Asymmetric Zn─N₃O₁ Single‐Atom Sites Promote Hydroxyl Radical Generation for Natural‐Light‐Driven Inactivation of Drug‐Resistant Bacteria

ABSTRACT

Waterborne drug‐resistant bacteria pose a serious global health threat, underscoring the urgent need for highly efficient disinfection strategies. The photocatalytic generation of hydroxyl radicals (•OH) represents a green and powerful route for bacterial inactivation, yet its efficiency is often limited by the sluggish kinetics of the stepwise single‐electron oxygen reduction pathway. Herein, we report an atomic‐level design strategy by anchoring asymmetric Zn─N₃O₁ sites onto ultrathin graphitic carbon nitride nanosheets (Zn₁/OCN) to accelerate •OH production. The introduced Zn─N₃O₁ sites create localized intermediate states that enable rapid trapping of photogenerated electrons at Zn single‐atom sites and prolong their lifetime, thereby driving a stepwise single‐electron oxygen reduction reaction (ORR) for efficient •OH generation. Simultaneously, adjacent C═O moieties act as hole‐trapping centers to drive water oxidation reaction (WOR), establishing a local reservoir of H₂O₂ and protons that couples with the ORR process, thus forming a cooperative redox pathway for enhanced •OH production. Additionally, these asymmetric sites effectively lower the formation energies of *OOH and *OH intermediates, thereby accelerating both ORR and WOR processes and facilitating •OH generation. Consequently, Zn₁/OCN achieves outstanding bactericidal performance, inactivating 99.9% of drug‐resistant bacteria, within 30 min under natural light, markedly outperforming representative photocatalytic antibacterial materials reported to date.

In this work, we present asymmetric Zn─N₃O₁ sites anchored on ultrathin carbon nitride (Zn₁/OCN) to steer a cooperative redox pathway that accelerates both stepwise single‐electron oxygen reduction and water activation for efficient •OH generation, thereby achieving rapid and effective photocatalytic inactivation of drug‐resistant bacteria under natural light.

1. Introduction

Waterborne pathogens pose a serious global threat to public health, causing infectious diseases and compromising the safety of drinking water supplies [1, 2, 3]. Conventional disinfection strategies, such as chlorination and ozonation, are effective but often suffer from critical drawbacks, including secondary pollution, the formation of hazardous by‐products, and limited long‐term sustainability [4, 5, 6]. In this regard, photocatalytic disinfection, particularly when driven by natural sunlight, has emerged as a promising alternative due to its environmentally benign nature, mild operational conditions, and broad‐spectrum antimicrobial activity [7, 8, 9]. Among the reactive oxygen species (ROS) generated during photocatalysis, including superoxide radicals (•O₂ ⁻), hydrogen peroxide (H₂O₂), singlet oxygen (¹O₂), and hydroxyl radicals (•OH), these species exhibit distinct reactivities and roles in antimicrobial processes. Notably, •OH is particularly attractive due to its extremely high oxidative potential, which enables nonselective and efficient microbial inactivation [10, 11, 12]. However, the selective and efficient production of •OH remains challenging, as the high redox potential prevents direct oxidation of H₂O to •OH by photogenerated holes [13, 14, 15]. An alternative strategy involves the H₂O₂‐mediated three‐electron oxygen reduction cascade (O₂ → H₂O₂ → •OH), which provides a viable pathway for sustained •OH generation [16, 17]. However, the efficiency and sustainability of •OH production in this cascade are severely limited by the sluggish kinetics of the required three‐electron transfer process. Overcoming this limitation requires the design of photocatalysts that can not only promote effective charge separation but also optimize the multistep reaction pathway, thereby driving O₂ reduction through a stepwise process for sustained and efficient •OH generation.

Among the diverse photocatalysts, graphitic carbon nitride (g‐C₃N₄) has garnered significant attention as a metal‐free photocatalyst for environmental purification, owing to its suitable band structure, excellent chemical stability, and visible‐light responsiveness [18, 19, 20]. However, bulk g‐C₃N₄ (BCN) suffers from rapid bulk and surface charge recombination as well as the limited exposure of catalytically active sites, which severely restrict its photocatalytic performance (Scheme 1a) [21, 22, 23]. To mitigate these limitations, exfoliated g‐C₃N₄ nanosheets (CN) have been developed, offering suppressed bulk recombination and enhanced exposure of surface‐active sites [24, 25, 26]. Nevertheless, the absence of efficient charge‐trapping centers in CN still leads to pronounced surface recombination and insufficient •OH generation (Scheme 1b) [27, 28, 29]. In recent years, single‐atom photocatalysts (SACs) have attracted increasing attention as a frontier in photocatalysis, owing to their unique electronic configurations, maximized atomic utilization, and capacity to function as highly efficient charge‐trapping and catalytic centers [30, 31, 32, 33, 34]. More importantly, SACs enable precise regulation of the adsorption and conversion of key oxygen reduction intermediates (e.g., *O₂ and *OOH), thereby facilitating the stepwise oxygen reduction reaction (ORR) pathway toward efficient •OH generation. However, most conventional SACs adopt symmetric coordination with a single nonmetal element, such as the M─N₄ motif, where the high symmetry leads to weak internal electric fields and consequently hampers charge separation and transfer [35, 36, 37, 38]. To address this limitation, deliberate modulation of the local coordination environment, specifically the construction of asymmetric coordination structures, has emerged as an effective strategy to fine‐tune the electronic structure of SACs, thereby lowering reaction barriers and enhancing catalytic efficiency and stability [39, 40, 41, 42]. Importantly, compared with electrocatalysis and thermocatalysis, photocatalysis presents additional challenges, including uneven photon absorption, charge accumulation, and the coexistence of multiple types of active sites (e.g., light‐harvesting, charge‐separation, and surface‐reaction sites), all of which give rise to more intricate reaction pathways. Consequently, the development of high‐performance SACs with asymmetric coordination anchored on CN, coupled with a mechanistic understanding of their reaction kinetics, is urgently needed to advance photocatalytic disinfection and enable sustainable environmental remediation.

