Co-Embedded N-Doped Carbon Composites Derived from CoZn-ZIFs for Peroxymonosulfate Activation Toward Efficient Tetracycline Degradation

Highlights

What are the main findings?

• A simple and environmentally friendly preparation method for CoZn-ZIFs was proposed.
• The novel PMS activator Co-CNTs@NC was successfully obtained via pyrolysis of CoZn-ZIFs.

What are the implications of the main findings?

• The Co-CNTs@NC/PMS system achieved efficient degradation of TCH over a wide pH range.
• The degradation mechanism dominated by ¹O₂ for TCH degradation was confirmed.

Abstract

Zeolite imidazolate frameworks (ZIFs)-derived carbon materials have garnered widespread attention as peroxymonosulfate (PMS) activators in removing antibiotics because of their excellent catalytic performance. However, most carbon materials derived from ZIFs exhibit limited efficacy in treating high-concentration (>10 ppm) antibiotic wastewater, and their synthesis methods are environmentally unfriendly. Herein, we develop a simple and environmentally friendly preparation method to synthesize a new type of nitrogen-doped carbon-supported carbon nanotubes coated with cobalt nanoparticle (Co-CNTs@NC) composites via high-temperature calcination of cobalt–zinc bimetallic ZIFs. The material characterization results confirm the successful preparation of Co-CNTs@NC composites featuring a high specific surface area (512.13 m²/g) and a Co content of 5.38 wt%. Across an initial pH range of 3.24–9.00, the Co-CNTs@NC/PMS catalytic system achieved over 84.17% degradation of 20 mg/L tetracycline hydrochloride within 90 min, demonstrating its favorable pH tolerance. The singlet oxygen-dominated degradation mechanism was confirmed by quenching experiments and electron paramagnetic resonance characterization. This work can provide technical guidance and reference significance for the preparation of metal–carbon materials derived from ZIFs with excellent efficiency of removal of high-concentration antibiotics.

1. Introduction

As a typical tetracycline antibiotic, tetracycline hydrochloride (TCH; the molecular structure of TCH is illustrated in Figure 1) is extensively employed in human medicine, animal husbandry, and aquaculture owing to its broad-spectrum antibacterial activity, cost-effectiveness, and excellent bioavailability [1]. However, tetracycline is poorly metabolized and absorbed by organisms, with over 70%–80% persisting in the environment as the parent compound or its primary metabolites [2]. Long-term exposure to water contaminated with TCH can promote the development of antibiotic resistance genes in microorganisms, which poses a serious threat to public health [3,4]. Due to its stable chemical properties and high-water solubility, traditional wastewater treatment technologies are inefficient in removing TCH [5]. Therefore, the development of efficient and cost-effective technologies for TCH removal is imperative.

Peroxymonosulfate-based advanced oxidation processes (PMS-AOPs) have emerged as promising technologies for the degradation of refractory organic pollutants, owing to their high oxidation potential, broad pH adaptability, and the long half-life of the generated radicals [6,7,8,9]. However, PMS alone exhibits limited efficacy in pollutant degradation, necessitating external activation to generate reactive oxygen species with strong oxidizing capacity [10]. Compared with other activators, metal–carbon composites supported by conventional carbon nanotubes (CNT_S) or graphene oxide (rGO) (e.g., Co₉S₈@S-N-rGO [11], Co@N-O-CNTs [12], and FeCoS@N-rGO [13]) have been extensively investigated because of their large specific surface area, excellent catalytic performance, superior stability, and good biocompatibility. However, these conventional carbon supports are typically synthesized from raw materials derived from fossil fuels, including coal and crude oil [14]. In light of the ongoing energy crisis, there is a pressing need to develop novel N-doped carbon materials as sustainable alternatives to these conventional carbon materials. Additionally, the activation mechanisms of PMS primarily involve radical pathways triggered by transition metal activation and nonradical pathways dominated by nitrogen (N)-doped carbonaceous materials [15,16]. Owing to its electrophilic properties, singlet oxygen (¹O₂) exhibits remarkable efficiency and selectivity in the abatement of antibiotic contaminants [17,18]. To date, ¹O₂ oxidation as the predominant non-radical mechanisms have been widely confirmed in N-doped carbon material/PMS systems [19,20]. Accordingly, the rational design and fabrication of novel N-doped carbon materials targeting PMS activation for the efficient degradation of high-concentration TCH wastewater are of profound research value.

