English

• The Fenton-like process has drawn considerable attention as a prospective technology for water decontamination [1,2]. Specifically, extensive research has been conducted on the advanced oxidation processes (AOPs) relying on peroxymonosulfate (PMS) due to their various oxidation pathways that can generate highly reactive radicals (e.g., ^•OH (1.9−2.7 V vs. NHE) and SO₄^•− (2.5−3.1 V/NHE)) or highly selective non-radical species including singlet oxygen (¹O₂, 2.2 V vs. NHE), high-valent metal, and metals-PMS complex [3,4]. It is notable that non-radical species (e.g., ¹O₂ and Co(Ⅳ)=O) have gained extensive attention in the selective oxidation of organic contaminants owing to their inherent merits like potent electrophilicity and selectivity, a long half-life, and strong resilience to environmental interference [5,6]. Nevertheless, the ordinary production procedures of the non-radical species through PMS activation by metallic nanoparticles or complexes presented unsatisfactory results with low selectivity [7].

  Defect engineering has emerged as a highly promising strategy for optimizing heterogeneous catalytic PMS activation, demonstrating significant potential to selectively modulate the PMS activation pathway toward non-radical species generation [8,9]. Vacancy defects can adjust the electronic structure of metal oxides, leading to the formation of unsaturated coordination sites [10]. Additionally, these defects optimize the adsorption energy of reactants on the catalyst surface. As a result, the reaction energy barrier between the catalyst and PMS is reduced, enhancing catalytic activation and accelerating the degradation of pollutants [11]. Recent researches have extensively explored various types of vacancy defects, such as oxygen vacancy (O_V), nitrogen vacancy (N_V) and sulfur vacancy (S_V), due to their substantial potential in water treatment applications [12–14]. Among them, O_V is among the most extensively investigated in PMS-activated systems [15]. In metal oxides, oxygen atoms within the crystal structure can be released under external conditions, such as elevated temperatures, leading to the formation of O_V. For example, Wu et al. synthesized a series of Fe-Co LDH catalysts with varying levels of O_V by modulating the calcination temperature and demonstrated the significant role of O_V in PMS activation [16]. Besides, it has been reported that O_V had the potential to improve site exposure and facilitate mass diffusion in cobalt-related catalysts, as well as enhance the adsorption of PMS on catalyst surfaces [17]. The localization of sufficient electrons around the O_V supports a single-electron-transfer non-radical process, which can induce the formation of ¹O₂ via direct electron transfer [18,19]. Whereas, some recent studies have revealed that ¹O₂, derived from the transformation of superoxide radicals (O₂^•−), may serve as the predominant reactive species responsible for pollutant degradation in O_V-enriched metal oxide/PMS systems [20]. Meanwhile, the role of O_V in O_V-rich catalysts for PMS activation remains controversial, as O_V typically coexist with reducing metal active sites. These metal sites can also directly activate PMS through electron transfer processes, complicating the attribution of catalytic activity [21]. Therefore, further studies are required to reveal the role of O_V on PMS activation and elucidate the underlying mechanism.

  Among transition metals, cobalt (Co) is recognized as the most efficient activator for PMS activation [17]. Recent studies have extensively documented Co²⁺-mediated PMS systems, with particular attention given to the non-radical species generation pathway [22]. However, some critical challenges remain, such as secondary pollution from Co²⁺ leaching and inefficient catalyst recovery [23], both of which hinder its practical application. While stable cobalt oxides can mitigate metal leaching and address recovery issues, they often suffer from rapid deactivation due to nanoparticle aggregation and sluggish Co(Ⅲ)/Co(Ⅱ) redox cycling [24]. Encapsulating cobalt oxides (e.g., Co₃O₄, CoO_x) onto the metal support presents a viable strategy to mitigate Co²⁺ leaching, suppress nanoparticle aggregation and enhance mass transfer kinetics. Natural montmorillonite (Mt) has emerged as a superior alternative to traditional template materials (e.g., carbonaceous materials, zeolites, and mesoporous matrices) for immobilizing metal oxide catalysts, owing to its unique combination of abundant availability, cost-effectiveness, layered architecture, and cation exchange capacity [25]. The two-dimensional lamellar structure of Mt, composed of outer tetrahedral silicon layers and an inner octahedral sheet, not only facilitates the efficient flow of aqueous matrices but also enhances the oxidative degradation of pollutants [26]. Crucially, native cations (e.g., K⁺, Na⁺, and Ca²⁺) adsorbed on basal surfaces or within interlayer spaces can be easily exchanged, enabling the strategic incorporation of active Co(Ⅱ) to tailor catalytic properties [27]. Furthermore, the interlayer domain of Mt functions as a nanoconfined "microchemical reactor", significantly improving mass transfer kinetics and prolonging the residence time for contaminant degradation [28]. Additionally, the dynamic expansion and contraction of these nanoconfined spaces provide a high surface area for the tunable dispersion of active species and create abundant adsorption sites for organic molecules, further optimizing catalytic performance [29].

  In this study, a series of nano-Co₃O₄-encapsulated montmorillonite catalysts with varying levels of O_V was synthesized for PMS activation. Ofloxacin (OFL), a frequently identified fluoroquinolone antibiotic in surface water and wastewater, was designated as the target pollutant to evaluate their catalytic efficiencies. The relationship between O_V levels and OFL degradation was established. Results showed that O_V in nano-Co₃O₄-encapsulated montmorillonite catalysts can selectively induce the activation of PMS to produce ¹O₂ and Co(Ⅳ)=O. The underlying mechanism was further revealed through comprehensive experiment and theoretical calculation. The findings offer significant insights into the O_V-mediated heterogeneous Fenton-like reaction for the selective oxidation of emerging organic contaminants in water and wastewater.

  The typical synthesis steps of Co₃O₄−Mt-O_V are shown in Fig. 1a. Typically, Co₃O₄−Mt was firstly synthesized via a combination of solvent-thermal and high-temperature calcination techniques (Text S1 in Supporting information). After that, Co₃O₄−Mt-xO_V with varied contents of O_V was obtained chemical etching reduction and subsequent oxidation (Text S1), where x was the reduction/oxidation cycle times. As shown in Fig. 1b, the X-ray diffraction (XRD) pattern of Na-Mt exhibited characteristic diffraction peaks of montmorillonite (JCPDS No. 29–1498) at 2θ = 7.08°, 19.76°, 29.40°, 36.04°, and 61.82°, corresponding to the (001), (100), and (440) crystal planes, respectively. For pristine Co₃O₄, the observed peaks at 19.0°, 31.3°, 36.9°, 38.5°, 44.9°, 55.7°, 59.4°, and 65.2° were assigned to the (111), (220), (311), (222), (400), (422), (511), and (440) planes of spinel Co₃O₄ (JCPDS No. 42–1467) [30]. Notably, the absence of impurity phases confirmed the high crystallinity and phase purity of the synthesized Co₃O₄. The Co₃O₄−Mt composite demonstrated a biphasic structure containing both montmorillonite and Co₃O₄ components, confirming successful immobilization of Co₃O₄ nanoparticles on the montmorillonite carrier. Whereas, the characteristic diffraction peak intensity of Co₃O₄ and Mt in Co₃O₄−MT was reduced in comparation with that of pure Co₃O₄, which meant that Co₃O₄ particles were well dispersed and crystallinity was reduced in the presence of montmorillonite carrier. The XRD patterns of Co₃O₄−Mt-4O_V and Co₃O₄−Mt were basically the same, which indicated that the crystal structure of the catalyst was not affected after chemical etching of O_V. However, Nevertheless, the slightly diminished peak intensity observed in Co₃O₄−Mt-4O_V may be attributed to partial lattice distortion induced by increased O_V concentration [31].