SCHEME 1.

Schematic illustration of the design rationale and progressive advantages of (a) BCN, (b) CN, and (c) asymmetric single‐atom sites anchored on CN.

Herein, we develop atomically dispersed Zn─N₃O₁ sites with an asymmetric coordination environment anchored on CN (Zn₁/OCN) for efficient photocatalytic water disinfection under natural sunlight. The asymmetric Zn─N₃O₁ coordination introduces localized intermediate states within the bandgap, which promote directional migration of photogenerated electrons to Zn single‐atom sites, thereby enhancing carrier utilization and suppressing surface recombination for efficient •OH generation (Scheme 1c). Mechanistically, Zn single‐atom sites act as electron‐trapping centers, promoting the three‐step single‐electron reduction of O₂ (O₂ → •O₂ ⁻ → H₂O₂ → •OH), ultimately yielding efficient •OH generation. In parallel, adjacent C═O moieties capture photogenerated holes to drive water oxidation into H₂O₂ and protons (H⁺), providing a continuous local source that couples with the ORR process and establishes a cooperative redox pathway for enhanced •OH production. In addition, the asymmetric coordination substantially lowers the kinetic barriers for both O₂‐to‐H₂O₂ and H₂O‐to‐H₂O₂ conversion, while greatly facilitating H₂O₂ activation and decomposition into •OH, thus accelerating •OH generation. As a result, Zn₁/OCN achieves outstanding bactericidal performance, inactivating 99.9% of drug‐resistant bacteria, including Escherichia coli, Staphylococcus aureus, and methicillin‐resistant S. aureus (MRSA), within 30 min under natural light, markedly outperforming BCN, CN, Zn₁/CN (featuring a symmetrical Zn─N₄ coordination environment), and representative photocatalytic antibacterial materials reported to date. This work highlights asymmetric single‐atom engineering as a powerful strategy to regulate charge dynamics and reaction kinetics, offering new design principles for high‐efficiency photocatalytic disinfection in sustainable environmental applications.

2. Results and Discussion

As shown in Figure 1a, Zn₁/OCN was synthesized via a typical two‐step thermal exfoliation process. Atomic force microscopy (AFM) analysis reveals that the average thicknesses of CN and Zn₁/OCN are approximately 4.48 and 3.75 nm, respectively (Figures 1b,c and S1), which are significantly lower than the 42.05 nm observed for BCN (Figure S2). These findings confirm the successful preparation of ultrathin CN and Zn₁/OCN nanosheets through the two‐step thermal exfoliation process. This conclusion is further supported by nitrogen adsorption‐desorption isotherms, which indicate that the specific surface areas of CN and Zn₁/OCN are significantly larger than that of BCN (Figure 1d). Notably, Zn₁/OCN exhibits a further increase in specific surface area compared to CN, which is likely attributed to the introduction of the Zn‐containing precursor (ZIF‐8) during thermal polymerization. The decomposition of ZIF‐8 can act as a sacrificial template and inhibit the restacking of CN nanosheets, resulting in a more porous and loosely packed structure. Transmission electron microscopy (TEM) images provide additional evidence, showing that while BCN exhibits a distinct stacked sheet structure (Figure S3a), CN, Zn₁/CN, and Zn₁/OCN display clear nanosheet morphologies following the two‐step thermal exfoliation (Figures S3b, S4a, and 1e). This observation aligns with the results from BET and AFM analyses. Notably, no Zn metal particles are detected in the TEM images of Zn₁/CN and Zn₁/OCN, suggesting that Zn exists as single‐atom sites within the CN framework. X‐ray diffraction (XRD) analysis further corroborates this conclusion, as no peaks associated with metallic Zn are observed in Zn₁/CN and Zn₁/OCN, compared to BCN and CN (Figure S5). Aberration‐corrected high‐angle annular dark‐field scanning TEM (AC‐HAADF‐STEM) imaging of Zn₁/CN and Zn₁/OCN reveals well‐dispersed bright spots, confirming the presence of uniformly distributed Zn single‐atom sites on the CN nanosheets (Figures 1f and S4b). Elemental mapping analysis further verifies the homogeneous distribution of Zn single‐atom sites within the CN framework (Figures 1g,h and S4c). Notably, compared with Zn₁/CN, Zn₁/OCN exhibits a clearly discernible O elemental distribution, suggesting the incorporation of O atoms into the coordination environment of Zn sites. Inductively coupled plasma optical emission spectroscopy (ICP‐OES) quantifies the Zn content in Zn₁/CN and Zn₁/OCN as 1.157 and 1.403 wt% (Table S1), respectively. Additionally, Fourier‐transform infrared (FTIR) spectroscopy indicates that the incorporation of Zn single‐atom sites does not affect the structural integrity of the CN framework (Figure S6).

FIGURE 1.

(a) Schematic diagram of synthesis Zn₁/OCN. (b) AFM image and (c) height profile of the Zn₁/OCN. (d) Nitrogen adsorption−desorption isotherms and (inset) pore‐size distribution curves of BCN, CN, Zn₁/CN, and Zn₁/OCN. (e) TEM, (f) AC‐HAADF‐STEM, (g) HAADF and (h) the corresponding elemental mapping images of Zn₁/OCN.

