Insights into highly effective catalytic persulfate activation on oxygen-functionalized mesoporous carbon for ciprofloxacin degradation

Nanocarbons have been demonstrated as promising carbon catalysts for substituting metal-based catalysts for the green treatment of wastewater. In this study, oxygen-functionalized mesoporous carbon (OCMK-3) was prepared by wet oxidation and exhibited high catalytic performance against ciprofloxacin (CIP) by activation of persulfate. The effects of environmental parameters (pH, temperature, coexisting ions) and process parameters (temperature, sodium persulfate concentration, catalyst agent dosage, initial concentration) on the removal of CIP were investigated. Compared with the pristine ordered mesoporous carbon (CMK-3), the removal efficiency of CIP by OCMK-3 was increased by 32% under optimal conditions. This rise in activity was attributed to the increase in oxygen-containing functional groups, porosity, and specific surface area of OCMK-3 with improved structural defects and electron transfer efficiency. Furthermore, based on active species scavenging experiments, a dual-pathway mechanism of the radical and nonradical pathways was discovered. The rational degradation pathway of CIP was investigated based on liquid chromatography–mass spectrometry (LC–MS). In addition, the OCMK-3/PS system exhibited high decomposition efficiency in pharmaceutical wastewater treatment. This study provides an in-depth mechanism for the degradation of organic pollutants by carbon-based PS-AOPs and provides theoretical support for further studies.


Introduction
Pharmaceuticals, mainly antibiotics, are emerging contaminants that are frequently found in the environment worldwide through continuous discharge from sewage pipes and environmental accumulation (Liu et al., 2015;Meena et al., 2015). The presence of antibiotics, even at low concentrations, could pose severe threats to the ecosystem and human health mainly through antibacterial resistance and possible toxicity. However, the chemical complexity and diversity make antibiotics difficult to remove in conventional wastewater treatment plants (Tello et al., 2012;Su et al., 2014). Therefore, in order to remove antibiotics from water bodies, several techniques such as advanced oxidation processes (AOP) (Chen et al., 2020;Du et al., 2021), adsorption (Torres-Perez et al., 2012;Li, et al., 2014;Al-Musawi et al., 2017), and biodegradation (Mueller et al., 2013;Fonseca and De, 2020) have been extensively researched. Among them, AOP technology is considered to be one of the most effective ways for antibiotic removal due to the reaction with Responsible Editor: Sami Rtimi * Yuanyuan Liu lyy8838@126.com 1 reactive radicals, including sulfate and hydroxyl radicals (·OH and SO 4 · − ) (Li et al., 2017;Thanekar et al., 2018). Compared with traditional AOP based on ·OH, novel AOP based on sulfate radicals (SR-AOPs) has been extensively applied in water treatment, as peroxymonosulfate (PMS) or persulfate (PS) using in SR-AOPs is more stable, easier to store, and cheaper than liquid H 2 O 2 . Besides, SO 4 · − possess higher standard oxidation potential and longer lifetime than ·OH (Yang et al., 2008;Duan and Xiaoguang et al., 2018), therefore drawing increasing attention. PS could be activated using many methods, such as illumination, heat activation, ultrasonic activation, microwave activation, metal-based material activation, and metal-free catalyst activation (Matzek and Carter, 2016;Shao et al., 2017). Compared with homogeneous activation system which requires intensive energy inputs or high dosages of chemical impost, heterogeneous systems based on transitional metals/metal oxides and carbonaceous materials have been widely investigated. In contrast with metal-based heterogeneous systems, carbonaceous materials could avoid secondary contamination by metals, heat, or alkalinity and exhibit high efficiency and stability, in line with the development strategy of green remediation (Hao and Carroll, 2016;Guo et al., 2017aGuo et al., , 2017b. Among metal-free carbonaceous catalysts, ordered mesoporous carbon is of great interest because of its regular spherical morphology, abundant mesopores, large specific surface area, high chemical stability, and good electrical conductivity (Zhu, et al., 2015;Du et al., 2020).
Previous studies have shown that catalytic materials play a synergistic role in specific applications due to their surface chemistry and pore structure. Both surface chemistry and pore structure modifications have been demonstrated to improve the performance of catalysts (Xu et al., 2003;Bazu et al., 2008). To increase active sites of the catalysts and thus enhance the catalytic performance, carbonaceous materials are usually post-treated by oxidation, polymer coating, grafting, etc. (Gómez-Serrano et al., 1999;Kim et al., 2011;Vara and Collazos-Castro, 2019). Among these modification methods, surface oxidation is one of the most convenient and commonly used methods. The oxygen-containing groups could be introduced into the surface of the catalysts through surface oxidation, which might offer more active sites for persulfate activation, and the hydrophilicity of the surface could also get improved (Bazu et al., 2008), making the catalysts more suitable for further applications. Although there have been some studies about oxygen-functionalized mesoporous carbon (OCMK-3), very few of them reported activating PS using OCMK-3.
Therefore, to reveal the mechanism involved in OCMK-3-based SR-AOPs for the removal of aqueous organic pollutants, the commonly used antibiotic ciprofloxacin (CIP) was selected as the target pollutant. The removal of CIP in the OCMK-3/PS system under different conditions was studied. Considering the superior adsorption performance of OCMK-3, the relationship between adsorption and catalytic processes was discussed. The reaction mechanism of the OCMK-3/PS system was discussed based on the experimental results of free radical burst, electron paramagnetic resonance (EPR), and liquid chromatography-mass spectrometry (LC-MS). The application prospects of the OCMK-3/ PS system were also evaluated using real pharmaceutical wastewater samples.