  Figure 1

  

  Figure 1.  (a) Schematic illustration of the synthesis of the Co₃O₄−MT-O_V. (b) XRD, (c) BET, (d) EPR, (e) O_V contents, (f) Co 2p HR-XPS and (g) O 1s HR-XPS of the catalysts.

  DownLoad: Full-Size Img PowerPoint

  The isotherm adsorption curves of all samples showed the characteristic of typical type Ⅳ curve (Fig. 1c), implying the mesoporous properties of the catalysts. The pore size distribution further verified the abundance of micropores and mesopores in these catalysts (Fig. S1 in Supporting information). The surface area of Co₃O₄−Mt-4O_V was detected to be 61.646 m²/g, which was about twice that of Co₃O₄−Mt (33.264 m²/g). Hence, the generation of O_V can increase the specific surface area of Co₃O₄−Mt, favoring the utilization of active sites in the catalytic reactions. The existence and content of O_V in different samples were further measured by electron paramagnetic resonance (EPR). From Fig. 1d, the asymmetric EPR signals at g = 2.004 were clearly observed in Co₃O₄−Mt-4O_V catalysts, which was attributed to the unpaired electrons on the capture surface O_V [32,33]. The intensity of O_V signals significantly enhanced by increasing the number of REDOX cycles in the preparation process. To be specific, the O_V content of Co₃O₄−Mt-6O_V reached 3.59 times than that of Co₃O₄−Mt (Fig. 1e). The full-scan spectrum of X-ray photoelectron spectroscopy (XPS) showed that Co, Al, Si and O were the dominant components in Co₃O₄−Mt-4O_V (Fig. S2 in Supporting information). The C 1s core level peak of 284.8 eV is used as energy reference for charge correction. The high-resolution XPS spectrum were further obtained to determine the valence distributions of Co in the different samples. Two components of Co 2p_3/2 and Co 2p_1/2 spin–orbit doublets were observed in the high-resolution Co 2p XPS spectra (Fig. 1f). The Co 2p_3/2 XPS spectra suggested the presence of Co(Ⅲ), Co(Ⅱ) and satellite species in Co₃O₄−Mt-4O_V with the peaks centered at around 794.8, 781.4 and 788.5 eV, respectively [34]. Based on the area intensity, the Co₃O₄−Mt-4O_V possessed the highest ratios of Co(Ⅱ)/Co(Ⅲ) compared to Co₃O₄ and Co₃O₄−Mt, which was hope to significantly facilitate the activation process. Such changing trend was also observed in the previous study, which might be attributed to the stepwise formation of O_V [35]. The O 1s XPS spectra of Co₃O₄ can be deconvolved into three main peaks corresponding to lattice oxygen (O_L), oxygen vacancy (O_V) and hydroxyl oxygen (O_OH) with peak centers of 529.7, 531.1 and 532.3 eV, respectively (Fig. 1g) [36]. The O 1s peaks of Co₃O₄−MT and Co₃O₄−MT-4O_V were slightly shifted towards higher binding energy, which might be due to the formation of chemical bonds between montmorillonite and Co₃O₄ nanoparticles [37]. The relative abundance of OV in Co₃O₄−MT-4O_V (41.4%) was higher than that in Co₃O₄−MT (33.7%), which aligned with the EPR results. As Fig. S2 shows, the positive binding energy shifts of Al 2p and Si 2p spectrum of Co₃O₄−Mt and Co₃O₄−MT-4O_V were observed, further suggesting that a Co-O-Al(Si) bond might be formed between Co₃O₄ and montmorillonite [38].

  From the scanning electron microscope (SEM) images, a massive scale structure and smooth surface was observed in the montmorillonite (Fig. 2a). The prepared Co₃O₄ nanoparticles are in a spherical shape of about 20 nm, but a large number of nanoparticles accumulate into large clusters and are easily agglomerated (Fig. 2b). Compared with Na-Mt, the surface of the Co₃O₄−Mt-4O_V sample became very rough and loose (Fig. 2c). In addition, the EDS element mapping showed that Co, Mg, O, Si, and Al elements were evenly dispersed on the surface of Co₃O₄−Mt-4O_V (Fig. 2d). The transmission electron microscope (TEM) images of Co₃O₄−Mt-4O_V further proved that Co₃O₄ nanoparticles with a diameter of 5–10 nm were uniformly fixed on the interlayer and surface of montmorillonite (Figs. 2e-g), which might be due to the confinement effect preventing the agglomeration of Co₃O₄ nanoparticles effectively. This is because nano-confinement effect and interaction in confined space can significantly affect the physical and chemical properties of Co₃O₄, leading to a smaller size and higher dispersion [39]. Characteristic lattice fringes of 0.23, 0.24, 0.28, and 0.46 nm were observed via HRTEM, corresponding to the (222), (311), (220), and (111) crystal planes of Co₃O₄ nanoparticles (Figs. 2h and i), respectively. These observations agreed with the XRD results. Additionally, evident grain boundaries between Co₃O₄ nanoparticles were observed, suggesting the presence of abundant defects due to the formation of numerous grain boundaries. This is consistent with the observation of significant lattice distortion in the catalyst, as indicated by the red circle in Fig. 2i, which implied the existence of O_V [40].

  Figure 2

  

  Figure 2.  SEM images of (a) Na-Mt, (b) Co₃O₄, (c) Co₃O₄−Mt-4O_V, and (d) element mapping of Co₃O₄−Mt-4O_V. TEM images of (e) Na-Mt, (f) Co₃O₄, (g) Co₃O₄−Mt-4O_V, and HRTEM images of (h) Co₃O₄, (i) Co₃O₄−Mt-4O_V.

