Catalytic ozonation of phenol by magnetic Mn0.7Ce0.3Ox/CNT@Fe3C

A high-efficient and stable catalyst is highly desired to catalyze ozone for refractory organic pollutants removal. In this work, Mn-Ce bimetallic oxide loaded CNT@Fe3C (Mn0.7Ce0.3Ox/CNT@Fe3C) was synthesized with Mn0.7Ce0.3Ox as the active component and magnetic CNT@Fe3C as the support. The catalytic performance of Mn0.7Ce0.3Ox/CNT@Fe3C towards catalytic ozonation was evaluated. The TOC removal efficiency of phenol degradation after 45 min of reaction was 98%, which was 1.5 times and 1.8 times that of monometallic CeO2/CNT@Fe3C (65%) and MnxOy/CNT@Fe3C (54%), respectively. The synthesized-Mn0.7Ce0.3Ox/CNT@Fe3C possessed good reusability during five successive cycles and remained efficient over a wide range of pH 4.2–8.3. The results of EPR measurements and quenching experiments demonstrated that hydroxyl radicals (·OH) were the dominant reactive oxidation species (ROS) for phenol mineralization in the Mn0.7Ce0.3Ox/CNT@Fe3C/O3 system. Moreover, the magnetic Mn0.7Ce0.3Ox/CNT@Fe3C is easily recovered and reused.


Introduction
Refractory organic pollutants in water generally exhibit low biodegradability, high toxicity, and ease of being enriched in organisms, which threaten the safety of human health and the ecosystem. Industrial wastewaters usually contain various refractory organic pollutants (such as phenolic compounds, chloro-organics, PAHs, etc). However, these compounds are difficult to be effectively removed by conventional biological wastewater treatment technologies. As one of the advanced oxidation processes (AOPs), heterogeneous catalytic ozonation is attracting a great deal of attention because of its high removal efficiency of refractory organic pollutants in water treatment [1][2][3].
The catalyst is crucial in determining the efficiency of removing organic pollutants in heterogeneous catalytic ozonation. Various catalysts have been constructed to enhance the conversion of O 3 to produce reactive species and degrade organic contaminants. Compared with homogeneous catalysts, heterogeneous catalysts are easier to be recycled and can be used under a wide range of pH [4,5]. Consequently, heterogeneous catalysts are more suitable for the actual water treatment processes. For example, Tahir Muhmood et al [6][7][8] successfully prepared ZrO 2 /Fe modified hollow-C 3 N 4 , graphene nanoplatelets/graphitic carbon nitride heterojunctions, and spherical-graphitic carbon nitride for photo-degradation of various organic pollutants. Up to present, the studied ozonation catalysts are mainly transition metal-based (such as Mn, Fe, Ce, Cu, etc) materials [9][10][11][12][13]. Among them, the catalysts based on Mn usually show good activity for catalytic ozonation. However, their stability during catalytic ozonation remains poor. The leaching of Mn ions could lead to the reduction of catalytic activity and cause secondary pollution of water, hindering the practical application of heterogeneous catalytic ozonation [14,15]. Consequently, developing highly efficient and stable Mn-based catalysts is becoming a priority for catalytic ozonation of organic pollutants.
Constructing bimetallic oxide-based materials is recognized as an effective approach to generating a synergistic effect, which thereafter improves the activity and stability of materials [16][17][18]. Shan and co-workers [19] employed Ce-Ti-Zr ternary oxide to degrade oxalic acid in catalytic ozonation. It has been demonstrated that the Ce-O-Ti bond in Ce-Ti bimetallic oxide can obviously increase the proportion of Ce 3+ , which could Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. enhance the decomposition of ozone to generate ROS and degrade organic pollutants. Zhang et al [20] found that loading PdO on CeO 2 can remarkably improve the catalytic efficiency of PdO for catalytic ozonation. The synergetic effect between PdO and CeO 2 makes the generated atomic oxygen reacts preferentially with the surface Ce-oxalate complex, which thereafter results in the high activity of PdO/CeO 2 . Feng and co-workers [21] investigated the catalytic performance of bimetallic oxide (CuFeO 2 ) towards peroxymonosulfate activation. It has been found that the stability of bimetallic CuFeO 2 was significantly higher than that of the monometallic (Cu or Fe) oxide because of the synergistic effects. Consequently, it was deduced that the construction of Mn-X bimetallic oxide could generate a synergistic effect and improve the activity and stability of a Mn-based catalyst in catalytic ozonation.
It is well-known that Ce-based catalysts possess excellent stability for catalytic ozonation. In this work, Mn-Ce bimetallic oxide was synthesized as the active component by a redox process between KMnO 4 and Ce(NO 3 ) 3 · 6H 2 O. Moreover, our previous study confirmed that CNT can accelerate the reduction of Ce 4+ to Ce 3+ in the CeO 2 /CNTs/O 3 system, and Ce 3+ catalyzes the conversion of O 3 to produce ·OH to degrade organic pollutants [22]. Employing CNT as a support could facilitate electron transfer and increase the exposure of active sites, which thereby enhance the catalytic efficiency of Mn-Ce bimetallic oxide. In addition, considering magnetic materials are easily recycled, the magnetic CNT@Fe 3 C was prepared as the support for Mn-Ce bimetallic oxide.