Preparation of CMK-3/OCMK-3
At first, hexagonal mesoporous silica (SBA-15) was synthesized according to the typical method as reported. SBA-15 was used as a hard template to prepare mesoporous carbon (CMK-3) (Jun, 2000). The details of the synthesis procedures of SBA-15 and CMK-3 are provided in Supporting Information (SI).
CMK-3 was oxidized by wet oxidation at different nitric acid concentrations, contact temperatures, and oxidation time. Typically, 0.1 g of dried CMK-3 powder was treated with 15 mL of HNO 3 solution (1, 2, 4, 6 M solution) under refluxing for 15 min to 60 min at a temperature range of 70-90 °C to prepare OCMK-3. After oxidation, mesoporous catalysts were filtered and washed with ultrapure water until solution was neutral (Nejad et al., 2016). OCMK-3 synthesized under optimal conditions was then used for characterization and the removal of CIP.

Material characterization
The morphological information of the two catalysts was obtained employing transmission electron microscopy (TEM, JEOL-1230). The nitrogen adsorption/desorption isotherm was obtained using Brunauer-Emmett-Teller (BET) method with a Micromeritics ASAP 2020, and the pore size properties of the materials were derived from the adsorption branches of isotherm using the BJH (Barret-Joyner-Halenda) method. To characterize the functional groups within the CMK-3 and OCMK-3 catalysts, Fourier-transform infrared spectroscopy (FTIR) on a PerkinElmer instrument (Spectrum GX, USA) within the 400-4000 cm −1 region was employed. The surface chemistry and chemical states of elements before and after the reaction were investigated, using X-ray photoelectron spectroscopy (XPS) by a Thermo ESCALAB 250XI instrument under an Al-Kα X-ray radiation. Zeta potential was determined by dispersing CMK-3 and OCMK-3 in 1 mmol/L NaCl solution with various pH values through ultrasonic treatment, and then measuring the supernatant with Zetasizer Nano (ZEN3600, Malvern).

Catalytic activity tests
Batch experiments were conducted in 100 mL conical flasks in a thermostatic shaker, with the shaking rate at 150 rpm and temperature maintained at 25 °C. Typically, OCMK-3 was first added to a series of conical flasks containing 50 mL of 200 mg/L CIP solution (pH 5 without any buffering). After shaking for 1 h to reach adsorption-desorption equilibrium, a certain amount of PS was added to activate the catalytic process and shaken for another 2 h. Generally, unless otherwise mentioned, 0.2 g/L of OCMK-3 and 8 mM of PS were used in the system, and. 0.1 M HNO 3 or NaOH was used to adjust the solution pH. The mixed solution was taken at each time interval, filtered, and then added to a centrifuge tube containing 1 mL of methanol, which was employed to scavenge the active free radicals and quench the further oxidation reaction. The experiment procedure for the pharmaceutical wastewater sample was conducted the same as above.
As a relatively crucial index for catalysts, after the completion of the experimental system reaction, OCMK-3 was recycled by ethanol and water, dried at 60 °C, and reused for three cycles. To study the roles of active free radicals, certain amounts of radical scavengers were added into the system. MeOH, EtOH, and 1,4-benzoquinone (BQ) were used (calibrated by molar ratio MeOH/PS, EtOH /PS, or BQ/PS) to scavenge SO 4 · − , ·OH, and O 2 · − , respectively. Experimental samples were conducted in triplicate, with the average values reported as the results.