  DownLoad: Full-Size Img PowerPoint

  The adsorption properties of the modified montmorillonite catalysts for OFL were evaluated. Merely 5.1% OFL can be adsorbed by Na-Mt after 60 min (Fig. S3a in Supporting information), since OFL as a macromolecular organic matter is hard to enter the montmorillonite layer. The adsorption of OFL with Co₃O₄ was 25.1%, while it was markedly increased to 69.1% with Co₃O₄−MT (Fig. S3a). With the increase of O_V content, the adsorption efficiency of Co₃O₄−Mt-2O_V, Co₃O₄−Mt-4O_V and Co₃O₄−Mt-6O_V for OFL gradually increased to 73.0%, 76.0% and 80.3% after 60 min, respectively, indicating that O_V might play an important role in the adsorption process. The adsorption of OFL by the catalysts conformed to the quasi-secondary kinetic curve (Fig. S3b in Supporting information), suggesting that the process was primarily governed by chemisorption [41]. Previous studies have demonstrated that the enhancement of the catalyst's adsorption capacity can facilitate the catalytic degradation of pollutants [42]. Given the significant adsorption capacity of these catalysts for pollutants, in this study, the catalysts were pre-mixed with the pollutants for 30 min prior to the catalytic experiments. This pre-treatment ensured that the adsorption capacity of the catalysts was nearly saturated, thereby minimizing the influence of adsorption on the outcomes of the catalytic degradation experiments. PMS alone can only degrade 5.5% of OFL within 60 min, and the removal efficiency slightly increased to 13.1% with the activation of PMS by Na-Mt (Fig. 3a). Under the same conditions, 83.9% of OFL was degraded in the Co₃O₄/PMS system with the reaction rates (k_obs) of 0.041 min^-1 (Fig. 3b). The k_obs value of OFL degradation significantly enlarged to 0.161 min^-1 by the Co₃O₄−Mt/PMS system, which might be attributed to the optimal dispersion of Co₃O₄ on montmorillonite surface abundant in hydroxyl groups. Besides, the interlayer structure of montmorillonite offers a localized environment for the accumulation of PMS and pollutants, facilitating the degradation of OFL by the nanoconfinement effect [43]. Notably, after chemical etching of OV, the catalytic performance of Co₃O₄−Mt-2O_V, Co₃O₄−Mt-4O_V and Co₃O₄−Mt-6O_V catalysts was further improved, and k_obs of OFL degradation were significantly increased to 0.441, 0.487 and 0.553 min^-1, respectively (Fig. 3b), suggesting that the augmentation of OV content could markedly enhance the catalytic reaction. As shown in Fig. 3c, the k_obs of OFL degradation was linearly correlated (R² = 0.98) with the O_V content in the Co₃O₄−Mt-xO_V, which evidenced the crucial role of OV for PMS activation. Due to their comparable catalytic performance between Co₃O₄−Mt-4O_V and Co₃O₄−Mt-6O_V, Co₃O₄−Mt-4O_V requiring a simpler preparation process was selected for further study. The catalytic activity of Co₃O₄−Mt-4O_V was compared with the reported relating catalysts of PMS. The modified k_obs (m-k_obs) was calculated by multiplying the observed rate constant by the pollutant concentration and then dividing by the catalyst dosage. As shown in Table S2 and Fig. S4 (Supporting information), the Co₃O₄−Mt-4O_V is one of the most efficient PMS heterogeneous activators, with an m-k_obs of 97.88 × 10^–3 min^-1, which is 1.1–36 times greater than those of reported Co-based and Ov-mediated heterogeneous catalysts.

  Figure 3

  

  Figure 3.  (a) Kinetics of OFL degradation in different systems, and (b) the corresponding degradation rate (k_obs). (c) The linear relationship between O_V content of Co₃O₄−Mt-xO_V and k_obs of OFL. (d) Effects of solution pH on the degradation of OFL in the Co₃O₄−Mt-4O_V/PMS system, and (e) the corresponding k_obs values. (f) Zeta potential of Co₃O₄−Mt-4O_V changed with varying solution pH. Experiment conditions: [OFL]₀ = 20 mg/L and T = 20 ℃, if not specified catalyst dosage = 0.2 g/L, [PMS]₀ = 1.0 mmol/L and pH₀ = 7.0.

  DownLoad: Full-Size Img PowerPoint

  As shown in Fig. 3d, the removal of OFL was not obviously changed within the pH range of 3.0–9.0, maintaining over 95% of OFL degradation. In a strong alkaline environment (pH 11.0), the degradation efficiency of OFL was reduced to 89.2% with the degradation rate decreased to 0.201 min^-1 (Fig. 3e). The pK_a1 and pK_a2 of PMS are 0 and 9.442, respectively [44]. Thus, PMS was dominantly existed in the form of SO₅²⁻ that was a less oxidized dianion and generated through deprotonation. Besides, the zeta potential test revealed that the surface of Co₃O₄−Mt-4O_V was negatively charged under strongly alkaline solution (Fig. 3f). Thus, the electrostatic repulsion between SO₅²⁻ and Co₃O₄−Mt-4O_V surface could prevent their catalytic reaction. The negative charge of OFL in alkaline conditions was not conducive to enter the montmorillonite layer. Additionally, PMS became more unstable under alkaline conditions, leading to increased unnecessary consumption (Fig. S5 in Supporting information). The above reasons together resulted to the inhibition of OFL degradation at pH 11.0.

  To identify the role of reactive oxygen species (ROS) in the Co₃O₄−Mt-4O_V/PMS system, quenching experiments were performed using different scavengers. Among them, MeOH can effectively quench ^•OH (k_^•OH = (1.2–2.8) × 10⁹ L mol^-1 s^-1) and SO₄^•− (k_SO4^•− = (1.6–7.7) × 10⁷ L mol^-1 s^-1), while tert–butyl alcohol (TBA) can be only used to quench ^•OH (k_^•OH = 10⁹ L mol^-1 s^-1) [45]. As Fig. 4a shows, the degradation of OFL was hardly inhibited with MeOH and TBA, indicating that the contributions of ^•OH and SO₄^•− were insignificant. EPR test was performed using DMPO as electron spin trapping agents to verify the generation of ^•OH and SO₄^•− in the Co₃O₄−Mt-4O_V/PMS system. From Fig. 4d, the peaks for DMPO-^•OH and DMPO-SO₄^•− adducts were not detected, demonstrating the absence of ^•OH and SO₄^•− in the activation of PMS with Co₃O₄−Mt-4O_V. Further, p-BQ and FFA were employed as typical scavengers for O₂^•− (k_O2^•− = 9.0 × 10⁸–1.0 × 10⁹ L mol^-1 s^-1) and ¹O₂ (k_¹O2 = 1.2 × 10⁸ L mol^-1 s^-1), respectively [46]. The degradation of OFL was obviously inhibited by p-BQ (Figs. 4a and b), while PMS consumption was barely affected (Fig. S6 in Supporting information). This suggested that the inhibition of OFL degradation by p-BQ was indeed caused by the scavenging for O₂^•−, and hence, O₂^•− might contribute to the removal of OFL. The DMPO-trapped EPR spectra of O₂^•− directly indicated the formation of O₂^•− (Fig. 4e). However, unlike SO₄^•−, ^•OH, and ¹O₂, O₂^•− typically acts as a reductant rather than an oxidant in aqueous systems due to its high solvation energy in water [47]. Previous studies have also demonstrated that O₂^•− exhibited limited reactivity in direct oxidation of organic contaminants except for the quinones due to its low redox potential (−0.33 V vs. NHE) [48], but can function as a precursor for ¹O₂ formation [49]. Hence, it is challenging for O₂^•− to directly contribute to the oxidation of OFL, but might emerge indirectly as an intermediate product of ¹O₂. Based on the above analysis, the degradation of OFL in the Co₃O₄−Mt-4O_V/PMS system is more likely to occur predominantly via non-radical pathways. As expected, the addition of FFA significantly suppressed OFL degradation (Figs. 4a and b) while exhibiting negligible effects on PMS consumption (Fig. S6), suggesting that ¹O₂ likely served as the predominant reactive species in this system. Moreover, replacing the solvent from H₂O to D₂O, which can increase the lifetime of ¹O₂ by approximately tenfold [50], led to enhanced degradation efficiency and reaction rate constants for OFL (Fig. 4c). This highlighted the essential role of ¹O₂ in facilitating the degradation of OFL. Notably, a typical 1:1:1 triple-state signal of TEM-¹O₂ was detected in PMS solution, and the peak intensity significantly increased once Co₃O₄−Mt-4O_V was introduced (Fig. 4f), strongly evidencing the generation of ¹O₂ during the catalytic reaction.

