Facile fabrication of MOF and natural polymer-derived carbon-aerogels with multiscale porosity for persulfate activation in water treatment

Abstract

Heteroatom-doped carbon materials have attracted significant attention as efficient catalysts for sulfate radical-based advanced oxidation processes (SR-AOPs). However, the practical application of powdered carbon materials has been hindered by challenges related to reusability. In this study, we developed a sustainable, reusable, and high-performance cobalt- and nitrogen-doped carbon aerogel (CoNCA) using agarose (AG) and zeolitic imidazolate framework-67 (ZIF-67) as metal-free catalysts for SR-AOPs. ZIF-67 acted as cobalt-doped carbon, while AG served as a porous, freestanding matrix to capture ZIF-67 during freeze-drying and pyrolysis. The CoNCA demonstrated excellent catalytic performance, achieving 95% methylene blue decomposition within 4 min through the activation of peroxymonosulfate. We systematically investigated the effects of PMS, catalyst, and organic pollutant dosages, along with pH and temperature variations, on catalytic performance. The CoNCA also exhibited excellent dye removal performance through immersion and pyrolysis processes, demonstrating high reusability. The material's high recovery rate using simple magnetic treatment highlights its feasibility as a purification filter. Moreover, the CoNCA exhibited catalytic efficiency in the degradation of 2,4,6-trichlorophenol, achieving 80% decomposition within 30 min. This approach presents a facile, reusable, sustainable, and high-performance carbon aerogel with a porous structure, offering significant potential as a next-generation metal-free catalyst.

---

1. Introduction

Environmental pollution caused by the rapid growth of industries has become a critical challenge.^1,2 Among various types of pollution, water contamination is particularly urgent, as it impacts both ecosystems and human health.^3,4 To address this, organic pollutants, the primary contributors to water pollution, are treated using methods such as adsorption, photocatalysis, electrochemical processes, and oxidation to obtain clean water.^5–7 Among these approaches, advanced oxidation processes (AOPs) have gained significant attention due to their rapid reaction rates and high effectiveness in degrading organic pollutants.^8,9 Recently, sulfate radical-based AOPs (SR-AOPs), which utilize sulfate radicals (SO₄˙⁻) for the degradation of organic compounds, have emerged as highly efficient methods due to their long half-life (30–40 μs), high selectivity, high redox potential (2.5–3.1 V), and stability across a broad pH range (2.0–9.0).^10–13 Sulfate radicals are typically generated by activating peroxymonosulfate (PMS) or persulfate using various techniques, including using transition metals, sonication, ultraviolet irradiation, heat treatment, and catalytic processes.^14,15 However, certain methods have limitations. For instance, sonication requires specific temperature and is primarily restricted to laboratory-scale applications.¹⁶ Similarly, ultraviolet light constitutes only 3–5% of the sun's output, limiting its broader applicability and warranting further research.¹⁷ Therefore, SR-AOPs using catalysts as activating agents, which require no additional energy input and are simple to prepare, have been widely studied.¹⁸ The performance of such systems depends heavily on the choice of catalyst material, leading to extensive research into material development. Recent studies have focused on generating sulfate radicals using transition metals (Co²⁺, Mn²⁺, Fe²⁺, and Cu²⁺) or metal oxide-catalyzed PMS activation.^19,20 However, these methods face challenges in environmental remediation due to metal leaching, which can result in secondary contamination.²¹

Metal–organic frameworks (MOFs), composed of organic ligands coordinated with metal centers, offer several advantages, including a large surface area and porous structure. The metal and organic linkers can be tailored to achieve specific properties, making MOFs versatile for applications in tissue growth, energy storage devices, adsorbents, and catalysts.^22,23 Among MOFs, ZIF-67, composed of cobalt (Co) and an imidazole linker, is one of the most effective and widely used materials in PMS catalysis. It has gained considerable attention due to its ease of large-scale production, simple synthesis process, high specific surface area, and environmentally friendly preparation methods that rely solely on water.²⁴ Catalysts based on MOFs can be developed through three main approaches: as pristine MOFs, as composites by integrating MOFs with other materials, or as derivatives obtained via doping or carbonization.²⁵ However, recovering MOFs without supporting matrices remains challenging, limiting their reusability and leading to secondary pollution caused by transition metal leaching. Additionally, the long-term stability of pristine MOFs in water is inadequate, as they are prone to leaching, which reduces their cycle stability during reuse.²⁶ To address these limitations and enhance catalytic performance, incorporating additional materials into MOFs has proven to be an effective strategy. Recent research has focused on MOF-derived materials and their composites as promising catalysts for PMS activation in SR-AOPs. Co-based catalyst systems have been explored for solar-assisted degradation and steam generation applications, while MOF-based membranes subjected to thermal modification have demonstrated improved PMS activation via non-radical pathways.^27,28 Additionally, composites such as PDINH/MIL-88A(Fe) and Co-doped Fe-MOFs have shown enhanced degradation performance through the synergistic generation of reactive oxygen species, as supported by theoretical calculations.^29,30 These studies collectively highlight the increasing attention on designing catalysts derived from MOFs specifically for advanced oxidation processes, particularly those driven by sulfate radicals.

Aerogels, characterized by high specific surface area, ultra-lightweight properties, and porous structures, are micro- and mesoporous solids containing air or gas within a three-dimensional framework. Their significant void volume makes them highly attractive as structural materials and carriers for various substances.³¹ The porous structure of aerogels enables the capture and retention of materials while providing ample space for reactions with external agents. These unique properties render aerogels versatile for applications such as catalysts, batteries, filters, and gas storage. Aerogels can be synthesized from various inorganic materials, including silica, metals, and oxides, as well as from synthetic and natural polymers.^32,33 Recent research has increasingly emphasized natural polymers due to their eco-friendly and biodegradable nature.³⁴ Among these, agarose (AG) offers several advantages, including non-toxicity, cost-effectiveness, a reversible sol–gel transition, and ease of size and shape control. Furthermore, AG can form a gel without requiring chemical additives, enabling faster and simpler fabrication compared to other natural polymers such as cellulose and starch.³⁵ Building on these properties, AG serves not only as a sustainable carbon precursor but also as a structural scaffold that facilitates the uniform dispersion of nanoparticles or precursors during synthesis. In addition, its encapsulating nature enables more precise control over the amount and distribution of embedded materials within the matrix, which is advantageous for tailoring the composition of the resulting carbon materials. Recent studies have reported the fabrication of carbon-based freestanding matrices from aerogels for use in structural and functional composites. During pyrolysis to form carbon aerogels, the macropore volume decreases while mesopore and micropore volumes increase, indicating the retention of the porous structure. These intrinsic characteristics and structures enable multiple functionalities, including (1) acting as an oxidation source, (2) serving as a support material for metal catalysts, (3) functioning as heterogeneous catalysts, and (4) facilitating adsorption.³⁶ However, since the catalytic performance of pure single-carbon materials is often inefficient, elemental doping has been explored to enhance performance, though the process is complex. Developing a facile strategy to fabricate doped carbon aerogels without additional post-carbonization treatment remains a significant challenge.

