Carbonization of N/P Co-Doped Resin for Metal-Free Catalytic Ozonation of Oxalic Acid

Abstract

In this study, a millimeter-scale N/P-doped carbonaceous catalyst was synthesized via facile carbonization of the N/P-doped resin at 800 °C (NPCR-800). This work aimed to investigate the performance of the NPCR-800 catalyst in heterogeneous catalytic ozonation and the mechanism of reactive oxygen species (ROS) generation. The NPCR-800 achieved the highest oxalic acid (OA) degradation efficiency of 91% within 40 min. The first-order kinetics of OA degradation in the NPCR-800/O₃ system was approximately twelve and three times higher than that in the O₃ and O₃/GAC system, respectively. In addition to excellent catalytic ozonation performance, the NPCR catalyst also exhibited good reusability and salt tolerance. The dominant ROS were identified by the electronic spin response and free radical quantitative experiments, being responsible for oxalic acid degradation in NPCR-800/O₃ system. The effect of the doped N and P elements on enhancing the catalytic activity was understood, what was ascribed to the efficient reaction of the O₃ molecule with the active site of the graphitic N, defect site and carbonyl/carboxyl groups of NPCR to generate the hydroxyl radical and singlet oxygen. A type of metal-free catalytic ozonation strategy was developed in this work, which is promising in the practical treatment of the refractory organic pollutants.

1. Introduction

Ozone is considered as a powerful eco-friendly oxidative agent for destructing refractory organic pollutants due to its relatively high redox potential (2.08 eV) and short half-life period (around 15 min in freshwater) [1]. However, molecular ozone is selective, merely attacking unsaturated bonds (such as alkene π-bond) [2]. The ozone oxidation of some organic pollutants is not expected, which limits the further application of the ozonation process. For instance, oxalic acid (OA) is a typically intermediate product generated during AOPs, performing ozone-resistant attributes [3]. In this case, various catalytic ozonation processes have been developed for boosting the oxidation performance of ozonation through generating reactive oxygen species (ROS) [4,5]. Catalytic ozonation has been a hot research topic in recent decades. Among ozonation catalysts, transition metal-based heterogeneous catalysts (e.g., waste iron shavings (Fe⁰) [6], CoFe₂O₄ [7], MnFe₂O₄ [8], MnO₂ [9] and Mn-CeO_x@γ-Al₂O₃) [10]) have been proven to be relatively efficient. Nevertheless, the inevitable heavy metal leaching will induce secondary pollution during the catalytic ozonation process. In working toward a sustainable future, it is of great significance to construct alternative metal-free catalysts with excellent performance for forming a green process on wastewater remediation.

Typical carbonaceous catalysts such as activated carbon [11], carbon nanotubes [12] and biochar [13] have been employed in the catalytic ozonation process. However, the catalytic activity of pristine carbonaceous catalyst is not always sufficient. In this regard, current studies have demonstrated that introducing nonmetallic heteroatoms (e.g., N, P, S and F) is a feasible strategy to enhance the catalytic activity of the carbonaceous catalysts [14]. Compared with raw carbon catalysts (e.g., activated carbon and biochar), much attention has been paid to the heteroatom-doped carbonaceous materials, which can activate ozone molecules to produce ROS for degrading organic contaminants because of their large specific surface area, unique electronic structure and electrochemical properties [15,16]. For instance, the waste MOF absorbent-derived N-doped spongy carbon catalyst [17] can provide extra electrons, resulting in the promotion of O₃ activation. The OA removal efficiency achieved 100% within 20 min, which is 10 times as much as that of sole ozonation. Moreover, the N/S co-doped biochar [18] catalytic ozonation achieved an Imazapic degradation reaction rate of 0.07011 min⁻¹, which is almost five-fold higher than that of non-catalytic ozonation (0.01283 min⁻¹).

In addition, it is well known that the ROS formation in carbonaceous materials’ catalytic ozonation is due to the ozone activation by oxygen-containing functional groups (e.g., C=O and COOH) [19]. In addition to oxygen-containing functional groups, carbon-based catalysts can provide more active sites after heteroatom doping. It comes down to the fact that typical heteroatoms (e.g., N and S) are more electronegative to C atoms, inducing more active sites by redistributing spin and charge density which results in the enhancement of catalytic activity of carbonaceous catalysts [18]. Despite these benefits, these non-metallic heteroatom-doped carbonaceous materials are typically employed in powdered forms, leading to poor separation and recovery performance. The inevitable loss of powdered catalyst will lead to an increase in treatment costs and the risk of secondary pollution. Moreover, synthesizing heteroatom-doping catalysts requires complex preparation methods and heteroatom additive sources (e.g., melamine [20] and urea [21]), thereby increasing the fabrication cost. Therefore, designing an easily separated metal-free catalyst without exogenous heteroatom additive sources is highly desirable. Recently, a millimeter-scale commercial resin was employed as a precursor to obtain heteroatom-doped carbonaceous catalyst, which has been applied in persulfate/peroxymonosulfate activation [22]. Importantly, the catalysts on a millimeter scale are easy to separate from aqueous solution. In this respect, resin would be a promising candidate as a precursor for synthesizing the ozonation catalyst with facile separation property.

Thus, in this study, the ion exchange resin D412, which is rich in the N and P elements, was hypothesized to be selected as a precursor to synthesize a novel in situ N/P co-doped carbonaceous ozonation catalyst (NPCR) via direct pyrolysis. This is the first time that the application of NPCR in catalytic ozonation was investigated. The objectives of this study were to (1) evaluate the catalytic activity of NPCR by the oxalic acid degradation efficiency; (2) explore the catalytic mechanism of ROS generation by electronic spin resonance (ESR) and free radical quantitative experiments. As a result, a new insight into the resin-derived metal-free catalytic ozonation process is understood, which presents a new solution for designing efficient and easily separable ozone catalysts in practical wastewater treatment.

