*Article* **Enhanced Degradation of Rhodamine B through Peroxymonosulfate Activated by a Metal Oxide/Carbon Nitride Composite**

**Yuanmin Mo 1,2, Wei Xu 1,2, Xiaoping Zhang 1,2,3,4,\* and Shaoqi Zhou 1,2,3**


**Abstract:** The development of high catalytic performance heterogeneous catalysts such as peroxymonosulfate (PMS) activators is important for the practical remediation of organic pollution caused by Rhodamine B (RhB). An economical and facile synthesized composite of copper–magnesium oxide and carbon nitride (CM/g-C3N<sup>4</sup> ) was prepared by the sol-gel/high-temperature pyrolysis method to activate PMS for RhB degradation. CM/g-C3N<sup>4</sup> exhibited a splendid structure for PMS activation, and the aggregation of copper–magnesium oxide was decreased when it was combined with carbon nitride. The introduction of magnesium oxide and carbon nitride increased the specific surface area and pore volume of CM/g-C3N<sup>4</sup> , providing more reaction sites. The low usage of CM/g-C3N<sup>4</sup> (0.3 g/L) and PMS (1.0 mM) could rapidly degrade 99.88% of 10 mg/L RhB, and the RhB removal efficiency maintained 99.30% after five cycles, showing the superior catalytic performance and reusability of CM/g-C3N<sup>4</sup> . The synergistic effect of copper and g-C3N<sup>4</sup> improved the PMS activation. According to the analyses of EPR and quenching experiments, SO<sup>4</sup> •−, •OH and O<sup>2</sup> •− radicals and <sup>1</sup>O<sup>2</sup> were generated in the activation of PMS, of which SO<sup>4</sup> •− and <sup>1</sup>O<sup>2</sup> were important for RhB removal. The toxicity of RhB was alleviated after being degraded by the CM/g-C3N4/PMS system. This study provides an efficient and promising strategy for removing dyes in water due to the hybrid reaction pathways in the CM/g-C3N4/PMS system.

**Keywords:** RhB degradation; heterogeneous catalyst; copper–magnesium oxide/carbon nitride composite; PMS activation

### **1. Introduction**

Rhodamine B (RhB) poses a potential threat to the aquatic environment and human health because of its high salt content and poor biodegradability [1]. RhB has been extensively applied in the fields of colored glass, as well as the textile and plastic industries. It can cause symptoms such as the red staining of human skin, mild congestion of blood vessels in the brain and breakage of myocardial fibers. According to the chemical carcinogenic risk assessment of the World Health Organization International Agency for Research on Cancer IARC), ingestion or skin contact with RhB will cause acute and chronic poisoning. Therefore, exploring efficient treatment technology is important for removing the highly toxic RhB.

Many methods have emerged to remove RhB, including adsorption [2], ion exchange [3], advanced oxidation processes (AOPs) [4,5], membrane filtration [6] and biodegradation [7].

**Citation:** Mo, Y.; Xu, W.; Zhang, X.; Zhou, S. Enhanced Degradation of Rhodamine B through Peroxymonosulfate Activated by a Metal Oxide/Carbon Nitride Composite. *Water* **2022**, *14*, 2054. https://doi.org/10.3390/w14132054

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

Received: 17 June 2022 Accepted: 25 June 2022 Published: 27 June 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/).

However, owing to its special chemical structural stability, light resistance, corrosion resistance and bacteriostasis, RhB cannot be fully degraded by traditional physical methods and biological technologies [4]. Among them, AOPs have garnered extensive attention for their high efficiency and mineralization in removing pollutants [8–10]. Traditional AOPs use hydroxyl radicals (•OH) to remove pollutants [11]. However, the problems of weak redox potential (1.8–2.7 V), short lifetime (<1 µs) and a narrow range of optimum pH restrict the application of •OH [12]. Recently, AOPs based on sulfate radicals (SO<sup>4</sup> •−) have received more and more attention for their high redox potential (2.5–3.1 V), long lifetime (30–40 ms) and a wide pH range [13]. Peroxymonosulfate (PMS) is one of the widely used oxidants activated by transition metal ions [14,15], heat [16], alkali materials [17], ultrasound [18], UV light sources (254 nm) [19] and heterogeneous catalysts [20] to produce SO<sup>4</sup> •− for pollutant degradation. Among various methods, PMS activated by heterogeneous catalysts is more promising for practical applications because of its mild reaction conditions, low energy consumption and easy recycling [20]. Previous research has employed heterogeneous catalysts including CuO [21], CuCo2O<sup>4</sup> [22], CuO@FeOx@Fe-0 [23] and CuO-Co3O4@CeO<sup>2</sup> [24] to activate PMS for pollutant degradation. All these studies have demonstrated that PMS activated by metal oxide for the generation of active substance is a prospective advanced oxidation process for the decomposition of toxic pollutants.

Copper is one of the transition metals for PMS activation. Copper oxide, as an efficient, economical and low toxic heterogeneous catalyst, has a high catalytic performance [25]. However, the problems with the aggregation and the dissolution of copper ion decrease the activity and limit the practical application of copper oxide. To overcome such drawbacks, Du et al. prepared CuO/rGO for 2,4,6-trichlorophenol removal, which exhibited much better catalytic performance than CuO [26]. Kiain et al. found that CuO@AC could be reused in three cycles, and the leaching of copper ions was negligible [27]. Li et al. synthesized BC-CuO for methylene blue, acid orange 7, atrazine and ciprofloxacin removal in a highly saline system, which achieved a high catalytic performance [28]. All these results indicated that loading the copper oxide on the supporter is an efficient way to improve the catalytic performance of copper oxide. However, the applications of these catalysts were restricted by the complex synthesis, high cost of precursor materials and insufficient reaction sites. Therefore, it is necessary to develop a method of low-cost and facile synthesis and increase the reaction site to enhance the catalytic performance of copper oxide.

Carbon nitride (g-C3N4), as a low-cost and stable material, is commonly applied in many fields [29–31], which could be easily prepared by being calcined in the air, without any chemical or physical pretreatment. The g-C3N<sup>4</sup> can not only be used as a supporter to disperse the metal oxide, with more reaction sites exposed, but also activate PMS to some extent. Therefore, the introduction of g-C3N<sup>4</sup> may decrease the aggregation and improve the catalytic performance of copper oxide. It is reported that the introduction of magnesium oxide can enhance the specific surface area and electron transfer capacity of the catalyst [32–34], which is beneficial for PMS activation. Furthermore, the large quantities of surface basic sites on magnesium oxide can facilitate the formation of an M–OH complex [34], thus increasing the reaction sites between the catalyst and PMS, which is the vital step for PMS activation. Hence, the combination of copper oxide, magnesium oxide and g-C3N<sup>4</sup> is an effective way to improve the catalytic performance of copper oxide. Moreover, it is an economical and eco-friendly catalyst worth exploring.

In order to develop a high catalytic performance catalyst with a low cost and abundant reaction sites for RhB degradation, copper oxide was combined with magnesium oxide and g-C3N<sup>4</sup> to form a composite (CM/g-C3N4). The composite was initially employed as a heterogeneous catalyst for PMS activation, which showed splendid catalytic performance and stability. When compared with previous reports, low-cost magnesium oxide and g-C3N<sup>4</sup> were introduced into the catalyst at the same time for increasing the reaction sites and decreasing the aggregation of copper oxide. Furthermore, the hybrid reaction pathways (radicals and non-radical) were produced in the CM/g-C3N4/PMS system.

In this study, the effects of preparation procedures on the catalytic performance of composite were investigated, and the effects of catalyst dosage, PMS concentration, RhB concentration and initial pH value on the RhB removal were also explored. The degradation intermediates of RhB were analyzed by LC-MS, and the toxicity was predicted by the Toxicity Estimation Software Tool. The activation mechanism of PMS was proposed according to the results of XPS, the EPR test and quenching experiments. The reusability of the catalyst was conducted through five cycle experiments.

### **2. Material and Methods**

### *2.1. Chemicals and Reagents*

The information relating to the chemicals and reagents is described in supplementary material—Text S1.

### *2.2. Preparation of Catalyst*

A preparation of pure g-C3N4: 5 g melamine was placed in a 30 mL quartz crucible and calcined at 550 ◦C (5 ◦C/min) for 4 h in an air atmosphere. After cooling down, it was fully ground with an agate mortar to obtain yellow powder.