As a representative category of metal–organic frameworks, zeolite imidazolate frameworks (ZIFs) possess permanent porosity, tunable pore size and structure, relatively low synthesis cost, and straightforward preparation procedures [21,22]. These distinctive advantages render ZIFs-derived materials promising catalysts for PMS activation [23,24]. For instance, Wang and co-workers successfully synthesized a cobalt-based catalyst featuring dual active sites (Co_0.75Zn_0.25-NC) through the calcination of cobalt–zinc bimetallic ZIFs (CoZn-ZIFs). The Co–Co site and Co–N site separately activated PMS to generate sulfate radicals and ¹O₂, thus enabling the efficient degradation of TCH [25]. Hua et al. reported that N-doped porous carbon-encapsulated Co nanoparticles (denoted as CZA-1000) were successfully fabricated via the pyrolysis of a Co-ZIF-8@anionic polyacrylamide network composite precursor. Benefiting from its abundant defect structures and encapsulated Co nanoparticles, CZA-1000 efficiently activated PMS to generate free radicals and ¹O₂, achieving a tetracycline removal efficiency of 99.8% within 30 min [26]. Yao and co-workers described the fabrication of Zn₄Co₁-C materials with Co-N₄ coordination configurations via the pyrolysis of Zn_xCo_1−x-ZIFs precursors. Co-N₄ sites can efficiently activate PMS to generate ¹O₂. Accordingly, Zn₄Co₁−C delivers a remarkable selectivity of 98% for phenol removal in aqueous systems containing a mixture of phenol and benzoic acid [27]. Despite the remarkable PMS-activation efficacy of ZIFs-derived metal–carbon composites, their preparation and practical deployment suffer from two major drawbacks. Firstly, the synthesis of ZIFs typically relies on toxic and volatile methanol as the solvent, along with lengthy solvothermal reactions under harsh high-temperature and high-pressure conditions, making the protocol environmentally unfriendly, unsafe, and time-consuming. Additionally, these materials are primarily applied for the removal of traditional organic pollutants including dyes, phenols, and bisphenol A. Nevertheless, their catalytic efficiency toward high-concentration antibiotic pollutants still warrants in-depth investigation [28].

To address these two major drawbacks, we report a facile, eco-friendly, and rapid self-assembly approach to synthesize a novel CoZn-ZIFs nanosheet precursor. Subsequently, N-doped carbon-supported CNTs coated with Co nanoparticles (Co-CNTs@NC) composites were obtained via the high-temperature calcination of CoZn-ZIFs. The physicochemical properties of Co-CNTs@NC were systematically investigated by means of X-ray diffraction (XRD), transmission electron microscopy (TEM), high-resolution transmission electron microscopy (HRTEM), X-ray photoelectron spectroscopy (XPS), and Raman spectroscopy. A series of control experiments were carried out to investigate the catalytic efficacy of Co-CNTs@NC toward PMS activation, as well as to identify the primary active sites. Systematic optimization was conducted on critical operational parameters (e.g., initial TCH concentration, PMS dosage, Co-CNTs@NC dosage, reaction temperature, and initial pH) to evaluate their impacts on TCH degradation efficiency. Concurrently, the reactive oxygen species accountable for TCH degradation were identified via a combination of quenching experiments and electron paramagnetic resonance spectroscopy, based on which a potential degradation mechanism was proposed. This study seeks to develop an eco-friendly, scalable synthetic protocol for high-performance PMS activators, while advancing the application of MOF-derived composites in the remediation of antibiotic-contaminated wastewater.

2. Materials and Methods

2.1. Materials

2-Methylimidazole, cobalt nitrate hexahydrate (Co(NO₃)₂·6H₂O), cetyltrimethylammonium bromide (CTAB), p-Benzoquinone (PBQ), tert-butanol (TBA), potassium monopersulfate triple salt (PMS, KHSO₅∙0.5KHSO₄∙0.5 K₂SO₄), L-histidine (L-His), and tetracycline hydrochloride (TCH) were obtained from Aladdin Biochemical Technology Co., Ltd. (Shanghai, China). Zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O), methanol (MeOH), and sodium hydroxide (NaOH) were supplied from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). All reagents were of analytical grade and employed without additional purification. Deionized water (18.2 MΩ·cm resistivity) was utilized throughout the entire experimental process.