2. Materials and Methods

2.1. Chemicals and Materials

Aminophosphonic acid (-CH₂NHCH₂PO₃⁻) chelating ion exchange resin D412 was chosen as the precursor to synthesize spherical NPCR, purchased from Zhengzhou Hecheng New Material Technology Co., Ltd., Zhengzhou, China. The chemical structure of D412 is illustrated in Figure S1. Commercial granular activated carbon (GAC, 20–50 mesh) was obtained from Macklin reagent Co., Ltd., Shanghai, China. Oxalic acid (OA, H₂C₂O₄·2H₂O) and sodium hyposulfite (Na₂S₂O₃, 99.5%) was purchased from Guangzhou Chemical Reagent Factory, Guangzhou, China. All the chemicals were of analytical grade and used directly without further purification. The initial reaction pH was adjusted by diluted NaOH and HCl solutions in the absence of buffer.

2.2. Synthesis of NPCR

Spherical N/P doped carbonized resin was prepared by direct calcination method. Typically, a certain amount of chelating ion exchange resin D412 was dried at 105 °C for 12 h, and then transferred to box furnace and calcined at 400 °C for 1 h with a heating rate of 5 °C/min under vacuum conditions. After that, the temperature was continuously increased to corresponding temperatures (600 °C, 700 °C, 800 °C and 900 °C) and maintained for 2 h, and ultimately, the samples were collected after the temperature dropped to room temperature. For eliminating residual Na⁺, the obtained catalysts were washed several times until the pH of the filtrate solution was nearly neutral, followed by drying and sieving with screens (30–50 mesh). The final products were denoted as NPCR-600, NPCR-700, NPCR-800 and NPCR-900. For comparison, the commercial GAC was also washed and dried as described above.

2.3. Characterization Methods

The morphology and elemental composition of catalysts were analyzed by scanning electron microscope (SEM, ZEISS, Oberkochen, Germany) equipped with energy-dispersive spectroscopy (EDS). The crystal structures of the samples were characterized using X-ray powder diffraction (D8 Advanced, Bruker, Karlsruhe, Germany) in a 2θ range of 10–90° with Cu Ka radiation. The composition and chemical states of elements on the catalyst surface were analyzed with the help of X-ray photoelectron spectroscopy (XPS, ESCALAB, Thermo Scientific, Boston, MA, USA) equipped with an Al Kα monochromatized source. The specific surface area (SSA) of catalysts was measured by a BET apparatus (ASAP-2020, Micromeritics, Norcross, GA, USA) via N₂ adsorption/desorption method.

2.4. Catalytic Ozonation for Mineralization of OA

An NPCR catalytic ozonation system (NPCR/O₃) was carried out using a 0.5 L glass cylinder (Φ 53 mm × 380 mm). Ozone was generated from high-purity oxygen (99.9%) by an ozone generator (QD-D5W, Guangzhou Qida technology Co., Ltd., Guangzhou, China). Typically, a catalyst (0.25 g) was added into 500 mL of a 150 mg/L OA solution and kept by steady magnetic stirring. The produced ozone with gas concentrations of about 60 mg/L was bubbled into the reactor through a titanium diffuser. The inlet flow rate was set at 100 mL/min by a rotary flowmeter unless otherwise specified. The initial pH value of the solution was around 2.8 without any adjustment. Particularly, in the experiments of OA degradation under different initial pH conditions, the initial pH was adjusted by diluted NaOH/HCl solutions (0.1 mol/L). A water sample of 5 mL was withdrawn from the reactor and filtered by a 0.45 µm PTFE filter at given time intervals. Sodium hyposulfite solution (0.5 mL) was added to each sample for quenching the radicals in NPCR/O₃ system.

The concentrations of organics remained in the samples were determined by total organic carbon analyzer (TOC-100, QIKUN Science, Hangzhou, China). In order to identify the active species during ozonation via NPCR catalysts, free radical quantitative experiments and electron spin-resonance spectroscopy (ESR, JES FA200, JEOL, Tokyo, Japan) were performed. Free radical quantitative experiments were detailed in the Supplementary Materials. DMPO (5,5dimethyl-1-pyrrolidine-N-oxide) and TEMP (2,2,6,6-tetramethylpiperidine) were employed as spin-trapping agents to capture the reactive oxygen species.

It is well known that the removal efficiency of oxalic acid is equivalent to TOC removal efficiency during the catalytic ozonation, since no other organic intermediates are generated during oxalic acid degradation [23]. Hence, in this work, the performances of different processes were estimated in terms of TOC removal. Additionally, the OA degradation rate (k_obs) was calculated according to the pseudo-first-order kinetic model, as given by Equation (1).

$$\ln ({\mathrm{C}}_{\mathrm{t}}/{\mathrm{C}}_0)=-k_{\mathrm{obs}}\mathrm{t}$$

(1)

where C_t and C₀ are the TOC concentrations (mg/L) at any time (t, min) and t = 0, and k_obs (min⁻¹) is the reaction rate constant.

3. Results and Discussion

3.1. Characterization of Synthesized NPCR Catalysts

The images of NPCR catalysts prepared at different calcination temperatures were obtained from electron microscope (Ruihong, China) and shown in Figure S2. When the calcination temperature varies from 600 °C to 800 °C, the NPCR catalysts obtained are about 0.3~0.6 mm in diameter with an undamaged spherical shape (Figure S2a–c). However, as can be seen in Figure S2d, a higher calcination temperature (900 °C) could cause the obvious breakage of prepared NPCR, which would lead to the undesired pulverization and poor recycling performance of catalysts. Hence, no further investigation into NPCR-900 was conducted. The morphology and structure of synthesized NPCR-800 were observed by SEM. As can be seen in Figure 1a, the as-obtained NPCR-800 reveals a spherical shape with an average diameter of about 0.3 mm. The corresponding EDS mapping images (Figure 1b–e) indicate that C, N, O and P atoms are uniformly distributed on the NPCR-800 sample, confirming the existence of N and P elements.

