Single-atom catalysts activate persulfate to degrade emerging organic contaminants in aqueous environments Open Access

In recent years, numerous emerging organic contaminants (EOCs) such as 1,4-dioxane (Tang et al. 2023a), per- and polyfluoroalkyl substances (Han et al. 2023a), pharmaceuticals and personal care products (Xia et al. 2023), as well as analgesics, antibiotics, and cytotoxic drugs in pharmaceutical wastewater (Wang et al. 2023b), due to their hydrophilic nature, low biodegradability, and high persistence, tend to persist in wastewater for extended periods (El Kateb et al. 2022). For instance, perfluorooctanoic acid (PFOA), with a half-life of 92 years, is carcinogenic and reproductive toxic, posing a significant threat to human health (Yu et al. 2024). Oxytetracycline (OTC) exhibits toxicity to bacteria, leading to antibiotic resistance that complicates eradication efforts and alters microbial community structures (Feng et al. 2022). Prolonged exposure to these pollutants in natural environments not only poses serious risks to human health but also disrupts ecological balance. Traditional wastewater treatment methods – physical, chemical, and biological – are increasingly inadequate for efficiently and environmentally safely treating wastewater containing EOCs like dyes (Elwakeel et al. 2020; Mumtaz et al. 2022). Persulfate-based advanced oxidation processes (PS-AOPs) are considered a promising wastewater purification technology due to their low cost, high redox potential, efficient pollutant removal, and wide pH applicability (Qiu et al. 2022). The heterogeneous catalysts in AOPs are primarily composed of metal particles and a substrate, with catalytic reactions typically occurring at surface active sites rather than predominantly within the interior active sites, resulting in lower atomic utilization efficiency. Reducing particle size to increase the effective active sites is an effective approach to enhancing mass transfer and charge transfer in AOPs (Xia et al. 2023).

Compared to other metal-loaded catalysts on biochar, single-atom catalysts (SACs) are a type of heterogeneous catalyst with relatively small particle size, possessing characteristics such as high reactivity, chemical stability, and environmental friendliness (Han et al. 2023a). They exhibit roles in local structure and catalytic performance similar to homogeneous catalysts, combining the advantages of both heterogeneous and homogeneous catalysts (Xia et al. 2023). Qiao et al. (2011) first investigated SACs in 2011, referring to catalysts where individual metal atoms are uniformly dispersed on a substrate. Due to the highly dispersed nature of the metal centers in SACs, they can form more uniform and specific active sites, allowing for the modulation of their activity and selectivity by altering the configuration around each individual atom. Meanwhile, the maximized atomic utilization (Kaur et al. 2024), unique electronic properties (Zhu et al. 2022), and special size-dependent quantum effects (Lee et al. 2023) of SACs endow them with broad application prospects.

The top-down approaches mostly involve high-temperature steps, resulting in SACs that exhibit excellent stability. Among these, the pyrolysis method is the most common, which is widely applicable and often used for the preparation of transition metal SACs (Huang et al. 2023).

The pyrolysis method also exhibits notable shortcomings, as elevated temperatures can induce easy migration of metal atoms, leading to the formation of metal clusters and NPs. This represents a widespread challenge that pyrolysis techniques typically struggle to circumvent (Li et al. 2021b).

In practical applications, pyrolysis methods also face significant challenges, particularly regarding secondary pollution issues. Temperature is a critical factor in pyrolysis, determining the quantity of ROS and affecting its direct electron transfer capability. However, SACs synthesized at higher temperatures may not necessarily be the most suitable materials. Despite their excellent catalytic activity and resistance to interference prepared at higher pyrolysis temperatures, they may also bring about serious metal leaching issues, which is a common problem and difficult to avoid in SACs. Wang & Wang (2023) prepared respective single-atom cobalt catalysts (Co–C–X) at 500, 600, and 700 °C. While Co-C-700 exhibited the highest catalytic activity and metal loading, it also showed the highest cobalt leaching (3.23 mg/L, compared to 0.14 mg/L for Co-C-600). Therefore, catalysts synthesized at 700 °C have the greatest environmental impact, resulting in severe heavy metal pollution. Hence, when synthesizing SACs via pyrolysis, it is necessary to adjust parameters such as temperature appropriately and carefully consider the balance between catalyst high performance and factors like secondary pollution and stability to achieve more balanced catalyst performance.

The bottom-up approach typically involves depositing a small amount of metal atoms on a carrier, mainly through deposition methods (atomic layer deposition (ALD), chemical vapor deposition, and electrochemical deposition), wet chemical methods (co-precipitation, impregnation, photochemical method), ball milling, sacrificial template method, etc.