2.2. Preparation

The synthesis processes of CoZn-ZIFs and Co-CNTs@NC are illustrated in Figure 2. CoZn-ZIFs were fabricated through a room-temperature self-assembly route in an aqueous medium, where the molar ratio of 2-methylimidazole to metal salts was maintained at 20:1. This procedure was adapted from our previously reported protocol with minor modifications [29]. For comparison, CoZn-ZIFs-30 and CoZn-ZIFs-40 were synthesized by adjusting the molar ratio of 2-methylimidazole to metal salts (Co²⁺ + Zn²⁺) to 30:1 and 40:1, respectively. Zn-ZIFs (without Co²⁺) were prepared using the same procedure of CoZn-ZIFs with a Zn²⁺:Co²⁺ molar ratio of 1:0. N-doped carbon-loaded CNTs coating Co nanoparticle composites were synthesized by calcining above precursors. Specifically, 1.0 g of CoZn-ZIFs was weighed, ground into fine powder, and loaded into a porcelain boat. The boat was then placed in a tube furnace. Under an argon atmosphere, the temperature was first increased up to 500 °C at a heating rate of 10 °C min⁻¹, followed by a subsequent increase to 1000 °C at 5 °C min⁻¹. After holding this final temperature for 3 h, the system was cooled down to 20 °C. The resulting black samples were collected, ground, and labeled as Co-CNTs@NC. Following the synthetic protocol for Co-CNTs@NC, Co-CNTs@NC-30, Co-CNTs@NC-40, and NC were successfully fabricated via the calcination of CoZn-ZIFs-30, CoZn-ZIFs-40, and Zn-ZIFs precursors, respectively. The synthetic details of all samples are summarized in Text S1 of the Supplementary Materials (SM).

2.3. Characterization

The crystal phase and morphologies of CoZn-ZIFs and Co-CNTs@NC were analyzed using XRD (D8 ADVANCE, Bruker, Karlsruhe, Germany), TEM (JEM 2100 plus, Electron Optics Laboratory, Tokyo, Japan), and scanning electron microscopy (SEM, Zeiss Sigma 360, Oberkochen, Germany), respectively. The XPS data of Co-CNTs@NC was obtained with a Thermo Scientific K-Alpha spectrometer (Thermo, Waltham, MA, USA). The XPS Peak 4.1 software was used to analyze the elemental composition and valence states of Co-CNTs@NC. The graphitization degrees of Co-CNTs@NC and NC were evaluated using Raman spectroscopy (LabRAM Odyssey, HORIBA, Tokyo, Japan) with an excitation wavelength of 532 nm.

2.4. TCH Degradation Experiments

The experimental procedures for TCH degradation, the investigation of reaction condition effects on TCH degradation, and the exploration of free radical types and roles are detailed in Texts S2–S4 of the Supporting Materials (SM), respectively. According to the UV spectral scanning data of TCH (Figure S1), the absorbance of the reaction mixture was determined at 358 nm with an ultraviolet-visible spectrophotometer (UV-Vis, T600A, Beijing Purkinje General Instrument Co., Ltd., Beijing, China).

3. Results and Discussion

3.1. Material Characterization

3.1.1. Morphological Characterization

The morphological features of CoZn-ZIFs and Co-CNTs@NC were investigated via SEM. Figure S2 clearly shows that CoZn-ZIFs possess an irregular lamellar microstructure. As shown in the SEM micrograph of Co-CNTs@NC at 1 μm magnification (Figure S3a), the carbonized material largely preserves its lamellar structure, concomitantly displaying the formation of partial CNT domains. At a higher magnification of 200 nm (Figure S3b), the SEM image distinctly visualizes the CNT morphology, where Co nanoparticles are encapsulated at the tips of the nanotubes. To further verify the successful fabrication of Co nanoparticles encapsulated within CNTs, TEM and HRTEM characterizations were performed. Figure 3a,b reveal that Co nanoparticles encapsulated within CNTs are successfully anchored on the surface of Co-CNTs@NC. This structural feature can be ascribed to the continuous migration of Co atoms during high-temperature calcination, which catalyzes the growth of CNTs [30]. The HRTEM micrographs of Co-CNTs@NC in Figure 3c show distinct lattice fringes with interplanar spacings of 0.20 and 0.34 nm, corresponding to the (111) facet of metallic cobalt and the (002) facet of graphitic carbon, respectively [31]. The characterization results of high-angle annular dark field scanning TEM (HAADF-STEM) and the corresponding element mapping images (Figure 3d) verify the uniform distribution of Co, N, and C in the Co-CNTs@NC with good dispersibility. The morphological analysis results confirm the successful preparation of Co-CNTs@NC.

3.1.2. XRD and Raman Analysis

To gain insights into the crystal structural characteristics of Zn-ZIFs and CoZn-ZIFs, XRD analysis was conducted on the two samples. As illustrated in Figure S4, both Zn-ZIFs and CoZn-ZIFs possessed well-resolved crystalline architectures with high crystallinity. The incorporation of Co species did not alter the crystal framework of Zn-ZIFs, confirming the remarkable structural stability of the as-prepared CoZn-ZIFs. Figure 4a shows that a broad diffraction peak at ~26° appears in all samples, and is corresponding the (002) crystal plane of graphitic carbon (JCPDS No. 75-1621) [32]. The intensity of this peak is significantly higher in Co-CNTs@NC, Co-CNTs@NC-30, and Co-CNTs@NC-40 than in NC, which demonstrates that the introduction of Co promoted the graphitization of carbon matrices during high-temperature calcination. Additionally, the XRD patterns of Co-CNTs@NC, Co-CNTs@NC-30, and Co-CNTs@NC-40 exhibit three distinct diffraction peaks located at 44.04°, 51.48°, and 75.71°, assignable to the (111), (200), and (220) crystal planes of metallic Co (JCPDS No. 15-0806) [33]. The diffraction peaks corresponding to cobalt oxides are not detected, which can be attributed to the enhanced stability of Co nanoparticles resulting from the CNT coating. The XRD results confirm that the three composite materials are successfully synthesized by tailoring the molar ratio of metal precursors to organic ligands in this work.