The obtained XRD patterns of the synthesized NPCR catalysts are presented in Figure 1g. All NPCR samples revealed two broad diffraction peaks at around 25° and 43°, corresponding to the typical reflections ((0 0 2) and (1 0 1)) of amorphous carbon, respectively [24]. As calcination temperature increased from 600 °C to 800 °C, the intensity of these two diffraction peaks increased gradually, suggesting that the graphitization level of NPCR increased [25]. The elevated graphitization level could be owing to the promoted conversion from amorphous carbon to graphitic carbon with increasing pyrolysis temperature [26].

Raman spectroscopy was leveraged to probe the defects and graphitization degree of carbon-based catalysts. As depicted in Figure 1h, two characterized peaks observed at around 1350 cm⁻¹ and 1590 cm⁻¹ were assigned to the disordered carbon structure (D band) and graphitized carbon structure (G band), respectively [27,28]. The intensity ratio of the D band and G band (I_D/I_G) is usually used to assess the defect density and disorder level of carbonaceous materials [28,29]. A higher I_D/I_G value denotes the higher defect degree and disorder level of carbonaceous catalysts. For pyrogenic temperatures of 600 °C, 700 °C and 800 °C, the I_D/I_G of the NPCR-600, NPCR-700 and NPCR-800 are 0.77, 0.81, and 0.89, respectively. A similar trend can also be observed in carbon-based catalyst-related reports [30,31,32]. It can be explained by the fact that higher pyrolysis temperature was conducive to the formation of graphitic-N, resulting in a significant downshift of the G band [32,33]. These results imply that high temperature can boost the generation of defects in NPCR and thus converting from ordered carbon into disordered carbon.

The structural properties of NPCR catalysts were investigated via N₂ adsorption/desorption isotherms and the results are summarized in Table S1. The values of S_BET for the NPCRs exhibit an increasing tendency with an increase in pyrolysis temperature. In addition, this tendency is also in accordance with their catalytic ozonation degradation efficiency, indicating that the SSA of NPCR is greatly responsible for the enhancement of catalytic performance.

As a powerful approach for investigating the composition and valence states of elements at the surface of the catalyst, XPS was employed and its spectra are displayed in Figure 2. The wide peak of all C1s spectra (Figure 2a–c) can be deconvoluted into four sections located at 284.7–284.8 eV, 285.9–286.4 eV, 278.8–288.5 eV and 290.4–291.3 eV, associated with C=C, C=O, COOH and π-π* shake-up groups [34]. In previous studies, carbonyl (C=O) [15,16] and carboxyl (COOH) groups [35] generally served as active sites in catalytic ozonation. To evaluate the prospective activity of carbonaceous catalysts, it is highly relevant to determine the amounts of carbonyl/carboxyl groups on NPCR. Based on the analysis methods described by Xiao et al. [34], the relative amounts of carbonyl/carboxyl groups on NPCR catalysts are determined and listed in

Table 1. Total content and relative amount of C=O/COOH on the surface obtained from XPS data and BET results of NPCR catalysts.

Sample	Carbon Content on the Surface (at.%)	Distribution (%)		C=O/COOH Content on the Surface (at.%)	SBET (m2/g)	Relative Amount of Surface C=O/COOH (SBET × Surface Content)
		C=O	COOH
NPCR-600	81.5	19.0	11.2	24.6	284.7	70.0
NPCR-700	84.1	16.1	10.7	22.5	397.1	76.1
NPCR-800	88.0	20.0	14.3	30.2	557.2	168.3

. The relative amounts of surface carbonyl/carboxyl groups increased as the pyrolysis temperature increased from 600 °C to 800 °C, implying that high temperature might contribute to the number of active sites on NPCR. The N1s spectra (Figure 2d–f) of all samples could be deconvoluted into three main peaks centering at 398.6 eV, 400.2 eV and 401.1 eV, corresponding to pyridine N, pyrrole N and graphitized N, respectively [36]. The graphitic N content increased from 19.8% to 60.3% while both pyridine N and pyrrole N contents reduced with the increase in calcination temperature, which was indicative of the transformation from pyridine N and pyrrole N to graphite type N.

The spectra of P2p display two characteristic peaks for NPCR-600 and NPCR-700 samples (Figure 2g,h), and the peaks located at about 134.5 eV and 133.6 eV could be ascribed to C-O-P and C-P-O bonds [37,38], respectively. Furthermore, only one characteristic peak assigned to the C-P-O bond was observed in the P2p spectrum of NPCR-800 (Figure 2i), and C-O-P bonds could barely be detected when the calcination temperature increased to 800 °C. It has been reported that the C-P-O bond was also regarded as an active site for catalytic oxidation reaction [39]. The promotion of content of the C-P-O bond might also boost the catalytic performance of NPCR catalysts.

Related studies have revealed that N and P doping favors the formation of structural defects [21]. To clarify the structural defects formation for NPCR, the correlation between the I_D/I_G ratio and different heteroatoms species was investigated. Taking N content into consideration (

Table 2. Surface N species distribution derived from the XPS data of NPCR catalysts.