  Figure 4

  

  Figure 4.  (a) OFL degradation in the Co₃O₄−Mt-4O_V/PMS process with and without the existence of different quenchers, and (b) the corresponding k_obs values. (c) Degradation of OFL in the D₂O solution (insert: the corresponding k_obs values). The EPR spectra of (d) DMPO-SO₄^•− and DMPO-^•OH, (e) DMPO—O₂^•− and (f) TEMP-¹O₂. Experiment conditions: catalyst dosage = 0.2 g/L, [PMS]₀ = 1.0 mmol/L, [OFL]₀ = 20 mg/L, pH₀ = 7.0 and T = 20 ℃, if applied [MeOH]₀ = [TBA]₀ = 100 mmol/L, [FFA]₀ = [p-BQ]₀ = 10 mmol/L.

  DownLoad: Full-Size Img PowerPoint

  Experimental evidence from N₂ purging confirmed that the primary source of ¹O₂ in the Co₃O₄−Mt-4O_V/PMS system stemmed from catalyst-mediated PMS activation rather than dissolved oxygen, as OFL degradation remained unaffected under anaerobic conditions (Fig. S7 in Supporting information). Subsequent investigation focused on elucidating the formation mechanism and evolutionary pathway of ¹O₂ within this catalytic system. Although the above results have identified O₂^•− as potential precursors for ¹O₂ formation, their direct conversion faces thermodynamic limitations [51]. Meanwhile, the classical Haber-Weiss reaction (O₂^•− → ^•OH→ ¹O₂) can also be ruled out, since ^•OH was not detected in the system (Fig. 4d). Recent studies have found that some metal active sites serve as critical mediators in the transformation process from O₂^•− to ¹O₂. For example, Yi et al. demonstrated the conversion of O₂^•− to ¹O₂ in a molybdenum-assisted Fenton system, where Mo⁶⁺ is reduced by O₂^•−, yielding ¹O₂ and Mo⁴⁺ (Mo⁶⁺ + O₂^•− → ¹O₂ + Mo⁴⁺) [51]. Similarly, Liu et al. revealed that high-valent cobalt species (e.g., Co(Ⅳ)/Co(Ⅲ)) can oxidize O₂^•−, producing ¹O₂ alongside Co(Ⅲ)/Co(Ⅱ) (≡Co(Ⅳ)/Co(Ⅲ) + O₂^•− → ¹O₂ + Co(Ⅲ)/Co(Ⅱ)) [52]. Building on these findings, the formation of high-valent cobalt species and their potential catalytic role in the O₂^•−/¹O₂ conversion was further explored.

  Notably, the addition of DMSO, a high-valent cobalt-oxo species (Co(Ⅳ)=O) scavenger, inhibited approximately 70% of OFL degradation (Fig. S8 in Supporting information), suggesting that Co(Ⅳ)=O played a significant role. The high conversion of PMSO to PMSO₂ (above 87%, Fig. S9 in Supporting information) further evidenced the generation of Co(Ⅳ)=O, since PMSO₂ is a known product of oxygen transfer between Co(Ⅳ)=O and PMSO. To further verify the role of Co(Ⅳ)=O, we examined the degradation of a ¹O₂ probe (1,3-diphenylisobenzofuran, DPBF) with and without the presence of DMSO. Notably, DPBF degradation decreased to 35% with DMSO, compared to 97% in the control (Fig. S10 in Supporting information), confirming that Co(Ⅳ) contributed to ¹O₂ generation. We subsequently examined the TEMP-¹O₂ signal in the presence of DMSO. Significantly, introducing DMSO resulted in a significant suppression of the TEMP-¹O₂ signal (Fig. S11 in Supporting information). Hence, considering that the interlayer confinement effect can drive quantum state transitions through valence electron redistribution, the generated Co(Ⅳ)=O species facilitates O₂^•− in surpassing the reaction energy barrier and subsequently converting to ¹O₂ [53,54].

  The content of O_V before and after catalyst use was determined by EPR method. The peak intensity of O_V was observed to decrease following the reaction (Fig. 5a), potentially due to the occupation of O_V sites by adsorbed oxygen, leading to the formation of relatively weakly bonded surface-adsorbed oxygen and lattice oxygen [55]. During the reaction process, pollutants or degradation intermediates can also cover the O_V on the catalyst surface, indicating that O_V was consumed in the activation of PMS. Similar results were obtained by O 1s high-resolution XPS spectra before and after the use of Co₃O₄−Mt-4O_V catalyst (Fig. S12c in Supporting information). This further implied that the O_V on the surface of Co₃O₄−Mt-4O_V catalyst played an important role in PMS activation.

  Figure 5

  

  Figure 5.  (a) EPR spectra of Co₃O₄−Mt-4O_V catalyst before and after use. (b) Electrochemical impedance spectroscopy and (c) cyclic voltammetry spectra of Co₃O₄, Co₃O₄−Mt and Co₃O₄−Mt-4O_V. (d) The calculated charge density difference of PMS adsorbed on (a) Co₃O₄−Mt and (b) Co₃O₄−Mt-4O_V.

  DownLoad: Full-Size Img PowerPoint

  Electrochemical impedance spectroscopy (EIS) analysis was conducted to evaluate the charge transfer capabilities of the catalysts. As shown in Fig. 5b, Co₃O₄ displayed the smallest semicircle diameter, suggesting the lowest charge transfer resistance among the tested samples. The introduction of montmorillonite (Mt) increased the charge transfer resistance in Co₃O₄−Mt. However, Co₃O₄−MT-4O_V, with its abundant oxygen vacancies, exhibited a significantly reduced charge transfer resistance, indicating a more efficient electron transfer between the catalyst and PMS. Further supporting this observation, cyclic voltammetry (CV) tests revealed that Co₃O₄−MT-4O_V demonstrated higher current density and superior reduction capability compared to Co₃O₄−Mt (Fig. 5c), highlighting the beneficial role of O_V in enhancing electron transfer performance.

  To further unveil the significance of O_V in Co₃O₄−Mt-4O_V, density functional theory (DFT) calculation was conducted to compare the adsorption energy and the charge transfer of Co₃O₄−Mt and Co₃O₄−Mt-4O_V in PMS activation. Results showed that Co₃O₄−Mt-4O_V had a higher adsorption energy (−3.47 kcal/mol) than that of Co₃O₄−Mt (−3.36 kcal/mol) (Fig. 5d), indicating that the introduce of O_V can enhance the adsorption of PMS, favoring the following catalytic reaction. The Bader analysis results indicated that the charge transfer to PMS was 1.092 e^- for the Co₃O₄−Mt configuration and 1.215 e^- for the Co₃O₄−Mt-4O_V configuration (Fig. 5d). This increased charge transfer can substantially enhance the cleavage of O—H bonds, thus promoting the formation of SO₅^•− and O₂^•− at the Co(Ⅱ) site [57]. Collectively, these findings indicated that the Ov in Co₃O₄−Mt-4O_V played a multifaceted and critical role in activating PMS. OVs acted as electron-rich sites that enhance PMS adsorption via electrostatic interactions or charge transfer, particularly the terminal oxygen (O_t) in PMS. Additionally, OVs synergized with neighboring metal sites (e.g., Co³⁺/Co²⁺) to form dual active centers, promoting the electron transfer and stabilizing intermediates.