Raman spectroscopy was further utilized to quantify the graphitization degree of Co-CNTs@NC and NC composites. As presented in Figure 4b, the Raman spectra of the Co-CNTs@NC and NC reveal two prominent peaks at ~1350 (D-band) and ~1580 cm⁻¹ (G band), yielding the I_D/I_G intensity ratios of 0.93 and 1.05, respectively. Significantly, the I_D/I_G ratio of Co-CNTs@NC is considerably lower than that of NC, which attests to the higher graphitization degree of the former. In general, a high graphitization degree is conducive to enhancing the electrical conductivity of carbon-based materials, and thus contributes to the improvement of their catalytic performance [34]. This finding is consistent with the XRD characterization results.

3.1.3. XPS Analysis

The element composition and chemical states of Co and N in the Co-CNTs@NC were systematically characterized using XPS. As shown in Figure 5a, the characteristic peaks of C 1s, N 1s, Co 2p, O 1s, and Zn 2p are resolved. Figure S5a displays a faint XPS signal corresponding to Zn 2p, which can be attributed to the evaporation of Zn (boiling point: 907 °C) during calcination at 1000 °C [35]. To further confirm the nearly complete evaporation of Zn, the contents of Co and Zn elements in Co-CNTs@NC were quantified via inductively coupled plasma mass spectrometry (ICP-MS), with the corresponding data summarized in Table S1. Co-CNTs@NC exhibited a high Co content (5.38 wt%), which is favorable for TCH degradation. Meanwhile, the Zn content was merely 0.03 wt%, indicating that Zn was almost completely volatilized during calcination at 1000 °C. The peak fitting of the O 1s XPS spectrum in Figure S5b resolves two well-defined peaks at binding energies of 531.62 eV and 532.91 eV, corresponding to carbonyl moieties (C=O) and hydroxyl groups (C–O–H) in the composite matrix, respectively [36]. The Co 2p XPS spectrum of Co-CNTs@NC (Figure 5b) are deconvoluted into two spin–orbit doublets (Co 2p₃/₂ and Co 2p₁/₂), each comprising peak pairs corresponding to metallic Co (778.78 eV/793.47 eV) and Co²⁺ (780.93 eV/796.72 eV), along with two satellite peaks located at 785.69 eV and 803.08 eV [37]. Similarly, no XPS peaks corresponding to cobalt oxides were observed, thereby verifying that the CNT coating improves the antioxidative stability of Co nanoparticles. The high-resolution N 1s XPS spectrum in Figure 5c displays three well-defined characteristic peaks, corresponding to pyridinic N (398.75 eV), pyrrolic N (400.93 eV), and graphitic N (405.57 eV), respectively [38]. As illustrated in Figure 5d, the high-resolution C 1s spectrum verifies the existence of C=C (284.51 eV), C–N (285.76 eV), and C=O (287.58 eV) moieties [39]. Collectively, the combined results from TEM, HRTEM, XRD, Raman, and XPS characterizations corroborate the successful fabrication of the Co-CNTs@NC composite.

3.1.4. Bet Analysis

The specific surface area and pore structure of NC and Co-CNTs@NC were characterized by N₂ adsorption–desorption isotherms. As depicted in Figure 6a, Co-CNTs@NC exhibits a reversible type IV isotherm at relative pressures (P/P₀) > 0.4, coupled with long and narrow hysteresis loops, which is characteristic of a mesoporous structure [40]. In the high-pressure region (p/p₀ > 0.9), NC shows a drastic upturn, confirming the presence of macro-porous structures [41]. Figure 6b provides complementary evidence confirming the macro-porous nature of NC and the mesoporous characteristics of Co-CNTs@NC. The Brunauer–Emmett–Teller (BET) surface area and pore volume of Co-CNTs@NC (512.13 m²/g, 0.78 cm³/g) are slightly lower than those of the N-C sample (738.07 m²/g, 1.17 cm³/g). Co-CNTs@NC still exhibits a substantial surface area and abundant pore structure, which are conducive to increasing the exposure of active sites.