Sample	N Content (at.%)	Distribution (%)			Absolute (at. %)
		Pyrrolic N	Graphitic N	Pyridinc N	Pyrrolic N	Graphitic N	Pyridinic N
NPCR-600	3.06	47.8	19.8	32.4	1.46	0.61	0.99
NPCR-700	3.49	39.3	34.5	26.2	1.37	1.20	0.91
NPCR-800	2.76	32.1	60.3	7.6	0.89	1.66	0.21

), a strong positive linear correlation between the I_D/I_G ratio and graphitic N can be observed in Figure S3a, suggesting the formation of the defect site could be attributed to the incorporation of graphitic N species. However, negative correlations and no correlations are observed between the pyrrolic N, pyridinic N, C-O-P, C-P-O and I_D/I_G ratio (Figure S3b–e), suggesting that these species had no significant effect on the formation of defect sites on NPCR. In summary, these results suggest that increased graphitized N with the pyrolysis temperature was the main reason for the increasing level of defect sites.

3.2. Performance of NPCR in Catalytic Ozonation

3.2.1. OA Degradation in NPCR/O₃ System

OA was employed as a target organic contaminant for evaluating the catalytic performance of NPCR. As seen in Figure 3a,b, the OA degradation and k_obs values increased from 61% to 91% and 0.023 min⁻¹ to 0.062 min⁻¹ as the calcination temperature increased from 600 °C to 800 °C, respectively. Hence, 800 °C could be deemed as the appropriate calcination temperature. As illustrated in Figure 3c, only about 12% of OAs were removed via adsorption on the NPCR-800 within 40 min. The OA concentration only dropped 20% in the sole presence of O₃ within 40 min with a k_obs value of 0.005 min⁻¹. The sum of OA removal efficiency by adsorption and ozonation processes was only 32%, while the combination of NPCR-800 and ozonation achieved a 91% degradation of OA within 40 min with a k_obs value of 0.062 min⁻¹. These results indicate the synergistic effect among the NPCR-800 and ozonation, which could be ascribed to catalytic activity of NPCR. The catalytic ozonation process using GAC (almost the same size as NPCR-800) and NPCR-800 resulted in 53% and 91% of OA removal within 40 min, respectively. Moreover, the first-order kinetics of OA degradation in the NPCR-800/O₃ system (0.062 min⁻¹) was approximately twelve and three times higher than that in the O₃ (0.005 min⁻¹) and O₃/GAC (0.019 min⁻¹) system, respectively. These results indicate that the catalytic activity of NPCR-800 was even higher than that of GAC, which is considered as one of best ozonation catalysts [11,40].

In addition, a brief comparison of OA degradation by NPCR-800 catalytic ozonation processes with other metal-free ozonation catalysts is shown in Table S2. Compared with active reduced graphene and N-doped nanocarbons, the advantage of NPCR-800 is that almost the same reaction rate can be realized with a smaller amount of O₃ consumption. Despite slightly better catalytic performance than that of NPCR-800, the previously reported catalysts were synthesized by complicated synthesis methods or expensive materials (MOF, graphene oxide or nanotubes), suggesting that a much higher cost was an inevitable problem for synthesizing these catalysts. In sum, NPCR-800 might be most likely ozonation catalysts for practical application.

3.2.2. Effects of Operational Parameters on OA Degradation via the NPCR/O₃ Process

As important operational parameters, ozone dosage, catalyst dosage and initial pH generally influenced the performance of catalysts. In this study, the ozone concentration of generated ozone gas was maintained around 60 mg/L. Ozone dosage was adjusted by regulating ozone gas flow. The effects of ozone dosage on OA degradation are shown in Figure 4a. It could be seen that OA degradation efficiencies were markedly facilitated from 66% to 91% as the ozone gas flow increased from 50 to 100 mL/min. This might stem from the fact that a moderately higher dosage of ozone could provide more oxidant to be induced into reactive oxide species, that favored OA mineralization. Nevertheless, when ozone dosage was further increased, no obvious enhancement of OA degradation was observed. This may be due to the limited catalyst dosage and insufficient utilization of excess ozone.

As seen in Figure 4b, the increase in catalyst dosage at relatively low concentration revealed a positive effect on OA degradation, and 91% of OA degradation efficiency was obtained with the presence of NPCR at 0.5 g/L. However, only a slight promotion on OA degradation was obtained when NPCR-800 dosage further increased from 0.7 g/L. These results indicated that the OA degradation efficiency depended well on NPCR dosage. Taking into account OA degradation efficiency and cost, the suitable dosage of NPCR was 0.5 g/L.

As shown in Figure 4c, the performance of NPCR demonstrated the pH-dependent feature. OA degradation efficiency gradually decreased from 91% to 35% in the NPCR-800/O₃ system when the initial solution pH varied from 2.8 to 11.2. The hindering effect of alkalinity on target organic pollutants also could be found in the reports on carbocatalytic ozonation [3,41], which could be attributed to fewer organic pollutants or ozone interacting with the catalyst. For elucidating the interaction between the OA molecular mechanism and catalyst, the surface charges of NPCR-800 under the pH range from 1 to 11 were analyzed by zeta potential measurement. As can be seen in Figure S4, the zeta potential of NPCR-800 decreased from 11.6 mV to −13.5 mV as the pH increased from 1 to 11, and NPCR-800 obtained a pH_pzc of 2.9. Once pH exceeded 3, the surface of NPCR-800 was negatively charged. Moreover, regarding its pKa1 and pKa2 of 1.23 and 4.19, respectively, the deprotonated forms (HC₂O₄⁻ and C₂O₄²⁻) were the dominant species of oxalic acid at all the selected initial pHs. Hence, at higher pHs (pH ≥ 3), the surface charge gave rise to a higher electrostatic repulsion between NPCR and oxalic acid, inducing the inhibition on the adsorption of oxalic acid onto NPCR. Although initial pH largely affected the production of ROS and the interactions between catalysts and pollutants, the NPCR-800/O₃ process still showed significant promotion on OA degradation under near-neutral conditions compared to the sole O₃ process (Figure S5). Thus, NPCR still can be regarded as a promising catalyst for treating organic wastewater.