The ALD technique enables precise control at the atomic level over the size and thickness of deposited materials, thereby facilitating the creation of uniformly distributed single-atom thin films. This capability endows SACs with exceptionally high catalytic activity (Cheng & Sun 2017). ALD mainly consists of four steps: precursor pulse, purge, oxidant pulse, and post-purge. The number of cycles in ALD determines the thickness and quality of the atomic layers. Within a certain range, more cycles result in a higher metal loading, which is advantageous for enhancing the catalytic performance of SACs (Fonseca & Lu 2021). Figures 3(b) and 3(c) illustrate the synthesis by Wang et al. (2020b) of high iron loading (>1.5 wt%) Fe SACs via ALD on different substrates (MWCNTs and TiO₂). After 15 cycles of Fe ALD, the Fe content in Fe/TiO₂ particles linearly increased with the ALD cycle number, and after 25 cycles of Fe ALD, no Fe NPs were observed on SiO₂ NPs, leading to an increase in active Fe sites and higher catalytic activity toward CO oxidation. The structural characterization of different catalysts is illustrated in Figure 4(c) and 4(d).

While increasing the cycling frequency benefits the catalytic performance of SACs, excessive cycles may transform single atoms into metal clusters or NPs, reducing the number of active sites on SACs. This limitation hinders further enhancement of metal loading. Therefore, developing SACs with higher loading capacity and better stability is currently a pressing issue to address (Cheng & Sun 2017).

Ball milling utilizes a ball mill to apply strong shear forces and localized high temperatures to the carrier, generating defect active sites. These active sites can capture metal atoms, leading to the formation of SACs (De Bellis et al. 2022). Wang et al. (2021) employed ball milling to prepare a high-performance catalyst (Co-SNC) with dual active sites, demonstrating significantly enhanced catalytic performance, accelerating amine adsorption, and O₂ activation. In the benzylamine coupling reaction, Co-SNC exhibited the highest conversion rate of 97.5% within 10 h, with a selectivity of 99% toward N-benzylbenzylamine. Tang et al. (2023b) successfully prepared single-atom Ni catalysts for photocatalytic CO₂ reduction using high-energy ball milling. This method ensures that highly dispersed Ni atoms are uniformly distributed on the carrier surface, compensating for deficiencies in other synthesis methods regarding single-atom dispersion and uniformity. These dispersed Ni single atoms significantly improve the photocatalyst's CO₂ adsorption capacity, reducing the activation energy barrier and facilitating the photocatalytic conversion of CO₂ to CO. Yu et al. (2022) synthesized single-atom Co-OH modified polymeric carbon nitride (Co-polymeric carbon nitride (PCN)) via ball milling assistance. With ball milling assistance, the Co content increased by 37 times, and Co-PCN exhibited significant enhancement in oxygen evolution reaction performance, with the highest rate reaching 37.3 μ mol h⁻¹, 28 times higher than that of common PCN/CoO_x.

Some drawbacks still exist in the ball milling method, such as requiring a long time to disperse single atoms, resulting in lower milling efficiency (Hu et al. 2023b).

The co-precipitation method is characterized by its simplicity, low cost, short synthesis time, high purity, and uniform elemental doping (Mansooripour et al. 2024). Qiao et al. (2014) dispersed Pt single atoms on iron oxides using the co-precipitation method and then calcined them at high temperatures. The highly dispersed Pt single atoms improved the atomic efficiency of Pt metal, showing excellent activity and selectivity for the preferential oxidation of CO in H₂-rich gas, completely removing CO from the gas in a wide temperature range of 20–70 °C. Using the co-precipitation method for single-atom modification not only significantly reduces the preparation cost of the catalyst but also improves the atomic utilization efficiency of precious metals. Sun et al. (2022) synthesized Ru₁/FeO_x SACs via co-precipitation, achieving high CO conversion rates with a low loading of 0.18 wt%, significantly higher than the turnover rates of Ru NPs and most Ru-based catalysts by several orders of magnitude. Shi et al. (2023) used a one-step co-precipitation method to prepare a single-atom Pt-CeO₂/Co₃O₄ catalyst with an ultra-low Pt loading capacity of 0.06 wt%. This catalyst exhibited catalytic performance comparable to that of the nanoscale Pt-loaded catalyst 0.41Pt-nanoparticles (NP) (with a loading ratio of 0.06Pt-single atom (SA) more than six times higher), showing ultra-long durability and excellent toluene degradation capability.

Co-precipitation often struggles to co-precipitate multiple cations, while the surface properties and activity of the catalyst are not sufficiently stable. It can be significantly influenced by various preparation conditions, such as pH and temperature (Sim et al. 2019).

The impregnation method, known for its straightforward preparation process, is commonly used to deposit single-metal atoms onto supports through ion exchange or adsorption processes. Effective interaction between the support and single atoms is crucial (Swain et al. 2022). Zhang et al. (2019) employed a simple impregnation-adsorption method to construct Pt SACs on N-doped carbon nanocages. The synergistic effect of micropore capture and N anchoring facilitates the capture of and derived Pt single atoms, resulting in highly stable Pt catalysts with unprecedented electrocatalytic hydrogen evolution performance far superior to conventional Pt-based catalysts. Chen et al. (2023b) prepared vacancy-rich nickel selenide-supported Pt SACs using a hydrothermal impregnation stepwise method. The formation of Pt–Se bonds, due to the combination of Pt atoms with highly electronegative selenium, acts as a ‘bridge’ for rapid electron transfer between single atoms and the substrate. This novel catalyst exhibits an extremely low overpotential of 45 mV at 10 mA cm⁻² and outstanding stability over 120 h. Gu et al. (2023) prepared Ru₁/CoO_x catalysts via impregnation, which, compared to pristine CoO_x, significantly enhanced the selectivity (13.8%) and yield (13.3%) of 2,5-furandicarboxylic acid.