In this study, we developed an effective method for preparing porous heteroatom-doped carbon aerogels with high catalytic performance through the pyrolysis of ZIF-67 and agarose (Z67@AG) composites. During pyrolysis, Co- and N-doped carbon aerogels (CoNCAs) were successfully fabricated, with their porosity and catalytic performance controllable by varying the ratio of ZIF-67 to AG. As a freestanding matrix, AG encapsulated ZIF-67 during gelation, contributing a microporous structure that facilitated the diffusion of organic pollutants. Building on these properties, AG serves not only as a sustainable carbon precursor but also as a structural scaffold that promotes the uniform dispersion of nanoparticles during synthesis facilitating rapid structure formation. Furthermore, its encapsulating nature allows for more precise control over the quantity and distribution of embedded materials within the matrix, which is beneficial for customizing the composition of the resulting carbon materials. The pyrolyzed ZIF-67 provided Co ions, which enhanced catalytic performance by activating PMS and degrading methylene blue (MB). Notably, CoNCAs exhibited enhanced catalytic performance compared to Z67@AG, demonstrating that pyrolysis and the incorporation of carbon with transition metals significantly improved functionality.³⁷ The CoNCA removed 98% of 10 mg per L MB within 4 min and degraded over 80% of various organic dyes, including RhB, CV, Rh6G, and MB, in just 5 min. Furthermore, the CoNCA is reusable, can be collected magnetically, and functions as a filter. For instance, the CoNCA removed 80% of 20 mg per L 2,4,6-trichlorophenol (2,4,6-TCP) within 40 min, demonstrating versatile decomposition capacity. These findings suggest that our approach yields a catalyst with high catalytic performance and significant promise for sustainable catalyst development.

2. Experimental section

The details of the materials, characterization, and evaluations are presented in the ESI.†

3. Results and discussion

Scheme 1 illustrates the manufacturing process of CoNCAs, highlighting the limitations of conventional materials. Initially, AG, a purified form of agar derived from natural sources, forms the hydrogel through gelation upon heating and subsequent cooling (Scheme 1a). This hydrogel is then freeze-dried to create a porous aerogel, which can be pyrolyzed to produce a carbon aerogel. However, its catalytic performance in decomposing organic dyes is limited due to single-element doping (Scheme 1b). Pyrolyzing ZIF-67 yields Co and N-doped carbon, but an additional recovery step is required, and shrinkage limits its standalone application (Scheme 1c).

Scheme 1 (a) The schematic illustration and photographic images of AG. (b–d) The schematic illustration of (b) carbonized agarose, (c) carbonized ZIF-67 NP, and (d) the CoNCA. (e) Proposed decomposition mechanism of MB under a PMS system using CoNCA as a substrate. (f) Photographic images of Z67@AG and CoNCAs before and after pyrolysis.

To overcome these issues, an aerogel was prepared by mixing ZIF-67 nanoparticles (NPs) with AG, followed by gelation and freeze-drying, as shown in Scheme 1d. This process effectively encapsulates ZIF-67 within the AG framework, as confirmed by SEM images in Fig. 1. Scheme 1e compares the reaction mechanisms of solid carbon and hollow carbon (CoNCA). Solid carbon, which lacks a porous structure, restricts mass transfer and ion diffusion of contaminants and PMS, as these species cannot penetrate its interior. Consequently, the reaction is confined to the surface of solid carbon. In contrast, hollow carbon aerogels possess a porous and hollow configuration, facilitating mass transfer and ion diffusion, which increases the number of active sites available for the reaction. This enhanced structure significantly improves catalytic performance compared to solid carbon. Finally, Scheme 1f presents photographic images of CoNCAs in various shapes, such as squares, cylinders, and disks, achieved by preparing the composites in different molds. These images demonstrate the ease with which the aerogel's shape can be controlled. Moreover, the aerogels retain their shape after pyrolysis, confirming the successful preparation of carbon-based aerogels.

Fig. 1 SEM images of as-synthesized (a and b) ZIF-67 nanoparticles, (c and d) Z67@AG (2 : 1), (e and f) the CoNCA (2 : 1), (g and h) the CoNCA (1 : 1), (i and j) the CoNCA (1 : 2), and (k and l) the CoNCA (1 : 4), and (m) the EDS element mapping of the CoNCA (2 : 1).

The surface structure and morphology of ZIF-67 NPs and CoNCAs were characterized as a function of the ratio of ZIF-67 to AG using SEM. As shown in Fig. 1a and b, the SEM images and photographic representation of ZIF-67 nanoparticles reveal a typical dodecahedron-shaped morphology, with particle sizes ranging from 900 to 1000 nm and a purple-colored powder. Z67@AG, a composite aerogel of ZIF-67 NPs and AG, shows ZIF-67 particles covered with AG on the surface, demonstrating that the porous structure of AG can effectively immobilize the ZIF-67 particles (Fig. 1c and d). Additionally, the color of Z67@AG, as shown in Fig. 1d, remains purple, indicating that the porous AG successfully immobilizes the ZIF-67 particles while allowing the NPs to maintain their shape. Fig. 1e–l present SEM images of Z67@AG after pyrolysis at 900 °C, revealing that the encapsulated ZIF-67 particles are visible on the surface, with no structural collapse observed after carbonization. As shown in Fig. 1e and f, it is evident that the color of the aerogel changed to black when Z67@AG was carbonized, and a rough surface could be detected. Furthermore, the CoNCA (2 : 1) exhibits a uniform size distribution, with particle sizes ranging from 0.86 μm to 0.96 μm, indicating the successful production of homogeneous carbon (Fig. S1†). When the ratio of AG is increased, the heteroatom-doped carbon is well-packed, with a uniform particle size distribution and a rough surface attributed to the presence of metallic NPs (Fig. 1g and h). The anchored surface due to AG is observable at both low and high resolutions when the weight ratio of AG to ZIF-67 is 1 : 2 (Fig. 1i and j). As shown in Fig. 1k and l, the carbonized particles are densely packed when AG and ZIF-67 are mixed at a 1 : 4 ratio and subsequently pyrolyzed, resulting in the smoothest surface compared to that of other CoNCAs. Elemental mapping of cobalt, carbon, oxygen, and nitrogen verified the presence of doped elements in the carbon (Fig. 1m). Cobalt mapping confirmed that the CoNCA (2 : 1), which contains the highest amount of ZIF-67, exhibits a large area of Co distribution, with nitrogen also detected, indicating that the aerogel with Co and N-doped carbon was successfully prepared.