Preparation of catalysts
The magnetic CNT@Fe 3 C was synthesized in terms of a method reported by Yang et al [23]. In brief, 7.5 ml of melamine solution (0.1 g l -1 ), 7.5 ml of Fe(NO 3 ) 3 solution (1.0 wt%) and 5.0 ml of P123 solution (0.10 g ml -1 ) were first mixed and stirred for 2 h. The mixed solution was then heated under 80°C with various stirring to evaporate the solvent. To obtain the magnetic CNT@Fe 3 C, the resulting powder was ground and calcined at 800°C for 1 h under Ar atmosphere with a heating rate of 3°C min -1 . The Mn-Ce bimetallic oxide loaded CNT@Fe 3 C (MnCeO/CNT@Fe 3 C) was synthesized by a redox process between KMnO 4 and Ce(NO 3 ) 3 · 6H 2 O. Typically, 0.2 g of synthesized-CNT@Fe 3 C was first dispersed into 45 ml of water by ultrasonic treatment. Then, a mixture of Ce(NO 3 ) 3 and KMnO 4 solution (Mn/Ce = 7/3, molar ratio) was added into the aforementioned suspension under vigorous stirring. The solution pH was adjusted to pH 6.0 with NaOH solution (0.1 M) and kept stirring at 60°C for 2 h. The precipitate was obtained by centrifugation and washed by ultrapure water till the filtrate pH ≈ 7. The obtained precipitate was then dried under 60°C for 2 h in a vacuum drying oven. In the calcination process, the Mn-Ce bimetallic oxide loaded CNT@Fe 3 C was obtained by heating at 300°C for 3 h. In addition, Mn/Ce monometallic oxide loaded CNT@Fe 3 C samples, including Mn x O y /CNT@Fe 3 C and CeO 2 /CNT@Fe 3 C, were synthesized using equalvolume impregnation method. The content of Mn and Ce in the precursor during the preparation of Mn x O y /CNT@Fe 3 C and CeO 2 /CNT@Fe 3 C is the same as that of MnCeO/CNT@Fe 3 C.

Characterization of catalysts
The morphology was characterized by an S-4800 field emission scanning electron microscopic (SEM) (Hitachi, Japan). The structure was characterized by an EMYPREAN XRD. The surface chemical composition was characterized using XPS (ESCALAB™ 250Xi). The BET surface areas were determined on a Quantasorb surface area analyzer (Quantachrome Corp., USA). The bulk contents of Mn and Ce of prepared samples were determined by ICP-AES (Optima 200 DV).

