*Article* **Catalytic Ozonation of Atrazine Enhanced by Mesoporous CeO2: Morphology, Performance and Intermediates**

**Jianlin Zhang 1,2 , Tao Zhuang <sup>2</sup> , Shanjun Liu <sup>2</sup> , Shan Sun <sup>2</sup> , Yongxin Wang <sup>2</sup> , Xinyu Liu <sup>3</sup> , Jin Wang <sup>4</sup> and Rutao Liu 1,\***


**Abstract:** Heterogeneous catalytic ozonation is an alternative approach for the removal of refractory pollutants from water, and the fabrication of mesoporous materials with high dispersibility would enhance the catalytic efficiency. A mesoporous CeO<sup>2</sup> was prepared by the nanocasting method with SBA-15 as a hard template, and was investigated in the catalytic ozonation of atrazine. The synthetical CeO<sup>2</sup> nanorods have a specific surface area of 95.08 m2/g, a diameter of 10.16 nm, and a spacing of 2.18 nm. The removal rate of atrazine was 85.5%, 64.8%, and 46.4% in the order of catalytic ozonation by synthetical CeO<sup>2</sup> > single ozonation > catalytic ozonation by commercial CeO<sup>2</sup> , respectively. The superior activity of the synthetical CeO<sup>2</sup> could be attributed to the well-ordered mesoporous structure, the high surface area, and the redox Ce3+/Ce4+ cycling. Moreover, eight organic intermediates were identified after one minute of catalytic ozonation of atrazine, and the cleavage of the ethylamino group was proposed as the main pathway of atrazine degradation.

**Keywords:** catalytic ozonation; atrazine; CeO<sup>2</sup> ; mesoporous

### **1. Introduction**

Atrazine as a chlorinated triazine herbicide has been widely used in agriculture, especially for the control of crops and grassy weeds [1]. It has a water solubility of 33 mg L−<sup>1</sup> at 25 ◦C, an organic carbon partition coefficient value of 100 L kg−<sup>1</sup> , and a half-life ranging from 21 days to two years, which makes it possible to leach into the aquatic ecosystem and persist for a long time [2]. Atrazine, even at low concentrations in water bodies, could cause serious negative effects on human health through the accumulation of it in the food chain [3], and has been listed as a priority substance by the European Commission. Notably, the detected concentration of atrazine in surface water and groundwater in Beijing ranged from 0.12 to 5.16 µg L−<sup>1</sup> , which exceeded the allowable level of 3 µg L−<sup>1</sup> for drinking water [4]. However, the traditional water purification methods are obviously not enough to remove atrazine from water efficiently and completely [5], and the by-products such as halogens might cause a more serious ecological impact [6]. Therefore, it is an urgent requirement to develop an effective removal method for reducing its environmental exposure.

Besides persulfate oxidation [7], photocatalytic oxidation [8], and electrochemical oxidation [9], heterogeneous catalytic ozonation technology is a promising technology for the removal of refractory pollutants [10], which could generate highly reactive hydroxyl radicals with a redox potential of 2.80 eV and rapidly and non-selectively degrade or even mineralize antibiotics with 106–10<sup>10</sup> L mol−<sup>1</sup> s −1 reaction rate [11]. Li et al. [12] evaluated the degradation of atrazine by the catalytic ozonation of iron scraps, and the results showed that the mineralization efficiency was as high as 86.6%, and the toxicity was efficiently

**Citation:** Zhang, J.; Zhuang, T.; Liu, S.; Sun, S.; Wang, Y.; Liu, X.; Wang, J.; Liu, R. Catalytic Ozonation of Atrazine Enhanced by Mesoporous CeO2: Morphology, Performance and Intermediates. *Water* **2022**, *14*, 3431. https://doi.org/10.3390/w14213431

Academic Editors: Huijiao Wang, Dionysios (Dion) Demetriou Dionysiou and Yujue Wang