Fig. 2 presents the TEM and HRTEM images of the CoNCA as a function of the weight ratio of ZIF-67 to AG. As shown in Fig. 2a–h, the black circles observed in the TEM images represent Co NPs encapsulated within the carbonaceous matrix. The results indicate that the quantity of Co NPs decreases as the AG content increases, consistent with the encapsulation process described in previous studies.³⁸ The CoNCA (2 : 1), which contains the highest ZIF-67 content, demonstrates a high and uniform distribution of Co within the carbon aerogel, as supported by the Co mapping in Fig. 1m and 2a, e. In contrast, Fig. 2d and h show that the distribution of Co NPs in the porous carbon is significantly reduced when AG content is increased, confirming that the level of Co doping in porous carbon can be effectively controlled. HRTEM analysis of CoNCA 2 : 1, which exhibits the highest ZIF-67 content, further elucidates the morphology of Co NPs, as shown in Fig. 2i–l. As shown in Fig. 2j, multilayered graphitic carbon surrounding the Co NPs is observed, with a lattice spacing of 0.208 nm corresponding to the (002) plane. Additionally, Fig. 2k and l reveal lattice fringes with spacings of 0.181 nm and 0.206 nm, corresponding to the (111) and (200) planes, respectively.³⁸

Fig. 2 TEM images of the (a and e) CoNCA (2 : 1), (b and f) CoNCA (1 : 1), (c and g) CoNCA (1 : 2), and (d and h) CoNCA (1 : 4). (i–l) HRTEM images of the CoNCA (2 : 1).

The specific surface area and pore size distribution of the CoNCA were analyzed, as shown in Fig. 3a–c and detailed in Table S2.† The CoNCAs exhibit type-IV isotherms with hysteresis loops, indicating mesoporous structures. The BET surface area of the CoNCA (2 : 1) is the lowest at 158.32 m² g⁻¹, while the CoNCA (1 : 2) has a slightly higher surface area of 168.08 m² g⁻¹. This decrease in surface area with increasing ZIF-67 content is likely due to the immobilization of ZIF-67 caused by the crosslinking and porous structure of AG. In contrast, the CoNCA (1 : 4), which contains the lowest amount of ZIF-67, demonstrates the highest BET surface area of 276.55 m² g⁻¹. In addition, all CoNCAs exhibit pore sizes between 2.7 nm and 5.1 nm, as indicated by the BJH model in Fig. 3b, c and Table S2,† confirming their mesoporous nature. Bulk densities of the CoNCA as a function of ZIF 67 loading was measured as shown in Fig. S2.† Interestingly, the CoNCA (1 : 4), which has the lowest ZIF 67 content, presents the highest density (143.6 mg cm⁻³). In contrast, the CoNCA (1 : 1), which possesses the second highest ZIF 67 loading, exhibits the lowest density (84.3 mg cm⁻³). The unusually high density of the CoNCA (1 : 4) might have been due to extensive matrix shrinkage and collapse of finer pores during carbonization, driven by its high agarose fraction, which led to a marked reduction in the total pore volume and average pore diameter. By comparison, the CoNCA (1 : 2) combines the largest total pore volume (0.2162 cm³ g⁻¹) and widest average pore diameter (5.144 nm), thereby maximizing void space within the monolith and yielding the lowest bulk density.

Fig. 3 (a) N₂ adsorption–desorption isotherms, (b and c) pore size distributions, and (d) Raman spectra of various CoNCAs, (e) XRD patterns of Z67@AG, AG, and CoNCAs. (f) XPS survey spectrum of CoNCAs with different ZIF-67 to AG ratios. High-resolution XPS spectra of (g) C 1s, (h) N 1s, and (i) Co 2p for the CoNCA (2 : 1).

Disordered carbon (D) at 1340 cm⁻¹ and graphitic carbon (G) at 1600 cm⁻¹ were identified in the Raman spectra of the prepared CoNCA (Fig. 3d). The I_D/I_G ratio for the CoNCA (2 : 1) is 0.99, while this ratio increases to 1.09, 1.10, and 1.20 for the CoNCA (1 : 1, 1 : 2, and 1 : 4), respectively, indicating an increase in carbonaceous components. The decrease in the I_D/I_G ratio with a reduction in AG content is likely due to the introduction of larger defects.³⁹ XRD patterns of CoNCAs were analyzed to investigate their structure, as shown in Fig. 3e. Compared to Z67@AG and neat AG, the CoNCA (1 : 4) spectrum, represented in pink, exhibits a peak at 25.9°, indicating successful carbonization and the formation of well-defined carbon corresponding to the (002) plane.²⁶ Peaks at 44.2° and 51.5° are observed for all CoNCAs, corresponding to the face-centered cubic metallic Co NPs' (111) and (200) planes, respectively. The peak intensity increases with higher AG content, reflecting an increased amount of Co NPs.^40,41 These findings are consistent with the HRTEM results presented in Fig. 2i–l. The chemical composition of CoNCAs was analyzed using XPS, as shown in Fig. 3f, which demonstrates strong intensities for carbon (C), nitrogen (N), oxygen (O), and cobalt (Co). High-resolution XPS data for the CoNCA (2 : 1) detailing C, N, and Co are presented in Fig. 3g–i, with additional oxygen data provided in Fig. S3.† The high-resolution C 1s spectrum for the CoNCA (2 : 1) displays three peaks at 284.6, 285.0, and 289.3 eV, corresponding to C–C/C=C bonds, C–O/C–N bonds, and C=O bonds, respectively. These results confirm the presence of active functional groups and sp²-carbon structures.⁴² The high-resolution N 1s spectrum in Fig. 3h can be deconvoluted into five peaks attributed to pyridinic-N (398.1 eV), Co–N (399.1 eV), pyrrolic-N (398.9 eV), graphitic-N (401.0 eV), and oxidized-N (403.7 eV).^43–45 During pyrolysis, transition metal-containing materials are known to enhance the content of graphitic-N and pyridinic-N, which are critical for oxygen reduction.⁴⁶ The presence of graphitic-N and Co–N facilitates heterogeneous electron transfer and enhances the catalytic activity of nitrogen-doped carbon materials, providing a strong basis for the exceptional catalytic performance of CoNCAs in PMS activation.⁴⁷ The XPS spectra of Co 2p are divided into six sub-peaks as shown in Fig. 3i.⁴⁸ The peaks at 780.1 eV, 781.9 eV, and 787.6 eV could be attributed to metallic Co, Co²⁺ 2p_3/2, and Co–N, respectively. Additionally, the Co 2p_1/2 spectrum contains three peaks corresponding to Co³⁺ 2p_1/2 (795.1 eV), Co²⁺ 2p_1/2 (796.9 eV) and a satellite peak (803.0 eV). These results confirm the coexistence of cobalt oxides and metallic cobalt, which contribute to the excellent catalytic performance.⁴⁹ The high-resolution O 1s spectrum (Fig. S3†) reveals three deconvoluted peaks at 533.9 eV, 532.1 eV, and 529.9 eV, corresponding to C=O, C–O, and Co–O bonds, respectively. These peaks indicate the generation of ¹O₂ and the presence of cobalt oxides, further supporting the catalytic functionality of CoNCAs.⁵⁰

The catalytic degradation of MB through PMS activation with various catalysts was investigated. In the presence of the CoNCA (2 : 1), PMS alone, or AG/PMS alone, the concentration of MB decreased by approximately 20% after 20 min, primarily due to the adsorption of MB molecules onto the surfaces of AG and carbon without significant degradation (Fig. 4a). In contrast, when PMS was combined with other catalysts, including ZIF-67-derived carbon (Z67C), Z67@AG, and CoNCAs, the catalysts activated PMS, leading to the decomposition of the organic dye and a marked reduction in MB concentration. Within 8 min, all catalysts demonstrated degradation efficiencies exceeding 95%, with performance improving as the Co content increased. Notably, the CoNCA (2 : 1) exhibited the highest efficiency, decomposing 95% of MB in 4 min. While Z67C with PMS achieved excellent catalytic performance, decomposing 97% of MB in just 3 min, its application is limited due to challenges in recovery after use, which involves additional cost and time.