3.2. Catalytic Degradation of TCH

3.2.1. Efficiency Comparison of Different Reaction Systems

To investigate material performance in PMS-activated TCH degradation and key active sites, a systematic evaluation of TCH removal efficiencies was performed across multiple catalytic systems. The adsorption efficiencies of Co-CNTs@NC for TCH were determined to be 6.01% and 8.21% at reaction time of 30 and 60 minutes (min), respectively (Figure S6). The results demonstrate that the adsorption of TCH onto Co-CNTs@NC can attain adsorption equilibrium within 30 min. As shown in Figure 7, the adsorption efficiencies of Co-CNTs@NC and NC for TCH reached 7.52% and 0.50% within 30 min, respectively. The adsorption efficiencies of the two reaction systems within 30 min were both below 10%, indicating that the adsorption role is almost negligible in TCH removal. The Co-CNTs@NC/PMS reaction system can remove 43.46% and 86.02% of TCH in 2 and 60 min, respectively. The removal efficiency at 60 min exhibited a remarkable superiority over that of PMS alone (44.04%) and the NC/PMS reaction system (43.92%), which can be attributed to the high Co content in Co-CNTs@NC. The high Co content in Co-CNTs@NC is responsible for the formation of abundant active sites, thereby enhancing PMS-activated TCH degradation efficiency. Concurrently, Figure S7 shows that the Co-CNTs@NC/PMS system possesses the highest pseudo-first-order rate constant (k_obs = 0.0859 min⁻¹) within 10 min, 13.22-fold and 14.32-fold higher than those of NC/PMS (0.0065 min⁻¹) and PMS alone (0.0060 min⁻¹), respectively. Consequently, the above results confirm that Co-CNTs@NC exhibits superior efficiency in PMS-activated TCH degradation, where Co nanoparticles are the dominant active sites. After comprehensive evaluation of key parameters including catalyst dosage, PMS dosage, and initial tetracycline concentration, Co-CNTs@NC/PMS delivers superior degradation kinetics and catalytic performance, comparable to the performance metrics of previously reported metal–carbon/PMS activation systems for tetracycline degradation (Table S2).

In addition, to investigate the effect of 2-methylimidazole to metal salt molar ratio on TCH removal efficiency, CNTs@NC-30/PMS and Co-CNTs@NC-40/PMS system were constructed separately, and the results are listed in Figure 7. The adsorption efficiencies of Co-CNTs@NC-30 and Co-CNTs@NC-40 for TCH were 28.14% and 28.67% within 30 min, respectively. The removal efficiencies of two reaction systems were determined to be 83.27% and 83.70% within 60 min, respectively, both lower than that of the Co-CNTs@NC/PMS system. As a result, increasing the dimethylimidazole to metal salt molar ratio only enhanced adsorption efficiency, whereas high adsorption efficiency impairs TCH deep removal. Therefore, considering both the degradation efficiency and economic cost, Co-CNTs@NC (2-methylimidazole: metal salts = 20:1) was selected as the optimal catalyst for subsequent experiments.

3.2.2. Effect of Different Experimental Conditions

Contaminant concentration is widely recognized as a key factor governing the performance of advanced oxidation systems. Accordingly, the influence of the initial TCH concentration on TCH removal was investigated first. As can be seen from Figure 8a, with the TCH concentration increasing from 10 mg/L to 30 mg/L, the TCH degradation efficiency decreased from 89.83% to 76.91%. Meanwhile, the k_obs of TCH decreased from 0.1143 to 0.0625 min⁻¹ (Figure S8a). The reason may be attributed to the limited availability of active sites on Co-CNTs@NC. Thus, 20 mg/L of TCH was set as the initial concentration for subsequent experiments to ensure good degradation efficiency.

The effect of PMS dosage on TCH degradation was investigated. After 90 min of reaction, the degradation efficiency of TCH increased from 77.14% to 86.33% and 88.15% with PMS dosage increasing from 0.05 g/L to 0.1 and 0.2 g/L (Figure 8b). Correspondingly, Figure S8b reveals that the k_obs increased from 0.0565 to 0.0962 and 0.1090 min⁻¹ with increasing PMS dosage. As shown by the results, increasing the PMS dosage can promote TCH removal, which is attributed to two factors. It strengthens the direct oxidation capacity of PMS for TCH removal and promotes Co-CNTs@NC to activate PMS for more active species generation [42]. In contrast, no notable changes were observed in the TCH degradation efficiency and k_obs value when the PMS dosage was increased from 0.1 to 0.2. Therefore, to balance degradation efficiency and cost-effectiveness, 0.1 g/L was determined as the optimal PMS concentration.