3.2.3. Salt Tolerance and Stability

With the further application of laboratory synthesized catalysts in real wastewater, high-salinity conditions are an inevitable problem. High-salinity organic wastewater produced from chemical companies (e.g., petrochemical, coal chemical and pharmaceutical industries) usually contains a high concentration of sulfates (SO₄²⁻) and chlorides (Cl⁻). However, these inorganic anions usually act as strong radical quenchers in radical-based oxidation processes, resulting in a noticeable inhibition effect on organic pollutant degradation [42,43]. Additionally, the reaction between chloride and ozone molecules may produce an undesired ozone depletion, thus reducing the degradation of organic pollutants [44]. Furthermore, many lab-synthesized catalysts also showed decreased activity under hypersaline conditions owing to their deficient salt tolerance [10,45]. Hence, the effect of chlorides (Cl⁻) and sulfates (SO₄²⁻) should be investigated to evaluate the performance of the NPCR/O₃ process in real industrial sewage. The OA degradation efficiency was examined in the NPCR-800/O₃ process with sulfates and chloride ions at concentrations ranging from 30 mmol/L to 300 mmol/L. As can be observed from Figure 4d, sulfate ion displayed a favorable effect on the catalytic oxidation process with all selected concentrations. The boost of OA degradation could be owing to the enhancement of ozone mass-transfer coefficient (K_La) at the gas–liquid interface [10]. As is well known, the ozone mass-transfer coefficient is generally negatively correlative to the size of ozone bubble in the solution. At the same time, salinity impeded the coalescence of bubbles, and the size of ozone bubble in high-salt solution was smaller than that in salt-free solution, suggesting that more dissolved O₃ molecules were utilized in catalytic ozonation. The positive effect of chloride ion on OA degradation was still slightly amplified when the concentration of Cl⁻ ulteriorly increased from 30 mmol/L to 300 mmol/L. Whereas, as reported by some researchers [46], obvious inhibitory effects were frequently observed in radical-based oxidation process as the concentration of Cl^- exceeded 200 mmol/L. Fortunately, these repression effects have been proved to be relieved via non-radical pathways [47]. Accordingly, the extraordinary salt tolerance of NPCR implies that non-radical reaction might play an important role in the NPCR/O₃ process.

In addition to catalytic performance, reusability is also an essential concern with respect to the practical application. A cycling experiment was conducted to evaluate the reusability of the NPCR-800 catalyst. As depicted in Figure 4e, the OA degradation efficiencies in five runs after 40 min of treatment were 93%, 94%, 92%, 92% and 83%, respectively. Obviously, the fifth run showed a slight decrease in OA degradation efficiency, more than 80% of OA removal was still obtained, which is still higher than that of ozone oxidation process alone. Moreover, NPCR can be easily intercepted and separated by screen mesh (<100 mesh) during the continuous flow process due to its millimeter-scale size. Hence, NPCR displayed great reusability in catalytic ozonation, and would be a promising catalyst applied in real wastewater treatment.

3.3. Mechanism

3.3.1. Identification of Generated ROS

Hydroxyl radicals (•OH) are usually regarded as the dominant reactive oxygen species (ROS) responsible for organic degradation in metal-based catalytic oxidation [5,48]. With the in-depth study of catalytic ozonation, some scholars have found that non-radical reactions also play an important role in the catalytic ozonation process, especially in metal-free catalytic ozonation [14,41,49]. Therefore, to further examine the degradation mechanism of NPCR catalytic ozonation, electron paramagnetic resonance (EPR) experiments were conducted to identify the reactive oxygen species generated in NPCR-800/O₃. Typical spin-trapping reagent DMPO (5,5-Dimethyl-1-pyrroline N-oxide) was employed in EPR experiments to detect •OH and •O₂⁻, and TEMP (2,2,6,6-Tetramethylpiperidine) was used to detect ¹O₂. As described in Figure 5, no obvious signals could be captured in sole ozonation, demonstrating the negligible ozone decomposition into ROS without catalysts. In contrast, strong characteristic peaks ascribed to DMPO-•OH adducts (quartet lines with height ratio of 1:2:2:1) were found in NPCR-800/O₃ systems (Figure 5a), suggesting the generation of •OH and distinct boosting of ozone decomposition. The significant signals (a_N = 14.5 G, ${\mathrm{a}}_{\mathrm{H}}^{\mathsf{\beta }}$ = 11.4 G, ${\mathrm{a}}_{\mathrm{H}}^{\mathsf{\gamma }1}$ = 1.3 G), which assigned to DMPO-•O₂⁻ also appeared in NPCR-800/O₃ systems (Figure 5b), indicating that •O₂⁻ was also generated in the catalytic ozonation process. In Figure 5c, the symbolic triplet spectrum which was assigned to TEMP-¹O₂ (a_N = 16.9 G) was apparently revealed in the NPCR/O₃ system and no obvious peak was observed in the sole ozonation system. The results confirm the existence of ¹O₂, implying a possible involution of non-radical reaction in OA degradation. Considering the occurrence of various ROS generated in catalytic ozonation, the contribution of each ROS to OA elimination needed to be further explored.