Fig. 4 (a) Degradation efficiency and (b) kinetic rates of MB under various catalytic systems within 20 min. Reaction conditions: catalyst dosage = 0.2 g L⁻¹, MB concentration = 0.01 g L⁻¹, PMS dosage = 0.5 g L⁻¹, temperature = 25 °C, and initial pH = 7.

To further analyze the catalytic degradation performance, the kinetic rate constant (k) of CoNCAs was evaluated using a pseudo-first-order reaction model (Fig. 4b). The pseudo-first-order kinetic equation is expressed as:

where k is the rate constant; t is the reaction time; C₀ and C_t are the initial and time-dependent concentrations of MB, respectively. The rate constant was the lowest for the CoNCA (1 : 4) at 0.226 min⁻¹, which contained the least amount of ZIF-67. As the ZIF-67 content increased, the rate constants increased to 0.310 min⁻¹, 0.509 min⁻¹, and 0.635 min⁻¹, respectively. These results indicate that Co in ZIF-67 is primarily responsible for PMS activation and the catalytic reaction. The catalytic performance of CoNCAs was significantly influenced by the precursor ratio of ZIF-67 to agarose, which determined the elemental distribution and microstructure of the resulting carbon material. As the ZIF-67 content increased, the relative concentrations of Co and N in relation to carbon also increased, leading to enhanced degradation efficiency of MB. This improvement can be attributed to the higher density of Co-based active sites that facilitate the activation of PMS through the generation of sulfate radicals (SO₄˙⁻). As shown in Fig. 3i, the increase in cobalt content is reflected by the relative intensities of Co oxides and metallic cobalt species, which are known to contribute to catalytic activity. In parallel, nitrogen doping plays a supportive role by stabilizing Co–N configurations and promoting charge transfer.⁴⁷ Furthermore, the agarose framework likely suppresses Co nanoparticle agglomeration during carbonization and enables a more uniform distribution of Co and N within the carbon matrix. Additionally, the presence of nitrogen species contributes to the stabilization of Co–N structures and promotes electron transfer. Our results demonstrate a consistent improvement in catalytic activity with increasing Co content. This enhancement may be due to the use of agarose as a structural matrix, which likely prevents Co agglomeration and supports the uniform distribution of Co and N during carbonization. These findings highlight the importance of compositional tuning through precursor ratio control and illustrate that optimizing the distribution of Co and N is critical for maximizing the performance of MOF-derived carbon catalysts in SR-AOPs. Additionally, Table S4† compares the catalytic performance of various catalysts and CoNCAs for MB degradation using PMS, confirming that CoNCAs achieve the highest degradation efficiency within a short period, highlighting their potential as effective catalysts for MB decomposition. Moreover, the performance of our SR-AOP system in the degradation of MB was found to be comparable to, or even superior to, that reported in previous studies employing photocatalysis, adsorption, electrochemical oxidation, and electro-adsorption (Table S5†). These results demonstrate the high catalytic efficiency and competitiveness of our system in relation to other decomposition methods. Consequently, the CoNCA (2 : 1) was selected for further experimentation.

For practical applications of MB degradation, we investigated the effects of several factors on the catalytic performance of CoNCAs: (1) PMS concentration, (2) catalyst concentration, (3) MB concentration, (4) pH, (5) temperature, and (6) the degradation performance of various organic dyes (Fig. 5a–f). As shown in Fig. 5a, the removal efficiency increased from 76% to 98% within 10 min as the PMS concentration was increased from 0.125 g L⁻¹ to 0.75 g L⁻¹.

Fig. 5 Effect of (a) PMS dosage, (b) catalyst dosage, (c) initial MB concentration, (d) pH, and (e) temperature on MB degradation, and (f) degradation efficiency of various organic pollutants. Reaction conditions: catalyst dosage = 0.2 g L⁻¹, concentration of MB and various other organic pollutants = 0.01 g L⁻¹, PMS dosage = 0.5 g L⁻¹, temperature = 25 °C, and initial pH = 7.

Correspondingly, the kinetic rate of CoNCAs (2 : 1) increased from 0.094 min⁻¹ to 0.789 min⁻¹ with the same increase in PMS concentration, suggesting that a higher PMS concentration enhances the production of free radicals, thereby improving the catalytic performance for MB degradation (Fig. S4a†). Next, the degradation performance of MB was evaluated as a function of catalyst concentration (Fig. 5b). Increasing the catalyst concentration from 0.1 g L⁻¹ to 0.3 g L⁻¹ provided a greater number of adsorption sites, thereby enhancing the degradation performance of MB. This improvement is evident from the adsorption graph at 0 min, prior to the onset of the catalytic reaction. As shown in Fig. S4b,† the kinetic rate increased from 0.27 min⁻¹ to 0.796 min⁻¹. Since the concentration of the contaminant can also affect catalytic performance, the degradation of MB was tested at varying MB concentrations (Fig. 5c). CoNCAs demonstrated a 98% decomposition efficiency within 10 min when the MB concentration increased from 0.01 g L⁻¹ to 0.02 g L⁻¹. However, when the MB concentration was further increased to 0.03 g L⁻¹, the decomposition efficiency decreased to 65%. Additionally, the rate constant decreased from 0.629 min⁻¹ to 0.085 min⁻¹ as the MB concentration increased from 0.01 g L⁻¹ to 0.03 g L⁻¹ (Fig. S4c†). These results indicate that the CoNCA is capable of effectively degrading substantial amounts of organic pollutants within a short time frame.

pH is a critical factor in the reactions occurring within catalytic systems. Since wastewater typically exhibits a wide range of pH values, the effect of pH on the catalytic performance was investigated. As shown in Fig. 5d, catalytic performance was evaluated under four pH conditions, ranging from pH 3 to pH 9. The lowest performance was observed at pH 9, with 97% degradation of MB achieved within 8 min. In contrast, at pH 3, 97% degradation was accomplished in just 4 min, indicating that acidic conditions are more favorable than alkaline ones. As the pH decreased from 7 to 3, the rate constant increased from 0.629 min⁻¹ to 0.803 min⁻¹ (Fig. S4d†). These results suggest that the PMS catalytic system is effective across a broad range of pH levels. The improved catalytic performance under acidic conditions can be attributed to the increased generation of sulfate radicals at low pH, whereas hydroxyl radicals are predominantly produced at neutral and alkaline pH values.⁵¹

Temperature plays a crucial role in activating PMS ions within PMS-based catalytic systems, facilitating bond cleavage and the formation of oxidizing SO₄˙⁻ radicals. To evaluate this effect, catalytic experiments were conducted at temperatures ranging from 25 °C to 35 °C (Fig. 5e). Adsorption was found to increase with decreasing temperature during the initial 40 min before degradation began. This phenomenon is likely due to the kinetic energy gained by MB molecules as the temperature increases, enabling them to overcome van der Waals forces with the catalyst, thereby hindering adsorption on the catalyst surface.⁵² In contrast to adsorption, degradation exhibited superior catalytic performance, achieving 95% and 97% degradation of MB within 4 min at 25 °C and 35 °C, respectively. The kinetic rate constant of CoNCAs increased from 0.629 min⁻¹ to 0.777 min⁻¹, confirming that catalytic performance improves with increasing temperature (Fig. S4e†).