  To characterize the evolution of cobalt species in Co₃O₄−Mt-4O_V during catalytic reactions, XPS analysis was performed to track the chemical state variations before and after the reaction. The high-resolution XPS spectra of Co 2p and O 1s are shown in Fig S12 (Supporting information), with characteristic peaks provided in Table S1 (Supporting information). The XPS analysis of Co 2p regions revealed significant cobalt speciation evolution during the catalytic process. In the Co 2p_1/2 spectra, the Co(Ⅲ) component at 798.16 eV exhibited an obvious increase from 32.0% to 38.9% after reaction, while the Co(Ⅱ) content correspondingly decreased from 68.0% to 61.0% (Fig. S12b in Supporting information). Parallel trends were observed in the Co 2p_3/2 region, where the Co(Ⅲ)/Co(Ⅱ) ratio displayed analogous progression. These redox dynamics strongly suggested that Co(Ⅱ) served as the catalytic initiator, driving the interconversion cycle among Co(Ⅳ)/Co(Ⅲ)/Co(Ⅱ) species. Notably, the absence of detectable Co(Ⅳ) species is attributed to the metastable nature of excited Co(Ⅳ), which undergoes ultrafast non-radiative decay to the Co(Ⅲ) ground state [56]. Compared with the fresh catalyst, the content of O_V in Co₃O₄−Mt-4O_V after use slightly decreased (Fig. S12c in Supporting information). It is noteworthy that the relative intensities of lattice oxygen and adsorbed oxygen increased from 12.5% and 46.1% to 15.6% and 50.1%, respectively. This phenomenon can be attributed to the consumption of oxygen vacancies, which facilitated the formation of relatively weakly bound surface-adsorbed oxygen (e.g., ≡Co(Ⅱ)−OH) and lattice oxygen [58]. Interestingly, previous studies reported that ≡Co(Ⅱ)−OH complexes could induce the formation of ≡Co(Ⅲ)-O-O-SO₃⁻ intermediates upon reaction with HSO₅⁻ of PMS, which subsequently evolved into Co(Ⅳ)=O species [13,59]. Hence, it can be inferred that ≡Co(Ⅱ) mediated the ultimate formation of Co(Ⅳ)=O through a two-electron transfer pathway. Accompanied by the continuous Co(Ⅱ)/Co(Ⅲ)/Co(Ⅳ) cycling, ¹O₂ was generated through the co-transformation of Co(Ⅳ)=O and O₂^•−. Beyond that, ≡Co(Ⅲ)-mediated electron donation to HSO₅⁻ facilitated its deprotonation to generate SO₅^•− (HSO₅⁻ → SO₅^•− + H⁺ + e^-), which subsequently underwent disproportionation reaction for the generation of ¹O₂ (SO₅^•−+ SO₅^•−→ 2SO₄⁻ + ¹O₂) [60].

  Based on the above results, the possible catalytic mechanism was proposed (Fig. 6). Initially, the curled and stacked structure of Co₃O₄−Mt-4O_V undergoes layer expansion in OFL solution, forming semi-confined microcavities that function as nanoreactors. Under dynamic aqueous conditions, both OFL molecules and PMS anions become concentrated within these expanded interlayer spaces through adsorption and confinement effects. Within the constrained environment, surface-bound ≡Co(Ⅱ) species rapidly couple with PMS through coordination, where HSO₅⁻ activation via ≡Co(Ⅱ) triggers the generation of SO₄^•−, alongside oxidized ≡Co(Ⅲ) species. The cleavage of PMS O—O bonds initiates chain reactions that yield O₂^•− through dual electron transfer pathways. The transient Co(Ⅲ)-OOSO₃ intermediate undergoes radical-mediated O—O bond reorganization, evolving into Co(Ⅳ)=O that ultimately cycle back to ≡Co(Ⅵ)/Co(Ⅲ) redox pairs. The generated Co(Ⅳ)=O can contribute to the degradation of OFL through both direct and indirect oxidation process. In indirect pathway, Co(Ⅳ)=O subsequently mediate the conversion of accumulated O₂^•− into ¹O₂ through electron transfer processes. As a powerful oxidizing species, Co(Ⅳ)=O can also directly contribute to the oxidation of OFL. Hence, the generated ¹O₂ and Co(Ⅳ)=O predominantly driven the degradation of OFL.

  Figure 6

  

  Figure 6.  The mechanism of PMS activated with Co₃O₄−Mt-4O_V for OFL degradation.

  DownLoad: Full-Size Img PowerPoint

  Various inorganic anions (e.g., Cl⁻, H₂PO₄⁻, HCO₃⁻/CO₃²⁻ and SO₄²⁻) existed in actual wastewater, which might affect the degradation of organic pollutants by AOPs to varying degrees [61,62]. It can be seen from Fig. S13 (Supporting information) and Fig. 7a that the presence of Cl⁻, H₂PO₄⁻ and SO₄²⁻ at low (1 mmol/L) and high concentration (50 mmol/L) showed negligible influence on the system, and the OFL removal efficiency still reached > 96% within 60 min. Upon the addition of 50 mmol/L HCO₃⁻ to the system, the degradation of OFL was slightly inhibited, resulting in a reduction of the degradation efficiency to 90.0%. HCO₃⁻/CO₃²⁻ can bind to the Co(Ⅱ) site on the catalyst surface to inactivate it, thus preventing the adsorption and activation for PMS. Moreover, as a representative organic matter, humic acid (HA) showed little effect on OFL degradation (Fig. S14 in Supporting information). In general, the Co₃O₄−MT-4O_V/PMS system showed a strong anti-interference ability to complex water matrices, which further evidenced the significant contribution of non-radical oxidation for OFL.

  Figure 7

  

  Figure 7.  (a) Effects of common anions in water on OFL degradation in the Co₃O₄−Mt-4O_V/PMS process. (b) Performance of OFL degradation during the cycling tests of Co₃O₄−Mt-4O_V, and (c) the leaching of cobalt ions during the cycling tests. (d) Degradation of OFL in the Co₃O₄−Mt-4Ov/PMS process in real waters. Experiment conditions: Catalyst dosage = 0.2 g/L, [PMS]₀ = 1.0 mmol/L, [OFL]₀ = 20 mg/L, pH₀ = 7.0 and T = 20 ℃, if applied [Cl⁻]₀ = [H₂PO₄⁻]₀ = [HCO₃⁻]₀ = [SO₄²⁻]₀ = 50 mmol/L.