3.3.2. Radical Quantification

In most of the studies, the contributions of each ROS involved in the catalytic ozonation process were distinguished via conventional quenching experiments [50,51,52]. Overall, TBA was selected as a scavenger for quenching •OH owing to its high •OH reactivity (6.0 × 10⁸ M⁻¹ • s⁻¹), low O₃ reactivity (3 × 10⁻³ M⁻¹ • s⁻¹) and no •O₂⁻ reactivity [53]. p-BQ could act as a scavenger for quenching both •O₂⁻ (3.5–7.8 × 10⁸ M⁻¹ • s⁻¹) and •OH (1.9 × 10⁹ M⁻¹ • s⁻¹) [2]. Thus, the relative contribution can be evaluated through comparing the different reductions of organic degradation efficiencies with the presence of TBA or p-BQ. Unfortunately, these quenching experiments using TBA and p-BQ are becoming increasingly controversial. The •OH radicals generated in catalytic ozonation cannot only react with target organic contaminants but also be the triggers for promoting O₃ decomposition to various ROS [54]. TBA can suppress the catalytic ozonation performance via inhibiting the decomposition of O₃, so it is reasonable to speculate that the contribution of •OH might be overestimated in certain cases. Moreover, the application of FFA (furfuryl alcohol) and NaN₃ also leads to a debatable conclusion on the investigation for the role of ¹O₂ [54]. Hence, in the current work, for a better understanding on the mechanism of ozone activation by NPCR, the generation of hydroxyl radical was quantified by fluorescence spectrophotometer through employing coumarin (COU) as the probe. Additionally, the concentration of superoxide radicals formed during the NPCR-800/O₃ process was determined using NBT (Nitrotetrazolium blue chloride) as a probe for superoxide radicals. The relative contributions of various ROS to OA degradation were evaluated by the equations described in Equations (2)–(5). The result of the •OH generation quantitative test using coumarin as a probe was presented in Figure 5d. Notably, the process of 7HC (7-hydroxycoumarin) generation can be approximately divided into two stages. The first stage from 0 to 2 min is the rapid increase in the 7HC concentration, suggesting that •OH was generated in the O₃/NPCR-800 system with a great speed. NPCR-800 could catalyze O₃ decomposition to ROS within 2 min. Nevertheless, at the second stage, the 7HC concentration unexpectedly decreased as reaction time increased. Similar phenomena were also recognized in other reports [55,56]. An explanation for the 7HC reduction could be that the formed 7HC was further attacked by •OH. Nevertheless, these results still demonstrate that at least 0.017 μmol/L of •OH was generated in the O₃/NPCR-800 system. For the quantification of •O₂⁻, Figure 5d showed that the concentration of •O₂⁻ in the NPCR-800/O₃ process was 29.5 μmol/L. These results manifest that the NPCR-800 catalyst can facilitate the translation of O₃ into •OH and •O₂⁻.

The reaction constants of oxalate ion with hydroxyl radical (7.7 × 10⁶ M⁻¹ s⁻¹) [57] are seven or more orders of magnitude higher than that with superoxide radicals (<0.2 M⁻¹ s⁻¹) [58]. It is unrealistic to expect that the contribution of •O₂⁻ for OA degradation can exceed that of •OH unless the concentrations of •O₂⁻ are seven orders of magnitude higher than that of •OH. As proven by the above results, the concentration of •O₂⁻ generated in the O₃/NPCR-800 process was 29.5 μmol/L, which was only three orders of magnitude higher than that of •OH (0.017 μmol/L), suggesting that the contribution of •O₂⁻ on OA degradation was negligible in O₃/NPCR-800 system.

As described by Equations (2) and (3) [59,60], ¹O₂ originated from the reaction between •O₂⁻ and H⁺ or H₂O. The k value of Equation (2) was about 10⁵ M⁻¹ s⁻¹ [61], which was much higher than the reaction constants between OA and •O₂⁻ (<0.2 M⁻¹ s⁻¹). Thus, it could be speculated that generated •O₂⁻ preferred to transform into singlet oxygen rather than react with OA molecular mechanisms. During the OA degradation, about 14 μmol/L of ¹O₂ was generated by assuming that the majority of •O₂⁻ was converted into ¹O₂ through the reaction described in Equation (2).

The relative contributions of various ROS (defined as $f_{•\mathrm{OH}}$, $f_{•{{\mathrm{O}}_2}^{-}}$ and $f_{{}^1\mathrm{O}_2}$) to OA degradation could be quantitatively evaluated using Equations (4)–(7) [54]. Where k_ad is the pseudo-first-order rate constant for OA adsorption onto the catalyst, $k_{{\mathrm{O}}_3}$, $k_{•\mathrm{OH}}$, $k_{•{{\mathrm{O}}_2}^{-}}$ and $k_{{}^1\mathrm{O}_2}$ are the second-order rate constant for the reaction of OA with O₃, •OH, •O₂⁻ and ¹O₂, respectively, wherein, the values of $k_{{\mathrm{O}}_3}$, $k_{•\mathrm{OH}}$, $k_{•{{\mathrm{O}}_2}^{-}}$ and $k_{{}^1\mathrm{O}_2}$ are 0.04 M⁻¹ s⁻¹, 7.7 × 10⁶ M⁻¹ s⁻¹, 2 × 10⁴ M⁻¹ s⁻¹ [62] and 0.2 M⁻¹ s⁻¹, respectively. t is the reaction time. As can be seen in Figure 4c, the $k_{\mathrm{ad}}\mathrm{t}$ value can be neglected owing to the insignificant OA removal by NPCR adsorption. Based on the above data, the $k_{•\mathrm{OH}}{\displaystyle \int [•\mathrm{OH}]\mathrm{dt}}$ value (0.131 s⁻¹) is almost of the same order of magnitude as the $k_{{}^1\mathrm{O}_2}{\displaystyle \int [{}^1\mathrm{O}_2]}\mathrm{dt}$ value (0.280 s⁻¹). Thus, it could be reasonably inferred that the contributions of •OH and ¹O₂ on OA degradation were basically the same. Furthermore, the $k_{{\mathrm{O}}_3}{\displaystyle \int {[\mathrm{O}}_3]\mathrm{dt}}$ and $k_{•{{\mathrm{O}}_2}^{-}}{\displaystyle \int [•{{\mathrm{O}}_2}^{-}}]\mathrm{dt}$ values are 2 × 10⁻⁴ s⁻¹ (the ${\displaystyle \int {[\mathrm{O}}_3]\mathrm{dt}}$ value is approximately equal to total O₃ input) and 5.9 × 10⁻⁶ s⁻¹, respectively, which are much lower than $k_{•\mathrm{OH}}{\displaystyle \int [•\mathrm{OH}]\mathrm{dt}}$ and $k_{{}^1\mathrm{O}_2}{\displaystyle \int [{}^1\mathrm{O}_2]}\mathrm{dt}$ values. In sum, the value of $f_{•\mathrm{OH}}$ and $f_{{}^1\mathrm{O}_2}$ are 0.319 and 0.681, respectively, while the value of $f_{•{{\mathrm{O}}_2}^{-}}$ is 1.44 × 10⁻⁵ which is negligible. Briefly, the above results suggest that both •OH and ¹O₂ played dominant roles in OA degradation.