The Arrhenius equation was applied to determine the activation energy (E_a) for the decomposition of MB using CoNCAs, expressed as follows (eqn (2)):

where E_a is the activation energy (kJ mol⁻¹), R is the universal gas constant (8.314 J mol⁻¹ K⁻¹), A is the Arrhenius factor, k is the kinetic rate constant, and T is the temperature. As shown in Fig. S5,† the value of 16.1 kJ mol⁻¹ in the PMS catalytic system using CoNCAs is lower than values reported in other studies, highlighting the exceptional catalytic activity of CoNCAs.⁵² To further assess its versatility, the CoNCA was tested for the degradation of additional organic pollutants, including Rh6G, RhB, and CV, with all pollutant concentrations set to 0.01 g L⁻¹. Fig. 5f shows that RhB exhibited a degradation efficiency of 95% within 3 min, while other pollutants achieved over 98% degradation within 8 min. The kinetic rate constants for Rh6G, RhB, MB, and CV were determined to be 0.269 min⁻¹, 0.805 min⁻¹, 0.629 min⁻¹, and 0.725 min⁻¹, respectively (Fig. S4f†). These results demonstrate that the CoNCA is highly effective as a catalyst for the degradation of a wide range of organic pollutants.

To detect the reactive oxygen species (ROS) involved in the degradation of MB using the CoNCA in the PMS system, scavengers such as tert-butanol (TBA), ethanol (EtOH), p-benzoquinone (p-BQ), and L-histidine (L-his) were employed to elucidate the underlying mechanism. Cobalt at active sites on the catalyst surface activates PMS to generate reactive species such as sulfate radicals (SO₄˙⁻), hydroxyl radicals (˙OH), and superoxide radicals (O₂˙⁻) in a heterogeneous system. Additionally, the radicals ˙⁻and ˙OH can produce nonradical singlet oxygen (¹O₂). The production of these radicals and the mechanism for MB degradation in the CoNCA/PMS system are described in eqn (3)–(5):

HSO₅⁻ + Co²⁺ → Co³⁺ + SO₄˙⁻ + OH⁻ (3)

HSO₅⁻ + Co³⁺ → Co²⁺ + SO₅˙⁻ + H⁺ (4)

SO₄⁻ + OH⁻ → ˙OH + SO₄²⁻ (5)

Among the scavengers, TBA specifically targets ˙OH, while EtOH scavenges both SO₄˙⁻ and ˙OH. The degradation of MB using the CoNCA was first evaluated by introducing these scavengers. As shown in Fig. 6a, in the absence of scavengers, 97% of MB degraded within 5 min. However, with TBA, the degradation rate decreased from 87% to 65% as the TBA/PMS molar ratio increased from 3 to 300. Similarly, with EtOH, the remaining concentration of MB decreased from 84% to 68% at 5 min as the EtOH/PMS molar ratio increased from 3 to 300. This slight reduction indicates that SO₄˙⁻ and ˙OH are not the primary contributors to PMS activation and MB degradation, suggesting that their roles are relatively minor. The degradation rate constant decreased from 0.630 min⁻¹ to approximately 0.180 min⁻¹ with the addition of TBA and EtOH (Fig. 6b). Despite this reduction, the rate constant remained relatively high, suggesting that other radicals, such as O₂˙⁻, or nonradical species, such as ¹O₂, play a significant role in MB decomposition. In AOPs, both free radical-based methods utilizing O₂˙⁻, SO₄˙⁻ and ˙OH, and nonradical methods employing ¹O₂ can effectively degrade contaminants into active species.⁵³

Fig. 6 Scavenger experiment results for (a) MB degradation and (b) kinetic rates in the presence of TBA and EtOH. Removal efficiency for (c) L-his and (e) p-BQ, along with their corresponding kinetic rates in (d) L-his and (f) p-BQ (reaction conditions: [catalyst] = 0.2 g L⁻¹, [MB] = 0.01 g L⁻¹, [PMS] = 0.5 g L⁻¹, T = 25 °C, initial pH = 7).

To further clarify this point, MB decomposition was investigated using L-his and p-BQ as quenching agents (Fig. 6c and e). L-His effectively scavenges ¹O₂, a typical ROS involved in nonradical processes, while p-BQ serves as a quenching agent for O₂˙⁻.^54,55 Upon the addition of 10 mM of p-BQ and L-his, 62% and 31% of MB degraded after 10 min, with kinetic constants of 0.127 and 0.022 cm⁻¹, respectively, both significantly lower than the control rate of 0.630 cm⁻¹ (Fig. 6d and f). These results indicate that ¹O₂ plays a more significant inhibitory role than O₂˙⁻ in MB degradation. This suggests that both the nonradical process (¹O₂) and the radical process (O₂˙⁻) are major contributors to the degradation reaction involving CoNCAs, whereas the contributions of SO₄˙⁻ and ˙OH are relatively minor. The formation of both radicals and nonradicals, as well as the MB decomposition mechanism, is detailed in eqn (6)–(12):

C-π + HSO₅⁻ → C-π⁺ + SO₄˙⁻ + OH⁻ (6)

C-π⁺ + HSO₅⁻ → C-π + SO₅˙⁻ + H⁺ (7)

HSO₅⁻ → SO₅²⁻ + H⁺ (8)

SO₅²⁻ + H₂O → O₂˙⁻ + SO₄²⁻ + H⁺ (9)

2O₂˙⁻ + 2H₂O → ¹O₂ + H₂O₂ + 2OH⁺ (10)

O₂˙⁻ + ˙OH → ¹O₂ + OH⁺ (11)

SO₄˙⁻/˙OH/O₂˙⁻/¹O₂ + MB → intermediates → H₂O + CO₂ (12)

SO₄˙⁻ can be produced from uniformly cleaved PMS through the migration of electrons from the carbon catalyst via delocalized π-electrons. The formed SO₄˙⁻ can further react with H₂O to produce small amounts of ˙OH. Moreover, O₂˙⁻ can be generated when PMS undergoes self-decomposition under natural conditions.⁵⁶ The formation of ¹O₂ may also result from the recombination of O₂˙⁻ or the reaction between O₂˙⁻ and ˙OH.⁵⁷ Building on these mechanistic insights, the enhanced reaction rate observed with increasing Co content clearly demonstrates that cobalt is the key active center for PMS activation (Fig. 4b and Table S3†). The quenching data presented in Fig. 6 indicate that the trapping of superoxide radicals results in a significant reduction in activity, while the scavenging of singlet oxygen also significantly inhibits degradation. This suggests that Co–N_x sites, which are stabilized by N doping, primarily drive the superoxide pathway, whereas the graphitized carbon framework, derived from agarose, promotes singlet oxygen generation through electron transfer. These kinetic and ROS-quenching results demonstrate that catalytic efficiency stems from the synergistic interaction of Co content (active-site density), N doping (site stabilization and electron mobility), and hierarchical porosity (mass transfer and electron delocalization), further modulated by PMS concentration and pH.