Catalytic activity tests
All experiments were conducted in a semi-batch mode reactor at room temperature. A COM-AD O 3 generator (Anseros, GER) was used to in situ produce O 3 from oxygen (O 2 ). A schematic diagram of the experimental setup for the catalytic ozonation is shown in figure 1. In brief, the catalyst was firstly dispersed in phenol solution (20 mg l -1 ) by sonication for 5 min. The aforementioned suspension was then transferred into a 0.5 l semicontinued reactor. And the generated gaseous

Analytical methods
The concentrations of total organic carbon (TOC) were determined by a multi-N/C ®2100 TOC analyzer (Germany). The concentrations of leached Mn and Ce ions in the systems were determined using ICP-AES (Optima 200 DV). A200 EPR spectrometer (Germany) with DMPO used as a spin-trapping agent was employed to evaluate the generated ROS during reaction.

Physicochemical characteristics of Catalysts
The morphologies of the as-synthesized CNT@Fe 3 C and MnCeO/CNT@Fe 3 C was characterized by SEM. As shown in figures 2(a)-(c), the prepared samples clearly illustrate a one-dimensional bamboo-like carbon nanotube structure, which is consistent with the work of Yang et al [23]. And the magnetic property of the prepared MnCeO/CNT@Fe 3 C was tested using a magnet. Figure 2(b) shows two suspensions containing the same concentration of MnCeO/CNT@Fe 3 C. It can be observed that when a magnet was placed outside the right bottle, the powder was quickly gathered on the magnet, while the left suspension without magnet remained dispersed. This phenomenon suggests that the as-synthesized MnCeO/CNT@Fe 3 C possesses a specific magnetic nature, which would make the catalyst easy to be recovered and reused.
The crystal structure of the synthesized materials was detected using XRD. As shown in figure 3(a), the patterns exhibit a typical broad peak located at around 26°, corresponding to the (002) crystal planes of graphite [24]. Additionally, the other peaks could be assigned to the characteristic peaks of Fe 3 C (JCPDS No. 35-0772). The N 2 adsorption-desorption isotherms of synthesized materials are shown in figure 3(b). The BET value of CNT@Fe 3 C is 425.9 m 2 g -1 , while that of CeO 2 /CNT@Fe 3 C, Mn x O y /CNT@Fe 3 C, and MnCeO/CNT@Fe 3 C is 146.1, 155.1 and 141.2 m 2 g -1 , respectively. Apparently, the presence of active components sacrifices some surface area of CNT@Fe 3 C. In addition, the pore structures of different materials are mesoporous, and the pore volume of the materials is in the range of 0.28-0.91 cm 3 g -1 . The bulk contents of Mn and Ce in MnCeO/CNT@Fe 3 C were determined to be 7.3 wt% and 7.6 wt%, respectively. The calculated atomic ratio of Mn/Ce was approximately 0.7/0.3. Hence, the chemical formula of the synthesized MnCeO/CNT@Fe 3 C can be further described as Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C.
The surface chemical components were analyzed by XPS measurements. As shown in figure 4(a), the C 1s, O 1s, and Fe 2p peaks can be observed in the XPS spectra of CNT@Fe 3 C, CeO 2 /CNT@Fe 3 C, Mn x O y /CNT@Fe 3 C, and Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C. For Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C, two typical peaks of Ce 3d and Mn 2p also can be observed, implying C, O, Fe, Ce, and Mn on the surface of Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C. The high-resolution Ce 3d spectrum is fitted and divided into ten spin-orbit peaks ( figure 4(b)), which are v, v 0 , v′, v″, v″′ peaks and u, u 0 , u′, u″, u″′ peaks. The v and u peaks represent the splitting peaks generated by Ce 3d 3/2 and Ce 3d 5/2 spin orbits, respectively [25,26]. The peaks of v 0 , v′, v″′, and u′ are the characteristic splitting peaks of Ce 3+ , while the peaks of v, v″, u 0 , u, u″, and u″′ corresponded to the characteristic splitting peaks of Ce 4+ . The characteristic peak areas are used for the quantitative analysis of the valence of Ce on the surface of Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C. The valence of Ce is calculated to be '+3' and '+4', and the atomic ratio of Ce 3+ /Ce 4+ is 0.66.     figure 5(c)). As the O 3 concentration increased from 10 m g l -1 to 14 m g l -1 , the TOC removal efficiency improved from 87% to 96%.
The impact of initial pH on TOC removal in the Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C/O 3 system was also investigated. The TOC removal efficiencies display rare changes in a wide solution pH range of 4.2-8.3 ( figure 5(d)). In addition, it could be observed that the TOC removal efficiencies at pH 6.2 and 8.3 are slightly higher than that of pH 4.2. The surface oxygen groups (-OH) are identified as the active sites of metal-based materials during catalytic ozonation. For the synthesized Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C, the pH zpc is determined to be 5.7, which means the surface of Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C is positively charged at pH 6.2 and 8.3 [29]. This result indicates the surface Me-OH 2 + groups could contribute to the high catalytic efficiency of Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C towards catalytic ozonation. This finding is in line with the work of Zhao et al [29] To further explore the feasibility of Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C towards catalytic ozonation, its reusability was studied. For each cycling experiment, the used-Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C was recycled by a magnet, washed with ultrapure water, and dried at 80°C for 6 h. The catalytic efficiency of Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C exhibits no significant reduction after five cycling runs, which could still reach 92% at the fifth run ( figure 6). This result demonstrates that the synthesized Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C possesses good stability and reusability towards catalytic ozonation, suggesting its promising potential in actual wastewater treatment.