However, there are still some drawbacks to the current impregnation method, such as the low metal loading on SACs. As the catalyst's surface area increases and its surface energy rises, metal atoms tend to aggregate into particles, which can diminish the catalytic performance of SACs (Zhang et al. 2014).

Bottom-up synthesis methods generally yield lower metal loadings in SACs, which restricts the number of available active sites compared to top-down methods. Therefore, when attempting to increase metal loading to enhance active sites using these methods, the low surface energy of single-atom surfaces often leads to the aggregation of metal atoms, forming NPs (Qin et al. 2024). This aggregation significantly reduces the number of active metal sites, resulting in a substantial decline in catalyst performance. Thus, balancing the enhancement of metal active sites while preventing metal atom aggregation poses a paradox and challenge. In addition, increasing metal loading can exacerbate issues such as metal leaching, necessitating a comprehensive approach to finding methods that effectively enhance the catalytic activity of SACs while preventing NPs formation and secondary pollution.

By modulating the microenvironment of SACs metal centers, their catalytic performance can be altered. These factors include the type of metal center, metal content, the influence of dopant atoms on coordination, coordination number, and the type of carrier, as detailed in Table 2, regarding their impact on the catalytic performance of SACs. Concurrently, we summarized key considerations for practical applications.

The type of active metal center directly influences the coordination structure, thereby affecting catalytic performance. Therefore, selecting the appropriate metal center is crucial for degrading organic compounds. Some noble metals, such as Ag (Wang et al. 2022b) and Ru (Yan et al. 2022), possess high conductivity and excellent catalytic reduction ability, making them more inclined to activate PMS without involving the generation of radicals. However, due to the rarity and high cost of noble metals, research on them is relatively limited. Yan et al. (2022) synthesized Ru-doped N-doped carbon (CNRu), which effectively activates PMS, and degrades and detoxifies dichlorofenac acid (DCF) through a non-radical pathway. The effective Ru–N coupling not only exposes more active sites but also enhances electron transfer, accelerating the activation of PMS. Within a wide pH range (3.0–9.0), DCF can be completely removed in just 10 min.

Transition metals such as Fe (Chen et al. 2022), Mn (Yin et al. 2023), Co (Pang et al. 2024), W (Gu et al. 2022), etc., exhibit excellent catalytic performance in sulfate radical-based AOPs (SR-AOPs) due to their low cost and the presence of active sites. Among them, Fe-based SACs have the significant advantages of low toxicity and high activity, making them more commonly used SACs (Yang et al. 2023b). Zeng et al. (2023) synthesized FeSA-NC, which contains dispersed Fe–N₄ active sites, through a strategy of micropore confinement. The single-atom Fe metal sites on the carrier provide uniformly distributed active centers for the activation reaction. Chemical adsorption and electron transfer between PMS and FeSA-NC were achieved through an internal electron shuttle mechanism, where Fe–N₄ acts as a conductive bridge. Therefore, the FeSA-NC/PMS system can completely remove bisphenol A (BPA) within 2.5 min, exhibiting a significantly high rate constant (k_obs = 2.373 min⁻¹). This result far exceeds the findings of Wang et al. (2017a), who utilized Fe–TiO₂ for photocatalytic degradation of BPA, achieving a 92.30% degradation rate in 180 min under visible light irradiation.

The content of metal centers is usually closely related to the active sites. Within a certain range, as the metal loading increases, so does the number of active sites. However, excessively high metal atom loading often leads to the clustering of metal atoms on the carrier, resulting in a decrease in active sites and consequently a decline in the catalytic performance of SACs (Cheng et al. 2021). Therefore, the preparation of metal SACs with high loading has become a hot topic in current research. Yin et al. (2023) used lignin as a raw material to prepare high Mn-loaded Mn SACs (Mn-SAC) through pyrolysis. The addition of Mn-SAC significantly improves the degradation efficiency of pollutants, with the k_obs (reaction rate constant) being two to three orders of magnitude higher than that of using PMS alone, Mn²⁺/PMS, MnO/PMS, and other catalytic systems. The Mn content reaches 4.07 wt%, indicating that Mn-SAC catalysts with a high Mn atomic content have strong catalytic capabilities for oxidizing pollutants. However, breakthroughs in achieving high metal center (>10 wt%) content still face challenges. Gu et al. (2022) obtained W-SAC (W-CN) with a high W loading of 11.16 wt% and unique O, N coordination through a thermal polymerization process to fix W atoms. When the W content is appropriately increased, SACs exhibit better photocatalytic activity. However, further increasing the W loading leads to a decrease in degradation activity, possibly due to the shielding effect of excess loaded W atoms. Despite the lower metal loading of SACs, their catalytic activity at active sites surpasses that of other single-metal catalysts. Guo et al. (2024a) synthesized Co-NC and CoSAC with metal loadings of 13.95 and 1.47%, respectively. The k_{per M} values for Co-NC and CoSAC were 1,688.9 and 5,487.2 min⁻¹ mol⁻¹, respectively, demonstrating that the catalytic activity of SACs often depends more on the increase in active sites rather than metal loading.