Received: 29 September 2022 Accepted: 25 October 2022 Published: 28 October 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

removed at 30 min. Ye et al. [13] identified that the prepared catalyst could effectively improve the degradation and mineralization rate of atrazine, and the mineralization strongly depended on the solution's pH. Zeghioud et al. [14] concluded that catalytic ozonation coupled with discharge plasma could improve the removal efficiency of refractory organic pollutants. He et al. [15] combined membrane filtration and catalytic ozonation for the removal of atrazine, 99.99% of atrazine could be removed within 40 min and hydroxyl radicals were inferred to be the main reactive oxygen species in the catalytic ozonation process. Three pollutant degradation pathways are normally proposed in the process of heterogeneous catalytic ozonation [16]. Firstly, ozone is adsorbed on the catalyst surface and catalyzed to generate radical oxygen species, which further degrades the organic pollutant adsorbed on the catalyst surface or in solution. Secondly, ozone directly attacks and degrades the organic pollutants adsorbed on the surface of the catalyst or the complex formed by organic pollutants and the catalyst. Thirdly, both ozone and organic pollutants are simultaneously adsorbed and reacted on the surface of the catalyst. Obviously, heterogeneous catalytic processes are really complicated, involving gas phase (ozone), liquid phase (water) and solid phase (catalyst), where catalysts would perform various roles, such as providing active sites for adsorption and catalytic ozonation. Therefore, the textural properties of catalysts, such as surface area, pore volume and porosity, are the keys to catalytic efficiency.

Numerous studies have been devoted to designing and optimizing the properties of catalysts to improve the efficiency of catalytic ozonation. Bing et al. [17] prepared Fe2O3/Al2O3@SBA-15 by wet impregnation and calcination, and characterized that Al2O<sup>3</sup> and Fe2O<sup>3</sup> were highly dispersed on SBA-15, and amounts of surface Lewis acid sites existed due to the substitution of Al3+ for the hydrogen of Si-OH. Yuan et al. [18] synthesized microporous iron silicate with poor crystallinity and found that Fe-Si binary oxides inhibited the futile decomposition of ozone and increased the probability of hydroxyl radicals generated through the reaction between ozone and α-Fe2O3. Our previous work prepared a mesoporous MnO<sup>2</sup> by the nanocasting method and found that the ordered nanorods with a high surface area facilitated mass transport and exposure of active sites [19]. Besides MnO2, CeO<sup>2</sup> is another widely studied heterogeneous catalyst for catalytic ozonation. The cerium redox pair (Ce(III) and Ce(IV)) performs a synergetic effect on catalytic ozonation, where Ce(III) could catalyze ozonolysis to generate hydroxyl radicals and would be regenerated through the reduction of Ce(IV) [20]. Notably, the surface basicity, defect density/oxygen vacancies and unsaturated coordination sites of CeO<sup>2</sup> crystal planes exposed are important factors in the catalytic performance, such as the ability of different crystal planes in the order of (100 + 110) > (100) > (111) to enhance the mineralization of nitrophenol [21].

In this study, an ordered mesoporous CeO<sup>2</sup> with a high surface area was prepared with SBA-15 as the hard template by the nanocasting method and used as an efficient catalyst for catalytic ozonation of atrazine. The textural properties of the synthesized samples were characterized by X-ray diffraction(XRD), surface area analysis techniques, scanning electron microscope (SEM), transmission electron microscope (TEM), and Xray photoelectron spectroscopy (XPS). The performance of CeO<sup>2</sup> was investigated by evaluating the degradation of atrazine under various conditions compared with control samples. Moreover, the degradation pathways were assessed through the identification of intermediate products. This study may provide an approach to effectively remove refractory pollutants in the process of wastewater treatment.