Analytical methods
The concentration of CIP was measured by the UV2600 ultraviolet spectrophotometer at wavelengths 275.5 nm (Tran et al., 2019). The generated active radicals, SO 4 • − , •OH, and O 2 • − , were detected on a Bruker ER200-SRC spectrometer using 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as a spin-trapping agent. The pharmaceutical wastewater samples were analyzed using three-dimensional excitation-emission matrix (3D-EEM) fluorescence spectroscopy with the excitation wavelengths ranging from 200 to 600 nm, and the emission wavelengths ranging from 200 to 900 nm. The intermediate products of CIP after OCMK-3/PS treatment were identified by LC/MS.

Pharmaceutical wastewater sample treatment
In order to test the practical potential, the OCMK-3 was applied in the treatment of pharmaceutical wastewater, which was gathered from Liu yang Bio-medicine Garden (Hunan province, China). The sample was brownish-yellow in color and had a strong pungent smell. The contents in the wastewater mainly contained alcohols, bufexamac, 1-bromobutane, hydroxylamine, and other organic matters.

Results and discussion
Characterization TEM was used to observe the morphology of CMK-3 and OCMK-3, with the results shown in Fig. 1. The wellordered stripe-like structures could be clearly observed from Fig. 1a and b, indicating that oxidation did not destroy the ordered mesoporous structure of CMK-3. The isotherm curve of both catalysts (inset of Fig. 2) exhibited a type IV feature, indicating the typical mesoporous structures. Based on the pore size distributions curves (Fig. 2), the most probable pore sizes of the two materials centered around 5 nm, indicating the uniform mesopores of them, in agreement with the TEM images (Wang et al., 2018). Meanwhile, the specific pore structure parameters of both catalysts are listed in Table S3.
The changes of CMK-3 and OCMK-3 in the composition of functional groups before and after reaction were analyzed using XPS. As shown in the wide scan XPS spectra ( Fig. 3a, b), The oxygen content of OCMK-3 was higher than that of CMK-3 before and after reaction, indicating that more oxygen-containing groups were introduced after the oxidation treatment, and resulted in the improvement of the catalytic effect. Meanwhile, both CMK-3 and OCMK-3 spectra showed increased oxygen levels after reaction owing to intermediate adsorption and surface oxidation, which was consistent with the previous studies (Duan et al., 2016). Figure 3c-f shows the de-convolution XPS C1s and O1s spectra before and after the reaction. Before reaction, high solution C1s XPS spectra of CMK-3 could be divided into three sections corresponding to t C graphite/C-C/C-H (283.85 eV), C-OH/C-O-C (284.77 eV) and π-π* (287.94 eV) shakeup, respectively. After reaction, a broad C1s peak of CMK-3 was fitted into four component peaks at 284.34, 285.92, 287.12, and 288.42 eV, indicating C graphite/C-C/C-H, C = O, COOH, and π-π* shakeup. Meanwhile, before reaction, C1s XPS spectra represented three C species on original OCMK-3 in the range of 282-292 eV: C graphite/ C-C/C-H, C-OH/C-O-C and π-π* shakeup generated at 283.58, 284.77, and 287.77 eV, respectively. After reaction, a broad C1s peak was fitted into four C species on OCMK-3 in the range of 282-292 eV: C graphite/C-C/C-H, C = O, COOH and π-π* shakeup generated at 284.34, 285.92 and 287.12 eV, 288.42 eV, respectively. After reaction, the C-OH/C-O-C on the surface of both CMK-3 and OCMK-3 decreased, and C = O and COOH increased. Therefore, it could be conferred that in the process of PS activation by the two materials, the hydroxyl groups on the surface were converted into carbonyl and carboxyl groups. Thus, hydroxyl on the surface could be regarded as an important functional group for catalyst activation. As shown in Fig. 3d, f, high solution O1s XPS spectra could be divided into three types of O species: O1 (530.98-531.9 eV), O2 (532.6-533.6 eV), and O3 (533.5-535.84 eV) indicating C = O, C-OH/C-O-C and chemisorbed oxygen and adsorbed water, respectively. For the specific oxygen groups, both C1s and O1s scans displayed the increase of C-OH/C-O-C and the decrease of C = O after reaction, which might be ascribed to the transformation between the two groups. Therefore, the changes of the specific groups might impose effects on the catalytic performances of the catalysts (Indrawirawan et al., 2015;Tang et al., 2018). The influence of trace amounts of nitrogen on the surface of the catalyst was also investigated. The de-convolution XPS N1s spectra of CMK-3 and OCMK-3 before reaction are shown in Fig. S1(a-b). Obviously, after oxidation of CMK-3 using nitric acid, nitrogen-containing functional groups were introduced on the surface. The N1s peak of OCMK-3 was fitted into two sections at 399 and 405 eV, indicating pyridine and − NO 2 . It was also conferred in other studies that under a high-temperature environment, oxidation using HNO 3 could effectively generate nitrogencontaining species on the surface (Yang, et al., 2016).
FT-IR spectroscopy ( Fig. 4) was also used for further studies of the surface chemistry of CMK-3 and OCMK-3. The absorption of stretching vibration of C-O bond was denoted in the absorption band at 1100 cm −1 . Peak around 1640 cm −1 corresponded to C = O and -OH stretch vibrations and the characteristic peaks around 2922 cm −1 and 3440 cm −1 indicated C-H/O-H and -OH stretch vibrations, respectively. The relative intensities of 1100 cm −1 , 1640 cm −1 , and 2922 cm −1 bands in OCMK-3 were higher than those of the CMK-3 sample, indicating that more C-O, C = O, -OH and C-H/O-H functional groups were generated after the oxidation (Zolfaghari, et al., 2011). Therefore, FT-IR spectra further demonstrated the presence of more oxygen-containing functional groups onto the OCMK-3 surface, which were favorable for catalytic activation of PS (Sharon et al., 2018;Liu et al., 2019aLiu et al., , 2019b.