The catalytic activity of SACs is often closely linked to the local electron configuration of the metal center. In the M-N₄ (M represents metal) configuration, the electronegativity of N atoms often limits the electron transfer between the central metal atom and PMS (Dai et al. 2024). To enhance catalytic performance, doping inorganic nonmetal atoms (such as B, O, P, S, etc.) into SACs can lead to changes in material charge density, thereby altering the electron structure (Zhang et al. 2021). Li et al. (2023b) doped N, P, and S atoms into Fe SACs, and density functional theory calculations showed that the doping of heteroatoms effectively changed the electronic structure of the catalyst (FeSA-NPS@C) in Fe–N₄, enhancing its coordination with PMS, promoting electron transfer to PMS, and facilitating the interaction between Fe^VN₄ = O and ofloxacin (OFX), leading to nearly 100% degradation of OFX within 3 min. Wu et al. (2023a) synthesized Co SACs with N/O dual coordination by doping O atoms, where the doping of O atoms optimized the electron distribution of Co 3d orbitals, reduced the electron density of the Co center, enhanced PMS adsorption at Co single-atom sites, lowered the energy barrier for the formation of key intermediates during the Co(IV) = O process, and achieved efficient pollutant removal. Zou et al. (2022) introduced P atoms into Co SACs, concentrating electron density and electron delocalization near the Co center, facilitating electron transfer and thereby activating PMS to produce ⁠, achieving a 96.9% degradation of sulfadiazine (SDZ) within 10 min. Meanwhile, Zheng et al. (2023) just achieved an 86.61% degradation of SDZ within just 30 min using heat-assisted Co–Bi₂₅FeO₄₀/PS, and with microwave-assisted Co-Bi₂₅FeO₄₀/PS, the degradation efficiency of SDZ only reached 94.40%. Therefore, it is evident that heteroatom doping plays a crucial role in enhancing SACs performance.

The coordination number is also an important parameter that affects the catalytic performance of SACs as it has a significant influence on the geometry and electronic structure of the metal center (Zhang et al. 2021). In the case of the M-N₄ configuration, each metal atom is anchored by four N atoms. This pattern gives SACs a unique geometric structure and uniform active centers, which are advantageous for activating PMS and facilitating the attachment of organic pollutants (Du et al. 2022). On the other hand, low-coordination M–N_x (x < 4) configurations exhibit better catalytic performance in certain areas and possess excellent stability. This is because the low-coordination configuration enhances the electronic interaction between the single-atom metal and PS molecules, thereby promoting the activation of PMS (Wang et al. 2023a). Liang et al. (2022) synthesized two different single-atom cobalt catalysts (CoSA–N_x–C) with varying N coordination numbers. The results showed that decreasing the Co–N coordination number from 4 to 3 increased the electron density of the individual Co atoms. In addition, the CoSA–N₃–C configuration was more favorable for PDS adsorption and the generation of active radicals, thereby facilitating pollutant degradation.

To prevent single-metal atoms from aggregating into NPs and to endow the metal atoms with unique electronic structures, selecting appropriate catalyst supports is crucial. Based on spatial structure, the carriers of SACs can be divided into two categories: 3D and 2D. 3D carriers include carbon materials, metal oxides, and metal-organic frameworks (MOFs), while 2D carriers include graphene, g-C₃N₄, and MoS₂ (Huang et al. 2020). Carbon materials are widely used due to their high stability and cost-effectiveness. In a dual-carbon background, synthesizing carbon materials from biochar not only achieves sustainable resource utilization but also helps with carbon reduction (Tian et al. 2024). Introducing N atoms into them can increase the catalyst's specific surface area, enhance active sites, and reduce the overflow of metal ions. Therefore, SACs loaded on N-doped carbon materials usually exhibit excellent catalytic performance (Jing et al. 2023). Polymeric nated carbon (CN) is an N-rich carbon material. Cui et al. (2023) synthesized Fe SAs loaded on CN in a single step. CN reduces the electron density of Fe sites, facilitating the transfer of high-valent Fe (IV) = O to organic pollutants, thereby rapidly degrading BPA. Other carrier materials, such as MOFs, are currently widely used. Composed of metal atoms and organic ligands, MOFs have extremely high specific surface areas, which are conducive to dispersing metal single atoms Yang et al. (2023a). ZIF-8, a type of MOF, has high surface area and porosity, as well as high carbon and N content, making it suitable for synthesizing various transition metal SACs (Wang et al. 2020a).

However, utilizing these carriers also presents challenges. For instance, g-C₃N₄ is prone to decomposition and curling at high temperatures, which may lead to the formation of metal atom clusters. To address these issues, Liu et al. (2023a) employed a pre-coordinated metal precursor (hematin chloride) to synthesize Fe SAC on g-C₃N₄. This approach prevented Fe atom aggregation and minimized g-C₃N₄ curling at high temperatures, due to hematin chloride's planar structure. In summary, choosing appropriate carriers to synthesize SACs with optimal catalytic performance necessitates careful consideration of carrier characteristics while mitigating the risk of secondary pollution.