Preparation of the copper–magnesium oxide/g-C3N<sup>4</sup> composite (CM/g-C3N4): First, 0.250 M Cu(NO3)2·3H2O and 0.250 M Mg(NO3)2·6H2O were dissolved in 40 mL of deionized water. Then, 0.50 M citric acid was added to the above solution and stirred for 10 min. Thereafter, the mixture was heated at 80 ◦C in the air until the water was completely evaporated to obtain a blue colloid. The obtained colloid was heated to 700 ◦C at a rate of 5 ◦C/min under a nitrogen atmosphere for 4 h to obtain black powder. Finally, the black powder and melamine were ground and mixed at a mass ratio of 1:2, and calcined in the air at 550 ◦C for 4 h (Scheme S1). In order to investigate the effect of the preparation procedures on the catalytic performance of CM/g-C3N4, the materials with different g-C3N<sup>4</sup> precursors (urea and pure g-C3N4), mass ratios of copper–magnesium oxide to melamine (1:1, 1:3, 1:4, 1:5), pyrolysis temperatures (500 ◦C, 600 ◦C, 800 ◦C, 900 ◦C), pyrolysis times (2 h, 3 h, 5 h, 6 h) and molar ratios of copper to magnesium precursor (1:0.25, 1:0.5, 1:1.5, 1:2) were also synthesized according to the preparation processes of CM/g-C3N4. The composites prepared with urea or pure g-C3N<sup>4</sup> as carbon nitride precursors were denoted as CM/g-C3N<sup>4</sup> (urea) and CM/g-C3N<sup>4</sup> (2), respectively. The preparations of copper oxide, copper–magnesium oxide, copper oxide/g-C3N<sup>4</sup> composite and magnesium oxide/g-C3N<sup>4</sup> composite were the same as the above processes, unless no Mg(NO3)2·6H2O and melamine, melamine Mg(NO3)2·6H2O and Cu(NO3)2·3H2O were added. They were denoted as CuO, CM, C/g-C3N<sup>4</sup> and M/g-C3N4, respectively.

### *2.3. Degradation Experiment*

The information relating to the degradation experiment was described in supplementary material—Text S2. Three groups of parallel samples were carried out in each experiment. All results are displayed as an average value and standard error.

### *2.4. Characterizations*

The information relating to the characterizations is described in supplementary material— Text S3.

### *2.5. Analytical Methods*

RhB concentration was analyzed by a UV-visible spectrophotometer at 554 nm. The total organic carbon (TOC) concentration was calculated by a TOC analyzer (Muti N/C 2100). Active substances were observed by using an electron paramagnetic resonance spectrometer (EPR, Bruker E 500-10/12). The intermediates of RhB degradation were identified using a liquid chromatography–mass spectrometry machine (LC-MS, Agilent 1290II). The ionization mode was chosen as ESI (positive ion mode), and the source temperature was 350 ◦C. The mixture of 0.10% formic acid and CH3CN (60:40, *v*/*v*) was the mobile phase, and the flow rate was 0.20 mL/min. The leaching concentrations of copper and magnesium ions were analyzed by an inductively coupled plasma optical emission spectrometer (ICP-OES, Agilent 720ES). As to the calculation and optimization of RhB molecule, the Fukui Function based on DFT calculation was adopted in Materials Studio software (MS). The Fukui function analysis was analyzed by LDA/PWC in Dmol3 of MS, without taking into account the influence of spin in the calculation [32]. The Fukui function representing free radical (*f* 0 ) attacks was used to evaluate the reaction sites of active substances on the molecules. The toxicity of RhB and its intermediates was evaluated by the Toxicity Estimation Software Tool (TEST) version 5.1.1 based on quantitative structure–activity relationships (QSAR) methodologies [35].

The RhB degradation kinetics were fit by the pseudo first order model and the apparent rate constant (*k*) was calculated according to Equation (1).

$$\text{Ln} \left( \mathbb{C}\_{t} / \mathbb{C}\_{0} \right) = -k \text{t} \tag{1}$$

where *C<sup>t</sup>* is the RhB concentration at a certain reaction time (t) and *C*<sup>0</sup> is the initial RhB concentration; t is the reaction time; *k* is the apparent rate constant. The crystallite size of CM/g-C3N<sup>4</sup> was calculated from XRD data using the Scherrer Equation (2) [36].

$$D = \frac{K\lambda}{B\cos\theta} \tag{2}$$

where *D* is the crystallites' size (nm), *K* is the Scherrer constant (0.9), *λ* is the wavelength of X-ray sources, *B* is the FWHM (radians), and *θ* is the peak position (radians).

### **3. Result and Discussion**

### *3.1. Degradation of RhB in Different Systems*

As shown in Figure 1a, the removal of RhB by CM/g-C3N<sup>4</sup> and PMS was 9.47 and 13.42% respectively, indicating that CM/g-C3N<sup>4</sup> and PMS show poor RhB removal efficiency. Pure g-C3N<sup>4</sup> could remove 28.01% of RhB in 5 min by activating PMS, demonstrating that g-C3N<sup>4</sup> is not only a supporter but also can activate PMS to some extent. CM showed a slightly better catalytic performance than CuO, and the catalytic performance of CM was enormously improved when it was combined with g-C3N<sup>4</sup> to form the CM/g-C3N<sup>4</sup> composite, with the removal of RhB increasing from 19.03 to 99.53%. Based on the characterizations of SEM and BET, the introduction of g-C3N<sup>4</sup> could increase the specific surface area and pore volume of CM, and decrease the aggregation of CM, with more reaction sites being exposed. Furthermore, more hydroxyl groups, which were important for PMS activation, were produced on the surface of CM/g-C3N<sup>4</sup> than that of CM (Sections 3.3.3 and 3.5). Therefore, CM/g-C3N<sup>4</sup> showed a better catalytic performance than CM. As shown in Figure 1b, the catalytic performances of C/g-C3N<sup>4</sup> and M/g-C3N<sup>4</sup> were worse than that of CM/g-C3N4. The reason is that the structure of CM/g-C3N<sup>4</sup> is more conductive for PMS activation than those of C/g-C3N<sup>4</sup> and M/g-C3N<sup>4</sup> (Section 3.3). These results indicate that the interaction among copper, magnesium and g-C3N<sup>4</sup> promotes the catalytic performance of CM/g-C3N4. Thus, the CM/g-C3N<sup>4</sup> is an efficient PMS activator.

**Figure 1.** Degradation of RhB under different systems (**a**,**b**). Reaction conditions: [PMS] = 2.0 mM, [catalyst] = 0.50 g/L, [RhB] = 10.0 mg/L, 25 °C. **Figure 1.** Degradation of RhB under different systems (**a**,**b**). Reaction conditions: [PMS] = 2.0 mM, [catalyst] = 0.50 g/L, [RhB] = 10.0 mg/L, 25 ◦C.

### *3.2. Effect of Preparation Procedures on the Catalytic Performance of CM/g-C3N<sup>4</sup>*

*3.2. Effect of Preparation Procedures on the Catalytic Performance of CM/g-C3N4* The effects of preparation conditions, such as g-C3N4 formed by different precursors, mass ratios of copper–magnesium oxide to melamine, pyrolysis temperatures and times, and molar ratios of copper to magnesium precursor, on the catalytic performance of The effects of preparation conditions, such as g-C3N<sup>4</sup> formed by different precursors, mass ratios of copper–magnesium oxide to melamine, pyrolysis temperatures and times, and molar ratios of copper to magnesium precursor, on the catalytic performance of CM/g-C3N<sup>4</sup> were investigated. CM/g-C3N<sup>4</sup> presented better catalytic performance than CM/g-

CM/g-C3N4 were investigated. CM/g-C3N4 presented better catalytic performance than CM/g-C3N4 (urea) and showed a similar catalytic performance to CM/g-C3N4 (2) (Figure

urea, CM/g-C3N4 will form more carbon nitride than CM/g-C3N4 (urea) in the preparation process, which is beneficial for PMS activation. Guan et al. (2020) also found that g-C3N4 prepared by melamine showed a better catalytic performance than that prepared by urea [37]. To simplify the synthesis process, melamine rather than pure g-C3N4 was utilized as

the g-C3N4 precursor.

C3N<sup>4</sup> (urea) and showed a similar catalytic performance to CM/g-C3N<sup>4</sup> (2) (Figure S1a). This is because the decomposition temperature of melamine is higher than that of urea, CM/g-C3N<sup>4</sup> will form more carbon nitride than CM/g-C3N<sup>4</sup> (urea) in the preparation process, which is beneficial for PMS activation. Guan et al. (2020) also found that g-C3N<sup>4</sup> prepared by melamine showed a better catalytic performance than that prepared by urea [37]. To simplify the synthesis process, melamine rather than pure g-C3N<sup>4</sup> was utilized as the g-C3N<sup>4</sup> precursor.

As the mass ratio of copper–magnesium oxide to melamine increased from 1:1 to 1:2, the degradation of RhB by CM/g-C3N4/PMS increased in 4 min, while RhB degradation decreased in an increase of the mass ratio from 1:2 to 1:5 (Figure S1b). In a low addition amount, g-C3N<sup>4</sup> will completely collapse the basic structure [38], and the dispersion of copper–magnesium oxide on the g-C3N<sup>4</sup> would be inadequate. Furthermore, for a high addition amount of g-C3N4, the dispersion of copper–magnesium oxide was excessive, causing the unsaturated contents of Cu, O and Mg in the CM/g-C3N4. All these phenomena would decrease the catalytic performance of CM/g-C3N4.

The pyrolysis temperature and time have an influence on the catalytic performance of catalysts. When the pyrolysis temperature of CM/g-C3N<sup>4</sup> was less than 700 ◦C, the removal of RhB increased with an increase in temperature. However, the removal of RhB decreased as the pyrolysis temperature of CM/g-C3N<sup>4</sup> reached higher than 700 ◦C (Figure S1c). It was possible that the structure of the catalyst would be destroyed or collapsed at a much higher pyrolysis temperature [39], thereby reducing reaction sites and decreasing the catalytic performance. The removal of RhB increased in 3 min with CM/g-C3N<sup>4</sup> heated from 2 to 4 h, while it decreased with CM/g-C3N<sup>4</sup> heated over 4 h (Figure S1d). This was possible because the structure of the catalyst would be sintered in the long pyrolysis time.