### **2. Materials and Methods**

### *2.1. Materials*

SBA-15 was purchased from XFNANO Co., Ltd. (Nanjing, China). The commercial CeO<sup>2</sup> and Ce(NO3)3•6H2O were supplied by Sinoparm Chemical Reagent Co., Ltd. (Shanghai, China). Atrazine was provided by the laboratory of the government chemist (Germany). NaOH was purchased from Rhawn Reagent (Shanghai, China). H2SO<sup>4</sup> was obtained from Yantai Far East Fine Chemical Co., Ltd. (Yantai, China). All chemicals

used were of analytical grade without further purification. Deionized water (18 MΩcm) produced by the Milli-Q purification system (Millipore, USA) was used to prepare solutions. Ozone was supplied by an ozone generator (COM-AD-01, ANSEROS), and the ozone stock solution was obtained by bubbling gaseous ozone into a 1 L cylindrical reactor filled with pure water for 30 min.

### *2.2. Preparation of CeO<sup>2</sup>*

The synthesis of mesoporous CeO<sup>2</sup> was based on a previously reported method [19]. Concisely, 4.51 g of SBA-15 was dispersed in an ethanol solution (15 mL of ultrapure water and 90 mL of ethanol) under stirring with a magnetic bar at room temperature. A total of 13 g of Ce(NO3)3•6H2O was added into the suspended solution and kept stirring for 2 h. The solid was filtered from the suspension, dried for 5 h at 105 ◦C, and subsequently calcined at 300 ◦C for 3 h. The CeO2/SBA-15 complex obtained was rinsed with 2M NaOH, then washed to neutral with ultrapure water and dried for 3 h at 105 ◦C.

### *2.3. Experimental Procedure*

The experimental procedure was conducted in batch mode in a 0.6 L cylindrical reactor with a diameter of 80 mm and a height of 120 mm at room temperature. Basically, appropriate amounts of catalyst and atrazine stock solution were added to the reactor under continuous magnetic stirring. Then, appropriate ozone stock solutions were added to the reactor to initiate catalytic ozonation (the initial mixing solution contained 50 µg L−<sup>1</sup> atrazine, 20 mg L−<sup>1</sup> catalyst and 1 mg L−<sup>1</sup> O3). The solution of an appropriate volume was sampled from the reactor at certain intervals, filtered with a 0.22 µm prefilter and quenched with a trace of Na2S2O<sup>3</sup> for further analysis. Two samples were repeated for each experiment.

### *2.4. Characterization*

The XRD patterns were recorded in the 2θ range of 5–60◦ using a Rigaku Ultima V diffractometer with Cu Kα as the radiation source. The surface area was measured on an automatic adsorption analyzer (ASAP2460, Micromeritics Instrument Corp., Norcross, GA, USA) and calculated by the Brunauer–Emmett–Teller (BET) method. The structure and morphology of the as-prepared CeO<sup>2</sup> were characterized by transmission electron microscopy (FEI Tecnai G2 F20, NL) and scanning electron microscopy (Zeiss Sigma 300, GER) equipped with an energy-dispersive spectrometer. The elementary compositions of the prepared samples were analyzed by X-ray photo-electron spectroscopy (Thermo Fisher Scientific, Waltham, MA, USA).

The atrazine concentration was detected by high-performance liquid chromatography (LC-20AT, Shimadzu, JPN) equipped with a diode array detector. A reversed-phase C18 column was used under isocratic elution with methanol and phosphoric acid solution (*v*/*v*, 55:45) at 1.0 mL/min at 35 ◦C. The intermediates were identified by an ultra-performance liquid chromatography tandem quadrupole time of flight mass spectrometer (LCMS-9030, Shimadzu). Gaseous O<sup>3</sup> concentrations were monitored using an ozone analyzer (GM-6000- OEM, ANSEROS). Aqueous O<sup>3</sup> concentrations were measured by iodometry and were calculated using the following equation.

$$\mathbf{C\_{(O3)}} = 24000 \mathbf{C\_{(Na2S2O3)}} \mathbf{V\_{(Na2S2O3)}} / \mathbf{V\_{(O3)}} \tag{1}$$

where C(O3) is the aqueous O<sup>3</sup> concentration, mg/L; C(Na2S2O3) is the concentration of Na2S2O<sup>3</sup> standard solution, mol/L; V(Na2S2O3) is the consumed volume of Na2S2O<sup>3</sup> standard solution, mL; V(O3) is the titration volume of the aqueous O3, mL.