Various conditions in real environments can potentially limit the application of SACs for activating persulfate. Consideration is necessary for their application in different water bodies, which contain numerous interfering substances such as various heavy metal ions, inorganic anions, and natural organic matter (NOM). These substances can inhibit the catalytic activity of SACs and reduce the efficiency of target pollutant degradation.

Researchers like Chai et al. (2024) utilized Fe-P, N-codoped carbon (PNC)/PMS systems to degrade BPA in tap water, sludge water, and aquaculture water. Results showed that in sludge water and aquaculture wastewater, the degradation efficiency of BPA decreased from around 100% to less than approximately 75 and 55%, respectively. This underscores significant reductions in removal rates when real water matrices contain high concentrations of interfering components. Regarding NOM interference, Guo et al. (2024b) utilized Mn-SAC/PMS systems to degrade nitenpyram (NPR) in the presence of NOM at 50 mg/L. The removal efficiency of NPR dropped from 100 to about 80%, primarily due to NOM consuming radicals, which hinders NPR's complete oxidation. Furthermore, reports from Guo et al. (2024c) indicate that NOM and anions in river and pond water can reduce the removal efficiency of BPA in Mo-NC-0.1/PMS systems. Within these water bodies, NOM components like humic acid compete with BPA for ROS, while competes for adsorption sites with BPA, altering the pH and affecting the catalyst's adsorption capability, thereby inhibiting BPA degradation. Therefore, various interfering components in real water bodies can differently inhibit the degradation of EOCs, especially with more complex backgrounds and higher concentrations, where this inhibition is more pronounced. It is crucial to adjust the parameters of catalytic oxidation systems appropriately to better adapt to real-world applications.

SACs/PS systems raise significant sustainability concerns. Most studies on SACs activating PS demonstrate excellent sustainability and catalytic activity. For instance, Zhang et al. (2023d) synthesized Co-SNC catalysts, which, after five cycles of activation with PMS, still removed 96.8% of carbamazepine (CBZ) within 60 min, with Co leaching only at 0.03 mg/L. In contrast, dual-metal catalysts (P–Fe/Co/N@BC) synthesized by Yu et al. (2024) showed a PFOA degradation rate of 93% after five cycles, accompanied by higher metal leaching (about 0.25 mg/L for Co). Moreover, Wang et al. (2022c) synthesized FeCoN₅P₁/C catalysts for ORR, whose performance rapidly degraded within 15 h and continued to decline over the next 20 h. According to An et al. (2024), after SACs deactivation, catalysts can be rejuvenated through heat treatment. Therefore, thermal treatment can restore catalytic activity after extensive cycles of SACs usage, thereby enhancing sustainability and practical value. Therefore, the SACs activating PS technique exhibits better sustainability and stability compared to other AOPs. However, some SACs show less optimistic stability, as observed by Gu et al. (2024) using single-atom cobalt catalysts (CoN4S-CB) for PMS activation, where after four cycles, the removal efficiency of sulfamethoxazole (SMX) was only 62.2%.

Substrate material plays a crucial role in sustainability in practical applications, necessitating careful consideration of stability and sustainability properties. Carbon-based SACs, as reported, can undergo detrimental reactions with strong oxidants during prolonged oxidation processes, risking damage to their structure and potential metal leaching or deactivation (Xu et al. 2020). Thus, selecting carbon materials with a larger adsorption surface area, robust pH stability, and superior physicochemical properties is crucial in enhancing the suitability, sustainability, and stability of SACs in real-world applications (Xiao et al. 2023).

The mechanisms of SACs activating PS for degrading organic pollutants cover adsorption mechanisms, ROS (radical pathways, non-radical pathways), and organic pollutant degradation mediated by ROS. The pathways through which SACs activate PS to produce ROS include radical pathways and non-radical pathways. The radical pathway mainly refers to a series of chain reactions initiated by SACs during the activation of PS, ultimately leading to the generation of radicals such as ⁠, ⁠, and ⁠. The non-radical pathway can generate ⁠, reactive high-valent metal oxides, electron transfer (e⁻), surface-bound radicals, and holes (H⁺), among others (Yang et al. 2022). In complex real aqueous environments, both pathways often coexist in complex AOP systems and demonstrate advantages in degrading organic pollutants (Zhang et al. 2024).

Adsorption, as the initial stage of the reaction process, plays a crucial role. According to Hu et al. (2023a), metal sites on SACs carry positive charges and adsorb negatively charged PMS through electrostatic interactions to immobilize PMS. Subsequently, spin-polarized PMS interacts magnetically with single-metal atoms, enhancing PMS adsorption. Under the influence of metal sites, PMS molecules break apart, producing various ROS. Concurrently, organic compounds are adsorbed onto SACs through weak molecular interactions. Finally, ROS degrade organic compounds by attacking specific chemical bonds and functional groups.