The best molar ratio of copper to magnesium precursors of CM/g-C3N<sup>4</sup> was explored (Figure S1e). Within 3 min, the removal efficiency of RhB was 48.30, 71.80, 98.93, 73.80 and 66.70% for the molar ratios of copper to magnesium precursor of CM/g-C3N<sup>4</sup> of 1:0.25, 1:0.5, 1:1, 1:1.5 and 1:2, respectively. The sample was prone to agglomeration and produced other impurities as the molar ratio of copper to magnesium precursor increased or decreased [40]. It could be inferred that the crystal structure of CM/g-C3N<sup>4</sup> was well constructed in the 1:1 molar ratio of copper to magnesium precursor.

Based on the above results, the catalyst synthesized with a 1:2 mass ratio of copper– magnesium oxide to melamine and a 1:1 molar ratio of copper to magnesium precursor at 700 ◦C for 4 h, was the best PMS activator.

### *3.3. Characterizations of the Catalysts*

### 3.3.1. SEM and TEM

For investigating the effect of the microstructure of catalysts on their catalytic performance, SEM characterization was performed, and the distribution of various elements of CM/g-C3N<sup>4</sup> was also tested by EDS mapping characterization. As shown in Figure 2a,b, CM was composed of particles with aggregation and an irregular shape, and the pure g-C3N<sup>4</sup> was a sheet packing structure. The aggregation of CM decreased when it was combined with g-C3N<sup>4</sup> (Figure 2c). The above results showed that CM was well dispersed on the g-C3N4, thus exposing more reaction sites and improving the catalytic performance. The C/g-C3N<sup>4</sup> were accumulated in a cluster, and no obvious particles had been seen, indicating that the CuO was covered and piled up with the g-C3N4; thus, fewer reaction sites were exposed (Figure 2e). The structures of CuO and g-C3N<sup>4</sup> in C/g-C3N<sup>4</sup> had been changed. On the contrary, M/g-C3N<sup>4</sup> was accumulated in the form of particles, and no obvious flaky structure had been seen (Figure 2f). The research by Hoai Ta et al. showed a similar phenomenon [41]. Compared with C/g-C3N<sup>4</sup> and M/g-C3N4, g-C3N<sup>4</sup> kept its structure in the CM/g-C3N4. These results indicated that the interaction among copper, magnesium and g-C3N<sup>4</sup> would not make the structure of g-C3N<sup>4</sup> change. The decreasing aggregation and the structure of CM/g-C3N<sup>4</sup> was more beneficial for exposing the reaction site to activate PMS than that of C/g-C3N<sup>4</sup> and M/g-C3N4. Therefore, the catalytic per-

formance of CM/g-C3N<sup>4</sup> was the best among them. As shown in Figure 2d, the structure of CM/g-C3N<sup>4</sup> did not change significantly after the reaction, but small cracks appeared in some places of CM/g-C3N4. It may be caused by the corrosion of CM/g-C3N<sup>4</sup> during the PMS activation. The results of SEM-EDS mapping showed that C, N, O, Cu and Mg elements were uniformly distributed on the CM/g-C3N4, rather than a random mixture of all substances (Figure 2g). Figure 2h clearly shows that the CM was uniformly dispersed on the g-C3N4. The HRTEM image of CM/g-C3N4 with clear lattice fringes is exhibited in Figure 2i. The lattice spacing was about 0.2325, 0.2113 and 0.1462 nm, corresponding to CuO (111) and MgO (200), (220) crystal planes, respectively. It was consistent with the XRD results (Section 3.3.2).

*Water* **2022**, *14*, x FOR PEER REVIEW 7 of 29

**Figure 2.** *Cont*.

**Figure 2.** SEM image of CM (**a**), g-C3N4 (**b**), CM/g-C3N4 before (**c**) and after the reaction (**d**), C/g-C3N4 (**e**), M/g-C3N4 (**f**), EDS mapping of CM/g-C3N4 (**g**), TEM image (**h**) and HRTEM image (**i**) of CM/g-C3N4. **Figure 2.** SEM image of CM (**a**), g-C3N<sup>4</sup> (**b**), CM/g-C3N<sup>4</sup> before (**c**) and after the reaction (**d**), C/g-C3N<sup>4</sup> (**e**), M/g-C3N<sup>4</sup> (**f**), EDS mapping of CM/g-C3N<sup>4</sup> (**g**), TEM image (**h**) and HRTEM image (**i**) of CM/g-C3N<sup>4</sup> .

3.3.2. XRD The XRD spectra of different catalysts are presented in Figure 3. The peaks at 13.2° and 27.3° in pure g-C3N4 were respectively attributed to the in-plane repeating units of the continuous heptazine framework ((100) crystal plane) and the interlayer stacking of the periodic conjugated aromatic structure ((002) crystal plane) [42–44]. The diffraction Figure 2h clearly shows that the CM was uniformly dispersed on the g-C3N4. The HRTEM image of CM/g-C3N<sup>4</sup> with clear lattice fringes is exhibited in Figure 2i. The lattice spacing was about 0.2325, 0.2113 and 0.1462 nm, corresponding to CuO (111) and MgO (200), (220) crystal planes, respectively. It was consistent with the XRD results (Section 3.3.2).

#### peaks of M/g-C3N4 at 42.74° and 62.18° corresponded to the (200) and (220) crystal planes 3.3.2. XRD

of MgO, respectively. The (111) crystal plane attributed to CuO could be observed in the C/g-C3N4 at 38.74°. SEM characterization showed that the morphology of g-C3N4 in the C/g-C3N4 and M/g-C3N4 had changed, therefore the g-C3N4 diffraction peaks were not The XRD spectra of different catalysts are presented in Figure 3. The peaks at 13.2◦ and 27.3◦ in pure g-C3N<sup>4</sup> were respectively attributed to the in-plane repeating units of the continuous heptazine framework ((100) crystal plane) and the interlayer stacking of

found in these two materials. The corresponding peaks of CuO and MgO appeared in the

the periodic conjugated aromatic structure ((002) crystal plane) [42–44]. The diffraction peaks of M/g-C3N<sup>4</sup> at 42.74◦ and 62.18◦ corresponded to the (200) and (220) crystal planes of MgO, respectively. The (111) crystal plane attributed to CuO could be observed in the C/g-C3N<sup>4</sup> at 38.74◦ . SEM characterization showed that the morphology of g-C3N<sup>4</sup> in the C/g-C3N<sup>4</sup> and M/g-C3N<sup>4</sup> had changed, therefore the g-C3N<sup>4</sup> diffraction peaks were not found in these two materials. The corresponding peaks of CuO and MgO appeared in the CM, and the aggregation of CM led to the low intensity. The crystal plane diffraction peaks of g-C3N<sup>4</sup> (002), CuO (111) and MgO (200, 220) were detected at 25.18◦ , 38.7◦ , 42.84◦ and 62.52◦ of CM/g-C3N4, respectively. It was in agreement with the HRTEM result. Due to the irregular arrangement of tri-s-triazine units [45,46], CM/g-C3N<sup>4</sup> showed a weak (100) crystal plane diffraction peak. The diffraction peak of the (002) crystal plane in CM/g-C3N<sup>4</sup> was weakened and shifted when compared with pure g-C3N4. This phenomenon indicated that the layered g-C3N<sup>4</sup> was peeled off, and copper–magnesium oxide was successfully introduced into the structure of g-C3N4. When compared with pure CuO (JCPDS: 80-1916) and MgO (JCPDS: 79-0612), the corresponding crystal planes in the CM/g-C3N<sup>4</sup> were slightly shifted, showing that there was an interaction between CuO and MgO in the CM/g-C3N4. Furthermore, the XRD peak of CM/g-C3N<sup>4</sup> was more broadening than that of CM. It could be ascribed to the interaction among copper oxide, magnesium oxide and g-C3N4. According to the calculation of the Scherrer equation, the size distribution of CM/g-C3N<sup>4</sup> was 22.01–63.54 nm, and the average size was 36.94 nm. The XRD spectra of CM/g-C3N<sup>4</sup> before and after the reaction were similar, but the peak of (002) crystal plane shifted from 25.18◦ to 26.2◦ as the intensity decreased. Furthermore, the diffraction peaks of (200) and (220) crystal planes at 42.74◦ and 62.18◦ of CM/g-C3N<sup>4</sup> disappeared after the reaction. The results show that the g-C3N<sup>4</sup> and MgO in CM/g-C3N<sup>4</sup> have important effects on the PMS activation. *Water* **2022**, *14*, x FOR PEER REVIEW 9 of 29 CM, and the aggregation of CM led to the low intensity. The crystal plane diffraction peaks of g-C3N4 (002), CuO (111) and MgO (200, 220) were detected at 25.18°, 38.7°, 42.84° and 62.52° of CM/g-C3N4, respectively. It was in agreement with the HRTEM result. Due to the irregular arrangement of tri-s-triazine units [45,46], CM/g-C3N4 showed a weak (100) crystal plane diffraction peak. The diffraction peak of the (002) crystal plane in CM/g-C3N4 was weakened and shifted when compared with pure g-C3N4. This phenomenon indicated that the layered g-C3N4 was peeled off, and copper–magnesium oxide was successfully introduced into the structure of g-C3N4. When compared with pure CuO (JCPDS: 80-1916) and MgO (JCPDS: 79-0612), the corresponding crystal planes in the CM/g-C3N4 were slightly shifted, showing that there was an interaction between CuO and MgO in the CM/g-C3N4. Furthermore, the XRD peak of CM/g-C3N4 was more broadening than that of CM. It could be ascribed to the interaction among copper oxide, magnesium oxide and g-C3N4. According to the calculation of the Scherrer equation, the size distribution of CM/g-C3N4 was 22.01–63.54 nm, and the average size was 36.94 nm. The XRD spectra of CM/g-C3N4 before and after the reaction were similar, but the peak of (002) crystal plane shifted from 25.18° to 26.2° as the intensity decreased. Furthermore, the diffraction peaks of (200) and (220) crystal planes at 42.74° and 62.18° of CM/g-C3N4 disappeared after the reaction. The results show that the g-C3N4 and MgO in CM/g-C3N4 have important effects on the PMS activation.