X‐ray photoelectron spectroscopy (XPS) was conducted to analyze the elemental composition of the prepared photocatalysts. In the high‐resolution O 1s spectrum (Figure 2a), two characteristic components centered at 532.0 and 533.4 eV are observed and assigned to C─OH and C═O species, respectively [43, 44]. The C─OH component may be associated with surface hydroxyl groups and adsorbed oxygen‐containing species, whereas the appearance of the C═O component suggests successful oxygen incorporation into the carbon nitride framework through partial substitution of lattice nitrogen atoms. In the high‐resolution C 1s spectra (Figure S7a), BCN, CN, Zn₁/CN, and Zn₁/OCN show three distinct peaks at 284.8, 286.3, and 288.0 eV, corresponding to C─C/C═C, C─N, and N═C─N bonding [45, 46], respectively. Similarly, in the high‐resolution N 1s XPS spectra (Figure S7b), three peaks at 398.5, 400.1, and 401.0 eV are observed for BCN, CN, Zn₁/CN, and Zn₁/OCN, assigned to C═N─C, N─(C)₃, and C─NH₂ bonding [47, 48], respectively. Furthermore, Zn₁/CN and Zn₁/OCN exhibit two new peaks at ∼1045.3 and ∼1022.2 eV compared with BCN and CN (Figure S7c), which are assigned to Zn 2p_1/2 and Zn 2p_3/2 [49, 50], respectively, suggesting that Zn single‐atom sites are successfully incorporated into the CN framework. Notably, the Zn 2p peaks of Zn₁/OCN show a slight shift compared to those of Zn₁/CN, indicating a modified electronic structure and an altered local coordination environment of the Zn single‐atom sites induced by oxygen incorporation.

FIGURE 2.

High‐resolution (a) O 1s spectra of BCN, CN, and Zn₁/OCN. (b) Zn K‐edge XANES spectra, and (c) EXAFS spectra of Zn foil, ZnO, ZnPc, Zn₁/CN, and Zn₁/OCN. (d) Zn K‐edge WT EXAFS contour plots of Zn foil, ZnO, and Zn₁/OCN.

X‐ray absorption spectroscopy (XAS) was conducted to probe the differences in the coordination environments of Zn single‐atom sites in Zn₁/CN and Zn₁/OCN photocatalysts. The Zn K‐edge x‐ray absorption near‐edge structure (XANES) spectra reveal that the absorption edges of both Zn₁/CN and Zn₁/OCN are located between those of Zn foil and ZnO references, indicating that the oxidation states of Zn in both samples lie between 0 and +2 (Figure 2b). Notably, compared with Zn₁/CN, the absorption edge of Zn₁/OCN exhibits a slight shift toward lower energy, suggesting a modulated electronic structure and altered chemical environment of the Zn single‐atom sites upon oxygen incorporation. Regarding the local coordination environment, the Fourier‐transformed extended x‐ray absorption fine structure (EXAFS) spectra of both samples show a complete absence of Zn─Zn coordination, unambiguously confirming the atomic dispersion of Zn single‐atom sites within the CN matrices (Figure 2c). Further analysis reveals that the main peak of Zn₁/CN at ∼1.5 Å closely matches that of the ZnPc reference, suggesting a typical Zn─N coordination environment. In contrast, the corresponding peak of Zn₁/OCN shows a discernible shift toward the Zn─O contribution observed in the ZnO reference, providing evidence for the incorporation of O atoms into the first coordination shell of Zn single atoms. These observations are further supported by wavelet transform (WT) analysis and quantitative EXAFS fitting (Figures 2d and S8). While the Zn single atoms in Zn₁/CN are coordinated by four N atoms (Zn─N₄), the best‐fitting results for Zn₁/OCN reveal a Zn─N₃O₁ configuration, indicating the successful incorporation of O atoms into the first coordination shell and the modulation of the local coordination environment around the Zn single‐atom sites (Figure S9 and Table S2).

The antibacterial activity of the prepared photocatalysts was systematically evaluated under natural‐light irradiation against drug‐resistant (MRSA), Gram‐positive (S. aureus), and Gram‐negative (E. coli) bacteria (Figure S10). As shown in Figures 3a–c and S11, a distinct difference in photocatalytic antibacterial activity can be observed between BCN and CN. The exfoliation of bulk BCN into nanosheet CN significantly enhances the inactivation efficiency against MRSA, S. aureus, and E. coli, which can be attributed to the enlarged surface area and improved charge separation efficiency associated with the two‐dimensional structure. The incorporation of Zn single‐atom sites further enhances the antibacterial performance. Specifically, Zn₁/CN with symmetric Zn─N₄ sites exhibits improved bacterial inactivation efficiency compared to pristine CN. However, its activity remains limited, with only partial bacterial reduction achieved after 30 min of illumination. In striking contrast, Zn₁/OCN featuring asymmetric Zn─N₃O₁ sites demonstrates markedly accelerated inactivation kinetics, achieving nearly complete bacterial eradication within 30 min under natural‐light irradiation. This substantial enhancement highlights the critical role of coordination environment engineering, where the transformation from symmetric Zn─N₄ to asymmetric Zn─N₃O₁ effectively modulates the local electronic structure and catalytic activity of Zn single‐atom sites. As a control, negligible bacterial inactivation is observed in the absence of catalysts, confirming that the bactericidal effect originates from the photocatalytic process. To further verify the antibacterial performance under standardized irradiation conditions, additional experiments were conducted under calibrated AM 1.5G simulated solar irradiation. As shown in Figure S12, the antibacterial activity of Zn₁/OCN closely matched that observed under natural‐light conditions, achieving nearly complete inactivation of MRSA, S. aureus, and E. coli within the same reaction timeframe. These finding confirm the excellent photocatalytic disinfection capability of Zn₁/OCN under standardized solar irradiation and further validate the robustness of its photocatalytic antibacterial performance. Furthermore, a comparison with previously reported photocatalysts (Figure 3d and Table S3) demonstrates that Zn₁/OCN achieves superior antibacterial efficiency even at relatively low catalyst concentrations, outperforming representative photocatalysts reported in the literature. Collectively, these findings demonstrate that the introduction of asymmetric Zn─N₃O₁ sites significantly enhances the photocatalytic antibacterial performance of CN, rendering Zn₁/OCN a promising candidate for broad‐spectrum and highly efficient antimicrobial applications.

FIGURE 3.