### **3. Results and Discussion**

*3.1. Characterization of CeO<sup>2</sup>*

The XRD patterns of the CeO<sup>2</sup> samples are shown in Figure 1. The synthetical and commercial samples both exhibited five distinguished diffraction peaks at 28◦ , 33◦ , 47◦ , 56◦ and 59◦ , which could be indexed to the (111), (200), (220), (311) and (222) reflection planes of CeO<sup>2</sup> (JCPDS No. 78-0694) [22]. The peaks of synthetical CeO<sup>2</sup> were wider and weaker than those of commercial CeO2, which implied that the crystal size of synthetical CeO<sup>2</sup> was smaller than that of commercial CeO2. The average crystal size of the synthetical and commercial CeO<sup>2</sup> was 9.8 nm and 37.4 nm, respectively, through the calculation of the Scherrer formula as below [23].

$$\mathbf{D} = \mathbf{K}\lambda / \beta \cos \Theta \tag{2}$$

where D is the average crystal size; K is a constant of 0.89; λ is the X-ray wavelength of 0.154 nm; β is the full width at half maximum of the diffraction peak; θ is the diffraction angle.

**Figure 1.** XRD pattern of CeO<sup>2</sup> samples.

The BET adsorption-desorption isotherms were measured to analyze the specific surface area and pore information of CeO2. In Figure 2a, the hysteresis loop of the synthetical CeO<sup>2</sup> corresponded to type IV, indicating the existence of mesopores. The commercial CeO<sup>2</sup> might be nonporous or microporous since its hysteresis loop belonged to type I. Moreover, the specific surface area of the synthetical and commercial CeO<sup>2</sup> shown in Table S1 was 95.08 m2/g and 1.44 m2/g, respectively, and the specific surface area of the synthetical CeO<sup>2</sup> was significantly higher than that of the commercial CeO2, which implied that the synthetical CeO<sup>2</sup> would provide more catalytic sites and facilitate catalytic ozonation as it could offer more active site exposure possibilities compared to the commercial CeO<sup>2</sup> of the same mass. Figure 2b showed that the synthetical CeO<sup>2</sup> had a relatively concentrated pore distribution while the commercial CeO<sup>2</sup> was almost non-porous, which echoed the analysis of BET adsorption-desorption isotherms and implied that the synthetical CeO<sup>2</sup> might be of ordered mesopores.

The morphologies of CeO<sup>2</sup> were characterized by SEM, and the images are shown in Figure 3. Apparently, the CeO<sup>2</sup> samples presented different degrees of aggregated morphology, and the synthetical CeO<sup>2</sup> seen in Figure 3a,c exhibited granular while the commercial CeO<sup>2</sup> seen in Figure 3b,d exhibited layered, which implied that the surface area of synthetical CeO<sup>2</sup> would be much higher than that of commercial CeO2. Figure 4 displays the TEM images of CeO<sup>2</sup> within different nanometer dimensions. Through the comparison of samples, the synthetical CeO<sup>2</sup> exhibited higher dispersibility than the commercial CeO2, which was consistent with the SEM results. Moreover, the synthetical CeO<sup>2</sup> was a nanorod with well-ordered channels, while the commercial CeO<sup>2</sup> was a nanocube with high agglomeration. Further, the nanorod was 10.16 nm in diameter, and the distance between each channel was 2.18 nm, implying the synthetical CeO<sup>2</sup> had an ordered mesoporous structure and would have large active sites exposed. The synthetical CeO<sup>2</sup> dominantly exposed (111) facets with an interplanar spacing of 0.31 nm, while the commercial CeO<sup>2</sup> mainly exposed (200) facets with an interplanar spacing of 0.19 nm. The lattice fringes revealed the high crystallinity of CeO2, which further proved the results of XRD.