**Figure 3.** XRD spectra of different catalysts, g-C3N4, M/g-C3N4, C/g-C3N4, used CM/g-C3N4, fresh CM/g-C3N4 and CM. Black: g-C3N4 (JCPDS: 87-1526); Red: CuO (JCPDS: 80-1916); Blue: MgO (JCPDS: 79-0612). **Figure 3.** XRD spectra of different catalysts, g-C3N<sup>4</sup> , M/g-C3N<sup>4</sup> , C/g-C3N<sup>4</sup> , used CM/g-C3N<sup>4</sup> , fresh CM/g-C3N<sup>4</sup> and CM. Black: g-C3N<sup>4</sup> (JCPDS: 87-1526); Red: CuO (JCPDS: 80-1916); Blue: MgO (JCPDS: 79-0612).

The FTIR spectrum of each sample is shown in Figure 4. The peak at 3375 cm−1 could be assigned to the stretching vibration of a hydroxyl group (−OH) [47]. This peak was

3.3.3. FTIR

### 3.3.3. FTIR

The FTIR spectrum of each sample is shown in Figure 4. The peak at 3375 cm−<sup>1</sup> could be assigned to the stretching vibration of a hydroxyl group (−OH) [47]. This peak was observed in the M/g-C3N4. It could be ascribed to the MgO in M/g-C3N4, which was in favor of –OH being absorbed on the surface. The peak of −OH was also observed in the CM/g-C3N4, with a higher intensity than that seen in the M/g-C3N4, but there were no peaks in the CM and C/g-C3N4. The result indicated that the large number of −OH on the CM/g-C3N<sup>4</sup> surface could be attributed to the interaction between MgO and g-C3N4. It may be because the aggregation of CM was not beneficial for −OH produced on the surface. However, the aggregation was decreased when CM was combined with g-C3N4, and in the action of MgO, −OH could be produced on the surface of CM/g-C3N4. −OH has been certified as an important PMS activation site [48], thus the excellent catalytic performance of CM/g-C3N<sup>4</sup> can be attributed to the numerous −OH (demonstrated in Section 3.5). The peak at 1635 cm−<sup>1</sup> of CM/g-C3N<sup>4</sup> could be ascribed to the <sup>−</sup>OH bending vibration in adsorbed water [49], which is also important for PMS activation. However, this peak was not observed on the other catalysts. The peaks at 1224–1647 and 3175 cm−<sup>1</sup> of pure g-C3N<sup>4</sup> were attributed to the stretching vibration of the heterocyclic ring (C−N=C) and N−H [50–55], respectively. This peak of the heterocyclic ring shifted and became weaker in the CM/g-C3N<sup>4</sup> (1083–1176 cm−<sup>1</sup> ), possibly because of the influence of Cu, Mg and O elements on the structure of the C−N=C. The peak at 810 cm−<sup>1</sup> was associated with the vibration of the triazine unit, but it was not observed in the CM/g-C3N<sup>4</sup> due to its irregular arrangement. These phenomena were in agreement with the XRD result. Due to the shape of the g-C3N<sup>4</sup> changing in the C/g-C3N<sup>4</sup> and M/g-C3N4, their FTIR spectrums showed no characteristic peak about g-C3N4, which was consistent with the results of SEM and XRD. The peaks below 700 cm−<sup>1</sup> of C/g-C3N4, M/g-C3N<sup>4</sup> and CM/g-C3N<sup>4</sup> corresponded to the stretching and bending vibrations of M−O (M: Cu and Mg) [39], respectively. observed in the M/g-C3N4. It could be ascribed to the MgO in M/g-C3N4, which was in favor of –OH being absorbed on the surface. The peak of −OH was also observed in the CM/g-C3N4, with a higher intensity than that seen in the M/g-C3N4, but there were no peaks in the CM and C/g-C3N4. The result indicated that the large number of −OH on the CM/g-C3N4 surface could be attributed to the interaction between MgO and g-C3N4. It may be because the aggregation of CM was not beneficial for −OH produced on the surface. However, the aggregation was decreased when CM was combined with g-C3N4, and in the action of MgO, −OH could be produced on the surface of CM/g-C3N4. −OH has been certified as an important PMS activation site [48], thus the excellent catalytic performance of CM/g-C3N4 can be attributed to the numerous −OH (demonstrated in Section 3.5). The peak at 1635 cm−1 of CM/g-C3N4 could be ascribed to the −OH bending vibration in adsorbed water [49], which is also important for PMS activation. However, this peak was not observed on the other catalysts. The peaks at 1224–1647 and 3175 cm−1 of pure g-C3N4 were attributed to the stretching vibration of the heterocyclic ring (C−N=C) and N−H [50– 55], respectively. This peak of the heterocyclic ring shifted and became weaker in the CM/g-C3N4 (1083–1176 cm−1), possibly because of the influence of Cu, Mg and O elements on the structure of the C−N=C. The peak at 810 cm−1 was associated with the vibration of the triazine unit, but it was not observed in the CM/g-C3N4 due to its irregular arrangement. These phenomena were in agreement with the XRD result. Due to the shape of the g-C3N4 changing in the C/g-C3N4 and M/g-C3N4, their FTIR spectrums showed no characteristic peak about g-C3N4, which was consistent with the results of SEM and XRD. The peaks below 700 cm−1 of C/g-C3N4, M/g-C3N4 and CM/g-C3N4 corresponded to the stretching and bending vibrations of M−O (M: Cu and Mg) [39], respectively.

*Water* **2022**, *14*, x FOR PEER REVIEW 10 of 29

**Figure 4.** FTIR spectra of different catalysts (g-C3N4; M/g-C3N4; C/g-C3N4; CM/g-C3N4; CM). 3.3.4. BET **Figure 4.** FTIR spectra of different catalysts (g-C3N<sup>4</sup> ; M/g-C3N<sup>4</sup> ; C/g-C3N<sup>4</sup> ; CM/g-C3N<sup>4</sup> ; CM). 3.3.4. BET

A Brunauer-Emmett-Teller (BET) gas adsorption analysis was used to investigate the porous structure and specific surface area (SSA) of the catalyst. All samples showed type A Brunauer-Emmett-Teller (BET) gas adsorption analysis was used to investigate the porous structure and specific surface area (SSA) of the catalyst. All samples showed type

IV isotherms and H3 type hysteresis loops, indicating that the mesoporous structure was

IV isotherms and H<sup>3</sup> type hysteresis loops, indicating that the mesoporous structure was dominant in these samples (Figure 5). When compared with CM and C/g-C3N4, the SSA and pore volume of CM/g-C3N<sup>4</sup> increased, and the average pore diameter slightly reduced (Table 1). These results demonstrated that the introduction of g-C3N<sup>4</sup> and MgO increased the SSA and pore volume of CM/g-C3N4, thus more reaction sites were exposed and the structure of CM/g-C3N<sup>4</sup> was more conducive to activating PMS. According to the SEM characterizations of CM, C/g-C3N<sup>4</sup> and CM/g-C3N4, the metal oxide in CM/g-C3N<sup>4</sup> was mostly dispersed. Thus, it could be inferred that the aggregation of the catalyst had a relationship with the SSA. However, the M/g-C3N<sup>4</sup> with the largest SSA (58.47 m2/g) was accumulated in the form of particles, illustrating that the relationship between SSA and the aggregation of catalyst was not linear. When compared with other materials, such as CuMg-MMO [40] and CuO-CN [45], the SSA, pore volume and average pore diameter of the materials prepared in this article were all the best. (Table 1). These results demonstrated that the introduction of g-C3N4 and MgO increased the SSA and pore volume of CM/g-C3N4, thus more reaction sites were exposed and the structure of CM/g-C3N4 was more conducive to activating PMS. According to the SEM characterizations of CM, C/g-C3N4 and CM/g-C3N4, the metal oxide in CM/g-C3N4 was mostly dispersed. Thus, it could be inferred that the aggregation of the catalyst had a relationship with the SSA. However, the M/g-C3N4 with the largest SSA (58.47 m2/g) was accumulated in the form of particles, illustrating that the relationship between SSA and the aggregation of catalyst was not linear. When compared with other materials, such as CuMg-MMO [40] and CuO-CN [45], the SSA, pore volume and average pore diameter of the materials prepared in this article were all the best.