possesses strong oxidation capabilities, with its redox potential relative to the normal hydrogen electrode ranging between 2.5 and 3.1 V, thus effectively decomposing many recalcitrant substances (Tian et al. 2022). Compared to •OH, the half-life of is longer, at 30–40 μs, which facilitates its diffusion in water, thereby enhancing the mineralization and oxidation of organic compounds (Chen et al. 2024). Its generation occurs as metal centers, according to Equation (1), disrupt the O–O bond in PS molecules through electron transfer (Cheng et al. 2024). For instance, Zr–Co₃O₄ activates PMS, and after PMS is adsorbed onto the catalyst, Zr active sites induce electron disruption of the O–O bond in PMS, leading to the formation of ⁠, as shown in Equation (2) (Zhang et al. 2023a). The application of transition metals can effectively activate PMS to generate ⁠, as demonstrated by Wang et al. (2017b), who prepared single-atom dispersed Ag-modified mesoporous graphitic carbon nitride (Ag/mpg-C₃N₄) hybrid materials via co-condensation. In the system, the primary ROS are and •OH. When using a 0.1 g/L catalyst and 1 mM PMS, 100% of BPA and 80% of total organic carbon (TOC) can be removed within 60 min.

The oxidation capability of is dependent on the reaction conditions, exhibiting significant differences in oxidative performance under acidic and alkaline conditions. Under acidic conditions, PMS is activated and decomposed into radicals, while under alkaline conditions, excess OH⁻ ions scavenge radicals, generating less active •OH radicals, thereby weakening the catalytic performance of SACs (Chen et al. 2023a). For instance, Chen et al. (2023a) synthesized xFe–MoS₂, which exhibited excellent catalytic performance in degrading rhodamine B (Rh B) over a wide pH range (3.0–11.0), with and as the main radicals. Analysis indicates that acidic conditions favor the degradation of RhB, with the highest degradation efficiency achieved at pH = 3.0 after 4 min of reaction, reaching a maximum of 99.4%. As the pH of the solution increases from 6.0 to 11.0, the degradation efficiency gradually slows down.

tends to react with organic pollutants containing aromatic π-electrons, electron-rich functional groups, or unsaturated bonds via electron transfer mechanisms, exhibiting a higher selectivity toward them. During degradation, directly abstracts electrons from benzene rings or unsaturated bonds, thereby disrupting the pollutant's structure and facilitating its degradation (Lian et al. 2017; Tian et al. 2022).

•OH possesses a high redox potential (E₀ = 1.9–2.7 V_{normal hydrogen electrode (NHE)}) and can degrade many EOCs into low-toxicity small molecules or directly convert them into CO₂ and H₂O, so it serves as a green and clean oxidant; however, its half-life is extremely short (<1 μs), making it difficult to persist in aqueous environments for extended periods (Zhao et al. 2023c; Liu et al. 2024). The mechanism of •OH generation, as depicted in Equation (3), involves the reaction of metal ions with to yield high-valent metals and •OH. This reaction essentially occurs through two processes. Taking Co-W₁₈O₄₉ activation of PMS as an example, initially, the active Co sites in Co-W₁₈O₄₉ induce electron transfer to PMS, leading to the formation of ⁠, as shown in Equation (4). Subsequently, reacts with H₂O or OH⁻ to generate •OH, as illustrated in Equation (5) (Zhu et al. 2024). However, compared to ⁠, •OH has a shorter half-life of 1 μs, resulting in a shorter diffusion distance and less effective oxidation of organic pollutants compared to ⁠, with poorer selectivity. This limitation constrains the selective degradation of pollutants (Chen et al. 2024). To prolong the effective duration of •OH, Chen et al. (2024) developed single-atom Co (SA-Co) sites within layered double hydroxides (LDHs) to activate PMS. This approach stabilized negatively charged PMS with positively charged LDHs, enabling Co single-atom sites to selectively and sustainably generate surface-bound •OH and radicals. These radicals remained effective for up to 48 h, suppressing PMS decomposition and self-quenching, thereby achieving prolonged radical generation and efficient oxidation of organic compounds.

exhibits mild oxidation potential with an oxidation–reduction potential E⁰ = +2.2 V^NHE (Li et al. 2022), possessing advantages such as high selectivity, long half-life (2 μs), wide pH range applicability, and strong resistance to inorganic ion interference. Therefore, it significantly enhances the degradation rate of organic pollutants, playing a positive role in the complete degradation of recalcitrant pollutants (Yuan et al. 2024). Zhao et al. (2023a) developed a single-atom co-anchored N-doped graphene (CoSAC-NG) for the degradation of BPA. Among these, is the primary active species for pollutant degradation, exhibiting excellent catalytic performance in BPA degradation, with the pseudo-first-order reaction rate constant increasing from 8.13 × 10⁻⁴ min⁻¹ without a catalyst to 4.42 × 10⁻² min⁻¹ at 5 mg L⁻¹ CoSAC-NG.

