. M. Quinta-Ferreira, Appl. Catal., B, 2009, 90, 268–277.
41. J. D. Xiao, Y. B. Xie, Q. Z. Han, H. B. Cao, Y. X. Wang, F. Nawaz and F. Duan, J. Hazard. Mater., 2016, 304, 126–133.
42. Y. Dong, L. Wu, G. Wang, H. Zhao, P. Jiang and C. Feng, Bull. Korean Chem. Soc., 2013, 34, 3227.
43. J. J. Lin, A. Kawai and T. Nakajima, Appl. Catal., B, 2002, 39, 157–165.
44. A. H. Lv, C. Hu, Y. L. Nie and J. H. Qu, Appl. Catal., B, 2012, 117, 246–252.
45. L. J. Xu and J. L. Wang, Environ. Sci. Technol., 2012, 46, 10145–10153.
46. H. Zhao, Y. Dong, P. Jiang, G. Wang, J. Zhang, K. Li and C. Feng, New J. Chem., 2014, 38, 1743–1750.
47. M. Q. Yang, Y. Zhang, N. Zhang, Z. R. Tang and Y. J. Xu, Sci. Rep., 2013, 3, 3314.
48. M. G. Antoniou and D. D. Dionysiou, Catal. Today, 2007, 124, 215–223.

---