$${{\displaystyle 2•\mathrm{O}}}_2^{-}+{{\displaystyle 2\mathrm{H}}}^{+}\to {\mathrm{H}}_2{\mathrm{O}}_2+{}^1\mathrm{O}_2$$

(2)

$$•{\mathrm{O}}_2^{-}{+2\mathrm{H}}_2\mathrm{O}\to {\mathrm{H}}_2{\mathrm{O}}_2+{}^1\mathrm{O}_2{+2\mathrm{OH}}^{-}$$

(3)

$$\ln ({\displaystyle \frac{\left[\mathrm{OA}\right]}{{[\mathrm{OA}}_{\mathrm{o}}]}})=k_{\mathrm{ad}}\mathrm{t}+k_{{\mathrm{O}}_3}{\displaystyle \int {[\mathrm{O}}_3]\mathrm{dt}+k_{•\mathrm{OH}}{\displaystyle \int [•\mathrm{OH}]\mathrm{dt}+k_{•{{\mathrm{O}}_2}^{-}}{\displaystyle \int [•{{\mathrm{O}}_2}^{-}}}}]\mathrm{dt}+k_{{}^1\mathrm{O}_2}{\displaystyle \int [{}^1\mathrm{O}_2]\mathrm{dt}}$$

(4)

$$f_{•\mathrm{OH}}={\displaystyle \frac{k_{•\mathrm{OH}}{\displaystyle \int [•\mathrm{OH}]\mathrm{dt}}}{k_{\mathrm{ad}}\mathrm{t}+k_{{\mathrm{O}}_3}{\displaystyle \int {[\mathrm{O}}_3]\mathrm{dt}+k_{•\mathrm{OH}}{\displaystyle \int [•\mathrm{OH}]\mathrm{dt}+k_{•{{\mathrm{O}}_2}^{-}}{\displaystyle \int [•{{\mathrm{O}}_2}^{-}}}}]\mathrm{dt}+k_{{}^1\mathrm{O}_2}{\displaystyle \int [{}^1\mathrm{O}_2]\mathrm{dt}}}}$$

(5)

$$f_{•{{\mathrm{O}}_2}^{-}}={\displaystyle \frac{k_{•{{\mathrm{O}}_2}^{-}}{\displaystyle \int [•{\mathrm{O}}_2^{-}]\mathrm{dt}}}{k_{\mathrm{ad}}{\mathrm{t}+\mathrm{k}}_{{\mathrm{O}}_3}{\displaystyle \int {[\mathrm{O}}_3]\mathrm{dt}+k_{•\mathrm{OH}}{\displaystyle \int [•\mathrm{OH}]\mathrm{dt}+k_{•{\mathrm{O}}_2^{-}}{\displaystyle \int [•{\mathrm{O}}_2^{-}}}}]\mathrm{dt}+k_{{}^1\mathrm{O}_2}{\displaystyle \int [{}^1\mathrm{O}_2]\mathrm{dt}}}}$$

(6)

$$f_{{}^1\mathrm{O}_2}={\displaystyle \frac{k_{{}^1\mathrm{O}_2}{\displaystyle \int [{}^1\mathrm{O}_2]\mathrm{dt}}}{k_{\mathrm{ad}}{\mathrm{t}+\mathrm{k}}_{{\mathrm{O}}_3}{\displaystyle \int {[\mathrm{O}}_3]\mathrm{dt}+k_{•\mathrm{OH}}{\displaystyle \int [•\mathrm{OH}]\mathrm{dt}+k_{•{\mathrm{O}}_2^{-}}{\displaystyle \int [•{\mathrm{O}}_2^{-}}}}]\mathrm{dt}+k_{{}^1\mathrm{O}_2}{\displaystyle \int [{}^1\mathrm{O}_2]\mathrm{dt}}}}$$

(7)

3.3.3. Catalytic Active Sites

Previous reports have disclosed that the function groups (such as graphical N, defect and oxygen contained group) of catalyst can serve as active sites for catalytic ozonation reactions [63,64]. Therefore, for investigating the catalytic mechanism of NPCR on ozone, the correlation between the functional groups and k_obs was analyzed.

The analysis of the structure–activity relationship between organic pollutant degradation rate constants and carbonyl/carboxyl group contents was performed to gain insight into the role of carbonyl/carboxyl groups on the NPCR surface. As reveal in Figure 6a, an ideal linear relationship (R² = 0.919) can be observed between the relative amount of carbonyl/carboxyl groups and OA degradation rate constant, suggesting that carbonyl/carboxyl groups acted as active sites for ozonation activation. Structural defects of catalysts have been reported as valid active sites for promoting electron transfer, thereby improving the catalytic performance. Hence, the relationship between k_obs and the I_D/I_G ratio of NPCR was also investigated. As shown in Figure 6b, the results show a strong positive correlation (R² = 0.979), indicating that defect site is also strongly associated with catalytic ozonation performance.