The reusability of catalysts is crucial for economic, environmental, and practical applications. To evaluate this, we tested the degradation of MB using CoNCA-based PMS over ten cycles (Fig. 7a). As shown in Fig. 7b, the used CoNCA after catalytic tests exhibited MB adsorbed on its surface. To remove the adsorbed dye, the catalyst was soaked in ethanol and annealed at 800 °C for 1 h. After annealing, the CoNCA retained its original shape with no visible residue, confirming that the organic dye and other adsorbed residues were successfully removed. The catalytic performance of the CoNCA was further assessed through repeated cycles during the reusability test. The CoNCA demonstrated a high degradation efficiency, achieving a 96% removal efficiency of the target pollutant within 7 minutes after reuse cycles. Even after ten cycles, a removal efficiency of 93% was observed within 8 minutes. This slight decrease in catalytic performance suggests the exceptional reusability and structural stability of the CoNCA catalyst, indicating its significant potential for practical applications in repeated PMS-based oxidation processes. To assess the long-term aqueous stability of CoNCAs, the catalyst was immersed in DIW as a function of time, and its catalytic activity was evaluated (Fig. S6a and b†). As the immersion time increased from 1 h to 7 days, a slight decrease in degradation efficiency was observed. Nevertheless, even after 7 days of immersion, the CoNCA retained a high MB removal efficiency of 93.4%, indicating excellent structural and functional stability. This observation is further supported by the SEM image shown in Fig. S6c,† which still confirms the preserved surface morphology after prolonged water exposure. Moreover, the stability of CoNCAs was assessed by measuring Co-leaching via ICP after immersion in DIW. After 1 h, the Co ion concentration was only 0.188 mg L⁻¹, confirming its initial stability. Even after 24 h, the concentration of Co ions increased to 2.647 mg L⁻¹, yet the CoNCA still maintained a 96% MB-degradation efficiency within 10 min (Fig. S7†), demonstrating its excellent long-term catalytic stability.

Fig. 7 (a and b) Cyclic tests and photographic images of the annealing-treated CoNCA. (c) Magnetic collection system used to recover catalysts after degradation. (d) Photographic representation of the MB degradation system employing the CoNCA as a catalytic filter through vacuum filtration with PMS. (e) UV-vis absorbance spectra comparing initial and purified MB solutions. (f) Degradation test results for 2,4,6-TCP (0.02 g L⁻¹). (g) Proposed mechanism illustrating the decomposition of MB and 2,4,6-TCP using the CoNCA and PMS (reaction conditions: [catalyst dosage] = 0.2 g L⁻¹, [MB] = 0.01 g L⁻¹, [PMS dosage] = 0.5 g L⁻¹, T = 25 °C, and initial pH = 7).

Recovery of the catalyst after use was also evaluated to ensure its reusability. The CoNCA could be easily collected using a magnet after the catalytic tests, and both suspended and floating CoNCA particles were successfully recollected, even after grinding (Fig. 7c). Additionally, we investigated the feasibility of using the CoNCA as a catalytic filter for practical applications. To develop a ground CoNCA filter, the catalyst was ground and placed on a membrane filter. Despite grinding, the CoNCA retained its original structure (Fig. S8†). As shown in Fig. 7d, the ground CoNCA filter demonstrates catalytic performance, as evidenced by the transparency of the filtered water, even after the CoNCA was ground. Fig. 7e shows that the characteristic absorption peak of MB at 664 nm was significantly reduced within 1 min. Additionally, the CoNCA can be placed in tea bags and used as a catalyst, highlighting its potential for practical applications (Fig. S9†). Chlorophenols, widely used in herbicides, fungicides, wood preservatives, and the production of paper, adhesives, dyes, and plastics, are commonly detected in agricultural and industrial wastewater due to insufficient removal during treatment processes.⁵⁸ Among these, 2,4,6-trichlorophenol (2,4,6-TCP) poses a major environmental challenge as it is bio-accumulative, highly toxic, and carcinogenic.⁵⁹ Effective degradation of residual 2,4,6-TCP is critical due to its potential health and environmental risks.

To address this, decomposition tests using CoNCAs were conducted with 2,4,6-TCP as an analyte. As shown in Fig. 7f, under conditions of 20 mg per L 2,4,6-TCP, 50% degradation was achieved within 6 min, and 80% degradation was observed after 30 min. The kinetic rate constant for 2,4,6-TCP degradation was determined to be 0.046 min⁻¹ (Fig. S10†) with a high correlation coefficient (R² = 0.91). This performance is comparable to that of electrochemical oxidation processes employing graphene and tin oxide (SnO₂) electrodes, as well as systems utilizing cuprous oxide and graphite phase nitrogen carbide (CuO@g-C₃N₄), and visible light-based degradation with graphitic carbon nitride (g-C₃N₄), which achieved 20%, 63%, and 80% degradation after 30 min, respectively.^33,60,61 The degradation mechanism of organic pollutants in the CoNCA/PMS system involves both radical and non-radical oxidation pathways, with their relative contributions varying based on the nature of the target compound and the structural properties of the catalyst. As illustrated in Fig. 7g, PMS is activated at Co–N coordination sites and within graphitic N-rich domains of the carbon framework, leading to the generation of multiple reactive oxygen species (ROS), including sulfate radicals (SO₄˙⁻), hydroxyl radicals (˙OH), superoxide radicals (O₂˙⁻), and singlet oxygen (¹O₂).^62,63 In the radical pathway, PMS is activated via electron transfer to form SO₄˙⁻ through homolytic cleavage of the O–O bond in PMS. This process is facilitated by the Co²⁺/Co³⁺ redox cycle at metallic Co or Co–N sites, where Co acts as an electron mediator. The generated SO₄˙⁻ can subsequently react with water molecules to produce ˙OH, contributing to additional oxidation potential. The non-radical pathway, primarily driven by ¹O₂, is initiated by the decomposition of PMS through sequential one-electron redox transitions involving Co centers and O₂˙⁻ intermediates.^64,65 The presence of positively charged carbon atoms adjacent to Co–N sites facilitates PMS adsorption and electron exchange, thereby promoting the generation of ¹O₂.⁶⁶ This pathway is crucial for the degradation of MB and 2,4,6-TCP, as evidenced by the significant inhibition observed upon the addition of furfuryl alcohol (Fig. 6). Furthermore, the high degree of graphitization of CoNCAs, achieved through Co-catalyzed carbon restructuring during pyrolysis, provides an extended π-conjugated system that enhances electron mobility across the carbon matrix. Combined with the mesoporous structure derived from agarose templating, this facilitates the rapid diffusion of PMS and pollutants, promoting catalytic contact and ROS generation. These integrated features, chemical functionality (Co–N and N-doping), electronic conductivity, and pore accessibility, collectively enable the CoNCA system to achieve efficient degradation of a wide range of pollutants through the synergistic activation of both radical and non-radical oxidative pathways.