Antibacterial performance against (a) MRSA, (b) S. aureus, and (c) E. coli under different irradiation times (n = 5, ***p < 0.001). (d) Comparison of the photocatalytic antibacterial efficiencies of various photocatalysts. (e) Recyclability test of Zn₁/OCN against MRSA (n = 5, ***p < 0.001). (f) XRD patterns of Zn₁/OCN before and after five photocatalytic cycles under illumination. (g) Bacterial survival rates of MRSA on the designed photocatalysts under illumination and in the dark (n = 5, ***p < 0.001). (h) MRSA inactivation by Zn₁/OCN in the presence of different scavengers (h⁺, •OH, and •O₂ ⁻) (n = 5, **p < 0.01).

Importantly, Zn₁/OCN retains its bactericidal performance over five consecutive cycles without any detectable loss of activity (Figures 3e, S13, and S14). This excellent durability is further corroborated by XRD patterns (Figure 3f), XPS spectra (Figure S15), and HRTEM images (Figure S16), all of which reveal no discernible structural degradation after repeated cycling. Furthermore, ICP‐OES analysis reveals minimal Zn²⁺ leaching (only 0.037 wt%) after the photocatalytic reaction (Table S4), confirming the high stability of the asymmetric Zn─N₃O₁ sites. Moreover, neither a Zn²⁺ solution with a concentration equivalent to the measured Zn leaching level nor the filtrate collected after the photocatalytic reaction exhibits observable antibacterial activity (Figure S17). These results indicate that dissolved Zn²⁺ species contribute negligibly to bacterial inactivation, confirming that the enhanced antibacterial performance primarily originates from photocatalytically generated ROS. To further elucidate the antibacterial mechanism, the photocatalytic performance under dark conditions was carefully evaluated. As shown in Figures 3g and S18, Zn₁/OCN exhibits pronounced antibacterial activity only under illumination, whereas negligible bacterial inactivation is observed in the dark. This distinct contrast demonstrates that photoexcitation is indispensable for activating the reactive sites of Zn₁/OCN and initiating the generation of bactericidal ROS, thereby establishing light irradiation as the essential driving force for the observed antimicrobial activity. Further mechanistic insight is provided by radical scavenger experiments (Figure 3h), which reveal that •OH plays the predominant role in bacterial inactivation, while •O₂ ^‒ and hole contribute to a lesser extent. The dominant role of •OH is attributed to its exceptionally high oxidation potential, which enables rapid oxidative damage to bacterial cell membranes, leading to membrane disruption, structural collapse, and leakage of intracellular components, as evidenced by SEM observations (Figure S19). Collectively, these results demonstrate that the superior antibacterial performance of Zn₁/OCN originates from the efficient photoinduced generation and utilization of •OH radicals.

The optical and electronic properties of the photocatalysts were systematically investigated to elucidate the origin of their enhanced antibacterial activity. As shown in Figure 4a, compared with BCN, CN, and Zn₁/CN, Zn₁/OCN exhibits comparable visible‐light absorption, but displays a more pronounced absorption tail extending into the 400–600 nm region. The corresponding Tauc plots (Figure S20) reveal a gradual narrowing of the bandgap from CN (2.73 eV) to Zn₁/CN (2.64 eV) and further to Zn₁/OCN (2.63 eV). The red‐shifted absorption edges and enhanced sub‐bandgap absorption suggest the formation of localized mid‐gap states. Notably, these states are significantly more pronounced in Zn₁/OCN, highlighting the critical role of asymmetric Zn─N₃O₁ coordination in modulating the electronic structure. These states not only broaden the photoresponse but also facilitate charge‐carrier separation and migration by serving as effective electronic bridges for electron‐hole transport [51]. The semiconductor characteristics were further probed by Mott–Schottky analysis (Figure S21). All photocatalysts show positive slopes, confirming their n‐type nature [52]. The flat‐band potentials of BCN, CN, Zn₁/CN, and Zn₁/OCN are determined to be −1.03, −1.00, −0.98, and −0.95 V versus Ag/AgCl (pH = 6.8), respectively. Based on these values and the measured bandgaps, the conduction and valence band positions are calculated to be −0.83/1.90 V for BCN, −0.80/1.93 V for CN, −0.78/1.86 V for Zn₁/CN, and −0.75/1.88 V for Zn₁/OCN (Figure 4b). Importantly, the band structures of all three photocatalysts are thermodynamically favorable for •OH formation via the conversion of H₂O₂ to •OH, which represents the primary reactive species driving photocatalytic disinfection.

FIGURE 4.

(a) UV‐vis absorption spectra and (b) band gap structure of BCN, CN, Zn₁/CN, and Zn₁/OCN. In situ XPS spectra of (c) Zn 2p and (d) O 1s for Zn₁/OCN under dark and light conditions (λ = 405 nm). 2D mapping TA spectra of (e) CN and (f) Zn₁/OCN using a 355 nm pump pulse. (g) TA kinetics probed at 410 nm within 200 ps. (h) Schematic illustration of the decay pathways of photogenerated electrons in CN and Zn₁/OCN.

The charge separation and transfer behaviors were investigated to substantiate the above electronic structure analysis. As shown in Figure S22a, CN nanosheets exhibit a markedly higher transient photocurrent response than BCN, underscoring the beneficial effect of exfoliation in enhancing charge generation and migration. This enhancement arises from the shortened charge‐transport pathways inherent to the two‐dimensional structure, which reduce bulk recombination and facilitate more efficient carrier separation. However, due to the absence of effective charge‐trapping and separation centers on the surface of pristine CN, the photocurrent density remains moderate, indicating that surface recombination still limits the overall performance. With the introduction of asymmetric Zn─N₃O₁ sites, Zn₁/OCN exhibits a markedly enhanced photocurrent response compared with both pristine CN and Zn₁/CN, the latter featuring symmetric Zn─N₄ coordination. This result highlights that the asymmetric Zn─N₃O₁ sites effectively promotes charge‐carrier generation and separation than its symmetric counterpart. Consistently, electrochemical impedance spectroscopy (EIS, Figure S22b) reveals a progressive decrease in the Nyquist semicircle radius from BCN to CN, then to Zn₁/CN, and finally to Zn₁/OCN, with the latter displaying the lowest interfacial charge‐transfer resistance. This trend suggests that the asymmetric Zn─N₃O₁ coordination in Zn₁/OCN gives rise to more effective localized mid‐gap states compared with the symmetric Zn─N₄ configuration. These states not only extend the photoresponse but also act as efficient electron‐trapping centers, thereby facilitating interfacial charge transfer.