Catalytic performances of OCMK-3
It has been reported that a high concentration of acid treatment might lead to the collapse of the pore structure of the carbonaceous material, and excessive oxygen-containing functional groups could block the micropores and thus reduce the specific surface area and pore capacity, which might inhibit the oxidative degradation activity of the catalysts (Bazu et al., 2008;Ambaye et al., 2021). To obtain the optimum conditions for OCMK-3 synthesis for CIP removal, CIP removal experiments were performed using OCMK-3 prepared under various conditions, with the results shown in Tables S1-S3. Compared with the pristine CMK-3, both the adsorption and removal of ciprofloxacin of OCMK-3 could be enhanced only after treatment with 4 mol/L of nitric acid solution. As shown in Table S2, neither the adsorption capacity nor the catalytic efficiency of OCMK-3 could get enhanced when the synthetic temperature was 70 °C or 90 °C. And Table S3 indicates that both the adsorption and catalytic performances experienced a rise when the modification time increased from 15 to 45 min, but started to decrease when the time increased to 60 min. Therefore, compared with the pristine CMK-3, the adsorption and catalytic performance of OCMK-3 could be enhanced under certain synthetic conditions.
In practical application, the catalytic performance of the OCMK-3/PS system is often affected by environmental parameters. Therefore, the removal efficiency of CIP and the corresponding kinetic performance were studied by changing reaction parameters (e.g., the dosage of OCMK-3 Fig. 3 a XPS survey of CMK-3 before and after reaction; b XPS survey of OCMK-3 before and after reaction; c C1s and d O1s high-resolution scans of CMK-3 before and after reaction; e C1s and f O1s high-resolution scans of OCMK-3 before and after reaction and persulfate, solution pH and CIP initial concentration, temperature, and coexistent ions). Firstly, the effect of the dosage of OCMK-3 was investigated, with the results shown in Fig. 5a. PS alone could hardly oxidize CIP without activation, with the removal rate only about 8%. Herein, with the increase of dosage from 0.1 to 0.4 g/L, the removal efficiency of CIP was enhanced, which could be attributed to more activators and active sites provided by the increased amounts of OCMK-3. However, K obs did not increase when the dose of OCMK-3 was above 0.3 g/L, which might be limited by the dose of PS. The influence of the PS dosage was also investigated. First, in order to eliminate the influence of adsorption, the adsorption experiment was carried out for 3 h without the addition of PS. As shown in Fig. 5b, the adsorption process reached equilibrium within 1 h and about 40% of CIP was removed. After adsorption reached equilibrium at 1 h, PS was added. It could be observed from Fig. 5b that the concentration of CIP decreased significantly. It means that in addition to adsorption, chemical reactions were also involved in the system, which was affected by the PS dosage. The Langmuir-Hinshelwood (L-H) kinetic model was used to fit the removal curve of CIP. Formulas (1) and (2) are (L-H) variable forms.
where C 0 and C t are the initial CIP concentration and the concentration at time t, respectively. K obs is the pseudo-firstorder reaction rate constant (min −1 ). When the PS dosage rose from 2 to 8 mM, K obs increased from 0.00189/min to 0.00364/min. An increase in K obs at the beginning might be resulted from more reactive free radicals through increasing PS dosage. However, K obs decreased when PS dosage continued to increase, which could be due to the self-quenching effect between excessive free radicals (Guo et al., 2017a(Guo et al., , 2017b. The effect of pH on the removal of CIP was studied over the pH range of 3.0 and 5.0, as such concentration of CIP was insoluble under neutral and alkaline conditions. As shown in Fig. 5d, the maximum removal of CIP occurred when pH was 4. To further understand the effects of pH, zeta potential for OCMK-3 under various pH values was analyzed. As shown in Fig. 5c, the isoelectric point of OCMK-3 is 2.1. Thus, when the solution pH is more than 2.1, the surface of OCMK-3 is negatively charged (Saleh et al., 2021). When the solution pH was 3 or 4, CIP existed as a cationic form CIP + as a result of the protonation of the secondary amine on the piperazine group, while the OCMK-3 surface was oppositely charged. However, OCMK-3 was more negatively charged when pH was 4, resulting in the enhanced adsorption amounts. When pH reached 5, CIP could exist as zwitterionic (CIP ± ). In this stage, both CIP molecules and OCMK-3 surfaces were negatively charged, resulting in a depressed adsorption process (Peng et al., 2015).
As demonstrated in Fig. 6a, CIP removal efficiency decreased from 89 to 57% with the increase of CIP concentration from 100 to 300 mg/L. As the concentration of pollutants rose, the increased competition of CIP molecules might be the reason for the reduced removal rate (Ye et al., 2019;Ahammad et al., 2021).
To compare CIP removal by the prepared CMK-3 and OCMK-3, comparative experiments were carried out, with the results described in Fig. 6b. About 42% and 74% of CIP were removed in 3 h by CMK-3 and OCMK-3, respectively. The removal efficiency of OCMK-3 was about 1.76 times that of CMK-3, which further verified that oxidation treatment could increase the activity of the catalyst. To evaluate the oxidative degradation performance, the mineralization degree of organic contaminants was investigated. It could be clearly observed from Fig. 6c that TOC removal by the OCMK-3/PS system was about 11% higher than that by the CMK-3/PS system. This suggested that the OCMK-3/PS system could not only enhance the removal of CIP but also exhibit higher mineralization efficiency.
The effects of temperature on CIP removal were further investigated, with the results shown in Fig. S2(a). At 25 °C, 74% of CIP could be removed. When the temperature increased to 45 °C, about 81% of CIP could be removed. Thus, it could be inferred that elevated temperatures could lead to enhanced PS activation. There might be two reasons to interpret this phenomenon. Firstly, a higher temperature could facilitate the decomposition of PS and therefore accelerated the removal of CIP. Secondly, the elevated temperature could also promote the adsorption process, which was favorable for the electron transfer between OCMK-3 and PS (Tang et al., 2018).
The trace amounts of inorganic anions often existed in real wastewaters, which might affect catalytic behaviors in very complex ways. Figure S2(b, c) shows the effects of the different concentrations of HCO 3 − and Cl − on CIP removal. As shown in Fig. S2(b), when the HCO 3 − concentration rose from 0 to 10 mM, the removal of CIP was also increased. The reason for the increase might lie in that the addition of HCO 3 − might increase the solution pH, thus reducing the solubility of CIP molecules (Wu et al., 2013). However, when the HCO 3 − concentration rose to 20 mM, CIP removal efficiency slightly decreased. As HCO 3 − could act as an electron donor, it might compete with CIP for reacting with SO 4 · − thus hindered CIP removal (Eqs. (3)-(4)) (Wen et al., 2018;Yong et al., 2018).
As demonstrated in Fig. S2 (c.), the presence of 5 mM and 10 mM chlorides slightly decreased the remove rate, as Cl − could act as radical scavengers and combine with SO 4 · − and ·OH to generate Cl· and Cl 2 · − (Eqs. (5)-(8)) (Guo et al., 2021), whereas an obvious increase of kobs could be observed with the chloride concentration continued to increase to 20 mM. Herein, besides scavenging SO 4 · − and ·OH, superabundant Cl − could also take part in the removal of CIP, which might act as an electron donor to afford electrons to PS; thus, the active chlorine species and sulfate radicals were generated under the catalysis of OCMK-3 (Oyekunle et al., 2021).