There are several pathways for generating by activating PMS. When is in excess, it can convert to (Dong et al. 2021), by using molybdenum selenide (MoSe₂) instead of molybdenum disulfide, established a visible light-driven process for PMS activation, where is the dominant ROS, and thus, the origin of primarily comes from the conversion of ⁠, as shown in Equation (6). In addition, can also originate from the composition of PMS itself or PMS oxidation. Du et al. (2022) demonstrated that when both FeSA and PMS are present in the system, the intensity of the DMPO- peak significantly increases, confirming the presence of in the FeSA/PMS system. However, no signal of was observed in the standalone FeSA system, while a weaker 2,2,6,6-Tetramethylpiperidine (TEMP)- signal was detected in the standalone PMS system, indicating self-decomposition of PMS generates ⁠: PMS self-decomposes to produce and H⁺, then reacts with to produce (Equation (7)). Wu et al. (2023b) further studied the evolution pathway of ¹O₂ based on transition state theory using Fe–N₄ and Co–N₄ sites. They found that PMS molecules preferentially adsorb to Fe–N₄ or Co–N₄ sites via a single O site on the *SO₄ side, then dissociate into *OH and *SO₄, with *OH interacting with the M-N₄ site (M = Fe or Co) to release H and generate *O intermediate. In the process of *O intermediate transforming into ⁠, two possible evolution paths exist: direct desorption and recombination of *O (Equation (8)), and generation of *OOH intermediate through *O and subsequent dissociation (Equation (9)). Furthermore, oxygen vacancies (O_v), i.e., the electron transfer from the catalyst to the PMS-generated catalyst-PMS complex, can also promote the generation of ⁠. Zhao et al. (2023b) successfully synthesized a novel SAC with single Co atoms anchored on CuO with O_v, Equations (10) and (11) demonstrate that O* generated from O_v can interact with PMS to form ⁠. In addition, in photocatalysis, the generation of h⁺ also facilitates the generation of ⁠, as shown in Equation (12) (Zhao et al. 2022).

In the electron transfer process (ETP), electrons mediated by catalysts migrate from pollutants (electron donors) to PMS (electron acceptors) on the surface of highly conductive catalysts, leading to the degradation of pollutants due to electron loss (Qi et al. 2021). The degradation process of pollutants via ETP primarily involves the binding of oxidants to surface groups of catalysts, forming metastable complexes through electron rearrangement effects. When electrons from electron-rich pollutants are accepted, metastable complexes re-accept electrons and undergo rearrangement, resulting in the oxidation of pollutants (Zheng et al. 2024). As previously mentioned, O_v can also expedite the transfer of electrons from pollutants to catalyst-PMS*. In the Co-CuO (O_v)/PMS system, electrons transfer from the catalyst to PMS, forming catalyst-PMS* complexes on the Co-CuO surface, leading to an increase in the oxidation potential of the catalyst. Upon addition of tetracycline (TC), the catalyst-PMS* complex is consumed by TC, causing a decrease in the surface potential of Co-CuO (O_v), thus demonstrating the degradation of TC through electron transfer (Zhao et al. 2023b). Zhang et al. (2023b) prepared a SAC using straw biochar as a raw material to rapidly degrade tetracycline hydrochloride (TCH) at different temperatures by activating PDS, indicating that when TCH is introduced into the system, TCH electrons transfer to the complex, leading to TCH degradation. Xie et al. (2023) prepared single-atom Co species with five-coordinated N atoms on carbon nanotube (CNT) catalysts (Co–N₅/CNT), which exhibited significant electron transfer capabilities. The electron transfer mechanism involves sulfamerazine (SMZ) pollutant molecules adsorbing onto Co–N₅/CNT, providing electrons to the Co–N₅/CNT catalyst, which continuously migrate from Co–N₅ sites to the C–O bond of the complex, and then to the O–O bond of PMS, ultimately achieving the decomposition of PMS and degradation of SMZ, which is a key factor for the excellent catalytic performance of Co–N₅/CNT catalysts.

It is worth mentioning that direct electron transfer can act in conjunction with radicals on organic pollutants. Li et al. (2021a) synthesized cobalt-based SAC (SACo@NG) with high Co content on N-doped graphene, primarily utilizing the electron transfer pathway to degrade benzyl alcohol (BzOH). On one hand, electron transfer between Co atoms and PMS induces the formation of •OH, ⁠, and radicals, which attack BzOH to form intermediate substances. On the other hand, electrons directly transfer from BzOH to PMS through the highly delocalized π electrons of the graphene framework. Through the combined action of indirect and direct electron transfer, BzOH can be more effectively converted into BzH.

High-valent metal oxo species (HVMOs) are powerful non-radical reactive species generated during the activation of PS (Miao et al. 2023). They generally possess the following advantages: long half-lives, high steady-state concentrations, resistance to nontarget substrate scavenging, and exhibit high selectivity toward recalcitrant pollutants with electron-donating groups, effectively enhancing the removal of pollutants via non-radical pathways (Li et al. 2023a).

Activation of PS by SACs can generate various high-valent metals such as Cu(III) (Li et al. 2022), Co(IV) (Wang et al. 2022a), Fe(IV) (An et al. 2022), etc. Cu(III) can directly oxidize electron-rich compounds without interference from chloride ions, thus selectively oxidizing and degrading micropollutants in water environments. Li et al. (2022) designed a single-atom copper catalyst with unsaturated Cu–N₂ sites (CuSA-NC) for PS activation. Due to the single-atom distribution of Cu (III) and selective oxidation, CuSA-NC exhibits high metal utilization, pollutant degradation selectivity, and resistance to matrix interference. For high-valent cobalt oxo species (Co^IV = O), Wang et al. (2022a) investigated the activation of PMS for BPA degradation using a CoSAC supported on graphitic carbon with Co–N₄ active sites (SA-CoCN). The Co atoms adsorb oxygen atoms from the O–O bond, leading to the formation of Co (IV) = O structure under acidic conditions, which then establishes an oxidation pathway with Co (IV) as the active species, crucial for BPA removal.