Both steady‐state photoluminescence (PL, Figure S23a) and time‐resolved PL (TRPL, Figure S23b) measurements further corroborate the above observations. Compared with BCN, CN exhibits reduced emission intensity accompanied by a shortened carrier lifetime, indicative of improved charge separation arising from the nanosheet structure. More importantly, Zn₁/OCN, featuring asymmetric Zn─N₃O₁ sites, exhibits the lowest PL emission intensity and the shortest carrier lifetime among all samples, outperforming Zn₁/CN. These results demonstrate that the asymmetric Zn─N₃O₁ sites effectively suppress radiative recombination and significantly accelerates charge separation and transfer kinetics. Collectively, these findings highlight the synergistic effects of the nanosheet morphology and the asymmetric Zn─N₃O₁ sites in minimizing charge‐carrier recombination and maximizing charge utilization efficiency.

To elucidate the role of asymmetric Zn─N₃O₁ sites in carrier trapping and charge transfer, in situ XPS measurements were performed (Figure 4c,d). Upon illumination, the Zn 2p binding energy undergoes a distinct negative shift, while the O 1s component associated with C═O displays a concomitant positive shift, collectively demonstrating electron transfer from the CN framework to the Zn single‐atom sites. These photoinduced spectral shifts provide direct evidence for spatial charge separation, with electrons preferentially accumulating at the Zn sites and holes accumulating at the C═O moieties. Such charge redistribution is expected to facilitate reduction reactions at Zn sites and oxidation reactions at the C═O moieties. This complementary functionality underscores the important role of the asymmetric Zn─N₃O₁ coordination environment in facilitating charge separation and promoting efficient photocatalytic redox reactions.

To further clarify how asymmetric Zn─N₃O₁ sites regulate the photogenerated carrier dynamics in CN, femtosecond transient absorption (fs‐TA) spectroscopy was conducted. As shown in the two‐dimensional TA spectra (Figure 4e,f), excitation with a 355 nm pump pulse produces a substantially stronger ground‐state bleaching (GSB) signal in Zn₁/OCN than in CN. The enhanced bleach intensity reflects a higher population of photoexcited electrons occupying the conduction band and adjacent states, indicating that Zn─N₃O₁ sites markedly improve light harvesting and electron excitation under identical irradiation conditions. Normalized ΔA spectra further highlight the kinetic divergence between the two photocatalysts (Figure S24). Zn₁/OCN exhibits a significantly slower recovery of the GSB signal, which demonstrates a prolonged lifetime of photogenerated electrons and a reduced recombination tendency. Kinetic traces monitored at 410 nm were fitted using a bi‐exponential model. The fast component (τ₁) describes the ultrafast trapping of photoexcited electrons into shallow intermediate states, whereas the slow component (τ₂) represents carrier recombination on longer timescales. As shown in Figure 4g, CN exhibits τ₁ = 1.56 and τ₂ = 3.67 ps, corresponding to the trapping of photoexcited electrons into shallow states, followed by charge recombination on the picosecond timescale. In contrast, Zn₁/OCN displays a reduced τ₁ of 0.83 ps, indicating that the introduced Zn─N₃O₁ sites act as efficient electron‐capture centers, enabling the rapid transfer of photoexcited electrons into intermediate states and facilitating effective charge separation. More importantly, τ₂ increases to 10.69 ps, which is nearly three times longer than that of CN (Figure 4h). This prolonged recombination time indicates that the Zn─N₃O₁ induced intermediate states effectively capture and stabilize photogenerated electrons and substantially suppress electron‐hole recombination, thereby extending the electron lifetime. Such stabilization of photogenerated electrons ensures a persistent electron reservoir that continuously drives the multi‐step oxygen‐reduction cascade, ultimately enabling significantly enhanced •OH generation.

Benefiting from efficient charge separation and transfer, Zn₁/OCN exhibits the strongest •OH signal, as detected by electron paramagnetic resonance (EPR) spectroscopy under light irradiation, with intensities 1.63, 2.82, and 6.27 times higher than those of Zn₁/CN, CN, and BCN, respectively (Figure 5a). This result demonstrates that the introduction of asymmetric Zn─N₃O₁ sites significantly enhances •OH generation. A similar trend is observed in the quantified •OH production (Figure S25), further confirming the superior •OH generation capability of Zn₁/OCN. Furthermore, comparison with representative photocatalytic systems under comparable conditions (Table S5) reveals that Zn₁/OCN exhibits outstanding •OH generation performance, further highlighting the advantage of asymmetric Zn─N₃O₁ coordination for efficient •OH generation. Notably, •OH radicals can still be detected under an Ar atmosphere, indicating that their formation is not solely dependent on dissolved O₂ but may also involve alternative pathways, such as the water oxidation reaction (WOR). This observation prompts further investigation into the origin and generation pathway of •OH in this system. Band structure analysis indicates that •OH primarily originates from the further decomposition of H₂O₂ (Figure 4b); therefore, elucidating the H₂O₂ generation process is essential for understanding the origin of •OH in this system. As shown in Figure 5b, Zn₁/OCN delivers a much higher H₂O₂ yield than Zn₁/CN, CN, and BCN, suggesting that photogenerated electrons at Zn single‐atom sites within the asymmetric Zn─N₃O₁ coordination environment efficiently drive the reduction of O₂ to H₂O₂. Remarkably, Zn₁/OCN still maintains the highest H₂O₂ productivity under an Ar atmosphere. Combined with in situ XPS results, this phenomenon can be attributed to the participation of photogenerated holes at C─O sites in the WOR, thereby providing an additional H₂O₂ source and further promoting •OH accumulation. To exclude the possible influence of H₂O₂ photolysis or decomposition on the measured H₂O₂ yield, additional control experiments were conducted using externally added H₂O₂. As shown in Figure S26, only a slight decrease in H₂O₂ concentration is observed after 60 min of light irradiation, indicating negligible photolysis under the reaction conditions. In contrast, the H₂O₂ concentration remains essentially unchanged in the presence of Zn₁/OCN under dark conditions, excluding significant catalyst‐induced decomposition. Therefore, the measured H₂O₂ concentration primarily reflects photocatalytic H₂O₂ generation rather than losses caused by photolysis or catalyst‐induced decomposition, confirming the reliability of the H₂O₂ production results. These results further support the role of H₂O₂ as a key intermediate in the subsequent •OH generation process.