  DownLoad: Full-Size Img PowerPoint

  The reusability and stability of catalysts are crucial for the long-term economic viability of wastewater treatment processes. Consequently, to assess both the reusability of Co₃O₄−MT-4O_V composites in degrading OFL within PMS systems and the stability of the catalysts, eight consecutive cycles of stability experiments were conducted under identical operating conditions. As illustrated in Fig. 7b, the Co₃O₄−MT-4O_V catalyst exhibited consistent high catalytic performance over eight cycles, with the removal efficiency of OFL consistently exceeding 94.6%. The degradation rate demonstrated a marginal decrease as the number of cycles increased, which might be caused by the loss of metal active sites. The intermediate products generated during the reaction can induce surface and interlayer passivation of montmorillonite. This passivation would hinder the entry of PMS and OFL into the interlayer of montmorillonite, thereby diminishing the catalytic efficiency. The Co ion leaching after the reaction was further detected. Results showed that the Co ion concentration after 60-min reaction in the first cycle was only 0.05 mg/L, and gradually decreased as the number of cycles increased (Fig. 7c), indicating the high chemical stability of Co₃O₄−MT-4O_V during PMS activation. Such a trace amount of Co²⁺ in the solution led to a limited degradation of OFL with PMS (Fig. S15 in Supporting information). Hence, the contribution of the leached Co²⁺ in the Co₃O₄−MT-4O_V/PMS system was insignificant, considering that the percolation is a gradual, cumulative process. The superior performance and stability of the Co₃O₄−MT-4O_V catalyst can be attributed to the nano-confinement effect of montmorillonite, which enhanced the adsorption of PMS and OFL and subsequently improved the utilization of catalytic sites within montmorillonite. Moreover, earlier research has indicated that carbon-based catalysts are typically prone to carbon corrosion issues [63]. In comparison, metal oxides and O_V active sites exhibit little sensitivity to such challenges. The application of nano-confinement techniques can improve the environmental resilience of catalysts, leading to enhanced stability and reusability. As a result, the Co₃O₄−MT-4O_V catalyst showcased superior performance in terms of stability and reusability. Moreover, the Co₃O₄−MT-4O_V/PMS system was almost not affected when operated in the actual waters (tap water, surface water, and municipal sewage) (Fig. 7d). The performance of degrading various organic contaminants with the Co₃O₄−Mt-4OV/PMS process was further evaluated. Results showed that all these six contaminants (including sulfamethoxazole (SMX), sulfadiazine (SDZ), carbamazepine (CBZ), ciprofloxacin (CIP), enrofloxacin (ENF) and tetracycline (TTC)) can be completely degraded within 10 min (Fig. S16 in Supporting information), suggesting that Co₃O₄−Mt-4OV/PMS process was effective for various organic contaminants degradation. These findings suggested that the Co₃O₄−MT-4O_V/PMS system held significant potential for practical applications.

  To better elucidate the degradation process of OFL, the main intermediate products of OFL degradation were detected using HPLC-MS-MS. Nine transformation products (TPs), namely TP-112, TP-150, TP-158, TP-200, TP-205, TP-279, TP-318, TP-338 and TP-378 were identified, and the key information of the intermediates was shown in Fig. S17 and Table S3 (Supporting information). Based on the results, three degradation pathways of OFL were proposed, as shown in Fig. S18 (Supporting information). In Pathway Ⅰ, the piperazine ring of OFL underwent hydroxylation to form TP-378, followed by ROS attacking the quinolone ring to generate TP-338. Further ROS attack led to piperazine ring cleavage, ultimately forming TP-158. In Pathway Ⅱ, the decarboxylation of OFL occurred to produce TP-318, after which ROS attack the quinolone ring, breaking the carbon-carbon double bond and inducing hydrolysis. The piperazine ring decomposed, and the oxazine ring opened to form TP-200, which further underwent defluorination to yield TP-150. In Pathway Ⅲ, the piperazinyl substituent of OFL was first attacked by ROS to form TP-279, followed by oxazine ring opening, addition reactions, and decarboxylation to generate TP-205. Finally, under further ROS reactions, TP-112 was produced.

  The Co₃O₄−Mt-4Ov/PMS process can effectively degrade OFL and ultimately mineralize it into CO₂ and H₂O. However, some intermediates retain an intact fluoroquinolone core, which may pose ecological hazards. Therefore, toxicity predictions were conducted using the ecological structure activity relationship (ECOSAR) program to analyze the acute and chronic toxicity of the degradation intermediates. As shown in Fig. S19 (Supporting information), the toxicity of three intermediates (TP-378, TP-158, and TP-112) was lower than that of OFL, while the toxicity of other byproducts became higher, indicating that some intermediates carry greater ecological risks. In particular, TP-318, TP-150, and TP-205 exhibited higher toxicity, likely due to decarboxylation at the quinolone substituent position or the presence of highly toxic aniline structures. Although some intermediates are more toxic, the system achieves a high mineralization efficiency of 90.3% (Fig. S20 in Supporting information). Therefore, by optimizing reaction conditions to ensure complete degradation, the potential toxicity risks posed by intermediate byproducts could be largely reduced. Moreover, an analysis of the degradation pathways revealed that all three intermediates in Pathway Ⅱ exhibit high toxicity. This observation suggested that future research should focus on regulating degradation pathways and stages to prevent the formation of toxic products.

  In this study, different levels of oxygen vacancy (O_V) were successfully constructed on the Co₃O₄-encapsulated montmorillonite (namely Co₃O₄−Mt-xO_V, x = 2, 4, 6) through a two-step process involving chemical etching reduction and oxidation. Results showed that the performance of activated PMS was markedly improved through the introduction of O_V on Co₃O₄−Mt. The degradation rate of ofloxacin (OFL) in the Co₃O₄−Mt-xO_V/PMS process was positively correlated with the content of OV, suggesting the significant role of O_V for PMS activation. Comprehensive experimental and theoretical analysis unveiled that O_V enabled the modification of the electronic structure in Co₃O₄ and enhanced the adsorption capacity for PMS and OFL. Moreover, O_V could modulate the catalytic decomposition of PMS, thereby predominantly generating ¹O₂ and Co(Ⅳ)=O, which behaved as the primary reactive species for OFL oxidation. This non-radical-mediated process enabled the system to selectively remove OFL from the contaminated water without interference from the background matrix of the water body. The Co₃O₄−Mt-4O_V also possessed outstanding chemical stability and reusability, making it a highly viable candidate for practical applications. The knowledge obtained in this study will advance the understanding of vacancy defect engineering in PMS activation, and guide the design of efficient catalysts. This research holds significant importance for the rational design and synthesis of environmentally friendly natural mineral materials rich in O_V, as well as for the advancement of highly efficient environmental catalytic materials.

  Declaration of competing interest

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

  CRediT authorship contribution statement

  Xiao-qiang Cao: Writing – review & editing, Writing – original draft, Visualization, Supervision, Funding acquisition. Xingyao Liu: Methodology, Investigation, Data curation. Yaqi Wang: Visualization, Methodology, Investigation. Haining Wang: Software, Investigation. Bo Wei: Software, Methodology. Yanan Shang: Validation, Supervision. Yizhen Zhang: Project administration, Data curation. Yujiao Kan: Project administration, Conceptualization. Yang Zhang: Methodology, Investigation. Xing Xu: Supervision, Resources. Longlong Zhang: Writing – review & editing, Writing – original draft, Supervision, Project administration, Methodology, Funding acquisition.