The Co-CNTs@NC dosage as an important variable was also investigated (Figure 8c). After 30 min of adsorption, the TCH adsorption efficiencies for Co-CNTs@NC at dosages of 0.05, 0.1, and 0.15 g/L were 3.01%, 7.52%, and 12.39%, respectively. In the degradation process, increasing Co-CNTs@NC dosage promoted the TCH degradation efficiency to rise from 27.62% to 42.21% within 1 min. Within 90 min, the TCH degradation efficiencies reached 85.74%, 86.33%, and 86.53%, respectively, with corresponding k_obs values of 0.0453, 0.0962, and 0.1059 min⁻¹ under the three dosage conditions (Figure S8c). No significant differences in degradation efficiency or k_obs were observed between the 0.1 and 0.15 g/L dosages. Therefore, considering both degradation performance and cost-effectiveness, 0.1 g/L was selected as the Co-CNTs@NC dosage for subsequent reactions.

The effect of initial pH on TCH degradation is illustrated in Figure 8d. The TCH adsorption efficiencies of Co-CNTs@NC under various pH conditions were fairly low, ranging from 5.23% to 10.16%, suggesting that pH had an insignificant impact on adsorption behavior. Within 90 min of reaction, the TCH degradation efficiencies reached 85.35% (pH 3.24), 86.27% (pH 4.65), 86.33% (pH 7.04), and 84.17% (pH 9.00), respectively. Figure S8d reveals that the k_1obs values for initial pH values of 3.24, 4.65, 7.04, and 9.00 were 0.0628, 0.0996, 0.0962, and 0.0816 min⁻¹, respectively, whereas the k_2obs values under these pH conditions were 0.0241, 0.0262, 0.0292, and 0.0161 min⁻¹. These results demonstrate that the Co-CNTs@NC/PMS system possesses excellent adaptability across a wide pH range. Comparison of TCH degradation efficiencies and k_obs across different pH values confirms that Co-CNTs@NC/PMS achieves the optimal degradation performance under neutral conditions. Considering that the pH of most real-world wastewater samples is close to neutral, an initial pH of 7.04 was determined to be the optimal operational parameter.

Finally, the influence of temperature on the TCH degradation was examined. Following a 30 min adsorption at 15, 25, and 35 °C, the TCH adsorption efficiencies of Co-CNTs@NC were measured to be 9.67%, 7.52%, and 12.87%, respectively, indicating that reaction temperature exerted a negligible impact on TCH adsorption behavior (Figure S9). The corresponding TCH degradation efficiencies reached 84.12%, 86.33%, and 88.68%, respectively, within 90 min. No statistically significant differences were observed in the degradation efficiencies across these three temperature conditions. In consideration of operational practicality and energy costs, 25 °C was designated as the standard reaction temperature for all subsequent experiments.

3.3. Reusability and Stability of Co-CNTs@NC

Cycling tests were conducted under optimal reaction conditions using the recovered Co-CNTs@NC catalyst to evaluate the cycling stability of Co-CNTs@NC. As shown in Figure 9a, the TCH degradation efficiency decreased from 86.02% to 51.59% after 60 min in the second cycle, indicating that Co-CNTs@NC exhibits unsatisfactory reusability.

To unravel the intrinsic cause of Co-CNTs@NC’s inferior reusability, the Co leaching concentration during the reaction process was first measured by ICP-MS. As documented in Table S3, Co-CNTs@NC exhibited substantial Co leaching (>4.54 mg/L), which induced irreversible depletion of Co active sites and thus a dramatic decline in TCH degradation efficiency. Figure 9b reveals that the characteristic diffraction signal of metallic Co diminished drastically after the reaction cycle, which further confirms the reduction in Co active sites. As a result, high Co leaching is a crucial factor responsible for the poor reusability of Co-CNTs@NC. Beyond metal leaching, the coverage of reaction intermediates can also impair the cyclic performance of catalysts. Thus, the variation in total organic carbon (TOC) during the reaction was monitored to clarify the mineralization degree of TCH and the evolution of intermediates. After 90 min of reaction with PMS addition, the TOC removal efficiency reached 21.85% (Figure S10), confirming partial mineralization of TCH. However, the relatively low mineralization efficiency demonstrates that most reaction intermediates were refractory to further degradation. The accumulation of a large number of intermediates in the reaction system would result in the masking of active sites [43]. A third factor responsible for performance deterioration is material loss incurred during the catalyst recovery process. Specifically, residual carbon particles present in the centrifugal supernatant inevitably result in material loss during the recycling of Co-CNTs@NC. A similar phenomenon has been previously reported in advanced oxidation processes employing cobalt-carbon composite materials [44,45]. Based on the aforementioned experimental results and corresponding analyses, the poor reusability of Co-CNTs@NC can be primarily ascribed to three key issues: excessive Co leaching, masking of active sites by reaction intermediates, and material loss during the recovery process. It has been demonstrated that the incorporation of a second metal (e.g., Fe and Cu) for fabricating bimetallic systems can effectively suppress metal leaching and enhance cyclic stability [37,46]. Accordingly, our future work will adopt established methodologies to tailor the Co-CNTs@NC derived from CoZn-ZIFs, aiming to optimize its long-term cyclic stability and mitigate metal leaching.