FIGURE 5.

(a) EPR spectra of •OH signals and (b) H₂O₂ production for BCN, CN, Zn₁/CN, and Zn₁/OCN (n = 5, ***p < 0.001). In situ FTIR spectra of Zn₁/OCN under different irradiation time after purging the cell with (c) O₂ and (d) Ar atmosphere. Gibbs free energy profiles for (e) O₂ reduction to H₂O₂, (f) H₂O oxidation to H₂O₂, and (g) H₂O₂ dissociation to •OH on CN, Zn₁/CN, and Zn₁/OCN surfaces. (h) Schematic illustration of the mechanism of •OH formation on Zn₁/OCN.

To further elucidate the regulatory role of asymmetric Zn─N₃O₁ sites in the ORR pathway, rotating disk electrode (RDE) measurements were conducted (Figures S27 and S28). As no significant difference is observed between dark and light conditions, the data obtained under light irradiation are presented for discussion. The Koutecky–Levich plots at −1.1 V (vs. Ag/AgCl) indicate that Zn₁/OCN exhibits an apparent electron‐transfer number close to 1 (1.06), which is higher than those of Zn₁/CN (0.84), CN (0.81), and BCN (0.63). To gain deeper insight into the ORR mechanism, rotating ring‐disk electrode (RRDE) measurements were further performed. As shown in Figure S29, Zn₁/OCN exhibits enhanced H₂O₂ selectivity (87.13%) and yield compared with Zn₁/CN (76.92%), CN (66.92%), and BCN (51.81%), as evidenced by the increased ring current and the calculated H₂O₂ production efficiency. The corresponding electron‐transfer number remains close to one, consistent with the RDE results, further supporting a predominant one‐electron ORR pathway. These results indicate that the asymmetric Zn─N₃O₁ sites favor a stepwise reduction of O₂, facilitating the formation of •O₂ ⁻ intermediates (Figure S30) and subsequent H₂O₂ accumulation. To further verify the formation dynamics of •O₂ ⁻, time‐resolved EPR measurements were performed on Zn₁/OCN under light irradiation. As shown in Figure S31, the •O₂ ⁻ signal appears rapidly and intensifies with irradiation time, providing direct evidence for the efficient generation and accumulation of •O₂ ⁻ intermediates. Collectively, the combined RDE, RRDE, and EPR results support a stepwise reduction pathway (O₂ → •O₂ ⁻ → H₂O₂), which provides a solid foundation for efficient •OH generation.

To elucidate the evolution of ORR and WOR intermediates, in situ FTIR spectroscopy was carried out on Zn₁/OCN under different atmospheres. Under an O₂ atmosphere (Figure 5c), the absorption bands at approximately 1117 and 1245 cm⁻¹ gradually intensify with prolonged light irradiation, which are assigned to adsorbed superoxide (*O₂ ⁻) and hydroperoxyl (*OOH) intermediates, respectively [53]. Meanwhile, a progressively emerging signal at around 1417 cm⁻¹ is attributed to adsorbed *H₂O₂ species [54]. The concerted growth of these characteristic peaks clearly demonstrates that O₂ is effectively activated at the asymmetric Zn─N₃O₁ sites and proceeds via a stepwise ORR pathway following the sequence O₂ → *O₂ ⁻ → *OOH → *H₂O₂, confirming that Zn₁/OCN can efficiently promote a selective oxygen reduction process. Simultaneously, the absorption bands at approximately 1658 and 3300 cm⁻¹ also increase with irradiation time, which are assigned to the vibrational modes of adsorbed *H₂O and *OH species, respectively [55]. These observations indicate that, in parallel with ORR, a light‐driven water activation and oxidation process also occurs on the catalyst surface, suggesting the involvement of a WOR.

To further verify the formation of key ORR intermediates, in situ Raman spectroscopy was performed. As shown in Figure S32, Zn₁/OCN exhibits a distinct band centered at approximately 855 cm⁻¹ under light irradiation, which is assigned to the Zn─OOH intermediate, whereas this feature is absent in pristine CN. This observation indicates that the introduction of Zn single‐atom sites enables the stabilization of *OOH intermediates. Moreover, the emergence and progressive intensification of this band with increasing irradiation time further confirm the dynamic formation of *OOH species during the photocatalytic process. These results provide direct spectroscopic evidence for the involvement of *OOH intermediates on Zn₁/OCN, in good agreement with the in situ FTIR results, and demonstrate that Zn single‐atom sites in the asymmetric Zn─N₃O₁ coordination environment serve as active centers for stabilizing *OOH intermediates and promoting the ORR process.

To verify the reactivity of water in this system, comparative in situ FTIR measurements were carried out under an Ar atmosphere (Figure 5d). Notably, even in the absence of O₂, distinct characteristic absorption bands corresponding to *H₂O₂, *H₂O, and *OH species are clearly observed, and their intensities gradually increase with light irradiation. These observations demonstrate that water molecules can be directly activated at the C═O sites and progressively converted into H₂O₂ via the pathway of H₂O → *OH + H⁺ → *H₂O₂, providing direct evidence for the occurrence of the WOR process [56]. More importantly, the H⁺ generated during water activation can further participate in the ORR occurring at the Zn single‐atom sites, facilitating the formation and transformation of *O₂ ⁻ and *OOH intermediates. These results reveal that Zn₁/OCN not only drives the reduction of O₂ to H₂O₂ at the Zn single‐atom sites but also promotes the oxidation of H₂O to generate additional H₂O₂ at the C═O sites. Therefore, a coupled WOR‐ORR dual‐channel mechanism is established, which synergistically supplies H₂O₂ and significantly enhances the overall generation efficiency of •OH radicals.