Catalytic ozonation mechanism
3.3.1. Reactive oxygen species in the Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C/O 3 system To investigate the generated reactive oxygen species involved in the Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C/O 3 system, EPR measurements were performed. As generally recognized, Mn/Ce -based catalysts can accelerate the conversion of O 3 into ·OH for catalytic ozonation [30,31]. For ozonation alone, no obvious peaks were observed ( figure 7(a)). This result indicates that no detectable radicals were produced in ozonation alone. Nevertheless, when the Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C was added, a 4-fold strong peaks with a relative intensity ratio of 1:2:2:1, which are assigned to the characteristic peaks of DMPO-·OH adduct, could be observed in the Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C/O 3 system [32]. It reveals that Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C could promote the conversion of O 3 to form ·OH in catalytic ozonation, resulting in the enhancement of phenol removal.
Moreover, radical quenching experiment was carried out to verify the contribution of ·OH to TOC removal efficiency in the Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C/O 3 system. Bicarbonate was added to the solution to capture the generated ·OH. As shown in figure 7(b), when 300 mg l -1 bicarbonate was added, the mineralization efficiency of phenol exhibits no noticeable reduction in ozonation alone. This phenomenon, along with EPR result displays that the direct oxidation of ozone molecules plays a crucial role in phenol mineralization during ozonation alone. Nevertheless, for the Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C/O 3 system, with the addition of 300 mg l -1 bicarbonate, the TOC removal efficiencies were significantly depressed to 50%. This result reveals that ·OH is the dominant ROS for the Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C/O 3 system. Notably, for the Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C/O 3 system, the phenol mineralization efficiencies were significantly inhibited with the presence of bicarbonate, but they still  obviously higher than those of ozonation alone. This result indicates that a nonradical pathway is accompanied by ·OH oxidation in phenol mineralization in the Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C/O 3 system. The nonradical pathway could come from the following two aspects. The first relies on the direct oxidation of ozone molecules: ozone possesses a strong oxidation capacity of E 0 = 2.07 V, which can directly oxidize various organic pollutants. The second could be contributed by the oxidation of high valence Mn and Ce: for example, the standard redox potentials of Mn 4+ and Ce 4+ are 1.23 V and 1.44 V, respectively, which can degrade some organic pollutants adsorbed on the catalyst surface and enhance the removal of TOC.