Additionally, no characteristic signals associated with cobalt oxides were identified in the Co 2p XPS spectrum, O 1s spectrum, and FT-IR spectrum (Figure S11) of Co-CNTs@NC, as well as XRD patterns of Co-CNTs@NC after the reaction. This observation can be ascribed to the protective CNT coating, which effectively enhances the oxidation resistance of Co-CNTs@NC and thus inhibits the formation of cobalt oxide species.

3.4. Degradation Mechanism

A series of radical quenching experiments were carried out to elucidate the reactive oxygen species (ROS) responsible for TCH degradation. Specifically, the sulfate radical (SO₄^•−), hydroxyl radical (•OH), superoxide radical (O₂^•−), and singlet oxygen (¹O₂) were quenched using methanol (MeOH), tert-butanol (TBA), p-Benzoquinone (PBQ), and L-histidine (L-His) as quenching agents, respectively. Comprehensive details regarding the experimental protocols and quenching mechanisms are documented in Text S4 of the Supplementary Materials.

Figure 10a illustrates that following 90 min of reaction, the TCH degradation efficiency declined from 86.33% to 84.49%, 59.32%, 64.80%, and 48.50% upon the addition of TBA, MeOH, PBQ, and L-His, respectively. These results demonstrate that ¹O₂ serves as the primary reactive species mediating TCH degradation, whereas SO₄^•−, O₂^•−, and •OH also participate in the degradation process within the Co-CNTs@NC/PMS system.

To further elucidate the species and relative contributions of ROS generated during the reaction, electron paramagnetic resonance (EPR) experiments were performed. 5,5-Dimethyl-1-pyrroline N-oxide (DMPO) and 2,2,6,6-tetramethyl-4-piperidone (TEMP) were utilized as spin-trapping agents to identify the corresponding ROS. A distinct characteristic signal of the DMPO oxidation product (DMPOX) can be observed in Figure 10b. In contrast, no detectable signals of SO₄^•− and •OH were observed. This observation can presumably be ascribed to the rapid oxidation of DMPO mediated by ¹O₂, which generates the short-lived DMPOX intermediate in the ¹O₂-dominant system, thereby impeding the efficient trapping and detection of the target free radicals [47]. As depicted in Figure 10c, faint DMPO–O₂^•− signals were detected in the Co-CNTs@NC/PMS reaction system, a phenomenon presumably arising from the rapid conversion of O₂^•− to ¹O₂ [48]. In addition, strong TEMP-¹O₂ signals were obviously observed with different reaction times in the Co-CNTs@NC/PMS system (Figure 10d). The results verify that Co-CNTs@NC is highly efficient at activating PMS to produce ¹O₂. Compared with the EPR signals of DMPOX and DMPO−O₂^•−, the signal intensity of ¹O₂ increased significantly as the reaction proceeded, which further demonstrates that ¹O₂ serves as the predominant ROS driving TCH degradation.

On the basis of the above results, a nonradical-dominated degradation mechanism is proposed in Figure 11. Co nanoparticles anchored within Co-CNTs@NC donated electrons to PMS, triggering cleavage of the O–O bond and concomitant formation of SO₄^•−, •OH, and O₂^•− according to Equations (1)–(7) [26,49]. ¹O₂ was generated via the further transformation of SO₄^•−, •OH, and O₂^•− according to Equations (8) and (9) [50]. Previous studies have demonstrated that electrons can migrate from nucleophilic PMS (electron donors) to electrophilic carbonyl (C=O) functional groups (electron acceptors). When PMS donates one electron, it generates the corresponding PMS anion radical SO₅^•− (Equation (10)). As depicted in Equation (11), SO₅^•− undergoes facile transformation into S₂O₈²⁻ or SO₄²⁻ with the accompanying production of ¹O₂ [51]. Ultimately, TCH was degraded by the in situ generated ROS (Equation (12)).

$$\begin{array}{l}{\mathrm{C}\mathrm{o}}^0+2{\mathrm{H}\mathrm{S}\mathrm{O}}_5^{-} \to 2{\mathrm{O}\mathrm{H}}^{-}+{\mathrm{C}\mathrm{o}}^{2+}+2{\mathrm{S}\mathrm{O}}_4^{•-}\end{array}$$

(1)

$$\begin{array}{l}{\mathrm{C}\mathrm{o}}^{2+}+{\mathrm{H}\mathrm{S}\mathrm{O}}_5^{-} \to {\mathrm{O}\mathrm{H}}^{-}+{\mathrm{C}\mathrm{o}}^{3+}+{\mathrm{S}\mathrm{O}}_4^{•-}\end{array}$$