Identification of radicals
Herein, to investigate the effects of radicals, radical quenching reaction was analyzed. SO 4 · − and ·OH have been regarded as two typical reactive species involved in PS activation. Studies had demonstrated the strong quenching ability of methanol for ·OH (k ·OH = 9.7 × 10 8 M −1 s −1 , kSO 4 · − = 3.2 × 10 6 M −1 s −1 ), while EtOH served as an effective quencher for SO 4 · − (k ·OH = (1.6-7.8) × 10 7 M −1 s −1 , kSO 4 · − = (1.2-2.8) × 10 9 M −1 s −1 ) (Buxton et al., 1988). Thus, methanol and EtOH were utilized as quenchers, with the results shown in Fig. 7a-b. The CIP removal was restrained from the existence of either MeOH or EtOH, with a decrease of kobs from 0.0044 to 0.00131 min −1 (n[MeOH/ PS] = 1000) and 0.00134 min −1 (n[EtOH/PS] = 1000), respectively. It could be referred that both SO 4 · − and ·OH were generated and involved in the reaction. In addition, 1,4-benzoquinone (BQ) was introduced as a quencher to identify the existence of superoxide anion radicals (O 2 · − ), with the results shown in Fig. 7c. Slight inhibition in the reaction by BQ could be observed, with a decrease of kobs from 0.0044 to 0.0034 min −1 , indicating that O 2 · − was also involved in CIP removal.
Electron paramagnetic resonance (EPR) spectroscopy was employed to further confirm the existence of radicals. Figure 7d shows the characteristic peaks of DMPO·OH (hyperfine splitting constants of aN = aH = 14.9 G) and DMPO·SO 4 (aN = 13.2 G, aH = 9.6 G, aH = 1.48 G and aH = 0.78 G) (Wang et al., 2015), in agreement with the quenching experiment, and it could be further inferred that ·OH and SO4· − were involved in the reaction (Jiang et al., 2020). Meanwhile, six characteristic peaks of DMPO-O 2 • − could be observed in Fig. 7e, further confirming the occurrence of O 2 · − (Cdab et al., 2021).