Moreover, several studies have suggested that higher electronegativity of heteroatom content (nitrogen and phosphorus) usually hastened the electron transfer of adjacent carbon atoms, thereby facilitating the ozone molecules transferring into ROS [15,65]. The structure–activity relationship between organic pollutant degradation rate constant and heteroatom species content was also analyzed. As shown in Figure 6c and Table 2, taking total N content into consideration, the absolute atom content of graphitic N species was positively linearly correlated with the k_obs significantly (R² = 0.986), whereas a significant negative correlation was observed between other N species and k_obs. As a result, graphitic N played a key role in O₃/NPCR system.

Likewise, in accordance with the data in

Table 3. Surface P species distribution derived from the XPS data of NPCR catalysts.

Sample	P Content (at. %)	Distribution (%)		Absolute (at. %)
		C-O-P	C-P-O	C-O-P	C-P-O
NPCR-600	0.58	62.5	37.5	0.36	0.22
NPCR-700	0.52	50.8	49.2	0.26	0.26
NPCR-800	0.24	0	100	0.00	0.24

, the correlations between k_obs values and the P species (C-P-O and C-O-P) are demonstrated in Figure S6. Unlike graphitic N species, the absolute atom content of C-P-O or C-O-P exhibited negative correlation or no significant correlation with k_obs. The low coefficient of determination demonstrates that the phosphorus groups are negligible contributors in NPCR/O₃ systems. In summary, it can be inferred that carbonyl/carboxyl groups, defect sites and graphitic N were the vital active sites responsible for the generation of ROS in the NPCR/O₃ system.

Based on the above results and analysis, the dominant catalytic ozonation reactions in the NPCR/O₃ process were identified as radical and non-radical pathways. Additionally, although the oxygen-containing groups, graphitic N and defects were confirmed to be the critical active sites on ozone activation via structure–activity relationship analysis, the roles of different active sites in catalytic ozonation still need more discussion.

$$•{\mathrm{O}}_3^{-}+{\mathrm{H}}_2\mathrm{O}\to {\mathrm{O}}_2+•\mathrm{OH}+{\mathrm{OH}}^{-}$$

(8)

$$•{\mathrm{O}}_3^{-}+•\mathrm{OH}\to •{\mathrm{O}}_2^{-}+•{\mathrm{HO}}_2$$

(9)

$$2•{\mathrm{HO}}_2\to {\mathrm{H}}_2{\mathrm{O}}_2+{}^1\mathrm{O}_2$$

(10)

$${2\mathrm{O}}_3{+\mathrm{H}}_2{\mathrm{O}}_2\to 2•{\mathrm{OH}+3\mathrm{O}}_2$$

(11)

$${\mathrm{Carbon}\text{-}\mathsf{\pi }+\mathrm{H}}_2\mathrm{O}\to {\mathrm{Carbon}\text{-}\mathrm{HO}}_3^{+}{+\mathrm{OH}}^{-}$$

(12)

$${\mathrm{O}}_3+{\mathrm{OH}}^{-}\to •{\mathrm{O}}_2^{-}+•{\mathrm{HO}}_2$$

(13)

According to previous reports, the oxygen-containing groups may be involved in generating ROS via the following pathways: (1) -C=O groups could facilitate the conversion of O₃ into •O₃⁻, a major precursor for radical chain reactions [66]. Furthermore, according to Equations (8)–(10) [63,67], •O₃⁻ also could be a trigger for ¹O₂. A higher amount of •O₃⁻ resulted in a higher yield of ¹O₂. Thus, -C=O groups mediately stimulated the non-radical reaction in the NPCR catalytic ozonation process; (2) –C=O groups could offer electrons for •OH generation via in situ decomposition of ozone [21]; and (3) -COOH was in favor of H₂O₂ generation, which was subsequently involved in the radical chain reactions described in Equation (11), leading to the boosting of •OH generation. As described previously, the graphitic N on NPCR could be considered to facilitate the formation of surface ROS, which subsequently converted into •OH, •O₂⁻ and ¹O₂ [21,63]. The delocalized electrons produced by defects of carbonaceous catalysts is often used to explain why defect sites can promote the activation of ozone [21]. The delocalized electrons (carbon-π) could facilitate O₃ decomposition to •HO₂ and •O₂⁻ via Equations (12)–(13), thus boosting radical chain reactions. Based on the above discussions, the possible pathways of ozone activation by NPCR are illustrated in Figure 7.

4. Conclusions

In this study, a millimeter-scale N/P-doped carbonaceous catalyst was successfully prepared via one-step pyrolysis of resin. The synthesized heteroatom metal-free catalyst exhibited excellent performance on catalytic ozonation of OA. Herein, the following conclusions could be obtained from the experimental results:

(1)

In the presence of the optimum NPCR-800 catalyst, OA degradation efficiency substantially increased by around 70% compared to that of sole ozonation within 40 min. Furthermore, the k_obs value of the NPCR-800/O₃ process is nearly three times higher than that of the GAC/O₃ process.

(2)

The performance of NPCR displayed noticeable dependency on initial pH, which might be owing to the change of electrostatic force between OA and NPCR under various pH conditions. Fortunately, a significant promotion of NPCR was still observed under neutral conditions compared to the O₃ process alone. Additionally, the NPCR catalyst displayed high salinity tolerance and exhibited superior activity after utilizing it five times.

(3)

An insight into the mechanism of the NPCR/O₃ system was also obtained via analyzing the results of EPR and radical quantification tests. NPCR was verified to be able to promote O₃ decomposition to produce •OH, •O₂⁻ and ¹O₂. Except •O₂⁻, both •OH and ¹O₂ played vital roles in OA degradation.

(4)

According to the results of the structure–activity relationship analysis, graphitic N, the defect site and carbonyl/carboxyl groups were affirmed as the major active sites towards OA degradation.

Overall, this work provides a notable strategy for synthesizing easily separable metal-free heteroatom catalysts which hold significant promise for practical application in wastewater treatment.