To further verify the direct WOR pathway, isotope‐labeled H₂ ¹⁸O was employed as the reactant instead of H₂O. Meanwhile, the ORR pathway was effectively suppressed by introducing NaIO₃ as an electron scavenger and replacing O₂ with N₂. Notably, H₂O can be oxidized to either O₂ or H₂O₂; therefore, both ¹⁸O₂ and H₂ ¹⁸O₂ are possible products and should be distinguished accordingly. As shown in Figure S33a, gas chromatography‐mass spectrometry (GC‐MS) analysis of the effluent gases after illumination reveals no detectable ¹⁸O₂ signal, indicating that H₂ ¹⁸O is not directly oxidized to ¹⁸O₂ over Zn₁/OCN. Interestingly, upon the addition of MnO₂ into the reactor followed by sonication for 5 min, distinct signals corresponding to ¹⁸O₂ emerge in the mass spectra (Figure S33b). Given that MnO₂ is known to catalyze the decomposition of H₂O₂ into H₂O and O₂, the appearance of the ¹⁸O₂ signal provides compelling evidence that H₂ ¹⁸O₂ is generated via the photooxidation of H₂ ¹⁸O.

Through density functional theory (DFT) calculations, the generation mechanism of •OH radicals is further elucidated at the atomic scale. Adsorption energy analysis reveals that Zn₁/CN exhibits stronger adsorption toward both O₂ and H₂O molecules than pristine CN, indicating that the introduction of Zn─N₄ coordination enhances interactions with reactant molecules (Figure S34). Notably, Zn₁/OCN further strengthens adsorption compared with Zn₁/CN. Specifically, O₂ preferentially interacts with Zn single‐atom sites, whereas H₂O adsorption is mainly associated with the C═O moieties in the carbon framework. This coordination‐dependent adsorption behavior highlights the critical role of asymmetric Zn─N₃O₁ sites in optimizing surface adsorption properties and providing stable precursor states for the coupled ORR and WOR processes.

Gibbs free energy calculations further reveal the kinetic regulation induced by asymmetric Zn─N₃O₁ sites along the ORR pathway (Figure 5e). Pristine CN encounters a considerable energy barrier (+0.22 eV) during the *OOH formation step (*O₂ + H⁺ + e⁻ → *OOH), which limits H₂O₂ generation. Upon the introduction of symmetric Zn─N₄ sites, Zn₁/CN exhibits a markedly reduced energy barrier (−0.06 eV), indicating that symmetric Zn─N₄ sites facilitate O₂ activation and render *OOH formation thermodynamically favorable. Notably, this step becomes even more favorable on Zn₁/OCN (−0.13 eV), demonstrating that further coordination modulation to asymmetric Zn─N₃O₁ significantly optimizes reaction energetics. This behavior can be attributed to the asymmetric Zn─N₃O₁ coordination environment in Zn₁/OCN, which strengthens the interaction between *O₂ and the catalyst surface, stabilizes the *OOH intermediate, and lowers the energy barrier for its formation. Meanwhile, the locally increased H⁺ availability, supplied by the adjacent WOR process, may further facilitate proton‐coupled electron transfer from *O₂ to *OOH, thereby contributing to the reduced energy barrier and accelerated H₂O₂ formation.

Along the WOR pathway, DFT calculations reveal that *OH formation on C═O sites is significantly more favorable than on Zn sites in Zn₁/OCN (Figure S35), indicating that water oxidation preferentially occurs on the carbon framework rather than at the Zn centers. Based on this finding, the evolution of reaction energetics across different samples was further analyzed (Figure 5f). Pristine CN requires overcoming a relatively high energy barrier (+0.71 eV) for *OH formation. This barrier is reduced upon the introduction of Zn─N₄ sites (+0.20 eV) and becomes thermodynamically favorable on Zn₁/OCN (−0.20 eV), demonstrating the strong promotional effect of asymmetric Zn─N₃O₁ coordination. This enhancement is closely associated with the presence of C═O groups, which act as hole‐accumulation centers to drive water oxidation. Compared with C═N sites in Zn₁/CN, C═O moieties exhibit superior activity toward H₂O activation and H⁺ generation. The synergistic interplay between asymmetric Zn─N₃O₁ sites and C═O groups not only lowers the kinetic barrier for *OH formation but also provides additional protons, thereby sustaining the coupled ORR process and reinforcing the overall ORR‐WOR synergistic cycle.

Furthermore, regarding the subsequent activation of H₂O₂ (Figure 5g), Zn₁/CN exhibits a reduced energy barrier (+1.15 eV) for H₂O₂ decomposition to •OH compared with pristine CN (+2.92 eV), indicating that the introduction of symmetric Zn─N₄ sites facilitates H₂O₂ activation. Notably, this barrier is further decreased to +0.66 eV on Zn₁/OCN, demonstrating that asymmetric Zn─N₃O₁ sites significantly promote this process. These results highlight that asymmetric Zn─N₃O₁ sites play a crucial role in facilitating H₂O₂ activation by constructing a low‐barrier pathway for rapid •OH release. Such synergistic thermodynamic and kinetic optimization accounts for the markedly enhanced •OH yield observed experimentally for Zn₁/OCN.