*Water* **2022**, *14*, x FOR PEER REVIEW 11 of 29

**Figure 5.** N2 adsorption–desorption isotherms of different catalysts (g-C3N4; M/g-C3N4; C/g-C3N4; CM/g-C3N4; CM). **Figure 5.** N<sup>2</sup> adsorption–desorption isotherms of different catalysts (g-C3N<sup>4</sup> ; M/g-C3N<sup>4</sup> ; C/g-C3N<sup>4</sup> ; CM/g-C3N<sup>4</sup> ; CM).

**Table 1.** Specific surface area and pore characteristics of different catalysts (g-C3N4; M/g-C3N4; C/g-C3N4; CM/g-C3N4; CM). **Table 1.** Specific surface area and pore characteristics of different catalysts (g-C3N<sup>4</sup> ; M/g-C3N<sup>4</sup> ; C/g-C3N<sup>4</sup> ; CM/g-C3N<sup>4</sup> ; CM).


M/g-C3N4 58.47 0.2041 13.96

As shown in Figure S2a, the removal efficiency of RhB increased from 36.88 to 99.90%

active substances can degrade RhB in the reaction system. As shown in Figure S2b, the removal of RhB raised from 91.83 to 99.55% at 5 min for the PMS concentration of 0.50–1.0

*3.4. Factors Impacting RhB Degradation by CM/g-C3N4/PMS*

3.4.1. Catalyst Dosage, PMS Concentration and RhB Concentration

### *3.4. Factors Impacting RhB Degradation by CM/g-C3N4/PMS* 3.4.1. Catalyst Dosage, PMS Concentration and RhB Concentration

As shown in Figure S2a, the removal efficiency of RhB increased from 36.88 to 99.90% at 5 min, and the *k* value raised from 0.0867 to 1.2916 min−<sup>1</sup> with increasing catalyst dosage (Table S1). More catalysts can provide more reaction sites for PMS activation, thus more active substances can degrade RhB in the reaction system. As shown in Figure S2b, the removal of RhB raised from 91.83 to 99.55% at 5 min for the PMS concentration of 0.50–1.0 mM, with the *k* value increased from 0.4939 to 1.1703 min−<sup>1</sup> (Table S1). As the PMS concentration increased from 1.0 to 3.0 mM, the *k* value decreased from 1.1703 to 0.8455 min−<sup>1</sup> . In fact, as the concentration of PMS increases, the activation reaction can be promoted rapidly. However, the self-quenching reaction of PMS will reduce the number of active substances and limit the expression of catalysts due to the overdose of PMS molecules in the reaction system [56]. Figure S2c showed that the degradation efficiency of RhB decreased from 99.90 to 65.69% in the RhB initial concentration range from 5.0 to 20.0 mg/L at 5 min, and the *k* value dropped from 1.4365 to 0.2187 min−<sup>1</sup> . This may be because the number of active substances were constant in an assured concentration of catalyst and PMS. The higher the content of RhB, the stronger the competition is among RhB molecules for active substances, which slows down the reaction rate of the CM/g-C3N4/PMS system. Considering environmental protection and the concentration of RhB in real life, 0.30 mg/L CM/g-C3N<sup>4</sup> and 1.0 mM PMS were suitable for 10.0 mg/L RhB degradation.

The *k* value of the system (0.30 mg/L CM/g-C3N4, 1.0 mM PMS, 10.0 mg/L RhB, *k* = 1.1703 min−<sup>1</sup> ) was much higher than those of other reaction systems under the optimal conditions, such as the Fe3O4/Co3S4/PMS/RhB system (0.302 min−<sup>1</sup> ), 5%-T/LDOs/PMS/RhB system (0.4537 min−<sup>1</sup> ), Co-G/PMS/RhB system (0.9438 min−<sup>1</sup> ) and Co1+xFe2-xO<sup>4</sup> /PMS/RhB system (0.260 min−<sup>1</sup> ) (Table S2). It indicated that CM/g-C3N<sup>4</sup> can efficiently and rapidly activate PMS to degrade RhB. The mineralization efficiency of RhB in the CM/g-C3N4/PMS system reached 43.43% within 5 min and slightly increased (48.93%) in 30 min (Figure S3). It is probable that RhB was decomposed into small molecule intermediates that can undergo sufficient degradation on the surface of CM/g-C3N4, then stabilizing the TOC value of the system.

### 3.4.2. pH

The effects of initial pH (3.0–11.0) on the RhB degradation were investigated under the optimal conditions of 0.30 g/L CM/g-C3N4, 1.0 mM PMS and 10.0 mg/L RhB. As shown in Figure S4, the removal of RhB was inhibited under the strong acid (pH = 3.0), weak acid (pH = 5.0) and strong alkaline (pH = 11.0) conditions, with the removal efficiency falling below 40.00% within 6 min. The reason was possible in that PMS was mainly presented in the form of H2SO<sup>5</sup> (SO<sup>5</sup> <sup>2</sup>−) under the acid (strong alkaline) condition [57]. These two forms were not activated by the catalyst to produce active substances for pollutant degradation. Under the neutral (pH = 7.0) and weak alkaline (pH = 9.0) conditions, over 99.00% of RhB was removed in 6 min. This means that the CM/g-C3N4/PMS system has good potential for RhB degradation in practical applications because of the neutral or weakly alkaline pH of most actual wastewater.

### *3.5. Possible Activation Mechanisms of PMS*

Quenching experiments were utilized to investigate the reactive oxygen species (ROS) produced in the CM/g-C3N4/PMS system. Figure 6a,b show that the introduction of MeOH (a scavenger for SO<sup>4</sup> •− and •OH) and IPA (a scavenger for •OH) inhibited the removal of RhB. The removal of RhB dropped from 99.88 to 60.50% (MeOH, 300.0 mM) and 82.08% (IPA, 300.0 mM), respectively. Furthermore, the removal efficiency of RhB respectively decreased from 99.88 to 85.78 and 18.14% by adding p-BQ (10.0 mM, O<sup>2</sup> •− scavenger) and L-histidine (10.0 mM, <sup>1</sup>O<sup>2</sup> scavenger) into the system (Figure 6c,d). EPR was carried out to verify ROS in the oxidation reactions. As shown in Figure 7a,b, typical signals of DMPO-OH and DMPO-O<sup>2</sup> <sup>−</sup> were respectively observed in the CM/g-C3N4/PMS/DMPO and

CM/g-C3N4/PMS/DMPO/MeOH systems [40]. However, the DMPO-SO<sup>4</sup> − signal could not be detected in the CM/g-C3N4/PMS/DMPO system. It may be because of the low sensitivity and short life of DMPO-SO<sup>4</sup> − signal [39,40,58], but the result of the quenching experiment had shown that SO<sup>4</sup> •− existed in the reaction system. A typical TEMP-1O<sup>2</sup> triplet signal peak with strong intensity was detected in the CM/g-C3N4/PMS/TEMP system (Figure 7c), showing the existence of <sup>1</sup>O<sup>2</sup> [59]. All these results indicated that SO<sup>4</sup> •−, •OH, O<sup>2</sup> •− and <sup>1</sup>O<sup>2</sup> were generated in the CM/g-C3N4/PMS system, and both SO<sup>4</sup> •− and <sup>1</sup>O<sup>2</sup> were the predominant ROS responsible for the removal of RhB. and 1O2 were generated in the CM/g-C3N4/PMS system, and both SO4•<sup>−</sup> and 1O2 were the predominant ROS responsible for the removal of RhB. The nitrogen was introduced into the CM/g-C3N4/PMS system to verify the contribution of dissolved oxygen in the O2•–generation. As shown in Figure S5, the removal of RhB decreased within 4 min, indicating that dissolved oxygen had contributed to O2•<sup>−</sup> generation to some extent.

**Figure 6.** *Cont*.

**Figure 6.** Effects of different scavengers on RhB removal. MeOH (**a**), IPA (**b**), p-BQ (**c**) and L-Histidine (**d**). Reaction conditions: [PMS] =1.0 mM, [CM/g-C3N4] = 0.30 g/L, [RhB] = 10.0 mg/L, 25 °C. **Figure 6.** Effects of different scavengers on RhB removal. MeOH (**a**), IPA (**b**), p-BQ (**c**) and L-Histidine (**d**). Reaction conditions: [PMS] = 1.0 mM, [CM/g-C3N<sup>4</sup> ] = 0.30 g/L, [RhB] = 10.0 mg/L, 25 ◦C.

The nitrogen was introduced into the CM/g-C3N4/PMS system to verify the contribution of dissolved oxygen in the O<sup>2</sup> •− generation. As shown in Figure S5, the removal of RhB decreased within 4 min, indicating that dissolved oxygen had contributed to O<sup>2</sup> •− generation to some extent.