**Figure 2.** Nitrogen adsorption-desorption isotherms (**a**) and pore size distribution (**b**) of CeO<sup>2</sup> .

**Figure 3.** SEM images of synthetical (**a**,**c**) and commercial (**b**,**d**) CeO2..

**Figure 4.** TEM images of synthetical (**a**–**c**) and commercial (**d**–**f**) CeO<sup>2</sup> within different nanometer dimensions.

The redox couple between Ce3+/Ce4+ or O2−/O<sup>3</sup> plays a crucial role in enhancing the catalytic activity as the electrons from Ce3+ would be transferred to O3, which decomposes O<sup>3</sup> into reactive oxygen species, and the electrons from O2<sup>−</sup> would reduce the Ce4+ to Ce3+ [21]. The chemical states of Ce/O and their concentrations on the sample surface were analyzed from the XPS spectra. As illustrated in Figure 5a, the Ce 3d XPS spectra consisted of two series of spin-orbit lines, where v represented the Ce 3d3/2 spin-orbit splitting peak while u represented the Ce 3d5/2 spin-orbit splitting peak. The spin-orbit was well-fitted with ten peaks through deconvolution [24], and the energy positions of these ten spin-orbit components were summarized in Table S2. Taking the synthetical sample as an example, three pairs of peaks, including v (882.46 eV) and u (900.98 eV), v00 (888.87 eV) and u00 (907.61 eV), and v000(898.18 eV) and u000 (916.83 eV), were related to the characteristic peaks of Ce4+. The remaining two pairs of peaks marked as v<sup>0</sup> (880.38 eV) and u<sup>0</sup> (898.38 eV) and v0 (884.84 eV) and u0 (903.05 eV) belonged to the characteristic peaks of Ce3+ [24]. The results indicated that cerium existed in CeO<sup>2</sup> in two valence states [25], Ce3+ and Ce4+, and the atomic ratio of Ce3+ to Ce4+ was 10.14% and 12.59% for the synthetic and commercial CeO2, respectively. The emergence of multivalent cerium would facilitate the charge transfer and the production of reactive oxygen species during the catalytic ozone process. The presence of Ce3+ would introduce the occurrence of oxygen vacancies on CeO2. As illustrated in Figure 5b, the O 1s XPS spectra exhibited a strong peak with a shoulder, and might be related to four kinds of surface oxygen states. The fitting peak at 529.47–529.51 eV could be attributed to the lattice oxygenis [26], the one at 530.51 eV could be assigned as the surface oxygen from the hydroxyl group [27], and the adjacent peaks at 531.46–531.47 eV and 531.81–531.97 eV should be characteristic of subsurface O−/O<sup>2</sup> 2− species, which belong to oxygen substances chemically adsorbed at oxygen vacancies and would improve the catalytic activity [28].

**Figure 5.** XPS spectra of Ce 3d (**a**) and O 1s (**b**) of CeO2..

### *3.2. Catalytic Ozonation of Atrazine*

To evaluate the catalytic superiority of the synthetical CeO<sup>2</sup> for ozonation, the atrazine degradation experiments were carried out under different conditions. The adsorption was performed for 10 min before atrazine degradation. As shown in Figure 6a, the adsorption amount of synthetical CeO<sup>2</sup> was as high as 9.9%, while that of the commercial one was only 4.0%. The good adsorption capacity is attributed to the high surface area of synthetical CeO2, which would be favorable for the reaction of atrazine with hydroxyl radical generated on the catalyst surface. The atrazine removal rate with synthetical CeO<sup>2</sup> could reach up to 85.5% after 10 min, while the removal rates with commercial CeO<sup>2</sup> and ozone alone were 64.8% and 46.4%, respectively. The performance of synthetical CeO<sup>2</sup> on catalytic ozonation of atrazine was statistically (*p* < 0.05) different from and better than that of commercial CeO<sup>2</sup> or ozone alone, indicating that the synthetical CeO<sup>2</sup> effectively enhanced the catalytic ozonation and its catalytic activity was obviously superior to commercial CeO2. The nano-ZnO [29] and mesoporous Fe3O<sup>4</sup> [30] were also prepared as ozonation catalysts to enhance the degradation of atrazine, and in contrast, the synthetical CeO<sup>2</sup> could obtain the high removal efficiency of atrazine at a dose of less than an order of magnitude. To further quantitatively analyze the catalytic activity of CeO2, the rate constants were calculated according to the pseudo-first-order kinetic equation [13]:

$$-\ln(\mathbf{C}\_t/\mathbf{C}\_0) = k\_{obs}\mathbf{t} \tag{3}$$

where C<sup>t</sup> is the atrazine concentration at the reaction time t, C<sup>0</sup> is the initial atrazine concentration, and *kobs* is the pseudo-first-order constant. As shown in Figure 6b, the apparent rate constant for ozone alone was 0.074 min−<sup>1</sup> , while the apparent rate constant increased to 0.175 min−<sup>1</sup> and 0.097 min−<sup>1</sup> after the addition of synthetical and commercial CeO2, which demonstrated that the applied CeO<sup>2</sup> could promote the removal of atrazine. Notably, the apparent rate constant of synthetical CeO<sup>2</sup> was significantly higher than that of commercial CeO2, which was associated with the superior aforementioned properties such as mesoporous structure, high surface area, as well as good dispersibility.

**Figure 6.** The curves (**a**) and pseudo-first-order plots (**b**) of atrazine degradation in different processes (atrazine concentration = 50 µg/L; catalyst dose = 20 mg/L; ozone concentration = 1 mg/L; initial pH = 7).

Combined with the experimental results and previous research [31], the atrazine degradation was a relatively complex process during the catalytic ozonation, which included adsorption, direct oxidation by molecular ozone and indirect oxidation by hydroxyl radical. The catalytic mechanism of synthetical CeO<sup>2</sup> on the ozonation of atrazine could be summarized as follows:

$$\text{CeO}\_2 + \text{Atrazine}\_{\text{(aqu)}} \rightarrow \text{CeO}\_2 \equiv \text{Atrazine}\_{\text{(ads)}} \tag{4}$$

$$\rm Ca^{3+} + O\_3 + H^+ \to Ca^{4+} + HO\_3\bullet \tag{5}$$

$$\bullet \text{ HO}\_3\bullet \rightarrow \bullet \text{OH} + \text{O}\_2 \tag{6}$$

$$\bullet \text{OH} + \text{O}\_3 \rightarrow \text{O}\_2^- + \text{HO}\_2 \\ \bullet \tag{7}$$

$$\text{O}\_2\text{}^- + \text{O}\_3 \rightarrow \text{O}\_3\text{}^- + \text{O}\_2 \tag{8}$$

$$\text{O}\_3\text{}^- + \text{H}^+ \rightarrow \text{HO}\_3\text{}\bullet \tag{9}$$

$$\text{HO}\_3\bullet \rightarrow \bullet \text{OH} + \text{O}\_2 \tag{10}$$

$$\text{Ce}^{4+} + \text{e}^{-} \rightarrow \text{Ce}^{3+} \tag{11}$$

$$\bullet \text{OH} + \text{CeO}\_2 \equiv \text{Atrazine}\_{\text{(ads)}} \rightarrow \text{byproducts} \tag{12}$$

The pH value has an important influence on the chemical characteristics of the catalyst surface, the charge of ionic organic molecules, and the decomposition of O<sup>3</sup> into reactive oxygen species in the catalytic process [32]. The influence of the initial pH value on the catalytic efficiency would be taken into account in the subsequent evaluation of the environmental influence on the catalytic performance.

### *3.3. Transformation Products and Proposed Pathways*

The molecular structure of atrazine mainly consists of the s-triazine ring, the isopropylamino group, the ethylamino group and the chlorine group, and the nitrogen atom in the amino group and the chlorine atom adjacent to the s-triazine ring are vulnerable to radical oxygen species [15]. It is widely recognized that the degradation of atrazine in catalytic ozonation involves the reaction pathways: dealkylation, dechlorination, deamination and olefination [33]. However, the priority of the pathways was basically not involved in previous studies on reaction pathways since the sampling time for the determination of intermediates was not considered.