  Acknowledgments

  The research work was supported by National Natural Science Foundation of China (Nos. 22476116, 52074176, 52400090), Natural Science Foundation of Shandong Province (Nos. ZR2024ME156, ZR2024QB138), and Qingdao Natural Science Foundation (No. 24-4-4-zrjj-70-jch).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2025.112112.

  
  
• 1. [1]

     J. Yang, S. Wang, X. Luo, et al., Chin. Chem. Lett. 36 (2025) 110996. doi: 10.1016/j.cclet.2025.110996

  2. [2]

     C. Zhang, N. Ding, Y. Pan, et al., Chin. Chem. Lett. 35 (2024) 109579. doi: 10.1016/j.cclet.2024.109579

  3. [3]

     Y. Lin, Y. Wang, Z. Wang, et al., Nat Commun. 15 (2024) 10032. doi: 10.1038/s41467-024-54225-x

  4. [4]

     Z. Weng, Y. Lin, S. Guo, et al., Angew. Chem. Int. Ed. 62 (2023) e202310934. doi: 10.1002/anie.202310934

  5. [5]

     Q. Fang, H. Yang, S. Ye, et al., Water Res. 245 (2023) 120614. doi: 10.1016/j.watres.2023.120614

  6. [6]

     Y. Wang, Y. Lin, S. He, et al., J. Hazard. Mater. 461 (2024) 132538. doi: 10.1016/j.jhazmat.2023.132538

  7. [7]

     J. Zhen, J. Sun, X. Xu, et al., Angew. Chem. Int. Ed. 63 (2024) e202402669. doi: 10.1002/anie.202402669

  8. [8]

     T. Zhang, S. Wu, N. Li, et al., J. Hazard. Mater. 449 (2023) 130971. doi: 10.1016/j.jhazmat.2023.130971

  9. [9]

     Z. Fang, J. Qi, W. Chen, et al., Appl. Catal. B 338 (2023) 123084. doi: 10.1016/j.apcatb.2023.123084

  10. [10]

      L. Kong, G. Fang, Z. Fang, et al., Chem. Eng. J. 416 (2021) 128996. doi: 10.1016/j.cej.2021.128996

  11. [11]

      F. Wang, Y. Xiang, Y. Zhang, X, et al., Chin. Chem. Lett. 37 (2026) 111315. doi: 10.1016/j.cclet.2025.111315

  12. [12]

      S. Chen, J. Li, W. Zhou, et al., Coord. Chem. Rev. 507 (2024) 215749. doi: 10.1016/j.ccr.2024.215749

  13. [13]

      J. Jiang, Z. Zhao, J. Gao, et al., Environ. Sci. Technol. 56 (2022) 5611–5619. doi: 10.1021/acs.est.2c01913

  14. [14]

      Q. Wang, J. Lu, M. Yu, et al., Environ. Pollut. 333 (2023) 121990. doi: 10.1016/j.envpol.2023.121990

  15. [15]

      S. Xue, C. Cheng, J. Kang, et al., Chin. Chem. Lett. 36 (2025) 110776. doi: 10.1016/j.cclet.2024.110776

  16. [16]

      P. Li, Y. Lin, S. Zhao, et al., Appl. Catal. B 298 (2021) 120596. doi: 10.1016/j.apcatb.2021.120596

  17. [17]

      J. Lim, Y. Yang, M.R. Hoffmann, Environ. Sci. Technol. 53 (2019) 6972–6980. doi: 10.1021/acs.est.9b01449

  18. [18]

      J. Wang, X. Duan, J. Gao, et al., Water Res. 185 (2020) 116244. doi: 10.1016/j.watres.2020.116244

  19. [19]

      L. Yang, Y. Jiao, X. Xu, et al., ACS Sustainable Chem. Eng. 10 (2022) 1899–1909. doi: 10.1021/acssuschemeng.1c07605

  20. [20]

      J. Hu, X. Zeng, G. Wang, et al., Chem. Eng. J. 400 (2020) 125869. doi: 10.1016/j.cej.2020.125869

  21. [21]

      X. Dong, X. Duan, Z. Sun, et al., Appl. Catal. B 261 (2020) 118214. doi: 10.1016/j.apcatb.2019.118214

  22. [22]

      Q. Liu, H. Qie, Z. Sun, et al., Chin. Chem. Lett. 34 (2023) 108397. doi: 10.1016/j.cclet.2023.108397

  23. [23]

      R. Yao, J. Pinals, R. Dorakhan, et al., ACS Catal. 12 (2022) 12227–12245. doi: 10.1021/acscatal.2c02525

  24. [24]

      J. Liang, L. Fu, K. Gao, X. Duan, Appl. Catal. B 315 (2022) 121542. doi: 10.1016/j.apcatb.2022.121542

  25. [25]

      P. Xu, H. Qin, Q. Tian, et al., Chem. Eng. J. 504 (2025) 158836. doi: 10.1016/j.cej.2024.158836

  26. [26]

      W. Gao, X. Cao, L. Hou, et al., Chin. Chem. Lett. 37 (2026) 111095. doi: 10.1016/j.cclet.2025.111095

  27. [27]

      N. Chen, M. Huang, C. Liu, et al., Water Res. 165 (2019) 114997. doi: 10.1016/j.watres.2019.114997

  28. [28]

      X. Li, I.C. Bourg, Environ. Sci. Technol. 58 (2024) 1109–1118. doi: 10.1021/acs.est.3c08253

  29. [29]

      X. Liu, C. Tournassat, S. Grangeon, et al., Nat. Rev. Earth Environ. 3 (2022) 461–476. doi: 10.1038/s43017-022-00301-z

  30. [30]

      X. Liu, W. Wang, C. Du, Y. Su, Appl. Clay Sci. 228 (2022) 106625. doi: 10.1016/j.clay.2022.106625

  31. [31]

      X. Zhang, X. Yang, S. Chen, et al., Chem. Eng. J. 452 (2023) 139192. doi: 10.1016/j.cej.2022.139192

  32. [32]

      X. Chen, X. Peng, L. Jiang, et al., Chem. Eng. J. 427 (2022) 130945.

  33. [33]

      B. Lei, W. Cui, P. Chen, et al., ACS Catal. 12 (2022) 9670–9678. doi: 10.1021/acscatal.2c02390

  34. [34]

      Y. Li, M. Wang, J. Tian, et al., Appl. Surf. Sci. 669 (2024) 160588. doi: 10.1016/j.apsusc.2024.160588

  35. [35]

      S. Yuan, X. Liu, W. Liao, et al., Geochim. Cosmochim. Acta 223 (2018) 422–436. doi: 10.3390/icem18-05277

  36. [36]

      Z. Wang, R. Lin, Y. Huo, et al., Adv. Funct. Mater. 32 (2022) 2109503. doi: 10.1002/adfm.202109503

  37. [37]

      X. Xue, R. Chen, H. Chen, et al., Nano Lett. 18 (2018) 7372–7377. doi: 10.1021/acs.nanolett.8b03655

  38. [38]