Future research should primarily focus on three aspects. First, efforts should aim to mitigate the environmental impact of SACs/PMS systems, enhancing their eco-friendliness. Secondary pollution remains a prevalent and significant concern. Despite the inevitable leaching of metal ions, there are numerous strategies available to reduce this secondary pollution. To mitigate metal leaching, selecting appropriate materials as catalyst supports is crucial. For instance, Yu et al. (2024) effectively controlled the issue of metal leaching in bimetallic catalysts by utilizing polytetrafluoroethylene (PTFE) to immobilize Co and Fe atoms. PTFE's strong viscosity plays a significant role in stabilizing metal centers. Furthermore, selecting low-temperature, high-performance SACs on suitable substrates can counteract increased metal leaching from high-temperature pyrolysis. An et al. (2024) synthesized single-atom Fe catalysts (Fe_SA-NC-500/1000) at 500 and 1,000 °C, with Fe_SA-NC-1000 exhibiting a higher SMX removal rate (96.1 versus 91.2% for Fe_SA-NC-500), yet Fe_SA-NC-500 approached Fe_SA-NC-1000 in overall catalytic performance while reducing metal leaching. The synthesis of catalysts at lower temperatures enhances their environmental sustainability, safety, and energy efficiency, thereby presenting substantial practical value.

Second, enhancing the applicability of SACs and improving the efficiency of EOCs removal are crucial. Strategies may involve combining SACs with other processes or synthesizing superior catalytic materials based on SACs, ensuring compatibility and adaptability with other SACs processes or materials. For instance, coupling with ceramic membranes (CM) can concentrate reactants and shorten transfer paths, enhancing mass transfer in dominant reactions. Yang et al. (2024) prepared CoSAC loaded onto CMs (Co₁-N-doped carbon nanotubes (NCNT)-CM), achieving complete removal of phenol within 15 min without heavy metal leaching, thus addressing metal leaching concerns with SACs. Moreover, Co₁-NCNT-CM/PMS systems maintain CM's high activity and stability, retain water transport characteristics, and suppress water medium interference, facilitating efficient, selective removal of phenols and sulfonamide drugs, expanding possibilities for application in complex water environments. Alternatively, due to the low surface free energy of single atoms facilitating agglomeration, Qin et al. (2024) synthesized catalyst systems coexisting with SACs and nano-clusters, namely Cu_SA/CoO_x–CeO₂, providing more active sites compared to SACs, enhancing adsorption and electron transfer capabilities toward PMS molecules, thus guiding applications of coexisting single-atom and nano-cluster catalytic reactions.

Lastly, SACs/PS research should not be limited to focus on singular organic pollutants. Realistic environments often host multiple high-concentration pollutants. Therefore, future studies should reference coexisting pollutants in actual water bodies, target the degradation of various pollutants, and explore their efficiency and mechanisms. This approach can enhance the practical value of SACs, offering more technical references for resolving real water pollution issues.

To summarize, SACs are distinguished by their uniformly dispersed metal centers, exhibiting high atomic efficiency, excellent catalytic performance, and the ability to activate PS for rapid removal of single EOCs from water. Moreover, their sustainability and stability underscore promising applications. Significant progress has been made in SACs activating PS technique, garnering widespread attention for its efficiency in EOC removal and activation/degradation mechanisms. From an applied perspective, this review critically examines and discusses this technology, highlighting current challenges and future research directions, aiming to advance theoretical foundations and research directions for future applications:

• (1) When selecting SACs synthesis methods, it is essential to consider both catalyst performance and issues of secondary pollution.

• (2) Factors limiting SACs catalytic performance and degradation effectiveness in practical applications are more complex, such as high concentrations of interfering substances in water and SACs sustainability.

• (3) The degradation of EOCs using ROS generated within the system involves a complex process where both radical and non-radical pathways play essential roles.

• (4) Combining SACs with other processes is necessary, along with researching more complex pollutant systems and elucidating their mechanisms.

Z. Q. developed the methodology, rendered support in data curation and formal analysis, and wrote the original draft. Z. Z. developed the methodology and rendered support in data curation and formal analysis. J. L. developed the methodology and rendered support in data curation. J. L. developed the methodology and rendered support in formal analysis. J. W. developed the methodology. X. C. rendered support in data curation, conceptualized the whole article, and developed the methodology. Y. W. wrote the review and edited the article, rendered support in data curation, conceptualized the whole article, and developed the methodology. L. W. wrote the review and edited the article, developed the methodology, supervised the work, rendered support in funding acquisition, and conceptualized the whole article.

This work was sponsored by the Shenzhen Polytechnic Project (6023310038 K and 6022312023K), the National Natural Science Foundation of China (52200167), and the Shenzhen Science and Technology Program (20231128105823001).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.