The previous results shown in Figure 5a indicated that atrazine could be mostly degraded within 10 min, and half could be degraded in one minute. The sample after one minute of catalytic ozonation could better reflect the early reaction process and was used for the determination of intermediates. Eight organic intermediates with m/z values of 232, 174, 214, 230, 172, 212, 188 and 146 were successfully identified (mass error within 0.5 ppm), and were shown in Table S3 and Figure S1. Based on the intermediates and literature reports, the degradation pathway for the catalytic ozonation of atrazine was proposed in Figure 7. Apparently, the structural changes of atrazine first occurred in the isopropylamino and ethylamino groups. Six intermediate products were detected during the change of the ethylamino group, and their mass spectral peak responses were relatively higher, which could be inferred that the change of the ethylamino group was the main pathway of atrazine degradation. The ethylamino group was attacked by hydroxyl radicals to form the hydroxylated intermediate P1(m/z = 232), and the formed hydroxyl group could be easily attracted to form carbon-centered radicals, which could further add oxygen and rapidly eliminate hydroxyl peroxide radical to generate ketones P4(m/z = 230) [34], and further attacked by hydroxyl radicals, leading to the loss of chlorine and the formation of hydroxylated intermediate P6(m/z = 212). The intermediate P3(m/z = 232) was formed from the olefination of P1, followed by further dealkylation to form the intermediate P5(m/z = 172) [35]. It is noted that the dealkylation reactions were ubiquitous, and the intermediates P2(m/z = 174), P5, P7 and P8(m/z = 146) were the products of dealkylation reactions since the nitrogen atoms on the isopropylamino group and ethylamino group were prone to being attacked, resulting in the heterolytic cleavage of the C–N bond [15]. These intermediates were also observed in other advanced oxidation processes, such as photocatalytic oxidation, persulfate oxidation and Fenton oxidation. All the above intermediates would be further oxidized to low molecular organics and even mineralized to CO<sup>2</sup> and H2O by prolonging the reaction time, which was widely investigated in previous studies [29]. The results indicated that the initiating reaction of catalytic ozonation of atrazine mainly occurred on the ethylamino group and proceeded to the dealkylation reactions.

**Figure 7.** Proposed degradation pathway of atrazine.

### **4. Conclusions**

The mesoporous CeO<sup>2</sup> was successfully obtained by the nanocasting method and investigated for catalytic ozonation of atrazine. Compared with commercial CeO<sup>2</sup> and single ozone, synthetical CeO<sup>2</sup> significantly enhanced the removal of atrazine due to its excellent properties such as high surface area, well-ordered mesoporous structure and redox Ce3+/Ce4+ couple. Eight organic intermediates were identified, and the initiating reaction of catalytic ozonation of atrazine mainly occurred on the ethylamino group, followed by the dealkylation reactions. This study might facilitate the application of mesoporous metal oxide and the understanding of the degradation pathways of organic pollutants in catalytic ozonation.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/w14213431/s1, Figure S1. HPLC-MS/MS chromatogram of atrazine and intermediates; Table S1. Physical properties of the catalysts; Table S2. Energy positions of Ce spin orbit components in different samples (unit: eV); Table S3. UPLC-MS/MS analysis of atrazine and intermediates. Reference [36] is cited in the supplementary material.

**Author Contributions:** J.Z.: Methodology, Software, Investigation, Writing-original draft; T.Z.: Writing—review & editing, Supervision, Data curation; S.L.: Investigation, Writing—review & editing; S.S.: Software, Writing—review & editing; Y.W.: Writing—review & editing; X.L.: Writing review & editing; J.W.: Writing—review & editing; R.L.: Resources, Writing—review & editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The authors would like to acknowledge the support for this work by the Beijing Key Laboratory of Aqueous Typical Pollutant Control and Water Quality Safegurad.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**