(2)

$$\begin{array}{l}{\mathrm{C}\mathrm{o}}^{3+}+{\mathrm{H}\mathrm{S}\mathrm{O}}_5^{-} \to {\mathrm{S}\mathrm{O}}_5^{•-}+{\mathrm{C}\mathrm{o}}^{2+}+{\mathrm{H}}^{+}\end{array}$$

(3)

$$\begin{array}{l}{\mathrm{C}\mathrm{o}}^0+2{\mathrm{C}\mathrm{o}}^{3+} \to 3{\mathrm{C}\mathrm{o}}^{2+}\end{array}$$

(4)

$$\begin{array}{l}{\mathrm{H}\mathrm{S}\mathrm{O}}_5^{-}+{\mathrm{H}}_2\mathrm{O} \to {\mathrm{H}}_2{\mathrm{O}}_2+{\mathrm{H}\mathrm{S}\mathrm{O}}_4^{-}\end{array}$$

(5)

$$\begin{array}{l}{\mathrm{C}\mathrm{o}}^{3+}+{\mathrm{H}}_2{\mathrm{O}}_2 \to {\mathrm{C}\mathrm{o}}^{2+}+{\mathrm{H}\mathrm{O}}_2•+{\mathrm{H}}^{+}\end{array}$$

(6)

$$\begin{array}{l}{\mathrm{C}\mathrm{o}}^{3+}+{\mathrm{H}}_2{\mathrm{O}}_2 \to {\mathrm{C}\mathrm{o}}^{2+}+{\mathrm{H}\mathrm{O}}_2•+{\mathrm{H}}^{+}\end{array}$$

(7)

$$\begin{array}{l}{\mathrm{S}\mathrm{O}}_4^{•-}+{\mathrm{H}}_2\mathrm{O} \to •\mathrm{O}\mathrm{H}+{\mathrm{H}}^{+}+{\mathrm{S}\mathrm{O}}_4^{2-}\end{array}$$

(8)

$$\begin{array}{l}{\mathrm{O}}_2^{•-}+•\mathrm{O}\mathrm{H} \to {}^1\mathrm{O}_2+{\mathrm{O}\mathrm{H}}^{-}\end{array}$$

(9)

$$\begin{array}{l}{\mathrm{H}\mathrm{S}\mathrm{O}}_5^{-}\to {\mathrm{S}\mathrm{O}}_5^{•-}+{\mathrm{H}}^{+}+{\mathrm{e}}^{-}\end{array}$$

(10)

$$\begin{array}{l}2{\mathrm{S}\mathrm{O}}_5^{•-}\to {{\mathrm{S}}_2\mathrm{O}}_8^{2-}+{}^1\mathrm{O}_2 \mathrm{or} 2{\mathrm{S}\mathrm{O}}_5^{•-}+{\mathrm{H}}_2\mathrm{O}\to {2\mathrm{S}\mathrm{O}}_4^{2-}+1.5{}^1\mathrm{O}_2+{2\mathrm{H}}^{+}\end{array}$$

(11)

$$\begin{array}{l}{\mathrm{O}}_2^{•-} /•\mathrm{O}\mathrm{H}/{\mathrm{S}\mathrm{O}}_4^{•-}/{}^1\mathrm{O}_2+\mathrm{T}\mathrm{C}\mathrm{H} \to \mathrm{I}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{m}\mathrm{e}\mathrm{d}\mathrm{i}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{s}+{\mathrm{C}\mathrm{O}}_2+{\mathrm{H}}_2\mathrm{O}\end{array}$$

(12)

4. Conclusions

In summary, CoZn-ZIFs were successfully synthesized via a simple room-temperature self-assembly method using water as the solvent. The TEM, HRTEM, and HAADF-STEM elemental mapping, XRD, and XPS characterization results demonstrate that Co-CNTs@NC with well-dispersed Co nanoparticles was successfully fabricated via the high-temperature calcination of CoZn-ZIFs. The in situ generated CNTs effectively enhanced the oxidation resistance of Co active sites. Benefiting from its high Co content (5.38 wt%), Co-CNTs@NC possessed 13.22-fold faster TCH degradation kinetics than NC. Within a reaction period of 90 min, the TCH degradation and mineralization efficiencies in the Co-CNTs@NC/PMS system reached 86.33% and 21.85%, respectively. Radical quenching experiments and EPR characterization demonstrate a ¹O₂-dominated nonradical mechanism for TCH degradation in Co-CNTs@NC/PMS, with SO₄^•−, •OH, and O₂^•− synergistically involved. With outstanding catalytic efficacy and cost-competitiveness, the Co-CNTs@NC composites demonstrate substantial practical potential for remediating antibiotic-contaminated aquatic environments.