**Figure 7.** *Cont*.

**Figure 7.** EPR spectra of SO4•<sup>−</sup> and •OH (**a**), O2•<sup>−</sup> (**b**) and 1O2 (**c**) in the CM/g-C3N4/PMS system. Reaction conditions: [PMS] = 1.0 mM, [CM/g-C3N4] = 0.30 g/L, [RhB] = 10.0 mg/L, DMPO/TEMP = 0.10 M, 25 °C. **Figure 7.** EPR spectra of SO<sup>4</sup> •− and •OH (**a**), O<sup>2</sup> •− (**b**) and <sup>1</sup>O<sup>2</sup> (**c**) in the CM/g-C3N4/PMS system. Reaction conditions: [PMS] = 1.0 mM, [CM/g-C3N<sup>4</sup> ] = 0.30 g/L, [RhB] = 10.0 mg/L, DMPO/TEMP = 0.10 M, 25 ◦C.

In order to clarify the surface components of CM/g-C3N4 before and after the reaction, XPS analysis was performed, and the atomic area ratios of different components were shown in Table S3. The full spectrum shows that CM/g-C3N4 contained C, O, N, Cu and Mg elements, and there was little difference in its spectrum before and after the reaction (Figure 8a), showing the stability of CM/g-C3N4. The C1s spectrum showed three peaks at 284.8 (C−C), 286.47 (C−O) and 288.72 ev (C=O) (Figure S6a,b). The atomic area ratios of C−C, C−O and C=O before (after) the reaction were 79.00 (77.42%), 11.68 (10.4%) and 9.32% (12.18%), respectively (Table S3). The result indicates that C−O may participate in the activation process of PMS. For O1s' spectrum, three binding energy peaks at 529.91, 531.93 and 532.84 ev could be classified as lattice oxygen (O2−), hydroxyl groups (−OH) and adsorbed oxygen [60,61], with atomic area ratios of 9.58, 43.27 and 41.99%, respectively (Figure 8b, Table S3). After the reaction, the atomic area ratios of O2<sup>−</sup> and −OH increased to 25.14 and 45.79%, with adsorbed oxygen being reduced to 13.78% (Figure 8c, Table S3). The result demonstrates that adsorbed oxygen was important for PMS activation. Adsorbed oxygen can promote the formation of −OH on the surface of the catalyst and facilitate the activation of PMS [62,63]. Previous studies have shown that −OH is an important reaction site for PMS activation, and H2PO4<sup>−</sup> has a strong chelating influence on the –OH of catalyst surfaces [64–66], thus impacting the PMS activation. Therefore, H2PO4<sup>−</sup> was introduced into the CM/g-C3N4/PMS system to investigate the role of −OH. As shown in Figure S7, the RhB removal was inhibited at the H2PO4<sup>−</sup> concentration of 1.0–8.0 mM at 4 In order to clarify the surface components of CM/g-C3N<sup>4</sup> before and after the reaction, XPS analysis was performed, and the atomic area ratios of different components were shown in Table S3. The full spectrum shows that CM/g-C3N<sup>4</sup> contained C, O, N, Cu and Mg elements, and there was little difference in its spectrum before and after the reaction (Figure 8a), showing the stability of CM/g-C3N4. The C1s spectrum showed three peaks at 284.8 (C−C), 286.47 (C−O) and 288.72 ev (C=O) (Figure S6a,b). The atomic area ratios of C−C, C−O and C=O before (after) the reaction were 79.00 (77.42%), 11.68 (10.4%) and 9.32% (12.18%), respectively (Table S3). The result indicates that C−O may participate in the activation process of PMS. For O1s' spectrum, three binding energy peaks at 529.91, 531.93 and 532.84 ev could be classified as lattice oxygen (O2−), hydroxyl groups (−OH) and adsorbed oxygen [60,61], with atomic area ratios of 9.58, 43.27 and 41.99%, respectively (Figure 8b, Table S3). After the reaction, the atomic area ratios of O2<sup>−</sup> and <sup>−</sup>OH increased to 25.14 and 45.79%, with adsorbed oxygen being reduced to 13.78% (Figure 8c, Table S3). The result demonstrates that adsorbed oxygen was important for PMS activation. Adsorbed oxygen can promote the formation of −OH on the surface of the catalyst and facilitate the activation of PMS [62,63]. Previous studies have shown that −OH is an important reaction site for PMS activation, and H2PO<sup>4</sup> − has a strong chelating influence on the –OH of catalyst surfaces [64–66], thus impacting the PMS activation. Therefore, H2PO<sup>4</sup> − was introduced into the CM/g-C3N4/PMS system to investigate the role of −OH. As shown in Figure S7, the RhB removal was inhibited at the H2PO<sup>4</sup> − concentration of 1.0–8.0 mM at 4 min. This indicated that –OH on the CM/g-C3N<sup>4</sup> surface was an important reaction site for PMS activation. Hence, the adsorbed oxygen on the surface of CM/g-C3N<sup>4</sup> can provide more reaction sites for activating PMS. The higher atomic area ratio of –OH after the reaction demonstrated that the rate of consuming –OH was slower than that of producing –OH. This

min. This indicated that –OH on the CM/g-C3N4 surface was an important reaction site for PMS activation. Hence, the adsorbed oxygen on the surface of CM/g-C3N4 can provide

tion demonstrated that the rate of consuming –OH was slower than that of producing – OH. This is mainly because of the constant formation of –OH by adsorbed oxygen. For N1s' spectrum, two peaks with the binding energy of 398.7 and 400.6 eV were related to sp2 hybridized N atoms (C=N−C) and N(C)3 groups (C−N), respectively (Figure 8d,e) [45]. The atomic area ratio of C=N−C groups decreased from 86.45 to 40.03%, and that of C−N is mainly because of the constant formation of –OH by adsorbed oxygen. For N1s' spectrum, two peaks with the binding energy of 398.7 and 400.6 eV were related to sp<sup>2</sup> hybridized N atoms (C=N−C) and N(C)<sup>3</sup> groups (C−N), respectively (Figure 8d,e) [45]. The atomic area ratio of C=N−C groups decreased from 86.45 to 40.03%, and that of C−N groups increased from 13.55 to 59.97% (Table S1). This shows that the g-C3N<sup>4</sup> of CM/g-C3N<sup>4</sup> has an important contribution to PMS activation, and the conversion of copper species on the surface of g-C3N<sup>4</sup> may occur [63]. For the Cu2p spectrum, the binding peaks located at 933.12 and 934.41 ev could be respectively ascribed to Cu<sup>+</sup> and Cu2+ [40,62], with atomic area ratios of 42.59 and 57.41% (Figure 8f, Table S1). After the reaction, the atomic area ratios of Cu<sup>+</sup> and Cu2+ were 51.14 and 48.86% respectively (Figure 8g, Table S1), indicating that the conversion between Cu<sup>+</sup> and Cu2+ were a major factor in activating PMS. For Mg1s' spectrum, the binding energy peak before and after the reaction was at 1304.59 ± 0.5 eV (Figure S6c,d). It is a typical MgO peak.

Based on the above analyses, the possible activation mechanism of PMS by CM/g-C3N<sup>4</sup> could be sufficiently expressed by Equations (3)–(17) (Scheme 1). They show the redox cycle of Cu2+/Cu<sup>+</sup> -activated PMS to produce SO<sup>4</sup> •−, OH− and SO<sup>5</sup> •− (Equations (3) and (4)). The results of FTIR and XPS analyses had shown that the interaction between MgO and g-C3N<sup>4</sup> and the adsorbed oxygen on the surface of CM/g-C3N<sup>4</sup> would enhance the production of –OH, which promoted the formation of Cu−OH complexes. PMS could react with these substances to form SO<sup>4</sup> •− and SO<sup>5</sup> •− (Equations (5) and (6)) [67]. The results of RhB removed by the g-C3N4/PMS system and XPS analysis had indicated that g-C3N<sup>4</sup> was not only a supporter but also participated in the PMS activation. The electron-rich structure of g-C3N<sup>4</sup> could provide electrons to promote the reduction of PMS. Moreover, PMS could act as an electron donor due to the presence of O−O bonds and the C atoms adjacent to the N atoms in g-C3N<sup>4</sup> could act as the electron acceptors (Equations (7) and (8)) [68]. The generated SO<sup>4</sup> •− reacted with water molecules or OH− to produce •OH (Equation (9)). In the oxidation system, PMS would be hydrolyzed to produce O<sup>2</sup> •− [40]. Furthermore, as mentioned earlier, dissolved oxygen also contributed to O<sup>2</sup> •− generation (Equations (10) and (12)). The generated O<sup>2</sup> •− could consume PMS to accelerate the production of SO<sup>4</sup> •− for improving the oxidation performance of system (Equation (13)). The self-reaction of SO<sup>5</sup> •− generated in the PMS activation and the self-decomposition of PMS would produce <sup>1</sup>O<sup>2</sup> (Equations (14) and (15) [69]. Moreover, the reaction between O<sup>2</sup> •− (SO<sup>5</sup> •−) and water molecules would also generate <sup>1</sup>O<sup>2</sup> (Equations (16) and (17)) [70–72]. RhB was transformed into small molecule intermediates and finally mineralized into H2O and CO<sup>2</sup> under the action of SO<sup>4</sup> •−, •OH, O<sup>2</sup> •− and <sup>1</sup>O2.