      J.H. Jhang, J.A. Boscoboinik, E.I. Altman, J. Chem. Phys. 152 (2020) 084705. doi: 10.1063/1.5142621

  39. [39]

      Z. Zhou, F. Yu, J. Ma, Environ. Chem. Lett. 20 (2022) 563–595. doi: 10.1007/s10311-021-01355-z

  40. [40]

      Y. Wang, M. Liu, Y. Zang, et al., J. Environ. Manage. 321 (2022) 115971. doi: 10.1016/j.jenvman.2022.115971

  41. [41]

      L. Cao, Y. Chai, Y. Zhang, et al., J. Environ. Manage. 375 (2025) 124217. doi: 10.1016/j.jenvman.2025.124217

  42. [42]

      T. Jiang, B. Wang, B. Gao, et al., J. Hazard. Mater. 442 (2023) 130075.

  43. [43]

      W. Li, Y. Nan, Z. Zhang, Q. You, Z. Jin, Chem. Eng. J. 398 (2020) 125449.

  44. [44]

      Y. Li, X. Zhao, Y. Yan, et al., Chem. Eng. J. 376 (2019) 121302.

  45. [45]

      F. Ghanbari, M. Moradi, Chem. Eng. J. 310 (2017) 41–62.

  46. [46]

      Z. Hao, W. Hou, C. Fang, et al., J. Hazard. Mater. 439 (2022) 129618.

  47. [47]

      Y. Guo, J. Long, J. Huang, et al., Water Res. 215 (2022) 118275.

  48. [48]

      Z. Luo, Y. Yan, R. Spinney, et al., Water Res. 261 (2024) 122023.

  49. [49]

      L. Zhang, X. Zhou, L. Yang, et al., Appl. Catal. B 349 (2024) 123897.

  50. [50]

      W. Chen, Y. Jiang, M. Zhao, et al., Anal. Chem. 95 (2023) 5340–5345. doi: 10.1021/acs.analchem.2c05459

  51. [51]

      Q. Yi, J. Ji, B. Shen, et al., Environ. Sci. Technol. 53 (2019) 9725–9733. doi: 10.1021/acs.est.9b01676

  52. [52]

      T. Liu, S. Xiao, N. Li, et al., Nat. Commun. 14 (2023) 2881.

  53. [53]

      Y. Zou, J. Li, J. Tan, et al., Chem. Eng. J. 471 (2023) 144531.

  54. [54]

      P. Xu, H. Qin, Q. Tian, et al., Chem. Eng. J. 504 (2025) 158836.

  55. [55]

      X. Hu, J. Wang, J. Wang, et al., Appl. Catal. B 318 (2022) 121879.

  56. [56]

      Y. Zong, H. Zhang, X. Zhang, et al., Appl. Catal. B 300 (2022) 120722.

  57. [57]

      M. Ta, T. Zhang, T. Wang, et al., J. Mater. Chem. A 12 (2024) 17529–17543. doi: 10.1039/d4ta02383h

  58. [58]

      Z. Su, W. Yang, C. Wang, et al., Environ. Sci. Technol. 54 (2020) 12684–12692. doi: 10.1021/acs.est.0c03981

  59. [59]

      Y. Zeng, F. Wang, D. He, et al., Chem. Eng. J. 459 (2023) 141642.

  60. [60]

      C. Ruan, L. Zhang, Z. Song, et al., ChemistrySelect 10 (2025) e202405300.

  61. [61]

      J. Wang, S. Wang, Chem. Eng. J. 411 (2021) 128392.

  62. [62]

      M. Zhao, Y. Liu, M. Feng, et al., Chem. Eng. J. 489 (2024) 151371.

  63. [63]

      H. Luo, H. Fu, H. Yin, Q. Lin, J. Hazard. Mater. 426 (2022) 128044.

• 

  Figure 1  (a) Schematic illustration of the synthesis of the Co₃O₄−MT-O_V. (b) XRD, (c) BET, (d) EPR, (e) O_V contents, (f) Co 2p HR-XPS and (g) O 1s HR-XPS of the catalysts.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  SEM images of (a) Na-Mt, (b) Co₃O₄, (c) Co₃O₄−Mt-4O_V, and (d) element mapping of Co₃O₄−Mt-4O_V. TEM images of (e) Na-Mt, (f) Co₃O₄, (g) Co₃O₄−Mt-4O_V, and HRTEM images of (h) Co₃O₄, (i) Co₃O₄−Mt-4O_V.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) Kinetics of OFL degradation in different systems, and (b) the corresponding degradation rate (k_obs). (c) The linear relationship between O_V content of Co₃O₄−Mt-xO_V and k_obs of OFL. (d) Effects of solution pH on the degradation of OFL in the Co₃O₄−Mt-4O_V/PMS system, and (e) the corresponding k_obs values. (f) Zeta potential of Co₃O₄−Mt-4O_V changed with varying solution pH. Experiment conditions: [OFL]₀ = 20 mg/L and T = 20 ℃, if not specified catalyst dosage = 0.2 g/L, [PMS]₀ = 1.0 mmol/L and pH₀ = 7.0.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) OFL degradation in the Co₃O₄−Mt-4O_V/PMS process with and without the existence of different quenchers, and (b) the corresponding k_obs values. (c) Degradation of OFL in the D₂O solution (insert: the corresponding k_obs values). The EPR spectra of (d) DMPO-SO₄^•− and DMPO-^•OH, (e) DMPO—O₂^•− and (f) TEMP-¹O₂. Experiment conditions: catalyst dosage = 0.2 g/L, [PMS]₀ = 1.0 mmol/L, [OFL]₀ = 20 mg/L, pH₀ = 7.0 and T = 20 ℃, if applied [MeOH]₀ = [TBA]₀ = 100 mmol/L, [FFA]₀ = [p-BQ]₀ = 10 mmol/L.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  (a) EPR spectra of Co₃O₄−Mt-4O_V catalyst before and after use. (b) Electrochemical impedance spectroscopy and (c) cyclic voltammetry spectra of Co₃O₄, Co₃O₄−Mt and Co₃O₄−Mt-4O_V. (d) The calculated charge density difference of PMS adsorbed on (a) Co₃O₄−Mt and (b) Co₃O₄−Mt-4O_V.

  下载: 全尺寸图片 幻灯片
  

  Figure 6  The mechanism of PMS activated with Co₃O₄−Mt-4O_V for OFL degradation.

  下载: 全尺寸图片 幻灯片
  

  Figure 7  (a) Effects of common anions in water on OFL degradation in the Co₃O₄−Mt-4O_V/PMS process. (b) Performance of OFL degradation during the cycling tests of Co₃O₄−Mt-4O_V, and (c) the leaching of cobalt ions during the cycling tests. (d) Degradation of OFL in the Co₃O₄−Mt-4Ov/PMS process in real waters. Experiment conditions: Catalyst dosage = 0.2 g/L, [PMS]₀ = 1.0 mmol/L, [OFL]₀ = 20 mg/L, pH₀ = 7.0 and T = 20 ℃, if applied [Cl⁻]₀ = [H₂PO₄⁻]₀ = [HCO₃⁻]₀ = [SO₄²⁻]₀ = 50 mmol/L.

  下载: 全尺寸图片 幻灯片
•