Degradation products and pathways
In this study, 9 transformation intermediates of CIP were identified by the LC-MS technique (Table S2 and Fig. S3). Through interpreting the molecular ion masses, analyzing ESI( +) MS spectra and comparing with previous studies (Guo and Li, 2013;Hong-Guang Guo, 2013;Guo et al., 2016), three major reaction pathways for CIP degradation were proposed and are illustrated in Fig. 8.
Pathway Ι was initiated by cleaving piperazine ring, and then ·OH, SO4· − , and O 2 · − attack caused stepwise oxidation. This route is in agreement with the study that the piperazine ring was determined to be the most active site for the free radical attack (Jiang, et al., 2016). Oxidation of piperazine led to ring opening, resulting in the dialdehyde derivative C1. C1 converted into C2 through removing one formaldehyde from the open piperazine ring, and then transformed into C3 through further loss of second formaldehyde in the open piperazine ring. The compound C3 could be further oxidized into C4 by losses of secondary amine nitrogen (Mahdi-Ahmed and Chiron, 2014).
In pathway II, CIP was firstly converted to C5 by hydroxylation, and then transformed into C6 through OH/F substitution and further hydroxylation. C7 was formed by substituting -COOH with -OH in C6. This pathway was also proposed in other studies (Zhu et al., 2019).
In pathway III, CIP might convert into C8 by defluorination. On the one side, product C9 was produced through piperazine epoxidation from C8. On the other side, the C8 might also form C5 through hydroxyl substitution of fluorine, and then with the similar route to pathway II, through further hydroxylation and decarboxylation process were similar to pathway II, the same intermediate C6 and C7 were generated. This pathway is in agreement with that proposed by Yja B. et al.(Yja et al., 2014).