Mechanism for the synergistic effect
For better disclosing the high efficiency of Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C towards catalytic ozonation, XPS measurements were carried out. The surface chemical composition of Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C after reaction was analyzed. As generally recognized, the decomposition of ozone converts into ROS catalyzed by Ce-based materials is accompanied by the conversion of Ce 3+ to Ce 4+ , and the conversion rate of Ce 4+ /Ce 3+ cycling is regarded as a pivotal step in catalytic ozonation. Figure 8(a) shows the Ce 3d XPS spectrum of Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C after reaction (used-Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C). The ratio of Ce 3+ to Ce 4+ was estimated by the area ratio of splitting peaks. The ratio is calculated to be 0.70, which is slightly higher than that of fresh-Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C. Our previous study investigated the catalytic efficiency of CeO 2 /CNT towards catalytic ozonation of various organic pollutants. The work revealed that the abundant π electrons on carbon nanotubes accelerate the electron transfer for catalytic ozonation and thereby bring the facilitated conversion of Ce 4+ to Ce 3+ [22]. Compared with CeO 2 /CNT, the reduction of Ce 3+ from Ce 4+ on Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C is further accelerated, which might be caused by the interaction between different metal oxides after the introduction of the second metal component 'Mn' [19]. Chen et al [33] found that in the MnFe 2 O 4 /O 3 system, in addition to 'Fe 3+ /Fe 2+ ' cycling and 'Mn 3+ /Mn 2+ ' cycling, 'Mn 3+ /Fe 2+ ' can also form a redox cycle, thus further improving the catalytic efficiency. For Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C, the redox potential of Mn 4+ (E Mn 4+/Mn 3+ = 0.95 V, E Mn 4+/Mn 2+ = 1.23 V) is lower than that of Ce 4+ (E Ce 4+/Ce 3+ = 1.44 V), which means the redox of 'Ce 4+ + Mn 3+ → Ce 3+ + Mn 4+ ' and 'Ce 4+ + Mn 2+ → Ce 3+ + Mn 4+ ' are thermodynamically feasible. Therefore, the reduction of Ce 4+ to Ce 3+ on Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C can be achieved not only by obtaining electrons from pollutants and oxygen vacancies, but also through the redox reaction between Ce and Mn. All these could facilitate the redox cycling of Ce 3+ /Ce 4+ , which is conducive to catalytic ozone generation of ·OH and improve the mineralization efficiency of phenol.
Similarly, the degradation of pollutants by Mn-based materials catalyzed ozonation is accompanied by 'Mn 4+ /Mn 3+ ' cycling and 'Mn 3+ /Mn 2+ ' cycling [31]. To gain insight into the change of Mn on the surface of Mn x O y /CNT@Fe 3 C during catalytic ozonation, XPS analyses of Mn for fresh-Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C and used-Mn x O y /CNT@Fe 3 C were also carried out. As shown in figure 8(b), for Mn x O y /CNT@Fe 3 C, the AOS of Mn decreased from '+3.19' to '+2.62' after 45 min of reaction. Andreozzi and co-workers [34] revealed that Mn 3+ could degrade the oxalic adsorbed on the surface of manganese dioxide, which thereafter resulted in the reduction of Mn 3+ to Mn 2+ . Therefore, the reduction of Mn could be attributed to the reaction of Mn with adsorbed organic pollutants. It is notable that compared with Mn x O y /CNT@Fe 3 C, the average oxidation valence of Mn on Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C (+3.05) shows no obvious change after the reaction. The redox of 'Ce 4+ + Mn 3+ →Ce 3+ + Mn 4+ ' and 'Ce 4+ + Mn 2+ →Ce 3+ + Mn 4+ ' in the Mn 0.7 Ce 0.3 O x /CNT@Fe 3 C/O 3 system could result in this phenomenon.