$$\equiv\mathrm{Cu^{+}} + \mathrm{HSO\_{5^{-}}} \rightarrow \equiv\mathrm{Cu^{2+}} + \mathrm{SO\_{4}^{\bullet-}} + \mathrm{OH^{-}} \tag{3}$$

$$\equiv\text{Cu}^{+} + \text{HSO}\_{5^{-}} \rightarrow \equiv \text{Cu}^{2+} + \text{SO}\_{4}\text{4}^{-} + \text{OH}^{-} \tag{4}$$

$$\equiv \text{Cu}^+ - \text{OH} + \text{HSO}\_5^- \ \rightarrow \equiv \text{Cu}^{2+} - \text{OH} + \text{SO}\_4^{\bullet-} + \text{OH}^- \tag{5}$$

$$\equiv \text{Cu}^{2+} - \text{OH} + \text{HSO}\_{5^{-}} \rightarrow \equiv \text{Cu}^{+} - \text{OH} + \text{SO}\_{5}^{\bullet -} + \text{H}^{+} \tag{6}$$

$$\rm HSO\_{5^{-}} + g - C\_{3}N\_{4} + e^{-} \rightarrow SO\_{4}{\bullet ^{\bullet -}} + OH^{-} \tag{7}$$

$$\text{HSO}\_{5^{-}} + \text{g} - \text{C}\_{3}\text{N}\_{4} \rightarrow \text{SO}\_{5}^{\bullet -} + \text{H}^{+} + \text{e}^{-} \tag{8}$$

$$\mathrm{SO\_4^{\bullet-}} + \mathrm{H\_2O/OH^-} \rightarrow \mathrm{HSO\_4^{-}}/\mathrm{SO\_4^{2-}} + ^\bullet \mathrm{OH} \tag{9}$$

$$\mathrm{HSO\_{5^{-}}} \rightarrow \mathrm{SO\_{5}^{2-}} + \mathrm{H^{+}} \tag{10}$$

$$\text{SO}\_5^{2-} + \text{H}\_2\text{O} \rightarrow \text{O}\_2^{\bullet-} + \text{SO}\_4^{2-} + \text{H}^+ \tag{11}$$

$$\text{O}\_2 + \text{e}^- \rightarrow \text{O}\_2\text{}^{\bullet-} \tag{12}$$

$$\rm{HSO\_{5^{-}}} + \rm{O\_{2}^{\bullet-}} \rightarrow \rm{SO\_{4}^{\bullet-}} + \rm{O\_{2}} + \rm{OH^{-}} \tag{13}$$

$$\rm{SO\_5^{\bullet-}} + \rm{SO\_5^{\bullet-}} \rightarrow \rm{2SO\_4^{2-}} + ^1\rm{O\_2} \tag{14}$$

$$\rm HSO\_{5^{-}} + SO\_{5}^{2-} \rightarrow \rm HSO\_{4^{-}} + SO\_{4}^{2-} + ^{1}O\_{2} \tag{15}$$

$$\rm{2O\_2^{\bullet-}} + 2H\_2O \rightarrow \rm{\cdot}^{1}O\_2 + H\_2O\_2 + 2OH^- \tag{16}$$

groups increased from 13.55 to 59.97% (Table S1). This shows that the g-C3N4 of CM/g-C3N4 has an important contribution to PMS activation, and the conversion of copper species on the surface of g-C3N4 may occur [63]. For the Cu2p spectrum, the binding peaks located at 933.12 and 934.41 ev could be respectively ascribed to Cu+ and Cu2+ [40,62], with atomic area ratios of 42.59 and 57.41% (Figure 8f, Table S1). After the reaction, the atomic area ratios of Cu+ and Cu2+ were 51.14 and 48.86% respectively (Figure 8g, Table S1), indicating that the conversion between Cu+ and Cu2+ were a major factor in activating PMS.

2SO<sup>5</sup> •− <sup>+</sup> <sup>H</sup>2<sup>O</sup> <sup>→</sup> 2HSO4<sup>−</sup> <sup>+</sup> 1.51O<sup>2</sup> (17) ± 0.5 eV (Figure S6c,d). It is a typical MgO peak.

*Water* **2022**, *14*, x FOR PEER REVIEW 17 of 29

**Figure 8.** *Cont*.

**Figure 8.** *Cont*.

**Figure 8.** *Cont*.

**Figure 8.** X-ray photoelectron spectroscopy spectra of full survey (**a**), O1s (**b**,**c**), N 1s (**d**,**e**), Cu 2p (**f**,**g**) in the CM/g-C3N4 before (**b**,**d**,**f**) and after (**c**,**e**,**g**) the reaction. **Figure 8.** X-ray photoelectron spectroscopy spectra of full survey (**a**), O1s (**b**,**c**), N 1s (**d**,**e**), Cu 2p (**f**,**g**) in the CM/g-C3N<sup>4</sup> before (**b**,**d**,**f**) and after (**c**,**e**,**g**) the reaction. 2SO5•- + H2O → 2HSO4- + 1.51O2 (17)

tions (10) and (12)). The generated O2•<sup>−</sup> could consume PMS to accelerate the production **Scheme 1.** The possible activation mechanism of PMS. **Scheme 1.** The possible activation mechanism of PMS.

der the action of SO4•−, •OH, O2•<sup>−</sup> and 1O2.

≡Cu<sup>+</sup>

#### of SO4•– for improving the oxidation performance of system (Equation (13)). The self-re-*3.6. Degradation Pathway of RhB and Toxicity Estimation 3.6. Degradation Pathway of RhB and Toxicity Estimation*

action of SO5•<sup>−</sup> generated in the PMS activation and the self-decomposition of PMS would produce 1O2 (Equations (14) and (15) [69]. Moreover, the reaction between O2•<sup>−</sup> (SO5•−) and water molecules would also generate 1O2 (Equations (16) and (17)) [70–72]. RhB was transformed into small molecule intermediates and finally mineralized into H2O and CO2 un-The oxidation intermediates of RhB were detected by LC-MS analysis. Generally, the degradation of RhB occurs through a series of reactions, including N-demethylation, deamination, dealkylation, decarboxylation, chromophore cleavage, ring-opening and mineralization [73–75]. The DFT calculation was used to analyze the molecular structure of The oxidation intermediates of RhB were detected by LC-MS analysis. Generally, the degradation of RhB occurs through a series of reactions, including N-demethylation, deamination, dealkylation, decarboxylation, chromophore cleavage, ring-opening and mineralization [73–75]. The DFT calculation was used to analyze the molecular structure

RhB (Figure S8). The result indicated that C10 was vulnerable to being attacked by active substances, and the atoms shown in Table S4 were also under attack. Depending on the

degradation pathway of RhB in the CM/g-C3N4/PMS system was proposed. First, under the intense attack of various active species, RhB (*m*/*z* = 443) was transformed into a series of intermediates (P1–P6) by N-demethylation and deamination. Then, P7 and P8 intermediates were produced by the decarboxylation and further oxidation of P6. Moreover, the chromophores of intermediates (P1–P8) were cleaved to form the P9-P13 intermediates.

≡Cu2+-OH + HSO5– → ≡Cu<sup>+</sup>

more, as mentioned earlier, dissolved oxygen also contributed to O2•<sup>−</sup> generation (Equa-

≡Cu<sup>+</sup> + HSO5- → ≡Cu2+ + SO4•- + OH- (4)



of RhB (Figure S8). The result indicated that C10 was vulnerable to being attacked by active substances, and the atoms shown in Table S4 were also under attack. Depending on the previous reports, the LC-MS analysis and the DFT calculation, nineteen intermediates including isomers were identified during the reaction [75–84]. As shown in Figure 9, the degradation pathway of RhB in the CM/g-C3N4/PMS system was proposed. First, under the intense attack of various active species, RhB (*m*/*z* = 443) was transformed into a series of intermediates (P1–P6) by N-demethylation and deamination. Then, P7 and P8 intermediates were produced by the decarboxylation and further oxidation of P6. Moreover, the chromophores of intermediates (P1–P8) were cleaved to form the P9–P13 intermediates. By a ring-opening reaction, the P9–P13 intermediates further formed small molecule products (P14–P19). Finally, RhB and its intermediates were mineralized into CO<sup>2</sup> and H2O. Many small intermediates of RhB were produced in the CM/g-C3N4/PMS system, showing the good degradation capacity of this system. By a ring-opening reaction, the P9–P13 intermediates further formed small molecule products (P14–P19). Finally, RhB and its intermediates were mineralized into CO2 and H2O. Many small intermediates of RhB were produced in the CM/g-C3N4/PMS system, showing the good degradation capacity of this system. The toxicity estimation is important for predicting the damage of pollutants to the environment and humans. Therefore, the toxicities of RhB and its intermediates were examined by assessing oral rat LD50 and bioaccumulation factors on the TEST. The results are shown in the Figure 10. The predicted oral rat LD50 of RhB was 924.86 mg/kg. Except for P5, P7 and P19, the oral rat LD50 of other intermediates decreased after the degradation of RhB, indicating that the toxicities of intermediates were reduced. The predicted bioaccumulation factor was 15.48 in the RhB, and decreased significantly in the intermediates, indicating the lower accumulation of intermediates in the organisms. In terms of oral rat LD50 and bioaccumulation factors, the CM/g-C3N4/PMS system was able to sufficiently reduce RhB toxicity by transforming it into a series of intermediates with lower toxicity.