4. Conclusions

In this study, we developed a facile, sustainable, and shape-controllable CoNCA catalyst through a pyrolysis process, which demonstrated high performance for SR-AOPs. The porous structure of AG effectively immobilized ZIF-67, retaining its integrity even after pyrolysis. The prepared CoNCA successfully activated PMS and exhibited outstanding catalytic performance, achieving efficient degradation of various organic pollutants. Catalytic performance under different conditions was thoroughly evaluated, and radical scavenger tests revealed that nonradicals play a critical role in the degradation of MB. Additionally, the CoNCA could be easily collected using a magnet, and it maintained excellent MB degradation performance in a filtration system after grinding, highlighting its potential as a filler material. Furthermore, the CoNCA achieved 80% degradation of 2,4,6-TCP within 30 min, underscoring its versatility for removing organic pollutants. This study provides a pathway for the sustainable development of carbon-based aerogels and proposes novel metal-free carbon catalysts for water remediation applications.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Author contributions

Minsoo Yoon: writing – original draft, investigation, formal analysis, data curation, conceptualization. Hyunuk Jeon: formal analysis, investigation, data curation. Jisoo Park: methodology, formal analysis, data curation. Jieun Jang: methodology, formal analysis, data curation. Hojoon Choi: formal analysis, data curation. Jinbo Kim: formal analysis, data curation. Donggyun Kim: formal analysis, data curation. Kyubin Shim: formal analysis, data curation. Teahoon Park: supervision, investigation, data curation, conceptualization. Goomin Kwon: writing – original draft, review & editing, supervision, investigation, data curation, conceptualization. Jeonghun Kim: writing – review & editing, supervision, investigation, funding acquisition, data curation.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This research was supported by the Principal Research Program PNK8740 in the Korea Institute of Materials Science (KIMS) and supported (in part) by the Yonsei University Research Fund (Post Doc. Researcher Supporting Program) of 2024 (project no. 2024-12-0024).