Electron transfer pathways
Based on the above analysis, the mechanism of CIP removal using the OCMK-3/PS system was proposed and shown in Scheme 1. The unique electronic circumstance on the surface of OCMK-3 resulted in the defective edge sites and free-flowing π electrons of the sp2-hybridized carbon network, which might impose effects on the electron configuration of PS and weaken the O-O bond to generate metastable complexes (Duan et al., 2015). In addition, the large specific surface area and porous structure of OCMK-3 could facilitate the adsorption of PS and CIP molecules onto its surface, and thus promoted their contact and subsequent reactions. Accordingly, two possible pathways were proposed in the following evolution.
The nonradical pathway In PS nonradical activation systems, when the electron-rich CIP molecule approached the activated PS molecule, a ternary system of electron donor (CIP)-mediator (OCMK-3)-acceptor (PS) was established. PS was used up in CIP oxidation by getting two electrons from organics (Eq. (12) (Yao et al., 2019)), and the radical reactions did not occur due to the steady state of PS (Chen et al., 2019;Zhang et al., 2020).
The radical and nonradical oxidation processes both took part in the reaction in this system. OCMK-3 acted as an electron transfer mediator, which could facilitate the transfer of electrons from CIP or H 2 O to PS. The electron transfer pathways were conformed to the CMK-3/PS system studied previously (Tang et al., 2018;Liu et al., 2019aLiu et al., , 2019b, indicating that OCMK-3 did not change the degradation pathway.

Stability of OCMK-3 and real wastewater application
The stability of the catalyst OCMK-3 for continuous PS activation was evaluated for further application. The experiment was carried out for four cycles to test the reusability of OCMK-3, with the results shown in Fig. S4. A slight decrease in CIP removal efficiency could be observed after each cycle, which might be attributed to the oxidation of the surface during the reaction, which changed its chemical composition. Moreover, the interaction between PS and the active sites was hindered owing to the adsorption of oxidation intermediates, thus imposing effects on the surface chemistry of OCMK-3 and the electron transfer from OCMK-3 to PS. However, the CIP removal by regenerated OCMK-3 after four cycles was 89.2% of that by the fresh OCMK-3, demonstrating that OCMK-3 could be regenerated and reused (Jya, et al., 2020a(Jya, et al., , 2020b. To test the practicability of the OCMK-3/PS system, real pharmaceutical wastewater was used as a target to test the degradation efficiency of the OCMK-3/PS system in the complex environment, with the results shown in Fig. 9. Multiple peaks could be observed from the spectrum of the original sample in Fig. 9a, showing complex components of the water with high concentrations. The high fluorescence intensities could also be identified in Fig. 9a, which were related to fulvic-like compounds, indicating the poor biodegradability and serious pollution of the wastewater (Wang,(12) al., 2016). After the adsorption, the number and area of peaks in Fig. 9b were reduced, indicating that most fulviclike compounds were adsorbed by OCMK-3. After 2 h of reaction with OCMK-3/PS (Fig. 9c), peaks were almost disappeared, indicating that the dissolved organic matter was almost removed after treatment, and the biodegradability of water was greatly improved (Yu et al., 2020). In addition, to further evaluate the mineralization of the organics, TOC contents of the wastewater were evaluated, with the results shown in Fig. 9d. It could be observed that only 5% of TOC were adsorbed, while about 75% of TOC were removed in the OCMK-3/PS system, indicating the high TOC removal efficiency of the system. The results further convinced the high removal efficiency of the OCMK-3/PS system in both laboratory-prepared polluted water and practical wastewater.

Conclusions
In this study, a novel carbon catalyst, OCMK-3, was synthesized for the activation of persulfate to remove CIP. Compared with CMK-3, OCMK-3 exhibited higher CIP removal efficiency. The introduction of oxygen-containing groups increased the specific surface area and pore volume, provided multiple oxygen-containing groups, and had good electron transfer ability, thus resulting in the excellent adsorption performance and catalytic performance of OCMK-3. During CIP degradation, various intermediates were identified and the degradation pathways were proposed. Both radical and nonradical pathways were involved in the oxidation process of CIP, and OCMK-3 acted as a mediator to promote the electron transfer between CIP (electron donor) and PS (electron acceptor), leading to the oxidative degradation of CIP. Additionally, the actual treatment of pharmaceutical wastewater using the OCMK-3/PS system further proved its high efficiency in organic wastewater treatment.