**Figure 9.** Proposed degradation pathway of RhB in the CM/g-C3N4/PMS system. **Figure 9.** Proposed degradation pathway of RhB in the CM/g-C3N4/PMS system.

The toxicity estimation is important for predicting the damage of pollutants to the environment and humans. Therefore, the toxicities of RhB and its intermediates were

examined by assessing oral rat LD<sup>50</sup> and bioaccumulation factors on the TEST. The results are shown in the Figure 10. The predicted oral rat LD<sup>50</sup> of RhB was 924.86 mg/kg. Except for P5, P7 and P19, the oral rat LD<sup>50</sup> of other intermediates decreased after the degradation of RhB, indicating that the toxicities of intermediates were reduced. The predicted bioaccumulation factor was 15.48 in the RhB, and decreased significantly in the intermediates, indicating the lower accumulation of intermediates in the organisms. In terms of oral rat LD<sup>50</sup> and bioaccumulation factors, the CM/g-C3N4/PMS system was able to sufficiently reduce RhB toxicity by transforming it into a series of intermediates with lower toxicity. *Water* **2022**, *14*, x FOR PEER REVIEW 23 of 29

**Figure 10.** Oral rat LD50 (**a**) and bioaccumulation factors (**b**) of RhB and its intermediates. **Figure 10.** Oral rat LD<sup>50</sup> (**a**) and bioaccumulation factors (**b**) of RhB and its intermediates.

#### *3.7. Stability of the CM/g-C3N4 3.7. Stability of the CM/g-C3N<sup>4</sup>*

The stability and repeatability of a material is an important parameter to measure its practical applications. Therefore, the CM/g-C3N4 was reused five times under the same The stability and repeatability of a material is an important parameter to measure its practical applications. Therefore, the CM/g-C3N<sup>4</sup> was reused five times under the same

conditions (0.30 g/L CM/g-C3N4, 1.0 mM PMS, 10.0 mg/L RhB). As shown in Figure S9, the removal of RhB decreased from 99.88 to 99.30% after five cycles, showing the good repeat-

The results showed that the loss of copper and magnesium ions were 5.16 and 31.97 mg/L after the reaction, respectively. As a non-toxic metal, magnesium ions are widely present in the water matrix with a concentration of about 0.0–130.0 mg/L. Unlike transition metal ions, the leaching of magnesium ions would not cause secondary pollution. In addition,

conditions (0.30 g/L CM/g-C3N4, 1.0 mM PMS, 10.0 mg/L RhB). As shown in Figure S9, the removal of RhB decreased from 99.88 to 99.30% after five cycles, showing the good repeatability of CM/g-C3N4. This indicates that the CM/g-C3N<sup>4</sup> was stable for long-term running. The leaching concentrations of copper and magnesium ions were monitored by ICP-OES. The results showed that the loss of copper and magnesium ions were 5.16 and 31.97 mg/L after the reaction, respectively. As a non-toxic metal, magnesium ions are widely present in the water matrix with a concentration of about 0.0–130.0 mg/L. Unlike transition metal ions, the leaching of magnesium ions would not cause secondary pollution. In addition, the copper ion leaching concentration of CM/g-C3N<sup>4</sup> was lower than those of other copper-based heterogeneous catalysts (Cu2FeSnS<sup>4</sup> (16.70 mg/L), such as Cu2S (131.60 mg/L) [85]). All the above results show that the eco-friendly and stable CM/g-C3N<sup>4</sup> system has great application potential in the field of environmental remediation. *Water* **2022**, *14*, x FOR PEER REVIEW 24 of 29 the copper ion leaching concentration of CM/g-C3N4 was lower than those of other copper-

#### *3.8. Practical Application of the CM/g-C3N4/PMS System* All the above results show that the eco-friendly and stable CM/g-C3N4 system has great application potential in the field of environmental remediation.

based heterogeneous catalysts (Cu2FeSnS4 (16.70 mg/L), such as Cu2S (131.60 mg/L) [85]).

The CM/g-C3N4/PMS system could degrade other pollutants, such as CR, MB and TC (Figure 11a). The removal of CR, MB and TC were all over 73.00%, indicating that the CM/g-C3N4/PMS system can oxidize various contaminants. Figure 11b reveals that RhB could be degraded in the tap water and the wastewater in 10 min, indicating that the CM/g-C3N4/PMS system can efficiently remove RhB in the actual environment. All these results indicate that the CM/g-C3N<sup>4</sup> system has the practical application potential for pollutant degradation. *3.8. Practical Application of the CM/g-C3N4/PMS System* The CM/g-C3N4/PMS system could degrade other pollutants, such as CR, MB and TC (Figure 11a). The removal of CR, MB and TC were all over 73.00%, indicating that the CM/g-C3N4/PMS system can oxidize various contaminants. Figure 11b reveals that RhB could be degraded in the tap water and the wastewater in 10 min, indicating that the CM/g-C3N4/PMS system can efficiently remove RhB in the actual environment. All these results indicate that the CM/g-C3N4 system has the practical application potential for pollutant degradation.

**Figure 11.** Degradation of different pollutants in the CM/g-C3N4/PMS system (**a**) and RhB removal in different water matrixes (**b**). Reaction conditions: [PMS] = 1.0 mM, [CM/g-C3N4] = 0.30 g/L, [RhB/CR/MB/TC] = 10.0 mg/L, 25 °C. **Figure 11.** Degradation of different pollutants in the CM/g-C3N4/PMS system (**a**) and RhB removal in different water matrixes (**b**). Reaction conditions: [PMS] = 1.0 mM, [CM/g-C3N4] = 0.30 g/L, [RhB/CR/MB/TC] = 10.0 mg/L, 25 ◦C.

### **4. Conclusions**

CM/g-C3N<sup>4</sup> was synthesized and initially used as a heterogeneous catalyst to activate PMS for the degradation of RhB. Owing to the splendid structure and abundant reaction

sites, CM/g-C3N<sup>4</sup> exhibited much a better catalytic performance than CuO, CM, g-C3N4, M/g-C3N<sup>4</sup> and C/g-C3N4. Under the low usage reaction conditions, RhB could be rapidly degraded at a natural pH in 5 min. PMS was activated by the copper and g-C3N<sup>4</sup> in CM/g-C3N4, with SO<sup>4</sup> •− and <sup>1</sup>O<sup>2</sup> as the main active substances for the removal of RhB. CM/g-C3N<sup>4</sup> showed good stability and reusability. This work enriches the family of copper-based heterogeneous catalysts used in the PMS activation, and provides a guiding role of enhancing the catalytic performance of copper oxide. The copper-based catalyst with a higher specific surface area, used for industrial applications, should be considered in the future work.

**Supplementary Materials:** The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/w14132054/s1. Scheme S1: Preparation of CM/g-C3N<sup>4</sup> . Figure S1: Effects of different g-C3N<sup>4</sup> precursors (a), mass ratios of copper-magnesium oxide to melamine (b), pyrolysis temperatures (c), pyrolysis times (d) and molar ratios of copper to magnesium precursor (e) on the catalytic performance of material. Figure S2: Effects of catalyst dosage (a), PMS concentration (b) and RhB concentration (c) for RhB degradation. Figure S3: Mineralization efficiency of CM/g-C3N4/PMS system. Figure S4: Effect of initial pH for RhB degradation in the CM/g-C3N4/PMS system. Figure S5: Effect of nitrogen on the degradation of RhB in the CM/g-C3N4/PMS system. Figure S6: X-ray photoelectron spectroscopy spectra of C1s (a, b) and Mg 1s (c, d) in the CM/g-C3N<sup>4</sup> before (a, c) and after (b, d) the reaction. Figure S7: Effect of H2PO<sup>4</sup> − on the degradation of RhB in the CM/g-C3N4/PMS system. Figure S8: The molecular structure of the optimized RhB (a) (oxygen, red; carbon, gray; nitrogen, green; hydrogen, white). Figure S9: Degradation of RhB with the recycled CM/g-C3N<sup>4</sup> . Figure S10: Degradation of different pollutants in the CM/g-C3N4/PMS system (a) and RhB removal in different water matrixes (b). Table S1: k value of CM/g-C3N4/PMS system in different conditions. Table S2: RhB degradation in different systems. Table S3: Atomic area ratios of different compositions in the CM/g-C3N<sup>4</sup> before and after the reaction. Table S4: Partial Fukui indexes of RhB molecule.

**Author Contributions:** Conceptualization, Y.M.; Formal analysis, Y.M.; Funding acquisition, X.Z.; Methodology, W.X.; Project administration, X.Z.; Resources, X.Z.; Software, W.X.; Supervision, X.Z. and S.Z.; Writing—original draft, Y.M.; Writing—review & editing, X.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Key Research and Development Program of China, (Grant No. 2016YFC0400702-2), the National Natural Science Foundation of China (Grant No. 21377041) and the Guangdong Science and Technology Program (Grant No.2020B121201003).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

### **References**