References

1. J. B. Aguiar, A. M. Martins, C. Almeida, H. M. Ribeiro and J. Marto, Sustain. Prod. Consum., 2022, 32, 35–51.
2. X. Zhang and F. Dong, Resour., Conserv. Recycl., 2021, 170, 105616.
3. D. P. Häder, A. T. Banaszak, V. E. Villafañe, M. A. Narvarte, R. A. González and E. W. Helbling, Sci. Total Environ., 2020, 713, 136586.
4. C. Zamora-Ledezma, D. Negrete-Bolagay, F. Figueroa, E. Zamora-Ledezma, M. Ni, F. Alexis and V. H. Guerrero, Environ. Technol. Innovation, 2021, 22, 101504.
5. H. Musarurwa and N. T. Tavengwa, Carbohydr. Polym., 2020, 237, 116142.
6. H. Sehil, M. Badaoui, A. Chougui, L. Bouhadjar, W. Aoun and L. Belkhier, Macromol. Res., 2024, 32, 1–12.
7. J. Wang and S. Wang, Chem. Eng. J., 2021, 411, 128392.
8. S. O. Ganiyu, S. Sable and M. G. El-Din, Chem. Eng. J., 2022, 429, 132492.
9. X. Yang, F. L. Rosario-Ortiz, Y. Lei, Y. Pan, X. Lei and P. Westerhoff, Environ. Sci. Technol., 2022, 56(16), 11111–11131.
10. S. Giannakis, K. Y. A. Lin and F. Ghanbari, Chem. Eng. J., 2021, 406, 127083.
11. J. Scaria and P. V. Nidheesh, Curr. Opin. Chem. Eng., 2022, 36, 100830.
12. C. Wang, S. Ren, T. Cao, J. Wang, R. Peng, Z. Lv, X. Zhu, Y. Song, J. Na and Y. Mao, Chem. Eng. J., 2024, 496, 154352.
13. Z. Zhao, S. Zhu, S. Qi, T. Zhou, Y. Yang, F. Wang, Q. Han, W. Dong, H. Wang and F. Sun, Biochar, 2025, 7(1), 24.
14. L. Su, K. Chen, Y. Cai, T. Sheng, S. Chen, H. Xiang, Y. Deng and C. Tan, J. Hazard. Mater., 2024, 461, 132670.
15. W. Tian, S. Chen, H. Zhang, H. Wang and S. Wang, Curr. Opin. Chem. Eng., 2022, 37, 100838.
16. P. Liu, Z. Wu, A. V. Abramova and G. Cravotto, Ultrason. Sonochem., 2021, 74, 105566.
17. X. Xia, F. Zhu, J. Li, H. Yang, L. Wei, Q. Li, J. Jiang, G. Zhang and Q. Zhao, Front. Chem., 2020, 8, 592056.
18. W. Zhu, M. Han, D. Kim, Y. Zhang, G. Kwon, J. You, C. Jia and J. Kim, Environ. Res., 2022, 205, 112417.
19. K. Li, S. Xu, X. Liu, H. Li, S. Zhan, S. Ma, Y. Huang, S. Liu and X. Zhuang, Chem. Eng. J., 2022, 438, 135630.
20. J. Xin, F. Zhang, S. Liu, Y. Liu, C. Han, X. Li, C. Shao, W. Li and Y. Liu, Chem. Eng. J., 2023, 455, 140774.
21. Z. Gao, J. Zhu, Q. Zhu, C. Wang and Y. Cao, Sci. Total Environ., 2022, 847, 157405.
22. Q. Huang, Q. Zhang, S. Zhao, C. Zhang, H. Guan and J. Liu, Biochar, 2025, 7(1), 1–17.
23. V. G. Oyervides-Guajardo, J. A. Claudio-Rizo, D. A. Cabrera-Munguía, M. Caldera-Villalobos, T. E. Flores-Guia, F. Soriano-Corral and A. Herrera-Guerrero, Macromol. Res., 2024, 32, 1–16.
24. K. Y. A. Lin and H. A. Chang, J. Taiwan Inst. Chem. Eng., 2015, 53, 40–45.
25. C. Wang, J. Kim, V. Malgras, J. Na, J. Lin, J. You, M. Zhang, J. Li and Y. Yamauchi, Small, 2019, 15(16), 1900744.
26. H. Li, J. Tian, Z. Zhu, F. Cui, Y. A. Zhu, X. Duan and S. Wang, Chem. Eng. J., 2018, 354, 507–516.
27. P. He, H. Lan, H. Bai, Y. Zhu, Z. Fan, J. Liu, L. Liu, R. Niu, Z. Dong and J. Gong, Appl. Catal., B, 2023, 337, 123001.
28. H. Zhang, J. Yang, Z. Sun, Y. Sun, G. Liu, D. Lu and J. Ma, Water Res., 2025, 268, 122783.
29. X. H. Yi, H. Ji, C. C. Wang, Y. Li, Y. H. Li, C. Zhao, A. Wang, H. Fu, P. Wang, X. Zhao and W. Liu, Appl. Catal., B, 2021, 293, 120229.
30. X. Li, J. Chen, Z. Liu, C. He, J. Pang, L. Zhang, F. Tang and X. Yang, Chem. Eng. J., 2024, 499, 156081.
31. N. Zion, D. A. Cullen, P. Zelenay and L. Elbaz, Angew. Chem., Int. Ed., 2020, 59(6), 2483–2489.
32. L. E. Nita, A. Ghilan, A. G. Rusu, I. Neamtu and A. P. Chiriac, Pharmaceuticals, 2020, 12(5), 449.
33. S. Wang, C. Yuan, W. Chen, Y. Niu, Y. Yan, F. Li and H. Jiang, Chem. Eng. J., 2024, 480, 148050.
34. R. M. Rego, G. Kuriya, M. D. Kurkuri and M. Kigga, J. Hazard. Mater., 2021, 403, 123605.
35. S. Nitta, S. Taniguchi and H. Iwamoto, Macromol. Res., 2024, 32, 1–13.
36. C. Moreno-Castilla and F. J. Maldonado-Hódar, Carbon, 2005, 43(3), 455–465.
37. M. Yoon, J. Park, J. Jang, H. Choi, H. Jeon and J. Kim, Carbohydr. Polym., 2024, 345, 122559.
38. Y. M. Li, Z. Y. Liu, Q. Y. Zhang, Y. J. Wang, G. Q. Cui, Z. Zhao, C. M. Xu and G. Y. Jiang, Pet. Sci., 2023, 20(1), 559–568.
39. K. H. Adolfsson, P. Huang, M. Golda-Cepa, H. Xu, A. Kotarba and M. Hakkarainen, Adv. Sustainable Syst., 2023, 7(3), 2200425.
40. Y. Suo, Z. Zhang, Z. Zhang and G. Hu, Ionics, 2021, 27, 289–303.
41. J. Tang, R. R. Salunkhe, H. Zhang, V. Malgras, T. Ahamad, S. M. Alshehri, N. Kobayashi, S. Tominaka, Y. Ide, J. H. Kim and Y. Yamauchi, Sci. Rep., 2016, 6(1), 30295.
42. S. Liang, H. Y. Niu, H. Guo, C. G. Niu, C. Liang, J. S. Li, N. Tang, L. S. Lin and C. W. Zheng, Chem. Eng. J., 2021, 405, 126686.
43. W. Peng, X. Yang, L. Mao, J. Jin, S. Yang, J. Zhang and G. Li, Chem. Eng. J., 2021, 407, 127157.
44. X. R. Wang, J. Y. Liu, Z. W. Liu, W. C. Wang, J. Luo, X. P. Han, X. W. Du, S. Z. Qiao and J. Yang, Adv. Mater., 2018, 30(23), 1800005.
45. G. Wu, C. M. Johnston, N. H. Mack, K. Artyushkova, M. Ferrandon, M. Nelson, J. S. Lezama-Pacheco, S. D. Conradson, K. L. More, D. J. Myers and P. Zelenay, J. Mater. Chem. A, 2011, 21(30), 11392–11405.
46. Y. Zhao, K. Watanabe and K. Hashimoto, J. Am. Chem. Soc., 2012, 134(48), 19528–19531.
47. G. Wang, X. Nie, X. Ji, X. Quan, S. Chen, H. Wang, H. Yu and X. Guo, Environ. Sci.: Nano, 2019, 6(2), 399–410.
48. J. Luo, S. Bo, Y. Qin, Q. An, Z. Xiao and S. Zhai, Chem. Eng. J., 2020, 395, 125063.
49. D. T. Oyekunle, B. Wu, F. Luo, J. Ali and Z. Chen, Chem. Eng. J., 2021, 421, 129818.
50. H. Shao, X. Zhao, Y. Wang, R. Mao, Y. Wang, M. Qiao, S. Zhao and Y. Zhu, Appl. Catal., B, 2017, 218, 810–818.
51. Q. Zhang, C. Li, Z. Li, N. Wang, X. Chen, C. Zhang, J. Xing, H. Qi and Q. Xing, J. Mater. Sci., 2024, 59(5), 1877–1895.
52. W. Zhu, D. Kim, M. Han, J. Jang, H. Choi, G. Kwon, J. Jeon, D. Y. Ryu, S. H. Lim, J. You, S. Li and J. Kim, Chem. Eng. J., 2023, 460, 141593.
53. J. He, Y. Wan and W. Zhou, J. Hazard. Mater., 2021, 405, 124199.
54. D. Dai, Z. Yang, Y. Yao, L. Chen, G. Jia and L. Luo, Catal. Sci. Technol., 2017, 7(4), 934–942.
55. Y. Long, S. Li, P. Yang, X. Chen, W. Liu, X. Zhan, C. Xue, D. Liu and W. Huang, Sep. Purif. Technol., 2022, 286, 120470.
56. C. Dong, Q. Yi, J. He, M. Xing and J. Zhang, EES Catal., 2023, 1(2), 103–116.
57. Y. Guo, L. Yan, X. Li, T. Yan, W. Song, T. Hou, C. Tong, J. Mu and M. Xu, Sci. Total Environ., 2021, 783, 147102.
58. X. Chen, X. A. Ning, X. Lai, Y. Wang, Y. Zhang and Y. He, J. Hazard. Mater., 2021, 416, 125721.
59. Y. Li, Y. Li, B. Jin, K. Zhang, L. Wang and J. Zhao, Bioresour. Technol., 2021, 323, 124627.
60. X. Duan, X. Sui, S. Tu, Z. Ning, Y. Li, L. Chang and P. Nie, J. Environ. Chem. Eng., 2023, 11(1), 109078.
61. H. Ji, F. Chang, X. Hu, W. Qin and J. Shen, Chem. Eng. J., 2013, 218, 183–190.
62. H. Dai, W. Zhou and W. Wang, Chem. Eng. J., 2021, 417, 127921.
63. X. Li, X. Huang, S. Xi, S. Miao, J. Ding, W. Cai, S. Liu, X. Yang, H. Yang, J. Gao, J. Wang, Y. Huang, T. Zhang and B. Liu, J. Am. Chem. Soc., 2018, 140(39), 12469–12475.
64. S. Zhu, X. Li, J. Kang, X. Duan and S. Wang, Environ. Sci. Technol., 2018, 53(1), 307–315.
65. Z. Yang, J. Qian, A. Yu and B. Pan, Proc. Natl. Acad. Sci. U. S. A., 2019, 116(14), 6659–6664.
66. X. Duan, Z. Ao, H. Sun, L. Zhou, G. Wang and S. Wang, Chem. Commun., 2015, 51(83), 15249–15252.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta03143e

‡ These authors contributed equally.

---