Based on the experimental results and theoretical calculations, Figure 5h illustrates the proposed mechanism for •OH radical generation on Zn₁/OCN. Under illumination, CN nanosheets facilitate efficient charge separation, while the introduced asymmetric Zn─N₃O₁ sites generate localized intermediate states within the bandgap, providing effective electron‐storage centers. This enables efficient migration of photogenerated electrons to Zn single‐atom sites, where they participate in subsequent surface reduction reactions, thereby enhancing charge utilization and suppressing surface recombination to drive •OH generation. Specifically, photogenerated electrons are transferred to Zn single‐atom sites, where they participate in a predominantly one‐electron ORR pathway, promoting the conversion of O₂ → *O₂ ⁻ + H⁺ → *OOH → *H₂O₂ and ultimately contributing to •OH generation. Meanwhile, photogenerated holes preferentially accumulate at C═O sites, promoting the oxidation of H₂O on the catalyst surface (2H₂O → H₂O₂ + 2H⁺ + 2e⁻). The enhanced water oxidation process not only increases the overall H₂O₂ yield but also supplies additional H⁺ for the coupled ORR pathway, thereby facilitating the proton‐coupled electron transfer from *O₂ to *OOH and further promoting H₂O₂ production. Overall, Zn₁/OCN optimizes the utilization of photogenerated charge carriers and reaction pathways, significantly boosting •OH radical production and exhibiting outstanding photocatalytic antibacterial performance.

3. Conclusion

In summary, we developed a SAC (Zn₁/OCN) featuring atomically dispersed Zn─N₃O₁ asymmetric coordination sites anchored on ultrathin CN, which achieves highly efficient water disinfection under natural sunlight. The asymmetric Zn─N₃O₁ sites generate localized intermediate states that facilitate rapid electron capture at Zn single‐atom sites, enabling a predominant one‐electron ORR pathway for efficient •OH generation. Simultaneously, adjacent C═O moieties act as hole‐trapping centers to promote the WOR, thereby sustaining continuous H₂O₂ and proton generation and establishing a cooperative redox pathway that further enhances •OH production. Moreover, the asymmetric coordination environment effectively lowers the formation energies of key *OOH and *OH intermediates, thereby accelerating the ORR and WOR processes. Consequently, Zn₁/OCN exhibits outstanding photocatalytic bactericidal performance, achieving 99.9% inactivation of drug‐resistant bacteria, including MRSA, S. aureus, and E. coli, within 30 min, significantly outperforming BCN, CN, Zn₁/CN, and representative photocatalytic antibacterial materials reported in the literature. This study not only demonstrates the effectiveness of asymmetric single‐atom engineering in modulating charge dynamics and surface reaction kinetics but also provides a generalizable strategy for the design of advanced photocatalysts for solar‐driven environmental applications.

4. Experimental Section

4.1. Preparation of Catalysts

Preparation of ZIF‐8: 0.513 g of Zn(NO₃)₂·6H₂O was dissolved in 15 mL of methanol to form solution A. Similarly, 0.577 g of 2‐methylimidazole was dissolved in 15 mL of methanol to form solution B. Solution A was then added to solution B under ultrasonic treatment for 15 min to obtain a homogeneous mixture (solution C). The resulting solution was transferred into a Teflon‐lined autoclave and heated at 120°C for 4 h. The precipitate was collected by centrifugation and washed three times with ethanol, followed by drying in a vacuum oven at 60°C overnight to obtain ZIF‐8 crystals.

Preparation of BCN: 2 g of melamine (MA) was pyrolyzed at 500°C for 4 h in air to obtain BCN.

Preparation of CN: CN was obtained by further pyrolyzing the as‐prepared BCN at 500°C for an additional 4 h in air.

Preparation of Zn₁/OCN: A mixture of 2 g of MA and 0.04 g of ZIF‐8 was thoroughly ground and then pyrolyzed at 500°C for 4 h in air to obtain the intermediate precursor. The obtained precursor was subsequently subjected to a second pyrolysis at 500°C for another 4 h in air to yield Zn₁/OCN.

Preparation of Zn₁/CN: Zn₁/CN was synthesized following the same procedure as that for Zn₁/OCN, except that all pyrolysis processes were conducted under an Ar atmosphere instead of air.

4.2. Antibacterial Performance Evaluation

Methicillin‐resistant S. aureus (MRSA, ATCC 43300), E. coli (DH5α), and S. aureus (ATCC 6538) were selected as representative models of drug‐resistant, Gram‐negative, and Gram‐positive bacterial strains, respectively. The bacteria were first cultured to the exponential growth phase and then collected by centrifugation to obtain bacterial pellets, which were subsequently resuspended and diluted with phosphate‐buffered saline (PBS) to a concentration of approximately ∼1.0 × 10⁷ CFU mL⁻¹. Subsequently, the photocatalyst (BCN, CN, Zn₁/CN, or Zn₁/OCN) was dispersed into the bacterial suspension at a concentration of 0.5 g L⁻¹, followed by exposure to simulated solar illumination. At designated time intervals, aliquots (1 µL) were withdrawn, serially diluted with PBS, and 100 µL of the diluted suspension was spread onto Luria‐Bertani (LB) agar plates. The plates were incubated at 37°C for 24 h, after which the number of bacterial colonies was counted. The antibacterial efficiency was evaluated based on the colony counts. The bacterial survival rate was calculated using Equation (1):

$$\mathrm{Survival}(\%)=\mathrm{C}/\mathrm{C}0\times 100\%$$ (1)

where C is the terminal concentration of bacteria and C₀ is the concentration at t = 0 min of the experiments.

4.3. Statistical Analysis

All quantitative data are presented as mean ± standard deviation (SD, n = 5 independent experiments). Prior to statistical analysis, the raw data were inspected for outliers, and no data points were excluded. The assumptions for parametric tests were verified by assessing the normality of the data distribution and the homogeneity of variance using the Shapiro‐Wilk test and Levene's test, respectively. No data transformation or normalization was applied unless otherwise specified. Statistical significance between two groups was evaluated using a two‐tailed unpaired Student's t‐test. For comparisons among multiple groups, one‐way analysis of variance (ANOVA) followed by Tukey's post hoc test was employed. A p value < 0.05 was considered statistically significant (*p < 0.05, **p < 0.01, ***p < 0.001). All statistical analyses were performed using IBM SPSS Statistics 26.0.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adma74222‐sup‐0001‐SuppMat.docx.

Data Availability Statement

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